HANDBOOK OF GEOPHYSICS

AND THE
SPACE ENVIRONMENT

SCIENTIFIC EDITOR

Adolph S. Jursa

AIR FORCE GEOPHYSICS LABORATORY

AIR FORCE GEOPHYSICS LABORATORY

AIR FORCE SYSTEMS COMMAND
UNITED STATES AIR FORCE
1985
Additional copies of this dwument can be obtained from
National Techtical Itiorrnation Service
5~5 Port Royal Road
SpringReld, VA 22161

Document Acc~on Number: ADA 167000

 ASSISTANT EDITOR
Marylou Tschirch

CONTRIBUTING AUTHORS

Chapter Chapter Chapter

Aarons, J. 10 Gallery, W.O. 18 Millman, G.H. 10
Altrock, R.C. 1 Garrett, H.B. 7 Muench, H. S.’ “21
Anderson, G.P. 21 Glass, M. 16 Murdoek, T.L. 25
Barnes, A.A. 16 Good, R.E. 18, 21 Neidig, D.F. 1
Barron, W.R. 11 Grantham, D.D. 16, 17 Novak, J.M. Capt. USAF 23
Basu, S. 10 Gtingorten, 1.1. 15, 16, 17 O’Neil, R.R. 12
Basu, Su 9 Guidice, D.A. 11 Philbrick, C.R. 21
Battis, J.C. 23 Hall, L.A. 2 Picard, R. H. 12
Bertoni, E.A. 16 Hardy, D.A. 8 Price, S.D. 25
Brown, H.A. 16 Hardy, K.R. 16 Radick, R.R. 1
Buchau, J. 10 Hecksher, J.L. 10 Rich, F,J. 9, 10
Burke, H.K. 20 Heroux, L.J. 2 Rothman, L.S. 18
Burke, W.J. 8 Hinteregger, H.E. 2 Rothwell, P.L. 5
Cabaniss, G.H. 23 Huffman, R.E. 22 Sagalyn, R.C. 20
Champion, K.S. W. 14 Izumi, Y. 15, 16 Sharma, R.D. 13
Cipar, J.J. 23 Jekeli, C. 24 Shea, M.A. 6
Cliver, E. W. 10, 11 Johnston, J.C. 23 Shearer, J.A. 23
Clough, S.A. 18 Kantor, A.J. 14, 15, 16, 17 Shettle, E.P. 18
Cohen, I.D. 16 Keil, S.L. Capt. USAF 1 Simon, G.W. I
Cole, A.E. 14, 15, 17 Klobuchar, J.A. 10 Smart, D.F. 6
Cronin, J.P. 11 Knecht, D.J. 4 Spjeldvik, W. N. 5
Dandekar, B.S. 10 Kneizys, F.X. 18 Shuman, B.M. 4
DeMastus, H.L. 1 Kossey, P.A. 10 Swider, W. 21
Dyer, R.M. 16, 19 Kunkel, B.A. 16 Tattelman, P. 15, 16, 17
Evans, J.W. 1 Lewis, E.A. 10 Vancour, R.P. 8
Falcone, V.J., Jr. 19 Marcos, F.A. 17 VOIZ,F.E. 8
Fenn, R.W. 18 McNamara, L.F. 10 Weber, E.J. 0
Feynman, J. 3 Mendillo, M.F, 10 Whalen, J .A. 2
Field, E.C. 10 Metcalf, J.I. 16
Fitzgerald, D.R. 20 Mill, J.D. Lt. Col. USAF 18

. ..
111
 FOREWORD

The space, atmospheric, and terrestrial environments influence the functioning of all Air Force systems. As technology
advances, the role of the environment becomes more important to system performance. In many cases, the environment
determines the limit in technical capability. This Handbook reflects the world of geophysics as honed and shaped by the
special needs of the Air Force.
Today’s operational systems were designed with the environmental knowledge made available by the Air Force Geophysics
Laboratory and its predecessor, the Air Force Cambridge Research Laboratories. Tomorrow’s Air Force depends on the
quality and breadth of today’s research as reflected in this Handbook. Although written primarily by Air Force Geophysics
Laboratory scientists, this work reflects the state-of-knowledge of thousands of collaborative investigators, both U.S. and
foreign. It is a mixture of basic research and exploratory development, and as such, represents the entire range of our
technology based efforts. The very breadth of Air Force operations from space, through the atmosphere, to the earth makes
it challenging, exciting, and rewarding to do front line research across this broad spectrum and to present a useable summary
of the results to a wide community via this publication. This Handbook is one of our deliverable end products.
The recognition by the Air Force of the value of geophysics research to its mission is clearly illustrated in this work.
The advances made by in-house scientists, the unique world class facilities created to do this work, and the excellent support
of space based experiments, all attest to the long term dedication required to advance our understanding of the environment.
Many significant players, military and civilian, have come and gone since the last issue of this Handbook. It is in the context
of pride in these people and pride in our accomplishments for our country that we take great pleasure in dedicating this
book to the twenty-fifth anniversary of the International Geophysical Year.

J. R. JOHNSON, Colonel, USAF A. T. STAIR, JR.

Commander Chief Scientist

v
 PREFACE

This fourth edition of the Air Force Handbook of Geophysics and the Space Environment has been completely revised.
It was conceived as a commemorative issue in recognition of the twenty-fifth anniversary of the International Geophysical
Year. This was particularly appropriate since the Air Force Geophysics Laboratory-at that time the Air Force Cambridge
Research Center—had been an active participant in the IGY in 1957–58 and the first edition of the Handbook had been
published in 1957. The complex task of preparing this edition involved several years of effort during which time the twenty-
fifth annivers~ of the IGY occurred.
The purpose of the Handbook remains the same, that is, to provide Air Force designers, engineers, and systems operators
with facts and data about the environment in which the Air Force operates. Extraordinary progress has occurred in this
scientific field since the first edition was published due both to rapid advances in computer capability and the development
of large rockets and satellites which have made available to experimenters platforms for upper atmosphere and space research
that did not exist twenty-seven years ago.
The first edition of the Handbook was published the same year the first artificial earth satellite was orbited. At the time
of the third edition, there was considerable activity in space but man had not yet gone to the rhoon. Since the publication
of the third edition, research in and use of space has grown at an astonishing pace and its importance in man’s future is an
accepted fact. The Air Force has been active in space since the development of intercontinental ballistic missiles and early
recormaissance satellites, and is preparing for an even more active role as the space shuttle becomes operational.
In keeping with the Air Force interest in space, the format of the Handbook has been Wanged to give those chapters
dealing with the space environment increased emphasis. The fourth edition begins with chapters on the sun and its emissions,
then freats the earth’s magnetic field and the radiation belts, and follows with chapters on the ionosphere and the aurora.
The subject of electrical charging of space vehicles has been of special concern to the Air Force an@has been included to
aid designers interested in that problem. Of no less importance are Air Force operations in the atmosphere and on the earth
itself. The next group of chapters deals with properties of the atmosphere, and the Handbook concludes with chapters on
the earth sciences and infrared astronomy.
Readers familiar with earlier editions will note that some chapters of the older versions have ken deleted, while new
chapters have been added. The choice of subject matter has been detertnined prim~ly by changes in today’s Air Force
requirements and operational activities. The contents also are related to the technical programs at the Air Force Geophysics
Laboratory and reflect the expertise of the scientific staff, but should not be considered a technical report on such programs.
In fact, the Handbook is only an introduction to several areas of geophysics. Subjects such as galactic x ray astronomy,
astrophysics, planet~ physics, meteors, some aspects of atmospheric pollution, oceanography, and others not of current
concern to the Air Force do not appear here. Such subjects are covered very adequately elsewhere.
Numerous references have been included in each chapter so that the reader may pursue a specific subject area to whatever
depth desired. The authors were not constrained by chapter format or length but were allowed to express themselves freely.
As a result the chapters range from brief statements on a subject to textbook treatment of others. The Handbook has been
brought as up to date as possible. Other government organizations such as NASA and NOAA have also published surveys
and technical memoranda that deal with aspects of the atmosphere and space environment. Together with this Handbook
they can provide the reader with an up-to-date picture of our understanding of the atmosphere and the near-earth space
environment.
A. S. J.

vii
 ACKNOWLEDGMENTS

Many people have contributed to make this edition of the Handbook of Geophysics and Space Environment possible.
The decision to proceed with the fourth edition of the Handbook was made by Colonels James Baker and Gerald D’Arty,
Commander and Vice Commander, respectively, of the Air Force Geophysics Laboratory in 1981. We are indebted to the
authors of the twenty-five chapters; without them there would be no Handbook. Recognition is due three co-editors who
undertook the necessary preliminary planning steps to organize the Handbook before I assumed responsibility as scientific
editor. They are Dr. John N. Howard, Dr. Morton Barad, and Dr. Jules Aarons, who also contributed as co-author and
scientific editor of Chapter 10. I would like especially to acknowledge the very important contribution of my assistant editor.
Marylou Tschirch, who has been primarily responsible for the detailed editing, coordination, and final preparation of the
manuscript for publication. Her tireless effort has been a major factor in the success of the Handbook. Mr. John Dempsey
assisted with the editing of several chapters as well as contributing his expertise to discussions of format, typesetting, and
printing.
Thanks are due to many secretaries in the Laboratory who typed the draft manuscripts of the chapters for the authors
and, in particular, to Mrs. Ann Turner for her competent and extensive typing, copying, and accomplishment of many other
administrative tasks that were necessary in the preparation of this Handbook.
A.S.J.

ix
 TABLE OF CONTENTS

CHAPTER 1. THE SUN

1.1 THE SUN’S “STEADY-STATE’ BEHAVIOR (THE QUIET SUN) ........................ 1-1
1.1.1 Basic Characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-1
1.1.2 SolarRotation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-5
1.1.3 The SolarAtmosphere ...................................................... 1-5
1.1.3.1 ThePhotosphere . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-5
1.1.3.2 TheChromosphere . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-9
1.1.3.3 TheTransition Region . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-9
1.1.3.4 The Corona . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-9
1.2 THEACTIVESUN ................................................................ 1-13
1.2.1 TheSunspotCycle ......................................................... 1-14
1.2.2 The Solar Dipole Magnetic Field . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-14
1.2.3 The Emergence and Organization ofMagnetic Flux . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-16
1.2.3.1 Active Regions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-16
1.2.3.2 Ephemeral Regions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-17
1.2.4 Prominences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-17
1.2.5 Coronal Magnetic Fields . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-18
1.2.6 Flares . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-18
1.2.7 Coronal Activity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-20
1.2.7.1 Coronal Holes and Solar Wind . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-20
1.2.7.2 Streamers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-21
1.2,7.3 The SolarCurrentSheet ............................................. 1-21
1.2.7.4 Transients . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-22
1.2.7.5 Coronal Induced Disturbances . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-22
1.2.8 Solar Variability and Climate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-22
1.2.9 Predictions ofGeophysical Disturbances . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-23
1.2.9.1 Flare Prediction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-23
1.2.9.2 Coronal Disturbance Prediction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-24

CHAPTER 2. SOLAR ULTRAVIOLET

IRRADIANCE

2.1 SOLAR UVIRRADIANCE INTHE STRATOSPHERE .................................. 2-1

2.1.1 Solar Spectrum Between 2000 and3000 ~. ............................. 2-1
2.2 SOLAR CONSTANT AND SPECTRAL IRRADIANCE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-1
2.2.1 SolarConstant . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-1
2.2.2 Solar Irradiance Between 1 and 3000 ~For Solar Minimum . . . . . . . . . . . . . . . 2-5
2.2.3 Solar Irradiance Below 1200~ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-14
2.3 ABSORPTION OFSOLARUV ...................................................... 2-16
2.3.1 Absorption Below 3200~ . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-16
2.3,2 UVAtmospheric/IonosphericProcesses ................................. 2-16
2,4 SOLAR UVIRRADIANCE VARIABILITY ........................................... 2-16
2.4.1 Sunspot Number and10.7cm Solar Radio Flux . . . . . . . . . . . . . . . . . . . . . . . . . . 2-16
2.4.2 Solar Cycle and UV Variability Below 2000 ~ . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-18

xi
 1

TABLE OF CONTENTS

CHAPTER 3. SOLAR WIND

3-1
3.1 BASIC THEORY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. ””””” ““ . . . . . . . . . . . . . . . . .
3.1.1 Simple Spherically Symmetric Treatment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . 3-1
3.1.2 FurtherConsiderations ....................................................... 3-3

3.2 PARTICLE AND FIELD OBSERVATIONS ........................................... 3-4

3.2.1 Coordinates ............................................................... 3-4

3-4
3.2.2 Magnetic Fields . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.2.3 Protons and Electrons ....................................................... 3-7
3-11
3.2.4 OtherIons . . . . . . . . . . . . . . . . . . . . ...-..”.”””.”.””””””””” ““”””””””””””””””””
Waves and Turbulence . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-12
3.2.5
3-15
3.3 LARGE SCALE STRUCTURES AND DISCONTINUITIES ..............................
3-15
3.3.1 Discontinuities .............................................................
3-17
3.3.2 SolarWindStreams ........................................................
3-22
3.3.3 SolarFlare Disturbances .....................................................
3-22
3.4 LONG TERMVARIATIONS . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.4.1 SolarCycle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...”””””””””” ““”.””””””-”.”””” 3-22
3.4.2 SecularVariations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...”””””””””..””””””””” ““”” 3-24

3.5 THE CLOSE NEIGHBORHOODOF THE EARTH ..................................... 3-24

---
3-24
3.5.1 Magnetosheath . . . . . . . . . . . . . . . . . ...”””””””” . “””””””””””””””””””””””””” ““”
3-29
3.5.2 BowShock ...............................................................
3-31
3.5.3 Upstream Waves and Particles ................................................

CHAPTER 4. THE GEOMAGNETIC FIELD

4-1
4.1 BASICCONCEPTS ................................................................
4-1
4.1.1 Units, Terminology, and Conventions ..........................................
4-2
4.1.2 Coordinate Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4-5
4.1.3 Sources ofthe Geomagnetic Field . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4-5
4.1.4 The Steady Interior Field ....................................................
4-7
4.1.5 Quiet Variation ExternalFields ...............................................
4-7
4.1.6 Disturbance ExternalFields ..................................................
4-8
4.2 MEASUREMENTS OFTHE GEOMAGNETIC FIELD ..................................
4-8
4.2.1 Instrumentation . . . . . . . . . . . . . . . . . . . ...”””” . “”-”.”””””-”””-””””””””””” ““””””
4-12
4.2.2 Ground Measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4-12
4.2.3 Satellite and Rocket Measurements ............................................
4.3 THEMAINFIELD ................................................................ 4-13

4.3.1 Basic Description . . . . . . . . . . . ... . . . . . . . . . . . . . . . . . . . . . . ““”..””””””””””.” 4-13

4.3.2 The SecularVariation ....................................................... 4-16
4.3.3 Paeleomagnetism . . . . . . . . . . . . . . . . . . . . . . . . -.....”””.”””””””””””””” ““””””. 4-17

QUIET VARIATION FIELDS ....................................................... 4-19

4.4

4.4.1 The Solar Quiet Daily Variation .............................................. 4-19

4.4.2 The Lunar Daily Variation ................................................... 4-20
4-21
4.4.3 Magnetospheric DailyVtiation ...............................................

4,5 DISTURBANCEFIELDS ........................................................... 4-21

4-21
4.5.1 Geomagnetic Storms andSubstorms ...........................................
4-23
4.5.2 Geomagnetic Pulsations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4-25
4.6 QUANTITATIVE MAGNETIC-FIELD MODELS .......................................
4.6.1 TheInternalField . . . . . . . . . . . . . . . . . . . . . . . . . . ...” ...”.”.”.””””””””””””””. 4-25
4.6.2 The ExternalFiel d. . . . . . . . . . . . . . . . . . . . . . . . . ...”””.”.””””””” “.” ””” .$”””””’” 4-27

xii
 a
TABLE OF CONTENTS
4.7 GEOMAGNETIC ACTIVITY INDICES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-27
4.7.1 General-Activity Indices K, Ks,Kp, andKm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-27
4.7.2 Related Indices ap, Ap, ak, Ak,am, and aa . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-29
4.7.3 Character Indices Cp, Ci, andC9 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...’..... 4-31
4.7.4 Auroral-Electrojet Indices AE,AU,, AL, andAo . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-31
4.7.5 Ring-Current IndexDst . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-31
4.7.6 OtherIndices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-31
4.8 SOURCES OFGEOMAGNETIC DATA .............................................. 4-32
4.8.1 Data Centers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-32
4.8.2 Internationa lOrganizations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-33
4.8.3 Magnetic Stations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-33
4.8.4 Magnetic Field Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-33
4.8.5 Activity Indices and Charts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-33

CHAPTER 5. THE RADIATION BELTS

5.1 THEORETICALPRELIMINARIES .................................................. 5-3

5.1.1 Single Particle Motion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-3

5.1.2 Adiabatic Invariants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-3
5.1.2.1 First Adiabatic Invariant . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-4
5.1.2.2 Second Adiabatic Invariant . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-4
5.1.2.3 Third Adiabatic Invariant . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-5
5.1.3 Particle Drift Motion ....................................................... 5-5
5.2 TRAPPED RADIATIONSOURCES .................................................. 5-8

5.2.1 Qualitative Description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-8

5.2.2 Simplifying Assumptions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-9

5.3 TRANSPORT PROCESSES INTHE RADIATION BELTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-1o

5.3.1 Convection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-1o

5.3.2 Stochastic Processes: Phenomenology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-1o
5.3.3 Effects ofField Fluctuations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-11
5,3.4 Radial Diffusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-12
5.3.5 Pitch Angle Diffusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-13
5.3.6 Energy Diffusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-13
5.4 LOSS MECHANISMS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-13

5.4.1 Exosphere . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-14

5.4.2 Coulomb Collisions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-14
5.4.3 ChargeExchange . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-15
5.4.4 Wave-ParticleInteractions ................................................... 5-18
5.4.4.1 Pitch Angle Scattering Into the Loss Cone . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-18
5.4.4.2 Scattering ofEnergetic Electrons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-19
5.4.4.3 Limit On Radiation Belt Particle Fluxes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-20
5.4.4.4 Strong DiffusionLimit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-21

5.5 THEORETICAL RADIATION BELT MODELS ........................................ 5-21

5.5.1 Quiet Time, Steady State Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-22

5.5.1.1 Formulation ofBoundary Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-22
5.5.1.2 Elec&onModel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-22
5.5.1.3 Proton Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-23
5.5.1.4 HeavyIonModels ................................................. 5-24
5.5.1.5 Theoretical Radiation Belt Ionic Composition . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-26
...
X111
TABLE OF CONTENTS
5.5.2 Geomagnetic Storms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-26
5.5.2.1 The Variability ofthe Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-26
5.5.2.2 Magnetic Topology Variations.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-26

5.6 EMPIRICALRADIATION MODELS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .../... 5-27

5.6.1 Data Acquisition and Processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-27

5.6.1.1 Protons (Ions) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-27
5.6.1.2 Heavy Ions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-30
5.6.1.3 Trapped Electrons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-31
5.6.1.4 Shell Splitting Effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-36
5.6.2 Geosynchronous Altitude Region . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-38

5.7 THERINGCURRENT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ! . . . . . . . . ” ””””””.... 5-42

5.7.1 Electrical Current Relations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-42

5.7.2 Compositionand Sources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...”. 5-44
5.7.3 Adiabatic Effects Produced bythe Ring Current . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-45

5.8 RADIATION EFFECTS ONSPACE SYTEMS ......................................... 5-46

5.8.1 DetectorMalfunctions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-47

5.8.2 Memory Alteration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-49
5.8.3 Control SystemFailure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...”” 5-49
5.8.4 Biological Effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-50

5.9 MAN’S IMPACT ON THE RADIATION BELTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-50

5.9.1 NuclearDetonations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ............c.....- 5-50

5.9.2 ReleaseofChemicals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...”””””””” 5-51
5.9.3 Transmission ofRadioWaves ................................................ 5-51
5.9.4 Effects ofSpaceStructures ................................................... 5-51

CHAPTER 6. GALACTIC COSMIC RADIATION

AND SOLAR ENERGETIC PARTICLES

6.1 NOMENCL.ATUREAND DEFINITIONS ............................................. 6-1

6.2 GALACTIC COSMICRADIATION .................................................. 6-3

6.2.1 Primary Cosmic Radiation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-3

6.2.1.1 Elemental Composition ofprimary Cosmic Radiation . . . . . . . . . . . . . . . . . . . . . 6-3
6.2.1.2 The Anomalous Cosmic Ray Component . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-4
6.2.1.3 Cosmic Ray Electrons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-6
6.2.2 Cosmic Ray Flux Modulations . . . . . . . . . . . . . . . . . . . . . . . . . . . ““”””””..””””””.””.” 6-6
6.2.2.1 SolarCycle Modulations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-6
6.2.2.2 ShortTerm Modulations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-8
6.2.3 Geomagnetic Effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-9
6.2.3.1 Cosmic Ray CutoffRigidities ........................................ 6-9
6.2.4 Charged-Particle Flux Within the Atmosphere . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-12
6.2.4.1 The Secondary Cosmic Ray Ionizing Component . . . . . . . . . . . . . . . . . . . . . . . . 6-13
6.2.4.2 The Secondary Cosmic Ray Neutron Component . . . . . . . . . . . . . . . . . . . . . . . . 6-14
6.2.5 CosmicRayAlbedo ........................................................ 6-15

6.3 ENERGETIC SOLAR PARTICLES (SOLAR COSMIC RAYS) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-16

6.3.1 The Solar Particle Source . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-16

6.3.2 The Size and Frequency of Solar Particle Events atthe Earth . . . . . . . . . . . . . . . . . . . . . . . 6-17
6.3.2.1 SolarProtonEvents ................................................ 6-19
6.3.2.2 SolarElectron Events . . . . . . . . . . . . . . . . . . . . . . . ...”” “-oo”””””””c.”0000” 6-22

xiv
 TABLE OF CONTENTS
6.3.3 Elemental Composition of Solar Particle Events . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-23
6.3.3.1 Charged States of Solar Particles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-24
6.3.3.2 Corotating Energetic Particle Streams . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-24
6.3.4 Prediction of Solar Proton Events . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... . . . . . 6-24
6.3.4.1 Propagation of Solar Particles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-24
6.3.4.2 Particle Increase Onset Time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-26
6.3.4.3 Time of Maximum P~icle Intensity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-26
6.3.4.4 Magnitude of Maximum Intensity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-27
6.3.4.5 ProtonEventDecay ................................................ 6-27
6.3.4.6 Differential Energy Spectra . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-27
6.3.4.7 Prediction ofPolar Cap Absorption . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-27

CHAPTER 7. THE CHARGING OF SPACECRAFT SURFACES

7.1 SPACECRAFT CHARGING-HISTORICAL PERSPECTIVE ............................. 7-1

7.2 Spi4CECRAF1’ CHARGING OBSERVATIONS ........................................ 7-2

7.2.1 Rocket Measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-2
7.2.2 Satellite Measurements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-3
7.2.3 The Effects ofSpacecraft Charging . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-6
7.3 CURRENTMECHANISMS ......................................................... 7-1o
7.3.1 Time Scales . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-11
7.3.2 IncidentParticle Fluxes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-11
7.3.3 Photoelectron Currents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-12
7.3.4 Backscattered and Secondary Electrons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-13
7.3.5 Magnetic Field—Induced Current Distortions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-14
7.3.6 Motion-InducedEffects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-15
7.3.7 Charge Deposition by Energetic Particles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-15
7.3.8 Artificial Charging Mechanisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-16
7.4 sPACECRAnCHARGINGTHEORY ............................................... 7-16
7.4.1 Analytic Probe Theory-General Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-16
7.4.2 Analytic Probe Theory-Thick Sheath Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-18
7.4.3 Analytic Probe Theory-Thin Sheath and Related Models . . . . . . . . . . . . . . . . . . . . . . . . . 7-20
7.4.4 GeneralProbeTheo~ ....................................................... 7-22
7.4.5 Numerical Simulation Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-25
7.4.6 NASCAP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-26
7.5 PREVENTION OF Spacecraft CHARGING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-28

7.6 CONCLUSIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-29

CHAPTER 8. MAGNETOSPHERIC AND HIGH LATITUDE

IONOSPHERIC ELECTRODYNAMICS

8.1 MAGNETOSPHERIC BOUNDARY INTERACTIONS ................................... 8-2

8.1.1 TheMagnetopause . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-2
8.1.2 Convection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-3
8.1.3 Boundary Layers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-5
8.2 THEPLASMASHEET ............................................................. 8-6
8.2.1 Adiabatic Motion in the Plasma Sheet . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-7
8.2.2 Pitch Angle Diffusion ofPlasma Sheet Particles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-10

xv
TABLE OF CONTENTS
8.3 ELECTRICAL COUPLING OF THE MAGNETOSPHERE AND 10NOSPHERE ............. 8-11

8.4 HIGH LATITUDE ELECTRIC FIELDS ............................................... 8-12

8.5 POLAR CAP ELECTRODYNAMICS ................................................. 8-13

8.5.1 Large Scale Electric Fields Patterns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-13

8.5.2 PolarCap Precipitation ...................................................... 8-16
8.5.3 PolarCapArcs ..................................................... ... 8-17

8.6 AURORAL OVAL ELECTRODYNAMICS .......................................... 8-19

8.6.1 Global Field Aligned Currents ................................................ 8-19

8.6.2 Equatorward Boundary ofthe Oval . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8-20
8.6.3 Inverted-V Phenomenology ................................................... 8-23
8.6.4 Substorms ................................................................ 8-25

8.7 CONCLUSIONS .......................................................... ...... 8-25

CHAPTER 9. IONOSPHERIC PHYSICS

9.1 STRUCTURE OFTHEIONOSPHERE ................................................ 9-1

9.1.1 Ionospheric Layers . . . . . . . . . . . . . . . . . . ............ . . . . . . . . . . . . . . . . . .. . . . . . . . . . 9-I

9.I.2 Chapman Theory forIonosphefic Layers ............ . . . . . . . . . . . . . . . . . .. . . . . . . . . . 9-2
9.1.3 Ionization Production, Loss, and Vertical Transport .. . . . . . . . . . . . . . . . . . .. . . . . . . . . . 9-2
9.1.4 Neutral Winds and Horizontal Transpofl ............ . . . . . . . . . . . . . . . . . .. . . . . . . . . . 9-4

9.2 HIGH LATITUDEPHENOMENA ................................................... 9-4

9.2.1 Total Ionization Trough and Light Ion Trough . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . 9-4

9.2.2 The Ionosphere in the Auroral Oval ......... . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . 9-4
9.2.3 SubstormEffects ................ ......... . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . 9-5
9.2.4 PolarCapStructure .............. ......... . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . 9-5

9.3 EQUATORIAL PHENOMENA ...................................................... 9-6

9.3.1 SqCurrentSystem .......... ...... ........ . . . . . . . . . ... . . . . . ......... . . . 9-6

9.3.2 Equatorial Electrojet ......... ...... ........ . . . . . . . . . ... . . . . . ......... . . . .. . . 9-6
9.3.3 Electrojet Irregularities ....... ...... ........ . . . . . . . . . ... . . . . . ......... . . . .. . . 9-7
9.3.4 Equatorial Anomaly and Fountain Effect ....... . . . . . . . . . ... . . . . . ......... . . . .. . . 9-8
9.3.5 Equatorial F Region Irregularities .... ........ . . . . . . . . . ... . . . . . ......... . . . .. . . 9-9

CHAPTER 10. IONOSPHERIC RADIO WAVE PROPAGATION

10.I MEASURINGTECHNIQUES ....................................................... 1o-1

10.1.1 Ionosonde . . . . . . . . . . . . . . . . . . . . . . . . . . . ................... .................. 10-1

10.1.1.1 Monogram . . . . . . . . . . . . . . . . . . . ................... .................. 1o-1
10.1.1.2Principles of Ionospheric Sounding ................. ................... 10-3
10.1.1.3 AnalogIonosonde ............ ................... ................... 10-4
10.1.l.4Analog/DigitalHybrid Ionosonde .................. ................... 10-5
10.I.1.5 Digital Ionosondes . . . . . . . . . . . . ................... ................... 10-6
10.1.1.6Digital Data Processing ....... ................... ................... 10-8
10.1.1.7 FM/CWorChirpSounder ..... ................... ................... 10-8
10.1.1.8 Topside Sounders . . . . . . . . . . . . ................... ................... 1o-1o
10.1.I.9 Ionogram interpretation . . . . . . . . ................... ................... 10-11
10. l.l.lOlonosonde Network.. ......... ................... ................... 10-13

xvi
 TABLE OF CONTENTS
10.1.2 Incoherent Scatter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... . . . . 10-14
10.1.3 LangmuirProbes ...................... ..................................... 10-17
10.1.4 Faraday Cups for Rockets and Satellites ... ..................................... IO-17
10.1.5 Optical Measurements .................. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . 10-18
10.1.5.1 Observing Techniques . . . . . . . . . . ..................................... 10-18
10.1.5.2 Ionospheric Structure from Optical Measurements ........................ 1O-I8

10.2 SOME ASPECTS OF LONG WAVE PROPAGATION .................................. 10-20

10.2.1 Groundwave Propagation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-22

10.2.1.1 Idealized Flat-EarthModels .......................................... 10-22
10.2.1.21dealized Spherical-Earth Models ..................................... 10-24
10.2.1.3 Models with Earth-Properties Gently Varying Along the Propagation Path . . . . 10-25
10.2.2 ELFPropagation ........................................................... IO-25
10.2.2.1 ELF Field-Strength Calculations ...................................... 10-26
10.2.2.2 Theoretical and Measured ELF Propagation Constants .................... 10-26
10.2.2.3 Anomalous ELFPropagation ......................................... 10-28
10.2.2.4 Analysis of Laterally Non-Uniform Ionospheric Disturbances .............. 10-28
10.2.3 Long Range VLFILFPropagation ............................................. 10-28
10.2.3.1 Waveguide Modes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-29
10.2.3.2 Waveguide Propagation Equations .................................... 10-29
10.2.3.3 TE/TMModeStructure ............................................. 10-30
10.2.3.4 NumericalModelingofVLF/LF Waveguide Propagation .................. 10-33
10.2.3.5 OtherVLF/LFPropagationPrediction Techniques . . . . . . . . . . . . . . . . . . . . . . . . 10-34
10.2.3.6 VLF/LFProbingoftheIonosphere .................................... 10-34
10.2.4 MFPropagation ............................................................ 10-36
10.2.4.I MFGroundwave Propagation ........................................ 10-36
10.2.4.2 MFSkywave Propagation ........................................... 10-36
10.4.2.3 Effect of MFWaves on the Ionosphere ................................ 10-36
10.2.5 Long Wave Propagation Through the Ionosphere . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-37

10.3 IONOSPHERIC MODELING ........................................................ 10-38

10.3.1 The Numerical-Phenomenological Models ...................................... 10-39

10.3.1.1 ThelTS-78Model ................................................. 10-39
10.3.1.2 TheBentModel ................................................... 10-40
10.3.1.3 The Ionospheric Communications Analysis and Prediction Program
(IONCAP) ....................................................... 10-41
10.3.1.4 The Bradley Model... .............................................. 10-42
10.3.1.5 The Air Force Global Weather Central 4-D Model . . . . . . . . . . . . . . . . . . . . . . . Io-43
10.3.1.6 International Reference Ionosphere—IRI-79 ............................. Io-43
10.3.2 The Theoretical Models ..................................................... 10-44
10.3.3 Comparison of the Phenomenological Models, Their Limitations addability .......... 10-45

10.4 HIGH FREQUENCY RADIO PROPAGATION ......................................... 10-45

10.4.1 Morphology ofthe Ionosphere .............................................. . . 10-47

10.4.2 Simple Ray Propagation ................................................... . . 10-51
10.4.3 Requirements for Successful Communications ................................. . . 10-54
10.4.4 Predictions for HFCommunications ......................................... . . 10-55
10.4.5 Problems with HFCommunications .......................................... . . 10-57
10.4.5.1 Multipath Propagation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-57
10.4.5.2 SporadicE ...................................................... . . 10-57
10.4.5.3 Problems in High Latitudes ........................................ . . 10-57
10.4.5.4 Problems in Low Latitudes ........................................ . . 10-58
10.4.5.5 Spread F and Irregularities ......................................... . . IO-58
10.4.6 Disturbances to Normal Communications ..................................... . . 10-58
10.4.7 Unusual Propagation Modes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-59

xvii
‘1’ABLE OF CONTENTS

10.4.8 Short-Term Forecasting of HF Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . i 0-60

10.4.9 Real Time Channel Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-61
10.4.10 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-62
10.5 IONOSPHERICDISTURBANCES ............................. ................... 10-62
10.5.1 Sudden Ionospheric Disturbance (SID) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.62
10.5.2 Polar Cap Absorption (PCA) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-63
10.6 HFRADAR IONOSPHERIC CLUTTER .............................................. 10-62
10.6.1 Signal Amplitude . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-63
10.6.2 Cross Section . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-67
10.6.3 AngularExtent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-68
10.6.4 DopplerFrequency Spectmm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-69
10.6.5 Frequency of Occurrence and Comelation with Solar-Geophysical Conditions . . . . . . . . . . 10-70
10.6.6 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-71

10.7 SCINTILLATION ON TRANS-IONOSPHERIC RADIO SIGNALS ........................ 10-71

10.7.1 GlobalMorphology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-72
10.7.2 Scintillation Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-72
10.7.3 Signal Characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-72
10.7.4 Frequency Dependency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-73
10.7.5 Fading Spectra . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-73
10.7.6 Geometrical Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-74
10.7.7 Spread F and Scintillation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-75
10.7.8 Equatorial Scintillations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-75
10.7.8.1 Patch Characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-75
10.7.8.2 Variation of Scintillation Activity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-77
10.7.8.3 In-situData . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-77
10.7.8.4 Sunspot Cycle Dependence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-77
10.7.9 Middle Latitude Scintillation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . IO-78
10.7.9.I Effect of Magnetic Index on Midlatitude Scintillation . . . . . . . . . . . . . . . . . . . . . 10-79
10.7.10 The High Latitude Region . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Io-79
10.7.10.1 The Plasmapause andthe Trough . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-79
10.7.10.2Auroral Scintillations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-80
10.7.10.3 Polar Scintillations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-81
10.7.11 Empirical Model ofGlobal Scintillation Behavior . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-82
10.7.11.1 WBMOD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-82
10.7.11.2Formulas in Atlantic Sector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-83

10.8 IONOSPHERIC TIME DELAY EFFECTS ON EARTH-SPACE PROPAGATION ............ 10-84

10.8.1 GroupPathDelay . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...””””””””.” “ 10-84

10.8.1.1 Two-Frequency Ionospheric Time Delay Corrections . . . . . . . . . . . . . . . . . . . . . 10-84
10.8.1.2 An Example of a Two-Frequency Ionospheric Time Delay System . . . . . . . . . . 10-85
10.8.2 RF Carrier Phase Advance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-85
10.8.2.1 Differential Carrier Phase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-85
10.8.2.2 Second Difference ofCamier Phase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-86
10.8.3 DopplerShift . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...”””””.. 10-86
10.8.4 Faraday Polarization Rotation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-87
10.8.5 Angular Refraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-87
10-88
10.8.6 Distortion ofPulse Waveforms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
10.9 IONOSPHERIC TOTAL ELECTRON CONTENT (TEC) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-89

10.9.1 AverageTEC Behavior . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-89

10.9.2 Temporal Variability ofTEC . . . . . . . . . . . . . . . . . . . . . ...” ““”””””””...””””” 10-91
10.9.2.1 Variability from Monthly Mean TEC Values . . . . . . . . . . . . . . . . . . . . . . . . . . . . Io-91
10.9.2.2 Short Term Temporal Variabilityof TEC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-92
10.9.2.3 Geographic Variabilityof TEC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-92
.. .
Xvlll
 TABLE OF CONTENTS
10.9.3 TECinthe Near-Equatorial Region . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-92
10.9.4 TECinthe Auroraland Polar Cap Regions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-94
10.9.5 Protonospheric E1ectronContent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-95
10.9.6 Short Term Variations inTEC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . 10-95
10.9.7 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-96
10.10 ARTIFICIALMODIFICATION ..................................................... 10-96
10.lO.l Chemical Releases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-96
10.10.2 High Power HFTransmissions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10-98

CHAPTER 11. SOLAR RADIO EMISSION

ll.l BASICDEFINITIONS ............................................................. 11-1

11.2 THE MINIMUM (ZERO-SUNSPOT) COMPONENT .................................... 11-2

11.3 THE SLOWLY VARYING COMPONENT ............................................ 11-2

11.4 THEBURSTCOMPONENT ........................................................ 11-2

11.4.1 Meter-Wave Range (25–580 MHz) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-2
11.4.2 Decimeter-Wave Range (500-2000 MHz) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-3
11.4.3 Centimeter-Wave Range (2000-35000 MHz) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-3
11.5 CORRECTIONS TO QUIET SUN AND BURST-FLUX DENSITIES ....................... 11-4

11.6 QUIET SUN FLUX-DENSITY MEASUREMENTS ..................................... 11-4

11.7 SOLAR RADIO-BURST CLIMATOLOGY BASEDON 1966-1978 SAGAMORE HILL

OBSERVATIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11-6

CHAPTER 12. THE AURORA

12.1 PHENOMENOLOGY, MORPHOLOGY, AND OCCURRENCE ........................... 12-1

12.1.1 General Characteristics ofthe Aurora . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-1
12.1.1.1 Particle Energy-Ionization and Excitation Height Relation . . . . . . . . . . . . . . . . 12-1
12.1.1.2 Particle Energy-Magnetospheric Regime Relation . . . . . . . . . . . . . . . . . . . . . . . 12-1
12.1.1.3 SpatialStructure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-2
12.1.1.4 Morphology ofthe ERegion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-2
12.1.1.5 Morphology ofthe FRegion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-3
12.1.1.6 Morphology ofthe DRegion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-3
12,1.1.7 Composite . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-4
12.1.2 Distribution of Auroras in Magnetic Latitude andkal Time . . . . . . . . . . . . . . . . . . . . . . 12-4
12.1.2.1 Statistical Distributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-5
12.1.2.2 Instantaneous Distributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-10
12.1.3 Empirical and Model Relationships . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-13
12.1.3.1 Auroral Ionospheric Profiles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-13
12.1 .3.2 Relations Between Auroral Electron, Optical, and Ionospheric Parameters . . . . 12-16
12.1.3.3 AuroralActivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-17
12.1.3.4 Heights ofDiscrete Auroras . . . . . . . . . . . . . . . . . . ... . . . . . . . . . . . . . . . . . . . . . . 12-18
12.2 AURORAL OPTICAL AND INFRARED EMISSIONS .................................. 12-19
12.2.1 Laboratory Studies and Auroral Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-21
*12.2.2 Artificial Auroral Experiments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-21

xix
TABLE OF CONTENTS

12.2.3 EmissionRates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...””.”””” 12-22

12.2.4 Auroral Processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-23
12.2.5 Infrared AuroralEmissions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12-27
12.2.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ””””””. ”..”””..” . 12-37

CHAPTER 13. INFRARED AIRGLOW

13.1 SOURCES OFTHE INFRARED AIRGLOW . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-1

13.1.1 Resonant Scattering ofthe Eafihshine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-1

13.1.2 Scattering ofSunshine ...................................................... 13-2
13.1.3 Photodissociation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...””.” 13-2
13.1.4 Chemical Reactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-2
13.1.5 Atomic and Molecular Collisions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-3

13.2 NON-LOCAL THERMODYNAMIC EQUILIBRIUM (NLTE) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-3

13.3 METHODS OFMEASUREMENT ................................................... 13-4

13.4 RESULTS ANDDISCUSSION ...................................................... 13-6

13.4.1 Thermosphere (above lookm) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-6

13.4.2 Lower Thermosphere (between loo kmand85 km) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-7
13.4.3 Upper Mesosphere (between85h and70km) .................................. 13-8
13.4.4 Lower Mesosphere (between 70km and40 km) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-9
13.4.5 InfraredRadiators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...””””””.. ““ 13-10
13.4.5.1 C02 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. O....”””””””””””” ““”” 13-10
13.4.5.2H*0 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. o.....”””””””””””””.”””.. . 13-12
13.4.5.3 NO . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...” .“-”.”..””””” 13-12
13.4.5.403 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...”””””””””””””” ““”” 13-12
13.4.5.5 Oz(lA~) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13-13
13.4.5.60H . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...”” ““”’””””..””” 13-13

CHAPTER 14. STANDARD AND REFERENCE ATMOSPHERES

14.1 STANDARD ATMOSPHERE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-1

14.1.1 U.SStandard Atmosphere, 1976 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-1

14.1.1.1 Equilibrium Assumptions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-2
14.1.1.2 Atmospheric Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-2
14.1.1.3 Atmospheric Structur e........ ....................................... 14-6

14.2 REFERENCEATMOSPHERES ...................................................... 14-7

14.2.1 Air Force Reference Atmospheres for Altitudes Below 90Km . . . . . . . . . . . . . . . . . . . . . . 14-7
14.2.1.1 Basic Assumptions and Formulas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-7
14.2.1.2 Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ....”””””..””””. 14-8
14.2.1.3 Seasonal and Latitudinal Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-8
14.2.1.4 Longitudinal Variations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-21
14.2.1.5 ColdandWarmWinterStiatosphere/Mesosphere ........................ 14-23
14.2.2 High Altitude Reference Atmospheres . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-26
14.2.2.1 Development ofReference Atmospheres . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-26
14.2.2.2 Mean CIRA Reference Atmosphere . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-28
14.2.2.3 Reference Atmospheres W–120km ................................... 14-32
14.2.2.4 Reference Atmosphemsabove 120 km . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14-36

xx
 a
TABLE OF CONTENTS

CHAPTER 15. ATMOSPHERIC TEMPERATURES, DENSITY AND PRESSURE

15.1 THERMAL PROPERTIES, SURFACET0 90KM ...................................... 15-1

15.1.1 Energy Supply and Transformation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-I

15.1.2 SurfaceTemperature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-2
15.1.2.1 Official Station Temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-2
15.1.2.2 DailyTemperature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-3
15.1.2.3 Horizontal Extent of Surface Temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-5
15.1.2.4 Runway Temperatures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-7
15.1.2.5 Temperature Extremes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-7
15.1.2.6 TheGumbelModel ................................................ 15-10
15.1.2.7 Temperature Cycles and Durations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-11
15.1.3 UpperAirTemperature ...................................................... 15-14
15.1.3.1 Seasonal and Latitudinal Variations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-15
15.1.3.2 Distribution Around Monthly Means and Medians . . . . . . . . . . . . . . . . . . . . . . . 15-16
15.1.3.3 Distributions at Pressure Levels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-18
15.1.3.4 Interlevel Correlation ofTemperature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-18
15.1.4 Speed ofSound vs Temperature. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-19
15.1.5 Earth/AirInterfaceTemperatures .............................................. 15-19
15.1.6 Subsoil Temperatures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-21
15.1.7 Degree-Day and Temperature-Wind Combinations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-29

15.2 ATMOSPHERIC DENSITYUP T090KM ............................................ 15-31

15.2.1 Seasonal and Latitudinal Variations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-31

14.2.2 Day-to-Day Variations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-31
15.2.3 Spatial Variations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-33
15.2.4 Statistical Applications to Reentry Problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-36
15.2.5 Variability with Time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-38

15.3 ATMOSPHERIC PRESSURE UP T090KM .......................................... 15-39

15.3.1 Sea-hvel Pressure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-46

15.3.2 Seasonal and Latitudinal Variations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-47
15.3.3 Day-to-Day Variations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-50
15.3.4 Diurnal and Semidiurnal Variations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15-51

CHAPTER 16. ~;;:t.APOR, PRECIPITATION, CLOUDS,

16.1 ATMOSPHERICWATER VAPOR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-1

16.1.1 Surface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-3
16.1.2 Troposphere . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-4
16.1.3 Stratosphere andMesosphere . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-6
16.2 PRECIPITATION ................................................................. 16-8
16.2.1 Surface RatesofRainfall .................................................... 16-8
16.2.1.1 ClockHourlyRates ................................................ 16-8
16.2.1.2 InstantaneousRates ................................................ 16-9
16.2.1.3 ExtremeRates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-14
16.2.1.4 Duration ofRainfall Rates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-15
16.2.1.5 Rainfall Rates Along Lines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-15
16.2.2 Hail . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-16
16.2.2.1 Hailstone Characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-16
16.2.2.2 HailstormCharacteristics ............................................ 16-16

xxi
TABLE OF CONTENTS

16.2.2.3 Frequency and Duration at the Ground . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-17

16.2.2.4 HailAloft . . . . . . . . . . . . . . . . . . . . . . . ...””. .......””””””..””””.””. 16-17
16.2.2.5 Area andLine Coverage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-18
16.2.2.6 Hail in Design and Operationof Equipment ““”””..””””.””””.”””. “:”” 16-18
16.2.3 Snowfall . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-19
16.2.4 Ice Accretion on Surface Stmctures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-20
16.2.5 Distribution ofPrecipitation Elements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-22
16.2.5.1 Raindrop Size Distributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-22
16.2.5.2 Snowflakes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-23
16.2.5.3 Distribution of Precipitation Elements with Height . . . . . . . . . . . . . . . . . . . . . . . 16-24
16.2.5.4 Extreme Values ofLiquid Water Content . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-24

16.3 CLOUDS . . . . . . . . . . . . . . . . . . . . . . . . . ...”” . “””””””””.”..”..””””.””””””””” ““”””””””” 16-26

16.3.1 SurfaceObservations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-26

16.3.1.1 Summaries of Surface Observations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-27
16.3.1.2 GlobalSkyCover .................................................. 16-30
16.3.1.3 Modeling ofCloud Distributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-30
16.3.2 Cloud Cover inLayers Aloft . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-32
16.3.3 Satellite and RadarObsemations .............................................. 16-32
16.3.3.1 Satellites . . . . . . . . . . . . . . . . . . . . . ...”””” ....”””””””.”””””.””.””””+ ““ 16-32
16.3.3.2 Radar . . . . . . . . . . . . . . . . . . . . . . . . . . . ...””””.”””””””” . “””.””.””-”””””” 16-34
16.3.4 Clear and Cloud-Free Lines-of-Sight . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-34
16.3.4.1 AircraftObservations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-34
16.3.4.2 Surface Observations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-34
16.3.4.3 Comparison of Estimates Between Aircraft andthe CFLOSModel Basedon
Surface Observations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-37
16.3.4.4 Cloud-Free Fields-of-View . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-38
16.3.4.5 Joint Probabilities ofCFLOS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-40
16.3.5 Water ContentofClouds .................................................... 16-40
16.3.5.1 Liquid Water Content in Clouds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-41
16.3.5.2 Measuring Liquid Water Content . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-41
16.3.5.3 Supercooled Clouds and Aircraft Icing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-41
16.3.6 CloudsAbove6Km ........................................................ 16-43
16.3.6.1 Observations ofCirrus Clouds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-43
16.3.6.2 Subvisible CirrUs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-45
16.3.6.3 Clouds Above theTropopause . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-45
16.3.6.4 TropopauseHeight . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-46

16.4FOG . . . . . . . . . . . . . . . . . . . . . . ...””... ““” ”.”””.””””””-”””-””””””””””” ““”””””””””””” 16-46

16.4.1 Microphysical Stmcture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-46

16.4.1.1 CondensationNuclei . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-46
16.4.1.2 Drop Size Distributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-47
16.4.1.3 Liquid WaterContent. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-47
16-48
16.4.2 Visibility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
16.4.2.1 Definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-48
16.4.2.2 Variability . . . . . . . . . . .......”.”.””.””.””.”””””” “-”””””’””””””””” 16-50
16.4.2.3 Relationship Between Extinction Coefficient and Liquid Water Content . . . . . . 16-53
16.4.2.4 Modeling of Visibility Distributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16-54

CHAPTER 17. WINDS

17.1 WIND AS AFUNCTION OF HEIGHT ............................................... 17-1

17.1.1 Variation ofWind Speed with Height (Lowest 100m) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-2

17.1.2 Wind Direction Shifts (Below 3000 m) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-3
17.2.3 Diurnal Variation and Low Level Jet Streams (Below 2000m) ..................... 17-4

xxii
 TABLE OF CONTENTS
17.2 LARGE SCALE WIND STRUCTURE ................................................ 17-6
17.2.1 Seasonal and Day-to-Day Variations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-7
17.2.2 Time and Space Variations . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-8
17.2.2.1 Time Variability upto30km ........................................ 17-11
17.2.2.2 Spatial Variability upto 30km . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-13
17.2.2.3 Time and Space Variations-30to 60km . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-13
17.3 WINDPROFILES ................................................................. 17-14
17.3.1 WindShear . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-14
17.3.2 Interlevel Correlations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-15
17.4 DESIGNDATAONWINDS ........................................................ 17-21
17.4,1 Hourly SurfaceWind Speeds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-21
17.4.2 SurfaceWindDirection ..................................................... 17-22
17.4.3 SurfaceWindGusts ........................................................ 17-23
17.4.4 Extreme SutiaceWindSpeeds ................................................ 17-25
17.4.5 Structure ofJetStrearns ..................................................... 17-27
17.5 THERMOSPHERICWINDS. ........................................................ 17-29
17.5.1 Observational Summary and Limitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-30
17.5.2 DiurnallyAveragedWinds ................................................... 17-31
17.5.2.1 Seasonal Variations at So]arMaximurn ................................ 17-31
17.5.2.2 Solstice Variationsat Solar Minimum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-34
17.5.3 Local Time Variations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-34
17.5.4 Magnetospheric Convection Effects on Neutral Winds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17-35

CHAPTER 18. OPTICAL AND INFRARED PROPERTIES

OF THE ATMOSPHERE

18.1 ATMOSPHERICGASES ........................................................... 18-1

18.1.1 Atmospheric Molecules, Models ofihe Atmospheric Composition . . . . . . . . . . . . . . . . . . . 18-i
18.1.2 Molecular Absorption and Spectroscopic Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . ,., 18-4
18.1.2.1 The Role ofMolecular Absoqtion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-4
18.1.2.2 The AtmosphericAbsorptionLine Parameter Compilation . . . . . . . . . . . . . . . . . 18-5
18.1.3 IndexofRefraction ......................................................... 18-7
18.1.4 MolecularRayleigh Scattering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-7
18.2 AEROSOLS, CLOUD, AND PRECIPITATION PARTICLES ............................. 18-9
18.2.1 Aerosols, Geographic, and Temporal Variations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-9
18.2.1.1 Composition, Sources and Sinks, Refractive Index, Particle Shapes . . . . . . . . . 18-9
18.2.1.2 Size Distributions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-10
18.2.1.3 Vertical Profiles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-10
18.2.1.4 Models ofAerosol Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-10
18,2.1.5 Aerosol Extinction, Scattering, Polarization . . . . . . . . . . . . . . . . . . . . . . . . . . . ., 18-15
18.2.2 Cloud, Fog, and Precipitation Particles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-25
18.2.2.1 Particle Types, Water-Ice Refractive Index, Particle Shapes . . . . . . . . . . . . . . . . 18-25
18.2.2.2 Size Distributions.. ................................................ 18-28
18.2.2.3 Cloud-PrecipitationModels .......................................... 18-29
18.2.2.4 Extinction, Scattering, Polarization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-30
18.3 SURFACE REFLECTANCEAND ALBEDO . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-35

18.4 ATMOSPHERIC TRANSMI~ANCE AND RADIANCE ................................ 18-37

18.4.1 Monochromatic Radiation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-37
18.4.1.1 Line-by-Line Computations (FASCODE) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-37
. ..
Xxlll
TABLE OFCONTENTS
18.4. l.2 Continuum Absorption .............................................. 18-39
18.4.1.3 Radiance and Transmittance for Atmospheric Paths ...................... 18-42
18.4.2 Broad-Band Radiation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . !8-flfl
18.4.2.1 Broad-Band Transmission (LOWTRAN) ............................... 18-44
18.4.2.2 Assumptions ofthe LOWTRAN Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . ‘. . . . 18-46
18.4.2.3 LOWTRAN Atmospheric Transmittance ............................... 18-46
18.4.2.4 LOWTRAN Atmospheric Radiance ................................... 18-47
18.4.2.5 Examples of Transmittance and Radiance Spectra . . . . . . . . . . . . . . . . . . . . . . . . 18-48
18.4.2.6 Comparisons of LOWTRAN With Measurements ........................ 18-48
18.4.2.7 Limitations ofLOWTRAN .......................................... 18-51

18.5 RADIATION TRANSFER BY ATMOSPHERIC SCATTERING ........................... 18-53

18.5.1 Multi~le Scattering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-53

18.5. i.l Formallolution to the Multiple Scattering Problem ...................... 18-54
18.5.1.2 Method of Successive Orders of Scattering ............................. 18-54
18.5.1.3 Gauss-Seidel iterative Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-54
18.5.1.4 The Matrix Operator or Layer Adding Methods . . . . . . . . . . . . . . . . . . . . . . . . . . 18-55
18.5.1.5 Monte CarloTechnique ............................................. 18-55
18.5.1.6 Discrete Ordinates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18-55
18.5.1.7 Dodecaton Approach to Radiative Transfer (DART) . . . . . . . . . . . . . . . . . . . . . . 18-56
18.5.1.8 Spherical Harmonic Method ......................................... 18-56
18.5.1.9 Small-Angle Approximations.. ....................................... 18-57
18.5. l.lOADDroximations ofImadiance . . . . . . . . . . ...”””.” ““”””””””””” 18-57
,.
18.5.1.11 Approximations forthe Radiance ..................................... 18-58
18.5.2 Background (Sky) Radiance .................................................. 18-58
18.5.3 Illuminance-lmadiance ....................................................... 18-60
18.5.4 Contrast Transmittance-Visibility .............................................. i 8-60

18.6 ATMOSPHERIC REFRACTION ..................................................... 18-65

18.6.1 Refractive Bending ......................................................... 18-65

18.6.2 OpticalAirMass ........................................................... 18-68

18.7 ATMOSPHERIC OPTICALTURBULENCE ........................................... 18-69

18.7.1 Amplitude Fluctuations (Scintillation) .......................................... 18-7 I

CHAPTER 19. ELECTROMAGNETIC WAVE PROPAGATION

INTHELOWER ATMOSPHERE

19. lREFRACTION IN THE LOWER TROPOSPHERE ...................................... 19-1

19.1.1 Optical Wavelengths ........................................................ 19-I

19.1.2 RadioWavelengths ....................................................... 19-2
19.I.3 Standard Profiles of Refractive Modulus ........................................ 19-2
19.1.4 Variations of Refractive Moduli . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19-3
19.1.5 Turbulence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...” 19-6

19.2 ATTENUATION AND BACKSCATTER1NG .......................................... 19-9

19.2.1 Backscattering and Attenuation Cross Sections . . . . . . . . . . . . . . . . ................... 19-10

19.2.2 Reflectivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...””””””” 19-1 I
19.2.3 Attenuation by Precipitation ............................... ................... 19-13
19.2.4 Total Attenuation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...”””””””’” 19-13

xxiv
 TABLE OF CONTENTS
CHAPTER 20. ATMOSPHERIC ELECTRICITY

20.1 FAIR WEATHER ELECTRICITY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ,20-I

20.1.1 Electrical Conductivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20-1
20.1.2 Electric Field . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20-3
20.1.3 Currents and Space Charge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20-4
20.1.3.1 Air-Earth Conduction Current . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20-4
20.1.3.2 Convection Current . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20-5
20.1.3.3 Displacement Current . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20-6
20.1.3.4 Electrode Effect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20-6
20.1.3.5 Earth Charge and Worldwide Cument System . . . . . . . . . . . . . . . . . . . . . . . . . . . 20-6
20.1.4 Atmospheric Ions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20-6
20.1.4.1 Definition and Relations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20-6
20.1 .4.2 Positive Ion Chemistry and Composition in the Stratosphere
and Troposphere . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20-8
20.1 .4.3 Negative Ion Chemistry and Composition in the Stratosphere
and Troposphere . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20-10
20.1.5 Electrical Equilibrium and Variations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20-1 I
20.1.5.1 Electrical Equilibrium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20- I I
20.1.5.2 Variations in the Exchange Layer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20-1 I
20.1.5.3 Variations in the Free Atmosphere . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20-13
20.1.6 Solar Influence on Earth’s Atmospheric Electrical Parameters . . . . . . . . . . . . . . . . . . . . . . . 20-15
20.1.7 Global Model ofAtmospheric Electricity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20-17
20.1.8 Recent Advances in the Middle Atmosphere . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20-17
20.1.8.1 Middle Atmosphere . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20-17
20.1.8.2 Middle Atmosphere Ion Concentration, Mobility, and Conductivity . . . . . . . . . 20-18
20.2 THUNDERSTORMACTIVITY ...................................................... 20-20
20.2.1 Thunderstorm Charge Distribution and Electric Field Pattern . . . . . . . . . . . . . . . . . . . . . . . 20-20
20.2.1.1 Surface Electric Field . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20-21
20.2.1.2Field PatkmsAloft ................................................ 20-22
20.2.2 Lightning Characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20-26
20.2.2.I General Phenomenology ofthe Discharge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20-26
20.2.2.2 Characteristics of Radio Frequency Signals Due to Lightning . . . . . . . . . . . . . . 20-27
20.2.3 Precipitation Static Characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20-32
20.2.4 Distribution and Duration ofThunderstoms ..................................... 20-32

CHAPTER 21. ATMOSPHERIC COMPOSITION

21.1 ATMOSPHERIC OZONE, BELOW 50 KM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-1

21.1.1 Production . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-2
21.1.2 Transport . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-3
21.1.3 Ozone Measurement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-5
21.1.4 Total Ozone: Its Global Distribution and Variability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-8
21.1.5 Vertical Profiles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-9
21.1.6 Spatial and Temporal Scalesof Ozone Variability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-12
21.1.7 Mathematical Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-13
21.2 MINOR CONSTITUENTS IN THE STRATOSPHERE ................................... 21-14
21.2.1 Nitrous Oxide, FC-ll, FC-12and Methyl Chloride (NzO, CC13F,CClzFzand CH3Cl) . . . . 21-16
21.2.2 Carbon-ContainingSpecies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-16

xxv
TABLE OF CONTENTS

21.2.2.1 Carboti Dioxide (COz) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...” . . . . . . . . . . . 21-16

21.2.2.2 Carbon Monoxide (CO) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-16
21.2.3 Hydrogen Containing Species . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-21
21.2.3.1 Molecular Hydrogen (Hz) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-21
21.2.3.2 Methane(CHQ) .................................................... 21-22
21.2.3.3 Ethane, Propane, and Acetylene (CzH6, C3HS,andC*Hz) ................. 21-23
21.2.4 Stratospheric WaterVapor(H20) .............................................. 21-23
21.2.4.1 Satellite Measurements .............................................. 21-24
21.2.4.20therMeasurements. ............................................... 21-24
21.2.5 OddOxygen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .....””””””””.”.” 21-25
21.2.5.1 Atomic Oxygen(O(3P)) ............................................. 21-26
21.2.6 OddNitro~en . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...”””””””””””” 21-26
21.2.6.1 Nitric Oxide(NO) ................................................. 21-26
21.2.6.2 Nitrogen Dioxide(NOz) ............................................. 21-30
21.2.6.3 NitricAcid(HNOs) ................................................ 21-35
21.2.6.4 Nitrogen Trioxide(NOs) ............................................ 21-35
21.2.6.5 Nitrogen Pentoxide(NzOs) .......................................... 21-37
21.2.6.6 Peroxynitric Acid(HOzNOz) ......................................... 21-37
21.2.7 OddHydrogen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .....”””””.””””””” 21-37
21.2.7.1 Hydroxyl Radical(HO) ............................................. 21-37
21.2.7.2 Hydroperoxyl Radical(HOz) . . . . . ...-.........””””””””” ““””.”.”” 21-38
21.2.7.3 Atomic Hydrogen(H) .............................................. 21-38
21.2.7.4 Hydrogen Peroxide(Hz02) .......................................... 21-38
21-38
21.2.8 OddChlorine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...”””””.””””””” ““””””””
21.2.8.1 Chlorine Oxide(CIO) ............................................... 21-38
21.2.8.2 Hydrogen Chloride (Hal) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-39
21-39
21.2.8.3 Chlorine Nitrate(C10N02) ..........................................
21.2.9 OtherHalo~ens . . . . . . . . . . . . . . . . . . . . . . . . . . . . .......”””.”””””””.”..”””” ““”” 2’-@
21-40
21.2.9.1 Hydrogen Fluoride (H~) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

21.3 MESOSPHERE . . . . . . . . . . . . . . . . . . . . . . . . . ...”” . ..”””””””””.,”””””””””” ““”” 21-41

21.3.1 The Oxygen Constituents ................................................. 21-42

21.3.2 Carbon-Oxygen Constituents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21-44
21.3.3 Hydrogen-Oxygen Constituents. ............................................... 21-46
21.3.4 The Nitrogen-Oxygen Constituents ............................................ ~~:fi
21.3.5 Meteoric and Dust Constituents ...............................................
21.3.6 TheInertGases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...”” ““”””” 21-50
21.3.7 ExcitedSpecies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...””” ““”... 21-50

2i.410NIZEDCONSTITUENTS. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .........””””c..”””.”””””” 21-51

21.4.1 Overview ............................................................... 21-51

21.4.2 Positive Ion Distributions .................................................... 21-53
21.4.3 Negative Ion Distributions ................................................... 21-56

CHAPTER 22. ATMOSPHERIC EMISSION AND ABSORPTION

OF ULTRAVIOLET RADIATION

22.lvuv/UVBACKGROUNDATMOSPHERIC RADIANCE ................................ 22-1

22.2 ABSORPTION AND IONIZATION CROSS-SECTIONS OFMAJORGASES ............... 22-2

22.3 RATE OF PHOTODISSOCIATIONIN THE ATMOSPHERE ............................. 22-4

xxvi
 TABLE OF CONTENTS

CHAPTER 23. GEOKINETICS

23.1 STRUCTUREOFTHEEARTH ..................................................... 23-1

23.2 PLATETECTONICS ............................................................... 23-2

23.2.1 Driving Mechanism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23-5

23.2.2 Plate Motions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23-5

23.3 SHORT PERIOD MOTIONS—SEISMOLOGY ......................................... 23-6

23.3.1 Seismic Waves . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23-7

23.3.2 Earthquakes ............................................................... 23-11
23.3.2.1 Spatial Distribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23-11
23.3.2.2 Measurement ...................................................... 23-11
23.3.2.3 Focal Mechanisms and Fault Plane Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23-14
23.3.2.4 Temporal Distribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23-15
23.3.2.5 Hazard Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23-15
23.3.2.6 Premonitory Phenomena . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23-18
23.3.3 Other Sources ofSeismic Radiation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23-19
23.3.3.1 Seismic Noise . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23-19
23.3.3,2 Other Sources ofSeismic Radiation ................................... 23-20

23.4 LONG PERIOD AND SECULAR EARTH MOTIONS ................................... 23-21

23.4.1 Tectonic Motions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23-21

23.4.1.1 Horizontal Motions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23-21
23.4.1.2 Vertical Displacements .............................................. 23-21
23.4.1.3 Measurement Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23-22
23.4.2 EarthTides ................................................................ 23-24
23.4.2.1 SolidEarthTides .................................................. 23-25
23.4.2.20cean Loading Effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23-27
23.4.2.3 Cavity, Topographic and Geological Effects ............................ 23-28
23.4.3 OtherMotions ............................................................. 23-29
23.4,3.1 Motion Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23-29
23.4.3.2 Surface Loading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23-29
23.4.3.3 Thermoplastic Effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23-30
23.4.3.4 Soil Moisture and Groundwater ....................................... 23-31
23.4.3.5 Conclusions ....................................................... 23-34

23.5 ROTATIONALMOTIONS .......................................................... 23-34

23.5.1 Definition ofAzimuth . . .’. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23-34

23.5.2 Conventional Azimuth Determination .......................................... 23-35
23,5.3 Inertial Azimuth Determination ............................................... 23-37
23.5.3.1 Gyrocompassing ................................................... 23-37
23.5.3.2 Wheel Speed Modulation ............................................ 23-38
23.5.3.3 Ring LaserGyroscope .............................................. 23-38
23.5.4 Geophysical and Local Environment Effects on Inertial Instruments ................. 23-39
23.5.5 AzimuthDataBase ......................................................... 23-40
23.5.6 Conclusions ............................................................... 23-42

CHAPTER 24. GEODESY AND GRAVITY

24.1 GEOMETRICGEODESY ........................................................... 24-1

24.1.1 Horizontal Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24-2

24.1.2 Vertical Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24-6
24.1.3 Inertial Positioning . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24-6

xxvii
TABLE OF CONTENTS
24-8
24.2 GRAVIMETRIC GEODESY .........................................................
24-8
24.2.1 NormalGravityField .......................................................
24-8
24.2.2 DisturbingGravity Field . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
24-12
24.2.3 Gravimetry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...”””””””””.”””””” ““:”””
24.2.4 Gradiometry ............................................................... 24-12

24-12
24.3 SATELLITEGEODESY ............................................................
24-13
24.3.1 DopplerPositioning ....... . .. . . . . . . . .... . .. . . .. .............................
24-14
24.3.2 Satellite InterferometrY .... . .. . . . . . . . .... . .. . . .. .............................
24.3.3 LaserRanging ........... . .. . . . . . . . .... . .. . . .. . . ...-..”””””.””””””””” ““”’ 24-14
24-14
24.3.4 Satellite Altimetry . . . . . . . . . .. . . . . . . . .... . .. . . .. ..........”.”””””””” ~

24.4 GEODET1CASTRONOMY ......................................................... 24-16

24.4.1 LunarLaserRanging ........................................................ 24-18

24-18
24.4.2 Very Long Baseline InterferometrY ............................................

CHAPTER 25. INFRARED ASTRONOMY

25.1 OBSERVATIONALPARAMETERS ................................................. 25-1

25.1.1 Coordinate Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...”.”.””””””” 25-1

25.1.2 Time . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...”” ““”””””””””””” 25-2
25.1.3 Intensities—Magnitude . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...”..””””” “ 25-3
25.1.4 Distance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...””.”””” ““””””..””””” 25-4
25.1.4.1 Parallax . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...””””””””” ~~ 25-4
25.1.4.2 StandardCandles ................................................. .. 25-4
25.1.4.3 Dynamic Distances. .............................................. .. 25-5
25.1.4.4 Extragalactic Distances . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 25-5

25-5
25.2 THESOLARSYSTEM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .......””””””.”””””””””””

Objects with Negligible Atmospheres .......................................... 25-5

25.2.1
25.2.1.1 Mercury . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...”””””””” ““”””””” 25-7
25.2.1.2 TheMoon ........................................................ 25-7
25.2.1.3 Asteroids and Planetary Moons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25-7
25.2.2 Objects with Optically Thick Atmospheres ...................................... 25-7
25.2.3 Objects with Semi-Transparent Atmospheres .................................... 25-8
25.2.3.1 Mars . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...”” .“””” 25-8
25.2.3.2 Comets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...”””””””.” ““” 25-8
25.2.4 TheZodiacalLight . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...”” . 25-9

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..”” “ 25-10
25.3 THEGALAXY

25.3.1 Stars . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...” ““”””” 25-10

25.3.1.1 SpectralClassification .............................................. 25-13
25.3.1.2 Stellar Luminosities, Colors, and Temperatures .......................... 25-13
25.3.1.3 InfraredStars . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...”..”” 25-13
25.3.1.4 MStars . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .“ 25-13
25.3.1.5 CarbonStars ...................................................... 25-15
25.3.1.6 Planetary Nebulae and Related Objects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25-15
25.3.1.7 Luminous Stars with Infrared Excesses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25-17
25.3.2 GasandDust . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...”””””””.”””””” “o”” 25-17
25.3.2.1 InterstellarExtinction ............................................... 25-17
25.3.2.2 HIIRegions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ...”” ....”””””” 25-18
25.3.2.3 MolecularClouds ................................................. 25-18

...
Xxvlll
 3

TABLE OF CONTENTS
25.4 EXTRAGALACTIC OBJECTS ....................................................... 25-18

25.5 THE INFRARED CELESTIAL BACKGROUND ....................................... 25-19

25.5.1 2–8~m .................................................................. 25-19
25.5.2 8-30~m ................................................................. 25-21

APPENDIX—UNITS, CONSTANTS, AND CONVERSION FACTORS

INDEX

xxix
 Chapter 1

THE SUN
R.C. Altrmk D.F. Neidig
H.L. DeMastus R.R. Radick
J.W. Evans G.W. Simon
S.L. Keil, Capt., USAF

Radiant energy from the sun determines the surface tem- energy. To produce the solar luminosity, about 4.3 x 1012
perature of the earth and supplies virtually all the energy g/s must be converted to energy.
for natural processes on the earth’s surface and in its at- The energy generated in the core is transferred outward
mosphere. The sun is about 5 x 109 years old and geological by radiation. Because of the high densities, the radiation is
and paleontological evidence indicates it has shone with its absorbed and re-emitted many times on its outward journey.
present intensity for about a billion years. The solar radiation Radiation that began in the core as high energy gamma rays
is, however, not absolutely constant. There are short-period is degraded by these successive abso~tions and re-emis-
fluctuations in ultraviolet and x-ray radiation and in cor- sions until they finally emerge as visible radiation charac-
puscular emission that indt.tce complex responses in the earth’s teristics of the solar surface. However, the energy is not
atmosphere. In this chapter, solar characteristics and activit y carried all the way to the surface as radiation. As it moves
are identified; other chapters deal with the geophysical re- outward from the core, the temperature, density, and pres-
sponses themselves. We shall discuss first the quiet sun, sure drop. As the temperature drops, free electrons can be
then solar activity. trapped by atoms into bound states, causing an increase in
the opacity. Thus radiation becomes less effective in trans-
porting energy. A large temperature gradient results and

1.1 THE SUN’S “STEADY-STATE”

Table 1-1. Solar structure as function of radius [Allen. 1973]
BEHAVIOR (THE QUIET SUN)
r T P M, L, log P
Although all observable solar features change on time-
scales of seconds to days, the statistical description of many RO 103 km 106 K g/cmq MO L,, dynlcmz
such features taken together remains unchanged over sub-
0.00 0 15.5 160 0.0000.00 17.53
stantial periods of time. These “steady-state” features of the
0.04 28 15.0 141 0.0080.08 17.46
so-called quiet sun are described here.
0.1 70 13.0 89 0.07 0.42 17.20

0.2 139 9.5 41 0.35 0.94 16.72

0.3 209 6.7 13.3 0.64 0.998 16.08
1.1.1 Basic Characteristics
0.4 278 4.8 3.6 0.85 1.00 15.37
0.5 348 3.4 1.00 0.94 1.000 14.67
The sun is a relatively small, faint, cool star at an average
distance of 1.49 x 108 km from the earth. Its mass is 0.6 418 2.2 0.35 0.982 1.000 14.01
1.99 X 1033 g. Composition of the outer layers by mass is 0.7 487 1.2 0.08 0.994 1.000 13.08
approximately 7370 H, 25% He, and 2% other elements 0.8 557 0.7 0.018 0.999 1.000 12.18
[Allen, 1973]. Table 1-1 gives the run of temperature, den- 0.9 627 0.31 0.0020 1.000 1.000 10.94
sity, mass concentration, energy generation, and pressure
0.95 661 0.16 0.034 1.000 1.000 9.82
as functions of radius [Allen, 1973]. The total solar lumi-
0.99 689 0.052 0.045 1.000 1.000 8.32
nosity is 3.83 x 1033 ergls.
0.995 692.5 0.031 0.042 1.000 1.000 7.68
A schematic cross-section through the sun is shown in
1.000 696.0 0.006 0.0 I.000 1.000 —
Figure 1-1. Energy is generated in the solar core by nuclear
reactions. The primary reaction is the fusion of hydrogen T = temperature; p = density; P = pressure; M, = mass within radius
4; L, = energy generation within radius q M., ~, ~ = mass, radius.
nuclei to fom helium nuclei (4H1 + He4). For every gram energy generation of whole Sun, with M. = 1,99 x 1013g and
of hydrogen entering the reaction, 0.007 g is converted into Lo = 3.83 x 1033ergs.

1-1
CHAPTER 1
coma

~Photos*ere
chrornosmere

%% .. . .
,.

.: ..’,,,, ,. (’,,

‘ - “’,
,’.’
,.
..,,’
,, ’,, ,.
,., . ,.
Convection zone ‘.
. / ,.
>~ ...,.. .,.,

7\

‘“” M #’K”

@
\%\.
‘Y A

Figure 1-1. Idealized general solar properties, structure, and modes of outward energy flow. The features shown are not to scale and provide a qualitative
picture only [Gibson, 1973].

1-2
 THE SUN
convection becomes the primary energy transport mecha- diameter have led to the discovery of global solar oscilla-
nism. The exact depth at which this occurs is model de- tions. The observed periods agree approximately with the-
pendent. However, observations clearly show thatthiscort- oretically predicted radial p-mode oscillations. Table 1-2
vection zone reaches the solar surface and that at least 3 or compares some observed periods of global oscillations with
4scales ofconvectiort are present. Near the solar surface, predictions based upon a theoretical solh model [Christen-
radiation can escape into space and it again becomes the sen-Dalsgaard and Gough, 1976]. A summary of attempts
primary energy transport mechanism. to measure global solar oscillations can be found in Hill
Attempts to make accurate measurements of the solar [1979]. ~ese modes of oscillation can be used in a manner

Table 1-2. Observed and theoretical modes of solar oscillations [Christensen-Dalsgaard and Cough, 1976].

Periods of Solar Oscillations (rein)

1=0 1=2 1=4 1=2

1973 1975
Mode Period Mode Period Mode Period Mode Period
PI 62.7
(63.9) gl 55.1
52 47.9 f 46.0
PI 42.2 39.6
P2 43.8 (42.3) ;1 38.0 g2 61.5
33 P3 32.6
(31.8) P2 34.3 g3 70.9
30.3 P2 29.3
p4 26.0 P3 26.7 g4 81.8
23.8 P3 23.4
21.0 P5 21.0
[21.51 P4 21.5 gs 93.0
P4 19.5
P6 18.2 P5 17.9 ge 105
16.7 17.1 P5 16.3
P7 15.8 P6 16.0 g7 117
14.6 [15.4] P6 14.6
P8 14.0 P7 14.1 gg 130
13.3 P7 13.0
P9 12.5 P8 12.6 g9 142
11.9 11.8 P8 11.7
plo 11.3 P9 11.4 glo 154
(147)
10.4 10.5 pll 10.4 Plo 10.4 gll 167
(159)
P12 9.55 g12 180
(171)
9.2 8.8 p13 8.84
[9.121
P14 8.52
7.9 P15 7.99
7.6 P16 7.53
7.0 7.2 p17 7.12
P18 6.75
P19 6.41
1320 6.10
. . . . . . -------
I ne penoas neaaea I Y / 5 ana
-“.
IY /3 are
. . . . ...- Stebbms,
Irom observations by HI1l,
. . ..— and Brown [1975]
. .. —---and are accurate to
- 5%. Those periods not repotied in
the publication were kindly supplied by Profesor Hill. Periods in the other cohmms were obtained theoretically using a solar model with Z = 0.02, except
those in parenthesis which are periods of oscillation of a model with Z = 0.04. The periods in square brackets were obtained by repeating the computation
with Z = 0.02 using a different distribution of mesh points. (Reprinted with permission from Macmillan Journals Ltd. @ 1976. )

1-3
CHAPTER 1
analogous to terrestrial seismology to probe the deep interior amount of radiant energy at all wavelengths received per
of the sun. This should eventually permit us to refine the unit time and area at the top of the earth’s atmosphere,
model shown in Table 1-1 and place better constraints on corrected to the mean earth-sun distance, and is customarily
theoretical models of the solar interior. expressed in units of watts per square meter (W/mz). The
The total solar irradiance is quantitatively defined as the term solar constant is also used occasionally to denote this

THE SOLAR SPECTRUM

—FREQUENCY
021HZ 10mMz 1019Hz 1018Hz 1017Hz 101oHz 10ISHZ 1014Hz 1013Hz ldGHz 102GHz 10GHz 103 MHz l&Mt
I I I I I I I I I I 1 I I I
. s vlaE
GAMMA-RAY X-RAY ULTRA- * INFRARED
VIOLET ; lo~
1

‘\
10’
NASA-GODDARD SPACE FLIGHT CENTER
GREENBELT, MARYLAND
AUGUST, 1976
ld

10’

lB FLA

J v --\
●

A IQ2A IW’A 1A IOA $02A l@A lpm 10pm l@#m

WAVELENOTN —
1O-1I

lmm lcm 10cm lm 1

Figure 1-2. Spectral distribution of solar irradiance [White, 1977]

14
 THE SUN
radiant flux but is somewhat misleading, since the total solar I I 1 ( ) 1
0 x 1
irradiance seems to fluctuate slightly. The best measurement
to date of the total solar irradiance, 1368 t 7 W/m*, was
obtained in 1980 using radiometers aboard the NASA Solar
e
Maximum Mission (SMM) satellite.
More than 70% of the solar irradiance is concentrated
o 0
in the near ultraviolet, visible, and near infrared portions
of the spectrum lying between the atmospheric transmission 0

cut-offs near 0.32 pm and 1.0 pm. Since most of this .

radiation reaches the earth’s surface, a large fraction of its — Recurrent s.njpots

energy enters the lower atmosphere through the evaporation-

o
precipitation cycle of water. About 2% of the solar irradi-
?ho!ospher,c
ance appears at ultraviolet and x-ray wavelengths shorter m.agneiic fields -.

than 0.32 pm. This radiation is all strongly absorbed in the 0

upper atmosphere where it drives several important pho-
tochemical reactions, including ozone production. The re- .
1 1 I I I 1 1 1 1
mainder of the solar irradiance appears at infrared and radio 40~ 30~ zoo I o. 0. I o. 20~ 30° 40~ 50° S
wavelengths longer than 1.0 pm. Much of this radiation, HELIOGRAPHIC LATITUDE

especially in the infrared, is strongly absorbed by atmos-

pheric water vapor and carbon dioxide and thus contributes Figure 1-3. Solar differential rotation for sunspots (solid curve), large
scale photospheric magnetic fields (cimles) arsd quiet pho-
to the energy budget of the lower atmosphere. tosphere (crosses) [Wilcox and Howard, 1970].
Figure 1-2 illustrates the spectral distribution of the solar
irradiance. A complete description of the ultraviolet and
radio portions of the solar irradiance will be found in Chap- 1.1.3 The Solar Atmosphere
ters 2 and 11.
Because the sun is gaseous throughout, there is no rigid
boundary marking the start of the atmosphere and end of
the interior. However, a definite optical boundary, below
1.1.2 Solar Rotation which we cannot see, exists and is commonly referred to
as the solar surface. As can be seen from Table 1-1, the
The rotation period of the sun is a function of solar solar temperature decreases outward from the core until we
latitude, varying from about 24 days at the equator to 34 reach this optical surface. For a few hundred kilometers
days at the pole. This variation is called dlflerential rotation. above this level the temperature continues to decrease. How-
Taking into account the earth’s orbital motion, an effective ever, at about 400 km one reaches the temperature mini-
period of 27 days is usually satisfactory for predicting ro- mum, above which the temperature again begins to increase
tation recurrences of solar-terrestrial disturbances. Because with increasing height. Although no rigid boundaries exist,
the sun’s axis of rotation is inclined only 7° to the plane of it is convenient to divide the solar atmosphere into layers
the earth’s orbit, most of the solar surface can be scanned having characteristic temperatures and densities. Moving
in the course of one rotation. outward from the optical surface these layers are known as
When solar rotation is measured by following tracers the photosphere, chromosphere, transition zone, and co-
(features on the solar surface such as sunspots) one observes rona. Each of these regions is discussed below.
faster rotation rates than from the Doppler shifts of solar
spectral lines. Figure 1-3 shows the latitudinal dependence 1.1.3.1 The Photosphere. The photosphere is the ap-
of rotation rate measured by tracers compared to the Doppler parent solar surface, with a very sharply defined limb, seen
shift of lines [Wilcox and Howard, 1970]. Furthermore, by ordinary observations in white light. Its location with
Doppler measurements show that the differential rotation respect to other solar features is shown in Figure 1-1. Its
varies with time, and seems to have large eddies or cells diameter, usually considered to be the diameter of the sun,
imbedded in it. These temporal variations can be as large is 1.391 X 106 km. Practically all of the solar mass is
as the variation in latitude. Evidence has also been found contained within the radius of the photosphere. Figure 1-4
for toroidal oscillations superposed on the solar rotation shows the variation of temperature, density, and pressure
[Howard and LaBonte, 1980]. in a standard model of the lower solar atmosphere. Height
Differential rotation probably has its origin in the in- is given in kilometers above a depth in the atmosphere for
teraction between convection and rotation. This same in- which the probability y of a photon escaping is l/e. The thick-
teraction underlies most of the magnetic activity observed ness of the photosphere is only about 100 km, or 0.014%
on the sun (Section 1.2.2). of ~, where & is the solar radius. Below the photosphere

1-5
CHAPTER 1

10s- Io-=r

105 - 10-7-

104- 10-0

~; ,.3 - Ptot~l
mE Io-s
\ <
:2 E
: 10 ~ Io-’”
--e m
%1 c-
~-10 10-11- $

10°- 10-12-
Height (km)
10-’
- 10-’3
-
I 1 I 1 I
10=- ,.-141
-500 0 500 Im la 2000 L.

(b) Height (km)

(a)

Figure 1-4. Gas properties versus height (HSRA model) [Gingerich, 197 1]. (a) Temperature as a function of height. (b) Pressure, density, electron pressure,
and hydrogen ionization as functions of height.

the solar gas is opaque. This opacity is primarily due to a ules) into the photosphere gives rise to the granulm ap-
small concentration of negative hydrogen ions in the region pearance of the solar surface shown in Figure 1-5. Similar
immediately below the photosphere. These negative ions opacity barriers which exist deeper in the sun cause other
act as continuous absorbers over a great range of wave- larger scales of convective motions. The most obvious of
lengths, absorbing most of the intense radiation from deeper these is the supergranulation shown in Figure 1-6. Figure
layers in the sun. The resulting accumulation of thermal 1-7 is a schematic diagram of the various scales of motion
energy in this layer sets up a system of convective currents present in the solar atmosphere.
that transports the excess energy through the negative hy-
drogen barrier. The energy is then reradiated into the rel-
atively transparent gases above the photosphere and, hence,
into space. Penetration of these convective elements (grarr-

Figure 1-5. Photospheric granulation and sunspot structure photographed Figure 1-6. Velocity spectroheliogram obtained in Ca 6103 ~, showing
with 180 ft. focal length tower telescope at the National Solar supcrganulation structure on solar surface. Supergramdes have
Observatory at Sacramento Peak. diameters of about 30 Mm and lifetimes of 1-2 days.

1-6
 THE SUN

Size (km)
102 703 ?04 705 ?06 ?07
. I I I ‘r I
––Resofution–– ~~

Super
Granula tion Granula tion

AAeso
Granula
————- tion

Turbufen ce Gion ————

--—— t Cells
——-

coA/VEC TION ———---

-—
--
High Frequency Waves
Acoust/c Rotation
Gravity Meridional Flow

Pulsations, Free Oscillations

—-- Radial and Non –Radial -—

Figure 1-7. This schematic diagram shows the horizontal size of various motions present in the solar atmosphere. Also shown is an approximate range
for the resolution limit of ground-based observations. The smaller sizes indicated for resolution are only achieved using special techniques
such as speckle interferometry or active optical systems. Typical resolution for conventional telescopes is limited by turbulence in the terrestrial
atmosphere to approximate one arcsecond or about 720 km on the sun. The scales shown with the solid lines indicate an approximate range,
while the dashed lines indicate sizes either predicted theoretically or that have been only tentatively observed.

Because the opaque gas composing the photosphere ab-

sorbs and reradiates approximately as a black body, the
photosphere emits an essentially continuous spectrum, or
continuum. Sharply tuned absorption by the various atoms
in and above the photosphere give rise to thousands of dark
lines (Fraunhofer lines) supe~osed on the continuum. This
line absorption spectrum, shown in Figure 1-8, is the source
of most information on the abundance of the various ele-
ments and the physical state of the solar atmosphere. Al-
though the lines of a number of atoms are unobservable
through the terrestrial atmosphere, all the natural chemical
elements known on earth probably exist in the sun. Their
relative abundances, however, differ radically.
The principal region of line absorption is a layer about
4W km thick just above the visible surface of the photo-
sphere. Spectra with high spatial resolution on the sun as
shown in Figure 1-9 show small-scale wiggles due to Dop-
pler shifts induced by mass motions in the line formation
region. Analysis of these wiggly lines has revealed several
physical processes occurring on the sun.
The granules of Figure 1-5 are evidence of convective Figure 1-8. The Fraunhofer line absorption spectmm of the solar atmos-
overshooting into the stable photosphere from deeper layers. phere.

1-7
CHAPTER 1
1.1.3.2 The Chromosphere. The chromosphere is a iayer,
several thousand kilometers thick, of transparent glowing
gas above the photosphere. Many of the phenomena oc-
curring in the photosphere also are manifest in the chro-
mosphere. Because the density in the’chromosphere con-
tinues to decrease with height (Figure 1-4) and is much
Iower than in the photosphere, magnetic fields and waves
can have greater influence on the structure. Thus the chro-
mosphere is even more inhomogeneous than the photo-
sphere. Its continuous spectrum is extremely faint, and like

—
,&,A (arc*c) ,8
5
0.0393 I f .67
*e-

Figure 1-9. Thespectral region around the MgB lines at5167~ and
5172 ~ is shown. The verticle axis is wavelength and the
horizontal axis is distance along the slit. The dark horizontal
bars are absorption lines.
;} P3 ,20

An average granule has a diameter of about 1.4 arcsec w /}p2

or 1 Mm. Granules typically live for 8 minutes and
reach velocities near2 km/s. Most granules lose their tern-
l.tm
perature excess after rising only 50 to 100 km into the
:} P,
photosphere, buttheir momentum camies them 1000r 200
km higher.
The supergranules shown in Figure 1-6 penetrate higher into ~
g
the atmosphere but have much smaller vertical motions than
.33
granules. They are primarily observed as horizontal flows
B
in the chromospheric layers above the photosphere. Super- f
granules have diameters of approximately 30 Mm and life-
times of 1-2 days. Evidence has recently been found for
two other scales of convection: mesogranulation [November
et al., 1981] with a scale of 5-10 Mm, and giant cells perhaps .00

100 Mm in diameter.
In addition to revealing various modes of convection, wiggly
line spectra show that the sun oscillates in a number of =0.0817
different modes. Oscillatory modes with large coherent wave
systems having periods near 300 s are observed in the pho- 5.00

tosphere and overlying chromosphere. Such data are com-

pared with a theoretical non-radial acoustic spectrum in
Figure 1-10. These modes have most of their energy con-
centrated in the underlying convection zone, and deviations
of predicted modes from observed modes can lead to better
models of the convection zone. The 300 s oscillations are
evanescent in the photosphere (below the temperature min-
Rgure I -10. Comparison of an observed k-~ diagram with theoretical
imum shown in Figure 1-4). At the higher temperatures Eigenfrequencies for two model solar envelopes [Rhodes et
found in the chromosphere, the waves become propagating al., 1977], Contour intervals are 0.125, 0.25, 0.75, 1.75,
and higher frequency acoustic waves are observed. The etc. mzs-z, Observations were made on the magnetically
insensitive line Fe 1 A 5576.10 using the vacuum tower
mechanical energy in these waves may be responsible for telescope at the National Solar Observatory at Sacramento
the temperature increase in the chromosphere and higher Peak. (Reprinted with permission from the American As-
layers. tronomical Society 0 1977. )

1-8
 THE SUN
all solar features above the photosphere, it can be observed
normally only by isolating the light of one of its strong
lines, usually the red Ha line at 6563 ~ (Fraunhofer C line)
or the line of CA 11 at 3934 ~ (Fraunhofer K).
The rise in chromospheric temperature with height can
be understood in terms of a nonradiative energy input into
the atmosphere via mechanical energy generated in the con-
vection zone. Acoustical waves propagate outward and
can form shocks in the low density chromosphere, creating
more energetic collisions between particles. The low den-
sity also leads to a relative enhancement in the ability of
magnetic fields to carry energy in the form of waves.
These mechanisms combine to heat the chromosphere
from approximately 4300 K at its base to over 20000 K
in about 2 Mm.
Similar to the granulation patterns observed in the pho-
tosphere, a chromospheric network can be observed in chro-
mospheric lines. Figure 1-11 shows the appearance of this
network in Ha and Ca II. The scale of this network cor-
responds fairly closely to that of supergranulation, and the
bright regions of the chromospheric network correlate well
with supergranular boundaries. At sub-photospheric den- Figure 1-11. Ca II spectroheliograms made on the violet side of Ca 11
sities the supergranules are capable of pushing magnetic line at 3934 ~. The network of bright emission is clearly
flux tubes to their boundaries while in the chromosphere the evident, and outlines the houndaries of supergranules seen
in Figure 1-6.
fields dominate the gas. If heating occurs preferentially near
the field lines, the appearance of the network can be ex-
plained. The network is also the location of spicules which
are brilliant flames of gas that project upward for approx-
imately 10 Mm and live for 2 to 5 minutes. Figure 1-12
shows how these spicules appear at the limb of the sun. than on the quiet sun. EUV photographs of the transition
Spicules may provide the primary path for moving energy zone obtained by Skylab are shown in Figure 1-14.
and material into the corona. For further information on this region, including models,
see Jordan [198 1], Zirker [1977], Athay [1976], and Bonnet
[1981].

1.1.3.3 The Transition Region. At the top of the chro-

mosphere, approximately 1.7 Mm above the temperature
minimum, the temperature begins a sudden steep rise to 1.1.3.4 The Corona. Physical Properties: Above the
coronal values. In the next few megameters it rises from transition region and extending out into interplanetary space
25000 K to near 2 x 10G K. This steep temperature rise is the faintly visible solar corona. Long seen only during
is called the transition region. Because this transition occurs the few seconds or minutes of total solar eclipses, the true
in such a narrow zone of t;~e solar atmosphere, it is very nature of this outer atmosphere has only in recent years been
difficult to measure the physical height boundaries accu- recognized. Dominated by temperatures of one to two mil-
rately. It is also very hard to observe this region, since its lion degrees, magnetically y-induced motions, sudden re-
spectral lines are in the EUV and can be studied only from leases of energy, and explosive expansion into the surround-
satellite and rocket observations. Thus the amount of data ing vacuum of space, this entity has strong implications for
available to date is very limited. We show in Figure 1-13 the energy balance of near-solar space and planetary envi-
a profile of temperature vs height from the photosphere ronments.
to the low corona with some transition zone spectral The corona is heated from below by mechanical, elec-
lines indicated. The slope of the rise and the location of its trical or magnetic dissipation. The exact nature of the heat-
onset are functions of solar activity, being steeper and ing has not yet been discovered. Favorite candidates are
lower in the atmosphere in active regions (Section 1.2.3.1) shock waves produced by the upward surging of granules

1-9
CHAPTER 1

+1+

+1

.?

+~
2

+;

o;

-i

--
i

--3
4

-1

-1+

Figure 1-12. Spicules at the limb observed in Ha. These narrow tongue-like flames extend 5-19 Mm above the limb and live for 2–5 minutes.

1-1o
 THE SUN

T( K) ronal holes, magnetic loops, prominence cavities, and time-

dependent variations in coronal transients.
106
Table 1-3 presents a simple model of a coronal hole and
the quiet corona. Further information should be sought in
Zirker [1977], White [1977], Allen [1973], Eddy [1978],
and Jordan [1981].

1 0 VI

~05
1$ c Iv

Si IV
I
Table 1-3.

I r/~
Model of corona in quiet sun and in coronal hole showing
electron density vs height above the solar limb (~ = 1.0)
[Zirker, 1977].

Quiet Sun
(T = 2.0 X 10’ K)

N. (cm-’)
Coronal Hole
(T = 1.5 X 106K)

N. (cm-3) I
LY~ (WING) 1.1 1.0 x 108 5.4 x 10’
CII - 1.2 7.1 x 10’ 1.6 X 107
MGII H&K 1.4 2.3 X 107 2.8 X 106
104 2.0 2.8 X 106 2.0 x 105
4.0 8.9 X 104 4.0 x 103
b 10.0 8.0 X 103

Lp
Methods of Observation: Coronal observations are ex-
tremely difficult because of the requirement to eliminate the
million-times-brighter light from the solar disk. Until the
development of the coronagraph in the 1930s, the only ob-
103
3000 2000 1000 0?M servations of the faint coronal light (about as bright as the
full moon) were made during total solar eclipses.
Figure 1-13. An average quiet sun temperature distribution from photo-
In the standard coronagraph, light spilling around an
sphere (O km), through the temperature minimum (400 km), occulting disk is either passed through the narrow slit of a
chromosphere (500-2000 km), and transition zone (2000- spectrograph to form a spectrum of the faint emission lines
25W km), and into the low corona (3000 km up), derived
from Vemazza et al. [ 1981].
or through a narrow-band birefringent filter to form an image
-.
ot the corona in an emission line, most often the green line
of Fe XIV (5303 ~), and at times the red Fe X (6374 ~),
into the photosphere or by propagating waves or oscillations or yellow Ca XV (5694 ~) lines. These lines have ionization
induced by them or by spicules. Other recent speculation temperatures of 1.5, 1.0, and 3.4 X 106 K, respectively.
has centered on the dissipation of Alfv6n waves being prop- Figure 1-16 shows two filtergrams of the corona.
agated upward from the turbulent photosphere along mag- Photoelectric coronal observations from the ground are
netic lines of force. The end result of this heating, whatever of two types. Observations of the emission-line corona made
its source, is an increase in temperature to near 2 x 106 K with a differential photometer at the National Solar Ob-
in the low corona. The temperature decreases slowly above servatory at Sacramento Peak in New Mexico consist of
that level. This extreme heat causes the outer parts of the circular scans around the limb at two or three heights. A
corona to “boil away” into the near-vacuum of interplanetary sequence of such daily scans can be used to build up a
space. Coronal heating is reviewed in Jordan [1981]. picture of the low corona over a half solar rotation (Figure
Observations of the corona (Figure 1-15) indicate that 1-17). Observations of the intermediate corona using light
it is high] y inhomogeneous. Different regions, such as scattered by coronal electrons (also called the “K” corona)
streamers (radially-elongated bright areas-Section 1.2) and are made daily with the K-coronameter on Mauna Loa,
coronal holes (dark areas—Section 1.2), show small vari- Hawaii.
ations in temperature (on the order of 3070) from area to The corona can also be observed with radio telescopes
area, but large variations in density, up to one to two orders (Chapter 11). Here the corona can be seen directly against
of magnitude. These variations become more pronounced the disk (there is no complicating “background light” as in
with height. the visible), and different heights can be probed by using
Models of the corona are controversial and complex. different frequencies.
Early one-dimensional models have now given way to two- The discovery of coronal holes (Section 1.2.7.1), one
and three-dimensional models representing streamers, co- of the major advances in solar physics in the past decade,

1-11
 \

.*- .

-.—.
.

$ I

Figure 1-14. Spectroheliograms (EUV) made with a slitless spectrograph of lines in the transition zone obtained with Skylab, 1973–74. The Magnetogram
is a ground-based observation.

Figure 1-15. The Solar Corona. This photograph was taken at a solar
eclipse with a filter that allows progressively more light to Figure 1-16, Two filtergrams of the solar limb as seen in the coronal lines
enter the camera as distance aw~y from the solar limb in- Fe X, 6374 ~ (above) and Fe XIV, 5303 ~ (below) obtained
creases. This greatly enhances the detail to be seen. Venus with the new “one-shot” coronagraph at the National Solar
is visible on the left side (High Altitude Observatory). Observatory at Sacramento Peak.

1-12
 A

THE SUN
VISIBLE CORONRLoIsa ON my 98.s0s snc. PE8K 6REEN-LINE IN7mITy R7 a ● f.is

IM7EN817Y
C*1.E6)
1>18

8<1<16

1<1

Figure 1-17. A “picture” of the ~en-line corona on 7 Feb 1984 obtained by allowing successive east-limb scans to rotate across the disk. Note the (black) pob
coronal holes from which high speed solar wind streams flow. (National Solar Observatory at Sacramento Peak).

was made possible by satellite observations. The lack of a near-solar interplanetary medium (1PM) is the scintillation
scattering and absorbing atmosphere at satellite altitudes of radio stars as observed through the corona/IPM. Semi-
made possible the first prolonged high-resolution solar ob- empirical relations between the signal fluctuation of a radio
servations in soft x rays. These Skylab observations and star and the turbulence generated at a given point in the
others revolutionized the understanding of coronal pro- 1PM by solar wind flow (Chapter 3) have allowed investi-
cesses. Astronomers also use satellite images of the corona gators at the University of California, San Diego (UCSD),
obtained in the extreme ultraviolet (EUV) with Skylab, and to produce low-resolution three-dimensional pictu~s of the
coronagraph observations in white light and emission lines solar wind flow within one astronomical unit (the distanpe
both from Skylab and the Solar Maximum Mission satellite from the etih to the sun) of the earth.
(SMM).
Indirect observations of the corona can be made on the
disk in some spectral lines of helium, most notably that of 1.2 THE ACTIVE SUN
10830 ~. The population of certain atomic states of chro-
mospheric helium is partially controlled by conditions in Although the quiet sun is of immense astronomical in-
the overlying corona, and so coronal holes are marginally terest because it is the only star whose spatial features are
detectable in this line. A daily 10830 ~ patrol (Figure 1- readily resolvable with earth-bound telescopes, most of the
18) is operated at the National Solar Observatory facility sun’s geophysical effects (except for the basic solar irra-
on Kitt Peak, Arizona. diance) result from solar activity. In this section we shall
Another source of data concerning the corona and the discuss phenomena related to the sun’s magnetic field, for
1-13
CHAPTER 1
more trailer spots to the east (Figure 1-20); the leader spot
is often the largest spot within the group. Magnetic fields
in large spots are typically 3000 gauss, with leader and
trailer spots generally having opposite polarities. During an
1l-year cycle all leader spots in the northern hemisphere of
the sun have one magnetic polarity, while in the southern
hemisphere this polarity is reversed. This pattern persists
throughout the rise and decline of the cycle. During the next
cycle, however, the entire polarity pattern is reversed. Thus,
the fundamental cycle of solar magnetic activity is actually
about 23 years in duration. During cycle No. 21 (maximum
in 1979-80), ieader spots in the northern hemisphere had
positive polarity (magnetic vector pointing outward from
the solar surface).
The first sunspots of a new cycle typically appear at
latitudes of 20° to 25° in both northern and southern hemi-
spheres. As the cycle progresses, additional spots appear at
Figure 1-18. A sample 10830 ~ scan obtained by the National Solar lower latitudes, until they are concentrated near the equator
Observatory at Kitt Peak on 24 Dec 1978. Active Regions at the end of the cycle (Figure 1-21).
and prominences are black and coronal holes are light. De-
Major variations in coronal structure take place during
fining a coronal hole usually requires a series of pictures
taken over several days as well as information about the the solar cycle. Several decades of observations have con-
magnetic field. firmed that green-line intensities, and to some extent the
red-line intensities, follow the sunspot number quite well.
In particular, averages of the green-line intensities show
it is probably through the emergence, growth, transport and
large maxima at midlatitudes in both hemispheres that ap-
destruction of magnetic flux that the sun produces the tran-
pear near or coincident with sunspot latitudes. These belts
sient behavior which causes large responses in the earth’s
move towards the equator with the progression of the solar
environment.
cycle and finally merge.
The total variation of the green-line intensity during a
solar cycle is significant. In addition, the appearance of the
1.2.1 The Sunspot Cycle two yellow lines is now known to be an excellent indicator
of regions which will produce large flares and perhaps even
The most obvious manifestation of solar activity is the proton events.
1l-year sunspot cycle. The relative sunspot number R (or
Wolf ‘or Zurich number) remains the single most important
index for the general level of solar activity; it is calculated 1.2.2 The Solar Dipole Magnetic Field
according to the formula
The sun has a weak dipole magnetic field aligned with
R = k(n + IOg), its axis of rotation. Because of the stronger, complex field
patterns associated with spot groups, however, this dipole
where n is the number of individual spots visible on the field is recognizable only at latitudes above 60°, where it
solar disk, g is the number of sunspot groups (see below), measures about 1 gauss on the solar surface. It is believed
and k is a station constant, or “personal equation” for the that the dipole field, although weak in general, becomes
particular observatory. concentrated in the midlatitudes due to the dynamo action
Records of sunspot counts have been kept since the mid- of the sun’s differential rotation. Magnetic ropes of strong
seventeenth century (Figure 1-19). The cyclic behavior of field may then rise to the surface due to magnetic buoyancy,
R with a period averaging 11.4 years is the most obvious forming the oppositely-poled leader and trailer portions of
feature in the historical record, although an approximately an active region where the flux loop intersects the photo-
80-year periodicity may also be present. The anomalous sphere. This would account for the spot-polarity law dis-
period of low activity, 1645-1700 (Maunder mimimum), is cussed above. The loops may eventually rise into the outer
recognized as a possible reduction in the level of solar ac- corona where they neutralize the original dipole field and
tivity; its existence has been independently supported by ultimately replace it with a new dipole field of opposite
measurements of the carbon-14 content in tree rings formed sign. The polarity reversal of the dipole field occurs near
during that era. sunspot maximum although the spot polarities themselves
Most sunspots occur in groups that are dominated by a remain consistent with the polarity of the dipole field that
leader spot on the western side of the group and one or existed at the beginning of the cycle.

1-14
 THE SUN

ANNUAL MEAN SUNSPOT NUMBERS FROM 161 OTO THE PRESENT

60 ‘
40
20

1670 1680 1690 1700 1710 1720 17:0

160
140
1 120
1 lm
80
60
40
20
0
1730 1740 1750 1760 1770 1780 1790 1800 1810 1820 1830 1840 1850

I I \ 180

“ 160

140
I \ =140

120 I A120
lCK)’
I 1 ) 100
A
80 80
1
60 ~ 60
)
40‘ 40

20‘ 20
01 u .J v. u v v. v v v
Jo
1850 1860 1870 1880 1890 1900 1910 1920 1930 1940 1950 1960 197o

fi~re 1-19. Annual mean sumspot numbersfrom 1610 to1975 (From Waldmeier [1961J and Eddy [1976] etal.)

1-15
CHAPTER 1
lengths of 1 to 5 Mm and narrow widths of 100 to 200 km.
They may, in fact, be smaller, but it has been impossible
to date to see smaller structures with solar telescopes. Fil-
igree reside in intergranular lanes and are visible with ground-
based telescopes only on rare occasions ‘when the atmos-
pheric seeing allows the 1/4 arcsec resolution required to
see them. Their locations are presumed to coincide with
extremely fine magnetic structures occupying the granule
boundaries, perhaps analogous to the somewhat larger flux
tubes of the supergranule network (Section 1. 1.3.2). Meas-
urements show that these small flux tubes have intense mag-
netic fields with strengths between 1000 and 2000 gauss.

1.2.3.1 Active Regions. fie development of a large-scale

sunspot-bearing active region usually begins with an in-
crease in the number and intensity of magnetic flux tubes
in the photospheric network. In response to this, a bright-
ening in the overlying chromosphere develops. These
~gure 1-20. Typical bipolar sunspot group showing leader spot (right)
and trailer spot. The background pattern of granulation is brightenings, or plages are the chromospheric counterparts
also visible (National Solar Observatory at Sacramento Peak). to the sunspt group that later becomes visible in the pho-
tospheric layers. The corona responds also, producing a
large area of bright emission visible, for example, in x-ray
1.2.3 The Emergence and Organization of and radio images (Figure 1-23). Within a day of plage for-
Magnetic Flux mation a small leader spot emerges which continues to grow
in size for several days. The trailer spots usually form within
The solar surface abounds with both large and small- a day of the leader, and many smaller spots ofien form
scale magnetic features. The smallest features, known as throughout the entire region (Figure 1-20). Generally a sun-
filigree (Figure 1-22) are bright, wiggly structures with spot begins life as a small pore which darkens as a con-

I I I I I 1 I I I I I I I
0 ,
40’ - 0. . e
>“” . ..” .
“. * : ?“.’”. ●:&e
.@ * “. ●*** “ *
.
30” - . . •@~ ~- 0- 0. L :;”. ?:* “,”/
“ :*.”:. : *.
. *. “. ““f>g%g,+
●
●
. ●S. *
● .
3 ““~c V“-””:””?? ““” ““”” .-e.
“.
209 - :4 ‘. m“2@*.:
9*”*
. #:,-
.%2;:.
● ●
?% ~~1.. .: ~
●** “ “;+;~r
●
::* ● ●

“ ●
U .*”
a ,4
~ ,; ‘.-. “. +“”” “ “” “; “!fi “ .8*4“. “-:;.
. “ “. -::: “
●
~~:
“s”. “.”~. . ;- @ b d“ ;e~-. .. ---
.. “.
E f*”” “ .“. . .. ‘* 4.“. .*%: .:%e ~t;xt:b“ 9...0 ..~ee*.. .
8 , ●.
●

“.
●
. .-
● ●

J O* .“” .
. “e. . :.d~” s: *“ “: : ‘:i$”{<.:~?$:’h ‘“ .“- ..:

●
●
. ●

..” ●*.*.,*8**”
●

z ““ ● ● ●
~ I& ;’ “
% e“”:”; - d ,~~”e:dl:~:.f”:” “‘“ : “ ~ [“;.: ::)?
2e -. ::44: : ● #.@ ‘“:”” 0 $
;-j~+~.:~
~a
“. ●
s “i~~”: ● ●
{A.”> .- ..; ;
.“• {?J”~.& , &.” ~: .;4”:. “. ~e::;2.0.
.&* “ ●“ -
30 “-”* “::*~”$ :.b”~,$*:
● *.* ● : “ ●*.. “ ‘*** ●

“*”%8 .* “~..” . %: 8..

4
4@ - MIN. ● . MAX.
.
“
.
““” MIN. “ .= .c”” *” ●
. 1 . ●

! 1 I I 1 1 i I 1 I i i
1933 1935 1937 1939 1941 1943 1945 1947
YEAR

F]gure 1-21. Typical “buttefly” diagram showing variation in distribution of sunspots with time. During a sunspot cycle the distribution migrates toward
the solar equator (Mount Wilson and Las Campanas Observatories).

1-16
 THE SUN

Figure 1-22. The solar filigree are visible as bright, wiggly structures
lying in the intergranrde zones. This series of filtergrams
was taken in different wavelengths (clockwise from upper
left): continuum, Ha + 2 ~ Ha + 7/8 ~, Ha – 7/8 ~
(National Solar Observatory at Sacramento Peak),
Figure 1-24. Complex sunspot group (National Solar Observatory at Sac-
ramento Peak).
sequence of temperature reduction. The temperature drop
is probably caused by the presence of the magnetic field,
which may inhibit the supply of convective energy to the is the last to dissolve and occasionally may last for several
photosphere at that location, and/or may enhance the pho- months. The chromospheric plage persists even after all
tospheric dissipation of energy at that point. The largest spots have disappeared.
spots attain areas of 3 x 109 kmz and temperatures 2000 A small number of active regions violate the simple
K lower than the surrounding photosphere. polarity rule, and instead form complex magnetic structures
Sunspot groups usually reach maturity within several with reversed or mixed polarities (Figure 1-24). These groups
days after which they begin a slow decline. The leader spot are responsible for nearly all the highly active solar phe-
nomena (Section 1.2.6).

1.2.3.2 Ephemeral Regions. A significant portion of the

total magnetic flux on the solar surface resides in ephemeral
active regions. These are miniature bipolar regions of size
-30 Mm with average lifetimes of 12 hours. Although not
usually associated with sunspots of their own, they are,
nevertheless, legitimate active regions in the sense that they
affat the chromospheric struc~s and develop visible plages.
There may be several hundred such regions on the sun at
any given moment, although like spots their numbers vary
with the 11-year cycle. Their magnetic orientation, how-
ever, appears to be far more random than the leader and
trailer polarities of sunspot groups.

1.2.4 Prominences

Prominences are large clouds of gas high above the

chromosphere, visible as bright flame-like objects with
emission spectra against the sky background at the limb, or
as long rope-like dark filaments with absorption spectra
against the solar disk. Quiescent prominences (Figure 1-25)
Figure 1-23. Soft x-ray photograph of the sun, showing the bright Iwp
structures of active regions (John Davis, American Science are not associated with active regions, but rather form
and Engineering, Inc. ). boundaries between very large-scale weak magnetic features

1-17
Figure 1-26c. Post-flare loop prominence (National Solar Observatory at
Sacramento Peak).

in the quiet sun. They are the largest visible structures on

the sun, sometimes reaching lengths exceeding a solar ra-
dius. Occasionally, quiescent prominences suddenly be-
come active, erupt and blow off to great heights in the
corona.
Other types of prominence-related phenomena include
surges, sprays, and loops (Figure 1-26). These structures
are often associated with flares as discussed below. They
are, nevertheless, similar to quiescent prominences in the
sense that their temperatures (10 000 K to 20000 K) remain
chromospheric even though their structures extend to co-
ronal heights where the ambient temperatures are 100 times
higher than their own.

1.2.5 Coronal Magnetic Fields

Careful observations also show a great spatial variation

in the nature, if not the magnitude, of the magnetic field
that threads its way through the corona. Bright regions tend
to correlate with closed-jield regions where models or ob-
servations indicate that the magnetic lines of force emerge
from and re-enter the photosphere over scales less than a
solar radius (Figure 1-23). Conversely, dark regions cor-
relate with open-field regions, mysterious magnetic mon-
opoles whose field lines may re-enter the sun again only
after some convoluted journey through interplanetary space.

1.2.6 Flares

A flare is a transient, localized release of energy usually

Occurnng in and above an active region. Flares produce
greatly enhanced emissions spanning the entire electromag-
netic spectrum and may accelerate particles to energies as
high as 50 GeV. The flare eruption affects the corona, chro-
 THE SUN
mosphere, and probably even the upper photosphere. The
energy released by a flare ranges from 1029 to 1032 erg;
durations vary from 1 to 100 minutes.
Flare importame has traditionally been assigned ac-
cording to the flare’s maximum area as observed in the Ha
line (Table 1-4). Following the importance class, a letter

Table 1-4. Flare optical impmtance

Area
Importance (Heliographic Square Degrtis)

S (Subflare) 2.06
1 2.06 to 5.15
2 5.15 to 12.4
3 12.4 to 24.7
4 >24.7

code (F, N or B) is given to designate the relative brightness “’’”’”+

of the flare as faint, normal, or bright. In recent years flare
importance has been supplemented by the soft x ray clas- Figure 1-27. Soft x-ray loop structures in a flare (American Science and
Engineering, Inc. ),
sification of the flare (Table 1-5). These x ray classes are
only roughly correlated with the optical importances.

fined to the footprints of these loops forming intense flare

Table 1-5. Flare x ray classification
kernels, or an arcade of loops may be involved forming a
pair of serpentine, parallel ribbons. Magnetograph obser-
1-8A x ray flux at IAU
vations reveal opposite magnetic polarities for the two mem-
Class (W/m2)
bers of a kernel or ribbon pair. Photospheric field strengths
c l~<E<lo-5 for the flare loop footprints are typically 1000-2000 gauss
M l@5<E <l@
x E>l&

Historically, flares have been studied in the H(Y line,

which is formed in the chromosphere. Observations in the
EUV and x ray regions, however, have shown that flares
may be more correctly characterized as coronal phenomena
with the primary energy release probably occurring in the
low corona or transition region. It is not surprising, there-
fore, that flares are best seen in wavelengths originating in
the corona. Flare images in soft x rays (1-10 ~) and XUV
lines (200-300 ~) obtained by the Skylab Mission show that
the basic flare structure is a low-lying coronal loop on the
order of 20-50 Mm in length (Figure 1-27), wherein tem-
peratures rise from the normal 1 x 106 K to about 20 x ld
K. Complex flares often reveal the presence of many such
loop structures. The loop structures are formed by magnetic
fields rooted in the deeper photospheric layers. The chro-
Figure 1-28. Hcs filtergram showing the cbromospheric component of a
mospheric flare (Figure 1-28), which is probably a second-
“’two-ribbon” flare (National Solar Observatory at Sacra-
ary response to the coronal energy source, is primarily con- mento Peak),

1-19
CHAPTER 1
for highly energetic events. Only rarely does the flare dis- flare emissions are responsible for a variety of immediate
turbance affect layers deep enough in the atmosphere to geophysical effects collectively known as sudden iono-
produce optical continuum (so-called white-light flares). spheric disturbances (SIDS) (Chapter 10). These distur-
About half of all flares show a sudden or explosive rise bances are a result of increased ionization at several levels
to maximum. This impulsive phase is most clearly defined in the ionosphere.
by hard x rays (photon energies 10 KeV - 1 MeV) or mi- Shock waves produced by explosive heating propagate
crowaves (wavelengths 1-30 cm) and may last only a minute upward through the corona where they accelerate electrons
in the case of simple flares. Soft x ray and Ha emissions, and protons to relativistic energies. The interaction of shocks,
even in impulsive flares, are usually less abrupt in their particles, and magnetic fields generates Type II and Type
onset, and often continue to rise after the impulsive com- IV meter-wave radio bursts (Chapter 11). The outward-
ponents are in decline. Complex flares show multiple im- propagating shock (velocity – 1000 km/s) is apparently also
pulsive phases, with each impulse apparently being asso- the origin of solar protons (energies >10 MeV) that produce
ciated with a particular point in the flare. The impulsive terrestrial polar cap absorption (PCA) and solar cosmic ray
components probably originate in a process which accel- ground-level events (GLE), while the shock wave itself ar-
erates electrons to energies exceeding 10 KeV, producing riving at the earth in about two days is a cause of the sudden
hard x rays via electron-proton bremsstrahlung and micro- commencement phase of geomagnetic storms (Chapter 4).
waves via gyrosynchrotron radiation in the ambient mag- The origin of flares, particularly the processes of energy
netic fields (Chapter 11). Type III meter-wave radio bursts storage and release as well as the triggering instability,
(Chapter 11), which may also occur during the impulsive remains a difficult theoretical problem. Obviously, the local
phase, have been attributed to the outward escape of some magnetic field is involved. Recent theories point to current-
of these electrons along open field lines. Recently obtained carrying magnetic configurations in which field lines re-
hard x ray images show the early stages of the hard x ray connect and simplify. The problem is observational] y dif-
bursts occurring at loop footprints similar to the first optical ficult because the flare plasma parameters cannot be tracked
kernel brightening. This suggests a common energizing in sufficient spatial and temporal detail. Without the aid of
source in the loop, from which electrons are guided down- in situ measurements we are forced to rely upon observations
ward into the chromosphere along the field lines. It is not of secondary effects which detect the flare energy only after
clear whether the population of electrons producing the im- disordering processes have already occurred.
pulsive emissions is best characterized as thermal or non-
thermal. Certainly the hard x ray kernels and Type III radio
bursts suggest a non-thermal directivity in the particles. If
the energy distribution is thermal, then temperatures ex- 1.2.7 Coronal Activity
ceeding 108 K, and possibly even 109 K, would be required
in order to produce the observed hard x ray continuum.. The 1.2.7.1. Coronal Holes and Solar Wind. During Sky-
presence of protons with energies exceeding 30 MeV, as lab, pictures of the sun were taken regularly with a soft x-
evidenced by nuclear gamma-ray lines in the impulsive phases ray telescope [Timothy et al., 1975], while at the same time
of large flares, lends additional support to acceleration pro- 27-day recurrent geomagnetic disturbances were occurring.
cesses which produce non-thermal distributions. It was noticed that when areas dark in x rays crossed the
A variety of phenomena occur in response to violent central meridian near the sub-earth point, a geomagnetic
heating or magnetic rearrangement associated with flares. disturbance occurred a few days later. This relationship
Surges (Figure 1-26) are spikes or filaments of chromo- persisted, and thus the source of the disturbances, the long-
spheric material sent outward (velocity - 100 km/s) by a puzzling solar M-regions, was finally discovered, solving a
sudden change in the distribution of magnetic field lines in riddle that had existed for decades. These dark areas named
the active region. Surprisingly, many surges are not asso- coronal holes (Figure 1-29) were shown to be sources of
ciated with flares, although conversely a large fraction of high-speed solar wind streams (HSS) that eventually im-
flares seem to generate surges. The material pushed upward pacted on the magnetosphere and caused the disturbances.
in a surge usually descends along its original trajectory. The For more information on HSS see Chapter 3.
latter characteristic distinguishes surges from sprays (Figure Subsequent work on the nature of coronal holes has
1-26) which are high velocity (exceeding the solar escape shown that they are somewhat cooler (by 500000 K) than
velocity, 670 km/s) explosive ejections associated with cer- surrounding areas and up to an order of magnitude less
tain flares. dense. See Table 1-3. They also coincide with large areas
The intense heating in flares usually “evaporates” the of open magnetic field. It is thus hypothesized that an open
chromosphere near the bases of magnetic loops. The heated field provides an unrestricted path for the propagation of
gas then rises to fill the loop with soft x ray-emitting plasma. solar wind into interplanetary space resulting in high solar
Optically, these post-flare loops or loop prominences are wind speeds that deplete the matter in the holes and carry
often observed to form in the late stages of large two-ribbon away internal energy (thus the lower density and tempera-
flares (Figures 1-26). The enhanced soft x ray and ultraviolet ture).

1-20
 THE SUN

CORONAL HOLE 1

1 JUN 73 28 JUN 73 25 JUL 73

21 AUG 73 17 SEP 73 14 OCT 73 10 NOV 73

Figure 1-29. The first coronal hole ever observed in extensive detail. This series of ~ictures from Skvlab shows the long-lived nature of the black coronal
hole in the center of the disk as seen over several solar rotations (Am~rican Science a;d Engineering, -Inc. )

1.2.7.2 Streamers. Earlier speculation on the source of this complicated pattern of fields is drawn out into space,
M-regions had centered on streamers, which are long, ra- a very complex neutral sheet or current sheet separating the
dially-oriented bright regions of the corona. Recent research two polarities develops (Figure 1-30).
has shown that streamers can be detected even at the orbit As this pattern sweeps by the earth (or any other point
of the earth. The signature of streamers [Bornni et al., 1981] near the ecliptic), it results in several days of one IMF
consists of low solar wind velocity (on the order of 350 polarity, followed by a sudden reversal and several days of
km/s), an increase in density and a decrease in the He/H
ratio. These events appear to be coincident with sector
boundary passages, which are sudden reversals of the in-
terplanetary magnetic field (IMF) that often separate HSS
from one another (Section 1.2.7. 3).

1.2.7.3. The Solar Current Sheet. In the upper corona,

magnetic fields tend to be mostly radially oriented and to
reflect the polarity of the large-scale average field underlying
them. Opposing fields in such large quasi-unipolar regions
tend to reconnect at low altitudes, resulting in a single po-
larity remaining at high altitude drawn out from the sun by
the solar wind. As the fields are stretched, they remain
attached to the rotating sun “beneath” them, and each field
line becomes a spiral (Chapter 3, Figure 2). Figure 1-30. Schematic illustration of the warped solar current sheet. This
Although the sun tends to have a dominant polarity in undulating surface divides regions in interplanetary space
having magnetic field pointing away from or toward the sun.
each hemisphere, polarities from one hemisphere often in- The region covered by this diagram is approximately 20
trude partially into the other. Thus as the sun rotates and astronomical units [Hakamada and Akasofu, 1982].

1-21
CHAPTER 1
the opposite polarity. This yields a regulm pattern of mag- electromagnetic and particle emissions connected with solar
netic secfors separated by sector boundaries, which as we flares. However, minor effects may occur due to the vari-
saw above are the signatures of coronal streamers sweeping ation of radiation from the very highly-ionized elements of
by the earth. Distortions of the IMF also occur as the flare the corona. As solar active regions and their associated
streams overtake slower streams and shocks are generated streamers grow and decay, rotate across the disk, and change
(Chapter 3). For further information refer to Hakamada and in number with the solar cycle, the amount of high-energy
Akasofu [1982]. radiation falling on the ionosphere varies slowly. Parameters
such as the Maximum Usable Frequency (MUF) for long-
range communications are partially controlled by the struc-
1.2.7.4 Transients. Skylab observations demonstrated ture of the corona, and knowledge of this structure aids in
the existence of a phenomenon never before observed out- the prediction of such parameters.
side of radio frequencies in the upper corona: the coronal Indirect effects of the corona are manifested through
transient. ,Fast motions had been observed in the lower propagation of coronal parameters to the earth by means of
corona, but the limited height range observable with ground- the solar wind (Chapter 3). Since solar wind conditions are
based coronagraphs gave no indication of their enormous directly controlled by the corona, the velocity, density, and
size. Skylab showed huge bubble-shaped disturbances rising magnetic field direction and magnitude of the solar wind
through the corona and expanding until their size approached are all determined at or near the solar surface. It is these
that of the sun itself (Figure 1-31). As they enter the inter- parameters that determine whether a geomagnetic disturb-
ance will occur. Geomagnetic disturbances generally indi-
cate the onset of a wide variety of particle and electro-
magnetic-radiation effects in the geosphere, which together
or separately have wide-ranging effects on a number of
systems. In addition, Altroek [1980] has found that even if
conditions in the solar wind are not conducive to the onset
of a geomagnetic disturbance, changes in upper-atmospheric
density may occur that affect satellite orbits. The change
may be connected with the passage of streamers at the earth.
Thus, in order to predict and/or protect against such detri-
mental effects, it is essential to have a thorough observa-
tional and theoretical understanding of processes in the solar
corona.

1.2.8 Solar Variability and Climate

While the spectacular and explosive events associated

with solar flares have immediate consequences upon the
geophysical environment, the radiant energy output of the
sun is a fundamental factor controlling the earth’s climate.
Although we do not yet have a firm, quantitative measure
of the sensitivity of the earth’s climate and weather to changes
in the solar irradianee, climatologists estimate that the earth’s
Flgure 1-31 White-light photograph of a coronal transient obtainti with
the High Altitude Observatory white-llght chronograph ex- mean surface temperature would change by about 1 K in
periment on Skylab. The black circle is not the sun but the response to a sustained 170 change in the total solar irra-
occulting disc of the instrument used to block out photos-
diance. The effects of a small change in the mean temper-
pheric light. The size of its image is two solar radii. This
particular transient traveled outward from the sun at about ature could be large. An increase of two or three degrees,
500 ktis. for example, would suffice to melt the polar icecaps, raise
the sea level by several hundred feet and flood large areas
of present coastland. On the other hand, a one degree de-
planetary medium, the transients may generate a shock wave crease would have a significant adverse impact on the major
that sweeps out into space. Their origin has been traced to grain producing regions of Europe and North America, and
surges or prominence eruptions. a decrease of perhaps ten degrees would probably result in
total glaciation of the earth.
1.2.7.5 Coronal Induced Disturbances. Direct effects Fortunately, the total solar irradiance appears to be rel-
of the corona on communications and electronic systems on atively constant, although this fact has “provenrather difficult
and near the earth come mainly through the very short-lived to verify. Ground-based irradiance measurements require

1-22
 a
THE SUN
1980 date
3f 1 3/15 411 4/15 5/1 5/15 6/1 6/15 7/1 7/15
1 I I I [ I 1 1 I 1

0.00 II
II I
I !’ t
!!
-0.05 - \
I ‘/i }
! 4
I
I \, i I
-0.10 - I i’
i 1’
I !/

-0.15 -1 \

Weighted mean 1 A.U. irradiance for period: 1368.31 W/m2

-0.20 +
45 60 75 90 105 120 135 150 165 180 195
1980 day

Figure 1-32. Total solar irradiance at 1 AU, shown as a percentage variation about the weighted mean for the first 153 days of SMM.

large corrections for atmospheric absorption, which is both 1.2.9 Predictions of Geophysical
highly selective and variable. Consequently, such measure- Disturbances
ments cannot achieve an accuracy better than + O.2%.
Measurements from spacecraft have until only recently been 1.2.9.1 Flare Prediction. Flare-induced geophysical
limited to a comparable accuracy due to detector drift prob- disturbances adversely affect many communications, nav-
lems caused by exposure to the harsh space environment. igation, and other electronic systems. Therefore daily pre-
The variability of the solar irradiance irt the ultraviolet dictions for the level of flare activity are issued by the joint
and radio portions of the spectrum is well-established (Chap- USAF/NOAA (National Oceanic and Atmospheric Admin-
ters 2 and 11) and is clearly associated with various man- istration) Space Environment Services Center (SESC) lo-
ifestations of solar magnetic activity such as flares. Fur- cated in Boulder, Colorado, working with Air Force Global
thermore, the SMM radiometer measurements have clearly Weather Central (GWC), Offutt Air Force Base, Nebraska.
detected short-term fluctuations of 0.1 -0.2% in the total Present techniques for preparing solar forecasts are based
irradiance (Figure 1-32) that correlate with the passage of on empirical relationships between flare occurrence and other
large sunspot groups across the face of the sun. However, active region parameters such as sunspot group type, mag-
longer term trends have been more difficult to establish. netic complexity, and region evolution characteristics.
Because there is no long term information from space-borne Most of the data used at SESC is provided by the Solar
instruments, ground-based measurements still provide the Electro-Optical Network (SEON) operated by the USAF Air
best information concerning such trends. There is some Weather Service. SEON consists both of optical telescopes
marginal evidence, based on measurements of light reflected (SOON - Solar Observing Optical Network) and radio tel-
from solar system bodies as well as direct radiometry, for escopes (RSTN - Radio Solar Telescope Network). The
a slight increase of about 0.4~0 in total irradiance between selection of the final SEON site (southeastern Italy) was
1968 and 1978. On the other hand, other ground-based announced in September 1982. SOON sites are located at
radiometric measurements show no evidence for any vari- five (four sites already operational in 1980) widely-spaced
ation exceeding O.2% over the 30 year interval from 1923 longitudes, providing continuous 24-hour monitoring ca-
to 1952. For more information refer to White [1977]. pability. SOON provides data on solar magnetic configu-

1-23
CHAPTER 1
rations, sunspot types and locations, and chromospheric capability of the data. One technique, using a variation of
characteristics. Four SEON sites compose RSTN, which rnultivariate discriminant analysis (MVDA) developed at
monitors solar background radio e~issjon (lo-cm flux) as AFG’L, has resulted in an objective flare forecast superior
well as the types and sizes of flare-associated radio bursts. to the presently-used, subjectively-derived forecasts.
SEON issues immediate reprts of flare occurrences, im-
poance, and locations for use at SESC tid GWC.
The near-real time acquisition of solar dat~ by SEON 1.2.9.2 Coronal Disturbance Prediction. SESC also
and other observatories, as well as present and planned predicts geomagnetic disturbances due to coronal fluctua-
satellite monitors, is making available gaily a lNW number tions as part of its daily program. Primary data sources come
of parameters for each active region on the solar disk. This from the National Solar Observatory’s Sacramento Peak
mass of data lends itself to computer programs using so- coronal photometer and Kitt Peak HeI 10830 ~ patrol, and
phisticated statistical procedures to maximize the predictive the UCSD’S radio scintillation program (Section 1. 1.3.4).

1-24
 THE SUN
REFERENCES

Allen, C. W., Astrophysical Quantities, Athlone Press, Lon- Hill, H. A., “Observational Evidence for Global Oscillations
don, 1973. of the Sun: A Review ,“ Lecture Notes in Physics, 125,
Altrock, R.C. , “Anomalous Satellite Drag and the Green- Nonradial and Nonlinear Stellar Pulsation Workshop,
Line Corona,” m “ Solar-Terrestrial Predictions Proc., Tucson, March 12–16, 1977, edited by H. Hill & W.
edited by R.F. Donnelly, NOAA, Boulder, 4: E1-E4, Dziembowski, Springer-Verlag, New York, 1979.
1980. Hill, H. A., R. T. Stebbins, and T. M. Brown, Proceedings
Athay, R. Grant (cd.), The Solar Chromosphere and Cor- Fifth International conference Atomic Masses andFun-
ona: Quiet Sun, Astrophysics & Space Science Libra~, damental Constants, Gorden and Breach, New York,
53, D. Reidel, Dordrecht, Holland, 1976. 1975.
Billings, D. E., A Guide to the Solar Corona, Academic Howard, R., and B.J. LaBonte, “The Sun is Observed to
Press, New York, 1966. be a Torsional Oscillator with a Period of 11 Years,”
Bonnet, R. M., “Contribution of 0S0-8 to our Knowledge Astrophys. J., 239: L33–L36, 1980.
of the Chromosphere and Transition Region,” Space Sci. Jordan, S. (cd.), The Sun as a Star, NASA SP-450, 1981.
Rev., 29: 131–200, 1981. Kuiper, G. P., The Sun, University of Chicago Press, Chi-
Bornni, G., J.T. Gosling, S.J. Bame, W.C. Feldman, and cago 1952.
J.M. Wilcox, “Solar Wind Helium and Hydrogen Struc- November, L. J., J. Toomre, K.B. Gebbie, and G.W. Si-
ture Near the Heliospheric Current Sheet: A Signal of mon, “The Detection of Mesogramdation on the Sun,”
Coronal Streamers at 1 AU,” J. Geophys. Res., 86: Astrophys. J., 245: L123–L126, 1981.
45654573, 1981. Rhodes, E. J., R.K. Ulrich, and G.W. Simon, “Observa-
Bray, R.J. and R.E. Loughhead, Sunspots, Wiley, New tions of Non-Radial P-Mode Oscillations ,“ Astrophys.
York, 1965. J., 218:910, 1977.
Bray, R.J. and R.E. Loughhead, The Solar Chromosphere, Sturrock, P. (cd.), Solar Flares S~fab Workhop, Univer-
Chapman & Hall, London, 1974. sity of Colorado Press, Boulder, 1980.
Cbristensen-Dalsgaard, J. and D.O. Gough, “Towards a Svestka, Z., Solar Flares, D. Reidel, Dordrecht, Holland,
Heliological Inverse Problem,” Nature, 259 89–95, 1976. 1976.
Donnelly, R.F. (cd.), Solar Terrestrial Predictions Pro- Timothy, A. F., A.S. Krieger, and G.S. Virma, “The Struc-
ceedings, Vol. 1-4, U.S. Dept. of Comm., NOAA/ERL, ture and Evolution of Coronal Holes,” Sol. Phys., 42:
Boulder, 1979–80. 135–156, 1975.
Eddy, J.A. (cd.), The New Solar Physics, Westview Press, Vemazza, J. E., E.H. Avrett, and R. Loeser, “Structure of
Boulder, 1978. the Solar Chromosphere 111.Models of the EUV Bright-
Eddy, J. A., P. A. Gilman, and D. E. Trotter, “Solar Ro- ness Components of the Quiet Sun,” Astrophys. J. Suppl.,
tation During the Maunder Minimum,” Sol. Phys., 45: 635–725, 1981.
46:3-14, 1976. Waldmeier, M. “The Sunspot Activity in the Years
Gibson, E. G., The Quiet Sun, NASA SP-303, NASA, U.S. 1610–1960,” Schulthess, Zurich, 1961.
GPO, Washington, D. C., 1973. White O.R. (cd.), The Solar Output and its Variation, Col-
Gingerich, O., R. W. Noyes, and W. Kalkofen, “The Har- orado Associated University Press, Boulder, 1977.
vard-Smithsonian Reference Atmosphere ,“ Sol. Phys., Wilcox, J.M. and R. Howard, “Differential Rotation of the
18: 347–365, 1971. Photospheric Magnetic Field,” Sol. Phys., 13:25 1–260,
Gough, D.O. (cd.), “Problems in Solar and Stellar Oscil- 1970.
lations,” Proc. 66th IAU Colloq., Crimean Astrophys. Willson, R. C., S. Gulkis, M. Janssen, H.S. Hudson, and
Ohs., USSR, 1-5 Sep 1981, Sol. Phys., 82, 1983. G.A. Chapman, “Observations of Solar Irradiance Vari-
Hakamada, K. and S.-I. Akasofu, “Simulation of Three- ability,” Science, 211:700-702, 1981.
Dimensional Solar Wind Disturbances and Result- Zirker, J.B. (cd.), Coronal Holes and High Speed Wind
ing Geomagnetic Storms,” Space Sci. Rev., 31: 3–70, Streams, Colorado Associated University Press, Boul-
1982. der, 1977.

1-25
 Chapter 2

SOLAR ULTRAVIOLET IRRAI)IANCE

Section 2.1 L.A. Hall
Sections 2.2-2.8 L.J. Heroux and H.E. Hinteregger

2.1 SOLAR UV IRRADIANCE IN THE Table 2-1 gives the solar spectrum in 1-angstrom av-
STRATOSPHERE erages from a measurement made on 19 April 1978 at Hol-
loman AFB, N.M. The data in Table 2-1 were taken at a
2.1.1 Solar Spectrum Between 2000 and 3000 ~ balloon float altitude of 41.4 km at 1022 h local time. The
Solar radiation incident on the earth’s atmosphere is solar zenith angle was 29. & and the ozone column density
absorbed in the spectral range 2000-3000 ~ by stratospheric above 41.4 km was about 2.7 X 1017 molecules/cm2. Fur-
ozone in the Hartley band. This absorption is an almost ther details and a description of the instrumentation and its
structureless continuum having a maximum near 2550 ~ calibration are given by Hall [1981].
and falling almost symmetrically on both sides. The ab-
sorption of solar radiation in this range provides the energy 2.2 SOLAR CONSTANT AND SPECTRAL
for stratospheric photochemistry and is the major part of the
IRRADIANCE
heat input which produces the local temperature maximum
in the stratopause around 50 km. 2.2.1 Solar Constant
Measurements of the 2000-3000 ~ spectrum in the strat-
osphere can be made up to about 40 km by means of high- The solar constant S is the total solar radiative energy
altitude balloons. Only about 370 of the total ozone column at the top of the earth’s atmosphere corrected for the mean
density remains above 40 km; but at 2550 ~ where the sun-earth distance. The solar constant or total solar irradi-
absorption coefficient of ozone peaks, only 3% of the in- ance expressed in units of watts per square meter (W/m2)
cident solar radiation penetrates to 40 km at local noon and is the largest contributor to the earth’s energy budget. There
30° latitude. This can be seen in Figure 2-1, where the bowl- have been several measurements of the solar constant as
shaped depression between 2900 and 2200 ~ is caused by summarized by Frohlich [1977]. The most probable value
ozone abso~tion (see Chapter 21). for S proposed by Frohlich is S = 1373 ? 20 Wlm2.

‘0’4~

7< 10”

10”

10’‘

10’0

I
1
2000 2200 2400 2600 2800 3000 32 10
WAVELENGTH -(~)

Figure 2-1. The solar irradiance in 1-~ intervals at 39.5 km in the stratosphere. Measurements made on 21 April 1977 with a solar zenith angle of 310
[Hall, 1981].

2-1
CHAPTER 2
Table 2-1. Solar irradiance at 40 km.

One Angstrom averages centered on the wavelength in units of 10” ph cm-z s-’ ~-’

A o 1 2 3 4 5 6 7 8 9

2000 0.470 0.551 0.519 0.688 0.649 0.623 0.661 0.748 0.674 0.695
10 0.802 0.623 0.727 0.678 0.642 0.821 0.760 0.766 1.03 0.934
20 0.907 0.809 0.897 0.983 1.00 0.648 0.613 0.827 0.861 0.940
30 1.04 1.09 0.909 0.952 0.929 1.13 0.950 1.13 0.905 0.732
40 0.879 1.15 1.21 1.05 1.01 0.968 1.25 1.08 1.13 1.09
50 1.09 1.23 1.09 1.30 1.10 0.943 1.21 1.26 0.981 1.15
60 1.16 1.07 0.754 1.20 1.14 1.12 1.04 1.02 1.12 1.25
70 1.27 1,36 1.35 1.42 1.36 1.54 1.39 1.55 1.29 1.35
80 1.34 1.30 1.32 1.43 1.21 1.35 1.70 1.51 2.22 1.39
90 1.50 1,99 1.86 2.61 1.71 2.06 2.22 2.00 2.72 2.05
2100 1.58 2.62 3.75 1.68 1.86 3.56 3.07 3.21 2.41 1.78
10 1.74 3.28 4.27 2.04 3.01 1.95 4.17 4.49 3.22 3.14
20 5.31 3.42 2.79 2.00 1.04 1.78 3.29 3.06 3.18 3.43
30 1.79 2.81 2.39 3.02 2.67 1.84 2.83 4.11 2.89 1.71
40 2.61 5.13 3.78 4.11 4.15 3.33 2.85 2.91 2.98 4.04
50 3.40 2.55 2.81 4.16 3.50 3.51 2.93 1.84 1.80 1.85
60 2.74 3.78 2.94 3.72 2.78 2.99 1.51 1.40 3.07 1.63
70 2.94 2.67 3.51 2.73 2.27 1.76 2.89 2.47 1.54 3.60
80 4.27 4.53 4.40 4.45 2.60 3.76 3.33 1.58 3.92 1.98
90 4.16 2.90 2.27 5.54 4.38 2.85 1.92 3.18 4.66 4.34
2200 3.03 1.83 3.80 4.60 4.64 4.01 3.47 3.15 1.68 3.18
10 1.55 1.10 1.74 4.23 2.26 3.34 2.17 1.26 1.60 3.20
20 3.29 3.67 3.85 2.32 3.64 3.13 2.60 2.49 3.40 3.10
30 3,04 2.01 4.06 5.33 4.42 4.65 4.52 4.12 2.75 3.27
40 4.20 2.86 3.25 2.73 2.20 3.48 3.29 4.04 4.00 2.43
50 2.97 2.22 3.37 2.65 2.13 2.82 2.12 2.90 2.12 3.41
60 1.44 2.12 2.62 1.76 1.78 1.45 1.60 1.33 1.57 1.25
70 1.34 1,47 1.40 1.79 1.35 1.92 1,28 1.86 1.90 2.17
80 1.39 2.75 2.87 1.95 2.12 2.83 2.02 1.18 2.30 1.94
90 1,45 1.61 1.75 1.79 1.89 1.77 2.55 1.08 0.858 1.25
2300 1.30 1.42 1.85 1.08 1.80 1.98 2.62 2.39 1.87 1.54
10 1.33 0.838 1.03 1.01 0.957 2.18 1.50 1.44 1.68 1.56
20 0.624 1.09 1.40 2.10 2.01 1.18 0.768 0.954 1.42 1.56
30 1.25 1.11 0.914 0.659 1.24 1.13 1.67 0.954 0.577 0.714
40 0.914 1.09 1.19 0.565 0.451 0.626 0.718 0.784 0.489 0.758
50 1.17 0.974 1.24 1.49 1.10 0.653 1.38 1.17 1.17 0.628
60 0.505 0.815 0.785 1.22 0.744 0.601 0.838 0.522 0.719 0.717
70 0.887 0.937 1.11 0.704 0.457 0.626 0.721 0.993 0.846 0.525
80 0.604 0.374 0.281 0.279 0.448 0.631 0.665 0.683 0.430 0.323
90 0.583 0.831 0.583 0.637 0.549 0.279 0.200 0.365 0.542 0.276
2400 0.472 0.567 0.479 0.4% 0.322 0.193 0.340 0.210 0.430 0.442
10 0.363 0.190 0.346 0.327 0.317 0.428 0.484 0.540 0.516 0.471
20 0.589 0.592 0.702 0.456 0.449 0.479 0.507 0.736 0.529 0.444
30 0.535 0.585 0.436 0.392 0.496 0.242 0.397 0.410 0.518 0.340
40 0.359 0.542 0.507 0.434 0.390 0.448 0.391 0.361 0.271 0.332
50 0.319 0.285 0.234 0.286 0.243 0.299 0.256 0.256 0.335 0.234
60 0.299 0.252 0.206 0.220 0.257 0.234 0.267 0.272 0.291 0.321
70 0.270 0.304 0.294 0.140 0.366 0.307 0.316 0.280 0.259 0.209
80 0.197 0.315 0.239 0.109 0.134 0.211 0.267 0.215 0.123 0.161
90 0.121 0.073 0.165 0.204 0.249 0.340 0.268 0.300 0.271 0.278

2-2
 SOLAR ULTRAVIOLET IRRADIANCE
Table 2-1. (Continued

One Angstrom averages centered on the wavelength in units of 1011 ph cm-2 S-l A-l

Ao 1 2 3 4 5 6 7 8 9

2500 0.323 0.151 0.241 0.289 0.326 0.292 0.218 0.129 0.245 0.294”
10 0.346 0.185 0.235 0.234 0.146 0.159 0.104 0.144 0.138 0.109
20 0.188 0.183 0.168 0.117 0.117 0.172 0.186 0.173 0.129 0.101
30 0.162 0.2+ 0.229 0.224 0.185 0.200 0.159 0.171 0.229 0.193
40 0.269 0.169 0.224 0.282 0.276 0.259 0.184 0.227 0.254 0.197
50 0.200 0.249 0.410 0.221 0.410 0.270 0.342 0.255 0.349 0.495
60 0.371 0.417 0.297 0.262 0.339 0.394 0.689 0.418 o.5q 0.575
70 0.725 0.509 0.593 0.681 0.682 0.413 0.288 0.509 0.594 0.623
80 0.856 0.962 0.781 0.441 0.641 0.370 0.189 0.422 0,470 0.788
90 0.866 0.561 0.558 0.529 0.320 0.685 0.605 0.526 0.258 Q. 195
2600 0.243 0.510 0.685 0.663 0.515 0.504 0.291 0.191 0.425 0.493
10 0.566 0.393 0.251 0.435 0.300 0.696 0.722 0.554 0.415 0.692
20 0.772 0.696 0.528 0.841 0.796 0.665 0.436 0.666 0.462 0,640
30 0.597 0.346 0.473 0.781 1.23 1.23 1.45 1.60 2.14 2.08
40 1.60 1.66 1.71 2.75 1.99 1.85 i.63 1.84 1.33 1.31
50 1.84 1.63 1.84 2.15 2.05 3.06 1.86 2.49 2.29 2.14
60 2.53 2.43 1.82 2.39 1.63 2.63 2.25 1.90 3.07 2.50
70 1.96 3.74 1.90 1.78 3:40 3.17 3.21 2.18 2.71 1.65
80 2.48 3.16 2.88 2.90 2.96 2.57 3.61 3.30 2.94 3.00
90 2.36 2.90 3.94 3.10 3.76 2.16 1.94 3.44 2.90 2.70
2700 5.25 4.00 2.96 5.02 3.74 5.70 3.86 3.62 4.34 3.19
10 4.45 4.35 3.79 5.92 2.91 3.27 3.98 4.62 3.49 1.21
20 2.44 1.81 3.78 4.08 2.89 3.58 4.13 3.96 3.48 6.37
30 6.98 4.15 5.56 4.42 3.52 4.30 5.25 1.63 4.18 3.32
40 2.47 3,90 3.14 2.20 2.55 4.65 3.55 1.73 4.11 2.38
50 2.03 4.25 6.45 6.19 4.60 4.22 1.80 4.23 8.10 9.87
60 7.34 8.71 4.90 6.28 7.22 11.7 9.70 7.86 7.55 ‘ 6.31
70 9.44 10.4 7.00 10.3 10.6 9.03 8,83 8.22 5.85 7.13
80 5.79 5.78 7.57 6.18 6.61 7.15 7.38 6.95 5.01 5.86
90 4.79 4.10 3.81 3.31 2.44 3.68 4.43 2,38 2.59 3.59
2800 3.79 3.07 2.56 5.27 2.95 3.97 5.39 5.61 7.01 8.34
10 8.58 8.87 8.80 8.66 12.9 15.7 14.2 13.8 15.0 15.7
20 19.1 17.1 18.0 15.7 22.7 21.1 13.9 21.1 19.5 19.7
30 24.4 24.6 17.8 26.6 23.3 19’1 12.4 24.9 23.9 26.0
40 21,9 19.0 26.7 18.4 15.8 25.4 17.2 14.7 11.9 10.9
50 7.65 5.92 2.60 4.87 7.64 10,8 12.2 15.8 18.7 21.5
60 29.5 24.4 31.8 28.1 30.7 29.3 38.0 27.1 26.5 21.8
70 33.6 40.8 42.1 39.7 24.6 37.9 30.3 33.7 38.4 22.3
80 25.1 12.2 10.2 24.2 29.0 47.5 41.4 41.8 39.2 44.9
90 40.2 38.0 50.4 34.8 47.7 46.7 62.4 60.3 54.6 59.9
2900 73.1 80.5 54.4 69.0 75.2 83.2 67.7 70.0 61.5 64.8
10 65.1 64.6 58.9 101 67.3 63.0 62.3 71.6 55.8 71.9
20 67.4 73.5 72.2 70.9 55.2 65.8 48.9 60.1 59.2 32.3
30 91.7 76.2 92.1 69.4 71.6 98.9 74.4 26.9 62.6 51.6
40 83.4 55.7 62.5 91.3 63.8 66.6 103 70.7 26.1 42.3
50 63.2 87.9 92.0 105 31.6 90.5 87.0 57.2 85.7 102
60 73.1 73.7 97.7 112 87.0 42.9 53.2 26.0 77.7 70.1
70 34.5 69.9 79.8 39.0 71.9 !06 81.6 113 123 85.4
80 86.0 67.8 41.8 68.1 34.6 45.3 51.7 63.2 67.3 76.0
90 110 106 87.0 82.9 56.9 40.3 94.0 92.4 110 69.1

2-3
CHAPTER 2
Table 2-1. (Continued
—
One Angstrom averages centered on the wavelength in units of 10” ph cm-2 s-’ ~-’

A o 1 2 3 4 5 6 7 ,8 9

3000 46.3 25.8 61.7 36.8 57.2 93.5 90.1 54.5 46.8 65.9
10 69.0 110 54.5 101 85.2 82.8 83.4 89.5 54.6 45.2
20 48.5 14.7 62.1 109 76.4 97.4 41.2 108 133 104
30 100 93.5 104 139 120 118 132 78.3 43.3 112
40 117 86.5 66.3 96.3 120 121 138 79.3 50.0 108
50 134 80.8 168 142 79.0 96.0 126 83.0 44.9 44.2
60 90.7 111 117 129 77.0 89.2 101 63.8 83.7 114
70 123 121 96.2 109 122 109 78.6 34 107 107
80 125 115 67.6 103 101 144 150 96.9 77.3 134
90 85.1 81.2 90.6 30.5 73.2 117 113 47.4 96.9 105

Two rocket flight measurements of the solar irradiance The solar spectral irradiance S AA, defined as the irra-
were also reported by Willson et al. [1980]. The total solar diance per wavelength interval AA, is plotted in Figure 2-
flux measurement for 29 June 1976 was 1368.1 f 0.5 W/mz 2 for wavelengths longer than 3000 ~. This figure is based
and for 16 November 1978 was 1373.4 * 0.5 W/m*. These on data from Labs and Neckel between 3000 ~ and 3 pm,
measurements are within the error estimates of Frohlich. and Thekaekara above 3 ~m. These sources of data are from

I I I I I I I I 1 I I I I I I I I 111111
I I

100

80

~ 60
a
: 40 —

20 —

I I I I 1111111 I I I
o !
3000A 5000 7000 Iym 2 3 45 678910

WAVELENGTH (j- ~m)

Rgure 2-2. The solar spectral irradiance for wavelengths longer than 3000 ~. Tb? data between 3000 ~ and 3 ~m are from Labs and Neckel and from
Tfrekaekara above 3 ~m. These data are from a summary given by Pierce and Allen [1977].

2-4
 a
SOLAR ULTRAVIOLET IRRADIANCE
a summary given by Pierce and Allen [1977]. The spectral activity near the minimum value of activity for the 11-year
region above 3000 ~ accounts for approximately 9970 of solar cycle. The irradiance values between 1 and 10 ~ are
the solar constant. Nearly all of this energy reaches the from Kreplin [1977], between 10 and 50 ~ from Manson
surface of the earth since atmospheric absorption of the [1977], and between 50 and 1850 ~ from the solar minimum
radiation is negligible. The irradiance in this wavelength reference spectrum of Heroux and Hinteregger [1978]. The
region does not appear to vary with solar activity. Because irradiance between 1850 and 2090 ~ from Semain and Si-
this region is the principal source of solar energy, the vari- mon and between 2090 and 2300 ~ from Simon is from a
ation in the total solar irradiance or solar constant is also summary of Delaboudiniere et al. [1978]. The irradiance
small and is now believed to be less than t 1Yo. between 2300 and 3150 ~ was measured by Mount and
Rottman [ 198 1] in a rocket flight flown near solar maximum
of the present solar cycle 21. However, because the spectral
2.2.2 Solar Irradiance Between 1 ~d 3000 ~ irradiance in this longer UV wavelength region does not
For Solar Minimum vary significantly with solar activity, these irradiance values
should also represent solar minimum conditions. In Table
The solar spectral irradiance for wavelengths below 3180 2-2, the solar spectral irradiance is given in units of photons
~ at the top of the earth’s atmosphere is tabulated in Table cm-2s-1 and in units of mW/m2, which is also equivalent to
2-2 for 10 ~ wavelength intervals. The values of irradiance ergs cm-zs-’. The fourth column in Table 2-2 gives the
for the wavelength range 1 to 2300 ~ are for levels of solar percentage of the total solar irradiance, s = 1373 W/m2,

Table 2-2. Solar irradiance and cumulative irradiasrce for quiet sun at the top of the earth’s atmosphere. S = 1373 W/m2, 1 mWm-2 = I erg cm-2 s-’

~ S.,ls
Wavelength Interval @A, s~h o
A (109 photons cm-’ s-’) mWm-2 %

1-1o 0.0005 0.002 0.15 x lH

10–20 0.0014 0.002 0.29
20–30 0.0076 0.006 0.73
3040 0.088 0.050 4.4
40–50 0.095 0.042 7.4

1–50 0.193 0.102

50-60 0.055 0.020 8.9 X l@6

60-70 0.069 0.021 1.0 x lrs
70–80 0.073 0.019 1.2
80–90 0.106 0.025 1.4
90–100 0.094 0.020 1.5

50–100 0.397 0.105

100-110 0.048 0.009 1.6 X l&5

110-120 0.017 0.003 1.6
120–130 0.014 0.002 1.6
130–140 0.009 0.001 1.6
140–150 0.062 0.008 1.7

100–150 0.150 0.023

150-160 0.082 0.011 1.8 X l&5

160–170 0.115 0.14 1.9
170–1 80 0.782 0.089 2.5
180–190 0.787 0.085 3.1
190–200 0.602 0.062 3.6

150-200 2.37 0.261

200-210 0.292 0.029 3.8 X 105

210-220 0.266 0.024 4.0

2-5
CHAPTER 2
Table 2-2. (Continued

~ S.,ls
Wavelength Interval a~k s~h o
A (10’ photons cm-’ s-’) mWm-2 %

220-230 0.268 0.024 4.1

230-240 0.206 0.017 4.3
240–250 0.531 0.043 4.6

200-250 1.56 0.137

250-260 1.233 0.095 5.3 x 105

260-270 0.204 0.015 5.4
270-280 0.426 0.031 5.6
280-290 0.262 0.018 5.7
290-300 0.225 0.015 5.8

250–300 2.35 0.174

300-310 7.700 0.504 9.5 x 10-5

3 10–320 0.375 0.023 9.7
320–330 0.020 0.001 9.7
330-340 0,140 0.008 9.8
340-350 0.430 0.025 9.9

300-350 8.67 0.561

350–360 0.110 0.006 1.0 x l&

360-370 0.840 0.045 1.0
370-380 0.000 0.000 1.0
38&390 0.000 0.000 1.0
390-400 0.014 0.001 1.0

350-400 0.964 0.052

400-410 0.161 0.008 1.0 x l&
41ti20 0.027 0.001 1.0
420-430 0.002 0.000 1.0
43W0 0.187 0.009 1.0
44W50 0.005 0.000 1.0

4W50 0.382 0.018

45W60 0.009 0.000 1.0 x i~

46M70 0.305 0.013 1.1
47W80 0.026 0.001 1.1
48W90 0.053 0.002 1.1
490-5m 0.173 0.007 1.1

450-500 0.566 0.023

500-510 0.163 0.006 1.1 x ld

510-520 0.025 0.001 1.1
520-530 0.111 0.004 1.1
530-540 0.120 0.004 1.1
540-550 0.021 0.001 1.1

500-5s0 0.440 0.016

550-560 0.786 0.028 1.1 x l&

560-570 0.093 0.003 1.1
57&580 0.031 0.001 1.1

2-6
 SOLAR ULTRAVIOLET IRRADIANCE
Table 2-2. (Conrinue@

~ S.,ls
Wavelength Interval OA, s~h o
A (10’ photons cm-2 s-’) mWm-2 %

580-590 1.270 0,043 1.1

590-600 0.155 0.005 1.1
550-600 2.34 0.080

600-610 0.530 0.017 1.1 x I&

610-620 0.017 0.001 1.1
620-630 1.832 0.058 1.2
630-640 0.021 0.001 1.2
640-650 0.062 0.002 1.2
600-650 2.46 0.079

650-660 0.026 0.001 1.2 x ld

660-670 0.014 0.000 1.2
670-680 0.014 0.000 1.2
680-690 0.131 0.004 1.2
690-700 0.044 0.001 1.2
650-700 0.229 0.006
700-710 0.369 0.010 1.2 x lW
710-720 0.072 0.002 1.2
720-730 0.015 0.000 1.2
730-740 0.019 0.001 1.2
740-750 0.024 0.001 1.2
700-750 0.499 0.014

750-760 0.117 0.003 1.2 x ld

760-770 0.340 0.009 1.2
770-780 0.323 0.008 1.2
780-790 0.587 0.015 1.2
790–800 0.516 0.013 1.2
750-800 1.88 0.048
800-810 0.111 0.003 1.2 x I@
810-820 0.143 0.003 1.2
820-830 0.184 0.004 1.2
830–840 0.857 0.020 1.3
840-850 0.305 0.007
800-850 1.60 0.037

850-860 0.392 0.009 1.3 x l&

860-870 0.504 0.012 1.3
870–880 0.643 0.015 1.3
880-890 0.833 0.019 1.3
890-900 1.071 0.024 1.3

850-900 3.44 0.079

900-910 1.487 0.033 1.3 x lb

910-920 0.502 0.011 1.4
920-930 0.274 0.006 1.4
930-940 0.447 0.009 1.4
940-950 0.408 0.009 1.4

2-7
CHAPTER 2
Table 2-2. (Continued

~ S.,ls
Wavelength Interval **, SAA o
A (10’ photons cm-’ S-l) mWm-2 %
900-950 3.12 0.068

950-960 0.048 0.001 1.4 x IN

960–970 0.058 0.001 1.4
970–980 5.069 0.103 1.5
980-990 0.253 0.005 1.5
99O–1OOO 0.439 0.009 1.5

95O–1OOO 5.87 0.119

1OOO-101O 0.119 0.002 1.5 x lW

1010–1020 0.222 0.004 1.5
1020–1030 3.671 0.071 1.5
1030–1040 3.794 0.073 1.6
1040-1050 0.244 0.005 1.6

1OOO-105O 8.05 0.155

1050–1060 0.292 0.006 1.6 X 1~

1060-1070 0.405 0.008 1.6
1070–1080 0.528 0.010 1.6
1080-1090 1.021 0.019 1.6
1090–1 100 0.599 0.011 1.6

1050–1 100 2.85 0.054

1100-1110 0.083 0.001 1.6x l&

1110–1120 0.022 0.000 1.6
1120–1130 0.679 0.012 1.6
1130–1140 0.045 0.001 1.6
1140–1150 0.077 0.001 1.6

1100-1150 0.906 0.015

1150–1160 0.129 0.002 1.6 X l&

1160-1170 0.219 0.004 1.6
1170–1180 2.872 0.049 1.7
1180–1190 0.493 0.008 1.7
1190–1200 0.675 0.011 1.7

1150–1200 4.39 0.074

1200-1210 4.855 0.080 1.7 x l@

1210–1220 251.774 4.114 4.7
1220-1230 0.636 0.010 4.7
1230-1240 1.480 0.024 4.8
1240–1250 0.640 0.010 4.8

1200-1250 259.39 4.24

1250–1260 1.080 0.017 4.8 X lti

1260-1270 0.930 0.015 4.8
1270–1280 0.750 0.012 4.8
1280–1290 0.500 0.008 4.8
1290–1300 0.860 0.013 4.8

2-8
 SOLAR ULTRAVIOLET IRRADIANCE
Table 2-2. (Continued

~ S.,ls
Wavelength Interval @A, s~h o
A (10’ photons cm-’ s-’) mWm-2 %

1250–1300 4.12 0.065

1300-1310 4.585 0.070 4.9 x I@

1310–1320 0.780 0.012 4.9
1320–1330 0.780 0.012 4.9
1330–1340 5.300 0.079 4.9
1340–1350 0.920 0.014 4.9

1300-1350 12.37 0.187

1350–1360 1.47 0.022 5.0 x lti

1360–1370 1.05 0.015 5.0
1370–1380 1.13 0.016 5.0
1380-1390 1.04 0.015 5.0
1390-1400 2.70 0.038 5.0

1350–1400 7.39 0.106

1400-1410 2.70 0.038 5.1 x lb

1410–1420 1.59 0.022 5.1
1420–1430 1.90 0.026 5.1
1430–1440 2.10 0.029 5.1
1440–1450 2.10 0.029 5.1

1400-1450 0.39 0.144

1450–1460 2.30 0.031 5.2 X 1~
1460-1470 3.00 0.041 5.2
1470–1480 3.80 0.051 5.2
1480-1490 3.70 0.049 5.3
149(-1500 3.40 0.045 5.3

1450-1500 16.20 0.217

15OC-151O 3.90 0.051 5.3 x l&

1510–1520 4.50 0.059 5.4
1520–1530 5.50 0.072 5.4
1530–1540 5.90 0.076 5.5
1540-1550 9.50 0.122 5.6

1500-1550 29.30 0.380

1550–1560 8.80 0.113 5.6 X lti

1560–1570 9.00 0.114 5.7
1570–1580 7.70 0.097 5.8
1580–1590 7.10 0.089 5.9
1590–1600 7.20 0.090 5.9

1550-1600 39.80 0.503

1600-1610 7.90 0.098 6.0 X 1~

1610–1620 9.40 0.116 6.1
1620-1630 11.10 0.136 6.2
1630–1640 12.00 0.146 6.3
1640-1650 15.30 0.185 6.4
E
CHAPTER 2
Table 2-2. (Continued

~ S.,ls
Wavelength Interval @A, SAA o
A (109 photons cm-’ s-’) mWm-2 %
1600-1650 55.70 0.681

1650–1660 25.00 0.300 6.6 x lW

1660-1670 18.40 0.220 6.8
1670–1680 23.00 0.273 7.0
1680–1690 26.99 0.318 7.2
1690–1700 37.02 0.434 7.6

1650–1700 130.41 1.55

1700-1710 43.00 0.501 7.9 x lN

1710–1720 43.99 0.510 8.3
1720–1730 45.98 0.530 8.7
1730-1740 41.99 0.481 9.0
1740–1750 50.02 0.569 9.4

1700-1750 224.98 2.59

1750–1760 56.99 0.645 9.9 x lW

1760-1770 61.97 0.698 1.0 x lrs
1770–1780 74.02 0.828 1.1
1780–1790 80.99 0.901 1.2
1790-1800 82.96 0.908 1.2

1750–1800 356.93 3.99

1800-1810 103.00 1.13 1.3 x 10-3
1810–1820 126.00 1.38 1.4
1820–1830 132.02 1.44 1.5
1830–1840 131.96 1.43 1.7
1840-1850 110.95 1.20 1.7

1800-1850 603.93 6.57

1850-1860 190.51 2.04 1.9 x lms

1860-1870 237.55 2.53 2.1
1870-1880 265.25 2.81 2.3
1880-1890 279.95 2.95 2.5
1890-1900 293.84 3.08 2.7

1850-1900 1267.10 13.41

1900-1910 293.48 3.06 2.9 X 103

1910-1920 332.62 3.45 3.2
1920–1930 347.92 3.59 3.4
1930-1940 254.26 2.61 3.6
1940–1950 446.52 4.56 4.0

1900-1950 1674.80 17.27

1950-1960 427.16 4.34 4.3 x 10-3

1960-1970 485.73 4.91 4.6
1970–1980 487.21 4.90 5.0
1980–1990 492.68 4.93 5.3
1990–2000 552.41 5.50 5.7

1950-2000 2445.19 24.58

2-1o
E a
SOLAR ULTRAVIOLET IRRADIANCE
Table 2-2. (Continued

~ S.,ls
Wavelength Interval @*, s~h o
A (10’ photons cm-2 s-’) mWm-2 %

2OOO-2O1O 624.83 6.19 6.2 X 103

2010–2020 629.98 6.21 6.8
20Z0–2030 642.27 6.30 7.1
2030–2M0 763.27 7.45 7.6
2040–2050 895.71 8.70 8.3

2000-2050 3556.06 34.85

2050–2060 919.75 8.89 8.9 X l@3

2060-2070 964.77 9.28 9.6
2070–2080 1128.23 10.80 1.0 x 10-2
2080–2090 1270.13 12.10 1.1
2090–2100 2562.98 24.30 1.3

2050–2100 6845.86 65.37

2100-2110 2.72 X 10+3 25.6 1.5 x 102

2110-2120 3.52 33.1 1.7
2120–2130 2.92 27.3 1.9
2130–2140 3.81 35.4 2.2
2140-2150 4.81 44.5 2.5

210(-2150 17.78 X 10+3 165.90

2150–2160 3.74 x 10+3 34.5 2.8 x l&*

2160-2170 3.48 31.9 3.0
2170–2180 3.72 34.0 3.2
2180–2190 5.07 46.1 3.6
2190–2200 5.60 50.7 3.9

2150–2200 21.61 X 103 197.2

220(-2210 5.42 X 103 48.8 4.3 x 10-2

2210-2220 4.19 37.6 4.6
2220-2230 5.80 51.8 5.0
2230-2240 7.96 70.7 5.5
2240–2250 7.17 63.4 5.9

2200-2250 30.54 x 103 272.3

2250–2260 6.95 X 103 61.2 6.4 X 102

2260-2270 5.34 46.8 6.7
2270–2280 5.50 48.0 7.1
2280–2290 8.88 77.2 7.6
2290-23W 7.18 62.1 8.1

2250–23& 33.85 X 103 295.3

2300-2310 6.88 x lV 59.3 8.5 x l&2

2310–2320 5.94 51.0 8.~
2326–2330 6.54 55,9 9.3
2330-2340 5.42 46.1 9.6
2340–2350 4.87 41.3 9,9

2300-2350 29.65 X 103 253.6

CHAPTER 2

Table 2-2. (Cc)ntinued

Wavelength Interval OA,

A ( 109 photons cm-’ s-’)

2350-2360 6.78 X 103 57.2 1.0 x I@’

2360–2370 6.18 51.9 1.1
2370-2380 6.15 51.4 1.1
2380-2390 5.36 44.6 1.1
2390–2400 5.85 48.5 1.2

2350-2400 30.32 X 103 253.6

240(-2410 5.10 x 103 42.1 1.2 x 10-’

2410-2420 7.00 57.6 1.3
2420-2430 9.39 76.9 1.3
2430-2440 8.49 69.3 1.4
2440-2450 7.73 62.8 1.4

240(-2450 37.71 x 103 308.7

2450-2460 6.46 X 103 52.3 1.4 x 10-’

2460–2470 6.79 54.7 1.5
2470–2480 7.50 60.2 1.5
2480–2490 5.43 43.4 1.6
2490–2500 8.18 65.1 1.6

2450-2500 34.36 X 103 275.7

2500-2510 7.66 X 103 60.7 1.6 X 10-’

2510–2520 5.69 44.9 1.7
2520-2530 5.55 43.7 1.7
2530-2540 7.16 56.1 1.8
2540-2550 8.17 63.8 1.8

2500-2550 34.23 X 103 269.2

“2550–2560 11.2 x 103 87.1 1.9 x 10-’

2560–2570 14.6 113 1.9
2570–2580 16.9 130 2.0
2580–2590 16.1 124 2.1
2590–2600 12.0 91.9 2.2

2550-2600 70.8 X 103 546

2600-2610 11.9 x 1(S 90.7 2.3 X 10’

2610-2620 12.9 98.0 2.3
2620–2630 13.5 102. 2.4
2630-2640 25.1 189 2.5
2640-2650 32.4 243 2.7

260(-2650 95.8 X 103 723

2650–2660 34.2 X 103 256 2.9 X l@’

2660-2670 31.7 236 3.1
2670–2680 33.0 245 3.3
2680-2690 31.8 235 3.4
2690–2700 32.4 238 3.6
2650–2700 163 X 103 1210

2-12
 ❑

SOLAR ULTRAVIOLET IRRADIANCE

Table 2-2. (Continued)

@A* s*h
$() S.,ls
Wavelength Interval
A. (10’ photons cm-’ S-l) mWm-2 70

27M-271O 35.7 x 103 262 3.8 X 101

2710–2710 29.3 214 4.0

2720–2730 26.0 190 4.1
2730–2740 28.6 208 4.2
2740–2750 18.2 132 4.3

2700-2750 138 X 103 006

2750-2760 21.2 x 103 153 4.4 x Im’

2760-2770 32.1 231 4.6
2770-2780 34.6 248 4.8
2780-2790 24.9 178 4.9
2790–2800 13.3 95 5.0

2750–2800 126 X 103 905

2800-2810 13.1 x 103 92.8 5.1 x lo’

2810-2820 28.9 204 5.2
2820–2830 40.1 282 5.4
2830-2840 44.0 308 5.6
2840–2850 33.9 237 5.8
2800-2850 160 X 103 1124

2850-2860 22.6 X 103 157 5.9 x Icr’

2860-2870 46.9 325 6.2
2870–2880 45.7 316 6.4
2880-2890 44.4 306 6.6
2890–2900 63.9 438 7.0

2850-2900 224 X 03 1542

2900-2910 81.6 x 103 558 7.3 x lml

2910–2910 77.1 525 7.7
2920–2930 68.3 464 8. I
2930-2940 73.9 500 8.4
2940–2950 69.2 467 8.8

2900-2950 370 x 103 2514

2950-2960 68.1 x 103 458 9.1 x im’

2960-2970 73.6 493 9.5
2970-2980 61.1 408 9.8
1980-2990 64.0 426 1.0 x I
2990-3000 64.4 427 I .0
2950–3000 331 x 103 2212

3OOO-3O1O 50.8 X 103 336 1.1

3010-3020 59.3 391 1.1
3020-3030 63.5 417 .1
3030-3040 77.7 509 .2
3040-3050 74.9 489 .2

3000–3050 326 X 10’ 2142

2-13
CHAPTER 2
Table 2-2. (Continued

~ s.A/s
Wavelength Interval @A, ~Ak o

A (10’ photons cm-2 s-’) mWm-2 %

3050–3060 73.5 x 103 478 1.2

3060–3070 71.9 466 1.3
3070–3080 79.0 510 1.3
3080–3090 76.4 492 1.3
3090-3100 58.5 375 1.4

3050–3100 359 x 103 2321

3100-3110 78.5 X 103 502 1.4

3110-3120 87.9 561 1.4
3120–3130 78.9 502 1.5
3130–3140 85.3 540 1.5
3140–3150 74.6 471 1.6

31W-3150 405.2 X 103 2576

3150-3160 66.5 X 103 419 1.6

3160–3170 77.8 488 1.6
3170-3180 105 657 1.7

contributed by the wavelength region shorter than the listed Iok 1 I 1

I 1I 1
I I
I I

wavelength. For example, the solar irradiance for the entire

wavelengthregion bel~w3000~is l%ofthesolw constant.
n
The solar irradiance based on Table 2-2 is plotted for
20 ~ wavelength internals in Figure 2-3. The pronounced
structure for the wavelength Region 1 to 1500 ~ arises from
emission lines of elements abundant in the sun. The high
levels of irradiance in the intervals 300-320 ~ and 1200-
1220 ~ originate primarily from the solar emission lines of
He 11 at 303.78 ~ and hydrogen Lyman-a (H Ly-a) at
1215.67 ~, respectively. The solar spectrum above about
15W ~ is essentially a continuum with absorption and emis-
sion lines superimposed on the continuum.

2.2.3 Solar Irradiance Below 1200 ~

The temperature of the visible surface of the sun is about

6000 K. If the temperature is measured as a function of
distance outward from the surface of the sun, the temper-
ature decreases to about 4500 K at the top O! the Photosphere
(near 500 km) ai’td increases gradually in the chromosphere
which extends to about 2000 km. As can be seen in Figure
2-4, the temperature increases rapidly from about 104 to
5 x 105 K in a very narrow transition region near 2000
km, and then continues to increase gradually in the corona
where the temperature can exceed 2 X 106 K. The solar
Wlllllll’1 ! l..~d J
UV spectrum between 1 and 3000 ~ is produced throughout (;00 1700 1900 2100 2300 2500 2700 2900
wAvELENGTH(i)
this spatial region which encompasses a temperature range
from the photospheric minimum temperature of 4500 K to
Figure 2-3. The solar spectral irradiance between 1 and 3000 ~ for a
he coronal temperature of 2 x 106 K. quiet sun at the top of the earth’s atmosphere. The plotted
The solar spectrum between 3000 and 1300 ~ is pro- values are averages over 20 ~ intervals af wavelength.

2-14
 SOLAR ULTRAVIOLET IRRADIANCE
duced predominantly in the upper photosphere and lower
chromosphere. The dominant feature in this wavelength re-
I 06
gion is a continuum that becomes apparent near 1300 ~ and
increases in intensity toward longer wavelengths as is evi-
dent in Figure 2-3. There are numerous emission and ab-
sorption lines superimposed on the continuum. At wave-
10s lengths shorter than about 1300 ~, the solar spectrum consists
predominantly of sharp emission lines that originate from
transitions in multiply ionized atomic species that are pro-
duced over the wide range of temperatures established in
10’ the chromosphere and corona. Figure 2-4 indicates the tem-

I I I I 1 perature regions in which several of the ion species observed

1000 2000 4000 10,000 20,000 40,000
in the solar spectrum are produced. An example of the solar
HEIGHT ABOVE SOLAR LIMB—(km) UV spectrum between 1300 and 270 ~, measured on the
NASA 0S0 3 satellite by Hall and Hinteregger [1970], is
Figure 2-4. The spatial distribution of temperature in the solar chromo- given in Figure 2-5. The scattered light background beneath
sphereand corona. The temperature region of prediction of the spectrum is estimated from a pre-flight calibration to
several multiply-ionized species is also given,
decrease gradually from 150 counts near 120 ~ to 50 counts

I05
II MARCH 1967 1

104

m
g 103
3

2
102 I
I

10’
flu’’’’’:’””
1300
1310
I 200 1100 I 000

WAVELENGTH (~)
900 800 700

NO. 610
I 0:
He I cent
2 Hell
I
104

102

10’
600 500 400 300
270
WAVELENGTH (~)

NO. 611

Figure 2-5. The solar UV spectmm between 1300 and 270 ~ observed from the satellite 0S0 3. The folding of the data for H Ly cr line at 1216 ~ is
introduced by the pulse counting circuitry [Hall and Hinteregger, 1970].

2-15
CHAPTER 2
near270 ~. The reversal of the center of the intense hy- to the temperature region of formation of He II and Fe XVI,
drogen Lyman- ci line at 1216 ~ is instrumental, caused by respectively. The reversal of the intense emission line of Fe
overflow of the photon counter. The hydrogen continuum XV at 284 ~ is again instrumental, caused by overflow of
which extends from about 911 ~ toward shorter wavelengths the photon counter.
is the only significant continuum in the spectrum for wave-
lengths shorter than 1300 ~, A detailed discussion of the 2.3 ABSORPTION OF SOLAR UV
spectrum shown in Figure 2-5 is given by Hall and Hinter-
egger [1970]. 2.3.1 Absorption Below 3200 ~
The solar UV spectrum between 300 and 100 ~ mea-
sured in a rocket flight on 4 April 1969 is given in Figure
Solar radiation at wavelengths shorter than about 3200
2-6. The figure is from a paper by Malinovsky and Heroux
~ is totally absorbed in the earth’s upper atmosphere. Al-
[1973] where the top of the atmosphere intensities of the
though this ultraviolet region accounts for Iess than 2% of
lines are also tabulated. For the wavelength interval 300-
the total solar irradiance, it is the principal source of energy
220 ~, one can see that this small interval of wavelength
in the upper atmosphere and controls the neutral and ion
contains multiples of ions produced over a wide temperature
composition, temperature, and photochemistry in the strat-
range extending from 2 x 104 to 2 X 106 K, comesponding
osphere, mesosphere, and thermosphere. The absorption of
solar UV in the earth’s atmosphere for an overhead sun is
t , 1 ( , r , 1 , r T , , , I illustrated in Figure 2-7, which gives the altitude at which
XI03
I F, Xl Fhm
I
the rate of absorption at a particular wavelength is a max-
2.0 - imum. At these altitudes, solar radiation is retiuced by a
factor of e-* = 0.37 of its value above the earth’s atmos-
phere. The primary atmospheric constituents that absorb the
: 1.5 -
radiation in the different wavelength regions are also given
;
o
a)
in the figure.
: 1.0 -
m 2.3.2 UV Atmospheric/Ionospheric Processes
1-
Z Fe 1
r
? 0.5 - Fe;
The absorption of solar UV radiation at wavelengths
shorter than 1027 ~ ionizes the primary atmospheric con-
0 stituents Oz, N2, and O. This radiation is absorbed at alti-
[60 (70 tudes above about 100 km. The rate at which the ions Oz +,
wAv~$NGTH.’~’ 200 2~o ‘ 22o N2 + and O+ are produced as a function of altitude is shown
in Figure 2-8. These rates were calculated by using rocket
1 1 , I , , 1 , , 1 1 I , , ,
measurements of solar UV intensities for 23 August 1972.
,Fom, ,, F, IIY. The corresponding total production rates of the three species
SIX of ions produced by radiation within several wavelength
*
XI03 He U
F, y intervals of the solar UV spectrum are given in Figure 2-9.
1.5 A detailed account of the calculations used to obtain these
figures is given by Heroux et al. [1974].

v
. I 2.4 SOLAR UV IRRADIANCE VARIABILITY

2.4.1 Sunspot Number and 10.7cmSolar Radio Flux

The sunspot number and the 10.7 cm (2800 MHz) solar

radio flux are used extensively as indices of solar activity.
Eddy [1977] gives a review of sunspot observations that
have been made for several hundred years. These records
clearly indicate that the sunspots are indicators of solar
220 230 240 250 260 270 280 290 activity. An additional diagram of the latitudinal distribution
WAVEL~TH (a) of sunspots during the period 1874-1976 is given by Yallop
and Hohenkerk [1980]. The sunspot numbers for the present
Ftgure 2-6. The solar UV spectrum between 100 and 300 ~ obtained solar cycle 21 are plotted in Figure 2-10. The values are
fmm a rocket flight on 4 April 1%9. The folding of the data
for the Fe XV line at 284 ~ is introduced by the pulse counting the monthly mean sunspot numbers obtained from the Sohir-
circuitry [Malinovsky and Heroux, 1973]. Geophysical Data Reports published monthly by the Na-

2-16
 SOLAR ULTRAVIOLET IRRADIANCE

02, N2,0

F REGION

E REGION

D REGION
t
Ly a

t \

o I I I I I I I I 1 ,:)0 1 /*
o 600 1200 1800 2400 6000
WAVELENGTH(A)

Flmrre 2-7. The altitude at which the rate of absorption of SOlar Uv radiation is at maximum. The principal atmospheric constituents that absorb the
radiation in the different wavelength bands are indicated.

tional Oceanic and Atmospheric Administration (NOAA). minimum of cycle 21. For comparison, the monthly mean
These reports also give the daily values of the sunspot num- values of the 10.7 cm radio flux for cycle 21 obtained from
bers. The smoothed mean values of the sunspot numbers the NOAA reports are also given in Figure 2-10. Both in-
for cycles 8-20 are also included in tie figure. Each cycle’s dices show a similar increase with increasing activity for
beginning minimum has been shifted to coincide with the cycle 21, which began in July 1976. The daily variations

300 f 300

250 – 250

200 –

150 –

100 –
t , # 1 1 1 I 11 0 0 I 1 J 1 1 1 # , d t 1 I I I 1111 I I I I 11111 1 I I I I
10 100 1000 10 100 7000
ION PRODUCTION RATE (cm-s see-l) ION PRODUCTION RATE (cm-3 see-’)

Figure 2-8. Production rates of O+, Oz +, and Nz + and their totals as a Figure 2-9. Total production rates of O+, 02+, and N2 + produced by
function of altitude. The rates were calculated by using rocket several wavelength regions of the solar UV spectrum. The
measurements of solar intensities between 1027 and 52 ~ of rates were calculated by using the solar intensities of 23
23 August 1972 [Heroux et al., 1974]. August 1972 [Heroux et al., 1973].

2-17
CHAPTER 2
1’1 I
I I
I I
I I

220

CYCLE 21
180

140

100

60 t--

160 CYCLE 21

120

80

1
40

0 1 I I I I I I I I I I I I 1 I 1 I I I
JUNE JUNE JUNE
1976 1978 1980 1982 1984 1986

Figure 2-10 Monthly mean sunspt numbers and 10.7 cm radio flux for solar cycle 21, which began in July 1976. The smoothed mean values of the
sunspot numbers for cycles 8-20 were obtained by shifdng each cycle’s minimum value to coincide with the minimum of cycle 21.

in both the sunspot number and the 10.7 cm flux can be Therefore, it is not surprising that the UV flux emitted from
significant. For November 1979, the mean value of the a wide range of regions in the solar atmosphere does not
sunspot number is 185, while the daily values during the correlate welI with the daily 10.7 cm flux emitted from a
month varied from a minimum value of 98 to a maximum limited region in the solar atmosphere.
value of 302. For the same month, the mean value of the
10.7 cm flux was 232, while the minimum and maximum
values were 154 and 374, respectively. 2.4.2 Solar Cycle and UV Variability Below 2000 ~
Although the 10.7 cm flux is a reliable indicator of the
general level of solar activity, the correlation between the The variation of the spectral irradiance from the mini-
solar UV flux in narrow wavelength regions and the daily mum through the maximum level of activity of a solar cycle
10.7 cm flux is highly unreliable. Solar UV emissions orig- can be significant for wavelengths shorter than about 1800
inate from a wide range of temperatures, densities, and ~. This variation is illustrated in Figure 2-11, where the
heights in the solar atmosphere extending from the lower ratio of the irradiance near solar maximum during January
chromosphere, the chromosphere-corona transition region, 1979 to the irradiance near solar minimum during July 1976
and the corona. The 10.7 cm flux, however, originates pre- for solar cycle 21 is plotted for 25 ~ intervals of wavelength.
dominantly from the upper chromosphere and lower corona. The data were obtained by Hinteregger [1981] from the

2-18
 SOLAR ULTRAVIOLET IRRADIANCE
AFGL spectrometer on the NASA Atmosphere Explorer greater than 100 frok solar minimum to maximum, although
(AE-E) satellite. Figure 2-11 shows that the variation de- the ratio of the averaged irradiance in the interval 325-350
creases with increasing wavelength towtid 1800 ~. At A is a factor of 12.
wavelengths longer than about .2300 ~ the ratio approaches Figure 2-12 illustrates the relative variation of the solar
unity, and therefore solar activity has a negligible effect on spectral irradiance for several 50 ~ intervals in the 1400-
the spectral irradiance at these longer wavelengths. The 1750 ~ range, along with the hydrogen Lyman-u line, at
large ratios that ara apparent .at wavelengfis blow 500 ~ 1216 ~. The vtiatioris for [he Fe XV 284 ~ line emission
originate from highly ionized atomic species produced in and for two inte~”als of the’ 170-205 ~ range are shown in
the solar corona. Individual lines that fall within the 25 ~ Figure 2-13. These data were obtained on the AE-E satellite
intervals can vary by a significantly higher ratio than that by Hinteregger [1980]. Each point represents the average
plotted in Figure 2-11 since the plotted ratio represents the of all measurements for the month. For comparison, the
irradiance averaged over several lines that fall within the monthly averages of the sunspot number R= and the 10.7
wavelength interval. For example, the ratio of the coronal cm radio emission are included in the figures. These data
emission line of Fe XVI at 335.41 ~ increases by a factor illustrate that the variation of the irradiance with solar ac-

r T I 11] 11111111-1/1,,,,,
T I
1-

loot’
11111, 111 [111,1,1,1,1,111 Illlllllj
200 400 600 800 1000 - 1200 1400 1600 1800
WAVELENGTH (h)

Figure 2-11. The ratio of the solar spectral irradiance near solar maximuim during Janu@ 1979 to the irradiance near solar minimum during July 1976
for solar cycle 21. The ratios are plotted for 25 ~ intervals of wavelength [Hinteregger, 1981].

2-19
 tivity between 1977 and 1980 depends upon the region of
300 production of the radiation in the solar atmosphere. The
variation of the H Lyman-a line that is produced in the
lower chromosphere is a factor of approximately 2.3, while
200
the variation of the Fe XV line produced in the corona is
I50 approximate y 8.0 The variation of the photospheric radia-
tion in the wavelength interval 1700-1750 ~ is only 1.16.
I00
This variation approaches 1.0 for wavelengths greater than
about 2200 ~.
70 In addition to the variability of the solar UV irradiance
i over an 1l-year cycle, there is also a 27-day variation of
solar UV associated with the period of solar rotation. The
1“ k 1400-1450
I
A ~ 50
magnitude of the variation at different wavelengths depends
upon the region of the solar atmosphere emitting the radia-

Hil
30 tion. The variations of the emission lines from highly ionized
1500-1550 A
species produced in the corona are significantly greater than
.z -
1600-1650 A 20 the variations of lines produced in the lower chromosphere.
---
This is illustrated in Figure 2-14 where measurements are
,-- 1700-1750A 15 plotted of the 27-day variation of several solar emission
---
lines for May and June 1967 from an experiment on the
10 0S0 3 satellite [Hall and Hinteregger [1970]. The variation
1977 1978 1979 1980
of the coronal Fe XVI lines at 335 ~ can be seen to be
significantly .- greater than the variations of the other chro-
Figure 2-12. Irradiance variations at wavelengths from 1216 to 1750 ~
from AE-E satellite measurements during solar cycle 21 mospheric lines shown in the figure. The 27-day variation
[Hinteregger, 1980]. of the unresolved He 11and Si XI lines near 304 ~ has also
been measured over a period of almost two years (1967 to
1969) from an instrument flown on the 0S0 4 satellite by
I I t Timothy and Timothy [1970].

I032A

630 A(XI.5)

I I I
7
MAY
~i
(3.81
1977 1978 1979 1980
I OATE I JUNE
No.612

Figure 2-14. The 27-day variations in several UV emission lines observed

Figure 2-13 lrradiance variations of some solar coronal emissions from from the 0S0 3 satellite in May and June 1967. The 10.7
AE-E satellite measurements during solar cycle 21 [Hinter- cm radio flux for the same period is also given for comparison
egger, 1980]. [Hall and Hinteregger, 1970].

2-20
 SOLAR ULTRAVIOLET IRRADIANCE
REFERENCES

Delaboudinier, J. P., R.F. Donnelly, H.E. Hinteregger, G. Hinteregger, H. E., “AE-EoExperiences of Irradiance Mon-
Schmidtke, and P.C. Simon, “Intercomparison/ Com- itoring for 1250-1850 A,” in Proceedings of the Work-
pilation of Relevant Solar Flux Data Related to Aeron- shop on Solar UV Irradiance Monitors, NOAA Envi-
omy” International Council of Scientific Unions, Com- ronmental Research Laboratories (ERL), Boulder, 1980.
mittee on Space Research, COSPAR Technique Manual Kreplin, R. W., K.P. Dere, D.M. H~ran, and J.F. Meekins,
Series, 7:48–5 1, Paris, 1978. “The Solar Spectrum Below 10 A,” in The Solar Output
Eddy, J. A., “Historical Evidence for Existence of the Solar and Its Variation, edited by O. R. White, Colorado As-
Cycle,” in The Solar Output and [ts Variation, edited sociated University Press, Boulder, 1977.
by O.R. White, Colorado Associated University Press, Malinovsky, M. and L. Heroux, “An Analysis of the Solar
Boulder, 1977. Extreme-Ultraviolet Spectrum Between 50 and 300 ~,”
Frohlich, C., “Contemporary Measures of the Solar Con- Astrophys. J., 181: 1009, 1973.
stant ,“ in The Solar Output and its Variation, edited by Manson, J. E., “The Solar Spectrum Between 10 and 300
O.R. White, Colorado Associated University Press, ~,” in The Solar Output and Its Variation, edited by
Boulder, 1977. O. R. White, Colorado Associated University Press,
Hall, L. A., “Solar Ultraviolet Irradiance at 40 Kilome- Boulder, 1977.
ters in the Stratosphere,” J. Geophys. Res., 86: 555, Mount, G.H. and G.J. ~Rottman, “The Solar Spectral Irra-
1981. diance 1200-3184 A Near Solar Maximum: July 15,
Hall, L.A. and H ,E. Hinteregger, “Solar Radiation in the 1980,” J. Geophys. Res., 86: 9193–9198, 1981.
Extreme Ultraviolet and Its Variation with Solar Rota- Pierce, A.K. and R.G. Allen, “The Solar Spectrum Between
tion,” J. Geophys. Res., 75:6959, 1970. 0.3 and 10 ~m,” in The Solar Output and Its Variation,
Heroux, L. and H. E. Hinteregger, “Aeronoomical Reference edited by O. R. White, Colorado Associated University
Spectrum for Solar UV Below 2000 A,” J. Geophys. Press, Boulder, 1977.
Res., 83:5305, 1978. Timothy, A.F. and J.G. Timothy, “Long-Term Intensity
Heroux, L., M. Cohen, and J .E. Higgins, “Electron Dens- Variations in the Solar Helium 11 Lyman Alpha Line,”
ities Between 110 and 300 km Derived from Solar EUV J. Geophys. Res., 75:6950, 1970.
Fluxes of August 23, 1972,” J. Geophys. Res., 79: Willson, R. C., C.H. Duncan, and J. Geist, “Direct Mea-
5237, 1974. surement of Solar Luminosity Variation ,“ Science,
Hinteregger, H. E., “Representations of Solar EUV Fluxes (London) 207:177-179, 1980.
for Aeronomical Applications,” Adv. Space Res., 1:39, Yallop, B .D. and C. Y. Hohenkerk, “Distribution of Sun-
COSPAR, 1981. spots 1874– 1976,” So/ . Phys. , 68:303, 1980.

2-21
 Chapter 3

SOLAR WIND
J. Feynman

The solar wind or interplanetary medium is a fully ion- where r is the radial distance, p the mass density, and u the
ized, electrically neutral plasma that carries a magnetic field radial velocity. The momentum equation for a fluid is written
and streams outward from the inner solar corona at all times. as
The wind is highly variable in both time and space and fills
the interstellar region in the vicinity of the sun. This chapter du –dP
—_ ~
PF> (3.2)
describes the observed properties of the solar wind in the ‘“%= dr
vicinity of the earth, including the undisturbed wind and
the magnetosheath and bow shock formed as the solar wind where P is the scalar pressure, G the gravitational constant,
flows around the earth’s magnetic field. and MO the mass of the sun. Energy conservation is given
by

( )1
3.1 BASIC THEORY ld
—— Ppu ;U2 +;;
? dr
(3.3)
1-
3.1.1 Simple Spherically Symmetric
Treatment — ;:(r2Pll)-pu GMO
T + S(r)

The solar wind is the outer part of the sun’s corona,

streaming past the earth. Chapman [1957] made an early where S(r) is an energy source or sink term. The corona is
estimate of the particle density expected in space at the orbit assumed to be a neutral electron proton gas so that
of the earth. He considered the expansion of a corona with
a temperature at the base on the order of 10s K. He calculated P=n(mp+m)=nm (3.4)
the coronal density profile assuming that the corona was in
hydrostatic equilibrium and found that the electron density where mP and m are the proton mass and u the mass of
at the earth’s orbit would be 102 or 103 particles/cm3. He the electron. For equal electron and proton temperature the
also found that the pressure at large heliocentric distances pressure is given by
would be of the order of l&5 dyn/cm2, whereas estimates
of the interstellar pressure were in the range of 1012 or 1013 P = 2nkT. (3.5)
dyn/cm2. Thus, the hydrostatic corona could not be matched
with its boundary conditions in the distant heliosphere. This Parker further simplified the problem by assuming that the
treatment, although describing the sun in a very simplified pressure and temperature were related by the polytropic law
manner, led Parker to postulate that the corona was not in
hydrostatic equilibrium and that, in fact, the coronal plasma
was flowing out from the sun continuously. Parker [1963] (3.6)
treated the problem as that of a steady state spherically
symmetric fluid expansion. The equations that govern the
\pO/
behavior of such a system are the equations of mass, mo- where a is the polytropic index. An isothermal gas corre-
mentum, and energy conservation. The mass conservation sponds to a = 1 and an adiabatic process to a = 5/3. The
is expressed by assumption of the polytropic law effectively assumes a so-
lution to the energy conservation equation with an implied
source term. Parker examined the solutions of the equations
for a range of selections for a. See Parker [ 1963] or Hund-
$:(r2pu)=o, (3.1)
hausen [1972] for further discussion and details.

3-1
CHAPTER 3
The simplest case, but one that illustrates important points I 1 I I I I I
concerning the behavior of the solutions, is that of the iso-
thermal corona. Near the sun the isothermal assumption is
probably valid. Using the expressions for the density and
the polytrope law with a = 1, the momentum conservation
equation becomes

du nmGm
–2kT: – ~. (3.7)
‘mu z =

Equation (3. 1) for mass conservation is integrated to give

v
4~nu? = constant, (3.8)

that is, the flux through a sun centered sphere is a constant.

Equation (3.8) can be used to eliminate the number density
from Equation (3.7)

1 du
——
u dr
( )~
U2 _2kT—
m
4kT
.—_%
P“
(3.9)
2 4
r/rc
6 8

Assume that the temperature of the corona is such that

Figure 3-1. Topology of solutions to the solar wind equations for the
spherically symmetric isothermal case. me only solution that
T<GMm
—— can satisfy the boundary conditions of low velaity at the sun
rO 4k’ and low pressure at large distances is the class 2 solution
[Hundhausen, 1972].

where rOis the base of the corona. The right hand side of
Equation (3.9) is negative in the range of rO< r < r. where either maximum or minimum near r.. These four classes of
solutions for the velocity are shown in Figure 3-1. The
——GMom physical solutions for the solar wind are members of the
(3.10)
‘c – 4kT ‘ class of solutions that satisfy the physically relevant bound-
ary conditions. Class 1 solution velocities approach zero
and r. is called the “critical radius.” For r > rc, the right near the sun and at great distances. Class 2 is the unique
hand side of Equation (3.9) will be positive. For r = rc, solution that has a low velocity near the sun, passes through
the right hand side must be zero. In that case, to make the the critical point, and has high velocity at great distances.
left hand side zero either Both class 3 and 4 solutions have high velocity near the
sun. Class 3 is the unique solution passing through the
2kT critical point and having low velocity at great distances,
uz(rc) = — (3.lla) whereas class 4 solutions have high distant velocities. Since,
m
for the physical solutions, the velocity near the sun is small,
or class 3 and 4 solutions are immediately eliminated. Class
1 solutions would give finite pressures at large distances
1 du and would result in the same problem of matching to the
—— = 0. (3.llb)
u dr interstellar medium as did the hydrostatic corona. For class
r=rc 2, however,it can be shown that the expansion speed con-
tinues to increase slowly as r + m, so that the density and
These two conditions expressed in Equation (3.1 la and hence the pressure approach zero at infinity. Thus the class
b) define four classes of solutions to the set of simultaneous 2 solution satisfies the boundary conditions and represents
equations governing the model corona. If Equation (3.1 la) the physically existing solar wind. Since according to Equa-
is satisfied then du/dr has the same sign for all r, that is, u tion (3. 1la) the velocity at r= is equal to the velocity of
either increases or decreases monotonically with distance sound, this solution is supersonic for r > r=.
from the sun. If Equation (3.1 lb) is satisifed, then u is Although the dicussion above is highly simplified [fol-

3-2
 SOLAR WIND

lowing Hundhausen, 1972] the essential characteristics are The velocity of the wind changes only slowly with radius
maintained in all more sophisticated treatments. That is, a beyond a few times the critical radius and will be approx-
family of solutions in every treatment exists but the im- imated by a constant. The path followed by the fluid element
position of the boundary conditions selects a finite number is determined by
of solutions that have low velocities at the sun, pass through
critical points where the flow becomes supersonic, and result ldr
.— .J= U u
(3.13)
in a high finite velocity and zero pressure at infinity. r d+ U+ – wrsin+”
The configuration of the solar wind magnetic field was
also treated by Parker [1963]. Since the conductivity y of the For constant u beyond rO this is integrated to give
solar wind is extremely high, there is essentially no diffusion
of the plasma transverse to the magnetic field and the field
is said to be “frozen in” to the flow. The plasma carries the r— r.= =(4 - +0) (3.14)
field with it into space. If the sun were not rotating, the
field would extend straight outward in all directions but the Then for a spherically symmetric geometry V o B = O
rotation causes the field to appear wound up as shown in gives
Figure 3-2. In a spherical system of coordinates (r, +, 6)
2

/ / TORBIT OF EARTH
B, (r, 4, 6) = B (r., 4., 0) f
()
\/
B@ (r, +, 0) = –B (rO, +., 0) ~ sin O (3.15)

Be = O.

The configuration near the solar equatorial plane is shown

in Figure 3-2.
Transforming to a stationary coordinate system leaves
the magnetic field configuration unchanged but in the sta-
tionary system there is an electric field

E=–ux B (3.16)

so that the direction of the plasma flow is radial rather than

along the field lines.

3.1.2 Further Considerations

Figure 3-2. The configuration of the interplanetary magnetic field in the

This simple spherically symmetric solar wind theory was
equatorial plane for a steady solar wind. The field winds up remarkably successful in predicting the existence of the wind
because of the rotation of the sun. Fields in low speed wind and estimating the velocity and density. However, it soon
will be more wound up than those in high speed wind [Parker,
1963]. (Reprinted with permission from J. Wiley and Sons
became apparent that the predicted velocity was lower than
O 1963,) that observed and the predicted density was too large. Re-
finements of solar wind theory have taken many approaches,
rotating with the sun, the velocity of a fluid element carrying none of them as yet completely successful. The earliest
the field is given by attempt at improvement was to consider the electrons and
protons as two fluids each expanding separately from the
Ur=u sun. The particles within each fluid were interacting with
each other through large scale electric fields. This two-fluid
U+ = – wrsin6 (3.12)
model produced results that were again too slow and too
Ue=o dense, and the predicted total energy flux at 1 AU was twice
that observed. Later, observations showed that the solar
where u is the speed of the wind and o is the angular velocity wind came from restricted areas of the sun so that the ex-
of solar rotation (w = 2.7 X 10-6 rad/s). pansion of the wind was faster than the # appropriate to the

3-3
CHAPTER 3
spherical case discussed by Parker [1963]. Studies of the ecliptic. The transformations among the systems are given
effect of this geometrical difference, however, showed that in Russell [1971].
although the acceleration of the solar wind at small distances
from the sun was increased markedly, the speed at 1 AU
was not raised appreciably. A possible cause for the speed 3.2.2 Magnetic Fields
of the wind being higher than expected from simple theory
is the effect of magnetohydrodynamic waves on the wind Observations of the magnetic field in the vicinity of the
in the supersonic region. Theory indicates these waves will earth show that the field is, on the average, in the spiral
exert a force on the solar wind tending to accelerate it. At direction predicted by the steady state model but that there
1 AU, outward propagating AIfv6n waves (Section 3.2.5) are often large variations. The variations are ascribed to
are the predominant mode observed. Fast mode MHD waves several sources: waves, discontinuities, and certain large
(Section 3.2.5) propagating outward near the sun would also scale disturbances. These topics will be described in Section
accelerate the wind. Since this latter mode is strongly damped, 3.3. Here the basic large scale structure and average prop-
it is quite reasonable that if there was a wave flux near the erties will be described.
sun it would be lost before the wind propagated to 1 AU. The magnetic field in the solar wind is, as expected,
This consideration is important because fast mode MHD drawn out in a spiral such that the angle the field makes
waves are rare] y seen at 1 AU. For both these wave accel- with the earth-sun line is a function of the velocity, low
eration mechanisms, an important problem is that the ex- solar wind speeds resulting in a tight spiral and faster speeds
istence of the required waves in the required region has not in a more radially aligned field. The angle increases with
been demonstrated. An alternative idea for wind accelera- distance from the sun. For the observed velocities of the
tion is that the electron thermal conductivity has been over- solar wind, the magnetic field has a spiral angle of about
estimated by using the classical values, and at the same 45° at the earth. The magnitude of the field is on the order
time, the thermal coupling between the electrons and protons of a few nanotesla (earlier literature used the term gamma,
has been underestimated. As with several other proposals ly = 1 nanotesla). An example of observed magnetic field
listed here, this seems to be an attractive possibility but the distributions is given in Figure 3-3a, b, and c [Hirshberg,
observations and analysis have not been canied far enough 1969]. The direction of the radial component (often referred
to make a definitive test. to as the x component) is, of course, either towards or away
from the sun. Wilcox and Ness [1965] found that the field
typically maintains one of these orientations for many days
3.2 PARTICLE AND FIELD and then rapidly changes to the other. The region of space
OBSERVATIONS in which an orientation is maintained is called a “sector”.
The sector orientation is usually determined on the basis of
one-hour or three-hour averages of the radial field compo-
3.2.1 Coordinates nent, but the maintenance of the sign of B, is typically seen
on a smaller time scale. Transitions from one sector to
There are two coordinate systems commonly used to another take place over a period of minutes or hours. At
describe observations of the interplanetary medium [Russell, transitions the direction of the interplanetary field usually
197 1]: geocentric solar ecliptic (GSE) and the geocentric changes sign rapidly while maintaining more or less the
solar magnetospheric (GSM). In the GSE system, X points same magnitude. Sometimes the sector boundary on the
from the earth towards the sun, Y is in the ecliptic plane scale of minutes or tens of minutes displays only a single
and negative in the direction of planetary motion, and Z is change of sign. Sometimes the sign changes back and forth
parallel to the ecliptic pole. In GSM, X is along the earth many times and over many hours before the field settles
sun line and positive in the direction of the sun, Z is positive down to the new orientation. When this happens the position
toward the north and is pe~endicular to X and in the plane of the boundary may be hard to define. An example of the
which contains X and the earth’s magnetic dipole axis, and observed structure is shown in Figure 3-4. Typically, two
Y completes the right handed system of coordinates. or four sectors occur during a solar rotation. The sector
GSM coordinates are usually used to describe interac- structure can also be inferred from high latitude surface
tions between the solar wind and the earth’s magnetic field magnetic variations [Svalgaard, 1973]. The accuracy of the
whereas GSE coordinates are used to describe the undis- procedure varies with season and station latitude. The sector
turbed solar wind. A second system sometimes used for the structure is stable over many months during the declining
undisturbed solar wind is the geocentic solar equatorial phase of the solar cycle, but changes from month to month
system in which the X axis remains in the ecliptic plane during the increasing phase. The sector structures and/or
pointing from the earth to the sun, but Y is parallel to the the high speed streams (Section 3.3 .2) apparently order a
sun’s equatorial plane and Z completes the system. The large amount of solar wind phenomenology. Studies of the
sun’s pole of rotation is tipped at 7° to the plane of the typical behavior of the interplanetary medium within sectors

3=4
 a
SOLAR WIND
E)tPLORER 33
796 THREE HOUR AVERAGES t
NOV. 1966 -FEB. 1967
NORTH

20

If
NOV. 1966 TO FEB. 1967

~ 16 F
1 ISOTROPIC

“z ~ABOVE (55Yo)
w 12 - F
v
fi8 - 75Y*
&
4 - !

-n
0246 8 10 12 ~14
B (nT)
(a)

~ BELOW (45 ‘/’)

t
TO SUN

t
e= -900

(b)

,2700

Ftgure 3-3. Observations of the interplanetary magnetic field. In this typ-

ical sample (a) the most common field magnitude is about 5
nT ( 1 y = 1 nT) but the average of the magnitudes is higher
because the distribution is skewed. (b) atid (c) show the rel-
ative frequencies of the directions of the field. It tends to lie
+
along the spiral direction appropriate to a velocity of about
+ = 180”
450 ksn/s (i. e., about 45°) and in the solar equatorial plane
(c) but there are significant deviations [adapted from Hirshberg,
1969] .

are relatively simple to carry out using sector boundaries as In the vicinity of the earth a typical magnetic field in-
convenient time markers. The results of some of these stud- tensit y IBI is about 6 nT, but hourly average values vary
ies are described in Section 3.3 on large scale disturbances. from about 1 nT to a high of 37 nT. The highest recorded
Observations from spacecraft that traveled to solar lat- single value was 170 nT observed on 4 August 1972 [Bur-
itudes of over 10° from the solar equatorial plane showed laga and King, 1979]. Figure 3-6 shows the distribution of
that the sector structure was dependent on latitude. The over 70000 hourly values collected from satellites at 1 AU
stmcture tends to disappear at higher latitudes, and the in- between 1963 and 1977. The distribution of the magnitude
te~lanetary magnetic field then has the sign of the solar IBI is skewed towards high vahtes. However, the distribution
magnetic field of the appropriate solar pole. The picture that of the log IBI is distributed normally with a mean of 0.75
has developed to explain these observations is shown in (IBI = 5.6 nT) and an rms deviation of 0.18. For about 3%
Figure 3-5. Plasma leaves the stin predominantly at high of these hours log IBI exceeded the mean by at least 20,
latitudes and flows out and towards the equator where a that is IBl > 13 nT. When the field is high it typically
current sheet is formed corresponding to the change in mag- remains so for many hours, forming distinct events. There
netic field polarity. The current sheet is tipped relative to were about 100 such events observed between 1963 and
the sun’s plane of rotation and waved so that as the sun 1974 of which 75% were associated with shocks or inter-
rotates the eafih passes through the current sheet, ‘experi- faces between interacting solar wind streams (Section 3.3.2).
encing periods of alternating polarities, that is, sectors. Periods of high-intensity field may persist for as long as 24

3-5
CHAPTER 3

● 00 ● ee be 1893

J16 B6 ● ** ● m ● * X*9 ● ea*a 1894

F12 D..s*xeeo ● xxoooooooooogo 1895

MIO be ● .*0 ● ooe ● **X ● oooa*e 1896

A6 D ● 000 ● eee ● 90 ● ● 1897

M3 ● *OO ● @eeee ● 08,,, ● 1898

M30 00000.0.. ● *OOOXOXO ● 006 1899

N
k
~ J26 B9e.0..xeeex9*ob ..9 . . . . . . 1900
Figure 3-5. Sketch of the current sheet responsible for the sector stmcture.
The sheet is warped and tilted relative to the plane of solar
J 23 B******* ● *O ● *oe**a 1901
rotation [Smith, 1979].

A19 B* ● *O*@eOe* ● 900 ● *XO ● ● 1902

S15 ● ****mix*..** ● **..*..* ● 1903 hours. The persistence of a given level of IBI can be found
from the autocorrelation function which falls to 0.5 at about
0/2 ● 000 ● *** ● ** .Oxo ● 0ea 1904
a 15-hour lag [Burlaga and King, 1979].
N8 )*** ● X******* ,o*e*oee*x 1905
An important magnetic field parameter is the component
of the field perpendicular to the X-Y plane. This component,
05 )X6 ● 000 ● 000 ● Ooooo.ooxoo 1906 B., plays a central role in the coupling of the wind and the
earth’s magnetic field such that when 13zis southward and

● **x...*.*g*a*a 866

Jl? ● 00000 ● e9xx*ooooo@og ● *.** 867 log B

B
F13 ● **m*** ● 00 ● *e, e*oee*aee ● 868 I 04 =

M12 869

Ar3 870

M5 871

872
2
~ “ x
as L
- J28 ●☛☛☛☛
*xoe8e**oe ● oe ● eex. . 873 o
& r’
J25 ●XX OOX*XO*OOXO’O ● m6mee ● * 874
2
=
A2 I B**e. ***m *.*b . . . . . . . . . . . 875 z

S17 ● 000000000 .0. . . . . . ...0.0 876

014 ●* 90)(. O* OQOOX*O .Ooooxoee 877 10’ I

I
I
NIO ● Xo XX* ..O ● X*O log B La
b O* O..OO. 878 i
~] o 0.4 0.8 I .2 1.6~
07 P* ● 00 ● * * - **O***.*. ● ** 879 # 1 I 1 I 1 I 1

I 0° e
0~~ 4 8 12 16 20 2r6
B (nT)
Figure 3-4. Two years of sector structure polarity. A small dot indicates
a day with polarity directed away from the sun; a big dot,
towards the sun; a cross, an ambiguous day; and a blank, a Figure 3-6. Distribution of 71431 hourly averaged values of the interplan-
data gap. Each line represents one solar rotation or 27 days. etary magnetic field magnitude IBI, and of log IBI. These data
The date of the first day of the rotation is given on the left- were collected between 1963 and 1977. The average value of
hand scale and the number of the rotation on the right-h~d B is 6.09 nT but the average of log B corresponds to a value
scale. Note the strong tendency for sector structure to repeat of 5.6 nT. ‘lIre most common value of B is about 5 nT [Burlaga
on subsequent rotations [adapted from Wilcox et al., 1975]. and King, 1979].

3-6
 SOLAR WIND
YEARLY DISTRIBUTIONS OF HOURLY AVERAGES
large, geomagnetic disturbances occur. Figure 3-7a shows
l=
distributions of GSE B. observed for each year from 1963
1963 “P &H
I nT
to 1974 [Siscoe et al., 1978]. The figure gives the number
,6

‘h%
1964

1965
, 58

k’%
32

.
of times IBzl fell within each 1 nT bin. The number of values
used in each histogram is also given. The distributions are
generally exponential, but with a high intensity tail. The
exponent changes in a systematic way with the solar cycle.
These data are collected in Figure 3-7b which gives the
‘L?\\.-_----.._--.-- probability that an hour chosen at random from the set of
over 50000 hours of observation will fall in a particular 1
nT bin. Of course, a figure of this type made using different
years of observation would differ somewhat from that shown.
The finding that the most probable value of B, is less than
1 nT is in general agreement with the notion that a steady
spherically expanding solar wind would have no B, com-
ponent when measured in the sun’s equatorial plane. The
I-I B, component is primarily due to disturbances such as waves
or interacting parcels of the interplanetary medium.

3.2.3 Protons and Electrons

‘ “x
The solar wind plasma consists of protons, doubly charged
helium ions, a small number of other positively charged
..~ L---—-
particles, and enough electrons so that the plasma is elec-
trically neutral. The Debye length is about 10 m. Since
commonly 95% of-the positively charged particles are pro-
tons, it is the protons that are usually referred to when
statements are made about the solar wind bulk properties
such as velocity, density, and temperature. In this section
the large scale general properties of the wind will be de-
- 1 1 I scribed for protons and electrons. Alpha particles and heavy
.- .,.
0 Iu ions will be dealt with in Section 3.2.4.
IBZI (nT)<”

(o)
I
[ TOTAL 1963-1974
52929 HOURS
o. I ~
1 PER DAY

I
: I
0.01 : I I PER WEEK
i
= I t
I I
a I I I PER MONTH
m
o 0.001 ~ I I I
K I I I
I 1-
n I
1 1
I 1
I I I PER YEAR
0.0001 ~ I t
I I I
I I -4
I I i I -----
I I L L -— ----
I I 1,
I I I I 1, I I
I I 1 11 I
0.000010
4 8 12 16 20 24 28
IBzl ‘nT) (~,

Figure 3-7, (a) The distributions of the magnitude of the hourly average value of the component of the interplanetary magnetic field perpendicular to the
ecliptic plane, lBzl. Data for each year from 1963 to 1974 are plotted. The data coverage varied from year to year. Note the changes in slope
and the high intensity tail. (b) Histogram of the occurrence frequency of IBzI based on the 52929 values used in the above distributions. The
ordinate gives the probabilityy that an entry selected at random will fall in the particular 1 y bin. Levels for selected average recurrence peritis
are marked. For example, the average time between observations of 137> BZ z 127 is about one month ISiscoe et al., 1978],

3-7
CHAPTER 3
Solar wind properties do not vary randomly on a time particles/cm’. Most of the extreme density values are due
scale of hours. Instead the wind is organized mainly into to rarefaction and compression arising through stream in-
periods of slow plasma and fast plasma, with the slow and teractions or solar flare disturbances discussed in Section
fast regions interacting. The structure of these streams will 3.3.3. The proton temperature TP, which increases with the
be discussed in Section 3.3.2. Here the interest will focus velocity, varies by a factor of twent y and has an average
on typical values of physical parameters. of 1.2 X 105 in this data set.
Early plasma probes detected protons and other posi- The proton distribution function varies widely and is
tively charged particles but did not permit detailed descrip- usually not well represented by a single Maxwellian. Figures
tions of particle distributions to be made. However, the first 3-9a and b show examples of two-dimensional proton ve-
three moments of the distribution could be determined, that locity distributions. Figure 3-9a shows a simple distribution
is, the proton velocity, density, and (less exactly) the tem- with a single peak, corresponding to a single velocity and
perature. A typical distribution of speeds is shown in Figure density. Figure 3-9b shows a distribution with a strong sec-
3-8. The direction of flow is almost always from the sun ond peak corresponding to a plasma having two beams with
different velocities and densities. The most usual distribu-
I I I I I 1 I I
tion is between these two and would show a single major
I

13,976 VELA 3 MEASUREMENT:

peak with evidence for a high velocity tail. Note that in
n JULY 1965 TO NOV. 1967 both main peaks the spread in contours is greater in one
AVERAGE FLOW SPEEO ‘ direction than at right angles, that is, the temperatures par-
400 km/s allel and perpendicular to some direction (the direction of
MEDIAN FLOW SPEEO = the magnetic field) are not equal and are often given sep-
380 km/s arately. The temperature TP is 1/3 (Tll + 2T1 ) where II and
L denote components parallel and perpendicular to the mag-
netic field. The average value of T1l/TL is 1.5 with a range
of 0.7 to 2.8. Distributions that have two velocity peaks
fi can be described as two streams of protons moving relative
1 I I I I I I to one another. The higher density streams are usually the
) 350 450 550 650 750 lower velocity ones.
FLOW SPEED (km/s) Solar wind electrons are more difficult to observe and
their properties have been described less fully than those of
Figure 3-8. A frequency distribution of solar wind speed [Hundhausen, the protons. In contrast to protons, electrons are subsonic
1972]. Both the average speed and the shape of the distri- at earth. Observed distributions of electrons are shown in
bution differ from year to year.
Figure 3-10 [Feldman et al., 1975]. Note that because of
the subsonic velocity, the spreads in particle speeds are
(within a few degrees). The major causes of non-radial flow much larger than the bulk speed. Very low velocity elec-
are aberration and interaction of fast and slow solar wind trons, which would allow an actual measurement of the bulk
structures. The amount of deviation from radial flow due to speed to be made, cannot be measured at all because of
the angular velocity of the source of the solar wind at the observational difficulties. The figure shows four cuts through
sun is still a matter of debate. Although there is considerable the distribution. Figure 3-10a shows particle fluxes in the
variation in the velocity distributions for different time pe- direction parallel and anti-parallel to the electron heat flux.
riods, they all have nearly the same range of values. In Since the electron heat flux is along the magnetic field di-
Figure 3-8 there are few measurements with velocities lower rection these two cuts are also in directions parallel and anti-
than 275 km/s and in data sets from several other years there parallel to the magnetic field. Note that fluxes for each speed
were none. In the sample shown, values higher than 650 are not the same in two directions, indicating that a sub-
km/s were rare but in other years the upper frequency bin stantial heat flux is carried by the electrons. In fact, it is
with non-negligible sample number was 775 krrt/s. Veloc- the electrons that regulate the heat flux in the wind as a
ities of over 1000 km/s occasionally have been reported whole. The curves in the upper panel represent fits to the
after extremely energetic flares occurred on the sun. Au- points, the fits to the lower speeds being Maxwellians. Note
tocorrelation among speed measurements is high for lags of that although the two Maxwellians are excellent fits at low
less than 40 h, reflecting the stream structure (Section 3.3.2). speeds, they fail decisively at high speeds. In much of the
Table 3-1 shows average values and ranges for 3-hour av- data reported in the literature, the higher speed particles are
erage speeds for about 3.5 years of data [Feldman et al., fit with a second set of two Maxwellians or a somewhat
1977]. The average velocity given there is 468 krrt/s, some- more complex function as shown here. In Figure 3- 10b the
what higher than that shown in Figure 3-8. data shown are from a plane pe~ndicular to the heat flux.
The density of the proton component is also ordered by Two data sets are shown: one for positive and one for neg-
the fast and slow velocity streams with an average of 8.7 ative directions along a line in the plane. The fluxes are the

3-8
 SOLAR WIND
Table 3-1. Average solar wind parameters

5–95% Range
Parameter Mean u Limit

speed (km/s) 468 116 320 to 710’

density 8.7 6.6 3.2 to 20

(proton/cm3)

MA 10.7 4.8 4.4 to 20

Ms 7.7 1.5 5.6 to 10

lB/ (nT) 6.6 2.9 2.2 to 9.9

T, (K) 1.2 x 105 9 x 104 lxl@to3xlo5

T. (K) 1.4 x 105 4 x 104 9xlo4t02xlo5

Te/TP 1.9 1.6 0.37 to 5.0

n.lnP 0.047 0.019 0.017 to 0.078

T. (K) 5.8 X 105 5X105 6 X ldto 15.5 X 105

TwiTP 4.9 1.8 2.3 to 7.5

P, 0.7 0.9 o.lltol.7

Be 1.3 1.9 0.3 to 3.1

p, + Be 2.1 2.4 0.5 to 4.6

C.(km/s) 50 24 30 to 100

Cs(km/s) 63 15 41 to 91 I
I
nformation from Feldman et al.,[ 1977])

same in these two directions showing that the electron dis- component electrons below 50 eV that can be well fit with
tribution function is symmetric about the magnetic field a bimaxwellian function are also known as the core elec-
direction. Again the data are fit with Maxwellians at low trons. The hot component or halo electron component is
speeds and more complex functions are fit to the residual defined as the remainder of the measured distribution func-
fluxes at high speeds. The low speed portions of electron tion after the cold component is subtracted. In this scheme
distribution functions are collision dominated, whereas the from 3 to 10% of the electrons belong to the hot component.
very high speed portions are collisionless. Observed distri- The energy at which f~ begins to rise above fC is about 60
bution functions are consistent with a theoretical model in eV. The hot and cold components generally move relative
which the entire distribution at all velocities arises from the to one another along the magnetic field direction so that no
effects of Coulomb collisions [Scudder and Olbert, 1979]. net current flows in the frame of reference moving with the
Although there is not yet general agreement concerning protons. The total bulk velocity is observed to be within
the best method for describing and interpreting electron about 50 km/s of the proton velocity. The total heat flux in
observations, the data presentation that is used here is that the solar wind is primarily carried by the hot electron com-
of Feldman et al. [1977], in which the measured flux dis- ponent. If both halo and core components are fit with bi-
tribution is broken up into low and high speed parts as shown maxwellian distributions then the average core electron tem-
in Figure 3-10. The electron and proton densities and bulk perature is 1.2 x 105 K, whereas the average halo electron
velocities are set equal, consistent with the observations. temperature is 6.9 X 105 K. The value of 1.4 x 105 K
The electron distribution function is described as a super- given in Table 3-1 is the temperature derived from the fit
position of two distributions, one a cold component fc and of the distribution function as a whole. As can be seen from
the other a hot component f~, so that f = f= + f~. The cold Table 3-1, the electron temperature does not vary as widely

3-9

\
“\
 0257 UTON MARCH 27, 1971

100 1916 UTON MAY 20, 1971

\
E
5 50

-50

-1oo
550 600
-loo~
450 500 300 350
RADIAL VELOCITY (km/s) RADIAL VELOCITY (km/s)

(a) (b)

Figure 3-9. Proton velocity distributions. Two examples (a,b) of two-dimensional proton velocity distributions contoured for equal number density.
Contours A. B. C and D correswnd to 0.001, 0.0032, 0.01 and 0.032 and contours E through N to 0. I to 1.0 in steps of 0.1 of the peak
contribution to the proton num&r density [Feldman et al., 1974]

,.-24
I I I
,. -24 IMP8
IMP8
11/05/73 1703 UT
-N 11/05/73 1703 UT
,0-26 _ . + ‘ 84”

-26
10

m
m ~: 10-28 -
@ w
1
E 10-28 ‘E
: :
L
k 10-30

,.-30 -s

,0-32 \
I
,0-32 I I I I I I 1

o 5 10 15 20 0 5 10 15 20

VELOCITY (103 kmls) vELOCITY (103 kmts)

(b)
(a)

Figure 3-10. Electron flux as a function of speed [Feldman et al., 1975]. (a) shows tbe flux parallel and anti-parallel to the direction of the electron heat
tfux. (b) shows the fluxes in a pe~ndicular plane.

3-1o
 ~

SOLAR WIND
as the proton temperature. The electron thermal pressure is helium abundances may be associated with impulsive in-
usually, but not always, greater than the magnetic field jections into the wind of hot or cool coronal plasma as a
pressure (~ = 8mnkT/B2, P= > 1). Thus, although the pro- result of solar activity but there does not appear to be a one
ton thermal pressure is usually less than the magnetic field to one correlation. Helium abundance also shows a trend
pressure (PP < 1), the average ~ calculated using the total toward larger values in higher velocity winds but this trend
thermal pressure is about 2 (Table 3-1). is time dependent and may change with the solar cycle. The
average abundances at velocities of 300 km/s vary between
390 and 4%, and at 500 ktis average abundances between
3.2.4 Other Ions 5% and 9% have been reported [Neugebauer, 1981].
The helium and hydrogen components of the solar wind
Since helium is the second most abundant solar con- have almost the same speed at 1 AU. However, differences
stituent, it was expected to be present in the solar wind in of up to a few tens of kilometers are common, and the
quantities that could be detected by the early solar wind helium is usually moving faster than the hydrogen. The
probes. This expectation was fulfilled with the flight of average reported value of the ratio of the helium to hydrogen
Mariner 11 in 1962 [Neugebauer and Snyder, 1966]. How- speeds (va/vP) ranges from 1.001 to 1.035 and is always
ever, the measured values of the abundance, velocity, and positive. The direction of helium motion relative to hydro-
temperature of the helium component of the wind were so gen is generally along the direction of the magnetic field.
different from what would be expected on the basis of simple At 1 AU the velocity difference Va – VPis in general pro-
theory that the behavior of helium is still not well under- portional to, but less than, the Alfv6n wave speed. Also,
stood. For a recent review of helium observations see
Neugebauer [1981]. Observationally, the helium properties
I I 1 I
are ordered by large scale structures such as sectors, velocity
streams, and flare disturbances. These properties will be VELA 3
JULY 1965- JULY 1967
discussed further in Section 3.3. The simplest helium pa-
rameters to measure are relative abundance, velocity, and
temperature. The major unexpected results are that in the
vicinity of the earth, relative abundances of helium by num-
ber, nm/nP, vary from less than l@3 to more than 0.3, helium
velocities tend to be a few km/s larger than hydrogen ve-
locities, and the temperature of helium is usually about 3
or 4 times that of hydrogen.
Helium in the solar wind is almost always completely
doubly ionized, as would be expected because of the high
coronal electron temperatures. However, a few events have
been reported in which He+ /He++ reached levels as high
as 0.1 [Schwenn et al., 1980] and 0.3 [Gosling et al., 1980].
These events have been interpreted as the inclusion into the
wind of material from relatively cool chromospheric prom-
inences.
The distribution of relative abundance values of He++
by number measured by the Vela 3 satellites is shown in
Figure 3-11 [Robbins et al., 1970]. For this data set the
values ranged from 8.1 x 10+ to 4.2 x 1~’. The yearly
average abundance between 1962 and 1976 varied from 370
to something over 5%. The helium abundances are not dis-
tributed randomly in the plasma but instead show systematic
behavior that has not yet been satisfactorily explained the-
oretically. Most values of na/np = O. 15 in Figure 3-11 were
observed in solar wind reliably associated with major solar
flares [Hirshberg et al., 1972]. Although most helium abun- 0.10 0.20
dance enhancements are associated with flares, regions of HELIUM TO HYDROGEN DENSITY RAT I O
solar wind containing between 970 and 1570 helium plasma
are sometimes seen in low speed wind with no association Figure 3-11. The distribution of solar wind helium to hydrogen number
density ratios [Robbins et al., 1970]. The helium enhance-
with major flares or interplanetary shocks (Section 3.3.3). ment events reflected in the high abundance tail are often
In general, the data suggest that unusually high solar wind associated with major solar flares.

3-11
CHAPTER 3
Va — VPis strongly dependent on the ratio of the solar wind another context in Section 3.1.1. The two uses are so dif-
expansion time to the Coulomb collision slowing downtime. ferent they almost never cause confusion. ) me heliocentric
The expansion time scale is defined as the time required for distance at which freezing in occurs depends somewhat on
a solar wind ion to move through one density scale height. the type of particle involved because of the details of the
It has been suggested that the helium is accelerated to higher interactions. In slow, low temperature regions of the wind
velocities than hydrogen by waves in the interplanetary me- the abundances of 06+ to 07+ ions in the available data
dium but as yet there is no direct observational confirmation. set indicate a freezing in temperature of 2.1 x 106 K and
The temperature of helium is almost always higher than the abundances of Fe’+ through Fe’s+ yield 1.5 x 106 K.
that of hydrogen, and it appears that there is a tendency for Other plasma, associated with solar wind accelerated by
the thermal velocities of the two components to be equal flares showed oxygen and iron freezing-in temperatures of
rather than for the temperatures to be equal. The most prob- 3 x 106 K [Bame et al., 1979], and iron charge states have
able value of Ta/TP is between 2 and 4 but values as high indicated temperatures as high as 1.6 x 107 K.
as 10 are occasionally observed. Average values of Ta~P
calculated from different data sets have varied from about
3 to 5. It is not known if this difference is due to actual 3.2.5 Waves and Turbulence
changes in the solar wind or to inaccuracies in the measure-
ments. There is a positive correlation between Ta/TP and Observations of the solar wind show that there are vari-
v. — VP.The behavior of T. is not understood theoretically ations in almost all quantities on all time scales. For ex-
but it has been suggested that the waves that can prefer- ample, as discussed in Section 3.2.2 the interplanetary mag-
entially accelerate helium may also heat the ions by an netic field of a uniform undisturbed wind should not have
amount roughly proportional to their masses. any Z component; however, hourly average Z components
In addition to 4He ions, solar wind 3He ions have been are about 1/3 the intensity of the field magnitude. This Z
detected. The 4He + ‘/3He + + ratio has been measured both component is due to the presence of waves and discontin-
from foils left on the lunar surface and directly using an ion uities in the plasma.
composition experiment on ISEE 3. The results of the two Considering the solar wind as a turbulent medium, power
techniques are in agreement. The ISEE experiment found spectra have been calculated and compared with the theo-
on average 4He + ‘/sHe++ of 2.1 t ().2 X 103using more retical predictions based on magnetohydrodynamic turbu-
than 4000 observations; however, the distribution was very lence theory. Spectra of the magnetic field, particle density,
broad. Low ratios, which lasted a few hours, had a tendency and particle velocity variations have been given in the lit-
to characterize periods of low 4He + + flux [Ogilvie et al., erature.
1980]. Magnetic field data lend themselves to these studies
Some ions heavier than helium have been observed in because of the good time resolution and accuracy of the
the interplanetary medium. The observations are difficult to measurements. Examples of spectra in the 1V’ to l&1 Hz
make because of the low ionic abundances. However, neon range are given in Figures 3-12a and 3- 12b [Hedgecock,
has been collected on foils on the moon, and plasma probes 1975]. If the field could be described as that ideally present
have been able to resolve oxygen, iron, and silicon at times in a spherically expanding constant velocity solar wind with-
when the plasma temperatures are low. The relative abun- out fluctuations, there would be no components in the two
dances are within a factor of ten of coronal abundances, transverse directions. The top curve in Figure 3-12a gives
and the speeds tend to be the same as the speeds of the the power in the variations of magnitude. The power levels
helium observed at the same time. There does not appear in the magnitude are much lower than the levels in the
to be any further acceleration of heavy ions. The temper- components. The spectra of the two transverse components
atures however, apparently increase further with higher ion (T and N) are very similar and have been averaged to pro-
mass and measurements of oxygen temperatures suggest that duce the transverse spectrum Tr. A composite spectrum of
the thermal speeds of all constituents tend to be equal in the transverse power is shown in Figure 3- 12b. Note the
the solar wind [Ogilvie et al., 1980]. different spectral indices used to fit the data in the different
The ionization state of heavy ions provides important frequency ranges. The magnetic field power spectrum for
information on the electron temperature at their source in periods greater than a day tends to flatten. The power spec-
the corona. The ionization state of the atom in the collision trum slope varies from day to day, or for lower frequencies
dominated corona is determined by a balance between ion- from sample to sample. In the IW5 to 1@3 Hz range both
ization by electron collisions and recombination due to ra- the slopes and the absolute value of the power variations
diative and dielectronic processes. As the solar wind moves are consistent with a factor of 10 variation, [Hedgecock,
away from the sun, collisions become more rare, and beyond 1975]. When the spectrum was fitted with a line given by
some heliocentric distance the ionization state of the particle F, the average spectral exponent N for 17 periods in 1972
remains unchanged. This process, called “freezing in,” per- was 1.27.
mits the particles to be used as tracers of their sources. Examples of power spectra of solar wind fluctuations in
(Note that the term “freezing in” has already been used in positive ion flux are shown in Figure 3-13a [Neugebauer,

3-12
 SOLAR WIND

‘~c -1

---”.% ~
~1
-- --:

x ,%
~z ~~ P(f)= A/( fO+f)’ ~
● 1- .h:\{l.5
c ASYMPTOTE
●
●
● ✚☛ 0—
‘%.
E ●
✘☛
●☛
~
=
>
4 ~
‘b,,
d“ \ ~-1.63
●* ● 1- \
G
●
●
L g3 =
.*
● J
\2
9
● * 2 .
N “8 1-
~2 .
$6 w
a
m .
0 ..
m “.
m- ●.O
0 ●. .
● gl ..+
:! ; %[ . .++.
L _FROM HOURLY+ ➤ FROM 6
>6 AVERAGES SECOND POINTS-””””””
* (WITH DATA GAPS) (NO DATA GAPS)
●
z o‘ -1
● -7 -6 -5 -4 -3 -2
5 FREQUENCY (log10 Hz)
n I
(b)
<6
E
+ ●
v ●
Figure 3-12. (a) Power spectral density (PSD) of the interplanetary mag-
w ●●H ● * netic field. The R direction is along the average spiral di-
=
VJ
. ●oeo: ‘*8, !R’1 rection of the field, N is in the solar meridian plane and T
~ :*
% completes the system. B is the magnitude of the field. The
a6 components are as marked. TR is the average of the trans-
“t. ‘~
g ● verse (N, T) spectra [adapted from Hedgecoek, 1975]. (b)
●
● ●:
●9 A composite mean transverse power spectrum for 1972
n ● o coO%T’
‘* [Hedgecock, 1975].
. ... ●

●
●
●., .
5 ~ ●**. ““%
●9
~
.6
gyroradius. A long period velocity spectrum is shown in
Figure 3- 13b from Coleman [1968].
4 : Frequently, fluctuations in the solar wind can be rec-
ognized as distinct types of waves andlor discontinuities.
4
:STANDARD ERRORS/BANDWIDTHS D;scontinuities will b; discussed in Section 3.3. Although
10 DAY I yAY a kinetic theory treatment is required for a proper description
I of the behavior of waves, the magnetohydrodynamic theory
I 1 1111111 1 1 I Illld I 1 1111111 1 I I Ill&
3
-7 -6 -5 -4 -3 is useful for classifying the waves [Barnes, 1979]. There
FREQUENCY (i OrJIO Hz)
are three types of MHD waves, the transverse Alfven (or
intermediate) wave, and fast and slow magnetoacoustic waves.
(a) The transverse Alfv&n wave is non-compressive with con-
stant magnitude of the magnetic field, but not constant di-
rection. The fluctuating magnetic field and velocity are re-
1975]. Curves A and D are for the noisiest and quietest 100 lated by
spectra of the 1728 spectra inthedata setused in this study.
In the 10second region these twopower densities differ by b = * (47rp)1’2v (3.17)
more than a factor of 100. Curves B and C are two different
types of averages of the entire data set. In curve B the values where b is the magnetic perturbation, v is the velocity per-
of the power density are simply averaged. In curve C the turbation, and p is the constant density. The velocity of an
logarithms of the power spectral density are averaged to Alfv6n wave is given by CA = B/(41rp)”2. Typical values
avoid giving too much weight to the shape of the spectra of CA are given in Table 3-2.
from the most turbulent intervals. The error bars represent The two magnetoacoustic modes are compressive and
90% uncertainty levels. The enhancement at high frequen- their magnetic fluctuations are nearly linearly polarized. The
cies is probably due to fluctuations at the proton thermal velocities of the fast and slow waves are given by

3-13
CHAPTER 3
I
A’

_pERloD~~:p~; ~ (cptt) _
FREQ
(days)
10 64321684 2 I 2 4816 ‘2.7 10.8
10

15 I +
10 ‘1=
7N
I B E
1-
N
‘m
*

E
v
c
NOISIEST
< POWER >
H 14
g i-
&

m
o
Z
Z106
u
c1
a
I MARINER 2
-1
8

E
o
< POWER> :,.5 — 29 AUG -29 SEP, 1962 f I ) ~
n ---30 SEP -31 OCT, 1962 (2)
v \ 3
“r

13
POWER>

QUIETEST
<POWER>
104
L -27
TOT
TOT

DAY PD
PWR ( I ) = 8360
PWR (2) = 9959
(kmlsec)~
(km/see)
‘i

I 1 1 1 1111 I I I 1111111
I 1111111 1 I Illlltl 1 Illm
103
0. I I .0
10-7 10-6 10-5 10-4 10-3 I 0-2
FREQUENCY (Hz)
FREQUENCY (Iiz)
(a) (b)
Figure 3-13. (a) Four different types of averages of the power spectral density of variations in the total charge flux of positive ions [Neugebauer, 1975].
(b) Power spectra of the solar wind veloeity in 1962 [Coleman, 1968], (Reprinted with permission from the American Astronomical Society
O 1968.)

vm=– ;‘[ (
(cl + c:) * c%+ c:
to be observed far from their place of origin. The Alfv6n
wave, however, is not damped.
By far the most commonly observed wave in the solar
wind is the AlfvLn wave. An example is shown in Figure
1/2

)1
3-14. In this figure R is the radial direction and N and T
—
2C;C: Cos 2 e (3.18) are transverse directions. The veloeit y and field component
variations are approximately related according to the expres-
sion above. The magnetic field magnitude and the plasma
density show very little variation. Interplanetary fluctuations
where C~ is the speed of sound, 0 is the direction of the are dominated by large amplitude aperiodic AlfvLn waves
wave normal relative to the field, and the plus is taken for more than 50% of the time [Belcher and Davis, 197 I ]. The
the fast mode and minus for the slow mode. Typical values clearest examples occur in high velocity streams (Section
of Cs are shown in Table 3-1 where Cs = [k(T. + 3TP)/ 3.3. 2) and their leading edges. These waves are propagating
m]”2. This expression assumes the electrons respond iso- outward in the rest frame which implies that they are pro-
thermally and the protons adiabatically to one dimensional duced in or near the sun below the critical point of the flow
sound waves as expected from the electrons role in heat r=. They have not yet been actually observed in the vicinity
conduction. In the limit of small amplitude waves, fluctua- of the sun and this is a very difficult observation to make
tions can be thought of as a superposition of these three without in situ detectors. In the vicinity of the earth, changes
types of MHD waves. However, the waves in space are not in the magnitude of B are small with AIB1/\Bl - 0.06 whereas
small amplitude so that their behavior is more complex than the variation of the components is relatively large with ABi/
described by small amplitude MHD theory. The magnetoa- IBI – 0.4 where i indicates a component IBurlaga and Turner,
coustic modes will be damped and so will not be expected 1976]. The direction of minimum variance of the field is

3-14
 m
SOLAR WIND
ALFViN WAVES IN SOLAR WIND

4 25

bR 0 0 VR (0)

-4 -25

4 25

bT 0 0 VT (b]

-4 -25

4 25

0 0 VN (c)
bN
-4 -25

4 6

B 0 .- -- 0 N (d)
8 Iz 16 Zu 24
TIME [hours)

F!gure 3-14. Alfven waves in the solar wind. Curves a, b&c show that the strong variations of the magnetic field compmrents occur simultaneously
with like variations of the veloeity. Curve d shows that the field magnitude and density were both very steady [Belcher and Davis, 1971].

along the main field. The observed frequencies of the waves purpose the hydromagnetic approximation is used, that is.
are from the order of minutes to hours. In Figure 3-12a the the plasma flow will be approximated by the flow of a
low power levels in the spectrum of IBI compared to the dissipationless perfect gas. The magnetic field will be as-
spectrum of the variations of components is due to the dom- sumed to follow the flow. The approximation will be suf-
inance of A1fv6n waves. ficiently accurate to organize the observations of interplan-
Observations also show fluctuations in the solar wind etary disturbances and discontinuities discussed here. The
that are non-AlfvEnic. For example, the leading edges of quantities that must be conserved across a hydromagnetic
high velocity streams contain large amplitude Alfv6n waves, discontinuity are mass, momentum, energy, the component
a non-Alfvenic component, and also perhaps an Alfvenic of the magnetic field perpendicular to the discontinuity, and
component propagating toward the sun [Belcher and Davis, the component of the electric field tangential to the discon-
197 1]. These latter two types of waves are probably of local tinuityy. The jump conditions have been given by Spreiter
origin. and Alksne [1969].
There are very few unambiguous identifications of in- For a tangential discontinuity:
terplanetary magnetoacoustic waves. One case reported by
Burlaga [1968] is probably a fast magnetoacoustic wave Vn =Hn=O
(3.19)
propagating across the magnetic field [Barnes, 1979]. [v,] + O, [H,] + O, [p] + O, [P + H2/87r] = O, ‘

for a contact discontiunity:

3.3. LARGE SCALE STRUCTURES
AND DISCONTINUITIES
v“ =0, H.+ O,[V]=[HI =[PI=O, [PI+ 0,
(3.20)
3.3.1 Discontinuities
for a rotational discontinuity:
Before discussing the types of disturbances that appear
v“ = t Hn/(4mp) “2, [v,] = [H,]/(4mp)”2
at 1 AU in the solar wind, it is helpful to describe the (3.21)
classifications of discontinuties more completely. For this [PI = [p] = [v.] = [v*I = [H*I = [H.]= 0,
3-15
CHAPTER 3
and fast and slow shock waves satisfy in the field magnitude. The distributions of these quantities
are shown in Figure 3-15, adapted from Tsurutani and Smith
[1979]. Small angle and magnitude changes occur most
v. + O, [p] >0, [P] >0, [Hn] = O
frequently. Discontinuities are usually quite abrupt. A study
(pvn) 2 (pvn) > (pvn) of Mariner 4 magnetic field data found’ that 50% of the
(3.22)
fast rot slow discontinuities analyzed had thicknesses less than 3500 km
and 10% less than 500 km [Siscoe et al., 1968].

‘1andH2(=)thr0ugh(~~)
shwkwaves A class of discontinuities called directional discontin-
uities (DD) has been defined as events that are either ro-
tational or tangential discontinuities and show abrupt changes
in magnetic field of at least 30° in consecutive 30-second
Here v, p, P, and H are the velocity, density, pressure, and magnetic field averages. It is not possible to unambiguously
magnetic field intensity respectively in a frame of reference distinguish between tangential discontinuities and rotational
at rest with the discontinuity. Subscripts t and n refer to discontinuities among DDs, but is was found that in low
components tangent to and normal to the discontinuity sur- velocity solar wind tangential discontinuities are most com-
face and [] denotes the difference in the enclosed quantity mon whereas DDs consistent with outward propagating ro-
on the two sides of the discontinuity y. The rotational dis- tational discontinuities become the predominant mode in
continuity is the steepening of an Alfv6n wave. Fast and high velocity winds [Solodyna et al., 1977].
slow shocks are the steepening of fast and slow magne- Shock discontinuities are also observed at earth and are
toaccoustic waves. Contact and tangential discontinuities do usually associated with solar events such as flares or perhaps
not propagate in a frame of reference at rest with the un- errupting prominences. The fast solar wind from such an
disturbed plasma. The three other discontinuities do prop- event accelerates the ambient wind in front of it. If the
agate in that frame. Contact discontinuities are not expected difference in velocity between the fast wind and the ambient
to be observed in the solar wind because they are difficult wind is larger than the speed of a fast magnetohydrodynamic
to form and would’ decay away quickly. Tangential and wave, a shock will form in the ambient wind at a position
rotational discontinuities and fast and slow shocks have been well in front of the boundary between the ambient wind and
observed. Slow shocks are very rare, and if not specifically the flare plasma. This shock is known as the forward shock.
stated otherwise, the word “shock” will be used here to refer As the fast plasma accelerates the ambient plasma, conser-
to a fast shock. vation of energy and momentum require the fast plasma
At earth, three types of discontinuities are common, fast itself to slow down. If the slowing down is severe enough
shocks and rotational and tangential discontinuities. Shocks another shock called the reverse shock will form within the
are formed in the solar wind by the steeping of the interaction fast plasma on the sun side of the boundary. When a detector
region between fast and slow solar wind parcels. They are passes through a forward shock, there will be a sudden
commonly associated with impulsive high speed solar wind increase in magnetic field, density, and temperature and a
from flare events. The interaction region between long lived decrease in velocity. Passage through a reverse shock will
streams of slow and fast solar wind rarely steepens into appear as a decrease in magnetic field, density, and tem-
shocks by the time the solar wind regions arrive at earth. perature and an increase in velocity. Both types of shocks
Rotational and tangential discontinuities are commonly ob- have been observed as shown in Figure 3-16 [Burlaga and
served at 1 AU and are apparently formed close to the sun King, 1979]. Forward and reverse shocks are found in flare
but the mechanism of formation is not known [Tsurutani associated events and reverse shocks occasionally are formed
and Smith, 1979]. Since the magnetic fields are measured during the interaction of long lived streams of slow and fast
in space with a smaller sampling interval than particles, solar wind. However, shocks are not as common as DDs
most studies of discontinuities have used observational cri- and occur once every few days at sunspot maximum and
teria based on changes in the magnetic fields. The frequency less frequently at low sunspot number.
of occurrence of discontinuities will depend on the selection The jump conditions at shocks (Rankine Hugoniot con-
criterion but with carefully designed criteria it is found that ditions) allow the strength of the shock and the direction of
tangential and rotational discontinuities occur at a rate of the normal to the shock to be determined if the plasma and
one or two per hour. The mean time interval between arrivals field properties are known accurately enough. For strong
of discontinuities can typically be described by a Poisson shocks the direction of the normal to the shock can be found
distribution over the interval of from perhaps 10 minutes to from
an hour, but the rate of occurrence varies strongly on a time
scale of several days. Even on a time scale of several solar n = f (Bl x B2) X (B2 – BI), (3.23)
rotations, the average rate varies by a factor of more than
2 [Tsurutani and Smith, 1979]. Two important parameters where B, and B2 are the observed fields before and after
that characterize the discontinuities are the angle through shock passage and the sign depends on whether the shock
which the magnetic field rotates and the percentage change is forward or reverse. A wide range of shock normals has

3-16
 SOLAR WIND

I I I 1 1 I I I I I I 1 I I I

1-
J 1 1 I I I 1 1 1 I I I I 1 I 1 FI 1 1 i

30 60 90 I20 I50 I I 0.5 I.0 1.5 2.0

q (degrees) A31BL

(0) (b)

Figure 3-15. Nomalized distributions of(a) tbefield rotation mgleqacross discontinuities andof(b) themlative field change lABl/BL, where BListbe
larger of the field magnitudes oneither side of the discontinuity. ~etwoplots ineachpanel wetbenomalized distribution plus or minus
onestandard deviation [Tsumtani and Smith, 1979].

been observed, and although most normals lie within45° variable, but both quantities also exhibited a good deal of
or so of the earth-sun line, almost any orientation can occur. order (Figure 3-17). During a 27-day rotation period of the
In addition, the flares that cause the strongest shocks tend sun, several intervals of higher than average speed occured.
to occur within 45° of the central meridian of the sun but As seen in the figure, the high speed intervals tended to
sometimes flares occurring near the solar limbs cause shocks recur on the next rotation. The proton density typically
to appear at earth’s orbit. The Mach numbers of most of maximized before the velocity. These high-speed corotating
the few shocks for which determinations have been made (that is, rotating with the sun) streams are one of the out-
tend to be rather low, about 2 or 3, but stronger shocks do standing organized features of the solar wind. A particular
occur [Hundhausen, 1972]. stream may persist for many months or longer than a year.
They are responsible for recurrent geomagnetic storms that
characterize the declining phase of the sunspot cycle. The
streams are imbedded within the sector structure, that is,
3.3.2 Solar Wind Streams magnetic sector boundaries typically precede the velocity
rise by hours or days. The solar sources of the streams are
The earliest continuous observations of the solar wind the low temperature, open magnetic field regions known as
were those from Mariner 2 [Neugebauer and Snyder, 1966] coronal holes [Neupert and Pizzo, 1974].
which found that the solar wind speed and density were The properties of the stream structure have been exten-

3-17
CHAPTER 3
INTENSE MAGNETIC FIELDS INTENSE MAGNETIC FIELDS
NEAR A SHOCK NEAR AN INTERFACE
NO SHOCK

—
30

7
LI

tll c

m
c
20

10

750

zg 650
zal 550 ; 550
m
s
; 450 ; 450
s
; 350 350

I 00 I 00

n n
1 1
E 10 E 10
u
s
c
I I
x I 05 : 105
1- 1-
4
ln
,“—
104
15 17 19 21 II 13 15 17 19 23252729311357 9
MAY 1972 JAN 1967 JAN 1974 FEB 1974

(a) (b)

Figure 3-16, Shocks and stream interfaces. (a)gives hourly average data fortwoexamples ofshmks detected inthesolw wind. Note therise in velocity,
temperature, density and magnetic field at the solid vertical line. The dashed Iineindicates theend of the shocked re~ion.
-.. (b) showsa stream
interface. The interface region ischaracterized byanabrupt drop intensity and rise in temperature. It is the region between the density and
temperature maxima and is bounded by the vertical lines. Although both the shocks and the interface region are characterized by very high
magnetic fields in these examples, that feature does not always occur (see Figure 3-23 forma~netic field hehavior across a shock).,,. lBurlaea
=
and King, 1979]

sively analyzed. The structure observed at 1 AU is partly material will form on both sides of the bounda~ between
due to the structure of the wind as it leaves the sun, and the streams.
partly to the evolution of that structure due to the interaction The boundary between low and high speed streams has
of low and high speed streams emmitted from a rotating a distinctive character and is called a stream interface region.
sun. An intuitive idea of the interaction is easily developed It is characterized by an abrupt drop in density, a sharp rise
by the following picture. Assume low speed wind leaves in temperature, and a rise in velocity that may be small. An
the sun’s neighborhood to the west of the source of high example of a stream interface is given in Figure 3-16 [Bur-
speed wind, that is, at a position further along in the di- laga and King, 1979]. This figure also shows an example
rection of rotation. This means that as the sun rotates the of an interface associated with a very intense magnetic field.
high speed region rotates to the angular position previously Magnetic field change is not a defining characteristic of
occupied by the low speed source region. Then, viewed interfaces and often is not present.
along a radius vector to the sun, the low speed parcel of A study of the characteristics of selected solar wind
wind will have a high speed parcel of wind overtaking and streams was made by Gosling et al. [1978] who superposed
compressing it. As the high speed material increases the data using 23 very abrupt interfaces as the zero time point
velocity of the low speed material, the low speed material for the data. The data were extremely well ordered by this
will become compressed. At the same time, due to conser- procedure, as is seen in Figure 3-18. The figure gives the
vation of momentum, the high speed material will be slowed one hour average data for 100 hours both before and after
and its density will increase. Thus a ridge of high density the interfaces. The error bars are the estimated error of the

3-18
 a
SOLAR WIND
800 I I I I I I I I I I I I I I I I I I I I I I I I I I I
x I 00
I I
ROTATION 1967
“ “P
600 ‘1
~j f) p 10
i J
~ /[ %
‘Y P
j n’ Y
400 -lJ
\. I
hi,,,,,+:
1 I I I
AuG. 27 SEPT. I SEPT. 6 SEPT II SEPT. 16 SEPT. 21
800 I 00

600 10

400 I

SEPT. 23 SEPT. 28 OCT 3 OCT 8 OCT. 13 OCT. 18

800 I 00
0
e
* ‘“E
\ v
E 600 10 ;
x
$
: &
G 400 1%
~
w
> OCT. 20 OCT. 25 OCT. 30 NOV. 4 NOV. 9 NOV. 14
800 I 00

600 10

400 I

NOV. 16 NOV. 2 I NOV. 26 OEC. I DEC. 6 DEC. I I

800 w I I I I 1 I I 00
‘ROTATION 1971
‘i’’’’’’’’’’’’’’’’’’’”-

600 wr~~il- “- ~’~L.”” ●

10

r I
1~~ 1<

II11
~{ -1.
400 -

L I 1 I I 1 I I I I I 1 I 1 I 1 I 1 1 1 I 1
DEC. 13 DEC. 18 DEC. 23 DEC. 28
1962

Rgure 3-17 Three-hour averages of the solar wind proton density (light line) and flow speed (heavy line) from Mariner 2 on its flight toward Venus in
1962. Each panel presents 27 days of data. Note the velocity stream structure and the tendency for recurrence of streams on each solar
rotation [Neugebauer and Snyder, 1966].

3-19
CHAPTER 3
N I I I I I I I I I I I ‘~
‘E
‘1 -lo

1
~ 12 x 10

$ 4

+J
I 1 I I I I I I 1
-4

I I I I I I I I I
.
-40 – ●

m
‘E d

230 – .
>
1-
G
$20 —
a :1 -

a -m
t I !
o~ -4 -2
POSITION RELATlk (E TO INTERFACE (days)
(a)

I I I I I I I I
●

5
● ; 12 — 10
t i x
T

4
a
x
n
J
o.031-
a
~1 --i
I 1 I , I , 1 I I 1 I , I I
, I ,
I I I I I I I I I 1 I I I I + I
I

E
Zz

-4 -2 0 2 4
POSITION RELATIVE TO INTERFACE (days)

(b)

3-20
 SOLAR WIND
, I
I , 1 , I
I I I I I I
●

5
x 10

: 1.4

I I , I 1 I I I 1 I 1 I 1 I 1 I I
r ,
I I I I I
5
-Xlo

I 1 I 1 I I t I , I
-4 -2 0 2 4 -4 -2 0 2 4
POSITION RELATIvE TO INTERFACE (cloys)

(c)

Figure 3-18. Su~~osed e~hanalysis ofone-hour average plasma pro~flies in 23 streams. The zero time is defined as being the time of an abrupt
interface. No streams with more gradual interfaces (Figure 3-16) were inchsded in the analysis [Goslinget al., 1978].

mean. Figure 3-18a gives the proton thermal pressure, the density of about 7 cm-3 at the interface. This discontinuity y
proton density, the flow angle in the ecliptic plane and the takes place on a time scale of minutes.
flow speed. The interface occurs when the speed is still Figure 3- 18b gives the temperatures of the protons and
quite low, about 400 km/s. Three hours after interface pas- electrons parallel and perpendicular to the magnetic field.
sage, the average speed is about 480 km/s, but the maximum All four quantities change discontinuously at the interface
average speed of 560 km/s is not reached until 1.5 days and both electron temperatures return to near their low speed
later. The plasma near the interface is accelerated by the values three hours after the interface passes, but both proton
stream-stream interaction. Since the interaction region has temperatures remain elevated. Note that the temperature
the shape of an Archimedes spiral (for the same reasons scales are not the same for the protons and electrons and
discussed in Section 3.2.2, for the magnetic field direction) that the electron temperature increase is relatively small.
the low speed plasma is deflected toward the west whereas The properties of the helium in the solar wind also show
the high speed plasma is deflected towards the east. A shear organization when analyzed according to stream structure.
flow occurs at the interface. Magnetic sector boundaries Figure 3- 18C shows that the abundance of helium minimizes
occur in the low speed wind preceding the interfaces. They about a day before the interface and that both the helium
typically appear from 1.5 days to 1.5 hours before the in- temperature and the ratio of alpha particle temperature to
terfaces. The total pressure (that is, particle plus magnetic proton temperature are discontinuous. In most of the stream
field) peaks approximately at the interface although the pro- structure the helium is, on average, flowing faster than the
ton density and proton thermal pressure are already declining hydrogen, but during the day before the interface is crossed
at that position. On the average there is discontinuity in the relative velocity changes sign, and in this region, the

3-21
CHAPTER 3
proton speed exceeds the alpha particle speed by a few the driven gas is shown in Figure 3-19, but its exact position
km/s. is often very difficult to identify observationally. A reverse
shock may form behind the discontinuity.
The driver gas itself, which typically arrives at earth
3.3.3 Solar Flare Disturbances about 5– 10 hours after the shock, is often characterized by
enhancement in the helium abundance that commonly ex-
The second major interplanetary structure that orders the ceeds 15% helium by number and may reach more than
solar wind is the solar flare disturbance. Major flares eject 20%–25~0 [Hirshberg et al., 1972]. On some occasions the
high speed plasma into the wind, which results in a dis- helium enhancement is not seen or it may be seen later in
turbance propagating to the earth and beyond. The disturb- the event. In Figure 3-19 this is interpreted as indicating a
ance is shown schematically in Figure 3-19 [Bame et al., patchy distribution of helium enhanced plasma. The number
of particles released into interplanetary space by major flare
events is estimated to be on the order of 1039–10@ particles
* POSSIBLE
PLASM.
GEOMETRY
onlv!NG
OF FL. RE EXPELLED
A SHOCK wav E
or 1015–10]6 g. The driver gas is often also characterized
by low proton [Gosling et al., 1973] and/or electron tem-
peratures and bidirectional streaming of electrons. These
observations suggest that the magnetic field can form closed
loops as shown in Figure 3-19. Such loops have apparently
been observed [Burlaga et al., 1981]. Examinations of the
ionization states of heavy ions in plasma associated with
flares indicate that at least part of the flare ejects originates
in coronal regions with temperatures of 3-4 X 106 K. The
driver gas has been slowed down on its way from the sun
to the earth by interaction with the ambient solar wind. Thus
the observed velocity of the driven gas will be less than the
mean transit velocity of that gas from the sun to the earth.
Typical values of the ratio of observed velocities to mean
transit velocities for a series of major flare-shock events is
about 0.8 [Hirshberg et al., 1972]. The interplanetary mag-
0 ‘“N netic field during flare disturbances often reaches very high
values (3040 nT) in the direction perpendicular to the eclip-
Figure 3-19 A schematic diagram of a possible geometry .— for the solar tic plane. Since the velocity of the driver and disturbed gas
wind disturbance caused by a major solar ttare [Bame et at., is high, and if these strong fields are southward, then the
1979]
solar wind will interact strongly with the earth’s magnetic
field and will drive impressive geomagnetic storms and au-
1979]. The rapidly moving flare plasma compresses the roras. The storms reflect the structure of the flare disturb-
ambient wind before it and a shock is formed. Studies in- ance. The classical storm begins with a sudden commence-
dicate that for major disturbances the shock expands on a ment caused by the abrupt increase in dynamic pressure on
broad front so that its shape when the shock arrives at 1 the magnetopause that occurs when the shock is crossed.
AU can be approximated by a half circle of radius 0.6 AU The initial phase of the storm, during which the earth’s field
and center 0.4 AU [Hirshberg, 1968; Rickett, 1975]. At the is enhanced but not disturbed is due to the shocked region
shock the density, proton temperature, magnetic field, and flow when the magnetic field in that region is northward.
velocity increase abruptly as the shocked gas is entered. When and if the field turns southward the main phase of
The density and velocity may or may not continue to in- the storm develops.
crease in the post shock region depending on the details of
the delivery of energy from the flare site to the solar wind
and also on the structures through which the shock propa- 3.4 LONG TERM VARIATIONS
gated on its way from the sun to the earth. For example, if
a shock is observed in the low speed wind preceding or
following a stream, the density and velocity change at the 3.4.1 Solar Cycle
shock will be relatively large. However, if observed near
the density peak or in the high speed wind, a disturbance It was expected that the interplanetary medium varied
with the same initial conditions at the sun will show a with the solar cycle even before in situ observations of the
relatively small change in velocity, although the change in solar wind began. Fluxes of particles from the sun were
density may still be considerable [Hirshberg et al., 1974]. believed to be the cause of geomagnetic activity and aurora
A discontinuity between the shocked ambient plasma and and it was well established that these phenomena exhibited

3-22
 SOLAR

a solar cycle variation. The variation of the solar output

during the eleven year solar cycle has many manifestations; I 50

for example, sunspot numbers, major solar flares, and the 100 - [a)

solar wind velocity. Distinctions should be made between

50 -
the solar cycle as expressed in each of the various quantities.
1
Although all solar cycle variations will have the same period 0
1964

they may have different phases, relative amplitudes, and

500 -
shapes. It is common practice to use the variations of sunspot t
numbers as the standard for the solar cycle and to compare \ [b)

450 -1
I
the changes in all other quantities to the sunspots. Here the k,

term “solar wind cycle” will be used to denote the changes 400
1 1
0.84 -
in the various interplanetary parameters.
1]
Before direct observations were begun, studies of geo- 0.76 - 11~111~~1 I
magnetic activity had shown that the variations in the in- [c)
terplanetary medium were such that minimum solar cycle 0.68
I 11

conditions at earth’s orbit occurred (on the average) about -,,1,,,,,,,,,,,,,,,

0.60
1 to 1.5 years after sunspot minimum. Also, geomagnetic
storms characterized by sudden beginnings were most com-
:: : ‘w
mon at sunspot maximum, but storms tending to recur on
successive solar rotations were more common about three
years after maximum. The sudden commencements are due
-.
m

-
I .75
/4
L la

to solar wind shocks which are most common at sunspot - “o~

:M
maximum. Recurrent storms are caused by long-lived rel-
atively steady high speed streams in the solar wind such as [e)
those observed in 1962 and 1973–74, that is, during the
phase of decreasing sunspot numbers.
The solar wind has now been observed at least inter-
1964 1968 1972 1976
mittently for almost 20 years and solar cycle variations of YEAR

several parameters have been established. Figure 3-20

[Feynman, 1983] shows the variations of some of these Long term variations of selected quantities, The top panel
quantities. Figure 3-20a gives the annual mean sunspot num- (a) shows the annual mean sunsfit number. The “second
panel (b) gives the annual mean velocity. The third panel
ber from 1962 to 1978. Figure 3-20b shows the annual mean (c) shows the yearly averaged values of logarithms of hourly
velocity of the solar wind. The first few values are joined interplanetary magnetic field magnitudes. The observed dis-
by a dashed line to indicate that the data sample was too tribution of hourly IMF is log normal. The fourth panel (d)
shows the yearly average of hourly values of the magnitude
small to give an accurate value. of the z component of the IMF. The final (e) panel gives
There is also a solar cycle variation in the magnitude of the percentage abundance of helium by number. The arrows
the interplanetary field that is most clearly seen in the yearly indicate that only limits on the value~ could be determined
[ Feynman, 1983].
mean value of the logarithm of IBI as shown in Figure
3-20c. The data exhibit relatively low mean values, cor-
responding to IBI about 5 nT in 1964, 1965, 1975, 1976 1966 were sparse but what information there is indicates
and 1977. The mean log IBI is remarkably constant through- that the mean velocity observed in 1962, during the declin-
out the cycle between 1966 and 1974 corresponding to IBI ing leg of solar cycle 19, was high with a velocity of about
about 5.8 nT. The large rise to [BI about 6.9 nT appearing 490 km/s whereas in 1964 and 1965 it was near 400 km/s.
at the beginning of the new sunspot cycle shows that the The distinction between the shape of the sunspot cycle vari-
solar wind magnetic field can differ from cycle to cycle. ation and the solar wind cycle variation is clearly shown by
The yearly averages of the hourly averaged magnitude the velocity. Here the highest yearly average velocities are
of BZ,(IBZI), are also compared to the sunspot number in observed during years dominated by recurring high speed
Figure 3-20d. Here B. is defined in GSE coordinates and solar wind streams. That this is nonetheless a solar cycle
the hourly average is taken by including the sign of B,, that variation is shown by the fact that the declining phase of
is, if B= were of constant magnitude and negative for half each sunspot number cycle exhibits such a series of recument
of the hour and positive for the rest, IBZI would be zero. geomagnetic storms.
There are two maxima in (IBZI), one before and one after The averaged helium abundance in Figure 3-20e shows
sunspot maximum. It is not known if this double maximum a solar cycle variation much more like that of the sunspot
is a common form of the (IBZI) solar cycle variation. number. The abundance near sunspot minimum in 1962-65
The variation in the yearly average solar wind speed (v) and 1976 is just above 3’%, whereas 5% was the abundance
from 1966 to 1978 is shown in Figure 3-20b. Data before typical of sunspot maximum conditions.

3-23
CHAPTER 3
3.4.2 Secular Variations typically 1 AU at earth’s orbit, the plasma acts as a fluid
on a much smaller spatial scale, because plasma instabilities
Little information exists on the variation of the solar and wave particle interactions produce a small equivalent
wind for time scales longer than a solar cycle. In Figure mean free path. For many applications it is good approxi-
3-20c it appears that the interplanetary magnetic field in- mation to treat the solar wind as a magnetohydrodynamic
tensity in 1979 near the maximum of the current solar cycle fluid interacting with the earth’s field. For a full discussion
was larger than at any time in the past, but Figure 3-20b of the interaction of the wind and the magnetosphere the
shows that the velocity was quite close to values typical of process of magnetic reconnection or merging is important;
the preceding maximum. It has also been inferred from however, its major effect is to feed energy into the mag-
geomagnetic records that the solar wind was in some sense netosphere rather than to affect the flow around the mag-
weaker during the early years of this century than it has netosphere and so it is beyond the scope of this chapter.
been recently [Feynman and Crooker, 1978], and from au- See Chapter 8. A further simplification is usually made in
roral records that it was also weak in the early years of the the calculations by assuming the Alfv&n Mach number of
19th century [Feynman and Silverman, 19801. The lack of the solar wind flow is large. This decouples the fluid equa-
aurora during the Maunder minimum in the 17th century tions from the magnetic equations so that the properties of
again indicates a weak wind [Suess, 1979]. A revaluation the flow can be determined by solving the gas dynamic
of the auroral record from 450 AD to 1450 AD convincingly equations, and subsequently the magnetic field can be found
shows a periodicity in occurrence frequency with an average from the flow.
period of about 87 years [Siscoe, 1980] which almost cer- The shape of the forward part of the magnetosphere is
tainly reflects a solar wind variation. calculated by neglecting the magnetic field press:re in the
solar wind outside the magnetosphere boundary and the
particle pressure inside the boundary. The remaining pres-
3.5 THE CLOSE NEIGHBORHOOD sures are then set equal on the two sides, that is, the dynamic
OF THE EARTH pressure of the solar wind flow outside is balanced by the
magnetic field pressure inside.
The flow around the magnetosphere can then be cal-
3.5.1 Magnetosheath culated. To simplify the calculations further the magneto-
sphere is often approximated as an axisymmehic body formed
As the magnetized solar wind flows past the earth, the by the rotation of the equatorial trace of the magnetosphere
plasma interacts with the earth’s magnetic field and confines boundary around the earth sun line. The resultant flow is
the field to a cavity, the magnetosphere. Since the proton quite sensitive to the ratio of specific heats y. Observations
flow of the solar wind is almost always highly supersonic favor a y of 5/3 [Auer, 1974] and this value was used in
and super-Alfv6nic, a shock, called the bow shock, forms the model calculations discussed here.
sunward of the magnetospheric cavity. The configuration is The model flow in the high Mach number approximation
shown schematically in Figure 5 of Chapter 8. The solar is illustrated by Figure 3-21 [Spreiter and Alksne, 1969].
wind flows across the bow shock in front of the earth where In Figure 3-21a, the lines of constant ratio of magnetosheath
it is slowed to subsonic velocities. It is deflected by the density to solar wind density are shown. The sheath density
obstacle presented by the earth’s field at the bounda~ des- is high in the forward part of the magnetosphere reaching
ignated as the magnetopause in Figure 8-5. The wind then a maximum in the subsolar region. As the plasma flows
flows generally antisunward down the flanks of the mag- back toward the magnetospheric flanks, it becomes less
netosphere carrying the interplanetary field with it. A typical dense and values less than one begin to appear near the
distance between the earth’s center and the subsolar point magnetopause at about 7 R~ behind the earth. Figure 3-21 b
on the magnetopause (the boundary between the solar wind shows both lines of equal velocity ratio and temperature
and the magnetosphere) is 10 earth radii (RE). The shock ratio. Since the fluid approximation is used, the temperatures
is about 3 R~ sunward of the magnetopause at the subsolar of the electrons and protons are assumed to be the same,
point. The region between the shock surface and the surface and for comparison with observation they are assumed equal
of the magnetosphere is called the magnetosheath. to the measured proton temperatures. The numbers labeling
The problem of calculating the flow of the solar wind the lines and written on the magnetopause side of the mag-
plasma past the earth using the full plasma equations is netosheath are the temperature ratios, and the numbers on
prohibitively difficult. Even in the magnetohydrodynamic the bow shock side are the velocity ratios. The stagnation
approximation, the problem has not been treated in general. point is labeled at zero velocity, which of course is un-
The high conductivity of the solar wind and the large kinetic realistic as far as expectations of observations are concerned.
energy density of the particles permits the approximation to However, slowing and heating of the plasma in the subsolar
be made that the magnetic field configuration can be treated stagnation region is observed. As the plasma moves back
as though the field is frozen into the flow. Although the along the flanks it begins to cool and increase in speed.
solar wind proton mean free path for Coulomb collision is Figure 3-2 lC shows the stream lines of the flow.

3-24
 SOLAR WIND
Y= 5/3

3
VELOCITY, V
TEMPERATuRE, T Vlvm
0.85

SHOCK

2-
0.75

0.7

0.6
I 14.66 MAGNETOSPHERE
0.5 BOUNDARY

0.4

0.3
0.2
0
I .0 0 -1.0 o— 1.0 0 -1.0
X/D
(Y- I)M: 2

()
_T 2
=1+ l-+ = 22.3- 21.3 ~
Tm 2
‘a () Vm

(cl) (b)

3-
- SHOCK

— STREAML INES
.
‘ --- VELOCITY .
CONTOURS

2
I
2- # /

*
a F
\ 08 0.85 V/V.
.
0.75 “
.7
L MAGNETOPAUSE
1-

:0 t

2+(Y-1) M~cos2a
‘o
—. -
PI (Y+I) M~cos2a
/ / 0
1,{1
o I I
-1 0 I
x/D
(c)

Figure 3-21. Flow in the magnetosheath inthegas dynamic approximation forthehigh Mach num&rcaw(M = 8). (a)gives the density and(b) the
temperature andveloeity contours. The subscript m denotes quantities in the undisturbed wind at “infinity”. Thenon-subseripted quantities
are the parameters within the magnetosheath. Lines ofconstant ratio of magnetosheath inundisturbed quantities are given. The shape of the
contours for velocity and tem~rature are the same but the ratios differ as marked. The jump relation for the temperature is given in (b). In
(c) the veloeity contours are given again, along with the stream lines. The formula at the bottom of (c) gives the jump relations across a
shock of arbitrary angle with the incident flow. Subscript 1 indicates values just behind the shock. The abscissas are normalized so the
distance from the center of the earth to the subsolar point on the magnetopause is taken as one [Spmiter and Alksne, 1%9; Spreiter et al.,
1%6].

3-25
CHAPTER 3

-3

3--
BO
s

2 -2

I -1
3.4 “’ /’
&=Pv 3.6 /f/ 1 ,/
0:, POv~

PI Vl” ‘ POVo”
o
I 0
-1 0 I -1 0 I
xID x/D

(0) (b)

3.4 3.2
/’/ 3.0
‘.5 /’,’,/
/ ,’ /
,, // ,/

=
‘1!
\

6
. .:’_-_._—_—
- ———_— —_—

,..,
&
3.82 ““d
4.:/5 -1 0
I

I
x/D

(c)

Figure 3-22.Magnetosheath magnetic fields for three orthogonal components of the interplanet,~ lield calculated by assuming the field is frozen into the
flow shown in Figure 3-21. The field directions are as shown in (a), (b), and (c). The jump relations at the shock are indicated and the
dashed lines indicate contours ofequal intensity ratio, whereas thesolid lines follow thedirections of the field lines [Alksne and Webster,
1970]. (Reprinted with permission from Pergamonpfess, Ltd. @ 1970.)
 SOLAR WIND
The magnetic field configuration is computed in this The observed flow and averages of the magnetic fields
approximation by assuming the field is carried by the flow. in the magnetosheath agree satisfactorily with the magne-
The results are shown in Figure 3-22 [Alksne and Webster, tohydrodynamic models. The magnetosheath magnetic field
1970]. The incident field has been broken up into three is typically very disturbed but the vectors tend to be aligned
orthogonal components with directions as shown on the with the magnetopause as shown in Figure 3-24a [Fairfield,
figure. An arbitrary incident magnetic field will be described 1967]. The average relative magnitudes of the magneto-
by an appropriate superposition of the three cases. The stream sheath field for 10 RE intervals along the earth-sun line are
lines are given by the solid lines and the lines of equal shown in Figure 3-24b [Behannon and Fairfield, 1969]. The
magnetic field ratio are given by the dashed lines. The ratio of the magnitude of the magnetosheath field to the
magnetic field lines become draped over the magnetosphere magnitude of the interplanetary field varies from 4 near the
and tend to be aligned with the magnetopause direction. most sunward portion of the bow shock to less than 1 far
Calculated values of magnetosheath electric fields are down the flanks.
shown in Figure 3-23 [Alksne and Webster, 1970] but these The variability of the magnetosheath fields is reflected
have not been measured. Since there is no electric field in the power spectra as shown in Figure 3-25 [Fairfield,
associated with field aligned flow, electric fields for only 1976]. The low frequency power levels of the variations of
two orthogonal magnetic field directions are shown, and the the field magnitudes differ by four orders of magnitude
electric field due to an arbitrary magnetic field can be found depending on the day. The September and November 1967
by combining the two cases shown. spectra are probably the most representative. Spectral peaks
Measurements of the shape of the magnetosphere show near 0.05 Hz and 0.07 Hz are typical but the exact frequency
some asymmetry that is not present in the results calculated and relative magnitude are variable. The 31 July spectrum
from the high Mach number approximation. This asymmetry was taken on a day when the solar wind Mach number was
probably arises because of the the asymmetric magnetic field low. At these times the power in magnetosheath variations
in the magnetosheath and the coupling between the fluid is also quite low. The dashed lines with slopes of l/f and
flow and the magnetic field. All of these effects will lead l/f3 are shown for comparison. The spectra typically steepen
to magnetosheath pres$~re asymmetries which will distort at frequencies above the proton gyro frequency. At frequen-
the magnetosphere to some extent. In addition, the orien- cies between 3.3 X Id and 2.6 X 103 Hz, which are
tation of the earth’s magnetic dipole relative to the direction lower than those shown in the figure, the average slope is
of incidence of the solar wind must be taken into account. slightly steeper than l/f. There is a tendency for the per-

Ez/Eoz
E/Eoy

w
3-

2- -
- 30
Xo
,0.0
o
\
a

2.25

1- 2.0
— FIELD LINES

‘ -- INTENSITY
1.5 CONTOURS

fi=lo sin a + cos a

1// E“
Ill Oz
0
-1 0 I 0.5’/-1 o I
xID u x/D
(0) (b]

Figure 3-23. Magnetosheath electric fields calculated under the same assumptions as in Figure 3-22. The relations at the shock jump are indicated in each
panel [Alksne and Webster, 1970]. (Reprinted with permission from Pergamon Press, Ltd. 0 1970. )

3-27
CHAPTER 3
0.
9 10
123~ 5678
!

— AVER&GE WELATIVE MbGNITuOE 1966- 1968 N : 8661

--- 8/00450 e”””” + 053

i
\\
m“
; Figure 3-24 Observations of the magnetosheath field: a) Averages of the
v
3 magnetosheath magnetic field intensity normalized to the
simultaneously measured interplanetary field. For each AX
bin measurements are taken from all available values of y
2 -
---
and z. The field is intense in the subsolar region and de-
I x
L creases to less than the interplanetary value in the distant
I
---- flanks [Behannon and Pdirtieid. 1969]. (Reprinted with per-
mission of Pergamon Press. Ltd. O 1969. ) b) Projections
1 of vectors measured in the magnetosheath into the X. y plane
-70 -80
Z. ,0 0 -IO -ZO -3o -40 “’o ’60 for four different t)rientations of the interplanetary field.
X( R<)
[Fairfield. 1967].

(a)

T T
-20 -15 -lo -5

‘SE

I
I -1
I
00 <$, < 1800 SOY
k I
-100 <6, < I 0°
‘SE-
20

T
T
-20 -15 -lo -5
-20 -15 -10 -5
1 , I I I ,
‘SE
‘SE k

(b)
 SOLAR WIND
(c pause or within the magnetosheath also contribute to ob-
served waves in the magnetosheath.
(c

,0
SEPT
550-
1, 1967
620
3.5.2 BOW Shock
(0

The flow of the solar wind past the earth is usually but
10 NOV 1, (961
not always highly supersonic and superalfvenic as far as the
300-330
protons are concerned. However, both theory and obser-
; ,0 vations imply that the treatment of the bow shock as if it
.
.
:
were a hydrodynamic discontinuity is a gross oversimpli-
m
.,0 fication. Since the solar wind at earth is collisionless on a
scale much larger than the magnetosphere, plasma prop-
10 erties are responsible for producing the bow shock discon-
\
\
\\
tinuity, which is a collisionless shock. The types of plasma
10
\
\\ phenomena involved will depend on the Mach number of
I 0.
000 3 k,
the flow, the ratio (~) of the internal energy of the particles
‘\\
to the energy in the niagnetic field, and the direction of the
,0- magnetic field relative to the shock surface. Figure 3-26
\ [Greenstadt et al., 1980] shows a conceptualization of col-
(
~ f(”zl lisionless shock phenomena at the earth’s bow shock. The
I 00 20 ,0 21 02 0, 0,02 0.01 T(s,. )
magnetic field is visualized as making an angle of about 45°
with the solar wind, which is flowing from the direction of
Figure 3-25. Composite of magnetosheatb magnetic field spectra. The
lower frequency data are magnetic field magnitude spectra
while the highest frequency data are from fluctuations in
the field components [Fairfield, 1976].

turbation vector to align itself with the shock surface for

variations observed near the shock surface. In the inner
portion of the magnetosheath, large amplitude waves oc-
curring primarily in the field magnitude are observed. Ro-
tational wave modes have been found to dominate over
magnetoacoustic modes for wave periods greater than 10
minutes. There also appear to be plasma clouds with scales
of several hundred to several thousand kilometers associated
with weak fields that deviate from a direction tangent to the
magnetopause. In addition to broad noise with a l/f3 spectral
shape there are often sporadic occurrences of quasi-mono-
chromatic bursts throughout the magnetosheath. ~ese waves
have a frequency of 50-200 Hz, durations of less than one
to tens of seconds and amplitudes of tenths of nanoteslas.
Several sources have been suggested for magnetosheath
field fluctuations. The waves and discontinuities in the solar
wind are sources of some of the variabilityy, as is the con-
vection into the magnetosheath of waves produced at the
bow shock or in the region of space just upstream from the
bow shock. Waves incident on the bow shock will be trans-
mitted across it and amplified and new waves of various
types will be produced in the process. The details concerning
the types of waves produced depend on the properties of
the incident wave. The power in the variations and com-
plexity of observed forms will be markedly increased by
transmission. Interplanetary discontinuities will contribute Figure 3-26 A schematic of the earth’s bow shock showing the quasi
to magnetosheatb variabili~ but they are not f~uent enough pe~ndicular and qoasiparallel regions. Perpendicular shocks
most closely obey the classical magnetohydrodynamic equa-
in the solar wind to explain a significant portion of mag- tions but parallel shock region behavior is dominated by
netosheath power. Processes taking place at the magneto- plasma effects [Greenstadt et al., 1980; Gary, 1981].

3-29
CHAPTER 3
the sun, that is, the sun is beyond the lower right hand netosheath so when parameters are averaged over periods
quadrant. Four regions of the shock surface are distin- of more than 5 min the shock should arrive at the shape
guished, the two extremes (strictly perpendicular and strictly given by steady state magnetohydrodynamic theory. The
parallel shocks) and the two intermediate regions (quasi- instantaneous shock is not in equilibrium. Figures 3-26 and
perpendicular and quasi-parallel shocks). A shock is called 3-27 show schematically the waves in the vicinity of the
parallel when the upstream magnetic field is perpendicular boundary that propagate back into the solar wind and into
to the plane of the shock and parallel to the upstream flow the magnetosheath. The properties of both the quasi-per-
velocity. Conversely, perpendicular shocks have normals pendicular shock and the quasi-parallel shock depend on the
perpendicular to the direction of the magnetic field in the range of ~ and M, where M is the Mach number of the
solar wind. The major part of the bow shock region has flowing gas. Since the fast wave mode is the MHD wave
been divided into quasi-parallel and quasi-perpendicular re- that steepens into a shock, the Mach number for a perpen-
gions based on empirical results. Quasi-perpendicular shmks dicular shock is given by M = V/(C~ + C:)”2 where CA
have a monotonic, sawtooth, or wave-step magnetic profile and Cs are the Alfv&n and sound wave velocities. In Table
and are sharply defined in space. Quasi-parallel shocks are 3-2 [Greenstadt and Fredericks, 1979] the characteristics of
thicker and do not show a clearly definable boundary be- shock phenomena are given for various combinations of ~
tween upstream and downstream parameters, either in the and M. A value of ~ = 1 is the dividing line for the regimes
field or particle properties. of ~ parameter space and M = 3 is the dividing level for
Figure 3-27 [Greenstadt and Fredericks, 1979] is a sche- the Mach number. An estimate of the incidence of the spe-
matic of a cut in the plane containing the interplanetary cific ranges of ~ and M the solar wind can be found from
field, showing the waves that appear in the parallel and Table 3-1. In the solar wind ~ varies from 0.5 to 4.6, and
quasi-parallel region in the vicinity of the classical position the ranges of sonic and Alfv&nic Mach numbers are from
of the bow shock. The shocked magnetosheath plasma takes 4 or 5 to over 10. However, solar winds with Alfv6n num-
about 5 min to flow around the forward part of the mag- bers less than 1 have been observed on occasion.

RETURN
PROTONS

Figure 3-27. Conceptualization of collisimrless shock phenomenology as manifested in the earth’s curved bow shock. Unshocked interplanetary field
direction B,W is indicated on the foreground field “platform. ” Field magnitude is plotted vertically. The superimposed three-dimensional
sketches represent solar wind proton thermal properties as number distributions in velocity space [Greenstadt and Fredericks. 1979].

3-30
 SOLAR WIND
Table 3-2. Macroscopic features of the bow shock bv., uarameter class

Parameter Plasma Name of Quasi-perpendicular Quasi-parallel

Values Conditions Structure Features Features .

p< Cold plasma, Laminar Clean field jump, sometimes with damped Multigradient field transition with
M<3 low Mach periodic waves, no turbulence. embedded, nearly-periodic
number wavetrains.
Relatively little proton temperature jump Upstream waves with strong
(Tp2/7’pl ) = 2 , maxwellian downstream periodic component, tens of
distribution. seconds period, unknown plasma
distributions.
@<l, Cool plasma, Quasi- Clean field jump, downstream, nearly- Multigradient field transition at
M>3 high Mach laminar periodic waves, little turbulence. least 2 RE thick, with large
number Appreciable proton temperature jump, amplitude pulses; extensive
bimodal distribution, nonmaxwellian high foreshock. Little change in solar
energy tail downstream. wind streaming velocity; some
heating with resulting non-
maxwellian proton distributions
differing from either solar wind or
magnetosheath forms.
p=l, Warm plasma, Quasi- Clean field jump, small scale turbulence. No known example.
~53 low Mach turbulent Little proton temperature rise, maxwellian
number downstream distribution.
p=l Warm plasma, Turbulent Irregular field fluctuations obscuring Multigradient, irregular field
~~3 high Mach definite average field jump. transition with large magnitude
number Bimodal or multimodal proton distributions, excursions, upstream waves.
nonmaxwellian downstream. Unknown plasma distributions, but
maxwellian deep in magnetosheath.
p>l Hot, high High-beta Irregular field fluctuations of extremely
M>3 velocity high peak magnitude, possibly lowering ~
plasma locally.
Extensive precursor region with appreciable
effect on approaching flow; proton
distributions unknown.
.fl. . .. . . ,.,. --1,
rom ureenstaaf ana Frearlclcs,[19 IYj )

3.5.3 Upstream Waves and Particles The upstream energetic particle population has been di-
vided into several types based on their observed velocity
The solar wind-magnetosphere interaction generates and phase distributions. Reflected ions are characterized by
substantial numbers of energetic particles with a high enough a beamlike distribution of solar wind ions that apparently
speed so that they can propagate back upstream guided by have been reflected and energized at the bow shock. They
the ambient interplanetary field. The upstream particle re- have a streaming velocity 2 or 3 times the solar wind velocity
gion has a strong local time asymmetry with the major effect and a density of about O. I/cmq. Their kinetic energy dis-
occurring usually on the dawnside on field lines connected tribution typically maximizes at about 5 keV and has spread
to the quasi-parallel shock. This region of space upstream of 1–30 keV. They are detected on interplanetary field lines
of the bow shock is sometimes called the foreshock and is that connect to the bow shock and are nearly tangent to the
characterized by many different types of particle distribu- shock surface. The sunward boundary of the region in which
tions and wave modes. The energy flux of particles into the these particles are observed is defined by the field-aligned
upstream region is found to be 10’7–10’8 ergs, which is flow back into the solar wind plus E X B convection, where
comparable to the peak energy released in a geomagnetic E is the solar wind electric field.
substorm. The presence of energetic heavy ions indicates In addition to these beam-like ions there is a popula-
that strong local acceleration is occurring. Observations of tion of diffuse ions that has a nearly isotropic velocity dis-
plasma waves, particles, and fields in the region has led to tribution with a typical temperature of 4 x 107 K, an up-
a partial understanding of some of the interrelationships stream flow speed of O.8–1.2 times the solar wind velocity
among the phenomena [Tsurutani and Rodriguez, 1981]. and a density that varies between 0.02 and O.8/cm3. The

3-31
CHAPTER 3
energy density of the diffuse ions is comparable to that be Doppler-shifted right-hand emissions being convected
of the reflected ions. These particles are found downstream back over the spacecraft. They propagate at small (- 25°)
(antisunward) of the reflected ions, almost filling the entire angles to the magnetic field. The rest-frame frequencies have
foreshock region. An excellent correlation exists between ben deduced to be 0.140 times the proton gyrofrequency.
diffuse ions and low frequency (0.01–1 Hz) MHD and They are related to the MHD waves and are detected during
ion-acoustic-like plasma waves. Recent observations indi- diffuse ion events. Whistler mode waves with frequencies
cate that the transition from distributions typical of reflected 10–200 Hz are observed upstream of the bow shock. These
ions to that of diffuse ions is a smooth one and there are emissions have a peak amplitude of O. 1 nT and occur at
ions with intermediate properties in a region of space be- about 1/4 to 1/2 the electron gyrofrequency. It is believed
tween the reflected and diffuse ion region. that the most intense waves are associated with low-energy
Upstream electrons have energies of 0.5 to 100 keV and electron spikes coincident with intense diffuse ion fluxes.
stream in a field aligned direction. The majority of events Other waves that have been observed include small am-
involve 1–2 keV particles with a peak flux of 104cm-2s-lsr plitude electrostatic waves in the 100 Hz–10 kHz range
keV-l. They appear on interplanetary field lines newly con- polarized primarily parallel to the ambient magnetic field.
nected to the bow shock and because of their high velocities They have short wave lengths, 30< k <215 m, and a low
can be found upstream of the ion foreshock boundary. relative energy density. Broadband impulsive electrostatic
The region upstream from the earth’s bow shock is also emissions with amplitudes proportional to f-z for frequencies
characterized by complex wave phenomena. Them m MHD less than 3 kHz are observed to almost always accompany
waves with an amplitude of 5 nT peak to valley and period electron plasma oscillations. The electron plasma oscilla-
of 10–60s, which have a left-hand polarization in the space- tions themselves have intensities up to 10 mV with wave-
craft frame. The particles involved in the wave show a f lengths of 400-780 m and frequencies of 20–30 kHz. They
15 km/s solar wind velocity variation as well as correlated are detected throughout the foreshock region and are driven
density fluctuations indicating a magnetosonic component. by 0.2 to 1.5 keV electrons at the boundary of the electron
There is a strong association between these waves and dif- foreshock. Low intensity electromagnetic emissions at twice
fuse ions. Wave packets of whistler mode waves having the electron plasma frequency having a namow band width
intensities < 1 nT and typically composed of 2 or 3 coherent are detected upstream of the bow shock to distances of at
cycles are observed. Long trains of more than 20 cycles least 240 RE. The emission frequency is correlated with
have been seen on occasion. The emissions, though left solar wind plasma density variations.
hand polarized in the spacecraft frame have been shown to

3-32
 SOLAR WIND
REFERENCES

Alksne, A.Y. and D.L. Webster, “Magnetic and Electric the Solar Wind and the Long Term Variation of Sunspot
Fields in the Magnetosheath,” Planet. Space Sci., 18: Activity,” J. Geophys. Res., 85:2991, 1980.
1203, 1970. Gary, S .P., “Microinstabilities Upstream of the Earth’s Bow
Auer, R. D., “Magnetohydrodynamic Aspects of the Earth’s Shock: A Brief Review,” J. Geophys. Res., 86:4331,
Bow Shock 1. Equilibrium Bow Shock Position,” J. 1981.
Geophys. Res., 79:5118, 1974. Gosling, J. T., V.Z. Pizzo, and S.J. Bame, “Anomalously
Bame, S. J., J.R. Asbridge, W.C. Feldman, E.E. Feni- Low Proton Temperatures in the Solar Wind Following
more and J.T. Gosling, “Solar Wind Heavy Ions from Interplanetary Shock Waves—Evidence for Magnetic
Flare-Heated Coronal Plasma,” Sol. Phys. 62: 179, Bottles?’, J. Geophys. Res., 78:2001, 1973.
1979. Gosling, J. T., J.R. Asbridge, S.J. Bame, and W.C. Feld-
Barnes, A., “Hydromagnetic Waves and Turbulence in the man, “Solar Wind Stream Interfaces,” J. Geophys. Res.,
Solar Wind,” in Solar System Plasma Physics, edited 83:1401, 1978.
by E.N. Parker, C.F. Kennel and L.J. Lanzerotti, North Gosling, J. T., J.R. Asbridge, S.J. Bame, W.C. Feldman
Holland, Amsterdam, 1979. and R. D. Zwicki, “Observations of Large Fluxes of He +
Behannon, K. W. and D.H. Fairfield, “Spatial Variation of in the Solar Wind Following an Interplanetary Shock,”
the Magnetosheath Magnetic Field,” Planet. Space Sci., J. Geophys. Res., 85:3431, 1980.
17:1803, 1969. Greenstadt, E. W. and R. W. Fredericks, “Shock Systems
Belcher, J. W. and L. Davis Jr., “Large Amplitude Alfv6n in Collisionless Space Plasma,” in Solar System Plasma
Waves in the Interplanetary Medium,” Y. Geophys. Res., Physics III edited by C.K. Kennel, L.J. Lanzerotti and
76:353, 1971. E.N. Parker North Holland, Amsterdam, 1979.
Burlaga, L. F., “Micro-Scale Structures in the Interplanetary Greenstadt, E. W., C.T. Russell and M. Hoppe, “Magnetic
Medium,” Sol. Phys., 4:67, 1968. Field Orientation and Suprathermal Ion Streams in the
Burlaga, L.F. and J.H. King, “Intense Interplanetary Mag- Earth’s Foreshock,” J. Geophys. Res., 85:3473, 1980.
netic Fields Observed by Geocentric Spacecraft During Hedgecock, P. C., “Measurements of the Interplanetary
1963–1975,” J. Geophys. Res., 84:6633, 1979. Magnetic Field in Relation to the Modulation of Cosmic
Burlaga, L. F., E. Sittler, F. Mariani, and R. Schwenn, Rays,” Sol. Phys., 42:497, 1975.
“Magnetic Loop Behind an Interplanetary Shock: Voy- Hirshberg, J., “The Transport of Flare Plasma from the
ager, Helios and IMP8 Observations,” J. Geophys. Res., Sun to the Earth,” Planet. Space Sci., 16(3): 309–3 19,
86:6673, 1981. 1968.
Burlaga, L.F. and J.B. Turner, “Microscale Alfv&n Waves Hirshberg, J., “The Interplanetary Field During the Rising
in the Solar Wind at 1 AU,” J. Geophys. Res., 81:73, Part of the Solar Cycle,” J. Geophys. Res. 74:5841,
1976. 1969.
Chapman, S., “Notes on the Solar Corona and Terrestrial Hirshberg, J., S .J. Bame, and D.E. Robbins, “Solar Flares
Ionosphere,” Smithsonian Contrib. Astrophys., 2:1, 1957. and Solar Wind Helium Enrichments, July 1965–July
Coleman, P.J. Jr., “Turbulence, Viscosity and Dissipation 1967,” Sol. Phys., 23:467, 1972.
in the Solar Wind Plasma,” Astrophys. J., 153:371, Hirshberg, J., Y Nakagawa, and R.E. Wellck, “Propagation
1968. of Sudden Disturbances Through a Non-Homogeneous
Fairtield, D. G., “The Ordered Field of the Magnetosheath,” Solar Wind,” J. Geophys. Res., 79:3726, 1974.
J. Geophys. Res., 72:5865, 1967. Hundhausen, A. J., Coronal Expansion and Solar Wind,
Fairfield, D. H., “Magnetic Fields of the Magnetosheath,” Springer-Verlag, New York, 1972.
Rev. Geophys. Space Phys., 14:117, 1976. Neugebauer, M., “The Enhancement of Solar Wind Fluc-
Feldman, W. C., J.R. Asbridge, S.J. Bame, and M.D. tuations at the Proton Gyroradius,” J. Geophys. Res.,
Montgomery, “Interplanetary Solar Wind Streams,” Rev. 80:998, 1975.
Geophys. Space Phys., 12:715, 1974. Neugebauer, M., “Observations of Solar-Wind Helium,”
Feldman, W. C., J.R. Asbridge, S.J. Bame, M.D. Mont- Fundamentals Cosmic Phys., 7: 131, 1981.
gomery and S .P. Gary, “Solar Wind Electrons,” J. Geo- Neugebauer, M. and C. W. Snyder, “Mariner 2 Observa-
phys. Res., 80:4181, 1975. tions of the Solar Wind,” J. Geophys. Res., 71:4469,
Feldman, W. C., J.R. Asbridge, S.J. Bame, and J.T. Gos- 1966.
ling, “Plasma and Magnetic Fields from the Sun,” in Neupert, W .M. and V. Pizzo, “Solar Coronal Holes as
The Solar Ouput and its Variations, edited by Oran R. Sources of Recurrent Geomagnetic Disturbances,” J.
White, Colorado Associated University Press, Boulder, Geophys. Res., 79:3701, 1974.
p. 351, 1977. Ogilvie, K. W., M.A. CopIan, P. Bochsler, and J. Geiss
Feynman, J., “Solar Cycle and Long Term Changes in the “Abundance Ratios of “He+ ‘/sHe + + in the SOIM wind,”
Solar Wind,” Rev. Geophys. and Space Phys., 21:33, J. Geophys. Res., 85:6021, 1980.
1983. Parker, E. N., Interplanetary Dynamical Process, Intersci-
Feynman, J. and N.U. Crooker, “The Solar Wind at the ence, New York, 1963.
Turn of the Century,” Nature, 275:626, 1978. Rickett, B. J., “Disturbances in the Solar Wind from IPS
Feynman, J. and S.M. Silverman, “Auroral Changes During Measurements in August 1972,” Solar Phys., 43:237,
the 18th and 19th Centuries and Their Implications for 1975.

3-33
CHAPTER 3
Robbins, D. E., A.J. Hundhausen, and S.J. Bame, “Helium the Interplanetary Medium,” J. Geophys. Res., 82:10,
in the Solar Wind,” J. Geophys. Res., 75: 1178, 1970. 1977.
Russell, C. T., “Geophysical Coordinate Transformations,” Spreiter, J. R., A.L. Summers, “and A. Y. Alksne, “Hydro-
Cosmic Electrodynamics, 2:184, 1971. magnetic Flow Around the Magnetosphe~,” Pbnet. Space
Schwenn, R., H. Rosenbauer, and K. H. Muhlhauser, “Sin- Sci., 14:223, 1966.
gly Ionized Helium in the Driven Gas of an Interpla- Spreiter, J .R. and A. Y. Alksne, “Plasma Flow Around the
netary Shock Wave,” Geophys. Res. Lett., 7:201, 1980. Magnetosphere,” Rev. Geophys. Space Phys., 7: 11,
Scudder, J. D. and S. Olbert, “A Theory of Local and Global 1969.
Processes Which Affect Solar Wind Electrons 1., The Suess, S. T., “The Solar Wind During the Maunder Mini-
Ongin of Typical 1 AU Velocity Distribution Func- mum,” Planet. Space Sci., 27: 1001, 1979.
tions, Steady State Theory” J. Geophys. Res. 84:275, Svalgaard, L., “Polar Cap Magnetic Variations and Their
1979. Relationship With the Interplanetary Magnetic Sector
Siscoe, G. L., “Evidence in the Auroral Record for Secular Structure. ,“ J. Geophys. Res., 78:2064, 1973.
Solar Variability,” Rev. GeopAys. Space Phys., 18:647, Tsurutani, B .T. and E.J. Smith, “Interplanetary Discontin-
1980. uities: Temporal Variations and the Radial Wlent from
Siscoe, G. L., L. Davis Jr., P.J. Coleman Jr., E.J. Smith, 1 to 8.5 AU,” J. Geopys. Res., W. 2773, 1979.
and D.E. Jones, “Power Spectra and Discontinuities of Tsurutani, B .T. and P. Rodriguez, “Upstream waves and
the Interplanetary Magnetic Field: Mariner 4 ,“ 3. Geo- particles: An Overview of ISEE Results,” J. Geophys.
phys. Res., 73, 1968. Res., 86:4319, 1981.
Siscoe, G. L., N. U. Crooker, and L. Christopher, “Solar Wilcox, J.M. and N.F. Ness, “Quasi-Stationary Corotating
Cycle Variation of the Interplanetary Magnetic Field,” Structure in the Interplanetary Medium,” J. Geophys.
Sol. Phys. 56:449, 1978. Res., 70:5793-5805, 1965.
Smith, E. J., “Interplanetary Magnetic Fields,” Rev. Geo- Wilcox, J. M., L. Svalgaard, and P.C. Hedgecock, “Com-
phys. Space Phys., 17:610, 1979. parison of Inferred and Observed Interplanetary Mag-
Solodyna, C. V., J. W. Sari, and J. W. Belcher, “Plasma and netic Field, 1970- 1972,” J. Geophys. Res., 80:3685,
Field Characteristics of Directional Discontinuities in 1975.

3-34
 Chapter 4

THE GEOMAGNETIC FIELD

D.J. Knecht and B.M. Shuman

The existence of the geomagnetic field has been rec- However, the most important connection, the fact that
ognized for a very long time. The usefulness of a magnet the geomagnetic field interacts with a continuous stream of
as a directional reference was probably known in China solar plasma, was established only within the last twenty
more than 1000 years ago and in Europe at least 800 years years. As a result of satellite investigations, these recent
ago. As early as the 15th century, the earlier belief that a years have seen drastic revisions in many fundamental ideas
compass needle points true north was found to be incorrect concerning the configuration of current systems and the
and mariners and mapmakers took account of this. Recorded magnetic field above the surface of the earth.
measurements of the magnetic declination (the deviation of But while satellite measurem~nts have expanded our
the compass from true north) at various locations on the understanding of the space above us, the development of
earth date back to the early 16th century, which also saw techniques for collecting and interpreting archaeological and
the discovery of the magnetic dip (the deviation of a compass geological data have led to some important discoveries about
needle from horizontal when unconstrained). Although ex- the earth below us: namely, that the continents have drifted
periments with magnets had been carried out since the 13th thousands of miles from their original locations and that the
century, the concept that the earth itself is a magnet was entire geomagnetic field undergoes periodic reversals.
not advanced until the end of the 16th century by Gilbert. Today, geomagnetism encompasses two broad areas of
From these beginnings, geomagnetism as a branch of theoretical study that are served by overlapping experimen-
science was developed. It was first assumed that the mag- tal data bases: the physics of the interior of the earth (which
netism of the earth was like that of a solid permanent magnet produces the steady and slowly varying field) and the physics
and was therefore expected to be constant in the absence of of the magnetosphere and ionosphere (which produce the
major geological changes, but this view was soon proved dynamic behavior of the field). The material in this chapter
wrong; the secular variation (changes in the field over time tends to neglect the physics in favor of a description of
intervals of years or centuries) was discovered in the 17th geomagnetic phenomena and their experimental observa-
century. Transient variations of the field (geomagnetic dis- tion, but this neglect is partly remedied in other chapters
turbance) were observed during the 18th century, and geo- on the ionosphere and magnetosphere.
magnetism was increasingly appreciated to be a dynamic
phenomenon.
By the early 19th century, a large number of magnetic 4.1 BASIC CONCEPTS
observatories had been established both in European coun-
tries and in the distant lands of their empires. Through
coordinated measurements by many stations, the geographic 4.1.1 Units, Terminology, and Conventions
dependence of some geomagnetic phenomena was discov-
ered and the worldwide nature of major disturbances was The geomagnetic field is characterized at any point by
established. The increasing volume and precision of accu- its direction and magnitude, which can be specified by two
mulated data made discouragingly clear how complex were direction angles and the magnitude, the magnitude of three
the phenomena being studied. Increasing international co- perpendicular components, or some other set of three in-
operation included investigations during the first Intern- dependent parameters. Angles are commonly measured in
ational Polar Year ( 1882–1 883). By this time, the correlation degrees, minutes, and seconds. Prior to widespread adoption
between the 11-year periodicities of sunspot occurrence and of mks units, the magnitude was usually given in units of
geomagnetic phenomena had been noted. Early in the 20th oersted (magnetic intensity) or gauss (magnetic induction).
century the intimate connection between solar and geomag- Since the field is less than one oersted everywhere, the unit
netic phenomena was further established by the correlation gamma was most useful; one gamma equals l~s oersted or
of recurrent disturbances with the 27-day solar rotation and 10-5 gauss and was used interchangeably for intensity and
later by the correlation of magnetic storms with solar flares. induction. Since the mks unit for field strength is very large

4-1
CHAPTER 4
(one tesla = 104 gauss), the nanotesla (nT), which very 4.1.2 Coordinate Systems
conveniently equals one gamma, is now almost universally
used. A number of coordinate systems are employed in the
Some of the angles and components commordy em- description of geomagnetic phenomena. Five of the most
ployed are shown in Figure 4-1. Standard terminology is as useful are the geographic, geomagnetic, geocentric solar-
follows. The vector geomagnetic field is the vector F. Its ecliptic, geocentric solar-magnetospheric, and solar-mag-
magnitude F is called the total intensity or the total field. netic systems. They are shown in Figure 4-2 and are defined
as follows.
A geographic coordinate system is one that is fixed with
respect to the rotating earth and aligned with the axis of
rotation. Most commonly used are the spherical polar co-
ordinates r, 6, and +, where r is geocentric distance, 6 is
colatitude (measured from the north geographic pole), and

Geographic Geomagnetic
= (Zr;;:
J ‘- i
I ,:,, ‘“

k
,/’
./ I
-05
,“ magn.tic
I I la
colatitude

--
1! I
,,~1

I
I
I
ii I

1// I

1//) i 1
1 ,/ “
,’
I /
I /’ %.-—---—-—-—
+*
<eti. I.ngit<

Solar- ecliptic Solar - magnetospheric

F
jzsm
F: total field
H: horizontal component axis
-
X: northward component
Y: eastward component
Z: vertical component
D: declination ,1
,1
I: inclination (1
II

Figure 4-1. Definition and sign convention for the magnetic elements.

The magnitude H of the horizontal component vector H is

called the horizontal intensity; the magnitude Z of the ver-
tical component vector Z is called the vertical intensity. The Solar-magnetic
northward, eastward, and downward components of the field
are designated by the magnitudes X, Y, and Z, respectively, direction of
geographic
the Cartesian components of the field. The magnitude D of north PO(E

the angle between X and H is called the declination, the

magnetic variation, or the variation of the compass. The
magnitude I of the angle between H or F is called the
inclination or the dip. The quantities F, H, X, Y, Z, D and
I are called magnetic elements. The sets of elements used D .
r Yam
,
most commonly to specify the field are (H, D, Z); (F, I,
D); and (X, Y, Z). The sign convention for each parameter
is shown in the figure, all vectors and angles being positive
as drawn. Figure 4-2. Several coordinate systems used in geomagnetism.

4-2
 THE GEOMAGNETIC FIELD
@ is east longitude (measured from the Greenwich merid- field line (that is, one computed from an accurate higher-
ian), with the earth assumed spherical. Sometimes altitude order model fitted to the actual geomagnetic field, as de-
(above the surface of the earth) is specified in place of r, scribed in Section 4.6. I ) is traced out to intersect the mag-
north or south latitude is specified in place of O, and west netic equatorial plane at the point A, which has polar co-
longitude is specified in place of + for values greater than ordinates LC and AC. From here a simple dipole field line
180 degrees. (Geodetic coordinates, which are defined rel- is projected back toward the earth to a “landing point” QC.
ative to the nonspherical earth ellipsoid, must be used with The corrected coordinates for point Q are the uncorrected
care. ) coordinates of point Q= (OC and AC). The value of @Cmay
The geomagnetic coordinate system is also a spherical be found from the dipole equation, L. = R sec2 @C, where
polar system fixed relative to the earth, but the polar axis R is the radius of the earth. Assigning the corrected coor-
is the axis inclined 1 I.5 degrees to the axis of rotation, dinates to the point Q permits an accurate description of
intersecting the earth surface at the point 78. 5“N, 29 1.O”E phenomena in the actual field as if it were a simple dipole
which defines the geomagnetic north pole. This was at one field; for example, points having the same coordinates in
time the axis of the best centered-dipole approximation to the northern and southern hemispheres will be actual con-
the field; current spherical-harmonic models of the main jugate points. A revised tabulation of corrected geomagnetic
field would place the pole approximately 0.25 degrees far- coordinates, using terms through order 7 in the spherical-
ther north and 1.6 degrees farther west. Geomagnetic co- harmonic model, has been published by Gustafsson [ 1970]
ordinates r, 0~, and +~ (and geomagnetic latitude and lon- for every 5° of geographic longitude and every 2° of geo-
gitude) are defined by analogy with geographic coordinates, graphic latitude.
with +~ (or geomagnetic longitude) measured from the The geocentric solar-ecliptic (GSE) coordinate system
American half of the great circle which passes through both is a right-handed Cartesian system with coordinates X~C,
the geographic and geomagnetic poles (that is, the zero- Y,=, and Z,e, and the center of the earth as origin. The
degree geomagnetic meridian coincides with 29 1.O”E geo- positive X,e axis is directed toward the sun. The Z,C axis is
graphic longitude over most of its length). directed toward the northern ecliptic pole, so both the X~C
The corrected geomagnetic coordinate system is a re- and Y.. axes lie in the ecliptic plane. This system therefore
finement (of the geomagnetic coordinate system) that has rotates slowly in space with the orbital period of the earth.
proved useful in considering phenomena that involve prop- In this system, field vectors are often resolved into two
agation along lines of force of the geomagnetic field [Hak- components, one lying in and the other perpendicular to the
ura, 1965]. It effectively provides a more accurate field-line ecliptic plane; the direction of the former is specified by the
connection from a point on (or near) the earth surface either angle + between it and the X., axis (positive counterclock-
to the equatorial plane or to its conjugate point than would wise when viewed from the northern pole). The direction
be afforded by any simple dipole approximation of the geo- of the total field is specified by @ and O, where O is the
magnetic field. Figure 4-3 shows how the corrected geo- angle between the vector and the ecliptic plane (positive
magnetic coordinates (latitude 0, and longitude A=) are northward). This system is particularly useful for referenc-
obtained for a point Q, at the earth surface, whose geo- ing data from interplanetary space, such as measurements
magnetic coordinates are @ and A. Starting at Q, an “actual” of the undisturbed solar wind and the interplanetary mag-
netic field.
The geocentric solar magnetosphetic (GSM) coordinate
system is also a right-handed Cartesian system, with co-
ordinates X,~, Y,~, and Z,~, and its origin at the center of
the earth. The positive X~~ axis is also directed toward the
sun. It differs from the solar-ecliptic system in that the Z\~
axis lies in the plane containing both the X,~ axis and the
geomagnetic dipole axis defined above. The system there-
fore not only rotates with the orbital period of the earth but
also rocks back and forth through 23 degrees (a rotation
about the X~~ axis) with a period of one day. This system
is particularly useful for referencing data from distant re-
gions of the magnetosphere, since time-dependent features
that result from the conical motion of the dipole axis are,
to a large extent, eliminated; that is, to a first approximation,
the entire magnetosphere, in its main features, may be ex-
pected to rock back and forth in this way.
A related frame is the solar-magnetic (SM) coordinate
Figure4-3. Method of finding corrected geomagnetic coordinates system. In this system the Z axis is directed to the north

4-3
CHAPTER 4
widely used is the B-L coordinate system of McIlwain. As
along the geomagnetic dipole axis, and the Y axis is per-
shown in Figure 4-4, surfaces of constant B (magnetic field
pendicular to the earth-sun line toward the d~sk side. The
intensity) are concentric, roughly ellipsoidal shells encir-
X axis is not always directed toward the sun in this system,
cling the earth, while surfaces of constant L (a magnetic
but rocks back and forth through 11.5 degrees about the
parameter) approximate the concentric shells generated by
earth-sun line. This system differs from the GSM essentially
by a rotation about the Y,~ axis. dipole field lines rotating with the earth. The mathematical
In addition to strictly spatial coordinate systems, several definition of L arises from the equations of motion of par-
ticles in the field; to some degree of approximation, particles
so-called magnetic systems have been found useful in study-
move to conserve three adiabatic invariants to which B and
ing the motion of charged particles trapped in the magnetic
L are simply related. Since these coordinates are more useful
field; these coordinates generally locate particles by refer-
to the study of trapped particles than to the study of the
ence to surfaces on which some magnetic parameter is con-
field itself, the reader is referred to standard texts on trapped-
stant, and since most particles are strongly controlled mag-
netically a great simplification of the data often results. Most particle physics for a complete discussion.

\ \ \ \ \
-90 -60 -30
NORTH LATITUDE (DEGREES)

Figure 4-4, The B-L coordinate system. The curves shown here are the intersection of a magnetic meridian plane with surfaces of constant B and constant
L. (The difference between the actual field and a dipole field cannot be seen in a figure of this scale. )

4-4
 THE GEOMAGNETIC FIELD
4.1.3 Sources of the Geomagnetic Field The steady (nonvarying or dc) component of the field
may be considered first. Although it is true that the entire
In considering a physical description of the field, a useful field has been varying drastically over geological time scales,
point of view to adopt is that of energy balance. A static that portion which varies with periodicities grqater than about
field represents an energy density B2/8v, and any change a year is customarily considered to be the steady field, while
in the field implies a transfer of energy to or from the field. the remainder is considered the variation field.
Understanding the field therefore implies identifying the Most of the steady field arises from internal terrestrial
energy sources and the causative physical mechanisms through sources (that is, below the surface of the earth, but excluding
which this energy generates (or is generated by) the field. currents induced in the earth by external current systems)
Except in the case of permanent residual magnetism, a mag- and is known as the main field. This field results primarily
netic field is generated only by the macroscopic motion of from convective motion of the core and is approximately
electric charge, so the final step in any physical process of dipole configuration, having a strength at the surface of
affecting the field will involve electric currents, though the the earth of several times 104 nT. The dipole is centered
energy driving the currents may be drawn from various close to the center of the earth, with its axis inclined about
sources. At present, the terrestrial and extraterrestrial sources 11.5 degrees to the axis of rotation. About 10% of the main
known to contribute appreciably to the geomagnetic field field, often termed the residual field, is nondipolar; it con-
are the following: sists of both large-scale anomalies (up to thousands of kil-
1. Core motion. Convection motion of the con- ometers), believed to be generated by eddy currents in the
ducting fluid core of the earth constitutes a self- fluid core, and small-scale irregularities (down to 10 km)
exciting dynamo. arising from residual magnetism in the crust. Changes in
2. Cmstal magnetization. Residual permanent mag- the main field (the so-called secular variation) are slow,
netism exists in the crust of the earth. with time constants of tens to thousands of years.
3. Solar electromagnetic radiation. Atmospheric If the earth were in a perfect vacuum, its dipole field
winds (produced by solar heating) move charged would extend outward without limit, merging smoothly with
particles (produced by solar ionizing radiation); the fields of the sun and other planets in a simple additive
this constitutes an ionospheric current which gen- fashion, the field strength declining inversely with the third
erates a field. power of geocentric distance. However, interplanetary space
4. Gravitation. The gravitational field of the sun and is not a vacuum but is filled with the ionized corona of the
moon produces a tidal motion of air masses that sun (the solar wind), which flows continuously outward past
generates a field in the same way as does the air the planets. On a quiet day, near the earth, this plasma
motion from solar heating. typically has a density of a few ions/cm3 and a velocity of
5. Solar corpuscular radiation and interplanetary field. about 400 ktis. An important feature of the plasma is its
A number of field contributions arise directly or high electrical conductivity. One result of applicable theory
indirectly from the interaction of the solar wind is that the magnetic field will be “frozen into” such a plasma;
and its imbedded magnetic field with the main that is, the ions, electrons, and magnetic field will move
field of the earth. Some important effects are the together as a compressible fluid medium. When such a mov-
compression of the main field by external plasma ing fluid encounters a stationary entity with which it can
pressure, the intrusion of solar plasma into the interact, such as the geomagnetic field, one or the other will
main field, the heating of plasma already within be deflected, swept away, or otherwise modified by the
the field, and the merging of magnetospheric and collision. The total pressure of the solar wind is the sum of
interplanetary fields. the pressure exerted by the momentum of the particles and
There are a number of other obvious possible sources that, the Maxwell pressure B2/8m of the frozen-in field. The geo-
in fact, do not contribute appreciably; examples are the magnetic field also contains highly conductive plasma, and
mantle of the earth and energetic cosmic rays. this medium similarly sustains a pressure equal to the sum
of the ambient-plasma and Maxwell pressures. When the
pressures of the interplanetary medium and the geomagnetic
4.1.4 The Steady Interior Field field are compared, it is clear that at great distances the
geomagnetic field will be swept away by the solar wind and
A geometric-temporal description of the field is con- that close to the earth the solar wind will be excluded by
structed from measurements made by observatories, sta- the field. At intermediate distances there must exist a region
tions, ships, rockets, and satellites, all of which are re- of interaction where the pressures are comparable and where
stricted in geometric coverage (geographic or spatial location) rather complicated features can be expected. In the last
and temporal coverage (time period and frequency re- decade, satellite experiments and theoretical development
sponse). Observed phenomena tend to be classified accord- have discovered and clarified the principal features of this
ingly. The traditional classification by frequency is very interaction. Some of them are illustrated in Figure 4-5. (The
useful and is retained here. 23° tilt of the rotational axis and the 11.5° tilt of the dipole

4-5
 //
CHAPTER 4
/0
0’

1~
0

20 0’
. . . . .. . .. .. . .. .. .. ... . . . . .. .. .. .. .. . . . .. . ... ...... . .. .....
/ .... .. .. .
/ ... .... ..........
/ .... .. . . ....
/ .....

TRAPPING L NEUTRAL
L PLASMA
BOUNDARY SHEET
SHEET
\ ~ PLASM APAUSE

NEUTRAL
POINT

>’.. \ \ ... ........... .. ................ . ..... . ...... .. . .. . ... ... ......

BOW -\ MAGNETOPAUSE
SHOCK ‘\
\
\\
“\
\\

Figu~4-5. Thegeneral cmsfiguration of the magnetosphere, shown ina noon-mendian section.

axis have been neglected for simplicity. ) The field of the the tail has been observed at a distance of more than 1000
earth extends to a geocentric distance of 10 RE toward the RE behind the earth. It might also be expected that within
sun (RE being a unit of length equal to the radius of the the tail the north-polar field lines would be smoothly con-
earth) at which distance it terminates abruptly in a thin layer nected across the equatorial plane to the corresponding south-
known as the magnetopause. The region interior to this layer polar field lines, but this also does not happert; these field
is known as the magnetosphere, The region exterior to it lines are drawn out into the tail, directed toward the earth
contains the solar wind, which is “piled up”, that is, com- above the plane and away from the earth below it. Thus,
pressed, deflected, heated, and made turbulent by the col- beyond a geocentric distance of about 10 RE in the tail, the
lision. These effects propagate some distance upstream, with equatorial plane (neglecting tilts) is a sort of neutral sheet,
the result that the wind is slowed over a distance of a few across which there is a fairly abrupt field reversal, and the
RE. Since the velocity of the undisturbed wind is “super- component perpendicular to the neutral sheet is very small.
sonic”, there exists a surface at which the velocity is just A surface of discontinuity in the magnetic field implies
“sonic” and a stationary shock front, the bow shock, is the existence of a current flow in the surface, and the current
created. The magnetopause is typically at about 10 RE and pattern can be inferred from the field. On the sunward mag-
the bow shock at about 14 RE on the sun-earth line. The netopause the flow is characterized by an eastward current
region between these is called the magnetosheath. Field lines sheet (dawn-to-dusk) across the nose (subsolar point) of the
from the high-latitude (polar cap) regions are swept back magnetosphere. In the neutral sheet the flow is westward
toward the night side and fotTn a long geomagnetic tail. across the tail (also dawn-to-dusk), with retu~ loops on the
Although it might be expected that solar wind pressure per- tailward part of the magnetopause. These currents are shown
pendicular to the direction of its bulk flow should close the schematically in Figure 4-6. (The composite magnetopause
magnetosphere within a few tens of RE behind the earth, current is not shown. )
that is not the case; the combined pressure of field and The current system of the magnetopause acts to cancel
plasma within the tail is sufficient to prevent closing, and the dipole field outside and enhance the field inside the

4-6
 THE GEOMAGNETIC FIELD
for separating field contributions have been somewhat ar-
bitrary and subject to personal judgment. These fields were
originally defined on the basis of data taken during a few
of the quietest days per month. With better understanding
of magnetospheric disturbance, improved’ measurements,
and an awareness that quietness is only relative (any day
being only more or less disturbed), it became more common
to consider them in an idealized sense as being those vari-
ations that would exist if the earth were subjected only to
an absolutely quiet external environment. More recently,
evidence of a direct connection between quiet and disturb-
ance variations has made this viewpoint less useful.
Quiet variation fields include several contributions. The
solar quiet (Sq) variation field, which has a peak-to-peak
amplitude of several tens of nanoteslas at most surface lo-
cations, is caused mostly by the generation of ionospheric
Figure 4-6. Flow patterns of the two principal current systems which
determine the configuration of the magnetosphere [adapted
currents by solar electromagnetic radiation. The L (lunar
after Axford, 1965] daily) variation field, which typically has an amplitude of
a few nanoteslas at the surface, results from the generation
magnetopause. This is equivalent to a compression of the of ionospheric currents by luni-solar atmospheric tides. An-
geomagnetic field by the cavity to which it is confined. other contribution of a few nanoteslas at the surface results
Because the cavity almost totally surrounds the earth, the from the confinement of the main field by the solar wind,
field is compressed on all sides, but since the tail is open the compression being stronger on the day side than on the
and the highest pressure is on the nose of the cavity, the night side. Quiet variation fields are discussed in Section
compression is somewhat less on the night side. The 4.4, which for both historical and practical reasons retains
compression results in an average increase of the equatorial a ground-based perspective.
surface field of about 0.170 (about 30 nT); just inside the
magnetopause the increase is 1007o (which is about 30 nT
when the magnetopause is at 10 R~ but about 60 nT if it 4.1.6 Disturbance External Fields
has been pushed in to 8 R~). About a third of the surface
increase results from diamagnetism in the solid earth. Variations in the geomagnetic field that do not have a
The steady field then consists basically of the main field simple periodicity and appear to result from changes in the
of the earth compressed by the cavity to which it is confined. interplanetary environment are called disturbed variation
In addition, most of time-varying field contributions dis- fields or geomagnetic disturbance and are denoted by D.
cussed below are also (like the magnetopause current) likely The D field is that remaining after the steady and the quiet
to have an average dc value which may be thought of as variation fields have been subtracted from the total. Large
part of the steady field. For example, convection of the disturbances of relatively long duration, the behavior of
outer magnetosphere and the flowing of a ring current are which suggests some magnetospheric events as the cause,
processes which continue even on the quietest days. are termed geomagnetic storms. Except for some fluctua-
tions attributable to irregular motion of the upper atmos-
phere, the sun is responsible for all disturbance effects rec-
4.1.5 Quiet Variation External Fields ognized at present, and with only two exceptions, it is the
solar wind with the frozen-in solar magnetic field that trans-
The earth with its core, atmosphere, and main field mits the disturbance to the vicinity of the earth. The two
rotates in the interplanetary environment and moves along exceptions are disturbances in which ionospheric conduc-
its orbit so that any point stationary in geographic coordi- tivity is enhanced as the result of solar flares: polar cap
nates experiences periodic variations in gravity force, solar absorption events (PCA), which result from low-energy pro-
illumination, and compression or other modification by solar tons from the flare, and solar-flare effects (SFE), which
wind effects. The field contributions that result from these result from x-ray emissions from the flare.
motions vary diurnally and seasonally. Field contributions Historically, the disturbance field has been studied by
that vary this slowly and regularly and do not result from ground observations, with the hope of separating the ob-
disturbances in the interplanetary environment are known served surface field into components that could be explained
as quiet variation fields. The analysis of experimental data in terms of current systems above the earth. A useful dis-
to determine the quiet variation fields is of course difficult tinction has been the separation of the component that de-
in the presence of magnetospheric disturbance, and criteria pends only on universal time (UT) from that which depends

4-7
CHAPTER 4
on local time (LT); the former, usually called the storm- 4.2.1 Instrumentation
time variation and denoted by Dst, is, by definition, sym-
metric about the polar axis, while the latter, called the dis- Instruments used over the past several hundred years to
turbance-daily variation and denoted by DS (or Ds), is asym- measure the intensity and direction of the magnetic field
metric. Then D = Dst + DS. The component Dst was have been few in number and simple in principle, but during
attributed to a ring current encircling the earth a few RE the past century they have been made very reliable and fairly
above the equator, while DS was attributed to ionospheric precise. Although greater precision and sensitivity were needed
currents generated by auroral particles precipitated from the earlier, major developments of new instruments came only
ring current. Although better knowledge of the magneto- in the past 35 years, partly because the older instruments
sphere has made clear that ionospheric and magnetosphenc were not adaptable for use on rockets and satellites. The
current systems are intimately related, this separation is still principal instruments currently in use may be listed as fol-
useful. Other separations have been made or proposed, usu- lows.
ally relating to a theoretical model of some postulated phys- Ground-based instruments exploit several physical prin-
ical process. Many are currently in use in the literature but ciples. Several of the older instruments are based on the
are likely to be revised as understanding improves. alignment or oscillation of a permanently magnetized needle
Except as noted above, geomagnetic disturbance results in the field; these include the compass, dip circle, and mag-
from the interaction of the solar wind with the geomagnetic netic theodolite, which measure D, I, and H, respectively,
field. While some minimum level of disturbance may be the three elements usually measured at observatories to de-
expected to result from turbulence generated by instabilities termine the field. Several others rely on the induction of a
in the flow of plasma around the magnetosphere even if voltage in a coil of wire. The coil may be rotated in the
solar wind properties were absolutely constant, most dis- field as in the dip inductor, or may be fixed as in a large
turbance phenomena having characteristic times of minutes induction-coil magnetometer used to measure geomagnetic
to days and observable in ground-based measurements of pulsations. Two magnetometers are based on the cancel-
the field result from variations (often abrupt) in one or more lation of a component of the geomagnetic field by the known
of the solar wind parameters (for example, the density or field of an electromagnet; these are the H-magnetometer of
velocity of the plasma or the direction or intensity of the Schuster and Smith and the Z-magnetometer of Dye. Of the
interplanetary field). The largest disturbances of the mag- newer instruments, several are based on atomic-resonance
netosphere are called magnetospheric storms and the cor- techniques; these are the proton precession (and proton vec-
responding disturbances of the geomagnetic field are called tor), rubidium-vapor, and helium magnetometers. Another
geomagnetic storms. While phenomena vary greatly from widely used instrument exploits the saturation characteristics
storm to storm it is possible to describe a typical or “classic” of a ferromagnetic core; this is the fluxgate or saturable-
magnetic storm (see Section 4.5. 1). Many other complex core magnetometer. Most recently, a number of extremely
dynamic processes in the magnetosphere are manifested in sensitive instruments have been developed which utilize the
magnetic-field disturbance; some of these are discussed in quantum-mechanical behavior of Josephson junctions in a
Chapter 8. superconducting loop; these are known as SQUID magne-
The dynamic behavior of the magnetosphere also in- tometers (for “superconducting quantum interference de-
cludes oscillations. especially in accompaniment to slower vices”).
magnetic disturbance, both because it is an elastic entity All of these magnetometers are in use for ground meas-
which can resonate and because a number of its dynamic urements. Satellite and rocket measurements rely mainly on
processes generate oscillatory currents. Periodic and ape- the rubidium-vapor, induction-coil (often called searchcoil),
riodic field fluctuations with frequencies covering nearly and fluxgate magnetometers, which inherent] y have small
eight decades (10-3 to 10$ Hz) are observed. In the lower size, low weight, modest power requirements, and an easily
frequency range (ULF up to 5 Hz) they are called geomag- telemetered output. A brief description of several of the
netic pulsations. These are discussed in Section 4.5.2. Higher most important of these instruments follows.
frequencies (ELF and VLF) are associated with the dynam- Fixed induction coils of various sizes are used to measure
ics of ionospheric and magnetospheric plasmas. rapid fluctuations in the field. To measure the vertical com-
ponent, horizontal coils with diameters of nearly 10 km are
laid out on the ground; for the other components, coils a
4.2 MEASUREMENTS OF THE few meters in diameter, but with many turns, are used. Also
GEOMAGNETIC FIELD used for this purpose are much smaller coils which are
wound around laminated mu-metal cores which concentrate
Geomagnetic phenomena are studied experimentally magnetic flux for increased sensitivityy. Since the quantity
through data obtained by ground stations, ships, aircraft, measured is the time rate of change of the field, the sen-
and space vehicles. This section discusses the instruments sitivity is inherently proportional to the frequency of the
used for geomagnetic measurements and reviews the prin- fluctuation. A metal-core coil of 30000 turns, having a
cipal sources of such measurements. diameter of 7 cm and a length of 2 m, can detect variations

4-8
 a

THE GEOMAGNETIC FIELD

of 0.001 nT at one Hz. A typical spacecraft searchcoil erated fields are known; an uncertainty as low as about 0.3
having a diameter of 2 cm and a length of 30 cm has a nT is possible. This instrument is used to measure H and
sensitivity 1000 times less. Z at many observatories.
The first and best developed of the atomic- or nuclear- In the last 25 years, a newer resonance instrument has
resonance instruments is the proton precession magnetom- been widely used. This is the alkali-vapor magnetometer,
eter. The physical principle on which it depends follows. which relies on the Zeeman effect and the phenomenon of
Individual protons in a hydrogenous material placed in a optical pumping. Any alkali vapor is suitable, but the ru-
magnetic field have both a magnetic moment and an angular bidium isotopes 85 and 87 have been most used. The energy-
momentum, which coincide in direction; the field exerts on level diagram for Rb-87, showing the Zeeman splitting in
the proton a torque tending to align its moment with the a magnetic field, is shown in Figure 4-7. When light having
field, but the existence of the angular momentum causes the a wavelength of about 0.79476 pm is passed through a
common vector to precess about the field direction. Nor- transparent cell filled with the vapor, resonance absorption
mally the processing vectors are random in phase and pro- and re-emission occurs involving transitions between the
duce no coherent signal, but if they are started with a com- various Zeeman sublevels of the ground and first excited
mon phase by suddenly releasing them after polarization by states. If the light is circularly polarized, the absorption
a strong field perpendicular to the field to be measured, they transitions must have Am = + 1, so no transitions from
precess for some time in unison, producing at the precession the groundstate sublevel with m = + 2 can occur. Even-
frequency a signal which can be detected by a pickup coil tually, all electrons are trapped in this substate and no further
surrounding the material. The precession (Larmor) fre- absorption can take place; the vapor becomes magnetically
quency is directly proportional to the field, the constant of polarized and transparent. This process is called optical
proportionality being l/2m times the proton gyromagnetic pumping. The polarization can be destroyed by impressing
ratio. This physical constant, known to an accuracy of better in a direction perpendicular to the ambient field a weak
than one part in 105, has the value 26751.9 X 105 T’ s-’ magnetic field oscillating at a frequency (the Larrnor fre-
so the frequency for a field of 30000 nT is 1.2773 kHz. quency) corresponding exactly to the Zeeman splitting (6.99
In a typical instrument, the hydrogenous material is a frac- Hz/nT). Forbidden transitions between the various m-sub-
tion of a liter of water, alcohol, or n-heptane around which levels of the groundstate are induced, electrons trapped in
is wound a single coil, used first to produce the polarizing the sublevel with m = + 2 are redistributed to other sub-
field of about 0.01 T and subsequently to detect the preces- levels, resonance absorption of the light is again possible,
sion signal. After the sample is polarized, the coherence of and the vapor is no longer transparent. The ambient mag-
the precession persists for a few seconds before being de- netic field is determined by measuring the Larrnor frequency
stroyed by thermal agitation. Several precautions and cor- (that is, the oscillating-field frequency that produces max-
rections are required, but the instrument is basically simple imum light absorption). The simplest magnetometer consists
and reliable. Absolute measurements of the field can be of the vapor absorption cell surrounded by a coil to produce
made with an uncertainty as low as 0.1 nT. The sensitivity the Larrnor-frequency field, an rf-excited vapor lamp with
of the instrument can be increased to 0.01 nT by adding a a filter to absorb all but the 0.79476 ~m line, a circular
microprocessor to process the precession signal. Versions polarizer between the lamp and the cell, and a photodetector
for use in observatories, aircraft, ships, and rockets have to measure the intensity of transmitted light. The frequency
been developed and a continuously self-oscillating version of the impressed field may be adjusted manually for mini-
is under development. mum transmission of light through the cell; there is a 20%
The proton vector magnetometer combines the proton- change in absorption between the pumped and unpumped
precession magnetometer with two sets of Helmholtz coils conditions. In more refined instruments, several corrections
arranged to null the H and Z components. To measure Z, and improvements are incorporated and they are usually self-
H is first annulled by producing – H in the H coils. The oscillating; that is, both the light intensity and the impressed
null cannot be detected directly but is produced by using field oscillate with the Larrnor frequency, which is estab-
just half the current required to generate – 2H; the latter lished using a feedback signal from the photodetector. The
condition can be detected since the total intensity is then absolute accuracy of alkali-vapor magnetometers is limited
exactly the same as that with zero current in the coil. The by the inherent line width of the resonance (several nT for
current required to annul Z is then measured. First-order rubidium) and a further splitting of the Zeeman levels by
instrument errors, of which leveling alignment is most crit- second-order effects in the coupling of moments; the un-
ical, can be corrected by appropriate checks with reversed certainty in weak field regions such as the distant magne-
coils. To keep the field gradient at the sensor low enough tosphere is negligible, but in strong-field regions, such as
with moderate coil dimensions, a four-element Fenselau or near the surface of the earth, it is seldom less than about 2
Braunbeck coil array may replace the simpler Helmholtz nT.
coil. Since this instrument uses the proton precession mag- The helium magnetometer also depends on optical
netometer simply as a null detector, the precision of the pumping. Its operation is similar to that of the alkali-vapor
measurement depends on the accuracy with which the gen- instrument, but since the groundstate of helium has zero

4-9
CHAPTER 4
GROUND STATE FINE HYPE RFINE ZEEMAN SPLITTING
AND FIRST STRUCTURE STRUCTURE IN MAGNETIC FIELD
4
EXCITED STATE J =;+~ F =;+~ mF=F, F 1, .,-F.

m~.
J = 3/!
2~3/2 +2
/.- +1
. .
,— F=2
P L=l . . o
‘. , -1
--.:P1/2 J=1\2,,0’” -2
z ‘.
~. -1
‘. , F=l
0
+1
1111Iliil 1111
D2 LI
FILTEI
OUT
D I ~mr= +1 o -1

I Ill
11111

&
Ill
D,

7947.6
LINE

A
SPLITTING
6.99
IS
HZ IGAMMA
],111
Ill
1,{11Ill
I 1
+2
F= 2 . ‘:-”-: +1
.—:~~ .- ——-- 4 o
.“
>.-.- - -1
.
.s
—-——
L=O 2s1/2
---—
-2
-,
‘. -1
‘< F=l ---
— E~~——- 0
--- +1

Figure 4-7. Energy-level diagram for Rubidium-87.

magnetic moment, electrons are trapped in a substate of the odd harmonics, the second harmonic is a measure of the
metastable 23P, state instead, being excited to the metastable ambient field. Its amplitude is proportional to the magnitude
state by an rf electric field. The inherent sensitivity is higher of the field component parallel to the core, and its phase
because of a greater Zeeman splitting (28 .Hz/nT compared indicates the sign. Several schemes may be used to eliminate
with 7 Hz/nT for Rb-87), but the line width of the resonance the drive signal in the secondary winding. The primary and
is much broader (about 100 nT). secondary coils can be wound on axes which are perpen-
The fluxgate magnetometer has been in use for nearly dicular, or if the core material is separated longitudinally
half a century, but its current usefulness results from ex- into two halves, the primary winding can encircle the two
tensive development in recent years. At present, it is prob- halves in opposite senses. In either case the net excitation
ably the most widely used instrument in both ground and flux through the secondary can be made zero, and to the
space measurements for geomagnetic and magnetospheric degree that this is achieved, only the desired even-harmonic
studies. In its simplest form, it consists of a highly perme- signal is detected. A magnetometer of the latter type is
able ferromagnetic core on which primary and secondary shown schematically in Figure 4-8. Since a single sensor
coils are wound. The primary winding is driven with a measures ordy one component, most applications uw a three-
sinusoidal current that has an amplitude sufficient to drive sensor orthogonal array. It is also common to combine the
the core material into saturation twice each cycle, thus instrument with coils, either external or wound in the sensor
changing the permeability of the material at a frequency (like the calibration winding shown), to annul all or most
twice that of the primary current. The flux in the core has of the ambient field. It has also been combined with a
two sources, the exciting field and the ambient external field. servomechanism to orient a three-sensor array such that two
The former produces a large component at the driving fre- sensors read zero, the third measures the total field, and the
quency and, because of the changing permeability, other servomechanism measures the field direction. This so-called
components at odd-harmonic frequencies. The latter pro- oriknting fluxgate was used extensively in world-wide mag-
duces a large component at the second-harmonic frequency netic mapping by aircraft. Since the fluxgate is an analog
and smaller components at higher even-harmonic frequen- electronic instrument, a number of characteristics of the
cies, all due to the changing permeability y; that is, the steady circuitry and core material limit its absolute accuracy; some
ambient field, which would otherwise produce no signal, is instruments have an uncertainty of less than 1 nT. The
“gated’ into the secondary winding by the changing per- sensitivity for both gror,rrrd and space measurements is better
meabilityy of the core. Since the exciting field produces only than 0.1 nT. Since the excitation frequency may be as high

4-1o
 -

THE GEOMAGNETIC FIELD

FIELD COMPONENT
TO BE MEASURED

I I
CURRENT
SOURCE

r
PHASE INTE -
AMPLIFIER I.
+ SENSITIVE GRATING
2 f“
DETECTOR AMPLIFIER OUTPUT
SIGNAL
4

FREQUENCY I OSCILLATOR
4
+2 FREQ = 2fo

Figure 4-8. Schematic diagram of a typical fluxgate magnetometer.

as 10 kHz, field fluctuations with frequencies as high as 100 corresponding to the existence of integral and half-integral
Hz or more can be measured. multiples of the flux quantum in the loop. The SQUID is
The most recent and most sensitive instrument is the not an absolute instrument; changes in the ambient magnetic
SQUID magnetometer, which operates in liquid helium at field are measured by counting the number of maxima which
about 4 K. One version uses two Josephson junctions in a occur in an output signal as the field changes. In practice
superconducting loop as shown in Figure 4-9. Each junction
consists of two superconductors (the two halves of the loop)
separated by a thin insulator across which a current flows
because of the quantum-mechanical tunneling of electrons SUPERCONDUCTOR
through the the insulator. Quantum mechanics requires that
all electron pairs in the superconductor be in the same state,
and therefore that a single wave function describe the entire A
loop. One result is that in a loop without junctions the
magnetic flux enclosed by the loop cannot change (being
“trapped’) and is quantized in units of h/2qe (having the
value 2.07 x 10-15 Wb), where h is Planck’s constant and
q. is the charge of the electron. Adding the junctions to the
loop results in a partial breakdown of the trapping; a non- S
g:
integral multiple of the flux quantum can exist in the loop,
?s
but the behavior of the loop in attempting to maintain the 3
trapping can be observed through its effect on the Josephson ;&
dc current flowing from A to B in the figure, which varies ~ ~
as shown with the applied magnetic field. Another version fg
requiring only one junction in the loop observes this be- - -400 -s00 -m -1oo 0 Ioom?iocsw
havior by its effect on the impedance of a coil arranged to MAGKTIC FIELD (MILL IGAUSS )

couple an externally generated rf magnetic field into the

loop. In either case, the measured quantity exhibits an in- - .=—..
Fi pure ($-9. The basic geometry and Josephson-cument response of a sim-
terference pattern, that is, a series of maxima and minima ple dc SQUID magnetometer.

4-11
CHAPTER 4
both dc and rf instruments are equipped with feedback cir- 4.2.3 Satellite and Rocket Measurements
cuitry to keep the detector locked on a single maximum,
and the amount of the feedback signal becomes the measure Sounding rockets carrying magnetometers have been used
of the magnetic field. The SQUID is a vector instrument, since 1949 to measure the intensity and height distribution
since only the component perpendicular to the loop con- of ionospheric current systems by measuring the change in
tributes. The sensitivity is extremely high; field changes as the transverse magnetic-field component as the current layer
small as 10-5 nT can be observed. The main limitation is was traversed and comparing it with that expected from an
the need to keep the thermal energy kT very small compared idealized infinite current sheet. The equatorial electrojet,
with the energy of a flux quantum so the interference pattern the midlatitude Sq currents, and the auroral electrojet have
is not obscured by thermal noise, and this limits the area all been studied in this manner using primarily proton-
of the loop to about 1 or 2 mm’. The magnetometer inher- precession or alkali-vapor magnetometers but occasionally
ently possesses excellent linearity, range, and frequency fluxgate or searchcoil instruments. Recently, attention has
response. Its disadvantages are that it must be calibrated to centered on the fine structure of field-aligned currents of the
some other standard, is susceptible to electromagnetic in- auroral oval, and a series of rocket-borne experiments has
terference (which can cause loss of lock to the reference attempted to determine the configuration of the lines of force
maximum and hence its calibration), and must be operated of the geomagnetic field by injecting energetic electron kams
in a liquid-helium cryostat. along the field line at rocket altitudes and detecting their
return.
The use of satellites for magnetic measurements began
in 1959 with the Sputnik 3 and Vanguard 3 spacecraft.
4.2.2 Ground Measurements Measurements since then have provided the drastically al-
tered concept of the magnetosphere described in Section
Since the early nineteenth century, the principal source 4.1.4. In the last decade, deep-space satellites with highly
of geomagnetic data has been a continually increasing num- elliptical orbits (Hawkeye 1; Heos 1 and 2; IMP H, I, and
ber of magnetic observatories throughout the world. Re- J; ISEE 1 and 2; OGO 5; and Prognoz 4, 5, 6, and 7) have
cently, the advent of scientific satellites has greatly advanced carried fluxgate magnetometers to the boundary regions of
the precision of global surveys and the study of magnetos- the magnetosphere. Other satellites in elliptical orbits of
pheric phenomena, but this has not diminished the need for lower apogee (Explorer 45, GEOS 1, and Jikiken) made
ground-based measurements. measurements largely within the magnetosphere.
By 1840 about 50 observatories were making coordi- Low-altitude polar satellites are ideal for making meas-
nated measurements of the declination by hourly intervals. urements of the geomagnetic field to derive a spherical-
In recent years over 250 stations have been in continuous harmonic model of the internal geomagnetic field, since they
or intermittent operation. Of these, over 130 are formal are nearly simultaneous and have a uniform global distri-
observatories, most of which publish data on a regular basis. bution. The OGO 6 satellite accomplished this during
Others are repeat stations which are carefully marked and 1969–1 970 using a rubidium-vapor magnetometer to obtain
used periodically for standard measurements with portable accurate scalar measurements of the field. These were im-
instruments. Still others are special stations set up for a proved on during 1979–1 980 by the MAGSAT spacecraft,
particular research problem. Their geographic distribution which obtained both scalar values, using a cesium-vapor
is shown in Figure 4-10, and a list of most stations in recent magnetometer, and vector component values, using an au-
operation is given in Section 4.8.3. Almost all observatories tomatic-field-offset fluxgate magnetometer. The scalar val-
measure three elements to define the vector field with an ues have an accuracy of 1 nT; instrument and attitude un-
accuracy of about 1 nT and a time resolution of one minute. certainties limit the accuracy of the vector values to 6 nT.
Many also derive K indices and make rapid-run magneto- In the last decade, high-resolution fluxgate magnetom-
grams for better time resolution. Many of these data are eters on the polar-orbiting TRIAD and Air Force S3-2 sat-
available from the data centers listed in Section 4.8.1. ellites measured the distribution and intensity of field-aligned
In recent years, a number of ground-station networks currents connecting the outer magnetosphere to the polar
have been constructed to provide the particular geographic and auroral regions of the ionosphere. By applying spin-
coverage needed for the study of ionospheric and magne- averaging techniques, data from the fluxgate magnetometers
tospheric dynamics. The largest number were in operation (carned for attitude determination on the ISIS 2, AE-C, and
for the three-year International Magnetospheric Study (IMS) S3-3 satellites) were also used to measure field-aligned cur-
commencing in 1977, including five in North America, three rents.
in Europe, and (through international cooperation) one which Magnetometers aboard geostationary satellites respond
ci. cled the globe. Some are still in continuous or intermittent primarily to time variations in the geomagnetic field and
operation. Data from many of these are in digital form, have have been useful as extraterrestrial observatories to monitor
10-second (or better) time resolution, and are available from the fluctuations caused by magnetic storms. In the last dec-
the National Geophysical Data Center. ade, such spacecraft, all equipped with three-component

4-12
 THE GEOMAGNETIC FIELD

90 I I I I I I I I I I I I
I I I I I I I I 1
●
● I* I I I I
b
8
m ●✛
●
●

● l I I I I wi Avl
I I ●

. I
✍❞
✍✎

&
●
●

-90 l-t , I i I I I i t 1 I I I
-1 o 30 60 90 120 150 180
-180 -150
East Longitude (degrees)

Figure 4-10. Geographical distribution of magnetic observatories.

fluxgate magnetometers, have included ATS 5 and 6; GEOS measurements and mathematical models (Section 4.6) are
2, GOES 1, 2, and 3; SMS 1 and 2; and (in an orbit slightly both sufficiently accurate that differences between them are
removed from the synchronous position) SCATHA. not detectable on world charts of moderate scale, and charts
are plotted from the models. Charts of the magnetic elements
F, H, Z, and D are presented in Figures 4-1 I through
4.3 THE MAIN FIELD 4-14, respectively. These charts for epoch 1980.0 are plotted
from the GSFC 9/80 model [Langel, 1982], but differences
The steady field includes both the main field of terrestrial between all recent models and the actual field are too small
origin and the nonvarying components of external current to detect on these small-scale plots. Crustal anomalies, which
systems. While only the former is discussed here, it should seldom exceed 100 km in extent but are resolved by current
be remembered that the latter contributes to the surface field models, are similarly unresolved in these plots.
an amount which exceeds the uncertainty of present satellite An explanation of the source of the main field must be
survey measurements. consistent with seismic and other geophysical data, and
many proposed explanations have been discarded on the
basis of convincing arguments. The most satisfactory theory
4.3.1 Basic Description is that the field is generated by a self-exciting dynamo sys-
tem in which an emf generated by the motion of a conductor
The detailed characteristics of the main field are most (molten iron) in a magnetic (excitation) field produces a
easily shown in world charts of the elements. Historically current so oriented as to produce the excitation field. Al-
such charts were prepared by the hand fitting of curves to though a detailed understanding of core circulations is still
observatory and survey data. Now, however, experimental lacking, it is believed that the dipole part of the field results

4-13
CHAPTER 4
MAGNITUDE

Longitude

Figure 4-11. Contours ofconstant total tield Fatthesurface of the earth from themodel IGRF1980.O

75.

60.

4s.

30.
—/

-15.

-30.

-45.

-60.
P 7..\\-

-75,
/ -. — —
,,, ,
-an
““.
Ii-i-l I I I I I I I 1 1 1 1 1 1 I I I I I I
180.
-180. -150. -120. -90. -60. -30. 0. 30. 60. 90. 120. 150.

Longitude

Figure 4-12. Contours ofconstant horizontal field Hatthesutface of the earth from themodel IGRF1980.O.

4-14
 THE GEOMAGNETIC FIELD

VERTICAL

1 1 1 I 1 I I I ! I I I I 1 I [ I I I 1
-90. -60. -30. 0. 30. 60. 90. 120. 150. 180.

Longitude

Figure 4-13 Contours of constant vertical field Z at the surface of the earth from themodel IGRF 1980.0.

DECLINATION
90. ,
--------------- . .

75.

60.

30.

15.
#
g o.
q
d

-15.

-30.

-45.

-60.

-75.

-90.
-180. -150. -120. -90. -60. -30. 0. 30. 60. 90. 120. 150. 180.
Longitude

Figure 4-14. Contours of constant declination D at the surface of the earth from the model IGRF 1980.0.

4-1$
CHAPTER 4

‘-3-
-,’ ( / A

>
\\\
-t’()~
-180 -150 -120 -90 -60
East
-30
Longitude
0
1/

30
(degrees)
lx
l/,’l

60
1/ I
T

l\ I
120
N 150 180

Figure 4-15. The vertical component of the nondipole field, extrapolated to the surface of the core (in oersteds). The locations of nine radial dipoles titted
to the field are indicated by solid circles [see Alldredge and Hurwitz, 1964].

from a major two-dimensional circulation while the non- The variation during the past 400 years or so can be studied
dipole regional anomalies arise from eddy circulations in on the basis of historical records. For most of this period
the outer layer of the core. The regional anomalies can be measurements separated by many years have been required,
fitted fairly well by assuming that there exist within the core but current satellite surveys are precise enough to show
about nine radial dipoles of various strengths that are equiv- changes occurring over an interval as short as a year. The
alent to circulation vortices [Alldredge and Hurwitz, 1964]. variation has traditional y been found by comparing world
These are shown in Figure 4-15. A new analysis of this type magnetic charts for successive epochs; it is now derived by
will be made of the MAGSAT data [Langel, 1980]. There computer programs that fit mathematical models to the ex-
are also local anomalies of limited extent caused by per- perimental data from different periods. Defined more nar-
manent magnetism in the crust; because this source is so rowly, the secular variation refers to the time derivatives of
close to the surface, these fields sometimes exceed that the magnetic elements of the main field. The secular var-
arising from the core. The largest is the Kursk anomaly, iation of the total field F is shown in Figure 4-17.
400 km south of Moscow, which is shown in Figure 4-16. The secular variation observed over recent decades and
centuries can be described as the sum of four components
that can be described individually: a decrease in dipole
4.3.2 The Secular Variation strength, a westward drift of regional anomalies, a north-
ward movement of the dipole, and residual nondrifting vari-
The secular variation refers to very slow changes in the ations.
main field, the time scale generally being measured in years. Over the entire period of recorded data, the dipole stingth

4-16
 h

THE GEOMAGNETIC FIELD

53 I I I

\’ \ ‘ ‘ ‘ \\\\\
\\\\l
0
Q A
Horizontal Intensity

I ,\l:co
Vertical Intens, ty

52

~ 51’

%
J

1 km

50

N51° A9
E 36°50’ 51’ 52’ 53’ 54’ 55’ 36”50’ 51’ 52’ 53’ 54’ 55’

Long#tude

Fizure 4-16. The northern wrtion of the Kursk anomaly. IsointensitYcontours of the horizontal and vertical field are shown in units of oersteds Iafter
Chapman and’Bartels, 1940].

has been decreasing at an average rate of about 0.0570 per archaeological samples is a measure of the geomagnetic field
year (16 nT per year at the equator); data for the past 150 that existed at the time of their production. If the NRM is
years are plotted in Figure 4-18. stable, its direction is the same as and its intensity is pro-
It has long been observed that the major regional anom- portional to the field in which the sample was formed; how-
alies in the field appear to be moving westward, and math- ever, only certain combinations of material, physical pro-
ematical analysis [Nagata, 1962] has confirmed that about cess, and conditions result in NRM that is stable enough
60% of the secular variation not attributable to dipole weak- for reliable results. The most reliable data result from ther-
ening can be accounted for by a westward drift of the non- moremanent magnetization, locked into the sample by cool-
dipole field by about 0.2 degree per year. The cumulative ing after formation at a higher temperature. The best ar-
drift over a period of 38 years is easily observed in the chaeological samples are baked earths, such as kilns or
equatorial profiles of the vertical field shown in Figure hearths, from earlier civilizations, and the best geological
4-19. samples include materials such as lava formed at high tem-
A smaller part of the variation can be described as a perature. Very careful experimental techniques are required,
northward movement of the dipole center with a velocity but the validity of a great many paleomagnetic data and
somewhat greater than 2 km per year. If this rate were to conclusions is well established.
continue, the center would be outside the core after about The study of the secular change of geomagnetic-field
1500 years, so it is likely that the northward motion is merely intensity has been extended backward in time about 5000
the current phase of an axial oscillation of the currents that years, as shown in Figure 4-20. Over the most recent 2000
generate the dipole. years the intensity has tended to decrease at an average rate
The remainder of the variation is relatively small (except of about 10 nT per year, while for the preceding 2000 years
in Antarctica). It seems to have about a dozen regional foci, an increase of similar magnitude is observed.
but it has not been accurately measured. The study of the secular change of geomagnetic-field
direction has been extended backward over both archaeo-
logical and geological time scales using data from rock
4.3.3 Paleomagnetism samples. During the past several tens of thousands of years
the direction appears to have made quasi-periodic oscilla-
Paleomagnetism, the study of the geomagnetic field in tions ranging over several tens of degrees. However, meas-
times earlier than those for which recorded data exist, is urements covering the past few millions of yem have yielded
based on the fact discovered more than a century ago that the important discovery that the dipole axis has not wandered
the natural remanent magnetism (NRM) of some rocks and over the entire earth but has remained quite closely aligned

4-17
 FIELD MAGNITUDE [nTl
w.
7s,

w.

4s.

30.

1s.

0.

-1s.

-m.

-4s.

-50.

-75.

-w.
-1oo. -150. -120. -w. -50. -a. o. m m. w. 120. 150. 150.

RATE OF CHANGE [nT/yr]

so

30

15

.15

-30

do

.m
a
.150 .1s0 .l’m .90 a -30 0 a w m I’m 1s0 lm

ACCELERATION lnT/ yr/yr ]

Figure 4-17. The secular variation of the total field for epoch 1980.0. Shown are the field itself (top panel), its first time derivative (center panel), and
its second time derivative (bottom panel) [Lange], 1982].

4-18
 THE GEOMAGNETIC FIELD
with the axis of rotation. There seems generally to be a
clockwise motion of the dipole axis about its mean position
with a period of roughly 10000 years suggesting a preces-
sion about the axis of rotation.
For earlier geological ages, paleomagnetic (as well as
paleoclimatological) data obtained on any one continent tend
to be consistent and yield a time history of the apparent
motion of the magnetic poles; over hundreds of millions of
years, such “virtual poles” seem to have moved systemat-
ically by many tens of degrees toward their present loca-
tions. However, there is a very large apparent disagreement
between traces for different continents. The explanation is
that during this time period the continents themselves have
drifted large distances from their original locations.
1 1 I 1 I 1 1 I I 1 1 1 I 1 1 , Measurements covering about 500 million years have
0
:~ : 2 revealed that there exist reversely magnetized rocks, and
2
Year A, D
careful study has established that these indicate that the field
has periodically undergone complete reversals, the latest
Figure 4-18. Change in equatorial field strength over the past 150 years being only about one million years ago. It is not clew whether
[historical data from Vestine, 1962].
the present weakening of the dipole field represents the
beginning of another reversal or a less drastic oscillation.

L , ! ( I ,
4
4.4 QUIET VARIATION FIELDS

A daily variation of the surface magnetic field was first

noted in 1722. Although the ionosphere was not discovered
until 1902, it was predicted as early as 1882 that the cause
of the variation was electric current in a conducting layer
of the atmosphere. Subsequent study has shown that the
quiet variation includes a large effect of solar origin, a
smaller effect of lunar origin, and a still smaller remainder
due to other magnetospheric processes.
t t 1 , , i

o 90 180 270 360

EAST LONGITUDE (DEGREES)

4.4.1 The Solar Quiet Daily Variation

Figure 4-19. Profiles of the vertical equatorial field in the yeara 1907 and
1945, showing the cumulative westward drift of the non- The solar quiet daily variation (the Sq field) results prin-
dipole field over 38 years [after Bullard et al., 1950].
cipally from currents flowing in the E layer of the iono-
sphere. To a first approximation, this current system is sta-
tionary in nonrotating coordinates, and the field variation is
0.6 -
1 1 1 1,, , 1,, ,,, ,,, ,!,.
observed on the ground as a function of local time because
.
“*.
the earth rotates under the currents; therefore, it is similar
. .. for all observers at the same latitude, having the dependence
,&a.
0 ,-: —-.
W“oo
0 “ shown in Figure 4-21. However, this is a poor approxi-
// “.
\o<: “* ●

A*
. k“
mation since the conductivities that determine the current
/’ .
>/’. . . pattern are controlled by the magnetic field, which is tilted,
:%
=. Key .JaPan .<
0Russia and the Sq variation is more nearly the same along contours
A France

——
— Ob,erva+orles
Genera! trevd
of constant dip latitude than at constant geographic latitude.
I
There is also a longitudinal dependence from such effects
as the influence of ocean areas upon the strength of induced
Year A. D.
currents.
Both the conductivity, which permits this current to
Figure 4-20. Equatorial field intensity in recent millenia, as deduced from
flow, and most of the electric field, which powers the cur-
measurements on archaeological samples and recent observ-
atory data [after Nagata and Ozima, 1%7]. rent, are produced by solar electromagnetic radiation. The

4-19
CHAPTER 4
T r
1 I \ 1 1 I is bounded in the vertical direction, the Hall current is in-

H 150~ z I‘or D
A
!’O’ hibited, and a polarization results. It is found experimentally
and theoretically that the polarization enhances the effective
conductivity in the direction of the electric field (Cowling
60
conductivity). At all other points, even slightly off the dip
50 equator, the conductivity along the slightly tilted field lines
I
40 -
is sufficient to allow the polarization to leak off partially,
z and the Cowling conductivity is much less enhanced.
LLl 30 -
w
K
Both the strength and pattern of the Sq variation show
~ 20 -
a dependence on longitude, season, year, and solar cycle
~ ,0 _
[Matsushita and Maeda, 1965a]. The dependence on season
W5 - is strong, with the current vortex in a given hemisphere
0
3 becoming more intense during its local summer. The de-
~o -
pendence on solar cycle is also strong; while E-layer ioni-
<
J zation increases by 50% from solar minimum to solar max-
-5 –
I imum, the Sq variation increases by about 100%, presumably
; -lo -
0 because the wind speed also increases. The Sq variation is
z -20 - enhanced and diminished by the changes in solar radiation
E -30 - produced by solar flares and eclipses, respectively.
n
-40 - The Sq field also exhibits large changes from day to
day, but the reasons for this are not yet well understood.
-50 –
The current evidence that the Sq current system is partially
-60
—
driven by magnetospheric processes associated with dis-
I ( 1 I 1 I 1 1 I turbance phenomena blurs the traditional distinction be-
00 12 24 00 12 24 00 12 24
LOCAL 50LAR TIME (HOURS)
tween quiet variation and disturbance variation fields.

Figure 4-21. Worldwide average of the solar quiet variation near the
equinoxes at solar maximum (March, April,September,and
October, 1958)[after Matsushha, 1967].
4.4.2 The Lunar Daily Variation

The lunar daily variation (the L field) is generated in

major part of the electric field appears to be generated in the same manner as the Sq field, except that the responsible
the manner of a dynamo by high-speed tidal winds produced winds are produced by luni-solar gravitational tides and
bysolar heating of the atmosphere, resulting in a variation there is currently no evidence for any contribution to the
termed SqO. However, it is also clear that part of the electric electric field by magnetospheric processes. The dominant
field, particularly the high-latitude part, originates in the behavior is a semidiumal variation; the amplitude is about
magnetosphere and is communicated to the ionosphere by an order of magnitude smaller than the Sq amplitude. As
field-aligned currents resulting in a variation termed Sqp in the case of Sq, about 30% of the L field is produced by
[Matsushita, 1975]. Figure 4-22 shows the SqO and total induced earth currents. Figure 4-23 shows the average L
Sq (SqO + Sqp) ionospheric current systems inferred from variation in the elements H, Z, and D near the time of an
a global array of measurements. Induced earth currents (not equinox and for a mean lunar age. Figure 4-24 shows the
shown) contribute roughly one-quarter to one-third of the inferred ionospheric (but not induced) currents. A lunar
total Sq field. equatorial electrojet is a principaJ feature, existing (and being
One notable feature is a concentration of current at the more intense than shown) for the same reasons given in the
magnetic dip equator. This so-called equatorial electrojet case of the Sq variation.
is, in fact, only a few hundred kilometers wide, more con- The dependence of L on several parameters is consistent
centrated than can be reproduced in the figure by the spher- with expectations [Matsushita and Maeda, 1965b]. There is
ical-harmonic model used to compute the currents. The elec- a seasonal dependence, 2s for Sq, and also a dependence
trojet exists because of a special circumstance. The fact that on lunar age. The solar-cycle dependence on L is smaller
the field at the dip equator is exactly horizontal creates a than that of Sq, the variation being about 30% (instead of
narrow belt of high conductivity in the following way. An 100Yo) greater at solar maximum than at solar minimum
electric field impressed perpendicular to a magnetic field because the increased activity increases only the conductiv-
(here eastward and northward, respectively) would normally ities and not the tidal wind speeds. The longitudinal de-
produce a Hall current flowing perpendicular to both (here pendence, if it exists, is too small to have heen established
vertically). However, in this case the conductive medium to date.

4-20
 THE GEOMAGNETIC FIELD

00

18 18

12

90, I xl I I I / I [ I

LOCAL TIME LOCAL TIME

Figure 4-22, Ionospheric currents inferred from the observed Sq variation. The right panels include only the tide-produced currents (Sq’)). The left panels
include the convection-produced polar currents (Sq = SqO + SqP). The current between adjacent solid (broken) contours is 10000 amperes
counterclockwise (clockwise) [Matsushita, 1975].

4.4.3 Magnetospheric Daily Variation 4.5 DISTURBANCE FIELDS

Except for the contribution to the high-latitude Sq field,

the daily variation in the surface field that results from the 4.5.1 Geomagnetic Storms and Substorms
rotation of the earth within its magnetosphere has not been
observed directly since it is similar to and certainly smaller The term geomagnetic storm refers to the geomagnetic
than the Sq and L variations. The dayside-nightside differ- effects of a magnetospheric storm, which, broad] y defined,
ence in compression of the field by a quiet solar wind results is any large prolonged disturbance of the magnetosphere by
in a surface diurnal variation computed to be about 3 nT, variations in the solar wind. These storms, observed in
and the surface diurnal components of other magnetospheric recordings (magnetograms) of the surface magnetic field,
fields (such as a quiet-time ring current) that could arise exhibit great variability and complexity, reflecting the com-
from asymmetric geometries are probably negligible. In the plexity of solar phenomena. However, a classic storm, the
outer magnetosphere, of course, such diurnal effects are features of which are frequently observed, can be described
large, but measurements made there are usually referenced as follows: It includes two energizing parts and a subsequent
to a coordinate system that does not rotate with the earth; recovery.
the effect of the rotating earth is then a somewhat different The first part consists of a sudden commencement and
problem. an initial phase. These result from a change in compression

4-21
 CHAPTER 4
I 1 1

D I 0.1’
of the magnetosphere following the passage of a disconti-
nuity such as a shock front propagating in the solar wind
- n and correlate well with the pressure exerted by the bulk
- - flow. The sudden commencement (SC)’ is seen at low-lat-
itude observatories as an impulsive increase in H, typically
having a rise time of one to six minutes and an amplitude
——— of several tens of nanoteslas and observed over the entire
n earth with a spread in arrival time of less than a minute.
4-W Depending on location and the particular storm, it may be
positive, negative, double-valued or absent. The rise time
corresponds to the time required for the discontinuity to
reach all points of the magnetopause and be transmitted to
the ground as a hydromagnetic wave. When not followed
by the later phases of a storm, this phenomenon is called a
-5 -
sudden impulse (S1). The initial phase typically lasts two
I
k to eight hours, during which the field remains compressed
~ 10
~- by the increased solar-wind pressure following the discon-
0.
-15 J-w-’ tinuity.
E
-20 The second part is the main phase. It results from an
inflation of the magnetosphere by a ring current and is best

1:~:
-30 —
correlated with a previous southward turning of the inter-
-40 ---— planetary field, which permits energy to be extracted from
-50 v-w- the solar wind by a merging of the interplanetary and geo-
magnetic fields at the magnetopause. It is seen at low lat-
itudes as a rapid decrease in the field to values which are
I I I I I I I I I I
00 ,.
1? 7A 00 12 2A co 12 2A below the prestorm level, often by more than 100 nT and
LOCAL L-UNAR TiME (HOURS)
infrequently by more than 1000 nT. It develops over a period
of a few hours to a day and is characterized by noise (large
Figure 4-23. Lunar semidiumal variation near the time of an eauinox.
fluctuations with a broad risetime spectrum) and an asym-
derived from data covering the Eriod 1841 -196~ [after
Matsushita and Maeda, 19;5b]. “ metry in local time (earliest development in the late-after-
noon sector). Since solar wind discontinuities usually in-
volve changes in both pressure and field direction, stems
typically show both compression and inflation effects, but
Ionospheric currents this is not always the case, and storms without sudden com-
~ 90 1 t 1 mencements or storms which fail to develop a main phase
\ I I I
w # I \ ,
are not uncommon.
The final part is the recovery phase. It consists of a quiet
increase of the field toward the prestorm level with a char-
acteristic time which is typically about one day but some-
times much longer; the recovery is often faster at first than
later. Recovery results from a decrease in the ring-current
plasma when the source is terminated and the existing plasma
is lost by various mechanisms.
The term magnetospheric substorm denotes a process
by which energy extracted from the solar wind and stored
in the magnetosphere is dissipated. It is so named because
the main phase of a large magnetic storm often appears to
be the superposition of many substorms, each of which
0.9005. 1 1 , t 1 I
18 24 contributes particles to the main-phase ring current. The
LOC:L LUNAR’; IME (HOURS) intermittent and impulsive nature of the substoms accounts
for the characteristic noise of the main phase. The substorm
Figure 4-24. Ionospheric currents inferred from the observed L variation. is the principal instability of the magnetosphere and very
Current between adjacent contours is 1000 amperes, and
each dot indicates a vortex center with the total current in
common; it is often observed almost daily and is seldom
thousands of amperes [after Matsushita and Maeda, 1965 b]. absent for many days. The process takes place near local

4-22
 THE GEOMAGNETIC FIELD
midnight and is manifested in auroral and geomagnetic ob- [1979], Shawhan [1979], Southwood [1979], Hughes [1982],
servations by the auroral breakup and the geomagnetic bay. Singer [1982], Southwood and Hughes [1983], and Hughes
It is now clear that it involves a short-circuiting through the [1983].
auroral-zone ionosphere (via field-aligned currents) of the Dungey [1954] suggested that the long~period ULF pul-
cross-tail neutral-sheet current. This permits closed field sations observed on the ground were hydromagnetic waves
lines that had been distended by the cross-tail current to resonating on geomagnetic field lines. In the idealized case,
relax earthward into a more dipolar shape, carrying with field lines can be considered fixed at both ends in a perfectly
them particles for the main-phase ring current. Substoms conducting ionosphere, and harmonic standing waves can
appear to be self-limiting, lasting for less than an hour larger exist on flux tubes. In space and on the ground, magnetic
amounts of energy are often dissipated in a sequence of observations of spatial variations in wave frequency and
substorms or one substorm with multiple onsets. polarization characteristics have supported this picture. An
Prior to the most recent decade, many studies of geo- approximate expression for calculating the resonance period
magnetic disturbance sought to explain various phenomena is given by the so-called WKB or time-of-flight approxi-
by constructing “equivalent” current systems confined to the mation
ionosphere. More recently, there has emerged a much better
understanding of the importance and general nature of field-
aligned currents in the magnetosphere and the intimate con-
nection between ionosphere and magnetosphere. Chapter 8
T=z —a
ds
‘fl J VA(S)
L’
[1
~
1
nimi “2 (4.1)

discusses some of the magnetospheric-ionospheric processes

and current systems, including substorms, that manifest where the Alfv6n speed is given by vA = B/(ponimi)”2, the
themselves in geomagnetic phenomena and that traditionally integration is carried out between conjugate ionospheres, q
have been studied from that perspective. is the harmonic number, ds is an element of length along
a field line, ni and mi are the number density and mass of
species i in the plasma, and L is the Mcllwain L-shell
4.5.2 Geomagnetic Pulsations parameter. This method is least accurate for the fundamental
mode but more accurate calculations of eigenperiods have
Variations of the geomagnetic field having periods from been earned out using the wave equations for low-frequency
less than one to several hundreds of seconds are observed hydromagnetic waves in a nonuniform magnetic field (for
both on the ground and in the magnetosphere. They are example, a dipole field). The full hydromagnetic wave equa-
ultra-low-frequency (ULF) waves with frequencies below tions are coupled and have not been solved in general.
the ion gyrofrequency that propagate as hydromagnetic waves However, with certain simplifying assumptions (such as
in the magnetosphere. They have commonly been called considering magnetic perturbations strictly transverse to the
micropulsations, geomagnetic pulsations, or simply pulsa- background field) they have been solved for poloidal-mode
tions. Wave amplitudes range from tenths to several hundred (radial field displacement) and toroidal-mode (azimuthti field
nanoteslas, with the largest amplitudes usually occurring in displacement) waves. A particularly useful table for cal-
the longer-period waves at high latitudes. Simultaneous pe- culating eigenperiods of the uncoupled poloidal and toroidal
riodic variations in particle precipitation, auroral intensity, mode waves is given by Cummings et al. [1969]. Solutions
electric fields, x ray bursts, and particle flux are often ob- in an arbitrary field geometry have been developed by Singer
served. The pulsations have been classified into two prin- et al. [1981]. The theoretical development of hydromagnetic
cipal types: Pc (continuous) pulsations, which often have waves has been discussed more fully by Southwood and
very sinusoidal waveforms, and Pi (irregular) pulsations. Hughes [1982].
More detailed classification is discussed by Jacobs [1970]; Several mechanisms have been suggested as sources of
further subdivisions and characteristics of the waves are magnetic pulsations; however, it is usually difficult to link
given in Figure 4-25 [Saito, 1978]. The shortest-period ULF a particular wave observation to a specific source defini-
waves are thought to result from cyclotron instabilities with tively. For waves external to the magnetosphere, sources
charged particles in the magnetosphere. An overview of include the Kelvin-Helmholtz instability (wind-over-water)
these waves can be found in Jacobs [1970] and Nishida at the magnetopause boundary, quasi-parallel wave exci-
[1978]. A description of the large variety of higher-fre- tation where waves upstream of the bow shock penetrate
quency (outside the ULF band) plasma waves in the mag- into the magnetosphere, and sudden impulses due to solar
netosphere is beyond the present scope; however, useful wind discontinuities encountering the magnetopause. For
summaries have been given by Jacobs [1970] and Shawhan waves internal to the magnetosphere, sources include mag-
[1979]. Several collections of papers presented for review netic substorms, real-space or velocity-space gradients in
at meetings have been published [Southwood, 1980; Orr, the magnetospheric plasma distribution, ionospheric con-
198 1; IAGA, 1982]. Other useful reviews have been given ductivity discontinuities, and unsteady large-scale plasma
by Saito [1969], Orr [1973], Lanzerotti and Southwood convection. Many dayside Pc pulsations probably originate

4-23
 (A) (B)
MATHEMATICAL PHYSICAL CLASSIFICATION N
— —
LAT. OIURM.
,“E. ‘ PEnloo
1 TYPE 1“ OCE vARIA. =F
IRANOE TYPE , NAME
‘R”’ (SEC) ~ A:kl.
I
PP PEARL PuLSATION

HM
c
CE
14y0R~AGNETIc
CONTINUOUS
cncmus
EMISSION
A

A
o 2

02-5 ~ Pcl
0
IPDP
3
0 -: fF;:kz’ONS
‘: ‘E ;: .:tiGh~l
AURORAL IRREGULAR \ A
3 AIP PULSATION
5-10 Pc 2
z OTHERS
. +
— Pc 3 Pc 3
10-45 PC3
OTHERS “ -
+- .
-- , ‘i”: ; :~~mmm~’
PC4 PC4
z
45-1s0 ~ pc4 Pg GIANT PULSATION
o oTHERS
i
v PC5 PC5
Ire-@ R5 -–
OTNERS
,
TF TAIL
:-- F~uTTERING ‘, i : + l~mmu:
6oo - ; Pc6 E
OTHERS – ‘“
s~t SNORT- PERlOO P!
PLB P, 6URST
~:- ~; : ~,jmrlm[,:
P,c P, (CONTINUOUS)

-40 PII p,D OAYTIME PO

p,:
SctSI)-ASSOCIATEO PcI.23 A
K i .,
OTNERS
a’ P12 P,2
__ (FORNERLY P1) ~: ;
I -I;nlmwl:
J Sfa-ASSOCIATEO
Psfo PuLSATION
3
40-’50 “2 PSC4 ~~).a_TEo PC4 s o .1 ,
o
Psi 4
u OTHERS
+
K“ psc:
SC(SC)-4SSOCIATE0 PcS A
m ~“ “0 l! :l~mmm::
— PSC6
Psi6
lw - P13
f’lp PULSATION

SUBSTORM. ASSOCIATEO .
: P,6 ‘2:’’;:”
LONA.PERIOO PULSATION ‘: i~:; ,immm;:
LOCAL TIME OF NAXIMUM OCCURRENCE
OTHERS

Figure 4-25. Classification of geomagnetic pulsations [Saito, 1978]

as surface waves generated by the KeIvin-Helmholtz insta- and the signals observed on the ground are due to Hall
bility that evanescently decay inside the magnetosphere where currents in the ionosphere. A consequence of this is that the
they couple to transverse-mode standing Alfv6n waves on polarization of the horizontal wave observed on the ground
resonant field lines [Southwood, 1974; Chen and Hasegawa, is rotated by 90° from that which would be observed in
1974]. Pi pulsations are observed primarily on the nightside space. In addition, the ionosphere acts as a spatial filter and
of the earth and have been clearly linked to the onset of screens short-scale horizontal signal variations <120 km
magnetospheric substorms [Southwood and Stuart, 1980]. (approximate height of the E region) from the ground.
The ionosphere influences long-period ULF waves in The reflection coefficient of an Alfv6n wave by the iono-
the magnetosphere as well as modifying the signal observed sphere is
on the ground from that which would be observed in space.
Wave energy is lost in the ionosphere through Joule heating
(4.2)
by Pedersen currents, and significant amounts of energy may
be deposited in the high-latitude ionosphere by waves
[Greenwald and Walker, 1980]. The ionospheric Pedersen where 2P is the height-integrated Pedersen conductivity in
currents shield the incident magnetic field from the ground, the ionosphere, and VA is the Alfv6n velocity along the

4-24
 THE GEOMAGNETIC FIELD
oscillating flux tube. The ratio of the effective field-line 4.6 QUANTITATIVE MAGNETIC-FIELD
conductance (p.v~)-’ totheionospheric conductance is one MODELS
control on wave damping in addition to wave-particle in-
teractions in the magnetosphere. For example, on the night-
side of the earth where 2P - (p.v~)-’, the wave may damp 4.6.1 The Internal Field
out after only a few cycles, forming Pi2-type pulsations,
whereas on the dayside, where 2P > (pOv~)-’, the reflec- A first approximation to the geomagnetic field near the
tance is high and waves can exist for many cycles as Pc surface of the earth is an earth-centered dipole with its axis
pulsations. The amplitude of the wave signals on the ground tilted to intersect the earth at 78. 5°N, 29 1.O”E, the geo-
is also controlled by the ionospheric conductivity and the magnetic north pole; the other intersection at 78. 5°S, 11 1.O”E
horizontal wave number of the waves. An expression re- is the geomagnetic south pole. (These values were adopted
lating the ground field strength b~ to that in the magneto- internationally to define the geomagnetic coordinate system;
sphere b~ is however, the dipole terms of the first International Geo-
magnetic Reference Field (epoch 1965 .0), also adopted in-
ternationally, place the pole at 78. 56”N, 290. 24”E, and this
(4.3)
axis is also sometimes used. ) In spherical coordinates, r, t3,
and @, with r measured from the center of the earth and f3
where h is the height of the E region, k is the horizontal measured from the dipole axis (geomagnetic colatitude), the
wave number, and ZH and XP are the height-integrated Ped- dipole field has the vector components
ersen and Hall conductivities, respectively.
Magnetic pulsations contain information about their source
B,= –$2cos0
and the regions through which they propagate. Accordingly,
attempts have been made to use ground observations of
pulsations as a diagnostic tool for infernng properties of the Be = –$ sin 8 (4.4)
solar-wind and magnetospheric environments. Ground-pul-
sation characteristics such as occurrence rate, amplitude, B+ = O.
and frequency have been correlated with various combi-
nations of the solar wind velocity and the interplanetary The total intensity is then
magnetic-field strength and its component values, param-
eters which are important for generating pulsations in the
solar wind or at the magnetopause boundary. Although this B = –$ [3 COS 6 + 1]”2 (4.5)
reserarch is still in its infancy, it may eventually be possible
to monitor solar wind properties from ground-based pul- where M is the dipole moment of the earth (about 7.9 x 10’s
sation measurements, a much easier and less costly tech- T/m3). The equation of a field line is obtained by equating
nique than in situ satellite measurements. Magnetospheric its slope to the slope of the field vector:
plasma density is another property that would be useful to
measure by ground-based observations, and attempts have rd6 Be tan 0
been made, using the notion of field-line resonance. As can
dr=~= 2
be seen from Equation (4.1), measurement of pulsation pe-
riod permits determination of plasma mass density along a
Integration yields
flux tube, providing the standing-wave harmonic and field
geometry are known. This problem is further complicated
r = rosinz 0 (4.6)
by the actual distribution of mass on the field line and it
does not distinguish plasma species. Nevertheless, it has
been used to give estimates of mass density on flux tubes. where the constant of integration r. is the geocentric distance
Another use of pulsations as diagnostics of the magneto- at which the field line crosses the geomagnetic equator. A
spheric environment is that of determining substorm onsets magnetic shell, which is the surface of revolution generated
using Pi2 pulsations. It appears that each substorm onset by rotating a field line about the dipole axis, is a useful
may be associated with a Pi2, although the converse is concept in studying charged-particle motion. In the B-L
probably not true. The Pi2 signal is particularly useful be- coordinate system, trapped particles which move adiabati-
cause it begins abruptly and can often be observed over a cally to conserve three invariants of the motion are confined
large portion of the earth. In space the signal propagates to a shell of constant L. The B-L coordinate system is,
faster than the plasma injected at substorm onset, and at roughly speaking, a transformation of the actual distorted
some distance from the injection source the Pi2 can be used field of the earth into an equivalent dipole field (employing
to foretell the arrival of energetic particles. a multipole expansion of the former, as discussed below),

4-25
CHAPTER 4
and the shells of constant L transform into the simple mag- integrated over the sphere is not unity but (2n + I )-1’2).
netic shells of the dipole field, with L = ro. The coefficients g:, h:, A:, and B: appropriate to this
The centered dipole is a poor approximation to the actual function, referred to as “Schmidt coefficients”, are now
field. It yields an equatorial field intensity of 30800 nT, quite universally used in quoting results of analyses, al-
while the actual intensity ranges between values of41 000 though earlier analyses have used other normalizations, par-
and 24000 nT, more than 25% above and below the cal- ticularly the Gaussian, in which the numerical multiplier is
culated value. Much of this discrepancy can be removed by replaced by [2” n!(n – ma)!/!].
taking the dipole to be eccentric, a fairly good fit being In the potential, those terms containing g; and h: arise
obtained by displacing the center of the dipole 436 km from sources internal to the earth, while those containing
toward 15 .6”N, 150.9”E and inclining the dipole axis to AN and B: arise from external currents; the potential func-
intersect the spherical earth at 8 1.OON, 275. 3°E and 75.0°S, tion is valid in the space above the surface and below the
120.4”E [Parkinson and Cleary, 1958]. However, discrep- external current system. The field is given by
ancies of about 10% remain, which is not good enough for
most purposes. Since much better models are readily avail- B = –VV. (4.9)
able for machine computation, the eccentric dipole is seldom
used. The northward, eastward, and downward components of the
Any field derivable from a potential function can be field are thus
expressed in terms of a multipole expansion of the potential.
The coefficients of the various terms can then be adjusted
by a least-squares method to give the best fit to the measured
field. The methodology for such analyses is well developed,
and sets of coefficients along with computer programs to 1 av
Y=–—— (4. 10)
calculate the field and a number of related parameters are r sin 8 13d)
readily available. In simplest form, the analysis is as fol-
lows. The magnetic scalar potential can be written as a
spherical-harmonic expansion:
z=:.

The construction of a model consists of adjusting the coef-

ficients g:, h:, A:, and B; to fit experimental measure-
ments taken over a network covering the entire earth, using
n+l

[()
a a least-squares procedure to minimize some quantity such
x– (g; cos m+ + h: sin mo) (4.7)
r as the weighted sum of the squares of the differences be-
—n tween calculated and measured X, Y, and Z values. As the
+:

() r
(A; cos m+ + B; sin m+)
1, degree of the multipole terms increases, the magnitude of
their effect decreases, and it is pointless to retain more terms
than are warranted by the accuracy of the experimental
where r, 6, and ~ are the geographical polar coordinates of measurements.
radial distance, colatitude, and east longitude, and a is the Twenty years ago all models were based on surface
radius of the earth (a = R~). The functions P: (COS0) are measurements. Since analyses were unable to identify con-
the Schmidt functions: tributions from external sources, the coefficients A: and
B; were usually assumed to be zero. Although measure-
ments were not simultaneous, the existing knowledge of the
secular variation was used to adjust them to a common epoch

[’::::’:1”2
P; (Cos 6) = to which the resulting model applied. The significance of
coefficients of order greater than about six was dubious.
Since then field models of much greater precision have
(1 – Cos’e)ti’ d“ + m (COS28 _ ~)“

Em[ 1
x been developed on the basis of data obtained by low-orbiting
2nn ! d(cost3)n+ m
polar satellites that can provide nearly simultaneous cov-
=2ifm>0 erage of the entire earth. A first major advance was made
possible by three spacecraft of the NASA POGO series,
Em = lifm=O.
(4.8) which carried accurate total-intensity magnetometers. A
second advance has now been afforded by the MAGSAT
The second quantity in brackets is the associated Legendre satellite, launched in late 1979, which made accurate three-
function Pn,~(cos 0). Its numerical multiplier makes the component measurements of the vector field. The increased
Schmidt function partially normalized. (Its mean-square value precision of these recent data require the retention of higher-

4-26
 THE GEOMAGNETIC FIELD
order coefficients in the multipole expansion; the latest models spherical-harmonic analysis. Therefore, a separate model
typically include terms to order 13. At satellite altitude (e .g., of the external field must be constructed, and contributions
500 km), these models specify the field to an accuracy of from the internal and external models can then be added to
about 10 nT and clearly resolve regional anomalies resulting specify the total field at any location in the magnetosphere.
from crustal magnetism. Also discernible in the data are the Large-scale current systems, often only partially under-
contributions of the external field, including that of the stood, are associated with many complex magnetospheric
magnetic-storm ring current. Since satellite surveys now processes. However, three major current systems appear to
span about 15 years, the secular variation can be increasingly explain the basic configuration of the magnetosphere: (1)
well determined. Current models derived from data span- the magnetopause current, which creates the magnetopause
ning such a time interval include the time dependence by and flows in it, confining the geomagnetic field to the mag-
specifying the coefficients for a reference epoch and also netosphere; (2) the neutral-sheet current, which is driven by
their first and second time derivatives. the dynamo-generated potential across the magnetosphere
Until recently, substantial discrepancies existed between and causes polar-cap field lines to be drawn out away from
models derived from different data bases. The value of a the earth into the tail; and (3) the ring current, which consists
standard reference field was recognized, and beginning with of plasma drifting around the earth in the region where field
epoch 1965.0 such a model, the International Geomagnetic lines have a more dipolar shape. Several useful models have
Reference Field (IGRF), has been adopted. IGRF 1965.0 been developed by Olson and Pfitzer [ 1974, 1977], Tsy-
was an average model; for any specific purpose another ganenko [1976], and others to compute the magnetic field
model was likely to provide greater accuracy, but the IGRF produced by these currents. One of the Olson-Pfitzer models
served well as a common reference. This difficulty no longer includes the tilt of the earth dipole; the other permits varying
exists; IGRF 1980.0 is the most accurate model available. the strengths of the three currents independently.
Its time-derivative coefficients permit its use until 1985 when External-field models describe a field configuration that
the next IGRF is scheduled. Retrospective studies have also is much less stable than the internal field. In studying some
resulted in a Definitive Geomagnetic Reference Field, con- magnetospheric process, the model can be used to describe
sisting of models at five-year intervals (DGRF 1965, DGRF the general field configuration while the process is modeled
1970, and DGRF 1975) with linear interpolation of coef- separately. However, the currents involved in the process
ficients between them. The DGRF is not expected to be may be intimately associated with the model currents. An
revised further, whereas the IGRF maybe (becoming DGRF example is the substom, in which part of the neutral-sheet
1980). A continuous bridge from the DGRF 1975 to the current is diverted along field lines and through the iono-
IGRF 1980 is made by the PGRF 1975 (“P’ for “provi- sphere causing an abrupt change in the configuration of the
sional”), a linear interpolation between the two. This pattern magnetospheric magnetic field. The development of self-
is expected to be followed for further updates. consistent models for such processes is just beginning.
The coefficients for the three DGRF models and IGRF
1980 have been published in EoS [IAGA, 1981] and else-
where; those for IGRF 1980 are listed in Table 4-1. These 4.7 GEOMAGNETIC ACTIVITY INDICES
and other field models can be provided by World Data
Center A for Rockets and Satellites (Section 4.8.1) in the The magnetograms and averaged data supplied by mag-
form of coefficients and computer programs to generate field netic observatories are too detailed for many purposes, and
values. it has long been found useful to compute some numerical
parameters that indicate the level of general magnetic ac-
tivity or of specific types of disturbance. The International
4.6.2 The External Field Association of Geomagnetism and Aeronomy (IAGA) has
formally adopted 19 of more than 30 such indices currently
The external field is that portion of the geomagnetic field in use. A brief description is given here of several of the
that results from currents flowing above the surface of the most useful indices; a complete review has been given by
earth. Even at the surface and at low altitudes, where the Mayaud [1980]. The availability and sources of indices are
contribution of the external field is a very small percentage, discussed in Section 4.8.5.
measurements are now sufficiently precise that it is included
in the spherical-harmonic models. As the strength of the
internal field weakens with increasing distance from the 4.7.1 General Activity Indices K, Ks, Kp,
earth, the external field becomes relative] y more important. and Km
However, even with external-field coefficients, the spheri-
cal-harmonic models described above lose their usefulness The K index, a measure of the irregular variations of
beyond about two or three earth radii. The reason is that standard magnetograms, is an indicator of the general level
the geometry of the external field does not lend itself to of disturbance at a given observatory. The index is defined

4-27
CHAPTER 4
T.hle
.-w . .
4.1
. -r......-.
Snherical.harmonic coefficients of the IGRF 1980.0.

n m g h dgldt dhldt n m g h dgldt dhldt

10 – 29988 22.4 80 20 0.8

11 – 1957 5606 11.3 – 15.9 81 7 7 –0.2 –0.1
82 1 –18 –0.3 –0.7
20 – 1997 – 18.3 83 –11 4 0.3 0.0
21 3028 –2129 3.2 – 12.7 84 –7 – 22 –0.8 –0.8
22 1662 – 199 7.0 –25.2 85 4 9 –0.2 0.2
86 3 16 0.7 0.2
30 1279 0.0 87 7 –13 –0.3 –l. I
31 –2181 – 335 –6.5 0.2 88 –1 –15 1.2 0.8
32 1251 271 –0.7 2.7
33 833 – 252 1.0 –7.9 90 6
91 11 –21
40 938 –1.4 92 2 16
41 783 212 –1.4 4.6 93 –12 9
42 398 – 257 –8.2 1.6 94 9 –5
43 –419 53 –1.8 2.9 9 5 –3 –7
44 199 – 298 –5.0 0.4 96 –1 9
97 7 10
50 –219 1.5 98 1 –6
51 357 46 0.4 1.8 99 –5 2
52 261 149 –0.8 –0.4
53 – 74 – 150 –3.3 0.0 10 0 –3
54 – 162 – 78 0.2 1.3 10 1 –4 I
55 – 48 92 1.4 2.1 10 2 2 1
10 3 –5 2
60 49 0.4 10 4 –2 5
61 65 – 15 0.0 –0.5 10 5 5 –4
62 42 93 3.4 –1.4 10 6 3 –1
63 – 192 71 0.8 0.0 10 7 1 –2
64 4 – 43 0.8 –1.6 10 8 2 4
65 14 –2 0.3 0.5 10 9 3 –1
6’6 – 108 17 –0.1 0.0 10 10 0 –6

70 70 –1.0
71 – 59 – 83 –0.8 –0.4
72 2 – 28 0.4 0.4
73 20 –5 0.5 0.2
74 – 13 16 1.6 1.4
75’1 18 0.1 –0.5
76 11 – 23 0.1 ‘ –0.1
77 –2 – 10 0.0 1.1

for each 3-hour interval on the basis of the largest value R ginia, has sometimes been used as a standard measure of
of the 3-hour ranges in X, Y, D, or H, where the range is activity for the United States. K values are integers ranging
the difference between the highest and lowest deviations from O through 9; the calibration for six representative sta-
from the regular daily variation. The K value for a given tions is given in Table 4-2; it may be noted that the K scale
value of R is found from a table in which the location of is roughly logarithmic with range.
the station has been taken into account to permit comparison The Kp index (“p” for “planetary”) is probabl y the most
between K values from different stations; the index reflects widely used of all the indices. It was intended as a measure
mainly auroral-zone activity, so stations nearer the auroral of the worldwide average level of activity; however, it is
zone are much more sensitive. A two-letter subscript iden- very sensitive to certain auroral-zone activity and insensitive
tifies the station. The KFR index, from Fredericksburg, Vir- to some other types of disturbance. It is based on the K

4-28
 THE GEOMAGNETIC FIELD
Table 4-2. Definition of the K scale for six represen~ative observatories

Name of Geomagnetic Lower Limit of R in nT for the K Value to be

Observatory Lat. Long. O 1 2 3 4 5 6 7 8 9

Godhavn 79.9 32.5 0 15 30 60 120 210 360 600 1000 1500

Sitka 60.0 275.3 0 10 20 40 80 140 240 400 660 1000
Huancayo –0.6 353.8 0 6 12 24 48 85 145 240 400 600
Fredericksburg 49.6 349.8 0 5 10 20 40 70 120 200 330 500
Tucson 40.4 312.2 0 4 8 16 30 50 85 140 230 350
Honolulu 21.1 266.5 0 3 6 12 24 40 70 120 200 300

Table 4-3. Value of theapindex foragivenvahreof Kp

IfKp= Oo O+ 1– 10 1+ 2– 20 2+ 3– 30 3+ 4– 40 4+
then ap= 023 4567 9 12 15 18 22 27 32

1fKp= 5– 50 5+ 6– 60 6+ 7– 70 7+ 8– 80 8+ 9– 90

then ap= 39 49 56 67 80 94 111 132 154 179 207 236 300 400

indices from 12 stations between geomagnetic latitudes48° Ap is then defined to be the average of the eight 3-hour ap
and 63°, selected for good longitude coverage. Values of indices for the day.
K are first used to find the Ks index (’<s”for “standardized”) In the same way, the single-station indices ak and Ak
from tables which remove the characteristic seasonal be- are found from K by use of Table 4-4. In this form, these
havior at the station. The Ks index ranges continuously from indices are normalized for station location (because K is),
0.0 to 9.0 but is quoted in thirds of an integer by use of
the symbols –, o, and +, such that the interval, for ex-
Table 4-4, Value of ak for a given value of K.
ample, from 3.5 to 4.5 includes the Ks values 4 – , 40, and
4+. Ks can therefore assume the 28 values 00, O+, 1 –, If K=O12 3 4 5 6 7 8 9
10,1 +,..., 9 –, 90. The Kp value for each 3-hour in- then ak= O 3 7 15 27 48 80 140 240 400
terval is the average of Ks from the 12 stations. The Kp
index is published in two forms: numerical tables and the
“musical-note” diagram reproduced in Figure 4-26; in the and comparison between stations is convenient. To remove
diagram each successive solar rotation is plotted under the the normalization, these indices are sometimes multiplied
previous one to expose 27-day recurrences of activity. by a calibration factor f, given by f = RJ250, where Rg
A newer index Km (“m” for “mondial”, meaning global) is the lower limit of R for K = 9. Indices so adjusted are
has recently been developed as an improvement on Kp. It quoted in units of nanoteslas. Thus at Fredericksburg,
is an average of two component indices Kn and Ks (not to f = (500 nT)/250 = 2 nT. As an example of the complete
be confused with the standardized K index Ks) derived from circle, when R = 27 nT, K = 3, and ak = 15 (dimen-
northern- and southern-hemisphere stations, respectively. sionless) cr ak = 30 nT.
Some of its advantage lies in a simpler and more direct The am index is similarly related to Km (there are also
conversion of the data, but the major improvement comes indices an and as corresponding to Kn and Ks), but in this
from a much better geographical distribution of stations. case the a-type indices are derived directly from the data
and the corresponding K-type index follows; that is, the
awkward conversion-reconversion to and from a logarithmic
4.7.2 Related Indices ap, Ap, ak, Ak, am, scale is avoided. The am index is probably, in many re-
and aa spects, the best worldwide index currently available.
The aa index is similar to the am index but is derived
Because the 3-hour K-type indices are defined with a from only two observatories situated at approximately an-
roughly logarithmic scale, they are not suitable for simple tipodal locations in England and Australia. Although its
averaging to obtain a daily index (though this is not un- development in the early 1970s was motivated by the avail-
commonly done). To convert to a roughly linear scale (that ability of records from two old observatories (Greenwich
is, reconversion to an equivalent range), the ap index is and Melbourne) and has resulted in an index which spans
defined from Kp by use of Table 4-3, and the daily index 115 years (1868 to the present), it has also proved to be an

4-29
 DAVS IN SOLAR ROTATION INTERVAL
Ti ‘- Z]114-T5i6 ‘ ; ] 8-r9~ 70 1 lf r~~7;1-)4T75 i 16~)8~9T20T211Z2 123”T2412ST%T~

1983
ROT

2045

●
*W”
%
?? * .+”””+,?a.:.fl+ -X* \.’
*** **, ● ●
2.,&~- “ + *p~
,4 % ~ ---- “+” %“—
20 25 A 30-1 8 ‘My
Apr h 12 A 15
F

m48

2049 9

d
Ju/ 2 7

2051

2053-

2m4

2ms

~6

30’31’7’2 m

PLANETARY MAGNETIC
THREE - HOUR-RANGE INDICES
Kp 1983
~+ :,:+ :+ :+ :+ :+ :+ :+ ;
(preliminary IMfces to 1984 Anuafy 18 )

Figure 4-26. Musical-note diagram of Kp values for the year 1983, produced by the International Service of Geomagnetic Indices
 THE GEOMAGNETIC FIELD
Table 4-5. Scale for finding Cp from the daily sum of ap. For each Cp interval the value listed is the upper limit of the daily-sum values in the interval.

If sum< 22 34 44 55 66 78 90 104 120 139 164 I 90 228

then Cp= 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2

If sum< 273 320 379 453 561 729 1119 1399 1699 1999 2399 3199 —
then Cp= 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0 2.1 2.2 2.3 2.4 2.5

easily obtainable but very good approximation to the am ically from a single combined plot of the deviations at all
index. stations of the element H from lts normal quiet-time value,
AE is the height of the envelope (that is, the difference)
between the curves AU and AL (for “upper” and “lower”)
4.7.3 Character Indices Cp, Ci, and C9 drawn through the maximum and minimum excursions of
the deviations, respectively. The values of AU and AL are
The Cp index, called the daily planetary character figure, also used as indices, and another index Ao is defined to be
is a number that ranges continuously from 0.0 to 2.5 and the mean deviation, the curve midway between AU and AL.
is quoted to tenths of an integer. It is found, using Table
4-5, from the daily sum (the sum of the eight 3-hour values)
of the ap index. It restores the quasilogarithmic relationship
originally introduced by the Kp scale and later removed by 4.7.5 Ring-Current Index Dst
the ap scale. The peculiarity y of its scale results from the
intention that it should replace and be equivalent to a val- The most widely used and available index of low-latitude
uable older index, the Ci index, called the international daily activity is the hourly Dst index. It is the magnitude of the
character figure, available for all days since 1884. Cp is normalized horizontal component of the Dst field (axially
more reliable and objective, but usually differs from Ci by symmetric disturbance), as determined from the data ob-
less than 0.2. Ci is obtained for each Greenwich day as the tained by four low-latitude observatories distributed in lon-
arithmetic mean of the C index values from a number of gitude. The Dst index was designed as a measure of the
observatories; at each observatory the C index, called the magnetospheric ring current of magnetic storms; therefore,
daily magnetic character index, is reported as O, 1, or 2 high-latitude and equatorial stations are avoided to minimize
according to whether the UT day is quiet, moderately dis- the effects of auroral and equatorial electrojets. The deri-
turbed, or highly disturbed, respectively. vation was described by Sugiura [ 1964], who since the IGY
The C9 index, which can be found from Cp or Ci by has been the principal source of this index.
use of Table 4-6, is sometimes used in their place especially
for graphing, since it uses the more convenient single-digit
zero-to-nine scale. Figure 4-27 shows a standard plot of the 4.7.6 Other Indices
C9 index that, like the Kp musical-note diagram, shows
recurrent 27-day activity. The Q index is an index which provides better time
resolution than those based on the 3-hour K indices, but its
availability is limited. It is a measure of high-latitude mag-
4.7.4 Auroral-Electrojet Indices AE, AU, netic activity for each 15-minute interval, assigned by a
AL, and Ao number of observatories above 58 degrees geomagnetic lat-
itude. It is computed, from the more disturbed of the ele-
The AE index called the auroral electrojet activity index ments X and Y, as the sum S of the absolute values of the
is derived from magnetograms from a number of auroral- maximum positive and maximum negative deviations from
zone stations distributed in longitude. It is intended to meas- the normal (quiet) curve, with the provision that if the “neg-
ure the strength of the auroral electrojet tabulated for hourly ative” (“positive”) deviation does not indeed become neg-
and (to some extent) 2.5-minute intervals. Obtained graph- ative (positive), it is considered to be zero. The conversion

Table 4-6. Scale for finding C9 from Cp or Ci,

If Cp or Ci = 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.5 1.9 2.0
or or thru thru thru
0.1 OO; 0: 0: 0: 1.1 1.4 1.8 2.5

then C9 = o 1 2 3 4 5 6 7 8 “9

4-31
CHAPTER 4

a ~’–
026 3 ,3 L4. L. ! .5. .~.....i...z~~
scale for obtaining Q from S, applicable to all stations, is
given in Table 4-7.
The indices U, u, u,, and A are all derived from the
J22 2,2. I.4L.545L,675L 5.33 .33 “$:3 difference between consecutive daily means of some ele-
F18 t 1565366’ 37353.. 36754 55> ~> ment of the field. The U figure is the ‘absolute magnitude
M17 552 22 ..664354653L 1.3,3224 $42 45L of the difference between the daily means of the horizontal
A 13u, L54m65577L5.4 .. 36 OL 357
component for the given and preceding days. The u figure
M 1004357747573 (5562’3 .224 L7L
is a reduction of U to an equivalent equatorial value by the
J6 474 $ . 35, ,32, (83 L4326445 )>3542
/3 ,,3541 .35z, .>62Z3&3~6J “235453 relation u = U/(sin ~~ cos $), where f3~ is geomagnetic
13 ,35 L53h L,. >44z>13165>>L37& 265544 colatitude and $ is the angle between H and the magnetic
AX ,65544 513L~ 33~5333z ~56>J L.2365 meridian. The u, figure is obtained from u by a nonlinear
Sn 4.2365 (LL)56L))65d347525t 576765 conversion intended to reduce the influence of large distur-
019 5767655235332 2!3 2564666 5L6&4
bances. These indices have been useful historically for monthly
NE 5L66L4L 35z6, ... 111L3i L3i “53 .2
and annual changes, but since they were intended to measure
012 53$.2L2 .$ 32, ,3665> z53 33 2 .. ..
7U65L
ring-current effects, the Dst index has largely superseded
J8 2.. .3434. ‘L655223L 36776
FL 76665 L477775 ,U657U771 3Z’L34 .3 them.
M3 434 3543233 ,,L53356z351z3 L55775 The W measure applies to quiet variation fields rather
MW ~55n555$ 337743,1353,565247,6565 than disturbance phenomena, being derived from the am-
AZ ,6566505 L).,, , ,5235322. ..zhn plitude of the Sq variation. It is a measure of solar radiation
M23 ..z6nU 55) 4Lz47~763, I 51555
effects and the intensity of the equatorial electrojet [Bartels,
J19 652555L4665653 ,..L543,OW 7663
W,,sa J16 765 U3 4375U 555 637&5 h$r 5U ,53t .$L
1946] .
36
765dm 37 A 12 53 ~ .Ld234645553366 L2356a 373444
Mm= 38 S8 373444 L333G06537L 236623 2565$4
5565L6W6 39 05 256544 i16624554b3z ,$L6L3U 665 ZZ3
ti6m 775 2040 N1 U5123.323545332 (33$6M7Q ‘ti3 ... 4.8 SOURCES OF GEOMAGNETIC DATA
53 m 756 41 ND U3 ..2 33.66575L ,23477067656,5554
765576U lg 0254,5554,],],( ..,57J5,z5&6LL 322543 The following paragraphs provide some specific infor-
556765322 J21 ,,,5L33,L 533,,7 W5555UU 4437a mation on the availabilityy of geomagnetic data. In recent
23.t 66542 E F17 LL37654L ,, 576632)2 b7u’ 335G3
years, the increasing emphasis on global studies of mag-
5655LL L45 2045 M16 335a322373 .#65LL44u 552 30765
546a7u 46 A 12 3677653,334 L7~L366541& 413 27765
netospheric phenomena and the correlation of satellite and
m677655 4? M9 ,277655,7 ]36767],3 )33, . ,3(356 ground-based measurements have encouraged the extensive
G5 & 765 he 15 ,3,15634734,4654554 ),3233 222 LL collection of such data at several data centers where they
55556764 u J2 .,: L431 45 565. 56 3’L52 ,64 are made readily available. For data that are not archived
055U5LL 205L 129 L5, 6L .573 )66,3 ,36465 ~ zL5 there, the centers can often provide a referral to an alter-
LLL555333 51 d 25 U,, L56L, ). ,5L43332$5U 375 33.26 native source.
ILL3’56775 52 S21 33.26L 31.35572534 ,$ .65537 72455
32, ,23553 53 018 7,( 6555.. 265>565..30605 65665
33t .2235 54 N% 656656,, .345,4554,.1463.5 655bL
5L3,),)2 20ss on h5564 ,>.33L34L 32u655&:t. .43.
J7 ,,.43$,89, 5233 242556545 574 222
4.8.1 Data Centers
F3 574,,, ,55305, L8LL ,41,4623 W2.6
Ml &2.6u3L23s 2L553 .2553645 7766 Two national data centers in the United States have
6 12059M28 7766~777L~63L L55 3 extensive holdings of geomagnetic data. The National Geo-
[2060!Z I preliminary i physical Data Center (NGDC) is operated by NOAA (En-
vironmental Data and Information Service, NOAA, D63,
S;,mboi 4 56 7:[ 325 Broadway, Boulder, CO 80303, phone 303-497-6323).
F= o 1-75 163011-45 L6-60 618081-100101130131-170 171 The NGDC is a recent successor to the National Geophysical
R9,C9=0 1 2 3 ~ 5 6 7 H g and Solar-Terrestrial Data Center (NGSDC) and also serves
Cp = 00-010.2-030L05 060708-0910-11 ?Z-ld 15-18 19 20-25 internationally as the World Data Center A for Solar-Ter-
restrial Physics (WDC-A for STP). The National Space
DAILY GEOMAGNETIC Science Data Center (NSSDC) is operated by NASA (God-
CHARACTER FIGURES C9 AND dard Space Flight Center, Code 601, Greenbelt, MD 20771,
3-DAY MEAN SUNSPOTNUMBERS R9 phone 301 -344-6695) and also serves as the World Data
(after Bartels ) Center A for Rockets and Satellites. NGDC has magneto-
grams (as plots and computer tapes) and magnetic indices
Figure 4-27. Plot of the C9 index and sunspot number R9 for the period
(in tabular, computer tape, and chart form). Both centers
1981-1984. have space measurements (mostly nonduplicative) of mag

4-32
 THE GEOMAGNETIC FIELD
Table 4-7. Scale for finding Q from S, The listed value of S (in nT) is the upper limit for the corresponding value of Q

IfS< 10 20 40 80 140 240 400 660 1000 1500 2000 —

then Q = o 1 234 5 6 7 8 9 10 11

netospheric magnetic fields, electric fields, and ULF/VLF nates) of Table 11of the WDC-A catalog. Stations are listed
phenomena. Both centers issue catalogs of their holdings in order of increasing uncorrected geomagnetic colatitude.
and publications [Allen et al., 1982; National Space Science
Data Center, 1981 and 1982].
Subcenters of the World Data Centers are also relevant 4.8.4 Magnetic Field Models
to geomagnetism. They are organized into WDC-A (sub-
centers in the United States, including the two mentioned A number of magnetic-field models and computer pro-
above), WDC-B (subcenters in the USSR, including one grams for their use are now available from the National
for solar-terrestrial physics at IZMIRAN, Moscow), and Space Science Data Center [see NSSDC, 1981, p. 55]. For
WDC-C (twenty subcenters in Europe, Japan, and India). many uses where the highest precision is not required hard-
WDC-A for STP now includes in its catalog the geomagnetic copy charts and tables may be more convenient. The IGRF
holdings of WDC-B2 (Moscow) and WDC-C2 (Kyoto). 1975.0 field is available in this form. Barraclough and Fa-
biano [ 1977] present 14 charts (elements X, Y, Z, H, D,
I, F and the secular variation in terms of their first time
4.8.2 International Organizations derivatives) and tables which list values for every five de-
grees of longitude and every two degrees of latitude. Pub-
The International Council of Scientific Unions (ICSU) lications resulting from the most current work with MAG-
includes 18 unions, which are nongovernmental organiza- SAT and other data can be expected from cooperative efforts
tions dedicated to scientific progress in various disciplines. at both NASA and the United States Geological Survey
Geomagnetism is a concern of the International Union of ILangel, 1980].
Geodesy and Geophysics (IUGG), which in the past has
initiated such international collaborations as the Intern-
ational Geophysical Year and the International Magneto- 4.8.5 Activity Indices and Charts
spheric Study. Within the IUGG a number of associations
have been formed; geomagnetism falls within the concerns At present, an excellent collection of indices has been
of the International Association of Geomagnetism and assembled at the NGDC (Section 4.8.1). The widely used
Aeronomy (IAGA). This association was originally the Sec- tables and charts furnished by the ISGI, as well as the
tion de Magnetism et Electricity Terrestres (19 19) and later publications of the IAGA, may be obtained through the
the International Association of Terrestrial Magnetism and IUGG Publications Office (see Section 4.8.2). A variety of
Electricity (1930) before taking its present name in 1954. sources current through 197 I was listed by Knecht [ 1972].
The IAGA has historically been a principal source of The periods for which various indices have been com-
publications of geomagnetic data, including its Bulletins 12 puted and published vary greatly, often depending on the
series (covering 1940 through 1969) and Bulletins 32 series resources available to the workers who have the capability.
(covering 1970 through the present) of geomagnetic data, Even some of the currently most useful indices, such as AE
indices, rapid variations, and special intervals. The Inter- and Dst, have been delayed. An enquiry as to the current
national Service of Geomagnetic Indices (ISGI) operates coverage of any index (beyond that listed in the catalog)
under supervision of the IAGA. Its services include the may be directed to the NGDC.
regular distribution of the Kp and C9 tables and charts from
its Gottingen office (Institut fur Geophysik, 34 Gottingen,
W. Germany). Publications of both the IAGA and the ISGI ACKNOWLEDGMENTS
are available from the IUGG Publications Office (39 ter rue
Gay-Lussac, 75005 Paris, France). A catalog is available. The section on magnetic pulsations was written by H.J.
Singer. Reference materials, recent data, and/or helpful
comment were generously provided by R. H. Lange], J.C.
4.8.3 Magnetic Stations Cain, and N.W. Peddie on current magnetic-field models;
S. Matsushita on recent work on the Sq variation; T. Saito
Nearly 400 magnetic stations, data from which are held on the classification of pulsations; and J. H. Allen on the
in the World Data Centers, are listed in Table 4-8, which activities and holdings of the WDC-A for STP. The authors
is a condensation (omitting two sets of corrected coordi- are very grateful for these contributions.

4-33
CHAPTER 4
Table 4-8. List of geomagnetic stations from which the WM-A has data holdings, listed in order of increasing geomagnetic colatitude. Successive columns
are observatory name, international three-letter code, and uncorrected geomagnetic colatitude and east longitude.

T HUI.E THL 1.?3 12. ~6 BIG ~LTA 8nF 25.4S 260.96 SEDDIN s~fs 37.87 943.11
TH!JLE AFB THII ?.16 9.RI NmTWAY M?w 25.62 265.52 GOTTINGEN GTT 37.99 94.82
ALFRT ALF 4.02 165.17 MAI I YE KARMAKULY MKL 75.62 143. R2 N IEM[ (’K NGK 38.01 91.67
RE$ 6.82 2Q2. R7 !4Ei V I LLE AFB t4!’ v 25.63 13.72 DOURBE S CKJu 38.30 88.88
R[$OLUTE HAY
NWTH ~LE 7 fiPG 6.ql 19’,,17 KOTZEHUE I(OT 25.99 243.55 MANHAY mB 38. >2 ~.04

9.24 l~z.aq I $LANO IAKE 1s1. 26.16 326.66 CASPER C SR 38.42 516.25
Nu?o
10.45 34.53 }lEA~. Y HEA 26.19 25 f3.17 M I NSK Ma 38.70 111.42
GOF)HAVN mH
S~ANKYLA Sm 26.3q 120.77 UGUT UGT 38.72 153.19
MOUL17 RAY MRc 10.59 258.54
11.17 322.55 E4ELIY ISLAND BE Y 26.40 155.68 COLLMBERG a. L 38.78 97.61
PELLY BAY mr4
GZH 11. Q2 183.95 NCRWAY HOUSE NHO 26.42 322.51 YAKIJTSK YAK 38.79 194.94
GUANGZHOU

12.13 311.67 ML47MANSK M 26.65 126.55 VAL JOYEUX VL J 39.01 85.70

GJOAHAVN GJO
12.93 272.79 CAPE Kt4410T CPS 26.9q 228.72 MAGAOAN NGD 39.02 211.57
JOHNSON ~lNT JOP
IJ.18 2q7.31 D I XOtl ISLAND DIK 26. qo 162.32 UKHTA UKN 39.21 153.04
CAV[V7 I Da BAY cBt3
AIF 13.64 20FJ.33 LOVf)ZERO LOZ 27.21 127.74 P . TUNGUSKA POD 39.23 165.40
ARC TII; ICE FLO
S UKKFR T@PEN SKT 14.17 30.74 KANDALAKSNA KND 27.56 125.02 MOSCOW Ws 39.28 121.51

M~CH I SON BAY MUR 14.57 139.00 LERW ICK LER 27.75 89.64 SW IUR ‘WI 39.61 105.69
SCCRESBYSUtJD SC() 14.60 82.63 TALKEETNA TLK 27.91 258. ~0 BELSK EtEL 39.80 105.13
SACHS MRRO~ SAH 14.60 267.47 CAPE WELLEN CWE 27.87 23 f4.4R NANTES NTs 39.82 81.39
NEM AALESUND N4L 14.69 1.31.33 D0!E4AS M13 Z7.99 Iol .05 CHAMBON-LA-F~ET CLF 39.83 85.58
I(AP TC81N KTG 15.26 274.22 MEAN~ MEA 28.03 302.97 ANEJREWS AFB AWS 39.95 352.15

SPITZ14ERGEN SFJB 15.44 131.69 KHARASAVEY KHS 2R.21 152. >2 CHELTENHAM CLH 40.06 352.24
CAW PARRY CPY 16.00 271.96 TM9EY TMfl 28.28 155.77 PRUHON ICE mu 40.36 90.43
BAKER LAKE ULC 16.17 317.73 A~t40f4AGE AMu 2R.78 259.73 CARROLLTON CAX 40.39 332.14
CHESTERFIELD INL CFI 16.49 326.72 AMDER?4A AMD ?9.77 147.69 FREDERICKSBURG FRO 40.58 351.58
N~TH ~LE 6 NPF 17.17 174.54 TIXIE BAY TIK 29.33 192.38 BELOIT BLT 40.75 326.63

RAM(IN INLCT t41T 17.19 324.72 SEYKHA SEY 29.69 155.52 KA7AN K/N 40.76 131.37
NOITti ~L[ 12 NF’L 17.50 21 R.~8 S I TKA SIT 2rA.75 277.08 FIOULMR nou 40. % 318.24
N~Ttl ~LE 1S N~.4 18.00 204.63 WHITE SHELL WHS 30.29 327.44 B\lDKOV BDV 41.12 97.51
T IKHAYA BAY TKH 18.48 153.60 WINNIPEG WNP 30.61 325.12 BALDWIN f3AL 41.14 330.25
HEISS ISLAND 1{1s 18.64 156.47 GLENLEA GLL 30.61 325.16 BURL I NGTON BRT 41.41 321.97

N~TH POLE 8 NP}i 18.68 22>.65 ARKHANGELSK ARK 31.16 129.25 FUR STENFELD8RUCK FUR 41.45 94.40
IV IGTUT Ivl 18.78 34.54 CAPE KA!4ENN I Y CKA 31.34 155.54 ART I ARs 41.52 139.57
RFAR ISLAND P JN IQ. f)O 124.95 NCR ILSK NW 31.40 165.68 SVERDLOVSK SVD 41.54 141.70
ESKIMO mlNT FKP 19.03 325.51 LC47 ING AFB L~ 31.77 2.97 CLIMAX CLl 41.83 317.35
NARS5ARSSUAQ NAo 19.22 58.70 ST JOHNS STJ 31.77 22.95 LEAOV I LLE LDV 41.93 317.26

I NUVl K INK 19.29 267.10 EM(OP I NG ENK 31.77 106.22 CHATONNAYE CTE 42.01 89.50
J U(. I ANEWAR JIJL 19.4q 57.24 E SKDALEMU I R ESK 31.83 84.05 TEVRIZ TEV 42.02 150.97
BARTER ISLAND RT I 19.71 254.77 L OVO LOV 32.15 106.77 L VOV LVV 42.19 106.96
LFIRVOG~ LR V 20.10 72.19 N’*MIJARVI NUR 32.32 113.53 PR ICE Pcu 42.23 312.07
NCRMAN WELLS Now 20.67 279.83 NYOA NDA 33.14 154.34 W I EN-KOBENZL WIK 42.33 99.34

rml RAE FRA 20.81 292.95 OTTAWA OTT 33.34 353.43 KIEV KIV 42.62 113.25
YELL~KNIFE YKC ?0.94 295.38 STONYHURST STO 33.36 83.87 NAGYCENK NCK 43.02 99.44
B ARROW Ww 21.12 242.60 VALENT IA VAL 33.68 74.70 ADAK ADA 43.04 241.60
F~T cH~Cttl LL FCC 21.28 325.00 LEN ING?4AD LNN 33.92 118.29 HIJRBANOVO NRE4 43.08 100.92
NERTH POLE 10 NP J 21.37 211.09 SLUTZK SLU 34.15 117.94 O GYALLA KY 43.08 100.92

ARCTIC VILLAGE AVI 21.77 256.17 RUFAE SKOV RSV 34.3q 99.58 BUOAKE SZ I BUZ 43.55 101.45
BACK BKC 22.59 325.46 COPE NHAC+N COP 34.58 99.60 T I HANY THY 43.95 100.25
f)ooc BOEJ 22.56 113.41 NCWF’ORT NEW 34.78 301 .B5 TC44SK TMK 44.04 160.64
FCRT S I!4PSON FSP 22.71 2t38.15 AG I NCOURT AGN 35.07 348.84 TULSA 44.06 330.06
CAPE ZHELAN 1A CZA 22.77 157.59 VI CTM IA Vlc 35.63 294.73 L~ONO ?& 44.25 78.46

TRCt4S0 TRo 23.04 117.42 HARTLANO ‘ MD 35.69 80.20 MONTE CHLL I NO MCP 44.46 90.7)
F~ T ‘YUKON F YU 23.07 258.45 WI NGST WNG 35.71 95.15 GENOVA EN 44.46 90.71
G I I. LAM Glf4 23.22 >25.64 W I LHF LMSHAVEN WLH 35.75 94.14 R@ URENT R~ 44.49 89.56
BOSSEKOP E40P 23.52 121.06 NU\fTO NMT 36.03 152.03 CASTELLACCIO CAO 44.57 90.64
GREAT WHALE R. Wc 23.54 349.46 GREENWICH mw 36.08 85.02 SAO MIGUEL 44.70 52.30

CA~ CHELYUSK IN Ccs 23.57 177.18 W ITTEVEEN WIT 36.12 92.32 PE TR@AVLOV SK PET 45.02 219.87
LYNN IAKE L YN 24.04 316.80 AB I NGER ABN 36.29 84.47 POLA POL 45.11 95.59
AR I SKO ABK 24.17 115.70 WESTON WES 36.30 358.86 CO I M8RA ml 45.29 71.55
THOtPSON TMP 24.73 321.60 DE 81LT E8N 36.51 90.48 E SPANOLA EPN 45.32 318.33
K IRUNA KIR 24.91 116.38 SREDN IKAN SRE 36.63 211.76 NOVOS 18 IRSK NV S 45.33 158.78

FWT SMITH F SM 24.92 299.87 HEL HLP 36.79 ~4.74 CASTLE ROCK CRC 46.36 300.28
cmLEC+ CMO 25.06 258.14 KAL I N I NGRAD KNG 37.04 D6.43 EBRO 46.41 80.91
E IFLSON AFB Em 25.09 259.03 Bmw Box 37.12 24.15 TOLEDO 46.44 75.95
PE TSAJ40 PTS 25.21 126.47 80CHUM 8W 37.50 92.15 ODE SSA 46.51 112.23
I VALO I VA 25.43 122.68 POTSDAM POT 37.79 98.21 GROCKA 46.69 102.08

4-34
 THE GEOMAGNETIC FIELD
Table 4-8. (Continued

DALLAS DAL 47.05 329.44 F UQUE NE F UQ ~>.~~ >56.55 WG 108.28 71 Q.!U

1. AQLI I LA AQU 47.41 94.14 PARANARIBO PA8 73.27 15.79 TSWF:8 TSU IOR.46 84.24
SIJRLARI S(JA 41.68 107.22 FREETWN FTN 75.54 59.15 P I LAR PIL 110.42 5.93
LA MAPDALENA LMD 47.75 8Q.92 L UNP I NG LW 76.11 190.80 TANANARI VE TAN 113. q9 113.96
MT WILSON Mwc 48.64 305.48 UJJAIN UJ J 76.51 148.22 LAS ACACIAS LAS 113.93 11.39

s IMFERNOL ?Iv 48.97 114.45 ZARIA zAy 76.73 80.47 MAURITIUS MRl 116.75 123.98
IRKIJTSK IQT 49.01 175.69 KONT\~A KNT 77.05 78.15 PLA I SANCF ns 117.11 124.04
SAN FEK{ANOO SF s 49.34 72.64 wAKE ISLAND WKE 77.18 234.42 klwUTO LMN 1!7.92 97.31
CAPRI CP I 49.35 94.38 C}IA PA CPA 78.92 174.57 TRELEW TRW 121.97 4.37
PANAGY~ I SHTE PAG 49.37 104.54 HONG KONG WC 79.09 1B4.35 CAPE ~ HWE Cw 127.97 BI .41

TUCSON TUC 49.49 31 J.Q4 KUMASI KLJM 79.35 70.46 Cm TNN CTO 122.99 81.39
KARAGANOA KGD 49.66 149.76 I BADAN I RD 79.66 75.03 HF~ANUS HER 123.59 81.99
ALMLRIA ALM 49.60 16.50 TATUOCA Tm 80.70 22.19 BRISBANE Ws 125.44 228.32
I S TANBUL-KAND I LL I SK 51.70 108.63 AL IBAG ABG 80.76 144.58 WATHEROO wAT 131.54 187.28
KHABAJ70VSK KHB 51.92 201.24 HYDERA13AD HYD 82.35 150.24 MAGALLANES MGS 131.88 359.61

AVERROES A,VE 52.20 70.45 TALARA TAL 83.51 349.08 GNANGARA 133.02 187.46
NOVOKAZAL I NSK NK K 52.41 139.55 Eus Fnlo EIJS 84.27 32.09 CANBERRA 133.56 225.74
YuZHNO SMHAL I NS YSS 5z.92 207.92 PALMYRA ISLAND PA I 84.56 267.21 SOUTH GE~GIA 134.51 27.19
ULAN BATCR UBA 53.46 17?.63 MIXA K!FP 84.58 79.94 TOOiANG I 136.34 222.32
TBILISI Fs 53.47 123.24 BAGUIO BAG 84.70 190.60 AN8ERLEY 137.34 253.80

PENOEL I PEG 53.64 102.81 ADD I S A8AJ3A ME 84.82 110.51 EYREWELL EYR 137.67 253.52
KZ 11 -AGACH K7A 54.59 153.76 8ANGU I BNG B5.46 B9.8”7 Cm I STCH~CH CHR 137.72 25>,84
CANARIAS TEN ‘)5.34 59.99 CH 1CLA% CCL 85.63 350.65 MAR 10N I SLANO ~N 139.24 95.90
SANTA CRUZ SI r 55.34 59.99 GUN4 GUA 85.73 214.26 LAUDER LAU 139.72 251.36
MF.M~BETSU MMB 55.70 209.70 FANNi M FAN 85.97 270.24 ORCADAS EL SIR mc 140.25 19.31

HAVANA HVN 55.97 346.94 ANT I POLO ANo B6.50 191.22 PORT ALFRED CZT 141.51 110.57
ALMA ATA MA 56.53 151.88 MAJUIU MAN 86.53 191.04 GOWALES V IOELA GVO 143.57 5.41
VLAT)IVOSTW VLA 56.90 199.36 MUNT I NLUPA MUT B6.73 191.09 ARff NTINE ISLANO AlA 143.93 4.35
TASHKENT TKT 57.68 145.19 BANGKOK BKK 87.46 171.23 CM~ELL ISLAND CA I 146.97 254.44
AShl(HABAD ASH 59.56 134.30 CHIMBOTE C?.IB 87.89 351.91 Pm T-AUX-FRANCA I KGL 147.43 129.83

KSARA KSA 60.00 112.96 MAJURO MJR 88.35 241.03 MACQUAR I E ISLAND KICQ 150.75 244.41
TEOLOYUCAN TF O 60.44 328.62 ANNA14ALA I NAGAR ANN 88.61 150.45 HEARD ISLAND Hll 151.48 131.91
VI EOUES VQS 60.58 5.41 KODA I KANAL KOO 89.40 14[!.45 SANAE SNA 153.95 45.40
SW JUAN S JG 60.59 4.65 JARVI S ISLAND JRV N.30 270.45 E IGNTS EGS 153.96 >56.08
TE*AN TE H 60.75 127.73 E TTAYAPURM ETT W.67 148.84 N~WAY STATION NWS 154.11 45.09

MI ZUSAWA MIZ 60.84 207.69 CEBU CCP 90.68 194.08 LAZAREV LZV 155.98 56.69
BEIJING BJI 61.23 185.75 }{UAWAYO HUA %.77 355.19 HALLEY BAY WA 156.06 25.25
ONAGAWA ONW 61.39 208.12 TR I VAN~UN TRD 91.08 147.75 NOVOLAZAREVSKAYA NVL 156.51 54.94
MOD I IM MOD 61.71 111.65 Cuzco Cus 92.21 358.48 RO I 8AUOOU I N RBO 158.33 64.67
NIT SANIM NSM 61.81 111.24 MOGAD I SC 10 MoG 92.90 115.68 SYOWA STAT ION SYO 159.96 79.38

AMATSIA NT 62.04 111.49 KORCU Km 92.95 204.73 MOL~Z HNA YA MOL 159.97 :7.36
HFLWAN NL W 63.02 107.69 BINZA Eil N 93.45 84.91 BYRD STATION BYR 160.62 336.71
M I SALLAT MLT 63.28 107.20 DA VAO OAV 93.81 195.90 MAWSON NW 163.31 105.08
KAKIOKA KAK 63.69 207.25 LW 11?0 LW I 94.01 9B.60 L ITTLE AMERICA LAA 163.95 312.23
TCUYO TC4( 64.21 206.90 YAUCA YAU 94.24 355.88 LEN I N@ ADSKAYA LEN 164.33 261.97

KANOZAN KNZ 64.68 207.19 NAIR@l NA I 94.63 106.57 CM HALLETT HLL 164.47 278.80
TWANRASSET TAM 64.92 80.88 CHARACATO ARE 95.17 35B.97 OUMONT D~V I LLE mv 165.33 232.38
MI YAZU MIU 65.02 202.83 LA PAz LPB 95.19 2.22 MIRNY MIR 167.02 149.66
LANZHOU LZH 65.17 174.87 PENANG ~N 95.68 170.81 OASIS OA S 167.37 163.53
MIDwAY WY 65.59 247.75 LUANDA LUA 97.49 81.96 PLATEAU PTU 167.53 53.79

S IMOSATO Sso 66.49 20?.74 DAR ES SALA~ nR s 100.44 108.00 WILKES WIL 167.57 181.35
HACHIJO HJJ 66.81 207.29 LA QU IACA ILQA 100.82 4.54 CHARCOT Crx 167.94 235.93
HA TIZYO HTY 66.86 207.32 VASSOURAS Vss 102.22 25.21 SWTH ~LE SPA 168.68 0.00
ASO ASO 67.74 199.36 HOLLANDIA HNA 102.25 211.70 SCOTT BASE WA 16B.80 294.35
QUETTA WE 68.40 140097 RIO DE JANEIRO RO J 102.77 25.60 MCMIROO 16B.82 294.31

HONOLULU t{oti 68.64 267.97 SAN JOSE S JB 102.86 23.07 PIONERSKAYA Plo 170.33 149.88
M BOUR MIIO 69.08 56.40 E L I ZARETHV 1LLE ELI 102,94 95.37 VOSTOK Vos 179.41 111.76
COSTA RICA SN J 69.09 345.69 KARAV I A KVA 102.94 95.37
KANOYA KNY 69.20 199.36 TAHITI TAH 105.11 2B4.04
SABHAWALA SAB 69.44 150.95 PAMATA I PPT 105.12 284.09

LIJK I APANG LKP 69.80 190.40 APIA WI 105.75 26! .58

Z IKAWE I ZKW 69.88 190.77 KUYPER KUY 107.34 177.92
SHESMN SSH 70.01 190.55 TANGERANG TNG 107.47 176.91
CHICHIJ IMA CR I 72.57 210.19 BATAVIA BTV 107.49 177.12
JAIPUR JAI 72.63 148.68 EASTER ISLANO EIC 108.18 323.88

4-35
CHA~ER 4
REFERENCES

Alldredge, L.R. and L. Hurwitz, “Radial Dipoles as the Langel, R. A., “Magsat Scientific Investigations” NASA
Sources of the Earth’s Main Magnetic field,”~. Geo- Technical Memorandum 80698, May, 1980.
phys. Res., 69:2631, 1964. Langel, R. A., private communication, }982.
Allen, J .H., C.C. Abston, Z.R. Kharin, and N. Papitashvili, Lanzerotti, L.J. and D.J. Southwood, “Hydromagnetic
International Catalog of Geomagnetic Data, Report UAG- Waves,” in Solar System Plasma Phvsics, 111, 109, ed-
36, World Data Center A for Solar-Terrestrial Physics, ited by E.N. Parker, C.F. Kennel, L.J. Lanzerotti, North
Boulder, CO, 1982. Holland, Amsterdam, 1979.
Axford, w. I., “Tail of the Magnetosphere,” J. Geophys. Matsushita, S., “solar Quiet and Lunar Daily Variation
Res., 70: 1231, 1965. Fields,” in Physics of Geomagnetic Phenomena, edited
Barraclough, D.R. and E. B. Fabiano (eds) Grid Values and by S. Matsushita and W .H. Campbell, Academic Press,
Charts for the IGRF 1975.0. IAGA Bulletin No. 38, New York, 1967.
1977. (Available not from IUGG but as NTIS No. PB- Matsushita, S., “Morphology of Slowly Varying Geomag-
276 63s.) netic External Fields—A Review ,“ Phys. Earth Plan.
Bartels, J., “Geomagnetic Data on Variations of Solar Ra- Int., 10:299, 1975.
diation: Part I—Wave Radiation,” Terr. Magn. Atmos. Matsushita, S. and H. Maeda, “On the Geomagnetic Solar
Elec., 5:181, 1946. Quiet Daily Variation Field During the IGY,” J. Geo-
Bullard, E. C., C. Freedman, H. Gellman, and J. Nixon, phys. Res., 70:2535, 1965a.
“The Westward Drift of the Earth’s Magnetic Field,” Matsushita, S. and H. Maeda, “On the Geomagnetic Lunar
Phil. Trans. Roy. Sot., 243A: 67, 1950. Daily Variation Field,” J. Geophys. Res., 70: 2559.
Chapman, S. and J. Bartels, Geomagnetism, Oxford Uni- 196~b.
-versit y Press, London, 1940. Mayaud, P. N., Derivation, Meaning, and Use of Geomag-
Chen, L. and A. Hasegawa, “A Theory of Long-period netic Indices, (Geophysical Monograph 22) AGU.
Magnetic Pulsations, 1, Steady State Excitation of Field Washington, D. C., ~980.
Line Resonance,” J. Geophys. Res., 79: 1024, 1974. Nagata, T., “Two Main Aspects of Geomagnetic Secular
Ctimmings, W. D., R.J. O’Sullivan, and P.J. Coleman, Jr., Variation-Westward Drift and Nondrifting Compo-
“Standing Alfv&n Waves in the Magnetosphere,” J. nents” in Benedum Earth Magnetism Symposium edited
Geophys. Res., 74:778, 1969. by T. Nagata, University of Pittsburgh Press, 1962.
~ungey, J. W., “Electrodynamics of the Outer Atmos- Nagata, T. and M. Ozima, “Paleomagnetism,” in Physics
phere,” Sci. Rep. 69, Ionos. Res. Lab., Pennsylvania of Geomagnetic Phenomena edited by S. Matsushita and
State University, 1954. (Also partially contained in Dun- W.H. Campbell, Academic Press, New York, 1967.
gey, J. W., “The Structure of the Exosphere, or Adven- National Space Science Data Center, NSSDC Data Listing,
tures in Velocity Space,” Geophysics: The Earth’s En- NSSDC Report 81-11, Goddard Space Flight Center,
vironment, edited by DeWitt, Hieblot and Lebeau, Gordon Greenbelt, Md., September 1981.
and Breach, New York, 1963). National Space Science Data Center, Data Catalog Series
Greenwald, R.A. and A.D.M. Walker, “Energetic of Long for Space Science and Applications Flight Missions,
Period Resonant Hydromagnetic Waves,” Geophys. Res. NSSDC Repor~ 82-2 I (VO1. 1A) and NSSDC Report 82-
Left., 7:745, 1980. 22 (Vol. 2A), Goddard Space Flight Center, Greenbelt,
Gtistafsson, G., “A Revised Corrected Geomagnetic Co- Md., September 1982.
ordinate System,” Ark. Geofis., 5:595, 1970. Nishida, A., Geomagnetic Diagnosis of the Magnetosphere,
Hakura, Y., “Tables and Maps of Geomagnetic Coordinates Springer-Verlag, New York, 1978.
Corrected by the Higher Order Spherical Harmonic Olson, W.P. and K.A. Pfitzer, “A Quantitative Model of
Terms,” Rep. Ionosph. Space Res. Japan, 19121, 1965. the Magnetospheric Magnetic Field ,“ J. Geophys. Res.,
Hughes, W. J., “Pulsation Research During the IMS,” Rev. 79:3739, 1974.
Geophys. Space Phys., 20:641, 1982. . Olson, W .P., and K.A. Pfitzer, “Magnetospheric Magnetic
Hughes, W. J., “Hydromagnetic Waves in the Magneto- Field Modeling,” McDonnell-Douglas Astronautics CO.
sphere”, submitted to U.S. National Report to 18th Gen- Rept, 1977.
eral Assembly of IUGG Hamburg, Aug. 1983. Orr, D., “Magnetic Pulsations Within the Magnetosphere:
IAGA Division 1 Working Group 1, International Geo- A Review,” J. Atmos. Terr. Phys., 35: 1, 1973.
magnetic Reference Fields: DGRF 1965, DGRF 1970, Orr, D., (cd.), “Geomagnetic Pulsations: Papers from 7th
DGRF 1975, and IGRF 1980, Eos Trans AGU 62:1169, Annual E.G.S. Meeting,” J. Atmos. Terr. Phys., 43(9):
1981. 1981.
IAGA, “Review Papers from the ULF Symposium held at Parkinson, W.D. and J. Cleary, “The Eccentric Geomag-
the 4th IAGA Scientific Assembly, Edinburgh, 14 Au- netic Dipole,” Geophys. J. Roy. Astron. Sot., 1:346,
gust 198 1,“ Planet. Space Sci., 30: 1199–1 258, 1982. 1958.
Jacobs, J .A., Geomagnetic Micropulsations, Springer-Ver- Saito, T., “Geomagnetic Nlsations,” Space Sci. Rev., 10:
lag, New York, 1970. 319, 1969.
Knecht, D. J., “The Geomagnetic Field,” AFCRL-72-0570, Saito, T., “Long-period h-regular Magnetic Pulsations, Pi
AD 759837, 1972. 3,” Space Sci. Rev., 21:427, 1978.

4-36
 THE GEOMAGNETIC FIELD
Shawhan, S .D., “Magnetospheric Plasma Wave Research Southwood, D.J. (cd.), “ULF Pulsations in the Magneto-
1975–1978,” Rev. Geophys. Space Phys., 17:705, 1979. sphere,” J. Geomagn. Geoelec., 32, Suppl. II, 1980.
Singer, H. J., “Multisatellite Observations of Resonant Southwood, D.J. and W .J. Hughes, “Theory of Hydrom-
Hydromagnetic Waves,” Planet. Space Sci., 30: 1209, agnetic Waves in the Magnetosphere,” Space Sci. Rev.,
1982. 35:301, 1983.
Singer, H. J., D.J. Southwood, R.J. Walker, and M.G. Kiv- Southwood, D.J. and W .F. Stuart, “Pulsations at the Sub-
elson, “Alfv6n Wave Resonances in a Realistic Mag- storm Onset ,“ Dynamics of the Magnetosphere, edited
netosphenc Magnetic Field Geometry ,“ J. Geophys. Res., by S.-I. Akasofu, D. Reidel, Dordrecht, Holland 1980.
86:4589, 1981. Sugiura, M., “Hourly Values of Equatorial Dst for the IGY,”
Southwood, D. J., “Some Features of Field Line Resonances Annals of the Int. Geophys. Year, 35:49, 1964.
in the Magnetosphere,” Planet. Space Sci., 22: 483, Tsyganenko, H. A., “A Model of the Cis-Lunar Magneto-
1974. spheric Field,” Ann. Geophys., 32: 1, 1976.
Southwood, D.J ., “Waves, Wave Particle Interactions and Vestine, E. H., “Influence of the Earth’s Core upon the Rate
Magnetic Oscillations,” ZAGA Report, International of the Earth’s Rotation ,“ in Benedum Earth Magnetism
Association of Geomagnetism and Aeronomy, Dec. Symposium, edited by T. Nagata, University of Pitts-
1979. burgh Press, Pittsburgh, 1962.

4-37
 Chapter 5

THE RADIATION BELTS

W.N. Spjeldvik and P.L. Rothwell

Interaction of the solar wind flow with the earth’s tor. Figure 5-2 illustrates principal aspects of a charged
magnetic field gives rise to a cavity in the interplanetary particle trajectory in magnetic mirror field geometry. The
medium known as the earth’s magnetosphere. Within this
cavity there exists a limited region where the motion of
energetic particles is confined by the earth’s magnetic
field. This region comprises the earth’s radiation belts as
depicted in Figure 5-1. The radiation belt region contains
electrons, protons, helium, carbon, oxygen, and other

Bi

Figure 5-2. A charged particle trajectory in a magnetic “bottle”. Con-

servation of the first adiabatic invariant can cause the spiral-
ing particle to be reflected where the magnetic field is
stronger. This causes the particle to be trapped by the mag-
netic field.

magnetic force (F ~ V x B) deflects the particle velocity

vector V so that the particle spirals around the magnetic
field B. The convergence of the magnetic lines of force
causes a tightening of the spiral angle and eventually a
reflection of the particle from the high magnetic field
region (mirror point); for this reason the earth’s magnetic
field is capable of confining charged particles. A detailed
account of single particle motion in magnetic fields is
Figure 5-I. Cross-section of the earth’s magnetosphere in the noon- found in Roederer [1970].
. .
midnight meridian showing relatlve Iocatlons (Ilghtly shaded To a fair approximation, the earth’s magnetic field in
regions) of the earth’s radiation belts in the overall magne- the radiation belt region can be described in terms of a
tospheric topology.
magnetic dipole located near the center of the earth. The
dipole moment is M = 0.312 G R;, and the dipole is
ions with energies from less than 1 keV to hundreds of directed so that the magnetic south pole on the earth’s
MeV. Particles below 200 keV energy represent the prin- surface is located in northern Greenland (geographic
cipal corpuscular energy density and form the extrater- coordinates: 78.5° N, 2910 E); on the earth, the northern
restrial ring current. Confinement (or trapping) of these end of the compass needle points to this location. The
particles results from the dipolar-like topology of the spatial distribution of the dipolar magnetic field strength
geomagnetic field which is characterized by magnetic is
field lines that converge at high latitudes towards the
[4-3 C052A] 1/2
()
R -3
poles resulting in a relative minimum magnetic field B=BE ~ (5.1)
COS6A
strength region in the vicinity of the geomagnetic equa-

5-1
CHAPTER 5
where R is the radial distance measured from the center sources, internal transport processes, and loss mecha-
of the earth, RE is the radius of the earth, BE = 0.312 G is nisms. The outer boundary of the radiation belt trapping
the equatorial field at R RE and A is magnetic latitude.
❑ region occurs at the point where the magnetic field is no
A detailed account of the earth’s magnetic field and its longer able to maintain stable trapping, and at low alti-
variability is found in Chapman and Bartels [195 1]. Since tudes the earth’s atmosphere forms an effective boundary
the geomagnetic field is inhomogeneous, a radiation belt for radiation belt particles. During geomagnetically quiet
particle experiences varying magnetic field strengths over conditions, the radiation belt region extends from the top
its trajectory. Field variations on a length scale of the of the atmosphere along dipole field lines to an equator-
order of the particle’s gyroradius cause a net drift across ial radial distance of at least 7 earth radii. Energetic ions
the magnetic field in the azimuthal direction around the and electrons that encounter the dense atmosphere col-
earth as illustrated in Figure 5-3. This is a direct result of lide with the atmospheric constituents and are readily
the field strength being greater closer to the earth causing lost from the radiation belts. Particles with mirror points
well inside the atmosphere (nominally below -100 km
altitude) are said to be within the atmospheric (bounce)
10.ss cone. Cinematically, the angle between the velocity
vector of such particles and the magnetic field direction
(the pitch angle) at the equator is below a certain value,
known as the loss cone angle. Figure 5-4 exemplifies this.

Figure 5-3. Schematic representation of the gyration and azimuthal

drift (solid curve) of an equatorially mirroring proton with
associated current patterns (dashed curves) [Shulz and Figure 5-4. Illustration of magnetic mirroring in a dipolar magnetic
Lanzerotti, 1974]. field. The single particle trajectory shown in the solid line is
for a particle outside the atmospheric bounce loss cone and
the dashed line represents the trajectory of a particle inside
the 10SScone. The latter particle will encounter the denser
the particle orbital radius of curvature to be less there. parts of the earth’s atmosphere (mirror point height nomi-
The direction of the magnetic force depends on the sign nally below IOf)km) and will thus precipitate from the radi-
ation belts.
of the particle charge: electrons drift eastward and posi-
tive ions drift westward. Thus the energetic trapped par-
ticles are spread out in a belt-like configuration around
the earth, forming the radiation belts. The earth’s radiation environment is best studied with
Radiation belt electrons move at very high speeds. combined experimental and theoretical means. On one
For example, electrons with a kinetic energy of 500 keV hand, it is impossible to encompass the entire magneto-
move at 85% the speed of light. Radiation belt electrons sphere by experimental techniques or even to measure all
must thus be studied using relativistic theory. Ions, the physical parameters that may have bearing on the
because they are substantially heavier, generally move at dynamical phenomen~ on the other hand, this environ-
subrelativistic velocities; at 500 keV the proton speed is ment is so complex that there can be little hope of theo-
3T0 of the speed of light, while the heavier ions are even retically predicting the total radiation belt behavior solely
slower at the same energy. from a set of mathematical postulates. For example, the
The composition and flux intensities of the earth’s governing diffusion equations describing the trapped
radiation belts are determined by the strength of the radiation phenomena may be known in analytic form,

5-2
 a
THE RADIATION BELTS

mm
PROTONS ELECTRONS
but the transport coefficients that enter into them must
be empirically determined. 001
This chapter gives an account of these and other ap- Id 01003
I 03
plied concepts. The current state of knowledge of the B. ,o 3kHz z
o
10° ~
geomagnetically trapped radiation is described both from a
3oom30Hzlo3 >
the theoretical perspective and from direct observations. w
We demonstrate how this knowledge is used to construct 102 100
physical models of the radiation belts. Empirical radia-
tion belt models based on data compiled from many
spacecraft and a brief survey of man’s interaction with
10
geospace are also presented. u
v
z
$
5.1 THEORETICAL PRELIMINARIES 3 Hz

mm
The difficult mathematical problem of the motion of 0.3
m
energetic charged particles in a dipolar magnetic field
3m
was extensively studied during the first half of the twen- 102 100
100
tieth century. A general analytic solution to the equation 30
30 10
10
of motion was never found, and in most cases particle Imiz
1-
10° I mH?
orbit tracing had to be, done numerically. The interested 0.3 ~
0.3
0. I 0.1
reader is referred to Stbrmer [1955]. Physical approxima- 0.03
0.01 0.8?3
tions that lead to great simplification have, however, 102 0.003
0.003
been found. This is known as the adiabalic theory for
I 2 4 8 2,4 8
trapped particles [Alfv6n and Ftilthammer, 1963], and L L

the earth’s radiation belts have now been successfully

described in terms of adiabatic invariants and their Figure 5-5. The gyration, bounce, and drift frequencies for equatorially
perturbation. mirroring particles in a dipole field as function of L-shell for
different-p~rticle energies- [Shulz and Lanzerotti, 1974].

these fundamental particle motion frequencies for pro-

5.1.1 Single Particle Motion tons and electrons in the earth’s radiation belts [Schulz
An ensemble of ions and electrons moving in space and Lanzerotti, 1974].
constitutes a plasma that can exhibit many modes of
collective as well as single particle behavior. In the pres-
ence of electric (E) and magnetic (B) fields these parti- 5.1.2 Adiabatic Invariants
cles are subject to the electromagnetic Lorentz force, In general, the motion of charged particles is such
F = q(E + V x B), where q and V are the particle charge that momentum and energy can be transferred between
and velocity vector respectively. For ions q = Ze, where the different particles, and between the particles and the
Z is the ionic charge state and e is the unit charge; for fields that influence their motion. Therefore, it is not
electrons q –e. This force controls the particle motion,
❑ always possible to identify constants of motion. How-
and collectively the ensemble of charged particles can ever, under certain conditions these energy and momen-
modify the fields through induction, charge separation, tum exchanges are very small, and it is possible to iden-
and electrical currents formed by differential ion and tify specific quantities that remain virtually unchanged
electron motion. When the latter effects are negligible, with the particle motion. These are called adiabafic
the particles move independently of each other and the invariants.
single particle motion approximation is applicable.. Associated with each of the three quasi-periodic
For geomagnetically trapped particles there exist modes of motion is an adiabatic invariant related to the
three quasi-periodic motions: gyro motion around the Iiamilton-Jacobi action variable:
magnetic field lines, bounce motion between the conju-
gate mirrorpoints, and drift motion around the earth. The (5.2)
fundamental physics in this approximation is described
in detail by Alfv~n and Falthammar [1963]. One should
note that the frequencies associated with each of these
periodic motions are such that fgyro >> fbounce >> ‘drift. where d Qis a vector line element along the path of inte-
For this reason the three types of motions are largely gration. Here P is particle momentum and A is the mag-
uncoupled. Figure 5-5 illustrates numerical values of netic vector potential (that is, B = V x A). The integration

5-3
 CHAPTER 5
is extended over the particle orbit for gyro motion, where the relativistic factor T = 1 <-. p is an
bounce motion, and azimuthal drift motion (for i = 1, 2, approximate constant of motion when both of the fol-
3 respectively). If the particle’s trajectory closed exactly lowing conditions are fulfilled:
on itself, then the action variables Ji would be absolute — The spatial scale of B-field variation is much
constants of motion. Finite spatial and temporal varia- larger than the particle gyroradius
tions in B prevent perfect closure, and thus the Jis are at
best approximate constants.
A >> pg ❑ P1/ Blql ❑ TmoV sincr/ Blql (5.6)

5.1.2.1 First Adiabatic Invariant. J1 is obtained by eval-

uating the integral in Equation (5.2) over the particle
gyro motion only, that is, over the particle orbit projec- — The time scale of change of the B-field is much
tion in a plane perpendicular to B. Using subscripts II and larger than the particle gyroperiod
L to denote directions parallel and perpendicular to B,
and by virtue of Stokes’ theorem [Schultz and Lanzer- T >> ~g = 27rpg/vL ❑ 2r~mo/Blql. (5.7)
otti, 1974] one derives

J1 = 27TpgPL + ; rrp:B . (5.3) 5.1.2.2 Second Adiabatic Invariant. J2 is obtained by

evaluating the integral in Equation (5.2) over the bounce
trajectory and averaged over the gyro motion, or equiv-
where B IBI, p is particle momentum and pg = P1/ Iql B
❑
alently along the magnetic field line (guiding center field
is the particle gyro (or cyclotron) radius. From this, one line) around which the particle gyrates, and thereby
defines the first adiabatic invariant defining the second adiabatic invariant:

PI P2 Sin2~
~=== (5.4)
2mOB ‘

also known as the relativistic magnetic moment. Here m. .Qm

is the particle rest mass, and tY= arc sin (B. P/ PB) is the
pitch angle illustrated in Figure 5-6. where d Q is an element of length along that field line
segment and Qm is the curvilinear distance of the mir-
rorpoints from the equator measured along the guiding
center magnetic field line. Since equatorially mirroring
particles (a. = 7r/2) do not have any bounce motion, it
follows that J=O for such particles.
Provided the particle mirrorpoints are above the
dense atmosphere, J will remain an approximate con-
stant when the time scale of B-field variation is much
larger than the particle bounce time between the conju-
gate mirrorpoints

‘Qm
Figure 5-6. A particle with its velocity vector inclined to the magnetic
field at an angle a. This angle is called the particle p“itch - T>>TB= dQ/Vll(Q) . (5.9)
angle. 1
Qm
For non-relativistic particles,

(5.5a) Constancy of the first adiabatic invariant p implies

that
where Cl is the particle kinetic energy associated with the
directions perpendicular to the local magnetic field direc- sin’2ao sinza 1
—= —= — (5.10)
tion, while for relativistic particles B. B Bm = constant

over the bounce motion between the mirror points. Here

tl.e subscript zero denotes equatorial quantities and Bm is
the magnetic field induction at one of the mirror points

5-4
 THE RADIATION BELTS

(where a ‘ T/2). Equation (5.10) is known as the mirror Although Equation (5. 17) cannot be solved explicitly for
equation. Using (5. 10) one finds Am(aO), a numerical solution is easily obtained, or one
may approximate as in Harnlin et al. [1961]

cosAm(ao) = [sinaO]l/4 . (5.18)

Pm
It should be emphasized that using a dipolar mag-
netic field representation explicitly disregards any azi-
muthal asymmetries of the geomagnetic field. Such
‘!2m
-’/2
asymmetries do exist and become significant beyond
=—
m97

P
1
Qm
dl(l-+)~ (5.11) L=5. Under such conditions a different magnetic field
representation should be used, and this is outlined in
Section 5.6.1.4.

and in dipolar coordinates (Equation 5.11) becomes

5.1.2.3 Third Adiabatic Invariant. J3 is obtained by
2mOy evaluating the integral in Equation (5.2) over the particle
TB . ~ T(aO) , (5.12) drift motion around the earth, and averaged over gyro
and bounce motion
where T(ao) is tlie bounce time integral given by

T(ao) =
Am(ao)

1- -
COS A [4-3 COSZA]
dA
‘ (5.13)
J3. : ~=!
c $ A“dl (5.19)

J [4-3 COSZA]112}112
o { where @ is the magnetic flux enclosed by the azimuthal
drift orbit and 1 is linear azimuthal distance. Using
Stokes’ theorem yields
and Am(ao) ii the magnetic latitude of the mirrorpoint
which depends on the equatorial pitch angle ao.

To a fair approximation
‘=$A”d’’lB”ds s
(5.20)

where S is a surface bounded by the azimuthal drift path.

T(ao) -1.30-0.56 sinao IHamlin et al., 1961] , (5.14)
In a dipolar magnetic field one calculates [Roederer,
1970]
or alternatively

T[ao) - 1.3802-0.3198 (sinao + [sinaO]l/2) (5.21)

‘ [Schulz and Lanzerrotti, 1974]. (5.15)
where L is the McIlwain [1961] L-parameter.
Jj will remain approximately constant when the time
Other approximations are given by Davidson [1973]. scale of B-field change is much longer than the azimuthal
Gradients of T(ao) should, however, not be derived from
such approximations. From Equations (5. 1) and (5.10) it drift time ?d = ds/ vd around the earth. Determina-
follows that $’”
dnft
tion of the azimuthal drift velocity is discussed in the
B. BE next section.
Bm = L=- , (5.16)
sinzao Lssinzao

where in a dipolar magnetic field L = (R/ RE) is the equa- 5.1.3 Particle Drift Motion. In a uniform magnetic
torial distance of a given field line, and field, charged particies execute a spiral motion such that
the angle ‘between the particle velocity vector and the
cos6Am(Qo) magnetic field direction (the pitch angle) remains con-
sinzao ❑ (5.17) stant. When the magnetic field lines converge, the parti-
[4-3 coszAm(ao)]l/2 “ cle will respond to an effective net magnetic force from

5-5
 CHAPTER 5
higher to lower magnetic field strengths. This is illus- In general, the magnetic field may also have an inten-
trated in Figure 5-7. The physical reason for this force is, sity gradient across the field lines. This is illustrated in
Figure 5-8. Charged particles moving in such a magnetic

(a)

(a)

(b)

IB I
I
1 ~llBl

(b)
(c)
F]gure 5-8. (a) A uniform B-field where field lines are represented as
evenly spaced.
Figure 5-7, (a) Particle motion in a uniform magnetic field (uniform
spiral motion).
(b) A magnetic field with increasing strength perpendicular
to B, represented as denser field lines with mcreaslng B.
(b) The tightening of spiral motion in a converging magnetic
field.
(c) Illustration of magnetic force with gyroaveraged net com-
ponent in the -VB direction in a converging magnetic
field.
field will have a smaller gyroradius in the higher field
region and a larger gyroradius in the lower field region.
that the particle gyro motion produces an elementary As a consequence, there will be a net drift velocity per-
current (which may be interpreted as a magnetic dipole pendicular to the magnetic field direction. In this figure
current loop). For each such loop the effective current is a positive charge would drift into the paper and a nega-
tive charge out of the paper. Defining the angular
._ dq . qP~ gyro frequency
1P -x
(5.22)
2rpgmoy ‘

(5.24)
where pg is the mean gyroradius over the loop. The mag-
netic moment of a current loop enclosing an area A is
one can express the instantaneous vector gyroradius as
~:
M.ip A.— (5.23) PxB CPXB
2moyB Pg=—=mo-yQB —’ (5.25)
~B2

where A = mp~, which is the particle magnetic moment

and the drift velocity is then the time rate of change of pg
itself. The particle will therefore, averaged over its gyro
motion, be subject to a net force F = –M v,, [BI in the
direction along the field lines away from the higher field vd=d~=:(%XB) (5.26)
region.

5-6
 THE RADIATION BELTS
where it is assumed that the magnetic field is constant in which non-relativistically becomes
dp
time. With F ~ being the net force due to the cross-B
❑
2cell
gradient, one obtains the “gradient drift” Vcd = — (BxvIIBI). (5.33)
qB3

Vgd ❑ A (-MV1 [Bl x B),

~B3
Although the effect of the earth’s gravitational field is
rather small compared to other forces on radiation belt
——
%2(vIIBI x B), (5.27a) particles, it can easily be included:

cmoy
CP: vgrav ‘~(gxB) (5.34)

2m oq ~B3(v~ IBI x B),

where g is the vector gravitational acceleration.
which non-relativistically is just The effect of a weak, externally imposed electric’ field
is also easily taken into account:

Vgd = #(BxvllBl). (5.27b) ExB

Ved=&(qEx B)=c —” (5.35)
~B2 B2

The electric field drift is independent of particle charge

The earth’s magnetic field is also curved (that is, the
and mass as long as either is non-zero. Thus, under the
dipolar-like field lines form loops from pole to pole), and
influence of an electrostatic field, ions and electrons drift
the field line radius of curvature is given by
together (plasma flow), while under the influence of an
inhomogeneous magnetic field, oppositely charged parti-
cles drift in opposite directions (causing current flow).
Rc ❑ & COSA (1+3sin2A)3/z
1+sinzA ‘
(5.28;
The total particle drift velocity is then the superposi-
tion of the contributing drifts:
where the individual field lines are described by the
dipole relation P? + ‘2P;
Vd=c (Bx VLIBI) (5.36)
2mOyqB3
R = R. COSZA (5.29)

with RO = REL. Thus a charged particle moving in that +2 B2(Ex B)+c~ (g x B)

field will experience a centrifugal force

Pi EL + 2CII
FC z —n (5.30) Vd . (B xVLIBI )
moyRc qB3
(5.37)
where n is a unit vector in the direction away from the moyc
+:( Ex B)+
~ (g’ B).
instantaneous field line center of curvature. This causes a
drift velocity
In the dipolar magnetic field representation an approxi-

Vcd . —
;2(&nxB) (5.31)
mate formula for the drift period is given by Davidson
[1977]

and in the absence

of significant plasma currents 1.43K,
Td % (5.38)
IB1 Ly (V/ c)2 (1 +0.42 sinao)’
v x B . (Jand v * IBI .- ~ n [Roederer, 1970], and one
c
can write
where Kt = 1.0308 x 104 seconds for electrons, Kt = 5.655
4~ Z: BE R;
CP; seconds for protons, and Kt ❑ for ions of
Vcd = (Bxv~lBl), (5.32) 3Mi C3
moyqBs mass Mi and charge state Zi.

5-7
 CHAPTER 5
The cartoon in Figure 5-9 illustrates the principal 5.2.1 Qualitative Description
drift effects associated with the different drift mechan- A number of sources are considered responsible, and
isms. In Equations [5.36) and (5.37) the terms are listed the effectiveness of each probably also varies with time:
in order of their impbrtarice in the radiation belts. Above
1. Particles from the-sun, including solar wind parti-
-10 keV the magnetic gtadient curvature drift is gener- cles and energetic solar particle emissions, possi-
ally strongest, and Static electric field and gravity effects bly via magnetotail storage. For a description of
are usually neglected in radiation belt studies. The
solar cosmic rays see Chapter 6.
gradient-curvature drift carries energetic electrons to- 2. Particles from the earth’s ionosphere, including
wards the east and ions to the west. Thus there will be a
the polar wind flow into the magnetotail, particles
net westward electrical current encircling the earth. This flowing up the magnetic field lines to fo~rn the
is the extraterrestrial ring current. These findings are plasm~phere and particles accelerated out of the
summarized in Figure 5-10. auroral ionosphere (Chapters 8 and 9).
3. Cosmic ray dbedo neutron decay (CRAND)
+ within the trapping region.
MAGNETIC FIELD CHARGED CHA;GEO 4. Particles arriving at the earth having been accel-
UPWARDS THROUGH PARTICLE PARTICLE
THE PAPER erated in interplanetary shock waves or in the
magnetospheres of other planets.

o o
5. Low energy components of galactic cosmic rays
(A)
HOMOGENEOUS (Chapter 6).
B-FIELD
O! STURBING
NO
6. In situ acceleration of pre-existing low ehergy
FORCE trapped particles within the radiatio~ belts.
The solar wind flows past the -rth’s magnetosphere
virtually at all times (Chapter 3). Some of these particles
may find their way through the outer regions of the
(01 magnetosphere to the stable trapping region [Hovestadt
HOMOGENEOUS
B- FIELO Wlt’H et al., 1978]; this process may be particularly effective
ELECTRIC FIELD
during periods of southward heliospheric magnetic field.
DRIFT
Direct transient injections of solar energetic particles
E DRIFT
I probably also occur, particularly in conjunction with
magnetic storms.
(c)
HOM06ENE6US
Ionospheric particles diffusing out of the polar iono-
E- FIELD WITH sphere (polar wind) escape into the magnetotail region
EXTERNAL FORCE
INDEPENDENT OF from which some may become energized and injected
SIGN OF CH.A*GE
(s. Q. GRAVITATION) into the trapping region. Auroral electric fields are
DRIFT DRIFT intermittent and can have a sigtiificant component paral-
lel to the magnetic field, and ions and electrons from the
FIELD ‘RDNGER
(D) topside aurora! ionosphere can be accelerated to multi-
INHOMOGENEOUS
B- FIELD
keV energies. The wave fields associated with plasma
waves may also cause particle acceleration. This could be
a source of H+, He+, 0+, and electrons provided other
GRAD /0/
l— processes act to trap the particles.
4

I
I DRIFT Cosmic rays impacting the earth’s atmosphere under-
‘R’FT FIELD k~AKER
go nuclear reaction!, and a flux of neutrons escapes from
Figure 5-9. Summary illustration of drift of both positive atid negative
the top of the atmosphere. Free neutrons are unstable
charged particles in a) a uniform B-field, no external force, and decay into proton, electron and neutrino triplets on a
b) a uniform B-field with a perpendicular electric field; c) a time scale of -IOW seconds. If the electrically charged
uniform B-field with an external force which is independent
of electric charge such as gravit~ and d) a B-field with a
decay products find themselves within the radiation belts,
gradient [Alfven and Falthammar, 1963]. they will immediately be subject to the magnetic force
and may become trapped. Empirically, this is an impor-
tant source of multi-MeV protons in the innermost part
of the inner radiation zone (L < 1.5). There is little direct
5.2 TRAPPED RADIATION SOURCES information about the efficiency of direct extraterrestrial
Precisely where the radiation belt particles come from energetic particle trapping in the radiation belts. One
and how tHey are accelerated to energies in the keV and may surmise, however, that time variability of the geo-
MeV range are still areas of research for which a c~m- magnetic field is needed for trapping to occur or that
prehensive answer is not yet available. incident extraterrestrial energetic ions in low charge

5-8
—

THE RADIATION BELTS

NORTH
~TRAJECTORY OF

MIRROR POINT
(PITCH ANGLE OF HELICAL TRAJECTORY = 90°)

DRIFT OF
PROTONS

MAGNETIC CONJUGATE POINT

Figure 5-10. A descriptive drawing of the three types of motion of particles trapped in the earth’s magnetic field.

states (for example C+ and 0+) may charge exchange to them towards lower L-shells on a time scale that pre-
higher charge states (for example, C6+ and 08+) within serves p and J but violates the constancy of 0. For equa-
the magnetosphere so that their gyro radii become small torially mirroring particles, a radial displacement from
enough (pg = P1/ qB) for trapping. L=5 to L=3 increases the particle energy by a factor of
The plasma sheet in the earth’s magnetotail is consid- -5. Furthermore, if the particle spectrum follows a
ered an important reservoir for radiation belt particles. power law distribution j(E) = (E/ Eo)-u with u = 3, for
However, the plasma sheet particles themselves are likely example, then the apparent particle flux enhancement
to be a mixture of particles from several of the prime seen at fixed energies will be a factor of 125. Seen at a
sources mentioned above. During magnetospheric sub- fixed L-shell, the observable flux increase will be even
storms plasma sheet particles convect inward toward the greater if the pre-storm radial distribution falls off
earth, and in the process can be accelerated and may toward lower L-shells (as may be the case at lower radia-
become trapped in the radiation belts. Unfortunately, lit- tion belt energies), and smaller if this flux gradient is
tle is known about the specific details of the time negative. Given the knowledge of the storm perturbation
dependent trapping process and it is not yet possible to electric field (magnitude, direction, azimuthal extent, and
make a quantitative evaluation of the strength and char- duration) together with observations of the pre-storm
acteristics of this source. radiation belt structure, this in silu “source” is in princi-
Current research also points to the earth’s ionosphere ple assessable.
as an important contributor to the lower energy particle
population below a few tens of keV, perhaps with a
roughly equal contribution from solar wind particles. In 5.2.2 Simplifying Assumptions
contrast, the high energy particles above ’500 keV For most of the radiation belt source mechanisms
appear to have an extraterrestrial source. The former accurate quantitative information is still lacking.
conclusion is derived from observations of dominant Ideally, one would like to know
oxygen fluxes at times, while the latter stems from 1. the source strength for different particle species as
observed carbon-to-oxygen ratios of order unity. For the a function of energy and pitch angle,
very important intermediate energy range where most of 2. the effective source locations within and on the
the radiation belt energy density is found, there is no boundaries of the trapping region, and
experimental result indicating the source. 3. the source strength as function of the different
A simple theory that seeks to explain observed storm- geophysical conditions during quiet and disturbed
time enhancements of radiation belt particle fluxes at times.
tens and hundreds of keV energies has been advocated Unfortunately, contemporary research has not yet
[Lyons and Williams, 1980]. Particles existing in the yielded quantitative answers to these requirements.
outer radiation zone may suddenly, during the storm Without this information, how can we understand and
main phase, be subject to an electric field that transports model the earth’s radiation belts?

5-9
 CHAPTER 5
For quiet time conditions one can solve the steady DAWN
———.
state radiation belt transport equations for the interior of
—._
the radiation belts subject to suitable outer zone bound-
ary conditions on the trapped fluxes. This amounts to
the assumption that the radiation belt source is capable
of supplying particles to the outer radiation zone bound-
ary at a rate sufficient to offset losses within the trapping
region. The existence of long-term approximate stability
of the radiation belts as a whole during extended quiet
periods supports this contention, and fortunately trapped
flux observations from geostationary spacecraft, such as SUN
ATS-6, of the outer zone flux levels at L 6.6 makes this
❑

a feasible solution.
This is not a satisfactory situation for magnetic
storms and other disturbances. As a consequence most
radiation belt modeling has been done for steady state,
quiet time conditions. Time dependent radiation belt
modeling would require time dependent boundary condi-
tions averaged over local time. It is conceivable that data
from several geostationary satellites may be used for this
purpose, but this has not yet been done.

5.3 TRANSPORT PROCESSES IN THE Figure 5-11, Equipotential contours for an electric field in the earth’s
equatorial plane (dashed lines). These are also drift paths
RADIATION BELTS for very low energy particles. The electric field is a super-
As we have seen, in the static geomagnetic field, radi- position of a corotational E-field due to the rotation of
ation belt particles execute the three periodic motions: the earth and its imbedded magnetic field and a uniform
dawn-dusk electric field. The spearatix (solid curve) is the
gyration around the magnetic field lines, bounce motion low energy particle Alfv/n layer which separates the open
between mirror points, and azimuthal drift around the and closed drift paths IKivelson, 1979], The outer oval
earth. The latter type of periodicity is caused by the represents the magnetopause in the equatorial plane.

gradient-curvature drift motion. Effects of gravity and

electric fields cause departure from this simple picture, tion, of course, take place during disturbed conditions
but both forces are primarily important at low energies when the imposed “convection” electric field is time vari-
typically below a few tens of keV. able and the low energy particle distributions are not in
equilibrium IHarel et al., 1981 a & b].
5.3.1 Convection
Magnetospheric convection results from externally
imposed electric fields. The solar wind flows past the 5.3.2 Stochastic Processes: Phenomenology
earth at a velocity in the range 200-600 km/ sec. This At higher energies, particularly above a few tens of
implies a “convection” electric field E = –V/ c x B across keV, the dominant drift of trapped particles is due to the
the earth’s magnetosphere directed from dawn to dusk. gradient and curvature effects of the geomagnetic field.
Combined with the electric field induced by the rotation The prime drift motion is therefore circular (with gyro
of the earth a characteristic magnetospheric convection and bounce motion superimposed) around the earth.
pattern is set up (Axford, 1964; Roederer, 1970). The Departures from this pattern are due to the fluctuating
radial corotational electric field is induced by the earth’s nature of the geoelectric and geomagnetic fields induced
magnetic dipole field corotating with the earth. Figure by variations in the solar wind flow and internal magne-
5-11 shows model electric equipotential lines around the tospheric processes. Radiation belt particles are also sub-
earth due to these electric fields in a time-independent ject to interactions with plasma waves and suffer colli-
situation. Particles of quite low energies (< 1 keV) are sions with exospheric neutral atoms and low energy
primarily controlled by the electric field drift, and their plasma particles. Common to these processes is their
motion approximately follows the equipotential lines as randomness in occurrence, and their effects are described
indicated by the arrows in Figure 5-11. Notice the topo- by stochastic analysis that can be reduced to diffusion
logically distinct regions: near the earth where the con- theory: radial diffusion and pitch angle diffusion. The
vective motion follows oval paths around the earth, and physical ideas are illustrated in Figure 5-12. Radial diffu-
at greater distances where the drift paths are open to the sion transports radiation belt particles across the dipolar-
magnetopause. Departures from this overall configura- like magnetic field lines in the radial direction, and pitch

5-1o
 THE RADIATION BELTS
/ mean change Ap, AJ, and A* takes place in the adia-
/ batic invariants per unit time (1Apl <<M, IAJI <<J, and
/
IA@l <<Q). The distribution function f(p, J, 0; t) aver-
I
aged over gyrophase, bounce phase, and azimuthal drift
SOURCE /—
phase is then governed by the relation ‘
\
\
\ f(p, J, Q; t) ❑ d(A~) d(AJ) d(A@)
\ /11
\ PITCH ANGLE
DIFFUSION f(p-A~, J-AJ, @-A@; t-At)
P(/.-Ap, J-AJ, @-A@; Ap, AJ, A@),
Figure 5-12. A conceptual representation of pitch angle and radial dif-
fusion in the earth’s radiation belts. Diffusion occurs in (5.39)
either direction, but in most cases there is a net diffusion
flux in the indicated direction towards the earth’s
and one expands f and P in Taylor series around the
atmosphere.
unperturbed quantities to arrive at
angle diffusion alters the particle pitch angle (or equival-
ently, the mirrorpoint location). In both cases the earth’s
atmosphere is a sink: for radial diffusion by transport to
very low L-shells, and for pitch angle diffusion by lower-
ing the mirrorpoints into the atmosphere. In addition to
diffusive processes, energy degradation from collisions
with exospheric particles also occurs.
J2 <( AJ)Z> f + ~ <( A@)2> f
+—
dJ2 ( 2At ) d~z ( 2At )
5.3.3 Effects of Field Fluctuations
Adiabatic invariants are useful substitutes for particle
constants of motion. It should be kept in mind, however,
that the earth’s magnetic field is never perfectly static.
Field fluctuations are associated with micropulsations d2
[Jacobs, 1970], magnetospheric substorms [Akasofu, <*fl@> f) (5.40)
+ dJd@ (
1968], geomagnetic storms and other phenomena, and
field oscillations vary considerably in magnitude, fre- + higher order terms,
quencies, and principal location. The adiabatic invariants
are said to be violated when electric or magnetic varia- where the stochastic coefficients are defined by
tions take place near or above the adiabatic motion fre-
quency in question. Even slow field variations may vio-
late the third adiabatic invariant @ (T– minutes to <Ai> = d(AM) d(AJ) d(A@)
hours) while p-violation requires wave-like fluctuations JJJ
on a time scale of -milliseconds. Microscopically, the
earth’s radiation belts are subject to field fluctuations P(p, J, @; Ap, AJ, A@) Ai
that occur at quasi-random times. Their effects are best (5.41)

Jll
described by stochastic methods [Chandrasekhar, 1965]
that treat the mean deviations in the adiabatic invariants <AiAj> z d(Ap) d(AJ) d(A@)
and the associated distribution function.
Define a particle distribution function f = f(g, J, @; t) P(w, J, Q; AM, AJ, A@) AiAj
such that the number of particles located within a
parameter space volume element given by P - 1/2 dp to with i and j being p, J or @ in all permutations.
p+l/2dp, J-l/2 dJto J+ l/2 dJ, @-1/2 d@ to@+
1/2 d@ is given by The transport equation (5.4 1) can be greatly simpli-
fied by recognizing that violation of one adiabatic invar-
dN ❑ f(p, J, @; t) dp dJ dQ iant is almost always uncorrelated with the process vio-
lating another. In that case all cross coefficients must
at a time t. The function f is thus the particle density in vanish:
this parameter space spanned by the adiabatic invariants.
Let P (K, J, Q; AK, AJ, A@) be the probability that a <ApAJ> ❑ <ApA@> = <AJAG> ❑ O. (5.42)

5-11
CHAPTER 5

Furthermore, <Ai> and ~ <( Ai)z> (i ❑ p, J, @) are In this case the Jacobian coordinate transformation is
just
related. In the absence of external sources and losses, dif-
fusion would proceed to transport particles from over-
abundant regions of parameter space until all gradients G(L; @)=g=-2TR: BE+, (5.50)
in the distribution function had vanished, and for each
diffusion mode
and the pure radial diffusion equation becomes

~Ai~ _ ~ <( Ai)2>

=O (i ❑ p, J, @). (5.43)
dl 2
:= ‘2$[D’’L-2%I ‘S-L ‘551)
Equation (5.4 1) then simplifies to
where —S and — L represent particle source and loss func-
tions. DLL is t~e pure radial diffusion coefficient at con-
stant p and J values. It has been estimated that

DLL ~ DLL(~o) = DLL (f) ~(~o) (5.52)

where

where ~(ao) is a function only of the particle equatorial

~., . <( Ai)2> (i=p, J,@). (5.45)
11
2At pitch angle and DLL = DLL ~ is the radial diffusion
()
coefficient for equatorially mirroring a. = Z particles
Equation (5.44) is the diffusion equation describing () 2
the earth’s radiation belts. It is valid whenever the per- where ~ ~ = 1. For a description of ~(ao) see Schulz
turbations are small (but accumulative). Transformation [1975a]. ()2
to other variables, such as @l, 42, r#J3is facilitated by the
Jacobian G = G(w, J, @; @l, 42, @3) such that
Both geomagnetic and geoelectric field fluctuations
contribute to DLL. For geomagnetic fluctuations it can
be shown that

D~~) (P, L) = U: H“J (@d) Llo (5.53)

where
where P(M) is the fluctuation power spectral density
evaluated at the azimuthal angular drift frequency
d~j 2 Vd 2~
DOj@j ❑ Dii (x (i= p, J, @), (5.47) — —. Empirically ~(”J (~) a w-r where the
)
‘d = LRE = td
value of r most often is r = 2 + 1. When r= 2, one obtains
and the very simple expression

F ❑ F(@l, 42, 43; t). (5.48) D~~) = K(M) L1O (5.54)

where K(M) is a factor dependent on the fluctuation

5.3.4 Radial Diffusion
magnitudes. For geoelectric field fluctuations it can be
Field fluctuations on a time scale comparable to the shown that
azimuthal drift time around the earth can violate the
third adiabatic invariant @ but preserve the p and J
invariants. This is the limit of pure radial diffusion. To
study this process requires knowledge of the diffusion
coefficient, sources, and losses. It is convenient to use the
dipole L-shell coordinates defined through where ~\E) is the n-th spatial Fourier component of the
electric field power spectral decomposition. For details,
see F~lthammar [1968], and Cornwall [1968].
(5.49)
The actual calculation of D~~ is fairly complicated
and the reader is referred to the research literature. How-
where BE z 0.312 G is the equatorial B-value at the sur- ever, for typical substorm conditions Cornwall [1972]
face of the earth where L = 1. derived the simple relation

5-12
 THE RADIATION BELTS
is the bounce averaged pure pitch angle diffusion coeffi-
cient related to the local pitch angle diffusion coefficient
Daa by
where PM is the fiagnetic moment in MeV/ G and Zi is
the particle charg6 state number.

The actual calculation of ~aoao from observed

Both K(M) and K(EJare dependent on the geophysical
plasma wave distributions in the magnetosphere is quite
activity. Likely values of K(M) fall in the range 2x 10-10
to 2 x 10-s (L-shells)z per day and K(E) may be found in complicated. The reader is referred to the research litera-
the range 10-6 to 10-4 (L-shells)z per day. The total radial ture [Lyons et al., 1971, 1972; Retterer et al. 1983 and
diffusion coefficient is then references therein].
In general, pitch angle diffusion can come about by
violation of p only (that is chatige in pl) by violation of
(5.57)
J only (change in PII),or by a combined violation of, both
p and J. The pure pitch angle diffusion limit is ideally
It should be emphasized that the relations in Equations rehlized when both ~ and J are violat!d so that the ratio
(5.54) and (5.36) represent simplified cpnsid~ations that of pL/ pll changes ivhile Ipl remains almost constant.
may not always be realized. If, for example, P(M) or ~(E~ When this happens, there is essentially no energy
do not follow OJ-Zdependence, then D~~) will involve exchange between waves and particles. Physically, this
dependence on p, and D~~ may have a differept form. can mean that the principal ifiteraction is between the
particle and the magnetic field of the wave.

5.3.5 Pitch Angle Diffusion

The presence of plasma and electromagnetic waves in 5.3.6 Energy Diffusion
the radiation belts implies fast low-amplitude field fluc- Particles can become energized when interacting with
tuations, and some of these waves (such as the ELF waves (wave damping) or can lose energy to the waves
whistler mode) can violate the Qrst adiabatic invariant w. (wave instability). One may write a pure energy diffusion
These fast fluctuations will, in principle, also violate @ equation in the form
and J; however, for these adiabatic invariarits the effect is
likely to be at least in part averaged out. It is convenient
to convert from p to equatorial pitch angle coordinates

p2 L3 sin2ao However, such an equation has not been used much in

P= (5.58) radiation belt physics, since almost inevitably a. also
BE
changes. The mote general case of coupled energy and
whereby the applicable Jacobian is angular difftision and the associated diffusion coefficients
needs further research. For information relating to such
coupled processes, We Schulz and Lanzerotti [1974].
pz L~ sin2&0
G(aO; P) = *O = BE (5.59)

5.4 LOSS MECHANISMS

at constant particle momentum. In the pure pitch angle
diffusion limit Ip[ is unchanged and oniythe ~article Energetic particles residing in the radiation belts are
direction of motion changes. subject to collisicmal interactions with coexisting particle
populations. The most important of these are the earth’s
main atmosphere, the atomic hydrogen exosphere, and
the plasmasphere. Near the earth (that is, at very low
The pure pitch angle diffusion equation becomes
L-shells or for small equatorial pitch angle:)such
collisions constitute a dominant energetic particle loss
g=
dt
1
~ EaO@Osin2aOT(u~)$
sin2aOT(aO) dp [ 1(5.60)
mechanism. Bu~even in the centtal parts of the radiation
belts Coulomb collisions and charge exchange” can be
quite significant. Pitch angle scattering of particles into
+g* -L*
the atmospheric bounce loss cone’ (wheie particle-pafiicle
where ~ and L* are the source and loss functions collisions are dominant) is also of great significance,
appropriate for the pitch angle diffusion process. DaOaO especially for radiation belt electrons.

513
CHAPTER 5
5.4.1 Exosphere nevertheless thought that this formula represents average
long-term cold plasma densities appropriate for long-
The terrestrial exosphere (or geocorona) is a contin-
term steady state radiation belts studies. Further im-
uation of the atmosphere to great altitudes where colli-
provement, including storm and substorm variability,
sions are infrequent and the constituents follow ballistic
will be needed when time dependent’ radiation belt
trajectories. The principal constituent is thought to be
models are developed.
atomic hydrogen with a density ranging from —104
atoms/ems at 103 km altitude to -102 atoms/ cms at
3 x 104 km altitude. Table 5-1 gives the mean atomic 5.4.2 Coulomb Collisons
Coulomb collisions are inelastic interactions between
Table 5-1. Number densitv of neutral hvdrogen at the equator for an charged particles. A radiation belt particle “colliding”
exospheric temperature of 950 K and for average geomag-
netic-conditions [Tinsley, 1976].
with an exospheric neutral hydrogen atom, for example,
will interact with the internal atomic electric field when-
T=950K ever the impact parameter is less than the atomic radius,
or with the electric field from a thermal (plasma) proton
L-Shell [H] (1 /cm3) L-Shell [H] (1/cm3) or electron out to the particle’s Debye shielding distance.
The encounter will result in energy transfer from the
1.1 15000. 1.5 16000. energetic (incident) particle, and in deflection (angular
2.0 3700. 2.5 1500. scattering) of both particles. Changes in ionization states
of either or both particles can also occur. On the average,
3.0 800. 3.5 470.
angular scattering is important for radiation belt elec-
4.0 300. 4.5 210.
trons, but usually not so important for the much heavier
5.0 148. 5.5 120. radiation belt ions. The collisional scattering process is
6.0 98. 6.5 83. qualitatively illustrated in Figure 5-13. - -

hydrogen number density [H] as function of L-shell at

the equator [for example, Tinsley, 1976] and it is thought
to be an average representation for an exospheric
temperature of –950 K. The exospheric temperatures
and densities will of course change with solar and geo-
magnetic activity. -y ~__ –__
Also overlapping the radiation belts is the terrestrial ‘–--’’ATOM

plasmasphere consisting of thermal ions and electrons

and contained within an L-shell range roughly beiow
about L-4 to 6 by the effect of the corotational electric Figure 5-13. Deflection of a positive and negative particle in the inter-
field of the earth. Figure 5-11 shows the electric equipo- nal atomic electric field of the target atom. b+ and b are
tential lines in the equatorial plane formed by the combi- impact parameters and rA is the atomic radius. If the
impact parameter is greater than rA, no interaction occurs.
nation of the corotational electric field and the solar For plasma particles rA is replaced by the Debye shielding
wind induced dawndusk “convection” electric field. The distance AD.
separatrix between closed (around the earth) and open
equipotential lines is related to the static plasmapause,
although in a dynamic, time variable situation no simple Radiation belt particles are much more energetic than
relation between the two exists. Empirically, there ‘is a exospheric particles and energy is transferred from the
much higher density of cold (1000 – 10000 K) plasma- energetic ion to the atomic bound electron(s) or plasma
spheric particles below the plasmapause than beyond it. electrons. The changes in the energy spectrum and direc-
Based on data deduced from ducted VLF wave propaga- tional characteristics of particles traversing a material
tion experiments, Cornwall [1972] estimated the follow- substance whose effective thickness is greater than the
ing average plasmaspheric particle densities: collision mean free path is a difficult mathematical prob-
lem. Great simplification is, however, obtained when the
[e] = 250 (L/ LO)K cm-s for L < LO differential collision cross sections are small enough to
(5.63) overwhelmingly favor very small energy losses and direc-
[e] = 13 (L/Lo)K cm-s for L > Lo tional changes in each collision. Fortunately, this is true
for Coulomb collisions in which the classical Rutherford
where Lo = 4.1 and K = 4.64. cross section is valid. See Rossi and Olbert [1970] or
It is well known that the plasmasphere deviates fre- Jackson [1975] for details.
quently and strongly from this functional form. It is One finds that the average energy loss rate for an

5-14
 THE RADIATION BELTS
energetic particle passing through a gas of atomic parti- The Coulomb collision 10SSfactor is given by
cles is given by

de
~ . -4Zi2 Zt [n] r, (mec3/ ~) F(~,Zt) (5.64) GC(L) ❑ (50 W m e4 fiH L912/ (meB#2)) [H] Z?
(5.71)
where Zi is the net charge state number of the incident
particle, Zt is the nuclear charge number of the target gas where mH = 1.67 x 10-*A grams is the hydrogen atom
atoms, me is the electron mass, re is the classical electron mass and e = (4.80286 * 0.00009) x 10-10 esu is the unit
radius, c is the speed of light, D =V/c and F(~,Zt) is a charge. Zi is the ionic charge state number for incident
slowly varying function determined quantum mechani- energetic ions [Cornwall, 1972; Spjeldvik, 1977]. For a
cally by fuller treatment of the collisional pracess the reader is
referred to the treatise by Matt and Massey [1952] or
Rossi and Olbert [1970].
F(P, Zt) ❑ -2.9+ in [~’2m~c4/(1-~2)3/2 I*(Zt)]

for energetic electrons, (5.65)

5.4.3 Charge Exchange
and Collisional encounters may also change the ionization
state of the colliding particles. This may entail ionization
F(B, Z,) ❑ -2~’2 + in [4m~cd~A/(1-p2)2 I*(Zt)] of the “target” atom and/or alteration of the net ionic
charge of the incident particle. The latter is of impor-
for energetic ions (5.66) tance for radiation belt ions since almost all af the physi-
cal processes depend directly on their charge state. The
where I(Zt) = 13.5 Zt (in eV) is an approximate value of simplest of the charge exchange reactions is the one that
the ionization potential for the gas atoms IRossi and neutralizes radiation belt protons (H):
Olbert, 1970].
Let f = f(p, J, ~; t) denote the distribution function H++ H- H+H’
— — (5.72)
for equatorial radiation belt particles. The changes due to
stochastic energy loss from the energetic particles can be
described by where underlying denotes energetic particles. The above
reaction is an example of the electron capture process in
af
()
~ Coulomb ‘
- ~ (< Ap/At>f)
dp
- & (< AJ/At>f), which the incident proton picks up (or captures) the
orbital electron from the thermal hydrogen atom, which
(5.67) then becomes a low energy proton. Once neutralized, the
incident proton (now fast neutral hydrogen) is no longer
and the stochastic time average <Ap/ At> and <AJ/At > subject to the magnetic deflecting force and escapes from
are then simply given by the trapping region. Microscopically these events occur
at random, and fast neutral atoms thus exit the radiation
belts in all directions. Since the speed of these particles
(5.68)
greatly exceeds the earth’s gravitational escape speed,
many disappear to outer space. Some of these fast neu-
and tral atoms move towards the earth where they produce
secondary interactions upon entering the atmosphere.
(5.69) In a similar manner, other singly charged ions can be
neutralized in such collisions:

Cornwall [1972] nated that it is possible to treat the

He++ H-
— He+H+
—
slowly varying logarithmic term in the expression for ~
as an approximate constant. By including contributions C++H -C+H+ (5.73)
— —
from energetic particle plasma electron collisions as well
as from energetic particle exospheric atom collisions, one O++ H- Q+ H+.
—
derives the expression valid for equatorially mirroring
(aO = T/2) particles: These newly generated thermal hydrogen ions (protons)
make a contribution to the earth’s plasma envelope;

()gdt Coulomb
however, their rate of formation is probably less than the
GC(L) p-1i2 (: (5.70)
= ) rate at which such ions are supplied from the topside

5-15
CHAPTER 5
ionosphere. One may note that the total charge is con- below about 20 keV, but falls off sharply witfi higher
served in the radiation belts under the charge exchange energies.
reactions.
There is aisoanother class of ion charge exchange;
this results from electron stripping reactions by which the
energetic ions (necessarily heavier than hydrogen) lose
one or more of the remaining bound electrons. For
example, there is a finite probability that an ion (say 0+)
in a collision with a thermal hydrogen atom loses several
of its electrons:
Figure 5-14. Flow diagram for energetic oxygen ion charge exchange.
~+ + H- 05+ + H+ + 4e- (5.74) This figure illustrates that while multiple charge state
— — increases are possible in a single encounter, only single
step charge state reductions can occtii in a hydrogen atom
gas. Oxygen ions in lower charge states may be further
ionized (through single or multiple electron stripping) in
In this example four electrons are stripped off. Whether single collisional encounters with the exospheric gas
the fast 0+ ion actually loses one or more of its electrons atoms. The ions may lose net charge in collisions by cqp-
or captures one from the hydrogen atom strortgly turing the bound electron from a ‘thekmal hydrogen atom.
To ~ecome neutralized, a fully ionized oxygen ion
depends on the energy of the incident ion and the details requires a minimum of eight separate collisions.
of the collision. Since this type of reaction preserves a
non-zero ionic state, the ion remains trapped within the
radiation &lts. For this reason one distinguishes b?tween
“internal” charge exchange reactions such & the stripping
reactions:

0’2++H+03++H+e-
— — i, ‘,
(5.75) :@ - ‘) ‘,,
* — F,,. ,,.l. (1%0) \
‘,,
03++H- 07+ + H + 4C- — Clofl,. (19m)
— — : m-m
– ‘— BImkfnmm .“8 K,.-”’ 11930)
;
‘,
..... Exl..ag.lal,cm
v ‘,,
~21 I I I I 1
al I 10 Im low lam w
with their reverse reactions such as Particle Energy (ke V)

08++H_07++H+
— — Figure 5-15. Proton charge exchange cross section as function of
prpton energy. It is wofih noting’ that this cross section is
Q7++H+06++H+ (5.76) high (uJOP - lo- 15)cmz) below a ~ew iens of keV causing
short proton lifetimes tit those energies, and it falls off
02J+H-O++H+ rapidly towards higher energies. THe values between 0.4
— — and 1 keV are from the experimental work of Fite et al.
[1960], the data from 1 keV are due to a compilation by
on one hand, and reactions that pr0dt2Ce immediate par- Claftip [1970], and above 1000 keV the theoretical results
ticle loss (lowest charge state to neutral) noted above on from Brinkmann and kramers [1930] have been used
[Spjeldvik, 1977].
the other. Notice also that to becqme neutralized, an 08+
ion requires a minimum of eight” separate collisions with
hydrogen atoms (since only one electron may be cap- For ions heavier than protons, multiple potential
tured in each collision). This should be contrasted with charge states are available. For helium ions, one must
electron stripping reactions in which tiie multiple charge consider not only the cross section for the neutralization
state changes occur in a single collisiot’sal encounter. The reaction but charge state changes: state 1 - state” 2 and
principal features of the charge exchange chemistry are state 2 - state 1 as well. Thus, for helium there are
illustrated in Figure 5-14. three important cross sections to be included. A point
The probability that ‘a given charge exchange process worth noting is that the process transforming He+ to
actually takes place in a collision is expressed as a re- Hez+ dominates over the charge state reducing reactions
action cross section. Such cross sections derive from in the high energy part of the radiation belts, essentially
laboratory studies of collision processes; however, for above 800 keV. This is explicitly depicted in Figure 5-16
many ions the pertinent cross sections have not yet been which shows the three charge exchange cross sections for
measured over a sufficiently large range of particle ener- radiation belt helium ions.
gies. Figure 5-15 shows the cross sections for the charge Charge exchange cross sections for the heavier ions
exchange of ~rotohs incident on atomic hydrogen from a such as carbon and oxygen have been measured over a
compilation by Spjeldvik [1977]. Notice that the charge small fraction of the radiation belt energy range. The
exchange cross section is quite high (U~o > 10-15 cmz) situation is particularly severe for carbon ions where

5-16
 THE RADIATION BELTS

Port lcle Energy (ke V)

Energy Ike Vl

Figure 5-16. Helium ion charge exchange cross section as function of

Figure 5-18. Charge exchange cross sections for (atomic) oxygen ions
energy. Solid lines show the values compiled by Claflin
in an atomic hydrogen gas: Charge gain/ Electron loss
[1970] from many experiments, and the dotted lines are
cross sections 0+ - Of I+J)++ j e The data shown are from
extrapolations, The dashed lines are computed from the
Lo et al., [ 1971] for O-targets. The solid lines are crude
theoretical work of Brinkmann and Kramers [1930]. The
estimates made by Spjeldvik and Fritz [1978a]. More reli-
three pertinent charge exchange cross sections are: He+ -
able cross sections for oxygen ions and other heavy ions
He (ion neutralization), Hez+ - He+ and Hez+ - He+. The
need to be established.
extrapolated curve above 1000 keV (for He+ - Hej+) is
progressively unreliable towards higher energies
[Spjeldvik and Fritz, 1978b]. charge exchange processes on radiation belt ion distribu-
tion functions is through loss and gain terms. For radia-
tion belt protons one has the expression
almost no measurements have been made. For oxygen
ions there exist a number of measurements, but unfortu-
df
nately the laboratory work used particles other than ❑ -Alo f (5.77)
()~ Charge Exchange
atomic hydrogen as targets. Figure 5-17 shows estimated

since the loss rate is proportional to the number of pro-

tons present (or more precisely, the distribution function
~, and also proportional to the factor Aij = <Vuij [H] >
where V is the ion speed and aij the charge exchange
cross section for transformation state i - state j. To de-
10‘,,
: Electron Copt.re/Charge Loss scribe the heavier ions where more than one charge state
Cross Secr,onof W ,n H
; mm
m — ,s!,.O !!, applies, one considers a distribution function for each
,0!0

charge state. For helium ions, let fl and f2 be the distri-

bution functions for He+ and He2+ ions respectively. The
,,,OL.-.-.-.7 –—–
charge exchange processes are then represented by
,0,

()
af,
= ‘Alo fl + A21 fz - A12 fl
Figure 5-17. Charge exchange cross sections for (atomic) oxygen ions ~ Charge Exchange
in an atomic hydrogen gas: Charge loss/Electron capture (5.78)
cross sections Oi+- O(i- I)+estimated from sparse available
data. Stebbings et al. [1960] give data with H-targets up to

()
8 keV, Lo et al. [1971] give data with O-targets and af2
MacDonald and Martin [ 1971] give data with He-targets. = -A21 f2 + A12 fl (5.79)
% Charge Exchange
Also shown are results with rarefied air-targets [Solov’ev
et al., 1972], The cross sections for reactions with i> 1
have not been measured below - 10 MeV [Spjeldvik and
Fritz, 1978a]. and this provides a coupling between the distribution
functions.
ion charge loss (electron capture) cross sections based on In general, for an ion species with s available charge
a compilation of a number of measurements using H, air states one obtains charge exchange expressions of the
and O as target particles [Spjeldvik and Fritz, 1978a]. form
Estimates of the charge gain (electron loss) cross secti’ons
have been made from limited data available. Examples dfi
are illustrated in Figure 5-18. For the remaining reactions
between the multiple charge states, no laboratory mea-
surements have been reported, and one must use crude
()
~ Charge Exchange = ‘X
j=o
‘ijfi + i
j= I
‘jifj
for i #j,

(5.80)
estimates [Spjeldvik and Fritz, 1978a].
The mathematical description of the effects of the and when the exosphere predominantly consists of

5-17
CHAPTER 5
atomic hydrogen (as in the case of the earth) the first and propagate to interact further with another particle
summation contains contributions fromj = i – 1 only. population elsewhere, then the latter process is termed
“parasitic”. Both types of processes are of major signifi-
cance in the radiation belts.
5.4.4 Wave-Particle Interactions
Plasma waves play an important role in radiation belt
physics. Different conditions for wave propagation, 5.4.4.1 Pitch Angle Scattering Into the Loss Cone. In-
growth and decay in the magnetosphere have led to teractions between energetic particles and plasma waves
delineation into numerous wave modes. For a detailed can significantly affect the trapped particle population’s
classification of these modes see Stix [ 1962] or Kennel et energy and angular distribution. If the pitch angle is
al. [1979]. In a previous section, the radiation belt trans- altered so that the particle finds itself within the atmos-
port equation was given; here we shall outline some of pheric bounce loss cone, it will have a high probability of
the concepts concerning the interactions. becoming lost upon entering the atmosphere. For ions
this probability is virtually 100% while for energetic
Angular scattering can result from interactions with
electromagnetic waves. As an example, consider gyro (or electrons a certain fraction of the precipitated particles
cyclotron) resonance: a radiation belt particle spiraling are backscattered up into the magnetosphere only to
around a magnetic line of force (the guiding center locus encounter the conjugate hemisphere during the succeed-
field line) will gyrate at a rate determined by the mag- ing bounce motion.
netic field strength, mass, and charge of the particle. An
The directional change in the angular scattering proc-
electromagnetic wave (with electric and magnetic wave
ess due to plasma wave or collisional interactions is ran-
vectors) also propagating along that field line will rotate
dom: to lowest order it is just as likely for a single parti-
according to its wave frequency. When both the sense of
cle to be scattered from lower to higher pitch angles as in
rotation and the rotation frequency match for both wave
the opposite direction. However, when the particles are
and particle, the particle will be subject to an essentially
anisotropically distributed in pitch angle (for example,
constant “wave” field for the duration of the encounter.
The particle can exchange energy with the wave through with $ > O), the number of particles scattered from the
.
the electrical interaction and/ or deflection can occur higher p~rticle density region towards the lower density
through the magnetic interaction. Higher order reso- region is greater than the number scattered the other
nances can also take place, for example, if the rotation way. As a consequence, the stochastic process is biased
rates differ by a factor of two. When the particle-wave by the particle pitch angle distribution and a net d[llu-
interaction is primarily via the magnetic wave vector (as sion flux occurs. This pitch angle diffusion flux is just
in the case of the radiation belt electron -— whistler mode
hiss wave interaction) the result is primarily angular scat-
tering; this is the pure pitch angle scattering limit. Other Fdiff = ~aoao
Sin Za. T(~.) go (5.81)
resonances including bounce resonance and drift reso-
nance can also be important.
so that the diffusive transport term in (5.45) may be
These waves in the radiation belts may remain in the written as simply
area where they were generated (very small group veloc-
ity) or propagate afar (large group velocity). They have 1 dFdiff
Jf ._ —. (5.82)
frequency and wave length, and there are different polar-
()% cliff sin2aOT(aO) da.
ization properties. Propagation properties are determined
by a dispersion relation [Stix, 1962] that in part &pends
on the density of the plasma in which the wave exists and Atmospheric particle losses within the bounce loss
the geomagnetic field. Stability or instability of the waves cone generally cause f(ao < @OLC) to be near zero
is frequently determined by the energetic particles with (except under conditions of extremely strong scattering).
which the waves can interact. This generally favors a positive pitch angle anisotropy
Plasma waves are said to be unstable to growth if (loss cone distribution), and particles can be lost to the
interactions with the charged particles transfer energy to atmosphere at any L-shell due to the pitch angle scatter-
the waves; if the transfer is from the waves to the parti- ing process. Of course, df/ dao may be negative under
cles the waves are said to be damped. If plasma waves certain conditions, and thus net reverse diffusion can
interact with radiation belt particles in such a way as to take place. Examples are an atmospheric source cone in
have their principal interactions locally, then it is possible the aurora] zone as a consequence of auroral electric
to self-consistently give a theoretical treatment of both fields, effects of L-shell splitting due to the deviations
wave and particle properties. On the other hand, if the from the dipolar azimuthal symmetry of the magentic
waves have significant spatial propagation so that they, field [Roederer, 1970], or particle injections during dis-
for example, gain energy from one particle population turbed times.

5-18
 THE RADIATION BELTS

Analysis of resonant wave-particle interactions violat-

I
ing the first adiabatic invariant have generally considered sdRpR

waves with frequencies near the gyro frequency of the

t
energetic particles. The condition for resonance between
waves of angular frequency w and particles at the gyro-
frequency Q is given by

w - kll Vll = n Q; n = O, +1, +2, +3, . . . [Lyons, 1979], ,IDSS CONE

(5.83)

where kll and Vll are the parallel (to B) wave vector and
particle velocity respectively. For a given wave mode
there is a dispersion relation linking w and k. The cycfo-
tron harmonic resonances have the (Doppler shifted)
wave frequency equal to a harmonic (n= +1, n = +2,
n.*3 , . ..) of the particle gyrofrequency, and the classi-
cal Landau resonance (n = O) has the wave parallel phase Figure 5-19. Radiation belt electron ELF whistler mode wave inter-
velocity Vll = oJ/ kll. All these resonances result in particle actions: Regions of cyclotron resonance in momentum
diffusion in both energy and equatorial pitch angle ao. space, Upper panel shows the effects of pure pitch angle
diffusion (conserving electron energy) for the resonance
The Landau resonance results in diffusion solely in with plasmaspheric ELF whistler mode turbulence, Lower
Vll, conserving V1. Thus the pitch angle change Aa at a panels show the effect upon progressively more energetic
given location along the particle trajectory is related to electrons, and that for the most energetic electrons the
high order resonances become increasingly important as
the parallel velocity change Avll by the electrons diffuse in pitch angle towards the atmos-
pheric bounce loss cone. The resulting pitch angle diffu-
Aa=-&Avll. (5.84) sion coefficient thus becomes a strong function of pitch
angle [Lyons et al., 1972].

The cyclotron harmonic resonance can likewise pro-

duce pitch angle changes resulting from the energy
The earth’s magnetic field is, however, quite in-
exchange, or the interaction can be primarily with the
homogeneous when considered in its totality, and the
magnetic wave vector causing pitch angle scattering more
wave energy is, in general, unevenly distributed over the
directly with little energy exchange. The latter process is
space of the radiation belts. During the particle bounce
particularly important for electrons.
motion, as the particles move away from the geomag-
In general, the study of wave-particle interactions
netic equator along its trajectory, the increasing magnetic
requires an extensive mathematical treatment, and the
field strength causes both particle pitch angle and parallel
necessary derivations and analysis are beyond the scope
velocity to change. The increase in the local pitch angle
of the present chapter. The interested reader is referred to
as a particle moves away from the equatorial plane (see
a number of works on the subject [Stix, 1962; Sagdeev
Equation 5. IO) implies that cyclotron resonance can
and Galeev, 1969; Lerche, 1968; Kennel and Engelman,
occur for all equatorial parallel particle energies greater
1966; Lyons et al., 1971, 1972; Retterer et al., 1983].
than a minimum value Ell~ln.
Figure 5-20 illustrates radiation belt electron and
5.4.4.2 Scattering of Energetic Electrons. Pitch angle ELF whistler mode wave propagation. The significant
scattering is particularly important for radiation belt elec- wave-particle interactions for energetic electrons are
trons. Here, we bypass the extensive mathematical treat- shown. Note that the wave propagation does not neces-
ment found in Lyons et al. [1971, 1972]. The different sarily follow magnetic field lines, and therefore the parti-
resonances are illustrated in Figure 5-19 which depicts cles may interact with waves generated over a significant
the regions of cyclotron resonance in velocity space volume of the magnetosphere. For comparison, the inner
resulting from waves distributed over a band of parallel and outer radiation zones (for electrons) are indicated.
wave vectors Akll. Resonance at each cyclotron harmonic Satellite measurements have shown that a band of
occurs over a band of parallel velocities Avll; for simplic- whistler mode waves centered around a few hundred
ity, relativistic effects for electrons >500 keV are not hertz exists essentially continually within the plasma-
included. No cyclotron harmonic interaction occurs for sphere. Because of the persistence and audio frequency
v < Vll~in, and the value of Vll~in depends on the actual range, this type of wave is called hiss. Ray tracing studies
kll band over which the wave energy is distributed. This have shown that the plasmaspheric hiss can readily prop-
is locally true where the geomagnetic field may be con- agate across the geomagnetic field lines [Lyons and
sidered uniform on the scale of the gyroradius. Thorne, 1970] and thus fill a great volume of the inner

5-19
CHAPTER 5
magnetosphere with wave energy. These waves are be-
lieved to be generated in the outer regions of the plasma-
sphere, and within the plasmasphere the ELF hiss turbu-
T “
lence is the dominant wave component that interacts <
with radiation belt electrons. Waves that may influence 0
%
trapped particles can also be generated by earth-based
radio sources.
200—-—.
..
KeV
___
------ ---,

CYCLOTRON

2000 KeV

=.=,

t
WHISTLER ~lVEN ELECTM DIFFUSION
L=4 BOUNCE AVERAGED
I 1 1 1 1 1 # I 1

Figure 5-20 Spatial illustration of radiation belt electron wave-particle 0° 30° w 90°
interactions. Characteristic locations for the inner and EWATORIAL PITCH-ANGLE
outer electron radiation zones are shaded, and the mag-
netic field line approximating the average plasmapause
location is also indicated. Left side: 0.5 kHz whistler- Figure 5-21. The bounce averaged electron pitch angle diffusion coeffi-
mode ray paths in the radiation belts showing internal cient Daoa(, calculated for all cyclotron harmonic reso-
reflection and cross-L propagation characteristics [from nances and the Landau resonance. Note that at each
Lyons and Thorne, 1970]. Right side: Typical energetic energy there is a region of very low pitch angle diffusion
radiation belt electron trajectories indicating the spiral (for 20 keV electrons, near 87”); this “bottleneck” in the
motion betwee,~ the magnetic mirror points. The spatial pitch angle diffusion coefficient is the cause of tbe
regions where the Landau and different cyclotron reso- “bumps” in the actual electron pitch angle distributions
nances are most effective are noted [Lyons et al. 1972]. giving rise to the so-called bell-shaped distributions
[Lyons et al., 1972].

Based on typical observed wave characteristics, Lyons It has been demonstrated that there is a limiting value
et al. [1972] calculated the pitch angle diffusion coeffi- 1’ that the radiation belt integral omni-directional flux I
cient for both cyclotron and Landau resonances shown cannot exceed without provoking a cyclotron wave
in Figure 5-21. They also computed lifetimes for ener- instability. Schulz and Lanzerotti [1974] estimated this
getic electrons subject to this wave-particle interaction limit to be about I’ = 101I L-’t particles cm ~ sec I inte-
process. An example of these lifetimes for average grated over all energies and pitch angles. The linear wave
modeled wave parameters and a normalized wave ampii- growth rate is Yg, and therefore the growth rate for wave
tude of 35 my is shown in Figure 5-22. These lifetimes energy is 2yg. An incipient wave undergoes a partial
take on great significance in the modeling of the radia- reflection (reflection coefficient R < 1) upon traveling
tion belt electron structure and will be discussed in more a distance d — L RE; the remaining fraction, l-R, of
detail in the modeling section. the wave is lost from the radiation belts. The time inter-
val between wave reflections is T = L RE/ Vg where
Vg = (Idw /dkl ) is the group wave velocity. The condition
5.4.4.3 Limit On Radiation Belt Particle Fluxes. Plasma
for marginal stability is that the waves on the average do
waves generated by radiation belt particles locally will
not grow further in time; that is, the decrease in the wave
have their growth rates in part controlled by the intensity
amplitude upon reflection is restored in one traversal
of the trapped energetic flux of those particles in reson-
between bounces:
ance with the waves. While radiation belt electrons
within the plasmasphere are controlled by the parasitic
R exp [2?g LRE/Vg] ❑ 1. (5.85)
type of interaction mechanisms, it has been found that
the radiation belt particle fluxes beyond the plasma-
sphere can be effectively limited by the self-generating This defines the marginal growth rate to be
wave mechanism. In the following a few principal aspects
of the radiation belt saturation process are outlined; for Yg ❑ (Vg/2LRE) Iln RI. (5.86)
a more detailed mathematical treatment the reader is
referred to Kennel and Petschek [1966] or Schulz If I exceeds 1’(that is, if Y exceeds Tg) the consequence is
[1975b]. a net growth of wave energy, and the stronger waves

5-20
 THE RADIATION BELTS
10$ .VSAS 15.
(Cmz. * jl .
.15. s A sm.
8
:.~----qo_ ----— _ .< ------- ------------ .
,, . ., .. . . 1,
0’ . “.-. \ .“..”*> . “ . ..
..\. . .. . * I .; “8*: ‘ o..
2. ,,
d. 10” . .
~ \
THEORETICAL LIMIT “’ E ‘“
Q’
I

.. . . . . “8---
10’ +.-.:.--_;_
. . .<-. . ..
-,>_ __ --.”- ”r.:7c: ---- A---.7-:
..
. ..
~6 .-. . . . . .. ““.. . . . . ..” “. .“
J. & .“.$ . . .. . “.

0’ ‘“”””” [

,n.lu
a
THEME TIcAL LIMIT .
!0”
“\”

0’ “:”. . :-----

1
.?-: “ :. “. . -.* __o=-- ~--- . ;

~6 ‘
.. . .7 _

-8-.— ------- / ------ ,. . .,. .;. ..

J. . . . . .$ . . .. . . . . :... . .
r“
. ... ~ ., .“ . .. “
..
I 2 3 4 5 *S . .. : .. . “.”.. , . :.. . .. . . “.
. . . . .. .“.
L %- . . .“ “m
n. ‘.
l,; ,1,, ,,1, ,,, l~llll,jl ,1,~, II
0’ Jllllll llllll\l!l illi
Figure 5-22. Radiation belt electron precipitation lifetimes due to elec- ,,%Tl ,,&J ,,%, ,,vN, ,T::l Tti ,TM!ITM2, ,, .2, ,,”2, –,
WY m JUL
tron pitch angle scattering into the atmospheric bounce m (w

loss cone. These lifetimes are given for an average whist~kr

wave amplitude of BWO= 35 myo. Notice that there is
generally a minimum in these lifetimes at each energy, and
that this minimum is displaced towards lower L-shells
with higher electron energy. At 2000 keV this minimum Figure 5-23, Comparison of the predicted whistler self-excitation limit
with ~>40 keV electron data recorded with the Explorer
occurs at L = 2.7 while at 500 keV it is substantially
broader and occurs at L = 4.2. This minimum e]ectron 14 spacecraft [Kennel and Petschek, 1966].
lifetime is ah important factbr for the electron “slot” lota-
tion defining the separation between the two radiation differential solid angle interval will be the same at all
belts.
pitch angles, even within the loss cone. Within the loss
scatter the particles in pitch angle resulting in precipitat- cone (ad < &OLC)the particles will be lost from the radi-
ion into the atmosphere and reduction in the particle ation belts on a time scale Tb, and for the strong dl~fu-
flux until I no longer exceeds I’. If an external particle sion [imit to be reached, the scattering mechanism must
injection source is strong enough to more than offset this be faster than this. The solid angle of the loss cone is just
maximum loss rate then the radiation belt particle fluxes ~LC = 2rr(1‘cosaOLC) and the maximum particle flux
will increase beyond the stably trapped flux limit I’; this loss rate ~ is then
may indeed happen for impulsive sources associated with
disturbed conditions. The aftermath of such an injection
(5.87)
will, however, be a rapid reduction to the stably trapfid
flux intensity level.
Figure 5-23 shows a comparison of the theoretical The only exception to this concept occurs when the
radiation belt flux limit with data [Kennel and Petschek, loss cone becomes a source cone due to the auroral
1966]. Notice that most of the time the outer radiation electric field parallel to the magnetic field, and a source
zone particle fluxes are just below the theoretical charac- cone distribution exiting one hemisphere can become
teristic flux limit. lost through direct precipitation into the conjugate
hemisphere.
5.4.4.4 Strbng Diffusion Limit. There is, however,
another important limit to consider, namely the upper
5.5 THEORETICAL RADIATION
limit on the rate at which radiation belt particles can
become lost to the atmosphere by scattering into the BELT MODELS ~
bounce loss cone. Cotisider the effect of very rapid pitch Two types of radiation belt models now exist: empiri-
angle scattering. The particle pitch angle distribution will cal models based on compiled observations and theoreti-
then become essentially isotropic; that is, the flux per cal models derived from our knowledge of radiation belt

5-21
CHAPTER 5
physics. The ideal empirical models represent the stan- (6) By symmetry f(ao)= f(rr - ao) so that one needs
dards with which the theoretical models must be com- to consider only the range O S a. ~ rr/ 2.
pared, and they are also useful for engineering purposes. The complete radiation belt modeling can be simpli-
However, they give only a smoothed statistical picture of fied in a way that retains much of the essential physics
the time period when the data were obtained, and the but greatly simplifies the mathematics,’ namely separating
data collection process is subject to experimental errors radial diffusion from pitch angle diffusion. One treats
and misinterpretations. Theoretical models, on the other radial diffusion for a. = m/2 particles only, and with
hand, can be used to simulate and predict radiation belt given f(ao = T/2) the pitch angle diffusion process at
behavior under a variety of conditions: quiet times, dis- fixed L-shells only. This approach is not strictly correct
turbed times, or magnetic storms. They can be used to when there is significant interaction between the pitch
study ionic species and charge state distributions for angle and radial diffusion modes, but it reduces an
which no experimental information is available. Such almost unmanageable numerical problem to cases where
models are only as valid as the physical approximation solutions can be found. The results should therefore be
they are based on, however. In the following sections treated with some caution.
simple theoretical models are outlined.

5.5.1.2 Electron Model. Radiation belt electrons are

5.5.1 Quiet Time, Steady State Models described by the diffusion equation
Studies of quiet geomagnetic conditions using theo-
retical radiation belt models serve as an important test of
our current understanding of the trapped radiation
sin2.:T(.o)
&[D~”~
%’L2k[DLLL-2El+
environment. It is assumed that under quiet time condi- (5.89)
tions the radiation belts can be described in terms of an df
sin2ao T(ao) ~
equilibrium balance between sources, internal transport, 1 + “(L)~-”2 ~ ‘
and losses. Symbolically one writes
where GC(L) is the Coulomb loss term for electrons.
Lyons and Thorne [1973] separated radial and pitch
%’(%)source+(%)transport +(~)loss=o ~ ““) angle diffusion by substituting for Equation (5.71) the
two equations
The sources are considered to be located on the
boundary of the trapping region (and thus describable
via a boundary condition on the trapped flux for each
species) as well as in the interior of the trapping region and
*’L2*[D’LL-2~1
-i-d (5.90)

(such as the CRAND-source). The transport mechanisms

af.
are’ radial diffusion and pitch angle diffusion. Particle
losses are caused by charge exchange ion neutralization,
energy degradation, and pitch angle diffusion into the
1
~ - sin2aoT(uo)
A
da.
[
DaOao sin2aoT(ao) ~
1“ (5.91)

They solved Equation (5.9 1) under an assumed angular

bounce loss cone. Particles diffusing outward may
distribution preserving exponential decay conditions and
encounter the magnetopause and become lost from the
obtained the lowest normal mode pitch angle distribution
magnetosphere. Radiation belt theoretical models consti-
and electron precipitation lifetimes TW. Their results are
tute the combination of these processes.
shown in Figures 5-24 and 5-25 respectively. The precipi-
tation lifetimes were calculated assuming a mean ELF
Whistler mode wave amplitude of 35nT. For other wave
5.5.1.1 Formulation of Boundary Conditions. For
amplitudes, these lifetimes scale as (BW/ 35my)-2.
steady state conditions one needs in general boundary
conditions on the particle distribution function f Using these characteristic electron scattering precipi-
(1) At L = Lrnax (outer boundary) one specifies the tation lifetimes it is possible to solve the steady state
particles energy spectra and angular distributions, radial diffusion Equation (5.90) for equatorially mirror-
which defines f at the outer boundary. ing radiation belt electrons. The results are shown in
(2) At L = Lmin = 1 (inner boundary formed by the Figure 5-25.
surface of the earth) the particle distribution func- A simplification in this work was a priori approxima-
tion vanishes (f= O). tion of the Coulomb collision “loss time” which was
(3) At ffo = rr/2 : df/aaO = o. defined as T.- - ~(d6/dt)- 1 where dE/dt is given by
(4) At a. =Oor~: df/dao =0. Equation (5.64); this is a rather crude approximation for
(5) At P = Pm.. (or E Ema.):f = O where Ems, ~100
❑ the last term in (5.90), since Tcc really depends on the
MeV. distribution function itself as given by Equation (5.66).

5-22
 :.
. m

a.:...
THE RADIATION BELTS

L.4

I 104
2W-6W Kev ..... “:::::
.....
... ... . . .
.:..
......
,..
20C0 KeV :.::::., .
500 Kev ?
/ :::::;:”::
-_ —.-
,/-- \ / ;;;;;;:;:;
/ 1 ..........
i 1500KeV . .. ......
II .. . ......
I50: ::V_,’

/
/“ ~
1

1/ :.. ... ..
E =0.10 mvlm

D
..... .. ..
/
1000 KeV Bw = 10my ;~;~~~”~”
. . . . . .....
........
.,.. . . ..
1000KeV
/- 1700-4000 Key::::”:...
/ .= ------- ..... . ..
,- I
1’ /
500KeV
/ / :;;:;:;:::
2

m
f’
1’
; 7
12~ / ), :;;:::::::
/ A .. . . .
500 KeV f
y;;”;;;;;;
/’
//’ IO-21234567 [234567
200 KeV
1~ L L
--
/

. /’
Figure 5-25. Theoretical and observed radiation belt electron radial

d
flux distributions at the geomagnetic equator. The theo-
,. 0- ;O-K:V retical results (dashed curves) are calculated from radial
.m, 6OOHz
1’
diffusion of energetic electrons subject to pitch-angle scat-
8ti=300Hz tering due to the (resonant) ELF whistler mode turbu-
N(L:4)= 1000 cm-3 lence (hiss) within the plasmasphere. The data (solid lines)
are taken from the OGO 5 spacecraft. Notice how well the
t 0 KeV
theory simulates the location and width of the observed
30” 60”
o~ w m“ 90”
radiation belt separation (slot) region over a wide range of
JATORIAL PITCH-y( E electron energies leaving essentially only the finer details
01 the equilibrium distribution for future investigation
[Lyons and Tborne, 1973].
Figure 5-24, Predicted radiation belt electron pitch-angle distributions
[Lyons and Thorne, 19731. The exam~le shown was calcu-
lated using a Gaussian E~F whistler mode hiss turbulence
frequency distribution peaked at 600 Hz and with a
proton flux radial distribution and this peak is located on
bandwidth of 300 Hz.
lower L-shells with higher proton energy. No specific
Much Of the essential radiation belt electron physics is wave mode has yet been identified that interacts in a
ret ained, however. Notice how well the electron slot dominant way with >100 keV protons below the average
region that separates the radiation zone into two belts is plasmapause location. On the other hand, protons (like
reproduced compared with actual data from the OGO-5 all ions) are subject to the charge exchange process.
Neglecting pitch angle scattering, the radial diffusion
spacecraft, both in width and L-shell location for differ-
ent energies. Perfect agreement in all details should not equation for equatorially mirroring (CYo= rr/ 2) protons
be expected since the theory treats average conditions may be written
based on parametrized wave characteristics, while the
data are for a specific time when the recordings were
made.
From such calculations we have learned that for
()
‘f
z
❑ Lzd/dL[DLLL-2(df/dL)] - Alof +

(5.92)
energetic electrons, wave-particle interactions are the G(L) p- l/2(df/dP)
cause of the division into two distinct belts, since elec-
trons in the “slot” region are lost to the atmosphere at a where Alo = < UFO[H] v > is the charge exchange loss
fast rate. Those few electrons that survive the cross-L frequency per unit distribution function f. To simulate
transport to arrive in the inner radiation zone became the steady state radiation belts (that is $ -’ O) this equa-
very stable since the time scale for wave particle scatter- tion can likewise be solved when sttitable boundary con-
ing is very long there. ditions are imposed. The boundary conditions are estab-
lished by using a measured outer zone proton flux spec-
5.5.1.3 Proton Model. Empirically, radiation belt pro- trum, and by covering an energy range sufficiently
tons are not distributed into two radiation zones. At con- large that f(L,p) = O at p > pmax where Nmax corre-
stant proton energy there is generally a single peak in the sponds to the upper limit of the radiation belt trapping

5-23
CHAPTER 5

m
,.s
energies considered. Figure 5-26 shows the computed
*3
E=IW keV

101

,.-1

,.-3

,Vs

,0-11 1,
12$456)
) I I I I 1 ~-n I I I I
12545$7

RADIAL DISTANCE (Earth Radii)

Figure 5-27. Theoretical proton radial distributions calculated from

proton radial diffusion subject to Coulomb collision
energy losses and charge exchange loss. Notice there is no
division (or slot) within the trapping region. The curves
ENERGY (keV) now show that the radial flux maximum is displaced
towards lower L-shells with higher proton energy. For
details about the numerical calculations see Spjeldvik
Figure 5-26. Radiation belt proton energy spectra obtained from a
[1977].
theoretical calculation b%ed on proton radial diffusion
subject to Coulomb collisions and charge exchange losses. DLL causes a substantial increase in the trapped fluxes
No wave-particle interactions were considered. Bohnd#ry
conditions were imposed at L-6.6 with data from the oh L-shells below the peak flux location. Proton models
ATS-6 spacecraft, and the computed spectra at lower that treat other than equatorially mirroring pitch angles
L-shells are compared with available equatorial data from in a comprehensive manner ~ave not yet been developed.
Explorer 45. The results of the theoretical calculations
below a few tens of keV energies may be inaccurate since The reason is partially that Daoao is not know for prot-
convection processes may dominate over diffusion at these ons, and partially that solving the simultaneous proton
low energies [Spjeldvik, 1977]. radial and pitch angle diffusion problem subject to the
simultaneous p-variations ii tnathbmatically difficult
radiation belt proton spectra obtained with this method,
owing to the great inhomogeneity in the coefficients.
and data are shown for comptison (taken from
Explorer 45 below L = 5.3). Notice that the spectral max-
imum found at L< 5 generally is displaced towards 5.5.1.4 Heavy Ion Models. Besides electrons and pro-
higher energies with lower L.shells. The flux values comp- tons, the earth’s radiation belts contain appreciable fluxes
uted below a few tens’ of keV we below radiation belt of energetic helium, carbon, oxygen, and other ions.
energies (convective processes also operate in that range) Theoretical modeling of these ion populations can be
and should not be considered reliable. The comparison carried out by a generalization of the proton model. One
with data is quite good, however, giving support to the must conside~ a distribution function for each charge
usefulness of the radial diffusion theory above at least state of a given radiation belt species, as well as the cou-
lW keV. pling between charge states imposed by the charge
A complementary view of the theoretical radiation exchange chemistry. For helium iom one may write the
belt proton fluxes is given in Figure 5-27 which depicts coupled equations for equatorially mirroring ions:
the cQmputed radial distributions for different proton

()
energies. Qualitatively similar to the obsemations, the afl
= L2d/dL [DLL IL-z(dfl /dL)] +
radial flux maxima are displaced towards lower L-shelis z
with higher proton energy. On L-shells well beyond the (5.93)
flux radial peak location, the characteristic time scales of
diffusive transport are substantially shorter than that of G](L) /.-112 (df1/dp) - AIQfl + A21f2 - A12fl
the loss processes. This situation is called diffusive equili-
and
brium. There is also an important connection with the
magnitude of the radial diffusion coefficient. As DLL
increases, the diffusive equilibrium radial range extends ‘~ = L2d/dL [DLL2L-2 (df2/dL)] +
towards lower L-shells, and the radial peak location is (5.94)
found at lower L. Also, the absolute magnitude of the
flux radial peak increases as DLL increases. An enhanmd G2(~) p-112 (df2/dp) - A21f2 + A21f1

5-24
 THE RADIATION BELTS
where the subscripts indicate the ion charge state. fourth charge state should be most abundant. A signifi-
These equations have also been solved numerically cant finding, illustrated in Figure 5-29, is that the radia-
[Cornwall, 1972; Spjeldvik and Fritz, 1978a], and some tion belt charge state distribution in the interior of the
of the findings are illustrated in Figure 5-28. Notice that radiation belts (L < 5) is almost completely independent
of the charge state distribution of the outer radiation
zone source. In fact, this figure shows that regardless of
the source being ionospheric (source charge state 1) or

,fi\,
solar (source charge state 6 dominant) a steady state
radiation belt distribution over ion charge states evolves
that is independent of the source characteristics in the

m
,.l interior of the trapping region.
m’
x ,b3

3
-1 105 I ‘–
u m’ ‘ 40 CQkeV

—-— 4 MeV OXYGEN - Ions: ON+

2(45F:
A: IONOSPHERIC sol CE B: SOLAR WIND sofJRcE

‘oo~
1 I 1 1 1
RADIAL DISTANCE (Earth Radii)

Figure 5-28. Theoretical helium ion radial distributions calculated

from radial diffusion theory with Coulomb collisions and
charge exchange. Solid curves show equatorially mirror-
ing fluxes of He++ while the dashed curves depict He+.
Notice that He+ is dominant below - 1 MeV while He++ is 4

4
dominant at higher energies ISpjeldvik and Fritz, 1978].

there is a transition from charge state 1 (He+) to the state 3

2 (Hez+) at energies around 1 MeV, so that the lower 5
charge state of helium ions is most abundant at the lower
2
energies and the higher charge state at the higher ener- o~ !
23456 7 23456
gies. The spectral features computed for helium ions
L- Shell L Shell
below – 1 MeV stem from the energy dependence of the
radial diffusion coefficient DLL and the loss rates. These
Figure 5-29. Relative charge state distribution for radiation belt oxy -
have yet to be verified or refuted by observations. gen ions. Left panel: an ionospheric source of O’ ions is
For even heavier ions a multitude of charge states are assumed; Right panel: a solar (wind) source of OfI+ions is
available to the radiation belt heavy ions: for carbon ions assumed. One finds the steady state radiation belt charge
state distribution in the interior of the trapping region
there are six positive charge states, and for oxygen ions becomes largely independent of the source charge state
there are eight. In general, for a given ion charge state of characterist its.
an elemental species with s available states, the radial dif-
fusion equation for equatorially mirroring ions may be
written
The findings concerning charge states of energetic
radiation belt ions may be summarized as follows:
dfi
_ = Lzd/dL [DLLiL-z(dfi/dL)] + Gi(L)~-112(dfi/d~) 1. Higher charge states are crucially important for
dt
the overall structure of the heavy ion component
of the earth’s radiation belts above - 100 keV.
—Aiofi 2. Charge state redistribution processes are of major
$Aijfi+$ Ajifj for i#j ‘5”95)
j:f) importance throughout much of the radiation
belts and at all energies.
where Aij denotes charge state transformation from state 3. Radiation belt charge state distribution becomes
i to state j, so that Aij O for i = j ISpjeldvik, 1979].
❑ largely independent of source charge state charac-
The radial diffusion equation has been solved for teristics because of the frequent “internal” charge
equatorially mirroring radiation belt oxygen ions exchange.
[Spjeldvik and Fritz, 1978 b], and one of the findings is 4. Relative charge state distribution is to a large
that oxygen ions can be more numerous than protons at extent independent of the diffusive transport rate
multi MeV energies. The oxygen ion charge state distri- in much of the inner magnetosphere.
bution is predicted to vary from a dominance of 0+ ions The last two conclusions are valid below a charge
at energies below - 100 keV to successively higher charge state redistribution zone adjoining the particle injection
states at the higher energies; for example at 4 MeV the region in the outer radiation zone.

5-25
CHAPTER 5
5.5.1.5 Theoretical Radiation Belt Ionic Composition. injection events. Lyons and Thorne [1973] have demon:
Because of experimental difficulties, only very limited strated that the radiation belt electron particle scattering
information is available concerning the actual composi- lifetime can explain quantitatively the restoration of the
tion of radiation belt ion fluxes. For this reason we pres- radiation belt slot region after it becomes filled in during
ent a theoretical prediction of radiation belt protons, the storm injection process. Spjeldvik’and Thorne [1975]
helium and oxygen ion fluxes at L 3.25 (in the center of
❑
subsequently demonstrated that the precipitating electron
the radiation belts). This is shown in Figure 5-30. It must flux into the middle latitude ionosphere following such
storms causes enhanced D-region ionization of sufficient
magnitude to explain VLF radio wave phase anomalies
Helcum Ions Atomic
Oxygen
Ions
observed at such times. Spjeldvik and Lyons [1980] have

kL
)l(HeT + I ~( He++)
,$,,(0”+)
LX3.25 L-325 suggested a simplified prediction model for these effects.
c
There is reason to think that magnetic storms differ
B considerably from one another, not only by the magni-
A tude of the ring current storm index D~t but also in the
injection characteristics of energetic ion fluxes. Once

,~ 4
IIII
\

r‘\ \
‘\
injected, radiation belt particles become subject to the
normal processes in the trapping region discussed earlier.
For specific magnetic storms, the post-storm

agreement between predictions and observations

decay of
protons and heavy ions has also been studied, and fair
has
ION ENERGY (keV)
been reported in limited energy ranges where the data
were available [Spjeldvik and Fritz, 1981a & b].
Figure 5-30. Theoretical model of the radiation belt ion fluxes: Com-
parison of theoretically predicted energy spectra of 5.5.2.1 The Variability of the Parameters. Within the
Proton, helium and oxygen fluxes at L=3.25 for different framework of radiation belt diffusion theory, the injec-
diffusion coefficients ISpjeldvik, 1979].
tion of particles is described by a (time variable) source
term S(L, P,ao;t). The boundary conditions may also be
be kept in mind that these predictions have yet to be
time variable, reflecting dynamic conditions on the outer
verified or refuted by experimental observation, particu-
edge of the stable trapping region f(Lnlax, P,ao;t), and
larly in the 100-1000 keV range. The prediction is that
the transport coefficients, DLL and ~aoao, will also be
protons should be the dominant ion species from 0.1 to
time variable reflecting the geomagnetic activity level.
I MeV, helium ions should dominate at 1 to several MeV
Based on limited data, Lanzerotti et al. [1978] estimated
per ion, and at higher energies oxygen (and possibly also
that the radial diffusion coefficient might vary with the
carbon) should dominate. These comparisons are done at
Kp geomagnetic index as
equal total ion energy. If comparison is made at equal
energy per nucleon, then protons would be the dominant
constituent at almost all radiation belt energies beyond a D(M)= 10-f L1O (5.96)
LL
few tens of keV.
with

5.5.2 Geomagnetic Storms t ❑ -(9.6-0.07 +,2 day Kp). (5.97)

Most radiation belt modeling deals with equilibrium

conditions appropriate for quiet time conditions. ‘There is It is not yet known how DLL ‘E) an d ~aoao might vary with
at the present time no fully comprehensive radiation belt geomagnetic conditions. It is known however that the
model for the effects of geomagnetic storms and other exospheric neutral density varies with activity (because of
disturbances. The reason is partially the incomplete the heating of the upper atmosphere), and the plasma-
knowledge of the radiation belt particle source mech- spheric densities certainly vary strongly. A first approach
anisms, and in part that disturbed time modeling is diffi- may be to solve the appropriate transport equations
cult because of the changes in magnetic field and time using perturbation theory for the different variables.
varying electric fields that must be also incorporated. Some geomagnetic conditions may, however, be too
Research is being conducted in this area, but no definite drastically altered to be treated as perturbations, so cau-
models are available as of this writing. tion is in order. At the present time there are many
Specific types of disturbances and associated time unknowns, and specific models have yet to be developed.
variability have, however, been analyzed. Among these
are studies of the post-geomagnetic storm decay of ener- 5.5.2.2 Magnetic Topology Variations. During the early
getic particle fluxes following the (yet not fully explained) phase of magnetic storms the earth’s magnetic field

5-26
 THE RADIATION BELTS

becomes compressed on the dayside. It has been considers the volume of space to be covered and the time
observed that the subsolar magnetopause can be pushed variation in the particle fluxes. Most of the data used in
inward from an average location of - 10RE to —5RE the NASA models were obtained in the 1960’s and early
during large storms. This implies a large B-field change. 1970’s. For example, the present AP-8 model for protons
Depending on the rapidness of the field change, the par- is a combination of 94 different instrument energy detec-
ticle population may respond adiabatically or non- tor channels from instruments flown on 24 satellites
adiabatically. However, most magnetic storms do not [Sawyer and Vette, 1976]. The combined data sets were
cause such a large perturbation of the magnetic field. smoothed to obtain omnidirectional flux distributions in
Increase in the lower energy (- 10-500keV) radiation the B, L parameter space. The fluxes were integrated
belt ion fluxes produces an enhanced ring current (see over all pitch angles and, therefore, directional informa-
Section 5.7) that depresses the earth’s magnetic field tion was not utilized. Other limitations in the data sets
earthward of the enhanced particle flux region, and are variations in instrument response and the lack of
causes adiabatic deceleration of the trapped radiation local time dependence considerations. Also note that the
belt particle fluxes. These effects must also be incorpo- flux models are compiled from a data base obtained over
rated into storm-time radiation belt prediction models. a brief time period in comparison with other geophysical
On the other hand, geoelectric fields penetrating into and solar time scales. Long term dynamical changes in
the trapping region during disturbed conditions [Harel et the radiation belts are not well understood so that
al., 198 1a, b] can cause cross-L non-diffusive transport extrapolation to different epochs must be done cau-
and thus adiabatic acceleration of the particle population tiously. For example, much higher fluxes than the mod-
[Lyons and Williams, 1980]. The relative influence of els give have recently been observed. One example of
these processes depends on particle energy. very long term effects is the adiabatic energization of
inner belt protons by the secular variation of the earth’s
magnetic field [Schulz and Paulikas, 1972].
5.6 EMPIRICAL RADIATION MODELS
5.6.1.1 Protons (Ions). It has become customary to
Since the discovery of the earth’s radiation belts, the refer to energetic protons located below L - 2 as inner
population of trapped particles has been measured with belt or inner zone protons. This is a misnomer, however,
ever improving instrumentation. The early Geiger coun- since the proton fluxes normally have only one radial
ters flown on the first few spacecraft had little or no par- flux peak. Sawyer and Vette [ 1976] have completed an
ticle identification capabilities. Subsequent instruments extensive model of the trapped “proton” environment out
used foil techniques, solid state detectors, magnetic to L 6.6 for energies above 100 keV. The measured ions
❑

and/ or electric deflection techniques and electronic signal were labeled “protons”, although no actual ion identifica-
discrimination. This led to a clear separation of electrons tion was made. The model is composed of two parts,
and ions (which were then assumed to be solely protons). AP8MIN and AP8MAX, which correspond to observa-
During the later years of space exploration the presence tions made during the 1960-1970 sunspot minimum and
of a muItitude of different ion species was established. maximum conditions, respectively. The difference be-
The purpose of this section is to provide a brief overview tween these two models is thought to result in part from
of existing radiation belt data, as a source guide for those differences in upper atmosphere heating during solar
who require crude numerical estimates of the radiation active periods such that the trapped particle collision
environment. It is emphasized that the older empirical rate (due to the exospheric expansion) was increased.
“proton” model in reality represents contributions from AP8MAX differs from AP8MIN for altitudes less than
many ion species, and that sometimes the heavy ion con- about 1000 km and for L values less than 2.9. Steep spa-
tribution may be dominant. tial gradients in the ion fluxes at lower altitudes are diffi-
cult to determine accurately.
An equatorial profile of the ion (proton) fluxes as
5.6.1 Data Acquisition and Processing given by AP8MIN is shown in Figure 5-31. Note the
Empirical radiation belt models are compiled by presence of >400 MeV protons for L <2. This is con-
NASA/ National Space Science Data Center, Goddard sistent with higher energy protons being produced by the
Space Flight Center in Maryland [Vette et al., 1979 and decay of neutrons produced in the atmosphere by cosmic
references therein]. These models represent a systematic rays (the CRAND source). The lower energy fluxes can
effort to compile many years of data containing a large arise from inward radial diffusion as discussed in the
number of disparate satellite observations into a few key theoretical modeling section.
models. These observations were separated in space and The AP8 proton models include no data after 1970
time, and made with highly varying instrumentation so and very little data above 100 MeV energies [Vette et al.,
that subjective judgments were necessary regarding data 1978]. Although the “proton” belt is considered far more
quality. The complexity of the task is appreciated if one static than the electron belts, significant enhancement

5-27
CHAPTER 5
AP8 MIN EQUATORIAL OMNIDIRECTIONAL RADIAL PROFILE
I
I ,
!
I 1
I I I

\\\\\\ \

3 4 5 6 7
1 2
L(RE )

Figure 5-31. Radial distribution of AP8MIN omnidirectional fluxes of protons in the equatorial plane with energies above threshold values between
0.1 and 400.0 MeV [Sawyer and Vette, 1976].

and depletions have been observed. A secondary equa- storm at L = 3.0 [Rothwell and Katz, 1973]. Beyond
torial proton peak (40- 110 MeV) of unknown origin was L = 5 earth radii, order-of-magnitude fluctuations occur
observed by McJlwain [1965]. This peak moved from on time scales as short as 10 minutes [Sawyer and Vette,
L = 2.25 to L = 2.1 earth radii between January 1963 and 1976]. Very large increases in MeV heavy ion fluxes at
January 1965. During the large May 1967 magnetic L = 2-5 were observed during the August 1972 magnetic
storm 2.2 – 8.2 MeV proton fluxes were observed to in- storm event [Spjeldvik and Fritz, 1981 a,b,c]. For this
crease more than a factor of 10 at L = 2.2 [Bostrom et al., reason engineering applications should consider large
197 1]. Lower energy (0.26-0.65 MeV) protons were deviations from the mean models.
observed to be preferentially enhanced during the same Off-equatorially mirroring ions intercept the upper

5-28
 THE RADIATION BELTS
atmosphere at higher equatorial pitch angles on lower L= 2.00
L-shells. For example, Figure 5-32 shows the B/ BO dis-
IOe
tribution for three different L-values just above the 00 = 0.0390
MeV

FLUX VS. B/B. DISTRIBUTIONS

L= I.17

BO= 0.1946

,f)o
.- Xu. .
IOQ 10’
Bt 00

Figure 5-32. Empirical inner zone radiation belt ~roton (ion) flux vs
B/ BO for three representative L-sheik, L= 1,17, ].50, and
J I i 1 1 2.0, and several representative energies as given by the
00 I .0; 1.10 1.15 1.20 1.25 AP8M IN proton model [Sawyer and Vette, 1976].
B!OO

L= 1.50
atmosphere. At L = 1.17 the equatorial pitch angle distri-
,06
butions have very steep loss cone gradients such that the
,0.25 MeV 00= 0.0923
1.50 MeV particle flux vanishes at aoc <73°. At L = 1.50 this “cut
L 8.50 MeV off” is aoc -37° and for L = 2.00, aoc -210. These
12 MeV cutoffs come about because the atmospheric bounce loss
/ 30 MeV
9 80 MeV cone is wide at the lower L-shells as seen in equatorial
l,,
E
u 170 MeV pitch angle.
% /
The South Atlantic Magnetic Anomaly has a control-
ling effect on the inner zone particle fluxes in the vicinity
of the loss cone. This anomaly arises from the earth’s
magnetic field being less intense at a latitude/longitude
region located near the coast of Brazil. Trapped particles
at these low L-shells will encounter their lowest mirroring
altitude (Hmin) and thus the densest atmosphere in this
longitude region. Figure 5-33 shows omnidirectional
proton ,flux cotttours in protons cm-z see-l MeV- 1 at
750 km altitude. These contours were derived from 5 to
7 MeV (dashed lines) and 28 to 45 MeV (solid lines) data
channels on the AFGL particle identifier instrument

,OOJ
1,0 i.2 I ,4 1.6 1.8 2.0
BIBO
2.2 2.4 2.6 2.0 :
flown on AF Satellite 72-1 in 1972. The data show that
the location of the proton peak flux in the South Atlan-
tic is dependent on energy. The 5 to 7 MeV peak is
located around L = 2 while the 28 to 45 MeV peak is
located around L = 1.3.

S-29
CH APTER 5
8

-----------
..-.-.--—
------
:-=---
~,,;;j.,,,-~~
;= —.
-80.0 , t
-180.0 -120.0 -W.o 0.0 60.0 Im.o 180
LONGITUOE
FLUX CONTOURS FOR H=750ENEffiY =35 MEV-

Figure 5-33. Proton isointensity flux contours asmeasured inthe South Atlantic anomaly atanaltitude of750km. Thesolid lines depict 2845MeV
proton (ion) fluxes and the dashed lines 5-7 MeV proton fluxes. The flux units are particles/( cm2-s-MeV).

The azimuthal drift around the earth of radiation belt radially inward. The observed ratio between helium ion
particles through the South Atlantic anomaly region and proton fluxes simultaneously observed are given in
produces a “windshield wiper” effect. Electrons present in Figure 5-35. Care must be exercised whether the ratio is
the enhanced loss cone created by the locally low B-field
region are lost. These electrons are subsequently replen-
ished by pitch angle diffusion during their drift around
the earth outside the anomaly.
Solar cycle variations and high altitude nuclear deto-
nations that modify the atmospheric composition signifi-

‘“
cantly ~affect the trapped particle populations. A 15-year
time study of inner belt 55 MeV protons concluded that
the observed flux variations
pected atmospheric
1981].
were consistent with ex-
loss processes [Parsignault et al.,
fftL2.25*’llL.2
“’IIL.27
5.6.1.2 Heavy Ibns. The abundance of trapped heavy
ions give clues to the origin of the radiation belt particles.
At high energies they also constitute a hazard to opera-
tional space systems. Over the last decade data have been
acquired to sketch the spectral and angular dependence
of the helium and carbon-nitrogen-oxygen (CNO) fluxes.
Helium [ohs; Helium ions are sometimes referred to
as alpha particles, although in a strict sense an alpha par-
ticle is only the totally ionized state (He2+) of a helium
ion. Figure 5-34 shows average equatorial helium ion
spectra over a range of L-shells during quiet-time condi-
tions [Fritz and Spjeldvik, 1979]. Notice that these spec- HELIUM ION ENERGY (MeV)

tra are very steep at L 24, become much harder with

lower L-shells, and are almost flat at L – 3. This is Figure 5-34. Energy spectra of equatorial radiation belt helium ions
deduced from mass ion observations on Explorer 45 dur-
expected since the lower energy helium ions are preferen-
ing the geomagnetically quiet period June 1-15, 1972. The
tially lost through the ion exchange mechanism and Cou- data are given at L = 2.25, 2.5, 2.75, 3, 3.25, 3.5, 4, 4.5,
lomb collision energy degradation, as the particles diffuse and 5 [Fritz and Spjeldvik, 1979].

5“30
 THE RADIATION BELTS
,~o strictly sinusoidal pitch-angle distribution did not com-
pletely describe the 98-240 keV/ nucleon ion fluxes at the

DBB,
L.2,25 L=2.5 L-2,75

,.-1 geosynchronous altitude (L – 6.6).

/
Carbon-Nitrogen-Oxygen (CNO): Trapped oxygen
,.-2 ions can at times be much more numerous than protons
particularly at L >5, when compared at equal total ion
,f3 energy. Figure 5-36 shows the radial flux profiles of car-
bon and oxygen ions as measured by the ISEE I space-

nR
It

,~+
craft [Hovestadt et al., 1978]. At equal total ion energies
in the MeV range the carbon to oxygen flux ratio is of
the order of 0.5, and at equal energy per nucleon the
L=3.25 L-3.5
carbon flux can dominate. This indicates an extraterres-
trial source for the very energetic trapped heavy ions
since the ionospheric C/O-ratio is < 105 [Blake, 1973].
The CNO flux pitch angle anisotropy is even more pro-
nounced than that of helium, having a value of the aniso-
tropy n-index typically between 12 and 16. At the higher
total energies (> 800 keV) and at geosynchronous alti-
tll[lll 1 1 1 II
tudes the CNO flux has been measured to be higher than
that of protons and helium.
Very Heavy Ions: Ions heavier than oxygen are also
present in the earth’s radiation belts in small quantities. It
has been demonstrated that substantial injection of ions
with nuclear charge Z 2 9 can take place during some
magnetic storms, and many orders of magnitude flux
intensity enhancements at E Z 10 MeV have been ob-
served lasting for many months [Spjeldvik and Fritz,
198 lc]. During such disturbed conditions the trapped
Emrgy p= lon ( MW )
fluxes of other ions (He, O, etc.) can also be greatly
enhanced [Spjeldvik and Fritz, 1981 a, b].
Figure 5-35. Quiet time He/ p ion flux ratios in—the equatorial radiation
.-
belts of the earth deduced trom bxplorer 4> observations
during June 1-15, 1972. The ratios are calculated from
flux observations at equal energy per ion. The shaded 5.6.1.3 Trapped Electrons. Empirical flux models have
areas depict the experimental uncertainty in the data
been developed that describe the inner and outer electron
[Fritz and Spjeldvik, 1979].
radiation belts. The National Space Science Data Center
defined at the same total ion energy or (as in this figure) at NASA/Goddard Space Flight Center has developed
at the same energy per nucleon. At higher total ion ener- two inner belt models: AE-6 [Teague et al., 1976] for
gies the heavy ion flux may be dominant over the proton sunspot maximum, and AE-5 epoch 1975 [Teague and
flux at some L-shells (Figure 5-35) while at the same Vette, 1974] for sunspot minimum conditions. There is
energy per nucleon proton fluxes usually dominate. also an outer belt electron model called AE-4. An
The equatorial helium ion pitch angle distribution is updated outer belt model, AE-7, is now under develop-
generally more anisotropic than the comparable proton ment. In this section a brief description of the models is
pitch angle distribution. For L-values between 2.5 and given and typical electron flux versus L-shell profiles,
4.5 the proton fluxes most often vary as sinnao where energy spectra, and pitch angle distributions are quoted.
n = 4 + 2 and a. is the equatorial pitch angle while for Inner Belt Electrons: The empirical data AE-5 model
helium ions n = 10 + 4. Beyond L -5 the helium fluxes is based on flux data from five satellites, OGO 1, OGO 3,
are quite dynamic and characteristic quiet-time values are 1963-38C, 0V3-3 and Explorer 26 [Teague and Vette,
difficult to define. During magnetic storms the fluxes of 1972]. This data base covered the period from December
energetic helium ions can increase by orders of magni- 1964 to December 1967, representing a transition from
tude in the heart of the radiation belts (L – 2-5), and the solar (sunspot) minimum towards maximum conditions.
relative abundance of the different ion species can vary. During this period the time-averaged Zurich sun spot
At higher L-shells substorm effects can be significant. number RZ ranged in value from about 10 (December
For example, during the 18 June 1974 substorm, helium 1964) to around 100 (December 1967). In const rutting
ions were more numerouk than protons at geosynchro- the AE-5 model it was assumed that the total inner belt
nous altitude for E > 800 keV/ ion [Fritz and Wiiken, electron flux is composed of four components: 1) quiet
1976]. Blake and Fennell, [1981] also have noted that a day flux at solar minimum, 2) quiet day flux at other

5-31
CHAPTER 5
E i I [ r 1 3 E 1 I i I 1 I 3
MPE/UoMd EXPERIMENT

E
M? E/ UoMd EXPERIMENT
ISEE -1 ISEE-1
2
CARBON ~z OXYSEN
IF 3 ULEZEQ
ULEZEQ Oti!Er
F’ERIOOS 7
OUIET PER:ODS
E 22 CCT 77-25 JAN 76

= Irrl

1. I I I i I I i
01 23 4 5 67
L—

L—

Figure 5-36, Equatorially trapped carbon and oxygen ion fluxes measured from October 1977-January 1978 during quiet times. Count rates may be
approximately convefled to omnidirectional flUX units: ions/ (cm2-s kev) by multiplying by 50/ AE where AE is the appropriate energy
passband in keV IHovestadt et al., 1981].

times during the solar cycle, 3) storm time flux and 4) quiet time values, and 3) the duration and characteris-
residual flux from the 1962 high altitude Starfish nuclear tics of the magnetic storms. The inner-belt electron flux
explosion. By late 1967 the Starfish generated energetic increases are infrequent but substantial and long-lasting,
electron flux at 6 < 1 MeV had decayed to insignificant so that a model storm effect was not extractable from
levels. It should be noted that for energies t >690 keV this NASA data base. However, the average long term
the available data for the inner belt AE-5 model were impact of magnetic storms was estimated by forming
quite limited. For example, within the observing range of the ratio R~ which is the average electron flux (June
the instruments the monthly averaged OGO 1 and OGO 1966- December 1967) divided by the quiet-time electron
3 data showed non-zero counts in this energy range only flux (October 1967). The results are shown in Figure 5-37
when the Starfish fluxes were still present or during and reflect a pronounced flux peaking at 1 MeV. Of
storm times. course, the largest relative storm time energetic electron
The quiet day solar cycle variation was defined by flux enhancements are found in the electron “slot-region”
taking the ratio of the omnidirectional flux measured located at L – 2-4, depending on energy.
from solar minimum to a standard reference epoch
An inner belt electron model for solar minimum was
(chosen as October 1967). Insufficient data necessitated
derived from the model discussed above [Teague and
the assumption of B-field (and therefore particle pitch
Vette, 1974]. This is called the AE-5 Epoch 1975 Pro-
angle distribution) independence for the geospace solar
jected Model and was constructed by removing the esti-
cycle variation. Also, the presence of Stafilsh residual
mated temporal variations. For example, the Starfish
energetic electrons restricted the analysis to later times
residual energetic electron flux component (L < 1.6 RE,
when these man made radiation belt electron fluxes had
E >700 keV) was removed, using the residual Starfish
decayed away. It should be emphasized that the solar
electron model of Teague and Stassinopolous [1972].
cycle variation has been determined for only one specific
cycle and that it may not apply to others. Similarly a solar maximum model, A E-6, was con-
The effect of magnetic storms on the time averaged structed using the AE-5 model at solar maximum values
inner radiation belt electron flux depends on 1) the fre- (epoch 1967) and with the estimated Starfish residual
quency at which magnetic storms occur, 2) the magni- (background) energetic electron fluxes also subtracted
tude of the storm time electron flux enhancement over out. This model is called A E-6 Epoch 1980.

5-32
 ‘lHM KADIA”l’lUN BEL”l’S

❑ L=2.8
Average Flux X L=2.4 1
June 1966. Dec. 1967 V L=2,2
RS =
O L=2,0
Quiet Day Flux Oct. ~
A L=l.9 .
,~,._ ~-. .
i ■ L=l.8 L= 1.4
100 –
~

.
\
: f
: )0, —

.-: ~
:
.
z
z
am ~
0
i
,
: j@ _—

10 – ❑ $
i

g _
~
:
~“ 6 OGO 5 Days 130-19D 1%8

A OV1-13 Days 139-1981968

0 0s04 Days 79-861966

,0 .–
o OGO 3 Days 250-3W 1966

● 0V3-3 Days 250-3801966

0 oGO 1 = Day 3001964

— Electron Model.
.5 6

, .—d..
10, 40
0% 03801 02 04 06 0010 10
~~ C ‘M, V)
26 28 30 32 34

loglo E (keV)
Figure 5-38. Empirical radiation belt electron fluxes at L = 1.4: Energy
1 1 1 I I I spectra comparison of the AE-5 1967 and AE-6 inner belt
03 0.5 10 1.5 2.0 30 electron models with experimental data taken from the
E (MeV) listed satellites for the times shown [Teague et al., 1979].

Figure 5-37. Estimate of the relative average effect of substorms on

inner belt electron fluxes (Teague and Vette, 1972). It is The quiet day inner belt electron fluxes for E <690
cautioned that different disturbances can have different keV at 1.3< L <2.4 can be represented by an analytic
effects. formula based on the empirical data [Teague and Vette,
1972]. The energetic electron flux is parametrized as
The electron models were verified by comparing them follows:
with additional data sets from the spacecraft 0V3-3,
OV1-13, OV1-19, OGO 5, OGO 1, OGO 3, 1963-038C, j (so, L, C) = A(ao, l-) E exp [-E/to(ao,L)] (5.98)
and OGO 4 [Teague et al., 1979]. These additional data
were also obtained before 1970 and substantiate the where L is the McIlwain L-shell parameter, a. is the
model mean values for that particular epoch. Figure 5-38 equatorial pitch angle and the parameters, A (ao, L),
shows ,a comparison between these data sets and the Eo(ao, L), are related to the equatorial values (a. 90°) ❑

AE-6 and AE-5 1967 models. This is the quiet time by the following empirical expressions:
energy spectra of equatorially mirroring electrons as
measured at L = 1.5. The OGO 1 day 300 1964 results are Sinm (~. – tYoLc)
A(~o, L) ❑ A(90°~ ‘) Sinrn ( @ - O!OLC) $> ffo>ac
evidently seriously contaminated by the Starfish detona-
tion residual energetic electrons. Otherwise, these results
show fair agreement between the AE-6 electron model = A(90° , L) 90°2a02@
and the different data. Figure 5-39 shows comparison of (5.99)
the model and measured equatorial pitch angle distribu-
tions at L 1.4. The AE-5 1967 (solar maximum) model
❑
60(tY0, L) ❑ 60(90°, L) sit-ancso/sinn @ ~>cso?ffc
and the AE-5 1975 (solar minimum) model bracket the
❑ CO(90”, L) 90°2 CY02 4.
data within a factor +2 to 3, except at very low equator-
ial pitch angles. (5.100)

5-33
CHAPTER 5
,.7

‘“”r’”-“N
T I I I 1
A

E AE-5 1967

m
\ 1
\
aoo- .04
,1

7.m - .25

6.CO -

L=I,4
5.m -
x
n
d

54.00 -

m -

‘1 O 0v3-3 711 keV Day 269-273, 1966

1
I X 0S0 4 7~ keV Day 60-90, 1968
2.CO -

I n OV1-13 670 keV Day 134-190, 1968

I — Model 575-775 keV

-– –- Stariish Model, Sept. 1%6 I.eo -

x

0
I.co 2.00 300 )
EQUATORIAL PITCH ANGLE a~(deg)
L

Figure 5-39. Empirical radiation belt electron fluxes: apitch angle dis-
tribution com~arison of the inner belt electron models Figure 540 Empirical radiation belt electron observations: Equatorial
with satellite “data taken at L=I.4 [Teague and Vette, flux versus L-shell for the AE5-MIN model as taken from
1974]. the National Space Science Data Center (NSSDC) com-
puter models. The flux is in units of electrons/ cm2-s), and
the energy range is 0.04-4 MeV.
Values for o, rsc, n, m, 60 (tYo = 90°, L) and A(ao ❑

90°, L) for a given L-shell are given in Table 5-2. Outer Belt Electrons: The outer belt tra~~ed electron
In Equations (5.99) and (5.100) the equatorial elec- fluxes are located between -3.5 and -11’ RE. Rapid
tron pitch angle distributions are fairly flat out to an changes in the magnetic field and background plasma
angle, ~, where they rapidly drop as sinn a. The parame- can modify the electron flux levels within minutes.
ter ~OLC is the minimum allowed equatorial pitch angle Because of the apparent coupling between magneto-
(the loss cone angle) and corresponds to a 100 km elec- spheric substorm process and the outer belt trapped elec-
tron mirroring altitude. These formulas should be used tron fluxes, time-averaged models have been developed.
with caution, however, since they represent extrapola- There are discernible changes in the average flux over the
tions based on an imperfect and incomplete data set. The time period studied so that the model is given for two
dipole approximation gives CSOLCin terrnS of BC from epochs. The NASA models are called AE-4 epoch 1964
Table 5-3 (solar min~.mum) and AE-4 epoch 1967 (near solar max-
imum) [Singley and Vette, 1972]. The data base was
acquired between 1959 and 1968 from 23 instruments on
BE 112
~OLC ‘ arcsin — (5.101) 11 satellites.
BCL3
[01 “ Because of the lack of azimuthal symmetry of the
geomagnetic field in the outer radiation zone, studies of
where Bc is the magnetic induction at the 100 km alti- the radiation belt electron structure beyond L – 5 re-
tude level. Figures 5-40 and 5-41 show the equatorial quires the conventional B-L coordinate system (calcu-
electron flux profiles as given by AE-5 (1975 projected) lated from the earth’s internal magnetic field) to be aug-
and AE-6. mented with the additional coordinate local time, LT.

5-34
E

THE RADIATION BELTS

Table 5-2. Flux model parameters for quietday inner belt electrons. See text for details [Teague and Vette, 1972]

Quiet-Day Model Parameters (Inner Belt Electrons)

Reference Pitch Angle = 90 Degrees

EPOCH = 10/67

J(ao = 90°) A(ao = 90° ) cO(a =900) M N PHI BC

(:E) (cm-2 see-l srl) (cm-2 see-] sr-l keV-l) (keV) (deg) (G)

1.30 1.20E07 1.71E03 83.7 2.80 0.670 67.1 0232

1.35 1.70E07 2.39E03 84.3 2.20 0.660 61.4 0.234
1.40 2.23E07 3.08E03 85.1 1.70 0.650 57.5 0.238
1.45 2.81E07 3.81E03 85.7 1.20 0.640 59.0 0.241
1.50 3.41E07 4.56E03 86.5 0.93 0.630 65.0 0.245
1.55 3.99E07 5.21E03 87.5 0.92 0.620 66.0 0.249
1.60 4.49E07 5.74E03 88.4 0.91 0.610 67.0 0.253
1.65 4.79E07 6.08E03 88.8 0.90 0.600 66.5 0.257
1.70 5.1OEO7 6.42E03 89.1 0.89 0.590 66.0 0.262
1.75 5.46E07 6.81E03 89.5 0.88 0.580 68.0 0.265
1.80 5.77E07 7.16E03 89.8 0.87 0.570 70.0 0.268
1.85 6.00E07 7.57E03 89.0 0.86 0.545 76.0 0.271
1.90 6.11E07 7.93E03 87.8 0.85 0.520 86.0 0.274
1.95 5.84E07 7.80E03 86.5 0.83 0.500 90.0 0.277
2.00 5.38E07 7.50E03 84.7 0.80 0.480 90.0 0.280
2.10 4.69E07 7.15E03 81.0 0.79 0.470 90.0 0.286
2.20 4.15E07 7.00E03 77.0 0.78 0.460 90.0 0.292
2.30 3.61E07 6.50E03 74.5 0.77 0.450 90.0 0.298
2.40 3.11E07 6.00E03 72.0 0.76 0.440 90.0 0.304

Theemuirical LT dependence of theouter radiation belt The parameters, BC and Boareall empirical func-
electron fluxes has been determined to be tions of L and are given in Table 5-3. The parameter B is
the magnetic field value at the desired location off the
geomagnetic equator, and BC is the value of the magnetic
log J (ao)
‘c(EL)cOs [m WI ‘5102) field at 100 km altitude on the same field line.
Figures 5-42 and 5-43 show the AE-4 equatorial
with LT in hours and C(E, L < 5) = O [Vette, et al., omnidirectional model electron fluxes from .04 to 4.50
1976]. This is only valid in a time average sense, and it MeV. Epoch 1964 represents solar minimum and epoch
was found that at a given local time the electron flux 1967 solar maximum. Using the above expressions with
intensity levels varied by at least factors of 10 to 50 over Table 5-3 and Figures 5-42 and 5-43, flux estimates can
the data acquisition period stated. be made at non-equatorial latitudes.
Given the equatorial flux (a. = 7r/2 or B= Bo) the off Measurements from the OV 1-19 satellite have indi-
equatorial outer belt integral electron flux (tYo # T/z) cated that the AE-4 model fluxes may be significantly
can be estimated by [Singley and Vette, 1972] too low, particularly at higher energies beyond 1 MeV
IVampoia, 1977]. These newer data were averaged over
J[>E, B, L]= J[>E, B= BO, L] GIB, LI (5.103) periods which included two magnetic storms in 1969. The
OV 1-19 instrumentation measured radiation belt elec-
and trons in the 53 keV to 5.1 MeV energy range in 24
differential energy bands, which significantly improved
Bc _ B ‘+1/2
the high energy data coverage over that which was avail-
G [B, L] = (B/ BO)-m
()
~
c o
;B<BC (5.104) able to construct the AE-4 models. The OV 1-19 data has
been incorporated into a new NASA model called AE-7
HI. Figure 5-44 shows the equatorial electron flux versus
GIBI L]= O ;B?BC. L-shell profile as predicted by the AE-7 HI model. Note,

5-35
CHAPTER 5
Table 5-3. Fiux model parameters for outer belt electrons. See text for AE6 - MAX
details ISingley and Vette, 1972].
9.00
I
AE-4 B/B Model Parameters (Outer Belt
Electrons)” am -

L B% BC
(RE) m (G) (G)
7.W-

3.00 1.12 0.01154 0.580

3.10 0.87 0.01046 0.582
600 -
3.20 0.71 0.009511 0.585
3.40 0.66 0.007929 0.588
3.60 0.63 0.006680 0.593 x
3500 -
4.00 0.60 0.004870 0.596
d
4.50 0.60 0.003420 0.599
o
5.00 0.60 0.002493 0.600 g
4,m -
5.50 0.60 0.001873 0.601
6.00 0.60 0.001443 0.601
6.50 0.60 0.001134 0.602
300-
7.00 0.60 0.000909 0.602
7.50 0.60 0.000739 0.603
8.00 0.60 0.000609 0.603
2.m-
8.50 0.60 0.000507 0.6035
9.00 0.57 0.000428 0.6035
9.50 0.52 0.000363 0.604
I.m
10.00 0.44 0.000312 0.604
10.50 0.35 0.000269 0.604
11.00 0.24 0.000234 0.604
o
t.m 2.00 3.00 4 o
*BO =0.311654 L
L3
Figure 5-4 I. Empirical radiation belt electron observations: Equatorial
flux versus L profiles for the AE-6 model as taken from
however, that no direct equatorial measurements have the (NSSDC) computer models, The flux is in units of
been included in the empirical model for electrons electrons/ (cmZ-s), and the energy range is 0.04-4 MeV.
(E> 300 keV) above 8000 km and below L = 5. Much of
range of L-shells and energies. Also plotted are the pre-
the data base was acquired from satellites orbiting at a
dicted electron lifetimes [Lyons et al., 1972] (see Section
significant inclination to the magnetic equator, making
5.5.1.2 on theoretical electron models) which in this
equatorial flux representations based on these data
comparison appear to agree with the experimental data.
uncertain.
Figure 5-45 shows the differences between the AE-4 5.6.1.4 Shell Splitting Effects. The outer belt particle
and the AE-7 models at energies above 1 MeV. The AE-7 pitch angle distributions are particularly interesting due
LO model is based on data taken on the AZUR satellite to a phenomenon called shell-splitting. Here we shall
and is shown in Figure 5-46. Note that 6< 1 MeV elec- qualitatively describe the physical process; for analytic
trons are most likely to penetrate spacecraft shielding and considerations, see Roederer [1970]. Shell-splitting arises
contribute to the accumulated radiation dosage and dam- from the lack of local time (azimuthal) symmetry of the
age. It is partially for this reason that the empirical model earth’s magnetic field, particularly at greater distances
uncertainties are of interest to spacecraft designers. from the earth. It is usually considered important for L
Magnetic storms may cause large energetic electron >4, but it should also exist at lower L-shells where the
flux enhancements that last for several weeks in the radi- earth’s magnetic multiples become significant. Shell-
ation belts. Figure 5-47 shows an example of an electron splitting arises because particles drifting in longitude pre-
flux enhancement as observed on OGO-5 [West et al. serving the first two adiabatic invariants modify their
198 1] during 1968. This example shows that the electron pitch angle and radial location according to the asym-
flux at 1.53 MeV increased by more than four orders metric magnetic field. While the concept of dipole L-shell
of magnitude during one particular magnetic storm. is useful to describe principal features of trapped parti-
Figure 5-48 shows the mean exponential decay time of C!SS, actual non-dipolar geometry with azimuthal asym-
these electron flux enhancements in days for a wide metry requires more generalized coordinates. One such is

5-36
 THE RADIATION BELTS

3.00 5.00 700 9.00 11.00

L ( EARTH RADII)
\\
IQO , Figure 543. Empirical radiation belt electron fluxes: AE-4 radial pro-
3.00 5.00 700 9.00 11.00 file of equatorial omnidirectional flux for different energy
L(EARTH RADII) thresholds, epoch f 967 [Singley and Vette, 1972].

Figure 5-42. Empirical radiation belt electron fluxes: AE4radia1 pro-

file of equatorial omnidirectionalflux for various energy encounter the magnetopause they may become lost and
thresholds, epoch f964[Singley and Vette, 1972]. there is a preferential depletion near a. = 90°. This gives
rise to the so-called outer zone btitterfl.v distribution
the (strictly non-invariant) McIlwain parameter Lm which is a pitch angle distribution with a minimum
[McIlwain, 1961; Stone, 1963]. Lm is defined as the around an equatorial pitch angle ~0 = ~/2. Figure 5-5 I
equivalent dipole L-shell of a test particle having the shows a survey of the energetic electron pitch angle dis-
same magnetic mirror field Bmj second adiabatic in- tributions in the near equatorial magnetosphere as mea-
variant, and energy as a corresponding particle in the sured by West [1979]. The butterfly distributions are
actual, nondipolar geometry (Figure 5-1). Shell splitting clearly predominant in the afternoon sector after the
can also result from asymmetric electric fields. For eastward (counterclockwise) drifting electrons have inte-
mathematical details see Schulz and Lanzerotti [1974]. racted with the magnetopause.
Figure 5-49 shows particles on the same L-shell at local Shell splitting also causes a coupling between pitch
noon in the noon-midnight meridian plane. When radia- angle and radial diffusion. Any type of pitch-angle dif-
tion belt particles drift around the earth to the midnight fusion may be accompanied by radial diffusion if the
sector they move to a lower L-shell and smaller equator- B-field is azimuthally asymmetric. The direction of the
ial pitch angles preserving their first adiabatic invarient radial displacement depends on the longitude at which
values. Those particles starting closer to the equator at pitch angle diffusion took place. Particles near the equa-
noon drift to lower L-shells at midnight. Conversely, tor that move to lower pitch angles on the dayside will be
Figure 5-50 shows the position of particles at local noon radially displaced further from the earth on the nightside.
having initially been on the same L-shell at local mid- Conversely, displacement to lower pitch angles on the
night. Those starting closer to the equator at midnight nightside leads to an inward particle flux on the dayside.
move closer to the magnetopause on the dayside. If they It is estimated that particles spend 2/3 to 3/4 of their

5-37
CHAPTER 5
108 I [ I I I I I I I I
7. 0

E
v

10’

●
o

v
. ●
,~5 A

\
10’ .\
\
Mar solar
Ma Ktm”m
( v
;
A
AZUR
Explorer
oGO 3
26
~ OGO 3
\
\
0 ATS 1
/ 0 [.plorer 6 \
● ATS 1 (1974)
A OGO 1 \v
A
OGO 1
$
Explorer lZ ,~ 3
O ATS 6 ( 1974)
V ATS 6 (1974) “ormal, zed to
\
Near Solar ● fxpbrer 14
0’.
M,n, mum ● ERS 13 L = 66 .stng AEI-7 cad!al
+ ERS 17 prof,le T.
HI Model
v IMP 1 \’\
AE-4 0967)
1 9 E“r,br,r 15 \ ‘,

_ AC-4
\ ___ AEI-7

\ ‘\
__. AKl-7 \ 102

LO Model \’
\ - \
\ \\
\
10‘ I 1 I \
1 2 3 4 5 6

E (MeV) E (MeV)
(a) (b)
Figure 5-44, Empirical radiation belt electron fluxes:
(a) Comparison of AEI-7 Model Spectra with a number of data sets at L=4. The HI model curve is mainly based on the OV 1-19
observations from Vampola.
(b) Comparison of AEI-7 Model Spectra with a number of data sets at L-6.6 [Vette et al., 1978].

drift period on the effective dayside so that pitch angle through radiation effects and spacecraft charging [Chap-
diffusion could lead to a net energy conserving outflow ter 7]. Many studies of the geosynchronous environment
of particles [Roederer, 1970]. First and second adiabatic have been made [for example, Paulikas and Blake, 1979;
invariant conserving inward radial diffusion as described Young, 1979; Garrett, 1979; Baker et al., 198 l; Mullen
in the theoretical modeling section would increase parti- and Gussenhoven, 1983].
cle energy with inward radial motion. After undergoing Near local midnight the magnetic field lines at geo-
many cycles of outward E-conserving diffusion and synchronous altitude often depart strongly from any
inward p-conserving diffusion, a significant local ener- resemblance of dipolar shape during magnetically active
gization of trapped particles could result [Schulz and periods. This effect is associated with changes in the pitch
Lanzerotti, 1974 and Theodoridis, 1968], but this needs angle distribution of the particle fluxes from being
to be investigated further. peaked perpendicular to the magnetic field line to a more
isotropic distribution. This and other flux changes have
been used as diagnostic devices to study underlying mag-
5.6.2 Geosynchronous Altitude Region netospheric processes [Higbie et al., 1978; Belian et al.,
(L= 6.6) 1978; Baker et al., 1978; Baker et al., 1980; Belian et al.,
Geosynchronous altitude is 3.6 X 104 km which corre- 1981].
sponds to an L-shell value of about 6.6. A satellite at this In this section the long term temporal behavior of
altitude in the plane of the earth’s equator will remain energetic (> 1 MeV) electrons and the plasma environ-
fixed over the same geographical location. This feature is ment is emphasized. Energetic electrons penetrate space-
highly useful for communication and surveillance satel- craft shielding and may cause radiation degradation of
lites. The natural geosynchronous charged particle envi- microelectronic components. The plasma environment,
ronment impacts the life-time and reliability of satellites of which the ions are an important component, modifies

5-38
 AE7-HI

Id

.
+ ,+.
. 2*.
●
E6 1530 keV
L = 3.00

I.*
:

.
b.

+.
t
●
.
I
+

*
●
✎

‘1”
. -..
‘.
..O

+, 1.
+. i

,0-1 I
300
, I
310 320
I
330
I I
340
I
3s0
I ,
360
I I I
370
1
3s0

01 I , t h i \ , I Oavs of Yesr- 196S

J
3W 4.CO Sm 6.W 7W am 9.m Krco IICQ 12m
L
Figure 5-47. Radiation belt electron observations: An intense injection
of energetic ( 1.53 MeV) electrons at L=3 as reported by
Figure 545. The NSSDC AE7-HI interim outer belt model for equa- West et al. [1981] during October 1968, Note the four
torial electron fluxes as a function of L. The listed energy
orders of magnitude increase and the subsequent expo-
is in MeV and the flux in units of electron /(cm2-s). The
nential decay.
discontinuous portions of these curves highlight the mod-
el’s areas of least accuracy.

9.
AE7-LO
Energetic trapped electron flux intensities (E > 1.0
T
MeV) at geosynchronous altitude have been shown to be
positively correlated with the average solar wind speed
[Paulikas and Blake, 1979]. The 3.9 MeV integral elec-
tron flux [J(> 3.9 MeV)], for example, has been
observed to vary by about a factor of 5 from a solar wind
speed of 400 km/see to one of 800 km/see. Lower energy
electron fluxes ( 140–600 keV), on the other hand show
little such correlation. Sufficiently long time averages (<
1 year) empirically remove the solar wind speed effects
and reflect the overall average stability of the electron
fluxes over longer time scales. For details see Paulikas
and Blake [1979].
Energy spectra of energetic electrons in the geosyn-
chronous altitude region measured with the SCATHA
spacecraft are shown in Figure 5-52 [Mullen and Gus-
senhoven, 1983]. These data represent 75-day averages
I .’~ taken between February 1979 and February 1980, and on
Q’J3.m 4.m 5.W 6.~ 7.CU 8.OQ 12CM
L
the average may be represented by a power law spectral
dependence. Integration of the fitted curves gives integral
Figure 5-46 The NSSDC AE7-LO interim outer belt model for equa-
flux levels that are consistent with the AE-4 and AE-7
torial electron fluxes as a function of L. The listed energy
is in MeV and the flux in units of electrons/(cm2-s). The models (Figures 5-42, 43, 45 and 46). This implies that
discontinuous portions of these curves highlight the model’s the long term temporal averages of the electron fluxes at
areas of least accuracy. geosynchronous altitude did not materially change dur-
ing the 1970’s. The scatter of the individual SCATHA
the voltage to which a satellite will charge. Results from flux data measurements about the mean time-averaged
the SCATHA (Spacecraft Charging at High Altitude) flux is substantial. At times the observed electron fluxes
satellite show that the ion composition at geosynchro- differed from the mean values by well over an order of
nous is a function of magnetic activity and local time magnitude. The flux models, therefore, should be used
[Mullen and Gussenhoven, 1983]. with caution.

5-39
CHAPTER 5
10 30 m m w m

k“ “’’’”l “’ ‘Y’”1 “ “’”q r. I /

. t

Figure 549. Theoretically computed shell splitting effects for particles

starting on common field lines in the noon meridian. Dots
represent the particles’ mirror points. The curved lines
give the position of mirror points for constant equatorial
pitch angle a. [Roederer, 1970].

so

I
“K m.. . ‘\
0.s 7 7

L Figure 5-50. Theoretically computed shell splitting effects for particles

starting on common field lines in the midnight meridian
[Roederer, 1970].

● LLL expect to find 1.4–2.6 MeV electrons with a flux inten-

o 0V3-3 2.25 sity greater than 10 electron cm-z see-1 sr- I keV) 50% of
I LTK-theoq the time in the 5.5–6.0 L-shell region. On the other
hand, from the same bar graph fluxes above 100 electron
1 , I 1 I [11111 I cm-2 See- 1 Sr- 1 keV are expected less than 10~o of the
10’ Id ld time.
Ions are a dynamic component of the radiation
Energv - keV
environment at geosynchronous altitudes. The ion com-
position at low energies varies with magnetic activity as
Figure 5-48, Radiation belt electron observations: Pitch angle diffusion
lifetimes at constant L were in part derived from the data illustrated in Figure 5-54 [Mullen and Gussenhoven,
in tbe previous figure, and the predicted electron precipi- 1983]. During magnetically active periods (high Kp) the
tation lifetimes are those of Lyons et al. [1972] [West et
0+ component becomes enhanced relative to protons. An
al., 1981].
explanation for 0+ enhancement [Kaye et al., 1981;
Fennell et al., 1981] is that the storm enhanced 0+ ions
It is useful to estimate the percentage of time that are accelerated up along the magnetic field lines from the
the electron flux will exceed a certain value. Figure 5-53 auroral ionosphere, while the protons probably also orig-
shows the cumulative probability distribution for the inate from the magnetotail plasma sheet that moves
SCATHA 1.4 – 2.6 MeV energy channel in four L closer to the eaoh during magnetically active periods.
intervals. For example, from this figure one would Oxygen ion enhancements also increase the particle

5-40
 THE RADIATION BELTS
=3 HIGH ENER6Y ELECTRON
Equatorial pitch angle CUMULATIVE OIS~BUTIW
. distributions (1419-2603 KeV)
@fl
100 100
L=60-70 go L)80
90 1
— 80
—
70
mfim
pM : 60
E 50
-M m E
u 40 ~ 40
— L 30 n 30
20 20
-20 -15
10 10
‘GSM
+ [
— ,.-2

— FLUX (elec/cm2-sec-sr- keV ) FLUX (elec /cm2-sec-w-keV )

— —
~m
——
90
— BO
F_
70
— m Butterfly
fi—
r’1 ~ 60
em–_ — Isotropic
— : 50
–w
f 40
—.
n 30
20

I I

Figure 5-51, Survey of energetic electron pitch angle distributions I62

~1 ,.O ,.l ,.2 ~3 102 103
observed in the near equatorial magnetosphere [West et FLuX (elec/cm2-sec- sr-keV) FLUX (elec/cm2-$w-w-kcV)
al., 1979].

Figure 5-53. Cumulative probability that the high energy electron flux
SC3 AVERAGE ELECTRON FLUX
(at energies 1419-2603 keV) is less than the shown levels.
, ~3 VERSUS ENERGY
This figure gives an estimate of the “spread” in the indi-
POWER LAW FIT
[ vidual measurements used to obtain Figure 5-5 I [Mullen
% and Gussenhoven, 1982].
‘\
~.
&
x., ‘\ x 2.5 !.5
& \\ ‘.
‘\\ ‘.K ‘\
‘\\
‘\ ‘\\ ‘\
1
K
\
\ ‘h, ‘\ 3 L=5.75
%. x,
\\\\’\ x L= 5-6
\ ix ‘\ ‘~
=6.25 2.0 / !.0
\\\ L ● L= 6-7 –
‘\ 1, 1,
+ L=7-8
b, \
h /
\ 1, L =6.75 A L= S- 0.5
\
\ L,
‘\ ~ L =7.25
“x L =7.75 1.5 .5

t m-
,0-, ~
‘E ~
e E
.5 .6 .7 .8 .9 I 2 345 + Ss

ENERGY (MeV)
glo / /’ .0

J/0/’”F
/
c / /’ ,/ 4 g
=
/’ /“
Flgure 5-52. Time-averaged energetic electron energy spectra measured
near =.
zeosvnchronous altitude from February 1979 to Feb- / --4
ruary 1980. These curves represent a 75 day average and .“

k
.5 x .5
approximate a power law curve. Individual data show,
however, that deviations two orders of magnitude from
“-&--J
these means are not uncommon [Mullen and Gussen-
hoven, 1982].
//
+“
I I I
energy density relative to that of the magnetic field. Fig- o 2460246
ure 5-55 [Mullen and Gussenhoven, 1982] shows the geo- UP UP
synchronous altitude ratio ~ of the particle energy deri-
sity to the magnetic field energy density as a function of Figure 5-54. Average Oxygen (left) and hydrogen (right) number densi-
ties es determined from the Lockheed ion composition
local time. The individual curves represent various levels measurements on the SCATH A spacecraft (for energies
of Kp where higher K p is a measure for higher magnetic 1-32 keV) versus K for various L-shell intervals [Mullen
activity [Jacobs, 1970]. A ~ value greater than unity and Gussenhoven, f 982].

5-41
CHAPTER 5

5.7.1 Electrical Current Relations

The ring current itself produces a magnetic field that
is superimposed on the earth’s magnetic field. Enhance-
ment in the ring current constitutes the cause of the
magnetic field depression observed at ‘mid-latitudes on
the surface of the earth during the main and recovery
phases of geomagnetic storms as well as magnetic field
enhancement beyond L — 6-7, as illustrated in Figure
5-3. For this reason we shall consider the magnetic effects
of the trapped particles.
As one can see from Equations (5.36) and (5.37), the
particle gradient<urvature drift velocity is proportional
to the particle energy. With the equatorial pitch angle,
cro, one may write

vd=AqB3 (1 +COS2~o)(B X v IBI) , (5.105)

where electric field and gravity effects have been disre-

garded. Here B is the magnetic field due t~ the m~in
(earth’s internal) dipole moment and M = M O where O is
a unit vector in the magnetic northward direction. At the
magnetic equator

(5.106)
~
LOCAL TIME For simplicity in the presentation, we shall consider
equatorially mirroring particles only, that is, ao = T/2;
F]gure 5-55. Time-average of the ratio of the particle energy density to for the mathematical treatment of the general case of an
the magnetic field energy density (Bz/8m) as a function of
local time. Higher KP implies higher magnetic activity.
arbitrary pitch angle ao, see Dessler and Parker [1959].
These data represent over 90 days averages at of geo- We then get
synchronous altitude (L=6.65) obtained during February
1979 to February 1980 [Mullen and Gussenhoven, 1982].
Vd =-c% qM
r2$. cqBE
* L2 $ (5.107)
implies that the particles are not strongly confined by the
magnetic field. A ~ value significantly less than unity
suggests particle confinement in this region. These results where $ is a unit vector in the eastward azimuthal direc-
indicate that the plasma processes that are operative dur- tion around the earth. From Maxwell’s equations (the
ing active periods at geosynchronous altitude are sub- Biot-Savart law; [Jackson, 1975]) the magnetic field
stantially different than those operative during quiet generated by the drift motion of each particle is
times. Garrett [1979] gives more details of the geosynch-
ronous plasma environment. 27rr
Bd. -i $4 (5.108)
5.7 THE RING CURRENT Jo
The gradient-curvature drift of radiation belt particles
where i is the magnitude of the single-particle drift
causes differential motion that is mass and charge
“current”:
dependent: electrons drift eastward and positive ions drift
towards the west. This constitutes an electrical current
around the earth in the westward direction, called the (5.109)
extraterrestrial ring current. During geomagnetic distur-
bances, such as magnetic storms, the population of
trapped particles at – 1-800 keV energies is substantially and thus
enhanced on L-shells between L = 3 and L = 6. As a
consequence, the ring current is intensified and magnetic
(5.110)
disturbances at the earth result.

5-42
 THE RADIATION BELTS

The minus sign indicates that the particle azimuthal drift where fi(~o, L, C) is the distribution function for particle
generated field opposes the main (internal) dipole field species i, expressed as function of equatorial pitch angle,
earthward of the ring current particle population. L-shell and energy.
There is also a magnetic effect of the particle’s spiral Parker [1957] developed a hydromagnetic formalism
motion around the field lines. Each gyro-loop may be that provides an alternative to this extensive integration.
considered a small dipole moment p = ~ (since fOr a~ = One may define the macroscopic particle pressures in the
T/2, E = 6L), and the associated magnetic field is direction parallel and perpendicular to the local magnetic
field direction
Bg=~=z=
r3 Br3
6 -~,
M
(5.111)
BER~ Pll = z Fi (X, V, aO)mvzcos’2aOdvda0 (5.117)
i 1
which is in the direction of the internal dipole. The total
perturbation at the origin due to a single equatorially pl=;~ Fi(X, V, ao)mvzsinzaodvdao (5.118)
mirroring particle is then J
i

AB,=Bd+Bg”-$. (5.112) where Fi is the particle distribution function for a particle

species i expressed in position, (X), speed, and pitch
angle coordinates. The summation is extended over all
At the equatorial surface of the earth the unperturbed particle species. The magnetic field pressure is
(internal) dipole field iS just BE ❑ # so that
E
p=~. (5.119)
m 87r
AB, 2e 26RE3 (5. 113)
—=. —= —— !
BE B2 R3 M2 With these pressure expressions the total gradient-
EE
curvature drift current can be written as

and noting that the total energy in the earth’s unper-

turbed dipole field above the earth’s surface may be writ- P~
Id =&B” 2Pm Vpm+~(B. P): ,
ten as {– }

(5.114) (5.120)

where c is the velocity of light [Williams, 1982]. The cor-

We may express the relative ring current single particle responding gyration current of the particle distribution as
perturbation as a whole is

P~
_ B. vp~-~vpm
4=-$. (5.115) lg = 8:P {
BE m
(5.121)
It turns out that this expression is valid for trapped par-
ticles in the radiation belts regardless of the equatorial
pitch angle a. [Dessler and Parker, 1959]. In defiving
Equation (5. 115) it was assumed that the total energy in
The two terms within the brackets of Equation (5.120)
the ring current is less than the magnetic field energy UM.
stem from the magnetic field gradient and field curvature
When that is not the case Equation (5.115) is no longer
respectively, and the three terms within the brackets of
strictly valid and may be in error by up to a factor of two.
By summing up the effects of all the individual parti- Equation (5. 121) represent currents driven by the particle
pressure gradient, the magnetic field gradient, and the
cle motions in the geomagnetic field, one arrives at the
magnetic field line curvature. The total current of all par-
total magnetic field perturbation
ticles then reduces to
rrj 2 max t max

AB=~ sinaodao dL d~ AB~ fi(ao, L, 6)

J ?J
io 1 Crnin

(5.116) (5. 122)

5-43
CHAPTER 5
As a rule of thumb, it requires a total of 4 x 1022 ergs of the H+ lifetimes are shorter than that of He+ and 0+ life-
particle kinetic energy to produce a surface magnetic times, and the opposite may be true. Figure 5-56 shows a
field depression of - 100nT (1 nT Iy 1O-5G). ❑ ❑ relative comparison of ring current ion flux observations
during four different time periods. These results pertain
5.7.2 Composition and Sources to E < 20 keV energies while the ionic composition at
The population of trapped particles is made up pri- higher energies remains to be investigated observationally.
marily of electrons, protons, helium ions, carbon ions, Ionospheric or atmospheric ions probably form a sig-
and oxygen ions. The composition is found to vary sub- nificant fraction of the ring current population. This is
stantially with location (such as L-shell) and with geo- inferred from the observations reported by Shelley et al.
magnetic conditions. The heavier ions, such as He+ and [1974] and Sharp et al. [1976]. Low-orbiting polar satel-
0+, may be dominant during disturbed conditions on L- lites detected the precipitation of oxygen ions (L ❑ 6.8)
shells in the range L 3 to L 5, while prolonged quiet
❑ ❑ during distrubed times and also detected field-aligned
periods tend to favor H+ (protons) above tens of keV energies. upward moving accelerated ions from the auroral iono-
The latter is also a reasonable expectation since at typical sphere. Sufficient pitch angle scattering at higher alti-
ring current energies (– 70 keV) the charge exchange life- tudes could cause these u~ward moving ions to become
time of H+ is longer. At lower energies, below a few keV, stably trapped and form part of the ring current.

w ,0-10

If
1 J, 1 1 I 1 1
1
L-VALUE 2 4 6 8
GM, LATfTUDE’ .::~ 2.10 ;56; 13.00 ~g:~
LTG 16.2 15.8
I I
PRoMlcs
79-02-13 78-11-03
-so -so
- 10”7 - (c) (d)
T -W DST (nl)
E @ Y
u

J 10-e -

g
g
,“
8 10-9 -

g
w
z
w 10’0 -

I 41
~
L-VALUE ~,e
fT~ LATITUDE’ 3.60 lo.@ j3.~
12.5 12.3 12.3 11.9 11.9

Figure 5-56. Observed energy densities of H+, He+ and 0+ ions in the radiation belts at ring current energies (0.2-17 keV/ion) versus orbital parame-
ters. Tbe data were derived from the PROM ICS experiment on the PROGNOZ-7 spacecraft. The dashed curves indicate apparent
energy densities calculated from a total ion (E/q) spectrometer (at 0.1-45 keV) under the assumrrtion that onlv Drotons were measured.
The;ksults show the importance of the heavier ions at different times and locations. The four panels represe”n~ data for four different
periods, and the D,,-index history is also shown as a guide to the ring current activity [Lundin et al., 1980].

5-44
 a
THE RADIATION BELTS

Based on total ion (no mass resolution) observations, fluctuate substantially for other geomagnetic conditions
it has become clear that the greatest contribution to the for which ID,t I generally remains less than 50 nT.
ring current comes from –20 -200 keV ions where the Figure 5-58 shows an example of the D,t index plotted
mass composition is yet unknown. This is illustrated in for June-December 1972, and the occurrence of four
Figure 5-57 [Williams, 198 1]. magnetic storm periods in June, August, September and
October/ November is evident. The D,t index is therefore
very useful to identify magnetic storms from surface
“~
magnetogram records.

5.7.3 Adiabatic Effects Produced

by the Ring Current
Much of the time the magnitude of the electrical cur-

‘!~[:?l1
rent set up by the azirnuthally drifting radiation belt par-
ticles changes slowly in comparison with the ion drift
period. Therefore, the third adiabatic invariant, which is
proportional to the enclosed magnetic flux threading the
I
~ drift path, is most often conserved following injection or
o JI
E(keV) 10 100
I I ‘m acceleration. S6raas and Davis [1968] have shown that
c~ significant adiabatic effects will take place even for mod-
-i
erate values of Dst. To separate the adiabatic and non-
Figure 5-57. Integral representation of the ring current energy density adiabatic features one can transform to a Dst O refer- ❑

as a function of ion energy. The contribution to the ring ence. If j 1 (f 1, L I) is the equatorially mirroring flux for
current energy density is greater where this curve is steep-
est. This is also where currently the ionic composition is E = El (L Ll, Dst = O) and j2 (62, L2) is the correspond-
❑

unknown [Williams, 1981 ] ing flux for Dst # O then one has

J2(C2,L>) ❑ (~2/cI) ji (EI, LI) . (5.123)

A practical measure for the overall strength of the

This follows from Liouvilie’s theorem which states that
extraterrestrial ring current is the D~t-index which mea-
the phase space density is constant along dynamical par-
sures the middle latitude spatially averaged decrease in
ticle trajectories (f = const. = j/ pz). The equatorially mir-
the horizontal component H of the earth’s surface mag-
roring energy E2 is mapped from the unperturbed ~I
netic field: Dst = <AH>. Under this definition the quiet
energy by conservation of the first adiabatic invariant:
time ring current corresponds to Dst = O. Hourly values
of the Dst index are published by NASA/National Space
6’2 = e, (B2/B~) , (5. 124)
Science Data Center, Goddard Space Flight Center,
Maryland. Magnetic storms generally have D~t depres-
sions on the order of 100 to 200 nT (very large storms where B 1 is the value of the quiet time magnetic field
may exceed IDst I = 300 nT), and the Dst index may also induction.

D=- INDEX FOR MYS 153-366 ( JUNE I - mCEMBER 31) 1972

100] I I I I I I I I I I I I I I I I 1 I I 1 I 1 I I I I I I I 1 I I I I I I I I I 1
I I

-300 I I I 1 I I I I I I I I I I I 1 I I 1 I I I 1 1 I I 1 1 I I I 1 I I I I I I I I I
1601mlm lw20021D 220230 2402602602702K) 290300310 w330340m 360

Figure 5-58. An example of the time-variation of the horizontal magnetic field component (D,,) at the equator. The large rapid drops in D,t corre-
spond to a build up of the ring current during magnetic storms followed by subsequent decay.

5-45
CHAPTER 5

For a dipole field where BE = 0.312 G

BE
Bl=~ (5.125)
1
and 0

-0.2 .
B2=~ + AB(r) . (5. 126)
‘2
-0.4 .
The magnetic flux enclosed by the drift path (the third
adiabatic invariant) is given by -0.6 .

LI w b
- -0.8 .
I

-1.0
(5.127)
-1.2 .
and
-1.4 .
L2
P
@2. -2 TBE/L2+2r A B(r) rdr , -1.6
J
0
(5. 128) EARTH RADII

Figure 5-59. The radial dependence of the ring current magnetic field
where A B(r) is the magnetic field change induced by the used in the calculation of adiabatic effects on trapped
~rotons by S6raas and Davis [1968]. Notice that the
ring current enhancement [Soraas and Davis, 1968]. decrease is-greatest near LA. -
Equation (5. 128) uses the fact that the integral from O to
L includes the return magnetic flux through the earth so ,.8
I I I I
I I I I
that the net dipole magnetic flux through the entire equa- ‘ --- UNCORRECTED

torial plane is zero. This fact allows the O to L1 integra- -.< — CORRECTED
‘. DST = ’47Y
--- -..
tion interval to be replaced by one for L 1 to @. The ions - 107 - ,’ --- >);., ‘\
will now equatorially mirror at L2 where by equating @1, -x \\\\ ‘\ ‘.
~.. Ile. ‘\\\\ ‘\\
and 62 7. y ‘\ \\\\ \\\\ \
L , ‘\ \\\\\\ \ .’”
\;\ >134KeV
\
~g 106 \ t, \ ‘\
BE/L2 z BE/L1 + A B(r) rdr (5.129) \ ,\ \\ 180 KeV
J l,, ,\\
0 z \ 51~KeV”,
o 1100KeV ,’ \ I b
and E ;;% 11* 345KaV
& 105 II
\ I 775KeV
~
m
A B(r) = 0.7 D,t f(r) . (5. 130) A I}00 KeV B,
=

“04=
The function f(r) is shown in Figure 5-59. For a given
D~t, L2 can be found, and hence, B2. Knowing B2, E2 can
be determined. In Figure 5-60 we show an illustrative
example as presented by S6raas and Davis [1968]. 2.0 3.0 4.0 5.0 6.0
R ~ (Earth Radii)

5.8 RADIATION EFFECTS ON SPACE Figure 5-60. The radial proton integral energy intensity profile as meas-
SYSTEMS ured on day 109 of 1965 when D,, = 46nT, together with
the transformed profile corresponding to D,( = O assuming
Among the known effects of particle radiation on the three adiabatic invariants of motion conserved. The
radial dependence of the ratio between the magnetic field
space systems are spacecraft charging phenomena and
after and before the build up of the ring current and the
effects of penetrating radiation on materials. Specific radial movement AR of the particles are shown in the
effects include detector malfunction and degradation, lower part of the figure [S6raas and Davis, 1968].

5-46
 THE RADIATION BELTS
optical system degradation, memory system alteration,
and control system malfunction or failure. For manned
space operations, biological effects are a major concern.
A crude measure for damage done by penetrating
energetic radiation is radia[ion dosage which is measured
in rad.s. This unit is defined as an energy deposition of
100 ergs (6.25 x 107 MeV) in one gram of a material sub-
stance. This definition does not distinguish between dif-
ferent kinds of incident radiation or different effects on
the material. Radiation dosage is thus only an overall
measure, and it is often necessary to examine specific
interaction cross-sections when studying radiation effects.
Energy is deposited through chemical (molecular
bond changes, bound electron excitation, and ionization)
and nuclear (element transmutation, nuclear excitation,
and induced radioactivity) interactions. The macroscopic
effects are evident in device failure after a critical level of
radiation exposure is reached. Most often this critical
level depends directly on the nature and energy character-
istics of the incident radiation.
A major concern is the on orbit lifetime of microelec-
tronic devices that are designed to a specific level of radi-
ation “hardness” (such as 104-105 rad). There is in many
cases a trade off between orbit choice and system lifetime
that must be determined.
A lowest order approximation to the expected radia-
tion exposure effects can be estimated by combining the ELECTRON ENE~ ( MEV)

energy deposition rate versus incident energy curves of

Janni and Radke [1979] with the expected radiation belt Figure 5-61. Model electron spectra for engineering applications:
Normalized integral omnidirectional electron flux as a
flux intensity deduced from previous observations, as in function of electron energy. By matching a model energy
Section 5.6 or from theoretical modeling. A simple (but spectra to the curves the appropriate E. can be quickly
very crude) approach to estimating the radiation dosage found. The flux at I MeV is normalized to I electron/cm2
[Janni and Radke, 1979].
follows.
It is assumed that shielding is equal in all directions A similar technique also applies to ions. For protons,
so that .a spherical shield approximation can be used. Figure 5-63 shows a similar set of exponential spectral
The shielding is also assumed to be aluminum or close to radiation curves extending to 300 MeV, and Figure 5-64
it in density. The incident omnidirectional particle fluxes gives the radiation exposure dosage as function of the
are normalized (or scaled) to unity at a selected energy so aluminum shielding thickness. Notice that the shielding is
that dosage need only be calculated as a function of spec- generally less effective in reducing the radiation dosage
tral shape. The dose rate for a given energy spectrum is due to the very energetic ions in the radiation belts.
found by multiplying the resulting dosage by the model
(or measured) omnidirectional flux at the selected energy.
Total dosage is determined by integrating over the 5.8.1 Detector Malfunctions
expected exposure time of the satellite. Single particle upsets occur when a single particle
Figure 5-61 shows a number of exponential spectral creates enough free electrons to simulate a device logic
radiation curves normalized to unity at 1 MeV for ener- state change. Single particle effects are particularly severe
getic electrons up to 6 MeV. By picking the one curve in small (<10 pm) sensitive regions of microelectronic
that most nearly approximates the actual expected radia- devices. Figure 5-65 illustrates the incidence of a cosmic
tion energy spectrum, one has a one-parameter spectral ray (very energetic heavy ion) in a single memory cell
represeritation, the spectral e-folding energy EO. One then commonly used for onboard information storage. Notice
proceeds to calculate the radiation exposure for the nor- that the volume where the ionization takes place is at
malized spectrum behind a certain thickness of shielding least comparable to the sensitive cell region itself. It is
by using the curves in Figure 5-62. This process has to be presently an area of controversy whether reduction in cell
averaged over the expected radiation conditions for the size will always increase the soft error or single event
expected spacecraft orbit during the period of the desired upset rate; it is conceivable that with very small memory
orbital operations. cells the ionization volume could encompass many cells.

5-47
CHAPTER 5
DOSE VERSUS SWERICAL WIELD THICKNESS

lo’~

4
Idz I I 1 1 8 I I 1 9 I 1 1
0 Ico Zm m 400 m m
ALUMINUM THICKNESS ( MILS)

Figure 5-62. Radiation dosage experienced behind various shielding

thicknesses of aluminum absorber for several incident
electron energy spectra normalized to I eiectron / cmz at
1 MeV. Multiplication by the I MeY flux gives the
expected radiation dose rate [Janni and Radke, 1979].

P~TON ENERGY ( MEV)

On the average, the effect of single-particle incidence

Figure 5-63. Model proton spectra for engineering applications: Nor-
on detector materials, such as aluminum and silicon, is malized integral proton flux curves for estimating E,,
the generation of one electron-hole pair per 3.6 eV of where the flux at 30 MeV is normalized to 1 proton/ cm2
energy deposition. Thus, 1 rad of incident radiation in 1 [Janni and Radke, 1979].

gram of material creates 1.74 x 1013 electron-hole Pairs.

Even a moderately energetic radiation belt particle (for
example 1 MeV) will create a large number of free charge
carriers in the detector material and may lead to false
signals. Prolonged exposure to energetic particles degrades
the detector performance by the accumulation of mate-
rial micro structural damage. For example, solid state
detectors of the A l-Si-Au variety are found to have a
factor of 10 increase in useful lifetime when the alumi-
num side is facing the radiation exposure (compared to
the gold side). Very energetic and very heavy cosmic ray
ions have a particularly devastating effect on detector
systems; for further details see Adams and Partridge
[1982] and McNulty [1981].
Figure 5-66 shows an example of a nuclear interac-
tion occurring near a sensitive region of a radiation par-
ticle detector or other solid state device. An incident
proton, for example, will stimulate a zgSi nucleus to emit
an alpha particle which has a short range. The recoiling
28Si nucleus stops in even a rnur,h shorter distance. The
~o,~ o 100 2C0
ALUMINUM
300 400
THICKNESS ( MILS)
500 600

combined effect can deposit tens of MeV in a small

(10 X 20 X 20) ~m volume element creating a “soft” Figure 5-M, Radiation dosage experienced behind various shielding
(data) error. Accelerator data [McNulty et al., 1980] thicknesses of aluminum absorber for several incident
proton spectra normalized to I proton/ cmz at 30 MeV.
indicate that the proton induced (E >40 MeV) soft error Multiplication by the omnidirectional flux at 30 MeV will
rate is less than 10–6 soft errors/ (protons-cmZ). An upper give the radiation dose rate [Janni and Radke, 1979].

5-48
 THE RADIATION BELTS
=~ RE-WAT~ OF SENSITIVEREGION deposition process. Similar effects can result from alpha-
IN A SlffiLE MEMORY C.SLL / particle emission from nuclear interactions and from
natural and induced radioactivity in the devices them-
selves.
If the effects are infrequent in occurrence, engineering
design emphasizing redundance of the critical compo-
nents could circumvent the problem. However, when the
effects are frequent and /or persistent this approach may
not be feasible.
Figure 5-67 [E. Petersen, personal communication,
1982] shows the energy deposited in a 10 pm — thick
Figure 5-65. Illustration of radiation effect: Electron-hole generation
sensitive region by different ions over a range of incident
near a sensitive region as a result of local ionization
produced by a traversing cosmic ray or energetic particle energies. The vertical scale on the right denotes the
[B. Blake, personal communication, 1982]. number of electrons produced. As a circuit becomes
smaller and more complex, less deposited energy (charge)
is needed to trigger errors.

.01

[ .005
; 01 1.0 10
KINETIC ENERGY (MeV/NUCLEONl

Figure 5-66 Illustration of the production of numerous secondary par-

ticles from the nuclear interaction of the primary particle Figure 5-67. Energy deposited in 10 pm of silicon by different ions
in a sensitive volume of a solid state detector device Tbe scale on the right shows the number of free electrons
[McNulty et al., 1980]. released. The low energy part of the curve occurs when
the ion penetrating range is less than 10 ~m so that it
deposits all its energy [E. Petersen, personal communica-
tion, 1982].
limit to the expected error rate (error/see), therefore, can
be found by using the proton flux models for E >40
MeV times 10-6. If trapped heavy ions are sufficiently 5.8.3 Control System Failure
abundant they could dominate the soft error rate [Adams Radiation induced errors in electronic circuitry can
et al., 1981]. This is one of several purely practical moti- be particularly damaging when they occur in critical cir-
vations for measuring the energetic trapped particle cuitry such as control systems or in decision making
composition to an adequate degree of accuracy. logic. While other non-critical circuits may continue to
function with false information, control systems can
latch-up, that is, be switched into an undesired mode
5.8.2 Memory Alteration from which there may be no reset option, or the space-
Certain microcircuitry used in current spacecraft craft may be damaged. Certain circuitry switching may
instrumentation has proven very susceptible to the effects cause burnout of electrical systems or even worse effects,
of energetic heavy ions in the radiation belts and in the particularly when propulsion, attitude, or weapons sys-
cosmic radiation. Memory chips and microprocessors are tems may be involved. For these reasons it is imperative
frequently found to have their logical states and informa- that proper safeguards and redundancy design be consid-
tion content severely altered by the localized energy ered in the early stages of spacecraft engineering.

5-49
CHAPTER 5

5.8.4 Biological Effects and shock processes. High altitude and space detonations
(more than 100 km above the earth) have the fireball size
There is extensive literature on space biology [see for
limited by the magnetic field. This occurs because the
example Bacq and Alexander, 1961]. Here we shall only
explosion generates electrically charged fragments that
point out that the quiet time radiation belts at some loca-
are susceptible to the magnetic force, q’V x B, where q is
tions present a lethal radiation dosage to a man in a
the particle charge and V its velocity. A nuclear detona-
space suit or even within a vehicle. But even outside the
tion of 1 megaton can have fireball expansion to -1000
main trapped radiation zone, there are intermittent high
km across the magnetic field when the B-field has a value
fluxes of solar energetic particles. For example, it is
B = 0.5 G. The expansion is not magnetically limited
believed that the energetic particle fluxes associated with
along the field lines.
the August 1972 solar flare/ magnetic storm event would
Depending on the location of the nuclear detona-
have been extremely harmful to humans almost any-
tions, a certain fraction of the neutrons will decay within
where in space. The method presented above can also be
the magnetic field trapping region, and the decay prod-
used to estimate human radiation exposure behind dif-
ucts will thus constitute artificially created trapped radi-
ferent shielding designs. A definitely lethal dosage is
ation. For nuclear fission, the fission fragments also
about 500 rads [Desrosier and Rosenstock, 1960].
emit particles (such as electrons and a-particles) before
reaching a nucleonic configuration as a stable isotope.
This process further contributes to the trapped radiation,
5.9 MAN’S IMPACT ON THE RADIATION and the characteristic electron energy is 1 to 8 MeV from
BELTS this source.
The activity of mankind can, to a significant degree, Studies of nuclear detonation effects have shown that
influence the earth’s radiation environment. Examples even small high altitude explosions (in the kiloton range)
are nuclear detonations (fission and fusion), accelerator affect the radiation belts considerably. See for example
particle beams (neutral and charged), release of chemical reviews by Hess [1968] and Walt [1977]. Table 5-4 gives
substances, injection of metallic powders, and electro- an overview of the known radiation belt effects of the
magnetic wave energy production. The effects of some of Teak, Orange, Argus 1, Argus 2, Argus 3, and Stafilsh
these modification sources have not yet been studied, but nuclear detonations carried out at high altitudes by the
for others a substantial body of data is available. United States, and the USSR-I, USSR-2 and USSR-3
high altitude nuclear detonations by the Soviet Union.
Energetic charged particles exiting the upper atmos-
5.9.1 Nuclear Detonations phere along the geomagnetic lines of force are generally
A vast number of free neutrons and other particles within the atmospheric bounce loss cone. In the absence
and a great pulse of electromagnetic energy are released of significant pitch angle scattering, such particles will
in nuclear explosions. In the nuclear fission process, the follow the field lines and precipitate into the conjugate
fission fragments also carry significant kinetic energy. hemisphere. Empirically, a significant fraction of the
The product of nuclear fusion is generally a stable particle nuclear detonation particles become trapped in the radia-
(He) which may be ionized. As a rule of thumb –IOZC tion belts. This implies that significant pitch angle scatter-
fast neutrons (each of which decay into a proton- ing must take place from the angular source cone region
electron-neutrino triplet) are released per megaton nuclear (a. < @oLc) to stably trapped panicle orbits (a. > @oL.C).
explosive yield. This pitch angle scattering must take place on the time
The size of the nuclear fireball depends not only on scale of a single half-bounce period ~b/ 2 (which is of the
the explosive yield, but also on the medium in which the order of seconds).
detonation occurs. In field-free empty space the fireball Following such an artificial injection of particles into
will expand without limits, but in the presence of mate- the radiation belts, the normal radiation belt radial and
rial substances or a magnetic field the fireball is effec- pitch angle diffusion mechanisms will operate. The initial
tively restrained. In a dense gas (such as below -100 km narrow injected radial distribution will broaden, and the
in the earth’s atmosphere), collisions between the explo- charge exchange (for ions) and Coulomb energy degrada-
sion products and the atmospheric constituents dissipate tion mechanisms will modify the characteristics of the
much of the detonation energy as heat. About half of this injected distributions. Depending on the location, the
energy is radiated away and the thermalized remainder is artificial radiation belts may last for days or years [Walt
typically at 6000 K to 8000 K [Zinn et al., 1966]. At an and Newkirk, 1966; Stassinopoulos and Verzariu, 1971].
altitude of 60 km in the earth’s atmosphere, a 1 megaton The effects of accelerator beams are likely to be sim-
fissional detonation will have a fireball radius of –4 km, ilar to those of the nuclear detonation particles, but the
and for the same nuclear explosive yield this radius will yield (in terms of number of particles) is likely to be
be smaller close to the ground. The fireball itself may much smaller. On the other hand, since the beam parti-
accelerate to velocities of several km/ sec due to buoyancy cles may be generated over a wide range of energies

5-50
 H

THE RADIATION BELTS

Table 5-4. Listing of high altitude nuclear detonations between 1958 and 1962 [Walt, 1977]

Approximate
Altitude Time L-Value Characteristics Approximate
Event (km) (UT) Date Latitude Longitude Detonation Yield of Band ‘ Decay Time

Teak 76.8 10:50:05 01 AUG58 17” N 169° W 1.12 MT Range Low Altitude few days
Orange 42.97 10:30:08 12AUG58 17° N 169° W 1.12 MT Range Low Altitude I day
Argus I 200 2:30:00 27 AUG58 38° S 12” W 1.7 I-2 KT Narrow Band 0-20 days
Argus 2 250 3:20:00 30AUG58 50° S 8° W 2.1 I-2 KT Narrow Band 10-20 days
Argus 3 500 22:10:00 06 SEP58 50° S IO”w 2.0 1-2 KT Narrow Band 10-20 days
Starfish 400 09:00:29 09JUL62 16.7” N 190.5” E 1.12 1.4 MT Wide Distribution I-2 yrs
USSR I — 03:40:46 220CT62 — 1.8 Wide Distribution 30 days
USSR 2 — 04:41:18 280CT62 — 1.8 Wide Distribution 30 days
USSR 3 — 09:13 01 NOV62 — 1.75 Narrow Band 30 days

(thermal to relativistic) a more precise study of their whistler-mode waves from VLF radio transmitters can
effects is warranted. perturb the energetic electron component of the earth’s
radiation belts. Correlative studies indicate that energetic
5.9.2 Release of Chemicals electron precipitation not only occurs from natural
sources [Spjeldvik and Lyons, 1979] but also in correla-
Chemical releases for research purposes have been
tion with strong terrestrial radio transmitter operations
carried out at high altitudes. In most cases barium or
IVampola and Kuck, 1978; Park et al., 1981; Imhof et
lithium was released to trace magnetic field lines locally
al., 198 1; Chang and Inan, 1983]. Precisely to what
and to assess the magnitude of electric fields and upper
extent man’s electromagnetic wave generation influences
atmosphere winds. Chemical releases into outer regions
the overall radiation belt structure is not known,
of geospace are also planned. Such programs may mod-
however.
ify the environment locally (for example, by altering
plasma wave dispersion characteristics), but are not
expected to impact the radiation belts seriously unless
large quantities of chemicals are used. 5.9.4 Effects of Space Structures
Extensive operations with rocket propulsion or spe- Proposed operations of large manmade metallic and
cial ion engines could, however, drastically alter the dif- electrically insulated space structures will produce local
ferent particle populations and could lead to profound “singular” regions in the magnetosphere. Associated with
changes in the radiation belt structure. See Chapter 7. To space shuttles, space platforms, or space power arrays
date no comprehensive environmental impact analysis will be a hydromagnetic wake in which the wave and par-
has been carried out. ticle behavior will go through a sudden change. It is not
known whether or not these cavity phenomena may have
5.9.3 Transmission of Radio Waves a significant effect on the radiation belts themselves. For
It has been suggested that electromagnetic wave some details see Garrett and Pike [1980] and references
energy from tropospheric thunderstorm activity and therein.
CHAPTER 5
REFERENCES

Adams. J .H., Jr. and K. Patridge, “Do Trapped Heavy Ions Claflin, E. S., “Change Exchange Cross Sections for Hy-
Cause Soft Upsets on Spacecraft?’” National Research drogen and Helium Ions Incident on Atomic Hydrogen:
Laboratory Memo. Rept. 4846, October 12, 1982. 1 to 1000 keV ,“ Rep Tr-0059 (6260-20)-1, Aerospace
Adams, J. H., Jr., R. Silverberg, and C.H. Tsao, “Cosmic Corp., El Segundo, Cal if.. 1970.
Ray Effects on Microelectronics, Part 1: The Near-Earth Cornwall, J. M., “Diffusion Processes Influenced by Con-
Particle Environment,” National Research Laboratory jugate Point Wave Phenomena,” Radio Sci., 3: 740,
Memo. Rept. 4506, August 25, 1981. 1968.
Akasofu, S.-1., Polar and Magnetospheric Substorms, D. Cornwall, J. M., “Radial Diffusion of Ionized Helium and
Reidel, Dordrecht, Holland, 1968. Protons: A Probe for Magnetospheric Dynamics,” J.
Alfv6n, H. and C .G. Falthammar, Cosmical Electrody- Geophys. Res. 77, 1756, 1972.
namics, Clarendon Press, Oxford, UK, 1963. Davidson, G. T., “The Motion of Charged Particles in the
Axford, W. I., “Magnetospheric Convection,” Rev. Geo- Earth’s Magnetic Field,” in The Trapped Radiation
phys., 1:421, 1969. Handbook, edited by J. B. Cladis, G.T. Davidson, and
Bacq, Z.M. and P. Alexander, Fundamentals of Radiology, L. L. Newkirk, Lockheed Palo Alto Research Labora-
Pergamon Press, New York, 1961. tory, DNA 25241H, 1977.
Baker, D. N., P.R. Higbie, E.W. Hones, Jr., and R.D. Desrosier, N. W. and H. M. Rosenstock, Radiation Tech-
Belian, “High Resolution Energetic Particle Measure- nology in Food, Agriculture and Biology, Avi, West-
ments at 6.6 RE 3. Low Energy Anisotropies and Short- port, Corm., 1960.
Term Substorm Predictions,” 3. Geophys. Res., 83:4863, Dessler, A.J. and E. N. Parker, “Hydromagnetic Theory of
1978. Geomagnetic Storms,” J. Geophys. Res., ~ 2239, 1959.
Baker, D. N., P.R. Higbie, and R.D. Belian, “Multispace- Falthammar, C. G., “Radial Diffusion by Violation of the
craft Observations of Energetic Electron Flux Pulsations Third Adiabatic Invariant,” in Earth’s Particles and
at 6.6 R~,” J. Geophys. Res., 85:6709, 1980. Fields, edited by B .M. McCormac, Reinhold, New York,
Baker, D.N. et al., “The Los Alamos Geostationary Orbit 1968.
Synoptic Data Set,” Los Alamos National Laboratory Fennell, J. F., D.R. Croley, Jr., and S.M. Kaye, “Low-
Report LA-8843, UC-34G, Los Alamos, N. M., August Energy Ion Pitch Angle Distributions in the Outer Mag-
1981. netosphere: Ion Zipper Distribution ,“ J. Geophys. Res.,
Belian, R. D., D.N. Baker, P.R. Higbie, and E.W. Hones, 86:3375, 1981.
Jr., “High-Resolution Energetic Particle Measurements Fite, W. L., R.F. Stebbings, D.J. Hummer, and R.T. Brack-
at 6.6 RE 2. High Energy Proton Drift Echoes,” ~. Geo- man, “Transfer in Proton-Hydrogen Atomic Collisions,”
phys. Res., 83:4857, 1978. Phys. Rev., 119:663, 1960.
Belian, R. D., D.N. Baker, E.W. Hones, Jr., P.R. Higbie, Fritz, T.A. and W .N. Spjeldvik, “Simultaneous Quiet Time
S.J. Bame, and J .R. Asbridge, “Timing of Energetic Observations of Energetic Radiation Belt Protons and
Proton Enhancements Relative to Magnetospheric Sub- Helium Ions: The Equatorial a/p Ratio Near 1 MeV ,“
storm Activity and Its Implication for Substorm Theo- J. Geophys. Res., 84:2608, 1979.
rks,” J. Geophvs. Res., 86: 1415, 1981. Fritz, T.A. and B. Wilken, “Substorm Generated Fluxes of
Blake, J. B. “Exper-imental Test to Determine the Origin of Heavy Ions at the Geostationary Orbit,” in Magneto-
Geomagnetically Trapped Radiation,” 3. Geophys. Res., spheric Particles and Fields, edited by B. M. Mc-
78:5822, 1973. Cormac, D. Reidel, Dordrecht, Holland, 1976.
Blake, J. B., and J .F. Fennell, “Heavy Ion Measurements Garrett, H. B., “Quantitative Models of the O to 100 keV
in the Synchronous Altitude Region,” Planet. Space Mid-Magnetospheric Particle Environment,” in Quan-
Sci., 29:1205, 1981. titative Modeling of Magnetospheric Processes, edited
Bostrom, C. O., D.S. Bean, and J.C. Armstrong, “Time by W.P. Olson, AGU, Washington, D. C., 1979.
History of the Inner Radiation Zone ,“ Models of [he Garrett, H. B. and C. P. Pike, (eds. ) Space Systems and Their
Trapped Radiation Environment, Vol. 111: Longterm interactions }vith Earth’s Space Environment, AIAA
Variations, NASA SP-3024, 1971. Press, New York, 1980.
Brinkmann, H.C. and H. A. Kramers, “Zur Theorie der Hamlin, D. A., R. Karplus, R.C. Vik, and K.M. Watson,
Einfangung von Elektronen Durch a-Teilchen,” Proc. “Mirror and Azimuthal Drift Frequencies for Geomag-
Akad Wetensch. Amsterdam Afd. Natuurk, 33:973, 1930. netically Trapped Particles ,“ J. Geophys. Res., M: 1• 1961.
Chandrasekhar, S., “Plasma Physics” (compiled by S.K. Harel, M., R.A Wolf, P.H. Reiff, R.W. Spiro, W.J. Burke,
Trehan), Phoe)lix Science Series, University of Chicago F.J. Rich, and M. Smiddy, “Quantitative Simulation of
Press, Chicago, 1965. a Magnetospheric Substorm 1. Model logic and over-
Chang, H.C. and U.S. Inan, “Quasi-Relativistic Electron view,” J. Geophys. Res., 86:2217, 1981a.
Precipitation Due to Interactions With Coherent VLF Harel, M., R.A Wolf, R.W. Spiro, P.H. Reiff, C.K. Chen,
Waves in the Magnetosphere,” J. Geophys. Res., 88: W.J. Burke, F.J. Rich, and M. Smiddy, “Quantitative
318, 1983. Simulation of a Magnetospheric Substorm 2. Compar-
Chapman, S. and J. Bartels, Geomagnetisrn, Oxford Uni- ison With Observations, ” J. Geophys. Res., 86; 2242,
versity Press, London, 1940. 1981b.

5-52
 THE RADIATION BELTS
Hess, W. N., The Radiation Belt andh4agnetosphere, Blais- flection of Whistlers,” Planet. Space Sci., 18: 1753,
dell, Waltham, Mass., 1968. 1970.
Higbie, P. R., R.D. Belian, and ~.N. Baker, “High-Res- Lyons, L. R. and R. M. Theme, “Equilibrium Structure of
olution Energetic Particle Measurements at 6.6 R~ 1. Radiation Belt Electrons,” J. Geophys. Res., 78:2142,
Electron Micropulsations,” J. Geophys. Res., 83:4851, 1973.
1978. Lyons, L. R., R.M. Thome, and C.F. Kennel, “Electron
Hovestadt, D., G. Gloeckler, C.Y. Fan, L.A. Fis~, F.M. pitch-Angle Diffusion Driven by Oblique whistler Mode
Ipavich, B. Klecker, J.J. O’Gallagher, and M. Scholer, Turbulence,” J. Plasma Phys., 6:589, 1971.
“Evidence for Solar Wind Origin of Energetic Ions in Lyons, L. R., R.M. Thome, and C.F. Kennel, “Pitch Angle
Earth’s Radiation Belts,” J. Geophys. Res. Letts., 5: Diffusion of Radiation Belt Electiohs Within the Plas-
1055, 1978. masphere,” J. Geophys. Res., 77:345,5, 1972.
Hovestadt, D., B. Klecker, E. Mitchell, J.F. Fennell, G. Lyons, L. R. and D.J. Williams, “A Source for the Geo-
Glaeckler, and C.Y: Fan, “Spatial Distribution of Z ~ 2 magnetic Storm Main Phase Ring Current,” J. Geophys.
Ions in the Outer Radiation Belt During Quiet Condi- Res., 85:523, 1980.
tions,” Adv. Space Rei., 1:305, 1981. MacDonald, J .R. and F. W. Martin, “Experimental Elec-
Imhof, W .L., E.E. Gaines, and J,B. Reagan, “High-Res- tron-Transfer Cross Sections for ~ollisians of Oxygen
olution Spectral Features Obsewed in the Inner Radia- Ions in Argon, Nitrogen and Helium at Energies of 740
tion Belt Trapped Electron ~apulatian ,“ J. Geop~ys. MeV,” Phys. Rev. (Sect. A), 4:, 1965, 1971.
Res., 86:2341, 1981: McIlwaitt, C. E., “Coordinates for Mapping the Distribution
Jackson, J .D., Classical Electrodymrnics, 2nd edition, Wiley, of Magnetically Trapped Particles ,“ J. Geophys. Res.
New York, 1975. 66:3681, 1961.
Jacobs, J. A., Geomagnetic Micrapulsations, Springer-Ver- McIlwain, C. E., “Redistribution of Trapped Protons During
lag, New York, 1970. a Magnetic Storm,” Space-Research V, North Holland,
Janni, J. and G. Radke, “Tlie Radiation Environment and Amsterdam, 1965.
Its Effects on Spacecraft in Quantitative Modeling of McNulty, P. J., R.C. Wyatt, G.E. Farrell, R.C. Filz and
Magnetospheric Processes, ” edited by W .P. Olson, P.L. Rathwell, “Proton Upsets in LSI Memories in
Geophysical Monagraph 21, AGU, Washington, D.C. Space,” in Prog. Astronaut. Aeronaut. Val. 71, Space
1979. Systems and Their Interactions with Earth’s Space En-
Kaye, S. M., E.G. Shelley, R.D. Sharp, and R.G. Johnson, vironment, edited by H.B. Garrett and ~. P. Pike, AIAA
“Ion Composition of Zipper Events,” J. Geophys., 86: Press, New York, 1980.
3383, 1981. Matt, N.F. and H. S. W. Massey, The Theory of Atomic
Kennel, C.F. and F. Engelmann, “Velocity Space Diffusiari Collisions, Clarendan Press, Oxford, UK, 1952.
from Weak Plasma Turbulence in a Magnetic Field,” Mullen, E.G. and M.S. Gussenhoven, “SCATHA Environ-
Phys. Fluids, 9:2377, 1966. mental Atlas,” AFGL TR-83-.0002, ADA13 1456, 1938.
Kennel, C.F. and H.E. Petschek, “Limit on Stably Trapped Park, C. G., C. S. Lin, and G. K. Parks, “A Graund-Satellite
Particle Fluxqs,” J. Geophys. Res., 71:1, 1966. Study of Wave-Particle Correlations,” J. Geophys. Res.,
Kennel, C. F., L.J. Lanzeratti, and E.N. Parker, (eds.) Solar 86:37, 1981.
System Plasma Physics, Vols. 1–111, North Holland, Parker, E. N., “Newtonian Development of the Dynamical
Amsterdam, 1979. Properties of Ionized Gases of Law Density,” Phys.
Kivelson, M. G., S .M. Kaye, and D.J. Sauthwood, “The Rev., 107:924, 1957.
Physics of Plasma Injection Events,” Dynamics of the Parsignault, D. R., E. Holeman, and R.C. Filz, “Solar Cycle
Magnetosphere, edited by S.-I. Akasofu, D. Reidel, Modulation of’ the 55-MeV Protan Fluxes at Low Al-
Dordrecht, Holland, 19$0. titudes?” J. Geophys. Res., 86: 11439, 1981.
Lanzeratti, L. J., D.C. Webb, and C. W. Arthur, “Geomag- Paulikas, G.A. and J.B. Blake, “Effects of the Solar Wind
netic Field Fluctuations at Synchronous Orbit 2. Radial on Magnetospheric Dynamics: Energetic Electrons at the
Diffusion,” J. Geophys. Res., 83:3866, 1978. Synchronous Orbit ,“ in Quantitative Modeling of Mag-
Lerche, I., “Quasilinear TBeary of Resonant Diffusion in a netospheric Processes, edited by W. P. Olson, AGU,
Magneto-Active Relativistic Plasma,” Phys. Fluids, 11: Washington D.C. 1979.
1720, 1968. Retterer, J. M., J.k. Jasperse, and T.S. Chang, “A New
Lo, H.H. and W .L. Fite, “Electron-Capture and Loss Cross Approach to Pitch Angle Scattering in the Magneto-
Sections for Fast, Heavy Particles Passing Through sphere,” J. Geophys. Res., 88: 201, 1983.
Gases,” Atomic Data, 1:305, 1970. Roederer, J. G., Dynamics of Geomagnetically Trapped Ra-
Lundin, R., L.R. Lyons, and N. Pissarenka, “Observations diation, Springer-Verlag, New York, 1970.
of the Ring Current Composition at L-Values less than Rossi, B. and S. Olbert, Introduction to the Physics of
4,” Geophys. Res. Lett., 7:425, 1980. Space, McGraw-Hill, New York, 1970.
Lyons, L. R., “Plasma Processes in the Earth’s Radiation Rothwell, P.L. and L. Katz, “Enhancement of 0.24+.96
Belts,” in Solar System Plasma Physics, edited by C.F. MeV Trapped Protans During May 25, 1967 Magnetic
Kennel, L.J. Lanzerotti, and E.N. Parker, North Hol- Storm,” J. Geophys. Res., 78t 5490, 1973.
land, Amsterdam, 1979. Sagdeev, R.Z. and A.A. Galeev, Nonlinear Plasma Theory,
Lyons, L.R. and R.M. Thome, “The Magnetospheric Re- W.A. Benjamin, New York, 1969.

5-53
CHAPTER 5
Sawyer, D.M. and J .1. Vette, “AP-8 Trapped Proton En- After Effects in the Middle Latitude D-Region,” J. At-
vironment for Solar Maximum and Solar Minimum,” mos. Terr. Phys., 37: 777, 1975.
NSSDC/WDC-A-R&S 76-06, NASA-GSFC TMS-72605, Stassinopoulos, E.G. and P. Verzariv, “General Formula
December 1976. for Decay Lifetimes of Starfish Electrons,” J. Geophys.
Sctiulz, M., “Geomagnetically Trapped Radiation,” Space Res., 76: 1841, 1971
Sci. Rev., 17:481, 1975a. Stebbings, R. F., W .L. Fite, and D.G. Hummer, “Charge
Schulz, M.,, “particle Saturation of the Outer Zone, A Non- Trarisfer Between Atomic Hydrogen and N+ and O+ ,“
Linear Model,” Astrophys. Space Phys., 29:233, 1975b. J. Chem. Phys., 33: 1226, 1960.
Schulz, M. and L.J. Lanzerottij, Parficle Di~usion in the Stix, T. H., The T~eo~ of Plasma Waves, McGraw-Hill,
Radiation Belts, Springer-Verlag, New York, 1974. New York, 1962.
Schulz, M. and G.A. Paulikas, “Secular MagnEtic Variation Stone, E. C., “Physical Significance and Application of L,
and the Inner Proton Belt ,“ J. Geophys. Res., 77: 744, B. and R,, to Geomagnetically Trapped Particles,” J.
1972. Geophys. Res., 68:4157, 1963.
Sharp, R. D., R.G. Johnson, and E.G. Shelly, “The Mor- Stormer, L., Polar Aurora, Clarendon Press, Oxford, 1955.
phology of Energetic O + Ions During Two Magnetic Teague, M. J., K. W. Chan, and J.I. Vette, “AE-6: A Model
Storms, Temporal Variations,” J. Geophys. Res., 81: Environment of Trapped Electrons for Solar Maximum,”
3283, 1976. NSSDC/WDC-A-R&S 76-04, May 1976.
Shelley, E. G., R.G. Johnson, and R. D. Sha~, “Morphol- Teague M. J., N.J. Schofield, K.W. Chan, and J.I. Vette,
ogy of, Energetic O + in the Magnetosphere ,“ in Mag- “A Study of, Inner Zone Electron Data and Their Com-
netospheric Physics, edited by B. ~. McCormac, D. parison with Trapped Radiation Models,” NSSDC/GSFC,
Reidel, Dordrecht, Holland, 1974. August 1979.
Singley, G. W. and J. I. Vette, “The AE-4 Model of the Teague, M.J. and E.G. Stassinopoulos, “A Model of the
Outer Radiation Zone Electron Environment ,“ NSSDC Starfish Flux in the Inner Radiation tine,” NASA-GSFC
72-06, 1972. TMX-662 11, December 1972.
Solov’ev, E. S., R.N. 11’in, Y.A. Oparin, I.T. Serenkov, Teague, M .J. and J .1. Vette, “The Inner Zone Electron
and N. V, Federenko, “Capture and Loss of Electrons M@el AE-5 ,“ NSSDC/WDC-A-R&S 72-10, November
by Fast Nitrogen and Oxygen Atoms and Ions in Air, 1972.
Nitrogen and Oxygen,” Sov. Phys.-Tech. Phys., 37: Teague, M.J. and J .1. Vette, “A Model of the Trapped
777, 1975. Electron Population for Solar Minimum,” NSSDCiWDC-
Soraas, F. and L.R. Davis, “Temporal Variations of the 10 A-R&S-74-03, NASA, April 1974.
keV to 1700 keV Trapped Protops Observed on Satellite Theodoridis,, G. C., “Bimodal Diffusion in the Earth’s Mag-
Explorer 26 During First Half of 1965,” NASA TMX- netosphere: 1. An Acceleration Mechanism for Trapped
63320, GSFC, August 1968. Particles,” Ann. Geophys., 24:944, 1968.
Spjeldvik, W. N., “Equilibrium Stticture of Equatorially Tinsley, B. A., “Evidence that the Recovery Phase Ring
Miiroring Radiation Belt Protons,” .I. Geophys. Res., Current Consists of Helium Ions,” J. Geophvs. Res.,
82:2801, 1,977. 81:6193, 1976.
Spjeldvik, W .N., “Expected Charge States of Energetic Ions Vampola, A. L., “A New Study of the Outer Zone Electron
in the Magnetosphere,” Space Sci. Rev. 23:499, 1979. Environment, A Hazard to CMOS,” SAMSO-TR-77-
Spjeldvik, W .N. and T.A. Fritz, ‘&Theory for Charge States 127, June 1977.
of Energetic Oxygen Ions in the Wh’s Radiation Belts,” Vampola, A.L. and G.A. Kuck, “Induced Precipitation of
J. Geophys. Res., 83: 1583, 1978a. Inner Zone Electrons,” J. Geophys. Res., 83: 2542,
Spjeldvik, W .N. and T.A. Fritz, “Energetic Ionized Helium 1978.
in the Quiet Time Rddiation Belts: Theory and Com- Vette, J., I. King, W. Chan, and M.J. Teague, “Problems
parison with Observation,” J. Geophys. Res., 83:654, in Modelling the Earth’s Trapped Radiation Environ-
1978b. ment ,“ AFGL-TR-78-O 130, ADA059273, 1978.
Spjeldvik, W .N. and T.A. Fritz, ‘~Observations of Energetic Vette, J. I., M.J. Teague, D.M. Sawyer, and K.~. Chan,
Helium Ions in the Earth’s Radiation Belts During a “Modelling the Earth’s Radiation Belts” in $olar-Ter-
Sequence of Geomagfietic Storms,” J. Geophys. Res., restrial Prediction Proceedings, Vol. 2, edited by R. F.
86:2317, 1981a. Donnelly, U.S. Department of Commerce, NOAA,
Spjeldvik, W.N. and T.A. Fritz, “Energetic Heavy Ions Boulder, 1979.
with Nuclear Charge Z > 4 in the Equatorial Radiation Walt, M., “History of Artificial Radiation Belts” in The
Belts of the Earth: Magnetic Storms,” ~. Geophys. Res., Trapped Radiation Handbook, edited by J. B. Cladis,
86:2349, 1981b. G.T. Davidson, and L.L. Newkirk, Lockheed Palo Alto
Spjeldvik, W .N. and L.R. Lyons, “On the Predictability of Research Laboratory, DNA 2524H, Revised January,
Radiation Belt Electron Precipitation into the Earth’s 1977.
Atmosphere Following Magnetic Storms,” Proceedings Walt, M. and L.L. Newkirk, “Addition to Investigation of
of the international Symposium on Geophysical Predic- the Decay of the Starfish Radiation Belt,” J. Geophvs.
tions, IV, B-59, U.S. Department of Commerce, Boul- Res. 71: 1966.
der, 1980. West, H. I., Jr., R.M. Buck, and G.T. Davidson, “The
Spjeldvik, W.N. and R.M. Thorne, “The Cause of Storm Dynamics of Energetic Electrons in the Earth’s Outer

5-54
 THE RADIATION BELTS

Radiation Belt During 1968 as Observed by the Law- Williams, D. J., “The Earth’s Ring Current: Causes Gen-
rence Livermore National Laboratory’s Spectrometer eration and Decav,” SuaCe Sci. Rev. 34:223, 1983.
OGO-5,” J. Geophys. Res., 86:2111, 1981. Young, D. T., “Ion ‘Composition Measurements in Mag-
West, H. I., Jr., “The Signatures of the Various Regions of netospheric Model ling” in Quantitative Modelling of
the Outer Magnetosphere in the Pitch Angle Distribu- Magnetospheric Processes, edited by W. P. Olson, AGU
tions of Energetic Particles,” in Quantitative Modelling Washington D.C. 1979.
of Magnetospheric Processes, edited by W. P. Olson, Zinn, J., H. Hoerline, and A.G. Petschek, “The Motion of
AGU, Washington D.C. 1979. Bomb Debris Following the Starfish Test,” in Radiation
Williams, D. J., “An Overview of Radiation Belt Dynamics” Trapped in the Earth’s Magnetic Field, D. Reidel, Dor-
in Proceedings of the Air Force Geophysics Laboratoq drecht, Holland, 1966.
Workshop on the Earth’s Radiation Belts: January 2 1–27,
1981, edited by R.C. Sagalyn, W .N. Spjeldvik, and
W.J. Burke, AFGL TR-81-0311, ADA113959, 1981.

5-55
 Chapter 6

GALACTIC COSMIC RADIATION AND

SOLAR ENERGETIC PARTICLES
D.F. Smart and M.A. Shea

6.1 NOMENCLATURE AND DEFINITIONS

J(>E) = X J(E)dE, (6.3)
JE
Energetic charged particle radiation is usually charac-
terized by the energy of the particle in electron volts. An with units of particles cm-2s-lsr-1. Omnidirectional inten-
electron volt is the kinetic energy a charged particle gains sities are J(E) or J( > E) integrated over the 41r steradians
by being accelerated through a potential difference of one solid angle, with units of particles cm-2s-l MeV- 1or particles
volt. The amount of energy in one electron volt is rather cm-2s-] respectively.
small; particle energies are usually given in units of keV A graph of the differential energy spectrum gives J(E)
(103 eV), MeV (106 eV), GeV (109 eV), and for extremely plotted against E; a graph of the integral energy spectrum
high energies, TeV (1012 eV). (An electron volt is equal to gives J(>E) plotted against E. The differential energy spec-
1.602 x l&12 erg. ) The total energy of a nucleon & is the trum J(E) and the integral energy spectrum J(>E) are com-
sum of the kinetic energy E plus rest-mass energy mOc2 monly expressed in two ways, either as a power law in
energy,
ET = E + ~cz = mc2, (6.1)
J(E) = JOE-?, (6.4)
where mO is the rest mass, m the relativistic mass, and c
E-y+]
the speed of light in vacuo. The rest mass energy of a proton
J(>E) = Jo— (6.5)
~c2 is equal to 938.232 MeV, the rest mass of a neutron y–l’
is 939.526 MeV, and the rest mass of an electron is equal
to 510.984 keV. It is customa~ in the literature to give where y is the differential spectral index, or in the expo-
values of the kinetic energy per nucleon for heavier particles. nential form
The total energy is related to the particle momentum p
by
J(E) = JO exp~ (6.6)
o
& = [p*c* + de’] “*. (6.2)
J(>E) = JOEOexp~ (6.7)
In the absence of electric fields and of time-varying magnetic 0

fields, & and p are constants of the particle’s motion.

Observations of the intensity of corpuscular radiation where EO is defined as the characteristic e-folding particle
are reported in various units depending upon the detection energy, that is, the value over which the flux will fall
method. The symbol “J” is normally used to designate the to I/e of its previous value. There is a considerable hazard
flux, but the precise units must always be specified to avoid in extrapolating any spectral form beyond the range of the
ambiguity. There are a number of ways to specify particle data from which it was derived, and many scientists give
flux. The unidirectional differential intensity J(E) is the flux the range of values for which specific spectral forms are
(number per unit time) of particles of a given energy per valid.
unit energy interval in a unit solid angle about the direction When the effects of corpuscular radiation on personnel
of observation, incident on a unit area perpendicular to the are of primary concern, particle flux intensities or counting
direction of observation; the units are usually particles rates must be converted to dose or dose rates. The rad is
cm-2s-isr-1 MeV-’. Unidirectional integral intensity J(>E) the unit of absorbed energy; one rad is 100 erg absorbed
is the intensity of particles with energy greater than a thresh- per gram of absorbing material. A rem (roentgen equivalent
old energy E man) is the dose absorbed that produces the same biological

6-1
CHAPTER 6
effect as one rad of x-rays or y rays. The relation between
the rem and the rad is one rem = rad X rbe where rbe
is the relative biological effectiveness. Conversion of ob-
served particle fluxes to absorbed dose rate, however, is not
straight-forward and is sometimes very difficult requiring
extremely detailed information on the particle composition
and energy spectrum. The conversion depends in a complex
way on the energies and kinds of particles and on geo-
metrical configurations of the absorbers and direction of the
incident particles, as well as on the different absorbing prop-
erties of materials. Computer codes exist [see, for example,
Seltzer, 1980] that will convert incident energetic particle
fluences into radiation dose behind a specified shield thick-
ness. See Haffner [1967] for a detailed discussion of radia-
tion dosage and the relative sensitivity of various organs to
radiation.
The magnetic rigidity P of a particle is a measure of its
resistance to a magnetic force that deflects the particle from
a straight-line trajectory. The rigidity, with units of mo-
mentum per unit charge, is defined as

(6.8)

where q is the charge of the particle. If pc is electron volts,

then q is the number of electronic charge units and P is in RIGIDITY (GV)
volts. Convenient units are MV ( 106V) and GV ( 109V).
Conversion from rigidity to energy can be accomplished Ftgure 6-1. Conversion from magnetic rigidity to kinetic energy per nu-
by solving Equation (6.2) for the desired quantity; however, cleon for electrons, protons and alpha particles.
the conversion can be done conveniently by employing the
relativistic parameter y. Conversion from rigidity to energy
in terms of kinetic energy per nucleon is Figure 6-1 shows the relationship between the rigidity
of protons, alpha particles, and electrons and their kinetic
energy per nucleon. Since most heavy nuclei have an A/Z
E. = (y – l)EO~ (6.9) ratio of approximately two, the alpha particle curve can be
used to approximate the rigidity to kinetic energy per nu-
where E~ is the kinetic energy per nucleon, and EO~is the cleon of the heavier nuclei. To obtain the total kinetic energy
rest mass energy per nucleon. Conversion from kinetic en- of a heavy nuclei the energy scale must be multiplied by
ergy per nucleon to rigidity is the total number of nucleons (as an example, by 4 to obtain
the total kinetic energy of an alpha particle).
On many occasions the magnetic rigidity is used in place
P = ~[(y2 – l)05]EO~ (6. 10) of the kinetic energy in describing the flux spectrum. In
particular, in lieu of Equation (6.6), a frequently used
expression is
where A is the atomic number and Z is the atomic charge.
The relativistic parameter T can be computed from either
the cosmic ray kinetic energy or the cosmic ray rigidity. J(P) = J. exp~. (6.12)
0

EA + EOA
y= (6. 1la) The earth’s magnetic field acts as a momentum analyzer
EOA on cosmic rays incident upon the earth’s atmosphere. Only
those cosmic rays having a momentum per unit charge (that
or is, rigidity) exceeding that of the threshold of the observation
point may be detected at any specified point on the earth’s

‘= 11”2
[(%)2+ (6.llb) surface or in the earth’s magnetosphere. The cutoff or threshold
rigidity of cosmic rays is the minimum rigidity that permits

6-2
 GALACTIC COSMIC RADIATION AND SOLAR ENERGETIC PARTICLES
a charged particle to arrive from a specified direction at a as the inverse of the solar sunspot number cycle. This is
given latitude and longitude; particles of lower rigidity are discussed more fully in Section 6.2.2.1. The isotropic flux
not observed at the specified Iocation and direction because exposure to galactic cosmic radiation in space at sunspot
of this geomagnetic cutoff. The geomagnetic field is con- minimum is -4 protons cm-2s-1 resulting in a yearly in-
figured such that particles of progressively lower rigidity tegrated exposure of - 1.3 X 108 protons/cm2. The iso-
are detected as the distance north or south of the equatorial tropic flux exposure to galactic cosmic radiation at sunspot
region increases. See Section 6.2.3 for more detailed dis- maximum is - 2 protons cm-2s-1 resulting in a yearly in-
cussions. More specialized concepts that consider the “bend- tegrated exposure of - 7 x 107 protons/cm2. The galactic
ing” of particle trajectories through the magnetic field and cosmic radiation converted to integrated dose results in ex-
allow a mapping of observational directions inside the mag- posures ranging from 4 rads per year at sunspot maximum
netosphere to directions in interplanetary space are referred to 10 rads per year at sunspot minimum [West et al., 1977].
to as asymptotic directions and asymptotic cones of accep- The differential energy spectra of all high energy cosmic
tance. A more detailed description of these concepts and ray nuclei above - 1 GeV/nucleon exhibit a spectrum pro-
their use is given by McCracken et al. [1968]. portional to a power law of the form dJ/dE a E-y, where E
is the kinetic energy per nucleon and y is the spectral index.
Below - 1 GeV/nucleon, the diffe~ntial spectrum of cosmic
6.2 GALACTIC COSMIC RADIATION ray nuclei observed at the earth’s orbit deviates from a
simple power law. The differential spectrum becomes flatter
Galactic cosmic radiation is composed of high energy with decreasing energy until a maximum in the differential
nuclei believed to propagate throughout all space unoccu- intensity is reached around a few hundred Me Vlnucleon.
pied by dense matter. Its origin is still a matter of scientific Below the maximum, the differential intensity decreases
debate and may have both galactic and extragalactic sources. monotonically to a few tens of MeV/nucleon. The primary
The flux of galactic cosmic radiation is believed to be es- cosmic radiation spectrum observed for protons and alpha
sentially isotropic outside the heliosphere; propagation ef- particles in the inner heliosphere around the earth’s orbit is
fects inside the heliosphere result in an anisotropy of - 1%. shown in Figure 6-2. At the vicinity of the earth, the low
Primary cosmic radiation is, by definition, the cosmic energy portion of the spectrum changes with time. These
radiation incident on the earth’s atmosphere. Cosmic rays changes are mainly the effects of solar modulation, and are
propagating through the atmosphere undergo nuclear col- illustrated by the shaded and hatched areas of Figure 6-2.
lisions and generate secondary cosmic rays; these secondary
cosmic rays consist of all known nuclear and sub-nuclear 6.2.1.1 Elemental Composition of Primary Cosmic
species. Radiation. Nuclei heavier than helium comprise only about
High energy cosmic ray particles contain a large amount 1% of the total primary cosmic radiation and have a total
of kinetic energy, and the deposition of this energy can integral intensity of about 25 particles m-2s-1 sr- 1. Cosmic
permanently affect the material through which the cosmic ray nuclei with Z > 2 are classified into various charge
ray nucleus passes. In the case of small, state of the art, groups such as light (L), medium (M), light-heavy (LH),
solid state electronic devices operating in space, the passage and very heavy (VH). The L-group nuclei include those
of a cosmic ray through a circuit element can generate enough with 3 < Z < 5, the M-group with 6 < Z < 8, the LH-
electrons in the sensitive volume to change the state of the group with 9< Z s 14, the H-group with 15 s Z s 19
circuit element and cause “soft errors” or permanent dam- and the VH-group with 20< Z s 28. The charge group
age. A recent examination of the effects of cosmic radiation from manganese (Z= 25) to nickel (Z= 28) is commonly
on microelectronics is given by Adams et al. [1981]. referred to as the iron group.
During their travel from the source regions to the vicinity
of the earth, the cosmic rays interact with the interstellar
6.2.1 Primary Cosmic Radiation medium, and some fragmentation occurs resulting in a de-
pletion of the heavy charged primary cosmic rays and an
The primary cosmic radiation observed at the earth’s increase of the lighter nuclei compared to the “Universal
orbit consists of approximately 83% protons, 13% alphas, Abundance” shown in Figure 6-3. The relative abundance
1% nuclei of atomic number Z >2, and 3% electrons. of the various nuclei observed at the earth for energies
This composition extends over an energy range from a few greater than 0.45 GeV/nucleon are shown in Table 6-1 [Lez-
hundred MeV to >1020 eV. There are no known local plan- niak and Webber, 1978].
etary sources for the high energy ions observed in the cosmic The abundance of the elements N, Na, Al, S, Ar, Ca, Cr,
radiation, but the electron component below about 20 MeV and Mn is enhanced in the cosmic radiation observed in the
is dominated by Jovian electrons. heliosphere as a result of fragmentation in the interstellar me-
The intensity of cosmic rays observed at the earth’s orbit dium. Almost all of the abundance of the elements Li, Be, B,
is solar cycle dependent undergoing a solar cycle modulation F, Cl, K, Si, Ti, and V is due to fragmentation of heavier cosmic

6-3
CHAPTER 6
I I , r , ! , I , , , , r
[

I&
II i

ATOMIC NUMBER

Figure 6-3. The universal abundance of the chemical elements in nature

relative to silicon = 1@. These ~sults were obtained from
studies of meteorites, our sun and other stars. [Data from
Cameron, 198 1.]

‘:lJI,
0.01 0.1

KINETIC ENERGY
10

(GeV/nut)
a
Table 6-1.

Charge

He
Cosmic-ray composition as observed at 1 AU.

>450 MeV/Nuc

44700 * 500
Li 192 * 4
Figure 6-2. Primary cosmic ray differential energy spectrum. The upper 94 * 2.5
envelope indicates the solar minimum spectrum while the Be
lower envelo~ indicates the solar maximum spectrum. The B 329 ? 5
shaded area indicates the range of the solar modulation over
a solar cycle. The hydrogen spectrum in this figure has been
c 1130 * 12
multiplied by a factor of 5 so the modulated portion of the N 278 ? 5
spectrum avoids merging with the top of the helium spectmm. 0 1000
F 24 * 1.5
ray nuclei (that is, higher charge number elements) during their Ne 158 ? 3
Na 29 Y 1.5
propagation through the interstellm medium. The ratios of the
cosmic ray elemental abundance to the solar system elemen- Mg 203 ? 3
tal abundancesare generally ordered by atomic parameters such Al 36 ~ 1.5
as the first ionization potential, at least up through Z = 40 (as Si 141 * 3
shown in Figure 6-4), with exceptions to this rule at H, He, C, P 7.5 ~ 0.6
s 34 * 1.5
N, and possibly Ne and Mo.
cl 9.f) ~ f).fj
A 14.2 t 0.9
6.2.1.2 The Anomalous Cosmic Ray Component. The K 10.1 t 0.7
low energy portion of the cosmic ray spectrum is quite Ca 26 ~ 13
variable reflecting its dependence on solar modulation. In Sc 6.3 ~ ().6
1972, during the decline of the 20th solar sunspot cycle, Ti 14.4 * 0.9
anomalies in the energy spectra and composition were noted v 9.5 ~ 0.7
at energies <70 MeV/nucleon. These differences persisted Cr 15.1 t 0.9
through sunspot minimum and continued throughout the Mn 11.6* 1.0
rising portion of the 21st solar sunspot cycle but seem to Fe 103 f 2.5
have disappeared after the maximum of the 21st solar cycle. Ni 5.6 * 0.6
This behavior leads to the suggestion that the anomalous

6-4
 a
GALACTIC COSMIC RADIATION AND SOLAR ENERGETIC PARTICLES

103~ .
T , 1 1 1 111, , , 11 (,] I ,,, ,,,
v
Quiet Time Energy Spectra

1~
1976-1977
Protons
U. Morylati/ Univ.
MPI Chicaqo
102 + Cal Tech
Protons: O A @
\

\
~ Helium: ● A
~
Carbon, ■

A
v
Helium o
Oxygen: ❑ v
i
4- \
1972-73 Measurewn?s
+*
\\ 1

t’
Mo CO ii(R)+
W Fe
Al tiGe Zn OR
Na SrCa Zr Cu S1 SeS C Kr Ar Ne He
t i tL
,t)+LJ 1 I 1 1 1 t 1 1 I

1
567 8910 15 20 25 102
FIRST IONIZATION POTENTIAL (eV)

Figure 6-4. Ratio of the galactic cosmic ray source (GCRS) to the “local
galactic” (LG) elemental abundan~s at 1 AU ordered by the
first ionization potential [after Mewaldt, 1981].
IO-3
10-’
I

Ioa Id
I I
,.2
1
Id

KINETIC ENERGY (MeV/nut)

component may be solar magnetic cycle de~ndent and only

Figure 6-5. The “anomalous” cosmic ray spectrum, The differential en-
present during alternate cycles. The anomalous component
ergy spectra of hydrogen, helium, carbon, and oxygen ob-
resulted in the low energy cosmic ray spectra shown in served in the interplanetary medium near 1 AU during the
Figure 6-5. The “anomaly” can be seen by comparing this solar minimum in 1976-1977 during quiet-times. The “anom-
alous cosmic ray” component appears between -2 and -30
spectra with that of the previous solar minima, the upper MeV/nucleon and is characterized by large overabundance of
envelope of the low energy portion of the s~ctra shown in He and O compared to H and C ,respectively [Gloeckler,
Figure 6-2. In a review of this anomalous component, 1979]
Gloeckler [1979] noted that the helium spectra above -2
MeV/nucleon was relatively flat, and between -5 and 30
MeV/nucleon He was more abundant than protons. The Table 6-2. Abundances relative to carbon of the anomalous component
oxygen spectrum possessed an unusual hump between - 2 and galactic cosmic rays.
and 20 MeV/nucleon with no comparable feature for carbon,
and at - 5 MeV/nucleon oxygen was some 10 to 20 times Attomalous Galactic
more abundant than carboti. In addition td helium and ox- Component Cosmic Rays
ygen, nitrogen and neon were observed to be more abundant Element (2-30 MeV/nucleon)(a) (> 100 MeV/nucleon)
in the anomalous component than in the z 100 MeV/nucleon
galactic cosmic rays. The abundances relative to carbon of H 40*4 275
the major elements in the anomalous ~omponent are com- He 90 * 10 50
pared to the composition of galactic cosmic rays in Table c =1 =1
6-2. Assuming that 2–30 MeV/nucleon carbon is of pre- N 3*1 0.225
dominately galactic origin during quiet times, the elements 0 lf3~4 1.0
He, N, O, Ne, and possibly Fe were from 5 to 20 times Ne 1.3 f f).4 0.175
more abundant in the anomalous component than in the Mg (),3 * 0.2 0.23
normal cosmic ray composition. Withitt statistical uncer- o z + 0.25
Si 0.17
tainties the composition of the other elements in the anom- – 0.15
alous component region were comparable to that expected Fe 06 ~ O.z 0.12
from the galactic cosmic ray flux. (a) Approximate energy range ot the measurements m MeV/nucleon.

6-5
CHAPTER 6
6.2.1.3 Cosmic Ray Electrons. The interplmew COSfiC earth is more efficient. However, the propagation charac-
ray energetic electrons of non-solar origin measured at teristics are still quite variable, depending on the status of
1 AU has two components, a galactic cosmic ray electron the interplanetary medium. The most favorable connection
component and a Jovian electron component. All electron of the interplanetary magnetic field lines between the earth
flux in the heliosphere below about 25 MeV are of Jovian and Jupiter repeats at approximate 13-month intervals with
origin [Eraker and Simpson, 1981]. At the earth’s orbit, for corresponding peaks in the electron flux during the “electron
energies greater than - 100 MeV, the galactic cosmic ray season” which may last for several months. Before the pres-
electron flux dominates. A composite primary electron spec- ence of the Jovian electrons was established, these “quiet
trum is shown in Figure 6-6. The solar cycle modulation time electron increases” were unexplained.
dominates the variability in the electron spectrum from about
100 MeV to about 5 GeV.
Jupiter’s magnetosphere is a source of relativistic elec- 6.2.2 Cosmic Ray Flux Modulations
trons that can be observed throughout the entire heliosphere.
Within several AU of the planet Jupiter, the flux is mod- Cosmic ray flux modulations can be divided into three
ulated with the characteristic 10-hour planetary rotation pe- groups. The long term modulation (millions of years), solar
riod. The MeV Jovian electron flux observed at the earth cycle modulations (usually referred to as the 11-yr cycle),
has a 13-month “seasonal” characteristic. This “seasonal” short term modulations (consisting of cyclic variations rang-
effect results from the relative position of Earth and Jupiter ing from a semi-diurnal variation to a 27-day variation, and
as the planets revolve around the sun in their respective transient variations consisting of decreases and increases
orbits. When the interplanetary magnetic field lines passing associated with various solar phenomena).
near the earth connect to the Jovian magnetosphere (the There is conjecture about variations in the galactic cosmic
Jovian magnetosphere being a minimum of 5 AU in length), ray sources over very long time scales. The analysis of
the propagation of the Jovian electrons from Jupiter to the cosmic ray exposure to meteorites and lunar rocks is both
limited and complicated. So far, statistics and uncertainty
in the determinations only allow the conclusion that the
current cosmic ray composition has not changed by more
COMPOSITE PRI MARY ELECT~ ePECTRUM
than a factor of two over the last 10a years. The isotopic
A ISEE-3 FEB 1979- MAR Ieel

v 000-5 JUNE 1971 composition of the cosmic radiation, specifically the 10Be
0 19?7 eALLOON (LOW mloI mcd.latlan) isotope presumed to be a spallation product of cosmic ray
O 1979 SALOON (high soiar tiufiattim)
❑ 1980 BAL_
interaction in the interstellar medium, results in a probable
❑ JOVIAN ELECT*
cosmic ray age of - 15–17 million years.
❑ SOLAR MODULATION
Cosmogonic isotopes produced by the cosmic ray in-
~&e&,:
teractions in air, such as 26A1, ‘“Be, and 14C, and then
frozen in sea sediments, ices or biological materials produce
\ ~:?’%~:+$$:s%.
““:$$::*.&qi&>

,,:: ~
somewhat better limits on possible cosmic ray variations.
A %..:...,..,
....
The 10Be isotope sedimettts over the last two million years
show less than a 30% variation in the cosmic radiation.
The carbon-14 variations are most likely due to the strength
of the solar modulation effect. These variations of a few
percent have the amplitude and phase expected from his-
u torical and recent sunspot data, cosmic ray modulation ob-
u servations, and’C02 transport theory [Fortnan and Schaffer,
— 1979].
u

❑
6.2.2.1 Solar Cycle Modulations. The galactic cosmic
radiation exhibits an intensity variation with - 11-yr pe-
riodicit y, as shown in Figure 6-7. The cosmic ray intensity
changes at the earth are inversely correlated with the sunspot
number and lag changes in sunspot number by 9 to 12
G months. The increase in solar activit y modulates the galactic
cosmic rays, through the agency of the solar wind, in such
I I 1 I [ I , I 1 a manner that an increase in solar activity corresponds to a
162 dl d I& 18
ELECTRON KINETIC ENERGY (GeV)
decrease in the cosmic ray intensity. From solar cycle min-
imum to solar cycle maximum, the energy density of the
Com~site primary cosmic-ray electron differential energy
spectrum [courtesy P. Meyer, University of Chicago, Private
primary galactic cosmic rays in the vicinity of the earth
Communication]. decreases by about 40%. The cosmic ray intensity, as ob-
 GALACTIC COSMIC RADIATION AND SOLAR ENERGETIC PARTICLES
where V is the solar wind velocity and K the effective
diffusion coefficient in the interplanetary magnetic field ir-
regularities. The diffusion coefficient is generally a function
of @ = (v/c), magnetic rigidity and radial distance. The
factor w ,= (y + 1)/y, where y is the relativistic factor.
Numerical techniques for solving Equation (6. 13) have been
developed by cosmic ray modulation theory specialists. These
techniques require the specification of the interstellar density
spectrum U(EO), the functional form of the diffusion coef-
ficient, and appropriate boundary conditions. Numerical so-
lutions of equation (6. 13) yield the modulated spectra at
1 AU.
Figure 6-7. Illustration of the - 1l-yr solar cycle modulation of galactic
The solar cycle modulation of the cosmic ray spectra is
cosmic rays observed at the earth by the Mt. Washington described quantitatively by the “modulation parameter,” ~.
neutron monitor from 1954 thro 1979. This parameter is defined as

served at the earth, does not change smoothly from mmi-

mum to minimum values. It decreases by a series of sharp
(6. 14)
drops followed by partial recovery until the minimum in-
tensity is reached near the maximum in solar activity. The
recovery to maximum cosmic ray intensities may be either
a slow rise, as observed from 1958– 1964, or a rapid increase where V is the solar wind velocity and K](r) is the radial
as observed from 1970-1972. These two variations in in- part of the diffusion coefficient. The diffusion coefficient
creasing cosmic ray intensity are also observed in the ion- K is usually treated as a separable function of radius and
ization chamber records going back to 1933, giving rise to rigidity. In practice the actual value of the cosmic ray mod-
a 22-yr solar cycle variation. The historical record of sunspot ulation parameter at a specific instant of time is difficult to
observations shows longer (80–200 yr) periods modulating determine. A reference level (such as the counting rate of
the amplitude of the sunspot number cycle. The variations a stable neutron monitor) is correlated with the historically
in the carbon-14 record seem to correlate with these same derived modulation parameters as shown in Figure 6-8, and
periods. See, for example, Stuiver and Quay [1980].
Theory of Solar Cycle Modulation. There has been con-
siderable theoretical work attempting to model the modu- t I I I I

lation of cosmic rays by the interplanetary medium. The

physical model is based on a solar system filled with an
expanding fully ionized and highly conducting plasma, the
solti wind (see Chapter 3), which contains frozen-in irreg-
ular magnetic fields. Cosmic rays undergo many scattering
from these irregularities and execute a random walk in the
solar wind. The cosmic ray population outside the helio-
sphere diffuses inward, and during this diffusive process it
undergoes deceleration by the adiabatic cooling associated
with the expansion of the solar wind. The parameters re-
quired to define the transport equation and its solution ‘are
the diffusion coefficient (generally a function of radius and
energy), the solar wind velocity, and the interstellar energy
spectrum.
The basis of current solar modulation theo~ is the Fok-
ker-Plarrck equation for the modulated number density U(r,E)
per unit kinetic energy at heliocentric radius r and kinetic
energy E. Gleeson and Axford [1967, 1968] have given this
equation as 1 I I I I
) 2100 2200 2300 2400 2500

j :(?VU Mt. Washingtan Caunting Rate

Figure 6-8. Correlation between the cosmic ray modulation parameter +

and the counting rate of the Mt. Washington neutron monitor.

6-7
 CHAPTER 6
interplanetary medium corotating with the sun. The ampli-

I
Mt. Wash
Counting Rote
Year of
Measurement
i tude of these variations, as observed by ground-level neutron
10’ +( M”)

“~
A 2055 1970 monitors, is quite variable, normally a few percent, ranging
1971 from about 1% to 5%.
1972
1972 For cosmic ray detectors located on the earth there are
1973
1974
also diurnal and semi-diurnal variations in phase with the
1974
earth’s rotation. The amplitude of these diurnal variations
1975
1976 is variable ranging from a maximum of a ‘few percent to a
1977
minimum of about O. 1%. Figure 6-10 shows an enhanced
daily variation continuing for several days. During stable
interplanetary conditions the maximum amplitude occurs at
-1800 h local solar time. This corresponds to an asymptotic
direction roughly along the direction of the sun-earth Arch i-
~edean spiral path continuing outward from the earth into
interplanetary space.
The semidiumal variation is usually less than 0.1 % and
maximizes in both directions along a line 135° west of the
earth-sun line, a direction perpendicular to the mean inter-
planetary magnetic field direction at the orbit of the earth.
The presence of a sidereal variation reflecting a galactic
anisotropy is still not firmly established, and if present would
have an extremely small amplitude of a few hundredths of
a percent.
Transient Variations. There can be both sudden de-
GeV/Nucleon creases, and more rarely, increases in the cosmic radiation
flux that result from solar activity. Sudden decreases in
Figure 6-9. Differential energy spectra for cosmic-ray iron nuclei. The
shaded area indicates the rattge of the solar cycle modulation. intensity, called Forbush decreases (named after the original
discoverer) are associated with sudden increases in the plasma
density and magnetic flux emitted from the sun. The large
the current neutron monitor counting rate is used as a ref- Forbush decrease events m generally associated with “large”
erence modulation level.
Solutions of Equation (6. 13) result in an analytic de-
scription of the cosmic ray spectrum as a function of kinetic
energy and solar modulation parameter.

[E(E + 2EOA)(E + ~ + M)-Y]

J(E,~) = A (6.15)
[(E + O)(E + 2EOA + ~)] ‘

where E is the kinetic energy of the cosmic ray nuclei in

MeV/nucleon, EOA is the rest mass energy of the cosmic
ray nuclei in MeV/nucleon and @ is the modulation param-
eter. To model the cosmic ray proton spectrum Garcia-
Munoz et al. [1975] present solutions of the cosmic ray
spectrum for various modulation levels and elements of the
cosmic ray flux. Figure 6-9 shows modulated differential
iron spectra for various values of the modulation pat-ameter
~. The shaded areas indicate the range of the modulation
observed during the past few solar cycles.

6.2.2.2 Short Term Modulations. Short term modula-

tions can be subdivided into two types, cyclic variations
and transient variations.
Cyclic Variations. There are - 27-day variations in the
I~ u I
I
5 6 7
observed cosmic ray intensity within the heliosphere, at least W-R 1977

in the regions explored from 0.3 to 20 AU. These variations

are statistical averages associated with the structure of the Figure 6-10, Illustration of an enhanced cosmic ray daily variation,

6-8
 GALACTIC COSMIC RADIATION AND SOLAR ENERGETIC PARTICLES
position in the geomagnetic field from a specified direction.
Rigidity is a canonical coordinate used in cosmic ray studies
which is element independent. When translating these to
the different elements of the cosmic ray spectmm, a different
curve is obtained for each of the elements contained in the
cosmic ray flux. If a primary cosmic ray particle is suffi-
ciently rigid (that is, having a rigidity higher than the cutoff
ngidit y), it can penetrate through the geomagnetic field to
an observation point; if the rigidity of the particle is less
than the cutoff rigidity, it cannot penetrate through the mag-
netic field to the specified location from a specified direction.
When considering charged particle propagation through the
geomagnetic field, it is convenient to use the rigidity (mo-
mentum per unit charge) of the particle. To determine the
a 15 16 17 18
energy per nucleon of a specific cosmic ray element, rigidity
FEWFN 1977
to energy tables can be utilized. See Section 6.1 for addi-
tional discussion.
Figure 6-11. Illustration of a cosmic ray Forbush decrease.
Cosmic ray cutoffs are a function of geomagnetic lati-
tude, altitude, and the zenith and azimuthal directions of
solar flares and interplanetary shock structures. The mag- the incident particle at the detection point. They have a
nitude of these decreases is quite variable ranging from a maximum value in the earth’s equatorial region and a min-
few percent to as great as 35% in the most extreme case imum value in the polar regions. Unfortunately, the cosmic
yet recorded by neutron monitors on the earth. An example ray cutoff is not a simple value. There is an upper cutoff
of a Forbush decrease is shown in Figure 6-11. above which all particles are allowed, a lower cutoff below
There are two types of “cosmic ray” increases attrib- which all particles are forbidden, and a penumbral region
utable to the sun. The solar “cosmic ray” event is discussed between these two values where the transmission of cosmic
in Section 6.3. There are other occasional cases when ex-
tremely energetic solar flare generated shocks propagating
through the interplanetary medium can accelerate the am-
bient particle population to higher energies. An example
is the 4 August 1972 cosmic ray event thought to be the re-
-
sult of Fermi type acceleration of particles between two . —
propagating interplanetary shock structures [Levy et al.,
1976].

’10
>-
0
6.2.3 Geomagnetic Effects ,.‘,,
C ‘\
Q–
Cosmic rays being charged nuclei experience a V X B 0
force that continuously alters their path as they propagate E– ..
through the geomagnetic field. When analyzing cosmic ray 5- 5
observations acquired on the earth to deduce the intensity ..
and anisotropy of the cosmic ray flux in interplanetary space,
it is necessary to make allowance for the “magnetic bending” ..
of the cosmic ray trajectory through the magnetosphere.
Methods that correct cosmic ray observations for the actual
amount of the “geomagnetic bending” each particle has .-1- -L-

undergone are very specialized and beyond the scope of this o- 70- k
GEOGRAPH!: LATITU~E
handbook. Detailed information on cosmic ray asymptotic
directions of approach, asymptotic cones, and variational
Figure 6-12, Illustration of the width (shaded area) of the cosmic ray
coefficients is given by McCracken et al. [1965, 1968]. vertical cutoff penumbra as a function of latitude along the
260”E meridian. The solid line indicates tbe effective geo-
magnetic cutoff rigidity along this meridian. The “sharp-
6.2.3.1 Cosmic Ray Cutoff Rigidities. Cosmic ray cu-
ness” of the penumbral width variations are exaggerated by
toff rigidities are values that specify the minimum rigidity only considering the vertical direction; however, the mag-
a charged particle must possess to be observed at a specified nitude of the variation is correct.

6-9
CHAPTER 6
radiation is chaotic. The cosmic ray penumbra exists near precision of these calculations is limited only by the ac-
the cosmic ray cutoff rigidity where there is a complex series curacy of the geomagnetic field model utilized.
of allowed and forbidden particle trajectories that has so far There are long term secular variations in the cosmic ray
defied all attempts to systematically order it. The effective cutoff rigidities directly reflecting the long term secular
cosmic ray cutoff is a practical value corrected for the trans- changes in the geomagnetic field. Over an approximate
mission through the cosmic ray penumbra. An example of 10-yr period these changes are sufficient to be experimen-
the variation of the vertical cosmic ray cutoff along the tally observed, and for very precise analyses these secular
260”E longitude meridian is shown in Figure 6-12. The variations should be considered.
width of the cosmic ray penumbra shown in this figure is Because of the complexity of cutoff rigidity calculations,
somewhat typical of its worldwide behavior. a number of approximations are generally employed. For
The primary reason that precise values of the cosmic experimental analyses, the specific approximation utilized
ray cutoffs are relatively difficult to obtain is that the equa- is usually determined by the precision of the measurement
tion of charged particle motion in a magnetic field does not being studied; as cosmic ray experiments become more pre-
have a solution in closed form. Relatively accurate geo- cise, more accurate cutoff rigidities are required.
magnetic cutoff rigidities can be derived by the numerical For the majority of cosmic ray studies on the earth’s
integration of cosmic ray trajectories in mathematical models surface or in the atmosphere, it is often sufficient to know
of the geomagnetic field; however, the calculation of these the vertical cutoff rigidity at each observational location.
values for a large number of locations and directions in- World grids of vertical cutoff rigidities from which vahtes
volves a formidable amount of computational time. The for intermediate locations and time periods can be inter-

Table 6-3. Trajecto~-derived effective vertical cutoff rigidities for Epoch 1980.

Geographic Longitude (E)

Geographic
Latitude f) 15 30 45 60 75 90 105 120 135 150 165

70 0.27 0.34 0.39 0.44 0.48 0.51 0.51 0.55 0.58 0.60 0.61 0.57
65 0.60 0.69 0.80 0.87 0.91 0.94 0.98 1.03 1.12 1.23 1.18 1.11
60 1.16 1.36 1.43 1.59 1.62 1.68 1.70 1.80 1.96 2.05 2.12 2.06
55 2.00 2.29 2.45 2.53 2.67 2.73 2.84 2.93 3.12 3.31 3.31 3.15
50 3.32 3.59 3.83 3.94 4.06 4.20 4.34 4.45 4.69 5.00 4.97 4.69
45 4.99 5.20 5.35 5.44 5.66 5.81 6.08 6.31 6.59 6.96 6.96 6.38
40 6.95 7.44 7.59 7.73 8.07 8.54 8.99 9.23 9.57 9.99 9.82 9.05
,35 9.77 9.74 10.01 10.42 10.88 11.27 11.39 11.63 11.95 12.18 11.69 10.67
30 1I .49 11.83 12.10 12.51 13.09 13.82 14.19 14.31 14.23 13.97 13.46 12.75
25 13.25 13.68 14.03 14.38 14.86 15.37 15.69 15.70 15.47 15.05 14.46 13.76
20 14.17 14.61 14.99 15.39 15.91 16.43 16.73 16.68 16.36 15.85 15.21 14.54
15 14.63 15.10 15.54 15.99 16.54 17.07 17.35 17.27 16.91 16.37 15.75 15.14
10 14.70 15.19 15.67 16.17 16.75 17.29 17.57 17.50 17.15 16.63 16.06 15.56
5 14.41 14.88 15.38 15.94 16.57 17.11 17.41 17.38 17.07 16.61 16.15 15.77
0 13.80 14.22 14.73 15.34 16.00 16.56 16.87 16.90 16.67 16.31 15.97 15.74
–5 12.94 13.27 13.77 14.41 15.10 15.64 15.97 16.07 15.94 15.68 15.50 15.42
– 10 11.86 12.11 12.57 13.23 13.90 14.40 14.71 14.88 14.84 14.69 14.65 14.77
– 15 10.45 10.63 11.08 11.75 12.32 12.80 13.06 13.24 13.17 13.18 13.39 13.69
– 20 8.87 8.89 9.26 9.74 10.24 10.45 10.55 10.69 10.75 10.66 10.62 11.87
– 25 7.28 7.29 7.63 7.93 8.02 7.71 7.28 7.26 7.42 7.64 8.36 9.49
– 30 6.11 5.84 5.84 5.80 5.58 5.40 5.19 5.14 5.09 5.38 5.84 6.60
– 35 5.05 4.49 4.37 4.38 4.12 3.85 3.47 3.41 3.34 3.55 4.10 4.90
– 40 4.03 3.62 3.38 3.26 2.89 2.58 2.21 2.04 2.04 2.22 2.53 3.29
– 45 3.33 2.88 2.53 2.38 2.00 1.54 1.28 1.12 1.10 1.22 1.47 2.01
– 50 2.76 2.27 1.97 1.64 1.30 0.94 0.71 0.56 0.53 0.60 0.81 1.15
– 55 2.17 1.72 1.45 1.12 0.82 0.56 0.35 0.24 0.21 0.25 0.38 0.57
– 60 1.69 1.29 1.03 0.76 0.69 0.30 0.15 0.09 0.06 0.08 0.14 0.26
– 65 1.29 0.96 0.72 0.48 0.29 0.14 0.06 0.00 0.00 0.00 0.03 0.10
– 70 0.84 0.66 0.46 0.30 0.16 0.08 0.00 0.00 0.00 0.00 0.00 0.02

6-10
 =

GALACTIC COSMIC RADIATION AND SOLAR ENERGETIC PARTICLES

polated for Epochs 1955, 1965, 1975, and 1980 have been EPOCH = 19800
..~—--——.--——--------- .-
calculated by numerical integration [Shea et al., 1968; Shea >+ ‘7

and Smart, 1975; 1983]. An iso-rigidity contour map of

effective vertical cosmic ray cutoffs for Epoch 1980.0, re-
flecting the use of the newer and more accurate internal
magnetic field model [Peddie, 1982], is shown in Figure
6-13. The values for the effective vertical cosmic ray cutoff
rigidities for this same epoch are tabulated in Table 6-3.
The Stormer Approximation. The classic work of Stormer
on the motion of charged particles in a dipole magnetic field
resulted in the following equation for the cosmic ray cutoff
rigidity P=
Figure 6-13. Iso-rigidity contours of vertical cosmic ray effective cutoff
cos4k rigidities, The units of cutoff rigidity are in GV.
PC=M (6.16)
R2[1 + (1 – sine sinb COS3A)1’2]2

where A is the geomagnetic latitude, e is the zenith angle,

and @ is the azimuthal angle measured from the magnetic
north, When this equation is normalized to the earth’s dipole

Table 6-3. (Corrfinued)

Geographic Longitude (E)

Geographic
180 195 210 225 240 255 270 285 300 315 330 345 Latitude

0.47 0.36 0.23 0.13 0.04 0.00 0.00 0.00 0.00 0.06 0.11 0.19 70
0.99 0.77 0.49 0.31 0.17 0.09 0.05 0.05 0.09 0.17 0.30 0.43 65
1.76 1.38 0.98 0.65 0.40 0.23 0.16 0.16 0.23 0.39 0.64 0.90 60
2.85 2.28 1.75 1.23 0.81 0.49 0.36 0.36 0.50 0.80 1.21 1.70 55
4.25 3.46 2.81 2.05 1.42 0.98 0.75 0.73 0.95 1.43 2.13 2.85 50
5.60 4.85 4.15 3.16 2.37 1.75 1.35 1.27 1.62 2.47 3.35 4.35 45
7.96 6.47 5.49 4.60 3.63 2.77 2.14 2.05 2.57 3.79 5.08 6.15 40
9.52 8.99 7.67 6.11 5.17 4.21 3.24 2.99 3.83 5.26 7.36 9.11 35
11.69 10.48 9.65 8.78 6.99 5.54 4.38 4.02 5.14 7.54 10.02 10.94 30
13.04 12.38 11.74 10.88 9.63 7.74 6.02 5.51 7.02 9.98 11.78 12.68 25
13.90 13.33 12.84 12.23 11.32 9.35 7.67 6.61 8.49 11.67 12.88 13.63 20
14.59 14.09 13.67 13.19 12.46 11.05 9.27 8.64 11.01 12.60 13.50 14.12 15
15.10 14.68 14.30 13.88 13.35 12.36 11.27 11.06 12.22 13.11 13.78 14.25 10
15.43 15.07 14.73 14.36 13.92 13.22 12.45 12.29 12.73 13.33 13.76 14.04 5
15.53 15.25 14.93 14.60 14.21 13.71 13.14 12.80 12.97 13.32 13.48 13.55 0
15.35 15.17 14.92 14.62 14.28 13.86 13.36 12.99 12.98 13.09 13.00 12.85 –5
14.87 14.82 14.66 14.43 14.15 13.81 13.37 12.97 12.81 12.69 12.36 11.92 – 10
14.01 14.17 14.15 14.03 13.84 13.57 13.19 12.77 12.47 12.15 11.53 10.79 – 15
12.72 13.15 13.36 13.42 13.35 13.17 12.85 12.42 11.99 11.47 10.49 9.47 – 20
10.19 11.24 11.98 12.56 12.67 12.63 12.38 11.94 11.40 10.52 9.35 8.04 – 25
7.98 9.45 9.40 10.75 11.79 11.92 11.78 11.34 10.56 9.57 8.15 6.88 – 30
5.65 6.62 8.01 8.73 9.62 11.05 11.09 10.55 9.73 8.50 6.88 5.95 – 3“5
4.15 4.84 5.60 6.76 8.18 9.73 10.08 9.63 8.80 7.35 6.18 5.00 – 40
2.69 3.30 4.28 4.99 6.01 7.87 8.89 8.52 7.74 6.80 5.34 4.15 – 45
1.65 2.24 2.94 3.79 4.58 5.61 7.05 7.41 6.77 5.47 4.27 3.42 – 50
0.95 1.36 1.94 2.64 3.35 4.29 4.90 5.18 4.90 4.25 3.48 2.77 – 55
0.51 0.76 1.20 1.77 2.27 2.98 3.75 4.01 3.82 3.39 2.72 2.14 – 60
0.22 0.42 0.69 1.05 1.55 1.96 2.46 2.72 2.63 2.40 2.02 1.61 – 65
0.10 0.22 0.40 0.60 0.90 1.20 1.51 1.67 1.75 1.59 1.33 1.11 – 70

6-11
CHAPTER 6
moment M and the distance from the dipole center R is 7231’”
expressed in earth radii, the constant terms evaluate to 59.6.
For locations on the earth the vertical cutoff rigidity is very
useful, and since the zenith angle is zero the dipole vertical
cutoff is given by

P= =
14.9 COS4A
R* “
(6.17)
W“ev
Variations of the Stormer cutoff equation can be em- 1
1
ployed when normalized to the earth’s actual magnetic field.
An improvement in accuracy can be obtained by correcting
for the displacement of the earth’s effective magnetic center
~ 74
from the geocenter. A number of magnetic coordinate sys- 13UN
; 72 . . . .. . #
. .. 15
tems can be employed; however, the McIlwain B-L coor- ~7w
.:. ..1-. . . 2.3 MeV
1.8
2.3
g w.....”..
dinate system [McIlwain, 1961] is very useful since in a 66 “.,**. J3
- .:
dipole field cos2h = R/L. The worldwide grid of trajectory- @ ..8
km
““.*. :

=rLL111t4ii111ch[llLrjl) S1 T,, ,,, ,,, ,,, ,,, ,,, ,,, ,,, , 1

derived effective vertical cosmic ray cutoff rigidities, nor- 0 6 12 la 24 0 6 12 Ie 24
MAGNETIC EQUATORIAL TIM E(hou,, )
malized to the L parameter, results in

Figure 6-14. Daily variation of the high latitude geomagnetic cutoffs.

P= = 16.0 L-2. (6.18)

This form is useful where the geomagnetic field retains its

basic dipolar structure; it can be used throughout the mag- 6.2.4 Charged-Particle Flux
netosphere wherever a useful L coordinate can be calculated. Within the Atmosphere
A practical limit for the region of applicability seems to be
within the domain of the L = 4 shell. At higher L-shells The intensity and composition of the cosmic rays ob-
increasing errors will be encountered due to the external served within the atmosphere depend on the quantity of
current systems of the magnetosphere. absorbing material traversed before observation in addition
Recent work indicates that the Stormer equation can to the cutoff rigidit y of the observation point. Atmospheric
be generalized to estimate the cutoff at other directions conditions, especially barometric pressure, have an appre-
and altitudes if a local normalization value is available. ciable effect on the measured intensity; hence, cosmic-ray
If a vertical (more properly, dipole radial) cutoff is known, intensities are usually reported in terms of atmospheric depth
the cutoff at other directions may be computed using a (mass of air per unit area above the observation point) or
normalized form of Equation (6. 16) and an appropriate of air pressure at the observation point rather than the al-
magnetic coordinate system. If a cutoff at one altitude is titude of the observation. The ionization rate measured within
known, the cutoff at another altitude can be computed. the atmosphere depends upon the amount of matter above
A convenient way is to employ the L-2 dependence of the point of observation and on its distribution in height.
Equation (6. 18). At a given altitude, the value of the pressure in millibars is
Cosmic Ray Cutoffs at High Latitudes. The earth’s mag- about 2% less than the atmospheric depth in grams per
netosphere has a dynamic topology reflected in the daily square centimeter. Figure 6-15 is a plot of atmospheric depth
variation of the cosmic ray cutoff at high magnetic latitudes. as a function of altitude.
In regions where the magnetic field lines are “open” (that Primary cosmic rays incident upon the top of the earth’s
is, traceable to the distant magnetospheric tail), the cosmic atmosphere interact with air nuclei producing high-energy
ray cutoff is essentially zero. In regions of the magneto- secondary cosmic rays. These secondary particles, in turn,
sphere where the magnetic field line topology changes from interact with other nuclei and produce additional secondar-
open to closed as the earth rotates (magnetic latitudes from ies. Figure 6-16 illustrates a nuclear cascade process initi-
-60° to - 750), there is a significant daily variation in the ated by a primary cosmic ray. The production of secondary
value of the cosmic ray cutoff rigidities. Current magne- components becomes significant at about 55 km (4-rob pres-
tospheric models are inadequate to precisely model these sure with the local intensity reaching a maximum (the Pfotzer
changes and are limited to providing upper limits of the maximum) at approximately 20 km (56 rob). The intensity
cutoff rigidity values. The best available average high lat- of secondaries then decreases from the Pfotzer maximum
itude cutoff values have been obtained by polar orbiting to the surface of the earth as the particles lose energy by
satellites [Fanselow and Stone, 1972] and are shown in additional collisions until the majority either decay or are
Figure 6-14. absorbed. An excellent compilation assembling cosmic ray

6-12
 GALACTIC COSMIC RADIATION AND SOLAR ENERGETIC PARTICLES
I&

Id :

,
4

. 10*
0.1

I -,

,n~ o 10 20 30 40
ALTITUOE(km)
50 so T(
ALTITUOE ( km)

Figure 6-15. Fhe mass of air per unit ma in a verticat column extending upward from a height H above sea level [Ely, 1962].

observations on the earth has been prepared by Allkofer and

Grieder [1984].
ItI theory, the intensity of any specified secondary com-
ponent can be derived from a knowledge of the primary
spectrum and the specific yield functions. Computer codes
exist that will convert a specified primary cosmic tay spec-
trum into probable secondary components at any location
and specified atmospheric depth. See O’Brien [1970],
O’Brien [1979] and references therein. An example of the
computed and experimentally measured ionization as a func-
tion ‘of altitude ii shown in ‘Figure 6-17
LOW ENERGY NUC L EONIC
COMPONENT
1DISINTEGRATION
NEUTRONS
PRODUCT
DCGC NERATE The
6.2.4.1 Secondary Cosmic Ray Ionizing
TO .SLOW- NEuTRONS)

Component. The cosmic ray secondaries most commonly

,--,
ELECTROMDGNCTIC MESON
,- ~o J
NUCLEONIC
.J
COMPONENT
measured on the earth’s surface are IT and p mesons, neu-
trons, protons, electrons, and y-ray photons. The secondary
OR “soFT” OR .HARO” N,P. HIGH ENERGY NuCLEONS
COMPONENT COMPONENT
0,0 . OISINTEGRhT IO.
cosmic rays are often classified into three major components:
-.— . PRW”C1 NUCLE~S
ENCRGY FEEDS ACROSS FROM SM. LL ENERGY FEEDObCN
. NUCLEAR DISINTEGRATKIN
the hard component comprised of primarily relativistic muons,
NuCLEAR 10 ELECTROMAGNETIC FROM MESON TO NuCLEONC
+
INTERACTIONS CWPONENT
the nucleonic component comprised of locally produced
protons and neutrons and the soft component, comprised of
Figure 6-16. Schematic diagram of a cosmic ray shower. N and P are electrons and y rays. The flux and composition of the sec-
high energy neutrons and protons while the lower case n ondary cosmic rays are the result of random interaction
and p are used to denote disintegration product neutrons and
protons. PIons, mesons, electrons, positrons, and gamma probabilities; the instantaneous flux of ionizing secondaries
rays are indicated by conventional symbols. can have a considerable variation from the average flux. For

6-13
CHAPTER 6
I I II I 1 —
L’ I I I I I I example the intensity in the core of the cosmic ray shower
500
400 k i,
\
A
may be orders of magnitude above the “average” for a few
microseconds.
Typical ionization rates in the polar atmosphere as a func-
300 – o EXPERIMENT - tion of atmospheric depth and solar-cycle variation are shown
❑
in Figure 6-18. The variation of ionization with latitude ob-
7“ 200 –00
–O’BRIEN (1970); - tained by Neher [ 1967] during the 1965 solar minimum, plot-
~E
Q e CALCULATED FOR ted as a function of cosmic ray cutoff rigidity, is illustrated in
(n SOLAR MINIMUM Figure 6-19. The most common high energy ionizing com-
fK Inn ,
— ponent of the secondary cosmic radiation is the cosmic ray p
$ ‘i; o
—
— meson. These muons must be relativistic to traverse the at-
_[\
$
— 60
+ —
—
mosphere (half life = 2 x 10- s), are only weakly ioniz-
O\ ing (the energy loss of a p meson in air is 2.2 MeV cmz g-i ) and
— at sea level have a pole to equator latitude dependence of the
1- order of - 10Yo. The typical K meson spectrum is shown in
d Figure 6-20. The p meson intensity as a function of altitude is
N
— shown in Figure 6-21.
The high energy secondary cosmic radiation can initiate
nuclear interactions in whatever matter they penetrate. The
h0 nuclear fragments of these interactions resemble “stars” when
10- recorded in photo~pbic emulsions and examined under high
8- magnification. “Star counts” can be used to estimate the prob-
6– able rate of interactiorts in semi-conductors; these interac-
—
tions can cause errors in solid state logic circuits. A compact
4– tabulation of observed star counts is given in Table 6-4.
—
6.2.4.2 The Secondary Cosmic Ray Neutron Compo-
k :
2- nent. The secondary cosmic ray neutron component is
generated within the atmosphere by the nucleat interaction
of high energy cosmic ray nuclei with atmospheric atoms.
1, 1 I I I i 1 I I II 1 1 There are no neutrons in the primary cosmic ray flux. At
0 200 400 600 800 1000 high energies the “knock-on” process dominates the neutron
ATMOSPHERIC DEPTH (gcti2) production, while at lower energies the “neutron evapora-
tion” process dominates the neutron production. These neu-
Figure 6-17, Computed and exwrimental altitride variation of cosmic rav
trons interact with other atmospheric atoms and produce
ionization. The d;ta are normalized to a sea level vatue o-f radioactive isotopes. The secondary neutron flux in the at-
2.15 ion pairs per second per cubic cent imeter at sea level mosphere is responsible for the generation of cosmogonic
[Kykler and LibOff, 1978].
isotopes such as 14C (half life of 5730 years), 26A1(half life

Table 6-4. Cosmic ray induced nuclear interactions (per cubic centimeter per day).

Type
of Star
(Number of 10 GV
Prongs) 2 GV Cutoff Rigidity Cutoff Rigidity

I Atmospheric
Atmospheric Depth Depth
(g/cmz) (g/cm’)
14.9 47.4 63.7 81.3 121 677 14.9 47.4

>2 2390 2030 2150 2040 1610 22.0 575 425

>9 618 352 372 235 290 1.2 238 132
>16 223 122 128 27 116 0.29 105 61
[Data sources: Rossi [ 1952; 1964] and references therein. ]

6-14
 m

GALACTIC COSMIC RADIATION AND SOLAR ENERGETIC PARTICLES

l\
500
h\ IONIZATION AT THULE

l\
*4 1965 to 1969
$/<
43D ALSO
‘~ *
3A3KATWN CANAOA, AUG 14,1937
* ***+*.
CARMICHAEL and DVMONOXI 03S
NEAR THULE, AUG 3+,1937
40D

. .
! I
A
0 m 100 Ino 200gm CM >0 30 100 130 200qm cm-z so
PRESSURE OF AIR WESSURE OF AIR FWESSURE OF AIR

Figure 6-18. Yearly averaged cosmic rayionization rate ~ratmosphere ofairas afunction ofatmosphetic depth [Neher, 1971].

of7.4 x l@years), adl%e(half life of 1.5 x 106 years). 6.2.5 Cosmic Ray Albedo
The carbon- 14 isotope enters the biosystem and can be used
todatedead biological materials. Other long lived isotopes Some small fraction of the secondary cosmic radiation
may collect in the polar ice or sea sediments. The differential will escape the atmosphere and contribute to the total ra-
energy spectrum of neutrons at various depths in the at- diation flux exposure to earth orbiting satellites. The elec-
mosphere is shown in Figure 6-22 [Armstrong et al., 1973]. trons and protons will spiral along the earth’s magnetic field
These results are typical of midlatitudes. The fast neutron lines to the opposite hemisphere. Measurements of the al-
flux in the atmosphere at various depths and cutoff rigidities bedo proton flux at the “top” of the atmosphere for high
for both solar minimum and solar maximum conditions is and midlatitudes are given in Figure 6-24 while the albedo
illustrated in Figure 6-23 [Light et al., 1973]. electron flux at the “top” of the atmosphere is given in Figure

\5 Q/cm’

15
k
0 400 30
7
E
z lCOg/cm2

rn zoo 200q/cm2
z
o
—
300g/cm2
IOJ
—

I 1 !,, !,1 , i , I i,.

2 4 12 I I 0-1 10°
1 ,.2 ( ~3 lo~
MUW M1&ENTUM PU [GeV(C)
CUT;FF RIGI:ITY(GVP

Figure 6-19. Cosmic ray ionization rate as a function of altitude and Figure 6-20. Differential vertical flux of muons at sea level. Measured
geomagnetic cutoff rigidity. values are indicated by the various symbols while the solid
line is a mathematical calculation. [Murakami et al., 1979].

6-15
 CHAPTER 6
1 f , r 1 I , ,
I 1 results from and is descriptive of a measurement technique.
Cosmic ray sensors on the earth recorded the first observed en-
ergetic solar flare particle outbursts, and hence the name solar
cosmic ray events was derived. Later,. when balloon borne
cosmic ray detectors observed p~icle events not detected at
the earth’s surface, the name ground level event was intro-
duced to distinguish between those events detected at ground
level and those observed only near the top of the atmosphere.
This name is still used to identify solar particle events contain-
ing relativistic protons. Beginning in the 1950s the earth’s po-
larionosphere was shown torespondto solar flare protons with
energies from about 1 to 50 MeV, and the term PCA events
became synonymous with solar proton events. Spacecraft
measurements are now many orders of magnitude more
~~ sensitive than earth-based measurements and allow meas-
urements of the composition of solar particle events. The
ATMOSPHERIC OEP7H (g. c m–z)
term solar electron event was introduced after 1965 to dis-
tinguish the species of particles present. However, the older
Figure 6-21. Altitude variation of the muon intensity >0.3 GeV. The
definitions are still used to provide continuity of historical
solid line represents older values ( circa 1948) [Murafcami
et al,, 1979]. data bases.
The most commonly measured components of solar par-
ticle events are protons and electrons. The first positive meas-
6-25. The electron albedo includes approximately equal urements of solar neutrons occurred in 1980 [Chupp et al.,
numbers of electrons and positrons. There appears to be a large 1982]. Solar neutron fluxes at the earth will be extremely
variability in the available electron albedo measurements that rare as the neutron half life of - 11.7 min implies all but
is not yet resolved. Typical numbers in the energy range 0.1 the most energetic solar neutrons will have decayed before
to2.5GeVaree+ = 77, e- = 67 m-2s-lsrl. The results of they can reach the earth’s orbit.
satellite measurements in the equatorial region at 230 km are
given in Figure 6-26. There is an approximate factor of 2 in-
crease in these numbers in the high latitude region. 6.3.1 The Solar Particle Source
About 10% of the secondq cosmic ray neutron flux in
the upper atmosphere will escape into space. The cosmic The source of the energetic solar particles is generally
ray albedo neutron flux escaping’ from the atmosphere is not the solar flare (see Chapter 1). There is a time correlation
bound by the earth’s magnetic field and will later decay between the observation of an increase of the particle flux
(neutron half life - 11.7 minutes) into an electron and in interplanetary space and the occurrence of a solar flare.
proton, which may be trapped in the earth’s magnetic field, At low energies, there is a correspondence between the level
and an ariti-neutrino. The albedo neutron flux is an important of solar activity (such as the number of flares and sub-flares)
source of the high energy proton flux trapped in the earth’s and the low energy solar particle flux in interplanetary space.
magnetosphere (see Chapter 5). A model “global averaged” Some flares are more copious energetic particle producers
albedo neutron spectrum is given in Figure 6-27. Near the than others, the more copious solar particle generators gen-
earth there is an approximate 1:7 ratio of the albedo neutron erally being solar flares rich with x-ray and high frequency
flux from the equatorial atmosphere as compared with the radio emission (particularly in the centimeter and millimeter
flux from the polar atmosphere. wave Imgths).
Many of the particles accelerated to high energies in a
solar flare will escape from the solar corona into the inter-
planetary medium. As the energetic charged particles move
6.3 ENERGETIC SOLAR PARTICLES into the interplanetary medium they will be guided along
(SOLAR COSMIC RAYS) the existing spiral magnetic field pattern. As a result, both
the intensity and the spectrum observed at the earth depend
Part of the energy released in solar flares maybe nuclei ac- somewhat on the relative positions of the earth and the flare
celerated to high energies and released into space. These solar on the sun. For example, a solar witid velocity of 400
particle events are commonly referred to by a number of de- km/s produces a spiral field that connects the earth to the so-
scriptive names such as solar cosmic ray events (SCR), solar lar longitude - 55° west of the central meridian of the sun
proton events, solar electron events, polar cap abso~tion events as viewed from the earth. The actual degree of “well con-
(PCA), and ground level events (GLE). Each of these names nectedness” between the earth and the sun depends on in-

6-16
 a

GALACTIC COSMIC RADIATION AND SOLAR ENERGETIC PARTICLES

I
— MEASURED, HESS, PATTERSON,
w &
10’
WLUCE, ANO CHUPP ( 1959)
10-” NEUTRON CURRENT I I 1 ~E

I O-s 1,,1,11 1 I 1 10 :B 11
,~~l~en-’,o+ l~sl~’l~~l~zl~’ 100 10’ Iozlo’ 10* 10’ 10 107 105 10’ 16’ 10’ 103 10’
ENERGY (MeV) ENERGY (MeV)

-=-––6-22.
Fieure ——Neutron soectra at various deDths from the top of the atmosphere. The data are adiusted to solar minimum conditions at the midlatitude
geomagnetic latitude of -42°: (equivalent to a-vertical cutof;rigidity of 4.5 GV)

terplanetary conditions at the time of the flare, and these events is correlated with the solar activity cycle; however,
conditions are highly variable and unpredictable. there is a large variance in the distribution. Significant solar
particle events have occurred during solar minimum. There
can be relatively long periods between significant solar par-
6.3.2 The Size and Frequency of Solar ticle events during the solar maximum. The elemental com-
Particle Events at the Earth position of the accelerated ions is not radically different
from the elemental composition of the solar atmosphere,
Major solar particle events occur at random, with a although composition differences as a function of energy
frequency that varies from approximately one every two exist as discussed in Section 6.3.3. The “size” of a solar
months to one every two years. Major solar particle events particle event depends on the energy of measurement, and
consist primarily of sizeable fluxes of energetic electrons consequently the “size” is dependent on the energy spectra.
and energetic protons. These solar flare initiated solar par- The particle events near the earth that result from these solar
ticle events may also contain a small and apparently variable flares may last from a few hours up to a maximum of
flux of heavier elements. The frequency of solar particle approximately ten days. The Catalog of Solar Proton Events

6-17
CHAPTER 6
SOLAR MINIMUM SOLAR MAXIMUM

I
A 0.0 GV
\ B 3.0 GV

&. C 4.6 GV
c’ D 6.0 GV
I .0 1.0
E 9.0 GV

k\\ \ F 12.0 GV
k G 17.0 GV 1

0.I 0.1

.
---- OBSERVED (PRIMED) ----OBSERVED (PRIMED)

t 7
0.01‘ I 1 I I I 1 0,0, ~

0 200 400 o 200 Soo

ATMOSPHERIC DEPTH G/cm2 ATMOSPHERIC DEPTH G/c#

Figure 6-23, Fast neutron flux versus atmospheric depth for various cutoff rigidities for solar minimum conditions (left panel) and solar maximum conditions
(right panel).

COSMIC RAY ALBEDO PROTONS

v r 1 , 1 1 1 , r 1 r 1 , 1 1 r, , r T IT, , ,
I I I 1 I 11 , 1 I i 111 1 1 1 I 111

I (a) _ _ + (b) _
●
x
Y *
--+
-f- ~, : ,’–}–’
+
++
6’
+; :
+
*
t +
HIGH LATITUDES PALESTINE, TEXAS
+ /
62r

0-3 , ,, mt 1 , 1 , 1 , , II , 1 1 , , 1, , , , , ,,‘, , , , I ,, I , I 1
I 10 I@ I 10 102 Id
KINETIC ENERGY (MeV)

Figure 6-24, Measured differential energy spectrum of the albedo proton flux at the “top” of the atmosphere, at two different locations [Wenzel et al,,
1975].

6-18
 GALACTIC COSMIC RADIATION AND SOLAR ENERGETIC PARTICLES
Ioc
SPLASH ALBEDO ELECTRONS
10 : I I I 1 I 1111 I I I 1 I I 11I J

L x -i

+ I 4

I ~
+1, ,

+
10-’ :
+“+:
L

e Iw Ic
ENERGY (MeV)

Figure 6-26. Spacecraft measurements of the low altitude alhcdo electron

flux [Kumosova, 1978].
10-2 r

,o-3~ 10 102
ENERGY (MeV)
103
“GLOBALLY AVERAGE””
COSMIC RAY ALBEOO
NEUTRON SPECTUM 1

Figure 6-25. Measured differential kinetic energy sPectmmofthedbedo

electron flux at the “top” of the atmosphere. The solid circles
are measurements at Palestine, Texas (verticat cutoff rigidity
-4.5 GV), andtheopen circles memeasumments at high
latitudes (Ft. Churchill, Canada), Adapted from Israel [1%9].

1955–1969 [Dodson et al., 1975] and Catalog of Solar

Proton Events 1970–1979 [Akiniyan et al., 1983] contain
considerable detail in describing solar particle events and
associated solar circumstances.

6.3.2.1 Solar Proton Events. A list of the major solar

proton events observed at 1 AU at the orbit of the earth is
given in Table 6-5. The solar proton event frequency dis-
tribution has a period of -11 years (the solar sunspot cycle),
and there is usually one anomalously large event in each
solar cycle. Apart from these anomalous events, the re-
,o-5~
maining events seem to be distributed as though the log (to 10 I00 600
base 10) of their sizes has a normal statistical distribution. NEUTRON ENERGY (MeV)
This is called a log-normal distribution. King [1974] has
found that the distribution of the solar proton fluences F is Figure 6-27. A model “global averaged” alhedo neutron spectrum [Claflin
a log-normal distribution above any energy threshold. The and White, 1974].

6-19
CHAPTER 6
Table 6-5. Maior solar uroton events observed at the earth.

Maximum Maximum Peak Flux Integrated

Neutron 30 MHz (cm-’ s-’) Intensity
Solar Particle Monitor Riometer (onmidirectional) ~ (cm-’)
Event Onset Increase Absorption >30 >30
Date PECS (Percent) (db) MeV MeV

1956 Feb 23 x34 4554 13 6 200* 6.5 X 108*

1956 Aug 31 X30 — 4.9 150* 2.5 X 107*
1957 Jan 20 X20 — 4.1 2 500* 3 x lo~*
1958 Mar 23 X20 — 3.2 1 200*
— [2 x 10’]J
1958 Mar 25 X30 10 1 500*( SC)
1958 Jul 7 X40 — 23.7 1 500* 3 x 108*
2 000*( SC)
1958 Aug 16 X40 — >15 200* 2 x lo’*
1958 Aug 22 X30 — 10.6 500* 5 x 107*
1958 Aug 26 X40 — 16.6 1 100* 5.3 x lo’*
1958 Sep 22 X30 — 10.6 50* 6 X 106*
1959 May 10 X40 22 6 000* 7 x 108*
8 000*( SC)
1959 Jul 10 X40 — 20 4 000* 8.8 x lo~*
1959 Jul 14 X40 — 23.7 11 OOO* 1.1 x 109*
1959 Jul 16 X42 10 21.2 16 000* 8.1 x 108*
18 000( SC)
1960 Apr 1 120 — 3 50* 2.7 X 10G*
1960 Apr 5 120 — 3.1 40* 2 x 106*
1960 Apr 28 (2)2 o — 3 30(3* 2.5 X 107*
1960 May 4 (1)2 4 290 5 230* 7 x 106*
1960 May 13 (1)2 o — 4.5 60* 4 x 106*
1960 Sep 3 221 4 2.7 240* 4 x lo’*
Maximum Maximum Peak Flux
Neutron 30 MHz (cm-’ s-’) Integrated
Monitor Riometer Intensity
Increase Absorption directional onmidirectional (cm-’)
“Date PECS (Percent) (db) >30 MeV >30 MeV >30 MeV

1960 NOV 12 444 135 21.2 12 000* 1.4 x 10’*

1960 NOV 15 443 88 >20 6 000* 5.2 x 108*
1960 Nov 20 (3)3 2 8 5 1 000* 6 X 107*
1961 Jul 11 Xlo — 1 20* 2 x lo’*
1961 Jul 12 (2)4 2 17 120* 1.0 x 108*
1961 Jul 18 333 24 11 2 500* 2.1 x 1O$*
1961 Jul 20 (1)2 2 7 5 3(30* 9 x lo’*
1961 Sep 28 220 3.3 11.0 I50* 2.4 X 10G*
1963 Sep 20 120 4 4.0 30 1 x 106
1965 Feb 5 110 — 1.3 50.0 250 1 x 106
1966 Mar 24 120 — 1.6 15.0 -1.0 x Iof
1966 Jul 7 121 3 2.1 30.0 4.4 x 105
1966 Aug 28 120 4 15.0 1.0 x 105
1967 May 25 330 11 32.0 1.7 x lob
1967 May 28 220 4.1 27.0 1.3 x 10’
1967 Dec 3 120 1.8 10.5 4.6 X 10s
1968 Jun 9 230 — 6.5 12.4 8.9 X 105
1968 Sep 28 110 1.2 0. I [6.9 X 10’]t
1968 Sep 29 121 1 1.7 19.0

6-20
 a
GALACTIC COSMIC RADIATION AND SOLAR ENERGETIC PARTICLES

Maximum Maximum
Neutron 30 MHz Peak Flux Integrated
Monitor Riometer (cm-z s-l sr-[) Intensity
Increase Absorption (directional) (cm-’)
Date PECS (Percent) (db) >30 MeV >30 MeV

1968 Ott 4–6 120 — 1.5 6.3 2.6 X 10s

1968 Ott 31 230 — 5.5 10.0
1.2 x lo6]t
1968Nov 1 230 — 4 11.7
1968Nov 18 233 14 12.5 404.0 1.7 x 10’
1968 Dec 3 120 — (slight) 1.7
[3.2 X 106]t
1968 Dec 5 230 — 4.7 31.0
1969 Feb 25 123 16 2.1 41.5 2.1 x 10’
1969 Mar 30 112 9 1.4 13.0 1.3 x Iofi
1969Apr 11 330 — >16 123.0 1.6 X 107
1969Nov 2 30 13 737.0 2.1 x 10’
1970 Jan 31 20 3 6.2 2.7 X 105
1970 Mar 6 30 — 5 0.9 1.0 x 105
1970 Mar 29 20 1.8 20.2 1.7 x 10’
1970 Jtl]y 24 20 — 4.5 0.8 5.8 X 104
1970Aug 14 20 — 3.0 2.7 3.9 x 105
1970 No; 5 120 — 3.5 1.7 2.8 X 105
1971 Jan24 333 26 14.5 408.0 2.7 X 107
1971 Apr 6 120 — 3.8 5.0 2.0 x 105
1971 Sep 1 233 16 5.2 162.0 1.3 x 107
1972 May 28 120 — 2.6 2.7 5.3 x 105
1972 Aug 4 442 20 >22 21000.0 6.2 X IOX
1972 Aug 7 342 8 3.1 384.0 3.0 x 107

PECS. The notation PECS stands for the three digit proton event classification system as defined by Smart and Shea
[197 1] as follows:

First Digit Second Digit Third Digit

E >10 MeV
Satellite Measured Daylight Polar 30 MHz Sea Level Neutron
Proton Intensity Riometer Absorption Monitor Increase
Digit cm-2s-’ sr-’

– 3 From 10-3 to l&2 — —

–2 From 10-2 to 1~’ — —
– 1 From 10-’ to 10° — —
O From 10° to 10’ No measurable increase No measurable increase
1 From 10’ to 102 Less than 1.5 dB Less than 3%
2 From 102 to 103 From 1.5 dB to 4.6 dB From 3% to 10%
3 From 103 to 104 From 4.6 dB to 15 dB From 10% to 100%
4 Greater than 104 Greater than 15 dB Greater than 100%
The first digit represents the measurement of the E > 10 MeV proton flux by a satellite within the earth-moon system, the second digit represents the 30
MHz absorption measured by a sunlit polar riometer, and the third digit represents the response of a high latitude sea level neutron monitor. An “X” in
the first digit indicates that not enough satellite data are available to make a determination of the maximum flux of protons greater than 10 MeV. A digit
within parenthesis is provisional, based on partial data.
*Proton flux data derived from Polar Cap absorption measurements.
(SC) Peak flux associated with geomagnetic storm.
[ ]’t Composite integrated proton flux of 2 separate events.

6-21
CHAPTER 6
Table 6-6. The parameters of the log-normal distributions for ordinary solar flares. Parameters are shown for the integral omnidirectional fluence
for the entire solar event in protons cm-z, and the peak omnidirectional flux in protons cm-z s-’.

E > 10 MeV E >30 MeV E>60MeV E > 100 MeV

Log Omnidirectional
Integral Fluence 8.27 + 0.59 7.28 ? 0.75 6.63 t 0.95 5.77 L 1.24
(mean t U)
Log Omnidirectional
Peak Flux 3.27 + 0.64 2.37 ? 0.82 1.88 f 0.78 —
(mean ~ U)
Data from King [ 1974].

means and standard deviations of the distributions for four Solar electron events are characteristically separated into
energy thresholds are shown in Table 6-6. two groups: low energy solar electrons (energy ranges in
the 10s of KeV) and high energy electrons (energies in the
6.3.2.2 Solar Electron Events. The data base for solar MeV range). (Since the rest mass energy of an electron is
electron events begins with their identification in 1965. Since 0.511 MeV, electrons with kinetic energy in the MeV range
it is difficult to measure electrons in a high flux of pene- are relativistic, and their velocity is nearly the speed of
trating protons, it was not until 1965 that the detector tech- light).
nology used on spacecraft was capable of uniquely distin- Low energy electrons are emitted by many solar flares
guishing between electrons and protons. Solar electrons are (even small ones), and there is a very good correlation
present in all major solar particle events and were probably between Type III radio bursts and the measurement of low
present, but not identified, in major events prior to 1965. energy solar electrons for “favorably” located solar flares.

Table 6-7. Average abundances relative to oxygen of energetic particle populations in the interplanetary medium.

Solar Flare Particle Events Solar Material

Iron-Rich Corotating
Normal (l-5 Particle Streams
Element (1-20 MeV) MeV) (1-10 MeV) Corona Photosphere

H 4600 300 2800 1780 1445

He 70 40 170 150 91*
Li -0.0005 -1.4 “ 10-”
Be <0.0007 -2 “ 10-x
B <0.0007 <1.8. 1U8
c 0.54 0.23 1.05 1.0 0.6
N 0.13 0.2 0.126
0 =1 =1 =1 =1 =1
F <0.002
Ne 0.16 0.39 0.2 0.063 0.054
Na 0.016 0.005 0.0028
Mg 0.18 0.54 0.13 0.079 0.050
Al 0.016 0.005 0.0048
Si 0.13 0.44 0.087 0.079 0.065
P <0.002 0.0005 0.0005
s 0.026 0.33 0.06 0.0251 0.023
Ar -0.004 0.01 0.0015
Ca 0.012 0.0045 0.0032
Fe 0.15 ‘ 1.3 0.095 0.093 0.05
---
Data rrom
. . . ------
uloe~ler 1IY /Yj.
*Note: The photosphere He abundance is not accurately measured and there is a large variance between the results of different measurement methods

6-22
 GALACTIC COSMIC RADIAI’iON AND SOLAR ENERGETIC PARTICLES
The frequency of the low energy solar electron events is @
highly correlated with solar flare frequency throughout the
Ill I I I I I I I 1 I I I I 1 I I I I I
solar sunspot cycle. The location of the solar flare is im-
portant since low energy solar electron event detection at
the earth is generally limited to heliolongitudes within ~ 90°
It? scfLAR mARE ARTICLES
of a position about 50° West heliographic longitude. 1 NORMAL
High energy solar electron events have a distinctly dif-
~ “Fe”, H@ rich
ferent time profile from the low energy solar electron events.
w ● SOLAR CORONA
The high energy solar electron event profile is similar to the
P % SOLAR ATMOSPHERE -
time-intensity profile of the 90 to 100 MeV protons. The
high energy solar electrons characteristically arrive at 1 AU E
after the observation of the low energy solar electrons emit-
ted from the same flare. See Simnett [1974] for a detailed
study of this phenomena.
There is a special class of solar electron events called
“scatter free” events. These events are very impulsive, with
short interplanetary propagation times, and short rise and
decay times. The data indicate that the solar electrons in
these events undergo very little scattering in the interpla-
netary medium between the sun and the earth. The scatter
free events have an upper energy limit of about 1 MeV, and
detection at the earth is restricted to a relatively narrow solar
flare source region about + 30° from the Archimedean spiral
path from the sun to the earth. See Section 6.3.4 for a
discussion of the most favorable propagation path between
the sun and the earth.

6.3.3 Elemental Composition of Solar I Illllllllj

Particle Events
1041
He Be BCNOFWw~ASi
I I 1 I i I I I
s a
I
H Li P Ar “Fe”
Average values of the elemental abundances of > 1
MeV/nucleon solar particles in relatively large solar flare Figure 6-28. Abundances of elements and element groups normalized to
oxygen in the solar corona (filled circles). solar atmosphere
events (peak intensities exceeding - 100 protons cm”2s-1 or photosphere (stars) and in solar flare particle events. The
sr-’ MeV- 1) with a “normal” composition are given in Table abundances for relatively large (peak intensities > – 100 pro-
6-7 and shown in Figure 6-28 as solid bars, open rectangles tons cm-2 s-’ sr] MeV-’ ) flare particle events with a normal
composition are indicated by solid vertical bars, sofid @-
(element groups) and solid triangles (upper limits). me angla (upper limits), open sq-s (single meas~ment with
abundances relative to oxygen of the individual elements or error bars) and open rectangles (element groups). The com-
element groups are derived from a number of individual position for “Fe’’-rich (atso 3He-nch) wlar flare particle even@
is indicated by cross hatched bars and rectangles and is baaed
fltie particle events. There is significant variability in the on abundance measurements in the 1 to 4.6 MeV/nucleon
composition of normal flare particle events, not only from range. The vertical extent of the bars for H, He, C, N, Ne,
one event to the next but sometimes during a given event. Mg, Si, S-Ca, and Fe in both the normal and Fe-rich events
indicates the degree of variability in the composition between
The vertical extent of the solid bars and open rectangles different solar flare particle events [Gloeckler, 1979].
indicates the degree of variability in the elemental com-
position. For the large events, variations in the abundances
from flare to flare, and with time during single flares, give
no indication that physical mechanisms preferentially ac- corona and photosphere whose abundances are also given
celerate any particular species over the energy range of in Table 6-7. There is overall agreement in the abundances
observations. Systematic chariges in the elemental compo- of He, C, N, and S through Fe; the elements Li, Be, B, F,
sition have been observed during individual solar particle and P are rare, and the odd Z (Z is the atomic number)
events and in successive flares from the same active region. elements N, F, Na, Al, and P are far less abundant than
In addition to the velocity dispersion there is a charge-to- their even Z neighbors. The energetic solar particle com-
mass dependent dispersion in the time-to-maximum for ele- position seems to be consistently enriched in the elements
ments having equal velocities. Ne to Si, and the solar flare C/O ratio, although consistent
Despite these variabilities the basic composition of solar with the solar photospheric abundance, is lower than the
flare particles is similar to the composition of the solar currently acceptable coronal abundance values. At this time

6-23
CHAPTER 6
little is known about elements heavier than Fe except that and plasma turbulence in the corotating interaction region
their abundances are very low. (CIR) provide evidence for the continuous acceleration of
The elemental composition of the solar particle flux is particles in the interplanetary medium between -2 and 5
basically similar for small, medium and large events; there AU. The acceleration mechanisms proposed most frequently
is, however, a tendency for the He/O ratio to be somewhat include both statistical processes resulting from magnetic
larger (- 70 to 90) for the small and medium events than and plasma turbulences that are generated when high and
the corresponding ratio for the large solar particle events low speed solar wind streams collide in the corotating in-
IGloeckler, 1979]. teraction region, and acceleration at the forward and reverse
Our knowledge of the isotropic composition of solar flare shocks which bound the corotating interaction region.
particles is at this time rather limited and generally confined Of the possible sources for the corotating energetic par-
to the lighter elements. Typical values of the H and He ticles, the most likely candidates are either the high-energy
isotropic ratios in the energy range of - 1 to 10 MeV/ tail of the solar wind or a possible but so far undetected,
nucleon averaged over a number of solar flare particle events low energy(< 200 keV/nucleon) residual particle popula-
are 2H/l H = 7(+ 10, –6) “ 10-6, ‘H/i H - 3 - 10-6 and tion, presumably of solar origin.
‘H/4H - (9 * 4) “ IW’. The isotropic abundances of solar
flare Ne in the energy range from 11 to 60 MeV/nucleon
have recently been reported to be 20Ne/22Ne = 7.6 * 2.
This value is distinctly lower than the comparable and well
6.3.4 Prediction of Solar Proton Events
established solar wind ratio of 13.1 * 0.6. The low
The capability to predict a phenomenon such as a solar
abundances of the carbon, oxygen, and neon isotopes
proton event illustrates our understanding of the processes
(“c/c - 1o-*,170/0 <3. 10-3, 180/0 -2. lm’, 21Ne/
controlling such a phenomenon. Solar protons are accel-
Ne < 10-2) indicate that element production by nuclear
erated in solar active regions during solar flare events. The
spallation reactions in the solar atmosphere is not significant.
x-ray, radio and optical emissions during the solar flare event
are the indicators that particle acceleration is occurring. The
6.3.3.1 Charged States of Solar Particles. Direct and “U-shaped” spectral signature in the radio emission peak
indirect determinations of the ionization states of solar par- power spectrum [Castelli et al., 1967; Castelli and Guidice,
ticles indicate incomplete stripping of the heavy ions over 1976] can be interpreted as the indicator that solar protons
an extended energy range up to at least - 5-10 MeVlnu- are being released from the solar active region (see Chapter
cleon. At high energies (z 10 MeV/nucleon), the ionic com- 11). After the energetic particles are released from the ac-
ponent of solar particle events seems to be completely stripped celerating region, they will travel through the solar corona
of electrons. Direct measurements have shown that, in gen- and some will propagate outwards along the interplanetary
eral, the charge states of C, O and Fe below 1 Mev/charge magnetic field lines to the position of the earth. A com-
in energetic solar particle events are (a) consistent with the puterized code to predict solar proton intensities at the earth
ionization states of these elements in the solar wind (charge after the occurrence of a specific solar flare has been de-
states of 6, 6 and 12 for C, O and Fe respectively), (b) do veloped by Smart and Shea [ 1979].
not vary substantially from event to event, and (c) remain
constant over the energy range of the measurements of - 50
6.3.4.1 Propagation of Solar Particles. The propaga-
to 1000 keV/charge. A number of indirect determinations
tion of solar protons from the flare site to the earth can be
of charge states of Fe at higher energies (- 1 to 10 MeV/
separated into two distinct phases. The first phase is dif-
nucleon) also indicate partial ionization.
fusion from the flare site through the solar corona to the
“foot” of the Archimedean sprial path formed by the inter-
6.3.3.2 Corotating Energetic Particle Streams. Coro- planetary magnetic field line between the sun and the earth.
tating energetic particle events are modest increases in the The second phase is the propagation in the interplanetary
intensity of protons and ci -particles in the range of < 1 to medium from the sun to the earth along the interplanetary
- 20 MeV/nucleon, that persist for three to six days and magnetic field lines. Figure 6-29 provides an illustration of
often reappear a number of times at - 27-day intervals. the propagation concepts.
These corotating or 27-day recurrent increases are not as- Propagation in the Solar Corona. The angular distance
sociated with solar flares and do not exhibit the velocity the solar particles travel in the solar corona from the pre-
dispersion often observed at the onset of solar flare particle sumed source (that is, the solar flare site) to the foot of the
events but are correlated with high-speed solar wind streams Archimedean spiral path from the sun to the earth is des-
and interplanetary magnetic field structures corotating with ignated by the symbol O in Figure 6-29. In a heliographic
the sun. Measurements of radial gradients and particle an- coordinate system the solar flare will have coordinates de-
isotropies in corotating events combined with the temporal- fined as solar latitude AE and heliocentric longitude $~.
spatial correlations observed at - 4 to 5 AU between the where ~~ is positive west of central meridian, and +~ is
-1 MeV/nucleon corotating particles and the magnetic field negative east of central meridian. If the “foot” of the Ar-

6-24
 a

COSMIC RADIATION AND SOLAR ENERGETIC PARTICLES

, 0-2 I 1 ( ,,! ,11 , ,1 ,1,1 , 1 , 1,11 t I ,,1,,1 ! , !ll ,111 , ,1

10-2 1o”’ ,00
I o) I 02 10’ I 04

RIGIDITY (MV)

Figure 6-30 Interplanetary mean free path Allas a function of charged

particle rigidity. These data have heen derived from an anal-
ysis of many solar particle events. The shaded areas indicate
the range of “scatter free” (upper region) and “normal” (lower
region) propagation in the interplanetary medium.

and are released into the interplanetary medium, they prop-

agate along the interplanetary magnetic field lines. (See
Chapter 3 for a detailed description of the solar wind. )
The spiral structure of the interplanetary magnetic field
shown in Figure 6-29 is highly idealized. On a microscale
there are many irregularities in the inte~lanetary medium.
The energetic solar particles propagating through the inter-
planetary medium will undergo random scattering and the
same general theory described in Section 6.2.2.1 applies.
Figure 6-29. Illustration of the concept of solar particle propagation from
the sun to the earth. The coronal propagation distance 0 is The mean free path between scattering centers is quite vari-
illustrated by the heavy arc on the sun. Interplanetary prop- able, depending on both the particle energy and the amount
agation proceeds along the Archimedean spiral path from
of turbulence in the interplanetary medium. A summary of
the sun to the earth.
the general range of mean free paths determined from a
~umher of sol~r-nafiicle events is-illustated in Figure 6-30.
chimedean spiral is at heliocentric longitude $*, then the During the sun to earth phase of their propa~ation, the
heliocentric longitudinal distance from the flare site to the MeV solar particles appear to have an average mean free
“foot” of the Archimedean spiral will be I+* – $~1. It is path length of the order of 0.1 to 0.3 AU; consequently,
assumed that coronal propagation is a function of 6, and over the 1 AU radial distance, diffusion perpendicular to
from diffusion theory, it is expected to be proportional to the interplanetary magnetic field line is negligible. The en-
62 [Wibberenz, 1974]. ergetic particles travel essentially along the interplanetary
As a result of diffusion in the solar corona and the magnetic field lines with a velocity that is a function of the
inherent assumption that some stochastic processes are ,op- particle energy. The minimum propagation delay will be for
erating, there will be a solar particle gradient existing in the particles that essentially travel along the inte~lanetary mag-
solar corona such that the proton intensity decreases as a netic field lines with miminal scattering, so for “scatter free”
function of distance from the flare site. The observational onsets the propagation time from the sun to the earth will
evidence shows that the gradient may vary from case to be the distance traveled along the Archimedean spiral path
case. An observer at one astronomical unit who is connected divided by the particle velocity.
via the interplanetary magnetic field to the heliographic po- Some scattering certainly occurs in the inte~lanetary
sition of the flaring region would observe the maximum medium, and to determine the time of maximum intensity,
possible particle intensity. An observer whose interplanetary diffusion theory is applicable. Almost all theones involving
magnetic field connection is at a distance of 6 from the solar particle transport in the interplanetary medium show
flaring location would observe a reduced intensity. Obser- that the time of maximum is proportional to the square of
vational data suggest an average gradient may be about one the distance traveled. The distance travel led can be obtained
order of magnitude per radian of heliocentric distance. by integrating along the Archimedean spiral path, Normal-
Propagation in the Interplanetary Medium. After the ized to the earth’s orbit at 1 AU, in a heliographic coordinate
energetic solar particles propagate through the solar corona system, the Archimedean spiral path length, D, is

6-25
CHAPTER 6

In+. + ~
$A 1 (6.19)

where ~~ is the heliographic longitude of the “foot point”

where the Archimedean spiral path to the earth connects to
the sun. If the solar wind plasma carries “frozen in” mag-
netic fields, then the heliographic longitude of the “foot
point” of the interplanetary magnetic field line that extends
from the sun to the earth is given by@* = (@,r)/V,W, where -.. L.– ..- L——J I I
90” 70” 50” 30” 10’ “ 10” 30” 50” 70” 90”
~, is the solar synodic rotation rate (13.3 degrees/day), r is East SOLAR LONGI1 )E OF FLARE West

1 AU, and V,W is the velocity of the solar wind in krrt/s at

the earth. If the Archimedean spiral equation in the polar Figure 6-31, Distribution of observed onset times of 30 MeV protons as
a function of solar longitude. The data points are the meas-
form is normalized to 1 AU, then ~~ = 404/V,W where
urements of Barouch et al. [1971].
404 is the nominal solar wind velocity when the solar ro-
tation is equal to one radian during the time it takes the
plasma element to travel 1 AU. where D is the distance along the length of the Archimedean
The velocity of a proton is given by spiral from the sun to the earth and ~ is the proton velocity.
This calculation can be repeated for each proton energy
1/2 (velocity) desired to obtain the velocity dispersion of the
p=

[
~-
(E/~c2
1
+ 1)2 1 (6.20) predicted onset times as a function of energy.

6.3.4.3 Time of Maximum Particle Intensity. The dis-

tribution of the time of the maximum of 20 to 80 MeV
where E is the kinetic energy of the particle in MeV, ~ protons observed at the earth for a number of proton events
is the velocity of the particle compared with the velocity of [Van Hollebeke et al., 1975] as a function of solar longitude
light, and the rest energy of the proton ~c2 has the value is illustrated in Figure 6-32 where the heavy solid line has
of 938.232 MeV. (Note: ~ = V/c. ) the functional form of 862. The minimum in the curve cor-
The minimum time in hours required for a proton to responds to a flare at the “foot point” of the Archimedean
propagate along the Archimedean spiral path from the sun spiral path between the earth and the sun which would be
to the earth with zero pitch angle and no scattering is 0.133 about 57.2° for a nominal solar wind of 404 km/s.
D/~., The time of maximum can be computed by assuming
the particles have an average bulk velocity that is about half
6.3.4.2 Particle Increase Onset Time. The delay time
from the solar flare until the onset of the particle increase
at the earth is the combination of the coronal propagation
time and the time for the particles to propagate from the
sun to the earth along the interplanetary magnetic field lines.
The distribution of onset times expected for 30 MeV
protons for nominal solar wind speeds is shown in Figure
6-31. In this figure the heavy solid line has the functional
form of 402; the data points shown on the figure are taken
from Barouch et al. [1971] and indicate typical variations
that may be expected. The minimum in the figure corre-
sponds to a flare at the heliographic longitude of the “foot
point” of the nominal Archimedean spiral path between the
sun and the earth (57. 2° West of central meridian). The
initial onset will be for the fastest (highest energy) protons.
The onset time is given by HELIOLONGITUOE WITH RESPECT TO THE CENTRAL MERIOIAN

Figure 6-32. Distribution of observed times of maximum of 20 to 80 MeV

0.133 D
T~ (in hours) = — + 402, (6.21) protons as a function of solar longitude. The data points are
P the measurements of Van Hollebeke et al, [ 1975].

6-26
 GALACTIC COSMIC RADIATION AND SOLAR ENERGETIC PARTICLES
of the particle speed that would be computed from the par- particle longitudinal gradient will combine to increase the
ticle energy. Most theoretical models set the time of max- flux decay rate at the earth.
imum proportional to the square of the distance traveled
along the magnetic field line from the sun. The time of 6.3.4.6 Differential Energy Spectra. The slope of the
maximum intensity is computed by differential energy spectrum of solar proton events observed
at the earth is summarized by Van Hollebeke et al. [ 1975]
0.133 D2 and shown in Figure 6-33. Assuming that the hardest spectra
T~ (in hours) = 2.0 + 802. (6.22)
P (the minimum in the solid curve of Figure 6-33) should be
along the interplanetary magnetic field lines leading away
from the flare site, the average spectral slope is observed
6.3.4.4 Magnitude of Maximum Intensity. The maxi-
to vary as
mum intensity is predicted by converting an electromagnetic
emission parameter (usually radio flux or x-ray flux) to a
proton flux. Many different algorithms can be used but
currently the best predictors are the relationships between
integrated radio flux and peak proton flux. The maximum
y~ = –2.7
() 1.0 + : (6.24)

6.3.4.7 Prediction of Polar Cap Absorption. It is pos-

particle intensity is expected to be along the Archimedean
sible to convert predicted proton flux to prediction of the
spiral path from the flare. The maximum particle intensity
profile of the riometer absorption in the polar ionosphere.
expected at the earth is derived by attenuating the flux through
the coronal gradient over the heliocentric distance in the It has been shown that riometer absorption is proportional
corona between the flare position and the solar equatorial to the square root of the intergal proton flux above some
specifi~ energy. A convenient relationship is Jf> 10) = 10A2,
longitude of the “foot point” of the Archimedean spiral path
where J(, ,0) is the flux of protons with energy greater then
from the sun to the earth.
10 MeV and A is the absorption in decibels for a 30 MHz
nometer. The relationships developed by Sellers et al. [1977]
6.3.4.5 Proton Event Decay. The decaying portion of a
are simple equations that are relatively independent of the
solar particle event is normally exponential in character with
slope of the solar proton energy spectrum. These authors
a nominal time constant for the flux to decay by a factor of
selected the integral proton flux greater than 5.2 MeV and
l/e although there is a slight energy dependence, and there
2.2 MeV to be a good predictor of the riometer absorption
can be a large variability between events.
in the sunlit polar ionosphere and the night time polar ion-
The decaying portion of the event can be modeled after
osphere respective y. These equations are
the principles of diffusive particle transport in the interpla-
netary medium. One such method is collimated convection
[Roelof, 1973]. This results in a l/e decay constant of A (day) = 0.6[J~>s,2)]1’2 (6.25)

A (night) = 0.25[J(>2,Z)] 1’2. (6.26)

3D
Td (in hours) = (6.23)
4v,w([7\
+ 1) 3600’

where T~ is the l/e decay constant, D is the distance along

the Archimedean spiral path, V,Wis the solar wind velocity,
and y is the differential energy spectral exponent.
Effect of Coronal Gradient on Apparent Decay Rate.
The assumption that there is a particle gradient in the corona
combined with the solar rotation, results in a decay rate at
the earth that is dependent on the solar flare location. After
the particle flux maximum, the flux along the interplanetary
magnetic field lines can be considered as a corotating struc-
ture being convected out into space. If the particle source * ~
1
is a flare on the eastern hemisphere or near central meridian
1 1 1 I 1 1 1 [ I 1 1 $ 1 , , j
of the sun, the particle coronal longitudinal flux gradient t’
EAST 120. 80. 40° 4W 80. 120. i60- wEST
0.
rotates toward the “foot point” of the Archimedean spiral HCLIOLONGITUDE WITH *CSPECT TO THC CENTRAL MC R1O!AN

from the sun to the earth; thus the apparent decay rate will
be slower than the radial convection rate out from the sun. Figure 6-33. Variation of the absolute value of spectral index y in the
If the particle source rotates away from the “foot point ,“ as 20-80 MeV range as a function of the heliolongitude. The
dashed contour lines enclose 92% of the “normal” events.
is the case for flares that occur on the extreme western The solid line is a least squares fit that can be represented
hemisphere of the sun, then the solar rotation and the coronal by 2.7[1 + 13/2].

6-27
CHAPTER 6
REFERENCES

Adams, J.H. Jr., R. Silverberg, and C.H. Tsao, “Cosmic Forman, M.A. and O.A. Schaeffer, “Cosmic Ray Intensity
Ray Effects on Microelectronics, Part I: The Near Earth Over Long Time Scales,” Ret’. Geophvs. and Sp. Phys.,
,-
Particle Environment,” NRL Rept. 4056, Naval Re- 17:552, 1979.
search Laboratory, Washington, D. C. 1981. Garcia-Munoz, M., G. M. Mason, and J .A. Simpson. “The
Akiniyan, S. T., G.A. Bazilevskaya, V.N. Ishkov, et al., Anomalous He component in the Cosmic Ray Spectra
Catalog of Solar Proton Events 1970—1979, Institut of s 50 MeV per Z Nucleon During 1972– 1974,” As-
Zemnogo Magnetizma, Ionosfery I Rasprostraneniya trophys. J., 202:265, 1975.
Radiovoln (Institute of Earth Magnetism, Ionosphere Gleeson, L.J. and W .1. Axford, “Cosmic Rays in the In-
and Radiowave Propagation), Academy of Science, terplanetary Medium,” Astrophys. J., 149: L1 15, 1967.
USSR, 1983. Gleeson, L.J. and W.1. Axford, “Solar Modulation of Ga-
Allkover, O.C. and P.K.F. Grieder, Cosmic Rays on Earth, lactic Cosmic Rays,” Astrophys. J., 154: 1011, 1968.
Physics Data 25-1, Fachinfomationszentrum Energie, Gloeckler, G. “Composition of Energetic Particle Popula-
Physik, Mathematik, Gmbh, Karlsruhe, Germany, 1984. tions in Interplanetary Space ,“ Rev. Geophys. Space
Armstrong, T. W., K.C. Chandler, and J. Barish, “Calcu- Phys., 17:569, 1979.
lations of Neutron Flux Spectra Induced in the Earth’s Ha ffner, J. W., “Radiation and Shielding in Space,” Nuclear
Atmosphere by Galactic Cosmic Rays,” J. Geophys. Science and Technology 4, Academic Press, New York,
Res., 78:2715, 1973. 1967.
Barouch, E., M. Gros and P. Masse, “The Solar Longitude Hess, W. N., H. W. Patterson, R. Wallace, and E.L. Chupp,
Dependence of Proton Event Decay,” Sol. Phys., 19: “Cosmic-ray Neutron Energy Spectrum,” Phys. Rev.,
483, 1971. 116:445, 1959.
Cameron, A. G. W., “Elementary and nuclidic abundances Israel, M. H., “Cosmic Ray Electrons Between 12 MeV and
in the solar system, “ in Essays in Nuclear Astrophysics, 1 Gev in 1967,” J. Geophys. Res., 74:4701, 1969.
edited by C.A. Barnes, D.D. Clayton, and D.N. King, J. H., “Solar Proton Fluences for 1977–1983 Space
Schramm, Cambridge University Press, Cambridge, Missions,” J. Spacecraft and Rockets, 11:401, 1974.
England, 1981. Kumosova, L. V., “Charged Particle Flux at 200 to 300 km
Castelli, J .P., and D.A. Guidice, “Impact of Current Solar above the Earth” in Cosmic Rays in the Stratosphere
Radio Patrol Observations,” Vistas in Astron., 19:355, and Near Space, edited by N. G. Basov, Lebedev, Trudy
1976. 88, Proceedings of the P.N. Lebedev Institute, 88:137,
Castelli, J .P., J. Aarons, and G.A. Michael, “Flux Density 1976. Engl transl Consultants Bureau, New York, 1978.
Measurements of Radio Bursts of Proton-Producing and Kykler, G.C. and A .R. Liboff, “Absolute Cosmic Ray Ion-
Nonproton Flares,” J. Geophys. Res., 72:5491, 1967. ization Measurement in a 900-Liter Chamber,” Y. Geo-
Chupp, E. L., D.J. Forrest, J.M. Ryan, J. Heslin, C. Rep- phys. Res., 82:5539, 1978.
pin, K. Pinkau, G. Kanbach, E. Rieger, and G.H. Share, Levy, E. H., S. P. Duggal, and M.A. Pomerantz, “Adiabatic
“A Direct Observation of Solar Neutrons Following the Fermi Acceleration of Energetic Particles Between Con-
0188 UT Flare on 1980 June 2 1,“ Astrophys. J. Lett., verging Interplanetary Shock Waves,” ~. Geophys. Res.,
263: L95–L99, 1982. 81:51, 1976.
Claflin, E.S. and R. S. White, “A Study of Equatorial Inner Lezniak, J.A. and W. R. Webber, “The Charge Composition
Belt Protons from 2 to 200 MeV,” J. Geophys. Res., and Energy Spectra of Cosmic-Ray Nuclei from 3000
79:959, 1974. MeV per Nucleon to 50 GeV Per Nucleon,” Astroph.vs.
Dodson, H. W., E.R. Hedeman, R.W. Kreplin, M.J. Martres, J., 223: 1978.
V.N. Obridko, M.A. Shea, D.F. Smart, and H. Tanaka, Light, E. S., M. Merker, H.J. Verschell, R.B. Mendell, and
Catalog of Solar Particle Events 1955-1969, Vol. 49 S .A. Korff, “Time Dependent World Wide Distribution
of the Astrophysics and Space Science Library, edited of Atmospheric Neutrons and of Their Products, 2. Cal-
by Z. Svestka and P. Simon, D. Reidel, Dordrecht, culations,” J. Geophys. Res., 78:2741, 1973.
Holland, 1975. Lingenfelter, R. E., “The Cosmic Ray Neutron Leakage Flux,”
Ely, J .T. A., “Atmospheric Depth and Effective Solid Angle J. Geophys. Res., 68:5633, 1963.
for Radiation Penetrating the Atmosphere,” Geophysics McCracken, K. G., U.R. Rae, B.C. Fowler, M.A. Shea,
Research Papers No. 74 (AFCRL 62-260), Air Force and D.F. Smart, “Cosmic Ray Tables,” lQSYInstruction
Cambridge Research Laboratories, 1962. Manua/ No. 10, IQSY Committee, London, 1965.
Eraker, J.H. and J .A. Simpson, “Origins of the LOW Energy McCracken, K. G., U.R. Rae, B.C. Fowler, M.A. Shea.
Relativistic Inte~lanetary Electrons,” Proceedings of the and D. F. Smart, “Cosmic Rays (Asymptotic Directions,
17th International Cosmic Ray Conference Paris 1981 etc)” in Chapter 14 of Annals of the IQSY, Vol. 1,
Conference Papers, 3:279, Centre d’Etudes Nucleaires, (Geophysical Measurements: Techniques, Observa-
Saclay, France, 1981. tional Schedules and Treatment of Data) edited by C. M.
Fanselow, J. L. and E.C. Stone, “Geomagnetic Cutoffs Minnis, 198, MIT Press. Cambridge, Mass., 1968.
for Cosmic-Ray Protons for Seven Energy Intervals be- McDonald, F. R., ed., So/ar Proton Manual, National Aer-
tween 1.2 and 39 MeV,” J. Geophys. Res., 77:3999, onautics and Space Administration Technical Report,
1972. NASA TR-R-169, 1963.

6-28
 w

GALACTIC COSMIC RADIATION AND SOLAR ENERGETIC PARTICLES

McIlwain, C. E., “Coordinates for Mapping the Distribution Rigidities for 1965 and 1975,” 14th International Cosmic
of Magnetically y Trapped Particles ,“ J. Geophys. Res., Ray Conference, Conference Papers, 4: 1298-1303,
66:3681, 1961. 1975. ,
Mewaldt, R. A., “The Elemental and Isotopic Composition Shea, M .A., and D.F. Smart, “A World Grid of Calculated
of Galactic Cosmic Rays,” Proceedings of the 17th In- Cosmic Ray Vertical Cutoff Rigidities’for 1980.0,” 18th
ternational Cosmic Ray Conference, Conference Pa- International Cosmic Ray Conference, Conference Pa-
pers, 13:49, Centre d’Etudes Nucleaires, Saclay, France, pers, 3:415, 1983.
1981. Shea, M. A., D.F. Smart, and John R. McCall, “A Five
Mur&ami, K., K. Nagashima, S. Sagisaka, Y. Mishima, Degree by Fifteen Degree World Grid of Trajectory-
and A. Inoue, “Response Functions for Cosmic-Ray Determined Vertical Cutoff Rigidities,” Can. J. Phys.,
Muons at Various Depths Underground,” Nuovo Ci- 46: S1098, 1968.
mento, 2C: 635, 1979. Simnett, G. M., “Relativistic Electron Events in Interpla-
Neher, H. V., “Cosmic Ray Particles that Changed from netary Space,” Space Sci. Revs., 16:257, 1974.
1954 to 1958 and 1965,” J. Geophys. Res., 72: 1527, Smart, D.F. and M.A. Shea, “Solar Proton Event Classi-
1967. fication System,” Sol. Phys., 16:484, 1971.
Neher, H. V., “Cosmic Rays at High Latitudes and Altitudes Smart, D.F. and M.A. Shea, “PPS76 - A Computerized
Covering Four Solar Maxima,” J. Geophys. Res., 76: “Event Mode” Solar Proton Forecasting Technique,”
1637, 1971. Solar-Terrestrial Predictions Proceedings, Vol. I: Pre-
O’Brian, K., “Calculated Cosmic Ray Ionization in the Lower diction Group Reports, edited by R.F. Donnelly, En-
Atmosphere,” J. Geophys. Res., 75:4357, 1970. vironment Research Laboratories, National Oceanic and
O’Brian, K., “Secular Variations in the Production of Cos- Atmospheric Administration, U.S. Department of Com-
mogenetic Isotopes in the Earth’s Atmosphere ,“ J. Geo- merce, Boulder, CO, 406, 1979.
phys. Res., 84:423, 1979. Stassinopoulos, E. G., and J. H. King, “An Empirical Model
Peddie, N. W., “International Geomagnetic Reference Field: of Energetic Solar Proton Fluxes with Application to
The Third Generation,” J. Geomag. Geoelectr., W 309, Earth Orbiting Spacecraft,” NASA X-601-72-487, 1972.
1982. Stuiver, M. and P.D. Quay, “Changes in Atmospheric Car-
Roelof, E. C., “New Aspects of Interplanetary Propagation bon- 14 Attributed to a Variable Sun,” Science, 207:11,
Revealed by 0.3 MeV Solar Proton Events in 1967,” in 1980.
Solar-Terrestrial Relations., 411, University of Cal- Van Hollebeke, M. A. I., L.S. Ma Sung, and F.B. Mc-
gary, Canada, 1973. Donald, “The Variation of Solar Proton Energy Spectra
Rossi, B., High Energy Particles, Prentice-Hall, New York, and Size Distribution with Heliolongitude,” Sol. Phys.,
1952. 41:189, 1975.
Rossi, B., Cosmic Rays, McGraw Hill, New York, 1964. Wenzel, K.-P., E.C. Stone, and R. E. Vogt, “Splash Albedo
Sellers, B., F.A. Hanser, M.A. Stroscio, and G. K. Yates, Protons Between 4 and 315 MeV at High and Low
‘&TheNight and Day Relationships betweeen Polar Cap Geomagnetic Latitudes,” J. Geophys. Res., 80: 358,
Riometer Absorption and Solar Protons,” Radio Sci., 1975.
12:779, 1977. West, G.S. Jr., S.J. Wright, and H.C. Euler, (eds. ), “Space
Seltzer., S. “SHIELDOSE: A computer code for Space- and Planetary Environment Criteria Guidelines for use
shielding Radiation Dose Calculations ,“ NBS Technical in Space Vehicle Developments,” NASA-TM-7811 9,
Note 1116, National Bureau of Standards, U.S. De- 1977.
partment of Commerce, Washington, D. C., May 1980. Wibberenz, G., “Interplanetary Magnetic Fields and the
Shea, M.A. and D.F. Smart, “A Five by Fifteen Degree Propagation of Cosmic Rays,” J. Geophys., 40: 667,
World Grid of Calculated Cosmic-Ray Vertical Cutoff 1974.

6-29
 Chapter 7

THE CHARGING OF SPACECRAFT SURFACES

H.B. Garrett

The buildup of static charge on satellite surfaces is an analysis is applicable to current spacecraft charging prob-
important issue in the utilization of satellite systems. The lems. This is true as in a very real sense the space vehicle
analysis of this spacecraft environmental interaction has re- itself can be considered as a “floating probe. ” This theme,
quired important advances in basic charging theory and the of the vehicle as a probe, will form the basis of much of
development of complex codes to evaluate the plasma sheaths this chapter.
that surround satellites. The results of these theories and Probe theory has developed into an important subfield
calculations have wide application in space physics in the of plasma physics in its own right, see review by Chen
design of systems and in the interpretation of low energy [1965]. It has only been with the advent of rockets and
plasma measurements. ultimately satellites that the charging of objects in space has
In this chapter, those aspects of charge buildup on sat- become a major separate area of concern. The first period
ellite surfaces relevant to the space physics community are of charging studies, in fact, was concerned with the potential
summarized. The types of charging processes, models of of interstellar dust grains. One of the earliest studies, that
charge buildup, satellite sheath theories, and charging ob- of Jung [1937], found that photoemission and electron ac-
servations are described with emphasis on basic concepts. cumulation were probably the dominant processes in inter-
As many books and monographs on specific aspects of stellar space. This subject was extended and put on a firm
the charging of bodies in space have appeared in the last physical footing by Spitzer [1941; 1948], Spitzer and Sav-
two decades, it is difficult to cover all areas in detail in a edoff [1950], and others [Cemuschi, 1947; Opik, 1956].
chapter of this nature. Rather, the chapter is limited to the Depending on the assumed “sticking” probability of the
charging of spacecraft surfaces in the near-earth magneto- electrons, the photoemission yield, and the ambient envi-
sphere. Rocket measurements are only briefly treated. The ronment, the estimate satellite to space potentials for these
reader is referred to books by Singer [ 1965], GrNd [1973a], early studies ranged between – 3 and + 10 V.
Rosen [1976], Pike and Loveil [1977], and Finke and Pike With the advefit of rocket-borne sensors in the early
[1979] that contain papers on the charging of natural bodies 1950s, spacecraft charging emerged as a discipline. Perhaps
such as the moon [Manka, 1973; Freeman et al., 1973], the first example of a spacecraft charging effect is in a paper
dust particles [Feuerbacher et al., 1973] and other planetary by Johnson and Meadows [ 1955]. They theorized that shifts
bodies [Shawhan et al., 1973]. A brief historical review of in the ion peaks measured by their RF mass spectrometer
spacecraft charging is followed by a discussion of obser- above 124 km could be explained by a negative rocket
vations. Following a description of the major charging potential of – 20 V. The first treatment of the ch~ging of
mechanisms, specific spacecraft charging models are ex: a macroscopic object was published the next year by Lehnert
amined with emphasis on basic concepts. A substorm worst [1956] who estimated satellite to space potentials of – 0.7
case environment is included. The chapter concludes with to – 1.0 V when ion ram effects were included.
a discussion of spacecraft charge mitigation techniques. A Not only did 1957 see the launch of Sputnik, but it
more detailed version of this chapter can be found in Garrett ushered in a second phase in spacecraft charging studies.
[1981]. Gringauz and Zejikman [1958] discussed the distribution of
charge (or sheath) around a space vehicle in the ionosphere
and the influence of photoemission fid satellite velocity.
7.1 SPACECRAFT CHARGING Jastrow and Pearse [1957], while neglecting photoemission
HISTORICAL PERSPECTIVE but including ram effects, computed the drag on a satellite
caused by charged particles in the ionosphere. Their study
The historical roots of spacecraft charging analysis can for the ionosphere assumed TE > T, so that the estimated
be traced to the early electrostatic probe work of Langmuir potentials were between – 10 V (night) and -60 V (day),
[Langmuir, 1924; Mott-Smith and Langmuir, 1926]. Not values which are too high as TE z T1 is the actual case
only is the Langmuir probe still an important space plasma [Brundin, 1963]. Another basic assumption of their study,
instrument, but as will be discussed, much of Langmuir’s as in the related study of Chrmg and Smith [1960], is that

7-1
CHAPTER 7
the ion density is little changed in the immediate vicinity effects (-O. 14 V). During this period more believable po-
of the vehicle. tential measurements were also reported for rocket probes
Zonov [1959] and Beard and Johnson [1960] analyzed as exemplified by Sagalyn et al. [1963] who found potentials
the effects of electric fields induced by the movement of a of –0.4 V (150 km) to –1.7 V (450 km).
satellite across the earth’s magrietic field; these fields can In a 1961 paper, Kurt and Moroz [1961] predicted po-
be quite important for large strictures. Beard and Johnson tentials of – 3.2 to 4 V outside the radiation belts. They
[1961 ] also discussed the effects of emitting charged par- also predicted potentials as high as – 20 kV in the radiation
ticles from a vehicle and the lirtiitations on vehicle potential belts. Although only crude estimates, their predictions an-
in the ionosphere, another issue which is still of concern ticipated the – 20 kV potentials observed in eclipse on the
(see Section 7.4. 1). In the same period, the first satellite geosynchronous ATS 6 Satellite in the 1970s. These and
potential measurements were made by Sputnik 3. Krassov- other results were compared in an excellent review of the
sky [1959] found a potential of – 6.4 V and TE = 15000 effects of charged particles on a satellite by Brundin [ 1963].
K at 795 km. Another good review from this period, in which the validity
The first review of spacecraft charging appeared in 1961 of various ionospheric measurements were discussed, is that
[Chopra, 1961]. Despite difficulties with this review (Cho- of Bourdeau [1963]. The most complete works of this pe-
pra predicted high positive potentials since the photoelectron riod, however, are in the first book concerned with space-
flux he assumed was too high), it can be used as a convenient craft charging [Singer, 1965] and the thesis of Whipple
marker for the end of the second phase of the study of [1965]. Whipple’s thesis brings together most of the pre-
spacecraft charging. By 1961 most of the elements of current ceding results in an analysis of the roles that secondary
spacecraft charging theory were in place. Preliminary ob- emission, backscatter, photoemission, and magnetic field
servations by rockets and satellites had confirmed that charg- effects have in spacecraft charging. As such, it completes
ing existed and, in agreement with some theones, was on the third period of spacecrti charge analysi= period marked
the order of a volt (typically negative) in the ionosphere. by a realization of the importance of spacecraft charging
Photoemission and the ambient electron flux were recog- for plasma measurements and of the importance of self-
nized as dominant sources and v X B effects had also been consistent calculations. Quantitative measurements also be-
considered. On the negative side, secondary emission and came available for the ‘first time. The period 1965 to the
backscatter had not really been adequately considered present has been primarily one of refinement and extension
[Whipple, 1965], charged particle drag (which was uhi- of these 1961 –1 965 results to higher altitude regimes and
mately shown to be of less importance than errors in the more complex geometric situations. It represents the “fourth
neutral drag coefficient (see reviews by Brundin [1963] and period” of spacecraft charging and is the concern of this
deLeeuw [1967])) was still of primary concern, and self- chapter.
consistent solutions of the particle trajectories and fields had
not been carried out. Only monoenergetic or Maxwellian
distributions were being considered. 7.2 SPACECRAFT CHARGING
The new phase of studies in spacecraft charging actually OBSERVATIONS
started somewhat earlier than 1961 with Bernstein and Ra-
binowitz’s probe study. Bernstein and Rabinowitz [1959]
introduced a means of calctdating particle trajectories in the 7.2.1 Rocket Measurements
vicinity of a probe for the case where collisions could be
ignored. The importance of their study (see Section 7.4.4) Rocket measurements of the ionospheric plasma have
was that their method allowed a self-consistent solution of been routine since the early 1950s. The first observations
the time-independent Vlasov equation and Poisson’s equat- of spacecraft potentials were probably the RF spectrometer
ion. Their work was later adapted to spacecraft by a number measurements on a rocket by Johnson and Meadows [ 1955].
of investigators. Davis and Harris [1961], by a more sim- They made use of the differences in the energy shifts of
plified method, calculated the shielding of a rapidly moving different ionized species to estimate a vehicle potential of
sphere in the ionosphere by estimating ion trajectories in a – 20 V above 120 km. Sagalyn et al. [1963], employing 2
fixed electron sheath. Their results concerning the distri- spherical electrostatic analyzers mounted on a Thor rocket,
bution of ions in the wake, despite their neglecting the ion found much lower potentials of – 0.4 V ( 150 km) to – 1.7
thermal velocity, are consistent with Explorer 8 measure- V (450 km). Their measurements are in good agreement
ments reported in the same year [Bourdeau et al., 1961]. with subsequent satellite measuretnents in the same region.
These latter results are probably the first accurate spacecraft Narcisi et al. [1968] also found rocket potentials of about
potential measurements between 425 km and 2300 km. The – 0.5 V in the D and E regions although their results may
ion current, which varied strongly as a function of angle have been affected by the potential distribution near the
relative to the satellite velocity vector, was found to agree rocket [Parker and Whipple, 1970]. As discussed by Parker
with Whipple’s [ 1959] theory (see Equation (7.23)). A sat- and Whipple [1970], however, there are for these and similar
ellite potential of – O. 15 V was observed along with v X B satellite measurements difficulties in interpreting the results

7-2
 CHARGING OF SPACECRAFT SURFACES
since the detailed particle trajectories must be considered in ‘ELECTRON oEt4sITY (electrons/cm3 )
,.5 1(
104
determining the actual instrument responses. Further, as 270.’~3

probably happened to Johnson and Meadows [1955], the 260 -

‘ 1 “’’’”
electric fields near the rockets may be perturbed by exposed 250 – /-Electron Density

potential surfaces. See, for example, the results of the recent 240 - /’

“tethered payload’ experiment of Williamson et al. [ 1980],

in which a potential of + 10 V on the main payload induced
a potential of – 5 V on the secondary payload 40 m away.
Olsen [1980] (see also Winckler [ 1980]) has described
~ 190 L
potential measurements from a number of rocket beam ex- .— THERMAL EMISSIVE pROBE
~ 180 -
periments. These ranged from the Aerobee flight [Hess, 6
----- BIPOLAR VOLTMETER
I 70 -
1969] through the Echo series [Hendrickson, 1972; Winck- 160 -
ler, 1976] down to the recent ARAKS [Cambou et al., I 50
1978], Precede, and Excede [0’Neil et al., 1978a, and b]. I 40 -
Although extensive literature exists on many of these flights, 130 -

the majority is concerned with the beam aspects. Potentials, 120 -

when observed, were typically a few tens of volts positive 110 -

I I I
or negative relative to the ambient plasma (Jacobsen and 1001 ,.O 10’ ,Oc ,
I
Maynard [ 1980], however, have reported potentials of POTENTIAL DIFFERENCE [volts]

hundreds of volts on the POLAR 5 rocket experiment). As

an example, voltages of + 4 V ( 108 km) to +28 V (122 Figure 7-1 Vehicle-to-ambient potential difference and electron density
km) were observed by Precede (electron beam voltage: -2.5 (modeled) as a function of altitude [Cohen et al., 1979],

kV; current: -O. 8A O’Neil et al., [1978a]).

An exception to the rocket beam flights just mentioned km (OGO 4); Sagalyn and Burke [ 1977], – 1.5 V to 4 V
was the “Spacecraft Charging Sounding Rocket Payload’ in the plasma trough and – 0.5 to – 1 V in the polar cap
[Cohen et al., 1979; Mizera et al., 1979]. This flight tested at 2500 km (INJUN 5); and Samir et al. [1979 a, b], –O. 1
prototypes of the positive and negative charge ejection sys- to – 1.3 between 275 to 600 km (AE-C). The highest values
tems, the transient pulse monitor, and the rocket surface observed at low aititudes were by Hanson et al. [ 1964; 1970]
potential monitor subsequently flown on the P78-2 SCA- who estimated potentials of – 6 V at 240 km to – 16 V at
THA satellite. Additionally, a thermal emissive probe, a 540 km and Knudsen and Sharp [ 1967] who recorded – 15
bipolar-intersegment voltmeter, and a retarding potential V at 516 km. Even higher values of – 40 V were observed
analyzer were also flown. The unneutralized beams repeat- by Sagalyn and Burke [ 1977] at 2500 km on INJUN 5.
edly varied the rocket ground potential between – 600 V These latter values are probably valid as they were observed
and + 100 V (Figure 7-1, Cohen et al. [ 1979]). Potentials in the aurora] zone during impulsive precipitation events
as high as + 1100 V were observed on conducting surfaces and at night. Although most of these results were for eclipse
and + 400 V on insulators [Mizera et al., 1979] relative to conditions, there was no unambiguous effect due to pho-
spacecraft ground. The variations in the vehicle to space toelectrons at these altitudes (this is not true for higher
potential correlated well with the ambient plasma density. altitudes).
Higher voltage variations, particularly in sunlight, are
seen in the plasmasphere proper at altitudes above 2500 km.
7.2.2 Satellite Measurements On OGO 5, Norman and Freeman [ 1973] found potentials
of –7 to – 10 V at 1.1 R~. Between 2 to 6 R~, as the
Satellite measurements at low earth orbit have been made satellite crossed the plasmapause, the voltage varied from
primarily by retarding potential analyzers and similar current – 5 to + 5 V. At 8 R~ the potential reached +20 V (note
collection probes. The earliest satellite observations of that in eclipse the potential fell below – 3.5 V). Abmed
spacecraft charging were by the ion trap experiment on and Sagalyn [ 1972], employing spherical electrostatic ana-
Sputnik 3 which measured potentials in the – 2 V to – 7 lyzers on OGO 1, calculated potentials of – 3 to – 6 V
V range [Krassovsky, 1959]; however, Whipple [ 1959] es- beyond the plasmapause and – 11 to – 8 V in the plas-
timated – 3.9 V. As has been discussed, the first well doc- masphere. On the same satellite, Taylor et al. [ 1965], em-
umented measurements of satellite potential at low altitudes ploying an RF spectrometer, estimated – 15 V at low al-
were by the Explorer 8 where potentials of – O. 15 V were titudes ( 1500-2700 km) to -O V at 30000 km. As discussed,
observed between 425 and 2300 km [Bourdeau et al., 1961]. however, this potential variation may have resulted from
Such low negative values are typical of this region: Reddy the interaction of the exposed positive electrodes on the
et al. [1967], –0.5 V at 640 km on TIROS 7; Samir [1973], spacecraft solar cells with the environment. Whipple et al.
– 0.71 V to – 0.91 V between 600 and 900 km (Explorer [ 1974] reported potentials between O and – 5.4 V in the
31); Goldan et al. [ 1973], – 0.7 V between 400 and 650 plasmasphere on OGO 3 in sunlight. Montgomery et al.

7-3
CHAPTER 7
[1973] observed potentials of + 100 V in the high latitude satellites inthe plasmasheet (TE - 10keV). The best doc-
magnetotail at 18.5 RE on Vela 6. (As there were apparently umented and most extensive set of such observations come
no ion measurements at the time of these estimates, there from the University of California at San Diego (UCSD)
may be some uncertainty in the method used as a result of particle experiments on the geosynchronous satellites ATS
possible differential charging [see Whipple, 1976b, or Grard 5 and ATS 6. The large potentials observed by these sat-
et al., 1977]. ) During eclipse in the same region, they es- ellites were first reported and explained by DeForest [1972;
timated the potential to be + 15 V. 1973], and provided a major impetus to the discipline of
As the satellite potential in eclipse is proportional to the spacecraft charging. A typical example of a –10000 V
electron temperature, it is not surprising that the most spec- eclipse charging event in spectrogram format for day 59,
tacular potential variations have been for geosynchronous 1976, for ATS 6ispresented in Figure 7-2a. In this type

:00” “cl

Figure 7-2a. Spectrogram of the UCSD particle detectors on ATS 6 for day 59, 1976 showing a 10-kV charging event between 2140 and 2200 UT

7-4
 CHARGING OF SPACECRAFT SURFACES

t later are the ATS 5 and ATS 6 beam experiments [Goldstein

I 00 ELEC~NS 105 PROTON S and DeForest, 1976; Olsen, 1980; and Purvis and Bartlett,
“E
-.r 1A
1, 1980]. These studies presented evidence that while electron
emission reduced satellite charging, neutral plasma emission
was necessary to achieve zero satellite to space potentials.
Limited observations are available in the solar wind and
in the vicinity of the other planets. The Vela satellites typ-
ically experienced potentials of + 3 to + 5 V in the solar
wind [Montgomery et al., 1973]. Whipple and Parker [1969a]
computed potentials of + 2.2 V and + 10 V for OGO 1 and
IMP 2, respectively, in the solar wind. The Voyager space-
craft observed potentials of + 1 V to + 10 V [Scudder et
al., 198 1]. Although preliminary, similar potentials were
apparently measured by the Voyager in Jupiter’s outer mag-
Electron and ion distribution functions versus energy for day netosphere while slightly negative potentials occurred inside
59, 1976. (Dashed lines represent the spectra before and the denser, cooler regions of Io’s plasma torus [Scudder et
after eclipse; solid lines represent eclipse spectra. ) al., 198 1; see also comments in Grard et al., 1977].
The most complete spacecraft charging measurements
are those being made currently by the P78-2 SCATHA sat-
of plot [DeForest and McIlwain, 1971] the Y axis is particle ellite (Figure 7-6). P78-2 SCATHA was launched in January
energy, the X axis is time, and the Z axis (shading) is particle of 1979 into a near-synchronous orbit (5 X 7 R~). The
count rates. The figure is interpreted as follows. When the satellite is specifically designed to study spacecraft charging
satellite entered eclipse (-2 140 UT), the photoelectron flux as is evidenced by the extensive list of scientific and en-
went to zero shifting the already (100 V) negative potential gineering instruments (Table 7- I ). P78-2 SCATHA has con-
even more negative. This negative satellite potential V, ac- firmed current ideas concerning the charging process. At
celerated the positive ions as they approached the satellite, the same time, information on the sheaths surrounding a
adding energy qV, to each particle. Zero energy positive satellite has been obtained. An abbreviated list of obser-
ions thus appear as a bright band between 2140 and 2200 vations follows:
UT at an energy (10 keV) equal to q times the potential V,. 1. Arcs were observed under different potential con-
Electrons, in contrast, were decelerated giving the dropout ditions (eclipse, sunlight, beam operations, etc.
in the electron spectra between 2140 and 2200 UT. The IKoons, 1980]).
spectra for before and after eclipse are presented in Figure 2. The satellite surface potential monitor (SSPM)
7-2b demonstrating this 10 keV shift. Potentials as high as has determined the response of a number of ma-
– 2000 V [Reasoner et al., 1976] in sunlight and – 20000 terials to both natural and artificial charging events
V in eclipse have been observed on ATS 6 [S .E. DeForest, [Mizera, 1980]. Sample potentials of over – 1000
1978, private communication]. V have been observed relative to spacecraft ground
Another effect of potential variations, differential charg- (see Figure 7-26b).
ing, is also visible in the ATS 5 and ATS 6 spectrograms. 3. Natural charging events of – 1000 V or greater
In Figure 7-3, a spectrogram from day 334, 1969, for ATS have been observed with one event in ex~ess of
5 is presented for the detector looking parallel to the satellite – 8000 V.
spin axis. Between 0500 and 1100 UT a feathered pattern These observations are the first to include simultaneous
is visible in the ions and a dropout in the electrons below data on the plasmas, magnetic and electric fields, surface
750 eV is observed. These patterns are not visible in the potentials of dielectrics and other surface materials, arcing,
detectors looking perpendicular to the satellite spin axis. and surface contaminants.
The explanation [DeForest, 1972; 1973] is that a satellite The SCATHA data set has been of particular value in
surface near the detectors has become differentially charged defining a “worst case” charging environment for the geo-
relative to the detectors resulting in a preferential focusing synchronous orbit. In Table 7-2 are listed two “worst case”
of the ion fluxes and a deficiency in the electron fluxes to examples from SCATHA as adapted from Mullen et al.
the parallel detectors. [1981] and Mullen and Gussenhoven [1982]. For compar-
The frequency of occurrence of the ATS 5 and ATS 6 ison with earlier estimates, a “worst case” example from
charging events in eclipse are presented in Figure 7-4 [Gar- ATS 6 has also been included [Deutsch, 1981]. The plasma
rett et al., 1978]. The daylight charging events on ATS 6 moments and single maxwellian temperatures [described in
have been found by Reasoner et al. [1976] (see also Johnson Garrett, 1979] are averaged over all angles while for SCA-
et al. [1978]) to be anticorrelated with encounters with the THA the 2-maxwellian values are for components parallel
plasmapause. The level of charging has, however, been ,and perpendicular to the magnetic field. The SCATHA ex-
found to correlate with Kp (Figure 7-5). To be discussed ample on Day 114 (24 April 1979, 0650 UT, 231 I MLT)

7-5
CHAPTER 7
..... ......... ....,.... ......*...... ..,.. .,. *..,*...*, .............. ......... ........ ......... ......... ..... ...
* *&L% ~-w /., . &
*g 9.x ~ ... .,.”. ”” -“>. . ..,.

*’>

m
.%.,.r.K’%=’’-,:.-..,:
-,””:. .,::’’.-:’.+.
...-.
___..-~. ---------- “ -’

1
—m
,:..’.” . . . . .. .. . . , . .. . . . . ... ..- . . . ---- ~]::.”:-” ““~:
, . . , . .,.- . . ......... ........ * _—.--’.. . .,,.,. .
. . ..
8
a t ?Lx H h ~~ ~ ea:a~t ~* .+a,&%G “
eLwm &&& .*.*%..*.,*, c 4“.,=,s ,* .Wbk, ,,..*, . ~
~~ $ @ mm%:zfxmw
-,.X. “ :*’=’*.
- ............... + -“-’ .........
......... ......... ‘“” “’”’
......... .................. ......... ........ ..........~~*f@~
......... ......... ......... ........... .............

E’:E4GY
m-
?,
,, E:

m
WO’fs
mJ-

m-
0-.

-2

-i . ..................\ .................
9 8 b
.....$&”;;
..*.* h;i&’&”;& “’”’ ~~ ‘*””’iS”””l#”””*
“’”” 1* ”’’ti’”’i#”””~”””K”””*
”””%””’%
012

[ I

Figure 7-3, Day 334, 1969, spectrogram from ATS 5 [DeForest, 1973]. The feathered pattern in the ions between 0700 and 1100 UT and the corresponding
loss of low energy electrons is real and the result of differential charging.

was particularly well documented. For this example, it was The average ratio of oxygen ions to hydrogen ions was 0.4
found that the vehicle frame potential closely followed the during the event.
electron current between 33 keV and 335 keV. It is this
current that apparently caused the observed high negative
vehicle potentials ( – 340 V in sunlight; estimated to be – 16 7.2.3 The Effects of Spacecraft Charging
kV in eclipse). Maximum surface material charging levels
for this event were: – 3.8 kV on quartz fabric, – 6.4 kV Aside from direct measurements of spacecraft charging,
on silvered Teflon, and – 1.5 kV on aluminized Kapton. indirect indicators that reflect the effects of spacecraft charg-

7-6
 -.
m
al
—
----
ATS-5
ATS-6
(1969-1972)
(1976)
cHARGING

●
OF SPACECRAFT

ATS - 5/ATS-6 ECUPSE POTENTIALS

SURFACES

‘1
g 40 ~
,- G ~ 0 2 MAXWELLIAN
E ~ -5000
0
— 0
> 0
~ 30
z -4000
w i
K
E z
~ 20 w
v ~---_-7-.7 E -3000 ●
v
2

OLA
o ‘1 [
i I

s 10

,1 .2 .4 1.0 2 4 lo---;
VOLTAGE ( keV)
—
1 1 1 1 1 1 1 1

Rgure 7-4. Occurrence frequency of ATS 5 (1969-1972) and ATS 6 00 ‘o 20 03 04 05 60 70 60 90

( 1976) eclipse potentials ( 10-min intervals), kp

Figure 7-5. Statistical occurrence frequency of observed variations of ATS

ing also exist. Typical is the plot of satellite operational
5 and ATS 6 eclipse potentials as a function of Kp (solid dots)
anomalies as a function of local time at geosynchronous and various theoretical predict ions [Garrett et al., 1979].
orbit presented in Figure 7-7 [McPherson and Schober, 1976].
The cause of this clustering near local midnight is believed
to be spacecraft charging—intense fluxes of energetic elec- defined power laws: peak current scales as the 0.50 power
trons associated with injection events are encountered near of the area, released charge as 1.00, energy dissipation as
local midnight which lead to charge buildup and arcing and 1.50, and pulse duration as 0.53. Although still preliminary,
hefice to control circuit upsets and operational anomalies. various attempts are also underway to theoretically model
Given that differential charging can take place, whether these arcing phenomena [Muelenberg, 1976; Beers et al.,
through potential diffe~nces on adjoining surfaces or through 1979].
charge deposition in dielectrics, arcing can occur. Arcing, Another result for surface arcs [Stevens, 1980; Nanevicz
defined as the rapid (-nanosecond) rearrangement of charge and Adamo, 1980] in the laboratory is that the breakdown
by punchthrough (breakdown from dielectric to substrate), potential on a negative surface varies from – 100 V at low
by flashover (propagating subsurface discharge), blowoff earth orbit to – 10000 V at geosynchronous orbit implying
(arc to surface), between surfaces, or between surfaces and that arcing should not be a common occurrence. In Figure
space, is not well understood. A typical arc discharge pulse 7-9 [Shaw et al., 1976] the arcing rate on a geosynchronous
[Balmain et al., 1977] is plotted in Figure 7-8. Balmain satellite shows a steady increase with the daily geomagnetic
[1980] finds that surface discharges display characteristics index ap. As arcing is common even at low levels of geo-
that scale with variations in specimen area according to well magnetic activity, a discrepancy exists between laboratory

P78-2

Figure 7-6, The P78-2 SCATHA satellite, The dimensions are approximately 1.3 m wide by 1.5 m high

7-7
CHAPTER 7
Table 7-1. PrinciDal
. investigators/suonsom
-. for F78-2 SCATHA

Experiment Principal Investigator/

Number Title Sponsor Address

Sc 1 Engineering Dr. H.C. Koons/ The Aerospace Corporation

Experiments USAF/AFSC/SD P.O. Box 92957
bs Angeles, CA 90009
SC2 Spacecraft Dr. J.F. Fennell/ The Aerospace Corporation
Sheath USAF/AFSC/SD P.O. BOX 92957
Electric Fields Los Angeles, CA 90009
SC3 High Energy Dr. J.B. Reagan/ Locld’teed Palo Alto Rsch Lab
Particle Office of Naval 3251 Hanover Street
Spectrometer Research Palo Alto, CA 94304
SC4 Satellite H.A. Cohen/ AFC3L/PHG
Electron and USAF/AF~C/AFGL Hanscom AFB, MA 01731
Positive Ion
Beam System
SC5 Rapid Scan Capt. D. Hardy/ AFGL/PHG
Particle USAF/AFSC/AFGL Hanscom AFB, MA 01731

SC6 Thermal R.C. Sagalyn/ AFGL/PH

Plasma USAF/AFSC/AFGL Hanscom AFB, MA 01731
Analyzer

SC7 Light Ion Dr. D.L. Reasoner/ NASA Marshall Space Flight
Mass Office of Naval Center, Code BS-23
S~ctrometer Research Huntsville, AL 35815

SC8 Energetic Ion Dr. R.G. Johnson/ Lockheed Palo Alto Rsch Lab
Composition Office of Naval 3251 Hanover Street
Experiment Research Palo Alto, CA 94304

SC9 UCSD Dr. E.C. Whipple/ University of California

Charged Office of Naval B019 Dept. of Physics
Particle Research/USAF/AFSC/ La Jolla, CA 92093
Experiment SD

Sclo Electric Field Dr. T.L. Aggsonl NASA Goddard Space Flight
Detector Office of Naval Center, Code 625
Research Greenbelt, MD 20771

Sell Magnetic Dr. B .G. Ledley/ NASA Goddard Space Flight

Field Office of Naval Center, Code 625
Monitor Research Greenbelt, MD 20771

ML12 Spacecraft Dr. D.F. HalU The Aerospace Corporation

Contamination USAF/AFSC/AFML P.O. BOX 92957
Los Angeles, CA 90009

TPM Transient Dr. J.E. Nanevicz/ SRI

Pulse USAF/AFSC/SD Menlo Park, CA
Monitor

7-8
 CHARGING OF SPACECRAFT SURFACES
Table 7-2. “Worst” case ~eosvnchronous environments. The moments. TAVG. and TRMS are averaged over all angles. The SCATHA 2-Maxwellian
parameters are-for fluxes parallel and perpendiculw to the magnetic field. ATS 6 2-Maxwefiian parameter; are averaged over all directions.

SOURCE DEUTSCH [1981] MULLEN ET AL. MULLEN AND

[1981] GUSSENHOVEN [ 1982]
DATE DAY 178, 1974 DAY 114, 1979
ATS 6 SCATHA SCATHA

ELECTRONS IONS ELECTRONS IONS ELECTRONS IONS

(ND) (cm-’) 0.112E + 01 0.245E + 00 0.900E + 00 0.230E + 01 0.300E + 01 0.300E + 01
(J) (ttA cm-’) O.41OE + 00 0.252E – 01 0.187E + 00 0.795E – 02 0.501E + 00 0.159E – 01
(ED) (eV cm-’) 0.293E + 05 O.1O4E + 05 0.960E + 04 0. 190E + 05 0.240E + 05 0.370E + 05
(EF) (eV cm-’ s-’ sr-’) 0.264E + 14 0.298E + 12 0.668E + 13 0.430E + 13 0.151E + 14 0.748E + 12

N1 (cm-’) parallel —0.882E – 02 0.200E + 00 0.160E + 01 O.1OOE + 01 O.11OE + 01

perpendicular — — 0.200E + 00 O.11OE + 01 0.800E + 00 0.900E + 00
T1 (eV) parallel —0.1 llE + 03 0.400E + 03 0.300E + 03 0.600E + 03 0.400E + 03
perpendicular — — 0.400E + 03 0.300E + 03 0.600E + 03 0.300E + 03
N2 (cm-’) parallel 0.122E + 01 0.236E + 00 0.600E + 00 0.600E + 00 0.140E + 01 0.170E + 01
perpendicular — 0.230E + 01 0.130E + 01 0.190E + 01 0.160E + 01
T2 (eV) parallel 0.160E + 05 0.295E + 05 0.240E + 05 0.260E + 05 0.251E + 05 0.247E + 05
perpendicular — — 0.248E + 05 0.282E + 05 0.261E + 05 0.256E + 05

TAVG (eV) 0.160E + 05 0.284E + 05 0.770E + 04 0.550E + 04 0.533E + 04 0.822E + 04

TRMS (eV) 0.161E + 05 0.295E + 05 0.900E + 04 0.140E + 05 0.733E + 04 0.118E + 05

and in situ measurements which underscores the need for IEEE Conference on Nuclear and Space Radiation Effects,
further analysis. 1979). Balmain [1980] and Nanevicz and Adamo [1980]
The effects of arcing are somewhat better understood have analyzed the effects of arc discharges on material sur-
than the process itself. The cument pulse (Figure 7-8) gen- faces. Balmain [ 1980] gives numerous examples of holes
crates an electrical pulse in spacecraft systems either by and channels of micron size in dielectric surfaces. Nanevicz
direct current injection or by induced cuments due to the and Adamo [1980] find additional, large scale physical dam-
associated electromagnetic wave (see Proceedings of the age to solar cells such as fracturing of the cover glass.
Of more immediate concern to the space physics com-
munit y than arcs, however, are the effects of spacecraft
LOCAL TIME DEPENDENCE OF ANOMALIES charging on plasma measurements. There are numerous ways
that charging can complicate the interpretation of low energy
13 12 II 10 9
plasma measurements. These can be loosely defined as shift-
ing of the spectra in energy, preferential focusing or exclu-

ARC CURRENT

.“~

V
❑
21

OSP LOGIC UPSETS

OSCS II RGA UPSETS
/
~

\
22 I 00 200 300
1
400
Y INTELSAT IV
2
O INTELSAT Ill T (nsec)

figure 7-7, Local time plot of satellite operations anomalies; radial dis. Figure 7-8 Oscilloscope tracing of discharge currents into a conductor
tance has no meaning [McPherson and Schober, 1977]. supporting a mylar specimen [Balmain et al., 1977].

7-9
CHAPTER 7
12 JANUARY 1974-02 FEBRUARY 1974 4
INCREASING
‘~ — -~ie

I
INCREASING ACTIVITY
60 GEOMAGNETIC _ s -
ACTIVITY i
u .
s 50 - . Al/Al
a . .
w (A1/AI) 2 -
0- N.
EC
<s40
z\ 4.: L...
Va . I
gl- .
. . SOUARSS Fll
::30 - .
~: SOUARES FIT
. .
~ .
20 -a JSo
%

10 - “a .

OL , , , , , t 1 I
Figure 7-10. Heating rate per day divided by the average heating mte for
051015202S30 35404550
the entire time ~riod as a function of the daily Dst index
[Nanevicz and Adamo. 1980].

Figure 7-9. Number of arcs per hour as a function of daily ap for a Contaminant ions, due to thrusters (ionic or chemical)
geosynchronous satellite [Shaw et al., 1976]. or outgassing of satellite materials, can be trapped within
the satellite sheath and preferentially deposited on nega-
tively charged spacecraft surfaces. Cauffman [ 1973] (see
sion of particles of a particular energy or direction, and
also Jemiola [ 1978; 1980]) has estimated that as much as
contamination of measurements by secondaries, backscat-
50 ~ of material can be deposited on charged optical sur-
tered electrons, and photoelectrons. Each of these effects is
faces in as little as 100 days. In Figure 7-10 [Nanevicz and
briefly described below.
Adamo, 1980] the heating rate of sensors on a geosynchron-
As described earlier in reference to the ATS 6 obser-
ous satellite apparently rose with geomagnetic activity. This
vations (Figure 7-2), a potential difference between the ve-
is believed to be due to increased contaminant deposition
hicle and the ambient plasma can raise or lower the energy
during periods of geomagnetic activity and, therefore, in-
of the incoming particles. In the case of differential flux
creased charging (Figure 7-5). Such deposition may also
measurements such as the ATS 6 electrostatic analyzers,
alter secondary emission and photoelectron properties.
the shift in energy is easily detected in the attracted species
Another effect related to charging observed in the lab-
(Figure 7-2). For devices that measure the total current (such
oratory is parasitic power loss. This is anticipated to be
as a Langmuir probe) more subtle techniques must be em-
important at low orbital altitudes due to interactions between
ployed which involve intimate knowledge of the current-
the ambient plasma and exposed high voltage surfaces such
voltage characteristic.
as solar cells. McCoy et al. [ 1980] and Stevens [ 1980] have
The most difficult effects to correct are due to particle
estimated this effect both theoretically and by laboratory
focusing or exclusion. As demonstrated in Figure 7-3, dif-
experiments in large vacuum chambers to result in a 10%
ferential charging of surfaces near a detector can result in
power loss for voltages in excess of 5000 V at low earth
focusing or defocusing at specific energies. Just as the ge-
orbit. They also find that arcing begins at low earth orbit
ometry of the vehicle can shadow the field of view of a
at – 100 V and is a significant prob]em for potentials of
detector, the potential gradients near the detector can distort
– 1000 V at geosynchronous orbit. Apparently, the power
the particle trajectories. Methods for estimating particle fluxes
loss and damage due to such arcing is much more serious
in the vicinity of such charged surfaces are discussed in
on high voltage arrays than parasitic power loss.
Section 7.4.
Secondary electrons, backscattered electrons, and pho-
toelectrons are emitted by spacecraft surfaces. if the vehicle
is positively charged, these low energy particles can be 7.3 CURRENT MECHANISMS
reattracted to the vehicle. They constitute an extra and un-
desirable current source that can confuse ambient electron The spacecraft charging effects reported in the previous
measurements. Even in the case of negative potentials, dif- section result from current balance, that is, all the various
ferential potentials or space charge potential minima can currents of charged particles to and from the satellite surface
result in significant secondary and backscattered electrons must balance. Such current balance is valid for a uniform,
and photoelectron currents into the detector. This current is conducting satellite surface or for a non-conducting surface
easily misinte~reted as a positive potential effect if only in the limit of a point. Unless otherwise stated, the former
electron measurements are availab]e. case (for a uniform conducting satellite) will be assumed (if

7-1o
 CHARGING OF SPACECRAFT SURFACES
a non-uniform satellite surface is considered, then the cur- s. These time scales imply at least three different ranges for
rent balance for each isolated surface and the resistive, ca- the validity of current balance. Except for the highest fre-
pacitive, and inductive currents between surfaces must also quency ambient electron variations, for plasma frequency
be considered). The calculation of the satellite surface po- variations at low altitudes, and arcing, satellite to space
tential consists of 2 steps. First, the currents to the satellite current balance can be realized. Thin dielectrics can respond
surface are determined as functions of ambient conditions, to typical environmental variations (note this includes am-
satellite geometry, and potential. Second, a satellite poten- bient variations due to satellite spin modulations, which are
tial is found so that current balance is achieved. The cal- usually a few seconds to a minute). Current balance of large
culation of the currents to the satellite surface is very de- surfaces relative to each other may, however, not be achieved
pendent on the fields in the vicinity of the space vehicle. (this behavior is illustrated in Figure 7-25b). Thus, current
These not only depend on the geometry of the vehicle but balance is expected to be valid in a number of interesting
also on the sheath or the cloud of charged particles trapped cases although care must be exercised in the vicinity of the
near the satellite. The formulas for calculating the current plasma frequency and, in the other extreme, of large isolated
sources, given the incident particle distribution at each sat- dielectric surfaces.
ellite surface, will be described in this section.

7.3.2 Incident Particle Fluxes

7.3.1 Time Scales
The major natural source of potentials of 10 kV or higher
In determining the validity of the assumption of current on satellite surfaces is the ambient space plasma. Although
balance, an important issue to consider is the time scales space plasma is seldom representable in terms of a single
for which it is applicable. Basic electrostatic considerations temperature and density, the Maxwell-Boltzmann distri-
give order of magnitude estimates of these time scales. As bution function is a useful starting point for describing the
an example, assume the satellite is a conducting sphere of ambient plasma conditions that generate these large poten-
radius r and has a capacitance of C= (Kr). The time SCale tials. Given the distribution function f for an isotropic Max-
~w, for the charging of a conducting sphere relative to space well-Boltzmann plasma
in the earth’s magnetosphere is then [Katz et al., 1977]

‘s-s
c= v
-—=2x10-3s
– 4nr2J
(7.1) ‘(vi)
= “(%)3’2exp(*)
’73)
where ni = number density of species i, mi = mass of
where r = 1 m, V is the satellite potential relative to space species i, Ti = temperature of species i, Vi = velocity of
(-1 kV), and J is the ambient flux (-O.5nA cm”’). species i, k = Boltzmann constant, and f = distribution
Unfortunately satellites normally are covered with ther- function.
mal blankets that consist of thin dielectrics deposited over The current flux to a surface in the absense of an electric
conducting substrates. The capacitance of a given dielectric field is
area A of thickness s can be estimated by CD (W A/4Trs) so
that the time scale ‘rD is [Katz et al., 1977]
JiO = qi vi . n fd3V. (7.4a)
JJJ
CD V
-—== 1.6s (7.2)
‘D – AJ Assuming isotropy,

where s = 0.1 cm, V = 1 kV, and J = 0.5nA cm-2.

Other important time scales are the charging time of
large, isolated surfaces relative to each other (from seconds
“0= (?)(%)”2 (7.4b)

to perhaps hours, depending on surface details), the plasma where JiOis the current density per unit area for O potential,
frequency ( 104–107 Hz), and the gyrofrequency ( 103–10’ n is the unit normal to surface, qi is the charge on species
Hz for electrons and 10–103 for H+); lower values are i, and d3V is the volume element in velocity space.
representative of geosynchronous orbit where n - 1 cm-3 As will be shown in Section 7.4, when the effects of
and B - IOOnT, and the higher values for the ionosphere, the spacecraft potential, sheath or plasma anisotropies, and
where n - 10’ cm-3 and B - 3 x 104nT. Typically envi- deviations of the ambient plasma from a Maxwellian dis-
ronmental changes take place in minutes or longer, although tribution are considered, the simple distribution function of
time scales on the order of the plasma frequency and gy - Equation (7.3) is no longer valid and the integration of
rofrequency are observed. Arcing durations are 10-9– 108 Equation (7.4) becomes difficult. Even so, Equation (7.4)

7-11
CHAPTER 7
or a modification of it, is accurate for many practical pur- photoelectron yield H(E) ( = W(E) S(E)) as functions of en-
poses. Approximate values of T and n for various space ergy for aluminum oxide [Grard, 1973b]. The total current
plasmas are tabulated in Table 7-3. density for zero or negative satellite potential and normal
incidence, as derived by Grard, is ,

7.3.3 Photoelectron Currents JPHO = “ W(E) S(E) dE = = H(E) dE. (7.5)

J 0 J 0

The photoelectron current from a surface is a function

of satellite material, solar flux, solar incidence angle, and JPHO for a variety of materials is tabulated in Table 7-4
satellite potential (see review by Lucas, [1973]). In Figure [Grard, 1973b].
7-11 adapted from Grard [ 1973b] are plots of the 4 functions If the satellite is positively charged, the ambient pho-
necessary to describe the photoelectron current. In Figure toelectron current is attracted to the surface. As this return
7-11 the solar flux S is plotted as a function of energy E current is a function of potential and geometry, the energy
(or wavelength). The details of the spectrum change with spectrum of the electrons for a given incident monochro-
solar activity and can vary greatly if the sunlight reaching matic photon must be known to calculate it accurately. Grard
the spacecraft is attenuated by the atmosphere [Garrett and [ 1973b] has carried out these calculations for several ma-
DeForest, 1979]. Also shown in Figure 7-11 are the electron terials and different probe geometries (see also Whipple,
yield per photon for normal incidence, W (E), and the total [1965]).

Table 7-3. Estimated plasma parameters for various environments. Most values are rough estimates, See appendix

Charac-
teristic
Energy (eV) h~(m) Potential (V)**

Sunlight Eclipse
JP~
Region Altitude NC,(cm-s) Ions 1+ E- 1+ E- V(krrr/s) (nA-cm-*) 1-D RAM 3-D 1-D RAM 3-D

Venus
200km 1OS 0+,0*+ 0.05 0.3 0.005 0.01 8 8 –1.2 –~ – 0.83 –1.8 –1.2 –0,88
1500km 102 0+ 0.2 1 0.33 0.74 8 8 6.0 Q 2.4 –5.6 –Q –2.9
Earth
150km 105D* 0+,02 +,NO+ o. I 0.2 0.007 0.01 8 2 –1.1 –~ –0.55 — — —
103N NO+ 0.05 0.1 0.05 0.07 8 2 — — — –0.58 –~ –0.37
1000km 104D 0+ 0.3 0.4 0.04 0.05 8 2 –1.3 –1.2 –1.2 — — —
l&N H+ 0.2 0.2 0.03 0.03 8 2 — —— — – 0.75 –~ – 0.52
3.5R. 10’ H+ 1 1 0.23 0.23 3.7 2 –1.6 –1.6 –1.4 –~ –5.2 –2.5
Geosyn- 5.62R~ 2 H+ 5m 2500 370 260 3 2 G 1.9 Q – 8500 – 23000 –~
chronous
High — 0.1 H+ 200 200 330 330 800 2 15. 15.
— 15. – 750 – 490 – 500
—
Latitude

e
Cold Torus 3.5R,- 50 S’,o+,o++ 0.5 0.5 0.74 0.74 44 0.08 – .75 – ,59 – .72 –2.3 –1.2 –1.6
5.5RJ 1000 2 1 0.33 0.23 69 0.08 –3.8 –~ –3.1 –4.2 –~ –3.3
Hot Torus 6.0R, - 1000 S+,o-+ 40 10 1.5 0.74 75 0.08 – 37 –~ – 33 – 39 –~ – 33
8.0R, - 100 80 20 6.6 3.3 100 0.08 – 65 –@ –60 – 78 –~ – 65
Plasma 8.0R, - 12 H+,s++ 50 50 15 15 150 0.08 –110 –110 –~ – 190 – 170 – 130
Sheet 15RJ

Outer — 0.01 H+ 1000 1000 230i3 2300 250 0.08 9.6 9.5 –~ – 3800 – 4400 – 2500
Magnetosphere

Solar Wind
0.3AU 50 H+ 40 65 6.6 8.5 500 20 4.6 4.9 4.4 – 2W –~ – 160
1.OAU 2 H+ 10 50 17 37 450 2 7.8 G 7.3 – 230 –120 –~
5.2AU 0.2 H+ 1 10 17 53 400 0.08 7.4 8.0 _
G – 50 –18 –~
*D = Day, N = Night
**see Appendix for description of computations and caPtiOns
Underlined values are “prefemed’ estimates
 CHARGING OF SPACECRAFT SURFACES

. ... . .. s
8
io22r It —H

Figure 7- 12a. Secondq electron yeild 8~ due to incident electrons of

energy E impacting on aluminum [Whipple, 1965].

electrons are emitted as a result of energy deposition by

incident electrons or ions [Stemglass, 1957; Willis and Skin-
ner, 1973, Chung and Everhart, 1974; Baragiola et al.,
1979]. Each of the three has a characteristic emission spec-
E (eV) trum.
The equation for the current density due to secondary
Figure 7-11. Composite plot of W(E), the electron yield per photon; S(E), emission, assuming an isotropic flux and ignoring other
the solar flux: and their product, H(E), the total photoelectron angular variations is (see, however, Whipple, [1965]):
yield, as function of e~ergy E for aluminum oxide [Grard,
1973].
Jsi = g,(E’ ,E) ~i(E)E f,(E) dE (7.6)
%IdE’r

7.3.4 Backscattered and Secondary Electrons where J~i = secondary electron flux due to incident species
i (usually assumed to be electrons, e-, or protons, H+),
The impact of ambient electrons and ions on a spacecraft gi = emission spectrum of secondary electrons due to in-
surface generates backscattered and secondary electrons— cident species i of energy E, ~i = secondary electron yield
backscattered and secondary ion fluxes being insignificant. due to incident species i of energy E, E’ = secondary elec-
These fluxes, though often neglected in charging calcula- tron energy, and f, = distribution function of incident par-
tions can exceed the incident fluxes under some circum- ticles at surface.
stances. Although a clear distinction between secondary and Typical curves for ~i for e- and H+ impacting on alu-
backscattered electrons is not always possible, backscattered minum from Whipple [1965] are plotted in Figure 7-12a
electrons are those ambient electrons reflected back from and 7- 12b (see also Stemglass [1957], Willis and Skinner
the surface with some loss of energy [Stemglass, 1954] (see [1973], Chung and Everhart [1974], and Baragiola et al.
also reviews by Dekker [1958], Hachenberg and Brauer [1979]). The function gi is assumed to be independent of
[1959], Gibbons [1966], and Lucas [1973]:) Secondary incident energy and incident particle species. The normal-

Table 7-4. Photoelectron emission characteristics of various spacecraft materials.

Photoelectron Saturation Flux Saturation Current

Material Work Function (eV) ( 10’2n./s-m2) Density (p A/mz)

Aluminum 3.9 260 42

Oxide
Iridium 4.8 190 30
Oxide
Gold 4.8 180 29
Stainless 4.4 120 20
Steel
Aquadog 4.6 110 18
LiF on Au 4.4 90 15
Vitreous 4.8 80 13
Carbon
Graphite 4.7 25 4

7-13
CHAPTER 7
10.0~ 0.5
I I I I I I I I 1

0.4

# 0.3
G
al A HillMd
V Gho8h aIKI Kh=d
h
● hwmd G
‘g- 0.2
m
[
I
0!01
103
1 k 1 1 11111
104
1 11111[ t 1 I 1 1{111
1 1

10s Id
1 1 I Illu
I 7
0.1
PRIMARY ENERGY (eV)

Figure 7- 12b. Secondary electron yield b, due to incident ions of energy

E impacting on aluminum IWhipple, 1965]. I I I I 1 I I I 1 1 Y
0.5 1.0
E~
ized curve [Whipple, 1965 ] for aluminum is plotted in Fig-
ure 7-12c. Figure 7-13, Graph of G (E’/E) as a function of (E’/E) where G is ap-
proximately the percentage of electrons scattered at a given
For backscattered electrons, the current density is given energy E’ as a result of an incident particle of energy E
by DeForest [1972] [Stemglass, 1954].

tual values are dependent on angle of incidence which has

J~~~ = B(E’,E) E f~(’) dE, (7.7)
~ld” k: been ignored in the preceding discussion. Further, the sec-
ondary and backscatter properties of actual satellite surfaces
where which are invariably oxidized or contaminated are not well
known. Currently the lack of knowledge in this area is one
E’ 1 of the major deficiencies in spacecraft charging theory.
B(E’, E)=G
()
~ ~

and G is the percentage of electrons scattered at fraction E’/ 7.3.5 Magnetic Field -Induced Current
E of the incident energy E. Distortions
Stemglass [1954] has published experimental measure-
ments of backscatter parameters for different materials. An A problem in low earth orbit often encountered by elec-
estimate of G for A 1 from his data is plotted in Figure 7- tric field experiments employ long (-10 m or longer) an-
13. For negative potentials, the backscatter flux is roughly tennas or booms is the induced electric field due to the
20% of the incident flux. satellite crossing the earth’s magnetic field lines. Briefly, a
For both secondary and backscattered currents, the ac- satellite moving relative to a plasma (assumed to have a
zero electric field with velocity v< will see an electric field
E in its rest frame given by
gm
In

E-
u
80 E=—
V,XB
c
= 10-’ (v. x B) V/cm (7.8)

:60 where E is in V/cm, v, is in cm/s, B is in G, and C is the

speed of light.
2 An earth-orbiting satellite will see a maximum induced
1- v X B field at low altitudes on the order of -0.3 V/m. The
?
local fields and current flows in the vicinity of the vehicle
0
024 16 18 2 will be distorted by this effect.
SECO6mRY tLEcT’;ON EERGY $v)
Besides the v X B current, the magnetic field also in-
bigure 7- 12c. Emission spectrum of secondary electrons due to incident
dllces anisotropies in the particle fluxes. Ambient fluxes,
electrons or ions impacting on aluminum [Whipple, 1965]. secondaries, beam fluxes, and charged particle wakes are

7-14
 .
a
CHARGING OF SPACECRAFT SURFACES
all controlled to a greater or lesser extefit by the magnetic
i j(8)/jo

w
j (0)/j.
field. Whipple [1965] and Parker and Murphy [1967 ]have
analyzed some of the effects of these magnetic field-induced
anisotropies on spacecraft charging (see also reviews by
Brundin [1963] and Guverich et al. [1970]) and find that
T[F- .
b
0.1
h= 530-640
nH+=0.09

. 0.1
I

.
h= 520-720
nH+. o.23
the electron flux can be reduced by as much as a factor of T .
2, but as a rule these are ignored in spacecraft charging
calculations. McCoy et al. [1980] has suggested that this
constraint of electrons to field lines may become of real
A
0 60 120 100 240300 360”
(0) 8
0 60 120 160 240 300 360”
(b) 8
concern for very large structures which are much larger than
a gyroradius.

7.3.6 Motion-Induced Effects

...
In low earth orbit, the velocity of a satellite is 7.5 km/s. o~60 120 160 240 300 -“
o~60 120 100 240 300 360”
Within the plasmasphere, where particle energies are 1 eV
or less, the ion thermal velocity is 10 km/s or less. This
j(8)/jo
implies that a plasma wake would be formed around the

X
vehicle. Chang et al. [1979] have estimated that motion-
induced effects on the satellite potential due to such a wake
h*llT5-lze5
are important for ion ram to thermal velocity ratios as low nH+. o.94
as 0.1. These motion-induced effects would be present eveo 0.1
for neutral particles and result in there being large asym- 0 60 120 I 240 300 360”
(e) o
metries in the flow between the leading surfaces of the
satellite and the rear for even small objects such as booms.
Figure 7-14 Normalized electron current versus angular position of the
The motion of the satellite can also induce space charge
plasma probe on Explorer 31 [Samir and Wrenn, 1969];
variatiotis in the vicinity of the vehicle [Parker, 1977]. The 180° would correspond to the center of the wake (that is,
depletiot-i of electrons and ions in the satellite wake can opposite the direction of movement). Also shown are the-
distoi-t the flow of the particles to the vehicle. A typical oretical fits. The altitude h and ambient density n are indi-
cated [Gurevich et al., 1970].
observation is presented in Figure 7-14 in terms of the nor-
malized electron fluxes in the wake of the Explorer 31 sat-
ellite [Samir and Wrenn, 1969]. Although simple models
of this phenomena will be discussed below, reviews by 3
Brundin [1963], Kasha [1969], Liu [1969], and Al’Pert
[1976] and papers by Gurevich et al. [1970] and Kunemann
[1978] should be consulted for details. 2

7.3.7 Charge Deposition ‘Q I

by Energetic Particles .
7
E
v
The deposition of charge in dielectrics by high energy :0
particles is a well known phenomena in nuclear physics (see, w 400 seconds

for example, Gross and Nablo [1967]; Evdokimov and Tub-

alov [ 1974]; and Fredrickson [1979]) and has been proposed
-1
as a source of satellite charging [Meulenberg, 1976]. As an 1-
example IFredetickson, 1980], eiectrons between 10 keV and
l@ keV lose energy at a rate of R = 106to 5 X 10 eVg-1cm2
depending on energy and the material involved (R o p, where
p is the density of the material, gives the energy loss rate per
cm). Electrons of 1 MeV will typically penetrate several mil-
limeters into a solid. Fields of the order of M V/mm are nec-
-2= 4
OEPTH ( mm )

essary to retard such incident fluxes. Fields of this magnitude

Figure 7-15. The electric field in 6 mm PVC sample as a function of
are more than adequate to cause electrical breakdown in a die- exposure time to a 4.66 x 10-’() A/cm*, I MeV electron
lectric. Figure 7-15 is a theoretical plot of the potential in a PCV beam [Fredrickson, 1980].

7-15
CHAPTER 7
sample exposed to 1-MeV electrons as a function of time to a much larger beam spread marked by an increase in the
[Fredrickson, 1980]. local plasma density. The critical current ~ when this occurs
As satellites in the earth’s radiation belts or the hostile is proportional to the beam energy, V~, such that Ic~V~3’2,
environment of Jupiter can experience high dosages of en- an equation resembling Equation (7. 14). The NASA Lewis
ergetic particles, this effect becomes important iflongmis- studies are part of a coordinated effort to validate the NAS -
sion lifetimes are desired. Recent evidence [Treadway et CAP code (see Section 7.4.6) and have included testing of
al., 1979] indicates that the effects of charging and arcing the P78-2 SCATHA experiments [Stevens et al., 1980a].
may be substantially altered by the buildup of charge in
dielectrics. As yet, this process has not been included in
spacecraft charge modeling. Itsinclusion, however, will be
7.4 SPACECRAFT CHARGING THEORY
critical to a complete understanding of the charging/arcing
process.
The basic equation expressing cumnt balance for a given
surface in an equilibrium situation is, in terms of the cument:

7.3.8 Artificial Charging Mechanisms 1~ (V) – [ll(V) + Is~ (V) + Is,(V) + I~s~ (V)
+ I~H (V) + 1~ (V)] = IT (7.9)
Artificial mechanisms that effect spacecraft charging are
numerous and include electron and ion beams and exposed,
where V = satellite potential, 1~ = incident electron cur-
high potential surfaces such as the junctions between solar cells
rent on satellite surface, 11 = incident ion current on sat-
[Stevens, 1980]. Recently, beam sources have been actively
ellite surface, 1~~ = secondary electron current due to 1~,
exploited both as probes of the satellite sheath and as a means
IsI = secondary electron current due to 1,,
of controlling the spacecraft potential [Goldstein and De-
I~s~ = backscatted electrons due to 1~, 1~” = photoelectron
Forest, 1976; Olsen, 1980; Purvis and Bartlett, 1980; Cohen
current, 1~ = active cument sources such as charged particle
et al., 1979]. Voltages in the KV range and currents between
beams or ion thrusters, and IT = total cument to satellite
mA and A are typical of these systems [Winkler, 1980]. The
(at equilibrium, IT = O).
Excede 2 test [0’Neil et al., 1978b], in fact, ejected 3 kV
In this section methods of solving Equation (7.9) for V
electrons at 10 A. An example of a theoretical beam calcula-
so that IT = O will be described. The basic problem is the
tion will be given in Section 7.4 as an illustration of the com-
solution of Equation (7.9) subject to the constraints of Pois-
plexities involved in analyzing such experiments.
son’s equation:
Parker [1979, 1980], Stevens [1980], McCoy et al. [1980],
and Rieff et al. [ 1980] have carried out calculations of the
V*V = 4 m q(n, + n~ – n,) (7.10)
currents to exposed potential surfaces for large, high voltage
structures. They find that a major effect is to iriduce large
and the time-independent collisionless Boltzmann (or Vla-
voltage gradients in the satellite sheath that must be con-
sov) equation:
sidered in the proper design and use of such systems. The
potentials can lead to multipacting and preferential depo-
sition of ion contaminants. Another possible difficulty as- V“Vfl–9V V(r)” Vvfi =0, (7.11
sociated with very high positive potentials is the “pinhole” ml
effect. As discussed in Kennerud [1974], the insulation on
such surfaces can become punctured. Even microscopic holes where n~ = local electron density, nl = local ion density
can result in exceptionally high focusing of electrons—in n, = surface-emitted electron density, V, and V, = gradient
some cases the pinhole can completely defeat the insulation. operators with respect to position and velocity space re-
Finally, a number of experiments have been conducted spectively.
in vacuum chambers. Currently, the most extensive pro-
grams are those at NASA Johnson [McCoy et al., 1980]
and NASA Lewis [Stevens, 1980; Purvis et al., 1977; and 7.4.1 Analytic Probe Theory
their colleagues]. The NASA Johnson chamber tests have General Considerations
involved testing of high potential surfaces and of rocket
beam sources prior to flight [Konradi, private communi- The most important concept in probe theory is that of
cation, 1980] in an environment resembling the ionosphere. the Debye length AD, the distance over which a probe or
As discussed in Bernstein et al. [1978; 1979], these latter satellite significantly perturbs the ambient medium. (Note
tests have concentrated on the analysis of the so-called “beam- that this is only one definition of the satellite sheath. The
plasma discharge” (BPD) in electron beam experiments. sheath thickness is not only dependent on satellite potential
They have found that the electron beam at a critical current and charge, but through the so-called “presheath”, influ-
transitions from one well defined by single particle dynamics ences the plasma up to the order of the satellite radius

7-16
 CHARGING OF SPACECRAFT SURFACES
[Parker, 1980]. More precisely, electrons and ions form a In order to solve these equations, it is assumed that V and
cloud around any charged surface that is a function of the dV(Y)/dY are O at some distance Y = S which determines
particles’ energies and densities and the potential on the the sheath thickness (called the “space-charge limited’ as-
surface. As an illustration, the solution to Poisson’s equation sumption). The solution becomes
for spherical symmetry is II*vo312

e— r/AD
V(r) = q ~ (7.12)
J.+%
() _
S*
(7.14)

This is the “space-charge limited diode model” solution or

where AD = (KT/4 m q2u) 1’2 and no = ambient charge Child-Langmuir model for a plane. If J is replaced by JO = K*
qn. (kT/m)”* (where (2 T)- 1/2< K* < 1, the lower limit
density. This is the classical Debye length for the assumption
of “linearized space charge.” The “linearized space charge” corresponds to Equation (7.4b); the upper limit is for a
approximation can be extended to spherical and other multi- monoenergetic flow of energy E = 1/2 kT) then the sheath
dimensional geometries; an analytical fit for the spherical thickness can be estimated:
case is discussed by Parker [1980]. Values for AD are listed

s‘i(&)’’4(v0)3’4
(:)’’2(+)”2
in Table 7-3.
The terminology “thick’ sheath and “thin” sheath de-
rives from the assumption that the region over which the
‘715)
satellite affects the ambient plasma or is screened from the This sheath thickness determines the region over which charge
ambient plasma is either larger or smaller than the char- is collected and is important in determining the maximum
acteristic dimensions of the spacecraft. Usually this reduces current that can flow to a probe for a given V.
to determining whether AD is long (thick sheath) or short Although seldom utilized in spacecraft charging calcu-
(thin sheath) compared to the spacecraft radius. From Table lations, a form of Equation (7. 14) was employed by Beard
7-3, for a 1 m radius spacecraft with no exposed potential and Johnson [196 1] to obtain the potential to which a vehicle
surfaces in low earth orbit, the term thin sheath is appro- can be charged by an electron emitter in the ionosphere.
priate. For geosynchronous orbit, unless the satellite is 10 This calculation placed an upper limit on the current that
m or larger, the thick sheath approximation is appropriate. could be drawn at a given potential ignoring magnetic field
If active surfaces (that is, exposed potentials driven by the shielding effects. When included [Parker and Murphy, 1967],
satellite systems) or a substantial photoelectron population the maximum current was reduced by as much as a factor
are present, these limiting criteria will change, but even so of 10. Linson [1969], including turbulence, found values
for most satellite studies they are very useful (Opik [19651 between these limits, however. These studies imply that a
has carefully considered this issue of the appropriate screen- 1 m sphere emitting a 0.5A electron beam at shuttle altitudes
ing criteria for the general case of an arbitrary central force). would have a potential of 104– 106V, depending on the mag-
As a first example, consider a large structure such that netic field [Liemohn, 1977], thereby seriously inhibiting a
the characteristic scale of the sheath is significantly smaller 10–100 keV beam. Recently, Parker [1980] (see also Ken-
than r,, the radius of the surface (the thin sheath assump- nerud [1974], McCoy et al. [1980], and Parks and Katz
tion). Assuming at the surface (X = rs) the potential is V. [1981]), in order to estimate the sheaths around large space
and that the surface is nearly planar relative to the sheath structures, have developed an analytic expansion for the
dimensions, at distance Y( = X – r,) from the surface (only thin sheath approximation to a sphere. He has investigated
the attracted species is considered) the applicability of this estimate of the sheath to extremely
high applied potentials in the ionosphere by comparing it
to the results of a self-consistent numerical calculation (see
Poisson’s Equation d2V(Y) Section 7.4.2). Although his rigorous computations (for a
— = –4 IT q n(Y) (7.13a)
dY2 Debye length to satellite radius ratio of 1 to 100) deviate at
low potential values (qV/kT = 10) from the space charge
Current continuity limited model, the results approach each other for very high
J = q n(Y) v(Y) = constant (7. 13b) potentials (qV/kT -400 000).
The preceding theory concerning the thin sheath assumes
The particles are assumed to have E = O in the ambient that sheath effects dominate the current flow to the satellite.
space: This places emphasis on Poisson’s equation and the space
charge around the satellite. In the opposite extreme, the
sheath and space charge are ignored so that, to first order,
Energy conservation Laplace’s equation holds. In practical terms this translates
into the assumption that AD> r,. For spherical symmetry,
~ mv(Y)2 + q V(Y) = O. (7.13C)
conservation of energy and angular momentum imply that

7-17
CHAPTER 7
for an attracted particle approaching the satellite from in- and for the repelled species Ji = Jio e ‘is where
finity

Qi, = ~~ ( + for electrons, – for ions) (7.21)

~mv~=~m v(r,)2+q V(r,) (7. 16a) kTi

m RIVO= m rsv (r,) (7.16b)

The current-voltage characteristics implied by Equations
where VOis the velocity in ambient medium, R is the impact (7.20) and (7.2 1) for the three geometries are plotted in
parameter, and RI is the R for a grazing trajectory for a Figure 7-16. Several important conclusions can be drawn
vehicle of radius r,. from Equations (7.20) and (7.2 1) and Figure 7-16. First
Solving for the impact parameter RI (only particles hav- Prokopenko and Laframboise’s results for a sphere are iden-
ing R < RI will reach r,): tical to Equation (7. 18) (with 1/2 mv~ replaced by kTi) for

“=’(1-%) (7.17) current

A
sphere
( infinite
(R1 – r,) is equivalent to the sheath thickness S defined for sheath )

a thin sheath as it also is the size of the region from which

particles can be drawn. cylinder
(infinite sheath )
The total current density striking the satellite surface for
a monoenergetic beam is infinite plane

Maxwellian
J(V)=&
s ‘J”(l -%) ‘718)

where I is the total current to spherical satellite equal, to the repulslve voltoge o attractive voltage
1
ambient current that would pass through an area 4 m R: and depends on independent (opproximotely) of
I
velocity distribution velocity distribution
JOis the ambient current density outside the sheath equal to I
I/4 mR~. This is the so-called “thick sheath, orbit-limited” independent of geometry : depends on geometry
current relation.
The thick sheath results are readily extended to more com- Figure 7-16. Qualitative hchavior of Equations (7.20) and (7.21) for a
plex distributions. For most spacecraft charging problems, Langmuir probe [Parker and Whipple, 1%7]. (Reprinted
Boltzmann’s equation [Equation (7.11)] in the absenm of with permission of Academic press O 1967)

collisions reduces to Liouville’s theorem. For a complex

distribution function, F, Liouville’s theorem states that: a thick sheath. It should also be readily apparent that the
planar solution is conceptually equivalent to a thin sheath.
Thus, Equation (7.20) additionally gives a qualitative pic-
F (V) = F’ (V’) (7.19)
ture of how the probe characteristics change as the ratio of
the Debye length to satellite radius is varied from small
where F’ and v‘ are the distribution function and velocity
values (thin sheath or planar) to large values (thick sheath
at the surface of the spacecraft. F and v are the ambient
or spherical). fiis was done explicitly by Whipple et al.
distribution and velocity at the end of the particle trajectory
[1974] and Whipple [ 1977] for the ratio of the Debye length
connected to the satellite surface where F’ and v’ are measu-
to satellite radius and is closely related to the parametrization
red. Prokopenko and Lafromboise [1977, 1980] have de-
method of Cauffman and Maynard [1974].
rived (based on Equation (7. 19)). in more general terms the
current density to a sphere, infinite cylinder, and infinite
plane (that is, three-, two-, and one-dimensions) for a Max-
well-B oltzmann distribution for the orbit-limited solution. 7.4.2. Analytic Probe Theory
Their results (see original derivation in Mott-Smith and Thick Sheath Models
Langmuir [1926]) for the attracted species are
At this point a complete analytic theory for the case of
(1 + Qi,) Sphere a thick sheath, spherical probe has been developed. Sub-
Ji = JiO “ [2(Qi,/~)”2 + eQ]serfc (Qi,’’2)] Cylinder stituting into Equation (7. 10) and assuming that the sec-
(1) Plane ondary and backscatter terms can be parametrized, for an
[
(7.20) ambient Maxwellian plasma

7-18
 w
CHARGING OF SPACECRAFT SURFACES
A, o J.. . [1 – SE (V, T~,n~) – BSE (v,TEn,)]

“ eXp(qv/kT,) (8)

~ qv (7.22)
– A, . JIO o [1 + SI(V,TI,nI)l -
() kT1
(D)
– A,H . JPHO . f(X~) = IT = O V<o ■ ■
8
c1
where ●

● 00

●
J~o = ambient electron current density
[Equation (7.4)] ! ATS-5
TO ( AVG
o
TO( RMS )
●

JIO = ambient ion current density [Equation ATS-6 o m

(7.4)]
1- nl I I I I I I
A~ = electron collection area (4 T & for a
“O -2 -4 -6 -8 -lo -12
sphere)
SATELLITE POTENTIAL (kV) IN ECLIPSE
AI = ion collection area (4 T < for a sphere)
Figure 7-17 Observed temperature of ambient electrons versus satellite
APH = photoelectron emission area (m r: for a potential in eclipse. T,,(AVG) is two thirds of energy density/
sphere) number density; T,,(RMS) is one half of energy flux/number
flux.
SE, S1, BSE = parametrization functions for secondary
emission due to electrons and ions
and backscatter threshold temperature below which charging does not occur
is real and is due to the fact that at an intermediate energy
JpHo = saturation photoelectron flux (Table 7-4)
(usually a few hundred eV), the secondary yield is greater
f(X~) = percent of attenuated solar flux as a than 1 [Rubin et al., 1978]. The electron temperature must
function of altitude X~ of center of sun be several times greater than this threshold energy before
above the surface of the earth as seen charge buildup occurs.
by satellite. The agreement can be significantly improved for Figure
7-17 if the actual ambient spectra are utilized in the inte-
Equation (7.22) is appropriate for a small (<10 m), uni- gration of Equations (7.4), (7.6), and (7.7) instead of a
formly conducting satellite at geosynchronous orbit in the Maxwellian in computing the currents. This method of using
absence of magnetic field effects. To solve the equation, V the actual particle spectra to estimate the currents has been
is varied until IT = O. A number of examples of this pro- extensively employed by DeForest [1972], Knott [1972],
cedure, assuming S1 = SE = BSE = O, are tabulated in Garrett and DeForest [1979], and Prokopenko and Lafram-
Table 7-3 for various plasma regions. boise [1977, 1980] in the calculation of satellite to space
If it is assumed that S1, SE, and BSE are constants (that potentials for geosynchronous spacecraft. The potentials of
is, -3., -0.4, and -0.2 for Al), then Equation (7.2) Figure 7-17 are recalculated in this manner and plotted in
predicts [Garrett and Rubin, 1978] that in eclipse, the po- Figure 7-18a [Garrett and DeForest, 1979]. Results are pre-
tential between the satellite and space is sented both for particle spectra immediately before entry or
immediately after exit from eclipse (“sunlit”) and for eclipse
v = – T~ (7.23) (“eclipsed”); differences are attributed to the digitization of
the spectra. As the exact secondary response of the satellite
The satellite to space potentials observed by UCSD elec- surface was not known, SE, S1, and BSE were assumed to
trostatic detectors for twenty-one ATS 5 and four ATS 6 follow the A 1 curves of Figures 7-12 and 7-13. Their ab-
eclipses are plotted in Figure 7-17 versus the electron tem- solute amplitudes were then varied until the observations
perature. As the geosynchronous plasma is not necessarily were fit in a least squares sense [Garrett and DeForest,
Maxwellian, two different “temperatures,” To (AVG) 1979].
(= 2/3 o energy density/number density) and To (RMS) Given the validity of the calibration method, the effects
(= 1/2 oenergy flux/number flux), are presented (these would of a varying photoelectron flux on the satellite to space
be equal if the plasma was actually Maxwellian). The agree- potential can be studied during eclipse passage. If the sat-
ment is good considering that the range of potentials is ellite position is known, the photon flux reaching the satellite
between – 300 V and – 10000 V. The existence of a can be calculated from first principles [Garrett and Forbes,

7-19
CHAPTER 7
(m) 198 1]. If the satellite surface materials were known, then
the photoelectron current, JpHo “ f(X~), could also be cal-
culated from first principles. Although the exact surface
[0) response is in actuality not known, adequate approximations
can be derived [Garrett and DeForest, ‘1979; Garrett and
ATS-5 ATS-6
SUNLIT o 0
Forbes, 198 1]. JpHo in Equation (7.22) is then varied to fit
ECLIPSED ● m observations. Estimates of the varying potential on ATS 6
●
during eclipse entry and exit by this technique are compared
with actual observations in Figure 7- 18b [Garrett and
0 DeForest, 1979]. This eclipse model has proven valuable
❑
in estimating photoelectron flux and potential variations for
ATS 5, ATS 6, Injun 5, and P78-2 SCATHA.
The results of Prokopenko and Laframboise [1977, 1980],
using the spectra suggested by Knott [1972], are particularly
1 I J I I
important because they established the existence of muiltiple
o -2 -4 -6 -8 -lo -12 -14
OBSERVED POTENTIAL (kV) roots for Equation (7.22) (see also Sandem and Inouye [ 1979]).
Multiple roots imply that a satellite can undergo rapid volt-
Figure 7-t8a. Predicted andobsewed ~tentials inecllpse for ATS5 and age variations in response to small environmental pertur-
ATS 6. Solid symbols are for calculations using the spectra bations and that adjacent surfaces can come to radically
in eclipse, Open symbols are for calculations using the different potentials for the same conditions. To obtain their
spectra measured in sunlight
result, they solved an equation equivalent to Equati6n (7. 22)
(that is, local current balance) for the spherical, the cylin-
drical, and planar assumptions of Equations (7.20) and (7.21)
and for eclipse conditions. They found that the potential is
markedly increased for a planar probe relative to a spherical
probe.

AE-6 7.4.3 Analytical Probe Theory

Thin Sheath and Related Models

> Analytic probe theory can also be utilized to estimate

;-103 satellite to space potentials in the 250–700 km range. As
a, discussed in Brtmdin [1963], in the absence of magnetic
i= forces, photoelectrons, and secondaries, Equation (7.9) re-
z
P duces to
o
a

& AE . J~o o C~ . eqv/kTE– AI “ JIR “ C, = O (V < O) (7.24)

i
-1

g -102 where
m

JEO = ambient electron current density (Equation (7.4)),

JIR = ion ram current density (ignoring ion thermal

o OBSERVED
velocity)
s PREDICTED

= q[ nl v,,
1 I I 1

1319 1320 1321 1322

v, = satellite velocity,
1284 1285 1286
UT(min. ) A~ = electron collection area (4 r r,z for a sphere),

Figure 7-18b. Observed and predicted potentials for the entry into eclipse Al = ion collection area (IT r,2 for a sphere),
of ATS 6 on day 66, 1966, and for eclipse exit on day 59,
1976. r, = satellite radius

7-20
 m

CHARGING OF SPACECRAFT SURFACES

!6
electron shielding factor ( = 1, no electron wake
assumed; =1/2, wake
on rear halo, 12

ion shielding factor {(=l, complete shielding of

ambient ions or thin sheath; e

no shielding or thick sheath,

4
=(1 - 3)9
(Equation (7. 18)).
0

Samir [1973] and Samir et al. [ 1979b], assuming no .180 .!35 -90 -45 0 45 90 135 I 80

ANGLE WITH vELOCITY vECTOR (DEGREES I

electron wake and no ion shielding (note that this is a thick
sheath assumption which is usually inappropriate in this
Figure 7-19 Variations in the positive ion current density with angle
region), have compared the predictions of Equation (7.24)
relative to the satellite veloclty vector. The data and figure
with observations. In spite of the simplicity of Equation are from Bourdeau et al. [1961]. The fitted line is given by
(7.24) and the assumption of a thick sheath, they predict Equation (7.29) with a assumed to he 3.6 km/s; the satellite
~tential is assumed to be O.
the satellite to space eclipse potential (typically - – 0.75V)
for a large range of ambient conditions to a factor of 2.5
or better. r, > AD, the Debye length in the ionosphere being much
A simple approximation for estimating the current to an smaller than the satellite’s characteristic dimensions. This
isolated point on a planar satellite surface in the ionosphere allows the neglect of the effects of the satellite potential
as a function of the surface normal relative to the velocity
except very close to the satellite surface. The left hand side
vector is that of Whipple [1959] and Bourdeau et al. [1961]
of Equation (7. 10) can then be ignored, so that it becomes:
(see also Tsien [1946] and Chang et al. [1979]):
n~—nl=O (7. lo*)
Ii = ~qniAi V, COS e ( 1/z
[ —x 2 This is the so-called “quasi-neutrality” assumption [for ex-
+ 1/2 erf (x)) + K
2G 1 (7.25) ample, Gurevich et al., 1966, 1968, 1973; Grebowski and
Fischer, 1975; Gurevich and Dimant, 1975; Gurevich and
Pitaveskii, 1975]. Equation (7.11) for the ions can be re-
where duced by making use of the “hypersonic” character of the
motion of a body in the ionosphere (that is, M, the square
v, Cos e qv “2
x=
a (-) kTi ‘
f3 = angle between sensor normal and velocity vector,
root of the ratio of the ion kinetic energy to electron thermal
energy, is much greater than 1). First, this implies that the
gradient of the potential perpendicular to the flow direction
Ai = collection area, is much greater than the gradient along the flow direction
a = grid transparency function, so that this latter term can be ignored. Second, the thermal
a = most probable ion thermal velocity. velocity of the ions in the direction along the flow direction
can be neglected relative to the satellite velocity v, in the
This equation, the so-called “planar approximation”, is good ionosphere. Based on these assumptions, Equation (7.11)
for short Debye lengths but becomes inaccurate for long for ions can be rewritten as
Debye lengths [Parker and Whipple, 1970] and more com-
plicated orbital trajectory calculations must be carried out. af, af, q dv aft
V, —+vl” ——— — “— = o (7.11*)
Even so, Bourdeau et al. [1961] found it to be a good az drl m, drl avl
approximation to the ion current relative to the velocity
vector measured by the ion planar probe on Explorer 8 where rl and V1 correspond to components normal to the
(425-2300 km altitude). Their results for V = O are plotted direction of motion and z is in the v, direction.
in Figure 7-19 (note that there is some question as to the For the electrons, a Boltzmann distribution
best value for a, see Bourdeau et al. [1961]). (nE = nEo o exp (qV/KTE)) is usually assumed. The re-
A simple analytic theory closely related to thin sheath sulting system of equations for the ions does not contain
probe methodology and capable, within limits, of explaining the ion thermal velocity along the direction of motion and
the satellite wake structure in the ionosphere has also been can be put into a dimensionless, self-similar form [A1’pert
developed. It (see reviews in Gurevich et al. [1970] and et al., 1965; Gurevich et al., 1970] that resembles classical
Al’pert [1976]) is based on the thin sheath assumption, hypersonic aerodynamic equations. Depending on the char-

7-21
CHAPTER 7
acteristics of the assumed plasma conditions, the wake vari- that dielectric surfaces can charge in tens of minutes but
ations for a number of simple geometries can be analytically may take hours to discharge. Further, charge up can be a
solved. These range from the extreme assumption of a true long term (days) process and be dependent on the time
neutral flow as reviewed in Gurevich et al. [1970] and Al’pert history.
[1976]-charged particle variations in the wake mirroring
the neutral variation~to plasma flows around infinite half-
planes, wedges, plates, cylinders, and discs. The predictions
for one such analytic solution [Gurevich et al., 1970] are 7.4.4 General Probe Theory
compared with Explorer 31 observations in Figure 7-14 from
Samir and Wrenn [1969] for various distances r/r, from the Whereas analytic probe theory is applicable to a number
satellite. The theory is not considered reliable at angles of practical problems, it does not allow for the complex
greater than -120° as the electron density can differ greatly geometric and space charge effects of the satellite sheath
from the ion density in this region of maximum rarefaction on particle trajectories. As a consequence, it is severely
[Gurevich et al., 1970]. limited in its quantitative accuracy (the parametrization method
Although severe constraints (primarily r, > AD and ne- of Cauffman [1973]; Cauffman and Maynard [1974], is one
glect of the ion thermal velocity) have been placed on the qualitative means of studying these effects using analytic
realm of applicability of this “hypersonic, ” quasi-neutral probe theory). In order to include the effects of the sheath
theory, it does allow the analytic study of the effects of on particle trajectories, it is necessary to seek simultaneous,
geometry, magnetic field, and, of more importance, ionic self-consistent solutions of Equations (7. 10) and (7.11).
composition on the wake. As discussed in Gurevich et al. Typically, this is not analytically possible and an iterative
[1970, 1973] and Samir et al. [1980], variations in ion procedure must be employed. First, V (r) is assumed so
composition play a critical role in the details of the expan- that fi (r,v) can be computed subject to @uation (7. 10).
sion of the ion population into the rarefied wake region The number densities, ni, are then found from
behind the satellite. Although these results are useful, in
most practical situations a finite Debye length is a necessary Ili (r) = fi (r,v) d3V (7.26)
assumption. This greatly complicates any theoretical com- JJJ
putation and requires the advanced probe theory of Section
7.4.4. Given the ni (r), V (r) is computed according to Equation
An application of analytic probe theory based on local (7. 10). The process is iterated until a consistent set of value
current balance is in the so-called circuit models. The single of ni(r), V(r), and fi (r,v) at grid points surrounding the
probe theory introduced so far does not explicitly consider surface are found. Then the Ji’s at the surface r, are found
the problem that satellites consist of a variety of surfaces, from the generalized form of Equation (7.4)
including dielectrics, and that each surface can charge to a
different potential if isolated from the others. In order to
Ji (r,) = qi v “ n fi(r,,v) d3V. (7.27)
explicitly model this differential charging effect, the cou- JJJ
pling currents between surfaces must be included in Equa-
tion (7.9). Circuit models [Robinson and Holman, 1977; The theories to be discussed assume the time-indepen-
Inouye, 1976; Massaro et al., 1977], as this class of models dent form of the Vlasov equation. Equation (7. 10) is then
is termed, consist of many “probes ,“ each representing a just a restatement of Liouville’s theorem, namely, that
particular point or surface on the satellite (a dielectric sur- fi (r,v) is constant along a particle trajectory in a potential
face, for example, would be approximated by one or more V(r). To determine fi (rp,v) at a point rp for a particle of a
individual points). Besides the ambient, secondary, back- given energy, all that is required is to find the intersection
scattered, and photoelectron currents considered in the sin- of that particle’s trajectory with a surface where fi is known.
gle probe model, the coupling currents to each point, J~~c, The trajectory may either be traced from the point backwards
are included in J to estimate the currents between surfaces. to the surface (the inside-out procedure) or from a surface
Time variations are explicitly handled by including induc- to the point (outside-in procedure). According to Parker
tive and capacitive elements which have finite charging times [1976a], the inside-out trajectory tracing method is pref-
(local current balance with space is assumed at each instant erable in that the points at which the density is calculated
in time) making these models applicable to a wider range can be picked at random and is suitable for both electrons
of problems than the single probe models. Inouye [1976] and ions. It has the disadvantage that the trajectory infor-
and Massaro et al. [ 1977] have utilized such models to assess mation computed for a given point is lost in moving to the
the effects of geomagnetic storm variation, varying solar next, increasing the computer time. The outside-in method
angle, isotropic fluxes, etc., on individual satellite surfaces can be readily adapted to time-dependent simulations but
as a function of time. Their results, although subject to the introduces difficulties in choosing trajectories so that the
difficulties associated with simple probe analysis, indicate density at an arbitrary point can be determined. As a result

7-22
 a
CHARGING OF SPACECRAFT SURFACES

many time-consuming trajectory calculations are required. r’. For various values of L we find that U has the shapes
The computation can be greatly reduced, however, by using plotted in Figure 7-20a. Looking at the contour LP, four
the “flux tube” method. As Parker [1976a] notes, this ad- types of orbits can be defined relative to point r = r,. Adopt-
aptation of the outside-in method, in which all the particle ing the terminology of Parker [1973, 1980], they are (Figure
trajectories between two reference trajectories are assumed 7-20b) as follows:
to be similar, ignores the possibility of orbit crossings or Type 1: The particle has sufficient kinetic energy and
reversals and is only suitable for axisymmetric bodies and small enough angular momentum to reach r, from infinity
cold ion beams [Davis and Harris, 196 1]. Unless otherwise or to reach infinity from point r,. These orbits would con-
stated Parker’s inside-out method is assumed in the follow- tribute only once to a density integral at r,.
ing. The integration of Equation (7.26) then reduces to Type 2: The particle starts at infinity but never reaches
determining the limits on tie trajectories that intersect rP as rs, being repelled at some minimum distance. If the mini-
these will be the limits on the integral. mum distance is inside the region being integrated over, it
The precise method of determining the trajectories that contributes twice as it goes both in and out [Parker, 1980]
intersect rpdifferentiates the VariOUS probe calculations [pw- and zero if it is outside the region of interest.
ker, 1978b, 1980]. Common to all, however, is the so- Type 3: The particle starts at r., but is reflected at some
called orbit classification scheme. Consider the following distance back to r,. These particles contribute twice to the
arguments from classical mechanics (see for example, Gold- integral if rs lies outside the turning point.
stein [1965], Bernstein and Rabinowitz [1959], Whipple Type 4: Type 4 orbits are trapped orbits that circulate
[1976a], Parker [1977, 1980]). The equation for energy around the satellite. These are normally ignored as the orbits
conservation in a spherically symmetric potential V(r) is can only be populated by collisions that are assumed zero
in most satellite studies. As Parker [1980] notes, this as-
sumption has never been justified rigorously.

–1
L* 1
(7.28) As a specific example of these calculations, we will
‘= [ qiv(r)+2rrlr2 ‘imv:
review the spherically symmetric models of Parker [1973,
1975, 1976b, 1980]. These models have been successfully
where E is total energy, L is m v,r = angular momentum, compared to the “Equivalent Potential Formulation” [Par-
v, is total tangential velocity, and v, is radial velocity. ker, 1975] and the particle-pushing code of Rothwell et al.
The two terms in brackets are called the “equivalent (or [1976] (Figure 7-23a). The typical procedure follows. As-
effective) potential” U(r,L) [Goldstein, 1965]. For illustra- suming spherical symmetry, Equations (7.26) and (7. 27)
tion assume the potential V(r) is attractive and of the form can be reduced to two-dimensional integrals. It is’also more

‘J
:“ c
=
I
I

I
I

Figure 7-20a. The “equivalent potential” U (r,L) as a function of r for Figure 7-20b Same as a, only for the LP contour illustrating the four
various values of L. :Iasses of orbits.

7-23
CHAPTER 7
convenient to work in terms of E and L2 so that the equations
become Lr

fi(E,L2) dE dL2
ni (r) = ~
JJ (2m [E – q V(r)] – L2/~)1’2
L:

(7.29a)
o

Ji (r) = % ~~ fi (E,L2) dE dL2. (7.29b) Figure 7-21. Turning-point formulation of Parker [ 1980]. Example illus-
trating domains of four types of orbits in (R, LK2) space.
Four in this plot is in units of r,.
The integrations are now over the allowable ranges of E
and L*. computed. V(r) is found by numerical integration from Pois-
There are two common ways of defining the allowable son’s equation given the ni(r) at grid points around the sur-
range of integration in the (E ,L2) space. The first scheme
face. The process is then iterated until a consistent solution
makes use of the fact that the energy E must be greater than
is found. Recently, Parker ( 1979, 1980) employed the spher-
U(r) if the trajectories are to exist:
ically symmetric model to calculate the sheath of a body of
radius 100 Debye lengths and for a voltage of 400000 kT/q,

1=
the most extreme combination of size and voltage solved
E> qi V(r)+& U(r). (7.30) rigorously to date.
[ More general geometric situations require actual nu-
merical trajectory tracing and are discussed in Parker [ 1973;
The maxima are found in U(r). This is the “Effective Po- 1977; 1978a, b; 1979], Parker and Whipple [1967, 1970],
tential Formulation” [Parker, 1980] and has been utilized Whipple [1977], and Whipple and Parker [ 1969a, b]. While
by Bernstein and Rabinowitz [1959], Laframboise [1966], the above procedure for spherical symmetry is qualitatively
and Chang and Bienkowski [ 1970]. representative of the calculations represented by these pa-
The other approach is to define a function: pers, numerical particle tracing is required if more realistic
geometries such as for the truncated cylinder or “pillbox”
illustrated in Figure 7-22 are to be studied. The final po-
tential contours, in this case for a directed plasma flow
showing the effects of a wake [Parker, 1978a], arrived at
by the iteration process are plotted in Figure 7-22.
where g is called the turning point function [Parker, 1980]. The preceding theoretical studies of Parker, Whipple,
To classify the orbits, the minima in g are found. This and others have been particularly useful in studying the
technique is termed the “Turning-Point Formulation” and effects of differential charging and space-charge potential
has been utilized by Bohm et al. [1949], Allen et al. [1957], minima. Differential charging, as distinct from space-charge
Medicus [1961], and Parker [ 1973, 1975, 1976b]. Although potential minima, has been demonstrated by these and sim-
the two methods are equivalent, Parker [1975] indicates that ilar efforts to result from wake effects [Parker, 1%7a, 1978a,
the Turning-Point Formulation is simpler and more efficient. b], from asymmetric photoelectron emission (see, for ex-
After obtaining equations of the form of Equations (7.29a) ample, Grard et al. [1973]; Fahleson [1973]; Whipple [1976b];
and (7.29b), solutions are found at a given point in space Prokopenko and Laframboise [ 1977b, 1980]; Besse and
assuming V(r) is known. That is, Equations (7.29a) and Rubin, [1980]), and from exposed potentials (Reiff et al.
(7 .29b) are broken up into integrals corresponding to the [1980]; Stevens [1980]). The effects of the space-charge
different orbit types (note that typ 4 orbits are normally potential minimum produced by emitted-electron space charge
ignored). fi (E,L2) is assumed known at the probe surface has been investigated by Soop [ 1972, 1973], Scbroder [ 1973],
(O for no emission or the appropriate values for secondary Parker [ 1976b], Whipple [1976a], and Rothwell et al. [1977]
or photoelectron emission) and, typically, assumed to be a [see also Guernsey and Fu, 1970; Grard and Tunaley, 1971;
Maxwellian in the ambient medium. and Grard et al., 1973]. Whipple [ 1976b] and Parker [ 1976a]
The next step is to determine the integral bounds on E found that such barriers, which are typically a few volts,
and L*. A typical example in terms of the Turning-Point are inadequate to account for the ATS 6 observations of
Formulation is presented in Figure 7-21 [Parker, 1980] and trapped photoelectrons and secondaries, and that a differ-
should be compared with Figure 7-20. For a given value of ential charging barrier must be invoked. Such charging bar-
r and E the limits on L2 are determined. Once the integral riers can significantly affect observations and their existence
ranges of E and L2 are known for r, ni(r) and Ji(r) are must be considered in designing satellite instrumentation.

7-24
 CHARGING OF SPACECRAFT SURFACES
incremented in time using this net potential. The next group
is then moved based on the other particles. The process is
iterated in time with the computer keeping track of surface
interactions (backscattered and secondary electrons, pho-
toelectrons, and the outer boundary) and interactions be-
tween the satellite fields and the particles. Although plasma
simulation codes of this type have been extensively em-
ployed in plasma physics, they are just beginning to be used
for the plasma probe problem.

rd
Although still limited by computer capacity to relatively

‘I L
simple spherical and cylindrical geometries, these models
have been invaluable in studying the detailed effects of space
charge and the time-history of the plasma sheath. Results
I 1--’ / for a spherical model of this type from Rothwell et al. [ 1976,
NON-
CONDUCTING _ 1977] are presented in Figures 7-23a and b. Of considerable
SPACECRAFT
SURFACES
interest are comparisons in Figure 7-23a between the code
and a steady state solution for the case of strong electron
emission [the “PARKSSG’ code of Parker, 1976b]. The
probe in this case was biased at + 2 V and the electrons
FRONT emitted at an energy of leV, the other parameters are as
indicated (secondary emission has been ignored). The re-
sults are meant to resemble isotropic photoemission and
$:;~E
show very good agreement between the two very different
types of codes. Figure 7-23b illustrates the important ability
that such codes have in simulating the time-dependent be-
havior of satellite charging following the turn-on of pho-
toemission as a function of different ambient conditions.
The rapid (2–10ps) rise time of the probe in response to
Figure 7-22. Differential charging of nonconducting spacecraft by di-
rected plasma flow (equipotential contours are in units of
kT/q) [Parker, 1978a].
STRONG EMISSION COMPARISON OF STEADY STATE ( PARKSSG)
WITH TIME SIMULATION (AFGL - SHEATH)

7.4.5 Numerical Simulation Techniques Ro=lm

EMISSION CURRENT = 10-5 amp/m2

EMISSION ENERGY = leV (monoenerget!c electrons)
Although general probe theory can be applied to a num- EMISSION OENSITY =211 /cm3
ber,of interesting and important cases, ithasnotbeenex- AMBIENT TEMPERATURE = 5eV(Maxwelllon)

tended much beyond spherical or cylindrical geometries nor AM BlENT OENSITY = I/cm’
@O = +2V= Sphere Potential (quosi-equilibrium)
does it take into consideration time variations in the sheath. vs. +lVequilibrium

Numerical techniques have been developed that, though 2.0 ❑ TIME - SIMULATION PUTENTIAL – 500
retaining many of the basic concepts of probe theory, allow @ TIME -SIMULATION [ISOTROPIC)
AVG. OF 100 ITERATIONS
explicit inclusion of time dependency or geometry. These t = Izp$

models are capable of handling time variations on the order ~ (PARKSSG - ISOTROPIC)

of the plasma frequency or complex shapes such as the @ ng

(volt s) (cm-’)
- 300
shuttle or the P78-2 SCATHA satellite. :+
1.0
The most straightforward numerical techniques concep-
tually, though perhaps the most demanding computation-
ally, are the so-called “particle pushing” codes. As origi-
nally presented by Albers [1973] and Rothwell et al. [1976,
1977] for a spherical geometry and by Soop [1972, 1973]
and Mazzella et al. [1979] for cylindrical geometry, many
individual or groups of similar particles are followed si-
o~
1.0 142
t ,5 2,0°
r/R.
multaneously by computer as they move through the satellite
sheath. At each time step the potential on a given particle
Figure 7-23a. Comparison of steady state sheath potential and density for
or set of particles due to all the other particles is computed 2 models: the steady state PARKSSG code and the AFGL-
by a fast Poisson equation solver. The particle group is SHEATH code [Rothwell et al., 1977].

7-25
CHAPTER 7
SATELLITE POTENTIAL FOR A STRONG PHOTOEMISSION
AND VARIOUS AMBlENT PLASMA DENSITIES

10 = 0.5eV E ~hOt = 6eV ( MAXWELLIAN) a = ix, /io, ~ue

14=
i Phot a IX 10-5 AMP/m*
12 -
NO SECONDARIES
f 10- a=527n. =1/cc

;8 -

:6 a = 52.7 nO = 10/cc
-
*“ 4 -

2 -

0
-2~

1 1 1 1 1 , t 1 1 1 1 1 1 1 1 1
0 2 4 6 B 10 12 14 16 18 20 22 24 26 28 30
t(p$) -

Figure 7-23b, Spacecraft voltage transients following the “turn-on” of the plasma at t = O. The results are for a sphere (no secondary emission). a is the
ratio of photoelectron to ambient electron temperature, and Emo, is the mean energy of the photoelectrons assumed to be Maxwellian
[Rothwell et al., 1977].

the turn-on of photoemission could be significant in causing

satellite transients as only -2 V are necessary to trigger
many circuits.
The cylindrical, particle pushing code AFSIM (Air Force
Satellite Interactions Model) [Mazzella et al., 1979] has been
employed to follow the effects of an electron beam emitted
by a satellite. A 300 eV beam at 200 PA was emitted into
a vacuum from a satellite initially at O potential. The “com-
putations showed that the satellite potential rose to + 300
Vat 36 ps, at which point a portion of the beam was trapped
in the sheath and began to orbit the satellite. A space-charge
potential formed four spacecraft radii away with most of ‘
the beam being reflected back to the satellite. Figure 7-24 ..
shows the configuration 65 ps after turn on. The circle
represents the satellite and the beam is visible as a collimated
source while the comet-like structure to the right is the cloud
of previously trapped beam electrons orbiting the satellite.

7.4.6 NASCAP

Particle pushing codes are very useful in studying the

detailed behavior of particles in time and in the sheaths
surrounding a probe. They are still, however, limited to
relatively simple geometries because of computer limitations
and are subject to numerical instabilities. Recently an al-
ternative model has become available that explicitly treats —.
Plgure I-24 Sheath simulation using the AFSIM code [Mazzella et al.,
the effects of geometry. This is the NASCAP (NASA Charg- 1970] at 65 us after tum-tm of a 30 eV, 200 UA electron
ing Analyzer Program) computer code [Katz el al., 1977, beam.

7-26
 =

CHARGING OF SPACECRAFT SURFACES

1979; Schnuelle et al., 1979; Roche and Purvis, 1979; Rubin
et al., 1980] for the thick sheath limit (the model is currently
being extended to thin sheaths). This code combines a so-
lution of the Poisson equation and a probe charging model
along with a complex graphics package to compute the -2
detailed time behavior of charge deposition on spacecraft
surfaces. An approximate circuit model of the satellite is
used to estimate voltage changes during time steps for an \
implicit potential solver. The propagation of particle beams
through the satellite sheath is computed by orbit tracing in
the sheath field. The code has several options available
ranging from a less-detailed code capable of calculating the
differential potential on a simple laboratory material sample
to a detailed code capable of modeling individual booms
and sensor surfaces on a complex satellite such as the P78-
-10 I I t I I
~~-3 10-2 1o-1 1 10 l(m
2 SCATHA satellite [Rubin et al., 1980]. The primary intent TIME (SEC)
of the model is to compute the effects of satellite geometry
on the satellite photosheath, sheath field, and surface ma-
Figure 7-25h. NASCAP computer simulation time-dependence of poten-
terial potentials. tial for two surfaces on the P78-2 SCATHA satellite
The first step in the NASCAP model is to insert the ISchnuelle et al., 1979].
satellite geometry and material content. Using simple
3-dimensional building blocks, the code has provisions for
modeling surfaces ranging from cubes and planes to com-
plete satellites with their booms and solar panels (Figure the circuit element model, the inside-out technique, and
7-25a). In the second stage the currents to the satellite sur- Langmuir probe theo~ in one model along with the added
faces are computed by treating each surface element as a advantage of detailed graphical results. Unfortunately, with
current-collecting probe. The inside-out technique may then this gain in capability, the code has become large and re-
be used to model the charge distribution in the photosheath. quires hours of computing time for the more detailed models.
The program is stepped forward in time, local current bal- Two steps have been taken to alleviate these problems. First,
ance being assumed at each time step. Time-dependency efficient versions have been developed for specific com-
and variations in differential potential can thus be modeled
as demonstrated in Figure 7-25b. The analysis is carried out
on successively larger, coarser nested grids [Katz et al.,
1977], allowing the trajectories of particles emitted from
the satellite to be traced in the space surrounding the satellite
(Figure 7-25c).
50
The NASCAP program combines the best elements of

-5C
SCATRA MoDEL

z.., size = 11.5 cm.

Figure 7-25c. Particle trajectories from an electron emitter as computed

Figure 7-25a NASCAP computer simulation of the P78-2 SCATHA sat- by the NASCAP code in the presence of a magnetic field
ellite [Schnuelle et al., 1979]. Compare with Figure 7-6. [Rubin et al., 1980]

7-27
CHAPTER 7
TIME IN SECONOS TIME, Sr3C
60000 S0500 65000 elsoo 6200Q o 50 100 150 m 250
o ,
~
-1 ~AO ‘!
ji . ,=/’?
Ii ‘\ ,~
-2 ?’
i /’ ‘.
/ -u’
/ \;’‘< !~; ‘
-3 /’
lb
I’1
* 1
, I ;,
,’
-4 -, I ‘.
, I
, 1 ‘k i.
.5 -, I
, I
, I
-6 -/ / NASCAP SIMULATED SCATHA CHARGING
RESPONSE FOR OAY 87 ECLIPSE
-7 -~ \

— NASCAP RESPONSE
-.. ‘/ “
---- –- OBSERVED RESPONSE

-9 -

1
Figure 7-26b. Observed potential difference measured by the P78-2 SCA-
, , ,
-10
THA satellite surface potential monitor experiment between
a Kapton sample and the satellite ground. Also shown are
Figure 7-26a, Observed satellite to space potential (P78-2 SCATHA) be- NASCAP predictions from same period [Stevens et al.,
tween 1630 and 1730 UT on day 87, 1979. Also shown 1980a].
are NASCAP predictions for the same period [Schnuelle et
al., 1981].

the potential between the Kapton and the satellite ground

at the spin frequency of the satellite. Second, it illustrate~
mutational tasks such as the SCATHA version discussed in the power of NASCAP in predicting such rapid variations
Rubin et al. [1980]. Secondly, in conjunction with a number (the slight time-lag between the data and theory is due to
of laboratory and in situ satellite experiments, an attempt the actual satellite spin period being slightly faster than the
is being made to verify all aspects of the code and determine assumed 1 rpm). This agreement is, in part, much better
what are the critical input parameters [Purvis et al., 1977; than the prediction of the satellite to space potential as the
Roche and Puwis, 1979; Stevens et al., 1980a]. The model Kapton sample properties were obtained prior to flight in
has been applied to many satellite charging problems such ground test simulations—an important consideration in fu-
as large cavities, exposed potentials, and arbitrary geome- ture studies.
tries not considered previously [Pttrvis, 1980; Stevens, 1980;
Stevens and Pttrvis, 1980].
Preliminary results of comparisons between the P78-2 7.5 PREVENTION OF
SCATHA measurements and NASCAP have recently be- SPACECRAFT CHARGING
come available. Some of these results are plotted in Figures
7-26a and 7-26b. In Figure 7-26a [Schnuelle et al., 1981], Although varying the satellite to space potential allows
the observed satellite to space potential between 1630 and the measurement of very low energy plasma, charge buildup
1730 UT on day 87, 1979 is plotted. The data, consisting on satellite surfaces is not in general a desired phenomenon.
of potential and plasma measurements during an eclipse, In order to eliminate or at least limit the worst effects of
were input into the NASCAP “one-grid” model (as the name spacecraft charging, several techniques have been devel-
implies, this computation makes use of only the inner most oped. Although the obvious solution is to develop systems
NASCAP computational grid—Figure 7-25c) at -1 minute that can withstand the worst effects, this is not always a
(1 spin period) intervals. Although the NASCAP simulation feasible or desirable method. Alternatives to this “brute
misses the two minor jumps in potential, it reproduces the force” method will be described in this section.
two major ones and is in excellent quantitative agreement The simplest method for preventing spacecraft charging
with the data considering the uncertainties in material prop- effects is to employ sound design techniques—use con-
erties. NASCAP also responds more slowly to environ- ducting materials where possible and proper grounding tech-
mental changes than the actual data. As Schnuelle et al. niques. These techniques are detailed in a design guideline
[198 1] note, this is due to the 1 minute time steps imposed handbook recently completed at NASA bwis [Purvis et al.,
on NASCAP by the data whereas the real environment is 1984]. Although a large satellite to space potential can oc-
changing continuously. cur, differential charging, the major spacecraft charging
Figure 7-26b [Stevens et al., 1980c], comparing the problem, is significantly reduced by these procedures. Sev-
NASCAP predictions with the surface potential of a Kapton eral different methods have been developed to assure an
sample (SC 1-2—see Table 7-1) on P78-2-SCATHA, is in- adequately conducting surface. As an example, non-con-
teresting for two reasons. First, it illustrates kV changes in ducting surfaces on the GEOS series of geosynchronous

7-28
 CHARGING OF SPACECRAFT SURFACES
satellites were coated with iridium oxide. As solar cells are dielectric materials may be altered by the arcing process in
the primary non-conducting surfaces on GEOS and as in- a manner which greatly reduces future arcing [A. R. Fred-
dium oxide is sufficiently transparent to sunlight so that it rickson, private communication, 1980]. It is currently
does not degrade their operation, this technique has been thought, however, that the techniques described above are
quite successful [G.L. Wrenn, private communication, 1980] adequate in reducing charging. Basically, spacecraft charge
at keeping satellite differential potentials near zero and, prevention is a matter of good design technique—ground
because of the secondary emission properties of iridium well, avoid cavities in which charge can be deposited, and
oxide, the satellite to space potential between zero and – 1000 avoid exposed potentials.
V even in eclipse. Such coating techniques, however, can
be expensive and difficult if large surfaces are involved.
Furthermore, they do not reduce the hazards associated with 7.6 CONCLUSIONS
charge deposition in dielectrics and, in the case of “pinholes”
(Section 7.3.8), may be ineffective. Before concluding this chapter, a brief summary of the
Another technique that may be applicable to large sur- major accomplishments of this fourth period of charging
faces involves the use of electron and ion emitters. Grard analysis is in order. Probably the major step forward has
[1977] and Gonfalone et al. [1979] discussed the application been the growing realization by the space physics com-
of such systems to actual satellite systems. The latter paper munit y of the role of spacecraft charging. Before the geo-
described the successful application of a low current (mA) synchronous observations of 10(kV) and higher potentials,
electron emitter on the ISEE- 1 satellite. The IS EE- 1 is in a spacecraft charging was considered to be a nuisance. Since
highly elliptical (300 km to 23 R~) orbit so that it spends a that time, however, spacecraft charging analysis has become
long time in the solar wind. The ISEE- 1 results indicated an important adjunct to plasma experiments and to satellite
that the electron cloud emitted by the satellite gun success- design. On the negative side, however, there is still apparent
fully clamped the potential of the satellite at a few volts confusion on the part of some experimentalists as to how
positive to the ambient solar wind plasma. Purvis and Bar- to correct low energy measurements and an unwillingness
tlett [1980] reported results from ion and electron emitters on the part of satellite designers to spend the necessary time
on the geosynchronous ATS 5 and ATS 6 satellites. These in properly designing their satellites. Both of these problems
results indicated that whereas electron emission alone re- have proven hard to solve.
duced the satellite to space potential to -O, it did not sig- In the area of plasma measurements, the spacecraft
nificantly reduce the charge on dielectrics. Use of an ion charging theory necessary for their interpretation can be said
emitter and neutralizer together not only clamped the sat- on the basis of this review to be quite sophisticated. The
ellite to space potential at O, but also, through the cloud of simple probe theory of Langmuir and his successors has
ions, neutralized the negative charge on the dielectrics. Sim- been shown to give adequate order of magnitude estimates
ilar success was demonstrated by the beam experiments on of the gross effects of spacecraft charging. The introduction
the P78-2 SCATHA satellite [H. A. Cohen, private com- of various sophistications such as the satellite velocity or
munication, 1980]. There may be some difficulties, how- the satellite sheath [Cauffman and Maynard, 1974, for ex-
ever, with these techniques as reducing the surface charge ample] have made this theory applicable to many practical
may enhance dielectric breakdown [A. R. Fredrickson, pri- cases. The development of finite element models has al-
vate communication, 1980] between the deposited charge lowed fairly sophisticated engineering studies. The original
and the surface. Also there is the possibilityy of contami- methods of Bernstein and Rabinowitz have grown into the
nation of the satellite environment by the beam ions. intricate, advanced trajectory codes of Parker, Whipple, and
Careful selection of satellite materials can reduce space- others. In combination with various simplifications these
craft charging. Although thermal control surfaces which are techniques have yielded straightforward methods of cor-
necessary on many satellites generally consist of dielectric recting Langmuir probe data. Even secondary emission,
materials, careful selection of the materials according to photoelectron emission, and velocity effects can be mod-
their secondary emission properties and bulk conductivity eled.
can reduce charge buildup. Rubin et al. [1978] have dem- With the advent of electrostatic analyzers and their abil-
onstrated that for materials with a secondary emission water ity to provide both detailed spectral information and mass
than 1, the plasma temperature must be several times the discrimination, experimental information on charging ef-
energy at which this occurs if a satellite is to charge up. fects has increased enormously. As a result, active studies
Again, however, the increased secondary electron popula- of the sheath population and fields in the vicinity of a space-
tion could contaminate low energy (E = 10 eV) plasma craft under a variety of ambient conditions have become
observation. feasible. These have been carried out in detail on the GEOS
Several other techniques have been proposed (for ex- and P78-2 SCATHA satellites for the geosynchro-nous orbit.
ample, see Beattie and Goldstein [1977] for methods of Similar experiments are planned for the early shuttle pay-
protecting the Jupiter probe) and recent results indicate that loads. Theoretical models of the sheaths and fields around

7-29
CHAPTER 7
typical geosynchronous satellites are available, ranging from where:
the simple thick sheath probe models to the NASCAP code qn~o 2T1 “2
IEo = 4mr~z . ~ —
which is capable of handling complex geometries. Ad- ~mE
()
vanced wake and sheath models are also available for shuttle
studies. Rapid time variations on the order of the plasma qn,o 2T, ‘“
1,0 = 47rr,z . ~ —
frequency have even been modeled. ()mml
In the area of engineering design, finite element models
IpH = Tr; o ‘PHO/( 1 + V/0.7)z v>”
have been applied in the design of a number of geosyn-
chronous and interplanetary missions. NASCAP has been ==m< o‘PHO v<”
applied to several systems and proven useful in designing
vehicles so as to avoid the worst effects of charging. On r. = satellite radius
the practical side, effort is’ beginning to be expended in
developing charge-reducing materials and techniques. More For the “RAM” case and assuming a thin sheath for the
importantly, the techniques learned on small spacecraft are electrons:
beginning to be applied in the design of the next generation

v‘+ Ln(lflo’rlpH)v<”
of large, high voltage vehicles.
Despite this impressive growth of spacecraft charging
technology since 1965, there are still a number of areas in
need of study. These can be grouped under the headings of
material properties, geometry, magnetic fields, wakes, arcs,
large size/high potential, and charge deposition in dielec-
trics. Although work is under way in each of these areas,
‘= [’- (lEOio’pH)l
“2m1v:q-’
‘>0
much still remains to be accomplished. Even so, more than
enough has been accomplished in this fourth stage of space- Where: 110 = mr,z . qnlov,
craft charging analysis so that, as a scientific discipline,
spacecraft charging can be said to have come of age. v, = satellite velocity

This assumes, for ion repulsion, that the ion ram cument,
APPENDIX—Table 7-3 Explanation 110, is reduced by a factor ( 1 – qV/ ‘/Zm,v?) and that the
ions have -0 thermal velocity (see Whipple [ 1965], p. 28).
The simple probe models of Section 7.4.1. can give first
The assumed environmental parameters have been adopted
order estimates of the satellite to space potential under a
from many sources. They should be treated at best as rough
variety of conditions. Given the plasma parameters listed
approximations as the actual environments can vary by fac-
in Table 7-3, this potential has been estimated using ap-
tors of X 10 to X 100. The Jupiter data are from Scudder
proximations to Equations (7.22) and (7.24). As only a first
et al. [1981] and J. Sullivan [private communication]. The
order estimate is desired, a conducting spherical satellite
solar wind data for less than 1 AU are from Schwenn et al.
( -1 m in diameter) has been assumed. Secondary and
[1977]. Values greater than 1 AU are estimated.
backscatter terms are ignored (these would tend to make the
potential more positive). For the planar, thin sheath as-
sumption ( I-D in Table 7-3): [T is in eV. ]
ACKNOWLEDGMENTS

‘=+ Ln(loylpH)
‘<0 This work represents the conclusion of 5 years of work
at the Air Force Geophysics Laboratory. During that time,

‘=?Ln(’EOi’pH)
For the spherical,
‘>0
thick sheath assumption (3-D in Table
my main source of advice and direction was C.P. Pike.
Grateful appreciation is also extended to A. Rubin, P. Roth-
well, L. W. Parker, and N. Saflekos, who helped me learn
7-3): the intricacies of spacecraft charging. N .J. Stevens and C.
Purvis have provided much practical assistance and advice.

‘=~Ln[I1o(l-::)
+Ip.l v<” T. Gindorf, P. Robinson, E.C. Whipple, Jr., and P. Leung
helped in the final revision. M. Spanos and B. Short pre-
pared the transcript. Finally, K. G-tt provided much needed
physical support through the long hours expended in pro-

V=—
– T,
Ln
1’0
()I+qy
.
TE
– 1~”-
v>”
ducing the manuscript. The work described in this chapter
was carried out in part at the Jet Propulsion Laboratory,
California Institute of Technology, under NASA contract
q
NAS7-1OO.

7-30
 CHARGING OF SPACECRAFT SURFACES
REFERENCES

Ahmed, M. and R.C. Sagalyn, “Thermal Positive Ions in Bohm, D., “Minimum Ionic Kinetic Energy for a Stable
the Outer Ionosphere and Magnetosphere from OGO 1,“ Sheath,” in The Characteristics of Eleqrical Discharges
J. Geophys. Res., 77: 1205-1220, 1972. in Magnetic Fields, edited by A. Guthrie and R.K. Wak-
Albers, N., “Computer Simulation of a Spherical Langmuir erling, 77–86, McGraw-Hill, New York, 1949.
Probe,” SU-IPR Report 449, Stanford University, Calif., Bohm, D., H.S. Burhop, and H.S. W. Massey, “The Use
1973. of Probes for Plasma Exploration in Strong Magnetic
Allen, J. E., R.L.F. Boyd, and P. Reynolds, “The Collection Fields ,“ in Characteristics of Electrical Discharges in
of Positive Ions by a Probe Immersed in a Plasma,” Magnetic Fields,, edited by A. Guthrie and R. K. Wak-
Proc. Phys. Sot. London, 70: 297–304, 1957. erling, McGraw-Hill, New York, 1949.
Al’pert, J. L., Wave-Like Phenomena in the Near-Earth Bourdeau, R. E., “On the Interaction Between a Spacecraft
Plasma and Interactions with Man-Made Bodies,” in and an Ionized Medium, ” Space Sci. Re~’., 1: 7 19–728,
Handbuch Der Physikj edited by S. Flugge, Geophysics 1963.
III, Part V, 217-350, Springer-Verlag, Berlin 1976. Bourdeau, R. E., J. L. Donley, G.P. Serbu, and E.C. Whip-
Al’pert, J. L., A.V. Gurevich, and L.P. Pitaevskii, Space ple Jr., “Measurement of Sheath Current and Equilib-
Physics with Artificial Satellites, Consultants Bureau, rium Potential on the Explorer VIII Satellite, ” J. Astron.
New York, 186–2 10, 1965. Sci., 8:65-73, 1961.
Balmain, K. G., “Surface Discharge Effects,” in Space Sys: Brundin, C. L., “Effects of Charged Particles on the Motion
terns and Their Interactions with the Earth’s Space En- of an Earth Satellite,” AIAA J., 1: 2529–2538, 1963.
vironment, (frog. Astronaut. Aeronaut., 71 ) edited by Cambou, F., R.Z. Sagdeev, and I.A. Zhulin, “Reviewing
H.B. Garrett and C.P. Pike, 276-298, AIAA Press, Artificial Radiation and Aurorae Between Kerguelen and
New York, 1980. Soviet Union,” Space Sci. Instrument., 4 117–121, 1978.
Balmain, K. G., M. Cuchanski, and P.C. Kremer, “Surface Cauffman, D. P., “Ionization and Attraction of Neutral Mol-
Micro-Discharges on Spacecraft Dielectrics,” in Pro- ecules to a Charged Spacecraft ,“ SAMSO TR-73-263,
ceedings of the Spacecraft Charging Technology Con- 1973.
ference, edited by C.P. Pike and R.R. Lovell, AFGL Cauffman, D. P., “The Effects of Photoelectron Emission
TR-77-005 IINASA TMX-73537, ADA045459, 519-526, on a Multiple-Probe Spacecraft Near the Plasmapause,”
1977. in Photon and Particle Interactions with Surfaces in
Baragiola, R. A., E.V. Alonso, and A.O. Florio, “Electron Space, edited by R.J.L. Grard, 153–162, D. Reidel,
Emission from Clean Metal Surfaces Induced by Low- Dordrecht, Holland, 1973.
Energy Light Ions,” Phys. Rev. B, 19:121-129, 1979. Cauffman, D.P. and N .C. Maynard, “A Model of the Effect
Beard, D. B. and F. S. Johnson, “Charge and Magnetic Field of the Satellite Photosheath on a Double Floating Probe
Interaction with Satellites,” J. Geophys. Res., 65: 1–7, System,” J. Geophys. Res., 79: 2427–2438, 1974.
1960. Cemuschi, F., “The Physics of Cosmic Grains,” Astrophys.
Beard, D.B. and F. S. Johnson, “Ionospheric Limitations J., 105: 241–254, 1947.
on Attainable Satellite Potential ,“ J. Geophys. Res., 66: Chang, K.W. and G.K. Bienkowski, “Effects of Electron
4113-4122, 1961. Emission on Electrostatic Probes at Arbitrary Pres-
Beattie, J.R. and R. Goldstein, “Active Spacecraft Potential sures,” Phys. Fluids, 13:902, 1970.
Control System for the Jupiter Orbiter with Probe Emis- Chang, H.H.C. and M.C. Smith, “On the Drag of a Spher-
sion,” in Proceedings of “theSpacecraft Charging Tech- ical Satellite Moving in a Partially Ionized Atmosphere,”
nology Conference, edited by C. P. Pike and R.R. Lov- J. Brit. [nterplanet. Sot., 17: 199–205, 1960.
ell, AFGL TR-77-oo51/NAsA TMX-73537, Chang, J. S., S.M.L. Prokopenko, R. Godard, and J.G.
ADA045459, 143–165, 1977. Laframboise, ‘-Prediction of Ion Drift Effects on Space-
Bernstein, I. B. and I.N. Rabinowitz, “Theo~ of Electro- craft Floating Potentials,” in Spacecraft Charging Tech-
static Probes in a Low-Density Plasma,” Phys. Fluids, nology—1978, edited by R.C. Finke and C.P. Pick,
2:112, 1959. NASA CP-207 l/AFGL TR-79-0082, ADA084626,
Bernstein, W., et al., “Electron Beam Experiments: The 179–187, 1979.
Beam Plasma Discharge at Low Pressures and Mag- Chen, F. F., “Electric Probes, “ in Plasma Diagnostic Tech-
netic Field Strengths,” Geophys. Res. Lett., 5: 127–1 30, niques, edited by R. H. Huddlestone and S .L. Leonard,
1978. 113–200, Academic Press, New York, 1965.
Bernstein, W., H. Leinbach, P.J. Kellogg, S .J. Monson, Chopra, K. P., “Interactions of Rapidly Moving Bodies in
and T. Hallinan, “Further Laboratory Measurements of Terrestrial Atmosphere,” Rev. Mod. Phys., 33: 153–189,
the Beam-Plasma Discharge,” J. Geophys. Res., 84: 1961.
7271–7278, 1979. Chung, M.S. and T.E. Everhart, “Simple Calculation of
Beers, B. L., et al., “First Principles Numerical Model of nergy Distribution of Low-Energy Secondary Electrons
Avalanche-Induced Arc Discharges in Electron-Irradi- Emitted from Metals Under Electron Bombardment,” J.
ated Dielectrics,” NASA-C-168 (Rev. 10-75), 1979. App/. Phys. , 45:707, 1974.
Besse, A.L. and A.G. Rubin, “A Simple Analysis of Space- Cipolla, J. W. and M. B. Silevitch, “Analytical Study of the
craft Charging Involving Blocked Photoelectron Cur- Time Dependent Spacecraft-Plasma Interaction,” in
rents,” J. Geophys. Res., 85: 2324, 1980. Spacecraft Charging Technology—1978, edited by R.C.

7-31
CHAPTER 7
Finke and C.P. Pike, NASA CP-207 l/AFGL TR-79- Photon and Particle Interactions with Surfaces in Space,
0082, ADA084626, 197-208, 1979. edited by R.J .L. Grard, 363–368, D. Reidel, Dordrecht,
Cohen, H. A., et al., “Design, Development, and Flight of Holland, 1973.
a Spacecraft Charging Sounding Rocket Payload,” in Garrett, H. B., “Review of Quantitative Models of the O to
Spacecraft Charging Technolog>-1978, edited by R.C. 100 KeV Near-Earth Plasma,” Rev. Geophys., 17:
Finke and C.P. Pike, NASA CP-207 l/AFGL TR-79- 397417, 1979.
0082, ADA084626, 80–90, 1979. Garrett, H. B., “The Charging of Spacecraft Surfaces,” Rev.
Davis, A.H. and 1. Harris, “Interaction of a Charged Sat- Geophys. Space Phys., 19: 577+16, 1981.
ellite with the Ionosphere, “ in Rarefied Gas Dynamics, Garrett, H. B. and S. E. DeForest, “Time-Varying Photoe-
edited by L. Talbot, 69 1–699, Academic Press Inc., lectron Flux Effects on Spacecraft Potential at Geosyn-
New York, 1961. chronous Orbit ,“ J. Geophys. Res., M: 2083–2088, 1979.
DeForest, S. E., “Spacecraft Charging at Synchronous Or- Garrett, H. B. and J.M. Forbes, “A Model of Solar Flux
bit,” J. Geophys. Res., 77:651, 1972. Attenuation During Eclipse Passage and Its Effects on
DeForest, S .E., “Electrostatic Potentials Developed by ATS Photoelectron Emission from Satellite Surfaces,” Planet.
5 ,“ in Photon and Particle Interactions with Surfaces Space Sci., 29:601-607, 1981.
in Space, edited by R.J .L. Grard, 263–276, D. Reidel, Garrett, H. B. and A.G. Rubin, “Spacecraft Charging at
Dordrecht, Holland, 1973. Geosynchronous Orbit-Generalized Solution for Eclipse
DeForest, S .E. and C.E. McIlwain, “Plasma Clouds in the Passage,” Geophys. Res. Lett., 5:865, 1978.
Magnetosphere,” J. Geophys. Res., 76:3587-3611, 1971. Garrett, H. B., E.G. Mullen, E. Ziemba, and S.E. DeForest,
Dekker, A.J. “Secondary Electron Emission,” in his Solid “Modeling of the Geosynchronous Plasma Environ-
State Physics, 41845, Prentice-Hall, Englewood Cliffs, ment—Part 2, ATS 5 and ATS 6 Statistical Atlas,”
N. J., 1958. AFGL TR-78-0304, ADA067018, 1978.
deLeeuw, J. H., “A Brief Introduction to Ionospheric Aero- Garrett, H. B., A.G. Rubin, and C.P. Pike, “Prediction of
dynamics ,“ in Rarefied Gas Dynamics, edited by C. L. Spacecraft Potentials at Geosynchronous Orbit ,“ in So-
Brundin, 2:1561 –1 587, Academic Press, New York, lar-Terrestrial Prediction Proceedings, edited by R. F.
1967. Donnelly, Vol. II, 104– 118, U.S. Government Printing
Deutsch, M.-J. C., “Worst Case Earth Charging Environ- Office, 1979.
ment ,“ J. Spacecraft, 19:473477, 1982. Gibbons, D. S., “Seconda~ Electron Emission,” in Hand-
Durrett, J .C. and J .R. Stevens, “Description of the Space book of Vacuum Physics, 2, edited by A.H. Beck, 301,
Test Program P78-2 Spacecraft and Payloads,” in Space- Pergamon Press, Oxford, U. K., 1966.
craft Charging Technology—1 978, edited by R .C. Finke Goldan, P. D., E.J. Yadlowsky, and E.C. Whipple Jr., “Er-
and C.P. Pike, NASA CP-207 l/AFGL TR-79-0082, rors in Ion and Electron Temperature Measurements due
ADA084626, 4-10, 1979. to Grid Plane Potential Nonuniformities in Retarding
Evdokimov, O.B. and N.P. Tubalov, “Stratification of Space Potential Analyzers,” J. Geophys. Res., 78:2907-2916,
Charge in Dielectrics Irradiated with Fast Electrons,” 1973.
Sov., Phys. Solid State, Engl. Trans., 15: 1869–1 870, Goldstein, H., Classicai Mechanics, Addison-Wesley,
1974. Reading, Mass, 63-69, 1965.
Fahleson, U., “Plasma-Vehicle Interactions in Space -- Some Goldstein, R. and S .E. DeForest, “Active Control of Space-
Aspects on Present Knowledge and Future Develop- craft Potentials at Geosynchronous Orbit, ” in Spacecraft
ment ,“ in Photon and Particle interactions with Surfaces Charging by Magnetospheric Plasmas, (Prog Astro-
in Space, edited by R.J. L. Grard, 563–570, D. Reidel, naut. and Aeronaut., 47) edited by A. Rosen, 159– 168,
Dordrecht, Holland, 1973. MIT Press, Cambridge, Mass., 1976.
Feuerbacher, B., R.F. Willis, and B. Fitton, “Electrostatic Gonfalone, A., A. Pedersen, U. V. Fahleson, C.G. Fal-
Charging and Formation of Composite Interstellar Grains,” thammar, F. S. Mozer, and R. B. Tobert, “Spacecraft
in Photon and Particle Interactions with Surfaces in Potential Control on ISEE- 1,“ in Spacecraft Charging
Space, edited by R.J .L. Grard, 415428, D. Reidel, Technology—1 978, edited by R.C. Finke and C. P. Pike,
Dordrecht, Holland, 1973. NASA CP-207 l/AFGL TR-79-0082, ADA084626,
Finke, R.C. and C .P. Pike (eds.), Spacecraft Charging 256-267, 1979.
Technology—1978, NASA CP-207 l/AFGL TR-79-0082, Grard, R.J.L. (cd.), Photon and Particle Interactions with
ADA084626, 1979. Su~aces in Space, D. Reidel, Dordrecht, Holland, 1973.
Fredrickson, A. R., “Electric Fields in Irradiated Dielec- Grard, R. J. L., “Properties of the Satellite Photoelectron
trics ,“ in Spacecraft Charging Technology—I 978, ed- Sheath Derived from Photoemission Laboratory Mea-
ited by R.C. Flnke and C.P. Pike, NASA CP-207 l/AFGL surements,” J. Geophys. Res., 78: 2885–2906, 1973b.
TR-79-0082, ADA084626, 554-569, 1979. Grard, R. J. L., “The Multiple Applications of Electron
Fredrickson, A. R., “Radiation Induced Dielectric Charg- Emitters in Space ,“ in Proceedings of the Spacecraft
ing,” in Space Systems and Their Interactions with the Chargning Technology Conference, edited by C. P. Pike
Earth’s Space Environment, (Prog. Astronaut. Aeron- and R.R. Love]], AFGL TR-005 l/NASA TMX-73537,
aut., 71) edited by H.B. Garrett and C.P. Pike, 38H12, ADA045459, 203-221, 1977.
AIAA Press, New York, 1980. Grard. R.J. L. and J .K .E. Tunaley, “’Photoelectron Sheath
Freeman, J.W. Jr., M.A. Fenner, and H.K. Hills, “The Near the Planar Probe in Interplanetary Space ,“ J. Ge[~-
Electric Potential of the Moon in the Solar Wind,” in phys. Res., 76: 2498–2505, 1971.

7-32
 CHARGING OF SPACECRAFT SURFACES
Grard, R. J. L., K. Knott, and A. Pedersen, “TheInfluence and Analysis of the Echo from Electrons Artificially In-
of Photoelectron and Secondary Electron Emission on jected into the Magnetosphere,” Tech. Report CR-160,
Electric Field Measurements in the Magnetosphere and School Physics and Astronomy, University of Minne-
Solar Wind,” in Photon and Particle Interactions with sota, Minneapolis, 1972.
.Sza--aces in Space, edited by R.J .L. Grard, D. Reidel, Hess, W. N., “Generation of an Artificial Aurora,” Science,
Dordrecht, Holland, 163–1 92, 1973. 164:1512, 1969.
Grard, R. J. L., S.E. DeForest, and E.C. Whipple Jr., “Com- Hinteregger, H. E., “Combined Retarding Potential Analysis
ment on Low Energy Electron Measurements in the Jo- of Photoelectrons and Environment Charged Particles
vian Magnetosphere,” Geophys. Res. Lett., 4:247, 1977. up to 234 km,” Space Res., 1: 304–327, 1960.
Grebowski, R. and T. Fischer, “Theoretical Density Dis- Inouye, G. T., “Spacecraft Potentials in a Substorm Envi-
tribution of Plasma Around a Cylinder,” Planet. Space ronment, ” in Spacecraft Charging by Magnetospheric
Sci., 23:287-304, 1975. Plasma, (Prog. Astronaut. Aeronaut., 42) edited by
Gringauz, K.I. and M. Kh. Zelikman, “Measurement of the A. Rosen, 103– 120, MIT Press, Cambridge, Mass.,
Concentration of Positive Ions Along the Orbit of an 1976.
Artificial Satellite,” Uspekhi Fiz. Nauk, 63: 239–252, Jacobsen, T.A. and N .C. Maynard, “Evidence for Signif-
1957; (translated in The Russian Literature of Satellites, icant Spacecraft Charging by an Electron Accelerator at
Part 11, International Physical Index, New York, 133–147, Ionospheric Altitudes,” Planet. Space Sci., B: 291-307,
1958.) 1980.
Gringauz, K. I., V. V. Bezrukikh, and V.D. Ozerov, “Mea- Jastrow, R and C.A. Pearse, “Atmospheric Drag on the
surements of the Concentration of Positive Ions in the Satellite.” J. Geouhys. Res., 62:413423,
1.
1957.
Ionosphere by Ion Traps on the Third Soviet Earth Sat- Jemiola, J. M., Proceedings of the USAF/NASA Interna-
ellite,” Iskusstv. Sputniki Zemli, 6: 63–100, 1961. tional Spacecraft Contamination Conference, AFML TR-
Gross, B. and S. V. Nablo, “High Potentials in Electron- 78- 1901NASA CP-2039, 1978.
Irradiated Dielectric,” J. Appl. Phys., 38: 2272–2275, Jemiola, J. M., “Spacecraft Contamination: A Review,” in
1967. Space Systems and Their Interactions with Earth’s Space
Guernsey, R.L. and J. H .M. Fu, “Potential Distributions Environment, (Prog. Astronaut. Aeronaut., 71) edited
Surrounding a Photo-Emitting Plate in a Dilute Plasma,” by H. B. Garrett and C. P. Pike, 680–706), AIAA Press,
1. Geophys. Res., 75:3 193–3 199, 1970. New York, 1980.
Gurevich, A.V. and Ya.S. Dimant, “Flow of a Rarefied Jew, H., “Numerical Studies of the Rarefied Plasma Inter-
Plasma Around a Disk,” Geomagn. Aeron., Engl. Transl., action at Meso-Thermal Speeds,” Ph.D. Thesis, Uni-
15:183-190, 1975. versity of Michigan, Ann Arbor, Mich., 1968.
Gurevich, A. V. and L. P. Pitaevskii, “Non-Linear Dynamics Jew, H., “Reply,” J. Geophys. Res., 78:6829, 1973.
of Raretied Gas, ” Prog. Aerospace Sci., 16:227, 1975. Johnson, B., J. Quinn, and S .E. DeForest, “Spacecraft
Gurevich, A. V., L.V. Pariiskaya, and L.P. Pitaevskii, “Self- Charging on ATS-6,” in Effect of the Ionosphere on
Similar Motion of Rarefied Plasma,” Sov. Phys. JETP, Space and Terrestrial Systems, edited by John Good-
Engl. Transl., 22:449, 1966. man, U.S. Government Printing Office, Washington,
Gurevich, A. V., L.V. Pariiskaya, and L.P. Pitaevskii, “Self- D. C., 322-327, 1978.
Similar Motion of Low Density Plasma,” Sov. Phys. Johnson, C. Y. and E.B. Meadows, “First Investigation of
JETP, Engl. Transl., 27:476, 1968. Ambient Positive-Ion Composition to 219 km by Rocket-
Gurevich, A. V., L.V. Pariiskaya, and L.P. Pitaevskii, “Ion Borne Spectrometer,” J. Geophvs. Res., 60: 193-203,
Acceleration Upon Expansion of a Rarefied Plasma,” 1955.
Sov. Phys. JETP, Engl. Transl., 36:274-281, 1973. Jung, B., “The Origin of Solid Particles in Interstellar Space,”
Gurevich, A. V., L.P. Pitaevskii, and V. V. Smimov, “Ion- Astron. Nach., 263:426, 1937.
ospheric Aerodynamics,” Sov. Phys. Usp., 99: 595, Kasha, M. A., The Ionosphere and its Interaction with Sat-
1970. ellites, Gordon and Breach, New York, 1969.
Hachenberg, O. and W. Brauer, “Secondary Electron Emis- Katz, I., D.E. Parks, S. Wang, and A. Wilson, “Dynamic
sion from Solids ,“ in Advances in Electronics and Elec- Modeling of Spacecraft in a Collisionless Plasma,” in
tron Physics, edited by L. Martin, 11: 413499, Aca- Proceedings of the Spacecraft Charging Technology
demic Press, New York, 1959. Conference, edited by C.P. Pike and R.R. Lovell, AFGL
Hanson, W. B., D.D. McKibbin, and G. W. Sharp, “Some TR-77-005 liTMX-73537, ADA045459, 319-330, 1977.
Ionospheric Measurements with Satellite-Borne Ion Katz, I., J.J. Cassidy, M.J. Mandell, G.W. Schnuelle, P.G.
Traps,” J. Geophys. Res., 69: 2747–2763, 1964. Steen, and J.C. Roche, “The Capabilities of the NASA
Hanson, W .B., S. Sanatani, D. Zuccaro, and T. W. Flow- Charging Analyzer Program,” in Spacecraft Charging
erday, “Plasma Measurements with the Retrading Po- Technology—1978, edited by R.C. Finke and C. P. Pike,
tential Analyzer on OGO 6,” J. Geophys. Res., 75: NASA CP-2071/AFGL TR-79-0082, ADA045459,
5483-5501, 1970. 101-122, 1979.
Henderson, C. L. and U. Samir, “Observations of the Dis- Kennerud, K. L., “High Voltage Solar Array Experiments,”
turbed Region Around an Ionospheric Spacecraft ,“ Planet, NASA CR-131280, Boeing Co., Seattle, March 1974.
Space Sci., 15: 1499-1513, 1967. Knott, K., “The Equilibrium Potential of a Magnetospheric
Hendrickson, R. A., “The Electron Echo Experiment, Ob- Satellite in an Eclipse Situation ,“ Planet. Space Sci.,
servations of the Charge Neutralization of the Rocket 20: 1137-1146, 1972.

7-33
CHAPTER 7
Knott, K., “Payload of the ‘GEOS’ Scientific Geostationary Lucas, A .A., Fundamental Processes in Particle and Photon
Satellite,” ESA Sci, Tech. Rev., 1: 173–196, 1975. Interactions with Surfaces, “ in Photon and Particle in-
Knudsen, W .C., “Evaluation and Demonstration of the Use teractions with Su@aces in Space, edited by R.J .L. Grard,
of Retarding Potential Analyzers for Measuring Several 3-21, D. Reidel, Dordrecht, Holland, 1973.
Ionospheric Quantities,” J. Geophys. Res., 71: 4669, Manka, R. H., “Plasma and Potential at the Lunar Surface,”
1966. in Photon and Particle Interactions with Su~aces in
Knudsen, W .C. and G. W. Sharp, “Ion Temperatures Mea- Space, edited by R.J .L. Grard, D. Reidel, Dordrecht,
sured Around a Dawn-Dusk AuroraI-Zone Satellite Or- Holland, 347-367, 1973.
bit,” J. Geophys. Res., 72:1061-1072, 1967. Massaro, M. J., T. Green, and D. Ling, “A Charging Model
Koorts, H. C., “Characteristics of Electrical Discharges on for Three-Axis Stabilized Spacecraft,” Proceedings of
the P78-2 Satellite (SCATHA) ,“ Paper 80-0333, AIAA the Spacecraft Charging Conference, edited by C.P.
18th Aerospace Sciences Meeting, AIAA Press, New Pike and R.R. Lovell, AFGL TR-77-005 l/NASA TMX-
York, 1980. 73537, ADA045459, 237-270, 1977.
Krassovsky, V. I., “Exploration of the Upper Atmosphere Mazzella, A., E. Tobenfeld, and A.G. Rubin, “AFSlM-
with the Help of the Third Soviet Sputnik,” Proc. IRE, An Air Force Satellite Interactions Model,” AFGL TR-
47: 289–196, 1959. 79-0138, ADA078032, 1979.
Kunemann, B., “The Collisionless Flow of Unmagnetized McCoy, J. E., A Konradi, and O.K. Garriott, “Current
Plasma Around Bodies, 1,“ J. Plasma Phys., 20: 17, Leakage for Low Altitude Satellites ,“ in Space Systems
1978. and Their Interactions with Earth’s Space Environment,
Kurt, P.G. and V.I. Moroz, “The Poterttial of a Metal Sphere Prog. Astronaut. Aeronaut., edited by H. B. Garrett
in Interplanetary Space ,“ Iskusstvennye Sputniki Zemli, and C.P. Pike, 71: 523–553, AIAA Press, New York,
7: 78–88, 1961; (translated in Planet. Space Sci., 9: 1980.
259-268, 1962). McPherson, D.A. and W .R. Schober, “Spacecraft Charging
Laframboise, J .G., “Theory of Spherical and Cylindrical at High Altitudes: The SCATHA Satellite Program,” in
Langmuir Probes in a Collisonless Maxwellian Plasma Spacecraji Charging by Magnetospheric Plasmas, Prog.
at Rest ,“ Report 100, University of Toronto Institute of Astronaut. Aeronaut., edited by A. Rosen, 47: 15–30
Aerospace Studies, Canada, 1966. MIT Press, Cambridge, Mass., 1976.
Laframboise, J. B. and E.C. Whipple Jr., Comments on Medicus, G., “Theory of Electron Collection of Spherical
paper by Uri Samir and Howard Jew, “Comparison of Probes,” J. Appi. Phys., 32:25 12–2520, 1961.
Theory with Experiment for Electron Density Distri- Meulenberg, A. Jr., “Evidence for a New Discharge Mech-
bution in the Near Wake of an Ionospheric Satellite,” anism for Dielectrics in a Plasma,” in Spacecraft Charg-
J. Geophys. Res., 78: 6827–6828, 1973. ing by Magnetospheric Plasmas, Prog. Astronaut.
Laframboise, J. G., R. Godard, And S .M.L. Prokopenko, Aeronaut., edited by A. Rosen, 47: 237–246, MIT Press,
“Numerical Calculations of High-Altitude Differential Cambridge, Mass., 1976.
Charging: Preliminary Results, ” in Spacecraft Charging Mizera, P. F., “Natural and Artificial Charging” Results from
Technology—1978, edited by R.C. Finke and C.P. Pike, the Satellite Surface Potential Monitor Flown on P78-
NASA CP-207 l/AFGL TR-79-0082, ADA084626, 2,” Paper 80-0334, AIAA 18th Aerospace Sciences
188-196, 1979. Meeting, AIAA Press, New York, 1980.
Langmuir, 1., “The Interaction of Electron and Positive Ion Mizera, P. F., E. R. Schnauss, R. Vandre, and E.G. Mullen,
Spacecharge in Cathode Sheaths,” Phys. Rev., 33&34, “Description and Charging Results from the RSPM,” in
1929 (reprinted in Collected Works of Irving Langmuir, Spacecraft Charging Technology—1978, edited by R.C.
edited by G. Suits, 5, MacMillan, New York, 1961). Finke and C.P. Pike, NASA CP-207 l/AFGL TR-79-
Langmuir, I., and K.B. Blodgett “Currents Limited by Space 0082, ADA084626, 91-100, 1979.
Charge Between Concentric Spheres,” Phys Rev., 24: Montgomery, M. D., J.R. Ashbridge, S.J. Bame, and E.W.
49, 1924. Hones, “Low-Energy Electron Measurements and
Lehnert, B., “Electrodynamics Effects Connected with the Spacecraft Potential: Vela 5 and Vela 6,” in Photon and
Motion of a Satellite of the Earth,” Tellus, 3:408409, Particle Interactions with Surfaces in Space, edited by
1956. R.J. L. Grard, 247–262, D. Reidel, Dordrecht, Holland,
Liemohn, H. B., “Induced Charging of Shuttle Orbiter by 1973.
High Electron-Beam Currents,” in Proceedings of the Mott-Smith, H.M. and I. Langmuir, “The Theory of Col-
Spacecraft Charging Technology Conference, edited by lectors in Gaseous Discharges,” Phys. Rev., B: 727-763,
C.P. Pike and R.R. Lovell, AFGL TR-77-005 l/NASA 1926.
TMX-73537, ADA045459, 271-286, 1977. Mullen, E.G. and M. S. Gussenhoven, “SCATHA Envi-
Linson, L. M., “Current-Voltage Characteristics of an Elec- ronmental Atlas,” AFGL TR-83-0002, ADA 131456,
tron-Emitting Satellite in the Ionosphere,” J. Geophys. 1983.
Res., 74: 2368–2375. 1969. Mullen, E. G., M. S. Gussenhoven, and H.B. Garrett, “A
Liu, V .C. and H. Jew, “Near Wake of the Rarefield Plasma ‘Worst Case’ Spacecraft Environment as Observed by
Flows at Mesothermal Speeds,” Paper 68-169, AIAA SCATHA on 24 April 1979,” AFGL TR-81-0231,
Press, New York, 1968. ADA10868O, 1981.
Liu, V. C., “Ionospheric Gas Dynamics of Satellites and Narcisi, R. S., A.D. Bailey, and L. Della Lucca, “Com-
Diagnostic Probes,” Space Sci. Rev., 9:423, 1969. position Measurements of Negative Ions in the D and

7-34
 CHARGING OF SPACECRAFT SURFACES
Lower E Regions (Abstract),” Eos Trans. AGU, 49: Parker, L. W., 6’Plasmasheath-Photosheath Theory for Large
149, 1968. High-Voltage Space Structures,” in spacesystemsand
Nanevicz, J. E. and R.C. Adamo, “Occurrence of Arcing Their Interactions tvith Earth’s Space Environment, (Prog.
and Its Effects on Space Systems,” in Space Systems Astronaut. Aeronaut., 71) edited by H. B. Garrett and
and Their Interactions with Earth’s Space Environment, C.P. Pike, 477–522, AIAA Press, New York, N. Y.,
(Prog. Astronaut. Aeronaut., 71) edited by H.B. Garrett 1980.
and C. P. Pike, 252–275, AIAA Press, New York, 1980. Parker, L. W., and B .L. Murphy, “Potential Buildup on an
Norman, K., and R. M. Freeman, “Energy Distribution of Electron-Emitting Ionospheric Satellite,” J. Geophys.
Photoelectron Emitted from a Surface on the OGO-5 Res., 72: 1631, 1967.
Satellite and Measurements of Satellite Potential,” in Parker, L. W., and E.C. Whipple, Jr., “Theory of a Satellite
Photon and Particle Interactions with Surfaces in Space, Electrostatic Probe,” Ann. Physics, 44: 126–1 61, 1967.
edited by R.J. L. Grard, 23 I–246, D. Reidel, Dordrecht, Parker, L. W., and E.C. Whipple, Jr., “Theory of Spacecraft
Holland, 1973. Sheath Structure, Potential, and Velocity Effects on Ion
Olsen, R.C. “Differential and Active Charging Results from Measurements by Traps and Mass Spectrometers,” ./.
the ATS Spacecraft,” Ph.D. Thesis, University of Cal- Geophys. Res., 75:4720, 1970.
ifornia, San Diego, 1980. Parks, D. E., and I. Katz, “Charging of a Large Object in
O’Neil, R. R., F. Bien, D. Burt, J.A. Sandock, and A.T. Low Polar Earth Orbit ,“ in Proceedings USAFINASA
Stair, “Summarized Results of the Artificial Auroral Ex- Spacecraft Charging Technical Conference, [11, 1981.
periment, Precede,” J. Geophys. Res., 83:32’73, 1978a. Pike, C.P. and M. H. Bunn, Spacecrtift Charging by Mag-
O’Neil, R. R., W.P. Reidy, J.W. Carpenter, R.N. Davis, netospheric Plasmas, (Prog. Astronaut. Aeronaut., 47)
D. Newell, J .C. Ulwick, and A.T. Stair, “Excede 2 edited by A. Rosen, 45-60, MIT Press, Cambridge,
Test, An Artificial Auroral Experiment: Ground-based Mass., 1976.
Optical Measurements,” Y. Geophys. Res., 83:3281, Pike C. P., and R. R. Loveli (eds. ), Proceedings of the
1978b. Spacecraft Charging Technology Conference, AFGL-
Opik, E. J., “Interplanetary Dust and Terrestrial Accretion TR-77-005 l/NASA TMX-73537, 1977.
of Meteoric Matter,” Irish Astron. J., 4: 84, 1956. Prokopenko, S. M .L., and J .G. Laframboise, “Prediction of
Opik, E. J., “Particle Distribution and Motion in a Field of Large Negative Shaded-side Spacecraft Potentials,” in
Force,” in Interactions of Space Vehicles with an Ionized Proceedings of the Spacecraft Charging Technology
Atmosphere, edited by S .F. Singer, 340, Pergamon Conference, edited by C.P. Pike and R.R. Lovell, AFGL-
Press, New York, 1965. TR-77-oo51/NAsA, TMX-73537, ADA045459,
Parker, L. W., “Computer Solutions in Electrostatic Probe 369-387, 1977.
Theory ,“ Mt. Auburn Res. Associates, Inc., Newton, Prokopenko, S. M. L., and J .G. Laframboise, “High-Volt-
Mass., AFAL TR-72-222, 1973. age Differential Charging of Geostationary Spacecraft,”
Parker, L. W., “Computer Method for Satellite Plasma Sheath J. Geophys. Res., 85:41254131, 1980.
in Steady-State Spherical Symmetry,” Lee W. Parker, Inc., Purvis, C. K., “Configuration Effects on Satellite Charging
Concord, Mass., AFCRL-TR-75-O41O, ADA015066, Response ,“ Paper 80-0040, AIAA 18th Aerospace Sci-
July 1975. ence Conference, 1980.
Parker, L. W., “Computation of Collisionless Steady State Purvis, C. K., and R.O. Bartlett, “Active Control of Space-
Plasma Flow Past a Charged Disk,” NASA CR-144159, craft Charging, ” in Space Systems and Their Interactions
1976a. With the Earth’s Space Environment, (Prog. Astronaut.
Parker, L. W., “Theory of Electron Emission Effects in Aeronaut., 71] edited by H. B. Garrett and C. P. Pike
Symmetric Probe and Spacecraft Sheaths,” Lee W. Par- 299–3 17, AIAA Press, New York, N. Y., 1980.
ker, Inc., Concord, Mass., AFGL-TR-76-0294, A037538, Purvis, C. K., H.B. Garrett, A. Whittlesey, and N.J. Ste-
1976b. vens, Handbook of Design Guidelines for Assessing and
Parker, L. W., “Calculation of Sheath and Wake Structure Controlling Spacecraft Charging Effects, NASA Lewis
about a Pillbox-Shaped Spacecraft in a Flowing Plasma,” Research Center, NASA TP 2361, September 1984.
in Proceedings of the Spacecraft Charging Technical Purvis, C. K., N.J. Stevens, and J.C. Oglebay, “Charging
Conference, edited by C.P. Pike and R.R. Lovell, AFGL- Characteristics of Materials: Comparison of Experimen-
TR-77-005 lINASA TMX-73537, ADA045459, 331-366, ‘ tal Results with Simple Analytical Models,” in Pro-
1977. ceedings Spacecraft Charging Conference, edited by
Parker, L. W., “Differential Charging and Sheath Asym- C.P. Pike and R.R. Lovell, AFGL-TR-77-005 l/NASA
metry of Nonconducting Spacecraft Due to Plasma Flow,” TMS-73537 ADA045459, 459486, 1977.
J. Geophys. Res., 83:48734876, 1978a. Reasoner, D. L., W. Lennartsson, and C. R. Chappell, “Re-
Parker, L. W., “Potential Barriers and Asymmetric Sheaths lationship Between ATS-6 Spacecraft Charging Occur-
Due to Differential Charging of Nonconducting Space- rences and Warm Plasma Encounters ,“ in Spacecraft
craft ,“ AFGL-TR-78-O045, ADA0536 18, 1978b. Charging by Magnetospheric Plasmas, (Prog. Astro-
Parker, L. W., “Plasma Sheath Effects and Voltage Distri- naut. Aeronaut., 47) edited by A. Rosen, 89– 102, MIT
butions of Large High Power Satellite Solar Array s,” in Press, Cambridge, Mass., 1976.
Spacecraft Charging Technology—1 978, edited by R.C. Reddy, B. M., L.H. Brace, and J.A. Findlay, “The Iono-
Finke and C. P. Pike, NASA CP-2071 /AFGL TR 79- sphere at 640 Kilometers on Quiet and Disturbed Days,”
0082 ADA084626, 34 1–375, 1979. J. Geophys. Res., 72:2709-2727, 1967.

7-35
CHAPTER 7
Reiff, P. H., J. W. Freeman, and D.L. Cooke, “Environ- Agena 10 Spacecraft,” in Photon and Particle inter-
mental Protection of the Solar Power Satellite ,“ in Space actions with Surfaces in Space, edited by R.J. L. Grard,
Systems and Their Interactions With the Earth’s Space 193-220, D. Reidel, Dordrecht, Holland, 1973.
Environment, (Prog. Astronaut. Aeronaut., 71) edited Samir, U., and H. Jew, “Comparison of Theory with Ex-
by H.B. Garrett and C.P. Pike, 554–576, AIAA Press, periment for Electron Density Distribution in the Near
New York, 1980. Wake of an Ionospheric Satellite,” J. Geophys. Res.,
Robinson, P. A., Jr., and A.B. Holman, “Pioneer Venus 77:6819-6827, 1972.
Spacecraft Charging Model, “ in Proceedings Spacecraft Samir, U., and G.L. Wrenn, “The Dependence of Charge
Charging Technical Conference, edited by C.P. Pike and Potential Distribution Around a Spacecraft on Ionic
and R.R. Lovell, AFGL-TR-77-O05 l/NASA TMX- Composition,” Planet, Space Sci., 17:693, 1969.
73537, 297-308, ADA045459, 1977. Samir, U., L.H. Brace, and H.C. Brinton, “About the In-
Roche, J. C., and C.P. Purvis, “Comparison of NASCAP fluence of Electron Temperature and Relative Ionic
Predications with Experimental Data,” in Spacecraft Composition on Ion Depletion in the Wake of the AE-
Charging Technology—1 978, edited by R.C. Finke and C Satellite,” Geophys. Res. Lett., 6: 101–104, 1979a.
C.P. Pike, NASA CP- 1072/AFGL-TR-79-oo82, Samir, U., R. Gordon, L. Brau, and R. Theis, The Near-
ADA084626, lM-157, 1979. Wake Structure of the Atmosphere Explorer C. (AE-C)
Rosen, A. (cd.), Spacecraft Charging by Magnetospheric Satellite: A Parametric Investigation,” J. Geophys. Res.,
Plasmas, (Prog. Astronaut. Aeronaut., 47) MIT Press, 84: 513–525, 1979b.
Cambridge, Mass., 1976. Samir, U., Y.J. Kaufmann, L.H. Brace, and H.C. Brinton,
Rosen, A., R. W. Fredericks, Inouye, N.L. Saunders, F.L. “The Dependence of Ion Density in the Wake of the
Scarf, W.M. Greenstadt, J.L. Vogl, and J.M. Sellen, AE-C Satellite on the Ratio of Body Size to Debye
“RGA Analysis: Findings Regarding Correlation of Sat- Length in an (O+ )-Dominated Plasma,” Y. Geophys.
ellite Anomalies with Magnetospheric Substorms and Res., 85: 1769–1772, 1980.
Laboratory Test Results,” Report 09760-7020-RO-W, Sanders, N.L. and G.T. Inouye, “Secondary Emission Ef-
Aug 1972, TRW Systems Group, Redondo Beach, Calif. fects on Spacecraft Charging: Energy Distribution Con-
Rothwell, P. L., A.G. Rubin, A.L. Pavel, and L. Katz, siderations,” in Spacecraft Charging Technology1978,
“Simulation of the Plasma Sheath Surrounding a Charged edited by R.J. Flnke and C.P. Pike, NASA CP-
Spacecraft ,“ in Spacecraft Charging by Magneto- 207 l/AFGL TR-79-0082, 747–755 , 1979.
spheric Plasmas, (Prog. Astronaut. Aeronaut., 47) ed- Schnuelle, G. W. et al., D.E. Parks, I. Katz, M.J. Mandell,
itedby A. Rosen, 12 1–1 34, MIT Press, Cambridge, Mass., P.G. Steen, J.J. Cassidy, and A. Rubin, “Charging
1976. Analysis of the Scatha Satellite,” in Spacecraft Chrging
Rothwell, P. L., A.G. Rubin, and G.K. Yates, “A Simu- Technology—1978, edited by R.C. Finke and C.P. Pike,
lation Model of Time-Dependent Plasma-Spacecraft In- NASA CP-207 l/AFGL TR-0087, 123–143, 1979.
teractions,” Proc. Spacecraft Charging Conf., edited by Schnuelle, G. W., P.R. Stannard, I. Katz, and M.J. Man-
C.P. Pike and R.R. Lovell, AFGL-TR-77-005 l/NASA dell, “Stimulation of the Charging Response of the
TMS-73537, ADA045459, 381412, 1977. SCATHA (P78-2) Satellite,” in Proceedings of the
Rubin, A. G., P.L. Rothwell, and G.K. Yates, “Reduction USAFINASA Spacecraft Charging Technical Confer-
of Spacecraft Chtiging Using Highly Emissive Surface ence 111, 1981.
Materials,” Proceedings of the 1978 Symposium on the Schroder, H., “Spherically Symmetric Model of the Pho-
Effects of the Ionosphere on Space and Terrestrial Sys- toelectron Sheath for Moderately Large Plasma Debye
tems, Naval Research Laboratory, Washington, D. C., Lengths,” in Photon and Particle Interactions with Sur-
1978. faces in Space, edited by R.J.L. Grard, 51-58, D. Rei-
Rubin, A.G. et al., I. Katz. M. Mandell, G. Schnuelle, P. del, Dordrecht, Holland, Mass., 1973.
Steen, D. Parks, J. Cassidy, and J. Roche, “A Three- Schwenn, R., H. Rosenbauer, and K.H. Muhlhauser, “The
Dimensional Spacecraft Charging Computer Code,” in Solar Wind during STIP 11 Interval: Stream Structures,
Space Systems and Their Interactions With the Earth’s Boundaries, Shocks, and Other Features as Observed
Space Environment, (Prog. Astronaut. Aeronaut., 71) by the Plasma Instruments on Helios- 1 and Helios-2,”
edited by H. B. Garrett and C.P Pike, 3 18–336, AIAA Contributed Papers to the Study of Traveling interpla-
Press, New York, 1980. netary Phenomenal1977, edited by M.A. Shea, D.F.
Sagalyn, R. C., and W.J. Burke, “lNJUN 5 Observations Smart and S.T. Wu, AFGL TR-77-0309, ADA056931,
of Vehicle Potential Fluctuations at 2500 km,” in Pro- 1977.
ceedings of Spacecraft Charging Technology Confer- Scudder, J.1)., E.C. Sittler, Jr., and H.S. Bridge, “A Survey
ence, edited by C.P. Pike and R. R. Lovell, AFGL TR- of the Plasma Electron Environment of Jupiter: A View
77-0051, NASA TMS-73537, 67-80, 1977. from Voyager,” J. Geophys. Res., 86:8157, 1981.
Sagalyn, R. C., M. Smiddy, and J. Wisnia, “Measurement Shaw, R. R., J .E. Nanevicz, and R.C. Adamo, “Observa-
and Inteqretation of Ion Density Distributions in the tions of Electrical Discharges Caused by Differential
Daytime F Region,” J. Geophys. Res., 63: 199–2 11, Satellite Charging,” Spacecraft Charging by Magneto-
1963. spheric Plasmas, AIAA (Prog. Astronaut. Aeronaut.,
Samir, U., “chargedParticle Distribution in the Nearest 47) edited by A. Rosen, 61-76, 1976.
Vicinity of Ionospheric Satellites - Comparison of the Shawhan, S. D., R.F. Hubbard, G. Joyce, and D.A. Gur-
Main Results from the Ariel I, Explorer 31, and Gemini- nett, “Sheath Acceleration of Photoelectrons by Jupiter’s

7-36
 CHARGING OF SPACECRAFT SURFACES
Satellite Io,” in Photon and Particle Interactions with Stationary Charged Sphere and a Rapidly Moving Charged
Su~aces in Space, edited by R.J.L. Grard, 504-414, Body,” in Interactions of Space Vehicles with an Ionized
D. Reidel, Dordrecht, Holland, 1973. Atmosphere, edited by S .F. Singer, 61 – 162, Pergamon
Singer, S.F. (ed. ), Interactions of Space Vehicles withan Press, New York, 1965.
lonizedAtmosphere, Pergamon Press, New York, 1965. Walker, E. H., “Plasma Sheath and Screening of Charged
Soop, M., “Report on Photosheath Calculations for the Sat- Bodies, “ in Photon and Particle Interactions with Sur-
ellite GEOS,” Planet. Space Sci., 20:859, 1972. faces in Space, edited by R.J.L. Grard, D. Reidel, Dor-
Soop, M., “Numerical Calculations of the Perturbation of drecht, Holland, 73-89, 1973.
an Electric Field Around a Spacecraft ,“ in Photon and Whipple, E. C., Jr., The Ion-Trap Results in the “Explo-
Particle Interactions with Surjaces in Space, edited by ration of the Upper Atmosphere with the Help of the
R.J.L. Grard, 127, D. Reidel, Dordrecht, Holland, 1973. Third Soviet Sputnik,” IRE Trans., 47:2023-2024, 1959.
Spencer, N. W., L.H. Brace, G.R. Corignan, D.R. Taeusch, Whipple, E. C., Jr., “The Equilibrium Electric Potential of
and H. Nieman, “Electron and Molecular Nitrogen Tem- a Body in the Upper Atmosphere,” NASA X-615-65-
perature and Density in the Thermosphere,” J. Geophys. 296, 1965.
Res., 70: 2665–2698, 1965. Whipple, E. C., Jr., “Theory of the Spherically Symmetric
Spitzer, L., Jr., “The Dynamics of the Interstellar Medium,” Photoelectron Sheath: A Thick Sheath Approximation
Astrophys. J., 93: 369–379, 1941. and Comparison with the ATS-6 Observations of a Po-
Spitzer, L., Jr., “The Temperature of Interstellar Material,” tential Barrier,” J. Geophys. Res., 81:601-607, 1976a.
Astrophys. J., 107: 6–33, 1948. Whipple, E. C., Jr., “Observation of Photoelectrons and
Spitzer, L., Jr., and M.P. Savedoff, “The Temperature of Secondary Electrons Reflected from a Potential Barrier
Interstellar Matter, III,” Astrophys. J., 111:593-608, in the Vicinity of ATS-6,” J. Geophys. Res., 81: 715,
1950. 1976b.
Sternglass, E. J., “Backscattering of Kilovolt Electrons from Whipple, E. C., Jr., “Modeling of Spacecraft Charging,” in
Solids,” Phys. Rev., 95:345, 1954. Proceedings of Spacecraft Technical Conference, edited
Stemglass, E. J., “Theory of Secondary Electron Emission by C.P. Pike and R.R. Lovell, AFGL TR-77-005 l~ASA
by High Speed Ions,” Phys. Res., 108: 1–12, 1957. TMX-73537, ADA045459, 225-236, 1977.
Stevens, N. J., “Space Environment at Interactions with Biased Whipple, E. C., Jr., and L. W. Parker, “Theory of an E1ec-
Spacecraft Surfaces, “ in Space Systems and Their In- tron Trap on a Charged Spacecraft ,“ J. Geophys. Res.,
teractions With the Earth’s Space Environment, (Prog. 74: 2962–297 1, 1969a.
Astronaut. Aeronaut., 71) edited by H.B. Garrett and Whipple, E. C., Jr., and L. W. Parker, “Effects of Secondary
C.P. Pike 455-476, AIAA Press, New York, N. Y., Electron Emission on Electron Trap Measurements in
1980. the Magnetosphere and Solar Wind,” J. Geophys. Res.,
Stevens, N. J., and C.K. Purvis, “NASCAP Modeling Com- 74:5763, 1969b.
putations on Large Optics Spacecraft in Geosynchronous Whipple, E. C., Jr., J.M. Wamock, and R.H. Winkler,
Substorm Environments,” NASA TM-81395, 1980. “Effect of Satellite Potential on Direct Ion Density Mea-
Stevens, N. J., J.V. Staskus, J.C. Roche, and P.F. Mizera, surement through the Plasmapause ,“ J. Geophys. Res.,
“Initial Comparison of SSPM Ground Test Results and 79: 179-186, 1974.
Flight Data to NASCAP Simulations,” NASA TM-81394, Wildman, P.J .L., ‘&Dynamics of a Rigid Body in the Space
1980a. Plasma,” in Space Systems and Their Interactions With
Stevens, N. J., J. V. Staskus, and J .C. Roche, “Initial Com- Earth’s Space Environment, (Prog. Astronaut. Aeron-
parison of SSPM Ground Test Results and Flight Data aut., 71) edited by H.B. Garrett and C.P. Pike, 633-661,
to NASCAP Simulations,” NASA TM-81314, 1980c. AIAA Press, New York, 1980.
Sub, P. K., and M.C. Stauber, “Photoelectron Drift and Williamson, P. R., W. Denig, P.M. Banks, and W.J. Raitt,
Multiplication due to Surface Potential Gradient; Ap- “Vehicle Potential Measurements Using a Mother-
plication to Solar Power Array s,” Paper 80-0044, AIAA Daughter Tethered Rocket (Abstract),” Eos Trans. AGU,
18th Aerospace Sciences Conference, AIAA, New York, 61: 1068, 1981.
1980. Willis, R. F., and D. K. Skinner, “Secondary Electron Emis-
Taylor, H. A., Jr., H.C. Brinton, and C.R. Smith, “Positive sion Yield Behaviors of Polymers,” Solid State Com-
Ion Composition in the Magnetionosphere Obtained from mun., 13:685, 1973.
the OGO-A Satellite,” J. Geophys. Res., 70:5769-5781, Winckler, J. R., “A Summary of Recent Results Under the
1965. ‘Echo’ Program for the Study of the Magnetosphere by
Treadway, M. J., C,E. Mallon, T.M. Flanagan, R. Denson, Artificial Electron Beams,” Cosmic Phys. Tech. Report
and E.P. Wenaas, “The Effects of High-Energy Elec- 168, School of Physics and Astronomy, University of
trons on the Charging of Spacecraft Dielectrics ,“ IEEE Minnesota, Minneapolis, September 1976.
Trans. Nut. Sci., NS-26, (6), 5102-5106, 1979. Winckler, J. R., “The Application of Artificial Electron Beams
Tsien, H. S., “Superaerodynamics, Mechanics of Rarefied to Magnetospheric Research,” Rev. Geophys. Space
Gases,” J. Aero. Sci., 13, 653–664, 1946. Phys., 18:659-682, 1980.
Tsipouras, P., and H .B. Garrett, “Spacecraft Charging Model Zonov, Yu, V., “On the Problem of the Infraction Between
2 Maxwellian Approximation,” AFGL TR-79-0153, a Satellite and the Earth’s Atmosphere and the Earth’s
ADA077907, 1979. Magnetic Field,” Iskusstvenyve Sputniki Zemli, 3, 1959,
Walker, E. H., “Plasma Sheath and Screening Around a (NASA Tech. Transl. TTF-37).

7-37
 Chapter 8

MAGNETOSPHERIC AND HIGH LATITUDE

IONOSPHERIC ELECTRODYNAMICS
W.J. Burke, D.A. Hardy, and R.P. Vancour

This chapter deals with the volume of space that is In dealing with the earth’s magnetosphere three things
bounded externally by the magnetopause and internally by quickly impress the mind. First, there is almost nothing
the plasmapause and the high latitude ionosphere at an al- there. Particle densities in the plasma sheet range up to about
titude of 300 km. The magnetopause separates regions of l/cm3. With present technology, laboratory vacuum systems
space dominated by the earth’s magnetic field (magneto- are able to get down to densities of 1010/cm3. Second, the
sphere) and by the shocked solar wind (magnetosheath). volume of space occupied by the magnetosphere is consid-
Earthward of the plasmapause, dynamics are generally con- erable. Typical magnetospheric dimensions are of the order
trolled by coronation rather than by solar wind driven con- of 10 R~ ( 1 R~ = 6.4 x 103 km). Third, when concen-
vection. The arbitrarily chosen, low altitude boundary in trated to global scales the effects of magnetospheric pro-
the ionosphere represents a transition below which the ef- cesses are impressive. The third point which is illustrated
fects of the earth’s neutral atmosphere are dominant. From in Figure 8-1 provides a convenient point of departure for
the viewpoints of both cause and effect, the chapter is some- this survey of magnetospheric and high-latitude ionospheric
thing less than self contained. Without the geomagnetic field processes. The figure exemplifies the spatial distribution of
(Chapter 4) and the solar wind (Chapter 3), there would be visible radiation observed by means of an optical imaging
no magnetosphere and no magnetospheric electrodynamics; system on a Defense Meteorological Satellite Program
without solar irradiance (Chapter 2), there would be much (DMSP) satellite. City lights provide an easily recognized
less of an ionosphere. Without magnetospheric electrodyn- map of the western half of North America. The total energy
amics, there would be no aurora (Chapter 12), no high- emitted by auroral fotms over the northern tier of Canadian
latitude currents (Chapter 4), no radiation belts (Chapter 5), provinces rivals or exceeds the combined ground emissions
and no problems with spacecraft charging (Chapter 7). from the United States and Canada. Auroral emissions are
The ,term “electrodynamics” encompasses a complex of largely due to plasma sheet electrons with energies of a few
processes by which charged particles move about in the keV impacting the E layer of the ionosphere. The instan-
magnetosphere-ionosphere system. The nature of the pro- taneous locus of plasma sheet electron precipitation is called
cesses varies from region to region within the system. Mag- the auroral oval. Global imagery from satellites such as
netic merging at the magnetopause and field-aligned poten- DMSP have shown that the auroral oval may be approxi-
tial drops above the aurora] ionosphere are examples of mated by circular bands surrounding the geomagnetic poles.
localized, electrodynamics processes. They are unified as The centers of the circles are offset by -3° to the night
electrodynamics processes in that they emerge, with appro- sides of the magnetic poles. The radii of the circle, the
priate boundary conditions, as solutions of the Vlasov-Max- widths of the bands and the intensity of emissions vary with
well equations. General solutions of the Vlasov-Maxwell the level of geomagnetic activity. However, at all times the
equations over the entire magnetosphere-ionosphere system auroral oval exists and acts as a major sink for magneto-
are well beyond present capabilities. Some success, how- spheric particles ahd energy. The particles and energy lost
ever, has been achieved by considering elements of the by the magnetosphere due to aurora] precipitation ultimately
system in relative isolation. This provides insight into how come from the solar wind. Thus, an estimate of global
system elements evolve in response to external inputs. Since precipitation loss is also an estimate of the efficiency of
the entire system is electrically coupled, the isolated element solar wind/magnetospheric interactions required to maintain
approach is self-limiting. As one element evolves it affects the auroras.
processes in other elements of the system. The main goals During periods of moderate geomagnetic activity the
of this chapter are to describe the various system elements auroral oval can be approximated as a circular band ex-
and indicate, in a qualitative sense, how they are electrically tending from 75° to 65° magnetic latitude. The area of such
coupled. a band is 1017cm2. The mean flux of electrons into the

8-1
CHAPTER 8
8.1 MAGNETOSPHERIC BOUNDARY
INTERACTIONS

In describing interactions between the solar wind and

the earth’s magnetosphere two coordinate systems are use-
fully employed: geocentric solar-ecliptic (GSE) and solar
magnetospheric (GSM) coordinates. Both coordinate sys-
tems, which we described in detail in Chapter 4, have their
origins at the center of the earth with the X axes positive
toward the center of the sun; that is, X,= = X,~. The Z,.
axis is normal to the ecliptic plane and positive toward the
north. me Y,. that completes the right hand system is pos-
itive toward local dusk. The Zs~ axis is coplanar with the
earth’s magnetic moment vector (M) and the X,~ axis. It
is positive toward ecliptic north. The Y,~ axis, which always
lies in the SM equatorial plane, completes the right hand
coordinate system. For a radially flowing solar wind
(v, = – V,xw= – V,xsm,where X= and X,~ are unit
vectors along X,e and X,~, respectively) the angle between
M and Z,~ gives the magnetic latitude of the magnetospheric
subsolar point. Note that due to 110 offset between M and
the earths rotational axis and to the 23.5° angle between the
equatorial and ecliptic planes the magnetic latitude of the
subsolar point is subject to A 34.5° combined seasonal and
diurnal variations. The GSM is superior to the GSE system
for ordering data relevant to interactions between the solar
wind and the magnetosphere.

8.1.1 The Magrketopause

The shape of the “steady state” magnetosphere is de-

termined from the force balance equation

v“[p+~]=o
— (8.1)

where ~ and ~ are the total pressure and the Maxwell stress
tensors; respectively. we total pressure tensor is made up
Figure 8-1. DMPS imagery from over western North America. of two parts due to the dynamic and thermal pressures of
the solar wind components:

(8.1)

auroral oval is - 109 cm-2s-1. Thus under steady state con- where n, is the solar wind density, mp the mass of a proton;
ditions the solar wind must supply electrons to the mag- psi and p= are the t~crrnal pressures of solar wind ions and
netosphere at a rate of 1026/s. The average energy of pre- electrons, respectively. The factor of 2 accounts for specular
cipitating electrons is of the order of 1 keV. The energy reflection of incoming particles. The shape of the magne-
loss due to electron precipitation alone is -1010 W. Similar topause on the dayside can be calculated by numerical means
or larger amounts of solar wind energy must be supplied to using a simplified force balance
account for ionospheric Joule heating and for maintaining
the ring current. The central focus of this chapter is to outline BT2
2n,mPV,2 ( – X,. o fi~)2 = — (8.3)
the present understanding of how 1026 electrons per second 2 ~o’
and tens of billions of watts are extracted from the solar
wind to drive magnetospheric and ionospheric electrody- where PO is the perrnitivit y of free space, fiM is an outward
namics processes. directed unit vector normal to the magnetopause, and BT

8-2
 MAGNETOSPHERIC AND HIGH LATITUDE IONOSPHERIC ELECTRODYNAMICS
the total magnetic field at the magnetopause. BT is a su- the calculated total field strikingly illustrates the effects of
perposition of fields due to the earth’s dipole B~, to the the solar wind on the overall magnetic topology. Magnetic
currents flowing on the magnetopause BM, and to other field lines on the dayside are compressed while those on the
currents distributed in the magnetosphere. Beyond the mag- nightside are elongated. Note that in this m,odel, field lines
netopause BM exactly cancels the internal fields. To a very intersecting the earth at magnetic latitudes greater than 83°
good approximation at the subsolar point of the magneto- are swept back to the nightside by the solar wind. There
pause are a pair of singular points on the magnetopause at separa-
trices between field lines closing on the day and night sides.
lB~l = 21fiM X B~l. (8.4) These points correspond to the dayside cusps.
The models just discussed do not describe the nightside
In the magnetic equatorial plane of the magnetotail adequately, One reason is apparent from
a consideration of Equation (8.2). On the dayside of the
magnetosphere, the dynamic pressure of the solar wind dom-
(8.5)
inates over the thermal pressures. On the nightside, with
plasma flow almost tangential to the magnetopause
where BO = 3 X 104 nT is the strength of the earth’s field (V, “ ftM = O), the converse is true. An early model [John-
at the surface on the magnetic equator and L is the distance son, 1950] of the magnetosphere had a tear drop shape with
from the center of the earth in earth radii (R~). Substitution the closing distance determined by the solar wind Mach
of Equations (8.4) and (8.5) into (8,3) gives the distance to number. Piddington [1963] suggested that, in flowing past
the magnetopause near the subsolar point the magnetosphere, the solar wind exerts tangential stresses
at the boundary. Such stresses draw the nightside of the
magnetosphere into an elongated magnetotail. In the ab-
(8.6)
‘M= (Po:;vs,)’”
sence of significant plasma within the magnetotail, the tan-
gential force exerted by the solar wind on the magnetosphere
For a solar wind density and velocity of 5/cm3 and 400 is
km/s, L~ = 9. The shape of the dayside magnetopause was
calculated by Mead and Beard [1964] and by Olson [1969] (8.7)
using iterative numerical techniques in which the tilt of the
dipole was ignored and included, respectively. Figure 8-2
shows a meridional cross section of the magnetosphere cal- where BMT and RMT are the field strength and radius of the
culated with LM = 10 in the Mead and Beard model. The magnetotail, respective] y. There are, however, distributed
locus of dipole field lines (dashed lines) in comparison with currents in the inner magnetosphere (the ring current) and
in the magnetotail (the neutral sheet) currents, whose effects
must be included in realistic stress calculations.

8.1.2 Convection

In many cases, the magnetosphere-solar wind interaction

is well described by steady state equations such as Equation
(8. 1). The equilibrium represented by these equations is,
however, dynamic rather than static. Only a dynamic sit-
uation is consistent with existing high latitude current sys-
tems. These currents result from ionospheric convection
which is driven by magnetosphenc convection [Gold, 1959].
Magnetospheric convection is in turn driven by the solar
wind. That is, energy is extracted from the solar wind by
the magnetosphere, and at least some of that energy is dis-
sipated in the ionosphere. Two mechanisms for transferring
energy to the magnetosphere have been developed over the
last two decades: viscous interactions [Axford and Hines,
196 1] and magnetic merging [Dungey, 1961]. Both models
explain many qualitative features of magnetospheric con-
Figure 8-2. Model of earth’s magnetic field distorted by the solar wind vection and auroral particle energization. Recent satellite
[Mead and Beard, 1964]. observations suggest that both mechanisms are operative,

8-3
CHAPTER 8
but in more complex ways than envisaged by early pro- Consider a magnetic field that at great distances above (be-
ponents. low) the X-Y plane points in the + ( – ) X direction. In the
The Axford-Hines model postulates that the magneto- presence of an electric field EY, magnetic field lines convect
sheath plasma exerts a viscous force on a layer of unspecified toward the X-Y plane. At the neutral line (X = O, Z = O),
thickness inside the magnetopause. Magnetic field lines magnetic field lines from the upper half space merge with
threading this layer are dragged in the antisolar direction field lines of the opposite polarity from the lower half space.
and are stretched to great distances in the magnetotail. As To the left (right) of the neutral line, merged magnetic field
elongated flux tubes move out of the viscous interaction lines cross the X-Y plane with a + ( – ) Z component, and
layer they snap back to a more dipolar configuration. In the E x B convect away from the neutral line in the + ( – )
rest frame of the earth this motion of magnetic field lines direction. Two necessary conditions for magnetic merging
appears as an electric field, E = – V X B. A magneto- are magnetic field of opposite polarity across some plane
spheric equatorial projection of the convection pattern gen- and an electric field component that is tangent to the plane.
erated in the viscous interaction model is given in Figure Before considering how magnetic merging might apply
8-3. When mapped to ionospheric altitudes, assuming that to the magnetosphere, it is useful to distinguish between
several possible magnetic topologies. It is well known that
a weak interplanetary magnetic field (IMF) is carried by the
solar wind. Except for a small correction term in the force
balance equation, the IMF plays no obvious role in the
viscous interaction model. The magnetic merging model
assigns important roles to the IMF because this model re-
quires three types of magnetic field lines: (1) IMF lines with
both “feet” in the interplanetary medium, (2) closed field
lines with both “feet” in the earth, and (3) open field lines
with one “foot” on earth and the other in the solar wind.
Dungey [1961] pointed out that when the IMF has a south-
ward component, magnetic merging can occur near the sub-
Figure 8-3, Equatorial projection of convection pattern in viscous inter- solar point of the magnetopause. The idea is illustrated in
action model.
Figure 8-5 which can be viewed either as a snap shot or as

E “ B = O, the model reproduces the general features of

the poladauroral current system (Chapter 4). Note that plasma
trapped on elongated flux tubes is adiabatically heated as
the flux tubes convect earthward and shrink in volume.
The second model postulates that the dynamic interac-
tion between the solar wind and the magnetosphere proceeds
by means of a magnetic merging process. The simplest
features of this phenomenon are illustrated in Figure 8-4.

z -b
4
I

Figure 8-5. snapshot of magnetic merging between southward IMF and

the earth’s magnetosphere.

a time history of an individual field line. As southward

directed IMF lines are convected up against compressed
dipolar field lines, merging occurs at time 0). Because one
foot of a newly merged field line is imbedded in the solar
wind, the whole field line is dragged in the antisolar ( – X,J
direction. In an earth-stationary frame of reference the mo-
tion of the ionospheric foot of the field line appears to result
F]gure 8-4. Magnetic field geometry and electric field required for mag- from a dawn to dusk electric field. Times @ through @
netic merging. show the various stages of antisunward motion of an open

8-4
 MAGNETOSPHERIC AND HIGH LATITUDE IONOSPHERIC ELECTRODYNAMICS
field line. At time @, a portion of the field line has con-
vected to the magnetic equatorial plane where it reconnects
with an open field line from the conjugate ionosphere. Un-
der the influence of the dawn to dusk electric field, the
field line then convects earthward (times @ through @).
Eventually, reconnected field lines move to the dayside
(time 0) where they are in position to continue the merging-
reconnection cycle.
The magnetospheric convection patterns predicted by :.,.:j ,.,,::.:,.
.,.,,,.,,..,
-.. .- —-
:!;:!!:::,,:’
/,,; ,,,,,, ; r7
the viscous interaction and magnetic merging models with PLASMA INTER IR LOW-LATITUDE PLASMA RING CURRENT
MANTLE CUSP (EL) BOUNDARY BWO N DARY mASMA SHEET
a southward IMF are quite similar. Empirical evidence from \ LAYER, LAYER
MA~ETOPAUSE BOUNDARY LAYERS
near the magnetopause indicates that both viscous and mag-
netic merging processes occur [Eastman et al., 1976; Russell
and Elphic, 1979; Mozer et al., 1979]. That many obser- Figure 8-6. (a) Schematic diagram of various observed magnetospheric
boundary layers; (b) their mapping down to the ionosphere
vational studies have shown high correlations between along magnetic field lines [Vasyliunas, 1979].
southward turnings of the IMF and the onset of magnetic
activity indicates a dominant role for energy transfer by a
The plasma mantle was first identified as a magneto-
merging process. As discussed in Section 8.5.1, this surmise
sheath-like plasma flowing along magnetic field lines inside
is supported by modifications observed in polar cap (open
the magnetopause in the near-earth lobes of the magnetotail
field line) convection patterns with variations in the YsM
[Rosenbauer et al., 1975]. The plasma density and the spa-
component of the IMF and when the ZsM component is
tial thickness of the mantle are greatest during periods of
northward. Finally, we mention that cumently merging is
southward IMF [Sckopke et al., 1976]. The mantle is also
thought to occur sporadically rather than as a steady state
observed in the lobes of the magnetotail at lunar distance
process [Haerendel et al., 1978], and in the vicinity of the
(XsM = – 60 R,) near the ecliptic plane [Hardy et al.,
dayside cusps rather than near the subsolar point [Crooker,
1975]. Mantle particles are believed to enter the magneto-
1977; Crooker, 1979].
sphere near the dayside cusp. Dawn to dusk electric fields
cause particles to convect in the antisunward direction so
that particles that mirror at low altitudes find themselves on
8.1.3 Boundary Layers
open field lines as they rise from their mirror points. The
same dawn to dusk electric field causes the mantle particles
Before considering the dynamics of the closed portion
to drift toward the equatorial plane as they move away from
of the magnetosphere, let us return briefly to our original
the earth. Pilipp and Morfill [1978] suggested that mantle
question of how 1026 particles/s gain entry to the plasma
particles may be one source of plasma sheet particles.
sheet. Between 1028 and 1029 solar wind particles/s impact
As the name suggests, the “interior cusp” refers to the
the dayside magnetopause. Thus, an entry efficiency of less
region of closed magnetic field lines passing through or just
than 1Yo is sufficient to maintain the plasma sheet. Our
equatorward of the cusp. Magnetosheath plasma diffuses
understanding of how magnetosheath plasma gains entry to
into this region. As opposed to the mantle or the LLBL,
the magnetosphere and influences its interior dynamics has
the plasma of this region, which is also called the entry
been evolving rapidly over the last decade. Under such
layer, is relatively stagnant. However, depending on the
circumstances, it is not unusual to encounter a multiplicity
strength and direction of convective electric fields in the
of nomenclatures that will probably be simplified as rela-
cusp, it is possible to think of the interior cusp plasma as
tionships between various boundary plasma regimes become
a partial source of both the mantle and the LLBL.
more evident. Vasyliunas [1979] has defined magneto-
Along the dawn and dusk meridians near the magne-
spheric boundary layers as regions of space threaded by
magnetic field lines of the magnetosphere but populated by
plasma similar to that found in the magnetosheath. The four
regions satisfying this definition are (1) the plasma mantle, —~
(2) the interior cusp, (3) the low latitude boundary layer
(LLBL), and (4) the plasma sheet boundary layer. Figure ENTRY LAYER

8-6a is a schematic representation of the magnetospheric

loci of these regimes. Somewhat speculative representations
of their ionospheric projections and their “source-relation-
~mw<ATiTuDE BOUNDARY LAYER =
ships” to the plasma sheet are given in Figures 8-6b and — ——

8-7, respectively. The plasma mantle is found on open mag-

netic field lines; the remaining three regions are found in Figure 8-7. A model of magnetosphenc circulation for filling the plasma
closed field line portions of the magnetosphere. sheet [Freeman, 1979].

8-5
CHAPTER 8
tospheric equatorial plane, the LLBL is characterized by Magnetopause
. —
magnetosheath-like plasma flowing in the antisunward di-
rection. The thickness of this layer ranges up to -1 RE. As
. .. . . . . . . . . . . . . . . . . . . . . . . . .
shown in Figure 8-6a, the LLBL has been observed to great =::... ..... . . . . .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. .. ..
distances in the antisolar direction. The density of plasma . . . . . . . . . . . . . . . . . . .. .. .. .
......................
......................
within the layer is about a factor of four less than that of .......................
.......................
the adjacent magnetosheath [Sckopke et al., 1981]. Elec- .......................
.......................
..................
trons within the layer have trapped pitch angle distributions . . . . . . . . . . . . . . . . . . .. .. .. .. ..
..................
. . . . . . . . . . . . . . . . . . .. .. .. .. .. .
[Eastman and Hones Jr., 1979]. Whether the LLBL is main- ........................
........................
tained through a diffusive and/or an impulsive [Lemaire and ..................
To . .. . . . . . . . . . . . . . . . . . . . .. .. .. .. ..
\. . . . . ......
Roth, 1978] entry process is currently a matter of debate. Sun \ Inner Boundary “
Sckopke et al. [1981] estimate that a diffusion coefficient ‘e \ + of PI asmb Sheet
Illv ):: . . . ... . . .. .. . .... .... ..
of 109m2/s is required to maintain the observed LLBL. Fig-
ure 8-6a shows the plasma boundary layer together with the .:....\....\\\ xfi:::!’~la~rna::::::::;’ .....
.....
LLBL as forming a continuous envelope surrounding the
hot plasma contained in the central plasma sheet. The phys- \4 ...............................
.... efi::::::: Sheet
.............................
.... .............

.........................
.........
.........
.........
ical processes which connect the LLBL and/or the mantle
with the plasma boundary layer and with the central plasma k::::::::::::::::::::::
1
.
. . . . . .
—..
. . . . . . . . .
.
. .

sheet are not known at this time. It is currently believed (Dusk)

that discrete arcs in the auroral oval map to the boundary
rather than to the central plasma sheet. Within the plasma Figure 8-8. The distribution of plasma sheet elwtrons in the equatorial plane
sheet boundary, rapidly flowing plasmas are observed. These [Vasyliunas, 1972].
flowing plasmas come from spatially limited acceleration
regions called “magnetosphenc fireballs” [Frank et al., 1976].
the dawn and dusk flanks of the tail. At lunar distance
Whether the energization process in fireballs results from
(XsM = – 60 R~) the tail radius is -25 R~. The average
magnetic reconnection or some other process is still another
half-thickness of the mid-tail plasma sheet is -3 R~. During
open question.
the expansion phase of substorms, the thickness of the plasma
sheet in the tail decreases and then expands during the re-
covery phase [Hones et al., 1973]. At lunar distance the
8.2 THE PLASMA SHEET average density of the plasma sheet is -0.1 cm3. The elec-
tron and proton temperatures are -0.25 and 2.5 keV, re-
The earth’s plasma sheet is the highly dynamical region
spectively [Rich et al., 1973].
of the earth’s magnetosphere that acts as a depository for
The remainder of this subsection is concerned with the
auroral particles. It is a region of closed magnetic field lines.
physical mechanisms responsible for particle energization
Before being first detected by the Soviet satellites Luna I
and precipitation in the plasma sheet. Energization processes
and Luna 2 [Gringanz et al., 1961], somewhat strangely,
are classified as either adiabatic or non-adiabatic. Examples
this important region of the magnetosphere was not antic-
of non-adiabatic energization are neutral sheet acceleration
ipated theoretically. Equatorial and noon-midnight merid-
[Speiser, 1967], stochastic wave-particle acceleration, and
ional projections of the plasma sheet are given in Figures
heating derived from magnetic field reconnection and an-
8-8 and 8-9, respectively [Vasyliunas, 1972]. Both projec-
nihilation. Although these are undoubtedly important sources
tions show that the plasma sheet extends for great distances
of particle energy, we limit ourselves to describing adiabatic
in the – X~Mdirection. The plasma sheet has a distinc~ inner
energization in some detail. Particle precipitation is main-
edge that varies as a function of local time and the level of
geomagnetic activity [Vasyliunas, 1968; Frank, 1971]. The
dynamics of the inner edge of the plasma sheet are well
understood theoretically and are discussed in Section 8.6.2.
The equatorial thickness of this boundary is -1 R~ and is
OGOI
marked by a cooling of electron temperatures. Just tailward
of this boundary plasma sheet electrons have an average
energy of -1 keV. During periods of substorm injections,
the temperature of electrons may rise to -10 keV. Tem-
peratures of plasma sheet ions tend to be higher than those
of electrons by a factor of 2 or more. ‘Electron 2 Vela
Figure 8-10 is a cross sectional view of the magnetotail
portion of the plasma sheet. It has a minimum thickness in Figure 8-9. The distribution of nightside, plasma sheet electrons in mer-
the mid-tail region and flares to a maximum thickness near idional plane [Vasyliunas, 1968],

8-6
 a
MAGNETOSPHERIC AND HIGH LATITUDE IONOSPHERIC ELECTRODYNAMICS
where pll and ds are momentum component and distance
hem lobe
along B, and
(3) the flux invariant

-Plasma sheet
where A is the magnetic vector potential and d{ a distance
element along a particle drift trajectory [Northrop, 1963;
Rossi and Olbert, 1970]. Most plasma sheet particles either
precipitate or follow drift trajectories that intersect the mag-
Magnetopause2 S ‘Southern lobe netopause before they can drift all the way around the earth.
Thus, in the plasma sheet only the first and second invariants
are of interest. The third invariant is important for under-
Figure 8-10. Cross-sectional viewofmagnetotail looking eafihward
standing the ring current and radiation belts (Chapter 5).
At this point it is useful to introduce the related concepts
tained or enhanced either by magnetic-field aligned electric of pitch angle and magnetic mirroring. The pitch angle of
fields (Ell) or by pitch angle diffusion. Ell is very important a particle (a) is defined as the angle between its instanta-
for discrete aurora] arc formation. Pitch angle diffusion re- neous velocity and the magnetic field
sults from wave particle interactions. Here we summarize
v*B
()
briefly the collective plasma modes responsible for these
a=cos–’ —
phenomena. Ivl IBI

Magnetic mirroring results from the constancy of a particle’s

8.2.1 Adiabatic Motion in the Plasma Sheet magnetic moment and total energy. The total energy of a
non-relativistic particle of mass m and charge q moving
The adiabatic energization of plasma sheet particles in-
with a velocity v in combined magnetic and electric fields
volves the related concepts of adiabatic invariance and guid-
is
ing center motion. The general concept of adiabatic invar-
iance comes from classical Hamilton-Jacobi theory [TerHaar,
e=l/2mv112+p B+q W (8.11)
1964]. If a system executes a periodic motion in a force
field (F) that changes slowly in time with respect to the
where W is the electrical potential. The component of force
period (T)
exerted along B is

dvll dB 8*
(8. 12)
‘x = –F%–qz”
then the quantity
The second term on the right hand side of Equation (8. 12)
I=~pdq, is due to field-aligned electric field components that are
discussed regarding aurora] arc formation. The first term on
where p and q are canonical momentum and coordinate the right hand side of Equation (8. 12) is the magnetic mirror
variables, is a constant of the motion known as an adiabatic force. A particle at the magnetic equator (s = O) with pitch
invariant. Charged particles moving in the earth’s magnetic ang]ecieq can move earthward along B until its pitch angle
field may have as many as three periodicities due to their reaches 90° (VII = v...~ = O). At this point, it is reflected
gyration, bounce, and drift motions. The three adiabatic by the mirror force toward the magnetic equator. The strength
invariants associated with these periodicities are of the magnetic field at the mirror point is designated BM.
(1) the magnetic moment The total kinetic energy of a particle at its mirror point is
l/2mv2 = PBM. Since in general p = 1/2mv2 (sin2m)/B,
mv12 the magnetic mirroring condition, in the absence of En, is
(8.8) often written
‘= 2B’

where VL is the component of velocity perpendicular to the B

(8.13)
magnetic field, ‘in2fi = ~“
(2) the longitudinal invariant
The line integration for the longitudinal invariant proceeds
J = $ pll ds, (8.9) from the magnetic equator to the mimer distance (s~).

8-7
CHAPTER 8
The notion of guiding center motion is more general With the expressions for the adiabatic invariants [Equa-
than that of adiabatic motion. Under many circumstances tions (8. 8) and (8.9)] and guiding center drifts [Equations
charged particle motions are well-approximated superpo- (8. 16), (8. 17), (8. 18)], we may understand(1) the adiabatic
sitions of motions of guiding centers and gyrational motion heating of magnetospheric plasma, (2). the existence of a
about the guiding center. For example, in an uniform mag- ring current, and (3) the position of the inner boundary of
netic field the equation of motion the plasma sheet.
Both the viscous interaction and the magnetic merging
models require that particles in the nightside plasma sheet
m~=q(vx B) (8. 14) be “adiabatically” heated as they convect earthward under
the influence of a dawn to dusk electric field. In this motion,
each particle’s magnetic moment l/2mvL 2/B is a constant.
describes a particle that gyrates with a circular frequency
Consider a particle mirroring in the equatorial plane, J = O.
Q = qB/m about a field line and moves along the field line
As it drifts from, say, 30 R~ in the magnetotail where B -20
(its guiding center) with a constant velocity. In the presence
nT to a distance of 6 R~ where B - 140 nT, its kinetic
of an external force field F that is perpendicular to B, the
energy must increase seven fold. Particles with J # O gain
equation of motion
kinetic energy, not only from the conservation of p, but
also from the apparent motion of magnetic mirror points
m$=q(vx B)+F (8.15) (conservation of J). In moving from equatorial crossings of
30 R~ to 6 R~, particles find themselves on shorter and
shorter field lines. Since $pllds must be a constant, pll must
can be reduced to the form of Equation (8. 14) by trans- increase as particles convect earthward.
forming to a coordinate system moving with a drift velocity From the conservation of energy, it is seen that an in-
crease in particle kinetic energy must be due to a decrease
FxB in potential energy. The potential energy is electrical; that
vD=— is,
qB2 “

In this frame of reference the motion of the particle about ~(1/2mv2)=q(V~. E). (8.19)
B is purely gyrational.
The most important drifts within the magnetosphere are
caused by electric fields The brackets in equation (8. 19) are used to represent time
averaging over a gyroperiod, and V~ is the total drift ve-
locity. Since V~ is perpendicular to E, only V~ and Vc
ExB
vE=— (8.16) contribute. Hines [ 1963] has shown that the energy gained
B2 ‘
by “adiabatically compressing” a magnetospheric plasma is
equivalent to the kinetic energy gained by gradient and
magnetic field gradients curvature drifting in the direction of an electrostatic potential
gradient.
pB X VB Guiding center motion is the simplest basis for under-
VG= (8.17)
standing the earth’s ring current. The existence of a west-
qB2 ‘
ward current encircling the earth can be inferred directly
and magnetic field line curvature from decreases in surface values of the horizontal compo-
nent of the earth’s field during the main and recovery phases
mv112R x B of magnetic storms (Chapter 4). The general expression for
v~=— — current density is given by a sum over plasma species
R* qB2 (8.18)
Bx(B. V)B
= mvllz
qB4

where R is the magnetic field line radius-of-curvature vec- Assuming, for simplicity, that the ring current is made up
tor. Note that V~ and Vc depend on both the particle’s of a single ion species (H+) then the current density is
energy and charge; V~ depends on neither. Protons (elec-
trons) gradient and curvature drift toward the west (east). j. = nq (vDi – Vh) (8.20)
In the nearly dipolar part of the inner plasma sheet, V~ and
Vc are of comparable magnitudes. Due to sharp magnetic Since V~ is independent of charge, only V~ and Vc con-
field line curvature across the neutral sheet, Vc > V~ in tribute to Equation (8.20). Recall that for protons (elec-
the magnetotail. trons), V~ and VC are westwards (eastwards). In an equiv-

8-8
 m
MAGNETOSPHERIC AND HIGH LATITUDE IONOSPHERIC ELECTRODYNAMICS
aient fluid description, j~ is driven by magnetospheric pressure equatorial plane of the magnetosphere, the form of the elec-
gradients. During the main phases of magnetic storms and tric potential is
the expansion phases of substorms, particles are energized
and injected into the inner magnetosphere by intense electric – QBOR:
W(R, +)= R + C RYsin @ (8.21)
fields. With the onset of recovery phases the electric fields
decrease in intensity andlor are shielded from the inner
magnetosphere. Injected particles find themselves on closed, where Q = 7.3 x 105s-’ is the angular spin velocity of
stably trapped orbits in which they gradient and curvature the earth, C is a constant to be determined, and y is a
drift around the earth. During recovery, phase these particles parameter that reflects the level of electrical shielding. The
are slowly removed from the ring current by precipitation case y = 1 corresponds to a uniform, dawn to dusk Ec that
or by charge exchange with iow energy neutrals (Chapter completely penetrates the inner magnetosphere. Best em-
5). pirical values of y are in the range of 2 to 3 [Ejiri et al.,
The position and shape of the inner edge of the plasma 1978; Gussenhoven et al., 1981]. The value of C is deter-
sheet is determined by the drift motions of plasma sheet mined by noting that EQ is directed radially inward and Ec
particles. V~ decomposes into drifts due to “convective” is mostly the YsM direction. In the dusk sector En and Ec
(V~c) and “corotational” (V~Q) electric fields. In the fol- are oppositely directed. Depending on the strength of Ec,
lowing discussion, we use the symbols & to represent the there is a stagnation point along the 1800 LT (+ = 31r/2)
“convective” electric field imposed by the solar wind on the axis of symmetry where the wo fields exactly cancel,
magnetosphere and E. to represent the coronation electric
field. The direction of (V~Q) is eastward for all particles.
For simplicity let us consider the drift motions of charged
particles having pitch angles in the equatorial plane of 90°.
The conservation of energy Equation (8.11) immediately = o. (8.22)
tells us that cold (p = O) particles are constrained to E X B R=Rs
drift along equipotentials. Particles with non-zero P drift
along surfaces of constant (T + ~B/q).
In the magnetotail, particles predominately drift earth- Substitution of Equation (8.21) into (8.22) gives
ward under the influence of a dawn to dusk E-field. As they
approach the earth, electrons acquire significant eastward
drifts due to both V~ and V~n. Since both of these drifts (8.23)
are eastward, cold electrons with VG = O drift closer to the
earth before their eastward drifts dominate over their earth- Thus ,
ward drifts. For this reason outbound satellites encounter
cold before hot electrons at the plasma sheet’s inner edge
[Kivelion et al., 1979].
The boundary between cold electrons that drift along
T(L, $)=–
*[ (:) -’(+))ysin’l
equipotentials from the tail and those that corotate on closed (8.24)
trajectories is called the zero-energy Alfv6n layer or the
inner boundary of the plasma sheet. Under steady convective
where we have made the substitution L = R/RE and
electric field conditions, cold electrons that drift in from the
magoetotail without precipitating eventually cross the day- L~ = R~/R~. The common term Q BOR~2 is -90 kV. The
side magnetopause. Thus, the inner bound~ of the plasma potential of the stagnation point is
sheet is the boundary between closed (coronation dominated)
and open (convection dominated) equipotentials.
Before calculating the shape of the last cIosed equipo-
tential, we note that for protons, V~ and VEQ are oppositely
W
()
Ls, ~ = -: [ 1 + l/y] kV. (8.25)

directed. This leads to more complex drift paths for protons Since along the @ = 31T/2 line EC < (>) E~I for L < (>)
than electrons. Protons with p # O can drift earthward of L~, Equation (8.25) gives the potential of the zero-energy
the zero-energy Alfv6n boundary in the evening local time Alfv6n boundary. By setting the term in brackets on the
sector. The different drift paths of protons and electrons right hand side of Equation (8.24) equal to [1 + I/y ], we
eventually lead to the build up of polarization electric fields arrive at the equation for the zero-energy Alfv6n bound~
near the inner edge of the plasma sheet. The main effects in the equatorial plane as a function of distance from the
of the polarization field are to shield Ec from the inner center of the earth (LA) and local time
magnetosphere and to distort its dawn to dusk orientation
in the plasma sheet.
As a function of distance R and local time $ in the

8-9
CHAPTER 8
Southwood and Kaye [ 1979] have shown that to an excellent tons into the aurora] ionosphere cannot be explained in terms
approximation of the individual particle model that we have been using.
The problem is illustrated simply in Figure 8-12 where we

‘%)
sketch sequential isocontours of particle ,distribution func-

“=’”(1 (8.27) tions in the magnetospheric equatorial plane. Figure 8-12a

where S = I cos (+ – Ir/2)/2 I . The solution is exact when , v,

A
(o)
Y = 1. Figure 8-11 shows that the shape of the last equi-

6
‘b)
V,l

“’XW7
Fiw 8-12. Isncontours of distribution functions with (a) isotropic and (b)
loss cone distributions.

12
5 represents an isotropic population at some initial time. Par-
ticles with a.q # 90° (vll # O) move along field lines to-
ward their mirror points. If their mirror points are suffi-
ciently deep in the atmosphere, typically altitudes < 200
km, the particles are lost. Otherwise they return to the
magnetosphere. If we designate the strength of the magnetic
field at an altitude of 200 km as B(200), then Equation
(8. 13) shows that particles initially with equatorial pitch
angles
18
I/2

[1
B
Figure 8-11. Shapes of last closed equipotential forvariousvaluesof y [Ejiri aeq < aL = sin–’ — (8.28)
et al., 1978]. B(2e&)

are lost after a few bounce periods. Such particles are said
potential varies from an elongated tear drop for y = 1 to to be in the atmospheric loss cone. For plasma sheet particles
a circle for y = CO.Note that Equations (8.26) or (8.27) aL is -2°. Figure 8- 12b shows the distribution after several
only allow calculations of the shape, but not the distance bounce periods as made up of the initial population minus
to the Alfv6n boundq. Ls depends on Ec, which varies an empty loss cone. Particle motion that conserves p and
with condition in the solar wind and with the level of mag- J allows no further loss.
netic activity. It is convenient to defer further comment on Kennell and Petschek [1966] pointed out that above
the Alfv4n bounda~ until we have discussed its ionospheric certain particle flux levels, loss cone distribution functions
projection, the equatorward boundary of diffuse auroral pre- such as shown in Figure 8- 12b are unstable to the growth
cipitation. of whistler waves. The waves grow in energy by causing
particles to diffuse into the loss cone. To produce strong
pitch angle diffusion, that is, maintain isotropy over the loss
8.2.2 Pitch Angle Diffusion of Plasma Sheet cone, a resonant condition must be fulfilled. Resonant scat-
Particles tering occurs for particles whose energy is equal to that of
the magnetic energy per particle
The final topic to be considered under the heading of
general magnetospheric processes is particle precipitation. B2
ER=== (8.29)
To anticipate our discussion of the ionosphere, we note that 2 p.n”
instrumentation on satellites passing through the diffuse au-
roral ionosphere measure fluxes of electrons and protons This model successfully explains flux limits observed for
that are isotropic over the downcoming hemisphere. The stably trapped ring current particles. The requirement for
continuous precipitation of plasma sheet electrons and pro- resonant pitch angle diffusion given in Equation (8.29) was

8-10
 w
MAGNETOSPHERIC AND HIGH LATITUDE IONOSPHERIC ELECTRODYNAMICS
empirically verified by observations of the proton ring cur- DRIVINL FIELD (OR C“B,t”T)

rent from Explorer 45 during the magnetic storm of De- t

cember 1971 [Williams and Lyons, 1974]. r 1 I I

IIAUI ETOS PHER 1C ,ONO>P”ERI,

An examination of the resonance condition given in ELECTRIC FIELD CENERALIZED EI.E,TE1< Frt Ll,

Equation (8.29) shows that a whistler mode instability can- 4

Om, s u.

not be responsible for strong pitch angle scattering in the

plasma sheet. At geostationary altitude (6.7 R~) in the plasma
sheet, B - 100 nT and n - 1 ems. This gives a magnetic
field energy density in eV/cm3 of 2.5 B2(nT). A resonant
energy of -25 keV is well in excess of mean thermal ener-
Y, LI.”-ALIcNE[)

gies for either electrons or protons in the plasma sheet. \ f

K,”E’rlc
cuRREm
EQUAT LM
Realizing that whistler mode interactions could not ex-
plain the isotropic precipitation of plasma sheet particles,
investigators in the 1970s concentrated on sources of elec-
trostatic wave energy. There are two important develop-
ments from the decade of which we take note. The first
concerns direct observations of broadband electrostatic noise
all along magnetic field lines connecting the auroral iono- PARTICLE PRESSURE ~
PERPENUICUUR

WENTW c“””ENT
sphere to the equatorial plasma sheet [Gumett and Frank, CONSERVATION 1
1977]. The amplitudes of observed waves are of sufficient
4
intensity to drive strong pitch angle scattering. The second BouNDARY
development concerns the theoretical recognition of the role SOURCE
played by cold plasma for making available free energy
contained in anisotropic pitch angle distributions. Cold plasma, Figure 8-13. Outline of self consistent calculation of magnetospheric con-
vection [Vasyliunas, 1970].
of ionospheric origin, in the plasma sheet can produce ve-
locity space gradients in the total distribution functions
( d f/ dvl > O) that are unstablp to the growth of (N + 1/2)
electron cyclotron waves [Young et al., 1973] and lower (a) First link: with knowledge of the electric field, we
hybrid ion waves [Ashour-Abdalla and Theme, 1978]. Nu- calculate the motion and distribution of protons and elec-
merical studies show that such waves are unstable over trons in the magnetosphere, and hence the total plasma
restricted ranges of parameter space. Measuring low den- pressure at any point.
sity, cold plasma embedded in a hot plasma sheet is ex- (b) Second link: from the plasma pressure gradients, we
perimentally difficult. A successful measurement of the cold calculate the components of the electric current perpendic-
plasma component is a critical requirement for our under- ular to the magnetic field. That is, from the force balance
standing plasma sheet processes. equation

Vp=jx B (8.30)

we calculate
8.3 ELECTRICAL COUPLING
OF THE MAGNETOSPHERE
BXVp
AND IONOSPHERE jl = (8.31)
B*
Before discussing the electrodynamics of the high lati-
tude ionosphere, it is useful to review briefly the theory of For simplicity we have assumed that the pressure is iso-
magnetosphere-ionosphere coupling. Vasyliunas [1970] has tropic.
developed a theoretical model that illustrates the physical laws (c) Third link: by calculating the divergence of the per-
describing how magnetospheric convention couples with the pendicular current and averaging over each flux tube, we
ionosphere. The model is presented in Figure 8-13 obtain Qll)the field-aligned currents (FAC) flowing between
in the form of a closed loop of equations (straight lines) and the magnetosphere and the ionosphere.
of quantities to be determined (boxes). External sources of
particles, cross-magnetospheric potentials, neutral winds in
(8.32)
the ionosphere, and so forth, are imposed boundary condi-
tions. Because the loop of equations is self-consistently closed,
it can be entered at any point. Let us assume that we have an (d) Fourth link: from the requirement that these field
initial idea about the distribution of magnetospheric electric aligned currents be closed by perpendicular Ohmic currents
fields and particles. in the ionosphere, we obtain the configuration of the electric

8-11
 CHAPTER 8
field in the ionosphere. The continuity of ionospheric current
requries that
(a)
V . I = jll SinX (8.33)

where I is the height integrated current and x the inclination

of magnetic field lines. In the ionosphere

1=2~. (E+ Vnx B) (8.34) DAWN

— DUSK

where Z —is the height integrated conductivity tensor and Vn

is the ~eutral wind velocity.
(e) Fifth and final link: the ionospheric electric field can
be map&d into the magnetosphere, and the requiremem that
it agree with the magnetospheric electric field assumed at
the outset determines the field, and thus closes the system
of equations. Except near discrete arcs, the mapping may
be done by assuming that magnetic field lines are equipo-
tentials. There is empirical evidence suggesting a functional (b)
relationship between jll in discrete arcs and field-aligned
DUSK N’P DAWN S’P DUSK
potential drops [Lyons et al., 1979]. We note in passing
Ex 1 ,,
that the Rice University group has successfully simulated // w
the ionospheric features of a magnetospheric substorm using I n ,/
this model [Hare] et al., 1981. ABY . 1
From Equations (8. 33) and (8. 34), it is clear that meas- UT +
urements of E and jll are critical for understanding the
magnetosphere-ionosphere circuit. Measurements of pre- Figure 8-14. (a) Field-aligned currents and electric fields as seen by polar
cipitating particle fluxes are needed to (1) understand spatial orbitting satellite in dawn-dusk meridian. (b) Idealized elec-
variation in E, (2) identify the dominant carriers of jll, (3) tric field and magnetic pe~urbation measurements.

calculate fiel~-aligned potential drops, and (4) help distin-

guish between topologically different regions.
centered coordinate system; X is positive along the satellite
velocity; ~ is positive toward local nadir; ~ completes the
8.4 HIGH LATITUDE ELECTRIC FIELDS right-hand system. In the dawn-dusk meridian Y is positive
in the antisunward direction. At high latitudes in the northern
Electric fields at ionospheric altitudes are measured from (southern) hemisphere we approximate B as being along the
potential differences between the ends of extended booms + ( – ) Z axis. Figure 8- 14a also shows dusk to dawn electric
on satellites [Fahleson, 1967] and from the E X B drifts fields in the auroral ovals and uniform dawn to dusk electric
of cold plasma [Hanson and Heel is, 1975]. In principle, fields across the polar caps. The convective electric field
jll can be determined from particle fluxes. With present tech- reverses directions near the poleward boundary of the au-
nology, full distribution functions of ions and electrons can- roral oval and goes to zero at the equatorward boundary.
not be measured with sufficient accuracy. For this reason, The governing equations are current continuity [Equa-
highly sensitive triaxial fluxgate magnetometers [Armstrong tion (8.33)], Ohm’s law [Equation (8. 34)] and the Maxwell
and Zmuda, 1973] are used to determine jll from magnetic equation
deflections. Particle fluxes in approximately the energy range
10 eV to 30 keV are measured by means of continuous VxB=~oj (8.35)
channel electron multipliers (channeltrons) placed behind
electrostatic energy-analyzers [Gold, 1959]. In the infinite current sheet approximation, these equations
It is useful to explain the format of high latitude data may be combined to eliminate jll and reduce to
presented in the following subsections by providing an ex-
ample of the simplest kinds of electric fields and magnetic
; [AB, – P.XPEX] = O (8.36)
field deflections expected to be measured by instrumentation
on a polar orbiting satellite. Figure 8-14a shows the trajec-
tory of a satellite in circular polar orbit with the ascending where ABY is the deflection of the magnetic field due to
node at the dusk (1800 LT) meridian. We define a satellite Jll and ~p iS the height-integrated Pedersen conductivity

8-12
 a
MAGNETOSPHERIC AND HIGH LATITUDE IONOSPHERIC ELECTRODYNAMICS
[Smiddy et al., 1980]. Except near local noon at the dayside The E field and FAC systems shown in Figure 8- 14b are of
cusp and near local midnight at the Harang discontinuity large scale.
where there are significant divergences of the ionospheric
Hall current [Rostoker, 1980], Equation (8.36) is a very
useful approximation. It tells us that for a uniformly con- 8.5 POLAR CAP ELECTRODYNAMICS
ducting ionosphere, fluctuations in the transverse magnetic
field component should track variations in the meridional This section treats three topics: ( I ) large-scale electric
component of the electric field. Deviations from correlated field patterns, (2) electron precipitation morphologies, and
variations are due to the presence of conductivity gradients. (3) characteristics of discrete, sun-aligned arcs in the polar
Figure 8- 14b is a plot of Ex and ABY expected over a full cap. Here we use the term “polar cap” to designate the
orbit assuming a uniform ionospheric conductivity. Positive portion of the high latitude ionosphere containing only open
(negative) slopes in AB~ correspond to regions of current magnetic field lines. Precipitating particles on these field
into (out of) the ionosphere. Thus, due to the divergence lines should be of direct magnetosheath origin. Except pos-
of ionospheric Pedersen currents, a field-aligned current sibly during periods of northward interplanetary magnetic
(FAC) should flow into the ionosphere at the equatorward field (IMF), cold ionospheric plasma should convect in the
boundary of the dusk side oval and out at the poleward antisunward direction under the influence of a dawn to dusk
boundary. The morning side oval currents have the opposite electric field.
polarity. If there are significant Pedersen currents across the
polar cap, the poleward FAC system should be of greater
intensity than the equatorward system. 8.5.1 Large Scale Electric Fields Patterns
In the following subsections, AB~ is given as a function
of time rather than distance. Since satellites at ionospheric Table 8-1 lists the six polar orbiting satellites launched
altitudes travel at speeds of -7 km/s, Equation (8.35) can to date that were capable of measuring ionospheric electric
be transformed to give a convenient expression for jll fields. Data from the double-probe experiment on Injun 5
confirmed the existence of convective reversals near the
= 0,113 d ABY(nT) poleward boundaries of the auroral oval. An inclination of
j~ (8.37)
d t(s) “ 68° allowed the Atmospheric Explorer (AE-C) driftmeter to
()
measure convective drifts in the oval but usually not in the
A current of 1 ~A/m~ corresponds to a locally unbalanced polar cap. The AE-D satellite had an inclination of 90° but
flux of - 109/cm-2 s- 1. failed -4 months after launch. The orbit was initially close
Finally, in both the polar cap and the auroral oval we to the noon-midnight meridian and precessed toward dawn-
refer to small- and large-scale structures. Small-scale struc- dusk. Because of the high altitude of S3-3, its near-apogee
tures have latitudinal dimensions of a few tens of km or data have been most useful for identifying the small-scale
less. They are traversed by satellites in a few seconds. features of auroral arc/inverted-V phenomena. The OGO 6
Discrete arcs and inverted-V’s are examples of small-scale and S3-2 satellites spent sufficient periods of time near the
structures. Somewhat arbitrarily, we define large-scale sys- dawn-dusk meridian to identify the main large-scale features
terns as having latitudinal dim~nsions greater t~an 100 km. of polar cap convection.

Table 8-1. Satellites capable of measuring electric fields.

INITIAL TYPE
SATELLITE LAUNCH DATE INCLINATION APOGEE (km) INSTRUMENT

Injun 5 Aug 1968 81° 2550 double

probe
OGO 6 June 1969 82° 1600 double
probe
AE-C Dec 1973 68° 4000 drift
meter
AE-D Ott 1975 90” 4000 drift
meter
S3-2 Dec 1975 96° 1550 double
probe
S3-3 Aug 1976 98° 8050 double
probe

8-13
CHAPTER 8
Two examples of Ex measured by OGO 6 at northern when BY > 0. For both OGO 6 and S3-2, A types are only
(summer) high latitudes are given in Figure $-15. Except found in the summer hemisphere and are associated with the
for small-scale variations, the main features of the expected IMF polarity that tends to produce strong E fields along the
Ex patterns are found in the auroral oval. The example in evening flank of the polar cap. The occurrence ration of A
the top trace is consistent with a uniform dawn to dusk to B was 2:1 in S3-2 data; the A to D ratio was 3:1 in June
electric field across the polar cap. In the bottom trace Ex 1969 OGO 6 data. The dependence of D types on BY (upper
has relatively high (low) values near the morning (evening) right plot) agrees with OGO 6. The type F patterns were found
flank of the polar cap. when BY -0. The distribution of B, D, and F patterns tend
A simplified summary of large-scale electric field pat- to confirm the critical role of BY [Friis-Christensen et al.,
terns identified by Heppner [1972] is given in Figure 8-16. 1972]. A survey of S3-2 measurements shows that A, B, D,
Type A: Uniform Ex across polar cap. These were and F types are found when Bz < 0.5 nT. Type I shows no
observed by OGO 6 only in the northern comlation with Bx or BY. The highly irregular electric fields
(summer) polar cap. are found only when Bz >0. Type I patterns are further
Type B: Strong Ex near the morning (evening) flank discussed below in connection with polar cap arcs.
of the northern (southern) polar cap. An example of an E-field pattern that was found in ap-
Type D: Strong Ex near the evening (morning) flank proximately half the summer polar cap passes of S3-2 when
of the northern (southern) polar cap. Bz >0.7 nT is shown in Figure 8-18. During S3-2, Rev 5215
Type F: Strong Ex along flanks of polar cap with southern high-latitude pass Bz was 4.9 nT [Burke et al., 1979].
weak fields in the central polar cap. As expected for driving sunward convection in the auroral
Type I: [rregular fields across the polar cap. oval, Ex was directed from dusk to dawn. At the poleward
Types A and D (B) were found when the IMF was in a boundary of the oval, Ex reversed plarity becoming dawn to
toward (away) structure. dusk near the morning and evening flanks of the polar cap.
During the last three months of 1976 the S3-2 orbit was Within the central polar cap Ex was directed from dusk to
close to the dawn-dusk meridian. Figure 8-17 is a scatter plot dawn. Sunward convection in the central polar cap is incon-
of Heppner’s patterns observed in S3-2 data as a function of sistent with a viscous interaction model. It was theoretically
IMF Bx and BY. It is seen that A and B types are found only anticipated as a consequence of the magnetic merging model

NORTHERN HIGH LATITUDES

- A+EX “SOUTilW-”

-60L

MLT
INV.IAT

NORTHERN HIGH LATITUDES

so
o
-w MORNWG
E I I -Ex ~WN TO DUSKI I
MLT 17ham ,7h ~m ~ h Zom ~h23m
Id mm
INV.MT. ~~o 75.8° +)W4 7W 6X9*

Figure 8-15. Two examples of large scale electric fields measured by OGO 6 [Heppner, 1972].

8-14
 MAGNETOSPHERIC AND HIGH LATITUDE IONOSPHERIC ELECTRODYNAMICS
S3-2 REV.5215
:ROSS-POLAR CAP JUNE,1969
SOU th Polar Cap
:LECTRIC FIELD PATTERNS OGO-6 - <Bx>= 39 “--
8or. B.>.-0.8 1

NORTHERN SOUTHERN IMF

HEMISPHERE HEMISPHERE STRUCTURE
I
w
VENING MORNING MORNING EVENING
,;::::,
~ A ,::::,:::
,,::,:::,,
99 ,,:::,::::

+/
I
180y
I

m 1’
1:

[6)
,L. L- —l-,
,1/ , .,(,
180” 1.
——— ——— ——— ——. —.—— ——— I ‘,

0600 ---{i~f:;:;’$_$_ ,800

F~ \
‘,
...
,1
. ..
,/,/

\ ‘. ‘., , / /

2400

Figure 8-16. Types ofelectric fieldpattems observed byoG06md their Figure 8-18. Example ofsunward convection, dusk todawrn electric field
dependence of IMF BX and BY [Heppner, 1972]. in the central polar cap during a period of northern IMF.

[Russll, 1972]. Sunward convection in the summer polar cap

-2 -4 -6 642 -2 -4 -6 during periods of northward B= is consistent with both ground
8,x TYPE
0,. TYPE
A
B
A TYPE D 6 magnetometer [Maezawa, 1976] and laboratory simulation
.

4
[Podgomy et al., 1978] results.
2
The cross polar cap potential AWP can be derived from
.
a. satellite electric field measurements and is an important

..
. 0

.
-2
parameter for magnetospheric modeling. This potential (in
.
volts) gives the rate (in webers/s) that magnetic flux is
-4
. ,
transferred from the day to the night side of the magneto-
-6
sphere. Based on two weeks of OGO 6 data Heppner [1977]
3E 3
1- found that on the average AVW increased from 20 to 100
:
kV as the magnetic index Kp increased from O to 6. There

E.
‘6 o TYPE F . TYPE 1 6
m were individual cases in which A~w significantly exceeded
4 “. 4
L +. 100 kV.
Z2 2
. *
. Reiff et al. [1981] analyzed 32 measurements of AWW
0 0
.
from AE-C and AE-D as a function of various solar wind
❑ *.-2
-2
❑
. . and IMF parameters. To calculate merging rates that account
-4
n
-4
for compression of the IMF in the magnetosheath, B at the
‘.
-6 . -6
magnetopause was set at the lesser of 8 times its solar wind
3
“B
642 0 -2-4-6 642 -2 -4 -6
value or 60 nT. The latter value was taken as typical of the
IMF By (nT) earth’s field near the subsolar magnetopause. Although the
best agreement was found with theoretical merging rates, a
high correlation was found with Akasofu’s [1978] E param-
Figure 8-17. Scatter plot of convective electric field patterns observed by
S3-2 as a function of IMF BX and BY. eter. Figure 8-19 shows that in the AE measurements VYW

8-15
CHAPTER 8
200 I I t ,.8 , I , , ,,, m
P ORBIT 1176
3

150
●
●

s
x ●☛
100 ●
-7 10’ CLEFT
u w
●e ~
~
(03:31:54 UT)
e’
@ = 30+ 000061 s T
Q 50 ●
k
q
E
u
oo~ 2,0
~ 106--

2 \
A
e = VB2 sin4(8/z) (104 nT2 km/s) g
UI
Figure 8-19. Cross polar cap potential as a function of solar wind param- n
eter E [Heppner, 1977]. s
n 10’-
- \
A

varied from 30 to 150 kV. This implies that 30 kV cannot POLAR RAIN
be accounted for by merging. Such a potential greatly ex- - (03’28:34 >\ UT) ~
ceeds the potential across the boundary layer theoretically
t i
estimated by Hill [ 1979] and measured at ionospheric al-
titudes by Smiddy et al. [ 1980]. The upper limit of -150 lo’~
10 10* 10’ 10’
kV is much less than the cross magnetosphere potential
ELECTRON ENERGY (eV)
drop (AV,W) in the solar wind. With V, = 400 km/s
and Bz = – 5nT, the Y component of the electric field
Figure 8-20 Typical differential spectra for the polar rain and cleft pre-
in the solar wind is 2 mV/m ( = 12.8 kV/R~). For a mag- cipitation observed on orbit 1176 on 15 May 1969 [Win-
netospheric diameter of 30 R~ at the dawndusk meridian, ningham and HeiUila, 1974].
A*,W = 384kV.
lobes of the magnetotail indicate that polar rain electrons
enter the magnetosphere at great distance downstream from
the earth. The efficiency of the entry process is modulated
8.5.2 Polar Cap Precipitation by the polarity of the IMF. Yaeger and Frank [ 1976] found
that fluxes of soft electrons in the northern lobe of the tail
Particle fluxes into the polar ionosphere are conveniently increased by more than an order of magnitude when the
divided into high and low energy components. Energetic IMF was in an away (Bx < O) sector. There is also evidence
particles from solar flares can seriously disrupt the polar suggesting a BY influence. Meng et al. [ 1977] found that
ionosphere; these are important during magnetic storm pe- the intensity of polar rain fluxes is strongest near the flank
riods. The flux levels of low energy protons in the polar of the polar cap along which convection is strongest.
cap are below the sensitivity levels of existing detectors. Polar showers are characterized by locally enhanced fluxes
Winningham and Heikkila [1974] identified three classes of of precipitating electrons with mean energies of -1 keV.
low energy electron precipitation: polar rain, polar showers, These electron structures are embedded in broader regions
and polar squalls. of polar rain. They are thought to be responsible for sun-
Polar rain is a relatively uniform type of precipitation aligned arcs in the polar cap and are discussed further in
that can fill the entire polar cap. Particles have mean thermal Section 8.5.3.
energies of -100 eV and are isotropically distributed outside Polar squalls are described by Winningham and Heikkila
the atmospheric loss cone. The energy fluxes carried by [1974] as localized, inte~se fluxes of electrons that have
these particles are in the 10-2 to 1@3 ergs cm-z s-’ range, undergone field-aligned accelerations of several kV. They
two to three orders of magnitude less than typical auroral are found in the polar cap during geomagnetic storms. Foster
fluxes. The highest energy fluxes for polar rain occur during and Burrows [ 1976; 1977] have reported observing fluxes
periods of geomagnetic activity [Meng and Kroel, 1977]. of electrons into the polar cap that are spectrally identical
Figure 8-20 is a plot of precipitating electron spectra mea- to those found in polar squalls. However, these fluxes were
sured in the dayside cusp and in the polar rain. The similarity observed to be widely and uniformly distributed over the
in spectral shape suggests that polar rain particles are of polar cap. Like squall particles, they were observed during
direct magnetosheath origin. Particle fluxes measured in the the recovery phases of magnetic storms. These fluxes also

8-16
 m
MAGNETOSPHERIC AND HIGH LATITUDE IONOSPHERIC ELECTRODYNAMICS
appear to be modulated by the polarity of the IMF. While are produced at F-layer altitudes by electrons with energies
intense fluxes of keV electrons were measured in the north- of a few hundred eV [Weber and Buchau, 1977]. Here we
ern polar cap, only polar rain was detected in the southern illustrate many of the known characteristics of polar cap
hemisphere. Figure 8-21 is a plot of particle fluxes measured arcs using data from the USAF satellites S3-2 and DMSP
(Defense Meteorological Satellite Program). At the times
of interest, both satellites were in orbits close to the dawn-
dusk meridian. S3-2 measured Ex, ABY (described in Sec-
I tion 8.5. 1), and fluxes of electrons with energies between
MARCH 09-11, 1972 50 eV and 17 keV. DMSP satellites are three-axis stabilized
/
1s1s 2
POLAR CAP and are in circular, sun-synchronous orbit at an altitude of
POLAR RAIN . . .
.’
,-
‘1
840 km. All DMSP satellites are equipped with scanning,
,,
\ optical imagers [Esther, 1979]. Some, but not all, are also
.,
equipped with spectrometers that look toward local zenith
\ ‘!
MAGNETOSHEATH ‘, and measure fluxes of electrons with energies between 50
eV and 20 keV.
VELA 5A
Figure 8-22 is a cartoon that represents, in magnetic
TAIL LOBE

../ ---, “,
‘. DECEMBER 12,1977
‘\
‘\ ,,
‘. !
‘, :
“-’~ ‘$ ,,
‘, I
,.

.L--L.-~~.
“,

I 2 3 4
LOG ELECTRON ENERGY (ev)

Figure 8-21 Differential electron energy spectra measured in the polar \ / .

‘L
70“ .“
.K70 .._.--& A
cap (solid curves) and in the tail lobe and magnetosheath \ ;

.0 ,6,
(dashed curves). The magnetosheath measurement was made .
.

at -13 hs on 9 March, the tail Iohe at -9 hs on 11 March,

the polar cap at -16 hours on IO March, and the polar rain
at -11 hs 9 March. The polar cap spectra display a more
or less pronounced high-energy tail relative to the magne-
tosheath [Foster and Burrows, 1976].

in the polar rain and extended squalls by ISIS 2 and nearly

simultaneous measurements of electron fluxes in the mag-
netosheath and the tail lobe from VELA 5. Foster and Bur-
rows [1977] argue from the near isotropy of the keV particles Figure 8-22, Cartoon representation of sequential DMSP high latitude im-
in the polar cap and the simultaneous absence of keV elec- agery on 12 kemher 1977.
trons in the magnetosheath that the electrons were accel-
erated along magnetic field lines at great distance from the latitude and local time, composites of visible imagery from
earth in the magnetosphere. DMSP/Fl and DMSP/F2 taken over the northern hemi-
sphere during a period of magnetic quieting on 12 December
1977. Straight solid lines give the portions of F2 trajectories
8.5.3 Polar Cap Arcs during which electron data were taken. To the right of each
cartoon, the hourly average values of IMF BY and Bz are
Investigations of discrete arcs in the polar cap have represented. During the initial period of southward Bz, the
shown that polar cap arcs tend to be sun-aligned and are polar cap was clear of visible emissions and only uniform,
most frequently observed during periods of magnetic qui- polar rain fluxes were detected. Approximately one hour
eting when the IMF has a northward component [Ismail et after the IMF turned northward sun-aligned arcs were found
al., 1977]. Visible arcs are caused by precipitating electrons in the polar cap. Polar cap arcs persisted until the IMF again
with energies of <2 keV. Another class of subvisual arcs turned southward. An hour after a second northward turning

8-17
CHAPTER 8
of Bz, arcs returned to the polar cap [Hardy et al., 1981]. it passed within 10 of the magnetic pole along the dawn-
The sun-aligned arcs were embedded in a region of high dusk meridian. The IMF X, Y, and Z components were
density (-O. 1 cm-3) polar rain. Within the arc the up-look- – 3.7, 3.8 and 7.4 nT, respectively. As compared with the
ing DMSP spectrometer detected three spectral components, idealized measurements of Figure 8- 14b, Ex and ABY were
a cold ( 100 eV) high densit y ( 1.5 cm-3) population, a peaked highly irregular. Recall that in the northern hemisphere Ex
primary distribution with a temperature of 350 eV that had positive corresponds to sunward convection; jll is out of the
been accelerated through a potential drop of -750 volts and ionosphere in regions where ABY has a negative slope. Eight
a secondary and/or degraded primary population. Burch et regions of negative slope in ABY accompanied by enhanced
al. [1979] found that over polar showers the low-energy electron fluxes are noted in Figure 8-23. As evidenced by
component was highly field-aligned. The secondary and their being embedded in polar rain, Events 3 through 7 lie
accelerated primary populations were nearly isotropic in in the polar cap.
pitch angle. Event 6 has been analyzed in detail by Burke et al.
Figure 8-23 gives a plot of Ex, ABY, the directional flux [1982]. It was shown that the FAC out of the ionosphere
of electrons (cm* s sr)-’ and electron pitch angles measured had an intensity of 2.8 ~ A/m* that was carried by electrons
during S3-2 Rev 5231 as functions of invariant latitude, with a temperature of 200 eV and had been accelerated
magnetic local time, and altitude. The pass occurred while through a potential drop of -1 kV. A nearly isotropic pitch
the satellite was near apogee over the north polar cap where angle distribution of electrons across event 6 suggests afield-

S3–2 REV 5231/N 8DECEMBER,1976

60 1 1 1 1 I 1 II I

!
I
i

30 500

300

100

0 -100 ~
c

-300 ->
m
-500 a

-30
(! ‘1 illl
II
II I Ill’
1/ I Ill’
I Ill
I

-60
,.9

,.8

,.7

..
184 184 187 058 06 I 06 i 062 062
MLT 183 183
1315 1312 1237 1267 1242 1209 1181
ALT 1295 1305 1313

Figure 8-23. The top panel gives the dawn to dusk electric field component and the transveme magnetic field deflection (heavy line). The bottom panels give
the directional electron flux and pitch angles. Data were taken over winter polar cap with IMF Bz northward.

8-18
 m
MAGNETOSPHERIC AND HIGH LATITUDE IONOSPHERIC ELECTRODYNAMICS
MAGNETOPALJSF roral physicist is replete with terms such as “inverted-Vs”,
“double-layers”, “diffuse aurora”, and “beams and conies”
[Mozer et al., 1980]. These terms describe distinct auroral
processes first reported in the 1970s. These new phenomena
defy adequate summation in this section. here we consider
three main topics: (1) global auroral morphology, (2) phe-
nomena associated with inverted-V events, and (3) sub-
storms. Under the heading of global morphology we treat &
systematic of large-scale, field-aligned current systems and
the equatorward boundary of diffuse auroral precipitation.

8.6.1 Global Field Aligned Currents

The average global field-aligned current (FAC) system

for periods of high and low magnetic activity m shown in
1 1
II
1
IONOSPHERE I
t
Figure 8-14. Iijima and Potemra [1978] define Region 1 (2)
I
I
I
1 as a region of FAC new the poleward (equatorWard) portion
1
100 –,
1
I of the auroral oval. In the evening sector current flows into
I
I the ionosphere in Region 2 and out in Region 1. The polar-
I
50 –1 I ity of current flow is reversed in the morning sector. The in-
I
1

E
( 1
1
tensities of these currents are of the order 1 pA/m2. Cur-
~.

1
1
rents out of the ionosphe~ m mostly carried by pwipitating
fi
1
1
I
I
electrons. It is believed that cold electrons moving from the
I 3.33
-50 –1 +e.4kv
ionosphere to the magnetosphere carry the current into the

1
c ionosphere. Klumpar [1979] found that the equatorward
-=3 =~.
-100
boundary of Region 2 is coterminus with diffuse auroral
4
electron precipitation in the post-midnight sector. In the
-+ -1oo

I I I
evening sector, Region 2 extends -2° equatorward of the
2002:40 02:55 03:10 electron precipitation boundary.
UT Near local midnight and noon large-scale FACS are the
most complex. Figure 8-25 shows that near midnight the
Figure 8-24. Idealized twodimensional projection ofelectric tields and
cm-rentsin vicinity of Event 6 in Figure 8-23. ,A, ,>, m,

aligned potential drop extending for large distances along

B. Themeasured electron energy fluxes of2.5 * 0.5 ergs
cm-2 s-1 sr-l were sufficient toproduce a visible arc. Figure
8-24 gives an idealized, two-dimensional projwtion of the
electric fields artd currents associated with Event6. Aspre-
dicted by Lyons [1980; 1981], tie arc is in a i-egion of neg-
ative electric fielddivergence. Similar, recently reported measu-
rements from the Dynamics Explorer satellites suggest that
such polar cap arcs may result from a bifurcation of the
distant plasma sheet. fiat is, some polar cap arcs are on
closed magnetic field lines [Frank et al., 1982].
Figure8-25. A sommaryof tbe distributionand flowktions of Iarge-ti
field-alignedcurrents determined from (a) data obtained from
439 passes of Triad during weakly distu~ conditions and
8.6 AURORAL OVAL (b) data obtained from 366 Triad passes dining active periods
ELECTRODYNAMICS [Iijima and Potemra, 1978].

The auroral oval is a region of closed magnetic field lines morning side Region 1 current overlaps the evening side
mapping through the plasma sheet. Ionospheric plasma con- Region 1. No simultaneous electric field and magnetic field
vection should be mostly in the sunward direction. Near lo- measurements have yet been reported from this region.
cal midnight (noon) strong equatorward (poleward) convec- However, it is expected that the morphology of these FACS
tion components are expected. can be understood in terms of latitudinal variations of the
The stock-in-trade vocabulary of a contemporary au- electric field just prior to local midnight. Maynard [1974]

8-i9
CHAPTER 8
showed that in the late evening sector the large scale con- field model of Fairfield and Mead [ 1975] the auroral bound-
vective electric field is directed poleward in the equatorward aries were projected to the magnetospheric equatorial plane.
part of the oval. It rotates through west across the Harang Figure 8-26 is a plot of projected boundary positions (open
discontinuity to equatorward in the poleward part of the circles) in comparison with predictions of the Volland-Stem
oval. Pedersen currents driven in the ionosphere converge electric field model [Ejiri et al., 1978] (solid line) and the
from both sides on the Harang discontinuity. To maintain injection boundary of Mauk and McIlwain [ 1974] (dashed
an overall divergence-free current system, current must flow lines). Best fits are obtained for a shaping factor y = 2.
into the ionosphere at both the equatorward and poleward Note that the stagnation points are offset from the dusk
boundaries of the oval and out of the ionosphere near the meridian. The offset angle +,, varies from 23° toward eve-
Harang discontinuity. These are the essential features found ning for Kp = O to 45° toward afternoon for Kp = 5. The
near midnight in Figure 8-25. potential distribution in the equatorial plane (Equation (8.24)
In the vicinity of the dayside cusp an extra FAC system of the previous chapter) takes the form
has been observed poleward of the Region 1. Its polarity is
L,

1
opposite to that of the nearby Region 1 current. In the
W (L, ~) = Q BORE’ — sin(~ – +.) – ~ .
northern (southern) hemisphere it appears only on the after- [ yLsY+ 1
noon (morning) side of noon when IMF By <0, and only
(8.38)
on the morning (afternoon) side when IMF BY >0
[McDiarmid et al., 1979]. Simultaneous electric and mag-
For y = 2 the term 1/ y (L~~+ ‘)is empirically related to
netic field measurements from the USAF satellite S3-2 in
Kp
the region of the dayside cusp suggest that the extra FAC
system lies entirely on open, newly merged magnetic field
1
lines that are being dragged toward the dusk or dawn flank — = (1.6 + 2.4Kp) .10-4
of the polar cap [Doyle et al., 1981]. yLs’+i

with a correlation coefficient of 0.97. Thus.

8.6.2 Equatorward Boundary of the Oval

In the previous section it was pointed out that the inner

edge of the plasma sheet, or zero-energy Alfv6n boundary,
maps to the equatorward boundary of diffuse auroral precip-
itation. More than four thousand crossings of this boundary
[
(1.6 + 2.4 Kp) 10-4L2sin(@ – @O) – ~
1 (8.39)

have been analyzed using data from an upward-looking A further analysis [Hardy et al., 1982] has been per-
electron spectrometer on the polar orbiting USAF satellites formed correlating the boundary data and hourly averaged
DMSP/F2 and DMSP/F4 IGussenhoven et al., 1981]. The solar wind and IMF measurements. The most significant
corrected geomagnetic latitudes of the boundaries (ACGM) results were obtained when Ac~~ was correlated with the
were studied as functions of magnetic local time (MLT) and interplanetary electric field VBZ for the hour preceding the
Kp. Kp is a three-hour index of magnetic activity compiled boundary measurement subject to the condition Bz < 1 nT.
from a world-wide network of midlatitude magnetometer Correlation parameters are summarized in Table 8-4. To
stations. Results of linear correlation analyses compare the results of the two studies a correlation analysis
was performed between Kp and VBZ.
Ac~~ = Ao + a Kp With Bz s 1 nT

from available MLT sectors along with correlation coeffi- Kp = 2.09 – 0.91 VBZ (8.40)
cients are given in Tables 8-2 and 8-3. Using the magnetic

Table 8-2. Regressi(m coefficients for auroral boundaries in the morning sector.

North South

MLT A,, a N cc A<, a N cc

04004500 67.4 –1.35 171 –0.58

0500-0600 67.8 – 1.87 365 –0.75
0600-0700 68.5 – 1.96 403 –0.82 67.4 – 1.67 376 –0.74
07~+800 70.2 –2.15 367 –0.83 68.3 – 1.97 411 –0.81
0800-0900 68.7 –1.88 302 –0.72
0900- I000 69.1 –1.64 217 –0.67

8-20
 a
MAGNETOSPHERIC AND HIGH LATITUDE IONOSPHERIC ELECTRODYNAMICS
Table 8-3. Regression coefficients for auroral boundaries in the evening sector.

North South

MLT AO a N cc AO a N cc

16W–1700 71.3 –1.19 107 –0.65

1700-1800 70.7 – 1.20 256 –0.69
1800-1900 71.6 – 2.00 103 –0.90 70.6 – 1.60 327 –0.80
1900-2000 71.2 – 1.96 426 –0.89 70.0 – 1.82 447 –0.87
2000-2100 69.4 – 1.85 452 –0.82 69.5 – 1.89 345 –0.84
2100-2200 68.7 – 1.66 556 –0.83
220C-2300 68.3 – 1.79 184 –0.63

with VBZ in mV/m. This allows us to express the magne- pare it with measured cross polar cap potential drops (A+w).
tospheric potential as a function of VBZ. To estimate AV~ we use the position and shape of the
magnetopause in the equatorial plane given by McIlwain
X(L, +) = QBORE2 [1972]. In this representation the magnetopause lies at a dis-

1
tance of 11RE near the subsolar point and flares to a distance
(6.6 + 2.2 VBZ) 10-4L2sin(@ – $0) – ~ . (8.41)
[ of 15RE at the dawn-dusk meridian. We ignore stormtime
compressions of the magnetosphere in the calculations. Two
To test the validity of this equation it is possible to calculate methods of calculating AW~ are used. The first method sets
the cross-magnetospheric potential drop (AW~) and com- 00 = O in Equation (8.40) making the axis of symmetry the

12

06

00

00

12 12

I KD=3 12 KP=5

, ?’=2

I
\
06 18 06
\

. .5

00
00
0’0

Figure 8-26. Volland-Stem injection boundaries for y = 2 and y = 3, rotated to tit tbe inner edge of the plasma sheet as determined by the DMSP/F2 au-
roral boundaries, for various Kp. The Mauk-McIlwain [1974] injection boundary is also shown (dashed line).

8-21
CHAPTER 8
Table 8-4. The intercepts (Ao), slopes, correlation coefficients (cc) and sample sizes (N) for the linear reg~ssion of the boundary location with VBZ in
each magnetic local time zone, with a 1 h delay in the value of the interplanetary magnetic field (IMF) used.

North Pole South Pole

MLT A. slop cc N L slope ‘cc N

04004500 65.1 3.28 0.52 43

0500-0600 63.6 1.94 0.61 110
0600-0700 64.2 1.80 0.70 105 64.4 1.70 0.56 124
0700-0800 66.1 2.56 0.68 85 63.7 1.95 0.61 94
0800-0900 65.1 2.27 0.69 68
09OO-1OOO 65.5 1.74 0.74 38
1000-1100 65.8 0.44 0.59 19
1600-1700 69.5 1.89 0.82 26
1700-1800 68.6 1.36 0.74 57
1800-1900 68,6 2.49 0.81 44 67.2 1.60 0.70 83
1900-2000 67.3 1.76 0.68 123 66.6 1.81 0.70 141
2000-2100 65.2 1.71 0.70 130 65.8 2.01 0.64 44
2100-2200 65.5 1.91 0.71 132

dawn-dusk meridian. The second method uses empirical of 400 km/s. The solid line calculations assume +0 = O,
values of +0 given in Figure 8-26. For Kp = O the axis is that is, the dawn-dusk meridian is the axis of symmetry.
tilted with the dusk stagnation point 23° to the Nghtward side For this case, AT. increases linearly from 6.5 kV for Kp = O
of the dawn-dusk meridian. For higher values of Kp the to 65 kV for Kp = 6. The dot-dash line indicates values
stagnation point rotates sunward reaching a constant value of A*~ using empirical values of +0 and axes of symmetry
of – 45” for Kp > 2. In Figure 8-27 we have sketched the indicated in Figure 8-27. In these calculations AW~ ranges
from 5.5 kV for Kp = O to 95 kV for Kp = 6. We note
that in these calculations there is an abmpt transition in the
trend of AW~ near Kp = 2. This is close to our indepen-
dently derived “breakpoint” in Bz.
For reference, in Figure 8-28 we have also plotted av-
erage values of BWP (dashed line) as a function of Kp based
on two w~ks of OGO 6 electric field measurements. Heppner
[1977] found that AWWincreased linearly from 20 kV for

I 00

80

KP)2
2
60

I
o

Figure 8-27. Equatorial magnetopause with distances along axes of sym-

metry for various values of Kp. 20

axes of symmetry indicating approximate distances to the 0 I I I I I I

2
dawn and dusk sides of the magnetopause for Kp = O, 1,2 K
P
4 6
1 1 1 1 1

and >2. VBZ o 2 4 m V/m

I I I
In Figure 8-28 AW~ is plotted as a function of Kp, VBZ o -5 -10 nl

and Bz. VBZ is positive when the interplanetary electric

field (IEF) is directed from dawn to dusk (positive Y in
Figure 8-28. Polar cap potential, from Heppner [ 1977] (dashed line), and
standard geocentric, solar-magnetospheric coordinates). from ~uation (8. 12) with +0 = O (solid line) and +0 with
Values of Bz are derived from VBZ using a solar wind speed empirical values (dot-dash line).

8-22
 H

MAGNETOSPHERIC AND HIGH LATITUDE IONOSPHERIC ELECTRODYNAMICS

Kp = O to 100 kV at Kp = 6. At low values of Kp the AE-D
DMSP values of AT~ are not sensitive to the value of +0 “INVERTED v“ EVENT OCCURRENCE MAP
and are considerably less than Heppner’s AWP. With values
of +0 = – 45° for Kp >2 VT. and A*W are of com- 02
parable magnitudes. In a time-average sense AWF and VW~
03
are coupled through the process of magnetic reconnection
in the magnetotail. The two quantities must be roughly the
04
same since on average, the amount of magnetic flux trans-
ferred to the nightside magnetosphere must equal the amount
05
being returned to the dayside magnetosphere.
We emphasize that these are “average” results in the
wnse that they represent least squm fits to the data. Heppner M
[1977] pointed out the AWP frequently exceeds 100 kV
during period of high Kp. Similarly, in both DMSP bound- 07
aries for cases of large negative Bz and high Kp respectively,
the latitude of the boundary is often much lower than the Oa
average. Since such cases correspond to the Alfv4n bound-
ary being closer than average to the earth, they also cor-
09
respond to potentials exceeding 100 kV.
10
1
11 12 13
8.6.3 Inverted-V Phenomenology LOCAL TIME

Diffuse auroral electron precipitation is fairly uniform and Figure 8-29. Spatial occurence map of 280 inverted-V events. Local time
and circles of constant invariant latitude are shown as refer-
isotropically distributed. In the poleward pordon of the oval,
ence [Lin assd Hoffman, 1979].
electron fluxes are more spatially structured and are often field
aligned. Discrete optical arcs are the most striking manifes-
tation of the structuring process. The dominant structural later. At the energy of the peak flux the electrons are field
features of electron fluxes observed with polar orbiting sat- aligned whereas those with higher energies are isotropic over
ellites are the so-called inverted-V structures. These s~c- the downcoming hemisphere. This distribution is consistent
tures are latitudinally narrow (- 1°) bands of electron pre- with the electrons having been accelerated through a field-
cipitation that inc~ase in average energy from a few hundred aligned potential drop at an altitude above the point of ob-
eV to several keV, then return to a few hundred eV [Frank servation [Evans, 1974]. Electrons with energies less than
and Ackerson, 1971]. On energy-time s~tragrams they have that of the peak show highly complex pitch angle distri-
the shapes of inverted-Vs. Although individual inverted-V butions. Some of these electrons are secondaries and de-
structures have been identified with ground-observed dis- graded primaries trapped between a magnetic mirror point
crete arcs [Ackerson and Frank, 1972], the general relation- and an electrostatic potential barrier. Some of the low-en-
ship between the two phenomena is not clear. Inverted-Vs ergy electrons are highly field-aligned suggesting that they
have latitudinal dimensions of several hundred km, whereas are accelerated electrons of ionospheric ongin. If so, the
optical arcs have typical widths of 1 to 10 km [Maggs and complexity of their distribution suggests that the field-aligned
Davis, 1968]. potential drop is varying temporally and/or spatially.
Lin and Hoffman [1979] have studied the global distri- One of the most exciting developments of the 1970s was
bution of inverted-V structures and the pitch angle distri- the development of direct evidence for the existence of field-
bution of electrons within them. Figure 8-29 is an event aligned potential drops from measurements by instrumen-
occurrence map of 280 inverted-V structures observed with tation on the S3-3 satellite. The observational work of the
the AE-D satellite. This map shows that inverted-Vs are University of California, Berkeley group is summarized by
found throughout the high-latitude region. They appear pre- Mozer et al. [1980].
dominantly in the late evening sector and are absent from The highly eccentric orbit of S3-3 carried it to altitudes
the late pre-noon MLT sector. of -8000 km above the auroral ionosphere. This is a hitherto
The pitch angle distribution of electrons measured over unexplored region. At these altitudes very intense electric
an inverted-V event is given in Figure 8-30. This spectro- field structures were observed. An example given in Figure
gram gives the ratio of field-aligned flux at a given energy 8-31 shows electric fields at 8000 km reaching 400 mV/m
to the flux at pitch angles of 60°. The heavy black line gives and undergoing rapid reversals. If mapped, assuming
the energy at which the maximum differential flux was meas- E “ B = O, to aurora] arc altitudes the electric field in-
ured. Note that the energy of the peak flux rises to -3 keV tensities would be of the order of 1000 mV/m and the lat-
at 1008:05 UT, then decreases to -0.5 keV 20 seconds itudinal width of the structure would be <10 km. This is a

8-23
CHAPTER 8

FLUX RATlO SPECTROGRAM FLUX RATlO

AE-D 75309
-.
75 A

I STEPPED -7/60
10 L
<0.9 M
(0.9, 1.1) m

1008:00 1008:30
UNIVERSAL TIME
Figure 8-30. Spectrogram displaying the directional differential flux ratios ( – 7°) and (60°) over an inverted-V event. The flux ratios are separated into four
categories as shown in the upper right comer [Lin and Hoffman, 1979].

typical width of a discrete optical arc. However, the electric

field intensity greatly ~xceeds that measured in the lower wave turbulence that are collocated with inverted-V precip-
ionosphere. For this reason it is argued that there must be itation structures. Within the electric field reversals we find
a potential drop along magnetic field lines that accelerate fluxes of field-aligned electrons going into, and H + and O +
plasma sheet electrons to form arcs. ions going out of the ionosphere. The upgoing ions are
These electric field structures have been observed at all decided according to pitch angle distributions into “beams”
altitudes down to 1000 km. They are mostly found above and “conies. ” Ion beams are highly directed along magnetic
4000 km. Comparisons of S3-3 electric field measurements field lines. They have only been observed at altitudes greater
with simultaneous measurements from other instruments on than 4000 km and in the region of discrete rather than diffuse
the spacecraft show that the electric field structures are aurora. They are thought to be accelerated upward by the

6 SEPTEMBER 1976 S3- 3 ORBIT 475

400
: 200 - (a)
Ex >
E .zo~ -
1 1 1 1
-400
(b)
Ey ●

P
1 I 1 I 1 I I I I
I -4C)0
UNIVERSAL TIME 0240 0300 0320
ALTITUDE, km 7593 8042 7465
MAG. LOCAL TIME 18.7 17.6 14.7
61.0 71.5 m.5
I NV. LAT. Degrees

Figure 8-31. An example of a rapid electric field reversal observed at 8000 km by S3-3 [Mozer et al., 1980]

8-24
 a

MAGNETOSPHERIC AND HIGH LATITUDE IONOSPHERIC ELECTRODYNAMICS

same parallel electric fields that accelerate plasma sheet 5. The recovery phase coincides with the period in which
electrons downward to form discrete aurora] arcs. Ion conies midnight sector arcs retreat equatorward.
have pitch angle distributions that peak at some intermediate The substorm signatures dealt with in the August 1978
angles bet ween 90° and OO.They have been detected at all meeting are observable from ground-based instrumentation.
latitudes in the auroral oval and at all altitudes greater than Except for the micropulsation bursts, the’ signatures are of
1000 km. They are thought to be accelerated through res- ionospheric effects whose causes lie in the magnetosphere.
onant interactions with upper hybrid and ion cyclotron waves A key observation for understanding how substorms may
[Gomey et al., 1981]. be triggered comes from numerous satellites at geostationary
altitude. During pre-substorrn periods the nightside mag-
netic field at 6.7 R~ takes on a tail-like configuration. At
8.6.4 Substorms substorm onset the magnetic field rapidly recovers its nor-
mal, nearly dipolar configuration. Simultaneously hot plasma
No overview of high-latitude electrodynamics would be sheet particles, with no dispersion, are injected.
complete without some comment, no matter how cursory, To explain the observations the following scenario has
on substorms. It is during substorms that the dynamic cou- been pieced together.
pling between the magnetosphere and ionosphere is most 1. At a southward turning of the IMF, magnetic flux
striking. Despite intensive studies of substorm processes is transferred from the day to the night side of the
over the last fifteen years, the richness of the observations magnetosphere. This process proceeds for about half
has made total agreement on what constitutes the essential an hour in which potential energy, in the form of
elements of a substorm elusive. stored magnetic flux, builds up in the tail. During
The evolution of discrete aurora] arcs in the late evening this period the neutral sheet current moves earthward
sector as synopsized from all-sky camera data by Akasofu to -10 R~ leading to a tail like field geometry at
[1964] is given in Figure 8-32. Under pre-substom con- 6.7 R~. Recall that discrete arcs map to the boundary
ditions, homogeneous arcs extending for thousands of kil- plasma sheet rather than the central plasma sheet.
ometers in the east-west direction are found in the oval and Thus, the most equatorward arc maps to a region
sun-aligned arcs are found in the polar cap. Substorm onset near the inner edge of the neutral sheet current.
is announced by a brightening of the most equatonvard arc. 2. At substorm onset, the neutral sheet current near the
The discrete arcs expand poleward and westward traveling inner edge of the boundary plasma sheet is diverted
surges develops in the evening sector. After expanding to via field-aligned currents through the ionosphere.
some maximum latitude, the arcs slowly retreat toward their This leads to a collapse of the inner portion of the
pre-substorm condition. tail. In the ionosphere, part of the energy released
To resolve differences between various schools of thought, in the collapse of the tail appears as an explosive
nine active investigators met in August 1978 to specify brightening of the most equatorward arc. As the in-
substom signatures and unambiguously define words com- ner-tail field lines snap back to dipolar, plasma sheet
monly .(oftep differently) used to describe substorm pro- electrons are rapidly accelerated by inductive electric
[Rostoker et al., 1980]. Major points of agreement fields and are injected to the vicinity of geostationary
were distance. The process continues while Bz remains
1. During extended periods of northward IMF, the mag- southward.
netosphere quiets and asymptotically approaches a 3. When the IMF turns northward the rate of flux trans-
ground state. fer decreases abruptly. If the IMF maintains a north-
2. As the IMF turns southward, magnetospheric con- ward component for considerable time, the potential
vection increases. This enhanced convection can ex- energy stored in the tail is slowly dissipated and the
ist for some time prior to substorrn onset. ‘ magnetosphere relaxes toward a “ground state. ”
3. Substorm onset is signaled by an explosive increase
in luminosity of the most equatorward arc, an in-
tensification of the auroral electrojet, and a burst of 8.7 CONCLUSIONS
Pi2 micropulsation. The burst of micropulsations tes-
tifies to the explosive nature of the onset process in This chapter provides an introductory summary of what
the magnetospheric source region. is known about the electrodynamics of the magnetosphe~ and
4. The expansion phase occurs from onset to the time the high latitude ionosphere prior to the launch of the Dy-
when the midnight sector arcs have undergone their namics Explorer satellites. Our knowledge is an amalgam of
most poleward excursions. Note that in this defini- theoretical models and in si~u observations. In a qualitative
tion, multiple intensifications of the substorm pro- sense, we are able to explain the gross features of magneto-
cess, each marked by a micropulsation burst, are spheric processes. For example, a set of MHD equations has
allowed. been used successfully to calculate the equilibrium shape of

8-25
CHAPTER 8

60°

G“*, G .\

B. T. O-5MIN

,/

/“ / ~ 60”
70”

/
/

c. T=5-10 MIN o. T= IO-30 MIN

,, /’ 70”
\

E. T=30MIN-IHR F T= I-2HR T=2-3 HR

F-DA

Figure 8-32, Schematic diagram to illustrate the development of the auroral substorm. The center of the concentric circles in each state is the north geomag-
netic pole, and the sun is toward the top of the diagram [Akasafu, 1964]. (Reprinted with permission from Pergamon Press, Ltd. O 1964. )

8-26
 MAGNETOSPHERIC AND HIGH LATITUDE IONOSPHERIC ELECTRODYNAMICS
the magnetopause [Mead and Beard, 1964]. However, the work of cartographers from the mid-seventeenth century. The
microphysics of how (i~ magnetic merging occu’rs at the mixed fruits of explorations spanning two hundred years are
magnetopause and how particles cross the magnetopause to evident. Europe, the Middle East, Africa, and the eastern
form the boundary layer are not yet understood. An essen- shores of the Americas are mapped with exquisite detail.
tially MHD model for coupling between the ionosphere and Shapes assigned to the American west coasts and to the na-
magnetosphere [Ashour-Abdalla and Thorne, 1978] has been tions of East Asia range from vague to fanciful. Our own
applied to successfully simulate substorm effects in the in- position twenty years into the age of satellite exploration is
ner plasma sheet and low latitude portion of the auroral oval analogous to that of seventwnth-century map-makem. Some
[Lyons et al., 1979; Harel et al., 1981]. Thus, adiabatic parts of the magnetosphere-ionosphere system are well ex-
energization is an important and well understood magneto- plored. Our understanding of other important parts is based
spheric process. Other energization processes associated with on fragmentary data samples and leaves much to be desired.
reconnection in the magnetotail ‘m not understood. The role Undoubtedly, readers a few decades hence will find many of
of parallel electric fields for the fo~ation of auroral arcs was our ideas as quaintly amusing as the seventeenth-century
experimentally established in the 1970s. Competing theo- image of Japan. To those who view catiography or space
retical models of how parallel electric fields form in a colli- physics as quasi-static processes, this is a scandalous situa-
sionless plasma fill the literature. tion. To those involved in the process, it crystallizes the ex-
With the gifi of hindsight, it is interesting to ex~e the citement of exploration.

8-27
CHAPTER 8
REFERENCES Fahleson, U. V., “Theory of Electric Field Measurements
Conducted in the Magnetosphere with Electric Probes,”
Ackerson K.L. and L,A. Frank, ’’Correlated Satellite Mea- Space Sci Rev., 7:238, 1967.
surements of Low-Energy Electron Precipitation and Fairfield, D.H. and G.D. Mead, “Magne(ospheric Mapping
Ground-Based Observations of a Visible Auroral Arc,” with a Quantitative Geomagnetic Field Model ,“ J. Geo-
J. Geophys. Res., 77: 1128, 1972. phys. Res., 80:535, 1975,
Akasofu, S.-I., “The Development of the Auroral Sub- Foster, J.C. and J .R. Burrows, “Electron Fluxes Over the
storm,” Planet Space Sci., 12:273, 1964. Polar Cap 1, Intense keV Fluxes During Poststorm Qui-
Akasofu, S .-I., “Interplanetary Energy Flux Associated with eting,” 1. Geophys. Res., 81:6016, 1976.
Magnetospheric Substorms,” Planet Sp. Sci., 27:425, Foster, J .C. and J.R. Burrows, “Electron Fluxes Over the
1978. Polar Cap 2, Electron Trapping and Energization on
Armstrong, J. and A.J. Zmuda, “Triaxial Magnetic Mea- Open Field Lines,” J. Geophys. Res., 82:5165, 1977.
surements of Field Aligned Currents at 800 km in the Frank, L. A., “Initial Observations of Low-Energy Electrons
Auroral Region: Initial Results,” J. Geophys. Res., 78: in the Earth’s Magnetosphere with OG03 ,“ J. Geophys.
6802, 1973. Res., 72: 185, 1967.
Ashour-Abdalla M. and R.M. Theme, “Toward a Unified Frank, L. A., “Relationship of the Plasma Sheet, Ring Cur-
View of Diffuse Auroral Precipitation,” Y. Geophys. rent, Trapping Bounda~, and Plasmapause Near the
Res., 83:4755, 1978. Magnetic Equator and Local Midnight ,“ J. Geophys.
Axford W .1. and C. O. Hines, “A Unifying Theory of High- Res., 76:2265, 1971.
Latitude Geophysical Phenomena and Geomagnetic Frank L.A. and K.L. Ackerson, “Observations of Charged
Storms,” Can. J. Phvs., 39:1433, 1961. Particle Precipitation into the Auroral Zone,” J. Geo-
Burch, J. T., S .A. Fields and R.A. Heelis, “Polar Cap Elec- phys. Res., 76:3612, 1971.
tron Acceleration Regions,” 1. Geophys. Res., 84:5863, Frank, L. A., K.L. Ackerson and L.P. Upping, “On Hot
1979. Tenuous Plasmas, Fireballs, and Boundary Layers in
Burke, W. J., M.C. Kelley, R.C. Sagalyn, M. Smiddy, and the Earth’s Magnetotail ,“ J. Geophys. Res., 81:5859,
S.T. Lai, “Polar Cap Electric Field Structure with a 1976.
Northward Interplanetary Magnetic Field,” Geophys. Res. Frank, L. A., J.D. Cravens, J.L. Burch, and J.D. Win-
Lett., 6:21, 1979. ningham, “Polar View of Earth’s Aurora with Dynamics
Burke, W. J., M.S. Gussenhoven, M.C. Kelley, D.A. Hardy, Explorer,” Geophys. Res. Lett., 9: 1001, 1982.
and F.J. Rich, “Electric and Magnetic Field Character- Freeman, J. W., “Internal Magnetospheric Plasma Flow,” in
istics of Discrete Arcs in the Polar Cap,” J. Geophys. International Sydney Chapman Conference on Magne-
Res., 87:243 1–2443, 1982. tospheric Boundary Layers, Alpbach, Austria, 1979
Crooker, N .U., “Dayside Merging and Cusp Geometry,” (ESA-SP-148), European Space Agency, Paris, 1979.
J. Geophys. Res., 82:3629, 1977. Friis-Christensen, E., K. Lassen, J. Willjelm, J.M. Wil-
Crooker, N. U., “The Magnetospheric Boundary Layers: A cox, W. Gonzalez, and D.S. Colbum, “Critical Com-
Geometrically Explicit Model ,“ J. Geophys. Res., 84: ponent of the Interplanetary Magnetic Field Responsible
951, 1979. for Large Geomagnetic Events,” J. Geophys. Res., 77:
Doyle, M. A., F.J. Rich, W.J. Burke and M. Smiddy, 3371, 1972.
“Field Aligned Currents and Electric Fields in the Re- Gold, T., “Motions in the Magnetosphere of the Earth,” J.
gion of the Dayside Cusp,” J. Geophys. Res., 86:5656, Geophys. Res. 64:1219, 1959.
1981. Gomey, D. T., A. Clarke, D. Croley, J. Fennell, J. Luh-
Dungey, J. W. “Interplanetary Magnetic Field and the Au- mann, and T. Mizera, “The Distribution of Ion Beams
roral Zones,” Phys. Rev. Lett., 6:47, 1961. and Conies Below 8000 km,” J. Geoph-ys. Res., 86:83,
Eastman T.E. and E. W. Hones Jr., “Characteristics of the 1981.
Magnetospheric Boundary Layer and Magnetopause Layer Gringanz, K. I., V.V. Bezrukikh, V.D. Ozerov and E.E.
as Observed by IMP 6,” Y. Geophys. f?es., 84: 2019, Rybchinskii, “A Study of the Interplanetary Ionized Gas,
1979. High Energy Electrons, and Corpuscular Radiation from
Eastman, T. E., E.W. Hones Jr., S.J. Bame, and J.R. As- the Sun by Means of the Three-Electrode Trap for
bridge, “The Magnetospheric Boundary Layer: Site of Charged Particles on the Second Soviet Cosmic Rocket ,“
Plasma Momentum and Energy Tranfer from the Mag- SOV.Phys. Dokl. 5:361, 1961.
netosheath into the Magnetosphere ,“ Geophys. Res. Lett., Gumett, D.A. and L.A. Frank, “A Region of Intense Plasma
3:685, 1976. Wave Turbulence on Aurora] Field Lines,” J. Geophys.
Eather, R. E., “DMSP Calibration,” J. Geophys. Res., 84: Res., 82: 1031, 1977.
4134, 1979. Gussenhoven, M. S., D.A. Hardy and W.J. Burke, “DMSPIF2
Ejiri, M., R.A. Hoffman, and P. H. Smith, “The Convection Electron Observations of Equatorward Auroral Bound-
Electric Field Model for the Magnetosphere Based on aries and Their Relationship to Magnetospheric Electric
Explorer 45 Observations,” J. Geophys Res., 83:4811, Fields,” J. Geophys. Res. 86:768, 1981.
1978. Haerendel, G., G. Paschmann, N. Sckopke, H. Rosenbauer
Evans, D. S., “Precipitating Electron fluxes formed by a and P. C. Hedgecock, “The Frontside Boundary Layer
magnetic field-aligned potential difference,” J. Geo- of the Magnetosphere and the Problem Reconnection ,“
phys. Res., 79:2853, 1974. J. Geophys. Res. 83:3195, 1978.

8-28
 MAGNETOSPHERIC AND HIGH LATITUDE IONOSPHERIC ELECTRODYNAMICS
Hanson, W.B. and R.A. Heelis, “Techniquesf orMeasur- Lyons, L. R., “Generation of Large-Scale Regions of Au-
ing Bulk Gas Motions from Satellites,’’ Space Sci1nst., roral Currents, Electric Potentials, and Precipitation by
1:493, 1975. the Divergence of the Convection Electric Field,” J.
Hardy, D. A., H.K. Hills, and J.W. Freeman, “A New Geophys. Res., 85: 17, 1980. .
Plasma Regime in the Distant Geomagnetic Tail,” Geo- Lyons, L. R., “ Discrete Aurora as the Direct Result of an
phys. Res. Lett. 2:169, 1975. Inferred High-Altitude Generating Potential Distribu-
Hardy, D. A., M.S. Gussenhoven, E. Holman, W.J. Burke, tion,” J. Geophys. Res., 86: 1, 1981,.
and N. Heinemann, “DMSP/F2 Electron Observations Lyons, L. R., D.S. Evans, and R. Lundin, “An Observed
of Equatorial Aurora] Boundaries and Their Relation- Relationship Between Magnetic Field Aligned Electric
ship to the Solar Wind Velocity and the North-South Fields and Downward Electron Energy Fluxes in the
Component of the Interplanetary Magnetic Field,” Y. Vicinity of Auroral Forms,” J. Geophys. Res., 84:457,
Geophys. Res., 86:996 1–9974, 1981. 1979.
Hardy, D. A., W.J. Burke, and M.S. Gussenhoven, “DMSP Maezawa, K., “Magnetospheric Convection Induced by the
Optical and Electron Measurements in the Vicinity of Positive and Negative Z Components of the interpla-
Polar Cap Arcs,” J. Geophys. Res., 87: 24 13–2430, netary Magnetic Field: Quantitative Analysis Using Po-
1982. lar Cap Magnetic Records,” ~. Geophys. Res., 81:2289,
Harel, M., R.A. Wolf, P.N. Reiff, R.W. Spiro, W.J. Burke, 1976.
F.J. Rich, and M. Smiddy, “Quantitative Simulation of Maggs, J. E. and T. N. Davis, “Measurement of the Thick-
a Magnetospheric Substorm, 1, Model Logic and Over- ness of Aurora] Structures,” Planet. Space Sci., 16:205,
view, and 2, Comparison with Observations,” J. Geo- 1968.
phys. Res., 86, 2217, 1981. Mauk, B .H. and C.E. McIlwain, “Correlations of KP with
Heppner, J. P., “Polar Cap Electric Field Distributions Re- Substorm Injected Plasma Boundary,” J. Geophys. Res.,
lated to the Interplanetary Magnetic Field Direction,” J. 79:3193, 1974.
Geophys. Res., 77:4877, 1972. Maynard, N .C., “Electric Field Measurements Across the
Heppner, J. P., “Empirical Models of High-Latitude Elec- Harang Discontinuity,” J. Geophvs. Res., 79: 4620,
tric Fields,” J. Geophys. Res., 82:1115, 1977. 1974.
Hill, T. W., “Generation of the Magnetospheric Electric Field,” McDiarmid, J. B., J.R. Burrows, and M.D. Wilson, “Large
in Quantitative Modelling of Magnetospheric Processes, Scale Magnetic Field Perturbations and Particle Mea-
AGU Monograph 21, edited by W.P. Olson, American surements at 1400 km on the Dayside,” J. Geophys.
Geophysical Union, Washington D. C., p. 297, 1979. Res., 84: 1431, 1979.
Hines, C. O., “The Energization of Plasma in the Magneto- McIlwain, C. E., “Plasma Convection in the Vicinity of
sphere: Hydromagnetic and Particle Drift Approaches ,“ Geosynchronous Orbit” in Earth’s Magnetospheric Pro-
Planet. Space Sci. 10:239, 1963. cesses, edited by B .M. McCormac, 268, D. Reidel,
Hones, Jr., E. W., J.R. Asbridge, S.J. Bame, M.D. Mont- Dordrecht, Holland, 1972.
gomery, and S. Singer, “Substonn variations of the Mead, G.D. and D. B. Beard, “Shape of the Geomagnetic
magnetotail plasma sheet from XsM = —6 RE to Field Solar Wind Bound~,” J. Geophys. Res., 69:1169,
XSM = – 60 RE,” J. Geophys Res., 78:1109, 1973. 1964.
Iijima, ~. and T.A. Potemra, “Large-scale characteristics of Meng, C.-I. and H. W. Kroel, “Intense Uniform Precipita-
field aligned currents associated with substorms,” J. tion of Low-Energy Electrons Over the Polar Cap,” J.
Geophys. Res., 83:599, 1978. Geophys. Res., 83:2305, 1977.
Ismail, S., D.D. Wallis, and L.L. Cogger, “Characteristics Meng, C.-1., S.-I, Akasofu and K.A. Anderson, “Dawn-
of Polar Sun-Aligned Arcs,” J. Geophys. Res., 82:4741, Dusk Gradient of the Precipitation of Low-Energy Elec-
1977. trons Over the Polar Cap and its Relation to the Inter-
Johnson, F. S., “The Gross Character of the Geomagnetic planetary Magnetic Field,” J. Geophys. Res., 82:5271,
Field in the Solar Wind,” J. Geophys. Res., 65:3049, 1977.
1960. Mozer, F. S., R. B. Torbert, U. V. Fahleson, C.G. Faltham-
Kennel C. F. and H.E. Petschek, “Limit on Stably Trapped mer, A. Gonfalone. A. Pedersen, and C.T. Russell,
Particle Fluxes,” J. Geophys. Res., 71:1, 1966. “Direct Observation of a Tangential Electric Field Com-
Kivelson, M. G., S.M. Kaye and D.J. Southwood, “The ponent at the Magnetopause,” Geophys. Res. Lett., 6:
Physics of Plasma Injection Events,” in Dynamics of the 305, 1979.
Magnetosphere, edited by S.-I. Akasofu, 385, D. Rei- Mozer, F. S., C.A. Cattell, M.K. Hudson, R.L. Lysak, M.
del, Dordrecht, Holland, 1979. Temerin, and R.B. Torbert, “Satellite Measurements and
Klumpar, D. M., “Relationships Between Auroral Particle Theories of Low Altitude Auroral Particle Accelera-
Distributions and Magnetic Field Perturbations Associ- tion,” Space Sci. Rev., 27: 155, 1980.
ated with Field Aligned Currents,” J. Geophys. Res., 84: Northrop, T. G., The Adiabatic Motion of Charged Paticles,
6524, 1979. Chap. 3, Wiley, New York, 1963.
Lemaire, J. and M. Roth, “Penetration of solar wind plasma Olson, W. P., “The Shape of the Tilted Magnetopause,” J.
elements into the magnetosphere ,“ J. Atmos. Ter. Phys., Geophys. Res., 74:5642, 1969.
40:331, 1978. Pilipp, W .G. and G. Morfill, “The formation of the plasma
Lin, C.S. and R.A. Hoffman, “Characteristics of the In- sheet resulting from plasma mantle dynamics,” J. Geo-
verted-V Event ,“ J. Geophys. Res., 84:1514, 1979. phys. Res., 83:5670, 1978.

8-29
CHAPTER 8
Piddington, J .H., “Theories of the Geomagnetic Storm Main Smiddy, M., et al., “Effects of High Latitude Conductivity
Phase,’’ Planet. Space Sci., 11:1277, 1963. on Observed Convection Electric Fields and Birkeland
Podgomy, I. M., E.M. Dubinin, and Yu. N. Potanin, “The Currents,” J. Geophys. Res., 85:6811, 1980.
Magnetic Field on the Magnetospheric Boundary from Southwood, D.J. and S .M. Kaye, “Drift Boundary Approx-
Laboratory Simulation Data,” Geophys. Res. Lett., 4: imations in Simple Magnetosphenc Convection Models,”
207, 1978. J. Geophys. Res., 84:5773, 1979.
Reiff, P. H., R.W. Spiro, and T.W. Hill, “Dependence of Speiser, T. W., “Particle Trajectories in Model Current
Polar-Cap Potential Drop on Interplanetary Parameters,” Sheets, II: Application to Auroras Using a Geomagnetic
J. Geophys. Res., 86:7639, 1981. Tail Model,” J. Geophys. Res. 72:3919, 1967.
Rich, F. J., D.L. Reasoner, and W.J. Burke, “Plasma Sheet TerHaar, D., Elements of Hamilronian Mechanics, Chapter
at Lunar Distance: Characteristics and Interactions with 6, North Holland, Amsterdam, 1964.
the Lunar Surface,” J. Geophys. Res. 78:8097, 1973. Vasyliunas, V. M., “A Survey of bw-Energy Electrons in
Rosenbauer, H., H. Grunwalt, M.D. Montgomery, G. the Evening Sector of the Magnetosphere with OGO 1 and
Paschmann, and N. Sckopke, “HEOS 2 Plasma Obser- OGO 3,” J. Geophys. Res., 73:2839, 1968.
vations in the Distant Polar Magnetosphere: The Plasma Vasyliunas, V. M., “Mathematical Models of Magneto-
Mantle,” J. Geophys. Res., 80:2723, 1975. spheric Convection and its Coupling to the Ionosphere,”
Rossi, B. and S. Olbert, Introduction to the Physics of in Particles and Fields in the Magnetosphere, edited
Space, Chapter 5, McGraw-Hill, New York, 1970. by B .M. McCormac, D. Reidel, Dordrecht, Holland,
Rostoker, G., “Magnetospheric and Ionospheric Currents 1970.
in the Polar Cusp and Their Dependence on the BYCom- Vasyliunas, V. M., Magnetospheric Plasma, in Solar Ter-
ponent of the Interplanetary Magnetic Field,” J. Geo- restrial Physics, edited by E. Dyer, D. Reidel, Dor-
phys. Res., 85:4167, 1980. drecht, Holland, p. 192, 1972.
Rostoker, G., S.-I. Akasofu, J. Foster, R.A. Greenwald, Y. Vasyliunas, V. M., “Interaction Between the Magneto-
Kamide, K. Kawasaki, A.T.Y. Lui, R.L. McPherron, and spheric Boundary Layers and the Ionosphere ,“ interna-
C.T. Russell, “Magnetospheric Substorms-Definitions tional Sydney Chapman Conference on Magnetospheric
and Signatures,” J. Geophys. Res., 85:1663, 1980. Bou&ry Luyers, Alpbach, Austria, 1979 (ESA-SP- 148),
Russell, C. T., “The Configuration of the Magnetosphere,” European Space Agency, Paris, 1979.
in Critical Problems of Magnetospheric Physics, edited Weber, E.J. and J. Buchau, “Polar Cap F-Layer Auroras,”
by E.R. Dyer, 1, National Academy of Science, Wash- Geophys. Res. Lett., 8:125, 1981.
ington, D. C., 1972. Williams, D.J. and L.R. Lyons, “The Proton Ring Current
Russell, C.T. and R.C. Elphic, “ISEE Observations of Flux and Its Interaction with the Plasmapause: Storm Recov-
Transfer Events at the Dayside Magnetopause,” Geo- ery Phase,” J. Geophys. Res., 79:4195, 1974.
phys. Res. Lett., 6:33, 1979. Winningham, ~.D. and W .J. Heikkila, “Polar Cap Auroral
Sckopke, N., G. Paschmann, H. Rosenbauer, and D.H. Electron Fluxes Observed with Isis 1,“ J. Geophys. Res.,
Fairfield, “Influence of the Interplanetary Magnetic Field 79:949, 1974.
Oneral and Thickness of the Plasma Mantle,” J. .Geo- Yaeger, D.M. and L.A. Frank, “Low Energy Electron In-
phys. Res., 81:2687, 1976. tensities at Large Distances Over the ~’s Polar Cap,”
Sckopke, N., G. Paschmann, G. Haerendel, B .U.O. Son- J. Geophys. Res., 81:3966, 1976.
nerup, S.J. Bame, T.G. Forbes, E.W. Hones, Jr., and Young, T. S. T., J.D. Callen and J.E. McCune, “High Fre-
C.T. Russell, “Structure of the Low-Latitude Boundary quency Electrostatic Waves in the Magnetosphere,” J.
Layer,” J. Geophys. Res., 86:2099, 1981. Geophys. Res., 78:1082, 1973.

8-30
 Chapter 9

IONOSPHERIC PHYSICS
Section 9.1 F.J. Rich
Section 9.2 F.J. Rich
Section 9.3 Su. Baau

9.1 STRUCTURE OF THE IONOSPHERE of the layers are shown in Table 9-1. This profile is valid
only for midlatitudes. In the equatorial region, the profile
is distorted by the geomagnetic field, and in the polar region,
9.1.1 Ionospheric Layers
the Profile is distorted by ionization by energetic particles,
magnetospheric coupling, and other effects.
The ionized atmosphere of the earth is composed of a
The D region is present only during daylight hours. me
series of overlapping layers. In each layer there is an altitude
altitude of the peak density is normally around 90 km, but
of maximum density, above and below which the ionization
this may decrease considerably to -78 km when the solar
density tends to drop off. The total ionization profile with
x-ray flux is enhanced. The E region peak density occurs
the layers indicated is shown in Figure 9-1. Characteristics
at a Deskaltitude of 110 km. At sunset, the E Region electron
den;ity drops by a factor of 10 or more in a ‘short period
MI DLATITUDE DENSITY PROFILES
I 000
I l\ \ \ \ (tens of minutes) before reaching a nighttime equilibrium
8oo- density. At night, the region of low density near 150 km
SOLAR MAx
600–
between the E region and the F region may have a sharply
SOLAR MIN lower density than shown in Figure 9-1 or the density may
4oo- be great enough so that there is no depletion region de-
% pending on geophysical conditions. The F region is a com-
: NIGHTTIME —
bination of two somewhat different regions. The Fl region
u ELECTRONS
a
has an altitude peak near 200 km, but is absent at night.
~ 2oo—
F
The F.Zregion has a peak near 300 km during the day and
< I 50,– at higher altitudes at night. Shortly after sunset, the absolute
density near the peak of the F region often increases due to
10o —
plasma transport processes before decreasing to a night time
80 value.
60 1 1 1 1 1 1
The Topside Ionosphere is the name given to the rest
10 I 102 103 lo~
,.5 ,.6 107 of the ionosphere above the F region peak, In a simple
-3,
DENSITY (cm model of the ionosphere, the density of the topside iono-
sphere decreases exponentially with height with some char-
Figure 9-1. Total ionization profile with ionospheric Iayem.

Table 9-1. Layers of daytime.midlatitude ionosphere.

Layer Altitude(km) Major Component Production Cause

D 70–90 km NO+, 02+ Lyman Alpha, x-rays

E 95–140 km 02+, NO+ Lyman Beta, Soft x-rays,
UV Continuum
F1 140-200 km 0+, NO+ He II, UV Continuum
(100-800~)
F2 2~ km 0+, N+ He II, UV Continuum
( 100-8CM)~)
Topside F >400km 0+ Transpofi from Below
Plasmasphere > 1200 km H+ Transport from Below

9-1
CHAPTER 9
acteristic scale height until the ionization density is below photons v by setting the production rate equal to the loss
detectable levels. In the mid and low latitude ionosphere, rate for a quasi-equilibrium ionosphere. The production rate
the geomagnetic field tends to trap ions, especially hydrogen is
ions, that would otherwise drift off into deep space. Thus,
somewhere between 800 and 2000 km altitude the scale Q. (h) = ~“ N(h) F“ (h) ionizations cm-3 s-’ (9.1)
height increases to a very large value ( > one earth radius).
If one follows a geomagnetic field line out to the equatorial where ~“ is the absorption cross section for photoionization,
plane and back into the conjugate ionosphere, the density N(h) is the neutral density at the height h, and F“ (h) is the
would change by less than 102 and in some cases less than photon flux of frequency v. The recombination rate is ap-
10. This region of trapped ionospheric ions is the Plas- proximately
masphere [Carpenter and Park, 1973]. The outer edge of
the plasmasphere where ionospheric ions are not trapped is LU (h) = ~R Ni N. (9.2)
the Plasmapause. The plasmapause is located approximately
along the geomagnetic field line that maps down to 60° where ~R is the recombination coefficient and Ni and N,
magnetic latitude. A representation of the topside iono- are the ion and electron densities. If multiply charged and
sphere and plasmasphere is shown in Figure 9-2. negative ions are not important, Ni = N.. For an appro-
priate set of assumptions, the density near the peak of the
layer is

N.(h) =
q
()[,,,
~R
~ _ (h-h~)’
4H2 1 (9.3)

where QM is the production at the peak of the layer, h“, is

the height of the layer and H is the scale height of the neutral
atmosphere. The parabolic variation of density with h around
the peak gives the layers of the ionosphere their character-
istic shape.
In principle, the above equation needs to be summed
over all appropriate frequencies of incident photons to obtain
the total ionospheric profile. In fact, many details about
---- ionospheric chemistry (especially in the D and E layers) and
/! f \\ PLASMASHEET+

about plasma transport (especially in the F layer) must be

considered. For example, NO is not a significant component
Figure 9-2. Representation of topside ionosphere and plasmasphere
of the neutral atmosphere, but NO+ is a significant ion
component due to a chain of chemical reactions induced by
For further information about ionospheric layers w Banks photoionization. Details about computing an ionospheric
and Kockarts [1973]; Kohnlein [1978]; and Chapters 12 and model with all factors included is given by Banks and Kock-
21. arts [1973].

9.1.2 Chapman Theory for 9.1.3 Ionization Production, Loss

Ionospheric Layers and Vertical Transport

The layering of the ionosphere was first discussed by Ionization production is generally a two-step process.
Sydney Chapman in the 1920s. Ionizing photons from the The first step is the creation of ions from the neutrals by
sun will produce more and more ions as they penetrate solar photons in the ultraviolet and x-ray spectrum, and to
deeper and deeper into an atmosphere with rapidly increas- a lesser degree by collisions with energetic particles. Most
ing density. As photoionization occurs, the flux of photons of the ionization is produced by solar radiation with wave-
is attenuated until a depth is reached that photoionization lengths less than 1026 ~, which ionizes O, 02, and Nz.
production drops. Thus, a layer of ionization near the al- There are a vast array of minor constituents, especially
titude of maximum production is created. molecular and metallic ions, that are important in under-
The process of quantitatively determining the height pro- standing the ionosphere. Some of these minor constituents
file of ionization is outlined as follows. We assume that play important roles in the absorption of solar UV radiation
direct photoionization and recombination are the only sources (especially in the D region), in the production of airglow,
of production and loss respectively. We can find the ion in the chemistry of the ionosphere, and in the role of tracers
density at a height h due to a single frequency of ionizing to indicate ionospheric and atmospheric transport. A review

9-2
 -

IONOSPHERIC PHYSICS
of photoionization is given by Hudson [ 1971] and Stolarski are important to the dynamics of the ionosphere, especially
and Johnson [1972]. above the density peak of the F region. Energetic ions can
The second step in the creation of the ionosphere is the undergo a charge exchange process, cross magnetic field
reaction between ions, neutrals, and electrons to create dif- lines as a neutral and be re-ionized by a second charge
ferent forms of ionization than that created by direct ioni- exchange. The result is a transfer of energy across field lines
zation. Ionospheric chemistry explains why NO+ is an im- that otherwise would be impossible.
portant ionospheric ion despite the low abundance of NO The principal loss of ionization comes from recombi-
in the neutrai atmosphere. ‘Two of the major reactions in nation. This is simply the reaction of positive ions and
the ionosphere which create NO+ are electrons to form neutrals. The most important recombi-
nation reactions are
N2+O+~ NO++ N
NO++e--N+O
Nz + +Oe NO+ + N (9.4)
02+ e--0+0. (9.6)
All of the major reactions are summarized in Figure 9-1 of
Tom [1979]. The reactions that result in minor constituents The charge exchange reaction
are quite numerous and not all of them are well understood.
Reactions that should be mentioned involve metastable atomic 0+ +02+02+ +0 (9.7)
states, negative ions, ionization by photoelectrons, energetic
is important to facilitate the recombination process. The
neutrals and vibrational states of molecules. Further details
major problem related to loss of ionization is not a lack of
of ionospheric chemistry are given in Chapter 21.
knowledge of the reactions but a lack of precise knowledge
After recombination of ions and electrons to form neu-
about the reaction rates. Reaction rates depend upon the
trals, the atoms are often in an excited state that is meta-
density of the ions and neutrals, the atomic state of the
stable. On the ground, such states are de-excited through
incident and resultant ions, the vibrational state of mole-
collisions, but with the lower collision frequency in the
cules, and the temperature of the ions and electrons. If one
thermosphere, forbidden atomic transitions occur which re-
reaction rate for any related part of the recombination pro-
lease photons not generally seen in the laborato~ such as
cess is not known precisely, it is difficult to determine the
the 5577 ~ and 6300 ~ emissions of atomic oxygen.
reaction rate of the complete recombination process. Solar
Negative ions are generally found in the D and lower E
eclipse data have been used to empirically determine some
regions. They are formed primarily through electron-neutral
of the effective recombination rates. The knowledge of re-
molecule collisions. Photodissociation of electrons from
action rates for both production and loss covers a vast array
negative ions provides the major source of D-region elec-
of atomic and molecular reactions [Tom, 1979; Ferguson,
trons shortly after sunrise. See Ferguson [1971] and Turco
197 1; and DNA Rate Book, 1972] known with varying de-
[1974] for a discussion of D-region chemistry.
grees of precision.
After electrons are removed from an atom by photoion-
Vertical transport of ions and electrons is principally
ization, they have an energy that depends upon the work
governed by collision frequency and gravity. Below 130
function of the atom and the energy of the photon. The
km, the collision frequency of ions and electrons is so large
energy distribution function of photoelectrons is a complex
that to a good approximation no ions or electrons enter or
function [Jasperse, 1977 and Doering et al., 1976] that can
leave a unit volume. The unit volume may move up or down
crudely be approximated by a combination of 2 eV and a
under the influence of pressure gradients. Between 130 km
20 eV Maxwellian distribution. In the D and E regions, the
and 300 km the mean free path of electrons, and to a lesser
photoelectrons lose their energy through collisions close to
degree ions, becomes comparable or larger than the scale
the location of their formation. In the F region, the photo-
height or layer thickness of the ionosphere. Diffusion be-
electrons can travel significant vertical distances before los-
comes an important factor in plasma transport and energy
ing all of their energy. As a rule of thumb, photoelectrons
exchange. Above an altitude of 300400 km, the plasma
traveling upward at altitudes above 300 km are considered
can be treated as collisionless for many purposes. At high
as escaping because of the low collision frequency. Except
and midlatitudes, vertical transport is approximately along
in the polar regions, these photoelectrons do not truly escape
magnetic field lines and particles will tend to diffuse upward
but follow magnetic field lines into the opposite hemisphere.
or downward according to their mass. Because the ions and
Some of their energy is lost to plasma in the plasmasphere
electrons are electrically coupled, ambipolar diffusion must
and the rest is used to heat and populate the opposite ion-
be considered. The ions and electrons diffuse upward to-
osphere [Mantas et al. 1978]. Photoelectrons from the op-
gether at a rate that must be less than a neutral particle with
posite hemisphere are most important when one hemisphere
the mass of an electron and slightly greater than a neutral
is in darkness and the other in sunlight.
particle with the mass of the ion. Near the equator, vertical
Charge exchange reactions such as
transport is impeded by the magnetic field lines. In the
O++ HSC)+H+ (9.5) daytime, however, due to electrodynamics effects ionization

9-3
CHAPTER 9
is transported upwards at the magnetic equator which sub- contour of the trough follows the contours of the auroral
sequently diffuses down magnetic field lines at ~ 15° mag- oval instead of paralleling the magnetic latitudes. As the
netic latitudes. This results in the Appleton anomaly as auroral oval moves poleward, the trough moves poleward
discussed in greater detail in Section 9.2.2.4. and expands in width. As the auroral oval moves equator-
ward, the trough moves equatorward and decreases in width.
The depth of the trough (the ratio of minimum density to
9.1.4 Neutral Winds and the density a few degrees equatoward and/or poleward) is
Horizontal Transport typically a factor of ten in density or total electron content
(TEC) but may vary from barely discernible to a factor of
Below 130 km altitude, the ion-neutral collision fre- 103.
quency is so high that ions will freely flow across field lines The equatorward wall of the trough tends to be found
with the neutrals but the electrons are relatively fixed with on the same field line or 1“–2° poleward of the plasmapause.
respect to the magnetic field. As a result, neutral winds tend From this observation has come the suggestion that the
to cause ionospheric currents. These ionospheric currents trough is formed by plasma flow upwards into evacuated
are detected on the ground by the diurnal variations in the flux tubes. The region equatorward lies on filled flux tubes
magnetic field at mid and low latitudes ( + 60° to – 600). that can replenish the plasma, and the region poleward of
The Sq (solar quiet) current system and the equatorial elec- the trough is replenished by ionization induced by precip-
trojet are the major current systems related to the neutral itating energetic particles. An alternate explanation is that
winds. The separation of ions and electrons in the E region the trough is created by the cancellation of the coronation
produce polarization electric fields that can cause E x B and convection electric field in the evening sector. Plasma
drift of plasma in both the E- and F-regions [Evans, 1978]. remains nearly stationery for several hours; depletion is due
Neutral winds are predominantly atmospheric tides with to recombination. After leaving the stagnation zone, the
periods of 24, 12, 8, 6, . . . hours. The zero order neutral trough is sustained through the night by the lack of pro-
wind consists of a steady flow away from - 1400 local time duction in the region of the trough.
toward -0200. The latitude of these high and low pressure The light ion trough is a sharp drop in H+ and He+
zones shifts seasonally with the sun. At high latitudes the density near the peak of F region and in the topside iono-
neutral winds are strongly affected by geomagnetic activity sphere. On the night side, the light ion trough or density
IRoble et al., 1981]. The ring current, which causes the gradient is collocated with the equatorward wall of the total
negative Dst midlatitude deflection is dissipated in the ion- ion trough. On the dayside, the light ion trough continues
osphere near 60° which, in turn, heats the thermosphere. to be found near 60° magnetic latitude while the total ion
The Joule heating of the auroral zones and polar caps and trough either ceases to exist or moves to higher latitudes.
the heating from precipitating particles also cause the high For a satellite traveling in the altitude range from -8~ to
latitude thermosphere to be heated. As a result of the heating -1500 km on the dayside and -600 to -1500 km on the
of the thermosphere, the neutral winds at high and mid- night side, the light ion trough is seen as a rapid transition
latitudes can be strongly affected and even reverse direction. between H+ and O + as the dominant ion [Titheridge, 1976].
The convection electric fields imposed upon the high Above -1500 km, the O+ density is so low even when it
latitude ionosphere from the magnetosphere also cause hor- is the dominant ion that the total density drops by a factor
izontal plasma transport. In the winter ionosphere, the drifts of -10 to >103 at all local times at the same latitude as
driven by the convection electric field are a major source the light ion trough.
of ionization in the polar cap drawing from the auroral zone The plasma temperature in the F region near the trough
and the dayside ISojka et al., 198 1]. The convection electric and the equatorward wall of the auroral zone is increased
field can also decrease ionization by enhancing the recom- substantially from the temperature in adjacent regions. This
bination rates as ions are driven through the neutrals. is partly related to energy from the ring current in the equa-
torial plane of the magnetosphere being transferred to the
ionosphere. When this energy deposition is large enough,
9.2 HIGH LATITUDE PHENOMENA the airglow is enhanced to form stable auroral red (SAR)
arcs. SAR arcs are generally subvisual. They tend to be
most intense following a major geomagnetic storm.
9.2.1 Total Ionization Trough and Light Ion
Trough.
9.2.2 The Ionosphere in the Auroral Oval.
The total ionization trough is a region of decreased F-
region ionization and/or total electron content found in a We have known of the visual displays called auroras for
latitudinally narrow band near 60°-650 magnetic latitude centuries and we have known for the past century that au-
[Ahmed et al., 1979]. The ends of the trough are typically roras are associated with electromagnetic disturbances.
found an hour after sunset and an hour before sunrise. The However, it has only been since the start of the space age

9-4
 d

IONOSPHERIC PHYSICS
we have known auroras are globs! and are the projection of An important feature of the high latitude ionosphere is
activity at great altitudes in the magnetosphere and solar that the plasma density is irregular on the scale of meters
wind (Chapters 3 and 8). The global nature of auroras is to kilometers vertically and on the scale of meters to hundred
generally acknowledged by referring to the auroral oval, a of kilometers horizontally. The small scale irregularities
band a few degrees in latitude and around both magnetic cause scintillation of radio signals passing ‘through the ion-
poles where auroral phenomena are found IFeldstein and osphere. The large scale irregularities or density gradients
Starkov, 1967; Meng, 1977; Gussenhoven et al., 1981 and provide a necessary condition for various plasma turbulence
1983; see also Chapter 12]. The optical emissions are a mechanisms that result in small scale imgultities (see Chapter
result of the energy from precipitating energetic particles 10). The causes of irregularities in the auroral zone are
(mostly electrons) being deposited in the ionosphere. Op- numerous [Fejer and Kelley, 1980] but are related to factors
tical emissions visible to the naked eye are mostly in narrow such as particle precipitation and E x B drifts. The region
latitudinal bands (1 to 10 km in width) called auroral arcs, of auroral zone scintillations extends equatorward of the
but there are precipitating particles and optical emissions optical auroral [Martin and Aarons, 1977], but are roughly
throughout the auroral zone. In the equatorward portion of collocated with the region of particle precipitation. Some
the auroral oval the optical emissions are spatially uniform of the strongest scintillations are observed when the ray path
and are known as the diffuse aurora. Along the poleward of the radio signal is aligned with the magnetic L-shell in
portion of the oval, the precipitating particles tend to be the F region. This has been analyzed to indicate that irreg-
grouped into bands about 10 wide. The intensity and max- ularities are in the form of sheets extended along the mag-
imum energy of the precipitating particles are greatest in netic field and in the magnetic E-W direction [Rino et al.,
the center of the band and fall off near the edges of the 1978].
band. These have been called “inverted-V” events due to
their signature in the records of polar orbiting satellites. If
9.2.3 Substorm Effects.
the maximum intensity is great enough, a visible auroral
arc will appear at the center of the inverted-V event.
In the ionosphere, a substorm is an intensification of
The precipitating energetic particles lose energy to the
processes and structures that are normally observed in the
atmosphere by ionizing neutrals in a manner similar to the
quiet time auroral zone ionosphere. See Chapter 8 for a
ionization caused by protons. If the precipitating particles
complete description. During a substorm the flux and av-
all had the same energy, a thin layer of ionization would
erage energy of the precipitating particles increase rapidly
be formed. The altitude of the ionization layer is determined
and substantially. This causes more visible features to ap-
from the energy of the precipitating particles: 10 keV elec-
pear. The increase in the energy of the particle causes ion-
trons produce an ionization layer near 110 km altitude; 500
ization in the lower E layer that is not present during quiet
eV electrons produce an ionization layer near 180 km al-
times. The increased activity causes an increase in the scin-
titude. The density of the ionization layer is determined by
tillation producing irregularities. The increased E x B drift
the intensity of the particle flux.
rates heat the ionosphere and increase the scale heights.
In the nighttime auroral zone, the electron density profile
During the early phases of a substorm, the aurora] zone
can be accurately estimated if the spectrum of precipitating
moves equatorward several degrees in 10–20 minutes. The
particles is known, or conversely the spectrum of precipi-
trough either moves equatorward with the auroral zone
tating particles can be estimated from the electron density
movement or ceases to exist during the substorm depending
profile [Vondrak and Baron, 1977]. This is especially true
on geomagnetic conditions and local time. During the late
in the E region where the ionization lifetimes are short. In
phases of a substorm the auroral oval contracts poleward
the F region where ionization lifetimes are long, the ioni-
and precipitation induced ionization tends to die away by
zation present at any given moment is influenced by the
recombination.
precipitation over the past few minutes to tens of minutes
as well as the instantaneous precipitation. Also, vertical and
horizontal transport has a major effect upon the structure of 9.2.4 Polar Cap Structure.
the F region. Regions of enhanced F region ionization can
drift many degrees from the production region [Vickrey et The polar cap ionosphere is relatively placid compared
al., 1980]. to the auroral zone, but soft particle precipitation known as
In the sunlit auroral zone, photoionization dominates the polar rain [Winningham and Heikkila, 1974] does affect the
production of ionization, but ionization from particles has polar cap ionosphere. In the summer months. the polar cap
a major effect on plasma irregularities. Also, in limited ionosphere is dominated by photoionization similar to the
altitude regions the ionization from particles can occasion- midlatitude ionosphere. In the winter months. ionization is
ally dominate photoionization. Even where the photoioni- generally maintained by the polar rain and by convection
zation is the dominant source, the heat from precipitating from the day side to the night side of the polar cap. In the
particles increases the scale height of the ionosphere and winter months. the F layer can be sunlit while the E layer
the neutrals. is in darkness. This can lead to He+ being the dominant

9-5
CHAPTER 9
ion in parts of the topside ionosphere. During times of weak where n is the electron or ion concentration, e is the elec-
convection and precipitation, the winter polar ionosphere tronic charge, v. and vi are the electron and ion cdllision fre-
near the midnight sector can decay to very low levels of quencies, and m, and rni are the masses of an electron and an
ionization; this area is called the polar hole [Brifiton et al., ion respectively. The longitudinal conductivity is indepen-
1978]. dent of magnetic field intensity and is identical to the con-
Since the magnetic field lines in the @lar cap diverge ductivity obtained in the absence of any magnetic field.
effectively toward infinity, the H + in the polar ionosphere When an electric field is applied perpendicular to the
escapes rapidly in a process known as the polar wind. Unlike magnetic field, the conductivity in the direction of the elec-
the midlatitude ionosphere, O+ is generally the dominant tric field is called the Pedersen conductivity and is given by
ion at all altitudes of the F-region.

9.3 EQUATORIAL PHENOMENA

u,
Q?)]
‘ne2[Q(Ve:+Q
‘IIl](v?:
(9.9)

9.3.1 Sq Current System where Q. and Qi are the electron and ion gyrofrequencies
respectively.
See Chapter 4. In such cases of a crossed electric and magnetic field,
a Hall current usually flows perpendiculti to both the electric
and magnetic fields and the resulting conductivity y, called
9.3.2 Equatorial Electrojet the Hall conductivity, is given by

The intense eastward ionospheric cumnt that flows by

day over a narrow latitudinal strip along the magnetic equa-
tor is known as the equatorial electrojet [Matsushita and
Campbell, 1967]. The electrojet causes the large daily vari-
ations of the horizontal component of magnetic field inten-
sity recorded by ground magnetometers near the magnetic
equator. The ionospheric cutient system is a result of a At the magnetic equator, an eastward electric field is
dynamo action of the horizontal wind system and the elec- developed by the dynamo action of the horizontal wind
trical conductivity of the ionosphere in the presence of the system, which gives rise to a motion of charged particles
electrons and ions. The concentration of ionospheric current in the east-west direction (X) due to the Pedersen conduc-
near the magnetic equator is a result of the high value of tivity u, and in the vertical direction (Z) due to the Hall
electrical conductivity of the upper atmosphere at the dip conductivity y U2. In view of the horizontal stratification of
equator, which arises from an inhibition of Hall current due the ionosphere, the flow of Hall current in the vertical di-
to the horizontal configuration of the earth’s magnetic field rection (Z) is totally inhibited and a polarization electric
and the horizontal stratification of the ionosphere. field (Ez) develops. In such a case the current densities Jx,
The electrical conductivity of the ionosphere is not only J= in the X and Z directions can be expressed as
governed by the concentration of the charged panicles but
by the neutral particles and the earth’s magnetic field as Jx = UIEX + U2EZ
well [Chapman, 1956]. Through collisions the neutral par-
ticles restrict t~e motion of charged pmicles under the action Jz = –U2EX + UIEZ = O
of any impressed electric field. The presence of the magnetic
field, on the other hand, restricts the motion of charged
particles across the magnetic field and therefore makes the
or,
‘x=(u’
‘:)Ex
conductivity anisotropic. 2

When a dc electric field is impressed parallel to the or. Jx/Ex = U1 + :

magnetic field, the longitudinal electrical conductivity y that
exists parallel to the magnetic field is given by =cr~. (9.11)

The resulting conductivity in the east-west direction is called

[1
1 1 the Cowling conductivity y cr3.
w. = ne2 —+— (9.8)
~ve miVi ‘ The variations of Pederson U1, Hall U2, and Cowling
cr3 conductivities with altitude at the magnetic equator are

9-6
 =

IONOSPHERIC PHYSICS
shown in Figure 9-3a for an assumed variation of electron 9.3.3 Electrojet Irregularities
density N and temperature T shown in Figure 9-3b. The
enhanced value of the Cowling conductivity in the dynamo Ionosondes first detected the existence of a distinct type
region is sufficient to account for the intensity of the equa- of sporadic E near the magnetic equator which is patchy
torial electrojet. Away from the equator, the geomagnetic and transparent to radio waves reflected from higher layers.
The intensity of equatorial E, (or E,~) is strongly correlated
T, deg with the strength of the electrojet current discussed in the
0.5 1.0 1. last section [Matsushita, 195 1]. A typical ionogram showing
250
(a) b) E,q echoes is shown in Figure 9-4. VHF forward scatter
T

I
/
I
\
200 ,

-—. ‘J ,
,”!.; .,.,’

X -—.
,d

I
A ,.
,!.
..’
,

J,//
‘1 I
2 5 10 m
I 2 3XI05 f(Mc /s )

ff, emu N, cm-’

(a) (b)
Figure 9-4. The typical equatorial sporadic E configuration on an iono-
gram recorded at Huancayo, Peru at 1229 hr (75”W) on 19
Figure 9-3. (a) Variation of Pedersen U1, Hall UZ, and Cowling U3 con- April 1960 [Cohen et al, ,“ 1962],
ductivities with altitude for an ionosphere in which electron
density N and temperature T vary as-shown at (b). Note that
scale for U3 is smaller than crI and U2 by a factor of 10
[Chapman and Raja Rae, 1965, based on Chapman, 1956]. experiments established that these echoes arise as a result
of scattering from field aligned irregularities of electron
field is no longer horizontal, which allows the Hall field to density immersed in the equatorial electroject [Bowles and
leak away. B&er and Martyn [ 1952] estimated that the half- Cohen, 1962].
width of the strip of enhanced east-west conductivity around The important characteristics of the electrojet irregular-
the dip equator is about 3° in latitude. The equatorial elec- ities as related to the physics of the scattering region have
trojet corresponds to an east-west electric field of 0.5 mV/m been probed by the VHF radar measurements performed at
and a vertical polarization field of about 10 mV/m. This the Jicamarca Radio Observatory. Radar spectral studies
gives rise to an eastward current or westward electron drift have shown the existence of two classes of irregularities
of several hundred m/s. The electron drift is westward by called Type 1 and Type 2, associated with the electrojet
day and eastward at night. This electrojet model was studied [Balsley and Farley, 1973; Farley and Balsley, 1973; Fejer
in detail by Sugiura and Cain [1966]. More complete models and Kelley, 1980].
allowing for vertical currents have been discussed by Untiedt Type 1 irregularities have a very narrow spectrum with
[1967], Sugiura and Poros [1969], Richmond [1973], and a Doppler shift corresponding approximately to the ion
others. acoustic velocity of about 360 m/s. Farley [ 1963] and Bune-
The equatorial electrojet current has been observed to man [1963] have explained the Type 1 irregularities by
reverse its normal direction during day or night and during showing that a plasma is unstable to waves when the relative
magnetically quiet or disturbed condtions; this reverse cur- electron-ion drift velocity in the direction of the wave ex-
rent system has been termed the counter electrojet [Gouin ceeds the ion acoustic velocity. As such, Type 1 irregular-
and Mayaud, 1967; Hutton and Oyinloye, 1970; Rastogi, ities are also called two-stream irregularities. Type 2 irreg-
1973; Fejer et al., 1976; Fejer and Kelley, 1980]. The rapid ularities on the other hand have phase velocities smaller
reversals during disturbed conditions have been related to than the ion-acoustic velocity and are observed even when
magnetospheric and high-latitude phenomena [Matsushita the eastward drift velocity is very small during the day.
and Balsley, 1972], whereas the reversals during quiet con- Type 2 irregularities are identified with E,~ echoes in ion-
ditions have been related to lunar tides [Rastogi, 1974]. ograms and both disappear under counter electrojet condi-

9-7
CHAPTER 9
tions during the daytime. At night the Type 2 irregularities equator and plasma diffusion along geomagnetic field lines
are almost always observed except for momentary disap- [Martyn, 1959]. Figure 9-6 illustrates how the eastward E-
pearance when the electrojet electric field reverses sign. The region dynamo electric field at locations slightly off the
Type 2 irregularities are explained by the gradient drift magnetic equator maps to F-region altitude over the equator.
instability mechanism. This is because the horizontal po- The eastward electric field in conjunction with the northward
larization field arising from a relative electron-ion drift in
the electrojet region can, in the presence of the earth’s
magnetic field, develop a drift in the direction of the ver-
tically oriented density gradient and give rise to these ir-
regularities [Fejer and Kelley, 1980].

F
9.3.4 Equatorial Anomaly and Fountain
Effect

During the equinox the sun is overhead at the equator,

/’ /’
Figure 9-6. The F region geomagnetic anomaly. Near rhe equator the
and in terms of solar control the ionization density is ex- electric fields of the atmospheric dynamo in the E layer are
pected to be maximum in that region. Instead, the daytime conveyed upwards along geomagnetic lines of force to the
motor in the F layer where they produce an upwards move-
ionization density at the Fz peak shows a pronounced trough
ment of the plasma during the day. The raised plasma then
at the magnetic equator and crests at about 30”N and 30°S diffuses down lines of force to produce enhanced concentra-
magnetic dip. This anomalous latitude variation of F2 ion- tion at places on each side of the equator and decreased
concentration at the equator itself [Ratcliffe. 1972].
ization near the magnetic equator obtained from bottomside
ionograms illustrated in Figure 9-5, was first recognized by
geomagnetic field gives rise to a vertically upward plasma
motion. At high altitudes over the equator, the plasma en-
counters field lines that connect to the Fz peak at 30°N and
30°S magnetic dip along which the plasma diffuses under
the action of gravity. Such plasma transport depletes the FZ
ionization at the equator and increases the density at loca-
tions 3WN and 30°S. Theoretical studies of the equatorial
anomaly on a more rigorous basis have been performed by
many workers [see Hanson and Moffett, 1966 and references
therein].
The transport processes involved in the formation of the
equatorial anomaly are best illustrated by the Alouette I
topside sounder results. Figure 9-7 shows the variation of
ionization density as a function of height and latitude in the
daytime topside ionosphere. At high altitudes over the mag-
netic equator, the density shows a dome-like structure fol-
lowing the shape of a magnetic field line. At lower altitudes,
below 700 km, a field-aligned double humped structure is
obtained, with the maxima being closer together at the greater
1 1 1 heights.
o(m)9~ ;0 3~ I
o -30 -60 -90 The diurnal development of the equatorial anomaly has
Magnetic dlp (deg ) been studied from ground based as well as topside sounders.
Ground based data indicate that during years of sunspot
Figure 9-5 Variation of N~Fl and of electron density (electron concen- minimum the anomaly is most pronounced at about 1400
tration) at fixed heights with magnetic dip, for noon on mag- LT and then declines steadily until it disappears around 2000
netically quiet days in September 1957. The zero level for
each curve is indicated on the left [Croom et al., 1959].
LT [Rastogi, 1959]. However, during the period of sunspot
Reprinted with permission from MacMillan Journals Ltd., @ maximum, the anomaly after an initial decay in the afternoon
1959. ) hours shows substantial redevelopment in the late evening
hours [Martyn, 1959; Appleton, 1960; Wright, 1960] and
Appleton [1946] and is known as the equatorial anomaly or often the ionization density at the crests in the evening period
Appleton anomaly. exceeds the daytime values. Latitudinal asymmetry of the
The equatorial anomaly is explained in terms of a~oun- equatorial anomaly in the northern and southern hemisphere
tain effect caused by vertical electrodynamics drift at the as a function of season and longitude has also been studied

9-8
 IONOSPHERIC PHYSICS

r 900
agated signal. The bulk of the information on magnitude
and occurrence of these irregularities came from amplitude
fluctuation or scintillation measurements at a host of equa-
torial stations (see Chapter 10 for further ,details).
During the last decade a determined effort has been made
700 to understand the nature and occurrence of nighttime equa-
torial F-region irregularities since the largest propagation
effects extending up to S-band frequencies are observed in
this region. Our insight has come from multi-technique ob-
m servations comprising satellite and rocket in situ measure-
‘E ment of irregularity amplitude and spectra, coherent and
$
‘o 0 incoherent radar backscatter measurements, total electron
=
z content and ground based and airborne multi-frequency scin-
g tillation, and all-sky imaging photometer measurements.
k
a 10 Together these techniques measure irregularities over scale
u
1-
Z lengths of 5 to 6 orders of magnitude from hundreds of
w
u kilometers to tens of centimeters, and given the right con-
z 8 0~
0 ditions the post-sunset equatorial F-region is indeed found
v
to contain irregularities over this enormous scale size range
[Basu and Basu, 1981]. A brief description of the different
6 techniques and their results are given below.
+ The rocket and satellite in situ measurements have de-
I
tected large scale irregular biteouts of ion concentration (Ni)
4 0~ HEIGHT (km) in the nighttime equatorial spread-F region associated with
small scale irregularities in Ni [Hanson and Sanatani, 1973;
Kelley et al., 1976; Morse et al., 1977]. A comprehensive
2
study of such depleted regions by McClure et al., [1977]
I
I indicates the presence of very sharp electron density struc-
—o 9! 1 tures (see bottom panel of Figure 9-8) and the existence of
0
-1o” 0° +10” +20” +30° ionic species near the F-peak that are normally obtained in
GEOGRAPHIC LATITUDE (degrees) the bottomside and valley region between the E and F layers.
This study also revealed the existence of a highly structured
Figure 9-7. Latitudinal variation of electron density across the equatorial upward velocity within these depleted regions on the order
anomaly at various altitudes above h~a, from topside iono- of 100 m/s (hence the name bubbles), and sometimes, in
grams [Eccles and King, 1969]. (Reprinted with ~rmission
from IEEE 0 1969.)
addition, a westward velocity of about 20 rnls as shown in

[Lyon, 1963; Lyon and Thomas, 1963]. Interhemispheric

neutral wind and variation of magnetic declination with
longitude have been invoked in theoretical models [Hanson
and Moffett, 1966] to account for such asymmetry.

9.3.5 Equatorial F Region Irregularities

Historically. the signature of equatorial nighttime F-

region irregularities was first obtained from the spread-F
signature on ionograms [Booker and Wells, 1938]. Equa- 1 I
A LT 498
torial spread-F has since been divided into two types, range 633 380 284 212
DLAT -40 27 98 168 23.8
and frequency spread [Calvert and Cohen, 1961; Rastogi, M LT 021 044 068 093 I .22

1980], the former type being attributed to strongly scattering

irregularities. Figure 9-8, Ion drift meter data, orbit 2282. Satellite altitude, dip latitude,
The advent of orbiting and geostationary beacons in the and magnetic local time are indicated on the figure. The sat-
ellite longitude was – 50° to – 65°. The observed pitch and
early 1960s provided another technique for monitoring equa-
yaw angles are shown in the upper and lower curves, res-
torial irregularities, that of measuring the phase, amplitude, pectively. Positive angles correspond to ions moving up or
and plane of polarization of the trdns-ionospherically prop- left with respect to the spacecraft IMcClure et al.. 1977].

9-9
CHAPTER 9
Figure 9-8. Woodman and LaHoz [ 1976] using the radar mensional maps of these structures show that the depletions
technique at Jicamarca observed plume-like structures in observed near the m,agnetic equator have typical E-W di-
backscatter power maps of equatorial irregularities at 3 m mension of 100 km and are as large as 1200 km in the
wave length. The maps they obtained are similar to that magnetic N-S direction. These observations establish that
shown in the top panel of Figure 9-9. They interpret the the bubbles are open at the bottom and confined within
plumes as being due to vertically rising bubbles and their magnetic field tubes. From measurements of both incoherent
wakes. Evidence for the fact that the plasma bubbles, most scatter and coherent backscatter using a steerable radar at
probably initiated by the Rayleigh-Taylor instability, are 155 MHz, Tsunoda [ 1980a] located the 1 m field aligned
probably extended in altitude has been obtained from con- irregularities at the top edge of a plasma bubble and mapped
ventional polarimeter observations of the total electron con- the bubble along the magnetic field tube.
tent (TEC). These observations [Yeh et al., 1979; DasGupta By performing careful coordinated studies of radar back-
et al., 1982] show that scintillation patches in the early scatter, high resolution satellite and rocket in situ and ground
evening hours occur in association with depletions of TEC scintillation measurements, the spatial and temporal coex-
which may be as large as 40% of the ambient value. istence of kilometer and meter scale irregularities have been
From topside sounder observations, Dyson and Benson studied [Basu et al., 1978; 1980; Rino et al., 198 I ]. Figure
[1978] have shown that plasma bubbles are confined within 9-9, taken from Basu and Basu [ 1981], shows that the 3 m
magnetic field tubes. By the use of an all sky imaging irregularities causing the radar backscatter and kilometer to
photometer on board an aircraft, Weber et al., [1978, 1980] several hundred meter irregularities causing scintillations
have detected 6300 ~ airglow depletions in the nighttime are simultaneously (within the limit of time resolution of
equatorial F-region in association with radar backscatter and the experiment - rein) generated in the onset phase, but
scintillation patches. The 6300 ~ airglow depletions signify the short scale irregularities are outlived by the large scale
depletions of integrated ionization density and the two-di- ones by several hours. A considerable effort has been made

7m- JICAMARCA ,% v .%
48 50 MHz BACKSCATTER
20-21 MARCH, 1977
600- 36 1
..
24% . . *-%
2 5m-
. 12 *
.
6 .. -
4oo- we-: ‘ .
:
:
< 3oo- .. . “. . .
.
u d-
-*-
2oe
-i.. +, . .. se._ , -“
100- ,
19 20 21 22 23 & 01
75- W TIME

36 MARCH 20-21, 1977 ANCON (GOES)

1 136 MHz

dB
18-

12-

6-

1 I I I 1 I I
19 20 21 22 23 00 01 02 EST

Figure 9-9. Temporal variati(m of mngc and intensity (different grcy tones) of 50 MHz backscattered power at Jicamarc~ ml 20–2 1 March 1977 (top
panel) and 137 MHz scintillations (bottt~m panel) over a nearly common ionospheric volume [ Basu et al., 1980].

9-1o
 IONOSPHERIC PHYSICS
over the past few years to detect irregularities shorter than [Tsunoda et al., 1979; Towle, 1980; Tsunoda, 1980b] and
the ion gyroraclius which is -5 m in the topside equatorial more recently the TRADEX radar at Kwajalein has been
ionosphere [Woodman and Basu, 19781 for radar system used to detect 11 cm irregularities which are approximately
applications tid for the understanding of the complex plasma 3 times the electron gyroradius and 30 times the Debye
processes in equatorial spread F. By the use of the ALTAIR length [Tsunoda, 1980c]. Thus equatorial spread F is found
radar at Kwajalein, Marshall Islands, irregularities with spa- to encompass irregularity wavelengths extending over 5-6
tial wavelengths of 1 m and 36 cm have been detected orders of magnitude.

9-11
CHAPTER 9
REFERENCES

Ahmed, M., R.C. Sagalyn, and P.J. L. Wildman,’’Topside Cormier, R. J., “Thule Riometer Observations of Polar Cap
Ionosphere Trough Morphology: Occurrence Frequency Absorption Events, ( 1962–1 972),” AFCRL TR-73-0060,
and Diurnal, Seasonal and Altitude Variations,” Y. Geo- 1973.
phys. Res., 84:489-498, 1979. Croom, S., A. Robbins, and J .0. Thomas, “A Review of
Appleton, E. V., “Two Anomalies in the Ionosphere,” Na- Topside Sounder Stlsdies of the Equatorial Ionosphere,”
ture (London), 157:691, 1946. Nature (London), 184:2003, 1959.
Appleton, E. V., “Two Anomalies in the Behavior of the DasGupta, A., J. Aarons, J.A. Klobuchar, S. Basu, and 1.
F2 Layer of the Ionosphere,” Some Ionospheric Results Bushby, “Ionsopheric Electron Content Depletions As-
Obtained During the ICY., edited by W .J .G. Beynon, sociated with Amplitude Scintillations in the Equatorial
p. 3, Elsevier, Amsterdam, 1960. Region,” Geophys. Res. Left., 9:147-150, 1982.
Baker, W.G. and D.F. Martyn, “Conductivity of the Ion- DNA Ratebook, Defense Nuclear Agency, U.S. Gover-
osphere, “ Nature, 170: 1090, 1952. nment Printing Office, Washington, D. C., 1972.
Balsley, B. B. and D.T. Farley, “Radar Observations of Doering, J. P., T.A. Potemra, W.K. Petersen, and C.O.
Two-Dimensional Turbulence in the Equatorial Electro- Bostrom, “Characteristic Energy Spectra of 1 to 500 eV
jet,” Y. Geophys. Res., 78:7174, 1973. Electrons Observed in the High Latitude Ionosphere from
Banks, P.M. and G. Kockarts, Aeronomy, Part B, Aca- Atmospheric Explorer C,” J. Geophys. Res., 81:5507,
demic Press, New York, 1973. 1976.
Basu, S. and Su. Basu, “Equatorial Scintillations—A Re- Donnelly, R. F., “Contribution of X-Ray and EUV Bursts
view,” J. Atmos. Terr. Phys., 43:473, 1981. of Solar Flares to Sudden Frequency Deviations,” J.
Basu, S., Su. Basu, J. Aarons, J.P. McClure, and M.D. Geophys. Res., 74: 1973, 1969.
Cousins, “On the Coexistence of Kilometer- and Meter- Domelly, R. F., “Extreme Ultraviolet Flashes of Solar Ffares
Scale Irregularities in the Nighttime Equatorial F Re- Observed Via Sudden Frequency Deviations: Experi-
gion,” J. Geophys. Res., 83:4219, 1978. mental Results,” Sol. Phys., 29: 188, 1971.
Basu, S., J.P. McClure, S. Basu, W.B. Hanson, and J. Dyson, P.L. and R.F. Benson, “Topside Sounder Obser-
Aarons, “Coordinated Study of Equatorial Scintillation vations of Equatorial Bubbles,” Geophys. Res. Lett., 5:
and In Situ and Radar Observations of Nighttime F Re- 795, 1978.
gion Irregularities,” J. Geophys. Res., 85:5119, 1980. Eccles, D. and J. W. King, “A Review of Topside Sounder
Booker, H.G. and H. W. Wells, “Scattering of Radio Waves Studies of the Equatorial Ionosphere,” Proc. IEEE, 57:
by the F-Region of the Ionosphere,” Terr. Magn. Atmos. 1012, 1969.
Elect., 43:249, 1938. Evans, J. V., “Incoherent Scatter Contributions to Studies
Bowles, K.L. and R. Cohen, “A Study of Radio Wave of the Dynamics of the Lower Thermosphere, ” Rev.
Scattering from Sporadic E. Near the Magnetic Equa- Geophys. Space Sci., 16: 195–216, 1978.
tor,” in Ionospheric Sporadic E, edited by E.K. Smith Farley, D. T., “A Plasma Instability Resulting in Field-Aligned
and S. Matsushita, Pergamon Press, New York, 1.962. Irregularities in the Ionosphere,” J. Geophys. Res., 68:
Brinton, H. C., J.M. Grebowsky, and L.H. Brace, “The 6083, 1963.
High Latitude Winter F-Region at 300 km: Thermal Farley, D.T. and B .B. Balsley, “Instabilities in the Equa-
Plasmas Observations from AE-C,” J. Geophys. Res., torial Electrojet ,“ J. Geophys. Res., 78:227, 1973.
83:47674776, 1978. Fejer, B., D.T. Farley, B.B. Balsley, and R.F. Woodman,
Buneman, O., “Excitation of Field Aligned Sound Waves “Radar Studies of Anomalous Velocity Reversals in the
by Electron Streams,” Phys. Res. Lett., 10:285, 1963. Equatorial Ionosphere,” J. Geophys. Res., 81: 4621,
Calvert, W. and R. Cohen, “The Interpretation and Syn- 1976.
thesis of Certain Spread-F Configurations Appearing on Fejer, B .G. and M .C. Kelley, “Ionospheric Irregularities ,“
Equatorial Monograms,” J. Geophys. Res., 66: 3125, Rev. Geophys. Space Sci., 18:401-454, 1980.
1961. Feldstein, Y.1. and G. V. Starkov, “Dynamics of Auroral
Carpenter, D. L. and C.G. Park, “On What Ionospheric Belts and Polar Geomagnetic Disturbances,” Planet. Space
Workers Should Know About the Plasmapause-Plas- Sci., 15:209, 1967.
masphere,” Rev. Geophys. Space Sci., 11: 133–1 54, Ferguson, E. E., “D-Region Ion Chemistry,” Rev. Geophys.
1973. Space Phys., 9:997-1008, 1971.
Chapman, S., “The Electrical Conductivity of the Iono- Gouin, P. and P. N. Mayaud, “A Propos de Existence Pos-
sphere: A Review,” Nuovo Cimento, 4( Suppl): 1385, sible d’un Contre-electrojet aux Latitudes Magnetiques
1956. Equatoriales,” Ann. Geophys., 23:41, 1967.
Chapman, S. and K.S. Raja Rae, “The H and Z Variations Gussenhoven, M. S., D.A. Hardy, and W.J. Burke,
Along and Near the Equatorial Electrojet in India, Africa “DMSP/F2 Electron Observations of Equatorward Au-
and the Pacific,” J. Atmos. Terr. Phys., 27: 559–581, roral Boundaries and Their Relationship to Magneto-
1965. spheric Electric Fields,” J. Geophys. Res., 86: 768,
Cohen, R., K.L. Bowles, and W. Calvert, “On the Nature 1981.
of Equatorial Slant Sporadic E,” J. Geophys. Res. 67: Gussenhoven, M. S., D.A. Hardy, N. Heinemann, and E.
965-972, 1962. Holeman, “ 1978 Diffuse Auroral Boundaries and a De-

9-12
 IONOSPHERIC PHYSICS
rived Auroral Boundary Index,” AFGL TR-82-0398, Rasmussen, J. E., P.A. Kossey, and E.A. Lewis, “Evidence
ADA130175, 1982. of an Ionospheric Reflecting Layer Below the Classical
Hanson, W. B. and R.J. Moffett, “Ionization Transport Ef- D Region,” J. Geophvs. Res., 85:3037-3044, 1980.
fects in the Equatorial F Region,” J. Geophys. Res., 71: Rastogi, R. G., “The Diurnal Development of the Anoma-
5559, 1966. lous Equatorial Belt in the F2 Region of the Ionosphere,”
Hanson, W. B. and S. Sanatani, “Large Ni Gradients Below J. Geophys. Res., 64:727, 1959.
the Equatorial F- Peak,” J. Geophys. Res., 78: 1167, Rastogi, R. G., “Counter Electrojet Currents in the Indian
1973. Zone,” Pianet. Space Sci., 21: 1355, 1973.
Hudson, R. D., “Critical Review of Ultraviolet Photoab- Rastogi, R. G., “Lunar Effects in the Counter-Electrojet Near
sorption Cross Section for Molecules of Astrophysical the Magnetic Equator,” J. Atmos. Terr. Phys., 36:167,
and Aeronomic Interest ,“ Rev. Geophys. Space Phys., 1974.
9:305406, 1971. Rastogi, R. G., “Seasonal and Solar Cycle Variations of
Hutton, R. and J.O. Oyinloye, “The Counter-Electrojet in Equatorial Spread-F in the American Zone ,“ J. Atmos.
Nigeria,” Ann. Geophys., 26:921, 1970. Terr. Phys., 42:593, 1980.
Jasperse, J. R., “Electron Distribution Function and Ion Ratcliffe, J. A., An [introduction to the Ionosphere and Mag-
Concentrations in the Earth’s Lower Ionosphere from netosphere, Cambridge University Press, 1972.
Boltzman-Fokker-Planck Theory,” Planet. Space Sci., Richmond, A. D., “Equatorial Electrojet, I. Development
25:743, 1977. of a Model Including Winds and Instabilities,” J. Atmos.
Kelley, M. C., G. Haerendel, H. Kappler, A. Valenzuela, Terr. Phys., 35:1083, 1973.
B.B. Balsley, D.A. Carter, W.L. Eklund, C.W. Carl- Richmond, A .D., and S. V. Venkateswaran, “Geomagnetic
son, B. Hausler, and R. Torbert, “Evidence for a Ray- Crochets and Associated Ionospheric Current Systems,”
leigh-Taylor Type Instability and Upwelling of Depleted Radio Sci., 6: 139, 1971.
Density Regions during Equatorial Spread-F,” Geophys. Rino, C. L., T. Tsunoda, J. Petriceks, R.C. Livingston,
Res. Lett., 3:448, 1976. M.C. Kelley, and K.D. Baker, “Simultaneous Rocket-
Kohnlein, W., “Electron Density Models of the Iono- Borne Beacon and in situ Measurements of ~uatorial
sphere,” Rev. of Geophys. Space Phys., 16: 341–354, Spread-F—Intermediate Wavelength Results,” J. Geo-
1978. phys. Res., 86:2411, 1981.
Lyon, A. J., The ionosphere, p. 88, Physical Society, Lon- Rino, C. L., R.C. Livingston, and S.J. Matthews, “Evi-
don, 1963. dence for Sheet-Like Auroral Ionospheric Irregulari-
Lyon, A.J, and L. Thomas, “The F2-Region Equatorial ties,” Geophys. Res. Left., 5: 1039, 1978.
Anomaly in the African, American and East Asian Sec- Roble, R. G., R.E. Dickinson, E.C. Ridley, and Y. Kamide,
tors during Sunspot Maximum,” J. Atmos. Terr. Phys., “Thermospheric Response to the November 8-9, 1969
25:373, 1963. Disturbance,” J. Geophvs. Res., 84:42074216, 1979;
Mantas, G. P., H.C. Carlson, and V .B. Wickwar, “Photo- “Thermosphere Circulation Models,” Eos, Trans. AGU,
electron Flux Build-up in the Plasmasphere,” J. Geo- 62:3, 19, 1981.
phys. Res., 83: 1–15, 1978. Sojka, J. J., W.J. Raitt, and R. W. Schunk, “A Theoretical
Martin, E. and J. Aarons, “F-Layer Scintillations and Au- Study of High Latitude Winter F Region at Solar Min-
rora,” J, Geophys. Res., 82: 2717–2722, 1977. imum for Low Magnetic Activity,” J. Geophvs. Res.,
Martyn, D. F., The Physics of the Ionosphere, p. 254, Phys- 86: 609–621 , 1981.
ical Society, London, 1963. Stolarski, R. S. and N .P. Johnson, “Photoionization and
Martyn, D. F., “The Normal F Region of the Ionosphere,” Photoabsorption Cross Sections for Ionospheric Calcu-
Proc. IRE N. Y., 47:147, 1959. lations,” J. Atmos. Terr. Phys., 34:1691, 1972.
Matsushita, S., “Intense E, Ionization Near the Magnetic Strascio, M.A. and B. Sellers, “The Calculation of Riom-
Equator,” J. Geomagn. Geoelect., 3:4446. eters Absorption and Approximate Connection Between
Matsushita, S. and B .B. Balsley, “A Question of DP-2,” Riometer Absorption and Solar Proton Fluxes during
Planet. Space Sci., 20: 1259, 1972. Night-Time PCA Events,” AFCRL TR-75-0469,
Matsushita, S. and W. H. Campbell (eds.), Physics of Ceo- ADA019656, 1975.
mgnetic Phenomena, Academic Press, New York, 1967. Sugiura, M. and J .C. Cain, “A Model Equatorial Electro-
McClure, J .P., W .B. Hanson, and J .H. Hoffman, “Plasma jet,” J. Geophys. Res., 71:1869, 1966.
Bubbles and Irregularities in the Equatorial Ionosphere,” Sugiura, M. and D.J. Poros, “An Improved Model Equa-
J. Geophys. Res., 82:2650, 1977. torial Electrojet with a Meridional Current System,” J.
Meng, C.-I., R.H. Holzworth, and W.-I. Akasofu, “Auroral Geophys. Res., 74:4025, 1969.
Circle Delineating the Poleward Boundary of the Quiet Titheridge, J. E., “Ion Transition Heights from Topside
Auroral Oval,” J. Geophys. Res., 82:164, 1977. Electron Density Profiles,” Planet. Space Sci., 24:
Morse, F. A., B.C. Edgar, H.C. Koons, C.J. Rice, W.J. 229–245 , 1976.
Heikkila, J.H. Hoffman, B.A. Timsley, J.D. Win- Torr, D. G., “Ionospheric Chemistry,” Rev. Geophys. Space
ningham, A. B. Christiansen, R.F. Woodman, J. Pom- Phys., 17:510-521, 1979.
alaza, and R.N. Teixeiro, “Equion: An Equatorial ion- Towle, D. M., “VHF and UHF Radar Observations of Equa-
ospheric Irregularityy Experiment, ” J. Geophys. Res., torial Ionospheric Irregularities and Background Den-
82:578, 1977. sities,” Radio Sci., 15:71, 1980.

9-13
CHAPTER 9
Tsunoda, R. T., “Magnetic-Field-Aligned Characteristics of Energy Distributiori of Auroral Electrons from Incoher-
Plasma Bubbles in the Nighttime Equatorial Iono- ent Scaker Radar Measurements ,“ Radar Probing of the
sphere,” J. Atrnos. Terr. Phys., 42:743, 1980a. Auroral’ Pla$w, edited by A. Brekke, pp 315-330,
Tsunoda, R. T., “On the Spatial Relationship of l-m Equa- Scandinavian University Books, Oslo, Norway, 1977.
torial Spread-F Irregularities and Plasma Bubbles,” J. Weber, E.J. j J. Buchau, R.H. Eathet,’ and R.S. Mende,
Geophys. Res., 85:185, 1980b. “North-South Aligned ~uatotial Airglow Depletions,”
Tsunoda, R. T., “Backscatter Measurements of 11-cm Equa* J. Geophys. Res., 83:712, 1978.
torial Spread-F Imegultities,” Geophys. Res., Lett., 7: Weber, E. J., J. Buchau, and J.G. Moore, “Airborne Studies
848, 1980c. of F Layer Ionospheric Irregularities ,“ J. Geophys. Res.,
Tsunoda, R. T., M.J. Baron, J. Owen, and D.M. Towle, 85:4631, 1980.
“Alti An Incoherent Scatter Radm for ~uatorid Sp@- Winninghm, J.D. and. W .J. Heikkila, “Polar Cap Auroral
F Studies,” Radio Sci., 14:1111, 1979. Electron Fluxes Observed with ISIS-I,” J. Geophys.
Turcoi R. P., “A Discussion of Possible Negative Ion De- Res., 79: 949–958, 1974.
tachment Mechanisms in the Sunrise D Region,” Radio Woodman, R.F. and C. LaHoz, “Radar Observations of F
Sci., 9:655, l~7A. Region ~uatorial Irregul@tieS,” J. Geophys. Res., 81:
Untiedt, J., “A Model of the Equatorial Electrojet with a 5447, 1976.
Meridional Cumnt System,” J. Ge@hys. Res., 725799, Woodrhan, R.F. and S. Basu, “Comparison Between in situ
1967. Spectral Measurements of Equatorial F Region Irregu-
Vickrey, J. F., C.L. Rino, and T.A. Potemra, “Chahn- larities and Backscatter Observations at 3-m Wave-
ika/TRIAD Observations of Unstable Ionization En- length,” Geophys. Res. Lett., 5:869, 1978.
hancements in the Auroral F-Regioo,” Geophys. Res. Wright, J. N., “A Model of the F2 Region Above h~ax F2,”
Lett., 7: 798–792, 1980. J. Geophys. Res., 65:185, 1960.
Vondrak, R.R. and M.J. Baron, “Radar Measurements of Yeh, K. C., H. Soicher, and C.H. Liu, “~biervations of
the Latitudinal Variation of Auroral Ionization ,“ Radio Equatorial Ionospheric Bubbles by the Radio Propaga-
Sci., 11: 939–946, 1976; “A Method of Obtaining the tion Method,” J. Geophys. Res., 84:6589, 1979.

9-14
 Chapter 10

IONOSPHERIC RADIO WAVE PROPAGATION

Section 10.1 S. Basu, J. Buchau, F.J. Rich and E.J. Weber
Section 10.2 E.C. Field, J.L. Heckscher, P.A. Kossey, and E.A. Lewis
Section 10.3 B.S. Dandekar
Section 10.4 L.F. McNamara
Section 10.5 E.W. Cliver
Section 10.6 G.H. Millman
Section 10.7 J. Aarons and S. Basu
Section 10.8 J.A. Klobuchar
Section 10.9 J.A. Klobuchar
Section 10.10 S. Basu, M.F. Mendillo

The series of reviews presented is an attempt to introduce in HF communications is leading to a rejuvenation of the
ionospheric radio wave propagation of interest to system global ionosonde network.
users. Although the attempt is made to summarize the field,
the individuals writing each section have oriented the work 10.1.1.1 Ionogram. Ionospheric sounders or ionosondes
in the direction judged to be most important. are, in principle, HF radars that record the time of flight or
We cover areas such as HF and VLF propagation where travel of a transmitted HF signal as a measure of its ionos-
the ionosphere is essentially a “black box”, that is, a vital pheric reflection height. By sweeping in frequency, typically
part of the system. We also cover areas where the ionosphere from 0.5 to 20 MHz, an ionosonde obtains a meas-
is essentially a nuisance, such as the scintillations of trans- urement of the ionospheric reflection height as a function
ionospheric radio signals. of frequency. A recording of this reflection height meas-
Finally, we have included a summary of the main fea- urement as a function of frequency is called an ionogram.
tures of the models being used at the time of writing these Ionograms can be used to determine the electron density
reviews. [J. Aarons] distribution as a function of height, N=(h), from a height
that is approximately the bottom of the E layer to generally
the peak of the F2 layer, except under spread F conditions
10.1 MEASURING TECHNIQUES or under conditions when the underlying ionization prevents
measurement of the F2 layer peak density. More directly,
ionosondes can be used to detemine propagation conditions
10.1.1 Ionosonde on HF communications links.
Two typical ionograms produced by a standard analog
For more than four decades, sounding the ionosphere ionospheric sounder using film recording techniques are shown
with ionospheric sounders or ionosondes has been the most in Figure 10-1. The frequency range is 0.25 to 20 MHz
important technique developed for the investigation of the (horizontal axis), and the displayed height range is 600 km,
global structure of the ionosphere, its diurnal, seasonal and with 100 km height markers. The bottom ionogram is typical
solar cycle changes, and its response to solar disturbances. for daytime, showing the signatures of reflections from the
Even the advent of the extremely powerful incoherent scatter E, F 1 and F2 layers. The cusps, seen at various frequencies
radar technique [Evans, 1975], which permits measurement (where the trace tends to become vertical) indicate the so-
of the complete electron density profile, electron and ion called critical frequencies, foE, foFl, and foF2. The critical
temperatures, and ionospheric motions, has not made the frequencies are those frequencies at which the ionospheric
relatively inexpensive and versatile ionosonde obsolete. On sounder signals penetrate the respective layers. These fre-
the contrary, modern techniques of complex ionospheric quencies are a measure of the maximum electron densities
parameter measurements and data processing [Bibl and of the respective layers. Since the densities vary with time,
Reinisch, 1978a; Wright and Pitteway, 1979; Buchau et al., ionospheric sounding is used to obtain information on changes
1978] have led to a resurgence of interest in ionospheric in the critical frequency and other parameters of the electron
sounding as a basic research tool, while a renewed interest density vs height profile.

1o-1
 x ox
km .*.,
700
600

500
400

300
200

100

NIGHT

km
600
500
400
300

200

100

.25 1 2 34 5678 9 10 5 20 MHz

DAY

Figure 10-1. Typical mid latitude day and nighttime ionograms, recorded by a C-4 ionosonde at Boulder, Colorado The dayt ime imrogmm shows retlcct itms
from E, Es, F1 and F2 layers; the nighttime ionogram those from Es and F2 layers.

The ionogram (Figure 10-1) shows signatures of various Finally, we see vertical bands in the frequency range
phenomena that complicate the process of ionospheric from 0.5 to 1.7 MHz, the signature of radio frequency
sounding or the ionogram analysis. Superimposed on the interference (RFI) in an ionogram, here from the AM band.
primary F layer echo trace is a similar but not identical RFI can become severe enough to prevent the recording of
trace, shifted up in frequency: the so-called extraordinary ionospheric echoes; for example, interference masks part or
or X component. The primary trace is called the ordinary all of the E layer trace below 1.7 MHz.
or O component. The echo trace is split into two traces due The top of Figure 10-1 shows a typical nighttime ion-
to effects of the earth’s magnetic field. A second trace sim- ogram. The E and F 1 traces have disappeared because these
ilar to the primary trace is seen at twice the range, a multiple layers dissipate after sunset. (Residual nighttime E region
reflection. Only a small fraction of the wave energy is re- ionization of low density can be observed in the absence of
ceived by the antenna after it has returned from the iono- low sporadic E layers at stations with low RF 1 and large
sphere. Most of the returned energy is reflected back from antennas. ) Echoes from a sporadic E layer (Es) and F2 layer
the ground and provides the first multiple (second order echoes and their multiples are clearly visible in Figure
echo) at twice the range. If the ionosphere is a good reflector, 10-1. At times a brushlike spreading of the F2 layer cusps
and losses in the D region are low, additional reflections is observed. It is called spread F and is caused by small
can be observed. Figure 10-1 (Night) shows a second mul- scale irregularities embedded in the ionosphere and ripples
tiple (third order echo) for part of the Es trace. It is easy in the equidensity contours on the order of hundreds of
to see that slopes increase by a factor that corresponds to meters to kilometers. For a detailed discussion of spread F
the order of the echo. see Davies [ 1966] and Rawer and Suchy [ 1967J; for a dis-

10-2
 a
IONOSPHERIC RADIO WAVE PROPAGATION
cussion of the occurrence and global distribution see Herman f. = 0.009 m (10.4)
[1966]. The nighttime ionogram also shows increased RFI
bands at higher frequencies. Because the D layer disappears N. = 1.24 X 104 f~ (10.5)
at night, HF propagation over large distances is possible.
This long distance propagation is heavily used for broad- where fN is in MHz and Nc in electrons/ctn3. The plasma
casting by commercial users and for shortwave radio com- frequency is the natural frequency of oscillation for a slab
munications by government services and radio amateurs. of neutral plasma with the density N, after the electrons
Fortunately the ionosonde’s own echoes also increase in have been displaced from the ions and are allowed to move
amplitude due to the disappearance of the D layer, reducing freely. For further discussions of the relation of K to the
to some extent the effect of increased propagated noise on wave propagation see Davies [1966].
the systems overall signal-to-noise ratio. Peak densities of the ionospheric layers vary between
104 and > 106 el/cm3. Inserting these numbers into Equation
10.1.1.2 Principles of Ionospheric Sounding. The con- (10.4) gives a plasma frequency range from 1 to >9 MHz;
cept of ionospheric sounding was born as early as 1924, this is the reason for the frequency range (O.5 MHz <f s
when Breit and Tuve [1926] proved the existence of an 20 MHz) covered by a typical ionosonde. The low densities
ionized layer with the reception of ionospheric echoes of of the D layer can only be probed with low frequencies
HF pulses transmitted at 4.3 MHz from a remote transmitter <250 Hz, requiring large antennas and complex processing/
(distance 13.8 km). This, during the next decade, led to the analysis techniques and are not directly measurable by the
development of monostatic ionospheric sounders by the Na- standard ionosondes (for details see Kelso [1964] and ref-
tional Bureau of Standards and the Carnegie Institution. erences therein). Indirectly the D region ionization is meas-
Even today the principles used by Breit and Tuves constitute ured by the integral absorption effects that it imposes on
the principles on which most ionospheric sounders are based. the HF waves propagating through it to the E or F region
These are the transmission of HF pulses and the measure- reflection levels (see discussion of fmin).
ment of their time of flight to the reflection level. For a The inclusion of the magnetic field in the formula for
short historical review of the development of ionospheric the refractive index leads to the well known Appleton dis-
sounders see Villard [1976]. persion formula (dispersion means that the refractive index
Ionospheric sounding takes advantage of the refractive depends on the propagating frequency) for a magnetized
properties of the ionosphere. A radio wave propagating into plasma, here given for the case of no collisions, generally
the ionospheric plasma encounters a medium with the re- valid for frequencies > 2 MHz, in the E and F regions.
fractive index (in the absence of the earth’s magnetic field
B, and ignoring collisions between electrons and the neutral F* =
atmosphere) 2X(1 – x)
l–
2 2(1– X)– YT2*V YT4 + 4 (1 – X)2YL2
~2=l_x=1_@
() f
(10.1)
(10.6)

where with

Nee2 eBLT
x= (10.2) YL.T = ~f (10.7)
4 m*Eomf*’
and
e, e. and m are natural constants, N, is the electron density, eB
—=f~ (10.8)
and f is the wave frequency. Below the ionosphere, N. =’ O, 2~m
and p = 1. Within the ionosphere, Nc > 0, and p < 1.
At a level where X = 1, where fH is the gyro frequency, the natural frequency at
which free electrons circle around the magnetic field lines.
Ne e2 BL,~ are the components of the magnetic field in the direction
G= —---= f2 (10.3)
4Tr2eOm of (longitudinal) or perpendicular to (transverse) the wave
normal. Inserting the constants into Equation (10. 8) leads
the refractive index p becomes zero. The wave cannot prop- to the useful relation for the gyrofrequency
agate any farther and is reflected. The quantity fN, which
relates the electron density to the frequency being reflected, fH = 2.8 X 104 B, (10.9)
is called the plasma frequency. Inserting the natural con-
stants into Equation ( 10.3) permits us to deduce the useful where fH is in MHz and B in gauss ( 1 gauss = Id tesla).
relation between electron density and plasma frequency (which The refractive index given in Equation (10.6) shows,
is identical to the probing frequency being reflected) by the t solution to the square root, that in a magnetized

10-3
CHAPTER 10
plasma two and only two “characteristic” waves can prop- detailed discussion see Davies [1966] and Chapter 10 of
agate. These two characteristic waves are called the ordinary Budden [ 1961].
or o-component and the extraordinary or x-component seen As a result, the actual reflection height h is smaller than
in the ionogram shown in Figure 10-1. A radio wave with the so-called virtual height h’, which is derived, assuming
arbitrary (often linear) polarization will split in the iono- propagation in the medium with the speed of light from
spheric medium into two characteristic or o-and x-compo-
nents, which in general propagate independently.
(10.15)
The reflection condition K = O gives two solutions for
X; for the + sign (o-component)
with t the round trip travel time of the pulse. Or since
X=l (10.10)
u < co, (10.16)
as in the no-field case, Equation (10.3); for the – sign (c-
omponent) then

X=1–Y. (10.11) h’ > h. (10.17)

At the reflection level for the O-component the plasma fre- As stated before, one of the main objectives of ionospheric
quency equals the probing frequency, f~ = f. The x-com- sounding is the determination of h(f), which through the
ponent is reflected at a lower level that depends on the local relation between f and Nc, Equations (10.3) and (10.4) rep-
magnetic field strength. It can be shown that the critical resents the desired function N=(h). Since the group travel
frequencies fo and fx, for fo > f~, are related by time is

f“
fx–fo=j, (10.12) h dz h
t=2—=2
J
cook co J o
p’dz, (10.18)

that is, the magneto-ionic splitting (due to the presence of

the magnetic field in the ionospheric plasma) depends on the virtual height is related to the group refractive index by
the local magnetic field strength and therefore varies, from
station to station. For a typical midlatitude station, B = 0.5G
and from Equation (10.9) we determine fH = 1.4 MHz, h’(f) = ~h(’) A’ [f~(z), fl dz. (10. I9)
leading to the fo-fx separation of -0.7 MHz seen in Figure ,)
10-1. A solution X = 1 + Y exists for frequencies below
the gyrofrequency f“. For details see Davies [1966]. If the electron density N=(h) is considered as a function of
Using ionograms to determine the true height electron the height h above the ground, p’ is also a function of h
density profile Nc(h) is further complicated by the slowing- and the problem is now to solve the integral equation (10. 19),
down effect that the ionization below the reflection level for given values of h’(f) obtained from the ionogram. The
has on the group velocity of the pulse. While the phase techniques used to solve this equation are known as true
velocity v of the wave is height analysis for which in general numerical methods are
used; they are discussed in detail in a 1967 special issue of
c,, Radio Science.
v.— (10.13)
~(h,f)

it can be shown that the group velocity u, defined as the

propagation velocity of the pulse envelope, is given for the 10.1.1.3 Analog Ionosonde. The general principle of an
no-magnetic field case by ionospheric pulse sounder is shown in Figure 10-2 [Rawer
and Suchy, 1967]. A superheterodyne technique is used to
c,, both generate the transmitted pulse of frequency f~ and to
u=—
= ~(h,f)co, (10.14)
p’(h,f) mix the received signals back to an intermediate frequency
or IF for further amplification. Tuning the receiver mixer
where p’(h,f) is the group refractive index. Therefore, while stage so that its output frequency is equal to the frequency
the phase velocity increases above the speed of light in a of the fixed frequency (pulsed) oscillator fc, and using a
plasma , the group velocity, the velocity at which the energy common variable local oscillator f,,, ensures that the receiver
propagates, slows down (p < 1 in a plasma). For a more and transmitter are automatically tuned for every value of

10-4
 a
IONOSPHERIC RADIO WAVE PROPAGATION

m ----
.--,
----

Receiver mixer
II
II
II
II

fi
if-
<

Amphfier
fi
Defec}or and
Udeo -Amphfier

II I
t
Helghf markers+ Recepfion I

II
II I
II
II
Pu/se - and c‘ +
II ‘YiYw/mth “
II 6enera/or ffecorder

Vor/ob/e osclllotor . t~!

~h Ior/zonfa/ deflection Synchro drive
Y
--

Figure 10-2. Schematic presentation of the major components of an Ionospheric Pulse Sounder

the oscillator frequency f.. The transmit pulse is amplified in synchronism with the transmission and pulling a film
in one” or several power stages and transmitted, using a slowly in the direction of the X-axis in the focal plane of
suitable wide-band antenna with a vertical radiation pattern. an imaging optic results in an ionogram recording such as
The same antenna can be used for reception using either a that shown in Figure 10-1. Since ‘sounders based on the
tap (voltage divider) of the transmitter tank circuit or a general concepts shown in Figure 10-2 rely completely on
Transmit/Receive or T/R switch, which protects the receiver analog techniques for signal generation, reception, and pro-
input from overloading during transmission of the pulse, cessing, they have more recently become known as analog
especially problematic for transistorized receivers. More re- sounders, in contrast to the digital sounders developed in
cently separate antennas for transmission and reception have several places during the last decade. Analog sounders are
become commonplace. This permits the use of smaller and still operated at many ionospheric observatories, especially
therefore less costly receiver antennas in phased arrays for the well-known C3 and C4 ionosondes, which were devel-
angle-of-arrival measurements and as polarized antennas for oped by NBS and which were distributed worldwide as the
polarization or mode ideritification [Bibl and Reinisch, primary ionosonde for the International Geophysical Year
1978a,b; Wright and Pitteway, 1979]. 1957/58.
The received signals are mixed down (or up) to the
intermediate frequency and amplified in m IF amplifier, that
is matched in bandwidth to the pulse width (overall bandwith
B = I/P, with P the width of the transmitted pulse). After 10.1.1.4 Analog/Digital Hybrid Ionosonde. While ver-
detection and amplification, the video signal modulates the tical sounding with the transmiuer and receiver and their
intensity of the CRTs electron beam (Z-axis modulation). respective antennas collocated made synchronization of
Deflecting the beam in the Y-axis with a sawtooth voltage transmitter and receiver relatively easy, a much more de-

10-5
CHAPTER 10
manding task arose when investigators attempted to sound ture of an ionogram trace simultaneously with the digital
the ionosphere over paths of varying distances to determine information. Preprocessing has largely eliminated the noise
the mode structure and the propagation conditions directly. background. The bottom part of Figure 10-3 shows a digital
If the transmitted signal is to be received within the amplitude ionogram, represented by all amplitudes above a
receiver bandwidth, the systems must be started at a precise noise level determined automatically and separately for each
time, and must have perfectly aligned frequency scans. This frequency. The noise level on each frequency can be esti-
was achieved using linear frequency scans and synchronous mated, since the unmodified signals of the lowest four height
motor drives, which derived their A/C voltage from crystal bins are shown at the bottom of the ionogram. The displayed
oscillators IBibl, 1963]. A large step forward was the de- range starts at 60 km and in 128 height increments with a
velopment of frequency stepping sounders such as the Gran- A = 5 km covers the range to 695 km. Each frequency
ger Path Sounder IGowell and Whidden, 1968] which com- step is in 100 kHz, which covers the range from (nominally)
bined digital and analog techniques. Digital techniques O to 13 MHz in 130 frequency steps. Ionograms of this type
generated ionograms by stepping synthesizer/transmitter and can be produced in between 30 s and 2 rein, depending on
receiver through the desired frequency range, providing se- the complexity of signal characterization selected. The num-
lectable frequency spacing (for example, 25, 50, or 100 ber of integrations required to achieve an acceptable signal-
kHz, linear or linear over octave bands). The frequency to-noise ratio, and the desired spectral resolution of the
synthesis itself and the data processing/recording however, Doppler measurements also affect the duration of the ion-
used the standard analog techniques. All digital and hybrid ogram sweep. The ionogram is similar in structure to the
pulse sounders currently available use these frequency step- daytime ionogram in Figure 10-1, showing clearly an
ping techniques. E-trace (foE = 3.25 MHz), an F1-cusp (foFl = 5.0 MHz),
and the F2 trace (foF2 = 8.2 MHz). The top part of the
10.1.1.5 Digital Ionosonde. The rapid development of figure was produced by printing only those amplitudes which
integrated circuits, microprocessors and especially Read- had a STATUS indicating o-polarization, vertical signals
Only-Memories, and of inexpensive storage of large ca- only. The resulting suppression of the x-component and of
pacity, has led to the development of digital ionospheric the (obliquely received) noise shows the effectiveness of
sounders. These systems have some analog components, these techniques.
but use digital techniques for frequency synthesis, receiver The digital “HF Radar System” developed at NOAA.
tuning, signal processing, recording, and displaying of the Boulder, Colorado [Grubb, 1979] is an ionospheric sounder.
ionograms. However, to the modem sounder, the digital built around a minicomputer. Appropriate software allows
control of all sounder functions, the ability to digitally con- freedom in generating the transmit signal phase coding and
trol the antenna configuration, and above all, the immense sequence, and in processing procedures. However, instruc-
power of digital real time processing of the data prior to tion execution times of the minicomputer limit this flexi-
recording on magnetic tape or printing with digital printers bility. The sounder with its present software uses an echo
are of special importance. detection scheme rather than a fixed FRB grid to obtain the
A, digital amplitude ionogram, recorded by a Digisonde information on the ionospheric returns. This scheme re-
128 PS at the AFGL Goose Bay Ionospheric Observatory quires that the return has to be identified beforehand, gen-
is shown in Figure 10-3. This system developed at the Uni- erally using a selectable level above the noise, and an “a
versity of Lowell [Bibl and Reinisch, 1978a,b] uses phase hits out of b samples” criterion. This system, by the use of
coding, spectral integration, polarized receive antennas for a receive antenna array, then determines on-line for this
o/x component identification, and fixed angle beam steering initially identified return, the echo amplitude, its polariza-
of the receive antenna array for coarse angle of arrival meas- tion, Doppler shift, and reflector location [Wright and Pitte-
urements to provide a rather complete description of the way, 1979].
properties and origin of the reflected echoes. Using a stan- The spectral information available in digital ionograms
dard set of 128 range bins for each frequency, the sounder has been used to track moving irregularities in the equatorial
integrates the sampled receiver output signals for a select- [Weber et al., 1982] and polar ionosphere [Buchau et al.,
able number of integrations, improving the signal-to-noise 1983]. An example of a Doppler ionogram recorded in the
ratio and providing the samples for spectral analysis. Since polar cap is shown in Figure 10-4. The right lower panel
for each frequency-range-bin or FRB only one return is shows a heavily spread amplitude ionogram and superim-
recorded, a search algorithm determines from the set of posed two oblique backscatter traces. The right upper panel
separate signals (o, x, several antenna directions, Doppler shows the Status or Doppler ionogram: each FRB displays
lines) the signal with the largest amplitude and retains am- the Doppler bin number instead of an amplitude. FRB’s
plitude and STATUS, that is, special signal characteristics. with an amplitude below an automatically determined noise
Using a special font [Patenaude et al., 1973], the resulting level show neither amplitude nor status. Separating the ion-
digital amplitudes are printed out providing the analog pre- ogram into positive and negative Doppler ionograms permits
sentation essential for the recognition of the detailed struc- the identification and subsequent tracking of approaching

10-6
 IONOSPHERIC RADIO WAVE PROPAGATION
.- .-
.- .
.:, .
,“
.! - .
-

: VERTICAL O-SIGNALS - : :- .
. .
.-
.-
.- ,
. . 7. . :.-, . .:

::. .

~ .

,.

.-
.- ..
., -
...-
. -,,
: :. — :- .“
-. ..,
. . .- .
.
.: -
200-: “. - .:
~. -
..“ - .:- . ...- . .. .
k
.- . .-
:. - ...-;. =
.
. ..
..- -- ~ -..
.- .
:. :- --
-. . . .. -----
.,-.
:.; _: -
---- --- —. . -.
loo-~- --
...-
, -- . --,.
..F
.. --. .. .
.- -
.-. . - :--- .,- : -: ..- .-
. .- ---.-
.- . .. . . - L~-L
.-m- .
.- -
g:~+-7- ..
..- .z -.. -.” .
. -.:-:. = .
.
ALL SIGNA-LS‘ -:: ‘! ‘; : ; _,;,,
- ,. ‘_
.. :,.
. . ..

:-:::=
...-..-.:. .
.-- ..:. ;..- .
.. -;----
,,,, ..::-; .: .-.
,..,, ,..
~oo=; .
..:
..;
.!
.*

,. :
300-; . -. .
..-
..- .. . -. .
., :: . .- .
“..-. ... . .
: .7<

200–: -.. z-
. .-- ._., - .
------ -.”

.-
.-

1234 56

Figure 10-3. Digital daytime amplitude ionogram recorded by a Digisonde 128PS at the AFGL Goose Bay Ionospheric Observatory 16 June 19801720
AST. Coarse angle of arrival and polarization information is used to separate the vertical ordinary trace shown in the upper part of the figure.
CHAPTER 10

::.,:,::F+#$:;;.: ,<...,,pm
f*:,-*:,
--T’ - ~+. ..:3+~g.”* : ‘
,oo_::::””:~* .~~:~~j
NEGATIVE , &.;s—% -
“’ ~“’’’~”h~”h~’
-;:: .........!.3”” ~–~~:.k!..!! ...... . . ~
~~”-’:yi~~>-
-’ * :-.=-
400 —...- ..... .. .... . ........... ...r~ - .—.. ,..–..–– —.,...-.—. -.::::: $-: ,>.~:. ‘ ‘ ‘

..-.—. . ... ........... .......... . m:::%’:’-’”

.... .... ..l ......fl .....! ....+ .. ........ ... . . . .. ..... . ._~!. -----3--- ‘------ ‘. 2 ..... ............ . .

... ...............l+’ -------- . ....~ “:...* ..... ..... .

‘-ER e:...::: :*”:’,””””””” ““’”s
200 ------i"-----------------------:-------*------.----------.--.---.-.--;=-. ..- ------!--------- —;~—-------------;:+. i--------- --- , ----...:, ......... -- .....!.... --
.. .
-.:.....—. -* . .,!,. i “ ............................. . --- ~i...— ..............~.-$-+ ..+ l..!... ..” .
. .
- * . * .“””””””~””””’””” ““” ““””””’”””’””

DOPPLER-SPLIT
AMRITUDE AND
IONOQRAMS
:- STATUS IONOGRAMS

AM~lTUDE
POSITIVE
.......—...——..<
......... ..r—.— ___________

DHER

_.&T .......... ._-..?.- .......-!.— ....__ ....___...

! i%

Figure 10-4, Amplitude/status ionogram taken by the AFGL airborne ionospheric observatory with a Digisonde 128PS at Thule, Greenland 9 December
2231 (UT). The lower right panel shows the amplitude ionogram after removal of radio noise. The Doppler ionogram shown in the upper
right pane] is produced by replacing each amplitude in the ionogram below with a number representing the measured Doppler shift. The
separation into positive and negative Doppler traces (approaching and receding reflection regions) is shown in the two panels on the left.

(traces marked B and C) and receding (trace marked A) the frequency extent of E and Es traces shows the typical
reflecting or scattering centers. The overhead trace (very cos X (X = solar zenith angle) pattern of the solar E-layer,
low Doppler) is marked O. maximizing at noon. Sporadic E events observed on all three
nights are typically observed at these high latitudes during
10.1.1.6 Digital Data Processing. The availability of auroral storms [Buchau et al., 1978].
ionograms in digital form has finally provided the basis for
successful automatic processing of these complex data. Real 10.1.1.7 FM/C W or Chirp Sounder. The availability
time monitoring [Buchau et al., 1978], survey of large data of very linear sweep-frequency synthesizers resulted in the
bases [Reinisch et al., 1982a], real time analysis of iono- development of FM/CW (frequency modulated continuous
spheric parameters [Reinisch et al., 1982b], automatic trace wave) or Chirp Sounders, initially for oblique incidence and
identification and true height analysis [Reinisch and Xue- in the 1970s also for monostatic vertical incidence sounding
quin, 1982, 1983] have been made possible by the availa- [Barry, 197 1]. A linear waveform with the constant sweep
bility of data in digital form. Analysis concepts for angle- rate df/dt is transmitted. Receiving the waveform after prop-
of-arrival determination and other parameters for the agation to the ionosphere and back and measuring the time
NOAA/SEL digital sounder have been presented by Wright delay of each frequency component against the original
and Pitteway [1982] and by Paul [1982]. waveform permits the determination of the travel time as a
An example of a data survey presentation using digital function of frequency. This is actually done by mixing the
ionosonde data from Goose Bay is shown in Figure 10-5. received waveform with the original, resulting in a differ-
The top row shows the integrated height characteristic, ence frequency which can be measured by spectrum anal-
obtained by collapsing each ionogram onto its height axis. ysis. The difference frequency as a function of frequency
This characteristic provides the history of E and F layer (the “Chirp”) is proportional to the travel time of the signal
(minimum) height variations over the course of three days. as a function of frequency; therefore, a graph of the dif-
The middle panel shows the temporal changes of the F layer ference frequency as a function of time or transmitted fre-
returns, with the lower envelope determined by foE (day- quency, through the known sweep rate dfldt, forms an ion-
time) or fmin (nighttime), while the upper envelope is de- ograrn. While initially transmitter and receiver were separated
termined generally by foF2. The bottom panel representing by a substantial distance, to avoid overloading of the re-

10-8
 =

IONOSPHERIC RADIO WAVE PROPAGATION

KM

600

500

-400

.300

-200

.100
MHZ
-13.0
-12.0
11.0
10.0
.9.0
-8.0
-7.0
.6.0
5.0
-4.0
-3.0
-20

*
“ 13.0
-12.0
-11.0
-100
- 9.0
–6.0
– 7.0
-6.0
- 5Q
+ 4.0
-3.0
2.0
L 1.0

Mm

Figure 10-5 Ionospheric characteristics spanning three days produced from digital ionogmms recorded at Goose Bay 28–30 April 1980. The intcgrtitcd
height characteristic shows the dynamic changes of the minimum height of the F layer and the apparance of the solar and spomdic E Iaycrs.
The F and E frequency characteristics show the diurnal variability of these layers as well as evidence of some auroral events.

ceivers with the unwanted direct signal, a monostatic system sweeprates of 20 kHz/s). This bandwidth is further reduced
was developed, using a T/R switch and a quasi-random by spectrum analysis to an effective bandwidth of the order
interruption of the linear waveform transmission. The main of lHz.
advantage of the FM/CW system is the very narrow instan- Although the digital integration and spectral analysis
taneous bandwidth of the transmitted signal, allowing a used in the modem digital pulsed ionosondes decreases the
similarly narrow receiver bandwidth (nominal 100 Hz at effective bandwidth of a pulse receiver significantly (by a

10-9
CHAPTER 10

rrT”T~!7”’” ““’””;’’”!
T-l

fltiz 28 26 24 22 20 18 16 14 12 10 8 6 4 2

figure 10-6. ~/CW or chirp-oblique propagation ionogram over a 2400 km path. Transmitted from an airerrrft over the North Atlantic and recorded at
a receiver site in Maine.

factor between 10 and 100, depending on the integration Alouette I, was launched into space on 29 September 1962.
time and processing procedures) from the original 20 kHz Since then a series of follow-on improved topside sounders
bandwidth for a typical 50 ps pulse and makes the actual were launched, either as orbiting ionospheric sounders (Al-
signal-to-noise ratio comparable to that of a chirp sounder, outie 11, 1965) or integrated with a complement of in situ
the FM/CW system allows a substantial reduction of the measuring geophysical instruments (ISIS I, 1969 and ISIS
peak power of the transmission. Good FM/CW ionograms 11, 197 1). For details on these sounders see Franklin and
have been obtained with transmit power as low as 1W (CW). MacLean [1969].
The FM/CW system is definitely a good solution for the Since groundb~ ionosondes obtain ionospheric echoes
already strained occupancy of the HF spectrum.
However, only up to the peak of the layer with the highest electron
several drawbacks have limited its application as a vertical density (generally the F2 layer), knowledge about the top-
sounder: motion of the reflecting region imposes a Doppler side ionosphere was based on theory, the few incoherent
shift, which introduces range errors that are intolerable un- scatter facilities, and sporadic high-altitude rocket flights.
der arctic conditions. The important Doppler shift of the The topside sounders therefore filled a gap in the techniques
echoes cannot be obtained. However, due to their low trans- available to measure the global ionosphere in its entirety.
mit power and superior S/N characteristics, many FM/CW During the approximately 15 years that these topside soun-
systems are used as backscatter sounders and as oblique ders have provided continuous data, the many millions of
incidence or propagation sounders. A typical propagation ionograms that have been accumulated have contributed
ionogram showing extensive mr.dtihop structure is shown in substantially to our picture of the global ionosphere [Chan
Figure 10-6. The ionogram was obtained over a 2400 km and Colin, 1969; Thomas and Rycroft, 1970] and of its
path with the transmitter located on board an aircraft and subregions, such as the high-latitude ionosphere [Jelly and
the receiver located at the AF OTH-B radar site at Columbia Petrie, 1969; Thomas and Andrews, 1968], the F region
Falls, Maine, U.S.A. trough IMuldrew, 1965; Pike, 1976], and the equatorial
ionosphere IEccles and King, 1969]. The topside sounder
10.1.1.8 Topside Sounders. Shortly after the onset of provides an h’f trace from the satellite altitude down to the
the satellite era, an ionospheric topside sounder satellite, peak of the F layer (except for those rare occasions where

1o-1o
 a
IONOSPHERIC RADIO WAVE PROPAGATION

G.c s?
4fB “ foF2 fxF2

0!5 0 i.5 3:5 4:5 5;5 6.5 ?0 75 8:5 9:5 10.5 li.5
FREQUENCY (Me/s)
DAY 319 ( 15 NOVEMBER 1962) 0731 /10 GMT ( 138”E , 30°S )
SATELLITE HEIGHT 1011 Km

Figure 10-7. An Alouette I topside ionogram illustrating Z-, O- and X-wave traces, cutoffs, resonance spikes, and earth echoes

ionization of lower layers exceeds the maximum density of therefore the final appearance of the h’ f-trace. This trace
the F2-layer). A typical topside ionogram is shown in Figure is sometimes further complicated by ionospheric irregu-
10-7 from the IJRSI Handbook of Ionogram Interpretation larities and oblique returns. All these factors combined en-
and Reduction [UAG-23, 1972]. A unique phenomenon sure an incredible variety of ionograms. To capture their geo-
observed in topside ionograms are the ionospheric resonance physically significant parameters, a large number of rules
spikes due to the excitation of the ambient plasma by the and definitions have evolved over the decades, which after
transmissions. The most frequently observed resonance spikes acceptance by the International Radio Science Union (URSI)
occur at the (loCal) plasma freqUenCy fN, at the 10Cal gy - have been published as the URSI Handbook of Ionogram
rofrequency f“ (labeled fB in Figure 10-7), at the hybrid Interpretation and Reduction, [UAG-23, 1972] governing
frequency the analysis of ionograms at all ionospheric stations.
This set of rules, resulting from the still continuing or ter-
fT=- (10.20)
minated operation of more than 300 ionosonde stations
and at certain harmonics of these frequencies [Hagg et al., distributed over the whole globe, has produced a rather uni-
1969]. form analyzed data base which is archived at the World Data
Many of the references given here and a large amount Centers for Solar Terrestrial Research located at Boulder,
of further material can be found in the special issue on Colorado (WDC A), lzmiran, USSR (WDC B), Tokyo,
topside sounding of the Proceedings of the IEEE [1969]. Japan (WDC C 1) and Slough, UK (WDC C2). With some
exceptions, the individual world data centers store data orig-
10.1.1.9 Ionogram Interpretation. The behavior of the inating in their respective regions. WDC A stores the data
ionosphere is often very dynamic. This fact and the large from the western hemisphere and also data from France and
range of electron densities, over which the ionospheric lay- India.
ers change from day to day, from day to night, with season To provide special instructions for the analysis of the
and with solar cycle result in a large variety of ionograms. extremely complex ionograms from high latitude stations.
There are also extreme differences in ionospheric variations a High-Latitude Supplement to the tJRSl Handbook on lon-
and structures from the equators to the poles and in the ogram Interpretation and Redliction has been published
regular or sporadic appearance and disappearance of the [UAG-50, 19751.
lower layers. Dynamic effects that shape the profile along For special research efforts. it is often essential to go
the ray path and specifically in the vicinity of the reflection back to the source data, the ionospheric films of a specific
region also affect the group delay at each frequency and station(s). For the western hemisphere, these films are stored

10-11
CHAPTER 10
at the World Data Center A for Solar Terrestrial Physics,
km
NOAA/NGSDC, Boulder, Colorado. A Catalogue of ion-
600
osphere Soundings Data [UAG-85, 1982] provides access 500
to this data base, which spans the period from 1930 through 400
today. The longest and still continuing operation of an io- 300

nosonde station started at Slough, UK in January 1930. 200 r

Continuous operation starting before 1940 is still ongoing I 00 - h’E

tronsmltted
at Canberra, Australia (1937); Heiss Island, USSR (1938); o ––––––——~—~~ ‘,–~,—,1
riTi TITiL Signal
Huancayo, Peru (1937); Leningrad, USSR (1939); Tomsk,
USSR (1937); and Tromso, Norway (1932).
To provide an overview of some of the more important
ionospheric parameters that can be derived from an iono-
gram and introduce their geophysical meaning, two iono-
fbEs fxEs
grams are provided in the form of a sketch (Figure 10-8), 14 16IS MHz
and the parameters are identified. Both ionograms depict 20 -

the same ionospheric conditions (taken from Figure 10- 1) 500 -

400 -
with the exception of an Es layer that can suddenly appear,
300 -
possibly as the result of a windshear at E layer heights. This
200 .
Es layer can obscure parts of the trace from reflections at I 1 h’Es
I00 —
higher regions of the ionosphere. A list of parameters and
0 ––——–—I–—T–T—I– rrr[~ rl TI rl u
their identification and interpretation is provided here as a 2 3 4 5 678910 15 I MHZ

general reference and not as a guide for ionogram analysis.

For detailed instructions in the evahtation of ionograms please Rgure 10-8. Line sketch of daytime ionogram shows definition of im-
refer to the URSI Handbooks UAG-23 and UAG-50. portant ionogram parameters.

Parameter MeaninglComments

a) Critical and characteristic frequencies

foF2 F2 layer ordinary wave critical frequency. A measure of the maximum density Ne~~, of this
layer [see Equation (10.5)].

fxF2 F2 layer extraordinary wave critical frequency. Can be used to infer foF2 using Equation (10. 12)
if foF2 is obscured by interference.

foil F1 layer ordinary wave critical frequency. This layer is often smoothly merging with the F2 layer
resulting in the absence of a distinct cusp and in difficulties of determining the exact frequency
(L condition).

foE solar produced E layer ordinary wave critical frequency.

Comment: Extraordinary wave returns exist for all layers. However, absorption of the
extraordinary component is stronger than that of the o-component and the x-trace of the E layer
is rarely, that of the F 1 layers not always observed.

fbEs Es layer blanketing frequency. Returns from higher layers are obscured by the Es layer up to this
frequency. This frequency corresponds closely to the maximum plasma density in the (thin) Es-
layer [Reddy and Rae, 1968].

fxEs Highest frequency at which a continuous Es trace is observed.

foEs foEs can be inferred, applying Equation (10. 12). If fbEs < foEs, the layer is semitransparent. Es
and higher layers are both observable. The determination of foEs and fxEs for all cases is subject
to a complex set of rules beyond the scope of this outline (see URSI Handbook on Ionogram
Interpretation). Modern Sounders, using polarized receive antennas, permit unambiguous foEs
determination.

10-12
 w

IONOSPHERIC RADIO WAVE PROPAGATION

Parameter Meaning/Comments

~~i~ Minimum frequency at which returns are observed on the ionogram. Since radio wave energy is
absorbed in the D region according to an inverse square law (Absorption - 1/P), the variation of
fmin is often used as a coarse indicator of the variation of D region ionization. f r.in is not an
absolute value (as for example foF2), but depends directly on the transmitted power and the
antenna gain. Comparison between stations, therefore, can be only qualitative.

b) Virtual heights
h’F The minimum virtual height of the ordinary wave F trace taken as a whole. Due to the effects of
underlying ionization and profile shape on the travel time of the pulse, these minimum virtual
heights are only useful as coarse and “relative” height classifiers (high, average, or low layer,
compared to a reference day). True height analysis must be made to give more meaningful height
parameters, such as the height of the layer maximum (h~.,F2).

hrF2 The minimum virtual height of the ordinary wave F2 layer trace during the daytime presence of
the F1 layer. When an F1 layer is absent, the minimum virtual height of the F2 layer is h ‘F,
defined above.

h’E The minimum virtual height of the normal E layer, taken as a whole.

h’Es The minimum virtual height of the trace, used to determine foEs.

hpF2 The virtual height of the ordinary wave mode F trace at the frequency 0.834 x foF2. For a
single parabolic layer with no underlying ionization this is equal to the height of the maximum of
the layer, h~.X. In practice hpF2 is usually higher than the true height of the layer maximum.
Useful as a rough estimator of h~,. but strongly affected by a low foF2/foFl ratio (< 1.3).

MUF(3000)F A set of “transmission curves” [Davies, 1966 and 1969] developed for a selected propagation link
distance (the URSI standard is 3000 km) permits the determination of the Maximum Usable
Frequency, which the overhead ionosphere will permit to propagate over the selected distance.
The MUF is detetiined from the estimated transmission curve tangential to the F-trace. For this
ionogram MUF(3000)F would be 17.0 MHz.

10.1.1.10 Ionosonde Network. Even though the rou- has been incorporated into INAG as of September 1984.
tinely operating ionosondes form;ng the worldwide network The INAG bulletin can be obtained from the World Data
are independent, generally operated as subchains or as in- Center A, Boulder, Colorado, 80303.
dividual stations by national or private organizations, their With the advent of modem digital ionosondes and on-
operation is coordinated by the “Ionospheric Network Ad- site automated processing, a carefully planned network of
visory Group (INAG)”, working under the auspices of Com- remotely controllable ionosondes can provide ionospheric
mission G (On the Ionosphere), a Working Group of the data and electron density profiles to a control location for
International Union of Radio Science (URSI). INAG pub- real time monitoring of ionospheric and geophysical con-
lishes the “Ionospheric Station Information Bulletin” at vary- ditions. Automatic oblique propagation measurements be-
ing intervals. The Bulletin provides a means of exchanging tween stations of the link can increase manyfold the number
experiences gained at the various ionospheric stations, dis- of ionospheric points that can be monitored. Considerate ions
ctisses in detail difficult ionograms for the benefit of all for the deployment of a modem ionosonde network have
participants, and disperses information on new systems, new recently been presented by Wright and Paul [i981 ]. Op-
techniques, special events (for example, eclipses), reievant erational and technical information on the individual stations
meetings, and g?neral network news. URSI’S International of the world wide network of ionosondes, as well as their
Digital Ionosonde Group (IDIG), which provides a forum respective affiliations and addresses, are available in the
for the disctission of standardization proposals, for the ex- Directo~ of Solar Terrestrial Physics Monitoring Stations
change of software, and for the general exchange of ex- [Shea et al., 1984]. Figure 10-9, taken fr~m the report in
periences with these rather new and still maturing systems preparation, shows the locations of all ionosondes reported

10-13
CHAPTER 10
180 150W 120W 90W 60W 30W o 30E 60E 90E 120E 150E 180
1 I 90
I I I I L I I I I I II

60
I

x
L — 30

— —+
0

30
I

-–t-–t-t–”-’+–}-+ e
60

~ --*
90

Figure 10-9. Map of vertical incidence ionospheric sounder stations 1984.

as operational or operating in 1984. World Data Center A per unit volume to be Nu,, where N is the electron number
Report, UAG-85, lists all past and present ionospheric ob- density. He also predicted that the spectrum of the scattered
servatories. signal will be Doppler broadened by the random electron
thermal motion. The spectrum of the scattered signal was
expected to be Gaussian with center to half-power width of
0.71 Af. where Af, is the Doppler shift of an electron ap-
iO.1.2 Incoherent Scatter proaching the radar at mean thermal speed so that

J.J. Thomson [1906] showed that single electrons can 1/2

1 8kTe
scatter electromagnetic waves, and that the energy scattered
by an electron into unit solid angle per unit incident flux is
given by (r. sin$)2 where r. is the classical electron radius
Af. =l
()—
m
Hz, (10.21)

( = e2/&.uc2 = 2.82 X 1~” m) and ~ is the polarization

angle, that is, the angle between the direction of the incident where k is the radar wavelength (m), k is Boltzmann’s
electric field and the direction of the observer. Thus the constant ( = 1.38 x 10-23 J/K), T, is the electron temper-
radar backscatter (+ = m/2) cross-section of a single elec- ature, and Q is the mass of an electron ( = 9.1 X l@3’
tron will be U. = 4rr~. Gordon [1958] first proposed that kg). At a wavelength k = I m, and Tc = 1600 K, 0.71 Af,
by the use of a powerful radar operating at a frequency = 200 kHz. Soon after Gordon [1958] proposed the fea-
f > foF2 where foF2 is the plasma frequency at the peak sibility of the incoherent scatter radar experiment to study
of the F2 layer, the backscattered power from the electrons the upper atmosphere, Bowles [ 1958] was able to detect
in the upper atmosphere should be detectable. The meas- radar echoes from the ionosphere. The echoes resembled
urement of scattered power and its characteristics as a func- the predicted ionospheric scatter signal except that the band-
tion of altitude was expected to provide a measurement of width of the signal was considerably less than the predicted
the various geophysical parameters both in the bottomside value. The decrease of the bandwidth of the scatter signals
and the topside ionosphere. Gordon assumed that the elec- contributing a larger signal power per unit bandwidth ob-
trons were in random thermal motion of the same type as viously made it easier to detect the signal. Bowles [1958]
the motion executed by neutral particles so that the radar correctly surmised that the presence of ions causes a re-
would detect scattering from individual electrons that are duction of the bandwidth of the scattered signal. Later the-
random in phase or incoherent. This is known as incoherent oretical work [Fejer, 1960; Dougherty and Farley, 1960;
scatter or Thomson scatter (for a comprehensive review, Salpeter, 1960; Hagfors, 1961] showed that the spectral
see Evans [1969]). Gordon calculated the backscattered power form of the scattered signal is dictated by the radar wave-

10-14
 IONOSPHERIC RADIO WAVE PROPAGATION
length in relation to the Debye length in the upper atmo-
sphere. The Debye length (D) for electrons is defined as
COMPLETE SPECTRA FOR VARIOUS VALUES C)F C

D = (EOkTC/4mN~)”2 m (10.22) a=o.oo3

where E. is the perrnittivity of free space ( = 8.85 X IW’2

‘2
F/m), e is the charge on an electron ( = 1.6 X 10-” C), k
is the Boltzmann’s constant, N. is the electron density (m-s)
and T. is the electron temperature (K). The Debye length
for the electrons in the ionosphere is typically of the order 10
of 1 cm or less below 1000 km and it is not possible to
sustain organized motion at scales smaller than these values.
It was shown that, in general, the spectrum of the scat-
I
tered signal consists of two parts, one due to the ions and L

the other to electrons. If the radar wavelength is much smaller

than the Debye length, the scattered energy is entirely due
0+
I \l

A
to the electronic component and the initial predictions of
Gordon [1958] for the scattered power (Ncre) and its spec-
trum (AfC) are valid. On the other hand, for radar wave-
lengths much larger than the Debye length, which represents
1-
the experimental situation, the electronic component de-
IC-8 162 10-’ 1.0 Io
creases and appear as a single plasma line at a Doppler shift
DOPPLER SHIFT (A fe)
approximately equal to the plasma frequency of the medium.
Under this condition, the largest part of the scattered energy
Figure 10-10 The variation of the overall spectrum for different values
resides in the ionic component and the spectral width is of the ratio rs( = 4mD/A). The ion has been assumed O+.
controlled by the Doppler shift Afi for an ion approaching These curves assume that collisions arc negligible and that
the radar at the mean thermal speed of the ions, given by T. = T, [Hagfors, 1961].

1 8kTi “2
Afi=~_
()
mi
Hz, (10.23) is encountered in the ionosphere, the total scattering cross-
section (u) may approximately be given by

where mi is the mass of the dominant positive ion and Ti is ue

the ion temperature. Considering T. = Tij and the dominant u. (10.24)
ion to be 0+, Afi = 2 x 10-2 Af,. The echo energy is, (1 + a’)(1 + : + CS2)
therefore, mainly concentrated in a relatively narrow spec- 1
tral window rendering the radar investigation $easible with
The incoherent scatter radar technique opened up the
apparatus of much lower sensitivity than initially envisaged.
possibility of in situ sampling of a wide range of upper
Figure 10-10 shows how the spectral shape depends on a
atmospheric parameters by the use of a powerful ground-
parameter cs = 47rD/k for the case T. = Ti. For ci z 10,
the scattered energy is entirely due to the electronic com-
ponent, whereas for very small values of ~, the electronic Ce.
component decreases and the energy appears mainly in the ~~e
ionic component with a much smaller bandwidth. The elec- z
tronic component now appears as a single line, known as i=
the plasma line, at a Doppler shift approximately equal to $
the plasma frequency of the medium. ~-
In the ionosphere, the electrons and the ions are at dif- 3
ferent temperatures and the spectrum of the ionic component ~ 01

changes for different values of the ratio T./Ti at a given

value of ci. This is shown in Figure 10-11 for the case $

C2 = 0.1 for O+ ions which illustrates the double-humped

form of the spectrum. By measuring the height of the hump 0 m 06 12 s 20 24

at the wing relative to the center of the spectrum, it is DOPPLER SHIFT (A f, )

possible to estimate Tc/Ti and the total scattering cross-
section due to the ionic component is simply obtained from Figure 10-11. Spectra of the ionic component for the case t)f
the area under the curves. For small values of T./Ti, which a( = 47rD/A) = O. I [Evans, 1969]. (Reprinted with per-
mission from IEEE G 1969. )
10-15
CHAPTER 10
based radar system [Evans, 1969]. The most obvious meas-
urement is the electron density (N) versus altitude (h) profile J
made by recording the variation of echo power P, as a 3000 .

function of delay by using a vertically directed pulsed radar. 4000 -

JICAMARCA
19 WL 1962
The echo power is given by — 1430-1535 EST
m - — 15W EST
+ 1543 EST
p = ~ N(h) u(h) 2500 -
5 (10.25)
h2 ‘ ?.ocro
-

1500 -
where C is a constant. The constant C can be determined
either by a careful determination of the radar parameters or
by an absolute determination of N at an altitude by an Zm _
~900 -
ionosonde or other techniques. However, as mentioned ear-
gaoo -
lier (Equation 10.24), the scattering cross-section u(h) de- g 700 -
pends both on a and Tefli which are both functions of ~600 -
altitude. From a measurement of the scattered energy spec- 500 -
trum, these corrections can be introduced and electron den-
4(xr -
sity profiles are determined. It has also been possible to
obviate this difficulty entirely by the use of Faraday rotation w -

technique. Figure 10-12 illustrates the electron density pro-

file extending to almost one earth radius obtained at Jica- 200 -
marca by this technique. In addition to the rather straight-
forward measurement of electron density profiles, electron 150

and ion temperatures, ion composition, and photoelectron

t i
flux, the ionospheric electric field and a variety of other up- loo~
per atmospheric parameters have been successfully meas- 10’ 10° lo” low 10”
ELECTRON
~ (N/m’)
ured at various locations extending from the magnetic equa-
tor to the auroral zone [Radio Science, special issue, 1974]. figure 10-12. An electron density profile obtained at Jicamarca that ex-
tends to almost one earth radius [Bowles, 1%3].

Table 10-1. Incoherent scatter facilities

Location Frequency (MHz) Power (MW) Antenna Dip latitude (“N)

licamarca, 50 6 290 m x 290 m array 1

Peru Pulsed
Arecibo, 430 2 300 m spherical 30
Puerto Rico Pulsed reflector
St. Santitt, 935 0.15 20m X100m 47
France Continuous reflector
Millstone Hill, 440 3 68 m 57
tJSA 1300 4 25 m parabola
Pulsed
Sondrestrom, 1300 5 32 m parabola 71
Greenland Pulsed
EISCAT*
Transmitter
Tromso, Norway 224 5 30m X40m 67
(monostatic) parabola cylinder
933.5 2
(tristatic) 32 m parabola
Receiver:
Tromso, Norway
Kiruna, Sweden 32 m parabola
Sodankyla, Finland
*European Incoherent SCATter facility

10-16
 a

IONOSPHERIC RADIO WAVE PROPAGATION

probe is replaced with the entire exposure conducting sur-
face of the rocket or satellite. If the exposed conducting
surface of the rocket or satellite is much greater than the
area of the probe, the potential of the surfaces will remain
fixed as the potential on the probe is swept. As a minimum,
the area of the conducting surfaces should be 100 times
greater than the area of the probe, and ideally the area should
be 10000 times greater then the area of the probe. By setting
SPAiEtiAF T the potential of the probe very positive ( + 1.5v to + 20v),
SURFACE
all electrons within a few Debye lengths of the probe will
Figure 10-13. Cylindrical Langmuir Probe. be drawn in and measured; this allows a direct measurement
of plasma density oscillations which are directly related to
plasma turbulence.
Table 10-1 adopted from Hargreaves [ 1979] gives a list of The two most common shapes for Langmuir probes are
the incoherent scatter facilities now in operation and the the cylindrical probe (Figure 10- 13) and the spherical probe
characteristics of the radar system. (Figure IO-14). Any shape probe is possible, but these shapes
are the easiest to analyze mathematically.

10.1.3 Langmuir Probes

10.1.4 Faraday Cups for Rockets
One of the simplest devices used on rockets or satellites and Satellites
to measure the ionospheric density in situ is the Langmuir
prok, named for Irving Lartgmuir, who pioneered the method The most commonly used device for measuring the ther-
at General Electric in the 1920s. The density is determined mal ions is the Faraday cup (see Figure 10- 15). It is usually
from a measurement of electric current passing between two an aperture that is a section of a flat, infinite surface in
conducting surfaces in contact with the environment. A contact with a plasma. A screen across the aperture shields
varying electrostatic potential placed between the two sur- electrostatic potentials inside the sensor from the outside
faces causes the current to vary. The magnitude of the cur- environment. The arrangement of grids or screens inside
rent indicates the density of the ionospheric plasma, and the the sensor is determined by the function of the sensor. Most
change in current with respect to changes in the potential Faraday cups use a suppressor screen in front of the col-
between the surfaces indicates the ion and electron tem- lector. This screen has a large negative potential ( – 10V to
peratures. The double-floating-probe, which is the closest – 100V) to repel electrons from the environment away from
version to an idealized Langmuir probe, usually consists of the collector and to drive secondary and photoelectrons from
a conductor at each end of a dipole antenna flown on a the collector back to the collector.
rocket or satellite for other purposes. The major disadvan-
tage of a double-floating-probe is that ion thermal velocity
is much lower than the electron thermal velocity and the
rocket or satellite speeds. Therefore, the usual Langmuir
probe is a single probe to measure only electrons; the other
.25
~’
.25
.07- DMSP SSIE

DMSP SSIE ELECTRON SENSOR t}

““’A 10N ‘ENsOR
GOLD PLATED ALUMINUM
1

R
COLLEC~OR 1.75” OIAMETER
!,
COLLECTOR
/—-. / :!,
/ UOUNTIN6 BOOM
,1

/ &pERTURE —, ,

SWEP7 GRID J!i~

( J.z;
SUPPRESSOR : ; , , TO 10N AMPLIFIER
;

\ ALL GRIDS GOLD PLATEO TuNGSTEN,

TO ELECTRON
\
AMPLIFIER ] CbLLECTOR GOLD PLATEO ALuMINuM.
\ ii~m-,j,j,,,:RFAcEsGoLD
\ PLATEO.
=—. “\
00LD PLATEO TUNOSTEN GNO.
TRANSPARENCY - 0.S0
2.2S” DIAMETER

Figure 10-14. The Spherical Langmuir Probe on the DMSP Satellite Figure IO- 15. The Fardday Cup used on the DMSP Satellite.

10-17
CHAPTER 10
10.1.5 Optical Measurements from 2-18 A. They can be scanned across the sky or operated
in the zenith. When properly calibrated, spectrometers pro-
Ground, airborne, satellite, and rocket based optical vide the absolute intensity of aurora] and airglow features
measurements are commonly used to determine ionospheric as well as some measure of spectral character.
structure and dynamics. While a number of different in- High Resolution Systern.s: Fabry-Perot interferometers
struments are employed, all analysis techniques must relate use multiple path interference to achieve high spectral res-
spectral emission features to ionospheric structure and dy- olution. These instruments are primarily used to measure
namic processes. This is done through a knowledge of the spectral line broadening and Doppler shift. From these pa-
atmosphere/ionosphere chemistry that leads to the measured rameters, atmospheric temperatures and drift velocities can
photon emission. Ionospheric domains are conveniently di- be derived. Primary spectral features of interest are 6300 ~
vided into regions that are produced or influenced by en- [01], 5577 ~ [01] for neutral winds. and 7320 ~ [0111 for
ergetic particle precipitation (auroral regions) and those con- plasma drift.
trolled mainly by solar ionizing radiation (equatorial and
midlatitude). Optica~ measurements have played important 10.1.5.2 Ionospheric Structure from Optical
roles in both regions in defining the spatial and temporal Measurements. Ionospheric structure at mid and equa-
characteristics of ionospheric plasma. Commonly used ob- torial latitudes is controlled by solar ionizing radiation, elec-
serving techniques will be discussed followed by a section tric fields, and neutral atmosphere dynamics. Airglow ob-
describing important results. servations of equatorial plasma depletions are one example
of optical measurements used to define ionospheric pro-
10.1.5.1 Observing Techniques. Optical instruments can cesses. A brief review of equatorial airglow chemical pro-
be classified according to spectral resolution as low, me- duction mechanisms is presented to illustrate the techniques
dium, and high. used to infer ionospheric plasma density variations from
Low Resolution Systems: The all sky camera has his- remote optical measurements.
torically been used to measure auroral structure. This is Two primary airglow spectral emission features are of
perhaps the lowest resolution system, measuring all wave- interest for nighttime, F region phenomena 6300 ~ [01] and
lengths over the sensitivity range for the type of film used 7774 ~ 01. The 6300~ atomic oxygen emission results from
(typically Kodak TRI X). The system uses a 160° field of the following sequence of reaction:
view lens to measure auroras over a circle of 1000 km
diameter in the lower ionosphere (110 km altitude). All sky
O++ OzHO1+e (10.26)
cameras typically measure only bright auroral features, pri-
marily at E region altitudes.
Photometers are low resolution systems. They rely on 0~ +eHO+O(’D) 10.27)
narrow band interference filters to isolate spectral lines and
0( ’D) ~ O (’P) + hv(6300) 10.28)
bands of interest. Meridian scanning photometers use a nar-
KQ
row (0.5° to 2.5°) field of view and -2 AOfilters to measure
0(’1)) + Nz ~ 0(3P) + Nz. 10.29)
absolute intensity of auroral and airglow along a vertical
circle, commonly aligned along a magnetic meridian. Tilt-
Since K2 > K,, and in regions where O”+ is the dominant
ing filters use the change in transmitted wavelength versus
ion (O+ = N.) the 6300A volume emission rate is given
tilt angle to perform a limited wavelength scan. This allows
by
separation of non-spectral continuum from the line or band
emission.
More recently, all-sky imaging photometers have been dI(6300) = 0.75 K,E [N.] [0,]/ ( 1 + ~)~dh
developed to perform alI-sky ( 155°) monochromatic meas- [
urements at high sensitivityy (20 Rayleighs). These employ (10.30)
slightly wider (-20 A“) interference filters because of the
lack of convergence of the extreme optical rays at large
where K2E = 1.4 x 10-” cm3 s-]
zenith angles. Image intensifiers are employed to achieve
the high sensitivity. Data are recorded either on a photo- KQ = 7.0 X 10-” cress-’
graphic image or by using a TV system to produce a video 10-2 s-l
signal. Typical system parameters are shown in Table A=
10-2.
Medium Resolution Systems: Ebert-Fastie type scanning (See Weber et al., [ 980] and Noxon and Johanson 1970]
spectrometers are used as medium resolution optical systems. for a more complete discussion).
These are effectively used over the (visible) wavelength The 7774 ~ 01 results from radiative recombination of
range of 3800 to 7900 ~ with variable spectral resolution 0+:

10-18
 a
IONOSPHERIC RADIO WAVE PROPAGATION
Table 10-2. Summary of system specifications.

Field of View 155°

Pass Band 25~atfl.4;5~atf8

Resolution 1/2° zenith, 2° horizon

Spectral Response S-20, exceeding 100 pA/lumen

Picture Storage No detectable degradation for up to 3 s

Tube Gain Photon noise granularity visible above tube noise

Threshold Sensitivity 20 R at 2 s exposure

1 kR at 30 frames/s

Dynamic Range 20 R to 10 kR covered by 3 preset HV settings

Flatness of Field 30 percent loss at edge of field

Repetition Rate Typically 20 s for complete filter cycle

Temporary Storage Video disc, three video tracks + one sync track

Permanent Storage Video tape deck, time-lapse type (9 h recording time

on a single reel): 16 mm color movie camera

Process Controller In-field programming capability

Display Systems Four black and white monitors, 9 in. diagnal; Color
monitor, RGB and A-B input, 12 in. diagonal

Real-Time Display Simultaneous fully registered display of three filter

channels. Capability of displaying difference of any
two pictures. Display of two or three filters as
pseudo-color on RGB monitor

Character Generators Date/time display on each frame for frame

identification

Digital Encoding Digital encoding of time and housekeeping data for

computer-controlled data handling

0+ + e+ 0(5P) (10.31)
?W ANOMALY - 0131 UT 15 DEC 1979
CUTS!DE DEPLETIDN

0(’P) ~ 0(5S) + hv(7774) (10.32)

~[ I(7774)=3DR

.l\
Q(7774) Ne

and the volume emission rate is given by j500

I (63DoI= 430R \

dI(7774) = KS [0+ ][N.] dh. (10.33)

!“[ \%
=~,-----’ >
To illustrate the altitude dependence, 6300 ~ and 7774 ~ — —
~~ m.——..
airglow volume emission rates were calculated for an elec- 0“’ 10° 0’ 10’ 10’

tron density profile representative of the Appleton anomaly P14DT0NS/Cd -xc EL /cm>

region, and are shown in Figure 10-16. The bottomside

profile was obtained from true height analysis of a digital Figure 10-16. Electron density profile derived from true height analysis
ionogram. This was matched to a modified Chapman func- of bottomside ionogram matched to a modified Chapman
function for the topside. Also shown are calculated 6300
tion [Tinsley et al. 1973] to represent the topside profile and 7774 ~ volume emission rates and column intensities
from hmax Fz to 690 km. Because of the exponentially in Rayleighs.

10-19
CHAPTER 10
decreasing 02 concentration, the 6300 A volume emission periods in the absence of “snapshot” satellite measurements.
rate is confined to the bottomside and reaches a maximum Optical measurements coordinated with VHF radar, iono-
value at 300 km, below hmax F2 (360 km). The 7774 ~ sonde, in situ density, and satellite radio beacon scintillation
volume emission rate is proportional to [Ne]2 and attains its observations have helped to provide a detailed description
maximum value at h ~ax F2. Because of the broad altitude of the development, structure, drift, and decay of these
exatent of the equatorial electron density profile the 7774 important equatorial ionospheric features.
~ volume emission rate displays a similar broad extent and Optical measurements have improved our understanding
falls to 50% of the maximum value at 300 km and 450 km. of auroral zone and polar cap ionospheric structure and
Thus the 7774 ~ airglow measurements provide information magnetosphere-ionosphere coupling processes. In this re-
over a broad altitude range, with approximately one half of gion dominated by the effect of precipitating electrons and
the emission produced above the F layer peak. ions over a wide energy range (few eV to 100’s of keV),
All sky imaging photometer measurements conducted optical measurements of impact excitation and chemical
within a few degrees of the magnetic equator and near the recombination aid in understanding a wide variety of pro-
Appleton Anomaly region (- 18° ML) have established the cesses. In this section, several examples of all-sky mono-
two-dimensional horizontal extent of equatorial plasma de- chromatic images are shown, primarily to demonstrate the
pletions. These are also the regions of post-sunset equatorial use of optical measurements, especially when coordinated
spread F, VHF radar backscatter plumes, and amplitude and with other ionospheric diagnostics. Figure 10-18 shows a
phase scintillation on transionospheric radio propagation. montage of auroral images at 10-rein intervals at 6300 ~
Figure 10-17 shows an example of all sky images at 6300 ~ [01] and 4278 ~ N,+. These images were recorded on an
and 7774 ~ near the equatorial edge of the Appleton Anom- aircraft which flew North-South legs along the Chatanika,
aly. The bright region over the southern twe-thirds of the Alaska Incoherent Scatter Radar magnetic meridian. The
image is airglow from the high-density anomaly region. The images provide a map of the instantaneous ~article precip-
North-South aligned dark band is a region of decreased itation patterns separately for the E (4278 A) and F (6300
airglow emission. Comparison with simultaneous in situ ~) layers. Measurements with the radar mapped electron
measurements from the Atmosphere Explorer satellite (AE- density structure and satellite UHF radio beacon scintillation
E) shows this airglow depletion to the collocated with a measurements mapped regions of ionospheric irregularities
region of significantly decreased ion density. Having estab- (from tens of meters to a few kilometers). [n this experiment,
lished the relation between airglow emission processes and optical measurements provide a continuous map of particle
F layer density, the all sky images provide a two-dimen- precipitation regions over a large area (1200 km diameter
sional map of these depleted regions. In addition, the dy- at F-region altitudes) for interpretation of magnetosphere-
namics of these regions can be monitored over extended ionosphere coupling and ionospheric dynamics.
Measurements in the polar cap have recently clarified
local particle precipitation effects from plasma transport (E-
15 DECEMBER 1979 field) effects. Local precipitation of low energy ( 100’s of
eV) electrons, during IMF Bz north conditions. leads to the
production of sun-aligned F region auroras. Some of these
auroras are characterized by F region plasma density en-
hancement and structuring within these auroras leds to am-
plitude and phase scintillation.
During Bz south conditions, large patches of F region
plasma are observed to drift across the polar cap in the anti-
sunward direction. Coordinated satellite measurements show
that these patches are not locally produced by precipitating
particles, but are convected from a source region at or equa-
torward of the dayside cusp. These patches are also subject
to structuring processes that lead to scintillation.
g,04 L

QI
103 -

,02
I

‘5035
L -L— ——-
11
10.2 SOME ASPECTS OF
LONG WAVE PROPAGATION
-12.9 -138 0-?43: -147 O-?/} ‘14,40:;: -13 I ?:; ‘1I I ‘;02: ;:f:
-21 ’34 ’46 ’59 ’71 ’83 ’95 -106-117 -127-137 It is convenient to refer to radio waves having frequen-
03 3.600:: 104231:: 17.3 207 242278 313 344 GLON (W I
0034 2306 2236 2208 L-(H V’
cies below 3000 kHz as “long waves”. These include Ex-
tremely Low Frequencies (ELF), Very Low Frequencies
(VLF), Low Frequencies (LF), and Medium Frequencies
Figure 10-17. All sky images at 6300 and 7774 ~ recorded near the
equatorial edge of the Appleton Anomaly. (MF), as outlined in Table 10-3. ELF has had very little

10-20
Figure IO- 18. Auroral images taken at 10 minute igtemals from 0620 to 0950 UT on 29 January 1979. The upper row under each hour shows the 6300
~ images; the lower row, the 4278 A images.

10-21
CHAPTER 10
Table 10-3, Long wave frequency bands.

Designation Abbreviation Frequency Range Wavelengths

Extremely Low Frequency ELF 0.W3-3 kHz 105– 102 km

Very Low Frequency VLF 3-30 kHz 10Z–lO1 km
Low Frequency LF 30–300 kHz 10–1 km
Medium Frequency MF 300-3000 kHz l–O. 1 km

use, except for communications that require wave penetra- phase. The latter is the basis of the long-range 100 kHz
tion beneath the surface of the ocean or earth. The VLF/LF groundwave navigation system, Loran-C.
bands are used extensively for navigation and military com- If the transmitted signal is a continuous wave, the am-
munication. The standard AM broadcast systems utilize part plitude and phase of the composite signal received at a fixed
of the MF band (535-1606 kHz). Long radio waves are also distance vary with time as the ionosphere changes. On the
used in basic ionospheric research, meteorology and thun- other hand, at a given moment the signal amplitude is a
derstorm study and tracking, standard frequency and time function of distance [Hollingworth, 1926], having maxima
distribution, geological studies, and minerals exploration. and minima typical of an interference pattern. The ground-
Long waves propagate by a number of different modes. wave component is stronger than the skywaves out to a
These ‘include propagation over the surface of the earth by distance that depends on the wave frequency, among other
diffraction modes, ELF propagation by transmission-line factors. This region of groundwave dominance is the most
type modes, propagation by ionospheric reflection (or earth- stable, or primary, coverage area of MF broadcast trans-
ionosphere waveguide modes) and propagation through the mitters.
‘---------
Iunuspncrc ‘-- -- --”-J “whistler” modes. Each type of
uy su-caIIGu As defined above, the groundwave exists at all radio
mode requires a separate physical description and mathe- frequencies, but at wavelengths comparable to the height
matical formulation. of the ionosphere or greater, the usefulness of the concept
begins to fade. Also, for transmitters high above the ground,
or at high frequencies where quasi-optical propagation anal-
ysis is appropriate, the term groundwave is seldom used.
10.2.1 Groundwave Propagation The earth often acts as a fairly good conductor for long
waves, in which case the electromagnetic boundary con-
The most general definition of the groundwave is the ditions permit electric fields perpendicular to the surface.
wave that would be excited by an antenna at or near the air- while tending to suppress electric fields tangential to the
earth bounds@ if there were no wave reflections from the surface. It follows that groundwave fields near the earth’s
upper atmosphere. At long wavelengths ionospheric reflec- surface tend to have transverse magnetic (TM) polarization
tions are important, and for continuous wave (CW) trans- rather than transverse electric (TE) polarization. In common
missions it is necessary to regard the total wave field as a usage the unqualified term “groundwave” implies TM po-
vector sum of the groundwave and skywaves. If an antenna larization.
radiates a very shprt pulse, however, it may be possible for
a distant receiver to resolve the groundwave and skywaves 10.2.1.1 IdeaIii Flat--h Models. In a simple model
individually. The time interval between the onsets of the the earth’s surface is regarded as a flat perfect conductor,
groundwave and the first-hop skywave is given by and the air is homogeneous with refractive index 1. The
most elementary source is a vertically-directed current
At = (2~h2 + 4a(a + h)sin2(d/4a) – d) (l/c), (10.34) I(t) = 10exp (i~t), at frequency f = m/2m, of infinitesimal
length d~, which has an electric dipole moment M(t) = I(t)
providing de (Note: complex antennas can be regarded as distributions
of such elementary currents). The fields of such a source
d < 2a cos-l{~(a+h)}, (10.35) may be found readily by the method of images. When the
current element is just above the surface, the fields in air
where d is the distance between the transmitter and receiver, are simply twice the homogeneous free-space fields. Be-
h is the effective height of reflection in the ionosphere, a is cause of symmetry the magnetic field is everywhere in the
the earth’s radius (=6370 km) and c is the wave velocity azimuthal direction @ (see Figure 10- 19), while the electric
(3 x 10’ km/s). If the transmitted pulse is short enough, field on the surface is constrained by the boundary conditions
At may be long enough (for example, At = 93 PS for to be strictly in the vertical direction z. In mks units the
d = 500 km and h = 80 km) to pe~it very accurate mea- magnetic and electric fields at a distance, d, on the ftat
surements of the groundwave, especially its arrival time or perfectly-conducting surface are, respectively,

10-22
 n
IONOSPHERIC RADIO WAVE PROPAGATION
z
I Ez I
# 1.0 , 10

IGI

Figure 10-19. Vector field-components at a point P in a cylindrical co-

ordinate system. The plane XOY represents the surface of
the earth.

-
M(t’) 1
H+= =$+$ At/m, /
[ (10.36)
0.05 0.10 o.? 0.5 1.0 2.0
( , 1 1 1

~1
I
2 5 10 20 50 100
+ 1201r arg G, degrees
EZ=-~~ 2 +
Vim,
2Tr [ i2EOwd3 d
(10.37)
Figure 10-20, Height variations in the amplitudes and phases of 100 kHz
E, = O, (10.38) groundwave fields for a source on a plane earth. Values
are shown at distances of 30, 100, and 300 km, for prop-
agation over fresh water, u = I&3 S/m, dco = 80
[Heckscher and Tichovolsky, 1981 ].
where t’ = (t – d/c), A is the vacuum wavelength
(=3 X lf)8/f m), E. is the permittivity o f free space
Zenneck wave, without any radiation field, requires an in-
( = 8.854 x 10-’ 2 F/m), and PO is the free space perme-
finitely long source distribution [Hill and Wait, 1978].
ability y ( = 41T x 1&7 H/m). The far-field components are
For a finitely conducting earth, Equation ( 10.39) is still
related by
true approximately, but the radial electric field component
E= = – ZOHO (10.39) EP has a finite value related to the loss of wave power into

,I,,,q —-T- r —r.—, —,—-,

where Z. is the impedance of free-space (~ 1201T ohm).
In a more realistic model, the plane earth is allowed to
““r~ r I
have finite conductivity u and permittivity E. The solution 100 -

of this boundary-value problem was given by Sommerfeld

[1909] in terms of an infinite complex integral [see Stratton, 80-
0
194 1]. A more complex problem, that of an elevated dipole, w
~
was solved by Weyl [1919], who expressed the free-space 6 60 -
a
field as a sum of plane waves that reflected at the earth’s K
3
surface in accordance with the Fresnel formulas. Norton * 40
I
[1941 ] and others have calculated numerical values from E“
3
the formal solutions. Height variations of the fields are shown >
; 20 ------PLANE EARTH
in Figure 10-20 for a source on a plane earth surface u —SPHERICAL EARTH

[Heckscher and Tichovolsky, 198 1], and curves illustrating

o
groundwave field amplitudes along the surface are given in
Figure 10-21 for both plane and spherical earth models.
A type of groundwave, the Zenneck surface wave, has
DISTANCE , km
fields expressed exactly in simple closed forms. Although
the Zenneck wave is important historically and concep-
Figure 10-21 Long wave groundwave field amplitudes as a function of
tually, it is generally difficult to excite because of its rather distance over plane and spherical earths, for propagation
slow decay with height. In fact, the excitation of a pure over good earth, u = 10-2 S/m, ‘E/E. = 20.

10-23
CHAPTER 10
Table 10-4. Ratio of groundwave radial and vertical electric fields for be sensed with a vertical monopole antenna, or a vertical
various earth surfaces and 100 kHz. dipole. The radial component E, may be sensed by a hor-
izontal dipole with its axis in the direction of propagation.
It follows from the principle of reciprocity that if a horizontal
dipole on the earth is driven with RF current, a TM-polarized
groundwave is radiated in directions along its axis. The
fields produced by horizontal and vertical current elements
have been discussed in mathematical detail by Wait [ 1954,
I Good Soil
1957, 1961, 1971].
Fresh Water ~0-3 80 0.0713 l=

Poor Soil 10-3 10 0.0745 l= 10.2.1.2 Idealized Spherical-Earth Models. A mathe-

matical treatment of the groundwave on a smooth spherical
Thick Ice 2 x l&5 5 0.403 1~” earth of homogeneous, isotropic material was undertaken
by Watson [1919] to determine if an atmospheric reflecting
layer (ionosphere) was required to explain the large fields
the ground. The ratio of the radial and vertical electric fields produced by distant transmitters. Such early theoretical anal-
at the surface is given by yses were handicapped by the poor convergence of the in-
finite series contained in the solutions. That difficulty was
E$EZ = ti (1 - I/p)/p, (10.40) largely overcome by Van der Pol and Bremmer’s [ 1938]
“residue series” solution, which has become a basis for
where modem numerical analysis of the groundwave.
Figure 10-21 shows examples of field-strength vs dis-
p = E/E. – iu/@Eo. (10.41) tance curves for waves of selected frequencies propagating
over “good soil ,“ assuming a vertical source on the surface.
Starting just below the surface of the assumed uniform The earth-curvature causes the wave amplitude to decrease
earth, EP and H+ decrease exponentially with depth. The with distance more rapidly than it would on a flat earth of
fields are l/e of their values on the surface at the “skin the same material, but near the source the flat- and round-
depth’ S, which may be estimated from earth models give essentially the same fields.
One way to present both amplitude and phase data is in
terms of the complex factor W, which is the ratio of the
actual field component to an idealized one calculated as if
the earth were flat and perfectly conducting. The flat earth
. propagation distance is taken to be the same as the curvi-
5031~f when arg p = – m12. (10.43)
linear distance on the sphere. Figure 10-23 shows curves
of W in the complex plane, for 100 kHz propagation over
Sample values of EP/EZ are given in Table 10-4 for various
sea water and good soil.
earth surfaces at a frequency of 100 kHz, and Figure
10-22 shows skin depths over the long wave spectrum.
The magnetic field of the groundwave H+ may be sensed
with a loop antenna having its axis parallel to the surface
and perpendicular to the direction of propagation. EZ may

o4- -
g
~ 0,3 =.4slm

02- -

I
-0.3 -o. 0.1 0.2 0.3 0.4 05 0.6 0,7 0.8 0.9 1.0
/
R. {W)
-o.l–

Figure 10-23. Complex values of the ratio W of actual groundwave field

FREQuENcY, HZ- components and ones obtained assuming a flat, perfectly-
conducting earth, for 100 kHz over sea water (U = 4 S/m)
and good soil (u = l&z S/m). Derived from Wait and
Figure 10-22, Long wave skin-depths for various earth surfaces. Howe [1956].

10-24
 a

IONOSPHERIC RADIO WAVE PROPAGATION

10.2.1.3 Models with Earth-Properties Gently Varying Groundwave perturbations caused by hills have been
Along the Propagation Path. Except for large bodies of studied with models having semielliptical bosses on oth-
water, the earth’s surface is too uneven, both in electrical erwise smooth surfaces [Wait and Murphy, 1957, 1958].
properties and in topology, to be represented well by the In the electrostatic case, the vertical electric field at the
idealized models discussed above. However, if the earth’s summit of a hemispherical boss is exactly 3 times that on
electrical properties and curvature do not vary much in a a flat plane, and for a semicylindrical ridge the factor is 2
wavelength, the groundwave amplitude and phase can be times. This field-enhancement effect carries over (with mod-
approximated by the solution of an integral equation for W ifications) to VLF groundwaves, and has been demonstrated
[Hufford, 1952]. The two-dimensional integral equation— experimentally [Harrison et al., 1971].
the most general form-is valid provided, first, impedance
boundary conditions can be applied; and second, terrain
irregularities are not too severe.
A much simpler, one-dimensional, integral equation is 10.2.2 ELF Propagation
more commonly used. It is derived by a stationary-phase
integration that reduces the dimensionality of the general ELF propagation has been the subject of theoretical study
version; being an approximation, it is not valid for all terrain for many years. The texts by Wait [ 1970], Galejs [ 1972]
types, particularly at low frequencies where wave lengths and Burrows [1978], along with review papers by Bernstein
and Fresnel zones have sizes comparable with terrain fea- et al. [1974], Wait [1974 and 1977] and Bannister, [1980],
tures. Field [1982a] compared solutions of the one- and provide comprehensive descriptions of the propagation char-
two-dimensional equations to quantify the errors incurred acteristics of ELF waves. Much of the discussion that fol-
by using the one-dimensional equation for terrain features lows is based on a review paper by Bannister [ 1982].
narrower than a Fresnel zone. For a frequency of 100 kHz ELF waves below a few hundred hertz propagate with lit-
and on-path features narrower than about 10 km, the two- tle attenuation, penetrate well into Iossy media and are very
dimensional equation is needed to properly account for the stable compared with higher frequencies. Nevertheless, ELF
signal’s dependence on obstacle width, recovery at long has limitations relative to conventional radio communication
distances, and transverse diffraction patterns. The one- and bands. Its restricted bandwidth allows only very low data rates,
two-dimensional solutions approach one another far beyond and because of the great wavelengths, ELF transmitting an-
wide terrain features. tennas are very inefficient (less than 0.5% is typical).
Considerable error can be incurred at low frequencies The energy of an ELF wave is confined principally to
by applying the one-dimensional equation, even for large the waveguide that exists between the earth and the iono-
terrain features. For example, for a path-length of 500 km, sphere. At ELF the effective height h of the waveguide is
that equation overstates by a factor of about four the effect much less than the wavelength h of the wave and the wave-
of an obstacle 6 km in diameter. It cannot give accurate guide is below cut-off for all but the lowest order mode,
results unless the diameter approaches a Fresnel zone width, the transverse electromagnetic (TEM) mode. On the other
which. for this example is several tens of kilometers. How- hand, at VLF tbe waveguide height exceeds the wavelength
ever, because numerical solution of the two-dimensional and several modes propagate. At LF the number of signif-
integral equation is costly, its use has been limited to highly icant propagating modes may exceed 20.
idealized irregularities. An alternate approach given by King The principal TEM fields are the vertical electric field
and Wait [1976] obtains an equivalent onedimensional model E,, and the horizontal magnetic field, H~. Secondary field
by averaging the terrain over the Fresnel zone. components arise because the surface impedance qg of the
If, instead of being homogeneous, the earth were com- ground—albeit small—is not zero; hence the term “quasi
prised of layers of different conductivity and dielectric con- TEM” mode. The secondary fields are small compared with
stant, it could still be characterized by a (frequency depen- the principal fields, but are important because horizontal
dent) surface impedance, so that the integral equation for antennas would not radiate if the secondary fields were zero
W can be formulated. Some progress has been made in this [&UTOWS, 1978].
and other ways of estimating earth-constants for ground- Attenuation of the ELF quasi-TEM mode in the earth-
wave phase prediction at 100 kHz, but uncertainties in these ionosphere waveguide is low, on the order of 1 or 2 dB/Mm.
constants remain a major source of prediction error. The attenuation is caused mainly by power absorption in
Special models of non-uniform terrain exist for which the ionosphere, since the surface impedance of the iono-
solution of the integral equation is unnecessary. Ground- sphere is typically much larger than the surface impedance
wave propagation from land to sea, and vice versa, has been of the ground. That effect is evidenced in the expression
modeled with an earth having a sharp discontinuity of con- for the waveguide attenuation rate a, which is inversely
ductivity along a horizontal straight line. A useful approx- proportional to h. That behavior indicates that the rate of
imate solution was given by Millington [1949] based on power leakage from the guide is proportional to the field
reciprocity arguments, and mathematical solutions have been intensity at the surface, whereas the rate of power flow along
given by Wait [1956] and Wait and Walters [1963]. the guide is proportional to the guide’s volume. Thus, as

10-25
CHAPTER 10
the guide’s height decreases, the ratio of power leakage to the exponential decay due to absorption and the p- ‘/2decay
total power flow increases [Burrows, 1978]. due to spreading; and (6) the directional dependence of the
ELF attenuation is low enough to support very long- radiated field. Once the current moment Id/, angular fre-
range propagation, and a planar model of the earth is in- quency w, and coordinates p, @ of the field point are spec-
adequate if the path length exceeds the earth’s radius. The ified, only two parameters are left undetermined, A and the
most important effect of earth curvature is the closure of attenuation rate, a. Both depend on the ionosphere.
the guide, which allows the field to return to the source Greifinger and Greifinger [1978, 1979], derived simple
point after one complete encirclement. Therefore, the total approximate expressions for ELF propagation constants that
field is the sum of the field arising from propagation over agree well with full wave numerical calculations. For day-
the shorter great-circle path from the source, and that arising time propagation, the approximate expressions for c/v and
from propagation over the longer one. The local effect of w are
the curvature is small, however, and the wave propagates
in the curved guide with nearly the same parameters as it Clv - 0.985- 10.46)
would in a planar one.
The geomagnetic field interacts with the charged par- and
ticles of the ionosphere to produce an anisotropic conducting
medium. However, the electrical mismatch between the at-
mosphere and the ionosphere is large at ELF, and the tran-
sition bet ween them abrupt, so the ionosphere acts much
ci- o.143fm
()$+5 , 10.47)

like a perfect reflector. The effect of the anisotropy is there- where ho is the altitude where al = WEO;h, is the altitude
fore small, and ELF attenuation and phase velocity depend where 4wpo~~ = 1; ~c, and ~, are the conductivity scale
only slightly on the direction of propagation. heights at altitudes hO and h,, respectively; and U, is the
conductivity y of the ionosphere, which varies with altitude.
10.2.2.1 ELF Field-Strength Calculations. The expres- Equations (10.46) and (10.47) show that the phase con-
sions most often employed for calculating ELF fields in the stant depends primarily on the two reflecting heights and is
earth-ionosphere waveguide are based upon an idealized nearly independent of the conductivity scale heights. On the
model that assumes the earth and the ionosphere to be sharply other hand, the attenuation rate depends on the scale heights
bounded and homogeneous. Experimental measurements of as well as the reflection heights.
the waveguide properties at ELF have shown consistently A simple exponential fit to the ionospheric conductivity
that they can be represented accurately by formulas based profile is given by Wait [1970] for determining propagation
on such a simple model. Complicated calculations that ac- parameters:
count for vertical structure of the ionosphere [Field, 1969
and Pappert and Moler, 1978] also confirm that the simple U1(Z)/EO= 2.5 x 105exp [(z – H)/~,,], (10.48)
model is adequate for many purposes.
For the idealized model, the magnitude of the magnetic where H is an (arbitrary) reference height. The correspond-
field of the signal from a horizontal dipole is approximately ing values for h and h, are

‘H’-’A[k(%)”2121
(%)”2 ~ lo-ap/2m x 10’ and
‘0=H- ‘0 Ln(2”5:f’05) (10.49)

Cos +, (10.44)
(k0P)”2
‘]‘ho+2’0Ln(2”39f: 104) ‘1050
where
Note that all heights are in kilometers,

A= ‘g (10.45)
h- 10.2.2.2 Theoretical and Measured ELF Propagation
Constants. It can be shown that the effective waveguide
is an excitation factor, ~~ is the surface impedance of the height of reflection is roughly hO, rather than the higher
ground, c is the speed of light, v is the phase velocity of reflecting height h,. This is in excellent agreement with the
the TEM mode and a is the radius of the earth. effective reflection heights inferred from ELF propagation
There are six distinct factors in Equation (10.44): (1) measurements [Bannister, 1975]. The most common values
the source strength Id( (2) the excitation factor A; (3) a of H and LOemployed in interpreting VLF daytime propa-
collection of free space parameters, all of which are known gation measurements are H = 70 km and ~,, = 3.33 km.
once the frequency is specified; (4) a spherical focusing By using these values the values of h., h,, CIV and a can
factor; (5) the radial propagation loss factor, including both be determined readily at ELF.

10-26
 =

IONOSPHERIC RADIO WAVE PROPAGATION

Under nighttime propagation conditions, an E region
W- bottom where the electron density increases very sharply is
2ao “ usually encountered below the altitude h,. For a simple
—CALCULATED
G MEASURED model that assumes the density above the E region bottom
E 10.0 -
to vary slowly on the scale of the local’ wavelength, the
e
*5D - propagation constant is [Greifinger and Greifinger, 1979]
w“
+ Clv - a, (10.51)
s 20 -
z and the attenuation rate is
2 Lo -
t-
3
= 0.5 - . &l+ I (10.52)
a- o. 143f m
~ ( ho lTk&hE )‘
<
0.2 - where h~ is the altitude of the E region bottom and k@~ is
I , the local wave number.
-0”’5710 20 50 100 200 500 1000 i Nighttime ELF attenuation rates are plotted in Figure
FREQUENCY, HZ 10-26 for frequencies from 40 Hz to 1000 Hz. The calculated
values were obtained using Wait’s nighttime ionospheric
Figure 10-24. Calculated and measured ELF daytime attenuation rates conductivity model (with a-reference height of 90 km and
[Bannister, 1982].
scale height of 2.5 km) in conjunction with Equations (10.51)
and (10.52), and assuming the height of the E region bottom
to be 90 km and its conductivity y to be 8 X 1 @b S/m. Also
Theoretical values of ELF daytime attenuation rates are plotted are various measured values of a. Figure 10-26
plotted in Figure 10-24 for frequencies from 5 Hz to 2000 shows that, for frequencies from 45 Hz to 800 Hz, there is
Hz. Also plotted are values of a determined from controlled excellent agreement between the theoretical and the meas-
source measurements [Bannister, 1982], or inferred from ured values.
Schumann resonance measurements [Chapman et al., 1966]. In addition to their very low attenuation rates, ELF radio
Figure 10-24 shows that the agreement between the theo- waves below about 200 Hz can penetrate lossy media and
retical and measured values of ELF daytime attenuation rates retain usable strengths to substantial depths. Those features
is excellent. make them attractive for communicating over great dis-
Theoretical values of ELF daytime relative phase ve- tances to sub-surface locations. For example, even for sea
locity are plotted in Figure 10-25 for frequencies from 5 Hz water, with a conductivity of 4 S/m, the skin depth is about
to 1000 Hz. Also plotted are various values of c/v deter- 36 m at a frequency of 50 Hz.
mined from measurements of atmospherics. Figure 10-25
shows excellent agreement between the theoretical and mea-
sured values of c/v for frequencies above 50 Hz, and fair
agreement for frequencies below 50 Hz.
—CALCULATED
{ -
0 MEASURED
m 5.0
=
DAYTIME u-
— CALCULATED ~
?1,6 - @ MEASURED a
o
z 2.0 -
>- g
1-
% NIGHTTIME
$ ‘o
w 1.2 - k
2 s
x
a
1$0 I t 1 I 1 1 I 1
o.~~
5710 20 50 100 200 500 1000 Ic
FREQUENCY, t’tz FREQUENCY , fiz

Figure 10-25, Calculated and measured daytime ELF phase velocities Figure 10-26. Calculated and measured ELF nighttime attenuation rates
[Bannister, 1982], [Bannister, 1982],

10-27
CHAPTER 10
10.2.2.3 Anomalous ELF Propagation Occasionally the agation in the earth-ionosphere waveguide where conditions
nighttime field strengths measured at 40 Hz to 80 Hz have change over transverse distances comparable with a Fresnel
decreased by 4 dB to 8 dB in the northeastern United States zone. They derived an expression for the relative errors
[Bannister, 1975, 1980]. Those relatively severe nighttime introduced by neglecting transverse ionospheric gradients
fades sometimes occurred during the several days following over the path and found that the WKB method is inaccurate
magnetic storms, when similar (but less pronounced) be- when the width of a disturbance is less than two thirds of
havior was found to coincide with phase disturbances on the width of the first Fresnel zone. Further, the WKB ap-
VLF paths across the northern United States. These short- proximation significantly overestimates the propagation
path (- 1.6 Mm) field strength reductions might have been anomaly when the disturbance is centered near the propa-
caused by enhanced ionospheric ionization due to precipi- gation path and underestimates the anomaly when the dis-
tating electrons from the radiation belts. However, attempts turbance is centered far off path.
to correlate the fades with geomagnetic indices have met Strong localized disturbances behave like a cylindrical
with limited success. Simultaneous measurements taken in lens tilling a narrow aperture. Lateral diffraction, focusing,
Connecticut and the North Atlantic area during the mag- and reflection cause the transverse electromagnetic (TEM)
netically quiet period of early March 1977 have indicated mode to exhibit a transverse pattern of maxima and minima
some of these anomalies might have been caused by a mov- beyond the disturbance and a standing-wave pattern in front
ing nocturnal sporadic E layer. of it. The focusing and diffraction diminish when the trans-
Calculations by Barr [1977] and Pappert and Moler [ 1978] verse dimension of the disturbance approaches the width of
show that nocturnal sporadic E can produce marked maxima the first Fresnel zone, typically, several megameters. Re-
and minima in the propagation characteristics of ELF radio flection from widespread inhomogeneities can be important
waves. One physical explanation for the effect is interfer- in two situations: first, for great-circle propagation paths
ence between waves reflected from the normal E region and that are nearly tangential to the boundary of the disturbed
from the sporadic E region. Pappert [1980] showed that a polar cap; and second, when the TEM mode is obliquely
sporadic E patch one square megameter in extent could incident on the day/night terminator, in which case a phe-
account for the 6-8 dB fades that have been observed. Sim- nomenon analogous to internal reflection can occur.
ilarly, patches 0.5 Mm* in extent could account for the more
commonly observed 3-4 dB fades.
Many other ionospheric disturbances can cause ELF
propagation anomalies, including those associated with solar 10.2.3 Long-Range VLF/LF Propagation
x-ray flares, energetic electrons, protons from solar particle
events (SPES) in the polar cap, and high-altitude nuclear Very low and low frequency (VLF/LF) waves are re-
bursts. The attenuation rate can increase easily by 1 or 2 flected from the lowest regions of the ionosphere (the D
dB/Mm relative to normal daytime conditions during such region during daylight and the lower E region at night). and
disturbances, depending on the wave frequency and severity apart from the sunrise and sunset periods, exhibit propa-
of the disturbance, and certain moderate solar proton events gation characteristics that are very stable in both phase and
can cause the attenuation rate to approach 4 dB/Mm at 75 amplitude. The LF band (30–300 kHz) is useful for com-
Hz [Field, 1982b]. munications to distances ranging from hundreds to several
thousand kilometers, shorter than the almost global ranges
10.2.2.4 Analysis of Laterally Non-Uniform Iono- achievable at VLF but much longer than the groundwave
spheric Disturbances. Most predictions of ELF fields in distances normally associated with the MF band.
the earth-ionosphere waveguide have used a WKB method Beginning in 1911 with Austin, various empirical for-
described by Pappert and Molar [1978]. To find E-fields mulas, deduced from numerous measurements, have been
along any particular radial from the transmitter, that m“ethod used to estimate the field strengths of these long waves. For
assumes the properties of the guide to depend only on dis- example, Pierce derived a semiempirical formula to describe
tance from the transmitter. However, even such large in- VLF propagation over water during the day. That formula,
homogeneities as sporadic E patches, the polar cap bound- which gives the vertical electric field strength at a distance
ary, and the day/night terminator can cause the properties d from a transmitter having a radiated power P (kw) at
of the earth-ionosphere waveguide to change markedly over frequency f (kHz), is [Watt, 1967]
the huge wavelength, or Fresnel zone, of an ELF signal.
Such inhomogeneities can cause lateral reflection, diffrac- 210 W
E(pV/m) = exp [ – O.2(d/a) f 075] , (10.53)
tion, and focusing of ELF modes. Those phenomena are -
usually unimportant at higher frequencies where the earth-
ionosphere waveguide can be assumed to be vary slowly in where a is the radius of the earth and the absorption term
the lateral directions. is the exponential.
To handle such situations, Field and Joiner [ 1979, 1982] Modem mathematical approaches for predicting the
employed an integral equation approach for analyzing prop- propagation characteristics of VLF/LF waves are formulated

10-28
 4

IONOSPHERIC RADIO WAVE PROPAGATION

TM -TRANSVERSE MAGNETIC TE -TRANSVERSE ELECTRIC
in terms of a conducting spherical earth surrounded by a
concentric electron-ion plasma (the ionosphere) into which ,.. ......>..>./.
... .. . . . ... . .
,... Ev
..-
waves are launched from a Hertzian dipole source. The EL TM
wAVE - >
application of Maxwell’s equations and the appropriate ‘DIRECTION k
GROUND’
boundary conditions allow the electromagnetic field to be
calculated everywhere. In doing this the earth and the ion- WAVE POLARIZAT IONS
osphere can be regarded as forming a waveguide (without
side-walls) in which propagation may be viewed either as
a series of wave reflections (wave-hops), or by the math-
ematical equivalent—traveling wave modes. Generally, it .,,.,,
,..
is more convenient to apply the waveguide mode approach
to the VLF and lower LF band, and the wave-hop approach EXCITEO BY EXCITED BY
to the higher LF band, as described below. In addition to VERTICAL CURRENTS HORIZONTAL CURRENTS

the references cited under the specific topics that follow, FAMILIES OF PROPAGATION MOOES
more detailed descriptions of the features of VLF/LF prop-
agation can be found in the works by Budden [1961], Wait Figure 10-27. Characteristics of transverse magnetic (TM) and transverse
and Spies [1964], Watt [1967], Pappefl [1%8], Wait [1970], electric (TE) waves IKossey et al., 1982].
dalejs [1972], Field et al. [1976], and the AGARD pro-
ceedings on long waves edited by Belrose [1982]. ionosphere, the TM and TE modes are not entirely inde-
pendent, but are coupled. In general, when an electron ac-
10.2.3.1 Waveguide Modes. In the waveguide-mode quires a velocity from the electric field of the wave the
formulation for VLF/LF propagation an arbitrary propagat- magnetic forces cause it to have a component of motion
ing field is regarded as being composed of a series of modal perpendicular to the electric field, thus causing polarization
patterns that propagate with characteristic phase velocities, conversion. The interaction is described by the Appleton-
little change in pattern shape, and gradually decreasing am- Hartree equation (for example, see Ratcliffe [1959]), and
plitudes. Generally, the more complex (or higher-order) the polarization conversion has been demonstrated experi-
modal patterns attenuate more rapidly: at great distances mentally (as reported by Bracewell, et al., [1951] Lewis,
only the simpler (lower-order) modes may be important. et al., [1973] among others). The coupling effect is most
The earth-ionosphere cavity exhibits such waveguide fea- pronounced at night when the waves reflect higher in the
tures as cut-off frequencies, and reflections from mis- ionosphere, where the electron-neutral collision frequency
matched sections caused by abmpt changes in the electrical is smaller than the electron gyrofrequency.
properties along its boundaries.
The field at a point in the waveguide depends on how 10.2.3.2 Waveguide Propagation Equations. The de-
strongly the various modes are excited, and upon their rel- tailed equations for VLF/LF waveguide propagation are
ative amplitudes and phases at the observation point. Al- described, for example, by Wait [1970], Pappert and Bickel
though there is evidence of wave focusing in the vicinity [1970], Galejs [1972], and Field et al., [1976]. To illus-
of the point antipodal to the transmitter, typical field strength trate the key dependencies and to define the commonly
versus distance curves generally show a decreasing trend, used notations, they are given here for the case when
upon which are superimposed local variations due to modal geomagnetic anisotrophy (that is, polarization conver-
interference. sion) effects can be neglected. That approximation is fairly
A transmitting antenna modeled by a vertical current- accurate for long-range VLFILF propagation under nor-
element produces transverse magnetic ~) polarized waves, mal daytime conditions.
which have a magnetic field parallel with the earth’s surface, Transverse Magnetic (TM) Modes
while the electric field is perpendicular to the magnetic field Conventional ground-based VLF/LF transmitters are
and not quite vertical. For TM waves, the earth-ionosphere vertical and their fields are composed of a superposition of
waveguide has a quasi cutoff frequency during the daytime TM waveguide modes. Following Field [1982c] the vertical
at about 2 kHz. Figure 10-27 illustrates idealized TM wave electric field is given by
vectors and mode patterns.
Airborne VLF/LF transmitting antemas with horizontal in IL cos +
current elements excite transverse electric (TE) waves with
horizontal electric fields, and magnetic fields in the vertical
EV = – 120Triexp
()
–1
6

plane. Figure 10–27 illustrates idealized TE wave vectors & ~ S~’2Aj exp -~ (10.54)
mode and waveguide patterns. In general, TE fields are very r ( .)
small at the ground, and the modes are difficult to excite
with groundbased transmitters. –d Gj(h~)Gj(h~) V/m,
Because of the .presence
. of the geomagnetic field in the exp(-+: J )

10-29
CHAPTER 10
where the shbscript j denotes quantities associated with the
jth TM mode, IL is the effective electric dipole moment of
“-’~”-g \
the transmitting antenna; k is the free-space wavelength; d \
\
is the distance from the transmitter; a is the earth’s radius; \
\
and c is the speed of light. Included is a factor cos & \
1* ●

where $ is the angle between the dipole orientation and the \*

vertical-to account for inclined transmitting antennas (COS

\* 2
$ = 1 for a vertical electric dipole). Although most qtzan- j ● 3-,--, ---- ~“- -“-
tities are in mks units, all distances (L, A, d, a) are expressed ___
.> .- ------- ‘-
in megameters. “‘Lvi:,
The quantity Sj is essentially the eigenvalue of the jth -.,
\
TM mode and must be computed numerically. At VLF, \ ‘.\
however, S has a magnitude close to unity, so the term \ ‘,
\ ‘\
S~ in Equation (10.54) does not appreciably influence the -j \\
field. The magnitude of the vertical electric field depends m—
\ ‘\
on the state of the ionosphere through three parameters: Aj \ ‘1
the excitation factor for the TM mode; ~j , the attenuation \ ‘\
*\ ‘\
rate in decibels per megameter of propagation (dB/Mm); ●
/\
and Gj , the height gain function for transmitter and receiver ‘\
heights h= and hR, respectively. The phase of the jth mode
is governed by the relative phase velocity, c/vj . These
I I I I [ I
propagation parameters must all be computed numerically 10 15 20 25 30 35 40 45--
for model ionospheres having arbitrary height profiles. FREQUENCY , kHz
Transverse Electric (TE) Modes
Airborne VLF/LF transmitters use long trailing-wire an- Figure IO-28. Excitation factors vs frequency for lowest three TM modes
tennas that radiate a complicated superposition of TM and Q = 1,2,3) and lowest two TE modes (m = 1,2): ambient
day, u = l@3 S/m [Field et al., 1976].
TE modes. Here much of that complexity is avoided by
considering broadside propagation, where the great-circle
path connecting transmitter and receiver is perpendicular to TEITM Excitation Factors
the plane containing the inclined electric-dipole transmitting Figure 10-28 shows the frequency dependence of the
antenna. The vertical electric field produced by the vertical excitation-factor magnitudes of the first three TM modes
component of the inclined transmitting antenna is given by and the first two TE modes. The first three TM modes are
Equation (10.55). The broadside horizontal electric field excited equally at the lower VLF frequencies, but above
produced by the horizontal component is given by about 30 kHz the higher-order TM modes are much more
effectively excited than the first one. The TE modes are
IL sin $ excited much more poorly than the TM modes, by four or
EH = –120miexp
(“) –~
a five orders of magnitude, as shown in Figure 10-28. The
efficiency of TE mode excitation relative to TM mode ex-

d &- SM’’2A. exp

( .)
-~ (10.55) citation improves as the ground conductivity y is reduced. For
example, at 20 kHz the TE mode excitation factors are
nearly two orders of magnitude greater for a 1&s S/m con-
-d G~(h~)G~(hR) V/m. ductivityy than for 1@3 S/m. On the other hand, the excitat ion
exp(-i: m ) factor for the lowest TM mode is less by almost an order
of magnitude if the conductivity is changed in the same
The symbols are the same as in Equation (10.54), except fashion.
that m denotes the mth TE mode. The excitation factors also depend on the state of the
ionosphere. The excitation factors as defined here and shown
10.2.3.3 TE/TM Mode Structure. Equations( 10.54) and in Figure 10-28 are inversely proportional to a quantity that
( 10.55) show that each mode’s contribution to the field is becomes the “height of the ionosphere” in the limit of a
proportional to the product of four quantities: the excitation sharply bounded ionosphere. For the diffuse ionospheres
factor A, the transmitter height-gain function G(h~), the the excitation factors at the lower VLF frequencies are roughly
receiver height-gain function G(hR), and the propagation proportional to the inverse of the average height at which
factor exp ( – ad/8.7). This section gives calculated values important reflections occur. Thus, one would expect these
of these four quantities for a nominal ambient daytime ion- factors to become somewhat larger under disturbed condi-
osphere and an assumed ground conductivity of 10-s S/m. tions.

10-30
 z
IONOSPHERIC RADIO WAVE PROPAGATION
70 – *8 .
%’ “’’(’1
-~,
60 – \\

\\
\
50 –
~ \rn*l -
‘b rn .2/
E
x ./ \ \
/“
w-
40 – .. .
n ‘\ ,~’ \
3
<, j=3
1-
c ,.
\
‘\
i ‘“ /
\.
a j=2

20
‘~ I
‘\. -\\
\/’
“\\
)/
“- -.-
10 – -.--, --. __,____
/
/
//
“
“10-2
10-1 1 10 lo~ 103

MAGNITUOE OF HEIGHT - GAIN FACTOR

Figure 10-29. Height-gain factors for lowest three TM modes (j = 1,2,3)

and lowest two TE modes (m = 1,2): am~lent day. 20 m=2
kHz, u = I@’ S/m [Field et al., 1976]. j=l
----- ----- _
---- ---- -
\_ _/~=
———. - I
TEITM Height-Gain Factors
The height-gain factor of the waveguide mode accounts
for the effects of non-zero transmitter and receiver heights.
The transmitter and receiver height-gain factors for a given
mode are identical and, therefore, are equal when the trans-
mitter and receiver are at the same altitude. Figure 10-29
I I I 1 I I
shows computed height-gain factors for the first three TM 10 15 20 25 30 35 40 45
modes and first two TE modes for a frequency of 20 kHz.
FREQUENCY, Ictiz
These height-gain factors exhibit the classic height-depen-
dences for antennas over a highly conductive ground; the ,.
Figure 10-30. Attenuation rates vs frequency for lowest three TM modes
TM mode height-gain factors are of the order of unity over (j = 1,2,3) and lowest two TE modes (m = 1,2): ambient
most of the waveguide, except for some rather sharp nulls; day, u = I&’ S/m [Field et al., 1976].
and, above a few kilometers, the TE mode height-gain fac-
tors increase sharply to values well in excess of one-hundti.
For elevated antennas, the large TE mode height-gain factor agation analysis than is mode theory. In intense distur-
mitigates the effects of the small excitation factor, and these bances, however, higher order modes m much more severely
modes can be excited about as effectively as TM modes. attenuated than lower, and the mode sum can be used well
At frequencies above about 30 kHz the first TM mode into the LF regime. Figure 10-30 shows that the first TE
develops a broad maximum in its height-gain factor in the mode is slightly less attenuated than the first TM mode,
40+0 km altitude range. Such “whispering gallery” be- although that result depends on the specific normal daytime
havior is not important for ground-based or airborne ter- conditions and ground conductivity assumed.
minals, but may be significant for very high, balloon-borne, Figure 10-31 illustrates the ground conductivity depen-
terminals operating at higher frequencies [Videberg and Sales, dence of the attenuation rates of the first TM and TE modes
1973]. at 20 kHz. Results are given for normal daytime conditions
TEITM Attenuation Rates and a moderate ionospheric disturbance, such as a solar
Figure 10-30 shows attenuation rates as a function of proton event (SPE). The disturbance increases the atten-
frequency for the first three TM modes and the first two TE uation rate of both polarizations over normal values. How-
modes. The higher order modes are more heavily attenuated ever, the TE attenuation rate is virtually independent of
than the lower, which often allows them to be neglected at ground conductivity, whereas the TM rate exhibits a strong,
VLF for long path-lengths. At the higher frequencies, the broad maximum for conductivities between 105 and lh
attenuation of the higher order modes can be mitigated by S/m, where the TM eigenangle is near the Brewster’s angle
efficient excitation. Under normal conditions, therefore, it of the ground. The TM mode propagates somewhat better
is usually necessary to retain many terms in the mode sums than the TE mode for most common ground conductivities,
for frequencies throughout the LF band. It that case, geo- but propagates much worse over low-conductivity ground,
metric optics is often a more convenient approach to prop- such as occurs throughout Greenland and much of Canada.

10-31
 CHAPTER 10

‘~
—TM I
\ ---TE

NT -

CONDUCTtVtTY,Sletnens/m

Figure 10-31, Attenuation rates of the lowest order TM and TE models

vs ground conductivity, for 20 kHz under normal daytime
and moderately disturbd ionospheric conditions [Field et
al., 1976].
INTENSE= ~
DISTllRBANCE
Although these results pertain to a frequency of 20 kHz, 1\
curves for other frequencies exhibit the same general be-
havior. The main difference is that the Brewster’s angle
peak in the TM attenuation rate occurs at higher values of
conductivity for higher frequencies, and vice versa.
Except for propagation over very low-conductivity ground,
TE modes are more vulnerable than TM mode to degradation Figure 10-33. T~M signal strengths vs distance for ambient day and
in disturbed ionospheric environments. In the VLF band, intensely disturbed ionospheric conditions: 20 kHz, u = I@’
S/m, 10 degree antenna inclination and transmitter and
for example, the TE mode attenuation becomes prohibitive receiver at 12.2 km [Field, 1982c].
for intense disturbances, as illustrated in Figure 10-32.
TE/TM Signal Strengths Versus Distance
Figure 10-33 shows 20 kHz TE and TM signal strengths, rect comparison of TE and TM polarized signals radiated
computed as a function of distance for ambient and intensely broadside from a trailing-wire antenna inclined 10 degrees
disturbed ionospheric conditions. The curves provide a di- with respect to the horizontal. Both the transmitter and the
receiver are assumed to be at an altitude of 12.2 km, and
the propagation path is over poorly conducting earth
,.
10 kHz/ k (u = lt33 S/m).
—TM
/ ,&
-— TE r In Figure 10-33 the ambient signals exhibit nulls and
enhancements at ranges up to several megameters, caused
by interference among several propagating waveguide modes.
At greater distances, the higher order modes, which are
more heavily attenuated than the lower order modes, di-
minish in importance and the signals fall off smoothly with
distance. The curves also reveal several differences between
signals in ambient and disturbed environments. First, and
most im~rtant, if at least 2 or 3 Mm of the path are dis-
turbed the disturbed signals fall well below the ambient
~~ ---
----- -—- signals. Such behavior is typical during strong solar proton
WEU MOOERATE INTEN32
events in the polar regions, and results when such wide-
1 I
01 spread ionospheric disturbances depress the effective height
SEVERITY OF DISTURBANCE
of reflection in the ionosphere significantly. Second, mode
Figure 10-32. Attenuation rates of the lowest order TM and TE modes,
interference patterns are nearly absent in disturbed environ-
for disturbed ionospheric conditions and a perfectly con- ments, indicating that heavy attenuation of the higher order
ducting earth [Field et al., 1976]. modes leaves only the lowest order modes to contribute

10-32
 a
IONOSPHERIC RADIO WAVE PROPAGATION
significant] y to the signal strengths. Third, at ranges under 10.2.3.4 Numerical Modeling of VLF/LF Waveguide
approximately 1.5 Mm, the disturbed signals can be stronger Propagation. Numerous sophisticated computer pro-
than the ambient signals, because the disturbed environment grams have been developed for making VLF/LF field strength
increases the mode excitation factors but destroys interfer- predictions. As described by Morfitt et al. [1982], the model
ence nulls. However, if the disturbance covers most of the developed at the United States Naval Ocean Systems Center
path, but not the transmitter and receiver, the attenuation (NOSC) is particularly attractive in that it incorporates (1)
rates increase but not the the excitations. arbitrary electron and ion density distributions and collision
In Figure 10-33 the ambient TE signal is stronger than frequency (with height), and (2) a lower boundary that is a
the TM signal at all distances, owing primarily to the rel- smooth homogeneous earth characterized by an adjustable
atively high transmitter and receiver altitudes and the nearly surface conductivity and dielectric constant. The model also
horizontal antenna orientation assumed in the calculations. allows for earth curvature, ionospheric inhomogeneity, and
Under the disturbed condition, the TE signal is more ad- anisotropy resulting from the geomagnetic field. In addition,
versely affected than the TM signal and falls below it at air-to-air, ground-to-air, and air-to-ground TE~M propa-
most distances; however, as indicated by Figure 10-31, if gation predictions can be made involving a horizontally
the surface conductivity was reduced sufficiently, the TE inhomogeneous waveguide channel. The NOSC waveguide
signal would again become stronger than the TM signal. model can be used for computing long wave fields at fre-

70 10.897 kHz

40 -
30 “
20 -
10 -
I I I 1 I 1 1 1 I
01234 01234 i

80 I

70 21794 kHz

40 -
30 -
20
10 II 1 1 i I I I 1 I I 1

012345 01234 5

DISTANCE, Mm
Figure 10-34. VLF signal calculations for a mostly seawater path from Hawaii to Sentinel, ArizonaAaytime, ----nighttime [Mortitt, 1977]

10-33
 CHAPTER 10
number of hops required to describe the signal strengths are
less than the number of waveguide modes that are required.
The most general wave-hop formulations describe the prop-
agation of TE/TM waves excited by an inclined dipole over
a spherical earth with an anisotropic ‘ionosphere [Lewis,
1970]. In addition to these sophisticated wave-hop tech-
niques, other (simpler) approaches have been developed that
provide good, quick, estimates of VLF/LF signal strengths,
but without showing such propagation features as wave in-
terference phenomena or polarization conversion effects [e.g.,
see Lewis and Kossey, 1975].

10.2.3.6 VLF/LF Probing of the Ionosphere. The

- \3\: ..--— - . propagation of long radio waves to great distances is con-
-lo
0
, I
I
I
2
“t,
3
, ‘\u
4
J trolled by the lowest regions of the ionosphere (usually the
5
lower E region and the D region). As such, the variations
DISTANCE . Mm
in the amplitudes and phases of propagating long waves are
very sensitive indicators of changes in the lower ionosphere.
Figure 10-35. LF daytime signal calculations for midlatitude propagation
[Pappert, 1981].
It is not surprising, therefore, that in addition to their uses
for long-range communication and navigation long waves
are used to assess the state of the lower ionosphere and as
quencies as high as 300 kHz for daytime propagation [Pap- a tool for characterizing some of its properties.
pert, 198 1] and as high as 60 kHz for nighttime propagation Because of the extremely long wavelengths, ELF waves
[Mortitt et al., 1982]. Figures 10-34 and 10-35 show rep- are affected by the electron and ion densities that are present
resentative TM signal strengths computed by NOSC for over a very large range of altitudes. At night, for example,
nominal daytime and nighttime models of the ionosphere that altitude range can extend from below 50 km, to well
appropriate for midlatitudes, and for propagation over water. up into the F region. Similarly, under disturbed conditions
At VLF, many of the propagation predictions have been the electrons and ions at altitudes appreciably below 50 km
validated by NOSC airborne measurements, such as those can play an important (if not dominant) role in ELF prop-
described by Bickel et al. [1970]. agation. However, owing to the difficulties involved in in-
terpreting long-path ELF propagation data (which tend to
10.2.3.5 Other VLF/LF Propagation Prediction represent an “average” of the state of the ionosphere), ELF
Techniques. In addition to the waveguide mode formu- has not been used extensively for ionospheric research.
lations a number of other mathematical techniques have been The propagation of VLF/LF radio waves is controlled
developed for describing the propagation of VLF/LF waves. by the region of the ionosphere below about 90 km at night
These include the zonal harmonic or spherical wave analysis and below 75 km during the day. Unlike ELF, the obser-
method [Johler, 1964, 1966] and the wave-hop method [Berry, vation of the signal characteristics of VLF/LF waves has
1964, and Berry and Chrisman, 1965]. The spherical wave provided a relatively simple ground-based technique for ex-
technique has the attractive feature that it can model vari- ploring the state and nature of the lower ionosphere, The
ations in height in the earth-ionosphere cavity, but it requires technique has proven to be especially sensitive for moni-
the use of large-scale, very fast, digital computers for its toring ionospheric disturbances, such as those produced by
implementation. For the higher VLF and LF frequencies the solar x ray flares [for example, see Reder, 1969 and Kossey
wave-hop method requires the least computing time, -but it and Lewis, 1974], geomagnetic storms [Belrose and Thomas,
has not been formulated in a way that lends itself to modeling 1968], electron precipitation events [Potemra and Rosen-
of discontinuities in the earth-ionosphere duct [Jones and berg, 1973], ionospheric substorrns [Svennesson, 1973],
Mowforth, 1982]. polar cap absorption events [Oelbemann, 1970], and high
In the wave-hop approach the field strength at any point altitude nuclear bursts [Frisius et al., 1964 and Field and
is the sum of the groundwave (see Section 10.2.1) and a Engle, 1965]. The observation of continuous-wave trans-
series of “hops ,“ which represent waves that have been missions over very long propagation paths has often been
reflected from the ionosphere and/or the ground. The hops used for monitoring because such paths provide coverage
are numbered according to the number of times they have over very large geographical areas. As with ELF, however,
been reflected from the ionosphere. Each reflection results a disadvantage of such long path observations is that the
in a reduction in hop amplitude so that usually a relatively effects of relatively localized disturbances are integrated,
small number of hops are needed to provide good field- or smoothed-out, making it difficult to obtain information
strength estimates. For VLF propagation at distances less on the severity, extent, and structure of the disturbed region
than about 1000 km and for LF propagation, in general, the of the ionosphere.

10-34
 a
IONOSPHERIC RADIO WAVE PROPAGATION
Steep-incidence (that is, short-path) VLF/LF propaga- the geomagnetic field are greatly diminished and the iono-
tion techniques provide data on more localized regions of sphere can be assumed isotropic. Under such conditions the
the ionosphere. However, with continuous-waves the direct mathematical inversion problem becomes somewhat simpler.
and reflected components (groundwave and skywaves) over- Field et al. [1983] have developed an inversion technique,
lap in space and time and can only be resolved indirectly appropriate for isotropic propagation, which has been used
by observing the interference pattern on the ground IHol- in conjunction with VLF/LF pulse reflection data to de-
lingworth, 1926] or by direct interpretation of diurnal phase rive conductivity profiles of the severely disturbed polar
and amplitude changes IBracewell et al., 1951]. For ex- ionosphere.
ample, the interference patterns produced by the ground- A problem with profiles calculated by inversion is that
wave and skywave from a 16 kHz transmitter were used by of nonuniqueness, which can be caused by either incom-
Bracewell and Bain [1952] to first suggest the presence of pleteness of data or the nonlinear dependence of the reflected
two ionized layers well below the ionospheric E-region. signal on the propagation medium. In addition, the profiles
Phase and amplitude observations can be used to char- characterize narrow regions of the ionosphere, since the
acterize the steep-incidence VLF/LF reflection properties of propagation data contain information about only those al-
the lower ionosphere. Of particular interest is the use of the titudes where the ionosphere interacts appreciably with the
data to determine effective heights of reflection and effective reflected wave.
plane wave reflection coefficients of the ionosphere [Brace- The altitude constraints are even more severe if long
well et al., 1951]. Such experimental data can be compared path propagation data are used, rather than steep-incidence
directly with that obtained theoretically, using full-wave reflection data. Nevertheless, some effort has been devoted
computational techniques in conjunction with electron den- to deducing the structure of the ionosphere from long path
sity and colIision frequency models of the ionosphere [for data. As described by Crain [1970], the data in this case
example, see Budden, 1961, Pitteway, 1965 and Inoue and are the attenuation rates and phase velocities of the prop-
Horowitz, 1968]. Thus, the experimental data can be used agating waveguide modes, and the analysis is a trial-and-
to validate theoretical models of the ionosphere, such as error technique effectively to find an ionospheric conduc-
those obtained from the chemistry of the upper atmosphere. tivity profile which provides a waveguide mode or wave
In addition the data can be used to develop and validate hop structure that agrees with the observed distribution of
phenomological models of the lower ionosphere, important radio field strength.
for long wave propagation prediction [for example, see Bain, In essence the long-path technique is similar to the steep-
1982]. incidence approach insomuch as ionospheric reflection coef-
More ~ntiy, with the advent of high-~solution VLFiLF ficients are calculated as an intermediate step in obtaining
pulse ionosounding [Lewis et al., 1973] it became possibIe the mode constants. In order to synthesize the total field as
to observe ionospheric reflections free of the ambiguities measured, one has to take care to add in as many modes as
of the groundwave and skywave interference phenomena contribute to the field. This can add a great deal of com-
characteristic of continuous-wave measurements. The tech- plexity to the application of the technique. Nevertheless,
nique has been used to obtain a variety of steep-incidence the technique has been applied with much success to develop
reflection data at low-, mid-, and high geomagnetic latitudes phenomenological models of the lower ionosphere. Al-
[Lewis et al., 1973 and Kossey et al., 1974], and to in- though such models may not be consistent in all respects
vestigate features of the C-layer of the daytime ionosphere with those derived from detailed analyses of the aeronomy
[Rasmussen et al., 1980 and Rasmussen et al., 1982]. The of the upper atmosphere, they have found widespread ap-
technique provides a relatively direct means for determining plication in long wave propagation prediction codes.
VLF/LF ionospheric reflection heights and effective plane The results obtained by Bickel et al. [1970], Morfitt
wave reflection coefficients, which then can be used to de- [1977] and Ferguson [1980] are especially noteworthy in
velop electron density models of the lower ionosphere [Kos- that regard. They have performed detailed analyses of a
sey et al., 1983]. large volume and a wide variety of VLF/LF propagation
The inversion of steep-incidence VLF/LF reflection data data and have derived analytic models of the lower iono-
to obtain electron density models of the lower ionosphere sphere for propagation prediction. Those models are simply
is not an easy task. Under quiet ionospheric conditions it exponential height-profiles of conductivity, which can be
is especially difficult, since usually the polarization con- specified by only two parameters, scale height and reference
version effects of the geomagnetic field cannot be ignored. height. Following Wait and Spies [1964] the conductivity
Nevertheless, mathematical approaches that employ full- parameter, o,, depends on the ratio of electron density to
wave and iterative computational techniques have been de- electron-neutral collison frequency, and is taken to be of
veloped and applied with some success [for example, see the form w,(z) = 2.5 x 105 exp (~(z-H’)), where z (km)
Shellman, 1970 and Field and Warber, 1984]. Under dis- is altitude, ~ is the inverse scale-height (km-}) and H’ (km)
turbed ionospheric conditions and certain daytime ambient is a reference height. The value of electron density N(z),
conditions, when the VLF/LF reflections are controlled in electrons/ems, is calculated as a function of height by the
primarily by ionization below about 70 km, the effects of equation N(z) = 1.43 x 107 exp(~(z-H’) –O. 15z). The

10-35
CHAPTER 10
Table 10-5. Suggested exponential profiles for use in long wave propagation prediction codes. Frequencies, f, are in kf-fz IMorfitt et al., 1982].

Seasonal-Diurnal
Propagation
Condition H’ (km) ~ (km-’) Magnetic Dip (0)

Summer day 70 0.5

Summer night 87 o.oo77f + 0.31
Winter day 74 0.3
Winter night 80 0.035f – 0.025 90–75
(10 < f < 35) (high latitudes)
Linear change between high and
middle latitudes (transition latitudes)
o.oo77f + 0.31 <70
(middle latitudes)

collision frequency v (collisions/s) is given by associated with ground-reflections on multi-hop paths (see
v(z) = 1.82 X 101] exp( – O. 15z). Table 10-5 gives ex- Figure 10-36). The problem is further compounded in that
ponential profiles, based on VLF/LF propagation data, which the ionospheric absorption losses alone show significant short-
are suggested for use in long wave propagation prediction period and day-to-day variations, as well as diurnal, sea-
codes [Mofitt et al., 1982]. sonal, latitudinal, and solar-cycle effects [Knight, 1982].
A number of techniques have been developed to estimate
MF field strengths. A relatively simple technique, based on
10.2.4 MF Propagation an empirical formula has been adopted by the CCIR [CCIR,
1978]. A more complex wave-hop method has also been
At night medium frequency skywaves can propagate to developed by Knight [1975]. Figure 10-37 shows computed
considerable distances with relatively little attenuation, but ionospheric losses over the 500-1500 kHz band for single-
during the day the skywaves are severely attenuated in pass- hop paths at mid- and low geomagnetic latitudes. In the
ing through the ionospheric D-region, so that only the auroral zones the ionospheric losses are somewhat greater
groundwave provides usable signals. Thus, during the day- than those shown in the figure, and the estimates do not
time MF signals are very stable, while at night they are apply if the waves penetrate the E-layer and are reflected
much less so owing to the variability of the lower E-region by the F-layer. The latter is most likely to occur at fre-
of the ionosphere, and to interferences between the ground- quencies above 1500 kHz [Knight, 1982].
wave and skywaves.
10.2.4.3 Effect of MF Waves on the Ionmphere. Even
10.2.4.1 MF Groundwave Propagation. The propaga- relatively small electromagnetic fields impart appreciable
tion of MF groundwaves can be described using the tech- energies to the electrons in the ionosphere causing their
niques discussed in Section 10.2.1. Because of the shorter temperatures, and consequently their thermal velocities, in-
wavelengths, however, such factors as the emh’s atmos- crease. This increases the effective electron-neutral collision
phere (and hence, the effective radius of curvature of the frequency; as a result, the complex dielectric constant of
earth), terrain elevation, conductivity y changes, and trees and
buildings along the propagation path usually influence MF
groundwave characteristics to much larger extents than they
do for VLF/LF waves. Most of these effects can produce
strong local interference patterns in the amplitudes of MF
groundwaves [Knight, 1982 and Hizal and Fer, 1982].

10.2.4.2 MF Skywave Propagation. Because there are

so many factors that affect the characteristics of MF sky-
waves, it is difficult to draw a representative set of propa-
gation curves. These factors include the losses associated -~? TRANSMITTER

with an imperfectly conducting earth at the transmitter and

receiver, polarization coupling losses that depend on the Figure 10-36. Losses associated with MF skywave propagation [Knight,
geomagnetic field, ionospheric absorption losses, and losses 1982].

10-36
 a
IONOSPHERIC RADIO WAVE PROPAGATION
I 10.2.5 Long Wave Propagation Through
m the Ionosphere
u
~“ I
8
~ 10 Electromagnetic waves cannot propagate in an ideal
~8 -
plasma unless the wave frequency is less than the plasma
frequency, or approximately 8980 N, where N is the number
~ 15 of electrons per cubic centimeter. Thus, a density only slightly
z 4-
g —— ________
0.s-15 more than 1 el/cm3 would suffice to completely reflect a 9
2-
kHz wave. The well-known “whistler” phenomenon, how-
04 ever, demonstrates that under certain conditions long waves
567 6910111218151720
W LEMGTH, km x W can penetrate even through the F-max region of the iono-
sphere, where the electron density is one hundred thousand
Figure 10-37. Computed MF ionospheric reflection losses:+ast-west times larger than would produce complete opacity if the
propagation at all latitudes, ---north-south propagation at ionosphere were a simple plasma. The long wave ionos-
the magnetic equator [Knight, 1982].
pheric transmission window is due to the geomagnetic field
of the earth, which constrains the electron motion produced
the medium becomes appreciably dependent on the field. by electromagnetic waves incident on the ionosphere. As
Thus, the associated physical processes, and the differential such the magnetic field provides a propagation mechanism.
equations which describe the radio wave propagation in the The term “whistler” refers to an audio-frequency phe-
ionosphere, become non-linear. This gives rise to various nomenon associated with lightning discharges in the lower
phenomena, including cross-modulation and de-modulation atmosphere. Electromagnetic energy at audio frequencies
which, in principle, can be observed experimentally. The emitted by such discharges propagate in the ionosphere in
effects depend on wave frequency and the collision fre- a highly dispersive manner. The higher frequencies travel
quency and are such that they have been mainly observed faster than the lower ones with the result that the signal,
in the MF band, particularly at night. which was originally impulsive, is received over a relatively
The main influence of the non-linear effects on the prop- long time interval with the frequency generally descending
erties of radio waves reflected from the ionosphere manifest with time (hence, a whistling sound).
themselves through self-interference of the wave; that is, its Extensive studies have been made to determine the prop-
influence on itself, and the interaction between a number erties of whistlers [for example, see Storey, 1953, Ratcliffe,
of waves. These waves can be of the same frequency, or 1959, Pitteway, 1965 and Helliwell, 1965]. Their results
two independent modulated or unmodulated waves of dif- [Watt, 1967] show that from the complete expression of the
ferent frequencies. An early observation of such an effect, Appleton-Hartree equations for the refractive index, ap-
the so-called Luxemburg effect, was reported by Tellegen proximate expressions can be developed which give insight
[1933] who noted that the signal received in Holland from into the nature of the whistler mode. For example, the en-
a 650 kHz Swiss station appeared to be modulated by the ergy transmitted through the ionosphere is well-coupled to
signals from a powerful station at Luxemburg (252 kHz). the whistler mode when the direction of propagation is in
This phenomenon of cross-modulation has been found to be the same direction as the earth’s magnetic field. For trans-
quite common when the unwanted, or disturbing, station is verse propagation, this coupling is very poor. Also, the
situated near the transmission path of the wanted wave [see coupling is increased for sharp gradients of refractive index
Davies, 1969 and Al’pert, 1960]. at the ionosphere boundary. When collisions are included,
Although the phenomenon of ionospheric cross-modu- there is a finite range of angles between the wave normal
lation or Luxemburg effect was discovered accidentally, it and the direction of the earth’s magnetic field for coupling
has been systematically investigated ever since its discov- to the whistler mode. In an anisotropic ionosphere, the wave
ery. The motives of the earliest theoretical and experimental normal and the direction of energy flow along the field lines
investigations were interest in the effect itself and the pos- are different depending upon the magnitude of the refractive
sible use of the effect for probing the ionospheric regions index and the static magnetic field. If ducts of ionization
in which the cross-modulation occurs [Fejer, 1970]. In the that are aligned with the earth’s magnetic field exist, the
latter case techniques have been developed to obtain D- waves can be guided in them.
region parameters such as electron density and collision VLF signals from terrestrial transmitters have been ob-
frequency profiles. served at satellite altitudes [Leiphart et al., 1962], and have
Another MF probing technique, the “partial reflection” been tracked from ground level to altitudes of 500 km or
experiment, has proven to be a valuable method for ob- more [Orsak et al., 1965 and Harvey et al., 1973]. Such
taining quantitative measurements of electron distributions probes showed that the polarization changes from linear to
in the 10wer ionosphere. It has been one of the most exten- circular as the wave penetrates the ionosphere, and that very
sively employed techniques for synoptic studies of the ion- significant delays in the signals occur.
osphere below 100 km IBelrose, 1970, 1972]. Such phenomena can be calculated using full-wave tech-

10-37
 CHAPTER 10
the layers, and the half-thickness widths** y~E, y~Fl; and
y~F2. These models are called phenomenological models
[Barghausen et al., 1969; Bent et al., 1972; Ching and Chiu,
1973; Chiu, 1975; and Kohnlein, 1978].
The ITS-78 model [Barghausen et al., 1969] based on
so the analysis by Jones et al. [ 1966] of world-wide, ground
based ionosonde data, predicts only the bottomside of the
40
ionosphere. The Bent model [Bent et al., 1972] predicts the
30
1
total electron content of the ionosphere in the altitude range
20 from 150 to 2000 km, without a direct consideration of the
t ,,
bottomside E and F] layers. The Ching and Chiu [ 1973]
0 -
model covers the altitude range from 110 to 1000 km. In-
L= 1 1 I
10-2n-’ xr’J 10’ 101 10=n~ 05 De 10’ log 09 loo
10-3 stead of parabolic layers [Barghausen et al., 1969] they
FREQUENCY , HZ assume Chapman functions for the electron density distri-
butions in the E, F1, and F2 layers. Later Chiu [ 1975]
Figure 10-38. Computed long wave transmission losses through ambient modified the Ching-Chiu model to incorporate the polar iono-
day and night ionospheres [Booker et al., 1970]. sphere. Their models are useful only for studying the large
scale phenomena such as global thermospheric and iono-
niques [Pitteway, 1965] or WKB methods [Booker et al., spheric calculations.
1970], in conjunction with appropriate models of the ion- Using ionospheric data from ESRO satellites, Kohnlein
osphere. Figure 10-38 gives computed long wave trans- [1978] extended the altitude range up to 3500 km. He sug-
mission losses for plane waves incident on the ionosphere gested a “differential approach” for the ionospheric mod-
in a direction parallel to the geomagnetic field. The results eling. He separated small scale spatial structures such as
are representative of those expected for ambient ionospheric the equatorial trough, the midlatitude trough and the polar
conditions at mid- and high geomagnetic latitudes. Under ionosphere, from the large scale global structure. He mod-
disturbed conditions or at very low geomagnetic latitudes eled these individual structures and added them into the
the penetration losses are much more severe, especially for global structure. His method reduces the number of coef-
frequencies above a few kilohertz [Booker et al., 1970 and ficients otherwise needed to model the complicated iono-
Harvey et al., 1973]. spheric behavior.
The other approach for ionospheric predictions is to use
theoretical models [Stubbe, 1970; Strobe] and McElroy,
10.3 IONOSPHERIC MODELING 1970; Nisbet, 1971; Oran et al., 1974; and Oran and Young,
1977] .These are based on the physical processes responsible
For successful radio communication, it is essetttial to for the production, maintenance and decay of the iono-
predict the behavior of the ionospheric region that will affect sphere. A theoretical model would thus rely on the process
a given radio communication circuit. Such a prediction will of ionization of neutral atmospheric constituents by the in-
identify the time periods, the path regions and the sections cident solar extreme ultraviolet radiation, the transport pro-
of high frequency bands that will allow or disrupt the use cesses such as diffusion and neutral winds, and also on the
of the selected high frequency communication circuit. This effect of electric and magnetic fields on the transport pro-
need for prediction leads to modeling of the ionosphere. cesses. Essentially the theoretical model tries to explain the
A model is a numerical statistical description of the experimental observations in terms of known physical pro-
ionosphere in terms of location (geographic or geomagnetic cesses. In addition, this approach seeks new physical pro-
latitudes and longitudes), time (solar zenith angle), seasons, cesses to explain the differences between the observational
and other factors such as the solar activity (10.7 cm flux, results and the predictions based on the theoretical consid-
sunspot number). The empirical equations are derived from erations.
the dependence of the observed phenomena on variables Radio communication can be divided in two main cat-
mentioned above. These observed phenomena include: the egories. Ground-to-ground radio communication is based
behavior of critical frequencies* foE, foF 1, foF2, and foEs on the reflection and scattering characteristics of the iono-
for the E, F 1, F2, and sporadic E layers; the altitudes (h~E, spheric layers. On the other hand, ground to satellite, satellite
h~F1, and h~F2) for peak (maximum electron densities for to ground, or satellite to satellite radio communications de-
pend on the transmission and refraction characteristics of
*The critical frequency is the limiting radio frequency be-
low which a radio wave is reflected by, and above which **The half-thickness width y~ of the ionospheric iayer is
it penetrates and passes through, the ionized medium (an determined with the assumption that the layer has a parabolic
ionospheric layer) at vertical incidence. shape [Appleton and Beynon, 1940].

10-38
 w
IONOSPHERIC RADIO WAVE PROPAGATION

the ionosphere. The main goal of any modeling effort is to Time, geographical location (latitude and longitude) of the
predict the periods of good or poor radio communications transmitter and receiver, and sunspot number. To compute
for the selected paths to enable a continuous undisrupted the system performance the model needs the antenna pa-
communication through the ionosphere or by some other rameters, the radiation &wer of the transmitter, and the
means. signal to noise ratio of the receiver.
We will consider several ionospheric models that are For the D region the model considers only the absorption
routinely used (or are available) for the prediction and spec- losses. The non-deviative absorption is in the form of a
ification of the ionosphere. The emphasis here is on ac- semi-empirical expression. It enables the user to compute
quainting the user with the modeling programs and their the losses for the HF frequencies penetrating the D layer.
limitations. We do not attempt to review the scientific lit- The deviative absorption losses are included in the loss
erature for a determination of the state of the art of modeling calculations as uncertainty factors.
efforts. Therefore only the essential references will be cited. For the E region the model computes the parameter foE.
First, we will consider the numerical-phenomenological It assumes a constant height of 110 km for the maximum
models. Then we will consider the theoretical models. This (peak) electron density of the E layer, with a constant semi-
will be folIowed by the modifications to models to take into thickness of 20 km. The numerical coefficients for foE are
account high latitude phenomena such as the auroral E layer based on the experimental ground ionosonde data during
and the midlatitude F region trough. In the concluding sec- high solar activity phase in 1958, and the low solar activity
‘;-- .s’= will look at thj limitations of these models and a
LIW1l, WC phase in 1964.
possible approach to overcome these limitations. For the F2 region the model computes the parameters
foF2, the height of maximum electron density h~F2 and the
semi-thickpess y~F2 of the F layer. These are in the form
10.3.1 The Numerical-Phenomenological
— of numerical coefficients for the high (1958) and low (1964)
Models phases of the solar activity. Both the E and F2 layers are
assumed to be parabolic in shape.
At present the three most widely used numerical models The sporadic E (Es) layer could be very helpful or harmf-
for ionospheric predictions are (1) The ITS-78 model, (2) ul to radio communications depending on the nature of the
The Bent model, and (3) the Ionospheric Communications Es layer. A blanketing, totally reflecting Es layer extends
Analysis and Prediction Program (IONCAP). In addition, the frequency range of the E-mode communications. How-
the 4-D model of the Air Force Global Weather Central and ever a semi-transparent or partly reflecting E, layer would
the Bradley model will be considered. We will also look at cause serious multipath and mode interference and would
the International Reference Ionosphere-IRI 79. be detrimental to communication systems. Using numerical
coefficients, the ITS-78 model computes foE, only for the
10.3.1.1 The ITS-78 Model. The main pu~se of this ordinary ray. (The earth’s magnetic field splits the incident
model [Barghausen et al., 1969] is to predict long term ray into the ordinary and the extraordinary rays. ) The nu-
perfo~ance of communication systems operating in the merical coefficients are for both the high (1958) and low
2–30 MHz frequency range. The ITS-78 model and its com- (1964) phases of solar activity. As the model does not pre-
puter program was developed by the Institute of Telecom- dict the occurrence of Es, the foEs maps are used only when
munication Sciences, ESSA, Boulder, Colorado. The model propagation via regulm E layer is not possible. To compute
is based on the presentation of the ionospheric characteristics the system performance, the model incorporates three kinds
in a form of synoptic numerical coefficients developed by of noise: galactic, atmospheric, and manmade.
Jones and Gallet [ 1960] and improved by Jones et al. [ 1966]. To determine the operational parameters such as the
The important features of the ITS-78 model are the pa- maximum usable frequency (MUF), the model computes
rameters for the D, E, Es (sporadic), and F2 layers of the the path geometry (between the transmitter and the receiver).
ionosphere. The parameters computed in the path geometry are the path
The model provides (output) circuit operational param- length, path bearing (azimuth), and the solar zenith angle
eters such as the maximum usable frequency (MUF), op- x of the sun. The model computes paths for reflections from
timum traffic frequency (FOT), and the lowest usable fre- E, Es, and F2 layers. These are called the E, Es, and F2
quency (LUF). In addition to the regular E-layer propagation modes. The paths could involve more layers (multiple modes)
mode, it takes into account propagation via the sporadic E and more reflections (multihop).
layer. The program computes all the probable modes. It To determine wave propagation the electron density dis-
computes the system @rformance. For that purpose it cal- tribution with altitude is needed. Both the E and F layers
culates the antenna patterns and gains for 10 most commonly are assumed to have ptiabolic shapes. The maximum usable
used antennas. It also has a program to determine MUF as frequency (MUF) is obtained by multiplying the critical
a function of the magnetic activity index Kp. frequency of the layer by the MUF factor M(3000). The
The inputs for the ITS-78 model are the date, Universal term in parentheses refers to the standard ground distance

10-39
CHAPTER 10
of 3000 km between the (hypothetical) transmitter and the The data base of the Bent model consists of 50000
receiver. me experimental data for the numerical factor topside ionospheric soundings, 6000 satellite measurements
M(3000) (in terms of coefficients) come from 13 ionosonde of electron density and 400 000 bottomside soundings of
stations covering the geomagnetic latitude range from 7 10S the ionosphere. The data extend from 1962 to 1969 to cover
to 88”N. the maximum and the minimum of the solar cycle.
The stability and predictability of the E layer results in The bottomside data are foF2 hourly values from 14
a 9990 probability (highest) of supporting radio propagation stations in the American sector covering geographic lati-
and communication via the E layer. The next highest prob- tudes 76° N to 12° S or geomagnetic latitudes from 85° to
ability is via the regular F layer. When neither of the above OO.The topside soundings cover the period 1962 to 1966,
modes is possible, the Es mode is considered for commu- with geomagnetic latitude range 85° to – 75°, and the elec-
nication. tron density profiles are from 1000 km down to the altitudes
For computing the system performance the program al- of the foF2 peak (h~F2). The satellite data are from the
lows a selection from 10 antenna patterns. The program Ariel 3 satellite covering the period May 1967 to April 1968
takes into account the ground losses, ionospheric losses, and are linked with real time foF2 from 13 ground stations.
free space losses, and the excess losses. The program com- Thus the data base of the Bent model refers to solar cycle
putes the radio communication circuit reliability, ‘service 20, while the data base of the ITS-78 model is from solar
reliability, and the multipath evaluations. cycle 19.
The ITS-78 model has several limitations. The results The Bent model uses foF2 from the ITS-78 model. In-
from the model are useful only when the operating frequency stead of the monthly median values, the Bent model com-
is below the maximum usable frequency. The model as- putes average values for every 10-day interval of the month
sumes that transmission will be by reflection from the ion- from the 10.7 cm solar flux input. For the height h~F2, it
osphere. For this the transmitter and the receiver must be uses M(3000) factors of NOAA (ITS-78) in terms of the
on the same side of the ionospheric layer (for example, sunspot number. It uses an empirical polynomial for M in
ground-to-ground communication. The model does not take computing h~F2, in place of the Shimazaki equation [ 1955]
into account the daytime F 1 layer which usually develops used by ITS-78.
between the E and F2 layers. The model does not adequately The distribution of electron density with altitude, as-
account for the electron density above the altitude of h~F2. sumed by the Bent model for the computation of the total
Finally, the model does not take into account the dependence electron density is shown schematically in Figure 10-39.
of absorption on the operating frequency in considering the Starting from the bottom, it divides the profile into five
D layer absorption. sections; a bottom hi-parabolic F2 layer; a parabolic F2 layer
above the peak; and three exponential sections to cover the
10.3.1.2 The Bent Model. The Bent Model [Bent et al., altitude above hO (h~F2 to 1000 km). The construction of
1972; Llewellyn and Bent, 1973] is basically for ground- the profile needs the parameters k 1, k?, ks, Y,. Y“,, foF2.
to-satellite communication but can be adapted for ground- and h~F2. The last two have already been explained. The
to-ground or satellite-to-satellite communication. The main dependence of the other parameters on geomagnetic latitude.
purpose of the Bent model is to determine the rota] electron solar flux, foF2, and the season is from work of Bent ILlew-
content (TEC) of the ionosphere as accurately as possible ellyn and Bent, 1973]. The topside and the first adjoining
in order to obtain high precision values of the delay and exponential section are matched at a height d (above h,,,)
directional changes of a wave due to refraction. Ground-to- by the equation
satellite communication demands operating frequencies which
are higher than the MUF. Thus the mode involves the trans-
mission refraction characteristics of the F2 layer and the
electron density distribution above the height of the F2 peak
d=ho– h.=:
[v
1 + Y:k:–
1
1 , (10.56)

must be known. where Y, is the half thickness of the F2 layer and k] is the
The model provides (output) the vertical total electron exponential constant.
content above the transmitter, the profile of vertical electron The remaining profile above the F2 peak [of altitude
density with altitude, and the total electron content along range (h~ + d to 1012) km] is divided in three equal in-
the path bet ween the satellite and the ground. It also provides tervals of altitude.
the refraction corrections to the elevation angle, the range, The model can predict with an accuracy of 75%-80%.
and the range rate. If the model is updated with observed recent data within a
The input parameters to the model are the date, Universal range of 2000 km radius (from the transmitters), the pre-
Time, locations of the transmitter and receiver (ground and dictability is improved to 90Yc.
satellite), rate of change in elevation and altitude of the Though the model predicts total electron content (TEC)
satellite, operating frequency, the solar flux (10.7 cm flux), with good accuracy, the model does not have separate E
and sunspot number. and F 1 layers. As the model was constructed for the TEC,

10-40
 -

IONOSPHERIC RADIO WAVE PROPAGATION

HEIGHT h In the system performance options, 22 performance pa-

.— rameters are available. The program for the antenna output
1000 kn
option computes the elevation angles and the operating fre-
quencies for optimum antenna geometry and its gain.
-k3a~
Inputs for the program are the date, Universal Time,
~a, N=N2e
h2 --- geographical locations of the transmitter and the receiver,
and sunspot number. The program can accept external iono-
spheric parameters as input to the program. For antenna
N= Nle-%a2 pattern, one can select the antenna from 17 antennas in the
/
h, .- fa’L--- program (7 antennas from ITS-78 have been modified). For
i 1
the system performance additional inputs such as radiation
power of the transmitter, and the S/N ratio of the receiver
are needed.
hoy The schematic for the electron density distribution with
hm altitude for the IONCAP program is shown in Figure
10-40. The model has 3 parabolic layers, E, F1, and F2.
Y The altitudes for the peak electron densities are h~E, h~Fl,
and h~F2. The half thickness widths for the layers are Y~E,
Y~Fl, and Y~F2 respectively. For the F1 layer, h~F1/Y~Fl
is assumed to be 4. The E layer has fixed altitudes h~E = 110
km and Y~E = 20 km. IONCAP improves on the ITS-78
model by incorporating D and F 1 layers. The D layer con-
il tribution is considered indirectly by adding an exponential
NON. ELECTRON
DENSITY, N tail for the E layer down to the altitude of 70 km. In the
transition region between the E and F1 (or F2 if F1 is absent)
Figure 10-39. Schematic for exponential and hi-parabolic profiles for the
electron density distribution with altitude for the Bent medel.

the E and F1 layers are included as the hi-parabolic bot-

tomside of the F2 layer. Also, the program does not take
into account the non-deviative absorption in the underlying
D layer.

10.3.1.3 The Ionospheric Communications

Prediction Program (IONCAP).
Amdysis and
The IONCAP [Lloyd
EIGHT, km

~mfq _________ .__–-–.

) VIRTUAL

et al., 1978] is essentially the latest, improved, and more

versatile and flexible version of the ITS-78 model.
It provides 30 output options which can be divided into
four categories, (1) ionospheric description, (2) antenna pat- PARABOLIC NOSE
——
hmFI—–– —-. -——
terns, (3) MUF predictions, and (4) system performance
predictions. t
mF! / LINEAR (or PARABOLIC) ,LEDGE
For ionospheric predictions it provides monthly median 0
/
0 I
values for the parameters foE, foFl, foF2, h&, h#l, h~, / I
I
/ I I
Y~E, Y~Fl, and Y~F2. It also provides the lower, median, 0
/ I
—— -- --- I
and upper decile values of the minimum foE or foE,. It can I
I
also provide a prediction in the form of a plot of operating (fv ,hv) LINEAR VALLEY I
frequency with virtual height and also with true height. I
The MUF option of the output provides the minimum I
I
radiation angle and the M factors for all four modes, E, F 1, 1
F2, and Es. The plots for the diurnal variation of the MUFS I
&
r
are also available. The MUFS provide the description of the I foE foF! foFz
state of the ionosphere and do not include any system pa- FREQUENCY
rameters. The operating frequency for a given radio com-
munication circuit is the critical frequency of the layer mul- Figure 10-40. Schematic for the electron density profile and virtual height
tiplied by the MUF factor. for the IONCAP model.

10-41
CHAPTER 10
layers, the electron density is assumed to be linear for the a reduction in electron density between the two peak values,
frequency range f, to fu where f, = xv X foE and fu = XU N~E and N~F—at altitudes h~E and h~F, respectively. The
x foE. Typical values for xv and Xu are 0.85 and 0.98 rocket observations have shown that in reality the electron
respectively. Thus the electron density decreases above the density at any altitude between the altitudes h~E and h~F
parabolic nose 0.85 foE (<foE) and continues upwards up is rarely smaller than the peak electron density N~E. Thus
to 0.98 foE (<foE) producing a linear valley in the transition the assumption of the parabolic shapes for E and F layers
region. When the x. = XV = 1 the valley is absent in the underestimates the electron density in the altitude region
transition region and the curve is a vertical line starting at between h~E and h~F. To correct such an underestimation
the tip of the parabolic nose of the E layer. The F1 layer Bradley and Dudeney [1973] suggested a linear distribution
forms a linear or parabolic ledge depending on the mag- of electron density from foE to 1.7 foE. The lower end is
nitudes of h~Fl, h~F2, foFl, and foF2. In the ITS-78 model, at h~E. At the upper end, the F layer is parabolic in shape
the F1 layer is assumed to be absent. In the IONCAP model down the altitude where the plasma frequency is 1.7 foE.
the numerical coefficients for foE are functions of geo- This linear interpolation has not yet been incorporated in
graphic latitude for both solar maximum and minimum from the IONCAP model (see Figure 10-40) of Lloyd et al. [1978].
the work of Leftin [ 1976]. The model uses foFl maps of In high frequency prediction it is essential to know the
Rosich and Jones [1973]. It also takes into account the probability of communication at any particular operating
retardation below the F2 layer. frequency f. For convenience the observed data are ex-
For the MUF computations the model uses the corrected pressed as follows: Fu is the upper decile, f~ is the median,
form of Martyn’s theorem. As the absorption equations us- and Fe is the lower decile of the ratios of f/f~. The distri-
ing the secant law do not work for lower frequencies at bution functions of F. and Fe are not simple Gaussian dis-
altitudes below 90 km, these equations have been modified tributions (Fu – MUF = MUF – Fe for Gaussian). The
in the IONCAP program. The IONCAP provides two pro- distributions are X2-distributions of Fu and Ft. For proba-
grams (1) the ITS-78 short path geometry and (2) the long bility determinations these two X2-distributions (of F. and
path (> 10000 km) geometry. In addition to the ITS-78 F?) have to be used. Two variables FU and Fe with their
model, the path computations now include the F 1 mode, associated degrees of freedom, and the need to integrate the
the over-the-MUF mode, D and E region absorption losses, X2-distribution curve makes the process of determining the
and sporadic E losses. A correction to frequency dependence probability distribution very cumbersome. Bradley and Bed-
is added for low frequencies reflected from altitudes below ford [1976] derived simple empirical equations for this prob-
<90 km. abilityy distribution. The equations are
The improvements over ITS-78 can be summarized (see
IONCAP) as follows:
Q= 130_ 80
1. The description of the ionosphere is now more com- ~ _ 1 – (f/fm)
plete.
1 –Fe
2. The loss equations have been supplemented. This (10.57)
includes E mode adjustments, sporadic E effects, over- or 100, whichever is smaller for f<f~
the-MUF losses, and losses for low reflection heights.
3. The ray path geometry calculations have been re- and
vised. This was an empirical adjustment to Martyn’s
theorem. Q= 80 – 30
4. The loss statistics were revised to include the effects ~ _ (f/fm) – 1
of the sporadic E layer and of over-the-MUF modes. FU–l
5. A separate long path model was developed. (10.58)
6. The antenna gain package was revised. or O, whichever is larger for f>f~,
All models pre~ict ;nly the quiet ionosphere, which
shows a large systematic dependence on latitude, longitude, where
season, time, and sunspot activity.
Q - is the cumulative probability,
10.3.1.4 The Bradley Model. The Bradley model con- f - is the operating frequency,
tains two modifications to the existing models: (1) the filling f~ - is the predicted median frequency,
of the valley between the E and F layers (Fl, or F2 if F1 Fe - is the lower decile (of f/f~),
is absent), by parabolic layers [Bradley and Dudeney, 1973] F. - is the upper decile (of f/f~).
and (2) a simple formulation of the prediction of the prob-
ability of the high-frequency propagation [Bradley and Bed- They note that the probability distribution from their
ford, 1976]. simple empirical equation is as good as, though not always
The assumption that the electron density distribution in better than, that from the X2 distribution. Therefore they
the E and F layers is parabolic in shape, results in a valley— highly recommend a replacement of X2 distribution proce-

10-42
 a
IONOSPHERIC RADIO WAVE PROPAGATION
dure by these equations for a determination of the present 10.3.1.6 International Reference Ionospher%IRI 79.
probability that signals will propagate at a given hour over The IRI 79 [Rawer, 1981 ] is the latest addition to the con-
a given sky-wave path. tinued efforts of ionospheric modeling. The emphasis of lRI
The latest computer model like the IONCAP has not 79 is to summarize the experimental data from rockets and
incorporated “Bradley Features” in its program. satellites to provide true height profiles of the ionosphere.
The model serves as a standard reference for various pur-
poses such as design of experiments, estimation of envi-
10.3.1.5 The Air Force Global Weather Central 4-D
ronmental and other effects, and testing theories. The model
Model. The input data to the 4-D numerical model [Flat-
gives the altitude dependence of four parameters: electron
tery et al., personal communication] are the critical fre-
density, electron and ion temperatures, and the composition
quencies for the layers and M(3000), real time or near real
of positive ions. It computes the density for atomic ions
time observations from 40 ground stations around the world,
0+, H+, He+ and for molecular ions Oz+ and NO+ .
and total electron content (TEC) from eleven stations. The
For the worldwide description of the peak electron den-
frequency of observations varies from hourly (best) at one
sity, the model uses foF2 from CCIR [ 1967] coefficients,
end to weekly (worst) at the other end. The desired purpose
with modified dip coordinates [Rawer, 1963]. As the foF2s
of the 4-D model is to produce a consistent ionospheric
are from the ground based ionosonde stations, the model
specification anywhere in the northern hemisphere for a 24-
really computes a relative distribution of electron density
h period. In that sense it is not a forecasting model like the
with true height, with respect to that of the foF2 peak. For
other models mentioned above.
a true peak height the model uses an empirical relation
This model has three ionospheric layers, E, F1, and F2.
[Bilitza et al., 1971] with M(3000)foF2 coefficients from
Each layer is represented by a Chapman distribution function
CCIR [ 1967]. This empirical relation is based on the in-
coherent scattering measurements which yield electron den-
N,(h) = Ne~aX exp {a[ 1 – z – exp ( – z)]} , (10.59) sity with true height. The model has an alternate procedure
based on the results of Chiu [ 1975] to replace the foF2 and
M(3000)foF2 coefficients from CCIR. This procedure when
used limits the ability of reproducing the complex iono-
where a refers to the loss mechanism and z is given by
spheric features available from the CCIR coefficients. The
z = h–h~,X
— and h, is the scale height for the layer. At a IRI 79 can also use direct data of the peak electron density
h,
and the peak altitude for computing the profiles.
given altitude the total contribution to electron density is
The schematic for the altitude dependence of the electron
the sum of contributions by all three layers.
density for the IRI 79 model is shown in Figure 10-41. The
For any height the electron density is approximated by
altitude range from 80 to 1000 km is divided into six sec-
tions: topside, F2 bottomside, F 1, intermediate, valley, and
Ne(t) = ~ a~W~((), E/D regions respectively. The topside region is modeled
k=l with the use of ‘harmonized Bent’ model [Ramakrishnan et
al., 1979]. The bottomside F2 is expressed as the sum of
where a is the weighting factor and W(t) is an empirically Epstein Transition Functions [Rawer, 198 1]. The F I region
derived set of discrete orthogonal functions for the altitude is based on the work of Eyfrig [ 1955] and Ducharme et al.
interval t. The 95 to 2000 km range is divided into 127 [1973]. The intermediate region fills the gap between the
intervals. The widths of the intervals range from 5 km at valley region and the F1 layer. The rocket measurements
the lowest altitudes to 50 km at highest altitudes. The em- compiled by Maeda [ 1971] determine the shape and the
pirically derived function W(t) is in two parts, spherical depth of the valley region. The foE is from Kouris and
harmonic functions for spatial dependence and trigonometric Muggleton [ 1973a,b]. The model also takes into account
functions (sine, cosine) for temporal dependence. the contribution from the D layer. The model does not
With the help of these variables a~ and Wk, the entire account for the highly variable Es layer.
data base for the ionosphere is reduced to a limited number The IRI 79 is the only numerical model with information
of coefficients. These can be used to construct the electron on additional parameters such as the electron and ion tem-
density profile for any location in the Northern Hemisphere peratures and the composition of positive ions. The com-
valid for a 24-h specification period. The model is still being position is determined with the assumption that the plasma
developed. The specification accuracy of the model will is electrically neutral above 84 km. The model also com-
depend strongly on the frequency and reliability of the input putes the distribution of cluster ions in the altitude range
data-real-time experimental observations from the 40 ground 80 to 90 km.
stations of the northern hemisphere. Also the quality of The inputs for the program are location (latitude and
specifications interpolated for locations inside the network longitude), sunspot number and time. The optional inputs
will be better than those extrapolated outside the observa- are the peak altitude and peak electron density. The output
tional network. consists of 11 parameters: absolute electron density, relative

10-43
CHAPTER 10
The programs are in FORTRAN-4 and ALGOL-60 com-
puter codes. These programs are available from the World
Data Center, Boulder, Colorado.

Topside (1)
\
10.3.2 The Theoretical Models
HMF2
I The theoretical models for the ionosphere are based on

I F2 12)
the physical processes responsible for the observed ionos-
pheric phenomena. The processes responsible for the ion-
osphere are production, maintenance, and decay of the ion-
HMF1 -. osphere. As the approach deals directly with the physical
F1 {3]
HZ processes, and not with the observed phenomena, the emerg-
HST - – —Intermed. (L)
ing model is called a physical model.
I
Four models are summarized in Table 10-6 to show
HEF -
several variations in the same processes considered by dif-
ferent workers. Strobel and McElroy [ 1970] considered only
--E-Valley (5)—
the F2 region (200 to 700 km), whereas others took into
HME account the altitude range from 120 to 1200 km. Nisbet
t
[197 1] constructed the first computer-based simple physical
HA model MK- 1 for the ionosphere. He considered only three
I neutral constituents NZ, OZ and O, whereas Stubbe [19701
and Oran and Young [1977] also considered the minor con-
NME NMF2
stituents He and H. For the dissociation and ionization of
log N ~
the neutral species, the incident solar EUV radiation in the
range 30 to 1912 ~ is used, along with the wavelength
Figure 10-41. The lRI 79 model profile. For details refer to Rawer [1981].
dependent absorption and ionization cross sections for the
neutral species. Nisbet considered three basic predominant
electron density, neutral temperature, electron temperature, ionic species: O+, NO+, and 02+. Oran and Young [ 19771
ion temperature, ratio of electron to ion temperature, percent took into account the additional ionic species H?+, Ne +.
concentrations of O + (and N+), H+, H.+, Oz + and NO+ N+ and H+. One has to consider the chemical reactions
ions. The accuracy (u) of predictions is as below: that produce ions by charge exchange processes. Nisbet
[197 1] used 5 reactions whereas Oran and Young [ 1977]
Peak used 24 chemical reactions [see Strobel and McElroy, 1970].
Height Density Temperature For maintenance of the ionosphere, the processes of dif-
F region t 15% f 3070 * 30% fusion and photoionization are assumed. The processes of
E region t 5~o t 10% * 1070 dissociative recombination and radiative recombination are

Table 10-6. Variations in the physical processes used in the theoretical models

Nisbet Stubbe Strobel and McElroy Oran and Young

Processes [1971] [ 1970] [1970] [1977]

In the Altitude Region (km) 120–1 250 120–1 500 200-700 120-1200
Neutral Constituents for Nz, 02, 0 N2, 02, 0, He, N2, 02, 0, He N2, 02, 0, He, H
Ionization H
Chemical Reactions 5 Reactions 10 Reactions 4 Reactions 24 Reactions
(Charge-Exchange)
Ionized Constituents 0+, NO+, 02+ O+, NO+, O 2?+ 0+ 0+, NO+, 0+29
H+ H+
N+, He+, N2+
Neutral Winds Horizontal Winds Horizontal Winds Horizontal Winds
Electric Fields Yes — —
Magnetic Fields — Yes — —
Additional Features — — Solar Flare Effects

10-44
 a
IONOSPHERIC RADIO WAVE PROPAGATION
responsible for the decay of the ionosphere. For his simple The additional limitations of these models are
model, Nisbet neglected the transport processes such as 1. All the models are poor in predicting the high
neutral winds, electric fields, and magnetic fields. The pro- latitude ionosphere.
cedure is further complicated because coupled simultaneous 2. None of the models take into account the effects
equations must be solved for neutral winds, mass transport, of particle precipitation in the auroral region which
and energy transport. For determining electron density in enhance the E(Es) and F layers.
the ionosphere, the gas consisting of both ions a~d electrons, 3. The mid-latitude trough which exhibits large hor-
is considered electrically neutral. Thus, in every elementary izontal gradients in electron density is not incor-
volume, the number of electrons is equal to the number of porated in these models.
ions. All the models reproduce many of the observed fea- 4. These models are good for latitudes f 20° tot 60°,
tures such as the diurnal variation, seasonal variation, and and are poor predictors for the equatorial region
solar cycle dependence of the midlatitude ionosphere under and the high latitude region.
quiet conditions. The accuracy of the theoretical models Nonetheless these models serve two useful functions:
depends upon the understanding of the physical processes (1) to predict ionospheric parameters, and (2) to determine
considered in the models. For accurate predictions from the physical phenomena and/or to modify existing coefficients
theoretical models, precise information on the large number for explaining the deviations between the experimentally
of variables used in the models is necessary. Also, the observed value and the predictions from these models.
models use several observed average boundary conditions The computer programs for the ITS-78 [Barghausen et
which could have a large variability dependent on other al., 1969] and the IONCAP program [Lloyd et al., 1978]
geophysical parameters such as solar activity and magnetic are available from the Institute for Telecommunications Sci-
activity. The results from the models are adequate for long ences, Boulder, Colorado 80303. The computer programs
term planning of science and engineering applications. Though for the Bent model [Llewellyn and Bent, 1973] are available
these models reproduce main observed average features of from the Atlantic Science Corporation, P.O. Box 3201,
the ionosphere, they are unable to specify the ionosphere Indialantic, Florida 32903. The computer programs for IRI
within an accuracy of *20% needed by the systems in 79 are available from World Data Center A, Boulder, Col-
operation. At present, the main input information of solar orado 80303.
EUV radiation needed for the theoretical models is not rou-
tinely available for predicting the ionosphere.
10.4 HIGH FREQUENCY RADIO
PROPAGATION

10.3.3 Comparison of the Phenomenological The high frequency (HF) band of the electromagnetic
Models, Their Limitations and Ability spectrum extends from 3 to 30 MHz, corresponding to a
wavelength range of 100 to 10 m. Many services have
In comparing the models one must note that IONCAP frequencies allocated in this band—Local/International
is the modified and more flexible version which replaces Broadcast, Amateurs, Standard Frequencies, Maritime and
the ITS-78 model. As the ITS-78 /IONCAP and the Bent Land Mobile, Point-to-point Communications, Industrial,
models serve entirely different purposes, it is essential to Scientific, Medical Diathermy, Aero Fixed, Citizens’ Band,
understand the difference in their approaches and final out- and so on. The band is also used for ionospheric sounding
put parameters computed by the models. These are sum- and over-the-horizon surveillance. Its use in most applica-
marized in Table 10-7. The left-hand column in Table 10- tions depends on the fact that HF waves are reflected by the
7 lists the parameter under consideration. The next four ionosphere.
columns summarize the features in each of the models, ITS- HF is used for broadcasting because of its greater area
78, IONCAP, the Bent, and the IRI 79 models, respectively. coverage relative to the bands on either side, which are
From the table it is seen that the selection of a model will restricted to either ground wave or line-of-sight propagation.
depend more upon the information sought under the param- Its use for communications stems mainly from the fact that
eter headings, than on accuracy. The IONCAP model is it is often the only means of communication. It is also very
basically useful for wave propagation using operating fre- often the simplest and least expensive form of communi-
quencies which would be reflected by the E, Es, F1, and cation.
F2 layers. On the other hand, the Bent model relies on the With the advent of satellite communications, which use
transmission, refraction, and absorption characteristics of signals of such high frequency that the normal ionosphere
the ionosphere, with the operating frequency much larger has little effect on them, and improvements to submarine
than the foF2 frequency. The IRI 79 model basically pro- cables, the proportion of traffic that goes by HF is signifi-
vides a distribution of electron density with altitude. All the cantly smaller than it used to be. However, the total use of
models predict quiet ionospheric conditions on]y. me models HF radio circuits is actually greater now than ever before,
do not hold for disturbed ionospheric conditions. and a substantial research effort is still being devoted to

10-45
CHAPTER 10
Table 10-7. Intercomparison of the empirical-computer based ionospheric models.

Parameter Ionospheric Models

ITS-78 IONCAP Bent IRI 79

D Region Non-deviative and Same as ITS-78 Not modeled Modeled

deviative absorp- + E Layer
tion only exponential
extension down
to 70 km

E Region
foE Modeled by hftin Same as ITS-78 Not modeled Modeled
et al. [1968] + exponential Kouris and
h.E 110 km fixed down to 70 km Muggleton 1973a,b]
Y.E 20 km parabolic Leftin [1976]
shape coefficients

F1 Region Not modeled Not modeled M~eled

foFl Rosich & Jones Eyfrig [1955]
coefficients [1973] Ducharrne et
al. [1973]

h~Fi
Y~Fl h~F1/Y~Fl = 4
(fixed)

F2 Region Hayden-Lucas Same as ITS-78 Bi-parabolic Modeled

Bottomside coefficients [1968] Rawer [198 1]
FoF2 Shirnazaki eq [1955] Bent co-
1~F2 + E layer efficients
Y.F2 retardation
Kelso [1964]

F2 Region TopsicJe Not modeled Not modeled Up to 1000 km Modeled

Rawer [ 1981]

E-F Transition Not modeled Modeled Not modeled Maeda [1971]

Region

Electron-Density Not computed Available up Available up Available up

Profile to h~F2 to 3500 km to 1000 km

Electron, Ion Not modeled Not modeled Not modeled Modeled

Temperatures

Ion Composition Not modeled Not modeled Not modeled Modeled

Total Electron Not computed Not computed Computed Not computed

Content (TEC)

MUF For short path Also for long Not modeled Not modeled
only path (> 10,000 km)

Short-Term Predic- Function of Kp Not modeled Not modeled Not modeled

tion of MUF

Input Parameters Sunspot number Sunspot number Sunspot number Sunspot number
required and 10.7 cm
solar flux

10-46
 IONOSPHERIC RADIO WAVE PROPAGATION
Table 10-7. (Conrinue@

Parameter Ionospheric Models

Noise Parameters Galactic Same Not modeled Not modeled

Atmospheric Modified
Manmade Same
MUF 5090 Modeled Modeled Not modeled Not modeled
FOT 90%
HPF 10%

System Performance Modeled for Also has a Not modeled Not modeled
short path lottg path
(<3000 km) option >
10,000 km

Antenna Patterns Uses ITSA- 1 Modified Not modeled Not modeled

Package with ITS-78
10 antenna package with
options 17 antenna
options

Sporadic E Modeled in terms Same as ITS-78 Not modeled Not modeled

of occu;ence
f;equency

Circuit Reliability Modeled Modeled as Not modeled Not modeled

Service tibability ITS-78
Multipath Evaluation

improving our knowledge of the ionosphere and HF prop- Lecture Series No. 127 on “Modem HF Communications”
agation. is also a valuable source of information [AGARD, 1983].
Some of the difficulties associated with using HF for It is the intention of this section to provide a broad
communications, broadcasting, or surveillance stem from overview of HF propagation, its relationship to the iono-
the ionosphere itself and success in any of these fields de- sphere, its problems, and to indicate those areas of current
mands a good knowledge of the ionosphere and its vagaries. interest to users of HF. A basic knowledge of the ionosphere
Irreducible difficulties associated with HF propagation can itself is assumed (see Chapter 9 of the present volume).
usually be traced to characteristics of the ionosphere or of Emphasis will be placed on the use of HF for communi-
radio waves propagating through any iightly ionized me- cations. The same concepts and problems also apply to the
dium. Thus it is essential for the professional user of HF use of HF for broadcasting and surveillance (over-the-ho-
to have a good knowledge of both the ionosphere and radio rizon radar). Section 10.6 covers the effects of over-the-
wave propagation. horizon radars in the HF band.
Much has already been written about the ionosphere and
radio wave propagation and the reader should look else-
where for details. See, for example, the books by Davies
[1966, 1969], David and Voge [ 1969], Rishbeth and Garriot 10.4.1 Morphology of the Ionosphere
[1969], Hargreaves [1979], Ratcliffe [1970], Lied [1967],
and Picquenard [1974]. The four volume report “Solar Ter- An understanding of the morphology of the ionosphere
restrial Predictions Proceedings” [Donnelly, 1979, 1980] is is an essential prerequisite for its successful use as a com-
an excelIent supplement to these books, providing more munications medium. The basic theory of the ionosphere
recent reports on the general problem of forecasting the and its variations has been outlined in Chapter 9 of this
solar-terrestrial environment. The reports of Study Group 6 volume-here we are concerned mainly with how the iono-
of the Consultative Committee for International Radio (CCIR) sphere varies, rather than why it does.
are also a very useful source of information, and are par- There are five main variations of the electron density of
ticularly valuable because they are regularly updated. AGARD the ionosphere that must be taken into account:

10-47
CHAPTER 10 SOLAR MAXIMUM

SUMMER 1958 WINTER 1958

F2 \
. T= 188 .

. .
F2

n
L

MNRISE SUNSE” SUNRISE SUNSET

I 1 II 150”E o I
06 12 18 : I ) 06 12 18 : 4
HOUR (LT) HOUR (LT)

SOLAR MAXIMUM

WINTER 1964

T=O

SUNRIS SUNSET
A
SUNRIS
I I II
SUNSET
150°Eo I I

“O 06 12 18 24 0 06 12 18 24
Figure 10-42. Diurnal variations of the critical frequencies of the E, F,, and F? layers for solar maximum (1958) and SOIW minimum (19@) and for
summer (January) and winter (June ), at a typical mid latitude station (Canbema). The parameter T is an ionospheric index related to the
level of solar activity.

10-48
 a

IONOSPHERIC RADIO WAVE PROPAGATION

60° E 120” E I 80” 120” W 60” W )“
90\

60 -

~,@ 20030°40”50”60”
Ioo090°80”70”6V 5004V Boe 90 100.
30
●
c) b o“
20”
o
?~u) 30”
40” ;0: !jtJN

30 - 50” SOLAR
60” X =ZENITH
ANGLE
70”&[
60 - 80”
90
1 I ~.
04 08 12 16 20
90-

0“ 60” E 120” E 180° 120”W 60” W

90 ‘

60 -

30 ~-o”

30

I
\\l,5-<.u —
60
t ‘.o~
I 1 I 1 I

90 04 08 12 16 20

0“ 60*E 120”E 100” 120”W 60”W

90

60 2.02.5 3.0
1.5
,57 30 2“5-

30 3.5
/
<~z’., ,4.0

0
G

30 ?~~ 3.5

3.0
60 - 2,5
2.0
~1.5
1 1 1 1 i
90 04 08 12 16 20

Figure 10-43. The geographical variation of the critical frequency of the E layer as a function of local time for June at solar minimum (center .tranel). and
solar maximum (botton panel). The top panel shows the variation of the solar zenith angle for the same month.

10-49
CHAPTER 10
1. DIURNAL-variation throughout the day, which is fs FI -JUNE, 1954, SOLAR MINIMUM
largely due to the variation of the solar zenith angle.
2. SEASONAL-throughout the year. a“~
3. LOCATION-both geographic and geomagnetic.
4. SOLAR ACTIVITY-both long term and disturbances
5. HEIGHT-the different layers.
These variations have all been deduced experimentally,
by world-wide observations of the ionosphere over the past
few decades. The reader may refer to Davies [1966, Chapter
3] and Hargreaves [ 1979, Chapter 5] for details. The diurnal,
seasonal, solar cycle and height variations of the ionosphere
Y/ lGROUND
may all be deduced by routine monitoring of the ionosphere # / DAY- NIGHT
at one location. Figure 10-42 shows these four variations ~ Z LINE
for a typical midlatitude station. 1958 was a period of high
8@
solar activity, as indicated by the high values of the ionos- I 1 I 1 I I I 1 I 1
pheric index, T (see Section 10.4.4). The figure also illus- 00 04 08 12 16 20 24
trates the mid-latitude seasonal anomaly, the name given to
the initially unexpected fact that foF2 is higher in the winter
than in the summer, in spite of the larger solar zenith angle.
This anomaly and others are described by Hargreaves [ 1979, fo FI - JUNE, 1958, SoLAR MAXIMUM
Chapter 5].
8 O“ -
Once the diurnal, seasonal, solar cycle and height vari- :.5
ations of the ionosphere at a given location have been de- I 60 \
5.5
/-
t- -\ ‘)/’ 0
duced, the next step is to measure and understand the vari- 4.0/
% 40 \
ation with location. This has been achieved through an z 4.5/
6.0
; \ r~ )$
international effort of observations and data exchange, and g 20
we now have reliable maps of the world-wide distribution \ 1 ~ /’
z
of the important ionospheric parameters. The accuracy of o
G I
a
these maps over the ocean areas, where no observations are -J20 - \ /
6.87 &6’4
available, still remains somewhat limited [Rush et al., 1983]. \ /
~40
The easiest part of the ionosphere to model on a world- o \ \ GROUND
wide basis is the E layer. Figure 10-43 shows the variation \ / DAY- NIGHT
w 60 \\ -/0 LINE
of foE, the critical frequency for the E layer, for June at
solar minimum and solar maximum. The figure also shows ~
the variation of the solar zenith angle for the month, and it 00 04 24
can be seen that the variation of foE follows closely that of
LOCAL TIME
the solar zenith angle. In fact, the variations are so close, .
indicating that foE is very largely solar controlled, that it
Figure 10-44. The geographical variation of the critical frequency of the
is possible to use a simple empirical representation to deduce F1 layer as a function of local time for June at solar min-
foE for a given zenith angle, x [for example, Hargreaves imum and maximum.
1979; Muggleton, 1971]:
10-44), except that the F 1 layer tends to disappear in winter.
Hargreaves [1979] gives the following formula for foFl:
foE = 3.3 [(1 + 0.008R) COS X] ’(4MHz .

foFl = 4.25 [(1 + 0.0015R) COS x]’” MHz .

Note that foE also varies linearly with sunspot number, R,
increasing by about 20Yc over a typical slow cycle for a The variations of foE and foFl with (COSx)’” identify both
zenith angle of zero. the E and F layers as well behaved Chapman layers. Other,
Sporadic E (Es) layers also occur in the height range of more accurate, world-wide representations of foF 1 have
the normal E region. These layers are patchy and only a been given by Rosich and Jones [1973] and Ducharme et
few kilometers thick at mid latitudes. They tend to appear al. [1971].
and disappear almost at random (hence the name), but have Moving on to the F2 layer, due to its large height and
well-defined gross seasonal and latitudinal variations. See, electron density the most important layer as far as HF prop-
for example, Hargreaves [ 1979, p.90]. agation is concerned, we find that the simple situation that
The F 1 layer is similar to the E layer (see Figure holds for the E and F1 layers does not hold very well for

10-50
 IONOSPHERIC RADIO WAVE PROPAGATION
N 90
IN tric and magnetic fields. Theoretical modeling studies of the
equatorial ionosphere have been performed by Anderson
[198 1], among others. Empirical maps of foF2 and other
ionospheric parameters have been published by CCIR [ 1966].
The morphology of the high latitude ionosphere is even
more complicated than that of the equatorial ionosphere and
much remains unknown about it. Probably the most im-
portant feature of the high latitude ionosphere is the mid-
latitude ionosphere trough, which lies equatorwards of the
~ 50 -
/
auroral oval. The trough is a narrow feature that moves in
70 -
step with geomagnetic activity and thus fails to appear in
1 1 1
S90
0. 30°E 60-E 90aE 12WE 150-E IW 150-W 120W 90-W SOOW 30-W @
monthly median maps of foF2. However, it can have very
GEOGRAPHIC LONGITUDE serious effects on HF communications at high latitudes be-
cause of the strong horizontal gradients associated with it.
The morphology of the high latitude ionosphere has been
reviewed, for example, by Hunsucker [ 1979], Hunsucker
et al. [1979] and CCIR [1981a].
Lastly, we must consider the D region. This region is
of no direct concern for HF radio propagation since the
electron densities are always too low to reflect HF waves.
However, the D region is very important from the point of
view of absorption of the energy of an HF wave, especial] y
at the lower end of the HF band. A review of the D region
and the prediction of its effects on radio propagation has
been presented by Thrane [ 1979]. Synoptic models of the
D region electron density are unreliable because of the com-
plexities of the D region and the difficulties encountered in
measurement of the electron density profi Ies.
Absorption of HF waves occurs mainly in the D and
Figure 10-45. The geographical variation of the critical frequency of the
F2 layer for June at solar minimum and maximum, for 00 lower E regions of the ionosphere. The free electrons absorb
UT. World maps such as these are made for each hour of energy from the incident wave and reradiate it in a contin-
each month (576 maps).
uous process. However, if an energetic electron collides
with a neutral particle before it can reradiate its energy, this
the F2 layer. Figure 10-45 shows, for example, how the F2 energy will be taken up by the neutral particle as kinetic
layer critical frequency, foF2, varies over the earth at 00 energy and will be lost to the HF wave, that is, energy will
UT in June, for low and high solar activities. It can be seen be absorbed by the medium.
that the simple structure obtained for the E and F1 layers, This type of absorption is known as non-deviative ab-
with the contours of foE and foFl closely following the sorption and is roughly proportional to 1/~ where f is the
contours of the solar zenith angle, no longer applies although wave frequency. Extra absorption, known as deviative ab-
a clear zenith angle dependence can be seen around sunrise sorption, also occurs near the reflection level. The non-
(-90 to - 12WE). In fact, the departures of foF2 from a deviative term usually dominates for oblique propagation.
simple R, cos “4X dependence are so great that it is necessary If the operational frequency, f, becomes too low, the ab-
to make world-wide observations to determine the actual sorption will increase to the point where the signal disap-
variations of foF2. Detailed studies of foF2 have shown that pears below the level of the noise at the receiver site and
as well as depending on R and cos x, foF2 also depends becomes unusable. This frequency is known as the lowest
on other factors such as electric fields, and neutral winds usable frequency, or LUF. A good treatment of absorption
to name a few, and its large scale morphology is controlled is given by Davies [ 1969, Chapter 6] while the variations
by the geomagnetic field. of absorption are discussed by Davies [1966, Chapter 3].
foF2 is also found to have variations with latitude which
are not seen in foE and foF 1. For example, Figure 10-45
shows that foF2 exhibits two afternoon peaks (-12 MHz 10.4.2 Simple Ray Propagation
in the solar maximum portion of the figure) situated on either
side of the equator. This feature is known as the equatorial Many of the operational aspects of HF propagation may
anomaly and is due to electrodynamics lifting of the layer at be studied using simple ray concepts. Figure 10-46 illus-
the equator under the combined influence of horizontal elec- trates the basic geometry of a one-hop HF circuit. Note that

10-51
CHAPTER 10
the diagram ignores the ground wave which is usable for
ranges up to about 50– 100 km, depending on the frequency,
antenna, ground conductivity, etc. The reflected ray is con-
tinuously refracted as it passes through the ionosphere and
if sufficient refraction occurs the ray will be bent down
sufficiently to reach the receiver. Figure 10-46 also shows
one of the most basic formulas of ray propagation:

MUF = f. x obliquity factor = fc X k X sec 4,

MUF= fck sec @

that is, the maximum usable frequency (MUF) is equal to the
product of the critical frequency, f=, of the reflecting layer
and an obliquity factor related to the geometry of the cir-
Figure 10-46. Simple geometry of an HF oblique circuit, illustrating the cuit. For a flat earth/ionosphere approximation, this factor is
essential elements of the circuit, The formula for the max-
sec +, where @ is the angle of incidence. For a curved earth
.
imum usable freauencv . (MUF) is afl that is required in
many calculations for HF propagation. and ionosphere, the factor k is introduced to allow for the
different geometry. This factor is typically of the order 1.1.
In practice, the obliquity factor for a given circuit relying
on reflection from the F2 layer for example, is obtained from

E-MOD ES F-MODES
9.0 90
I 111111/ I

80 a .0 80 . . 8.0

70 7,0 70 . – 7.0

ELEVATION ANGLE
’60 6.0 60 . – 6.0

/
/
h7 I.*D -z
50 ELEVATION ANGLE 2-HOP ;’””j5.0 : 50 -
t J
a
——

‘x]
40 4,0 ; 40 -
\ I z
,2-HOP l-HOP
I - HOP
30 3.0 30 .-
I
I
2.0 20 . I
K*sec (PHI)

10

0 I I 1111111I I I Ill Ill I I

OS o
~ ~~
PATH LENGTH (km) PATH LENGTH (km )

Figure 10-47. Plots showing the elevation angle and obliquity factor (k sec @) for propagation via the E and F layers. For a given circuit length and
number of hoDs. these olots show the reauired elevation angle (necessary for selecting an appropriate antenna) and obliquity fdctor ft)r the
two layers (u;eful, am;ng other things, ;or consideration o; possible E layer screening)

10-52
 1[IONOSPHERIC RADIO WAVE PROPAGATION

t
T
MAXIMUM_EL~TQN ~NSITY IONOSPHERE

--
/

Figure 10-49. Sample ray path for a fixed frequency but varying elevation
angle. If the ooeratin~ freauencv is above the local critical
fre~uency, higlr eleva;ion ;ays ;ill penetrate and there will
be a “skip zone” around the transmitter which cannot be
reached by an ionospheric all y propagated ray.

Figure 10-48. Sample ray paths for a fixed distance but different fre-
quencies, As the frequency increases, the ray must pene- The area surrounding the transmitter, which is defined by
trate further into the ionosphere. If the frequency is too
k sec 4< f/foF2, is known as the skip zone for that fre-
high, that is, abve the MUF for the circuit, the ray perp-
etrates, quency (and location and time). Signals at the frequency f
cannot penetrate into the skip zone, although the ground
wave would propagate out to about 50–80 km. This phe-
values of the obliquity factor for a 3000 km hop, M(3000)F2, nomenon can also be used to advantage by ensuring that an
which is scaled routinely from vertical incidence ionograms unwanted receiver lies in the skip zone of the transmitter.
[Piggott and Rawer, 1972]. The obliquity factor for a dis- When the value of foF2 above the transmitter exceeds the
tance D is related in an empirical fashion to M(3000)F2. operating frequency there is, of course, no skip zone.
Figure 10-47 shows how the value of k sec + varies with HF-communication via ground wave is important in
circuit length for one and two hop E and F layer modes. many areas, particularly over sea and flat land with high con-
Typical values for the E and F one-hop modes are -5 and ductivities, where reliable circuits may be established up
-3. Figure 10-47 also shows the elevation angles corre- to distances of several hundred kilometers. The conductivity
sponding to the different propagation modes. of the surface is strongly frequency dependent with rapid
Figure 10-48 illustrates ray propagation for different attenuation at the higher frequencies. In the past CCIR has
frequencies on a fixed circuit. As the frequency increases, published a set of curves of ground wave field strength
the ray must penetrate further into the ionosphere before it versus distance. CCIR is in the course of implementing a
is refracted to the horizontal and thence back to the ground. computer program to estimate ground wave field strengths.
The highest frequency that can be reflected back to the Ground wave propagation may be quite complex, particu-
ground is the MUF for that circuit. Note that the ray cor- larly over rough terrain and over mixed land-sea paths.
responding to the MUF does not reach the altitude of the There is a need for better charts of ground conductivity, and
peak density of the layer, h~.,. It is only for vertical inci- in some cases terrain modeling may be useful and impor-
dence that the ray actually reaches h~,X. tant. Large topographical features such as mountain ranges
Figure 10-49 illustrates the concept of the skip zone, and glaciers may cause reflections and strong attenuation,
which is a zone into which an ionospherically reflected and vegetation, soil humidity and snow cover also influ-
signal cannot propagate. The figure illustrates the effect of ence the propagation characteristics.
different elevation angles for a fixed frequency. As the el- We have seen that the MUF for a given circuit is set by
evation angle increases, corresponding to a shorter circuit the density of the ionosphere at the reflection point and the
length, the ray must penetrate deeper into the ionosphere in geometry of the circuit. For a multi-hop circuit, the MUF
order to be reflected. However, as the elevation angle in- is set by the lowest of the MUFS for the individual hops.
creases the obliquity factor k sec + decreases. While the The lowest usable frequency (LUF), on the other hand, is
product foF2 . k sec + remains greater than the operating set by absorption of the signal by the ionosphere and by the
frequency f, the signal will be reflected. When f exceeds generally poor performance (low gain) of most HF antennas
foF2 . k sec +, the signal will penetrate the ionosphere. at low frequencies.

10-53
CHAPTER 10
10.4.3 Requirements for
Successful Communications I CHOICE OF MODE

Under normal operating conditions, there are three fac-

tors that must be considered to achieve successful com-
munications. These are
1. Choosing a suitable operating frequency.
2. Choosing a suitable antenna system.
3. Ensuring that the wanted signal is at a level above
that of the local radio noise at the operating frequency.
The choice of a suitable operating frequency is the sub-
ject of main concern in the present context since this is
where a knowledge of ionospheric physics and radio wave
propagation is required. This choice will be discussed in Figure 10-50. Sketch illustrating the fact that the antenna pattern should
Section 10.4.4. The choice of a suitable antenna system match the required propagation mode for a given circuit.

will require matching the antenna pattern to the propagation

angles of the HF signals—these can be deduced using the
ionospheric models developed for ch~sing a suitable op- the situation. Note that the radiation pattern of an antenna
erating frequency. However in many cases practical con- is a function of frequency so that an antenna appropriate
siderations will intervene and a far from ideal antenna will for a low HF frequency may have a very poor performance
be used. The ideal antennas tend to be very expensive and at a higher frequency. Selection of the correct antenna will
mobile operators especially will often be forced to use a not only ensure that the bulk of the transmitted power goes
random wire hung over a tree. As a general rule of thumb, at the required elevation angle, but can also be used to select
horizontal antennas are required for short circuits and ver- a partici.tlar propagation mode and thus avoid multipath in-
tical antennas are required for long circuits. The ubiquitous terference. Multipath interference arises when the trans-
whip antenna is absolutely the worst (but probably the most mitted signal arrives at the receiver over two or more sep-
common !) choice for short sky-wave circuits since its an- arate propagation paths with different time delays.
tenna pattern has a null in the vertical direction. Antenna Figure 10-50 illustrates propagation via lF and 2F prop-
patterns for the common antennas can be obtained from agation modes, with an antenna pattern whose lobe favors
Lloyd et al. [1978]. Table 10-8 gives a brief summary of the higher angle 2F mode and almost completely prohibits

Table 10-8. Suitable simole antennas for use on paths from O to mom than 3000 km.

Required Radiation
Path Length (km) (Elevation) Angles Suitable Simple Antenna

0-200 60°–900 Horizontal dipole: broadside

is required azimuth. 0.25
wavelength (h) above ground.

200-500 40°-700 Horizontal dipole: broadside

to required azimuth. 0.3 A
above ground.

500–1000 25°–500 0.25 A vertical monopole, or

and 10°–200 horizontal dipole, broadside to
required azimuth 0.5 A above
ground.

1000-2000 10°–300 Vertical monopole up to 0.3 k

and low angles long with ground screen.

2000-3000 5°–150 Vertical monopole up to 0.3 k

and 20°–300 long with ground screen.

>3000 low angles Vertical monopole up to 0.6 k

long with ground screen.

10-54
 a
IONOSPHERIC RADIO WAVE PROPAGATION
If an adequate signal-to-noise ratio cannot be obtained
I CHOICE OF ANTENNA
I by lowering the noise level, it is necessary to increase the
signal level. Some increase in signal level can be achieved
by choosing an antenna that has more gain at the given
operating frequency and elevation angle. This is one ap-
proved solution. An alternate solution is to increase the
transmitter power.

10.4.4 Predictions for HF Communications

The first step in predicting HF communication condi-

tions is to set up an appropriate model of the ionosphere.
Figure 10-51. Sketch illustrating the fact that the antenna should be chosen
to match the circuit length. The antenna pattern illustrated To have practical application, this model must include all
is not appropriate for short circuits (R - 300 km). five main variations of the electron density distribution of
the ionosphere (altitude, location, diurnal, seasonal, solar
propagation via the 1F mode. The impact of the choice of cycle), must include some measure of the remaining vari-
an antenna for a given circuit is illustrated in Figure 10-51 ability of the ionosphere after these main variations have
for hop lengths of 300 and 1000 km. In this case, the antenna been accounted for, must exist as a reasonably efficient and
heavily favors the longer circuit and is quite inadequate for fast computer code, and must possess some method for
short-haul circuits. projection forward in time. Ideally, it should also be capable
Given a frequency that the ionosphere will support and of a modelled response to short time scale events such as
an antenna which emits sufficient power in the direction shortwave fadeouts and ionospheric storms. (These are dis-
taken by the signal that arrives at the receiving site, the cussed in Section 10.4.6. )
third thing to ensure is that the signal strength is above the Most of the ionospheric models used for communication
strength of the local radio noise. This noise can be natural predictions are empirical models based on world-wide ob-
or manmade. Below about 20 MHz, natural noise is caused servations of the ionosphere over the past four decades.
by either distant thunderstorms, which cause a general in- Observations of the main parameters of the ionosphere, foE,
crease in background noise level, or local thunderstorms foFl, foF2, M(3000)F2, and Es have been used to produce
which are usually much more obvious causes of poor signal- worldwide contour maps of monthly median values of these
to-noise ratio. Galactic cosmic noise becomes the domi- parameters for each hour of each month and for low and
nating natural noise above -20 MHz when it penetrates the high levels of solar activity. An example of such maps is
ionosphere from ahve at frequencies above foF2. The world- given in Figure 10-45. To calculate the maximum usable
wide distribution and characteristics of atmospheric radio frequency (MUF) for a given circuit, hour and month, the
noise can be obtained from the CCIR Report 332 [CCIR, values of foF2 and M(3000)F2, for example, are determined
1963]. Manmade noise includes such things as industrial by interpolation in the appropriate world map for the ex-
noise due to welders, diathermy machines, car ignition, pected point(s) of reflection of the signal. For a level of
power lines and so on. Interference from other communi- solar activity other than the low or high levels for which
cators using the same operating frequency can also be re- the maps are drawn, values of the ionospheric parameters
garded as noise. are obtained by interpolating in each of the low and high
There are many techniques available to ensure an ade- activity level maps and then interpolating again in these
quate signal-to-noise ratio. The more environmentally ac- results to find the values appropriate to the given level of
ceptable method is to aim at lowering the noise level. This activity.
can be achieved by choosing as a receiver site some location In practice, the only prediction actually performed in
that is remote from the major sources of manmade noise, making predictions of suitable HF operating frequencies is
which usually entails being away from major cities. Some that of the general level of solar activity that can be expected
noise rejection is also possible using horizontally polarized to pertain at the epoch for which the predictions have been
antennas which de-select the local noise that tends to be made. This prediction is usually made in the form of some
vertically polarized. Careful attention to the azimuth of the ionospheric or solar index, which is a single parameter de-
main lobe of the receiving antenna can also result in the scribing the gross behavior of the ionosphere, and which is
beam not being aimed at a nearby source of noise. In heavily used to drive the computer program. The most common
urbanized areas (Europe, East and West Coast of the United index used for prediction programs such as the U.S. pro-
States), the noise level is often set by the large number of grams IONCAP and ITS-78, and by the CCIR, is the average
other communicators. sunspot number, R, z. However, Wilkinson [1982] has shown

10-55
CHAPTER 10
A PU)T FOR BRISBANE JUNE OOUT an appropriate signal-to-noise level at a receiver site is a
much more complicated procedure than the calculation of
the MUF. The required transmitter power is of no small
matter since the cost of the transmitter increases dramatically
with the power. The noise level may be either measured at
‘oo~
the site or estimated from empirical data bases such as
provided by CCIR. See for example Davies [1966, Chapter
7] and CCIR [1963].
There are two other main concerns here, the antenna
and the transmitter, but let us assume that an appropriate
antenna with a known gain has already been selected. To
calculate the required transmitter power, we need to cal-
culate the losses. These include deviative and non-deviative
absorption losses, basic free-space transmission loss, mul-
tihop ground reflection loss, polarization coupling loss, spo-
radic E obscuration loss and horizon foeus gain (a negative
loss). The largest of them is the free-space loss, and this
increases with the length of the propagation path.
61 Because the circuit losses depend on the propagation
path, all possible paths must be considered and compared.
%/
73
Reliable determination of the actual propagation paths re-
quires a reliable model of how the electron density varies
,52 years
with altitude up to the peak of the F2 layer. Suitable models
of this N(h) variation have been developed by Bradley and
Dudeney [ 1973], among others. As well as being an accurate
o [ I I [ I
o 2 4 54Y56 8 10 12 14 representation of the real N(h) distribution, such a model
must also be computationally simple because it is used many
FREQUENCY (Mliz) times in the determination of the virtual reflection heights,
h’, and thence the possible propagation modes. The Inter-
Figure 10-52 The variation of the monthly median value of foF2 for national Reference Ionosphere [Lincoln and Conkright, 1981]
June, 00 UT, at Brisbane, Australia, as a function of the
ionospheric index, T. Each data point represents the value may also be used to deduce an N(h) profile, but this model
for June for years 1943-1980. has not been designed for the speed that is required in routine
field strength calculations.
The ionospheric models currently available for use in
that any index based on the ionosphere itself, such asIF2 studies of HF propagation have been reviewed by Dan-
[Minnis and Bazzard, 1960] is usually preferable to a purely dekar [ 1982] and Goodman [ 1982] (see also Section 10.3).
solar index such as R,z. Figure 10-52 illustrates the rela- Full descriptions of the techniques used to calculate field
tionship between the ionospheric index, T [Turner, 1968; strengths at a receiving site are given by Davies [ 1966,
Turner and Wilkinson, 1979] and the monthly median values Chapter 7], Lloyd et al. [1978] and CCIR [1970, 1978].
of foF2 at Brisbane, Australia for 00 UT in June (10 LT). A straight forward approach to obtaining MUF’S for var-
Each data point represents data for June in the years 1943 ious modes and their respective path losses which also
to 1980. The low dispersion of the data points about the permits the inclusion of realistic antenna patterns for
regression line indicates the usefulness of the T index for transmit and receive sites is available in the 10NCAP
describing ionospheric conditions. The selection of the cor- program [Lloyd et al., 1978]. The model computations
rect index for some future epoch will become more uncertain are based on a three layer (F2, F 1, E) representation of
as the lead time increases because of our general inability the electron density profile with ITS78 coefficients being
to predict the detailed behavior of the sun. used to determine the (RZ-dependent) foF2 value. The ion-
Predictions of the main critical frequencies and heights ospheric parameters are evaluated at the reflection points
of the ionosphere, together with some simple geometry (Sec- and two dimensional ray tracing is applied, assuming lo-
tion 10.4. 2), are adequate for the calculation of the appro- cal horizontal stratification. For the most reliable results,
priate frequencies to use for communication on a given simplistic approaches to propagation mode determination
circuit. It is also possible to determine elevation angles of must be replaced by ray tracing techniques such as de-
the possible propagation modes (and thence choose a good scribed by Davies [ 1969, Chapter 7]. These techniques
antenna) and to determine the conditions under which some are essential when large horizontal gradients exist in the
adverse propagation conditions will exist. ionosphere, but are rarely used for routine calculations
The calculation of the transmitter power required to yield because they are very time consuming.

10-56
 a
IONOSPHERIC RADIO WAVE PROPAGATION
a circuit by choosing a frequency that is supported by only
one propagation mode (for example, use a frequency above
\ the 2F MUF but below the 1F MUF) or by choosing an
antenna that heavily favors one propagation mode. Diversity
techniques, in which the same information’ is sent in two or
more ways and recombined at the receiver, can also be used
to overcome the effects of fading. Example are space (spaced
antennas), frequency, angle of arrival, and polarization di-
:-
“\. / XX versit y.
\,/.\./’ x *X x
x
xx
xx x 10.4.5.2 Sporadic E. Sporadic E is a two-faced phenom-
. xx %1
enon having both advantages and disadvantages. There is
xx-
no doubt that when a strong sporadic E layer is present, it

I
____ UPPER DECILE
x
x — MEDIAN
—.–._ LOWER DECILE presents an ideal propagation mode suitable even for me-
2
xxxxxx MEASURED VALUE
FOR A MAGNETICALLY
dium speed digital data transfer. However, a dense Es layer
DISTURBED PERIOD can screen the Flayer, preventing signals from ever reaching
that layer and causing Es-F combination propagation modes
that may have none of the desired characteristics-the sig-
150* EMT
nals may miss the target or may arrive at an angle not
specified in the original receiving antenna design. Many
Figure 10-53. Plot of theobserved valuesof foF2amid1atitude station, prediction schemes for HF propagation include statistical
December 1980, asafunction oflocal time. At each hour
there are 31 data points. 8070 of observations lie within occurrence rates for Es propagation modes, but the Es layer
the dashed upper and Iowerdecile curves. remains essentially unpredictable.

10.4.5.3 Problems in High Latitudes. After major solar

10.4.5 Problems with HF Communications flares with particular characteristics, high energy protons
penetrate into the lower ionosphere at the poles and causes
The use of the ionosphere for communication, broad- wide-spread and long-lasting disruptions to HF communi-
casting, and surveillance is fraught with difficulties, some cations. The polar cap absorption event is discussed briefly
of which have already been mentioned. Predictions of the in Section 10.4.6. The penetration of electrons with energies
MUF, LUF, and field strengths rely on median models of in excess of 10 keV into the D region leads to increased
the ionosphere and can therefore specify only mean prop- absorption in the auroral zone. This auroral absorption may
agation conditions. A knowledge of the spread of values have severe consequences for circuits crossing the auroral
about the median is required for successful communication zone, but strong absorption is usually limited geographical y
for more than 50% of the time, and such statistics are in to patches a few hundred kilometers in extent and the du-
fact usually provided by the better prediction programs. For ration is typically half an hour to a few hours.
quiet, undisturbed conditions, this spread about the median An important feature of the undisturbed high latitude
is typically 15%–2070. See for example Figure 10-53. ionosphere is the great variabilityy in space and time. This
Even when conditions are undisturbed and the com- variability severely limits the usefulness of a median model
municator has chosen an appropriate operating frequency, of the ionosphere. The F region trough, which marks a
antenna, and transmitter power, propagation problems are transition between the midlatitude and high-latitude iono-
still encountered. Some of these are associated with events spheres, can have severe and detrimental effects on signals
on the sun and will be discussed in Section 10.4.6. Here propagating through it. The sharp gradients at the walls can
we wish to consider some problems that can occur even cause reflections and result in off-great-circle propagation.
when the ionosphere is being cooperative. Propagation in the auroral region may introduce rapid
and severe fading of HF signals. Diversity techniques or
10.4.5.1 Multipath Propagation. We have already seen some sort of real time channel evaluation technique (see
in Section 10.4.3 that propagation will normally occur by Section 10.4.9) therefore become almost mandatory.
several paths, for example, the 1F and 2F modes (see Figure The polar cap ionosphere in darkness, of importance to
10-50). The received signal will be the vector sum of all long distance HF communication between higher latitude
waves arriving at the receiver. If the different signal paths stations such as in Canada and northern Europe, is essen-
change with time in different ways, deep and rapid fading tially unpredictable for HF purposes. Enhanced F ionization
may occur, sometimes causing the signal level to drop below regions resulting from particle precipitation are randomly
the local noise level. Different modulation techniques are distributed within a low density background ionization. Es-
affected in different ways by this multipath fading. pecially during slightly disttttid conditions polar cap plasma
Multipath propagation may sometimes be eliminated fim convection moves high density particles anti-sunward at

10-57
CHAPTER 10
high velocities (Buchau et al, 1983]. These changes will unusual complaints in the Indian Subcontinent [Lakshmi
result in rapid MUF and mode variations and will in general et al., 1980].
lead to poor channel performance. MUF variations of typ-
ically several MHz from hour to hour and day to day were 10.4.5.5 Spread F and Irregularities, Small scale ir-
observed on an arctic link [Petrie and Warren, 1968]. In regularities of ionization seem to exist at every level of the
summertime the F 1 mode tends to dominate the mid-mom- ionosphere superimposed on the background of ionization
ing MUF, and during winter/solar minimum conditions Es discussed in previous sections. They affect the propagation
modes over transpolar circuits (for example, Andoya, Nor- of radio waves and their characteristics may, therefore, be
way to College, Alaska) have a greater than 5090 occurrence determined by radio techniques (see for example, Rishbeth
[Hunsucker and Bates, 1969]. and Garriott, [1969], Section 6.4). When seen on iono-
grams, the presence of irregularities is described as “spread
10.4.5.4 Problems in Low Latitudes. HF communica- F“, the spreading being either in the range or frequency
tion problems at low latitudes due to steep spatial and tem- domain. The corresponding descriptions are “range spread-
poral gradients have been discussed by Lakshmi et al. [1980]. ing” and “frequency spreading”, and parameters describing
The very steep gradients in foF2 during sunrise hours give this spreading are scaled routinely from ionograms [Piggott
rise to several difficulties [Lakshmi et al., 1980]: and Rawer, 1972]. Many mo~hological studies of spread
1. HF link operators are expected to get their frequencies F, showing its diurnal, seasonal, solar cycle and latitudinal
cleared from the appropriate governmental authority well in variations, have been published (see for example Herman,
advance and it is usual practice to fix one frequency for the [1966] or Singleton, [1980], and references cited therein).
daytime and another for the nighttime. The use of the night An earlier review of spread F and some of the effects on
frequency during sunrise will require much more power than radio propagation has been given by Newman [1966]. Most
is normally permitted while the frequency allocated for the morphological studies of spread F have been concerned with
daytime will be higher than the MUF during the transition frequency spreading. The characteristics of range spreading
period. are, however, quite different and only a few studies of this
2. Point-to-point links normally use inexpensive tuned type of variation have been made (see for example Cole and
directional antennas, and frequent change of operational McNamara, [1974]). At equatorial regions, range spreading
frequency is deleterious from the point of view of antenna is often associated with the large scale electron density de-
efficiency. pletions known as bubbles or plumes [Basu et al., 1980c;
3. In the case of long distance circuits in the east-west Tsunoda, 1980; and Tsunoda et al., 1982].
direction involving multi-hop F region propagation, the The effects of irregularities on radio propagation are
problem of the sunrise period will extend to a large number most important on paths that cross the equator when prop-
of hours, because the different F region reflection points agation actually relies on the presence of the irregularities
will fall in the sunrise transition location at different periods. [Nielson and Crochet, 1974]. Most recent efforts on de-
The steep gradients associated with the equatorial anom- scribing the effects of irregularities on communications
aly cause problems with north-south circuits. For example, have concentrated on the effects of irregularities on trans-
if we consider the anomaly peak in the northern hemisphere ionospheric propagation, where the problem of scintillation
to be at 15”N geomagnetic latitude, if a north-south is encountered. This subject is covered in Section 10.7, this
HF circuit is operating such that the reflection point is on Chapter.
either of the sides of the peak, and if the frequency of the
link is very close to the MUF, a peculiar situation arises. 10.4.6 Disturbances to Normal
If the point of reflection is equatorward of this anomaly Communications
peak, the radio waves incident on the ionosphere for the
northern circuit will continuously encounter increasing lev- We have previously seen that HF communications are
els of electron density due both to the vertical gradient as subject to sudden and often large disturbances due to events
the radiowave penetrates higher into the ionosphere and the that originate on the sun. Solar flares can cause immediate
horizontal gradient as the wave progresses in the direction and complete absorption of HF radio waves by greatly en-
of increasing electron density. On the other hand for the hancing the absorption suffered as the waves pass through
same link in the return direction, the horizontal gradient is the D region. These events, called short-wave fadeouts (SWF),
reversed. Thus the real MUF values for the two opposite are due to increased ultraviolet and x-ray fluxes, and can
directions in the same circuit can vary by a large margin therefore occur on on] y the sunlit side of the earth. Major
depending on the angle of incidence and on the magnitude flares can also eject a stream of protons which can penetrate
of the horizontal gradient. In fact, rather frequently, espe- the ionosphere near the magnetic poles and give rise to
cially when the operating frequency is close to the MUF complete blackout of HF communications in polar regions.
(calculated ignoring horizontal gradients), only one way Such events are called polar cap absorption events (PCAS)
communication would be possible. This has been one of the and can last for days depending on the size of the flare and

10-58
 a
IONOSPHERIC RADIO WAVE PROPAGATION
how well it is connected magnetically to the earth. The through an international effort coordinated through the IUWDS
SWFS, onthe other hand, usually last from a few minutes (International Ursigram and World Days Service) and many
to a few hours at the greatest and are most severe at low national agencies issue forecasts of solar and geophysical
latitudes (see Section 10.5). conditions. This effort has been described, for example, by
Many large flares also eject a cloud of plasma, which several authors in Volume 1 of Donnelly’[ 1979], including
if geometrical and interplanetary magnetic field conditions Heckman [1979]. CCIR Report 727 [CCIR, 1981 c] gives a
are favorable, can intersect the earth’s magnetosphere and review of the subject and refers to many of the latest avail-
cause both geomagnetic and ionospheric storms. The geo- able papers.
magnetic storm, in which the earth’s magnetic field is usu- The USAF Air Weather Service (A WS), through its
ally depressed below its normal quiet day value, is a result operational centers of the Air Force Global Weather Central
of a strong enhancement of the ring current and of no in- (AFGWC), provides space environmental support to the
trinsic interest to the HF communicator, but is a useful and entire Department of Defense. The overall driving require-
readily available indicator of an accompanying disturbance ment is to minimize system effects caused by impulsive
to the ionosphere and thus to communicators. The iono- solar/geophysical activity and ionospheric variations. The
spheric storms are of concern, however, especially when techniques for geophysical forecasting used at AFGWC have
the effect of the storm is to decrease the MUF on a circuit been described by Thompson and Secan [1979]. A major
to well below the predicted levels. advance in observational equipment during the last decade
The physics of ionospheric storms is not yet completely was the deployment of the new solar observing network
understood, but it is well established that both electric fields (SOON) and radio solar telescope network (RSTN). With
and thermospheric winds (see Chapter 17) play a role. See the data from SOON, RSTN and x-ray data from satellites,
for example Hargreaves [ 1979, Chapter 11]. AFGWC can provide a real-time comprehensive analysis of
The effect of a storm at a given location can be either a flare and its effects on the space environment.
to increase the value of foF2 (a so-called positive phase or The action taken by an HF user to overcome the effects
enhancement) or to decrease the value of foF2 (a negative of these disturbances depends on the nature of the disturb-
phase or depression), or to do both, a long depression fol- ance. During a SWF, a move to higher operating frequencies
lowing short enhancement. What actually happens depends is appropriate since the absorption decreases with increasing
on such things as local time of onset of the storm, station frequency. During the negative phase of an ionospheric
latitude, and season. During winter, most storm effects tend storm, lower frequencies must be used. The effect of a PCA,
to be enhancements, whereas in summer and equinox on the other hand, is normally so severe that it becomes
depressions often follow short-lived enhancements. It is the necessary to reroute the traffic around the disturbed area.
depression that causes the major communications prob- This usually requires avoiding the whole high latitude ion-
lems—the enhancements often go unnoticed by communi- osphere.
cators. Major depressions last typically for a day and can
decrease the MUF by a factor of two. Storm effects are 10.4.7 Unusual Propagation Modes
much more marked at higher latitudes (see for example
Rishbeth and Garriott, [1969], Chapter 8). The usual monthly-median HF predictions normally as-
The lower regions of the ionosphere are not usually sume simple propagation modes such as the 1F and 2F
affected during ionospheric storms and the lower frequency modes and in general these predictions are quite successful.
limit for HF communication remains unchanged. Some prediction systems also include Es and such propa-
Ionospheric storms due to solar flares are a high solar gation modes as 1Es and Es-F. However, it is found in
activity phenomenon. However, ionospheric storms also oc- practice that other unusual propagation modes can also exist.
cur away from the peak of the solar cycle. These are at- In general these unusual modes have one feature in com-
tributed to enhanced solar wind streams emanating’ from mon—they rely for their support on some particular feature
magnetically open features in the corona known as coronal of the ionosphere that is restricted in latitude. We shall
holes, As the stream sweeps over the earth, the electric consider here a few examples of such modes.
currents flowing in the magnetosphere and ionosphere are Possibly the most useful propagation mode not normally
modified, yielding both geomagnetic and ionospheric storms. predicted is the F2 super-mode encountered on transequa-
These storms are called recurrent storms because they tend torial circuits during periods when the equatorial anomaly
to recur every 27 days (solar rotation period as seen from is well developed, that is, during the afternoon and early
the earth) as the solar wind stream passes over the earth evenings, during equinoxes, at high levels of solar activity.
again. The effects on the ionosphere are usually less marked This mode involves two F region reflections without an
than those of a flare-induced storm but can last longer (a intervening ground reflection and is characterized by high
few days) because of the time taken for the stream to pass signal strength, low fading rates and an MUF 10-15 MHz
over the earth. higher than for the normal (2F) mode. It is often described
The scientific community keeps a 24-h watch on the sun as the aftemoon-ty~ TEP (transequatorial propagation) mode.

10-59
CHAPTER 10
A second mode associated with the equatorial iono- ditions. For a survey of the historical development of fore-
sphere, and which can coexist with the afternoon-t ype TEP casting methods for HF propagation see Rawer [1975].
mode in the early evening, is called the evening-type TEP The effects on the ionosphere and communications of
mode. It is characterized by strong signals, rapid flutter solar disturbances which cause ionospheric, storms have been
fading and frequencies well above the normal 2F MUF. A discussed in Section 10.4.6. However daily values of foF2
propagation frequency of 102 MHz has been regularly ob- and MUF are known to vary by about 15% to 2070 from
served on a Japan-Australia circuit. The propagation mode the monthly median values during quiet times as well as
in this case is probably a ducted mode, the signals traveling during ionosphenc/magnetic storms. See, for example, Fig-
within the walls of equatorial “bubbles” [Heron and ure 10-53. These variations may be superimposed on slower
McNamara, 1979 and Winkler, 1981]. This mode is also upward or downward drifts in values over several days. It
most likely to occur during the equinoxes at solar maximum. is desirable to predict all of these variations for the purpose
A review of TEP has been published by Nielson and Crochet of efficient radio communications.
[ 1974]. There are several possible approaches to the short-term
Enhancements of the MUF on transequatorial circuits forecasting of ionospheric parameters.
have also been attributed to combination modes in which 1. Using associated geophysical parameters.
one hop is via a reflection from the equatorial Es belt which 2. Using the ionospheric parameters themselves.
stretches a few degrees either side of the magnetic equator 3. Using ionospheric indices.
[McNamara, 1974a]. In an Es-F combination mode, the The day-to-day variability of foE and foFl is so small that
MUF is enhanced because the F layer hop is longer than monthly median values of these parameters can be used to
the usual hop length on a 2F mode. Similar MUF enhance- represent the daily variation in the E and F1 regions [Rush
ments on nighttime circuits have been attributed to a com- and Gibbs, 1973] and therefore offer little difficulty.
bination mode in which one hop relies on scatter from F Short-term forecasts of foF2 have been made by relating
layer irregularities [McNamara, 1974b]. Propagation above changes in foF2 to corresponding changes in selected solar
the MUF due to scatter by small scale irregularities can and geophysical variables such as the 10.7 cm solar flux
routinely be observed [Rawer, 1975]. and the geomagnetic index Kp [Bennett and Friedland, 1970;
The high latitude also produces its share of unusual Ichinose et al, 1980]. The disadvantage of this type of ap-
propagation modes, mostly because of the presence of the proach is that the independent variables upon which changes
midlatitude trough. Strictly speaking, the modes are not in foF2 are assumed to depend must themselves first be
different from those expected-they are just heavily affected predicted. Even if this is done successfully, only limited
by the presence of the trough. success is possible because the correlations of these param-
Another interesting propagation mode is the round-the- eters with parameters of the ionosphere are not very high.
world (RTW) propagation mode. This mode relies on as- A more successful approach to the short-term forecasting
pects of the vertical distribution of electron density rather of foF2, or alternatively of an operational MUF on a given
than on latitudinally localized features of the ionosphere. circuit, is a prediction scheme based directly upon imme-
Consideration of observed elevation angles, signal losses diate past observations of foF2 or an operational MUF. Such
and time delays have led to the conclusion that RTW signals prediction schemes for foF2 are described, for example, by
propagate via a chordal mode of propagation, rather than Rush and Gibbs [1973], Lyakhova and Kostina [ 1973],
by a uniform multihop mode. McNamara [ 1976], and Wilkinson [ 1979].
Rush and Gibbs [ 1973] used a five-day weighted mean
value of foF2 to predict daily and hourly values of that
10.4.8 Short-Term Forecasting parameter. The method of Lyakhova and Kostina [ 1973] is
of HF Conditions based on the observation that correlation coefficients be-
tween the deviations of foF2 from median values remain
Long-term predictions of monthly median parameters of greater than 0.5 for up to four hours. The high correlation
the ionosphere such as described in Section 10.4.4 are the between hour-to-hour variations of foF2 has been discussed
traditional approach to frequency management. Frequencies by Lyakhova [ 1960], Radinov [ 1963], Gautier and Zachar-
are selected which should ensure at least a 9070 success rate isen [1965] and Rush [ 1972].
for communications at all times of day, during any season, McNamara [1976] made predictions of foF2 at a par-
and at all parts of the solar cycle. Engineering decisions ticular location up to 3 hours ahead by projecting forward
such as transmitter power and antenna configuration are also the trend in the departures of the last few hours’ observations
made at this stage. Optimum working frequencies (OWF) from a 15-day running median. Wilkinson [ 1979] on the
are derived for each hour of the day, usually for a month other hand, simply projected forward in time the deviation
at a time, using empirically derived frequencies and signal- of an observed foF2 value from the predicted monthly me-
to-noise ratios. Use of predictions in this manner continues dian value of foF2. He found this technique to be effective
to provide acceptable, though not high quality, communi- for lead times of up to about 3 hours.
cations for many purposes such as voice and telegraph con- Similar techniques have been applied to oblique circuits

10-60
 a
IONOSPHERIC RADIO WAVE PROPAGATION
by Ames and Egan 1967], Ameset al. 1970], Krause et ditions requires the projection forward in time of the effec-
al. [1970, 1973a, b], D’Accardi [1978], nd Uffelman and tive ionospheric index. In the limit of zero lead-time, the
Harnish [1982]. “forecast” in fact becomes a real-time assessment of the
The success of any of these forecasting schemes will ionosphere.
depend on the particular circumstances of its intended use, A brief review of short-term forecasting for HF prop-
especially as regards the requried accuracy of the forecast agation has been given by CCIR [198 Ib]. The special case
and the lead time required. Most schemes are reasonably of high latitude propagation is considered in CCIR [ 1983].
successful in forecasting an operational MUF that is closer
to the actual value than is the predicted monthly median
value, but only for lead times of the order of an hour or 10.4.9 Real Time Channel Evaluation
less.
Forecasts can be made with different lead times, de- To take full advantage of the HF communications po-
pending on how closely the variations in the ionosphere tential of the ionosphere and to overcome its inherent vari-
need to be tracked, and the sampling interval for actual ability, frequency management should be implemented in
observations must match that lead time. In general, the error three stages, namely long-term predictions (Section 10.4.4),
in the forecast values will increase with the lead time and short-term forecasting (Section 10.4. 8) and real-time chan-
if this error becomes too great, the forecasting scheme would nel evaluation (RTCE). It is at the first stage that engineering
offer no advantages over the use of a monthly median value. decisions Sbch as site location, antenna configuration, and
To make short-term forecasts for circuits for which no transmitter pwer are made, and suitable operating fre-
real-time observations exist, the behavior of the ionosphere quencies applied for from the appropriate regulating body.
must be inferred from such data as are available, using these The second stage, that of short-tern forecasting, determines
observations to infer the values along the required circuit. which of the allocated frequencies will actually propagate
Numerous studies have reported correlation coefficients that now or a short time ahead. These two stages are generally
illustrate the degree to which hourly deviations from median all that would be required for voice and low-speed teleg-
values of ionospheric parameters at two or more locations raphy circuits.
me related [Gautier and Zacharisen, 1965; Zacharisep, 1965; This two-stage approach has several limitations:
Zevakina et al., 1967; Rush, 1972]. These correlations may 1. The signal-to-noise data are not always reliable and
be used to infer values of foF2 at locations remote from the this ratio is not necessarily a useful criterion for choosing
observing site, with an error depending on the separation frequencies for some forms of communications.
of the two locations in latitude and longitude, and on the 2. Reliable long-term predictions of sporadic E are not
local time and season. Rush [1976] has used the observed available. (Es modes are often the best to use. )
correlations to establish the usefulness of, and requirements 3. No account is taken of interference from other users.
for, a network of ground-based and satellite-borne iono- 4. The forecast available frequency range will be un-
spheric observations whose measurements are to be used certain to some extent because of inevitable errors in the
for short-term forecasting of radio propagation conditions. forecasting models.
However, it should be noted that the relatively high cor- 5. The approach does not indicate which of the assigned
relations used by Rush originate from the very disturbed frequencies propagating at a given time would be the best
days (that is, ionospheric storm days) and these same high to use for a given form of communication.
correlations are not always obtained for days when the de- RTCE is the third stage of a frequency management
viations from the median values are relatively small. system required to maintain reliable high quality HF com-
A third approach to short-term forecasting is to use ob- munications even under the most adverse conditions. It be-
servations at all available ionospheric stations to dete~ine comes especially important for medium and high speed dig-
an effective ionospheric index, which can then be used in ital data transfer. With an RTCE system, all channels are
conjunction with synoptic monthly median maps of foF2 sounded in turn to determine which is actually the best to
and other parameters (see Section 10.4. 4) to predict the use at a given time for the particular type of communica-
values of these parameters at the reflection points of the tion/modulation system.
given circuit. The limitations of this technique are set by Damell [1978] has given the following definition of
the accuracy with which an appropriate index can be de- RTCE: “Real-time channel evaluation is the term used to
termined from a restricted subset of observations and the describe the processes of measuring appropriate parameters
accuracy with which the ionospheric model, driven by this of a set of communication channels in real time and of
index, can reproduce the actual ionosphere at the reflection employing the data thus obtained to describe quantitatively
points. An example of the use of ionospheric indices to the states of those channels and hence the relative capabil-
update models of the ionosphere has been given by Mc- ities for passing a given class, or classes, of communication
Namara [1979]. Other, more complicated, methods have traffic:”
been described by Thompson and Secan [1979] and Tas- A review of RTCE has been prepared by Study Group
cione et al. [1979]. Actual forecasts of propagation con- 6 of CCIR [CCIR, 198 ld], and the interested reader is

10-61
 CHAPTER 10
referred to that report. One of the more recently developed transient propagation modes, for example, sporadic E layer
techniques of RTCE, which can serve as an example of the propagation.
technique, is that of pilot tone sounding. 3. RTCE evaluation allows a more efficient use of the
In pilot tone sounding [Betts and Damell, 1975], low frequency spectrum by tending to select frequency channels
level CW tones are either inserted into the data spectrum higher than those which would be chosen via prediction
or transmitted in potentially available alternative channels. techniques. Thus spectrum congestion is reduced.
At the receiver, simple measurements on the tones enable 4. RTCE will provide a means of automatically selecting
the relative states of the channels to be specified in terms the best frequency and simultaneously indicating preferred
of predicted (digital) data error rates. In a situation where stand-by channels.
a multicomponent broadcast is being used to radiate identical 5. Transmitter power can be minimized, consistent with
data simultaneously on each of several frequencies, this type providing an acceptable quality of received traffic.
of RTCE system allows the best component to be selected 6. In the longer term, RTCE data can be used to adapt
automatically at the receiver. other parameters of a communication system, apart from
One of the main advantages of the pilot-tone RTCE frequency, optimally for the prevailing path conditions.
technique is the extreme simplicity of the concept and the
implementation when compared with other techniques for
RTCE, for example, oblique incidence ionosondes. A fur- 10.4.10 Conclusion
ther advantage of the technique is that it permits RTCE and
data signals to be combined in a simple format, rather than It has not been possible within the present limitations
requiring a separate stimulus. The technique also lends itself of space to present more than a cursory overview of the
readily to automation. With an automated system, there will subject of HF propagation. However, it is hoped that most
be little difficulty in principle in including into the algorithm readers will be able to follow up on topics of special interest
for selecting the optimum working frequency any measure- by going to the references cited. In this regard, the value
ments of other pertinent properties of the pilot tone, for of the proceedings of the “Boulder Workshop” (Donnelly,
example, amplitude, Doppler shift, and spectral width. The 1979; 1980) cannot be stressed too highly. The CCIR doc-
noise level in the channel could also be included. uments and the AGARD Lecture Series No. 127 should also
The pilot-tone sounder technique does not permit the prove very useful, but possibly somewhat harder to obtain.
determination of the operational MUF or identification of
propagation modes. The former may cause difficulties when
the MUF is decreasing (for example, due to normal diurnal 10.5 IONOSPHERIC DISTURBANCES
variation) if the working frequency is just below the MUF.
Difficulties could be severe if the MUF changes significantly
during the RTCE time (typically 3 min per channel). 10.5.1 Sudden Ionospheric Disturbance (SID)
The signal optimization problem at HF is potentially
very complex and to date RTCE has been used almost ex- The SID is caused by x-ray and ultraviolet emissions
clusively as a frequency selection aid. Damell [ 1975b] lists from solar flares. These emissions produce increased ioni-
the various parameters on an HF link which could be under zation of the sunlit ionosphere. This excess ionization typ-
the control of the communicator and hence be made adaptive ically lasts for a period of 10 to 60 minutes, depending on
in response to the RTCE data: the intensity and duration of the responsible flare. SIDS are
Transmission frequency observed by monitoring manmade and natural High Fre-
Transmitter power level quency (HF) or Very Low Frequency (VLF) radio signals
Bandwidth that propagate via the ionosphere. The various types of SIDS
Data rate are named according to the method of their observation:
Modulation type 1. SWF—short wave fadeout: a decrease (either sudden
Time at which the transmission is made or gradual) in the signal received from a distant HF
Signal processing algorithm at the receiver (2–32 MHz) transmitter,
Elevation angles for antenna array beams 2. SCNA—sudden cosmic noise absorption: a decrease
Diversity type in the intensity of the constant galactic radio noise as
To this list may be added antenna polarization for near measured by riometers (relative ionospheric opacity
vertical incidence propagation. meters) operating between 15+0 MHz,
The potential advantages accruing from the use of RTCE 3. SPA—sudden phase anomaly: a change in phase of
techniques have been summarized by Damell [ 1975a]: a received VLF ( 10–150 kHz) signal relative to a
1. The effect of man made noise and interference can frequency standard at the receiving site,
be measured and specified quantitively [Damell, 1978]. 4. SES—sudden enhancement of signal: an increase in
2. The facility for real-time, on line measurement of the strength of an incident VLF signal occurring at
propagation and interference allows the use of relatively the same time as an SPA,

10-62
 a
IONOSPHERIC RADIO WAVE PROPAGATION
5. SEA—sudden enhancement of atmospherics: an in- of an over-the-horizon (OTH) radar designed for the sur-
crease in the background VLF noise from distant veillance of aircraft at ranges of approximately 1000 to 3500
thunderstorms, km.
6. SFD-sudden frequency deviation: a short-lived in- The radar reflections are the result of scattering from
crease in the frequency of the signal from a distant electron density irregularities aligned alorig the lines of force
HF transmitter, of the earth’s magnetic field. The characteristics of the field-
7. SFE—solar flare effect or geomagnetic crochet: sud- aligned scatterers are such that the radar echoes originate
den variation in the H component of the earth’s mag- in a small range of angles about perpendicular incidence to
netic field. the magnetic field lines and that the amplitude of the auroral
The first five of these effects are all attributable to increased echo is aspect-angle sensitive. In addition to the orthogo-
ionization in the D layer, primarily by soft ( 1–8 ~) flare x- nality condition, it is necessary that this geometric config-
rays. The absorptive effects (S WF and SCNA) are caused uration take place at ionospheric altitudes, that is, 80 km
by a thickening of the absorbing layer through which the and above.
HF waves pass, while the effects on VLF signals are due The probability of observing radar reflections from ion-
to the lowering (SPA) and strengthening (SES and SEA) of ization irregularities is also dependent upon the frequency
the D layer from which these lower frequency waves are of occurrence of E layer auroral or spread-F irregularity
reflected. In contrast, SFDS arise from the flare associated activity in the region of interest. Reflections can be expected
increase in ionization in the E and F regions prod~ced by to occur in regions where both the conditions of near-per-
extreme ultraviolet (EUV) radiation in the 10–1030 A range. pendicularity at ionsophenc heights and high auroral or spread-
Finally, the geomagnetic crochet (SFE) appears to be a F irregularity activity are satisfied. The fulfillment of only
hybrid effect, having an impulsive component associated one requirement is not sufficient to warrant a radar reflec-
with the flare “flash phase” EUV emission, and a more tion.
gradual component associated with the flare soft ( 1–8 ~) x- An appropriate model to use to define the condition of
rays [Richmond and Venkateswaran, 1971]. These ionizing auroral activity is the Feldstein-Starkov [1967] auroral belt
emissions temporarily increase the D and E layer conduc- (oval) model. Since the location and extent of the auroral
tivity and alter the normal ionospheric currents to give rise oval are a function of time and geomagnetic activity, E layer
to the SFE. For a comprehensive review of the SID phe- auroral echoes can be expected to appear over a wide area
nomenon, see Mitra [1974]. during magnetically disturbed periods for an HF backscatter
radar located at high latitudes.
The F layer irregularities can be described in terms of
the probability of occurrence of spread-F derived by Penn-
10.5.2 Polar Cap Absorption (PCA) dorf [1962] or in terms of Aarons’ [1973] irregularityy scin-
tillation region.
While SIDS are caused by flare electromagnetic emis- In this section, an estimate is made of the characteristics
sion, PCAS result from bombardment of the ionosphere by of ionospheric clutter that could be observed by an HF
flare-accelerated protons with energies s 10 MeV. These backscatter radar operating in the midlatitude with the an-
particles stream into the earth’s polar regions along geo- tenna beam oriented towards the polar region. The topics
magnetic field lines and produce increased ionization in the to be discussed are the amplitude, the backscatter cross-
D layer (as well as aurora at E layer heights). PCAS are sectional area, the angular extent, the Doppler frequency
normally observed by means of riometers. In contrast to the variation, the frequency of occurrence, the diurnal and sea-
SID, the PCA is a long-lived effect, with durations ranging sonal variation and the correlation with solar-geophysical
from tens of hours to several days. In general, PCAS follow conditions.
only the most intense solar flares and begin within a few
hours after flare maximum, dependent on the flare longitude.
A review of the PCA phenomenon has been published by 10.6.1 Signal Amplitude
Hultqvist [1969].
The amplitude of E layer auroral clutter that could be
encountered by an HF backscatter radar is deduced by ex-
trapolation of radar-auroral data recorded by SRI Intern-
10.6 HF RADAR IONOSPHERIC ational at Fraserburgh, Scotland, during 1959 and 1960
CLUTTER [Leadabrand et al., 1965]. The radar measurements were
made simultaneously at frequencies of 30, 401 and 800
Backscatter reflections from E layer auroral ionization MHz.
and F layer electron density irregularities can be an impor- To predict the SRI International auroral data effects on
tant source of clutter for a radar operating in the high lat- an HF backscatter radar in terms of the corresponding radar,
itudes. The ionospheric clutter can degrade the performance it is necessary to determine the relative sensitivities of the

10-63
CHAPTER 10
For aurora which fills the antenna beam, a is replaced
by the volume scattering coefficient, Uv, that is, radar cross
section per unit volume, such that, for a rectangular antenna,

(J = UVV = U,R2PHPV; > (10.61)

where V is the volume filled by the scatterers, c is the free

space velocity, T is the pulse length, and ~“ and ~v are the
horizontal and vertical antenna beamwidth, respectively.
-30 L
Figure 10-54 is a plot of the relative signal strengths of
auroral echoes recorded by SRJ International at the three
frequencies in April 19@. The region of uncertainty be-
Pigure 10-54. Relative signal strength of aurora] echoes at 30, 41, and
800 MHz obtained by the SRI International-Scotland radars tween 30 and 401 MHz results from the fact that there was
[hadabrand et al., 1965]. no way to specify to what degree the auroral scatterers filled
the 30-MHz antenna beam. At401 and 80 MHz, the antenna
beamwidth was 1.2°, while at 30 MHz it was 15°. Thus, it
radars to the aurora. This is accomplished by comparing the was assumed that the scatterers completely filled the beam
various parameters that enter into the radar equation. at the two higher frequencies. However, it was doubtful
According to the radar equation, the signal-to-noise ratio that the beam was completely filled at 30 MHz. The 10 X 1°
of the received signal is given by aurora is assumed to be a point target for the 30-MHz an-
tenna.
Pr P,G,G,A2U The data in Figure 10-54 are normalized with respect to
(10.60)
k= (4m)3R4 KTB ~F L,LP 401 MHz. For a filled beam antenna, the relative signal
strength of the 30-MHz echoes was 6-dB greater than those
where Pt is the transmitted power, Gt and G, are the gains at 401 MHz, while, for the point target case, there was a
of the transmitting and receiving antennas, respectively, A 28-dB difference between the 30- and 401-MHz data.
is the wavelength, u is the radar-target cross section, R is The absolute magnitude of the 30-MHz auroral echoes
the radar range, K is Boltzmann’s constant, T is the ambient can be deduced by comparing the system sensitivities of the
temperature, B is the receiver noise bandwidth, ~F is the 30- and 401 -MHz radars. According to Table 10-9, the401 -
receiver noise figure, L, is the system loss and ~ is the MHz radar was 48 .9-dB more sensitive than the 30-MHz
two-way loss due to the propagation medium. This rela- radar assuming that the aurora was a point target and 26.9-
tionship is applicable to a point target, that is, the antenna dB more sensitive, assuming that the 30-MHz antenna beam
beamwidth is much larger than the dimensions of the ttiget. was completely filled with scatterers. Thus, it follows that

Table 10-9. Comparison of SRI International-Scotland 30-MHz and 401 -MHz radar sensitivities utilized in the April 1960 radar-auroral observations

Scatterers Fill
Point Target Antenna Beam
30-MHZ 401-MHZ (401 MHz/30 MHz) (401 MHz/30 MHz)
Parameter Radar Radar (dB) (dB)

AZ(mz) (10.0)2 (0.748)2 – 22.5 –22.5

P, (Mw) 0.0015 0.12 +19 + 19
G,G,(dB) 40 84 +44 +44
15 1.2 — –11
~H

(DEG)
15 1.2 . –11
Pv

(DEG)
?(ps) 300 300 0 0
B (kHz) 6 6 0 0
NF (dB) 9 4.5 +4.5 +4.5
L, (dB) 7 3.1 ~ ~
+48.9 +26.9

10-64
 a
IONOSPHERIC RADIO WAVE PROPAGATION

w 99
45

40
/
m ~90
: 35
2!
s 30 /
a 5 70 - P
3
: 25 E 30 MHz
o : 50
L 20
0 E
~ 30 /
g 15

10 E

5
~ 10
F
0 ~
-115 -105 -95 -85 -75 -55 -55 -45
AMPLITUDE [dBm)
“z 1
3
u ●SRI EXPERIMENTAL DATA
Figure 10-55. Amplitude histogram of SRl Intemationa1401 -MHz Scot-
0.1 STANDARD DEVIATION = 14.8 dB .
land radar auroral data [badabrand et al., 1965].

0.01 , 1 , 1 , ! 1
the magnitude of the 30-MHz radar-auroral data was 20.9- -120 -104 -88 -72 -56
dB less than that of the 401-MHz data. AMPLITUDE ( dBm )
The amplitude distribution of the401 -MHz radar-auroral
echoes is presented as a histogram form in Figure 10-55. Figure 10-56. Cumulative distribution function of SRI Intemational-Scot-
The system noise level was approximately – 133 dBm and land radar auroral data.
that the system saturated at – 60 dBm, which accounts for
the large number of echoes at that amplitude. ure 10-55 was computed utilizing the value of – 77.2 dBm
The histogram data, when replotted as a cumulative for the mean as obtained from Figure 10-56. The 30-MHz
distribution on probability paper, shown in Figure 10-56, amplitude data, also shown in Figure 10-56, was derived
closely fit a straight line. This characteristic is indicative on the basis that the 30-MHz auroral echoes followed a
that the 401 -MHz amplitudes are log-notmally distributed. Gaussian distribution and were 20.9-dB less than the 401-
The theoretical normal distribution function plotted in Fig- MHz results.

Table -.
. . . .. 10-10. Comparison of the sensitivity of an HF backscatter radar to that of the SRI Intemational-Scotland 30-MHz radar.

Scatterers Fill
HF SRI- Point Target Antenna Beam
Backscatter Scotland (HF Radar/SRI) (HF Radar/SRI)
Parameter Radar Radar (dB) (dB)

A* (m*) (10.0)2 (10.0)2 o 0

P, (MW) 0.8 0.0015 +27.3 +27.3
G, G, 45 40 +5 +5
(dB)
5 15 — –4.8
~H
(DEG)
15 15 — o
p“

(DEG)
400 300 — +1.2
~ (Ws)
10 – 14.8
B (kHz) 2.5 6 +3.5 +3.8
100 – 12.2 – 12.2
~ (dB) 12 9 –3 –3
L, (dB) 3 7 +4 +4
(? = 400 PS) +37.1 dB +33.5 dB
(T = lops) +21.1 dB + 1.5dB

10-65
CHAPTER 10
Table 10-11. Estimate of radar-auroral clutter levels of a hypothetical HF backscatter radar at 30 MHz based on SRI International data.

Scatterers Fill
Point Target Antenna Beam

Statistical T=400~S T=lo~S T=400ps T=io~S

Parameter (dBm) (dBm) (dBm) (dBm)

Upper Decile –42.0 – 58.0 –45.6 –77.6

Upper Quartile –51.0 –67.0 –54.6 – 86.6
Median –61.0 –77.0 –64.6 –96.6

Lower Quartile –71.0 – 87.0 – 74.6 – 106.6

Lower Decile – 80.0 – 96.0 –83.6 –115.6
Standard 14.8 14.8 14.8 14.8
Deviation

To estimate the magnitude of the auroral echoes that imately the 10- to 20-MHz range. From simultaneous au-
could be observed by an HF backscatter radar, it is necessary roral echo measurements at frequencies of 49.7, 143.5 and
to compare the sensitivity of the HF backscatter radar to 226 MHz, Flood [1960] has deduced a value of n = 3.5
that of the SRI International 30-MHz radar. The parameters between 49.7 and 143.5 MHz and n = 6.5 between 143.5
of an hypothetical HF radar, considered in this analysis, are and 226 MHz.
given in Table 10-10. For an assumed pulse length of 400 Measurements of E region radar aurora at frequencies
ps, the HF radar is 37.1 -dB more sensitive than the SRI in the HF band and at VHF (143.8 MHz) from locations in
International 30-MHz radar for the point target case and Finland and Germany indicated a frequency dependency of
33 .5-dB more sensitive for the filled beam case. When the n = 3 [Oksman et al., 1979].
pulse length is decreased to 10 ps, the difference in sen- Since there is a wide discrepancy in the experimental
sitivities decreases to + 21.1 dB and + 1.5 dB for the point measurements of the frequency dependence of auroral back-
target and filled beam case, respectively. scatter, which could be due to the characteristics of the
The predicted HF backscatter radar-auroral clutter levels auroral ionization, that is, inhomogeneous distribution of
at 30 MHz are presented in Table 10-11. The SRI Inter- aurora] electrons, varying scale sizes of ionization imegu -
national 30-MHz radar-auroral data given in Figure 10-56 larities and different scattering altitudes, a value of 4 for
and radar sensitivities in Table 10-10 are used in the cal- the exponent in Equation ( 10.62) is assumed in this analysis.
culations. The estimated median aurora] clutter amplitudes that
In, deducing the radar-aurora] signal levels at other fre- could confront the hypothetical HF backscatter radar in the
quencies in the HF band, it is assumed that the auroral echo 5-to 30-MHz band are shown in Figure 10-57. The external
power is frequency dependent according to the law

p = k f-n, (10.62) -20 — ; “ ‘––T

where k and n are constants. Utilizing the data in Figure

m’
~
10-54, it can be shown that for this power law n = 2.5 for
u .60
a point target and n = 0.5 for the filled beam case. How- 0
3
+
ever, an examination of other radar-auroral data taken at ~‘“’[--;’:LFS
_m —————
; ~ ~

Scotland revealed a value of n = 6.8 for a point target and z —POINT TARGET ‘ ~ ‘.. ‘
a ‘— FILLED BEAM ‘:.
n = 2.7 for a filled beam [Leadabrand et al., 1965]. An o
x -1oo — y= ~uL~E LENGTH. . + ,— ‘i.
v 1
analysis of auroral echoes from simultaneous multiple fre- . t I
,,
quency observations in Alaska by Leadabrand et al., [ 1967] <-120 .-+--. .{ +4.
L,
E
revealed than n = 2 for frequencies between 50 and 850 ~
MHz and n = 5 between 850 and 3000 MHz.
<-140 EXTERNAL NOISE
7’
LEvEL>. J- ‘ ‘ ‘
\
Radar backscatter measurements of artificial electron -160 [ I 1
clouds in the E region of the ionosphere by Gallagher and ,00 2 46 8 ,.l 2 46 8,02

FREQUENCY (MHz)
Barnes [1963] yielded a constant, n, of 4 for frequencies
between approximately 20 and 50 MHz and – 4 between 5
Figure 10-57 Estimated median auroral clutter amplitude based on SRI
and 20 MHz. At times, however, it was found that the International-Scotland radar data, f4 frequency depen-
amplitude returns were insensitive to frequency in approx- dency.

10-66
 a
IONOSPHERIC RADIO WAVE PROPAGATION

110 11o-

100 100
z s
~ s
————.
- -- ----
290 990
-.. , ;7 = 400 ps +
G 6 1“” ; “-““ “-~~=’~’- ‘-
K K
.J:- :--, --:.. ___
#so -
:60
----- +1 i
5 E
z ---
z \ PDINURGET: ~ ._
1 70 ; 70
----FILLED BEAM -–+-” - ““”
p — POINT TARGET I
y 7= lops
––- FILLED BEAM T= 1Op 7 = PULSE LENGTH
A K
7 = PULSE LENGTH .— —
:60 . -— —--- :60
+ - . . / + /
3 -. . 3
2 --- ---
dso Uso =.-.—..—. .—.. .
/
1-1 ~1 ‘ -{ ~
/’
401 I I I 1,1 I I I 1<] 401 II
,~o 2 4 6 8 ,~1 2 4 6 8,.2 ,00 2 46B ,01 2 466 ,02

FREQUENCY (MHz) FREQUENCY (MHz)

F]gure 10-58. Estimated median radar-auroral clutter-to-noise ratio based Figure 10-59. Estimated median radar-aurora] clutter-to-noise ratio based
on SRI Intemational-Scotlmd radar data, f4 frequency de- on SRI Intemational-ScotIand radar data, f-z frequency de-
pendency. pendency.

noise level of a rural environment also shown in Figure agation and magnetic field lines are not orthogonal, the
10-57 is the average median value as predicted for a location auroral clutter amplitude should decrease.
in Maine at an azimuth of 60° for all seasons, all times of Bates and Albee [1969] report that, assuming a simple
day and a sunspot number of 70. It is seen that the auroral model of aurora] backscatter from the E region that includes
clutter amplitude increases with decreasing frequency and ionospheric refraction effects, a lower limit of 6 dB/deg in
bandwidth and that the amplitude of the point target case the 15- to 50-MHz range was observed for the aspect-sen-
is greater than that of the filled beam case. sitive decrease in backscatter cross section with off-pe~n-
Figure 10-58 is areplot of thedata in Figure 10-57 in dicular angle from the geomagnetic field. The aspect-sen-
terms of theauroral clutter-to-noise ratio. Itisof interest to sitive decrease of the cross section of F layer backscatter
note that, for a fq frequency dependency, the clutter-to- echoes was found by Bates and Albee [1970] to be on the
noise ratio maximizes in the vicinity of 10 MHz. This is order of 5 dB/deg of off-perpendicularity from the magnetic
due to the fact that the external noise level increases with field in the 6- to 15-MHz region. An aspect angle decay of
decreasing frequency. 2 dB/deg at 30 MHz was deduced from the Scotland-auroral
When a frequency dependency of f-2, that is, n = 2 in measurements by Leadabrand et al. [ 1965].
Equation (10.62), is assumed, the ionospheric clutter levels
over the HF band are decreased, with respect to the f4
estimates, by an amount of 15.6, 9.5, 6.0, 3.5 and 1.6 dB 10.6.2 Cross Section
at 5, 10, 15, 20 and 25 MHz, respectively. As shown in
Figure 10-59, the clutter-to-noise ratio for the f-2 case, mon- HF radar observations conducted at Caribou, Maine,
otonically increases with increasing frequency. The clutter- under the Polar Fox 11program, at frequencies of 8.125 and
to-noise ratio at 30 MHz is 12.2 dB greater than that at 5 14.875 MHz, during the period between December 1971
MHz. The upper and lower decile values of the data pre- and November 1972 are used to estimate the cross section
sented in Figures 10-58 and 10-59 are + 19 dB and – 19 of ionospheric clutter.
dB with respect to the median level. The radar cross section given by Equation (10.61) can
The height distribution of the auroral echoes observed also be expressed by
by SRI International in Scotland was peaked at about 100
to 120 km, that is, in the E region, although heights as great
as 200 km were observed. Thus, the data presented in Fig-
ures 10-58 and 10-59 can be considered to apply only to E
region reflections. where Ah is the thickness of the aurora in the vertical di-
It should also be mentioned that the amplitude distri- rection. Since Ah was an unknown factor in the Polar Fox
bution of the 401-MHz auroral echoes shown in Figure 10- 11experiment, it was assumed to be unity ( 1m). Thus, Equa-
55, which was used as a basis of extrapolation to the HF tion (10-63) can be written as
backscatter radar, was applicable to data obtained when the
direction of propagation at the reflection point was perpen- UAA =
u = uARpH ~ (10.64)
dicular to the magnetic field lines. When directions of prop-

10-67
CHAPTER 10
Table 10-12. Average area scattering coefficient of E layer and F layer The cross section of ionospheric clutter that could be
irregularities based on ‘Polar Fox II data, Jhuary 1972~ observed by the HF backscatter radar is presented in Table
10-13. The calculations are based on an assumed radar range
Area Scattering
of 700 km for the E layer irregularities and 1500 km for
Coefficient (dBsm/m2)
the F layer irregularities. These ranges correspond to a ray
Frequency Statistical path oriented at an elevation angle of 5° and intersecting
(MHz) Parameter E Layer F Layer the ionosphere at altitudes of approximately 100 and 300
km, respectively. For this configuration, the E layer scat-
8.125 Upper Quartile – 34.2 –27.2 tering area evaluates to 95.6 dBsm and 79.6 dBsm for a
Median –47.6 –33.6 pulse length of 400 ps and 10 ps, respectively. For the
Lower Quartile –58.6 –42.0 corresponding pulse lengths, the F layer scattering area is
14.875 Upper Quartile –56.5 –37.1 99.0 dBsm and 82.9 dBsm. According to Table 10-13, it
Median –63.9 –47.2 is evident that the F layer clutter cross section is about 10
Lower Quartile –74.0 –54.3 to 20 dB greater than the E layer cross section at 8.125
MHz and 20 to 23 dB at 14.875 MHz. The differences
between the F layer and E layer cross sections are most
likely due to the fact that the F layer ionization level is
where U* is the area scattering coefficient and A is the area many times that of the E layer.
containing the scatterers. lt should be noted that the Polar Fox II estimates of the
The area scattering coefficients of E layer and F layer E layer and F layer area scattering cross section contain all
irregularities as deduced from the Polar Fox II data are given the propagation loss terms, that is, the ionospheric propa-
in Table 10-12. The data show the averages over all times gation losses had not been removed from the calculations.
in the month of January 1972, and over all azimuths in the Because of this, it is not .Nssible to use the Polar Fox II
90° sector between –-300 and + 60° with respect to true data for extrapolation to other frequencies in the HF band.
North. It is seen that, for all the statistical parameters, the
8. 125-MHz data are from 12- to 21-dB greater than the
14. 875-MHz results. The most interesting feature of the 10.6.3 Angular Extent
basic data, from which Table 10-12 was derived, was the
lack of evidence of the influence of the geographic locations The angular extent of HF ionospheric clutter was de-
of the ionospheric scatterers on the magnitude of the area termined from Polar Fox II radar data recorded in January
scattering coefficient. It was hypothesized that the scattering 1972 at six frequencies ranging between 8 and 23 MHz.
coefficient would be greater in the northerly directions since Figure 10-60 contains plots of the cumulative distribution
this is the region encompassed by the auroral oval [Feldstein of the angular extent of E layer, F layer, and the combined
and Starkov, 1967]. E and F layer clutter. According to Table 10-14 which

Table 10-13. HF backscatter radar+ stimated average cross-section of E layer and F layer irregularities based on Polar Fox 11data, Janwary 1972

Average Cross Section

Cross Section Ratio
(dBsm) (dB)

Frequency Pulse Length Statistical

(MHz) (ps) Parameter E Layer F Layer F Layer/E Layer

8.125 400 Upper Quartile 61.2 71.8 10.6

Median 48.0 65.4 17.4
Lower Quartile 37.0 57.0 20.0
10 Upper Quartile 45.4 55.7 10.3
Median 32.0 49.3 17.3
Lower Quartile 21.0 40.9 19.9
14.875 400 Upper Quartile 39.1 61.9 22.8
Median 31.7 51.8 20.1
Lower Quartile 21.6 44.7 23.1
10 Upper Quartile 23.1 45.8 22.7
Median 15.7 35.7 20.0
Lower Quartile 5.6 28.6 23.0

10-68
 a
IONOSPHERIC RADIO WAVE PROPAGATION
1.0 120 I I I I t60
— DOPPLER SPREAD (LEFT SCALE)
$ / ‘ /’ ——— DOPPLER SHIFT (RIGHT SCALE)
~ 0.8
m FL AYER / - - ---
- _- -- 100 - *5O
z --. ---
~ 0.6 / /--
,> - /A
E
/“
A ~
? 0,4 /- m AND F .LAYE R 80
F / 1-
a ~ E. LAY ER
/ /’
~ 0.2

5 60 -
u
o
081624324046 666472W66
ANGULAR EXTENT (DEG)
40

Figure 10-60. Cumulative distribution of the angular extent of ionospheric

/
clutter based on Polar Fox II data.
/
tlo

OL , 1 1 I I I
10
summtizes the statistical characteristics of the three curves 5 10 15 20 25 30
in Figure 10-60, the median values evaluate to 60.2°, 18.8° FREQUENCY (MHz)

and 27.6°, respectively. The angular dimensions character-

ized in Table 10-14 were deduced from ionospheric clutter Figure 10-61, Estimate of E layer HF radar-auroral doppler frequency
present only in the 1-hop propagation mode. spread and shift based on SRI International-Scotland mea-
surements at 401 MHz,

10.6.4 Doppler Frequency Spectrum [1967], the maximum Doppler frequency spread at 401 MHz
is on the order of 1.90 kHz at – 6 dB below the peak, 4.00
The Doppler spectrum of radar pulses reflected from the kHz at –12 dB, and 4.75 kHz at –18 dB.
aurora differs from that of the original transmitted pulse. The expected (one standard) deviation of Doppler fre-
The changes that can take place are (1) the center frequency quency shift and spread at HF is presented in Figure 10-61.
can be shifted, and (2) the spectral width can be increased. The estimates are based on the fact that the Doppler fre-
The shift in the center frequency corresponds to a drift quency variations are directly proportional to frequency and
motion of the auroral ionization. Radar-auroral data indicate on the assumption that shifts and spreads are normally dis-
an east-west drift before magnetic midnight and a west-east tributed. ~us, the maximum shift and spread are equivalent
drift after magnetic midnight [Leadabrand et al., 1965]. to the 3-sigma (standard deviation) value of the distribution.
Drift velocities of the order of 500 m/s are typical. Ac-
cording to the analysis of E layer radar-auroral echoes re-
corded in Scotland at a frequency of 401 MHz, the maxi-
mum Doppler shift is normally *2.15 kHz [Larson and
Hodges, 1967].
The spread in the Doppler spectrum is due to the random,
turbulent motion of the imegularities of electron density in
the auroral ionization. According to Larson and Hodges

Table 10-14. Statistical distribution of the angul~ extent of ionosphe~c

clutter based on Polar Fox II da~a.

Angular Extent (Deg)

Statistical E Layer F Layer E and F Layer

Parameter Clutter Clutter Clutter

Upper Decile 87.7 76.0 87.0

Upper Quartile 83.8 42.4 74.4 5 10 15 20 25 30
FREQUENCY (MHz)
Median 60.2 18.8 27.6
Lower Quartile 19.0 10.0 13.6 Estimate of E Ipyer HF radar-aurorat total doppler frequency
Figure 10-62.
Lower Decile 11.6 2.8 3.4 deviation based on SRI International-Scotland measur-
ements at 401 MHz.

10-69
CHAPTER 10
At 10 MHz, the Doppler frequency shift evaluates to *18 in sporadic E. The data did not reveal a seasonal depen-
Hz and the Doppler frequency spread becomes 16 Hz at the dence.
–6 dB level, 33 Hz at –12 dB and 40 Hz at –18 dB. Under quiet magnetic conditions, the field-aligned F
The estimates of E layer HF radar-auroral total Doppler layer echoes were found to be a sunset phenomenon and
frequency deviation are plotted in Figure 10-62. At 10 MHz, correlated with the sunspot cycle [Basti et al., 1974]. In
it is possible that the 1-sigma value of the total Doppler general, the echoes occurred in the range from 1050 to 1500
deviation could be on the order of +26 Hz at the – 6 dB km. However, during the daytime, long range echoes, at
level, *35 Hz at – 12 dB, and f38 Hz at – 18 dB. times, were observed at ranges of 3000 to 3300 km. An
E layer radar-auroral echoes, having radial velocities as interesting disclosure was the fact that the occurrence of the
high as 1450 m/s, have been observed at 17 MHz by Brooks F layer field-aligned echoes increased directly with the in-
[1966]. This corresponds to a Doppler frequency shift of crease in magnetic activity until the level K~, (Fredericks-
164.3 Hz. The maximum radial velocity of F region field- burg, Maryland, three-hour K index) attained a value of
aligned irregularities detected at 17.3 MHz [Baggaley, 1970] four. Beyond this level, there was a decrease in the occur-
was found to be on the order of 165 mls with a mean of 65 rence of the F layer echoes with no echoes being observed
m/s. In terms of the Doppler frequency shift, these velocities when K~, 2 7. Radar auroral measurements made at Stan-
evaluate to 19.0 and 7.5 Hz, respectively. ford University at a frequency of 17.3 MHz reveal that
The extrapolation to the HF band of the Doppler meas- maximum auroral activity occurred between 1900 and 0400
urements of E region irregularities conducted by Hofstee hours local time [Peterson et al., 1955].
and Forsyth [1969] at a frequency of approximate] y 40 MHz, According to the HF radar echo observations conducted
Balsley and Ecklund [ 1972] at 50 MHz, Balsley et al. [ 1973] at 12 and 18 MHz in the state of Washington, there was a
at 50 MHz, Greenwald et al. [1975] at 50 MHz and Hal- pronounced peak in the frequency of occurrence of the echoes
doupis and Sofko [1976] at 42 MHz are in agreement with from field-aligned irregularities near the time of local sunset
the results presented in Figures 10-61 and 10-62. [Hewer et al., 1966]. The characteristics of the echoes de-
tected at sunset were different from those occurring at night.
It was found that the local sunset echoes appeared at ranges
of about 2000 km and were generally spread in range (dif-
10.6.5 Frequency of Occurrence fuse). The occurrence of the sunset echoes was relatively
and ‘Correlation with independent of magnetic activity. The nighttime echoes, on
Solar-Geophysical Conditions the other hand, appeared at ranges of about 1000 km and
tended to display little spread in range (discrete). The oc-
The experimental observations of field-aligned echoes currence of nighttime echoes was highly dependent on mag-
at 19.4 MHz conducted over a three-year period ( 1961–1963) netic activity.
at a site located in the vicinity of Boston disclosed that the Hewer et al., [1966] noted a decreasing echo activity
echoes were present for as long as 11 h/day [Malik and with solar cycle, the maximum percentage occurrences dur-
Aarons, 1964]. Although a seasonal pattern was clearly ing the nighttime being approximately 80% in December
defined, there appeared to be a tendency for the reflections 1958, 37% in December 1963, and 35% in March 1964.
to occur on a greater number of days during the summer The (Zurich) relative sunspot number during the correspond-
than during the winter. The correlation of echo activity with ing period was 185.2, 11.8, and 14.5, respectively. Hewer
sunspot number was found to exist. For example, in 1961 and Makhijani [ 1969] have concluded that HF F layer back-
when the sunspot number was 53, echoes appeared on 73T0 scatter echoes arise from the same general irregularity re-
of the days at an average of 3.0 h/day. In 1963 when the gions in which spread F is detected.
sunspot number was 28, the echoes appeared on 55% of the Sprenger and Glode [ 1964] have reported that the diurnal
days at an average of 1.5 h/day. Maximum echo activity variation of the frequency of occurrence of radar-auroral
occurred between 1800 and 2000 h local time. echoes recorded from late 1958 to mid 1962 at 33 MHz at
A more thorough analysis of the 19.4 MHz field-aligned Kuhlungsbom (geographic coordinates: 54”N, 12“E; geo-
echoes recorded from 1961 through 1965 had been per- magnetic latitude: 54”N) showed a double maximum at ap-
formed by the Air Force Cambridge Research Laboratories proximately 0100 and 1700 hours local time. The probability
[Basu et al., 1973, 1974]. It was found that under quiet of occurrence of E layer auroral echoes was found to be a
magnetic conditions field-aligned E layer echoes showed a function of magnetic activity. No auroral echoes were ob-
summer evening maximum and appeared to be associated served at a magnetic index, Kp < 5, and 100% occurrence
with the ground-backscatter echoes from sporadic E [Basu at Kp = 9. With regard to the correlation of aurora] echo
et al., 1973]. A weak secondary maximum existed in the activity with sunspot cycle, Sprenger and Gltie [ 1964] have
winter with no detectable field-aligned echoes during the concluded that the maximum of auroral activity occurred in
daytime. During disturbed magnetic conditions, the field- 1960 about two years after the sunspot maximum and that
aligned E layer echoes increased with magnetic activity and the activity decreased to zero within a period of only one
appeared during the daytime with a corresponding decrease and a half years.

10-70
 a
IONOSPHERIC RADIO WAVE PROPAGATION
Brooks’ [1965, 1966] investigation of radar-auroral echoes For an assumed f-2 dependence, the radar-auroral clutter-
at 17 MHz conducted in 1959 and 1960 indicated that, for to-noise ratio at 5 MHz is 15 .6-dB less than that derived
the discrete echoes, the maximum number appeared bet ween on the basis of an f~ law. Assuming that the radar aurora
OOOC+300 hours local time, and for the diffuse echoes, is a point target, the clutter-to-noise ratio for a 400 ps pulse
between 1900-2100 hours local time. is estimated to be 3. 6-dB greater than’ that deduced for
A maximum number of backscatter echoes at 13.866 auroral scatterers that completely fill the antenna beam. For
MHz from the F region on 75% of the nights for March a 1O-PS pulse, the corresponding difference is 19.6 dB.
1958 was reported by Weaver [1965]. A minimum number An analysis of the HF ionospheric backscatter data ob-
of echoes was observed in June, the data being collected tained in Maine under the Polar Fox II program reveals that
during 1957 and 1958 at Ithaca, New York, with the antenna the F layer cross section could be 10 to 20 dB greater than
beam oriented in a northward direction. the E layer cross section at 8.125 MHz and 20 to 23 dB at
Backscatter observations of F region field-aligned irreg- 14.875 MHz.
ularities made at 17 MHz near Sheffield, England (geo- The median value of the angular extent of E layer, F
graphic coordinates: 53 .43”N, 1.58°W; geomagnetic lati- layer and combined E and F layer ionospheric clutier appears
tude: 65 .4”N) between mid-October, l@, and mid-January, to be on the order of 60.2°, 18.8° and 27.6°, respectively.
1966, [Baggaley, 1970] which was a period of low sunspot Extrapolating from the SRI International-Scotland UHF
activity, showed that, on the average, the echoes were pres- radar-auroral data, it is estimated that for 5-MHz radar trans-
ent on 23.370 of the days while only i. 570 of the total missions reflected from E layer auroral ionization, the
observing time was occupied by the echoes. A summer l-sigma value of the total Doppler frequency deviation, that
maximum and a winter minimum were found to exist. This is, Doppler frequency shift and spread, could be on the
seasonal variation correlates with Malik and Aarons’ [1964] order of A 13 Hz at the – 6 dB normalized signal level,
results. *17 Hzat –12dB, and +19 Hzat –18dB. At 30 MHz,
Baggaley [1970] noted that there was a correlation be- the 1-sigma Doppler deviations should increase to approx-
tween the onset of F layer echoes and the solar zenith angle imately *78 Hz, t 105 Hz, and ~ 114 Hz, respectively.
at the reflection point in the F layer (assuming a height of HF field-aligned backscatter echoes usually occurs most
300 km). That is, a maximum number of echoes occurred often near local sunset. However, experimental observations
at a solar zenith angle of 9W–95°, and a minimum at angles have shown, at times, peak activity near the midnight hours.
less than 85° and greater than 125°. In addition, no corre- There is a strong correlation of backscatter echoes with
lation was found to exist between F layer echo occurrence the solar cycle. That is, the percentage of days displaying
and magnetic activity. Of the total time for which F layer backscatter reflections decreases with decreasing sunspot
echo activity was present, only 11% was associated with number.
magnetic index Kp >4. The characteristics of sunset echoes are found to be
Field-aligned echoes from the F region observed at Bris- different from those occurring during the night. The sunset
bane, Australia, at 16 MHz occurred preferentially during echoes are generally of the diffuse type, and are independent
geoma~etic disturbances and correlated strongly with spad- of magnetic activity while the nighttime echoes are discrete
F and radio star scintillations occurring in the same region and correlate with magnetic activity.
of the ionosphere [Swenson, 1972].
An analysis of 29-MHz backscatter measurements car-
ried out in Northern (West) Germany and Scandinavia (geo-
magnetic latitudes of 55°–770N) by Czeehowsky et al., [ 1974] 10.7 SCINTILLATION ON TRANS-
confirmed the close correlation between radio-auroral oc- IONOSPHERIC RADIO SIGNALS
currence and geomagnetic activity. Most of the radio auroras
appeared in the afternoon and evening hours. A radio wave traversing the upper and lower atmosphere
of the earth suffers a distortion of phase and amplitude.
When it traverses ionospheric irregularities, the radio wave
experiences fading, phase deviations, apd angle of arrival
10.6.6 Conclusions variations. These signal ”fluctuations, known as ionospheric
scintillation, vary widely with frequency, magnetic and so-
The radar-auroral clutter-to-noise ratio in the 5–30MHz lar activity, time of day, season, and latitude.
frequency range, as predicted from the SRI International The irregularities producing scintillations are predomi-
radar data acquired in Scotland, is dependent on the fre- nantly in the F layer at altitudes ranging from 200 to 1000
quency dependence law of the auroral backscatter echo power. km with the primary disturbance ‘region for high and equa-
Assuming a frequency variation of f~, the clutter level max- torial latitude irregularities between 250 and 400 km. There
imizes at approximately 10 MHz. When a f-2 law is as- are times when E layer irregularities in the 90 to 100 km
sumed, the maximum clutter level is shifted to a frequency region produce scintillation, particularly sporadic E and au-
of 30 MHz. roral E.

10-71
CHAPTER 10
SCINTILLATIW AND FADE DURATION ANALYSIS

+, r I MIN
.; A
do.:
-* “-7
‘MEAN
LOWELL TIME
SIGNAL
., 9 SECS

h~
LEVEL
100
r’

ExCEEDED ORDINATE

Figure 10-64. Sample of intensity fading prcduced by signal passing through

irregularities. Fade duration and cumulative probability dis-
tribution are also shown,
DEPTH OF SCINTILLATION FADING ( PROPORTIONAL TO
DENSITY OF CROSSHATCHING ) DURING LOW N MODE~TE sOL~ AcT1vl~

Figure 10-63, Global depth of scintillation fading (proportional to density 10.7.3 Signal Characteristics
of crosshatching) during low wd moderate solar activity.

The amplitude, phase and angle of arrival of a signal

will fluctuate during periods of scintillation. The intepsit y
of the scintillation is characterized by the variance in re-
10.7.1 Global Morphology ceived power with the S4 index commonly used for intensity
scintillation and defined as the square root of the variance
From the global point of view there are three major of received power divided by the mean value of the received
sectors of scintillation activity (Figure 10-63). The equa-
torial region comprises an area within * 20° of the magnetic povverthat is, S2 = ‘12)~2(’)2 [Briggsand Parkin, 1963].
equator. The high latitude region, for the purposes of the
An alternative, less rigorous but simple measure of scintil-
scintillation description, comprises the area from the high
lation index has been adopted by many workers in the field
latitude edge of the trapped charged particle boundary into
[Whitney et al., 1969] for scaling long-term chart records.
the polar region. We shall term all other regions “middle
The definition is
latitudes”.
In all sectors, there is a pronounced nighttime maximum. P
At the equator, activity begins only after sunset. Even in SI = ‘ax – ‘M’” (10.65)
P ~,. + P~in
the polar region, there appears to be greater scintillation
occurrence during the winter months than during the months
where Pmax is the power level of the 3rd peak down from
of continuous solar visibility.
the maximum excursion of the scintillations and P~in is the
level of the 3rd peak up from the minimum excursion,
measured in dB [Whitney et al., 1969].
The equivalence of selected values of these indices is
10.7.2 Scintillation Examples indicated below.

The intensity fading can bes$ be characterized by an

idealized example such as in Figure 10-64. The mean signal s. dB
level is modulated by the passage through the irregularities 0.075 1
so that the signal level very rapidly increases and decreases.
In Figure 10-64 the mean signal level at times fades below 0.17 3
the 3 dB level and below the 6 dB level. The number of 0.3 6
fades and the fade duration for a typical 15 minute length
of signal from a synchronous satellite is shown in Figure 0.45 10
10-64 along with the cumulative probability distribution
function (pdf). In this example the signal was above the 6 Scaling of the chart records is facilitated by simply meas-
dB fade level 91.7% of time. uring the decibel change between the Pm,, and P~in levels.

10-72
 z
IONOSPHERIC RADIO WAVE PROPAGATION
The phase variations are characterized by the standard de- cartesian coordinate system (x, y,z), and a plane wave of
viation of phase, U+, over a given interval of fluctuation wavelength k traveling in the z-direction is incident on the
frequency. layer, then for weak scattering (rms phase +0 < 1) the spa-
Attempts have been made to model the observed am- tial spectrum, @( Kx,Ky), of phase (s) and log amplitude (x)
plitude pdf. Whitney et al. [1972] and Crane [1977] have (Yeh and Liu, 1982) on the ground (z = z) is given by
constructed model distribution functions based upon the use
of the Nakagami-m distribution (m = (S4)-2) and have shown

=‘(reA)2L{1;
&sin(%)
that the empirical models provide a reasonable approxi-
mation to the calculated distribution functions. Fremouw et ‘XS(KXKS) (10.66)
al. [1980] showed that the Nakagami distribution for inten-
sity and the normal distribution for phase may be used to
characterize the statistics of the scintillation signal. In ad-
-i)l}@~(KxKy
xCos[:(z
dition, the Rayleigh pdf provides a good fit to the data under
conditions of very strong scintillation (S4 - 1.0). the upper and lower signs within the bracket referring to x
and s respectively. In the above equation,

KX, KY,KZ spatial wave numbers in x, y and

10.7.4 Frequency Dependence z directions
.—
K: K: + K;

Observations IFremouw et al., 1978] employing ten fre- @AN(KX>Ky, KZ) - power spectrum of irregularity electron
quencies between 138 MHz and 2.9 GHz transmitted from density deviation
the same satellite, show a consistent wavelength (h) de- k the wavenumber of the radio wave
pendence of the form A15 of S4 for S4 less than about 0.6.
.— 2Tr
The frequency dependence becomes less steep for stronger
scintillation, as S4 approaches a maximum value near unity ( -)A
r. the classical electron radius
with a few rare exceptions. When S4 exceeds 0.6 (peak to
peak values > 10 dB) the frequency dependence exponent (= 2.818 x 10-15 m)
L irregularity y layer thickness.
decreases. (If two frequencies are being compared and both
experience strong scattering to the extent that each displays
In the presence of a relative motion between the prop-
Rayleigh fading (S4 - 1.0), then there is effectively no
agation path and the irregukuities the Spatial spectrum @( K,, Ky)
wavelength dependence over the frequency interval. ) When
is convected past the observer and a temporal variation of
strong scattering occurs but is not constant over the fre-
signal phase and amplitude is observed. If the irregularity
quency interval, the wavelength dependence is difficult to
structure does not change during the convection (‘frozen-
determine. The observations IFremouw et al., 1978] also
in’ hypothesis) and the irregularities have a uniform velocity
show. that the phase scintillation index, U@, varies as A for
u in the x-direction then the power spectra of log amplitude
both weak and strong scattering. However, in extremely
and phase in the frequency domain are given by
strong scattering environment, the frequency dependence of
phase scintillation is also weakened.

10.7.5 Fading Spectra

During their passage through the ionospheric irregular-

ities of electron densit y, radio waves from satellites undergo
spatial phase fluctuations. As the wave emerges from the
irregularity layer and propagates towards the ground, these where
phase fluctuations cause interference to occur and a dif-
fraction pattern in both intensity and phase develops on the *angular frequency of phase and amplitude fluctuation
ground. In the presence of a relative motion between the
diffraction pattern and the observer, a temporal variation of
intensity and phase results. By transforming the temporal
pattern to the frequency domain, the frequency spectra of
intens.ty and phase fluctuations are obtained.
If a thin irregularity layer lies in the plane z = O of a The term within the bracket with the upper sign is known

!
10-73
CHAPTER 10
as the amplitude or Fresnel filter function and that with the bandwidth of the frequency spectra, implying the devel-
lower sign as the phase filter function. The Fresnel filter opment of shorter scales in the diffraction pattern. It is found
function oscillates with the variation of frequency f and at- that the correlation lengths get progressively smaller with
tains its first maximum at the frequency f~ = u/& . increased strength of scattering [Rino and Owen, 1981; Basu
The behavior of the phase filter function is very different and Whitney, 1983]. Under conditions of strong scattering,
from the amplitude filter function as it fails to attenuate the the phase scintillation spectra are also believed to suffer
low frequency regime. Equation (10.67) provides a rela- from the refractive scattering from very large scale irreg-
tionship between the irregularity power spectrum in the ion- ularities [Booker and MajidiAhi, 1981]. Since wave prop-
osphere and the amplitude or phase scintillation spectrum agation through a strong irregularity environment has con-
obtained on the ground. Since the irregularity power spec- siderable systems applications, intensive work in this area
trum has a power law variation [Dyson et al., 1974; Phelps is in progress.
and Sagalyn, 1976] of the form K-p, the power spectrum of
amplitude scintillation shows a maximum at the frequency
f~ due to the Fresnel filter function. On the other hand, the
phase fluctuations are dominated by the low frequency re- 10.7.6 Geometrical Considerations
gime. At f > f~, both amplitude and phase scintillation spectra
show an asymptotic variation f]-p when the three-dimen- The intensity at which scintillations are observed de-
sional irregularity spectrum is of the form K-p. Thus from pends upon the position of the observer relative to the ir-
a study of weak scintillation spectra the spectral form of the regularities in the ionosphere that cause the scintillation.
irregularities causing scintillations may be deduced. Keeping both the thickness of the irregularity region and
Figures 10-65a and 10-65b show two samples of weak AN, the electron density deviation of the irregularity, con-
phase and amplitude spectra obtained from 244 MHz scin- stant, geometrical factors have to be considered to evaluate
tillation observations made at Goose Bay, Labrador by the data and to predict scintillation effects at a particular lo-
use of the Fleetsat geostationary satellite [Basu et al., 1982]. cation. Among these are:
The phase and amplitude scintillation data were detrended (a) Zenith distance of the irregularity at the ionospheric
by a high pass detrend filter with a cut-off frequency of layer. One study [Wand and Evans, 1975] found the inten-
0.0067 Hz and the data sample yields an rms phase deviation sity of scintillation related approximately to the secant of
of 2.4 radians and amplitude scintillation index Sq = 0.51 the zenith angle up to 70°; at angles >70° the dependence
conforming to weak scintillation criterion. Both spectra show ranges between 1/2 and the first power of the secant of the
an asymptotic variation in the high frequency region, the zenith angle.
amplitude spectrum showing f-235 variation, and the phase (b) Propagation angle relative to the earth’s magnetic
spectrum showing f-zb variation. If we consider that the field. Performing this calculation demands the use of an
scintillation spectra have an average f-25 variation, the cor- irregularity configuration and the consideration of Gaussian
responding three-dimensional irregularity power spectrum or a power law model for the irregularities. At high latitudes,
is expected to have a power law wavenumber spectrum of irregularities in one study were elongated along the earth’s
the form K-35. The decrease of power spectral density at magnetic field with a cylindrical form of axial ratio of 5
the low frequency end of the phase spectrum is caused by along the lines of force. Sheet-like irregularities with forms
the detrend filter. It may be noted that in the amplitude of 10 : 10 : 1 have also been found in recent auroral studies
scintillation spectrum the high frequency roll-off starts at [Rino et al., 1978]. For equatorial latitudes, this elongation
the Fresnel frequency of about 40 mHz. For an irregularity along the lines of force may be of the order of 50 to 100
layer height of 350 km, the observed Fresnel frequency [Koster, 1963].
yields the irregularity drift velocity as 37 ~sec. (c) The distance from the irregularity region to the source
When scintillations become intense, the theory outlined and to the observer (near the imegularities, only phase fluc-
above does not hold, and strong single scattering as well as tuations are developed). As noted in Mikkelsen et al., [1978]
multiple scatter theories appropriate to such cases have been and Crane, [1977] the theoretical scintillation index can be
developed [Yeh and Liu, 1982 and references therein]. In expressed in terms of the above factors when dealing with
such cases it becomes difficult to relate in a straightforward ionospheric irregularities represented by a Gaussian power
manner, the scintillation spectra to the irregularity structures spectrum.
in the ionosphere. Figures 10-65C and 10-65d show a sample Mikkelsen et al. [1978] have attempted to determine the
of intense amplitude scintillation data and its spectrum theoretical scintillation index, Sq, when the imegularities are
(S1 = 0.88) analyzed from 257 Hz scintillation data ob- described by a power-law power spectrum.
tained at Ascension Island near the crest of the equatorial Mikkelsen assumed the approximate dividing line be-
anomaly by the use of transmission from the geostationary tween weak and strong scintillation is -9 dB, with S1 <9
satellite, Marisat [Basu and Whitney, 1983]. In contrast to dB denoting the weak case.
the weak amplitude scintillation spectrum shown earlier, The geometric variation of Sd is provided in Mikkelsen
these spectra show a flat low frequency portion and increased et al. [1978].

10-74
 a
IONOSPHERIC RADIO WAVE PROPAGATION

20. -
r B 11111111 I 11111111 I 11111111 I I,,,,,,

T:
r

-2S8
r

m-
r ‘ “’’’”r‘“’’’’”‘ “’’’’”‘ “’’’”r‘“1 T= -276
i

P = -2,s0 P: -235 I

10. .
10-

0. .
0.

-lo -
-lo -

:
8
~.m _ , -m -

-m -
-30 -

-40 -
-40 -

MYSO~-~UT -50 -
-50
FS .5.0 Hz FC:O.~ HZ

t 1
,, , , , ,tvl, , , I ,,,
Om Ooi 01 1 10

FKO - HZ

a b

MAR, UHF 122079

7 MAR. UHF
257 MHZ 2158UT
122079
S4 =0.88
20

10
257 Mtiz
ASC. IS
2158 UT S4 = 0.88
n=-6.0+ 0.2

-lo
2
; -20
n
-30

-40

-50

-60

I - I

162 0-’ 10° 10’ 102

FREQUENCY- Hz

Figure 10-65. S~ctraof (a)phase scintillation and(b) intensity scintillation under weAscafier conditions at2#MHz obsewedat Gwse Bay, Labrador
on 31 March, 1979. (c) Data of 257 MHz intensityscindlladcmand (d) its spectrunrumierstrong scatterconditionsobserved at Ascension
Island on20 December. 1979.

10.7.7 Spread F and Scintillation 10.7.8 Equatorial Scintillations

The evidence from the correlation of scintillation occur- 10.7.8.1 Patch Characteristics. Ithas been established
rence and spread F [Rastogi, 1980] is that equatorial range that nighttime ionospheric equatorial irregularity regions
spread is associated with scintillation activity and frequency emerging after sunset develop from bottomside instabilities,
spread is not. Thus the available spread F maps cannot be probably of the Rayleigh-Taylor type. The depleted density
used for scintillation observations in these regions; they are bubble rises into the region above the peak of the F2 layer.
dramatically misleading in many cases. In the high latitude Steep gradients on the edges of the hole help to generate
region no statistical study has been made to correlate types the smaller-scale irregularities within the patch which pro-
of spread F with scintillation activity. It might be noted that duces intense scintillation effects [Basu and Kelley, 1979].
even range spread occurrence and scintillation have impor- A plume-like irregularity region develops, finally form-
tant differences. ing a patch of irregularities that has been likened to a banana

10-75
 .EOGRAPH8C
LATITUDE
,0.. -. 30- N
–—+
I
-----
,5.. _____ {- ~,:” ----------

km
o.—
_~[

(
-*Z
H“ANCA1O

-“’ /~’i~~:~ON
,..%—— .+
‘( ?.-~..~;”;
..--..,.- ,0.w 0.
.%.s

,,0-, 60. w
\
THREE DIMENSIONAL LO”C(l UDE

PATCH MODEL
Figure 10-67. Map of equatorial regions using the 1975 epoch of the
DMA magnetic inclination map. X marks sub-ionospheric
Figure 10-66. Three dimensional model of an equatorial irregularity patch
in the form of a banana or orange segment. intersection.

HUANCAYO, PERU

80 r
* --- K= O-I’ GUAM
; ‘1 — K=3+-9
,\
60 - I NOV - JAN NOV - JAN

40 - 40
r
20 - 20 -

t
12 18 24 6 12 LT 12 18 24 6 12 LT

-- A ,
●

60 - 60- ,41 \
\
1 !
FEB - APR I $ FEB-APR
I’ I \
40 - 40 -
1’
/’
w ;
20 - ;
v 20 -
z s’
tt
w 1’
K
a. 1 ,1
3 12 18 24 6 12 LT 12 18 24 6 12 LT
:
0
60- ,.--. a.
t
It
40 -

a
40
r MAY - JUL

I
1’
:

20 - 20 -
;

I I I
24 6 12 LT 12 18 24 6 12 LT
12 18

60- 60 -

AUG -OCT AUG - OCT

,..
40 - \ 40 -
#

20 - 20 –
f

I I 1
12 18 24 6 12 LT 12 18 24 6 12 LT

Figure 10-68. Comparison of seasonal patterns of occurrence of scintillation activity > 10 dB at 250 MHz at Guam and Huancayo under very quiet
(Kp = O – 1+ ) and disturbed (Kp = 3 + – 9) magnetic conditions.

10-76
 a
IONOSPHERIC RADIO WAVE PROPAGATION
or an orange segment. A cut through the center of the “ba- PERCENT KCURRENCE GREATER THAN 2da (S4 = 13) 1
nana” is shown in Figure 10-66. SUNSET (350 kin)

The characteristics of the patch development, motion i JUL

I

and decay can be summarized as follows: - AUG

1. A new patch forms after sunset by expanding west- ~ SEP

ward in the direction of the solar terminator with velocities - CCT

probably similar to that of the terminator. It comes to an - NOV

abrupt halt after typically expanding to an east-west di- ~ OEC

mension of 100 to several hundred kilometers. It appears I

JAN
4
to have a minimum size of -100 km.
j FE,
2. It is composed of field-aligned elongated rod or sheet-
- MAR
like irregularities. The vertical thickness of the patch is 50
to several hundred kilometers. The patch has maximum
intensity irregularities in a height region from 225 to 450
km, with irregularities extending to over 1000 km.
3. Its north-south dimensions are of the order of 2~
km or greater.
4. Once formed, the patch drifts eastward with velocities
LJ’~ APRIL
HUANCAYO
76
03
1 54 GH,
- OCT 77
09 15 LT

ranging from 100 to 200 fls.

Figure 10-69. Percentage occurrence of 1.5 GHz scintillation Z 2 dB
5. The patch duration as measured by scintillation tech- during Apr 197~t 1977.
niques is known to be greater than 21/2 h; individual patches
have been tracked by airglow techniques up to 3 hours where by the occurrence of L band, 1500 MHz activity, at Huan-
they have maintained their integrity [Weber et al., 1978]. cayo, Peru. That evidence is shown in Figure 10-69 [Basu
Effects have been seen over 8 h. et al., 1980b] L-band activity at Huancayo does not suffer
Studies of the variation in electron content in the patches from strong scattering or from saturation (as does 136 MHz
have been made by measuring the change in relative phase and 250 MHz data on occasion); the data show clear equin-
between the two characteristic waves (ordinary and extra- octial maxima.
ordinary modes) with polarimetem. It is found that the patches From available data it appears as if geomagnetic control
are regions of depletion in electron content. While the elec- of the occurrence of scintillation vanes with longitude. The
tron content depletions are found to be only of the order of generalization can be made that increased magnetic activity
20%, the satellite in situ data may indicate density depletions inhibits scintillation activity before midnight+xcept dur-
as large as two or three orders of magnitude at one fixed ing those months with very low scintillation activity (May–July
altitude. In a strong irregularity environment, however, fast for the region (0”–70”W) and November–January in the
fluctuations in polarization are often obtained. Lee et al. Pacific longitudes ( 1350–180”E). After midnight the scin-
[ 1982j have shown that scattering suffered by each char- tillation activity in general increases slightly with the pres-
acteristic wave may induce fast polarization fluctuations and ence of magnetic storms. The data shown in Figures 10-70
obtained expressions for the variance of these fluctuations and 10-71 are for a year’s observation in each case. The
for irregularities with Gaussian and power law type spectra. complexities of the magnetic control of scintillation occur-
rence are illustrated by the variations in the curves of oc-
10.7.8.2 Variation of Scintillation Activity. A variety cumence at each station in each season. For further details
of observatories used data taken over the same time period see Mullen [1973].
to compare scintillation activity at 250 MHz [Aarons et al.,
1980b]. One set of data was taken at Huancayo, Peru, Natal, 10.7.8.3 In-Situ Data. Basu and Basu [ 1980] have de-
Brazil, and Accra, Ghana with all observations made at veloped a model from in. situ, theoretical, and stint i]Iat ion
elevation angles greater than 20° and with distance between studies. In their morphological model of scintillations, in
the most separated stations about 70° of longitude; a map situ measurements of irregularity amplitude, AN/N, as com-
of both geographical and magnetic coordinates is shown on puted from T sees of data are utilized in conjunction with
the right side of Figure 10-67. simultaneous measurement of electron density N. A com-
A second comparison of data at 250 MHz was made bination of AN/N and N data provides the required AN
between observations from Huancayo and from Guam. The parameter as a function of position and time.
data are shown in Figure 10-68; activity minima occur from
May-July in Huancayo and from November-January in Guam. 10.7.8.4 Sunspot Cycle Dependence. From the view-
The conclusion is that the occurrence patterns are longitu- point of electron density variations, the equatorial region
dinally controlled. around the magnetic equator displays a complex pattern.
It should be noted that in general maximum intensity During the day an increase in maximum electron density
occurs in the equinoctial months. This can best be illustrated occurs away from the equator. The electron density contours

10-77
CHAPTER 10
--- K= O-l+
— K=3”-9
FEB. - APR.

--- K= O-1* NOV. - JAN. 80-

— K = 3+-9 ,.---, 2.57 MHz
\
: \
60 - 257 MHz \
60 - \
,,-% ;
ACCRA, GHANA
\ 1’
t, ACCRA , GHANA
40 - : ‘, 40 - :
I
#
20 - 20 -

L L I
12 18 24 6 12 LT 12 18 24 6 12 LT
,-.,
80 - \
1’ \
$
i ,,
60 - I NATAL, BRAZIL
1’
1’ NATAL , BRAZIL
40 - 40
1’ I
i

w
w
a
20 –
I
20, ~ ,2LT
+
z 1 I
w 12 LT
o 12 18 24 6 12 18 24 6
a
w
n
80
r 1’
i \,
~,
60 - \
i’ s HUANCAYO , PERU
1’ t
\
40 - 1’
I
:
20 - 1
I

12 18 24 6 12 LT

Figure 10-70. Seasonal patterns ofwcumence ofscintillation activity >6dB(S4 = 0.3)at 250 MHz forveryquiet(Kp = O – I+)and for disturbed
(Kp =3; – 9) magnetic conditions for Nov–Apr.

display a distinct trough of electron density in the bottomside 1. The equatorial anomaly has considerably higher elec-
and topside ionosphere at the magnetic dip equator with tion density values in high sunspot number years than in
crests of ionization at + 15°–200 north and south dip lati- years of low solar activity.
tudes; this is the Appleton anomaly with the region within 2. The occurrence of maximum electron density for
+ 5° dip latitude of the magnetic equator termed the elec- anomaly latitudes is near sunset in the years of high sunspot
trojet region. number and in the afternoon in years of low solar activity.
From the solar cycle minimum in 1974 and maximum Thus the post sunset irregularity patches attain high AN
in 1969–1 970, Aarons [ 1977] found that there was a higher levels in the years of high solar flux.
occurrence of deep scintillations during a year of high solar
flux than during a year with low solar flux for observations
at both Accra, Ghana and Huancayo, Peru.
Recent observations of L band scintillations from both 10.7.9 Middle Latitude Scintillation
MARlSAT and GPS during the period of maximum solar
flux ( 1979–1981 ) [DasGupta et al., 19811 have revealed The middle latitude scintillation activity is not as intense
that scintillation intensities maximize in the Appleton anom- as that encountered at equatorial, auroral or polar latitudes.
aly region rather than near the magnetic equator. However, activity may reach levels, primarily at VHF and
The conclusion in the study [Aarons et al., 1981a] was UHF, that will increase error rates of systems with low fade
that the intense scintillation activity during years of high margins. The reader is referred to Aarons [ 1982] and Bram-
solar flux are due to two factors: ley [1974].

10-78
 IONOSPHERIC RADIO WAVE PROPAGATION
‘-- K= O-l”
‘K =3%9
--- K= O-l+
—K =3=9 AUG. -OCT.
MAY - JULY
60 r 60-

40
I 257 MHz
257 MHz

40 - ACCRA, GHANA
ACCRA , GHANA #
t
20 - 20 -

12 18 24 6 12 LT 12 18 24 6 12 LT

60 60-

NATAL , BRAZIL
40 NATAL, BRAZIL
40 -

[
20 20 -
t

11 , -A I I 4
12 18 24 6 12 LT 18 24 6 12 LT

60

,o”\ HUANCAYO, PERU

,b
40 HUANCAYO, PERU :~’ ‘~ tt
\
,[ 5
\\
It $
20
t A .:
\
\
\
[L 18 24 6 12 LT
I
12
1
‘.
I
12 18 24 6 12 LT

Figure 10-71. Seasonal patiemsofwcumence ofscintillation activi~>6dB(S4 = 0.3) forve~quiet(Kp = O – l+)andfordisturbed (Kp = 3+ – 9)
magnetic conditions for May~t,

10.7.9.1 Effect of Magnetic Index on Midlatitude 10.7.10 The High Latitude Region
Scintillation. At latitudes below the auroral oval, various
sets of data have yielded behavior indicating little correlation Figure 10-72 depicts the intensity of scintillation at high
with magnetic conditions. Evans [1973] found no correla- latitudes in a very broad manner for the period of time
tion of their 400 MHz radar scintillations with magnetic around midnight.
index south of their shtion at 56° injariant latitude. Aarons
and Martin [1975] found that during the August 4–10, 1972 10.7.10.1 The Plasmapause and the Trough. The pres-
magnetic storms there was a negative correlation of scin- ent evidence is that there is a boundary at the high latitudes
tillation and magnetic index for Athens, Greece and Camp where weak irregularities’ commence. It is probably a few
Parks, California and little correlation at the 45° invariant degrees equatorwards of the plasmapause; between 45°–550
latitude intersection for Aberystwyth, Wales.. Bramley [1974] Corrected Geomagnetic Latitude (CGL), the system used at
found that except for the December 1971 magnetic storm high latitudes in this review.
(when the irregularity region probably encompassed the in- At night a trough or region of low electron density and
tersection point of -45°), there was no correlation between total electron content exists between the end of nornial ion-
magnetic activity and scintillations. ospheric plasma behavior and the auroral region where en-
This type of data essentially corroborates the early radio ergetic electron precipitatiori and current systems are dom-
star obvservations in the U.K. which found little correlation inant factors in producing both the normal ionospheric layers
with magnetic index except in paths to the north (with the and the irregularities.
exception of some intense magnetic storms). All observers of irregularities see a dramatic change in

10-79
CHAPTER 10

600 km

AU RORAL
/ OVAL

cORRECTEO GEOMAGNETIC
LATITUDE
JUL +

Figure 10-72. Depiction of high latitude irregularities -22-02 LT. Sheet-

like irregularities are seen in the aurora] oval, rod irregu-
larities at higher and lower latitudes.
AUG

sEP

OCT
[

II15
+
*

1
{( 1
II

( I
,,/ ~

I
:7

1
I

1 I
i
1
21 03 09 15 LT
CGM

irregularities in the auroral oval at the poleward edge of the MEAN SCINTILLATION INOEX (dB)
NARSSARSSUAII 1968-1974 KP=4-9
trough. In the auroral oval, the intensity of scintillations is
a function of local magnetic activity. Poleward of the aurora
there may again be a lowering of scintillation activity until Figure 10-74. Contours of monthly mean scintillation index in dB at 137
MHz as a function of local time for disturbed (Kp =
the observing path transits the polar region [Aarons et al., 4 – 9) magnetic conditions obtained at Narssarssuaq dur-
1981b] ing 1968– 1974.

10.7.10.2 Auroral Scintillations. From studies of radio

star and low altitude satellite scintillations, a series of height Perhaps the most consistent studies of long term behav-
measurements have pointed to F layer heights as the primary ior of scintillations have been made in the auroral zone, at
seat of the irregularities producing the signal fading. Alaskan longitudes and along the 70°W meridian.
Maximum irregularity intensity appears above the region Both the diurnal pattern of scintillation activity and the
showing maximum intensity aurora [Martin and Aarons, seasonal behavior as observed from one site can be noted
1977]. Vickrey et al. [1980] have shown that there is a in Figures 10-73 and 10-74. The data used for this long
collocation of scintillation patches in the auroral oval and term study [Basu and Aarons, 1980] were taken over a
F region ionization enhancements. period of 6 years from Narssarssuaq by observing 137 MHz
scintillations of the ATS-3 beacon; the propagation path
traversed the ionosphere at -63° CGL.
The long term study used for Figure 10-75 inco~rated
SUNRISE
+
data from three observatories (Narssarssuaq, G~nland; Goose
‘r Bay, Labrador; and Sagamore Hill, Massachusetts). The
*
contours are of reduced data for one season (May–July) for
+
magnetically active periods of time [Basu and Aarons, 1980].
*
The boundary of active scintillation is pushed equatorwards
.? extending into what was the quiet trough and plasmapause
latitudes. Thus during magnetic storms scintillations and

65- MAY THRu JULY

*
“c’- 15
(
09
,
15 L

MEAN SCINTILLATION INDEX (dB) MEAN S![60) 27 PTS KP~4-9

NARSSARSSUAQ 1966-1974 Kp= O-3

Figure 10-75. Variation of mean scintillation index during the northern

Figure 10-73. Contours of monthly mean scintillation index in dB at 137 solstice in dB at 137 MHz with local time and invariant
MHz as a function of local time for quiet conditions (Kp = latitude derived from hourly data at the 3 stations under
O – 3) obtained at Narssarssuaq during 1968-1974. disturbed magnetic conditions (Kp = 4 – 9).

10-80
 a

IONOSPHERIC RADIO WAVE PROPAGATION

.4 I I
GEOMETRICAL ENHANCEMENT FAeTOR FOR rms PHASE (8:8:1) I I I I I I

8 —

6 -
POKER FLAT ~
‘“i,= DIP
65

LAT - deg
70 75 SD 85

Figure 10-78, S4 at 50% exceedance level vs magnetic latitude for daytime

G data during 1977-1978.

4 —

ture—at least as observed from Alaska. Figure 10-78 illus-

trates the daytime increase with increasing latitude.
2 —

10.7.10.3 Polar Scintillations. A long term consistent

series of measurements has been taken at Thule, Greenland
with observations at 250 MHz IAarons et al., 1981 b]. The
OIP LAT – deg scintillations for this study ranged from very low values of
34 dB peak to peak on occasion during a period of low
sunspot number to saturation fading of 28 dB peak to peak
for hours during winter months of years of high sunspot
Figure 10-76. Model computations of phase geometrical enhancement number.
factor for sheet-like structures with an 8:8: 1 anisotropy. One set of measurements was taken between April and
Because of the meridional pass trajectory, the location of
October 1975. During this period of low solar activity, there
the enhancement is independent of the pass elevation.
[Rino and Owen 1980] was an absence of strong scintillation activity to such an
extent that only the occurrence of scintillation greater than
6 dB could be plotted. Figure 10-79 shows the contrast
optical aurora can be noted farther south than 55°. In the
between the 1975 period when solar flux was low ( 10.7 cm
70”W longitude region this extends below the latitude of
flux was -75) and the same months in 1979 when the solar
Boston.
flux was high (- 150–225).
Geometry and Enhancement. Sheet-like irregularities
A contour plot of the percent occurrence of scintillation
produce strong enhancements when observations are made
index greater than 10 dB is shown in Figure 10-80. The
in specific directions. For two sites in Alaska, Rino and
plot was developed from hourly average values of the 15
Owen [1980] have constructed the theoretical geometrical
minute S1 for each month for low magnetic activity (Kp = O-
enhancement factor for rms phase fluctuations for an 8:8:1
3). Two patterns emerge: (1) Maximum occurrence of ac-
irregularity (Figure 10-76) [Rino and Matthews, 1980]. They
tivity takes place in the months of little or no sunlight at F
found this enhancement in phase fluctuations as can be seen
by the data in Figure 10-77. The amplitude enhancement,
less dramatic but present, is also shown in Figure 10-77.
Daytime scintillation does not show the sheet-like struc-

“o
~
O 80

G 60
“>?050 929
, Om
888 7’4

/’
,/’—
4,8 040

0 ‘0

020
s, E
10
, .0 ~.. ,.- APR-OCT 1975
536 ------
--- /. ‘. /- -----=
----- ______ I
0 10 + I , ._...2*
329 ]
‘m
0 “4 8 12 ‘~ 16 24 UT
r
2;;0 ,200
, I I I I 1,
,d 55 60 ,5 ,. ,5 ,0 ,5

,1P LA, I,. ”E

Figure 10-79. Percentage occumence of scintillation greater than 6 dB for
low solar flux period April-October 1975 is contrasted with
Figure 10-77. RMS phase and S1 at 50% exceedance level vs magnetic that for high solar flux period April-October 1979 for troth
latitude for nighttime data during 1977-1978. quiet and disturbed magnetic conditions,

10-81
CHAPTER 10
PERCENT OCCURRENCE GREATER THAN 10d B_
of subvisual F layer (k = 6300 ~ 01) polar cap arcs. Kil-
r -7
ometer-size irregularities within the arcs produced intense
(saturated) amplitude scintillation at 250 MHz as the arcs
drifted through a satellite to ground ray path. Outside the
arcs, scintillation frequently persisted at a lower level (S1
-6 dB).
A pictorial representation of both the small scale anti-
sunward irregularity drift and the patch motion (predomi-
nantly dawn to dusk) is shown in Figure 10-81 (E. Weber,
private communication). Results point to two irregularity
components in the polar cap; antisunward drifting irregu-
larities which produce a background level of weak to mod-
erate scintillation and intense irregularities within F layer
polar cap arcs that produce more discrete (-1 h duration)
L, t
15.00N 21
CG 2400
–- .L..
03
t
;:~;nT 09
–L. c’“~u,

intense scintillation events as the arcs drift through the ray

LOCLL
K, O-3 THULE 1979
path.

Figure 10-80. Contour plot of diurnal pattern of monthly percent occur-

rence of scintillation greater than 10 dB for low magnetic
activity (Kp – O – 3). Observations were taken during 10.7.11 Empirical Model Of Global
Mar 1979-Feb 1980. Scintillation Behavior

region heights. Much lower scintillation occurrence takes

place in the sunlit months. (2) The diurnal variation is weak, 10.7.11.1 WBMOD. Over a period of yem, starting from
and apparent only during the winter months. available data and from weak scintillation theory, a model
Auroral arcs in the polar cap are approximately aligned of scintillation termed WBMOD has been developed [Fre-
with the noon-midnight magnetic meridian [Davis, 1962]. mouw and Bates, 1971; Fremouw and Lansinger, 1981;
These arcs generally drift in the dawn to dusk direction Fremouw and Rino, 1978, 1976; Fremouw, 1980; Fremouw
[Danielson, 1969]; however, reversals have been noted et al., 1977]. The program provides for phase and amplitude
[Akasofu, 1972; Weber and Buchau, 1981]. Recently We- information. Input user parameters include frequency, lo-
ber and Buchau 198 1] described the orientation and motion cation, local time, sunspot number, and planetary magnetic
index, Kp. The user also must specify the longest time the
system needs phase stability. Scintillation indices are the
output. A model of the irregularity drift velocit y is contained
/’7’0 ~ ,,,, in the program.
. Program WBMOD permits a user to specify his oper-
ating scenario. The code returns the spectral index p for
power-law phase scintillation, the spectral strength param-
eter T, the standard deviation u+ of phase, and the intensity
scintillation index, S4, as functions of a changing indepen-
dent variable chosen by the user.
The descriptive imegularity model is based on numerous
18
obwrvations [Fremouw and Bates, 1971; Fremouw and Rino,
1978], but most particularly on observations of phase scin-
tillation performed in the DNA Wideband Satellite Exper-
iment [Sagalyn et al., 1974]. The most significant caveat
about use of WBMOD, however, is that it has been cali-
brated quantitatively against Wideband data from only a
single station in the northern auroral zone (Poker Flat, Alaska).
The descriptive model was developed by iterative compar-
ison with most of the Wideband data population from Poker
Flat, with a portion of the population reserved for final
< w
comparative tests.
DAWN-DUSK ARC DRiFT
The basic calculations are made of two central quantities
T and p. T is the spectral strength of phase at a fluctuation
Figure 10-81. Schematic of small scale anti-sunward irregularity drift and
the patch motion. frequency of 1 Hz, p is the power-law spectral index of

10-82
 IONOSPHERIC RADIO WAVE PROPAGATION
phase; T is highly variable, unlike p. The program calculates more limited than WBMOD as ( 1) they are applicable only
T and p and the two commonly used indices of scintillation for the frequency of the data base, 137 MHz, (2) there is
activity based on them, one for phase, U+, and one for an equipment-biased limited excursion of the scintillations
intensity, SG. and, (3) these data have an implicit dependence on the
In order to calculate T, p, U+, and S4, one must have geometry of the observations, namely, observing ATS-3
values for eight parameters describing ionospheric irregu- from the stations detailed above. This does not permit other
larities. They are (1) the height, h; (2) vector drift velocity, viewing geometries or taking into consideration the config-
Vd, of the irregularities; (3) an outer scale, a; (4,5,6,7) four uration of the irregularities unless correcting factors are
“shape” parameters describing the imegularities’ three-di- included.
mensional configuration and spatial “sharpness”, a, b, b, With these caveats, the equations for each station are
and v; and (8) the height integrated strength of turbulence, Narssarssuaq (63° CGL intersection)
C,L. Program WBMOD contains models for the foregoing
eight parameters, but the degree of detail is very much less S1 (dB) = –6.4 + 9.2(1 – 0.2FD){1 + 0.23(1 – 0.3FD)
for some than for others.
The most variable and the most important of the eight X COS (HL + 2.0 + 0.34Kp) + 0.03
is the height-integrated strength of turbulence, C,L. The X COS [2(HL – 0.6)] + 0.02 COS [3(HL
irregularity strength is modeled by
+ 3 .0)]}2[0. 14KP(I+0. 12FD)+0.09As(I+ 1.76FD)]

@ = E(k.,k,,T,D,~) + M(k.,T) FD = COS (DA + 15.6) + 0.56 COS [2(DA – 22.4)]

+ H(Am,Tm,Kp>~) (10.68)
(10.69a)
where Am = geomagnetic invariant latitude,
Goose Bay (60° CGL intersection)
kg = geographic latitude,

T= local meridian time, SI(dB) = — 1.3 + 1.1(1 – 0.77FD){1 + 0.5(1 – 0.2FD)

D= day of the year, X COS(HL + 2.1 – 0.6Kp) + 0.06

R= smoothed Zurich sunspot number, x Cos [2(HL – 2. 1)] + 0.02 Cos 13(HL
+ 5,z)]}z[o.3KP(l+0. IFD)+O.8AS(I+ I 2FD)]
Tm = geomagnetic time,
FD = cos (DA + 0.5) + 0.2 COS [2(DA – 99)]
Kp = planetary geomagnetic activity index.
(10.69b)
The three terms in Equation (10.68) respectively describe
the strength of equatorial, midlatitude, and high-latitude Sagamore Hill (53° CGL intersection)
irregularities. The first two have not been tested extensively
against Wideband data but the high latitude term H has been. SI(dB) = 0.33 + 0.02(1 + 0.2FD) {1 + 1.2(1 – O.OIFD)
The high-latitude term is based on the observation that
x cos (HL – 0.4 – 0.15 Kp) + 0.3
there often is a more-or-less abrupt boundary [Aarons et al.,
1969] between the midlatitude region of relatively smooth X COS [2(HL – 0.8)] – 0.1 COS [3(HL
ionosphere and the highlatitude scintillation region. It is
+ 6. l)l}z[o.38Kd1+o.3~)+3.lAs(l -0,2FD)]
located, typically, equatorward of discrete-arc auroras in
the general vicinity of the diffuse auroral boundary. FD = COS (DA + 56) + 0.7 COS [2(DA – 143)].
(10.69c)
10.7.11.2 Formulas In Athmtic Sector. Since WBMOD
has been developed and calibrated against data from only DA is day number, As = S~/100, HL is local time (hours)
one longitude sector (Alaska), it is appropriate to note em- at subionosphenc point (350 km), and SF is solar flux at
pirical formulas that, though not as complex, have been 2695 MHz in solar flux units; all angles are in radians.
developed for another longitude sector, along the 70”W Arguments of the cosines with diurnal and yearly terms
meridian. These formulations have been made [Aarons et should be converted by factors of 2n124 and 2m/365, re-
al., 1980a] for Narssarssuaq, Greenland, Goose Bay, La- spectively.
brador and Sagamore Hill, Massachusetts based on 3–7 In Aarons et al. [1980a] corrections for frequency de-
years data base of 15-rein scintillation indices. The forcing pendence are given thus allowing higher frequency scintil-
functions are time of day, day of the year, magnetic index lations to be estimated. In addition, corrections for geometry
and solar flux. However, these individual models are much are also given similar to those cited in Section 10.7.3.1.

10-83
CHAPTER 10
10.8 IONOSPHERIC TIME DELAY
EFFECTS ON EARTH-SPACE
PROPAGATION

One of the most important effects of the ionosphere on

radio waves that traverse it is a retardation, or group delay,
on the modulation or information carried on the radio wave,
due to its encounter with the free, thermal electrons in the
earth’s ionosphere. Other effects the ionosphere has on radio
waves include (1) RF carrier phase advance (2) Doppler
shift of the RF carrier of the radio wave (3) Faraday rotation
of the plane of polarization of linearly polarized waves (4)
angular refraction or bending of the radio wave path as it
travels through the ionosphere (5) distortion of the waveform
of transmitted pulses, and (6) amplitude and phase scintil-
lation. With the exception of scintillation effects (see Sec-
tion 10. 7), all the other effects listed here are proportional,
at least to first order, to the total number of electrons en- FREQUENCY

countered by the wave on its passage through the ionosphere

or to their time rate of change. In fact, phase scintillation Figure 10-82. Time delay vs fiquency for various values of TEC.
also is merely the short term, time rate of change of total
electron content (TEC) after the longer term variations have
been removed. km! obviously, the TEC parameter is of potentially great
In this section a short description is given of each iono- importance to precision satellite ranging systems.
spheric TEC effect upon radio waves, along with a repre-
sentative value of the magnitude of each of the these effects 10.8.1.1 Two-Frequency Ionospheric Time ~lay
under normal ionospheric conditions. This is followed by a Corrections. If the navigation or ranging system band-
discussion of the important characteristics of average io- width is large enough so that two, fairly widely spaced bands
nospheric TEC behavior and the temporal and spatial vari- can be used for ranging, the ionospheric time delay error
ability of TEC. can be reduced to an acceptable level automatically and can
be made transparent to the system user. Because the io-
nospheric time delay is a function of frequency we can write:
10.8.1 Group Path Delay
Atl=Xx TEC, At2&= xTEC, (10.71)
The additional time delay, over the free space transit Cfi Cfi
time, of a signal transmitted from above the ionosphere to
a user on, or near, the earth’s surface is given by where Atl is the ionospheric error on frequency fl, and Atz
is the ionospheric error on the frequency f2. If the normal
40.3 system operational frequency is fl and we choose fz at a
At=? x TEC (S), (10.70)
lower frequency for ionospheric correction purposes, we

where the TEC is the total number of electrons along the obtain: 8(At) = ~ X TEC (1/fi – 1/6)
path from the transmitter to the receiver, c is the velocity
of light in rn/sec., and f is the system operating frequency = Atl (H – B)/ E (10.72)
in hertz. The TEC is generally expressed as the number of
electrons in a unit cross section column of one square meter or At, = ~/(fi – E) X 8(At)
area along this path.
A plot of time delay versus system operating frequency The value 8( At) is obtained from the difference of the si-
for TEC values from 1016 to 1019 el/m2 is given in Figure multaneous measurements of the total range, including ion-
10-82. These two values represent the extremes of observed ospheric time delay, at the two frequencies, f 1 and f2, since
TEC in the earth’s ionosphere. Note that, at a system op- the geometric distance is, of course, the same at all fre-
erating frequency of 1 GHZ, for example, a TEC of 1018, quencies. The quantity fi/ (H – C) is called the ionospheric
a value frequently exceeded in many parts of the world, scaling factor. For ratios of f2/f, near unity, the required
would produce a time delay of 134 ns or 40.2 m of range precision of the differential measurement may be unreason-
error. At a system operating frequency of 100 MHz this ably large. A plot of this quantity, normalized by f 1, is given
same TEC values would produce a range error of over 4 in Figure 10-83. In this derivation the contribution of re-

10-84
 IONOSPHERIC RADIO WAVE PROPAGATION
6- 1 assuming the satellite transmitted modulation phase at L 1
I and L2 is known and the receiving system frequency dis-
1
5 -
I
persive characteristics can be independently measured and
K I corrected for.
0
+
34 -
I For a typical daytime high solar activity TEC value of
I
u
z I
10’8 el/m2 column the 8( At) measured by a GPS receiver
i would be 35 ns or 10.5 m of ionospheric error. For a direct
53 - I
m I
measure of absolute TEC from the modulation phase delay
~
I
K at L2 minus L 1 we have
yz - I
n
m I
0
z
Q,
I TEC = 2.852 X 10’b X 8(At), (10.74)
I
I

, -L’ .~
I where 8( At) is measured in nanoseconds (ns). Since, at
oL
‘0,3 0.4 05 06 07 08 09 10 10.23 MHz, one complete cycle of modulation phase of
t
RATIO OF SECONOARY FREQUENCY fz I
y
PRIMARY 360° is 97.75 ns, we obtain TEC = 0.7745 X 10’6 el/m2
TO PRIMARY FREQUENCY
FREQUENCY
per degree of 10.23 MHz modulation phase difference, or:
TEC = 278.8 x 10’6 el/m2 per cycle of modulation phase
Figure 10-83. Ionospheric scaling factor vs ratio of primary (higher) to
secondary (lower) frequency.
difference. Thus, the cycle ambiguity in absolute values of
TEC is trivial to resolve using the GPS system as a means
of determining ionospheric time delay.
ceiver noise to the differential measurement accuracy has Absolute ionospheric time delay measurements can be
not been considered. made with an accuracy approaching 1 to 2 ns, depending
upon the received signal to noise ratio on both frequencies.
10.8.1.2 An Example of a Two-Frequency Ionospheric For the power levels transmitted by the ~PS satellites, an
Time Delay System. The Department of Defense is cur- omnidirectional receiving antenna, and a receiver with a
rently testing an advanced navigation system, called the modulation tracking bandwidth of approximately 15 Hz, the
NAVSTAR-Global Positioning System (GPS), [Demaro, differential modulation phase has been measured to within
198 1; Milliken and Zaller, 1978] which uses coherently approximately + 2 ns. The contribution of receiver noise
derived, identical modulation on two catier frequencies, for the two-frequency ionospheric time delay corrections on
called L 1 and L2, to measure the ionospheric group path the GPS system has been considered by Cretcher [1975].
delay directly and thereby correct for ionospheric time de-
lay. The ratio of frequencies used in the GPS system is
exactly 154/ 120, with the higher frequency (L 1) at 1575 10.8.2 RF Carrier Phase Advance
MHz. The two carrier frequencies transmitted by the GPS
system ‘are the 154th and 120th harmonics of 10.23 MHz. In addition to group path delay, or modulation time
This 10 MHz frequency is hi-phase modulated on both car- delay, over the free space delay, the phase of the carrier of
riers with a psuedo random code resulting in a [(sin x)/x ]2 radio frequency transmissions is changed by the ionosphere.
shaped spectrum of width 20 MHz to the first nulls. A user The RF phase is advanced with respect to its phase in the
with knowledge of the transmitted code collapses the re- absence of an ionosphere. This effect is extremely important
ceived spectrum to equivalent carriers with 10 MHz mod- in determining space object velocities by means of range
ulation. The 10 MHz modulation is transmitted with a known rate measurements. The amount of phase increase or phase
phase difference on the two carriers, and the received mod- path decrease can be expressed as
ulation phase difference is a direct measure of the ionos-
pheric group path delay. 1.34X 10-7
A+ = TEC (cycles), (10.75)
For the GPS carrier and modulation frequencies the ion- f
ospheric group path delay at frequency L 1, as obtained from
Equation 10.73 is where f is the system operating frequency in hertz, and TEC
is in el/m2 column. In practice, the amount of this phase
At, = – 1.5457 5(At), (10.73) advance cannot readily be measured on a single frequency
and two, coherently derived, frequencies are required for
Where 8( At) is the difference between the ionospheric time this measurement.
delay measured at the two frequencies. This difference in
range is directly related to absolute ionospheric time delay 10.8.2.1 Differential Carrier Phase. In addition to the
as, of course, the satellite is at the same range at both dual frequency identical modulation transmitted from the
frequencies. The only frequency dependent parameter in GPS satellites for ionospheric group path correction, these
range measurements is the ionospheric time delay effect, satellites also transmit two, coherently-derived carrier fre-

10-85
CHAPTER iO
quencies for ionospheric differential carrier phase measure- A24=(4u– @c)–(4c–4L)=@u+4L –2x4C
ments. For the pair of frequencies used by GPS, approxi-
mately 1.2 and 1.6 GHz, the differential carrier phase shift,
referenced to the lower frequency, is from Equation (10. 75)

A~=~ ~ ,~-, ~ (m;; 1)

A+=? x 10-7 x TEC (cycles)

x TEC (cycles), (10.76)

thus
where m = fl/f2 = 1.2833, A+ = 4.31 X 1~” X TEC
or 2.32 X 10‘b el/m2 per complete 2T cycle of differential 2.68 X 10-7 E
A24= TEC (cycles).
carrier phase between L 1 and L2, measured at L2. The f(F – %)
differential carrier phase [Equation (10.76)] is related to the
differential modulation phase, [Equation (10.72)] simply by When
the ratio of carrier to modulation frequencies. With a rea-
sonable carrier signal to noise ratio, this differential carrier F>>% (10.77)
phase can be measured to within a few degrees, or less than
approximately 0.04 x 10’6 el/m2. Since the TEC is gen- 2.68 X 10-7fi
erally much greater than 2.32 x 10’6, comesponding to 21r A,+ = TEC (cycles).
F
of differential carrier phase, there is a 2nm ambiguity in the
differential phase measurement. For a carrier frequency of 100 MHz a modulation fre-
The differential carrier method of measuring TEC can- quency of 1.93 MHz would be required to give 2T of second
not, in practice, be used to measure absolute values of TEC differential phase for a TEC value of 10’8 el/m.2 A value
by itself due to the large 2nm phase ambiguity in the meas- of A2@of 21r for 10’8 el/m2 is a reasonable compromise
urement, but this is not important for navigation systems between the requirement for minimizing chances of an am-
which require a correction only for range rate errors due to biguity in absolute TEC and accuracy in measuring TEC
the ionosphere between two measurement times. relative changes. The second difference of carrier phase has
The US Navy Navigation Satellite System, NNSS [Black, been used with the DNA-002 satellite to make estimates of
1980; Kouba, 1983], determines position for stationary and the absolute value of TEC [Freemouw et al., 1978].
slowly moving vehicles by measuring satellite transmitted
RF canier phase changes as a function of low-orbit satellite
motion across the sky. This method of positioning requires
only range rate information. The primary NNSS frequency 10.8.3 Doppler Shift
is 400 MHz. A second RF carrier at 150 MHz is used only
for ionospheric range rate corrections. While various tech- Since frequency is simply the time derivative of phase,
niques have been proposed for determining the absolute TEC an additional contribution to geometric Doppler shift results
from the differential carrier phase information received from due to changing TEC. This additional frequency shift is
the NNSS satellites, they all involve assumptions concern- generally small compared to the normal geometric Doppler
ing some a priori knowledge of the ionosphere, and they shift, but can be computed by
cannot be used in the general case.
As an ionospheric monitoring tool the combination of 1.34 X 10-7 d
Af=%= f ~ TEC (Hz). (10.78)
differential carrier phase and differential modulation phase dt
provides an excellent means of determining ionospheric
electron content along the ray path to the satellite. The For high orbit satellites where the diurnal changes in
absolute value of TEC can be determined by the group delay TEC are greater than geometric ones, an upper limit to the
technique and relative TEC changes can be measured with rate of change of TEC is a~xirnately 0.1 X 10’6el m-2s-’.
great accuracy by the differential carrier phase technique. This value yields an additiond frequency shift of less than
0.1 Hz at 1.6 GHz which would not be significant compared
10.8.2.2 Second Difference of Carrier Phase. The sec- with a typical required receiver loop bandwidth of at least
ond difference in phase between an RF carrier and that of a few hertz. At 400 MHz a similiar rate of change of TEC
its upper and lower sidebands can be used to measure ab- would produce a frequency shift of approximately 0.3 Hz,
solute values of TEC, as described by Bums and Fremouw probably still not significant.
[1970]. If three coherently derived frequencies, f – f~, f, During times of severe phase scintillation, which can
and f+ f~ are transmitted the second difference of phase is occur even at GHz frequencies, the TEC likely does not
given by change in a consistent, rapid manner to yield greater ion-

10-86
 J

IONOSPHERIC RADIO WAVE PROPAGATION

ospheric Doppler shifts, but the phase of the incoming RF
signal can have a large random fluctuation superimposed 1000
upon the changes associated with the normal rate of change zz
in TEC. This large, random component may actually spread a
out the spectrum of the received signal sufficiently to cause m
the receiver to lose phase lock, as the receiver signal phase
5
may have little energy remaining in the carrier, and instead 1=
* ,0
may be spread over several Hz, with little recognizable 0
w
carrier remaining. A knowledge of phase scintillation rates 2
c1
is required to determine the spread of received signal phase. 2
~

,1
10.8.4 Faraday Polarization Rotation

When a linearly polarized radio wave traverses the ion- .01

IOOMHZ 200 300400500 IGHz 2
osphere the wave undergoes rotation of the plane of polar-
FREQUENCY
ization. At frequencies of approximately 100 MHz and higher
the a~ount of this polarization rotation can be described by:
Figure 10-84. Faraday polarization rotation vs frequency for various val-
ues of TEC.
Q= 2“36 j 10-5 ~ B COS 6Ndl. (10.79a)

satellite transponder frequency band, the amount of Faraday

where the quantity inside the integral is the product of elec- rotation can be a tenth of a radian, well in excess of that
tron density times the longitudinal component of the earth’s required for dual, linear orthogonal channel separation.
magnetic field, integrated along the radio wave path. Many The Faraday rotation problem is overcome by the use
ionospheric workers have used this effect, named for Mi- of circular polarization of the comect sense at both the sat-
chael Faraday who first observed polarization changes in an ellite and at the user’s receiver. Generally the mobile user
optical experiment, to make measurements of the TEC of finds it difficult to utilize circular polarization due to the
the ionosphere. Since the longitudinal magnetic field inten- continual vehicle directional changes; thus he settles for a
sity changes much slower with height than the electron received linear polarization. The 3 dB loss between trans-
density of the ionosphere, the equation can be rewritten as mitted circular polarization and received linear polarization
is a necessary price to pay for user antenna maneuverability
and simplicity.
Q=~BLx TEC, (10.79b)

where BL = B cos 6 is taken at a mean ionospheric height, 10.8.5 Angular Refraction

usually near 400 km, K = 2.36 x l@5 and TEC is JNdl.
Typical values of polarization rotation for northern midla- The refractive index of the earth’s ionosphere is re-
titude stations viewing a geostationary satellite near their sponsible for the bending of radio waves from a straight
station meridian are given in Figure 10-84 as a function of line geometric path between satellite and ground. This an-
system frequency and total electron content. In fact, the gular refraction or bending produces an apparent higher
largest portion of TEC data available today from stations elevation angle than the geometric elevation. Millman and
throughout the world have come from Faraday rotation meas- Reinsmith [1974] have derived expressions relating the re-
urements from geostationary satellite VHF signals of op- fraction to the resultant angular bending. Perhaps the easiest
portunity. expressions to use, as given by Millman and Reinsmith
For satellite navigation and communication designers, [1974] relate the ionospheric range error to angular refrac-
however, the Faraday polarization rotation effect is a nuis- tion:
ance. If a linearly polarized wave is transmitted from a
satellite to an observer on or near the surface of the emh, *E = R + r. sinE. (r. cos E.) ~ ~
(10.80)
the amount of polarization rotation may be nearly an odd hi (*r. + hi) + d sin Eo R ‘
integral multiple of 90 degrees, thereby giving no signal on
the receiver’s linearly polarized antenna, unless the user is where EO is the apparent elevation angle, R is the apparent
careful to realign his antenna polarization for maximum range, AR is computed from AR = (40.3/P) X TEC, ro is
received signal. the earth’s radius, and hi is the height of the centroid of the
As shown in Figure 10.84 at 4 GHz, a commercial TEC distribution, generally between 300 and 400 km.

10-87
CHAPTER 10

—.
r
20
-lo18el/rn2
‘ECVERTICAL

15

10

I
5
200MHz, 2GHZ

0
o 5 10 15 20 25

ELEVATION ANGLE AT SURFACE

Figu~ 10-85. Refraction in elevation angle vs elevation angle for indicated frequencies and values of TEC.

For low elevation angles and satellites well above most 10.8.6 Distortion of Pulse Waveforms
of the ionization, R>rOsinEO, and the angular refraction can
be expressed as: Two characteristics of the ionosphere can produce dis-
tortion of pulses of RF energy propagated through it. The
cos EO dispersion, or differential time delay due to the normal ion-
AE = ~ AR. (10.81)
1 osphere, as derived by Millman [1965] is proportional to
l/~, and produces a difference in pulse tival time across
Typical values of elevation refraction error fora TEC a bandwidth Af of
of 10’8el/m2 column are shown in Figure 10-85 for several
frequencies. Note that, atthe lowest frequency, 100 MHz, 80.6 x 106
near the horizon the refraction is well over 1.5 degrees! The At = Af X TEC, (10.82)
c fs
curves shown in Figure 10.85 have been constructed using
the approximation derived by Millman and Reinsmith [1?74]
for low elevation angles given in Equation (10.81). where c is the velocity of light in m/s, f and Af are expressed
Generally, the range error itself is the main ionospheric in Hertz, and TEC is in el/m2 column. The dispersive term
problem for advanced navigation systems, and elevation for pulse distortion is thus proportional to TEC. When the
angle errors are insignificant. Satellite detection radar sys- difference in group delay time across the bandwidth of the
tems, on the other hand, do have the requirement to know pulse is the same magnitude as the width of the pulse it will
accurate pointing elevation angles for their large aperture be significantly disturbed by the ionosphere. Millman and
arrays, though generally the accurate tracking is done by Bell [197 1] also derived mathematical relationships for ion-
using range rate information, and elevation angle is of sec- ospheric dispersive effects on an FM Gaussian shaped pulse.
ondary importance as long as the beamwidth of the antenna In addition to pulse distortion by the dispersive effects
is large enough to see the target. due to the TEC of the normal background ionosphere, radio
Errors in the azimuth of radio waves transmitted through pulses are also modified by scattering from ionospheric ir-
the ionosphere can also occur; they depend upon azimuthal regularities. Yeh and Liu [ 1979] have computed pulse mean
gradients in TEC which are generally small and which can arrival time and mean pulsewidth due to both dispersion
usually be neglected in practical cases. and scattering.

10-88
 a
IONOSPHERIC RADIO WAVE PROPAGATION

I700 1700

1600 1600

1500 1500
L
1400 BOULDER, COLORADO
1
MARCH 1980
1300- -
(F ,.,7= 167)
1200- -

I 100” -

1000 ““

~ goo. .

+
z 800 .-
0
m 700 --
1

600 --

500 -- 500. -

400 -- 400- -

300 ‘- 300- “

200

100
t
.2 .4 .6 .8 1.0 1.2 1.4 1.6
104 105 106

el /cm3 el/cm3 units of 106

Figure 10-86. Typical profile ofelectron density vs height. In(a) log N,isplo~ed; in(b) N,isplotted ona1inearscaie.

10.9 IONOSPHERIC TOTAL ELECTRON ionosondes have been used since the 1930s to make con-
CONTENT (TEC) tinuous, routine measurements of the density at the peak of
the F2 region, measured by ionosondes as foF2, and equated
to N~.X by
10.9.1 Average TEC Behavior
(foF2)2 = 80.6 N, (10.83)
The ionospheric parameter responsible for the effects
described in section 10.8 is the total number of free elec- where foF2 is in MHz, and N is in units of 10b cl/cc.
trons, TEC, or its rate of change, along the path from a In the 1950s and 1960s, continuing to a more limited
satellite to a ground station. The greatest contribution to extent even today, upwards of 150 ionosondes were operated
TEC comes from the F2 region of the ionosphere. A typical to provide improved prediction capability for long distance
daytime midlatitude, high solar maximum electron density high frequency propagation by means of ionospheric re-
profile is illustrated in Figure 10-86. The curve on the left fraction. Various models of foF2 were developed for this
side of Figure 10-86 is the log of N, plotted versus height purpose, one of the more popular ones being commonly
as normally shown by ionospheric workers. Since the TEC known as ITS-78 [Barghausen et al., 19691 after the report
is represented by the area under the curve of a linear plot number which described the model. This model, among
of N, versus height, the right hand plot of Figure 10-86 other things, characterized the 10 day average worldwide
illustrates the actual linear plot. Note that most of the con- behavior of foF2 by Fourier temporal components and Le-
tribution to TEC occurs near the peak of the F2 region. The gendre polynomial geographic coefficients ordered by mag-
reason for making this point is as follows: ground-based netic, rather than geographic, latitude. The success of this

10-89
CHAPTER 10
experimental, data based, or empirical model, in repre- puting TEC. A representation of world-wide average be-
senting the actual worldwide foF2 is due to the large amount havior of TEC is illustrated in Figure 10-87 for 2000 hours
of data available from ionosondes in many regions of the UT. To first order the TEC contours shown in Figure 10-
world. Other characteristics of this model are discussed by 87 move westward along magnetic, rather than geographic,
Dandekar [ 1982] and in Section 10.3 of this chapter. latitude lines, at the earth’s rotation rate. The Bent model
For the TEC parameter, data availability have been, and was constructed using solar maximum data from the 1968-
will likely continue to be, much more sparse. First, TEC 1969 period and had to be adjusted upward somewhat to
measurements have generally been calculated from mea- account for the much higher 1979-1980 solar maximum than
surements of Faraday polarization rotation using VHF sig- that of 1968-1969. This adjustment was necessary to ade-
nals of opportunity transmitted from geostationary satellite quately represent the actual TEC values from stations mak-
telemetry transmitters. A few lunar reflected Faraday ro- ing observations in March 1980, which was near the max-
tation measurements in the late 1950s and early 1960s and imum of the second highest solar cycle ever recorded in the
the TEC obtained from a few low orbit satellites did not more than 200 year history of solar cycle observations. The
contribute significantly to our knowledge of world-wide TEC Bent model, appropriately adjusted for high solar cycle val-
behavior, at least not for modeling average ionospheric con- ues, does however, represent fairly well the average be-
ditions. Only since the early to mid- 1960s have TEC values havior of TEC for many locations tested. Other worldwide
been obtained on a more-or-less regular basis. Even today ionospheric electron density profile models from which av-
fewer than one dozen stations regularly contribute TEC data, erage TEC can be obtained include ones by Ching and Chiu
which can be used in TEC modeling purposes, to a world [1973], and Chiu [ 1975], Kohnlein [ 1978], the 4-D model
data center. [VonFlotow, 1978], and the International Reference Iono-
Fortunately, most of the contribution to TEC comes from sphere (IRI)[Rawer, 1981]. The characteristics of some of
near the F2 region density peak where models of foF2 are these models are described in Dandekar [1982] and in Sec-
available. These foF2 models can be combined with some tion 10.3.
limited knowledge of topside ionospheric thickness obtained Other empirical models of TEC have been developed
from topside sounders, along with topside in situ density directly from TEC data alone, though these have necessarily
measurements, to produce a complete ionospheric height been limited in temporal and geographic extent by the avail-
profile model. The most well known of these models is the able data base. These include models of TEC over Europe
one by Bent [Llewellyn and Bent, 1973] which uses ITS- and the Mediterranean [Klobuchar, 1973] for low and me-
78 coefficients for foF2, and topside exponential for com- dium activity portions of the 11 year solar cycle, and a

60

40

20

LATITUOE 0

-20

.
o~ ,:0 I
150
I
180
I
150
I
120
I
90
I
60
I
30
I
0

LONGITUOE

CONTOURS OF TOTAL ELECTRON CONTENT UNITS OF 10’6 el /m2 COLUMN

MARCH 1980 ( F,O,, = 167)
2000 UT

Figure 10-87. Contoursof vertical TEC in units of 10 16~l/mZ column for 2~ UT, Mwch 1980.

10-90
 a
IONOSPHERIC RADIO WAVE PROPAGATION
model of TEC over the Indian subcontinent for both solar of curves of diurnal changes in TEC for a northern mid-
minimum and for an average solar maximum IKlobuchar, latitude station for twelve months during a solar maximum
et al. 1977]. Models of the slab thickness parameter, the period is shown in Figure 10-88. The standard deviation
ratio of TEC/N~a, have been developed for specific regions from monthly mean diurnal behavior is approximately 20~o-
such as the one for northern Europe by Kersley [1980], and 25%, during the daytime hours when the absolute TEC
one for the eastern USA by Klobuchar and Allen [1970], values are greatest. Figure 10-89 shows the standard de-
from which TEC can be obtained from a model of foF2. viation from monthly average TEC behavior for the mid-
An algorithm designed for an approximate 50% correction day hours for a number of stations during the solar maximum
to world-wide TEC, for use in an advanced navigation sys- period 1968-1969. Again 20%-25% is a good value for the
tem, has been developed by Klobuchar [1975]. standard deviation from the monthly average behavior. The
All of the models listed here, and the list is by no means standard deviation is somewhat higher during the nighttime
complete, are empirical models that attempt to correct for hours, but the absolute TEC values are much lower during
average TEC behavior only. However, the variability from these periods.
average TEC behavior can be large and may be important If a satellite ranging system has error requirements such
to some radio wave systems that must propagate through that it must correct for monthly average ionospheric time
the ionosphere. delay, but still can tolerate the approximate 20%-2570 vari-
ability of TEC from monthly average conditions, approxi-
mately 70%-8090 of the ionospheric effect on the system
10.9.2 Temporal Variability of TEC can be eliminated by the use of an average TEC model such
as the one constructed by Bent ILlewellyn and Bent, 1973].
10.9.2.1 Variability from Monthly Mean TEC Values. If the system only requires an approximate 5070 rms cor-
The ionosphere is a weakly ionized plasma and the resultant rection of the ionospheric time delay, the algorithm devel-
TEC is a function of many variables including solar ionizing oped by Klobuchar [1975] can be used. On the other hand,
radiation, neutral wind and electric field effects, neutral if corrections for some portion of the remainder of the ion-
composition, and temperature changes. A monthly overplot ospheric time delay are required, after a state of the art TEC

. FEBRMY

+ JWE

/“

b —.-. + —4

12 SEPTEMSER + OCTOSER
4 ~
1 Od

80 L

t
J

‘,’”’’’’’ ,.’ d4+’d4

TOTM EWIVALENT VERTICAL ELECTRON ~TENT

F~ HAMILTON (ATS- 5)

(TIME UT)

Figure 10-88 Monthlyoverplotsof TEC diurnalcurves for Hamilton,Mass. for 197’9.

10-91
CHAPTER 10
80

70
. EDMONTON 1969

60 . RBERYSTWYTH 1969
x HFIMILTON 1969

50 . STRNFORO 1969
, HONOLULU 1969
1-
: 40 HONG KONG 1968

0
m 30

20

10

1 ,
0 , , , 1 , , &
i

JRN FEB MRR RPR MFIY JUN JUL RUG SEP OCT N13V OEC

Figure 10-89. Percentage standard deviations for daytime TEC from the stations indicated.

model, such as the Bent one, has been used to takeout the direction. Their results are shown in Figures 10-90a and b
monthly mean TEC, then the shortterm (a few hours) tern- for the east-west and the north-south station alignments,
poral variability as well as the geographic variability, of respectively. No significant difference in correlation dis-
TECmust reconsidered. tance was found with season.
The percent improvement, P. I., in TEC from the average
value is related to the correlation coefficient rby
10.9.2.2 Short Term Temporal Variability of TEC.
The correlation time of departures of TEC from monthly
P.I. = 100 x [l–(l– P)O’5]
average curves has been studied by Donatelli and Allen
[1981], They concluded, for the midlatitude station they (see Gautier and Zacharisen, [1965]) (10.84)
studied, that the useful prediction time was a function of
localtime, season ,andlongtermsunspot activity. However, Note that a correlation coefficient of 0.7 explains only 29%
in most cases they gained no significant improvement over of the variance between the data at station pairs; hence a
the use of monthly mean predictions when they used actual measurement at one station, with a correlation coefficient
data more than 3 hours old. The longest useful prediction of 0.7 between data sets with a second station would result
interval occurred, fortunately, during solar maximum day- in an improvement at the second station over the average
time hours when absolute TEC values are highest. During predicted value of only 29%.
solar minimum periods their useful prediction time interval
was often as short as one hour.
10.9.3 TEC in the Near-Equatorial Region
In their study, Donatelli and Allen [ 198 1] predicted TEC
data for the same geographic location and direction in the
All of the preceding sections have concentrated on the
skyas their measurements. Iftheprediction is foradifferent
behavior of TEC in the midlatitude regions of the world,
location, the temporal correlation will be lower.
mainly because most of the available data are from that
region. The near-equatorial region deserves special mention
10.9.2.3 Geographic Variability of TEC. Thevariabil- due to the fact that the highest TEC values in the world
ity of TEC at the same local time, but as a function of occur in this region, as shown in Figure 10-87. This region
distance has been studied by Klobuchar and Johanson [1977]. extends to approximately ~ 20°-250 either side of the mag-
They utilized TEC data from two sets of stations, one aligned netic equator, with the highest TEC values not at the equa-
approximately along an east-west direction, with the other tor, but rather at the so called “equatorial anomaly” regions
set of stations aligned along an approximate north-south lccated at approximately ~ 15° from the magnetic equator.

10-92
 a
IONOSPHERIC RADIO WAVE PROPAGATION

1.0

.8
—— ——
[
.6

.4
[ SEASONAL MEAN DAYTIME ‘\s

I CORRELATION COEFFICIENT

1 I I 1
01 1

0 I000 2000 3000 4000 5000

IONOSPHERIC DISTANCE (km)

LATITUDE SEPARATION =70” W LONGITUDE

1,0 >-
o
a
z
a
m
.8 a
3

w
,6 s
r
SEASONAL MEAN DAYTIME
CORRELATION COEFFICIENT
.4
E

.2 -

1 I I I 1
0-
0 I000 2000 3000 4000 5000
IONOSPHERIC DISTANCE (km)
LONGITUDE SEPARATION AT NORTHERN MID-LATITUDES

Figure 10-90. Comelation c~fficient vsstation wpation in(a) latimde red(b) longitude.

10-93
CHAPTER 10
)60+ .
;4cf MAY 19S0 JUNE JULY AWUST

)6 +
SEPTEMBER t OCTOBER WVEMBER
L
I 40

12 1 /---

100

80

6L

‘E 40
*Z
Q 20

&

g
~ 200

1 80 T JANUARY 1%1 + FEBRuARY . MARCH .

I
APRIL f

60 J ,

40 1

I 2: t

120

do

6.

A,7

).

-.. J...2J. <.+..+ ., . .,, . . ...3.

..-3 ‘“:7-’”---”’” ‘r”?! -’!’ ”~””;’’’.’ ”z””’’z”’:””-’

TOTM EWIVALENT vERTW ELECTRON ~TENT

FROM ASCENSION ISLANO (SIRIOI

TIME IUT)

Figure 10-91. Monthly overplots of TEC diurnal curves for Ascension Island, May 1980-April 1981.

The regions of highest TEC values at 2000 hours UT are 10.9.4 TEC in the AuroraI and
clearly seen in Figure 10-87 near 100° west longitude. Polar Cap Regions
Most of the day-to-day geographic variability of TEC
in the equatorial anomaly region during solar minimum con- Since most available TEC values have been measured
ditions can be explained by the variability of equatorial using radio signals transmitted from geostationary satellites,
electrojet strength. Unfortunately, no such similar TEC data which can be viewed only at low elevation angles from high
are available for solar maximum. latitudes, knowledge of the variability of TEC in the auroral
An example of the high temporal variability of TEC for and polar cap regions is sparse. In the American longitude
solar maximum conditions for Ascension Island, a station sector, where the magnetic latitudes are lowest for a given
located near the peak of the southern TEC equatorial crest geographic latitude, there is considerable TEC data from
region, is shown in Figure 10-91. Note the extremely large Goose Bay, Labrador, over which the aurora passes south-
day-to-day TEC variability in the afternoon and evening ward, even during moderately magnetically disturbed pe-
hours in some months. Any satellite ranging system re- riods. The behavior of TEC during those periods can be
quiring ionospheric TEC corrections in the near-equatorial highly irregular, especially during the nighttime hours. TEC
region should not use the midlatitude standard deviation values often exhibit rapid changes and occasionally even
values of approximate] y 20%-25 Yo to represent the vari- e~ceed the daytime maximum values briefly. While the oc-
ability of the near-equatorial region. currence of general auroral activity may be predictable, the

10-94
 i

IONOSPHERIC RADIO WAVE PROPAGATION

JULY AUGUST SEPTEMBER OCTOBER

67 3
I I
xx XxXxxxxxX%xx
x 2
x
xxx x
I
L
l—-l~Lo
xxx KIRUNA, SWEDEN

ID NOVEMBER FEBRUARY
6 DECEMBER JANUARY 3
-o

N
x

\
E

al
4
1--&d *--\
----
,--

<
---
--a.a

G---”--”
~. -.. . .-.

‘--~
,, ,L[ o
1 1 I I i
;

ik;@Wr’?-:-: 18 24

LOCAL TIME
6 1’ 18 24
XXXSAO PAULO, BRAZIL
6 1’ 18 24

Figure 10-92 Monthly average plasmaspheric electron content vs. local time for Akr’ystwyth, Wales (dashed line) and for Hamilton, Mass. (solid line).
Also plotted arevalues from Kimna, Sweden for Octokr 1975 mdfrom Sao Paulo, Brazil for May 1975.

specific large increases in TEC, likely due to auroral pre- Davies [1980] has reviewed the overall results of the ATS-
cipitation, are not individually predictable, but maybe sta- 6 experiment. A summary of typical protonospheric electron
tistically characterized as a function of magnetic activity. content data is shown in Figure 10-92 taken from Klobuchar
In the polar cap region a negligible amount of TEC data et. al. [1978]. Note that the protonospheric values are fairly
exists. The absolute TEC values are probably lower in this low in absolute value.
region than in the midlatitudes, and the variability of the During the nighttime hours when the ionospheric TEC
polar cap TEC is probably very high. is low, the protonospheric contribution may become a fairly
large percentage of the total number of electrons between
a satellite at geostationary height and an observer on, or
10.9.5 Protonospheric Electron Content near the earth’s surface. Unfortunate y, no protonospheric
electron content data are available during solar maximum
Most of the available TEC data has been taken by meas- conditions.
urements of Faraday rotation of single frequency tilo waves
transmitted from geostationary satellites to ground observ-
ers. The electron content obtained from Faraday rotation 10.9.6 Short Term Variations in TEC
observations, while made from radio waves transmitted from
satellites at geostationary satellite height above the earth’s The time rate of change of TEC, in addition to the normal
surface, only includes the contribution of electrons up to diurnal variations, also has periodic variations due to per-
heights of approximately 2000 km. This is because the in- turbations of the ionospheric F region from various potential
tegrated product of the longitudinal component of the earth’s sources as geomagnetic substorms, meteorological sources
magnetic field times the electron density, above approxi- such as weather fronts, shock waves from supersonic air-
mately 2000 km, is negligible. The only measurements of craft, volcanic explosions, rocket launches, and other mis-
the additional contribution of electrons above the Faraday cellaneous sources. While these short term variations in TEC
maximum height have been made using signals from an cover a large range of periods and amplitudes, common
ionospheric beacon on the geostationary satellite ATS-6. periods range from 20 to over 100 minutes with amplitudes

10-95
CHAPTER 10
of a few percent of the background TEC. A 10% ionospheric Experiments [Russell and Rycroft, 1980] dealt with a broad
disturbance with respect to the background TEC is uncom- spectrum of experiments: energetic ptiicle injections, plasma
mon, while a 1Yo TEC perturbation is common. Titheridge wave (VLF) injections, mass (neutral gas) injections, as well
[1968] and Yeh [1972] have made studies of the statistics as with laboratory and computer simulation experiments.
of traveling ionospheric disturbances (TIDs), in TEC for The most recent summary of Active Ex&riments in space
midlatitude regions. treated particle beams, neutral gas injections, wave injec-
A system that requires correction for the rate of change tions and high power heating experiments [Burke, 1983].
of TEC cannot rely Ori models of TEC to provide reliable The common thread that binds all of these methods is the
infofiation on short term rate of change of TEC informa- use of well-defined input/output experiments to probe the
tion, and can use available TID information only in a sta- system response @nctions for specific atmospheric and space
tistical manner. The only recourse for a system significantly plasma systems.
affected by rate of change of TEC is to use a dual frequency In terms of purely ionospheric phenomena, the modi-
measurement technique to directly measure the ionospheric fication of ambient eleetins and ions ~ most often achieved
contribution to range rate. by chemical injections or by radiowave heating experiments.
Each of these areas is treated in the following sections.

10.9.7 Conclusions
10.10.1 Chemical Releases
There are at least three categories of systems potentially
affected by ionospheric time delay. For the first category The history of chemical release experiments dates from
of user the potential systems effects may be small, at least the the earliest days of space exploration when, shortly after
under any naturally occurring worst case ionospheric con- Sputnik- 1 in 1957, rocket-borne payloads of highly reactive
ditions. In the second category, a user may require a nominal chemicals were injected into the upper atmosphere in at-
correction for average ionospheric time delay, but is able tempts to use artificial perturbation techniques as a way of
to tolerate the 20 Y0-25Y0 standard deviation from average investigating the structure and dynamics of the neucal and
conditions. He should expect at least a 5070 correction for ionized components of the upper atmosphere. The Air Force
ionospheric time delay effects using a relatively simple time Cambridge Research Laboratories (ARCRL) carried out the
delay algorithm, and up to 70%-80% for a state of the art, initial work with plasma cloud injections [Marmo et al.,
fairly complex model. These model corrections can be im- 1959], and later a pioneering and comprehensive series of
proved by the use of actual ionospheric measurements within chemical injection experiments under PROJECT FIREFLY
a reasonable temporal and spatial frame. For the third cat- [Rosenberg, 1964]. Experiments using barium releases (or
egory of user ionospheric model corrections, even updated similar, easily ionized species) have formed the major ac-
with near-real-time measurements, may not be sufficient to tivity in this field, tracing andlor modifying ionospheric
correct for ionospheric time delay, and the system must then processes from auroral locations [Holmgren et al., 1980] to
make. its own ionospheric correction. Fortunately, the ion- the equator [Kelly et al. 1979]. The symposium proceedings
osphere is a dispersive medium and the use of identical referenced above [Alb~ht, 1976; Russell and Rycroft, 1980;
modulation on two, widely-spaced frequencies will allow a Burke, 1983] offer comprehensive summaries of these ex-
direct measurement to be made of ionospheric range delay. periments.
Two coherently-derived carrier frequencies may be used to The field of neutral mass injections was first concerned
obtain accurate time rate of change information for TEC. with the environmental impacts that might result from the
Details of measuring ionospheric effects directly by a sys- larger and more powerful rockets being developed for space
tem’s use of multiple frequencies are available in Bums and exploration [Kellogg, 1964]. In 1973, when the last Saturn
Fremouw [1970]. V rocket to be used in the U.S. Space Program launched
NASA’s Skylab Workshop, the resultant deposition of ap-
proximately 1000 kg/s of Ha and H20 exhaust molecules
10.10 ARTIFICIAL MODIFICATION into the 200-440 km altitude region initiated a rapid and
large-scale depletion of the ionosphere to an extent never
The field of ionospheric modifications is a subset of a seen before (see Figure 10-93). The artificially-created “ion-
more general class of research today called “Active Exper- ospheric hole” amounted to nearly a 50% decrease in the
iments” in space plasmas. This field was initiated early in total electron content (TEC) of the ionosphere over an area
the space program by using rocketbome chemical releases of approximately a million square kilometers. Mendillo et
as tracers and/or modifiers of upper atmospheric processes. al. [ 1975 a,b] attributed the effect to the explosive expansion
The physical basis for such experiments was reviewed in of an exhaust cloud of highly reactive molecules that ini-
some detail by Haerendel [1976] during the first intern- tiated a rapid recombination of the ionospheric plasma.
ationalmeeting devoted entirely to artificial modification studies The introduction of such typical rocket exhaust products
[Albrecht, 1976]. A second major symposium on Active as Hz, H20 and C02 into the upper atmosphere causes the

10-96
 J
IONOSPHERIC RADIO WAVE PROPAGATION
ATs-3 observations FROM SAGAMORE HILL, HAMILTON, MASS the lack of large-scale/long-lived modification effects upon
“E 20- the lower regions of the ionosphere is due primarily to the
: SKYLAB high neutral densities and molecular ion chemistry already
G LAuNCH
. 16 -
dominant at D and E region heights, as discussed in detail
-0
by Forbes [ 1980].
I
t-
Z 12 -
Computer simulation models for the F region effects
/“
E have been constructed by Bemhardt et al. [ 1975], Mendillo
z
o
0 and Forbes [ 1978], Anderson and Bemhardt [ 1978] and Zinn
z 8 -
o and Sutherland 11980]. The emphasis in these studies has
a \
b \
ranged from environmental impacts of proposed in-space
24
A
w construction scenarios [Rote, 19791, to laboratory-in-space

L~
A
$ experiments using “dedicated engine-bums” of the space
p
00 03 06 09 12 15 18 21 24 EST shuttle as part of the Spacelab-2 mission in 1985, to a series
14 MAY 1973 of chemical modification experiments planned for the Com-
bined Release and Radiation Effects Satellite (CRRES)
Figure 10-93 Total Electron Content (TEC) data used to detect the “SKY- scheduled for the late 1980’s. Some of these concepts have
LAB effect” on 14 May 1973. The dashed curve gives the been tested using rocketbome chemical payloads during pro-
anticipated diurnal TEC behavior based upon a monthly
median prediction updated for geomagnetic storm effects jects LAGOPEDO [Pongratz and Smith, 1978], WATER-
[Mendillo et al., 1975b]. HOLE [Whalen et al., 1981], BIME [Narcisi, 1983] and
COLOURED BUBBLES [Haerendel et al., !983]. In the
atomic ion F region plasma to be transformed to a molecular AFGL Ionospheric Modification Study [Narcisi, 1983], at-
ion plasma at rates 100 to 1000 times faster than occur with tempts were made to study effects associated with SFC in-
the naturally present molecules of nitrogen (Nz) and oxygen duced negative ion plasmas [Mendillo and Forbes, 1978].
(o,). During so-called “experiments of opportunity,” where
These important reactions are scheduled rocket launches are monitored by a variety of
techniques, satellite radio beacon observations have been
reported by Mendillo, et al. [ 1980] incoherent scatter meas-
0+ +HZO:HZO+ +0 K, = 2.4 X 10-9 C~3/S urements by Wand and Mendillo [ 1984], and optical di-
agnostics by Kofsky [ 1981] and Mendillo and Baumgardner
(10.85)
[ 1982]. Figure 10-95 offers an example of the artificial
K2
0+ +HZ -OH+ +H KZ = 2.0 x 10-9 cm~/s airglow clouds associated with F region hole-making ex-
periments.
(10.86)

0+ ,+ C02: 0; + co K3 = 1.2 x 10–9 cm3/s .

(10.87) 10.10.2 High Power HF Transmissions
Once a molecular ion is formed, its dissociative recom- Ground based high power high frequency transmitters
bination with an ambient electron occurs rapidly, operating below the critical frequency of the ionosphere
have been used to artificially modify the ionospheric electron
H20+ + e- ~ OH + H a, = 3.0 X 10-7 cm3/s themtal budget and plasma characteristics [Utlaut, 1970;
(10.88) Gordon et al., 1971; Shlyger, 1974; for comprehensive re-
views, see Carlson and Duncan, 1977 and Gurevich and
oH++e-~H+o CY, = 1.0 X 10-7 ems/s Fejer, 1979]. The power aperture product of these high
(10.89) power transmitters have been typically of the order of 104
Mwm2 providing power densities of about 10-100 pwm-2
at ionospheric heights. The ionospheric “modification”
OJ+e-; C)+O 0!, = 2.0 X 10-7 cm3/s,
or so-called “heating” experiments have been observed to
(10.90) cause not only the initially intended enhancements of elec-
tron gas temperature with associated plasma redistribution
and hence an “ionospheric hole” is formed. A review of but give rise to a variety of nonlinear plasma phenomena.
rocket induced ionospheric disturbances has been given by Figure 10-96 [after Carlson and Duncan, 1977] summarizes,
Mendillo [1981]. in a schematic form, the srnking variety of observed ef-
Figure 10-94 contains a schematic showing the many fects of ionospheric heating. The enhancements of electron
physical and chemical processes associated with artificially- gas temperature have been observed to be a few hundred
induced depletions in the F region. It should be noted that degrees K [Gordon et al., 1971] caused by the deviative

10-97
CHAPTER 10

SPACELAB-2 PLASMA DEPLETION EXPERIMENTS

I 1

I I I 1
t
I

m
I 1
h
i TRGGERING of PLASMh FL~
KWSTIC lfQUATC91AL1 INTO DfPLET[O
MAVITY PLAW REGION
wAv[S INSTA81LIT!ES

THEORY u
Lu,,b
(WO1ONOSPHER!C)
DIPLf TlON

n CW)UG&Tf
POINT
DfPl iTIONS

IN - SITU
\ PROBES

wASMA

“a
HOLES

RASMA
@o@ UAGNOSTICS
o ~CKAGE

*
)
)) \
\ \

&;/’’’’”& “..9$
lNeoHERENT
SCATTER & A

DIAGNOSTICS
GROUND-BASED -- AIRBORNE -- IN-SPACE
Figure 10-94. Schematic summary of possible rocket effluent effects upon the upper atmosphere (h >200 km), associated with the NASA Space lab-2
mission scheduled for Spring–Summer 1985.

absorption of the heater wave near the altitude ofHF re- the high-power HF transmissions; and (5) strongly enhanced
flection. The following manifestations of plasma instabilities airglow at 6300 ~; some enhancement at 5577 ~ is also
have been observed: (1) artificially created spread F; (2) observed (see the Special Issue of Radio Science, [1974]).
strongly enhanced radio wave absorption; (3) the creation The short-wavelength ( 1 cm -10 m) field-aligned irreg-
of field-aligned density irregularities which scatter (this phe- ularities produced by ionospheric heating are a result of
nomenon is sometimes called field-aligned scatter of FAS) parametric decay instability or wave interaction between the
an incident HF, VHF, or UHF wave with virtually no fre- high power radio wave (pump) and the ion-acoustic and
quency change and make certain types of scatter commu- Langmuir waves. This was predicted from theory [Perkins
nication circuits possible; (4) scattering process in which and Kaw, 1971; Perkins et al., 1974; DuBois and Goldman,
the frequency of the scattered wave differs from the fre- 1972] and experimentally confirmed at Arecibo [Carlson et
quency of the incident wave by roughly the frequency of al., 1972].

10-98
 P
w
IONOSPHERIC RADIO WAVE PROPAGATION

000
is now attributed to either a thermal self-focusing mecha-
nism [Perkins and Valeo, 1974; Theme and Perkins, 1974]
or the alternative mechanisms of stimulated Brillouin scat-
tering [Cragin and Fejer, 1974] and stimulated diffusion
scattering [Goldman, 1974].

0000
A-2 :57:37 ❑-2,B:05 C-2.9 36
The above range of irregularity scale sizes has sufficient
power spectral intensity to cause scintillation of radio signals
received from radio stars and artificial satellites. This was
demonstrated when VHF/UHF signals transmitted through
the artificially heated ionospheric F region were found to
>2:5WB E.259 :35 F~rnV5 G..9~:42
exhibit scintillations [Rufenach, 1973; Pope and Fritz, 1974;
Bowhill, 1974]. Radio star scintillation measurements at 26
Figure 10-95. The growth of an ionospheric hole is shown in this sequence MHz during ionospheric modification indicated the presence
of image-intensified, wide-angle photographs of the ex- of either rapid and random or deep iong-period (-5 reins)
panding 6300 ~ airglow cloud produced by excited oxygen
atoms created from the recombination of free electrons and fluctuations. In order to avoid some of the difficulties of
molecular ions (O2 +, OH+, HZO +) produced by exhaust radio star observations, Bowhill {1974] performed scintil-
molecules (C02. H2, HzO) and ambient atomic ions (O+).
lation measurements with both geostationaiy and orbiting
Times are a.m., PST [Mendillo and Baumgardner, 1982].
satellites and established the field-aligned nature of the ir-
regularities causing VHF and UHF scintillations, their trans-
verse scale and drift speed. One feature common to all the
The Iong-wavelength (-1 km) field-aligned irregulari- above studies was the fact that the heater frequency was
ties giving rise to artificial spread F [Utlaut et al., 1970; below the plasma frequency of the F region. The magnitude
Utlaut and Violette, 1972; Wright, 1973] could not, how- of scintillations observed on transionospheric communica-
ever, be explained in terms of the above instability process. tion channels is found to be of the order of 5 dB at 250
The causative mechanism for the generation of long wave- MHz when the nighttime F region is heated by an incident
length irregularities remained obscure for quite a while and power density of about 50 pwm-2 [Basu et al., 1980a].

r
AWELERATED ELECTRONS
(WE ESCAPE TO C~JUGATE

fP
400

-\

..
.

300 -
<.3, (-km’s :,:,
THI~~, L’s /
v
HF REFLECTION ALTITUDE

< loom THICK SLAB OF

e g O ?m WAVELENGTH
PLAW INSTABILITIES
AI R6LOW EXCITATION:

I ( ‘IO’S km THICK) / ‘.
r I J I I I
5 6 7 8 9
IONOSPHERIC PLASMA FREQUENCY fp ( MHz)

Figure 10-96 Effects produced by ground-based transmitter of power aperture of the order of 104 Mwm2 in the 4-12 MHz frequency range. Energy
deposited in the ionospheric plasma alters both the thermal and nonthermal properties of its charged particle population. Controlled
experiments have applications to aeronomy, chemical rates, atomic cross sections, communicant ions, and a number of areas of plasma
physics [Carlson and Duncan, 1977].

10-99
CHAPTER 10
REFERENCES

Aarons, J., “A Descriptive Model of F-Layer High-Latitude Baggaley, W. J., “BackScatter Observations of F-Region Field-
Irregularities as Shown by Scintillation Observations,” Aligned Irregularities During the I. Q.S.Y. ,“ J. Geo-
J. Geophys. Res., 78:7441–7450, 1973. phys. Res., 75: 152–158, 1970. ‘
Aarons, J., “Equatorial Scintillations: A Review,” IEEE Bain, W. C., “Models of the Mid Latitude D Region at
Trans. Antennas Propagat., AP-25, 729, 1977. Noon” in Medium, Long and Very Long Wave Propa-
Aarons, J., “Global Morphology of Ionospheric Scintilla- gation, AGARD Conf. Proc., 305: 34- 1–34-8
tions,” Proc. IEEE, 70:360-378, April 1982. ADAl 13969, 1982.
Aarons, J. and R.S. Allen, “Scintillation Boundaries During Balsley, B .B. and W .L. Ecklund, “VHF Power Spectra of
Quiet and Disturbed Magnetic Conditions,” J. Geophys. the Radar Aurora,” J. Geophys. Res., 77: 474H760,
Res., 76: 170, 1971. 1972.
Aarons, J. and E. Martin, “The Effects of the August 1972 Balsley, B .B., W .L. Ecklund, and R.A. Greenwald, “VHF
Magnetic Storms on Ionospheric Scintillations,” Radio Doppler Spectra of Radar Echoes Associated with a
Sci., 10:547-554, 1975. Visual Aurora] Forrrr: Observations and Implications,”
Aarons, J., J. P. Mullen, and H.E. Whitney, “The Scintil- J. Geophys. Res., 78:168 1–1687, 1973.
lation Boundary,” J. Geophys. Res., 74:884-889, 1969. Bannister, P. R., “Variations in Extremely Low Frequency
Aarons, J., E. MacKenzie, and K. Bhavnani, “High-Lati- Propagation Parameters,” J. Atmos. Terr. Phys., 37:
tude Analytical Formulas for Scintillation Levels,” Ra- 1203-1210, 1975.
dio Sci., 15: 115–127, 1980a. Bannister, P. R., “Extremely Low Frequency (ELF) Prop-
Aarons, J., J.P. Mullen, J.P. Koster, R.F. daSilva, J.R. agation,” NUSC Scientific and Engineering Studies, Na-
Medeiros, R.T. Medeiros, A. Bushby, J. Pantoja, J. val Underwater Systems Center, New London, Corm.,
Lanat, and M.R. Paulson, “Seasonal and Geomag- February 1980.
netic Control of Equatorial Scintillations in Two Longi- Bannister, P. R., “Overview of ELF Propagation,” in Me-
tudinal Sectors,” J. Atmos. Terr. Phys., 42: 861–866, dium, Long, and Very Long Wave Propagation, AGARD
1980b. Conf. Proc., 305: 10-1–10-13 ADA113969, 1982.
Aarons, J., E. MacKenzie, and S. Basu, “Microwave Equa- Barghausen, A. L., J.W. Finney, L.L. Proctor, and L.D.
torial Scintillation Intensity During Solar Maximum,” Schultz, “Predicting Long-term Operational Parameters
Radio Sci., 16: 939–945, 198 la. of High-Frequency Sky-Wave Telecommunication Sys-
Aarons, J., J.P. Mullen, H. Whitney, A. Johnson, and E. tems,” ESSA Tech. Rept. ERL 110-ITS 78, 1969.
Weber, “VHF Scintillation Activity Over Polar Lati- Barr, R., “The Effect of Sporadic Eon the Nocturnal Prop-
tudes,” Geophys. Res., Lett., 8: 277–280, 198 lb. agation of ELF Radio Waves, ” J. Atmos. Terr. Phys.,
AGARD, “Modem HF Communications,” AGARD Lecture 39: 1379–1387, 1977.
Series 127, (avail. through NTIS, Springfield, Va) 1983. Barry, G. H., “A Low Power Vertical Incidence Ionosonde”
Akasofu, S.-I., “Midday Auroras and Polar Cap Auroras,” IEEE Trans. Geosci. Elect., GE-9(2): 86–89, 1971.
Geofisiske Publikasjoner (Norske Videnskaps-Akademi Basu, S. and Su. Basu, “Modeling of Equatorial Phase and
i Oslo), 29: 73–85, 1972. Amplitude Scintillations from OGO-6 and AE Irregu-
Albrecht, H. J., (cd.), Artificial Modification of Propagation larityy Data, Low Latitude Aeronomical Processes ,“
Media, AGARD Conf. Proc. 192, Brussels, 26-29 April COSPAR Symposium Series, 8: 187–1 99, Pergamon Press
1976. 1980.
Al’pert, Ia. L., “Ionospheric Propagation of Radio Waves,” Basu, S. and H.E. Whitney, “The Temporal Structure of
Acad. of Sci. of USSR, Moscow, Section on Long Waves Intensity Scintillation Near the Magnetic Equator,” Ra-
(avail) as Translation T5-60, National Bureau of Stand- dio Sci., 18: 263–271, 1983.
ards, Boulder, 1960. Basu, S., Su. Basu, A.L. Johnson, J.A. Klobuchar, and
Ames, J. W. and R.D. Egan, “Digital Recording and Short- C.M. Rush, “Preliminary Results of Scintillation Mea-
Term Prediction of Oblique Ionospheric Propagation,” surements Associated with Ionosphere Heating and Pos-
IEEE Trans. Antennas Propag. AP-15: 382, 1967. sible Implications for the Solar Power Satellite,” Geo-
Ames, J. W., R.D. Egan and G.F. McGintie, “Short-Tern phys. Res. Lett., 7:609, 1980a.
Prediction of HF Communication Circuit Performance” Basu, S., Su. Basu, J.P. Mullen, and A. Bushby, “Long-
in Ionospheric Forecastingj AGARD Conf. Proc., 49, term 1.5 GHz Amplitude Scintillation Measurements at
AD700896, 1970. the Magnetic ~uator,” Geophvs. Res. Lett., 7: 259–262,
Anderson, D. N., “Modelling the Ambient Low Latitude F- 1980b.
Region Ionosphere—A Review,” J. Atmos. Terr. phys., Basu, S., J.P. McClure, Su. Basu, W.B. Hanson, and J.
43:753, 1981. Aarons, “Coordinated Study of Equatorial Scintillation
Anderson, D.N. and P.A. Bemhardt, “Modeling the Effects and in-situ and Radar Observations of Nighttime F Re-
of an H Gas Release the Equatorial Ionosphere,” ~. gion Irregularities,” J. Geophys. Res., 85:5119-5130,
Geophys. Res., 83:4777, 1978. 1980c.
Appleton, E. V. and W .J.G. Beynon, “The Application of Basu, Su. and J. Aarons, “The Morphology of High-Lati-
Ionospheric Data to Radio-Communication Problems: tude VHF Scintillation Near 70”W,” Radio Sci., 15:
Part l,” Proc. Ph]s. Sot., 52:518, 1940. 59-70, 1980.

1o-1oo
 a
IONOSPHERIC RADIO WAVE PROPAGATION
Basu, Su. and M.C. Kelley, “A Review of Recent Obser- Tones” in “Radar Systems and the Ionosphere,” AGARD
vations of Equatorial Scintillations and their Relation- Conf. Proc., 173 ADA022439, 1975.
ship to Current Theories of F-Region Irregularity Gen- Bibl, K., “Fixed and Variable Frequency Pulse Transmis-
eration,” Radio Sci., 14:471, 1979. sion at Oblique Ionospheric Incidence,” AFCRL 64392
Basu, Su., R.L. Vesprini, and J. Aarons, “Field-Aligned AD603045, 1963.
Ionospheric ~-Region Irregularities and Sporadic E,” Bibl, K. and B. W. Reinisch, “Digisonde 128 PS ,“ Uni-
Radio Sci., 8: 235–246, 1973. versity of bwelI, Center for Atmospheric Research,
Basu, Su., R.L. Vesprini, and J. Aarons, “F-Layer Irreg- Lowell, Mass., 1978a.
ularities as Determined by Backscatter Studies at 19 Bibl, K. and B. W. Reinisch, “The Universal Digital Ion-
MHz over Half of a Solar Cycle,” Radio Sci., 9: 355–371, osondes,” Radio Sci., 13:5 19–530. 1978b.
1974. Bickel, J. E., J.A. Ferguson, and G.V. Stanley, “Experi-
Basu, Su., S. Basu, R.C. Livingston, E. MacKenzie, and mental Observations of Magnetic Field Effects on VLF
H.E. Whitney, “Phase and Amplitude Scintillation Sta- Propagation at Night ,“ Radio Sci., 5: 19–25, 1970.
tistics at 244 MHz from Goose Bay using a Geosta- Bilitza, D., M. Noor Sheikh, and R. Eyfrig, “A Global
tionary Satellite, ” AFGL TR-82-0222, ADA 124291, Model for the Height of the F? Peak using M3000 Values
1982. from the CCIR Numerical Map,” Telecomm. J., 46:
Bates, H.F. and P.R. Albee, “Aspect Sensitivity of HF 549-553, 1971.
Auroral Echoes,” J. Geophys. Res., 74: ~164-1168, Black, H.D. ‘me Transit System, 1977; Performance, Plans
1969. and Potential,” Philos. Trans. R. Sot. Lend. A294:
Bates, i-I.F. and P.R. Albee, “Aspect Sensitivity of F-Layer 217–236. 1980,
w Backscatter Echoes,” J. Geophys. Res., 75:165-170, Booker, H .G., C.M. Crain, and E.C. Field, “Transmission
1970. of Electromagnetic Waves Through Normal and Dis-
Belrose, J. S., “Radio Wave Probing of the Ionosphere by turbed Ionospheres,” RAND Rept. R-558-PR, RAND
the Partial Reflection of Radio Waves (From Heights Corp., Santa Monica, Calif., 1970.
below 100 km),” J. Atmos. Terr. Phys., 32:567-596, Booker, H.G. and G. MajidiAhi, “Theory of Refractive
1970. Scattering in Scintillation Phenomena,” J. Atmos. Terr.
Belrose, J. S., “COSPAR Symposium Record on D- and E- Phys., 43:1199, 1981.
Region Ion Chemistry,” Aeron. Rept. No. 48, Univ. of Bowhill, S. A., “Satellite Transmission Studies of Spread-
Illinois, Urbana, Ill., 1972. F Produced by Artificial Heating of the Ionosphere,”
Belrose, J. S., “LF Propagation: An Overview” in Medium, Radio Sci., 9:975, 1974.
Long, & Very Long Wave Propagation, AGARD Conf. Bowles, K. L., “Observations of Vertical Incidence Scatter
Proc., 305:22-1-22-10 ADAl 13969, 1982. from the Ionosphere at 41 Me/see,” Phys. Rev. Let?.,
Belrose, J.S. and L. Thomas, “Ionization Changes in the 1:454, 1958.
Middle Latitude D-Region Associated with Geo- Bowles, K.L. and Staff of Jicamarca Radio Observatory
magnetic Storms,” J. Geophys. Res., 30: 1397–1413, Institute, “Profiles of Electron Density Over the Mag-
1968. netic Equator Obtained Using the Incoherent Scatter
Bennett, S .M. and A.B. Friedland, “Prediction of Daily Technique,” Tech. Rept. Note 169, National Bureau of
Fluctuations of the F-Region Plasma Frequency,” ion- Standards, Boulder, Colo., 1963.
ospheric Forecasting, AGARD Conf. Proc., 49 Bracewell, R.N. and W .C. Bain, “An Explanation of
AD700896, 1970. Radio Propagation at 16 kc/see in Terms of Two Layers
Bent, R. B., S.K. Llewellyn, and M.K. Walloch, “Descrip- Below E Layer,” J. Atmos. Terr. Phys., 2: 2 16–225,
tion and Evaluation of the Bent Ionospheric Model,” I, 1952.
SAMSO-TR-72-239, 1972. Bracewell, R. N., K.G. Budden, J.A. Ratcliff, T. W. Straker,
Bemhardt, P. A., C.G. Park, and P.M. Banks, “DepIetion and K. Weekes, “The Ionospheric Propagation of Low
of the F Region Ionosphere tid Protonosphere by the and Very Low Frequency Radio Waves Over Distances
Release of Molecular Hydrogen,” Geophys. Res. Let., Less than 1000 km,” Proc. lEE, 98:221, 1951.
2:341, 1975. Bradley, P,A. and C. Bedford, “Prediction of HF Circuit
Bernstein, S. L., M.L. Burrows, J.E. Evans, A.S. Griffiths, Availability ,“ Elec. Lett., 12:32, 1976.
D.A. McNeill, C. W. Neissen, I. Richer, D.P. White, Bradley, P.A. and J.R. Dudeney, “A Simple Model of the
and D. K. William, “Long Range Co~munication at Ex- Vertical Distribution of Electron Concentration in the
tremely Low Frequencies,” Proc. IEEE, 6X3), 292–3 12, Ionosphere,” J. Atmos. Terr. Phys., 35:2131, 1973.
1974. Bramley, E. N., “Fluctuations in Direction and Amplitude
Berry, L. A., “Wave Hop Theory of Long Distance Prop- of 136 MHz Signals from a Geostationary Satellite,” J.
agation of LF Radio Waves,” Radio Sci., 68D: Atmos. Terr. Phys., 36: 1503–1513, 1974.
1275–128~, 1964. Breit, G. and M.A. Tuve, “A Test of the Existence of the
Berry, L.A. and M.E. Chrisman, “The Path Integrals of Conducting Layer,” Phys. Rev., 28:554-575, 1926.
LFIVLF Wave Hop Theory,” Radio Sci., 64D: Briggs, B.H. and I.A. Parkin, “On the Variation of Radio
1469–1480, 1965. Star and Satellite Scintillation with Zenith Angle,” J.
Betts, J.A. and M. Damell, “Real Time HF Channel Es- Atmos. Terr. Phys., 25:339-365, 1963.
timation by Phase Measurements on Low-hvel PiIot Brooks, D., “Observations of Radio Auroral Echoes and

10-101
CHAPTER 10
Simultaneous Magnetic Disturbances,’’ Y. Atmos. Terr. CCIR, “Short-Term Prediction of Solar-Induced Variations
Phys.,27: 1151-1158, 1965. of Operational Parameters for Ionospheric Propagation,”
Brooks, D., “Observations of Auroral Echoes Following Study Program 27A/6, Dec. 6/1016-E, 14 Ott 81, XV
Worldwide Storm Sudden Commencements,” J. Atmos. Plenary Ass., ITU, Geneva, 198 lc.
Terr. Phys., 28: IO3–109, 1966. CCIR, “Real-Time Channel Evaluation of Ionospheric Ra-
Buchau, J., W.N. Hall, B.W. Reinisch, and S. Smith, dio Circuits,” Study Program 27B/6, Dec. 6/1018-E of
“Remote ionospheric Monitoring from Effects of the XVth Plenary Ass., Geneva, 198 ld.
Ionosphere on Space and Terrestrial Systems,” Proc. CCIR, “Prediction and Short-Term Forecasting of HF Radio
IES (Ionospheric Effects Symposium), edited by John Propagation Conditions at High Latitudes,” Draft Report
H. Goodman, U.S. Government Printing Office, Wash- submitted ITU, Geneva, August 1983.
ington, D. C., 1978. Chan. K.L. and L. Colin, “Global Electron Density Dis-
Buchau, J., B.W. Reinisch, E.J. Weber, and J.G. Moore, tribution from Topside Soundings,” Proc. IEEE, 57:6,
“Structure and Dynamics of the Winter Polar Cap F 1969.
Region,” Radio Sci., 19: 995–1010, 1983. Chapman, F. W., D.L. Jones, J.D.W. Todd, and R.A. Chal-
Budden, K. B., Radio Waves in the Ionosphere, Cambridge Iinor, “Observations on the Propagation Constant of the
University Press, Cambridge, UK, 1961. Earth-Ionosphere Waveguide in the Frequency Band 8
Buneman, O., “Scattering of Radiation by the Fluctuations c/s to 16 kc/s,” Radio Sci., 1(11), 1273–1282, 1966.
in a Non-Equilibrium Plasma,” Y. Geophys. Res., 67: Ching, B .K. and Chiu, Y. T., “A Phenomenological Model
2050, 1962. of Global Ionospheric Electron Density in the E, F],
Burke, W. (cd.), Proceedings of International Symposium and Fz Regions,”J. Atmos. Terr. Phys., 35:1615, 1973.
on Active Experiments in Space, 23–29 May 1983, Alp- Chiu, Y. T., “An Improved Phenomenological Model of Ion-
bach, Austria, European Space Agency Report ESA- ospheric Density,” J. Atmos. Terr. Phys., 3? 1563, 1975.
SP-195, Paris, 1983. Cole, D.G. and L.F. McNamara, “Variations of Spread-F
Bums, A.A. and E.J. Fremouw, “A Real-Time Correction Occurrence Rates at Near-Equatorial Stations in the Aus-
Technique for Transionospheric Ranging Error,” IEEE tralian Zone,” Aust. J. Phys., 27:249, 1974.
Trans. Antennas Propag., AP-18:(6), Nov. 1970. Cragin, B .L. and J.A. Fejer, “Generation of Large-Sca]e
Burrows, M. L., ELF Communications Antennas, Peter Per- Field-Aligned Irregularities in Ionospheric Modification
egrinus Ltd., Stevenage, England, 1978. Experiments ,“ Radio Sci., 9: 1071, 1974.
Carlson, H.C. and L.M. Duncan, “HF Excited Instabilities Crain, C. M., “Ionospheric Probing with Long Wavelength
in Space Plasmas,” Radio Sci., 12: 1001, 1977. Radio Waves,” J. Atmos. Terr. Phys., 32: 55 1–566,
Carlson, H. C., W.E. Gordon, and R.L. Showen, “High Fre- 1970.
quency Induced Enhancements of the Incoherent Scatter Crane, R. K., ‘&Spectraof Ionospheric Scintillation,” J. Geo-
Spectrum at Arecibo,”J. Geophys. Res., 77:1242, 1972. phys. Res., 81:2041-2050, 1976.
CCIR (Comite Cosultatif International des Radiocommun- Crane, R. K., “Ionospheric Scintillation,” Proc. IEEE, 65:
ications = International Radio Consultative Commit- 180, 1977.
tee), “World Distribution and Characteristics of At- Cretcher, C. K., “Ionospheric Effects in NAVSTAR-GPS,”
mospheric Radio Noise,” Report 332, Dec. Xth Plenary in Effects of the Ionosphere on Space Systems and Com-
Ass., International Telecommunications Union (ITU), munications, edited by J .M. Goodman, Naval Research
Geneva, 1963. Lab., Washington, D. C., U .S. Government Printing Of-
CCIR, “CCIR Atlas of Ionospheric Characteristics,” Report fice, Stock No. 08-051-00064-0, 1975.
340 and 340-2, Dec. XIth Plenary Ass., ITU, Geneva, Czechowsky, P., W. Dieminger, and H. Kochan, “Back-
1966. scatter Results from Lindau-I. Observations of Radio
CCIR, “Interim Method for Investigating Sky-Wave Field Auroras,” J. Atmos. Terr. Phys., 36: 955–966, 1974.
Strength and Transmission Loss at Frequencies Between D’Accardi, R. J., “Time Series Modelling and Analysis of
the Approximate Limits of 2 and 30 MHz,” Report 252-2, HF Vertical and Short-Path Oblique Incidence Ion-
Dec. XIIth Plenary Ass., ITU, Geneva, 1970. ospheric Soundings,” U.S. Army Communications R &
CCIR, “Second CCIR Computer-Based Interim Method for D Command Tech. Rept. CORADCOM 78-7, NTIS
Estimating Sky-Wave Field Strength and Transmission ADA058630, 1978.
Loss Between 2 and 30 MHz,” Suppl. Rept. 252-2, Doe. Dandekar, B. S., “Ionospheric Modeling,” AFGL TR-82-
XIVth Plenary Ass., ITU, Geneva, 1978. 0024, ADA 115243, 1982.
CCIR, “Prediction of Sky-Wave Field Strength Between Danielson, C., “Aurora] Observations at Thule 1961 -65,”
150 and 1500 kHz,” CCIR Recommendation 453-3, Doe. Danish Met. Inst., Geophys. Papers R-9, 1969.
XIV Plenary Ass., ITU, Geneva, 1978. Damell, M., “Channel Evaluation Techniques for HF Com-
CCIR, “Special Properties of the High Latitude Ionosphere munication” in Radio Systems and the Ionosphere,
Affecting Radio-communications,” Study Program 25B16, AGARD Conf. Proc., 173, 1975a.
Dec. 6/1005-E, 14 Ott 81, XVth Plenary Ass., ITU, Damell, M., “Adaptive Signal Selection for Dispersive
Geneva, 1981a. Channels and its Practical Implications in Communi-
CCIR,’’Short-Term Forecasting of Operational Maximum cations System Design” in Radio Systems and the Ion-
Usable Frequencies and Total Electron Content ,“ Study osphere, AGARD Conf. Proc., 173, 1975b.
Program 27C/6, Dec. 6/1017-E, 14 Ott 81, XVth Plen- Damell, M., “Channel Evaluation Techniques for Disper-
ary Ass., ITU, Geneva, 1981b. sive Communications Paths,” Communication System

10-102
 IONOSPHERIC RADIO WAVE PROPAGATION
and Random Process Theory, edited by J. K. Skwirzyn- gation Study, Lincoln Laboratory, MIT, Lexington,
ski, Sijthoff and Noordhoff, Alphen aan den Rijn, Hol- Mass., 1973.
land, 1978. Evans, J. V., “High-Power Radar Studies of the Iono-
DasGupta, A., A. Maitra, and S. Basu, “Occurrence of sphere, “ Proc. IEEE, 63: 12, 1636-1650, 1975.
Nighttime VHF Scintillations near the Equatorial Anom- Eyfrig, R., “Contribution a I’Etude de ~l’Effect du Geo-
aly Crest in the Indian Sector,” Radio Sci., 16: 1455–1458, magnetisme sur le Couche F, ,“ Comptes Rendus Aca-
1981. demie des Sciences, 241: 759–761, 1955.
David, P. and J. Voge, Propagation of Waves, Pergamon Fejer, J. A., “Scattering of Radiowaves by an Ionized Gas
Press, London, U. K., 1969. in Thermal Equilibrium,” Can. J. Phys., 38:1114, 1960;
Davies, K., Ionospheric Radio Propagation, National Bu- also J. Geophys. Res., 65:2635, 1960.
reau of Standards Monograph 80, U.S. Government Fejer, J. A., “Radio Wave Probing of the Lower Ionosphere
Printing Office, Washington, D. C., 1966. by Cross-Modulation Techniques,” J. Atmos. Terr. Phys.,
Davies, K., Ionospheric Radio Waves, Blaisdell, Waltham, 32: 597~7, 1970.
Mass., 1969. Fejer, J .A., “Ionospheric Modification and Parametric In-
Davies, K., “Recent Progress in Satellite Radio Beacon stabilities,” Rev. Geophys. Space Phys., 17: 135, 1979.
Studies with Particular Emphasis on the ATS-6 Radio Feldstein, Y.I. and G. V. Starkov, “Dynamics of Auroral
Beacon Experiment,” Space Sci. Rev., 25: 357+30, Belt and Polar Geomagnetic Disturbances,” Plan. Space
1980. Sci., 15: 209–229, 1967.
Davis, T. N., “The Morphology of the Auroral Displays of Ferguson, J .A., “Ionospheric Profiles for Predicting Night-
1957–1 958. (2) Detailed Analysis of Alaska Data and time VLF/LF Propagation,” Tech. Rep. 530, Naval Ocean
Analysis of High Latitude Data,” Y. Geophys. Res 67: Systems Center, San Diego, Calif., 1980.
75-110, 1962. Field, E. C., “Propagation of ELF Waves Under Normal and
Demaro, R. P., “NAVSTAR: The All-Purpose Satellite,” Naturally Disturbed Conditions,” J. Geophys. Res., 74:
IEEE Spectrum, 18:3540, May 1981. 3639–3650, 1969.
Donatelli, D.C. and R. S. Allen, “Time Cells for Adaptive Field, E. C., “Low-Frequency Ground Wave Propagation
Predictions of Total Electron Content ,“ Radio Sci., 16: Over Narrow Terrain Features ,“ IEEE Trans. Antennas
261–269, 1981. Propag., 30: 831–836, 1982a.
Donnelly, R. F., ed., Solar Terrestrial Prediction Proceed- Field, E. C., “ELF Propagation in Disturbed Environments”
ings, U.S. Dept. of Commerce, U.S. Government Print- in Medium, Long and Very Long Wave Propagation
ing Office, Washington, D.C. 20402 AGARD Conf. Proc., 305: 11-1–11-10, ADA113969
vol. 1 Prediction Group Reports, No. 003-023-0041-9, 1982b.
1979. Field, E. C., “VLF propagation in Disturbed Environments”
vol. 2 Working Group Reports and Reviews, No. 003- in Medium, Long and Very Long Wave Propagation
017-00471-6, 1979. AGARD Conf. Proc., 305: 17-1-17-10, ADA113969,
vol. 3 Solar Activity Predictions, No. 003-017-00473- 1982c.
2, 1979. Field, E.C. and R.D. Engel, “The Detection of Daytime
vol. 4 Predictions of Terrestrial Effects of Solar Activ- Nuclear Bursts Below 150 km by Prompt VLF Phase
ity, No. 003-017-00479-1, “~980. “ Anomalies,” Proc. IEEE, 53:2009-2017, 1965.
Dougherty, J.P. and D.T. Farley, “A Theory of Incoherent Field, E.C. and R.G. Joiner, “An Integral Equation Ap-
Scattering of Radio Waves by a Plasma,” Proc. Roy. proach to Long Wave Propagation in a Non-Stratified
Sot. (London), A259: 79, 1960. Earth-Ionosphere Waveguide,” Radio Sci., 14:
Dubois, D.F. and M. V. Goldman, “Spectrum and anom- 1057–1068, 1979.
alous Resistivity for the Saturated Parametric Instabil- Field, E.C. and R.G. Joiner, “Effects of Lateral Ionospheric
ity,” Phys. Rev. Lett., 28:218, 1972. Gradients on ELF Propagation,” Rdio Sci., 17: 693–700,
Ducharme, E. D., L.E. Petrie, and R. Eyfrig, “A Method 1982.
for Predicting the F, Critical Frequency,” Radio Sci., Field, E.C. and C.R. Warber, “Calculation of Ionospheric
6:369, 1971. Electron Density Profiles by Inverting VLF/LFReflection
Ducharme, E. D., L.E. Petrie, and R. Eyfrig, ‘6A Method Data, 2. Anisotropic Propagation” RADC-TR-84-242,
of Predicting the F1 Layer Critical Frequency Based on Rome Air Development Center, Griffiss AFB, N. Y., 1984.
Zurich Smoothed Sunspot Number,” Radio Sci., 8: Field, E. C., M. Lewinstein, and M.A. Dore, “Effects of
837-839, 1973. Antenna Elevation and Inclination on VLF/LF Signal
Dyson, P. L., J.P. McClure, and W. B. Hanson, “In situ Structure RADC-TR-76-375 ADA0355 10, 1976.
Measurements of the Spectral Characteristics of Iono- Field, E. C., R.E. Warren, and C.R. Warber, “Calculation
spheric Irregularities,” J. Geophys. Res., 79:1497, 1974. of Ionospheric Conductivity Profiles by Inverting VLF/LF
Eccles, D. and J. W. King, “A Review of Topside Sounder Reflection Data, 1. Isotropic Propagation,” Radio Sci.,
Studies of the Equatorial Ionosphere” Proc. IEEE 57(6): 18:452460, 1983.
1012-1018, 1969. Flattery, T. W., T.F. Tascione, J.A. Secan, and J.W. Tay-
Evans, J. V., “Theory and Practice of Ionosphere Study by lor, Jr., A Four-Dimensional Ionospheric Model (private
Thomson Scatter Radar,” Proc. IEEE, 57:496, 1969. communication).
Evans, J. V., “Millstone Hill Radar Propagation Study” Sci- Flood, W .A. “Simultaneous VHF Auroral Backscatter Mea-
entific Results,” Tech. Rept. 509; Joint Radar Propa- surements,” J. Geophys. Res., 65:2261 –2268, 1960.

10-103
CHAPTER 10
Forbes, J. M., “Upper Atmosphere Modifications Due to Free Electrons with Applications to Space Exploration
Chtonic Discharges of Water Vapor from Space Vehicle by Radar,” Proc. IRE, 46:1824, 1958.
Exhausts,” iri Space Systems and their Interactions with Gordon, W. E., H.C. Carlson, and R.L. Showen, “Iono-
Earth’s Space Environment edited by H.B. Garrett and spheric Heating at A~ci@: First Tests,” J. Geophys.
C.P. Pike, American Institute of Aeronautics and Astro- Res., 76:7808, 1971.
nautics, New York, 1980. Gowell, R. W. and R. W. Whidden, “Ionospheric Sounders
Franklin, C.A. and M.A. MacLean, “The Design of Swept in Aircraft ,“ AFCRL-68-0369, AD678047, 1968.
Frequency Topside Sounders,” Proc. IEEE, 57(6): Greenwald, R. A., W.L. Ecklund, and B.B. Balsley, “Dif-
897-929, 1969. fuse Radar Aurora: Spectral Observations of Non-Two-
Fremouw, E. J., “Geometrical Control of the Ratio of In- Stream Irregularities,” ~. Geophys. Res., 80:131-139,
tensit y and Phase Scintillation Indices,” Y. Atmos. Terr. 1975.
Phys. , 42:775-782, 1980. Griefinger, C. and P. Greifinger, “Approximate Method for
Fremouw, E.J. and J.F. Bates, “Worldwide Behavior of Determining ELF Eigenvalues in the ~b-Ionosphere
Average VHF-UHF Scintillation,” Radio Sci., 6: Waveguide,” Radio Sci., 13:83 1–837, 1978.
863-869, 1971. Griefinger, “C. and P. Griefinger, “On the Ionospheric Pa-
Fremouw, E.J. and C.L. Rino, “Modelling of Transion- rameters which Govern High htitude ELF Propagation
ospheric Radio Propagation,” RADC-TR-76-35, Rome in the Earth-Ionosphere Waveguide,” Radio Sci., 14:
Air Development Center, Griffiss AFB, NY, 1976. 889–895, 1979.
Fremouw, E. J., C.L. Rino, A.R. Hessing, and V.E. Hat- Grubb, R. N., “The NOAA SEL HF Radar Systems (Iono-
field, “A Transionospheric Communication Channel spheric Sounder),” NOAA Tech. Memo ERL SEL-55
Model,” Quart. Tech. Rep. 7, SRI International (Stan- NOAA Environmental Research Laboratories, Boulder,
ford Research Institute), Menlo Park, Calif., 1977. co, 1979.
Fremouw, E.J. and C.L. Rino, “A Signal-Statistical and Haerendel, G., “Modification of Ionized Media by Chemi-
Mo@hological Model of Ionospheric Scintillation,” Proc. cal Substimces-A Review of Physical Processes,” in
AGARD Conf. Operational Modelling of the Aero- Artificial Modification of Propagation Media, edited
space Propagation Environment, Ottawa, Canada, by H.J. Albrecht, AGARD Conf. Proc., 192, Brussels,
1978. 1976.
Fremouw, E. J., R.L. Leadabrand, R.C. Livingston, M.D. Haerendel, G., O.H. Bauer, S. Cakir, H. Foppl, E. Reiger,
Cousins, C.L. Rino, B.C. Fair, and R.A. Long, “Early and A. Valenzuela, “Coloured Bubbles—An Experi-
Results from the DNA Wideband Satellite Experiment- ment for Triggering Equatorial Spread F,” in Active
Complex Signal Scintillation,” Radio Sci., 13: 167–187, Experiments in Space edited by W. Burke, European
1978. Space Agency Report SP-195: 295, 1983.
Fremouw, E. J., R.C. Livingston, and D.A. Miller, “On we Hagfors, T., “Density Fluctuations in a Plasma in a Mag-
Statistics of Scintillating Signals,” J. Atmos. Terr. Phys., netic Field with Applications to the Ionosphere,” J. Geo -
43:717-731, 1980. phys. Res., 66:1699, 1961.
Fremouw, E.J. and J.M. Lansinger, “Recent High Latitude Hagg, E. L., E.J. Hewens, and G.L. Nelms, “The Inter-
Improvement in a Computer Based Scintillation Model,” pretation of Topside Sounder Monograms,” Proc. IEEE,
Proc. Ionospheric Effects Symposium, 3rd, 1981, Al- 57:949-959, 1969.
exandria, Va., edited by J. M. Goodman, Naval Re- Haldoupis, C. and G. Sofko, “Doppler Spectrum of 42 MHz
search Lab., Washington, D. C., 1982. CW Auroral Backscatter,” Can. J. Phys., 54: 1571,
Frisius, J., A. Ehmert, and D. Stratmann, “Effects of High 1976.
Altitude Nuclear Tests on VLF-Propagation,” J. Atmos. Hargreaves, J. K., The Upper Atmosphere and Solar-Ter-
Terr. Phys., 26:25 1–262, 1964. restrial Relations, Van Nostrand Reinhold, New York,
Galejs, J., Terrestrial Propagation of Long Electromagnetic 1979.
Waves, Pergamon Press, Oxford, UK, 1972. Harrison, R. P., J.L. Heckscher, and E.A. Lewis, “Heli-
Gallagher, P. B. and R.A. Barnes, “Radio-Frequency Back- copter Observations of Very Low Frequency Radio Waves
scatter of Artificial Electron Clouds ,“ Y. Geophys. Res., Over Ce@in Mountains and Shorelines,” J. Atmos. Terr.
68:2987-3010, 1963. Phys., 33: 101–110, 1971.
Gautier, T.N. and D.H. Zacharisen, “Use of Space and Harvey, R. P., R.B. Harrison, V.C. Fields, G.C. Hirst, P.A.
Time Correlations in Short-Term Ionospheric Predic- Kossey, and E.A. Lewis, “Rocket Investigations of the
tions,” Conference Record, 1st Annual JEEE Commu- VLF Ionospheric Transmission Window,” AFCRL-TR-
nications Convention, 671-676, 1965. 73-0293, AD767185, 1973.
Goldman, M. V., “Field Aligned Instability Due to Stimu- Haydon, G.W. and D.L. Lucas, “Predicting Ionosphere
lated Scattering of Intense Radio Waves from Diffusion Electron Density Profiles,” Radio Sci., 3: 111, 1968.
Quasimodes,” Radio Sci., 9:1077, 1974. Headrick, J .M. and M.I. Skolnik, “Over-the-Horizon Radar
Goodman, J. M., “A Survey of Ionospheric Models. A Pre- in the HF Band,” Proc. IEEE, 62:664, 1974.
liminary Report on the Development of an Ionospheric Heckman, G., “Predictions of the Space Environment Ser-
Model Thesaurus and Users Guide,” NRL Memorandum vices Center,” STPP, (Solar Terrestrial Predictions
Report 4830, Naval Research Lab., Washington, D. C., Workshop) 1:322 (see R.F. Donnelly, cd. ) 1979.
1982. Heckscher, J.L. and E.J. Tichovolsky, “Height-Gain Atlas
Gordon, W. E., “Incoherent Scattering of Radio Waves by for an Elemental Vertical Dipole Above a Flat Earth,”

10-104
 a
IONOSPHERIC RADIO WAVE PROPAGATION
RADC-TR-81 -39, Rome Air Development Center, Grif- gation Between the Ionosphere and the Ground,” Radio
fiSS AFB, N. Y., 1981. Sci., 68D:67–78, 1964.
Helliwell, R. A., Whistlers and Related Ionospheric Phe- Johler, J. R., “Zonal Harmonics in Low Frequency Terres-
nomena, Stanford University Press, Stanford, Calif., trial Radio Wave Propagation ,“ NBS Tech. Note TN335,
1965. U.S. Government Printing Office, Washington, D. C.,
Herman, J. R., “Spread F and Ionospheric F-Region Irreg- 1966.
ularities,” Rev. Geophys., 4: 244–299, 1966. (Includes Jones, T. B. and K. Mowforth, “A Review of the Analyti-
important bibliography. ) cal Techniques for Determining the Phase and Ampli-
Heron, M.L. and L.F. McNamara, “Transequatorial VHF tude of a VLF Radio Wave Propagating in the Earth-
Propagation through Equatorial Plasma Bubbles,” Radio Ionosphere Waveguide,” in Medium, Long, and Ve~
Sci., 14, 897, 1979. Long Wave Propagation, AGARD Conf. Proc., 305:
Hill, D. A., and J. R. Wait, “Excitation of the Zenneck Sur- 16-1-16-16 ADA113969, 1982.
face Wave by a Vertical Aperture,” Radio Sci., 13: Jones, W .B. and R.M. Gallet, “Ionospheric Mapping by
969–977, 1978. Numerical Methods,” ITU Telecomm. J., 12:260, 1960.
Hizal, A. and A.F. Fer, “Medium Wave Multi-Section Jones, W. B., R. P. Graham, and M. Leftin, Advances in
Groundwave Propagation in Marinara and West Black- Ionospheric Mapping by Numerical Methods, NBS Tech.
Sea Region—Numerical Results” in Medium, Long & Note 337, U.S. Government Printing Office, Washing-
Very Long Wave Propagation AGARD Conf. Proc., 305: ton, D. C., 1966.
31-1–31-7, ADA113969 1982. Kelley, M. C., K.D. Baker, and J.C. Ulwick, “Late Time
Hofstee, J. and P.A. Forsyth, “Ion Acoustic Waves in the Barium Cloud Striations and Their Possible Relationship
Auroral Plasma,” Can. J. Phys., 47:2797, 1969. to Equatorial Spread F,”J. Geophys. Res., 84:1898, 1979.
Hollingworth, J., “The Propagation of Radio Waves,” J. Kellog, W.W. “Pollution of the Upper Atmosphere by
IEE, 64:579, 1926. Rockets,” Space Sri. Rev., 3:275, 1964. .
Holmgren, G., R. Bostrom, M.C. Kelley, P.M. Kintner, Kelso, J. M., Radio Propagation in the Ionosphere, Mc-
R. Lundin, U.V. Fahleson, E.A. Bering, and W.R. Graw Hill, New York, 1964.
Sheldon, “Trigger and Active Release Experiment that Kersley, L., “An Empirical Model of Ionospheric Slab
Stimulated Aurora] Particle Precipitation and Wave Thickness,” AGARD-CPP-284, May 1980.
Emissions,” .1. Geophys. Res., 85:5043, 1980. King, R.J. and J .R. Wait, “Electromagnetic Groundwave
Hewer, G.L. and A.B. Makhijani, “Further Comparison of Theory and Experiment ,“ Symposia Mathematical (In-
Spread-F and Backscatter Sounder Measurements,” J. stituto Nazionale di Alta Mathematical, Bologna, Italy),
Geophys. Res., 74: 3723–3725, 1969. 18: 107–208, 1976.
Hewer, G. L., D.M. Ranz, and C.L. Allison, “Compari- Klobuchar, J. A., “Numerical Models of Total Electron Con-
son of HF Radar Echoes and High-Lattitude Spread-F tent over Europe and the Mediterranean and Multi-Sta-
Measurements,” J. Geophys. Res., 71: 32 15–3221, tion Scintillation Comparisons,” AGARDOGRAPH: AG-
1966. 166A, Nov. 1973.
Hufford, G. A., “An Integral Equation Approach to the Klobuchar, J. A., “A First-Order, Worldwide, Ionospheric
Problem of Wave Propagation Over an Irregular Sur- Time Delay Algorithm,” AFCRL-TR-75-0502,
fa~,” Quart. Appl. Math., 9:391403, 1952. ADA018862, Sep 1975.
Hultqvist, B., “Polar Cap Absorption and Ground Level Klobuchar, J.A. and R.S. Allen, “A First-Order Prediction
Effects,” in Solar Flare and Research, edited by G. de Model of Total-Electron-Content Group Path Delay for
Jager and Z. Svestka, p. 216, North Holland, Amster- a Midlatitude Ionosphere ,“ AFCRL-70-0403, AD711 365
dam, 1969. 1970.
Hunsucker, R. D., “Morphology and Phenomonology of the Klobuchar, J. A., M.J. Buonsanto, M.J. Mendillo, and J.M.
High Latitude E and F Regions,” STPP, 2:543 (see R.F. Johanson, “The Contribution of the Plasmasphere to
Donnelly, cd.), 1979. Total Time Delay” in Effects of the Ionosphere on Space
Hunsucker, R.D. (Chairman), “High Latitude E- and F- and Terrestrial Systems, edited by J.M. Goodman, U.S.
Region Ionospheric Predictions,” STPP, 2:513 (see R.F. Government Printing Office stock no. 008-051-00069-
Donnelly, cd.), 1979. 1, 1978.
Hunsucker, R. D. and R.F. Bates, “Survey of Polar and Klobuchar, J .A. and J. M. Johanson, “Correlation Distance
Aurora] Region Effects on HF Propagation,” Radio Sci., for Mean Daytime Electron Content,” AFGL TR-77-
4:347-365, 1969. 0185, ADA048117, 22 Aug 1977.
Ichinose, M., R. Maeda and S. Ito, “MUF Short-Term Klobuchar, J. A., K.N. Iyer, H.O. Vats, and R.G. Rastogi,
Prediction for HF Radio Propagation Circuits over Japan “A Numerical Model of Equatorial and Low Latitude
Area,” J. Rad. Res. Labs. (Japan), 27: 179, 1980. Total Electron Content for use by Satellite Tracking
Inoue, Y. and S. Horowitz, “Numerical Solution of Full- Systems for Ionospheric Corrections,” Indian J. Radio
Wave Equation with Mode Coupling,” Radio Sci., 1: Space Phvs., 6, 1977.
957, 1968. Knight, P., “MF Propagation: A Wave-Hop Method for
Jelly, D. H. and L. E. Petrie, “The High-Latitude Iono- Ionospheric Field-Strength Prediction,” BBC Engineer-
sphere,” Proc. IEEE, 57: 1005-1018, 1969. ing, 100: 22–34, 1975.
Johler, J. R., “Concerning Limitations and Further Correc- Knight, P., “Medium Frequency Propagation: A Survey,”
tions to Geometric Optic Theory for LF, VLF Propa- in Medium, Long, and Very Long Wave Propagation,

10-105
CHAPTER 10
AGARD Conf. Proc., 305:28-1-28-17, ADAl 13969 quency Observations of the Aurora at Fraserburgh, Scot-
1982. land,” J. Geophys. Res., 70:42354284, 1965.
Kofsky, 1., “Further Evaluations of Infrared Simulation Data,” Lee, M. C., A. DasGupta, J.A. Klobuchar, S. Basu, and
Tech. Rept. DNA Contract 001-81-C-003, Photo Met- S. Basu, “Depolarization of VHF Geostationary Satellite
rics, Inc., Wobum, Mass., 1981. Signals near the Equatorial Anomaly ’Crest,” Radio Sci.,
Kohnlein, W., “Electron Density Models of the Iono- 17:399, 1982.
sphere,” Rev. Geophys. Space Phys., 16:341, 1978. Leftin, M., “Numerical Representation of Monthly Median
Kossey, P.A. and E.A. Lewis, “Detection and Classification Critical Frequencies of the Regular E Region (F&),”
of Solar X-Ray Flares Using VLF Phase and Amplitude OT Rept. 76-88k, U.S. Office of Telecommunications,
Information,” AFCRL-TR-74-0468, ADAO03383 1974. Boulder, Colo., 1976.
Kossey, P. A., E.A. Lewis, and E.C. Field, “Relative Char- Leftin, M., S .M. Ostrow, and C. Preston, “Numerical Maps
acteristics of T~ Waves Excited by Airborne VLF~F of foE, for Solar Cycle Minimum and Maximum,”
Transmitters, “ in Medium, Long, and Very Long Wave ESSA Tech Rept. ERL 73-ITS-63, Boulder, Colo.,
Propagation, AGARD Conf. Proc., 305: 19- 1–19- 10 1968.
ADAl 139691982. Leiphart, J. P., R.W. Z.eek, L.S. Bearce, and E. Toth, “Pen-
Kossey, P. A., J. E. Rasmussen, and E.A. Lewis, “VLF etration of the Ionosphere by Very-Low-Frequency Ra-
Pulse Ionosounder Measurements of the Reflection dio Signals-Interim Results of the LO~I I Experiment ,“
Properties of the Lower Ionosphere,” COSPAR, Lower Proc. IRE, 50:6-17, 1962.
Ionosphere Structure, 133–1 38, Akademie Verlag, Ber- Lewis, E.A. and P.A. Kossey, “POWERFLUX I: A Method
lin, 1974. of Estimating Wave Intensities at Large Distances from
Kossey, P. A., J.P. Turtle, R.P. Pagliarulo, W.I. Klemetti, Ground-Based Low Frequency Transmitters,” AFCRL-
and J.E. Rasmussen, “VLF Reflection Properties of the TR-75-0338, ADA014787, 1975.
Normal and Disturbed Polar Ionosphere in Northern Lewis, E. A., J .E. Rasmussen, and P.A. Kossey, “Mea-
Greenland,” Radio Sci., 18:907-916, 1983. surements of Ionospheric Reflectivityy from 6 to 35 kHz,”
Koster, J. R., “Some Measurements of the Irregularities Giv- J. Geophys. Res., 78:3903-3911, 1973.
ing Rise to Radio-Star Scintillations at the Equator,” J. Lewis, R. L., “The Wave-Hop fields for an Inclined Dipole
Geophys. Res., 68: 2579–2590, 1963. Over a Spherical Earth with an Anisotropic Ionosphere,”
Kouba, J. “A Review of Geodetic and Geodynamic Satellite Tech Rept. OT/ITS RR 5, Institute for Telecommuni-
Doppler Positioning, Rev. Geophys. and Space Phys. cation Sciences, Boulder, Colo., 1970.
21:27-40, 1983. Lied, F., (cd.), H.F. Radio Communications, Technivision,
Kouris, S. S. and L.M. Muggleton, “Diurnal Variation in Maidenhead, UK, 1967.
the E Layer Ionization,” J. Atmos. Terr. Phys., 35: Lincoln, J. V. and R.O. Conkright (eds.), “International
133–139, 1973a. Reference Ionosphere-IRI 79,” Report UAG-82, World
Kouris, S.S. and L.M. Muggleton, “A Proposed Prediction Data Center A for Solar-Terrestrial Physics, Boulder,
Method for Monthly Median foE,” CCIR Report 252- Colo. , 1981.
2, Contrib. No. 6/3/07 to Interim Working Party 6/3, Llewellyn, S .K. and R. B. Bent, “Documentation and De-
1973b. scription of the Bent Ionospheric Model ,“ AFCRL-TR-
Krause, G. E., R.J. D’Accardi, and E.L. Roswell III, “Field 73-0657, AD 772733, 1973.
Test of a Near-Real-Time Ionospheric Forecasting Scheme Lloyd, J. S., G.W. Haydon, D.L. Lucas, and L.R. Teters,
(60 km),” U.S. Army ECOM Rept. 3345, 1970. Estimating the Performance of Telecommunication Sys-
Krause, G. E., R.J. D’Accardi, and E.L. Roswell 111,“Field tems Using the Ionospheric Transmission Channel, Na-
Test of a Near-Real-Time Ionospheric Fo~casting Scheme tional Telecommunications and Information Adminis-
(200 km),” U.S. Army ECOM Rept. 4144, 1973a. tration, Boulder, Colo., 1978.
Krause, G. E., R.J. D’Accardi, and E.L. Roswell III, “Field Lyakhova, L. N., “Short-Range Forecasting of the State of
Test of a Near-Real-Time Ionospheric Forecasting Scheme the Ionosphere,” Trudy IZMIRAN, 17:240, 1960.
(500 km),” U.S. Army ECOM Rept. 4145, 1973b. Lyakhova, L.N. and L.I. Kostina, “On Quantitative Iono-
Lakshmi, D. R., S. Aggarwal, P.K. Pasricha, and B .M. spheric Forecasting,” Geomag. Aeron., 13:50, 1973.
Reddy, “HF Communication Problems at Low Latitudes Maeda, K. I., “Study on Electron Density in the Lower
due to Steep Spatial and ~emporal Gradients,” STPP, Ionosphere,” J. Geomag. Geoelect., 23:133-159, 1971.
4: D2-58 (see R.F. Donnelly, cd.), 1980. Malik, C. and J. Aarons, “A Study of Auroral Echoes at
Larson, A.C. and J .C. Hodges, “VHF/UHF Auroral Radar 19.4 Megacycles per Second,” J. Geophys. Res., 69:
Measurements,” Stanford Research Institute Intern- 273 1–2736, 1964.
ational, Cal if., June 1967. Marmo, F. F., L.M. Ashcenbrand, and J. Pressman, “Ar-
Leadabrand, R. L., A.G. Larson, and J.C. Hodges, “Pre- tificial Electron Clouds—I, Planet ,“ Space Sci., 1:277,
liminary Results on the Wavelength Dependence and 1959.
Aspect Sensitivity of Radar Auroral Echoes Between 50 Martin, E. and J. Aarons, “F-Layer Scintillation and the
and 3000 MHz, ” J. Geophys. Res., 72: 3877–3887, Aurora,” J. Geophys. Res., 82:2717, 1977.
1967. McNamara, L. F., “Ionospheric Predictions on Transequa-
Leadabrand, R. L., J.C. Schlobohm, and M.J. Baron, “Si- torial Circuits,” Proc. lREE (Aust .), 35: 17, 1974a.
multaneous Very High Frequency and Ultra High Fre- McNamara, L. F., “Unusual Nighttime Propagation Modes

10-106
 d

IONOSPHERIC RADIO WAVE PROPAGATION

on the Guam and Manila—N. W. Cape Circuits,” Iono- in Medium, Long and Ven Long Wa\’e Propagation,
spheric Prediction Service Series R Reports, IPS-R27, AGARD Conf. Proc., 305: ADA 113969 32-1–32-14,
Sydney, Australia, 1974b. 1982.
McNamara, L. F., “Short-Term Forecasting of foF2,” Iono- Muggleton, L. M., “Solar Cycle Control of N.,(E),” J. AI-
spheric Prediction Service Series R Reports, IPS-R33, mos. Terr. Phys., 33: 1307, 1971.
Sydney, Australia, 1976. Mullen, J. P., “Sensitivity of Equatorial Scintillation to
McNamara, L. F., “The Use of Ionospheric Indices to Make Magnetic Activity,” J. Atmos. Terr. Phvs., 35: 1187,
Real and Near Real Time Predictions of foF2 Around 1973.
Australia,” STPP, 1:249 (see R.F. Donnelly, cd.), 1979. Muldrew, D.B. , “F-Layer Ionization Troughs Deduced from
Mendillo, M., “The Effect of Rocket Launches on the Iono- Alouette Data,” J. Geophys. Res., 70(11): 2635-2650,
sphere,” Adv. Space Res., 1:275, 1981. 1965.
Mendillo, M. and J. Baumgardner, “Optical Signature of Narcisi, R. S., “Overview of Project BIME (Brazil ion-
an Ionospheric Hole,” Geophys. Res. Lett., 9215, 1982. ospheric Modification Experiments) ,“ in Active E.rper-
Mendillo, M. and J. M. Forbes, “Artificially Created Holes iments in Space edited by W. Burke, European Space
in the Ionosphere,” J. Geophys. Res., 83: 151, 1978. Agency Report SP- 195, p. 295, 1983.
Mendillo, M. and J. Forbes, “Theory and Observation of a Newman, P. (cd.), “Spread F and its Effects upon Radi -
Dynamically Evolving Negative Ion Plasma,” J. Geo- owave Propagation and Communication,” Technivision,
phys. Res. 87:8273, 1982. Maidenhead, UK, 1966. (Also AGARDOGRAPH 90).
Mendillo, M., G.S. Hawkins, and J.A. Klobuchar, “A Large- Nielsen, E. and J. Aarons, “Satellite Scintillation Obser-
Scale Hole in the Ionosphere Caused by the Launch of vations over the Northern High Latitude Regions,” J.
Skylab,” Science, 187:343, 1975a. Atmos. Terr. Phys., 36: 159–165, 1974.
Mendillo, M., G.S. Hawkins, and J.A. Klobuchar, “A Sud- Nielson, D. L. and M. Crochet, “Ionospheric Propagation of
den Vanishing of the Ionospheric F-Region Due to the HF and VHF Radio Waves Across the Geomagnetic
Launch of Skylab,” J. Geophys. Res., 80:2217, 1975b. Equator,” Rev. Geophvs. Space Phvs., 12:688, 1974.
Mendillo, M. D. Rote, and P. Bemhardt, “Preliminary Re- Nisbet, J. S., “On the Construction and use of a Simple
port on the HEAO-Hole in the Ionosphere,” Eos, Trans. Ionospheric Model,” Radio Sci., 6:437, 1971.
AGU, 61:529, 1980. Norton, K. A., “The Calculation of Ground-Wave Field In-
Mikkelsen, 1.S., J. Aarons, and E. Martin, “Geometrical tensity Over a Finitely Conducting Spherical Earth,”
Considerations of 136 MHz Amplitude Scintillation in Proc. IRE, 29:623-639, 1941.
the Auroral Oval,” J. Atmos. Terr. Phys., 40:479483, Noxon, J .F. and A.F. Johanson, “Effects of Magnetically
1978. Conjugate Photoelectrons on 01 (6300 A),” Planet. Space
Milliken, R.J. and C.J. Zaller, “Principle of Operation of Sci., 18: 1367, 1970.
NAVSTAR and System Characteristics,” Navigation: J. Oelbermann, E., “Solar Particle Effects on Polar Cap VLF
Institute of Navigation, 25(2): 95–106 Summer 1978. Propagation,” J. Franklin Inst., Special Issue on the
Millington, G., “Ground-Wave Propagation Over an In- Polar Cap, 290:281-296, 1970.
homogeneous Smooth Earth,” Proc. IEE, 96, Part III: Oksman, J., H .G. Moller, and R. Greenwald, “Compari-
53–64, 1949. sons between Strong HF Backscatter and VHF Radar
Millm,an, G. H. Modern Radar, ed. R. S. Berkowitz, Wiley, Aurora,” Radio Sci., 14: 1121-1133, 1979.
New York, 1965. Oran, E.S. and T.R. Young, “Numerical Modeling of Iono-
Millman, G. H., “An Evaluation of HF Ionospheric Back- spheric Chemistry and Transport Processes ,“ J. Phvs.
scatter Echoes ,“ G.E. Tech. Information Series Rept. Chem., 81:2463, 1977.
No. R75EMH19, NOV 1975. Oran, E. S., T.R. Young, D.V. Anderson, T.P. Coffey,
Millman, G.H. and C.D. Bell, “Ionospheric Dispersion of P.C. Kepple, A. W. Ali., and D.F. Strobel, A Numerics/
an FM Electromagnetic Pulse ,“ IEEE Trans. Antennas Model of the Mid-Latitude Ionosphere, Naval Research
Propag., AP-19( 1), Jan 1971. Laboratory Memorandum Rept. No. 2839, 1974.
Millman, G. H. and G.M. Reinsmith, “An Analysis of the Orsak, L. E., L.H. Rorden, G.B. Carpenter, and B.P. Fick-
Incoherent Scatter-Faraday Rotation Technique for Ion- Iin, “VLF Propagation and Noise in the Ionosphere Ob-
ospheric Propagation Error Correction ,“ Rept. No. served by Sounding Rockets,” Final Report on Contract
R74EMH2, G. E., New York, Feb 1974. NASr-49(01 ), Stanford Research Institute, Menlo Park,
Minnis, C.M. and G.H. Bazzard, “A Monthly Ionospheric Calif., 1965.
Index of Solar Activity Based on F2-Layer Ionization Pappert, R. A., “A Numerical Study of VLF Mode Structure
at Eleven Stations,” J. Atmos. Terr. Phys., 18: 297, and Polarization Below an Anisotropic Ionosphere,” Ra-
1960. dio Sci., 3: 219–233, 1968.
Mitra, A. P., ionospheric E~ects of Solar Flares, D. Reidel, Pappert, R. A., “Effects of a Large Patch of Sporadic E on
Dordrecht, Holland, 1974. Nighttime Propagation at Lower ELF,” J. Atmos. Terr.
Morfitt, D.G. , “Effective Electron Density Distributions Phys., 42:417-425, 1980.
Describing VLF/LF Propagation Data,” Tech. Rept. 141, Pappert, R. A., “LF Daytime Earth Ionosphere Waveguide
Naval Ocean Systems Center, San Diego, Calif., 1977. Calculations ,“ Tech. Rept. 647, Naval Ocean Systems
Morfitt, D. G., J .A. Ferguson, and F.P. Snyder, “Numerical Center, San Diego, Calif., 1981.
Modeling of the Propagation Medium at ELF/VLF/LF,” Pappert, R. A. and J. E. Bickel, “Vertical and Horizontal

10-107
CHAPTER 10
VLF Fields Excited by Dipoles of Arbitrary Orientation Ramakrishnan, S., D. Bilitza, and H. Thiemann, Limita-
and Elevation,” Radio Sci., 35: 1445–1 452, 1970. tions of the IRI-78 Model, presented at COSPAR Gen-
Pappert, R. A. and W .F. Moler, “A Theoretical Study of eral Assembly, Bangalore, India, 1979.
ELF Normal Mode Reflection and Absorption Produced Rasmussen, J. E., P. A. Kossey, and E.A. Lewis, “Evidence
by Nighttime Ionospheres,” .J. Armos. Terr. Phys., 40: of an Ionospheric Reflecting Layer Below the Classical
103 I-1045, 1978. D region,” J. Geophys. Res., 85: 3037–3044, 1980.
Patenaude, J., K. Bibl, and B. W. Reinisch, “Direct Digital Rasmussen, J .E., P.A. Kossey, and J .P. Turtle, “VLF/LF
Graphics,” American Lab., 9:5-101, 1973. Pulse Reflections from Layers Below the Ionospheric
Paul, A. K., “Processing of Digital Monograms,” NOSC- D-Region,” AGARD Conf. Proc., 305:5- 1–5-10 ADA
TD-529, Naval Ocean Systems Center, San Diego, 113969, 1982.
1982. Rastogi, R. G., “Seasonal and Solar Cycle Variations of
Penndorf, R., “Geographic Distribution of Spread-F in the Equatorial Spread-F in the American Zone,” J. Atmos.
Arctic,” J. Geophys. Res., 67: 2274–2288, 1962. Terr. Phys., 42:593-597, 1980.
Perkins, F.W. and P.K. Kaw, “On the Role of Plasma Ratcliffe, J. A., “The Magneto-Ionic Theory and its Appli-
Instabilities in Ionospheric Heating by Radio Waves,” cation to the Ionosphere,” Cambridge University Press,
J. Geophys. Res., 76:282, 1971. London, 1959.
Perkins, F. W. al,l E.J. Valeo, “Thermal Self-Focusing of Ratcliffe, J .A., Sun, Earth and Radio, World University
Electromagnetic Waves in Plasma,” Phys. Res. Let?., Library, London, 1970.
32:1234, 1974. Rawer, K., “Propagation of Decameter Waves (HF-Bands)”
Perkins, F. W., C. Oberman, and E.J. Valeo, “Parametric in Dynamical and Chemical Coupling between the Neu-
Instabilities and Ionospheric Modification,” J. Geophys. tral and Ionized Atmosphere, edited by B. Grandal and
Res., 79: 1478, 1974. J.A. Holtlet, Reidel, Dordrecht, Holland, 1963.
Peterson, A. M., O.G. Villard, R.L. Leadabrand, and P.B. Rawer, K., “The Historical Development of Forecasting
Gallagher, “Regularly Observable Aspect-Sensitive Ra- Weather for Ionospheric Propagation of HF Waves,”
dio Reflections from Ionization Aligned with the Earth’s Radio Sci., 10(7): 669+79, 1975.
Magnetic Field and Located within the Ionosphere Layer Rawer, K., International Reference Ionosphere—IR1-79,
at Middle Latitudes,” J. Geophys. Res., 60: 497–5 12, edited by J. V. Lincoln and R.O. Conkright, World Data
1955. Center A, for Solar Terrestrial Physics (STP), NOAA,
Petrie, L.E. and E.S. Warren, “The Propagation of High Boulder, Colo., 1981.
Frequency Waves on the Winnipeg-Resolute Bay Oblique Rawer, K. and K. Suchy, “Radio Observations of the Iono-
Sounder Circuit ,“ in Ionospheric Radio Communica- sphere ,“ in Encyclopedia of Physics, edited by S. Flugge,
tions, Plenum, New York, 1968. Springer-Verlag, New York, 1967.
Picquenard, A., Radio Wave Propagation, MacMillan, 1974. Reddy, C.A. and M.M. Rae, “On the Physical Significance
Piggot, R. W. and K. Rawer, “URSI Handbook of Ionogram of the Es Parameters fbEs, fEs and foEs ,“ J. Geophys.
Interpretation and Reduction,” Report UAG-23 and UAG- Res., 73:215, 1968.
23A, World Data Center A for Solar-Terrestrial Physics, Reder, F. H., “VLF Propagation Phenomena Observed Dur-
NOAA, Boulder, Colo., 1972. ing Low and High Solar Activity,” XVI General As-
Pike; C. P., “An Analytical Model of the Main F-Layer sembly of URSI, Ottawa, Canada, 1969.
Trough ,“ AFGL-TR-76-O098, ADA026031, 1976. Reilly, M. H., L.O. Hamish, and J.M. Goodman, “Polar-
Pitteway, M. L. V., “The Numerical Calculation of Wave imetry Studies of Ionospheric Modifications by Rocket
Fields, Reflection Coefficients and Polarizations for Long Boosters ,“ NRL Mere. Rpt. No. 4517, Naval Research
Radio Waves in the Lower Ionosphere, I,” Phil. Trans. Laboratory, Washington, D. C., 1981.
Roy. SOC., A275: 219-241, 1965. Reinisch, B. W. and H. Xuequin, “Automatic Calculations
Phelps, A.D. R. and Sagalyn R.C. “Plasma Density Irreg- of Electron Density Profiles from Digital Ionograms. 1.
ularities in the High Latitude Topside Ionosphere,” J. Automatic O and X Trace Identification for Topside
Geophys. Res., 81:5151976. Ionograms, “ Radio Sci., 17:421434, 1982.
Pongratz, M. B. and G.M. Smith, “The LAGOPEDO Ex- Reinisch, B. W. and H. Xuequin, “Automatic Calculation
periments—An Overview ,“ Eos, Trans. AGU, 59:334, of Electron Density Profiles from Digital Ionograms. 3.
1978. Processing of Bottomside Monograms,” Radio Sci., 18:
Pope, J.H. and R.B. Fritz, “Observations of Artificially 477492, 1983.
Produced Scintillation using Satellite Transmissions,” Reinisch, B. S., J.S. Tang, W.N. Hall, and J. Buchau,
J. Geoph>>s. Res., 79: 1074, 1974. “Automatic Ionogram Processing Shows Seasonal Vari-
Potemra, T.A. and T.J. Rosenberg, “VLF Propagation Dis- ations in the Aurora] Ionosphere,” AFGL-TR-82-O 122,
turbances and Electron Precipitation at Mid-Latitudes,” ADA 119911, 1982a.
J. Geophys. Res., 78: 1572-1580, 1973. Reinisch, B. W., J. S. Tang, and R. R. Gamache, “Automatic
Proceedings IEEE, 57(6), 1969. Scaling of Digisonde Monograms,” Test and Evaluation
Radinov, Ya. S., “Autocorrelative Characteristics of Crit- Report, AFGL TR-82-0324, ADA 121666, 1982b.
ical Frequency and Ionization Density Fluctuations, ” Richmond, A.D. and S. V. Venkateswaran, “Geomagnetic
Geomag. Aeron., 3:985, 1963. Crochets and Associated Ionospheric Current Systems,”
Radio Science, 2( 10), 1967. Radio Sci., 6: 139, 1971.
Radio Science, 9(11 ), 1974. Rino, C.L. and S .J. Matthews, “On the Morphology of

10-108
 a
IONOSPHERIC RADIO WAVE PROPAGATION
Auroral Zone Radio Wave Scintillation,” J. Geophys. the Maximum Electron Density of the Ionospheric Fz
Res., 85:4139, 1980. Layer,” J. Radio Res. Labs. (Japan), 2; 86, 1955.
Rino, C .L. and J. Owen, “The Structure of Localized Night- Shlyger, 1.S., “Self-Modulation of a Powerful Electromag-
time Auroral Zone Enhancements,” J. Geophys. Res., netic Pulse Reflection from the Upper Layers of the
85:2941-2948, 1980. Ionosphere,” JETP [Engl. Transl. ), 19: 162, 1974.
Rino, C.L. and J. Owen, “On the Temporal Coheren&e Loss Singh, M., E. P. Szuszczewicz, and J .C. Holmes, “The
of Strongly Scintillating Signals,” Radio Sci., 16: 31, STP/S3-4 Satellite Experiment,” Equatorial F-Region
1981. Irregularities,” Froc-ionospheric Effects Symposium
Rino, C. L., R.C. Livingston, and S .J. Matthews, “Evi- 1981, Alexandria, Va. edited by J. M. Goodman, Naval
dence for Sheet-Like Auroral Ionospheric Irregulari- Research Lab., Washington D. C., 1982.
ties,” Geophys. Res. Lett., 5: 1039, 1978. Singleton, D. G., “The Morphology of Spread F Occumence
Rishbeth, H. and O.K. Garnott, ]titroduction to Ionospheric over Half a Sunspot Cycle,” J. Geophys. Res., 73:
Physics, Academic Press, 1969. 295–308, 1968.
Rosenberg, N. W. (cd.) PROJECT FIREFLY 1962-1963, Singleton, D. G., “Dependence of Satellite Scintillations on
AFCRL-64-364, 1964. Zenith Angle and Azimuth,” J. Atmos. Terr. Phys., 32:
Rosich, R.K. and W .B. Jbnes, “The Numerical Repre- 789–803, 1970.
sentation of the Critical Frequency of the F] Region Singleton, D. G., “Transionsopheric Propagation Predic-
of the Ionosphere,” OT Rept. 73-22, Boulder, Colo., tion. An Improved Ionospheric Irregularity Model ,“ STPP,
1973. 4: D1-1 (see R.F. Donnelly, cd.), 1980.
Rote, D. M., “Proceedings of the Workshop on Ionospheric Sinno, K. and M. Kan, “Ionospheric Scintillation and Fluc-
and Magnetospheric Effects of Satellite Power Systems tuations of Faraday Rotation Caused by Spread-F and
(SPS),” Conf. Rept. Nb. 7807115, U.S. Department of Sporadic-E over Kokobunji, Japan,” J. Radio Res. Lab.,
Energy, December 1979. 27: 53–77, 1980.
Rufenach, C. L., “Radio Scintillation of Steliar Signals dur- Somrnerfeld, A., “The propagation of Waves in Wireless
ing Artificial Ionospheric Modification,” J. Geophys. Telegraphy,” Ann. Physik, 28:665, 1909.
Res., 78:5611, 1973. Sprenger, K. and K. Glode, “Some Propefiies of Radio
Rush, C. M., “Improvements in Ionospheric Fo~ecasting Ca- Aurorae in Medium Latitudes,” J. Atmos. Terr. Phys.,
pability,” AFCRL-72-0138, AD742258 1972. 26: 193–198, 1964.
Rush, C. M., “An Ionospheric Observation Network for use Storey, L. R. O., “An Investigation of Whistling Atmos-
in Short-Tern Propagation Predictions ,“ Telecomm. J., pherics,” Phil. Trans. Roy. Sot., London, A246:
43:544, 1976. 113–141, 1953.
Rush, C.M. and J. Gibbs, “Predicting the Day-to-Day Vari- Stratton, J. A., Electromagnetic Theo~, McGraw-Hill, New
ability of the Mid-Latitude Ionosphere for Application York, 1941.
to HF Propagation Predictions,” AFCRL-TR-73-0335, Strobel, D.F. and M.B. McElroy, “The Fz Layer at Middle
AD764711 1973. Latitudes,” Planet. Space Sci., 18:1181, 1970.
Rush, C. M., M. Pokemp~er, D.N. Anderson, F.G. Sew- Stubbe, P., “Simultaneous Solution of the Time Dependent
ard, and J. Perry, “Improving Ionospheric Maps using Coupled Continuity Equations, and Equations of Motion
Theoretically Derived Values of foF2,” Radio Sci., 18: for a System Consisting of a Neutral Gas, and Electron
95, 1983. Gas, and a four Component Ion Gas,” J. Atmos. Terr.
Russell, C.T. and M.J. Rycroft (eds.), Acfive Experiments Phys., 32:865, 1970.
in Space Plasmas, Proc. Syti. 9, COS~AR XXIII, Bu- Svennesson, J., “Effects on VLF ~opagation During Iono-
dapest, 2–14 June 19?0, Pergamon Press, 1980. spheric Substoims,” J. Aws. Terr. Phys., 35:761-773,
Sagalyn, R. C., M. Smiddy, and J. Wisnia, “Measurements 1973.
and Interpretation of Ion Density Distribution in the Swenson, E. M., “Aspect-Sensitive Reflections from Ioni-
Daytime F-Region,” J. Geophys. Res., 68: 199, 1963. zation Itiegularities in the F-Region,” J. Atmos. Terr.
Sagalyn, R. C., M. Smiddy, and M. Ahrned, “High Lati- Phys., 34: 1469–1476, 1972.
tude Irregularities in the Topside ionosphere based on Tascione, T. F., T. W. Flatte~, V.G. Patterson, J .A. Secan,
ISIS—Thermal Probe Data,”J. Geophys. Res., 79:4252, and J.W. Taylor, Jr., “Ionospheric Modelling at Air
1974. Force Global Weather Central,” STPP, 1:367 (see R.F.
Saltpeter, E. E., “Electron Density Fluctuations in a Plasma,” Donnelly, Ed.) 1979.
Phys. Rev., 120: 1528, 1960; also J. Geophys. Res., Tellegen, B. D. H., “Interaction Between Radio Waves,”
65:1851, 1960. Nature, 131:840, 1933.
Shea, M. A., S.A. Militello, and H.E. Coffey, “Directory Thomas, J.O. and M.K. Andrews, “Transpolar Exospheric
of Solar-Terrestrial physics Monitoring Stations ,“ Plasma,” J. Geophys. Res., 73:7407-7417, 1968.
MONSEE Special Publication No. 2, AFGL TR 84- Thomas, J.O. and M.H. Rycroft, “The Exospheric Plasma
0237,1984. tig the Jntemational Years of the Quiet Sun,” Pfanet.
Shellman, C. H., “Electron Density Distributions in the Lower Space Sci., 18:41--63, 1970.
Ionosphere with Associated Error Limits Derived from Theme, G.D. and F. W. Perkins, “Production of Ionospheric
VLF and LF Sounder Data,” Radio Sci., 5: 1127–1 135, Striations by Self-Focusing of Intense Radio Waves,”
1970. Phys. Rev. Lett., 32: 1238, 1974.
Shimazaki, T., “World Wide Variations in the Height of Thompson, R.L. and J. A. Secan, “Geophysical Forecasting

10-109
CHAPTER 10
at AFGWC,” STPP, 1: 355 (see R.F. Donnelly, cd.), Electromagnetic Waves from an Electrical Point Source
1979. Round a Finitely Conducting Sphere,” Phil. Msg. Series
Thomson, J .J., Conduction of Electrici~ Through Gases, 7, 24:141-176, 1937; 24:825-864, 1937; 25:817-834,
p. 321, Cambridge University Press, London, 1906. 1938; and 26:262-275, 1938.
Thrane, E.V. (Chairman), “D-Region Predictions,” STPP, Vickery, J. F., C. L. Rino, and T.A. Potemra, “Chatanika
2:573 (see R.F. Donnelly, cd.), 1979. Triad Observations of Unstable Ionization Enhance-
Tinsley, B. A., A.B. Christensen, J.A. Bittencourt, p.D. ments in the Auroral F-Region,” Geophys. Res. Lett.,
Angreji, and H. Takahashi, “Excitation of oxygen per- 7:789-792, 1980.
mitted Emissions in the Tropical Nightglow,” J. Geo- Videberg, J. 1. and G.S. Sales, “Long Range Survivable MF
phvs. Res., 78: 1174, 1973. Radio Communication Study Using High Altitude Whis-
Tither~dge, J .E., “Periodic Disturbances in the Ionosphere,” pering Gallery Modes,” AFCRL-TR-73-0552,
J. Geophys. Res., 73: 243–252, 1968. AD774989, 1973.
Trizna, D. B. and J .M. Headrick, “Ionospheric Effects on Villard, O.G. Jr., “The Ionospheric Sounder and Its Place
HF Over-the Horizon Radar,” in Effects of the Iono- in the History of Radio Science ,“ Radio Sci., 11: 847–860,
sphere on Radiowave Systems, Proc. Ionospheric Effects 1976.
Symposium 1981, Alexandria, Va. edited by J.M. Good- von Flotow, C. S., “Ionospheric Forecasting at Air Force
man, Naval Research Lab., Washington, D. C., 1982. Global Weather Central ,“ in Effects of the Ionosphere
Tsunoda, R. T., “On the Spatial Relationship of 1-m Equa- on Space and Terrestrial Systems, edited by J.M. Good-
torial Spread F Irregularities and Plasma Bubbles,” J. man, U.S. Government Printing Office, Stock No. 008-
Geophys. Res., 85: 185, 1980. 051-00069-1, Jan 1978.
Tsunoda, R. T., R.C. Livingston, J .P. McClure, and W .B. Wait, J. R., “Radiation from a Ground Antenna,” Can. J.
Hanson, “Equatorial Plasma Bubbles” Vertically Elon- Tech., 32: 1-9, 1954.
gated Wedges from the Bottomside F-Layer,” J. Geo- Wait, J. R., “Mixed-Path Ground Wave Propagation, 1. Short
phys. Res., 387:9171, 1982. Distances,” J. Res. NBS, 57: 1–15, 1956.
Turner, J. F., “The Development of the Ionospheric Index Wait, J .R., “Excitation of Surface Waves on Conducting,
T,” Ionospheric Prediction Science Series R Reports, Stratified, Dielectric-Clad, and Corrugated Surfaces,”
IPS-R12, Sydney, Australia, 1968. J. Res. NBS, 59:365-377, 1957.
Turner, J .F. and P.J. Wilkinson, “A Weekly Ionospheric Wait, J. R., “The Electromagnetic Fields of a Horizontal
Index,” STPP, 1:279 (edited by R.F. Donnelly), 1979. Dipole in the Presence of a Conducting Half-Space,”
UAG Report 10, Atlas of Ionograms, (avail. World Data Can. J. Phys., 39:1017-1028, 1961.
Center A, U.S. Dept. of Commerce, NOAA/EDIS, Wait, J. R., Electromagnetic Waves in StratijedMedia, Per-
Boulder, Colo.), 1970. gamon Press, New York, 1970.
UAG Report 23, URSI Handbook of Ionogram Interpre- Wait, J. R., (cd.), Electromagnetic Probing in Geophysics,
tation and Reduction, (avail. World Data Center A, U.S. The Golem Press, Boulder, Colo., 163-207, 1971.
Dept. of Commerce, NOAA/EDIS, Boulder, Colo. ), Wait, J. R., (cd. ) “Special Issue on Extremely Low Fre-
1972. quency (ELF) Communications,” IEEE Trans. Com-
UAG Report 23a, URSI Handook of Ionogram Interpreta- mun., COM-22, 1974.
tion and Reduction, Revision of Chapters 1A, (avail. Wait, J. R., ‘&Propagation of ELF Electromagnetic Waves
World Data Center A, U.S. Dept. of Commerce, NOA~ and Project Sanguine/Sea farer,” IEEE J. Ocean. Eng.,
EDIS, Boulder, Colo. ), 1978. OE-2: 161-172, 1977.
UAG Report 50, High Latitude Supplement to the URSI Wait, J .R. and H.H. Howe, “Amplitude and Phase Curves
Handbook of Ionogram Interpretation and Reduction, for Ground Wave Propagation in the Band 200 Cycles
(avail. World Data Center A, U.S. Dept. of Commerce, per Second to 500 Kilocycles,” NBS Circular 574, Na-
NOAA/EDIS, Boulder, Colo. ), 1975. tional Bureau of Standards, Washington, D. C., U.S.
UAG-85 Catalogue of Vertical Ionosphere Soundings Data, Government Printing Office, 1956.
(avail. World Data Center A, U.S. Dept. of Commerce, Wait, J. R. and A. Murphy, “Influence of a Ridge on the
NOAA/EDIS, Boulder, Colo.), 1982. Low Frequency Ground Wave,” J. Res. NBS, 58: 1–5,
Uffelman, D. R. and L.O. Hamish, “Initial Results from 1957.
HF Propagation Studies During Solid Shield,” Naval Wait, J .R. and A. Murphy, “Further Studies of the Influence
Research Lab. Memorandum Report 4849, 1982. of a Ridge on the Low Frequency Ground Wave,” J.
Utlaut, W. F., “An Ionospheric Modification Experiment Res. NBS, 61:57-60, 1958.
using Very High Power, High-Frequency Transmis- Wait, J.R. and K.P. Spies, “Characteristics of the Earth-
sion,” J. Geophys. Res., 75:6402, 1970. Ionosphere Waveguide for VLF Radio Waves,” NBS
Utlaut, W .F. and E.J. Violette, “Further Observations of Tech. Note 300, National Bureau of Standards, Wash-
Ionospheric Modification by a High-Powered HF Trans- ington, D. C., U.S. Government Printing Office, 1964.
mitter,” J. GeophYs. Res., 77:6804, 1972. Wait, J.R. and L.C. Walters, “Curves for Ground Wave
Utlaut, W .F., E.J. V“iolette, and A.K. Paul, “Some Iono- Propagation Over Mixed Land and Sea Paths,” IEEE
sonde Observations of Ionospheric Modification by Very Trans. Antennas Propag., 11:3845, 1963.
High Power High-Frequency Ground-Based Transmis- Wand, R. H. and J. V. Evans, “Morphology of Ionospheric
sion,” J. Geophy.s. Res., 25:6429, 1970. Scintillation in the Auroral Zone, “ in Efect of the Iono-
Van der Pol, B. and H. Bremmer, “The Diffraction of sphere on Space systems and Communications, edited

10-110
 IONOSPHERIC RADIO WAVE PROPAGATION

by J. Goodman, NTIS CSCL 04/1 N75-307 14, Naval Winkler, C., “Radio Wave Guidance at VHF Through Equa-
Research Laboratory, Washington, D. C., 1975. torial Plasma Bubbles,” J. Atmos. Terr. Phys., 43:307,
Wand, R. H. and M. Mendillo, “Incoherent Scatter Obser- 1981.
vations of an Artificially Modified Ionosphere, ” J. Geo- Wright, J. W., “Kinesonde Observations of Ionosphere Mod-
phys. Res., 89:203, 1984. ification by Intense Electromagnetic Fields from Platte-
Watson, G .N., “The Transmission of Electric Waves Round ville, Colorado,” J. Geophvs. Res., 78: 5622, 1973.
the Earth,” Proc. Roy. Sot., 95: 546–563, 1919. Wright, J. W. and A. K. Paul, “Toward Global Monitoring
Watt, A. D., VLF Radio Engineering, Pergamon Press, New of the Ionosphere in Real Time by a Modem Ionosonde
York, 1967. Network: The Geophysical Requirements and Techno-
Weaver, P. F., “Backscatter Echoes from Field-Aligned Ir- logical Opportunity,” NOAA Special Report, Space En-
regularities in the F-Region,” J. Geophys. Res., 70: vironment Lab., July 1981.
5425–5432, 1965. Wright, J. W. and M .L. V. Pitteway, “Real-Time Data Ac-
Weber, E.J. and J. Buchau, “Polar Cap F-Layer Auroras,” quisition and Interpretation Capabilities of the Dyna-
Geophys. Res. Lett. 8:125, 1981. sonde, 1. Data Acquisition and Real Time Display, 2.
Weber, E. J., J. Buchau, R.H. Eather, and S.B. Mende, Determination of Magnetosonic Mode and Echolocation
“North-South Aligned Equatorial Airglow Depletions,” using a Small Spaced Receiving Array ,“ Radio Sci., 14,
J. Geophys. Res., 83:712, 1978. 815–835, 1979.
Weber, E. J., J. Buchau, and J.G. Moore, “Airborne Studies Wright, J. W. and M. L. V. Pitteway, “Data Processing for
of Equatorial F-Layer Ionospheric Irregularities,” J. the Dynasonde: The Dopplionogram ,“ J. Geophys. Res.,
Geophys. Res., 85:4631-4641, 1980. 87:1589, 1982.
Weber, E. J., H.C. Brinton, J. Buchau, and J.G. Moore, Yeh, K. C., “Traveling Ionospheric Disturbances as a Di-
“Coordinated Airborne and Satellite Measurements of agnostic Tool for Thermospheric Dynamics,” J. Geo-
Equatorial Plasma Depletions,” J. Geophys. Res., 87: phys. Res., 77: 709–719, 1972.
503–510, 513, 1982. Yeh, K.C. and C. H. Liu, “Ionospheric Effects on Radio
Wemik, A. W., C.H. Liu, and K.C. Yeh, “Model Com- Communication and Ranging Pulses,” IEEE Trans. An-
putations of Radio Wave Scintillation caused by Equa- tennas Propag., AP-27(6), Nov 1979.
torial Ionospheric Bubbles,” Radio Sci., 15:559, 1980. Yeh, K.C. and C.H. Liu, “Radio Wave Scintillations in the
Weyl, H., “Propagation of Electromagnetic Waves Above Ionosphere, “ Proc. IEEE, 70, 1982.
a Plane Conductor,” Ann. Physik, 60:481 –500, 1919. Yeh, K. C., H. Soicher, and C. H. Liu, “Observations of
Whalen, B. A., et al., “Preliminary Results from Project Equatorial Ionospheric Bubbles by the Radio Propaga-
Waterhole—An Auroral Modification Experiment ,“ Cam. tion Method,” J. Geophys. Res., 84:6589, 1979.
J. Phys., 59: 1175,1981. Zacharisen, D. M., “Space-Time Correlation Coefficients
Whitney, H. E., C. Malik, and J. Aarons, “A Proposed for use in Short-Term Ionospheric Predictions,” NBS
Index for Measuring Ionospheric Scintillation,” Planet. Report 8811, 1965.
Space Sci., 17:1069-1073, 1969. Zevakina, R.A. E. V., Lavrova, and L.N. Lyakhova, Man-
Whitney, H. E., J. Aarons, R.S. Allen, and D.R. Seemann, ual on Short-Term Predictions of Ionospheric Geomag-
“Estimation of the Cumulative Amplitude Probability netic Storms and Radio-Propagation Forecasting, (text
Distribution Function of Ionospheric Scintillations ,“ Ra- in Russian) Nauka, Moscow, 1967.
dio Sci., 7: 1095–1 104, 1972. Zinn, J. and C.D. Sutherland, “Effects of Rocket Exhaust
Wilkinson, P. J., “Prediction Limits for foF2,” STPP, 1: Products in the Thermosphere and Ionosphere,” Space
259 (see R.F. Donnelly cd.), 1979. Solar Power Rev., 1: 109, 1980.
Wilkinson, P. J., “A Comparison of Monthly Indices of the Zinn, J., C.D. Sutherland, S. Stove, and L. Duncan, “Iono-
Ionospheric F-Region,” Ionospheric Prediction Service spheric Effect of Rocket Exhaust Products: HEAO-C
Series R Reports, IPS-R41, Sydney, Australia, 1982. and Skylab,” J. Atmos. Terr. Phvs., 44: 1143, 1982.

10-111
 Chapter 11

SOLAR RADIO EMISSION

W. R. Barron
E. W. CIiver
J. P. Cronin
D. A. Guidice

Since the first detection of solar radio noise in 1942, If the frequency f is in cycles per second, the wavelength
radio observations of the sun have contributed significantly A in meters, the temperature T in degrees Kelvin, the ve-
to our evolving understanding of solar structure and pro- locity of light c in meters per second, and Boltzmann’s
cesses. The now classic texts of Zheleznyakov [1964] and constant k in joules per degree Kelvin, then Br is in W
Kundu [1965] summarized the first two decades of solar m-2Hz-lsr-1. Values of temperatures T~ calculated from
radio observations. Recent monographs have been presented Equation ( 11.1) are referred to as equivalent blackbody tem-
by Kruger [1979] and Kundu and Gergely [1980]. perature or as brightness temperature defined as the tem-
In Chapter 1 the basic phenomenological aspects of the perature of a blackbody that would produce the observed
sun, its active regions, and solar flares are presented. This radiance at the specified frequency.
chapter will focus on the three components of solar radio The radiant power received per unit area in a given
emission: the basic (or minimum) component, the slowly frequency band is called the power flux density (irradiance
varying component from active regions, and the transient per bandwidth) and is strictly defined as the integral of B~Q,
component from flare bursts. between the limits f and f + Af, where Q, is the solid angle
Different regions of the sun are observed at different subtended by the source. In solar radio astronomy the re-
wavelengths. At millimeter wavelengths, the radiation is lationship used is
from the photosphere. Centimeter wavelength radiation
originates in the chromosphere and low corona. Decimeter F. = BrQ, = 2kT~Q,h -2, (11.2)
and meter wavelengths have their origin at increasing heights
in the corona; at meter wavelengths the observed radiation
comes from heights ranging from 100000 to 700000 km where the apparent or disk temperature T~ is that temperature
above the photosphere. For receiving equipment on the e@, which a uniform source of the same angular size as the solar
the low-frequency limit for observation is the frequency at optical disk must have in order to produce the power flux
which radio waves are reflected by the ionosphere (for prac- density F, received from the sun. Values of power flux
tical purposes, around 20 MHz). The high-frequency limit density are usually given in solar flux units (1 solar flux
is set by absorption of radiation by atmospheric oxygen and unit, sfu = l&22 W m-z HZ-l).
water vapor. Recently, radio experiments on satellites have The power P, received at the antenna due to solar ra-
observed hectometric wavelength (<2 MHz) emission that diation is given by:
originate at heights >10 solar radii (Ro). This leaves only
the frequency range form 2 to 20 MHz, corresponding to
P, = F,A=, (11.3)
emission heights of 2– 10 RO, unexplored by radio methods.

where A= is the effective area of the antenna. P. is also

11.1 BASIC DEFINITIONS conveniently expressed in terms of the effective antenna
temperature T~ corrected for any RF losses. T~ is defined
If the sun radiated only,as a thermal source, the emitted by
energy density would vary with frequency and temperature
according to Planck’s radiation law. In the radio region, the
P, = kT,Af. (11.4)
Rayleigh-Jeans approximation for blackbody radiation is
valid; the brightness, radiance per unit bandwidth, is
T~ is readily measured with suitably calibrated instruments.
Bf = 2kT~c-2 = 2kTA-2. (11.1) The equation for calculating the solar power flux density

11-1
CHAPTER 11
from a given antenna temperature measured at a given in- wavelengths that solar radio emission approximates a ~
stallation is K blackbody. At wavelengths longer than 1 cm, the equiv-
alent blackbody temperature ranges between 104 to 106 K
for the spotless sun and from 104 to 10’0 K for the disturbed
F, = 2kTA+, (11.5)
e sun depending on the condition of the sun and the time in
relation to the 11-year sunspot cycle. Meter wavelengths
where A. is the effective area of the antenna in square are characterized by much burst activity; thus, the basic (or
meters, and L is a dimensionless correction factor related minimum) sun temperatures at these wavelengths are de-
to the antenna response shape and to the diameter and tem- termined by making observations over a period of weeks or
perature distribution across the source. The value of L is months. The low temperature to which the sun periodically
unity only if the antenna half-power beamwidth is large returns, but never goes below, during this period is taken
compared to the source. L exceeds unity when the ratio of to be the zero-sunspot value; it is of the order of 106 K.
the antenna half-power beamwidth to the solar angular di-
ameter drops below about five; thus, it is desirable to use
a parabola small enough so the half-power beamwidth is 11.3 THE SLOWLY VARYING
more than five times the solar angular diameter. Once the COMPONENT
solar flux density is known, the apparent temperature and
the brightness of the solar disk may be calculated om The slowly varying component (SVC) exhibits a well
Equation (11 .2). defined 11-year cycle variation and a 27-day solar-rotation
variation, since this emission originates principally in co-
ronal condensations overlying active regions, and is well-
11.2 THE MINIMUM (ZERO-SUNSPOT correlated with sunspot number. The routine daily mea-
COMPONENT surement of the combination of the basic (or minimum)
component and the SVC of solar radio emission is referred
The standard method for obtaining the basic radio flux to as the quiet-sun flw densi~. The SVC of the sun as a
density of the unspotted sun is to make a scatter plot of whole is obtained by subtracting the basic component from
solar temperature at a given frequency against the projected the quiet-sun flux density. The SVC of individual active
sunspot area; the extrapolation of the curve that best fits regions can be obtained by either eclipse observations or
these data to zero sunspot area determines the minimum or interferometric measurements.
basic flux density at that frequency. By doing this for all
frequencies, one determines the spectrum of the basic com-
ponent of solar radiation. Figure 11-1 shows the distribu- 11.4 THE BURST COMPONENT
tions of power flux density for the sun and for black bodies
at various temperatures. It is only at millimeter and shorter During solar flares (Chapter 1) there may be large in-
creases (bursts) in radio emission lasting anywhere from a
few seconds to several hours. These bursts originate by
Ip 10p IOOP I mm I cm 10cm bremsstrahhtng, gyrosynchrotron, and plasma radiations.
10-’0 Characteristics of the bursts vary with wavelength. Bursts
10-’2 in the meter-wave range (12 m to about 50 cm) are classified
by spectral type. No spectral classification exists for the
10-K decimeter or centimeter wave regions.

10-’6
11.4.1 Meter-Wave Range (25 -500 MHz).
10-’8

10-m Most information on solar bursts in the range from 12

m to about 50 cm is obtained from swept-frequency obser-
l(j22 vations. Dynamic spectra are displayed on a cathode-ray
tube and recorded photographically as a series of intensity-
I(--24
modulated traces that give intensity as a brightening in the
‘v 10-7 10-6 105 104 103 10-210-’ I 10 100
frequency-time plane. Figure 1I-2 is an illustration of ideal-
WAVELENGTH (m)
ized dynamic spectra of various types of bursts. These spec-
tral types are discussed below in the order of their occurrence
Figure 11-1. Solar spectmm and spectra of blackbody radiation at various
[emperdtures, The solar power-flux density (power per unit in the flare event.
area per unit bandwidth) is plotted against wavelength. Type III bursts are the most common type of solar radio

11-2
 SOLAR RADIO EMISSION
optical flare. The storm continuum can last for several hours
and often degenerates into the Type I noise storms whose
durations range from hours to days. In contrast to the flare
continuum and moving Type IV emissions that are only
weakly polarized, storm continuum is strongly polarized in
the ordinary mode. This suggests plasma radiation as the
source of the storm continuum, while gyrosynchrotron ra-
FLARE STARTS diation from energetic electrons spiraling in weak coronal
TIME (minutes)

magnetic fields are generally cited as the source of the flare

continuum and moving Type IV emissions.
Figure 11-2. Idealized illustration of the record of a complete solar radio
outburst as recorded by a dynamic spectrograph at meter Type I events are distinguished from the relatively smooth
wavelengths, broad-band Type IV emission by the presence of a great
number of short (-1 s) intense bursts superimposed on the
activity. They can occur either singly (duration -5 s) or in background continuum. These short intense bursts are dis-
groups. Only about one-third of Type III bursts are asso- tributed more or less randomly over the frequency range of
ciated with flares. When associated with flares, however, the underlying continuum. Both the background continuum
the timing agreement with flare impulsive hard x-ray and and the superimposed bursts are strongly circularly polar-
microwave emission is often quite good (within seconds). ized, usually in the ordinary mode. Type I radiation appears
Type III bursts are caused by streams of -100 keV electrons to be more closely associated with certain active regions
propagating outward through the solar atmosphere and ex- than with flares, although they can be flare-initiated. At
citing plasma waves. Because of their relatively high drift present, the mechanism and origin of Type I emission are
rates (20 MHz/s) to lower frequencies, they are referred to not well understood.
as fast drift bursts.
Type V bursts consist of a broad band continuum of
short duration (-1 rein) that is preceded by a Type III burst 11.4.2 Decimeter-Wave Range
and accompanied by centimeter-wave and hard x-ray bursts. (5M -2000 MHz).
The Type V burst may indicate the presence of a particularly
rich stream of electrons, part of which is trapped in the At decimeter wavelengths the emission is highly variable
corona and becomes visible either through gyrosynchrotron and complex. Rapid time structures (several peaks per min-
radiation or plasma waves. ute) are often observed in the time profiles of decimetric
The Type II burst, or slow-drift (-0.2 MHz/s) burst is radio bursts observed at discrete frequencies, and the re-
presumed due to the presence of a shock wave propagating lationship of these fast structures to the source of the smoother
outward through the solar corona with a characteristic ve- emission observed at centimeter wavelengths is not clear.
locit y of 1000 km/s. The disturbance excites plasma waves Individual peaks in complex events are often strongly cir-
at the local plasma frequency. In the spectrograph record, cularly polarized in the ordinary mode.
these bursts appear as a narrow band of intense radiation
that drifts gradually and often irregularly from high to low
frequency. About 60% of all Type II bursts show emission 11.4.3 Centimeter-Wave Range
at the second harmonic. Type 11bursts are closely associated (2000 -35000 MHz).
with solar flares.
Type IV emission has at least three distinct components; Solar emission in the 15–1 cm range does not show as
these components can not be separated on the spectrograph rapid fluctuations as emission in the meter and decimeter
record and can only be distinguished by interferometers. ranges. There appear to be at least two basic morphological
Flare continuum is the broad band emission Occurnng at types of centimeter-wave bursts. The first of these is the
meter and decameter wavelengths during the flare impulsive simple impulsive event that reaches a maximum peak-flux
phase. Moving Type IV bursts involve a variety of forms, density ranging from 101 to > I& sfu in a few minutes.
although three basic types have been recognized. These are Impulsive bursts are interpreted in terms of gyrosynchrotron
the magnetic arch, the advancing front, and the isolated emission. Complex bursts may consist of several impulsive
source (ejected plasma “blob”). Both the flare continuum events in sequence, although the appearance of a relatively
and moving Type IV emissions have durations <10-60 min. smooth broad band microwave Type IV component in the
While the moving Type IV burst travels outward through later stages of many complex events may indicate an ad-
the corona and can reach. heights of -106 km above the ditional or prolonged acceleration process. The second of
photosphere, the final stage of Type IV emission, the storm these, the gradua[ rise andfall microwave burst, may occur
continuum, originates low in the corona near the corre- by itself or may follow an impulsive event (monotonic decay
sponding plasma level and directly above the site of the only), in which case it is referred to as a post-burst event.

11-3
CHAPTER 11
quencies) with different electron energy distributions and
02 JUNE 1974 different magnetic fields [Kundu and Vlahos, 1982].

MANILA T
OBSERVATORY 1,250
11.5CORRECTIONS TO QUIET-SUN AND
8800
BURST-FLUX DENSITIES
1
In the final two sections of this chapter, charts and tables
of burst and quiet-sun flux densities from Sagamore Hill
Solar Radio Observatory are presented. It is appropriate to
1f
discuss errors of absolute calibrations in these measure-
I 1 ments. In 1973, a report [Tanaka et al., 1973] by an absolute
calibration working group fomed by Commission V of URSI
was published. It contained corrections for Sagamore Hill
for the years 1966-1971. For the years 1972–1976, cor-
rection factors were taken from the IAU’s Quarterly Bulletin
of Solar Activity. For 1977– 1979, correction factors were
derived by extrapolation from previous years and compar-
ison of the Sagamore Hill quiet-sun flux densities with those
of other solar-patrol observatories. For 245 MHz, a sub-
1415 stantial correction factor (1.55) was applied to all the data
as a result of an absolute calibrations measurement (using
the radio source Cassiopeia A) carried out at Sagamore Hill
in 1980. The multiplicative flux-density correction factors
for the five frequencies for which data are presented are
listed in Table 11-1. Before processing the data presented
606 below, these correction factors were applied to the Sagamore
Hill burst and quiet-sun data.
I 1 I I 1

0420 0430
UNIVERSAL TIME 11.6OUIET-SUN FLUX-DENSITY
MEASUREMENTS
Figure 11-3. Thetime-intensity profile ofanimpulsive microwave burst.
Solar microwave emission correlates well with solar EUV
flux [Forbes and Straka, 1973], a fact which makes it useful
Peak flux densities of these gradual events seldom exceed as an input to ionospheric models in lieu of the more difficult
50 sfu; these events are generally explained in terms of to obtain sunspot number and EUV flux. For this reason,
bremsstrahlung from a Maxwellian distribution of electrons. the 1965 edition of the Handbook of Geophysics included
Examples of the time profiles of microwave bursts observed a table of the daily quiet-sun (non-bursting) flux densities
at Sagamore Hill Solar Radio Observatory are contained in observed at 2800 MHz (10.7 cm wavelength) by the Na-
the burst atlas compiled by Barronet al. [1980]. Atypical tional Research Council in Ottawa, Canada from 1947 to
impulsive event is presented in Figure 11-3. mid- 1963. We continue the table in this handbook using
Instantaneous spectra of events in the centimeter fre- Ottawa data through 1965 and Sagamore Hill data thereafter.
quency range are relatively smooth and tend not to have In addition, the 8800 MHz quiet-sun flux-density values
narrow-band emission features. For moderate sized events from Sagamore Hill are included.
(> 100 sfu), the peak-flux-density spectral maximum of the In Table 11-2 we present the observed daily solar flux
emission generally occurs at frequencies >9 GHz. For the density value measured at 2800 MHz at Ottawa for the years
largest events with centimeter wave peak flux densities >1000 1963–1965 and the 2695 and 8800 MHz flux density values
sfu, the emission usually extends to the meter-wave range measured at the Sagamore Hill Radio Observatory from
where it often exhibits a second spectral maximum with a January 1966 through December 1979. It should be noted
minimum occurring at the intermediate wavelengths. This that these are observed values and not values adjusted to 1
U-shaped spectral signature (Figure 11-4) is thought to re- AU. The values are taken at local noon at the Sagamore
flect the fact that there are two different sources of burst Hill meridian. This is done so that the radiation will have
radiation (one at centimeter waves and one at lower fre- passed through the shortest path in the earth’s atmosphere

11-4
R
SOLAR RADIO EMISSION
1 1 I I I 1 111 , 1 1 I 1 1 Ill 1 1 I 1 1 [ lr

1981 APRIL 15

●
Ill
10 1 1 1 1 1 1 1 1 1 1 1 1 1 11 1 1 1 1 I 1 I 1

0.1 1.0 10.0 100.0

FREQUENCY (GHz)

Figure 11-4. The peak flux-density s~ctrum of a large microwave burst. The U-shape is characteristic for large microwave bursts with peak flux density
> 1000 Sfu.

and be subject to the minimum atmospheric attenuation. In Note the asterisk (*) between the 2695 MHz flux-density
some instances this meridian-transit observation is not pos- value and the 8800 MHz flux-density value on some days.
sible because a solar radio burst is in progress at that time. This indicates that there were adverse weather conditions,
When this happens, the observation is taken as, soon as it usually rain or snow, present when the observations were
has been determined that the radio burst has ended. In vir- being made. These conditions will cause the signals received
tually all cases this was within an hour of the time of me- to be further attenuated. The asterisk indicates that some
ridian transit so that the added attenuation due to the in- adjustment has been made to the observed value to com-
creased atmospheric path length was negligible. pensate for this problem.

Table 11-1. Calibration correction factors, Sagamore Hill, 19W1979.

Year 245 MHz 610 MHz 1415 MHz 2695 MHz 8800 MHz
1966 — 0.91 1.16 0.90 0.91
1967 — 0.91 1.16 0.90 0.91
1968 — 0.91 1.16 0.90 0.95
1969 1.55 0.91 1.16 0.94 1.00
1970 1.55 0.91 1.16 0.94 1.00(2)
1971 1.55 0.91 1.16 0.94 1.00(2)
1972 1.55 0.91 1.08 0.93 0.95
1973 1.55 0.91 1.03 0.91 0.90
1974 1.55 0.91 1.05 0.90 0.90
1975 1.55 0.91 1.05 0.90 0.87
1976 1.55 0.92 1.08 0.90 0.85
1977 1.55 1.00 1.M 0.88 0.88
1978 1.55 1.00 1.06 0.85 0.87
1979 1.55 1.00 1.06 0.85 0.84
Nores: (1) For the ~riod 9 June to 3 I August 1967, all 2695 MHz burst and daily flux-density values should be multiplied by 1.13.
(2) For the period 1 November 1970 to 31 August 1971, all 8800 MHz burst and daily flux-density values should be multiplied by 1.14,

11-5
 CHA~ER 11

11.7 SOLAR RADIO-BURST Table 11-2a. Daily values of observed solar flux densitv at 2800 MHz
recorded at the Algonquin Radio Observato’~ (ARO) of the
CLIMATOLOGY BASED ON National Researeh Council. Ottawa, Canada.
1966-1978 SAGAMORE HILL
okY Jti
OBSERVATIONS FE~-WWJW JW-66P%T- 02C

1963 ~

01 7* 72 e2 w 87 74 87 79
In addition to its role in maintaining the ionosphere, the n2
03 77
m
78
74
74
e2
al
el
al 7e
e7 73
74
m
e3
m
?9
sun is the most powerful natural source of radio interference w B w 70 e2 79 7e n 03 77
w 77 ea 72 M 7e 7s 74 w 76

to ground-based or near-earth (aircraft or satellite) radio 06

*7
77
77
8s
8*
7e
20
97
m
77
w
77
77
74
76
7s
76
76
77
09 76 ez et % 90 77 n 75 77
receivers. Therefore, it is of interest to designers and op- G9 7s 22 e2 ee 9?, 77 76
10 eo eo 62 87 99 z n
erators of radio receiving equipment to have information on 11 al 7a ee w 103 75 z 76
12 70 77 93 e7 109 74 77 77
what sort of interference one might expect from the sun at 13 79 74 e9 89 107 e9, 77
14 86 eo e7 9s Im z 98 7e
different frequency ranges in the radio spectrum and over 15
16
25
w
m
79
m
m
9e
100
s
e9
?6
76
99
Im
●t
al
what percentage of time the interference might occur (above 17 e2 ?9 *7 100 66 74 99 w
lb m eG 68 9e 82 ?4 97 m
19 7* 77 79 m
a certain intensity level). Since radio burst emission from m 7s ?7
w
?e
99
n
74
?7
102
109 79
21 76 76 74 2 73 m 90 79
the sun varies unpredictably over a wide range of frequen- 22 75 76 72 m 72 73 lW 77
Z2 74 m 71 93 ?2 72 99 76
cies, one can only give the statistical likelihood of certain 24
. ..
73 m 73 89 72 72 9s 72
74 7s 72 ez 74 ?4 e6
levels being exceeded at various frequencies (or frequency 26
.!7
73 7?,
74
72 76 74
72
73 84
7e
e2 74
81 n w ?4 ei 74
ranges) of interest. We call this body of statistics on solar 2e
29
m
79
7%
m
7e
7e
?9
eo
?4
7>
?3
77
?4
71
79
79
73
72
radio-burst occurrence “climatology.” Burst climatology is m
21
7e
62
74
71
w 83
29
76 u
w
69 79 71
7!
mm 72 7a 83 76 77
provided by Sagamore Hill observations at 245,610, 1415, 79 88 m al

2695, and 8800 MHz [Castelli et al., 1973]. Observations

at 245 MHz began in March 1969; observations for the other
listed frequencies began in January 1966. 01 73787 7696I3 676770 72 = 76
To produce a useful solar-burst “climatology”, one must 02
03
71
7?,
727s7S
717477706e
6668
z6e7G
66 69 72
72
7A
;i
76
7?

first determine what intensity (flux-density) level a burst w 7: 71 % 77706a 686e 70 71 7?,
w 74 72 72 76 7268666970 73
% n 7; 74 76 6e 70 7?, 74
must reach to interfere with operational systems. Since ra- 07 75 72 73 76 66 71 74 7?,
Oe 73 73 74 74 66 71 77 72
dio-burst emission occurs over only a very small fraction U9 7? 72 72 75 68 71 7?, 72
10 73 7?, 73 ?3 69 72 7?, 72
of the time that any system is exposed to the sun’s steady- 11
12
n
76
72
7Z
7s
77
7*
73
6e
70
72
72
72 72
72
state (quiet-sun) radio emission, the “disruptive” level of lJ
14
76
76
7;
73 z
7?.
72
74
76
72
72
72
7!
72
73
burst emission must be sigfiificantly greater than the quiet- 15
16
75
74
7;
73
79 71 m
7Z
71
70
71
71
72
72
17 72 74 ;: ;: 72 68 71
sun flux-density levels which vary over the 11-year solar- 18 74 ?6 m 72 70 60 72
76
7s
19 7s 76 74 71 70 71 69 7?, n
activity cycle (Figure 11-5). 20 76 76
~~
.74 71 70 69 76
21 75 7* 72 69 69 74
Since the quiet-sun flux densities are different at dif- 22 7s w 69 69 73
23 75 e* E z 69 69 72
ferent frequencies, one could choose a burst level greatly 24
25
74
73
85
w
77
74
72
71
6e
6s
. 68
69
71
71
exceeding the quiet-sun level at any frequency covered in 26
27
74
7?.
es
es
74
7s
70
70
66
66
69
70
70
72
the climatology study (1000 sfu, for example). One could 2e
29
77
78
24
01
76
n
7rJ
69
67
67
70
?1 R
30 7s 78 71
also choose a burst level at each frequency that is a fixed 31 74 77
69
6e 66
69
69
74 77
7e
mm 74 76 76 72 69 69 67 69 70 ?3 ?e
multiple of the mean monthly quiet-sun emission (for ex-

,31 ?6 71 71 71 ?2 n 92 79
02 75 72 71 74 73 76 93 60
~z e2 73 71 71 n 77 77 % el
w
m z z
71
71
70
69
78
7e
76
76
7? 96
92
60
m
74
7s

t. t
.,’”
%:MHz. “.=.
,,/. .—. - ‘.,
..1 06
07
60
78
77
?7
71
71
70
71
78
77 e2
79
W
77
70
aY
e4
al
m
76
n
77
I
~ ,m
8W ,. –/
.% .._ . . . . ,.’ /
w
09
7e
7? 75
?4
?>
70
72
72
72
7a
79
81
81
77
7e
79
76
a?,
ez
60
e2 m
SW”” 1415 MU,
., . . . . . Lo 76 76 7?, 7?. 72 78 m 76 76. eo e4 n
z “1 ., ,. . . . . . .
................. ,/ 11 n 74 ?2 ?* 71 76 eo 77 76 76 e~ 76
Rw .+. , i2 75 73 74 73 72 76 7s 76 ?5 e! 76
f- \ “’l.. _.. A.. 13 7s 72 76 74 74 7? 76 75 76 77 74
$W ~ 610 MHz
-—-
\. / 14 74 72 75 n Ts 76 75 74 m 76 75
“ / ‘i 77
15 75 7-$ ?3 m eo 7? 74 72 74 76
lE 74 73 ?1 n 96 76 72 74 72 74 79
17 74 7? 72 73 91 76 72 72 72 74 78
10 7* 72 m 73 90 78 72 73 ?2 73 7e
19 74 72 ?7 74 92 ?5 72 74 ?2 73 77
20 76 ?4 72 B 76 73 74 73 73 74
21 77 ;; ?4 ?: p> 76 73 74 72 73 72 74
2/ 77 72 7: 72 7a 72 73 71 76 72 72
22 76 ,, ~ 73 72 2 eo 71 73 72 79 71 73
2* 76 74 72 71 m 78 70 73 76 ?6 71 71
~., 76 74 74 70 el 79 70 72 ?6 ?e 71 72
26 76 7* 7?, 69 76 ?9 69 ?2 77 7e 72 77
Km 27 79 76 72 69 76 7a 70 74 78 78 74
2a 77 76 72 70 74 77 71 73 20 77 ?7
29 ?9 ?2 70 74 76 71 7* e7 77 ?4 :
30 71 70 74 77 71 n 69 76 m e2
Z* 2 ?2 72 71 n 7e el
Figure 11-5. Yearly quiet-sun flux densities for several frequencies mea- 71Enl 70 n 7* 72 78 77 74 7Y 76 m 77 76
sured at Sagamore Hill.

11-6
 SOLAR RADIO EMISSION
Table 1I -2b. Daily values of observed solar power flux densities at 2695 and 8800 MHz. Recorded at the AFCRL/AFGL Sagamore Hill Radio Observatory
(SGMR); Hamilton, Massachusetts

1366

96 270 86 259 S7 24?

74 253 72 262 97 27e 79 2m 9?, 2s7 es 247 103 261 9(, 253
77 25? 69 252 92 27?. 81 253 90 249 S4 247 103 260 90 267 92 88* 259
e2 2s3 e7 2s0 90 269 91 86* 250 92 2S6
57 6e 2s2 86 2s9 S5 245 105 261
93 96 26(I 1(,2 260
72 25E se 251 265 82 2S6 86 250 07 2s5 101 3s9
97 264 115 274
7Z 250 es 263 eo 2s0 2s 66 266 93 2s6 263 92
7s 24?. 66 2s2 94 2s3 77 2m es n? 98 Zse w 2s1 92 2s4 % 9s. 114 273
27?,
1<)3 280 117
74 25? 67 249 93 26e 60 2s2 e?. 251 94 281 SY 249 SC 2ss 94
271
67 2s0 92 2?0 76* 236 66 2s1 94 2?2 2s2 94 2s6 99 107 285 119
es 93 1<19 279 128 20:
6e 252 ee 2s6 74* 23s 2s0 es 29? 01 248 03 2s3
146* 277
249 68 2s2 e4 262 76 2S0 S4* 91 2s5 79 249 04 2s3 93 112 278
116 278 151 290
%55 69 3s3 82 265 77 254 e2 249 ae 2s6 70 240 97 255
1•8
eo 262 77 2s3 eo 249 e4 247 79 231 91 263 122* 277 27s
256 71 2s6 276
77 251 79* 240 e4* 236 246 02 2s2 1!7 mo 153
261 73* 244 e2
79 2s4 79 254 el 24s 240 79 24e 103 119 2SS 1•9 277
267 71 a3
~65 107 117 2s2 142 271
76 2s3 78 2se 82 253 81 260 81 244 ea 249 SS 240
251 81 2s0 96 240 m 2s3 109 107- 114 2s1:, 136 267
265 7s 2s7 eo =9 81 83 260
e7 271 03 252 e2 2s6 86 248 2U 83 no 114 109 1c,9 27S 119 263
267 7?, 262 e4
120 Ioe 106 273 1,05 261
101 268 75 260 92 271 79 2s1 E1 257 60 249 B: 2*9 e3 236
lnl 26e 76 262 93 270 75 24e 89* 250 79 248 as 249 S4 3s6 I2e 1[37* 106 26(, ,10
1,:,6
266
259
94 263 75 264 94 270 ex 2s3 9?. 264 80 245 89 2s3 e7 2s7 125 Ioe 104 257
1<,9 ,<,4 257 161 26!:,
264 73 268 99 278 79 2s0 94 271 el 264 es 2s2 es 2ss 120
264 77 26e 09 263 e2 2s1 % 276 e2 243 91 m. 2ss 114 1- 167 2S9 10 I 262
SY 107 25S I02 265
e4 249 72 2s8 84 260 e8 259 91 264 e4 24e % 261 1*2* 262 115
76 25e 260 900 299 90 262 64 249 100 270 1C)9 271 111 92 103 2s5 104
75 251 90 259 e9 2% % 25e 1n6 260 110 277 1(14 90 1(I3 252
74 74 25; 91 259 92 263 Ss 2s6 11[> 271 112 27s 101> e7 104 2s6 I m
268
7: 254 e2 24e 1<,7 360 11s 2s7 103 e2 ,<>0. 264 1<,3
26CI 74 252 e6 09 257
78 262 85* 254 e7* 82 249 119 396 es e4 261 1M, 254 1,:,s
74 263 1[,3
ee 272 e4 256 Ee. 249 84 24e 112 274 1C9 2m ee e9 267 92 259
e8 277 79 2m b5 232 m 249 Ioe ,C,7 2s0 04 267 92 2S4 112 263
92 252 1C16 269 102 27S z> 116 26?
9s 291
E2 254 7e 260 8S 239 e4 255 e5 249 92 2S9 96 2s1 lVO 260 96 263 lW 267 116 266
262

120* 273 140 295 ie4 344 134 28e 113 269 t43 2s0 106 255 147 3ss 133 394 11* 2s3 I ie 2e5 t 2e z%
: I39 2ee 160* 292 Iee
~~~
347
~z~
117 280 114 269 120 2s2 103 2s5 13s 2s1 130 291 10s 273 120* 297 120
lCA*
2s6
279
3 154 297 132 290 112 273 109 2S0 119 27e 109 2s7 132 277 122 291 107 274 lie 2ss
157* 299 134 287 189 342 111 26e 106 262 i24 273 107 2s1 126 2s4 lie 27e 104 27o 113 2es 109 293
: 164 301 136 2% 174 ?.24 105 267 106 %0 ill 263 i04 24e 127 3SS Ioe 27s 103 273 107 279 1<6 277
6 157 291 190 29e 165* 313 10E* 26S 107 260 101 25s 100 24s 127* 2S0 107 26s 107- 273 Ios 2n 111 320
7 152 292 163* ‘~ lW* 297 11s+ 26e 10s 260 93 2s1 97 2*6 115 274 10S 266 104 271 99 274 115 29 i
8 134* 3s7 I46 2e3 141 2SS 11? 2?* 10I3* 248 e7* 242 69 240 119 27I2 107 2n 101 270 92 271 116* 279
9 1% 2s9 140 2ss 140 282 116 264 %* 2s3 SS 242 93 244 112 2s6 106 2n 99 273 9s0 27S 116 293
10 142 283 137* 293 12S 277 103 26s 91* 244 - 237 92 243 112 262 104* 277 104. 229 101 200 ils m
11 172 2eo 12S* 2s3 120 270 liS 269 E8 249 73 240 92 240 112 364 Ios 27e Itk 110 291 125* 2ee
12 1~a 2s0 132 22s 117 269 112 2ss es 2s3 e2 2.2 90 240 lie 2ss 105 27s 11s 2s3 112 123* 2W
13 1;0 274 127 2E0 114 27s 107 26S 2S 249 es 249 93 249 119 em 103 276 109 2n 117 2s9 126 ma
1* 120 272 120 2s2 ill 271 10s 27s es 2ss m 262 106 2s9 lM 263 101 27s t07 279 121 22s 136 393
15 116 264 112 279 119 2M lICU w 930 2s1 91 2s1 109* 2s7 99 237 I04 27e llm 27s 12s. I47 31S
16 116 262 IM* 279 lie 276 112* 269 94 2s3 92 293 106 2s9 103 2s1 97 269 92 274 i34 299 173* 329
17 106 2s9 114 2s0 123 2s2 109 2ss % 2s7 96 2s?, 10s a 115 264 92 27t 96 273 140 177 332
I* 111 262 116 292 i22 2e2 111* 262 103 260 106 2s3 111 262 123 279 102 262 96 272 140 207 19s 3s5
19 120 2= 111 2s1 121 2m 113* 27s 11s 269 106 2s7 10s 263 137 291 103 2s7 9s 273 143 31s lee 369
m 127 2S? 116* 2e3 123 279 107 26e 120 27s 104* 2s0 113 .267 134 297 106 2s7 107 276 169 312 192 339
21 132 272 11* 2ss 129 2ss 111 270 142 297 1OS* 113 27t 14s 293 104 264 1OS* 271 1•2 306 173 337
22 i3S 276 132 29s 137* 296 117 220 161 = 111 264 121 22s 1*1 297 101 26* 115 223 129. 30s 1m 322
2> 139 291 136* 299 1449 296 110 26s i76 w 11> 262 12S 293 t4e 291 100 267 114 2s3 140* 1446 394
24 125 276 139 723 160 293 1130 264 177 337 107 264 141 307 140 29s lW 274 113 2n 134 292 129 290
2s 13s 2s2 1s2 329 142 227 112 270 ie6* W3 1m 249 1s6 330 13s 296 110 274 119 397 122* 143 299
26. 139 299 1se 337 1S2 294 106 264 201* Wz 102 24S l!I1 332 136 294 112 27s 124 a 142 30s 1•9 315
27 1+3 29s 1ss 331 149 291 110 264 1@7 306 10? 247 160 330 Is 296 114 2ss 131 312 i42 302 1s6 313
2C 146 296 133 >64 1ss 299 11s a t74 303 112 257 179 m 110 27e 14s 3i9 129 164*
29 1s1 302 162 316 11s 269 161 m 112 n7 179 20s 141 303 1‘)7* 272 137 314 13s 397 1370 313
w 1ss 29s 13s 309 112 26s 1s1 22s 114 2SS 162 301 1213 2* 109 272 126 12s 29s 144 202
Zt 147 296 143 296 1s3 290 149 291 132 29s 122 296 162 309
m I37 293 1-A 29s 143 296 112 269 12s 279 103 2s5 119 271 112 293 12s 292 141 206
127 2W 109 27s

1s2 =603 137 * 1296 123 3S1 11s 279 12s 27s 104 260 107 269 106 269 114 27e 137 303 !23 294
: 1s1 2000391 139 m 130 2s0 122 239 121* 27s 96 109 267 111 270 113 290 131 29S
: 1%* 194 370 122 292 114 279 129 292 121 27s m 2s7 113 220 117 276 120 a ix g: 127 -
164* 179 -a 120 1io* 279 129 2s6 121 ● 269 941 )3s6 110 27s 119 27e 12s 2s3 121 2s0 126 ● 276
: 174 161 3?,?. 119 2ss 106 ● 27s 127 291 115 269 96 26s 109 2S0 116 277 12s 2s1 11S 276 13s 293
6 lei t44 m7 119 Zes 104 264 122 2s1 121 279 m 269 117 3SS 113 279 123 29s 109 27S 126 2n
7 192 143 W2 111 27s 104 269 117 2W 129 2s0 106 276 112 2ss 11s 279 123* 269 114 2eo 123 276
e 212* 134 294 lm 273 111 274 117 2s1 126 2e2 112 293 lle 293 124 22s 121 292 lle 2s4 122 277
9 209 136 29s 103 276 116*226 11s 2s2 122 261 120 226 115*2S0 124 293 11S 273 120 299 i2e 2s2
10 191 137* ~ms 102 274 116 2SS 113*223 124* 279 127 lie 2s1 12s 297 114 276 116* 273 131 2S7
11 130 lza 302 113 279 11s 2s4 107 27s 124* 2s3 122 322 120 392 127 27s ii3 279 t17 293 127 203
12 166 137 297 111 117 267 lmO 272 114* 27s 139 190 29s 130 2s1 117 2s2 116* 296 122 2?6
1? 192 123 263 110$ 119 297 104 26s Iie 271 13s 3s6 140 303 126 2eo 110 273 113* 294 113 272
14 174 126 2E4 110> 2e I 11s 2ES 111 2s1 116 275 12s 2s6 1s2 319 123 277 110 269 111 27S 116 27S
15 161 124 2s7 10s 276 116* 2S3 113 2S4 113 27Q 122 279 i4e 311 112 273 117 277 116 2S7 119 276
16 15? I22 2e5 107 274 116 2E4 122 2e7 114 2s6 123 2s1 166 309 1oe 269 117 2S1 112 299 112 267
1? 130 ile*2so 114 277 109 281 137 29s 115* 2?6 11s 27e 139 292 I09 27<, 123 290 124 294 119 276
,*
1?
;o
144
133
127
117 2ei
lie 2-
119 297
110
112
110
277
271
t06 28e
107 27e
101 26e
145* 293
1s2 ms
157* 30s
120
119
124
2m
2ss
m9
107
10s
109 266
x 135 3ss
131 292
134 2s2
109 270
1<,5 26e
105 266
121
131*
i3* ●
29o
294
299
127 ●
122
121
289
279
291
119
126
131*
273
279
Z7S
21 122 126 23s 120 276 102* 270 1s6 30s 126 263 107 265 132 296 111 274 139 297 119 279 132 279
22 119 129 294 1228 277 97 262 151 312 124 2s3 112 270 133 2s3 1<>4 247 143 2s9 IIS 279 133 294
z: 120 131 2s4 127* 2S2 9s 260 i4e ms 1*A 3n2 117 26S 112 267 lr.,3 267 144 293 11S 279 131* 27S
24 131 t4i -mo I-A 2e5 lm* 261 149 30s 129 292 123 273 11% 247 117 275 1so 29s .17 291 139 Sm
25 134 1s4 ~m3 142 291 tm 2s5 143 =2 I2e 292 129 279 102 266 135 29- IS]* 296 119 208 142 29S
.,
.6 120*
144 1s5 m9 i37 2e3 94 2s9 142* 2S8 277 124 27S SS 26S 134 296 132 303 119 290 141 2s6
27 154 i●e m6 141 2e5 97 263 I3e 2s9 123* 2e1 lte 271 9S 2S7 13CI 279 1S6 310 12iI 295 i49 294
2E 164* 149 ~A7 13S 266 1132 2s6 12e 2s7 119* 272 119 275 90 2S6 132 2e6 1S10 310 115 279 139* 296
29 1e70 143 301 134 2e6 lm 27S 126 * 279 114 26e 121 287 1<ICI 266 ,3(, 292 166 341 I I(, 279 127 2ee
m 203 415 1=X* 290 107 272 12e 27e 1[17 264 112 274 104 267 119 292 ,.(, 313 IIS 2e4 120 291
:1 204 417 134 2e6 126 277 109 267 I1:11 266 129 3v6 I24 2ea
m 16? 342 143 +ms 121 2e2 102 275 129 2EE 121 II* 274 119 2E1 277 I3(, zgr., 118 2S4 I2s 2e2

11-7
CHAPTER 11
Table 11-2b. (Continued)

1w. 1 20* 146 160 = lW m 94 1230 277 1•9 293 136 273 114 27CI 126 209 )Ie 270
130 126 1 -* 1639 332 lW 27a lm 13* 297 160 302 130 273 121 276 121 205 122 276
144 201 1SS 293 128 272 115* 273 113* 270 12? 275
1 ?.s *1 129 132* 167 K 113 27* 110
127 263 279
144 i 27 131 134 306 1 is 271 126 129 296 1610 299 111 267 113. 2? 1 129
4 12[, 276 267
152 304 131 13a t* 306 131 222 161 122 291 1s+ 293 116 eee 112* i2e 293
s
i4t 297 116 272 165 137 290 144 263 i14 271 123 276 116* 272 122 20 I
6 124 316 129 127
107 260 126 2e6 116* 264 111 271
7 in m 131 123 133 & 112 271 126 265 1= 279
169 322, 124 128 12e 263 11* 273 2 137 266 129 274 104 =9 I 24* 27e t 15 2?0 1(,(,. 237
6
122* 127 lZY 36? i 190 271 1= I?m 262 123 272 10CI* 244 12r.# 260 113 265 1<$4 261
9 172 340
160 123 1260 229 t 17 274 m 134 Beo 116 264 96 257 116 263 107 262 1C13 265
10
119 1*3 126 * 129 276 2W 131 ?76 112 267 99 256 110 276 106 271 104. 269
11 1% 321
311 112 i36 m *27 lee 122 272 167 266 101 261 111 276 105 26? 113 279
t2 1s1
i4e 11% :~ 131 w tae 279 166 121* 267 io6 =S 112 269 107 272 111 265 119 205
la
I ●s 206 it4 1s3 * 137 I 82 111 a 97 257 109 265 105 267 121* 270 121 2e2
14
let 1 t+ 143 1ss 31s 1-A 292 164 314 106 257 97 m 115 272 100 263 123 27e i 2e 205
?5 303
105 2% 92 240 115 264 96 263 125 296 12? 2es
ie 140 117 1= 142 267 132 290 144 396
17 141 m 123 101 134 262 136 1~?. 200 103 ~ 69 247 113 267 106 273 143 310 132 290
127 162 132 261 140 292 129 99= 36 246 117 269 111 276 161 33a 134 2%
ie 133
le6* 1- 279 1-?. ~ 119 :Z 99 =3 m 244 112 265 125 293 160 360 135 2e7
19 121 *7 !=
m 1 t+ 146 194 133 - 1-?. 297 117 271 %230 91 25fJ 117 273 141 3<,6 165 34* 142 295
207= 145 11s 267 97 =3 % =3 119 278 1s1 321 176 361 1s<, 29?
21 122 279 124 133 269
121 169 197 130 290 154 293 111 96 244 101 26C) llE 273 160* 320 1e6 362 1494 200
2a
2> 119 261 1b7 196 121 279 149 247 109 96 252 110 269 12Y 277 172 329 1B6 336 147 3,:..,
1e3* 174 1249 ml 265 107 264 95234 121 202 134 262 17e 326 176 339 145 296
26 120* 264
29 124 101* 162* 122 260 :: 293 99 99236 131 281 14s 292 le* 320 176 340 I 43 256
lW A7 174 121 27? 160 291 97 Z9 99 257 131 287 141. 262 lm 33<$ 325 143. 299
g 122 176 332 133 117 266 130 266 96 252 10s a 140 269 13<s 277 100 325 ;2 319 155* 2e5
26 161 1s1 11?, 267 121 276 92 a7 116 263 147 299 122 274 166 320 1●4 3oe 13s 207
122 324
1s9 102 102 116* 265 151 299 110 277 155 312 I 2e 205 14,:! 209
122 106 261
166 109 267 97 262 115 2? 1 120 271 144 m 12o 277 144 205 124 2?7 14? 29(,
116*
162 93 260 139 265 139 279 136 293 144. za3
115
tsz 13s 291 127 261 29$ 117 269 122 272 llE 271 133 292 136 300 129 2e4
136 137

1970

141 Zee 122 m 161 326 145 302 135 291 125 274 149 301 114 264 130 269 I1O 265 1* 334 140 279
133 m 124 292 is- zie 14* 29a 141 297 121 274 156 207 109 265 132 105 262 143 329 I 4* 2e4
127 2n II- m 1s1 = 145 3Q5 1- 237 114 269 164 315 110 240 137 290 104 267 140 313 141 204
126 2n ill 27E 153 309 149 a 139 292 111 267 16S 313 107 266 144* 302 ile m 133 207 154* 291
1 i7 271 11s 261 1S5* 277 164 --4 i 42 297 113 269 160 =3 107 264 I ●e 302 113 264 122* 2MI 15s 3,X)
tlo 272 122 Ze9 1% *a 174 336 139 299 1M* 1% 306 107 262 124 111 266 129 296 136 303
102 271 126 22? 156 293 i95 36s 140 299 117 ?2 141 294 107 261 139 266 114 270 137 297 1S3 302
107 267 141 m7 159 303 lel 339 13s 292 114 269 136 BeY 110 265 132 279 121 277 126 291 133 305
11s 272 142 312 149 m 166 343 133 m 117 266 125 276 117 267 136 223 126 277 130 297 157 324
1= 264 160 347 i42 299 197 =5 136 290 124 274 la 266 113 262 122 274 i B6e 277 126 Ws 1•O 29?
1s 305 174 w 144 296 177 332 148 -A3 144 297 115 266 1OS* a2 11s 269 122 273 144 212 160 3~;I1
164 302 184 326 i51 299 i74 ~zi 1s4 30s 143 301 110 265 120 279 106 263 119 270 152 3i6 160 31$>
164 292 169 334 140= -?.2 164 ?,20 163 317 169 330 i 09 264 130 297 103 113 26S 16,> 333 I 40. 3$;UJ
167 313 1s7 333 135 269 1s?, >10 169 322 177 338 107 260 155 326 1OS* 233 114 270 177 372 14! 3[, 1
In 221 16s4 327 131 286 141 301 170 32* 176 330 103 257 143 267 100. 239 120 263 179, U7 136 293
166 ~ lei 334 126 207 12i 293 17s 322 163 312 101 260 140 296 100 117* 269 109 426 132 2%
171 300 166 344 120 283 129 292 165* ‘A2 162 ~A7 t 03 262 121 261 96 me 113 271
1750 m i 79 ?.33 119 206 122 292 164* 2e6 151 297 109 26$ 12* 262 102* 259 120 275 142 301
166 309 19Y 334 116 261 117 279 166 301 149 293 113 277 123 201 114 271 120 27S lel 3s3 143 296
20 16S 309 1E2 327 127 283 114* 274 1% 294 137 2e7 116 207 123 276 123 274 27Y 163 324 143 31>1
21 172 320 192 3X 135 264 111 270 14S 29U 134 266 126 296 119 2?0 127 272 120 279 144* 332 1 3s 304
ze 1s7 m 173 326 138 2U9 111 273 I● 208 I 20 278 134 292 lle 270 132 277 124 ~ 166 207 127 297
23 143 m 179* 331 140 265 112 274 1?,9 287 122 270 142 292 124* 269 139 279 129* - 134 291 121* 2s4
24 133 * i 75 3?,3 151 287 110* 274 139 292 121 269 I 39 204 117 270 1- 294 140 299 126 202 112* 2el
23, 142 S 166 3s 150 293 11s a 137 292 126 273 143 265 11S 266 132 270 1SS 312 121 276 I 13 20(>
26 154 x 171 327 lm -l II* 270 140 290 123 279 150 2W 116 260 133 169 319 122 278 111 266
27 1s3 301 162 311 139 239 tie 270 133 262 131* 162 293 11S 274 127* 277 166 310 132 290 107 279
w 1s1 s 160* 312 137 232 120 2s3 I 32 292 133 133 293 119 279 126 263 166 m3 140 ti7 107 2?6
39 152* 306 140* 291 1Z3 291 1*A 286 1+5 % 136 279 124 282 110 272 167 301 143 291 113 279
151 304 133 x lm 266 i39 290 1468 132 276 126 207 114 267 1% 294 147 391 120 209
E 151 291 131* 277 140 267 121 269 131 291 lW 267 119 2WJ
?94 146 ~ 160 319 142 293 143 303 t46 2% 133 293 120 279 123 274 126 260 146 313 137 292

1971

11- ,206 1s3 304 li7 265 83 264 64 243 261 i 2* 271 94 B!n 79 247 100 264 99 273 110 277
121 a 149 31s 107 279 95 269 e7 231 11s 272 234 79 24* 101 246 W 273 1<>7 273
124 * 136 306 104 261 94* 261 930 236 ’233 110 260 :? 20 266 1C17 269 1(120 274 1C,* 277

6
127*
132*
124
,302
, 301
3fn
139
137*
132
301
ml
=
1oo1 I 273
%
97
276
276
96266
97
97
*S
269
w
104
111
271
2B0
%

2s6
11s
114
109
276
2n
27o
z91
22
●1
we
240
246
24s
9?
96
93
269

263
106*
1(I9
101
272
273
269
99
1cl?
I C*
260
2?9
294
7 136 = 124 294 90. s 272 101 266 114 261 102 270 29 93 253 69 260 96* 269 1<,7 291
e 136 ~a 114 233 39 274 115 2n 233 99 269 97 9CJB36 91 239 n 264 1(!9 266
9 141 30s 1~ 2?3 25 273 s~ ~ 1160 274 m 92 256 % 262 B523CI 91 265 93 264 112 292
10 1?,9 314 102 273 90 274 122 271 233 92 93 254 M 246 66* 264 9? 339 290
113
11 137 w 104 273 92* 246 110 276 122 2?6 247 94 261 91 zn 79 266 64 2* 97 260 115 293
12 139 309 105 a74 93 117 276 114 271 252 94 26s Be 79* 261 m 261 96 267 117 291
13 1-4 * 106 276 :% 126 262 111* 96 262 66 234 66 244 66 ms 26?
94 269 110
14 141* 312 109 273 % 272 127* m 107 266 251 Be 26? - 253 %252 ●l A = 271 113 297
15 140 31e 110 277 92* 127 263 104 261 24S i06 276 63 251 102 =9 76 249 96 27r~ i12* 200
16 I 40 317 121 266 % 272 126 Zel 102. la 2n m 99- 79 2*9 B 2?2 111> 209
1? 141 319 12s 293 % 277 127 2W 102 237 2% 103 266 90 101 25? 66 ml 96 276 123 2%
1* 144 319 120 297 95 273 127 280 96 2ss 247 106 271 97 % 104 255 69 260 97 270 126* 293
19 Iv 32Y 132 295 91 271 123 220 96 237 a 111. 267 109 99 260 294
9? 273 I <J(I 20 I 124
20 159 214 124* 277 91 274 123 274 94 252 101 26s 123 269 96 a? 110 265 103 279 126* 292
21 164 327 126= 92 270 114 269 90 243 247 95 129 293 99 =2 I 06 2e2 j<* a4 Z9,:8
129
22 165 134 2% 93 273 108 269 76* 243 247 99 270 lm 27? 97 m 110 266 113 207 131 293
+: 1s30 g: 132* Zm 69 109 279 76 247 244 Iw 279 129 262 97 =9 I 09 2er, 29(:<
113 206 129
163 133 293 90 270 96 275 n 266 246 lla I 30 96 2S9 116 291 116 290 123 zgr>
% 134 31s 130 297 9r 272 92 259 79 246 293 105 :: 124 273 97 264 1 lb. 206 284
1169 209 122
151* 310 121 267 90 264 92 360 64 24* 255 104 271 110 271 97 260 1*B 275 I 2c, 204 115 281
2% 152 314 II- 2n 90 w Zse 91 257 105 262 101. 267 119* 264 1(!3 27S 10S 274 I 2,:1 292 115. 279
192 314 117 s 07 26S e7 ~i * 113 273 96 262 109s a7 103 274 99 266 i 24 204 1,,7 278
x 1* =3 65 262 03= 24a 94 239 123 274 97 260 97 259 1[,1 27<> 96 265 IIe 270 ,,:,4 2 ?9
m i 49 312 el 262 0?. 243 94* 259 126 279 100. 260 99 2s4 1(j9 275 1[>6
1*1 267 96 269 282
31 f= 310 95 269 m 259 96 256 03 249 90 272 , (, , 296
m 144 310 124 290 93 271 109 269 98 2s9 91 254 103 101 262 92 255 96 267 104 2?6 114 296

11-8
 SOLAR RADIO EMISSION
Table 11-2b. (Continued)

JAWV FE~ *CM -IL WV JIM3 JULY Aur,us’r SEPTEMSER OcTOeER NOVEMBER DECWBCR
2695 8800 2695 - 2695 - 2693 - 3695 - 26= 6800 269; B% 2695 98(10 2695 BOOII 269S 80CICI 269S E96C, 2695 eSOCi
-.. ..-
DAYm47w7m421962w WWwww ?66? Wz HHZ mz I(nz Wz HH7. MHZ MHz MHZ mn2 MHZ

1972

% 263 101 =7 116 276 M 246 e?, 244 107 259 115 243 136 310 126 2 ?2 1<,1 26? 117 3[10 ea 271
93* 264 101 lie* 27? 26 250 e2 23e 114 263 118 2a7 132 312 12* 270 97 262 i 12. 2e7 el 269
96 2s2 %* 2% 1120 276 e7 =5 64* 237 122 271 121 244 I ze m5 11O* 269 93 261 1(12 274 77 264
92* 258 9s9 123 22 237 e7. 2s1 132 277 131 2s4 I 2e 3r.,4 1<,7 237 9<, 23e 95 273 73 262
94* 2s6 94 z 127= 221 96 s % 2s0 143 279 130. 254 12e 3CB7 100 266 ee 255 90. 27CI 71. 265
9s2s2 94 23a 131 261 lm 266 102 256 1?6 277 127 255 121 Z9e 104 263 ee 236 e4 262 71* 262
% 2ss 95 261 122 221 107 267 1Oz 252 131* 276 125 233 136* %12 % 257 e?. 256 77 254 a<. 26e
i 942%
91 2s5
100
99
263
260
126*
123 27s
107
116
266
366
115*
113*
2s8
259
123
130
286
2S7
117
111
246
245
I oe
106
276
27e
1(I1
1<,0
264
263
ee
09
257
264
73.
75
231
25Z
e6
me
2?%
Z?*
9
25 291 1- 123 275 115 264 121 262 120 278 Im 236 102 271 93 263 Bb 260 ?5 253 9(J* 26e
10
22= =3 112 z 122 274 116* 261 124 275 129 277 109 247 % 264 29 =9 26 256 BG 259 99 26e
11
93 2s3 114 271 116 27?, 115 260 129 277 129 276 104 24s 92 253 e7 256 e4 2S6 e<, 260 I 02 273
12
13 95- 1 ie 2= 114 270 113* 355 140 288 121 257 lW 240 29 24? ee 232 e3 26 I 81:v 261 1<,5 2e2
14 104 m 126 2E1 117 27Y 111 146 29? 119 252 106 244 26 240 26* =4 m 261 75. 239 1[,9 262
15 111 224 134 2a6 122* 273 111 252 146 2ee 139 270 100 245 m 243 w 2% 90* 27, 77 26Z 1<SE 2e3
16 11s 245 147 2E3 119 221 109 2S6 145 204 121 =e % 229 * 249 69 2% 96 27<, 91 265 1(,3. 209
17 107 Zao 161 m 118* 221 107 261 147 284 126 237 99 237 93 24s 22 253 101 2?6 e2 26:, 104. 2s3
te 104 237 181 324 118 273 106 255 142 2m 131 a m 237 96 =5 D9 99 277 e6 261 1(!3 267
111 265 1m* m 116 273 100 256 134* 274 126 255 234 lm a9 92* 262 11:6 287 92 263 1os 206
x 114 =3 169= 319 122 27s %* 129* 270 120 251 234 101 9a 266 106 29r, 99 269 1,)2. 279
21 121 266 tee = 121 276 93 244 127 26a 123 256 231 102 262 104 272 ile 303 100 275 1(,7. 277
22 lle 270 167 307 120* 276 97 248 125 263 118* 234 112 264 lW 272 12s0 304 11)3* 2e2 99* 271
2?, 127 276 164 296 127 278 97 247 lie 269 114 249 9a 243 111 11<8 273 133 325 1(>7 2e6 9e 26a
126 277 153 293 111 267 98 2s1 117 275 11O* 249 96 266 I oe 112 273 137 3;7 11<$ Zee 93 263
t 22* 272 141 287 103 261 98 2s5 109 277 107 246 97 249 112 112 27* 153 364 1(!2 285 97 263
119 266 141* 267 99 96 2ss 100 263 103 244 106 12* 273 113 272 153 357 99* 2B(, 91 264
114 26s ix 276 93 259 % 262 98 = 107 244 109 262 126* 269 I 07 270 1 S(I 36 f., 94 273 9(, 265
113* 26S lle 273 e9 2s9 90 257 100 269 112 247 117 2n 133 277 lW 269 , so. 359 B9 271 e6 266
109 264 lle 277 es 249 26 252 100 262 114 247 121 221 I ●e 291 267 136. 335 B4 267 ea 27o
110 270 26* 232 03 2s0 102 256 114* 247 12s 293 146 247 264 133 325 ee 271 06. 27e
107 259 m 249 260 12a 306 135 276 12a 302 ee. 2? 1
10s
10S 262 130 281 114 272 100 2s6 11!3 266 122 262 109 w 113 27s 101 264 I oe Zgr.+ % 269 92 273

1973

66 243 79 262 66264 97 266 96 274 72 239 74 2*2 73. 236 96 265 97 265 79 2% 8\ 261
69 252 79* 261 63 263 1m* 269 101 2n 7t 247 77 2*7 738 233 111 275 96 261:1 77 257 81 266
96 260 e4 262 a7 = 113 2?9 101 270 69 243 77 247 72 231 116 271 92* 259 ?4 232 76 23?
100* 26n 262 659 263 lK 271 106 274 69 24?, 7s 243 73 236 119 277 u 232 71 251 7* 252
103 259 : 263 24 263 102 369 99 265 71 248 W* 244 72 246 123:1 269 60 24a 67 252 7<,. 24e
106 270 66 269 e7 267 99 263 94 m 71 245 93 247 74 239 114* 263 7a 232 67 230 66 24a
102 265 69 272 al. ~ % 26s 69 =7 72 242 aa 249 76 2* 109 2% n 2WJ 66 23\ 66 249
% 261 91* 270 a7. m 106 az 251 73 247 M 7* 242 101 252 71 251 65 251 67 250
9 m 264 92 273 92266 103 277 al= 247 74 247 26 =7 73 239 1<,* 24a 7,> 251 6b 232 63* 247
10 94 2s9 9Z* 274 97 270 102* 266 73 247 76 249 26 249 7(> 243 % 249 70 247 64 232 6e 245
11 93 263 a9. 274 102 271 99 269 73* 24?, 78 247 78 236 67 23a 24 239 67 246 6S 252 67. 251.5
12 93 94 27s 101 261 30 261 73 241 7e a 71 241 65 233 79 240 64 242 6a 249 69 269
1?. 93 93 273 lm 262 a7 265 71 243 al* 249 67 2= 63 233 74 244 63 242 6? zea 69 249
14 87 % 2n 1- 274 el 260 76 245 a2 249 6a 23a 66 233 74 244 63 245 62 2*3 69* 251
15 26 %= 273 %Z aa 76 2s9 m 244 83 - 62* 23a 60. 23a 73 242 67 247 ?0 247 72 251
16 e3 29 962ss 73 77 243 m 24e 6a 240 62 239 73 2*1 67 2*8 70. i49 75 261)
17 85 az 266 -* 260 71 249 7a 242 76 246 63 2*O 63 22a 72 247 72 251 69 233 ??* 26b
1e, W el 26S m 263 69 249 81* 247 77 244 70 241 66 22a 720 247 70 2%> 70 254 at 263
19 65 77 263 W 271 72 24a as a3 253 6a 242 64 239 76 2*7 66 247 7<, 253 ao 261
20 e7 60 m 262 77 6a 251 69 al w 23a 65 237 7s 247 69 nz 72 256 a3 263
21 62 79 =7 77 =9 a3 2s4 a7 249 92 247 66* 23a 70 243 64 72 252 76 a~. 272
22 69 79 222 alo =9 30 2% m 91 =3 65 23a 70 243 e7 2% 76 78. 231 92 269
22 86 al* 77 Zn 94 262 el 24a 9om 6S 237 71 243 92 256 79 237 at 2% 92 2?4
24 a7 26 264 7a =7 % 266 e7 252 W 24e 62 2*2 73 2*7 92. =7 90 266 255 93 273
25 a7 66 76 as m 264 = 255 W 247 69 243 77 zea 100 262 93 26e 91 271
26 m 26 265 749 m w 26S 86 254 az. 245 71 243 7a 2% 102 263 93* 268 25a a4. 263
27 e3 87 270 60 263 274 24 25Y 72w 244 72+ 242 ei 2YI 102 267 9a 27a 253 8<> 264
28 82* 90 265 a3 264 279 79 244 70 241 26. =3 104 267 92 270 236 75
29 - SW 277 7a 240 n 244 69* 24S a7 2% 106 266 91 263 2% 71 k?
m al 90 263 27s ?6 241 81 247 71 242 a7 2% 1*3 267 93 264 260 67 249
31 79 97 m 73 242 73 240 9Y 265 8* 260 66* 24a
m 90 8a 267 26263 92 265 24 79 247 73 242 7a 2*2 96 2% 79 25s 73 253 76 2S7

1974

63= 249 65 249 69 243 64 244 242s2 W* , 246 69 264 71 240 67 a2 252 72
eo 2ss 259
: 67
26
231
232
6* ,2s6
66 252
69
7I
246
24a
63*
62
236
2-A
too
100
264
261
e4 245 W 275 7i Zel 6a* a? n* e4 261 6e. 2%
3 ez 242 99 274 70 23a 65* a3 255 e7 264 69 254
26- w 252 69 260 66 23? 102 2% 60 24S 72 24 i 69. a9 263 8a* 256 68 232
: 66 249 70 71 242 67 2X 102 2s7 79 2Z9 107 267 71 241 69 236 93 274 66. 256 67 252
6 69 2S3 71 251 69 243 70 244 106 261 78 243 102 269 74 243 70 2?,2 99 262 85 257 6? 249
7 69 255 72* ,2s2 6e 24?, 70 244 103 256 77 242 m 261 72 243 73 232 1,>,:, 263 ?0. 25: 64 ;?,.>
a 74 259 74 257 69 249 71 242 100 257 91 a6 M 75 247 112 274 75. 240 64. 245
9 76. 265 n 2s6 72 251 78* Z5i % =5 77 241 70 2ea 7* 243 a2 =r> llZ 274 70 2s1 62 243
10 al. 266 m 76* 251 82* =6 a9. 253 78 247 73 242 73 245 86 255 116* 2?2 7(> 256 65 25e
11 64* 265 73 2s2 77 294 88 256 90 ml 60 244 70 242 79 251 9<1 264 127 29? 69 25; 69 25?
12 26 265 70 247 7e 255 93 262 83* 2SCI 74 245 70 23e a3 2S W., 267 121 291 72. 254 ?i 25 ~
13 a7 s 70 247 = 254 91* 267 79 244 73 24?, 69 237 e6 231 9e 275 113 20(, 68 251 74 Z56
14 62252 70 75 2s0 99 277 73 245 72 241 69 243 77 247 90 265 10<1 26e 71 232 ,<* 26:
87 249 76 2s9 73 2s2 107 m 66 239 69 2-A 71 246 76 230 66 26! 9s ?36 71 255 77 262
26* 2s3 77 75 2s4 99 264 67 237 71 239 77. ao al 2S 93 274 79. 254 73 261 B2* 261
03 259 77 254 74 251 aa 267 63 232 71 2-A 76 231 a4 254 86 25(, 74 248 77 263 81 263
ai 25s G 2s6 73 2s1 e3 262 62 233 70 23Z 71 245 20 2*9 %1 231 71 25(, 8(. 265 S3 267
20 249 74 2s3 70 247 77 255 61 2-4 66 2-A 73* 245 74 25<, 8<s 245 67 249 e? 27<s a4 271
7a 254 73 24s 70 247 71 247 59 236 62 232 73 243 70 246 74 243 67 247 S7. 267 B1 26<1
74 247 76 m 67* 247 67 241 20 27A 67 226 n =2 70 237 72. 233 66 240 e9. 272 79 262
72 247 m 24a 71 24e 63 236 S9 237 66 229 60 251 67 24,> 69 65 243 94 280 fi2= 267
76 244 71 24e 62 235 60* 27A 66 2= az =3 69 23a 6S 64 245 ae 274 77 2%
72 246 73 67* 246 63* 240 60 23s 66 235 7e 24e 65 233 64 224 66 248 ee 266 73. 256
67 261 73 66 24s 65 242 62* 232 67 236 n 243 64 239 64 235 69. 245 W 265 71. 2s5
64 27A 22 63 240 64 239 68* 2?.9 67 237 7a 245 62 233 65 237 72 233 83 272 69 235
65 2X m 65 244 a 242 73 246 69 23s 74 2*2 62 233 67 237 71 254 e3 26S 67 257
64 236 m 62 24S 65 238 74 245 70 27s 71 23S 61. 230 69 235 72 2S5 e<) 261 68 252
6S 242 62 245 68 241 74. 246 76. 241 70 23e 57* 229 77 249 72 252 76 239 69 251
66 245 63 2?.9 76 250 = 246 79 244 7i 230 61 226 79 249 73 24e 76 261 6e 25,:’
63 24?. 24 2-a 79 247 73 244 64 236 ai 257 7~t 253
74 251 72 70 247 75250 7e 246 74 79 ao 71 242 77 248 86 259 e<> 261 72 256

11-9
CHAPTER 11
Table 1 I -2b. (Continued)

DAv

72 2*5 253 44* 223 64229 66 234 63= 214 63 231 n 240 7* 239 @ 233 63 64* 247
; 70 24s 351 66* ’224 63 227 243 m Zle 63 77 242 7** 240 67* 69 243
?, 71 231 2m 67 226 66 z= 222 63 229 92230 * 241 67 240 : 67 244
62244 66 -226 6% 232 224 63 231 96230 79 241 6? .- 67 46 207
: 69= 24S 244 64 234 660226 66 235 224 62 231 99 al 79 24 66 239 m 70 246
6 67 243 ’42 63 -224 4s 231 64 102 262 79* 242 64 237 72 230
7
8
m
7i
243
243
243
244
65 :Z
669 231
66224
a 227
6a 231
61 226
220
224
61
61
22?
229
99* 23s
96* 24*
77
73 243
64
64
240
239
?2
71*
247
69
66
242
244
242 66* 24,
9 72.244 243 67 232 65 230 60 227 223 224 96 242 241 66 244 66 69* 246
10 m 347 67* 231 64 230 40 227 21* 2a4 % 246 69 239 64 728 239 67* 243
It 73244 249 63 = 59 224 227 90229 69 233 244 71 67 244
12 76 243 241 231 61 924 36 224 21e 223 96 240 69* 2= 244 67 247
13 730 242 233 : 62224 59= 224 216 231 ● 239 64 237 24* 64 243
14 73=240 66 231 61 223 223 221 n 239 64 z= 246 331 63* 241
15 71 239 344 630 227 61 224 s 221 232 71 236 63 242 6S 240
16 70 237 241 67 232 61 22S 221 69 237 630 231 241 m 6S 244
17 639 240 239 67 230 61 Z3i % 220 %: 64* 230 66 233 M 233 244 239 64 24S
te 69= 241 241 39 221 36 223 219 69 63 236 6S 64 24S
19 *239 63 23? : % 59 220 33 222 227 64 233 6* Zml R 236 64 246
20
21
22
69
-
69
238
2=
236
z
63
242
244
243
64
6-
336
z= -
3’9
221
220
:
223
218
21s
324

220
:*
6*
69
231
233
233
64 233
67 233
66* 226
::
234
236
64 230
239
237
239
230

2a7
●

6**
331

246
23 67 236 226 60 210 60 213 69 234 64 231 65* 230 239 2*4 63 2*3
24 46 2= 242 227 60 219 60 210 61 224 65 64* 231 6S0 230 66 249
2s 6% 239 64= 236 62 323 223 63 23t 66* 229 67 236 66* 230 :Z :: 66* 260
26 64* 236 63= 230 63 221 219 235 66 6* 233 66* 233 231 344 66* 2*7
27 6* 243 63 236 62 231 64 221 220 233 66 232 69 234 64 237 243 64 247
64 241 63 242 62 231 65 219 66 234 63 234 69 239 64 232 239 341 64 245
: 64* 239 62 230 64 224 217 233 63 234 71 243 m 234 61 237 343 67 246
6S 236 40* 229 63 222 60 215 236 64 234 72* 241 6S 233 62. 239 249 67 243
z 64 2= 62 229 211 69 233 73 239 62 2*2 66 244
m 69 241 67 243 6s 230 62224 61 224 60 224 6S 7? 240 69 66 ,1 244 66 245

1S76

1 66 239 63 237 * 233 71* 237 67 232 224 216 64. 226 63 230 224 64 237 63 237
2 * 241 63= 237 600 2= 663m 44 223 224 217 66 227 64* 230 226 237 6S 241
3 670 240 63 236 610 236 67* 246 - 230 222 219 70 232 64 239 330 :: 237 66 241
4 67 240 63 234 61. 2= 67 245 64 233 223 222 70 233 6S 227 239 62 23Y 66* 246
5 67 243 62 233 60 233 a 241 62 226 Zlm 71 230 64* 223 236 630 231 65 237
6 66 239 63= 60 231 43 2= 61 224 218 71* 232 63 224 22Y 39* 233 6S 2* i
7 66 237 63 241 61* 231 63 2= 230 210 210 6* 232 63 226 227 42 232 64. 234
e 6* 233 62 241 61 232 6S 231 62= 232 221 219 7W 231 63 230 227 63 236 63 237
9 a 240 63 241 60 231 67 233 61 230 2= 220 690 234 63 227 46* 223 6S 237 60. 236
10 6S 24i 63 240 63 233 67 237 61 224 70* 64 66 229 6* 236 70 242
11 - 242 62 233 63 331 720 239 62 s: 222 66 230 6S 223 6S 63 2*O
12 710 246 61 233 64 232 719 239 620 219 61 226 66 66 6S 230 6S 237 720 236
13 74 249 233 649 230 71 237 63 6t ;: 210 64
66 64 63 233
14 62 2= 64 230 71 z- 63 224 420 224 218 61 216 6S 220 67. 231 73
1s 74 247 63 239 66* 233 70 236 63 229 42 223 21E 216 64 226 67 233 : :~ 74 242
16 72!w 247 63 236 * 240 70 230 67 236 64 227 21, % 221 66 230 69 223 66 239 72 239
17 71 245 63= 237 720 243 70 231 6s 44 23s 218 64 232 67* 227 70. 237 67 260 710 241
le 7i 247 m 233 74 244 70 234 m 226 64 219 66 224 70 241 69 241 67. 243
19 69 2% 64 2= 76* 244 69 232 63= 219 63 223 219 66 237 : 224 66 239 69 244 67 243
20 6- 246 64 233 = 243 69 231 63= 221 46 221 210 63 234 61 223 68* 234 242 64* 236
21 46 241 64 - 241 69 23S 64* 226 6s 222 219 60 230 62 224 6?* 235 H 242 66* 234
22 63= 239 73 243 69 236 62 66 224 220 61 222 61 227 68 237 66 24 i 67 240
23 6S 243 79 24s 66 227 61 229 43 224 36 217 60 222 63 234 73 239 46 241 63 237
659 24 t 43 79 242 67 234 60 227 62 z= 37* 210 59 223 2= 73 241 239 69 2=
3 64 241 61 77 24s 690 236 60 226 - 220 39 223 39 227 % 226 770 239 : 239 69 237
26 62* 240 62 73 2s3 67* 230 40 224 62 224 36 220 39 222 64 z= 74* 237 44 240 70. 239
27 610 236 61 76233 67* Z= m 60 222 37 21e 60 223 63 223 71 236 64 239 70 240
39 610 233 60 60256 u 232 2ie 60 219 39 223 61* 6S 229 70 46 239 69 239
29 42 234 61 7Y 267 69 233 223 219 36 Z’m 62* 22e 63 233 64 :2 64* 260 69* 239
m 63239 74 263 71 23s 223 219 59* 2/6 62 63 226 64 64 241 70 239
31 64 239 74 344 223 60- 226 64 230 66* :% ?0 239
m 67 242 42 237 69 24 i 66 236 61 223 36 220 6S 227 64 227 68 233 m 239 69 239

1977

66 246 n 343 67 242 66 241 236 77 z= 66 261 69 233 72 237 al 231 a? m 30. 290
; 71 249 77 =9 69 245 67= 239 P 770 242 - 243 69 237 70 236 03* 244 96 240 S7 233
3 76 244 76* 24e 70 * m. 237 242 82 246 32 240 69 230 7s 24 i 690 26s 83 246 es 232
4 71 247 73 25e 708 24t 67 244 m z% 01 237 71 236 72 239 62254 62 92 262
5 69 246 73 246 71 237 - 243 239 77* 231 76 239 73 239 73 237 99 262 81. 246 96. 265
6 66 246 70 ml 70s 69 242 231 ~ 241 76 243 74 233 73 237 94* 23s 04 245 9e* 271
7 66* 69 = * 341 69 243 233 81* 237 73 233 73 23i 77 24S 93 -3 63 240 103 276
B 42 73 247 n = 237 240 62* 239 6% 232 71 23a 03 246 93 26o 82* 242 101 276
9 64 76 =7 73 246 62 243 - 237 6S 2W 73 232 93. 23s V3~ 242 105* 262
10 a 77 246 72 243 67 244 w 27A m 236 74 243 94* 233 e7 230 63 2*6 112 2W
11 64 246 782s3 70 241 70 239 70 233 73= 231 63 230 74 239 92 269 96 249 96 =7 107 202
12 M 246 79 244 66 2% 69 241 67 2-A ?1 230 64* 230 74* 236 96 24s 91* al 99 259 109 204
1?. 7i 244 77= 241 6* 238 71 24i 69 2= 71 231 64* 231 77 236 100* 263 90. =2 070 263 96* 264
14 66* =9 76 244 700 239 73 244 73 242 69 2-A 65 23o n 233 104 262 69* 246 92 266 99* 262
Is 69 2S9 73. mo 67* 2S 76 243 71 66229 63 231 73 236 106 263 63* 2*O 94 267 m 236
16 73 260 79* 44= 231 74 24 t 73 241 66 234 63 29o n 240 11O* 263 96 243 92 ~ 96 260
17 72 =9 64360 660 234 n 24S 70 237 69 2-?. 64 226 73* 236 109* 237 63. 236 94* 237 66 266
te 69 331 79- 670 2= 69 234 70 244 71. 233 74 103 234 66 240 94236 69* 253
19 72 2S3 76 244 66 24i 71 249 70 233 76 233 65 226 74 :Z 91 -* 234 94 363 03* 260
20 70 2s9 74* 244 67 236 71 240 241 76 232 67 231 76 246 92* 253 93. 249 92 261 02. 254
66334 7-A 244 a 232 69 236 z 79 2M 69 229 79 244 93* 249 91 25s 07. 2S7 77* 253
21
22 70 m6 7i 244 46 240 66 2-A 44 e7 242 74 2= 77. 240 94 253 63 ml 63 233 e3* 26*
23 7.? 3s6 70 245 * 233 69 2s 94 24S ?2 243 73 234 91 2* 79 253 m as 93 2S9
24 72 =9 66= 236 - 23S 66 236 90 253 72 240 73 2= 92* 246 01 =1 W* 333 29 260
a mmo 66 Zm :% 690 241 2s 9e 266 730 239 n 244 92* 246 60 231 91 936 90. 264
26 7i 234 67 237 66 239 66 234 64 240 101 262 7’2 239 76 240 92* 247 01* 231 76. 236 9Y 271
27 70 247 67 242 67 232 63 233 67 236 99m6 73 237 72 240 a7 231 S3 240 60236 93 360
ze 71 234 67 246 66 24i 64 231 67 2= 74 237 71 236 99 233 92 231 al 256 97 270
29 7i 249 63 237 63 Z?n 70 236 76 233 76 233 w 230 m3 230 ●l 233 103 273
-4 71 245 a 236 64 2- 76 242 71 23i 74 236 m 2*9 36 260 60 232 105 268
31 n a? 640 231 76 233 69 229 74 239 90 260 117 273
m 69 = 74 246 66 239 66 233 69 239 71 234 73 237 69 231 e7 231 96236 94 366
m
SOLAR RADIO EMISSION
>Ie.II-?h
--- (Continlled)
.. . ....... . . .
FE~ ~ -IL w JULY AUGUST SEPT~BER OCTOBER MOVE?IBER D6C2M62R
dF- 2633 - 26% eeoo 2635 eeoo 26% eeoo 26L-m 266: BBeq 2~69 Beme 2~y e~eo 266) ~ 2633 Beoo 26* m
Y mzf3ut34z ?31zt6’lzw muwwmz Wz Wz Wz mz mz NH7

1972

1224 277 133 233 133 3UI 11* 222 1631b31e 122 27e 119 233 e4* 2s0 143 294 121 2eo l= 303 153 316
127* 276 14e 303 133 312 117* 227 166 324 122 276 104 267 e4* 247 144 3ie 119 276 i4s 302 Is 314
122 206 1s9 321 132 303 122 230 136 321 106 264 % 2S9 e4* 230 149 319 119 2e3 1S2 316 134 3i7
iii 227 i4e m i3B zte 133 237 149 322 33 266 94* 261 92+ 26Q 136 32e 117+ 279 164s 310 1329 314
113 272 1s3 311 144 317 134 292 1S9 316 36* m 99 263 97 257 152 320 122* 276 168* 311 1s3 311
112 2m 151* 31s 134 317 137 226 Im 311 23 ml 99 265 lo2* as 159 303 13S* 2BI 137 306 1S9 316
106* 272 157 319 134 263 146 322 103 264 100 259 155 3i3 127 290 1** 304 165 3=
%* 26s 161 326 ~.~ 2B4 123 206 m* Z3 110 279 114 23s 149 296 133 299 1%* 304 l?oe 337
i4e 315 170 331 142 236 121* 273 - = 121 299 116 264 140 293 142 305 15s 306 176+ 347
141 “& 171 322 14a 236 117 272 63 236 123 319 107 2S7 130 229 142 304 lS3 316 ie3 3se
e7 262 146 303 16S ~~ 136 307 iie 277 22 =7 136 321 101 256 123+ 2e4 157 312 ~4e 3t3 133 323
23 26t 153 724 156 310 1- ~ 122 273 % m 144 336 101* 260 ~ie* 225 132 307 141 310 200 366
e3* 26s 137 317 147 303 126 27e 122 274 101 261 133 317 111 263 i21 2e4 136 302 13s 303 1% 377
139 323 145* 233 la 227 123 270 114 270 134 311 107 266 127 226 164s 303 122 230 iB3 376
ei =3 126 310 144 231 t27 290 127* 276 tIo 27e i37* 307 109 271 132* 298 172 314 123 233 130 337
73 235 122 300 134 2- 123 224 126, 273 117 276 i36* 302 100 261 133 2% 133 309 lie 263 166 347
7e* 233 lie 233 124 222 117 277 126 271 122 274 127+ 292 9B 232 146 314 159 310 117* 2ee 161 339
73* 231 117 221 121 236 116 261 120 264 131 232 i2e 239 94 1s1 313 153 307 113 2es
7e 234 112 279 115 272 133 m 113 233 137 233 120 230 9i 146 310 1ss 307 i19 292 144 321
11s 224 119 230 1- ~~ 114 269 149 m 116 234 e7 14e 29e 153 306 121 303 123 306
34232 114 23s 107 266 12s 277 114 267 164 236 110 2BI e7 146 293 133 313 1164 291 127 305
03236 119 226 106 271 ~30 ~ 119 223 165* 302 102 264 ~ 1S% 302 141* 311 116* 230 121 309
91 264 12s 222 10s 266 147 m 126 23s 167 30S 922626s 149 30s 142* 307 i13* 2ee t20+ 299
93 269 126 209 102 272 139 307 170 *?.4 w 237 B9* 240 136 293 144 30s 113* 291 126* 304
92* 271 133 230 107 273 149 303 124* 224 1s0 232 92 = 36 242 139 291 143 300 11S 291 126* 301
l.~ ~ 101* 269 146 306 126 w tse 231 36 261 64 253 133 229 140* 294 i24 301 134 299
110 278 130 230 100* 266 1s6 319 127 272 1s1 276 ee 2s3 91 259 137 29e 130 291 133* 309 143 31e
121 220 123 292 100 264 tel 364 123 294 146 222 07 260 92* 232 131* 29S 1= 29S 143* 316 ~4e 319
122 27e 30 264 164 331 122 274 141 232 e7 34 238 123 209 137 300 146 313 1% 321
126 23s 107 272 124 355 123 272 133 224 e9 232 106 26e 124 204 t33 302 154 3~e 164 336
120 291 109 122 274 26 2S7 117+ 2eo 131 295 i7e. 34i
1~ 270 134 300 130 333 137 237 130 206 123 277 110 221 97 233 140 300 142 299 136 302 157 330

1979

~e7* 33t 173 324 146 3~ i75* 314 160 297 160 343 136 324 i~e 2e2 144 321 174 334 lee 366 149 322
163* 333 169 323 i4e 296 17e* 322 132 293 171 3* iei 33s Iis 27e i4e 307 ~7e 333 lee 336 1s9 339
193 3= 172 310 147 303 171 ~m ~~~ ~ 130 344 173 322 116 291 153* 311 177+ 349 1269 344 130 36s
i7e 333 170 3ie 15s ?.07 163 314 152 297 132 333 le4 339 112 279 149 324 176 337 lee 334 201 369
126 342 123 ?.16 173 317 1s2 307 144 2e7 134 344 166 333 113 276 149 303 163 334 220 391 203 371
le2* 3.3 193 323 161 ‘~ 159 306 149 303 133 339 174 342 li7 279 130* 3io 163 330 239 432 1% 332
i7e* ~a ~93 321 1ss 303 144 308 157 297 197 332 169 327. 119 297 151 316 173 329 as 433 1% 360
176* 326 199 331 150 302 151 302 134 303 202 339 169 329 126 313 152 302 17* 341 273 462 219 372
17S “?.27 172 336 130 307 151* 306 151 -?17 139 344 162 329 129 307 1* 314 166 333 m 445 210 370
t73 323 let xl 162 306 1s2 308 14s ?.02 197 349 1s1 319 123* 301 132 315 169* 337 319 479 206 370
167 *24 133 ?,70 136* 317 1s1 304 148 311 1- ?.32 143 307 i25+ 299 154 332 lee 3= 233 459 206 364
159 332 160 3% 174 326 137 306 1s0 3to 101 336 134 297 117* 301 153 312 193* 347 232+ 423 201 373
177* ?.30 136 34s 132 323 134 310 136* -a7 157 130 299 127 303 1SS 322 20s 34s 221* 236 217 363
1720 330 202 373 i64 3ie i4e* 303 153* 2.04 1s7 323 126 291 141 311 160 3ie 207 372 217 370 2ie 376
176 330 ~gg .- 134* 314 151* 297 I* 303 144 -s 119 2e4 136 307 166 340 206 376 22s 320 2ie 373
te2 324 z 363 162 323 1334 293 153* 303 137 314 113 291 134 306 166 35t 200 36s 209 371 201 371
is4* 3ie 2Q0364 la 31e 144 -m 164 3ie 131 ?.02 112 2e2 131 312 170 331 203 336 209 371
166 339 223 326 173 3= 144 2% 144* 304 129 296 112 270 143 327 let 3s3 203 373 211 3eo 173 346
130 333 217 322 164 321 142 26S 132* 29i 119 233 11S 209 131+ 344 193 374 212 326 1% 366 166 339
179 344 202 ~&3 132 318 133 231 131 2e9 129 300 lle 2e6 166 3s7 191 360 207 369 ~e3* 33B 173 369
1- 239 203 333 1S7 312 127 292 ,.~ 2es 121 233 120 202 170 344 136 340 191 349 162 349 166 349
193 333 123 ~Al 132 3ie 14?. 294 133 294 110 286 122 297 179 360 ies~ 344 lee 349 166 340 LS2 327
207 349 1634 331 160 3ie 137 302 ~~. 392 11s 226 132 291 100 3S2 lw 347 123 34s 160 333 ~4e 313
190 32s 1s% 31s 167 333 134 300 ~30* 277 120 291 13s 3oe 179 340 199 343 200 334 1SS 321 144 307
182* 323 144 297 ,6-> -~ 140 266 127* 276 124 239 131 301 130 360 1* 332 166 3s7 i46 307 142* 307
175* -21 ~~. ~, 173 333 l= 233 1= 276 I23 307 133 312 ~e3 351 193 329 179 333 143* 31e 140 307
17e* VA4 133 274 16S 333 127 277 127 ‘A7 123 2e9 176 340 ie7 346 106 344 132 313 140 3ie
iei 32B 147 301 163 323 161 3i2 123* 274 135 302 129 2ee 172 334 ie9 355 le9* 345 134 311 142 323
ie7 339 173 3ie 136 233 i27 273 137 312 126 292 1S9+ 316 180 341 lee 330 130 297 130 342
le6* 3*A 160* 323 162 306 133192e3 t41 31s 12’3 294 1s4 302 177* 31e 193- =3 139 320 166 324
188* 336 1* 326 143 231 121 291 147 312 192 3S2 173 3s3
iel 332 ie2 337 161 317 132 303 143 294 1% 319 133 305 143 315 169 331 132 3S2 201 371 179 349

11-11
CHAPTER 11
ample, five times the quiet-sun level). This climatology level). We have also shown the “above-the-level” time,
presents results using both approaches. I“ae.
With regard to the statistics of burst levels over a certain For Sm. < S<, the longest possible length of time that
threshold (flux-density level), one might want to know how a burst with mean flux density S~e could exceed the Se
many level-exceeding bursts occur in a given period (in a threshold is S~e/Se X Td. This is the worst case; it comes
24-hr day, for example). One might also like to know for nowhere near representing the results for actual bursts.
what fraction of time (rein per 24-hr day, for example) do In similar fashion if S~e > Se, the longest possible length
bursts remain above a certain threshold. Again, both kinds of level-exceeding time would be T~. Obviously, this max-
of statistics are presented in this climatology. imum result is not representative of Tae. A burst with an
The radio-burst climatology presented here is based on exmmely large very-steep peak and long low-level tail could
routine, daily, whole-sun patrol observations made at the have its mean Sm. greater than St, yet its flux density could
Sagamore Hill Radio Observatory over the years 1966 to lie below S( for the overwhelming majority of the burst
1978. This period covers the decline from the maximum of duration time T~.
the 20th solar cycle (a relatively low-amplitude cycle) through For each burst there is a factor by which (S~JSe)Td [for
the minimum (around 1976) to the rising portion of the new Sm < St] or T~ [for S~e > S1] exceeds the actual “above-
(and higher-amplitude) 21st solar cycle. Data from the the-level” time Te. We do not have the resources to go over
1966-1978 period should provide a reasonable measure of the time profiles of every burst at each frequency. Therefore
the intensity/duration parameters of solar-burst radio emis- we had to look at a representative sample of bursts and
sion during the least and most disturbed portions of the solar derive an overall factor to multiply the (S~JS~)Td or T~
cycle. obtained from the SGD data.
The Sagamore Hill data on radio bursts, published rou- Analysis of a limited number of relatively large events
tinely in Solar Geophysical Data(SGD) are given in terms from a Burst Atlas compiled by Barron et al. [1980] shows
of the following parameters: the peak flux density SP (sfu), that the factor by which (S~e/S~)T~ or Td overstates T,e
the mean flux density S~. (sfu) and the duration (rein) T~ varies between 1.5 and 10, independent of frequency, with
from start to end. These parameters are illustrated for the a mean of about 3. Therefore, we have used the following
idealized burst in Figure 11-6. The area under the curve is relationships to derive the “above the level” times of bursts
from the flux-density and whole-burst duration data listed
in Solar Geophysical Data;

Sv

1[ miadxf

Having discussed the data limitations and the approxi-

— -- mating adjustments, we now present the results of the burst
climatology study. In Figure 11-7, we show the number of
bursts (statistically) per 24-hr day exceeding 500% of the
I –T.i— TIME —
I
quiet-sun (flux-density) level. The numbers are given for
Td five frequencies (245 MHz, 610 MHz, 1415 MHz, 2695
MHz and 8800 MHz) over the years 1966 to 1978. In Figure
F]gure 11-6. Burst parameters depicted for a typical microwave event. 11-8 we show the number of bursts (statistically) per 24-hr
The symbols are explained in the text,
day exceeding a flux-density level of l@’9 W/m*Hz ( 1000
Sfu) .
the time-integrated flux density. The mean flux density S~e In Figure 11-9, we show the time duration (rein per 24-
is defined as the amplitude (flux density) of a rectangle with hr day) that bursts exceed (statistically) 500% of the quiet-
area equal to the time-integrated flux density and duration sun flux density level. These durations are given, as in the
equal to the burst duration T~. previous graphs, for the five patrol frequencies over the
In Figure 11-6, we have also shown an arbitr~ flux years 1966 to 1978. In Figure 11-10 we show the duration
density level, Se. In the climatology study, we eventually (rein per 24-hr day) that bursts exceed (statistically) the 10-’9
want to derive (from the SGD-listed data alone) the time W mz HZ-l flux-density level.
that bursts exceed a certain level. To illustrate the discussion The burst number and duration statistics (in Figures 11-
that follows, we have shown the level Se between SP and 7 through 11-10) are given on a “per 24-hr day” basis.
S~e (if Se > SP, the burst would never exceed the arbitrary Obviously, the sun is not in view (available to provide

11-12
 1

SOLAR RADIO EMISSION

I A BuR2T EIuRATION, MINUTES =R M-HR

,31EE
/\”’TriE’’”

,
.-,.
?
01 , 1 1 , , ,

Ias2 19s7 192s 19*S Isiu 1971 Im *73 1974 1975 !97s 1977 1979
YEAR

Figure 11-9. ‘lIre number of minutes per day that solar burst radiation,
Figure 11-7. The number of bursts per day that exceeded 500% of the
at various frequencies, exceeded 500% of the quiet-sun flux
quiet-sun flux-density level for the years 1966-1978. ‘
density.

interference) 24 hours a day from anY ground-based OPr- “outage” times obtained (- 1 rein/day) result from sporadic
ating location. On sites f~ from the eqttator (high latitudes), events during which the burst ,Pdiation from the sun may
sun-in-view hom per day vary widely with season. AIso> exceed the indicat@ thresholds for durations ranging from
the system susceptible to solar radio interference maybe in -10 min to 2 hours.
opration only during certain hours of the day (or night). At the higher radio frequencies (above 500 MHz), strong
Hence, the best method of pmsetitation is to give me sta- solar radio interference (above the quiet-sun level) occurs
tistics on a “per 24-hr day” basis @d let the user deteimine as distinct, individually-identifiable radio bursts. me bursts
what fraction of his operating time tlie sttn is actually akve have different spectral (frequency) characteristics, but at a
the horizon to provide intefierente to his system. Finally, given frequency the burst has a st~time, a peak (or several
it is important @ point out that the relatively”low ave~ge Paks), and a time of decay. At lower frequencies (245

10?w/m* m
‘t243 Nm— 1

YEAR

Figure 11-S. me numkr of bursts during a 24-hr day that exceeded, at Figuk 11-10. The number of minutes per day that solar-burst radiation,
vatious frequencies, i~ sfu. at various frequencies, exceeded 1000 sfu.

11-13
CHAPTER 11
1750 Iaoo 1810 UT la I=
I I 1 I I
I I
245 MHz I T II t
31 JAN 1970

Figure 11-11. The time-intensity profile of a typical 245 MHz noise storm observed at Sagamore Hill.

MHz, for example) there is, in addition to distinct radio

bursts, another phenomenon known “assolar noise storms.
~ese storms, more prevalent during the sunspot-maximum
portion of the 1l-year cycle, often last for many hours and
sometimes last for a period of days. The noise sto~s consist
of very irregular lo”w-to-moderate increases in solar flux
density above the quiet-sun le~el. For only a ve~’ small
portion of the noise Storm’s duration does the added flux
density exceed the moderate level (for example, 100 sfu).
Figure 11-11 shows a typical example of such a noise storm
observed at 245 MHz.
Since ~ese noise storms constitute another significant
type of solar radio interference for systems with operating
frequencies below 500 MHz, we have also included a sta- m
tistical graph on noise storm duration, based on our 245
MHz data. See Figure 11-12. The statistical duration is given
in terms of minutes of noise storm occurrence per 24-hr Figure 11-12. The average number of minutes per 24-br day that a noise
storm was reported by Sagamore Hill observatory for the
day. The shortened period plotted (only 1970 to 1978) gives years 197&197g. The large peak in 1978 may be due (at
some indication of the sunspot-cycle variation. least partially) to a change in reporting procedures.

11-14
 SOLti RADIO EMISSION
REFERENCES
Barron, W. R., E.W. Cliver, D.A: Guidice, and V.L. Bad- kundti, M.R. and T.E. Gergely (eds.), Radio Physics of
illo, “An Atlas of Selected Multi-~quency Radio Bursts the Sun, IAU Symp. No. 86, D. Reidel, Dordrecht,
from the Twentieth Solar Cycle,” AFGL-TR-80-O098 Holland, 1980.
ADA088220 1980. Kundu, M.R. and L. Vlahos, “Solar Microwave Bursts-
Castelli, J. P,, J. Aarons, D.A. Guidice, and R.M. Straka, A Review,” Space Sci. Rev., 32:405, 1982.
“The Solar Radio Patrol Network of tie USAF and its Ttiaka, H., J.P. Castelli, A.E. Covington, A. Kruger, R.L.
Application,” Proc. IEEE, 61:1307, 1973. Mndecker, tid A. Tlarnicha, “Absolute Calibration of
Forbes, J.M. and R.M. Straka, “Correlation BetW&n Ex- Solar Radio Flux Density in the Microwave Region,”
ospheric Temperatuk and Various Indicators of Solar Solar Phys., 29:243, 1973.
Activity,” AFCRL-TR-73-0378, AD766421 1973. Zheleznyakov, V. V., Radio Emission of the Sun and Planets
Kruger, A., Introduction to Solar Radio Astronomy and (in Russ.) Izdat. ‘Nauka’, Moscow 1964 Engl. Transl.
Radio Phys., D. Reidel, Dordrecht, Holland, 1979. Pergamon Press, 1970.
Kundu, M. R., Solar Radio Astronomy, Interscience, New
York, 1965.

11-15
 Chapter 12

THE AURORA
Section 12.1 J. A. Whalen
Section 12.2 R. R. O’Neil and R. H. Picard

12.1 PHENOMENOLOGY, km), E region (90-160 km), and F region (160-500+ km).
MORPHOLOGY, AND The particle energies and morphologies of each are depen-
OCCURRENCE dent on the source regions in the magnetosphere.

12.1.1 General Characteristics of the Aurora

The aurora is the name given to the light resulting from

‘“”~
the precipitation of electrons and protons from the mag- 300-
F REGION
netosphere into the earth’s atmosphere. This light consists
of atomic line spectra and molecular band spectra charac- E
*
teristic principally of oxygen and nitrogen, the chief con-
200-
stituents of the upper atmosphere ionized or excited by col-
—— —— —
lisions with these precipitating particles.
!-
Associated with auroral precipitation and ionization, A
●

currents called aurora] electrojet currents can flow in the 1oo-

—— —— ——

atmosphere. These currents produce magnetic fields de- D REGION

tectable at ground level by magnetometers and form the 0.1 1 10 100 1000

basis for magnetic indicators of auroral activity (see Chapter ELECTRON KINETIC ENERGY (keV)

4).
Subsection 12.1.1 is a qualitative introductory discus- Figure 12-1 Altitude of maximum production of ionization in the upper
sion of the general classes of aurora in terms of the precip- atmosphere by auroral electrons as a function of incident
electron kinetic energy. Calculation by Rees[1964]assumes
itating particles, the source of each type of particle in the
mono-energetic eleetrons of isotropic incidenee. Dashed Iiis
earth’s magnetosphere, the energy spectrum of each, the are the approximate boundaries of the D, E, and F regions
height at which each produces maximum ionization and which are the principal ionospheric altitude regions.
excitation in the earth’s atmosphere, the resulting electron
density for each, the spatial structure of each, and a map
of the general global morphology of each. The succeeding 12.1.1.2 Particle Energy—Magnetospheric Regime
subsections provide tables, graphs, and definitions that spec- Relation. These general source regions are the magne-
ify and characterize auroral properties in as quantitative tosheath, the plasmasheet, and the Van Allen or so-called
terms as are currently available. trapped radiation belts. These regions are occupied by elec-
trons of kinetic energies of -0.1 keV, 1–10 keV, and >40
12.1.1.1 Particle Energy—Ionization and Excitation keV, respectively. (Protons also exist in these regions, the
Height Relation. Particles penetrate into the atmosphere energies of which are greater than electron energies by a
to a height dependent on their initial kinetic energy, the factor of -5).
higher the energy the deeper the penetration and therefore Quite fortuitously, electrons that precipitate into the at-
the lower the altitude of the resulting ionization and exci- mosphere with energies characteristic of these three mag-
tation. netospheric regimes produce maximum ionization and ex-
The condition for electrons (monoenergetic and iso- citation in the three different altitude regimes; that is, -0.1
tropic) is shown in a graph of height of maximum production keV electrons ionize principally at heights in the F region,
of ionization versus electron energy in Figure 12-1. Note 1–10 keV in the E region, and >40 keV in the D region.
that the height scale is divided into three sections which are A cross section of the magnetosphere through the noon-
the three fundamental altitude regimes: the D region (50-90 midnight meridian schematically defines these regions to-

12-1
CHAPTER 12
gether with their particle populations and associated ionos- (IGY) and 50% in solar minimum (IQSY) [Sandford, 1968].
pheric height regimes (Figure 12-2). These two structural types can be distinguished in the
The origin of the magnetosheath plasma is the solar wind large scale photographs of the aurora by the Defense Me-
proton-electron plasma, which is approximately thermalized teorological Satellite System (DMSP) (Figure 12-3). The
in passing through the bow shock on the solar side of the bright structured forms at high latitude are the discrete au-

SOLAR

WIND

Figure 12-2, Schematic of the magnetosphere of the earth in noon-mid-

night cross section showing principal magnetospheric do-
mains, the associated particle populations via representative
electron kinetic energy, Ee, and the ionospheric region (F,
E or D) in which these particles produce maximum ioni-
zation.

magnetosphere. These magnetosheath particles are thought Figure 12-3. Two DMSP photographs showing the midnight aurora across
to be the source of the plasma sheet particles and they, in the North American continent defined by city lights,
turn, to be the source of the Van Allen radiation belt par-
ticles. However, the route of entry of the particles into these
two regimes and the processes by which they are accelerated roras; the relatively faint glow at lower latitudes is the con-
are not understood. As such, they remain the fundamental tinuous aurora. City lights show the location and scale of
unresolved questions of magnetospheric physics and the the photographs.
subject of much active research (see Chapter 8). It is not
thepurpose hereto deal with these questions, but it is well 12.1.1.4 Morphology of the E Region. The most fa-
to note that the fundamental links between the magneto- miliar model of the aurora is the Feldstein and Starkov
sphere and the ionosphere make the ionospheric measure- [1967] aurora] oval, a statistical compilation as determined
ments important ways of probing the magnetosphere in pur- from auroral all-sky camera photographs taken during the
suit of these answers. IGY. The oval is a band encircling the magnetic pole but
displaced eccentrically to higher latitude in the day sector
12.1.1.3 Spatial Structure. Another key distinction that as a consequence of the distortion of the earth’s magnetic
one must make bet ween auroras is their spatial structure, field by the solar wind. Figure 12-4 shows the oval projected
that is, whether they are relatively structured or unstruc- on a map of the northern hemisphere in magnetic coordinates
tured. This distinction is most evident in the 1 to 10 keV [Whalen, 1970]. The pattern, fixed with respect to the sun,
plasmasheet precipitation, which represents the bulk of the is shown in four different orientations with respect to the
energy of auroral precipitation. Here the structured com- earth at the indicated values of UT. With increasing mag-
ponent is known as the discrete aurora. Since it is the most netic activity, the oval expands and moves equatorward.
visible, it is popularly known as “the aurora”. The unstrttc- The oval typical of moderate conditions (that designated
tured component is known as the continuous aurora (also by the magnetic index Q = 3) will be used to form a ref-
called mantle and diffuse aurora). Although often neglected erence system for other types of auroras. The coordinate
because its faintness and absence of contrast make it difficult system of the oval, Corrected Geomagnetic Latitude and
to observe, this unstructured aurora comprised 8070 of the Corrected Geomagnetic Local Time, is shown in six dif-
total energy input into the polar region at solar maximum ferent plots in Figure 12-5. The Q = 3 oval is drawn in

12-2
 a
THE AURORA
aurora results from electrons alone since protons effectively
diffuse throughout distances of a few hundred kilometers in
the atmosphere. The continuous aurora will be examined in
detail in a subsequent section.

12.1.1.5 Morphology of the F Region. The F precipi-

tation appears to have structured and unstructured compo-
nents analogous to the E region (for example, all-sky pho-
tometric observations of Mende and Eather [1976]). The
location of each appears to be statistically well represented
by the oval—whether there are differences in the morphol-
ogy of the two is not known. The ionosphere at F layer
heights exhibits irregular and spread conditions throughout
these regions of soft precipitation. In addition, the polar cap
(the circular area poleward of the oval) is also a region of
spread F.
An additional ionospheric region of importance is the F
layer trough, a region of low ionization 50–10° wide located
equatorward of the oval in the night and evening sectors
[Muldrew, 1965]. The poleward edge or wall of the F layer
trough is a region of large electron density gradients that
can cause large refractions and thus large departures from
great circle propagation paths for HF. Buchau et al. [1978]
treat a specific case. Pike [ 1976] has developed an analytical
model of the F layer trough wall.
These F region precipitation patterns are shown in the
left column of Figure 12-5. Therrnalized solar wind from
the magnetosheath impinges directly on the atmosphere to
Figure 12-4. The auroral oval, the locus of maximum discrete (E) auroral
activity, projected on the north polar region as mapped in produce the dayside region but not the night sector regions,
Corrected Geomagnetic latitude and longitude. The orien- the field lines for which map down from the plasma sheet.
tations of the oval at four different values of UT are shown. In any case the regions are continuous in local time as
shown.
One of the complications in the F layer is that the regions
each plot bounded by the heavy lines. The shaded areas of production of ionization are not necessarily the bounds
schematically map the types of auroras differentiated as to of the location of the ionization. This is because ions in the
structured (top row) and unstructured (bottom row), and to F-layer have relatively long lifetimes during which they can
particle energy/ionospheric regime by columns. be transported considerable distances from their origin.
The E type auroras appear in the center column of Figure
12-5, the discrete auroral oval itself (above) and the con- 12.1.1.6 Morphology of the D Region. The unstruc-
tinuous aurora (below). The continuous aurora tends to over- tured component of the D region precipitation exists as a
lap the oval in the night sector [Whalen et al., 1977], but band generally a few degrees wide that extends from mid-
to locate equatorward of the oval in the day sector [Whalen night to morning and sometimes is entirely circumpolar [for
et al., 1971]. In quiet times the discrete aurora disappears example, Hook, 1968]. In quiet times its locus near midnight
entirely and the continuous aurora remains [Wagner et al., is the equatorward edge of the continuous aurora; in active
1973]. Although the continuous aurora forms a band that times the distribution can be quite wide. The temporal char-
encircles the pole for all levels of activity, its location can acteristics have been the subject of many studies [for ex-
change and the total energy fluxes can vary by several orders ample Hartz and Brice, 1967]. The occurrence of this pre-
of magnitude. In terms of the magnetosphere, the continuous cipitation causes D region HF absorption of relatively short
aurora appears to be precipitation from the plasma sheet, duration (15 minutes) near midnight, longer ( 1–2 hours) in
both having energy spectra which can be described as Max- the morning sector. Structured D region auroras seem to be
wellian [Rearwin and Hones, 1974]. associated with activity in the night sector as, for example,
The discrete aurora] particles also apparently arise from with poleward expanding arcs during substorms. Thus they
the plasma sheet but undergo additional accelerations. A are transient in nature and occupy the midnight sector of
further distinction is that the continous aurora can result the auroral oval. The patterns for the D region are shown
from both electron and proton precipitation. The discrete schematically in the righthand column of Figure 12-5.

12-3
CHAPTER 12

F E D
12 12

. ..- T--

IB 06 18 06 18 06

00 00 00

12

8 06

00 00

F]gure 12-5, A schematic representation of the regions of stmctured (top row) and unstructured (bottom row) auroras differentiated as to ionosr)heric
regime (particle” energy): from left to fight, F( -0.1 keV), Etl – 10 keV) and D(>40 keV). Coordinate systems are CC latitude and Cd local
time. The Q = 3 aurora] oval is defined by the heavy solid lines; the location of the specific auroras, by the shaded areas.

The unstructured pattern marks the region in which ra- paradoxical situation that the F region has been correctly
diation belt particles nun-trap” so as to precipitate into the defined as being the Feldstein oval in the dayside because
atmosphere. Berkey et al. [1974] have described a number the photographs have detected structured E auroras (un-
of events in which the dynamics of this precipitation are accountably) in the same location as the unstructured F
detailed. auroras.
The foregoing is not intended to be a complete descrip-
12.1.1.7 Composite. In Figure 12-5 two generalizations tion of the auroral ionosphere but is intended to identify the
can be made regarding the unstructured auroras. First, in minimum distinctions necessary for such a description. Fig-
the day sector the D, E, and F precipitation and ionization ure 12-5 presents, schematically, all six of these phenomena
production are at different latitudes (D lowest, E interme- together in hopes of clarifying their existence and inter-
diate, and F highest) so that there should be little question relation about which a great deal of confusion exists.
as to their separate identities. Second, in the night sector
the three overlap, a condition which had led to the confusion
that they are a single phenomenon since they can be in the 12.1.2 Distribution of Auroras in Magnetic
same place. Latitude and Local Time
For structured auroras, E and F categories are apparently
collocated at all local times and with D near midnight. Two The global distribution of auroras has proved to be a
features are of note regarding the relation between structured difficult subject to quantify. A principal reason is that most
and unstructured auroras. All six categories overlap in the studies have failed to distinguish among the large number
night sector. In the day sector, unstructured F auroras due of auroras, each of which can exhibit quite independent
to particles of magnetosheath origin are collocated with brightness and location on a small and large scale, short
structured E and discrete aurora, the energies of which cor- and long term.
respond to those of the plasma sheet. This has lead to the In addition, studies have been necessarily piecemeal,

12-4
 THE AURORA

/
c..r.ct, d 6.. . . ...1.. L... I Tn.. c.....*.4 G.. -.,..,,. L. Is*”4.
ground based studies being too small in scale and satellite
studies being too short in time. The Dynamics Explorer
(DE) lmager addresses these contradictory elements, and if
it can distinguish between the classes of auroras, promises
,.~l.
to provide valuable information in this area.
At the present time, the available distributions are gen-
erally statistical in nature and fall into two categories: mor-
phology of the aurora on the one hand and average values
of parameters on the other. However, the fo~er is accom-
plished without regard to intensity and the latter confuses
intensity with location or frequency of occurrence so that
the numerical values do not represent actual values. In ad-
dition these statistical measurements are ordered through a
magnetic index (usually Kp) that encompasses a large va-
riet y of auroral conditions.
A second approach has shown progress in defining quan-
titatively the instantaneous distribution of one of these classes
of aurora—the continuous aurora—throughout nearly its en-
tire latitude-local time extent. It finds that no single param-
eter, magnetic or otherwise, can quantitatively describe the
intensity and location of this aurora since both can vary
-.. -__--._----”-
quite independently. Rather this approach measures both
intensity and location and takes these parameters to be fun-
Flgure 12-6. Percentage of occurrence of discrete aurora in the zenith as
damental auroral indices.
recorded by ASCA in the Northern Hemisphere during the
Unfortunately neither of these approaches can be con- period 1957-1959.
sidered to give adequate quantitative descriptions of the
aurora in this scale at the present time: the statistical ap-
proach because of inherent limitations and the instantaneous response to increasing activity, the night sector expanding
approach because it has been demonstrated to date only on in width and day sector movitig equatorward. The standard
a limited basis. Nevertheless, this section will endeavor to deviation of these boundaries is about 10. No clear distinc-
present in each of these categories the best results which tion is made between discrete and continuous auroras. How-
are available at the present time. ever, they appear not to be exclusively discrete auroras.
Since the measurements were generally not in the zenith,
the continuous aurora would tend to be visible particularly
12.1.2.1 Statistical Distributions.
at large zenith angles due to VanRhijn enhancement. Thus
Frequency of Occurrence—The frequency of occurrence
the oval borders could be the continuous aurora and probably
of aurora in the zenith as determined by all-sky camera
are in the equatorward region since the discrete aurora tends
(ASCA) in one-half hour intervals during the IGY is shown
to be absent there [Lui et al., 1977; Whalen, 1983].
in Figure 12-6. The 70% contours define a band which was
Although no intensity info~ation was recorded, the
originally named the aurora] oval [Feldstein, 1966] and has
oval pattern has been found in a number of cases to be quite
since been called the auroral oval zone or the statistical
representative of the instantaneous pattern of the aurora (for
auroral oval. This is apparently the discrete aurora since the
example, the DE image—Figure 12-25 in Section 12.2).
continuous aurora has generally too low an intensity and
The solar cycle dependence of the oval is such that at
presents too little contrast to be seen in the zenith.
solar minimum the midnight sector (the only sector reported)
The Aurora/ Ova/—The auroral oval (originally called
was displaced 10 to 2° poleward of its location at solar
the auroral oval belt) is a band encircling the pole, the high
maximum [Starkov and Feldstein, 1970].
and low latitude borders of which are the average locations
The equatorward boundary &~, of the auroral oval cart
of the edges of auroral luminosity seen in ASCA photo-
be represented by the expression
graphs during the IGY. Such a band has been determined
for each value of magnetic index Q, where Q was deter-
mined within the same 15 minute intervals of UT in which 360°
Aw = 72° – 0.9°Q – 5.l”cos ~ k~ – 12°
the ASCA photographs were taken but from magnetometer
( )
measurements in the midnight sector near 65”CG latitude
[Feldstein and Starkov, 1967]. Figure 12-7 shows examples (12.1)
of ovals for three values of Q.
The ovals are thinner and displaced poleward in the day for Q >1 where Q is the magnetic index, ~~ is the CG
sector. In addition, the day and night sectors differ in their local time [Starkov, 1969]. This border can be closely ap-

12-5
CHAPTER 12
QUIET MODERATE ACTIVE

14

Is cm m
....

.... .... .... .. ......... .

Figure 12-7. Variation in the size of the auroral oval with activity as denoted by the magnetic index Q. Coordinate system is CC latitude and CG local
time.

proximate by a circle centered at 84.9”CG latitude 00:48 in producing large scale spatial integrals. However, it is
CG local time having a radius R = 18° + 0.9”Q [Gass- important to keep in mind Sandford’s [1968] caveat that
man, 1973]. Holzworth and Meng [1975] have derived sim- such distributions “give an average gross representation of
ilar expressions for the ovals. parameters that show rapid and large variations in space and
An example of the nature of the fit of such an offset time. ”
circle will be shown in Figure 12-11 for the Q = 2 oval. The most extensive set of distributions of important op-
Also shown in Figure 12-11 are other examples of other tical auroral parameters are due to Sandford [1968]. In-
auroral parameters that can be ordered in such an offset pole cluded in Figures 12-8 and 12-9 are contours in CG lati-
reference frame and which will be discussed below. tude-CG local time of intensity of emissions at 6300 ~,
The magnetic Q index is not generally available, an 5577 ~, and 3914 ~, and frequency of occurrence of dis-
unfortunate circumstance because it has been the most suc- crete aurora for three levels of activit y (quiet, moderate and
cessful single parameter in ordering auroral morphology. In disturbed based on magnetic local K indices of O-2, 34,
the absence of the magnetically measured Q, it is possible and >4 respectively) for solar cycle maximum ( 1958–1 959)
to infer Q by taking a measurement of A.~ at a single tcG, and minimum (1963).
substituting these values into Equation (12. 1) and solving The latitudinal distributions of these same parameters
for Q. By substituting this effective Q into Equation (12.1) are shown at midnight and at noon for values of K from O
one can calculate A.~ at other local times. to 8 for solar maximum (Figure 12-10, left) and solar min-
Average Values—A number of studies have provided imum (Figure 12-10, right).
distributions in magnetic latitude-local time of the average In addition, Sandford [ 1968] has shown conclusively
values of auroral parameters. One of the difficulties in this the existence of a non-discrete form of the aurora which he
approach is that the measurements are ordered in terms of termed mantle aurora. The term continuous aurora [Whalen
magnetic indices that typically span three hours. During et al., 1971] was coined to distinguish the E-region com-
such an interval auroral intensity and location are never fixed ponent of this aurora from the possible D-region component
so that a given Kp represents a large variety of aurora] which Hartz and Brice [1967] had named diffuse aurora.
conditions. In particular the aurora can move throughout a Sandford found that at solar maximum the mantle aurora
range that is generally much larger than its instantaneous contributed 80% of the total overall auroral energy flux and
extent. As a result instances of non-occurrence of aurora the discrete aurora accounted for only 20%. At solar min-
within a spatial interval are considered to be zero intensity imum, on the other hand, the two energy fluxes were equal.
and these are averaged together with cases of finite intensity The principal solar cycle difference in auroral energy fluxes
within that interval. Accordingly the resulting average can was that the mantle auroral energy flux was lower by a
produce values of intensity that never occur in nature; the factor of four at solar minimum than it was at solar maxi-
average spatial distribution is one than can never occur mum. Energy flux in the discrete aurora on the other hand
instantaneously. was the same at solar minimum as it had been at solar
Nevertheless these average distributions have been use- maximum although the spatial distributions at solar mini-
ful in revealing some overall trends related to activity and mum were narrower and more contracted toward the pole.

12-6
 MAGNETIC ACTIVITY 1958-59
DISTURBED MODERATE OUIET
00

12

Figure 12-8. Contours of occurrence of discrete aurora (in percent) and average intensity of auroral emissions (in kilo-Ray leighs) for 3 levels of activity
from Southern Hemisphere measurements in years 1958–1 959. Coordinates are CC latitude and CG local time. [Sanford, 1968] (Reprinted
with permission from” Pergamon Press. Ltd. @ 1968. )

12-7
 MAGNETIC ACTIVITY 1963
DISTURBED MODERATE

?2
00

\ 1

\ 1. /

Fl~ure 12-9.
.0 .—
Contours of occurrence of discrete aurora (in percent) and average intensity of auroral emissions (in kilo-Ray leighs) for 3 levels of activity
from Southern Hemisphere measurements in 1963. Coordinates ar= CG latitude and CG local time, [Sanford, 1968] (Reprinted with permission
from Pergamon Press Ltd. O 1968. )

12-8
— m

THE AURORA

1958-59 1963
I 1 I 1 I r I 1 1 1 1 1
400

-4
0 34
0 1 00 {$
m 6T
9
w
40 - \ :
/
‘\
u 1
&+,&:
x r 1 1 1 r 1
-4-

lm

— 40

10

z
— 4

r6\

‘O
1

A
34

1
4

50 60 70 eo 90 80 70 I 60 70 80 90 80 70 60

MIDNIGHT I NOON MIDNIGHT I NOON.

CORRECTED GEOMAGNETIC LATITUDE

Figure 12-10. The percentage occurrence of discrete aurora and the median intensities of 3914, 5577 and 6300 ~ emissions as a function of local K-
index and CG latitude along the noon-midnight CG time meridian for the years 1958– 1959 and 1963. [Sanford, 1968] (Reprinted with
prmission from Pergamon Press Ltd. @ 1968. )

Other examples of distributions of average parameters A.. = AO + aKp (12.2)

will be shown in Figure 12-11.
Equatorward Boundaries—The equatorward boundary where A.~ is the CG latitude of this equatorward boundary,
of precipitating electrons measured by DMSP satellites has AO is the latitude at Kp = O and a is the straight line slope
been related to magnetic index Kp via [Gussenhoven et al., 1983]. Values of these parameters are

12-9
CHAPTER 12
Table 12-1. Parameter values measured by DMSP satellites tector would place the apparent edge as much as 4° equa-
torward of that determined by a less sensitive detector. For
A= Ao+cYKp
statistical distributions the situation is worse judging by the
MLT Number* AO c1 Cct 3914A distributions of Sandford [1968] shown in Figure
12-10 (right).
Oooc+loo 312 66.1 – 1.99 –0.80 Offset Pole Coordinates—A number of auroral phenom-
0100-0200 220 65.1 – 1.55 –0.68 ena have been found to be ordered by a coordinate system
267 67.7 – 1.48 –0.57 that lies within the frame of reference of magnetic latitude—
0400-0500
1123 67.8 – 1.87 –0.71 magnetic local time but has an effective pole offset from
0500-0600
2462 68.2 –1.90 – 0.74 the magnetic pole. This offset pole lies near the midnight
0600-0700
68.9 –1.91 –0.76 meridian and has a colatitude near 4°.
0700-0800 3159
0800-0900 2159 69.3 – 1.87 –0.73 The auroral oval was noted earlier as having an equa-
0900-1000 1178 69.5 – 1.69 –0.66 torward boundary which is such an offset circle. The Q = 2
1000-1100 864 69.5 – 1.41 –0.57 oval is shown in Figure 12-11a together with a circle whose
1100-1200 513 70.1 – 1.25 –0.52 circumference is defined by the small circular points and
1200-1300 353 69.4 –0.84 –0.35 whose center is the circled cross. As noted earlier, this is
better approximated by an analytical expression then by a
1500-1600 63 70.9 –0.81 –0.34 circle.
1600-1700 204 71.6 – 1.28 – 0.66 Three other examples are shown in Figure 12-11 (b)
1700-1800 526 71.1 –1.31 –0.69 average intensity of auroral emissions at 3914 ~ determined
1800-1900 997 71.2 – 1.74 –0.82 by Sandford [1968] for IGY, (c) blanketing frequencies,
1900-2000 2469 70.4 – 1.83 –0.82 fbEs, of the ionized auroral E layer determined by Bes-
200(-2 100 3309 69.4 – 1.89 –0.82 prozvannaya and Shchuka [1976] for Kp O and 1 in IGY,
2100-2200 3092 68.6 – 1.86 –0.79 and (d) average precipitating electron energy flux measured
220(-2300 1482 67.9 – 1.78 –0.77 by satellites AE-C and AE-D in years 1974–1976 for
230C-2400 461 67.8 – 2.07 –0.81 AE < 100 y [Spiro et al., 1982].
*Number of boundaries. That the offset pole reference frame is not an artifact of
tCorrelation coefficient. the statistical averaging process is indicated by the existence
of auroral arcs seen nearly instantaneously (within approx-
imately 10 rein). Meng et al. [1977] have found a group of
given in Table 12-1. This relation is plotted for the 2300-2400 about 50 quiet auroral arcs imaged by DMSP satellites which
MLT sector in Figure 12-12 where the points are mean are offset circles. These arcs extend continuously over 12
values and the error bars are standard deviations. Also shown hours of magnetic local time and occur at all local times.
in this figure arethemidnight sector boundties determined Average radius of the arcs was 1Y ? 5°. Average offset
by the following means: (1) ASCA [Feldstein and Starkov, pole location was 85.8° CG latitude, 00:15 h CG local time.
1967], (2) DMSP Images [Sheehan and Carovillano, 1978],
and(3) 6300~ at 100R [Slateret al., 1980]. The trend in 12.1.2.2 Instantaneous Distributions. Recently prog-
all four cases is that latitude decreases with increasing Kp. ress has been made in quantitatively formulating the in-
However, there is considerable scatter which is due pri- stantaneous distribution of the energy flux in the continuous
marily to the fact that Kp encompasses a large range of aurora throughout much of its entire latitude-local time ex-
auroral conditions [Feldstein and Starkov, 1967]. An ad- tent. This occurred in an 8-hour case study in which this
ditional uncertainty can arise in particle data since the elec- energy flux was found to be constant instantaneously along
trons are measured at a single pitch angle and pitch angle contours that were defined analytically and approximated
distributions candepart appreciably from isotropy [for ex- by offset circles. Changes that took place did so simulta-
ample, Sharber, 1981]. neously (within 15 minutes time resolution) throughout local
This boundary has been found to be solar cycle depen- time, the extent of which was from evening through mid-
dent as noted earlier in relation to the auroral oval. In ad- night to noon [Whalen, 1983].
dition it has been reported to vary by several degrees as a The latitudinal distribution was therefore the same in-
function of UT, being at highest latitude near 0600 UT and stantaneously at all local times within the observed extent.
lowest latitude near 1800 UT [Sheehan and Carovillano, Two distributions were defined, the principal one being
1978; Meng, 1979; Whalen, 1983]. Gaussian. Furthermore, the Gaussian scale factor was con-
Finally the instantaneous distribution of energy flux typ- stant in all 32 consecutive determinations. Therefore, the
ically has a gradient with latitude that makes the location dynamics of the entire nearly global distribution could be
of an equatorward boundary a function of detector threshold. described by the variation of only two parameters, the Gaus-
Strickland et al. [1983] show a case where a sensitive de- sian maximum energy flux and its offset pole latitude. These

12-10
 THE AURORA

18 06 06

00
00

(a) AU RORAL OVAL (b) 1(3914~)

12 12

18 06

00

(c) fbEs (d) 0=

Figure 12-11. Examples of iso-energy flux contours of the continuous aurora which are approximately circles about a pole offset from the CC pole.
Circumference of each offset circle is defined by the 12 small circular points, and the offset pole, by the circled cross. (a) The auroral oval
for magnetic index Q = 2 [Feldstein and Starkov, 1%7]. (b) Median intensity of 3914 ~ auroral emissions for ICY [Sandford, 1967]. (c)
Auroral E ionization measured by median values of fiEs for Kp of O and I for 1958 [Besprozwnnaya and Shchuka, 1976]. (d) Average
precipitating electron energy flux for AE s 100 for 1974-1976 [Spiro et al., 1982].

12-11
CHAPTER 12
70”7 ages (85. 8°CG latitude, 00:15 h CG time). Although named
+ ELECTRONS
the offset pole coordinate system, the formulation is an
analytical expression similar to that found by Starkov [1969]

h
0 OMSP
(
A ASCA
[Equation (12. 1)] for the low latitude boundary of the auroral
oval to which the offset circles themselves are approxi-
‘--6300~
65”- mations. In this formulation the latitude r of a point on the
earth is given by
A eq

(CG Lat. )
8 r = Ac~ + ep cos ~ (UT – UT~N) (12.3)
60°-
A

k
‘.,
. where ACG is the average latitude and has the value of the
.

t
+ CG latitude of the point, ep is the amplitude of the lati-
tudinal variation and has the value of the CG colatitude of

I
●
55°- the offset pole (4. 2° here), UT is universal time and UTMN
is the universal time corresponding to CG local midnight
for the point considered. During this case study, contours
of equal latitude r calculated by this expression were mea-
●
sured to be contours of equal energy flux Q, which were
50° 1
accurate to within 0.4° over the CG local time extent of 10
2 6 8 hours.
KP4
The function that describes the distribution of energy
flux, Q, in offset pole latitudes r is
Figure 12-12. Examples of auroral equatonvard boundaries near midnight.
Electrons: Precipitating electrons [Gusseohoven et al., 1983]
ASCA: All-sky camera photographs [Feldstein and Starkov,
1967] DMS~: DMSP images [Sheehan and Carovillano,
1978] 6300 A: Photometer [Slater et al., 1980]. Error bars Q(r)= Qmaxexp[-12(r-fl) (12.4)
indicate standard deviations for the precipitating electrons
but are also comparable for the other measurements.

where Qmax and r max define the peak value as shown in

Figure 12-13.
parameters are therefore global indices of this aurora ex-
The Gaussian scale parameter u has been found to have
pressing independently the location and intensity of this
aurora. a typical value of 1.4° (3. 2°FWHM) and extremes of 1.2°
A second latitudinal distribution was found at higher and 2.0° [Whalen, 1981, 1983]. The maximum energy flux
latitude. Gaussian and high latitude distributions corre- Qmax has a typical value of 1 erg cm-’s-’ and extremes of
sponded to the regions of particle precipitation named Cen- 0.25 to 12 erg cm-’s-’ [Omholt, 1955; Hilliard and Shep-
herd, 1966; Whalen, 1981, 1983].
tral Plasma Sheet and Boundary Plasma Sheet by Win-
ningham [1975]. The distributions in magnetic latitude and The latitude of the Gaussian maximum, rmax, has a
local time are shown schematically in Figure 12-13. median vlaue of 710 and extremes of 64° to 74° [Whalen
1981, 1983]. The circular aurora] arcs of Meng et al. [ 1977]
These distributions have additional significance in that
had an average latitude of 71° * 5°, which is consistent
the latitude of the Gaussian maximum, rmax, is at or near
with their being interface arcs [Robinson et al., 1982] at or
the interface between positive and negative field aligned
near rmax.
currents [Robinson et al., 1982]. In addition, rmax is the
The spectrum of precipitated (that is, loss cone inte-
equatorward limit of discrete auroras for a large range of
grated) electrons in the Gaussian/CPS region is Maxwel-
conditions [Lui et al., 1977; Whalen, 1983]. Thus the Meng
lian, so that the differential number flux (in electrons
et al. [1977] auroral circles apparently also locate rmax.
cm-zs-’ keV-’ ) is
Finally, Heelis et al. [1980] find the convection reversal
region near the CPS-BPS boundary.
The latitudinal coordinates used in this case study are
those approximated by the offset pole at 85.8° CG latitude
and 00:00 hr CG time. This pole is determined from the
@(E)=& Eexp
()
~ , (12.5)

ionospheric data of Besprozvannaya and Shchuka [ 1976]

(Figure 12- I It), and is also very close to the average value where E is electron kinetic energy and EM the characteristic
which Meng et al. [1977] determined from the DMSP im- energy.

12-12
 THE AURORA
12

rmax
I II i Ii HIGH

FT”DE
‘Qmaxti “., :-
06

“.
“.
\l
“.,
:
::
...”
:“
..”

4
KEY:

v 00

Figure 12-13. Schematic diagrms illus&ating theinstant~eous distibution ofenergy flux intiecontinuous aumraas dexribd inthetext. Le~r Energy
flux, Q, as a function of offset pole latitude, r, showing the Gaussi@CPS and high latitude/BPS distributions. Right: Contours of constant
energy flux approximated as circles centered at the offset pole (the circled cross) in CG latitude and CG local time. The offset pole latitudinal
distribution, which is therefore the same at all Iwaf times except for a region in the day sector, isthe Gaussian/CPS distribution shown at
tlte left. The contour of the Gaussiti irraximum energy flux, Q~,,, as well as the contours of 1~2 Q ~.x and 1/10 Q~,, are identified via the
sector marked “Key” in the lower Ieti. The dashed contours in the dav sector indicate tlie approximate region in which statistical studies
find energy fluxes ~o b lower than at other Iwat times,

EM has been found near midnight also to W Gaussian 12.1.3 Empirical and Model Relationships
in latitude with a scale parafneter close to that of Q.
E~max has been measured at 1 keV [Shtiber, 1981]. 12.1.3.1 AuroraI-Ionospheric Profiles. This section de-
Infernng particle energies from the pr~uction heights mea- scribes the ionospheric effects of auroral p~icle precipi-
sured by Hunten et al. [1963] for this aurora (which they tation calculated by a numbr of workers. Tlie effects of
termed diffuse surfaces), E~&ax was gpneraljy between 0.4 continuous and discrete electron auroras are presented as
and 1.5 keV with extremes of 0.3 and 9 keV. The depen- well as those of proton auroras. In general, isotropic pitch
dence of E~max on local time however has not &n es- angle distribution of particles is assumed. Other distribu-
tablished. tions have been studied as well as ehergy spectra that w
Because of the equivalence of the particle, optical, and different tim those assumed he~ [for ex~ple, Rees, 1%3].
ionospheric parameters for the continuous aurora and auroral Corrtihuous Aurora—The electron spectrum of the con-
E-layer (as described in Section 12.1 .3.2 and Figure 12-19) tinuous aurora has been found to & Maxwellian. Sharber
the following is true: contours of equal loss-cone ;nt#grated [1981] has found good agreement between precipitating
electron energy flttx Q= are also contours o} equal column electrons fluxes and their simultaneously measured ionos-
integrated NJ emission at 3914 ~, 1(3914 ~), as Well as pheric E-layer effects provided the particles are in@grated
cofitours of equal foEa4; where foEa is the aurotal E layer over the actual loss cone and averaged over intervals of time
critical frequency. Furthermore, because these three param- and space appropriate to the response of the ionosphere itself
eters are pro~rtional to one another, for a Gaussian lati- and of the characteristics of the ionospheric measurements.
tudinal distribution, the three have the same Gaussian scale The energy spectrum which results from this procedure is
parameter. well represented by a Maxwellian function.
Contours of equal foEa4 are also contours of equal max- The Maxwellian differential electron number flux as a
imum electron density, Ne~aX,as well as contoufi of equal function of energy has the form
foEa. However, for a Gaussian latitudinal distribution, the
scale parameters are related by the following multiple pro-
portionality:

u (foEa4): u (Ne~,X): u (foEa) = i: ~: 2. (12.6)

+(E) = ~ E exp
()~ electrons cm – 2s- ‘eV-’,

(12.7)

12-13
CHAPTER i2

n
200 [ 1 200
\ \ \ 1
MAXWELLIAN ~
Em: .5 keV
I 80 - I 80

MAXWELLIAN +0
160 -
$
g w
3 140 s 140 1
+ l.OkeV
G I
d
I 20 -

100 - 100

1 J
10’ 04 10. 1Os 106
PRoOUCTl~ RATE (cm-’-s-’) ELECTRON DENSITy(Cm-’)

ELECTRON piECIPITATION

MAXWELLIAN SPECTRA (CONTINUOUS AURORA)

Figure 12-14 Altitude profiles of ion production rates (left) and electron densities (right) resulting from electron precipitation which produces the continuous
aurora. Calculations [Strickland et al., 1983] assume Maxwellian spectra, isotropic incidence and ener~v -. flux of 1 erz cm-zs-’. Curves for
4 values of Maxwellian characteristic energy, E~, are shown.

where energy E is itl eV, energy flux QM is in eV cm”zs-’ Discrete Aurora—The electron spectra that produce the
and EM is the characteristic energy (or temperature). discrete aurora have been modeled by either monoenergetic
Since the average energy for the Maxwellian is 2EM, or Gaussian s~ctra. Although the actual spectra are more
the total integral number flux, N~ (electrons cm-2s-’ ) is accurately modeled by a Maxwellian spectrum that has been
given by NM = QM/2E~. accelerated through a fixed potential, the Gaussian is an
Other useful relations can k derived from adequate approximation for the calculation of the iono-
NM = eE~@(EM) where @(EM) is the maximum value of spheric response.
+(E).
Altitude profiles of the rates of ion-electron pair pro-
duction per unit volume resulting from this Maxwellian s , 1 , !11[ 1 1 1 I 11111 1 1 I , 1,1,1 , t I 1 lrn-
spectrum have been calculated by Strickland et al. [19831
for four values of EM and are shown in Figure 12-14 (left). 250 -
An energy flux of Q~ = 1 erg cm-2s-’ has been assumed; z
=
profiles for other values of Q~ can be determined from those w 200 -
n
shown since ionization rate is proportional to QM. In ad- a
~
dition these profiles can be converted to energy de~sition ~ 150-
● ELECTROIS8
per cm3 by multiplying the ion production rate by the factor
34 eV per ion pair. 100 -

The electron density profiles resulting from these pro-

duction profiles have been calculated by Strickland et al. 50-’ 111111 t , 1 11,,1 I , 1 ,,11,1 1 1 ,1lG
0.1 I 10 100
[1983] and appear in Figure 12-14 (right). In these calcu- CHARACTERISTIC ENERGY, EM (kev)

lations equilibrium conditions have been assumed since they

are appropriate to the slowly varying nature of this aurora Figure 12-15. Altitude for the maximum ion production rate versus Max-
and the E-region. wellian characteristic energy for ionization resulting from
A plot of the altitude of the peak ionization rate as a the precipitation of particles with Maxwellian spectra and
at isotropic incidence.
function of characteristic energy is shown in Figure 12-15 Le~t: Electrons which produce the continuous aurora.
displaying the agreement in the results of several workers. Rig/rf: Protons.

12-i4
 THE AURORA
200
GAUSSIAN q

L
I 80

)2.0
EG=I keV

keV

1-
,33keV
2
120

. I
.fan 1 1

102 103 104 104 d 106

PRODUCTION RAT E(cm-3-s-’) ELECTRW WNSITY(cm-s)

ELECTRON PRECIPITATION
GAUSSIAN SPECTRA (DISCRETE AURORA)

Flgu~ 12-16. Altitude profiles of ion production rates (left) and electron densities (right) resulting from electron precipitation which produces the discrete
aurora. Calculations [Strickland et al., 1983] assume Gaussian spectra, isotropic incidence and energy flux of 1 erg cm-2s-’. Curves are
shown for6 vahsesof Gaussiati maximum energy, &

The representation of the discrete differential number Basu [1982] assuming isotropic incidence. Figure 12-18
flux is show:s a profile for each of five values of chhacteristic
energy, EM for energy flux of 1.0 erg cm-2s-1. A plot of
the altitude of the peak ionization rate as a function of
+(E)=&exp - electrons cm-2s-leV-]
m1’2uE~ [ (-)21 characteristic energy is shown in Figure 12-15.

(12.8)

where & is the maximum energy in eV. The Gaussiatt scale ELECTRON PRECIPITATION

parameter u has values near 0.2 &. MO NO ENERGETIC SPECTRA

Altitude profiles ofionization rates forvalues of&and (DISCRETE AURORA)

1
for Q~ = lergcm-2s-1 appear in Figure 12-161eft and the
i
resulting electron density profiles are shown in figure
12-16 right [Strickland etal., 1983] .Again electron density
has been calculated assuming eqtiilibnum and so applies
specifically to auroras that are constant in all respects for

_%J
peri@sgreater than about 1 min. Forauroras that are more
1 1
transitory, these electron density prctfiles can be considered
to be upper limits.
The altitude of maximum ionization rate (points) and
the altitudes at which the rate has fallen to 10% of maximum o.I 10 lofJ Iwo
(bars) is shown in Figure 12-17 for the case of monoener- ENERGY ( KaV )

getic electrons of isotropic incidence [Rees, 1964].

Proton Aurora—Proton spectra have been found to be Figure 12-17. Altitude of the maximum ionization rate (points) and the
consistent with Maxwellian spectra [for example, Sharber, altitudes at which this rate has fallen to 10% of the max-
imum (bars) versus energy, for ionization produced by
198 1]. Ionization rate profiles for precipitating protons with electrons which form the discrete aurora as’ modeled by
Maxwellian spectra have been calculated by Jasperse and monoenergetic spectra of isotropic incidence [Rees 1964].

12-15
CHAPTER 12

..—>

102 10’ 104

IONIZATION RATE ( cm-’s-’) ,,
I
,
,

Figure 12-18. Altitude profiles of ion production rates resulting from 1

1
precipitation of protons of isotropic incidence and Max- 1

wellian spectra of characteristic energy EM.Energy flux is 1

I
1.0 erg cm-zs;’ [Jasperse and Basu, 1982]. o
#
#
1
1
I
I
1
12.1.3.2 Relations Between Auroral Electron, Optical, 1
1

and Ionospheric Parameters. Auroral studies m con- I

t

ducted in three major disciplines: pdicle, optical and ion- I

1
ospheric. Under some conditions parameters in all three are ;
closely enough related so that a measurement in one dis- I
1
1

cipline permits inference of parameters in the other two. ,

I
,
This section presents a graph interrelating these parameters I

in a convenient form as well as a description of the con-

ditions under &hich these relations apply.
The parameters plotted in columns in Figure 12-19 are
as follows: Figure 12-19. Inter-relations of particle, optical and ionospheric pm-
IB,~—~e International B~ghtness Coefficient, a semi- eters for the continuous aurota and aurora] E layer.
quantitative index for visible auroras. lBC: International Brightness C~fficient.
Q: Ener Y flux of precipitating auroral electrons.
Q-Energy flux in erg cm-2s-’ of precipitated electrons. 1(3913 1 ): Column’integrated intensity of the N; emission
These are loss cone integrated and subject to band at 3914 ~.
the space and time constraints described in foEa4: Fourth pdwer of aumral E layer critical frequency
for the ordinary ray.
Section 12.1.3.1. Ne_,: Electron detrsity of the auroral E layer at its max-
1(3914~)-Column integrated intensity of the NJ emis- imum in altitude.
sion band at 3914 ~ produced by auroraI par-
ticle precipitation.
The above relationships are well established [Dalgarno The ionospheric parameters obey these quantitative re-
et al., 1965] and apply in the E-region to both discrete and lations only for the auroral E layer, which is produced by
continuous auroras. In addition, the relation between Q and the continuous (or diffuse) auroral precipitation. This layer
1(3914 ~) is correct for the D-region although IBC is not is extensive, highly uniform and is in equilibrium. foEa,
defined there. the critical frequency of the ordin~ ray measured by the

12-16
 THE AURORA
ionospheric sounder, is the maximum plasma frequency of the relation of occurrence frequency to magnetic indices Kp
the E-layer and has been related empirically to both Q and and AE and dependence on UT. In Figure 12-20e the col-
1(39 14 ~) via its fourth power, foEa4 [Omholt, 1955; Shar- umns are headed by activity observed on a given orbit; the
ber, 1981] as shown in column 4. rows are the occurrence frequencies observed on the fol-
Ne~,X is the peak electron density in the auroral E-layer, lowing orbit.
being derived via foEa = 8.98 X 103 fi~aX where foEa Auroral images have been recorded by ASCA at 1-rein
is in Hz and Ne~,X in electrons per cm3. intervals aboard an aircraft flying at local midnight for du-
rations of 5– 10 h [Krukonis and Whalen, 1980]. This data
12.1.3.3 Auroral Activity. The discrete aurora on the has permitted analysis of aurora] activity over a considerable
small scale is frequently very unstable and chaotic. How- range of time durations as distinct from the DMSP images,
ever, certain ordering of brightness and extent has been the sampling rate of which is limited by the orbital period.
achieved on the large scale using DMSP auroral images the Twelve flights totaling 93 hours provided 5600 photo-
breadth of which is 2200 km, and airborne ASCA photo- graphs in the midnight sector near 69 ( ~ 4°) CG latitude.
graphs, with a diameter of 900 km. The two data bases When mounted in a montage format these photographs re-
represent very different time scales, DMSP images being vealed individual events of aurora] activation with meas-
separated in time by 100 min and the ASCA by 1 min. urable lifetimes. Categorizing the auroras in the above man-
Auroral images by the DMSP satellite have been cate- ner produced the histograms of the lifetimes or aurora] activity
gorized as active, moderate, or quiet as described in Table shown in Figure 12-21. The combined A, M, and Q cate-
12-2 [Sheehan et al., 1982]. The distribution of frequency gories indicate a characteristic lifetime of auroral activity
of occurrence appears in Figures 12-20a, b, c, and d showing near 15 min (Figure 12-22).

Table 12-2. Description of the criteria used to rate auroral activity in DMSP images,

Latitudinal
Category Aurora] Morphology Extent Longitudinal Extent Comments

N No aurora visible in a properly Seldom observed in a good

exposed image having adequate quality image
aurora] zone coverage.
Q Quiet auroras consiting of a single, s 1° CGL Up to the entire longitudinal
thin, structureless discrete arc (discrete arc) extent of the image (= 3– 12
or broader band of diffuse aurora hours of local time depending
[Lui et al., 1973] also called on coordinate orientation)
continuous aurora [Pike and
Whalen, 1974].
M Moderate state of auroral s 5° CGL A few hours of magnetic Statistically most common
activity typified by patches, knots, local time, usually not the situation
or loops in discrete auroras, in- entire oval
eluding ‘westward surges’ and
bulges along otherwise thin arc
structures: also, polar cap arcs
extending from oval auroras.
Gradations in the moderate range
(M –, M, and M+) attempt
further refinement, but distinctions
are often difficult to make,
particularly between M + and A.
A Very active auroraI conditions with = 5° CGL The entire longitudinal M + and A categories often
(apparent) large westward surges, extent of the oval visible difficult to distinguish
poleward bulges, and a broad band in an image
of discrete aurora. With linear
amplifiers employed in older
satellites, bright auroras can
saturate and wash out all detail.
Newer satellites with logarithmic
detectors can distinguish detail
over a wide range of intensities.

12-17
CHAPTER 12
DMSP ACTIVITY ON INITIAL ORBIT
A M Q

(a) DMSP AURWAL ACTIVITY

Kp< 2- 2- SKPS3+

Mu
u>
Zv (219) ( 376)
Wz
au
~ ~ 0.5
“u
“m
0“

NMQA

(b) RELATION OF ACTIVITY TO KP

1
AE>30Q
(233)

(e) RECURRENCE OF ACTIVITY

Figure 12-20. (Continued)

(c) RELATION OF ACTIVITY TO AE

12.1.3.4 Heigh@ of Discrete Aurorm.
Dependence of Heights on Magnetic Latitude and Local
Time—The altitude and latitude of peak emission at 5577
~ have been measured for homogeneous arcs by photom-
eters using triangulation methods IBoyd et al., 1971]. The

L
04-08 08-12
(195) (153) resulting heights versus magnetic latitude are plotted in Fig-

LIFETIMES OFAURORALACTIVITY

20- No of %}:1
NQMA 10-
TyF.e Events

r
I .0 0
31 22

E 20-24
r
12-16 16-20

L
(106) ( 186) (74)
}
37

N QMA
24

(d) DEPENDENCE OF ACTIVITY ON UT

Figure 12-20. Frequency of occurrence of auroral activity observed in the

DMSP images in categories of Active, Moderate, Quiet 17

and No visible aurora as defined in Table 12-2 [Sheehan 10 30 50 70 90 110 130 150 ~!””t.,

et al., 1982].
a. All cases.
b. Relation to Kp. Figure 12-21, Distributions of lifetimes of auroral activity within cate-
c. Relation to AE. gories descrikd in Figure 12-20 as determined by airborne
d. Dependence on UT. all-sky photographs nea midnight. Number of events witiln
e. Recurrence of activity where columns are headed by each category and tbe pereent of the total time represented
activity obsemed on initial orbit; the 3 rows are the distri- by each category m listed at the right [Kmkonis and Whalen,
bution of occurrence frequency on the following orbit. 1980]

12-18
 THE AURORA
nique requires relatively stable arcs, hence the results apply
Llfetlmes of Combined Aurorol Events
to relatively quiet conditions.
(A, M8Q) Height-Brightness Relations—A number of workers have
reported that discrete auroras increase in brightness with
decreasing heights as determined by optical triangulation
[for example, Starkov, 1968 and Harang, 195 I: Boyd et
total no events = 186 al., 197 1]. Evidence for this effect, which is more quan-
titative in nature and not as dependent on aurora] stability,
has resulted from a study of discrete aurora by airborne
ionospheric sounder and ASCA during I I north-south cross-
ings of the auroral oval near midnight [Wagner and Pike,
1972; Gassmann et al., 1972]. Auroral sporadic E echoes
recorded by the ionospheric sounder were selected for times
when the ASCA films showed aurora to be in the zenith.
Virtual height for these echoes, h’Esa, is plotted versus the
0
-1
30 60 90 120 150 m,nutes
frequency fEsa in Figure 12-24, together with the qualitative
designation of brightness from the ASCA photographs. The
Figure 12-22. Percentage occurrence of lifetime of auroral activity for trend that brightness increases with decreasing height can
combined active, moderate and quiet categories (from F]g-
ure 12-21) showing Poissonian lifetime of -15 minutes. be described in a more quantitative way by the approximate
translation of frequency to intensity of 39 i4 A emission via
Figure 12-19. Height can be translated into approximate
ure 12-23 in four 3-hour intervals of local time encom- monoenergetic electron energy via Figure 12-17.
passing evening through midnight to morning. Heights have
been related to the energy at which monoenergetic electrons
produce maximum emission via the relations of Rees [1963] 12.2 AURORAL OPTICAL AND
and these energies are plotted on the right hand scale. INFRARED EMISSIONS
Two relations are apparent: heights decrease with lati-
tude within each local time interval and heights decrease as The optical emissions that characterize an aurora are
local time increases from evening to morning at each latitude produced by the deposition of energy in the atmosphere by
where the aurora is observed. energetic particles, primarily electrons and, in some in-
The limitation of this study is that the triangulation tech- stances, protons. The energetic particles are derived from

666768 @707772 G676869~71?2

D@ole Latitude

Figure 12-23. Distributions of the height of the peak emission in 5577 ~ as a function of dipole latitude for four 3-hour time periods shown as 150”WMT.
Local magnetic midnight is approximately 0115 150”WMT. The median point of the distribution in each latitude interval is sh[)wn by a
circle,

12-19
CHAPTER 12
h’EW
o
~ 150
0
a
w
y 140 .
. ..
0 .
m
u
130 . :. ● .

;m
E . ●.
x . .
[

2345676 9 10 MHz
faint modl:~AMght brigM fE~
au?wa aurora

Figure 12-24. Relations between brightness (from air’bome ASCA), au-

roral spradic E frequency, ~sa, and virtual height of
auroral sporadic Elayer, h’Esa(from airbomeionospberic Figure 12-26. Photo of the earthlimb seen from Skylab 3 Iwking toward
sounder) for discrete aurora in the zenith. from 11 lati- the southern auroral zone, The airglow layer runs in a thin
tudinal scans across the oval near midnight by aircraft band across the center of the photo, while a bright auroral
[Wagner and Pike, 1972]. arc crosses the earthlimb at lower right and ends in the
foreground in the middle of the scene [Garnott, 1979]

the interaction of the solar wind, a gaseous plasma contin-

satellite images of auroral luminosity in the polar regions
uously streaming from the sun’s surface, and the earth’s
have contributed significantly to the understanding of the
magnetic field [Chapter 8]. The currents of charged particles
interaction of the solar wind and the magnetosphere, auroral
generated by the energy coupling between the solar wind
substorms, and the morphology of the diurnal cycle of the
and the magnetosphere produce energetic electrons surging
auroral ovals surrounding the geomagnetic poles [Akasofu,
earthward along geomagnetic field lines in the auroral zone.
198 1]. Aurora] optical emissions have been examined by
Collisions of the precipitating energetic auroral particles
instrumentation located in ground-based observatories
with the increasingly denser atmospheric neutral atoms and
[Vallance Jones, 1974], research aircraft [E.R. Huppi et al.,
molecules in the 200 to 100 km altitude range produce the
1974], rocket probes [Stair et al., 1975; K. D. Baker et al.,
luminosity associated with auroras illustrated in Figures
1977; Stair et al., 1980; Feldman and Gentieu, 1982], and
12-25, 12-26, and 12-27. As reviewed in Section 12.1,
satellites [Hanson, 1973; Huffman et al., 1980; Frank et
al., 1982]. These studies involve measurements of spectral
distributions, photon emission rates, and altitude profiles.
The data are then analyzed to identify the excited states of

Figure 12-25. View from space of intense aurora] emission distributed

around entire auroral oval. Image was taken at 0241 UT,
8 Nov 1981, with University of Iowa’s ultraviolet auroral
imaging instrumentation akard Dynamic Explorer 1 while
orbiting 20000 km over North America. (Courtesy of L.A. Figure 12-27. Bright auroral arcs recorded from the ground at Poker Flat,
Frank. ) Alaska. (Courtesy of AL. Snyder. )

12-20
 THE AURORA
the radiating atomic and molecular species and to determine by the primary electron. Once the ionization rates are known,
their auroral production and loss processes. the production rate of secondaries as a function of energy
Auroral phenomenology, morphology, and occurrence and the seconda~ electron flux can be calculated [Rees et
are the subject of Section 12.1. The current section will al., 1969], using analytic or experimental values for sec-
concentrate on the overall photofi emission rates. spectral ondary electron production cross sections and electron-neu-
distributions, and altitude profiles of auroral optical emis- tral inelastic cross sections. From the secondary electron
sions and the identification of the radiating species and their flux, excited-state production rates may be deduced. Other
production and loss mechanisms. Particular emphasis will descriptions of electron energy deposition and excited-state
be placed orI the infrared portion of the spectrum. Only production have used range-theoretic methods [Stolarski and
emissions from aurorally disturbed atmospheres will be Green, 1967; Peterson et al., 1973], Monte-Carlo tech-
discussed; quiescent atmospheric emissions are treated in niques [Berger et al., 1970], and Fokker-Planck [Banks et
Chapter 13. al., 1974] and Boltzmann [Strickland et al., 1976] transport
This section begins with a discussion of laboratory stud- equation methods. In addition to the above time-independent
ies and their impact on aurotal models and continues with approaches, recent time-dependent models taking atmos-
a general discussion of emission rates in terms of production pheric motion into account have been formulated IRoble
and loss mechanisms for atiroral species. Processes giving and Rees, 1977].
rise to various aurora] visible and ultraviolet emissions are The models have been tested against the results of field
discussed, and the section concludes with a discussion of experiments. Advances in instrumentation design and sensor
the dominant aurorally enhanced infrared bands, especially technology have resulted in steady improvement of the pre-
the important C02 4.3 pm emission and the NO 2.7 Am cision, spectral resolution, and minimum detectable emis-
and 5.3 pm emissions. sion level in the field and laboratory experiments. The re-
sulting improved and expanding data set has broadened the
scope and increased the precision of our understanding of
12.2.1 Laboratory Studies and Aurora] auroral processes.
Models

Auroral mechanisms are studied in laboratory experi- 12.2.2 Artificial Aurora] Experiments
ments involving particle beams incident on atmospheric gases
or gas mixtures. Such experiments measure specific param- In recent years, rocketbome electron accelerators have
eters that quantify the incident electron energy loss, energy been used in a variety of active experiments in the upper
deposition profile, and excited-stated production [Cohn and atmosphere [Winckler, 1982], including the investigation
Caledonia, 1970], as well as the spectral, temporal, and of wave-particle interactions, beam-plasma discharge, par-
spatial radiative characteristics of the gases. Gas dischtiges ticle trapping in the geomagnetic field, generation of radio-
have been used in flowing afterglow experiments [Piper et frequency electromagnetic waves, and artificial auroral
al., 198 1a, b] and chemiluminescent reaction cells [Rawlins experiments (Figure 12-28). A series of artificial auroral
et al., 198 lb] to determine rate coefficients and photon experiments using rocketbome electron accelerators has &n
yields for specific auroral reactions. The laboratory exper- conducted at AFGL under the EXCEDE program [0’Neil
iments provide an opportunity to study a specific production et al., 1978a and b, 1979, 1982]. In artificial aurora] ex-
or loss process for an excited species under controlled con-
ditions of temperature and gas density. GEooss 1- w 1977
Measured parameters, including cross sections for en- 10kW ReCX~ ELEm ~RA~
ShV 8E~, FrlYLWD WITUOE 102 h
ergy loss by incident electrons and for production of specific PiTeH ANGLE = W-SXI -ES

new species or excited states [Wadzinski and JasperSe, 1982],

have contributed to the generation of a number of aurotal
models. Such models typically include descriptions of elec-
tron energy deposition and excited-state production, usually
followed by chemistry codes to describe production of ad-
ditional species and perhaps by an atmospheric motion code
and a radiative transfer code. Descriptions of electron in-
teraction with the ambient atmospheric gas have used a 06 30

variety of techniques. Simple semi-empirical models of TtME AFTER PUL% INMATION [%-)

electron energy deposition have been constructed [Rees,

1963] using energy dissipation functions measured in the Rgure 12-28, Ground-based image of emission from EXCEDE artificial
laboratory. The secondary electron production rate in such aurora] experiment taken through 3 1-inch GEODSS tele-
scope at White Sands, New Mexico. The prompt emissions
models is calculated from the observation that an electron- and afterglow are shown at two different times after pulse
ion pair is produced for every 34 eV of energy deposited initiation.

12-21
CHAPTER 12
periments such as EXCEDE, the dose conditions (primary The intensity of atomic and molecular transitions from
electron energy, ionization rate, dose time, and dose alti- upper state u to lower state / is given by the relation
tude) are carefully controlled and defined. By contrast, in
natural auroras the dose conditions must be inferred from I.e = N. A.e, s (12.9)
the observations and may be less certain because the effects
are integrated over a range of altitudes and dose levels. where I.e is the optical intensity in photons cm-3 s-’, N. is
Consequently, artificial aurora] experiments provide a useful the population of the upper state in molecules cm-3, and
adjunct to observations of naturally occurring auroras in the A.t is the Einstein A coefficient, or radiative transition prob-
attempt to understtind auroral phenomena. ability, in s-’. If auroral activity has continued long enough
for the production and loss processes of a given state to
come to equilibrium, the number of excited molecules per
cm3 at steady state, Nti,,, is given by
12.2.3 Emission Rates
NUss = PUIL, (12.10)
The kinetic energy of the auroral particles is converted
to optical and infrared emissions of discrete wavelengths where P. is the production rate per unit volume in cm-3s-’
through processes initiated by inelastic collisions with am- and ~ the loss rate in s-’. The loss rate of a given state
bient atmospheric species. These species may spontaneously may be described by the relation
relax to lower energy states directly while emitting radiation.
Alternatively, they may undergo one or more chemical re-
L = ~ A“e + K“o, [021 + K.N, [N,]
actions, finally resulting in an excited product which also e
emits energy by spontaneous radiative decay. The specific + K“O IO] + . . . , (12.11)
auroral processes involved in the production of various ex-
cited species continue to be the focus of much aeronomic where KUMis the rate coefficient in cm3s-’ molecule-’ for
study. Excellent general reviews of auroral processes are collisional deactivation (or quenching) of the excited state
given by Chamberlain [1961], Vallance Jones [1974], and by species M.
Bates [1982]. The reader is also referred to the textbooks The production of a given excited state maybe the result
on atmospheric radiation by Goody [1964] and on infrared of electron impact in inelastic collisions of the form
and optical spectroscopy by Herzberg [1950, 1945, 1967]
and Huber and Herzberg [1979]. e+ N~N*+e, (12.12)

AURORAL RADIATIVE PROCESSES

~N2+ first negotive~ NO+hv (5.3+)

ho ton
N2+ Meine I
emission
02+ first negative

~:!-
‘2+ ‘N: ‘ charge exchonge ‘+’02+’ dissociative
~ 02+ ,02+’, 0+, Nz+,
and reorrangemen wombinot ion (3P)
$
N+ ,OT (with N0,02 ,N2,.. ) ‘+, ~~ (with electrons) (lD )–-----()(sP)+hv (6300,63641)
*3
c .>8
p
~#*a 0( ’S)---- –-O(’O)+hv(5577~)
*
dis~tiotive ionization
At m~heri

eP
ourciol
primary
electrons N2 Lymon-B!rge-Hop field

N2 secoti Pasltlve

N2 first posihve

N2 Vegord-Kaplon

o~ atmosmer!c
~ 02 [“f,~,~d ~tm~spher,c
\

Figure 12-29. Schematic diagram of auroral radiative

b~
processes, showing sources of most important infrared emissions at 2.7, 4.3, and 5.3 ~m.

12-22
 a
THE AtJRORA
the result of energy transfer figure is schematic and indicates only examples of selected
auroral production and loss mechanisms. All primary elec-
M*+ N~M+ N*, (12.13) trons with energies in excess of 100 eV partition their kinetic
energy in the atmosphere in approximately the same branch-
or the result of a chemiluminescent reaction ing ratios. That is, regardless of the primaty electron energy,
approximately the same fraction of the initial kinetic energy
A+ BC-AB*+C. (12.14) is spent in forming various ions (N;, N;*, ~+, N+*,
oi, oj*, o+, . . .) or various neutral excited states (N;,
Here the asterisk indicates an excited state of the atom or 0;, Nj, COj, NO*, . ..). Here and in the remainder of this
molecule. In the case of energy transfer, production of the chapter, the asterisk is used to indicate electronically excited
excited state N* has resulted from the collisional deacti- species, while the symbol $ indicates vibrational excitation.
vation or quenching of the excited M* to some lower energy Excited-state ions produced when the primaries ionize
state. For the chemiluminescent reaction, some fraction of atmospheric neutrals may relax by spontaneous decay to a
the exothermicity of the reaction has been converted to
internal energy of the product molecule. A comprehensive
review of auroral production and loss processes is given in ENERGY LEVELS AND ELECTRONIC BAND
the monograph by Vallance Jones [1974]. Production and SYSTEMS FOR Q AND 0;
loss processes for’metastable states were recently reviewed
by Torr and Torr [1982].
b4 X;

12.2.4 Aurora] Processes ~Z7u Ist NEGATIVE

A flow diagram indicating some physical and chemical a41ru

processes leading to selected ultraviolet, visible, and in- T ~!

‘1
frared auroral emissions is presented in Figure 12-29. This 2nd NEGATIVE 15

ENERGY LEVELS AND ELECTRONIC BAND SYSTEMS

FOR N2 AND N;
02z;

--L
1
Ist NEGATIVE x2Hg
A2rru N;
MEINEL
~G Xzxi
15 15
10

N2
D’z; 4th POSITIVE
b’~u
C’37TU
ESZ;
GOL;:;:;NN-
BIRGE-
HOPFIELD C37TU
HERMAN- 2nd - 10
KAPLAN POSITIVE
—B3Zi
a’ *g B“xi
B3rg :HUMANN -
I AFT E~RGLOW
—W3AU JNGE 5
Ist POSITIVE
I — A3 Xt
, LYMAN -BIRGE -
F@ A32;
/ HOPFIELD
/“
/ 5 HERZBERG
i
/ vEGARO -
tI / KAPLAN

1- 11 /“ 4
,’

H
I
/
/X’xi
— —. —— —— —— -.— —— —— ~ 0
—
o

Figure 12-30. Electronic energy levels and principal electronic band sys- Figure 12-31. Electronic energy levels and principal electronic band sys-
tem of Nz and Nz + [adapted from Vallance Jones, 1974]. tems of 02 and 02+ [Vallance Jones, 1974]. Forbidden
Forbidden transitions are indicated by dashed lines. transitions are indicated by dashed lines.

12-23
CHAPTER i2
lower state emitting a quantum of energy or photon in the secondary electrons, which themselves result from ion-pair
process. The NJ first negative (B2XJ – X*ZJ ), Nl Mei- production by the more energetic prim~ auroral electrons.
nel (A2HU – X*ZJ ), and O; first negative (b4Z; – a4H.) These excited neutral species, in turn, radiate, giving rise
systems (Figures 12-30 and 12-31) are examples of elec- to aurorally enhanced emissions. Some of the affected band
tronic transitions of molecular ions producing significant systems of N2 and Oz are shown in Figures 12-30 and 12-
levels of auroral emissions. The secondary electrons re- 31.
sulting from ionization of \he ambient atmosphere account The fraction of the primary electron epergy radiated in
for a large fraction of the primary electron’s energy as it is a given optical transition as an energetic electron and its
slowed and stopped in the atmosphere by Successiveionizing secondary electrons are stopped in the atmosphere is defined
collisions. The secondary electrons lose most of their energy as the electron-induced radiant efficiency for that transition.
in inelastic collisions which result in excited states of neutral The electron-induced radiant efficiency is independent of
atmospheric species, N**, O**, and Nz*. The Nz (a’~.) electron energy for energies greater than approximately 1W
state, giving rise to the Lyman,-Birge-Hopfield system, and eV. Radiant efficiencies for selected auroral emissions ti
vibrationally excited Nz$“are representative of the excited- presented in Table 12-3, which indicates that 21 to 34% of
state neutral species produced by secondary electrons, as the electron energy radiates in these auroral emissions, with
illustrated in Figures 12-29 and 12-30. Thus, the neutral the infrared emissions of C02 and NO accounting for large
excited species are produced by impact with neutrals of fractions of the incident electron energy. Table 12-3 is based

Table 12-3. Atiroral lBC III intensities and electin-induced radiant efficiencies.

Electron-induced
Z4mith radiant efficiency
Emission brightness= (x lm2)

N2 first positive systemb m 1.6

N2 second positive system’ 90 0.5
N2 Lyman-Birge-Hopfield system 380 4.4
N2 Vegard-Kaplan system 55 0.3
N2 + first negative system 150 0.7
Nz+ first negative (O-O)band 98 o.45d
Nz+ Meinel system’ 770 1.3
Oz atmospheric system 1300 3.0
0.2 infrared atmospheric system 2500 3.5
OJ first negative system 26 0.1
NO y band 60 0.5
NI 8680 ~ (4P–4D0) 10.5 0.022
NII 2144 ~ (3P–5S0) 32f 0.3
017774 ~ (5S”–5P) 9.6 0.022
01.8446 ~ (3S0–3P) 11.5 0.024
015577 ~ (lD–lS) lood 0.32
016300, 6364 ~ (3P–lD)g 2–loo 0.0056-0.28
01 1302–1306 ~ (23P–33S0) 30 0.4
NO fifit overtone 2.7 pmh 0.5–1 .4 x 103 0.4-1.0
C02 (001~) 4.3 pmi’J 3–30 x 103 1–12
NO fundamental 5.3 pmk 6-12 x 103 24
Total 20.8–34.7
‘Intensities are from Valhmce Jones [1974], unless otherwise indlca(ed.
bVallance Jones tid Gattinger [1976].
cVallance Jones aisd Gattinger [1975].
‘Reference value.
CGattinger and Vallance Jones [1981].
‘Sharp [ 1974].
slntensity and effici~nc y @ strongly altitude (or primary electron energy) dependent.
‘R.J. Huppi and Stair [1979].
‘Values are typicat for rocketbome measurements of this strongly self-absorbed band. The primary production pmess, vibrational energy transfer from
N*$, is a slow process, ~d,steady-state radi~ce efficiencies are difficult to establish. Scc Stair et al. [1975].
jSee also Gordiets et aJ. [1978].
‘Inferred from Caledonia and Kennealy [1982].

12-24
 a
THE AURORA
principally on the auroral multiplet and band system in- tion rules [Herzberg, 1950], the most important of which
tensities presented by Vallance Jones [1974] for an IBC III are the electric-dipole selection rules. Transitions obeying
aurora and assumes an electron-induced radiant efficiency electric dipole selection rules have high probability for decay
of 4.5 X 1~3 for the N; first negative (O-O)band at 3914 (large Au?values) and are designated as allowed or permitted
A. transitions. The Nj first negative, N; ~einel (Figure 12-
The band system intensities of Table 12-3 are given in 30), and O: first negative (Figure 12-31) transitions are
Rayleighs (R). The unit, proposed by Hunten et al. [1956], examples of allowed electronic transitions. Transitions vi-
is an equivalent column radiance of 106 photons cm-2 s-’ olating the selection rules have small transition probabilities
and was named in honor of the fourth brd Rayleigh, who and are denoted as forbidden transitions. Excited states from
made the initial measurement of the absolute intensity of which only forbidden transitions originate are metastable.
the night airglow. Standards of auroral optical intensity were Forbidden transitions in atomic oxygen, atomic nitrogen,
proposed by Seaton [1954] and Hunten [1955] based on the molecular oxygen, and molecular nitrogen are important
brightness of the most intense visible feature of the aurora, auroral transitions (Figure 12-29), and their intensities and
the 0(’S – ‘D) 5577 ~ auroral green line (Figure 12-32). altitude profiles have been used as indices of both auroral
brightness and auroral color type.
ENERGY-LEVEL DIAGRAM OF THE TERMS The color of an aurora varies markedly and has been
used to classify auroras into color classes according to the
OF THE 01 2s 22P4 CONFIGURATION
International Auroral Atlas (IAA), published by the Inter-
national Union of Geodesy and Geophysics in 1963. Since
energetic electrons penetrate to lower altitudes, the color
variations are primarily due to differing mean energies and

I
~
I I
‘s ‘0”8s”) energy distributions for the primary electrons incident on
the atmosphere. Metastable species with small Einstein tran-
sition probabilities are vulnerable to collisional deactivation
at relatively low gas densities (high altitudes). The 0( ’D)
state (Figure 12-32) with an A value of 6.8 X l@3 S-l is
quenchd in collisions with N2 at altitudes less than 300
km. Thus, at altitudes less than 300 km, the 6300 ~ auroral
I occll~
011*
ii I
I
I
red line emissions from this state to 3PZand 3P] levels are
diminished relative to other visible emissions. As a con-
oll~ I sequence, ground-based measurements of the 4278 ~ N2+
t *11#
@ I
tirst negative (O– 1) band intensity 1(4278)and of the atomic
oxygen red line intensity 1(6300) can be used to determine
the total auroral electron flux and characteristic electron
Figure ‘12-32. Lowest-lying energy levels and forbidden transitions in energy, or equivalently the altitude profile of energy dep-
atomic oxygen. The radiative lifetimes of the ‘S and ‘D osition [Rees and Luckey, 1974; Shepheti and Esther, 1976].
levels are indicated. While 1(4278) is proportional to the total energy deposited,
1(6300)/1(4278) is sensitive to the altitude profile of energy
The result was the establishment of a logarithmic scale, the deposition and decreases in value for lower altitude auroras
International Brightness Coefficient (IBC), defined in Table (more energetic primary electrons).
12-4. In Figure 12-19 the IBC is related to other auroral Similarly, the 0(’ S) state, which is the source of the
parameters, such as the electron energy flux, peak E region green line (Figure 12-32), has an A coefficient of 1.25 s-’
electron density at steady-state, and optical radiance at 3914 and is quenched by Oz at altitudes less than 100 km. The
A. collisional destruction of this species contributes to the red
The probability of radiative decay, or Einstein transition lower border of the type-B red aurora. TheNZ(A32. +) state
probability Au~, is governed by quantum-mechanical selec- (Figure 12-30) is the major source of 0(’S) in the aurora
through the energy-transfer process (Figure 12-29)

table 12-4. International Brightness Coefficient (IBC). N2(A3 X:) + 0(3P) - 0(’S) + Nz, (12.15)

IBC Class Brightness (kR), 0( ’S-’D) 5577 ~ as reported by O’Neil et al. [1979] and Piper [1982]. Ein-
stein A coefficients for molecular transitions of auroral in-
I 1
!() terest, in some cases summed over all possible lower states,
II
are given in Table 12-5, while A coefficients for atomic
III 100
oxygen transitions are the reciprocals of the lifetimes listed
IV 1000
in Figure 12-32. Ultraviolet and visible auroral spectra re-

12-25
CHAPTER 12
.
Tahl,= 1?.<
..”,- , - #
Ein.f.=in
. . . . . . . . . .
,--ffi,-ientc
. . . . . . . . . . . .
fnr
---
211mml
-------
molecular
---------—
transitions.
-— . . . . . . . ”

System v’ ; AV,VH(S-’)

N2 first positive (B3H.-A3ZJ ) o ~ 1.12 (+5)b

(2.53-0.478 pm) 1 1.29(+5)
2 1.43(+5)
3 1.54(+5)
4 l.a (+5)
5 1.73(+5)
6 1.81 (+5)
7 1.98 (+5Y
8 2.12(+5Y
9 2.27 (+5Y
10 2.31 (+5)’
11 2.39 (+5)’
12 2.43 (+5)’

Nz second positive (C311u-B3Hg) o 2.73 (+7)

(0.546-0.268 pm) 1 2.75 (+7)
2 2.73 (+7)
3 2.67 (+7)
4 2.50 (+7)

N, Wu-Benesch(W3A.-B3H,) o 6.00 (–2)

(2.24.3 pm) 1 5.00 (+2)
2 2.04 (+3)
3 4.28 (+3)
4 7.26 (+3)
5 1.08(+4)
6 1.45 (+4)
7 1.84(+4)

Nz Vegard-Kaplan (A3ZJ -X’ZJ ) o 5.30(–1)

(0.532-1 .25 pm)

Nz Lyman-Birge-Hopfield (a’~g-X’~~) d 6.7 (+3)

(0.260-0. 100 pm)
N2 Herman-Kaplan (E3Z~ -A3ZJ ) e 3.0 (+3)
(0.274-0.213 pm)
NJ first negative (B2ZJ -X2Z~ ) o 1.41 (+7)
(0.5874.286 pm) 1 1.38(+7)
2 1.4 (+7)

NJ Meinel (A2~.-X2Z~ o 6.04 (+4)

(O. 177-0.550 pm) 1 7.20 (+4)
2 8.30 (+4)
3 9.34 (+4)
4 1.03(+5)
5 1.12(+5)
6 1.21 (+5)
7 1.30(+5)
8 1.37(+5)
9 1.45 (+5)
10 1.52(+5)

12-26
 a
THE AURORA

Band origin
System v ‘–v” Avv,, (S-’) (pm)

02 atmospheric (b’x~ -X3Z; )f o-o 8.5 (–2) 0.7619

1-0 6.9 ( – 3) 0.6882
2-o (1.636 * 0.040) (–4) 0.6287
1–1 (7.04 * 0.42) ( -2) 0.7708
0-1 4.16 (–3)g 0.8645
02 IR atmospheric (a’Ag-X3X~-)f o-o 2.58 (–4)g 1.269
0-1 1.75 (–6)13 1.580
0, Noxon (blZJ-alAg)f o-o 1.5 (–3) 1.908
‘All data are from Loftus and Krupenie [1977] unless otherwise indicated
%e notation “a(n)” means “ax 10””.
cEvler and Pirrkin ~19831.
‘C~rrected fo; a = a’ cascade [Freund, 1972].
‘Using relative transition probabilities for E + A, E + C, and E + B from Freund [1969] and lifetime of E state from Borst and Zipf [1971].
‘Kru@nie [1972].
gCalculated from upper-state lifetime and Franck-Condon factor.

suiting from the EXCEDE artificial auroral experiment are 12.2.5 Infrared Aurora] Emissions
presented in Figures 12-33a and b. Natural auroral spectra
in the ultraviolet and visible are shown in Figures 12-34 The development of infrared cryogenic instrumentation
through 12-37. and sensors has facilitated the measurement of infrared au-
Caution is urged in applying artificial auroral spectra to roral emissions in recent years [Stair et al., 1983]. Infrared
situations involving natural aurora. The former are biased artificial auroral spectra from EXCEDE are shown in Figures
toward prompt emitters, and delayed emitters such as the 12-38a and b, while natural airglow and auroral emissions
COZ4.3 pm band may not be measured efficiently. This is in the 0.88 to 5.4 pm range are illustrated in Figures
true of ground-based spectra such as Figure 12-33b due to
continual motion of the electron gun through the atmosphere
into unclosed regions. However, the remark applies espe-
cially to rocketbome spectra, where the spectrometer can
only observe a dosed region for a short while.

EXCEOE SPECTRAL UV SPECTROMETER

w J: wAVELENGTH (4)

+
i!’
,1
,1 I

wAVELENGTH [i)

WMLENGTH (f)

Figure 12-33b Visible spectrum taken from the ground with an image-
Figure 12-33a. Ultraviolet spectrum taken during EXCEDE SPE~AL intensified spectrograph during the PRECEDE artificial
artificial aurora] experiment at 125 km altitude. Principal aurora] experiment with payload at 92 km. Principal band
band systems seen are the Lyman-Birge-Hopfield (LBH), systems seen are N*+ first negative (IN) and Meinel (M)
Herman-Kaplan (HK), Vegard-Kaplan (VK), and second systems and the Nt first positive (I P) system [0’Neil et
positive (2P) systems of N, [0’Neil et al., 1982]. al., 1978a].

12-27
CHAPTER 12

DAYGLOW

1’ 1 1r
l“’’”

AURORA
160-180 km

WAVELENGTH (~)

1300 1400 I 500

Figure 12-34a, Extreme ultraviolet side-looking s~ctrum(525-1250~)
WAVELENGTH (a)
of a natural aurora taken at lW180 km altitude (bottom),
with a comparable measured dayglow spectrum (top).
Principal band system is the Nz Birge-Hopfield (BH) sys- Figure 12-34b. Far ultraviolet auroral spectrum (1 I 75–1520 ~) under the
tem. Prominent N I and N II lines and off-scale O I, O same conditions as Figure 12-34a. Dominant features are
II, and Ly-a lines are also present [Feldman and Gentieu, the Nz Lyman-Birge-Hoptield (LBH) band system and
1982]. atomic emissions from N I (1200 ~), Ly a (1216 ~), and
O I (1304 and 1356 ~) [Feldman and Gentieu, 1982].

12-39, 12-40, and 12-41. Spectra of the near-infrared por-

tion of this range are also shown by D. Baker et al. [1977].
Emissions in this wavelength range correspond to energies 1.0
varying from 1.4 eV at 0.88 pm to 0.23 eV at 5.3 pm. ,,1-
The energy levels involved in these less energetic transitions
producing infrared emission are both electronic states of Nz r~ [01]

k../
and 02 (Figures 12-30 and 12-31) and vibrational transitions
within the ground electronic state of minor atmospheric
species. Pure vibrational transitions are forbidden in the
homonuclear diatomic molecules Nz and 02 forming the
major constituents of the atmosphere. Thus infrared-active
minor atmospheric species, such as the diatomics OH, CO,
I I 1 1 1 1 I 1 1 1 I I 1 I I 1 1
NO, and NO+ and the triatomic species COZ, NZO, NOZ, m m
1
2400 2W 1600

and 03 become the principal sources of infrared airglow WAVELENGTH (i)

[Chapter 13] and auroral emissions at longer wavelengths.

The potential energy curve and fundamental and first-over- Figure 12-35. Far ultraviolet aurora] spectrum from a nadir-lwking sat-
tone vibrational transitions of NO are shown in Figure ellite (1700-3200 ~). Principal band systems are the NZ
Lyman-Birge-Hopfield (LBH), Vegard-Kaplan (VK), and
12-42, and the band origins and Einstein coefficients are
second positive (2POS) systems [Sharp and Rees, 1972;
given in Table 12-6. The vibrational energy levels of COZ Sharp, 1978]. The dotted curve is a synthetic spectrum.

12-28
 N;lN
3,11 12,0 3,212)i 10[ 2,21 !,1[ Qoll
i--

3100 3200 3300 3400 3500 3600 3700 3600 3900

WAVELENGTH (i)

Figure 12-36a. Observed ultraviolet ground-based s~ctrum of an aurora from 3100-3920 ~ (top) along with synt~etic spectrum (bottom). The synthetic
spectrum is also shown as a dotted line in the top panel, and various components of Ihe synthetic spectrum are shown in the bottom panel.
SevekaI band systems m shown witti the sarire notation as in Figures 12-33a,b [Vallance Jones and Gattinger, 1975].

3,41 2:31 1.21 0,11

,51 1,41 0,31 4,81 3,d 2,61 I,51 0,41 b,g 3.e
,:~; x
~14; N2VK
,141 1,121 2,131 1,131 5,161 , 16,17 13,15 4,161

OU6 on 2 .On 5 on 15 on I

013
II

L- ~
. .
40100 4100 4200 1300 44m 4500 46’~ 4700

WAVELENGTH [~)

Figure 12-36b. Ground-based visible aumralspc~m from 392W4720 ~. Rttilsm asinfigure 12-36a[Vallance Jones and Gattinger, 1975].

12-29
 I I T

Hg14
01 3F Iig Is’
NR5
HqI 1 I
T I Nlll ‘y’ Mn9
r
Iie I i! Na28
001
I ‘y’‘? N!,3 NIJ3F IT NI127
NCI14 of.~2
N19 ~ %E29 ND 3
r“
I 7 N~9
N1 IF
1 -1
\ T0,40 ‘“~’;f’ o’lJ2rf10*l’o
ND20 ti~oo280114° ‘ IL-
on15 01!2 01,1 I :) I
-~ T
II h
. A ~ -s on - ---

19,2
OH
17.1 18.2 5.01 1

I 3,0 ‘,,, ,2,0 ‘:]NG 2,11 11,0 3,3 I [2,211,1 10,0

-u ~ ..:. .

D 5000 - 5500 6000

WAVELENGTH (h)

Figure 12-37a. Ground-based visible auroral spectrum from 4500- 62W~. Details areas in Figure 12-36a. Additional band systems shown arethe Oz’
first negative (lN)and the N,infrared afterglow (IRA) [Gattinger and Vallance Jones, 1974].

k 6000
t

WAVELENGTH (i)
7050

Figure 12-37b. Same as Figure 12-37a, but for spectral range 58W 7400 ~. One additional band system, the O* atmospheric (At), is present

12-30
 THE AURORA

014 ,
011

N13

i
v

o
WAVELENGTH (A)

Figure 12-37c. Same as 12-37a, but for the spectral range 74W9000 ~.

I04

N2 (W3AU–
B311a)
EXCEOE SPECTRAL LHE CVF

L
m-l No
AV=I
—
z
ALTITUDE 74 k m
ACCELER~TION ON, 20 KW
I

NOAV=I

I01
II
2 3 4 5
WAVELENGTH (PM )
I I I
3 4 5 6 7

Figure 12-38a. Short-wave infrared spectrum taken of electron-induced WAVELENGTH (PM)

emission with liquid-nitrogen-cooled spectrometer during
EXCEDE artificial aurora] experiment at altitude of 74
km. Dominant emissions are the Nz Wu-Benesch band Figure 12-38b. Electron-induced infrared sfictra taken with liquid-he-
system, as well as NO and CO* vibrational bands [0’Neil lium-cooled spectrometer at 74 km during EXCEDE ar-
et al., 1982]. tificial auroral experiment [G’Neilet al., 1982].

12-31
 WAVELENGTH (i)

Fieure
-. 12-39. Observed ~round-based auroral near-infrared spectrum from 88W11400~ (top), along with synthetic spectrurir (bottom). Wtails are as
shown in~lgure 12-36a [Vallance Jones and Gattihger, 1976].

areschematically represented in Figure 12-43, andthe band pm and 5.3 pm emission from NO and the 4.3 pm emission
ongins and transition probabilities for CO, C02, NZO, NOZ, from COZ.
H20,0~, and NO+ arepresented in Table 12-7. Nitric oxide is an important minor constituent in the
Not all the above atmospheric infrared radiators give thermosphere. It is infrared-active, radiating near 5.3 pm
rise to emissions that are aurorally enhanced. Modest two- in the fundamental Av = 1 sequence and near 2.7 pm in
fold enhancements of the 1.27 pm emissiorr from the Oz the first overtone (Av = 2) (Figure 12-41). me 5.3 pm
(0-O) infrared atmospheric band (a’A, – X’Z;) (Figure fundarnentil band emission is an important cooling mech-
12-3 1) have been reported in a very strong IBC III+ aurora, anism for the thermosphere [Chapter 13; Kockrtrts, 1980].
and the enhancement has been explained through direct elec- The v = 1 state is predominantly produced, in the nighttime
tron excitation of 02 [D.J. Baker et al., 1978]. Due to the quiescent atmosphere, by collision with atomic oxygen and
long lifetime (3900s) of the 02(a’ A~) state, horizorital winds by absorption of earthshine [Chapter 13],
cart trartsprt the excited species downwind by ~ved hundred
kilometers before it decays [Rees and Roble, 1980], serving NO(v=O) +O@NO(v=l)+O (12.16)
to complicate the analysis. The Wu-Benesch bands of N2
(W3AU – B3~~) (Figitre 12-30) are strong features of the NO (v = O) + hv (earthshine) ~ NO (v = 1). (12.17)
EXCEDE artificial auroral spectra (Figure 12-38a) in the
24 pm region, but have ;ot been observed in natu~al au- These processes primarily produce the NO (v = 1) level
roras, probably due to lack of sensor senditivjtj. Recent with insignificant population of levels v > 2.
observations [Stair et al., 1983] during the HIRIS experi- Vibrationally excited NO is also produced by chemical
ment indicate enhancements of the 15 vfi COZ VZband reactions initiated by precipitating auroral electrons. Lab-
eniission as well as the 9.6 pm 03 *3 band emission in a oratory studies have shown that the major chemiluminescent
very strong (IBC 111+) aurora. However, no mechanism for auroraI process that produces NO (Figure 12-29),
these ,enhancemerits is known at the present time. Conse-
quently, these observations await elaboration through fur- N(2D) + Oz ~ NO(v) + O, (12.18)
ther experiments and data analysis. The remainder of this
section deals with the most firmly established and best populates NO levels up to v = 12 [Kennealy et al., 1978].
understood infrared auroral enhancements, namely the 2.7 The N(2D) is produced directly from N2 by precipitating

12-32
 THEAURORA

WAV~&E’JGTH-A POTENTIAL ENERGY CURVE AND VIBRATIONAL

1mm 10500 I! 500 !29GG 125W

100
ENERGY LEVELS FOR NO

50 N(4S0)+O(3P)

I
0,(
>
~N21p 5,6] [0,0 4,5 2,31 !.2
g4
—N; M
0,0
I 1
,,, 2,2, 3.3
‘1 w
50- ,,
z
w
,!{:!
,.::,. ,

i’::;:
~’~
t
[;,
:{!!
*,4 ‘1
L.
<2
..:,..,{,.’ ,’ : . ~
1-
Z
07-
1Owo
. /1. ..-”:%

9500
~ . . . . .. ..

9000 8500 8000

w x 211r
WAVENUMBER-CM “
+
0
,3500 WA~~~ENG~;;oA ‘O
12500 13030 15000 1550C 16333 165m

100- - -AV= I FUNDAMENTAL VIBRATIONAL TRANSITIONS

--AV=2 OVERTONE VIBRATIONAL TRANSITIONS
01- N!-Nl-
-2
50-
!
I i I I I I
( } 0.8 1.2 1.6 2.0 2.4 2.8
INTERNUCLEAR DISTANCE (i)
0 -
7500 7C.XI 6500

Figure 12-42. Potential energy curve for lowest electronic state of nitric
oxide, showing vibrational energy levels as well as fun-
50 - 4“4’
damental and first overtone vibrational transitions [adapted
from Gilmorc, 1%5].

0,
8000 7500 7000 65W 6~ electrons and by dissociative recombination and charge-
WAVENUMBER-CM-’
exchange reactions of aurorally produced ionic species,

Figure 12-40. Infrared spectrum from 1.0 to 1.65 pm taken from the
ground (top panels) along with corresponding synthetic N2 + e + ze+ f1N(2D) + (1 – f,) N(4S) + N+
spectra (bottom panels) [Gattinger and Valiance Jones,
1981]. N2 + e ~ e + 2f2N(2D) + 2(1 – f2)N(4S)

N; + e ~ 2f~N(2D) + 2(1 – f3)N(4S)

NO+ + e ~ 0.76 N(2D) + 0.24 N(4S) + O

N; + o + fd N(2D) + (1 – f~) N(4S) + NO+

SWIR/MWIRAURORAL EMISSION SPECTRUM N+ + 02- fs N(2D) + (1 – fs) N(4S) + OJ .

= 100, I T I I [ T I
~ (12.19)
VERTICAL VIEW
a ROCKETALTITUOE:86 km C02 (u, )
~ 75 -
AuRORA:IBC Ill The branching ratios fi are of crucial importance, but they
IAl ) are uncertain for all reactions except NO+ recombination
v
z
a NO (f,4 and f~ are probably near unity). Nevertheless, the pro-
SW - {AV=I)
a
portion of N(2D) that reacts with 02 to produce NO(v) should
-1 vary with altitude as the reaction of Equation (12. 18) com-
NO
~ 25- petes with quenching of N(2D) by atomic oxygen. The vari-
1- (AV=2)
v
u (::=2) ation of the [0]/[02] ratio with altitude above 100 km then
$ 1 leads to an altitude-dependent efficiency for auroral NO(v)
1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5
production. The altitude dependence of this production ef-
WAVELENGTH (pm )
ficiency can then account for the variability of 5.3 pm
auroral emission with auroral strength since auroras pro-
Figure 12-41. Zenith spectrum of natural aurora in the 1.5 to 5.4 #m
spectral range taken from 86 km showing principal emitters
duced by more energetic electrons have lower peak ioni-
[Stair et al.; 1975]. zation altitudes.

12-33
CHAPTER 12
Table 12-6. Band origins and thermally-averaged Einstein coefficients for The situation is somewhat complicated by the large vari-
the fundamental and first overtone vibrational transitions of ability of NO and by its transport. Since the rates of reaction
nitric oxide,
of NO [Equations (12. 16) and (12. 17)] are proportional to
Av=l Av=2 the NO density, the nighttime quiescent component of the
fundamental band should be stronger at high latitudes, es-
Av ‘ v“a BAND A“,jv,ra BAND pecially in aurorally predosed regions where the NO density
ORIGINb ORIGINb is expected to be large [Swider and Narcisi, 1977]. Nitric
v’ (s-’) (cm-’) (pm) (s-’) (cm-’) (pm) oxide is a long-lived species in the thermosphere, and model
calculations have pointed out the importance of therrno-
1 10.8 1876 5.33 — — — spheric winds in redistributing aurorally produced NO [Ro-
2 20.4 1848 5.41 0.460 3724 2.68 ble and Gary, 1979].
3 29.1 1820 5.50 1.51 3668 2.73 The NO 5.3 ~m band radiance and the spectral shape
4 36.5 1792 5.58 3.10 3612 2.77 of the emission depend on whether the NO(v) is produced
5 42.9 1764 5.67 4.90 3556 2.81 by the quiescent mechanisms [Equations (12.16) and (12. 17)]
6 48.5 1736 5.76 7.28 3500 2.86 or the auroral mechanism [Equation (12. 18)]. As alluded to
7 53.5 1708 5.86 9.63 3444 2.90 previously, the quiescent processes can only excite the 1-O
8 57.5 1680 5.95 12.5 3388 2.95 transition, while the auroral process should yield in addition
9 60.6 1652 6.05 15.7 3332 3.00 appreciable emission in the hot bands 2-1, 3-2, . . . ,
10 62.8 1624 6.16 19.1 3275 3.05 12-11. Evidence of these components of the NO Av = 1
11 64.2 1596 6.27 22.9 3219 3.11 emission in an aurora of only moderate intensity (IBC II)
12 64.6 1567 6.38 27.2 3163 3.16 is clearly seen in the spectra resulting from the launch of a
13 64.2 1539 6.50 31.9 3107 3.22 high-resolution (1.2 cm-l) rocketbome field-widened
14 63.2 1511 6.62 36.9 3050 3.28 interferometer (FWI) from Poker Flat, Alaska, in March
15 61.7 1483 6.74 41.9 2994 3.34 1983 [Steed et al., 1983; Murphy et al., 1983]. Preliminary
16 59.8 1455 6.87 46.4 2938 3.40 analysis of the FWI flight data indicates that the hot band
17 57.6 1427 7.01 50.5 2882 3.47 component increases with auroral energy deposition in a
18 55.0 1399 7.15 54.3 2825 3.54 manner consistent with the mechanism of Equation (12. 18).
19 51.9 1370 7.30 58.1 2769 3.61 In contrast to the 5.3 pm NO fundamental, the overtone
20 48.5 1342 7.45 61.8 2713 3.69 band near 2.7 pm arises only when levels v >2 are excited.
m.,,. , ,. A-,,
‘rslllurgsley[lylol
Thermal collisions and earthshine do not populate these
bCalculated from vibrational constants of Goldman and Schmidt [1975]. levels significantly, and thus the 2.7 pm emission is neg-
ligible during quiet nighttime conditions. However, the
Av = 2 sequence is observed even for moderate IBC 11
auroras [Caledonia and Kennealy, 1982], the emission oc-
curring with an efficiency of 0.4-1.090 [R.J. Huppi and
Stair, 1979; Reidy et al., 1982].
The altitude profiles of zenith spectral radiance measured
4000 during several earlier auroral rocket experiments in the ICE-
CAP series involving lower-resolution spectrometers than
the FWI, are shown in Figure 12-44 [Stair et al., 1975].
The 5.3 pm radiance is seen to be enhanced for the IBC
III + aurora, but essentially at quiescent levels for the weaker
IBC 11 and 11+ auroras. On the other hand, the 2.7 pm
. radiance is enhanced for both the IBC II and 111+auroras
(Ilo)nu
recorded.
(030)lru The failure to observe any enhancement of NO 5.3 pm
radiance in moderate IBC II-II + auroras during the ICECAP
(020)x~
series of flights contrasts sharply with the enhancement of
E
the Av = 1 hot band component seen in the FWI experiment
N= 1-
b during a moderate aurora. The lack of enhancement in the
ICECAP observations is due to the existence of two con-
:
:
ditions: (1) The ICECAP detector’s long-wavelength cutoff,
g
(Ooo)xj shown at 5.4 ~m in the data of Figure 12-41, passes all of
Y
the Q and R branches of the NO (1-O) band, but misses
Figure 12-43. Lowest vibrational energy levels of COZ molecule showing more than half of the hot-band emission. (2) Quiescent-
some transitions of aeronomic interest. atmosphere emission dominates auroral emission in the re-

12-34
 m
THE AURORA
Table 12-7. Data on infrared bands of atmospheric interest.’

Rotational Band Origin Band

constant strength S Einstein
B Wave number V. Wavelength AO (cm-2atm-’ coefficient A
Species (cm-’) Transition (cm-’) (pm) at 273 K) (s-’) Reference

co* 0.390 1o”o-oooo 1388 7.20 Inactive Inactive c

01‘O-oo”o 667.4 15.0 205 1.28 c
00°1-0000 2349 4.26 2380 368 c
10’J1-00°0 3715 2.69 37.2 14.4 c
20°1-0000 4978 2.01 0.868 0.603 c
00°1-1000 961.0 10.4 o.o170b 0.411 c

HZO d 1OO-OOO 3657 2.73 12.0 4.52 e

01 O-OOO 1595 6.27 257 18.30 f
Oo1-ooo 3756 2.66 172 67.90 e
01 la 5331 1.88 19.9 15.90 e

N20 0.419 00°1-0000 2224 4.50 1420 196 g

0100-00”0 588.8 17.0 24.4 0.119 e
1ooo-oo”o 1285 7.78 218.0 10.1 e

d 1OO-OOO 1103 9.07 16.6 0.568 f

01 O-OOO 700.9 14.3 15.6 0.214 f
Oo1-ooo 1042 9.60 345 10.5 f

N02 d loCf-oOO 1320 7.58 — — h

O1O-OOO 749.6 13.3 13.1 0.207 i
Oo1-ooo 1617 6.18 1520 111 i
101-OOO 2906 3.44 64.0 15.2 i

co 1.931 1-0 2143 4.67 243 31.3 e

2-o 4260 2.35 1.86 1.12 e

NO ‘ 1.997 1-0 2344 4.27 70.8 10.9

2-o 4656 2.15 1.15 0.697
2-1 2312 4.32 135b 20.2

NO 1.705 1-0 1876 5.33 109 10.8 k

2-o 3724 2.68 1.18 0.460 k

OH 18.871 1-0 3570 2.80 24.0 8.58 i

2-o 6974 1.43 8.98 12.2 i

‘Table adapted and updated from Kennealy and Del Greco [1972].
bBand strength is strongly temperature-dependent, since the lower level is not the ground state.
CRothman and Young [198 I ].
‘Molecule is an asymmetric rotor with three distinct rotational constants.
‘Rothman [1981].
‘Rothman et al. [ 1983b].
gMcClatchey et al. [1973].
‘Herzberg [1945].
‘Rothman et al. [1983a].
JWemer and Rosmus [1982].
‘Billingsley [1976].

12-35
CHAPTER 12
SPECTRAL R~ANCE (Wth cm-2 ar-’~m-’ x 10 “e)

— IBC IU +(24 MAR. 1973 ~UTE-TOMAHAWK)A

-- lBC ~(~ I E~~R)~ f=E~~74~CK ~ANl

‘~

80

70 m

60

50 i
5
I
10 15
I

SPECTRAL
I
20
RADIANCE
1
25
(*pm
1

??
)
=
I I

“0
1[!30
60

50
o
I I
5
1
w
I
SPECTRAL
I I
Is
I

RADIANCE(MR
I
20
1

pm-l)
I
28
I
m

Figure 12-44. Altitude profiles of zenith peak spectral radiance observed in the NO fundamental 5.3 pm and overtone 2.8 #m bands for auroras of different
brightness [Stair et al., 1975].

gion of the (1-O) band for moderate auroras and masks “any We turn now to the 4.3 ~m radiation from the us asym-
aurora] contribution. On the other hand, the long-wave- metric-stretch mode of C02, which results from the emission
length cutoff for the recent FWI launch was sufficiently high process
to show clearly resolved lines of the NO hot bands on the
low-frequency side of the NO (l-O) band, whose strength CO,(k{ m) @ COz(k,(,m – 1) + hv (12.20)
varied in altitude with the energy deposition [Steed et al.,
1983; Murphy et al., 1983].
and, in particular, from the fundamental band emission
In summary, the intensity in the NO Av = 1 system
for v’ z 2 responds to energy input from moderately strong
auroras in the same way that the NO overtone does, in accord COZ(OO1) @ COZ(OOO) + hv (12.21)
with model predictions. By contrast, for moderately strong
auroras the NO ( 1-O) emission is dominated by airglow from at 4.26 pm. The altitude profile of this emission in the
the background atmosphere, as described by Equations (12.16) zenith is indicated in Figure 12-45, showing results from
and (12.17). four ICECAP rocket launches [Stair et al., 1975], one launched
As a final note on auroral NO(v), we mention that the under quiet conditions and three under various levels of
HIRIS experiment previously referred to found the NO fun- electron precipitation in discrete arcs. The increase of spec-
damental emission to be rather strongly enhanced for the tral radiance between 75 and 90 km and the peak at 92 km
very strong IBC 111+ aurora which it observed. This is are due to the fact that the VJ band becomes optically thick
consistent with models of NO emission [Caledonia and Ken- in the zenith at about 90 km. There is a general correlation
nealy, 1982], which indicate that chemiluminescent contri- between the peak radiance level and the electron flux, or
butions to the NO Av = 1 bands become comparable to the 5577 ~ radiance, except that the IBC II+ aurora is
emission from the background atmosphere under such au- considerably less bright at 4.3 pm than the IBC II aurora.
roral conditions. The explanation for this apparent anomaly is that a very

12-36
 THE AURORA
the delay as a major factor in accounting for enhanced 4.3
pm etnission, namely NO+ (Av = 1) chemiluminescence
(Table 12-7). The NO+ 4.3 pm emission results from vi-
brationally excited NO+ produced (Figure 12-29) in reac-
tions such as

02+ + NO+ Oz + NO+ (12.24)

.....-
..I
and the fourth reaction in Equation (12. 19). However, it
~ cannot account for the observed time delay since the time
~ constant for NO+ formation is s 10 s [Kumer, 1975].
4 .
On the other hand, the reaction sequence of Equation
7, (12.23) has the slow step (b). Assuming a nitrogen vibra-
“.\
...
.. .
tional temperature T, = 750 K, one obtains a time constant
““ ------ of 200 s at an altitude of 115 ‘km for the V-V transfer. The
““-------- -.. -.. -.O

effective time constant for response of the V3emission to

0 30 100 1% 200 2s0 electron @ipi@tion is lengthened further by repeated transfer
sPECTRAL
RACs~E IMRpm”8 I of excitation back and fodh between N2 and C02 before
photon emission occuis and radiative trapping of the emitted
Figure 12-45. Altitude profiles of ped spectral radiance observed in the photon. The net effect is an appreciable delay in the 4.3
zenith in the COZ 4.3 pm band for several classes of aumm pm emission. To model the process [Kumer, 1975, 19771,
[Stair et al., 1975].
one must add quenching of both N; and the COZ excited
states to the ~action set of Equations (12.20) and (12.23).
The reverse reactions, indicated by backward arrows in the
bright IBC 111arc had occupied the observation region in
reactions of Equaiions (12.20) and (12. 23b), account for
the latter case before the rocket arrived. repeated collisional trhsfer of the excitation between COZ
There are two possible sources of aurofially enhanced
and N2 and for radiative transfer of the V3photons. Kumer
C02 excited vibrational states: (1) direct ex~itation of CO*
[1977] has shown that up to 11 combination bands and
by precipitating second~ electrons,
isotopic bands must be considered in Equation (12.20) and
that an error of a factor of three can result if only the
COZ (OCIO)+ e ~ COZ (k(m) + e, (12.22)
fundamental 001-000 band is used.

and (~) vibratiod~l pumping of Nz by secondary electrons

followd by vibration-to-vibmtion (W) etirgy transfer from
12.2.6 Sumtiary
N2 to CO* (Figure 12-29),
Our current understanding of infrared auroral processes
N2+ee Nj+e (12.23a)
suffers from a lack of sufficient high-resolution, high-sen-
N; + COZ (k,t,m – 1) - Nz + COZ (kern). sitivity data on infrared emissions. This deficiency should
(12.23b) be partially alleviated as analysis of data from the March
1983 rocketbome ~1 experiment proceeds. In addition,
Due to the low COZ mixing ratio, the ratie for the direct several planned rocket-probe and shuttle-based experiments
process [Equation (12.22)] is too small by several orders of will carry aloft telescoped high-resolution interferometers
magnitude to explain the observed radiance. Moreover, since to view the infrared aurora in earthlimb geometry. The com-
the radiative lifetime of the 001 state is A-’ = 2.3 X 103 bination of improved sensor sensitivity and the prefemed
s (Table 12-7), the 4.3 pm emission due to the direct re- earthlimb viewing aspect with greater auroral brightness will
action of Equation (12.22) should be prompt, following the provide a comprehensive spectral survey of aurora] infrared
energy deposition closely in space and time, However, the emissions. The observation of additional auroral emissions
actual altitude profiles and time histories for the”3914 ~ is anticipated as detection thresholds are improved and as
Nz+ first negative (O-O)radiance and for the 4.3pm pdiance infrared measurements are extended to higher spectral res-
are quite different with the COZ emission having a time olution and longer wavelengths. These experiments should
delay = 5 min [Kumer, 1975]. The delay arising from the contribute significantly to our understanding of the produc-
N2-C02 transfer is responsible for the apparently anomalous tion and loss processes for infrared auroral emitters.
brightness at 4.3 pm of the IBC 11auroral profile in Figure There is also a great need for data on the variability of
12:45. the infrared aurora in order to identify seasonal, diutnal,
Another potential mechanism can also be ruled out by and latitudinal trends, as well as depndences on solar and

12-37
CHAPTER 12
geomagnetic activity. Such a data base may be obtained infrared satellite experiments such as IRAS ( InfraRed As-
from a satellite platform or, in pati, from ground-based tronomical Satellite) [Neugebauer et al., 1984] portend well
observatories and series of aircraft or balloon flights. None for the future of infrared auroral studies.
of these experiments are without difficulty; the satellite ex-
periments require solving the problem of maintaining de-
tectors at cryogenic temperatures for extended @riods, while
ground-based and atmospheric platform expe~ments suffer ACKNOWLEDGEMENTS
from the masking and/or absorption of the auroral emission
by the lower atmosphere and from the impossibility of using The authors are indebted to A.J. Ratkowski, J.R. Win-
any limb enhancement technique. Nevertheless, the exis- ick, and R. D. Sharma for reading Section 12.2 and making
tence of the space shuttle and the success of unmanned useful recommendations.

12-38
 THE AURORA

REFERENCES

Akasofu, S.-I., “Energy Coupling Between the Solar Wind Fluorescent Radiation Emission Caused by an Electron
and the Maghetosptiere,” Space Sci. Rev., 28:121, 1981. Beam,” J. Appl. Phys., 41:3767, 1,970.
Andriyenko, D. A., “Some Vertical Distribution Pattern s,” Dalgamo, A., I.D. Latimer, and J. W. McCo~key, “Cor-
Geomag. and Aeron., 5:838, 1965. puscular Bombardment and N; Radiation,” Planet. Space
Baker, D., W. Pendleton, Jr., A. Steed, R. Huppi, and Sci., 13:1008, 1965.
A,T. Stair, Jr., “Near-Infrared Spectrum of A Aurora,” Eather, R.H. and K.M. Burrows, “Excitation and Ionization
J. Geophys. Res., 82: 1601, 1977. by Auroral Protons,” Aust. J. Phys., 19:309, 1966.
Baker, D. J., W.R. Pendleton, Jr., G.J. Romick, V.B. Eyler, E.E. and F.M. Pipkin, “Lifetime Measurements of
Wickwar, and M.J. Baron, “An Aurora] Enhancement the B 3H~ State of N2 Using Laser Excitation,” J. Chem.
of 01 h 1.27-pm Emission,” Planet. Space Sci., 26: Phys., 79:3654, 1983.
619, 1978. Feldman, P.D. and E.P. Gentieu, “ne Ultraviolet Spec-
Baker, K. D., D.J. Baker, J.C. Ulwick, and A.T. Stair, Jr., tmm of an Aurora 530-1520 ~,” J. Geophys. Res., 87:
“Measurements of 1.5- to 5.3-pm Infrared Enhance- 2453, 1982.
ments Associated with a Bright Auroral Breakup,” J. Feldstein, Y. I., “Peculiarities in the Auroral Distribution and
Geophys. Res., 82:3518, 1977. Magnetic Disturbance in High Latitudes Caused by the
Banks, P. M., C.R. Chappell, and A.F. Nagy, “A New Asy~etrical Form of the Magnetosphere,” Plunet. Space
Model for the Interaction of Auroral Electrons with the Sci., 14; 121, 1966.
Atmosphere: Spectral Degradation, Backscatter, Optical Feldstein, Y.I. and G.V. Starkov, “Dynamics of Auroral
Emission, and Ionization,” ~. Geophys. Res., 79:1459, Belt and Polar Gmmagnetic Disturbances,” Planet. Space
1974. Sci., 15:209, 1967.
Bates, D. R., “Airglow and Auroras,” in, Applied Atomic Frank, L. A., J.D. Craven, J.L. Burch, and J.D. Win-
Collision Physics, 1, Academic Press, New York, 1982. ningham, “PoIw Views of the Earth’s Aurora with Dy-
Berger, M. J., S.M. Seltzer, and K. Maeda, “Energy Dep- namics Explorer,” Geophys. Res. Left., 9:1001, 1982.
osition by Auroral Electrons in the Atmosphere,” J. Arm. Freund, R: S., “Molecular-Beam Measurements of the
Terr. Phys., 32:1015, 1970. Emission Spectrum and Radiative Lifetime of NZ in the
Berkey, F. T., V.M. Driatskiy, K. Henriksen, B. Hultqvist, Metastable E 3Z{ State,” J. Chem. Phys., 50:3734,
D.H. Jelly, T.,1. Shchuka, A. Theander, and J. Yliniemi, 1969.
“A Synoptic Investigation of P@icle Precipitation Dy- Freund, R. S., “Radiative Lifetime of Nz (a lH~) and the
namics for 60 Substorms in IQSY (1-1%5) tid ~SY Formation of Metastable N2 (a’ ‘~u- ),” J. Chem. Phys.,
(1969),” Planet. Space Sci., 22:255, 1974. 56:4344, 1972. “
Besprozvonnaya, A.S. and T.I. Sheh~ka, “Distribution of Garriott, O. K., “Visual Observations from Space,” J. Opt.
Anomalous Ionization in the High Latitude E Region Sot. Am., 69:1064, 1979;
According to Ground-Based Sounding,” Geomag. Aeron., Gassmann, G. J., “Analog Model of the Arctic Ionosphere,”
16:430, 1976. AFCRL-TR-73-0151, AD762280, 1973.
Billingsley, F. P., II, “Calculated Vibration-Rotation In- Gassman, G.J.., J. Buchau, R.A. Wagner, C.P. Pike, md
tensities for N0(X2~) ,“ Y. Mol. Spectrosc., 61: 53, M.G. Hurwitz, “Arctic Ionospheric Modelling-Five
1976. Related Papers,” AFCRL-72-0305, AD775084, 1972.
Borst, W .L. and E.C. Zipf, “Lifetimes of Metastable CO Gattinger, R.L. and A. Vallance Jones, “Quantitative Spec-
and N2 Molecules,” Phys. Rev. A, 3:979, 1971. troscopy of the Aurora. II. The Spctrum of Medium
Boyd, J. S., “Rocket-Bome Measurements of Auro~l Elec- Intensity .Aurora Between 4500 and 8~ ~,” Can. J.
trons,” Rev. Geophys. Space Phys., 13:735, 1975. Phys. , 52:2343, 1974.
Boyd, J. S., A.E. Belon, and G.C. Romick, “Latitude and Gattinger, R.L. and A. Vallance Jones, “Quantitative Spec-
Time Variations in Precipitated Electron Energy Inferred troscopy of the Aurora. V. The S ctrum of Strong
from Measurements of Auroral Heights,” J. Geophys. Aurora between 10,000 and 16,000 r ,“ Can. J. Phys.,
Res., 76: 7694–7700, 1971. 59:480, 1981.
Buchau, J., G.J. Gassmann, C.P. Pike, R.A. Wagner, and Gilmore, F. R., “Potential Energy Curves for Nz, NO, Oz
J.A. Whalen, “Precipitation Patterns in the Arctic Ion- and Corresponding Ions,” J. Quant. Spectrosc. Radiat.
osphere Determined from Airborne Observations,” Anti. Transfer, 5:369, 1965.
Geophys., 28:443, 1972. Goldman, A. and S.C. Schmidt, “infrared Spectral Line
Buchau, J., C.P. Pike and M. Wong, “Detailed Specifi- Parameters and the Abso@tance Calculations of NO at
cation of the Arctic Ionosphere and an Application to Atmospheric and Elevated Temperatures fQr the Av = 1
Three-Dimensional Raytracing,” AFCRL-TR-73-0726, Bands Region,” J. Quant. Spectrosc. Radiat. Transfer,
AD775084, 1973. 15:127, 1975.
Caledonia, G.E. and J.P. Kennealy, “NO Infrared Radiation Goody, R. M., Atmospheric Radiation, 1, Clarenden Press,
in the Upper Atmosphere,” Pbnet. Space Sci., 30:1043, Oxford, U. K., 1964.
1982. Gordiets, B. F., M.N. Markov, and L.A. Shelepin, “1.R.
Chamberlain, J. W., Physics of the Aurora and Airglow, Radiation of the Upper Atmosphe~,” Pknet. Space Sci.,
Academic Press, New York, 1961. 26:933, 1978.
Cohn, A. and G. Caledonia, “Spatial Distribution of the Gussenhoven, M. S., D.A. Hardy, and N. Heinemann,

12-39
 1

CHAmEk 12
“Systematic of the Equatorward Diffuse Auroral second edition, edited by M. H. Bofier AndT. Batirer,
Boundary,” J. Geophys. Res., 88:5692, 1983. DNA 1948H, 1972.
Hanson, W. B. (cd.), “Special Issue on Atmospheric Ex- Kennealy, J. P., F.P. DelGreco, G.E. Caledonia, and B.D.
plorer Mission,” Radio Sci., $(No. 4): 19?31 Green, “Nitric Oxide Chemiexcitation Occurnng in the
Harang, L., The Aurora, Wiley, New York, 1951. Reaction Betweeti Metastable Nitrogen Atoms and Ox-
Hartz, T.R. and N.M. Brice, “The General Pattern of Au- ygen Molecules,” J. Chem. Phys., 691574, 1978.
roral Particle Precipitation,” Planet. Space Sci., 15301, Kockarts, G., “Nitric Oxide C@ling in the Terrestrial At-
1967. mosphere,” Geophys. Res. Lett., 7: 137, 1980.
Heelis, R. A., J.D. Winningham, W.B. Hanson, @ J.L. mkonis, A.P. and J.A. Whalen, “Occurrence and Life-
Burch, “The Relationships Between High-Latitude Con- times of Discrete Auroras Ne@ Midnight ,“ J. Geophys.
vection Reversals and the Energetic Particle Morphology Res., 85:119, 1980.
Observed by Atmosphere Explorer,” J. Geophys. Res., Krupenie, P. H., “The Spectrum of Molecular Oxygen,” J.
85:3315, 1980. Phys. Chem. Ref. Data, 1:423, 1972.
Herzberg, G., Molecular Spectra and Molecular Structure, Kymer, J. B., “Summary Analysis of 4. 3pm Data,” in At-
2, Van Nostrand, Princeton, N. J., 1945. mospheres of Earth and the Planets, edited by B. M.
Herzberg, G., Molecular Spectra and Molecular Structure, McCormac, D. Reidel, Dordrecht, Holland, 347–358,
1, second edition, VW Nostrand, New York, 1950. 1975,
Herzberg, G., Molecular Spectra and Molecular Structure, Kutier, J. B., “Theory of the C02 4.3-pm ,Aurora and Re-
3, Van Nostrand, Princeton, N. J., 1967. lated Phenomena,” J. Geophys. ~es.; 82:2203, 1977.
Hilliard, R.L. and G.G. Shepherd, “Upper Atmospheric Loftus, A. and P.H. Kmpenie, “ne Spectrum of Molecular
Temperatures from Doppler Line Widths-IV,” Planet. Nitrogen,” J. Phys. Chem. Rej Ddta, 6:113, 1977.
Space Sci., 14:383, 1966. Lui, A.T.Y. tid C.D. Anger, “A Uniform Belt of Diffuse
Holzworth, R.H. and C.-I Meng, “Mathematical Represen- Aurora and Airglow from ISIS 2 Spacecraft,” Planet.
tation of the Aurora] Oval,” Geophys. Res. Lett., 2: Space Sci., 21:819, 1973.
377, 1975. Lui, A. T. Y., P. Perreault, S.-I. Akasofu, and C.D. Anger,
Hook, J. L., “Morphology of Auroral Zone Radiowave Ab- “The Diffuse Aurora,” Planet. Space Sci., 21: 857,
sorption in the Alaska Sector,” J. Atmos. Terr. Phys., 1973.
30:1341, 1968. Lui, A. T. Y., D. Ve@atesan, C.D. Anger, S.-L Akasofu,
Huber, K.P. and G. Herzberg, Molecular Spectra and Mo- W.J. Heikkila, J.D. Winningharti, and J.R. Burrows,
leczdarStructure, 4, Van Nostrand Reinhold, New York, “Simultaneous Observations of Particle Precipitations
1979. and Auroral Emissions by the 1S1S 2 Satellite in the
Huffman, R. E., F.J. LeBlanc, J.C. Lmabee, and D.E. 19–24 MLT Sector,” J. Geophys. Res., 82:2210, 1977.
Paulsen, “Satellite Vacuum Ultraviolet Airglow and Au- McClatchey, R. A., W.S. Benedict,. S.A. Clough, D.E.
roral Observations,” J. Geophys. Res., 85:2201, 1980. Burch, R.F. Calfee, K. Fox, L.S. Rothman, and J.S.
Hunten, D. M., “Some Photometric Observations of Auroral Garing, “AFCRL Atmospheric Absorption Line Param-
Spectra,” J. Atmos. Terr. Phys., 7:141, 1953. eters Compilation,” AFCRL TR-73-0096, AD762904,
Hunten, D. M,, E.G. Rawson, and J.K. Walker, “Rapid 1973.
Measurement of NY Rotational Temperatures in Au- Mende, S.B. and R.H. Esther, “MonWhromatic All Sky
rora,” Can. J. Phys., 41:258, 1963. Observations and Auroral Precipitation Patterns,” J.
Hunten, D. M., F.E. Roach, and J. W. Chamberlain, “A Geophys. Res., 81:3771, 1976.
Photometric Unit for the Airglow and Aurora,” J. At- Meng, C.-I., “Diurnal Variation oi the Auroral Oval Size,”
mos. Terr. Phys., 8:345, 1956. J. Geophys. Res., 84:5319, 1979.
Huppi, R.J., and A.T. Stair, Jr., “Aurorally Enhanced In- Meng, C.-I., R.H. Holzwofi, and S.-I. Akasofu, “Auroral
frared Einissions,” Appl. Opt., la: 3394, 1979. Circle—Delineating the Polew~d Bound~ of the Quiet
Huppi, E. R., J. W. Rogers, and A.T. Stair, Jr., “Aircraft Auroral Belt,” J. Geophys. Rei., 82:164, 1977.
Observations of the Infrared Emission of the Atmos- Muldrew, D.B., “F-Layer Ionization Tmghs Deduced from
phere in the 700-2800 cm-’ Region,” Appl. Opt., 13: Alouette Data,” J. Geophys. Res., 702635, 1965.
1466, 1974. Murphy, R., R. Shma, R. Picard, J. Winick, K.D. Baker,
International Union of Geodesy and Geophysics, interna- and J .C. Ulwick, “~eliminary Interpretation of Infrared
tional Auroral Atlas, Edinburgh University Press, 1963. Interferometer Rocket Measurements,” Eos, 64: 785,
Jasperse, J.R. and B. Basu, “Transport Theoretic Solutions 1983.
for Auroral Proton and H Atom Fluxes and Related Neugebauer, G., et al., “The Infr~ed As&onomical Satellite
Quantities,” J. Geophys. Res., ?7: 811, 1982. (I~S) Mission,” Astrophys. J., 278: L1, 1984.
Jones, R.A. and M.H. Rees, “Time Dependent Studies of Omholt, A., “The Auroral E-Layer Ionization and the Au-
the Aurora, I. Ion Density and Composition ,“ Planet. roral Luminosity,” 3. Atmos. Terr. Phys., 7:73, 1955.
Space Sci., 21:537, 1973. O’Neil, R. R., F. Bien, D. Burt, J.A. Sandock, and A.T.
Kennealy, J .P. and F.P. Del Greco, “The Kinetics of At- Stair, Jr., “Summarized Results of the Artificial Auroral
mospheric Radiative Processes in the Infrared,” Chapter Experiment, Precede,”J. Geophys. Res., 83:3273, 1978a.
11, Defense Nuclear Agency Reaction Rate Handbook, O’Neil, R. R., O. Shepherd, W.P. Reidy, J.W. Carpenter,

12-40
 THEAURORA
T.N. Davis, D. Newell, J.C. Ulwick, and A.T. Stair, Within the Region 2 Field-Aligned Current Sheet,” J.
Jr., “Excede2Test, an Artificial Auroral Experiment: Geophys. Res., 87:731, 1982.
Ground-Based optical Measurements,” J. Geophys. Res., Roble, R.G. and J .M. Gary, “The Effect of Horizontal
83:3281, 1978b. Transport on Auroral NO Densities,” Geophys. Res.
O’Neil, R. R., E.T.P. Lee, and E.R. Huppi, “Auroral 0(’S) Lett., 6:703, 1979.
Production and Loss Processes: Ground-Based Mea- Roble, R.G. and M.H. Rees, “Time-Dependent Studies of
surements of the Artificial Auroral Experiment Pre- the Aurora: Effects of Particle Precipitation on the Dy-
cede,” J. Geophys. Res., 84: 823, 1979. namic Morphology of Ionospheric and Atmospheric
O’Neil, R. R., A.T. Stair, Jr., W.R. Pendleton, Jr., and Properties,” Planet. Space Sci., 25:991, 1977.
D.A. Burt, “The EXCEDE SPECTRAL Artificial Au- Rothman, L. S., “AFGL Atmospheric Absorption Line Pa-
roral Experiment: An Overview ,“ in Artlfcial Particle rameters Compilation: 1980 Version,” Appl. Opt., 20:
Beams in Space Plasma Studies, edited by B. Grandal, 791, 1981.
Plenum, New York, pp. 207–215, 1982. Rothman, L.S. and L.D.G. Young, “Infrared Energy Levels
Peterson, L. R., T. Sawada, J.N. Bass, and A.E.S. Green, and Intensities of Carbon Dioxide—II,” J. Quant. Spec-
“Electron Energy Deposition in a Gaseous Mixture,” trosc. Radiat. Transfer, 25:505, 1981.
Computer Phys. Commun., 5:239, 1973. Rothman, L. S., A. Goldman, J.R. Gillis, R.R. Gamache,
Pike, C. P., “An Analytical Model of the Main F-Layer H.M. Pickett, R.L. Poynter, N. Husson, and A. Chedin,
Trough,” AFGL TR-76-0098, ADA026031, 1976. “AFGL Trace Gas Compilation: 1982 Version,” App/.
Pike, C.P. and J.A. Whalen, “Satellite Observations of Opt., 22:1616, 1983a.
Auroral Substoxms,” J. Geophys. Res., 79:985, 1974. Rothman, L. S., R.R. Gamache, A. Barbe, A. Goldman,
Piper, L.G., “The Excitation of O(lS) in the Reaction Be- J.R. Gillis, L.R. Brown, R.A. Toth, J.-M. Flaud, and
tween N*(A 3Z~ ) and 0(3P),” J. Chem. Phys., 77:2373, C. Camy-Peyret, “AFGL Atmospheric Absorption Line
1982. Parameters Compilation: 1982 Edition,” Appl. Opt., 22:
Piper, L.G., G.E. Caledonia, and J.P. Kennealy, “Rate 2247, 1983b.
Constants for Deactivation of N*(A) v’ = 0,1 by O*,” Sandford, B.P., “Variations of Awral Emissions with Time,
J. Chem. Phys., 74:2888, 1981a. Magnetic Activity and the Solar Cycle,” J. Atmos. Terr.
Piper, L.G., G.E. Caledonia, and J.P. Kennealy, “Rate Phys., 30:1921, 1968.
Constants for Deactivation of Nz(A 3Z~, V’ = O,1) by Seaton, M.J., “Excitation Processes in the Aurora and Air-
0,” J. Chem. Phys., 75:2847, 1981b. glow. 1. Absolute Intensities, Relative Ultraviolet In-
Rawlins, W.T,, G.E. Caledonia, J.J. Gibson, and A.T. Stair, tensities, and Electron Densities in High Latitude Au-
Jr., “Infrared Emission from NO (Av = 1) in an Aurora: rorae,” J. Atmos. Terr. Phys., 4:285, 1954.
Spectral Analysis and Kinetic Interpretation of HIRIS Sharber, J. R., “The Continuous (Diffuse) Aurora and Au-
Measurements,” J. Geophys. Res., 86:1313, 198la. roral-E Ionization,” in Physics of Space Plasmas, edited
Rawlins, W. T., G.E. Caledonia, and J.P. Kennealy, “Ob- by T.S. Chang, B. Coppi, and J.R. JasperSe, SPI Con-
servation of Spectrally Resolved Infrared Chemilumi- ference Proceedings and Reprint Series, 4, Scientific
nescence from Vibrationally Excited 03(v3) ,“ J. Geo- Publishers, Cambridge, Mass., 1981.
phys. Res., 86:5247, 1981b. Sharp, W. E., “The Ultraviolet Aurora: The Spectrum Be-
Rearwin, S. and E. W. Hones, “Near Simultaneous Mea- tween 2100 ~ and 2300 ~,” Geophys. Res., Lett., 5:
surement of Low-Energy Electrons by Sounding Rocket 703, 1978.
and Satellite,” J. Geophys. Res., 79:4322, 1974. Sharp, W .E. and M.H. R=s, “Auroral Spectrum between
Rees, M. H., “Auroral Ionization and Excitation by Incident 1200 and 4000 Angstroms,*’J. Geophys. Res., 77:1810,
Energetic Electrons,” Pknet. Space Sci., 11:1209, 1%3. 1972.
Rees, M .H., “Note on the Penetration of Energetic Electrons Sheehan, R.E. and R.L. Carovillano, “Characteristics of
into the Earth’s Atmosphere,” Planet. Space Sci., 12: the Equatoward Auroral Boundary Near Midnight De-
722, 1964. termined from DMSP Images,” J. Geophys. Res., 83:
Rees, M.H. and D. Luckey, “Auroral Electron Energy De- 4749, 1978.
rived from Ratio of Spectroscopic Emissions 1. Model Sheehan, R. E., R.L. Carovillano, and M.S. Gussenhoven,
Computations,” J. Geophys. Res., 79:5181, 1974. “Occurrence and Recumnce in Auroral Activity in DMSP
Rees, M.H. and R.G. Roble, “Effect of a Horizontal Wind Images,” J. Geophys. Res.. 87:3581, 1982.
on 1.27-pm Auroral Emission from 02 ( ‘Ag)Mole- Shepherd, M.M. and R.H. Eather, “On the Determination
cules,” J. Geophys. Res., 85:4295, 1980. of Auroral Electron Energies and Fluxes from Optical
Rees, M. H., A.I. Stewart, and J.C.G. Walker, “Secondary Spectral Measurements,” J. Geophys. Res., 81: 1407,
Electrons in Aumra,’’Phwt. Space Sci., l% 1997, 1%9. 1976.
Reidy, W. P., T.C. Degges, A.G. Hurd, A.T. Stair, Jr., Slater, D. W., L.L. Smith, and E. W. Kleckner, “Correlated
and J. C. Ulwick, “Auroral Nitric Oxide Concentration Observations of the Equatorward Diffuse Auroral
and Infrared Emission,” J. Geophys. Res., 87: 3591, Boundary,” J. Geophys. Res., 85:531, 1980.
1982. Spiro, R. W., P.F. Reiff, and L.J. Maher, “Precipitating
Robinson, R, M., R.R. Vondrak, and T.A. Potemra, “Elec- Electron Energy Flux and Auroral Zone Conductances—
trodynamicsProperties of the Evening Sector Ionosphere An Empirical Model,” J. Geophys. Res., 8% 8215, 1982.

12-41
CHAPTER 12
Stair, A. T., Jr., J.C. Ulwick, K.D. Baker, and D.J. Baker, Wadzinski, H.T. and J.R. Jasperse, “Low Energy Elec-
“Rocketbome Observations of Atmospheric Infrared tron and Photon Cross Sections for O, N2, and Oz,
Emissions in the Auroral Region,” in Atmospheres of and Related Data,” AFGL-TR-82-OO08, ADA 118921,
Earth and the Planets, edited by B.M. McCormac, D. 1982.
Reidel, Dordrecht, Holland, pp. 335-346, 1975. Wagner, R.A. and C.P. Pike, “A Discussion of Arctic Ion-
Stair, A.T., Jr., R. Nadile, J.C. Ulwick, K.D. Baker, and ograms” in Radar Propagation in the Arctic, Chapter
D.J. Baker, “Infrared Measurements of Aurora, Airglow 4, edited by J. Frihagen, Technical Editing and Repro-
and the Upper Atmosphere,” in Proceedings of the 8th duction, London, 1972.
Annual Meeting on Upper Atmospheric Studies by Op- Wagner, R. A., A.L. Snyder, and S.-1. Akasofu, “The
tical Methods, edited by E. O’Mongain, University Col- Structure of the Polar Ionosphere During Exceptionally
lege, Dublin, pp. 27–35, 1980. Quiet Periods,” Planet. Space Sci., 21:1911, 1973.
Stair, A. T., Jr., J. Pritchard, I. Coleman, C. Bohne, W. Werner, H.-J. and P. Rosmus, “Ab Initio Calculations of
Williamson, J. Rogers, and W .T. Rawlins, “Rocket- Radiative Transition Probabilities in the X 1X+ Ground
bome Cryogenic ( 10 K) High-Resolution Interferometer State of the NO+ Ion,” J. Mol. Spectrosc., %: 362,
Spectrometer Flight HIRIS: Auroral and Atmospheric 1982.
IR Emission Spectra,” Appl. Opt., 22:1056, 1983. Whalen, J. A., “Auroral Oval Plotter and Nomograph for
Starkov, G. V., “Auroral Heights in the Polar Cap,” Geo- Determining Corrected Geomagnetic kal Time, Lat-
mag. Aeron., 8: 28, 1968. itude and bngitude for High Latitudes in Northern
Starkov, G. V., “Analytical Representation of the Equatorial Hemisphere,” AFCRL 70-0422, AD713 170, 1970.
Boundary of the Oval Auroral Zone,” Geomag. Aeron., Whalen, J.A. “Characteristics of the High Latitude Iono-
9:614, 1969. sphere Produced by Auroral Particle Precipitation ,“ in
Starkov, G.V. and Y.I. Feldstein, “Position of the Auroral AGARD Conference Proceedings, No. 263, Special
Belt at Night During the IGY and IQSY” in Complex Topics in HF Propagation, 1979.
Investigations of the Polar Ionosphere, edited by Isaev, Whalen, J. A. “General Characteristics of the Auroral Ion-
Nauka Press, Leningrad, 1970. osphere” in Physics of Space Plasmas, edited by T. S.
Steed, A., J.C. Ulwick, C. Harris, F. Cook, and R. Straka, Chang, B. Coppi, and J.R. Jasperse, SPI Conference
“Rocketbome Interferometer Measurements of SWIW Proceedings and Reprint Series 4, Scientific Publishers,
MWIR Spectra,” EOS, 64:784, 1983. Cambridge, Mass., 1981.
Stolarski, R.S. and A.E. S. Green, “Calculations of Auroral Whalen, J. A., “A Quantitative Description of the Spatial
Intensities from Electron Impact ,“ J. Geophys. Res., Distribution and Dynamics of the Energy Flux in the
72:3967, 1967. Continuous Aurora,” J. Geophys. Res., 88:7155, 1983.
Strickland, D. J., D.L. Book, T.P. Coffey, and J.A. Fedder, Whalen, J.A. and J.R. Sharber, “TheNighttime Auroral-E
“Transport Equation Techniques for the Deposition of Laye~ Partical Production, Latitudinal and hngitudinrd
Auroral Electrons,” Y. Geophys. Res., 81:2755, 1976. Structure and Dynamics” in The Effects of the iono-
Strickland, D. J., J,R. Jasperse, and J.A. Whalen, “Depen- sphere on Radiowave Systems, edited by J. Goodman
dence of Auroral FUV Emissions on the Incident Elec- and F. Clarke, U.S. Government Printing Office, Wash-
tron Spectrum and Neutral Atmosphere,” J. Geophys. ington, D. C., 1981.
R-es., 88:8051, 1983. Whalen, J. A., J. Buchau, and R.A. Wagner, “Airborne
Swider, W. and R.S. Narcisi, “Auroral E-Region: Ion Com- Ionospheric and Optical Measurements of Noontime Au-
position and Nitric Oxide,” Planet. Space Sci., 25:103, rora,” J. Atmos. Terr. Phys., 33:661, 1971.
1977. Whalen, J. A., R.A. Wagner, and J. Buchau, “A 12-Hour
Tom, M.R. and D.G. Tom, “The Role of Metastable Species Case Study of Auroral Phenomena in the Midnight Sec-
in the Thermosphere,” Rev. Geophys. Space Phys., 20: tor Oval, Polar Cap and Continuous Auroras,” J. Geo-
91, 1982. phys. Res., 82:3529, 1977.
Vallance Jones, A., Aurora, D. Reidel, Dordrecht, Holland, Winckler, J. R., “The Use of Artificial Electron Beams as
1974. Probes of the Distant Magnetosphere,” in Artificial Par-
Vallance Jones, A. and R.L. Gattinger, “Quantitative Spec- ticle Beams in Space Plasma Studies, edited by B. Gran-
troscopy of the Aurora. III. The ~Spectrum of Medium dal, Plenum, New York, pp. 3–33, 1982.
Intensity Aurora Between 3100 A and 4700 ~,” Can. Winningham, J. D., F. Yasuhara, S.-I. Akasofu, and W .J.
J. Phys., 53: 1806, 1975. Heikkila, “The Latitudinal Morphology of 10 eV to 10
Vallance Jones, A. and R.L. Gattinger, “Quantitative Spec- keV Electron Fluxes during Magnetically Quiet and Dis-
troscopy of the Aurora. IV. Th~ Spectrum of Medium turbed Times in the 2100 to 0300 MLT Sector,” J.
Intensity Aurora between 8800 A and 11400 ~,” Can. Geophys. Res., 80:3148, 1975.
J. Phys., 54:2128, 1976.

12-42
 Chapter 13

INFRARED AIRGLOW
R.D. Sharma

Infrared radiance observed by a detector looking at the Table 13-1. OH Vibrational Transition Band On~ins
-. [L.S. Rothman
et al., 1983]
earth’s undisturbed atmosphere is called infrared airglow.
This radiation may arise from rotational transitions, as is the Av=l Av=2 Av=3
case for radiation from water vapor around 17 pm, or from
electronic transitions, for example, the Oz (alA~ + X3Z~-) v’ (cm-i) (~m) (cm-’) (pm) (cm-’) (Win)
transition causing the airglow at 1.27 pm (O-O band) and
1 3570 2.80
at 1.58 pm (O-1 band). However, the most common sources
2 3404 2.94 6974 1.43
of infrared airglow are the vibrational-rotational (V-R)
3 3240 3.09 6644 1.50 10214 0.98
transitions within the ground electronic state. This class of
4 3078 3.25 6318 1.58 9722 1.03
transitions includes emission from OH, COZ, 0s, NO, and
5 2915 3.43 5993 1.67 9233 1.08
H20. Figure 13-1 gives the energy level diagram of COZ
6 2752 3.63 5667 1.76 8745 1.14
and points out the more commonly encountered transitions.
7 2585 3.87 5337 1.87 8252 1.21
Table 13-1 gives the band origins and Table 13-2 the Ein-
8 2415 4.14 5000 2.00 7752 1.29
stein A coefficients for some OH V-R bands. For addi-
9 2237 4.47 4652 2.15 7237 1.38
tional information on the energy levels and the radiative
lifetimes of the infrared active species see Chapter 12. The
extent of the wavelength range covered by the term infrared tral region from 1.2 to 17 pm. This rather arbitrar-
is not well defined. We will limit this discussion to the spec- ily restricted wavelength range leaves out several important
airglow emissions, including the 63 pm emission from ox-
-1
x 103cm ygen atoms [Offermann and Grossmann, 1978]. This latter
process together with the 5.3 pm (1 + O) emission from
NO and the 15 pm radiation from the bending mode of COZ
are the important processes in the cooling of the thermo-
sphere [Craig and Gille, 1969; Kockarts, 1980; Gordiets et
. al., 1983; Zachor, Sharma, Nadile and Stair, 1981].
Infrared airglow has been the subject of a number of
stidies [Stair et al., 1981; Gordiets et al., 1978; Ogawa,
1976]. It is a quickly changing field because of the rapidly
improving measurement and analysis twhnology and it is
z$:g:; not surprising, therefore, that the information in the liter-
— 20003
ature is a few years behind that available to workers in the
field. For this reason, the emphasis in this chapter has been
2 7pM[ 2.0PM
placed upon results obtained at AFGL.
13.9pM 16 2PM
01 101

15.0PM
— > 13.1 SOURCES OF THE
‘1 ‘2 ‘3 VI “2 ‘2 ‘3 VI V3 INFRA-RED-AIRGLOW

Figure 13-1. The energy levels of carbon dioxide. The lowest levels of
bending mode and asymmetric stretch, which lead to tran-
sitions at 15.0 and 4.26 pm, are shown. Also shown are
13.1.1 Resonant Scattering of the Earthshine
the upper and lower levels for some of the important tran-
sitions observed in the infrared airglow of the earth’s at- Radiation upwelling through the atmosphere from the
mosphere. The notation in the figure is the same as in AFGL
line tape [Rothman and Young, 1981]. (Reprinted with per-
surface of the earth is called earthshine. Although the av-
mission from Pergamon Press Ltd. O 1981.) erage temperature of the earth’s surface is 296 K, the tem-

13-1
 a

CHAPTER 13
Table 13-2. Thermallv-avera~ed OH (X2 m,) vibrational Einstein coefficients AV,V (S-’) for 2M K [Mies, 19741 (Reprinted with Permission from Academic
Press O i974) -

v“ = o 1 2 3 4 5 6 7 89

V’=1 20.15
2 14.07 25.24
3 0.92 39.93 20.90
4 0.079 4.29 72.61 12.25
5 0.050 0.392 10.62 108.4 4.47
6 0.053 1.274 20.99 141.5 2.347
7 0.183 2.907 37.32 163.4 9.142
8 0.030 0.569 5.67 60.84 166.8 25.84
9 0.130 1.189 10.81 90.28 146.6 50.65

perature of the earthshine depends upon altitude and wave- A process capable of producing large amounts of in-
length. For example in the 12-13 pm region, where the frared airglow radiation is the electronic excitation of mol-
absorption due to the earth’s atmosphere is minimal, the ecules by sunshine followed by emission of a longer wave-
upwelling radiation can be described by 296 K blackbody length photon. The molecule is thus left in an excited
radiation independent of altitude. On the other hand, radia- vibrational state, the final vibrational level being determined
tion in the 15 pm region up to the lower mesosphere (about by the Franck-Condon integrals. Several quanta of infrared
70 km altitude) can be described as radiation from a black- photons may be emitted before the molecule relaxes to the
body with temperature given by the local translational tem- ground state. This process, however, has not been identified
perature. At higher altitudes, because of the lack of sufficient as the source of any airglow from the quiescent atmosphere.
collisions, the upwelling 15 pm radiation can be described
by about 200 K blackbody temperature.
Whatever the variations in the temperature of the ear- 13.1.3 Photodissociation
thshine, it is clear that this radiation has appreciable intensity
only at longer wavelengths (k > 5 pm). It can, therefore, Dissociation of molecular species by solar radiation con-
excite only the lower vibrational levels of the infrared active tributes to the airglow by producing fragments in excited
molecules. The excited molecules can reradiate isotropically states. An example is the airglow of 02 (al A~ ~ X3Zg-) at
without change of frequency causing infrared airglow. Res- 1.27 pm (O-O transition) and at 1.58 pm (O-1 transition).
onant scattering of earthshine by the V3 mode of ozone is The Oz(alA~) state is mainly produced by the photodisso-
an important contributor to the airglow at 9.6 pm. The 15 ciation of 03 by solar ultraviolet radiation around 255 nm.
~m radiation from the bending mode of carbon dioxide also Photodissociation of molecular species can also modify
contains an important contribution from the earthshine. the infrared airglow due to reduced density of the photo-
dissociating molecules and/or increased density of the prod-
uct species. Photodissociation of 03 also leads to reduced
13.1.2 Scattering of Sunshine daytime airglow of OH not only because of the reduced
production of OH by the reaction H + Os ~ OH + Oz
In contrast to the earthshine which populates only the but also due to the more rapid OH + O ~ Oz + H de-
lowest lying vibrational levels, solar radiation (- ~ K pletion reaction. The latter reaction becomes faster because
blackbody temperature) plays an important role in exciting of the enhanced oxygen atom concentrations produced by
higher vibrational levels corresponding to k s 5 pm. The photodissociation of 03. The net result is that daytime OH
vibrational levels thus produced can lose their energy by emissions around 1.5 pm (Av = 2) and around 3.0 pm
optical transitions to the ground state (resonant scattering) (Av = 1) are much weaker than their nighttime intensities.
or by radiation to other lower excited vibrational levels
(nonresonant scattering or fluorescence). The effect of sun-
shine is shown in the dramatic increase in the airglow near 13.1.4 Chemical Reactions
4.3 pm during the day due to scattering by the VJ mode of
C02 [James and Kumer, 1973]. Scattered solar radiation, Creation of molecular species in the excited state by
resonant and nonresonant, also contributes to the COZ air- chemical reactions and their subsequent decay contributes
glow at 2.7 pm, the important absorption bands being lo- to airglow. The airglow produced by this mechanism is also
cated near 2.7 ~m [Kumer, 1981], 2.0, and 1.6 pm [Sharma called chemiluminescence. A well-known example of this
et al., 1981]. process is the OH radiation around 1.5 and 3.0 pm and at

13-2
 INFRARED AIRGLOW
shorter wavelengths, even visible regions, OH being pro- because of the very small cross sections involved. In ad-
duced by the chemical reaction dition, there is Rayleigh scattering of solar radiation im-
H+03+OH*+ O*, portant only below about 40 km altitude.

the asterisk denoting the vibrational excitation. Another ex-

13.2 NON-LOCAL THERMODYNAMIC
ample of tfiis process [Stair et al., 1981, 1985] is the ra-
EQUILIBRIUM (NLTE)
diation around 9.6 ~m from the V3mode of OS produced
by the three body recombination reaction
We note again that except for 1.27 pm and 1.58 pm
0+02+ M~O; +M. Oz(a ‘A~+ X3Z~-) radiation the ihfrared airglow from 1.2
to 15 pm involves molecular vibrational-rotational transi-
tions. In the lower p& of the atmosphere (below about 40
13.1.5 Atomic and Molecular Collisions km altitude) these vibrational transitions can generally be
described by local thermodynamic equilibrium (LTE), that
Collisions of infrared radiating species with other mol- is, vibrational excitation of the molecules can be described
ecules lead to transfer of vibrational energy from one col- by a Boltzmann distribution with a vibrational temperature
lision partner to the other. This transfer is much more ef- equaling me translational tem~rathre. Even at these alti-
ficient if the vibrational @uencies of the ~sitions involved tudes the 02(a1A~) electronically excited state population
are not too far apart (near-resonant), that is W7 - 1, where cannot be described by a distribution with a temperature
o is the frequency separation between the transitions in- equaling the translational temperature and is much larger
volved and 7 the time duration of the collision. Because of than predicted on the basis of LTE. The reason for this lies
the near-resonance, Nz transfers its vibrational energy ef- in the metastable nature of the 02(a’Ag) state (its radiative
ficiently to the V3mode of CO, [Sharma and Brau, 1969]. lifetime is about 65 rein) and small rates of collisional
This energy transfer process plays an important role in the quenching by major atmospheric species [McEwan and Phil-
enh~ced nighttime 4.3 pm airglow of C02 around 70 km lips, 1975]. This metastable nature and the fact that colli-
[Kumer et al., 1978]. sions are not a rapid process for depopulating this state,
Another important type of collision producing infrared account for the great abundance of 02(a’A~) in the strato-
airglow is that in which the relative translational energy of sphere and mesospliere.
atom-molecule or molecule-bcilecule is converted into vi: At higher altitudes (starting about 70 km) the vibrational
brational excitation of the molecule. These collisions lead population of the radiating species, even those that we not
to establishment of local thermodynamic equilibrium (LTE) metastable and are neither produced by chemical reaction
between various degrees of freedom, the only requirement nor pumped by sunshine or earthshine, becomes different
being that enough of these take place during the effective from what would be predicted on the basis of LTE. Situ-
radiative lifetime of the emitter (taking into account selfab- ations deviating from LTE behavior are called NLTE cases.
sorption). Even when the frequency of collisions is not Sometimes it is useful to introduce the concept of vibrational
sufficient to establish LTE, this type of collision may be an temperature. The concept is based on the fact that near-
important source of vibrational excitation. Examples are resonant vibration-to-vibration transfer processes are much
collisions above 90 @ of oxygen atoms with COZ and NO faster than nonresonant vibration-to-translation processes.
which play a very impomnt role in the airglow at 15 pm The time needed for the vibrati~nal energy to be partitioned
and 5.3 pm, respectively [Degges, 1971; Zachor et al., according to a Boltzmann dist~bution in a collision between
1981; Shpa and Na~ile, 1981]. identical molecules is usually several orders Qf magnitude
Another important source of vibrational energy is the shorter than the time needed to establish LTE in gases [Sharma,
collisional excitation of molecules by electronically excited 1969; Taylor, 1974]. The vibrational temperature T, is de-
atoms. For example, oxygen atoms produced by photodis- fined by tie relation
sociation of OS are created in the excited ‘D state with unit
efficiency. These atoms are very efficiently quenched by ,n g= nl Cz E
molecular nitrogen. Runng the quenching process, a frac-
tion of the electronic excitation of oxygen atoms is cdnverted
(–)
glne ‘— T.

into vibrational energy of N2 [McEwan and Phillips, 1975] where u and n, are the excited state afid low state number
and can be collisionally transferred to C02 and emitted as densities, g= and gl are the respective statistical weights,
4.3 ~m airglow. C2 = 1.4399 IUcm-’ is the second radiation constant and
The processes described in the above sections are the E is the energy difference between the upper and lower
most important but by no means the only sources of infrared states in cm-’.
radiation from the undisturbed atmosphere. Raman scatter- The concept of vibrational temperature has been used
ing of the solar radiation is a less important soutce of airglow to calculate [Sharrna and Nadile, 1981] the contribution to

13-3 --
CHAPTER 13
the 15 pm radiation by the transition 02201 + 01101 of gains altitude, less and less of the gas is being looked at so
COZ. Only when this contribution is added to that of the that the signal during ascent usually decreases with time
01101 + 00001 transition from all the isotopes is a rea- (except during atmospheric disturbances); during the descent
sonable agreement obtained between the 15 pm radiation the reverse is true. However, both in ascent and, descent,
observed by the rocket experiment SPIRE in the 70 and 90 the magnitude of the derivative of the signal with altitude
km tangent height and the calculated C02 radiances [Stair is a measure of the local radiance—provided we are dealing
et al., 1981, 1985]. with an optically thin and steady state one-dimensional at-
The concept of vibrational temperature cannot be used mosphere. Figure 13-2 shows a typical spectral profile of
in all NLTE situations. For example, the vibrational pop-
ulation of OH around 80 km altitude cannot be described ENERGY BUDGET
by a Boltzmann distribution. This concept, however, is used HS4 CVF
where possible because it also gives an idea of the extent SCAN 5
TAL 71.85 sec
of deviation from LTE.
ALT 87,75 km
It should be pointed out that while fewer collisions lead t [ 1 1 1 I .- 1 I I
to NLTE situations, more frequent collisions do not guar-
antee LTE behavior. This is illustrated by the 2.7 pm day-
time COZ emission that can still be characterized around 50

1
km altitude by vibrational temperature about 100 K higher
than the translational temperature. This emission, in part,
is shown to arise from the absorption of sunlight at 1.5 pm
by very weak 00001 + 3001 (x = 1,2,3) bands (Figure
13- 1) [Sharma et al., 1981]. Although each absorbed photon
has a rather small probability of being emitted in the 2.7
pm region, the large number of absorbed photons yields a
12
detectable signal. WAVELENGTH (pm)
Having stated the differences between LTE and NLTE
situations in some detail, we will focus for the remainder
Figure 13-2. Plot of radiance vs wavelength from the zenith looking CVF
of this discussion on NLTE airglows, specifically 1–18 pm “Energy Budget” rocket at 87.75 km altitude. The signal
emissions originating from altitudes higher than 40 km. The from COZ at 15 pm is at least an order of magnitude larger
than that from NO at 5.3 ~m or ozone at 9.6 pm.
LTE situation has been discussed in great detail by other
workers [see, for example, Clough et al., 1981].

zenith emission from the NO vibrational-rotational (V-R)

13.3 METHODS OF MEASUREMENT fundamental transition (Av = 1) at 5.3 pm, the001 + 000
V-R band of 03 at 9.6 pm, and the 01101 + 00001 band
Infrared airglow has been studied with rockets [Stair et of COZ at 15 pm obtained by a rocket launched from An-
al., 1974; Nadile et al., 1977; Stair et al., 1981; Markov, doya, Norway during the Energy Budget Campaign [Ulwick
1969; Offermann and Grossmann, 1978] and satellites [Mar- et al., 1983]. This rocket was launched in November 1980
kov et al., 1976]. In this chapter the primary emphasis will at 0314 UT. Since the solar shadow height was 400 km
be, for reasons cited earlier, on the work at AFGL. Many above the launch site, the experiment was essentially con-
of the AFGL rockets were launched from Poker Flats, Alaska ducted in a night atmosphere.
[Stair et al., 1975]. The rocket payloads contained either Another approach useful for viewing weak infrared at-
limb-viewing or zenith-viewing light collecting and analyz- mospheric emissions is to add a telescope to the optical
ing instruments. In these experiments the instrument pack- system and observe the infrared radiation from the earth-
age included a protective cover that was opened at a spec- limb. The telescope with a small fraction of degree full
ified altitude during the free ballistic flight phase. The altitude viewing angle (usually about 1/4 degree) and excellent off-
for cover opening, determined by the requirement that the axis rejection keeps the footprint of the LOS a few kilo-
infrared emission due to shock heating by the rocket be meters wide. The limb-looking SPIRE (SPectral Infrared
sufficiently small, was 50 km or higher. Rocket Experiment) has obtained very useful results yield-
The simplest infrared sounding rockets measure the in- ing many new insights into the infrared airglow [Nadile et
frared radiance reaching the payload from the atmosphere al., 1977; Stair et al., 1981]. In the SPIRE experiment with
above it. If the payload is not provided with an attitude the rocket in the exosphere, the eamhlimb was scanned ver-
control system, the rocket usually cones around an axis that tically through a telescoped circularly variable filter (CVF)
makes a small angle with the vertical direction. When mea- spectrometer. The resolution of the instrument was 370 to
suring the infrared fluorescence, care is taken that the sun 4% depending on the wavelength. The tangent height (the
is never directly in the line-of-sight (LOS). As the rocket distance of closest approach to the earth’s surface) is varied

13-4
 a
INFRARED AIRGLOW
and limb experiments. Two important ones are the amount
of the gas sampled and the altitude dependence of the vol-
ume probed. In the zenith looking experiment, the length
of the LOS between any 1 km altitude interval is sec 6 km,
where 6 is the zenith angle of the LOS as long as (3 < 75°
so that the curvature of the earth is not important. The total
volume of the gas sampled is relatively small. Moreover,
the zenith experiment weights equal increments of altitude
USSR in the LOS equally. Such is not the case in a limb experi-
CANADA
ment. For example, a detector positioned at an altitude of
285 km looking at 100 km tangent height gathers photons
from a path length in excess of 3000 km. Further, for density
decreasing monotonically with altitude the volume of the
gas sampled drops sharply as we move away from the tan-
gent height. The length of the gas column with altitude
between 100 and 101 km is about 227 km. The length of
Figure 13-3, Geographical location of projection on earth’s surface of
the path sampled with each km increase in altitude decreases
tangent heights for various SPIRE scans. Tangent height
increases going outwards from the center of figure. Farthest in proportion to [fi – ~(n – 1)] where n and (n – 1)
tangent point is 1500 km away from launch site. are the upper and lower altitudes in km measured from the
tangent height. In the example cited above, the length of
the LOS between 110 and 111 km is 0.154 times that be-
by changing the telescope zenith artgle. As the tangent height tween 100 and 101 km. Thus, equal increments of altitude
is varied, the location of the LOS and the tangent point in the LOS are not equally weighted in the limb experiment
(point closest to the earth’s surface) also change (Figure and the radiation from tangent height usually predominates.
13-3). One result of this spatial scanning is the ability to The above remarks are valid only for an optically thin gas.
obtain a day, twilight or night LOS (Figure 13-4). This is
A further characteristic of the limb experiment is that
in contrast to the zenith-looking rocket where in ascent or because the Yolurne of the gas sampled is so large, the local
descent the region sampled is less than 100 km over the variations in the atmospheric structure are averaged out. On
surface of the earth. While the possibility of investigating the other hand, the large column sampled also permits the
day-night variations in a zenith-looking experiment is rather limb experiment to see some emissions missed by the zenith
limited, it does offer the possibilities of determining the experiment when perfotmed at the same sensitivity. For
morphology on a small scale. example, Figures 13-5 and 13-6 show 15 pm radiance from
There are numerous other differences between the zenith V2 band of CO~ seen by the SPIRE and Energy Budget
rockets as a function of altitude. It can be seen in Figure
13-5 that between 95 and 110 km tangent height the ~and
SPIRE SHADOW HEIGHT
[ 18~ SPATIAL SCAN NUM8ER
240
7
TO SUN
,6 ,5 *1
200 i
[

SCAN NUMBER

RADIANCE (W cm-2 sf1)

Figure 13-4. Plot showing visible ul~violet terminators for eight of twelve
Figure 13-5, Integrated radiance observed by SPIRE in the 14– 16 #m
SPIRE scans. Ultraviolet shadow height is 23 km higher than
interval as function of tangent height. Night scan 8 has 30%
visible shadow height due to thickness of ozone layer. Scans
9–12 are completely sunlit. higher radiance above 100 km.

13-5
CHAPTER 13
ENERGY BUDGET covered the 1.4 to 16 pm region. The telescopes had 1/4°
180 . .*OA~ I I
~ ~%n *
fields of view with high off-axis rejection (point source
A ~t CARBON DIOXIDE rejection of l@5 at 10 off-axis) and were coaligned to better
A
B 15 MICRON BAND RADIANCE
a than 0.1°. The payload reached an altitude of 285 km, and
‘8 o ASCENT the earthlimb was spatially scanned from hard earth to local
-’t B ‘
A DESCENT horizontal at 0.5 deg/s while taking spectra at 2 scans/s.
%
~ 140 – $
&
The spectrometers were built at Utah State University and
the telescopes at the Honeywell Radiation Center. With a
1- 1
1/4° field of view, the footprint of the LOS varied from 4
,-
+ to 8 km, depending upon the tangent height and the payload
A
INBAND
< altitude.
NOISE
EQUIVALENT Before we give the profiles of the emitted band radiance
’00 RADIANCE as a function of tangent height (for the limb-looking ex-
1
periment) or altitude (for the zenith-looking experiment),
i
L I I I I we would like to discuss ai~glow in general terms as a
-lo -e
10 10 function of altitude. For this purpose we would like to divide
BAND RADIANCE (~ cm-2sr-1) ! the atmosphere above 40 km into several regions. This di-
vision is created to facilitate discussion of the SPIRE data.
Figure 13-6. Integrated bandradiance around 15~mobsemed by zenith- Taking ihto account the fact that it was an experiment of
looking Energy Budget rocket during ascent and descent as
about 5 min duration at dawn, it is quite possible that the
function of altitude.
boundaries of the regions created here may have to be mod-
ified in context of other data bases. Still we believe it is a
radiance stays constant although the COZ concentration de- useful exercise to point out underlying sources common to
creases ip this altitude range by a factor of 30 to 40. Detailed the airglow from different bands.
modeling has shown this phenomenon to be due to oxygen
atoms, which are much more efficient at collisionally pump-
ing the COZ (01 101) level than Nz or Oz [Sh-a and Nadile, 13.4.1 Thermosphere (above 100 km)
198 1]. The zenith experiment (Figure 13-6) does not give
any hint of this behavior in the measured radiances. This region is characterized by a rapid increase of tem-
Finally, the tangent height in the earthlimb experiment perature with altitude. Important emitters in this region are
can be varied from zero (earth’s surface) to 250 km, whereas restricted to NO and C02. The airglow, at 5.3 and 15 pm,
the range of altitudes attainable in the zenith experiment is in this region can be explained primarily in terms of oxygen
more limited. The two types of experiments emphasize di- atom collisions with the emitters. The oxygen atom con-
ferent aspects of infrwed airglow and each have their own centration in this region shows no diurnal variation. The
advantages and disadvantages. result is that infrared radiation from the excited states pro-
duced during the collision of the radiating species with the

13.4 RESULTS AND DISCUSSION

In this section we introduce the prominent features of

the measured airglow and compare them with model cal-
culations. Because of the nature and scope of this ch?pter,
we omit the details of the calculations and approach the
subject as an introduction to infrared airglow from the un-
disturbed atmosphere. Where possible, the published work
is referenced; however, a great deal of work is available so
far only in the form of non-refereed reports and unpublished
work.
We will use SPIRE to illustrate the broad features of
the data. SPIRE was launched on September 2g, 1977 [Na-
!Owu
E4 6

WAVELENGTh
9

(~m)
13 I 7°
~
dile et al:, 1977; Stair et al., 1981, 1983]. This payload
Figure 13-7. Nighttime radiance seen by the long wavelength SPIRE
had three coaligned primary sensors: (1) a telescoped two- spectrometer as function of wavelength at 124 km tangent
channel photometer, (2) a cryogenic (77 K) telescoped short height. NO (1 ~ O) band at 5.3 ~m and C02
(01 101 + 00001) band at 15 ~m are the only reproducible
wavelength infrared (SWIR) CVF spectrometer, and (3) a
spectra seen. Scale for the NO band is 10 times larger than
cryogenic (10 K) telescoped long wavelength iti (LWIR) that for COZ band. (Daytime results were essentially the
CVF spectrometer. The spectrometers were coaligned and spree. )

13-6
 INFRARED AIRGLOW
oxygen atoms shows no diurnal variation. A spectrum ob-
tained with the SPIRE liquid-helium-cooled spectrometer at
124 km tangent height is shown in Figure 13-7. Although
this spectrum is obtained at night, the post-dawn results are
essentially the same. Emission from the 1-Ovibrational band
of NO and the 15 pm (O1101 + 00001) bending mode of
C02 are the only two features seen at this altitude. Detailed
8-

m TANGENT HEIGHT
80 km
NIGHT

C02(U2)
11

d
modeling of these emissions indicates that (1) the emissions 03

originate from the NLTE atmosphere with molecular vibra-

tional temperatures lower than the local translational tem-
peratures, (2) the dominant mechanism for vibrational
pumping is the collisional excitation by oxygen atoms, and 6 13

(3) radiative cooling due to NO emissions is the most im- WAVELENGTH (pm)

portant atmospheric cooling mechanism (-500 K/day) at

these altitudes. Figure 13-9. Nighttime signal for long wavelength SPIRE spectrometer
as function of wavelength at tangent height of 80”km.Notice
appearanceof 01 (VS)signal. The apparent signal in the 4-6
ym region is due to dielectric relaxation response of the
detector to the 15 pm radiation. The 5-6 ~m region has
13.4.2 Lower Thermosphere (between 100 been magnified 10-fold to highlight the NO spectral features.
km and 85 km)

In this region the oxygen atom concentration begins to

decrease without showing any diurnal variation while the 100 km tangent height at night, the V3transition of 03 begins
temperature shows little change. In this region infrared air- to appear in the spectrum between 9 and 11 pm (Figure
glow is characterized by the effect of the solar radiation and 13- 10). At a slightly lower tangent height, this transition
oxygen atoms on the radiating species. In addition to the appears during the dawn observations as well. While the
NO and COZ bands mentioned above, we see emission from 001-000 transition of ozone is centered around 9.6 pm, the
ozone at 9.6 pm (VJ band) and the 4.26 pm v~ band of C02 profile of the emitted radiation appears, in addition, to con-
(Figures 13-8 and 13-9). Because of the large absorption sist of some hot-band transitions. These transitions are not
coefficient of 00011 + 00001 transition at 4.26 pm, this unambiguously identified as yet and are referred to in the
transition is severely self-absorbed. The v~ photons reaching following discussion as the long-wavelength tail of the
the detector most likely originate from the transition between
hot bands and have slightly longer wavelength (around 4.3
pm). The diurnal airglow variations also begin to manifest
120 I I I I I
themselves in the 9.6 pm emission from ozone. At about f?~ ●

I I I I I I
10
WAVELENGTH (pm) 10-8 10-6 10-4

RADIANCE (W cm-2 sr”’ )

Figure 13-8. Daytime signal from the long wavelength SPIRE spectrom-
eter as function of wavelength at tangent height of 80 km. Figure 13-1 (). 9.6 ~m (v,) ozone ( 19.20-10.50 ym) band radiance as
The 15 ~m C02 and 5.3 ~m NO signals are about the same function of tangent height. Note stronger nighttime radiance
as in Figure 13-9 (nighttime). The 4.3 ~m COZ signal is and peak around 80 km. A terminator crossing is evident
enhanced and 9.6 ~m OJ signal is reduced over nighttime in scan 1 (0) where the radiation decreases by an order of
values (Figure 13-9). magnitude from 85 to 90 km.

13-7
CHAPTER 13
(v s) SPECTRAL RADIANCE *
03 ‘ +% j
. %
I I
I 100
m
y>
TAN. HT. = 84 km +~.x

/& 1
X =106deg
+’
75 ~q
* *m
❑ SPATIAL SCAN 1 -%%
50 o SPAT I AL SCAN 2
E
P 76 km A SPAT I AL SCAN 3
“~
+ SPATIAL SCAN 4 x

)// 82 km
78deg
\ 25
x
0
SPATIAL
SPATIAL
SCAN
SCAN
5
6 f
+ SPAT I AL SCAN 7
X SPAT I AL SCAN E i
[ $

!-

9 10 II 12
WAVELENGTH (~m)

Flgurc 13-13. Nighttime integrated 4.3 pm band radiance observed by

Figure 13-11. 9.6 pm (v~) ozone spectmm at 80 km tarsgent height for shortwavelength SPIRE s~trometer. Notice much faster
different solar elevation angles. ‘t’he signal for wavelengths decrease of the signal above 70 km tangent height than for
longer than 10.2 ~m shows no diurnal variation.. the daytime case.

9.6 ~m emission. The night and day long wavelength tion. The COZ 4.3 pm V3band daytime radiance wn clearly
tails are essentially the same (Figure 13- 11) even though in Figure 13-12 is over ten times more intense during the
the total 0~ emission decreases dramatically in the day- day than during the night (Figure 13- 13) in agreement with
time. This appears quite reasonable since the day and the earlier theoretical prediction of James and Kumer [1973].
night oxygen atom concentrations at these altitudes are the As pointed out earlier, the daytime 4.3 pm signal is due to
same and the vibrationally excited OJ formed by the resonance fluorescence as well as nonresonant fluorescence
O + Oz + M ~ Oq + M reaction is the suggested source excited by a shorter wavelength, for example, solar radiation
of the long wavelength V3photon emission. The peak at 9.6 (2.7 pm). Near the top of this layer, the collisions of oxygen
~m results mainly from resonant scattering of earthshine by atoms with C02 and NO also play art important role in 15
the Vq mode of OS and is a measure of OS concentration; and 5.3 pm emissions.
the smaller daytime value of this scattered radiation reflects
the lower daytime OS concentration due to photodissocia-
13.4.3 Upper Mesophere (between 85 km
and 70 km)
I 50

Infrared airglow is primarily determined by processes

125 involving solar radiation during the day (Figure 13- 14) and
the airglow reaction, H + OJ ~ OH + Oz and HOZ
+ O - OH + 02 duringtheslight(Figure 13-15). Large
100

I I I I I I r
75

1
SUNLIT 74 km TANGENT HEIGHT
C02(V3)
50 02 FLUORESCENCE /
t x SPATIAL SCAN 9 SCALE X 10
\ (1.58#ma4ND)

25
E x
x
8
SPATIAL
SPATIAL
SPATIAL
SCAN
SCAN
SCAN
10
11
12
~

G
* i
i
2.7pm
FLUORESCENCE r
t S* :

WAVELENGTH (pm)
(W cm-2 stil )
Figure 13-14. The daytime signal from near infiarcd s~ctrometer at 74
Daytime integrated 4.3 ~m band radiance observed by short km tangent height. The nighttime OH radiation is reduced.
wavelength SPIRE spectrometer as function of tangent height. 02(aAg v = O) ~ ‘Zg- v = 1) radiation appears at 1.58
Larger signal in scan 12 is due to the fact that a deli~rate pm. The 2.7 ~m radiation is fluorescence signal from
effort was made in this last scan to test the telescope. COZ(VI + V3combination bands) and HzO (VI and V3bands).

13-8
 INFRARED AIRGLOW
(Figures 13-8 and 13-9). This emission is due to the tran-
I I 1 1 1 I I
sition from 10001 and 10002 levels to the 01101 level of
TANGENT HT 80h,n
NIGHT
C02; the transition from the third level, that is, the
h OH [AV=2) 02201 ~ 01101 transition at 668.3 cm-’ (compd to 667.3
cm-’ for the 01‘O ~ 000 transition) makes a substantial
contribution to the band radiance but is not resolved from
the 01101 ~ 00001 emission. Also not separated are the
trace isotope emissions from that of the main isotope. Mod-
) 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
WAVELENGTH (~m) eling studies indicate that the contribution of the trace iso-
topes to the radiance at 15 ~m is much larger than expected
Figure 13-15, The nighttime signal from near infrared spectrometer as
on grounds of abundance because of the fact that the emis-
function Notice the progression of OH over-
of wavelength. sion from the O1‘O~ 000 transition of the main isotope is
tone bands (Av = 2) between 1.5–2.4 ~m and the OH self absorbed at tangent height below 110 km.
fundamental bands (Av = 1) peaked at about 3 ~m. The
fluorescent emission at 4.3 pm is shown here without the Around 80 km, the rates for collisional relaxation of
effects of the dielectric relaxation problem encountered in ozone, Oq(OO1) + M ~ 0~(~) + M, become compa-
the long wavelength spectrometer. rable to the Einstein A coefficient for the radiative relaxa-
tion. Rates of collisional relaxation from higher vibrational
diurnal differences in the density of oxygen atoms are pre- levels of ozone are not known. A clear description of the
sent. Collisions of the radiating species with oxygen atoms underlying processes, therefore, cannot be formulated.
become an unimportant mechanism for depopulating the Around 75 km tangent height, the daytime scans show
radiating states because of the small oxygen atom concen- the a’A~ ~ X32~_ (O-1) 02 transition at 1.58 pm; the O-O
tration even in daytime. The collisions of the radiating spe- band of the same electronic transition at 1.27 ~m is outside
cies with Nz are believed important for OH and COZairglows the range of detection and, therefore, not observed. No
at night [Kumer et al., 1978]. signal at 1.58 pm is observed at night. The metastable a
At a tangent height of about 80 km, the emission around state has a radiative lifetime of 3.88 x 103 s. The Franck-
4.3 ~m from C02 emission is about 30 times stronger during Condon integral for the O-1 transition is 0.0131 times that
the day than at night due to absorption of solar radiation of the O-Otransition, making the Einstein coefficient of the
(Figures 13-12 and 13-13). Nighttime 4.26pm fluorescence O-1transition about 1.76 x lti’ s-’ [Krupenie, 1972]. That
is primarily due to energy transfer from vibrationally excited a finite signal is observed is due to the large length over
Nz [Kumer et al., 1978], in turn produced by energy transfer which light is collected and the large densities of Oz(’A~)
from vibrationally hot OH produced by chemical reactions, (about 5 x 109 cm-3). Total ground state 02 density at 74
for example, H + OS - OH + 02. km is 1.7 x 10’4/cc and that in vibrational level 1 of the
In contrast to the strong daytime emission from C02 ground electronic state is about 3 x 109cm-s. We thus see
around 4.3 ~m, a strong night-time airglow is seen from that at 74 km tangent height there is complete optical in-
OH (Figure 13-15). This airglow is composed of many version between the a(v = O) and X(v = 1) states. The
vibration-rotation bands extending over a wide spectral re- gain in the electronic transition is about 108 per km and
gion from visible to 4 pm. Fundamental vibration-rotation assuming a path length of 103 km we get signal amplifi-
bands (Av = 1) and first overtone (Av = 2) bands are ob- cations of e-.ooool.Although the laser action on this transition
served in the infrared [Murphy, 1971] and are shown in does not appear likely, the signal shows a large amount of
Figure 13-11. The overtone hydroxyl emission is broad and energy stored [Oz(a’A~)~ 02(X3Z~), O-l]. The lack of 1.58
appears to be a composite of several vibrational transitions. pm emission at night indicates that the main source of
The fundamental (Av = 1) band is less broad and is peaked Oz(a’Ag) for the region of the atmosphere discussed here is
around 2.9 pm indicating strong contributions from 1-Oand the photodissociation of 03.
2-1 vibrational transitions (Table 13-l). It should be pointed
out that the Einstein A coefficients for OH overtone emission
are larger than those for OH fundamental emission for all 13.4.4 Lower Mesosphere (between 70 km
upper vibrational levels except v = 2 (Table 13-2). The and 40 km)
observations cited above appear reasonable if the collisional
deactivation rates are closer to the larger of the two sets of Solar radiation in the strong absorption bands starts be-
Einstein A coefficients. The daytime hydroxyl (Figure 13- coming depleted as the sunlight traverses this region. In-
14) emissions appear weak. The 1.6 pm region is dominated frared airglow in this region is therefore very dependent on
by Oz(a lA~) emission and the 2.7 pm region by COZ and the LOS and three-dimensional atmospheric calculations are
H20 emission below 70 km. needed to model the radiance in this region. Figure 13-16
The emission around 15 pm continues to show no diur- shows the O-1 band radiance of the electronic transition
nal variations (Figure 13-15). Also, emission at 13.8 and (a’~~ + X3~~-) of 02 as a function of tangent height. It is
16.2 pm in the wings of COZ 15 pm band is observed seen that the larger the solar elevation angle, the greater the

13-9
CHAPTER 13
230 I I i I I I I I I

02 (~g– 3Z~(0.1~l,58~m o ● C02 (V2)

0 ● VIBRATIONAL TEMPERATURE
INTEGRATED BAND RADIANCE
0 ●

-1
I I I I I I o KIO1 = :13cm3/sy
I90 0 ●
o SPATIAL SCAN 9 (13.0°) = 10 ficm /see
o ● ● K[0]
co A SPATIAL SCAN 10(11.3°) ~ 0 ●
I
+ SPATIAL SCAN I I (6.8°) m I o ● I
& SPATIAL SCAN 12( 9.9°)
+XAo x
: !50 o ●

x o ~ 0° : KINETIC TEMPERATURE
+
XA 1- ●
●
-i 0° JACCHIA 77
iAO ●
a
x 0° ● Tm= 1050‘K
+ ●
A
A
AO 110 :0 ●
A
A
[ ● A 1
A ‘o 0 ● “ A
1 00
O*
0*
A
A
a
O* A
00 A
I 1 1 I 1 1 1 1
7920 160 200 240 280 320
1 I I I I I
TEMPERATURE (K)
0.04 0.08 0.12

RADIANCE (KW cm ‘2sr-*) Figure 13-17. Vibrational temperature of bending mode of COZ for two
rate constants for the deactivation by oxygen above. Kinetic
temperature is denoted by triangles.
Figure 13-16 Integrated (1.5 1-1.65 &m) Oz (a’ Ag) radiance at 1.58 ~m
as function of solar elevation angle. Peak signal around 40
km shows inverse correlation with solar elevation (x) angle.
Solar radiation plays an important role in the funda-
mental (Av = 1) emission from the V3mode at 4.3 pm.
Figure 13-18 gives a plot of the vibrational temperature and
1.58 pm radiance. Also in contrast to the higher altitudes
kinetic temperatures [Degges and Smith, 1977] as functions
where nighttime 9.6 pm V3radiance from ozone is larger
of altitude. Figures 13-12 and 13-13 give a plot of the 4.3
than the daytime value, the diurnal variation vanishes in the
pm emission observed by SPIRE as a function of tangent
lower part of this altitude range (below about 50 km).
height during day and night, respectively. me daytime ra-
We again point out that this division of the atmosphere
into four parts for the discussion of the airglow is rather
arbitrary and with a different data base the boundaries be- 1 I 1 1 I I

tween the regions may have to be slightly shifted. However NIGHT DAY
we believe that basic qualitative features of infrared airglow 000 II 000 I I
are well described by this division. This method of pre-
senting the data also stresses the common sources that give KINETIC
(’
rise to airglow in different bands and species.

T
13.4.5 Infrared Radiators

We now present a brief discussion of the airglow clas-

sified according to the radiating species in order to sum-
marize the previous discussion as well as to bring out several
new points.

13.4.5.1 C02. The COZ radiation is seen in the limb-

scanning SPIRE experiment at wavelengths of 15, 4.3, and
2.7 pm. 15pm radiation (Figure 13-5) shows no diurnal
variation and is excited mainly by collisions with oxygen J I I I 1

atoms above 90 km. Below 90 km altitude the main mech- 100 200 300 400
TEMPERATURE (K)
anisms of excitation are the collisions with N2 and 02 with
the earthshine also playing a role. Above 70 km the 15 pm
Figure 13-18. Daytime and nighttime vibrational temperatures of asym-
vibrational temperature lags behind the translational tem-
metric stretch of COZ as ti,mction of altitude. Kinetic tem-
perature, whereas below 70 km LTE forms a good basis for perature is also piotted for comparison [Dcgges and Smith,
describing the 15 pm radiation (Figure 13-17). 1977].

13-10
 a
INFRARED AIRGLOW

100
SUNLIT 2.5-2 9#fn INTEGRATED RADIANCE

$80 o SCAN 9
1- A SCAN 10
❑ SCAN I I
:
❑ 60
r
1-
‘%
~ 40 .. A @
< RAYLEIGH
b ‘~ 0 / SCATTERING
20 ❑ b. 10-’0 IO-8 10-6

“AA ,
\ RADIANCE (W cm ‘2ir”’ )
0
10-8 10-6 Figure 13-21. Integrated 5 .W5.70 ~m NO band radiance as function of
RADIANCE (W cm-2 sf’ ) tartgent height. Night scan 8 gives systematically larger
signals around 100 km.

Figure 13-19. Daytime integrated 2.5–2.9 ~m band COZ and HzO radi-
asice as function of tangent height. Signal helow 30 km is
due to Rayleigh scattering.

diance at around 80 km tangent height is about a factor of ENERGY BUDGET

30 larger thari in nighttime value. Also, Scan 12 was pro- 200[
‘q..
grammed tp look almost at the sun at higher altitude. The
increased sigtlal around 125 km tangent height is due to
telesco~ leakage.
Radiation at 2.7 pm (Figure 13-19) is observed from I NBAND
I ‘~
y~
5.3
NiTRIC
MICRON
RADIANCE
O
BAND
OXIDE

ASCENT
C02 during daytime only. In addition to C02, there is con- NOISE
EQUIVALENT 1 A DESCENT
tribution from H20 in this band [Kumer, 1981; Sharma et i 60 RADIANCE
al., 1981].

SPIRE
I I ! I I
[ I
WATER 6.3 MICRON
T 80 - INTEGRATED
&
k BAND RADIANCE
s
u “+.
z %:=
I ● *%
+
z 40 – ‘ ~?r
LLl
u IN BAND NOISE “y 80
z EQuIvALENT
< RADIANCE ~.k
1-
,*;.
J I I I
n 1 4 -8
“-
-10 -8 10-’0 10
10 10 10-6
BAND RADIANCE (W cm-2 sF’ )
RADIANCE (W cm ‘2sr-’ )
Rgure 13-22. Integrated band radiance for the 5.3 prn NO signal observed
Figure 13-20. 6.3 pm (vl) water band radiance (integrated 5.50-6,74 in the Energy Budget campaign during ascent and descent
pm) as function of tangent height. as a function of tangent height.

13-11
CHAmER 13
I I “T I I
13.4.$.3 NO. The 5.3 pm radiation trom NO in the SPIRE
data (Figure 13-21) &ws no diurnal valuation. Figure
13-22 shows NO radiation from the Energy Budget zenith-
looking rocket as a function of altitude. Above 90 km, the
oxygen atoms play a dominant role in the 5.3 pm radiation
process because of the large rate constant [Degges, 1971;
Fernando and Smith, 1979] for the collisional excitation of
NO. The 5.3 pm radiation from NO is the major cooling
mechanism above 100 km [Kockarts, 1980; Zachor et al.,
i981].

13.4.5.4 03. Ozone emission is observed around 9.6pm

(v, band) and around 4.8 pm (v, + V3combination band).
Ozone emission at 9.6 ~m has been one of the better studied

o I 1 I I I I
I00 200 300 400
TEMPERATURE (K)

Figure 13-23. Daytime andnighttime vibrational tempemtures ofvsmode

ofozone as function ofaltitude [Wggesmd Smitfr, 1977].

13.4.5.2 H20. Radiation from H20 is observed at 6.3

pm (vZbending mode) and dufing the daytime around 2.7
pm (vI and V3modes). The radiation at 6.3 ~m as a function
of tangent height is shown in Figure 13-20. The LOWTMN
4 radiance model [Selby et al., 19781 using 5 ppm HzO
adequately represents the data between 60 and 70 km t~gent
heights. Above 70 km he daytime 2.7 m emission is partly
explained at 1.5 ppm of water in a non-LTE model by
Degg4s (private communication, 1984).

03iv[ + v3) COMBINATION

INTEGRATED BANt) RADIANCE
60

,.
40
NIGHT

20 l’ ,
)/ J’ 1
(“
x RES, 5 cm-1

I ‘ 000 1000 2000 3000 4000

I I J
-lo WAVELENGTH (~m)
10 10-8 10-6
TANGENT HEIGHTS (60, 80, ---n Okm)
RADIANCE (W cm ‘2sr”i)
,
Figure 13-25. Day and night signals predicted using model calculations
Figure 13-24. 4.8 pm (v, + VS)combination band signal for ozone. of T. Degges at SPIRE and 5 cm-’ resolutions.

13-12
 m
INFRARED AIRGLOW
NIGHTTIME - RADIANCE DAYTIME - RADIANCE
200
200 ~
~ H20( v,) 6.lpm ~H20(v2) 6.1pm
?OH(AV =2) 1.9pm .NO(AV=l) 5.3pm
175
●NO(AV=l) 5.3pm
AH N03(v5) 11.4pm \
\
AHN03 (vs) 11.4~m

150 k Ocoz (v~)4.3+rn

*COZ ( vz) 15#m
00H (AV=l) 3,3pm
V03(V3) 9.6Am
125
+03( Vl+V3)4.0pm
A

100

75 :

50

25 :

o ok 1 I I I 4,,,,,,,,~ EJ

, Q-1o ,Q-9 , Q-8 ,~-7

10-6 ,~-9 10* ,Q-lo , Q-s , Q-a , Q-7 , Q-6 ,Q-5 , Q-4

(W cm’2 SF1 ) (W cm-2 sr ‘1)

Figure 13-26. Summary of the nighttime signals observed b, SPIRE. Figure 13-27. Summary of the daytime signals observed by SPIRE.

infrared airglow processes [Rawlins et al., 1981]. Figure The model for hydroxyl emission is complex and involves
13-23 shows the VS vibrational temperature [Degges and a set of several aeronomic reactions of ozone in an atmos-
Smith, 1977] as a function of altitude. Figures 13-10 and phere containing hydrogen (Moreels et al., 1977).
13-24 show the radiance from the 9.6 and 4.8 pm bands In addition to the principal radiators mentioned above,
as functions of tangent height. there are several minor radiators in the altitude range covered
in this article, for example, NOZ, NZO and ~~.
13.4.5.5 02(1Ag). The 0-1 band emission at 1.58 pm has Figure 13-25 gives a summary of the emission in the
been already described in detail. Most of the earlier rocket 2-40 pm regions as computed by the AFGL Limb Model
work has studied the O-O emission [Paulsen et al., 1972]. [Degges and Smith, 1977] at the resolution of the SPIRE
The conclusion of the studies is that the concentration of experiment (AA - O.WA). ~ncipal features discussed above
O*(’ A.) in the atmosphere is 109 – 10” cm-3 depending are seen, that is, enhanced daytime emission at 2.7 and 4.3
on the altitude. pm, enhanced nighttime ‘emission at 9.6 #m, as well as
diurnally independent emission at 5.3 and 15 pm, and long
13.4.5.6 OH. The hydroxyl airglow emission occurs over wavelengths. Also clearly seen is the region (11 to 13 Am)
a relatively narrow altitude region of the mesosphere [Baker characterized by the absence of infrared airglow. Figures
et al., 1977]. The peak emission occurs around 85 km and 13-26 and 13-27 give a summary of the nighttime and day-
the thickness of the emitting layer is about 10km. The origin time radiances observed by SPIRE. At high altitudes, NO
of the hydroxyl emission is believed to be chemilumines- (5.3 pm) is the most intense emission observed while at
cence from two chemical reactions, low altitudes COZ ( 15 pm) and OS (9.6 pm) are the dominant
radiators. It should be noted that in the absence of a data
H+03+OH(VS 9)+02
base in the 17-40 pm, the radiances in Figure 13-25 can
and be regarded as theoretical predictions. ln the light of future
O + HO* ~ OH(v = 6) + 02. work these predictions may have to be modified.

13-13
CHAPTER 13
REFERENCES

Baker, D.J., T. Conley, and A.T. Stair, “On the Altitude Its Application to the OH Emission Problem,” J. Atmos.
of OH Airglow,” (Abstract), Eos Trans. AGU, 58:460, Terr. Phys., 39:551, 1977.
1977. Murphy, R. E., “Infrared emission of Oh in @e Funda-
Clough, S.A., F.X. Kneizys, L.S. Rothman, and W.O. mental and First Overtone Bands ,“ J. Cheti. Phys., 54:
Gallery, “Atmospheric Spectral Transrniqance and Ra- 48524859, 1971.
diance: FASCODEIB,” SPIE, 277:152-166, 1981. Nadile, R. M., N.B. Wheeler, A.T. Stair Jr., D.C. Frod-
Craig, R. A., and J.C. Gille, “Cooling of the Thermosphere sham, and L.C. Wyatt, “SPIRE-Spectral Infrared Ex-
by Atomic’Oxygen,”J. Atmos. Sci.,, 26: 205–209, 1969. @riment” SPIE, X 18-24, 1977.
&gges, T. C., “Vibrationally Excited Ni@c Oxide in the Offerman, D., and K. V. Grossman, “Spectrornetric Meas-
Upper Atmosphere,” ApP1. ~pt., 10:1856, 1971. urement of Atomic Oxygen 63 ~m Emission in the Ther-
Degges, T.C. and H.J.P. Smith, “A High Altitude Infrared mosphere,” Geophys. Res. Lett., 5:387, 1978.
Radiance Model,” AFCRL:TR-77-0271, ADAC)59242,. Ogawa, T., “Excitation Processes of Infrared Atmospheric
1977: Emissions,” Planet. Space Sci., 24: 749–756, 1976.
Fernando, R.P. and I.W.M. Smith, “Vibrational Relaxation Paulsen, D. E., R.E. Huffman, and J.C. Larabee, “lm-
of NO by Atomic Oxygen,” Chem. Phys. Left., 66:218, pmved Photoionization Rates of 02( ‘A~) in the D Re-
1979. ~ion,” Radio Sci., 7:51, 1972.
Gordiets, B.F., Y.N. Kuli~ov, M,N. Markov, pnd M.Ya. Rawlins, W .T., G.E. Caledonia, and J.P. Kennealy, “Ob-
Marov, “The Influence of Radiative C~ling and Tur- servation of Spectrally Resolved Infrared Chemilumi-
bulence on the Heat Budget of the Thermosphere,” (To nepcence from VibrationaJly Excited Q3(M3),*J. Geo-
be published in J. Geophys. Res.). phys. Res., 86:5297, 1981.
Gordiets, B.F., M.N. Markov, and L.A. Shelpen, “IR Ra- Rothman, L.S. and L.D.G. Young, “Infrared Energy Levels
diation of the Upper Atmosphere,” Planet. Space Sci., and Intensities of Carbon Dioxide,” J. Quant. Spectrosc.
26: 933–948, 1978. Radiat. Transfer, 25: 505–524, 1981.
James, T.C. and J.B. Kumer, “~uorescence of C02 near ~othman, L. S., A. Goldman, J.R. Gillis, R.R. Gamache,
4.3 pm: Application to Daytime Limb Radiance Cal- H.M. Pickett, R.L. Poynter, N. Husson, dnd A. Chedin,
culations,” J. Geophys. Res., 78: 8320-8329, 1973. “AFGL Trace Gas Compilation: 1982 Version,” Appl.
Kockarts, G., “Nitric Oxide Cooling in the Terres@al Ther- Opt., 22:1616, 1983a.
mosphere,” Ce~phys. Res. L.ett., 7: 137, 1980. Rothman, L. S., R.R. Germache, A. Bar&, A. Goldman,
Krupenie, P. H., “Spectrum of Molecular Oxygen,” J. Phys. J.R. Gillis, L.R. Brown, R.A. Toth, J.-M. Flaud, and
Chem. Ref. Data, 1: 423–534, 1972. C. Camy-Peyret, “AFGL Atmospheric Absorption Line
Kumer, J.B. ,’“Arlalysis of Non-LTE C02 SPI~ Earthlimb Parameters Compilation: 1982 Edition,’! Appl. Opt., 22:
Data,” AIAA Reprint 81-0425, Aerospace Sciences 2247, 1983b.
Meeting, Jan 1981. Selby, J. E. A., F.X. Kneizys, J.H. Chetwynd, Jr., and R.A.
Kumer, J. B., ‘A.T. Stair Jr., N. Wheeler, K.D. Baker, and McClatchey, “Atmospheric Transmittance/Radiance:
D.J. Baker, “Evidence for an OH$”” ~ N~”””- Computer Code LOWTRAN 4,” AFGL TR-78-0053,
CO; - CO* + hY(4.3 pm) Mechanism for 4.3 pm ADA058643, 1978.
Airglow,” J. Geophys. Res., 83:47434747, 1978. Sharnia, R. D., “Near-Resonant Vibrational Energy Transfer
Markov, M. N., “Molecular Radiation of the Upper At- Among Isotopes of C02,” Phys. Res., 177:102, 1969.
mosphere in the 3–8 Km Spectral Region,” Appl. Opt., Sharrna, R.L3., and C.A. Brau, “Energy Transfq~ in Near-
8:887, 1969. Resonant Molecular Collisions due to Long-tige Forces
Markov, M.N., G.M. Grechko, A.A. Gubarev, Yu.S. Ivti- with Application to Transfer of Vibrational Energy from
nov, and V. $. Petrov, “Infrared Spectrum QfNitric Ox- V3Mode of C02 to Nz,” J. Chem. Phys., ,50:924, 1969.
ide Obtained in the Upper Atmosphere at Middle Lat- Shwa, R.D. and R.M. Nadile, “Car@n Dioxide (vz) Ra-
itudes by the Orbiting Scientific Station Salyut 4:’ Cosmic diance Results Using a New Nonequilibriurn Model,”
Res., 15:102,1977. AIAA Reprint 81-0426, Aerospace, Sciences Meeting,
McClatchey, R. A., W.S. Benedict, S.A. C1ough, D.E. 1981.
Burch, R.E. Calfee, K. Fox, L.S. RothmM, ‘md J.S. Sharrna, R. D., R.M. Nadile, A.T. StairJr.l and W. Gallery,
Garing, “AFCRL Atmosphe~c Absorption Line Param- “mimb Emission Analysis of SFtml ~ Rocket
eters Compilation ,“ AFCRL-T-73-0096, AD762904, Experiment (SPIRE) Data at 2.7pm Mimmeters,” SP[E,
1973. 304:139, 1981.
McEwan, M;J. and L.F. Phillips, Che~istry of the Atmo- Stair, A. T., Jr., J.C. Ulwick, D.J. Baker, C.L. Wyatt, and
sphere, The rate constants cited in the book for the K.D. Baker, “Altitude Profiles of Infrared Radiance of
deactivation of 02( lAE)state by Nz, Oz and O tie s 3 Os (9.6 pm) and C02 (15 pm),” Geophys. Res. Left.,
x 10-21,2 x l~]x and < 3 x 10-lbcmrnolecule-lsl, 1:117,1974.
respectively, Wiley, New York, p. 140, 1975. Stair, A.T. Jr., J.C. Ulwick, K.D. Baker, and D.J. Baker,
Mies, F. M., “Calculated Vibrational Transition Probabili- “Rockethome Observations of Atmospheric Infrared
ties of OH (x*m),” J. Mol. Spec., 53: 150, 1974. Emissions in the Auroral Region,” in Atmosphere of
Moreels, G., G. Megie, A. Vallance Jones, and R.L. Gat- Earth and the P~anets, edited by B .M. McCormac, D.
tinger, “An Oxygen-Hydrogen Atmospheric Model and Reidel, Dordrecht, Holland, p. 335-346, 1975.

13-14
 INFRARED AIRGLOW
Stair, A.T. Jr., R.M. Nadile, J.C. Ulwick, K.D. Baker, Taylor, R. L., “Energy Transfer Processes in the Strato-
and D.J. Baker, “Infrared Measurements of Auro- sphere, “ Can. J. Chem., 52:1436-1451, 1974.
ral, Airglow and the Upper Atmosphere,” AIAA Re- Ulwick, J. C., K.D. Baker, A.T. Stair Jr., W. Frings, R.
print 81 -W21, Aerospace Sciences Meeting, Jan Hennig, K.V. Grossman, and E.R. Hegblom, “Rock-
1981. etbome Measurements of Atmospheric Infrared Fluxes,”
Stair, A.T. Jr., R.D. Sharma, R.M. Nadile, D.J. Baker, J. Atmos. Terr. Phys., In Press, 1985.
and W .F. Grieder, “Observations of Limb Radiance with Zachor, A. S., R.D. Sharma, R.M. Nadile, and A.T. Stair
Cryogenic Spectral Infrared Rocket Experiment (SPIRE),” Jr., “Inversion of SPIRE NO Data,” AFGL TR-8 1-0325,
J. Geophys. Res., In Press, 1985. ADA108866, 1981.

13-15
 Chapter 14

Standard and Reference Atmospheres

K.S.W. Champion, A.E. Cole, and A.J. Kantor

14.1 STANDARD ATMOSPHERE (ICAO) Standard Atmosphere. The mean annual midlatitude
vertical temperature profile for the portion between 32 and
The World MeteorologicalOrganization (WMO) has de- 55 km was developed from both routine radiosonde obser-
fined the Standard Atmosphere as follows: vations and meteorological rocketsonde observations. The
“A hypothetical vertical distribution of atmospheric tem- portion between 55 and 86 km is based primarily on mea-
perature, pressure and density which by international agree- surements derived from grenade, pitot static tube and falling
ment and for historical reasons, is roughly representative of sphere experiments.
year-round, midlatitude conditions. Typical usages are as a The temperature-height profile between 86 and 1000 km
basis for pressure altimeter calibrations, aircraft perform- is expressed in geometric altitude. It is not a series of linear
ance calculations, aircraft and rocket design, ballistic tables functions as it is at altitudes below 86 km. Rather, it is
and meteorological diagrams. The air is assumed to obey defined in terms of four successive functions chosen not
the perfect gas law and the hydrostatic equation which, taken only to provide a reasonable approximation to observations,
together, relate temperature, pressure, and density with geo- but also to yield a continuous first derivative with respect
potential. Only one standard atmosphere should be specified to altitude over the entire regime. Observational data of
at a particular time and this standard atmosphere must not various kinds provided the basis for determining the four
be subjected to amendment except at intervals of many segments of the temperature-height profile between 86 and
years.” 1000 km. The observed temperatures at altitudes between
Because of the interest of aerospace engineers and at- 110 and 120 km were particularly important in imposing
mospheric scientists in conditions at much higher altitudes limits on the temperature-height structure for that region,
than those currently being considered by the WMO, mem- while the observed density at 150 km and above strongly
bers of the U.S. Committee on Extension to the Standard influenced the selection of both the temperature and the
Atmosphere (COESA) agreed to add the following para- extent of the low-temperature isothermal layer immediately
graph to the above definition in describing the U.S. Standard above 86 km. Values of the thermodynamic properties for
Atmosphere, 1976, which extends to 1000 km: the region between 140 and 1000 km were determined al-
“The atmosphere shall also be considered to rotate with most exclusively from satellite related observations and ra-
the earth, and be an average over the diurnal cycle, semi- dar incoherent scatter techniques. For altitudes between 90
annual variation, and the range of conditions from active to and 140 km, however, there is only a very limited amount
quiet geomagnetic, and active to quiet sunspot conditions. of data from rocket soundings and incoherent scatter ob-
Above the turbopause (about 110 km) generalized forms of servations, and almost none from satellite observations.
the hydrostatic equations apply .“ The 1976 Standard, like its predecessors, uses the fol-
lowing sea-level values of temperature, pressure, and den-
sity which have been standard for decades:
14.1.1 U.S. Standard Atmosphere, 1976
Temperature 288.15 K, 15°C or 59°F
The U.S. Standard Atmosphere, 1976 is an idealized,
steady-state representation of mean annual conditions of the Pressure 1013.25 mb, 760 mm of Hg, or 29.92 in.
earth’s atmosphere from the surface to 1000 km at latitude of Hg
45”N, as it is assumed to exist during a period with moderate
solar activity. The defining meteorological elements are sea- Density 1225.00 g/cm3 or 0.076474 lbs/ft3.
level temperature and pressure, and a temperature-height
profile to 1000 km. The latter includes a linearly segmented In spite of the various independent data sets upon which the
temperature-height profile in geopotential heights between several temperature-height segments are based, it is desir-
the surface and 86 km. The first 32 km portion of this profile able, for purposes of mathematical reproducibility of the
is identical to the International Civil Aviation Organizatiorl tables of the 1976 Standard, to express the temperature in

14-1
CHAPTER 14
a series of consecutive height functions from the surface to port of individual gas species lead to the need for a dynam-
1000 km, with the expression for each successive function ically oriented model including diffusive separation. Under
depending upon the end-point value of the preceding func- these conditions, Colegrove et al. [1965] found it convenient
tion as well as upon certain terms and coefficients peculiar to express the variations with altitude of the atmospheric
to the related altitude interval. This total temperature-height number density in terms of the vertical’ component of the
profile applied to the fundamental continuity models (hy- flux of the molecules of individual gas species. In terms of
drostatic equation and the equation of motion) together with the ith gas species, this expression is
all the ancillary required constants, coefficients, and func-
tions, defines the U.S. Standard Atmosphere, 1976. dni ni “ (1 + ai) .—
dT +g”ni” Mi

14.1.1.1 Equilibrium Assumptions. The air is assumed

ni”vi+Di.
( —+
dz T dz R*. T
)

to be dry, and at altitudes below 86 km the atmosphere is dni ni dT g.ni” M DO

+K. ~+–”—+
TdZ R*. T
assumed to be homogeneously mixed with a relative volume ( )
composition leading to a constant mean molecular weight (14.6)
MO. The air is treated as if it were a perfect gas, and total
pressure P, temperature T, and total density p at any point where
in the atmosphere are related by the equation of state, that
is, the perfect gas law, one form of which is ni = number density of the ith gas species,

vi = the vertical transport velocity of the ith species,

~=p. R*. T
(14.1)
MO ‘ Di = the altitude-dependent, molecular-diffusion

where R* is the universal gas constant. An alternate form coefficient of the ith species diffusing
of the equation of state in terms of the total number density through N2,
N and the Avogadro constant N~ is
ai = the thermal-diffusion coefficient of the ith species.
N- R*” T
P= (14.2) Mi = the molecular weight of the ith species,
N~ “
M= the molecular weight of the gas through which the
This form represents the summation of Pi, the partial pres-
the ith species is diffusing,
sures of the individual gas species, where Pi is related to ni
the number density of the ith gas species in the following
and
expression:
K = the altitude-dependent, eddy-diffusion coefficient.
Pi=ni”k” T, (14.3)

where k is the Boltzmann constant. 14.1.1.2 Atmospheric Properties. Because of limited

Within the altitude region of complete mixing, the at- space, the 1976 Standard Atmosphere Tables are not in-
mosphere is assumed to be in hydrostatic equilibrium and cluded in this handbook. For detailed tables consult U.S.
to be horizontally stratified so that dP, the differential of Standard Atmosphere, 1976 [COESA]. The temperature-
pressure, is related to dZ, the differential of geometric al- height profile of the 1976 Standard for altitudes up to 86
titude, by the relationship km is shown in Figure 14-1. The profile is the same as the
1962 Standard for altitudes up to 51 km. From 51 to 70 km
dP=–g”p”dZ, (14.4) it is colder, and from 70 to 86 km it is warmer than the
1962 Standard. Temperature profiles to 500 km for the 1962
where g is the altitude-dependent acceleration of gravit y. and 1976 standard and pressure and density profiles from
The elimination of p between Equations (14.1) and (14.4) the surface to 1000 km for the 1976 Standard are presented
yields another form of the hydrostatic equation, which serves in Figures 14-2 and 14-3. The equations used for computing
as the basis for the low-altitude pressure calculation: other atmospheric properties tabulated in the 1976 Standard
are discussed below.
–g. MO. dz
dln P=~= (14.5)
R*. T “ Mean Particle Speed
The mean particle speed V is the arithmetic average of
Above 86 km the hydrostatic equilibrium of the atmo- the speeds of all air particles in the volume element being
sphere gradually breaks down as diffusion and vertical trans- considered. All particles are considered to be neutral. For

14-2
 i

STANDARD AND REPERENCE ATMOSPHERES

, I , I I I I I DENSITY (kg/m’)
1/’ 10
-14
10
-12
10
-10
10-’ 10
-s
10
-4
la
-2
10°
THERMOSPHERE 1
I 000 r.~ I T I I I

II :
I

960 -’l
I
\
800 - \
\
\
~ 700 -

20 200
t
I 00
10 t

o L, o-e ,0-6 ,o-4 ,’o-2 ,02 ,04 ,06

I QrJ
1s0 200 220 24il 260 260
PRESSURE (N/m2)
TEMPERATURE (K)

Flgrs~ 14-3. Total pressure andmass densi~as afUnction ofgeometic

Figure 14-1. Tempratu~-height profile f~r U.S. Stasidard Atmosphere altitude.
O-86 km.

500 I 1 1 I 1 I I avalidaverage towcur,t here must be a sufficient number

ofp~icles involved torep~sent mean conditions. Rssure
I .:
I
fidtemperatqre gradients within the volume must M neg-
i ligible. The analytical expression for V is closely related to
I
400 - that for the speed of sound and is pro~rtional to the ratio
I
I of T/M.
~ I
/
d
/’
~= [8;R*MT],2 (14.7)
$ 300 -
1- 1962~
~
t’
a ! where M is the mean molecular weight at a specific Altitude.
~ /’ A graph of V as a function of geometric altitude is given
g 200 /
/
/ in Figure 14-4.
11.1
z /
-/
a ---
w ---
a Mean Collision Frequency
100 he mean collisioq frequency, V(S-l), between the hy-
t pothetical mean neutral particles is a function of pressure
P, temperature T, and molecular weight M, as well as of
:( an effective mean collision dipmeter U. The expression used
0I lx 1 I 1 I I I
for computing this proprty !s
o --=
7nn ---
400 ---
600 800 1000 1200 1400

KINETIC TEMPERATURE (K)

Figure 14-2. Kinetic temperature as a function of geome~c altitude. ‘=4u2NAp(Mi*T) (14.8)

14-3
CHAPTER 14
1 I I 1 where m = the ratio of the circumference to the diameter
of a circle, that is 3.14159 (dimensionless) and u = the ef-
feetive collision diameter, 3.65 X 10’Om. Figure 14-5 shows
the variation in mean collision frequency with altitude.

Mean Free Path

The mean free path L is equal to the ratio of mean particle
speed to mean collision frequency,

L=Y” (14.9)
v

This quantity represents the average travel distance between

collisions of the hypothetical mean neutral particles in any
direction along which the total change of number density is
small over the mean free path length. Thus at the altitude
of 130 km where L is less than 10 m, the concept is ap-
plicable in all directions. At altitudes above 175 km, where
L is the order of 100 m or greater, the concept of mean free
path, at best, becomes blurred and must be used with cau-
tion. The values of L versus altitude are plotted in Figure
500 I000 1500 2000
14-6.

MEAN PARTICLE SPEED ( m/s)

Mean Molecular Weight
The mean molecular weight M of a mixture of gases is
Figure 14-4. Mean particle speed as a function of geometric altitude. by definition

~(ni Mi)

(14.10)
1000 T I I 1 1 1 1 M= ~Ili

900 -
1000

800 -
900

700 - 800

7
z 600 - ; 700
x
:
z g 600
500 -
~
a ~
~ 500
; 400 - a
L
z : 400
o 1-
w 300 -
m
2
0 300
u
200 -
w

200

I 00 -
100

0 1 I I I I I
10 0
10-4 10-2 10° IO* 10’ 106 108 10
10-e 10-6 10-4 10-2 10° 102 10’ 106 106
COLLISION FREQUENCY (s-’) MEAN FREE PATH (m)

Figure 14-5. Collision frequency as a function of geometric altitude. Figure 14-6. Mean free path as a function of geometric altitude.

14-4
 STANDARD AND REFERENCE ATMOSPHERES

1000 I I , I I I

900

Soo

700

?
5 600
w
o
: 500
r
2
v 400
E
L
z
o 300
w
w
200

I 00
t 1 1 1 I 1

I I 1 I t I
2T0 280 290 300 310 320 330 to
c
2 6 10 14 18 22 26 30 SPEED OF SOUND (m/s)

MOLECULAR WEIGHT (kg/kmol)

Figure 14-8. Sped of sound as a function of geometric altitude.

Figure 14-7. Mean molecular weight asafunction ofgeometicdti@de.

cessively lower frequencies as atmospheric pressure de-

creases or as the mean free path increases. For this reason,
where ni and Mi are &e number density and molecular
the concept of speed of sound (except for frequencies ap-
weight, respectively, of the ith gas s@cies. In that part of
proaching zero) progressively loses its range of applicability
the atmosphere between the sufiace and 86 km altitude,
at high altitudes. Hence the main taples listing the values
mixing is dominant, and the. effect of diffusion and pho-
of S.@ed of sound are temiinated at 86 km. The variation
tochemical processes upon M is negligible. Consequently;
of the speed of sound with altitude is shown in Figure
Mrernains constant at itssea level value, M. = 28.9644
14-8.
kg/kmol. The variation of the mean molecular weight with
altitude is shown in Figure 14-7.
Dynamic Viscosiq
The coefficient of dynamic viscosity P(N s/m2) is de-
Speed of Sound fined as a coefficient of internal friction developed where
The expression adopted for the speed of sound C, is gas regions move adjacent to each other at different veloc-
ities:
~ . T3/’

CS=(Y”H”T)”2 (14.11) ‘=T+S

(14.12)

where ~ is a constant equal to 1.458 X 10-6 kgs-]m-] K -”2

where y is the ratio of specific heat of air at constant Pressure and S is Sutherland’s constant, equal to 110.4 K. The equa-
to that at constant volume and ii taken to be 1.40 exact tion fails for conditions of very high and very low temper-
(dimensionless). The above equation for speed of sound atures and under conditions occurring at great altitudes.
applies only when the sound wave is a small perturbation Variation of dynamic viscosity with altitude is shown graph-
on the ambient condition. Calculated values for c, hive ically in Figure 14-9.
been found to vary slightly from experimentally determined
values. The limitation of the concept of speed of sound due Kinematic Viscosity
to extreme attenuation is also of concern. The attenuation Kinematic viscosity q is defined as the ratio of the dy-
that exists at sea level for high frequencies applies to suc- namic viscosity of a gas to the density of that gas, p,

14-5
 I

CHAPTER 14
I I I I I I 90 , I 1 , 1 1

I 1 I i 1

1.2 1.3 1.4 1.5 1.6 1.7 1.8

DYNAMIC VISCOSITY (IO-* N- S/h*)

FIguti 14-9. Dynamic viscosity as a function of geometric altitude. 10-5 10-4 10-3 10-2 10-’ 10° 10’
KINEMATIC VISCOSITY (m2/s)

Figure 14-10. Kinematic viscosity as a function of geometric altitude.

(14.13)

monatomic gases agree well with observation. The vtiation

~imitations of this equation are comparable to those asso-
of the coefficient of the~al conductivity with altitude is
ciated with dynamic viscosity. A graphical representation
shown in Figure 14-11.
of the variation of kinematic viscosity with altitude is shown
in Figure 14-10.
14.1.1.3 Atmospheric Structure. Tetiperatu~ varia-
tion with. altitude is the most pronounced fea~re df the
Coeflcient of Therniul Conductivity earth’s atmosphere and the various layers and shells can be
The empirical expression adopted for developing tabular descrikd mQre easily by tempe~ture than by any other
values of the coefficient of thermal conductivity kt for al- element. This featuro is illustrated in Figure 14-1 with the
titudes up to 86 km is as follows: tem~ra~re height profile for the 1976 Standard up to 86
km. Usually; temperature decreases witi increasing altitude
in the lowermost several kilometers. Except for the first
2.65Q19 x 10-3 T3/2
~t = (14.14) kilometer above the earth’s surface, the normal lapse rate
T“+ 245.4 X “lO-(]n) “ is about 6.5 K/km in the troposphere, which extends from
the surface to the tropopause.
This expression differs from that used in the U.S. Standard The tropopause is the atmospheric surface at which the
Atmosphere, 1962 in that the numerical constant has been tem~rature decrease with altitude either stops abruptly or
adjusted to accommodate a conversion of the related energy the lapse fate becomes less than some arbitrary low value.
unit from the temperature-dependerit kilogram calorie to the The height of the tropopause varies with latitude, season,
invariant joule. Thus, the values of kt in units of J m-ls-lK-l and weather situation; in general it is lowest (8 to 10 km)
or W rrt-lK-l are greater than the values of k~ in units of in arctic regions in winter, and highest (16 to 18 km) in the
kcdl m-ls-lK-l by a factor of exactly 4.19002 X 103. Ki- tropical and equatorial regions. Above the tropopause, tem-
netic theory determinations of therinal conductivity of some peratures increase with altitude, slowly at first, then more

14-6
 STANDARD AND REFERENCE ATMOSPHERES
90, I 1 I I I I I I I 14.2 REFERENCE ATMOSPHERES

00 - 14.2.1 Air Force Reference Atmospheres for

Altitudes Below 90 Km
70 -
z The Reference Atmospheres presented in this section
x
- 60 - [Cole and Kantor, 1978] were developed to provide Air
u
n Force engineers and designers of aerospace systems with
2 up to date information on the seasonal, latitudinal, and lon-
~ 50 -
gitudinal variations in the vertical distributions of the ther-
a
modynamic properties (pressure, temperature, and density)
v
40 - of the atmosphere in the Northern Hemisphere for altitudes
E
!-
up to 90 km. Special models are included to illustrate the
:
0 30 - magnitude of the changes in the vertical distributions of
z
0 these atmospheric properties during extreme winter warrn-
LIJ
w ings and toolings of the stratosphere and mesophere in arctic
20 -
and subarctic regions.

10 - 14.2.1.1 Basic Assumptions and Formulas. The at-

mospheric models are defined by temperature-altitude pro-
I 1 I 1 I I I I files in which temperatures change linearly with respect to
0
.017 .019 .021 .023 .025 geopotential altitude. It is assumed that the air is dry, is in
hydrostatic equilibrium, and behaves as a perfect gas. The
COEFFICIENT OF THERMAL CONDUCTIVITY W/(m. k)
molecular weight of air at sea level, 28.9644 kg/kmol, is
assumed to be constant to 90 km. Actually, dissociation of
Figure 14-11. Coefficient of thermal conductivity asa function of geo-
metric altitude.
molecular oxygen begins to take place near 80 km and
molecular weight starts decreasing slowly with height. Con-
sequently, the temperatures given in the tables for altitudes
above 80 km are slightly but not significantly larger (less
rapidly up to about 50 km, where the average temperature than one degree K) than the ambient kinetic temperature.
is within a few degrees of the freezing point. This surface Numerical values for the various thermodynamic and
of maximum temperature is designated the stratopause, and physical constants used in computing the tables of atmos-
the entire layer below it down to the tropopause is the pheric properties for these Reference Atmospheres are iden-
stratosphere. tical to those used in the preparation of the U.S. Stadard
The layer above the stratopause in which the temperature Atmosphere, 1976, with two exceptions. Surface conditions
normally decreases with increasing altitude is called the for the atmospheres are based on mean monthly sea-level
mesosphere. It ends at the mesopause at 80 to 90 km, where values of pressure and temperature for the appropriate lat-
the lowest temperatures of the atmosphere are found (135 itude rather than on standard conditions. The accelerations
to 190 K). Above the mesopause lies the thermosphere; its due to gravity at sea level for the latitudes were obtained
upper limit varies with solar and geophysical conditions. from the following expression by Lambert [List, 1968] in
The temperature in the thermosphere increases rapidly with which gravity, g, varies with latitude $:
altitude up to about 200 km, then, more slowly, becoming
nearly isothermal above 300 to 500 km. The temperature gb =
above 200 km varies widely according to the degree of solar
9.780356 (1 + 0.0052885 sin’ @ – 0.0000059 sin’ 2+).
activity; it is about 600 K when the sun is quiet and possibly
(14.15)
2000 K during sunspot maxima. The exosphere, sometimes
called region of escape, is the outermost, or topmost portion
The relationship between geopotential altitude and geo-
of the atmosphere. Its lower boundary is the critical level
metric altitude is the same as that used for the U.S. Standard
of escape, estimated to be 500 to 1000 km above the earth’s
Atmosphere, 1976:
surface. The air density is so low that the mean free path
of individual particles depends upon their direction with
respect to the local vertical, being greatest for upward mov-
ing particles. It is only from the exosphere that atmospheric
H=($.)(2)9 (14.16)

gases can escape into outer space. In this region temperature where H is the geopotential altitude in geopotential meters
can no longer be defined in the usual way. (m’), Z is the geometric altitude, r+ is the effective earth

14-7
 1

CHAPTER 14
Table 14-1, Sea level acceleration of gravity and the effective radius, r6, based on Meteorological Rocket Network (MRN) observa-
of the earth for each latitude,
tions taken during the period 1964 through 1976 and tem-
Latitude Sea Level Gravity Effective Earth Radius peratures from sets of 5, 2, and 0.4 mb constant pressure
r$, (km) maps prepared by the Upper Air Branch of the National
+ g+ (~s’)
Weather Service for the years 1964 through 1968 and from
o 9.78036 6334.984 January 1972 through June 1974. The temperature distri-
15 9.78381 6337.838 butions between 55 and 90 km are based primarily on tem-
30 9.79324 6345.653 peratures derived from grenade, falling sphere, and pressure
45 9.80665 6356.766 gauge experiments. A detailed description of the analysis
60 9.81911 6367.103 of these data and the development of the monthly models
75 9.82860 6374.972 is contained in Air Force Surveys in Geophysics No. 382,
90 9.83208 6377.862 Air Force Reference Atmospheres [Cole, 1978].
14.2.1.3 Seasonal and Latitudinal Models. Sets of mean
radius, g+ is the sea-level value for acceleration of gravity monthly Reference Atmospheres for altitudes UD to 90 km.
at a specific latitude @ as given by Lambert’s equation, and [Cole, 1978] have been d;veloped for 15° inte~als of lati:
G is the unit geopotential set equal to 9.80665 mz S-2 (m’)-]. tude, including the equator and north pole, to provide infor-
Values of r+ and g+ are given in Table 14-1. mation on the seasonal and latitudinal variations in the ther-
modynamic properties of the atmosphere. Due to limitations
14.2.1.2 Data. Initial sea-level pressures for each at- on the length of individual presentations in the handbook, ody
mosphere were taken from mean monthly sea-level charts the properties of the January, April, July, and October models
for the northern hemisphere. Temperature-height profiles are presented by geometric altitudes in Table 14-2.
for altitudes up to 30 km were obtained for specific latitudes The distribution of atmospheric temperature between the
and months by giving equal weight to radiosonde temper- equator and north pole, based on the January and July Ref-
atures interpolated for each 10° of longitude. erence Atmospheres for latitudes 0°, 15°, 30°, 45°, 60°, 75°
Temperature distributions between 30 and 55 km are and 9@N, is shown in Figure 14-12 for geometric altitudes

I I I I I I I

m m

I I 1 I I I I 1 I I I 1 I 1 I
-m w 4W w 15= EQ 15* w e“ w w w
LAllTU~

Figure 14-12. Latitudinal temperature-height cross-sections of monthly temperatures for January and July.
 1

STANDARD AND REFERENCE ATMOSPHERES

Table 14-2a. Reference Atmosphere Temp (K+JaMIw

Alt (km) 0° 15”N 30”N 45°N 60”N 75”N 90°N

O.000 299.15 296.65 287.15 272.15 257.15 248,15 237.15

2.000 288.78 287.27 281.16 265.15 255.94 251.37 245.88
4.000 278.41 277.90 268.20 255.66 247.73 240.36 236.61
6.000 268.06 266.28 255.24 243.68 234.13 229.35 226.60
8.000 256.40 252.35 242.28 231.71 220.54 218.35 216.59

10.OOO 240.49 238.42 229.34 219.74 217.15 214.14 212.64

12.000 224.58 224.50 216.40 218.65 217.15 212.14 210.64
14.000 208.69 210.59 211.08 217.66 217.15 210.14 208.63
16.000 197.34 199.46 205.91 216.67 216.66 208.15 204.64
18.000 195.03 195.15 203.15 215.67 215.66 206.16 200.65

20.000 201.37 203.65 207.73 215.15 214.66 204.16 196.67

22.000 207.71 212.57 212.49 215.15 213.67 202.93 196.65
24.000 214.05 217.22 217.25 215.15 212.68 204.12 196.65
26.000 220.38 221.58 221.79 215.15 212.98 205.32 196.65
28.000 226.71 225.93 225.36 217.08 214.77 206.51 196.65

30.000 232.70 230.28 228.92 221.44 216.55 209.00 203.04

32.000 236.65 234.63 233.13 225.79 218.34 212.97 209.59
34.000 240.60 238.97 237.67 230.15 220.12 216.94 216.14
36.000 244.55 244.29 242.24 235.54 222.07 220.90 222.69
38.000 249.66 250.20 246.98 241.47 224.25 224.86 229.23

40.000 254.98 256.12 251.71 247.40 226.43 228.82 232.98

42.000 260.30 262.02 256.45 253.32 232.87 232.78 236.55
44.000 265.61 265.70 261.17 259.24 239.78 236.74 240.11
46.000 270.92 268.84 265.65 262.82 245.16 240.69 243.67
48.000 272.15 271.15 265.65 264.65 247.93 244.64 247.22

50.000 272.15 271.15 265.65 264.65 250.69 248.59 250.78

52.000 269.12 268.95 261.73 261.41 251.15 252.53 252.15
54.000 265.01 264.43 256.62 255.51 251.15 255.15 252.15
56.000 260.71 259.92 251.65 250.83 250.93 254.06
58.000 255.61 254.49 247.13 247.29 250.54 250.72

60.000 250.52 249.01 242.62 243.76 250.11 247.37

62.000 245.44 243.53 238.12 240.22 245.59 244.03
64.000 240.35 238.05 233.61 236.69 241.08 240.69
66.000 235.27 232.58 229.11 233.17 238.65 238.52
68.000 230.19 226.62 224.62 229.64 238.65 240.09

70.000 225.12 228.57 220.12 226.12 238.65 241.66

72.000 217.25 214.88 218.15 225.65 238.07 239.43
74.000 206.72 210.20 218.15 225.65 234.16 235.52
76.000 196.20 205.53 216.35 223.44 230.25 231.60
78.000 195.65 200.85 212.84 219.54 226.34 227.69

80.000 195.65 198.15 209.33 215.63 222.43 223.77

82.000 195.65 198.15 205.83 212.50 218.53 219.87
84.000 195.65 198.15 202.32 210.16 214.65 215.96
86.000 194.59 196.97 198.82 207.82 214.65 214.15
88.000 191.29 193.48 195.33 205.49 214.65 214.15

90.000 187.99 189.99 191.83 203.15 214.65 214.15

14-9
CHAPTER 14
Table 14-2b. Reference Atmosphere Temp (K>April

~o
Alt (km) 15°N 30”N 45°N 60°N 75”N 90”N

O.000 300.15 297.15 292.15 279.15 269.15 255.15 248.65

2.000 288.98 286.97 282.16 273.15 263.14 254.88 252.12
4.000 277.82 276.81 272.19 263.66 257.13 244.86 242.11
6.000 266.67 264.78 260.24 250.68 243.13 233.85 232.10
8.000 254.63 250.85 246.29 237.71 229.14 222.85 222.09

10.OOO 239.71 236.92 232.35 224.75 222.15 222.56 221.91

12.000 224.81 223.00 218.42 218.15 222.15 224.15 224.15
14.000 209.91 209.09 213.47 218.15 222.15 224.15 224.15
16.000 198.56 200.02 208.69 218.15 222.15 224.15 224.15
18.000 196.12 196.15 206.15 218.15 222.15 224.15 224.15

20.000 202.66 204.84 210.92 218.15 222.15 224.15 224.15

22.000 209.20 213.96 215.88 218.15 222.15 224.15 224.15
24.000 215.74 221.02 220.46 220.26 222.15 224.15 224.66
26.000 222.27 225.38 225.02 223.24 223.17 224.15 226.85
28.000 228.39 229.73 229.57 226.21 225.35 224.15 229.04

30.000 233.93 234.08 234.13 229.54 227.53 227.61 231.23

32.000 239.46 238.43 238.68 234.49 231.97 231.97 233.41
34.000 244.98 242.77 243.05 239.44 236.53 234.76 236.16
36.000 250.51 247.24 247.39 244.54 241.59 238.33 241.12
38.000 256.03 252.77 251.73 250.47 247.33 241.89 246.08

40.000 261.54 258.28 256.07 256.40 253.07 248.03 251.03

42.000 267.06 263.80 260.41 261.09 258.80 254.36 255.98
44.000 269.72 268.24 263.77 265.04 264.53 260.69 260.93
46.000 271.69 270.21 266.92 268.08 269.24 267.02 265.87
48.000 272.15 272.15 269.15 271.65 272.20 273.34 270.15

50.000 272.15 272.15 269.15 271.65 274.15 274.15 270.15

52.000 269.27 269.52 267.98 269.06 274.15 274.15 270.15
54.000 265.35 265.99 263.46 264.34 270.80 272.01 270.15
56.W 261.42 261.78 258.95 259.62 264.90 265.71
58.000 256.93 254.72 253.74 254.31 259.00 259.41

60.000 249.49 247.67 248.25 249.01 253.10 253.11

62.000 242.05 240.62 242.76 243.71 247.21 246.82
64.000 234.62 233.58 237.28 238.42 241.31 240.53
66.000 227.20 226.54 231.80 233.13 236.24 234.24
68.000 219.76 219.51 226.20 229.09 232.12 230.47

70.000 211.95 212.48 220.53 225.53 228.01 226.75

72.000 204.15 205.86 214.87 221.26 223.90 223.02
74.000 196.35 199.62 209.21 217.35 219.79 219.30
76.000 188.55 193.39 203.55 213.44 215.33 215.58
78.000 190.73 193.15 200.17 209.54 210.25 209.96

Bo.000 193.45 193.15 199.20 205.63 205.17 204.09

B2.000 196.40 195.44 198.22 201.11 200.09 198.23
B4.000 199.70 197.97 197.25 196.13 195.02 192.37
B6.000 203.00 200.49 196.28 191.27 189.95 186.51
B8.000 206.30 203.02 195.31 189.65 184.88 180.66

90.000 209.60 205.54 194.34 189.65 179.81 176.15

14-10
 I

STANDARD AND REFERENCE ATMOSPHERES

Table 14-2c. Reference Atmosphere Temp (K)--Ju1Y

Alt (km) 0° 15”N 30”N 45°N 60°N 75°N 90°N

0.00s) 298.65 299.65 301.15 294.15 287.15 275.65 273.15

2.000 287.68 288.87 288.16 285;15 276.34 273.62 271.12
4.000 276.72 278.11 277.19 273.16 265.53 260.60 260.09
6.000 265.77 267.35 266.22 261.18 253.13 247.59 246.88
8.000 254.8Z 255.38 252.29 248.21 239.14 234.59 233.67

10.W 238.99 238.46 238.35 235.25 225.17 228.55 228.76

12.000 223.08 223.55 224.42 222.29 225.15 229.35 231.15
14.000 207.19 207.65 210.49 215.65 225.15 230.15 231.15
16.000 198.59 200.30 203.15 215.65 22S:15 230.15 231.15
18.000 201.05 202.87 207.38 216.78 225.15 230.15 231.15

20.000 207.99 208.82 211.75 219.17 j25.15 230.15 231.15

22.000 214.92 214.76 216.il 221.55 225.15 230.15 231.15
24.00(1 219.97 220.71 220.47 223.75 227.03 230.15 231.15
26.000 224.13 224.81 224.83 225.93 231.00 232.62 233.48
28.000 228.28 228.77 229.19 229.16 234.97 237.79 236.67

30.000 232.43 232.72 233.55 233.72 239.18 242.95 239.84

32.000 236.58 236.68 237.90 238.28 244.33 248.12 243.02
34.000 240.72 240.63 237.90 238.28 244.33 248.12 249.82
36.000 2ti.87 244.57 246.59 247.38 254.@ 258.64 256.76
38.000 249.01 248.59 251.45 251.93 259.?8 263.29 263.70

40.000 253.14 252.78 256.38 256.47 264.83 268.24 270.63

42.000 257.28 256.86 261.31 261.01 270.07 273.19 275.90
M.000 261.84 260.99 264.44 265.55 275.20 278.14 278.86
46.000 266.76 265.13 267.19 270.09 278.13 283.08 281.83
48.000 269.15 268.15 269.15 273.15 279.15 283.65 283.65

50.000 269.15 268.15 269.15 273.i5 279.15 283.65 283.65

52.000 266.56 266.71 267.34 272.91 277.60 283.65 283.65
54.m 263.03 263.77 264.98 267.01 272.87 283.23 283.65
56.000 258.78 258.71 261.50 260.61 268.15 278.30
58.~ 251.33 251.~7 254.04 253.73 263.43 273.38

60.000 243.89 243.43 246.59 246.86 258.71 268.46

62.000 236.45 235.80 239,14 239.99 252.79 263.41
64.000 229.02 228.17 231.70 233.13 243.17 252.01
66.000 221.57 220.54 224.27 226.27 233.56 240.62
68.000 213.76 212.92 216.83 2i9.41 223.95 229.23

70.000 205.95 205.31 209.41 2i2.56 214.34 2J7.86

72.000 198.15 201.81 204.50 205.72 204.75 2M.49
74.000 192.59 199.47 201.58 198.88 195,16 195.12
76.000 194.54 i97.65 198.63 192.7i) 185.57 183.76
78.00U 196.49 197.65 19S.72 188.01 17s.99 172.41

80.000 198.44 197.65 192.80 183.33 lM.42 161.07

82.000 199.65 197.65 189.88 178.65 157.09 149.73
84.000 199.65 197.65 186.96 173.97 155.13 147.01
86.000 197.85 193.52 184.04 169.30 153.i8 146.23
88.000 192.22 188.66 181.13 164.63 151.23 145.45

90.000 186.59 183.81 178.22 164.15 149.29 144.67

14-11
CHAPTER 14
Table 14-2d. Reference Atmosphere Temp (K Wctober

~.a 15“N 30”N 45”N 60”N 75”N 90”N

Alt (km)

O.000 299.15 298.15 293.65 284.15 275.15 262.65. 252.65

2.000 288.38 288.57 283.66 274.15 267.14 259.44 254.37
4.000 277.62 277.81 273.69 263.16 259.13 248.85 243.35
6.000 265.69 266.66 261.74 251.18 246.13 235.84 232.34
8.000 252.55 254.56 247.79 239.21 233.14 222.84 221.34

10.000 239.43 239.44 233.85 227.24 220.16 219.65 220.16

12.000 223.58 224.33 219.92 215.28 220.15 219.65 222.14
14.000 207.69 209.23 211.85 215.15 220.15 219.65 220.14
16.000 196.77 197.92 203.89 215.15 220.15 219.65 218.14
18.000 199.19 200.82 203.65 215.15 219.86 217.66 216.15

20.000 206.33 207.76 205.96 215.15 218.67 215.66 214.16

22.000 213.09 214.70 210.92 217.07 217.47 213.67 212.16
24.000 218.24 219.59 215.88 219.06 218.30 211.68 210.17
26.000 223.39 224.35 219.52 221.04 219.89 211.79 210. I5
28.000 228.53 229.09 223.09 223.02 221.48 214.17 210.15

30.000 233.67 223.84 226.66 225.00 223.06 216.56 210.15

32.000 238.78 238.58 230.22 227.91 224.65 218.94 210.15
34.000 243.72 243.32 233.78 232.07 227.87 221.32 213.99
36.000 248.65 248.30 237.34 236.22 231.83 225.02 217.96
38.000 253.58 253.43 241.77 240.37 235.79 230.37 221.93

40.000 258.50 258.55 246.70 244.52 339.75 235.71 225.89

42.000 263.32 263.67 251.63 249.68 244.59 241.06 231.71
44.000 266.47 266.72 256.56 256.59 249.53 246.40 237.65
46.000 269.62 269.28 261.49 261.82 254.47 251.73 243.58
48.000 271.15 271.15 266.25 266.75 259.40 257.06 249.51

50.000 271.15 271. i5 267.65 267.65 260.15 262.39 253.15

52.000 267.84 267.07 267.65 266.08 260.15 263.15 253.15
54.000 263.33 261.57 263.43 260.77 257.14 263.15 253.15
56.000 258.60 255.93 257.93 255.47 251.82 259.05
58.000 252.92 249.66 252.52 249.72 246.51 251.96

60.000 247.24 243.39 247.42 243.83 241.20 244.88

62.000 240.21 236.37 242.32 237.94 235.90 237.80
64.000 232.39 228.94 232.13 233.42 232.20 230.72
66.000 224.57 221.51 227.04 229.50 230.24 227.15
68.000 216.76 214.08 221.96 225.59 228.28 227.15

70.000 208.95 206.66 216.88 221,67 226.32 227.15

72.000 201.15 201.09 211.80 217.76 224.36 227.15
74.000 195.15 197.19 206.72 213.85 222.40 225.52
76.000 195.15 195.15 201.65 209.94 218.75 221.60
78.000 195.15 195.15 200.67 206.04 214.84 217.69

80.ti 195.15 195.15 199.70 202.15 210.93 213.77

82.000 195.15 195.15 199.65 202.15 207.03 209.87
84.000 195.15 195.15 199.65 202. i5 204.15 205.96
86.000 195.15 i95.15 199.65 202.15 204.15 205.15
88.000 195.15 195.15 199.65 202.15 204.15 205.15

90.000 195.15 195.15 199.65 202.15 204.15 205.15

14-12
 STANDARD AND REFERENCE ATMOSPHERES
Table 14-2e. Reference Atmosphere Wnsity (kg/m3)-Januq

~o 15°N 45”N 60°N 75”N 90”N

Alt (km) 30”N

O.ti 1.1761 1.1891 1.2363 1.3013 1.3742 1.4244 1.4910. + o*

2.000 9.6636 9.7230 9.9321 1.0357 1.0590 1.0711 1.0854
4.000 7.8839 7.8982 8.1234 8.2700 8.3466 8.4810 8.4962 –1
6.000 6.3834 6.4218 6.5793 6.6018 6.6507 6.6428 6.6028
8.000 5.1518 5.2127 5.2712 5.2114 5.2286 5.1417 5.0740

10.OOO 4.1779 4.1820 4.1727 4.0631 3.8811 3.8185 3.7553

12.000 3.3404 3.3114 3.2591 2.9928 2.8353 Z.7984 2.7457
14.000 2.6276 2.5835 2.4310 3.2009 2.0717 2.0452 2.0020
16.000 1.9873 1.9567 1.7992 1.6165 1.5172 1.4905 1.4674
18.000 1.4185 1.4158 1.3073 1.1859 1.1126 1.0832 1.0692

20.000 0.9761 9.6547 9.1885 8.6716 8.1486 7.8497 7.7432 –2

22.000 6.7966 6.6802 6,5057 6,3254 5.9605 5.6510 5.4786
24.000 4.7853 4.7721 4.6425 4.6148 4.3545 4.0226 3.8768
26.000 3.4046 3.4372 3.3411 3.3675 3.1592 2.8697 2.7438
28.000 2.4462 2.4921 2.4296 2.4375 2.28i4 2.0517 1.9256

10.000 1.7766 1.8183 1.7760 1.7546 1.6522 1.4615 1.3392

32.000 1.3103 1.3348 1.3015 1.2715 1.2000 l.m 9.3378
)4.000 0.9715 0.98~7 0.9582 0.9272 0.8740 0.7448 0.6584
16.000 7.240i 7.2933 7.0967 6.7763 6.3791 5.3685 4.6932 –3
38.000 5.3993 5.4221 5.2832 4.9797 4.6657 3.8925 3.3786

Io.000 4.0487 4.6598 3.9559 3.6876 3.4235 2.8388 2.4805

12.W 3.0546 3.0605 2.9787 2.7508 2.4799 2.0819 1.8316
?’4.000 2.3182 2.3396 2.2549 2.0662 1.8ioo 1.5352 1.3589
$6.000 1.7692 1.7981 1.7174 1.5752 1.3411 1.1379 1.0128
$8.000 1.3760 1.3906 1.3334 1.2119 1.0088 0.8477 0.7582

50.000 1.0753 1.0856 1.0354 0.9399 0.7615 0.6347 0.3700

52.ti 8.4906 8.5422 8.1503 7.3740 5.8134 4.7744 0.4338 –4
54.m 6.7090 6.7568 6.4162 5.8169 4.4471 3.6259 3.3209
56.000 5.2860 5.3239 5.0251 4.5435 3.4051 2.7964
58.000 4.1600 4.1913 3.9109 3.5198 2.6086 2.1707

60.000 3.2586 3.2837 3.0303 2.7172 1.9082 1.6796

62.000 2.5402 2.5592 2.3371 2.~ 1.5522 1.2952
M.000 1.9702 1.9835 1.7939 1.@16 1,2002 0.9954
M.000 1.5200 1.5285 1.3701 1.2227 0.9162 0.7587
68.000 1.1663 1.1730 1.0410 0.9297 0.6920 0.5695

70.m 8.8979 8.9430 7.8672 7.M13 $.2271 0.4284

72.000 6.8278 6.7570 5.8499 5.2477 3.9583 0.3275
74,~ 5.2406 5.0487 4.3090 3.9033 3.0319 2.5119 –5
16.f300 3.9678 3.7483 3.1989 2.9301 2.3122 1.9182
18.ot3f) 2.8323 2.7644 2.3840 2.2067 1.7555 1.4583

10.000 2.0164 2.0059 1.7683 1.6538 1.3267 1.1037

12.000 1.4358 1.4343 1.3053 1.2290 0.9979 0.8313
14.000 1.0226 1.0258 0.9587 0.9069 7.4677 0.6231 –6
16.000 7.3229 7.3799 7.0051 6.6705 5.4751 4.60S2
18.000 5.2827 5.3488 5.0910 4.8902 4.0150 3.3735

)0.000 3.7900 3.8549 3.6793 3.5731 2.9448 2.4717

*Power of 10 by which preceding numbers should be multiplied.

14-13
CHAPTER 14
.-- .. 14.7f
T~hle . .. ----------- Atmosphere
Reference .....--r .._-_ Densitv
_____ ~ (ke/m3>Auril
> ~ , ,

Alt (km) w 15°N 30”N 45°N 60”N 75°N 90”N

O.000 1.1716 1.1857 1.2113 1.2678 1.3116 1.3884 .1.4297 + o*

2.000 9.6565 9.7223 9.8903 1.0117 1.0375 1.0635 1.0745
4.~ 7.8992 7.915.9 8.0172 8.1340 8.1648 8.4225 8.4831 –1
6.000 6.4096 6.4388 6.4947 6.5615 6.5709 6.6274 6.6313
8.000 5.17’28 5.2202 5.2446 5.2336 5.2213 5.1557 5.1285

10.OOO 4.1737 4.1823 4.1832 4.1224 3.9711 3.7924 3.7669

12.000 3.3219 3.3066 3.2908 3.1158 2.9216 2.7756 2.7470
14.m 2.6032 2.5754 2.4578 2.2808 2.1499 2.0475 2.0262
16.000 1.9726 1.9299 i.8223 1.6700 1.5823 1.5106 1.4947
18.000 1.4115 1.3954 1.3288 1.2229 1.1648 1.1147 1.1029

20.000 0.9725 9.5265 9.3798 8.9578 8.5768 8.2281 8.1399 –2

22.000 6.7816 6.6002 6.6703 6.5625 6.3162 6.0742 6.0084
24.000 4.7828 4.6845 4.7883 4.7676 4.6523 4.4850 4.4260
26.000 3.4092 3.3930 3.4617 3.4649 3.4129 3.3122 3.2438
28.000 2.4585 2.4733 2.5195 2.5294 2.4976 2.4465 2.3849

30.000 1.7923 i.8140 1.8456 1.8516 1.8336 1.7839 1.7589

32.000 1.3165 1.3383 1.3603 1.3540 1.3391 1.3068 1.3011
34.000 0.9741 0.9929 1.0092 0.9968 0.9836 0.9621 0.9631 –3
36.000 7.2575 7.4043 7.5300 7.3817 7.2550 7.1172 7.1019
38.000 5.4428 5.5315 5.6476 5.4858 5.3740 5.2896 5.2700

40.000 4.1077 4.1592 4.2575 4.1059 4.@89 3.9134 3.9347

42.000 3.1189 3.1468 3.2253 3.1075 3.oio8 2.9155 2.9550
44.000 2.4~ 2.4043 2.4632 2.3692 2.2758 2.1882 2.2319
46.000 1.8624 1.8595 1.8890 1.8139 1.7371 1.6540 1.6949
48.000 1.4527 1.&12 1.4582 1.4007 1.7371 1.6540 1.6949

50.000 1.1352 1.1262 1.1360 1.0933 1.0396 0.9812 1.0104

52.000 8.9597 8.8794 8.8895 8.6128 8.1322 7.6739 C!.7872 –4
54.000 7.0759 7.0044 7.0239 6.8140 6.4353 6,0473 6.1347
56.000 5.5694 5.5228 5.5281 5.3690 5.1181 4.8196
58.000 4.3781 4.3994 4.3441 4.2210 4.0501 3.8208

60.000 3.4619 3.4507 3.4001 3.3023 3.1882 3.0122

62.OQO 2.7184 2.7Ci07 2.6471 2.5703 2.4960 2.3609
64.000 2.1189 2.0987 2.0495 1.9898 1.9429 1.8391
66.000 1.6387 1.6186 1.5776 1.5319 1.4984 1.4234
68.000 1.2568 1.2384 1.2075 1.1670 1.1456 1.0840

70.000 9.5682 9.3944 9.1854 8.8452 8.7i79 0.8218 –5

72.000 7.2109 7.W88 6.9386 6.6727 6.6024 0.6203
74.000 5.3759 5.2331 5.2Ci32 5.W4 4.9755 4.6613
76.000 3.9611 3.8493 3.8718 3.7419 3.7359 3.4859
78.~ 2.7561 2.7309 2.8279 2.7806 2.7954 2.6145
30.m 1.9224 1,9355 2.0362 2.0551 2.0773 1.9477
B2.(ioo 1.3465 1.3584 1.4640 1.5152 1.5325 1.4389 –6
84.000 o.947i 0.9568 1.0512 1.1105 1.1220 1.0535
B6.000 6.7014 6.7706 7.5368 8.0774 8.1494 7.6418
B8.000 4.7692 4.8129 5.3958 5.7400 5.8691 5.4874

90.000 3.4132 3.4364 3.8574 4.0433 4.1894 3.8701

. . .. .
*Power of 10 by which preceding ntimkrs should be mult]plled

14-14
 STANDARD AND REFERENCE ATMOSPHERES
Table 14-2g. Reference Atmosphere Density (kg/m3)--July

Alt (km) 0° 15”N 30”N 45”N 60”N 75”N 90”N

O.000 1.1797 1.1751 1.1717 1.2003 1.2252 1.2778 1.2919 + o*

2.000 9.7077 9.6694 9.7103 0.9780 0.9987 1.0032 1.0120
4.000 7.9287 7.8986 7.9314 7.9946 8.0757 8.1533 8.1528 –1
6.000 6.4239 6.4017 6.4266 6.4769 6.5120 6.5568 6.5576
8.000 5.1594 5.1685 5.2140 5.2145 5.2231 5.2120 5.2124

10.OOO 4.1775 4.1875 4.1827 4.1502 4.1345 3.9747 3.9519

12.000 3.3352 3.3444 3.3109 3.2612 3.0546 2.9399 2.9076
14.000 2.6191 2.6275 2.5823 2.4580 2.2570 2.1773 2.1645
16.000 1.9556 1.9493 1.9214 1.7932 1.6679 1.6192 1.6116
18.000 1.3723 1.3730 1.3521 1.3021 1.2329 1.2043 1.2001

20.000 0.9525 9.5976 9.5807 9.4313 9.1149 8.9601 8.9390 –2

22.000 6.6925 6.7780 6.8380 6.8564 6.7399 6.6671 6.6592
24.000 4.7926 4.8334 4.9144 5.0062 4.9463 4.9618 4.9518
26.000 3.4691 3.5029 3.5556 3.6673 3.6143 3.6569 3.6647
28.000 2.5265 2.5552 2.5891 2.6835 2.6556 2.6803 2.7081

30.000 1.8509 1.8743 1.8969 1.9638 1.9598 1.9781 2.0097

32.000 1.3637 1.3823 1.3981 1.4460 1.4495 1.4694 1.4975
34.000 1.0103 1.0248 1.0363 1.0712 1.0791 1.0985 1.1062
36.000 7.5248 7.6366 7.7248 7.9815 8.0839 8.2646 8.2372 –3
38.000 5.6332 5.7168 5.7779 5.9797 5.0918 6.2554 6.1828

40.000 4.2381 4.2984 4.3456 4.5040 4.6169 4.7601 4.6763

42.000 3.2037 3.2475 3.2867 3.4100 3.5184 3.6410 3.5820
44.000 2.4293 2.4650 2.5144 2.5946 2.6955 2.7989 2.7782
46.000 1.8493 1.8795 1.9321 1.9836 2.0908 2.1619 2.1604
48.000 1.4275 1.4450 1.4931 1.5315 1.6358 1.7007 1.6911

50.000 1.1125 1.1250 1.1632 1.1971 1.2850 1.3408 1.3332

52.000 8.7481 8.8047 9.1197 9.3670 1.0150 1.0572 1.0511 –4
54.000 6.8831 6.9145 7.1504 7.4642 8.0867 8.3501 8.2891
56.000 5.4137 5.4579 5.6185 5.9204 6.4179 6.6864
58.000 4.2874 4.3247 4.4591 4.6911 5.0734 5.3337

60.000 3.3722 3.4028 3.5152 3.6881 3.9941 4.2378

62.000 2.6331 2.6574 2.7514 2.8804 3.1454 3.3550
64.000 1.0402 2.0588 2.1373 2.2339 2.4948 2.7023
66.000 1.5679 1.5815 1.6469 1.7196 1.9607 2.1552
68.000 1.1961 1.2038 1.2581 1.3134 1.5256 1.7003 -5

70.000 9.0350 9.0750 9.5234 9.9476 1.1742 1.3256

72.000 6.7522 6.6490 7.0578 7.4673 8.9324 1.0199
74.000 4.9325 4.8263 5.1557 5.5522 6.7071 7.7331
76.000 3.4623 3.4821 3.7497 4.0732 4.9649 5.7676
78.000 2.4394 2.4867 2.7148 2.9407 3.6178 4.2228
80.000 1.7251 1.7763 1.9564 2.1061 2.5905 3.0274
82.000 1.2285 1.2690 1.4031 1.4957 1.8162 2.1187
84.000 0.8809 0.9068 1.0013 1.0528 1.1999 1.3730 –6
86.000 6.3733 6.6002 7.1100 7.3426 7.8878 8.7593
88.000 4.6691 4.7844 5.0219 5.0703 5.1588 5.5760

90.000 3.3898 3.4399 3.5279 3.3925 3.3563 3.5420

. . . ... ... .. ,
‘ower of 10 by which preceding numbers snoulst oe mumpllea.

14-15
CHAPTER 14
Table 14-2h. Reference Atmosphere Density.- (kg/m3)--October

Alt (km) 0° 15°N 30°N 45”N 60”N 75”N 90”N

O.000 1.1769 1.1816 1.2063 1.2474 1.2795 1.3385 1.3981 + O*

2.000 9.6821 9.6775 9.8598 1.0122 1.0240 1.0425 1.0617
4.000 7.9068 7.9044 8.0015 8.1810 8.1419 8.3111 8.4294 –1
6.000 6.4329 6.4133 6.4897 6.5738 6.5407 6.6134 6.6221
8.000 5.2053 5.1764 5.2471 5.2270 5.1931 5.1955 5.1425

10,OOO 4.1652 4.1790 4.1912 4.1081 4.0696 3.8648 3.7877

12.000 3.3261 3.3274 3.3023 3.1874 2.9860 2.8327 2.7569
14.000 2.6135 2.6080 2.5021 2.3246 2.1912 2.0766 2.0434
16.000 1.9701 1.9722 1.8579 1.6946 1.6082 1.5226 1.5107
18.000 1.3779 1.3810 1.3313 1.2356 1.1821 1.1252 1.1140

20.000 0.9524 9.5802 9.3893 9.0114 8.7162 8.2945 8.1936 –2

22.000 6.6793 6.7278 6.6771 6.5238 6.4173 6.0980 6.0101
24.000 4.7657 4.8169 4.8106 4.7364 4.6790 4.4711 4.3964
26.000 3.4279 3.4765 3.4862 3.4493 3.4076 3.2395 3.1819
28.000 2.4848 2.5268 2.5398 2.5196 2.4879 2.3289 2.3031

30.000 1.8144 1.8489 1.8599 1.8460 1.8208 1.6807 1.6673

32.000 1.3343 1.3617 1.3687 1.3512 1.3358 1.2175 1.1889
34.000 0.9885 1.0091 1.0122 0.9890 0.9760 0.8852 0.8528
36.000 7.3687 7.5165 7.4956 7.2803 7.1469 6.4248 6.1550 –3
38.000 5.5256 5.6298 5.5731 5.3888 5.2620 4.6616 4.4688

40.000 4.1672 4.2419 4.1689 4.0101 3.8948 3.4080 3.2409

42.000 3.1613 3.2145 3.1366 2.9879 2.8883 2.5095 2.3687
M.000 2.4237 2.4664 2.3732 2.2277 2.1543 1.8607 1.7450
46.000 1.8643 1.9011 1.8063 1.6837 1.6164 1.3888 1.2952
48.000 1.4467 1.4728 1.3971 1.2809 1.2197 1.0431 0.9767

50.000 1.1295 1.1498 1.0869 0.9928 0.9386 0.7882 0.7482

52.000 8.9199 9.1029 8.5800 7.7661 7.2462 6.0836 0.5732 –4
54.000 7.0493 7.2119 6.7740 6.1400 5.6554 4.7098 4.3930
56.000 5.5537 5.6898 5.3200 4.8317 4.4337 3.7001
58.000 4.3709 4.4754 4.1526 3.7897 3.4585 2.9235

60.000 3.4219 3.4993 3.2253 2.9575 2.6836 2.2949

62.000 2.6779 2.7270 2.4920 2.2944 2.0710 1.7889
64.000 2.0863 2.1125 1.9149 1.7597 1.5788 1.3842
66.000 1.6118 1.6230 1.4632 1.3402 1.1915 1.9476
68.000 1.2341 1.2360 1.1114 1.0161 8.9728 0.7798

70.000 9.3589 9.3244 8.3%6 7.6683 6.7413 0.5805 –5

72.000 7.0241 6.9050 6.2932 5.7589 5.0531 0.4323
74.000 5.1671 5.0394 4.6880 4.3034 3.7788 3.2416
76.000 3.6738 3.6236 3.4675 3.1991 2.8382 2.4457
78.000 2.6127 2.5767 2.5022 2.3654 2.1236 1.8363
BO.000 1.8585 1.8326 1.8032 1.7391 1.5808 1.3719
B2.000 1.3222 1.3036 1.2925 1.2508 1.1705 1.0196
84.000 0.9409 0.9276 0.9264 0.8998 8.5777 0.7537 -6
36.000 6.6974 6.6017 6.6420 6.4750 6.1893 5.4678
38.000 4.7680 4.6993 4.7628 4.6600 4.4668 3.9511

JO.000 3.3951 3.3458 3.4160 3.3544 3.2244 2.8557

.-
*Power of 10 by which preceding numbers should be multiplied.

14-16
 1

STANDARD AND REFERENCE ATMOSPHERES

Table 14-2i. Reference Atmosphere Pressure (mb)-January

Alt (km) w 15°N 30°N 45°N WN 75”N 90”N

0.000 1.0100 1.0126 1.0191 1.0166 1.0144 1.0147 1.0150 +3*

2.000 8.0106 8.0179 8.0160 7.8835 7.7806 7.7293 7.6613 +2
4.000 6.3009 6.3007 6.2540 6.0693 5.9354 5.8516 5.7706
6.000 4.9119 4.9087 4.8205 4.6180 4.4699 4.3734 4.2948
8.000 3.7918 3.7760 3.6661 3.4662 3.3102 3.2228 3.1547

10.OOO 2.8842 2.8621 2.7470 2.5629 2.4192 2.3472 2.2922

12.000 2.1535 2.1340 2.0245 1.8785 1.7673 1.7041 1.6602
14.000 1.5741 1.5618 1.4729 1.3751 1.2914 1.2337 1.1990
16.000 1.1258 1.1203 1.0634 1.0054 1.9436 1.8906 0.8620 +1
18.000 7.9414 7.9315 7.6236 7.3420 6.8878 6.4106 6.1588

20.000 5.6424 5.6442 5.4791 5.3555 5.0213 4.6004 4.3714

22.000 4.0525 4.0763 3.9683 3.9065 3.6559 3.2919 3.0926
24.000 2.9403 2.9757 2.8952 2.8501 2.6584 2.3570 2.1884
26.000 2.1538 2.1863 2.1271 2.0797 1.9314 1.6913 1.5488
28.000 1.5920 1.6163 1.5717 1.5189 1.4065 1.2162 1.0869

30.000 1.1867 1.2020 1.1670 1.1153 1.0271 1.8768 1.7805

32.000 8.9018 8.9907 8.7098 8.2413 0.7521 0.6358 0.5618 +0
34.000 6.7103 6.7618 6.5377 6.1259 5.5230 4.6386 4.0855
36.000 5.0825 5.1143 4.9348 4.5816 4.0665 3.4042 3.0000
38.000 3.8695 3.8943 3.7456 3.4517 3.0035 2.5126 2.2231

40.000 2.9634 2.9848 2.8584 2.6188 2.2252 1.8647 1.6590

42.000 2.2824 2.3019 2.1927 2.0003 1.6577 1.3912 1.2437
44.000 1.7675 1.7844 1.6905 1.5376 1.2458 1.0432 0.9366
46.000 1.3759 1.3876 1.3096 1.1884 0.9438 0.7862 0.7084 –1
48.000 1.0749 1.0823 1.0168 0.9207 7.1802 0.5953 0.5380

50.000 8.4007 8.4501 7.8958 7.1403 5.4799 0.4529 0.4103

52.000 6.5594 6.5948 6.1234 5.5334 4.1911 3.4610 3.1400
54.000 5.1036 5.1289 4.7265 4.2665 3.2060 2.6556 2.4037
56.000 3.9560 3.9723 3.6300 3.2713 2.4528 2.0394
58.000 3.0525 3.0619 2.7745 2.4986 1.8761 1.5623

60.000 2.3434 2.3472 2.1105 1.9012 1.4346 1.1926

62.000 1.7897 1.7890 1.5975 1.4412 1.0942 1.9073
64.000 1.3593 1.3554 1.2030 1.0882 1.8306 0.6877
66.000 1.0265 1.0204 0.9011 0.8184 0.6276 0.5195 –2
68.000 7.7069 7.6307 6.7122 6.1291 4.7405 0.3925

70.000 5.7500 5.6623 4.9711 4.5704 3.5808 2.9720

72.000 4.2580 4.1680 3.6632 3.3991 2.7051 2.2512
74.000 3.1098 3.0464 2.6983 2.5283 2.0379 1.6982
76.000 2.2347 2.2114 1.9867 1.8794 1.5282 1.2753
78.000 1.5906 1.5938 1.4565 1.3907 1.1406 0.9531
BO.000 1.1324 1.1409 1.0626 1.0237 0.8471 0.7089 –3
B2.000 8.0642 8.1586 7.7126 7.4973 6.2598 5.2469
B4.000 5.7436 5.8351 5.5682 5.4712 4.6013 3.8628
B6.000 4.0905 4.1728 3.9981 3.9795 3.3735 2.8309
B8.000 2.9008 2.9707 2.8545 2.8846 2.4738 2.0738

90.000 2.0453 2.1023 2.0260 2.0837 1.8144 1.5194

*Power of 10 by which preceding numbers should be multiplied.

14-17
CHAPTER 14
Table 14-2i.
—.’
Reference Atmosphere
1
Pressure (mb>Auril

Alt (km) 0° 15°N 3VN 45”N 60”N 75”N 90”N

0.000 1.0095 1.0114 1.0159 1.0159 1.0134 1.0169 i .0205 +3*

2.000 8.0104 8.0090 8.0108 7.9327 7.8374 7.7818 7.7772 +2
4.000 6.2997 6.2899 6.2640 6.1563 6.0265 5.9201 5.8956
6.000 4.9065 4.8940 4.8518 4.7216 4.5860 4.4489 4.4181
8.000 3.7810 3.7589 3.7079 3.5712 3.4344 3.2982 3.2696

10.000 2.8720 2.8444 2.7901 2.6596 2.5323 2.4228 2.3996

12.000 2.1437 2.1167 2.0633 1.9511 1.8631 1.7859 1.7675
14.000 1.5686 1.5457 1.5060 1.4283 1.3709 1.3174 1.3037
16.000 1.1243 1.1081 1.0917 1.0457 1.0090 0.9720 0.9617 +1
18.000 7.9466 7.8573 7.8633 7.6583 7.4282 7.1729 7.0967

20.000 5.6576 5.6017 5.6791 5.6094 5.4693 5.2942 5.2374

22.000 4.0725 4.0537 4.1336 4.1095 4.0277 3.9083 3.8660
24.000 2.9619 2.9721 3.0303 3.0144 2.9667 2.8857 2.8544
26.000 2.1751 2.1952 2.2361 2.2204 2.1863 2.1311 2.1123
28.000 1.6118 1.6311 1.6604 1.6425 1.6156 1.5741 1.5680

30.000 1.2035 1.2189 1.2404 1.2200 1.1976 1.1656 1.1674

32.000 9.0499 9.1598 9.3203 9.1144 0.8917 0.8672 0.8718 +0
34.000 6.8505 6.9201 7.0416 6.8516 6.6787 6.4835 6.5297
36.000 5.2189 5.2550 5.3475 5.1817 5.0313 4.8691 4.9156
38.000 4.0001 4.0135 4.0811 3.9442 3.8154 3.6729 3.7226

40.000 3.0840 3.0837 3.1296 3.0220 2.9122 2.7863 2.8353

42.000 2.3909 2.3829 2.4110 2.3290 2.2367 2.1288 2.1714
44.000 1.8616 1.8513 1.8650 1.8025 1.7281 1.6375 1.6717 –1
46.000 1.4525 1.4423 1.4473 1.4006 1.3426 1.2678 1.2935
48.000 1.1349 1.1259 1.1266 1.0922 1.0466 0.9876 1.0058

50.000 8.8690 8.7980 8.7773 8.5260 8.1813 0.7722 0.7835

52.000 6.9254 6.8698 6.8383 6.6521 6.3997 6.0390 6.1049
54.000 5.3897 5.3482 5.3121 5.1704 5.0026 4.7219 4.7573
56.000 4.1795 4.1501 4.1092 4.0012 3.8918 3.6761
58.000 3.2290 3.2022 3.1641 3.0814 3.0111 2.8452

60.000 2.4793 2.4533 2.4229 2.3604 2.3164 2.1886

62.000 1.8888 1.8654 1.8447 1.7981 1.7712 1.6727
64.000 1.4271 1.4072 1.3959 1.3618 1.3459 1.2698
66.000 1.0687 1.0526 1.0497 1.0252 1.0161 0.9571 –2
68.000 7.9288 7.8036 7.8411 7.6744 7.6335 0.7171

70.000 5.8214 5.7300 5.8149 5.7172 5.7060 5.3494

72.000 4.2257 4.1655 4.2797 4.2381 4.2435 3.9715
74.000 3.0300 2.9988 3.1248 3.1254 3.1391 2.9343
76.000 2.1440 2.1369 2.2623 2.2926 2.3092 2.1572
78.000 1.5090 1.5141 1.6249 1.6725 1.6871 1.5758

80.000 1.0675 1.0731 1.1643 1.2131 1.2234 1.1411

82.000 7.5916 7.6212 7.3308 8.7433 8.8026 8.1878 –3
84.000 5.4293 5.4373 5.9522 6.2526 6.2813 5.8179
86.000 3.9051 3.8967 4.2465 4.4348 4.4435 4.0914
88.000 2.8243 2.8048 3.0251 3.1248 3.1148 2.8457

90.000 2.0536 2.0275 2.1519 2.2011 2.1624 1.9569

*Power of 10 by which preceding numbers should be multiplied

14-18
 i

STANDARD AND REFERENCE Atmospheres

Table 14-2k. Reference Atmosphere Pressure (mb)-July

Alt (km) o“ 15”N 30”N 45”N 60”N 75”N 90°N

O.000 1.0114 1.0108 1.0129 1.0135 1.0099 l.olli 1.0(130 +3*

2.000 8.0167 8.0182 8.0323 8.0057 7.9221 7.880~ 7.8763 +2
4.000 6.2981 6.3057 6.3110 6.2688 6.1556 6.0994 6.0870
6.000 4.9007 4.9130 4.9112 4.8560 4.7319 4.6602 4.~73
8.000 3.7740 3.7889 3.7768 3.7154 3.58S5 3.5098 3.4964

10.OOO f,8659 2.8784 2.86i8 2.8026 2.6724 2.6077 2.59S1

12.000 2.1358 2.1462 2.1329 2.0810 1.9742 1.9355 1.9293
14.000 1.5577 1.5662 1.563 1.5215 1.4587 1.4384 1.4362
16.000 1.1148 1.1208 1.1204 1.1100 1.0780 1 .M97 i .M93 +1
18.000 7.9201 7.9959 8.ti91 8.1030 7.9683 7.9568 7.9632

20.000 5.6870 5.7531 5.8235 5.9336 5.8909 5.9195 5.9312

22.000 4.1290 4.1786 4.2420 4.3606 4.3560 4.4046 4.4185
24.000 3.0263 3.0622 3.glo3 3.2154 3.2235 3.2780 3.2922
‘26.000 2.2319 2.2W 2.2948 2.3784 2.3966 2.4419 2.4562
28.000 1.6556 1.6780 1.7034 1.7653 1.7912 1.8295 1.8398

30.oi)o 1.2349 1.2521 1.2717 1.2175 1.3455 1.3795 1.3836

32.000 9.2614 9.3917 9.5480 9.8912 1.0167 1.0466 1.0447 +0
34.000 6.9815 7.0790 7.2069 7.4672 7.7286 7.9867 7.933s
36.000 5.2893 5.3614 S.4680 5.6678 5.9090 6.-1289 6.0713
38.000 4.0266 4.0794 4.1705 4.3244 4.5428 4,7279 4.6802

40.ti 3.07b7 3.1183 3.1982 3.3159 3.5111 3.6654 3.6329

42.~ 2,3661 2.3945 2.4654 2.5549 2.7277 2.8553 2.8376
44.000 1.8260 1.8468 1.9087 i .9778 2.1294 2.2346 2.2240 –1
46.000 1.4161 1.4304 1.4819 1.5379 1.6693 1.7567 1.7478
48.000 1.1028 1.1123 1.1536 1.2008 1.3108 1.3847 1.3770

50.000 8.5953 8.6599 8.9876 9.3866 1.0297 1.0917 1.0855

52.000 6.6938 6.7410 6.9986 7.3382 8.0887 8.6086 8.5588
54:000 5.1970 5.2354 5.4389 5.7211 6.3343 6.7889 6.7492
56.000 4.0215 4.0533 4.2174 4.4357 4.9401 5.3417
58.000 3.0932 3.1169 3.2518 3.4168 3.8364 4.1856

60.000 2.3609 2.3778 2.4883 2.6i35 2.9661 3.2658

62.000 1.7873 1.7987 1.8888 1.98d 2.2823 2.5368
64.000 1.3412 1.3484 1.4216 1.4949 1.7415 1.9549
66.000 0.9972 1.0012 1.0602 i.1169 1,3145 1.4886 –2
68.000 7.3396 7.3582 7.8313 8.2726 9.8076 1.1189

70.000 5.3414 5.3484 5.7247 6.0699 7.2253 8.2902

72.000 3.8406 3.8519 4.1432 4.4096 5.2499 6.0456
74.W 2.7270 2.7636 2.9833 3.1697 2.~8 3.0425
76.000 1.9335 1.9756 2.1382 2.2531 2.6448 3.0425
78.000 1.3759 1.4i08 1.5253 1.5871 1.8277 2.0899

80.000 0.9826 1.0078 1.0827 1.1083 1.2375 1.3997 –3

82.000 7,0408 7.2002 7.6480 7.6707 8.1900 9.1071
84.000 5.0489 5.1453 5.3742 5.2582 5.3435 5.7944
86.000 3.6196 3.6665 3.7563 3.5684 2.2396 2.3281
88.000 2.5763 2.5911 2.6111 2.3961 2.2396 2.3281

90.000 1.8156 1.8150 1.8048 1.5985 i .4383 1.4709

‘owerof 10 by which preceding numbrs should be multiplied.

14-19
 CHAPTER 14
Tnhl+
.“”. -
I. A.71
. . . .
Refererce
. . . . . . . . ..- .
Atmncnhere
. . . ..-wr . . . . .
Pressl]re
. .. ----- (mhMctober
\. ..- ,

Alt (km) @ 15”N 3@N 45”N 60°N 75°N 90”N

0.000 1.0107 1.0113 1.0169 1.0175 1.0106 1.0092 ‘1.0140 +3*

2.000 8.0149 8.0165 8.0286 7.9662 7.8528 7.7639 7.7527
4.000 6.3010 6.6036 6.2863 6.1801 6.0564 5.9371 5.8884 +2
6.000 4.9062 4.9091 4.8760 4.7399 4.6213 4.4773 4.4166
8.000 3.7737 3.7826 3.7323 3.5891 3.4755 3.3235 3.2674

10.OOO 2.8627 2.8723 2.8135 2.6797 2.5720 2.4368 2.3937

1~.m 2.1347 2.1427 2.0847 1.9697 1.8870 1.7860 1.7580
14.000 1.5581 1.5664 1.5216 1.4356 1.3847 1.3093 1.2913
16.W 1.1128 1.1205 1.0974 1.0466 1.0163 0.9600 0.9460 +1
18.000 7.8791 7.9611 7.8713 7.6313 7.4607 7.0305 6.9124

20.000 5.6412 5.7136 5.6849 5.5654 5.4712 5.1350 5.0370

22.000 4.0858 4.1464 4.1378 4.0651 4.0062 3.7403 3.6603
24.000 2,9857 3.0364 3.0315 2.9783 2.9321 2.7169 2.6524
26.@ 2.1982 2.2389 2.2325 2.1886 2.1509 1.9695 1.9194
28.000 1.6300 1.6617 1.6525 1.6131 1.5817 1.4318 1.3893

30.000 1.2170 1,2411 i.2291 1.1923 1.1659 1.0448 1.0058

32.000 9.1464 9.3262 9.1856 8.8406 0.8614 0.7652 0.7303 +0
34.000 6.9158 7,0486 6.8961 6.5885 6.3843 5.6244 5.3358
36.000 5.2595 5.3575 5.2020 4.9367 4.7562 4.1500 3.9211
38.000 4.0221 4.0956 3.9467 3.7183 3.5616 3.0827 2.8978

40.000 3.0923 3.1483 3.oi13 2.8147 2.6804 2.3059 2.1556

42.000 2.3896 2.4330 2.3100 2.i415 2.0279 1.7365 1.6158
44.000 1.8540 1.8884 1.7813 1.6408 1.5431 1.3161 1.2201 –1
46.000 1.4429 1.4695 1.3805 1.2654 1.1807 1.0035 0.9277
48.000 1.1260 1.1463 1.0734 0.9808 9.0823 0.7697 0.7097

50.000 8.7921 8.9499 8.3510 7.6277 7.0097 0.5937 0.5437

52.000 6.8580 6.9786 6.4881 5.9319 5.4112 4.5954 4.1658
54.000 5.3285 5.4152 5.0155 4.5962 4.1744 3.5576 3.1923
56.000 4.1227 4.18~0 3.8563 3.5432 2.4473 2.7515
58.000 3.i734 3.2013 2.9493 2.7166 2.ti73 2.1145

60.000 2.4286 2.4449 2.2435 2.0700 1.8581 1.6132

62.000 1.8465 1.8503 1.6969 1.5671 i .4024 1.2211
64.000 1.3917 1.3883 1.2760 1.1791 1.0524 0.9168 –2
66.000 1.0390 1.0320 0.9536 0.8829 0.7875 0.683i
68.000 7.6789 7.5958 7.0818 6.5802 5.8798 0.5084

70.000 5.6135 5.5316 5.2237 4.8795 4.3796 4.7856

72.000 4.0557 3.9860 3,8261 3.5999 3.2544 2.8189
74.000 2.8945 2.8526 2.7819 2.6417 2.4125 2.0985
76.000 2.0580 2.0299 2.0072 1.9279 1.7822 1.5558
78.000 1.4636 1.4434 1.4414 1.3990 1.3096 1.1475
80.000 1.0411 1.0266 1.0336 1.0091 0.9571 0.8419 –3
82.000 7.4071 7.3031 7.4076 7.2585 6.9561 6.1426
84.000 5.2710 5.1963 5.3096 S.2217 5.0267 4.4562
86.000 3.7517 3.6981 3.8065 3.7572 3.6270 3.2199
88.000 2.6709 2.6324 2.7295 2.7040 2.6176 2.3267

90.000 1.9019 1.8742 1.9577 1.9465 1.8895 1.6817

*Power of 10 by which preceding numbers should be multiplied

14-20
 i

STANDARD AND REFERENCE ATMOSPHERES

up to 90 km. The largest seasonal changes in mean monthly metrical about the pole. During the winter season, changes
temperatures occur norfh of 60° latitude in arctic and sub- with longitude remaitt relatively small at low latitudes but
arctic regions. At altitudes up to 60 km, coldest temperatures become as important as changes with latitude and season in
occur in January and warmest in July. Conditions are re- arctic and subarctic regions. The magnitude of the longi-
versed at altitudes between 60 and 90 km with coldest tem- tudinal variations in arctic and subarctic regions during win-
peratures occurring in July and warmest in January. ter is illustrated by a set of atmospheric models depicting
Cross sections of the January and July densities between Janttary conditions between the surface and 54 km at 10°,
the equator and north pole are shown in Figure 14-13. Den- IW, and 140”W for 60”N, and 10° and 140°W for 75”N.
sities at the various altitudes are expressed as percentages The models are based on radiosonde observations, rocket-
of the densities in the U.S. Standard Atmosphere. The larg- sonde observations, constant pressure maps between 5 and
est departures from standard in January and July densities 0.4 mb, hydrostatic build-up techniques from the 5- and 10-
occur north of 65° latitude near 70 km, where mean January mb levels, and the thermal wind equation. The atmospheric
densities, 51 Yoof standard, are a third of the July densities properties for these January models at 60° and 75”N are
which are 168% of standard. given in Table 1~-3. Temperature-height profiles for the
individual models are shown in Figures 14-14 and 14-15.
14.2.1.4 Longitudinal Variations. In summer, longi- The density-height profiles for the 60”N January models
tudinal variations in the structure of the atmosphere are developed for lW, 100° and 140”W (Figure 14-16) indicate
relatively small at all latitudes and at all altitudes above 20 that the longitudinal variation in bean monthly densities at
km. Isotherms and contour lines on constant pressure-height 40 km in winter ranges from 5% less than standard at 140”W
charts in the stratosphere and mesosphere parallel the lati- to 2070 less at 10°W. Density profiles for longitudes 10°
tude circles, and the associated circulation pattern is sym- and 140”W at 75”N (Figure 14-17), indicate that the lon-

I I I I I I 1 I I 1 I I
mY

Wwmo

,-----------
-----------
-/ ‘ ;

---
0
1 -*’
,4’

100
------ --
:=...J
------------ /
100

-
LATITU)E

Figure 14-13, Latitudinal density-height cross-sections for January and July.

14-21
Table 14-3a, Temperature (K) in January for 60”N and 75”N at specified Table 14. 3b Density (kg/mJ) in January for 60”N and 75”N at Specified
bngitudes

Alti@de 60°N 75”N Altitude WN 75”N

(km) 10”W lQOOW 140”W 10”W 140”W (m) 10”W 100”W 140bw 10”W 140°w

O.000 278.15 246.15 269.15 257.65 242.15 O.000 1.2555 1.4400 1.3075 1.3ti8 1.4681 +0*
2.000 26S.73 248.64 261.14 252.64 246.95 2.000 1.0219 1.0829 1.0412 1.0612 1,0880
4.000 293.33 242.13 251.73 244.11 242.12 4,000 8.2374 8.4250 8.2788 B.9467 8.4007 – 1
6.000 240.93 232.14 238.13 232.10 232.34 6.000 6.5693 6.5882 6:6212 6.5870 6.5674
8.00U 228.54 222.14 224.54 220.10 217.34 8.000 $.1775 5.0972 5.2271 5.1341 5.1802

10.OOO 216.16 217.15 221.15 213.64 213.65 10.OOO 4.0275 3.815> :.9010 3.8527 3.8306
12,000 216.15 217.15 221.15 212.64 213.65 12.000 2.9382 2.7874 2.8660 2.8103 2.7832
i4.000 216.15 217.15 221.15 211.65 213.65 14.000 2.1438 2.0367 2.1061 2.0473 2.0226
16.000 216.15 215,96 221.15 210.65 213.65 16.000 1.5645 1.4954 1.5479 1.4895 1.4701
18.W 214.17 214.76 221.15 209.65 213.65 18.000 1.1508 1.0963 1.1379 1.0822 1.0688

20.000 212.18 213.57 221.15 206.40 212.16 20.000 8.*35 8.0254 8.3672 7.9325 7.8198 –2
22.000 207.2’7 212.37 218.22 200.82 210.17 22.000 6.2514 5.8655 6.2233 5.8370 5.7198
24.@ 202.29 211.18 215.24 195.25 208.18 24.000 4.5969 4.2802 4.6120 4.2589 4.1721
26.000 204.84 213.07 217.46 195.15 211.27 26.000 3.2519 3.0803 3.3352 3.~83 2.9732
28.000 207.62 215.06 219.84 199.22 214.45 28.000 2.3090 2.2230 2.4190 2.0881 2.1291

30.000 210.40 2~7.04 222.22 203.39 217.63 30.000 1.6474 1.6094 1.76101 1,4600 1.5325
32.000 213.18 219.03 224.@ 207.56 220.50 32.000 1.1808 1.1689 1.2865 1.0284 1.1085
34.000 215.95 221.94 226.98 211.73 223.98 34.000 8.5018 8.4851 9.4326 7.2973 8.0574 – 3
36.OCKJ 220.38 224.91 229.36 215.89 227.15 36.000 6.1087 6.1851 6.9392 5.2134 5.8840
38.000 224.94 227.88 231.73 221.17 230.32 38.000 4.4174 4.5282 5.1221 3.73324.3165

40.000 229.49 232.65 234.03 226.51 233.49 40.000 3.2158 3.3069 3.7944 2.6943 3.1806
42.000 234.03 237.59 236.20 231.86 236.66 42.000 2.3561 2.4296 2.8208 1.9597 2.3537
44.000 238.58 242.53 238.37 237.20 239.82 44.000 1.7370 1.7967 2.1031 1.4361 1.7491
46.000 243.12 247.47 240.55 242.53 242.80 46.000 1.2881 1.3370 1.5724’ i .0599 1.3059
48.000 247.66 249.85 242.72 244.89 245.77 48.000 0.9607 1.0100 1.1790 D.7959 0.9788

50.000 ?52.19 251.82 245.83 246.87 248.73 50.000 7.2057 7.6608 8.8318 6.0021 7.3628 –4
52.000 256.73 252.15 250.76 248.84 251.69 52.000 5.4328 5.8578 6.6017 4.5369 5.5581
54.000 257.55 252.15 254.15 250.15 252.15 54.000 4.1700 4.4858 4.9924 3.4462 4.2473
*Power of 10 by which precdlng numbers should be multiplied.

60----1-, 60°N’
, ,
JANUARi
1 , I

TEMPERATURE
, !

PROFILES
I I
1
60 I
1’
1 1
75”N
1 I
JANUARY
I , I
TEMPERATuRE
I ,
PROFliE~
, r
‘ ‘1

((’
i)
\

‘:~ I70 190 210

TEMPER%~URE
2!30
( K)
270 290

Figure 14-14. Mean monthly temperature-hpight profiles for the WON Figure 14-15. Mean monthly temperature-height profiles for the 75”N
models at longitudes 10°, 100°, and 140W. models at 10” and 140W.

14-22
 a
STANDARD AND REFERENCE ATMOSPHERES
Table 14-3c, Pressure (rob) in January for 60”N and 75”N at Specified 60 I , I , I
Longitudes 60”N JANuARY OENSITY

Altitude 60”N 75”N 50 –

(km) lmw lWW 140”W 10”W 140”W

O.000 1.0025 1.0175 1.0102 1.0065 1.0205 +3* 40 -

2.000 7.7954 7.7292 7.8054 7.6963 7.7132 +2 E

~
4.000 5.9903 5.8559 5.9823 5.8488 5.8387 w
g 30 —
6.000 4.5435 4.3901 4.5261 4.3886 4.3801 ~

8.000 3.3967 3.2504 3.3693 3.2438 3.2318 b

a Ioo”w ~
1/
20 –
10.OOO 2.4991 2.3783 2.4764 2.3628 2.3493
~1
12.000 1.8230 1.7375 1.8194 1.7154 1.7069 \,\

14.000 1.3301 1.2695 1.3370 1.2438 1.2404 10 –

\J

16.000 9.7072 9.2705 9.8268 9.0069 9.0164 + 1

18.000
20.000
7.0753
5.1428
6.7589
4.9201
7.2240
5.3116
6.5134
4.6998
6.5550
4.7625
o~ -60 -40
PERCENT
-20
DEPARTURE
‘\\
\ ;.

FROM
I
20
1976
1 I
40
STANDARD
, I
60

22.000 3.7194 3.5758 3.8983 3.3648 3.4508

24.000 2.6695 2.5947 2.8495 2.3870 2.4933 Figure 14-16. Density-height profiles for the 60”N models at 10°, 100°,
26.000 1.9122 1.8840 2.0819 ~1.6852 1.8032 and 140°W.
28.000 1.3762 1.3723 1.5266 1.1941 1.3106
30.000 0.9949 1.0027 1.1233 D.8524 0.9573
these altitudes since the day-to-day variations in temperature
32.000 7.2258 7.3493 8.2951 6.1279 7.0263 +0
are in some cases as great or greater than the seasonal or
34.000 5.2703 5.4058 6.1459 4.4351 5.1805
latitudinal changes. Although these warmings and toolings
36.000 3.8645 3.9933 4.5687 3.2309 3.8367
occur throughout the arctic and subarctic region, the largest
38.000 2.8523 2.9621 3.4072 2.3701 2.8539
changes generally occur between latitude 60° and 75”N; they
40.000 2.1184 2.2084 2.5490 1.7519 2.1318 have been observed much more frequently at some longi-
42.000 1.5829 1.6570 1.9126 1.3043 1.5990 tudes than at others.
44.000 1.1896 1.2508 1.4391 2.9778 1.2041 A family of warm and cold atmospheric models typical
46.000 0.8990 0.9498 1.0858 0.7379 0.9102 of the region between 60° and 75”N has been prepared to
48.000 6.8305 7.2441 8.2146 5.5957 6.9055 – 1 provide an indication of the magnitude of the variations that
can occur in the vertical distributions of temperature, den-
50.000 5.2165 5.5378 6.2323 4.2534 5.2570
sity, and pressure in winter for altitudes up to 90 km. The
52.000 $.0037 4.2399 4.7520 3.2407 4.0156
54.i300 3.0829 3.2468 3.6422 2.4746 3.0742
*Power of 10 by which preceding numbers should be multiplied. 60 I I I 1 I ,
I
1 I , I I I
75”N JANUARY DENSITY PROFILES

\
\
Io”w \140”w
gitudinal variability is slightly smaller at 75”N than at 60”N.
The lowest mean monthly densities between 35 and 55 km I \

\
occur at 10OW at both 60° and 75°N.
\

\
14.2.1.5 Cold and Warm Winter Stratosphere/Meso-
[
sphere. In arctic and subarctic regions, sudden warmings
and toolings of the winter stratosphere and mesosphere pro-
duce large changes in the vertical and horizontal structure
\
of the atmosphere. Both the magnitude and altitude of max-
imum temperature and density fluctuations during major
warmings and toolings vary considerably. Some of the larg-
est changes in the vertical temperature profiles have been
10
I
observed in the upper stratosphere between 35 and 45 km.
The observed 35- to 45-km temperatures have a range of
0 I
-60
1 I
-40
PERCENT
1
-20
I I

DEPARTuRE
0
FROM
20
1 I
40
1976 STANDARD
1 I
60
1
roughly 85 K in winter compared with 20 K in summer. As
a result, mean monthly atmospheric models for the winter Figure 14-17. Density-height profiles for the 75”N models at 10° and
months are of limited value for specifying temperatures at 140”W.

14-23
atmospheric properties representative of one cold and three nades, falling spheres, and pressure gauges) taken at Poker
warm stratospheric regimes that occur at these latitudes are Flat, 64”N, Ft. Churchill, 59”N, Pt. Barrow, 71°N and West
given in Table 14-4. The temperature-height profiles are Geirinish 57”N during the period 1965 to 1976. Mean Jan-
shown in Figure 14-18. They are based on radiosonde, me- uary conditions at 60”N are assumed belo,w 9 km since the
teorological rocket, and experimental observations (gre- temperature-height profiles for this region during the various

Table 14-4. High-latitude thermodynamic properties of warm and cold winter stratospbere/mesosphere.

Altitude Model A Model B Model C Model D Model A Model B Model C Model D Model A Model B Model C Model D
(km) (Warm) (Warm) (Warm) (Cold) (warm) (warm) (w~) (Cold) (WanrI) (Warm) (Warm) (Cold)

Temperature (K) Density (kg/m3) Pressure (rob)

O.000 257.15 257.15 257.15 257.15 1.3742 1.3742 1.3742 1.3742 +0 1.0144 I .0144 1.0144 1.0144 +3*
2.000 255.94 255.94 255.94 255.94 1.0590 1.0590 1.0590 1.0590 7.7806 7.7806 7.7806 7.7806 +2
4.000 247.73 247.73 247.73 247.73 8.3466 8.3466 8.3466 8.3466 –1 5.9354 5.9354 5.9354 5.9354
6.000 234.13 234.13 234.13 234.13 6.6507 6.6507 6.6507 6.6507 4.4699 4.4699 4.4699 4.4699
8.000 220,54 220.54 220.54 220,54 5.2286 5.2286 5.2286 5.2286 3.3102 3.3102 3.3102 3.3102

Io.000 217.15 217.15 217.15 219.39 3.8811 3.8811 3.8811 3.8460 2.4192 2.4192 2.4192 2.4221
12.m 215.16 215.66 214.17 222.38 2.8495 2,8534 2.8716 2.7866 I .7661 1.7664 1.7654 1.7789
14.000 211.17 212.66 208.18 223.15 2.1163 2.1050 2,1394 2.0461 1.2829 1.2850 1.2785 1.3106
16.000 207.19 209.68 202.21 223.15 1.5577 1.5465 1.5806 1.5080 9.2646 9.3087 9.1750 9.6600 + I
18.000 203.20 206.69 196.23 223.15 1.1399 1.1314 1.1574 1.1116 6.6496 6.7132 6.52CH3 7.1210

20.000 199.22 203.70 196.15 223.15 8.2929 8.2419 8.1870 8.1965 –2 4.7425 4.8194 4.6097 5.2503
22,m 195.24 200.72 204.93 222.17 5.9953 5.9767 5.5825 6.0670 3.3601 3.4436 3.2840 3.8693
24.000 194.56 201.50 213.88 221.18 4.2395 4.2401 3.8671 4.4895 2.3677 2.4525 2.3742 2.8481
26.000 197.54 206.46 222.82 220.18 2.9535 2.9667 2.7200 3.3130 1.6747 1.7583 1.7397 2.0940
28.000 200.51 211.43 231.75 219.19 2.0692 2.0939 1.9402 2.4439 1.1910 1.2708 1.2908 1.5377

30.000 203.49 216.39 240.68 218.20 1.4576 1.4901 1.4021 1.8006 0.8514 0.9256 0.9687 1.1278
32.000 215.87 227.36 253.37 217.21 0.9942 1.0447 1.0122 1.3251 6.1609 6.8189 7.3624 8.2621 +0
34,000 228.75 238.66 266.25 214.35 6.9184 7.4429 7.4226 9.8119 –3 4.5429 5.0991 5.6730 6.0374
36.000 241.63 248.52 279.13 211.38 4.9118 5.4164 5.5241 7.2403 3.4068 3.8640 4.4261 4.3933
38.000 254.49 256.44 275.59 208.41 3.5505 4.0156 4.3868 5.3208 2.5938 2.9560 3.4704 3.1832

40.m 267.35 264.35 270.64 211.39 2.6084 3.0049 3.4878 3.8002 2.0018 2.2802 2.7096 2.3060
42,0(XI 282.87 272.26 265.70 214.95 1.9283 2.2683 2.7617 2.7224 1.5657 1.7727 2,1063 1.6798
44.OCKI 298.67 278.41 260.76 219.78 1.4478 1.7360 2.1775 I .9506 1.2413 1.3874 1.6299 1,2306
46.000 296.64 280.65 255.83 226.69 1,1640 1.3537 1.7093 1.3979 0.9912 1.0906 1.2553 0.9097
48.000 285.79 278.24 250.89 233.60 0.9585 1.0732 1.3357 1.0121 7.8635 8.5719 9.6204 6.7872 – 1

50.ti 274.94 274.30 246.43 240.50 0.7834 0.8530 1.0370 0.7398 6.1836 6.7169 7.3358 5.1080
52.000 264.10 270.36 242.89 245.76 6.3533 6.7591 7.9903 5.4900 –4 4.81ti 5.2456 5.5710 3.8730
54.000 253.27 266.42 239.34 250.68 5.1074 5.3380 6.1340 4.1041 3.7132 4.0823 4.2143 2.9533
56.~ 244.48 260.58 235.80 255.60 4.0350 4.2290 4.6913 3.0861 2.8317 3.1633 3.1754 2.2643
58.000 240.55 254.29 232.26 255.87 3.1084 3.3379 3.5739 2.3716 2.1464 2.4365 2.3828 1.7419

60.000 236.61 248.00 228.72 254.89 2.3847 2.6194 2.7118 1.8302 1.6197 1.8647 1.7804 1.3391
62.000 232.69 241.71 225.18 253.91 1.8217 2.0431 2.0492 1.4112 1.2168 1.4176 1.3246 1.0285
64.000 228.76 235.43 221.65 252.92 1.3855 1,5835 1.5419 1.0872 0.9098 1.0701 0.9810 0.7893
66.000 224.83 229.15 218.11 251.94 1.0489 1.2190 1.1551 0.8369 6.7699 8.0188 7.2325 6.0527 –2
68.ooil 220.91 222.87 214.59 250.28 7.9039 9.3184 8.6146 6.4526 –5 5.0122 5.%16 5.3064 4.6359

70.m 216.99 216.60 211.32 248.32 5,9267 7.0697 6.3867 4.9718 3.6917 4.3957 3.8742 3.5440
72.000 213.07 210.33 209.36 246.36 4.4216 5.3212 4.6899 3.8235 2.7045 3.2128 2.8186 2.7040
74.000 209.16 204.07 207.40 244.40 3.2815 3.9716 3.4346 2.9348 1.9702 2.3266 2.0448 2.0590
76.000 205.35 199.73 205.45 242.45 2.4210 2.9126 2.5084 2,2482 1.4271 1.6699 1.4793 1.5646
78.000 202.41 196.60 203.49 239.80 1.7700 2.1123 1.8268 1.7234 1.0284 1.1921 1.0671 1.1863

80.000 199.48 193.47 201.54 236.67 1.2884 1.5244 1.3266 1.3195 7.3779 8.4664 7.6749 8.9645 – 3
82.000 198.15 190.35 199.59 233.55 0.9268 1.0945 0.9605 1.W8 5.2718 5.9806 5.5034 6.7501
84.000 198.15 187.23 197.63 230.43 6.6205 7.8172 6.9347 7.6561 –6 3.7657 4.2013 3.9342 5.0642
86.CQO 198.15 184.68 196.65 227.31 4.7301 5.5370 4.9689 5.8011 2.6904 2.9354 2.8049 3.7852
88.000 198.15 182.73 196.65 223.82 3.3802 3.8948 3.5417 4.3861 1.9226 2.M30 1.9992 2.8181

90.000 198.15 180.79 196.65 219.93 2.4160 2.7299 2.5250 3.3072 1.3742 1.4167 1.4253 2.0879

*Power of 10 by which preceding numbers should be multiplied.

14-24
 STANDARD AND REFERENCE ATMOSPHERES
90 ~ 1 1 1 I I I I I I I
1’ I
P
~.
I 1 1 I 1 1 I 1 I I
f Y’
80- —.- MODEL C WMOOEL A——-
WMODEL B —-—--
------ MODEL D w~ —
. —60”N JAN MEAF ror 60*N
70 -

60 -

~
- 50 -
w ,.~
n .. .
)
3 ,,
1-
~ 40 -

d ‘,

30 -

20 -

10 -

o~ 40
I I I I I
60
300
180 200 220
E FROM20~
STO
TEMPERATURE (K)

Figure 14-19. Density profiles (warm) associated with extreme temper-

Figure 14-18. Temperature-height profilesassociated
withextremewarm atures in the upper stratosphere and mean January condi-
andcoldregimesinthewinterstratosphere
andmeaoaphere tions at 60”N. Circled points form an envelope of high and
near60”N. low densities equaled or surpassed 5% of time.

warmings andcoolings are not significantly different from profile in Figures 14-19 and 14-20. The densities are por-
the mean Janu~60°N atmosphere. trayed as percent departures from the 1976 Standard At-
The three warm models could all occur during various mosphere. Envelopes of the high and low values of density
stages of one large-scale warming. However, available ob- which are equaled or surpassed 570 of the time at 60”N in
servations indicate that a temperature of 300 K at 45 km is January are also shown. ~ey are envelopes rather than
equaled or exceeded 2% of the time at West Geirinish and realistic profiles since 570 values do not occur simultane-
0.4% of the time at Ft. Churchill during January, whereas ously at all altitudes. The density profiles for the warm and
a temperature of 280 K at 45 km is equaled or exceeded cold models illustrate the negative correlations that exist
10% of the time at West Geirinish and 4% of the time at between the densities at various levels in the atmosphere
Ft. Churchill. A temperature of 280 K near 36 km is equaled [Quiroz, 1971; Labitzke, 1971; Cole, !972]. For example,
or exceeded 0.6% of the time at West Geirinish and 0.170 when the density is much less than the mean monthly value
of the time at Ft. Churchill. Frequencies of occurrence were at altitudes between 25 and 40 km (Figure 14-20), it is
obtained by plotting the observed temperature distributions greater than the mean value between 45 and 75 km. In most
on probability paper. cases the departures of density from the monthly mean fall
The cold profile, Model D, is based on an average of within the 5% envelope. However, as shown in Figure
five observations in which the temperature at 45 km was 14-20, density profiles associated with an extreme winter
within 2° of 223 K. Observed data indicate that a temper- warming or cooling will approach both the 570 maximum
ature of 223 K or colder occurs at Ft. Churchill 6% of the and 570 minimum values at different altitudes.
time, at West Geinnish 4% of the time, and at Poker Flat The altitudes of the maximum density departures from
9% of the time in January. the monthly mean are related to the altitudes of maximum
The portions of the temperature-height profiles between temperature deviations in that the maximum density depar-
55 and 85 km are based on estimates obtained by using tures are roughly 10 to 20 km above the maximum tem-
interlevel temperature correlations with the temperatures perature deviations. For example, the largest positive den-
adopted at 40, 45, and 50 km. sity departure for profile C (Figure 14-20) occurs near 49
The density profiles associated with both the warm and km, whereas the maximum stratospheric temperature, 280
cold models are provided along with the mean January 60”N K for profile C (Figure 14-18), is at 36 km. The largest

14-25
CHAPTER 14
90
I I I 1/ I I I I I I 1 I the U.S. Standard Atmosphere Supplements, 1966 [COESA]
11 were published. This publication included models up to
~ ~~ g—_._=...-
---
80 — 0
1000 km. The lower thermosphenc models were based on
WAN MNJARY
far60” N work by Champion and the upper thermospheric models on
@
work by Jacchia. An updated set of CO~PAR models was
m — o ,1 published in 1972 as CIRA J972 (COSPAR). In these models
0/ Champion and Schweinfurtb [1972] prepared the Mean CM
I
60 – @\ for altitudes between 25 and 500 km and Jacchia [1971]
\
prepared a set of models for the altitude range 110 to 2000
2\\
km. More recent models include the U.S. Standard At-
g 50 –
mosphere, 1976 (COESA), and the Jacchia [1977] model.
$ o The preceding models have all been based principally
~ 40 – 0 on density data from orbital drag and in situ measurements
of density and composition. Recently, models have been
developed which are also based on airglow temperatures
30 –
r (DTM) [Barlier et al., 1978] or incoherent radar scatter
temperatures (MSIS) [Hedin et al., 1977]. A number of
20 – theo~tical models m being developed, including the NCAR
I
thermospheric general circulation model by Roble and col-
leagues [Dickinson et al., 1981]. This is a three-dimensional
Io–
model of the global neutral gas temperatures and circulation
of the thermosphere. Rees and colleagues [Fuller-Rowell
1 I I I I I I I I I I I
-w - 4G -20 0 20 40 60 and Rees, 1980] are also developing a three-dimensional
ffWFNT WPARTI lXf F- ~ STD
global model of the thermosphere. They start with a realistic
steady-state model and then progressively modify it in re-
Figure 14-20. Density profiles associated with extreme temperatures in sponse to the sources and sinks of energy and momentum
the upper stratosphe~ and mean January conditions at ~.
Circled points form an envelope of high and low densities
and to the winds that result from the various driving forces
equaled or surpassed 5% of time. redistributing mass, momentum, and energy. Important
components of theoretical models are the amplitudes and
phases of the tidal effects. The principal tides are the solar
negative density departure for the same profile occurs near diurnal, semidiumal and terdiumal. The magnitudes of these
33 km and its minimum stratospheric temperature, 196 K tides and their effects on density and composition are re-
is at 18-20 km. viewed by Champion [1981].

14.2.2.1 Development of Reference Almaa@. Since

14.2.2 High Altitude Reference Atmospheres the earliest upper atmosphere models, it has been customary
to define them in part by means of temperature profiles
Although there were early high altitude model atmos- which, by means of the appropriate physical relations, yield
pheres based on upward extrapolation of rocket data, they density profiles, and in the more modem models, compo-
were not accurate and the first official U.S. models of the sition and other properties.
satellite era, including density values derived from satellite At this point we will review the temperature profiles of
drag, were the ARDC Model Atmosphere, 1959 [Minzner, some models of particular interest. Figure 14-21 shows the
1959] and the U.S. Standard Atmosphere, 1962 [COESA]. temperature profiles of the U.S. Standard 1962, Mean CIRA
These models accurately represented thermosphenc density 1965, Mean CIRA 1972 and U.S. Standard 1976. It can be
at the time the observations were made, but since the ob- seen that as time progressed the mean model temperatures
servations were made at the peak of solar cycle 19 which in the thermosphere have become lower. There is a good
had unusually high activity, they represented conditions that reason for this which can be understood by referring to
are not frequently observed. Figure 14-22. This figure shows a plot of the mean annual
The next generation of reference atmospheres started number of sunspots from 1820 to 1976. The 1l-yr cycle of
with the COSPAR International Reference Atmosphere activity can be seen and also the wide variation in the number
(CIRA) 1965 (COSPAR). These atmospheres included a of sunspots at the cycle peaks. Table 14-5 shows that the
Mean CIRA between 30 and 300 km prepared by Champion succeeding models, which are based on satellite orbital drag
and a set of models for 120 to 800 km prepared by Harris and in situ measurements, have sampled the atmosphere
and Priester. The latter are semi-theoretical models but with starting from very non-typical conditions in 1957 to in-
some free parameters whose values were chosen so that the creasingly representative conditions with time. In addition,
models reproduced observed densities. The following year Slowey [1979] has done a study using both sunspot numbers

14-26
 STANDARD AND REFERENCE ATMOSPHERES
1 Table 14-5. Average exospheric temperature.

‘“r
I
1
---US STD 62 I AVERAGE
MEAN CIRA 65 ; t
— MEAN CIRA 72 1 1
I
US STANDARD ATMOSPHERE 1962 1500 K
US STD 76 !
- RROPOSED I (3 YRS)
MEAN CIRA 1 I
MEAN CIRA 65 12W K
I MEAN CIRA 72 1000 K
I
I
I
I US STANDARD ATMOSPHERE 1976 1000 K
1
I I (1* SOLAR CYCLES)
I
I
/’ SPECIAL ~TU~Y (12 SOLAR CYCLES) 882 K
i i
fl /
I
back to 1847 (12 cycles) and tie Jacchia 1971 model to
/“
/
i
/’
/’” determine relative values and has found that the expected
,/’”
.,” mean exospheric temperat~ averaged over all conditions
./ ---
----- for this pe~od would be 882 K. A“histogram showing the
/ predicted exospheric temperature distribution is given in
Figure 14-23. The exact mean tempe~ture will depend on
the time period chosen (for example, including the Maundet
01 1 1 ,

minimum would have a profound effect) and the model used

0 500 1000 1500
KINETIC TEMPERATURE(K) to determine relative values. Hcrwever, it is recommended
that 900 K be used for the exospheric temperature for the
Figure 14-21. Kinetic temperature profiles forseveral standard andref- next Mean CIRA as being appropriate to this phase of solar
eknee atmosphem, including a pm- new inears CIRA. activity. The corresponding density profiles of the four pre-
vious standard or mean models and that of the proposed
new model are shown in Figure 14-24.
The most up-to-date set of approved reference #atmo-
spheresfor altitudes above 90 km is that con@ined in C1~
72. There are plans to prepare revised versions of both CIRA
and the U.S. Standard Atmosphere Supplements. The latest
published version of the latter is the”U.S. Starufard Atmos-
phere Supplements, 1966.
The CIRA 1972 models up to 120 km were prepared
by Groves of University College, London, England and the
Figure’ 14-22. The mean annual number of sunspots from 1820 to 1976. Air Force Catnbridge Research Laboratory (AFCRL). An
independent set of models for the region 110 to 2000 km
were prepared by Jacchia of the Smithsonian Astrophysical

\ ---uSSTD62
MEAN CIRA 65
-- MEAN CIRA 72
— US STO 7S
‘-~ PR~SED MEAN CIRA

OL
10”’410-” 10-’2 ,0”’( ,0”’0 ,0”9 ,0”8 ,0” ‘ ,0”6 ,0-’ ,0-” 10”’ 10

OENSITY (kg /m3)

EXOSPHERIC TEMPERATURE (Te, )

F]gure 14-23. Histogram of frequency per degree of predicted exospheric Figure 14-24. Density profiles of several standard and reference atmo-
tem~ratures for a mean solar cycle. spheres, including a proposed new mean CIRA.

14-27
CHA~ER 14
observatory. Since these two sets of models do not, in where
general, match at i 10 km, Champion and Schweinfurth of
AFCRL developed the Mean CIRA which is continuous p= pressure at altitude z
from 25 to 500 km and applies under specified conditions. p, = pressure at reference altitude z, ~

M. . sea level value of mean molecular weight = 28.96

14.2.2.2 Mean CIRA Reference Atmwphere. The ba-
sis of the mean reference atmosphere is as follows: Between R= 8.31432 x 107 ergs K-’ g mol-’
25 and 75 km the model represents annual mean conditions
g= acceleration due to gravity at 30° latitude
for latitudes near 30°. Between 120 and 500 km the model
corresponds to diurnal, seasonal, and semi-annual variation TM = molecular-scale temperature
average conditions for a latitude near 3@ and a SOIWflUX
~ of 145 x l&22 W m-2 HZ-l. Betweeh 75 and 120 km a — MOT
model has been develo~d which provides a smooth con- M
nection between the lower and upper sections of the mean
atmosphere. This atmosphere contains kinetic temperature, where
molecular-scale ‘temperature, density, pressure, density and
pressure scale heights, mean tnolecular weight, densities of M= mean molecular weight
major constituents, and total number densities.
T= kinetic temperature.
The data used to develop the model between 25 and 75
km were the values at 30° latitude of the annual mean pres-
sure at 25 km and the annual mean temperature between 25 The model between 75 and 120 km was constrained to
and 75 km. The pressure equation was integrated numeri- provide a transition between the low altitude model and the
cally. high altitude model. These two models are not only different
but are functions of different parameters. Thus a compro-
mise was devised. A temperature profile was developed with
a constraint that it yield a specified density value at 120
km. The equations and method used to determine the com-
position are given in Champion and Schweinfurth [1972].

Table 14-6. Mean Reference Atmosphere SWcture Parametem 25 to 120 km.

Mol Number Pressure

Height Temp Density Pressure Density Scale g
km K kg/m3 N/m2 m-’ Ht(km) mlsz

25 221.7 3.899E – 02 2.483E + 03 8.111E+23 6.55 9.716

30 230.7 1.774E – 02 1.175E+03 3.690E -t 23 6.83 9.701
35 241.5 8.279E – 03 5.741E+02 1.722E +23 7.16 9.686
40 255.3 3.972E – 03 2.911E+02 8.265E + 22 7.58 9.671
45 267.7 i .995E – 03 1.535E+02 4. 148E+ 22 7.96 9.656
50 271.6 1.057E – 03 8.241~+01 2.198E+22 8.09 9.641
55 263.9 5.821E–04 4.406E+01 1.21 OE+22 7.87 9.626
60 ‘249.3 3.206E-04 2.296E+01 6.669E+21 7.45 9.611
65 232.7 1.718E–04 1.146E+01 3.568E+21 6.96 9.596
70 216.2 8.770E – 05 5.#5E+O0 1.822E+21 6.48 9.581
75 205.0 4.178E–05 2.460E+O0 8.696E + 20 6.15 9.566
80 195.0 1.905E-05 1.067E+O0 3.964E+20 5.86 9.551
85 185.1 8.337E–06 4.426E–01 1.736E +20 5.57 9.536
90 183.8 3.396E–06 1.795E–01 7.087E + 19 5.54 9.521
95 190.3 1.343E – 96 7.345E – 02 2.808E+ 19 5.75 9.506
lm 203.5 5.297E – 07 3.090E – 02 1.125E+19 6.16 9.492
105 228.0 2.173E–07 1.422E – 02 4.768E+ 18 6.91 9.477
110 265.5 9.661E–08 7.362E – 03 2.182E+ 18 8.06 9.462
115 317.1 4.645E-08 4,236E – 03 1.076E +18 9.64 9.448
120 380.6 2.438E–08 2.667E-03 5.772E+ 17 11.58 9.433
—. ——

14-28
 STANDARD AND REFERENCE ATMOSPHERES
Table 14-7. Kinetic Temperature and Composition of the Mean Reference Atmosphem, 75 to 120 km

Height Temp Mean Log N(N*) Log N(02) Log N(0) Log N(Ar) Log N(HE) Log N(Oq)
~-’ ~-3
km K Mol Wt m-’ m-’ m-’ m-’

75 205.0 28.96 20.832 20.261 18.910 15.659’

80 194.9 28.95 20.492 19.914 16.794 18.563 15.315 14.495
85 184.9 28.93 20.135 19.550 17.144 18.199 14.954 14.098
90 i83.4 28.89 19.747 19.153 17.220 17.802 14.>65 13.425
95 189.3 28.81 19.346 18.739 17.280 17.389 14.178 12.665
100 199.4 28.37 18.940 18.299 17.618 16.945 13.928 12.099
105 216.6 27.51 18.556 17.823 17.647 16.435 13,811 11.178
110 245.1 26.73 18.200 17.398 17.509 15.961 13.711 10.132
115 285.2 26.05 17.872 17.036 17.332 15.539 13.620 9.098
120 334.5 25.45 17.579 16.734 17.153 15.173 13.538 8.167

The exospheric temperature for the conditions specified density, density scale height, pressm, p~ssure scale height,
above is 1000 K. The Jacchia models were recomputed using and acceleration due to gravity for the altitude mnge 12&500
the acceleration due to gravity for 30° latitude. The original km are given. Table 14-9 contains the corresponding values
models use a value valid only near latitude 45°. The model of kinetic temperature, mean molecular weight, number
values were then changed at all altitudes so that fiey matched density, and log number density of NZ,OZ, O, Ar, and He
the values at 120 km for the intermediate altitude model. for the altitudes 120–500 km.
The properties of the Mean Reference Atmosphere are Some of the properties are illustrated in Figures 14-25
presented in abbreviated forin in Tables 14-6 to 14-9. (The to 14-30. Figure 14-25 shows the kinetic temperature of the
complete tables are provided in CIRA, 1972 [COSPAR, mean atmosphere plus curves indicating low extreme and
1972]. Table 14-6 contains values of molecular-scale tem- high extreme temperatures whose frequency of occurrence
perature, density, pressure, number density, pressure scale is 1Yoor less. Figure 14-26 shows the pressure scale heights
height, and acceleration due to gravity over the altitude
range 25–1 20 km. Table 14-7 contains values of kinetic
temperature, mean molecular weight and log number den- 500
sities of N2,02,0, Ar, He, and 03 over the altitude range
75–120 km. In ~~ble 14-8 molecular-scale temperature,

LOW MEAN HIGH

400
ExTREME CIRA EXTREME

o 400 800 1200 1600 2000

0 10 20 30 40 50 60 70
TEMPERATuRE ( K)
PRES$URE SCALE HEIGHT (km)
Figure 14-25. Mean CIRA temperatures and low extreme and high ex-
treme temperatures. Figure 14-26. Pressure scale heights of the mean CIRA.

14-29
CHAPTER 14
Table 14-8. Mean Reference Atmosphere Stmcture Parametem 120 to 5W b.

Molec Density Pressure

Height Temp Density Sc$e Pressure Scale g
km K kg/rn3 Ht (km) N/mz Ht(km)’ m/s*

120 380.6 2.440E-08 8.17 2.666E-03 11.58 9.433

125 452.3 1.382E –08 9.51 1.795E –03 13.79 9.418
130 526.9 8.484E –09 11.05 1.283E –03 16.08 9.404
135 600.9 5.563E –09 12.70 9.597E–04 18.37 9.389
140 672.4 3.845E –09 14.44 7.423E–04 20.59 9.375
145 739.8 2774E – 09 16.21 5.891E–04 22.69 9.360
150 802.7 2.070E – 09 17.98 4.770E–04 24.66 9.346
155 861.0 1.587E – 09 19.70 3.923E–04 26.49 9.332
160 914.8 1.244E-09 21.37 3.267E-04 28.19 9.317
165 9@.7 9.927E– 10 22.% 2.749E–04 29.77 9.303
170 1011.1 8.040E– 10 24.48 2,334E–04 31.25 9.289
175 1054.4 6.5~3E– 10 25.92 1.996E-04 32.64 9.274
180 1095.0 5.464E– 10 27.29 1.718E–04 33.95 9.260
185 1133.2 4.568E– 10 28.W 1.468E–04 35.19 9.246
190 1169.3 3.850E– iO 29.85 1.292E–04 36.36 9.232
195 1203.4 3.267E– 10 31.05 1.129E–04 37.48 9.218
200 1235.8 2.789E– 10 32.21 9.8%E–05 38.55 9.204

210 1295.9 2.066E– 10 34.40 7.686E – 05 40.55 9.176

220 1350.3 1.55~E– 10 36,45 6.039E –05 42.38 9.148
230 1399.6 1.192E– 10 . 38.39 4.792E –05 44.06 9.120
240 1444.4 9.246E– 11 40.21 3.834E –05 45.61 9.092
250 1484.9 7.248E–11 41.92 3.090E-05 47.03 9.065
260 1521.5 5.735E–11 43.52 2.505E – 05 48.33 9.037
270 i554.6 4.576E– 11 45.03 2.042E-05 49.54, 9.010
280 1584.5 3.677E– 11 46.43 1.673E – 05 50.64 8.983
290 1611.5 2.974E–11 47.73 1.376E – 05 51.& 8.955
3W 1636.0 2.418E–11 4!3.95 1.136E–05 52.60 8.928
310 1658.1 1.976E–11 50.07 9.405E-06 53.48 8.902
320 1678.2 1.621E–11 51.12 7.812E–~ 54.29 8.875
330 1696.5 1.336E–11 52.09 6.507E–06 55.05 8.848
340 1713.4 1.104E-11 52.98 5.432E –06 55.76 8.822
350 1728.8 9.158E–12 53.81 4.545E–06 56.43 8.795
360 1743.2 7.615E– 12 54:59 3.811E–06 57.07 8.769
370 1756.7 6.348E– 12 55.31 3.202E–M 57.69 8.743
380 1769.5 5.304E– 12 55.98 2.694E–06 58.28 8.717
390 i781,8 4.MIE– 12 56.60 2.272E–06 58.86 8.691
400 1793.7 3.725E– 12 57.20 1.918E–06 59.43 8.665
410 1805.4 3.130E–12 57.75 1.622E–06 60.00 8.639
420 1817.2 2.635E– 12 58.28 1.374E–06 @.57 8.614
430 1829.1 2.221E– 12 58.79 1.166E-06 61.14 8.588
440 1841.3 1.875E– 12 59.26 9.910E–07 61.74 8.563
450 1854.0 1.585E–12 59.73 8.435E – 07 62.35 8.537
460 1867.4 1.341E– 12 60.18 7.191E–07 62.98 8.512
470 1881.6 1.137E–12 ~.63 6.140E-07 63.65 8.487
480 1896.9 9.644E– 13 61.08 5.252E – 07 64.36 8.462
490 1913.4 8.192E– 13 61.52 4.500E-07 65.11 8.437
5oi3 1931.4 6.967E– 13 61.95 3.863E – 07 65.91 8.412

14-30
 STANDARD AND REFERENCE ATMOSPHERES
Table 14-9. Kinetic Temperatu~ and Composition of the Mean Reference Atmosphere, 120 to 500 km.

Number
Height Temp Mean Density hg N(N*) Log N(02) Log N(0) Log N(Ar) Log N(HE)
M-3 m-3 m-3
h K Mol Wt m-3 m-3 m-3 L

120 334.5 25.45 5.772E+ 17 17.579 16.734 17.153 15.173 13.538

125 389.7 24.95 3.336E+ 17 17.322 16.449 16.978 14.835 13.469
130 445.4 24.48 2.087E+ 17 17.098 16.203 16.826 14.541 13.410
135 499.0 24.05 1.393E+ 17 16.903 15.987 16.693 14.284 13.358
140 549.0 23.65 9.793E+ 16 16.731 15.796 16.577 14.056 13.314
145 594.5 23.27 7.178E+ 16 16.577 15.624 16.474 13.850 13.275
150 635.2 22.92 5.439E + 16 16.437 15.469 16.381 13.663 13.241
155 671.3 22.58 4.233E + 16 16.308 15.325 16.298 13.490 13.212
160 703.1 22.26 3.366E+ 16 16.189 15.192 16.221 13.328 13.185
165 731.3 21.95 2.723E+ 16 16.077 15.066 16.150 13.176 13.161
170 756.2 21.66 2.236E+ 16 15.971 14.947 16.083 13.031 13.139
175 778.4 21.38 1.857E+16 15.870 14.834 16.020 12.892 13.118
180 798.1 21.11 1.559E+ 16 15.773 14.724 15.960 12.758 13.099
185 815.9 20.85 1.319E+ 16 15.679 14.619 15.902 12.629 13.081
190 831.9 20.60 1.125E+ 16 15.589 14.517 15.847 12.504 13.064
195 846.3 20.37 9.661E+ 15 15.501 14.418 15.794 12.382 13.048
200 859.3 20.14 8.342E+ 15 15.415 14.321 15.742 12.262 13.033

210 882.0 19.71 6.312E+ 15 15.250 14.133 15.642 12.031 13.004

220 900.7 19.32 4.856E+ 15 15.090 13.952 15.547 11.807 12.977
230 916.4 18.96 3.788E+ 15 14.936 13.777 15.456 11.590 12.951
240 929.4 18.63 2.988E+ 15 14.785 13.606 15.367 11.378 12.927
250 940.2 18.34 2.380E+ 15 14.638 13.438 15.281 11.171 12.903
260 949.3 18.07 1.912E+ 15 14.494 13.274 15.197 10.967 12.881
270 956.8 17,82 1.546E+ 15 14.352 13.112 15.114 10.766 12.859
280 963.1 17.60 1.258E+15 14.212 12.953 15.033 10.567 12.837
290 968.4 17.40 1.029E + 15 14.073 12.795 14.953 10.371 12.816
300 972.8 17.22 8.456E+ 14 13.937 12.639 14.874 10.177 12.796
310 976.5 17.06 6.977E+ 14 13.801 12.485 14.796 9.984 12.776
320 979.7 16.91 5.776E+ 14 13.667 12.332 14.719 9.793 12.756
330 982.3 16.77 4.798E+ 14 13.533 12.179 14.642 9.604 12.736
340 984.6 16.64 3.997E+ 14 13.401 12.028 14.566 9.415 12.717
350 986.5 16.52 3.338E+14 13.269 11.878 14.490 9.228 12.698
360 988.1 16.42 2.794E+ 14 13.139 11.729 14.415 9.041 12.679
370 989.5 16.31 2.344E+ 14 13.008 11.580 14.341 8.856 12.660
380 990.7 16.21 1.970E+ 14 12.879 11.432 14.267 8.672 12.641
390 991.7 16,12 1.659E+14 12.750 11.285 14.193 8.488 12.622
400 992.6 16.03 1.400E+ 14 12.621 11.138 14.119 8.305 12.604
410 993.4 15.93 1.183E+14 12.494 10.992 14.046 8,123 12.585
420 994.1 15.84 1.00IE+ 14 12.366 10.847 13.973 7.941 12.567
430 994.7 15.75 8.493E + 13 12.239 10.702 13.901 7.761 12.549
440 995.2 15.65 7.213E+ 13 12.113 10.558 13.829 7.581 12.531
450 995.7 15.55 6.136E+ 13 11.987 10.414 13.757 7.401 12.513
460 996.1 15.45 5.229E+ 13 11.862 10.271 13.685 7.222 12.495
470 996.4 15.34 4.464E+ 13 11.737 10.128 13.613 7.044 12.477
480 996.8 15.22 3.817E+ 13 11.612 9.986 13.542 6.867 12.459
490 997.0 15.09 3.269E+ 13 11.488 9.844 13.471 6.690 12.441
500 997.3 14.95 2.806E+ 12 11.364 9.703 13.400 6.513 12.423

14-31
CHAPTER 14
COLD TEMPERATURES WARM TEMPERATuREs
I I I I MEAN I I I I I I
90% MIEOIAN 1%
90 / MEOIAN “RA ‘. lo%. \
-.
t\ 99?. /’ \ . \\

400

z 300
:

:
;
: zoo

100

I [ I I I 1
30

I 1
0 1, / I 1
-1
I J
,.. ,+ ,.. ,, ,.. ,O ,.-O ,.-, ,.-4 ,.-,
130 150 170 190 210 230 250 270 29o
DENSITY (kg/m’]
TEMPERATuRE Tm ( K)

Figure 14-27. Mean CIRA densities and curves of extreme densities Figure 14-29. Mean CIRA temperatures and temperatures which are ex-
ceeded 50%, 10%, 1% of the time during warmest months,
and temperatures exceeded 50%, 90%, and 99% of the
time during coldest months, at latitudes between O“ and
of the mean atmosphere as a function of altitude. Figure 80”N
14-27 contains low extreme, high extreme, and mean den-
sity values. Total number densities and densities of Nz, 02,
0, 0s, Ar, He, and H are given in Figure 14-28. Figure pending on the availability of data. Above 90 km the data
14-29 contains the Mean CIRA temperature profile, median are primarily from rocketbome falling sphere, grenade and
warm temperatures and those exceeded 10% and 1% of the chemical release techniques. Table 14-10 contains average
time and, similarly, median cold temperatures and those values of zonal (W - E) winds at latitudes from the equator
above which 90% or 99%, respectively, of the temperatures to 50° for the first day of each month. Data used are from
lie. The corresponding density curves are shown in Figure all longitudes, with southern hemisphere data shifted by six
14-30. months. Mean temperature values at latitudes from the equa-
tor to 70° are given in Table 14-11. Less accurate values
14.2.2.3 Reference Atmospheres 90 to 120 km. These (based on few data points) are indicated by an asterisk. Table
empirical atmospheres are based on a report entitled, “At- 14-12 gives the corresponding pressure values and Table
mospheric Structure and its Variations in the region from 14-13 the density values. - -
25 to, 120 km” by Groves [1971]. In this section only the
properties above 90 km are presented. The upper altitude
of the models varies with latitude and time of year, de- 1 , r
IEOIAN 10% I%
‘\
~.
. ~.
---
r 1 1 1 1 r 1 1 1 )
500 - ., \,*\” I
TOTAL —
‘. \ /’
‘. “’\, \
1’
‘. 1
400 - ‘\
/’
‘\
‘\ 1’
z ‘. ,. I
. HO . ~~~~ I
‘\ ,, ~ l.,
/’
I
- 300 - ‘.
u
n ‘\
.
:
? ‘\ :, \.,,
~ ~oo - ‘.;\ /’
‘.: ..1
a . .:. l!
.,.
1’

I 00 - 1’
... , t
20 40 60 60

I
10*
I
109
1
1o’” lo”
I I
10”
I
10”
I
10”
I
,.20
DENSITY- PERCENT DEPARTURE FROM CIRA 1972

NUMBER DENSITY (M-3) Figure 14-30, f3ensities relative to Mean CIRA exceeded 50%, 10%, and
1% of the time during months with highest densities, and
densities exceeded 50%, 90%, and 99% of the time during
Figure 14-28. Total number densities and densities of N2, 02, 0, 03, A, months with lowest densities, at latitudes between 0° and
He, and H. 80”N

14-32
 STANDARD AND REFERENCE ATMOSPHERES
Table 14-10. W-E Winds 90 to 130 km based on data from all longitudes with S. Hemisphere data shifted six months in time. Winds to the east are
positive in m/s. Values apply to the first day of each month.

km Jan Feb Mar Apr May Jun Jul Aug Sep Ott Nov Dec

O Degrees N
90 –9 –31 – 22 – 34 – 15 –23 –9 –31 – 22 – 34 – 15 – 23
95 – 28 – 23 – 13 –21 – 22 – 32 – 28 – 23 – 13 –21 – 22 – 32
100 –28 –3 7 –3 – 22 – 29 –28 –3 7 –3 – 22 – 29
105 –7 22 26 3 –11 –11 –7 22 26 3 –11 –11
110 17 37 30 –13 10 14 17 37 30 – 13 10 14
115 31 37 20 – 29 32 33 31 37 20 – 29 32 33
10 Degrees N
90 – 14 – 30 – 36 –54 –5 –7 o –21 4 3 –2 – 22
95 – 39 –41 – 43 – 52 –15 –8 –14 –5 15 14 –5 – 36
100 –40 – 29 – 27 – 25 – 19 1 –21 19 28 15 – 10 – 43
105 – 20 8 7 –3 – 10 5 1 36 31 2 – 10 –21
110 1 41 27 – 20 9 3 22 40 19 – 17 2 20
115 18 49 27 –44 30 13 30 32 1 – 27 19 41
20 Degrees N
90 –8 – 29 – 23 – 32 16 17 12 5 19 22 19 –6
95 – 23 – 36 – 34 – 40 6 21 –6 23 25 28 18 – 18
100 – 28 – 16 – 48 –36 –2 22 – 12 45 29 15 9 – 32
105 – 18 17 –41 –28 -3 11 –3 55 20 –11 –3 – 24
110 0 34 –21 –31 5 2 7 46 –1 – 27 :8 5
115 20 27 -7 – 36 19 8 12 26 –19 –25 –1 31
30 Degrees N
90 5 –18 15 0 29 34 22 14 24 27 45 11
95 0 – 19 17 – 16 25 42 1 31 12 21 38 –3
100 –8 5 0 – 30 18 40 –1 50 6 –5 21 – 19
105 –9 39 –8 –31 4 15 0 63 –6 –39 –2 – 25
110 2 36 1 –26 –4 – 12 0 52 – 28 – 43 – 19 –11
115 11 0 12 –22 –1 –18 –5 26 –44 –21 – 30 16
120 7 – 27 10 –21 -1 –7 –7 –1 –42 –5 – 28 34
125 – 12 –39 –1 – 15 –12 –4 – 13 – 17 – 25 – 16 – 17 37
130 – 33 – 50 –13 –8 –24 –4 – 19 –32 –9 –32 –2 38
40 Degrees N
I 90 31 32 20 –6 – 10 28 34 15 29 24 29 37
95 17 15 17 –3 20 54 43 27 22 23 20 37
100 7 14 32 6 57 60 58 33 19 28 –4 31
105 1 17 20 12 73 36 65 28 9 29 – 19 17
110 1 6 – 33 3 56 –7 49 16 –21 5 – 15 9
115 1 – 18 7 – 16 19 – 43 13 –4 – 55 –39 –9 4
50 Degrees N
90 28 21 21 – 18 –18 – 13 8 22 27 8 22 20
95 14 21 19 –20 –6 12 18 18 20 –2 7 10
100 8 8 13 –3 27 18 43 11 12 –5 – 13 11

The preceding tables contain mean climatological val- the corresponding amplitude and phase of the solar semi-
ues. One variation of these values that is important and can diurnal tide. Note that the tide amplitudes are quite small
be readily modeled is that due to solar tides. Figure 14-31 at low altitudes, but rapidly increase at higher altitudes.
contains the amplitude and phase of the solar diurnal !ide These values were calculated theoretically by Forbes [Pri-
in density as a function of latitude and Figure 14-32 contains We Communication, 1982].

14-33
Table 14-11. Temperatures (K) 90 to 110 km.

km Jan Feb Mar Apr May Jun Jul Aug Sep Ckt Nov DC

O Degrees N
90 185 193 198 200 193 189 185 193 198 200 193 189
95 187 200* 204* 199* 187* 184* 187 200* 204* 199* 187* 184*
100 204 219* 220” 209” 193* 191* 204 219* 220* 209* 193* 191*
105 231 251* 255* 239* 219* 215* 231 251* 255* 239” 219* 215*
110 273 296* 306* 293* 269* 261* 273 296* 306* 293* 269* 261*

10 Degrees N
90 185 191* 193 197 194 189 186 194 199 200 192 188
95 187* 197* 199 193 187* 186 189 199 207* 203* 191 184
100 203* 215* 214* 201* 191* 193* 204 216* 225* 217* 200 193
105 232* 246* 245* 229” 215* 214* 229 247* 257* 247* 227* 220*
110 276* 290” 291* 278* 259* 253* 265 291* 307* 301* 282* 270*

20 Degrees N
90 186* 189 190 194 192 188 186 191 197 198 191 189
95 190* 197 195 190* 187* 189 193 197 ‘20fj* 206* 196 187
100 204* 213 210 195* 194* 198* 206 211* 222* 222* 210 197*
105 228* 237* 236* 220* 216* 221* 229 236* 248* 250* 239* 225*
110 271* 278* 275* 2j4* 255* 253* 260 276* 294* 299* 286* 271*

30 Degrees N
90 189* 190 189 190 185 179 179 187 193 195 194 194
95 195* 197 195 189* 187 189* 195 197 205 205 202 196
100 204* 210 208 196* 197* 206* 214 208 216 221 219 204*
105 222* 230* 227 216” 220” 232* 237 226 236* 249* 245* 225*
110 259* 266* 265 256* 259* 264* 265 263 278* 29(3* 284* 263*

40 Degrees N
90 199* 195 191 185 172 162 167 180 191 197 200 203
95 203* 197 196 191 184 182 192 199 203 202 208 207*
100 206” 207 207 201 204 213 222 212 211 217 222 212*
105 217* 225* 226 218 230 247 246 225* 226* 241 241* 222*
110 248* 258* 263 262 274 285 276 256* 261* 276* 272* 250”

50 Degrees N
90 208* 202 192* 179* 161* 147 153 170 185 198 207 213
95 21O* ZOO* 197* 191* 180* 173 184 194 196* 201 214 219”
100 21O* 206* 209” 21O* 212* 219 221 211* 204* 209 220* 220*
105 215* 224* 231* 237* 253* 265 255 224* 216* 228* 231* 219*
110 240* 259* 274* 284* 304* 313 293 254* 244* 256* 253* 238*

60 Degrees N
90 214 207 196* 181* 159* 1~* 145 159 177 197 211* 217
95 214* 206* 199* 193* 180* 171* 174* 181 188* 199 215* 222
100 214* 21O* 215* 221* 223* 222* 216* 203* 195* 201 215* 221
105 217* 228* 245* 263* 279* 282* 258* 220” 202* 209” 217* 215*
110 235* 261* 291* 317* 343* 343* 3ofj* 250* 223* 227* 231* 227*

70 &grees N
90 214 208 199* 183* 162* 145* 141* 150 173* 196* 212* 217
95 216* 209* 205* 197* 184* 173* 168* 170* 179* 195* 212* 219*
100 215* 214* 223* 232* 234* 226* 212* 193* 184* 193* 209* 217*
105 215* 230* 256* 285* 302” 295* 260* 213* 187* 192* 2of’j* 211*
110 229* 261* 303* 344* 373* 366* 314* 241* 198* 199* 212* 216*
‘temperature data lacking (i.e., less than two data points within about one month and 10deg latitude)
Values apply to the first day of each month.

14-34
 STANDARD AND REFERENCE ATMOSPHERES
Table 14-12. Pressu~ (~/m2) 90 to 110 km. Insert decimal point on the right of the three digits and multiply by 10N

km Jan Feb Mar Apr May .tun ‘Jul Aug Sep Ott Nov NC N

O Degrees N
90 197 200, 206 212 213 208 197 200 206 212 213 208 –3
95 803 855? 903* 920* 887* 853* 803 855* 903* 920* 887* 853* –4
100 350 395* 419* 416* 379* 36Q* 350 395* 419* 416* 379* 360* –4
105 168 199* 213* 202* 172* 162* 168 199* 213* 202* 172* 162* -4
110 090 113* 122* 112* @l* og4* 090 113* 122* 112* W1* 084* –4
10 Degrees N
w 194 188* 191 206 214 206 198 214 230 228 216 204 –3
95 079* 080* 081 088 N* 085 081 093 101* 099* 090 083 –3
100 345* 364* 374* 386* 379* 362* 357 42i* 480* 462* 395 353 –4
105 164* 182* 183* 182* 172* 164* 170 213* 246* 230* 184* 161* -4
110 089* 101* 104* 098* Q88* 084* 090 117* 141* 130* 100* 085* –4
20 Mgrees N
90 187* 188 191 213 215 201 191 214 236 233 212 199 –3
95 077* 080 080 09f)* 090* 083 079 090 j 04* 102* 089 081 -3
100 338* 362 362 386* 382* q~l* 354 412* 487* 4go* 405 353* -4
105 160* 178* 175* 178* 175* 167* 168 199* 247” 2424 195* 161* –4
110 086” 096* 096” 093* 089* 087* 089 109* 138* 137* 109* 087* –4
30 Degrees N
90 184* 187 188 212 199 183 170 189 200 202 192 194 -3
95 778* 784 789 873* 819 737* 688 796 872 876 826 822 –4
100 345* 357 354 380* 351* 327” 317 359 m 413 384 368* –4
105 163* 171* 169 172* 163* 156* 153 171 199* 207* 192* 173* –4
110 084* 091* 089 089* 084* 084* ot33 090 107* 116* 106* 091* -4
40 Degrees N
90 19a* 191 187 198 181 156 145 158 166 172 175 191 –3
95 833* 814 785 813 700 577 569 660 721 751 774 849* –4
100 379* 366 354 359 308 260 262 301 328 343 365 393* -4
105 177* 174* 167 165 145 127 132 145* 158* 172 183* 188* –4
110 899* 908* 885 858 784 719 725 748* 823* 928* 988* 962* -5
50 Degrees N
90 191* 192 191* 192* 169* 141 131 140 150 158 169 185 –3
95 873* 843* 804* 769* 619* 480 479 567 633* 694 772 867* –4
100 401* 377* 365* 348* 276* 214 218 255* 280* 310 364* 412* -4
105 190* 179* 174* 167* 1,35* 108 110 1,22* 131* 151* 180* 21O* –4
110 939* 929* 937* 923* 785* 640 626 629* 656* 778* 933* 992* –5
60 Degrees N

90 174 175 182* 184* 170* 145* 131 134 153 169 178* 179 –3
95 813* 792* 783* 754* 622* 489* 458* 504 619* 737 819” 841 -4
100 376* 359* 355* 341* 280+ 216* 198* 216* 265” 325 386* 405 -4
105 181* 174* 177* 176* 145* 112* 101* 102* 119* 150* 184* 195* –4
110 089” 090* 098” 102* 089* ofj9* 058* 052* 056* 072* 091* 095* –4

14-35
 CHAPTER 14
Table 14-12. (Continued)

km Jan Feb Mar Apr May Jun Jul Aug Sep Ott Nov Dec N

70 Degrees N
90 138 133 146* 158* 170* 152* 135* 138 164* 170* 176* 162 –3
95 646* 600* 645* 658* 649* 526* 456* 489* 638* 735* 811* 764* _4
100 3oi* 278* 298* 305* 292* 23Q* 191* 199* 261* 316* 372* 361* –4
105 145* 135* 153* 164* 162* 124* 098* @l* 11O* 139* 174* 172* –4
110 070* 071* 086* 099* 100* 077* 056* 045* 048* 061* 081* 082* –4
*n”.. I“,.l.:-.. ,: - 1-”. .. . . ....- J... --:-.. .....L.- -L_... --- —-—..-—. In J-– ,–..-.,-.
Values apply-to the first day of each m~nth.

14.2.2.4 Reference Atmospheres Above 120b. These The models described in this section include both long and
atmospheres are based onareportby Jacchia [1971]. Pre- short-term solar activity effects and diurnal, but not semi-
vious thermospheric models that assumed a single, constant diumal, variations. Atmospheric response to storms is rep-
set of boundq conditions at 120 km did not accurately resented in terms of Kp, the 3-h geomagnetic planetary
~present v@atioqs inpro~rties inthe altitude region 120 index. This provides reasonable representation of the re-
to 180 km. To attempt to remedy this problem, Jacchia sponse to most storms, but not all. For example, the Auroral
lowered the boundary to 90 km where it is known fiat there Electrojet (AE) index is better for high latitude response,
is an approximate isopycnic (constant density) layer, al- both because it provides a better measure of the change in
though the temperature shows importmt variations. In his energy input in that region, and because of its much smaller
models, all temperature profiles start from a constant value time resoh.ttion. Unfortun~tely, values of the AE index are
183 K at 90 km. He did not print out the model values not always available.
below 110 km in order to disregard the less accurate region. The average semi-annual variatioh is reasonably mod-
The following are the major ty~s of variation of prop- eled, but it must be pointed out that the amplitude of this
erties of the thermosphere: effect, whose cause is not completely understood, varies
1. Variations with the solar cycle from year to year. The seasonal-latitudinal variations of the
2. Variations with dally changes in solar activity lower thermosphere are not modeled, except for an empirical
3. Diurnal and semidiurnal variations expression for density but not temperature changes. Cham-
4. Atmospheric response to storms of various kinds pion [COESA, 1966] developed a limited set of models for
5. Semi-annual variation this portion of the atmosphere, but further modeling efforts
6. Seasonal-latitudinal variations of the lower ther- have been delayed because of a lack of information on the
mosphere systematic variations of this part of the atmosphere. At-
7. Effects of ltige scale circulation tertipts have been made to model empirically some of the
8. Fluctuations associated with gravity waves. effects of l~ge scale circulation which p~marily affect the

,, --
120 -

lw - ..-

j 80 -

z
.
I
~ 60
Y 1
:.-. -.-’-yd
40 -.’
,,-”

20 -

.~
Cal 01
tl,
04812
l,t, l,, l(l t62024 02468012
8P/p. PHASE(LTI
Bp,po PHA2E [ LT )

Figure 14-31. Relative amplitude and phase of density variation for the Figure 14-32. Relative amplitude and phase Of density variation for the
diurnal tide at 0°, 30°, and 60° latitude under equinox semidiurnal tide at 0°, 30°, and ~“ latitude under equinox
conditions. conditions.

14-36
Table 14-13. Densities (kg/m3) 90 to 110 km. Inserf decimal point on the right of the three digits and multiply by 10N

km Jan Feb Mar Apr May Jun Jul Aug Sep Ott Nov &C N

O Degrees N
90 370 360 361 368 384 384 370 3@ 361 368 384 384 – 8
95 148 147* 153* 159* 163* 160* 148 147* 153* 159* 163* 160* –8
100 580 610* 644* 673* 663* 636* 580 61O* 644* 673* 663* 636* – 9
105 242 264* 279* 282* 262* 252* 242 264* 279* 282* 262* 252* – 9
110 108 125* 131* 126* 111* 106* 108 125* 131* 126* 111* 106* –9

10 Degrees N
90 365 343* 345 364 383 380 370 384 403 396 390 378 – 8
95 146* 140* 140 157 166* 158 148 161 169* 169* 162 156 –8
100 574* 571* 591* 648 * 670” 633* 591 658* 721* 719* 668 617 –9
105 236* 246* 249* 265* 266* 256* 247 287* 318* 31O* 270” 244* – 9
110 106* 114* 117* 116* 112* 109* 112 133* 151* 142* 116* 103* –9

20 Degrees N
90 349* 347 350 382 390 372 358 390 417 409 386 366 – 8
95 139* 139 142 163* 166* 152 142 158 174* 170* 156 149 –8
100 560* 574 583 fjfj9* fjfj5* 616* 581 659* 741* 731* 651* 605* –9
105 234* 249* 247* 269* 270* 252* 245 281* 332* 322* 272* 239* – 9
110 104* 114* 114* 115* 115* 113* 113 129* 154* 150* 125* io6* – 9

30 Degrees N
90 339* 342 345 388 375 355 329 351 361 361 345 347 – 8
95 138* 137 139 159* 151 134* 122 139 147 147 141 145 –8
100 571* 574 574 654* ~j* 535* 500 583 633 631 593 609* – 9
105 244* 247* 248 265 * 246* 224* 215 252 280* 276* 2~* 256* – 9
110 107* 113* 110 114* 107* 104* 103 112 127* 132* 123* 114* –9

40 Degrees N
90 333* 341 340 373 366 335 302 305 302 303 304 327 –8
95 141* 142 138 147 131 109 102 114 122 128 128 141 –8
100 622* 598 578 603 510 412 399 479 526 534 555 627* –9
105 271* 257* 247 253 210 171 178 214* 233* 237 252* 281* –9
110 119* 116* 110 108 094 083 086 096* 104* 110* 119* 126* –9

50 Degrees N
90 320* 331 346* 374* 365* 333 297 287 282 277 283 302 –8
95 143” 145* 141* 139* 119* 096 090 101 111* 119 124 136* –8
100 645” 619* 589* 560” 439* 330 333 408* 464* 502 559* 633* –9
105 294* 266* 251* 234* 178* 136 143 182* 202* 220” 259* 305* –9
110 128* 118* 112* 107* 085* 067 070 081* 088* 100* 121* 137* –9

60 Degrees N
90 283 295 323* 353* 372* 350* 314 293 300 298 294* 286 –8
95 131* 133* 136* 135* 119* 099” 091* 096 113* 128 131* 131 –8
100 593* 578* 559* 522* 424* 328* 31O* 360+ 459* 546 fjo6* 619 –9
105 278* 254* 241* 223* 173* 132* 130* 154* 196* 239* 282* 3Q2* –9
110 124* 114* 111* 106* 085* ofj6* 062* ofj8* 083* 104* 129* 137* –9

70 Degrees N
90 225 222 255* 301* 365* 365* 332* 321 329* 302* 288* 260 –8
95 103* 099* 108* 115* 122* 105* 094* 099* 123* 130* 132* 120* –8
100 474* 439* 451* 445* 422* 345* 305* 349* 479* 553* fjo2* 563* –9
105 224* 195* 200* 192* 178* 140* 125* 142* 197* 241* 281* 272* –9
110 101* 089* 094* 095* 088* 069* 058* ofj2* 080* 10I* 125* 124* –9
*Data lacking (i.e. less than two data points within about one month and 10 deg. Iatltude)
Values apply to the first day of each month.

14-37
CHAPTER la
composition of the thermosphere. For example, the strong where TD is the maximum daily temperature and TN is the
increase in helium concentration above the winter pole (the minimum daily temperature. R is a coefficient with a value
helium winter bulge) is a result of large scale circulation, of approximately 0.3,
which causes preferential transport of lighter constituents
away from the summer hemisphere. Density and cotipo-
sition fluctuations caused by gravity waves are not modeled.
The tables in this section are given with the exospheric
teinperature as the defining parameter. To determine the
where $ is the latitude and 80 is the sun’s declination. A
exospheric tem~rature for a particulw time and location
value of 2.2 can be used for m.
the following equations are used. The temperatuk T= is the
global night-time minimum of the exospheric temperature For Kp above zero the exospheric temperature is in-
creased by
when K~ is zero.

AT. = 28 Kp + 0.08 exp Kp . (14.20)

T. = 379 + 3.24 ~Io.7 + 1.3 (FIo.7 – ~10,7) (14.18)
Formulas to give the exospheric temperature at any time of
day, to include the effect of the semi-annual variation, and
where F10,7is in units of l@22 W m-2 HZ-l. The diurnal
to include the winter heliutn bulge are given in the reference
variation of the exospheric temperature can be represented
[Jacchia, 1971].
by
Table 14-14 contains model values of kinetic temper-
ature, logarithms of the concentrations of N2, 02, 0, Ar,
T~ = T= (1 + R COSmq) He, and H, mean moleculti weight, and density as functions
(14.19) of altitude for exospheric teriiperaturos ranging from 500 to
TN = TC(1 + R sinme), lm K.

Table 14-14. Atmospheric temperature, density and composition as fonctions of height and exospheric temperature

Height Temp LogN(N2) LogN(Oz) bgN(0) Lo&_~A) LogN(HE) ~:N:H) Mean Density
m-3 m-3
km K m-3 m-3 Mol Wt kg/m3

Exospheric Temperature = 500 K

120 267.1- 17.536 16.672 17.169 15.073 13.579 25.10 2.25E–08
130 328.3 16.981 16.051 16.814 14.320 13.457 23.74 6.80E – 09
140 378.8 16.531 15.545 16.530 13.704 13.363 22.51 2.67E–09
150 411.9 16.149 15.114 16.296 13.174 13.291 21.39 1.25E–09
160 433.1 15.804 14.723 16.089 12.691 13.231 20.40 6.51E– 10
180 457.7 15.170 14.~2 15.717 11.798 13. i29 18.78 2.13E– 10
200 471.4 14.576 13.325 15.372 io.956 13.038 17.62 8.13E–11
250 488.1 i3.169 il.721 14.562 8.956 12.830 15.79 1.d5E-11
300 494.6 11.823 10.184 13.791 7.039 12.635 13.30 1.72E– 12
350 497.3 10.512 8.686 13.041 12.ti6 8.33 3.29E– 13
400 498.5 9.226 7.218 12.306 12.262 3.82 8.1 OE–14
5@ 499.5 10.873 11.903 12.860 1.44 1.94E– 14
600 499.8 9.484 11.555 12.772 1.19 1.24E– 14
8@ 499.9 6.822 10.889 12.* 1.06 7.24E– 15
1000 5@.o 10.2W 12.445 1.03 4.79E– 15
1500 5W.O 8.827 12.084 1.01 2.04E– 15
2000 5oil.o 7.565 11.767 1.01 9.78E– 16
Exospheric Temperature = 600 K
120 285.2 17.549 16.6W 17.165 15.103 13.567 25.21 2.3CiE-08
130 359.6 17.018 16.09s 16.818 14.389 13.*3 23.98 7.29E–09
140 423.6 16.597 15.627 16.547 13.817 13.349 22.87 3.00E-09
150 468.4 16.246 15.232 16.328 13.336 13.278 21.89 1.48E –@
160 498.6 15.937 14.883 16.140 12.906 13.221 20.99 8.1OE– 10

14-38
 a
STANDARD AND REFERENCE ATMOSPHERES

Table 14-14. (Continued

Height Temp LogN(N2) LogN(Oz) LogN(O) LogN(Ar) LogN(HE) Lo:N$H) Mean Density
m-3
km K m-3 m-’ m-3 m-3 Mol Wt kg/m3

180 534.9 15.381 14.252 15.809 12.127 13.127 19.49 2.93E– 10

200 555.8 14.869 13.670 15.510 11.404 13.046 18.35 1.23E– 10
250 581.6 13.676 12,310 14.820 9.711 12.866 16.67 1.99E–11
300 591.6 12.546 11.021 14.171 8.103 12.701 15.72 4.15E–12
350 595.8 11.450 9.769 13.544 6.541 12.543 14.41 9.69E– 13
400 597.7 10.377 8.543 12.931 12.389 11.79 2.46E– 13
500 599.2 8.285 6.154 11.735 12.090 12.126 4.81 2.49E– 14
600 599.6 6.257 10.577 11.800 12.053 2.37 7.08E– 15
800 599.9 8.359 11.245 11.913 1.54 2.54E– 15
1000 600.0 6.262 10.720 11.780 1.25 1.36E– 15
1500 600.0 9.525 11.484 1.04 5.27E– 16
2000 600.0 8.474 11.215 1.01 2.76E– 16

Exospheric Temperature = 700 K

120 300.5 17.559 16.705 17.161 15.126 13.558 25.29 2.35E–08

130 386.2 17.046 16.134 16.821 14.440 13.433 24.15 7.67E–09
140 462.0 16.644 15.686 16.558 13.900 13.338 23.14 3.27E–09
150 518.2 16.315 15.317 16.349 13.452 13.267 22.24 1.67E–09
160 557.8 16.029 14.995 16.172 13.058 13.211 21.43 9.46E– 10

180 607.5 15.526 14.426 15.869 12.357 13.122 20.03 3.67E– 10

200 636.8 15.072 13.910 15.601 11.718 13.047 18.92 1.65E– 10

250 673.5 14.030 12.723 14.995 10.242 12.886 17.15 3.16E–11

300 688.0 13.054 11.610 14.434 8.854 12.742 16.24 7.80E– 12
350 694.0 12.111 10.533 13.894 7,511 12.606 15.53 2.17E– 12

400 696.8 11.189 9.481 13.367 6.198 12.473 14.50 6.46E– 13

500 698.8 9.395 7.431 12.341 12.216 11.559 10.02 6.99E– 14
600 699.5 7.656 11.348 11.968 11.496 5.19 1.26E– 14
800 699.8 9.446 11.492 11.376 2.77 2.54E– 15
1000 699.9 7.649 11.042 11.263 2.13 1.04E– 15
1500 700.0 10.018 11.005 1.29 2.39E– 16
2000 700.0 9.117 10.778 1.07 1.09E– 16

Exospheric Temperature = 800 K

120 313.5 17.567 16.716 17.158 15.145 13.550 25.35 2.39E–08

130 408.8 17.067 16.162 16.823 14.481 13.424 24.28 7.99E–09
140 495.1 16.680 15.731 16.566 13.964 13.329 23.35 3.50E–09
150 562.1 16.366 15.380 16.363 13.539 13.258 22.52 1.82E–09
160 611.3 16.096 15.078 16.194 13.171 13.202 21.77 1.06E-09

180 675.4 15.632 14.553 15.910 12.526 13.115 20.46 4.34E– 10

200 714.4 15.218 14.084 15.664 11.948 13.044 19.39 2.06E– 10

250 764.0 14.288 13.026 15.120 10.633 12.897 17.57 4.46E– 11

300 783.7 13.427 12.043 14.623 9.410 12.769 16.59 1.25E–11
350 791.9 12.598 11.098 14.148 8.230 12.648 15.96 3.96E– 12

400 795.6 11.791 10.175 13.686 7.079 12.532 15.33 1.34E– 12

500 798.4 10.219 8.381 12.788 12.307 11.108 12.86 1.77E–13

14-39
CHAPTER 14
Table 14-14. (Continued

Height Temp LogN(N2) LogN(02) LogN(O) LogN(Ar) LogN(HE) LogN(H) Mean Density
~-3 m.’ ~-’
km K m-’ m-’ m-’ Mol Wt kg/m3

Exospheric Temperature = 800K

600 799.3 8.697 6.642 11.918 12.089 11.053 8.43 3.04E– 14

800 799.8 10.254 11.673 10.948 3.92 3.75E– 15
1000 799.9 8.681 11.279 10.849 3.22 1.39E– 15
1500 800 10.383 10.623 2.10 2.31E– 16
2000 800.0 9.594 10.425 1.39 7.06E– 17

Exospheric Temperature = 900 K

120 324.7 17.574 16.726 17.156 15.160 13.544 25.41 2.41E–08

130 428.4 17.084 16.184 16.825 14.513 13.417 24.39 8.25E–09
140 523.9 16.708 15.766 16.572 14.013 13.321 23.51 3.68E–09
150 600.9 16.404 15.428 16.373 13.606 13.249 22.73 1.95E–09
160 659.6 16.147 15.140 16.209 13.257 13.193 22.03 1.16E–09

180 738.9 15.710 14.648 15.938 12.654 13.107 20.81 4.93E– 10

200 788.5 15.328 14.216 15.708 12.122 13.039 19.79 2.44E– 10

250 852.9 14.483 13.255 15.211 10.931 12.902 17.97 5.83E–11

300 878.6 13.710 12.374 14.764 9.835 12.786 16.91 1.80E–11
350 889.3 12.971 11.530 14.339 8.783 12.678 16.26 6.29E– 12

400 894.2 12.251 10.709 13.927 7.758 12.574 15.74 2.36E– 12

500 897.9 10.853 9.112 13.128 12.373 10.741 14.22 3.76E– 13
600 899.0 9.500 7.566 12.355 12.179 10.692 11.08 7.04E– 14
800 899.7 6.910 10.876 11.809 10.599 5.03 6.35E– 15
1000 899.9 9.478 11.459 10.511 3.81 2.05E– 15
1500 900.0 6.294 10.663 10.310 3.08 3.40E– 16
2000 900.0 9.962 10.133 2.21 8.36E– 17

Exospheric Temperature = 1000 K

120 334.5 17.579 16.733 17.154 15.172 13.539 25.45 2.44E-08

130 445.4 17.098 16.202 16.826 14.539 13.411 24.48 8.47E–09
140 549.0 16.730 15.794 16.577 14.053 13.315 23.64 3.84E–09
150 635.2 16.435 15.467 16.381 13.660 13.242 22.91 2.06E-09
160 703.1 16.187 15.189 16.220 13.325 13.186 22.25 1.24E–09
180 798.1 15.770 14.721 15.959 12.755 13.100 21.10 5.44E– 10
200 859.3 15.413 14.317 15.741 12.258 13.034 20.13 2.78E– 10

250 940.2 14.634 13.434 15.279 11.164 12.904 18.33 7.21E–11

300 972.8 13.932 12.633 14.872 10.169 12.796 17.21 2.40E– 11
350 986.5 13.263 11.871 14.488 9.219 12.698 16.52 9.09E– 12

400 992.6 12.614 11.130 14.116 8.294 12.604 16.02 3.69E– 12

500 997.3 11.355 9.692 13.396 6.499 12.423 10.438 14.94 6.89E– 13
600 998.8 10.137 8.300 12.699 12.249 10.393 12.85 1.45E– 13
800 999.6 7.805 11.368 11.915 10,309 6.55 1.17E–14
1000 999.9 10,109 11.601 10.229 4.24 3.02E– 15
1500 1000.O 7.244 10.884 10.049 3.62 5.28E– 16
2000 1000.o 10.253 9.890 3.10 1.32E–16

14-40
 i

STANDARD AND REFERENCE ATMOSPHERES

Table 14-14. (Continue@

Height Temp LogN(N2) LogN(()*) LogN(O) LogN(Ar) LogN(HE) LogN(H) Mean Density
m-3 m-3 m-3 m-3 m-3
km K m-3 Mo] Wt kg/m3

Exospheric Temperature = 1200K

120 350.9 17.587 16.745 17.150 15.192 13.530 25.52 2.48E–08

130 473.8 7.119 6.229 16.827 14.580 13.401 24.61 8.82E–09
140 591.1 6.764 5.837 16.583 14.115 13.304 23.84 4.09E–09
150 693.3 6.481 5.524 16.391 13.741 13.231 23.17 2.24E–09
160 778.3 6.245 5.262 16.235 13.426 13.173 22.57 1.37E–09

180 905.0 5.857 4.828 15.985 12.901 13.087 21.54 6.28E– 10

200 991.4 15.533 14.463 15.783 12.455 13.021 20.65 3.36E– 10

250 110.3 14.851 13.691 15,373 11.504 12.901 18.95 9.84E–11

300 159.1 14.253 13.011 15.023 10.659 12.806 17.79 3.70E–11
350 179.6 13.691 12.370 14.699 9.860 12.722 17.01 1.57E–11

400 188.9 13.147 11.750 4.387 9.087 12.643 16.47 7.19E– 12

500 195.9 12.096 10.549 3.785 7.589 12.491 9.966 15.67 1.70E– 12
600 198.2 11.080 9.388 3.205 6.140 12.346 9.929 14.62 4.46E– 13
800 199.4 9.136 7.169 2.095 12.068 9.858 10.17 4.09E– 14
1000 199.8 7.300 1.046 11.806 9.792 5.74 7.21E– 15
1500 200.0 8.658 11.208 9.642 3.96 1.09E– 15
2000 1200.0 6.557 10.682 9.509 3.82 3.25E– 16

Exospheric Temperature = 1400 K

120 364.1 17.593 16.754 17.147 15.207 13.523 25.57 2.51E–08

130 496.8 17.134 16.250 16.827 14.611 13.393 24.71 9.1 OE– O9
140 625.2 16.789 15.869 16.587 14.160 13.296 23.99 4.29E–09
150 740.8 16.514 15.566 16.398 13.800 13.222 23.36 2.38E–09
160 841.0 16.286 15.313 16.245 13.498 13.163 22.81 1.48E –09

180 998.5 15.916 14.901 16.001 13.003 13.075 21.86 6.93E– 10

200 1111.9 15.613 14.562 15.809 12.591 13.009 21.05 3.82E– 10
250 1274.4 14.998 13.867 15.431 11.738 12.893 19.45 1.22E– 10
300 1342.6 14.473 13.271 15.122 10.999 12.807 18.30 5.00E -11
350 1371.4 13.986 12.716 14.840 10.309 12.733 17.48 2.32E–11

400 1384.4 13.518 12.182 14.571 9.643 12.664 16.89 1.15E–11

500 1394.3 12.614 11.151 14.054 8.356 12.534 9.621 16.09 3.23E– 12
600 1397.4 11.743 10.155 13.555 7.113 12.409 9.588 15,39 9.98E– 13
800 1399.2 10.077 8.252 12.603 12.171 9.528 12.79 1.17E–13
1000 1399.7 8.503 6.454 11.704 11.946 9.471 8.37 1.93E– 14
1500 1399.9 9.658 11.434 9.342 4.18 1.93E– 15
2000 1400.0 7.856 10.983 9.229 3.96 6.44E– 16

Exospheric Temperature = 1600 K

120 375.4 17.598 16.762 17.144 15.219 13.518 25.62 2.53E–08

130 516.4 17.147 16.266 16.827 14.635 13.387 24.79 9.32E–09
140 654,1 16.808 15.894 16.590 14.196 13.290 24.10 4.45E–09
150 781.1 16.540 15.598 16.404 13.846 13.215 23.51 2.50E–09
160 894.6 16.318 15.353 16.252 13.554 13.155 23.00 1.56E–09

14-41
CHAPTER 14
Table 14-14. (Conrinue@

Height Temp Log(N(N,) Log(N(O,) bg(N(0) Log(N(A) Log(N(HE) bg(N(H) Mean Density
km K /m3) /m3) /m3) /m3) /m3) /m3) Mol Wt kg/m3

180 1081.3 15.960 14.956 16.012 13.080 13.065 22.11 7.46E– 10

200 1222.4 15.672 14.635 15.825 12.692 12.998 21.36 4.20E– 10

250 1433.1 15.103 13.995 15.470 11.910 12.884 19.87 1.43E– 10

300 1523.5 14.633 13.461 15.190 11.250 12.804 18.75 6.27E– 11
350 1561.9 14.201 12.970 14.939 10.640 12.737 17.91 3.1 OE– I1

400 1579.2 13.790 12.500 14.702 10.055 12.676 17.28 1.64E-11

500 1592.4 12.997 11.596 14.248 8.926 12.561 9.361 16.43 5.21E– 12
600 1596.5 12.234 10.724 13.811 7.838 12.452 9.333 15.82 1.82E– 12
800 1599.0 10.776 9.059 12.978 12.243 9.280 14.21 2.67E– 13
1000 1599.6 9.398 7.485 12.192 12.046 9.230 11.01 4.88E– 14
1500 1599.9 6.263 10.401 11.598 9.117 4.71 3.30E– 15
2000 1600.0 8.824 11.204 9.018 4.03 1.08E– 15

14-42
 STANDARD ANI) REFERENCE ATMOSPHERES
REFERENCES

Btilier, F., C. Berger, J,L. Falin, G. Kockarts, and G. Fuller-Rowell, T.J. and D. R~es, “A Three-Dimensional
Thuillier, “A Thermospheric Model Based on Satellite Time-Dependent Global Model of the Thermosphere,”
Drag Data,” Ann. Geophys., 34:9, 1978. J. Atmos. Sci., 37:2545, 1980.
Champion, K.S.W., “Properties of the Mesosp~ere and ~wer Groves, G. V., “Atmospheric Structure and its Variations in
Thermosphere,” 4dv. Space Res., 1: 133, 1981. the Region from 25 to 120 km,”AFCRL-71 -0410, 197i.
Champion, K. S.W., “Properties of the M&sosphere and Hedin, A. E., J.E. Salah, J.V. Evans, C.A. Reber, G.P.
~ermosphere and Comparison with CIRA 72,” Adv. Newton, N. W. S~ficer, D.C. Kayser, D. Alcayde, P.
Space Res., 3:45, 1983. Bauer, L. Cogger, and J.P. McClure, “A Global Ther-
Champion, K.S.~. and R.A. Schweinfutih; “A New Mean rnospheric Model haied on Mass Spectrometer and In-
Reference Atmospher6 for 25 to 500 km,” AFCRL-72- coherent Scatter Data. 1. Nz Density ad Temperamre,”
0579, AD757483, 1972. J. Geoptzys. Res., 82:2139, 1977.
COESA, U.S. Standard Atmosphere, 1969, U.S. Gover- ICAO (International Civil Aviation Organization), Manual
nmentPrinting office, Washington, D. C., 1%2. of the ICAQ Standard Atmosphere to 32 km (105,000
COESA; U.S. Standard Atmosphere Supplements, 1%6, X), Second Edi~on, Montreal, Canada, 1964.
U.S. Govemkent Printing Office, Washington, D. C., 1S0 (International Standards Organization), International
19fi. Sta&rd Atmosphere, 1S0 Standard No. 2533,
CQESA, U.S. Standard Atmosphere, 1976, U.S. Gover- ISO/TC20/SC6 Secretariat USSR, 1973.
nmentPrinting Office, Washington, D.C. 1976, Jacchia, L. G., “Revised Static Models of the Thermosphere
Cole, A, E., “Models of extreme arctic and subarctic at- and Exosphere with Empirical Tem@ratu~ Profiles,”
mospheres between 20 and 90 km,” Space Res. X11, Smithson. Aitrophys. Ohs. Spec. Rep., 332, 1971.
629, Akadernie-Verlag, Berlin, 1972. Jacchia, L.G., “Thermospheric Temperature, Density, and
Cole, A.E. and A.J. Kantor, “Air Force Reference Atmos- Composition: New M@els,” Smithson. Astrophys. Ohs.
pheres,” AFGLTR-78-0051, ADA05g505, 1978. Spec. Rep,, 375, 1977.
Colegrove, F. D., W.B. Hanson, and F.S. Johnson, “Eddy List, R. J., Smithsonian Meteorological Tabies, Smithsonian
Diffusion and Oxygen Trans~rt in the Lower Ther- Institution Press, Washington, D. C., 1968.
mosphere,” J. Geophys. Res., 70: 49314941, 1965. Labitzke, K., “Synoptic-Scale Motions Above the Strato-
COSPAR Working Group IV, COSP~ Interrtational Ref- pause,” NC* Ms. No. 71-39, 1971.
erence Atmosphere (CIRA) 1965, North-Holland, Am- Minzner, R. A., K.S.W. Champion, and H.L. Pond, “me
sterdam, 1965. & Model Atmos@m, 1959,” AFCRC-TR-59-267,
COSPAR Working Group IV, COSPAR International Rej-
erence Atmosphere (CIRA) 1972, Akadeqie-Verlag, Quiroz, R. S., “The Determination of the Amplitude and
Berlin, 1972. Altitude of Stratospheric Warrnings from Satellite-Mea-
Dickinson, R, E., E.C. Ridley, and R.G. Roble, “A Three- sured Radiance Changes,” JAM, 10(3), 1971.
Dimensional General Circulation Model of the Ther- Slowey, J. W., “Global Frequency Distribution of Exospheric
mosphere,” J. Geophys. Res., 86: 1499, 1981. Temperature,” AFGLTR-79-0143, ADA077293, 19?9.

14-43
 Chapter 15

ATMOSPHERIC TEMPERATURES, DENSITY,

AND PRESSURE
Section 15.1 1.1. Giingorten, A.J. Kantor, Y. Izumi, and P.I. Tattelman
Section 15.2 A.J. Kantor and A.E. Cole
Seetion 15.3 A.J. Kantor and A.E. Cole

The three physical properties of the earth’s atmosphere, 15.1.1 Energy Supply and Transformation
temperature T, density p, and pressure P are related by the
ideal gas law P = p TR where R is known as the gas The prime source of energy that produces and maintains
constant for air. Except for the one thousandth of 1% of the atmosplieric motions and the spatial and temporal variations
atmosphere by mass above 80 km, various gases comprise of meteorological elements is the solar radiation intercepted
the atmosphere in essentially constant proportions. The prin- by the earth. In comparison with solar radiation, other r.n-
cipal exception is water vapor discussed in Chapters 16 and ergy sources such as heat from the interior of the earth,
21. radiation from other celestial bodies, or the tidal forces of
the moon and sun are pmctjcally negligible. Manmade sources,
such as the heat island of a city, can be neglected although
their by-products, such as the increasing amounts of carbon
15.1 THERMAL PROPERTIES, SURFACE dioxide in the atmosphere, have been subjected to intense
TO 90 KM scrutiny in recent years with respect to heat balance and
climatic trends.
In the following sections the units of measurement are Tlie rate at which solar energy is received on a plane
metric. Abbreviations are used whenever quantitative mea- surface perpendicular to the incident radiation outside of the
sures are presented. The temperature is in degrees Kelvin atmosphere at the earth’s mean distance from the sun is the
(K), density in kilograms per cubic meter (kg/m’), ~reisure solar consfant; the approximate value used in this section
in millibars (rob), time in seconds (s) or hours (h), length is 2.0 L/rein, or about 1400 W/m2. (A detailed description
in centimeters (cm), meters (m) or kilometers (km), and of the solar constant and its empirically detennirted value
speed in meters per second (m/s) or kilometers per hour is given in Chapters 1 and 2. The rate at which direct solar
(km/h). The main unit of energy is the calorie (cal): energy is received on a unit horizontal plane at the eartli’s
surface or in the atmosphere above the earth’s surface is
1 cal = 4.1860 joules (J) called the itisolation. The planetary albedo, which is the
reflected radiation divided by the total incident solar radia-
—
– 1.163 x 10-3 watt-hours (Wh). tion, varies primarily with angle of incidence of the radia-
tion, the type of sutiace, and the amount of cloudiness. On
For energy per unit area an additional unit, the Langley (L), the average, 30% to 40% of the incident solar energy is
is introduced: reflected back to space by the cloud surfaces, the clear
a~osphere, the earth/air inte~ace, and particles such as
1 L = 1 cal/cm2 dust and ice crystals suspended in the atmosphere. The
remaining 60% to 7070 of the soIar radiation is available as
= 11.62 Wh/m2
the energy source for maintaining and driving atmospheric
= 41.84 kJ/m2. processes.
Less than twenty years ago we could confidently con-
The main unit of power is the watt (W), but the unit of sider the earth and its surrounding atmosphere as a self-
solar power per unit ~ea is given as Langleys per hour contained thermodynamic system. No major energy changes
(L/H). In terms of the British Thermal Unit (BTU) 1 in the system within the 50 to 100 year period of our cli-
L/H = 3.686 BTU “ ft-2 “ h-’. matological records were apparent. Globally there had been

15-1
 CHAPTER 15
no obvious systematic shofi-term change in (1) heating of point varies during the earth’s’ rotation about its axis and
the earth’s surface or the atmosphere, (2) ‘the intensity of revolution about the sun. A consistent feature of this vari-
the atmospheric circulation, or (3) the balance between ation on a global scale is the driving force produced by
evaporation and precipitation. The processes affecting the differential latitudinal solar heating of fhe eai-th’s surface.
internal and latent heat and’ the mechanical energy within The reaction “of the atmosphere’ to the solar driving force
the earth-atmosphere system had appeared virtually bal- on an hourly, a daily, or an annual basis is observed most
anced. easily in the temperature field at low levels.
Over the past twenty years there has been much ago- The solar energy input varies according to season, lat-
nizing by many experts and authors over the possibility of itude, orientation of terrain to the incident energy, soil struc-
climatic change. Since there have been changes in the cli- ture, all of which cap change the balance between the in-
mate throughout geological history, it is inevitable that there coming solar and sky radiation (short wave) and the outgoing
will be long-term and large-scale changes in the future. Man- atmosphere-terrestrial radiation (long wave). The difference
produced local changes ‘through the use of fossil fuels, de- ~tween short-wave an~ long-wave radiation is the net ra-
struction of forests and desertification, irrigation on one diation.” Locally, net radiation is decreased primarily by
hand and drainage of swamps on the other hand, urbani- atmospheric moisture (vapor and clouds). Evaporation of
zation and the introduction of pollutants in the air all have soil moisture diminishes by the latent heat required the por-
telling effects on local climate. The broader implications, tion of net radiation available for heating air aqd soil at the
however, over large regions and over decades or centuries ground. The importance of moisture in establishing general
have been the subject of many extensive and ongoing in- climatic zones is shown by comparing desert climates with
vestigations by agencies worldwide with only one univer- adjacent climates at roughly the same latitude. Table 15-1
sally accepted conclusion. The carbon dioxide content of gives the effect of soil mdisture ori the heat budget of the
the atmosphere is increasing, which may lead to a global earth/air interface.
Warming [WMO, 1979]. The next 5 tq 10 years might pro- Slopes facing south dive m~irnum solar energy. Slopes
duce a valid prediction. facing west are usually warmer and drier than those facing
A consensus among climatologists on heating or cooling east because the time of maximum insolation on a west
of large regions of the earth or changes in rainfall patterns slope is shifted to the afte,rqoon when the general level of
in response to natural or manmade influences is lacking. air temperattire is higher than in the forenoon.
For this chapter the climate is considered to be stationary The energy balance of the eatih/air interface requires
stochastic. It is stochastic because there is much variability that net radiation equals the sum of heat fluxes into the air
in weather events and conditions that can be fitted into and soil plus the heat equivalent of evaporation. In general,
probability distributions assuming paiially random pro- the maximum of heat flux into the soil precedes the maxi-
cesses. It is stationary because derived statistics or param- mum of heat flux into the air. ~h6 temperature maximum
eters, such as averages and standard deviations, are assumed at standard instrument height (aWut 1.8 m) follows the
to be unchanging. Their true values are constant and are maximum of heat flux into the air by roughly one to two
best e~timated by large samples. hours.
The main features of the global energy transformation
are summarized in a flow chart, Figure 15-1, from which
the relative importance of the major energy cycles within
the earth-atmosphere thefiodynamic system can be deter- 15.1.2 Surface Temperature
mined. The numerical data presented “inthis figure are useful
for various quantitative estimates. For example, if all energy 15.1.2.1 Official Sktion Temperature. The stand~d
inputs for the system ceased and rates of energy expenditure station temperature used in meteorology [NWS, 1979] is
were maintained, the reservoir of mechanical energy ‘(mo- rneasuti by a thermometer enclosed within a white-painted,
mentum) would be depleted in about 3 days, the reservoir louvered instrument shelter or Steyenson Screen. The shelter
of latent heat (precipitable water) in about 12 days, and the has a base about 1 m (1.7 to 2.0 m in Central Europe) above
reservoir of internal energy (heat) in about 1~ days. the grourid and is mounted in an open field (close-cropped
Although an absolutely dry and motionless atmosphere grass surface): The standardized height of the thermometer
is conceivable, it is difficult to imagine an atmosphere at is abut 1.8 m. The shelter permits air circulation past the
zero degrees K. It is perhaps moke appropriate to interpret thermometer and is designed to exclude direct solar and
the above time intervals as fictitious cycles of turnover of terrestrial radiation. However, the shelter unavoidably ab-
the atmospheric prope~ies. The relative magnitudes of these sotbs and radiates some heat energy which, although mini-
life cycles show that, in comparison with rainfall and winds, mal, causes some deviation of the thermometer reading
temperature is the most conservative and will exhibit the from the “tree” air temperature. On calm, sunny days the
relatively smallest, and thus most regular, temporal and recorded daytime shelter temperature will normally b 0.5
spatial large-scale variation. to 1 K higher than ~hefree air temperature outside the shelter
The solar energy input into the atmosphere at any one at the same level. On calm, clear nights it will most likely

15-2
 a
ATMOSPHERIC TEMPERATURES, DENSITY, AND PRESSURE

A A
100
INCOMING SOLAR RADIATION ‘ REFLECTED SOLAR RADIATION OUTGOING TERRESTRIAL RADIATION
100
L 4 i
’34 ‘ ~

I AVAILABLE SOLAR ENERGY

I
1-20 -46 I
! v

46 -42 -24

I
<-1 <1
RESERWIR- :F INTERNAL E:gRGY (500 000
4

A
W h m-2 )
<1

20 <1 <-1
RESERVOIR OF UTENT HEAT
(20000 Wh m-~)
[
1 20 I
RELEASED2&
PRECIPITATION
20 INVOLVED IN DRIVING -
HYDRO@GICAL AND THERMODYNAMICAL
CYCLES AND CIRCULATl~S ~

I WORK DONE AT ~WER B~~

<1 <-1 <-1

RESERVOIR OF MECHANICAL ENERGY (140 W h m-z)
4- I <1

T v A
IN;OLVED IN CONTINUOUS C:&GES & POTENTIAL EHGY
(BUILD-UP MD DECAY * MmE-ScALE HORlmNTAL PRE=U= DIFFERENCES)

Figure 15-1. Global mean energy cycles of the atmosphere.All values are relativeto 100 units of incomingsolar radiation, which average350 W/m*,
or l/4 of the solar constant. [Revised from httau, 1954a].

be cooler by 0.5 K. Therefore, the thermometer should be 15.1.2.2 Daily Temperature. The official station tem-
artificially ventilated. Spatial variations of the ambient air perature taken every hour on the hour reveals a fairly ~gular
temperature, especially in the first meter above the ground, diurnal cycle. This is true despite several superimposed
are dependent upon the type of soil and ground cover. Ground phenomena such as frontal passages and thunderstorms.
effects decrease with height and for this reason the inter- Usually there is a temperature maximum in midaftemoon
national standard heights of temperature measurement are and a temperature minimum near sunrise. The amplitude of
a compromise between suppressing ground-cover effects the diurnal cycle varies with location and season from as
and maintaining ease of access. little as 1 K to more than 17 K.

15-3
CHAPTER 15
Table 15-1. Short-wave radiation on horizontal plane, net radiation, and estimated constituents of heat budget at the earth/air interface showine effect
of difference in soil moisture caused by rains before 9 August and a dry spell before 7 Septem-kr 1953 [Davidson and bttau, 195;].

Radiation (W/m2)

Mean Local Time 04h 06h 08h 10h 12h 14h 16h ’18h 20h

9 August 1953*

Short-wave o 141 544 733 796 823 537 144 —

Net – 59 47 364 497 540 525 273 –13 —
Flux into soil – 40 29 186 63 74 73 28 –65 —
Flux into air –11 — 81 158 176 190 64 –17 —
Heat of evap. –8 — 97 276 290 262 181 69 —

7 September 1953**

Short-wave o 54 441 765 870 735 407 44 0

Net – 54 – 32 181 403 488 398 154 – 69 – 77
Flux into soil –44 – 25 36 84 95 66 13 – 29 – 28
Flux into air –6 –6 98 230 303 299 114 – 30 – 39
Heat of evap. –4 –1 47 89 90 33 27 – 10 – 10
... .. . . ---- .A- . . . .
“Mean SOII moisture m U to 1U cm layer,’ about tUV6 wet weight IJasls.
**Mean soil moisture to O to 10 cm layer, about 4% wet weightbasis.

The annual cycle of daily mean temperature ranges from endures for several months of the year, the 24-h cycle is
practically zero near the equator to as much as 40 K in the minimal and the small diurnal variations are controlled pri-
temperate zone. As an example, Figure 15-2 shows tem- marily by changing winds and cloudiness. In summer, nearly
peratures at Hanscom AFB, Mass. The middle curve reveals all of the solar energy is expended in melting ice; hence,
the annual cycle of the daily mean temperature (actually the the maximum temperature seldom exceeds 273 K. Extra-
monthly mean is plotted) and shows an annual range of 25
K. The diurnal range, given here by the difference between
mean daily maximum and minimum in Figure 15-2, is fairly
~ +$
uniform throughout the year, averaging 12 K.
310 -
Superimposed on both the diurnal and the annual cycles ~$k
~+
of temperature are many influences including cloudiness, O\O

precipitation, wind speed and direction, type of soil, ground \

300 - ~c+
cover, and aerodynamic roughness of the terrain. In the ~Q
example of Figure 15-2, there is a range from the uppermost z
1% of the daily maximum to the lowermost 1% of the daily - 290 -
minimum that is 3 times the mean diurnal cycle. The stan- &
dard deviation of hourly temperature averages 5 K. The 2
range from the uppermost 1% of the maximum temperature ~ 280 -
w
of the hottest month to the lowermost 1Yoof the minimum n
z
temperature of the coldest month in Figure 15-2 is about w
+ 270
2 1/2 times the mean annual cycle.
The pattern of surface temperature varies with geo-
~+
graphic location. This is illustrated by the statistics of some 2m -
~o+
widely scattered stations and even by the statistics of neigh-
boring stations (Table 15-2). The annual mean temperature
is generally lowest in the polar regions and highest in the 25t) -
equatorial belt. In addition, the mean temperature decreases
generally with elevation. The diurnal range is greatest in I I 1 I I I I I I I I
J FMAMJ JASOND
desert climates and least in oceanic or maritime climates.
MONTH
The mean annual range tends to be greatest in temperate
climates and least in equatorial climates. Figure 15-2. Surface temperature at Hanscom AF8, Mass. throughout
In polar regions, where continuous darkness (daylight) the year.

15-4
 ATMOSPHERIC TEMPERATURES, DENSITY, AND PRESSURE
Table 15-2. Temperatures at various stations around the world

Mean Mean Hottest Coldest

Annual Diurnal Annual Month 1% Month 1%
Elev. Mean Range Range of Daily of Daily
Station Lat Long m K K K Max K Min K

Hanscom AFB, Mass. 42”28’N 71”17’W 43 282.3 11.7 25.6 311 247
Boston, Mass. 42°22’N 71”02’W 5 283.8 8.6 24.3
Blue Hill Ohs., Mass. 42”13’N 71”07’W 192 282.3 9.6 24.4 —
Nantucket, Mass. 41 °15’N 70”04’W 13 282.7 7.2 20.4 —
Pittsfield, Mass. 42°26’N 73”18’W 357 280.2 11.4 25.6 — —
Worcester, Mass. 42°16’N 71”52’W 301 281.2 9.5 25.4 — —
Thule, Greenland 7@32’N 68°42’W 59 261.8 6.4 31.9 289 233
Eieison AFB, Alaska W41’N 147”05’W 170 270.2 10.8 39.6 303 224
Keflavik, Iceland 63 °58’N 22”36’W 50 278.1 4.4 11.2 291 258
Goose Bay, New Foundland 53°19’N 60°25 ‘W 44 2’73.2 9.5 33.7 307 237
Berlin, Germany 52”28’N 13°26’E 50 282.8 7.2 20.6 307 254
Limestone, Maine 4&57’N 67°53’ W 230 276.9 9.4 29,4 308 241
Boiling AFB, Wash. D.C. 38”49’N 76°5 1‘W 20 287.0 10.2 23.0 311 259
Scott AFB , Ill. 38”33’N 89°5 1‘W 138 286.1 6.2 26.3 311 250
Blytheville, Ark. 35”58’N 89”57’W 81 278.4 6.8 17.9 312 255
Riverside, Calif. 33 °54’N 117”15’W 461 292.7 16.8 14.4 314 269
Tucson, Arizona 32°10’N 11O”53’W 809 289.4 11.7 19.0 316 267
Ft. Huachuca, Arizona 31 °25’N 110”2O’W 1439 290.0 14.8 17:0 312 264
Dharan, Saudi Arabia 26”17’N 5@09’E 22 299.8 11.8 19.8 321 276
Wheeler, Hawaii 21”29’N 158”02’W 256 295.8 7.5 4.0 305 283
Honolulu, Hawaii 21”20’N 153”55’W 12 297.7 6.7 4.2 308 291
Guam, Phillipines 13”29’N 144”48’E 82 300.8 1.7 1.7 306 297
Diego Garcia Island 07018’S 72”24’E 2 300.7 3.9 2.0 305 296
Canton Island 02°46 ‘S 171”43’W 3 300.7 1.2 0.8 305 297

tropical regions characteristically have distinct diurnal and version, as much as 4 or 5 K, in the air within several feet
annual cycles. These cycles are superimposed over tem- of the ground.
perature variations caused by shifting air masses and frontal The induced temperature in military equipment exposed
passages. In tropical regions, the diurnal range rarely ex- to the sun’s heat will vary greatly with physical properties
ceeds 6 K. such as heat conductivity, reflectivity, capacity, and type
Depending on circumstances and ground characteristics, of exposure. Surface and internal temperatures, such as are
the surface air temperature could differ by several degrees induced in a boxcar, make the reading of the shelter ther-
over short distances ranging from a few meters to a few mometer only the beginning of the engineering problem.
kilometers. Also, vertical temperature variations are ob- Table 15-2 is only an initial guide to the effects of
served from a few millimeters above the ground to the top various influences on station temperature. Detailed temper-
of the instrument shelter. On windy, cloudy days or nights, ature information should be obtained from the climatological
the differences bet ween thermometer readings, within short record of each station or of stations close by. The latter
distances of one another in either the horizontal or the ver- should be modified for the influences of terrain proximity
tical, will be minimal. In high temperature regimes, how- to water, and elevation.
ever, with a bright sun and light winds, the ground surface,
especially if dry sand, can attain temperatures 17 to 33 K 5.1.2.3 Horizontal Extent of Surface Temperature.
higher than the free air. The temperature of air layers within Horizontal differences in surface temperature can arise both
a few centimeters of the surface will differ only slightly from large-scale weather disturbances and from local influ-
from the ground, but the decrease with height is rapid. The ences. Weather disturbances such as cold and warm fronts,
temperature at 0.5 to 1 m above the ground will be only thunderstorms, and squall lines account for unsystematic
slightly warmer than that observed in the instrument shelter changes in the horizontal temperature gradient. Nonuniform
at 1 or 1.5 m above the ground. Conversely, on calm, clear radiational heating and cooling of the ground also contribute
nights the ground radiation can produce a temperature in- to turbulent mixing, cloudiness, and vertical motions in the

15-5
CHAPTER 15
Table 15-3. Estimates of the horizontal scale of certain local meteorological conditions.

Horizontal Scale Temperature Differences

Local Conditions (km) (K)

Changes in Air Mass 160to 1600 3 to 22

Weather Fronts 16to 160 3 to 22
Squall Lines 8 to 80 3to 17
Thunderstorms 8 to 24 3to 17
Sea Breezes 8 to 16 ltoll
Land Breezes 3t08 lto6

lower troposphere, resulting in constantly changing tem- lower. Also, some pronounced horizontal temperature gra-
peraturesat the surface. dients occur along coastlines in temperate latitudes due to
Horizontal transport by air currents, referred to as ad- the cooling effect of coastal sea breezes.
vection, is a key factor in surface temperature differences. Generally, temperatures between two stations become
Large-scale advection will bring both the relatively dry cold more independent of one another with increasing distance
arctic air masses and the relatively moist warm tropical air (Figure 15-3). One model curve [Gringorten, 1979] for fit-
masses alternately to the temperature zones. This can pro- ting the correlation coefficient p as a function of distance s
duce large changes in the day’s mean and the diurnal range between stations is given by
of temperature.
Table 15-3 gives estimates of surface temperature dif- 2
P = ;[(COS “’ a) – V“G2] , (15.1)
ferences over varying horizontal distances associated with
several kinds of weather phenomena. Large-scale differ-
ences are greatest in winter due to the more substantial where
differential heating by solar radiation from equator to pole
and, consequently, the more intense large-scale motion of s
~=— (15.2)
the atmosphere. In summer the north-south gradients in solar 128r’
insolation are much less, but the general increase in the
amount of insolation results in more thunderstorms and other where r is the model oarameter and is in the same units as
1

air-mass activity. the distance s. The value of r is, in fact, the distance over
Systematic differences in the surface temperature be- which the correlation coefficient is 0.99. For the curve in
tween neighboring stations are due to five prime factors: (1) Figure 15-3, r = 17.7 km. While this curve could be fitted
aspect or orientation of the terrain with respect to incident by other models, the given model curve has the quality that
solar radiation, (2) type of surface structure and of soil cover
underlying the stations, (3) proximity to the moderating
influences of large water bodies, (4) elevation, and (5) dif-
1.0
ference in solar time for stations that are several hundred
.9 -
kilometers apart. Sometimes the topography permits “pools”
of cold air to drain locally at night into lower basins or
valleys. Also nonuniform distribution of water vapor and F .7
g
cloudiness will result in uneven distributions of short--wave ~ .6
w
and long-wave radiation and, consequently, uneven cooling o
v .5
and heating at the surface. z
o .4
A striking example of local influences on surface tem- F
u
perature gradients is found in the temperature contrasts be- _l .3
w
c
tween”cities and the surrounding countryside. The sheltering ~ .2
●

o
effect of buildings, their heat storage, products of fuel com- 0 .1
●
bustion, smog, rain water drainage, and snow removal all
o
*.
act to make the city a relative heat source. Thus, the city’s ..
1 I 1 1 1 1 I I 1 I 1 t
nightly minimum temperature might be 5 to 14 K higher “ ‘ o 200 400 Soo 800 1000 1200 1400 1600 Iaoo 200022002 )0

than that of surrounding suburbs. As another example, in DISTANCE (km)

hilly or mountainous terrain the valley floor could have a

Figure 15-3. The correlation coefficient of the daily mean temperature of
diurnal temperature range 2 to 4 times as great as that over Columbus. Ohio with that of nine other U.S. stations at
the peaks, and a temperature minimum from 5 to 17 K indicated distances from Columbus.

15-6
 ATMOSPHERIC TemperatUreS, DENsITy, AND PRESSURE
the correlation coefficient decreases exponentially with dis- northerly latitudes during the winter months when the tem-
tance between stations for the first few kilometers of sep- perature distributions are substantially bimodal. Thus the
aration. Eventually, the co~elation coefficient drops to zero straightforward method for determining the frequency dis-
at distance 128r. tribution of hourly temperatures is to obtain a representative
In the United States the separation between weather sample of observations for each location and compute ‘the
stations averages about 160 km, with the exception of the distributions. Estimates of the freqbency distribution from
eastern states where it is 30 to 80 km. The root mean square such data can be made using the Blom formula given by
difference of temperatures, as a function of the correlation
coefficient p between two StatidhS is approximated by n~ – 3/8
P(T) = (15.4)
N + 1/4’
rrnsd = S, ~, (15.3)
where P(T) is the estimated cumulative probability of the
where St is the standard deviation of the hourly temperature temperature T, nT is the number of observations equal to
(estimated as 5 K for Hanscom AFB). For stations 150 km or less than T, and N is the overall sample Size. Since,
apart, with p = 0.91 (Equatiod 15.1), the rmsd should be representative samples of data are not easily obtained for
approximately 2 K. regions outside North America, an objective tnethod has
been developed by Tattelman and Kantor [1977] so that the
15.1.2.4 ~unway Temperatures. At airports the desired frequency distribution of surface temperature can be esti-
length of the landing strip or runway is directly related to mated at all locations from data in climatic summaries that
air temperature. Any discrepancy, therefore, between free are available for most locations throughout the world.
air temperature over runways and shelter temperatures is Because the warmest temperatures in the world are found
important in establishing safe aircraft payloads and runway at locations where the monthly beans are high and the mean
lengths. It had been thought, on days when insolation is daily range is large, Tattelmah et al. [1969] developed an
strong, that the free air temperature over airfield landing index using these valves. The index is expressed by
strips is significantly higher than standard shelter temper-
ature over the surrounding grassy areas. Results of obser- IW = ~ + ~~~aX – ~~in), (15.5)
vations, however, over four airstrips (Easterwdod Airport,
Hearne Air Force Satellite Field and Bryan Air Force Base where IW is the warm temperature index, ~ is the monthly
in Texas, and an auxiliary naval airstrip near Houma, Lou- mean, ~~,X is the mean daily maximum, and ~~in is the
isiana) have shown that the air between 0.3 and 6 m above mean daily minimum temperature (K – 273) f~r the warmest
a landing strip is about 0.5 K cooler than indicated by the month. The index was related to temperature occuhing 1%,
shelter thermometer over adjacent grassy areas. The relative 5% and 10% of the time during the warmest months at a
smoothness of the runway surface is the physical cause of number of locations; it appears in the following regression
daytime flow of air from grass to runway. Dudngfthe tran- equations for estimating m~n~hly 1Yo, 5% and 10~o warm
sition from flow over the r~ugh grassy surface the wind temperature extremes [Tattelman and Kantor, 1977]:
speeds up and entrains the cooler air immediately above tie
runway: When a daytime equilibrium state is established, ~lx = 0.6761W + 10.657, (15.6)
there will be a large lapse rate close to the ground. This is
the effect over both concrete and blacktop .airs~ps with ~~m = 0.7331W + 5.682, (15.7)
surrounding grass having only a slightly modifying effect.
*,OW = 0.7621W + 2.902. (15.8)
In exceptional cases the free air temperature over the
runway exceeds the shelter temperature but by nd more than
where ~ is the esiidated temperature in (K – 273) occurring
0.5 ~ when averaged over 10 rein, 1 K when averaged over
1, 5, and 10% of ttle time, respectively. The same principle
1 min with a dry soil environment, and 0.25 K (5-rnin mean)
can be used to estimate cold temperature extremes. The cold
with a swamp environment. Thus the standard method of
temperature index is
temperature measurement in a properly exp~ked shelter over
grass provides a representative temperature for the esti- . . —
mations of runway length and aircraft payloads. 1= = T – (T~=X – T~in), (15.9)

15.1.2.5 Temperature Extremes. A khowledge of the where IC is the cold temperature index, ~ is the monthly
occurrence of hot and cold temperature extremes is impor- mean, ~~,. is the mean daily maximum, and ~~in is the
tant for the design of equipment and the selection of material mean daily minimum temperature (K – 273) for the coldest
that will be exposed to the natural environment. The hourly months. The corresponding regression equations [Tattelman
temperature observations at most locations are not normally and Kantor, 1977] are
distributed around the mean monthly values. Departures
from a normal distribution are largest in the temperate and t,% = 1.069L – 7.013, (15. io)

1s-7
CHAPTER 15

~~% = 1.0841= – 3.050, (15.11) Most extreme high temperatures have been recorded
near the fringes of the deserts of northern Africa and south-
~,o% = 1.0821c – 0.704. (15.12) western U.S. in shallow depressions where rocks and sand
reflect the sun’s heat from all sides. I,n the Sahara, the
This technique has been used by Tattelman and Kantor greatest extremes have been recorded toward the Mediter-
[ 1976a,b] to map global temperature extremes using loca- ranean coast, leeward of the mountains after the air has
tions for which monthly climatic temperature summaries are passed over the heated desert. The highest temperature on
available. Estimates of the 170 warm and cold temperature record is 331 K at Al Aziziyah, Libya (32”32’N, 13°1 ‘E,
extremes for the Northern Hemisphere are shown in Figures elevation 112 m). Northern Africa and eastward throughout
15-4 and 15-5. most of India is the hottest part of the world. Large areas

Figure 15-4. Temperature equaled or exceeded 1% of the time during the warmest month (K – 273) [Tattelman and Kantor, l~76a].

15-8
 ATMOSPHERIC TEMPERATURES, DENSITY, AND PRESSURE
@ w, -*
~\ @ b,

II,,
‘<.
%,

‘%

,*. &

2.*. )*
,,0”, -

Figure 15-5. Temperature equaled or colder 1% of the time during the coldest month (K – 273) [Tattleman and Katrtor, 1976a].

attain temperatures greater than 316 K more than 10% of equal to or greater than 316 K for 1Yo of the time in the
the time during the hottest month. Sections of northwest hottest month. Death Valley, within this area, has temper-
Africa experiertced temperatures greater than 322 K as much atures equal to or greater than 322 K for 1Yo of the time in
as 170 of the time during the hottest month of the year. the hottest month and it once had a record temperature of
Regions in Australia and South America have temperatures 330 K.
at and above 311 K much of the time, but do not experience Geographic areas of extreme cold include the Antarctic
temperatures greater than 316 K more than 1Yoof the time Plateau (2700 to 3600 m in elevation), the central part of
during the hottest month. The southwestern U.S. and a the Greenland Icecap (2500 to 3000 m), Siberia between
narrow strip of land in western Mexico are exceptionally 63°N and 68°N, 93”E and 160”E (less than 760 m elevation)
hot. A substantial part of the area experiences temperatures and the Yukon Basin of Northwest Canada and Alaska (less

15-9
CHAPTER 15
than 7@ m elevation). The gefierally accep[ed record low bility PT of the annual extreme temperature, it is equal to
temperature (excluding readings in Anarctica) is 205 K in 1/(1-P~) years. The return period is not to be confused with
Siberia. the planned lifetime (n) of the equipment. Roughly speak-
ing, the temperature with the 100-yr return period or the
15.1.2.6 The Gumbel Model. For equipment that is either annual 170 (PT = .99) is approximately the 10% temper-
in continuous operation or is on standby status, thereby ature of a 10-yew planned life.
continuously exposed to all temperatures, the statistic of The Gumbel distribution with a se! of periodic extremes
interest is the extreme temperature that is likely to occur is the easiest model to use, but there are reservations in its
during a full month, season, year, decade, or whatever application. Theoretically the basic distribution, such as the
period is considered to be the useful lifetime of the equip- station temperature taken hourly, should be an exponential
ment. type, such as Pearson Type 111or Gaussian. However, this
Many extreme values have been estimated effectively condition’ may not be sufficient because the record may not
by a model that has become known as the Gumbel distri- be long enough to make the annual exveme fit into a Gumbel
bution. Let us assume the annual highest temperature (Ti) distribution. The Gumbel distribution is only the limiting
has been recorded for each of N years (i = 1,N) with av- form over long times and may not be adequately reached
erage ~ and stapdard deviation s,. The Gumbel eslirnate of over short periods. It is advisable, therefote, to test the data
the cumulative probability P.r of the annual extreme high to determine if the Gurnbel distribution is applicable. Figure
temperature (T) is then 15-6 illustrates the use of special-purpose “Extreme Prob-
ability Paper” in which the cumulative probability PT is read
P~ = exp[ – exp( – y)], (15.13) on the vertical axis to correspond to T on the horizontal
axis. Alongside the scale of PT is the scale of the reduced
where variate y, which is uniform on this paper. A Gum&l dis-
hibution appears as a straight line.
Let us suppose a set of N extreme temperatures Ti for
(15.14)
each of N years (i = 1,N) is ordered from lowest to highest
value. The cumulative probability of the ith lowest tem-
where, to five decimal places perature since it is an extreme is best estimated by

~ = 0.57722 and UY= 1.28255. (15.15)

(There are pther estimates to the Gumbel distribution. This

one is preferred for its simplicity as well as degree of ac- .98
curacy. ) The quantity y is referred to as the ~educed va~iate.
.97
If one. is interested in the cold temperature, these formulas
hold with the T and ~ reversed in Equati@n(15.14). .90
If the lifetime of a piece of equipment is intended to be
n years, tiien the cumulative probability pT(n) that’the tem- a
.8
perature T will not be exceeded in the n yew is ‘

P~(n) = exp[ – exp( – y + #n n)] (15.16) .7

.5
where y is given by Equation (15. 14). Assuming, for ex-
ample, that ‘we want to estimate the te,m)erature T that has o
.3 e
only a 10% probability or risk of being excee@edover n CL
yeari, we set pT(n) equal to 0.9 in ~uation (15. 16) and .1
solve Equation (15. 16) for y obtaining
.05
y = (nn – /n(–~n P), (15.17) .01

which we in turn use in Equation (15. 14) to obtain .001

-2 I 1
i= T+st(y
-~)
Uy
(15.18) 300 305 310 315
TEMPERATURE, T (K)
The return period is a term sometimes used in associ- Figure 15-6. a) Annual highest temperature, Hanscom AFB, Mass., 2 I
ation with the extreme. In tetis of the cumulative proba- ordered values ( 1944– 1964).

15-10
 ATMOSPHERIC TEMPERATURES, DENSITY, AND PRESSURE

I I
4-
-.98
-.97

3- - .95
290K - OEW POINT
>.
c t
–.9 z
280K
u 2—
1-
a
E -.8 ; 40Y. -

s I_ -,7 a RELATIVE
a 20Y. –
n HUM IOITY
w m
$ -.5 ~
R o— 10
WI ND SPEED
a -.3 t MPS

-1- -:;5

-.01
-2- -1 .001 I I I I
265 260 255 250 245
TEMPERATURE, (K)

o~
Figure 15-6. b) Lowest temperatures of22winter seasons (1943-1965) 02 04 06 08 10 12 14 16 IS 20 22 00 02 04
ordered from warmest to coldest, Hanscom AFB, Mass.
HOUR OF THE DAY

Figure 15-7. Yuma, Arizona typical July diurnal cycles when maximum
i – 0.44 daily tempemture equals or exceeds 317 K (based on
pT = (15.19)
N +0.12 1961-1968 data).

rather than Equation (15 .4). Now in the example of Figure

15-6a, we have the plot of the annual highest temperatures The dewpoint has a median of 287 K with a small diurnal
of 21 years ( 1944– 1964 at Hanscom AFB, Mass. ) ordered range. Relative humidity, consequently, has a large-ampli-
from lowest to highest value and having cumulative prob- tude diurnal cycle. Wind speed at anemometer levels of 6
ability estimates P= given by Equation ( 15. 19). The mean to 8 m above ground averages approximately 4 ds with
is ~ = 309 K and the standard deviation is s, = 1,9 K. little diurnal range. Solar insolation, on the other hand, has
The solution of Equation (15. 14) gives the straight line plot a large diurnal range with a maximum clear-sky value of
between y and T as shown. Whether the straight line and 88.2 L/h and a minimum value of zero from 2000 LST in
therefore the Gumbel distribution adequately fits the distri- the evening till 0500 LST in the morning. For the hottest
bution is a matter of judgment. If accepted, and it should areas on earth (for example, Sahara Desert) Table 15-4
be in this example, then the 99th percentile (P-r = .99) or presents the associated cycles of temperature, relative hu-
the 19. extreme is estimated by Equations (15. 17) and ( 15. 18) midity, windspeed and solar insolation when the afternoon
with n = 1 as 315 K. For a lifetime of 25 years (n = 25) temperature in the middle of a 5-day period reaches 322 K
the temperature of 109. risk (PT = 0.9) is given by Equa- which occurs about 1YO of the time in the hottest month.
tions (15.17) and (15.18) as 316 K. Death Valley, California is also one of the hottest areas
As another example, Figure 15-6b shows the plot of the but is close to 60 m below sea level resulting in extreme
extreme low temperatures of 22 winter seasons ( 1943–1965) absorption of solar radiation before it reaches the ground.
at Hanscom AFB, Mass. The mean is ~ = 251 K and the Consequently its maximum clear-sky solar insolation of 82.5
standard deviation is s, = 3.68 K. A straight line fit of L/h is less than that shown in Figure 15-7. Solar insolation
these data is not satisfactory. Possibly a concave curve would I increases with elevation roughly in accord with the ex-
be more appropriate. The Gumbel model is not acceptable ponential model given by
in this case, and consequently another model should be tried.
1 = 11 e-’(p-pl’, (15.20)
5.1.2.7 Temperature Cycles and Durations. High tem-
perature extremes are inevitably part of a well pronounced where p and p] are, respectively, atmospheric surface pres-
diurnal cycle, modified by wind and by moisture content. sures for a given station and another reference station at
Typical of a hot climate, the record of Yuma, Arizona roughly the same latitude 11 is the solar insolation at the
(32°5 1‘N, 114°24’ W) (Figure 15-7) reveals a mean diurnal reference station, and the value for a is dependent on the
temperature range of 15.3 K for the middle 20 days in July. location. For Yuma and Death Valley, where the mean

15-11
CHAPTER 15
Table 15-4. Diurnal cycles of temperature and associated other elements for days when the maximum temperature equals or exceeds the operational 1%
extreme temperature (322 K) in the hottest month in the hottest area.

Time of Day (h)

Item 1 2 3 4 5 6 7 8 9 10 11 12

Temperature (K)
—
Hottest Day 308 307 307 306 306 305 306 308 311 314 316 317
1 day before 309 308 307 306 306 305 306 309 311 313 315 317
or after
2 days before 307 307 306 306 305 305 306 308 310 312 314 315
or after

Other Elements

Relati~e Humidity (%) 67788 886654 4

(dp = 266 K)
Windspeed (m/s) 2.7 2.7 2.7 2.7 2.7 2.7 2.7 2.7 2.7 4.3 4.3 4.3
Solar Radiation o 0000 5 23 43 63 79 90 96
(L/H)

Time of Day (h)

Item 13 14 15 16 17 18 19 20 21 22 23 24

Temperature (K)

Hottest Day 320 321 321 322 321 321 319 315 314 312 311 310
1 day before 318 320 320 321 320 319 317 315 313 311 310 309
or after
2 days before 316 317 319 320 319 318 317 314 312 311 310 309
or after

Other Elements

Relative Humidity (%) 33333 334566 6

(dp = 266 K)
Windspeed (tn/s) 4.3 4.3 4.3 4.3 4.3 4.3 4.3 4.3 4.3 4.3 4.3 2.7
Solar Radiation 96 90 79 63 43 23 5 0 0 0 0 0
(L/~)

atmospheric surface pressures are about 1006 mb and 1020 amount in June. In contrast, Table 15-6 gives corresponding
mb respectively, a ~ 0.00461 rob-’. results for the insolation at Caribou, Maine where there is
The hottest locations in the Sahara Desert are relatively much more frequent cloudiness and precipitation.
high (about 300 m above sea level) with atmospheric pres- The operability of equipment in a cold climate is very
sure about 977 mb. Thus Equation (15.20) yields an estimate much dependent on the duration of extreme cold. Unlike
for the peak solar insolation at these elevations of about 100 the hot extremes, cold extremes are usually accompanied
L/h. Most countries, however, including the U. S., Canada, by very small diurnal ranges, if any. The direct approach
and the United Kingdom, have adopted a peak figure for for determining the duration of cold temperature is by an
solar insolation for operational and design purposes of 96 analysis of hourly data. Such data are available for many
L/h. stations in North America but are not generally available
Heavy clouds and precipitation reduce the incident solar for other regions of the world. Data from 108 stations in
insolation. At a few stations the National Weather Service the U.S. and Canada have been analyzed [Tattelman, 1968]
has taken records of incoming solar insolation. Table 15-5 to obtain information on the longest period of time during
gives the results of processing such data from Albuquerque, which the temperature remained at or below eight “thresh-
N.M. It presents estimates of the probabilities with which old” values (from 273 K to 220 K) during a 10-yr period.
daily incoming solar insolation equals or exceeds the given Figure 15-8 from that report shows the results for a threshold

15-12
 ATMOSPHERIC TEMPERATURES, DENSITY, AND PRESSURE
Table 15-5. Probability of daily solar insolation equaling or exceeding
given amounts for given number of consecutive days in June,
at Albuquerque, N.M. Station elevation is 1620 m. Peak
clear sky solar insolation was observed at 910 L/day.

Insolation No. of Consecutive Days

L/Day 1 2 4 8 15 30

850 0.03
800 0.24 0.06
750 0.49 0.28 0.09
700 0.71 0.54 0.31 0.11
650 0.81 0.70 0.52 0.29 0.09
600 0.90 0.84 0.77 0.52 0.29 0.08
550 0.935 0.90 0.81 0.65 0.44 0.20
500 0.955 0.93 0.87 0.75 0.57 0.33
450 0.971 0.95 0.91 0.82 0.67 0.43
400 0.985 0.972 0.946 0.87 0.81 0.65
350 0.9933 — 0.973 0.945 0.902 0.82
300 0.9946 — 0.980 0.958 0.928 0.86
250 0.9975 — — 0.980 0.963 0.932
200 0.999 — — — — 0.970

Table 15-6. Probability of daily solar insolation equaling or exceeding , Figure 15-8. Longest duration (h) of temperature < 250 Kin ten winters
given amounts for given number of consecutive days in June, [Tattelman, 1968].
at Caribou, Maine. Station elevation is 190 m. Peak clear
sky solar insolation was observed at 843 L/day.

Insolation No. of Consecutive Days temperature of 250 K. The report also presents the expected
L/Day 1 2 4 8 15 30 (approximately 50% probability) duration of the temperature
at or below six “threshold” temperatures (from 273 K to
850
232 K) during a single winter season. Figure 15-9 shows
800 0.019
the single winter results for a threshold value of 250 K.
750 0.085 0.02
Estimates of duration have been made using data that
700 0.20 0.057
consisted mainly of daily, monthly and annual average max-
650 0.31 0.13 0.02
imum and minimum, and monthly and annual absolute max-
600 0.40 0.20 0.05
imum and minimum, for some 35 to 50 years at Siberian,
550 0.47 0.26 0,086
Yukon and Alaskan stations. The mean January temperature
500 0.55 0.33 0.13 0.024
in eastern Siberia (Verkhoyansk and Oimyakon) is 225 K.
450 0.59 0.40 0.18 0.037
Table 15-7 presents estimates of the lower 20% of the av-
400 0.66 0.50 0.26 0.075
erage temperature (averaged for durations ranging from one
350 0.72 0.56 0.33 0.12 0.02
hour to 32 days), the maximum temperature for the durations
300 0.78 0.67 0.47 0.22 0.062
shown, and the minimum temperature for the same dura-
250 0.82 0.72 0.54 0.30 0.10
tions.
200 0.88 0.80 0.66 0.42 0.22 0.05
The duration of temperature anywhere, hot or cold, is
150 0.921 0.87 0.77 0.59 0.37 0.13
of general interest. In the midlatitude belt the temperatures
100 0.965 0.943 0.90 0.79 0.63 0.40
of Minneapolis, Minn. are typical (Figure 15-10). The Jan-
90 0.975 0.96 0.92 0.84 0.72 0.51
uary probability distribution of all hourly temperatures has
80 0.980 0.962 0.93 0.86 0.75 0.56
a 1% value of 244 K, and a 50% or median value of 263
70 0.987 0.978 0.956 0.912 0.83 0.70
K. That is, the range from the lower 1% to the median is
60 0.9931 — 0.971 0.947 0.903 0.82
19 K. The 24-h averages, as expected, have a narrower
50 0.9961 — — 0.977 0.95 0.90
range, 17 K. The range of monthly averages (768 h) is much
40 0.99906 — — — — 0.97
narrower, 7 K. Similarly, the July hourly temperatures have

15-13
CHAPTER 15

.2 1 HOURS

Figure 15-10.
3 6 12 24 46 96 192 364

The distribution of the averages of corraeeutive hours of

temperature at Minneapolis, Minn. (The upper haff is for
768

mid-summer month, from I July to 1 August; the lower

half is for mid-winter month, 1 January to 1 February.
Each curve is labeled with percent probability of oecur-
mnce. )

month of January, and 0.919 in the midsummer month of

July.
Hourly observations have been taken at Minneapolis for
many years, making many useful summaries possible. Fig-
ure 15-11 shows a sample distribution (1949-1958) of hourly

;’-”m:,;$:’::(:\”’’”’”-
January temperatures alongside the left axis and the distri-
bution of m-hour minima over the body of the graph, m
1oo. ‘+ . go. .. ’.: ’’70. -. ..-” from 1 hour to 768 hours (1 Jan to 1 Feb inclusive). Figure
15-12 shows the sample distributions of hourly temperatures
in January of m-hour maxima. As an example of the use-
Figure 15-9, Expected longest duration (h) of temperature s 250 K dur- fulness of such a chart, freezing conditions (s273 K) are
ing a single winter season ITattelman, 1968].
shown as 9490 frequent for 1-h durations. For 24 consecutive
hours this frequency reduces to 83%, for 8 days (192 h) to
42% and for 16 days (384 h) to 10%.
a 12 K range from the 5090 value of 295 K to the upper
1% value of 307 K. The 24-h averages have an 8 K range
15.1.3 Upper Air Temperature
and monthly averages have only a 2 K range. These figures
imply a relatively high hour-to-hour correlation. Correlation
The temperature data discussed in this section are from
analysis has provided estimates of 0.982 in the midwinter
direct and indirect observations obtained from balloon-borne

Table 15-7. Durations of cold temperatures associated with the 222 K .96
extreme. Each temperature in this table is the maximum,
,95 -
average, or the minimum in an operational time exposure of
m hours, with 2070 probability of occurrence, during Janu- .9 -
aV, in a Siberian cold center.
: .a -
Time m(h) i

a
263 u
I 136 12 24 48 96 192 384 768 m
0
.6 -

E .5 =

Maximum y .4 -

k
Temperature 222-’-223 224 225 226 228 230 334 238 241.
-1
K ~ .2-
3
v
Average .1-
Temperature 222 222 222 222 222 222 222 223 223 224
.0s -
K
.Oe
t HOUR6 3 6 12 24 4a 96 192 S64 766
Minimum
Temperature 222 222 220 219 217 216 215 213 211 210
K Figure 15-1 I. The cumulative probability of the M-hour minimum tem-
perature (1 January to 1 February) at Minneapolis, Minn.

15-14
 ATMOSPHERIC TEMPERATURES, DENSITY, AND PRESSURE

60 -

L \ \ 52 -

1- \“~n. \
\ I
44 -

F
.
536 –

.05
z

z 28 -
JAN —
1 HOURS 3 e 12 24 48 96 192 204 766 Am –––––
JULY— —
ocT -------
Figure 15-12. The ffequency of dumtion (h) of the temperati (= T) in 20 –
the mid-winter month (I January - I February) at Minne-
apolis, Minn. (based on 1943-1952 data.)
12 –

instruments, primarily radiosondes, for altitude Up to 30 km

and frqm rockets and inst~ments released from rockets for 4–

altitudes between 30 and 90 km. I I I I I I I I 1

1s0 200 220 240 260 2s0 300
TEMPERATURE (K)
15.1.3.1 Seasopal and Latitudinal variations. The
Reference Atmospheres presented in Chapter 14 provide
tables of mean monthiy temperature-height profiles, sufiace
to 90 km, for 15° intefials of latitude between the equator
and North Pole. These profiles depict both the seasonal and

‘7
r
~—~-– -,—-, ~-–
.7 ..– ~--.~– -- ~q

ASCENSION TEMPERATURE

60 –
1
60 –
‘%

52 –
52 –’

44 –
44 –

~ $3 _
36 –
g
E
~ ~

>2 8-
52 8- Arn -––––
JULY — — JWY— —
~T_. .–...
m-----

2 o–
2 o–

I2–
I2–

4–
4–
I I I I I I I I
I I I I I I I I I I I
180 200 220 24Q 260 260 300
Ieo Z@ 220 240 260 2W 300
TEMPERATURE (K)
TEtiPERATURE (K)

FlguR 15-13. Seasonal differences in the temperature-altitude profiles at Ascension Island, Wallops Island, and Ft. Chumhill.

15-15
CHAPTER 15
Table 15-8a. Median, high, and low values of tem~ratures for January Table 15-8b. Median, high, and lowvalues oftemperatures for Januaw
. . ..
“..

and July at 30”N. and July at 43-N.

170 10% 2070

kltltude Median High Low High Low High Low Altitude Median High Low High bw High Low
(km) (K) (K) (K) (K) (K) (K) (K)
~
January January

5 262 272 251 267 256 265 258 5 250 263 233 257 239 254 242
10 229 239 219 235 223 233 225 10 220 233 206 227 212 225 214
15 208 221 198 216 203 214 205 15 217 231 202 225 208 222 211
20 208 222 200 216 203 214 204 20 215 227 203 222 208 220 210
25 220 231 210 226 216 224 217 25 215 233 197 226 205 224 209
30 229 239 218 236 224 234 226 30 221 240 209 230 214 226 219
35 240 254 222 248 232 245 235 35 233 258 215 251 223 243 226
40 252 270 240 262 249 258 250 40 247 272 226 264 236 257 240
45 264 283 253 277 258 272 260 45 262 288 240 283 250 271 254
50 266 281 256 276 260 273 262 50 265 282 249 274 256 270 258
55 254 272 231 267 243 263 248 55 253 275 229 267 239 263 245
60 243 254 223 248 232 246 235 60 244 266 220 263 230 257 241
65 231 254 218 242 226 238 228 65 235 255 214 246 223 243 228
70 220 235 198 227 204 225 210 70 226 246 206 238 211 234 217
75 218 253 197 237 203 227 208 75 225 261 197 245 205 235 210
80 209 243 187 230 194 217 197 80 216 248 185 237 197 228 202

July July

5 272 278 262 274 266 275 268 5 267 277 255 274 259 272 262
10 238 249 227 246 232 242 234 10 235 247 222 240 227 239 230
15 204 216 196 211 200 210 200 15 216 227 205 222 206 220 212
20 212 223 203 218 206 216 206 20 219 233 207 227 213 225 215
25 223 230 216 ~27 218 226 219 25 225 233 216 229 217 228 221
30 234 241 226 238 229 236 231 30 234 242 228 239 231 237 232
35 244 254 237 250 240 247 242 35 245 254 238 250 241 248 243
40 256 267 247 263 251 261 253 40 256 268 250 265 254 263 255
45 266 275 259 272 264 269 265 45 268 280 260 276 263 272 265
50 269 282 258 278 262 275 264 50 273 283 264 279 268 277 270
55 264 273 247 269 253 267 256 55 264 273 249 269 255 267 260
60 247 262 231 255 240 252 243 60 24? 270 230 2M 235 260 238
65 228 240 215 236 219 234 222 65 230 245 216 241 223 238 220
70 209 222 186 219 194 214 200 70 213 226 188 219 196 216 202
75 200 218 178 214 192 209 196 75 195 210 175 205 186 201 i90
80 193 207 182 200 189 198 191 80 183 203 154 195 163 191 170

latitudinal variations in mean monthlv temperatures. The Temperature-altitude profiles, surface to 60 km, for the mid-
largest seasonal variations in “temperat~re oc~ur at altitudes seas~n months at As~ension Island, 8°S, Wallops Island,
between 70and 80kmne~75”N latitude. In this region 38°N, and Ft. Churchill, 59°N, aregiven in Figure 15-13
the mean monthly temperature fluctuates from 230 K in and illustrate the magnitude of the seasonal and latitudinal
January to 160K in July. Intheupper mesosphere, 60to variations in mean monthly temperatures.
85 km, mean monthly temperatures decrease toward the pole
in summer and towards the eqtiatorin winter. In the up~r 5.1.3.2 Distribution Around Monthly Means and
stratosphere, 20 to 55 km, conditions are reversed; tem- M@w. Thedistributionsofobservedtempemti~sarouiid
perature decreases toward the @le in winter and toward the the tiedian values for altitudes upto 80km in January and
equator in summer. At altitudes between 15 and 20 km July at 3W,45°, 600and750N areshown in Tables 15-8a
temperature decreases toward the equator in all seasons. to 15-8d. Median, andhigh andlowvalues that are equaled

15-16
 a
ATMOSPHERIC TEMPERATURES, DENSITY, AND PRESSURE
Table 15-8c. Median, high, and low values of temperatures for January Table 15-8d. Median, high. and low values of temperature for January
and July at-600N. and July at-750N.

Altitude Median (K) (K) (K) (K) (K) (K)

m ~
January January

5 240 255 225 249 231 246 234 5 235 246 222 241 229 238 230
10 217 231 203 224 209 222 211 10 214 224 202 219 207 217 209
15 217 231 203 225 209 222 212 15 209 219 195 213 201 211 203
20 215 236 194 226 204 222 208 20 204 225 179 215 189 210 194
25 212 241 185 229 197 223 203 25 205 233 181 221 193 216 198
30 216 253 203 235 204 225 210 30 209 255 194 231 198 224 202
35 221 277 204 259 209 238 214 35 219 256 199 249 210 236 213
40 227 300 206 278 211 246 219 40 229 284 207 256 219 248 224
45 243 303 219 282 225 255 231 45 239 281 203 264 224 260 233
50 251 289 226 280 240 271 245 50 249 282 201 265 225 259 229
55 251 283 225 275 233 256 238 55 255 291 208 262 221 253 226
60 243 271 210 261 224 253 234 60 247 303 206 263 213 255 219
65 238 262 208 258 218 249 222 65 238 310 186 277 202 263 209
79 239 264 212 253 219 249 225 70 242 297 166 277 201 261 207
75 232 255 180 249 203 246 213 75 234 289 183 259 201 261 207
80 223 248 173 243 195 239 204 80 224 277 165 254 194 240 201

July July

5 260 271 250 266 254 264 256 5 254 264 244 259 248 257 250
10 225 238 214 233 219 231 221 10 229 238 219 234 223 232 225
15 225 235 217 231 221 229 223 15 230 237 225 235 228 233 229
20 225 233 219 230 222 229 223 20 230 237 227 235 228 234 229
25 229 236 222 233 225 232 226 25 230 240 226 238 227 237 229
30 239 245 232 243 234 241 235 30 243 262 233 247 235 246 240
35 252 258 243 256 247 253 248 35 256 262 238 260 246 258 250
4Q 265 272 259 269 263 268 262 40 268 275 252 271 260 270 262
45 277 287 271 283 274 280 275 45 281 292 268 287 275 284 278
50 279 290 273 286 277 284 279 50 284 296 270 291 279 288 280
55 271 278 257 275 264 273 266 55 281 288 254 284 270 283 275
60 259 273 212 265 250 263 253 60 268
65 238 259 225 253 230 248 233 65 246 (insufficient data above 55kmin
70 214 239 202 226 208 222 211 summer)
75 190 202 178 196 182 194 186 70 218
80 166 180 142 176 153 174 155 75 189
80 161

or more severe 1%, 10%and 20% of the time during these

months are given for 5-km altitude increments between the the paucity of data and larger observational errors at the
surface and”80 km. Distributions below 30 km are based higher altitudes. Only median temperatures are given above
on radiosonde observations taken in the Northern Hemi- 55 km at75°N for July dueto the small number ofobser-
sphere, and those above 30kmare based on meteorological vations that are available for the higher altitudes over the
and experimental rocket observations taken primarily from polar regions in summer.
launching sites in or near North America. The 1% values In tropical regions, 0° to 15° latitude, the day-to-day
are considered to be rough estimates as they are based on variations of temperature are normally distributed about the
the tails of the distributions of observed values plotted on mean at altitudes upto50 km. Consequently, a reasonably
probability paper. Estimates of values for altitudes above accurate estimate of the distribution of temperature at a given
50 kf-r-tare less reliable than those below 50kmbecauseo{ altitude can be obtained from the standard deviations and

15-17
CHAPTER 15
Table 15-9. .
Standard deviations of observed day-to-dav. variations in 5.1.3.3 Distributions at Pressure Levels. The mean
temperatures (K) at Ascension Island (8°S) at altitudes up
January and July temperatures over North America for stan-
to 50 km during the midseason months.
dard pressure levels up to 10 mb (=31 km) are presented
S. D. of Temperature (K) in Table 15-11. Standard deviations of, the daily values
Altitude around these means are also shown, thereby providing in-
(km) Jan April July Ott formation on seasonal changes in monthly mean tempera-
5 0.8 0.6 0.7 0.6 tures and interdiumal (day-to-day) variability at various
10 0.8 1.0 0.8 1.1 pressure levels and latitudes. Standard deviations are not
15 1.6 2.0 1.9 1.5 shown above 100 mb north of 50° latitude because a bimodal
20 2.2 2.2 2.4 2.1 temperature distribution exists in the winter stratosphere in
25 2.2 2.2 2.7 2.1 arctic and subarctic regions over eastern North America. As
30 3.1 2.8 3.8 3.6 a result, the standard deviations do not provide reliable
35 3.7 3.2 3.7 3.8 information on the temperature distributions at these levels.
40 5.2 3.9 3.3 3.5
45 3.6 2.8 3.2 3.3
50 5.8 2.9 3.9 3.0 15.1.3.4 Interlevel Correlation of Temperature. The
manner in which the correlation between temperatures at
two levels decreases (or decays) with increasing separation
the monthly means. The standard deviations of observed between the levels is an example of the general problem of
temperatures around the mean monthly values for the mid- correlation decay. Correlation decay is similar for most me-
season months at Ascension Island, Table 15-9, are typical teorological elements as the horizontal or vertical distance
of the day-to-day variations found in the tropics. Values are between the points of observations increases. As yet, no
not given for altitudes above 50 krn as there are too few fully satisfactory description of the decay rate, based on
daily observations on which to base the monthly temperature fundamental properties or assumptions, is available. Con-
distributions. The observed standard deviations includes the sequently, many empirical models that are valid for specific
rrns instrumentation errors as well as the actual rms climatic elements over restrictive ranges have been proposed.
variations. Consequently, the observed variations are some- Profiles of correlation coefficient r, of surface temper-
what larger than the actual values. ature with temperature at other altitudes are shown in Figure
Day-to-day variations of temperature around the annual 15-14 for the midseason months at Ascension Island, Kwa-
mean at levels between 50 and 90 km in tropical areas (Table jalein, Wallops Island, and Ft. Churchill. At most locations,
15- 10) were computed from data derived from grenade and the correlation between surface temperatures and tempera-
pressure-gage experiments at Natal, 6S, and Ascension tures at other altitudes decreases rapidly with increasing
Island, 8°S. These data were not uniformly distributed with altitudes, reaching a minimum or becoming negative be-
respect to season or time of day. An analysis of the relatively tween 12 and 16 km and then remaining near zero, plus or
sparse data that are available for individual months indicates minus 0.3, from 20 to 60 km. Individual arrays of the mean
that if the seasonal and diurnal variations are removed from temperatures, standard deviations and interlevel correlation
the data, standard deviations around monthly means due to coefficients for altitudes to 60 km are given in Table 15-
day-today changes in synoptic conditions would be roughly 12a to 15- 12f for the months of January and July at Ft.
50% of those given in Table 15-10. Churchill, Wallops Island, and Kwajalein. Additional in-

Table 15-10. Standard deviations of observed densities (%) and tem~ratures (K) around the mean annual values of AscensionIsland (8°S)/Natal (6°S).

Altitude Density Temperature

(km) S.D. (% of mean) S.D. (K) No. of Observations

50 4.1 6 33
55 4.3 3 33
60 4.8 6 33
65 4.7 7 33
70 6.4 9 32
75 8.6 10 31
80 7.8 10 30
85 10.2 13 29
90 12.3 21 28

15-18
 ATMOSPHERIC TEMPERATURES, DENSITY, AND PRESSURE
Table 15-11. Mean temwrature and standard deviation at standard pressure levels over North America.

Mean Temperature and Standard Deviation (K)

Pressure 2WN 30”N 40”N 50°N 60”N 70”N ~ 80”N

(rob) Mean S.D. Mean S.D. Mean SD. Mean SD. Mean SD. ‘can ‘“D. ‘can ‘“D.

January

700 280 2 275 5 267 6 256 8 251 8 247 7 245 6

500 264 3 259 4 252 6 242 7 238 7 234 5 232 5
300 236 3 232 3 227 4 221 4 220 4 217 4 214 5
200 217 3 216 5 216 6 219 7 219 7 216 6 213 6
100 198 3 204 4 212 4 218 –5 219 6 216 7 210 6
50 208 3 209 3 213 3 215 4 216 * 213 * 206 *
25 218 2 218 3 216 4 212 5 212 * 208 * 203 *
15 225 2 223 3 221 4 218 6 215 * 211 * 207 *
10 230 2 227 3 224 4 221 –6 217 * 213 * 209 *

July

700 283 2 283 2 282 3 275 4 270 3 268 4 265 4

500 267 2 267 2 264 3 258 4 255 4 253 4 249 4
300 240 2 240 2 237 3 232 4 229 4 228 4 227 4
200 218 2 218 2 219 3 221 5 224 5 226 5 230 4
100 200 3 203 3 210 3 220 4 226 3 228 2 231 2
50 213 2 215 2 218 3 221 3 226 3 228 3 230 3
25 222 2 222 2 225 2 227 2 229 2 232 2 233 2
15 228 2 228 2 229 2 232 2 235 3 236 3 236 3
10 232 2 233 2 234 2 237 2 239 3 240 3 241 3

Not normally distributed.

formation useful in design studies is given by Cole and considerable variability produced by small-scale terrain fea-
Kantor [ 1980]. tures, differences in soil moisture and cultivation. A snow
surface is markedly affected by aging. The physical con-
ditions of water in a shallow puddle are quite different from
the open ocean. All these conditions reflect themselves in
15.1.4 Speed of Sound vs Temperature the micro-climatological aspects of natural or unnatural sur-
faces.
The speed of sound is primarily a function of temper-
As discussed in Section 15.1.2, the use of ordinary
ature. An equation for computing the speed of sound and
thermometers to measure surface temperature, will result in
the limitations of such computations are presented in Chap-
meaningful values only in the rare cases of a flat, uniform,
ter 14. Figure 15-15 shows the relationship between tem-
and homogeneous surface. In general, area averages of tem-
perature and the speed of sound. It can be used with the
perature obtained by an integrating method over certain
various temperature presentations given in this section to
defined sections will be more representative than any one
estimate the probable speed of sound for various altitudes
of a multitude of widely varying point values. Bolometric
and geographical areas.
temperature measurements from an airplane cruising at low
altitude provide a more reasonable approach to the problem
of surface temperature determination than a series of ther-
15.1.5 Earth/Air Interface Temperatures mometric point measurements. Table 15-13 lists some re-
sults of bolometric measurements from an airplane. The data
The earth/air interface is either a land, snow, or water illustrate the great horizontal variability of surface temper-
surface. At many locations, the physical structure of the ature even when effects on the scale of less than 6 m linear
interface is overwhelmingly complex. The land surface can dimension are averaged out.
be covered with seasonally varying vegetation of great di- The processes that determine the temperature of the
versit y, and even without plant cover there is normal! y a earth/air interface and the surface characteristics that influ-

15-19
CHAPTER 15
I I I I I I I I I I I 1 I I I I I 1 I I —1
JANUARY TEMPERAT~ APRIL TEMPERATURE

60 60 –

52 I 52 -
,. ( ‘, 1

.’ [\
,,,
:/ /’
44 –
j
<,

F
z 36
\
‘“‘[ z
] 36 –

u
0
2
ASCENSION
KwxlN
‘—
‘ --— A z
~
~
~ 28
WALLOPS
CHURCHILL — ;;’\ a 28 –

I ‘-----
,.-!
)’” > ;. -
ASCENSION —— ‘-
t / ‘ /’ KWMLEIN —-.—
20 20 – WALLoPs ______
CMCHILL —
t t
<:
... .
t ----
12 –
--
2
} F <:’ “
4 4 -

I I I I ~ P+-+= I I I I I
-1.0 -0.8 -0,6 -0,4 -0.2 0 0.2 0.4 0.6 0.8 1.0 -1.0 -C18 -0.6 -0.4 -0,2 0 0.2 0.4 0.6 0.8 I.O
COEFFICIENT OF ~L~lON WITH SURFA~ TEMPERATURE =FFICIENT OF -ELATION WITH SURFACE TEMPERATE

,——.r. -. ~—~~ ~~
I
I
I IJUtY TEMPERATURE OCTOBER TEMPERATR
1
I
60 A
60 . “ \ ,/
,; .Y / /(
.,’
/, ‘{
‘\ ,, 52
52 “ < 4
‘\ ‘,
:. ./”
< ‘,
\
‘\ ,, .:.
‘.\ j
44
44 .- ‘\
\\\ . .
<“”” ~
. . .. -’7 ,’
1 ~ ~~”” w 1: :/
F 0
3s .,>
~ 36 <---- )
‘. ..
u E
n AXENSION —— > <
d
( KWULEIN — -. —
<
ASCENSION —— ‘.> ‘ WALLOPS ______ ,,‘\ \
b 28 (
a 28 KW~LEIN — -‘— CHURCHILL —
------ .
WALLOPS 1
1 CWRCHILL —
\“ ~J,
~“’’’”>: :;’
I. 20
20 ,/
,/ ‘.,
. ..
... “ >
‘,.
.. -----
12 .:
12 ; .\ ------
‘.\ -,

. .
‘.\.’!.
‘.. 4 . “~.
4 (
‘ . ... . Q / ‘.,

~ -- [Ill; ~:
[l! i
-1.0 -0.8 -0,6 -0,4 -0.2 0 0.2 0.4 0.6 0.8 1.0 -1.0-08 -0,6 -0,4 -0.2 0 k+ 0.6 0.8 1.0
=FFICIENT OF mRELATION WITH SURMCE TEMPERATE C~FFICIENT OF -ELATION WITH SURFACE TEMPERATURE

Figure Figure 15-14. Vertical profiles of interlevel coefficients of correlation of surface temperature with tempratum at other altitudes up to 60 km for
the mid-season months at Ascension Island, Kwajalein, Wallops Island, and Ft. Churchill.

15-20
 ATMOSPHERIC TEMPERATURES, DENSITY, AND PRESSURE
ence these processes may be separated into the following perature is shading. Thin roofs (metal, canvas), however,
four classes: may attain a temperature so high that the under surface acts
1. radiative energy transformation (or net radiation in- as an intense radiator of long-wavelength radiation, thus
tensity), which depends upon the albedo and selective ab- acting to warm the ground. In hot climates< multilayer shades
so~tion and emission; with natural or forced ventilation in the intermediate space,
2. turbulent heat transfer into the air (by both convective or active cooling of the outer surface by water sprinkling,
and mechanical air turbulence); can be used to cool the ground with some success. Table
3. conduction of heat into or out of the ground, which 15-15 compares temperature measurements of various ma-
depends upon the thermal admittance of the soil; and terial surfaces with corresponding air and soil temperatures.
4. transformation of radiant energy into latent heat by
evaporation, which depends upon the dampness of the sur-
face or available soil moisture at the ground level.
The aerodynamic roughness of a natural surface strongly 15.1.6 Subsoil Temperatures
influences the momentum exchange between ground and air
flowing past it. The momentum exchange establishes the The thermal reaction of the soil to the daily and seasonal
low-level profile of mean wind speed. The mechanical tur- variations due to the earth’s rotation and its revolution about
bulence produced by surface roughness also determines to the sun of net radiation is governed by the molecular therma[
a certain degree the relative amount of heat transported into conductivi~ of the soil, k, and by the volumetric heat ca-
or from the air at mean ground level. Other conditions being pacity of the soil, C = pc (where p is the density and c is
equal, an increase in roughness and hence mechanical tur- the heat capacity per unit mass). For a cyclic forcing func-
bulence will cause lowering of maximum surface temper- tion of frequency n, the quotient (nk/C) 1’2(which has the
ature during daytime and raising of minimum surface tem- physical units of velocity) determines the downward prop-
perature during nighttime. For ordinary sandy soil, under agation or amplitude decrement with depth of the soil-tem-
average conditions of overall aiflow and net radiation on perature response. The product (nk C)-1’2, which has the
summer days in temperate zones, the diurnal range of sur- physical units of degrees divided by Langleys per unit time
face temperature is about 17 K if the roughness coefficient (1 L/s equals 4.186 X 104 W/m’), governs the amplitude
is 0.06 mm or 14 K if it is 6.35 mm (roughness coefficient, of the temperature profile in time at the soil surface. The
also called roughness “length,” is E/30 where e is the av- ratio k/C is the thermal dlflusivi~ (physical units of length
erage height of surface irregularities). squared per unit time). The expression (kC) 1’2defines ther-
A special and rather extreme case of the influence of mal admittance of the soil.
surface characteristics is represented by forests. The trees The continuous flow of heat from the earth’s hot, deep
intercept solar radiation and the heat absorbed is given off interior to the surface is the order of 10-5 L/rein. This is
into the air that is trapped between the stems. Although deep very small compared with a solar constant of 2 L/rein,
snow may lie on the ground, daytime temperatures in waled average net-radiation rates of 0.2 L/rein, and induced soil-
areas ‘in spring can reach 289 K. heat fluxes in the uppermost several feet of the earth’s crust
The thermal admittance (Section 15. 1.6) of most soils of 0.1 L/rein. Only for depth intervals in excess of about
depends on porosity and moisture content. Because both the 30 m must the heat flow from the earth’s interior be con-
thermal conductivity and heat capacity of soils increase with sidered, inasmuch as it results in vertical temperature gra-
soil moisture, the thermal admittance may be significantly dients of the order of 2.5 to 25 K/km.
affected by humidity variations during rainy or wet weather Table 15-16 gives experimental data on thermal admit-
periods, whereas the normal diffusivity may remain unal- tance and theoretical values of the half-amplitude depth
tered. These effects are difficult to assess, however, because interval based on experimental thermal diffusivity data for
the dampness of the surface is also a major factor in the diverse ground types. The smaller the thermal admittance,
utilization of solar energy for evaporation. If soil moisture the larger the surface-temperature amplitude for a given
is readily available at the earth’s surface, part of the net forcing function. This latter inverse proportionality is valid
radiation that would have been used for heating air and only when turbulent heat transfer into the atmosphere is
ground is used instead for latent heat of evaporation, Table negligible.
15-14 lists observed temperatures in the air and soil at levels In a simple theoretical model of thermal diffusion, an
close to the earth/air interface. effective atmospheric thermal conductivi~ K is introduced.
Engineers must consider the effect of albedo and color For air, K is many times larger than the molecular thermal
or net radiation in artificially changing surface or ground conductivity y of the air. For the same forcing function, the
temperature. In India, a very thin layer of white powdered surface-temperature amplitudes at two different kinds of
lime dusted over a test surface made ground temperatures ground follow the ratio
up to 15 K cooler; the effect was felt at a depth of at least
20 cm. (TAR)2 + (K/k)j~:
(15.21)
Another effective method of co~trolling surface tem- (TAR), + (~)l!;

15-21
CHAPTER 15

15-22
ATMOSP~lZIC TEMPERA~, DENSmY, AND PRESSURE

15-23
CHAPTER 15
 a
ATMOSPHERIC TEMPERATURES, DENSITY, AND PRESSURE
I

15-25
CHAPTER 15

15-26
 a
ATMOSPHERIC TEMPERATURES, DENSITY, AND PRESSURE

15-27
 a
CHAPTER 15
collations” [Lettau, 1954b]. The best insulator is still air or
any porous material with air-filled pores, such as feathery
200 -
snow; materials such as leaf litter have similar insulating
properties [Geiger, 1957].
220 - Much information is available on so~l-temperature var-
G
iations in various climatic zones. Table 15-17 gives annual
w
Kz60 - and daily temperature cycles in different soil types. In ad-
:
K
dition to the type of ground, certain meteorological factors
~2’” -
such as rainfall and melting snow have marked effects on
1- the soil temperature. Snow cover is a leading factor in pro-
Zzo - tecting the soil from severe frost. On one extreme occasion
with an air temperature of 255 K, the temperature was 272
200 - K under a 13 cm snow cover, whereas on bare soil it was
251 K.
270 200 290 300 310 320 ~ 34o :
The soil-temperature variations illustrated in Figure 15-
16 were obtained at a station cleared of pine trees but in
SPEED OF SOUNO (m /see)
generally wooded country. Topsoil and brown sandy loam
Figure 15-15. Speed of sound vs temperature (K). (Oto 0.6 m) changed to brown sand and gravel that varied
from medium (0.6 to 2 m), to coarse (2 to 4 m), and again
to medium (4 to 18 m). The water level was at 15 m. The
where TAR represents the ratio of the thermal admittance figure illustrates the amplitude decrease and phase retar-
of the ground to that of air. For diurnal cycles of net ra- dation of the annual cycle with depth. Amplitudes of weather
diation, the ratio (K/k).ir is the order of 104. disturbances with periods of several days, as illustrated by
The most extreme surface temperature oscillation occurs the temperature curve of the O.75-m level, decrease with
over feathery snow where the amplitude may reach ap- depth more rapidly than the annual amplitudes. Qualita-
proximately four times that over still water or sandy soil, tively, this agrees with the theoretical prediction of an am-
and is at least 100 times as large as that over the turbulent plitude decrement proportional to the square root of the
ocean. An amplitude ratio of about 3.5 can be expected for length of the period of oscillation. The actual half-amplitude
surface temperature over dry vs moist sand surfaces. The- depth interval of the annual cycle can be estimated from
oretically, the penetration of thermal “oscillations” into the Figure 15-16 as being nearly 3 m, which is much larger
soil is inversely proportional to the frequency of the “os- than the depth inferred from experimental values of thermal

Table 15-13. Bolometric records of area (approximately 37 mz) surface temperature from an aitplane cruising at approximately 370 m along a constant
flight path, April 1944 [condensed from Albrecht, 1952].

Surface Temperature (K) – Bolometric Data

Standard
Sun’s Shelter Temp.
Elevation Sky at Airport Baltic Sand Down Opening
Day Hour (degree) Cover K Sea Beach Land City woods in Woods

9 13 to 14 40.4 10/10 282 275 285 287 281 280 281

11 10 to 20 –1.6 1/10 283 275 280 275 276 276 278
16 19 to 20 –1.8 9/10 287 277 281 280 280 277 275
20 05 to 06 0.5 1/10 275 278 271 267 272 270 266
29 14 to 15 42.1 4/10 280 280 319 315 290 289 296

Wind Speed Clear Cutting Dry

ds woods in Woods Peat Swamp Pond

7 19 — 4/10 0.5 275 271 267 273 274

20 20 2/10 1.5 273 272 269 273 273
26 20 — 3/10 2.6 275 270 272 272 274

Albedo values as determined by Albrecht: 5% 8% 8% 7% 5V0

15-28
 a
ATMOSPHERIC TEMPERATURES, DENSITY, AND PRESSURE
Table 15-14. Temperature of the air 10 cm above, and of the soil 0.5 cm below, the earttiair interface measured by thermocouples [Davidson and Lettau.
1957].

Temperature (K) at Mean Local Time

Condition 0400 0600 0800 1000 1200 1400 1600 1800 2000

*Air 287.7 289.8 296.0 300.7 303.7 304.7 305.0 301.9 297.5
*Soil 290.6 290.9 296.0 304.1 308.6 309.2 306.8 302.3 298.5
**Air 281.1 282.5 291.5 297.9 301.7 303.0 301.1 296.1 292.3
**Soil 284.8 284.6 291.4 303.1 310.8 310.3 304.3 297.8 293.8
*Mean soil moisture in O to 10 cm layer 10% wet weight basis.
**Mean soil moj~tum in o to ](I cm layer about 4~o wet weight basis

diffusivity. The discrepancy may becaused by seepage or to conserve on the air conditioning load, or to dissipate heat
downward migration of rain water and the accompanying generated by power cables. The determination of frost pen-
advection or transfer of heat. This process could increase etration depths is usually the principal concern.
the apparent or effective thermal diffusivity for annual soil-
temperature variations by factors of 4 to 8 times the ex-
perimental values obtained in soil of constant moisture. The 15.1.7 Degree-Day and Temperature-Wind
data in Table 15-17 are more in line with experimental Combinations
findings than the curves in Figure 15-16. The limitation of
Table 15-17 is that the data are for clearly defined and nearly A degree-day is a unit adopted to measure the departure
ideal soil types that are seldom matched by actual ground of the daily mean temperature from a given standard. In the
conditions. United States the number of hearing degree-days, on any
Factors that must be investigated and assessed for any one day, is the number of Fahrenheit degrees of the 24-h
one set of soil-temperature observations are (1) type and mean temperature below 65°F (291 K). Cumulated, day by
state of compaction of the soil, (2) moisture content and day, over the heating season, the total number of degree
seepage of the soil during the test, (3) position of the water days becomes an index of heating fuel requirements. In such
table during the test, (4) type and color of surface cover, cumulation, the days on which the mean temperature ex-
(5) amount and nature of traffic over the site, and (6) local ceeds 65°F (291 K) are ignored. When the centigrade scale
climatic conditions. is used, the base is usually 19°C (292 K). The United States
Subsoil temperature information is useful in computing Army Corps of Engineers computes “freezing-degree days”
thermal stresses and loads. Some examples are the deter- as the departure of the daily mean temperature from 32°F
mination of the depth to which a structure should be buried (273 K), a negative departure when above 32°F (273 K).
when proximity to natural isothermal conditions is desired The National Weather Service supplies “normal degree-days,”
both monthly and annual totals. A few examples of the 30-
year annual normals are 9274(F) for Fargo, N. D., 5634(F)
Table 15-15. Comparison of air and soil temperature with surface tem- for Boston, Mass. and 108(F) for Key West, Florida.
peratures of materials exposed on a tropical island with
normal trade winds. Air and material surface temperatures The wind-chill concept was introduced in 1939 by the
at 1.2 m above, soil temperature at 2.5 cm below, the earth/ famous antarctic explorer, Paul Siple, to measure the cool-
air interface. Exposed surface area about 930 cmz [Draeger ing effect of low temperature and strong wind combined.
and Lee, 1953].
The wind-chill index is the equivalent temperature, in a
I Temperature (K) normal walk (1.9 ds) in calm air, comesponding to the
combination of actual air temperature and windspeed. It can
Highest Average be related to the heat loss H from a nude body in the shade.
Material Recorded Max. Min. H is given by
Air (1.2 m) 302 301 299
307 299 H = (10 ti + 10.45 – V)(306 – T,), (15.22)
Soil (2.5 cm) 307
wood 314 310 298
309 298 where H is the heat loss in kilogram calories per square
Aluminum 313
318 311 298 meter of body surface per hour, T, is the air temperature
Galvanized Iron
Black Iron 324 315 298 (K), and V is the windspeed (tn/s). Neutral skin temperature
310 307 298 is roughly 306 K. For windspeeds greater than 1.9 m/s the
Concrete Slab
wind-chill index (TWC)in K is given closely by

15-29
 CHAPTER 15
Table 15-16. Physical thermal parameters of diverse ground types [Lettau, 1954b].

Half-Amplitude Depth Interval

Thermal
(theoretical)
Admittance
Ratio (TAR), Annual Cycle ‘Diurnal Cycle
Ground Type Ground to Air (m) (m)
SOILS
Quartz sand,
medium-fine dry 110 1.0 0.05
870 moisture 230 1.6 0.08
2270 moisture 360 1.5 0.08
Sandy clay, 15% moisture 280 1.3 0.07
Swamp land, 90% moisture 340 1.0 0.05
ROCKS
Basalt 350 1.8 0.09
Sandstone 380 2.2 0.12
Granite 440 2.5 0.13
Concrete 440 2.3 0.12
SNOW, ICE, AND WATER
Feathery snow 10 0.67 0.04
Packed snow 100 1.4 0.07
Still water 280 0.82 0.04
Ice 320 1.4 0.07
Turbulent ocean 103 to 105 61 to 610 3 to 30

T Wc = (306 – H/22). (15.23) not used, or needed, with wind speeds less than 6 km/h
(2 ds).
This formula gives only an approximation because of in- Extreme temperature-wind combinations are frequently
dividual body variations, incoming radiation, and other fac- important in thermal equilibrium design problems, requiring
tors affecting heat loss from the body. The fomula is not estimates of the maximum steady wind speeds likely to be

Table 15-17. Annual and daily temperature cycles. Annual values are
averages for the years 1939 through 1940 at Giessen, Ger-
300 I I 1 I i i 1 [ 1 I I
Im
298 J954
many [Kreutz, 1943]. Daily values are averages of clear
weather, 10 through 12 August 1893, Finland, after Homen
[Geiger, 1957]. z=
293 -
Temperature (K)
290 :
~ 2e8 -

w
K
~ 205 -
~ 283 -
w
Surface 282.1 282.3 283.1 289.6 298.0 297.6 %
f 280 -
1.0 m above 283.8 284.3 284.3 ———
278 -
0.6 m above —— — 284.5 287.0 293.4
Surface 275 –
Amplitude 283.5 283.7 284.4 283.4 290.0 283.2 273 1 I I I I I I 1 : 1 I
ocr Nov oce JAN FED MAR APR MAY JUM JIJL AU@ 22P

Half-Amplitude Depth Interval (m)

Figure 15-16. Variations of soil temperature at indicated depths; North
Depth 1.8 1.6 1.4 0.05 0.08 0.15 Station, Brookhaven, bng Island, Getober 1954 through
September 1955, [afier Singer and Brown, 1956].

15-30
 ATMOSPHERIC TEMPERATURES, DENSITY, AND PRESSURE

I I I 1 II I 15.2 ATMOSPHERIC DENSITY

330
●-RECOMMENDED IN MI L- STO-210B UP TO 90 KM
WR WORLD WIDE =S6N ~lTERIA .
-1
320
The density data discussed in this section are from direct
and indirect observations obtained from balloon-borne in-
310 strumentation for altitudes up to 30 km, and measurements
from rockets and instruments released from rockets for al-
300 titudes between 30 and 90 km.

290
15.2.1 Seasonal and Latitudinal Variations
280
The Reference Atmospheres presented in Chapter 14
provide tables of mean monthly density-height profiles, sur-
270
face to 90 km, for 15° intervals of latitude between the
equator and the North Pole. Densities at altitudes between

/1
260 . ●
10 and 90 km are highest during the months of June and
● July and lowest in December and January at locations north
250 of 30° latitude. In tropical and subtropical areas seasonal
variations are relatively small with highest densities at levels
240 .0 above 30 km occurring in the spring and fall.
/ Mean monthly density profiles, surface to 60 km, ob-
I I I I I 1 I
10 20 30 served during the midseason months at Ascension Island,
8°S, 14°W, Wallops Island, 38”N, 75”W, and Ft. Churchill,
5YN, 94°W, are plotted in Figure 15-18. Densities are
WI NDSPEED (reps)
shown as percent departure from the U.S. Standard At-
Figure 15-17. Extremes of temperature in combination with windspeed. w.rphere, 1976. The individual mean month] y profiles cross
Windspeeds, in general, were observed 12 to 30 m above or converge near 8 km and between 22 and 26 km. Both
the surface. The 35 observations were taken over a 5-year
period at some 22 stations widely scattered in the United are levels of minimum density variability. The level near 8
States. The envelope is for the recommended U.S. design km is considered an isopycnic level because mean monthly
criteria. densities depart from standard by no more than 1% or 2%
regardless of the geographical location or season. Between
22 to 26 km, however, there is a marked seasonal variability,
encountered at various temperatures. Figure 15-17 was pre- even though there is very little longitudinal or latitudinal
pared from 4 years of 6-hourly and 1 year of hourly data variability during individual months. Seasonal differences
for 22 stations in the United States [Sissenwine and Court, in the density profiles at the same three locations are shown
195 1]. It shows maximum steady (5-rein) wind speeds that in Figure 15-19. The minimum seasonal variability of the
occurred with temperatures in the range from 236 K to 319 mean monthly values, 1Yo to 2?70, occurs at 8 km, and the
K during this period. The stations used in this study were maximum seasonal variabilityy occurs above 60 km. The
selected as representative of climatic areas in the United seasonal variations are largest at Ft. Churchill and are small-
States. Mountainous stations were unrepresentative of gen- est at Ascension Island.
erally operational areas and were not among those selected.
Also, the high winds of hurricanes and tornadoes were omit-
ted from the figure. 15.2.2. Day-to-Day Variations
The wind speeds of Figure 15-17 occurred at anemo-
meter heights, usually at 12 to 30 m above ground level The density at a specific altitude may differ from the
during the years of observation. The wind speeds at the 3- seasonal or monthly mean at that altitude due to day-to-day
m level are approximately 2090 less and even 50% less for changes in the weather pattern. The distribution of observed
the extreme low temperature (less than 252 K). densities in January and July at the most climatically extreme
The combination of values of temperature and wind- locations for which data are available near 30°,45°,60° and
speed, recommended for extreme U.S. thermal equilibrium 75”N are shown in Table 15-18a to 15- 18d for altitudes up
design criteria, are shown by the envelope in Figure 15-17. to 80 km. Median, and high and low values that are equaled
This recommendation is not valid in mountainous areas or or more severe 1YO, 10%, and 20% of the time are given
in Death Valley. For the latter the criteria are the same as as percent departures from the U.S. Standard Atmosphere
for world-wide criteria, as plotted in Figure 15-17. at 5-km altitude increments. The 1Yo values for altitudes

15-31
CHAPTER

60
15

\ ‘\
\ i
I I I I I I

APRIL
I

DENSITY
I I I

‘ ‘7

)
\
\
\
52
\ I “j ‘ \
\

/’
1’
/’ ) 1 /’
/’
)

)
44 1

(
44
/’

/\
/ t
1’
! :
i
T
.
: 36 –
,’
I /
n

$“[
ASCENSION ASCENSION — —
g \ WOPS ——
—-— (
WALLCfS
—————
5 \ CHURCHILL
< 28 – ‘1 Lo
i CHURCHILL — 1

4
20 – 20 –
‘!, \
>
12 – 12 – )/
4– 4–
/!\
l~\
~
-30
I I
-20
I I
-10
1
0 0
I I I
20
I I
30
1
PERCENT OEPARTURE FROM 1976 STANMRD PERCENT DEPARTURE FROM 1976 STAN~RD

I 1 1 I [ I r I I 1 I I I I I I I I
I i~ I I
1

60
JULY DENSITY
I OC=R DENSITY

I
52

44 [

I
36 –

ASCENSIW — — ASCENSION — —
WLWPS WALLOPS .-- —---
28 –
CH~CH ILL ~ CHURCHILL —

(\
i
20

12 –
‘\

/’
>
//’.

4–

PERCENT OEPARTURE FROM 1976 STANOARO PERCENT OEPARTURE FROM 1976 STANDARD

Figure 15-18. Latitudinal differences in the density-altitude profiles for the mid-season months at Ascension Island, Wallops Island, and Ft. Churchill.
 a
ATMOSPHERIC TEMPERATURES, DENSITY, AND PRESSURE
I I 1 I I I I I I I I 1 I I I 1 I I 1 1 I 1 i~

ASCENSION OENSITY
I CHURCHILL OENSITY

60 I
1 I
52 –
/
/
44 –
/
F z
. .
JAN —
: 36 –
i
; 36 – AW —————
0
0
3 JULY — — 3 /
h
k OCT —.- —---
< 5 / JAN —
< 28 – 4 28 – APR –––––
/ JMY — —
OCT ---

20 – /
20 –

12 – 12 –

4– 4
[
t 1 I I I I I
I I I I I
0 20 30
-30 -20 -10

PERCENT DEPARTURE FROM 1976 STANDARD PERCENT DEPARTURE FROM 1976 STAN~RO

above 30 km are considered rough estimates as they are

based on the tails of the distributions of,observed values
r“’’’’’’’” WOPS OENSITY plotted”on probability paper. Estimates above 60 km are
less reliabIe than those at lower levels because of the paucity
of data and larger observational errors at the higher altitudes.

52

1 In tropical regions the monthly density distributions are

nearly normal for altitudes up to 50 km. Consequently,
reasonable estimates of the distributions of density in the
tropics can be obtained from monthly means and standard
deviations. Stand&d deviations of the observed densities
around the mean monthly values at Ascension, given in
Table 15-19, are typical of the day-to-day variations found
in the tropics [Cole and Kantor, 1980].

15.2.3 Spatial Variation

The rate of decay of the correlation coefficient between

densities at two points with increasing horizontal separation
is directly related to the scale of the major features of the
weather patterns that are experienced at a specific latitude
4L -30 -20
and altitude. Figure 15-20 provides information on the decay
of density correlations with distance near 6&N for altitudes
PERCENT DEPARTURE FROM 1976 STANDARD up to 60 km. The decay in density correlations below 20
km are based on an inte~retation of data from studies of
the spatial correlations of pressure, temperature, density and
Figure 15-19 Seasonal differences in the density-altitude profiles at As-
cension Island. Wallops Island, and Ft. Churchill. wind at radiosonde levels at locations between 30° and 70°N

15-33
CHAPTER 15
Table 15- 18a. Median, high, and low values of densities given as percentage departure from U,S, Standard Atmosphere 1976 for January and July at
30”N.

Median 1% 10% 20% U.S. Std

Altitude (% of Density
High
(km) Std) High Low High Low Low ‘ (kg/rn’)

January

5 –1 +1 –3 o –2 o –2 7.3643 – 1
10 +1 +4 –3 +3 –1 +2 o 4.1351
15 +7 + 15 –1 + 12 +4 + 10 +5 1.9476
20 +3 +7 –2 +5 +1 +4 +2 8.8910 – 2
25 –2 +4 –6 +3 –4 +1 –2 4.0084
30 –4 +2 – 10 –2 –8 –3 –6 1.8410
35 –3 +3 – 12 0 –8 –1 –6 8.4634 – 3
40 –1 +2 – 10 +1 –7 o –5 3.9957
45 0 +8 – 10 +3 –7 +2 –5 1.9663
50 +1 + 12 –8 +7 –4 +5 –2 1.0269
55 0 +9 – 10 +5 –6 +3 –4 5.6810 – 4
60 –2 + 12 –15 +5 –9 +2 –6 3.0968
65 –4 +21 – 25 + 13 –13 +7 –6 1.6321
70 –5 + 16 – 26 +9 – 17 +6 – 12 8.2828 – 5
75 –7 +21 – 25 + 13 – 15 +8 – 10 3.9921
80 –4 +21 – 22 + 15 –13 +8 –7 1.8458

July

5 –3 o –5 –1 –4 –2 –4 7.3643 – 1
10 +1 +3 –1 +2 o +2 o 4.1351
15 + 16 i-20 +11 + 17 + 13 + 17 + 14 1.9476
20 +8 +11 + 14 + 10 +5 +9 +6 8.8910 – 2
25 +4 +9 o +7 +2 +6 +3 4.0084
30 +3 +7 –1 +5 +1 +4 +2 1.8410
35 +6 + 10 +2 +8 +3 +7 +4 8.4634 – 3
40 +9 + 15 +2 +11 +5 + 10 +7 3.9957
45 + 12 + 19 +4 + 14 +7 + 13 +9 1.9663
50 + 13 +23 +6 + 17 +8 + 15 + 10 1.0269
55 +11 +20 +2 + 15 +5 + 13 +7 5.6810 – 4
60 + 13 + 14 –1 +21 +3 + 19 +7 3.0968
65 + 15 +43 –6 +38 o +30 +6 1.6321
70 + 15 +32 –9 +23 +1 +20 +8 8.2828 – 5
75 + 10 +24 –11 +20 –6 + 15 +1 3.9921
80 +6 +22 – 15 + 17 –6 + 14 +1 1.8458

latitude IBertoni and Lund, 1964]. Information on the spatial 50 km, zero correlation is attained at more than twice that
correlations at altitudes above 20 km is from a study by distance, or 4450 km. This analysis indicates the presence
Cole [1979]. In that paper, data from constant pressure maps of disturbances with wavelengths of roughly 18500 km at
for 5.0, 2.0 and 0.4 mb levels were used together with 50 km, close to planetary wavelength number one at 60”N.
nearly simultaneous rocket observations at several pairs of Information from Kantor and Cole [ 1979] on the correlations
stations near 60”N to determine the rates of decay of density between densities at points up to 370 km apart in tropical
correlation at levels between 30 and 55 km. As Figure 15- regions is provided in Table 15-20, for levels between 10
20 indicates, the rate of decay in density correlation with and 60 km.
distance decreases substantially with altitude. At 10 km, for The rms difference between the densities at two points
example, zero correlation is attained at about 2000 km at can be estimated by

15-34
 a
ATMOSPHERIC TEMPERATURES, DENSITY, AND PRESSURE
Table 15- 18b. Median, high, and low values of densities given as percentage departurefrom U.S. StandardAtmosphere1976 for January and July ~t
45”N.
Median 1% 1070 20% U.S. Std
Altitude (% of Density
High
(km) Std) High Low High Low Low ‘ (kg/m’)

January

5 0 +4 –3 +3 –2 +2 –1 7.3643 – 1
10 –2 +6 – 10 +3 –6 +1 –4 4.1351
15 –3 +4 – 12 +1 –8 –1 –6 1.9476
20 –2 +2 –8 o –6 –1 –5 8.8910 – 2
25 –2 +2 –8 0 –6 –1 –5 4.0084.
30 –5 +1 – 17 –2 – 13 –4 –9 1.8410
35 –6 +2 – 20 –2 – 16 –4 – 12 8.4634 – 3
40 –8 +5 – 23 o – 17 –4 – 13 3.9957
45 –9 +8 – 22 +2 – 16 –3 – 14 1.9663
50 +8 +11 – 20 +4 – 16 –3 – 14 1.0269
55 –9 +9 – 25 +2 – 18 –4 – 16 5.6810 – 4
60 – 12 +7 – 28 o – 23 –7 – 20 3.0968
65 – 14 o – 38 –5 – 34 – 10 – 28 1.6321
70 – 15 +2 – 38 –9 – 30 – 12 – 26 8.2828 – 5
75 – 16 –3 – 38 –9 – 30 –12 – 26 3.9921
80 – 23 –2 – 42 –8 – 36 – 10 – 30 1.8458

July

5 –2 +1 –5 –1 –4 –1 –3 7.3643 – 1
10 o +3 –4 +2 –2 +1 –1 4.1351
15 +8 + 17 +2 + 15 +4 + 13 +5 1.9476
20 +6 +11 o +8 +2 +7 +3 8.8910 – 2
25 +7 + 10 +4 +9 +5 +8 +6 4.0084
30 +7 + 12 o +9 +2 +8 +4 1.8410
35 +9 + 16 0 + 12 +3 + 10 +6 8.4634 – 3
40 + 13 +21 +4 + 16 +3 + 14 + 10 3.9957
45 + 15 +26 +6 +20 + 10 + 18 + 12 1.9663
50 + 17 +31 +9 +25 + 12 +21 + 14 1.0269
55 + 17 +32 +8 +25 +11 +22 + 14 5.6810 – 4
60 + 19 +30 +4 +26 + 10 +24 +13 3.0968
65 +20 +40 +4 +35 + 10 +30 +13 1.6321
70 +20 +37 o +32 +9 +27 + 12 8.2828 – 5
75 + 19 +40 –2 +30 +7 +26 +11 3.9921
80 + 14 +32 –4 +30 +4 +25 +9 1.8458

u xv = Vat + U? — 2rXYu. aY, (15.24) observed simultaneously at locations 90, 180 and 360 km
apart in the tropics are presented in Table 15-21 for altitudes
between 10 and 60 km. For a given month, the rms differ-
where u,, is the estimated rrns difference between densities ences provided in Table 15-21 can be considered to represent
at points- x and y, m~and @are the variances of density variability around the mean monthly density gradients, which
around the monthly mean values, and and rXYis the corre- are given in Table 15-22 [Cole and Kantor, 1975] for the
Iation coefficient between the densities at points x and y. indicated latitudinal differences. Longitudinal difference re-
For short distances (up to 550 km) u; and a; can usually main near zero in tropical areas. Information on the spatial
be assumed to be equal. variability of density is useful in determining how accurately
The estimated rms difference between densities that are a density observation taken 75 to 500 km from the point of

15-35
CHAPTER 15
Table 15-18c. Median, high, and low values of densities given as percentage departure from U.S. Standard Atmosphere 1976 for January and July at
WN.

Median 1% 10% 20% U.S. Std

Altitude (% of Density
High
(km) Std) High bw High Low Low ‘ (kg/m’)
January

5 +1 +6 –3 +4 –1 +2 o 7.3643 – 1
10 –6 +3 – 15 +2 – 15 –3 – 10 4.1351
15 –9 –2 – 15 –5 – 12 –6 –11 1.9476
20 –8 –1 – 15 –5 –11 –6 – 10 8.8910 – 2
25 –7 +3 – 16 –2 – 12 –4 – 10 4.0084
30 “ – 10 +7 – 32 +2 – 18 –2 – 15 1.8410
35 – 12 +8 – 35 –3 – 27 –3 – 19 8.4634 – 3
40 – 15 + 10 – 36 +5 – 30 –4 – 20 3.9957
45 –21 + 12 – 39 +5 – 34 – 10 – 24 1.9663
50 – 26 + 14 – 43 +3 – 36 – 15 – 29 1.0269
55 – 32 +9 – 48 – 10 – 39 – 20 – 35 5.6810 – 4
60 – 36 +4 – 54 –12 –40 – 25 – 39 3.0968
65 – 36 –5 – 50 – 16 –46 – 27 – 42 1.6321
70 – 37 – 12 – 54 – 25 – 49 – 32 – 43 8.2828 – 5
75 – 35 – 10 – 53 – 24 – 47 – 30 – 42 3.9921
80 – 28 –11 – 53 – 17 – 47 –21 –40 1.8458

July

5 –2 +2 –5 +1 –4 o –3 7.3643” – 1
10 0 +7 –8 +4 –5 +2 –3 4.1351
15 0 +6 –7 +3 –4 +2 –2 1.9476
20 +3 +7 –2 +6 o +5 +1 8.8910 – 2
25 +5 +8 +1 +7 +2 +6 +3 4.0084
30 +7 + 12 –1 +9 +2 +8 +4 1.8410
35 + 10 + 18 0 + 14 +3 + 12 +7 8.4634 – 3
40 + 15 +23 +5 + 19 + 10 + 17 +12 3.9957
45 +20 +28 +7 +25 + 13 +23 +16 1.9663
50 +25 +35 + 10 +30 + 16 +28 +22 1.0269
55 +27 +35 +11 +30 + 16 +29 +22 5.6810 – 4
60 +28 +42 +11 +39 +16 +33 +22 3.0968
65 +35 +50 +11 +44 +18 +39 +28 1.6321
70 +42 +52 + 12 +46 +20 +44 +30 8.2828 – 5
75 +44 +58 + 12 +52 +20 +48 +35 3.9921
80 +40 +56 + 10 +50 + 18 +44 +30 1.8458

vehicle reentry represents the conditions encountered in the surface to 60 km, together with interlevel correlation coef-
reentry corridor. ficients between levels have been developed for tropical,
temperate and arctic regions [Cole and Kantor, 1980]. Ta-
15.2.4 Statistical Applications to bles 15-23a to 15-23f contain statistical arrays of density
Reentry Problems for the months of January and July at Kwajalein (9”N),
Wallops Island (38”N), and Ft. Churchill (59°N).
The relatively large number of available radiosondes and Variations in the range or deceleration of free falling
meteorological rocket observations permit a detailed anal- objects or ballistic missiles that arise from day-today changes
ysis of the characteristics of atmosphere density profiles at in atmospheric density can be estimated from Tables 15-
altitudes below 60 km. Arrays of means and standard de- 23a to 15-23f. The integrated effect, E, of mean monthly
viations of density at 2-km intervals of altitude from the density on the trajectory or impact point of a missile can

15-36
 a
ATMOSPHERIC TEMPERATURES, DENSITY, AND PRESSURE
Table 15- 18d. Median, high, and low values of densities given as percentage deuarture from U.S, Standard Atmosphere 1976 for January and July at
75”N. -

Median 1% 10% 20% U.S. Std

Altitude (% of Density
High
(km) Std) High Low High Low Low (kg/m’)

January

5 +2 +6 –1 +5 o +4 +1 7.3643 – 1
10 –8 +2 – 18 –3 – 13 –5 – 10 4.1351
15 – 10 –1 –18 –6 – 14 –8 –13 1.9476
20 – 12 –1 – 22 –6 – 17 –8 – 15 8.8910 – 2
25 – 15 –2 – 28 –8 – 20 – 10 –18 4.0084
30 –21 –4 – 36 –9 – 26 – 16 – 24 1.8410
35 – 25 0 – 43 – 10 – 32 – 16 – 30 8.4634 – 3
40 – 29 +4 – 48 –9 – 38 – 16 – 38 3.9957
45 – 33 +8 – 52 –6 – 45 – 16 – 39 1.9663
50 – 38 +4 – 56 –8 – 48 – 20 – 42 1.0269
55 – 44 +5 – 65 – 10 – 56 – 23 – 50 5.6810 – 4
60 – 46 0 – 70 – 16 – 60 – 32 – 55 3.0968
65 – 47 +1 – 66 – 27 – 62 – 35 – 58 1.6321
70 – 48 –1 – 69 –21 – 62 – 35 –60 8.2828 – 5
75 – 45 – 10 – 65 +25 – 57 – 35 – 53 3.9921
80 – 40 –8 – 55 – 24 – 50 – 34 – 45 1.8458

July

5 1 +4 –2 +3 –1 +2 o 7.3643 – 1
10 –4 +5 – 12 +3 – 10 o –7 4.1351
15 –4 +2 –9 o –7 –2 –6 1.9476
20 +1 +6 –4 +4 –2 +3 –1 8.8910 – 2
25 +1 + 10 –8 +6 –3 +5 –2 4.0084
30 +7 + 13 +2 + 10 +5 +8 +6 1.8410
35 + 12 +25 +3 + 18 +8 + 16 + 10 8.4634 – 3
40 + 19 +27 +6 +23 + 13 +21 + 16 3.9957
45 +25 +35 +10 +30 + 18 +28 +21 1.9663
’50 +27 +40 + 10 +35 +20 +32 +24 1.0269
55 +32 +42 + 10 +39 +20 +35 +25 5.6810 – 4
60 +37 3.0968
65 +48 (Insufficient data above 55 km 1.6321
70 +60 in summer) 8.2828 – 5
75 +67 3.9921
80 +64 1.8458

be determined for a specific location by computer “flights” due to day-to-day variations from the mean seasonal or the
through mean monthly or seasonal density profiles if the mean monthly density profile can be obtained from
proper influence coefficients, Ci, for the missle at various
levels are given. For example, we can write

where u ?“, is the integrated variance for all layers being

considered, C, and Cl are influence coefficients at the ith
where pi is the mean monthly density at the ith level. The and jth levels, U, and al are the standard deviations of densit y
influence coefficients depend upon aerodynamic character- at the two levels, and r,J is the come! ation coefficient between
istics, reentry angle, and the speed of the vehicle. The densities at the two levels. In these computations density is
integrated standard deviation in range or deceleration Uint, assumed to have a Gaussian distribution at all levels. As a

15-37
CHAPTER 15
Table 15-19. Standard deviations (%) of observed daY-to-~dY variations Table 15-20. Correlation coefficients between densities at points up to
in density around tbe monthly mean at Ascension Island 370 km apart in [be tropics.
(8°s).
Altitude Correlation Coefficient
S.D. of Density (% of Monthly Mean) (km)
Altitude
Apr July Ott 90 km 180 km’ 370 km
(km) Jan

0.3 0.3 0.4 10 0.97 0.95 0.90

5 0.4
0.4 0.4 0.4 20 0.98 0.97 0.92
10 0.4
0.7 0.8 0.7 30 0.98 0.97 0.92
15 0.8
1.3 1.8 1.3 40 0.98 0,97 0.92
20 1.5
1.3 1.2 1.3 50 0.98 0.97 0.92
25 1.3
1.2 1.4 1.2 60 0.98 0.97 0.92
30 1.2
35 1.8 1.8 1.4 1.2
40 2.3 2.1 1.8 1.8
45 2.3 2.3 2.6 2.3
50 2.7 2.5 2.6 2.7 The decrease in the number of available observations above
60 km and the larger random observational errors at the
higher altitudes make it difficult to obtain reliable estimates
of the magnitude of the diurnal variations at altitudes be-
result, the error in the CEP (the circle within which 5070 tween 60 and 90 km.
of the events are expected to occur) will be generally less The 50-km densities from a series of soundings taken
than 10%. at Ascension during a 48-h period in April 1966 [Cole and
Kantor, 1975] are plotted versus local time in Figure 15-
21. Densities are given as percent departure from those for
15.2.5 Variability with Time the 1976 U.S. Standard Atmosphere. The crosses represent
averages of observations taken within two hours of each
Studies based on radiosonde observations have shown other. Harmonic analysis of the eight average values pro-
that there are no significant diurnal variations in density at duced the solid curve when the first and second harmonics
altitudes up to 30 km. The analysis of meteorological rocket for the 48-h period were added together. An F-test indicates
observations, however, indicates the presence of a signifi- that the second harmonic, which represents the diurnal os-
cant diurnal oscillation in density at altitudes between 35 cillation in density, has an amplitude of slightly less than
and 60 km. The phases and amplitudes of the diurnal os- 4% (a range of almost 8%) and is significant at the 1% level;
cillation at these altitudes are best defined in the tropics. it reduces the observed variance by 91%. Maximums occur
at 1600 and minimums near 0400 local time. From this
analysis it is apparent that the diurnal oscillation is the
I I 1 1 1 I dominant short-period fluctuation at 50 km.
60 CORRELATION The rms differences between density observations taken
from 1 to 36 hours apart also provide a measure of the rate
of change in density with time at a given altitude. Computed
rms values from the Ascension series mentioned above are
shown as a function of time in Figure 15-22 for altitudes
from 35 to 60 km. The number of pairs of observations
available for each time interval is also shown. Since at time
T = O the rms change in space is zero, an estimate of the
random observational error can be obtained from the ob-
servations themselves by extrapolating curves in Figure 15-
22 back to zero hours. This procedure indicates that the
random rms errors are approximately 1% at 35 and 40 km,
and 1.5% to 2.0% at altitudes between 45 and 60 km.
If there are no well-defined periodic oscillations within
a 24-h period, the rms variability would be expected to
increase smoothly with time until it reached a value rep-
DISTANCE (km) resenting the climatic or the day-to-day variability around
the monthly mean. However, a well-defined 24-h oscillation
Figure 15-20. Decay of density correlations with distance at various al-
titudes in midlatitudes. can be seen (Figure 15-22) in the rms density variations at

15-38
 A’i’MOSPHERIC
TEMPERATURES, DENSITY,AND PRESSURE
Table 15-21 Estimated rms differences (% of mean) between densities at Ioeations 90, 180, and 360 km apatt during the midseason months in the tropics.

January April July October

Altitude 90 180 360 90 180 360 90 180 360 90 180 360
(km) km km km km

10 0.10 0.13 0.18 0.10 0.13 0.18 0.]0 0.13 0.18 0.10 0.13 0.18
15 0.13 0.17 0.25 0.11 0.14 0.21 0.16 0.20 0.30 o.i6 0.20 0.30
18 0.50 0.61 1.00 0.34 0.42 0.68 0.30 0.37 0.60 0.34 0.42 0.68
20 0.28 0.34 0.:6 0.28 0.34 0.56 0.24 0.29 0.48 0.24 0.29 0,48
25 0.28 0.34 0.56 0.28 0.34 0.56 0.24 0.29 0.48 0.26 0.32 0.52
30 0.30 0.37 O.io 0.30 0,37 0.60 0.28 0.34 0,56 0.30 0.37 0.60
35 0.34 0.42 0.68 0.30 0.37 0.60 d.30 0.37 0.60 0.36 0.4 0.72
40 o.4a 0.49 0.80 0.44 0.54 0.88 0.48 0.59 0.96 0.44 0.54 0.88
45 0.46 ;.56 0.92 0.40 0.49 0.80 b.60 0.73 i.20 0.52 O.M 1.04
50 0.56 0.69 1,12 0.54 0.66 1.08 0.72 0.88 1.44 0.54 0.66 1.08
55 0.66 0.81 1.32 0.56 0.69 1.12 0.84 1.03 1.68 0.78 0.96 1.56
60 0.84 1.03 1.68 0.66 0.81 1.32 l.ti 1.22 2.00 0.82 I .00 1.64

all altitudes between 40 arid 60 km with maximums at 12 Unlike the tropics, a 24-h oscillation in, density is not ap-
and 36 hours and a minimum at 24 hours. An analyses of p~ent from this analysis which is based on eight years of
meteorological rocket observations t&en at Kwajalein (9°N) data at Ft. Churchill and ten years at Waljops ~slqnd. The
and Ft. Sherman (9°N) show similar results [Kantor and diurnal oscillation is relatively small and is ~robably masked
Cole, 1981]. The diagram in Figure 15-22 and the results bj instrumentation errors and changing synoptic patterns.
of similar studies show that in the tropics an observation The tis variability at boti locations increases with time
24 hours old is more representative of actual conditions than until the climatic values of day-to-day v~ations around the
one 12 hours old. monthly means me reached.
The observed ms variati~ns of density with time lags
of 1, 2, 4 and 6 hours are shown in Figure. 15-23 for levels
between 60 and 90 km at Kwajalein. This information, from i5.3ATMOSPHERIC PRESSURE
Cole et al. [1979], is based on a July 1978 series of high- UP TO 90KM
altitude ROBIN falling sphere flights at Kwajalein. The first
profile represents the estimated rms obse~atioiial emor. Pressure data provided in this section dk based on (1)
The rms variations of density with time at the 50-km routine radiosoride observations taken by national weather
level are shown in Fi@re 15-24 for Wallops Island (38”N) services and extending to approximately 30 km, and (2)
and Ft. Churchill (59°N) for the months of January and July. measurements from rockets and instmments released from

Table 15-22. Mean monthly latitudinal density gradients (% change per 180 km) in the tropics.

Altitude January April July October

(km) Gradient (%) Gradien~ (%j Gradient (%) Gradient (%)

10 0.01 0.02 0.03 0.04

15 0,15 0.17 0.08 0.05
20 0.12 0.23 0.08 0.06
25 0.04 0.14 0.10 o.i4
30 Q.26 0.13 0.14 0.21
35 0.13 0.22 0.16 0.23
40 0.03 0.16 0.16 O.qo
45 0.14 O.Q1 0.17 0.21
50 0.11 0.09 0.12 0,20
55 0.08 0.12 0.04 0.27
60 0.09 0.04 0.10 0.25

15-39
CHA~ER 15
 a

ATMOSPHERIC TEMP~RATVBEs, DENSITY) A~~ PRESSURE

-G f-..n
~leel
CHA~ER 15

15-42
 ATMOSPHERIC TEMPERATURES, DENSITY, AND PRESSURE

.
r
c
,
CHAPTER 15

z
<>

15-44
ATMOSPHERIC TEMPERATURE, DENSITY, AND PRESSURE

15-45
.
CHAPTER 15

8
*1—
n I 1 I I I I I 1 1
& 16 ASCENSION 50 km
i
●

LOCAL TIME

Flgum 15-21. Diurnal density (50 km) variation at Ascension Island. (Dots indicate observed values, x’s represent 3-h averages, and the solid line depicts
the computed diurnal cycle.)

rockets at altitudes between 25 and 90 km. Both data sources 15.3.1 Sea-Level Pressure
are supplemented with pressures derived from measure-
ments made from earth-orbiting satellites. Although atmos- The variations of sea-level pressure normally have little
pheric pressure above radiosonde altitudes is occasionally effect on the operation of surface equipment. However, in
measured directly, it normally is calculated hydrostatically the design of sealed containers that could possibly explode
(as discussed in Chapters 14) from observed temperatures or collapse with pressure changes, the range of surface pres-
or densities for altitudes above 30 km. These data are in- sures likely to be encountered should be considered. Surface
tended for use in design problems involving variations in pressures vary with the height of the station above sea level
the heights of constant pressure surfaces and/or changes in as well as with changing weather patterns. Standard at-
pressure at specific altitudes. mospheric pressure at sea level is 1013.25 mb, but there

I I I I I I I I I

ASCENSION

6 -
50 km

11111111 ,1,,111,,1
16

2
21

6
18

12 18
12

24
9

30
6 ~;;

36.
TIME LAG (HOURS)

Figure 15-22. Root mean square (rms) lag variability of density with time at Ascension Island.

15-46
 H

ATMOSPHERIC TEMPERATURES, DENSITY, AND PRESSURE

DENSITY Table 15-24. Sea-1evel pressures exceeded 997c and I% of the time in
January.

I HR/A /’x /)6HR

2 HR /’4HR
/ / ,/
// +/ ~
Aleutian low 965
Icelandic low 953
Exceeded 0/0of time
Siberian high 1057
Pacific high 1038
Canadian high 1052

in areas under the influence of anticyclones. In the Northern

Hemisphere extreme values, exclud;ng tropical cyclones and
tornadoes, are most likely to occur in these regions during
A 1 1 1

z
a
+ 6 8 10 January. Table 15-25 lists for comparison actual worldwide
pressure extremes, including those resulting from storms of
RMS VARIATION (PERCENT OF MEAN)
tropical origin. Examples of mean sea-level pressures and
Rgure 15-23. The rms variation in density for time lags of I to 6 h at typical fluctuations are given in Table 15-26 which contains
Kwajalein.
mean sea-level pressures for the four midseason months and
the standard deviations of daily values around these means
are sizable variations from this value with both time and for 16 specific locations in the Northern Hemisphere.
location.
Table 15-24 indicates extreme sea-level pressures that
may be encountered in the Northern Hemisphere. During 15.3.2 Seasonal and Latitudinal Variations
the month of January, pressures exceeded 9970 of the time
are given for areas under the influence of semipermanent The Reference Atmospheres presented in Chapter 14
cyclones, and pressures exceeded 1~0 of the time are given provide tables of mean monthly pressure-height profiles,

RIIIS DIFFERENCES BETWEEN DENsITIEs OBSERmtI I TO 72 HOURS APART; x 25 WIRS, os 5 WIRS

I Y. 0.015X + 1.900 CHURtilLL 1 POKER FLATS JULY

6 -

0
4 -

0
2 .
0

t Y. 0.069x+ 2.3S9 CHURCHILL / POKER FLATS ~ JANUARY

6 - xx
x
t-
Z
4 -
u ox o
0 2 = 0
a
g 0

Y.o.010x+2.931 WALLOPS JULY

0
e -

6 - 0
x o
6 - xxx
x x
o 0
2 Xx.
0 x

V. O.017X+4.144 WALLOPS JANuARY

0 o
6 -0 x
x x
x
x 0
x
4 0

z o
I I 1 1 1 1 1 1
0 1 I 1 !

6 12 la 24 30 36 42 48 54 60 66 72

TIME LAG (HOURS)

Figure 15-24, I+ms differences between densities observed 1 to 72 h apart at 50 km (x ~ 5 pairs, o ~ 5 pairs).

15-47
CHAPTER 15
Table 15-25. Worldwide bressure extremes.

Pressure
(rob) kation Date

LOW
World 870* 17”N, 138”E, T~hoon Tip [Wagner, 1980] 12 Ott 1979
876* 13”N, 141”E, Typhoon June [Holliday, 1976] 19 Nov 1975
877* 19”N, 135”E. Typhoon Ida [Riordars, 1974] 24 Sep 1958
877* 15”N, 128°E, Tfihoon Nora [Holliday, 1975] 6 Ott 1973
No. America 892.3 Mataumbe Key, Horida, hurricane [Riordan, 1974] 2 Sep 1935
HIGH
World 1083.8 Agata, Siberia [Riordan, 1974] 31 Dec 1968
1075.2 Irkutsk, Siberia, [Valley, 1965] 14 Jan 1893
No. America 1067.3 Medicine bat, Al&tia [Riordan, 1974] 24 Jan 1897
*Dropsonde measurements

surface to 90 km, for 15° intervals of latitude from the ences are small and do not display a systematic seasonal
equator to the north pole. These atmospheric models de- pattern. Departures from standard generally increase with
scribe both seasonal and latitudinal variation of pressure. latitude. Thus, largest seasonal differences are shown at
Figure 15-25 contain the vertical pressure profiles for the 60”N where mean monthly pressures at 60 to 70 km are
midseason months at each of four latitudes: 1Y, 30°, 45° nearly 40% greater than standard in July ~ 30% to 36~o
and 6@N. The profiles are displayed as percent depar- less than standa~d in January. Consequently, July values
tures from the U.S. Standard Atmosphere, 1976. Pres- are roughly twice those in January between 60 and 70
sures at altitudes between 2 km and 70 or 80 km are km. Pressures at these levels at 75”N (not shown) are
highest in summer and lowest in winter over regions roughly 10% lower than jhese values in winter and 15%
poleward of 30°N. In tropical latitudes, seasonal differ- higher in summer.

Table 15-26. Mean monthly sea-level pressures and standard deviations of daily values.

Location January April July October

Mean S.D. Mean S.D. Mean S.D. Mean S.D.

Latitude Longitude (rob) (rob) (rob) (rob) (rob) (rob) (rob) (rob)

10”N 70”W 1013 2 1012 1 1012 1 1011 1

20”N 70”W 1018 2 1017 2 1018 2 1013 2
30”N 70°w 1022 6 1019 5 1021 3 1018 4
4Q”N 70”W 1018 10 1017 9 1016 5 1018 8
50°N 70”W 1016 12 1014 10 1011 6 1013 11
60°N 70°w 1008 11 1014 10 1008 7 1008 10
70°N 70”W 1004 11 1014 10 1009 6 1006 10
80”N 70”W 1o11 11 1020 9 1o11 6 1013 8
10”N 20”E 1012 4 1008 3 1009 2 1009 2
20”N 20°E 1017 4 1o11 3 1008 2 1012 2
30”N 20”E 1019 5 1014 4 1012 3 1015 3
40°N 20”E 1018 9 1013 6 1013 3 1016 5
50”N 20°E 1020 12 1013 7 1013 5 1017 8
60”N 20”E 1015 16 1012 10 1o11 7 1011 11
70”N 20”E 1004 15 1010 10 1012 7 1005 11
80°N 20”E 1008 15 1016 10 1014 7 1010 10

15-48
 a
ATMOSPHERIC TEMPERATURES, DENSITY, AND PRESSURE
60” N 45” N 30” N 15” N

.// ;
“/
./ ;’

/’i ,’

‘1 \
1;/!

II’

\;! \
E JAN OT APR JUL JAN WT APR JuL
. f,
\
- 50 -,
u \
n. \
g \

~ 40 \’;
a \
\ 1;
\
30
\ Ii
\
I \;
20 II
Ij
I /:
10 \ II
\ ,.

\i
( \

-20 0 +20

PERCENT DEPARTURE FROM US STANDARD ATMOSPHERE

Figure 15-25. Pressure-altitude profiles for midseason months.

Table 15-27. Average height and standard deviation at standard pressure levels over North America, 90 to 100”W
1
Average Height and Standard Deviation

20”N 30°N 40°N 50”N 60°N 70°N 80”N

Pressure Mean S.D. Mean S.D. Mean S.D. Mean S.D. Mean S.D. Mean S.D. Mean S.D.
(rob) (km) (m) (km) (m) (km) (m) (km) (m) (km) (m) (km) (m) (km) (m)

January

700 3.165 30 3.115 55 3.015 85 2.865 100 2.770 100 2.710 90 2.690 75
500 5.845 40 5.745 85 5.565 125 5.340 145 5.180 I 50 5.075 I40 5.055 125
300 9.595 55 9.425 125 9.150 175 8.825 195 8.585 230 8.425 210 8,380 180
200 12.280 70 12,090 130 11.765 165 11.430 175 11.180 190 10,995 200 10.920 I95
100 16.455 55 16.325 95 16.110 110 15.890 145 15.655 175 15.400 195 15.195 I90
50 20,540 130 20.500 200 20.415 215 20.280 215 20.075 200 19.775 180 19.440 180
25 24.900 210 24.865 245 24.790 335 24.555 275 24.380 245 23.905 245 23.425 230
15 28.100 245 28.050 335 28.000 365 27.750 365 27.650 350 27,000 335 26.350 305
10 30.600 250 30.550 380 30.500 400 30.250 380 30.150 380 29.500 320 28.750 260

July

700 3.185 15 3.190 20 3.170 35 3.080 55 3.005 60 2.975 65 2.930 65

500 5.890 20 5.910 25 5.875 50 5.720 85 5,600 90 5.540 90 5.465 95
300 9,675 30 9.705 35 9.630 80 9.405 130 9.215 125 9.125 125 9.010 I30
200 12.395 40 12.430 50 12.345 100 12.080 135 11.870 I 35 11.790 135 11.710 135
100 16.570 45 16,625 50 16.605 65 16.515 80 16.455 85 16.420 80 16.390 75
50 20.765 75 20.865 90 20.940 105 20.975 105 21.005 105 21.045 90 21.050 75
25 25,180 150 25.330 150 25.440 175 25.530 175 25.625 165 25.715 I50 25.765 I50
15 28.300 I70 28.450 170 28.650 175 28.800 175 28.900 175 29.100 I70 29.200 165
10 30.800 190 30.950 190 31.200 190 31.400 I90 31.550 190 31.750 185 31.950 180

15-49
CHAPTER 15
Table 15-28. Departures from mean monthly pressures (%) exceeded less than 5% of the time in January and July. Values below 30 kti are bdsed on
radiosonde observations. Those values above 30 km are ba$ed on roeketsonde observations.

Height January July

(km) 75”N 60°N 45”N 3&N 15°N 15“N 3@N 45”N M“N 75°N

o ?2.5 *3 ?2.5 tl ?0.4 to.4 fo.5 *1 *1 *1.5

10 7 4 3 2 0.8 0.7 0.8 2 3 4
20 10 10 10 7 4 2 2 3 3 3
30 20 16 14 12 7 4 4 4 5 5

40 25 20 15 8 7 8 8 10
50 30 25 18 10 10 12 13 14
60 35 30 20 12 12 14 16 18
70 30 25 18 10 10 12 15 16
80 20 16 12 8 8 9 10 10

Table 15-29. Amplitudes of systematic pressuti variations and time of

maximum at Tereeira, Azores [Harris et al, 1%2].

15.3.3 Day-to-Day Variations Diurnal Semidiumal

Changing synoptic situations, which incltide movements Pressure Level Ampl. Time Ampl. Time
of high and low pressure centers and their associated ridges (rob) (m)* (rob) (h) (rob) (h)
and troughs, and variations in the energy absorbed directly
by the atmosphere, cause day-to-day changes in the height Sfc d 0.10 2100 0.50 0948
of constant pressure surfaces. Information on the magnitude 1000 122 0.10 1904 0.53 W50
950 570 0.12 1824 0.46 0956
of day-to-day variations in the heights of si.tchsurfaces are
900 1033 0.16 1612 0.49 1002
provided in this section. Detailed information for specific
850 1454 0.18 1604 0.47 I002
levels and locations should be requested if conditions appear
800 2027 0.20 1612 0.44 1002
critical.
750 25G9 0.20 1616 0.38 1010
Table 15-27 lists monthly mean heights of pressure sur-
700 3127 0.25 1548 0.37 1002
faces in January and July and their standard deviations for
650 3731 0.18 1608 0.40 1030
middle North America. These data indicate the variation in
600 4365 0.25 1608 0,33 1020
the mean heights of constant pressure surfaces between 7~
550 5051 0.27 1508 0.33 1034
and 10 mb with latitude and season, and the estimated dis-
500 5782 0.28 156 0.29 1032
tributions of day-to-day variability around the monthly mearis.
450 6587 0.27 1424 0.24 1036
Estimated departures from mean monthly pressures, which
400 7449 0.31 1504 0.24 1046
are equaled or exceeded less than 590 of the time between
350 8409 0.31 1504 0.20 1046
15° and 75°N, are shown in Table 15-28 as percentages of
3m 9482 0.32 1444 0.18 1108
the mean January and July values, surface to 80 km.
250 10708 0.33 1420 0.16 1102
As can be seen in Table 15-28, day-to-day variability
2W 12149 0.32 1408 0.14 1110
generally increases with latitude and altitude in both January
175 12991 0.32 1352 0.13 1120
and July, although to a much smaller extent in July. The
I50 13948 0.30 1348 0.11 1100
estimated 5% extremes are largest at 60 to 65 km at all
125 15066 0.28 1328 0.11 1124
latitudes, reaching t 35% during 60°N winter. Variability
100 16423 0.26 1304 0.09 1128
appears to decrease above 65 km to a probable minimum
80 17776 0.24 1300 0.09 1120
value near 85 km. The estimated departures shown in Table
60 19547 0.23 1256 0.07 1124
15-28 include some diurnal and semidiumal fluctuations due
50 20668 0.21 1256 0.07 1114
to solar influences, particularly since the basic pressure data
40 22077 0.20 1244 0.06 1116
were not observed at the same time every day. Envelopes
30 24012 0.18 1256 0.05 1110
of these estimated 95% values should not be used as profiles
20 2G673 0.16 1256 0.04 1128
since such pressures would not necessarily be found at all
15 28005 0.15 1252 0.03 1136
altitudes at any one given time and/or location. Decreases
10 30507 0.12 1304 0.01 1204
in atmospheric pressure in one layer, for example, normally
are associated with increases in another layer. Estimated mean annual height

15-50
❑ a
ATMOSPHERIC TEMPERATURES, DENSITY, AND PRESSURE
15.3.4 Diurnal and Semidiurnal Variations Upper-air pressures appear to follow a systematic diur-
nal/semidiumal cycle similar to that at sea level; however,
extremes occur at somewhat different hours. Table 15-29
Mean hourly sea-level pressures follow a systematic lists amplitudes and times of occurrence of diurnal and sem-
diurnal and semidiumal periodicity somewhat variable in idiumal maxima to 10 mb (roughly 30 km) over Terceira,
amplitude and phase according to location and season. The Azores, which provides an estimate of mean annual sys-
sea-level pressure cycle is generally characterized by min- tematic pressure variations at a maritime location near 40”N.
ima near 0400 and 1600 hours and maxima near 1000 and The semidiurnaf variations at climatically and geographi-
2200 hours local time. The amplitude approaches 1 mb, cally different locations such as Washington, D.C. and Ter-
which is small relative to synoptic changes in middle lati- ceira, Azores, appear to be similar [Valley, 1965]. The
tudes. In the tropics only minor synoptic changes occur from diurnal maxima and minima, however, that result from solar
day to day, so that interdiumal pressure changes are small insolation and terrestrial radiation, may differ considerably
compared to the systematic daily variations in these lati- in time of occurrence and amplitude at various locations,
tudes. particularly at or near surface levels.

15-51
CHAPTER 15
REFERENCES — CHAPTER 15

Albrecht, F., “Micrometeorologische Temperaturmessun- denen Boeden unter Gleichen Witerungs Verhaeltnis-
gen von Flugzeug aus,” Ber. Deutsch Wetterd, Bad Kis- ser,” Zeitschr, f angewandte Meteorologic, ~ 65, 1943.
singen, 38:332, 1952. httau, H., “A Study of the Mass, Momentum and Energy
Bertoni, E.A. and I.A. Lund, “Winter Space Correlations Budget of the Atmosphere,” Archiv fur Meteorologic,
of Pressure, Temperature and Density of 16 km,” AFCRL- Geophysik and Bioklimatologie, Serie A, 7:133, 1954a.
64-1020, ADA611OO2, 1964. Lettau, H., “Improved Models of Thermal Diffusion in the
Cole, A.E. “Review of Data and Models of the Middle Soil,” Trans. AGU, 35:121, 1954b.
Atmosphere,” Space Research, 201979. National Weather Service, “Surface Observations,” Federal
Cole, A.E. and A.J. Kantor, “Tropical Atmospheres, O to Meteorological Handbook 90-1, Government Printing
90 km,” AFCRL TR-75-0527, ADA019940, 1975. Office, Washington, D. C., 1979.
Cole, A.E. and A.J. Kantor, “Interlevel Correlations of Riordan, P., Weather Extremes Around the World, Report
Temperature and Density, Surface to 60 km,” AFGL ETL-TR-74-5, U.S. Army Engineering Topographic
TR-80-0163, ADA0905 15, 1980. Laboratories, Ft. Belvoir, Va., 1974.
Cole, A. E., A.J. Kantor, and C.R. Philbrick, “Kwajalein Singer, I.A. and R.M. Brown, “Annual Variation of Sub-
Reference Atmospheres, 1979,” AFGL TR-79-0241 , Soil Temperatures About a 600-Foot Diameter Circle,”
ADA08 17801979. Trans. AGU, 37:743, 1956.
Davidson, B. and H. Lettau (eds.), Great Plains Turbulence Sissenwine N. and A. Court, “Climatic Extremes for Mil-
Field Program, Pergamon Press, New York, 1957. itary Equipment ,“ Env. Protection Br., Rpt. No. 146,
Draeger, R.H. and R.H. Lee, “Meteorological Data Eni- Dept. of Army, Washington, D. C., 1951.
wetock Atoll,” Naval Medical Res. Inst. Memo Rept. Tattelman, P., “Duration of Cold Temperature Over North
53-8, Bethesda, Md, 1953. America,” AFCRL-68-0232, AD669567, 1968.
Geiger, R., The Climate Near the Ground, Harvard Uni- Tattelman, P. and A.J. Kantor, “Atlas of Probabilities of
versity Press, Translated by M.N. Stewart et al., 1957. Surface Temperature Extremes: Part I, Northern Hemi-
Gringorten, 1.1., “Probability Models of Weather Conditions sphere, “ AFGL TR-76-0084, ADA027640, 1976a
Occupying a Line or an Area,” J. Appl. Meteorol., 18: Tattelman, P. and A.J. Kantor, “Atlas of Probabilities of
957, 1979. Surface Temperature Extremes: Part II, Northern Hemi-
Harris, M. F., F.G. Finger, and S. Teweles, “Diurnal Vari- sphere, “ AFGL TR-77-0001 , ADA038237, 1976b.
ation of Wind, Pressure and Temperature in the Trop- Tattelman, P. and A.J. Kantor, “A Method for Determining
osphere and Stratosphere over the Azores,” J. Atmos Probabilities of Surface Temperature Extremes,” J. App/.
Sci., 19:136, 1962. Meteorol., 16: 175, 1977.
Holliday, C. R., “An Extreme Sea-level Pressure Reported Tattelman, P., N. Sissenwine, and R. W. Lenhard, Jr., “World
in a Tropical Cyclone,” Monthly Weather Rev., 103: Frequency of High Temperature,” AFCRL-69-0348,
163, 1975. AD696094, 1969.
Holliday, C. R., “Typhoon June — Most Intense on Re- Valley, S. (cd.), Handbook of Geophysics and Space En-
cord,” Monthly Weather Rev., 104: 1188, 1976. vironments, AFCRL, 1965.
Kantor, A.J. and A.E. Cole, “Time and Space Variations Wagner, A. J., “Weather and Circulation of October 1979,”
of Density in the Tropics ,“ AFGL TR-79-O1O9, Monthly Weather Rev., 108: 119, 1980.
ADA074472, 1979. WMO, “Proceedings of the World Climate Conference, Ge-
Kreutz, W., “Der Jahresgang der Temperature in Verschie- neva, 12–23 Feb 1979,” WMO Report No. 537, 1979.

15-52
 Chapter 16

WATER VAPOR, PRECIPITATION, CLOUDS, AND FOG

Section 16.1 D.D. Grantham, 1.1. Gringorten, and A.J. Kantor
Section 16.2 R.M. Dyer, 1.1. Gringorten, and P. Tattelman
Section 16.3 A.A. Barnes, E.A. Bertoni, I.D. Cohen, Capt. M. Glass, D.D. Grantham, 1.1. Gringorten,
K.R. Hardy, Y. Izumi, A.J. Kantor, and J.I. Metcalf
Section 16.4 H.A. Brown and B.A. Kunkel

One of the most important constituents of the atmosphere in a one-to-one correspondence to the dew point. Table
is that all-pervasive substance, water, which appears in the 16-1 shows this correspondence between dew point and frost
solid state as snow, hail, frost, and cloud particles; in the point, vapor pressure, and mixing ratio. The table also gives
liquid state as rain, clouds, fog, and dew; and in the gaseous the absolute humidity (vapor density) corresponding to the
state as water vapor. In addition to water, the atmosphere other measures when the air is saturated.
contains in suspension many types and sizes of aerosols of In a unit volume of air, the water vapor exerts a vapor
varying composition, including ions, salt, dust, smoke par- pressure, which in itself is another measure of the water
ticles and other contaminants. This chapter deals with a vapor content of the air. When the air is saturated, the vapor
variety of conditions or events that are related to atmospheric pressure is a direct function of air temperature.
water content, including surface rainfall rates; the vertical The absolute humidity is the mass of water vapor in a
distribution of precipitation intensity; drop-size distributions unit volume of air; it is actually the water vapor density
in rain, clouds and fog; types of cloud data; and water and within the air. The Gas Law shows that at a given air
water-vapor content in and out of clouds. temperature, there is a saturation water vapor contenf per
unit volume that the air can hold, which rough] y doubles
for each 10 K increase of air temperature. As a consequence
16.1 ATMOSPHERIC WATER VAPOR of the Gas Law, an isobaric increase of the temperature will
decrease the absolute humidity (vapor density) while the
There are various ways to express the water vapor con- other measures in Table 16-1 remain the same.
tent of the atmosphere. The dew point is the temperature to If the mass of water vapor is given in proportion to the
which’ a parcel of air, at constant pressure and water vapor mass of the ambient dry air, typically in grams per kilogram,
content, must be cooled for saturation to occur. Further the resulting measure is called the mixing ratio. To relate
cooling produces condensation to liquid water. Condensa- mixing ratio to the other measures of water vapor content,
tion could be delayed as the parcel is cooled below the dew the atmospheric pressure must be given. When the mixing
point, in which case the air becomes supersaturated. When ratio remains constant, as it does in adiabatic lifting of
the air is supersaturated, the condensation of water vapor unsaturated air, the dew point will decrease with increasing
can take place abruptly and the released latent heat of va- altitude.
porization will raise the air temperature. If the water vapor Relative humidity is the ratio of the actual vapor pressure
content is low enough, then the parcel of air might be cooled of the air to the saturation vapor pressure at the ambient air
below 273 K until saturation with respect to ice occurs, at temperature; it is given in percent. Relative humidity can
which point there usually is sublimation directly from water be derived with sufficient accuracy as the corresponding
vapor into ice crystals. This temperature is called the ~rosf ratio of the mixing ratio to saturation mixing ratio. At a
poinf. Sometimes the air can be cooled below 273 K, to as relative humidity of 100% the dew point is equal to the
low as 233 K without sublimation, but there could be con- ambient temperature. This can occur by the addition of
densation onto existing supercooled water drops. Because moisture into the air and/or by the decrease of temperature
the saturation vapor pressure over water is greater than that to the dew point. Thus, relative humidity usually varies
over ice, the dew point for a given water vapor content is considerably during the course of a day, increasing during
always less than the frost point. The frost point-dewpoint the cool of the night and decreasing in the heat of the day.
difference, although essentially zero at 273 K, becomes Precipitable water, W, is the amount of water contained
increasingly large as the temperature lowers. in a vertical air column of unit cross-section extending be-
Several other measures of water vapor content are related tween two specified levels. While the physical unit is the

16-1
CHAPTER 16
Table 16-1. The correspondence between the several measures of water vapor content (based on Smithsonian Meteorological Tables [List, 1968]).

Mixing Ratio (g/kg)

Dew Frost Vapor Absolute$
Point Point pressu~ Humidity 1000 850 700 500 400 100 50 10 1
(K) (K) (rob) (g/rn’) mb mb mb mb mb mb mb mb mb

313 7.378+ 1 5.119+1 4.98+ 1 5.941 + 1 7.361 + 1 1.080 + 2 1.411+2 * * * *

308 5.624+ 1 3.963 + 1 3.725 + 1 4.427+ 1 5.456+ 1 7.910+ 1 1.020+2 B.008+2 * * *
303 4.243+ 1 3.038+ 1 2.769+ 1 3.282+ 1 4.029 + 1 5.786+ 1 7.399+ 1 $.590+2 * * *
298 3.167+1 2.305+1 2.044+ 1 2.417+ 1 2.959+ 1 4.219+ 1 5.363+ 1 2.886+2 * * *
293 2.337+ 1 1.730+ 1 1.495 + 1 1.766+1 2.156+1 3.059+ 1 3.870+ 1 1.899+2 5.462 +2 * *
288 1,704+1 1.283+ 1 1.083+ 1 1.278+ 1 1.557+ I 2.201 + 1 2.775+ 1 1.279+2 3.217+2 * *
283 1.227+ 1 9.399 7.762 9.146 1.113+1 1.569 + 1 1.973+ 1 B.707+ 1 2.024+2 * *
278 8.719 6.797 5.495 6.471 7.870 1.107+1 1.389+1 5.946+ 1 1.314+2 * *
273 273.0 6.108 4.847 3.839 4.519 5.492 7.710 9.664 4.049+ I 8.659+ 1 9.764+2 *
268 268.6 4.215 3.407 2.644 3.112 3.780 5.3W 6.637 2.739+ 1 5.728+ 1 4.533+2 *
263 264.1 2.863 2.358 1,794 2.110 2.562 3.590 4.492 1.834+ 1 3.779+ 1 2.495+2 *
258 259.6 1.912 1.605 1.197 1.408 1.709 2.393 2.993 1.213+1 2.474 + I 1.470+2 *
253 255,1 1.2540 1.074 7.847–1 9.227–1 1.120 1.568 1.960 7.903 1.601+1 8.919+ 1 *
248 250.5 8.070– 1 7.047– 1 5.048-1 5.936– 1 7.204– I 1.W8 1.260 5.603 1.021 + 1 5.461+1 *
243 245.8 5.088– I 4.534– 1 3.182–1 3.742– 1 4.540– I 6.352–1 7.938– I 3.183 6.397 3.335+1 6.443+ 2
238 241.2 3.139–1 2.856– 1 1.%3–1 2.308– 1 2.801 – 1 3.918–1 4.896– I 1.960 3.931 2.016+ 1 2.846+2
233 236.5 1.891–1 1.757– 1 1.183–I 1.390-1 1.687–1 2.360– 1 2.948– 1 1.179 2.362 1.199+1 1.450+2

273.0 273 6.107 4.847 3.839 4.518 5.492 7.709 9.668 4.048 + 1 8.658+ 1 9.759+2 *
267.3 268 4.015 3.246 2.518 2.963 3.599 5.047 6.322 2.W+ 1 5.433 + 1 4.722+2 *
261.8 263 2.597 2.139 1.627 1.913 2.324 3.255 4.075 1.660+1 3.409+ 1 2.182+2 *
256.2 258 1.652 1.387 1.034 1.216 1.476 2.067 2.592 1.045+1 2.126+1 1.231 +2 *
250.8 253 1.032 8.835– 1 6.456– 1 7.592– 1 9.214– 1 1.289 1.613 6.490 1.311+1 7.158+ 1 *
245.3 248 6.323– 1 5.521 – 1 3.955 – 1 4.650 – 1 5.M3–1 7.895– 1 9.872– 1 3.961 7.969 4.199+ 1 *
239.9 243 3.798– 1 3.385–1 2.375–1 2.792– 1 3.388–1 4.740– 1 5.926– I 2.373 4.763 2.456 + 1 3.809+2
234.6 238 2.233–1 2.032–1 1.396– 1 1.&2–l 1.993 – 1 2.787– 1 3.483– I 1.393 2.791 1,42+1 1.788+ 2
229.3 233 1.283–1 1.192–1 8.026–2 9.434–2 1.1+–1 1.600-1 2.001 – 1 7.%– 1 1.601 8.084 9.154+ I
224.1 228 7.198–2 6.836–2 4.503–2 5.293–2 6.422–2 8.981 –2 1.122–1 4.483 – 1 8.970– 1 4.510 4.824+ I
223 3.935–2 3.821 –2 2.463 – 2 2.895–2 3.512–2 4.910–2 6.135–2 2.450 – 1 4.901 – 1 2.457 2.548+ 1
218 2.W2–2 2.078–2 1.3W–2 1.539–2 1.867–2 2.611 –2 3.261 –2 1.302– 1 2.604– 1 1.304 1.329+ I
213 1.080–2 1.098-2 6.761–3 7.947–3 9.640-3 1.347-2 1.684–2 6,723–2 1.344-1 6.725 – 1 6.791
208 5.006-3 5.627–3 3.386–3 3.979–3 4.826–3 6.749–3 8.427–3 3.365–2 6.728–2 3.362–1 3.381
203 2.615–3 2.784–3 1.639–3 1.926–3 2.336–3 3.265–3 4.076–3 1.628–2 3.254–2 1.627– 1 1.631
198 1.220–3 1.334–3 7.M6–4 8.986–4 1.090-3 1.524–3 1.902-3 7.593–3 1.518–2 7.590-2 7.597–1
193 5.472–4 6.138–4 3.423 – 4 4.023 – 4 4.882–4 6.828–4 8.530–4 3.406-3 6.810–3 3.404-2 3.405– 1
188 2.353–4 2.710–4 1.472–4 1.730–4 2.@9–4 2.936–4 3.668-4 1.465–3 2.928–3 1.4ti-2 1.464-1
183 9.672–5 1.144–4 6.051 –5 7.111–,5 8.629–5 1.207–4 1.508–4 6.020 – 4 1.2W–3 6.016–3 6.016–2
*At saturation only
*Atmospheric saturation is not pssible at this ambient temperature and pressure.

mass per unit area, it is commonly reported as the height lute humidity (g/m3) and r mixing ratio (@g), then in terms
at which the liquid would stand if it were all condensed in of the frost point, the vapor pressure is approximated by
a vessel of the same unit cross-section as the air column.
For example, a height of 1 cm corresponds to 1 gm/cm2 of logloe = – 2485 .O~~ + 3.5665 log]oT~
precipitable water.
A programmable computer can use a set of equations – 0.0032098 T, + 2.0702 (16.1)
or algorithms to generate the numbers given in Table 16-1.
Where TD represents dew point (K), T~ frost point (K), e In terms of the dew point, the vapor pressure is approxi-
vapor pressure (rob), p atmospheric pressure (rob), p, abso- mated by

16-2
 a
WATER VAPOR, PRECIPITATION, CLOUDS, AND FOG
loglOe = – 2949. l/T~ Table 16-2. Diurnal cycle of dew &int, temwrature, relative humidity,
and solar insolation associated w;th the 1% high operational
– 5.0281 loglOT~ + 23.832 (16.2) .
extreme of moisture content (dew .ooint 304 K) in a coastal
desert location,

In terms of the vapor pressure and the air temperature T(K),

Time Dew point Temperature R.H. Insolation
the absolute humidity (g/m3) is approximated by (LST) (K) (K) (%) L/h

p, = 216.68 e/T . (16.3) m 302 304 88 o

0300 302 304 88 0
This will be the same as in Table 16-1 for saturated air only. 0600 302 304 88 4
In terms of the vapor pressure (rob) and the atmospheric 304 310 73 68
pressure (rob), the mixing ratio (g/kg) is approximately by 1200 304 313 63 93
1500 3M 314 60 61
r = 621.97 e/(p-e) . (16.4) 1800 304 310 69 1
2100 303 306 85 0

16.1.1 Surface
a Department of Defense Mili& . Standard. Cliwric Ex-
Water vapor at the earth’s surface has a range of more tremes for Military Equipment [U.S. DOD, 1973] and are
than three orders of magnitude. The maximum amount of further discussed in the background studies document [Sis-
water vapor that the atmosphere can contain is regulated by senwine and Cormier, 1974]. Table 16-2 gives details of
its temperature. However, the amount actually present in the diurnal cycle of humidity, temperature, and solar in-
the atmosphere is a complex function of various atmospheric solation associated with the upper 1% dew point (304 ~)
and geophysical features. The highest value of atmospheric in a hot cQastal climate. In this kind of weather the dew
water vapor, observed in the Persian Gulf, is a dew point point remains nearly constant throughout 24 hours, while
of 307 K, translating to 38 g/m3 of water vapor or a mixing the relative humidity changes from nearly 9090 to 60%,
ratio of about 35 g water vapor per kg dry air. For extremely respectively, as the temperature climbs from 304 K in the
cold air (213 K) the moisture content, at best, can be no morning to 314 K in the afternoon. Table 16-3 presents a
more than 0.02 g/m3, or, at sea level, 0.01 g water vapor repeated diurnal cycle for an extreme moisture-laden case
per kg dry air. Even this amount, however, can be impofi”mt of 30 days duration. The conditions are maritime, with
in the transmissionof infrared radiation overlong path lengths. temperatures 300 K to 303 K, dew points 299 K to 301 K
Information on the distribution Qf water vapor in the and relative humidity at 91% to 9490. Table 16-4 presents
northern hemisphere was collected in the Atmospheric Hu- a dium’al cycle that has high relative humidity associated
midity Atlus [Gringorten et al., 1966], which maps sutiace with high temperature. This case differs from that in Table
mixing ratios and dew points from the 5th to the 95th per- 16-2 by having less absolute humidity, less insolation and
centile for the midseason months. Two atlases, for the north- lower temperature, but greater relative humidity throughout
ern and southern hemispheres [Crutcher and Meserve, 1970; the 24 hours. Relative humidity varies from 7590 to a more
Taljaard et al., 1969] have been published that inchtde mean persistent 100% as the temperature varies from an afternoon
monthly dew points and standard deviations of the monthly maximum of 308 K to a nighttime minimum of 299 K.
averages. Table 16-5 depicts the persistent high relative humidity of
Figure 16-1, taken from the Atmospheric Humidity At- a steaming jungle or rain forest. While the temperature
las, presents 95th percentile sutiace dew hints in JuIy to
illustrate high values of dew points over the northern hemi- Table 16-3. A repeated diurnal cycle during a month of sustained extreme
sphere. The peak values are found in the Persian Gulf above sustained moisture content in ~ugust at Belize City, Belize.
exceptionally warm waters. Over deserts the water vapor
content is lQwer, and coupled with high tem~ratures, the Time Dew Point Temperature R.H. Insolation
relative humidity drops to as low as 590 or less. Figure (LST) (K) (K) (%) L/h
16-2, also from the Atlas, presents 5th percentile surface
300 301 91 o
dew points in January to illustrate low moisture content
0300 299 301 92 0
during the coldest time of the year. As expected, the dew
299 300 94 4
point is lowest in polar and subpolar regicms, but it is also
0900 300 302 92 54
low on high mountains. On the other hand, relative humidity
1200 301 303 91 83
can be high since it would not require much moisture to
1500 301 302 91 68
saturate the air at low temperatures.
1800 300 302 91 20
World-wide extremes of humidity and the diurnal cycles
2100 300 302 91 0
of humidity in several distinctive climates are presented in

16-3
CHA~ER 16

Figure 16-1. Dew poirtt (K-273) at the surface in July, 95th percentile [Gringorten et al., 1966].

consistently remains moderate at 297 K, the relative hu- northern hemisphere is provided in the Atmospheric Hu-
midity persists above 95Y0. Such conditions may prevail midiry Atlas [Gringorten et al., 1966]. Dew points aloft are
several days a month throughout the year. mapped at 850, 700, 500 and 400 mb in percentiles from
5th to 95th for mid-season months. Also, two northern and
southern hemispheric atlases [Crutcher and Meserve, 1970;
16.1.2 Troposphere Taljaard et al., 1969] have been published showing mean
monthly dew points and standard deviations of the monthly
Climatic humidity data are relatively plentiful for alti- averages for altitudes up to 500 mb.
tudes up to 7 or 8 km (approximately 400 rob). Information Aloft, relative humidities of 100% can occur at any
on the distribution of water vapor at these levels in the altitude below the tropopause over the entire globe. ~ow-

16-4
 WATER VAPOR, PRECIPITATION, CLOUDS, AND FOG
— ;,, ~

,/

(’
‘,

Figure 16-2. Dew point (K-273) at the surface in January, 5th percentile [Gringotienet al.. 1966].

ever, the greatest moisture content is imbedded in the mon- sing low concentrations. Consequently, it is not possible at
soonal flow over northern India. Table 16-6 presents high this time to describe spatial and temporal variations of water
dewpoints with 1~o, 5%, 10%, and 2090 estimated proba- vapor in the upper troposphere in a manner comparable to
bilities in July from 1 to 8 km over northwestern India. that given for altitudes below 8 km.
Table 16-7 presents the low frost points with 1%, 5%, 10%, For altitudes from 8 km up to the tropopause, atmos-
and 20% estimated probabilities in January from 1 to 8 km pheric moisture can be estimated using information obtained
in the Canadian Northwest Territories. from several different experimental sensors. In the UPWr
Water vapor measurements above radiosonde humidity troposphere the median mixing ratio appears to decrease
altitudes (roughly 8 km) are not made on a regular basis approximately logarithmically with increasing altitude up to
because of a lack of radiosonde instruments capable of sen- the tropopause. As a result, median dewpoints, given at 400

16-5
CHAPTER 16
Table 16-4. A diurnal cycle associated with temperature and high relative mb (7 to 8 km) in the Atmospheric Humidity Atlas [Grin-
humidity in the rainy tropics. gorten et al., 1966], cah be converted to median mixing
ratios and extended upward logarithmically to medians of
Time Dew Point Temperature R.H. Insolation
2 to 4 ppmm (parts per million by mass) at the appropriate
(LST) (K) (K) (%) L/h
tropopause height. The altitude of the tropopause ranges
300 300 I& o from 16 or i7 km in the tropics to 9 or 10 km near the pole
0300 299 299 100 0 [Cole and Kantor, 1978]. Estimated median mixing ratios
0600 299 299 100 4 at various latitudes are shown in Table 16-8.
0900 300 304 82 54 World maps of the mean precipitable water above the
1200 302 307 75 83 surface and above 850, 700, and 500 mb have been pub-
1500 302 308 74 68 lished for the midseason months [Bannon and Steele, 1960].
1800 302 305 82 20 More recently, reports on “~ipitable Water Over the United
2100 300 301 95 0 States” were compiled by the Natignal Weather Service
ILott, 1976; Ho and Riedel, 1979]. Based on radiosonde
observations, the range of the maximum precipitable water
Table 16-5. A typical diurnal cycle in a tropical jungle with high relative content over the contiguous United States is about 1 cm
humidity. (Such conditions may occur several days a month
throughout the year. )
over the Rocky Mountains to nearly 7 cm along the Gulf
of Mexico in the surface-to-400 mb layer. Table 16-9 pre-
Time Dew Point Temperature R.H. sents mean monthly maximum precipitable water content
(LST) (K) (K) (%) for Key West, Florida and Lander, Wyoming.

297 297 100

0300 297 297 100
16.1.3 Stratosphere and Mesosphere
0600 297 297 100
0900 296 297 95
Water vapor measurements in the stratosphere and me-
1200 296 297 95
sosphere are not made at regular intervals because standard
1500 296 297 95
instrumentation is not available for use at these altitudes.
1800 296 297 95
Measurements at very low concentrations typical of the strat-
2100 297 297 100
osphere can be made only with research qualit y instruments.
Such sensors include frost point hygrometers, optical de-
Table 16-6. High absolute humidity (dew point) extremes. 1 to 8 km aloft
vices (such as spectrometers, radiometers), cryogenic air
(ty~ical of notihem I~dia in july). samplers and absorption devices. Roughly two dozen in-
vestigators have each provided from a few to over 100
I Dew Point I observations during the last 35 years, with the bulk of strat-
(K) ospheric humidity measurements coming since 1965.
Altitude I I ,
(km) 1% Probable 5% 10% 20% The earliest measurements, made by British aircraft
[Brewer, 1949], provided evidence that the tropopause acts
1 302 300 299 298 as a cap to water vapor, preventing it from penetrating into
2 297 295 294 293 the stratosphere. General atmospheric circulation theory in-
4 289 286 283 282 dicates that tropospheric air passes into the stratosphere over
6 276 273 272 271 equatorial regions from where it moves northwm’d. It sinks
8 265 262 261 Z60 back into the troposphere at higher latitudes, closing the
cycle. Tropical tropospheric air following this route into the
stratosphere would have to pass through the very cold tro-
Table 16-7. Lowabsolute humiditv (dew ooint) extremes. lto8km aloft popause of the tropics ( 193 K) where th~ water va~r density
(typical of Canadian Northw;st Territories in January).
would be forced to extremely low concentrations due to

Altitude
1 Frost Point
(K) I condensation and sublimation. As this air continues upward
from this level, the mixing ratio in the stratosphere must
remain constant if there is neither source nor condensa-
(km) 1% Probable 5% 10% 20% tion/sublimation of water vapor. The mixing ratio that most
closely depicted these British findittgs is 2 mg per kg which
1 223 231 233 236
may be interpreted as 2 ppmm. Some investigators report
2 222 228 231 234
their measurements in terms of volume mixing ratio, de-
4 217 221 222 224
noted as ppmv. The relationship between ppmv and ppmm
6 208 212 213 214
is ppmv = 1.607 ppmm. Other possible sources that have
8 202 202 204 205
been suggested for water vapor to be introduced into the

16-6
 H

WATER VAPOR, PRECIPITATION, CLOUDS, AND FOG

Table 16-8. Estimated distribution of median mixing ratios in the upper troposphere.

Altitude 15°N 30°N 45”N 60”N 75”N

(km) Jan Jul Jan Jul Jan Jul Jan Jul Jan Jul

(ppmm) (ppmm) (ppmm) go ~:ppmm) (ppmm)

400 mb* 433* 105O* 295* 722* 155* 540” 401* 63.4* 295*
7 . — — — 148 — 57.5 — 34.2 —
8 346 010 213 510 38.5 368 11.3 48.7 8.3 24.6
9 205 513 130 239 9.8 136 2.0 13.2 2.0 2.0
10 122 250 80.0 113 2.5 50 2.0
11 71.2 119 49.0 52.5 18.3
12 42.0 57.2 29.8 24.6 6.7
13 24.6 27.5 18.3 11.5 2.5
14 14.5 13.2 11.2 5.4
15 8.6 6.3 6.8 2.5
16 5.3 3.0 4.1
17 3.0 2.5
$W-mbmixing ratios calculated from Gnngofienet al. [1966].

stratosphere are the transport of tropospheric air upward of water vapor with mixing ratios up to about 4 ppmm
through the overlapping tropical and polar tropopauses, cu- occurring in 1- to 2-km layers at altitudes up to about
mulonimbus clouds penetrating the tropopause, and chem- 30 km.
ical reactions such as the oxidation of methane. The extent Based on three independent series of field measure-
to which such mechanisms influence the stratospheric hu- ments, an annual cycle for stratospheric water vapor has
midity regime continues to be debated. However, the com- been suggested, in which the amplitude of the seasonal
bined effects of the Hadley cell circulation and tropical variation is largest in the lowest part of the stratosphere,
thunderstorms, both of whose inputs of moisture into the about 0.4 ppmm at 15 km, and decreases with altitude to
stratosphere must penetrate the tropical “cold trap ,“ are re- about 0.1 ppmm above 20 km. The phase of the annual
sponsible for the bulk of stratospheric mixture. In fact, most cycle also changes with altitude, with the maximum oc-
investigators agree that all other sources are smaller by a curring in November at 15 km and in April at 20 km. This
factor of 5 to 10. annual cycle may be related to the annual variations in the
In the early 1960s there was a significant controversy temperature and altitude of the tropical tropopause.
concerning the amount and distribution of water vapor in In the lower stratosphere (below 20 km) there seems to
the stratosphere [Gutnick, 1961]. Much uncertainty contin- be a small but discernible latitudinal variation in mixing
ues, but the extent of uncertain y has narrowed. The bulk ratio, with higher values, about 3 ppmm, occurring near the
of credible observations of stratospheric mixing ratio now equator and decreasing with latitude to somewhat less than
range from about 0.5 to 5 ppmm. The stratospheric water 2 ppmm near 45° latitude in both hemispheres. The lack of
vapor measurements made over the previous decade or so data in polar regions does not allow estimates of variation
have been summarized in two recent articles [Harris, 1976 at higher latitudes.
and Penndoti, 1978]. The conclusions are that the mean There have been even fewer observations of water vapor
midlatitude stratospheric mixing ratio is 2.5 t 0.7 ppmm in the mesosphere than in the stratosphere; consequently,
and is nearly constant with altitude up to at least 30 km. these distributions and variations continue to be controver-
There is some evidence that from 30 km the mean may sial. However, there does seem to be reasonably consistent
gradually increase to 3 ppmm at 45 km. This increase could hypotheses of the general envelope of the vertical profile of
be the result of an increase in water vapor due to the oxi- mixing ratio from the stratopause to the mesopause. Deguchi
dation of methane in the middle atmosphere. There is also and Muhleman [1982] compared their ground-based radio
limited evidence for the existence of weak, shallow layers telescope measurements to rocket measurements by Scholz

Table 16-9. Mean monthly maximum precipitable water content—two extremes.

Layer, Surface to
Station Month 850 mb 700 mb 500 mb 400 mb

Key West, Fla. Sep 2.7 cm 4.2 cm 5.2 cm 5.4 cm

Lander, Wyo. Jan — 0.4 0.7 0.8

16-7
CHAPTER 16
et al. [1970], Ehhalt et al. [1975] and Rogers et al. [1977]; 16.2.1 Surface Rates of Rainfall
airborne measurements by Farmer et al. [1980] and Waters
et al. [1977]; and current photochemical models by Crutzen Frequency distributions of various rainfall rates from
[1974] and Allen et al. [ 1981]. A concensus of these ob- actual data are available for only a small number of mete-
servations and models provides a midlatitude mesospheric orological stations. Precipitation data for most areas of the
water vapor profile of 4 to 5 ppmm at 50 km decreasing world are limited mainly to average monthly and annual
logarithmically with altitude to roughly 1 ppmm at 90 km. totals, and to the number of days on which precipitation
Such a profile is consistent with microphysical observations fell. Clock hourly precipitation data (totals on the hour every
of noctilucent clouds. An experiment in which noctilucent hour) are available for numerous stations in the United States
cloud particles, captured by a rocket at the 80 km mesopause and Europe, but for only a few stations elsewhere. Monthly
level included ice [Soberman, 1963, and Michaels, 1965], tabulations of the amounts of precipitation which fall in 5
implies the existence of water vapor through a layer at this to 180 min periods are also available for National Weather
high altitude. Such clouds are not found at tropical latitudes Service first-order stations in the United States and its island
and rarely found at middle latitudes where temperatures are possessions. Special 1-, 2-, and 4-rein rainfall rates have
seldom colder than 173 K [Cole and Kantor, 1978] at this been taken at a number of locations in Southwest Asia,
level. They are normally seen at high latitudes in the summer Europe and the United States.
where extremely cold mesopause temperatures prevail. Dur- These data have been used by scientists to develop math-
ing one experiment [Anonymous, 1963] in which such clouds ematical models for estimating the frequency of occurrence
were physically sampled, the temperature observed was 130 of various duration-rainfall rates at locations where only
K. During another sampling, when the clouds had dissi- routine rainfall observations are available. Some selected
pated, a temperature of 153 K was measured. Assuming the methods for estimating clock hour and instantaneous (1 rein)
temperature and pressure structure of the Air Force Refer- rainfall rates are presented in the following sections.
ence Atmospheres [Cole and Kantor, 1978], frost points of
130 K and 153 K at 80 km convert to saturation mixing 16.2.1.1 Clock Hourly Rates. Several researchers have
ratios of about 0.01 ppmv and 6 ppmv, respectively. used available clock-hour rainfall data to derive methods
It is interesting to note that there is a region of the middle for estimating the number of hours that specific rates occur
atmosphere, roughly 40 to 65 km, in which the combined at other locations. In 1965, Cole et al., [1965] presented
temperature and pressure structure does not support the ex- three linear equations for determining the frequency of 3
istence of clouds [Schilling, 1964]. The relatively warm clock-hour precipitation rates ( 1.5, 3.0, and 4.6 mm/h) vs
temperature structure at these altitudes requires a saturation an index of rainfall intensity. The index is the average per
vapor pressure greater than the total atmospheric pressure— day of measurable precipitation obtained by dividing the
an obvious impossibility. This condition is denoted by the total annual precipitation by the number of days with meas-
asterisk entries in Table 16-1. urable precipitation of >0.25 mm. Winner [ i 968] elabo-
rated on this and other earlier studies to develop techniques
for estimating clock-hours of rainfall at rates of trace, 0.25,
0.5, 2.5, 6.4, 12.7, 25.4 and 50.8 mm/h, based on data
16.2 PRECIPITATION from 123 National Weather Service stations with 5 to 10
year periods of record. The technique incorporates a mois-
Knowledge of the frequency of occurrence of given rates ture index, Y, where
of precipitation and associated vertical distributions of pre-
cipitation particles is important to the design and operation
of many types of equipment. It is a critical input for eval-
uating and deploying systems which employ surveillance
‘=lm[*-ll (16.5)

radar, and for the selection of microwave frequencies used P is the mean annual rainfall in mm, and PE is the mean
in communicating systems. The frequency and intensity of annual potential evapotranspiration determined by the method
rainf~l must also be considered in the design and operation developed by Thomthwaite et al. [ 1958].
of helicopter rotor blades, and leading edges of aircraft and Figures 16-3, 16-4, and 16-5 are nomograms for esti-
missiles which are subject to erosion by precipitation par- mating the mean annual clock-hours per year that a rainfall
ticles. Water ingested by jet engines can cause flame-outs rate of 6.4, 12.7, and 25.4, mm/h is equaled or exceeded
and penetrate protective coverings of exposed electronic and in temperate latitudes. Potential evapotranspiration data for
mechanical material. Precipitation in the form of hail is a Equation (16.5) can be calculated using Thomwaite et al.,
serious hazard for aircraft, whereas snow loads and ice [1958]. It is desirable to use the same period of record for
accretion on surface structures can destroy towers, radomes, all variables in the nomograms. If these publications are not
and buildings. This section contains examples of the type available, clock-hour estimates of more limited value can
of precipitation information available. be made by assuming Y to be zero.

16-8
 a
WATER VAPOR, PRECIPITATION, CLOUDS, AND FOG
RUR@L MODELS 15.000 PHRT/Ct19

0.8

06

0.4

0,2

o. ~
o 1020304050~ 70
CLOCK WR2(Z)

Figure 16-3. Nomogram for determining average number of clock hours

per year a rainfall rate of 6.4 mm/b is equaled or exceeded.

The ratio of the average annual rainfall to the number

of days with precipitation =0.25 mm for a location is then
obtained from climatic data. Knowing this value and the
value for Y, the number of clock-hours can be determined
from the nomogram.
The technique described above provided poor results
when applied on independent data from tropical locations. WAVELENGTH ( ~m)
Therefore, Winner [1968] developed an alternate method
for use in the tropics using data from 32 stations in the Figure 16-5. Nomogram for determining average number of clock hours
Panama Canal Zone. The best results were obtained from per year a rainfall rate of 25.4 mm/his equaled or exceeded.
a linear correlation of the annual rainfall vs the number of
clock-hours at specified rates. Figure 16-6 presents the clmk-
hours per year that rates of 6.4, 12.7, 19, and 25.4 mm/h I I I 1 ! 1 1 I 1 I
are equaled or exceeded in the tropics. 200

I 80
16.2.1.2 Instantaneous Rates. For most applications, the
distribution of rainfall rates for short periods, usually for 1
rein, are of the greatest importance since these can greatly

E
E 1
I 1 I
x I I

08 – Z-6.15+51.64X+0.OIIY
CORRELATION CCJEFFICIENT= 0.828
sTANDARD ERROR = 3.75 HOUI?3

06 -
... .. . . .
S7,”0..0 E..OR OF EST,M.lE
CMF W CORRELAT80M

RATE, I 2.7 mmlh

/ “:685.042.
sTAN04R0 ERFIOROF ESTIMATE
-100 CORRELATION
04 – 19 mm/h
0 D 23X
D ER~ OF ESTIMATE
IOO(Y) ;OEF”OF CMRELAT(W
R,TE: 25.4 mmlh
’20 “:2.,0.0 ,2,
02 – SIANDAROER.W OF ESTIMATE
COEFW CO. RELATKU

1 I I 1 1 1 1 I
40 60 80 100 I 20 I 40 160 I 80
CLOCK-HOURS (Y]

CLCCK - WUR (Z 1

Figure 16-6. Nomogram for use in tropical regions to determine mean

number of clock hours occurrence of a specified precipitation
Figure 16-4. Nomogram for determining average number of clock hours
rate when the mean annual precipitation amount is known.
per year a rainfall rate of 12.7 mm/his equaled or exceeded.

16-9
CHAPTER 16
Table 16-10. Percent contribution of instantaneous precipitation rate to clock-hour precipitation rates for Urbana, Illinois (annual).

Instantaneous Rates
Clock-hour (mm/h)*
Rates o.oo– l.o– 2.5– 6.4– 12.7– 25.4– 50.8 – 127– ‘ Tot No
mm/h* 0.76 2.3 6.1 12.4 25.1 50.5 126.7 254 >254 Clock-h

Trace 96.02 3.43 0.51 0.04 0.0 0.0 0.0 0.0 0.0 576
0.25 87.56 10.38 1.96 0.06 0.04 0.0 0.0 0.0 0.0 114
0.5 – 2.3 59.17 29.16 10.17 1.16 0.25 0.08 0.01 0.0 0.0 581
2.5 – 6.1 27.34 19.24 37.58 11.82 2.82 1.05 0.15 0.0 0.0 I79
6.4 – 12.4 23.50 11.28 23.01 23.14 12.16 5.33 1.58 0.0 0.0 61
12.7 – 25.1 20.45 8.53 14.74 16.28 16.99 13.53 8.65 0.83 0.0 26
25.4 – 50.5 6.67 5.56 8.89 13.33 19.44 20.00 25.56 0.56 0.0 3
50.8 – 76 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0
76.2 – 101.3 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0
101.6 – 126.7 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0
=127 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0
*This work was originally presented in English units. Direct conversion to metric is the reason for small gaps between each range of rates.

exceed the hour]y rate. A 1-rein rainfall rate is generally contribution of the instantaneous rate to clock-hour rate for
referred to as being “instantaneous.” Bussey [1950] found Urbana, Illinois and Miami, Florida.
that about 20% of the time during an hour the instantaneous A more direct method for determining instantaneous
rate was zero, or a very low amount called a “trace.” About rainfall rates was developed by Jones and Sims [1978]. They
35% of the time the hourly rate was exceeded, and it was used instantaneous rainfall rate data for 9 observation sites,
common for the hourly rate to be exceeded by 5 or 6 times plus 4-rein rainfall rate data at another 10 sites to evaluate
for a few minutes. the frequency of rainfall rates based on a world rain climate
Davis and McMotTow [1976] developed tables of clock- classification used by Herbstritt [1973]. Figure 16-7 from
hour rate versus instantaneous rate for 6 locations, and clock- the Jones and Sims report shows the climatic classifications
hour rate versus 4-rein rate for 7 locations from original and the data stations. The data for stations within each of
data. They indicate that these can be used to predict the 1- four of the five climatic zones were summarized and are
or 4-rein rate for any region having a similar precipitation presented in Figure 16-8. Although the curves in this figure
regime as determined from available clock-hour data. A are intended to represent instantaneous rainfall rates, mixing
variety of climates, including the wet tropics, is represented. of the 1- and 4-rein data has the net effect of giving a slightly
Tables” 16-10 and 16-11 from this study give the percent low estimate for less than 1~’ exceedance-frequency~

Table 16-11, Percent contribution of instantaneous precipitation rate to clock-hour precipitation rates for Miami, Florida (annual).

Instantaneous Rates
Clock-hour (mtih)”
Rates o.oo– l.o– 2.5 – 6.4 – 12.7– 25.4– 50.8 – 127– Tot No
mtih * 0.76 2.3 6.1 12.4 25.1 50.5 126.7 254 >254 Clock-h

Trace 93.55 4.60 1.49 0.33 0.04 0.0 0.0 0.0 0.0 92
0.25 85.31 9.44 4.01 1.17 0.06 0.0 0.0 0.0 0.0 27
0.5 – 2.3 65.67 19.97 11.13 2.30 0.66 0.26 0.01 0.0 0.0 170
2.5 – 6.1 39.21 16.40 29.82 9.84 2.21 2.14 0.38 0.0 0.0 74
6.4 – 12.4 29.81 11.52 18.64 22.42 9.47 5.87 2.23 0.04 0.0 44
12.7 – 25.1 34.04 6.32 16.67 10.61 10.00 11.14 10.18 1.05 0.0 19
25.4 – 50.5 15.50 12.17 7.50 12.33 13.17 13.67 20.83 4.83 0.0 10
50.8 – 76 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0
76.2 – 101.3 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0
101.6 – 126.7 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0
=127 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0
This work was originally presented in English units. Direct conversion to metric is the reason for small gaps between each range of rates

16-10
 WATER VAP6R, PRECIPITATION, CLOUDS, AND FOG

ITuC)E I!/TERIOR

Hgure 16-7. Rain climates of theworld [after Herbstntt, 1973].

10-1 A more comprehensive model for determining instan-

taneous precipitation rates exceeded for a specified percent
of a year was developed by Crane [1980]. The world was
divided into 8 regions, as shown in Figure 16-9, on the
basis of total rain accumulation, and maps showing the
numkr of thunderstorm-days published by Landsberg [1974].
Additional guidance was taken from the Koppen world cli-
mate classification. Boundaries were adjusted to accom-
modate variations in terrain, predominant storm type and
1A \\\ motion, general atmospheric circulation, and latitude. Sat-
Q 10-1
k ellite and precipitation frequency data were used to extend
the rain climate regions over the ocean; however Feldman
[1979] presents an elaborate expansion of the model for
ocean areas. The U.S. is spanned by 5 regions; but one of
them is divided into 3 sub-regions as shown in Figure
16-10.
The measured instantaneous rain rate distributions avail-
able for each of the 7 regions and 3 sub-regions were pooled
_MAHlll ME >UB-l MUYILSl- to construct median rain rate distributions. These are pre-
F
1-
——CONTINENTAL
----MARITIME
TEMPERATE
TEMPERATE
\\

1
sented in Table 16-12. No data were available for Region
A, and the author does not explain how its distribution was
—MID-LATITUDE l~TERIOR
I I 1111111 1 1 Ill
derived.
I I {111111 I 1 1111111
I 0-4 Tattelman and Grantham [1982] discuss sources of 1-
100 10’ I 02 103
10-’
min data and present a survey and comparison of available
THRESHOLD PRECIPITATION RATE, mm hr-i
m@els for estimating 1-rein rates. They found that the best
estimates of annual statistics on 1-rein rates are made using
Figure 16-8. Average rainfall ratefrequency relationships for four rain
climates.
the Davis and McMorrow model for locations where clock-

16-11
 IAPTER 16
RHn Ram ClmltiR*Om
Polw Tunwrate Sub T?wual lrw~l

DA Tundra IDV) Q c Marl,,w ❑E Wet ❑G M&vme

~B T.tw (Mod.rzte} ❑ L) Ce.nt,nenta, E F Ar,d ~ H Wet

75

POLAR

—
60

45~
TEMPEPATE
fc

-b , —

~, TROPICAL

IATE

Le H
A
60— “ POLAR

180 w 150 120 w 60 30 0 20 60

LONGITUDE ( DEG)

Figure 16-9. Rain rate climate regions [Crane, 1980]. (Reprinted with permission from IEEE 0 1980)

hour data are available. Tattelrnan and Grantham also found Two tables represent “convective” and “stratiform” precip-
that good estimates from clock-hour data could be made itation. The third table, which was found to be most ac-
using a model developed by Huschke et al. [ 1982]. Huschke curate, simply includes all the data. Where clock-hour data
et al. consolidated the tables for the individual locations are not available, Tattelman and Grantham found the Crane
from Davis and McMorrow into 3 tables for all locations. model produced estimates of annual 1-rein rain rate statistics
at least as good as any of the other available models.
A model for estimating instantaneous rainfall rates eqtialed
or exceeded 2.OYO, 1.OYO,0.5Y0, O. lYo, 0.05% and O.OlqO
of the time during a month from readily available clima-
tological data was developed by Lenhard [ 1974] and im-
proved by Tattelman and Scharr [ 1983].
The Tattelman-Scharr model was developed using step-
wise multiple regression analysis for each of six exceedance
levels (p = 2.OVO, 1.090, 0.5Y0, O. IYo, 0.05% and 0.01%
of time during the month). The multiple regression equation
is expressed by

RP = A, + BPT + CPI + DP f(L,T) (16.7)

where RP is the estimated precipitation rate (mm/min) for

exceedance level p, T is the monthly mean temperature (°F
or 1.8 K – 459), 1 is a precipitation index (monthly mean
Figure 16-10. Rain rate climate regions within the United States [Crane.
1980]. (Reprinted with permission from IEEE G 1980) precipitation divided by the monthly mean number of days

16-12
 a
WATER VAPOR, PRECIPITATION, CLOUDS, AND FOG
Table 16-12. Point rain rate (R.) distribution values (mm/hr) versuspercent of year rain rate is exceeded

Rain Climate Region

Percent
of Year A BIC D, D2 D~ E F G H

0.001 28 54 80 N 102 127 l% 66 129 251

0.002 24 40 62 72 86 107 144 51 109 220
0.005 19 26 41 50 64 81 117 34 85 178
0.01 15 19 28 37 49 63 98 23 67 147
0.02 12 14 18 27 35 48 77 14 51 115
0.05 8.0 9.5 11 16 22 31 52 8.0 33 77
0.1 5.5 6.8 7.2 11 15 22 35 5.5 22 51
0.2 4.0 4.8 4.8 7.5 9.5 14 21 3.2 14 13
0.5 2.5 2.7 2.8 4.0 5.2 7.0 8.5 1.2 7.0 13
1.0 1.7 1.8 1.9 2.2 3.0 4.0 4.0 0.8 3.7 6.4
2.0 1.1 1.2 1.2 1.3 1.8 2.5 2.0 0.4 1.6 2.8

Number of o 25 44 15 99 18 12 20 2 11
Station
Years of
Data

with precipitation), and f(L,T) is a latitude-temperature term regression anal ysis for each of the six exceedance levels,
defined as including regression coefficients and constants to be used
in Equation (16.7), are provided in Table 16-13, ‘16-14 and
o L --23,5”
16-15 for indices based on rainy day threshold values of
[ 2.54 mm, 1 mm, and 0.25 mm, respectively.
f(L,T) = { (L – 23.5)T 23.5°< L C40°
The Tattelman-Scharr model was subjectively evaluated
( (40 – 23.5)T L>40°
(16.8)
by estimating rates at independent locations representing a
wide variety of the earth’s climates. Results indicted several
circumstances when the model is either invalid or should
AP is the constant for exceedance level p, and BP, Cp, and be used with discretion. This occurred for very dry or cold
D, are multiple regression coefficients for T, 1 and f(L,T), months for which there were little or no data among the
respectively, for exceedance level p. dependent stations. The model was found to be generally
The minimum threshold amount of precipitation to de- invalid when any of the following conditions exist for a
fine a rainy day varies with country. Three of the most specific month at a location:
common threshold values used wQrldwide to define a rainy 1. T=273K
day are 0.25, 1, and 2.54 mm. The number of days per 2. I < 2mm/day
month with another frequently used threshold called a “trace” 3. Number of rainy days < 1 .
differs only slightly with the number of days equal to or IncQnslstencies such as negative rates, or increasing rates
greater than 0.25 mm. Results of the stepwise multiple with increasing exceedance level occasionally occur when

Table 16-13. Result$, of stepwise multiple regression analysis for exceedance levels p = 0.01%, 0.05%, O. 10%, 0.50%, 1.0%, and 2.0% based on a
threshold value of 2.54 mm for I. (The regression coefficients are given for each independent variable. )

Constant T 1(2.54) f(L,T) SEE

P Ap Bp Cp D, R (mdmin)

0.01 –0.91 2.8 X l@2 2.3 X 10-2 –3.4 x lb 0.83 0.43

0.05 –0.50 1.6 X 102 1.9 x 10* –3.1 x i& 0.86 0.24
0.10 –0.31 1.1 x 102 1.4 x lf3* –3.0 x l& 0.85 0.19
0.50 –0.01 2.5 X l@3 5.4 x lf33 –1.5 x l& 0.76 O.w
1.0 0.03 7.4 x l& 2.9 X 10-3 –7.6 X l~s 0.67 0.06
2.0 0.04 –2.0 x 104 1.5 x l&3 –3.2 X 105 0.64 0.02

16-13
CHA~ER 16
Table 16-14. Results of steuwise multiple regression analysis for exceedance levels P = 0.01%, 0.05%, O. 10%, 0.50%, 1.OYO,and 2.OYr based on a
threshold valuk of 1.00 m-m for-I. (The regression c~fficients are giveri for each independent variable. )

Constatlt T !(I.00) f(L,T) SEE

P Ap Bp Cp D, R (mm/min)

0.01 –1.00 2.8 X l~z 3.6 X l@2 –2.2 x Iti 0.84 0.41
0.05 –0.56 1.6 X l@2 2.5 X I@z –2.4 X l& 0.88 0.23
0.10 –0.36 1.1 x l@* 2.0 x 1~* –2.4 X l& 0.87 0.18
0.50 –0.03 2.4 X 103 7.8 X 10-3 –1.2 x lW 0.79 0.09
1.0 0.02 6.9 X l& 4.2 X l@3 –6.2 X l@5 0.71 0.05
2.0 0.04 –1.8 X Id 2.0 x lr~ –2.6 X l@5 0.67 0.02
.

T is between 273 K and ’277 K. When T is in this range, Sissenwine [1973] found Pensacola, Florida to have the
estimated rates are acceptable if they are positive and de- most intense rainfall of the 203 stations; the curves for
crease with increasing exceedance level. If the estimated Pensacola are reproduced in Figure 16-11 for 2- to 100-yr
rates are consistent, btit a negatiye rate is estimated for return Wriods. Smoothed analyses of the data from indi-
p = 2%, this may indicate that it rains less than 2% of the vidual stations were subsequently published by the National
time during that month. Weather Service because me variable nature of rainfall can
When there are between 1 and 3 rainy days in the month, result in substantial differences in rainfall intensity statistics
the model may estimate rates for each of the exceedance between stations separated by a relatively short distance [for
levels which, when integrated, indicate a total rainfall two example, Frederick et al., 1977]. A return period tipresents
or more times greater than the monthly mean precipitation. the time interval that an event would be expected to occur
Under these circumstances heavy, but infrequent, convec- at least once. Mathematically, a return period of n years is
tive precipitation accounts for just about all of the precip- the reciprocal of the probability of the event Occurnng in
itation in the month. Therefore, the most accurate estimates one year. If we denote the probability of the event not
are for the two or three iowest exceedance levels, ~d sig- Occurnng in one year by P, the P =(2 – l/n). The prob-
nificant rainfall might not occur more often than O.5% of ability that it will not happen in n years is P“. That is, the
the month. These possibilities should be considered before probability of the event occurring at least once in its return
accepting rate estimates in arid locations. ~riod approaches 63% asymptotically as n increases. Com-
parison of the these curves with estimates of world-wide
extremes using the model presented in the Lenhard and
16.2.1.3 Extreme Rates. As noted in the previous sec- Sissenwine study indicate that, except for tropical moun-
tions, there are many stations in the world where precipi- tainous areas with orographically enhanced rainfall, rainfall
tation”is measured, but there are few outside of the U.S. intensities at Pensacola are comparable to those at some of
and Europe where amounts are tabulated and publiihed for the more extreme locations in the world. This is not sur-
time periods of less than 24 hours. In tropical wgions, where prising, since thunderstornis occurring in tropical maritime
the most intense precipitation &curs, data ?n exwmes of air over the United States are considered as severe as most
short-duration rainfall are even rarer than for the middle places in the world.
latitudes. United States Weather Bureau Technical Paper The most extreme rainfall intensities that have ever been
No. 25 [1955] presents rainfall intensity-duration curves for recorded in the world for time periods of 1 minute to 1 year
200 stations in the contiguous United States plus 1 slation are shown in Figure” 16-12, with h envelope of their ex-
each in Alaska, Hawaii, and Puerto Rico. Lenh*d and tremes. Data used were obtained from Riordan [1974].

Table 16-15. Results of stepwise multiple regression analysis for exceedance levels p = 0.01%, 0.05%, O. 10%, 0.50%, 1.0%, and 2.0% based on a
threshold value of 0.25 mm for I. (The regression coefficients are given for each independentvariable.)

Constant T 1(0.25) f(L,T) SEE

P AP BP c, DP R (mm/min)

0.01 –1.00 2.8 X l&2 4.2 X l&2 –2.2 x ld 0.85 0.41

0.05 –0.56 1.6 X l&2 3.0 x lm* –2.3 X l& 0.88 0.22
0.10 – 0.36 1,1 x l&* 2.4 X l@2 –2.3 X l& 0.88 0.17
0.50 –0.03 2.3 X 10-3 1.0 x l&* –1.2 x lti 0.82 0.08
1.0 0.01 S,6 X l& 6.0 X l@3 –5.6 X l@5 0.77 0.05
2,0 0.03 –2.3 X ld 2.8 X l@3 –2.4 X l&5 0.74 0.02

16-14
 WATER VAPOR, PRECIPITATION, CLOUDS, AND FOG

4/6/~
UN10?4VlLLE , MD
25
I 2460 S0406) 3612N 5 15%3 6 12

MINUTES Hw MY s WTHS

Figure 16-12. World record rainfall values.

1 I I 1 1 I 1 i 1 1 1
5 the table does not indicate how often various rates occur,
5 10 15 30451 2 34 691224
MINUTES WRS but how long various rates continued at that rate or higher
CURATION without dropping to a lower rate.

Figure 16-11. Rainfall intensity-duration-frequency curves for Pensacola, 16.2.1.5 Rainfall Rates Along Lines. Jones and Sims
Florida.
[197 1] present data for 3 rain gage networks, one near
Miami and two in England. Data for a larger network in
16.2.1.4 Duration of Rainfall The question of
Rates. central Illinois are presented by Sims and Jones [1973].
how long a given rate is likely to persist was investigated Average intensities along paths of various lengths, based
by Sims and Jones [1973]. They derived duration tables for on 2-rein rates, were derived for all the networks. The
Miami, Fla.; Majuro Island; Island Beach, N. J.; Franklin, cumulative frequencies of average rates for 2-gage (5-km),
N. C.; and Urbana, Ill. based on an analysis of at least one 8-gage (33-km), and 14-gage (62-km) north-south lines in
year of l-rein rainfall rates at each location. The tables the central Illinois network are shown in Figure 16-13. This
present the conditional frequency distribution of occurrences figure illustrates the effect of line length on the frequency
of various durations from 1 to 90 minutes for rainfall rates of the line-averaged rates. For rates water than 0.2 rnm/min,
from 0.2 to 3.0 mm/min. Table 16-16 presents the results the highest rates occur most often on the shortest line. This
for Miami, Fla. based on 1 year of data. The table indicates, is because larger paths are likely to extend beyond the high
for example, that when it was raining the rate exceeded 0.8 rates at the core of a rain cell into an area of light or no
mm/min, it continued at that rate or greater for at least 4 rain. This situation reverses during light rain (<0.2 mtn/min
min 52.1 Yo of the time. It equaled or exceeded that rate in the figure) since a short path is more likely to be associated
continuously for 25 min only 1.1% of the time. Note that with th~ fringe of a rain ~ell.

Table 16-16. Conditional frequencies of durations of various precipitation rates for about one year of data from Miami, Florida.

Duration Rate Equaled or Exceeded

Equaled or (mm/min)
Exceeded
(rein) ~ ~ ~ ~ Q ~ ~ ~ 2.0 ~ ~
1 100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0 ltio 100.0 100.0
2 85.7 67.1 72.3 77.7 81.9 69.0 62.8 74.1 59.1 45.5 50.0
4 53.3 47.9 43.1 52.1 43.1 29.3 20.9 22.2 22.7 0.0 0.0
6 37.3 27.7 26.2 28.7 19.4 8.6 11.6 11.1 4.5
8 25.4 20.7 20.0 16.0 8.3 6.9 4.7 3.7 4.5
10 19.7 14.9 11.5 11.7 6.9 5.2 4.7 0.0 0.0
15 11.9 8.0 6.9 6.4 5.6 1.7 0.0
20 7.8 4.8 4.6 4.3 0.0 0.0
25 5.7 1.6 0.8 1.1
30 4.1 0.5 0.8 0.0
45 1.2 0.0 0.0
60 0.4
90 0.0

16-15
CHAPTER 16
frequency from 0.25 cm to several centimeters. In Septem-
ber 1970 a hailstone diameter of 14 cm was reported to
have fallen in Coffeyville, Kansas. In hailstorms, the con-
ditional probability of occurrence of maximum-size hail-
stones of 7.5 cm or greater is approximately 0.570, of 5 cm
or greater 2%, of 2.5 cm or greater 7% or 8%. Diameters
of less than 3 mm occur roughly 5% of the time [Gringorten,
1971].
The density (weight per unit volume) of hailstones is a
variable with little doeumentation. For some large natural
hailstones in several storms in the midwestem U. S., esti-
mates ranged from 0.828 to 0.867 g/cm3. Hence, a rounded
value of 0.9 g/cm3 is deemed acceptable in calculations of
impact energy. Such large hailstones have an onion-like
structure, attesting to the up-and-down motion to which they
have been subjected during their growth in severe storms.
Terminal veloeity (w), according to most authors, can
be related directly to the hailstone diameter (d) by

w= Kti. (16.10)

For w in cm/s and d in cm, the value of K at the surface

Figure 16-13. Cumulative frequency distributions of average rainfall rates has been found to range from 1150 to 1990, with a “best”
for 5-, 33-, and 62-km north-south lines in central Illinois.
estimate at 1630.

16.2.2 Hail 16.2.2.2 Hailstorm Characteristics. The ratio of hail-

storms to thunderstorms varies greatly with geography, so
Hail consists of ice particles with a diameter of 5 mm that thunderstorm frequency is no criterion of hail fre-
or more. Smaller particles are classified as ice pellets. Gen- quency. The month of greatest hailstorm frequency also
erally transparent globular ice pellets are called sleet, while varies. In the United States the hail season extends from
generally translucent snow pellets encased in a thin layer April to September, with June and July the most severe
of ice are called graupel or soft hail. If the diameter of the months. Most hail occurs between 1200 and 1900 LST with
ice particles is 2 cm or more, it is called large hail. a peak for damaging hail at 1600 LST.
The potential for damage of hailstones depends on the The width of a swath of hailstone incidence has been
size, density, hardness, number density, impact velocity, variously estimated from a few yards to 120 km, but rarely
and duration of the hailstorm. For a vehicle in flight the more than 8 km. The length of the path is considerably
horizontal extent of the hailstorm also becomes important. greater than the width, but even more difficult to define.
However, it is difficult or impossible to take all factors into The term “hailstreaks” has been coined for hailstorms in the
account. Consequently, for this review the criterion of hail- midwest, which are normally 2 to 25 km long in Illinois.
storm intensity is arbitrarily restricted to maximum stone The median width is 1.5 to 3 km, the length is 10 km in
size. Illinois and 25 km in South Dakota. The areal extent of
Hailstone diameters of 1.9 to 2.5 cm are the minimum these hailstreaks varies from 2 to 2000 km2.
for damage to aircraft, especially in flight when the impact Large hail (greater than 2 cm), is found only in well-
is greater due to airspeed. The metal surface of a transport developed thunderstorms whose cloud tops may reach 15
aircraft at rest on the ground resists damage from hailstones km. Small hail and soft hail are thought by some meteor-
up to 7.5 cm in diameter. Damage to crops reportedly occurs ologists to be an essential feature of all thunderstorms. How-
with hailstones exceeding 3.5 cm in diameter. However, a ever, ice particles less than 1 cm in diameter, which may
great number density of smaller 6 mm stones will damage be present aloft in the cores of the thunderstorms, are likely
crops of wheat, com and soybeans IGringorten, 1971]. Ter- to melt completely before reaching the ground. Although
minal velocity, an important factor in the destructive power, thunderstorms are most frequent in the tropics and subtrop-
is related directly to the size of the hailstones. ics, occurring up to 180 days per year in several places,
hail is rarely found on the ground in the tropics. Hail on
16.2.2.1 Hailstone Characteristics. Although hailstone the ground is most likely found in midlatitude mountainous
diameters are usually less than 1.3 cm, there is a distribution and adjacent areas.

16-16
 WATER VAPOR, PRECIPITATION, CLOUDS, AND FOG
16.2.2.3 Frequency and Duration at the Ground. The Table 16-17. Diameter of the maximum-size hailstone in pried of I to
25 years in the most severe location
most severe hail activity in the world occurs in the United
States, and the most significant center within the United Period Hailstone Diameter (mm)
States is in western Nebraska and southeastern Wyoming. of
Although the greatest number of hailstorm days is at Chey- Years Average 10% Extreme
enne, Wyoming, averaging 9.4 days per season, it is so
localized that an average of 7 days per year may be con- 1 25 55
sidered an acceptable extreme. By comptiison, in some 2 35 66
areas thunderstorms average 40 to 50 days per year. In the 5 53 76
most severe month and in the most severe location, the 10 63 89
average number of hailstorms is 2.9 per month. However, 25 76 102
a more acceptable number for our use, especially for ap-
plication to flight paths, is 2 hailstorms per month.
Using the average of 7 hailstorms per year in the Poisson
month, the single-point probability of encountering a hail-
distribution for the frequencies of zero, one, and two hail-
storm becomes 0.000448. The probability of encounter in
storms per year, together with the conditional frequency
terms of hailstone size is shown in Table 16-18.
distribution of the maximum hailstone size in a storm (Sec-
tion 16.2.2. 1), we obtain the cumulative probability of the
annual largest hailstone diameter in the most severe location
16.2.2.4 Hail Alofi. Nearly every thunderstorm has some
shown in Figure 16-14. Application of the Gumbel distri-
hail, snow, or ice pellets at one stage or another of its
bution further leads to estimates of the greatest diameters
development, particularly at levels between 6 and 8 km. At
in periods of 1 to 25 years (Table 16-17).
any given time, however, hail is considered to occur in a
The duration of hailstorms has been reported variously
shallow layer. Above 9 km hail has been encountered in-
from a few minutes to a few hours. The best estimate for
frequently, but cannot be discounted.
the average time, at the point of most severe occurrence
In the absence of sufficient and objective data, a certain
and in the most severe month, is 10 min IChangnon, 1970].
amount of inference must be used in considering the fre-
When this is coupled with the average of 2 hailstorms per
quency of hail as a function of altitude. Hailstones must
form and grow in size above the freezing level in the at-
mosphere. Over northeastern Colorado, western Nebraska,
or southeastern Wyoming, the average height of the 273 K
isotherm is 4.1 km during the hail season (summer). Once
formed, the hailstones will fall, be buffeted vertically up
.995 and down, or become suspended at a “balance” level. This
should become a level of high concentration and, therefore,
-a .66 ● r
a level with a relatively high probability of hail occurrence.
u 4— . .* E Balance levels may be either at or below the level of updraft
5 ~ maximum, but above the freezing level.
z
$ 3— —.9s; With our present knowledge. it is expedient to assume
n
~ .9
: 2—
Table 16-18. Estimates of the probabilityy of encountering hailstones of
2 — .8
U given diameter at a single-point location from ground to
a l— — .7 1500 km in the most severe Iwation in the worst month.
----- .
Hail Diameter Single-station
o— -
(cm) Probabilityy
.1
-1 — : Any size 0.448 X 10-’
* 0.6 0,354 x 10-3
1
-2—’ = 1.3 0.161 X l&3
o ,2.5 5.0 7.5 10.0 12.5 0.605 X 10-4
> 1.9
~ HAILSTONE OIAMETER ( cm.), h s 2.5 0.314 x 10-4
a 5.1 0.851 X 10-5
Figure 16-14. Cumulative probability of the annual largest hailstone di- z 7.6 0.170 x 10-’
ameter in the most severe location, plotted on extreme =10.2 0.25 X 10-b
probability paper.

16-17
CHAPTER 16
that the probability of encountering hail is uniformly the Table 16-20. Estimates of the probability of occurrence of a hailstorm in
an area (A), in the worst area during the worst period, when
same at any level from 3 to 6 km, and that any level in this
2 hailstorms per month are expected at any single pint.
interval can become a level of hail concentration. Concom-
itantly, it is assumed that the probability of hail encounter Area
decreases from 3 downward to 1.5 km and also from 6 A Probability
upward to 14 km. Since hailstones do not form or grow at (km’) (in %)
the 1.5 km level, and since hailstones do not melt appre-
ciably when falling from that level to the surface, the prob- o 0.045
ability of hail encounter at levels at and below 1.5 km is 26 0.06
approximately the same as that found at the surface 260 0.17
(0.000448). 2600 0.58
For the significantly large hailstones (>2 cm diameter), 26000 3.1
it has been found that their diameters remain virtually un- 260000 29
changed when falling from aloft to the ground. As a result,
the conditional probability distribution of hailstone diame-
ters aloft can be adequately estimated by the conditional
16.2.2.5 Area and Line Coverage. Ten years of obser-
probability distribution at the surface.
vations at a network of some 50 stations in a 400 kmz area
Estimates of the ratio of probabilities of hail aloft to hail
around Denver [Beckwith, 1960] gave the average ratio of
at a lower level are available for thunderstorms in Ohio and
occurrence of hailstorm days in that area to the oecumence
Florida. In these storms hail was encountered about 7 times
at a single station as 4.4 to 1. Thus, if the probability of
more often bet ween 3 and 6 km than at 1.5 km. These data
occurrence at a single station is 0.000448, then the prob-
also indicated that at 7.6 km the probability of encountering
ability of one or more occurrences in a 400 km2 area would
hail is about 3 times as great as at 1.5 km. Above 7.6 km
be 0.00197. A subsequent investigation in Illinois and South
the probability of encountering hailstones diminishes stead-
Dakota [Changnon, 1971] has led to the conclusion that the
ily to nearly zero at 13.7 km (Table 16-19).
area-to-point ratio of hail-day frequencies is related to the
The probability of hail occurring at the ground increases
area size, but is reasonably independent of geography.
with increasing height of the radar-echo tops of the asso-
With the above information a recent model of area]
ciated thunderstorm. For example, hail has been reported
coverage [Gringorten, 1979] has yielded an estimate of the
on the ground for about half of the New England thunder-
correlation coefficient of 0.99 for hail events at points 1.3
storms when radar-echo tops have exceeded 15 km. In ad-
km apart. With this estimate the probability of a hailstorm
dition, hailstoms in the southern and midwestem U.S. occurrence somewhere in an area (A) as a function of the
sometimes have echo tops up to 18 and 21 km [Kantor and
size of that area becomes as shown in Table 16-20. The
Grantham, 1968]. These extremely high echo tops extend
model also has yielded the probability of encountering hail
well above the cirrus anvil and penetrate several kilometers
along a flight path (at an altitude where the single-point
into the stratosphere. Because these giant storms must be
probability is 0.00314), as 0.025 in a path length of 160
accompanied by exceedingly high vertical velocities, hail
km, and 0.042 in a path of 320 km. The probability of
above the tropopause is possible.
encountering hailstones of at least 2.5 cm in diameter is
smaller by 7/100, becoming 0.0018 for 160 km and 0.0029
for 320 km.
Table 16-19. Estimates of the probability of encountering hail of any size
at a single-point-location by altitude
16.2.2.6 Hail in Design and Operation of Equipment.
Altitude Since the probability of hail of even the smallest size is very
(km) Probability slight between the ground and 1500 m and since the effect
of hail on most surface equipment will not result in endan-
Ground level 0.000448
gering human life, there is no need to specify a hail-size
1.5 0.000448
extreme as an operational design criterion. If hail interferes
3.0 0.00314
with a surface operation, postponement until the storm ends
4.6 0.00314
is advisable.
6.1 0.00314
Aloft (between 3 and 6 km), probabilities of hail are as
7.6 0.00134
high as 0.00314, which is still sufficiently low so that there
9.1 0.00100
is no need to specify a hail-size extreme for equipment
10.7 0.00067
whose failure would not endanger human life. If such danger
12.2 0.00034
exists, however, the estimates provided in this section should
13.7 0.000
be considered.

16-18
 WATER VAPOR, PRECIPITATION, CLOUDS, AND FOG
16.2.3 Snowfall 75 cmof snow has been recorded in a24-hperiod in 23
states in the U.S. including locations as far south as Arizona,
Snow is important to the design of equipment and struc- New Mexico, North Carolina, and Virginia. The greatest
tures that must bear an increased structural load imposed Canadian 24-hour snowfall was 112 cm at Livingston, Al-
by snow accumulations. It can also im~de movement over berta. North America’s greatest single storm accumulation
r,mcleared ground. In areas where the mean winter ternper- of 480 cm occurred at Mt. Shasta Ski Bowl in the California
atures are near or below freezing, the maximum depth of mountains [Riordan, 1974]. The greatest measured snow
snow during the season is usually the accumulation from depth in the U.S. was 1145 cm at Tamarack, California.
many individual snows. In warmer climates extremely heavy Snow loads on structures are not routinely measured and
snowstorms can occur, but these are Usually followed by must be estimated. This requires information not only on
melting and amounts on the ground at any one time are not the depth of snow, but also its density. Even then, the shape
as great as in colder areas farther north. The most favored and type of structure, and its wind exposure must be taken
locations for heavy snowstorms and excessive snow depths into account. The density of freshly fallen snow is most
are mountain locations, especially those in proximity to commonly between 0.05 and 0.15 g/cm3. Consequently an
large bodies of water. accumulation of 76 cm with a “normal” density of 0.1 g/cm3
A survey on North American snowfall extremes through would result in a snow load on the ground of about 76
the 1969-70 season by Ludlum [1970] indicates that each kg/m2. Depths in much of North America can greatly exceed
state in the U. S., except Florida and Hawaii, has recorded 76 cm, and rain on top of the snow significantly increases
24-h snowfalls of at least 45 cm. North America’s heaviest its density. Boyd [1961] analyzed snowfall statistics for
24-h accumulation of 193 cm occurred at Silver Lake, Col- more than 200 Canadian stations. His analysis of the max-
orado, a mountain location (elevation 3115 m). A study of imum depth of snow is shown in Figure 16-15 for a 30-
intense snowfalls by Riordan [1973] indicates that at least year return period, the time interval in which the indicated

Cplsl
1
MAXIMUM ~PTH OF SN~
M - YEAR RETURNkRIOO
cm.

Figure 16-15. Maximum depth of snow for a 30-year return period for Canada expressed in metric units IBoyd, 1961]

16-19
CHAPTER 16

A 200

=’--’%’-’%[:“h’”
Kg/m2

1 w — I

Figure 16-16. Estimated maximum snowload in Canada in a 30-year return period expressed in metric units IBoyd, 19611

maximum depth would be expected to occur (a 63% chance) the burden of accumulated ice. Nevertheless, it remains one
at least once. He also estimated the amount of rain that of the few meteorological conditions not quantitatively ob-
could occur and be retained by the snow. These findings served on a routine basis by any national weather service.
are incorporated into Figure 16-16, which indicates that a There are three basic types of ice formed by accretion
maximum snow load on the ground of 600 kg/m2 can occur in the atmosphere: glaze, hard rime, and soft rime. Glaze
in non-mountainous areas in eastern Canada. He tentatively occurs most commonly when rain or drizzle freezes on ob-
indicates that snow loads on flat or low-slope roofs would jects. It is clear and nearly as dense as pure ice (which has
be about 80% of the ground load based on limited obser- a density of 0.917 g/ems). Hard rime is less transparent than
vations. In a 1951 U.S. Weather Bureau study, the maxi- glaze, and at times is opaque, depending on the quantity of
mum non-mountainous ground snow load in the U.S. is air trapped in the ice. Its density varies from about 0.6-0.9
estimated at 300 kg/m2, and the 1 year in 10 maximum at g/cmJ. Soft rime, which is white and opaque, is feathery
200 kg/m2. Snowfall extremes in North America are con- or granular in appearance, with a density less than 0.6 g/cmq.
sidered to be as severe as anywhere in the world, so that Rime ice, both hard and soft, is most often formed by the
these values are considered representative of non-moun- freezing of supercooled cloud or fog droplets. For this reason
tainous worldwide extremes. it is a localized phenomenon, occurring most frequently in
mountainous areas or hilltops exposed to low, stratus type
clouds, or on land areas adjacent to bodies of water.
16.2.4 Ice Accretion on Surface Structures Most of the information available on surface icing is for
glaze; it is usually recorded because of extensive human
Ice accretion, or icing, refers to ice accumulating on distress or a large amount of damage. Even then, measure-
stationary objects located near the earth’s surface. Such ice ments have been subjective, non-standardized, and difficult
can disrupt transportation, cause power and communication to interpret. The most comprehensive study on the geo-
outages, and result in severe damage to structures bearing graphical distribution and frequency of glaze. worldwide.

16-20
 WATER VAPOR, PRECIPITATION, CLOUDS, AND FOG

REGION ~ NORTHEAST

MAX

.9!5l--

.s0

t
Figure 16-17. Number of times ice at least 1.25 cm thick was observed .20
during the 9-year period of the Association of American t
Railroads Study [Bennett, 1959].

was done by Bennett [1959]. The information presented

indicates that glaze is as frequent and as severe an occur-
rence in parts of North America as it is anywhere in the
world. Figure 16- i 7 from that report shows the number of
times that ice at least 1.25 cm in thickness was observed
in the U.S. during a 9-yr period in a study by the Association
of American Railroads. Although the length of record is
short for climatological purposes, the relatively high fre- r(cm)
quency of heavy ice storms in the central and northeastern
U.S. is supported by other data. Figure 16-19. Estimated probability of ice thickness, r, occurring at least
Tattelman and Gringorten [1973] reviewed many sources once in ahy year at a representative point in region 111
(AVG), and in the most severe part of region 111(MAX)
for information on ice storms in the contiguous U.S. for the
ITattelman and Gringorten, 1973].
50 winter seasons from 1919-20 through 1968-69. The data
were analyzed to determine the probability of occurrence
of ice thicknesses in each of eight regions in the U .S .“The of occurrence in a.year for the nopheast (region 111in Figure
regions shown in Figure 16-18 are areas with similar glaze 16- 18), which had the greatest frequency of ice storms. The
characteristics subjectively determined on the basis of lat- curve labelled AVG represents an average for all states in
itude, geography, climatology and the distribution of ice the region, whereas the curve labelled MAX represents the
storms. Figure 16-19 shows ice thickness versus probability estimate for the state in the region with the highest frequency
of ice storms. The authors stress the subjectivity involved
in arriving at their estimates.
Eastern Canada, from Nova Scotia and Newfoundland
northward to southern Baffin Island, has the highest fre-
quency of freezing rain and drizzle in North America [McKay
and Thompson, 1969]. Extreme ice storms in this area have
produced accumulations of ice 20 to 40 cm in thickness.
This is not surprising because the area is frequently subject
to the passage of slow moving “mature” cyclones, and as-
sociated winds can exceed hurricane force (33 mls). Strong
winds increase the number of precipitation drops impinging
on a vertic~l surface, so that the accumulation of ice can
exceed the depth of precipitant ion by a factor of 3 or more.
Ice accretion is difficult to measure because the shape
u and size of the collecting surface, and its orientation to the
wind can result in large differences. Furthe~ore, the con-
Figure 16-18 Regions of similar glaze characteristics ITattelman and figuration and density of ice on a surface are dependent
Gringorten, 1973]. upon a number of atmospheric variables. Tattelman [1982]

16-21
CHAPTER 16
Table 16-21. Mean and standard deviation of the ratios of the mass of 105
l\ I
ice on the 3-, 13-, and 50-mm diameter cylinders to that
on the 25 mm diaineter cylinder.
7

I
E
*E ,.3 ,----

Cylinder Diameter Standard ‘-

,,,

Category (mm) Mean Deviation

In-cloud icing 3.2 0.21 0.059

In-cloud icing 12.7 0.59 0.062 \
x’
In-cloud icing 50.8 1.47 0.116 %
~ 01 -
z
Freezing rain 3.2 0.18 O.ql 001 I 1 I I I
I 2 3 4 5 6
Freezing rain 12.7 0.54 0.056 o
DROP DIAMETER (mm)
Freezing rain 50.8 1.6.5 0.091

Both combined 3.2 0.20 0.054 Figure 16-20. Four typical sha~s of raindrop size distributions measured
at the ground.
Both combined 12.7 0.57 0.063
Both combined 50.8 1.56 0.139
Examples of some drop size distributions are shown in
Figure 16-20. When d = O and ~ = 1, &uation (16. 11)
gives an overview on the dynamics of ice accretion and reduces to the simple exponential distribution first proposed
research in the field. He also presents an analysis of icing by Marshall and Palmer [1948]. This is shown in the solid
measurements on cylinders, for differing--- synoptic condi- line of Figure 16-20. This distribution was derived from
tions, made during ~limatic chamber tests of an ice n?eas- data obtaitied in stratiform rain, and tends to overestimate
urement system. Results include a comparison of the mass the number of small drops. When cs = 2 and ~ = 1, the
of ice that accu~ulates on cylinders which tie 3, 13, 25, resulting expression is an approximation to the log-normal
and 50 mm in diameter (Table 16-21). distribution proposed by Levin [1954]. In this case, there
is a realistic decline in the number of drops at the lower
end of the spectrum, Setting a = – 2 and @ = 2 results
in a gamma distribution first derived by Best [1950]. This
16.2.5 Distributions of distribution holds if, as some measurements indicate, the
Precipitation Elements liquid water content per drop size interval follows an ex-
ponential distribution. The fourth distribution is one pro-
Precipitation paramet~rs vary appreciably with type of
posed by Litvinov [1956] for drizzle-~ rains, setting a = O
storm, geographic location, and even from storm to storm
and ~ = 3/2. Although all the distributions in Figure
of the same type and in the same geographic region; for this
16-20 have been extended to 5-mm diameter drops (the usual
reason, no model storms are presented in this section. How-
maximum drop size observed), the distribution for drizzle-
ever, individual profiles or averages that are derived from
type rains should be truncated at approximately 2-mm di-
observations at several locations are given, and, wherever
ameter.
possible, the applicability and representativeness of the data
The total liquid water content M, the Rayleigh radar
are indicated. Great care must be taken in extrapolating the
reflectivity factor Z, and the median volume diameter Do
results to geographical regions that are characterized by a
can be expressed as weighted integrals of the size distri-
climatology which differs from that of the region from which
bution. The weighting factor in deriving Z is D6, while that
the data were obtained.
for deriving M is ti/6 pD3, where p is the density of water
in g/cm3. Assuming an exponential distribution ( u = O,
16.2.5.1 Raindrop Size Distributions. Numerous equa-
P = 1), and integrating between D = O and D = W (an
tions have been proposed to express the size distribution of
approximation sufficiently accurate for most purposes), we
raindrops measured at the ground as a function of rainfall
obtain
rate. They can all be expressed in the fom
DO = 3.67/ A (mm) (16.12)
ND AD = NOD “exp( – AD ‘), (16.11)
M = 10-‘ m(pNJA4) (g/m3) (16.13)

where ND AD is the number of drops per unit volume with Z = 720 NJA7 (mmG/m3). (16.14)
diameters between D and D + AD, N. is the value of ND
where the curve crosses the D = O axis, Ais a parameter Other parameters (for example, the median power di-
that depends on the type and intensity of the precipitation, ameter) can also be found using the appropriate weighting
and CYand ~ are parameters which determine the shape of factor. Such i~tegrations, however, become cumbersome
the distribution. for distributions in which a # Oand ~ # 1. For this reason,

16-22
 a
WATER VAPOR, PRECIPITATION, CLOUDS, AND FOG
I 00

Figure 16-21. Radar reflectivity factor vs rainfall rate (shaded area is the
range of the measured relations). 0. I
0.01 0. I I 10
LIQUID WATER CONTENT(gm-3)
the simple exponential distribution is often used, even when
it is known to deviate from measured size distributions.
Numerous empirical relations have been proposed ex- Figure 16-23. Liquid water content of precipitation vs rainfall rate.

pressing radar reflectivity as a function of rainfall rate or

liquid water content, or rainfall rate as a function of liquid
ND AD = No exp ( – AD) (16.15)
water content. The relations differ significantly, according
to the location and type of precipitation. Figures 16-21, where NO = 3.8 x 103 R–087(m-3mm-’) (16.16)
16-22, and 16-23 summarize the results of measurements
from several observers. The shaded area of Figure 16-21 A = 2.55 R-048 (mm- ‘), (16.17)
encompasses the range of measured relations between radar
reflectivity and rainfall rates, and illustrates the variability where R is precipitation rate in millimeters of water per
that can be encountered. Figure 16-22 expresses the same hour. For snowflakes, the diameter D refers to the melted
data in terms of radar reflectivity versus liquid water content. spherical diameter of the snowflake.
If rainfall rate is the measured quantity, Figure 16-23 can Recent measurements [Lo and Passarelli, 1982] of the
be used to find the expected range of liquid water content. evolution of snow size spectra with decreasing altitude dem-
onstrate the existence of three growth regimes. At the top
16.2:5.2 Snowflakes. Gunn and Marshall [1958] found of the snow region, growth is by deposition only. The size
that an exponential law similar to that shown by the solid distribution is typified by curve 1 in Figure 16-24. Pro-
line of Figure 16-20 was applicable to the size distribution gressing down through the atmosphere, N,, increases while
of aggregate snowflakes. The spectral parameters in snow A remains constant. Then aggregation begins, and there is
are related to the precipitation rate by a rapid decrease in both No and A. Curve 2 of Figure

10 I 1 1 I

10-’

I
10<

163

IO-4

10-~ 1 I 1 I
10’ 102 103 104 105 Id o. I 0.2 0.3 0.4
RADAR REFLECTIVITY FACTOR, Z ( mm6 m-3) Particle Diameter (cm )

Figure 16-22. Radar reflectivity factor vs liquid water content of rain Figure 16-24. Snow size spectra as a function of growth regime and
(shaded area is the range of observed relations). altitude.

16-23
CHAPTER 16
16-24 illustrates a typical spectrum during this phase. Fi- Height variations of gross parameters such as water con-
nally, breakup counteracts the effects of both deposition and tent or radar reflectivity factor are functions of type of storm
aggregation, and the distribution remains constant, with a and geographic location. However, when there is a sufficient
slope and intercept typical of curve 3 in Figure 16-24. The body of data for a particular region, it is possible to define
minimum value of A observed in the atmosphere is ap- a climatological model. Two such models are shown in
proximately 10 cm-’. Figures 16-25 and 16-26. Figure 16-25 shows the results
for stratifom storms at Wallops Island, Virginia. The dotted
16.2,5.3 Distribution of Precipitation Elements with line may be considered an average vertical profile of liqiud
Height. The vertical distribution of precipitation param- water content for stratiform storms at this location, and the
eters can be inferred from the vertical distribution of radar solid lines indicate the variability of the data. Figure 16-26
reflectivityy, using the appropriate conversions. When the shows the same analysis for convective storms at Wallops
precipitation particles are ice crystals or snowflakes, it is Island. The major difference between these two figures is
necessary to determine the type of particle and general size the high water content between 2 and 4 km in convective
regime before choosing the equation to convert the particle storms. This storage zone aloft is characteristic of most
size or mass to the equivalent diameter or mass of a liquid convective storm situations, regardless of location. Strati-
drop. Table 16-22 shows factors for converting from a phys- form storms, on the other hand, are characterized by rela-
ical length, L, of the solid particle to an equivalent spherical tively smooth vertical profiles of liquid water content, with
diameter, D. A power relation of the form D = a Lb has only a slight bulge at the melting zone.
been assumed, and a and b are listed for general particle
types.
16.2.5.4 Extreme Valu@ of Liquid Water Content.
The maximum size of hydrometers is related to the
Measurements of the maximum concentration of liquid water
slope of the distribution by the empirical relation
in severe convective storms have not been made on a sys-
tematic basis. Isolated reports of concentrations as high as
D.A=C, (16.18)
30 g/m3 ISulakvelidze et al., 1967] and even 44 g/m3 IRoys
and Kessler, 1966] are found in the literature. There have
where D~ is the maximum equivalent diameter, Ais the
been occasions when investigators have suspected the oc-
slope of the exponential distribution, and C is a constant
currence of abnormally high concentrations but lacked the
dependent on the type of hydrometer. Using data collected
over several years, AFGL scientists have found that average
values of C range from 12 for ice crystals to 10.5 and 9 for 12
small and large snow, respectively, and 7.5 in rain.
If an exponential distribution is assumed, the particle II

size distribution of hydrometers can be completely de-

10
scribed from the slope of the distribution (A), the intercept
(No), and the total number of drops (N,), which depends — STORM TOP
9.,
L ..
upon the minimum and maximum particle size assumed.
ICE REGt~
The size is dependent upon the slope of the distribution as
shown in Equation (16. 18). ICE TO SMALL

FF
=7 j I SNOW TRANSITION
Table 16-23 gives the characteristics of average distri- 1~E
butions for various hydrometer types. In each case, N,, No : ...
.. sM*L!gow
U6
and A are expressed as power functions of M, the total Q
.. ,SMALL TO LARGE
water content in g/m3. The units of A are lmm”’; of No, g
55 i __ ——.— ..— ,S~W TRANSITION
m-smm-’; and of N!, m-3. Expressions connecting other sets a{
..
— ZONE

of parameters may be obtained from this table and manip- 4

r “...,

2*
., LA~EES;OW
ulation of Equations (16. 12) through (16. 14), or integration ..
of Equation (16. 11) with the appropriate weighting factor. 3 ...
,., ” / ~E~:lNG
z .,. .
,..’ I
Table 16-22. Factors for converting from physical length, L, to equivalent ..
diameter, D. .. RAIN REGION

Particle type Size range a b OUL

0.1 0.2 0.3 0.4
~_LL –.L-L.-~
0.5 0.6 0.7 0.8 08 1.0 1.1 1.2
M(g rri3)
Ice crystals <0.2mm 0.438 1.000
>0.2mm 0.256 0.667 Figure 16-25. Liquid water content in grams @r cubIc meter vs altitudes
Small snow <0.5mm 0.370 0.670 in kilometers for stratified-storm situations (dotted line is
Large snow >0.5mm the mean and the solid lines represent plus and minus 1
0.400 0.782
standard deviation).

16-24
 d

WATER VAPOR, PRECIPITATION, CLOUDS, AND FOG

Table 16-23. Distribution parameters as a function of hydrometer type and of total mass, M (N, is the total number of particles per cubic meter; N,, and
A are
.. -.. the
. . 7e;o interceDt . for an exwnential
-. and slooe size distribution).

Hydrometer
Temp. Range
Min. Diameter N,=e Mf No=i Mj i=k M~”

Sub Types e f i j k q

[ce Crystal
<248 K
d = 0.055 mm
Tropical 61522 0.073 514541 –0.583 6.343 –0.394
Temperate 98719 0.303 1104536 –0.201 7.675 -0.299

Small Snow
258 to 243 K
d = 0.065 mm
Tropical 50723 0.032 403471 –0.739 5.968 –0.431
Temperate 19727 0.266 109OO2 –0.148 4.295 -0.287

Large Snow
273 to 253 K
d = 0.065 mm
Tropical 4122 –0.373 10839 – 1.089 2.403 -0.522
Temperate
Agg. of Plates 6833 0.384 23966 0.096 2.930 –0.226
Agg. of Stellars 1649 0.155 3251 –0.202 1.778 –0.301
& Dendrites
Agg. of Columns 14005 0.455 67147 0.164 3.791 –0.209
& Bullets

Rain
>273 K
d = O.lmm
Tropical 4033 0.443 12619 0.161 2.471 –0.210
Temperate
Widespread 1934 0.133 4382 0.018 1.897 –0.245
Drizzle 5850 0.366 21509 0.017 2.824 –0.246
Thunderstorm 649.9 0.310 945.6 0.017 1.293 –0.245

equipment for measuring them. In considering many aspects It is dangerous to extrapolate from such a small sample.
of the storm. Revs and Kessler could not find evidence Nevertheless, the actual occurrence of extreme values of
which indicated that the measured concentration 44 g/m3 liquid water concentrations probably follows the general
should be either rejected or accepted. Omitting the value of shape of Figure 16-27 with possible modification of the
44 g/m3, their 26 other measurements of the maximum liquid constants in the distribution function. Figure 16-27 indicates
water content in Oklahoma thunderstorms fit very well the that 75~0 of Oklahoma thunderstorms have maximum liquid
distribution illustrated in Figure 16-27. According to this water contents exceeding 4.3 g/m3, and that liquid water
distribution, the probability that the maximum water con- contents in excess of 9.4 g/m3 occur in only 25% of the
tent, M, in an Oklahoma thunderstorm will exceed a given storms. Substitution into Equation (16. 19) shows the values
value, x, is of 30 g/m3 or higher me literally “one in a million” occur-
rences in Oklahoma thunderstorms, and the occurrence of
P(M > x) = exp ( – x2/64), (16.19), a vahte of 44 g/m3 has a probability of 10-12.
It is of interest to compare these extreme values of water
where M and x are in g/cm3. content to the water content corresponding to a record rain-

16-25
 CHAPTER 16
12r 0.01 I I
I
II
0.I
t
10
k STORMTOP

:L
ICE REGION
●
5
ICE TO SMALL
)W TRANSITION 10
‘ ZONE
SMALL SNW m
REGION

40

60

2 ““”’”””””
.................. 80

L-
4 :
LA= w
... ... REGION 90
3 ..”
‘ “. 0. ......... ● 00 ROyS a KESSLER( 1966y
, . . . . . ...””
2 ... . . ...””””” ,:&;lNG
,.-.””
. . .. . 99

H)
Ou
1;

I I I I I I 1
RAtN

I
REGION

I J
99.9
r
O 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.6 0.9 I.O I.1 1.2
M(g m-’)

~x=MAXIMUM WATER CONTENT ( gin-3)

Pigure 16-26. L]quid water content in grams per cubic meter vs altitudes
in kilometers for convective-storm situations (dotted line
is the mean and the solid lines represent plus and minus I Figure 16-27. Probability of maximum liquid water content witbin Okla-
standard deviations). homa thunderstorms.

fall rate. Over a 1-rein interval, the world record rainfall Secondarily, visual estimates of amount and height of
amount is 3.1 cm; this fell at Unionville, Maryland, in 1956 bases of individual cloud types and layers are reported. But
{Riordan, 1974]. Assuming a Marshall-Palmer distribution, because low clouds obscure higher clouds, it is often im-
this rainfall rate corresponds to a liquid water content of possible to obtain accurate information on the distribution
approximately 55 g/m3. Consequently, extremely large val- of clouds at the higher levels. Studies of the accuracy of
ues of water content (for example, greater than 30 g/m3) visual cloud observations show that the greatest errors occur
may occur on rare occasions either at the surface or aloft in estimating cloud amounts when the sky is 3/10 to 7/10
in severe thunderstorms. covered. These errors however, tend to average out over
long periods. The average error in visual estimations of
cloud heights ranges from 300 m for clouds at 750 m, to
1500 m for clouds with bases near 7 km. At many points,
16.3 CLOUDS however, particularly at airport stations, ceiling heights can
be measured with a higher degree of accuracy with ceilo-
meters, clinometers, or balloons.
16.3.1 Surface Observations Observed values for a particular time give amount of
sky covered but not the cloud distribution. For example, if
Cloud observations taken regularly by ground observers 5/10 cloud cover is reported, one large cloud formation may
at weather stations throughout the world primarily contain cover half of the sky or small individual cloud cells may
the following information: be equally distributed over the sky. There is no direct way
(1) Visual estimate of total amount of sky covered by to obtain the geometry of clouds in the sky from the reported
clouds, standard observations.
(2) Cloud ceiling, which is the estimated or measured There can be considerable variation in the frequency of
height above ground of the lowest layer of clouds that cover occurrence of given cloud amounts and ceiling heights be-
more than half of the sky. tween stations located within several kilometers of each

16-26
 H

WATER VAPOR, PR~CIPITA~ION, CLOUDS, AND FOG

other. Because local topographical effects are the primary Table 16-24. Sample Wrcent frequencies of sky cover, by tenths.
(example is for Bedford, Mass).
cause of such variations, caution must be used in applying
clirnatological data on clouds of one or more stations to Hours January Sky Cover (tenths) Avg.
Sky
an entire region, particularly in mountainous and coastal
Cover
regions. o 123456789t0 (tenths)

00-02 3423321232246 5.5

16;3.1. 1 Summaries of Surface Observations. Tabu- 03-05 3223322333245 5.7
lations Of surface cloud observations me available for sta- 06-08 2154432235447 6.3
tions in most areas of the world. A Guide to Statiard 09–1 1 i755442236448 6.5
Weather Summaries and Climatic Semites [NOC, 1980] 12-14 1645443346S46 6.6
has been prepared by the Naval Oceanography Command 15–17 1665533345545 6.4
in Asheville, North Carolina, which lists some 23 types of 18–20 2835432334243 5.7
cloud summaries. The station list includes locations in var- 21–23 3234421334242 5.5
ious countries, continents, islands, and oceans. There are
close to 10@ weather stations hound the world> although Mid-
not all of therh @vide routine observations of clouds. Among season Seasonal Sky Cover
the better summaries of cloud information are those prepared Month
by the United States Air Force Environmental and Technical
Jan 2444432344345 6.0
Applications Center (USAF/ETAC) as Revised Uniform
Apr 184454334744 6.4
Summaries of Su~ace Weather Observations (RUSSWO).
Ju1 1556765569531 5.9
Similar summaries are issued by the Naval Weather Service
Ott 3055543345432 5.0
as Summary of Meteorological Observations, Surface
(SMOS), and Summary of Synoptic Meteorological Obser-
vations (SSMO).
Table 16-24 is a sample of the frequencies (in tenths) In the RUSSWO, as in many summaries, the frequencies
ofs~ cover, from clear (0/ 1O)to overcast ( 10/10) at a single are given for ceiling heights jointly with the visibility in
station (Bedford, Mass. ) in a single month (January) for matrix form. Table 16-25 is a sample of ceiling height
eight 3-h periods of the day. It also shows the average sky information as gleaned ftom eight such matrices for Bed-
cover, but does not subdivide the cloud cover by height Of ford, Mass., one for each 3-hour period. Table 16-25 also
clouds. Table 16-24 also shows the frequencies for the mid- shows the cumulative percent frequencies of ceilings for the
season months for all hours combined. mid-season months for all hours combined.

Table 16-25. Sample percent frequencies of ceiling heights equal to or higher than, including ceiling unlimited (example is for Bedford, Mass)

January Ceiling (m)

Hour 61W 4900 3700 3000 2400 1800 1200 900 600 300 150 30

00-02 55 55 58 60 62 66 72 76 80 86 92 98.4
03-05 51 52 55 57 60 63 70 74 80 86 93 98.3
06-08 48 48 53 56 60 63 70 74 79 86 94 98.9
09–1 1 50 51 56 59 63 66 71 75 80 86 94 99.8
12–14 50 51 56 60 64 66 70 75 80 86 94 99.4
15-17 52 53 58 61 64 67 72 77 81 86 93 99.2
18–20 56 57 60 62 65 67 7i 76 81 87 94 99.2
21-23 57 57 60 62 64 67 72 77 81 87 94 99.1

Seasonal Ceiling (m)

Mid-
season
month 6100 4900 3700 3000 2400 1800 1200 900 600 300 150 30

Jan 52 53 57 60 63 66 69 76 78 86 94 99.0
Apr 49 50 53 56 60 64 73 77 81 88 95 99.5
Jul 61 62 68 72 75 77 82 85 88 92 97 99.5
Ott 62 62 65 6& 70 72 77 81 84 89 95 98.6

16-27
CHAPTER 16

Figure 16-28. Mean sky cover. January, 1200-1400 LST

16-28
 WATER VAPOR, PRECIPITATION, CLOUDS, AND FOG

Figure 16-29. Mean sky cover, July, 120(-1400 LST.

16-29
CHAPTER 16
16.3.1.2 Global Sky Cover. While there we numerous 16.3.1.3 Modeling of Cloud Distributions. In the fol-
tabulations of sky cover compiled from surface observa- lowing presentations a distinction is made betweeti the terms
tions, there tie relatively few maps. One of the Sources for cloud cover and sky cover. Sky cover is an obsetier’s view
these maps is the chapter “Climqfology” by Landsberg [1945] of cover of the”sky dome, whereas cloud, cover can be used
in the Handbook of Meteorology. The chapter has six world to describe areas that are smaller than the fl~r space of the
maps of sky cover, one for every other month beginning sky dome, or larger.
with Janua~. Figures 16:2? and 16-29 show average clou- Cloud Cover Models. There have been several alter-
diness for 1200-1400 LST in January and July, respectively, native mathematical formulas proposed for the probability
based on Landsberg’s charts; subsequently publisbed tables dis(ri~ptio~ of sky cover. Each uses the variable (x) ranging
of the National Intelligence Suvey, RUSS WO, SMOS and from’ zero, for clear, to 1.0 for overcast. Each m~el is
SSMO; and data collected at the Woods Hole Oceanographic claimed to have versatile statistical characteristics to sim-
Institute for the Atlantic and Indian oceans.’ ulate U-shaped curves of sky cover. That is, large frequen-
Meao sky cover is significantly cotilated with the fre- cies of clear artd/or overcast skies, with small likelihoods
quency of clear and of overcast conditions. Table 16-26 of partial cover, can be fitted; yet the models also will fit
presents a sample of RUSS WO info~ation for some U.S. the bell-shaped distributions in which the sky cover is mostly
stations; it contains the percent frequencies of sky cover, scattered to broken.
from clear to overcast, in January at noontime. Among these Alternative 1: The Beta distribution is an parly cloud
stations the mean sky cover ranges from 4.7 to 8.4 tenths. model [Falls, 1973] whose density function is given by
For low mean ‘sky cover the fre+y of clew skies is high
arid the frequency of overcast low. The opposite is true for r(a + b) xa_,(l _ ~) b-1.
f(x) = ,
a high mean sky cover. In extra-tropical regions the inci- r(a) r(b) (16.20)
dence of 1/10 to 7/10 sky cover remains generally low. O<x<l, a,b>o.
However, the U-shaped distribution is less markd in the
summer months (Table 16-24). In tropical and sub-tropical Falls gives pairs of values of the two parameters in some
regions partial sky cover also is more frequent at the expense 29 regional types that cover the world, for the four mid-
of c]e~ and overcast. This is evident from the frequencies ~ason months, for two times of the day.
for Key West, Florida, and is particularly well illustrated Altertifive 2: More ~ntly, Somerville and B- [ 1979]
by the data from Diego Garcia in the Indian Ocean at latitude have used a model called the S-distribution in which the
7° south (Table ?6-26). cumulative probability F(x) of sky cover (x) is ‘estimated by

Table 16-26. Pereent frequencies of sky cover, by tenths, at a representativesample of stations. (For uniformity, the cases are all for January 1200 LST. )
. .
Sky Cover (tenths) Mean
Sky
Cover
Station o I 2 3 4 5 6 7 8 9 10 (tenths)

Extra-tropical

phoenix, Ariz. 33 6 4 43345 7 5 26 4.7

Washington, D.C.
(Dunes Airport) 22 8 6 33323 6 4 40 5.7
Bridgeport, Corm. 21 3 4 54235 6 6 41 6.2
Austin, Texas 18 5 4 33335 6 644 6.4
Chicago, 111. 14 4 4 33325 6 6 50 6.9
Great Falls, Mont. 11 4 3 33235 8 8 50 7.3
Boise, Idaho 11 4 3 32224 5 6 58 7.5
Niagara Falls, N.Y. 333 33324 9 8 59 8.2
Seattle, Wash. 52 2 2 122 4 9 764 8.4

Sub-tropical
and tropical

Key West, Florida 14 9 10 12 8 8 4 7 9 7 12 4.7

Diego Garcia
7018’S, 72”24’E 000 457616 18 14 30 7.8

16-30
 WATER VAPOR, PRECIPITATION, CLOUDS, AND FOG
F(x) = 1 – (1 – xa)~; (16.21) gorten, 1979] which, like the above models, requires two
parameters for a description of the probability distribution
O<xsl, a,p>o.
of cloud cover. The parameters in Model B have physical
meaning. One parameter (Po) is the mean cloud cover as
Pairs of values of the two parameters (cY, ~) have been given in RUSSWO and other climatic summaries; it is taken
determined to make the distribution F(x) fit the data in the to be the single-point probability of a cloud intercept when
sky-cover summaries; they have been published for some looking up from the ground. The second parameter (r),
23 stations around the world, for each of eight periods of known as the scale distance, is the distance between two
the day in each month of the year. The procedure has been stations whose correlation coefficient of cloud covers is
to use 10 climatic cumulative frequencies F(xi) of sky cover 0.99. At Bedford, Mass., it varies from 0.5 km in summer
for xi = 0.05(0. 1)0.95. A first guess is made for a, and ~ noontime to as much as 10 km in winter midnight. Re-
is then estimated by method of least squares: grettably, as developed so far, estimations of the probability
of cloud cover by this model must be made by use of 11
~ = S{tn [1 – F(xi)] “ en[l – x?]} charts as published [Gringorten, 1979]. It does not yield
(16.22)
Z{{n[ 1 – x;]}’ “ estimates succinctly by formula. On the other hand, Model
B avoids the difficulties of the previous two alternatives by
After the pair of values (a, ~) a~e entered into the basic providing finite estimates of all-clear and full overcast for
equation to obtain the estimates F(xi), the sum of squares varying areal extent. As an example, Table 16-27 shows
of the differences between ~(xi) and F(xi) is obtained. The the percent probability of cloud cover at Bedford, Mass.,
best values for the parameters (cY,~) are those that minimize January noontime; the parameter values are Po = 0.66, and
the sum of squares of errors. These were determined by r = 2.6 km. In Table 16-27 the cloud cover is given, not
trial and error, such that the best pair of values for the sky only as a ground observer sees it, but also for a small area
cover at Bedford, Mass., in January at noontime, were found of 100 kmz and for a large area of 100000 kmz. A small
to be a = 0.1468, ~ = 0.1721. These estimates are dif- area is likely to be either all-clear (0/10) or overcast ( 10/ 10),
ferent from those in Table 16-24 by rinse = 0.01. but a large area is nearly always partially covered. For the
In the above-mentioned models, strictly speaking, the ground observer’s sky cover, the estimates differ from those
probabilities of absolutely clear and of complete overcast in Table 16-24 by rinse = 0.02.
are each computed to be zero; hence they are not truly Ceiling Model. The Burr curve provides one of the best
realistic. The difficult y is overcome, in practice, by stipu- models for ceiling height cumulative distributions:
lating that the integration for clear be taken from zero cover
to 0.5 tenths, for overcast from 9.5 tenths to 10 tenths.
There is another difficulty, however, in that each solution F(h) = 1 – {1 + (h/cY}-b ; a,b,c >0, (16.23)
is given for the sky dome as seen by a ground obsemer,
considered to have a radius of 28 km. The above alternatives
have not been developed to give cloud-cover distribution where h is the ceiling height and a,b ,C are parameters.
for varying areal coverage. However, simplicity is in their At the University of Florida [Bean et al., 1979] sets of
favor, especially with the S-distribution, which is readily values for a, b, and c have been determined (for each of
programmed on a desk-top computer. eight periods of the day in each month of the year at some
Alternative 3: A “Model B“ for linear and areal coverage 23 station> around the world) to make the estimated distri-
of a weather element has been described recently [Grin- butions, F(h), fit the data for some 30 ceiling heights as

Table 16-27, An example of model estimates of percent probabilities of cloud cover, by tenths, depending upon size of the area (example is for Bedford.
Mass., January 1200 LST).

Cloud Cover (tenths) Mean

Areal Cloud
Size Cover
(km’) o 12345 6 7 8 9 10 (tenths)

100

2424
29

16
11121

45 33
1

5
2

4
2

6 9
I 59

41
1 6.6

6.6
(Observer’s
sky dome)
100000 0.5 3 4 7 9 9 12 13 16 19 8 6.6

16-31
CHAPTER 16
given in the climatic summaries. In this procedure, a rounded 10
figure is chosen for c, such as 300, 600, or 3000 m. Trial
estimates for a and b are then made to minimize the sum 9
of squares of the differences between the model estimates
8
~(h) and the cumulative frequencies F(h) of the climatic
summaries. An initial trial estimate of a can be made by a 7
linearization of the Burr curve through Taylor expansion,
resulting in 6

(16.24) 4

where xi = (n(hi/c), yi = tn F(hi), and N is the number 2

of ceiling heights for which the climatic tables give the
I
frequencies F(hi). Corresponding to each trial estimate of
a,
0 0. I 0.2 0.3 0.4
1

0.5
1

0.6
1

0.7
1

0.8
t

0.9
I
I.0
[- ~(n {1 - F(hi)} tn {1 + (hi/c)a}l AVERXE CLOW COVER ( tenths)
b=
~{en(l + (hi /C~}’ . (16.25)
Figure 16-30. Average cloud cover by layers, December, 12W1400
LST, Grissom AFB, Indiana.
For Bedford, Mass., January noontime, values found were
a = 1.1678, b = 0.1927 when c = 0.305 km, giving the
put of the model in December, noontime for Grissom AFB,
estimates of probability that differed from those in Table
Indiana. The abscissa is for the cloud cover from O (clear)
16-25 for 1200-1400 hours by rinse = 0.01.
to 1.0 (overcast). The ordinate is for elevation above the
ground, from O (surface) to 10 km aloft. The broken curve
16.3.2 Cloud Cover in Layers Aloft is drawn for the mean cloud cover at levels, and the solid
curves are drawn for the mean cloud cover in layers. The
Nephanalysis, concerning the type, amount, and scatter broken curve shows, for example, mean cloud cover of 0.1
of clouds both in the horizontal and vertical directions, is at 1 km, O. 14 at 2 km, 0.07 at 5 km. From each level the
important, particularly in connection with line-of-sight re- solid curve, when followed to a higher level, gives the
quirements. For example, knowledge of the probability of average cloud cover in the layer between the two levels.
cloud-free lines-of-sight (CFLOS) from one level of the For example, in the layer from 2 km to 4 km, the cloud
atmosphere to another at selected angles of elevation, can cover averages 0.32, from 2 km to 8 km it averages 0.38,
be vital to electro-optical communication systems. Although and so on.
there is now a sizable body of literature on this subject, the In similar manner, the distribution of cloud cover in
statistics are none too satisfactory. Clouds are not only vol- layers aloft can be reconstnscted for any station that has
atile but are often hidden by other clouds, so that an analysis RUSSWO or similar records. Grissom AFB, Indiana, was
of cloud presence, or a climatology of cloud frequencies as chosen for illustration because of its relatively high average
a function of level or layer above the ground is crude and sky cover. With a lower average, a station should have lower
speculative, especially if information is limited to the ob- frequencies in most or all of its layers aloft.
servations from the ground. An observer sees the “sky dome”
with an uncertain radius, assumed to be 27 km in the fol-
lowing discussion and analysis. 16.3.3 Satellite and Radar Observations
A valuable study on the frequency of occurrence of
clouds in the vertical was based on observations from aircraft 16.3.3.1 Satellites. Weather satellite images are provid-
over Germany [deBary and Moller, 1963]. It resulted in ing detailed information on the structure and distribution of
estimates of the average cloud cover as a function of altitude clouds over large areas of the earth. Figure 16-31 illustrates
and the thickness of the atmospheric layer. Recently a model a variety of cloud and fog formations over Europe and
has been developed that yields similar estimates of cloud Northern Africa; the image was obtained from one of the
cover, without using aircraft observations. The model uses satellites of the Defense Meteorological Satellite Program
only RUSSWO statistics on the total sky cover and ceiling (DMSP). In addition to images in the visible portion of the
heights [Gringorten, 1982]. Figure 16-30 illustrates the out- spectrum as shown in Figure 16-31, corresponding images

16-32
 WATER VAPOR, PRECIPITATION, CLOUDS, AND FOG

Figure 16-31. DMSP LF low enhancement, 1015 GMT, 13 October 1978 [Fett et al., 1981].

16-33
CHAPTER 16
in the thermal infrared are available. These simultaneous 16.3.4.1 Aircraft Observations. In a test program, Ber-
images or brightness fields can be used to deduce the height toni [ 1967] determined that realistic estimates of the prob-
of the cloud tops. From the accumulation of such data, a ability of clear and cloud-free lines-of-sight could be derived
regional and, eventually, a global cloud climatology will be from a large sampling of aircraft in-flight observations. This
developed. resulted in a major effort in which observations were made
from aircraft, at various angles to the sky ( + 30° and + 600),
the horizon (00), and the earth’s surface ( – 3@ and – 60°).
16.3.3.2 Radar. Because of the electromagnetic scatter-
Observations were acquired over most of the northern hemi-
ing characteristics of water drops and ice particles, detection
sphere during all months, but were limited to the daylight
of non-precipitating (less than 200 pm diameter), hydro-
hours. (Nighttime observations tended to overestimate clear
meters in clouds is effectively limited to the shorter radar
conditions).
wavelengths (less than a few cm). For example, radars op-
Figure 16-32 is an example of the analyses shown in
erating at O.86-cm wavelength have been used for opera-
the resulting report [Bertoni, 1977a, b]. The large numbers
tional cloud detection and continue to be used in research
in the figure are percent probabilities of CFLOS; the smaller
for deducing physical and dynamical characteristics of clouds
numbers, in parentheses, indicate the total number of ob-
prior to the development of precipitation. Precipitation-sized
servations taken within each of the 10° latitude-longitude
particles are readily detected by longer wavelength (3–10
sectors. Figure 16-33 is similar to Figure 16-32 in that it
cm) radars, as discussed in Section 16.2.5. Optical and
shows probabilities of CLOS which are generally smaller
infrared wavelength (O.5 to 15 pm) radars, or lidars, have
than probabilities of CFLOS. This is illustrated more dra-
also been used to observe clouds; theoretical comparisons
matically in Figure 16-34 in which frequencies of CLOS
of optical and microwave cloud detection capabilities have
and CFLOS are compared for an area around Columbia,
been made by Derr [1978].
Mo, about 1000 kmz. Values for CLOS are reduced by some
Observations by the vertically pointing TPQ- 11 radar
20% to 30%, due mostly to haze. Figure 16-34 also confirms
were recorded by the Air Weather Service between 1964
that the probabilities of CLOS and CFLOS depend on the
and about 1975 at 43 locations around the world. The mul-
angle of the line-of-sight as well as the percent of sky cover.
tiple-layer ceiling observations are archived at the National
CLOS and CFLOS probabilities increase as the lines-of-
Climatic Center, Asheville, N.C. Two of these radars, ac-
sight approach the vertical.
quired by the NOAA Wave Propagation Laboratory and the
University of Washington, respectively, have been modified
16.3.4.2 Surface Observations. Whole-sky photographs
for improved reliability and sensitivity and are used in multi-
at Columbia, Missouri and conventional National Weather
sensory field programs. Early results were presented by Weiss
Service observations of sky cover were used by Lund and
et al. [ 1979], and by Pasqualucci and Miller [1981]. The
Shanklin [1973] to develop a universal model for estimating
Wave Propagation Laboratory subsequently built a trans-
probabilities of CFLOS through the entire atmosphere. Ear-
portable fully-scanning coherent 0.86-cm radar with a dual-
lier CFLOS studies based on these same observations are
polarization receiver which was first used in the spring of
described by Pochop and Shanklin [1966], Bundy [ 1969],
1981. The dual polarization capability permits derivation of
Shanklin and Landwehr [1971], and Lund and Shanklin
hydrometer microphysical characteristics, while the coh-
[1972].
erent reception permits measurement of Doppler winds in
Probabilities of CFLOS based on the Lund and Shanklin
clouds. There is increasing interest in the detection of clouds
model can be estimated through the use of the following
and aerosols by 0.3 cm radar [Lhermitte, 1981].
formula:
.
.P1 = .C,K1, (16.26)
16.3.4 Clear and Cloud-Free Lines-of-Sight
where .P1 is a column vector of a rows for each angle
To determine the utility of various communications, sur- considered; ~C, is a matrix for ct rows and s columns, one
veillance, and weapons systems, the probability of clear/cloud- row for each elevation angle, one column for each sky
. cover
free lines-of-sight (CLOS/CFLOS) must be known for the category; and ,Kl is a column vector ofs rows. The P values
geographical areas and climatic regimes in which these sys- are estimates of probabilities of CFLOS through the at-
tems must operate. (CLOS is defined as LOS not obscured mosphere, the C values are probabilities of CFLOS at angles
by either haze or clouds). Since CLOS and CFLOS are (x given k tenths of sky cover and the K values are proba-
neither direct] y observed nor measured on a routine basis, bilities of each k tenths of sky cover. The aC, matrix is
various methods have been developed to estimate their fre- contained in Table 16-28.
quency of occurrence. Some of these methods are described Using the historical records for Columbia, Missouri as
in this section and examples of data, analyses, and results an example, the following probabilities of 0/1 O, I/10, . . .,
are provided on the frequency of occurrence of CLOS and 9’10, 10/10 sky cover at Columbia are 0.187,0.047,0.047,
CFLOS . 0.049, 0.037, 0.031, 0.045, 0.045, 0.055, 0.065, 0.392.

16-34
 WATER VAPOR, PRECIPITATION, CLO~S, AND FOG

$,: *
,%
“~% -— ‘F AFCR, W[AR EQuAL. AREA MAP PEA 1
(N-TNERN HEMISPHERE I

Figw 16-32. Estimates of the probability of a cloud-b line-of-sightto the @und at 30° below the horizon ( – 30”) from altitudes betw~n 7.6 and
10.Qkm in summer.

Then the matrix multiplication indicated in Equation (16.26) Consequently, there is an estimated 50.9~ pr~bability Ofa
yields CFLOS at Columbia looking toward the zenith (900), and
a 50% and 46.370 probability of a CFLOS at 60° and 3@
elevation angles, respectively. (See Section 16.3.4.3 for
Elevation Angle discussion of a correction to a suspected bias toward higher
(deg) Probability CFLOS probabilities @uced by the Lund-Sh~in model.)
Using this model, atlases of CFLOS probabilities have
90 0.509’”
been compiled for mahy parts of Europe, Asia and Africa,
80 0.5M
as well as for the U.S. and the USSR. (Lund et al., 1975,
70 0.5M
1976, 1977, 1978, and 1979). A sample page from the atlas
60 0.500
for the United States of Ame@ (~ud et al., 1977) is
50 0.496
s~own in Figure 16-35. This shows the probability of having
40 0.483
a CFLOS in July at 1200 to 1400 LST at a 3~0 elevation
30 0.463
angle. This atlas is based oq at least a 15-yr perid of record
20 0.437
for the majority of U. S. stations shown by the dots in Figure
10 0.392
16-35.
16-35
CHAPTER 16

/
1
\ -...

Figure 16-33. Estimates of the probability of a clear line-of-sight to the ground at 30” below the horizon ( – 30°) from altitudes between 7.6 and 10.6 km
in summer.

Table 16-28. Robabilities of CFLOSas afunction ofelevation angleand obsewed TotalSky Cover [Lund and Shanklin, 1973].

Elevation Sky Cover (Tenths)

Angle
(deg) o 1 2 3 4 5 6 7 8 9 10

90 1.00 0.97 0.92 0.87 0.81 0.77 0.70 0.62 0.48 0.31 0.08
80 0.99 0.97 0.92 0.87 0.81 0.77 0.69 0,61 0.47 0.31 0.08
70 0.99 0.97 0.91 0.86 0.80 0.76 0.68 0.61 0.47 0.30 0.08
60 0.99 0.96 0.90 0.85 0.80 0.75 0.66 0.60 0.46 0.29 0.08
50 0.99 Q.96 0.90 0.85 0.78 0.73 0.64 0.58 0.45 0.29 0.08
40 0.99 0.95 0.88 0.83 0.76 0.71 0.62 0.55 0.42 0.27 0.07
30 0.98 0.93 0.86 0.80 0.73 0.66 0.57 0.50 0.38 0.24 0.06
20 0.98 0.90 0.83 0.75 0.67 0.59 0.50 0.42 0.33 0.21 0.05
10 0.97 0.86 0.76 0.65 0.55 0.47 0.39 0.32 0.24 0.16 0.03

16-36
 m
WATER VAPOR, PRECIPITATION, CLOUDS, AND FOG
Rapp, Schutz, and Rodriques [ 1973] extended the Lund-
[899)
Shanklin CFLOS model to include finding the probability
of a CFLOS between the earth and any given altitude in the
atmosphere. They accomplished this by u$ing 3-h synoptic
(1140)
reports of clouds in addition to total sky cover.

16.3.4.3 Corn@son of Estima* Between AK* tid

(747)
the CFLOS Model Based on Surface Observations.
CFLOS probabilities computed from sky-cover observations
(434) taken at 6 stations (Columbia, Mo., Atlanta, Gh., Fort WOfi,
(514)
Tx., North Platte, Neb., Albuquerque, N. M., and Ely,
Nev.) have been compared with CFLOS probabilities based
(104)
on observations taken from aircraft by Lund and Bertoni
00 I ( 1 , , , 1 ,

90 1, o [ 1980].
~00 .10 20 30 40 50 Go ~ so
2
5 ~; RELATIVE FREQUENCY CFLOS probabilities were computed at selected altitudes
.
for each of the 6 stations, each midseason month, and each
depression angle. The probabilities for the four midseason
Figure 16-34. Relative frequencies of clear or cloud-free lines-of-sight
months were averaged and plotted, and were usually higher
~bse~ed in the area between 34° to 43”N and 88° to 97”W
as a function of altitude. (The number of observations is than those based on inflight observations. Figures 16-36 and
shown in parentheses on the right hand margin. ) 16-37 show the model and inflight CFLOS probabilities for

Figure 16-35. CFLOS probabilities for July, 12~lW LST, 30° elevation

16-37
CHAPTER i6

I I I I I I I I I estimates as well as the model and inflight CFLOS proba-

bilities.
?

16.3.4.4 Cloud-Free Fields-of-View. An instantaneous

\l
“\ t CFLOS is often not sufficient for acqu~ring and tracking a
Xo
\.\\ target, or for estimating the probability of detecting a target
)\\
when there is an op~rtunity to search for favorable cloud
conditions. Therefore, probabilityy distributions of cloud-
x\ ‘k I free areas of fields-of-view (FOV) are needed. Grantham et
\ ‘\ ‘, al. [1979] and Lund et al. [1980] completed studies for
\
---o MODEL CFLOS

1
.k estimating the probability of cloud-free fields-of-view
–.–x A/c CFLOS b
.1
●
,\
(CFFOV) between earth and an airbo~e or space platform.
2.5 —oCORRECTED MODEL
\ Using whole-sky photographs taken at Columbia, Mo., the
\
I I I I I I I I I , average cloud-free fraction of each of 185 annular sectors
0 20 40 60 80 100 [see Figure 16-38] was computed, plotted, and analyzed to
% PROBABILITY determine how it varied with position on the photograph.
Figure 16-38 depicts the average cloud-ftie fraction of each
Figure 16-36. Six-station average probabilities of a clotid:free iine-of-
sector when 2/10 total cloudiness was observed. Clearly,
si~ht detemined from airerafi observations (X’ s), from the the average cloud-free fraction decreases from the center of
C;LOS m~el (.’s), and from the rncdel corrected for&as the picture, which corresponds to directly overhead, to low
(0’s) at a depression ang]e of 30°.
elevation” angles of the edges of the picture. The maxima
(highs) are not exactly centewd at the zenith (the center of
the picture) as one would extect. The failure of the isolines
depression angles 3W and 60°, respectively. The differences to be symmetric about the center of this figure is due in part
suggest that the Lund-Shanklin model CFLOS probabilities to sampling variability. However, othef studies of the pho-
are biased toward higher values. This apparent bias can be tographs have suggested that sun angle or local effects,
essentially eliminated by a linear reduction of the inodel perhaps from the city of Columbia, may also contribute to
estimates as follows: the lack of symmetry.
Circular fields-of-view centered at a ground observers
P’ = (0.99 – 0. W5a)P ; O s a <45, (16.27) zenith were investigated. Figure 16-39 shows circles with

where “a” is the altitude in kilometers and ~ is the model

N
estimate. Flgurek 16-36 and 16-37 show the corrected CFLOS

---, MODEL CFLOS

–.–x A/C CFLOS “~k 1’”
—-o CORRECTED MODEL
%. 15
B
I 1 I 1 I I 1
0 20 40 60 80 100

“k PROBABILITY
s
Figure 16-37. Six-station average probabilities of a cloud-free line-of-
sight determined from aircraft observations (X ‘s), from the Figure 16-38. Average fraction of each annular sector that was cloud-free
CFLOS model (*’s), and from the model corrected for bias on the 131 photographs when 2/10 sky cloud cover was
(o’s) at a depression angle of W. observed.

16-38
— —
H

WATER VAPOR, PRECIPITATION, CLOUDS, AND FOG

N

E w

Figure 16-39. Circular fields-of-view with angular diameters of 10°,50°,

90°, 130°, and 170° centered at the ground observer’s
zenith.

angular diameters, centered at the observer’s zenith, of 10°,

50°, 90°, 130°, and 170°. Because the probability distri-
butions are functions of the sky cover, they were stratified
on that basis.
Figure 16-40 depicts cumulative frequency distributions
of the fraction of the area that is cloud free in the 10° circle 0.2 Q4 0.6 0.8 1.0
as a function of the observed sky cover. The points on the X (CLOUD-FREE FRACTION)
figure were obtained from the data and the curves were
subjectively drawn through the points. This figure shows, Figure 16-40. Probability that a 10° FOV at zenith will have a cloud-free
fraction =x (curves for each tenth of sky cover were sub-
for example, that if one wants to be 80% sure that at least
jectively drawn to the data points).
0.5 of the 10° FOV is cloud free, operation can be allowed
only when 0/10, 1/10, 2/10, 3/10, or 4/10 sky cover is
observed. If he relaxes his requirements to a 50% probability probabilities of CFFOV from climatic probabilities of the
that at least 0.2 of the 10° FOV is cloud free, he can ad- tenths of cloudiness. Figure 16-42 shows a comparison of
ditionally operate when 5/10 or 6/ 10 sky cover is observed. the probability of a CFLOS as a function of elevation angle
Figure 16-41 shows how the cloud-free fraction varies when twe-tenths sky cover is reported, with corresponding
for five fields-of-view ( 10°, 50°, 90°, 130° and 170°) when CFFOV probabilities when the FOV is 10° and 20°. The
the weather observer reports 5/10 sky cover. Let us assume figure shows good agreement between CFLOS probabilities
that at least 0.2 of a FOV must be cloud free. The curve and FOV probabilities for cloud-free fractions >0.5 and
labeled 10° shows there is a 73% chance that at least 0.2 >0.75. However, there is only a 61% chance that a 10°
of the FOV is cloud free. The curve labeled 90° shows a FOV centered at a 30° elevation angle will be totally cloud
94% chance, and the curve labeled 170° shows a 100% free, and only a 48% chance for a 20° FOV to be totally
chance that at least 0.2 of the 170° FOV will be clear when cloud free. These differences between the CFLOS and CFFOV
the observer reports 5/10 sky cover. If, however, the re- model probabilities are larger than expected. Therefore, users
quirement is that at least 0.8 of the area must be cloud free, of the CFFOV model and the CFLOS model should un-
the probability will be near zero for the large 170° FOV and derstand that the models yield probability estimates which
will increase to 44% for the small 10° FOV. are believed to be good first approximations. Better ap-
Lund et al. [ 1980] modified the Lund-Shanklin CFLOS proximations await higher quality whole-sky photographs
model described in Section 16.3.4.2 to estimate climatic or more precise observing techniques.

16-39
 CHAPTER 16
100
90
I

80

80 p cz 70

:
a 60
u
n
70 - / .’
50 -
g
-1
P
/’”
z. 40
60 - m
o
1’
a
I
/’
n30-
/ 1’
r f
50 “ 20 x’
t

40 “
J 1 I I t 1 1 I
o 102” 3”4”5060~ eogo
ELEVATION ANGLE (“)
30 “
Figure 16-42. The probability of a CFLOS as a function of elevation
angle when 2/10 sky cover is reported (heavy curve), and
20 corresponding CFFOV probabilities when the FOV is 10°
(solid lines) and 20° (dashed lines. ).
t \ ‘!l- ‘i i
I3U” \\l
\ \
\

/!
10 , 7oo’\ ‘ 16.3.5
Water Content of Clouds
\
I \ \
Moisture in clouds can exist as a vapor, liquid, or solid.
()~
Some water vapor is always present in the atmosphere. The
0.2 04 0.6 0.8 1.0
amount of water vapor a parcel of air can hold is a function
X (CLOUD- FREE FRACTION) of the temperatu~ [Hatuwitz, 1941]. Under conditions where
the atmosphere becomes saturated the water vapor conden-
Figure 16-41. The probability that a 10°, 50°, 90°, 130°, and 170° FOV
at zenith will have a cloud-free fraction =x when the weather
ses and forms clouds.
observer reports 5/10 sky cover. Knowledge of the amounts of liquid water and ice con-
tained within clouds is important to aviation due to its direct
and indirect effects. The quantity “liquid water content” is
16.3.4.5 Joint Probabilities of CFLOS. Lund [ 1973] the concentration, expressed in mass per unit volume, pre-
extended the Lund-Shanklin [ 1973] CFLOS model de- cipitation. Although many methods of determining liquid
scribed in Section 16.3.4.2 to address the problem of es- water content have been used, none are completely satis-
timating joint probabilities of cloud-free lines-of-sight from factory.
more than one observing sight. This model requires a cli-
matic record of sky cover observations taken simultaneously
Table 16-29. Estimated 3-way joint probabilities of CFLOS at 40° ele-
from all sites.
vation angle for August at Minot P(C~); Grand Forks, Minot,
Table 16-29 gives estimated CFLOS probabilities for ~nd Fargo, P(CG CM C~); none of the three sites, P(?G ~~
Minot, N. D., P(C~); three-way joint probabilities of CFLOS CF).
for Grand Forks, Minot, and Fargo, P(CC CM CF); and,
Hour
three-way joint probability for none of the three sites, P(~~
0369 12 15 18 21
~~ ~F). The probabilities apply to August, the month when
the highest CFLOS probabilities occur, and the elevation P(CM) 0.73 0.74 0.65 0.62 0.59 0.58 0.60 0.67
angle is 40°. Table 16-30 is the same as Table 16-29 except P(CG CM c~) 0.48 0.48 0.37 0.35 0.34 0.31 0.35 0.41
that probabilities apply to November, the month when the
p(~~ CM ~~) 0.07 0.08 0.12 0.13 0.15 0.14 0.13 0.10
lowest CFLOS probabilities occur.

16-40
 -
H

WATER VAPOR, PRECIPITATION, CLOUDS, AND FOG

Table 16-30. Same as Table 16-29 except the probabilities apply to No- 1981a] and Large Scale Cloud Systems study [Cohen, 198 lb]
vember, the month when the lowest CFLOS probabilities used this method. Knollenberg [ 1972] notes that this method
occur.
, [ compared favorably to results given by direct measurement.

I I I
Hour During recent years, attempts to relate liquid water con-
0369 12 15 18 21 tent to the intensity of radar returns have been made. Stickel
and Seliga [ 1981] describe one such attempt. At this time,
P(c~) 0.50 0.49 0.50 0.36 0.38 0.39 0.40 0.49
however, radar is not able to give the definition that either
P(c.CMc,) 0.23 0.22 0.20 0.14 0.13 0.14 0.17 0.23 direct measurements or computation from particle distri-
P(CM CM c,) 0.27 0.28 0.28 0.40 0.38 0.37 0.34 0.29 butions can provide.

16.3.5.3 Supercooled Clouds and Aircraft Icing. The

principal meteorological parameters associated with aircraft
16.3.5.1 Liquid Water Content in Clouds. The actual
icing are the ambient temperature, the cloud liquid water
amount of liquid water from droplets or ice crystals in a
content, and the drop size distribution. The extent to which
cloud will vary, depending on the type, altitude, and tem-
ice forms on an aircraft (that is, icing intensity) and whether
perature of the cloud. Clouds formed by strong vertical
or not rime or glaze ice is formed depend additionally on
motion will generally contain large amounts of liquid water.
the aerodynamic shape of the airplane stmctures, its true
Lamb et al. [1976] measured as many as 1.5 g/m3 in syn-
air speed, and the length of time the aircraft is in the icing
optic-scale winter clouds that contained convective ele-
environment.
ments. Generally, clouds of this type range in amounts that
The source of most meteorological information collected
are illustrated in Figure 16-25. Aufm Kampe and Weick-
on icing parameters, including the data presented here, are
mann [1957] suggest 10.0 g/rnq as the maximum value of
from stratiform clouds (stratus, stratocumulus, and alto-
liquid water content in cumulonimbus clouds. Higher values
stratus). Ninety percent of these clouds are 1 km thick [Fig-
have been reported. The range of values and probability of
ure 16-43]. Stratiform clouds are of principal concern with
occurrence are discussed in detail in Section 16.2.5.4.
respect to icing since convective clouds (cumulus, cumu-
In large-scale storm systems, liquid water content in
Iocongestus, and cumulonimbus) are generally avoided by
most of the clouds is in the range O. 1 g/m3. The highest
aircraft for other reasons. Cumulus clouds are also relatively
values of liquid water content (see Figure 16-25) occur near
limited in horizontal dimension and duration. However,
the freezing level (or melting layer).
conditions in cumuliform clouds can result in severe icing
The values quoted in the preceding paragraphs have
been computed from particle-size distributions and assump-
tions relating crystal size and equivalent melted diameter. 1.0
They should be considered as best estimates. Increased ac-
curacy will be possible only when a reliable method of
directly measuring water content is developed.

16.3.5.2 Measuring Liquid Water Content. Three

methods of measuring liquid water content have been tried: 0.I
direct measurement, computation from particle distribu-
tions, and attenuation of radar or lidar signals.
Although several possible instruments for direct mea-
surement have been devised, only one, the Johnson-Wil-
liams (JW) indicator has been used on an operational basis.
This instrument operates on the principle that the change in 0.01
resistance of a heated wire as it evaporates impacting par-
ticles is proportional to the liquid water content. The probe
is sensitive only to liquid droplets having diameters of
<30 pm.
The most frequently used method of determining liquid
water content is by integration of the particle-size spectra
obtained by use of an optical array spectrometer. This pro- 0 0.5 1.0 1.5 2.0 2.5
cedure is straightforward when particles are liquid. How- ICING-CLOUD DEPTH, H.(km)
ever, when particles are ice crystals, assumptions concer-
ningthe relation of crystal size and their equivalent melted Figure 16-43. Cumulative frequency distribution of depth of icing-cloud
diameter are necessary. The AFGL Cirrus study [Cohen, layer [after Perkins, 1978].

16-41
CHAPTER 16
of aircraft since supercooled water in excess of 2.0 g/m3
[Mason, 197 1] have been observed as well as drop sizes
occurring over the complete range of the size spectrum.
Additional information concerning specific cloud types, as-
sociated synoptic conditions, and their relation to icing is 15
given by Lewis [ 1951] and the Air Weather Service [AWS,
1980]
Clouds containing supercooled droplets (that is, water
in the liquid state at temperatures below freezing) can exist 10
in the atmosphere to temperatures as low as 233 K. How-

1L
ever, the probability that supercooled droplets will remain
unfrozen below the temperature range of 258 K to 253 K
decreases rapidly since the number concentration of active 5
freezing nuclei becomes significant at these temperatures.
Perkins [1978] observed that approximately 90% of the clouds
sampled that contained supercooled water were at temper-
atures warmer than 253 K. The likelihood of encountering 0 —
o 0. I 0.4 0.5 0.6 0.7
clouds containing supercooled droplets as a function of tem-
LWC (g /m3)
perature is shown in Figure 16-44.
The liquid water content (LWC) in wintertime stratiform
Figure 16-45, Distribution of LWC from 65 sample cases [Glass and
clouds is a highly variable parameter. This is due principally
Grantham, 1981].
to the horizontal variations in cloud structure. LWC aver-
aged over long path lengths (>25km) usually do not exceed
0.3 g/m3. Altostratus and altocumulus clouds forming at with horizontal dimensions of 0.5 km to 1.0 km, peak LWC
higher altitudes have proportionately lower LWC amounts values of 1.3 g/m3 have been observed, although values not
due to the temperature dependency of saturation vapor pres- exceeding 0.8 g/m3 are more typical. The distribution (in
sure of water In individual stratocumulus cloud elements, percent) of supercooled water content amounts is shown in
Figure 16-45. These data are based on sample measurements
[Glass and Grantham, 1981] from a day considered repre-
I.0 sentative of clouds during winter icing conditions. The dis-
tribution of median volume diameters for these data is shown
F in Figure 16-46.

0.I

0.01
—

0.001

Figure 16-44.
L
273
I 1
263

ICING-CLOUD
I I
253

TEMP
I

Cumulative frequency distribution of tem~rature

K )
1
243

of icing- Rgure 16-46.

4 8

MEOIAN
12

VOLUME
L
DIAMETER
20
(#m)

Distribution of average median volume diameter during

24 28

clouds [after Perkins, 1978]. each of 65 sample cases [Glass and Grantham, 1981].

16-42
 a
WATER VAPOR, PRECIPITATION, CLOUDS, AND FOG
Table 16-31, Characteristics of size and mass spectra for the indicated
trgures,

Figure 16-47 Figure 16-48

LWC(g/m’) 0.27, 0.04
Number density (cm-3) 321 192
Mean Vol Diameter (~m) 12 8
Median Vol Diameter (pm) 16 14

pm in di~meter) can be found in these clouds. Their Oumber

density, however, is much less than 1.0/cm3.
Techniques developed by the Air Weather Service
[AWS, 1980] for forecasting icing conditions are based on
Ix radiosonde observations of tern~rature and the dew-point
10-1 spread (difference between temperature and *W pint). fiese
9 10 m 30 40

DIAMETER @ml parameters are used since formation and maintenance of

clouds containing supercooled droplet s~ctra require a sat-
47. Number (*) and Mass (. .x..) distribution with size for data urated or near-saturated environment. This report quotes
sets with >0.2 g/m-’ iiqui~ water. studies indicating that when only dew-point spread is con-
sidered, there is an 80% probabilityy of icing when the spread
is <3 K, and 84% probabilityy of no icing when the spread
Distributions of the number and mass density of droplets exceeds 3 K. A northern hemisphere climatology of icing
with respect to size obtained from the same data set are probabilities, utilizing the Air Weather Service analysis
sh~wn in Figures 16-47 and 16-48. Both figures are based method [AWS, 1980] and applied to 380 northern hemi-
on long:duration data sets, Figure 16-47 is drawn from data sphere radiosotlde stations, has been published by the Air
sets where the average LWC was equal to or greater than Weather Service [Heath and Cantrell, 1972].
0.2 g/m3, whereas Figure 16-48 is based on data sets from
which LWC values did not exceed 0.1 g/m3. Some param-
eters obtained from these distributions are shown ih Table
16-31. Past observations [Perkins, 1978] of droplet-size dis- 16.3.6 Clouds above 6 Km
tribution in supercooled stratiform clouds indicated that val-
ues can range up to 50 pm. More recent observations have Almost all clouds found above 6 km are ci,rnform clouds
suggested that large supercooled drops (approaching 300 composed of ice crystals. Cirriform clouds have been ob-
sefied at altitudes in excess of 15 km.

16.3.6.1 Obseivations of Cirrus Clouds. In 1896-97,

various observatories made double-theodolite measurements
of cirrifo~ clouds. Table 16-32a lists the average heights
of the cloud bases in different seasons and at different lat-
itudes. Table 16-32b shows the maximum base heights of
cirrus and cirrostratus clouds at these locations [Suring,
1941; Appleman, 1961]. This table indicates that the bases
_“ of cirriform clouds are higher at lower latitudes and during
the s~mmer, and that the bases of cirrus clouds tend to be
higher than those of cirrostratus layers. These observations
were confirmed by aircraft flights in the 1950s. The flights
1- I-. \ I ma~e at that time also provided data on the heights of tops,
x

,&
11
L ,.-,I 1 1 1
\
1 Y. 1 ! , I
I
I
thickness, and frequency of occumence of cirnform clouds.
Cirnform clouds in the ternmrate zone are more com-
mon in cold weather and at high latitudes. koth bases and
o io m 40 !s0
tops vary in height, being higher in summer and lower in
OIAMETER (pm] wihter. Tops of cirriform clouds are eioselyassociated with
the trompause. Generally, the tops occuf 1.2 to 1.5 km
F[gufe 16-48. Number (*) and mass (,.x ..) distribution with size for data Mlow the tropopause. The thickness of high clouds averages
sets with <0.1 g/ml liquid water. 2.1 to 2.2 km and shows little or no variation with season.

16-43
—
CHAPTER 16
Table 16-32a. Mean heights of cirms bases for international cloud year 1896-1897.

Northern Northern
summer winter
Latitude Place (km) (km)

78”N Cape Thordsen 7.3 —

70”N Bossekop (Sweden) 7.5 —
60°N Pavlovsk plus Uppsala 7.9 7.1
51”N Potsdarn plus Trappes (average) 8.7 7.7
40”N Blue Hill plus Washington (average) 10.2 9.1
35”N Mera (Japan) 11.0 9.1
14°N Manila 12.0 11.1
7°s Batavia 11.0 —
35°s Melbourne 8.5 9.6

Table 16-32b. Maximum heights of cirrus bases for international cloud year 18961897.

Cirrus Cirrostratus
Latitude Place (km) (km)

78”N Cape Thordsen 8.6 —

70”N Bossekop (Sweden) 11.8 10.4
60”N Pavlovsk plus Uppsala 11.7 10.1
51°N Potsdam plus Trappes (average) 12.7 11.9
40”N Blue Hill plus Washington (average) 15.0 13.6
35”N Mera (Japan) 16.8 15.5
14”N Manila 20.4 17.1
7°s Batavia 18.6 14.2
35°s Melbourne — —

In the 1970s, both particle density and liquid water con- Heymsfield did find occasional particles as large as 1900
tent of high-altitude clouds were investigated in airborne pm, but these were very rare, and generally their number
studiw of the microphysics of clouds. Aircraft studies of density was less than l/cm3. Most cirrus particles are smaller
cirrifofi clouds have been reported by Heymsfield [1975], than 600 pm. Varley et al. [ 1980] found some larger par-
Ryan et al. [ 1972], and McTaggert-Cowan et al. [1970]. In ticles in the vicinity of a surface storm, but this was an
more recent years, AFGL has published a series of reports exceptional case.
on this sttbject [Cohen, 198 la, b]. Although comparison of the work of different investi-
AFGL research flights have shown that cirrus cloud par- gators is risky due to differences in instrumentation and
ticles are almost always smaller than 1200 pm in diameter. presentation, an analysis of the results of three studies are

Table 16-33. Water content (melted ice) and particle density observations in cirriform clouds.

Water Maximum
Content Particle
Altitude (Melted Ice) Density
Observer Date Cloud Ty~ (km) (grn/m3) (Number/cm’)

Heymsfield 9 $Ul 70 Cirrus 7 0.30 Not reported

Heymsfield 4 Apr 72 Cirrostratus 8 0,03 0.20
Ryan 12 Jun 70 Cirrus 9 Not Reported 3.60
McTaggart-Cowan 18 Jun 69 Cirrus 12 Not Reported 0,53
AFGL 28 Jan 79 Heavy 8 0.23 0.13
Cirrostratus
AFGL 2 Feb 79 Cirrostratus 9 0.03 0.11

16-4i
 a
WATER VAPOR, PRECIPITATION, CLOUDS, AND FN
,.O

o
0
I
0
—— COHEN- tiRNES
. 18362 4 APR 78
0
10-‘ 0
0
e 00 vARLEY-COHEN-6ARNES
0
0
18462 29 JAN 79
0
0 — RYAN 12 JUN 70
0
0 . . ..o HEYMSFIELO(CS OECK)
0

‘E 0 ---- HEYMSFIELO(CS UNICUS)

y o
I 0-2 -0
z 0
e
z .
0 l<ss, 1,, ,, ,
o ,..3 I ss0 1,

F
0

0 0 50 IW 150 200 250 3DD

a
\J~v~
0
a EQulv MEU3D D!AMET3R (pm)
+
z 0:
u
~
-3
10 ‘t
g
:‘,
.$
:1
.1
‘\ 0:0 \ 00
Figure 16-50. Particle size dis@bution of small-p~icle subvisible cirrus.

:1

~ \ 00

:1 \ can fall to the ground [Kikuchi and Hogan, 1979] as “dia-

10-4
.1
:,
:1
\,q 0“0 mond dust” snowfalls. Cirrus particles can survive falls in
[ “.: ,<*1 ‘L,, 00
.-, .\ ~s.- excess of 2 km in subsaturated air [Hall and Prttpacher,
“.$.:.- \O ..,C..

\*
‘** 1976] .
... O.’;, Hall [1964] implies that subvisible cirrus can reduce the
... \ .>,
.. . lift on lamin~ flow wings. Barnes [1978] notes hat cirrus
165 1 I I
{ I 1
0 400 m 1200 particles cause erosion of hypersonic vehicles, such as reen-
LENGTH DIAMETER (pm) try vehicles,, atid that the presince of unre~rted thjn and
subvisible ci~s reduces the efficiency of solti collectors
Figure 16-49. Cirrus particle concentrations seen by diffeFrmt obkwers. and attenuates energy in laser beams [Bties, 1982].

16.3.6.3 Clouds Akve the Tropopause. While’ almost

shown in Figure 16-49. This figure gives an idea of different all of the water and ice clouds in die atmosphere ti in the
size distributions and is not intended to compare the results tropshere, there are some occasions in which clouds pen-
of the different studies. etrate into the stratosphere: This is generally the result of
Typical water content (melted ice) and particle densities strong convective activity in which severe thunderstorms,
reported by different observers are shown in Table 16-33. caused by unstable conditions and high moisture content,
have enough energy to break through the tro~pause (Figure
16,3.’6.2 SubVisible Cirrus. Citis-like ice ptiicles oc- 16-52). In this vein, the frequency and percent occurrence
curring in apparently cloudless air are called subvjsible cir- of radar precipitation echoes have above 15 km have ken
rus. Thin cirrus clouds that are only visible from cetiain estimated [Kantor and Grantham, 1968]. This climatology
directions or at certain times may also @ called subvisible of radar precipitation echoes also provides information on
cirrus: Most subvisible cirrus is found in the 3 or 4 km the month of occurrence and estimated vertical extent of
immediately below the tropopause. The maximum occur-
rence of cirrus, both visible and subvisible, is usually 1 to
105
2 km below the tropopause. Recent observations suggest = AUG 1976
that subvisible cirrus is present more than 7570 df the time; !0+ 161001-161630

one rarely finds cloudless areas in the cirrus region in which 10 ~

there are fewer than one particle greater than 2 pm in di- ,.2
F
ameter per cm3:
There are two kinds of subvisible cirrus. The most com- o’

mon consists of small particles with diameters of less than o“

50 ~m, with most particles being less than 20 pm, in di-

o“
ameter (see Figure 16-50). Barnes [1980] describes the sec-
@
ond type of subvisible cirrus as consisting of isolated ice
crystals with diameters of 100 pm to 3~ tire. These pti- ,.3~mo
o w 1?0 1s0
ticles may occur with the first type (see Figure 16-51), or EQUV MELTED DIAUETER (pm I

they may be found in layers devoid of the smaller particles,

having fallen from higher-level clouds (often cirrus unicus), Figure 16-51. Particle size distribution showing ~th types of subvisible
where they were generated. In Arctic regions, these crystals cirrus.

16-45
CHAPTER 16
Table 16-34. Distribution of cirrus tops relative to the tropopause.

Distaoce From Tropopause Frequency of Occurrence

(m) (%)

British Isles

>1200 above 10
1200 to O above 10
0 to 1200 below 50
1200 to 2400 below 16
>2400 below 14
I

Canada

~gure 16-52. Average monthly tropopause penetrations by thunder- >1500 above 1

storms, 1961–1964 [Long et al,, 1965]. 1500 to O above i2
O to 1500 below 52
1500 to 3000 below 23
cumulonimbus clouds over the United States. As a further
<3000 below 12
example, Table 16-34 indicates that cims tops can build
well beyond the tro~pause. The cirrus produced by thun-
derstorms can often extend about 185 km downwind of the
cell that produced it. Thus, stratospheric clouds, while not 16.4. FOG
common, do exist and can be expected in and neti regions
of strong convective activity. Fog, as well as clouds, affect the propagation of elec-
The size and shape of high altitude convective clouds tromagnetic radiation through the atmosphere, especially in
have been estimated [Long et al., 1965; Roach, 1967]. For the frequency spectrum >30 GHz (<1 cm wavelength).
thunderstorms that have penetrated the tropopause in the The absorption and scattering of electromagnetic radiation
midwestem United States, the shape of the cloud is a cone by water spheres and fogs have been extensively studied.
with a diameter-to-height ratio of roughly five to one for Deirmendjian [ 1964], ,Herman [ 1962], and Platt [ 1970] have
the portion of the cloud above the trcrpopause. However, discussed extinction and scattering by water droplets in the
much of the cirrus blowoff (shield) that is observed from visible, infrared and subrnillimeter regions of the spectrum,
the ground remains below the tropopause. Mean diameter respectively. Gunn and East [1954] have discussed absorp-
of high altitude convective cells, including the associated tion and scattering of microwaves by clouds and rain. The
anvil, has been estimated at 32 to 50 km. amounts of absorption and scattering are functions of the
microphysical structbre of the fog.
16.3.6.4 Tropopause Height. Because there are sea-
sonal, latitudinal, and daily variations in the height of the
tropopause, both the latitude and season of the year must
16.4.1 Microphysical Structure
be considered in studying the height of the tropopause. Table
16.4.1.1 Condensation Nuclei. The atmosphere con-
16-35 gives mean seasonal heights of the tropopause in the
tains a large concentration of dry aerosol particles ranging
Notihern Hemisphere; however, mean seasonal heights may
in size from abotit 1&lpm to 10pm radius. Mason [1971]
be several kilometers higher than indicated in Table 16-35
divides the spectrum of particles into three parts: (1) Aitken
on the east side of semistationary lows such as the Icelandic
and Aleutian lows during the winter and spring months.
In middle latitudes the day-to-day variations in the height
Table 16-35. Mean seasonal heights of the tropopause
of the tropopause exceeds the seasonal variations; large vari-
ations in height occur as a result of horizontal oscillations Altitude
in the tropopause. The arctic tropopause from the north is
brought southward in the rear, and the tropical tropopause 3ec–Feb Mar–May lun–Auf Sep-Nov
from the south is projected northward in advance of well- Latitude (m) (m) (m) (m)
developed migratory cyclones. Although data are not suf-
60” to 80”N 8400 8800 9400 9200
ficient to establish exact limits, available data indicate that
50° to 60°N 9200 10000 11300 10000
in middle latitudes (40° to 60”N) the tropopause can fiuctuate
40° to 50°N 10000 11OOO 12700 11800
between about 6 and 15 km. North of 60° and south of 30°
30° to 40”N 14000 13800 14000 14200
latitude, the range is much smaller, the order of 900 to 1800 17000 16800 15700 16300
10° to 30”N
m around the seasonal mean heights.

16-46
 i

WATER VAPOR, PRECIPITATION, CLOUDS. AND FOG

nuclei (r <0.1 ~m), (2) large nuclei (O. 1 < r< 1.Opm), the droplets, (2) the low collection efficiencies of the smaller
and (3) giant nuclei (r > 1.Opm). All of these particles can droplets resulting in unreliable counts below about 4 ~m
act as condensation nuclei provided the supersaturation is radius, and (3) fog-laden air must be drawn through a small
large enough (300%). However, in the real atmosphere there opening, thus creating problems with obtaining represen-
is a sufficient number of hydroscopic particles so that con- tation samples. In recent years the light-scattering technique
densation occurs at supersaturations of less than 0.1 %. ~ese has become much more reliable and popular. The advan-
hydroscopic particles are either large or giant nuclei. In other tages of this technique are the near real time readout of the
words, the Aitken nuclei are not an important factor in the drop size distribution and the ability to measure the smaller
formation of fog even though concentrations of Aitken nu- droplets. The main disadvantage of most light scattering
clei up to 106/cm3 are possible in urban industrial areas. instrumentation is that, like the impaction technique, drop-
Concentrations of large and giant nuclei are considerably lets must be drawn into the sample volume through a narrow
less, averaging about 102/cm3, with a maximum concentra- opening.
tion of about 103/cm3. Most impaction techniques show droplet radii ranging
Several workers have derived analytical expressions de- from 2 to 40 Lm with a peak or mode in the number con-
scribing the nuclei size spectrum. The most popular expres- centration occurring around 5 to 10 pm. Recent light-scat-
sion is that derived by Junge [1963] for particles greater tering techniques, however, reveal that the droplet count
than r = 0.1 pm frequently continues to increase at radii below 5 to 10 pm.
An example of this is described in Figure 16-53 which shows
— dN . Cr-E, (16.30)
four comparisons of fog droplet spectra at Otis AFB, Mass.,
d(log r) each measured simultaneously with a Calspan droplet sam-
pler (which uses gelatin coated slides), and a Particle Meas-
where N is the number of nuclei per cm3 in the radius interval uring System forward scatter spectrometer probe (FSSP-
d (log r), and B and C are functions of the drop-size dis- 100) [Kunkel, 1981].
tribution. Measurements from several different parts of the Using the FSSP- 100, Kunkel [1982] shows three basic
world show, on average, an r-~ law, especially in the size types of distributions in advection fogs at Otis AFB (Figure
range 0.05 pm through 1.0 pm. Below a radius of 0.05 16-54). The common feature of the three distributions is
pm, the number of Aitken nuclei begins to decrease with that the maximum concentration per micron diameter in-
decreasing size. terval occurs below 2 ~m, apparently the result of inactive
There are 3 primary sources of origin of nuclei: (1) soil or haze nuclei. The middle spectrum (Type B) was the most
and vegetation, (2) sea spray, and (3) combustion. De- typical of the distributions observed at Otis AFB, showing
pending on their origin, the nuclei normally contain various a primary mode between 0.5 and 2 pm, a secondary mode
quantities of SOX, NH, NO, Cl, and Na. The SOX and NOX or plateau between 5 and 10 pm and a plateau between 15
are primarily from industrial or combustion sources, Cl is and 30 pm. Type A has a mode in the 15 to 30 pm plateau.
from industrial sources as well as the sea, NHX is from This type of distribution was quite common in radiation fogs
decaying organic matter, and Na is from the sea and exists observed at Hanscom AFB, Mass. Type C spectra show no
primarily as sodium chloride. plateau or peak but a steady decrease in concentration which
Depending on the size and hygroscopicity of the parti- can be represented by a power law curve [Junge, 1963].
cles, condensation may begin when the relative humidity is
as low as 7090. As the humidity rises due either to additicmal
moisture or a lowering of the temperature, the particles 16.4.1.3 Liquid Water Content. The liquid water con-
continue to grow and form a haze. When a slight super- tent (LWC) of fog can be measured through a variety of
saturation occurs, the particles quickly grow into fog drop- techniques. One technique is to collect the liquid water on
lets. The type and concentration of hydroscopic particles some type of absorbent filter paper over a specific time
and the amount of superasaturation are the primary factors period. The paper is weighed before and after collection,
that affect the fog drop size distribution and resulting vis- and the increase in weight can then be converted to LWC.
ibility. There are various types of hot-wire devices in which the
droplets impact on electrically heated wire and subsequently
16.4.1.2 Drop Size Distributions. There are two pri- evaporate and cool the surface of the wire. The magnitude
mary techniques for measuring fog drop-size distributions: of the cooling is a measure of the LWC. The LWC also can
impaction and light scattering. In the impaction technique, k determined by integrating the dropsize distribution curves,
droplets are impacted on a glass slide which is usually coated assuming the volume of air sampled is known. A small wind
with gelatin or some oily substance, the droplets leaving tunnel or blower is necessary with all of these techniques,
impressions on the slide. The impressions are then later in order to draw the sample into the sampling surface or
counted and sized to determine the drop size distribution. through the sampling area.
The 3 primary disadvantages of this technique are (1) the Measurements of LWC in fog have been very sparse.
tedious and time consuming process of counting and sizing Therefore, it is difficult to obtain any type of statistics on

16-47
CHAPTER 16
DHTE : 3 dUL BE DRTE : 3 dUL 90
-1 -1 TIME : 2109 EDT
g TIME : ~2S EDT
2
1
g 100

--lrl

u
v

DROP RADIUS (pm) DROP RADIUS (pm)

DRTC II LJUL9RI
: DRTC rJULS0
s II
TIME : 0319 EDT TME : 2310.EDT

00 1

“v
v

DROP RADIUS (pm) DROP RADIUS (pm)

Figure 16-53. Four examples of droplet concentration spectra from the Calspan sampler (hatched area) and the FSSP- 100

the frequency of LWC amounts. Recently however, Kunkel 16.4.2 Visibility

[1982] determined the LWC from 400 samples during 7
advection fogs at Otis AFB by integrating the drop-size 16.4.2.1 Definitions. There are several terms commonly
distributions which were averaged over 5-rein periods. Fig- used to describe the opacity of the atmosphere. Some of
ure 16-55 shows the percentage of time that the LWC at these terms and their definitions are
5 m and 30 m above the ground is below a given value. Visibility-A subjective evaluation of the greatest dis-
The data include all samples with a liquid water content of tance in a given direction at which it is just possible to see
at least 0.01 g/ins at the 5 m level. The average LWCS were and identify with the unaided eye in the daytime a prominent
0.03 and 0.17 g/m3 and the maximum LWCS were 0.33 and dark object against the sky at the horizon, and at night a
0.48 g/m3 at the 5- and 30-m levels, respectively. known, preferably unfocused, moderately intense light source.

16-48
 J

WATER VAPOR, PRECIPITATION, CLOUDS, AND FOG

MEAN MEAN MEAN
CONC DIA
1000 CONC DIA
1000 CONC DIA
(cm-3) (pm) {cm-3’ (pm) (cm-3) (~m)
171 6.5 574 5.0 604 5.2
100 100
LWC EX COEFF EX COEFF LWC EX COEFF
(g/m3) (KM-I) (g/m3) (KM-l)
.390 40.5 10 10 ~ .143 39.4

I I

DATE DATE DATE

6/27/80 0.1 7/12/80 0.1 7/3/80

\
0.01 0.01 ●

o 10 20 30 40 50 10 20 30 40 50 10 20 30 40 50
A DROP DIAMETER (pm) B DROP DIAMETER (pm) c DROP DIAMETER (pm)

Figure 16-54. Exampleof three types of clropsize distributions1ssadvectionfogs at otis A~, Mass. [Kunkel, 1981]

Prevailing Visibili~—The greatest horizontal visibility

(16.31)
that is equaled or surpassed throughout half of the horizon
circle. This value is entered in surface weather observations.
Daytime Visual Range—The distance at which the ap-
parent contrast between a specified type of target and its Nightime Visual Range—The greatest distance at which
background becomes just equal to the threshold contrast of a point source of light of a given candle-power can be
an observer. Daytime visual range V is a function of the perceived by an observer under given atmospheric condi-
extinction coefficient of the atmosphere k, the threshold tions. Nighttime visual range V is a function of extinction
contrast of the observer’s eyes E, and the viewing angle. coefficient of the atmosphere k, the candle-power of the
The relationship can be described by Koschmieder’s Law point source I and the threshold illuminance for the ob-
[1924] server’s eyes E. The relationship can be described by Al-
lard’s Law [1876]

90 - V2 = ~ exp (–kV). (16.32)

So -

70 -
Threshold Contrast (E)—The smallest contrast of lu-
<
2. 60 - minance that is perceptible to the human eye. The threshold
&
contrast value can vary from 0.005 to 5.0 depending on the
?Xl -
z
u observer and the brightness and visual angle of the target.
g 40 -
An average threshold contrast of 0.055 is normally used.
E
w - Threshold Illuminance (E)—The lowest value of illu-
minance which the eye is capable of detecting under spec-
20 -
ified conditions of background luminence and degree of dark
10 - adaption of the eye. A value of 101 lumens per km2 is
normally used.
0 alo 0.20 am 0.40 D
Extinction Coe@cient (k)—A measure of the space rate
LIOUIO WATER CONTENT (g/m’)
of extinction of any transmitted light. The extinction coef-
Figure 16-55. The percentage of time that the liquid water content at 30 ficient is a function of the concentration N and radius r of
and 5 m above the ground is be)ow a given value [Kunkel, particles in the atmosphere; it can be calculated by the fol-
1981].
lowing.

16-49
CHAPTER 16

k = IT~ KiN,r~, (16.33)

i=)
where K is the scattering area ratio foragivenptiicle size
i. K is normally given a value of 2.

16.4.2.2 Variability. The variability of visibility in fog

has been the subject of numerous recent studies, such as
Chisholm and Kruse [1974a], Chisholm [1974], Chisholm
and Kruse [ 1974b], Brown [1979], and Jiusto and Lala
[ 1980]. The first four papers present analytical results of
data collected in AFGL research projects during the last ten
years. For example, an automated mesoscale weather net-
work was in operation in eastern Massachusetts from 1972 In
to 1976 and was followed by the establishment of a weather
facility at Otis AFB, Mass. on Cape Cod. All visibility
measurements were obtained using forward-scatter-meters ii
IT
(FSM) whose characteristics have been previously described
[Muench et al., 1974]. The visibility measurements in this
101 I L ~~
section will be expressed either in terms of atmospheric o 10 203040 W607080 90
extinction coefficient (x 10q/m) as directly measured by the TIME (minutes)
FSM, or a daytime sensor equivalent visibility (SEV) as
defined by George and Lefkowitz [1972]. The daytime SEV Figure 16-56. Temporal variability of SEV persistence error as a function
of restriction cause [Chisholm and Kruse, 1974a].
is equivalent to the daytime visual range defined in the
previous section.
The presentation of the variability of visibility in fog indicating more rapid changes in visibility over short dis-
will be divided into three parts: ( 1) temporal variability; (2) tances and time intervals.
horizontal variability and (3) vertical variability. Meas- Figure 16-57 is a time section of visibility (SEV) ob-
ures of variabilityy are the simple correlation coefficient r, served at the three instruments at Hanscom AFB during a
and the standardized error of estimation (SE) or persistence radiation fog. The rapid and erratic fluctuations of visibilit y
error computed from in time and space are clearly shown.
A 45-m tower located about 1000 m north of the west
SE = [(1 –~) U2]”2, (16.34) runway instrument (W) was also instrumented with forward-

where-cr2 is the percent variance of the dependent variable.

Temporal Variabili@-Chisholm and Kruse [1974a] ex- 9 MAR 1973
amined the question of obtaining visibility at an airport with w
less than three small-volume visibility instruments along a N /
runway and produced an insight into the temporal variability
of fog. Figure 16-56 shows the resulting plots of persistence
1/8
error for a single instrument (W) and a three-instrument
mean based on autocorrelations with time lags from 1 to 90 1/4

min. They restricted the data to night-time observatio”ns to z

x
1/2

avoid sunrise-sunset effects. The instruments were situated

>’
on both sides of the runway. Distances between the central %2
site and the two extremes were 1240 and 1280 m; the dis-
5
tance between the two extreme sites was 2400 m.
The single-station and three-station mean curves are nearly 10

identical for advection fog. They show a rapid increase in

persistence errors during the first 10 min and consistently
I I 1
lower persistence errors than the radiation fog cases. The 8 PM 9PM
similarity of the single- and three-instrument means vali- TIME (EST)
dates existence of a higher degree of homogeneity in ad-
vection fog. On the other hand, in the radiation fog cases,
Figure 16-57. Shoti time plot of SEV at three locations along Hanscom
the single- and three-station means differ widely and have AFB mnway in a radiation fog episode [Chisholm and
consistently higher persistence errors than advection fog, Kruse, 1974a].

16-50
 a
WATER VAPOR, PRECIPITATION, CLOUDS, AND FOG
125 05T
r
I 00 -
90 -
< 80 –
L 70 -
a 60 -
0
~g 50 -
w
~ 40 -
0
z
w 30 -
b
w
z o STATION 2 (4 m ABOVE RUNWAY)
m
w 20 X STATION I (45m ABOVE RUNWAY)
a.

,o~ 10 20 30 4Q 50 60 70 80 90 23 JAN 1973

I
TIME (minutes) m STAT ION-ERY
F5M-I 0
FSM-2 +
Figuk 16-58. Time dependent persistence error of SEV as a function of
sensor height above ground [Chishoim arid fise, 1974b].

scatter-meters. Figure 16-58 shows for advective fog cases 0730Ati (EST) 08d0AM
that the persistence error is consistently less at higher levels
indicating a greater degree of homogeneity in advection fog Flgurc 16-59. Time plot of SEV from two sensors colloeakd at east end
with height. of runway in a radiation fog episode [Chisholm and Kruae,
1974a].
Horizontal Variability-Chisholm and Kruse [1974a]
compared simultaneous readings of two forwatd-scatter-me-
ters separated horizontally by 3 m to obtain a measure of for 5 classes of visibility restriction at the touchdown point.
the accurhcy and representativeness of the individual sensor Data from the radiation and advection fogs have been further
observations. They found that the advective and radiation stratified so that each end of the runway (W and E) becomes
fog variability at this scale is on the order of 8% and .13%, a touchdown point. The principal diagonal in each table
respectively. Figure 16-59 shows a 1-h sample of the si- represents the homogeneous condition, that is conditions at
multaneous observations during a period of radiation fog mid- and endpoints of runway are the same as at the touch-
when visibility varied from 10 km to less thad l@ m.
Figure 16-60 is a plot of forward-scaper-meter extinction
c~fficient measurements made at the Otis Weather Test A ,00 ❑ 0100 A PIOO Q100
+ x 8NO05

Facility at Otis AFB, Mass., during an advection fog on 5 91

Nov 197$. For comparative purposes, an extinction coef-
ficient of 30 x l&/m denotes a daytime SEV of approx-
imately 1 km while a reading of 210 x 1~/m represents
an SEV of about 140 m, assuming a contrast ~reshold of
0.055. These measurements were made by instruments lo-
cated at the 30-m level of four towers (A, B, P and Q).
Towers P, A, and Q were oriented southeast-northwest at
500-m intervals. Tower B was located about 85 m southwest
of Tower A, the center tower of the line. The graph shows
the degree of homogeneity that existed over these distances
in one pmicular case. N-

Chisholm and Kruse [1974a] also evaluated the degree

of homogeneity in radiation and advection fogs aiotig a
.—.- -..—.....–.
2
. ...4. .;. ..-

TIME (GMT)
_+* . . . . . . .

5
..& 6

runway (locations W, N, and E; see temporal variability)

by examining observations at two points on the runway
Figure 16-60. Time series of atmospheric extinction coefficient (1-rein
given the conditions at a third. Table 16-36 presents a sum- averages) measured at the 29 m levels at towers A, B, P,
mary of the two distributions, labeled midpoint and rollout, and Q oh 5 November 1978 during fog.

16-51
CHAPTER 16
Table .-
16-?6. Conditional freauencv
,. distribution along Hanscom AFB Runway. of equivalent
. daytime SEV

Midpoint Distribution Rollout Distribution

SEV
Restriction Touchtiown Range > 80(– 400– 200- 0- > 800- 400– 200- 0-
Cause Location (m) 1600 1600 800 400 200 1600 1600 800 400 200

Radiation w >1600 80 9 354 77 8 2 5 7

Fog 800-1600 49 14 12 8 17 54 8 8 8 22
400-800 23 5 18 21 32 42 12 6 21 20
200400 10 * 7 32 50 16 11 23 35 15
0-200 8 0 7 31 55 9 5 7 42 37

E >1600 84 5 344 84 6 3 33
800-1600 49 10 13 16 11 62 6 7 14 12
400-800 26 6 8 34 27 24 8 5 40 23
2WO0 23 5 4 24 44 19 3 6 23 49
0-200 11 15 12 17 45 28 9 6 10 47

Advection w >1600 96 4** o 93 7 * 00

Fog 800-1600 16 78 5 1 0 12 78 10 1 0
400-800 1 18 78 3 0 2 22 63 13 0
200-400 0 0 10 85 5 0 0 4 79 17
0-200 0 0 0 24 76 0 0 0 2 98

E >1600 98 2** o 96 4 * 00
800-1600 19 76 500 20 71 9 00
400-800 0 21 75 3 0 * 25 71 40
2moo O 0 22 76 2 0 2 16 81 *
0-200 0 0 0 47 53 00 0 44 56

All w >1600 94 4 1 1* 90 7 1 1 1
Restrictions 800-1600 18 74 7 1 * 11 79 8 1 I
400–800 3 15 76 4 2 5 25 57 12 I
200400 2 * li 73 14 4 3 10 71 12
0-200 4 .0 4 24 68 74 5 32 52

E >1600 96 3* 1* 95 3 1 * *
800-1600 23 67 9 1* 22 66 10 1 1
400-800 5 17 69 6 3 6 19 64 83
2M-400 5“ 1 19 62 12 6 2 14 62 ]6
0–200 6 7 6 28 53 16 5 4 22 53
-, .. . ,“
-less man I 70

down point. The significantly higher variability of the ra- by Muench et al. [1974] and was found to have an inherent
diation fog is very apparent, for example, with a touchdown overall error of f 2070 due to calibration drift and instrument
visibility, between 400 and 800 m the corresponding visi- noise as compared to the *4% of the FSM.
bility at the midpoint and rollout are in the same range only Figyre 16-61 shows a comparison of spatial variability
189. and 6% of the time for the West (W) touchdown and at Hanscom AFB with variabilities reported by Hage and
8% and 570 for the East (E) touchdown. For advective fog Enirekin [1965] for Kennedy Airport, N. Y., and by Hage
the midpoint frequencies are in the same range 78% and and Wilson [1967] for Atlantic City, N.J. Chisholm and
767. and rollout frequencies 63% and 57%. Disparity be- Krase [1974a] imply that a significant part of the variability
tweeen these results and those reported by Schlatter and at Kennedy and Atlantic City is due to use of the trans-
Lefkowitz [ 1968], which showed more variability at ai~rts missometer as a measuring system rather than to natural
located at Atlantic City, N.J, and Los Angeles, Calif., can causes. Figure 16-61 illustrates horizontal variability over
be attributed to different climatic regimes and to different distances ranging from 0.4 to 3 km in radiation and advec-
observing systems. The transmissometer used in the Schlat- tion fog. Figure 16-62 represents a compilation of space
ter and Lefkowitz [ 1968] report was compared with the FSM correlations in the range 5 to 40 km for station pairs carefully

16-52
 m

WATER VAPOR, PRECIPITATION, CLOUDS, AND FOG

ATLANTIC CITY
and 45 m levels. Figure 16-63 from Chisholm and Kruse
[1974a] shows the cumulative frequency distribution of vis-
ibility in advective fogs (and rain cases) at three levels:
ground, 30 m, and 45 m. ‘The figure shows a dramatic
‘-:>2;N”Y increase of reduced visibility with height. For example,
A
during advective fog, risibilities (SEV) greater than 400 m
occur about 90% of the time at the surface, while at the 45
m level, they occur only 65% of the time.
HA NSCOM
GROUND FOG Figure 16-64 shows a typical time section of the vertical
“ /
variation of visibility (extinction coefficient) at Otis AFB,
Mass., on 5 Nov 1’978 during an advection fog. Five levels,
from 3 m(X) to 57 m(A200), of visibility were plotted. The
\ striking decrease in visibility (increase in extinction coef-
:=::”\F
ficient) with height is readily app~ent. Figure 16-65, on
\KF the other hand, shows a time section of visibility when a
I I I I I
0.5 I 15 2 2.5 3 typical ground fog occurred. The same levels of the tower
STATION SEPARATION (km) are shown. The stiking feature in this figure is the complete
reversal of visibility gradient, with the surface instrument
Figure 16-61. Companion ofhorizontal vtiability of SEVatt&eai~fl (X) recording the lowest visibility.
locations [Chisholmand Kmse, 1974a].

16.4.2.3 Relationship Betw~n Extinction Coefficient md

selected from the AFGL Mesonetwork to eliminate coastal Liquid Water Content. A direct relationship exists be-
and hilltop effects and to insure similar elevations. The tween the extinction coefficient and the amount of liquid
decay in correlation as a function of station separation is water in a cloud. However, the extinction coefficient in the
clearly evident in the advection fog and quite dramatic in visible part of the spectrum 1s also inversely related to the
the case of radiation fog.’ Because radiation fog rarely is radii of the droplets. In’ other words, a large-droplet fog
found in the eastern part of the network, the station pairs with a given liquid water content will yield a lower extinc-
for this class restriction were lirn’ited to 15 km. tion (higher visibility) than a small-droplet fog with the same
Verti~al Variabi/iq—The AFGL Mesonetwork, in ad- liquid water content. Eldridge [1966] suggested that the
dition to numerous surface weather ‘stations, contained one extinction coefficient (u) and the liquid water content (W)
tower at Hanscom AFB which was instrumented at the 30 can be related by the empirical formula

SPACE CORRELATION DECAY AS A

I .0 FU NCT ION OF STATION SEPARATION

09 ,f
v/ ,
0.8 /$

,/
// f’

~/
z <~.
o +

/
x
— 0.7
+ / / ,x,’
/,
//
a
-1 /$ ,0 4
=.
’06
“/’”’
!
*
m *
x
o
u
~o.5
x \ *
/9°’.,.fl,’
g \
x
.. /:’p’ ‘“ /x
a
m X RAUIATION
7 ,$$yx.. X ADVECTION FOG

o RAIN
0,4 8 Syow
$’/,,/
o Rhl N — SURFACE 3-STATION MEAN
● AD VECTION ,, /
/
—-. 30 m LEVEL OF TOWER
\

0.3 I I I I I I I I
4 —.-– 45 m LEVEL OF TOWER

5 10 15 20 25 30 35 40 I I I I
STATION SEPARATION ( km ) 1600 800 400 200

LOWER LIMIT OF SEV (meters)

Figure 16-62. Average spatial correlation of SEV in tbe Hanscom me-

sonetwork during various types of restrictions [Chisholm Figure 16-63. Cumulative frequency distribution of SEV at three levels
and Kruse, 1974b]. above the Hanscom AFB nmway [Chisholm, 1974].

16-53
CHAPTER 16
4200 ❑ A150 A AIOO + A50 )( X ~ 8NO05

9 *

2 5
3 TIME(GMT) 4

Figure 16-64. Time series of atmosplieric extinction coefficient (I-rein averages) measured at the 57-m(A-200), 48-m(A- 150), 29-m(A-100) and 16-m
(A-SO)levels of tower A and at the 3-m (x) level of site x on 5 Novemkr 1978 during fog.

u = 128W. (16.37)

where u is in l/km and W in g/m3. The Eldridge relationship Since the relationship is unique to all fogs, one can
was derived from spniewhat questionable drop-size data @d determine the extinction coefficient at a wavelength of 11
would probably hold true only in a highly polluted, foggy pm from the LWC measurements and vice versa.
environment sqph as might occur Dear an industrial tia.
Kunkel [1982], using 1525 drop-size distributions from 16.4.2.4 Modeling o! Visibility Distributions. The
eight advection fog cases, @veal at the following foimula: Weibull Distribution has been used for visibility

rJ = 155 W094 (16.36) 6(x) = 1 – exp( – ax ‘), (16.38)

He found, however, that me coefficient varied frop 110 in where x stands for visibility, and a and ~ are parameters.
a large-droplet fog with drizzle to 253 in a small-droplet Estimates for a and ~ We made to minimize the sums of
fog and the expohent varied from 0.96 to 1.14. For all squares of the differences between the model estimates ~(x)
practical purposes a linear relationship can be assumed; but and ~he cumulative frequencies
. . . F(x) of the climatic sum-
the slope will de~nd on the drop size distribution. maries.
Chylek [1978] found that at a wavelength of 11’pm At the University of Ceptral Florida [Somerville et al.,
there is ‘an unambiguous relation between the extinction 1979] sets of values for a and ~ have been determined for
coefficient and the liquid water content, independent of the each of the eight periods of the day in each month at some
size distribution, in fogs with droplets up to 28 ~m diameter. 22 stations around the world, to make the estimated cu-
That relatiomhip is as follows: mulative distributions ~(x) fit the data [or some 15 visibil-

16-54
 WATER VAPOR, PRECIPITATION, CLOUDS, AND FOG

A 200 ❑ AIL A A 100 + A50 )( X x 8NO04

360- “

,
TIME (GMT)

Figure 16-65. Time series of atmospheric extinction coefficient at the same levels as shown in Figure 16-64 (observations taken on 4 November 1978
during ground fog).

Table 16-37. A sample of cumulative frequency distribution of visibility ities (in kilometers) corresponding to those in the climatic
in kilometers. (The station is-Bedford, Mass., January noon- summaries.
time. Probability estimates were made by the Weibull
distribution, with parameter values a = 0.04679,
A simpler method for solving for Q and ~ utilizes the
~ = 0.81856.) “ linear relationship between (n( – (n[ 1 – F(x)]) and tnx.
This permits a solution for ~ and end by linear regression.
Visibility RUSSWO Model As an example, at Bedford, Mass., in January at noontime,
(km) Frequency Probability w and ~ were found to be 0.04679 and 0.81856, respec-
0.4 0.018 0.022 tively; the resulting visibility probabilities are compared
0.5 0.028 0.026 with RUSSWO data in Table 16-37.
0.8 0.034 0.038
1.0 0.053 0.046
1.2 0.056 0.053
1.6 0.072 0.067
2.0 0.094 0.080
2.4 0.099 0.W2
3.2 0.116 0.115
4.0 0.135 0.136
4.8 0.146 0.156
6.4 0.184 0.193
8.0 0.221 0.227
9.7 0.254 0.259
16 0.377 0.365

16-55
CHAPTER 16
REFERENCES

Air Weather Services, “Forecasters’ Guide on Aircraft Ic- Brewer, A. W., “Evidence for a World Circulation Provided
ing,’’AWS-TR-80-OOOl Scott AFB, Ill., 1980. by Measurement of Helium and Water Vapor Distri-
Al lard, E., Memoire sur L‘ intensite et la Portee des Phares, bution in the Stratosphere,” Quart. J. Roy. Meteorol.
Paris, Dunod, 1876, quoted in Middleton, W. E., Vision SOC., 75:351-363, 1949.
Through the Atmosphere, University of Toronto Press, Brown, H. A., “Preliminary Assessment of an Automated
1963. System for Detecting Present Weather,” AFGL TR-79-
Allen, M., Y.L. Yung, and J.W. Waters, ’’Vertical Trans- 0137, ADA07831, 1979.
port and Photochemistry in the Terrestrial Mesosphere Bundy, D. S., “A Two-Year Study of the Probability of
and Lower Thermosphere (50–1 20 km),” J. Geophys. Cloud-Free Lines-of-Sight at Columbia, Mo. ,“ Unpub-
Res., 86: 3617–3627, 1981. lished Masters Dissertation, University of Missouri, Co-
Anonymous, “Lowest Measured Temperature in the Earth’s lumbia, 1969.
Atmosphere Recorded in U.S. - Sweden Experiments,” Bussey, H. E., “Microwave Attenuation Statistics Estimated
Bull. Am. Meteorol. Sot., 44:806, 1963. from Rainfall and Water Vapor Statistics,” Proc. lRE,
Appleman, H. S., “Occurrence and Forecasting of Cirro- 38:781, 1950.
stratus Clouds ,“ Tech. Note 140, World Meteorological Changnon, S. A., “Hailstreaks,” J. Atmos. Sci., 27:109-125,
Organization (WMO), 1961. 1970.
aufm Kampe, H. J., and H .K. Weickmann, “Physics of Changnon, S .A., “Means of Estimating AreaI Hail-Day Fre-
Clouds,” Meteoro[. Monographs, 3:182, 1957. quencies,” J. Wea. Mod@ca., 3:154-158, 1971.
Bannon, B.A. and J .P. Steele, “Average Water Vapor Con- Chisholm, D. A., “The Mesoscale Variability of Visibility
tent of the Air,” Geophys. Memoirs No. 102, Meteor- in Time and Space,” Proceedings 5th Conference on
ological Office, Air Ministry, London, 1960. Weather Forecasting and Analysis, St. Louis, Mo.,
Barnes, A. A., “New Cloud Physics Requirements,” Pre- American Meteorological Society, Boston, Mass., 1974.
prints 4th Symp. on Meteorological Observations and Chisholm, D.A. and H. Kruse, “The Variability of Airfield
Instrumentation, American Meteorological Society, Visibility: A Preliminary Assessment ,“ AFCRL TR-74-
Boston, Mass., 1978. 0027, AD777219, 1974a.
Barnes, A. A., “Observations of Ice Particles in Clear Air,” Chisholm, D.A. and H. Kruse, “The Variability of Airfield
Journal de Recherches Atmospheriques 14:3 11–3 15, Visibility: A Preliminary Assessment ,“ AFCRL TR-74-
1980. 0265, AD784791 , 1974b.
Barnes, A. A., “Ice Particles in Clear Air,” Preprints 8th Chylek, P., “Extinction and Liquid Water Content of Fogs
Internatioml Conf. on Cloud Physics, Clermont-Fennad, and Clouds,” Y. Atmos. Sci., 35:290-300, 1978.
France, 15–19 July 1980, pp 189–190, AFGL TR-81- Cohen, I. D., “Development of a Large Scale Cloud System,
0009, AD094444, 1981. 23–27 Mar78,” AFGLTR-81-0127, ADA106417, 1981a.
Barnes, A. A., “The Cirrus and Subvisible Cirrus Back- Cohen, I. D., “Cirrus Particle Distribution Study, Part 8,”
ground,” Preprints 2nd Symp. on Composition of Non- AFGL TR-81-0316, ADA118715, 1981b.
Urban Troposphere, American Geophysical Union, Cole, A. E., R.J. Donaldson, D. Atlas, A.J. Kantor, R.K.
Williamsburg, Va., May 1982. Soberman, and J.E. Manson, “Precipitation, Clouds,
Bean, S .J., P.N. Somerville, M. Heuser, “Some Models and Aerosols ,“ Chapter 5, Handbook of Geophysics and
for Ceiling,” AFGL TR-79-0221, ADA078033, 1979. Space Environments, AFCRL, 1965.
Beckwith, W .B., “Analysis of Hailstorms in the Denver Cole, A.E. and A.J. Kantor, “Air Force Reference Atmos-
Network, 1949–1958 ,“ Physics of Precipitation, edited pheres, “ AFGL TR-78-0051, AD058505, 1978.
by H. Weidkman, American Geophysical Union, Wash- Crane, A. K., “Prediction of Attenuation by Rain,” IEEE
ington, D. C., 348–353, 1960. Trans. Comm. COM-28: 1717–1733, 1980.
Bennett, 1., “Glaze, Its Meteorology and Climatology, Geo- Crutcher, H.L. and J.M. Meserve, “Selective Level Heights,
graphic Distribution and Economic Effects,” Tech. Rept. Temperatures and Dew Points for the Northern Hemi-
EP- 105, Quartermaster Research and Engineering Cen- sphere ,“ Naval Weather Service Command, NAVAIR
ter, Envir. Prot. Res. Div., Natick, Mass., 1959. 50-IC-52, U.S. Government Printing Office, Washing-
Bertoni, E. A., “Clear Lines-of-Sight from Aircraft,” AFCRL ton, D. C., 1970.
67-0435, AD657801 , 1967. Crutzen, P. A., “A Review of Upper Atmospheric Photo-
Berton.i, E. A., “Clear and Cloud-Free Lines-of-Sight from chemistry,” Can. J. Chem., 52: 1569–1581, 1974.
Aircraft ,“ AFGL TR-77-O141, ADA046352, 1977a. Davis, A.R. and D.J. McMorrow, “Stochastic Models for
Bertoni, E. A., “Clear and Cloud-Free Lines-of-Sight from Deriving Instantaneous Precipitation Rate Distribu-
Aircraft,” AFGL TR-77-0141, ADC 015057 1977b. tions,” USAF Air Weather Service, AWS-TR-76-263,
Best, A. C., “The Size Distribution of Raindrops,” Quart. 1976.
J. Roy. Meteorol. Sot., 76: 16, 1950. deBary, E. and F. Moller, “The Vertical Distribution of
Boyd, D. W., “Maximum Snow Depths and Snow Loads Clouds,” J. Appl. Meteorol., 2:806-808, 1963.
on Roofs in Canada,” Research Paper 142, Div. of Deirmendjian, D., “Scattering and Polarization Properties
Building Research, National Research Council, Ottawa, of Water Clouds and Hazes in the Visible and Infrared,”
Canada. 1961. Appf. Opt., 3: 187–196, 1964.

16-56
 -

WATER VAPOR, PRECIPITATION, CLOUDS, AND FOG

Derr, V. E., “A Basic Comparison of Lidar and Radar for Gunn, K.L.S. and J. S. Marshall, “The Distribution with
Remote Sensing of Clouds,” Tech. Rept. ERL 397-WPL Size of Aggregate Snow flakes,” J. Meteorol., 15:542,
52, Wave Propagation Laboratory, NOAA, Boulder, 1958.
Colo. , 1978. Gutnick, M., “How Dry is the Sky?”, J. Geophys. Res.,
Deguchi, S. and D.O. Muhleman, “Mesospheric Water Va- 66:2867-2871, 1961.
por,” J. Geophys. Res., 87: 1343–1346, 1982. Hage, N. and H. Entrekin, “Further Studies of Space and
Ehhalt, D. H., L.E. Heidt, R.H. Lueb, and E.A. Martell, Time Variations in Atmospheric Transmission Along
‘{concentrations of CH~, CO, COZ, Hz, H20,andN02 Airport Runway s,” TR74 17-143, The Travelers Re-
in the Upper Stratosphere ,“ J. Atmos. Sci., 32: 163–169, search Center, Inc., Hartford, Corm., 1965.
1975. Hage, N. and J. Wilson, “Analyses of the Variability of
Eldridge, R. G., “Haze and Fog Aerosol Distributions,” J. Clouds, Fog, Wind and Rain in the Atlantic City Me-
Amos. Sci., 23:605-613, 1966. sonetwork,”TN7481 -262, The Travelers Research Cen-
Falls, L. W., “The Beta Distribution: A Statistical Model ter, Inc., Hartford, Corm., 1967. ,
for World Cloud Cover,” NASA TM X-64714, Marshall Hall, G. R., “On the Mechanics of Transition Produced by
Space Flight Center, Ala., 1973. Particles Passing Through on Initially Laminar Bound-
Farmer, C. B., O.F. Raper, B.D. Robbins, R.A. Toth, and ary Layer and the Estimated Effect on the LFC (Laminar
C. Muller, “Simultaneous Spectroscopic Measurements Flow Control) Performance of the X-21 Aircraft ,“
of Stratospheric Species 02, CH4, CO, COZ, NZO, HZO, Northrop Aircraft Report, October 1964.
HCL, and HF at Northern and Southern Mid-Latitudes,” Hall, W.D. and H.R. Prupacher, “The Survival of Ice Par-
J. Geophys. Res., 85:1621-1632, 1980. ticles Falling from Cirrus Clouds in Saturated Air,” J.
Feldman, N. E., “Rain Attenuation Over Earth-Satellite Atmos. Sci., 33: 1995–2006, 1976.
Paths,” Science Applications, Inc., Final Report on Htis, J. E., “The Distribution of Water Vapor in the Strat-
Contract NOO039-79-C-0136 for Naval Electronic Sys- osphere,” Rev. Geophys. Space Phys., 14: 565–575,
tems Command, ADA075390, 1979. 1976.
Fett, R.W., J.J. Bates, W.A. Bohan, R.J. Englebretson, Haurwitz, B., Dynamic Meteorology, McGraw-Hill, New
J. Rosenthal, and G.A. Stevenson, Navy Tactical York, 1941.
Applications Guide, Vol. 3, North Atlantic and Heath, E. D., and L.M. Cantrell, “Aircraft Icing Climatol-
Mediterranean Weather Analysis and Forecast Applica- ogy for the Northern Hemisphere ,“ Tech. Rept. 220,
tions, NEPRF Tech. Report 80-07, Monterey, Calif., Air Weather Service (MAC) USAF, 1972.
1981. Herbstritt, R. L., “Rain Climate Considerations for Domes-
Frederick, R. H., V.A. Meyers, and E.P. Auciello,4’Five- tic Satellites—Continental U.S. ,“ Report R-7301, FCC,
to 60-Minute Precipitation Frequency for the Eastern Washington, D. C., 1973.
and Central United States,” NOAA Tech. Memo. NWS Herman, B. M., “Infrared Absorption, Scattering and Total
HYDRO-35 , 1977. Attenuation Cross-Sections for Water Spheres,” Quart.
George, D. M., and M. Lefkowitz, “A New Concept: Sensor J. Roy. Meteorol. Sot., 88: 153–250, 1962.
Equivalent Visibility,” Proceedings International Con- Heymsfield, A. H., “Cirrus Unicus Generating Cells and the
ference on Aerospace and Aeronautical Meteorology, Evolution of Cirriform Clouds, Part 1. Aircraft Obser-
Washington, D. C., American Meteorological Society, vations of the Growth of the Ice Phase,” J. Atmos. Sci.,
1972. 32:799-808, 1975.
Glass, M. and D.D. Grantham, “Response of Cloud Mi- Ho, F.P. and J.T. Riedel, “Precipitable Water over the
crophysical Instruments to Aircraft Icing Conditions,” United States, Vol. 11, Semimonthly Mariver,” U.S.
AFGL TR-81-0192, ADA112317, 1981. Dept. of Commerce, NOAA/NWS, Silver Springs, Md.,
Grantham, D. D., I.A. Lund, and R.E. Davis, “Estimating 1979.
the Probability of Cloud-Free Fields-of-View Between Huschke, R.E., H.H. Bailey, R.J. Nelson, and J.P. Bibbo,
Earth and Airborne or Space Platforms,” Air Weather “An Improved Weather-Attenuation Model for Earth-
Service Tech. Rept. 79-001, 1979. Satellite Communication Links,” Science Applications
Gringorten, 1.1., “Hailstone Extremes for Design,” AFCRL Inc., Final Report on Contract DAAK80-80-C-O049,
72-0081, AD743831, 1971. U.S. Army Satellite Communications Agency, Ft. Mon-
Gringorten, 1.1., “Probability Models of Weather Conditions mouth, N. J., 1982.
Occupying a Line or an Area,” J. Appl. Meteorol., 18: Jiusto, J,E. and G.G. Lala, “Thermodynamics of Radia-
957-977, 1979. tion Fog Formation and Dissipation—A Case Study,”
Gringorten, 1.1., “Climatic Probabilities of the Vertical Proceedings 8th International Cloud Physics Con-
Distribution of Cloud Cover,” AFGL-TR-82-0078, ference, Clermont-Ferrand, France, 15–19 June
ADAl 18753, 1982. 1980.
Gringorten, 1.1., H.A. Salmela, 1. Solomon, and J. Sharp, Jones, M.A. and A.L. Sims, “Climatology of Instantaneous
“Atmosphere Humidity Atlas-Northern Hemisphere,” Precipitation Rates,” AFCRL TR-72-0429, AD748283,
AFCRL 66-621, AD642429, 1966. 1971.
Gunn, K.L.S. and T. W.R. East, “The Microwave prop- Jones, M.A. and A. L. Sims, “Climatology of Instanta-
erties of Precipitation Particles,” Quart. J. Roy. Mete- neous Rainfall Rates,” J. Appl Meteoro/., 17: 1135–1140,
orol. Sot., 80: 522–545, 1954. 1978.

16-57
CHAPTER 16
Junge, C. E., Air Chemistry and Radioactivity, Academic Lund, I. A., “Haze Free and Cloud-Free Lines-of-Sight
Press, New York, 1963. Through the Atmosphere,” AFCRL 72-0540, AD75 1264,
Kantor, A.J. and D.D. Grantham, “A Climatology of Very 1972.
High Altitude Radar Precipitation Echoes,” AFCRL 68- Lund, I. A., “Joint Probabilities of Cloud-Free Lines-of-
0630, AD68 1107, 1968. Sight Through the Atmosphere at Grand Forks, Fargo,
Kikuchi, K. and A. W. Hogan, “Properties of Diamond Dust and Minot, N. Dak. ,“ AFCRL TR-73-O 178, AD763083,
Type Ice Crystals Observed in the Summer Season at 1973.
.4mundsen-Scott South Pole Station, Antarctica,” J. Me- Lund, I.A. and M.D. Shanklin, “Photogrammetrically
teorol. Sot. of~apan, 57: 180–190, 1979. Determined Cloud-Free Lines-of-Sight Through the
Knollenberg, R. G., “Comparative Liquid Water Content Atmosphere,” J. Appl. Meteoro[., 11: 773–782,
Measurements of Conventional Instruments with an Op- 1972.
tical Array Spectrometers,” J. Appl. Meteorol.,11: Lund, I.A. and M.D. Shanklin, “Universal Methods for
501–508, 1972. Estimating Probabilities of Cloud-Free Lines-of-Sight
Koschmieder, H., “Theorie Der Horizontalen Sichtewite,” Mugh the Atrnosphe~,” J. Appl. Meteorol., 12: 28–35,
Beitr. Phys. Atmos., 12: 33–53, 1924. (Cited in Mid- 1973.
dleton, W .E.K., Vision Through the Atmosphere, Uni- Lund, I.A. and E.A. Bertoni, “Some Comparisons Between
versit y of Toronto Press, 1963. ) Probabilities of Cloud-Free Line-of-Sight Estimated
Kunkel, B., “Comparison of Fog Drop-Size Spectra Mea- from Aircraft and from Sky Cover Observations,” AFGL
sured by Light Scattering and Impaction Techniques ,“ TR-80-0046, ADC02 1847, 1980.
AFGLTR-81 -0049, ADA1OO252, 1981. Lund, I. A., D.D. Grantham, and R.E. Davis, “Estimating
Kunkel, B., “Microphysical Properties of Fog at Otis AFB, Probabilities of Cloud-Free Fields-of-View from the Earth
MA,” AFGL TR-82-0026, ADAl 19928, 1982. Through the Atmosphere,” J. App. Meteorof., 19:
Lamb, K., K.W. Nielson, H.E. Klieforth, and J. Hallett, 452463, 1980.
“Measurements of Liquid Water Content in Winter Con- Lund, I. A., D.D. Grantham, and C.B. Elam Jr., “Atlas of
ditions Over the Sierra Nevadas,” J. Appl. Meteorol., Cloud-F~ Line-of-Sight Probabilities”, in several parts
15: 763–775, 1976. Part 1, Germany, AFCRL TR-75-0261, ADAO1 3784,
Landsberg, H., “Climatology,” Chapter XII, in Handbook 1975. Part 2, U. S. S.R., AFGL TR-77-~5,
of Meteorology, 1945. ADA040705, 1976. Part 3, U. S.A., AFGL TR-77-0188,
Landsberg, H. E., World Survey of Climatology, Vols. ADA0511 12, 1977. Part 4, Europe, AFGL TR-78-0276,
I–XV, Elsevier, Amsterdam, The Netherlands, 1974. ADA065 167, 1978. Part 5, North Africa and the Middle
Lenhard, R. W., “Precipitation Intensity and Extent,” East, AFGL TR-79-0275, ADA088 156, 1979.
Journal de Recherches Atmospheriques, 8: 375–384, Marshall, J.S. and W. M. Palmer, “The Distribution of
1974. Raindrops with Size,” J. Meteorol., 5:165, 1948.
Lenhard, R. W. and N. Sissenwine, “Extremes of 1, 12 and Mason, B. J., The Physics of Clouds, 2nd Ed., pp 113–121,
24 Hour Rain for MIL-STD-210B ,“ AFCRL TR-73- Clarendon Press, Oxford, England, 1971.
0329, AD76621O, 1973. McKay, G.A. and H.A. Thompson, “Estimating the Hazard
Levin, L. M., “size Distribution Function for Cloud of Ice Accretion in Canada from Climatological Data,”
Droplets and Raindrops,” TR-263R, DSNS, Can- J. Appl. Meteorol., 8: 927–935, 1969.
ada, 1954. McTaggert-Cowan, J. D., G.G. Lala, and B. Vonnegut,
Lewis, W., “Meteorological Aspects of Aircraft Icing,” “The Design, Construction and Use of an Ice Crystal
Compendium of Meteorology, American Meteorol. So- Counter for Ice Crystal Studies by Aircraft,” J. Appl.
ciety, Boston, Mass., 1951. Meteorol., 9:294-299, 1970.
Lhermitte, R. M., “Millimeter Wave Doppler Radar,” Pre- Michaels, D. W., “Noctilucent Cloud Research,” Foreign
prints 20th Conf. Radar Meteorol., American Meteorol. Sci. Bull., 6:5 1–55, Library of Congress, 1965.
Society, Boston, Mass., 1981. Muench, H. S., E.Y. Moroz, and L.P. Jacobs, “Develop-
List, R.J. (cd.), Smithsonian Meteorological Tables, Smith- ment and Calibration of the Forward-Scatter-Meter,”
sonian Institution Press, Washington, D. C., 1968. AFCRL TR-74-0145 , AD4783270, 1974.
Litvinov, 1.V., “The Function of the Distribution of Par- NOC (Naval Oceanography Command), Guide to Standard
ticles of Liquid Precipitates,” IZV. Acad. Sci.—Geo- Weather Summaries and Climatic Services, NAVAIR
phvs., 12: 1474–1483, 1956. 50-l C-534, Asheville, N. C., 1980.
Lo, K.K. and R.E. Passarelli, “The Growth of Snow in Pasqualucci, F., B. W. Bertram, R.A. Kropfli, and W .R.
Winter Storms: An Airborne Observational Study,” J. Moninger, “A Millimeter Wavelength Dual-Polarization
Atmos. Sci., 39: 697–706, 1982. Doppler Radar for Cloud and Precipitation Studies,” J.
Long, M. J., H. H. Hanks, and R.G. Beebe, “Tropopause Clim. Appl. Meteorol., 22: 758–765, 1983.
Penetrations by Cumulonimbus,” AFCRL 65-561, Pasqualucci, F. and L.J. Miller, “Dual Wavelength Radar
AD621573, 1965. Observation in Clouds and Precipitation,” Preprints 20th
Lott, G. A., Precipitab[e Water Over the United States, Vol. Conference Radar Meteorology, American Meteorol.
1, Monthly Means, U.S. Dept. of Commerce, Society, Boston, 1981.
NOAA/NWS, Silver Springs, Md., 1976. Penndorf, R., “Analysis of Ozone and Water Vapor Field
Ludlum, D.M. (cd.), “Extremes of Snowfall; United States Measurement Data,” FAA Tech. Rept., FAA-AEE-78-
and Canada,” Weatherwise, 23:286-294, 1970. 29, ADA072721, NOV 1978.

16-58
 WATER VAPOR, PRECIPITATION, CLOUDS, AND FOG
Perkins, P. J., Proceedings, Second Atinual Workshop on Somerville, P. N., and S.J. Bean, “A New Model for Sky
Meteorological and Environmental Inputs to Aviation Cover,” AFGL-TR-79-02 19, ADA078368, 1979.
Systems, NASA Conf. Publication 2057: 85–99, 1978. Somerville, P. N., S.J. Bean, “and S. Falls, “Some Models
Plank, V. G., “Hydrometer Data and Analytical-Theoret- for Visibility,” AFGL-TR-79-O 144, ADA075490, 1979.
ical Investigations Pertaining to the SAMS Missile Flights Stickel, P. G., and T.A. Seliga, “Cloud Liquid Water Con-
of the 1972–73 Seasons at Wallops Island, Virginia,” tent Comparisons in Rain Using Radar Differential Re-
AFGL TR-77-0149, ADA05 1192, 1977. flectivity Measurements and Aircraft Measurements,”
Platt, C. M. R., “Transmission of Submillimeter Waves Preprints 10th Conference Radar Meteorology, Amer-
firough Water Clouds and Fog,” J. Atmos. Sci., 27: ican Meteotol. Society, Boston, 30 Nov-3 Dec 1981.
421425, 1970. Sulakvelidze, G. K., N.Sh. Bibilashvili, and V.F. Lap-
Pocliop, L.O. and M.D. Shanklin, “Sky Cover Photograms, cheva, “Formation of Precipitation afid Modification of
A New Technique,” Weatherwise, 19: 198–200, 1966. Hail Process,” Gidrometeorologicheskoe Izdatel’-stvo,
Rapp, R. R., C. Schutz, and E. Ronriques, “Cloud-Fr:e Leningrad, Translation by Israel Program for Scientific
Line-of-Sight Calculations,” J. Appl. Meteorol., 12: Translations, Tel-Aviv (No. 1806) 1967.
484493, 1973. Suring, R., Die Wolken, 2nd Ed., Leipzig, 1941.
Riordan, P., “Weathpr Extremes Around the World,” U.S. Taljaard, J. J., H. vanLoon, H.L. Crutcher, and R.L. Jenne,
Army Engineer Topographic Laboratories, Ft. ‘Belvoir, Climate of the Upper Air-Southern Hemisphere, Vol.
Va., Report ETL-TI?-74-5, 1974. 1, Temperatures, Dew Points atid Heights Selected Lev-
Riordan, P., “’Extr&rne 24-Hour Snowfalls in the United els, U.S. Dept. of Commerce, ESSA, Environmental
States: Accu.mylation, Dis@bution, and Frequency,” U.S. Data SeVice, NAVAIR 50-IC-55, 1969.
Army Engineer Topographic Laboratories, Ft. Belvoir, Tattelman, P., “An Objective Method for Measuring Sur-
Va., Special Report ETL-SR-78-4, 1973. face Ice Accretion,” J. Appl. Meteorol., 21:153-166,
Roach, W .T., “On the Nature of the Summit Areas of Se- 1982.
vere “Storms in Oklahoma,” Quart. J. Roy. Meteorol. Tattelman, P. and D.D. Grantham, “A Survey of Tech-
SOC., 93: 318–336, 1967. niques for Estimating Short-Duration Precipitation Rate
Rogers, J. W., A.T. Stair, T.C. Degges, C.L. Wyatt, and Statistics,” AFGL-TR-82-0357, ADA125705, 1982.
K.D. Baker, “Rocketbome Measurement of Mesos- Tattelman, P. and 1.1. Gringorten, “Estimated Glaze Ice and
pheric H*O in the Auroral Zone,” Geophys. Res. Lett., Wind Loads at the Earth’s Surface for the Contiguous
4: 366–368, 197,7. United States,” AFCRL-TR-73-0646, AD775068, 1973.
Roys, G.P. and E. Kessler, “Measurements by Aircrafi of Tattelman, P. and K. Scharr, “A Model for Estimating 1-
Condensed W“ater in Great Plains Thunderstorms,” NSSL Minute Rainfall Rates,” J. Climate Appl. Meteorol., 22:
Tech. Note 49, ESSA, U.S. Dept. of Commerce, 1966. 1575–1580, 1983.
Ryan, R. T., H.H. Blau, P.C. Von Thuna, M.L. Cohen, Thomwaite, C. W., J.R. Mather, D.B. Carter, and C.E.
and G .D. Roberts, “Cloud Microstructure as Determined Molineux, “Estimating Soil Moisture and Tractionabil-
by an OpticaI Cloud Particle Spectrometer,” J. Appl. ity Conditions for Strategic Planning, ” AFCRC TN-58-
Meteorol., 11: 149–156, 1972. 201 (1 and 2), AD146789, 1958.
Schilling, G. F., “Theoretically Permissible Altitudes and U.S. Dept. of Defense, “Military Standard Climatic Ex-
Seasons for the Occurrence of Clouds Near the Meno- tremes for Military Equipment,” MIL-STD-21OB, U.S.
pause,” Rand Corp. Paper P-2899, AD435637, 1964. Govemnlent Printing Office, 1973.
Schlatter, E. and M. Lefkowitz, “Evaluation ,of Multi- U.S. Weather Bureau, “Rainfall Intensity -Duration-Fre-
Transmisometer Systems,” RD-68-49, ESSA-Weather quency Curves,” Tech. Paper 25, 1955.
Bureau, Atlantic City, N. J., 1968. , U.S. Weather Bureau, Determination of Snow Loa& for
Shanklin, M.D. and J .B. Landweh~, “Photogrammetrically Building Construction, Div. of Climatological afid Hy-
Determined Cloud-Free Lines-of-Sight. at Columbia, drological Services, Washington, D. C., 1951.
Mo. ,“ Final Report under Contract F19628-68-C-0140, Varley, D. J., I.D. Cohen, and A.A. Barnes, “Cirrus Particle
University of Missouri, AFCRL TR-71-0273 AD725758, Distribution Study, Part 7,” AFGL-TR-80-0324,
1971. ADA1OO269, 1980.
Scholz, T. G., D.H. Ehhalt, L.E. Heidt, and E.A. M~ell, Waters, J. W., J.J. Gustincic, R.K. Kakar, A.R. Kerr, H.K.
“Water Vapor, Molecular Hydrogen, Methane, and Trit- Rosc&, and P.N. Swanson, “The Microwave Limb
ium Concentrations Near the Stratopause,” J. Geophys. Sounder Experiment: Observations of Stratospheric and
Res., 75:3049-3054, 1970. Mesospheric H20 in Interhemisphetic Survey of Minor
Sims, A.L. and D.M.A. Jones, “Climatology of Instanta- Upper Atmospheric Constituents during October-No-
neous Precipitation Rates ,“ AFCRL TR-73-O171, vember 1976,” NASA Tech. Memo TMX 73630, 1977.
AD760785 , March 1973. Weiss, R. R., J.D. Locatelli, and P. V. Hobbs, “Simulta-
Sissenwine, N. and R. Cormierj “Synopsis of Background neous Observations of Cloud and Precipitation Particles
Material for MIL-STD-2 IOB, Climatic Extremes for with Vertically Pointing X-Band and Ka-Band Radars,”
Military Equipment,” AFCRL-TR-74-0056, AD780508, IEEE Trans. Geosci. Electr., V, GE-17: 151, 1979.
1974. Winner, D. C., “Climatological Estimates of Clock-Hour
Soberman, R. K., “Noctilucent Clouds,” Sci. Am., pp 5 1–59, Rainfall Rates,” Air Weather Service Tech. Rept. 202,
June 1963. 1968.

16-59
 CHAPTER 17

WINDS
Section 17.1 D.D. Grantham and 1.1. Gringorten
Section 17.2 and 17.3 A.E. Cole and A.J. Kantor
Section 17.4 1.1. Gringorten and P.I. Tattelman
Section 17.5 F.A. Marcos

The atmosphere’s motions defy rigorous classification the surface. Once established, however, the model formula
or modeling. The application of a particular feature of wind can be used to describe wind speed and gusts along a vertical
structure to a given engineering problem should be dictated profile in the lowest 50 or 100 m of the atmosphere.
by its physical dimensions. Extrapolation of data beyond its Wind flows in response to pressure gradients in the at-
indicated limitations is risky. Local conditions may not al- mosphere. Such pressure gradients change slowly with al-
ways be well represented by the data in this chapter, and titude, negligibly within the first 100 m. Yet the changes
may produce extreme wind variations in excess of those in wind speed with height are pronounced. Air motion near
presented, even when the local wind structure is free from the surface does not obey the simple pressure gradient law.
perturbations such as fronts, thunderstorms, or squalls. The Anemometer near the ground may be hardly turning, whereas
practicing engineer should avail himself of applicable local those on tall buildings or towers may show moderately strong,
meteorological records whenever possible. Information other gusty winds. Kites may be difficult to launch, but once
than that given below is available through the various na- several hundred feet high, they may fly without difficulty.
tional weather services and the World Meteorological Or- Friction caused by terrain is one of the main factors
ganization [NOCD, 1980]. Special studies prepared by En- affecting the horizontal wind speed up to an altitude known
vironmental Technical Applications Center (ETAC) or Air as the gradient level. At this height, 300 or 600 m, the
Force Geophysics Laboratory (AFGL) might provide the pressure gradient is said to be dynamically balanced against
best answers for specific design problems. two other influences: the earth’s rotation and the curvature
The measure of wind, in speed and direction, presents of the wind path. A theoretical wind speed that closely
an immediate problem in the time interval for the obser- approximates the observed wind at gradient height can be
vation: A conventional observation of wind speed is the computed from the isobaric spacing and curvature on surface
wind travel in 1 rein, 5 rein, or 1 h, that is, the l-rein, weather maps.
5-rein, or 1-h average. The current standard averaging time The height of the gradient level and the velocity profile
period in the United States is 1 min. In England and Canada up to that level vary greatly, mainly with the type of surface
a 10-rein period is customary when wind speed records are and the stability of the air. Stability is chiefly a function of
available, otherwise the averaging period is something over the temperature structure in the boundary layer. One extreme
15 s. However, in published climatic data, hourly (60 rein) of temperature structure is represented by a super adiabatic
averaged winds are often given. lapse rate in which temperature decreases rapidly with al-
As might be expected, the variance of the 1-rein wind titude so that air displaced upward will continue upward
is greater than the 5-rein, which in turn is greater than that because it is warmer than its surroundings. The opposite
of the 1-h wind, but not significantly so. On the other hand, extreme is a negative lapse rate or inversion, in which tem-
wind speeds of shorter duration than 1 min are subject to perature increases with height, so that air displaced upward
significantly greater variability. When the wind speeds peak is cooler than its surroundings and tends to sink back to its
between lulls, in 20s or less, they are conventionally termed original level. A neutral condition (adiabatic) exists when
gusts. the temperature lapse rate is such that a parcel of air dis-
placed vertically will experience no buoyant acceleration.
In general, a neutral lapse rate is established by the turbulent
17.1 WIND AS A FUNCTION OF HEIGHT mixing caused by strong winds at the surface.
A popular model for the shift of wind speed and direction
A major problem with pooling the various surface wind with altitude in the boundary layer, originally developed for
observations is in the determination of the best method or ocean depth, is termed the Ekman Spiral. Since its intro-
model for adjusting wind speeds to a common height abov.: duction, micrometeorologists have studied the energy trans-

17-1
CHAPTER 17
Table 17-1. Ratio of wind speed at height H to speed at 91 m over open Table 17-2. Typical values of the roughness parameter, H,,
prairie obtained during [be Great Plains Turbulence Field
Type of Surface Ho (cm)
1
VI,IV9, Smooth (mud flats, ice) 0.0009
H Lapse Rate Inversion Isothermal Lawn, grass up to 1 cm 0.09
(meters) (dt/dz < O) (dt/dz > O) (dt/dz = O) Downland, thin grass up to 10 cm 0.61
Thick grass, up to 10 cm 2.25
91 1.000 1.000 1.000
Thin grass, up to 240 cm 4.9
30 0.965 0.689 0.872
Thick grass, up to 240 cm 9.1
21 0.944 0.608 0.829
15 0.915 0.538 0.792
9 0.866 0.452 0.733
6 0.825 0.403 0.686 (17.2)
3 0.749 0.339 0.604
1.5 0.662 ~.275 0.518
0.6 0.556 0.231 0.424 where the exponent p depends on the height, terrain, thermal
0.3 0.470 0.200 0.336 stratification and speed of the overall aifiow. The parameter
0.15 0.383 0.166 0.300 p is larger for rough ground, for altitudes below H,, and
for relatively small V,. It varies within limits, approximately
0.05 to 0.8, and averages between 0.1 and 0.3. Table
17-4 lists mean values of p, determined for several locations
fer and diffusion phenomena in the boundary layer and have
and two types of terrain. Table 17-5 lists mean values of p
found improved empirical relationships to fit wind speed
that were determined for typical daytime (dT/dz < O) night-
data gathered at various heights above ground.Table 17-1
time (dT/dz > O), and isothermal conditions in the lowest
shows the ratio of wind speed at various heights to that at
9 mover open prairie country (Great Plains Turbulence Field
91 m over open prairies. Values are based on actual wind
Program). The exponent p is larger when there is a stabi-
measurements taken during the Great Plains Turbulence Field
lizing inversion and smaller when there is a positive lapse
Program conducted at O’Neill Nebraska [Lettau and Dav-
rate. According to Sherlock [1952] a typical value for p is
idson, 19571. The ratios are shown for typical daytime lapse
1/7 (or O. 143). Early workers had recognized that his p-
rates (dT/dz < O), night-time (dT/dz- > O) and for iso-
value was applicable to typical steady or mean winds but
thermal conditions (dT/dz = 0).
not applicable to gustiness. Sherlock [1952] noted that gusts
were better described with a value of p = 0.0625. Shellard
[1965], in reducing high wind speeds and gusts to a common
17.1.1 Variation of Wind Speed with Height height of 10 m, used Equation (17.2) with p = 0.17 for
(Lowest 100 m) mean speeds and 0.085 for gusts.
A special study was made with data from the Windy
Two alternative classes of models have been used to Acres Project [Izumi, 1971] consisting of 39 h of 1-s wind
estimate the increase of wind speed with height: logarithmic
and power models.
In one logarithmic model Table 17-3. ,.. H to wind at height H, for [w{>
Ratios of wind at height G.

reference levels and three roughness parameters HI), calcu-

lated by Equation ( 17.1).
v = /n(l + H) ,/n(l + H,)
(17. i) H
< H. HO
(m) HO = 30 cm HO=3cm Ho = ().3 Cm

V is the mean wind speed at height H, V, is the mean wind V/Vl for HI = 6 m
speed at the reference level H, (anemometer level), and Ho 100 1.91 1.53 1.37
is the roughness parameter a length dete~ined by the char- 10 1.16 1.10 1.07
acteristic ground surface. The boundary condition at H = O 1 0.48 0.67 0.76
is V = O. This model has the advantage that the effect of Q.5 0.32 0.54 0.67
terrain is included explicity. Typical values for Ho are given
V/V[ for H1 = 3 m
in Table 17-2. Table 17-3 lists V/VI for a variety of rough-
100 2.42 1.76 1.51
ness parameters HO, for two reference levels, and for various
10 1.47 1.26 1.17
heights H.
1 0.61 0.77 0.84
ln the simplest power model the mean wind speed V at
0.5 0.41 0.62 0.74
height H is approximated by

17-2
 WINDS
Table 17-4. Mean values for heights rdnging from 10 to 100 m of exp(merrt Y(NjO,l 1

p in Equation ( 17.2) . .
1- MINUTE WI NDSPEEDS (VH) lsoPLETHS OF EXPONENT(P) ,0001
.999~

E\ \ \ ‘Yhg / L:’
Location and Terrain P 2

Fairly Level Open Country

,995
Ann Arbor, Michigan 0.14
.99 : -,0 I
Sale, Victoria, Australia 0.16
Cardington, Bedfordshire, England 0.17
.95
Leafield, Oxfordshire, England 0.17
.90. -
Fairly Level Wooded Country ~
Quickborn, Germany 0.23
Upton, Long Island, New York 0.26
o
Akron, Ohio 0.22 g
3

0
\4 0,18 .6

k’ 4
:
.7
speeds, taken in the summer of 1968, at 8 heights on a k
~ .20
lmt*rs/*oc .8
32 m tower in southeast Kansas. The terrain was very flat a -1 3
0.20

z ,,
.9
and partly covered with wheat stubble 16 to 22 cm tall. The
0.22
1-s wind was never more than 15.3 m/s, Figure 17-1 (right .05 .s5

side) presents resulting isopleths of p as a function of wind- .02 -2

2\ + 0.7
.90

speed, from 1-s winds including gusts to 1-h integrated wind

speeds. The exponents p were found for wind speeds of the
same probability, from level to level, as opposed to the
mean wind speed. Nevertheless previous results were sup-
ported. There is a systematic decrease of p, from 0.7 down
to 0.12, with increasing wind speed for either 1-s, l-rein, /
HEIGHT-METERS (H) OURATION - SECONOS (m)
5-rein or 1-h winds, when the lower 90% of the wind speeds
are considered. For winds equal to or greater than the 90th
Figure 17-1. A nomogram to obtain windspeeds of duration 1 to 1000 s
percentile, the exponent p is almost uniform at 0.12 except
at a height of 1 to 32 m. (The right hand side of the Nom-
for gusts or short-duration winds of 1 min or less. The Windy ogram shows the exponent (p) in the wind power m~el
Acres winds became turbulent above the 95% speeds. In (Equation 17.2) as a function of the duration of the wind,
including 1-s, I-rein, 5-rein, and l-h averages, and as a
gusts, the value of p varies from 0.11 down to 0.08, sug-
function of the strength of the wind (in percentiles). Left-
gesting a tendency toward a common speed throughout the hand side shows percentiles of 1-rein windspeed as a func-
turbulent layer. tion of height above ground level. )
Equation (17 .2) can be used to standardize the height
of the wind data of individual stations to one level even
though they have differing anemometer levels. In the pub- surement at the reference level. If the anemometer mast is
lication “Climatic Extremes for Military Equipment” IMIL- in a poorly defined terrain, the use of the wind profile for-
STD-210B] the value of the exponent p was adopted at 0.125 mula is questionable. Uniformity of the terrain would im-
when the wind is strong but steady, and at 0.08 when the prove the result. Winds observed with a land-based ane-
wind is strong and gusty. mometer cannot be used to estimate the wind speed over
The validity of a wind extrapolation to another height an adjacent water surface. In certain cases the wind speed
is dependent on the representativeness of the wind mea- over terrain may attain a maximum speed at a level signif-
icantly below 100 m. Such cases usually occur in cold air
flow, for example, nocturnal down-slope winds or sea breezes.
Table 17-5. Mean values of exponent (p) for the lowest 9 m in Equation
(17 .2), determined from wind profile measurements.

P
17.1.2 Wind Direction Shifts (Below 3000 m)
H Lapse Rate Inversion Isothermal
(m) (dt/dz < O) (dt/dz > O) (dt/dz = O) Normally wind direction changes with height. Changes
in direction with height are termed veering if the wind turns
9.0 0.11 0.38 0.14
clockwise and backing if the wind turns counter clockwise.
4.6 0.13 0.31 0.16
Veering is usually expressed as the rate of turning in degree
0.9 0.18 0.23 0.21
per height intervals (negative for backing).

17-3
CHAPTER 17
Wind direction shifts with height are caused by surface ~dble 17-7. Average angle formed by the wind and the gradient wind,
and avemce veering for weather stations in Gerrmany. grou~d
frictional effects and by height changes of the horizontal
according- tc~topo~rophy. [H. Lcttau. 1939. ]
pressure patterns controlling the mean airflow. The follow-
ing discussion excludes such phenomena as slope winds,
land and sea breezes, nocturnal low-level airflow conditions Altitude Coastal Rolling Hilly
in the equatorial zone, and rapid, small-scale wind fluctua- (100m) Plains Country Land
tions.
Angle (deg)
In the free atmosphere, the mean horizontal airflow is
9.0 to 15.0 0 2 3
a gradient wind, approximately parallel to the isobars, the
6.0 to 9.0 2 5 10
lower pressure being to the left. Surface friction reduces the
3.0 to 6.0 10 17 25
speed, causing a component of the surface wind to blow
1.5 to 3.0 22 30 36
across the isobars toward lower pressure. Thus the wind
0.0 to 1.5 29 36 43
direction changes with height to align itself with the gradient
wind in the free atmosphere. Veering (deg/ 10Om)
To a first approximation, it can be said that under strong 9.0 to 15.O 0.3 0.7 I .0
mean wind conditions in the Northern Hemisphere, the winds 6.0 to 9.0 0.7 2.0 3.9
will veer with height in the lower 1000 m, the magnitude 3.0 to 6.0 4.3 5.9 5.2
of the veering being determined by the intensity of thermal 1.5 to 3.0 4.9 4.6 4.3
advection processes. W am-t air advection increases veering, 0.0 to 1.5 4.9 3.0 4.3
and cold air advection either decreases the magnitude of
frictional veering or causes backing of the winds. The con-
ditions for veering and backing are reversed in the Southern
of the earth surface: the average for oceans is 5° to 15°, for
Hemisphere. North of approximately 20”N latitude, winds
continents, 25° to 45°. The avel”age veering is usually greater
in the lowest 1000 m of the atmosphere will usually display in winter than in summer, and greater at northern stations
varying degrees of veering with relatively few cases of back-
than at southern stations. The averaging process masks the
ing. Southerly surface winds will veer more with height
variability of veering that would be encountered with iso-
than northerly surface winds. Above approximately 1000 lated observations in time and space. In the first 1000 m of
m, southerly winds will continue to veer with height, while
the atmosphere, however, the importance of the general
northerly winds will begin to show backing with height.
direction of the surface wind in obtaining reasonable esti-
Tables 17-6, 17-7 and 17-8 indicate the order of mag-
mates of veering appears doubtful. Above this layer south-
nitude of veering with associated general meteorological
erly winds veer with height, northerly winds back with
conditions. Because no advection is permitted for the sit-
height. Maximum values of veering in the frictional layer
uation described in Table 17-6, the veering with height is
are near 3°/30 m; isolated cases of backing of the same order
almost constant and represents the gradual decrease of the
of magnitude are observed.
surface frictional effect with height. Once the gradient level
In summary, the average total veering (or backing) in
of the free atmosphere is reached, frictional effects and
the lower 1000 m is 20° to 40°, with isolated cases of 70°
hence veering become negligible in the case of no advection.
to 90°. To a first approximation, it may be assumed that
Strong surface winds generally make an angle with the gra-
this veering (or backing) is evenly distributed throughout
dient wind of 10° to 30°; this is the overall veering found
the layer. Above this layer, primarily dependent on hori-
in the frictional layer of the atmosphere when little or no
zontal advection conditions, winds will show veering or
advection exists. The direction of the surface wind is in-
backing with approximately the same average order of mag-
significant. Total veering over the entire layer of frictional
nitude as in the frictional layer.
influence depends primarily on the roughness characteristics

Table 17-6. Typical frictional veering of wind over plain land with mod- 17.1.3* Diurnal Variation and Low-Level Jet
erately strong gradient winds (18 m/s) and no temperature Streams (below 2000 m)
advectitm: 5 1.3°N. 12.5°E. 20 October 1931. [H. Lettau.
1950}.
The mean diurnal variation of wind speed at various
Altitude Average Speed Angle* Veering heights above any given site is caused by diurnal variations
(100m) (m/s) (deg) (deg/ 10Om) of both the horizontal pressure gradient force and the fric-
tional force. The regular variations of the former are con-
6.0 to 9.0 18 7 2.3
trolled by tidal effects (solar and lunar) which produce pre-
3.0 to 6.0 16 15 2.6
1.5 to 3.0 13 21 2.6
*This is based on the section by H. H. Lettau and D. A.
0.0 to 1.5 9 25 2.3
Haugen in the Handbook qf Geophysics for Air Force De-
*Angle between the wind :md the gmdient wind. signers, 1957.

17-4
 WINDS
Table 17-8, Average veering (deg/100 m) for various ranges of height; means ( 1918 to 1920) for three stations in U.S.A. in the latitude zone 31“ to 36°
(average 33°N), and three in latitude zone 40° to 45” (average 43”N). [Condensed from W. R. Gregg, 1922. ]

Southerly Surface Winds Northerly Surface Winds

Summer Winter Summer Winter

Altitude
(100 m) 33°N 43”N 33°N 43”N 33”N 43”N 33°N 43”N

27.0 to 36.0 2.3 3.3 2.3 –3.3 –3.0 –2.3

15.0 to 27.0 1.0 3.3 2.3 3.3 –1.0 –3.3 –1.0 –2.3
6.0 to 15.2 1.0 2.3 2.3 3.0 –1.6 –1.3 –1.0 –2.3
3.6 to 6.0 3.0 2.3 3.3 3.3 –1.0 –1.0 3.0 3.0
1.5 to 3.6 2.3 2.3 3.0 4.3 2.3 2.3 3.0 4.3
0.0 to 1.5 2.3 2.3 3.6 4.3 1.6 2.3 3.6 4.3

dominantly a 12-hourly wave, and by differential solar heating of stable thermal stratification near sunset. Average wind
of the air over different locations and subsequent horizontal profiles showing the development of a low-level nocturnal
density gradients. In the troposphere the tidal motions of jet stream over O’Neill, Nebraska, are given in Table
the atmosphere are small (range less than 0.5 m/s). The 17-9. A particularly large value of wind speed variation
barometric effects of differential solar heating produce with height, recorded during the development of a low-level
marked diurnal wind variations only in special locations jet stream, was obtained by kite observation at Drexel, Ne-
(along coast lines and the rims of extended high plateaus). braska on the night of 18 March 1918; the surface wind
Over most parts of the continents, the diurnal variation of speed was 26 m/s, but at an elevation of 238 m the reported
wind speed is controlled by the horizontally uniform effects wind speed was 36 rn!s, yielding an average shear of
of the cycle of solar heating and nocturnal cooling of the 0.14 s-’.
earth’s surface. Consequent changes in the vertical thermal
stratification of the atmosphere at 1000 to 2000 m signifi-
cantly influence the effective frictional force in large-scale
air flow.
Over relatively smooth land the daytime thermal strat-
ification intensifies the vertical mixing and the nocturnal
thermal stratification weakens it. This causes a wind speed
maximum near the ground at about mid-afternoon and a
minimum in the early morning hours. As seen in Figure
17-2, the phase of diurnal wind variation is reversed ap-
proximately 100 m above the ground, a level that varies
with climatic zone, season, and surface roughness from
100 to 200 m. The amplitude of the diurnal variation of
wind speed normally has two maxima, at approximately 6
and 600 m. The vertical extent of such diurnal variation
varies roughly as the vertical extent of convective activity
(2000 m).
Over the midwestem United States the nocturnal max-
imum of wind speed above approximately 100 m frequently
leads to a sharp peak of the velocity profile at heights of
300 to 900 m. The peak, usually at the top of the nocturnal
inversion, is significantly stronger (supergeostropic) than I I I I
explained by a balance between the horizontal pressure gra- % 06 12 n 24
dient and the Coriolis forces; it is often associated with EASTERN STANDARD TIME (h)
extremely large values of wind shear. These peak winds are
called low-level jet streams. The supergeostrophic wind speeds Figure 17-2. Diurnal curves (slightly smoothed) of annual mean wind
(peak speeds up to 26 m/s in a pressure field resulting in speed at Oak Ridge. Tennessee, average for 1948 to 1952.
10 to 15 ds of geostrophic speed) suggest that an inertial (The wind speed data at heights of 1.8 and 16 m are from
anemmneter recordings: for other heights, data were ob-
oscillation of the air masses is induced when the constraint tained from pilu-halk)mr sc~undings at 4-h intervals [dashes
imposed by the daytime mixing is released by the initiatio~ arc for clarity of illustration ].)

17-5
 a

CHAPTER 17
Table 17-9. Average wind orofiles (sDeed and direction at various heights) showinz development t]t’ nocturnal luw-level iet stream (Avemrc of five niuhts
of obs;rvation~, Great Piains Turbulence Field Program,-O’Neill, Ne~raska. )’

Mean Local Time

1800 2000 2200 2400

Height Speed Dir. Speed Dir. Speed Dir. Speed Dir.

(km) (m/s) (deg. ) (m/s) (deg. ) (m/s) (deg. ) (m/s) (deg. )

2.0 6.1 223 4.3 225 3.5 223 3.1 23 I

1.8 7.2 219 5.0 226 4.0 227 3.4 233
1.5 9.4 213 8.3 225 5.6 229 5.0 232
1.2 10.9 207 12.3 217 10.8 225 8.9 223
0.9 11.6 194 14.7 202 13.9 211 13.1 219
0.8 11.7 I90 15.5 193 15.9 204 16.2 213
0.6 11.6 182 15.8 190 17.2 196 19.6 203
0.3 10.8 177 14.7 178 18.0 183 18.9 189
0.2 9.7 174 13.6 171 14.9 175 15.1 183
0.1 9.2 173 12.2 167 12.8 I 70 12.9 I70

0200 0400 0600 0800

Speed Dir. Speed Dir. Speed Dir. Speed Dir.

(m/s) (deg.) (m/s) (deg. ) (m/s) (deg.) (m/s) (deg. )

2.0 4.0 227 3.7 214 3.3 208 4.5 208

1.8 4.2 226 3.9 220 3.7 212 4.9 208
1.5 5.4 226 4.5 227 4.8 213 6.2 204
1.2 8.5 227 8.0 228 6.7 218 8.3 203
0.9 13.5 224 12.1 228 10.5 220 11.2 213
0.8 16.5 219 14.5 225 12.7 220 12.4 215
0.6 18.9 211 17.5 217 15.6 217 13.9 215
0.5 19.9 203 18.9 209 17.2 213 15.9 213
0.3 19.0 196 18.7 199 17.3 204 14.0 204
0.2 17.1 189 16.2 191 14.5 195 12.6 I95
0.1 13.4 187 13.9 189 12.7 191 11.8 191

Over the ocean, the diurnal variation of wind speed is here occur quite frequently. In general, they are easily ob-
negligible because diurnal variations in thermal stratifica- served during conditions of clear skies and light to moderate
tions are extremely small. In coastal regions where land intensity of the large-scale aitiow. It should be remem-
and sea breezes are experienced, the amplitude and phase bered, however, that when the diurnal variations are not
of the diurnal variation of wind SW are comparable with readily observed, it may be because they are superimposed
those over land. A reversal of phase with height is ques- on the large-scale airflow. Empirical wind structure models
tionable, however. The vertical extent of the sea breeze is not accounting for possible diurnal variations could then be
roughly 1000 m, the land breeze roughly 500 m, the over- in error.
all land-sea breeze system including return flow, about
3000 m.
In mountainous regions, the daytime, up-slope valley
winds ( 1 to 3 m/s) are normally stronger than the night-time 17.2 LARGE SCALE WIND STRUCTURE
down-slope mountain winds (0.2 to 2 m/s). The mean am-
plitude of the variation decreases very slightly with height. This section provides information on the vertical and
A reversal of phase with height is occasionally found. The horizontal distribution of winds for altitudes up to 90 km.
vertical extent of the valley breeze is about that of the height The wind data for altitudes up to about 26 km are based on
of the surrounding ridges, the mountain breeze, 60 to routine rawinsonde observations taken by the National
100 m. The whole system extends about 300 m above the Weather Service (NWS). The root mean square (rms) ob-
height of the ridges. servational errors in vector wind measurements using FPS-
The various features of local wind variation discussed 16, T-19, or similar tracking radar for altitudes up to 26

17-6
 WINDS
km are 1 rnJs plus 2% the vector wind speed. The infor-
mation presented on winds between 26 and 60 km was m -
---- -
obtained from Meteorological Rocket Network (MRN) ob-
0
servations. Most of the available observations are for lo- . . . . . . . . . . . . . . ... . . . . . . ,.,

cations in the Northern Hemisphere. The estimated rms

observational errors at these altitudes are 2 ds plu: 3% of
---
the vector wind. The wind distribution at altitudes between
60 and 90 km are based on data obtained from grenade and
inflatable falling sphere experiments. The estimated ms g
m
I ZONAL WIND
[50 KM) /
/
I

~
error for the falling sphere observations, which make up the ...... \
s
largest portion of the data, is 2 to 3 m/s between 60 and
90 km. s
*m -

-20 -
— WUITWDS
17.2.1 Seasonal and Day-to-Day Variations --- Wum ISWNO +
-40 -

The broad features of the seasonal change in winds be-

-so
tween the surface and 90 km are reasonably well established. 1 1 1 1 1 1 1 1 I 1 1 1 i
4 MAM J& SON 0
Figure 17-3 is a meridional cross section of the observed &H

mean monthly zonal wind compone~ts for Januq and July

from the surface to 80 km. The seasonal change in the Figure 17-4, Annual variation of meridional and zonal wind at White
Sands, Wallops Island and Ft. Churchill.
stratospheric and mesospheric wind fields differs from tro-
pospheric seasons in timing and length of season. At middle
and high latitudes, long periods of easterly and westerly meridional components (Figure 17-4) are less variable and
flow are separated by shorter periods of transition as shown maintain a slight southerly (positive) component at most
by the curves of the mean monthly 50-km zonal wind com- locations throughout the year.
ponents in Figure 17-4 for White Sands, 32”N, Wallops Profiles of mean monthly zonal wind speeds (m/s) for
Island, 38”N, and Ft. Churchill, 59”N. The mean monthly each of the midseason months at Ascension Island, Wallops

JULY

Figure 17-3. Meridional cross section of zonal wind speed (M/s),

17-7
 /
~..
CHAPTER 17
r , , r ,
I I

1“’’’’’’’’’’’” AsCENSl~ -N WINOS MLLOPS MEN WINOS

601 60 -
\,
\ ‘)
\

‘“/
S2 -
\
\ ‘!
\ /
/
44 - \ /
\ ‘\ /
\ /
~ ‘, f
\
:6s - \ j /’
~ \\ ,/ /
h ,,/
4 ‘\ ,//
\ ,, /
28 -
\ ‘/
\ I JAN—
\ APu ------
JUL ‘—
20 - \\ ~ _.. –
‘\
JAN — ‘b,+
~p~ . -----
JUL —— ‘\ ‘.*
Oel —--— \, k.
12 -
/ / :,)
\ /,/’
.):,/’
4t
, 1 , I I , 1 -~ >
lo20s04050a Tu -m-60 -50* -so-20-Kl o203040m60m
EAST (msin-!) Wsl EAST (nl We+) WEST

,.J
‘,, \
/
\
‘, 1
,/’ /
,,! /
/’
,
,/’

/
/
/
/
I
/
I
I
I JAN—
>1 Am ------
\\ \ JUL — —
“) OeT –--—
12

4
[
1
-60 @
I 1
-40
1
-m
I
-20
I
-lo
1
1020W
1 1 1
4060
, I J
*
EAST (m=-1) WEST

Figure 17-5a. Seasonal effects on the zonal wind profiles at Ascension Island, Wallops Island and Ft. Churchill.

Island and Fort Churchill, from a report by Kantor and Cole 17.2.2 Time and Space Variations
[1980]are plotted in Figure 17-5 for altitudes up to ~ km.
The seasonal variations in the mean monthly zonal winds The change and movement of pressure patterns in the
are largest at Wallops Island, a midlatitude location, and at atmosphere cause standard wind observations (for example,
altitudes above 30 km. The largest day-to-day variations the mean wind over 1 rein) to vary from time to time at a
around the mean monthly values (Figure 17-6) occur during given place, and from place to place at a given time. In
the winter months at altitudes above 20 km at middle and g~neral, as shown in Figure 17-7, the amount of change in
high latitudes. wind between two observations increases with the time in-

17-8
 =

WINDS
1 1 I , 1 1 1 I I 1 I 1 I ! I I 1 I
1 1 1 ! 1 1

AsCENSICN MEAN WINDS I wLLOPS MEAN WINOS

60

52

44
I
I

1-
t 20

L
20
t
JAN —
w—
Am ------
APR ------ ,4 ~L ——

JUL —— 12
~ —-.—
12 ‘ ~~ _..—
1 ‘,
t
t

4~
4

I 1 1 1 1 1 I
1 1 I 1 I

’60 ’50 ’40 ’30 ’20 ‘IO I020W40S060 -60 -s0 -40 -s0 -20 -lo I0203040WW
NRTH (m au-l) SOUTH NU?TH (m m-’) =TH

I , I , , I I I , 1 , !

C+URCHILL MEAN W~S

J
60 -

52 -

44 -

F
=36 -

~
~
d 28 -

20 -

12 -

4 -

*-50-40-SO-20-10 o lomso 405060

NORTH (m u-’) SOUTH

Figure 17-5b. Seasonal effects on the meridional wind profiles at Ascension Island, Wallops Island and Ft. Churchill.

terval between them and with the distance between observ- atmosphere below 30 km are derived from standard pilot
ing points. The rate of increase in wind change with in- balloon or rawinsonde observations. Lower limits of reso-
creasing time or space interval between observations is, in Iution of these observations are such that minimum intervals
turn, a variable depending upon season, geographic loca- for wind variability are about 15 minor 1.6 km. The small-
tion, average wind speed, nature of the sample, and, to scale fluctuations appear to be fairly random, and their com-
some extent, height above the ground. bined effect on ballistic or synoptic-scale forecasting prob-
Most of the wind variability data pertaining to the free lems normally cancel out and, thus, are neglected.

17-9
CHAPTER 17
, 1 r
ASSENSION wALLOPS CHURCHILL
1“’’’’’’’”1

L
i

w-

44 -

$6s -
F
~

28 -

20 -

12 -

M .,
4812
I I I

16
I

20
I /,
46
t

12
1

16 20
,

24 2S
,

~
1 I
4 1312162024
,

~32s640
1

44
J
SMRO OEVIATION OF THE ZONAL WINO (m UC+)

Figure 17-6a. Day-to-Day variability around mean monthly zonal winds for the midseaso~ months at Ascension Island, Wallops Island, and Ft. Churchill.

I r
1 , I I 7 , 1 ,
I
I AWiNSION

STANOARO OEVIATl~ OF TK MERIDIONAL WIND (M =-11

Figure 17-6b. Day-to-Day variability around mean monthly meridional winds for the midseason months at Ascension Island, Wallops Island and
Ft. Churchill.

17-10
 m
WINDS
usually a good approximation to the frequency distribution
of wind changes.

17.2.2.1 Time Variability up to 30 km. The time rate

of change of wind in the frictional layer is affected by the
topography and the thermal structure. The results in Table
17-10 are for steady southerly flow conditions from seven
observational periods, all but one of which extended over
24 h (pilot balloon, rawinsonde, and smoke-puff observa-
tions are combined). The effect of thermal stratification is
indicated by comparison of the day and night values for
time differences up to 8 h.
The change in wind variability above the frictional layer
with increasing time between observations and with altitude
is illustrated in Figure 17-7. Figure 17-8 shows the effect
of latitude and season for a 24-h lag between wind obser-
vations.
For relatively short periods during which the pattern of
the winds is fairly stable, the variability of the wind is given
directly by

S, = Ktp, (17.3)
Fizure
-= 17-7. IsoDleths of time and sDace variabilities of changes in the
me;n vector wind with’ altitude: the rms (63rd &rcentile) where S, is the rms change in wind during the time interval
values, in ki Iometers, of observed changes for various time t, and K is a constant. The exponent p depends on r,, the
lags between observations, and of derived changes for var-
ious distances are given on each curve [Ellsaesser, 1960]. correlation coefficient between winds separated by the time
interval t. At short lags where r( is 1, p is 1; at greater lags
where r, is O, p is 0.5. For t in hours and SI in m/s
Observations of VHF and UHF radar backscatter from
turbulence in the clear atmosphere, conducted at several S, = 1.8 t05 (17.4)
locations during the past several years, permit much more
detailed measurement of the time variation of vector winds is a suitable generalization for middle latitudes and for lag
than was previously attainable. Depending on the radar intervals of 30 min to about 12 h. Although this empirical
wavelength and the power-aperture product [Gage and Bal- relation is an acceptable average, K actually depends on the
sley, 1978], measurements are obtainable from near the mean wind speed. The mean wind varies with season, al-
surface to an altitude of about 100 km. Several radars which titude, and geographic location (see Figure 17-5); hence,
were originally designed for ionospheric research have been values of K other than 1.8 will on some occasions be more
used for observations of the neutral (non-ionized) atmo- applicable to engineering problems. Values of K from 1.5
sphere, and a few radars have been built specifically for to 6.3 are tabulated by Arnold and Bellucci [ 1957]. An
tropospheric and stratospheric wind measurements. The lat- analysis of the relationship of K to the mean wind in a stable
ter include NOAA radars in Colorado and Alaska and the flow pattern shows that K increases from about 0.5 at speeds
SOUSY-VHF-Radar operated by the Max-Planck-Institut of 2 mJs to perhaps 2 or 3 at speeds of 16 rnls and higher.
fur Aeronomie in the Federal Republic of Germany. Ex- An indirect model relates the variability in time to r’ and
amples of data are presented by Green et al. [1979], Balsley to the climatological dispersion of the winds. This relation-
and Gage [1980], Rottger [1979], and in numerous other ship is given by
publications. A radar of this type has been incorporated into
the NOAA Prototype Regional Observation and Forecasting
System (PROFS).
s, = ut~, (17.5)
Several measures of wind variability are possible. The
most useful measure is the (RMS) of the vector change in where U, is the standard vector deviation of the winds (rms
wind; others commonly used are the mean and the median deviation from the vector resultant wind). The vector stretch
absolute vector differences. All of these are scalar measures correlation coefficient between the initial wind having com-
computed from the magnitudes of the difference vectors. ponents u and v and the wind after a time interval t, having
They are related in a circular normal distribution which is components x and y, is given by

17-11
 a
CHAPTER 17
Table 17-10. Mean and standard deviation of absolute value of vector velocity differences at various time intervals, A t, in the lower 1829 m over smuoth
open terrain (Great Plains Turbulence Field Program).

Velocity Differences (m/s) 1

Height At = 2 hours At = 4 hours At’= 8 hours
(1000 m) Day Night Day Night Day Night

1.8 3.0 t 1.7 3.3 ? 2.4 5.0 ? 2.6 4.5 ? 3.2 7.4 ? 2.4 4.0 + 2.9
1.4 2.6 * 1.5 2.8 & 19 4.2 ? 2.5 4.2 + 2.6 5.4 ? 2.8 5.1 ~ 3,3
0.9 2.6 * 1.7 3.9 Y 2,5 4.1 * 2.2 5,6 ~ 26 5.4 * 1.9 7.5 f 3.1
0.5 3.1 * 2.2 4.0 ? 2.8 4.8 + 31 6.5 Y 3.5 6.8 ? 2.8 9.6 & 3.1
0.4 3.0 * 1.9 4.f) ~ 3.f) 4.8 + 3.1 6.9 * 4.5 7.1 * 3.5 6.2 + 3.5
0.2 2.6 * 24 3.4 t 3.1 4.3 + 2.9 6,0 ~ 5.4 6.2 ? 3. I 8.9 ? 4.9

At = 12 hours At = 16 hours At = 20 hours At = 24 hours

1.8 5,6 Y 2,9 6.4 ~ 3.1 7.5 t 3.5 6.5 ? 3.3

1.4 6.3 k 3.1 6.6 ? 3.0 5.9 * 2.1 4.8 ? 2.6
0.9 7.8 t 3.0 7.7 + 3,5 5.9 ? 2.8 4.4 * 2.6
0.5 10.2 * 3.9 8.8 * 4.1 6.9 ? 4.0 4.9 f 3.3
0.4 10.4 i 4.1 8.8 ~ 3.6 7.1 * 4.9 4.6 ? 2.9
0.2 8.3 ? 3.7 7.2 ? 3.8 6.0 t 4.9 4.3 y 3.9

Z(UX + Vy) of the wind variability for a desired lag interval that pertains
(17.6) specifically to the place, season, and altitude of interest.
“ = VZ(U2 + v’) ~(x’ + y’)” This model has two serious limitations; it will not permit r,
to become negative, nor does the constant coefficient of t
This parameter undergoes an exponential decay with in- allow for variations in the rate of decay of the correlation.
creasing lag for intervals up to 24 h or more, and then At sufficiently large time intervals, r, does become negative,
appears to oscillate about zero. The relation, and rt is so close to zero for lags in excess of 72 hours that
the model becomes unreliable. In some cases r, becomes
negative at shorter lags, and the lag at which this occurs
r, = exp ( – 0.0248 t), (17.7)
varies from place to place. Investigations of the rate of decay
of correlation show that it varies geographically, seasonally,
with t in h, is widely used. This equation, in conjunction and probably with altitude. The variation has been mapped
with Equation ( 17.5) and values of at allows an estimate only for the United States at 5.5 km [Ellsaesser, 1960].

I 1
) 60 50 40 ~ ~ 10 “70 60 50 3020100
LATITuDE (deg) LA:?UDE (dog)

Figure 17-8. Distribution in latitude and altitude of the vector wind variability.,.—
(St) for a 24-h time lag between observations alone the meridian 75”W in
summer and winter (August 1954 data are for East Africa); rms values in knots are shojn on curves [Ellsaesser, lj60]

17-12

ACKET 702-058 CHAPTER 17

 WINDS

Attempts to develop a more precise model of the vari- The indirect model provides several empirical curves for
ability of winds with time have resulted either in only mod- the decay of the correlation coefficient of winds with in-
erate improvement or extremely complex models. Thus creasing separation between observing points. These curves
Equation (17 .7) is considered the most useful approxima- indicate differing rates of decay depending upon latitude
tion. When precision in estimating the wind variability with and upon the orientation of the line connecting the observing
time is required, a special climatological study must be points. The analogy between time and space variability of
made. the winds extends to these curves. Figure 17-7 indicates
that, for temperate latitudes, a general approximation to the
17.2.2.2 Spatial Variability up to 30 Km. In general, space variability of winds can be obtained by taking 3 hours
variability increases with increasing distance between ob- as equivalent to 92.6 km. The space variabilityy is then
servation points, and the rate of increase with distance de- estimated in a manner similar to time variability from
pends on geographic location, season, and altitude. A change ad ~; where r, is the correlation coefficient be-
in the wind with time can be thought of as resulting in part tween winds separated by the distance interval d.
from the movement of wind-field patterns over the observing
point, and thus as analogous to the spatial variability.
Extending this analogy to the models, the space vari- 17.2.2.3 Time and Space Variations-30 to 60 km.
ability of wind Sd is given by K’dp, where d is the distance Observations show that pronounced diurnal and semi-diur-
between observing points. The parameter K’ varies with nal oscillations exist in the winds at altitudes above 30 km.
season, geographic location, and altitude, but no detailed Sufficient data, however, are not available to permit the
examination has been made of the way in which these factors development of a satisfactory model for all latitudes and
act. Arnold and Bellucci [1957] tabulated values of K’ from seasons. The results of an analysis of a series of wind ob-
1.1 to 6.1. They consider the expression, representative for servations taken at Wallops Island during a 48-h period in
middle latitudes. May 1977, shown in Figure 17-9, provide an indication of
the magnitude of the diurnal oscillation at a midlatitude
Sd = 1.5d05, (17.8a) location. The X’s in Figure 17-9 are average values of the
north-south and east-west wind components at the indicated
where Sd is in mph and d in miles, or converted to metric local times. The curves are based on a harmonic analysis
units of the meridional (N-S) and zonal (E-W) wind components.
The amplitude of the combined diurnal and semidiumal
Sd = 0.53d05 (17.8b) oscillations of the N-S component, increases with altitude
above 30 km, reaching a maximum of 10 to 1I m/s between
where Sd is in m/s and d in km. 50 and 55 km. The diurnal variations of the E-W component

12 - 60km
4 -

-12 - x

s~
~ = I 732-5.385 cos15° t-7 884s,. 15”t

12 - 55 km x 35972+0~2 CDS 15-t ‘6966 sin 15-?

4 -

~= -,,40~. *,752 COS15 et- 7.115 s,. 15”t

z -40
~ 12 -
E4
z -4 -
- x
z
Q
G
-44

~ -24 –
Y= ‘28757+5 249 CO, 15°t -3o44 s,n 15-t
1
0 WKM
p -12 - x a
> -20 -
~ Y = 1.819-0.805 COS 15”1-10885 s,n 15°!
IJl -2 – ~, -z66w + 5Z9S cm 15°t+ 1721 S1. 15°t
>
m

-16 – 45KM

-m –
y = 1.908+ 2.4zE cos 150t- 3.186 s,n15Qt -24 -
~:~ V= ‘20257-1999 cm 15°t ,2734,,” 15°t

.-
IP L 40km -! -16
40KM
Y
4 - v x F -20 –
-4 - y = 2.560-0.693 cos 15 °t- 3,497 sln 15”t -24
y, -18514+ 2 % CO, 15at -23~sIn 15”t
, 1 t I 1 I 1 1
1 1 1 1 I i
19 22 01 04 07 10 13 16 19 18 21 w 09 12 15 18
LO:AL TI MWE ( h )
LOCAL TIME ( h)

Figure 17-9a. Diurnal wind variations at Wallops Island in May (N-S). Figure 17-9b. Diurnal wind variations at Wallops Island in May (E-W).

17-13
CHAPTER 17
, ,
16 60 km 17.3 WIND PROFILES
x x
12 L x x x x x
. xx x c
8 -x
4 -
Y = 10 857+0 236COS 15”t -0 363s,” 15-+
17.3.1 Wind Shear
55 km x x
12 - x

:8 x
m
x
. Wind shear is the derivative of the wind vector with
:4
b Y=9 5oo–0956cos 15”t+0 371s, n 15”t i
5 o
respect to distance and is itself a vector. The shear of the
50 km x
z 12 -
0 horizontal wind is of primary interest and is the one dis-
x 25 PAIRS x
cussed in this section. The terms vertical wind shear and
Y=7 886–0 459COS 15”1+1 029s,n 15°t x ~< ~ p~,R~ o
2 horizontal wind shear are commonly used in referring to the
shear of the horizontal wind in the vertical and horizontal
=8 -
directions, respectively. Horizontal wind shear is the deriv-
“,7653-I ,59CIJS ,5°+-(10,2s,n ,5”+ J
40 km
ative of the horizontal wind with respect to an axis parallel
x
8 -
x x x--
x to the earth’s surface. Its applications are restricted largely
4 - x Y=5.992-0.403cos 15”1+0.322s,n 15$t
to meteorological analysis. Vertical wind shear is expressed
6 12 18 24 30 36 42 48 54 60 66 72
TIME LAG (h )
as AWIAy, where A is the change of the horizontal wind
in the altitude interval Ay; the unit of shear is s-1. Although
direction is also necessary to specify the shear vector, it is
Figure 17-10, Rms differences between north-south winds observed 1 to
72 h apart in Ascension, Ft. Sherman and Kwajalein. usually ignored; for design purposes shear is normally ap-
plied in the most adverse possible direction.
The climatology of vertical wind shear is applicable to
are not as well defined at 40 and 45 km. At altitudes of 50 problems dealing with design and launch of vertically rising
km and above, however, the amplitudes are between 12 and vehicles and jet aircraft, radioactive fallout investigations,
14 ds. The magnitude of the diurnal oscillations are some- and many phases of high-altitude research. Most investi-
what smaller in tropical and arctic regions. gations of shear climatology are for specific locations in
Rms differences between wind components observed order to satisfy design and operational requirements of mis-
from 1 to 72 h apart have been analyzed at altitudes between siles and other vehicles at or near launch sites. For vehicles
30 and 60 km to obtain estimates of the variations in these with other than a vertical flight path, vertical wind shear
components with time. Rms differences between the N-S can be determined by multiplying the shear by the cosine
wind components observed from 1 to 72 h apart at Ascen- of the angle between the vertical axis and the vehicle tra-
sion, Ft. Sherman and Kwajalein are shown in Figure jectory.
17-10. Data for all seasons for these three stations have been Measurements of vertical wind shear indicate that av-
combined. The values display a relatively stable rms vari- erage shear behaves inversely with layer thickness (scale-
ation of roughly 10 to 12 rnls for periods up to 72 h at 55 of-distance). This is illustrated in Figure 17-11, which is
and @ km and about 6 or 7 ds at altitudes of 40 and 45 based on a relatively large number of observations during
km. There is an indication of a diurnal cycle with an am-
plitude of 1 to 2 m/s at these altitudes as shown by the
24-h harmonic curves which have been fitted to the indi-
vidual values. Most of the observed variability can be at-
tributed to random measurement errors as there is little dif-
ference in the rms variations between observations taken 6
h apart and those taken 72 h apart.
In summer conditions are approximately the same at all
latitudes, with little change in rms values with time. In
winter, however, the rate at which the rms values increase
with time varies with latitude. The larger day-to-day changes
in the synoptic patterns at middle and high latitudes are
reflected in well-defined increases in therms variability with
time. At 50 km, for example, rms values for the N-S com-
ponents at White Sands Missile Range (32”N) increase with
time from 6 to 8 m/s to 16 or 18 m/s in 72 h. At Poker SCALE OF DISTANCE, AY,(m)
Flats (64”N) and Ft. Churchill (59”N), they increase from
6 or 8 ds to 25 mls. The diurnal cycle in observed data is Figure 17-11, Selected vertical wind shear spectrums (4. 10-14.20-35.
and 60–80 km altitude) for use with 57c and 17Cprobability
masked by the synoptic changes and instrumentation errors level wind profile envelope, Cape Canaveral, Florida
during the winter months at middle and high latitudes. [Scoggins and Vaughan, 1962],

17-14
 WINDS
0.3

Au = Aulm
()
1000
—
Ay ‘
(17.9)

where Au is the shear, Au ,m is the 1000 m shear, and Ay

is the scale-of-distance in m for thickness < 1000 m.
Wind shear statistics for various locations differ pri-
marily because of prevailing meteorological conditions, or-
ographic features, and data sample size. As a result, sig-
nificant differences exist in the shear structure for different
11111 II GIVEN ALTITUDE(Y) locations. Consistent shear data for five vehicle launch and/or
landing sites are presented in Tables 17-11 through 17-15
for the Eastern Test Range, Florida; Vandenberg AFB, Cal-
ifornia; Wallops Island, Virgihia; White Sands Missile Range,
New Mexico; and Edwards AFB, California. To get actual
shear (s-l ) from the indicated wind speed changes, divide
by the appropriate scale-of-distance. Table 17-16 gives en-
VERTICAL WIND SHEAR (80C-1) velopes of the 99% wind speed change for the five locations
combined. The data contained in Table 17-16 are applicable
Figure 17-12. 1% probability-of-occurrence vertical wind shear spectrum when design or operational capability is not restricted to a
as function of altitude and scale-of-distance for association specified launch site or may involve several geographical
with the 5% and 1% wind speed profi Ie envelope for cape
Canaveral, Florida IScoggins and Vaughan, 1962].
locations. Equation (17.9) was used to construct Tables
17-11 through 17-16 for scales-of-distance < 1000 m.

the windiest months over Cape Canaveral, Florida. Figure

17-12, based on the same data, shows the variation of shear 17.3.2 Interlevel Correlations
with altitude and with layer thickness; it provides, as an
example, the vertical wind shear spectrum at Cape Can- Deviations in the assumed vertical wind profile over a
averal with a 1Yo probabilityy of occurrence. target or reentry point affect the range and cross-range of a
Wind shears for scales-of-distance Ay z 1000 m in ballistic missile. These effects must be considered in the
thickness are computed directly from rawinsonde and rock- design of guidance systems for entry vehicles and for tar-
etsonde observations, whereas smaller scale shears can be geting ballistic missiles.
calculated directly only from special fine-scale observations. The mean effect E of mean monthly or mean seasonal
Howe”ver, shears associated with scales-of-distance winds on the range and cross range of a missile can be
Ay < 1000 m can be estimated from the following rela- determined for a particular location by computer-simulated
tionship [Fichtl, 1972]: flights through mean monthly component wind profiles if

Table 17-11. Envelopes of 99% wind speed change (m/s), 1–80 km altitude region, Eastern Test Range [Kaufman, 1977].

Scales of Distance (m) Thickness

Wind Speed at Reference

Altitude (m/s) 5000 4000 3000 2oofl 1000 800 600 400 200 100

2 90 77.5 74.4 68.0 59.3 42.6 36.4 29.7 22.4 13.8 8.5
= 80 71.0 68.0 63.8 56.0 40.5 34.7 28.5 21.4 13.2 8. I
= 70 63.5 61.0 57.9 52.0 38.8 33.1 27.0 20.3 i2.5 7.7
= 60 56.0 54.7 52.3 47.4 36.0 31.0 25.3 18.9 11.7 7.2
= 50 47.5 47.0 46.2 43.8 33.0 28.3 23.2 17.5 10.7 6.6
= 40 39.0 38.0 37.0 35.3 29.5 25.3 20.6 15.5 9.6 5.9
= 30 30.0 30.0 29.4 26.9 22.6 19.4 15.8 11.9 7.3 4.5
= 20 18.0 17.5 16.7 15.8 14.6 12.5 10.2 7.5 4.7 2.9

17-15
CHAPTER 17
Table 17-12. Envelopes of 99% wind speed change (m/s), 1-80 km altitude region, Vandenberg AFB [Kaufman, 1977].

Scales of Distance (m) Thickness

Wind Speed at Reference

Altitude (m/s) 5000 4000 3000 2000 1000 800 600 400 ‘ 200 I00

2 90 66.9 62.5 57.8 51.5 37.5 32.1 26.1 19.7 12.0 7.4
= 80 64.1 60.8 56.6 48.8 36.9 31.5 25.6 19.1 11.6 6.8
= 70 62.0 59.2 54.8 48.1 36.0 31.0 25.0 18.6 11.2 6.5
= 60 57.1 54.5 51.3 45.4 32.7 28.5 23.0 17.1 10.2 5.3
= 50 49.6 47.8 45.7 42.1 30.1 25.9 21.8 15.6 9.2 5.0
= 40 39.4 38.8 37.9 35.5 25.9 23.5 19.6 14.9 8.8 4.8
= 30 30.0 29.4 28.3 26.3 20.5 18.6 15.8 12.2 8.0 4.6
= 20 20.0 19.8 19.5 18.4 15.0 13.1 10.9 9.0 6.3 4.3

Table 17-13, Envelopes of 99% wind speed change (m/s), 1-80 km altitude region, Wallops Island [Kaufman, 1977],

Scales of Distance (m) Thickness

Wind Speed at Reference

Altitude (m/s) 5000 4000 3000 2000 1000 800 600 400 200 I00

2 90 72.5 67.0 60.2 50.5 37.6 32.3 26.3 19.8 12.2 7.5
= 80 66.5 62.5 57.5 48.8 37.0 31.7 25.9 19.5 12.0 7.4
= 70 61.2 58.5 53.8 46.5 35.8 30.7 25.1 18.9 11.6 7.1
= 60 54.4 52.5 50.0 44.4 34.5 29.6 24.2 18.2 11.2 6.9
= 50 46.2 44.2 42.3 38.8 33.0 28.3 23.2 17.4 10.7 6.6
= 40 36.7 35.6 34.5 32.3 27.6 23.7 19.3 14.5 8.9 5.5
= 30 27.2 26.3 25.3 24.2 20.6 17.7 14.7 10.8 6.7 4.1
= 20 17.8 17.3 16.8 16.4 15.2 13.0 10.6 8.0 4.9 3.0

Table 17-14, Envelopes of 99% wind speed change (M/s), 1-80 km altitude region, White Sands Missile Range [Kaufman, 1977].

Scales of Distance (m) Thickness

Wind Speed at Reference

Altitude (m/s) 5000 4000 3000 2000 1000 800 600 400 200 100

a 90 70.7 67.0 61.2 52.4 42.0 36.0 29.4 22.1 13.6 8.4
= 80 66.0 63.0 57.7 50.0 40.2 34.5 28.1 21.2 13.0 8.0
= 70 60.2 57.0 53.0 46.5 38.0 32.6 26.6 20.0 12.3 7.6
= 60 52.6 50.0 46.5 42.3 35.5 30.5 24.9 18.7 11.5 7.1
= 50 45.0 43.0 40.2 37.0 32.0 28.3 23.1 17.4 10.7 6.6
= 40 36.5 35.5 34.8 33.5 29.3 25.1 20.5 14.5 9.5 5.5
= 30 27.4 27.0 26.4 24.8 22.0 19.3 15.8 11.8 7.3 4.5
= 20 18.4 17.7 17.3 16.5 15.0 12.9 10.5 7.9 4.9 3.0

17-16
 a
WINDS

Table 17-15. Envelopes of 99% wind speed change (m/s), 1–80 km altitude region. Edwards AFB [Kaufman, 1977].

Scales of Distance (m) Thickness

Wind Speed at Reference

Altitude (m/s) 5000 4000 3000 2000 1000 800 600 400 200 100

2 90 75.2 72.0 67.3 59.0 42.8 36.7 30.2 22.5 13.9 8.5
= 80 68.0 66.3 62.5 55.5 40.8 35.0 28.6 21.5 13.2 8.1
= 70 60.4 59.0 56.8 51.4 38.7 33.2 27.0 20.4 12.5 7.7
= 60 53.0 51.8 49.3 45.0 36.0 30.9 25.2 19.0 11.7 7.2
= 50 44.8 43.6 41.5 38.4 32.0 27.5 22.4 16.9 10.4 6.4
= 40 36.5 35.5 34.5 33.0 27.0 23.2 18.9 14.2 8.8 5.4
= 30 28.0 27,3 26.9 26.3 21.4 18.4 15.0 11.3 6.9 4.3
= 20 18.0 17.7 17.4 16.7 15.2 13.0 10.6 8.0 4.9 3.0

the appropriate influence coefficients for the missile at var- The two components can be combined and used to determine
ious levels are given: the probabil~y of occurrence of deviations of any desired
magnitude from a trajectory or impact point based on mean
E=~Ci Vi (17.10) monthly winds. In these computations, it is assumed that
the cross-component correlations are zero at and between
where Ci is the influence coefficient at the ith level that levels, and the wind frequency distributions are essentially
describes the portion of the total response of a missile as- circular normal.
signable to that level, and Vi represents the mean of the The wind climatology necessary for determination of
component wind speed at that level. The variation around ballistic effects consists of statistical arrays of mean monthly
this average effect caused by day-to-day fluctuations in the north-south and east-west wind components, their standard
winds is obtained from deviations, and interlevel correlations between the same
component at different levels. This information has been
lJ~N~ = ~i,J C C, ri, vi u, (17.11) prepared in matrix form for a relatively large number of
locations. Examples for levels between the surface and
where U~~Tis the integrated variance for all levels consid- 60 km are shown in Table 17-17 for the month of January
ered, Ci and C, are the influence coefficients at the ith and at Ft. Churchill, Wallops Island and Ascension Island. Winds
jth levels, Ui and Uj are the standard deviations of the com- are lighter in summer at middle and high latitudes. For
ponent wind at these levels, and rij is the correlation between targeting purposes, the wind data must pertain to the lo-
the component wind at the ith level and that at the jth level. cations of interest. For design purposes, however, a rep-
The square root of the solution to this equation yields the resentative sample of data from the various climatic regions
standard deviation for each component of the ballistic wind. should be used.

Table 17-16. Envelopes of 99Yc wind speed change (m/s), 1–80 km altitude region, for All five locations [Kaufman. 1977].

Scales of Distance (m) Thickness

Wind Speed at Reference

Altitude (tis) 5000 4000 3000 2000 1000 800 600 400 200 100

2 90 77.5 74.4 68.0 59.3 42.8 36.7 30.2 22.5 3.9 8.5
= 80 71.0 68.0 63.8 56.0 40.8 35.0 28.6 21.5 3.2 8.1
= 70 63.5 61.0 57.9 52.0 38.8 33.2 27.0 20.4 2.5 7.7
= 60 57.1 54.7 52.3 47.4 36.0 31.0 25.3 19.0 1.7 7.2
= 50 49.6 47.8 46.2 43.8 33.0 28.3 23.2 17.5 0.7 6.6
= 40 39.4 38.8 37.9 35.5 29.5 25.3 20.6 15.5 9.6 5.9
= 30 30.0 30.0 29.4 26.9 22.6 19.4 15.8 12.2 7.3 4.6
= 20 20.0 19.8 19.5 18.4 15.2 13.1 10.9 9.0 6.3 4.3

17-17
CHAPTER 17
 a
WINDS

17-19
CHAPTER 17

;
>

I
 WINDS

17.4 DESIGN DATA ON WINDS where (Y - 3) v, – v,

(17.13)
T= s,
Wind statistics are presented in a variety of ways, each
of which is intended for maximum usefulness for particular The cumulative probability of wind speed, P(V,), equal to
aspects of design and operational problems. In “Upper Wind or less than V. is given by
Statistics,” Crutcher [1959] presents northern hemisphere
charts of some 15 wind variables or statistics. For surface
winds, in the “Climatic Atlas of the United States,” the
National Weather Service [1968] presents monthly and an-
nual charts of the prevailing direction, mean wind speed,
‘(vS)=
1 -e-y
(l’’+y+~)
‘1714)
As an example, in the month of January, the noontime
the fastest wind on record and its direction, and wind roses Bedford, Mass., wind has a mean hourly speed ~, of
that give the frequency distribution of the wind by direction 5.1 m/s and standard deviation s, of 3.5 ds. Equation
including frequency of calm, the mean wind vector and
(17. 14) gives the cumulative probability for several values
direction. There is also a table of the frequencies of 9 cat- of the wind speed as follows:
egories (Beaufort) of wind speeds for some 120 U.S. sta-
tions.
Generally speaking, any data source on surface wind v, Estimated Observed
speed is limited to providing several key parameters from ds Y P(v.) Distribution
which the general frequency distribution of the wind, in
calm 0.445 0.011 0.115
speed and direction, can be estimated or reconstructed. But
1.5 1.219 0.125 0.168
to do so, a practical model of the distribution of the wind
3.1 1.993 0.321 0.322
speed must be adopted. Three such models are the gamma
5.2 3.03 0.583 0.592
distribution, Weibull distribution, and the circular normal
8.2 4.57 0.834 0.838
bivariate distribution. Each has its merits and drawbacks.
10.8 5.89 0.932 0.946
13.9 7.41 0.978 0.989
17.O 8.96 0.9936 0.999
17.4.1 Hourly Surface Wind Speeds

The distributions of surface wind speeds, observed every This kind of approximation is good for moderate to strong
hour on the hour, have been studied by many climatologists wind speeds and is generally useful except for very light
and statisticians whose conclusions and consequent models winds or calm conditions.
differ. Some favor the distribution of wind speeds as given Model Alternative 2: A Weibull Distribution. A recent
by the circular bivariate normal distribution, others the log-
study [Bean and Somerville, 1979] has produced the model
normal, still others the gamma distribution. A historical
record of at least 5 years should be used to obtain good P(V,) = c + (1 – c) [1 – exp(–av?)] (17.15)
estimates of wind speed distributions, especially small prob-
abilities (for example 1%). Although there are about 600 where c is the probabilityy of calm, and a and @ are param-
city and airport stations in the United States where hourly eters of the Weibull distribution that are determined either
records are kept, they generally represent wind fields in- to make the model fit the observed frequencies of wind speed
adequately. A location that is close to a weather reporting (in a least squares sense), or are estimated in other effective
station should have similar wind characteristics, but terrain ways. The records of many stations indicate a high prob-
effects including manmade effects, and proximity to large ability of “calm, ” which makes this formulation desirable.
bodies of water superimpose a spatial variability that is Records like the “Revised Uniform Summaries of Sur-
difficult to generalize. Fairbanks, Alaska, for example, has face Weather Observations” (RUSSWO) (Compiled by the
a mean wind speed in January of 0.9 m/s with standard Environmental Technical Applications Center, Scott AFB,
deviation 1.7 m/s. Yet Big Delta, only 121 km away, has 111.)contain the relative frequencies f, of each of 11 or fewer
a mean wind speed of 5.4 m/s and standard deviation winds speeds V,. Where x,, the middle value in the category,
4.9 mls. is used for the ith category of V,, formulas have been found
The following are alternative models of windspeed fre-
for u and ~ as follows:
quency distribution:
Model Alternative 1: A gamma distribution. In terms of ~ f,
the mean wind speed ~, and standard deviation s,, the prob-
ability density function of the wind speed V, is given in ~=~fixp
(17.16)
terms of a transformed variable y by [~ f, - X? In Xi ~filnXil-l

f(y) = 0.5 y2 e-y, (17.12)

17-21
CHAPTER 17
To solve for a and ~, an initial guess is entered for ~ in however, can be given in terms of ~ and s or s.. One of
the right-hand side of Equation ( 17. 16), and the equation the most applicable formulas is
is solved for a first estimate of ~. With this revised estimate,
I(I “

the equation is solved again for a second and better estimate,

P(V.) = e-yn~o~ 1 – e-x ~:o;], (17.20)
and so on, until the value of ~ is stabilized. .[

The RUSSWO for Bedford, Mass., January at noon-

time, .,.
gives f, for categories of V, (or xi) which, when entered where x = V;ls:,
in Equation (17. 16) produced the following values after
several iterations: y = /v[’/s:.

~ = 1.961 In the example of Bedford, Mass. January noontime

winds, the RUSSWO table of wind speed versus wind di-
& = 0.02512 rection yields

When these values for a and ~ are entered in Equation

(17. 15) together with c = O.115, the probabilities P(V,) ITI = 2.9 dS
become s“ =5.lmls

or S = 3.6 dS
as opposed to V, = 5.1 mls
v, Observed
s. = 3.7 m/s.
(m/s) P(v,) Distribution

calm 0.115 0.115

When the values for 1~1ands, are used in Equation (17.20)
1.5 0.166 0.168
the probabilities P(V,) become
3.1 0.297 0.322
5.2 0.527 0.592
8.2 0.816 0.838 v,
10.8 0.9395 0.946 m/s P(v,) Observed
13.9 0.989 0.989
17.0 0.9987 0.999 1.5 0.063 0.168
3.1 0.231 0.322
5.2 0.521 0.592
Model Alternative 3: The circular normal distribution. 8.2 0.853 0.838
The assumption of the circular normal distribution for the 10.8 0.964 0.946
wind implies that the zonal u and meridional v components 13.9 0.9936 0.989
each have a normal Gaussian distribution with individual 17.0 0.9965 0.999
means ti and V but with the same standard deviation s. The
two components are independently distributed. As a con-
Thejistribution of the wind V around the mean wind
sequence the probability density function of each is
vector (V) is sometimes of prime interest as it would be in
f(u) = {exp [ - (u -U)2/2]} targeting. It is the difference between the actual vector wind
and the mean vector wind and has the distribution
6
(17.17) v’
f(v) = {exp [ - (v -7)’/2]}
6“
P(V) = 1 – exp
() –~
v
(17.21)

Thus a circle of radius equal to one standard vector devia-

The magnitude of the mean wind vector (~) is given by tion, drawn about the tip of the mean wind vector, would
encompass 63% of the population of winds. Half of the
Ivl = ~. (17.18) wind vectors will fall within a circle of radius 0.83 times
the standard vector deviation.
The standard vector deviation (s,) is simply

s. = Sti (17.19) 17.4.2 Surface Wind Direction

The mean wind speed ~, and the standard deviation of Winds tend to have a modal direction, like the trade
the wind speed s, are not easily derived from the compo- winds or the prevailing westerlies. Using the model of the
nents. The cumulative probability p(v.) of the wind speed> bivariate normal distribution, the estimate of the frequency

17-22
 =

WINDS
P(6, < A <62 of the wind direction A between directions 17.4.3 Surface Wind Gusts
01 and 02 is given by
Studies on the relationship of gusts to the steady wind
speed are in general agreement; however, quantitative re-
P((3, ~ A < 02) = (l/m)
sults have varied depending on the data ‘and the analytical
methods used. Sissenwine et al. [ 1973] reviewed the results
. “y ,exp _ (n - U)’ 8u][F(x,))
2
– F(x,)] (17.22) of many of these; he also presented a comprehensive analysis
“=0
of gustiness based on actual wind records during periods of
strong winds, observed at 14 airfields in the Northern Hemi-
where u = Vls,
sphere between 14° and 77° latitude. Tattehnan [1975] ana-
V is the wind com~nent in the direction (60) of the
lyzed the same data to develop 50%, 75%, 90%, and 98%
mean wind vector ~,, afid s is the standard deviation of the
gust factor curves for 5-rein, 1-rein, and 30-s average (or
appropriate wind component,
steady) wind speeds. Gust speeds reported in weather ob-
servations are normally considered to be about 2-s averages,
ii = IvI/s, but for designers of various sized equipment, other short
bu is a small increment of u, duration gusts may be applicable.
Xz = tan (02 – O.), Since resolution of the w~nd records was approximately
xl = tan (61 – do), 2 s the shortest period gusts were considered to be 2-s wind
speeds. Gust factor (GF) curves were fitted using the equa-
and F(x) is approximated by tion

F(x) = 1 - 0.5(1 + C,x + C*X2 + C3X3 + C.X’)-’ GF = 1 + Ae-Bv, (17.23)

where
where A and B are constants, and V is the speed of the
Cl = 0.196854 steady wind. The values for B were determined by a least
squares fit of all the data to the 5-rein, I-rein, and 30-s
c~ = 0.115194
speeds. Using these values of B, the value for A was de-
C3 = 0.000344 termined for each percentile curve by a least squares fit to
the mid-class wind speed values (for exatiple, 12.6 m/s for
C4 = 0.019527 V = 10. 3–14.9 m/s) and weighted for the number of ob-
servations at each point. Figure 17-13 shows curves of the
In the example for Bedford, Mass., January noontime wind 30-s gust factor as related to the 5-rein steady speed, and
00 = 300°. After correcting for calm frequenc:~ of 0.115, the l-rein gust factor related to the 5-rein speed using all
the computed frequencies of wind directions were as fol- the data. Figure 17-14 shows the relationship between the
lows: 2-s gust factor and the 5-rein steady speed. Curves in both

Direction P(o, = k < O*) Observed

N 0.082 0.067
NNE 0.056 0.041
NE 0.027 0.040
ENE 0.027 0.013
E 0.022 0.019
ESE o.oi9 0.010
AIRFIELD DATA
SE 0.018 0.0.14
SSE 0.020 0.009 bl I
0.025 0.022

‘::~
s
Ssw 0.032 0.051
Sw 0.036 0.067
Wsw 0.070 0.063
10 20 30 40 50
w 0.098 0.108 5-Ml N WIND SPEED (m/see)
WNW 0.122 0.143
NW 0.126 0.110 Figure 17-13. Curves of the 30-s and I-rein gust factor to 5-rein wind
NNW 0.110 0.106 speed for the airfield data fitted to GF = 1 + AeB’ [Tat-
telman, 1975].

17-23
 m
CHAPTER 17
I I I I I figures extend beyond the limit of the wind data, for which

k
(173) (133) (119) (70) (42) NUMBER OF
the highest steady speed is 35 MIS.
2.1 OBSERVATIONS
AIRFIELD DATA Sherlock [1947] found that a gust must have a duration
such that its size is about 8 times the downwind dimension
of a structure in order to build a force on the structure
commensurate with the gust speed. Because the orientation
of most equipment will not be taken into consideration, the
shortest horizontal dimension of the equipment should be
considered as the downwind dimension. For example, if a
structure has a 5-m downwind dimension, a gust must be
40 m long to build up full dynamic pressure. Therefore, a
speed of 40 m/s would build up full force on such a structure
in only 1 s; a speed of 20 m/s would require 2 s. Figure
17-15 shows the expected (50%) gust factors versus gust
duration and 5-rein steady speed to aid in calculating the
design windspeed for structures of differing downwind di-
mensions. The curves were drawn to four points: a gust
factor of 1.0 at 300 s, the 60-s and 30-s gust factors for
each speed from Figure 17-13, and the 2-s, 50% gust factors
from Figure 17-14. Lines on Figure 17-15 were extended
,.l~ from the 2-s gust factors to include 1-s gust factors.
Tattelman [1975] also presented curves of the 50%,
5-MINUTE WIND SPEED (Mlsec)
75%, 90Y0, and 98% wind speed range (the difference be-
tween the maximum and minimum 2-s speed) as a function
Figure 17-14. Relationship between the 2-s gust factor and the 5-rein
wind speed at indicated percentiles for the airfield data. of time interval and mean 5-rein wind speed. The 9070 wind
(Data were fitted to GF = 1 + Ae”Bv and weighted for speed range is shown in Figure 17-16. The dashed lines
the number of observations at each mid-class vaiue).
indicate extrapolation beyond the limits of the data. The
gust factor and wind speed range curves in this section are
300
considered applicable to n ost airport locations at the average
20Q
height of the data used, 5 m.

100 I I I I I
5- MINUTE SPEEI
( m/,ec)
60
500E

;
~ 30

z
o
G
s
~

q ,0
w

I
!
I I , , ,
GUST FACTOR 10 20 m 40 50
90- PERCENTILE WIND SPEED RANGE (Mjsec)

Figure i 7-15. Expected (50%) gust factors versus gust duration and Figure 17-16. 90% wind speed range versus time interval and 5-rein
5-rein steady speed. sped.

17-24
 WINDS

17.4.4 Extreme Surface Wind Speeds As an example, a !9-yr record at Denver, Colorado
shows an average annual 5-rein wind speed maximum
Various statistics are used for the daily maximum wind 15 m above ground level of 19.6 mJs and standard deviation
speed. One statistic is the maximum 5-rein wind of the 288 2.5 m/s. If a tower is to be built to last for 25 years, with
5-rein intervals of one day. Another popular measure is the 99% certainty of survival throughout the 25 years, then the
fastest mile, which is the reciprocal of the shortest interval critical wind speed, given by Gumbel model (Equations
(in 24 h) that it takes one mile of air to pass a given point. (17.28), (17.27)), is estimated at 33.7 m/s.
The strongest gusr is the highest reading of an instantaneous An estimate is subject to sampling variations, in the
recording anemometer. There is no unique relation between above example a 19-yr sample. One standard error of this
these units. The Washington, D.C. records show that, for estimate s(x) is approximate y
speeds greater than 15 ds, the maximum 5-rein wind speed
is roughly 0.93 times the fastest mile and roughly 0.67 times s(i) = Sxa
the strongest gust. The maximum I-rein wind speed is roughly
0.71 times the strongest gust. Canadian and British records where N is the number of years, and u is given by
show the maximum 1-h wind speed to be 0,62 times the
strongest gust. a = 0.710 + 0.l16y + 0.669 y2
A model distribution of the annual extreme has been the
subject of conflicting viewpoints. The most favored distri- In the Denver example, the standard error of the estimate
bution is the Fisher-Tippett Type I, now generally known (33.9 m/s) is approximately s(x) = 3.8 m/s, or 11% of the
as the Gumbel Distribution. If x is the symbol for the annual estimate itself. A brief meteorological record of annual ex-
extreme, then the cumulative probability y of x is given by tremes permits estimates only within confidence bands that
vary with the size of the record.
P(x) = exp[ – exp( – y)] (17.24) Publications by climatic agencies such as NOAA En-
vironmental Data Service and the Canadian Meteorological
where y is the reduced variate, whose mean is ~ = 0.5772 Service give extreme values. Yet each permanent installa-
and whose standard deviation is UY = 1.28255. tion might require a special study for a particular site and/or
A relatively simple equation can be used to relate x to for a particular design problem. For example, the extreme
y: annual wind speed at Logan airport (Boston) would be an
underestimate, by about 33%, of the extreme on Blue Hill,
19 km to the south-southwest. Even the record at the exact
(Y – Y) .— (x - x)
(17.25) location of interest may not be representative of the wind
~Y Sx
effects on a tall (or short) structure.
Table 17-18 gives means and standard deviations of
where Y and s, are the mean and standard deviation, re-
annual extreme wind speeds and peak gusts at some rep-
spectively, of the annual extremes (x).
resentative stations around the world including stations that
The work that is implied to make adequate estimates of
are typically in the paths of hurricanes. The 170 risk in
percentiles of x is to find the sample mean and standard
10 yrs was computed by the above formula, assuming the
deviation of a set of N extremes for N years. Other ap-
Gumbel distribution. This column is virtually the same as
proaches, although more rigorous, require ordering the N
for the 1000-yr return period, and thus is tbe speed that has
extremes by magnitude and/or estimating parameter values
only 1/10 of 170 chance of being exceeded each year. There
by successive approximations. Estimates also might depend
is a noticeable difference between the extreme speeds at
directly on sample size,
airports and at city stations, attesting to the effects of tall
To judge the risk of an extreme event in n years, then
buildings in the cities. The reduction of speed might be as
Equation (17. 24) above is replaced by
high as 55%, but averages roughly 25%.
Tornadoes are severe whirlwinds, typically a hundred
P(x;n) = exp[ – nexp( – y)]. (17.26) meters iil diameter, which fortunately are rare accompani-
ments of severe thunderstorms. They are most common in
To estimate x for probability P in n years, or an n-year risk Australia and the United States where they number some
(l-P), and knowing the annual extreme mean X and standard 140 to 150 per year, most frequent in spring in the plains
deviation (s.), Equation (17 .25) then gives area of the United States between the Rockies and the Ap-
palachians. The wind speeds are generally less than 110 m/s
in the lowest 10–20 m above the ground. Speeds up to
(17.27)
140 ds occur occasionally in very severe tornadoes; higher
wind speed estimates are generally suspect [Golden, 1976].
where y = tn (n) – tn(–tn (P)). (17.28) These high wind speeds place them outside the distributions

17-25
CHAPTER 17
TableIT.18. Extreme annual wind speed (fastest M/s) at 15.2 m above ground at the given stations; (A) denotes airpofl station

1%
Years Risk
of Mean S.D. ‘ in 10 yr
Station Record (M/s) (M/s) (M/s)

Tampa, Fla (A) 1941-56 23 3.9 42

Miami, FIa 1943-58 24 8.0 64
Wilmington, N.C. (A) 1951–58 30 7.1 65
Hatteras, N .C. 19 12–57 28 6.0 58
Dallas, Tex (A) 1941-58 23 2.9 38
Washington, D.C. (A) 1949-58 22 3.8 41
Dayton, Ohio (A) 1944-58 27 3.8 46
Atlanta, Ga. (A) 1933-58 22 3.3 39
Abilene, Tex (A) 194:–58 28 6.1 59
Columbia, Mo (A) 1949-58 25 2.8 39
Kansas City, Mo 1934-58 25 3.1 40
Buffalo, N.Y. (A) 1944–58 26 3.7 44
Albany, N.Y. (A) 1938–58 23 3.8 42
Boston, Mass (A) 1936–50 26 5.4 53
Chicago, 111.(A) 1943-58 23 2.5 35
Cleveland, Ohio (A) 1941-58 26 2.6 39
Detroit, Mich (A) 1934–58 22 2.6 34
Minneapolis, Minn (A) 1938–58 23 5.0 48
Omaha, Nebraska (A) 1936-58 26 2.9 55
El Paso, Tex (A) 1943-58 26 2.0 36
Albuquerque, N.M. (A) 1933-58 27 4.6 50
Tucson, Ariz 1948-58 22 3.2 38
San Diego, Calif 1940–58 16 2.7 30
Cheyenne, Wyo 1935-58 28 3.1 43
Rapid City, S.D. 1942–58 30 3.0 44
Bismarck, N.D. 1940–58 30 2.3 41
Great Fall, Mont. 1944-54 29 1.6 37
Portland, Ore 1950-58 25 3.0 41
New York, N.Y. 1949-58 26 2.1 37
Pittsburgh, Pa. 1935–52 23 2.8 37

Number
of Years
of Data

Vairbanks, Alaska 9 17 3.7 35

Nome, Alaska 11 27 4.1 47
Slmendorf AFB, Alaska 14 20 3.2 36
$hemya Island, Alaska 10 31 7.8 45
flickam AFB, Hawaii 17 20 3.8 38
21ark AB, Philippines 13 17 5.5 45
~ajes Field, Azores 13 28 7.6 65
41brook AFB, Canal Zone 18 12 1.8 21
San Pablo, Spain 11 34 6.8 68
Wheelus AB, Libya 14 22 5.3 48
Stuttgart, Germany 15 18 2.1 29
Keflavik, Iceland 9 38 4.8 62
Thule, Greenland 14 36 5.5 63
rainan, Formosa 39 24 9.5 71
raipei, Formosa 39 26 9.8 75

17-26
Table 17-18. (Continued)

1%
Years Risk
of Mean S.D. in 10 yr
Station Record (m/s) (m/s) (m/s)

Itazuke AB, Japan 14 19 4.5 42

Misawa AB, Japan 11 21 3.2 37
Tokyo Intl. Airport, Japan 15 23 5.5 46
Kimpo AB, Korea 8 19 3.6 37
Bombay, India 6 22 6.3 54
Calcutta, India 6 25 3.3 42
Gaya, India 6 23 3.0 38
Madras, India 6 20 3.4 37
New Delhi, India 6 23 1.7 31
Poona, India 6 17 2.7 31
Central AB, Iwo Jima 17 35 16.9 119
Kadena AB, Okinawa 14 37 11.3 93

discussed previously. Their small dimensions and infrequent is 1915 m above mean sea level and the anemometer was
occurrence make it impossible to state precise probabilities mounted at 11.6 m.
of occurrence at specific locations. Statistics of occurrence Operationally speaking, the greatest wind extremes are
on an areal basis have been developed by Abbey [1976] and typically in northern Scotland. In the windiest month the
Kelly et al. [1978]. 1Yo extreme is set at 22 m/s (a 1-rein wind at 3 m above
Hurricanes, while much more frequent, larger, and more ground level). Gusts accompanying this extreme are esti-
predictable, still cause problems for-the statistical analyst. mated, by the shortest dimensio-n of equipment, at
Recent studies at the National Bureau of Standards suggest
that the Gumbel (Type I) distribution of extreme winds does
not provide as good a description of the extreme wind dis- I Dimension (m) Speed (m/s)
tributions when they must include hurricane winds. The <0.6 32
departures of the winds from the Gumbel estimates tend to
1.5 30
be on ,the weaker side. This means that the estimates of
3.0 29
Table 17-18 should be used with caution. Hurricanes may
7.6 27
strike at some stations while missing others, so that the
15.2 26
record extreme wind speeds of one station may not represent
30.5 25
the potential risk of still stronger winds at another nearby
station or even at the same station. We conclude, however,
that the Gumbel distribution is the best single model for all
extreme winds data including hurricane winds. 17.4.5 Structure of Jet Streams
Directly recorded data for record wind extremes are rare
due to damage or destruction of the wind measuring instru- Belts of exceptionally strong winds are jet streams. (The
ments and power outages, during an event such as a tornado term as commonly used applies to the strong westerly winds
or hurricane. Newspaper reports give estimates whose re- found at the base of the stratosphere. ) Because the mete-
liability is unknown and variable. Some might be from a orological disturbances vary in” size and intensity and are
weather station where visually observed wind-speed dials moving, the jet streams vary in extent, severity, and loca-
were still in operation even though the recorder lost power. tion. Geographically-oriented descriptions obscure signifi-
Others might be calculated from the amount of force re- cant features. As Figure 17-17 shows, this is overcome by
quired to blow over a building or tree or to drive a metal relating the features to the position of the core. In winter,
rod through a wooden post. the centers (or cores) of jet streams are found at altitudes
The recognized worldwide maximum wind speed mea- between 9 and 12 km in latitudes 25° to 70° in each hemi-
sured at a surface station is a 5-rein speed of 91 m/s and a sphere. In summer, jet streams are weaker, higher (11 to
1-s gust of 101 m/s measured at Mt. Washington, New 14 km), and farther poleward (35° to 75° latitude). Generally
Hampshire Observatory on 12 April 1934. Mt. Washington speaking, a jet stream is several thousand kilometers long,

17-27
CHAPTER 17
.

.~-S~......---........----------
‘“
lW

~1
Sm .5
2W ..
---
~- =-- z----
.---1
-
2

---
—— - c.- ----- -------- -------- “-”

,------- ------ .- ------- ----, ~

------- ------ z
~x-a - -------- ------”””- --
~ ------- ---------
. . ? ---- :
----- ---------------- .
: ------ ----- --8 k
> . --------- “p-r
. . . -- 5
:W a% ,/--<-:------- .-- .--<
-----
..-.. -------” ., --
:
. ...------”-- ‘- >... * 6
-------- -----
5& -m -.:.;.=-~~”
. . ..------- ”---
, *,m , .-. -e---
a . . . . . . . . ..-
. . ..------- ”--”- : ~*o*r
---------
. ------------ ,’ . . . -------
4
NORTH~O ’00 0 ’00 200 30~&~”TH
----------
ma . . . ------- . . . ------- . . .

— lsor~cns ---tWTHEnUS (XHEUbTtC1 ----- ,,’

_ lROWWUSE —LbvER 80uNOARIES 2
830 --- LEVEL OF MhXIUUU wIUD@ JEISTREAUCOnE ‘ ----- Figure 17-18. Example of wind and temperature fields near the jet stream.
WRIH W7H

4—321) 1234; b

LbTITUDE [Ma)

(for example, southern India) at an altitude of about 14 km

Figure 17-17. Idealized model of the jet stream, average structure in a in summer.
cross section perpendicular to the flow. (Percent of the core
speed is given for each isotach; wind direction is into the
Frequently it is too expensive to design equipment to
page. ) operate under extreme conditions. In such cases, calculated
risks are assumed or alternate methods of operation are
devised for periods when such conditions exist. If high wind
several hundred kilometers wide, and several kilometers speeds or strong shears normally associated with them ad-
deep. versely affect the operation of a particular vehicle that is
The typical structure of jet streams is shown in Figure released vertically into the atmosphere, it may be desirable
17- 17; this is a cross section showing isotachs (lines of equal to select several release sites spaced far enough apart to
wind speed blowing “into the page”) labeled in percent of ensure that operational conditions will exist above one of
the wind speed at the jet core. Core wind speeds range from the stations when the critical wind speed or shear is exceeded
50 mls to as much as 130 mls on rare occasions. Speeds of above the others. In such cases, Figure 17-17 will be useful
75 ds are common in winter. Figure 17-17 also shows the in selecting alternate sites; it indicates the average rate with
thermal field, the tropopause, and the level of maximum which wind speed decreases with distance on both sides of
wind. Typical wind and temperature variations across jet the jet stream core. For example, with a 100 m/s wind at
streams at a constant altitude or vertically through jet streams the jet stream core, Figure 17-17 indicates that, on the
can be determined from this figure. average, a belt of winds exceeding 70 mls can be expected
Individual jet streams are variable and may differ con- to extend more than 300 km south and nearly 200 km north
siderably from the typical. In some cases, two or even three of the core. Consequently, if the maximum wind likely to
jet cores may be found paralleling each other, separated by occur over a region is 100 m/s and the vehicle is only
distances of several hundred km. In addition to these de- designed to withstand speeds up to 70 m/s alternate release
viations from the average, eddies of various sizes are present
in the wind and temperature fields, as shown in Figure
17-18. Extreme horizontal changes of wind and temperature
are of the order of 20 m/s and 5 K in a 16-km distance
measured perpendicular to the jet stream. Variations parallel
to jet streams are an order of magnitude smaller.
Clear air turbulence is closely related to jet stream struc-
ture; measurements show that it is more likely in certain
regions than in others [McLean, 1962]. At the level of
maximum wind (where vertical shear changes sign), tur-
bulence was observed about 14% of the time. From Figure
17-19 it can be seen that turbulence is most likely to be i $ JETYSTR;AM ,:ORE
w r P 3“ 4“ ~u,”
found north of the jet stream core near the polar tropopause MORTW
400 200 0 200400
and above the core along the tropical tropopause. Indication
DISTANCE FROM JET STREAM CORE (km)
of a maximum is also found in the jet stream front.
Other wind currents which exhibit properties similar to
Figure 17-19. Turbulence in various sectors of a typical jet stream cross
those shown in Figure 17-17 are the low-level jet and the section. (Frequent y ( 7C ) of occumence is shown for each
tropical easterly jet stream which is observed in certain areas contour. )

17-28
 WINDS

sites should be selected at least 500 km apart. As the jet wind velocity. It is inversely proportional to density and is
stream axis occasionally approaches a north-south orienta- large enough above 300 km to prevent development of large
tion, selected sites should be more than 500 km apart in vertical and horizontal shears. Ion drag is a collisional in-
longitude as well as in latitude. teraction between charged particles and neutral particles.
Outside the auroral zone, where electric fields are small,
the ions above 120 km can be assumed to corotate with the
earth. A neutral wind flowing through the ions experiences
17.5 THERMOSPHERIC WINDS a collisional drag that is a maximum at the peak of the
ionospheric layer. A global distribution of electron density
Neutral air motions are a major key to our understanding [for example Ching and Chiu, 1973] is used to determine
of the upper atmosphere at therrnospheric altitudes. A va- the ion drag in various model calculations.
riety of phenomena associated with the neutral and ionized The important terms in the equations of motion of the
atmosphere need to be explained in terms, of a unifying neutral air velocity U are
description of thermospheric dynamics. Neutral density,
composition, and temperature characteristics are related to dU
=F–2QXU –~ni(U_V)+~V2U +g
geomagnetic storm-induced circulation processes and sea- x
sonal and diurnal tides. Aspects of ionospheric structure
depending directly on winds and compositional transport (Acceleration) = (Pressure gradient force) – (Coriolis
include maintenance of the nighttime ionosphere, spread F- term) – (Ion Drag) + (Viscous Drag) + (Gravity)
layers, scintillations, and geomagnetic storm effects. where
The dynamics of the upper atmosphere are controlled
mainly by solar EUV and UV heating, waves and tides dU au
—+(U” V)U
propagating upward from the mesosphere, ion drag inter- X=at
actions with the ionosphere, and energy and momentum
sources associated with auroral region processes (magne- and V is the ion drift velocity, Q is earth’s angular velocity,
tospheric convection and particle precipitation). The major F is the driving force per unit mass due to pressure gradients
heat source, due to solar electromagnetic radiation, is de- determined from a model, ~ni is the appropriate neutral-ion
posited mainly at low and midlatitudes and is reasonably collision frequency, w is the coefficient of kinematic vis-
well understood. At high latitudes, magnetospheric con- cosity, p is the atmospheric den sit y, and g is the acceleration
vection effects are important even during geomagnetically due to gravity. Since this equation contains the ion velocity
quiet times. They dominate the high latitude circulation and the collision frequency that is proportional to the ion
during geomagnetic storms. Magnetospheric convection concentration, it is necessary to take into consideration the
processes are extremely variable and are not well understood equation of motion and continuity equation for the ions.
(see Qapter 8). The phenomenon involves ionospheric plasma Methods of solutions to this set of equations for thermos-
driven into motion by electric fields interacting with the phenc altitudes have been extensively described by Rishbeth
neutral atmosphere via collisions. The neutral constituents [1972]. (Note that the definition of winds does not include
gain energy through Joule dissipation of currents and gain the coronation of the atmosphere. Results in this chapter are
momentum through ion-neutral drag [Straus, 1978; Mayr et presented in a coordinate system that rotates with the earth).
al., 1978]. The accuracy of all models of global winds systems is lim-
Most of our present knowledge of thermospheric neutral ited by lack of experimental data and self-consistent model
winds is based on calculations made by using semiempirical solutions. Complete specification of the wind fields requires
dynamic models of the neutral atmosphere. The pressure solving a large set of non-linear partial differential equations
gradients inherent in these density models provide the pres- in three dimensions that relate the hydrodynamic and elec-
sure forces that drive the therrnospheric circulation. At ther- trodynamics properties of the neutral and ionized components
mospheric altitudes the collision frequency is sufficiently in the atmosphere to the energy, mass,, and momentum
high that the air may be regarded as a fluid subject to hy- sources of the magnetosphere-ionosphere-themosphere
drodynamic equations of motion. Therefore, the equations system.
of fluid motion can be integrated to determine the wind A brief review of thermospheric wind properties de-
system that should result from the pressure distribution. duced from experimental data is given in Section 17.5.1.
These equations are the same as those used by meteorolo- The remainder of this chapter describes winds data derived
gists to study the lower atmosphere, but are modified to from various models, which are only briefly described here.
include two additional terms that are important in the ther- References provided should be consulted for the theoretical
mosphere: a viscous force and an ion drag force. Kinematic aspects and for detailed descriptions of the results. Numer-
molecular viscosity transfers momentum between various ous plots excerpted from the relevant journal articles, sum-
altitude regions and thus smoothes out vertical gradients in marize these data. The development over the past decade

17-29
CHAPTER 17
of a general circulation model at the National Center for afternoon. A nighttime transition occurs in early morning
Atmospheric Research (NCAR) is traced historically. Vari- (winter) or near midnight (summer). During geomagnetic
ations in circulation as a function of season, solar cycle, storms, data show the influence of high latitude heat sources.
altitude, and latitude are given in Section 17.5.2. Local time Observed effects include enhanced nighttime winds, earlier
variations are shown in Section 17.5.3. Results of including nighttime transition of zonal winds from eastward to west-
the effects of magnetospheric convection are shown in Sec- ward, a tendency for a midnight “surge” of winds, and
tion 17.5.4. This section also includes the recent work, occasionally enhanced poleward daytime winds. Fabry-Perot
described in a series of papers, of the University College data at Fritz Peak, Colorado, showed nighttime peak equa-
of London group in determining the influence of magne- torial winds of 650 m/s were linearly related to the time
tospheric convection. At present, the effects of heating by rate of change of the equatorial ring current, and peak equa-
particle precipitation at high latitudes and by upward prop- torial meridional and westward zonal winds were related to
agating atmospheric tides and momentum deposition by ti- the square of the auroral electrojet index.
des at lower latitudes are not included in the above models. At near-equatorial latitudes the circulation is primarily
There are insufficient data on which to build an adequate controlled by solar heating, upward-propagating tides from
model of particle precipitation. Tidal phenomena have been below 100 km, large-scale neutral gas-plasma interactions
extensively described by Forbes and Garrett [1979] and by (such as E x B drifts), and effects related to geomagnetic
Forbes [ 1982a,b] and should eventually be incorporated as storms. A unique feature in this region is the equatorial
part of therrnospheric dynamic models. midnight temperature bulge discovered in incoherent scatter
and satellite data. This persistent but variable feature has
been attributed to a complex interaction between the up-
ward-propagating semidiumal tide and ion-neutral momen-
17.5.1 Observational Summary tum coupling associated with the diurnal variation of ion
and Limitations density. A general trend observed is that nighttime equa-
torward winds weaken and sometimes reverse after midnight
Direct measurements of winds in the thermosphere are in response to this bulge. Othei features of the low latitude
difficult to make and there have been insufficient observa- circulation include seasonal variations with stronger equa-
tions for development of empirical models. Techniques in- torward winds in summer than in winter, and zonal winds
clude ground-based Fabry-Perot spectrometer measure- that are eastward throughout the night during winter and
ments of the 6300 ~ line in the airglow [for example, Hays spring but have a westward reversal during early morning
and Roble, 1971; Sipler et al., 1982] and incoherent scatter hours in the summer.
radar techniques [for example, Evans, 1975], rocketbome At high latitudes, F region data indicate magnetospheric
chemical release experiments [for example, Kelley et al., convection is the dominant factor controlling wind systems.
1977], and satellite data from mass spectrometers [for ex- Fabry-Perot measurements [Hays et al., 1979] show early
ample, Spencer et al., 1976] and accelerometers [DeVries, evening westward winds in the direction of magnetospheric
1972; Marcos and Swift, 1982]. These limitations are being convection and give evidence of a heat source located near
ameliorated at F region altitudes with Dynamics Explorer local midnight. Heppner and Miller [1982], analyzing 39
satellite data from the Fabry-Perot Interferometer [Hays et chemical-release rocket experiments at high latitudes, found
al., 1981 ] and Winds and Temperature Spectrometer [Spen- that winds in the magnetic dusk sector above 160 km follow
cer et al., 198 1]. An extensive review of recent thermo- the magnetospheric convection pattern with a time lag of
spheric wind measurements has been given by Roble [1983]. about two hours. They also observed a 140-160 km altitude
General features of wind patterns deduced from these ex- transition region in the wind pattern. Meriwether et al. [1982]
perimental data in this reference are summarized below for used airglow data to show strong magnetospheric convec-
different latitude regions. For more detailed information on tion-driven winds with superimposed variability probably
various experimental techniques and results obtained, the due to particle precipitation. Dynamics Explorer (DE) sat-
reader should consult Roble [ 1983] and references therein. ellite data analyzed by Killeen et al. [1982] show wind
The general circulation at midlatitudes at F region heights velocities of 600-800 m/s. They further show that the DE
is controlled mainly by solar heating during quiet geomag- data can be described as a superposition of a “background”
netic conditions and high latitude heat sources during storms. solar-driven wind field and a wind field driven by high
Experimental data obtained when geomagnetic activity is latitude energy and momentum sources. DE data of Spencer
low indicate a flow from the high temperature dayside of et al. [1982] showed zonal winds of about 1 km/s related
the thermosphere to the low temperature nightside. Daytime to convecting ions. These strong zonal winds frequently
poleward meridional winds are 50–100 m/s and nighttime correlated with vertical motions having velocities of
equatorward winds are about 100-200 m/s. These values 100–250 m/s. Knutson et al. [ 1977] also found evidence of
depend on ion drag and season. Zonal winds are generally large vertical winds, up to 65 ds, in the early morning sec-
found to be westward before local noon and eastward in the tor of the auroral zone, with no simple comelation with Kp.

17-30
 WINDS

17.5.2 Diurnally Averaged Winds circulation properties are schematically represented in Fig-
ure 17-20. This figure shows the mass flow of the cellular
17.5.2.1 Seasonal Variations at Solar Maximum. The motion above about 80 km. Figure 17-20a shows the cir-
zonal mean (diurnally averaged) thermosphenc circulation culation during a very quiet geomagnetic period. Upward
has been calculated for solar maximum equinox and solstice motion occurs over the equatorial subsolar point and flows
conditions respectively by Dickinson et al. [1975, 1977] poleward to high latitudes where sinking motion occurs.
using a two-dimensional (time-dependent) model run under The small reverse circulation at high latitudes is due to a
steady state conditions. Their model utilized a zonally av- small high-latitude heat source as well as persistent solar
eraged version of lower atmospheric dynamic meteorology heating over the polar regions. During average geomagnetic
equations. Heat and momentum sources used were solar activity (Figure 17-20b) the equatorward circulation is larger.
EUV and UV heating, high latitude heating due to auroral Below -150 km the strong solar UV heating still maintains
processes (assumed distributed unifo~ly in longitude) and a poleward flow. At higher altitudes the flow is equatorward
a momentum source resulting from correlation of the diurnal to about 20° latitude. During large geomagnetic storms (Fig-
variation of winds and ion drag. This model uses a 5° lat- ure 17-20c) the equatorward circulation extends to the equa-
itude-by-longitude grid with 24 constant-pressure surface tor and circulation below 150 km is also modified. The
layers extending in altitude from about 100 to 500 km. (Note situation is similar during solstice. Under geomagnetically
that the altitude of a constant pressure surface can vary. ) quiet conditions (Figure 17-2 1a) the asymmetry in solar
Their results showed that while the winds are driven pri- heating drives a summer-to-winter circulation. For average
marily by heating due to solti EUV and UV absorption, a geomagnetic conditions (Figure 17-21 b), the high-latitude
high latitude heat source was also required to give an equa- heat source reinforces the summer-to-winter circulation in
torward mean meridional circulation in agreement wjth in- the summer hemisphere but forces a reverse circulation in
coherent scatter observations [Evans, 1972]. Equinoctial the high latitude winter hemisphere, with transition from

EQUINOX
i
SOLSTICE
I
I I /~i I K I I
600
500
400
300
200
100
01 1 1 I 1 1 1 1 ) I 1 1 1 1 1 I 1 I J
0

400

I 00
01 1 1 1 1 1 1 I 1 1 1 1 1 I 1 I 1 1
01 I 1 1 1 I 1 1 I I 1 I 1 I t I t I I

I I I 1 I [ I 1 1 t 1 1 I I f I
600 ~---------- s
600 (c) /.& {
500 500 \
400
300 ; :t-**g~’~
200
I 00
01 I 1 I 1 1 i I I I 1 1 I I 1 I 1 I 1 I 1 I t I 1 I t I 1 I , I 1 I 1 I
n
S00 60 40 20 0 20 40 60 80 N “S 00 60 40 20 0 20 40 60 80 N
LATITUDE (Degrees) LATITUDE (Degrees )

Figure 17-20. Schematic diagram of the zonal mean meridional circulation Figure 17-21. Schematic diagram of the zonal mean meridional circulation
in the earth’s thermosphere during equinox for various ~ev- in the earth’s thermosphere during solstice for various levels
els of auroral activity (a) extremely quiet geomagnetic ac- of auroral activity (a) extremely quiet geomagnetic activity,
tivit y, (b) average activit y ( 101xerg/s), and (c) geomagnetic (b) average activity ( 10’x erg/s), and (c) geomagnetic sugs-
substorm (1019 erg/s). The contours schematically illustrate torm (101“ erg/s). The contours schematicallyy illustrate the
the mass flow, and the arrows indicate the direction of the mass flow. and the arrows indicate the direct ion of the
motion [Roble, 1977]. motion [Robie, 1977].

17-31
CHAPTER 17
one cell to another in the midlatitude winter hemisphere. the summer hemisphere. The summer-to-winter circulation
Below -150 km the summer-to-winter circulation is main- intensifies from 50 to 200 m/s as the season progresses
tained at all latitudes. During large geomagnetic storms toward the solstice conditions of Figure 17-22c and d. Above
(Figure 17-2 Ic) the circulation is equatorward above 300 -200 km an intensification of the summer hemisphere equa-
km in both hemispheres, with an asymmetry at lower alti- torward circulation penetrates into the winter hemisphere
tudes. and forces the boundary between the two oppositely directed
Roble et al. [1977] have extended this analysis to de- winds poleward as the season progresses toward solstice.
termine seasonal circulation patterns for solar maximum This boundary is near the equator at equinox and at -60°
(F = 165 units) and solar minimum conditions (F = 80 latitude for solstice. Also, the strongest equatorward winds
units). A solstice-type circulation is maintained except for in the high latitude winter hemisphere decrease from about
about a week either side of equinox. This circulation is most 50 mls at equinox to about 30 mls at solstice.
intense at solstice and weakens as equinox is approached. The derived zonal winds during equinox (Figure 17-23a)
The transition period is controlled in part by the high-latitude are relatively weak above 200 km. As the season progresses
heat source which in turn is related to geomagnetic activity. (Figure 17-23b and c) there is a gradual increase in wind
For higher geomagnetic activity the transition from equinox speeds above -150 km in the summer hemisphere and be-
to solstice circulation occurs more rapidly. The derived diur- low 150 km in the winter hemisphere. At solstice (Figure
nally average meridional, zonal, and vertical winds for solar 17-23c) there are generally westward winds in the summer
maximum conditions are shown in Figures 17-22–17-24 hemisphere and eastward winds in the winter hemisphere.
respectively. These results correspond to average geomag- In the winter hemisphere the velocity at -120 km increases
netic conditions. to 120 m/s. Above 200 km the maximum solstitial velocities
During equinox the meridional flow (Figure 17-22a) be- increase to 60 m/s in the summer hemisphere and 40 m/s
low 150 km is directed from the equator to both poles at in the winter hemisphere.
about 5 mls. At higher altitudes the pole-to-equator flow Vertical velocities shown in Figure 17-24 are given in
has a maximum velocity of about 50 m/s near 60° latitude units of cm/s. Below -120 km the vefiical velocity at equi-
in both hemispheres. By October 6 (Figure 17-22b) the nox (Figure 17-24a) is upward in the equatorial region and
circulation below 150 km is almost completely reversed in downward at middle and high latitudes. By October 6 (Fig-

SUMMER WINTER SUMMER WINTER

POLE SEPT 21 POLE POLE OCT 6 POLE

6 6W
600 - 600 -

500 - 5m
500 - 4
400
400 - 400 -
2 z
z 300.
: 3W – w -
z 250:
+ 250 - 250 -
0
2W - 200 z
$200 -
u G
I -2 150 I
I 50 - t50 -

120
‘2”1 -4L I i
120 120 -4

100 ,~ 100
100 -~ -6
o 0
t L-lIGA 1
-90 -60 -30 0 30 60 90 -90 -60 -30 0 30 60 90
(o) LATITUOE (DEGREES) (b] LATITUOE (OEGREES)

SUMMER WINTER SUMMER WINTER

POLE OCT 21 POLE POLE DEC 21 POLE
700 -
m 500
6 600 - 6
600 -
500
500 - 4eu
500 - 4 4
400 400 -
400 -
300 ~
2 z 2
:300 300=300 -
2505
~ 250 - z 250 ~ 250 - z
o 0 200:
~ 2W - 200 z 200 -
~ E G
150 - -2 150 x ,50 – -2 150 x

lm - 120 I20 - 120

-4 -4

100-
-6~’Oo
-90 -w -30
LATITuOE
0 30
[OEGREES)
60 90
001 -6

-90
[d)
-60 -30
LATITUOE
0 30
[OEGREES)
60 90
IOQ

(c)

Figure 17-22. The calculated zmral mean contours of meridional wind com~nent (in m/s) for 4 days representing the seasonal variation during solar
maximum (a) equim]x, September 21. (b) October 6, (c) October 21. and (d) winter solstice, December 21 [Roble et al. 1977].

17-32
 SUMMER
POLE OCT 6
WINTER
POLE
WINDS
6-

0 0 0
4 -

z
2

1s0

120 -4
{

Im -6 –
I ~ t I 1 1 1 1 1 I
–90 90 -90 -6a -so o w 6a 9a
{a) LATIT~ (DEGREESl (b) LATITE (DEGREESI

SUMMER SUMMER WINTER

POLE OCT 2 I POLE DEC 21 POLE

Soo -

son -

400 -
F
:s00 -
1- 2s0 -
3200 -
~
Isa -

120–

Um-

-(c) LATITW (OEGREESI [d) LATITUDE (DE~EES)

FIgrrE 17-23. The calculated zonal mean contours af zonal wind component (in m/s) for 4 days representing the seasonal variation during solar maximum
(a) equinox, September 21, (b) October 6, (c) October 21, and (d) winter solstia, hmher 21 [Roble et al, 1977].

WINTER SUMMER WINTER

SUMMER OCT 6 POLE
POLE POLE
---
PO1 F SEPT 21

600
500 –

400 -
F
:300 -
+ 2s0 -
:Zoo -
G
I
150 –

120–

Im -
-61,,,,,,,Lt..t.’oooo
-90 -60 -30 :306090

-[-a ) -- LATITUDE (DEGREES) {b) LATITUW (OEGREES)

,., _6~1m I.L:w&.#;w

-90 -w -w o 30s090
[d) LATITUW (OEGREES)
(c) LAT ITUOE !OEGREES)

Figure 17-24. The calculated zonal mean contours of vefiical wind component (in C~S) for 4 days ~P~=nting the ~a~nal variation during snlar
maximum (a) equinox, September 21, (b) October 6, (c) October 21. and (d) winter solstice, December 21 [Roble et al. 19771.

17-33
CHAPTER 17

~ L. ~1 t : 1A i _w~ .$...$ L ‘$
o 30 60 90
[0) LATITUDE [DEGREES) [b) LATITUDE (DEGREESI
WINTER POLE SUMMER E
~–T
~1-~11
I ~6
1,
1 1
I :;
4U0 ~
: .& I
-20
I : 3~_m z
I F
250 E .
: t ~-
1 5250 ‘
!
+
3200 0 $
1 u
h
1501, l–’ x

120 120 .4

,_6w<,,,,,24:”~
100 ,00
-6

-90 -60 -30 60 90 -90 -60 -30 o 30 60 90

LATITuDEOIOEG%ES) (d) LATITuDE [DEGREESI
(c)

Figure 17-25. Cdcrdated eontoura of tie -tints zomd mean temperature and winds for kmber solstim during solar mtilmum: (a) meridional
winds (in m/s), (b) zonal winds (in m/s), (c) tempera- (in K) and (d) vefical winds (in cds) [Roble et al, 1977].

ure 17-24b) these motions become upward in the summer sufficiently large to drive a reverse circulation. Poleward
hemisphere and downward in the winter hemisphere. This winds are encountered at all latitudes in the winter hemi-
pattern intensifies as solstice is approached (Figure 17-24c sphere. The zonal winds (Figure 17-25b) are westward in
and d). Maximum upward velmities occur at high latitudes the summer hemisphere and eastward in the winter hemi-
and maximum downward velocities occur at mid-latitudes. sphere. These winds m similar in magnitude to those of
The main seasonal effect above -200 km is the increasing solar cycle maximum. In general, the zonal jets near 120
asymmetry between hemispheres as the season progresses. km are slightly greater during solar maximum than dting
At high latitudes the summer hemisphere upward veloci- wlar minimum. At higher altitudes the converse is true.
ties are about six times greater than those in the winter Perturbation temperature data are shown in Figure 17-25c
hemisphere. but are not discussed he~. Vertid velocity mults are shown
in Figure 17-25d. In the summer hemisphere the winds are
generally directed upward except for a small region of down-
17.5.2.2 Solstice Variations at Solar Minimum. Re- ward velocities near 40° atu)ve -175 km. Downward winds
sults of calculations for SOIWminimum (F = 80) Member occur at all latitudes in the winter hemisphere. Vertical
solstice condhions are presented as Figures 17-25 a-d for velocities are smaller during solar minimum than during
the meridional wind, zonal wind, perttirbation temperature, solar maximum.
and vertical wind, respectively. The high latitude heat source
required to obtain reasonable agreement between the model
and experimental data was 4.4 times smaller than that re-
quired for solar maximum conditions. 17.5.3 -I Time Variations
The meridional winds (Figure 17-25a) are directed from
the summer to winter hemisphere at all altitudes and lati- A time-dependent “Three-dimensional General Circu-
tudes. A maximum velocity of 90 tn/s occurs near 330 km lation Model” @M) developed by Dickinson et al. [1981]
at 60° latitude in the summer hemisphere. The magnitude permits specification of winds in latitude-local time coor-
of the winds in the summer hemisphere are lower than those dinates. This model uses the same geophysical inputs as in
at solar maximum by about a factor of two. The derived Dickinson et al. [1975, 1977] and Roble et al. [1977].
high latitude heat source in the winter hemisphere is not Results are presented in Figures 17-26 and 17-27 for solar

17-34
 WINDS

Z=lo Z.)o

90

60

-60

-90 .-
0 6 12 16 24 0 6 12 18 24
LOCAL TIME LOCAL TIME

Z-4 o Z.40

r
.;-.-=..-
90 90 I
.“ .-...,. ..”,lf ~~.<. -. .’-. .- > _.. -.

:--
--=---
-:.-
.--,:.
. .
..: 0-,-,-7. --
--1
L--:--------- “- 7 :/ -/ -?>:>__-: -. . .
60

‘~~”w
60
%-.

‘-”
J -30

-60 -60

..
Qn -s0
,
0 6 12 18 24 0 6 12 16 24
LOCAL TIME LOCAL TIME

Figure 17-26. Calculated global distribution of winds and perturbation Figure 17-27. Calculated global distribution of winds and perturbation
temperature (K) along two constant pressure surfaces: (a) temperature (K) along two constant pressure surfaces: (a)
~ = I at approximately 3N km and (b) z = – 4 at ap- z = 1 at approximately 300 km and (b) z = – 4 at ap-
proximately 120 km for equinox conditions during solar proximately 120 km for solstice conditions during solar
cycle maximum. Both solar heating and a high-latitude cycle maximum. Both solar heating and a high-latitude
heat source are included. The maximum wind arrow is 130 heat source are included. The maximum wind arrow is 180
m/s in (a) and 90 nr/s in (b). [Dickinson et al., 1981]. rrr/s in (a) and 75 m/s in (b) [Dickinson et al., 1981].

maximum equinox and solstice conditions, respectively, at 17.5.4 Magnetospheric Convection Effects
altitudes of 120 km and 300 km. Solid lines shov~n in these on Neutral Winds
figures represent a calculated perturbation temperature from
global mean values derived from the model and are fully More detailed and higher resolution models of ther-
described in Dickinson et al, [1981]. The present discussion mospheric dynamics have resulted from incorporating elec-
considers only the wind results, depicted by arrows. tric field models to provide both an energy and momentum
At 300 km the equinoctial winds blow from the daytime source at high latitudes. This permits including the important
high pressure region near 1700 LT into the nighttime low influence of high-latitude plasma convection. The magni-
pressure region near 0400 LT as shown in Figure 17-26a. tude of the magnetospheric convection is defined by the
Maximum winds of 130 m/s occur at night when ion drag electric potential drop across earth’s polar cap. Values to
is smallest. At 120 km (Figure 17-26b) the maximum arrow 150 kV are possible during strong geomagnetic storms. Cor-
corresponds to 90 m/s. The low latitude dynamic structure relation of polar cap convection potential drop with more
exhibits a semi-diurnal variation. At high latitudes the cir- conventional indicators of geomagnetic activity including
culation remains nearly diurnal. the Kp and AE (auroral electrojet) indices has been studied
Results for solstice (southern hemisphere summer) at PY Reiff et al, [ ] 9g 1], They found that AE had a correlation
300 km are shown in Figure 17-27a. High latitude heating coefficient of 0.71 with the electric field data while Kp had
greatly enhances the flow toward the winter hemisphere. a relatively weak correlation coefficient of 0.55.
The maximum wind vector is in the nighttime summer hemi- Roble et al. [1982] have used the electric field model
sphere with an equatorward velocity of 180 m/s. At 120 km of Volland [1975, 1979a,b] in their Thermospheric General
(Figure 17-27b) the complex circulation pattern again shows Circulation Model. Results have been initially presented to
a pronounced semi-diurnal oscillation at low latitudes and compare two cases: (1) solar EUV and UV heating only and
a dominent diurnal variation at high latitudes. The maximum (2) solar heating plus a 20 kV cross-tail potential to simulate
velocity of 90 m/s occurs for high latitude night conditions. low geomagnetic conditions. These runs assume equinox

17-35
CHAPTER 17

and solar cycle maximum conditions and coincident geo- imum winds increase from 120 m/s to 200 m/s over the
graphic and geomagnetic poles. Circulation and temperature poles. Also, the pattern of circulation in the polar cap is in
data are shown on polar plots for 50”N to 90”N geographic the direction of the plasma drift associated with magneto-
latitude. As in Figures 17-26 and 17-27, arrows represent spheric convection. At lower latitudes the influence of mag-
wind magnitude and direction, and heavy solid lines rep- netospheric convection is much smaller.
resent perturbation temperatures. The TGCM [Roble et al., 1982] was then modified to
Figures 17-28 a and b show the circulation pattern pre- include displaced geographic and geomagnetic poles. Data
dicted at 130 km and 300 km, respectively, in response to are given below for a 20 kV and 60 kV (moderate geo-
solar heating only. These results are the same as obtained magnetic activity) case. Figure 17-30 shows circulation pat-
for Figures 17-26 and 17-27, but in polar plot format. The terns for the 20 kV case at 130 km and 300 km for 0600
characteristic dayside to nightside flow is found at the 130 UT and 1800 UT. At 130 km (Figures 17-30a and b) the
km level (Figure 17-28a). At high latitudes the daily vari- counter-clockwise circulation on the momingside is near
ation pattern is mainly diurnal while a semidiumal com- 80° latitude and 0° longitude at 0600 UT. Figure 17-30b
ponent is indicated at low latitudes. A counter-clockwise (1800 UT) shows it has a UT dependence due to the dis-
circulation is centered near 0600 LT and 75”N latitude. placed poles. The 300 km circulation patterns are shown in
Maximum winds of 75 m/s occur in the polar region. At Figures 17-30c and d. These are analyzed as zonal and
the 300 km level (Figure 17-28b) the strongest winds are meridional components in Figure 17-31. Maximum zonal
again in the polar region. However, the day-to-night flow winds (Figures 17-3 la and b) are about 200 m/s in the polar
is now rotated to the left of that for the 130 km case. region at 1800 UT and are about 150 m/s at 0600 UT.
The effect of including a 20 kV cross-tail potential is Maximum meridional winds (Figures 17-3 lC and d) at high
shown in Figure 17-29. In the lower thermosphere (Figure latitudes are also generally about 50 m/s greater at 1800 UT
17-29a) the dawn-dusk temperature contrast is enhanced in than at 0600 UT. Model results obtained by increasing the
the polar region but the winds are only slightly changed cross-tail potential to 60 kV (moderate geomagnetic activity)
(from a maximum of 75 m/s to 80 m/s). However, in the are shown in Figure 17-32. At 120 km (Figures 17-32a and
upper thermosphere (Figure 17-29b) the overall temperature b) the maximum wind velocity increases to 102 m/s. The
contours are not significantly modified, but the polar region counterclockwise wind rotation around the low temperature
winds are greatly affected in magnitude and direction. Max- vortex also increases in magnitude. At 300 km (Figures

Z=-4.O
Z=l.o
18
18

12
00 12 00

06 06

LOCAL TIME (h) LOCAL TIME (h)

Figure 17-28 Polar plots giving the direction and magnitude of the calculated nofihem hemisphere high-latitude circulation and contours of perturbation
temperature (K) along (a) the z = – 4 (130 km) and (b) z = + 1 (300 km) constant-pressure surfaces for the case of solar heating alone.
Coincident poles are assumed. The wind speed associated with the maximum arrow is 75 m/s in (a) and 150 m/sin (b) [Roble et al., 1982].

17-36
 WINDS

Z9-4.O Z=l.o
18 18

20

22 22
/4

12 00

[-
“‘X6-’”
-40.” --

06 06

LOCAL TIME (h) LOCAL TIME (h)

Figure 17-29. Polar plots giving the direction and magnitude of the calculated northern hemisphere high-latitude circulation and contours of perturbation
temperature (K) including magnetosphetic convection with a cross-tail potential of 20 kv along (a) the z = – 4 (130 km) and (b) z = + I
\——–
f300 –., crmstant-Dressure surfaces for the case of solar heating alone. The wind speed associated with the maximum arrow is 80 m/s in
km)
(a) and 225 ms” in’(b) IRoble et al., 1982].

17-32c and d) the winds in the magnetic polar cap increase The maximum wind speed of 89 m/s occurs directly over
to 380 rnls. These winds are enhanced by ion drag with the the polar region. Vortex development occurs on the mom-
incremental winds tending to follow the twin ion drift vor- ingside of the polar cap, obstructing the seasonal flow from
tices associated with magnetospheric convection. The dou- the high-to-low temperature regions. The effect of mag-
ble vortex wind system at F-region heights extends down- netospheric convection is to redirect the overall flow pattern,
ward into the lower thermosphere. However, there is a resulting in a jet over the polar cap.
significant shift in the pattern with the main features at Results for winter conditions in the northern hemisphere
120 km associated with the cold low-pressure and warm at 120 km are given in Figure 17-34. In each case, a counter-
high-pressure regions. clockwise circulation is driven throughout the high latitude
Roble et al. [1983] have extended the TGCM to examine region by a low-pressure system over the pole. Maximum
seasonal variations for solar cycle maximum conditions. wind speeds increase from 108 m/s to 117 m/s to 120 m/s
This version of the model assumes coincident geographic as the cross-tail potential increases from zero kV to 20 kV
and geomagnetic poles; universal time effects are not in- to 60 kV. Magnetospheric convection effects are less than
cluded. Results are shown for the three cases: (1) solar EUV in summer because of the lower electron density and con-
and UV heating only, (2) solar heating plus a 20 kV cross- sequently decreased ion drag momentum source.
tail potential, and (3) solar heating plus a 60 kV cross-tail Results for the altitude of approximately 300 km for
potential. southern hemisphere summer winds are given in Figure
Results for southern hemisphere summer at the altitude 17-35. Maximum winds increase from 110 ds for the zero
of approximately 130 km are given in Figure 17-33. With kV case (Figure 17-35a) to 200 rrtls for the 20 kV case
solar heating only (Figure 17-33a) the circulation is again (Figure 17-35b). For both cases the neutral winds tend to
from the high temperature noon sector toward the midnight follow the two-cell pattern. For the 60 kV case (Figure
sector where maximum wind velocities of about 75 m/s 17-35c) anti-sunward winds increase to 380 m/s. Further,
occur. With a 20 kV cross-tail potential (Figure 17-33b) at the boundary of the polar cap, the wind direction is
there is very little change in the circulation pattern. The sunward, following the pattern of magnetospheric convec-
maximum wind speed is 77 rrt/s. For the 60 kV case (Figure tion. An equatorward surge in wind velocities is also ob-
17-33c) the circulation pattern changes, reflecting the influ- served near local midnight.
ence of the larger magnetospheric convection ion velocities. Circulation patterns for the 300 km winter case are given

17-37
CHAPTER 17

UT=6 Z=-4.O UT=18 Z=-4.O

-180 -180
-r
-150 -— \lso

-1

-9%

// /20
0 o
(a) LONGITUDE (b) LONGITUDE

UT,6 Z=l.o UT=18 Z=l.o

-180 -180
-150

“0=0
(c) LONGITUDE
0
(d)
-.\-

LONGITUDE
o
J /20

Figure 17-30 Winds and contours of perturbation temperature (K) along the z = -4(130km) surface fortwouniversal times. 0600UT(a) and 1800
UT(b), and along the z = +1 (3W km) surface at 0600 UT(C) and 1800 UT (d) for the case where the geographic and geomagnetic
poles aredisplaced andthecross-tail potential forthemagnetosphefic convection model is20kV. Thecirculation isindicated bythearrc}ws
giving direction, and the length is the magnitude of tbe winds. The length of the maximum arrow represents 76 m/s in (a), 87 m/s in (b).
219m/sin (c), and222m/s in(d), respectively. Local noon isinuicated byanopen circle on the boundary [Robleet al.. 19821.

17-38
 WINDs
UT=6 Z=l.o UT=18 Z=l.o
-180

-30
o 0
LONGITUDE (b) LONGITUDE
(0)

UT=6 z =1.0 UT= 18 Z=l. o

-180 - Im

1’ \ 150

-’o~” o
0
LONGITUDE (d) LONGITUDE
(c)

Flgurc 17-31. Contours of the zonal and meridional wind components at 0600 UT, (a) and (c), and 1800 UT, (b) and (d), mpectively, along tbe z – + 1
(3~ km) surface for the caw where the poles are displaced and the cross-tail potential is 20 kV. kal noon is indicated by an open circle
on the boundary [Roble et al., 1982].

17-39
 CHAPTER 17
UT=6 Z=-4.O UT ❑18 Z=-4.0
-180 -180

-30 ---

0 0
(a) LONGITUDE (b) LONGITUDE

UT=6 Z=l.o UT=18 z= I .0

-180 -1s0

-1

-90

-60\ ,’t’~. ”

o 0
(c) LONGITUDE LONGITUDE
(d)

Figure 17-32. Winds and contours of perturbation temperature (K) along the z = – 4 (i 30 km) surface for two universal times, 0600 UT (a) and 1800
UT (b), and along the z = + 1 (300 km) surface at 0600 UT (c) and 1800 UT (d) for the case where the geographic and geomagnetic
poles arc displaced and tbe cross-tail potential for the magnetospheric convection is 60 kV. me circulation is indicated by tbe mws giving
direction, and the Iengtb is tbe magnitude of the winds. The length of the maximum arrow represents 1~ m/s in (a), 109 m/s in (b), 395
m/s in (c), and 373 nr/s in (d). Local noon is indicated by an open circle on tbe boundary [Roble et al., 1982].

17-40
 WINDs
,2 ~=.40 z: -40

.-- .
LOCAL T!uE (b) LOCAL T,W
{.) L0C4L TM [M (01 LOCAL Tl=

60 xv a z= -40

06
a

00
LWAL 11=
(c) -L TIUE (cl

figure 17-33. Polar plots giving the dircedon and magnitude of“the cat- Flgurc 17-34. Polar plots giving the di=tion and magnitude of the cal-
crdated southern hemisphere (s~r) high-latitude cir- culated northern hemisphere (winter) high-latitude circrs-
culation and contom of perturbation tc~ (K) atong lation and contours of perturbation temperature (K) along
the z = – 4 ( I30 km) constant-pressm surf- at -m- the z = – 4 ( I30 km) constant-pressure surface at f3eccm-
ber solstice for(a) solar heating only, (b) solar heating plus bcr solstice for(a) solar beating only, (b) wlar heating PIUS
magoctoapheric mnvcction with across-tail potential of 20 ~gOCtOSpbCriCconvection with a cross-tail potentiat of 20
kV, and (c) solar heating plus magrsetospheric convection kV, and (c) solar heating plus magnctosphcric convection
with a cross-tail potential of 60 kV. The wind speed as- with a cross-tail potential of do kV. The wind speed as-
sociated with the maximum arrow is 75 m/s in (a), 77 sociated with the masimum arrow is 108 m/s in (a), I 17
m/s in (b) and 89 m/s in (c) [Roble et at., 1983]. (Reprinted nr/s in (b) and 120 0s/s in (c) [Rotde et at., 1983]. (Re-
with permission from Pergamon hss Ltd. O 1983.) prisrtcdwith pcrmissioo from~Press Ltd. o 1983.)

in Flgttrc 17-36. As for the corresponding case at 120 km, Roble et al. [1983] also give meridional winds and tem-
the circulation patterns for & zero and 20 kV cases” are perature data in altitude-latitude coordinates. Results are
similm’, with the maximum wind increasing from 181 (Flg- given for the cases of a zero and a 60 kV potential for local
ttre 17-36a) to 200 trt/s (Figure 17-36b). For the 60 kV case times of 1200 h and 0000 h. In Figures 17-39 (1200 LT)
(Figure 17-36c) polar winds increase to 330 tn/s. These and 17-40 (0000 LT) the top two plots show temperatures
winds also show a surge near 0200 LT. There is again a for the two heating cases and the bottom two plots show
tendency to follow the two-cell pattern of magnetospheric the corresponding wind fields. At 1200 LT, m=imum ve-
convection with a reduced effect due to the lower winter locities for solar heating only are 75 m/s and 125 m/s over
electron densities. the summer and winter poles msptively. The inchtsion of
Pole-to-pole cylindrical equidistant projections of winds magnetospheric convection increases the flow over the pole
and temperatures as a function of local time as given by to 150 m/s and 300 m/s for the summer and winter cases
Roble et al. [1983] are shown in Figure 17-37 for 120 km respectively. Also, the higher summer electron densities
and Figttm 17-38 for 300 km. Ftgttre 17-37 shows that the permit the ion drag acceleration source to accelerate the
major effect of magnetospheric convection in the lower ther- winds down to about 130 km in the summer hemisphem
mosphere occurs at high latitudes in the summer with a 60 compared to about 180 km in the winter hemisphere. Along
kV potential as discussed previously. In the upper ther- the ~ LT meridian, com~arison of Figures 17-40a and
mosphere (Flgttre 17-38) the 60 kV potential has a much b shows the magnetospheric convection ~atly enhances
more pronounced influence on the high-latitude circulation. eqttatorward flow. As the cross-tail potential is i~
Maximum winds of about 400 tn/s occur at high latitudes from zero to 60 kV, winds over the summer pole increase
near local midnight and arc eqttatorward. Changes ~ ob- from 100 tn/s to 400 m/s while over the winter pole they
served even at lower latitudes, indicating that the high- increase from 100 tn/s to 300 trt/s. The sharp wind gradient
latitude energy input is redistributed globally by the ther- near 60° in both hemispheres indicates the equatorward limit
mospheric wind system. of magnetospheric convection forcing for this local time.

17-41
 ,2 2=10 2CIKV 12 Z=lo

00 w 12
b) (b) LWL TIME
(01 LWAL TM w LOCSL ~

60Kv 12 Z=1O
m Z.lo

m m

[c) LOCAL l=
[cl LOCAL TtuE

Figure 17-35. Polar plots giving the direction sod magnitude of ti cal- Flgurc 17-36. Polar plots giving the direction and magnitude of the csd-
culated southern hemisphere (summer) high-latitude cir- culatcd northern hemisphere (winter) high-latitude circu-
culation and contours of perturbation temperature (K) along lation and conturs of pcrturbadon tempemturc (K) along
the z = + 1 (-300 km) cosrstant-~ssurc surf= at Dc- the z = + 1 (-300 km) constant-pressure surface at f3e-
cembcr solstice for (a) solar heating only, (b) solar heating mmhcr solstice for (a) solar heating only, (b) solar beating
plus rnagnctospheric convection with a cross-tail potential plus magnetospheric convection with a cross-tail Pete.ntid
of 20 kV, and”(c) solar heating plus magnetosphcric con- of 20 kV, and (c) solar heating plus rnagrsctospheric coss-
vection with a cross-tail potential of 60 kV. The wind speed vection with a cross-tail potential of 60 kV. Tbc wind speed
astiiatcd with the rnazimum arrow is 110 m/sin (a), 200 associated with the mazimum asTow is 181 m/s in (a), ~
nr/s in (b), and 380 rrr/sin (c) [Roble et aL, 1983]. (Re- ds in (b), and 339 m/s in (c) [Roble et aL, 1983]. (Re-
printed with permission from *amosr Press Ltd. @ 1983.) printed withpermimiifiom~k~. 0 1983.)

20 KV z ❑ 4.0

w
n
2
1=
a
-1

.-
04 8 12 16 20 24 0 4 8 12 16 20 24
l“)
,“,
(h)
LOCAL TIMk-(hours)

@ KV
m“
60
w 30
o
20
i=
a -30
J
–60
-90
(c)
LOCAL TIME (hours)
Figure 17-37 Cylindrical equidistant projections giving the direction and magnitude of the calculated circulation and contours of perturbation tempcrahue
(K) along the z = – 4 ( 130 km) constant-prcssw surface for December solstice for (a) solar ktirsg only, (b) solar heating plus mag-
nctospheric convection with a cross-tail potential of 20 kV, and (c) solar heating plus magnetosphcric convection with a cross-tail potential
of 60 kV. The wind speed associated with the maximum arrow is 108 m/s in (a), 117 m/s in (b) and 120 m/s in (c) [Roble et al., 1983].
(Reprinted with permission fmm Pergarnort Press Ltd. 0 1983.)

17-42
 20 KV z= 1.0
90

60

30

-30

-60
-Q~
.V

o 4 8 12 16 20 24 0 4 8 12 16 20 24
(o) (b)

LOCAL TIME (hours)

90

60
w 30
o
? 0
i=
a -30
-t
–60

-90

(c)
LOCAL TIME (hours)

Figure 17-38. Cylindrical equidistant projections giving the direction and magnitude of the calculated circulation and contours of perturbation temperature
(K) along the z = + 1 (300 km) constant-pressure surface for &cember solstice for (a) solar heating only. (b) solar heating plus mag-
netospheric convection with a cross-tail potential of 20 kV, and (c) solar heating plus magnetospheric convection with a cross-rail potential
of 60 kV. The wind speed associated with the maximum arrow is 181 m/s in (a), 220 m/s in (b) and 380 m/s in (c) [Roble et al., 1983].
(Reprinted with permission from Pergamon Press Ltd. 0 1983.)

The TGCM model has been run to simulate conditions of about 300 m/s. The TGCM simulations used a 60 kV
encountered by the Dynamics Explorer (DE-2) satellite. Re- potential and displaced geographic and geomagnetic poles.
sults from two on-board wind sensors [Killeen et al., 1982; Results for the summer and winter cases are given in Figures
Spencer et al., 1982] were combined to provide thermos- 17-42a and b respectively. The model data generally agree
pheric winds along the spacecraft path, at an altitude of with the direction of the experimental data. However, the
about 300 km. These data were then compared to the TGCM model wind magnitudes underestimate the measured winds.
predictions. Experimental data obtained on the four orbits Maximum model values are 400-500 rrt/s. Roble et al. [1983]
over high southern latitudes, shown in Figure 17-41a are note that better agreement could be obtained by increasing
considered representative of summer conditions. Local times the cross-tail potential and ion-drift velocities in the mag-
are between night and earl y morning hours. Data are shown netospheric convection model, but that such a modification
in polar plots and in geomagnetic coordinates. Further de- is not warranted until the ion-drag momentum source is
tails of these data are in the paper by Roble et al. [1983]. better defined (from DE-2 satellite plasma measurements).
This figure shows strong control of the winds by magne- Roble et al. [1983] have also examined whether a linear
tospheric convection and the characteristic two-cell con- combination of solar-driven and high-latitude auroral-driven
vection pattern. Maximum winds are over the polar cap and circulation and temperature structures reproduce the total
have velocities between 500 and 750 m/s. Winter (northern structure. Such a linear combination would allow high-lat-
hemisphere) polar data for three passes are shown in Figure itude plasma-dynamic interactions with neutral gas to be
17-41 b. Local times are during twilight and daytime hours. analyzed independently, with the results superimposed lin-
Magnetospheric convection control of the winds is evi- early on the solar-driven results. This would have the ad-
denced by their similarity with those for summer conditions. vantages of simplifying analyses of Dynamics Explorer data
Maximum speeds of 500–800 m/s are found in the polar and placing more confidence in linear perturbation model
cap. In the auroral oval, the winds are sunward with speeds approaches to thermospheric dynamic studies. Their finding

17-43
CHAPTER 17

LT = 12 AT LT=12 60 KV AT
5 5
I I 600
500
600 4
4
500
3 - 400 500 3

2 - 2
300 I
I
o o
200 z -1
z -1
-2 -2

-3 -3

-4 -4
-5
-5 >50 100 100
100 -6
-6
_, ~ _7

-87.5 -62.5-37.5 -12.5 12.5 375 62.5 875 -’875 -6z,5 -37,5 -12.5 12.5 375 62.5 875

LATITUDE (a) LATITUDE (b)

LT=12 v LT=12 60KV v

5
511:1 1’ I I I 11 II II J 500
600
600
4
Soo
400 500 3
1
2
300 I
0
z 200 z -1

-2
-3

1 -4
-5
100 100
-6
_, ~ I _, ~
-87.5-62.5-375 -12.5 12.5 37.5 62.5 875 -875 -62.5-37.5 -12.5 12.5 375 62.5 075

LATITUDE (c) LATITUDE (d)

Figure 17-39. Meridional cross sections giving contours of the calculated perturbation temperature (K) [(a) and (b)] and meridional wind speed (M/s.
positive northward) [(c) and (d)], respectively. The local time of the meridional slice is 1200 LT. (a) and (c) are the perturbation temperature
and meridional wind for the case of solar heating only, and (b) and (d) are the perturbation temperature and meridional wind for the case
of solar heating plus magnetospheric convection with a cross-tail potential of 60 kV [Roble et al., 1983]. (Reprinted with permission from
Pergamon Press Ltd. O 1983. )

was that a linear superposition is accurate to within 1070–20% conditions and coincident geographic and geomagnetic poles
at F region heights, where motion is mainly controlled by are assumed. Winds are calculated for altitudes 120 and 240
the ion drag momentum source. In the lower thermosphere km from onset of the storm to 4.5 h after onset at 20-min
the linear combination is not a good approximation, appar- intervals. Results are shown here at t = O, 1, and 2 h to
ently because of the non-linear nature of energy transport indicate the complex nature of the time-varying response.
by winds and of Joule heating which is a function of the Longer term effects are also noted in the text. Data are again
square of the difference between neutral and ion velocities. presented as polar plots from 50°–900 north geographic lat-
It therefore appears that non-linear models are required for itude. Arrows representing winds are scaled to 40 m/s = 2°
lower therrnospheric studies. latitude at 120 km and 160 m/s = 2° latitude at 240 km.
Thermospheric wind variations as a function of time due Pre-storm conditions, t = O, at 120 km (Figure 17-43a)
to a substorm corresponding to moderate geomagnetic ac- show a strong anti-solar flow in the evening polar cap which
tivity have been calculated by Fttller-Rowell and Rees [ 1981]. partially returns via a -40 m/s sunward flow in the morning
These results were derived using the global three-dimen- auroral oval. It is partially deflected in the midnight region
sional, time-dependent numerical model of Rees et al. [1980]. into the westward flow of the evening auroral oval (2000-0200
Fuller-Rowell and Rees [1981] estimate that their input cross- LT) to eventually rejoin the general anti-sunwti flow. These
tail potentials for the example given here represent a vari- results are in general agreement with the 20 kV data of
ation in Kp from about one to about six and back to about Roble et al. [ 1982] shown in Figure 17-29a. At t = I h
one in successive 3-h intervals. Equinoctial solar maximum (Figure 17-43b) peak wind amplitudes of -180 ds occur.

17-44
 WINDS

LT=O AT LT=O 60 KV AT
5r, 5.
II 1 1 1 I 1 I II
500 600
I 4 4
400
400 500 3 -
350
‘\\ 300
400
2 -
I

‘1
,j i!oo
300 o -

200
z -1
-2

-4
-5 - /50
‘5
-6
_, ~
P 50 100 100 -6
-7-
-
1 I 1 1
r50-
-875 -62.5-37.5 -12.5 12.5 37.5 62.5 87.5 -8?.5 -62.5-375 -12.5 12.5 3?.5 62.5 8T5

LATITUDE (a) LATITUDE (b)

LT=O v LT=O 60 KV v
5 I I I 5
I I 600
\ll 500
4 - 600 4
500 - I 50 ~{ 500
3 - 100 400 500 3
400 -
2 - 2 400
400
I 300 I
300 -
300
o - 300 0
200
200 - z -1
200
z -1 200
-2 - -2
-3 - -3
-4

100 -
;~’oo
-875-625-375 -12.5 12.5 37.5 62.5 87.5
I 00
:~’m
-875 -62.5-37,5 -125 12.5 375 625 075

LATITUDE (c) LATITUDE (d)

Figure 17-40, Meridional cross sections giving contours of the calculated perturbation temperature (K) [(a) and (b)] and meridional wind speed (m/s.
positive northward) [(c) and (d)], respectively.The lmal time of the meridional slice is woo LT. (a) and (c) are the per’rurbati~n temrre~ature
and meridional wind for the case of solar heating on] y, and (b) and (d) are the ~rturbation temperature and meridional wind or the case
of solar heating plus magnetospheric convection with a cross-tail potential of 60 kV IRoble et al.. 1983]. (Reprinted with permission fmm
Pergamon Press Ltd. 0 1983.)

Other significant features are poleward winds in the after- Greater wind magnitudes are observed at 240 km. The
noon sector south of 70”N latitude, equatorward winds from anti-sunward flow over the polar cap at t = O (Figure
midnight to midday in the morning region south of 65”N 17-44a) is about 300 m/s (compared with 150 m/s when
latitude and a steadily increasing surge from dusk to dawn solar heating is the only source in the thermosphere). Also,
over the polar cap. At t = 2 h (Figure 17-43c) velocities in the late evening auroral oval (1600-2400 LT) there is a
have fallen to about half their maximum values. Also, an significant sunward flow which forms a cell centered about
anti-cyclone circulation has been established with center 72° latitude with closure over the polar cap. South of 60°N
near 2200 LT and 74° latitude. There is also a complimen- latitude the wind circulation is only slightly modified by the
tary cyclonic cell centered near 1000 LT and 74° latitude. electric field. One hour into the substorm (Figure 17-44b),
These cells have velocities of 160 m/s and decrease very winds in the evening auroral oval and over much of the
slowly with time. At 4.5 h after onset the winds south of polar cap exceed 300 tis. Near 1500 LT and 73° latitude
65°N latitude revert to essentially pre-storm conditions in the velocity is 450 m/s. Horizontal winds converge near
the afternoon and evening. In the morning sector equator- 70°–750 latitude and 1200 LT due to the auroral oval jets.
ward winds of 60 m/s persist. Poleward of 65° latitude the While not evident in the limited data shown here, there are
anticyclonic nighttime and weaker cyclonic daytime vortices strong oscillations due to gravity waves which have already
continue, along with the connecting dusk to dawn flow. propagated to 50° latitude in the early morning hours, af-
Fuller-Rowell and Rees [ 1981] indicate that these vortices fecting both zonal and meridional winds. Wind magnitudes
may persist for several more hours. in the midday region have been significantly decreased but
/

17-45
CHAPTER 17
DE-2 FPi/w ATS 20.85 LT TO 8 85 LT
DO I E ,.74C. 12b 0 11 1052 UT 1181 2Q5
2, 1232 UT 1182 DE-2 FPI/wATS 6.02 L’ TO 18.02 LT
295
FROM -40 TO -gc !2
31 173~ UT 1200 2~6 POLE ( 80.0, 282.0) 11 1929 UT 1810. 337
FROM 40 TO 90 21 2243 UT 1813 337

“-–--’ 5’3“ ’2222“

12
31 332 UT 1816, 338
..-

4)

b
b

M/s

(a) O HRS (b) o bius

Figure 17-41 Polar plots of the measured wind vector at F-region heights along the DE-2 satellite path: (a) winds on four orbits measured over the
southern magnetic polar cap on 22-23 Ott 1981 and (b) winds measured on three orbits over the northern magnetic polar cap on 3-4 Dec
1981 [Roble et al., 1983]. (Reprinted with permission from Pergamon Press Ltd. @ 1983.)

TGCM 0(1 81 bOKV 20 85 LT TO 8.85 L1

POLE [-74 5 127 O) 11 11 OUT
21 13 0 LIT
FROM -40 TO 90 12
31 170u T

(b) O HRS

Figure 17-42. Polar plots of the calculated wind vector at F region heights along the same DE-2 satellite paths shown in Figure 17-31: (a) winds for the
same four orbits over the southern hemisphere magnetic polar cap in October and (b) winds for the same three orbits over the northern
magnetic polar cap in December [Roble et al., 1983]. (Reprinted with permission from Pergamon Press Ltd. O 1983. )
 00.00 00.00
+ +

oh lb

+ m.m
mm+ 06.m +

+
12+C0 12.oo

Oo.w
+

2h

e6.oa + +m.m

+
moo

Flgore 1743. Polar plot of dynamical response of winds at 120 km to a large geomagnetic storm at (a) t = O h, (b) t = I h, and (c) t = 2 h. Winds
m seated to 40 ds = 2° latitude [Fuller-Roweil and Rees, 19811. (Re@~ wi~ pe~ission fs’omPerg-n ~SS Ltd. @ I~ 1. )

not reversed. Wind vectors over the polar cap appear to are again observed. On the nightside (2200 LT to 0900 LT)
converge toward the midnight part of the auroral oval. ~is strong oscillations of both meridional and zonal winds occur
is due to the poleward gravity wave launched by the dayside w,th latitude and longitude. Also, the dayside anti-solar
auroral oval. The high velocity ion flow over the polar cap circulation has been stopped, or “deflectccl” since the late
also provides momentum to drive surges of equatorward morning winds are strongly enhanced relative to the case
winds at midlatitude, particularly betw~n midnight and for solar heating only. At t = 4 h, the model runs show
dawn. At t = 2 h (Figure 17-44c) the morning circulation that the late morning circulation is still enhanced and that
cell is tending to collap=, but the evening cell is hardly gravity waves reach mid to low latitudes. Simulations run
reduced. Polar cap winds have steadily in-seal to maxi- out to 12 h after onset show that the wind circulation has
mum vahtes greater than 600 rrs/s in the night sector between not yet returned to prestorm conditions.
MO–700 latitude. Evidence of gravity wave characteristics An improved version of this model has been developed

17-47
 CHAPTER 17

moo Woo
+ +

Oh

Ill

%.00 + +m.oa 0s.00”+ + Is.m

12+W
Q+oo

moo
+

2h

06.00 +
+ m.w

Is+w

Flgum I7-44. Polar plot of dynarnicat response of winds at 240 km to a large geomagnetic storm at (a) t = O h, (b) t = I h, and (c) t = 2 h. Winds
are sealed to 160 ns/s = 2° latitude [FuI1er-Rowell and Rees, 1981 ]. (Reprinted with permission from Pergamon press Ltd. 0 198 i. )

[Rees et al., 1983]. It incorporates the theoretical treatment The DE-2 wind data, denoted “KVEC”, m shown for
by Quegan et al. [1982] of the polar and auroml ionosphere. each orbit in Figures 1745 and 17-46 as a function of
This results in a much more realistic elatron density dis- geographic latitude. The left side of the graph shows evening
tribution at high latitudes during other than low geomagnetic data and the right side of the gmph shows morning data for
conditions since ionization enhancements due to magnetos- the October data (Figure 17-45) while the reverse occurs
phCtiC phenomena= reflected. ~S second dynamical model for the kmber data (Figure 17-46). Model data for quiet
~d the emlier version [RWS et al., 1980] have been com- geomagnetic conditions (labeled “CHIU) and moderately
pared to the same DE-2 Ott 81 and Dec 81 data shown in disturbed conditions (labeled “SHEFFIELD) are shown re-
Figure 17-41. Only a brief description of the extremely spectively above and below the experimental data for each
comprehensive analyses given by Rees et al. [1983] is pre- orbit. The “SHEFFIELD results are from the new dynami-
sented below. cal model using variable electron densities. The following

17-48
 ,I,,j,, -...~,.k \ IHIII 0( ‘. lb

P
1 111///1,1111. / / / /K/f CiOFlb:I:1181

// / II*’:,,,,.
-~~ \ =FFIELD MT. 16

\ CHILI al. 16
Ifljf ,,,/-- -=-..~,~,~~

/// /KdC/~:1: 1182

,i_...\\\\\~\\i\* \ WFFIELC !lCI. 16

/[

21 LST 09 LST

\ k-~ /// ////] /ll\\\\\\k . . ..w....\ CHIU OCT. 16

. jlll~llll.1111 I I I IKVEC
OR5.T,
la

\ ~— // /lo\\\\\>\\\ \.\\l,, I WFFIELO IXT. 16

\\ . . . Ctil U OCT. :6

KWC ~.~, I ?<,2

\ Wfrlf:; Orl. lf.

I I I I I
-so -70 -90 -70 -50

Ldtuda s-POLE

Figure 1745. Comparison hctwcen esch individmd southern (Ckt 1981) polar pass (marked K= ofilt No.) and the comspotilng polsr slices from
two thm~ticsl global (threedlmcnsional time-dependent) model simulations at 320 km. Chi~let geomagnetic conditions. Sheffiel&
Modcratelydlstuti geomagnetic conditions. Scale lW latitude -667 ms [Rees et al., 1983]. (Reprinted with permission from Pergamnn
~SS Ltd. @ 1983.)
 \lll,. ,i,, -\\\\\\
\
CHIU MC. ?1

a KWC ~11: IB1O

v-

. I 1 \\ ~\ WFFIELO OEC. 21
\-

Ctslu OEc. ?1

KVEC OMIT: 181;

11”

P
\\\. \\\, \ \\ -FFIELD KC. 21
0/
r /

06 LST 10 LST

. Cnlu LIEc. 21

KVEC ORBIT: :815

\ b~~- / /[/ ///// \ \ \\ wrFltl.o OFr. ,,1

Ill” “

I 1 I I I
+50 +70 +90 +70 +50
htitsssia N.FOLE
tiee81

Figure 17-46. Comparison between each individual nostbem (W 1981) fsolar pass (marked KVEC osbb No.) and tfse corresponding polar slices fmm
two tbeomticd global (tfuecdlmensiomd time-dependent) model simulations at 320 h. ~l~iet geomagnetic Codltions. Sbeffiel&
Moderatelydistuti geomagnetic conditions. Scsde: I@ latitude = 667 ms [Rms et af., 1983]. (Reprinted with prmission fmm Pergamon
hSS Ltd. @ 1983.)
 UCL 3-DIMENSIONAL T 1ME-DEPENDENT NEUTW HI NO HOCIEL
SEPWTE GEOGRRPH 1C RND GEOMRGNE T I C POLES
SOUTH GEOttRGNET ] c PDLE : 7U CIEG LFIT. ; 126 OEG LONG.

SHEFF lELD ELECTRON DENS111 f40DEL

VOLLF?NO V2 ELECTRIC FIELD MODEL
OCTOBER 16 SIttULflTION
ftLT 1TUDE (KM] : 320
LOCRL TIME SECTOR : 20. u tiR -- Et. U HR

--- -
. --- -
. ---
. ---
. . . .
.,, .
.,. >
.,, >
.,, \
- ,,, ,
.,, <
. ‘ d.<
. ,.-<
. --- -e
. , ‘—
+<
. .- *-

.
/
,
4

1
\-
?
1
\
1
1
—
d
/ r r
\ T /.
$ P ,.*
,’ .
‘9
\ r ? ,.
. 1 ,
.
\ \ ,
d , \ . .
, ., . .
, ,, . .
,,

1,
. . , .
. ,/ . . . ,
. ., . . . .
. /, . . . .
. ., . . . .
. . . . . . . .
, ., ,, \ , . . .
. . . ,,, b-<k-~-, . . . .
. ., ,,, \ \\ --- . . .
. . . . . . ,,% ~-. . . .
. ., .,, ,,, \. . .
. .’. .,, , ,,, . . .
. . .,, ,,, ,,. - .
. . . ... , ,,, , . . . . .
-
,, r

Figure 17-47 University College London three-dimensional, time-dependent neutral wind model. Separate geographic and geomagnetic ~les. South
geomagnetic pole: 74° Lat., 126° Long. Sheffield Electron Density Model plus Volland V2 Electric Field Model. 16 October simulation.
Altitude 320 km; local time sector 2040-0840 h. This presentation allows the UT geomagnetic dependence of vector winds from DE-2 to
be compared orbit by orbit. Note this is not a flow pattern. Each slice is taken from the model at the same local time at 1.2 (U.T. ) intervals
as the earth rotates under the plane of the satelite orbit [Rees et al., 1983]. (Reprinted with permission from Pergamon Press Ltd. 0 1983. )

17-51
CHAPTER 17

UCL 3-DIWNSIW TIW-OEPENOENT WUT~ HIM) ~L

3E~TE GE~HIC RNO GEmRGNET [C P~ES
NORTH GE~WTIC POLE : 80 KG LRT. ; 272 OEG LONG.

.,,
,..
,.,
,
WFF1 ELO ELECTRON OENSI TY MOOEL

,,,
VOLL~ V2 ELECTRIC FIELO MOOEL

.,,
.
OECEtiR 21 SIMULRTION

,..
.,.
,
RLTITuOE (KM) : 320
L~W TIME SECTOR : 18.0 HR -- 6.0 HR

,,,
.
*,.
/“ \,~&

x.ti+
*,.
\,. &
L

.
w
.
.
.
.

#
P

\,,
,,. +
. . . .
..Z ,

. .#, .

. ..#> .
. . ..z

. .>, - .
. ,., > .

L
!,, . .,, > .
,,, . ,,, , * . . .
,,, , /), , . . . .
,,, / ,,, , , -.,
,., , .,, , . -..
,,, , P/, , . . . .
/,, , P,, , / .,,
/,, . , /?. , , . . .
—
I I I I r I I I “+— l-t-l---T
0 6 12 18 24

._) = 1~ M/SEC

Figure 17-48. University College London three-dimensional, time-dependent neutral wind model. Separate geographic and geomagnetic poles. North
geomagnetic pole: 80° Lat., 272° Long. Sheffield Electron Density Model plus Vollsmd V2 Electric Field Model. 21 December simulation
allows the UT geomagnetic dependence of vector winds from DE-2 to be compared orbit by orbit. Note this is not a flow pattern. Each
slice is takm from the model at the same local time at 1.2 (UT) intervals as the earth rotates under the plane of the satellite orbit [Rees et
al., 1983]. (Reprinted with permission from Pergamon Press Ltd. @ 1983. )

17-52
 UCL 3-D IMENSICINOL TIME -OEPENOENT NEUTRRL HINO MODEL
SEPFIRRTE mOGRaPHIC RNO GEOMRGNET IC ?OLES
NORTH GEOMRGNE TlC ?OLE : 80 OEG LRT. ; 272 DEG LONG.

SHEFFIELO ELECTRON OENSITT MODEL

LEF2 LRRGL ELECTRIC F IELD MOOEL

~~TITu~E ,fiflj ; 320

UN IVERS9L TIME 22.8

90

+
.

. -
. .+>
. .-j >
. ,AA

. !,.

., ,

. ,,,

. ,,,

. !,,

. ,.z

. . .

. ,.

. ,.

. . .

. . .

.
. ,,.

.
.

0 ,,1 II, ,, ( II I ,,, r

o 6 12 18 24

. —+ = 1000 M/SEC

Figure 17-49. Storm-time simulation, University College London three-dimensional time-dependent neutral wind model. Separage geogr~hic and geo-
magnetic poles. North geomagnetic pole: 80° Lat.; 272° bng. Sheffield Electron Density Model; LEF2 Large Electric Field Model;
Expanded aurora] oval and cross-cap potential, Altitude 320 km; U.T. 22.80. Note: This diagram shows the actual thermospheric flow
pattern at 22.80 U.T. In December 1981, the DE-2 satellite was (approx.) in the 0600-1800 Local Solar Time plane and thus crossing the
region of peak westward winds (-50”-60° geographical latitude at 22.80 U .T., over North America) [Rees et al,, 1983]. (Reprinted with
permission from Pergamon Press Ltd. @ 1983.)

17-53
CHAPTER 17
wind features were identified for each polar pass: (a) a region related to locations of auroral oval’ boundaries. There is a
of strong antisunward flow over the geomagnetic polar cap modulation of about 45° in the geographic location of these
with considerable variability within an orbit and from orbit boundaries as a function of Universal Time. Results of a
to orbit, (b) sunward winds (or strongly reduced antisunward simulated polar electric field enhanced to 130 kV, corre-
winds) in the dawn and dusk auroral ovals, (c) antisunward sponding to a Kp value between -5 and 7, are given in
winds below the latitude of the auroral oval, as expected Figure 17-49. The predicted winds have been compared to
from a solar-driven source, and (d) highly variable high those obtained during the disturbed Dec 81 period. At au-
latitude winds with only a general correlation of velocity roral oval latitudes the predicted winds exceed the measured
with geomagnetic activity. data, but in the polar cap the maximum DE-2 winds (orbit
The wind velocities in the different flow regimes and 1813) are greater than the model values. The figure illus-
their dependence on Universal Time and geomagnetic events trates that during very disturbed periods the wind patterns
were studied further. The relation between ion drifts and and magnitudes accommodate to the expansion of the au-
neutral velocities showed that wind velocities are 3090-70 roral oval and the increasing ion convection. Future studies
of ion velocities in the aftemoodevening auroml oval, greater are planned to examine these affects and to attempt to dis-
than ion velocities over the polar cap, and not well correlated tinguish between effects of high latitude momentum and
with ion velocities in the dawn/rooming auroral oval. More heating (Joule and particle heating) processes.
detailed studies of the relationship between ion convection
and neutral winds are planned using simultaneous electric
field and ion drift data from DE-2. The UT variations in ACKNOWLEDGEMENTS
neutral winds as calculated from the moderate activity model
for the local time planes of the Ott 81 and Dec 81 data are The critical comments of R. G. Roble, National Center
given in Figures 17-47 and 17-48 respectively. These plots for Atmospheric Research, Capt. D. V. Ridge, AFGL, and
are necessary to explain many features of the wind data J. M. Forbes, Boston University, are greatly appreciated.

17-54
 WINDS

REFERENCES-CHA~ER 17 Fuller-Rowell, T. J., and D. Rees, “A Three-Dimensional

Time-Dependent Simulation of the Global Dynamical
Abbey, R. F., Jr., “Risk Probabilities Associated with Tor- Response of the Thermosphere to a Geomagnetic Sub-
nado Windspeeds,” Proc. Symp. on Tornadoes, Institute storm,” J. Atmos. Terr. Phvs., 43:701, 1981.
for Disaster Research, Texas Tech. Univ., Lubbock, Gage, K. S., and B. B. Balsley, “Doppler Radar Probing of
Texas, 1976. the Clear Atmosphere,” Bull. Amer. Meteor. Sot., 59:
Amayenc, P., ‘Tidal Oscillations of the Meridiofial Neutral 1074-1092, 1978.
Wind at Midlatitudes,” Radio Sci., 9:28 1–294, 1974. Golden, J. H., “An Assessment of Windspeeds in Torn-
Amo]dj A., and R. Bellucci, “Variability of Ballistic Me- adoes,” Proc. Symp. on Tornadoes, Institute for Disaster
teorological Parameters,” Tech Memo. No. M-1913, Ft. Research, Texas Tech. Univ., Lubbock, Texas, 5-42,
Monmouth, U.S. Ai-my Sig. Eng. hb, 1957. 1976.
Balsley, B .B., and K.S. Gage, “The MST Radar Technique: Green, J. L., K.S. Gage, and T.E. Van Zandt, “Atmospheric
Potential for Middle Atmospheric Studies,” Pure APP1. Measurements by VHF Pulsed Doppler Radar,” IEEE
Geophys., 118:452493, 1980. Trans. Geosci. Elec., GE-17: 262–280, 1979.
Bean, S. J., and P.N. Somerville, “Some Models for Wind- Gregg, W. R., “An Serological Survey of the United States,”
speed,” 4, AFGL-TR-79-01 80, AD A077048, 1979. Monthly Weather Review, Supplement No 20:1922.
Brekke, A. J., J.R. Doupnik, and P.M. Banks, “Observa- Hays, P. B., and R.G. Roble, “Direct Observations of Ther-
tions of Neutral Winds in the Auroral-E Region During mospheric Winds during Geomagnetic Storms ,“ J. Geo-
the Magnetosphetic Storm of Aug 3–9, 1972,” J. Geo- phys. Res., 76:53165321, 1971.
phys. Res., 79: 2248–2256, 1974. Hays, P. B., J. W. Meriwether, and R.G. Roble, “Nighttime
Ching, B. K., and Y.T. ~lu, “A Phenomenological Model of Thermospheric Winds at High Latitudes,” J. Geophys.
Global Ionospheric Electron Density in the E, F1, add F2 Res., 84: 1W5–1913, 1979.
Regions,”J. Atmos. Terr. Phys., 35:1616-1630,1973. Hays, P. B., T.L. Killeen, and B.C. Kennedy, “The Fabry-
Crutcher, H. L., “Upper Wind Statistics Charts of the North- Perot Interferometer on Dynamics Explorer,” Space Sci.
ern Hemisphere,” NAVAIR 50-l C-535, 1959. Instr., 5:395416, 1981.
DeVries, L. L., “Analysis and Interpretation of Density Data Heppner, J. F., “Empirical Models of High-Latitude Electric
from the Low-G Calibration System (hgacs),” Space Fields,” J. Geophys. Res., 82: 1115, 1977.
Res., 12: 777–789, 1972. Heppner, J. F., and M.L. Miller, ‘~ermosphenc Winds at
Dickinson, R. E., E.C. Ridley, and R.G. Roble, “Meri- High Latitudes from Chemical Release Observations,”
dional Circulation in the Thermosphere. I, Equinox Con- J. Geophys. Res., 87: 1633–1647, 1982.
ditions,” J. Atmos. Sci., 2: 1737–1754, 1975. Hemandez, G., and R.G. Roble, “Direct Measurements of
Dickinson, R. E., E.C. Ridley, and R.G. Roble, “Meri- Nighttime Thermospheric Winds and Temperatures. 1.
dional Circulation in the Thermosphere. 11, Solstice Seasonal Variations during Geomagnetic Quiet Pe-
Conditions,” J. Atmos. Sci., 34: 178–192, 1977. riods,” J. Geophys. Res., 81: 2W5–2074, 1976a.
Dickinson, R. E., E.C. Ridley, and R.G. Roble, “A Three- Izumi, Y., “Kansas 1%8 Field Program,” AFCRL-72-0041,
Dimensional General Circulation Model of the Ther- AD739165, 1971.
mosphere,” J. Geophys. Res., M 1499–15 12, 1981. Kantor, A. J., and A.E. Cole, “Wind Distributions and In-
Ellsaesser, H. W., “Wind Variability; AWS Tech. Rept. No. terlevel Correlations, Surface to 60 Km ,“ AFGL-TR-
105-2, HQ Air Weather Service, Scott AFB, 111., 1960. 80-0242, ADA092670, 1980.
Evans, J .V., “Ionospheric Movements Measured by inco- Kaufman, J. W. (cd.), “Terrestrial Environment (Climatic)
herent Scatte~ A Review,” J. Atmos. Terr. Phys., * Criteria Guidelines for Use in Aerospace Vehicle De-
175, 1972. sign,” NASA Tech. Memor. 78118, MSFC, 1977.
Evans, J. V., “A Review of F Region Dynamics,” Res. Kelley, M. C., T.S. Jorgensen, and 1.S. Mikkelsen, “Ther-
Geophys. Space Phys., 13:887, 1975. mospheric Wind Measurements in the Polar Region,”
Evans, J. V., “Incoherent Scatter Contributions to Studies J. Atmos. Terr. Phys., 39:21 1–219, 1977.
of the Dynamics of the Lower Thermosphere,” Rev. Kelly, D. L., J.T. Schaeffer, R.P. McNulty, C.A. Doswell
Geophys. Space Phys., 16:195, 1978. III, and R.F. Abbey Jr., “An Augmented Tornado Cli-
Fichtl, G. H., “Small-Scale Wind Shear Definition for Aero- matology,” Mon. Wea. Rev., 106: 1 I72–1 183, 1978.
space Vehicle Design,”J. SpacecrajiRockets, 979, 1972. Killeen, T. L., P.B. Hays, N.W. Spencer, and L.E. Whar-
Forbes, J. M., “Atmospheric Tides 1. Model Descriptions ton, “Neutral Winds in the Polar Thermosphere as Meas-
and Results for the Solar Diurnal Components,” J. Geo- ured from Dynamics Explorer,” Geophys. Res. Letf., %
phys. Res., 87: 5222–5240, 1982a. 957–960, 1982.
Forbes, J. M., “Atmospheric Tides 2. The Solar and Lunar Knutson, J. R., D.C. Kayser, and W.E. Potter, “Mass Spec-
Semidlumal Components,” J. Geophys. Res., m: trometric Measurement of Tbermospheric Wind,” J.
5241–5252, 1982b. Geophys. Res., 82:5253, 1977.
Forbes, J. M., and H.B. Garrett, “Theoretical Studies of Lettau, H., Atmospharische Turbulenz, Akademische Ver-
Atmospheric Ti~es,” Rev. Geophys. Space Phys., 17: Iagsgesellschaft M.B. H., Leipzig, 114, 1939.
1951–1981, 1979. Lettau, H., “A Re-examination of the ‘Leipsig Wind Pro-
Fuller-Rowell, T. J., and D. Rees, “A Three-Dimensional file’ Considering Some Relations Between Wind and
Time-Dependent Global Model of the Thermosphere,” Turbulence in the Frictional Layer,” Telfus, 2: 125,
J. Atmos. Sci., 37: 2545–2567, 1980. 1950.

17-55
CHAPTER 17
Lettau, H. H., and B. Davidson (eds. ), E.rp/oring ~heAtmo- Latitude Circulation and Temperature Structure of the
sphere’s First Mile. Pergamon Press. New York, 1957. Thermosphere Near Solstice,” Planet. Space Sci., 31:
Marcos, F. A.. and E. R. Swift, “Application of the Satellite 1479–1499, 1983.
Triaxial Accelerometer to Atmospheric Density and Wind Rottger, J., “VHF Radar Observations of a Frontal Pas-
Studies,” AFGL-TR-82-O091. ADA 120852, 1982. sage ,“ J. Appl. Meteorol., 18: 85–9 ~, 1979.
Mayr, H. G., I. Harris, and N. W. Spencer, “Some Prop- Scoggins, J. R., and W. W. Vaughan, “Cape Canaveral Wind
erties of Upper Atmosphere Dynamics ,“ Re}’. Geophys. and Shear Data ( 1 through 80 km) for Use in Vehicle
Space Phys., 16: 539–566, 1978. Design and Performance Studies,” Tech. Note D-1274,
Mcban, G. S.. “The Jet Stream and Associated Turbu- NASA, 1962.
lence,”’ in Wind.f for Aerospace Vehicle Design, Sissen- Shellard, H. C., ‘~ Estimation of Blgn Wind Speeds, Wln
wine and Kasten, editors, AFCRL-62-273(I), 1%2. Effects on Buildings and Structures,” Nat’1 Phys. Lab.
Meriwether. J. W., C.A. Tepley, F.A. Price, P.B. Hays, Symp., Teddington, England, 16: 3&51, 1965.
and L.L. Cogger, “Remote Ground-based Observations Sherlock, R. H., “Variation of Wind Velocity and Gusts with
of Terrestrial Airglow Emissions and Thermospheric Height,” Paper No. 2553, Proc. Am. Sot. Civil Eng.,
Dynamics at Calgary. Alberta,” Opt. Engr., 1982. 1: 463–508, 1952.
NOCD (Naval Oceanography Command, Detachment), Gui& Sherlock, R. H., “Gust Factors for the Design of Buildings,”
to Standard Weather Summaries and Climatic Services, Int. Assoc. for Bridge and Structural Engineering, 8:
NAVAIR 50- IC-534. Asheville, N. C., 1980. 207–235, 1947.
Quegan, S., G.J. Bailey, R.J. Moffett, R.A. Hwlis, T.J. Sipler, D. P., and M.A. Biondi, “Midlatitude F-Region Ne-
Fuller-Rowell, D. Rees, and R. W. Spiro, “A Theoret- tural Winds and Temperatures during the Geomagnetic
ical Study of the Distribution of Ionization in the High Storm of March 26, 1976,” J. Geophys. Res., U 37W,
Latitude Ionosphere and the Plasmasphere: First Results 1979.
on the Mid latitude Trough and the Light-Ion Trough,” Sipler, D. P., B.B. Lukkala, and M.A. Biondi, “Fabry-Perot
J. Atmos. Terr. Phys., 44:619, 1982. Determinations of Mid-Lathude F-Region Neutral Winds
Rees, D., T.J. Fuller-Rowell, and R. W. Smith, “Measure- and Temperatures from 1975 to 1979,” Planet. Space
ments of High-Latitude Thermospheric Winds by Rocket Sci., * 1025–1032, 1982.
and Ground-Based Techniques and Their Interpretation Sissenwine, N., P. Tattelman, D.D. Grantharn, and 1.1.
using a Three-Dimensional Time-Dependent Dynamical Gringorten, “Extreme Wind S-s, Gustiness, and
Model,” Plane(. Space Sci., =: 919, 1980. Variations with Height for MIL-STD-210B,” AFCRL-
Rees, D., T.J. Fuller-Rowell, R. Gordon, T.L. Killeen, TR-73-0560, AD774044, 1973.
P.B. Hays, L. Wharton, and N. Spencer, “A Compar- Spencer, N. W., R.F. Theis, L.E. Wharton, and G.R. Car-
ison of Wind Observations of the Upper Thermosphere ignan, “kal Vertical Motions and $inetic Temperature
from the Dynamics Explorer Satellite with the Predic- from AE-C as Evidence for Aurora-Inducecl Gravity
tions of a global Time-Dependent Model,” Planet. S~ce Waves,” Geophys. Res. Lett., 3:313, 1976.
Sci., 31: 1299–1314, 1983. Spencer, N. W., L.E. Wharton, H.B. Niemann, A.E. Hdln,
Reiff, P. H., R. W. Spiro, and T. W. Hill, “Dependence of G.R. Carignan, and J.C. Maurer, ‘~ Dynamics Ex-
Polar Cap Potential Drop on Interplanetary Parameters,” plorer Wind and Temperature Spectrometer,” Space Sci.
J. Geoph~s. Res. 86: 7639–7648, 1981. lnstr., 5:417428, 1981.
Rieger, E., “Neutral Air Motions Deduced from Barium Spencer, N. W., L.E. Wharton, G.R. Carignan, and J.C.
Release Experiments. 1. Vertical Winds,” J. Atmos. Maurer, “Thermosphere Zonal Winds, Vertical Motions
Terr. Ph>~s., 36: 1377, 1974. and Temperature as Measured from Dynamics Ex-
Richmond, A. D., and S. Matsushita, “Thermospheric Re- plorer,” Geophys. Res. Lett., 9 953–956, 1982.
sponse to a Magnetic Substorm,” J. Geophys. Res., 80: Straus, J. M., “Dynamics of the Thermosphere at High Lati-
2839–2850, 1975. tudes,” Rev. Geophys. Space Phys., 16: 183–194, 1978.
Rishbeth, H. “Thermospheric Winds and the F-Region: A Tattelman, P., “Surface Gustiness and Wind S~d Range
Review.” J. Atmos. Terr. Phys., 34: 148, 1972. as a Function of Time Interval and Mean Wind Speed,”
Roble, R. G.. “The Thermosphere,” in Ph~sics ofthe Upper J. Appl. Meteorol., 14: 1271–1276, 1975.
Atmosphere and Magnetosphere, National Academy of Volland, H., “Models of Global Electric Fields Within the
Sciences, Washington. D. C., 68: 1977. Magnetosphere,” Ann. Geophys., 31:154-173, 1975.
Roble, R. G., “Dynamics of the Earth’s Thermosphere,” Volland, H., “Semiempirical Models of Magnetospheric
Rev. Geophys. Space Phys., 21:2 17–233, 1983. Electric Fields,” in Quantitative Modeling of Magne-
Roble; R. G.. R.E. Dickinson, and E.C. Ridley, “Seasonal tospheric Processes, Geophys. Monogr. Ser., VO1. 21,
and Solar Cycle Variations of the Zonal Mean Circu- edited by W .P. Olson, 26 I -280, American Geophysical
lation in the Thermosphere,” J. Geophys. Res., Union, Washington, D. C., 1979a.
82:5493–5504. 1977. Volland, H., “Magnetospheric Electric Fields and Currents
Roble, R. G., R.E. Dickinson. and E.C. Ridley, “Global and Their influence on Large Scale Thermospheric Cir-
Circulation and Temperature Structure of Thermosphere culation and Composition,” J. Atmos. Terr. Phys., 41:
and High-Latitude Plasma Convection ,“ J. Geophys. 853–866. 1979b.
Res., 87: 1599–1614. 1982. Wu, S. T., S. Matsushita, and L.L. DeVries, “An Analysis
Roble. R. G.. R.E. Dickinson. E.C. Ridley. B.A. Emery, of the Upper Atmospheric Wind Observed by Logacs ,“
P.B. Hays. T.L. Killeen. and N.W. Spencer, “The High Planet. Space Sci., 22: 1036, 1974.

17-56
 Chapter 18

OPTICAL AND INFRARED PROPERTIES

OF THE ATMOSPHERE
R. W. Fenn J. D. Mill Lt. COL USAF
S. A. clou@ L. S. Rothman
W. O. Gallery E. P. Shettle
R. E. Good F. E. Voiz
F. X. Kneizys
*

This chapter deals with the absorption, scattering, emis- 1920s. Early Standard Atmosphere Models represented mean
sion and refractive properties of the natural atmosphere from conditions of temperature, density, and pressure as a func-
the ultraviolet through the ticrowave region. Nonlinear tion of altitude. During the 1950s and 1960s extensions and
propagation processes (such as thermal blooming) are not supplemental atmospheres were derived that describe in ad-
covered. dition the seasonal and latitudinal viability of atmospheric
The natural atmosphere includes the range of those av- structure (see Chapters 14 and 15).
erage conditions which are not directly affected by limited For the purpose of modeling the optical properties of
local processes or sources, such as gases or particulate from the atmosphere, specifically as described in Sections 18.4. I
a fire or industrial plant, a dust cloud from vehicular traffic (FASCODE) and 18.4.2 (LOWTRAN), the 1962 U.S. Stan-
etc. The natural atmosphere does include the effects of mol- dard Atmosphere and supplemental model atmospheres were
ecules, aerosol (haze) particles, clouds, fogs, rain, and snow. used as models. Although updated Standard Atmo-
The vertical extent of the atmosphere is from the surface spheres have heen published as recently as 1976, differences
up to 100 km altitude. A brief discussion is included on the are limited to altitudes above 50 km where the adopted
earth’s surface reflectance properties since they also affect radiance models were not well characterized.
the atmospheric optical properties. Pressure, temperature, water vapor density, and ozone
The chapter includes descriptions of the basic physical density profiles as a function of altitude are provided to
relationships of optical propagation as well as the current describe these molecular atmospheric models. These pro-
state of experimental knowledge and modeling of the at- files, taken from McClatchey et al. [ 1972], correspond to
mospheric propagation medium. An extensive annotated the U.S. Standard Atmosphere, 1962 and five supplemen-
bibliography covering much of the material of this chapter tary models [Cole et al., 1965 or U.S. Standard Atmosphere
was compiled by Wiscombe [ 1983] in Reviews of Geo- Supplements 1%6]: Tropical ( I S“N), Midlatitude Summer
physics and Space Physics. (45”N, July), Midlatitude Winter (45”N, January), Subarctic
Summer (60”N, July), and Subarctic Winter (60°N, Janu-
ary). These profiles are identical to the ones used in the
18.1 ATMOSPHERIC GASES LOWTRAN code [Kneizys et al., 1980]. The water vapor
and ozone latitude profiles added to the 1962 U.S. Standard
Atmosphere by McClatchey et al. [ 1972] were obtained
18.1.1 Atmospheric Molecules, Models from Sissenwine et al. [ 1968] and Hering and Borden [ 1964]
of the Atmospheric Composition respectively, and correspond to mean annual values. The
water vapor densities for the 1962 U.S. Standard Atmo-
me propagation of optical and infrared radiation through sphere correspond to relative humidities of approximately
the atmosphere depends on the composition and variability 50% for altitudes up to 10 km, whereas the relative humidity
of the atmosphere. Systematic variations in the density, values for the other supplementary models tend to decrease
pressure, temperature, water vapor, and ozone as a function with altitude from approximately 80Yc at sea level to ap-
of latitude and season have been known for many years. proximately 30% at 10 km altitude. Above 12 km, the water
The development of model atmospheres goes back to the vapor density profi Ies depicted here have been replaced by

18-1
CHAPTER 18
100 ~ 100

90- - — TROP 9 — TRDP

. . ..Ms .- .-MS
---- Mw ---- Mw
80- - --- Ss 6 --- Ss
––SW —– Sw
—- US STD —- US STD

70- - 7

z
x 60- - :6
,h \

u u
0 SF - n 5
3
z +

+ 40- - e 4
J -1
a a

30- - 3

\
2@ - 20

lo- - 1

1
, ‘

! 60 140 260 2io 240 240 240 2$0 260 2?0 2A0 240 3 1 +0-0 io-” to-’ t’o-s t’o--’?o’ (0-’
TERPERRTURE [K] H20 DENSITY [GM M-31
(a)
Figure 18-1. Temperatu~ vs altitude for the six model atmospheres tNSp
ical (TROP), midlatitude summer (MS), midlatitude winter —
(MW), subarctic summer (SS), subarctic winter (SW), and
US standard (US STD).
— TROP
. . ..Ms
---- Mw
more recent measurements [Remsberg et al., 1984] and are --- Ss
––SW
approximately 5 parts per million by volum~ppmv.) For all —- US STD

model atmospheres, the gases COZ, NZO, CO, CH4, Nz,

and 02 are considered uniformly mixed, with mixing ratios
of 330, 0.28, 0.075, 1.6, 7.905 x l@, and 2.095 x 1~
ppmv, respectively.
The temperature profiles for the six model atmospheres
as a function of altkude are shown in FIgttre 18-1. The
pressure profiles are given in Figure 18-2. Figures 18-3 a
and b show the water vapor density versus altitude from O \
to 100.km, and an expanded profile from Oto 30 km. Figures

100

9 —
..-.
TROP
MS
..#o.,.
---- mw
--- Ss H20 OENSITY ( Gti H-31

I ‘% ––SW
—- US STD (b)
7
Figure 18-3. (a) Water vapor density profiles vs attitude for the six model
atmospheres.
6
(b) Water vapor density profiles vs altitude for the six model
atmospheres with the ~gion from O to 30 tmr expanded.
5

4
18-4 a and b and Figures 18-5 a and b show similar profiles
3 for ozone and relative air density.
Measurements made from balloon flights [Murcray et
2
al., 1968] have shown the existence of nitric acid in the
earth’s atmosphere. Although nitric acid (HNOS) is of only
1
minor importance in atmospheric transmittance calculations,
/
it has been shown to be a significant source of stratospheric
emission, particularly in the atmospheric window region
PRESSURE (MB 1
from 10 to 12 ~m. The concentration of atmospheric nitric
Figure 18-2. Pressure vs altitode for the six model atmospheres. acid varies with altitude and also appears to depend on

18-2
 w

OPTICAL AND INFRARED PROPERTIES OF THE ATMOSPHERE

100 ure 18-3a and b are now known to be too high. The model
atmospheres can still be considered representative of their
90- -
respective conditions up to about 50 km for temperature,
80- -
30 km for ozone densities, and the tropopause (approxi-
mately from 8 km in the article to 15 km in the tropics) for
70- - water vapor. These models are still used for cases dominated
— TROP
. . ..M$
E ---- MW
x 60- -
--- Ss
–– Sw
IIJ —- US STD
0 50- -
3 100
t-
+ 40- -
-! so-- — TROP
m . . ..MS
---- MW
30- - --- Ss
80- -
–– Sw
—- US STD
2F - 7@ -

r
10-- x 60- -

u
0 n 50- -
I’D-{’
z
OZONE DENSITY [Gtl M-3)
1- 4& -
(a) A
u

30- -

— TROP 20- -
. . ..MS
---- Mw
--- Ss 10- -
–– Sw
—- USSTD

C4 ,
1 0-+ ire-5 “ ~o-i i’~-i “-
20- -

~ UNIFORMLY MIXED GASES

u
(a)
0
2
t-
J
— TRW
= lo- - . . ..Ms
--- MW
*
--- Ss
–– Sw
\ —- US STO

\.

I
0 1
10+ 1

OZONE DENSITY (GM M-3)

(b)
\

Figure 18-4. (a) Ozone density profiles vs altitude for the six model
atmospheres.
(b) Ozone density profiles vs altitude for the six model
atmospheres with the region from 0t030km expanded.

latitude and season. Figure 18-6 shows the volume mixing

ratio profile of atmospheric nitric acid as a function of al- Iu

titude from the measurements of Evanset al. [1975]. This UNIFORMLY MIXED GASES
profile has been chosen to represent a mean nitric acid profile (b)
for the six model atmospheres.
Figure 18-5. (a)Profileof (P/Po) (Tfl), the relative air density, vs altitude
Since all of the above profiles were constructed, knowl- for the six model atmospheres. The density of the uniformly
edge of the state of the atmosphere has increased. This is mixed gases is proportional to this quantity. P“ = 1013 mb
particularly true regarding the stratosphere and the concen- and To = 273 K.
(b) Profile of (P/Po) (Tfl), the relative air density, vs altitude
tration of minor constituents. For example, stratospheric for the six model atmospheres with the regions O to 30 km
water vapor concentrations for the six profiles given in Fig- expanded.

18-3
CHAPTER 18
I 1 18.1.2 Molecular Absorption and
Spectroscopic Parameters

18.1.2.1 The Role of Molecular Absorption. Molecular

.
absorption, that is, attenuation of electromagnetic radiation
by the mechanism of molecules absorbing quanta of energy
to alter vibrational and rotational states, is of prime signif-
icance in propagation in the atmosphere. Although the ter-
restrial atmosphere consists primarily of nitrogen, oxygen,
and argon, absorption in the infrared region is dominated
by species with very low concentrations but very active
vibration-rotation bands such as water vapor, carbon diox-
ide, ozone, and nitrous oxide. Other active species also play
an important role depending on the region of the electro-
magnetic spectrum being investigated or the layer of the
atmosphere being probed.
In order to calculate transmittance due to a given spectral
line in the atmosphere, it is useful to describe the absorption
. .
,0-m ,0-” ~-.
coefficient as a function of frequency for each line. Assum-
VOLE WXIW RATIO ing superposition of the contribution from individual spec-
tral transitions, the absorption coefficient k(v) as a function
Figure 18-6. Volume mixing ratio profile for nitric acid vs altitude from
the measurements of Evans et al. [19751.~Is single profile of wavenumber v (cm-1) can be considered in general to be
is used with all of the six model atmos-pheres. - “ a product of the radiation field, the molecular system, and
the coupling between them. This can be expressed [Clough
et al., 1981] as
by conditions in the troposphere. For cases dominated by
stratospheric conditions or where the distribution of minor
constituents is significant, modelers should supply their own ~(V) = V tanh(hcv/2kT)
profiles. X ~ p(mi) si(T) [f(v, vi) + f(v, – ~i)], ( 18.1)
There are several recent sources for profiles of temper-
ature and minor constituent density. The U.S. Standard
Atmosphere 1976 updates the 1962 Standard for temperature
above 50 km and provides revised estimates for the surface where p(mi)(molec/cm3) is the density for the molecular
concentrations of what was termed previously the “uni- species mi with transition wavenumber vi responsible for
formly mixed gases”. The new values for the volume mixing the i’th transition. Si(T) (molec/cm2)-1 is the intensity of
ratios of C02, N20, CO, and CH4 are 330, 0.27, 0.19 and the transition at temperature T(K) appropriate to the line
1.50 ppmv. Actually the concentrations of these gases do shape f(v,vi) (l/cm-l). For most molecular species, the line
show significant variations from these values, particularly intensity is a function of the expectation value of an ex-
with altitude in the stratosphere. Cole and Kantor [1978] pansion of the electric dipole moment. For some species,
provide sets of monthly mean temperature profiles up to 90 for example nitrogen and oxygen [Rothman and Goldman,
km at 15° intervals between the equator and the pole. Along 198 1], the much weaker electric quadruple and magnetic
with statistics on the variability of these profiles, they also dipole moments are the means by which the quantum states
provide models that portray longitudinal variations in monthly change. In terms of intensity definition used in the AFGL
mean values of temperature during winter months and the line compilations, Si(T), discussed in the next section, we
ve ical variation that occurs during stratospheric warming have the relationship
an cooling events in the winter arctic and subarctic. Hough-
!
toq [1977] provides seasonal profile temperatures at 10”N, Si(T) = vi tanh (hcv/2kT) S(T)
40@N and 70”N up to 105 km plus the original references
= v, 1 – e-hcq,2kT -
for the data. I ~ + e - hcvi/2kT
Si(T). (18.2)
For profiles of the minor constituents including ozone
and stratospheric water vapor, WMO [19821 provides an
up-to-date and exhaustive source. See also Chapter 21. Much The line shape function f(vjvi) is dependent on molecular
of the profile data from this source plus some more recent species, broadening density and temperature. For collisional
measurements have been compiled as annual averages in 2 broadening in the impact limit, the form factor is given by
km steps in Smith [ 1982]. the Lorentz line shape

18-4
 OPTICAL AND INFRARED PROPERTIES OF THE ATMOSPHERE
O!i vibration-rotation lines of naturally occurring molecules of
f(Vi) = : (18.3)
IT (V – Vi)2 + (@i)2 significance in the terrestrial atmosphere. Since that time
where tii (cm-’) is the collision broadened halfwidth at half these data have been expanded in terms of the frequency
maximum (HWHM), The monochromatic transmittance covered, addition of trace species, and inclusion of weaker
through a layer of homogeneous medium of thickness ( is transitions [Rothman, 1981; Rothman et al., 1981, 1983a,b].
obtained by the Lambert-Beers law Historically, these data have been divided into two atlases,
a Main Compilation comprising transitions of the most ac-
7. = e– ‘(’”e (18.4) tive terrestrial absorbers H20, COZ, 0s, NZO, CO, CH4,
and 02, and a Trace Gas Compilation. The species presently
where the exponent is called the optical thickness. covered by these data bases are summarized in Table 18-1.
The line intensity is temperature dependent through the The isotopes are abbreviated by the code 161 = Hz “O,
Boltzmann factor and the internal partition function 162 = HD]’O, etc. The first seven species (incorporated in
the Main Atlas) include all transitions contributing to at
e – hcE:/kT least 10% absorption over a maximum atmospheric path.
Si(T) = di(l – e-hcti’kT) — (18.5) This has been determined from standard atmospheric models
Q(T)
of the uniformly mixed gases and the two nonuniformly
where Q(T) is the internal partition sum, E; (cm-’) the en-
mixed gases, water vapor and ozone. The standard tem-
ergy of the Iowerstate of the transition, and di is the transition
perature chosen was 296 k. The gases in the Trace Gas
strength. The collisional halfwidth depends on the specific
Compilation were made available for a variety of problems
molecule and can be expressed in terms of density and
including stratospheric probing, pollution monitoring, tem-
temperature as
perature profile retrieval, and laboratory studies. The Main
XT Atlas covers the frequency range Oto20000 cm-’. The low
T
tii (p,T) = ~“
()()
&
PO m
(18.6) frequency or millimeter and submillimeter limit represents
primarily pure rotational transitions while the high fre-
(the exponent XT is 1/2 with the usual classical theory as- quency domain represents for the most part excited vibra-
sumption of temperature-independent collision diameters). tion-rotation transitions of water vapor. The Trace Gas Com-
The ratio of partition sums does not require the knowledge pilation covers the range O to 10000 cm-’. In addition to
of any further parameters aside from the fundamental fre- the four basic parameters, there are included for each tran-
quencies of the vibrational modes. Thus, at least for local sition the unique quantum identification, the molecule and
thermodynamic equilibrium conditions, calculations of mo- isotope, and a reference code.
lecular absorption (transmission and emission) at high res- The goal of generating these parameters has been to
olution basically require the knowledge of four essential create a theoretical y self-consistent set. Ideally, the per-
parameters (1) the resonant frequency of the transition v, formance of quantum mechanical calculations from the de-
(2) the intensity per absorbing molecule Si(TO), (3) the Lor- velopment of mathematical models of the molecular system
entz line width parameter ti:, and (4) the energy of the lower would provide the necessary parameters. This has been sat-
state E;. The question of line shape will be discussed in isfactorily accomplished in many cases, the hydrogen hal-
more detail in Section 18.4.1. The effects of molecular ides being a prime example. In cases involving complex
(Rayleigh) scattering and of aerosol scattering and absorp- resonances or anomalous behavior, theory has lagged behind
tion will be discussed in Sections 18.1.4 and 18.2.1.5. experiment and it has been necessary to utilize available
observed parameters. The latter method does not directly
allow for interpolating the unobserved or weaker transitions
18.1.2.2 The Atmospheric Absorption Line Parameteh that might play a role in long path or high temperature
Compilation. From the preceding discussion it can be observations. Nevertheless, a fair amount of success has
seen that a compilation of spectroscopic data on individual been achieved deriving the tabulated parameters vi, S,, ~i,
molecular transitions would facilitate the so-called line-by- E: and their unique quantum identifications within the frame-
line calculations of spectra. These high resolution calcula- work of general theories of molecular spectroscopy applied
tions, that is, calculations capable of reproducing spectra to experimental observations. Details of the basic theory,
obtained by such instruments as Michelson interferometers, specific data, and treatment of exceptional cases can be
tunable diode lasers, and heterodyne spectrometers have found in references contained in the articles on the com-
been developed along with the advances in computer tech- pilations [Rothman, 1981; Rothman et al, 1981; 1983a;
nology. In the 1960s a program was initiated at the Air 1983b]. As of this writing, the compilations, representing
Force Cambridge Research Laboratories (now AFGL) to the culmination of international efforts in molecular spec-
create such a compilation and the associated codes to pro- troscopy, contain about a third of a million transitions. The
duce synthetic spectra [McClatchey et al., 1973]. The scope maintenance, updating, and improvement of these data is
originally aimed at providing a complete set of data for all ongoing.

18-5
CHAPTER 18
Table 18-1. Molecular s~cies on atmospheric absorption line parameter atlases.

Relative Relative
Natural Natural
Molecule Isotope Abundance Molecule Isotope ‘ Abundance

H,O (1) 161 0.9973 HNO, (12) 146 0.9891

181 0.0020
171 0.0004 OH (13) 61 0.9975
162 0.0003 81 0.0020
62 0.00015
co* (2) 626 0.9842
636 0.0111 HF (14) 19 0.99985
628 0.0040
627 0.0008 HCI (15) 15 0.7576
638 0.00044 17 0.2423
637 0.000009
828 0.000004 HBr (16) 19 0.5068
728 0.000002 11 0.4930

0, (3) 666 0.9928 HI (17) 17 0.99985

668 0.0040
686 0.0020 C1O ( 18) 56 0.7559
76 0.2417
N,O (4) 446 0.9904
456 0.0036 Ocs ( 19) 622 0.937
546 0.0036 624 0.0416
448 0.0020 632 0.0105
447 0.0004 822 0.0019

co (5) 26 0.9865 H2C0 (20) 126 0.9862

36 0.011 136 0.0111
28 0.0020 128 0.0020
27 0.0004
HOCI (21) 165 0.7558
CH., (6) 211 0.9883 167 0.2417
311 0.0111
212 0.00059 N, (22) 44 0.9928

o* (7) 66 0.9952 HCN (23) 124 0.9852

68 0.0040
67 0.0008 CHSC1 (24) 215 0.7490
217 0.2395
NO (8) 46 0.9940
H,02 (25) 1661 0.9949
so, (9) 626 0.9454
646 0.0420 C*H2 (26) 1221 0.9776

NO, ( 10) 646 0.9916 C2H. (27) 1221 0.9770

NH, (11) 4111 0.9960 PHq (28) 1111 0.99955

5111 0.0036

18-6
 H

OPTICAL AND IN~D PROPERTIES OF THE ATMOSPHERE

18.1.3 Index of Refraction Table 18-2. Retilve modulus of dry air at I atmosphereand 288 K.

Wavelength Wayenumber
The theory of molecular optics [Born and Wolf, 1975]
(pm) I(cm,4) N = (m-1) x 106
derives the index of refraction for gases from the scattering
0.23 4.35 x Id 308.0
properties of molecules (Lorentz-hrenz formula). In the
0.25 4.00 x lV 301.5
atmosphere the index of refraction m is very close to 1 so
0.30 3.33 x ld 291.6
that it is convenient to define the refractive modulus N as
0.40 2.50 X Id 282.7
0.50 2.00 x lm 279.0
N=(m–l) xl@. (18.7)
0.75 1.33 x lm 275.4
1.0 l(moo 274.2
N is a functiofi of both wavelength and density and is dif-
2.0 5~ 272.9
ferent for diy air and water vapor. For optical wavelengths
4.0 2500 272.7
greater than 0.23 pm and for infrared wavelengths, N for
10.0 1000 272.6
atmospheric conditions is given by the following formula
from Mien [1966]:

particles of any size. If the size of a particle is very small

N=
[
~+
1 –
a’
(v/b,)*
+ az
1 – (v/b2)2
1P (TO + 15.0)
~ T
compared to the wavelength of the incident light, certain
simplifyirtg conditions exist that also simplify the theoretical
description of the scattering ptiess. ~ls type of scattering

[1
P. of light, which causes the blue sky light, was first described
— co – (v/c,)’ ~ (18.8) by Lord Rayleigh in 1871 [McC~ey, 1976].
0
The total volumetric scattering coefficient k, A for mol-
ecules is defind by
where P is the total air pressure in mb, T is the temperature
in K, PO = 1013.25 mb, TO = 273.15, PWis the p~ial dI, = – 1A- k.(A) “ dx , (18.9)
pressure of water vapor in mb, and v = l@/A is the fre-
quency in cm-’ for the wavelength A in micrometers. with

2
G = 83,42 24m3 m’–l
k~(A) = ~ ~ (18.10)
a, = 185.08 ()
a2 = 4.11

b, = 1.140 X l& where m is the refractive index of the gas (such as air), N‘
the concentration of molecules per unit volume, IA the in-
tensity at wavelength A, and dx @e path element. ~Is
= 43.49 expression is often simplified by Wing advantage of ti
co
refractive index of air being nearly 1, and writing Equation
c, = 1.70 x ld (18.10) as

fie fotmsila is valid from 0.23 pm to the infrared. For

k~(A)= * (m’ – 1)2, (18.1 la)
millimeter and microwaves, the refractive modulus is much
m complicated and includes a strong deence on water
vapor density. For referefice, Table 18-2 lists the index of or
refiction for dry air at 1 atmosphere and 288 K for ---U---:
Vxluus
wavelengths.
k~(A) =~(m– 1)’. (18.1 lb)

18.1.4 Molecular Rayleigh Sca@ering For standard temperature and pressure, using Equation
(18. 1la) introduces an error on the order of 0.04% in the
when air molecules are subjected to an oscillating elec- visible and using Equation (18. 11b) introduces an error of
tric field such as in a light wave, the molecule temporarily about 0.025%. Since air molecules are not completely iso-
absorbs and immediately re-emits this radiation as a point tropic, a small correction factor for anistoropy [see, for
source. This process is known as scattering. It occun on example, Chandrasekhar, 1950; Kerker, 1969] must be ap-

18-7
CHAPTER 18
‘f’able 18-3. Depolarization factor A of atrnos~nc gases for incident unpolarized light [Young, 1980].

I Gas I Rowell et al. [1971] Alms et al. [19751 Baas and van den Hout [1979] Young [1980] I

N2 0.0214 and 0.0242 0.0210 * 0.0004 0.0200 * 0.006 —

o* 0.0566 and 0.0587 — 0.0580 * 0.~2 —
co* 0.073 and 0.075 0.0797 * 0.0010 0.077 t 0.002
Dry Air — — — 0.0279

plied to the molecular scattering in Equations (18. 10) or pressure). Values of u, [Equation (18. 13)] and k. [Equation
(18. 11). With this correction, Equation (18. 10) becomes (18. 12)] arc presented in Table 184 for wavelengths be-
tween 0.25 and 4.0 pm. The largest uncertainty in the
2

()
tabulated values is due to the uncertainty in the values of
24m3 m2 – 1 6+3A
k,(k) = — — — (18.12) the depolarization factor A used. Young [ 1980] indicates an
N’k4 m2 + 2 6–7A’
error of a “few percent” in his value of A for air (Table 18-
3). This corresponds to an uncertainty of f 0.1 to 0.2% in
where the depolarization A is the ratio of the two polarid US(A) or ks(A).
intensities iJi, at a 90” scattering angle (see Section 18.2.1.5). The angular distribution of the light scattered by at-
Table 18-3 gives depolarization values for atmospheric g- mospheric molecules is given by the Rayleigh scattering
[Young, 1980]. It is often convenient to use the scattering phase function:
cross section per molecule

1
32

u,(k)
= ~
N’ = ~
24T3 —.
m2—1

mz + 2
2
6+3A
6 _ 7A (18.13)
P(e) = —
161T(2 + A) [
(1 + A) + (1 – A)cos26

(18.14a)
()

which has the advantage over the scattering coefficient in Neglecting the correction for depolarization, this simplifies
that it is independent of air density (or temperature and to

Table 18-4. Rayleigh aeatteriog m aeetion u.(A) and Rayleigb ~g coefficient ks(k).

Wavelength a,(A) h(A)

(pm) (cm2/molec.) (for t = 273.15 K and p = 1013.25mb)

0.25 1.243E-25 3.339E-1

0.30 5.605E-26 1.506E-1
0.35 2.913E-26 7.829E-2
0.40 1.668E-26 4.482E-2
0.45 1.025E-26 2.754E-2
0.50 6.650E-27 1.787E-2
0.55 4.505E-27 1.211 E-2
0.60 3. 161E-27 8.4%E-3
0.65 2.284E-27 6.139E-3
0.70 1.692E-27 4.547E-3
0.80 9.864E-28 2.651 E-3
0.90 6. 135E-28 1.649E-3
1.0 4.014E-28 1.079E-3
1.2 1.929E-28 5.184E4
1.4 1.039E-28 2.793E-4
1.6 6.083E-29 1.635E-4
1.8 3.794E-29 1.020E-4
2.0 2.488E-29 6.695E-5
2.5 1.018E-29 2.736E-5
3.0 4.906E-30 1.318E-5
4.0 1.552E-30 4. 169E-6

18-8
 OPTICAL AND INFRARED PROPERTIES OF THE ATMOSPHERE

HEIGHT iKhl) FCSSIBLE DuST CLOUD IN ORBIT

P(0)= (18.14b) I
+[’“0s2’13
2m hlETEOR

FLO\b
which is often used for the Rayleigh phase function. The
DECELERATION
phase function gives the probability distribution for the scat- Im
NOCTILULENT C1OUOS
tered light, so that P(o) do is the fraction of the scattered 80

radiation that enters a solid angle dQ about the scattering 60 ZONE OF SLOtVLY SETTLING

Cosrvllc OUST
angle 9. 40
t SGhlE LAYEREb STRUCTURE
30

AEROSOLS OF VCLCANIC ORIGIN

20 JLENGE LAYER(SI
18.2 AEROSOLS, CLOUD, AND 15 L
( SOhlE PHOTOCHEhll CAL FORP>lATION

PRECIPITATION PARTICLES
:
8
==””~ --”--””-
–
POSSIBLE LAYER BE1OLV TRCPOf’AUSE
““”-=
Propagation of electromagnetic radiation through the at- 6
mosphere at optical/infrared frequencies is affected by ab- ExPONENTI AL OECREASE

4 WITH HEIGHT OF
sorption and scattering from air molecules and particulate
3 AEROSOL CONCINTRATIOk
matter (haze, dust, fog, and cloud droplets) suspended in
1 ““
the air. Scattering and absorption by haze particles or aer- 2
AEROSC1 cONTENT OETERh!l NEO BY ME TECRCLOGICAL CONDITIONS

osols becomes the dominant factor in the bounda~ layer I

WATER VAPOR CONTENT OF ATNSPHERE CHARACTER OF THL SURFAC[

near the earth’s surface, especially in the visible, and under o

low visibility conditions at all wavelengths. Atmospheric

Figure 18-7. Charactenstim of atmospheric aerosols.
aerosol particles in the atmosphere vary greatly in their
concentration, size, and composition, and consequently in
their effects on optical and infrared radiation.
and agglomeration; this component is called the accumu-
lation mode. The second mode is made up of larger particles
18.2.1 Aerosols, Geographic, approximately 1 pm diameter and larger which are produced
and Temporal Variations by mechanical processes such as resuspension from soil or
sea spray; this component is called the coarse particle mode.
18.2.1.1 Composition, Sources and Sinks, Refractive
Index, Particle Shapes. Figure 18-7 summarizes the gen-
Table 18-5. Estimates of particles smaller than 20 ~m radius emitted into
eral characteristics of atmospheric aerosols. Aerosols in the or formed in the atmosphere (106 metric tons/year) [SMIC,
boundary layer of 1–2 km have the greatest variability. 1971].

These aerosols consist of a variety of natural and manmade

Natural
chemical compounds, inorganic as well as organic. Particles
Soil and rock debris* 100-500
are transported into the atmosphere from their sources at or
Forest fires and slash-burning debris* 3–150
near the surface, or they may be formed in the atmosphere
Sea salt 300
by chemical reactions from gaseous components, often with
Volcanic debris 25-150
the influence of solar radiation through photochemical pro-
Particles formed from gaseous emissions:
cesses. Since more than two-thirds of the earth’s surface is
Sulfate from H’S 130-200
covered by oceans, the maritime aerosol component, which
Ammonium salts from NHS 80–270
consists largely of sea salt particles from the sea water,
Nitrate from NOX 60-430
forms the most uniquely identifiable aerosol. Over land areas,
Hydrocarbons from plant exudations 75-200
soil particles and dust are an important component. Organic
particles from vegetation sources are also an important aer- Subtotal 773-2200
osol component. Table 18-5 gives a breakdown of the global
Man-made
atmospheric aerosol composition [SMIC, 1971].
Particles (direct emissions) 10-90
In the troposphere above the boundary layer the distri-
Particles formed from gaseous emissions:
bution and composition of aerosol particles becomes less
Sulfate from S02 130-200
dependent on the geography and on the variability of sources
Nitrate from NOX 30-35
near the surface. Experimental data [Whitby and Cantrell,
Hydrocarbons 15-90
1976 and Whitby, 1973] point toward an aerosol which is
composed of several modes, each having a different origin Subtotal 185415
and history. The most clearly identifiable two components
Total 958-2615
are submicron size particles formed from gaseous compo-
nents and are still going through processes of coagulation *Includes unknown amounts of indirect manmade contributions.

18-9
CHAPTER 18
In the stratospheric region from 10–30 km, measure- 18.2.1.2 Size Distributions. Over the past 20 to 30 yem,
ments have shown a background aerosol that has a rather much emphasis has been put on measurements of the aerosol
uniform global distribution. This aerosol is considered to particle size distribution, realizing that natural aerosols were
be mostly composed of sulfate particles formed by photo- far from being monodisperse. Junge [ 1963] suggested that
chemical reactions. The stratospheric aerosol background most aerosol size distributions over the radius range from
can occasionally be increased by factors of 100 or more due 0.1 to 10 pm, and even larger sizes, could be described by
to the injection of dust and SOZ from massive volcanic an inverse power law dN/d \og r = cr - Y with c being a
eruptions. Once these volcanic particles and gases have been constant and y being between 2 and 4. This means that the
injected into the stratosphere they are spread out over large number of aerosol particles decreases rapidly with increas-
portions of the globe by the stratospheric circulation. While ing particle size. It has already been mentioned that more
the dust particles may settle out within a few months, the detailed measurements point towards a multimodal distri-
SOZ is converted to sulfuric acid through photo-chemical bution. This becomes especially noticeable if the distribu-
reactions. This enhanced layer of approximately 7590 sul- tion is presented as particle volume distribution dv/dr rather
furic acid droplets may remain for one to two years. For a than dN/dr. One can assume that these size distributions are
complete review of the stratospheric aerosols see Turco et the result of a dynamic equilibrium between the various
al. [1982]. processes which act on the aerosol population. However the
Only a very small portion of the total aerosol content interaction of these various processes is complex and no
of the atmosphere exists above 30 km. However, when the general theory for aerosol fortnation processes exists at this
effects of the lower atmosphere are small, such as in satellite time.
observations of the limb of the earth’s atmosphere or de-
termining the ozone distribution by inverting measurements 18.2.1.3 Vertical Profiles. The change in aerosol prop-
of backscattered ultraviolet from the atmosphere, the effects erties and distribution with altitude is very closely tied to
of the aerosols in the upper atmosphere can be significant. the vertical structure of the atmosphere, and in the tropo-
Newkirk and Eddy [1964] and later Rosen [1969] con- sphere especially to the “weather” processes. In general the
cluded that the major component of the normal upper at- number of aerosol particles decreases much more rapidly
mospheric aerosols is meteoric dust. Meteoric or cometary with altitude then the molecular air density; however, wide
dust also form layers occasionally observed in the upper variations in the ve~ical profile do occur. Within the bound-
atmosphere. The optical effects of small particles are de- ary layer of the atmosphem, aerosol vertical mixing is strongly
termined largely by their refractive index. influenced by the temperature profile.
Aerosol samples from all over the world have been col-
lected and analyzed for their optical and infrared refractive 18.2.1.4 Models of the Aerosol Properties. There are
index properties. Based on such measurements, the aerosol many scientific and technical reasons why it is necessary to
in rural, non urban areas can be assumed to be composed develop models for atmospheric aerosol and cloud particles.
of a mixture of water soluble substances (ammonium and They are needed to make estimates of the transmittance,
calcium sulfates and organic compounds) and dust-like aer- angular light scattering distribution, contrast reduction, sky
osols [Volz, 1972 a, b]. In urban areas the rural aerosol radiance, or other atmospheric optical properties or effects
gets modified by the addition of aerosols from combustion (see Section 18.4).
products and industrial sources. There is no experimental Models for the optical properties of aerosols have been
evidence for any systematic change in aerosol refractive developed at AFGL and elsewhere. [Elterman, 1964, 1968,
index as a function of altitude through the troposphere. The 1970; Ivlev, 1967; Deirmendjian, 1964, 1969; McClatchey
refractive index of the volcanic components in the strato- et al., 1970 and 1972; Shettle and Fenn, 1976; Toon and
spheric aerosols was measured by Volz [1973] on volcanic Pollack, 1976; Hanel and Bullrich, 1978; and Nilsson, 1979].
dust samples. The refractive index for meteoric dust was This chapter describes aerosol models and their optical prop-
determined by Shettle and Volz [1976] based on the com- erties for the lower and upper atmosphere [Shettle and Fenn
position of meteorites falling on the earth. 1976, 1979]. The models presented below are based on a
Although the aerosol particle size (see Section 18.2. 1.2) review of the available experimental data on the nature of
is the most important parameter determining small particle aerosols, their sizes, distribution, and variability. However,
scattering properties, the effect of particle shape cannot be it must be emphasized that these models represent only a
ignored; it becomes significant especially for particles where simple, generalized version of typical conditions. It is not
size is large compared to the waelength of the incident practical to include all the details of natural particle distri-
radiation. However, because of poor understanding of these butions nor are existing experimental data sufficient to de-
shape effects, essentially all applications-oriented aerosol scribe the frequency of occurrence of the different condi-
models assume spherical particle shapes. For more infor- tions. While these particulate models were developed to be
mation on nonspherical particle scattering properties see as representative as possible of different atmospheric con-
Schuerman [ 1980]. ditions, the following point should be kept in mind when

18-10
 OPTICAL AND INFRARED PROPERTIES OF THE ATMOSPHERE
Table 18-6. Characteristics of the aerosol models as a function of relative humidity.

Aerosol Model Size Distribution Type

Ni ri ui

Rural 0.999875 0.03 0.35 Mixture of Water Soluble

0.000125 0.5 0.4 and Dust-Like Aerosols
Urban 0.999875 0.03 0.35 Rural Aerosol Mixtures
0.000125 0.5 0.4 with Soot-Like Aerosols
Maritime
Continental 0.99 0.03 0.35 Rural Aerosol Mixture
Origin
Oceanic 0.01 0.3 0.4 Sea Salt Solution in Water
Origin
Tropospheric 1.0 0.03 0.35 Rural Aerosol Mixture

using-. any such model: given the natural variability of the radius and the number density of r,. This form of distribution
atmosphere, almost any particle model is supported by some function represents the multimodal nature of the atmospheric
measurements and no model (or set of models) will be aerosols. Following the usual convention, log is the loga-
consistent with all measurements. rithm to the base 10 and in is the logarithm to the base e.
The size distributions for the different aerosol models Four different aerosol models for the atmospheric
are represented by one or the sum of two log-normal dis- boundary layer have been developed. They differ in particle
tributions: size distribution and particle refractive index. Table 18-6
lists the parameters defining the size distributions in ac-
cordance with Equation (18. 15) for these models. These
mode radii correspond to moderate humidities (7070 to 80%);
Ni
1
dN(r) values of ri as a function of humidity are given in Table
n(r) = — =;
dr i=] {[ ln(lO) “ r . u, G 18-7.
The choices of N in Table 18-6 are normalized to cor-

11
(lOg r – 10g r,)’
.[
x exp –
2 Ui2 ‘ (18.15)
respond to 1 particle/cm 3. The actual size distributions can
be renormalized to give the comect extinction coefficients
for the altitude and the visibility being used. The continental
and oceanic components of the maritime model can be used
in various proportions depending on the prevailing winds—
where N(r) is the cumulative number density of particles of particularly in coastal regions.
radius r, u is the standard deviation, and ri, Ni are the mode As the relative humidity increases, water vapor con-

Table 18-7. Mode radii for the aerosol models as a function of relative humidity

Rural Urban
Relative Tropospheric
Humidity rl rl r2 Maritime r, r2

o% 0.02700 0.02700 0.4300 0.1600 0.02500 0.4000

50% 0.02748 0.02748 0.4377 0.1711 0.02563 0.4113
70% 0.02846 0.02846 0.4571 0.2041 0.02911 0.4777
80% 0.03274 0.03274 0.5477 0.3180 0.03514 0.5808
90% 0.03884 0.03884 0.6462 0.3803 0.04187 0.7061
95% 0.04238 0.04238 0.7078 0.4606 0.04904 0.8634
98% 0.04751 0.04751 0.9728 0.6024 0.05996 1.1691
99% 0.05215 0.05215 1.1755 0.7505 0.06847 1.4858

18-11
CHAPTER 18
denses out of the atmosphere on the particulate suspended
2.4
in the atmosphere. This condensed water increases the size
of the aerosols and changes their composition and effective 2.2 1
L
— OCEANIC AEROSOL
------- WATER ( HALE .9 QUERRY)
refractive index. The resulting effect of the aerosol on the SEA SALT (VOLZ)

absorption and scattering of light will correspondingly be ,., , ,.

modified. There have been a number of studies on the change ?

. . .. . . .
of aerosol properties as a function of relative humidity. The
‘ ---_-—___ ~
most comprehensive of these, especially in terms of the .
resulting effects on the aerosol optical properties, is the work I .0
of Hanel [1976].
The “Rural Model” is intended to represent the aerosol
under conditions where it is not directly influenced by urban
and/or industrial aerosol sources. The rural aerosols are
assumed to be composed of a mixture of 70~0 water soluble
,..-Q
Iv
substances (ammonium and calcium sulfate and also organic .1 1, 10 100

compounds) and 30% dust-like aerosols. The refractive in- WAVELENG~ ( ~ m )

dex for these components based on the measurements of

Figure 18-9. Refractive index of oceanic aerosol, water, and sea salt.
Volz [ 1972a,b; 1973] is shown in Figure 18-8. The refrac-
tive index is in general a complex quantity with a real and
imaginary part. These refractive index data weighted by the modified by humidity changes (Figure 18-10). The effective
mixing ratio of the two components are consistent with other refractive indices for the two size components were then
direct measurements and with values inferred from in situ computed.
measurements. For the refractive index of water, the survey In urban areas the air with a rural aerosol background
of Hale and Querry [1973] was used. While there are some is primarily modified by the addition of aerosols from com-
minor differences between the optical constants in Hale and bustion products and industrial sources. The urban aerosol
Querry’s survey and the more recent measurements, these model therefore was taken to be a mixture of the rural aerosol
differences are comparable to the experimental errors and with carbonaceous aerosols. The proportions of the soot-
are small compared with the other uncertainties in the model like carbonaceous aerosols and the rural type of aerosol
parameters. These refractive index data are shown in Figure mixture are assumed to be 20% and 80%, respectively. The
18-9. The resulting number density distributions n(r), are soot-like aerosols are assumed to have the same size dis-
shown in Figure 18-10. To allow for the dependence of the tribution as both components of the rural model. The re-
humidity effects on the size of the dry aerosol, the growth fractive index of the soot-like aerosols are based on the soot
of the aerosol was computed separately for the accumulation data in Twitty and Weinman’s [ 1971] survey of the refrac-
and co~se particle components. In accounting for the aer- tive index of carbonaceous materials. As with the rural
osol growth, changes in the width of the size distribution model, a composite urban aerosol refractive index was de-
were assumed negligible so only the mode radius ri was termined at each wavelength.
The aerosol compositions and distributions of oceanic
origin are significantly different from continental aerosol
26
. types. These aerosols are largely sea-salt particles produced
— WATER SOLUBLES (VOLZ ) /’
24
- --–- DUST UKE (VOLZ ) /’ by the evaporation of sea-spray droplets that have grown
22 ----- SOOT (TWITTY e WEINMAN) .“ again due to aggregation of water under high relative hu-
/ 7
20
,’
/’ midity conditions. However, even over the ocean there is
“1
IB
-- —------- -
..’
i a more or less pronounced continental aerosol background
/
16 // J that, mixed with the aerosol of oceanic origin, forms a fairly
1
!4 . uniform maritime aerosol. It is representative for the bound-
12
.
ary layer in the lower 2–3 km in the atmosphere over the
Fi ‘i ‘ “i
oceans, but also may occur over the continents in a maritime
air mass. This maritime model should be distinguished from
the fresh sea-spray aerosol that exists in the lower 10–20
m above the ocean surface and is strongly dependent on
wind speed.
,.4 ;. , .J.ul—-.— —L—.. L , i
The maritime aerosol model, therefore, is composed of
I I 10 100 two components: ( 1) the sea-salt component and (2) a con-
WAVELENGTH U- 1
tinental component assumed to be identical to the rural aer-
Figure 18-8. Refractive index for the dry rural and urban aerosol com- osol with the exception that the very large particles are
ponents. eliminated since they will eventually be lost due to fallout

18-12
 OPTICAL AND INFRARED PROPERTIES OF THE ATMOSPHERE
rural model (70% water soluble and 30% dust-like). The
size distribution is modified from the rural model by elim-
inating the large (or coarse) particle (ri = 0.5 ~m) com-
105= - ponent of the size distribution because of ~the longer resi-
dence of aerosols above the boundary and the expected
differential loss of the larger particles. This leaves the log-
normal distribution with the small (or accumulation particle
ti
n component). This is consistent with the changes in aerosol
\ size distribution with altitude suggested by Whitby and Can-
z
o trell [1976]. The dependence of particle size on relative
humidity is the same as for the small particle component of
10’ the rural model, and is shown in Figure 18-12.
Based on observations of stratospheric aerosol behavior
four different vertical distribution models for the strato-
sphere were developed: a background model and three dif-
ferent volcanic aerosol profiles (moderate, high, and ex-
treme). These models represent either different amounts of
volcanic material injected into the stratosphere or the de-
crease over time from the extreme conditions following a
major event such as the Krakatoa explosion. The experi-
mental data also confirm a seasonal trend in the stratospheric
aerosol distributions. Figure 18-13 shows these stratospheric
models and also for comparison the 1968 Elterman model

~0-6
:.-3 ““ I

~’o-i
1 !

““ ~ro-i “>’ ;’.O’ “’”

1

~’ol’
\\

“A
“. .
~ ~2
M9RITIME REROSOL MODELS
I , , ,
RADIUS (pm) 1“’”
i
Figure 18-10. Aerosol number distribution (cm-3 pm-’) for the rural model
at different relative humidities with total particle concen-
trations fixed at 15000 cm-3.

as the air masses move across the oceans. For the size
distribution of the oceanic component, a log normal distri-
bution is used with r. = 0.3 pm for moderate relative hu-
midity ( = 80Yo) and u = 0.4. The relative proportions of
aerosol of oceanic or continental origins will vary prtrticu-
larly in coastal regions. To account for these variations, the
n
model permits the user to adjust the relative amounts of the
oceanic and continental types of aerosol. The number den-
sity distribution is shown in Figure 18-11. The refractive
index is based on that for a solution of sea salt in water,
using a weighted average of the refractive indices of water
and sea salt. The refractive index of the sea salt is primarily
taken from the measurements of Volz [1972 a, b]. A model
for maritime aerosols that accounts for production of oceanic
aerosol at the sea surface by white caps and spray as a
function of wind speed and the changes in drop size due to , 1 , \ \
varying relative humidity was developed recently by Gath- ‘“-3 ,’”-2 ~ n-1 * no , nl 1(

man [1983]. With some modifications this model has been

RADIUS ( pm)
incorporated into LOWTRAN 6 (see Section 18.4.2).
The tropospheric aerosol model represents the aerosols
Figure 18-1 I. Aerosol number distribution (cm-~ yin-’) for the maritime
within the troposphere above the boundary layer. These model at different relative humidities with total particle
aerosols are assumed to have the same composition as the concentrations fixed at 4000 cm-3.

18-13
CHAPTER 18
TROPOSPHERIC RER13S0L MODELS VERTICAL AEROSOL DISTRIBUTIOAS
.— .- . .
‘0’ ~ ——1-- ‘-- T ‘—- , ,

70 ‘. WECULAR
.RAYLEIGH EXTREME ATMOSPHERIC AEROSOL MODELS
60 .,, FOR OPTICAL CALCUtATIONS .

50 b. .
. . ..
. .. . . .
4or NORMAL . . .. .
. .. ..
. . . . .. .,
..
3’A . . .. MESOSPHERE
CH~GE IN
: SCALE

<
.
3-

l“,,’,
I “., , // WINTER
STRATOSPHERE

‘Y
BAGKGVUUND
15 1.
SPRING SUMMER ‘/
FALL WINTER
‘./
~ .,’/
\ ,{

lo- .
\ :.
v &E’’’’”A” “’se)
\
\: q,
\. TROPOSPHERE
5+
“h
‘~. : \

RADIUS (pm)
MARITIME, RuRAL AND URBAN
WITH SURFACE VISIBILITIES 50,23,10,5,2km
..—
10-’ ,.-4
L.
,..3

AT1 LNuATION COEFFICIENT AT 0 55PM IKm-rl

.,...
‘\h-
...

,..2
‘,,
:.
‘b... .,,310
.F.
!0
..=—
5,~

Ftgure 18-12. Aerosol number distribution (cm-3 ~m-’) for the tropos-
Egure 18-13 The vertical distribution of the aerosol extinction (at 0.55
pheric model at different relative humidities with total par-
microns) for the different models, Also shown for com-
ticle concentrations fixed at 10000 cm-3.
parison are the Rayleigh profile (dotted line) and Elterman’s
[ 1968] model. Between 2 and 30 km. where a distinction
on a seasonal basis is made, the spring–summer conditions
are indicated with a solid line and fall–winter conditions
based on data collected during the period several years after are indicated by a dashed line.
the Agung eruption in 1963 andtherefore representative for
the moderate volcanic conditions.
The stratospheric background aerosols were assumed to dN
= n(r) = Ar” exp ( – br”) (18.16)
be a 75% solution of sulfuric acid in water following the dr
work of Rosen [1971] and Toon and Pollack [1973]. The
complex refractive index as a function of wavelength is whose parameters are given in Table 18-8. Either of the two
based onthemeasurements of Remsberg [1971 and 1973] volcanic size distribution models are appropriate to use with
and Palmer and Williams [1975]. The refractive index for the “moderate” or “high” volcanic vertical profiles, de-
the volcanic models is based on the measurements of Volz pending on whether these profiles represent a major eruption
[1973] on volcanic dust. after a couple of years or a recent weaker eruption. It should
There are two volcanic aerosol size models: a “Fresh be noted that the profiles may have a much more layered
Volcanic Model” representing the size distribution of aer- structure than is shown in the model distributions, partic-
osols shortly after a volcanic eruption, and a “Volcanic ularly shortly after a volcanic injection of dust into the
Model” representing the aerosol about a year after an erup- atmosphere.
tion. Both size distributions were chosen mainly on the basis The major component of the normal upper atmospheric
of Mossop’s [1964] measurements following the eruption
of Mt. Agung. The size distribution was also made con-
sistent with the observed wavelength dependence of ex- Table 18-8. Modified gamma distribution parameters.
tinction due to volcanic aerosols, and in the case of the
“Fresh Volcanic Model” consistent with the observation of I A a Y bl
optical phenomena such as Bishop’s rings, and a blue or 37,4.0 I I 18
Background Stratospheric
green sun sometimes observed following major volcanic
Fresh Volcanic 341.33 1 0.5 8
eruptions. These size distributions are represented by a mod-
Aged Volcanic 5461.33 I 0.5 16
ified gamma distribution:

18-14
 OPTICAL AND INFRARED PROPERTIES OF THE ATMOSPHERE
aerosol is considered to be meteoric dust. Meteoric or com- il — i2
P=— (18.18)
etary dust also forms some of the layers occasionally ob- il + i2
served in the upper atmosphere. The refractive index of
meteoric dust is based on the work of Shettle and Volz Integration of the scattered light over all directions (47r
[1976] who determined the complex refractive index for a steradians) gives
mixture of chondrite dust which represents the major type
4T
of meteorite falling on the earth. The size distribution has
I*d~=Io,A”C,. (18.19)
been represented by a log-normal distribution (equation 18.15 J 0

with i = 1) whose parameters are given as

The ratio of C, (scattering cross section) to the geometrical
N, = 1 cm-3, r = 0.03 pm, U! = 0.5 cross section of the scattering particle, r2m, is called scat-
tering efficiency factor Q,. If light is traveling through a
Figure 18-13 shows in addition to the normal Upper atmos- scattering medium containing N particles per unit volume,
pheric model an extreme model similar to Ivlev’s [1969) the loss of light due to scattering per unit path length dx is
model for the upper atmosphere. Attenuation coefficients
this large occur in layers with thicknesses no greater than dl, =–IA” C,- N”dx. (18.20)
a few kilometers. At different times these layers can be
either micrometeoric dust or noctilucent clouds. The wave- Integration of Equation ( 18.20) gives the Bouguer extinction
length dependent refractive indices for the different aerosol law:
types are listed in Table 18-9.
I ~,transmitted = 10,A exp ( – k,x) (18.21)

18.2.1.5 Aerosol Extinction, Scattering, Polarization. with k, = N “ C, called the scattering coefficient. Similarly
Once the size distribution and refractive index of the aerosol to k, one can define an absorption coefficient ka for ab-
models are specified, the optical properties can be deter- sorption by absorbing aerosol particles.
mined. The effec@ of small particles on the propagation of
radiation are defined by the coefficients for extinction, scat- h = k, + k. is called the extinction coefficient.
tering and absorption, the angular scattering function, and
the polarization of scattered radiation. For Mie scattering calculations, the aerosol particles are
The Mie [1908] theory treats the scattering of light by assumed to be spherical, which in general is not true. While
homogeneous particles of arbitrary size and refractive index liquid aerosols are approximately spherical, dry particles
(for Rayleigh scattering see Section 18. 1.4). For detailed usually are irregularly shaped. However, it can be argued
discussions on particle scattering see textbooks by van de that many of the measurements of aerosol size distributions
Hulst [1957], Kerker [1969], and McCartney [1976]. The directly measure scattered light from the aerosol particles,
spatial distribution of scattered light for incident unpolarized and the size “assigned” to the particle is the size of a sphere
light of intensity 10,Ais that has similar scattering properties to the measured par-
ticle. Furthermore, the irregularly shaped particles are clos-
1 est to “equivalent” spheres in their scattering properties in
IA=10,AOOk2. R2 [i, (a, m, 8) + i, (a, m, 6)],
L the forward direction generally used for size measurements.
(18.17) Therefore, the resulting size distribution is the size distri-
bution of spheres that have similar optical properties to those
where k = 2 T/A, of the actual aerosol particles. Also, studies by Chylek et al.
[1976] and Holland and Gagne [1970] indicate that for par-
R= Distance from the scattering particle center,
ticles of equal overall dimension but different shapes, the
a= 2 mr/A, the size parameter, r is particle radius, spherical particle extinction has the highest values.
m. The computed attenuation coefficients (extinction, scat-
complex refractive index of the particle
tering, and absorption) for the model aerosols above are
relative to its surrounding medium,
shown in Tables 18-10 a thru c and Figures 18-14 thru
o= angle between the incident and the scattered 18-22.
beams (6 = 0° is defined as forward scattering. The attenuation coefficients for the rural aerosol model
at 5070 relative humidity as a function of wavelength are
il = perpendicular polarized component
shown in Figure 18-14. To show the effect of variations of
iz = parallel polarized component relative humidity on the aerosol extinction, the rural model
extinction has been calculated over a range of humidities
The dimensionless intensity functions define degree of linear for a constant total number density. The resulting extinction
polarization P: versus wavelength is shown in Figure 18-15 for several

18-15
CHAPTER 18
Table 18-9a. Refractive indices for different aerosol types

Wavelength Sea Salt Oceanic Water Ice Meteoric

(pm) real imag. real imag. real imag. real imag. real imag.

0.200 1.510 – 1.00E-04 1.429 – 2.87E-05 1.396 – 1.1 OE-O7 1.394 – 1.50E-08 1.515 – 1.23E-05
0.250 1.510 –5.00E-06 1.404 – 1.45E-06 1.362 – 3.35E-08 1.351 –8.60E-09 1.515 –2.41E-05
0.300 1.510 – 2.00E-06 1.395 – 5.83E-07 1.349 – 1.60E-08 1.334 –5.50E-09 1.515 –4.18E-05
0.337 1.510 –4.00E-07 1.392 – 1.20E-07 1.345 – 8.45E-09 1.326 –4.50E-09 1.514 –5.94E-05
0.400 1.500 – 3.00E-08 1.385 –9.90E-09 1.339 – 1.86E-09 1.320 –2.71E-09 1.514 –9.95E-05
0.488 1.500 – 2.00E-08 1.382 –6.41E-09 1.335 –9.69E-1O 1.313 –1.75E-09 1.513 –1.81E-04
0.515 1.500 – 1.00E-08 1.381 –3.70E-09 1.334 –1.18E-09 1.312 –2.19E-09 1.513 –2.13E-04
0.550 1.500 – 1.00E-08 1.381 –4.26E-09 1.333 – 1.96E-09 1.311 –3.llE-09 1.513 –2.61E-04
0.633 1.490 –2.00E-08 1.377 – 1.62E-08 1.332 – 1.46E-08 1.308 – 1.09E-08 1.512 –3.99E-04
0.694 1.490 – 1.00E-07 1.376 – 5.04E-08 1.331 –3.05E-08 1.306 – 2.62E-08 1.511 –5.30E-04
0.860 1.480 – 3.00E-06 1.372 – 1.09E-06 1.329 –3.29E-07 1.303 –2.15E-07 1.509 – 1.02E-03
. 1.060 1.470 –2.00E-04 1.367 –6.OIE-05 1.326 –4.18E-06 1.300 – 1.96E-06 1.506 – 1.95E-03
1.300 1.470 –4.00E-04 1.365 – 1.41E-04 1.323 –3.69E-05 1.295 – 1.32E-05 1.501 –3.72E-03
1.536 1.460 –6.00E-04 1.359 –2.43E-04 1.318 –9.97E-05 1.290 –6.1OE-O4 1.495 –6.34E-03
1.800 I .450 –8.00E-04 1.351 –3.llE-04 1.312 –1.15E-04 1.282 – 1.13 E-M 1.488 – 1.06E-02
2.000 1.450 – 1.00E-03 1.347 – 1.07E-03 1.306 – 1.1 OE-O3 1.273 – 1.61E-03 1.482 – 1.51E-02
2.250 1.440 – 2.00E-03 1.334 –8.50E-04 1.292 –3.90E-04 1.256 –2.13E-04 1.474 – 2.24E-02
2.500 1.430 – 4.00E-03 1.309 – 2.39E-03 1.261 – 1.74E-03 1.225 –7.95E-04 1.467 – 3. 18E-02
2.700 1.400 –7.00E-03 1.249 – 1.56E-02 1.188 – 1.90E-02 1.163 –2.93E-03 1.462 –4.1OE-O2
3.000 1.610 – 1.00E-02 1.439 –0.197 1.371 –0.272 1.045 –0.429 1.456 – 5.73E-02
3.200 1.490 – 3.00E-03 1.481 –6.69E-02 1.478 – 9.24E-02 1.652 --0.283 1.454 – 6.94E-02
3.392 1.480 – 2.00E-03 1.439 – 1.51E-02 1.422 – 2.04E-02 I.51O –4.OIE-02 1.454 –8. 15E-02
3.500 1.480 – 1.60E-03 1.423 –7. 17E-03 1.400 – 9.40E-03 1.453 – 1.61E-02 1.455 – 8.82E-02
3.750 1.470 – 1.40E-03 1.398 – 2.90E-03 1.369 – 3 .50E-03 1.391 –7.00E-03 1.459 –o. 103
4.000 1.480 – 1.40E-03 1.388 – 3.69E-03 1.351 –4.60E-03 1.361 – 1.00E-02 1.466 –0.116
4.5W 1.490 – 1.40E-03 1.377 – 9.97E-03 1.332 – 1.34E-02 1.340 – 2.87E-02 1.485 –O. lj~
5.000 1.470 – 2.50E-03 1.366 – 9.57E-03 1.325 – 1.24E-02 1.327 – 1.20E-02 1.500 –0.135
5.500 1.420 – 3 .60E-03 1.333 –9.31E-03 1.298 – 1.16E-02 1.299 –2.17E-02 1.508 –0.132
6.000 1.410 – 1.1 OE-O2 1.306 –7.96E-02 1.265 –0.107 1.296 – 6.47E-02 1.507 –0.126
6.200 1.600 –2.20E-02 1.431 –6.91E-02 1.363 –8.80E-02 1.313 –6.83E-02 1.504 –0.124
6.500 1.460 –5.00E-03 1.374 – 2.94E-02 1.339 – 3.92E-02 1.320 – 5.59E-02 1.497 –0.121
7.200 1.420 – 7 .00E-03 1.343 –2.49E-02 1.312 –3.21E-02 1.318 –5.44E-02 1.469 –0.119
7.900 1.400 – 1.30E-02 1.324 – 2.79E-02 1.294 – 3.39E-02 1.313 –4.79E-02 1.422 –O. 130
8.200 1.420 – 2.00E-02 1.324 –3.08E-02 1.286 –3.51E-02 1.306 –3.90E-02 1.395 –0.142
8.500 1.480 – 2.60E-02 1.336 – 3.36E-02 1.278 – 3.67E-02 1.291 –3.91E-02 1.363 –0.162
8.700 1.600 – 3.00E-02 1.366 – 3.56E-02 1.272 – 3.79E-02 1.282 –4.00E-02 1.339 –O. 182
9.m 1.650 – 2.80E-02 1.373 – 3.65E-02 1.262 – 3.99E-02 1.269 – 4.29E-02 1.302 –0.228
9.200 1.610 –2.60E-02 1.356 –3.71E-02 1.255 –4. 15E-02 1.261 –4.46E-02 1.281 –0.273
9.500 1.580 – 1.80E-02 1.339 – 3.68E-02 1.243 – 4.44E-02 1.245 –4.59E-02 1.272 –0.360
9.800 1.560 – 1.60E-02 1.324 – 3.88E-02 1.229 – 4.79E-02 1.219 –4.70E-02 1.310 –0.450
10.OOO 1.540 – 1.50E-02 1.310 –4.06E-02 1.218 –5.08E-02 1.197 –5.1OE-O2 1.355 –0.488
10,591 1.500 – 1.40E-02 1.271 –5.22E-02 1.179 –6.74E-02 1.098 –0.131 1.419 –0.547
11.m 1.480 – 1.40E-02 1.246 –7.31E-02 1.153 –9.68E-02 1.093 –0.239 1.509 –0.691
11.500 1.480 – 1.40E-02 1.227 –O. 105 1.126 –0.142 1.176 –0.360 1.847 – 0.634
12.500 1.420 – 1.60E-02 1.208 –0.190 1.123 –0.259 1.387 –0.422 1.796 –0.252
13.000 1.410 – 1.80E-02 1.221 –0.223 1.146 –0.305 1.472 –0.389 1.711 –0.219
14.000 1.410 –2.30E-02 1.267 –0.271 1.210 –0.370 1.569 –0.283 1.641 –0.217
14.800 1.430 – 3.00E-02 1.307 – 0.292 1.258 –0.396 1.579 –0.191 1.541 –0.198
15.000 1.450 – 3 .50E-02 1.321 –0.297 1.270 –0.402 1.572 –0.177 1.510 –0.206
16.400 1.560 –9.00E-02 1.407 –0.331 1.346 –0.427 1.531 –0.125 1.478 – 0.467
17.200 1,740 –0.120 1.487 –0.341 1.386 –0.429 1.534 –0.107 1.441 –0.400
18.000 1.780 –0.130 1.525 –0.341 1.423 – 0.426 1.522 –8.39E-02 1.354 –0.557
18.500 1.770 –0.135 1.536 – 0.339 1.448 –0.421 1.510 –7.60E-02 1.389 –0.705
20.000 1.760 –0.152 1.560 –0.324 1.480 –0.393 1.504 – 6.70E-02 1.803 –0.765
21.300 1.760 –0.165 1.568 –0.318 1.491 –0.379 1.481 – 3.85E-02 1.797 –0.556
22.500 1.760 –0.180 1.579 –0.316 1.506 –0.370 1.455 –2.91E-02 1.661 –0.592
25.000 1.760 –0.205 1.596 –0.313 1.531 –0.356 1.414 – 2.99E-02 1.983 –0.861
27.900 1.770 –0.275 1.612 –0.320 1.549 –0.339 1.358 – 4.90E-02 2.023 –0.666
30.000 1.770 –0.300 1.614 –0.320 1.551 –0.328 1,325 – 6.50E-02 2.149 –0.665
35.000 1.760 –0.500 1.597 –0.383 1.532 –0.336 1.226 –0.155 2.146 –0.380
40.000 1.740 – 1.000 1.582 –0.561 1.519 –0.385 1.202 –0.344 1.979 –0.359

18-16
 m
OPTICAL AND INFRARED PROPERTIES OF THE ATMOSPHERE
Table 18-9b. Refractive indices for different aerosol types

Water
wavelength Soluble Dust-Like soot 75% H*S04 Volcanic
(~m) real imag. real imag. real imag. real imag. real imag.

0,200 1,530 –7.00E-02 1.530 –7.00E-02 1.500 –0.350 1,498 –1.00E-08 1.500 –7.00E-02
0.250 1.530 –3.00E-02 1.530 –3.00E-02 1.620 –0.450 1.484 –1.00E-08 1.500 –3.00E-02
0.300 1.530 –8.00E-03 1.530 –8.00E-03 1.740 –0.470 1.469 –1.00E-08 1.500 –1.00E-02
0.337 1.530 -5.00E-03 1.530 –8.00E-03 1.750 –0.470 1.459 –1.00E-08 1.500 –8.00E-03
0.400 1.530 –5.00E-03 1.530 –8.00E-03 1.750 –0.460 1.440 –1.00E-08 1.500 –8.00E-03
0.488 1.530 –5.00E-03 1,530 –8.00E-03 1.750 –0.450 1.432 –1.00E-08 1.500 –8.00E-03
0.515 1.530 –5.00E-03 1.530 –8.00E-03 1.750 –0.450 1.431 –1.00E-08 1.500 –8.00E-03
0.550 1.530 –6.00E-03 1.530 –8.00E-03 1.750 –0.440 1.430 –1.00E-08 1.500 –8.00E-03
0.633 1.530 –6.00E-03 1.530 –8.00E-03 1.750 –0.430 1.429 –1.47E-08 1.500 –8.00E-03
0.694 1.530 –7.00E-03 1.530 –8.00E-03 1.750 –0.430 1.428 –1.99E-08 1.500 –8.00E-03
0.860 1.520 –1.20E-02 1.520 –8.00E-03 1.750 –0.430 1.425 –1.79E-07 1.5M –8.00E-03
1.060 1.520 –1.70E-02 1.520 –8.00E-03 1.750 –0.440 1.420 –1.50E-06 1.500 –8.00E-03
1.3m 1.510 –2.00E-02 1.460 –8.00E-03 1.760 –0.450 1.410 –1.00E-05 1.5N –8.00E-03
1.536 1.510 -2.30E-02 1.400 –8.00E-03 1.770 –0.460 1.403 –1.37E-04 1.490 –8.00E-03
1.800 1.460 –1.70E-02 1.330 –8.00E-03 1.790 –0.480 1.390 –5.50E-04 1.480 –8.00E-03
2.000 1.420 –8.00E-03 1.260 –8.00E-03 1.800 –0.490 1.384 –1.26E-03 1.460 –8.00E-03
2.250 1.420 -1.00E-02 1.220 –9.00E-03 1.810 –0.500 1.370 –1.80E-03 1.460 –8.00E-03
2.500 1.420 –1.20E-02 1.180 –9.00E-03 1.820 –0.510 1.344 –3.76E-03 1.460 –9.00E-03
2.700 1.400 –5.50E-02 1.180 –1.30E-02 1.830 –0.520 1.303 –5.70E-03 1.460 –1.00E-02
3.000 1.420 –2.20E-02 1.160 –1.20E-02 1.840 –0.540 1.293 –9.55E-02 1.480 –1.30E-02
3.200 1.430 –8.00E-03 1.220 –1.00E-02 1.860 –0.540 1.311 –0.135 1.480 –1.40E-02
3.392 1.430 –7.00E-03 1.260 –1.30E-02 1.870 –0.550 1.352 –0.159 1.490 –1.20E-02
3.500 1.450 –5.00E-03 1.280 –1.1OE-O2 1.880 –0.560 1.376 –0.158 1.490 –l. IOE-02
3.750 1.452 –4.00E-03 1.270 –1.1OE-O2 1.900 –0.570 1.396 –0.131 1.500 –9.00E-03
4.000 1.455 –5.00E-03 1.260 –1.20E-02 1.920 –0.580 1.398 –0.126 1.500 –7.00E-03
4.500 1.460 –1.30E-02 1.260 –1.40E-02 1.940 –0.590 1.385 –0.120 1.520 –7.50E-03
5.000 1.450 –1.20E-02 1.250 –1.60E-02 1.970 –0.600 1.360 –0.121 1.510 –9.00E-03
5,500 1.440 –1.80E-02 1.220 –2. IOE-02 1.990 –0.610 1.337 –0.183 1.510 –1.20E-02
6.000 1.410 –2.30E-02 1.150 –3.70E-02 2.020 –0.620 1.425 –0.195 1.480 –1.50E-02
6.200 1.430 –2.70E-02 1.140 –3.90E-02 2.030 –0.625 1.424 –0.165 1.460 –1.80E-02
6.500 1.460 –3.30E-02 1.130 –4.20E-02 2.040 –0.630 1.370 –0.128 1.450 –2.40E-02
7.200 1.400 –7.00E-02 1.400 –5.50E-02 2.060 –0.650 1.210 –0.176 1.440 –4.50E-02
7.900 1.200 –6.50E-02 1.150 –4.00E-02 2.120 –0.670 1.140 –0.488 1.380 –7.20E-02
8.200 1.010 –0.100 1.130 –7.40E-02 2.130 –0.680 1.200 –0.645 1.340 –9.70E-02
8.500 1.3~ –0.215 1.300 –9.00E-02 2.150 –0.690 1.370 –0.755 1.620 –0.121
8.700 2.400 –0.290 1.400 –0.100 2.160 –0.690 1.530 –0.772 1.950 –0.170
9.000 2.560 –0.370 1.700 –0.140 2.170 –0.700 1.650 –0.633 2.200 –0.215
9.200 2.200 –0.420 1.720 –0.150 2.180 –0.700 1,600 –0.586 2.230 –0.240
9.500 1.950 –0.160 1.730 –0.162 2.190 –0.710 1.670 –0.750 2,250 –0.275
9.800 1.870 –9.50E-02 1.740 –0.162 2.200 –0.715 1.910 –0.680 2.280 –0.304
10.OOO 1.820 –9.00E-02 1.750 –0.162 2.210 –0.720 1.890 –0.455 2.300 –0.320
10.591 1.760 –7.00E-02 1.620 –0.120 2.220 –0.730 1.720 –0.340 2.200 –0.305
11.000 1.720 –5.00E-02 1.620 –0.105 2.230 –0.730 1.670 –0.485 2.150 –0.270
11.500 1.670 –4.70E-02 1.590 –0.100 2.240 –0.740 1.890 –0.374 2.050 –0.240
12.500 1.62Ci –5.30E-02 1.510 –9.00E-02 2.270 –0.750 1.740 –0.198 1.800 –0.155
13.000 1.620 –5.50E-02 1.470 –0.100 2.280 –0.760 1.690 –0.195 1.760 –0.148
14.000 1.5641 –7.30E-02 1.520 –8.50E-02 2.310 –0.775 1.640 –0.195 1.700 –0.145
14.800 1.440 –0.lti 1.570 –0.100 2.330 –0.790 1.610 –0.205 1.650 –0.157
15.m 1.420 –0.200 1.570 –0.100 2.330 –0.790 1.590 –0.211 1.650 –0.170
16.400 1.750 –0.160 1.600 –0.100 2.360 –0.810 1.520 –0.414 1,750 –0.200
17.200 2.080 –0.240 1.630 –0.100 2.380 –0.820 1.724 –0.590 1.850 –0.240
18.000 1.980 –0.180 1.W –0.115 2.400 –0.825 1.950 –0.410 2.m –0.305
18.500 1.850 –0.170 1.640 –0.120 2.410 –0.830 1.927 –0.302 2.100 –0.325
20.000 2.120 –0.220 1.680 –0.220 2.450 –0.850 1.823 –0.235 2.250 –0.318
21.3p 2.060 –0.230 1.770 –0.280 2.460 –0.860 1.780 –0.292 2.400 –0.290
22.500 2.000 –0.240 1.900 –0.280 2.480 –0.870 1.870 –0.315 2.5W –0.350
25.000 1.880 –0.280 1.970 –0.240 2.510 –0.890 1.930 –0.200 2.600 –0.400
27.900 1.840 –0.290 1.890 –0.320 2.540 –0.910 1.920 –0.180 2.500 –0.430
30.000 1.820 –0.3~ 1.800 –0.420 2.570 –0.930 1.920 –0.180 2.400 –0.450
35.m 1.920 –0.400 1.900 –0.500 2.630 –0.970 1.900 –0.190 2.300 –0.520
40.000 1.860 –0.500 2.100 –0.600 2.690 –1.000 1.890 –0.220 2.250 –0.650

18-17
 CHAPTER 18
Table 18- 10a. Vertical distribution of aerosol extinction (km ‘1) for a wavelength of 0.550 pm with rural model aerosol bundary layer

Height 50 km 23 km 10 km 5 km 2 km
(km) Met. Range Met. Range Met. Range Met. Range Met. Range
0.0 6.62E-02 1.58E-01 3.79E-01 7. 70E-02 1.94E+O0
1.0 4.15E-02 9.91E-02 3.79E-01 7.70E-01 1.94E+O0
1.5 3.26E-02 7.92E-02 3.79E-01 7.70E-01 1.94E+O0
2.0 2 .60E-02 6.21E-02 6.21E-02 6.21E-02 6.21E-02

. Fall-Winter Profiles Spring-Summer Profiles

50 km Surface 2 to 23 km Surface 50 km Surface 2 to 23 km Surface

Met. Range Met. Range Met. Range Met. Range
2.0 2.60E-02 6.21E-02 2.60E-02 6.21E-02
3.0 1. 14E-02 2.72E-02 1.46E-02 3.46E-02
4.0 6.43E-03 1.20E-02 1.02E-02 1.85E-02
5.0 4.85E-03 4.85E-03 9. 30E-03 9. 30E-03
6.0 3.54E-03 3 .54E-03 7.71E-03 7.71E-03
7.0 2.30E-03 2.30E-03 6.22E-03 6.22E-03
8.0 1.41E-03 1.41E-03 3. 36E-03 3. 36E-03
9.0 9.80E-04 9. 80E-04 1.81E-03 1.81E-03

Background Moderate High Extreme Background Moderate High Extreme

Stratosphere Volcanic Volcanic Volcanic Stratospheric Volcanic Volcanic Volcanic
9.0 9. 80E-04 9. 80E-04 9. 80E-04 9. 80E-04 1.81E-03 1.81E-03 1.81E-03 1.81E-03
10.0 7.87E-04 1.38E-03 1.70E-03 1.70E-03 1. 14E-03 1.85E-03 1.85E-03 1.85E-03
11.0 7. 14E-04 1.79E-03 2.31E-03 2.31E-03 8.00E-04 2. IIE-03 2.1 lE-03 2. I IE-03
12.0 6.63E-04 2.21E-03 3. 25E-03 3. 25E-03 6.42E-04 2.45E-03 2.45E-03 2.45E-03
13.0 6. 22E-04 2.75E-03 4.52E-03 4.52E-03 5. 17E-04 2. 80E-03 2.80E-03 2.80E-03
14.0 6.45E-04 2. 89E-03 6.40E-03 6.40E-03 4.43E-04 2.89E-03 3.61E-03 3.61E-03
15.0 6.43E-04 2.92E-03 7. 80E-03 1.OIE-02 3 .95E-04 2.92E-03 5.23E-03 5.23E-03
16.0 6.41E-04 2. 74E-03 9.42E-03 2.35E-02 3. 82E-04 2.74E-03 8. 1OE-O3 8. 1OE-O3
17.0 6. OIE-04 2.46E-03 1.07E-02 6. 1OE-O2 4. 25E-04 2.46E-03 1.20E-02 1.27E-02
18.0 5.63E-04 2. 1OE-O3 1. 1OE-O2 1.00E-01 5 .20E-04 2. 1OE-O3 1.52E-02 2.32E-02
19.0 4.92E-04 1.71E-03 8 .60E-03 4.00E-02 5. 82E-04 1.71E-03 1.53E-02 4.85E-02
20.0 4.23E-04 1.35E-03 5. 1OE-O3 9.15E-03 5 .90E-04 1.35E-03 1. 17E-02 1.OOE-Oi
21.0 3 .52E-04 1.09E-03 2.70E-03 3.13E-03 5.03E-04 1.09E-03 7. 1OE-O3 5.50E-02
22.0 2.96E-04 8 .60E-04 1.46E-03 1.46E-03 4. 20E-04 8.60E-04 4.50E-03 6. 1OE-O3
23.0 2.42E-04 6.60E-04 8 .90E-04 8.90E-04 3.00E-04 6. 60E-04 2.40E-03 2.40E-03
24.0 1.90E-04 5.15E-04 5. 80E-04 5. 80E-04 1.98E-04 5. 15E-04 1.28E-03 1.28E-03
25.0 1.50E-04 4. IOE-04 4. 1OE-O4 4. 1OE-O4 1.31E-04 4. 1OE-O4 7.75E-04 7.75E-04
26.0 1. 15E-04 3 .20E-04 3. 20E-04 3. 20E-04 9. OIE-05 3. 20E-04 4.45E-04 4.45E-04
27.0 8.95E-05 2.51E-04 2.51E-04 2.51E-04 6.78E-05 2.51E-04 2.90E-04 2.90E-04
28.0 6.70E-05 2. 1OE-O4 2. 1OE-O4 2. 1OE-O4 5.18E-05 2. 1OE-O4 2.1 OE-O4 2.1 OE-O4
29.0 5. 20E-05 1.24E-04 1.24E-04 1.24E-04 4. 12E-05 1.24E-04 1.24E-04 1.24E-04
30.0 3.32E-05 7 .60E-05 7.60E-05 7.60E-05 3.32E-05 7 .60E-05 7.60E-05 7.60E-05

Transition from Volcanic to

Normal Extreme
Upper Atmos. Normal Extreme Upper Atmos.
30.0 3.32E-05 7.60E-05 7.60E-05 3 .32E-05
35.0 1.65E-05 2.45E-05 7.20E-05 4.25E-05
40.0 8 .00E-06 8.00E-06 6.95E-05 5 .60E-05
45.0 4.02E-06 4.02E-06 6.60E-05 6.60E-05
50.0 2. IOE-06 5.04E-05
55.0 1.09E-d6 3 .40E-05
60.0 5.78E-07 2. 30E-05
65.0 3 .05E-07 1.62E-05
70.0 1.60E-07 1.03E-05
75.0 6.95E-08 6. 70E-06
80.0 2 .90E-08 4. 30E-06
85.0 1.20E-08 2.78E-06
90.0 5. 1OE-O9 1.55E-06
95.0 2.15E-09 8. 30E-07
100.0 9.3OE-10 4. 50E-07

18-18

JACKET 702-058 CHAPTER 18

 u
OPTICAL AND INFRARED PROPERTIES OF THE ATMOSPHERE
Table 18- 10b. Vertical distribution of aerosol scattering (km ‘1) for a wavelength of 0.550 Wm with rural model aerosol boundary layer

Height 50 km 23 km 10 km 5 km 2 km
(km) Met. Range Met. Range Met. Range Met. Range Met. Range
0.0 6.35E-02 1.52E-01 3 .64E-01 7.39E-01 1.86E+O0
1.0 3.98E-02 9.51E-02 3 .64E-01 7.39E-01 1.86E+O0
1.5 3. 13E-02 7 .60E-02 3 .64E-01 7.39E-01 1.86E+O0
2.0 2.49E-02 5. 96E-02 5 .96E-02 5 .96E-02 5 .96E-02

Fall-Winter Profiles Spring-Summer Profiles

50 km Surface 2 to 23 km Surface 50 km Surface 2 to 23 km Surface

Met. Range Met. Range Met. Range Met. Range
2.0 2.49E-02 5 .96E-02 2.49E-02 5.96E-02
3.0 1.09E-02 2.59E-02 1.39E-02 3. 30E-02
4.0 6.13E-03 1. 14E-02 9.72E-03 1.76E-02
5.0 4. 62E-03 4.62E-03 8.86E-03 8.86E-03
6.0 3.37E-03 3. 37E-03 7.35E-03 7.35E-03
7.0 2. 19E-03 2. 19E-03 5.93E-03 5.93E-03
8.0 1.34E-03 1.34E-03 3 .20E-03 3 .20E-03
9.0 9.34E-04 9. 34E-04 1.72E-03 1.72E-03

Background Moderate High Extreme Background Moderate High Extreme

Stratosphere Volcanic Volcanic Volcanic Stratospheric Volcanic Volcanic Volcanic
9.0 9. 34E-04 9.34E-04 9. 34E-04 9. 34E-04 1.72E-03 1.72E-03 1.72E-03 1.72E-03
10.0 7.87E-04 1.31E-03 1.45E-03 1.45E-03 1. 14E-03 1.75E-03 1.58E-03 1.58E-03
11.0 7. 14E-04 1.70E-03 1.98E-03 1.98E-03 8.00E-04 2.00E-03 1.80E-03 1.80E-03
12.0 6.63E-04 2.09E-03 2.78E-03 2.78E-03 6.42E-04 2.32E-03 2. 1OE-O3 2. 1OE-O3
13.0 6. 22E-04 2.61E-03 3.87E-03 3.87E-03 5. 17E-04 2.65E-03 2.39E-03 2.39E-03
14.0 6.45E-04 2.74E-03 5.47E-03 5.47E-03 4.43E-04 2.74E-03 3.09E-03 3.09E-03
15.0 6.43E-04 2.77E-03 6.67E-03 8.64E-03 3.95E-04 2.77E-03 4.47E-03 4.47E-03
16.0 6.41E-04 2 .60E-03 8.06E-03 2. OIE-02 3. 82E-04 2.60E-03 6.93E-03 6.93E-03
17.0 6. OIE-04 2.33E-03 9.15E-03 5 .22E-02 4.25E-04 2. 33E-03 1.03E-02 1.09E-02
18.0 5.63E-04 1.99E-03 9.41E-03 8.55E-02 5 .20E-04 1.99E-03 1.30E-02 1.98E-02
19.0 4.92E-04 1.62E-03 7. 36E-03 8.42E-02 5 .82E-04 1.62E-03 1.31E-02 4.15E-02
20.0 4.23E-04 1.28E-03 4. 36E-03 7.83E-03 5.90E-04 1.28E-03 1.00E-02 8.55E-02
21.0 3.52E-04 1.03E-03 2.31E-03 2.68E-03 5.03E-04 1.03E-03 6.07E-03 4.70E-02
22.0 2.96E-04 8. 15E-04 1.25E-03 1.25E-03 4.20E-04 8.15E-04 3.85E-03 5.22E-03
23.0 2.42E-04 6.25E-04 7.61E-04 7.61E-04 3.00E-04 6. 25E-04 2.05E-03 2.05E-03
24.0 1.90E-04 4.88E-04 4.96E-04 4.96E-04 1.98E-04 4. 88E-04 1.09E-03 1.09E-03
25.0 1.50E-04 3.88E-04 3.51E-04 3.51E-04 1.31E-04 3. 88E-04 6.63E-04 6.63E-04
26.0 1. 15E-04 3 .03E-04 2.74E-04 2. 74E-04 9. OIE-05 3 .03E-04 3.81E-04 3.81E-04
27.0 8 .95E-05 2. 38E-04 2.15E-04 2.15E-04 6.78E-05 2.38E-04 2.48E-04 2.48E-04
28.0 6. 70E-05 1.99E-04 1.80E-04 1.80E-04 5.18E-05 1.99E-04 1.80E-04 1.80E-04
29.0 5 .20E-05 1. 17E-04 1.06E-04 1.06E-04 4. 12E-05 1. 17E-04 1.06E-04 1.06E-04
30.0 3. 32E-05 7. 20E-05 7.20E-05 7 .20E-05 3.32E-05 7 .20E-05 7.20E-05 7.20E-05

Transition from Volcanic to

Normal Extreme
Upper Atmos. Normal Extreme Upper Atmos.
30.0 3. 32E-05 7. 20E-05 6.50E-05 2.84E-05
35.0 1.64E-05 2.44E-05 7. 16E-05 4.23E-05
40.0 7.96E-06 7.96E-06 6.91E-05 5. 57E-05
45.0 4.00E-06 4.00E-06 6.57E-05 6.57E-05
50.0 2.09E-06 5. OIE-05
55.0 1.08E-06 3. 38E-05
60.0 5.75E-07 2.29E-05
65.0 3.03E-07 1.61E-05
70.0 1.59E-07 1.02E-05
75.0 6.91E-08 6.67E-06
80.0 2. 89E-08 4.28E-06
85.0 1. 19E-08 2.77E-06
90.0 5 .07E-09 1.54E-06
95.0 2. 14E-09 8 .26E-07
100.0 9.25E-10 4.48E-07

18-19
 CHAPTER 18
Table 18- IOC. Vertical distribution of aerosol absorption (km ‘1) for a wavelength of 0550 ~m with rural m{)del aer{)s{)l htlLInd:i~ layer

Height 50 km 23 km 10 km 5 km 2 km
(km) Met. Range Met. Range Met. Range Met. Range Met. Range
0.0 2.70E-03 6.45E-03 1.55E-02 3. 14E-02 7.92E-02
I .0 1.69E-03 4.04E-03 1.55E-02 3. 14E-02 7.92E-02
1.5 1.33E-03 3.23E-03 1.55E-02 3. 14E-02 7.92E-02
2.0 1.06E-03 2,53E-03 2.53E-03 2.53E-03 2.53E-03
Fall-Winter Profiles Spring-Summer Profiles

50 km Surface 2 to 23 km Surface 50 km Surface 2 to 23 km Surface

Met. Range Met. Range Met. Range Met. Range
2.0 1.06E-03 2.53E-03 1.06E-03 2.53E-03
3.0 5.38E-04 1.28E-03 6.89E-04 1.63E-04
4.0 3.03E-04 5.66E-04 4.81E-04 8.73E-04
5.0 2.29E-04 2.29E-04 4.39E-04 4.39E-04
6.0 1.67E-04 1.67E-04 3.64E-04 3.64E-04
7.0 I.08E-04 I.08E-04 2.93E-04 2.93E-04
8.0 6.65E-05 6.65E-05 1.58E-04 1.58E-04
9.0 4.62E-05 4.62E-05 8.54E-05 8.54E-05
Background Moderate High Extreme Background Moderate High Extreme
Stratosphere Volcanic Volcanic Volcanic Stratospheric Volcanic Volcanic Volcanic
9.0 4.62E-05 4,62E-05 4.62E-05 4.62E-05 8.54E-05 8.54E-05 8.54E-05 8.54E-05
10.0 4.66E-I1 7.27E-05 2.46E-04 2.46E-04 6.75E-I1 9.75E-05 2.68E-04 2.68E-04
11.0 4.22E-II 9.44E-05 3.34E-04 3.34E-04 4.73E-I1 1.IIE-04 3.05E-04 3,05E-04
12.O 3.92E-II I I6E-04 4.70E-04 4,70E-04 3.80E-11 1.29E-04 3.54E-04 3.54E-04
13.0 3.68E-II 1.45E-04 6.54E-04 6.54E-04 3.06E-11 1.48E-04 4.05E-04 4.05E-04
14.O 3.82E-II 1.52E-04 9.26E-04 9.26E-04 2.62E-11 1.52E-04 5.22E-04 5.22E-04
15.0 3.80E-II 1.54E-04 1.13E-03 1.46E-03 2.34E-II 1.54E-04 7.57E-04 7.57E-04
16.0 3.79E-I1 1.44E-04 1.36E-03 3.40E-03 2.26E-I1 1.44E-04 1.17E-03 1,17E-03
17.0 3.56E-II 1.30E-04 1.55E-03 8.83E-03 2.51E-11 1.30E-04 1.74E-03 1.84E-03
18.0 3.33E-11 1.IIE-04 1.59E-03 1.45E-02 3.08
E-II l.l IE-04 2. 20E-03 3.36E-03
19.0 2,91E-11 9. OIE-05 1.24E-03 5. 79E-03 3.44E-I 1 9. OIE-05 2.21 E-03 7.02E-03
20.0 2.50E-I 1 7. 12E-05 7,38E-04 1.32E-03 3.49E-I 1 7. 12E-05 1.69E-03 1.45E-02
21.0 2.08 E-I I 5.75E-05 3.91E-04 4.53E-04 2.98 E-I I 5.75E-05 1.03E-03 7.96E-03
22.0 1.75E-11 4.53E-05 2. I IE-04 2.llE-04 2.49E-1 1 4.53E-05 6.51E-04 8.83E-04
23.0 1.43 E-I I 3.48E-05 1.29E-04 1.29E-04 1.78E-11 3.48E-05 3.47E-04 3 .47E-04
24.0 1.12E-11 2.71E-05 8.39E-05 8. 39E-05 1.17E-11 2.71E-05 1.85E-04 1.85E-04
25.0 8.88E-12 2. 16E-05 5.93E-05 5.93E-05 7.75E-12 2, 16E-05 1.12E-04 1.12E-04
26.0 6.80E-12 1.69E-05 4.63E-05 4, 63E-05 5.33E-12 1.69E-05 6.44E-05 6.44E-05
27.0 5.30E-12 1.32E-05 3.63E-05 3.63E-05 4. OIE-12 1.32E-05 4.20E-05 4. 20E-05
28.0 3.96E-12 l.l IE-05 3 .04E-05 3 .04E-05 3.07E-12 1. IIE-05 3.04E-05 3.04E-05
29.0 3.08E-12 6,54E-06 1.79E-05 1.79E-05 2,44E-12 6.54E-06 1.79E-05 1,79E-05
30.0 1.96E-12 4. OIE-06 4. OIE-06 4. OIE-06 1.96E-12 4. OIE-06 4. OIE-06 4. OIE-06

Transition from Volcanic to

Normal Extreme
Upper Atmos. Normal Extreme Upper Atmos.

35.0 8.35E-08 1.24E-07 3 .64E-07 2. 15E-07

40.0 4.05E-08 4.05E-08 3.52E-07 2,83E-07
45.0 2.03E-08 2.03E-08 3. 34E-07 3.34E-07
50.0 1.06E-08 2.55E-07
55.0 5.51E-09 1.72E-07
60.0 2.92E-09 1. 16E-07
65.0 1.54E-09 8. 20E-08
70,0 8.1 OE-10 5.21E-08
“75.0 3.52E-10 3.39E-08
80.0 1.47E-10 2.18E-08
85.0 6.07 E-I 1 1.41E-08
.
90.0 2.58 E-I I 7. 84E-09
95.0 1.09E-11 4. 20E-09
100.0 4.71 E-12 2.28E-09
*

18-20
 a
OPTICAL AND INFRARED PROPERTIES OF THE ATMOSPHERE
Rti=50Z URBFIN. 20.000 PRRT/CM3
Rtt=50Z RURRL . 15.000 P9RT/cf13

10° -~
: ‘oo-~

z
w

II ‘ ,, ‘..

\
. ..,.” [/
: .. \
. ..
b
!

EX11NCTION
6Cfll
TERIN0
......... OBSORPTlON

10-4 I I
~ 0.4 I I t
1 ~’o-l’ ‘ “’’”;
10-1 10° 101 1 12 10° 10’ 102

WAVELENGTH (pm) WAVELENGTH ( pm)

Figure 18-16. Attenuation coefficients vs wavelength for the urban aerosol

Figure 18-14 Attenuation coefficients vs wavelength for the rural aerosol
model at 5070 relative humidity.
model at 5070 relative humidity.

URBflN MODELS. 20.000 PRRT/Cf13

RLIR6L l10DfL6 15.000 PflRT/CnS
“Y
1:’
4
10’ ~
J 1
4

1
10° - .-

1 “-l --

~ ~-z --

u
1- t-
x X
LIJ w
‘
10-3~’o-l”
f I
10-31 I
10° 101 02 ~’o-l : ‘ ,0 , ,,1
. .

WAVELENGTH ( /im ) WAVELENGTH (pm)

Figure 18-15 Extinction coefficients vs wavelength for the rural aerosol Figure 18-17 Extinction coefficients vs wavelength for the urban aerosol
model for different relative humidhies and constant number model for different relative humidities and constant number
density of particles.
density of particles.

18-21
CHAPTER 18
Rtl.50% MRRTTIME, 4000 PRRT/Cf13 RH=50:! TROPO. , 5000 PHRT/CH3

.-.
10° ~

:,.-4%02
\\, :
‘. ...‘ ‘
‘. .. ,

\ ,’1,
1
1
1o-1
10°
~-1’
10° 101 1T2
.

WAVELENGTH (pm )
WAVELENGTH (pm )
Figure 18-20. Attenuation coefficients vs wavelength for the tropsphenc
Figure 18-18. Attenuation coefficients vs wavelength for the maritime
aerosol model at 50% relative humidity.
aerosol model at 50% relative humidity.
TROPO. f100CLS. S000 PRRT/CnS
tl~RITItlE l100EL6. 4000 PRRT/Ct19

10° ~
101
i t
;

RELATIVE ‘
1
HUMIDITY

u
*
x
u
10-3‘ 1 1 ~ 0-4 I I
1
1
4
~~”+ 10° 101 102 ~’o-l’
10° 101 102
WAVELENGTH ( ~m ) WAVELENGTH (pm )

Figure 18-19. Extinction coefficients vs wavelength for the maritime aer- Figure 18-21. Extinction coefficients vs wavelength for the tropospheric
osol mtiel for the different relative humidities and constant aerosol model for different relative humidities and constant
number density of particles. number density of particles.

18-22
 w
OPTICAL AND INFRARED PROPERTIES OF THE ATMOSPHERE
RURnL 0.80.95 .C99Z Rli
STRFITOSPHERIC 8EROSOL MOOELS
~ ~-1 + 1 .0’
FRESH VOLCANIC
------ AGED VOLCANIC
. . . . . . . .. . . BACKGROUND
STRATOSPHERIC CJ
m 0.8- -
—

---
+.
\ 0.6- -
\
\
.... ..O
\

...
. ..
\

\
a
“.. \ u
“.. \ m 0.4 --
\
4 “.. \
z ‘.
\
o “..
\
$ \
“. \
1- 10-4 “.
$A\ z

m
\
u “.
“. \ 2: \ u
-z/

0“21
“. .*. \ l\\- n
:: :“””..2+:
:: ‘ ~’
:

1
1- :. :,
. .. : :$
x ::
w “: ; : ::
:.:
.,+: 1 . ..’ :
lo-s I t 1--
0.0’ ~’o-l’ 1 1 <
~’o-l’
10° 101 10
10° 101 02
WAVELENGTH (pm ) WAVELENGTH(pm)
Figure 18-22 Extinction coefficients for the different stratospheric aerosol
models (background stratospheric, volcanic, and fresh vol- Figure 18-23. Single scattering albedo for the mral aerosol model for
canic). The extinction coefficients have been normalized to different relative humidities.
valves around wok levels for these models (see also Figure
18-13). UREaN O .80.9S .C99L Rfl

l.Oi
.~.—._
./ I

I
‘Y
relative humidities between O% and 9990 where the number /
‘“.
.. . . . . . . . . . . .
density is fixed at 15000 particles/cm3, which corresponds ... . ..<
..- ...
to a meteorological range of about 25 km for the dry aerosols ....
~ 0.8 ---
and about 5 km at 99% relative humidity. / \“

\
u / \
The attenuation coefficients for the urban aerosol model m
-1 \
as a function of wavelength are shown in Figures 18-16 and u \
18-17 for different humidities. The corresponding results I
for the maritime and tropospheric aerosol models are shown
in Figures 18-18 and 18-19 and Figures 18-20 and 18-21,
respectively.
The extinction coefficients for the background strato-
?’ \
+

spheric and volcanic aerosol models are shown in Figure

18-22. Frequently the ratio of the scattering coefficient to
total extinctiuh k,/L (single scatter albedo) is used as a
measure for the relative contribution of scattering and ab-
sorption to the total extinction. Figures 18-23 through 18-
27 give the single scatter albedo for the aerosol models
discussed above. Notice the rapid change in the relative
importance of aerosol scattering and absorption in the mid-
I I
dle infrared spectral region. This is due to two factors, an 0.0’~’o-l’ I

increase in the absorption because of an increase in the 10° 10’ 102

imaginary part of the refractive index in the infrared, and WAVELENGTH (pm)
a decrease in the contribution from scattering as the wave-
length becomes significantly larger than the size of the aer- Figure 18-24. Single scattering albedo for the urban aerosol model tor
0s01s different relative humidities.

18-23
 CHAPTER 18

—...
,,.—
flRRINE 0.60.9 S.C99Z RII uPPER ATMOSPHERE AEROSOLS

,.~..’ . . .. .. ..

..”””””

“
,.,..
/ .

/
/ ;..’” \
\
I
\
l\

i
I +
~\./”
.
. “’”\
..”
.$/ -
““.. .
..
I
~;:/ ‘; I

\ ,1

O.OJ
,l\l
BKGRD. SIRATO
\ AGED VOLCANIC ,1 ‘,
. ... . . . .. FRESH VDLCANI
\ 1 ‘, ““
‘._ ,
1
1
1o-1
101 102 1 o“ ; 0’
,~-1
10° 13’

WAVELENGTH (~m) WAVELENGTH ( ~m )

Figure 18-25. Single scattering albedo for the maritime aerosol model for Figure 18-27. Single scattering albedo for the different stratospheric aer-
different relative humidities. osol models versus wavelength.

TROPO 0.80.95 .499Z Rli

There are also some noticeable differences between the

angular scattering functions of these various aerosol models.
Figures 18-28 a, b and 18-29 a, b give examples for two
wavelengths 1.06 and 10.6 pm. Both figures give the phase
function, defined such that the integral from 0° to 180°
becomes 1. For many purposes (see Section 18.5.1) a con-
venient one-parameter representation for the angular de-
pendence of scattering by aerosols is the asymmetry param-
eter g. It is defined as the cosine weighted average of the
phase function P(6):

+1

g=
J COS8 P(O) d(cos 6)
-’+, (18.22)
P(6) d(cos 6)
J –1

where P(6) is the differential probability y of scattering at an

02 REL. ““”.
angle 8. The asymmetry factor for the various aerosol models
-----
SOZ RCL. HUN.
. . . . . . . . . 352 RCL. IUJfl. is shown in Figures 18-30 to 18-34.
—. —.- 992 *EL. HUM.
Any quantitative treatment of light scattering must in-
0. o
1

clude the polarization of light since in general light becomes

~o-l’
10° 10’ 12
polarized in the scattering process, and the angular distri-
WAVELENGTH (p) bution of scattered light depends on the state of polarization
Figure 18-26. Single scattering albedo for the tropospheric aerosol model
of the light incident on a particle. The degree of polarization
for different relative humidities. P can vary between – 1 and + 1 (see Equation (18.18)).

18-24

J1O JACKET 702-058 CHAPTER 18

 OPTICAL AND INFRARED PROPERTIES OF THE ATMOSPHERE
BOUNOARY LAYER AEROSOL MODELS 1.061 (~) SOUNOARY LAYER AEROSOL mOELS 10.6 (pm)
, ~2 , * . . . . . . . . . . .

— RURAL RURAL
A .---- MARINE ----- MARINE
UJ
. . . . . . . .. . URBAN ..... .... URBAN

kl
10’
----- TROPO.
5
A
1-
0

,-, .-. -.%.

L

w t; ’-..,
w ,~------ “
< -, -. _.-. -
= 10-’.: ‘$.
-
& ~%(
..
..,.
n
w
N
; 10-2.
.-
‘.
~..
-----
.
~:~,,::=~o
<
E
w
0
z ,0-3
+
o 30 60 90 120 150 180 o 30 60 90
SCATTERING ANGLE SCATTERING ANGLE
(a) (a)
UPPER ATMOSPHERIC AEROSOLS 1.06 (pm) UPPER ATMOSPHERIC AEROSOLS 10.6 (
- ,.l _ ,.O

_BKGDSTRAT. .
&
.i --___ AGED VOLC. &
m -, m
. \:., . . . . . . . . . . FRESH VOLC”
m
z
–.–.- METEORDUST :-
o ....... \,, —!--- METEORDUST
..
1-
0
z
3
u

Ln
<

;,o~:,: ;:jo
4
a
~ ,0-2

o 30 60 90
SCATTERING ANGLE SCATTERING ANGLE
(b) (b)
Figure 18-28. (a) Angular scattering functions of low altitude aerosol Figure 18-29. (a) Angular scattering functions of low altitude aerosol
models at 1.06 pm. modelsat 10.6pm.
(b) Angular scattering functions of high altitude aerosoi (b) Angular scattering functions of high altitude aerosol
modelsat 1.06 pm. modelsat 10.6 ~m.

Figure 18-35 gives examples of the angular variation of i, acteristics make modeling of these particles much more
and i2. The polarization of light scattering is dependent on realistic than for aerosols. Also the formation processes and
the particle size and the wavelength of the incident light. life cycle for these particles are better understood than those
for aerosols. For additional information see Chapter 16. A
survey on fog properties was compiled by Stewart and Es-
18.2.2 Cloud, Fog, and senwanger [1982].
Precipitation Particles
18.2.2.1 Particle Types, Water-Ice Refractive Index,
Cloud, fog, and precipitation particles are discussed sep- Particle Shapes. Liquid, fog, and cloud droplets are the
arately because of their rather unique and special physical result of condensation of water vapor on condensation nuclei
characteristics and life cycle that distinguish them from haze (aerosol particles). This condensation occurs at relative hu-
or aerosol particles. First, their composition is uniform, midities near 10070 and is dependent on the physicoehemical
either water or ice, or under some circumstances a com- properties of the condensation nuclei. However, once the
bination of both. Second, in liquid phase these particles are condensation has occurred the effect of the condensation
drop-shaped, that is, close to spherical. These two char- nucleus on the total droplet properties becomes negligible.

18-25
CHAPTER 18
RUR~L 0.90.95 .C99% RM HRRINE 0.60.95 .C992 RM

l“o~

0.8- -
-’
‘.
—”----
—----
— .<.:
. . . . . . . . . . . . . ., . . ...>.
-
~ ..”

$
~>, ,
~,,

=\
.
$.:j .....”’..
~!\\
\

;.
\~:.
.

Y
“\
\
,’. “\
\““.... “\
0.6- - \ “. “\
\ \ ““...
\ ..
\ 1“
\ “\
\
0.4- -

0% REL . UUn .
---- - 602 REL Wll .
0.2- -
......... 9s/ *EL. IuJm.
—. —.- 99z RCL liUll .

i
0.0’ 1 1
0.0’ ~’o-l” I i

~’o-l’
10° 10’ 12 10° 10’ 12
WAVELENGTH (pm) WAVELENGTH (pm )

Flgure 18-30. Asymmetry parameter of mral aerosol model of different Figure 18-32. Asymmetry parameter of maritime aerosol model of dif-
relative humidities. fer;nt rela~iv~ hunlidities.

TROrO 0.60.95 .C99X Rli

URISQN 0.60.95.4897 RH

m
,.

H ~-1’
1 I
4
0.0’ ~’o-l” 1 1

10° 101 102

10° 101
WAVELENGTH @m)
wAVELENGTH (pm)
Figure 15-33. Asymmetry parameter of tropospheric aerosol mudc ! n{
Figure 18-31. Asymmetry parameter of urban aerosol model of different different relative humidities,
relative humidities.

18-26
 d

OPTICAL AND INFRARED PROPERTIES OF THE ATMOSPHERE

UPPER ATMOSPHERE AERCSOLS The concentration and size distribution of fog and cloud
1.0! particles depend in a complex way on the existing mete-
orological conditions. If fog or cloud droplets become so
large that atmospheric buoyancy can no longer balance the

~
w
0.8-
‘ -
“\.. E
\
.
,.,
gravitational forces, they will fall out as rain.
At temperatures sufficiently below freezing, ice particles
will form. Depending on the conditions, they may be single
+ ice crystals such as in ice fog or high altitude cirrus clouds,
w amorphous ice pellets as in sleet or hail, or complex shaped
z
snowflakes. If ice particles fall from cold air into warmer
z 0.6- -
< air, their surface may be coated with a water skin. The
a refractive index of water at optical and infrared wavelengths
> is well known [Hale and Querry, 1973; Downing and Wil-
m liams, 1975]; see Table 18-9.
~ o.4- - The complex refractive index of ice in Table 18-9 was de-
x rived from several sources [Schaaf and Williams, 1973; Irvine
z
> and Pollack, 1967; Grenfell and Perovich, 198 1; Hobbs, 1974;
m and Warren, 1984]. The detailed procedures were as follows:
< 0.2- - A. For the real part of the refractive index:
1. 0.2+.4 pm: Warren [1984] using the refractive
index of water for guidance.
2. 0.4-0.7 pm: Table 3.1 in Hobbs [1974].
I I 3. 0.7-2.5 pm: interpolation between Hobbs [ 1974]
0.0
,- ~-1’ 10’ )2
10° data at 0.7 pm and the value of Schaaf and Wil-
WAVELENGTH (pm ) liams [ 1973] at 2.5 ~m using a two term disper-
sion equation follows Irvine and Pollack’s [ 1967]
data between 0.95 and 1.3 ~m but deviates for
Figure 18-34. Asymmetry parameter of upper atmosphere aerosol models
of different relative humidities. longer wavelengths.
~. ..–__ ,

70-7. ,,, ,,---------------

lx ‘
L
a= 0,1

10-1 ------ _’
.-.-, --------
------ --l

.. . I
~~ ~=,

,01 _ I

‘ ... ,,.m=4
I
0“
n = 155
I
30”
I
60°
I
90”
I
120” 150”
I
180” 0“
1 n = 1,59-0,66j
I
joo
I
60°
[
90”
‘“[
120”
!
150” 180”
—v

Figure 18-35, Angular dependence of i, ---, iz... and (i, + i*)/2— for different aerosol particles. n = complex refractive index: a = si7e parameters.

18-27
CHAPTER 18
4. 2.5 to 25 pm: the measurements of Schaaf and wavelengths 0.2-0.4 pm and 2540 pm. For most of the
Williams [ 1973], since more accurate than any tabulated data (0.4–25 pm) the differences are the order of
available at the time of Irvine and Pollacks’ [1%7] experimental uncertainties in the refractive index. For wave-
survey. lengths outside this range, see Warren [1984].
5. 2540 pm: Warren [1984]. Cloud and rain drops can be considered more or less
B. For the imaginary part of the refractive index: perfect spheres. The shape of single ice crystals, although
1. 0.2-0.4 pm: Warren [1984] variable depending on conditions, is fairly well known.
2. 0.4-0.95 ~m: Grenfell and Perovich’s [ 198 1] meas- Under most conditions, simple snow crystals have an ap-
urements. proximate hexagonal symmetry; however, size and shape
3. 0.95–1 .40 pm: a weighted average between (habit) are influenced by the temperature and humidity con-
Grenfell and Perovich [198 1] and Irvine and Pol- ditions extant during their formation and growth. The stan-
lack’s [1967] survey. dard classification scheme is that of Magono and Lee [ 1966],
4. 1.40–2.7 ~m: Irvine and Pollack’s [1967] values who describe 80 crystal types and combinations. Further
were used since these were based on transmission complicating the picture is the essentially infinite variety of
measurements that are more accurate than the agglomerates (what most people refer to as “flakes”) that
reflectance measurements of Schaaf and Williams can exist, and the fact that any of the basic crystals can be
for weak absorption. modified by the attachment of super-cooled drops in a pro-
5. 2.7–25 pm: Schaaf and Williams [1973]. cess called riming. For a more detailed discussion of crystal
6. 2540 pm: Warren [1984] formation and habit, see Chapter 16.
Since the original compilation of Table 18-9, a review
of the optical constants of ice has been completed by Warren 18.2.2.2 Size Distributions. There have been extensive
[1984]. His values have replaced the preliminary ones for measurements of the the size of fog and cloud droplets.

22 JAN 1400 hrs

1 1530
1
1800
(
L
.

8 FEB 100lhrs 1201 I 400

,
o 3 6 90 3 6 9 0 3 6 9

MAXIMUM DIMENSION (mm)

Figure 18-36 PtirtIclc size distributions fur selected times during snowstorms on 22 January and 8 February 1981 near Burlington, Vermont. Each
distributimr is for a 5-minute sample period taken with an optical array probe [Berger, 1983].

18-28
 OPTICAL AND INFRARED PROPERTIES OF THE ATMOSPHERE
Sufficient data exist to differentiate between drop size dis- 18.2.2.3 Cloud–Precipitation Models. Cloud and tog
tributions for different types of fogs and clouds. Drop sizes drop size distributi{~ns have been m(~deled by Several re-
from I to 100 ~m can occur: the smaller sizes (I–IO ~m searchers. Most commonly used are the models given by
range) are found in fogs and stratus type clouds, the larger Deirmendjian [ 1969] for cumulus clouds and Silverman and
sizes in convective clouds with strong updraft. Ice crystals Sprague [ 1970] for a number of fog types and several types
inclouds (or ice fog) may be as small as a few micrometers of clouds. Table 18-1 I is based on Silverman”s compilation,
although the size of cirrus cloud particles is more typically The distribution functions for models I– 12 are shown in
of the order of 0.5 to 1 pm. Rain drops are distinguished Figures 18-37 through 1X-39. Advective fog is produced by
from cloud drops only by the fact that they have grown the transport of moist air over a colder surface resulting in
large (heavy) enough to fall to the surface of the earth (that the cooling of the surface layers below their air dew points
is, to precipitate out). Raindrops maybe as Iargeas 5 mm, with condensation taking place in the t’orm of fog. Most sea
larger drops have an increasing tendency to break up. Snow fogs are advective. Both the size r~nge of particles and the
flakes present an interesting problem with regard to optical liquid water contents are large (see M(>dels 1 and 2, Table
effects because of their extremely complex shapes and di- 18-1 I). Radiative fog is produced when stagnant moist air
mensions. Various data (such as Lo and Passarelli [ 1982]; becomes progressively cooler during the night due to rJ-
Berger [ 1983]; and Gray and Male 1198 11) suggest that a diative cooling. This type of fog has both a small size range
“typical” snowfall consists of particles rdnging from single of particles and a small liquid water c(~ntent.
crystals O. 1 mm ( I00 pm) or smaller to agglomerates several Cloud droplet size distributions are more difficult to
millimeters in diameter. model. However. the optical thickness ot most clouds is so
There is little information on the exact nature of the size large that their transmittance at visible and infrared wave-
distributions. Lo and Passarelli [ 1982] assumed a simple lengths becomes completely diffuse due to multiple scat-
exponential form and found average diameters of 0.5 to 2 tering (see Section 18.5. l).
mm for several aircrdft measurements using optical array The size distribution of rain drops is best expressed in
probes. Ground-level measurements with aspirated optical terms of rain rate, since this is the quantity usually measured
array probes by Berger [ 1983] and Formvar replicating data by the weather services. This formulation goes back to ex-
of Koh and O’Brien [ 19821 show a wide range of size perimental work by Laws and Parsons 11943] and Marshall
distributions; typical examples are shown in Figure 18-36. and Palmer [ 1948]. According to their work the number of
The reasons for this wide variation are not clearly under- drops of diameter D can be expressed as
stood, though ongoing modeling work by Martinez-Sanches
N,, = N<,exp( – a . D)
et al. [ 1983] is beginning to unravel some of the causes.
The majority of the distributions peak between 0.2 and 0.5 where N,, = 0.08 cm ‘ (18.23)
mm. The area-weighted averdge radius, the size measure
and a = 41 R (l~lcm-l
with the most influence on extinction, is generally in the
range 0.1 to 0.3 mm. with R being the rain rate in mm/hr.

Table 18-11. Cloud models: drop size distributions and liquid water contents.

n(r) = a r’ exp [ – br]

Model
Cloud Type Number x b No (cm-s) a M(.g/m7)

Heavy Advection Fog I “3 0.3 20 o,~7 0.37

Moderate Radiation Fog 2 6 3.0 200 607.5 0.02
Cumulus 3 3 0.5 250 ~, 604 1.00
Altostratus 4 5 1.11 400 6,268 0.41
Stratocumulus 5 5 0.8 200 0.4369 0.55
Nimbostratus 6 1 0.333 100 11.089 0.27
Stratus 7 3 0.667 250 8.247 0.42
Stratus 8 2 0.6 250 27.00 0.29
Stratus-Stratocumulus 9 2 0.75 250 52.734 0.15
Stratocumulus 10 2 0.5 150 9.375 0.30
Nimbostratus 11 2 0.425 200 7.676 0.65
Cumulus-Cumulus Congestus 12 2 0.328 80 1.4115 0.57
M y‘ Llquld water content
N,, = Total number of particles per unit volume

18-29
CHAPTER 18
FOG rIODELS RFGL CLOUD f100ELS

‘“’~

In
Cumabs 3
----- tiTsl RuIIIS 4
1 0’ ............ STR6TC-bM1tU5 5
.-. —.—.—
—..—
.
N1M!30$IRA1[S 6
/ ‘. STRAI.C 7

~. ‘.
\. 02
Ii \. 1
i \.
j \
/ i
1
i i
i

1
;
I
.i 1

!
!“
1
i’ \
i

ii’
1

/ .
1 I - - ! O-Y I 1’ : 1 1
1.
\
-1
1’ 10° 10’ 102 10-’ 10° 101
RADIUS (pm) RADIUS (PM)

Figure 18-37. The drop size concentration (cm-s pm-’) to particle drop Figure 18-38. The drop size concentration (cm-’ ~m-’ ) to particle drop
radius (~m) for fog models 1–2 in Table 18-11. radius (pm) for cloud models 3–7 in Table 18-11.

This relationship is strictly empirical and does not explain pendent of wavelength in the visible and infrared. Trans-
the physical significance of the distributions. missometer measurements over the past few years by a
number of investigators [summarized by Seagraves and
18.2.2.4 Extinction, Scattering, Polarization. The at- Ebersole, 1983], however, have found that the measured
tenuation coefficients for the fog models given in Section infrared extinction coefficient varies from 1.0 to 1.45 times
18.2.2 .3arepresented in Figures 18-40 and IS-41. For the that in the visible. generally increasing with wavelength,
droplet concentrations given in Table 18-II (20and200 for and depending somewhat on the type of instrument used.
Model 1 and 2. respectively), the resulting meteorological Mill and Shettle [ 1983] have shown that most of the
ranges (visibility) are 130m and 450m. However, the models apparent wavelength dependence can be explained by in-
can be scaled to risibilities from less than 50m to over 1 strumental effects. The scattering phase functions of par-
km. The extinction coefficients for different cloud models ticles that are large compared to the wavelength are char-
are shown in Figures 18-42 and 18-43. The single scatter acterized by extremely narrow forward lobes due to Fraunhofer
albedo and the asymmetry parameters are given in Figures diffraction. For snow crystals with average areas equivalent
18-44 through 18-49 to describe further the scattering prop- to a sphere of O. I to 0.3 mm radius, this peak is comparable
erties of fogs and clouds. Because of the much larger size to the field-of-view (FOV) of typical transmissomcter re-
of fog and cloud droplets compared to aerosol (haze) par- ceivers. Therefore. for a given transmissometer FOV. the
ticles, the polarization of light scattered by these particles longer the wavelength the less forward-scattered radiation
is very small and has little practical importance. is sensed resulting in a lower “apparent” transmission. The
No theory presently exists to calculate exactly the ex- relationship between the measured transmittance 7’” and the
tinction coefficient for complex shapes such as snow crys- true radiance transmittance T isgiven by lMill and Shettle.
tals. Mie theory in the geometric limit (the minimum Mie 1983]
size parameter at 10 pm wavelength is about 30) suggests
that extinction in snow (and rain) should be nearly inde- T ‘“ = T(1 + D’in T) (18.24)

18-30
 OPTICAL AND INFRARED PROPERTIES OF THE ATMOSPHERE
RFGL CLOUD MODELS FoG t100EL 1
1 !

STRATUS 8
‘- —-- STRAT/STRUICU 9
............ SIRA1OCUHUUS 10
-.-.—.
NIMwsTRflTvS 11
—,.—
cu/cu cc~mslus 12

.:
.’. :
.. .
.. .. ..
. ..
. .
::

, ,
,-1’
10° lC1 :
WAVELENGTH (pm )

Figure 18-40. Attenuation coefficient versus wavelen~th: heavv advection

fog model 1.
WUS (pm)
Figure 18-39, Thedrop size concentration (cm-3 ~m-])to particle drop FOO ROOEL 4
radius (~m) for cloud models 8–12 in Table 18-11.

0 0 Jl[xsin (e’ + +’)] 2f(8,)d0,d@,

D’” =x’
JJ{
00 xsin (e’ + +’) }
(18.25)

e = half angle ofreceiver field of view % :. .’

. . : ... ..
::
. . , ..~ ,@\\\
@ = halfangle of source ::>:
:...>.

x = 2m r/A, theparticle size parameter ::;

: LI \
\

—
-
.
..---
.
r = particle radius .: \,

~m
::
..
k = wavelength : ::
::
J] = Bessel function of the first kind, first order, and ::
cxrl~rlon
scnrrcntffi
f(0) = relative response of thedetector over the FOV. .BSO*P,
IM
..
...
(= 1 for flat detector response)
i

Ic’ nz
10-1 100 1d
This equation assumes that multiple scattering is negligible,
6 and @ are small (less than 50), the phase function is
WAVELENGTH ( ~m )
adequately represented by the Fraunhofer diffraction, and
the size distribution can be represented by a single effective Figure 18-41. Attenuation coefficient vs wavelength; light to moderate
scattering radius. Values of D’ have been tabulated [Shettle radiation fog model 2.

18-31
cHA~ER 18 ,
VAR1OUS CLOUO MOOELS
Stal.; .1985] for various 6, +; and x’s and are summarized
in Table 18-12. To use the takle, calculate the two table
paratneters as follows:
: ‘“’~

P = Max (6, @)/Min (9, +) (18.26) I

v = x QMin (6, +), where 6 and + are in radians,

whwe the expressiori Mux rheans use the lagest of he

argurhents in the parentheses: Similtily, Min means use the
smallest, Max and Min ~, respectively, the l~g~r and
smaller of the two arguments. A comparison of uncorrected
\.,
and diffraction-cbrrected measu~ trarismittances is shown
in Figures 18-50 amd 18-51. Seagraves [1983] has compared
measured visible and 10.5 Wm~ans~ittances to those cal-
wlated from geomettic optics theo~. The calculations as-
sumed a modified gamma size distribution of particle sizes
between 84 and 761 km radius based on replicator data,
~nd used the airborne sfiop mass concentrations measured
by CRREL Jsee Lacombe, 1983]. These studies strongly
suggest that the extinction c~fficient in.snow is very nearly
wavelength independent in the visible and infrared, though
particle sha~ may have a small effect below the resolution 1o-’l~ r [ 1
1 4
of the measurement precision ~d ,cannot be ruled out. 10° 10’ 102 103
Studies by a number ot investigators have addressed the wAVELENGTH (pm)
question of the relationship ~tw~n snow fall rate and ex-
tinction. Results summarized in Figure 18-52 vary widely Figure 18-43. Extinction coefficients for cloud models 3, 4, 6, 11, 12
(Table 18-11).

STRATUS TYPE CLOUO MODELS

FOG MOOEL ALBEOO
103 T
1.0 \\ \
n

m t 02 __— ----
i--
z —..—. $ —.
WI —. —----

I o’

“\
“\
z I
0 \
I

STRATOCUMULUS 5

3
----- iTRbi~S MbDEL 7
......... STRATUS M9DEL E
-. —,. ST ,SC HOOEL 9
—..— ST3iTOCUHV. US!O

\
\
,.-1 1 1 1 ..
I r
/o-i ‘ “’’” “ “’”’” )3 O.c 1

Ib” 10’ 102 ~-i “’’’”

10° 10’ 102 03
WAVELENGTH ( jm) WAVELENGTH (pm)

FigiIre 18-42. Extinction coefficients for straks cloud models 5,7, 8,9, Figure 18-44. Single scatter albcdo for fog models 1 and 2 (Table 18-
10 (Table 8-11). 11).

18-32
 O~ICAL AND INFRARED PROPERTIES OF THE ATMOSPHFRF
STRATUS TYPE CLOUO MOOELS FOG MOOEL ASYMMETRY PARAMETER

‘“”v
l“o~
-t

0.8
‘1
L I \
\
‘1
0.6 0.6- - I
i I

1 I
,~.

\cl
.. ..’: 0.4- -
\
Li ~il ‘?.:% \.”.). ,..
\
d \,;:.~. \
z STRATOCUMULUS 5 \
; 0.2 , . . . . . . ..- ;;:;: %: ; “\. “~\ 0.2- - \

“..=
—.—.— ST/SC MOOEL 9 \, “it”.

,
STRAr OCUMULUS1O ,.j>

\
“,

1 1 I
0.0’ ,’o-i “’’’”
1

~-l
i 0° i’o’ 102 i 05 10’ 102 )3
WAVELENGTH (pm) WAVELENGTH (pm )

Figure18-45. Single scatter albedo for stratus cloud models $, 7, 8, 9, Figure 18-47. Asymmetry parameters for fog models 1 and 2 qable 18-
10 (Table 18-11). 11).

VARIOUS CLOUO MOOELS

VARIOUS CLOUO MOOELS

\/

w
&
z
; 0.2
IT) \;>. -----
. .. .. . .. .
—.—. -
CUMULUS
AL1OSTRATUS
NIUSOSTRATUS
NIMSOSTRATUS
~OEL
II

4
6
! 1
,,---”.

,/-,

3
\

\
“.-

\
\
\
\
\
\
““”..>
., a 0.2

\
—., — CUMULUS 00 NOEST12
\
\
\ \
! I /
0.0, 1 .
0.0! I I 1

10-’ ,00 102

,,01,,,,,,,,, 10-1 100 Iol 1’02” 103
wAVELENGTH (pm)
WAVELENGTH (pm )

Figure 18-46, Single scatter albedo for cloud models 3, 4, 6, 11, 12 Figure 18-48. Asymmeq parameters for stratus cloud models 5, 7, 8,
(Table 18-11). 9, 10 (Table 18-11).

18-33
‘ CHAPTER 18
S:? A’2S ‘YPE CLOUD MODF. S 1
r SNOW-ONE- A TRANSMITTANCE
—. -—— —
1 .0 t-’— 0,9

0,6

07

u
o
0,6
z
<
+ 0.5
~
z
o 0.4
z
a
a
3.6-
I 1- 0.3

0.2

0.1
I I 1 1 1 1 I , , I
0.4+ o
0 30 60 90 120 150 180 200 240 270

1 TIME IN MINUTES AFTER 1530

>- Figure 18-50. Transmittance in the visible, 3–5 Am and 8– 12 +m bands

during a 31 Januw 1982 snowstom near Burlington, Ver-
0.2 mont. The apparent increasing extinction with wavelength
is typical of measurements in snow [Mill and Shettle, 1983).

1 due perhaps to both measurement error, particularly in the

0.0
/o-i I

“’’” ,,’. O
10’
1 !

102 )3 snow rate data, and to crystal habit or size distribution.

Recent results of Lacombe [1983] show that for a given
WAVELENGTH (pm) airborne snow mass concentration visible extinction de-
creases with increasing riming (Figure 18-53) and appears
Figure 18-49, Asymmetry parameters for cloud models, 3,4,6, 11, 12 to be greater for plate-like crystals (spatial and platie den-
(Table 18-1 1). dritic, broad branched crystals and plates) than for the more
compact types (columns, bullets, side planes, needles and
assemblages of plates).
The most extensive set of snow phase function mea-
surements have been made at Michigan Technological Uni-
versity [Winchester et al., 1981]. A typical example is shown
Table 18-12. Forward scattering correction factors D’ (x, 0, ~) for a in Figure 18-54. These measurements are of high angular
transmissometer with flat receiver field of view, wherep=
max (e, O)/min (13,dr) and v = x – min (e/0), for o and resolution but required one to two hours per phase function,
@ in radians

p 100.0050.00 20.00 10.00 5.00 2.00 1.00

SNOW-ONE- A TRANSMITTANCE
v 0.9 OATE : 31 JAN 82
TIME: 1530-2000
0.8 CORRECTED
0.04 0.016 0.015 0.008 0.004 0.0020.001 0.000 h
0.06 0.025 0.024 0.0160.009 0.0040.0020.001 07
0.10 0.041 0.040 0.0360.023 0.0120.0050.002
u 0.6
0.15 0.061 0.060 0.0590.046 0.0260.011 0.006 u
z
a
0.20 0.080 0.08~ 0.0780.070 0.0450.0190.010 . 05 Iv
,_
L I I

0.30 0.116 0.116 0.1150.113 0.0890.0420.022 z

m 04
z
0.40 0.150 0.150 0.1490.146 0.1330.0720.038 a
e
1- 03
0.60 0.210 0.210 0.2100.208 0.2040.1430.081
1.00 0.304 0.304 0.3040.303 0.2990.2760.189 0.2

1.50 0.378 0.378 0.3780.378 0.3760.3640.310 01

2.00 0.418 0.418 0.4180.418 0.4170.4050.381 . I I I I I 1
“

3.00 0.446 0.446 0.4460.446 0.4460.4400.423 0 30 60 90 120 150 180 210 240 270
4.00 0.455 0.455 0.455 0.455 0.455 0.4520.438 TIMEIN MI NUTE3 AFTER 1530

6.00 0.468 0.468 0.4680.468 0.4680.4680.461 Figure 18-51. Data from Figure 18-44, corrected for forward scatter into
10.00 0.475 0.475 0.475 0.475 0.475 0.475 0.474 the transmissometer field of view, assuming an effective
scattering radius of 0.2 mm. Residual wavelength depen-
20.00 0.476 0.476 0.4760.476 0.4760.4760.476
dence may be due to multiple scattering or shape effects.

18-34
 OPTICAL AND INFRARED PROPERTIES OF THE ATMOSPHERE
EXTINCTION VS SNOWFALL RATE I 1 I 1 I
I I I I I I I I I I I I I I I I
FOR DIFFERENT MODELS

‘~

FALLING SNOW
X - WINCHESTER
x O- AFGL

x
9
xx
x)( )
x
‘xxx
xx
xx
‘Xxox x

x )(Xxx Fx

I I I I I 1 I I I I I I I I I i I I I 1 J
o 90 180

SC ATTERI NG ANGLE
l--t I 1 I 1 I 1 I I I 1 I
0 I 5 6 Figure 18-54. Phase function of falling snow measured in Houghton,
PREiPATION iATE (m4m hi’) Mississippi on 26 January 1981 with a polar nephelometer
(EQUIVALENT LIQUID WATER) [Winchester et al., 1981]. The dominant crystal was a six-
sided dendritic of about 2 mm diameter. Circled points are
Figure 18-52. Models of visible extinction coefficient versus equivalent data from the 31 January 1982 SNOW-ONE-A episode
snow fall rate. Curve R is for rain while the others are for averaged over the 1530–2000 period. Data were taken each
snow asmeastrred byvarious investigators [Mason, 1978]. minute with a fixed-angle nephelometer at 30°, 110°, and
150° scattering angles.

raising the question whether they represent any one snowfall

“type”. Measurements made by AFGL are also shown in
Figure 18-54. They were made simultaneously at three fixed
angles and are shown normalized to the Winchester data at
30°; they show good agreement in relative scattered inten-
sity. An absolute comparison is not possible since Win-
chester’s data are not normalized to total extinction.

18.3 SURFACE REFLECTANCE

AND ALBEDO

a A fraction of a light beam that is incident on the earth’s

* surface or on clouds will be reflected back into the atmo-
In
q sphere where it may undergo further scattering and absorp-
02
b- tion. Surface reflectance consequently affects sky radiance
seen from the ground, the radiance of the atmosphere as
(.) Moderote to Heavy Riming seen from space, and other atmospheric optical quantities
(+) Graupel or “graupelized” (see Section 18.5). The percentage of light reflected from
natural terrain or water surfaces varies with angle of inci-
I I I I
0.4 0.6 dence and look angle and is also a function of the wavelength
C, Airbarne Snaw Concentration (g/m3) of the radiation. Details of these reflection processes, es-
~cially laboratory measurements on surfaces, are often treated
Figure 18-53. Relationship between airborne snow concentration and 0.55
~m extinction for 3 degrees of riming. Data are from SNOW-
by the bidirectional reflectance function (BDRF), while in-
ONE-A. Curves are power law fits within each category. tegral reflection effects—for example the brightness of a

18-35
CHAPTER 18
planet—are usually characterized by “albedo”. For mirror SALT FLAT CORN
surfaces (specular reflection). the incident and reflected beam
and the normal to the surfi~ce at the point of incidence lie
in the same plane: the angle between the normal and the
direction of’ incidence and between the normal and the re-
flected beam are identical.
A brief derivation of the BDRF following Nicodemus
[ 1967] and Leader [ 1979] will now be given. The radiant
power incident on a surface element dA through a beam of
solid angle d(l, (from polar angle 0, azimuth angle +) is
10” 44° 16° 5P
d Pi = NI dA dfl’l , (18.27) SOLAR ELEVATl~

Figure 18-55. Bidirectional reflection function of Bonneville. Utah, Salt

where N, is the radiance (W cm z sr”l) and d(l’, = cos 0, Flat (left) and a com field (right), each for a low and a
dfli is the projected solid angle. The radiant power reflected high solar elevation. F,(Tf’ )~a~,, = albedo for a horizontal
into do, is receiver. Spectral range 0,3 to 3 Vm [Eaton and Dirmhim,
1979].

d P, = dN, dA dfl’, ( 8.28)

w

Thus the ratio of the radiant power is

dP, dN,
— dQ’, = f’ dfl’,, ( 8.29)
d P, = N, do’, ARCTIC SUMMER ICE

.’
(sr 1) is called the bidirectional reflectance
f’ (6,, 4,, 0,, b,)
function (BDRF). This definition is especially useful for
narrow (collimated) illuminating beams (sun. laser). If the
incident radiance comes from all directions of the hemi-
sphere, then Ii
“ 0.2 0.6 1.0 1.4 1.8 2.2 2.6 3.0 3.4 3.8 42
rn
WAVE LENGTH (~m )
271 n,?
N, = f’ N, sin O cos O d Odd. (18.30)
lJ() <) Figure 18-56. Tv~ical s~ctral diffuse reflectance of snow, ice. soil. and
v;~etatio; after Suits [ 1978. Figure 3- 19].

A perfectly diffuse reflector is characterized by a con-

stant f’ in all directions. If such a surface is diffusely ir-
radiated and the reflected radiance in any direction is meas- 0.10 CALCULATED EXTINCTION

ured, the directional reflectance is f~,,f = IT f’ [dimension- ... .. BY PHYTOPLANKTON

‘\. . “.
less ]. Hence, the BDRF of a reflection standard (barium
sulfate) with f,,,,, = 1 is f’ = I/m.
Comparison of incoming and outgoing shortwave ra- R
. . . . . .. S..
diance of the earth’s surFdce or the earth’s atmosphere in
the simplest way is made by a “’flat” receiver turned up (for
downwelling radiation) and down (for upwelling radiation). 0.01 CRATER LAKE
The rdtio of the fluxes, P ~ /P ~ , is often called albedo or
\ \.
diffuse reflectance. GULF OF \
Eaton and Dirmhirn [ 1979]. for example. have made CALIFORNIA \
measur~rnents of BDRFs and albedo of severdl surfaces for
\\
shortwave radiation (0.3–3 pm). Those of a salt flat and a \\
field of corn are shown in Figure 18-55 for two solar ele- \:\\
vat ions. Due to the an isotropy. the nadir values mf’ of the \
BDRF are to be multiplied by a Pdctor F to obtain the albedo. 0.001
0.4 0.5 0.6 0.7
At low solar elevation angles, F is generally >1. but sur-
faces with vertical structure may make F < I at high sun WAVELENGTH (pm)
(when partial shadowing occurs).
Figure 18-57. Nadir reflectance R 0.4-().7 pm of some water bodies.
Spectral reflectance from 0.2 to 4~m for several surface Measured value raken frt)nl Tyler [ 197X ], calculatit)ns fr{]n)
types is gi~en in Figures 18-56 through 18-58 [Suits, 1978]. Suits [ 1978. Figure 3-1241.

18-36
 OPTICAL AND INFRARED PROPERTIES OF THE ATMOSPHERE

1.0 AFGL a computational model, FASCODE (Fast Atmo-

spheric Signature Code), has been developed and made
R WATER SURFACE available to the scientific community. Line-by-line calcu-
lations are discussed in the following section in the general
0.3 context of atmospheric calculations and the FASCODE model
SMOOTH IClough et al., 1981; Smith et al., 1978]. Specific details
of the model are not discussed, but are available in the
ROUGH literature [for example, Clough and Kneizys, 1979; Clough
0.1 et al., 1977]. A general reference covering the topic of
atmospheric molecular absorption that the reader will find
particularly useful is the work by Goody [ 1964].
0.03
18.4.1.1 Line-by-Line Compu@tions (FASCODE). The
0.016 I monochromatic optical depth, k(v), at wavenumber value
o 30 60 90 v(cm-] ), assuming superposition of the contribution from
the individual spectral transitions, is given by
NADIR ANGLE
Figure 18-58. Variation with nadir look angle of specular reflectance R i(v) = ~ 11 –+ e -
e -
““/kT
hcv/kT
of smooth water surface and surface roughened by a bcau- (18.31)
fort 4 [Suits, 1978, Figure 3-1 14].
X ~ W(mi) ~i(T) [f(v,vi) + f(v, – ~,)1
1

where W(m,) (mol/cm2) is the column density for the mo-

Further examples, citations, and references to data files can lecular species, mi, involved in the i’h transition. v, is the
be found in Suits [1978]. Fitch [1981] made a theoretical transition waventtmber and Si(T)( 1/mol/cm~) is the intensity
study of the polarized radiance (at 0.4 and 0.6 pm) from at temperature T (K) appropriate to the FASCODE line
the top of the earth’s atmosphere for different surfaces and shape, f(v, vi) (l/cm-l). We note that W(mi) = p(mi)l for
aerosol conditions. path length 1 and molecular density (p(mi). This line shape
formulation is an extension of that used by Van Vleck and
Huber [ 1977] and discussed by Clough et al. [ 1980; 1983].
Equation ( 18.3 I ) may be conveniently rewritten in terms of
18.4 ATMOSPHERIC TRANSMITTANCE the radiation field dependent term v tanh(hcv/2kT) to obtain
AND RADIANCE
~(~) = v tanh(hcv/2kTi) (18.32)
x ~ W(m,) Si(T) [f(~,vi) + f(~, – vi)]
18.4.1 Monochromatic Radiation.

For many purposes, knowledge of the monochromatic The intensity, ~i, is expressed in terms of the transition
optical depth for a specific optical path in the earth’s at- strength \p?l debye2, as
mosphere is required. The type of calculation necessary to
obtain this information is referred to as a line-by-line cal-
8Tr3 X 10-36
culation since it takes into account the contribution of each Si(T) =
3hc
spectral line to the monochromatic absorption. Conse-
quently, line by line calculations require detailed knowledge
of the spectral line parameters associated with each contri-
buting transition. Such computations, performed over an
x Iw?l 1
(
e-hcE,/kT
+ e - hcv,/kT

)
(18.33)

extended wavenumber interval (of the order of 100 cm”]),

Qv(m,,T)Q~(mi,T)’
may require significant computational effort. Several meth-
ods for performing such computations have been discussed
in the literature. Some are particularly suitable for special- where Ei(cm-l ) is the lower state energy for the transition
ized calculations [Scott and Chedin, 1981; McMillin et al., and Qv(miT) and QR(m, .T) are the vibrational and rotational
1976; Kunde and Maguire. 1973], some make approxi- partition functions for the appropriate molecular species,
mations that in certain cases maybe acceptable [Karp, 1973; mi. In terms of the intensity definition, S,(T), used for the
Mankin 1979], and others are accurate and direct but not AFGL line compilations [Rothman et al., 1983a,b: Mc-
computationally efficient [Susskind and Searl, 1978]. At Clatchey et al., 1973] we have the relationship

18-37
CHAPTER 18
, + ~ - hcvi/kT

Si(T) = Si(T). (18.34)

vi 1 _ ~ h..i/kT 0.8

( ) z 0.6
3 +
\
0.4
The line shape function must satisfy the normalization con- %
5 i
dition, 0.2

x 0.0
-4 -3 -2 -1

J 0
dv [f(~,vi) + f(~, –vi)] = 1, (18.35)
Z(HRLF WIDTH)

As a result, the expression for optical depth, Equation ( 18.32) ‘“o-~

satisfies two important conditions: the Nyquist condition,

. k(v)
J o
dv
v tanh (hcv/2kT)
= ~lp;l
,
(18.36)

that is, the preservation of transition strengths; and the con-

dition that radiation balance be satisfied between emission
Z(HRLF WIDTH)
and absorption for a system in thermal equilibrium.
The line shape function f(v,vi) is dependent on molecular
species, broadening density, and temperature. For colli -
sional broadening in the impact limit, the form factor is
given by

a:
f(V, Vi) = 1 (18.37)
‘iT(V – Vi)* + (~f)2’

where ~i’ (cm-’) is the collision broadened halfwidth at half

maximum (H WHM). It proves convenient for computational
purposes to define a dimensionless parameter z such that
Figure 18-59. (a) The line shape profile appropriate for collisional broad-
V—Vi
ening.
z=—
(18.38) (b) The line shape profile appropriate for veloeity broad-
a; ‘ ening.
(c) The function utilized in FASCODE for the construction
of the Voigt line profile.
Thus, f(v,v,) becomes f(z) = ( l/aY) L(z) where

L(z) = + +. (18.39)
The extensive computational effort required for atmos-
pheric line by line calculations is a result of three principal
The functional dependence of L(z) is shown as function L factors: the large number of spectral lines contributing to
in Figure 18-59a. the absorption, the small sampling interval necessary to
The dependence of the collision broadened halfwidth on model the attenuation at higher altitudes and the slow con-
the number density, n, and temperature, T, of the absorbing vergence to zero of the Lorentz function given in Equation
media is given by the relation, (18.39). The latter consideration has been treated through
the algorithms utilized in FASCODE and the details have
a: (n ,T) = a; (n/n,,)(T/T,,) ‘T (18.40) been made available in the literature. A computational sav-
ings of sixteen is achieved over direct methods for a spectral
where a? is the collision halfwidth at reference number interval of 512 halfwidths.
density n,, defined for 1013 mb and reference temperature For the Doppler regime, in which velocity broadening
T<,(296 K). The quantity XT, characterizing the temperature dominates, the line shape is
dependence of the halfwidth, typically ranges between O
and 0.5, depending on molecular species and temperature. f(VVi) = +(~) ’’2exp[-ln2(~)2]
Classical collision theory gives a value of 0.5 for XT. Values
of XT appropriate to each molecule type are stored in the
program. (18.41)

18-38
 OPTICAL AND INFRARED PROPERTIES OF THE ATMOSPHERE
Letting 0.0 0.2 0.4 0.6 0.8 1.0

V—Vi
z= (18.42)
a? o .95- -

then f(z) = ( l/aP) G(z) where

1/2 o .90
G(z) =
()
: exp [ – ( ln2)(z2)]. (18.43)
a
> 1

o .85
The functional dependence of G(z) is Gaussian and is shown
in Figure 18-59b. The Doppler halfwidth (HWHM) is given
by I
0 .801 , ,I (1.8G
0.0 0.2 0:4 0.6 0.8 1.0

ZETQ
“= :[2(1n2’Rl”2 (18.44)
Figure 18-dO. The constant relating the Voigt width to the sum of the
Dopu]er and collision widths: collision broadened width
where M is the gram molecular weight of the molecular ov;r’ the doman 0.5.
species of the transition and NO is Avogadro’s number.
In the intermediate regime between collisional broad- where z is defined as
ening and velocity broadening, the line shape is obtained
from the convolution of the collisional and Doppler line
shape giving the Voigt line profile. The computation of the z=— v—v{ (18.49)
m“
Voigt line shape using a direct approach is too time con-
suming for general utilization. These direct methods gen-
The function V~(z) is an analytic function having the de-
erally give results to significantly higher precision than the
pendence on z shown in Figure 18-59c. The constants CC(L),
uncertainty in the line parameters and the uncertainty in the
CD(L), and Cv(~) are determined from least square fits of
line shape itself. An effective approximation to the Voigt
V(~,z) to the exact Voigt function. The Voigt line p~files
line shape may be obtained by usihg an extension of the
for four values of ~ and the FASCODE approximati~,
method proposed by Whiting [1968]. A Voigt parameter,
V(~,z), are shown in Figure 18-61. The computational sav-
~, is defined in terms of the collisional and Doppler widths
ings using this approximate approach is on the order of 100
compared with direct methods. The largest error, -3Y0,
(18.45) occurs for ~ = 0.05 at a value of z corresponding to 16
halfwidths from line center at which the function is four
orders of magnitude less than the peak value.
for which ~ = O in the Doppler limit and ~ = 1 in the
collision broadened limit. For ~ = 0.5 the collision and
18.4.1.2 Continuum Absorption. FASCODE includes
Doppler widths are equal. The Voigt width a“ (HWHM) is
continuum contribution from self and foreign density de-
obtained through the relation
pendent water vapor absorption, foreign density dependent
carbon dioxide absorption, and the collision induced band
a v = Av(~) (aD + a“). (18.46)
of nitrogen. A more extensive treatment of the approach
used in FASCODE in which the continua are develo~d
The function Av(~) shown in Figure 18-60 is obtained from
from the wings of strong lines appears in the literature [Clough
the relations given by Kielkopf [ 1973] where
et al., 1980 and 1983]. For atmospheric applications it is
advantageous to express the density dependence in terms of
Av(~) = (~) ~ + (L)* a self and foreign component. The continuum contribution
[ 1/2
to the optical depth, ~c, is given by the expression
1–E2
() 2
+1–(1–C)*
1 ,
(18.47) ~c = W v tatth(hcv/2kT) (18.50)

with E = 0.0990 1n2. An appropriate

V(~,z), is given as
Voigt line shape, x
[
(n,/nO) 6.(v,T) + (n,/n,,) Cf(v,T)
1
,

V(~,z) = CC(L) L (Z) + CD(C) G (z) where W is the column density of the absorbing molecular
+ C,(L) VE(Z), (18.48) species, (n,/no) and (n~/n,,) the number density ratios for the

18-39
CHAPTER 18
100

10-’7 -

, 1 I

3.0’
,
1 1
, ,
, 1

+
t

, , , , t t
-3.0
-i8 -i2 -16 b 1’6 3’2 4~ 64
Z(HRLF-HIIITH)
Z[HRLF-WIDTH)
*Oo t , , , t ,
, 1
, I , , ,
100
i 1

, , , ,
, +
3.0 3.0 ,

m
0
K
0.0” m o.G- ~ :
u
h“
, , , ,
-3.0
-d4 -48 -42 -{6 b 1’6 32 48 64

Z[HRLF-WIDTHI ZIHRI.F-WIDTHI

Figure 18-61. Voigt line shape profiles for representative values of zeta. The solid line represents the correct function and the symbol + the FASCOD 1
approximation.

18-40
 OPTICAL AND INFRARED PROPERTIES OF THE ATMOSPHERE
p~o 500 1000 I 500 2000 2500 3op?o.20
10- 3 , , , t !
296 K

308 K
, 0-21
, ~-21
322 K

k 338 <
+ 353 K , 0-22

,
, o-22_ -

~1
x 358 K
I
\ CALC 296

f
0-23
z 0-23 __
\
E

‘E 0-24 __
, 0-24
v

o-25
, 0-25 __
1/’

0-26
, 0-26 _-
\*

, 0-27
, 0-27

—---1* ---- , , ,~~o~o-z,

10 -28 10 “---~o~o 1500 2000 2500
WAVENUMBER (cm-’)

Figure 18-62. fies~lf density depndent continuum values C, forwater vaprasaknctionof wave number. Theexpenmental values me from Burch
[1981].

!O-zo
Q ?Qo 490 6Q0 8qo 1Opo 12po 14po 16p0 iBpo 2opo 22po 24po 2600 Zepo 30 Ppo-,o

‘–-–0–260K
296K
............338K

,
, ~-28 , ,
~. o~o-”
2b0 4b0 6b0 8b0 lobo 12’00 14bo 16bo 18’00 20’00 22bo 24bo 26bo 2abo

WAVENUMBER (cm-’)

Figure 18-63. The self density dependent continuum values C. for water vapr as a function of wave number at 260 K, 296 K, and 338 K. The values
from 296 K and 338 K are fits to experimental results, the 260 K is extrapolated.

18-41
CHAPTER 18
wlf and foreign corttinuum, and C, and cf [(cm-’ mO~crn2)-’ ~ Only values near room temperature are available for the
the wavenumber dependent continuum abso~tion parame- foreign dependence of the water vapor continuum. The con-
ters for the self and foreign components. The density n, is tinuum values Cf at 296 K are shown in Figure 18-64 and
the density of the absorbing molecular species and nr is the have ken obtained by a fit to the data of Burch. There is
density of all other rnoleculti s~ies; con~uently, n, + nf considerable uncertainty in the values for the spectral win-
represents the total density. The quantity, %, is the reference dow regions at 1000 and 2500 cm-’.
nurnbef density defined at 1013 rnb and 296 K. In sotne For njtrogen, continuum values at 296 ~ are included
cases the dependence of C on temperature is not known. in FASCODE for the collision induced absorption band at
The present formulation has the advantage that the contii- 2350 cm-]. For this case C, is taken to equal Cf, so that
uum’ contribution to the optical depth decreases with in- the effect is dependent on the total density.
creasing temperature tti~ough the number densi~y ratio term.
The quantities C, and Cf for water vapor and Cf for carbon 18.4.1.3 Radiance and Transmittance for Atmospheric
dioxide are’ stored in the program for the spectral range O Paths. The layering of the atmosphere is primarily de-
to 20000 cm-’. pendent on two considerations: the ratio of the Voigt widths
The values for ~, for water vapor at 296 K ~e shown at the layer boundaries and the temperature across the layer.
in Figure 18-62 together with the experimental values ob- The Voigt width as a function of altitude for the U.S. Stan-
tained by Burch [burch’and Gryvnak, 1978 and 1979; Burch dard Atmosphere is shown in Figure 18-65’. The samp]ing
et al., 1971; and Burch, 1970]. The 260 K result was ob- interval, DV, for FASCODE is indicated on the figure where
tained by extrapolating the fits to the 353 K and 296 K data nominally DV = av/4. All calculations are performed com-
of Burch. The results for the three temperatures are shown mencing at the lowest altitude involved in the problem.
in Figure 18-63. The strong temperature dependence of the Consequently the sampling interval decreases monotonically
self density dependent water vapor continuum is treated by in the altitude regime for which collisional broadening pre-
using exponential interpolation of the 260 K and 296 K dominates and becomes essentially constant at altitudes for
values. which the Doppler broadening is dominant. This method of

, ~-21 Q Sqo 1000 1500 2opo 25?0

, 1 1 1 r ??O-2’
o 296 K

k
A 308 K
O-22 , 0-22
+ 353 K

CALC 296
0-23 , 0-23

@
, 0-24

, 0-25

, 0-26

, 0-27

, 0-28

1 0-29+ 1 I
1
I 1 t LO-29
560 1000 I 500 2000 2500 3

WAVENUMBER (cm-’)

Figure 18-64. The foreign density dependent continuum values Ci for water vapor as a function of wave nqmtrer. The experimental values are from Burch
[1981].

18-42
 a
OPTICAL AND INFRARED PROPERTIES OF THE ATMOSPHERE
recursion relation for the accumulated optical depth, K(,
.-— after ~ layers is expressed as
90. 90. K< = it., +i{ (18.51)
80. 80.
I where kc is the optical depth forlayert and K. = O. This
;70 .
relationship is implemented layer by layer to obtain the total
-:60. optical depth for the path.
-;50.
Fortheradiance/transmittancemode, the radiant inten-
sity, If (W cm-2 sr-’/c),’), originating in layer (, is ex-
-:40.
pressed as
:30.

}20.
1< = (1 – Tt)B(T~), (18.52)
20.

10. / 10.
\{ where ~< is the transmittance and B (W cm-2 sr’/cm-’ ) is
o. ,O-i ““’ l’.-s ““’j’2-~ “jc.i “’~’o.j “’ ““’o-i
+0 o-i the Planck black body function at temperature Te for the
layer. Since all calculations are performed from lower to
VOIGT NilJ7H [CM-ii
higher altitudes, different recursion algorithms are required
F]gum 18-65. The Voigt width (HWHM) as a function of altitude for the for the looking down, the tangent path, and the looking up
U.S. Standard Atmosphere. geometric situation. For the looking down geometry, we
obtain the recursion relation for the contribution of the cur-
rent layer t to the upwelling radiation as
performing the calculation ensures that the minimum amount
of information is handled at any given stage of the calcu-
I? = Tc It_l + (1 – T<) B(T() (18.53)
lation. Figure 18-66 shows the Voigt parameter, ~, as a
function of altitude for the U.S. Standard Atmosphere.
with the condition that l., the boundary radiance, is
The atmospheric layering is performed by program
FSCATM [Gallery et al., 1983]. For specified boundaries
10 = E. B(To), (18.54)
and geometric parameters, the program computes the col-
umn densities and the density weighted temperature and
where T. is the boundary temperature and E. the boundary
pressure for each layer. Alternatively, the maximum ac-
emissivity. The transmittance relationship for this case be-
ceptable ratio of Voigt line widths for the layer boundaries
comes
and the maximum temperature differential across the layer
may be selected to perform the atmospheric layering. These
Tt = ~~T/.l (18.55)
two parameters are directly related to the accuracy of the
radiance and/or the transmittance result.
where Tt is the total transmittance after layer (.
FASCODE operates in either of two modes: optical depth
For tangent path geometry with symmetry about the
or radiance/transmittance. In the optical depth mode, the
tangent height, we obtain the recursion relations

0.2 0.4 0.6 0.8 1.0 It = TC It-, + (1 + T? Tt-, ) (1 – 7,-, ) (B(T,)

100! 100.
(18.56)
90. 90.

80. for the radiance with 10 = O and

x 70.
x T( = ~eT(_, (18.57)
60.
u
n 50. for the transmittance with T. = 1. This algorithm for the
3
1- tan-
40.
+ gent path case involves only a single line-by-line calcula-
-1 30.
u
tion for each layer and enables this case to be handled consis-
20.
tently with the other cases, that is, the calculation commences
10. with the tangent layer and proceeds to higher altitudes.
0.
( 0.2 0.4 0.6
For looking up geometries, the recursion relation for the
ZETR downwelling radiance is
Figure 18-W. The Voigt parameter E as a function of altitude for the U.S.
Standard Atmosphere. 1{ = 1(-, +T(-, (1 –70B (T/) (18.58)

18-43
CHAPTER 18
with 10 = O and for the transmittance sists of an array of equally spaced identical lines of Lorentz
shape. This model has been applied to absorption bands that
T( = 7e_, Tt_, (18.59) have a regular line structure, for example, to some bands
of COZ, NZO, CO, CHq, and 02. The Goody model, on the
with T’O= 1. These algorithms are implemented layer by other hand, assumes that the band is composed of spectral
layer starting with the initial layer at the lowest altitude and lines with an exponential intensity distribution and with
proceeding to the higher altitude. random spacing between lines. Again the lines are assumed
In Figure 18-67 we show the comparison of a FASCODE to have a Lorentz shape. This model has generally been
calculation with experimental results obtained by Rice and applied to bands that have an irregular line structure, for
Ade [Rice and Ade, 1979] in the millimeter spectral region. example, some HZO and Os bands.
The data were taken with an interferometer measuring the In practice the wavelength dependent absorption coef-
downwelling radiance at the ground from which atmospheric ficients are determined for each absorbing gas separately
transmittance was inferred. The calculation was performed using laboratory transmittance values measured under known
with a two layer atmosphere and the water amount given conditions. The absorption coefficients obtained in this way
by the measured amount for the Rice and Ade experiment. are then used in the band model transmittance function to
In Figure 18-68 we show the results for a fifteen layer determine the average transmittance for each absorber as a
calculation with a geometry in which the observer is at 100 function of frequency for other values of path length and
km looking through a 15 km tangent height. Figure 18-69 pressure. Finally, the total mean transmittance for molecular
shows an expanded view of the Q branch region shown in absorption is given by the product of the mean transmit-
Figure 18-68. The radiance calculation is plotted in terms tances of the individual absorbers at each frequency.
of equivalent black body temperature. Exact analytical expressions have been obtained for most
of the band models. However, they are sometimes difficult
to use and simpler approximations have been found to apply
18.4.2 Broad-Band Radiation in two limiting conditions common to all band models.
Those simpler expressions are the well known “weak line”
Band models assume an array of lines having chosen and “strong line” approximations [see Goody, 1964, and
intensities, half-widths, and spacings which can be adjusted Plass, 1958] for which the transmittance is a function of the
to represent the molecular line structure in some part absorber amount, and the product of the pressure and ab-
of a real band. For a particular band model, the mean trans- sorber amount, respectively (for a given temperature).
mittance can be represented by a mathematical expres- The weak line approximation, which corresponds to the
sion (transmittance function) expressed in terms of pressure, exponential law, is valid when the absorption is small at
temperature effective path length (or absorber concen- the line centers (generally for high pressures and low ab-
tration), and one or more frequency dependent absorption sorber amounts). Unfortunate y this case is rarely applicable
coefficients. to conditions existing in the terrestrial atmosphere. The strong
Several band models have been developed [see Goody, line approximation is applicable where the lines are com-
1952 and 1964; Elsasser, 1942; Plass, 1958; King, 1964; pletely absorbing at their centers; the effect of increasing
and Wyatt et al., 1964]; the Elsasser and Goody models the amount of absorber is then confined to the edges or
being the most well known. The Elsasser band model con- wings of the lines. The regions of validity of the strong and
weak line approximation for the Elsasser and Goody models
are discussed by Plass [ 1958]. For practical purposes most
problems fall in either the strong line approximation region
or the intermediate region.
In the LOWTRAN band model discussed in the follow-
ing section, empirical transmittance functions for HzO, 0s,
and the combined contributions of the uniformly mixed gases
have been determined from laboratory and calculated trans-
mittance data. Over a wide range of pressures and absorber
amounts, the empirical transmittance functions were found
to give better agreement with laboratory and calculated
,., transmittance data than the commonly used band models.
0.0 ‘ k’
.-’<
[
An excellent review of the LOWTRAN model and other
01 234567891011 1213
band models is given in La Rocca [ 1978].
WAVENUMBER ( Cm- ‘)

Figure 18-67. Spectral transmittance inferred from a radiance measure- 18.4.2.1 Broad-band Transmission (LOWTRAN). The
ment looking up from the ground. The symbols o are the
data of Rice and Ade [1979]. and the continuous curve is LOWTRAN model was developed to provide a fairly ac-
a two-layer FASCOD 1B calculation. curate, simple and rapid means of estimating the transmit-

18-44
 OPTICAL AND INFRARED PROPERTIES OF THE ATMOSPHERE

WAVELENGTH ( urn)
4.8s 4.00
1 .H :1.0

.8

.6

.4

.2

WAVENUMBER (cm-’ )

WAVELENGTH ( pm)
2s0 -

f60 . 1 c -7
w
~ f40.

i-
cf 220. 1 c -8
a
LsJ
n ‘zoo.
s 1 E -9
w Iso.
+
1 E -10
160.
1 E -11
140. 1
2030 Co’ss toko Zoks Z070 20’75 zobo zobs Zobo z 0’9s Zlilo

WAVELENGTH (cm-’)
Figure 18-68. The transmittance and equivalent temperature radiance plot at 100 km and a 15km tangent height. The results are from a 15 layer FASCOD I B
calculation.

WAVELENGTH (pm)
280. ‘
1

Z60 - 1

-v&
L
140. ! t I # i
Z07S Z076 Z077 Z078 Z079 z06

WAVENUMBER (cm-l)
Figure 18-69. An expanded s~ctral region from the results shown in Figure 18-68
CHAPTER 18
tance and background radiance of the earth’s atmosphere at 6. Atmospheric and earth radiance are calculated by a
moderate resolution (20 cm-] ) over a broad spectral interval numerical evaluation of the integral form of the equation of
(350 to 40000 cm-’). The model was derived using a single radiative transfer. The emission from aerosols and the treat-
parameter band model for molecular absorption and includes ment of aerosol and molecular scattering are considered only
the effects of cent inuum absorption, molecular scattering, in the zeroth order.
and aerosol extinction. It is based on an empirical graphic
prediction scheme of transmittance using some techniques
18.4.2.3 LOWTRAN Atmospheric Transmittance. In
originally suggested by Altshuler [ 1961]. The prediction
the LOWTRAN model, the total atmospheric transmittance
scheme, derived from laboratory transmittance measure-
at a given wavenumber averaged over a 20-cm-1 interval is
ments complemented by line-by-line transmittance calcu-
given by the product of the average transmittances due to
lations was reported by McClatchey et al. [1972] and, sub-
molecular band absorption, molecular scattering, aerosol
sequently, together with atmospheric models developed into
extinction, and molecular continuum absorption. The mo-
the model and computer code LOWTRAN [Selby et al.,
lecular band absorption is composed of four components,
1972; 1975; 1976; 1978; and Kneizys et al., 1980; 1983].
namely the separate transmittances of water vapor, ozone,
The model contains representative (geographical and sea-
nitric acid and the uniformly mixed gases (C02, N20, CH4,
sonal) atmospheric models and representative aerosol models
CO, 02 and N2).
with an option to replace them with user derived or measured
The average transmittance; over a 20-cm-i interval (due
values and computes atmospheric transmittance or both at-
to molecular absorption) is represented by a single parameter
mospheric transmittance and radiance for any given slant
model of the form
path geometry.

? = f(Cv w * DS) (18.60)

18.4.2.2 Assumptions of the LOWTRAN Model. The
basic assumptions made in the LOWTRAN Model are as
where C. is the LOWTRAN wavenumber-dependent ab-
follows:
sorption coefficient and m * is an “equivalent absorber den-
1. The atmosphere can be represented by a 33 layer
sity” for the atmospheric path DS at altitude z, defined in
model between sea level and 100 km. Each model atmo-
terms of the pressure P(z), temperature T(z), concentration
sphere contains the variation of pressure, temperature, water
of absorber w and an empirical constant n as follows
vapor, and ozone with altitude. A mean nitric acid profile
is incorporated into the model. It is assumed that the mixing
ratios of C02, NZO, CH4, CO, Oz and N2 remain constant
with altitude. Each layer is assumed to be in thermal equi-
w*=w{y&[ (18.61)

librium.
2. Refraction and earth curvature effects are included in where PO and To correspond to STP ( 1 atm, 273 K). If
the calculation of absorber amounts for slant atmospheric Equation (18.61 ) is substituted in Equation (18.60) and n
paths. is set to zero and unity respectively, Equation ( 18.61) reverts
3. The average transmittance over a 20 cm-’ interval to the well-known weak-line and strong-line approximations
(due to molecular absorption) can be represented by a single common to most band models.
parameter empirical transmittance function. The argument The form of the function f and parameter n was deter-
of the transmittance function is the product of wavenumber mined empirically using both laboratory transmittance data
(or wavelength) dependent absorption coefficient, and “an and available molecular line constants. In both cases, the
equivalent absorber amount” for the atmospheric path. transmittance was degraded in resolution to 20 cm-’ throughout
4. The total transmittance at a given wavenumber av- the entire spectral range covered here. It was found that the
eraged over a 20 cm-] interval is given by the product of functions f for H20 and the combined contributions of the
the average transmittances due to molecular absorption, mo- uniformly mixed gases were essentially identical, although
lecular scattering, aerosol extinction, and continuum ab- the parameter n differed in the two cases. Mean values of
sorption. It is further assumed that the molecular absorption n were determined to be 0.9 for H20, 0.75 for the uniformly
is composed of four components, namely the separate trans- mixed gases, and 0.4 for ozone.
mittances of water vapor, ozone, nitric acid and the uni- Figure 18-70 shows the LOWTRAN empirical trans-
formly mixed gases. mittance functions defined by Equation (18.60) versus the
5. The variation of aerosol optical properties with alti- log,0 of the effective optical depth (C” w * DS). The solid
tude is modeled by dividing the atmosphere into four height function shown is used for water vapor and the uniformly
regions each having a different type of aerosol. The aerosol mixed gases. The dashed function is applicable to ozone.
models for the lower atmosphere (RURAL, URBAN, MAR- Absorption coefficients for water vapor, ozone, and the
ITIME and TROPOSPHERIC) are given as a function of combined effects of the uniformly mixed gases are included
relative humidity. Radiation and advection fog models are as data for LOWTRAN. The absorption coefficients for
also included. water vapor, ozone, and the uniformly mixed gases are

18-46
 W

OPTICAL AND INFRARED PROPERTIES OF THE ATMOSPHERE

\\
,0 5. 5.0
1.

\\
t t

w
o.

4 ‘,,“\
!8

I
v 16
z 0.6
a
!- ‘, “\,,
H20 AND UNIFORMLY \ ‘i,
~— MIXED GASES \
:;.4 ‘,, .4
\
K
1- ––o~
‘, \,

c.? 1
\\
i\>\“\, . ?.

.5

LOG IO (EQUIVALENT OPTICAL DEPTH) WRVENU1’lBER [cK-l]

FIgurc 18-72. Absorption coefficient C, for ozone from 3$0 to 5000 cm-’.
FI~18-70. LOW~Nempirical transmittanm tinctionsvemus~glo
of the effective optical depth (CW*DS).

tered one or more times are neglected. Local therinodynamic

shown in Figures 18-71, 18-72 and 18-73 forth~ SPC~l equilibrium is assumed”in the at~ospqere.
region from 350t05000crn-’. The average atmospheric radiance (over a 20-cm-’ in-
terval) at the wavequmkr, v, along a given line-of-sight in
18i4.2.4 LOWTRAN Atmospheric Radi~ce. me terms of the LOwTRAN transmittance paramete~ is given
LOWRAN rn@el and code also calculate atmospheric a~d by
earth radiance. A numerical evaluation of the in~egral form
of the equation of radiative transf?r is used. The emission
from aerosols and the treatment of aerosol and moleculw I(v) = ~:b
d F. B(v,
a
T) 7, + B(v, T~] ? ~, (18.62)
scattering is considered only ‘in the zeroth order; Additional
conrnbutions to”atmospheric emission ftorn radiation scat-

UNIFORMLY MIXED GASES .0

5. C02fl
C02 II
.0

.0

1.0

-3.0

-5. ~s. o
-5. 5.0
1000 2000 3ioo 4000 So’oo k
Lit
WRVENUfl BER [cm-l]
WRVENUflBER [CM-1 )

Figure 18-71. Absorption coefficient C. for water vapr from 350 to 5~ Figure 18-73. Absorption coefficient C, for the uniformly mixed gases
from 350 to 5000 cm-’.
cm-’.

18-47
CHAPTER 18
where the integral represents the atmospheric contribution
and the second term is the contribution of the boundary,
(for example, the surface of the earth), and

7, = average transmittance due to absorption,

T. = average transmittance due to scattering,
?t=?a?s= average total transmittance,
7:, ?; = average total gansmittances from the
observer to boundary,
B(v,T) = average Planck (blackbody) fitiction
corresponding to the
frequency v and the temperature T of
an atmospheric layer, Hfl VENUMBEk [C M-1 I

T~ = tem~rature of the boundary. Figure 18-75. Transmittanw s= for a 10-ton horizontal path at sea
level for the h, maritime, urban, and kopoaphenc aer-
osol models using the U.S. Stanti Model Atmosphe~
The emissivity of the boundary is assumed to be unity. anda VISof23ti.
The numerical analogue to Equation (18.62) has been
incorporated in the LOWTRAN mddel.
km). This calculation fot the s@tral region from 400 to
18.4.2.5 Examples of Transrnit@nce and Radiance 4000 cm-] used the rural aerosol model (23 km met. range)
Spectra. Some ex~plei of transmittance and radiance and the U.S. Standard model atmosphere.
spectm obtained from the LO~RAN model are presented
in Figures 18-74 through 18-77. 18.4.2.6 Comparisons of LOWTRAN with Measure-
Figure 18-74 shows the transmittance from ground to ments. Comparisons of LOWTRAN with measurements
space from 0.25 to 4pm. This calculation used the U.S. are Presented in Figures18-78 though 18-87.Figures
Standard model atmosphere and the rural aerosol model with 18-78 and 79 show transmittance comparisons of LOW-
a 23-km met. range. ~N with laboratory measurements of Burch et al. [1%2]
Figure 18-75 shows the variation in transmittance in the for some important water vapor and carbon dioxide ab-
spectral region between 4@ and 4000 cm-’ for the rural, sorption bands. ‘
maritime, urban, and ~pospheric aerosol models. The cal- Figure 18-80 shows a transmittance comparison with a
culation is for a 10-krn horizontal sea-level path using the sea-level measurement by Ashley et al. [1973] (General
U.S. ,Standard model atmosphere and a 23-km meteorolog- Dynamics). The measurement, made with an interferometer
ical range.
Figure 18-76 shows the transmittance of the two fog
models in LOWTRAN for a O.2-km horizontal sea-level ‘“”~ T’””
path and a 1-km meteorological range in the spectral region 0.9 .9

I
from 400 to 4000 cm-l. .....
Hl=OKM
0.8 ,.. ‘.. .8
Figures 18-77 a and b show the transmittance and ra- ;.: ..
RANGE= 02 KM

.: ‘.. — COG 1 (AOVECTlON)

diance spectra for a vertical path at O“zenith angle looking 0.1
...! .:
FOQ 2 (RADIATION) .7

to space as a function of observer altitude (O, 20 tid 40 0.6 I i ““””

.6

).or — — .) .,,

HRVENUtlBER [C M-I 1
WAVELENGTH ( p m )

Figure 18-76. Transmittance spectra for the advwtion fog (Fog 1) and
Figure 18-74. Transmittance s~tra for a vertical path from grmmd to the radiation fog (Fog 2) models, for a 0.2-krn horizontal
space from 0.25 to 4 p, oaing the rurat aerosol model, 23- path at sea level, with the U.S. Standard Model Atmos-
Ion VIS and the U.S. Standard Mdl Atmosphere. phere and a l-km VIS, from 400 to 4000 cm-1.

18-48
 a
OPTICAL AND INFRARED PROPERTIES OF THE ATMOSPHERE

I
y I
vi

v:. .
y“

~ J’ —
---
BURCH, et al (1962)
LOWTRAN
g .: 1
m.
m
u
+
z 2:
.

2
u 42
a

6?

03

,.. —–
,..
? ?00
~~~c
,200 ,.00 ,600 1800

W9VENUfl BER [CM-II

WAVE NLJMBER ,. cm
a
Figure 18-78. Representative absorption curves for the 6.3 pm H20 Band.

terferometer was 5 cm-’ as compared to the 20 cm-’ reso-

lution of LOWTRAN. As shown in Figure 18-82, two the-
oretical models (tropical and midlatitude summer) were used
for comparison and are displaced two divisions above and
below the measured radiance for clarity. Both models as-
sumed a 23 km met. range and used the temperature at O
km in the model atmosphere as the boundary temperature.
Figure 18-83 shows the comparison of atmospheric ra-

-. ..
-.
WRVENUfl BER [Ci--ll
0.8 - ‘.

u
b v
~ 0.6 -
1-
Figure 18-77 Transmittance and,radiance spectra for a vefiical path look- b
s
ing to space from HI (H1 = 0, 20 km, akin, Hz ~ 100 lno4 -
z
km, ANGLE = 0°) the rural aerosol model (IHAZE = 1, a ;W in atm-cn
VIS = 23 km) and the U.S. Standard Atmosphere a
+02 -
(MODEL = 6), from 400 to 4000 cm”’: (a) transmittmce,
(b) radiance. ---- LOWTRAN
I 1
0.0
2200 2300 2400
WAVENUMBER ( CM-l )

of -4 cm-’ resolution from 1.8 to 5.4 pm, is for a 1.3 km 1.0

sea-level horizontal path.
Figure 18-8 I shows a comparison with Chancy [19691 ~ 0.8 - ‘\
v .
z
of the calculated upward atmospheric radiance and an in- : 0.6 -
~
terferometer measurement from a balloon flight over north-
s
m 0.4
ern Nebraska. The measurement was taken at a float altitude ~
of 34 km. The calculated radiance used the midlatitude :0.2 -
BURCH, et al [1962)
0.748 ---- LOWTRAN
winter model with a 23 km met. range and a ground temp- 6.15 I
erature of 280 K. 0.0
575 625 675 725 775

Figure 18-82 shows a comparison of an interferometer WAVENUMBER (C M-l )

measurement made from the Nimbus 3 satellite Iconrath et

al.. 1970] looking down over the Gulf of Mexico with the Figu~ 18-79. Comparison LOWTRAN cafcuh~tions and Burch et al. [ I%2]
calculations for C02 bands at 4.3 #m and 15pm.
calculated atmospheric radiance. The resolution of the in-
18-49
CHAfiER 18

---- LOtiRAN

— GENERAL DYNAMICS

1.

..

. A
90 2.20 2.60 3.00 3.40 3.81J ’20 4.60 S.oo 5
WAVELENGTH ( pm )

F!gure 18-80.
Comparison bctwLen LO*Nmd Geneti~mics memmtink; ~ge = 1.3krnataea Ievel,
——— — LO~TRAN
12

WAVENUMBER (cm-1)
Figure 18-81. Comparison between LOWTRAN prediction and University of Michigm balloon measurement of atmospheric radiance over northern
Nebraska.

18-50
— -

OPTICAL AND INFRARED PROPERTIES OF THE ATMOSPHERE

l“’’I’’ l’’’’’ l’’’’ l’’’’’lll “’’’’’’’’’’’’[’” I
1

2.0 —-—— LOWT R AN

I

.‘::.L4
\ \./\ /’:
1’ ,..,.
/’:

\l
I
1.
l.’

,:

... .
... . .
,..
.,,.‘,
,. -.+.
05 -
““”p “,

.,
CH4
n -.

.. . .
,,
. . ,,,
. . .
,,,
.,,

0.0 ~
I 1 , I h 1 I 1 t I I I & I , I I I , , I I 1 I 1’
400 600 800 1000 1200 1400
WAVENUMBER (cm-i)
I I I I I 1 1
25 20 15 14 13 12 11 10 9 8 7

WAVELENGTH (pm)

Figure 18-82. Comptison ktween LOWTRAN pmtiction mdNIMBUS 3satellite measumment ofatmosphenc radimce over the Gulf of Mexico,

diance as seen from space between the LOWTRAN cal- order of the grating. The resolution was 0.03 pm in the 8
culation and measurements from the Nimbus 4 satellite [Ha- to 14 pm region. The data in these figures are presented as
nel and Conrath, 1970] for three different geographic locations. a function of altitude and as a function of zenith angle. The
The spectra, obtained with a Michelson interferometer of LOWTRAN radiance calculation used the pressure, tem-
resolution 2.8 cm-’, were measured over the Sahara Desert, perature, ozone, and nitric acid profiles from the -ray
the Mediterranean, and the Antarctic. The calculated LOW- report, and the midlatitude winter water vapor profik csm-
TRAN radiances used the midlatitude winter model and a tained in LOWTRAN.
ground temperature of 320 K for the Sahara, the midlatitude
winter model and a ground temperature of 285 K for the 18.4.2.7 Limitations of LOWTRAN. The overall ac-
Mediterranean; and an arctic winter cold model taken from curacy in transmittance from LOWTRAN is better than 10%.
the Handbook of Geophysics and Space Environments [Cole The largest errors may occur in the distant wings of strongly
et al., 1965] and a ground temperature of 190 K for the absorbing bands in regions where such bands overlap ap-
Antarctic comparison. All three calculations assumed a 23 preciably. The reason for this error is twofold. First, the
km met. range for aerosols. LOWTRAN spectral coefficients are based on a single ab-
Figures 18-84 through 87 show comparisons of calcu- sorber parameter and cannot be defined for a wide range of
lated and observed atmospheric spectral radiance versus atmospheric paths without some loss in accuracy.
wavelength in the 8 to 14 ~m spectral region. The meas- Secondly, the transtiittance in the window regions be-
urements were made on a balloon flight launched from Hol- tween strong bands generally lies in the weak line approx-
loman AFB, New Mexico by Murcray et al. [ 1977] Uni- imation region where the transmittance is a function of the
versity of Denver. The instrument used for these observations quantity of absorber present and not of the product of ab-
was a grating spectrometer, operated in the first and second sorber amount and pressure. The one-dimensional prediction

18-51
CHAPTER 18
– – – – LOWTRAN —MURCRAY ET AL, HOLLOMAN AFB; NEW MEXICO,
19 FEBRUARY 1975
--- LOWTRAN

‘0-3 ~
REC 97
:%
~ IO-4 - n- . ,
/’
N
/
E

LLt 1
I
v \
z
~ ,.-6
4

2
E

: ,0-7 -
1-
V
k
m ,0-8 1 1 1 I I
891011 12 13 14
WAVELENGTH (pm)

Figure 18-85, Sample spectrumof short wavelength region observed at

an altitude of 13.5 km and a zenith angle of 63° on 19
1 1 1 February 1975 and LOWTRAN comparison.
I
ANTARCTIC

------
I 1
scheme in LOWTRAN is less accurate for such conditions.
0 1 —
400 600 800 1000 1200 1400 1500 The digitized spectral data used in LOWTRAN were ob-
WAVE NUM8ER-cm-1 tained for conditions representative of moderate atmospheric
paths and will tend to overestimate the transmittance for
Figure 18-83. Comparison between LOWTRAN predictions and NIM- very long paths and underestimate the transmittance for very
BUS 4 satellite measurements of atmospheric radiance over
the Sahara Desert, the Mediterranean, and the Antarctic.
short paths in the spectral regions described above. As the
transmittance approaches 1.0, the percentage error in trans-

— MURCRAY ET AL, HOLLOMAN AFB. NEW MEXICO.

19 FEBRUARY 19”75 — MURCRAY ET AL, HOLLOMAN AFB, NEW MEXICO,
‘ -- .LOWTRAN 19 FEBRUARY 1975
--- LOWTRAN

I 0-3

.’0-3r————— ;*
-IO-4
m
N
k
;10-5
w
u
z
~10-6
n
a
u
-t
d
=
-7
10
:,. -7
1- t
b

!lo-a~ 89 10 II 12 13
w
n
m ,.-8
8 9 10 12 13 14
WAVELENGTH (pm) WAVELENGTH (pm )

Figure 18-84. Sample spectmm of short wavelength region observed at Figure 18-86. Sample spectrum of short wavelength region observed at
an altitude of 9.5 km and a zenith angle of 63” on 19 an altitude of 18.0 km and a zenith angle of 63” on 19
February 1975 and LOWTRAN comparison. February 1975 and LOWTRAN comparison.

18-52
 d

OPTICAL AND INFRARED PROPERTIES OF THE ATMOSPHERE

—MURCRAY ET AL. HO LLOMAN AFB, NEW MEXICO, for multiple scattering events increases with increasing op-
19 FEBRUARY 19”75 tical thickness t = k. . x [see Equation (18.21)]. Van de
-—. -LOWTRAN
Hulst [ 1957] states that if the optical thickness t <0.1 single
scattering prevails, for O. I < t < 0.3 a correction for dou-
ble scattering may be necessary. For stili larger values of
the optical depth the full complexities of multiple scattering
have to be considered.
Monographs by Chandrasekhar [ 19501, Sobolev [ 1975 ],
and van de Hulst [ 1980] are classical references on theory
G

w --- and approaches to radiation transfer. Recent reviews of mul-

I
v
z / tiple scattering have been presented by Hansen and Travis
I
~ ,.-6
n
[1974], Herman et al. [ 1978], and by the Radiation Com-
<
u
mission of the International Association of Meteorology and
I
Atmospheric Physics (lAMAP/RC) [Lenoble, 1977 and
~ ,0-7 \

1-
Fouquart et al., 1980].
: The basic radiative transfer equation in a scattering at-
a
u) lo-e 1 mosphere defining the diffuse radiance 1. can be written as
8 9 10 II 12 13 14

WAVELENGTH (pm)
dl.(~, P, +) =
P – IV(L, p, 0) + J.(tw, P 4)
Figure 18-87. Sample spectrum of short wavelength region observed at d~
an altitude of 24.0 km and a zenith angle of 63° on 19
February 1975 and LOWTRAN comparison.
(18.63)

where the source term, Jv, is given by

mittance decreases toward zero but tbe uncertainty in the
absorptance (or radiance) increases.
Additional constraints on both the validity of the model
as well as the range of applicability are introduced for at-
mospheric radiance calculations. As mentioned above, the
atmospheric radiance becomes less accurate for very short
paths. In addition, the radiance calculations assume local
thermodynamic equilibrium exists in each layer of the model
atmospheres. This assumption will break down for radiance and where the following notation is used:
calculations in the upper atmosphere.
For the shorter wavelengths (<5 pm), scattered Solar .
radiation becomes an important source of background ra-
diation. This contribution for radiance calculations at the
~.
Jz
k“(z’)dz’ optical depth

shorter wavelengths with a sunlit atmosphere would add to k.(z) = extinction coefficient
the atmospheric radiance calculation using the LOWTRAN @u(L) = albedo for single scattering
model. Single-scattered solar and lunar radiation has been
added to the latest LOWTRAN model [Kneizys et al., 1983]. z= height
p= Coso
g= angle between a given direction and
the direction of increasing optical
18.5 RADIATION TRANSFER BY depth, t (that is, towards the ground).
ATMOSPHERIC SCATTERING
+= azimuth angle

P,(p,+, p’, +’) = phase function defining the function

18.5.1 Multiple Scattering
of the light incident at p’, 4’ which
is scattered in the direction w, 0.
Scattering of visible and infrared radiation occurs in the
atmosphere by air molecules, haze particles, and water drop- BV(t,T) = Planck blackbody function at
lets in the air (see Sections 18.1.4 and 18.2). Scattering temperature T.
effects from molecules and aerosols are additive. Multiple
scattering occurs when scattered light undergoes subsequent mFW, = solar irradiance perpendicular to the
scattering on other particles or molecules. The probability direction of incidence, (PO. $0)

18-53
CHAPTER 18
The first term on the right side of Equation (18.63) is the
energy lost from the beam of radiation ‘traveling in the di-
rection (~, ~). This loss is due to either absorption or
t* t)]
‘-(t*~*+)exp[-
‘t*li
scattering out of the beam. The second term (known as the
source term) represents the diffuse radiation contributed to
+~J k,
J(t’,p,@)exp
[“vldt’
the direction (p, ~) by scattering processes or self emission
by the atmosphere. This source term is expanded in Equation p<o (18.66a)
(18.64) into three components (in order): the diffuse radia-
tion scattered into the beam from other directions, the direct p=()
I(t,p,+) = J (t,p,+), (18.66b)
sunlight scattered into the beam direction (p, ~), and the
thermal radiation emitted by the atmosphere at frequency
u. For simplicity, the frequency u will be omitted in sub-
sequent equations. However it should be kept in mind that
these equations only apply to spectral intervals sufficiently
narrow that the atmospheric absorption and scattering pro-
cesses do not change significantly with wavelength.
The full radiative transfer Equation (18.63) with the
source term as in Equation (18.64) is too cumbersome to L ~>o (18.66C)
solve in general. However for most scattering problems in
the earth’s atmosphere, there is a natural separation of the where t* is the total optical depth of the atmosphere and
parts of the source term. In the temperature region of interest where I-(t*, ~,@) for p <0 and I+(0,p,4) for ~ >0 (de-
for the earth’s atmosphere, BV(t,T) contributes only in the noting the upward radiance at the ground and the downward
longwave region (infrared and beyond). For wavelengths radiance at the top of the atmosphere respectively) are spec-
less than 3 to 4 Am, generally the thermal radiation BV(t,T) ified by the boundary conditions: At the ground, the up-
can be neglected compared to the solar irradiation Fw, and welling radiance is related to the downward radiance by an
for longer wavelengths FM, can be neglected compared to appropriate surface reflection law (see Section 18. 3), and
Bv(t,T). at the “top” of the atmosphere the downward diffuse radi-
Also, for the longer wavelengths (greater than 3 to 4 ance I + (O,p,+) for p >0 is usually taken to be zero (that
~m) scattering is negligible except in the presence of clouds is, the downward radiance at the top of the atmosphere is
or fog. These cases where the source term is limited to the due only to the incident sunlight). The formal solution
thermal radiation BU(t,T) are discussed in Section 18.4. Sec- [Equation (18.66)] does not completely solve the radiative
tions 18.5.2–1 8.5.4 will focus on the solution to Equation transfer equation since J is defined in terms of I [Equation
(18.63) when the thermal radiation can be neglected al- (18.65)]. However it is used in the development of several
though many of the methods of solving the multiple scat- of the methods discussed below.
tering ,problem can easily be generalized to include atmos-
pheric emission. The reduced expression for the source term 18.5.1.2 Method of Successive Orders of Scattering.
is given by The method of successive orders of scattering expands the
radiance as [see e.g. Dave, 1964, 1965, or Irvine, 1965]

Q(t) 211 +1
.
J(t,~,@) = ~ “
JJ –1
P(p,@;p’,$’)I(t, p’, @’)dp’d@’
I(t,p,+)
= ~
“Cl
1’”)(t,p,+) (18.67)
m(t)
+ — – dP,,)F..
P(p,+;p,,,+,,)exp( (18.65)
4 where 1(”) is the part of the radiance scattered exactly n
times. I’”)(t, ~,~) is found by evaluating the formal solution
The phase function P used here is a weighted average of [Equation (18.66)] using J = J(n)(t,p,@) which is given by
the phase functions for molecular scattering and aerosol Equation ( 18.65) with I = 1(”- ‘)and starts with 1’”) = O.
scattering discussed in Sections 18. 1.4 and 18.2. 1.5 re- This method has the advantage of easily handling vertically
spectively. The weights are proportional to the scattering inhomogeneous atmospheres. The main disadvantage is that
coefficients for scattering by air molecules and by aerosol the computation time becomes very long as o approaches
particles. 1.0 (absorption becomes weak) or as the optical depth be-
comes large.

18.5.1.1 Formal Solution to the Multiple Scattering

Problem. The formal solution to the radiative transfer 18.5.1.3 Gauss-SeideI Iterative Method. The Gauss-
equation is given by Seidel iterative method originally developed by Herman and

18-54
 OPTICAL AND INFRARED PROPERTIES OF THE ATMOSPHERE
,,
Browning [1965] and others [Dave and Gazdagj 1970] also in temis of, the scattering and transmission functions for the
starts with Equations (18 .65) and (18 .66). Here they initially individual layers. For a homogeneous atmosphere, it is most
assume 1 = O and numerically integrate Equation (18.66) efficient to start with t, = tz = h, where to is small, and
for small steps At for downward radiances (K > O) until build up to t = 2P ~ in p steps, by a doubling procedure.
they reach the ground. They then integrate stepwise from The Neumann series method (iteration in orders of scat-
the ground back to the top of the atmosphere using the tering [see Irvine, 1965] has been used to find S(k) and
downward radiances just calculated for the integral in Equa- T(b) for ~ = 0.125 by van de Hulst and Grossman [ 1968]
tion (18. 65). Th& process is then repeated until the solution and Ifiine [ 1968b]. Hansen started with an initial optical
~onverges. R&centl y Herman et al. [1980] improved the thickness of to = 2-25 where he could approximate S(k) and
computation speed by approximating I(t, p,+) by polyno- T(h) by the appropriate expressions for single scattering.
mials in t which allowed the use of larger step sizes At for The formalism developed by van de Hulst and Hansen has
optically thick atmospheres. only been applied to radiation ~m~rging from an atmosphere
and in its present form cannot be used to calculate the radiation
18.5.1.4 The Matrix Operator or Layer Adding within ah atmosphere. The approach developed by Grant
Methods. The radiative transfer problem has been solved afid Hunt overcomes this drawback by incorporating an
by using different formations of the matrix operator or layer algorithm for determining the radiances between the layers
adding techniques reviewed by Plass, et al., [1973]. One of used in the calculation. However, this can be quite time
the first applications to light scattering in, the atmosphere consuming for a number of different depths. The advantage
was the doubling method discussed by van de Hulst in an in computation time is also weakened if the approach is ex-
unpublished report ih 1963 and later used by him and others tended to vertically inhomogeneous atmospheres.
[e.g. van de Hulst and Grossman, 1968; van de Hulst,
197 1]. Hansen [1969a and b] has developed a modified 18.5.1.5 Monte Carlo Technique. The Monte Carlo
version of ~his method and used it extensively [Hansen, method basically utilizes a random walk approach, follow-
197 1a and b; Hansen and Pollack, 1970]. ing a large numkr of photons in turn as they scatter through
Ari equivalent procedure was independently developed the atmosphere until they are absorbed or emerge from the
by Twomey et al. [1966] and has been used by them to atmosphere. This procedure was first extensively applied to
study multiple scattering of light [Twomey et al., 1967]. A the study of light in the atmosphere by Collins, Wells, and
similar approach has been fomiulated by Grant and Hunt their associates [Collins and Wells, 1965; Thompson and
[1968, 1969d, b] based on invariance principles [Chandra- Wells, 1971; Collins et al., 1972; Blattner et al., 1974].
sekhar, 1950] and the discrete space theory [Preisendorfer, Subsequently, PlaSs and Kattawar [1968a, 1968b & 1972]
1965]. They have applied this matrix operator method to a have independently developed a Monte Carlo type of pro-
variety of problems. gram. The Russian work in this area is covered in the mono-
If we consider a plane-parallel ~edium of optical depth graph by Marchuk et al. [1980].
t and an incident radiation field Iinc(O,p’, +’), (O < p’ < 1), The major advantage of the Monte Ctilo approach is its
it is convenient to introduce a scattering function S (t; p, great flexibility. It can allow for all of the complications of
o, p’, 0’) and a transmission function T (t; A, ~, w’, ~’) a realistic atmosphere usually neglected by other methods;
[see Chandrasekhar, 1950]. Then the angulti distribution these include tlie spherical nature of the atmosphere, mo-
of the reflected and transmitted iight will be given by lecular band model type absorption, and horizontal irihomo-
geneities. The major drawback is that the errors in the cal-
Ire” (O,p,+) = culations are proportional to N – “2, where N is the number
of photons counted. This means that to increase the accuracy
by an order of magnitude, the number of photons foilowed
(and the computer time) must be increased by two orders
()> P> .-l of magnitude.
(18.68a)

Itran,(t,~,+) = 18.5.1.6 Dissrete Ordinates. One of the &st known so-

lutions of the transfer equation is Ctiandrdsekhv’s [1950]
method of discrete coordinates, Here, the phase function is
expanded as a series of spherical harmonics:
I>w>() (18.68b)

The basic principle of the layer-adding methods is that given P (p, +, p’, +’) = i (2 – ao.m)
m=O

x,~m
@T
p(P)PF(p)
? 1
the scattering functions S(t I) and S(t2) and transmission (18.69)
function T(t, ) and T(tz) of two layers, it is possible to find cos m(+’ - 4),
S(t, + t2) and T(t, + tz) for a combination of the layers [

18-55
CHAPTER 18
where direction is p. Application of this integral operator to the
radiative transfer equation leads to a system of linear dif-
O,m#O ferential equations in terms of the streams, S~, in the
8 c).m = (n + 1)(n + 2) different directions ~.
{ l,m=O For n = 2, there are 12 streams that can be arranged
to be centered on the faces of a regular dodecahedron. This
and the P? are associated Legendre polynomials. Then the arrangement minimizes the number of different scattering
radiance is, expanded in a cosine series, leading to the fol- angles that must be considered. Whitney has used her do-
lowing system of equations: decaton approach in combination with exact single-scatter-
ing calculations and has used the shape of the single-scat-
dIm(t,p) tering results to estimate the multiple-scattered radiances in
P — = – Im(t,p)
directions other than for the 12 streams.
di
Whitney has compared the computation speed of her
DART method with the Monte Carlo technique, which is
the only other method that has been applied to radiative
transfer calculations in realistic spherical atmospheres. She
found the DART method to be considerably faster, although
it is hard to judge how much faster because of the differences
m= 0,1, . . ..N. (18.70) in computers and programming languages and the question
of whether more angles and photon histories than necessary
where were used for the Monte Carlo calculation. In a comparison
with other methods such as spherical harmonic or matrix
n
operator for a homogeneous plane-parallel atmosphere [Len-
I (t, ~, ~) = ~ 1’” (t, p) cos m (+0 – +). (18.71)
oble, 1977], the DART method appeared to be slower and
“1=0
less accurate.
The primary advantage of the DART method is its ap-
After suithble manipulations and evaluation of the integral
plicability to a spherical geometry with reasonable accuracy
in Equation (18.70) by Gaussian quadrature of order
and computer time. It is not as advantageous for plane
n(4n > 2N – 1), Equation ( 18.70) is replaced by a system
parallel geometries or irradiance calculations.
of linear differential equations in terms of I’M)[t, pi), where
~i(i=*l, ~2, ..., ~ N) are the quadrature points.
This system of equations can be evaluated quite readily. In 18.5.1.8 Spherical Harmonic Method. The develop-
principle, as N (and n) - =, the solution becomes exact. ment of the spherical hamonic method goes back half a
Various forms of this approach have been used by Samu- century [Jeans, 1917 and Eddington, 1926]; however, it has
elson [ 1967 and 1969]; Lieu [ 1973]; Shettle and Green been used more widely for neutron transport problems. In
[ 1974}; Stamnes and Swanson [ 198 I]; and Stamnes and Dale recent years, it has begun to be used more for atmospheric
[1981]. light scattering problems (for example, Devaux and Her-
man, 197 1; Canosa and Penafiel, 1973, Dave, 1975a,b; and
18.5.1.7 Dodecaton Approach to Radiative Transfer Karp et al., 198 l). Like the discrete ordinate method, the
(DART). A different discrete stream approach has been spherical harmonic method expands the phase function as
developed by Whitney [ 1972, 1974], and Whitney and a sum of associated bgendre polynomial [Equation ( 18.69)].
Malchow [ 1978]. This approach appears to have consider- However, it differs from the discrete ordinate method in
able potential especially for problems requiring consider- that the radiance is also expanded in a series of spherical
ation of the curved geometry of the earth’s atmosphere such harmonics. Equivalently, the Fourier components I“’ are
as observations of the earth’s limb or twilights. This me- expressed as an expansion of Legendre functions:
thod has several important conceptual differences from the
discrete ordinate, or two-stream and multichannel methods Im (t, p) = ~ Cy(t) PT(~). (18.73)
(=()
discussed elsewhere in this section. The DART method
achieves discretion by angular integration of the radiance Substituting Equation (18 .73) allows the integrals in Equa-
I in a given direction, ?, with a cosine to the nth power tion ( 18.70) to be evaluated analytically using the orthog-
weighting function:
onality of the Legendre functions. After additional manip-
ulations of the equations, a linear system of first order
differential equations for the C~(t) is obtained.
The major advantage of the spherical harmonic method
where p ? is the dot product of the unit vectors p and f is that it can give the radiances at all depths in the atmosphere
and S; is a radiation stream whose nominal propagation and not only those emerging from the atmosphere. Those

18-56
 w
OPTICAL AND INFRARED PROPERTIES OF THE ATMOSPHERE
timing comparisons made with other methods (the matrix function as the sum of a Dirac delta-function and a Legendre
operator method, successive orders of scattering, and Gauss- expansion (with 2M terms). This is a generalization of the
Seidel iteration) tend to favor the spherical harmonic method approach for the deha-Eddington Approximation [Joseph et
[see Lenoble, 1977], especially for cases where the sun is al., 1976]. Potter [ 1970] had basically just truncated the
directly overhead and the azimuthal symmetry allows the phase function by eye. McKellar and Box’[!981 ] have pro-
system of Equation ( 18.70) to be limited to m = O. The vided a discussion showing the relationship between these
discrete ordinate method should be comparable to the spher- delta-function approximations and the necessary resealing
ical harmonic method in terms of computational efficiencies. of the transfer equation, with other scaling laws and simi-
The primary disadvantage of these methods is that the com- larity relations used in radiative transfer theory.
putational load increases with the number of terms kept in
the expansion of 1 and P (or the asymmetry of the phase 18.5.1.10 Approximations of Irradiance. Several dif-
function used); also, the radiance values calculated from the ferent methods of calculating radiances within. or emerging
basic solution often have appreciable oscillations about the from, a scattering atmosphere have been discussed. When
correct solution unless some smoothing process is used [Dave interest is mainly in the irradiances F. less accurate solutions
and Armstrong, 1974 or Karp, 1981]. of the transfer equation can be used because the errors may
cancel out in averaging the radiances over a hemisphere
18.5.1.9 Small-Angle Approximations. When light is
scattered by a particle whose dimensions are larger than the
F (t) = pl(t, p, +]d@dp (18.75)
wavelength of the incident radiation, it is diffracted in a JJ
narrow forward cone. It is this sharp diffraction peak that
greatly increases the difficulty of solving the multiple scat- where p > 0 corresponds to F ~ and p < 0 to F ~ . Often
tering by increasing the number of angles required to eval- it is adequate to have the irradiances computed accurately
uate the integral in Equation (18.65). Romanova [1962 and to a few percent, such as calculating heating rates or the
1963] has developed a procedure in which she separates the irradiance incident at the ground. Several approximations
radiance for the small-angle scattering i from the rest of the for irradiances often used are basically simplifications of
radiance: the closely related discrete ordinate and spherical harmonic
methods discussed above. When considering the irradiance.
i(t,p, +) = i(t,p, 0) + i(t,w,4) (18.74) only the first term (m = O) in the cosine expansion of the
radiance in Equation ( 18.7 I ) must be included, since the
and solves for j exactly. The rest of the radiation field (i) higher order terms drop out in doing the azimuthal or $
can then be found from a modified form of the transfer integration in Equation ( 18.75). The most common are the
equation. Irvine [1968a and b] has made a comparison of Eddington approximation [Irvine, 1968b; Shettle and Wein-
Romanova’s approach with Neumann series and a doubling man, 1970], various formulations of the “two stream” ap-
procedure and concluded that her method gives radiance proximation [Chu and Churchill, 1955; Sagan and Pollack,
accurate to at least 570. 1967; Coakley and Chylek, 1975], the delta-Eddington [Jo-
Weinman [ 1968] developed an approach in which the seph et al., 1976], and the delta-two-stream [Schaller, 1979]
diffraction peak of the phase function was represented by a approximations. These have been discussed and unified in
Gaussian distribution and the rest of the phase function by the studies by Meador & Weaver [ 1980] and by Zdunkowski
a Legendre expansion. He then solved for the small-angle et al. [1980], [also see Wiscombe & Grams, 1976 and Welch
scattering i, using a Neumann solution. lt was possible to & Zdunkowski, 1982].
do the integration analytically because of the use of the They showed that the different Eddington and two-stream
Gaussian distributions. The large-angle scattering I was then approximations all reduce the radiative transfer equation to
fotind by applying the discrete coordinate solution. a pair of coupled differential equations for the upwelling
Potter [ 1970] approximated the forward peak as a Dirac and downwelling irradiances:
delta distribution. He then considered this forward-scattered
radiation as not being scattered at all but as part of the dFi =
transmitted radiation and used a truncated phase function. –VI F$ + yzF~ + ~mylexp(–t/p,,) F,,
dt
This necessitated a resealing of the scattering coefficient and (18.76a)
corresponding y, the optical depth and the albedo for single
scatter, the resealing depending on the fraction of radiation
included in the forward peak that was eliminated. His ap-
proach gave quite good results except in the immediate
vicinity of the diffraction peak in contrast with Weinman’s
approach which is most accurate for the diffraction peak.
W iscombe [1977 ] discusses the deha M Method which where the values of the y, depend on the choice of approx-
provides a systematic procedure for representing the phase imation used.

18-57
CHAPTER 18
18.S.1. 11 Approximations for the Radiance. The dif- M= secant of the solar zenith angle 90
ferent approaches to solving the radiative transfer equation
M’ = secant of the line of sight angle 6
described above generally must be solved numerically on a
computer and permit tradeoffs between the accuracy of the t= total optical thickness per unit air mass (scattering
solution and the computer memory and time. These methods and absorption by molecules and aerosols
in principle can be made arbitrarily accurate by increasing
t, = scattering optical thickness
the number of terms kept in the expansions (such as Equation
(18.6g) or Equation (18.71)), by decreasing the step size F(W) = angular scattering intensity (molecular + aerosol)
used in numerically evaluating any necessary integrals, or per unit air mass
by increasing the number of photon histories following the
The scattering angle V is defined by
Monte Carlo Method.
There are also several techniques [Kaufman, 1979; 13av-
cos Y = cos 60 cos 6 – sin 60 sin 6 cos ($ – @o)
ies, ]980; and Hering, 1981] developed for solving the
transfer equation which are aimed at retaining reasonable
accuracy (10% to 2070) for the radiance field while mini- where @ – $0 is the angular azimuth difference between
mizing the computer time requirements. These methds have the sun direction and the line of sight. However, if tM or
a commonality in that they all use one of the two- tM’ becomes larger than 0.10, higher order scattering can
stream/Eddington type solutions for the radiances to allow no longer be neglected; this means that illumination of the
the integral in Equation ( 18 .65) to be evaluated analytically. scattering volume from the sky and the earth’s reflecting
They differ in the details of the solution beyond that point; surface becomes increasingly important. Higher order scat-
for examplej Davies [1980] and Hering [1981] both ex- tering becomes dominant for tM or tM’ >0.5, particularly
plicitly include the exact single scattered radiances in their for high ground albedo. This is particularly true for the
solutions whereas Kaufman [1979] does not. radiation leaving the top of the atmosphere (radiation to
space). The assumption of a homogeneous (constant mixing-
ratio) atmosphere is usually not valid and leads to errors.
me most recent revision of the LOWTRAN transmit-
18.5.2 Background (Sky) Radiance tance radiance codes, LOWTRAN 6, [Kneizys et al., 1983]
contains a subroutine which performs solar (or lunar) single
Atmospheric scattering of solar radiation is responsible scattering sky radiance calculation for either the atmospheric
for the skylight. For a pure molecular atmosphere, analytic models built into LOWTRAN or any user provided atmos-
solutions were given by Coulson et al. [1960]. They provide pheric data with nonhomogeneous vertical profiles. Since
tables of sky radiance and polarization including the Stokes LOWTRAN also allows a calculation of thermal emitted
parameters. Since then, much effort has been put into de- radiation from atmospheric molecules, the continuity from
veloping capabilities to calculate and model sky radiance visible scattered to infrared thermal emitted sky radiance is
for real aerosol containing atmospheres. These attempts wek provided in LOWTRAN for any lookangle, up- or down-
based on the different calculation methds for radiative transfer ward. One must always keep in mind that single scattering
discussed in the previous section. The intensity, wave- calctilations underestimate sky radiances; for a scattering
length, and polarization of skylight have been studied over optical depth of less than 0.7, the ratio of multiply to single
the decades to also derive information on atmospheric aer- radiation is in general less than 1.5. Single scattering is a
OsOis. good approximation, for lookangles near the sun.
For some applications, simplifying assumptions can lead Procedures for multiple scattering calculations are gen-
to useful results. As long as only single scattering is im- erally time consuming and complex, especially for cases of
@rtant and a homogeneous atmosphere is assumed (aero- strongly anisotropic aerosol scattering. Examples of Monte
sol/air mixing ratio constant with height), the angular dis- Carlo calculations for the distribution of the radiance of the
tribution of skylight at a specific wavelength can be obtained hazy atmosphere, seen from both the ground and from space
simply from are shown in Figure 18-88 [McClatchey et al., 1972]. Dotted
curves of the radiance (per unit solid angle and unit incident
B(M, M’, W) solar flux) are for the sun in the zenith and solid lines are
for the sun at 8 = 86.3°. In Figure 18-88a, the downward
= ‘(*) (e-’M
-e’M’)?
‘1877) or “transmitted” radiance general] y increases with decreas-
ing wavelength when the sun is in the zenith. Forward
where scattering causes the radiance to peak near the sun. Near
the horizon, the radiance increases again, except at short
wavelengths, and the albedo influence is large. At low solar
10 = extraterrestrial solar irradiance (W cm-2)
elevations, downward radiance is generally much smaller
B = sky radiance (W cm-z sr-’ ) especially near the zenith at long wavelengths, and the al-

18-58
 OPTICAL AND INFRARED PROPERTIES OF THE ATMOSPHERE

SOLAR AN TISOLAR
HORIZON ZENITH HORIZON
I I I I I I I I [ I I I
l“’’’’’”+
I
XO.4pm -z
l-” I I I 4
-..
‘.
-.~- .~~.”.” k
-1 -1
10 10
-------- ~/\+Q -.~-..-.< _\
\
k-
D +\ A
*-

-2
10
. 10-2 10-2 –

0r~l”o
I I I I I I I I 1 tllllllllli0 0 o
I I I I I ,0, I I
0
COS8 0 0 C’o: e do: 8
90 7084020

SOLAR A::;NT:;;R
HORIZON NADIR
[ 1 I I I I I 1 I I I
.0~------ ----

m
,/ ‘\\

-1 A A 0.7~m
-1
-1
10 10 ------ .2 ------
10 1
B
4
.\
\\ .8
= \ )

M “’..ow
\\,y
-2 _2
10 10-2 10
0
1- --i

0
;- 0
c’0: 8
Figw 18-88. Downward (a) and upward (b) tilance computed by Monte Carlo techniques. Solid curves are for the sun at 0 = 86.3”. The indicated
parameter is the surface albedo and A is the wavelength. Radiances are per unit solid angle and unit inci&nt solar flux.

bedo influence is significant only at short wavelengths. Since (a) (b)

,Ou , r I I I
the sky radiance values have been averaged over intervals
I“’’’’r”l
1 1

;NN A= I y
of O. 1 in cos 6 (that is, over angular intervals ranging from ----------- -.
5° to 250), the radiance near the sun actually should be much