Modern Structural Practice

A structural geology laboratory manual

for the 21st Century
Richard W. Allmendinger
v. 1.6.0, 2015-2016
Download most recent version from:
http://www.geo.cornell.edu/geology/faculty/RWA/structure-lab-manual/

Preface ....................................................................................ix
Chapter 1 ................................................................................1
Measurement: What and How
Introduction

Primitive Geometric Objects

Data Collection

Instruments Used in the Field

Measuring Lines and Planes
Uncertainties

Graphical Representation of Orientation Data

2
3
8

10

Maps
Stereonets

11
17

Computing

21

ExercisesChapter 1

24

Chapter 2 ..............................................................................29
Coordinate Systems and Vectors
Coordinate Systems

29

Vectors: A Review

32

Vector components, magnitude, and unit vectors

Vector Addition, Subtraction, and Scalar Multiplication
Dot Product and Cross Product

Geological Features in Cartesian Coordinates

Direction Cosines from Trend and Plunge
Trend and Plunge from Direction Cosines

Some Practical Applications Using Vectors

Mean Vector
Rake of a Line in a Plane
True Dip from Two Apparent Dips

ExercisesChapter 2

32
33
34

36
36
38

39
39
44
45

47

Chapter 3 ..............................................................................49

TABLE OF CONTENTS

Extracting Information from Geological Maps & Folds

Geologic Maps

49

Three Point Problems

50

Stratigraphic Thickness from Maps

54

Analysis of Folded Rocks

57

Cylindrical Folds
Dip Isogons
Folds in Map View

57
60
60

Structural Contour Maps

62

ExercisesChapter 3

64

Chapter 4 ..............................................................................69
Transformations
Introduction

69

Transformations of Coordinates and Vectors

70

Coordinate Transformations
Transformation of Vectors
A Simple Coordinate Transformation

70
72
74

Structural Geology Applications of Transformations

Stratigraphic Map Thicknesses
Down-Plunge Projection
Rotations

74
75
77
80

Summary

83

ExercisesChapter 4

84

Chapter 5 ..............................................................................87
The Orientation and Stress Tensors
Introduction

87

Matrices and Indicial Notation

88

Tensors

90

Tensors as Linear Vector Operators

Principal Axes of a Tensor
Tensor Transformations and the Mohrs Circle

MODERN STRUCTURAL PRACTICE

"iii

90
91
93

R. W. ALLMENDINGER 2015-16

TABLE OF CONTENTS

The Orientation Tensor

96

Least Squares Best Fit Fold Axis

Types of Line Distributions

97
98

The Stress Tensor

99

Cauchys Law
Mohrs Circle for Stress
Mean Stress & Deviatoric Stress

99
102
104

ExercisesChapter 5

105

Chapter 6 ............................................................................107
Faulting and Stress
Introduction

107

Fault Geometry Review

108

Terminology
Determining Slip from Piercing Points

108
110

Stress and Faulting

113

Failure in the Brittle Realm

Reactivation of Pre-existing Planes of Weakness
The Effect of Pore Fluid Pressure
Calculating the Normal and Shear Stress Vectors on any Plane
The Principal Stress Ratio

ExercisesChapter 6

113
115
118
119
124

126

Chapter 7 ............................................................................131
Deformation and Infinitesimal Strain
Introduction

131

Strain

132

One Dimensional Measures of Strain

Three Dimensional Deformation
Infinitesimal Strain

132
134
136

Some Geological Applications

142

Strain from GPS

Analyzing Brittle Faults

142
145

ExercisesChapter 7

149

MODERN STRUCTURAL PRACTICE

"iv

R. W. ALLMENDINGER 2015-16

TABLE OF CONTENTS

Chapter 8 ............................................................................151
Large Strains
Introduction

151

Comparison to Infinitesimal Strain

152

A Plethora of Finite Strain Tensors

Multiple Deformations
Mohrs Circle for Finite Strain in the Deformed State
Progressive Strain

Strain from Geological Objects

152
154
155
157

159

Finding Principal Axes

Deformed Spherical Objects
Three Deformed Lines Graphically
Three Deformed Lines with Inverse Methods

ExercisesChapter 8

159
160
163
166

169

Chapter 9 ............................................................................173
Rheology, Stress in the Crust, and Shear Zones
Introduction

173

Relationship Between Stress and Strain

173

Elasticity
Plasticity
Strain Rate and Viscosity

174
175
176

Environmental Factors

177

Deformation Mechanisms

179

State of Stress in the Earth

179

Frictional Slip on Preexisting Fractures

Power Law Creep
Stress Variation in the Lithosphere

180
181
183

Shear Zones

184

Brittle Shear Zones

Ductile Shear Zones
Displacement in Heterogenous Shear Zones from Foliations
General (2D) Shear

ExercisesChapter 9

MODERN STRUCTURAL PRACTICE

185
188
191
193

195

"v

R. W. ALLMENDINGER 2015-16

TABLE OF CONTENTS

Chapter 10 ..........................................................................199
Fault-Related Folds and Cross Sections
Introduction

199

Uncertainty in a Simple Depth Projection

200

Drawing Cross Sections

202

Preparation for drawing a cross section

Step-by-Step Section Construction

203
204

Fold-Fault Models

206

Fold Kinematics
Fold Type and the Propagation/Slip Ratio
Fault-bend Folds
Fault-propagation Folds
Extensional Drape Folds
Roll-over Anticlines and Listric Normal Faults

206
208
210
213
217
218

Balanced Cross-sections

220

The Fundamental Bases for Balanced Cross-sections

Line Length Balancing
Area Balancing
A Final Word About Balancing

220
223
226
229

ExercisesChapter 10

231

Chapter 11 ..........................................................................237
Structural Interpretation of Seismic Reflection Data
Introduction

237

Echo Sounding

238

Common Mid-Point (CMP) Method

241

Data Redundancy and Signal to Noise Ratio

Correction for Offset from the Source
Migration

241
242
245

Resolution of Seismic Reflection Data

246

Vertical Resolution
Horizontal Resolution

248
248

Diffractions

249

Artifacts

250

MODERN STRUCTURAL PRACTICE

"vi

R. W. ALLMENDINGER 2015-16

TABLE OF CONTENTS

Velocity Pullup/pulldown
Multiples
Sideswipe
Buried Focus

250
250
252
253

Structural Interpretation of Seismic Profiles

Truncations and Discontinuities
Kink Axes
Kink Axes and Growth Strata

253
254
256
256

Exercises Chapter 11

259

Chapter 12 ..........................................................................265
Solid Mechanics and Structural Geology
Introduction

265

The Mechanical Approach

266

Physical Principles
Constitutive Equations

267
269

Elasticity

269

Viscosity

271

Plasticity

271

Boundary Conditions and Initial Values

Some Simple Geological Examples

Mechanics of Thrust Belts

272

273
273

Hubbert and Rubeys (1959) Force Balance for Thrust Plates

273

Critically-Tapered Wedges (Dahlen, 1990)

276

Step-up Angle of Thrusts

278

Holes and Cracks: Some Important Results from Linear Elastic Fracture Mechanics 280
Circular Holes

280

Cracks

283

Final Thoughts: Simulation vs. Illumination

286

Exercises Chapter 12

288

Postface ................................................................................291
What youve accomplished and where to go next

MODERN STRUCTURAL PRACTICE

"vii

R. W. ALLMENDINGER 2015-16

TABLE OF CONTENTS

References Cited .................................................................295

Quick Reference ..................................................................301

MODERN STRUCTURAL PRACTICE

"viii

R. W. ALLMENDINGER 2015-16

PREFACE

Preface
The subtitle, "A structural geology lab manual for the 21st Century", as well
as the rationale for a new manual, needs a little explanation (and, perhaps, some
justification). There are several decent structural geology lab manuals out there and
the fact that a new one has not appeared in many years would suggest that the field
is mature and not changing rapidly, so why write a new one? Because the field of
structural geology, indeed all of geology, has changed enormously in the last
decade or two. Consider the following:
Large online digital databases for topography, satellite imagery,
scanned geologic maps, GPS data, earthquake data, etc have become readily available
Digital devices are seeing increasing use in the field. We collect
data using tablets and, increasingly, smart phones will replace analog compasses
Using programs, apps, and web apps to analyze data is now common place. Most structural geologists use such programs in preference to paper, whether drawing sections or plotting stereonets
The world has gone digital, but our teaching of structural geology has lagged
far behind: many of the exercises that we ask students to do in structure lab have
changed little over the last fifty years. We are not teaching our students how to take
advantage of this new world, nor how the tools that they will use in their profession
(e.g., stereonet programs, etc.) actually work internally. Digital means that, suddenly, lots of numbers are available to us. To analyze those numbers, requires some
math and computing. Most of the math is just additions and multiplications done
in a very systematic way as well as, of course, trigonometry. Students have much of
the needed math background already (vectors, calculus) so why don't we encourage
them to use those skills in their chosen major? Additional concepts for example
matrices can be learned in the context of problems that are actually of interest
to us.
Computing can take many forms but the lowest common denominator is the
humble spreadsheet program, which is whatModern Structural Practiceuses. Everyone
MODERN STRUCTURAL PRACTICE

"ix

R. W. ALLMENDINGER 2015-16

PREFACE

already knows the basics of using a spreadsheet and it is almost certain that everyone already has a spreadsheet on their computer. This enables us to implement algorithms without the overhead of learning a computer language. Students who already know how to program should be encouraged to use their skills rather than
the spreadsheet. For the rest, once you see the power of algorithms, you can decide
what flavor of computing to learn and use in your subsequent endeavors.
A more mathematical approach benefits all of us in another way: it helps lay
the necessary groundwork for understanding difficult concepts from mechanics.
This manual has several chapters devoted to stress and strain which, of course,
constitute only a very small part of the field of solid mechanics. However, the
background gained will help the motivated student to understand the entirety of
mechanics in subsequent studies.
The other thing that is almost entirely lacking from current lab manuals is
even a passing emphasis on uncertainty and error analysis. Modern Structural Practice
does not delve into these topics in great detail but does, at least, mention them.
Even something as seemingly well know as the elevation of a point on the surface
of the earth can vary by 15 m or more depending on the dataset used. This is particularly important when a student first encounters a geologic map or makes a measurement with a compass. If you can make a lot of measurements, which digital
tools now enable us to do, then you can begin to evaluate uncertainty.
Some might question whether this approach removes us from classical geology but quite the opposite is true: if you go digital, there is a huge amount of quantitative information that can be extracted from a geologic map or a digital data set.
In the past, the manual methods of extracting that information were so tedious that
it was too painful to explore the map. Some of the software tools that are used in
Modern Structural Practice make it trivially easy to extract quantitative 3D information
which can then be analyzed as the student and professor wish.
This manual is a work in progress: The current version has been used by my
own students to whom I am deeply grateful for their feedback and their willingness
to try something different. I also appreciate the time that Nstor Cardozo, Haakon
Fossen, David Pollard, and Ray Fletcher have taken to read over parts of this material, though they are neither responsible for any errors nor do they necessarily en-

MODERN STRUCTURAL PRACTICE

"x

R. W. ALLMENDINGER 2015-16

PREFACE

dorse the overall approach! I welcome comments and suggestions from faculty and
students. I know that some of the material is hard nothing ventured nothing
gained so the type of feedback that really helps me is to help identify sections
that are unclear or missing entirely. Students, please note that, while I would like to
help with the solution to the exercises at the end of each chapter, I really dont have
time to do so.
My current plan is that Modern Structural Practice remain a free resource, updated periodically. It is more important to me that it is widely used (nothing like
free to help make that happen) than that it is published. Besides, students already
pay too much for textbooks. Think of it as crowd-sourcing the revisions. The
downside of this self-published version is that the reader has to suffer with my
minimal page layout skills and limited artistic ability with the figures. An advantage
is that color can be used with abandon and it is still free!
This manual is for teaching and learning. Most of the concepts herein are
delivered more rigorously and in more detail in:
Allmendinger, R. W., Cardozo, N. C., and Fisher, D., 2012, Structural Geology
Algorithms: Vectors & Tensors: Cambridge, England, Cambridge University
Press, 289 pp.

Please refer to that published resource if you use the concepts here in your own research.
Rick Allmendinger
Ithaca, New York, August 2015

MODERN STRUCTURAL PRACTICE

"xi

R. W. ALLMENDINGER 2015-16

Chapter 1
Measurement: What and How
Introduction
The purpose of the laboratory portion of most structural geology courses is
to learn how to define quantitatively the geometry of deformed rocks. Furthermore, we would like to develop the tools to allow us to convert, easily and precisely,

between the geometries of rock bodies at two different times in their history. These
times may be the present, deformed, state and the initial, undeformed state or the
two times may capture only a small part of the rock's history. Finally, the lab should
lay the groundwork for at least a basic understanding of mechanics. All of these
goals require that the structural geologist be very precise about geometry and the
coordinate systems used to define the geometry, so that is where we will start.

Primitive Geometric Objects

Rocks are composed of linear and planar elements: bedding and secondary
foliations such as cleavage can be defined, at least locally, by planes whereas things

MEASUREMENT & DISPLAY

CHAPTER 1

like fold axes, mineral lineations, and paleocurrent directions are lines. It turns out
that, because there is one line oriented perpendicular to any particular plane, all
planes can also be represented by their corresponding perpendicular line which is
known as the normal or the pole to a plane.
The previous paragraph is an oversimplification because, in the vast majority
of cases, all of these lines have a direction in which they point. The paleocurrent,
for example, flowed towards a particular direction. The pole to bedding can point
in either the direction in which strata become younger or become older. The direction matters and thus our geological lines commonly have an arrowhead and tail;
that is, they are a geometric and mathematical quantity known as a vector. In
some cases, we care about the length, or magnitude, of our vector as in the case
of the displacement of a fault or the thickness of a stratigraphic unit. But in many
other cases, we only care about its orientation in space. We will use vectors extensively in subsequent chapters but before we can do that there are some more basic
things to address: how do we measure lines and planes and what coordinate systems
do we use, because it is impossible to talk about vectors without reference to a coordinate systems.

Data Collection
Instruments Used in the Field
Traditionally, a structural geologist used a geological compass/clinometer to
measure the orientation of features of interest. These are precision analog instruments that enable the geologist to measure the orientation of a feature of interest
to within a degree or less. The Brunton Pocket Transit (Fig. 1.1), most commonly
used in North America, excels at taking taking bearings horizontal angles measured with respect to North and inclinations angles in a vertical plane measured with respect to the horizontal over long distances; thus the word "transit"
in its formal name, though most people just refer to it as a "Brunton compass". In
Europe and other parts of the world, a Freiberg compass is more common. This
type of instrument is less suitable for sighting bearings over long distances but is
very efficient at measuring planar and linear features by placing the top or edge of
the compass flush against the rock.
MODERN STRUCTURAL PRACTICE

"2

R. W. ALLMENDINGER 2015-16

MEASUREMENT & DISPLAY

CHAPTER 1

Precision analog compasses have

been around for centuries but it is likely
that, within a few years, they will largely be
replaced by digital devices, most commonly
in the form of smart phone programs or
apps. Most smart phones contain accelerometers, gyroscopes, and electronic
magnetometers which enable apps to deFigure 1.1 Two ways of measuring
termine the exact orientation in space. Adthe orientations of geological lines and
ditionally, such devices also keep accurate
planes. Smartphone app (FieldMove
Clino from Midland Valley) on left and
track of the time and date and can detertraditional Brunton Compass on right.
mine their position every accurately using
the Global Positioning System (GPS) as well
as triangulation on cell phone towers and wireless networks. Thus, one can hypothetically capture a large number of measurements quickly using a compass app on
a smart phone. Several such apps are already available for iOS and Android operating systems and, with proper calibration, can yield excellent results. To date, most
smart phone compass apps for geological use have taken to mimicking analog compasses and thus use the same terminology and coordinate system. There is a great
deal of room for innovation in this space and it is likely that future apps will provide substantially more information. As a simple example, it is not possible to measure directly the pole to a plane with a traditional analog compass; the geologist
must do a simple calculation in the field or more likely back in the office to get this
value. A smart phone app, however can instantaneously calculate, record, and display this value in the field. The approach that this manual takes to structural geology lab will enable you to take advantage of data from this new digital world.

Measuring Lines and Planes

Both types of compasses and most smart phone compass apps measure bearings or azimuths clockwise with respect to the rotation axis of the Earth, i.e., north,
and the angle downward or upward from the horizontal. The previous sentence has
several important assumptions built in that are seldom explicitly stated: what is our
coordinate system and what are the conventions used to indicate whether a number
MODERN STRUCTURAL PRACTICE

"3

R. W. ALLMENDINGER 2015-16

MEASUREMENT & DISPLAY

CHAPTER 1

vertical plane

upward pole or
normal vector

horizontal plane
trend

rik
st

e
rak

plung

line A

dip

Figure 1.2 Orientations

of lines and planes. Lines
are defined by their azimuth or bearing the angle with respect to North
measured in a horizontal
plane and their plunge,
and angle measured
downwards from the horizontal in a vertical plane.
Likewise, planes are represented by the orientation
of a horizontal line within
the plane (the strike) and a
vertical angle (the dip). Alternatively, you can specify
the orientation of a plane
by its dip direction and dip,
or the trend and plunge of
its pole.

is positive or negative? Traditionally in structural geology, horizontal angles are

positive measured clockwise from North and vertical angles are measured positive
downwards. If you think about it for a minute, this is just the opposite of the convention that you use when you make a graph or plot data on a map. Graphs follow
an engineering convention where angles are measured positive up from the x-axis
(i.e., counterclockwise) and on maps, elevations are positive upward and depths are
negative! The point is, it doesn't matter what convention you choose as long as you
are clear and consistent. In a subsequent chapter, we will learn how to change from
one convention and corresponding coordinate system to another, an operation that
will help us solve lots of interesting problems.
A plane can be defined by the azimuth of a horizontal line contained within
it, known as the strike and the maximum angle measured downward from the horizontal to the plane, a quantity knows as the dip (Fig. 1.2). The azimuth of the dip
that is, the projection of the dip onto the horizontal is known as the dip direction or dip azimuth. The true dip direction is always 90 from the strike. The
most convenient way to think of the strike line is to imagine the plane half-submerged in a body of water. Because the surface of the water is necessarily horizontal, the waterline on the plane is a horizontal line in the plane (Fig. 1.3).

MODERN STRUCTURAL PRACTICE

"4

R. W. ALLMENDINGER 2015-16

MEASUREMENT & DISPLAY

CHAPTER 1

str

Figure 1.3 The strike

visualized as the water
line on an inclined plane
partially submerged in
water.

ike

water body

ke
tri

d
tren

pp
of a

aren

t dip

app

e!

true
dip

no

ap

e
ar

nt

di

in

ts
cu

pa

ra

lle

o
lt

ri k
st

true
dip

are

ip
nt d

Figure 1.4 True dip

and apparent dip. True
dip is the largest angle
between the horizontal
and the plane of interest (left block). Apparent dip is the angle between the horizontal
and the plane observed
in a vertical cut parallel
to the trend of the apparent dip (expanded
right block). The apparent dip in cuts parallel
to the strike is zero.

The dip deserves further consideration. Lets say that you are studying a vertical road cut in a region where the rocks have a uniform strike and dip. The orientation of the road cut will determine what you see: if the road cut is parallel to the
strike, then the strata will appear to be flat because the strike is a horizontal line in
the plane of bedding (Fig. 1.4, left block). If the road cut is perpendicular to strike,
then you will see the true dip of bedding. At any other orientation of the road cut,
you will see an apparent dip of bedding, which is always less than the true dip
(Fig. 1.4, right block). Problems involving apparent dips are extremely common in
structural geology. In subsequent chapters, we will learn quantitative methods for

MODERN STRUCTURAL PRACTICE

"5

R. W. ALLMENDINGER 2015-16

MEASUREMENT & DISPLAY

CHAPTER 1

15

15

ra

ke

Plane 1!
1
W
N 15 E, 45 W!
015, 45 W
W!
195, 45
45!
285, 45
45!

30

Format
Format!
Quadrant
Quadrant!
Azimuth
Azimuth!
Right-hand Rule
Rule!
Dip
Dip direction & Dip!

line A

45

Plane 2!
2
E
N 15 E, 30 E!
015, 30 E
E!
015, 30
30!
105, 30

Line A rake!
rake
SE
30 SE!
150
150!
150
150!

Figure 1.5 Conventions for

how to specify the orientations of two planes that
share the same strike line.
Specifying the rake of a line
is also given. Stick figure is
looking away from the
reader and holding out his/
her right hand. Avoid using
the quadrant format!

calculation of true dip from two apparent dips or apparent dip given true dip and
apparent dip direction but for now a little practice in visualization is in order.
The specification of a strike and dip for a plane is subject to a number of potential ambiguities that do not plague the dip direction and dip format. Which direction of the strike line do you use? How do you make sure that the plane dips in
the correct direction? Because in the past structural geologists have tried to accommodate all possible combinations of strike direction and dip direction that is
they did not settle upon a standard and stick to it a number of inefficient and
potentially error-prone formats have sprung up (Fig. 1.5). The quadrant format is
perhaps the worst offender: one must write down (correctly!) a combination of five
numbers and letters such as N 47 W, 22 S. The student is strongly encouraged
not to use this format even though you will need to know how to read it. On can
eliminate two letters simply by using the azimuth of the horizontal bearing: 313, 22
S. In this book, we will use either dip direction/dip or more commonly a format
known as the right hand rule (RHR). Using the RHR, one gives the strike direction such that the dip is to the right when looking in the direction of the strike az-

MODERN STRUCTURAL PRACTICE

"6

R. W. ALLMENDINGER 2015-16

MEASUREMENT & DISPLAY

CHAPTER 1

imuth (Fig. 1.5)1. Thus, we can further reduce the plane orientation given previously to: 133, 22.
The orientation of a line is also specified by a horizontal azimuth, or trend,
defined by the projection of the line onto the horizontal plane and the plunge, a
vertical angle measured from the horizontal downwards to the line (Fig. 1.2). Note
that for lines, or vectors, where the the arrowhead points upwards, their plunge has
a negative value. There are many times in geology when we are interested in lines
that are contained within planes: for example, a paleocurrent direction in sedimentary bedding or slickensides on a fault plane. In these cases, it is often most convenient to measure the angle between the strike of the plane and the line of interest,
which is known as the rake or pitch (Figs. 1.2, 1.5). Although strikes and trends are
measured in horizontal planes and dips and plunges in vertical planes, the rake is
commonly measured in an inclined plane. Traditionally, the rake has been measured from either end of the strike line, yielding two possible values, with the correct value identified by give the general quadrant direction. This is a recipe for confusion! The convention followed here is that the rake is always measured from the
given strike azimuth and thus varies between 0 and 180.
There are several ways to measure the orientations of lines and planes. The
two most common have their origins in the unique capabilities of the two types of
compasses mentioned earlier: Freiberg compasses, as well as certain modern versions of the Brunton compass, can measure dip and dip direction in a single operation. To measure a plane, the top cover of the compass is rotated so that it is flush
against the rock layer. An air bubble is brought to the center of the bullseye level on
the face of the compass, ensuring that the compass face defines a horizontal plane.
The geologist can then read the dip azimuth of the plane from the number that the
North needle points at on the face of the compass. The vertical angle that represents the dip is read directly off a scale on the side of the compass.
If one is using a traditional Brunton compass, a different approach is used:
First, the side edge of the main part of the compass is placed on the rock and ad1

Two other reasons for using the RHR will be seen in a later chapter: (a) the strike, true dip, and pole (SDP) define a
right handed cartesian coordinate system that is convenient for solving problems related to bedding; and (b) double
couple focal mechanisms of earthquakes are commonly giving using the Aki-Richards convention where the plane is
specified using RHR and the sign of the rake gives the sense of slip (positive for thrust faulting and negative for
normal faulting.

MODERN STRUCTURAL PRACTICE

"7

R. W. ALLMENDINGER 2015-16

MEASUREMENT & DISPLAY

CHAPTER 1

justed by rotating it until the bullseye level is level. One can then read the strike of
the plane as the number to which the north arrow points. It takes a second operation to measure the dip angle: the side edge of the compass is placed flush on the
rock so that the brass arm of the compass points either up or down dip. The lever
on the back side of the compass is then rotated until the bar level is horizontal and
the dip value (or apparent dip value) read off of the clinometer scale on the face of
the compass.

Uncertainties
Of course, rocks seldom define perfectly smooth planes or exactly straight
lines. They have irregularities that can be quite significant. Several field methods
have been developed for dealing with this fact. Students are commonly admonished
to place a field book or other non-magnetic planar object on the surface of the
plane to be measured to smooth out the irregularities. Using a more advanced field
technique, the geologist positions themselves so that their eye is in the plane of
bedding and they see the plane edge-on. They then set the clinometer of a Brunton-style compass to horizontal and sight through the hole in the mirror to located
a horizontal line in the plane. Finally, they take a bearing on the horizontal line in
the plane to get the strike. This method does a better job of smoothing out the irregularities at the scale of the entire outcrop and can also be used to measure a
planes orientation where one cannot physically access the plane (e.g., across a roaring river or a busy six lane highway). It does, however, take considerable practice to
master!
Without knowing why one is measuring the plane in the first place, it is impossible to know which is the best method to use. For a regional mapping job, one
commonly wants the best orientation at the scale of the entire outcrop. On the other hand, the variations, themselves, maybe the subject of study as for example
where one is trying to define asperities on a fault plane. However, the scale-invariant nature of irregularities in any natural surface or linear feature means that,
whatever the scale of the problem being addressed, the measurement of its orientation will not be perfect.

MODERN STRUCTURAL PRACTICE

"8

R. W. ALLMENDINGER 2015-16

MEASUREMENT & DISPLAY

CHAPTER 1

In addition to natural irregulariIn past times and even today, some structurties, there are other reasons why it is
al geologists acknowledge(d) that their strike
impossible to define the correct strike
and dip measurements were only good to
and dip of the plane. Two different geabout 5 and thus they would round their
dip values on their maps to the nearest five
ologists may use slightly different
degree increments. It is relatively easy to tell
methods for measuring a line or plane
on a published map when they did this even
and even the same geologist who makes
though it is seldom stated explicitly because
multiple measurements of the same
all of the dips on the map end in either a 0
plane at the same location will come up
or a 5! Needless to say, this is a poor substiwith a range of values because of
tute for actually determining the natural
seemingly trivial changes in measurevariation in our measurements, and even
stating that a dip is 335 is better than
ment methodology. Thus, ideally one
recording the dip as 35. Maps with strike
would make multiple measurements of
and dip value recorded to the nearest 5
a plane or a line and average them.
leave something to be desired!
That way, although we cannot state
with certainty what the correct orientation is, we could at very least come up with the best estimate of the orientation
and even use standard statistical techniques to calculate the uncertainty (standard
deviation) in the orientation. We will do exactly that in a following chapter.
The uncertainties we have discussed so far are of the type known as random, uncorrelated errors (Fig. 1.6a, b). There is no way to predict whether the
next measurement will be higher or lower than the previous one or, within some
range, by how much. They result from stochastic variations in natural features
and our own inability to measure them accurately. These errors are amenable to
calculation of a mean and a standard deviation (which we will do in the next chapter). The smaller the uncertainty, for these types of errors, the more reliable the result. Non-random or correlated errors are an entirely different beast (Fig. 1.6c,
d). Take for example, the geologist who unwittingly measured a bunch of strikes
and dips near a large magnetic body or with the declination set incorrectly on his
or her compass. In the case of nonrandom errors, even though the statistics might
be good (e.g., Fig. 1.6c), the results will be crummy.
This discussion brings up two important terms: accuracy and precision. Accurate means that our measurements are a reasonably good representation of the
MODERN STRUCTURAL PRACTICE

"9

R. W. ALLMENDINGER 2015-16

MEASUREMENT & DISPLAY

CHAPTER 1

Random
small

large
(b)

(c)

(d)

large

Systematic

small

(a)

Figure 1.6 Illustration of the

difference between random and
systematic errors. Each diagram shows a target with a
number of shots. Note that, in
this case, we know what the
right answer was (i.e., the
center of the target) and therefore the systematic error is
clear. In the natural features
that we study, we dont know
what the right answer is (i.e.,
there is no target) and thus we
can evaluate whether the random errors are large or small
but we cannot evaluate the systematic errors. Based on a concept by Taylor (1997).

true value, even if we can never know exactly what the true value is. Precise, on
the other hand, means that a measurement has been made with instruments that
record the observation to a large number of significant figures. Ideally, our measurements should be both accurate and precise (Fig. 1.6a). Even experienced scientists, however, often fall into the trap of thinking that because a measurement is
precise, it must also be accurate; that is a fallacy. In the case of nonrandom errors,
our measurement can be very precise but not very accurate (Fig. 1.6c). Earlier in
this chapter, we mentioned the advent of data collection via digital devices such as
smart phones. With any digital device, however, one must be careful to acknowledge that, although the answer produced is precise, its accuracy must not be assumed but proven. Digital sampling systems are highly subject to nonrandom errors!

Graphical Representation of Orientation Data

There are two fundamental ways that structural geologists display their orientation data with quantitative rigor. In maps, we are concerned about both orientation and the spatial relation of one feature to another. In the second type, stereo-

MODERN STRUCTURAL PRACTICE

"10

R. W. ALLMENDINGER 2015-16

MEASUREMENT & DISPLAY

CHAPTER 1

Horizontal Surface

Figure 1.7 Map as a projection onto the horizontal surface. In this case,
we show contour lines (lines of equal elevation) on an oblique view of the
topography and projected onto a horizontal surface

graphic projections, commonly called stereonets, we are just concerned with orientations, alone.

Maps
Geological maps are one of the most fundamental, original data sources in
our profession. They show the distribution, known and inferred, of rock units on
the surface of the earth, the nature of contacts separating them (e.g., stratigraphic,
unconformities, intrusive contacts, faults), and are the basic way that we evaluate
the distribution and geometry of deformed rock layers. Maps commonly show
strike and dip symbols in the layered units, and may show the orientations of secondary structures such as cleavage, lineations, and joints. A map is a formal scientific document and should be treated as such.
All maps are a projection of features that lie on the irregular, but approximately spherical, surface of the earth onto a horizontal plane (Fig. 1.7). This has an
immediate practical implication: when you measure a distance on a map (i.e., a
map distance), it is a horizontal distance not the longer distance that you actually
MODERN STRUCTURAL PRACTICE

"11

R. W. ALLMENDINGER 2015-16

MEASUREMENT & DISPLAY

CHAPTER 1

slo

pe

dis

tan

Change in
elevation

Figure 1.8 Map distance, slope distance, and slope angle. The slope distance is always larger than or equal to
the map distance.

ce

slope
angle
map distance

travel along a slope (i.e., the slope distance; Fig. 1.8). The slope angle is measured between the slope and the horizontal and is usually calculated by:
elevation
map distance

" slope angle = tan 1

where means change in. A related measure of steepness of slope is the percent
grade, which you are likely to have seen on highway signs in mountainous terrain:
elevation
100 .
map distance

" percent grade =

Humans are notoriously poor at estimating slopes. On a road over a steep mountain pass, you are likely to see a sign reporting 8% Grade Ahead. If you take the
arctangent of 0.08, that steep grade has a slope angle of just 4.6!
Beyond this practical implication, there is a more profound aspect to the fact
that maps represent projections: Although the Earth is approximately spherical, the
piece of paper or computer screen that displays the map is flat. There is inevitably
some distortion that occurs when we project a spherical surface onto a flat medium.
There are a myriad of map projections that accomplish this task depending on
the size of the map area and the purpose for which the map is needed (Snyder,
1987). For now, you mostly need to know that all maps represent tradeoffs of the
following factors:
Conformal preserves the same scale in every direction, locally,
thus maintaining the correct shape of the features

MODERN STRUCTURAL PRACTICE

"12

R. W. ALLMENDINGER 2015-16

MEASUREMENT & DISPLAY

CHAPTER 1

Equal Area preserves the area throughout, but distorts the shape
Equidistant depicts the correct distance between a point at the
center of the projection and points in any direction away from the
center
Equal Angle Shows true angles or bearings, locally
You cannot have all of the above in a single map! For example, a map that is both
conformal and equal area is impossible, as is a map that is equidistant and equal
angle. As we shall see below, stereonets have exactly the same limitations because
they are fundamentally the same thing: a projection of spherical data onto a flat
screen or paper.
Map coordinates are usually given in terms of longitude and latitude. You
are, of course, familiar with a globe with its lines of longitude running from pole to
pole and lines of latitude running around the globe perpendicular to its rotation
axis (Fig. 1.9). Lines of longitude are known as great circles; if you slice the globe

Figure 1.9 Left: the globe with lines of longitude and latitude. Right: a cut-away view of
the earth showing how the position of New York City is defined by a vector from the center of the Earth that pierces the surface of the sphere in a point

MODERN STRUCTURAL PRACTICE

"13

R. W. ALLMENDINGER 2015-16

MEASUREMENT & DISPLAY

CHAPTER 1

through the middle, the intersection on the surface is a great circle. A more concrete way of visualizing great circles is to think of a peeled orange or grapefruit:
the lines made by the segments of the fruit are great circles because all of the segments meet in the middle. You have probably heard people speak of great circle
routes as the shortest distance between two points. That is true because that line is
an arc or segment of a great circle. The lines of latitude are small circles because
they do not cut through the center of the globe. The complete coordinates of any
point on a globe, say New York City, are given by its longitude the angle between the great circle that goes through Greenwich, England and that which goes
through New York and the latitude the angle between the equator and a line
drawn from the center of the Earth to New York (Fig. 1.9 right). Lines through the center of a sphere intersect the surface at a point (e.g., New York City). Finally, you can see
that if you were to rotate the vector from the center to New York City about the rotation axis of the Earth, it would sweep out a cone and the intersection of the cone
with the sphere produces a small circle (Fig. 1.9), which is why they are sometimes
called conic sections.
Mapping has been revolutionized in the last decade or two by three trends:
(1) the advent of geographic information systems (GIS) of varying complexity, (2)
the availability of large digital data sets, especially digital topography and ubiquitous satellite imagery with resolutions of a few meters, and (3) small portable devices like tablets, smart phones, and ruggedized laptops equipped with GPS receivers for accurate positioning. Mapping has gone digital and with it, the need for
geologists who know how to manipulate and extract information from large collections of numbers has increased dramatically.
Ironically, as the power of the tools and data sets we use has increased, the
number of geologists creating new maps has plummeted. Many students in structural geology today will never get a chance to make a geologic map beyond whatever they experience in field camp. Even so, todays geologists need to know how to
extract quantitative information from the reams of paper maps published by state
and federal geological surveys for more than 100 years. A few years ago, this would
have meant spreading the map out on a drafting table with scale, compass, and
protractor in hand and carefully carrying out graphical constructions. Today, we
have a variety of software tools available to do these tasks and many government
MODERN STRUCTURAL PRACTICE

"14

R. W. ALLMENDINGER 2015-16

MEASUREMENT & DISPLAY

CHAPTER 1

30

15

30

15

Figure 1.10 Strike and dip symbol for bedding dipping 30 to the east-southeast and a lineation symbol showing the trend and plunge of a linear feature in
the plane of bedding in map view (left) and visualized as a block diagram
(right).

entities have made raster images of maps available online at little or no cost. We
will work with these scanned maps extensively in this course.
The way we represent the orientation of a planar feature on a map is with a
strike and dip symbol (Fig. 1.10). The exact nature of the symbol varies with the
feature that it represents, but they all have the same basic design: a long line drawn
parallel to the strike of the planar feature and a short tick mark indicating the dip
direction. The magnitude of the dip is commonly shown whereas the strike value is
not. Linear features, like paleocurrent directions, are shown with a lineation symbol which is an arrow with the value of the plunge shown at the arrowhead (Fig.
1.10).
In addition to strike and dip symbols on maps, there is another way to extract orientation data: because surface of the Earth is irregular that is, it has
topography the way planar units cross the surface of earth reflects their orientation. We call this the rule of Vs because, when seen in map view, contacts that
represent planar surfaces make the form of the letter V when they cross topography. Where a bed crosses a valley, the tip of the V points downstream when the
bed is inclined or dips downstream (Fig. 1.11) and it points upstream when the bed
dips upstream, is horizontal, or dips downstream at an angle less than the stream

MODERN STRUCTURAL PRACTICE

"15

R. W. ALLMENDINGER 2015-16

MEASUREMENT & DISPLAY

CHAPTER 1

200

0!

1000 m

Figure 1.11 The Rule of Vs depicted using

the same geologic map visualized in different
ways. Upper left: the map has been draped over
a digital elevation model and shaded with an
artificial sun in the northeast. Note how the
geological units, especially Js cross the valley
in the middle of the map, V-ing down
stream. Upper right: oblique block diagram of
the same map area, illustrating how the beds
dip down stream to the East (right). Lower left:
traditional three-point construction for calculating the dip. Two points along the lower surface of Js are at the same elevation (7600 ft.)
and thus determine the strike direction. The
third point is at a lower elevation. See Figure
1.12 for construction.

gradient. Vertical surfaces cross topography in a straight line, V-ing neither up nor
downstream.
Of course, to interpret correctly the orientation of beds using the rule of Vs,
one has to be able to determined hills, valleys, and which way the rivers flow. Most
geologic maps published by geological surveys contain topographic contours,
lines of equal elevation. If you were to walk exactly along a contour line, you would
go neither up nor down hill; walk perpendicular to the contours and you are going
in the direction of maximum slope. Modern mapping increasingly uses shaded
digital elevation models (DEMs) to portray topography but for quantitative
analysis students still need to know how to use topographic contour maps.

MODERN STRUCTURAL PRACTICE

"16

R. W. ALLMENDINGER 2015-16

MEASUREMENT & DISPLAY

CHAPTER 1

Figure 1.12 Classical three point

construction to determine the dip of
bed Js in Figure 1.11. The diagram is a
vertical plane oriented in the true dip
direction (parallel to the line labeled
1025.5 in Figure 1.11. Note that maps
are a projection onto a horizontal
plane.

7600 ft

Js

6800 ft

800 ft
= tan 1
= 38
1025.5 ft

1025.5 ft
(a) Equal Angle

(b) Equal Area

Figure 1.13 Two types of stereonets: (a) the equal angle, or Wulff, net; and (b) the equal
area, or Schmidt, net. All of the blue shaded small circles are the same size and shape on
the surface of the sphere (they all have a 10 radius). You can see that the equal angle net in
(a) is conformal (shape is preserved, that is they are all circles) but is not equal area or
equidistant (the 10 spacing gets bigger as you go from the center to the edge of the net).
Conversely, in the equal area net (b) the circles are distorted except at the center but they
all have the same area and the 10 spacing from the center outward is constant (equidistant).

Stereonets
Structural geologists use stereographic projections to display orientation data
when the spatial relation of the individual observations with respect to each other is
not important. A stereonet can also used to do complex calculations such as rotations, calculating lines of intersection, etc. by placing a sheet of tracing paper over
the grid of lines and rotating it about the center. That function has been largely reMODERN STRUCTURAL PRACTICE

"17

R. W. ALLMENDINGER 2015-16

MEASUREMENT & DISPLAY

CHAPTER 1

placed by computer stereonet programs that are far more powerful, accurate, and
precise in terms of computation and provide publication quality graphics automatically. Here, we emphasize only the use of stereonets to display orientation data.
Ones first view of a stereonet is instantly familiar: it looks like a globe or a
map of the entire Earth (Fig. 1.9 left). There are great circles, just like lines of longitude, running from pole to pole, and small circles, similar to lines of latitude, that
run across the globe from east to west. Of course, our picture of a stereonet, just
like the globe, on the screen is a projection of a sphere onto a flat surface and, as
shown in Figure 1.9, left, can only display half of the sphere. As with any mapping,
depicting a sphere on a flat surface inevitably involves distortion. Two types of
stereonets are commonly in use: the equal area (Schmidt) net and the equal angle
(Wulff) net (Fig. 1.13).
Stereonets used in mineralogy, structural geology, geophysics embody all of
the same concepts as our globe. Planar features are plotted as great circles and linear features plot as points and there are different types of distortion depending on
the type of projection used. In a similar fashion to our example of a globe, miner-

(a)

(b)

e
strik

ike

90

str

90

great
circle

a line pierces the

hemisphere in a
point

great
circle

Figure 1.14 The lower hemisphere projection (a) oblique view, and (b) top view.
Planes intersect the hemisphere as great circles and lines plot as points. is the true
dip, is the apparent dip, and is the angle between the strike and the trend of the
apparent dip. Note similarity of (b) with Figure 1.15b.

MODERN STRUCTURAL PRACTICE

"18

R. W. ALLMENDINGER 2015-16

MEASUREMENT & DISPLAY

CHAPTER 1

(b)

s tr i

ke

05

90

(a)

= 35
90

strike

str

ike

90

90

gr
o f e at
pla cir
n e cl e

great circle
of plane

Figure 1.15 Plotting a plane and doing an apparent dip calculation by hand,
by rotating a piece of tracing paper over the stationary grid of the stereonet.
(a) shows the paper rotated so that one can plot the great circle that represents
the plane, and (b) shows the tracing paper rotated back so that N on the paper
and on the grid are aligned.

alogists use an upper hemisphere projection whereas in structural geology and

geophysics we use a lower hemisphere projection (Fig. 1.14). The latter projection means that we are looking down into a cereal bowl and seeing the inside of the
bowl projected onto the horizontal plane through the center of the sphere (Fig.
1.14). Visualizing this geometry is fundamental to understanding how stereonets
display data.
Plotting stereonets by hand is tedious and imprecise. The basic idea is shown
in Figure 1.15. An equal angle, or more likely equal area, grid is mounted on a
piece of cardboard or some other firm but thin surface. The geologist then places a
sheet of tracing paper over the grid and put a push pin or thumbtack through the
tracing paper into a hole in the center of the firm backing to which the grid is
mounted. The tracing paper can now rotate about the center of the grid. After
marking North on the tracing paper, the geologist is ready to begin plotting. To plot
MODERN STRUCTURAL PRACTICE

"19

R. W. ALLMENDINGER 2015-16

MEASUREMENT & DISPLAY

CHAPTER 1

a plane, rotate the tracing paper by an amount equal and opposite to the strike of
the plane. For example, in Figure 1.15, the plane has a strike of 056 so the geologist rotates the tracing paper by 56 counterclockwise. Once this rotation is accomplished, one counts in on the east axis of the grid an amount equal to the dip value
and then traces by hand the corresponding great circle. Rotating the tracing paper
back to alignment with the grid so that north on the tracing paper is aligned
with north on the grid puts the great circle that represents the plane in its correct
position (Fig. 1.15b).
The great circles and small circles on the stereonet grid are read in degrees,
just like longitude and latitude are marked off in degrees on the globe. In Figure
1.15, the grid is drawn in 10 increments, whereas most paper stereonets are in 2
increments (most computer programs allow you to adjust the grid increment). If
you read in along the EW grid line (Fig. 1.15), youll see that you can measure from
0 at the edge of the net (known as the primitive or horizontal) to 90 at the center;
a vertical line will plot as a point exactly at the center of the net (a plunge of 90).
Angles measured in the plane (e.g., rakes) can be determined by counting off the
number of degrees along the great circle. For example, in Figure 1.15a, you can see
that the rake of the apparent
dip line (that has a plunge of
) is 40; likewise, the angle in
the plane between the apparent dip line and the true dip
line is 50.
If all that sounds complicated, it is but nonetheless
the procedure is a skill that
can be learned with a small
amount of practice. The
problems with this approach,
however, are two-fold: first,
several of the operations have
nothing to do with learning
how to interpret these displays
MODERN STRUCTURAL PRACTICE

Figure 1.16 Two planes and the line of intersection

between the visualized in three dimensions using the
Macintosh program Stereonet3D (Cardozo and Allmendinger, 2013). Programs like this can help if you
are having trouble relating a stereonet plot to the
three dimensional objects it represents.

"20

R. W. ALLMENDINGER 2015-16

MEASUREMENT & DISPLAY

CHAPTER 1

but only with the limitations of drawing great circles by hand. How does rotating
the tracing paper in a direction opposite to the rotation of the strike with respect to
north make this process clearer? Hint: it doesnt! Second and more importantly, few
structural geologists use paper stereonets now, anyway, as there are a number of
good programs available for Mac, Windows, Android, and iOS. The computer certainly isnt virtually rotating a piece of paper with respect to a grid; it is using simple but powerful mathematical algorithms and displaying the results on a stereonet
because thats what structural geologists are used to! Thus, in most of this book, we
are going to learn how to do structural calculations the way the computer does and
only use stereographic projects to visualize our results (Fig. 1.16).

Computing
Structural geologists use a large number of computer programs to speed up
their work. These programs may be written in a variety of languages: a few
decades ago the language was likely to be Fortran, Pascal, or Basic; today it is more
likely to written in some flavor of C, Python, Java, or Matlab. What is important,
however, is not the language but the algorithm behind the language. In this book,
we will use a very simple computing environment, the humble spreadsheet program, usually exemplified by Microsoft Excel. There are several advantages to
this approach:
1. You have probably used spreadsheets before so they are familiar.
That means that you can just focus on the algorithm and not worry
about language syntax, development environment, etc.
2. It is quite likely that you already have Microsoft Office installed
on your computer and thus you already have Excel. No additional purchase necessary!
3. A spreadsheet is naturally in the form of a table, or matrix, and
many of the things we would like to calculate in structural geology
are best thought of as tables of numbers.
Computers have a couple of traits that you will probably find either exhilarating or
frustrating: they are dumb and they are logical. Because computers are dumb machines, they are extremely literal: they will do exactly what you tell them to do, even
MODERN STRUCTURAL PRACTICE

"21

R. W. ALLMENDINGER 2015-16

MEASUREMENT & DISPLAY

CHAPTER 1

if it is not really what you wanted! Second, because they are logical, you have to understand
the logic of your calculation before you can tell
the computer what to do. Nonetheless, computing is a profoundly powerful skill for a scientist
to possess, and it is in that spirit that this book
emphasizes simple computing via spreadsheets
to solve structural geology problems. Once you
understand the power of the calculation, that is
the algorithm, you will have greater motivation
to learn a programming environment much
more powerful than a spreadsheet program.
A spreadsheet, like any program, has certain rules that one must follow. For those of you
who have never used a spreadsheet for anything
more than making a table, we offer the following
abbreviated set of rules:

Those of you who already know a

programming language should,
by all means, do the exercises in
this book using that language
rather than in a spreadsheet.
More advanced programing languages have many features that
simplify the overall task of implementing an algorithm. For example, a simple do/for loop in a
programing language is more
cumbersomely implemented in a
spreadsheet by manually copying
and pasting rows containing formulae and nested loops are even
more difficult.

To fill a cell with a computed value, the first character in the cell
must be the equals sign, =. Following the equals sign is the formula that you want to compute.
In formulas, one refers to the input data, whether a number or another computed value, via its cell address. The cell address is composed of a letter that indicates the column of the cell followed by a
number that indicates the row. For example, C3 refers to the value in column C, row 3. When entering a formula, after the equals
sign you can usually enter the cell address by just clicking on the
cell that you want.
When you copy a formula from one part of the spreadsheet to another, the cell addresses are adjusted automatically that is, the
cell addresses are relative. If you have a formula in C3 that uses the
value in A1, when you copy the formula to D4 (one column over
and one row down from the original), it will automatically adjust to

MODERN STRUCTURAL PRACTICE

"22

R. W. ALLMENDINGER 2015-16

MEASUREMENT & DISPLAY

CHAPTER 1

use the value in B2 (one over and one down from A1). Relative cell
addresses are one of the reasons why spreadsheets are so powerful.
Sometimes, however, you want to keep using the original value
even though you have copied the formula. For this you need an absolute cell address which is indicated with a $ in front of the column and/or row letter/number. If, in the above example, you
wanted to keep using the value in A1 when the formula was copied
to a different cell, you would type the address as $A$1 in the
formula.
Finally, for now, all computer languages, including spreadsheets, do
trigonometric calculations using radians rather than degrees. Useful formulae for converting from radians to degrees are:

"

180
= 57.2958;

3
180 = ; 270= ; 360=2
2

1 radian =

90 = ;
2

(1.1)

Say you want to calculate the sine of the angle in cell A3. In the
cell where you want the answer to appear (e.g., A4) you would type:
=SIN(RADIANS(A3))

To convert the value in A4 back to degrees you would enter

=DEGREES(ASIN(A4))

MODERN STRUCTURAL PRACTICE

"23

R. W. ALLMENDINGER 2015-16

MEASUREMENT & DISPLAY

CHAPTER 1

ExercisesChapter 1
1. For the following block diagrams, fill in the bed geometry on each exposed side
of the block. For each side, indicate whether one would see a true dip or an apparent dip of the bed if you observed that side of the block head on (looking
perpendicular to that face of the block).

(a)

(b)

30

90

(c)

(d)

10
30

MODERN STRUCTURAL PRACTICE

"24

R. W. ALLMENDINGER 2015-16

MEASUREMENT & DISPLAY

CHAPTER 1

2. For each of the map snippets below, state qualitatively which way the bedding
dips. Start by using the topographic contours to identify ridges and valleys and
then determine which way the creeks in the valley flow (i.e., the up- and downstream directions). Then, use the way that the stratigraphic contacts cross the
valley to determine the dip direction. For each map, it is sufficient to state that
the bedding dips to the east or the bedding dips to the SW.

MODERN STRUCTURAL PRACTICE

"25

R. W. ALLMENDINGER 2015-16

MEASUREMENT & DISPLAY

CHAPTER 1

3. Below are five orientations of planes. Convert them into right-hand rule format:
a. Convert from Quadrant to RHR format: (i) W 15 S, 61 S; (ii) N 56 E, 20
NW; (iii) S 30 E, 33 W;
b. Convert from Dip direction and dip to RHR format: (i) 237, 74; (ii) 099, 48
4. The instructor will set up a tilted layer or rock in the lab or will take you out to
a place where you can measure strikes and dips of a natural planar surface.
Each person in the class, including the instructor or other experienced geologist, should measure the plane the same number of times, at least10 or 15 times
would be good.
a. Describe the variation in your answers: How much variation in strike is
there? In dip? Does it appear easier to measure the strike or the dip more
accurately? What factors might determine which is easier?
b. Now compare your answers to those of your classmates and instructor. Can
you tell which answers are more precise or whose are more accurate? How?
5. Plot your results from Question 4 in a stereonet program as (a) great circles, and
(b) poles to the planes.

MODERN STRUCTURAL PRACTICE

"26

R. W. ALLMENDINGER 2015-16

MEASUREMENT & DISPLAY

CHAPTER 1

6. The following stereonet plots show planes and or lines. Describe the orientations in each and answer any additional questions written next under the plot.

what is the approximate orientation of the line of intersection of

the black and blue planes?

MODERN STRUCTURAL PRACTICE

"27

R. W. ALLMENDINGER 2015-16

MEASUREMENT & DISPLAY

CHAPTER 1

All of the questions below apply to the following table. Each row in the table is a
single datum, i.e. a single plane with one or more lines in it. Planes are given in azimuth format and lines in trend, plunge format. Asterisks show missing values; ***
is a missing azimuth or bearing, and ** is a missing plunge or dip. Some of the values are redundant: for example, once you know the true dip and dip direction, you
also know the strike and dip.
Datum

Strike & Dip of Plane

True Dip (T&P)

Apparent Dip(s)
(T&P)

050, **, *

***, **

090, 25

***, **, *

***, **

159, 34 and 270, 43

***, **, *

010, 48

321, ** and 090, **

7. Solve for the missing values for each of the three datums using a stereonet to
solve for each of the missing values in each row.
(a) In the second datum, above, you are given two apparent dips. Use a stereonet to determine the angle between those to apparent dips, measured in
the plane that contains them.
(b) Assume that the plane in datum three is a bedding plane. Restore the plane
to horizontal. What were the original bearings of the two apparent dip directions prior to the tilt of the bedding?
Rotate the plane in row 1 by returning the plane in row 2 back to its original horizontal position. What is the orientation of the plane in row one after you have performed this rotation?

MODERN STRUCTURAL PRACTICE

"28

R. W. ALLMENDINGER 2015-16

Chapter 2
Coordinate Systems and Vectors
Coordinate Systems
When we use a compass to measure features of interest to the structural geologist, we are implicitly using a spherical coordinate system defined by the rotation
axis and surface of the Earth. Another spherical coordinate system based on the

Earth is given by latitude and longitude. Geologists have successfully used these coordinate system for a couple hundred years; they are accurate and work well because the Earth is a nearly spherical body. As we saw in the last chapter, any rectangular coordinate system applied to the Earth must necessarily have some distortion. The most obvious example of this are maps where, via various algorithms, the
sphere of the earth is projected onto a flat piece of paper (or computer screen). A
map can display areas or angles correctly but not both, something that also applies
to Stereonets.
Even so, just as it is more convenient to carry around a flat map that can be
folded, rolled up, or displayed on a computer screen rather than a physical globe, it

COORDINATE SYSTEMS & VECTORS

CHAPTER 2

is commonly more convenient to do structural

calculations in rectangular Cartesian coordinates. Those are the main focus of this section. Before we get to Cartesian coordinates
commonly used in geology, a few formalities
are in order.

The three axes of a Cartesian coordinate system are commonly referred to as X, Y, and Z. We start
out the same way here because that
is most familiar, but when we start
talking about coordinate systems
more formally, we will switch to
using X1, X2, and X3, which are
more convenient for numerical calculations in a computer.

The axes of any graph have positive and

negative directions, of course, and most of us
are used to seeing two dimensional graphs
where the horizontal axis, commonly referred
to as the X axis or abscissa, is positive to the right and the vertical Y axis (the
ordinate) is positive upwards. This means that angles are measured positive in a
counterclockwise direction from the horizontal axis, just the opposite of how angles
are measured on a compass rose (clockwise from the top). But, what happens if we
add a third axis? How do we determine the positive direction for that axis?
Convention suggests that we should follow a right-handed naming convention: If you hold you hand so that your thumb points in the positive direction
of the first axis, your fingers should curl from the positive direction of the second
axis toward the positive direction of the third (Fig. 2.1). To follow this convention in
the case of our graph above where X is horizontal and points to the right, we have
two options: the new third, or Z, axis can be horizontal and point off the page

+X3 (+Z)

+X2 (+Y)

+X3 (+Z)

+X2 (+Y)

+X1 (+X)
(a) Right handed

MODERN STRUCTURAL PRACTICE

Figure 2.1 (a) Right-handed

and (b) left-handed coordinate
systems. As the disembodied
hand shows, for a right-handed system, if the thumb of the
right hand points in the direction of positive X1, then the
fingers curl from positive X2
towards positive X3.

+X1 (+X)
(b) Left handed

"30

R. W. ALLMENDINGER 2015-16

COORDINATE SYSTEMS & VECTORS

CHAPTER 2

+X1 = North

+X3 = Up

+X2 = North

Figure 2.2 Right-handed

coordinate systems commonly used in structural geology,
geophysics, and mapping.

+X2 = East

+X3 = Down
(a) Structural Geology

+X1 = East
(b) Geophysics, Topography

towards the viewer, or we can rename the vertical axis the Z-axis and add a new
horizontal Y-axis pointing into the page away from the viewer. Though neither are
wrong, the second option is the more common of the two and is, in fact, the coordinate system used in maps: the east direction is the first axis, the north direction
the second, and elevation (positive upwards) is the third. This means that a car
traveling southward is going in the negative direction. The point is, the labeling of
axes is not arbitrary but follows well-established conventions. This map convention of axes is used extensively in geophysics as well as mapping and is sometimes
referred to as an East-North-Up (ENU) system (Fig. 2.2).
Structural geologists, because of our reliance on magnetic compasses and
out predilection for treating angles measured downwards from the horizontal as
positive, use a different convention: We treat the North direction as the first axis,
the East direction as the second axis and the Down axis as the third axis, an NED
coordinate system (Fig. 2.2). This system is consistent with angles like strikes and
trends measured clockwise from North, and because down is positive, angles measured downwards from the horizontal (e.g., dips, plunges) are positive. Of course it
is easy to change between the two coordinate systems: A point that has coordinates
(100, 200, 350) in an ENU system would have coordinates (200, 100, and 350) in
a NED system.

MODERN STRUCTURAL PRACTICE

"31

R. W. ALLMENDINGER 2015-16

COORDINATE SYSTEMS & VECTORS

CHAPTER 2

Vectors: A Review
Most structural geology students have learned the basics of vectors in their
math courses. Well first review those basic concepts and then put that knowledge
to work because, as mentioned in the last chapter, most linear features that we
might wish to measure in structural geology are vectors.

Vector components, magnitude, and unit vectors

Because structural geometry is three dimensional, all of our vectors will have
three components. Each of the three numbers that define a vector refer to a specific
coordinate axis: for example, v2 (or vy) is the value of our vector, v, projected onto
the second axis of the coordinate system, X2 or Y. In a NED coordinate system, v2
is the projection of v onto the East axis, but in an ENU coordinate system, v2 is v
projected onto the North axis. Therefore, the numbers that define a vector depend
on the specific coordinate system. We write out vector as:

" v = v1 v2 v3 = vx vy vz

(2.1)

One of the most fundamental characteristics of a vector is its length or magnitude. Magnitude is a scalar quantity because it has no directional significance and
furthermore, it is the same in all coordinate systems. The magnitude is given by:

" v = v12 + v22 + v32

(2.2)

In two dimensions, you can see that the magnitude is calculated from the Pythagorean theorem which gives the length of a hypotenuse as the square root of the
sum of the squares of the two sides (Fig. 2.3). The extension to three dimensions is
straightforward.
But, what if we dont care about the magnitude? What if we are only interested in the orientation of our vector? In this case, we simply assume that the vector has a length of one and call it a unit vector. A special symbol is reserved for a
unit vector, a triangular hat or circumflex accent over the vector symbol: " v . Unit
vectors have a special property that makes them especially useful for orientations.
The projection of a unit vector onto a coordinate axis is just equal to the cosine of
MODERN STRUCTURAL PRACTICE

"32

R. W. ALLMENDINGER 2015-16

COORDINATE SYSTEMS & VECTORS

CHAPTER 2

X3

X2

(v

1
2 2
2

v = v12 + v22 = ( v12 + v

+v 2
2

1
2

v3

v
v2 = v cos

2
1

v1

v2

X2

X1

v1 = v cos

X1

(a)

(b)

Figure 2.3 The length or magnitude of a vector, v (in red) is calculated

using the Pythagorean Theorem in both two (a) and three (b) dimensions.

the angle that the vector makes with that axis (Fig. 2.3a). Thus, the components of
a unit vector are:
" v = cos cos cos

(2.3)

where is the angle with respect to the X1 axis, with X2, and with X3. cos,
cos, and cos are known as direction cosines and, as with the magnitude, the
proof is straightforward in two dimension and easily extended into three dimensions.

Vector Addition, Subtraction, and Scalar Multiplication

Vector addition or subtraction from another is no more complicated than
adding, or subtracting, the individual components (Fig. 2.4):

" v + u = v1 v2 v3 + u1 u2 u3 = ( v1 + u1 )

( v2 + u2 ) ( v3 + u3 ) (2.4a)

And likewise

MODERN STRUCTURAL PRACTICE

"33

R. W. ALLMENDINGER 2015-16

COORDINATE SYSTEMS & VECTORS

CHAPTER 2

uv

u+v

u
v

v
(a)

(b)

Figure 2.4 Vector addition (a) and subtraction (b) via the parallelogram
rule.

" v - u = v1 v2 v3 u1 u2 u3 = ( v1 u1 )

( v2 u2 ) ( v3 u3 ) (2.4b)

In scalar multiplication, each component of the vector is multiplied by

the same scalar quantity:

" av = a v1 v2 v3 = av1 av2 av3

(2.5)

The concept of a unit vector and vector addition allows us to define vectors
in a more explicit and elegant way: base or reference vectors are unit vectors
parallel to the three axes of the coordinate system. Any vector can thus be written
as the sum of the three base vectors multiplied by the scalar components of the
vector:

" v = v1i + v2 j + v3k

(2.6)

Dot Product and Cross Product

There are two operations that are unique to vectors and uniquely useful to
structural geology: the dot product (also known as the scalar product) and the
cross product (or vector product). The dot product of two vectors yields a single,
scalar number:

" u i v = v i u = u v cos = u1v1 + u2 v2 + u3v3

MODERN STRUCTURAL PRACTICE

"34

(2.6)

R. W. ALLMENDINGER 2015-16

COORDINATE SYSTEMS & VECTORS

CHAPTER 2

If u and v are both unit vectors, then the term " u v = 1 . We can then rearrange
equation 2.6 to give us an extremely convenient equation for the angle between two
lines, which we will be using a lot:
" = cos1 ( v1u1 + v2u2 + v3u3 )

(2.7)

The cross product is similarly useful because it yields a third vector which is
perpendicular to the first two vectors (Fig. 2.5). There will be many times in structural geology when we will want to calculate a vector which is perpendicular to two
other vectors. Notice that the direction that the third vector points depends on the
order in which you cross the first two. The formula for the cross product is:

"

v u = u v = v u sin
= ( v2u3 v3u2 ) ,

( v3u1 v1u3 ), ( v1u2 v2u1 )

(2.8)

If v and u are unit vectors, then the length or magnitude of the new vector is equal
to the sine of the angle between the two vectors. This brings up an important
point: the cross product of two unit vectors produces a third vector, perpendicular
to the first two, whose length varies between zero and one. The only time that the

Figure 2.5 The cross product produces a

third vector that is perpendicular to the first
two vectors. If u and v are unit vectors, then
the magnitude of the cross product is equal
to the sine of the angle between u and v. The
cross product follows a right-hand rule convention.

MODERN STRUCTURAL PRACTICE

"35

R. W. ALLMENDINGER 2015-16

COORDINATE SYSTEMS & VECTORS

CHAPTER 2

cross product produces a unit vector is when the first two vectors are at 90 to each
other. It turns out that this is a surprisingly common occurrence in some, but not
all, of the problems we will face in structure.

Geological Features in Cartesian Coordinates

Direction Cosines from Trend and Plunge
As mentioned before, all lines in geology are vectors so it makes sense that, in
Cartesian coordinates, the orientation of a line should be represented by direction
cosines. We will also represent planes by their pole or their normal vector. So, the
question becomes, how to convert from the familiar trend and plunge, or strike and
dip, to direction cosines? As shown in Figure 2.3, we will by convention represent
the angle that our vector makes with the first axis (North) by the Greek letter , the
angle with the second axis (East) with
and the angle with the third axis
N
(Down) as . We can get some idea
of the orientation of a vector simply

by looking at the signs of the direccos (+)!

cos ()!
tion cosines (Fig. 2.6). The cosine
cos (+)

function is positive (or zero) between90 (or 270) and 90 and

D
E
negative (or zero) between 90 and
cos ()!
cos ()!
270. Thus, a line that has a trend
cos ()!
cos (+)!
cos (+)
cos (+)
and plunge in the northeast quadrant
will have all positive direction cosines
because the angle that it makes with
each of the three axes is between
Figure 2.6 Lower hemisphere projection
zero and 90 (red point in Fig. 2.6). A
showing how the sign of each direction copoint in the southwest quadrant will
sine varies with quadrant. Note that and
are angles from N and E respectively meabe more than 90 from East and
sure in inclined (i.e., non-vertical) planes,
more than 90 from North and thus
whereas is measured in a vertical plane.
In the northeast quadrant where the red
cos and cos will both be less than
point is, all three direction cosines are posizero.
tive.
MODERN STRUCTURAL PRACTICE

"36

R. W. ALLMENDINGER 2015-16

COORDINATE SYSTEMS & VECTORS

CHAPTER 2

cos

trend

ge
plun cos

cos

90

e)

lung

p
cos(

Figure 2.7 Perspective view of the

geometry needed to derive the direction
cosines of a unit vector, v, given its trend
and plunge. The pink plane is the vertical
plane that contains the vector.

nge

plu

However, what we would really like are formulae for converting from spherical coordinates (trend and plunge or strike and dip) to direction cosines and back
again. To calculate these exact values, were going to have to do a bit of trigonometry because and are not measured in vertical planes (Fig 2.6). The basic geometry and the angles involved are shown in Figure 2.7. The easiest direction cosine to
calculate is cos because it is just equal to the sine of the plunge. From the relationships on the horizontal plane of Figure 2.7, you can see that:

"

cos = cos(trend)cos( plunge)

cos = sin(trend)cos( plunge)

(2.9)

With a similar set of derivations, one can also calculate the directions cosines for
the pole to a plane from the planes strike and dip. Table 2.1 gives the complete set
of equations for determining the direction cosines from the spherical coordinates.

MODERN STRUCTURAL PRACTICE

"37

R. W. ALLMENDINGER 2015-16

COORDINATE SYSTEMS & VECTORS

CHAPTER 2

Table 2.1
Axis

Direction Cosine

Lines

Poles to Planes
(strike & dip using RHR)

North

cos

cos(trend)cos(plunge)

sin(strike)sin(dip)

East

cos

sin(trend)cos(plunge)

cos(strike)sin(dip)

Down

cos

sin(plunge)

cos(dip)

Trend and Plunge from Direction Cosines

Doing the reverse calculation trend and plunge given the direction cosines
is a little less straightforward. The plunge is a piece of cake:

" plunge = sin 1 (cos )

If the trend varied only from zero to 180 there would be no problem, but the
trend in fact varies from 0 to 360. For any inverse trigonometric function arcsine, arccosine, or arctangent there are two possible angles between 0 and 360.
How do you choose the correct one? The answer is to use the signs of the direction
cosines to determine which quadrant the trend lies within.
There are several possible ways to solve for the trend. The simplest, though
with a caveat, is to observe from Equation 2.9 that:

"

cos sin(trend)cos( plunge)

=
= tan(trend)
cos cos(trend)cos( plunge)

cos
trend = tan 1

cos

(2.10)

From the graph in Figure 2.8 and by inspection of Figure 2.6, you can see that:
cos
cos

" trend = tan 1

" trend = 180 + tan 1

cos
cos

if

cos > 0

if

(2.11a)

cos < 0

(2.11b)

The caveat, of course, is that you have to test for the special case of cos = 0:

" trend = 90 if

MODERN STRUCTURAL PRACTICE

( cos = 0
"38

and cos 0 )

(2.11c)

R. W. ALLMENDINGER 2015-16

COORDINATE SYSTEMS & VECTORS

CHAPTER 2

" trend = 270 if

( cos = 0

and cos < 0 )

(2.11c)

There are several other ways of solving for the trend, but all involve testing the sign
of either cos or cos. Despite the extra test required for cos = 0, the tangent is
convenient because it involves a simple addition of 180 ( radians) when cos < 0.

Some Practical Applications Using Vectors

Okay, it is time to dust off your spreadsheet skills because even the basic
vector operations we have looked at so far enable us to do some very powerful
calculations. At the start, here, we will see how to set up the spreadsheet and what
formula to type into each cell. After this section, it will be assumed that you can do
so on your own and we will just concentrate on the equations, themselves.

Mean Vector
Suppose we have group of vectors, how do we calculate the average or mean
vector that represents the entire population of vectors? Our vectors may represent

180 ( radians)

Tangent (trend)

cos > 0

30

60

90

120

150

180

210

240

270

300

330

360

Trend ()

cos > 0

cos < 0

Figure 2.8 Plot of the trend versus the tangent of the trend.

MODERN STRUCTURAL PRACTICE

"39

R. W. ALLMENDINGER 2015-16

COORDINATE SYSTEMS & VECTORS

CHAPTER 2

(a)

(b)

(d)

(c)

r/5

Figure 2.9 The steps in a mean vector calculation, depicted graphically. (a) a
group of five unit vectors. (b) Adding the five vectors together and drawing the
resultant vector (red). (c) Normalizing the resultant vector. Note that the red vector is somewhat shorter than the five unit vectors. (d) the mean vector (unit vector
parallel to the resultant vector) calculated by dividing r by its own magnitude.

paleocurrent directions, paleomagnetism vectors, slip vectors on a fault surface, or

any other geological feature that can be represented as a vector. This first example
involves nothing more than vector addition, with some scalar division at the end.
Even so, this example is impossible to do via a paper Stereonet or by any other approach. For the very final step, we will see how to calculate the uncertainty about
the mean vector.
To determine the mean of any scalar measurement, say temperature, we
simply add together all of the measurements and divide by the number of measurements. The procedure is exactly the same with unit vectors, except now you
sum each of the three scalar components of the vector with the corresponding
components of every other vector. The resultant vector, r, is what you get when
you add several vectors together. Graphically, vector addition is just like placing the
tail of one vector at the head of a previous vector; the resultant vector is drawn
from the tail of the first vector to the head of the last vector (Fig. 2.9b). The length
of the resultant vector is, itself, a measure of the preferred orientation of the
group of unit vectors. If all of the unit vectors had exactly the same orientation,
then the magnitude of resultant vector, |r|, would be equal to the number of vectors (if there were five vectors, |r| = 5). When comparing different groups of unit
vectors with different numbers, it is convenient to normalize r by dividing each of
its components by the number so that r will always vary between zero and one (Fig.

MODERN STRUCTURAL PRACTICE

"40

R. W. ALLMENDINGER 2015-16

COORDINATE SYSTEMS & VECTORS

CHAPTER 2

Figure 2.10 Simple spreadsheet to calculate the mean vector for five vectors
whose orientation in spherical coordinates is shown in columns A and B, rows 2-6.
Areas with calculated values shown in blue shading and the answer in yellow. The
formulae for those cells are given in Table 2.2. See text for discussion.

2.9c). Eventually, though, we will have to convert r

into its equivalent unit vector, " r (Fig. 2.9d), so that
we can extract the mean vector orientation. Now,
lets do the same thing numerically.

Spreadsheets and full-fledged

computer programs are not
very useful if they are not well
organized and clearly labeled
(i.e., commented). People who
use or modify the program
after you, not to mention your
Professor and Teaching Assistant, will thank you!

The completed spreadsheet to calculate the

mean vector for five lines is shown in Figure 2.10.
The original data are outlined in the box beneath
the trend and plunge and all data calculated by
formulae are in blue. Note that calculating the
magnitude of the unit vector from the original direction cosines is redundant: we know that they have a magnitude of one already.
Nonetheless it is a useful check on your calculations to make sure that you didnt
enter any equations incorrectly. For this particular data set, you can see that the
normalized resultant vector magnitude is 90% of a unit vector indicating a fairly
strong preferred orientation.
The equations in each of the cells in row 2-11 is given in Table 2.2 (the
equals sign, =, has been omitted from in front of each equation). Note that, for
rows 2-6, you would use the appropriate row number in each of the equations; all
you have to do is fill in the equations for row 2 and then copy them to rows 3-6 and
the spreadsheet program will automatically adjust the row numbers. Line 8 does
the vector addition simply by summing the columns of individual direction cosines.
MODERN STRUCTURAL PRACTICE

"41

R. W. ALLMENDINGER 2015-16

COORDINATE SYSTEMS & VECTORS

CHAPTER 2

Finally, note that we have used an absolute cell reference, $G$9, in row 11 because all three formulae in columns D, E, and F reference the same cell in column
G. If we had not used an absolute reference, then when the formula in D11 was
copied to E11, the spreadsheet would have adjusted it automatically to read E9/H9
not what we want at all.
Table 2.2
row(s)

D (cos(alpha))

E (cos(beta))

F (cos(gamma))

G (Magnitude)

COS(RADIANS(B2))
*COS(RADIANS(C2))

SIN(RADIANS(B2))
*COS(RADIANS(C2))

SIN(RADIANS(C2))

SQRT(D2^2 + E2^2 + F2^2)

SUM(D2:D6)

SUM(E2:E6)

SUM(F2:F6)

SQRT(D8^2 + E8^2 + F8^2)

D8/5

E8/5

F8/5

SQRT(D9^2 + E9^2 + F9^2)

D9/$G$9

E9/$G$9

F9/$G$9

SQRT(D11^2 + E11^2 + F11^2)

2-6

11

To get the trend of the mean vector in row 11, column B, we need to use a
conditional statement because we have to check the sign of cos as described
above. The formula for the trend is:
=IF(D11>=0, DEGREES(ATAN(E11/D11)), DEGREES(ATAN(E11/D11))+180)

Between the parentheses following the IF are three statements separated by commas
(this is how spreadsheets do these things): the first is the logical test (i.e., if the
value in D11 is greater than or equal to zero), the second statement is what to do if
the logical test is true, and the third is the calculation if the logical test is false. This
formula takes a shortcut because it does not check for whether or not cos is equal
to zero, which you would clearly want to do in any general purpose program. The
formula for the plunge in C13 is:
=DEGREES(ASIN(F11))

In both formulae, DEGREES converts from radians to degrees because the inverse
trigonometric functions always return the angle in radians.
Like any calculation that involved determining a mean, we can also determine the statistical standard deviation. For three dimensional mean vector data,
the standard deviation is given by:

MODERN STRUCTURAL PRACTICE

"42

R. W. ALLMENDINGER 2015-16

COORDINATE SYSTEMS & VECTORS

CHAPTER 2

N R 1 N 1
" cos p = 1
1
R 1 p

(2.12)

where p is the probability, N is the number of vectors, R is the magnitude of the resultant vector. p is the apical angle of the cone of uncertainty at probability, p.
For example, if you wanted to calculate the probability at two standard deviations,
or 95%, you would set p = 0.95. You can see that, for the perfect case of all vectors
having the same orientation, cosp = 1 and p = 0 because N=R. Though not
terribly useful with so few measurements, the 2 (or 95) cone of uncertainty
around our mean vector in the data set shown if Figure 2.10 is 28.2: we are 95%
certain that the correct answer lies within 28.2 of the mean vector that we calculated.
What we have just seen is undeniably powerful: it is exactly the calculation
that geologists who study paleomagnetism make when they determine a paleomagnetic pole. However, be careful: the mean vector calculations is very literal. That
raises a common pitfall: the case where the direction of the vector doesnt matter
and we are plotting exclusively in the lower hemisphere. We commonly think of
lines that plot on the opposite sides of a stereonet say, 090, 05 and 270, 05 as
having very nearly the same orientation, that is they are very nearly parallel. However, the mean vector calculation treats them as vectors pointing in opposite directions as shown in Figure 2.11. When we add the two together, the resultant vector is
vertical and very short. This is the right thing to do if we are summing anything
where the direction matters but it is not what we want if, say, intersection lineations
-u

v - u
v
u
u + v

MODERN STRUCTURAL PRACTICE

"43

Figure 2.11 Oblique

view of a lower hemisphere projection showing the result of adding
to vectors with shallow
plunges on opposite
sides of the net (red vectors). When the vectors
have no directional significance, we usually
want the blue vector.

R. W. ALLMENDINGER 2015-16

COORDINATE SYSTEMS & VECTORS

CHAPTER 2

are being analyzed. For these lines that have no directional significance, there are
other methods but well need to develop some additional concepts before we get
there.

Rake of a Line in a Plane

Recall that the rake of a line is just the angle between the line and the strike
of a plane that contains the line. Here, we are going to follow the format that the
rake is always given with respect to the right-hand rule strike. The dot product
(Eqn. 2.7) will serve us nicely for this calculation. One line is given by its trend and
plunge; the second line is the strike line which has a trend equal to the strike and a
plunge of zero.
The spreadsheet for this calculation is shown in Figure 2.12. The formulae
for the direction cosines of the line in row 2 are straightforward and the same as
used in the previous example. The plane in row 3 requires a bit more explanation:
we are treating the strike as a horizontal line in the plane so we dont actually use
the dip value at all! Instead, the direction cosines are calculated for a line with a
trend equal to the strike and a plunge of zero (because the strike is a horizontal
line you get the idea). That is why the third direction cosine for the strike in F3 is
equal to zero. If a line has zero plunge, the angle that it makes with the down axis,
gamma, is 90 and cosine of 90 is equal to zero.
The formula for the rake in cell B5 is just:
=DEGREES(ACOS(D2*D3 + E2*E3 + F2*F3))

Figure 2.12 Spreadsheet for calculating the rake of a line in a plane. Note that
the direction cosines in row 3 are for a line with a trend of 053 and a plunge of 00.

MODERN STRUCTURAL PRACTICE

"44

R. W. ALLMENDINGER 2015-16

COORDINATE SYSTEMS & VECTORS

CHAPTER 2

As before, we dont really need and dont even use the magnitude column but it is a
useful check to make sure that our formulae are entered correctly.

True Dip from Two Apparent Dips

sin

There are many times when we observe in nature two cuts or slices through a
plane, but neither cut is in the true dip direction (i.e., perpendicular to strike). This
can happen in quarry walls, road cuts, mine tunnels, etc. In these cases, we may
have to calculate the orientation of the plane that is, the strike and true dip
from the available observations of apparent dip. In terms of the concepts developed in this Chapter, we know two vectors that lie in the same plane and want to
calculate the orientation of the plane, itself. You may recall that one way to specify
a plane is by the normal to the plane or the pole (Fig. 2.13). The pole is, by definition, perpendicular to every line in the
plane, including the two apparent dip
n
lines, so our solution to this problem
u v
will use the cross product (Eqn. 2.8, Fig.
2.5).
As before, the procedure that we
will follow is similar to what weve done
in other examples in this chapter:

1. Calculate the direction cosines

of the two apparent dip lines
from their trends and plunges

v u
n

2. Calculate the cross product using Equation 2.8

3. If the cross product gives us an upper hemisphere pole, which you
can determine from the sign of the
third direction cosine, then convert
it to a lower hemisphere pole by
multiplying all three direction
cosines by 1.
MODERN STRUCTURAL PRACTICE

"45

Figure 2.13 Any two vectors in a plane

can be used to find the pole to the plane
using the cross product. As shown, the
order in which you cross the vectors determines whether the pole is upward or
downward pointing. The cross product
of any two unit vectors has a length
equal to the sine of the angle between
the two vectors. You must calculate the
unit pole vector, n, parallel to the cross
product before converting it back to
trend and plunge.

R. W. ALLMENDINGER 2015-16

COORDINATE SYSTEMS & VECTORS

CHAPTER 2

4. Convert the vector product to a unit vector by dividing by the magnitude of the vector (which, remember, is equal to the sine of the angle
between the two vectors).
5. Determine the trend and plunge of the pole from the direction
cosines of the unit vector calculated in step 4.
6. The right hand rule strike of the plane will just be equal to the pole
plus 90 and the dip will be 90plunge of the pole.
The spreadsheet that carries out these steps is shown in Figure 2.14. The two lines
in rows 2 and 3 are processed just the way they were in the previous examples. The
cross product in row 5 is an upward pointing vector as shown by the fact that the
third component is negative. To reverse the direction of the vector so that it points
into the lower hemisphere, each component is multiplied by 1 in row 6. The direction cosines for the unit vector parallel to the cross product are calculated in row 8
by dividing the components in row 6 by the magnitude of the vector. The trend of
the pole to the plane is calculated from these direction cosines by using the same
formula that we used in the mean vector example:
=IF(D8>=0, DEGREES(ATAN(E8/D8)), DEGREES(ATAN(E8/D8))+180)

As before, we havent checked for the special case of cos = 0. The resulting trend
of the pole is a negative number, which translates to an azimuth of 308.35; we
dont bother to change it because we need to add 90 to it, anyway, to get the righthand rule strike, which is done in row 10.

Figure 2.14 Spreadsheet for calculating the orientation of a plane from two
apparent dips, labeled line 1 and line 2. The strike of the plane is given using
right-hand rule format.

MODERN STRUCTURAL PRACTICE

"46

R. W. ALLMENDINGER 2015-16

COORDINATE SYSTEMS & VECTORS

CHAPTER 2

ExercisesChapter 2
1. Practice thinking about coordinate system labeling.
(a) Devise three different right-handed coordinate systems that start with West
as the positive X1 axis.
(b) Likewise, determine three different left-handed coordinate systems that start
with South as the positive X1 axis.
2. If you multiply a vector times 3, does the resulting vector have a magnitude
three times greater than the original vector? Prove your answer.
3. From the poles to bedding that you calculated for your measurements in exercise 1.4 in Chapter 1, calculate the mean pole vector and its uncertainty (Eqn
2.12).
You should recognize the table below; they are the same problems as in Week 1 of
the course. The table below lists a series of orientations. Planes are given in azimuth format and lines in trend, plunge format. Asterisks show missing values; ***
is a missing azimuth or bearing, and ** is a missing plunge or dip. Some of the values are redundant: for example, once you know the true dip and dip direction, you
also know the strike and dip.
Datum

Strike & Dip of Plane

True Dip (T&P)

Apparent Dip(s)
(T&P)

050, **, *

***, **

090, 25

***, **, *

***, **

159, 34 and 270, 43

***, **, *

010, 48

321, ** and 090, **

4. Solve each of these problems with a spreadsheet program or Matlab using the
cross product or dot product as necessary. You will need to convert all of the
values to direction cosines, do the calculations, then convert the values back to
trend and plunge or strike and dip. Remember that the cross product does not
normally give you a unit vector so you must convert your answer to a unit vector before you can get its orientation in trend and plunge.
MODERN STRUCTURAL PRACTICE

"47

R. W. ALLMENDINGER 2015-16

COORDINATE SYSTEMS & VECTORS

CHAPTER 2

MODERN STRUCTURAL PRACTICE

"48

R. W. ALLMENDINGER 2015-16

Chapter 3
Extracting Information from Geological Maps & Folds
Geologic Maps
Geologic maps are one of the most fundamental types of geologic documents and yet they are a strange mixture of data and interpretation; it is often not
easy to tell one from the other. Published maps are all solid colors and bold, confident lines; they look like they are representing data. In this Chapter, we will build
on the visualization that we began in Chapter 1 and the vector methods from

Chapter 2 to introduce you to some of the types of information that you can extract from a geologic map.
Most of the features that we deal with at a map scale are approximately planar features stratigraphic contacts, faults, dikes, etc. and, more often than
not, they are not horizontal. At a regional scale, the surface of the earth where we
make most of our observations is pretty planar but, at more detailed scales, topography is very irregular. Mapping the outcrop patterns of these planar features is basically an exercise in locating the line of intersection between the planar feature
and the irregular 3D surface of the Earth.

GEOLOGICAL MAPS

CHAPTER 3

Three Point Problems

Some Common Map Symbols

Contact

A geologist cannot always go the the

field and put a compass on the rock to determine a strike and dip. Even when s/he
can do so, we frequently want to know the
orientation of a planar rock unit at a scale
larger than the outcrop scale. That is, we
want to know the average orientation at a
map scale. Fortunately, our geologic maps
provide sufficient information to enable the
structural geologist to determine orientation
independent of outcrop and compass.

Fault
Thrust fault, saw teeth on upper plate
Fault, bar and ball on downthrown side

Anticlinal trace of axial surface

synclinal trace of axial surface

Overturned anticline trace of axial

surface with trend and plunge of hinge
Overturned syncline trace of axial surface
Strike and Dip of Bedding, of overturned
bedding
Strike and Dip of Cleavage or foliation,
of joint

The orientation of a plane can be determined if we know the positions of three non-collinear points within the plane.
Because the surface of the earth has variable topography, we can commonly find
three points on a plane at different topographic elevations. Geologic maps com-

7600 ft

slo

pe

Js

dis

tan

ce

800 ft
= tan 1
= 38
1025.5 ft

6800 ft
map distance = 1025.5 ft

200

0!

1000 m

Figure 3.1 The classical three-point problem where a line is draw between two points of
equal elevation on a planar surface, yielding the strike (left). On the right is the construction in a vertical plane parallel to the line labeled 1025.5 (feet) on the map for calculating
the dip using the elevation values (in feet) and map distance.

MODERN STRUCTURAL PRACTICE

"50

R. W. ALLMENDINGER 2015-16

GEOLOGICAL MAPS

CHAPTER 3

monly depict topography with contour lines and, even if they dont, modern digital
elevation models are available for most of the earths surface and the elevation of
any point can be determined via the Internet.
Figure 3.1 shows the classical way of determining strike and dip from a geologic map with topographic contours. This simple method takes advantage of the
fact that a line connecting two points of equal elevation along a mapped contact of
a planar feature define the strike of a plane. The dip can then be calculated from a
third point at different elevation from a simple geometric construction. Both the
map distance, and the elevation difference to the third point, perpendicular to the
strike, can be read directly off the geologic map.
A more general method takes advantage of the vector methods weve just
learned in the previous chapter. It is more flexible because all three points can be at
different elevations and can be used wherever we have spot elevations but no topographic contours (e.g., in Google Earth). We use the cross product of two vectors in
a plane just like we did in the previous Chapter to determine the true dip from
two apparent dips but this time we are not using unit vectors describing orientations of two lines in the plane, but are using position vectors whose magnitude is
much greater than one (Fig. 3.2).
A position vector is a line connecting a point in space to the origin of the coordinate system (P1, P2, and P3 in Fig. 3.2). The coordinates of the position vector
are just the scalar components of the vector projected onto the coordinate system
axes. In the case of our geologic map, we could use the UTM coordinates (eastings

Figure 3.2 Three points in a plane can be

used to calculate the orientation of a plane.
The coordinates of the three points are the
coordinates of the three position vectors,
P1, P2, and P3. To get the vectors that lie
within the plane, v and u, we use vector
subtraction as described in the text. The
pole to the plane is calculated from v u.

v
P2
u
P1

P3

MODERN STRUCTURAL PRACTICE

"51

R. W. ALLMENDINGER 2015-16

GEOLOGICAL MAPS

CHAPTER 3

and northings) plus elevation to define the position vector in an East-North-Up coordinate system. Or, we could use any other local Cartesian coordinate system.
To calculate the pole to the plane, we will use the cross product of two vectors in the plane, v and u. These two vectors can be calculated from the position
vectors using vector subtraction. The complete sequence of steps is given, below:
The first step is to subtract the position vectors to get v and u. Note that at
this point, we are working in an ENU coordinate system so the subscripts in the following equation correspond to the axes of our coordinate system: 1=E, 2=N, and
3=U.

v = ( P21 P11 )

( P2 2 P12 ) ( P2 3 P13 )

u = ( P31 P11 )

( P32 P12 )

(3.1)

( P33 P13 )

To convert these to a lower hemisphere NED centric coordinate system which

we will need in order to calculate our orientations we switch the order of the
first two components of the vector and multiply the third by 1:

v = ( P2 2 P12 )

( P21 P11 )

( P2 3 P13 )

u = ( P32 P12 )

( P31 P11 )

( P33 P13 )

(3.2)

The cross product is defined as:

" v u = ( v2u3 v3u2 )

( v3u1 v1u3 ) ( v1u2 v2u1 )

= s
1

s2

s3

(3.3)

The cross product gives us the pole to the plane but we need to convert it to a unit
vector before it can be transformed back into geographic orientations like trend
and plunge or strike and dip. We start by calculating the magnitude of cross product, smagn:

MODERN STRUCTURAL PRACTICE

" smagn = s12 + s22 + s32

"52

(3.4)

R. W. ALLMENDINGER 2015-16

GEOLOGICAL MAPS

CHAPTER 3

And now we calculate the unit pole vector, " p , by dividing each component of s by
its magnitude, smagn.

" p = p1 p2 p3 = 1
smagn

s2

s3

smagn

smagn

(3.5)

If the third component is negative, i.e., p3 < 0, then the unit pole vector we have
calculated points into the upper hemisphere. To covert to the lower hemisphere in
this case, multiply each component by 1.
Were now ready to convert our unit pole vector back to trend and plunge.
The plunge is straightforward because it is just the arcsine of the p3 component of
the pole to the plane:
" plunge = sin 1 ( p3 )

(3.6)

As we saw in Chapter 2, the trend is a function of p2 and p1 and the sign of p1. If p1
0 then you use the equation on the left, below; otherwise use the equation on the
right:

p2
p1

" trend = tan 1

p
trend = 180 + tan 1 2
p1

or

(3.7)

To get the right-hand rule strike, just add 90 to the trend; the dip is 90plunge.

Figure 3.3 A
three point calculation in GMDE.
This program can
also calculate the
uncertainties;
those shown here
are given a horizontal and vertical
uncertainty of 40
ft.

MODERN STRUCTURAL PRACTICE

"53

R. W. ALLMENDINGER 2015-16

GEOLOGICAL MAPS

CHAPTER 3

Figure 3.4 The spreadsheet to

do the same calculation as shown
in Figure 3.3. Note that the switch
from an ENU to an NED coordinate system occurs in row 11.

The program GeolMapDataExtractor (GMDE) can do this calculation for

you automatically (Fig. 3.3), and the equations we have just seen are exactly how
the program does it. As a scientist, you are, of course, not content to trust your precious data to a canned program, and thus will want to calculate these values yourself. Figure 3.4 shows you how to set up your spreadsheet using the same values as
in Figure 3.3.

Stratigraphic Thickness from Maps

Stratigraphic thickness is defined as the thickness of a unit measured perpendicular to the upper and lower surfaces of the unit. For people truly interested
in stratigraphic sequences, there is still no substitute for going out in the field and
measuring a stratigraphic section with tape, compass, Jacob staff, or whatever. Measuring section that way is certainly the most accurate way to determine the thickness of a unit but the process is time-consuming. For the rest of us, especially when
drawing cross sections, we are more likely to calculate the thickness of a stratigraphic unit from a geologic map, commonly called a map thickness. In this section, we will see how to determine the map thickness in the way that geologists have
for many decades. In the next chapter, well learn how to do the calculation in a
much more flexible way after learning a new concept.
To calculate a map thickness, t, using traditional methods (Fig. 3.5), you must
determine the map distance, h, between the top and bottom of the unit measured
in the true dip direction (i.e., perpendicular to strike); the vertical distance, v, be-

MODERN STRUCTURAL PRACTICE

"54

R. W. ALLMENDINGER 2015-16

GEOLOGICAL MAPS

CHAPTER 3

t = v cos() + h sin()
h
v

t
t = v cos() + h sin()
h
v
h

t
t

t = v cos() h sin()

Figure 3.5 Cross section illustrating the traditional way of calculating map thickness. In all
three cases, the cross section is
perpendicular to the strike of the
unit so that the true dip, , is used.
There are three cases (from left to
right): the unit dips in the same
direction but less steeply than the
slope; the unit dips in the opposite direction from the slope; and,
the unit dips in the same direction
as the slope but more steeply.

tween base and top along the same transect; and the true dip, . There are three
equations and the correct one to use depends on the relationship between the dip
direction and the slope direction (Fig. 3.5). For cases where the stratigraphic unit
dips more steeply and in the same direction as the slope, one uses:

" t = v cos + hsin

(3.8a)

When the slope and dip direction are the same, but the slope is steeper, the correct
formula to use is:

" t = v cos hsin

(3.8b)

And finally, where the dip and the slope are in opposite directions, the equation is:

Figure 3.6 The two right triangles

necessary to solve for the thickness on
your own. The equations shown in the
graphic will eventually simplify to
Equation 3.8c, though it is not necessary to take it that far.

= tan (v/h)
s = h/cos()#
t = s sin(+)

MODERN STRUCTURAL PRACTICE

"55

R. W. ALLMENDINGER 2015-16

GEOLOGICAL MAPS

CHAPTER 3

" t = v cos + hsin

(3.8c)

I highly recommend that, rather than memorizing these equations, you simple derive the equation from the geometry when you need it. Trying to recall which special case to use can lead to errors and it is usually faster to derive the correct equation than to look up and use an equation out of a book, cookbook style. To work
out the equations yourself, note that there are two right triangles as shown in Figure
3.6. You first calculate the slope angle, , and slope distance, s, and then you use
those quantities along with the dip to derive the thickness. Similar graphics can be
devised for the other two special cases.
The necessity to construct these diagrams in the true dip direction and remember the three special cases (Fig. 3.5) is limiting when it come to writing a single
general equation to accomplish the task of determining map thickness. In the next
Chapter, we will learn a powerful set of methods known as coordinate transformations that will enable, not only a general solution to this problem but to a whole set
of interesting problems typically encountered in structural geology.

cr

l
st

in

of

tr a

ce

ax

s
ia l

u r fa

ce

in

in

lin

o
tr

lin

lin

Figure 3.7 Basic fold terminology illustrated with two folded surfaces.
The hinge line connects points of maximum curvature on a surface
whereas the crest and trough lines connect the topographically highest
and lowest points, respectively, on the surface. The axial surface is the
surface that contains all of the hinge lines on all of the folded surfaces.

MODERN STRUCTURAL PRACTICE

"56

R. W. ALLMENDINGER 2015-16

GEOLOGICAL MAPS

CHAPTER 3

Analysis of Folded Rocks

Folded rocks (Fig. 3.7) are some of the most aesthetically pleasing of all geological structures and are quite important economically as hydrocarbon traps.
When folded rocks intersect the irregular surface of the Earth, the resulting outcrop patterns can be quite complex. On a map, the fold is usually represented by
the trace of the axial surface (the intersection between the surface of the earth and
the axial surface) or sometimes the crest line. For fold with shallowly dipping axial
surfaces (i.e., recumbent folds) the trace can be quite complex.

Cylindrical Folds
The simplest fold model is that of a cylindrical fold (Fig. 3.8) in which
there is a line of zero curvature, known as the fold axis, and non-zero curvature at
any point along the fold in a direction perpendicular to the fold axis. the resulting
structure is completely two dimensional such that a measurement of bedding orientation in one part of the fold should be identical to bedding measured in any other
part of the fold parallel to the fold axis. If you were to measure bedding throughout
a cylindrical fold and plot the poles. They would lie within a single plane known as
the profile plane (Fig. 3.8). The profile plane is perpendicular to the fold axis and
provides the truest view of the fold. Any profile plane constructed anywhere
along the fold will show the same view as long as it is perpendicular to the fold axis.

poles

Fo

ld

ax

to be
d

ding

Figure 3.8 A perfect cylindrical

fold illustrating the concept of a
fold axis (line of zero curvature
shown in red) and profile plane. All
of the poles to bedding should be
oriented parallel to the profile
plane. In the real case, bedding is
not measured along a single arc,
but can be measured anywhere on
the surface and the projected parallel to the fold axis to a common
plane.

is

pro

MODERN STRUCTURAL PRACTICE

file
p

lan

"57

R. W. ALLMENDINGER 2015-16

GEOLOGICAL MAPS

CHAPTER 3

(a)

(b)

Figure 3.9 Finding a cylindrical fold axis by plotting bedding orientations on

a stereonet. (a) -diagram of great circles representing the strikes and dips of
bedding and the region of intersection defines the fold axis. (b) -diagram of
poles to bedding. The pole to the best fit great circle through the bedding poles
defines the fold axis.

The cylindrical fold model is highly oversimplified and very few natural folds are
perfectly cylindrical. Nonetheless, the model provides a quick way of determining a
fold axis and it allows us to project a plunging fold geometry to depth.
Determining the fold axis is simple: as implied in the previous paragraph, we
measure a number of bedding orientations around a fold and plot them on a stereonet. Two different approaches can be used: in a -diagram, we plot the great circles of bedding and locate the fold axis where the maximum number of intersections occur (Fig. 3.9a). In the -diagram, the poles to the bedding planes are calculated and a great circle fit to the bedding poles. The pole to the best fit great circle defines the fold axis (Fig. 3.9b). As you can imagine looking at Figure 3.9a, the
-diagram can get messy very quickly with a lot of bedding measurements. For that
reason, and because the numerical solution to this problem uses the poles anyway
as we shall see in a later chapter, most structural geologists prefer the -diagram.
Either type of diagram is easy to use in a stereonet program (look for the Cylindrical best fit menu option).
Most folds are not as nicely curved as the idealized folds depicted in Figures
3.8 and 3.9. Instead, they tend to have mostly planar limbs and a narrow hinge
zone where the dip changes occur over a narrow area. In this case, the geologist
MODERN STRUCTURAL PRACTICE

"58

R. W. ALLMENDINGER 2015-16

GEOLOGICAL MAPS

CHAPTER 3

(a)

(b)

fold
axis

a xi a l

a xi a l p l a n e

p la

ne

fold
axis

Figure 3.10 Determining the fold axis and axial plane on a stereonet. At the top of each
diagram is a profile view of a bed (in green) with the axial plane in blue. In the stereonet
views, below, the poles to bedding are shown with green dots, the cylindrical best fit and
fold axis in red and the axial plane in blue. Dashed lines with arrows show where each
limb plots on the stereonet. (a) and upright asymmetric fold. The downward pointing
poles for both limbs are plotted. (b) an overturned asymmetric fold. The poles of the overturned limb are plotted as upward pointing poles.

who conscientiously measures bedding orientations at regular intervals is more likely to see two point clusters, corresponding to the two limbs, with a relatively small
number of measurements in-between (Fig. 3.10). Furthermore, if the fold is asymmetric, then there will be fewer measurements from the shorter limb. In these cases,
the resulting -diagram can be used to locate not only the fold axis but also the fold
axial plane. Where bedding thickness is preserved across the fold, the axial plane
bisects the angle between the limbs of the fold and contains the fold axis. This calculation is straightforward for upright folds (Fig. 3.10a); you can see on the stereonet that the axial plane plots between the two point clusters. However, overturned
folds are more problematical because, whereas for the upright limb we plot the
downward pointing poles, for the overturned limb we are plotting the upward
pointing poles so they will appear in the lower hemisphere (Fig. 3.10b). In this case
the axial plane will bisect the large obtuse angle between the two point clusters.

MODERN STRUCTURAL PRACTICE

"59

R. W. ALLMENDINGER 2015-16

GEOLOGICAL MAPS

CHAPTER 3

Figure 3.11 The construction of dip isogons and the three general classes of
folds: (a) Class 1 dip isogons where the radius of curvature of the lower bedding
surface is smaller than the upper surface. (b) Class 2 dip isogons where both bedding surfaces have the same curvature; and (c) class 3 dip isogons where the lower
surface has a larger radius of curvature than the upper surface.

Dip Isogons
Cylindrical folds come in several flavors. A common way to classify them is
with dip isogons, which are used to describe the geometric relationship of one
bedding surface to the next. A dip isogon connects two points on adjacent bedding
surfaces which both have the same dip relative to the axial plane of the fold. As
shown in Figure 3.11, there are three general classes of folds based on dip isogons.
Where bedding thicknesses are preserved, Class 1 dip isogons are also known as
parallel folds (Fig. 3.11a). Class 2 dip isogons result from similar folds (Fig.
3.11b). Finally, Class 3 dip isogons indicate a fold where the inner surface is less
curved than the outer surface. We will return to fold geometry and kinematics in
Chapter 10 because cross section construction relies heavily on choosing an appropriate fold model.

Folds in Map View

A cylindrical fold is just about the simplest fold form you can imagine and
yet, when you convolve the form with the topography of the earth, the resulting
map pattern can be highly complicated. Even in the absence of topography, correct
interpretation of the fold axis and axial trace can be counterintuitive. Because the
axial plane is, well, a plane, the trace that the axial plane makes across the land surface is plotted on the map, i.e., the intersection between the plane and the topograMODERN STRUCTURAL PRACTICE

"60

R. W. ALLMENDINGER 2015-16

GEOLOGICAL MAPS

CHAPTER 3

(a)

(b)

Figure 3.12 Folds and topography in (a) map view, and (b) down-plunge view. In the map
view (a), the heavy black lines are index topographic contours and the light gray lines are
other topo contours; a hill with NW-trending ridge is depicted. The yellow lines are traces of
fold axial planes with standard symbols for over-turned anticlines and synclines. the block
diagram in (b) shows the down-plunge view of the structure tilted so the viewer is looking
parallel to the trend and plunge of the folds where the true fold geometry is revealed. In
this view, axial traces are approximately straight lines. In (b) only the index contours are
shown as discontinuous black lines. Diagrams created with Visible Geology (thanks,
Rowan!).

phy (Fig. 3.12a). The fold axis is a line that, in general, intersects the surface in a
point (really a point for each geological surface). Thus, we show the trend and
plunge of the fold axis as a small arrow pointing in the direction of the trend of the
axis with a number next to it for the plunge (Fig.
3.12a). The down-plunge view of a cylindrical
There is a spectacular web-based
fold, which can be approximately achieved by
resource, Visible Geology,
tilting the map so that you are looking in the diwhere you can create your own
rection direction of plunge, provides the most
3D models, look at those created
accurate perspective on the fold geometry and
by other people or visualize a
can help make sense of a seemingly complicated
stereonet in 3D:
map pattern (Fig. 3.12b). In the next chapter, we
http://app.visiblegeology.com
will see how to calculate down-plunge views of
folds.

MODERN STRUCTURAL PRACTICE

"61

R. W. ALLMENDINGER 2015-16

GEOLOGICAL MAPS

CHAPTER 3

Working out the geometry of fold axis and axial plane is not always straight
forward, even in the absence of significant topography. A couple of tricks will help
you to interpret folds more quickly. Most importantly, as shown in Figure 3.10, the
fold axis is a line contained within the axial plane, but it can have any rake within
that plane. Thus, for some types of folds known as reclined folds, the trend of the
fold axis can actually be as much as 90 to the strike of the axial plane. If we know
the strike of the axial plane and the trend and plunge of the fold axis, we can determine the dip of the axial plane. Alternatively, by using the trace of the axial
plane in down-plunge view and the trend and plunge of the axis from a -diagram,
we can calculate the orientation of the axial plane using the cross product or on a
stereonet.
The other trick is equally useful, enabling rapid determination of the
trend and plunge of the fold axis: the strike of any vertical bedding must be parallel
to the fold axis trend and the dip of any bedding that strikes perpendicular to the
fold axis must be equal to the plunge of the fold axis. Thus, if you find bedding
dipping 90 in one part of an area that has experienced one episode of cylindrical
folding, you know that the trend of the folds is parallel to the strike of that bed!

Structural Contour Maps

Just about any value can be contoured to
show its spatial variation and structure data are no
exception. In a structural contour map, one contours lines of equal elevation on a surface of interest. Surfaces can be smoothly varying or they can be
disrupted by faults. Automatic contouring algorithms can do a good job where the elevations are
smoothly varying but tend to have trouble where
contours terminate abruptly against a fault. Furthermore, the geologist has a better idea of what deformed geological surfaces actually look like so is in
a better position to draw realistic contours, though
one must always be aware of unintended bias.

MODERN STRUCTURAL PRACTICE

"62

3752

3500
3000
B

2500

2940

2000

1475

1500

Figure 3.12 Contouring via

linear interpolation along the
sides of a triangle with apices
of know elevation.

R. W. ALLMENDINGER 2015-16

GEOLOGICAL MAPS

CHAPTER 3

A geologist doing a first pass at contouring a surface would operate very

much as a computer program does: s/he would construct a mesh of triangles
such a mesh is frequently accomplished using Delaunay triangulation with
apices at the points of known elevation and would then linearly interpolate between the points (Fig. 3.12). This sort of contouring is tedious to do by hand but
easy to implement in a computer program.
A significant difference between contours of the topographic surface and a
geologic surface of interest is that the surface of the Earth is continuous whereas
the geologic surface can have gaps and overlaps across faults. A normal fault will
produce a gap (Fig. 3.13a) and a thrust fault will produce an overlap (Fig. 3.13b).
Folded surfaces that are also faulted yield the most complex patterns (Fig. 3.13c).
Clearly, a good structure contour map, except in the simplest case of an inclined
planar surface with constant strike and dip, requires extensive subsurface data to
construct. Such data may be available where there has been extensive hydrocarbon
industry exploration.

4000

3000

(b)

5000

3000

4000

3000

6000

4000

(a)

4000

5000

6000

4000

500

5000

2000

5000

6000

2000

6000

600

5000

700

6000

4000

1000

(c)

Figure 3.13 Examples of structure contour maps. (a) A normal fault produces a
gap, (b) thrust or reverse fault an overlap, and (c) a cylindrically folded surface cut by
a thrust fault striking perpendicular to the fold axis.

MODERN STRUCTURAL PRACTICE

"63

R. W. ALLMENDINGER 2015-16

GEOLOGICAL MAPS

CHAPTER 3

ExercisesChapter 3
These exercises and those of the next Chapter will rely heavily on the program
GeolMapDataExtractor (GMDE) and a digital copy of the Poker Peak geologic
map from the US Geological Survey (Albee and Cullens, 1975). GMDE, in either
Macintosh or Windows version can be downloaded from:
http://www.geo.cornell.edu/geology/faculty/RWA/programs/strikedipthickness.html

and the Poker Peak Quadrangle map can be downloaded from the class Web site.
GMDE will allow you to make quick accurate measurements on the map. Note
that you will need an Internet connection for the program to get elevations automatically. For Exercises 1 and 2, you will also need a copy of Google Earth, which
you can download for free from:
http://www.google.com/earth/download/ge/agree.html

1. Open the Poker Peak Quad in GMDE and make sure it is georeferenced (it
should automatically be unless you have separated the .jpg file of the map from
the .txt file of the same name (the .txt file contains the georeferencing information.) Locate the part of the map shown to the right. Notice that a fault is
shown offsetting the stratigraphic
units.
(a) Across the entire segment of the map
shown, digitize the following contacts: base of Kp, top of Kp, top
of Kb, top of Kd.
(b) Save your contacts as a .kml file
(File>Export KML>contacts),
and then open the .kml file in
Google Earth.
(c) Kp and Kd are two very light
gray to nearly white limestone
units. Trace these white units on
the Google Earth image from
the northern end of your digitizing to the south.
(d) Evaluate the mapping job and
MODERN STRUCTURAL PRACTICE

"64

R. W. ALLMENDINGER 2015-16

GEOLOGICAL MAPS

CHAPTER 3

the evidence for a fault, based on the agreement or lack of agreement of

your digitized contacts with the units that you see in Google Earth. It will
probably help to tilt and pan the Google Earth view to evaluate different
perspectives. Turn in a paragraph with your assessment of the mapping.
2. With the Poker Peak Quad loaded into
GMDE, digitize about 60 strikes and
dips around the Big Elk Anticline, the
area outlined in the red polygon to the
right.
(a) In GMDE, select
Settings>Mode>Strike & Dip Only.
Each time you want to measure a
strike and dip, select
Operations>Drag Strike Line or select D from the keyboard (easier).
To drag a strike line, click at the intersection of the strike line and dip
tick mark and then drag the mouse in
the direction of the right-hand rule
strike (you can correct it if you drag
in the wrong direction with the popup menu). You will have to type the
dip value in by hand. Once you are
happy with the strike and dip, press the Record Strike & Dip Button. The
program should draw a new strike and dip symbol over the one that you just
measured.
(b) When you have measured and recorded all of the strikes and dips in the outlined area, save your data. You should save it in three ways:
(i) first, select File>Save Data>Thickness/Orientation Data. This will allow
you to read your data back in to GMDE easily which will be useful in
next weeks lab.
(ii) Next select File>Export Strikes & Dips. This file can be easily read into
Stereonet.
(iii) Finally, select File>Export KML>Strikes & Dips to make a file for visualization in Google Earth. In the dialog box that appears, just click Okay
which will use the default values.

MODERN STRUCTURAL PRACTICE

"65

R. W. ALLMENDINGER 2015-16

GEOLOGICAL MAPS

CHAPTER 3

(c) Open the KML file in Google Earth. The strikes and dip symbols are rendered in three dimensions so you can clearly see the relationship between
strike and dip and the actual stratification. Be sure that you have topography
turned on in Google Earth and tilt and pan to get a good feeling for the 3D
nature of the geology. In Google Earth, save an image of the screen
(File>Save image) which you think shows especially well the relationship between the measured strike and dip and the geology
and turn it in with a sentence about why you liked it.
(d) Open the second file in Stereonet. Calculate the poles to the plane, make a
-diagram, and determine the fold axis of the Big Elk Anticline. Turn in
your plot with the fold axis clearly annotated.
3. In GMDE with the Poker Peak
Quad loaded, find the area in the
northeast corner of the map shown
to the right.
(a) Select Settings>Mode>Strike &
Dip Only and uncheck the
check box that says user entered strike and dip. Use
GMDE to find the East, North,
and Up coordinates of three
points on the base of the purple
TRa where it crosses the ridge
with the number 14 on it. To
do so click the click button
next to one row, then click on the
map exactly where you want to place the point. Repeat for the other two
points. If you have an Internet connection enabled, the program will get the
elevation for you automatically; otherwise, you will have to type in the elevation coordinate by hand.
(b) Copy the ENU coordinates of the three points you have defined into a
spreadsheet and calculate the strike and dip of the unit in the spreadsheet
using vector subtraction and the cross product. You can compare your answer to the correct answer that GMDE calculates automatically.
(c) Compare your calculated value to the strike and dip symbol that the geologist measured on the ridge for the base of TRa and discuss the uncertainties
and reasons for the difference in two values, if any. Once you record your
MODERN STRUCTURAL PRACTICE

"66

R. W. ALLMENDINGER 2015-16

GEOLOGICAL MAPS

CHAPTER 3

strike and dip, you can see its location in a satellite image by choosing Datum Details from the Window menu. This may help you in your discussion.
Turn in both the spreadsheet with the calculation of the strike
and dip and the your comparison with the mapped value and your
discussion.
4. GMDE allows you to make accurate distance measurements: just click and drag
the mouse from the top (or base) of the unit in the direction of the true dip to
the base (or top) of the unit. The distance and azimuth of the line you drag is
shown in the box at the lower left corner of the map view. Be sure to read the
distance before releasing the mouse! You can read the elevations off of the topo
contours on the map or, if you have an Internet connection, hold down the option key while you click the mouse at the point for which you want the elevation.
(a) On the same ridge that you used for problem 3, calculate the thickness of
TRa using the appropriate Equation 3.8. Do the calculation for both the geologist measured orientation and the orientation that you determined from
the three-point problem in Exercise 3.
(b) Make a drawing and derive a set of equations for the two right triangles
that you would use to get a the answer in part (a). You dont have to derive
Equation 3.8 but the equations that you do get have to give you the same
answer as Eqn. 3.8!
5. On the next page is a map showing a number of borings that were made to determine the elevation of the Jurassic Twin Creek Limestone in the subsurface
around the nose of the Big Elk anticline. Contour these elevations by hand using a contour interval of 500 ft and describe the resulting surface. You do not
need to do a formal linear interpolation between adjacent points.

MODERN STRUCTURAL PRACTICE

"67

R. W. ALLMENDINGER 2015-16

GEOLOGICAL MAPS

CHAPTER 3

7,072

4,112

5,771

2,910

38,500
4,933

6,865

2,013

964
6,414

3,704

36,500
7,615

2,838
5,712

34,500
7,234

4,600

2,193

6,066
1,538

4,721

32,500

Jtc outcrop

7,619

3,610
1,971

North

5,576

30,500

7,166

3,437

5,536

8,078
6,717
6,787

28,500

7,335

7,478

2,788

5,247

6,348
4,801

6,655

5,958

6,482

3,496

5,391

5,491

3,986

4,502

26,500

4,713

5,720

3,523

4,549
3,727
2,619

4,394

4,333

3,718

5,168
3,890

3,405
2,914

24,500

4,560
3,529

3,352

3,366

00
,0
17

00
,0
15

13

,0

00

0
11
,0
0

0
00
9,

0
00
7,

5,

00

22,500

East
East
17,000
15,000
13,000
11,000
9,000
7,000
5,000
22,500
24,500
26,500
28,500
30,500
32,500
34,500
36,500
A
38,500
North
4,933
6,414
3,704
5,712
4,600
7,234
6,066
7,619
5,576
3,610
2,838
2,013
7,166
5,536
3,437
7,335
6,348
4,713
5,391
3,986
3,890
5,168
5,720
3,529
3,405
4,333
6,655
7,478
5,491
5,958
6,787
6,482
4,549
2,788
1,971
9
2,193
2
964
5
2,910
5,771
7,615
8
6,717
8,078
4
4,721
4,112
3
3,718
1
3,366
7
1,538
6
A
7,072
6,865
,933
,414
,704
,712
,600
,234
,066
,619
,576
,610
,838
,013
,166
,536
,437
,335
,348
,713
,391
,986
,890
,168
,720
,529
,405
,333
,655
,478
,491
,958
,787
,482
,549
,788
,971
,193
64
,910
,771
,615
,717
,078
,721
,112
,718
,366
,538
,072
,865
MODERN STRUCTURAL PRACTICE

"68

R. W. ALLMENDINGER 2015-16

Chapter 4
Transformations
Introduction
Many times in geology, as in other disciplines, the problem we are interested
in solving is simpler when viewed from a different perspective. Anyone who plays
computer games or flight simulators knows this to be true. So far, we have been
dealing with geographic-centric coordinate systems either NED or ENU because we measure our world with respect to geographic coordinates. But there is a

whole class of problems that benefit from a change in coordinate system or transformation. In the last chapter, we noted that the profile view of a plunging fold is
the most accurate representation of the structure. The profile view is defined by a
different coordinate system the fold coordinate system rather than the geographic view. In this Chapter, we will learn how to switch back and forth between
these two views and in the process begin to lay the foundation for understanding
entities more complicated than geometry such as stress and strain. Before getting to
interesting geological problems, however, well need to develop the simple mathematical foundation of transformations.

TRANSFORMATIONS

CHAPTER 4

X2

X2

X 2

X 2
j
22

X 1

12

X 1

i
a22

21

a12

11
X1

a21

(a)

a11

X1

(b)

Figure 4.1 A two-dimensional rotation of a coordinate system between an old,

unprimed coordinate system in black and a new primed coordinate system in red.
(a) The rotation defined by four angles, and (b) defined by four direction cosines
which are the projections of the base vectors of the new coordinate system on the
axes of the old system. In both cases, the first subscript (in red) identifies the new
axis and the second subscript (in black) the old axis

Transformations of Coordinates and Vectors

Coordinate Transformations
At the simplest level, a transformation involves a change in orientation of the
coordinate system. Like many things, we will visualize this first in two dimensions
(Fig. 4.1) and then in three dimensions (Fig. 4.2). In the text that follows, we will always refer to the new coordinate axes as the primed coordinate system, X, and the
old coordinate system as the unprimed system, X. To define the rotation of a coordinate system in two dimensions, there are four angles between the two new coordinate axes and the two old axes. One needs to be very careful systematic about naming these
angles: The first subscript refers to the new coordinate axis and the second subscript
to the old coordinate axis. For example, 21 is the angle between the X2 axis and
the X1 axis (Fig. 4.1a). After the last few chapters, it should come as no surprise that
we will typically be using the direction cosines rather than the angles themselves
(Fig. 4.1b). The subscript convention is exactly the same and the relations between
the angles and their direction cosines are:

" a11 = cos11 ; a12 = cos12 ; a21 = cos 21 ; a22 = cos 22

MODERN STRUCTURAL PRACTICE

"70

(4.1)

R. W. ALLMENDINGER 2015-16

TRANSFORMATIONS

CHAPTER 4

X3

23

X2
22

21

X1
New Axes

X3

X2

Old Axes
X2
X3

X1

a11

a12

a13

X2

a21

a22

a23

X3

a31

a32

a33

X1
X1

(a)

(b)

Figure 4.2 Coordinate transformation in three dimensions: (a) the new axes
(primed) in red and the old in black. As before the first subscript refers to the new
axis and the second to the old axis. (b) The array of directions cosines: each row
refers to a new axis and each column to an old axis.

The extension to three dimensions is quite straightforward (Fig. 4.2). Because

we now have three coordinate axes in the new system and three in the old system,
there will be nine angles and nine direction cosines. This 33 table, or array, of direction cosines is known as the transformation matrix:

a11 a12

" aij = a21 a22

a31 a32

a13 cos11 cos12

a23 = cos 21 cos 22
a33 cos 31 cos 32

cos13

cos 23
cos 33

(4.2)

We will be using the shorthand aij a lot in the following sections because it is much
more compact than writing out those matrices all the time. Just recall that i and j
can each have values ranging between 1 and 3.
As you no doubt notice in the case of the 2D transformation, not all of the
angles, or direction cosines, are independent. In that case, only one angle is independent and the rest can be calculated from it. In the 3D case, only three angles are
independent and the rest can be calculated from those three. Matrix a is technically
known as an orthogonal matrix and the equations that relate the direction
cosines are known as the orthogonality relations:

MODERN STRUCTURAL PRACTICE

"71

R. W. ALLMENDINGER 2015-16

TRANSFORMATIONS

CHAPTER 4

a112 + a122 + a132 = 1

a21a31 + a22 a32 + a23a33 = 0

" a + a + a = 1 and " a31a11 + a32 a12 + a33a13 = 0
2
2
2
a11a21 + a12 a22 + a13a23 = 0
a31
+ a32
+ a33
=1

2
21

2
22

(4.3)

2
23

While we wont bother to do so in detail here, both of these sets of equations are
easy to prove: the left equations come from the fact that the base vectors of the new
coordinate system have magnitudes of 1, and the right hand equations are due to
the fact that, because the base vectors are perpendicular to each other, the dot
product of each of the base vectors with the other is zero.

Transformation of Vectors
The transformation matrix by itself is not very useful. Its power comes from
the fact that, once we know the transformation matrix and the components of a
vector in one coordinate system, we can immediately calculate the components of
the vector in the other coordinate system. Figure 4.3 shows how this works in two
dimensions to calculate the v1 component of the vector in the new coordinate system as a function of the old components v1 and v2. In terms of direction cosines we
could write the equation as:

" v1 = a11v1 + a12 v2

X2

(a)

X2

(b)
D

X2

v2

X2

v1

v2

X1

X1

v
A

12
v2

11
O

12

v1
v2
v1

X1

os 1
v 1c
11
v1

12
B

C
2
os 1 X1
c
v2

Figure 4.3 the transformation of a vector. (a) shows the components of vector v in the
old (black, unprimed) and new (red, primed) coordinate systems. (b) the graphical construction to calculate the value of v1 in terms of the old coordinates. The blue and yellow
highlighted triangles are used to trigonometrically calculate v1 in terms of v1 and v2.

MODERN STRUCTURAL PRACTICE

"72

R. W. ALLMENDINGER 2015-16

TRANSFORMATIONS

CHAPTER 4

Perhaps the most important thing to notice is that, although the coordinate systems
have changed, and therefore the values of the components of the vector in the two
coordinate systems are different, the fundamental nature of the vector itself has not
changed at all. This seems trivially obvious but it is exactly this property that determines whether or not something is a tensor property! Yes, a vector is sometimes
referred to as a first order tensor. More on that in later chapters.
In three dimensions, there are three equations that describe the components
of the vector in the new system as a function of the old components:
v1 = a11v1 + a12 v2 + a13v3
" v2 = a21v1 + a22 v2 + a23v3
v3 = a31v1 + a32 v2 + a33v3

(4.4)

And, there are another three equations that give the components of the vector in
the old coordinate system as a function of the new coordinates:
v1 = a11v1 + a21v2 + a31v3
" v2 = a12 v1 + a22 v2 + a32 v3
v3 = a13v1 + a23v2 + a33v3

(4.5)

These equations consist of nothing more than multiplications and additions

done in an extremely systematic way. If you look at Equations (4.4), you can see
that the order in which the a terms occur is that same as in the matrix in Equation
(4.2). In each equation, you can see that the subscript of the v terms varies from 1
to 2 to 3; In fact, the subscript of the v term in each
X3 = U
equation matches the second subscript of the a term
that multiplies it. In Equation (4.5), youll notice that
X2 = N
the subscripts of a appear flipped relative to the a
X1 = N
subscripts in (4.4). In technical terms, the matrix a has
been transposed. The reason for the systematic variation in subscripts is because Equations (4.4) and (4.5)
180
X1 = E
represent matrix multiplications.
X2 = E

Figure 4.4 The transformation from ENU (a map coordinate system) to NED (structural geology coordinate system).

MODERN STRUCTURAL PRACTICE

"73

X3 = D

R. W. ALLMENDINGER 2015-16

TRANSFORMATIONS

CHAPTER 4

A Simple Coordinate Transformation

There is a coordinate transformation that we already know how to do but,
nonetheless it is illustrative to see that it works as expected: the transformation from
an ENU to a NED coordinate system. It is simple because all the angles involved
are 0, 90, or 180 but important because it is a transformation that we do frequently. The geometry is shown in Figure 4.4. You can see that, for example 11 = 90,
12 = 0, and 13 = 90. Writing it out in the form of the matrix, we have:
cos 90 cos0 cos90
" aij = cos0 cos90 cos90

cos90 cos90 cos180

This simplifies to:

0 1 0
" aij = 1 0 0

0 0 1

(4.6)

When we use this transformation matrix in Equation 4.4, we get:

" v1 = v2 ;

v2 = v1;

v3 = v3

(4.7)

Structural Geology Applications of Transformations

The equations (4.4) and (4.5) give us some real power to solve interesting
problems in structural geology. In all of these problems, the key is determining
what the transformation matrix, a, looks like and that is where we will spend most
of our time, below. As a general rule, the components of the transformation
matrix are just equal to the direction cosines of the new axes in the old coordinate system. Thus, we will be seeing combinations of the equations in Table
2.1 a lot. Most importantly, though, recall that the math required is nothing more
than multiplications and additions done in a very systematic way.

MODERN STRUCTURAL PRACTICE

"74

R. W. ALLMENDINGER 2015-16

TRANSFORMATIONS

CHAPTER 4

X3 (U)
X2, X1 (N)

X1 (Strike)

Figure 4.5 Diagram showing the coordinate transformation from an ENU to

Strike-Dip-Pole (SDP) coordinate system.
The thickness can be calculated from any
point on the top and any point on the base
of the bed by subtracting there X3 coordinates.

top
e

bas

X1, X2 (E)

X3
(Pole)

X2 (Dip)

X3 (D)

Stratigraphic Map Thicknesses

We start with a problem first introduced in the previous chapter: determination of the thickness of a geological unit from map data. The general method introduced here (Allmendinger and Judge, 2013), like all methods to determine thickness assumes that the top and bottom surfaces of the bed are planar and parallel.
Our basic approach will be a transformation into the bedding coordinate system
where the strike is the first axis, the dip the second axis and the pole the third axis,
an SDP system (Fig. 4.5). In this coordinate system, the thickness is just the difference between the X3 coordinate of a point on the top of the bed and X3 coordinate of any point on the base (not necessarily in the dip direction).
A slight wrinkle of this method is that the position vector of the points on the
top and base of the bed are given in an ENU coordinate system but the direction
cosines of the strike, the dip direction and the pole will be in a NED coordinate system (Fig. 4.5). Lets assume that the point (i.e., position vector) on the top has coordinates [tE, tN, tU] and the point on the base has coordinates [bE, bN, bU] in the
ENU system. Our first step will be to transform that into the coordinates in a NED
system as we have seen before. In the NED system the two vectors are:

" t = t N t E tU and " b = bN bE bU

MODERN STRUCTURAL PRACTICE

"75

R. W. ALLMENDINGER 2015-16

TRANSFORMATIONS

CHAPTER 4

Now we can use the direction cosines of the strike, dip and pole to construct the
transformation matrix. The strike is our X1 axis so the first row of the transformation matrix will be its direction cosines in the NED system:

" cos ( strike) cos ( 0 ) sin ( strike) cos ( 0 ) sin ( 0 )

Recall that the strike is a horizontal line with a plunge of 0. The second row of
the transformation matrix are just the direction cosines of the dip and dip direction
(X2) in an NED system (assuming strike in right hand rule format):

" cos ( strike + 90 ) cos ( dip ) sin ( strike + 90 ) cos ( dip ) sin ( dip )

Finally, the pole is the third axis of our new coordinate system, X3 and its direction cosines in the NED system are:

" cos(strike 90)cos ( 90 dip ) sin ( strike 90 ) cos ( 90 dip ) sin ( 90 dip )

Assembling all three rows of transformation matrix and making the necessary
trigonometric simplifications, we get:

cos(strike)
sin(strike)
0
" aij = sin(strike)cos(dip) cos(strike)cos(dip) sin(dip)
sin(strike)sin(dip) cos(strike)sin(dip) cos(dip)

(4.8)

Now we can transform the top position vector, t, in the NED coordinate system into the new bedding system (the bottom vector will look identical):
t1= a11t1 + a12t 2 + a13t 2 = cos ( strike) t N + sin ( strike) t E + 0 ( tU )

" t 2 = a21t1 + a22t 2 + a23t 3 = ( sin ( strike) cos ( dip )) t N + ( cos ( strike) cos ( dip )) t E + sin ( dip )( tU ) (4.9)
t 3 = a31t1 + a32t 2 + a33t 3 = ( sin ( strike) sin ( dip )) t N + ( cos ( strike) sin ( dip )) t E + cos ( dip ) ( tU )

The thickness is " t 3 b3 so only the third line above matters and we can write the
thickness as:

MODERN STRUCTURAL PRACTICE

"76

R. W. ALLMENDINGER 2015-16

TRANSFORMATIONS

CHAPTER 4

" t 3 b3 = ( sin ( strike) sin ( dip ))( t N bN ) ( cos ( strike) sin ( dip ))( t E bE ) + cos ( dip )( bU tU ) (4.10)
The order of t and b in the third term are reversed because it is really: [tu (bu)].
To reemphasize, strike and dip must be in right hand rule format for this equation
to work.
The spreadsheet solution to this problem is shown in Figure 4.6. The 33
block of numbers labeled Transformation Matrix are calculated using Equation
4.8. The actual transformation from NED to SDP coordinate system occurs in
rows 14 and 15; you can see the formula for cell D15 expanded and color coded to
show which values are being used. In those formulae, we use absolute cell references (i.e., the $ signs) to the transformation matrix so their cell positions will not
change when we copied the formulae from row 14 to row 15. Finally, note that because NED is positive downwards, we actually subtract the top from the base!

Figure 4.6 The spreadsheet for

calculating stratigraphic thickness using the Equations 4.6 and
4.7.

Down-Plunge Projection
As we saw in the previous Chapter, cylindrical folds have a profile plane that
is perpendicular to the fold axis (Fig. 3.8). The profile plane contains the least distorted view of a cylindrically folded surface, known as a down-plunge projection.
Constructing down plunge projections by hand usually involves orthographic projection and can be particularly tedious, so much so that most structural geology lab

MODERN STRUCTURAL PRACTICE

"77

R. W. ALLMENDINGER 2015-16

TRANSFORMATIONS

CHAPTER 4

pro

p
file

lan

X3
X3 (U)

X1
X2 (N)

Figure 4.7 The two coordinate systems in the

down plunge projection.
See text for discussion.

horizontal
X1 (E)

fold
2 (

axis

manuals omit the 3D version entirely and instead only teach students how to do a
two dimensional construction that assumes that the surface of the earth is flat!
However, a full three dimensional down-plunge projection can be calculated using
the transformation methods in this chapter.
The two coordinate systems are shown in Figure 4.7: the first is our usual geographic coordinate system East-North-Up. It makes sense to use a map coordinate
system because the contacts that describe the bedding surfaces to be projected will
be digitized from a geologic map. The second coordinate system is defined by the
fold axis2. As before, if we want to use the trick that the direction cosines of the
transformation matrix are just the direction cosines of the axes in NED given by
Table 2.1, we first have to transform the position vectors of the digitized bed from
ENU to NED, like we did in the previous example. Lets assume that a digitized
point along a bedding surface has coordinates x1, x2, x3 in an ENU coordinate system. Our first transformation using Equation (4.7) gives us:

" x1 = x2 ;

x2 = x1;

x3 = x3

(4.11)

The second transformation matrix will be based on the trend and plunge of the
fold axis (X2), TFA and PFA, respectively. As you can work out from Figure 4.7,
2

The derivation that we do here is different than that in Allmendinger et al. (2012), including the numbering of the
axes. In Allmendinger et al. they derive a single transformation from ENU directly to the fold coordinate system.
Here we use an easier to follow two step transformation in which the fold axis is the X2 axis, not the X3 axis.

MODERN STRUCTURAL PRACTICE

"78

R. W. ALLMENDINGER 2015-16

TRANSFORMATIONS

CHAPTER 4

the trend and plunge of X1 will be (TFA+90), 0 and the trend and plunge of X3
will be TFA, PFA-90. The plunge of X3 will be negative because the arrow points
upwards in a NED coordinate system. Now, we can use the equations in Table 2.1
to construct the second transformation matrix from NED to the fold coordinate
system. Without bothering with the intermediate steps, it is:

"

cos(TFA + 90)cos(0)
sin(TFA + 90)cos(0)
sin(0)

bij =
cos(TFA)cos(PFA)
sin(TFA)cos(PFA)
sin(PFA)
cos(TFA)cos(PFA 90) sin(TFA)cos(PFA 90) sin(PFA 90)

sin(TFA)
cos(TFA)
0

= cos(TFA)cos(PFA) sin(TFA)cos(PFA) sin(PFA)

cos(TFA)sin(PFA) sin(TFA)sin(PFA) cos(PFA)

(4.12)

The second transformation is then

" xi= bi j x j

(4.13a)

x1= b11 x1 + b12 x2 + b13 x3

" x2 = b21 x1 + b22 x2 + b23 x3
x3 = b31 x1 + b32 x2 + b33 x3

(4.13b)

Where " xi is the coordinate of a point on a digitized bed in fold coordinates," x j is

the coordinate of the same point in NED coordinate system which was transformed from its original ENU coordinates via Equation. 4.11.
To implement these equations, you will first need to digitize some contacts,
specifying an East coordinate, a North coordinate, and an elevation for each vertex
of the polygon drawn along the contact. A program like GMDE makes the process
relatively painless and will also show you what your calculated projection should
look like. Then, you will import the coordinates of your digitized polygon along the
bedding contact into a spreadsheet program (Fig. 4.8). The fold axis presumably
determined by constructing a -diagram (Fig. 3.9b) must also be entered. The
first transformation (Fig. 4.8) follows Equation 4.11; the transformation matrix
comes from Equation 4.12 and the actual transformation from Equation 4.13. No-

MODERN STRUCTURAL PRACTICE

"79

R. W. ALLMENDINGER 2015-16

TRANSFORMATIONS

CHAPTER 4

Figure 4.8 Part of the spreadsheet for doing a down-plunge projection. The first
transformation switches from ENU to NED, and the second changes from NED to
the fold coordinate system defined by the fold axis.

tice in the spreadsheet in Figure 4.8 that the equations implementing the second
transformation should have absolute cell references to the transformation matrix.
To see your down-plunge projection in the spreadsheet, select the columns G
and I, representing coordinates along the X1 and X3 axes respectively, and choose
scatter plot (you can do non-contiguous selections in most programs using the
command key rather than the shift key). The plot will probably be distorted because spreadsheet programs to do not try to scale the two axes equally. Thus you
will have to, either by hand or using the chart dialog box, adjust the scales so that
they are equal.

Rotations
Rotations are a way of life in structural geology because we frequently want
to restore some feature to its initial, pre-deformation orientation: What direction
did the current flow prior to tilting of bedding? How can well tell whether magnetic vectors were imprinted on the rock before folding or after? what orientation was
the dike when it was intruded? These questions have been answered for decades by
using a stereonet with paper and pencil. Rotations done in this manner are straightforward when the rotation is about a horizontal axis (e.g., returning dipping bedding to horizontal) but rotations about inclined axes require breaking the rotation
up into two or more separate rotations. Computer stereonets, of course, do rota-

MODERN STRUCTURAL PRACTICE

"80

R. W. ALLMENDINGER 2015-16

TRANSFORMATIONS

CHAPTER 4

X1
X3

X1

22

X3
X2

X2

Figure 4.9 Rotation as a coordinate

transformation. The old axes are NED
and the new coordinate axes are shown
in red. The rotation axis orientation is
defined by the angles that it makes with
the NED axes; the magnitude of rotation
is . Note that the rotation angles are
defined along small circles centered on
the rotation axis whereas the angles between axes such as 22 are measured
along great circles.

X1

tions completely differently and thus do not suffer any of the limitations of paper
stereonets. Lets see how computers accomplish this task.
Given its appearance in this Chapter, it should come as no surprise that most
computer programs use coordinate transformations for rotations. The type of
transformation, though, is a bit more complicated to derive than in previous examples because the rotation axis does not coincide with either the new or the old coordinate system. The old coordinates, of course, are our NED coordinate system,
but we do not know a priori what the orientations of the new axes will be; we have
to calculate them.
Figure 4.9 shows the angular relations involved. The stereonet grid has been
repositioned to be centered on the rotation axis. All points will move parallel to the
small circles in this projection. The givens are the orientation of the rotation axis,
specified in terms of the angles , , and that it makes with North, East, and
Down, respectively, and the magnitude of the rotation, . As shown (Fig. 4.9), the
amount of rotation and the angle between new and old axes are different things.
From those values, the direction cosines of the transformation matrix must be
calculated. Here, well just give you the matrix; the derivations are not particularly
difficult but are tedious to follow. You can find the derivation in terms of spherical

MODERN STRUCTURAL PRACTICE

"81

R. W. ALLMENDINGER 2015-16

TRANSFORMATIONS

CHAPTER 4

trigonometry in Allmendinger et al. (2012) and in terms of matrix multiplications

in any linear algebra textbook.
The following equations (4.14) comprise the transformation matrix for a
general rotations:
a11 = cos + cos 2 (1 cos )

a12 = cos sin + cos cos (1 cos )

a13 = cos sin + cos cos (1 cos )

a21 = cos sin + cos cos (1 cos )

" a22 = cos + cos2 (1 cos )

(4.14)

a23 = cos sin + cos cos (1 cos )

a31 = cos sin + cos cos (1 cos )
a32 = cos sin + cos cos (1 cos )
a33 = cos + cos 2 (1 cos )

These equations are put into practice in the spreadsheet in Figure 4.10: (a) enter
the trend and plunge of the lines to be rotated as well as the rotation axis and the
magnitude of the rotation (columns A and B). (b) Calculate the direction cosines of
those lines and the rotation axis (columns C, D, E). (c) Calculate the transformation
matrix (F14-H16) using the Equations (4.14), above. Dont forget that you have al-

Figure 4.10 Spreadsheet to rotate the eight lines in rows 3-10. See Text for discussion

MODERN STRUCTURAL PRACTICE

"82

R. W. ALLMENDINGER 2015-16

TRANSFORMATIONS

CHAPTER 4

ready calculated the cos, cos, cos for the rotation axis so you do not need to do
that again! Although it is not necessary, a quick check on your equations is to calculate the magnitude of the vector in each row of the transformation matrix; as
shown in I14-I16, they should all be equal to 1. (d) Transform the direction cosines
in columns C-E to the rotated direction cosines in columns F-H using the transformation matrix. As you can see from the exposed formula, be sure to use absolute
cell references. (e) Finally, convert your rotated direction cosines back to trend and
plunge (columns I, J) using Equations (2.11).

Summary
This chapter covers a hugely important technique transformations and
the illustrations here only scratch the surface of what you can do with this approach. The theory is easy to grasp in two dimensions and the extension to three
dimensions is straightforward. One only needs to be very systematic in making sure
that the subscripts are in the correct order and apply to the axes we think they apply to. Most importantly, the subscripts always apply to a specific coordinate axis.
Understanding this will help us a lot when we get to stress and strain. Finally, although the approach is very systematic and logical, the math is very simple: just
multiplication and additions, something that you have know how to do since middle
school!

MODERN STRUCTURAL PRACTICE

"83

R. W. ALLMENDINGER 2015-16

TRANSFORMATIONS

CHAPTER 4

ExercisesChapter 4
Once again, these exercises rely heavily on the program GeolMapDataExtractor
(GMDE) and a digital copy of the Poker Peak geologic map from the US Geological Survey (Albee and Cullens, 1975). GMDE, in either Macintosh or Windows
version can be downloaded from:
http://www.geo.cornell.edu/geology/faculty/RWA/programs/strikedipthickness.html

And the Poker Peak Quadrangle map can be downloaded from the web site.
GMDE will allow you to make quick accurate measurements on the map. Note
that you will need an Internet connection for the program to get elevations automatically. For Exercises 1 and 2,you will also need a copy of Google Earth, which
you can download for free from:
http://www.google.com/earth/download/ge/agree.html

1. Open the Poker Peak Quad in GMDE and

make sure it is georeferenced (it should automatically be unless you have separated the
.jpg file of the map from the .txt file of the
same name (the .txt file contains the georeferencing information.) Locate the part of
the map shown to the right. Notice that a
fault is shown offsetting the stratigraphic
units.
(a) Digitize the tops of the following contacts all the way around the fold: Jn, Jp,
Js, Ke, Kp, Kb, Kd.
(b) Save the contacts as a text file (File>Save
Data>Digitized Contacts), and then Save
the contacts as .kml file (File>Export
KML>Contacts).
(c) Open the file you just saved in a spreadsheet program. Use the coordinate and
vector transformation that we derived in class, and the fold axis that you determined last week, to produce a down-plunge projection of the Big Elk an-

MODERN STRUCTURAL PRACTICE

"84

R. W. ALLMENDINGER 2015-16

TRANSFORMATIONS

CHAPTER 4

ticline. Beside to adjust the scales of the two axes so they are approximately
the same.
(d) Open the KML file in Google Earth and rotate the image around so that
you see the same down plunge projection that you produced in your spreadsheet. Capture a screen shot in
Google Earth and turn in with
your spreadsheet.
2. Back to the NE corner of the Poker
Peak Quad: find, again the ridge
shown to the right in GMDE. In
this question, we will determine the
thickness of the following units using a coordinate transformation:
TRa, JTRn, and Jt.
(a) In GMDE, select
Settings>Mode>Strike Dip
Thickness. GMDE can and will
calculate the thickness for you
(which you can use as a check on your calculations), but
for this problem, we are mostly interested in using the
program to get accurate coordinates for two points, one
on the top surface and one on the bottom surface. We
will just use the strike and dip information measured by
the field geologist for this problem.
(b) Once your have determined the coordinates of the
points at the top and bottom of each surface, in a
spreadsheet program, calculate the thickness of each
unit using a coordinate transformation from geographic coordinates to bedding plane coordinates.
3. Using the transformation matrix for rotations about an
arbitrary axis (equations 3.11, pp. 55-56 in Allmendinger
et al.), rotate the data to the right by a magnitude of 56
about an axis that has a trend and plunge of 113, 29. You
should do the rotation in a spreadsheet. You may check
your answer in a Stereonet Program.

MODERN STRUCTURAL PRACTICE

"85

Trend
113
76
175
229
75
111
78
316
25
21
81
80
58
40
42
229
110
278
264

Plunge
73
78
71
62
62
77
85
53
78
57
77
58
62
57
71
23
72
61
78

R. W. ALLMENDINGER 2015-16

TRANSFORMATIONS

CHAPTER 4

MODERN STRUCTURAL PRACTICE

"86

R. W. ALLMENDINGER 2015-16

Chapter 5
The Orientation and Stress Tensors
Introduction
The topic of tensors typically produces significant anxiety for students of
structural geology. That is due, at least in part, to the fact that the term is studiously
avoided until it is sprung on the student when introducing the topic of stress, thus
conflating two difficult to grasp concepts. In our case, the previous chapters have
already introduced the concept of vectors as a type of first order tensor so the
X3
X2
X3

X2

X1

X1

fundamental definition an entity that can be transformed from one coordinate

system to another, changing its components in predictable ways such that its fundamental nature doesnt change is already understood on an intuitive level. In
the case of our vector, it has the same magnitude and orientation on the page regardless of the orientation of the coordinate axes (Fig. 4.3). We are now fully prepared to tackle second order tensors, which will allow us to quantify the relations
between different families of vectors. These second order tensors will take us to the
very core of structural geology. We will still be doing just multiplication and additions (mostly), but first, well need to introduce some concepts about matrices.

ORIENTATION & STRESS TENSORS

CHAPTER 5

Matrices and Indicial Notation

In the last chapter, we saw that the transformation of a vector could be described with three simple equations (4.4), repeated here:
v1 = a11v1 + a12 v2 + a13v3
" v2 = a21v1 + a22 v2 + a23v3
v3 = a31v1 + a32 v2 + a33v3

(5.1)

As was pointed out at the time, the subscripts in these equations, all of which refer
to specific coordinate axes, vary in an extremely systematic and precise way. That is
because these equations are an alternative way of writing down a matrix multiplication:

v1 a11 a12

" v = av = v2 = a21 a22
v a
a
3 31 32

a13 v1

a23 v2
a33 v3

(5.2)

Note that the subscripts of the matrix a vary in exactly the same way as the subscripts, or indices, of aij in (5.1). Equation (5.2) works because there are the same
number of columns of a as there are rows of v. For this reason, you cannot write
the left side of the equation as v = va because v has only one column but a has
three rows. That is, matrix multiplication is non-commutative: the order of the
multiplication matters. This will be an extremely important insight when we come
to strain in a few more chapters; strain is mathematically represented as a matrix
multiplication so the order in which strain is superimposed determines the final
outcome.
Matrix multiplication gives us another way to write the dot product of two
vectors, where we use the transpose of one of the vectors (i.e., one of the vectors
is flipped):

" u i v = uv T = u1 u2

MODERN STRUCTURAL PRACTICE

v1

u3 v2 = u1v1 + u2 v2 + u3v3

v
3

"88

(5.3)

R. W. ALLMENDINGER 2015-16

ORIENTATION & STRESS TENSORS

CHAPTER 5

You can imagine that it gets pretty tedious to keep writing out the equations
in (5.1) or the right side of (5.2) and the bold text on the left side of (5.2), known as
matrix notation, is one type of shorthand. There is a second type of shorthand,
known as the summation convention, that is more convenient, especially when it
comes to implementing these equations in a programing language. Using the summation convention, we would write Equations (5.1) and (5.2) as:

" vi = aij v j

(5.4)

Lets break this down and see how it works. As we have already seen i and j can
vary from 1 to 3 in value because there are three axes to our Cartesian coordinate
system. There is one i one each side of the equation and it is referred to as the free
suffix; that means there will be three equations and, in each, i will have a constant
value. The index j is know as the dummy suffix and appears twice on the right
hand side, only, of Equation (5.4). Thus, the summation occurs with respect to j in
each of the three equations as shown in the following equation:

" vi = aij v j = ai1v1 + ai 2 v2 + ai 3v3

(5.5)

j=1

The recipe for the summation convention takes a little practice but it turns out to
be quite powerful and so is worth learning.
We will encounter various properties of matrices in the subsequent chapters
but for right now, well wrap up with three important terms: and matrix is symmetric if there are six independent values and the three off-diagonal components
above the principal diagonal are the same as the three below the principal diagonal
(Fig. 5.1). That is, T12 = T21, T13 = T31, and T23 =
T32. If a matrix is asymmetric then there are nine
T11
T12
T13
offindependent components and T12 T21, T13 T31,
diagonal
T21
T22
T23
and T23 T32. Finally, a matrix can be antisymT31
T32
T33
metric with three independent components where
principal
offthe values along the principal diagonal are all zeros
diagonal
diagonal
and T12 = T21, T13 = T31, and T23 = T32.

Figure 5.1 Anatomy of a 3

3 matrix.

MODERN STRUCTURAL PRACTICE

"89

R. W. ALLMENDINGER 2015-16

ORIENTATION & STRESS TENSORS

CHAPTER 5

Tensors
In this Chapter, when we use the term tensors, we are specifically referring
to second order or second rank tensors. All tensors can be expressed as matrices (a
3 3 matrix in the case of second order tensors) but not all matrices are tensors. In order to be a tensor, the matrix must represent an entity or quantity that
transforms like a tensor; that is, the components must change in a logical and systematic way during a coordinate transformation. Note that the transformation matrix, a, is itself not a tensor. A common convention is that tensors are written with
brackets, [ ], whereas matrices that are not tensors are written with parentheses, ( ).
We used this convention, for example, in Equation (5.2). Because all tensors can be
written as matrices, they can be symmetric, asymmetric, etc. and any operation that
can be performed on a matrix can likewise be performed on a tensor.

Tensors as Linear Vector Operators

The best way to think of a second order tensor is that it relates two families
of vectors. In the case of stress, discussed later in this chapter, the stress tensor relates the stress vector (or traction) on a plane to the orientation of the plane. In the
case of strain, as we shall see in a subsequent chapter, the displacement gradient
tensor (one of many tensors related to strain!) relates the position vector of a point
to the displacement of the point during a deformation. If we know the stress tensor,
we can calculate the stress vector on a plane of any orientation within a body. You
can imagine that this is hugely important in any study of earthquakes, induced
seismicity, faulting, etc. Likewise, with the displacement gradient tensor, we can calculate how all points within a body are displaced as a function of position during a
deformation.
The basic way that this relationship is written is as follows:

" ui = Tij v j

(5.6a)

Where u and v are generic vectors and T a generic tensor that relates the two vectors. This equation expands following the rules of summation or tensor notation:

MODERN STRUCTURAL PRACTICE

"90

R. W. ALLMENDINGER 2015-16

ORIENTATION & STRESS TENSORS

CHAPTER 5

u1 T11 T12

" u2 = T21 T22
u T
T
3 31 32

T13 v1

T23 v2
T33 v3

u1 = T11v1 + T12 v2 + T13v3

" u2 = T21v1 + T22 v2 + T23v3
u3 = T31v1 + T32 v2 + T33v3

(5.6b)

(5.6c)

You already know how to do this expansion because it looks a lot like Equations
(5.1-5.3). All of these equations represent the multiplication of a 3 3 and a 3 1
matrix. Beyond this mathematical similarity, there are entirely different. First, in
Equation (5.1) it is the same vector, v, on both sides of the equation whereas u and
v are two entirely different vectors in Equation (5.6). Secondly, T is a tensor as
shown by the brackets in (5.6b), where as a in Equation (5.2) is not.
There is a second way to make a second order tensor out of two vectors by
taking the dyad (tensor) product of those two vectors. In indicial notation, we
can write:

" T = u v = uT v

or

Tij = ui v j

(5.7a)

The right hand side of (5.7) does not involve any summation because there is no
dummy (i.e., repeated) suffix. So T simply works out to:

u1v1 u1v2

" Tij = u2 v1 u2 v2
uv uv
3 2
31

u1v3

u 2 v3
u 3 v3

(5.7b)

The dyad product has important applications in earthquake and faulting studies
and also in constructing the orientation tensor as we shall see below.

Principal Axes of a Tensor

Because our second order tensor is, well, a tensor, we can rotate the axes of
the coordinate system and the values of the tensor will change. It turns out that, for

MODERN STRUCTURAL PRACTICE

"91

R. W. ALLMENDINGER 2015-16

ORIENTATION & STRESS TENSORS

CHAPTER 5

X3
X2
X3

X2

Figure 5.2 The magnitude ellipsoid of a tensor and two different

coordinate axes through the ellipsoid. The red, primed coordinate
system coincides with the principal
axes of the ellipsoid. In that coordinate system, we can define the principal axes of the tensor.

X1

X1

symmetric tensors, there is one set of axes where all the off-diagonal components
will be zero and and there will only be values along the principal diagonal. Because
symmetric tensors can be represented by ellipsoids, this special set of axes occurs
when the coordinate system is aligned with the principal axes of the ellipsoid (Fig.
5.2). Not surprisingly, these are called the principal axes of the tensor and, when
referring to magnitude and not orientation, are labeled from largest to smallest
T1, T2, and T3. For the magnitude ellipsoid shown in Figure 5.2 we would write:

T11 0

" Tij = 0 T22

0
0

0 T2

0 = 0
T33 0

0
T1
0

0
T3

(5.8)

The magnitude does not have to coincide with the number of the same axis! In figure 5.2, the long axis of the magnitude ellipsoid is parallel to the X2 axis. Double
subscripts and the position in the matrix always refer to the coordinate system axes
and not to the magnitude. Thus, in this example T22 = T1.
The transformation from a random coordinate system to one where the axes
are parallel to the principal axes is usually determined numerically by solving the
eigenvalue problem. The derivation and solution to the eigenvalue problem is be-

MODERN STRUCTURAL PRACTICE

"92

R. W. ALLMENDINGER 2015-16

ORIENTATION & STRESS TENSORS

CHAPTER 5

yond the scope of this manual; see Allmendinger et al. (2012) or any book on Linear Algebra. The mathematical procedure gives us a cubic polynomial:

" 3 I 2 II III = 0

(5.9)

The three roots of are the three eigenvalues which corresponds to the magnitudes
of the three principal axes. Once one has the eigenvalues you can calculate the
eigenvectors, which give the orientations of the three principal axes. The three
coefficients, I, II, and III are known as the invariants of the tensor and they turn
out to have important physical significance in many cases. Their values are:
I = T11 + T22 + T33 = T1 + T2 + T3

(T T
" II =

ij ij

I2

) = (T T

1 2

+ T2T3 + T3T1 )

(5.10)

III = det T = Tij = T1T2T3

Where " det T = Tij is the determinant of tensor T. The significance of the invariants is that they have the same values in any coordinate system. When we say that
the fundamental nature of a second rank tensor doesnt change upon transformation, we are referring to the invariants.

Tensor Transformations and the Mohrs Circle

Given that we just mentioned tensor transformations, how does one go about
transforming a tensor? In principle, it is similar to a vector transformation: if we
have a transformation matrix and a tensor in the old coordinate system, we can
calculate the same tensor in the new system. However, it is more involved because a
second order tensor is a more complicated entity than a first order tensor (i.e., a
vector). The derivation of the tensor transformation equation is relatively straightforward; you can find it in Nye (1959), Allmendinger et al. (2102), or a variety of
other sources. Here it is using the summation convention, yielding the new tensor
in terms of the old:

MODERN STRUCTURAL PRACTICE

"93

R. W. ALLMENDINGER 2015-16

ORIENTATION & STRESS TENSORS

CHAPTER 5

" Tij = aik a jlTkl

(5.11)

This equation looks complicated and it does represent nine equations with nine
terms each! The expansion is not so much complex as it is tedious. Here, for example is the expansion of one of the nine terms:
(l eq. 1) (l eq. 2) (l eq. 3)
T = a a T +a a T +a a T
" 13 11 31 11 11 32 12 11 33 13
+a12 a31T21 + a12 a32T22 + a12 a33T23
+a13a31T31 + a13a32T32 + a13a33T33

(k eq. 1)

(k eq. 2)
(k eq. 3)

(5.12)

Dont worry! Youre not going to be asked to expand these equations. This type of
systematic equation is what computers live for and nobody solves these by hand
anymore. If you ever want to impress your friends (uh sure), the parenthetical
statements in Equation (5.12) show you how k and l vary systematically. With the
single equation arranged the way it is, you can see that k increments by one in each
row whereas l increments by 1 in each column.
Long before there were computers to do these calculations for us, in the late
1800s a German engineer by the name of Otto Mohr introduced a clever graphical
solution to a common, but special case of tensor transformation which has become
known as Mohrs Circle. In the Mohr construction, the old coordinate system is
parallel to the principal axes of the tensor. The coordinate system is then transformed by rotating by about one of the principal axes, commonly T2 (Fig. 5.3).
Thus the initial form of the tensor in the old coordinate system is:

X3

X3

X2 , X2

T3
T2

X1

Figure 5.3 The coordinate transformation

involved in constructing Mohrs circle for a
generic tensor, T, by rotating the coordinate
system around the intermediate principal axis
of the tensor

T1

MODERN STRUCTURAL PRACTICE

X1

"94

R. W. ALLMENDINGER 2015-16

ORIENTATION & STRESS TENSORS

CHAPTER 5

T1 0

" Tij = 0 T2
0 0

0
T3

(5.13)

And the transformation matrix is:

cos
cos 90 cos ( 90 )

" a = cos 90
cos 0
cos 90
cos ( 90 + ) cos 90
cos

cos

0
=
sin

0 sin
1
0
0 cos

(5.14)

For example, expanding Equation (5.12) we get:

T13 = cos ( sin ) T1 + cos ( 0 ) ( 0 ) + cos cos ( 0 )
+ ( 0 ) ( sin ) ( 0 ) + ( 0 ) ( 0 ) T22 + ( 0 ) cos ( 0 )

"

+ sin ( sin ) ( 0 ) + sin ( 0 ) ( 0 ) + sin cos T3

= cos sin T1 + sin cos T3
= (T3 T1 ) cos sin

The entire tensor, T, in the new coordinate system is:

T cos 2 + T sin 2 0
)
3
( 1
" Tij =
0
1

(T3 T1 ) sin cos 0

(T3 T1 ) sin cos

0

(T sin + T cos )
2

(5.15)

These equations can be put into a more useful form by substituting some common
double angle formulae:

" sin 2 = 2sin cos ;

sin 2 =

1 cos 2
;
2

cos 2 =

1+ cos 2

2

(5.16)

The coordinates of the tensor in the new coordinate system are:

MODERN STRUCTURAL PRACTICE

"95

R. W. ALLMENDINGER 2015-16

ORIENTATION & STRESS TENSORS

CHAPTER 5

T +T T T
T11 = 1 3 + 1 3 cos 2
2 2
T1 + T3 T1 T3

cos 2
2 2

" T33 =

(5.17)

T T
T13 = T31 = 1 3 sin 2
2

These equations are in the parametric form of a circle with

T1 + T3

T T
, 0 and a radius, r = 1 3
2
2

" center at:

Well see just how this works in the section on stress, below. However, you should
realize that just about any second order tensor can be represented by a Mohr Circle
construction. Time for some geological applications.

The Orientation Tensor

The first application harkens back to Chapter 2 and the limitations of the
mean vector calculation (Fig. 2.11) and Chapter 3 and the -diagram (Fig. 3.9). The
solution is to calculate the least squares best fit to a distribution of lines (poles, really). The result, however, will allow us to characterize any line distribution and it will
not suffer from the lower hemisphere artifacts that plague the mean vector calculation.

p [i]

Figure 5.4 Finding the best fit fold axis given

a number number of bedding poles. The dot
product of a single bedding pole unit vector p[i]
and and the fold axis unit vector, f, is the cosine
of the angle between them, . A perfectly oriented pole should be 90 from the fold axis and
so the cosine should equal zero. The deviation
from zero is the deviation we are trying to minimize.

MODERN STRUCTURAL PRACTICE

"96

R. W. ALLMENDINGER 2015-16

ORIENTATION & STRESS TENSORS

CHAPTER 5

Least Squares Best Fit Fold Axis

In Chapter 3, we simply eye-balled in a great circle which fit the bedding
poles in the -diagram (or perhaps you used a cylindrical best fit plotting option
without knowing how it worked). Here, we will actually calculate the best fitting
great circle and fold axis by minimizing the sum of the squares (Charlesworth et al.,
1976). If a unit bedding pole " p [i ] is perfectly oriented, then the angle between it and
the unit vector parallel to the fold axis, " f , is 90 (Fig. 5.4); it is the deviation from
this perfect case that we will minimize. Fortunately, there is a simple function that
we can use: the dot product of ' f and ' p [i ] , written as in Equation (5.3) is:

" cos[i ] = p [i ]f T

(5.18)

where ' p [i ] is the ith bedding pole in the data set. The sum of the squares of the
deviations of all of the poles, S, is:
" S = cos2 [i ] = ( p [i ]f T )

i=1

i=1

(5.19)

Because the dot product is commutative, it can be written as:

" p [i ]f T = f p [iT ]

(5.20)

And we can rewrite Equation (5.19) as:

" S = f p [iT ] p [i ]f T =f T f T where " T = p [iT ] p [i ] = ( pi p j )[i ]

n

i=1

i=1

i=1

(5.21)

T is known as the orientation tensor and it is composed of the sum of the dyad
products of each unit pole vector with itself (Eqn. 5.7a). Expanding for one vector:

cos
[i ]
" p T p = cos
[i ] [i ]
[i ]

cos [i ]

cos [i ]

MODERN STRUCTURAL PRACTICE

cos [i ]

cos [i ]

cos 2 [i ]

= cos cos
[i ]
[i ]

cos cos
[i ]
[i ]

(
(

"97

( cos

[i ]

cos [i ]

cos
)
) ( cos cos
2

[i ]

[i ]

) ( cos cos )
( cos cos )
cos
)
[i ]

[i ]

[i ]

[i ]

[i ]

[i ]

R. W. ALLMENDINGER 2015-16

ORIENTATION & STRESS TENSORS

CHAPTER 5

So the entire orientation tensor, T, will be:

"T=

cos
cos
cos

[i ]

[i ]

cos [i ]

[i ]

cos [i ]

cos

[i ]

cos
cos

cos [i ]

[i ]

[i ]

cos [i ]

cos [i ]

cos [i ] cos [i ]
cos2 [i ]

cos

[i ]

(5.22)

The smallest eigenvalue of this symmetric matrix is the minimization of the sum of
the deviations; thus the corresponding eigenvector should be the fold axis.

Types of Line Distributions

The orientation tensor was derived for fitting a fold axis to the distribution of
lines in a plane (i.e., the profile plane). It turns out that it is a much more generally
useful calculation. Figure 5.5 depicts three very different end member types of line
distributions and the resulting eigenvalues and eigenvectors of the orientation tensor. The bipolar distribution (Fig. 5.5a) is the one that fails miserably in a mean vector calculation because the vectors plunging in opposite directions cancel each other out. However, the orientation tensor captures the preferred orientation quite
nicely with the eigenvector of the largest eigenvalue coinciding with the NW-SE
(a)

(b)

(c)

1
1

1
2

3
3
Eigenvalue
1. 0.9278
2. 0.0597
3. 0.0125

2
Trend & Plunge
307.9, 01.1
048.3, 83.7
217.8, 06.2

Eigenvalue
1. 0.5672
2. 0.4229
3. 0.0098

Trend & Plunge

306.7, 53.0
211.8, 03.7
119.0, 36.7

Eigenvalue
1. 0.3710
2. 0.3507
3. 0.2784

Trend & Plunge

352.8, 59.6
106.5, 13.3
203.3, 26.8

Figure 5.5 Different types of lines distributions: (a) bipolar; (b) girdle; and (c)
random. Each diagram has 30 lines.

MODERN STRUCTURAL PRACTICE

"98

R. W. ALLMENDINGER 2015-16

ORIENTATION & STRESS TENSORS

CHAPTER 5

orientation of the point cluster. The relative values of the three eigenvalues clearly
identify a single cluster because there is one large eigenvalue (near 1) and two small,
near zero eigenvalues. In the girdle distribution (Fig. 5.5b), there is one small, near
zero eigenvalue that corresponds to the fold axis and two large, relatively equal eigenvalues (near 0.5) whose eigenvectors lie within the profile plane through the
points. Finally, the random distribution (Fig. 5.5c) has three relatively equal eigenvalues. Note that it would be quite possible to have a non-random distribution
produce three equal eigenvalues if you had three clusters of points representing an
equal number of lines with each cluster 90 from the others. In general, though,
the eigenvalues give you a first order way of characterizing point distributions.
A spreadsheet to calculate the orientation tensor is shown in Figure 5.6. Note
that eigenvalues suggest something between a girdle and a point distribution.

Figure 5.6 Spreadsheet for calculating the orientation tensor from trends and
plunges of ten poles to planes. Note that you only need to calculate six components
because the tensor is symmetric.

The Stress Tensor

Cauchys Law
Whereas the orientation tensor is created by the dyad product of a pole vector with itself, the stress tensor is a more normal type of linear vector operator. In
this section, Ill assume that youve already learned about stress the mechanics

MODERN STRUCTURAL PRACTICE

"99

R. W. ALLMENDINGER 2015-16

ORIENTATION & STRESS TENSORS

CHAPTER 5

type, that is in the lecture part of your structural geology class. When most people say that stress is equal to force [a vector] divided by area [a scalar] what they
are really referring to is the stress vector or traction on a particular plane of interest. The stress tensor is different: it relates the stress on planes of all possible
orientations through a point in a body to the tractions on those planes.

" pi = i j j

(5.23a)

Where pi is the traction on a particular plane, j is the direction cosines of the pole
to the plane and and ij is the stress tensor. Equation (5.23a) is very important and
is known as Cauchys Law. It expands as follows:

p1 = 111 + 12 2 + 13 3
" p2 = 211 + 22 2 + 23 3
p3 = 311 + 32 2 + 33 3

(5.23b)

We can get insight into the meaning of the components of the stress tensor by looking at some special cases. Lets assume first that the plane were interested in is perpendicular to the X1 axis. In this case, the pole direction cosines will be [1, 0, 0].
Substituting these values into Equation (5.23) we see that the tractions on the plane,
p, are pi = 11 21 31 . Likewise, a plane perpendicular to the X2 axis will
have direction cosines of [0, 1, 0] and the tractions on that plane are
12 22 32 and the plane perpendicular to X3 will have tractions of

" 13 23 33 . In other words, the components of the stress tensor are

simply the tractions on the planes that are perpendicular to the three coordinate axes.
Written as they are in the preceding paragraph, derived from Cauchys Law,
the subscripts of the stress tensor have the following meaning: the first subscript indicates which axis the traction vector is parallel to and the second subscript indicates that axis that the plane is perpendicular to. This is opposite to the convention
that one sees more commonly in structural geology textbooks. Thats okay, though,
because it turns out that the stress tensor is symmetric. That is,

MODERN STRUCTURAL PRACTICE

" 12 = 21, 13 = 31, 23 = 32

"100

(5.24)

R. W. ALLMENDINGER 2015-16

ORIENTATION & STRESS TENSORS

CHAPTER 5

X3
X3

X3

23

13

33

X1

11

X1

22

e
an

21

X2

X2

pl

plan

32

12

31

X1
p

la

X2

X1

Figure 5.7 The meaning of the components of the stress tensor. They are the normal
and shear stresses on planes perpendicular to the axes of the coordinate system. Normal
stresses have equal subscripts whereas shear stresses have unequal subscripts

Thus, it is also true that the first subscript indicates the axis perpendicular
to the plane and the second subscript
the axis to which the traction vector is
parallel (Fig.5.7). Thus each of the
three planes normal to the coordinate
system has one stress vector perpendicular to the plane and two stresses vectors parallel to the plane. The perpendicular stress, or normal stresses,
have two equal subscripts and occur
along the principal diagonal of the
stress tensor. Those parallel to the
plane, or shear stresses, have unequal subscripts and are in the off-diagonal position in the tensor. We will
use n to indicate the normal traction
(or stress vector) on a plane and either
(tau) or s for shear stress vectors on a
plane The symmetry of the stress ten-

MODERN STRUCTURAL PRACTICE

X2
p2

12
11
area = A1

po

le

p1

area = A

21

X1

area = A2

22
Figure 5.8 The traction vector, p, on a
plane whose pole that makes an angle of
with respect to X1 and with respect to X2.
The plane is viewed edge on so it appears as
a diagonal line. The traction vector is scaled
correctly given the components of the stress
tensor and angles shown.

"101

R. W. ALLMENDINGER 2015-16

ORIENTATION & STRESS TENSORS

CHAPTER 5

sor and the equivalence of the shear stresses (Eqn. 5.24) is sometimes called the
theorem of conjugate shear stresses.
Note that when we calculate the traction on a plane using Cauchys Law
(Eqn. 5.23), the vector is in the most general case neither normal nor parallel to the
plane. The components of p are always parallel to the axes of the coordinate system (Fig. 5.8). As we will see later on, one way to calculate the normal and shear
stress on the plane is to transform the vector into a coordinate system where the X1
axis is perpendicular to the plane. The normal and shear stress vectors on a plane
are key pieces of information to determine whether or not faulting, or earthquakes,
will occur.

Mohrs Circle for Stress

Like any symmetric second order tensor, stress can be represented on a Mohr
Circle construction. This graphic is a very common way of displaying stress on
planes, even if behind the scenes the values are calculated using tensor transformations. In Mohrs Circle for stress, one plots normal stress the horizontal axis and
shear stress on the vertical axis. Before making a plot, though, we need to discuss
sign conventions for both of these quantities. Although the engineering sign convention is that tension is positive and compression negative, within the Earth, almost all normal stresses are compressions except close to the surface. Thus, in geology, compression is treated as positive and tension negative. The sense of
(s)
b
si

le

id
ut s

e of

t h is c i r c le

no s
tr e

ss

es

po

50

MODERN STRUCTURAL PRACTICE

he

stress po

(n)
1

100

de

no

si

ble

50

insi

se

circ

les

"102

Figure 5.9 Basic Mohrs Circle

construction for stress where 1 =
120 MPa, 2 = 60 MPa, and 3 = 20
MPa. The only possible normal
and shear stress combinations are
along the edges of the circles or in
the pink shaded region. Two
planes are shown: that in red is
parallel to 2 and its pole is 30 (2
= 60) from 1. The blue plane is
not parallel to any principal stress;
its pole is 65 from 1, 50.74 from
2, and 49.55 from 3. Dashed
lines are construction lines.

R. W. ALLMENDINGER 2015-16

ORIENTATION & STRESS TENSORS

CHAPTER 5

shear also has a sign and this, too, depends on your perspective. If one derives the
Mohrs Circle in the traditional way as a force balance, counterclockwise (i.e., leftlateral) shear is positive and clockwise (i.e., right-lateral) shear is negative. However,
if it is derived as a tensor transformation, then the opposite convention applies. We
will use the geological sign convention.
The equations for the Mohrs Circle for stress for planes that are parallel to
2 are:

" n =

1 + 3 1 3
+
cos 2
2 2

" = s =

1 3
sin 2
2

(5.25a)

(5.25b)

is the angle between the pole to the plane of interest and 1. Similar equations

can be derived for the two inner circles. There are many ways of deriving the
Mohr Circle equations: as a force balance, using Cauchys law and a vector transformation, or using a tensor transformation (Eqns. 5.11 - 5.17).
In the most general case of Mohrs Circle for stress, there will be three nested
circles defined by the three principal stresses (Fig. 5.9). The simplest case for plotting normal and shear stresses is when the plane of interest is parallel to one of the
three principal axes (e.g., the red dot in Fig. 5.9). In that case, one simply determines the angle between the pole to the plane and the largest principal stress, ,
and then measures 2 counterclockwise from the side of the circle with largest
principal stress as shown in Figure 5.9. When the plane is not parallel to any of the
principal stresses, the manual plotting procedure is more cumbersome and consists
of drawing arcs of concentric circles defined by the three angles, , , and , that
the pole makes with the three principal stresses (see Allmendinger et al., 2012, if
you want to know how to do this). A program like MohrPlotter can do this calculation and plot directly. But, of course, it is more satisfying to calculate the stresses
directly! We can use Equation (5.23) to calculate the tractions on the plane directly
and this is done in the spreadsheet in Figure 5.10 for the same angle values as blue
plane in Figure 5.9. In the next chapter, well actually calculate the normal and
shear stresses.
MODERN STRUCTURAL PRACTICE

"103

R. W. ALLMENDINGER 2015-16

ORIENTATION & STRESS TENSORS

CHAPTER 5

Figure 5.10 spreadsheet for calculating the

tractions on a plane
whose pole makes angles
of alpha, beta, and gamma with respect to the
three principal stresses.

A couple of things bear emphasizing in the example in Figure 5.10. First,

this calculation takes place in the principal stress coordinate system, which is why
the stress tensor has such a simple appearance. Without knowing the orientation of
the plane or of the principal stresses, we have no idea the values of the tractions in
a geographic coordinate system. The same is true of the Mohrs Circle construction: it represents a principal stress coordinate system; without knowing the orientations of the stresses, a plane represented could be a normal, thrust, or strike-slip
fault plane. Second, the angles , , and are not independent of each other; only
two out of the three can be independently specified. Their direction cosines must
yield a unit vector (thus the magnitude test in column H).

Mean Stress & Deviatoric Stress

Earlier in the Chapter, we saw that symmetric tensors have invariants, and
the first invariant of the stress tensor is the sum of the stresses along the principal
diagonal. This allows us to define the mean stress, which is the same in all coordinate systems:

" m =

11 + 22 + 33 1 + 2 + 3
=

3
3

(5.26)

The total stress can then be defined as the mean stress plus the deviatoric stress:

m 0
0

" ij = 0 m 0
0
0 m

MODERN STRUCTURAL PRACTICE

11 m
12
13

21
22 m
23
+

31
32
33 m

"104

(5.27)

R. W. ALLMENDINGER 2015-16

ORIENTATION & STRESS TENSORS

CHAPTER 5

ExercisesChapter 5
All of the following exercises should be done either in a spreadsheet or Matlab. If
you use a spreadsheet, you will need to download the simple utility program
EigenCalc in either Macintosh or Windows version:
http://www.geo.cornell.edu/geology/faculty/RWA/programs/utility-programs/eigencalc.html

This program will take the matrix that you calculate in the spreadsheet and calculate the eigenvalues and eigenvectors for you, which is much less cumbersome than
trying to figure out how to get the spreadsheet to do it!
You might also find useful a program to plot Mohr Circles for stress, MohrPlotter, which can be downloaded from:
http://www.geo.cornell.edu/geology/faculty/RWA/programs/mohrplotter.html

Even though MohrPlotter can do some of the calculations for you, you must still do
the calculations in a spreadsheet or Matlab.
1. In a spreadsheet or in Matlab, calculate the orientation tensor for the strikes
and dips that you used to construct a pi-diagram of the Big Elk anticline last
week. Note that you do not have to find the eigenvalues or eigenvectors of the
resulting matrix, just construct the orientation tensor in a North-East-Down
coordinate system. If you choose to use Matlab, please write your own code
rather than just cutting and pasting the code from Allmendinger et al.
2. A state of stress with the following principal stress magnitudes 1 = 40 MPa, 2
= 20 MPa, 3 = 10 MPa, has 1 axis oriented vertically, 2 aligned in a horizontal E-W direction, and 3 in a horizontal N-S direction. Using Cauchys law,
calculate the magnitude and orientation of the tractions acting on a plane striking 060 and dipping 55 SE.
3. For the same state of stress described in question 2, but a different plane that
has a strike in right hand rule of 270, 60,
(a) Use Cauchys law to calculate the tractions on the plane
(b) Use Mohrs Circle for stress to calculate the normal and shear stress on the
plane
MODERN STRUCTURAL PRACTICE

"105

R. W. ALLMENDINGER 2015-16

ORIENTATION & STRESS TENSORS

CHAPTER 5

(c) Reconcile your answers in parts (a) and (b) by showing that they are actually
the same answer.
(d) If the plane with strike of 270, 60 were a fault plane, what type of fault
would it be?
4. The following matrix gives the stress tensor in a NED coordinate system (units
are MPa):
20 6 15
" ij = 6 45 20
15 20 30

(a) Use Cauchys Law to determine the tractions on a plane with a strike and
dip (right hand rule) of 045, 22.
(b) The principal stresses for this stress tensor, given by calculating the eigenvalues and eigenvectors, are 1 = 63.23 MPa with a trend and plunge of
068, 36; 2 = 24.4 MPa, 317, 26; and 3 = 7.39 MPa, 200, 43. Plot the
normal and shear stress on the 3D Mohr circle for stress as shown on page
110 of ACF (2012).
5. Two planes are both parallel to 2. The first has a normal stress, n = 55 and
shear stress, = 17 MPa. For the second, n = 27 and shear stress, = 13 MPa.
(a) Use a Mohr Circle construction to calculate the magnitude of the principal
stresses, 1 and 3. [Hint: the perpendicular bisector of a chord in a circle
goes through the center of the circle].
(b) If plane 1 has a strike and dip (RHR) of 084.7, 40.3 and plane 2, 046.3,
70.7, use a stereonet to determine the trends and plunges of the three principal stresses.

MODERN STRUCTURAL PRACTICE

"106

R. W. ALLMENDINGER 2015-16

Chapter 6
Faulting and Stress
Introduction
We now have the tools necessary to begin to understand faulting. Faulting is
probably the most important type of deformation in the upper crust of the earth
and, as well see in later chapters, is often accompanied by folding. An earthquake
is just a single, rapid (i.e., geologically instantaneous) movement on a fault and induced seismicity are earthquakes cause by humans, often due to subsurface

pumping of fluids and the interaction of those fluids with pre-existing faults. Needless to say, these are phenomena of great, and increasing, societal interest!
As you know from the lecture part of your course, a fault is a break or fracture in the rock across which there has been movement parallel to the fault surface.
We know that movement parallel to a surface is called shear and so it wont come
as a surprise that the shear stress on a fault plane is an important parameter to
quantify. However, as we shall see, the normal stress on a fault plane is equally important. We now have the ability to analyze these parameters and will do so after a
brief review of fault geometry and the introduction of two additional key concepts:
friction and pore pressure.

FAULTING & STRESS

CHAPTER 6

Mode I!
(opening)

Mode II!
(sliding)

Mode III!
(tearing)

Figure 6.1 The three modes of cracking, with the crack surface colored
red. Mode I cracks correspond to joints or veins because they have opening but no shear. Modes II and III are faults.

Fault Geometry Review

Terminology
Cracks or fractures in a rock can occur in one, or a combination, of three
fundamental modes (Fig. 6.1). In mode I, there is opening perpendicular to the
crack but no shear and we call the resulting geological structure a joint or, if it is
filled with mineral precipitate, a vein. Most faults in the subsurface are elliptical
surfaces that terminate in a tip line (really a loop) which separates faulted from unfolded (i.e., intact) rock. Depending on the position along the fault surface, the fault
may resemble either a mode II (shearing perpendicular to the tip line) or a mode
III (shearing parallel to the tip line) crack. For example, near where a dip slip fault
dies out along strike, you are likely to see mode III style behavior whereas in the
middle of the fault trace, mode II would be more common.
Faults are classified in a variety of ways (Fig. 6.2). For faults in which the
movement is parallel to the dip direction, they are further subdivided in terms of
whether the hanging wall block moved down (normal faults, Fig. 6.2a) or up (reverse faults, Fig. 6.2b). Faults that moved parallel to the strike likewise are subdivided into whether the opposing block (i.e., the one opposite the the block the observer is on) appears to move to the right (Fig. 6.2c) or left (Fig. 6.2c). Of course,
few faults fit these end members exactly and most have combinations of movement
parallel to strike and to dip; these are known as oblique-slip faults.
Faults can also produce a component of rotation of one block relative to the
other. If the two blocks are to remain in contact (a constraint known as strain comMODERN STRUCTURAL PRACTICE

"108

R. W. ALLMENDINGER 2015-16

FAULTING & STRESS

CHAPTER 6

Dip Slip Faults

(a)

(b)

Figure 6.2 Basic endmember fault types. In

each case, the red arrow is
a slip vector which shows
the offset of two originally
adjacent points. The slip is
in the dip direction in (a)
and (b), and in the strike
direction in (c) and (d). The
stick figure is holding out
their right arm in (c) and
their left arm in (d).

footwall
hanging
wall

Normal

Reverse

Strike Slip Faults

(c)

Right-lateral

(d)

Left-lateral

patibility, which we will see later on), then only two different orientations of the rotation axis are possible: (a) perpendicular to the fault plane, or (b) parallel to the
fault plane. The latter case is only possible if the fault plane is a curved surface.
Curved fault surfaces where the dip diminishes with depth are known as listric
faults. A rotation axis perpendicular to the plane will produce a scissors fault
(Fig. 6.3)
Planar features like stratification are common in geology and it is tempting to
think that the apparent offset that one sees of such features, either on the surface or
in cross section, represents the actual movement of the fault. We call such apparent
offset of a planar feature separation (Fig. 6.4) The actual vector displacement of
(b)

(a)

Figure 6.3 Rotational faults: (a) listric normal fault where the rotation axis
(in red) is parallel to the curved fault surface. (b) Scissors fault where the rotation axis is perpendicular to the planar fault plane.

MODERN STRUCTURAL PRACTICE

"109

R. W. ALLMENDINGER 2015-16

FAULTING & STRESS

CHAPTER 6

(a)

strike separation

(b)

slip

dip separation

Figure 6.4 Slip versus separation. (a) Block diagram illustrating an offset planar layer
that contains a linear feature shown as a dashed line. The linear feature intersects the
fault plane in two red points known as piercing points and the slip vector (heavy red arrow) connects the piecing points. Note that the strike and dip separation shown are neither
parallel to, nor the same magnitude as the slip vector. (b) schematic map view of a plunging syncline offset by a fault. Note that the strike separation is in the opposite direction on
the two limbs of the fold!

two originally adjacent points across a fault surface is called slip. It is critically important to distinguish clearly between these two concepts. Separation is not very
useful (e.g., Fig. 6.4b) except when accompanied by independent evidence such as
observation of striations or slickensides on the fault surface. It is seldom possible to
find two originally adjacent points that have been offset across a fault so, instead,
we use offset linear features. A line intersects a plane in a point, and therefore our
offset lines intersect the fault plane in what are known as piercing points. The slip
can be determined if piecing points on the fault surface can be calculated.

Determining Slip from Piercing Points

As we have just seen, separation, alAs you begin to derive your own
though easy to determine, is not very useful.
functions related to geometry, you
Slip, on the other hand, is incredibly useful for a
may find the web site by Paul
whole range of problems in structural geology.
Bourke (http://paulbourke.net)
Slip is most effectively determined from offset
very useful. The derivation in this
piercing points of linear features so the question
section is modified from an algorithm on his site.
becomes: how does one determine a piercing
point, especially given that, in general the piercing point will either be in the subsurface or in the air and thus not directly observable? The vector methods we learned in earlier chapters come to the rescue for the

MODERN STRUCTURAL PRACTICE

"110

R. W. ALLMENDINGER 2015-16

FAULTING & STRESS

CHAPTER 6

p2
p1!
(fw)

p!
(fw)
p!
(hw)

uv

p1!
(hw)

Figure 6.5 Calculation of piercing points, p, in the hanging wall

(hw) and the footwall (fw) of a
fault plane shown as a gray-shaded surface. n is the normal or the
pole to the fault plane and v is the
unit vector defined by the trend
and plunge of the line that produces the piercing points. Only
p1, p2, n and v are needed to calculate the piercing points, p.

solution to this problem (Fig. 6.5). To solve this problem, we need to know one
point, p1, on the line with a known trend and plunge that pierces the plane, and
one point, p2, on the plane as well as the strike and dip of the plane. Points p1 and
p2 will usually be points on the surface of the earth or well determined (e.g.,
drilled) points in the subsurface. The line, of course could be any geological feature
that can be identified fold axis on a particular stratigraphic horizon, the intersection of a dike and bedding, the intersection of a stratigraphic horizon beneath
an angular unconformity as long as we can determine its trend and plunge and
at least one point on each sides of the fault plane.
For simplicitys sake, our derivation only treats one of the two piercing
points. As you can see in Figure 6.5, the dot product of the normal vector, n, and
the vector between the position vectors p2 and p (the point in which we are interested) is zero because n is perpendicular to all vectors in the plane:

" n i [ p - p2 ] = 0

(6.1)

Point p also occurs along the line that pierces the plane. The vector distance between p1 and p is " uv where " v is the unit vector calculated from the trend and
plunge of the linear feature. We can write that:

" p = p1 + uv

(6.2)

Substituting Equation (6.2) into (6.1) and rearranging, we get:

MODERN STRUCTURAL PRACTICE

" n i [ p1 + uv ] = n i p2

"111

(6.3)

R. W. ALLMENDINGER 2015-16

FAULTING & STRESS

CHAPTER 6

And solving for u we have:

"u =

n i [ p2 p1] n1 ( p21 p11 ) + n2 ( p2 2 p12 ) + n3 ( p2 3 p13 )

=

n i v
n1v1 + n2 v2 + n3v3

(6.4)

There is one special case that will result in an error: if the pole to the plane is perpendicular to the line then the denominator of (6.4) will be zero. This means that
the line is parallel to the plane and either does not intersect it at all or lies within
the plane and there are an infinite number of intersections! Once we know the
scalar u, we can substitute it into Equation (6.2) and solve for the the piercing point,
p:

" p1 = p11 + uv1;

p2 = p12 + uv2 ;

p3 = p13 + uv3

(6.5)

We can repeat this process for the piercing point on the other block of the
fault. Note that the trends and plunges of the lines on the two sides of the fault do
not have to be equal to each other, thus accounting for faults across which there has
been rotation. The final slip magnitude on the fault is just the magnitude of the
vector connecting the two piercing points:

" slip =

( p(hw)1 p( fw)1 )2 + ( p(hw)2 p( fw)2 )2 ( p(hw)3 p( fw)3 )2

(6.6)

Where p(hw) is the piercing point in the hanging wall and p(fw) is the offset piercing
point in the footwall (Fig. 6.5). A sample calculation is shown in the spreadsheet in
Figure 6.6. The expanded equation in the spreadsheet is the realization of Equation (6.4). The only special note is that one must remember to calculate the direction cosines of the trend and plunge of the line and the pole to the plane in EastNorth-Up coordinates rather than North-East-Down. The formula for Equation
(6.6) is in cell B12 of the spreadsheet.
Note that the calculation we have just done can be used for a variety of other
useful tasks in geology. For example, you could use this same approach to determine the distance that you would need to drill along a particular trend and plunge,
starting at some point on the surface, to intersect a fault plane, knowing nothing
more than the strike and dip of the plane and a single well determined point on the
plane. The scalar distance, u, is given by Equation (6.4).
MODERN STRUCTURAL PRACTICE

"112

R. W. ALLMENDINGER 2015-16

FAULTING & STRESS

CHAPTER 6

Figure 6.6 spreadsheet to calculate the slip on a fault, given the strike and dip of the
fault plane and a position vector of a point in the fault plane (p2), and the trend and
plunge of a line in the hanging wall and footwall, position vectors somewhere along
those lines, and the trends and plunges of the lines. In this example, we have assumed
that the trend and plunge of the line on either side of the fault is the same but that is
not necessary. The calculation of the scalar quantity, u, has been expanded for view.

Stress and Faulting

Faulting occurs when the yield stress of a material is exceeded and the material fails locally. Ductile failure is an extremely important process in geology but we
are going to leave it for a later chapter. In this chapter, we are only concerned with
brittle failure, friction, reactivation, and the influence of fluid pressure in the pores
of the material. Thus, the material here corresponds to the upper 105 km of the
Earths crust, an area of particular interest for humans because of earthquakes, as
well as exploitation of the subsurface for resources (hydrocarbons, water) and storage of waste (e.g., water water, carbon sequestration, etc.). There are two separate
questions: (a) what are the macroscopic conditions under which a rock will break?
and (b) under what conditions will preexisting fractures be reactivated? Pore fluid
pressure will be important in both cases, but we will first examine the case without
pore pressure.

Failure in the Brittle Realm

You have, no doubt, in the lecture part of your structural geology course
been introduced to the concept of a failure envelope. This envelope is most often
depicted on a plot of shear stress versus normal stress as a line, symmetric about
the normal stress axis, with several distinct parts (Fig. 6.7). A combination of normal and shear stress that lies outside of the envelop is one that will cause the materMODERN STRUCTURAL PRACTICE

"113

R. W. ALLMENDINGER 2015-16

FAULTING & STRESS

CHAPTER 6

ial to fail whereas those inside the envelope are stable. The important points on the
failure envelop are: To tensile strength; 1. transitional-tensile behavior; So
cohesive strength; 2. Coulomb failure; and 3. von Mises failure. The last
of these is equivalent to ductile failure and will not be discussed further in this
chapter. The field of Coulomb behavior is the most important for upper crustal
faulting whereas tensile strength and transitional-tensile behavior are most important for near surface jointing or for cases of small differential stress ( = 13)
and large pore fluid pressures, Pf. We will start with Coulomb failure, the linear part
of the envelope.
For Coulomb failure, only the 1-3 part of the Mohrs circle matters because, in plane strain, planes parallel to 2 will always fault before planes in other
orientations (except for the cases where 2=3 or 2=1). At failure, the Mohrs circle
touches the failure envelope in two places (Fig. 6.8). Somewhat paradoxically at first
glance, the plane that actually breaks is not the plane at 45 to 1 which has the
maximum shear stress, max, on it! Instead, there are two potential conjugate fault
planes with their poles at angles of = (45+/2) to 1; , the angle of internal
friction, is the slope of the Coulomb part of the failure envelope and tan() = =
/n is the coefficient of internal friction. The complete equation for the
Coulomb part of the failure envelope is:

or (s)
3. von Mises failure

2.

ou

b
lo m

u
fail

re

So
1.
3

(n)

To

MODERN STRUCTURAL PRACTICE

"114

Figure 6.7 The anatomy of the

failure envelope, plotted on axes
of shear stress (vertical axis) and
normal stress (horizontal axis).
The Mohrs circle for stress is
also shown.

R. W. ALLMENDINGER 2015-16

FAULTING & STRESS

CHAPTER 6

(s)

max

50

So
2
To

n 50

100

(n)

Figure 6.8 Coulomb failure. The

Mohr circle intersects the failure envelope in two places (two red dots) so two
different conjugate planes are equally
likely to fracture in shear. is the angle
of internal friction and is the angle
between the pole to the fault plane and
1.

" = s = So + n

(6.7)

Where So is the cohesion, essentially the shear strength of a material under zero
normal stress load. The reason why the plane with max does not fault is that it has
much higher normal stress on it than the plane that does fault. We call this balance
between shear stress and normal stress friction.
Equation (6.7) is written in terms of the normal and shear stress at failure
but there are times when we want to know the Coulomb failure criterion in terms
of the principal stresses. There are several ways to achieve this result. For example,
you could substitute Equations (5.25) into (6.7) or you can do a simple graphic construction based on the Mohrs Circle and Coulomb failure envelopes. The failure
criterion in terms of the principal stresses is (Eqn. 6.8):

" 1 = Co + K 3

where

K=

1+ sin
1 sin

and

Co = 2So K

(6.8)

Reactivation of Pre-existing Planes of Weakness

The upper crust of the earth is anything but homogeneous and isotropic:
even a casual glance at any outcrop will reveal numerous pre-existing planes of
weakness. The rocks are beset with fractures both joints and faults in a variety of orientations and stratification itself can commonly represent a significant anMODERN STRUCTURAL PRACTICE

"115

R. W. ALLMENDINGER 2015-16

FAULTING & STRESS

CHAPTER 6

isotropy. These pre-existing planes of weakness commonly have little or no cohesion and, because the fracture is already there, the coefficient of static friction on
the plane, s, is different than the coefficient of internal friction. Thus, the failure
envelope is different for pre-existing fractures than for Coulomb failure. The formula for the failure envelope for slip on preexisting planes is:

" = s n

(6.9)

As you can see in Figure 6.9, the Mohrs Circle crosses the envelope for failure along pre-existing planes with zero cohesion but does not touch the Coulomb
part of the failure envelope for intact rock. Just because it crosses the pre-existing
failure envelope, however, does not mean that failure will actually occur; the preexisting weak planes must lie within a range of orientations defined by the intersections. In the case of Figure 6.9, the poles to the preexisting planes must be oriented
between 48.75 and 79.25 to 1 (of course, the doubles of those angles are whats
(s)
50

Co

m
ulo

ail
bF

ure

tan s

1
79.25
48.75

2 = 97.5

(n)

2 = 158.5

50

100

red dot
corresponds
to this plane
range of orientations
of reactivated faults

Figure 6.9 Left: Mohrs circle for stress and failure envelopes for preexisting fractures (in red) and for Coulomb failure. The circle intersects the preexisting fractures
envelope but not the Coulomb failure envelope. Planes with normal and shear stress
that plot in the pink region will be reactivated. Right: the physical orientation of 1
and the range of orientation of potential planes (in cross-section view so they appear
as lines) that will be reactivated under these stress conditions in the shaded pink region. The red plane on the right corresponds to the red dot in the Mohr diagram on
the left.

MODERN STRUCTURAL PRACTICE

"116

R. W. ALLMENDINGER 2015-16

FAULTING & STRESS

CHAPTER 6

plotted on the Mohrs Circle). In the rectangular rock sample on the right hand side
of Figure 6.9, the regions where reactivation will occur are shown in pink color. If
there are no weak planes with those orientations, then failure will not occur. If
there are (e.g., the plane represented by the red dot), and loading occurs slowly
enough, the failure on existing planes will relieve the stresses and prevent them
from getting large enough to cause new faults to form.
A geologist at the U.S. Geological Survey, James Byerlee (1978), summarized
a large amount of experimental data on rock friction and demonstrated that, to a
first order, s, is independent of rock type (Fig. 6.10), a relationship that has become
known as Byerlees Law. For a wide variety of rock types, the data show that:

" = 0.85 n for n 200 MPa

and

= 0.5 + 0.6 n for n > 200 MPa

(6.10)

The only significant deviations from Byerlees Law are rocks composed of the clay
minerals, illite, vermiculite, and montmorillonite. However, these clay minerals are
commonly found in fault gouge.

Figure 6.10 Experimental data

summarized by
Byerlee (1978)
showing that s is
commonly relatively independent of
rock type.

MODERN STRUCTURAL PRACTICE

"117

R. W. ALLMENDINGER 2015-16

FAULTING & STRESS

CHAPTER 6
(s)
50

Pf

Pf
*1

50

Figure 6.11 Mohrs Circle

and failure envelopes in the
presence of pore fluid pressure, Pf (blue circles and arrows). The initial stress state
in the absence of pore pressure
is shown as light gray circles
which do not intersect either
failure envelope.

(n)

100

The Effect of Pore Fluid Pressure

One of the most important controls on the strength of rocks in the upper
crust is the pressure of fluid in the pores of the rocks. A fundamental property of
fluids is that they cannot support any shear stress in any direction. Thus, every
plane in a fluid is free of shear stress and thus every direction in a fluid is a principal stress direction. This is known as a spherical state of stress because the stress,
or pressure, is the same in every direction. Because pressure in the pores of the
rock pushes outward in every direction, it affects the normal stress but not the shear
stress. Thus, we can write the effective stress tensor as:

P
11
f

" ij = 21

31

12

22

Pf

32

13

23

33

Pf

(6.11)

where Pf is the pore fluid pressure. This has a profound effect on the deformation,
most easily illustrated with the Mohrs Circle (Fig. 6.11). The differential stress, 1
3 or the diameter of the circle, is unaffected by pore pressure but the entire circle is
shifted to the left because the normal stresses are reduced. Thus a rock which exists
in a stable stress state without pore pressure can be made to fracture, or slip on pre-

MODERN STRUCTURAL PRACTICE

"118

R. W. ALLMENDINGER 2015-16

FAULTING & STRESS

CHAPTER 6

existing weak planes, simply by raising the pore fluid pressure. All of our failure criteria in Equations 6.7, 6.8, and 6.9 really ought to be written in terms of effective
normal stresses. For example, Coulomb failure should be written:
" = s = So + ( n Pf ) = So + n*

(6.12)

This process is known as hydraulic fracturing, sometimes referred to by

the increasingly pejorative term fracking, the new f-word in environmental politics. Many natural processes can produce elevated pore pressure which can lead to
natural hydraulic fracturing. Many human enterprises water well improvement,
oil and gas development, enhanced geothermal, carbon sequestration also either
rely on hydraulic fracturing (commonly referred to as stimulation), or inadvertently weaken the rock while storing fluids in the subsurface (waste water storage,
carbon sequestration). Weakening of pre-existing buried fault zones has been implicated in human induced seismicity as you will see in the exercises.
An important parameter for many studies is the pore fluid pressure ratio,
, a term first introduced by Hubbert and Rubey (1959). This ratio is defined as:

" =

Pf
Plithostatic

where

Plithostatic = gz

(6.13)

varies between 0 and 1; when it equals 1, the pore fluid pressure is high enough to
support the weight of the overlying column of rock. In natural settings such as the
Gulf Coast or the Barbados accretionary prism, values of approaching 1 have
been measured. Hydrostatic pressure is the weight of a column of fluid in the
interconnected pores in a rock. In that case, = 0.4, assuming an upper crustal
rock density of 2500 kgm3.

Calculating the Normal and Shear Stress Vectors on any Plane

Clearly, from all of the preceding discussion, the normal and shear stress vectors on a plane is a fundamental parameter to determine. In Chapter 5, we saw
how to calculate stress on a plane for a special case where the principal stresses
were parallel to the axes of our coordinate system. Here, we tackle the more gener-

MODERN STRUCTURAL PRACTICE

"119

R. W. ALLMENDINGER 2015-16

FAULTING & STRESS

CHAPTER 6

N
co
s 1
a1

s
1
1

s
co

cos c13

cos

c 11

co
s

a1

a 13

n
cos 1
c

12

Figure 6.12 the coordinate

systems used to calculate normal and shear stress on a plane
(black great circle). The first is
the North-East-Down (NED)
system, the second is defined
by the principal stresses (in
red), and the third is the fault
plane system (nbs) in blue.The
direction cosines of the new
coordinate system in the old
coordinate system define the
transformation between the
two. t is the traction vector
on the plane, which is coplanar
with the pole (n) and the slip
vector (s) on the plane.

al case where the principal stresses, and the plane, can have any orientation. This
general case is depicted in Figure 6.12.
There are three coordinate systems and thus two transformations. We start in
our usual NED coordinate system because the orientations of the principal stresses
and planes are specified by their trend and plunge or strike and dip. The second
coordinate system is defined by the principal stresses. The first transformation matrix, a, is composed of the direction cosines of the principal stresses in the NED
coordinate system (Fig. 6.12). The third coordinate system is defined by the pole to
the plane, n (X1), the slip direction on the fault plane, s (X2), and null or b axis (X
3), on the fault plane. The second transformation matrix, c, is given by the direction cosines of nbs in the principal stress coordinate system. Here are the steps to
calculate the normal and shear stress orientations and magnitudes on the plane of
interest:
1. Calculate the direction cosines of the principal stresses, and the pole
to the plane, in the NED coordinate system (i.e., X1X2X3). We assume
that the trends and plunges of the three principal stresses are already

MODERN STRUCTURAL PRACTICE

"120

R. W. ALLMENDINGER 2015-16

FAULTING & STRESS

CHAPTER 6

known, so we can just use the equations in Table 2.1 of Chapter 2.

Lets assume the those trends and plunges are labeled trdi and plgi
where i is the number of the principal stress (e.g., trd1 is the trend
of 1, etc.):

cos(trd 1 )cos( plg 1 ) sin(trd 1 )cos( plg 1 ) sin( plg 1 )

" aij = cos(trd 2 )cos( plg 2 ) sin(trd 2 )cos( plg 2 ) sin( plg 2 )
cos(trd 3 )cos( plg 3 ) sin(trd 3 )cos( plg 3 ) sin( plg 3 )

(6.14)

2. Transform the pole, n, into the principal stress coordinate system

(X1X2X3). We use our standard vector transformation to accomplish
this:

" ni = aij n j

(6.15)

where nj are the direction cosines of the plane in the NED coordinate
system and aij is the transformation matrix composed of the direction
cosines of the principal stresses calculated in Equation (6.14).
3. Calculate the traction vector, t, in the principal stress coordinate system using Cauchys Law (Eqn. 5.23). Because we are in the principal
stress coordinate system, the stress tensor has the form:

1 0 0

" ij = 0 2 0
0 0
3

(6.16)

" ti = ij n j

(6.17)

Where nj is the pole direction cosines calculated in Equation (6.15).

4. Calculate b' and s' in stress coordinates from the following cross
products (from Eqn. 2.8):

MODERN STRUCTURAL PRACTICE

"121

R. W. ALLMENDINGER 2015-16

FAULTING & STRESS

CHAPTER 6

"

b = n t = ( n2t 3 n3t 2 ) ,

( n3t1 n1t 3 ), ( n1t2 n2t1)

s = n b = ( n2b3 n3b2 ) ,

( n3b1 n1b3 ), ( n1b2 n2b1)

(6.18)

Because t is not a unit vector, b and s are also not unit vectors. We
need to convert these to unit vectors by dividing by their magnitude
in order to use them as the second transformation matrix, c.

" cij =

n1

n2

b1
b

b2
b

s1
s

s2
s

n3

b3
b

s3
s

(6.19)

Note that, unlike b and s, we dont need to divide n by its magnitude because n is already a unit vector.
5. Transform t into the fault coordinate system using the transformation
matrix, c.

" ti= ci j t j

(6.20)

In this coordinate system, t1 is the normal stress magnitude on the

plane and t3 is the shear stress magnitude. We could do this step using a tensor transformation, but we have already calculated the traction vector so we do this simpler calculation that is easy to implement
in a spreadsheet.
6. Of course, we want to know the orientations of the shear stress vector
(we already know the orientation of the normal stress because it is
parallel to the pole to the plane). We already know s in the principal
stress system so we can use a reverse transformation from that system
to the NED system using the transpose of the transformation matrix,
aT. Assuming that we have calculated s in Equation (6.19), the for-

MODERN STRUCTURAL PRACTICE

"122

R. W. ALLMENDINGER 2015-16

FAULTING & STRESS

CHAPTER 6

mula for getting the orientation of the shear stress on the fault plane
is:

" si = a ji s j

(6.21)

Note that the indices of a have been switched in Equation (6.21).

7. The final step is to convert the direction cosines of s from Equation
(6.21) back into trend and plunge format. At this point, you have had
lots of practice with that but if you have forgotten, check out Equations (2.11).
The spreadsheet that accomplishes all of these steps is shown in Figure 6.13. It has
been set up for the same plane and principal stress orientations as shown in Figure
6.12. When you enter principal stress orientations in a spreadsheet, make sure that
the three axes are perpendicular to each other! This can be accomplished with a
stereonet program. Programs such as MohrPlotter ensure that the axes are orthogonal by only allowing you to enter the truly independent parameters. This may
seem to be a particularly tedious calculation, but as you will see in the Exercises, it
is very powerful with a lot of practical importance for serious issues of our times.

Figure 6.13
The spreadsheet to calculate the normal
and shear
stress on a
plane of any
orientation,
given a randomly orientated set of principal axes.

MODERN STRUCTURAL PRACTICE

"123

R. W. ALLMENDINGER 2015-16

FAULTING & STRESS

CHAPTER 6

The Principal Stress Ratio

There is one final issue that is important for understanding the relationship
between stress and faulting: the importance of 2 relative to 1 and 3. This relationship is called the principal stress ratio and it takes the form of:

"R=

2 1 c13c23
=

3 1 c12 c22

(6.22)

Remarkably enough, this ratio depends on nothing more that the direction cosines
of the second transformation matrix, c, as you can demonstrate to yourself by calculating R from c in Figure 6.14. R varies between 0 and 1; when R = 0, 2 = 1
and when R = 1, 2 = 3. For planes that are not parallel to a principal stress, R can
have a very significant effect on whether or not the plane is likely to be reactivated
because the magnitude (but not the orientation) of the normal stress on the
plane varies with R as well as the orientation of shear stress on the plane. Because R is a function of the direction cosines of c, it is the one additional parameter that can be determined when fault planes or earthquake focal mechanisms are
inverted for principal stress orientations.

R=0

R=1

Figure 6.14 The variation in orientation of shear stress resolved on a plane

with change in the parameter, R. The
magnitudes of 1 and 3 are held fixed
and the only thing that varies is the magnitude of 2. The fault plane and orientation and magnitudes of the stresses are
the same as in Figures 6.12 and 6.13. The
blue dot at R = 0.6 is the actual value calculated in Figure 6.13. In this case, the
variation in shear stress on the plane is
about 40; with other geometries the
variation can approach 90.

n
2

MODERN STRUCTURAL PRACTICE

"124

R. W. ALLMENDINGER 2015-16

FAULTING & STRESS

CHAPTER 6

Although we will not do a full inversion for stress in this course, it is a common calculation amongst structural geologists and geophysicists. Some authors use
a different formulation of the principal stress ratio:

"=

2 3

1 3

(6.23)

Thus, = 1R. As were talking about inverting fault planes and earthquakes for
stress, we should wrap up this section with a note of caution. All methods to do this
calculation have two assumptions in common: (a) that faults slip in the direction of
maximum resolved shear stress on the plane, and (b) that the stress orientation and
magnitude does not vary during the faulting. Both of these assumptions require
careful evaluation as there are many instances where they could be questioned. In a
subsequent chapter, we will see a simpler approach to analyzing faults using infinitesimal strain.

MODERN STRUCTURAL PRACTICE

"125

R. W. ALLMENDINGER 2015-16

FAULTING & STRESS

CHAPTER 6

ExercisesChapter 6
All of the following exercises should be done either in a spreadsheet or Matlab. You
will need the MohrPlotter program that you downloaded last week for some of the
exercises.
1. The map, below, is a modified version of the Mystic Lake Quadrangle. The objective is to calculate the slip on the unnamed fault that crosses the image and
has a strike and dip of 299, 38 (RHR). Coordinates of three points are given,
one on the fault plane and the others on the fold axis at the contact between
Kmt and Kk in the hanging wall and the footwall. Southwest of the fault, the
fold axis has a trend and plunge of 056, 25 and to the northeast it is 042, 12.
Construct a spreadsheet (or use Matlab) to calculate the slip on the fault as in
Figure 6.5, using Equations 6.4 and 6.5.

18199, 47815, 6126

Fold axis trend &

plunge: 056, 25

18982, 47749, !6137

16699, 47224, 6515

MODERN STRUCTURAL PRACTICE

Fold axis trend &

plunge: 042, 12

Fault strike & dip

(RHR): 299, 38

"126

R. W. ALLMENDINGER 2015-16

FAULTING & STRESS

CHAPTER 6

2. Principal stresses are shown below in a North-East-Down coordinate system.

They have the following values and their orientations are known in the stereonet, below. You can check your answer in MohrPlotter but must do the problem in a spreadsheet or Matlab:

fault plane

(=50.0)

Stress Trend

Plunge

Magnitude

20

30

50

255.8

44.2

30

130

31

10

s = shear
stress on
plane

(=30.0)

N=pole to plane

(=10.0)

a. Calculate the orientation and magnitude of the shear stress on the plane
which is oriented 307, 50 following the steps in section 6.5 of Allmendinger
et al. (2012) or as outlined in this chapter. You do not have to do the tensor
transformation described in section 6.5.5; instead, transform the traction
vector, t, into the fault plane coordinate system. Note: stresses, above, define
a left-handed coordinate system but you want a right handed system!
b. Part (a) uses a principal stress ratio of 0.5; try the problem again using values
of 0.1 and 0.9 to see how that affects the orientation of the shear stress on
the plane.
3. A geologist measures 8 conjugate faults in a region; the strikes and dips are listed, below. All have normal displacement.
090.0
088.4
304.4
298.7

68.6
52.6
68.3
57.6

S
S
N
N

083.1
097.1
311.6
296.5

66.0
64.3
58.0
69.6

S
S
N
N

a.

Determine the orientation of 1, 2, and 3 from these data. Be as accurate

as possible

b.

What are the angle and the coefficient of friction for this data set?

MODERN STRUCTURAL PRACTICE

"127

R. W. ALLMENDINGER 2015-16

CHAPTER 6

-98.5 -98.0 -97.5 -97.0 -96.5 -96.0

37.5
37.5
FAULTING & STRESS

37.0

36.5

36.0

N em

4. The state of Oklahoma has seen a

-98
-97 Kansas
37
37.0

very substantial increase in induced

Oklahoma
seismicity since 2005 as large volumes of waste water related to conventional oil production have been
reinjected into the subsurface.
36.5The

map to the right shows known faults

in red, injection wells in black target
circles scaled by injection volume
and earthquakes in blue. The red
36
36.0
star labeled OKC is Oklahoma City.
Earthquake and in situ stress measurements suggest that, at 3 km
depth (most of the seismicity occurs
35.5
between 2 and 4 km depth)
1 is
OKC
horizontal and trends 070, 3 is also
horizontal and trends 340, and 2 is
0
20 40 60 80 100
vertical and equal to the lithostatic
load, for which you can 35.0
assume a
km
35
3
density of 2500 kg m . The faults
are high angle reverse faults; their
detailed dips are poorly know but for this problem we will assume that they dip
70. A downloadable data
file with
34.5
orientations of about 40 faults can be found
on the course web site. This problem will also require your use of MohrPlotter.
ah a

t
Faul

35.5

35.0

34.5
-98.5 -98.0 -97.5 -97.0 -96.5 -96.0

(a) The magnitudes of the principal stresses, especially 1 and 3, are not well
known. You should calculate 2 assuming that it is equal to the lithostatic
load. Assume that 1 = 100 MPa and 3 = 40 MPa.

(b) Open MohrPlotter and enter the stresses and their orientations as described
above in the stress tensor window (Calculate>Enter as Tensor). To complete
the entry, press the Calculate button. In the MohrPlotter Window, enter the
coefficient of static friction of 0.85 which corresponds with Byerlees law for
these depths. Choose File>Import Planes to read in the data file on fault
plane orientations that you downloaded from the course Web site. You will
probably want to turn off plotting of construction lines (Plot>Show Construction Lines is unchecked). In the Inspector Palette (Window>Inspector),
select the Envelopes & Points tab and make sure that the slip tendency
checkbox is checked. Slip tendency will color the fault planes both in the
stereonet view and the Mohrs Circle view according to how likely the fault
is to slip under these conditions: the warmer (i.e., more red) the color, the

MODERN STRUCTURAL PRACTICE

"128

R. W. ALLMENDINGER 2015-16

FAULTING & STRESS

CHAPTER 6

more likely slip is to occur. Once you have done all that, does any plane appear that it will slip (and produce an earthquake) under these circumstances?
(c) Now use the up-arrow next to the P(fluid) = text box to increase the pore
fluid pressure gradually. Watch both the Mohr Circle plot and the stereonet
view to see how the color of the planes changes as the pore fluid pressure
increases. When the pressure reaches hydrostatic (which you will need to
calculate), stop increasing the pore fluid pressure. Save your plot to print and
turn in. The program saves plots as .svg files that can be read by most modern vector graphics programs. SVG files can also be displayed and printed at
high resolution by opening them in any web browser.
(d) Which faults are highly susceptible to reactivation? What are their orientation(s) and how do they compare to the streaks of seismicity visible in the
map that accompanies the beginning of this question?
(e) The Nemaha fault which runs right next to Oklahoma City has some of the
largest injection wells located right along its trace. Given the length of the
fault, it would be capable of generating a large, destructive earthquake.
What is the likelihood that it will be reactivated and why?
(f) A group of large injection wells are located along faults immediately south
of the Kansas-Oklahoma border, and yet they do not appear to be associated with significant seismicity. Why is that the case and does that mean that
companies can continue to inject waste water there with impunity?

MODERN STRUCTURAL PRACTICE

"129

R. W. ALLMENDINGER 2015-16

FAULTING & STRESS

CHAPTER 6

MODERN STRUCTURAL PRACTICE

"130

R. W. ALLMENDINGER 2015-16

Chapter 7
Deformation and Infinitesimal Strain
Introduction
In this chapter, we turn our focus to deformation the quantification of
changes in shape and/or volume of a rock, plus any associated translations and rotations that is the bread and butter of most structural geologists. It will be a significant departure from the topics of the last two chapters in several fundamental
ways: stress is an instantaneous property that exists only in the instance that the

force is applied. Deformation, on the other hand, still exists in rocks hundreds of
millions, or even billions, of years after the associated stresses have dissipated. Deformation is also cumulative: one can superpose many episodes of deformation
produced by completely unrelated geologic events. Finally, because deformation
represents a change in shape or size, it is a comparison between two different states,
an initial and a final state of the material. Eventually, we will have to come to grips
with the fact that these represent two possible reference states. Whenever we consider the change in some property, we are really talking about mathematical derivatives: local slopes along a curve of a continuous function. In our case, we will be
looking at the change in position, or the change in displacement, with respect to
position.

INFINITESIMAL STRAIN

CHAPTER 7

Strain
One Dimensional Measures of Strain
A structural geologist can make three different types of measurements in order to quantify strain: changes in line length, angles, or volumes. In the discussion
that follows, well use a capital X to indicate the coordinate of a point in the initial state (the material coordinate system) and a small x for the final state (the
spatial coordinates, implicitly at time, t). Well start with the change in length of
a line (Fig. 7.1). As you can see in Figure 7.1, the initial and final lengths of the line
are:

" i = Xb Xa = X

and

f = xb xa = x

(7.1)

Thus, the stretch, S can be defined as:

"S =

f x
=

i X

(7.2a)

"s =

i X
=

f x

(7.2b)

The initial length, i, in the denominator of Equation (7.2a) means that the initial
state is the reference state. In (7.2b), f occurs in the denominator so the final state is
the frame of reference. Another way to describe this deformation is by looking at
the displacement, u, of the end points of the lines:

Xa

Xb

xa

xb

ua
ub
Figure 7.1 The change in length of a line. The initial state is shown in
blue with capital Xs and the final state in red with small xs. ua and ub
are the displacements of the end points of the line.

MODERN STRUCTURAL PRACTICE

"132

R. W. ALLMENDINGER 2015-16

INFINITESIMAL STRAIN

CHAPTER 7

ub

Xb, Yb

xb, yb

Figure 7.2 the change in angle of a line that

is originally perpendicular to the X-axis and
the displacement, u.

Xa, Ya

ua

xa, ya

" u a = xa X a

and

ub = xb Xb

(7.3)

And the change in displacement, u, is:

" u = ub ua = ( xb Xb ) ( xa Xa ) = ( xb xa ) ( Xb Xa ) = f i =

(7.4)

Thus, the extension, E or e, can be defined as the change in length over the initial
length (an initial state frame of reference) or the final length (a final state frame of
reference):

"E =

f i u
=
=

i
i X

(7.5a)

"e=

f i u
=
=

f
f x

(7.5b)

Now lets turn our attention to changes in angles and to do so we need to

bring a second dimension which we will (temporarily) call Y and y. Consider a line
initially perpendicular to the X direction that is displaced in the X direction by an
amount that varies with the distance, or length, Y (Fig. 7.2). We can write:

"

u a = xa X a

and

ub = xb Xb

u = ( xb Xb ) ( xa Xa ) = ( xb xa ) ( Xb Xa )

But, (Xb Xa) = 0, so

(7.6)

u = ( xb xa )

We define the shear strain, , and the angular shear, , as:

MODERN STRUCTURAL PRACTICE

"133

R. W. ALLMENDINGER 2015-16

INFINITESIMAL STRAIN

CHAPTER 7

" =

u
Y

and

u
= tan 1

Y

(7.7)

As we will see a bit later, as defined in Equation 7.7 is actually known as the engineering shear strain to distinguish it from a similar but distinct quantity known
as the tensor shear strain.

Three Dimensional Deformation

Now it is time to extend these concepts to three dimensions. The ratios in
Equations (7.2) and (7.5) are gradients of change in position with respect to the
new or old position and gradients of displacement with respect to position. In an
arbitrarily deformed body, these should vary in the three directions of our Cartesian coordinate systems. That is:

"

x
x
= lim i = Dij
X
X j

and

u
u
= lim i = Eij
X
X j

(7.8)

where Dij is the deformation gradient tensor and Eij is the displacement gradient tensor. The equations in (7.8) are referenced to the initial state and, as you
might expect, there are equivalent forms referenced to the final state. We need to
use partial derivatives because the displacement is a function of gradients along the
three axes of the coordinate system.
Q
ui
Q

ui

xi

Xi

Figure 7.3 The deformation of line PQ

in undeformed state (in blue) to P Q in the
final state (in red).

Q
Q

Xi
Xi

ui

P
P

Xi

MODERN STRUCTURAL PRACTICE

"134

R. W. ALLMENDINGER 2015-16

INFINITESIMAL STRAIN

CHAPTER 7

Figure 7.3 shows a more general case. The difference in displacement vectors
is:

" ui = Q ui P ui =

ui Q
u
ui
Xj i PXj =
Xj
Xj
Xj

Xj PXj =

ui
X j
Xj

(7.9)

Thus, we can write:

" ui =

ui
X j = Eij X j
X j

or

dui =

ui
dX j = Eij dX j
X j

(7.10)

We can integrate the right side of equation (7.10) to yield a general expression for
the displacement at any position:
" dui = Eij dX j

ui = ti + Eij X j

(7.11)

where u is the displacement vector at position X, and t is a constant of integration

that represents the displacement of a point at the origin of the coordinate system.
This equation holds as long as the strain is homogeneous; that is, Eij is the same
throughout the deformed body. Because Equation (7.11) represents three linear
equations, it follows that a line that is straight before deformation will also be straight after the
deformation; likewise parallel lines in the initial state will remain parallel in the final state.
Just as we did for the displacement gradient tensor, the relationships depicted
in Figure 7.3 can also be used to derive an expression for mapping points in the initial state into the final state:

" xi = Q xi p xi =

xi Q
x
x
Xj i PXj = i
X j
X j
X j

Xj PXj =

xi
X j = Dij X j
X j

(7.12)

Once again, assuming homogeneous strain, we can integrate both sides of the
equation to get:

MODERN STRUCTURAL PRACTICE

" dxi = Dij dX j

"135

xi = ci + Dij X j

(7.13)

R. W. ALLMENDINGER 2015-16

INFINITESIMAL STRAIN

CHAPTER 7

X2, x2

(X1, X2)

u1
u2
(x1, x2)

X1, x1

Figure 7.4 The displacement of points in an initial circle depend on their initial positions. Overall, the circle deforms to an ellipse. Note that the arrows represent finite displacement and not the path that the point follows.

Where x is the new position, X is the old position, and c is a constant of integration that represents the coordinates of a point initially at the origin of the coordinate system.
Recall that a second order tensor is a linear vector operator. The displacement gradient tensor relates the displacement of a point to its position, whereas the
deformation gradient tensor relates the position in the initial state to the position of
the same point in the final state (Fig. 7.4). If we know the tensor, then the displacement of the point and its new position can be calculated from Equations (7.11) and
(7.13). All of the equations that we have developed so far hold for any magnitude
of deformation. To understand the nature of these tensors, and explore some
common applications, some simplifying assumptions are in order.

Infinitesimal Strain
If we assume that the distortions are small, a number of simplifications can
be made. At the most basic level, a small or infinitesimal strain assumption permits us to consider that the initial and final states are identical, thus cutting in half
the number of tensors to worry about. Additional benefits will become apparent,

MODERN STRUCTURAL PRACTICE

"136

R. W. ALLMENDINGER 2015-16

INFINITESIMAL STRAIN

CHAPTER 7

X2, x2

M'

Q'
M

P'

X1

u2

X1 u1

Figure 7.5 The deformation of two lines

originally perpendicular to the axes of the
coordinate system.

X1, x1

below, but first lets take a closer look at the components of the displacement gradient tensor, Eij.
It will probably come as no surprise the values along the principal diagonal,
E11, E22, and E33 represent the extensions of lines that are parallel to the corresponding axes. The off-diagonal components are more interesting, which will be
illustrated by examining a special case in two dimensions (Fig. 7.5). From the geometry, you can see that:

" tan =

u2

X1 + u1

(7.14)

and by making our infinitesimal strain assumption, you can see that X1 u1.
Thus, we can write that:

" tan

u2

X1

(7.15)

For very small angles, the tangent of an angle is equal to the angle itself measured
in radians, so we can further write:

"

u2 u2
=
= E21
X1 X1

(7.16)

Thus E21 is the counterclockwise rotation of a line parallel to the X1 axis towards
the X2 axis. Likewise, E12 would be the clockwise rotation of a line parallel to the
X2 axis towards the X1 axis. The first subscript indicates the axis that the rotation is

MODERN STRUCTURAL PRACTICE

"137

R. W. ALLMENDINGER 2015-16

INFINITESIMAL STRAIN

CHAPTER 7

M'

X2, x2

!
P'

if ! is small, u1 0

Q'

Q'

M
!

X1

P
!

u2

X1

Figure 7.6 The pure rotation of two perpendicular vectors parallel to the axes of the
coordinate system.

X1, x1

towards and the second subscript indicates the axis to which the line is, initially,
parallel. Note that the E21 + E12 combined equal the change in angle, or shear strain,
of two lines initially at 90 to each other. They are tensor shear strain components and each is equal to one-half of the engineering shear strain of Equation
(7.7).
There is one more thing to learn about the displacement gradient tensor
and, by extension, the deformation gradient tensor as well. Again, well examine a
special case (Fig. 7.6). As in Figure 7.5, there are two vectors in the initial, undeformed configuration that are parallel to the coordinate axes. This time, however,
we introduce a pure rotation by a small angle, , with no deformation; that is, the
vectors P Q and P M are still perpendicular after the rotation. You can see that:

" E11 =

u1
=0
X1

and

" E21 =

u2
= tan
X1

(7.17)

Likewise, E22 = 0 but E12 = because it is also a counterclockwise rotation even

though positive E12 should be a clockwise rotation. Thus our two dimensional displacement gradient tensor for the case of pure rotation is:

0

0

" Eij =

(7.18)

Weve just learned two really important things: First, the displacement gradient
tensor is an asymmetric tensor (because ), and second, the tensor includes both strain and rotation.

MODERN STRUCTURAL PRACTICE

"138

R. W. ALLMENDINGER 2015-16

INFINITESIMAL STRAIN

CHAPTER 7

Any asymmetric matrix can be additively decomposed into a symmetric matrix and an antisymmetric matrix. An antisymmetric matrix has zeros along the
principal diagonal and the values below the principal diagonal must be the negative
of those above the principal diagonal. For the displacement gradient tensor, we can
write:

" Eij = ij + ij

(7.19)

where

" ij =

1
Eij + E ji
2

and

" ij =

1
Eij E ji
2

(7.20)

ij is the symmetric infinitesimal strain tensor and ij is the antisymmetric rotation tensor or axial vector. ij can be turned into a rotation vector, ri, as follows:

" r1 =

( 23 32 )
,
2

r2 =

( 13 + 31 )
,
2

r3 =

and

( 12 21 )

2

(7.21)

the magnitude of r gives the amount of rotation (in radians) and the unit vector, " r ,
gives the orientation of the rotation axis. Equation (7.19) basically says that deformation is equal to a strain plus a rotation, thought that is strictly true only for small
strains.
Like any symmetric tensor, the infinitesimal strain tensor has principal axes,
found by solving the eigenvalue problem, and invariants. The first invariant of the
infinitesimal strain tensor:

" I = 11 + 22 + 33 = 1 + 2 + 3

(7.22)

is the infinitesimal volume strain or the dilatation. The infinitesimal strain ellipsoid is defined by the equation:

"

x12

x22

x32

(1+ 1 )2 (1+ 2 )2 (1+ 3 )2

MODERN STRUCTURAL PRACTICE

x12 x22 x32

+ +
S12 S22 S32

"139

x12 x22 x32

+ +
1 2 3

=1

(7.23)

R. W. ALLMENDINGER 2015-16

INFINITESIMAL STRAIN

CHAPTER 7

3
,x

X3, x3

X 3

1
,x
X 1

maximum shear
strain at 45 to
principal axes

13
2

X1, x1

11

(a)

(b)

Figure 7.7 The construction of the Mohrs Circle for infinitesimal strain. (a) The infinitesimal strain ellipsoid with the old coordinate system in blue and the new, primed coordinate system in red, rotated about the intermediate principal axis by an angle . (b) the
Mohrs Circle for infinitesimal strain. Elongations are plotted on the horizontal axis and
the shear strain on the vertical axis.

where S1, S2, and S3 are the stretches along the principal axes and 1, 2, and 3 are
the quadratic elongations. If the intermediate principal stretch is 1, the deformation is two dimensional and we refer to it as plane strain.
Just like any symmetric tensor, the infinitesimal strain tensor can be represented by a Mohrs Circle construction (Fig. 7.7). We start with the axes of the infinitesimal strain ellipse parallel to the axes of the coordinate system:

1 0

" ij = 0 2
0 0

0
3

(7.24)

And transform the tensor by a rotation of about the intermediate principal axis
(Fig. 7.7a). The transformation matrix is:

cos
" aij = 0

sin

0 sin
1
0
0 cos

(7.25)

The tensor transformation equation is:

MODERN STRUCTURAL PRACTICE

"140

R. W. ALLMENDINGER 2015-16

INFINITESIMAL STRAIN

CHAPTER 7

" ij = aik a jl kl

(7.26)

So, the infinitesimal strain tensor in the new coordinate system is:

11 0

" ij = 0 2
0
31

2
2
13 ( 1 cos + 3 sin ) 0

0 =
0
2

33
(( 1 3 ) cos sin ) 0

((

1 ) cos sin )

(1 sin2 + 3 cos2 )
3

(7.27)

And the equations for Mohrs Circle for infinitesimal strain (Fig. 7.7b) are:
11 =

"

(1 + 3 ) + (1 3 ) cos 2
2

( )
13 = = 1 3 sin 2
2
2

(7.28)

One of the most important features of infinitesimal strain, which is made especially clear by the Mohrs Circle construction, is that the maximum infinitesimal shear strain is oriented at 45 to the principal axes of strain.
This has profound implications for our practical study of common structures, from
brittle to ductile shear zones (Fig. 7.8).

1
45

45

(a)

(b)

(c)

Figure 7.8 Three geological/geophysical examples demonstrating the importance of the

fact that the principal axes of infinitesimal strain are at 45 to the planes of maximum
shear strain. (a) a heterogeneous ductile shear zone in granitoid rocks, (b) sigmoidal extension fractures in a brittle shear zone, (c) P & T axes for an earthquake or fault slip analysis.
Despite their names, P and T axes are infinitesimal principal strain axes.

MODERN STRUCTURAL PRACTICE

"141

R. W. ALLMENDINGER 2015-16

INFINITESIMAL STRAIN

CHAPTER 7

Some Geological Applications

Many structural geology problems we face break the assumptions of infinitesimal strain and thus require a more complicated analysis. However, the fields
of active tectonics and brittle fault analyses are of significant importance for modern structural geology and are amenable to an infinitesimal strain approach. That
is because the deformation accrues over a short, commonly geologically instantaneous, period of time and thus is very small in magnitude.

Strain from GPS

The Global Positioning System (GPS) has revolutionized earth sciences in
the last 25 years by providing geologists and geophysicists with real time monitoring
of active deformation. Modern continuous geodetic GPS provides sub-centimeter
resolution of the displacement of monuments or stations relative to a stable reference frame. Because the changes in displacements measured are on the order of
centimeters between stations separated by tens of kilometers, the deformation measured by GPS certainly qualifies as infinitesimal. GPS data present a common circumstance: we know displacements and want to calculate the strain rather than
knowing the strain tensor and calculating displacements at different points.
We start with a simpler task: calculating the one-dimensional extension given
a transect of GPS stations. Recall that, from Equation (7.7), the one dimensional
extension is just:

"E =

Displacement, u

slope =

u
X

u
X

Figure 7.9 Plotting displacement against

position in a GPS transect in order to determine the 1D extension. Because the slope
is positive, the data represent an elongation;
a negative slope would indicate shortening
in the direction of the transect.

Position, X

MODERN STRUCTURAL PRACTICE

"142

R. W. ALLMENDINGER 2015-16

INFINITESIMAL STRAIN

CHAPTER 7

X2 (N)
X1

Figure 7.10 Map view of GPS vectors

and a coordinate transformation so
that the new coordinate system is parallel to the GPS mean displacement
vector.

X2

X1 (E)

Therefore, one can plot the displacement at a station on the Y-axis against the position of the station on the X-axis and fit a straight line (for homogeneous strain) or a
curve (for heterogeneous strain) to the data points as in Figure 7.9. The equation
for a straight line on the graph is:

"u = t +EX

(7.29)

where t is the intercept on the vertical axis. Note how similar this equation is to
Equation (7.11).
To get to the point of making a graph like that shown in Figure 7.9 you will
probably have to do several intermediate steps, which we have the background to
do. It is unlikely that the GPS displacement (or velocity) vectors, or the transect,
will be parallel to North or East. While you could eyeball the best orientation of the
transect and then calculate the distances between each station, there is a more elegant way. First, calculate the orientation of the mean GPS vector just like we did in
Chapter 2. Then rotate the coordinate system so it is parallel to the mean vector
and transform both the station positions and the displacement vectors into this new
coordinate system (Fig. 7.10). The two dimensional transformation matrix for this
operation will be:

" aij =

cos

cos ( 90 + )

MODERN STRUCTURAL PRACTICE

cos ( 90 ) cos
=
sin
cos

"143

sin

cos

(7.30)

R. W. ALLMENDINGER 2015-16

INFINITESIMAL STRAIN

CHAPTER 7

And the vector and station transformations will be, respectively:

" ui = aij u j

and

Xi = aij X j

(7.31)

Now, you can plot the component of the displacement parallel to X1, u1, against
the stations X1 coordinate (ignoring the X2 components completely), as in Figure
7.9. Thus, expanding Equation (7.31) for those components only, we get:

" u1 = a11u1 + a12u2

and

X1 = a11 X1 + a12 X2

(7.32)

To get the two (or three) dimensional strain from displacement vectors requires more involved methods. Recall that the multidimensional equivalent of
Equation (7.29) is equation (7.11) which is repeated here:

" ui = ti + Eij X j

We know the displacement vectors, u, and the positions of the stations, X. We

dont know Eij or ti; in two dimensions (ignoring the vertical component of the GPS
data), there are thus six unknowns: E11, E12, E21, E22, t1, and t2. Each station and
displacement vector pair provides two equations (one for u1 and one for u2), so in
two dimensions, we need three non-collinear stations and displacements to solve for
our six unknowns. You can visualize the three stations defining a triangle with a circle inscribed in it; when the points are displaced the triangle becomes distorted and
the inscribed circle becomes a strain ellipse (Fig. 7.11).
To calculate the unknowns, we need to gather them into a single matrix:

MODERN STRUCTURAL PRACTICE

Figure 7.11 In 2D, three stations and displacement vectors

are necessary to define the
strain produced by the displacements. The strain here is
finite (i.e., large), but the same
principle holds for infinitesimal
deformations like that measured by GPS.

"144

R. W. ALLMENDINGER 2015-16

INFINITESIMAL STRAIN

CHAPTER 7

1u
1
1u 2
2
u1
2
" u2
!
!
n
u1
n
u2

! !
! !
1 0
0

X1

0
2

X1

X2
0

0
1

X2
0

!
!
n
X1

!
!
n
X2

X1
0

X1

!
!
0
n

X1

1
X2

2
X2

n
X2

t1
t2
E11
E12
E21
E22

(7.33)

This equation is written, not for three stations but for n stations and vectors
more information than we actually need. Thats okay, though, because with data
redundancy, we can calculate the uncertainties using a least squares approach.
Solving Equation (7.33) is non-trivial because it requires calculating the inverse of
the large matrix full of zeros, ones, and Xs. The techniques required to do this
generally referred to as inverse methods are extremely powerful and form the
basis of many very important calculations in earth sciences and other fields but are
beyond the scope of this book. If you want to be a modern earth scientist, you will
probably eventually want to learn these methods.

Analyzing Brittle Faults

The upper crust of the earth is full of faults discontinuous deformation
features. If we assume that the faults are small relative to the volume of earth that
contains them, we can analyze them using infinitesimal strain assumptions. The
simplest approach is to leverage the fact that the plane of maximum shear strain is
at 45 to the principal axes of infinitesimal strain. In fault slip (and earthquake)
analysis, the principal infinitesimal shortening axis is referred to by the letter P
and the principal elongation axis with the letter T (Fig. 7.8c). Faults are particularly easy to analyze because there is no movement perpendicular to the slip vector
and thus each individual fault represents deformation in plane strain. The basic
geometry is shown in Figure 7.12.
If one is only interested in calculating the orientations of the individual P
and T axes, the calculation is quite straightforward, involving nothing more than
MODERN STRUCTURAL PRACTICE

"145

R. W. ALLMENDINGER 2015-16

INFINITESIMAL STRAIN

CHAPTER 7

en

tp

la

mo

ve

45

pole

en

tp

la

45

ne

ne
pla
t
l
fau

null or
b-axis

ne

po

le

em

fau
lt pla
ne

ov

Figure 7.12 Anatomy of a fault plane. Right side: 3D perspective view

showing the movement plane which contains the pole to the fault and the
slickenside (or other movement indicator on the fault). Left side: lower
hemisphere view of the same geometry. Arrow at the intersection of the
fault and movement planes shows the movement of the hanging wall.

ne
la
tp
ul
fa

slip!
vector, s
2

pole, n

Figure 7.13 Edge-on view of the fault

plane looking perpendicular to the movement plane. P and T can be calculated from
a simple vector addition because a single
fault is a plane strain deformation. Note
that the vector addition yields vectors which
have a magnitude of square root of two because they are 45-45-90 right triangle.

vector addition and subtraction. Figure 7.13 is a view perpendicular to the movement plane so that one sees the fault plane edge on and it appears as a line. The
unit pole vector, " n , points down into the footwall and the unit slip vector, " s , also
points down into the lower hemisphere (i.e., we are using a NED coordinate system). We need some way to specify the sense of slip, which we will do with a scalar
value, k: for faults with a normal component of slip the hanging wall moves down
and k = 1; for those with a reverse component, k = 1, because the hanging wall
moves up whereas down in positive. Thus, we can write:

MODERN STRUCTURAL PRACTICE

"146

R. W. ALLMENDINGER 2015-16

INFINITESIMAL STRAIN

CHAPTER 7

" Pi =

ni + ksi
2

and

Ti =

ni ksi

2

(7.34)

The square root of two in the denominator is necessary to ensure that P and T are
unit vectors.
The preceding analysis is fine if we are concerned about calculating P and T
axes of single faults, but what if we want to calculate the strain produced by a
group of faults? For this case, we introduce a new concept from earthquake seismology, the scalar seismic moment, Mo, and its sibling, the geometric moment, Mg:

" M o = As

and

M g = As

(7.35)

where A is the surface area that slipped, " s is the average displacement, and is the
shear modulus. The moment tensor is the sum of the scalar moment times the
dyad product of the unit slip and normal vectors:

" M ij =

n faults

M g ui n j

(7.36)

where ui is the unit slip vector and nj the unit normal (or pole) vector. This equation
is for the geometric moment tensor. If you calculated uinj for a single fault and calculated the eigenvalues and eigenvectors, it would give us the P and T axes, just like
Equation (7.34). It turns out that our old friend, the displacement gradient tensor, is
equal to (Molnar, 1983):
n faults

" Eij =

M g ui n j
V

M ij

V

(7.37)

The summation of moment tensors of all of the faults in the data set is a simplification afforded us by the infinitesimal strain assumption. When the faults are large
enough to cause finite strain, you can no longer add the tensors together. Instead,
matrix multiplication is involved as we will see in a subsequent chapter.

MODERN STRUCTURAL PRACTICE

"147

R. W. ALLMENDINGER 2015-16

INFINITESIMAL STRAIN

CHAPTER 7

Eij of Equation (7.37) is, of course, an asymmetric tensor and in infinitesimal

strain can be additively decomposed into a symmetric strain and antisymmetric rotation tensors:

" Eij = ij + ij =

M (u n
g

+ u j ni

2V

) + M (u n
g

2V

u j ni

(7.38)

The symmetric part of (7.38), before being divided by 2V is Kostrovs (1974) symmetric seismic moment tensor. When you see a moment tensor reported in an
earthquake catalog, it is Kostrovs symmetric tensor and the infinitesimal strain, ij
in that case is:

" ij =

M (u n
o

2V

+ u j ni

(7.39)

If you dont have any scaling information that is, you dont know Mo, Mg, V, or
you can still add up the moment tensors and get the principal axes of the group
of faults assuming that each fault contributes equally to the overall strain in the region. If, however, some faults (or earthquakes) are markedly larger than others, that
is a poor assumption; the moment tensor sum is always dominated by the largest
features.

MODERN STRUCTURAL PRACTICE

"148

R. W. ALLMENDINGER 2015-16

INFINITESIMAL STRAIN

CHAPTER 7

ExercisesChapter 7
All of the following exercises should be done either in a spreadsheet or Matlab. You
will need the EigenCalc program that you downloaded last week for some of the
exercises. All of the exercises have datasets that can be downloaded from the course
web page.
1. Calculate the 1D coseismic strain for the 1995 Antofagasta earthquake, using
the data from Klotz et al. (1999). The data will be provided to you in a spreadsheet. Carry out the following tasks:
(a) Determine the average or mean vector that characterizes, as best possible,
the overall orientation of the vectors.
(b) Determine the two-dimensional transformation matrix, aij, needed for a new
coordinate system where the X1 axis is parallel to the mean vector direction.
(c) Transform the East and North coordinates of the GPS stations, vectors, and
errors into the new coordinate system.
(d) Plot the u1 component of each displacement vector, and its error, against the
X1 component of the station position.
(e) Fit a straight line to approximately linear segments of the resulting curve using the relations or built in functions from the previous sections.
2. Fifteen measurements of faults and their slickensides are given in the table be
low. Calculate the P and T axes of the individual faults and then calculate an
unweighted moment tensor summation. Use EigenCalc to determine the principal axes of infinitesimal strain. In the downloadable spreadsheet, the sense of
slip (SOS) has been specified using k=+1 for normal faults and k= 1 for reverse faults as described in the text (Eqn. 7.34).
Strike
149.5
127.6
189.4
328
22.9
108.8
184.6
93.7
297.6

Fault Plane
Dip
Direction
47.2
W
60
S
34.6
W
42.5
E
50.2
E
31.1
S
39.8
W
65
S
64.1
N

MODERN STRUCTURAL PRACTICE

Slickensides
Trend
Plunge
164.4
15.4
134.6
11.9
349.6
13.1
335.3
6.6
182
23.2
169.2
27.7
317.1
31.6
269.6
8.8
300.2
5.4

"149

Sense
of slip
Normal
Normal
Thrust
Normal
Thrust
Normal
Thrust
Normal
Thrust

R. W. ALLMENDINGER 2015-16

INFINITESIMAL STRAIN

CHAPTER 7

272.5
151.6
302.7
349.4
90.9
189.7

34.5
58.1
47
33.7
71.1
36.6

284.4
154.9
105.3
145.2
96
247.9

N
W
N
E
S
W

8
5.3
17.7
15.3
14.6
32.3

Thrust
Normal
Normal
Thrust
Thrust
Thrust

3. Youll be given a spreadsheet containing 57 foreshocks of the Mw 8.2 Pisagua

earthquake (northern Chile March-April 2014) which have nodal planes and
scalar moments.
(a) Calculate the moment tensor sum for the foreshocks and use EigenCalc to
determine the orientation of the principal axes.
(b) Plot the P and T axes of the individual events in Stereonet and compare
them to the moment tensor sum that you did in part (a).

MODERN STRUCTURAL PRACTICE

"150

R. W. ALLMENDINGER 2015-16

Chapter 8
Large Strains
Introduction
Most geological deformation, whether distorted fossils or fold and thrust belt
shortening, accrues over a long period of time and can no longer be analyzed with
the assumptions of infinitesimal strain. Fortunately, these large, or finite strains
have the same starting point that infinitesimal strain does: the deformation and displacement gradient tensors. However, we must clearly distinguish between gradi-

ents in position or displacement with respect to the initial (material) or to the final
(spatial) state and several assumptions from the last Chapter small angles, addition of successive phases or steps in the deformation no longer hold. Finite
strain can get complicated very quickly with many different tensors to worry about.
Most of our emphasis here will be on the practical measurement of finite strain
rather than the details of the theory but we do have to review a few basic concepts
first, so that we can appreciate the differences between finite and infinitesimal
strain. Some of these differences have a profound impact on how we analyze deformation.

FINITE STRAIN

CHAPTER 8

Comparison to Infinitesimal Strain

A Plethora of Finite Strain Tensors
There are lots of finite strain tensors and they come in pairs: one referenced
to the initial state and the other referenced to the final state. The derivation of
these tensors is usually based on Figure 7.3 and is tedious but straightforward; we
will skip the derivation here but you can see it in Allmendinger et al. (2012) or any
good continuum mechanics text. The first tensor is the Lagrangian strain tensor:
1 u

u u

" Lij = i + j + k k = Eij + E ji + Eki Ekj

2 X j Xi Xi X j 2

(8.1)

where Eij is the displacement gradient tensor from the last Chapter. Recall that the
infinitesimal strain tensor, , is (from Eqn. 7.20):

" ij =

1
Eij + E ji
2

(8.2)

This highlights the first distinction between infinitesimal and finite strain: the former ignores the higher order term, " Eki Ekj . Expansion of the different terms in
Equation (8.1) follows the usual summation convention rules. For example, to expand for L11, we write:

" L11 =

1
1
2
2
E11 + E11 + E11E11 + E21E21 + E31E31 ) = E11 + ( E112 + E21
+ E31
(
)
2
2

In contrast,

" 11 =

(8.3)

1
( E11 + E11 ) = E11
2

Just for practice, lets also expand for L13:

" L13 =

MODERN STRUCTURAL PRACTICE

1
1
E13 + E31 ) + ( E11E13 + E21E23 + E31E33 )
(
2
2

"152

(8.4)

R. W. ALLMENDINGER 2015-16

FINITE STRAIN

CHAPTER 8

As before, the higher order non-linear term on the right hand side of the equation
is ignored in infinitesimal strain. The Eulerian finite strain tensor, " Lij , is the
same as the Lagrangian strain tensor but referenced to the final state:

1 u

u u

" Lij = i + j + k k = eij + e ji + eki ekj

2 x j xi xi x j 2

(8.5)

where eij is the displacement gradient tensor referenced to the final state.
Alternatively, we can start with the deformation gradient tensor, Dij (Eqns.
7.8 and 7.13) and derive the Green deformation tensor:

" Cij =

xk xk
= Dki Dkj
Xi X j

(8.6)

referenced to the initial or material state. The equation (8.6) expansion for C11 and
C13 (as before) is:

" C11 = D11 D11 + D21 D21 + D31 D31 = D112 + D212 + D312

(8.7a)

" C13 = D11 D13 + D21 D23 + D31 D33

(8.7b)

And, finally, the Cauchy deformation tensor, " Cij , is similar to Equation (8.6), but
referenced to the final or spatial state:

" Cij =

Xk Xk
= dki dkj
xi x j

(8.8)

where dij is the deformation gradient tensor referenced to the final state. For a vector parallel to the X1 axis, the quadratic elongation, , is equal to the square of the
stretch, S, is equal to C11.

2
" (1) = S(1)
= C11 = 1+ 2L11

(8.9)

And, a line parallel to the x1 axis in the final state:

MODERN STRUCTURAL PRACTICE

"153

R. W. ALLMENDINGER 2015-16

FINITE STRAIN

CHAPTER 8

1
1
= 2 = C11 = 1 2L11
(1) S(1)

"

These tensors are not all independent of each other. The relations between
them are:

(8.9)

(8.10)

" Lij =

1
Cij ij
2

" Lij =

1
ij Cij
2

And

Where " ij is the Kronecker delta that represents the components of the identity
matrix. All of these tensors are symmetric. Thus " Lij and Cij have the same principal
axes, Likewise, " Lij and Cij also have the same principal axes, which are different than
those for " Lij and Cij . The difference in orientation between the principal axes of " Lij
and " Lij is the amount of rotation during the deformation.

Multiple Deformations
In infinitesimal strain, the displacement gradient tensors are simply added
together to sum the total deformation as we saw in the case of the moment tensor
for earthquakes or faults in the previous chapter (Eqn. 7.37):

" total E = n E = 1 E + 2 E ++ n E

(8.11)

Lets see how the finite strain version compares. Using the deformation gradient
tensor for the first deformation, we get:

" 1 dx = 1 DdX

(8.12)

The final state of the first deformation is the initial state of the second deformation
that is, " 2 dX = 1dx so for the second deformation, we can write:

MODERN STRUCTURAL PRACTICE

"154

R. W. ALLMENDINGER 2015-16

FINITE STRAIN

CHAPTER 8

" 2 dx = 2 D 2 dX = 2 D ( 1 DdX ) = 2 D 1 DdX

(8.13)

We know that E = D I, so the expansion of the 2D1D term in Equation (8.13) is:
" 2 D 1 D = ( 2 E + I)( 1 E + I) = 1 E + 2 E + ( 2 E 1 E) + I

(8.14)

So, superposing deformations really has a higher order term, " ( 2 E 1 E ) , that is ignored in the infinitesimal strain summing of displacement gradient tensors. This
term is a matrix multiplication which is non-commutative and thus, because
" 2 E 1 E 1 E 2 E , the order in which the strain and/or rotation occurs makes a
difference in the final result. Figure 8.1 illustrates this important principle for a
stretch and a rotation.

(a)

(b)

Figure 8.1 The order in which strains and rotations occur make a difference! (a) A
horizontal stretch of 2 followed by a rotation of 45. (b) a rotation of 45 followed by
a horizontal stretch of 2. In both cases, the initial square is shown in pink.

Mohrs Circle for Finite Strain in the Deformed State

The Mohr Circle construction for the Cauchy deformation tensor (Eqn. 8.8)
is particularly useful in practical strain analysis. This construction is derived, just
like all Mohrs Circles, by rotating the coordinate system about a principal axis and
you can see the derivation in Allmendinger et al. (2012). The equations are given
by:

MODERN STRUCTURAL PRACTICE

"155

R. W. ALLMENDINGER 2015-16

FINITE STRAIN

CHAPTER 8

= /

max

2LNFE

a
S3 = 3

max

S1 = 1

2a
1

2b

b
(a)

(b)

Figure 8.2 (a) the finite strain ellipse (red) and the initial circle with radius of one (blue).
Two lines are shown in the undeformed (a and b) and deformed (a and b) state. Because
the initial length is 1, the major and minor axes of the ellipse are S1 and S3, respectively.
Line a is drawn in the orientation of the line that experiences the maximum shear strain
in the body. (b) The Mohrs Circle for finite strain in the deformed state. Lines a and b are
plotted in their correct positions. Note that the angular shear is measured by drawing a
line for the origin of the plot to the point on the circle, with the maximum angular shear
determined by the line from the origin that is tangent to the circle (line a). The orientation
of a line of no finite elongation (LNFE), which has a = 1, is also shown.

" C11 =

(C

+ C3 )
2

(C

C3 )
2

and

cos 2

" C13 =

(C

C3 )
2

sin 2

(8.15)

You will more commonly see these written as:

" =

( 1 + 3 ) + ( 1 3 ) cos 2
2

where

" =

and

" =

( 1 3 )
=
sin 2

(8.16)

1 1
=
S2

Figure 8.2 shows a finite strain ellipse on the left and the corresponding Mohrs
Circle construction on the right. This Mohrs Circle construction is very useful,
particularly where the deformation is in plane strain.

MODERN STRUCTURAL PRACTICE

"156

R. W. ALLMENDINGER 2015-16

FINITE STRAIN

CHAPTER 8

Progressive Strain

1a

LN
F

(b)
LN
IE

(a)

In Figure 8.2a, you see the finite strain ellipse (in light red) superimposed on
the initial circle (in light blue). The points of intersection between the ellipse and
circle define lines that are currently at the same length as they were when they
started out; we call these lines of no finite elongation (LNFE) and they are symmetric with respect to the principal axes of strain. You can imagine making the
same sort of diagram for the infinitesimal strain ellipse, which would be only very

FE
LN

1a

LN
F

LNIE, LNFE

LNIE

IE
LN

(c)

(d)
3: initially lengthened,
will continue to lengthen

Extension (+)

2: initially shortened but

now longer than initial

Time
1b: shorter than initial but

Extension ()

will lengthen in next step

1a: will shorten

in next increment

Figure 8.3 Illustration of progressive strain. (a) A pure shear strain path where the principal axes of finite strain and incremental infinitesimal strain remain parallel throughout
the deformation. (b) A simple shear strain path where infinitesimal and finite principal
axes are non-coaxial. (c) The meaning of fields 1-3 showing the history of shortening and
lengthening of lines at time t in those fields. (d) Cartoon illustration of what a layer of
rock might look like in each of the four fields.

MODERN STRUCTURAL PRACTICE

"157

R. W. ALLMENDINGER 2015-16

FINITE STRAIN

CHAPTER 8

slightly different than the initial circle. In that case, there would be two lines of no
infinitesimal elongation (LNIE) and they would be at 90 to each other. These
concepts allow us to understand how strain progresses with time (Fig. 8.3).
In the simplest of strain paths, there are two end-member cases: in pure
shear, the infinitesimal strain axes in each incremental step of the deformation are
parallel to each other and to the finite strain axes representing the entire deformation (Fig. 8.3a). In simple shear, the infinitesimal strain axes are not parallel to
each other in each increment, nor are they parallel to the finite strain axes. A key
aspect of simple shear is that one of the two LNFE is parallel to one of the two
LNIE and these are parallel to the zone of simple shear (Fig. 8.3b). To understand
how this works, imagine a deck of cards: You inscribe a circle onto the deck and
then you progressively shear the deck by sliding the cards over each other (Fig. 8.4).
The cards themselves do not change length; they are parallel to the LNIE and
LNFE, but the circle deforms into an ellipse and the principal axes of the ellipse
have a different orientation in each step. Simple shear is particularly important in
shear zones (Chapter 9) as well as in parallel folding (Chapter 10). The inherent
asymmetry of the simple shear case (Fig. 8.3b) is what we look for to determine the
sense of shear in fault zones.
As the strain progresses, material lines will experience different histories of
shortening and/or lengthening, depending on their initial orientations (Fig. 8.3c).
Some experience only shortening, others only lengthening, and still others are first
shortened and then lengthened. As you will see in the exercises, this can have a profound effect on different lines, or layers, that start out with different orientations in
the rock (Fig. 8.3d). Note that it is impossible to distinguish between pure and
simple shear based on the finite strain alone!
(a)

(b)

(c)

Figure 8.4 Illustration of at the classic card deck experiment to demonstrate

simple shear. The cards are viewed edge-on and thus appear as parallel lines. (a)
initial stack of cards with a circle inscribed on it. (b) and (c) progressive steps of
shearing parallel to the cards, deforming the circle into an ellipse.

MODERN STRUCTURAL PRACTICE

"158

R. W. ALLMENDINGER 2015-16

FINITE STRAIN

CHAPTER 8

Figure 8.5 More competent layers in a deformed

and metamorphosed
limestone have been
stretched, a process
known as boudinage. We
can frequently put the
boudined pieces back together to give us a linear
extension.

Strain from Geological Objects

In geology, there are a variety of different objects that can be used for the
calculation of strain. The problem, however, is two fold: first, they are generally in
random orientations and second they provide different types of information. A layer that has been pulled apart in boudinage (Fig. 8.5) may provide evidence of linear extension but nothing quantitative about shear strain. A deformed fossil, on the
other hand, may furnish information about shear strain but none about the original
dimensions of the fossil and thus none about extensions. Mathematically, determining strain from such measurements is, like our GPS example (e.g., Eqn 7.33), another case where inverse methods are needed. In the absence of computing resources and knowledge of inverse methods, a large number of clever graphical constructions and methods have cropped up. Presented here are just a few of those
methods; for a much more comprehensive treatment, see the works by Ramsay
(1967) and Ramsay and Huber (1983).

Finding Principal Axes

Sometimes, all we want to do is find the orientations of the principal axes of
finite strain. If you examine Figure 8.2b, it is clear that there are just two sets of
lines in a deformed body that have zero angular shear, . Those sets of lines are
parallel to the principal axes of finite strain. In other words, the principal axes of
strain are the only lines that have no shear strain or angular shear. FortuMODERN STRUCTURAL PRACTICE

"159

R. W. ALLMENDINGER 2015-16

FINITE STRAIN

CHAPTER 8

(a) Undeformed Trilobite

(b) rock slab with deformed Trilobites

Figure 8.6 A trilobite fossil is just one of many types of fossils with bilateral
symmetry. (a) In a tracing of the undeformed fossil, the medial line and line drawn
at the base of the head (cephalon) are perpendicular (two lines in red). (b) On a tracing of a deformed rock slab which contains several trilobites, initially in different
orientations, in most cases those lines will no longer be perpendicular because the
fossils are in an orientation where they have experienced angular shear. However, in
three of the trilobites (highlighted in blue), those two lines are still be perpendicular
and therefore must be parallel to the principal axes of strain (dashed lines).

nately, there are many different types of features in geology that are initially perpendicular, for example: fossils with bilateral symmetry (Fig. 8.6a), worm burrows
perpendicular to bedding, etc. Where such features are present in deformed rocks,
they give us a rapid way of finding the orientations of the principal axes: just look
for the features where perpendicular lines in the undeformed state are still perpendicular in the deformed state (Fig. 8.6b). Those lines must be parallel to the principal axes.

Deformed Spherical Objects

If you are fortunate, your field area will contain deformed, originally approximately spherical objects like the deformed quartz pebble conglomerate shown at
the beginning of this chapter. Other examples include deformed oolites or pellets in
carbonate rocks, reduction spots in slates, or even deformed vesicles in volcanic
rocks. Each object is now distorted into its own strain ellipsoid. In the simplest type

MODERN STRUCTURAL PRACTICE

"160

R. W. ALLMENDINGER 2015-16

FINITE STRAIN

CHAPTER 8

(a)

(b)

(c)

S1 S
2

S2

Figure 8.7 (a) A hypothetical rock sample bearing deformed spherical objects (light
gray ellipses). By cutting the sample (b) perpendicular and (c) parallel to the long axis of
the strain ellipse and measuring the aspect ratios, the full shape of the strain ellipsoid
can be determined and plotted on a Flinn diagram (Fig. 8.8).

of analysis, one would make two cuts of the sample, perpendicular to the long axis
(Fig. 8.7b) and perpendicular to the short axis (Fig. 8.7c) of the ellipsoids. In many
textbooks, the long, intermediate and short axes of the strain ellipsoid are referred
to as the X, Y, and Z axes, but we refer to them here as the X1, X2, and X3 axes
with magnitudes (i.e., principal stretches) S1, S2, and S3. For each ellipse in each cut,
we measure the long and short axis and calculate the ratio, S1/S2 or S2/S3 (i.e., X/Y
or Y/Z). If the strain is homogeneous, these ratios should ideally be the same for all
of the ellipses, but measurement error and the variability inherent in naturally deformed particles means that you will want multiple measurements to get the best
determination of the ratios.
Once you have the best fit ratio of S1/S2 and S2/S3, it is particularly convenient to visualize the result in a graphical plot known as a Flinn diagram (Fig. 8.8).
This diagram can be further quantified by calculating a parameter, k, given by:

"k =

( )
( )
S1

S2 1

S2

S3 1

(8.17)

Rocks in the prolate field with k > 1 are dominated by a linear fabric and are
commonly referred to as L-tectonites; those lying on the plane strain line (k = 1)

MODERN STRUCTURAL PRACTICE

"161

R. W. ALLMENDINGER 2015-16

FINITE STRAIN

CHAPTER 8

k=

k=1

ln (S1/S2)
prolate

Figure 8.8 The Flinn Diagram

depicting different types of strain.
In the prolate field, the strain ellipsoids are cigar shaped whereas
in the oblate field they are pancake
shaped. The original cube is shown
in blue and its deformation into
the appropriate shape for the field
of the Flinn diagram is shown in
light red.

oblate
ln (S2 /S3)

k=0

are called LS-tectonites, and those in the oblate field display a pronounced planar
fabric and a called S-tectonites.
In rocks where the spheroids have different mechanical properties than the
matrix in which they lie, or their boundaries are poorly defined, measuring the aspect ratio of the ellipses in the two cuts may give a misleading answer. In that case,
the method of choice for most structural geologists is the Fry plot (Fry, 1979). This
method works by plotting the centers of the deformed spheroids rather than measuring their aspect ratios. The centers should be most closely spaced in the S3 (or Z)
direction of the strain ellipse. You can make this plot manually by placing a piece
of tracing paper over a photograph of a section of the rock with the deformed
spheroids in it. The center of the paper is marked and placed over an ellipse near
the center of the image and all of the centers of the other ellipses are marked with
points. Subsequently, the paper is shifted to a neighboring ellipse, taking care not to
rotate the paper, and all of the centers of the other ellipses are again plotted on the
paper. The process is repeated until all of the ellipses have formed the center of the
plot. Although it is possible to construct a Fry plot by hand, there are many fine
computer programs available that will make the plot automatically once the ellipse
centers have all been marked (Fig. 8.9).

MODERN STRUCTURAL PRACTICE

"162

R. W. ALLMENDINGER 2015-16

FINITE STRAIN

CHAPTER 8

(a)

(b)

Figure 8.9 (a) tracing of deformed ooids from a photo in Ramsay and Huber (1983) and
(b) a basic Fry plot of the same data. Both graphics were produced with Frederick
Vollmers (2015) excellent program, EllipseFit.

Three Deformed Lines Graphically

Three lines in different orientations, for which you can determine both the
initial and the final length, can be used in combination with the Mohrs Circle for
finite strain (Fig. 8.2) to determine the magnitude and orientation of the strain ellipse in two dimensions (i.e., assuming plane strain). The graphical procedure is depicted in Figure 8.10. To calculate the strain proceed as follows:
1. Find three lines in the sample for which you can determine the initial
and final lengths. In Figure 8.10, they are the heavy black lines in the
upper left corner, labeled a, b, c.
2. Measure the final linear and initial sinuous (or discontinuous) lengths
of the three lines (red lines superimposed on the black) and calculate
the stretch for each line (table in upper right corner of the figure.
3. Select the shortest final line to use as a reference line (we have used c
in Figure 8.10), and use it as one side of the deformed state triangle
(lower left). Draw the other two sides of the deformed triangle with
the lines overlapping (e.g., lines a and b).

MODERN STRUCTURAL PRACTICE

"163

R. W. ALLMENDINGER 2015-16

FINITE STRAIN

CHAPTER 8

4. Our restoration to the undeformed state will not use the full lengths
of a and b; instead we will restore the triangle shaded in gray (A and
B). Measure the lengths of just those parts of a and b at constitute the
side of the deformed triangle.
5. Construct the undeformed triangle by first restoring side C to its undeformed length, keeping in parallel to c in the deformed triangle.
Using a compass, swing arcs from the two ends of C with lengths appropriate to the restored lengths of A and B. You calculate the
b

a
c

100

200

Deformed (final)
geometry

Line

lf

li

S=lf/li

344

273.5

1.26

272

354

0.77

208

229.3

0.91

Start
x1
Line a 2.8
Line b 62.4
Line c 275.8

x2
138.9
272.5
74.9

End
x1
318.0
167.6
298.8

x2
275.4
22.4
282.4

300
Undeformed
(initial)
geometry

26.5

10.5
26

c
we actually analyze
the yellow highlighted
triangle

Figure 8.10 Graphical procedure for determining the strain ellipse from
three deformed lines in different orientations (heavy black lines). Final
lengths are shown in blue and initial lengths in red. In this construction, we
have chosen C as the reference line so it has the same orientation in all three
diagrams. The perpendiculars from the apices to the opposite side in the undeformed state are reconstructed in the deformed state triangle so that we
calculate
the angular shears that each
MODERNcan
STRUCTURAL
PRACTICE
"164 line has experienced.
R. W. ALLMENDINGER 2015-16

FINITE STRAIN

CHAPTER 8

restored lengths using the inverse of the stretches in the table in the
upper right of Figure 8.10.
6. Once you have the undeformed triangle, draw a line from each apex,
perpendicular to the opposite side. For each side of the triangle, measure the distance from the nearest apex to the perpendicular. Using
the stretch for the appropriate line, calculate those lengths in the deformed state and measure the resulting distances from the appropriate
apex on the deformed triangle.
7. Construct lines from each vertex to the line segment that you calculated in step 6 on the deformed triangle (dashed red lines on the blue/
gray triangle). In the initial state triangle these lines were perpendicular to the side, but they are not in the final state triangle. The difference in angle is the angular shear that each line experienced during
the deformation.
8. You now have enough information to construct the Mohrs Circle for
finite strain in the deformed state. From your measurements, calculate
the inverse quadratic elongations, , and = /. Recall that the
shear strain, , is the tangent of the angular shears that you calculated
in step 7.
' = /

0.5

b
a

b
2' = 93.5
' = 1/
0.5

1.5

a
-0.5

Figure 8.11 Mohrs Circle (left) for the deformation shown in Figure 8.10 and
(right) the strain ellipse determined from the Mohrs Circle plotted on top of the
original deformed lines from Figure 8.10. Blue lines are shorter than they started
out and red lines are longer.

MODERN STRUCTURAL PRACTICE

"165

R. W. ALLMENDINGER 2015-16

FINITE STRAIN

CHAPTER 8

9. Plot the three points on your Mohrs Circle construction and fit a circle, centered on the axis, to the three points. You can now read the
principal inverse quadratic elongations from the Mohrs Circle and
you can calculate the orientation of the maximum principal axis of
finite strain relative to line c by measuring the 2 angle on the circle.
The final plot for Figure 8.10 is shown in Figure 8.11.

Three Deformed Lines with Inverse Methods

If this graphical construction method strikes you as particularly tedious and
prone to small errors (and you would be right, it is), it should provide excellent motivation to learn inverse methods, first mentioned in the preceding Chapter. In this
final section we outline just how to do that to demonstrate that it is not too scary!
Malvern (1969) and Allmendinger et al. (2012) give the solution for the inverse
quadratic elongation as a function of the Cauchy deformation tensor, " Cij , as:

"

dx
1
1 dxi
=
=
Cij j
2
S

ds
ds

(8.18)

where ds is the scalar length of the deformed line, f, and dxi are as shown in Figure
8.12. This equation expands to:
=

"

1 dx1
dx dx
dx dx
dx
=
C11 1 + 1 C12 2 + 1 C13 3
ds
ds ds
ds ds
ds
+

dx2
dx dx
dx dx
dx
C21 1 + 2 C22 2 + 2 C23 3
ds
ds ds
ds
ds
ds

dx3
dx dx
dx dx
dx
C31 1 + 3 C32 2 + 3 C33 3
ds
ds ds
ds
ds
ds

(8.19)

The simple two-dimensional form of the equation, equivalent to the graphical

analysis, above, is:

MODERN STRUCTURAL PRACTICE

"166

R. W. ALLMENDINGER 2015-16

FINITE STRAIN

CHAPTER 8

x2
final length, ds =

initial length =

Figure 8.12 The meaning of ds, dx1 and

dx2 in Equations 8.18-8.22.

dx2
dx1

x1

"

1 dx1
dx dx
dx
=
C11 1 + 1 C12 2
ds
ds ds
ds
dx
dx dx
dx
+ 2 C21 1 + 2 C22 2
ds
ds ds
ds

(8.20)

Because " Cij is a symmetric matrix, " C12 = C21 , we can simplify Equation (8.20) as:
2

"

1 dx1
dx dx dx dx
dx
=
C11 + 1 2 + 2 1 C12 + 2 C22

ds ds
ds
ds
ds ds
2

dx dx
dx
dx
= 1 C11 + 2 1 2 2 C12 + 2 C22
ds
ds
ds

(8.21)

There are three unknowns, so we need the quadratic elongations of three lines (a,
b, and c) to solve for the two dimensional strain ellipse:

" b =
c

dx1

f

dx1

f

dx1

f

dx1dx2
2 2

dx1dx2
2 2

dx1dx2
2 2

dx2

f

dx2

f

dx2

f

C11

C12

C22

(8.22)

If we call the 33 matrix, M, then this equation has the general form of:
MODERN STRUCTURAL PRACTICE

"167

R. W. ALLMENDINGER 2015-16

FINITE STRAIN

CHAPTER 8

" = MC

which can be solved by inverting matrix M (Fig. 8.13):

" C = M 1

(8.23)

With more than three lines, one can do a least squares best fit (Menke, 1984):
1

" C = M T M M T

(8.24)

In the three dimensional case, we would need the quadratic elongation of six lines
on at least two different planes to solve for all six independent components of the
Cauchy deformation tensor.
The x1 and x2 in Equation (8.22) are the endpoints of the lines in the deformed state. So, dx1 would be the x1 coordinate of the end of the line minus the x1
coordinate of the start of the line as depicted in Figures 8.12 and 8.13.

Figure 8.13 Spreadsheet solution to the same 3 lines problem solved graphically in Figures 8.10 and 8.11, using inverse methods (eqns. 8.22 and 8.23).

MODERN STRUCTURAL PRACTICE

"168

R. W. ALLMENDINGER 2015-16

FINITE STRAIN

CHAPTER 8

ExercisesChapter 8
For some of these problems, you will need to download the StrainSim program
from: http://www.geo.cornell.edu/geology/faculty/RWA/programs/strainsim-v-3.html
1. Using the StrainSim 3 program,
(a) Draw a circle and triangle object and then choose animate from the model.
Select simple shear and specify 100 steps of 2 each and click okay. You will
see the objects progressively deformed. After the animation is finished, select
the Info tab and copy the entire contents of the table on the right into a
spreadsheet or graphing program. Do this for each of the three triangular
bisectors, which you can select from the popup menu. In the spreadsheet or
graphics program, plot the following:
The incremental extension for all three lines on a single plot
The finite extension for all three lines on a single plot
The angle for all three lines on a single plot
(b) Now, start over and draw a circle with 4 lines at 90, 75, 45, and 30 (but
no triangle). Select Animation from the Model menu and specify pure shear
with 100 steps of 0.014 horizontal extension. Again, plot the following:
The incremental extension for all three lines on a single plot
The finite extension for all three lines on a single plot
The angle for all three lines on a single plot

Using your graphs as illustrations, discuss the similarities and differences between the progressive simple and pure shear deformation paths.
2. Use the strainSim 3 program to perform a simple shear deformation with an
angular shear of 35. Plot the result on a Mohr Circle for finite strain in the deformed state. Use the measurements from the Mohr Circle to figure out what to
enter so that you get the exact same size and shape strain ellipse using a pure
shear deformation.
3. The figure, below shows a tracing of a slab of deformed rock with a number of
brachiopods and a crinoid stem that has been pulled apart, though the individual segments have not been deformed. Assuming plane strain and no volume
change, plot the Mohr Circle for finite strain in the deformed state and deterMODERN STRUCTURAL PRACTICE

"169

R. W. ALLMENDINGER 2015-16

FINITE STRAIN

CHAPTER 8

mine the orientation of the principal axes of strain. What directions in the fossils have the same length as they had before the deformation?

"

MODERN STRUCTURAL PRACTICE

"170

R. W. ALLMENDINGER 2015-16

FINITE STRAIN

CHAPTER 8

4. The photograph above is an annotated image of a Pennsylvanian limestone

from the Blackpine Mountains in southern Idaho, part of the cover sequence
of the Raft River metamorphic core complex. The folded layers and those
stretched in boudinage have been highlighted in black to be easier to see. The
actual initial and final measurements of A, B, and C are shown in expanded
view on the next page.
(a) Determine the magnitude and orientation of the finite strain ellipse following the steps outlined earlier in this chapter, either graphically or numerically. Be sure to show your Mohrs Circle for finite strain with the three lines
plotted in correct orientation. Show the orientations of the principal axes
of the strain ellipse on the diagram on the next page.
(b) Can you tell whether the deformation was in plane strain? Explain.
(c) Can you determine the orientation of the infinitesimal strain axes? Explain.

MODERN STRUCTURAL PRACTICE

"171

R. W. ALLMENDINGER 2015-16

MODERN STRUCTURAL PRACTICE

"172

li = 13.043

(6.408, 6.864 )

10

(12.036, 8.916)

li = 12.032

15

li = 13.467

20

Blackpine sample tracings and measurements

25

x1

(24.081, 0.919)

(26.003, 10.820 )

(2.296, 1.519)

(0.170, 12.167 )

x2

CHAPTER 8
FINITE STRAIN

R. W. ALLMENDINGER 2015-16

Chapter 9
Rheology, Stress in the Crust, and Shear Zones
Introduction
Why is it that some rocks break whereas other rocks appear to flow seamlessly? Sometimes, one can observe these contrasting types of behavior in different
minerals in the same rock? Rheology is the study of flow of rocks and to delve into
this topic, and understand the question raised in the first sentence, requires us to
understand understand the relations between stress and strain (or strain rate), the

effects of environmental factors on deformation, and how materials actually deform at the scale of the crystal lattice.

Relationship Between Stress and Strain

At its simplest, deformation can be either non-permanent or permanent.
In the former case, the deformation exists only while the stress is applied and the
material returns to its undeformed state upon removal of the stress. Permanent deformation, on the other hand, is forever.

SHEAR ZONES & STRESS IN THE CRUST

CHAPTER 9

Elasticity
Elastic deformation is, by definition, non-permanent and instantaneous. The
material suffers distortion only while stress is applied and quickly returns to normal
when he stress is removed. Many processes in geophysics and geology are mostly or
completely elastic: the propagation of seismic waves, the earthquake cycle, or flexure of the lithosphere beneath a load such as a mountain belt or a sedimentary
basin. The GPS data that you analyzed in Chapter 7 is largely non-permanent deformation that occurs when the earth on one side of a fault snaps back during an
earthquake, something known as the elastic rebound theory. What distinguishes all
of these deformations is that they are very small even though the stresses are larger.
From your physics courses and experiments with springs, you probably remember Hookes Law, in which there is a linear relationship between force and
displacement. However, you also know that both stress and infinitesimal strain are
second order tensors and, therefore, the relationship between them should be a
fourth order tensor:

" ij = Cijkl kl

(9.1)

where, Cijkl is the stiffness tensor. Although Equation (9.1) looks nasty with 81 terms,
they re not all independent. In fact because of symmetry, there are at most 36 independent parameters and for all practical purposes, we only refer to a few elastic
moduli. They are:
Youngs modulus, E, for axial strain (elongations or shortenings),
where = E. A material with a high Youngs modulus is very rigid.
The shear modulus or modulus of rigidity, G, is appropriate
for simple shear deformations.
The bulk modulus, or incompressibility, K, is likewise the one
to use for simple contractions or dilations (i.e., volume strains).
These moduli can be related to one another if we know and independent parameter, known as Poissons Ratio, , which describes to what extent a shortening in one
direction is balanced out by a lengthening in an orthogonal direction. Poissons ratio
is the ratio of the transverse to the longitudinal extension:
MODERN STRUCTURAL PRACTICE

"174

R. W. ALLMENDINGER 2015-16

SHEAR ZONES & STRESS IN THE CRUST

CHAPTER 9

wi

Figure 9.1 The blue rectangle is elastically

deformed into the red rectangle (strains are
highly exaggerated). Poissons ratio is defined
as the ratio of the transverse extensions to the
longitudinal extensions, calculated from the
ws and s, respectively as shown in equation
(9.2).

wf

w f wi
w
e
i
" = t =

e
f i

i

(9.2)

For volume constant deformation, = 0.5 but for most rocks, 0.25 0.33. All
of these parameters are related by the following equation:

"G =

3K (1 2 )
E
=

2 (1+ )
2 (1+ )

(9.3)

Rocks only experience a very small amount of elastic strain before permanent deformation ensues. That permanent deformation can be in the form of a
fracture or fault, something we saw in Chapter 6. When a fracture cuts across the
material, there is a loss of cohesion and the sample falls apart. However, the material can also deform permanently without losing cohesion, a type of deformation we
call

Plasticity
Plastic deformation results when a critical threshold stress, known as the
yield stress (y), is exceeded and the ratio of the change in differential stress to the
change strain decreases drastically. Three different behaviors are possible (Fig. 9.2):

MODERN STRUCTURAL PRACTICE

"175

R. W. ALLMENDINGER 2015-16

SHEAR ZONES & STRESS IN THE CRUST

CHAPTER 9

st

h
r a in

a rd e n

in g

strain sof
tening

Figure 9.2 Idealized differential

stress()-strain() curves exhibiting different types of plastic deformation once
the yield stress, y, is surpassed. Strain
hardening occurs at lower temperatures
where as strain softening at higher temperatures. If the stress is removed (dashed
line), the initial elastic deformation is recouped but the deformation beyond the
yield stress is not.

perfect
plastic

stress
removed

(a) the slope of the stress strain curve decreases but remains positive, which is
known as strain hardening, (b) strain increases continually without any further increase in stress (perfect plastic behavior), or (c) the strain increases with decreasing
differential stress, known as strain softening. Temperature largely controls which of
these behaviors will occur.

Strain Rate and Viscosity

So far, we havent said anything about time except that elastic deformation is
instantaneous. Consider the deformation shown in Figure 9.3a. The same material
deforms continuously for a constantly applied differential stress, but the rate of deformation tensor3, ! , increases with increasing stress. So, we can make a new
curve (Fig. 9.3b) where the differential stress is plotted against strain rate; the ratio
between stress and strain rate is known as the viscosity, , which is a measure of a
fluids resistance to flow. The material shown in Figure 9.3b exhibits a constant viscosity and thus is known as a Newtonian fluid. Over a long period of time, even
rocks within the earth can exhibit fluid-like behavior, but in contrast to Newtonian
fluids, they tend to exhibit non-linear, power law viscosity as we shall see, below.

The rate of deformation tensor is commonly confused with the time derivative of the strain tensor. For infinitesimal
strain, the two are equivalent but in finite strain, the former is defined with respect to the spatial coordinates whereas
the latter is defined with respect to the material coordinates (Malvern, 1969).

MODERN STRUCTURAL PRACTICE

"176

R. W. ALLMENDINGER 2015-16

SHEAR ZONES & STRESS IN THE CRUST

CHAPTER 9

(a)

(b)
differential stress,

strain,

c
b
a

c
b
a

time, t

strain rate,

Figure 9.3 (a) A material that accrues strain over time at constant stress. a b
c and thus the strain rate varies differential stress. (b) Same data plotted with differential stress against strain rate. The slope of the line is known as the viscosity and
the simple material shown, with constant viscosity, is known as a Newtonian fluid.

Viscous and elastic idealized models are combined in various ways. For example, viscoelastic deformation is non-permanent but develops over time and is
recovered over time as well. Likewise, there are viscoplastic models that combine
elements of viscosity and plasticity. There are many additional hybrid mechanisms.

Environmental Factors
The type of deformation that a rock experiences is due primarily to its composition and the environmental conditions under which the deformation occurred.
You have undoubtedly reviewed the environmental factors in the lecture part of
your course. The most important are:
Confining Pressure This is the uniform pressure surrounding
the rock at the time of deformation. It commonly corresponds to
the vertical stress or lithostatic load, that is the weight of the
overlying rocks. An increase in confining pressure makes rocks
stronger (i.e., the yield stress increases), as reflected by the slope of
the Coulomb part of the failure envelop. Because confining pressure increases with depth, rocks should get stronger deeper in the
earth. The formula for lithostatic load is:

MODERN STRUCTURAL PRACTICE

"177

R. W. ALLMENDINGER 2015-16

SHEAR ZONES & STRESS IN THE CRUST

CHAPTER 9

" Plith = g dz gz

(9.4)

The right side of this equation is a common approximation for the

case where the density, , does not vary with depth, z, and the
change in gravitational acceleration, g, is small with changes in
depth corresponding to crustal conditions. The confining pressure
effect is relatively insensitive to rock composition, except of course
to the extent that composition determines density.
Temperature With increasing temperature, mechanisms of
crystal plasticity described below, which depend on composition,
begin to kick in and reduce the yield stress. The increase in conductive temperature with depth varies with tectonic setting, being
about 15-20C/km with depth (a heat flow of about 60 mW/m2)
in stable continental interiors, ~30C/km in rift provinces (~90
mW/m2), and >40C/km in active volcanic provinces (~120 mW/
m2). Because temperature increases with depth, its effect is opposite
to that of confining pressure. Thus these are the two great competing factors, but they are modified by other factors as well, including
Fluids There are three distinct ways in which fluids act to
weaken rocks: (a) increasing the pore fluid pressure counteracts
confining pressure by reducing the effective normal stress; (b) pressure solution dissolves soluble minerals, especially at high stress
grain-to-grain contacts, redepositing the material locally in low
stress pressure shadows or flushing the dissolved material completely out of the rock; and (c) hydrolytic weakening where water in the crystal structure weakens the bonds of the crystal.
Strain Rate When the strain is high, rocks are more resistant to
deformation and when they do deform, they are more likely to do
brittlely. A slower strain rate allows the rocks to creep or flow under lower differential stress conditions, effectively lowering the yield
stress.
MODERN STRUCTURAL PRACTICE

"178

R. W. ALLMENDINGER 2015-16

SHEAR ZONES & STRESS IN THE CRUST

CHAPTER 9

Deformation Mechanisms
Although the crust is complex, there are a relatively small number of deformation mechanisms which you have probably already reviewed in the lecture part
of your class. Here is a commented list:
Elastic deformation Very low temperature, small strains
Fracture Very low temperature, high differential stress, pore
fluid pressure important
Frictional slip on preexisting fractures Low temperature, high
differential stress but less than that required for fracture, pore fluid
pressure important
Pressure Solution Low temperature, fluids necessary
Dislocation glide Low temperature, high differential stress.
Produces strain hardening behavior
Dislocation glide and climb Higher temperature, high differential stress. Requires increased lattice diffusion to permit dislocations
to climb around obstacles.
Grain boundary diffusion Low temperature, low differential
stress, slow strain rates
Crystal lattice diffusion Very high temperature (T 0.85Tmelt),
low differential stress. Probably only effective as a primary mechanism in the mantle of the earth.

State of Stress in the Earth

The state of stress in the earth is important for a number of reasons ranging
from where earthquakes nucleate to the thickness of the elastic lithosphere that
flexes under loads. There is a fundamental principle underlying the determination
of stress in the lithosphere: the dominant deformation mechanism is the one
that requires the lowest differential stress to activate. If mechanism A requires 500 MPa at a particular set of environmental conditions, but mechanism B
only require 200 MPa, not only will B dominate, but the differential stress will nev-

MODERN STRUCTURAL PRACTICE

"179

R. W. ALLMENDINGER 2015-16

SHEAR ZONES & STRESS IN THE CRUST

CHAPTER 9

er significantly exceed 200 MPa. We tend to think of stress magnitude as something

that is imposed externally and independently of the rocks but in fact, the strength
of the rock controls the maximum differential stresses at any depth.

Frictional Slip on Preexisting Fractures

Although we examined Coulomb failure in an earlier chapter, even a casual
look at almost any outcrop will demonstrate that rocks are full of pre-existing fractures and in fact the brittle crust is much more likely to slip on this pre-existing fractures than by making new fractures. There is no such thing a pristine rock! Thus,
the place to start is the failure envelope for preexisting fractures (following Sibson,
1974, 1985):

" = n*

(9.5)

where the asterisk indicates effective normal stress. We expand this equation by
substituting in the equations for Mohrs Circle for stress:

" ( 1* 3* ) sin 2 = ( 1* + 3* ) ( 1* 3* ) cos 2

(9.6)

and after some algebra, we get an equation for the ratio of principal stresses:

" =

1* (1+ cot )
=

3* (1 tan )

(9.7)

To fine the minimum ratio of effective principal stresses that is necessary for reactivation, we set the derivative of with respect to equal to zero:

"

d
=0
d

min =

(1+ ) + )
2

(9.8)

We now have the basis for calculating the minimum differential stress at
which reactivation will occur but first, however, what about ? Recall from Chapter
6 that Byerlees Law shows that friction is relatively independent of rock type (Fig.
6.10). This result holds that at confining pressures of less than 200 MPa (a little
over 8 km depth for a density of 2500 kg/m3) = 0.85 and at greater confining

MODERN STRUCTURAL PRACTICE

"180

R. W. ALLMENDINGER 2015-16

SHEAR ZONES & STRESS IN THE CRUST

CHAPTER 9

pressures, = 0.6. The only significant exceptions to this rule are the clays illite,
montmorillonite, and vermiculite. Thus, to a first order, we can ignore composition
in the upper crust, except for its obvious control on density.
Assuming that one of the principal stresses is vertical and equal to the lithostatic load minus the pore fluid pressure (here expressed as the pore fluid pressure
ratio, ):
" v* = gz (1 )

(9.9)

we can derive expressions for the minimum differential stress that will activate slip
on pre-existing weaknesses for the three basic conditions of Andersons Law (Table
9.1)
Table 9.1: Frictional Strength of the Crust
Vertical Stress

Tectonic Environment

Minimum Differential Stress

thrust faulting

" 1

3* ( min 1) gz (1 )

normal faulting

" 1

3*

1* 3*

strike-slip faulting
"

( min 1) gz (1 )
min

( min 1) gz (1 )
( min + 1) + 1

where =

2 3
1 3

Power Law Creep

We have seen that dislocation glide and climb is the dominant mechanism of
crystal plasticity in the crust of the earth. Unlike frictional deformation, which is
relatively independent of composition and strongly dependent on confining pressure, crystal plastic deformation depends strongly on composition and temperature
but not on confining pressure. Through much experimental work, flow laws have

MODERN STRUCTURAL PRACTICE

"181

R. W. ALLMENDINGER 2015-16

SHEAR ZONES & STRESS IN THE CRUST

CHAPTER 9

been developed to describe this type of deformation for many different rock types.
The basic equation is:
Q

" e! = Co ( 1 3 ) exp
RT
n

(9.10)

Where
e! = strain rate [s1]

Co = a constant [GPans11; experimentally determined]

1 - 3 = the differential stress
[GPa]

Table 9.2: Power Law Creep Parameters

n = a power [experimentally
determined]

Rock Type

Log(Co)
GPans1

Q = the activation energy

[kJ/mol; experimentally
determined]

Albite Rock

6.1

3.9

234

Anorthosite

6.1

3.2

238

Aplite

2.8

3.1

163

2.6

335

R = the universal gas constant

= 8.3144 10-3 kJ/mol
K

"

14.4

6.4

444

6.5

3.4

260

11.72

356

14.4

9.1

349

Clinopyroxenite
Diabase
"
Dolomite

T = temperature, K [K = C
+ 273.16]
It is called power law creep
because the strain rate is proportional
to a power of the differential stress.
Equation (9.10) describes a non-linear
viscous rheology. Because temperature
occurs in the exponential function, this
sort of rheology is going to be extremely sensitive to temperature. To
think of it another way, over a very

Olivine (dry)

14.4-15.3

3.4-3.5

Olivine (wet)

14.2,16.1

3.4,4.5

528-544
444,498

-0.2

2.9

106

Granite (dry) 2

1.6

3.4

139

1.9

137

9.9

2.1

210

Marble

25.8

7.6

418

"

20.6

4.2

427

Quartz Diorite

4.3

2.4

219

Quartzite (dry)

1.2

1.9

149

Quartzite (dry)

2.9

184

Quartzite (wet)

1.8

134

Granite (wet)
Limestone

Quartzite (wet)

"182

Q
(kJ/mol)

Granite (dry) 1

Salt (halite)

MODERN STRUCTURAL PRACTICE

3.7 to 2.6
16.7

to 167
5.3

102

R. W. ALLMENDINGER 2015-16

SHEAR ZONES & STRESS IN THE CRUST

CHAPTER 9

small range of temperatures, rocks change from being very strong to very weak.
The exact temperature at which this occurs depends on the lithology. Also unlike
the equations for frictional strength, there is no depth term in Equation (9.10). Instead, we must use temperature as a proxy for depth by assuming a geothermal
gradient. Some of the experimentally determined parameters for different common rock types are shown in Table 9.2. There are many more that are available in
the literature.
Generally, one will want to calculate the differential stress as a function of
temperature. That requires rewriting Equation (9.10) as:
1

e!
"1 3 =

Q

Co exp

RT

(9.11)

Stress Variation in the Lithosphere

To plot the variation of stress in the crust one simply plots the appropriate
equation in Table 9.1 and Equation (9.11) against depth (Fig. 9.4). The temperature
in (9.11) will have to be converted to depth by assuming a geothermal gradient and
here we have also assumed a continental strain rate of 1015s1. You can see from
Figure 9.5 that the transition from frictional to crystal plastic occurs where the two
lines cross and depends on tectonic regime, pore fluid pressure, geothermal gradient, strain rate and composition.
Care should be taken not to interpret these curves too literally! In Figure
9.4, for example, we assume that the crust has a quartz diorite composition when,
in fact, crustal composition is highly variable. Likewise the mantle here is assumed
to be dry olivine but in some tectonic environments, wet olivine would be more appropriate. Furthermore, we have not even taken into account mechanisms like pressure solution which are known to be quite important in the crust.
What this analysis does demonstrate is that the middle crust can support the
highest stresses, on the order of several hundred megapascals, whereas the lower
crust is very weak, supporting < 20 Mpa of differential stress. In some areas where

MODERN STRUCTURAL PRACTICE

"183

R. W. ALLMENDINGER 2015-16

SHEAR ZONES & STRESS IN THE CRUST

CHAPTER 9

Figure 9.4 Spreadsheet illustrating the calculation of differential stress magnitude

in the crust of the earth assuming that strength is controlled by frictional slip on preexisting fractures and power law creep. Note that we have assumed a single average
coefficient of friction rather than the two from Byerlees Law. The power law creep parameters are for quartz diorite from Table 9.2. Note that Equation (9.11) gives differential stress in GPa whereas the equations in Table 9.1 yield stresses in Pa. Conversion
factors have been added to the spreadsheet to cover all values to MPa.

the crust is < ~40 km thick and/or heat flow is not very high, there will be a jump
in strength of the uppermost mantle. In the case of oceanic lithosphere or continental rift provinces with low heat flow, the upper mantle may be strong enough to
support upper mantle earthquakes. This notion of a strong middle crust and strong
(-ish) upper mantle has given rise to the concept of the lithosphere as a so-called
jelly sandwich: two strong layers separated by a weak lower crust.

Shear Zones
As faults cut down through the earth, they change from simple fracture
planes, to zones of anastomosing fault planes to ductile shear zones that may be
tens to thousands of meters in thickness. One of the great accomplishments of the
last thirty years in structural geology is the understanding of genesis of shear zones,
especially the minor structures which are key to understanding the sense of shear.

MODERN STRUCTURAL PRACTICE

"184

R. W. ALLMENDINGER 2015-16

SHEAR ZONES & STRESS IN THE CRUST

CHAPTER 9

90/2
P

/2

R
R

Figure 9.5 Crosssection showing

Riedel shears in a
right lateral shear
zone. The gray
dashed lines are Pshears that form
after Riedel shears
at the time of
through-going rupture.

45

Brittle Shear Zones

Brittle shear zones that form in the upper ten kilometers or so of the Earths
crust develop a suite of structures that reflect their evolution from pre- to post-rupture. The pre-rupture structures remain preserved in the wall rocks long after the
fault surface is fully developed. The most important of these structures are the
Riedel shears. These are conjugate Coulomb fractures that form in an en echelon
manner within an overall shear zone that is, ideally, oriented at ~45 to the principal stresses (Fig. 9.5). Because they are Coulomb fractures, the two planes have opposite senses of shear: the R, or synthetic Riedel shears, form at an angle of /2
with respect to the shear zone and have the same sense of shear as the main shear
zone, itself (clockwise in Fig. 9.5) whereas the R or antithetic Riedel shears form
at 90/2 and have a sense of shear that is opposite to that of the main shear zone
(counterclockwise in Fig. 9.5). The angle is the angle of internal friction (Chapter
6). P-shears may form at the time of rupture and have the same sense of shear as
the main shear zone.
A number of publications emphasize the use of Riedel shears and P-shears
to determine the sense of shear in fault zones (e.g., Petit, 1987). Some imply that
any fracture at a low angle to the shear zone is an R-shear and any at a high-angle
is an R shear. These are dangerous assumptions because rocks are full of fractures
that existed prior to, and have nothing to do with, the formation of the fault. To
improve your chances of a correct interpretation, one should always determine that
(a) there has been shear on the minor fractures (e.g., by observing slickensides, etc.)
MODERN STRUCTURAL PRACTICE

"185

R. W. ALLMENDINGER 2015-16

SHEAR ZONES & STRESS IN THE CRUST

CHAPTER 9

Figure 9.6 Synthetic

Riedel shears (red arrows) showing correct
sense of offset associated with a larger fault
zone (yellow arrows).
Sense of shear (white
arrows) is opposite
that shown in Figure
9.5.

and (b) the shear on the minor fractures is consistent with that associated with
Riedel shears (Fig. 9.6).
Where fluids flow and precipitate minerals commonly quartz or calcite
in crevices and fractures associated with faulting, the resulting geometries can be
very indicative of the sense of shear. Two types of structures are very useful. The
first are sigmoidal veins, which are sometimes referred to in older literature as
tension gashes (Fig. 9.7). The asymmetry of the veins and their shape reflects the
rotation of the older, central parts of the veins that occurs during large magnitude
finite strain. The tips of the veins propagate parallel to the shortest principal axis of

Figure 9.7 Sigmoidal veins in

cross-section in a right lateral
shear zone. Two generations
are shown: an older inactive set
in light gray and a younger
smaller set at 45 to the shear
zone boundary in white.

MODERN STRUCTURAL PRACTICE

"186

R. W. ALLMENDINGER 2015-16

SHEAR ZONES & STRESS IN THE CRUST

CHAPTER 9

infinitesimal strain which, we know from Chapter 7 is at 45 to the shear zone

boundary. Eventually, the central part of the vein may rotate so much that it is no
longer favorably oriented for opening. When that happens, new veins at 45 may
form, crosscutting the older generation of veins (Fig. 9.7). Sigmoidal veins are highly reliable indicators of sense of shear.
The second case of minerals precipitating from fluids in shear zones is the
situation where an undulatory fault surface opens up a void during fault motion
(Fig. 9.8). These voids are ideal places in which to precipitate minerals such as calcite or quartz during fault movement. Under these conditions, the mineral fibers,
sometimes called slickenfibers, are oriented parallel to the direction of motion.
On an exposed surface (Fig. 9.8b), one can usually identify the step from which the
fibers grew, usually by the sharp contact between the step and the fibers which
marks the upstream (with respect to the movement of the upper plate) side of the
void. Like sigmoidal veins, slickenfibered steps are highly reliable indicators of fault
motion (Fig. 9.9).
(a)

movement opens a voi

d
along irregular fault pla
ne

mineral fibers grow from

the step into the void

(b)

slickensides on
the fault surface

Figure 9.8 The formation

of mineral fiber steps on an
undulatory fault plane. (a)
The void produced by the
fault slip is an ideal place in
which to precipitate fibrous
minerals such as calcite or
quartz. (b) The fibers grow
in the direction of fault slip
as the void grows in size.
The key to correct interpretation of fault slip from the
exposed geometry in (b) is
to identify the sharp contact between the top of the
step in the footwall and the
fibers. A right-lateral or top
to the right shear sense is
indicated.

sharp contact indica

tes
beginning of step

MODERN STRUCTURAL PRACTICE

"187

R. W. ALLMENDINGER 2015-16

SHEAR ZONES & STRESS IN THE CRUST

CHAPTER 9

Figure 9.9 Calcite

slickenfibers on a
fault surface in limestone. Large red ar rows indicate the
sense of shear in this
oblique view with the
upper (missing) block
having moved down
towards the lower left
corner of the photo.
Small yellow arrows
indicate the sharp
contact between the
fibers and the fault
surface which indicates the upstream
side for the void.

Finally, although we tend to think of foliations forming in metamorphic

rocks, brittle fault zones can likewise have local foliation developed in the high
strain zone that parallels the fault surface. Two morphologies of fault-related foliations are common (Fig. 9.10): (a) where there is a discrete zone of gouge claylike material produced by mechanical grind along the rock surface the largest
movements and thus highest shear strains occur between the gouge and the much
less deformed wall rock. The foliation in the gouge reflects this heterogeneous simple shear strain and produces curving foliation planes (Fig. 9.10a). (b) Where the
fault plane is in carbonate rocks, a localized pressure solution cleavage can form; in
this case the highest shear strains are at next to the discrete fault surface and dies
out upward into the country rock on either side (Fig. 9.10b). Both of these cases are
examples of a type of brittle S-C fabric, a term we will define in the next section.

Ductile Shear Zones

To appreciate how far structural geology has come consider that, in the middle of the last century, it was common place for structural geologists to interpret
that every planar foliation in a rock represented a separate deformation, labeling
them S0 to Sn, where S0 was the original stratification. We now understand that
progressive deformation in shear zones can develop two more more planar folia-

MODERN STRUCTURAL PRACTICE

"188

R. W. ALLMENDINGER 2015-16

SHEAR ZONES & STRESS IN THE CRUST

CHAPTER 9

(a)

Figure 9.10 Crosssections of fault related foliations in (a)

fault gouge in siliciclastic rocks, and (b)
zones of enhanced
pressure solution
next to faults in carbonates.

(b)

Figure 9.11 Idealized diagram

showing the arrangement of S
and C planes in an SC mylonite.
Red ovals represent the strain
schematically in different parts
of the structure.

tions. The basic shear zone foliations (Fig. 9.11) are known as S-C fabrics; with the
C standing for French word for shear, cisaillement, and the S the French
word for schistosit or foliation in English. As you can tell, French geologists were
amongst the first to correctly describe S-C fabrics (Berth et al., 1979)! The geometry of S and C planes develops by extremely heterogeneous simple shear. From
your experiments with simple shear in the last chapter, you know that, the larger
the shear strain, the closer and closer the long axis of the finite strain ellipse rotates
towards the plane of shear, itself (though technically never being exactly parallel,
because there is no shear strain on planes parallel to a principal strain axis). The C,
or shear, planes are region of extremely high shear strain and thus the S or foliation
planes become asymptotic to the C planes as they approach them (Fig. 9.11).
Basement rocks in which one tends to find mylonites commonly lack easily identifi-

MODERN STRUCTURAL PRACTICE

"189

R. W. ALLMENDINGER 2015-16

SHEAR ZONES & STRESS IN THE CRUST

CHAPTER 9

-type

-type

Figure 9.12 Asymmetric porphyroclasts in a mylonite from NW Argentina (left) and

schematically on the right. Representative clasts are labeled on the photo. The topto-the-right sense of shear is appropriate for both photo and sketch.

able offset features, so the reliable sense of shear indicated by S-C fabrics is quite
important.
Before moving on to other sense of shear indicators, a word of caution:
shear zone shapes commonly exhibit curving, sigmoidal geometries (re. Figs. 9.7,
9.10 and 9.11). However, if you compare Figures 9.7 and 9.11, the sense of curvature is opposite to each other even though both diagrams have the same sense of
shear. To interpret these fabrics and shapes correctly, it is essential that you understand the kinematics (i.e., the strain and how it was produced) rather than simply
trying to do pattern recognition. The curved feature in Figure 9.7 is a vein and
therefore the long principal axis of the strain ellipse is perpendicular to it. The
curving features in Figures 9.10 and 9.11 are foliations and therefore the long axis
should be approximately parallel to them.
There are a variety of other shear sense indicators in mylonitic rocks based
on the geometry of individual grains and grain aggregates. These include - and shaped asymmetric porphyroclasts (Fig. 9.12) and fractured an rotated mineral
grains (commonly feldspars as in Fig. 9.13). Although beyond the scope of this
manual, crystallographic c-axis fabrics and other microscopic preferred orientaMODERN STRUCTURAL PRACTICE

"190

R. W. ALLMENDINGER 2015-16

SHEAR ZONES & STRESS IN THE CRUST

CHAPTER 9

Figure 9.13 Fractured and rotated plagioclase crystal from a mylonite in the
Sierra Chango Real of NW Argentina. Note the domino-style deformation of the
feldspar: the overall rotation gives the sense of shear, even though the micro faults
have the reverse sense of displacement

tions are likewise commonly used to determine sense of shear in plastically deformed mineral grains.

Displacement in Heterogenous Shear Zones from Foliations

So far, we have looked mostly at features that give us the sense, but not the
magnitude, of displacement in a shear zone. Displacement is commonly not easy to
determine due to lack of identifiable piercing points in basement rocks. If we make
the assumption that only heterogeneous simple shear is responsible for displacement, there is quite a nice method published by Ramsay and Graham (1970). Consider a homogeneous simple shear zone (Fig. 9.14). In the field, we cant measure
directly, but we can measure , which is just the angle between the foliation (assumed to be kinematically similar to S-planes) and the shear zone boundary. If the
foliation is parallel to the XY plane of the strain ellipsoid then there is a simple relationship between and :

MODERN STRUCTURAL PRACTICE

" tan 2 =

"191

(9.12)

R. W. ALLMENDINGER 2015-16

SHEAR ZONES & STRESS IN THE CRUST

CHAPTER 9

area = d

Figure 9.14 Hypothetical

homogeneous simple shear
zone demonstrating how the
area under the shear strain-y
curve is equal to the displacement, d.

shear strain,

area = d Figure 9.15 Variation in foli-

foliation

ation across a heterogeneous

simple shear zone (left) and the
calculation of displacement by
determining the are under the
curve of shear strain versus
distance across the shear zone.

shear strain,

Although it is trivial in the case of a homogeneous shear zone, we could compute

the displacement graphically by plotting as a function of the distance across the
shear zone y and calculating the area under the curve (Fig. 9.14):

" d = dy =

2
dy
tan 2

(9.13)

For a heterogeneous shear zone the usual case in geology the situation is
more complex, but you can still come up with a graphical solution as above. The
basic approach is to (1) measure the angle between the foliation and the shear zone
boundary, , at a number of places, (2) convert those measurements to the shear
strain, , (3) plot as a function of perpendicular distance across the shear zone,
and (4) calculate the displacement from the area under the resulting curve (Fig.
9.15).

MODERN STRUCTURAL PRACTICE

"192

R. W. ALLMENDINGER 2015-16

SHEAR ZONES & STRESS IN THE CRUST

CHAPTER 9

General (2D) Shear

Alas, most natural shear zones probably do not fit the simplified assumptions
of Ramsay and Graham (1970). Instead, they probably have a combination of pure
and simple shear (sometimes known as sub-simple shear), and may also have volume changes, both of which would negate the elegant analysis of Ramsay and
Graham (1970). We will only briefly look at the theory here, following the work of
Tikoff and Fossen (1993) and Fossen and Tikoff (1993). Because we have studied
tensors and strain already, this part should be (relatively) easy to follow. We start
with the deformation gradient tensor, which relates the positions in the initial state
to those in the final state:

" xi = Dij X j

(9.14)

Neglecting the constant of integration, this is the same equation as (7.14). In a coordinate system parallel to the stretches and the simple shear is parallel to the X1
axis, the tensor D that captures simultaneous pure and simple shear can be written
in two dimensions as:

S1 S1
=
" Dij =
0 S3
0

( S1 S3 )

2 ln ( S1 )

S3

(9.15)

is known as the effective shear strain. This equation is appropriate for a constant
volume deformation where S1 = 1/S3. The deformation matrix for simultaneous
simple shear, pure shear, and volume change is:

S1
" Dij =

( S1 S3 )
S
ln 1 S
3

( )
S3

(9.16)

Dij, of course, is an asymmetric tensor. If you multiply D by its transpose, you get
the symmetric Green deformation tensor, C (referenced to the initial state):

MODERN STRUCTURAL PRACTICE

"193

R. W. ALLMENDINGER 2015-16

SHEAR ZONES & STRESS IN THE CRUST

CHAPTER 9

" DDT = Dki Dkj = Cij

(9.17)

From C you can determine the principal axes of strain, etc.

There are an infinite number of potential paths that lead to any finite strain
state. The deformation in Figure 9.16, for example, could be produced by (a) pure
shear with stretch, S1 = 1.75, followed by a simple shear with = 4.0 (76); (b) a
simple shear with = 1.31 (52.6), followed by a pure shear with S1 = 1.75; or (c) a
general shear with S1 = 1.75 and effective = 2.29 (66.4). One would need to extract
information on the path perhaps from curved mineral fibers to distinguish
these three possibilities.

Figure 9.16 Any finite strain may arise from an infinite number of possible strain paths. The one shown here was produced by a general shear with
horizontal stretch of 1.75 and an effective shear strain for 2.29, but many
other paths will produce exactly the same final state from the given initial
state.

MODERN STRUCTURAL PRACTICE

"194

R. W. ALLMENDINGER 2015-16

SHEAR ZONES & STRESS IN THE CRUST

CHAPTER 9

ExercisesChapter 9
1. Calculate the variation of differential stress with depth in the crust for Granite
(dry) 2 (Table 9.2) assuming a strain rate of 1015s-1, geothermal gradient of
15C/km, an average density of 2750 kg*m3, in an extensional environment
with hydrostatic pore fluid pressure, and a coefficient of static friction of 0.70.
You can to the calculations either in a spreadsheet or in Matlab.
2. The diagram, below, shows the case of simple shear of a square where the angular shear, , is defined as the change in angle of two originally perpendicular
lines (i.e., the sides of the box). A dashed line in the initial state makes an angle
of with respect to the plane of shear. After the deformation, the same dashed
line makes an angle of . Derive an equation which shows as a function of
and .

"

3. Once Riedel shears form, they become material lines (really, cross-sections of
planes) that rotate in the shear zone. For the questions that follow, assume int =
0.625, s = 0.75, 1 = 120 MPa, 3 = 14.7 MPa, and the cohesion, So=20 MPa.
Also assume simple shear, not general shear.
(a) Calculate how much synthetic and antithetic Riedel shears will rotate before
they become inactive. Show your calculations and include any plots that you
used to come up with your answer.
(b) What is the effect of changing the magnitude of the differential stress?

MODERN STRUCTURAL PRACTICE

"195

R. W. ALLMENDINGER 2015-16

SHEAR ZONES & STRESS IN THE CRUST

CHAPTER 9

(c) Use your result from the previous problem to calculate the angular shear
necessary to make the synthetic Riedel shears inactive.
(d) Given their different orientations within the shear zone, which shear, R or R
do you anticipate would rotate faster and why?
4. The following questions relate to the sketch, below, of a ductile shear zone in in
granitoid rocks. A grid has been laid over the sketch to help you make your
measurements.
(a) Use the Ramsay and Graham (1970) relationship, described above, between
orientation of foliation in a shear zone, , and shear strain, , to determine
the displacement across the shear zone.
(b) What is the sense of shear in the shear zone?

"

(c) What assumptions does this calculation involve?

4
Figure 3.4'

6m

Fabric traie

Figure 3'3'
ctories of

See Questio

ns 3'3 an

d 3'4'

5. The photographs on the next pages are rock samples showing fault related deformation. For each one, interpret the sense of shear (relative to its orientation
on the page) and make a quick sketch depicting the key features that helped you
#il11fi:TJJ
:"
"
J
come to the conclusion that you did.
f;
:f
::
i:
H
::
T
*:"

T,lili3l,l1
"
f
:"
if
T
iJ
:'
i"
"
::
"
J
lx
T
"
1
,i
fi
ij
if
:l
.:fJ::::"p"":y J;"#

MODERN STRUCTURAL PRACTICE

"196

$L

R. W. ALLMENDINGER 2015-16

SHEAR ZONES & STRESS IN THE CRUST

CHAPTER 9

MODERN STRUCTURAL PRACTICE

"197

R. W. ALLMENDINGER 2015-16

SHEAR ZONES & STRESS IN THE CRUST

CHAPTER 9

MODERN STRUCTURAL PRACTICE

"198

R. W. ALLMENDINGER 2015-16

Chapter 10
Fault-Related Folds and Cross Sections
Introduction
In this chapter, we change scale considerably from the previous few chapters,
zooming out and looking at systems of structures. In the process, you will gain experience carrying out one of the most fundamental activities with which a structural geologist is commonly tasked: drawing cross sections. Cross sections are the primary way that we project the geometries of rock bodies into the subsurface and are

motivated by a variety of reasons both practical and intellectual. This task occurs at
the intersection between data and interpretation. The difference between a good
and a bad cross section can mean millions of dollars wasted or embarrassing gaffes
in publications and public presentations. Cross sections can be informal freehand
affairs when one is trying to understand different alternatives, but the final version
can be significantly improved by application of rules and procedures based on
our understanding of fold geometrics and their relations to faults. Regardless of the
rigor with which one constructs a cross section, however, one should never forget
that the cross section is only an under constrained model and not data. Uncertainty
is where we will start.

CROSS SECTIONS

CHAPTER 10

= dip

90

Figure 10.1 The calculation of

the depth to a layer at a horizontal distance x from a surface
outcrop with measured dip .
Quite on purpose, we do not assume that the borehole will be
perfectly vertical.

90 +
r
d

Uncertainty in a Simple Depth Projection

Uncertainty is a topic that we visited briefly in the early chapters of this book
but have not revisited since. You may recall that in Chapters 2, we saw the formula
for the mean vector calculation (Eqn. 2.12). But now, lets say we know the mean
vector of a series of poles to bedding and want to use that data in a calculation of
the depth to that layer some distance away where the layer is in the subsurface (Fig.
10.1). This is the simplest sort of subsurface projection of geometry that one can
imagine. Now, a question of great import for anyone who has ever paid for a drill
hole to a particular layer: what is the uncertainty, or error, in our knowledge of the
depth to the layer?
The first part of the problem, of course, is to calculate the depth, d, as a
function of the horizontal distance, x, the dip, , and the angle that the borehole is
drilled at, . Because the borehole is not necessarily vertical, we have to solve the
two right triangles with sides rx and rd, above. After some trigonometry, we get the
expression for d:

"d =

x sin( )

cos +

(10.1)

Note that we are using /2 which is 90 in radians, because most spreadsheets and
programing languages require the use of radians rather than degrees. You can see
that, for the special case of a vertical borehole, equation (10.1) would simplify to
the much simpler:
MODERN STRUCTURAL PRACTICE

"200

R. W. ALLMENDINGER 2015-16

CROSS SECTIONS

CHAPTER 10

" d = x tan ( )

(10.2)

But we are going to use Equation (10.1) because that is more general and because it
will allow us to incorporate uncertainties in the angle that the borehole is drilled.
Lets assume that x is perfectly known but there are uncertainties in the dip,
, of the layer were are interested in (the gray layer above) and in the angle of the
borehole, . The standard error propagation equation for Gaussian errors (random
and uncorrelated) for this case looks like (Taylor, 1997; Bevington and Robinson,
2003):
2

d
d
" d = +

2

(10.3)

Where d is the uncertainty in depth in the borehole, the uncertainty in dip, and
the uncertainty in the angle of the borehole. So, we have to differentiate equation (10.1) with respect to and . You probably dont remember how to do this off
the top of your head but there is an online tool that can help: WolframAlpha
(http://www.wolframalpha.com). To differentiate with respect to , enter this line
into the search field: d(x*sin(a)/cos(g-(pi/2)+a))/da. And for , enter: d(x*sin(a)/
cos(g-(pi/2)+a))/dg. Notice we are using the letters a and g in place of and .
This gives us the following equations:

"

d
= x sin( )csc 2 ( + )

and "

d
= x sin ( ) ( cot ( + )) csc ( + )

(10.4)

We can now simplify these and substitute into Equation (10.3) to get the propagated error in depth to the layer in the borehole:
x sin ( )

x sin ( )
" d = 2
+

sin ( + ) tan ( + ) sin ( + )
2

(10.5)

Lets put this to work in a simple test example. Our input is:

MODERN STRUCTURAL PRACTICE

"201

R. W. ALLMENDINGER 2015-16

CROSS SECTIONS

CHAPTER 10

Parameter
Horizontal distance , x
Dip,
Borehole angle,

Magnitude

Uncertainty

3000 m

35

90 (i.e., vertical)

Using equations (10.1) and (10.5) above, we calculate that uncertainty in the depth
at which the top of the layer is pierced in the borehole is 2101 411 m! The 5 of
uncertainty in the borehole and dip angles yields an uncertainty equivalent to 20%
of the depth of the borehole.
But these methods are even more powerful than that because we can now
ask the question, which factor has a bigger effect on the uncertainty, the dip or the
borehole angle. In other words, you now have the tools in hand to answer the question: Should I invest more time/money in reducing the error on the dip or controlling the angle that the borehole is drilled? You may be tempted to answer: 5
is five degrees but in fact the trig functions are non-linear, so lets try it. If we set
the borehole uncertainty to zero, we can then calculate that the 5 uncertainty in
dip yields a depth uncertainty of 390 m, close to the total error. In contrast, perfectly known dip and 5 error in the drill hole orientation will give us 128 m of uncertainty in the depth. So now we know where to devote our time and resources:
determining the dip better.
This will not always be the case: for some geometries, the angle of the borehole could be more important. If one has programed in these relationships in a
spreadsheet or more complicated code (say, by propagating the errors in Equations
6.4 to 6.6), it becomes trivially easy to play what if games. With this lesson
that even the simplest type of subsurface projection contains considerable error
firmly in mind, let us return to the initial objective, that of drawing cross sections.

Drawing Cross Sections

Cross sections are the projection of surface geology, supplemented by commonly sparse subsurface data, onto a vertical plane. Although similar to, they differ

MODERN STRUCTURAL PRACTICE

"202

R. W. ALLMENDINGER 2015-16

CROSS SECTIONS

CHAPTER 10

from the down-plunge projections that we studied earlier, in that the profile plane
of a fold is only vertical if the fold axis is horizontal. The data that one has available varies, but commonly includes:
Stratigraphic contacts
Intrusive contacts and metamorphic aureoles
Structural information such as faults and fold axes
Orientation data such as strikes and dips of bedding, cleavage, etc
Borehole and/or seismic reflection data (Chapter 11)

Preparation for drawing a cross section

Prior to starting any task, it is always worth examining your motives; otherwise your time can be wasted. Types of questions to ask yourself before you start to
draw your cross section include:

What is my purpose in drawing a geologic section?

Is this for a scientific study of the tectonics of a region or a detailed process-oriented study of a single feature

Is the section to be a balanced section?

Is it for some applied use such as oil and gas or minerals exploration?

Do I need to show the distribution of surficial features or interpret

the upper 10 km of the crust? Etc.

What types of rocks and structures will the section depict (because
there are rules for different types)?

Unmetamorphosed sedimentary rocks with excellent lateral continuity?

Metamorphic and igneous rocks with irregular contacts?

An area with lots of pre-existing structure with abrupt lateral

changes?

Where is the best place to draw the section?

MODERN STRUCTURAL PRACTICE

"203

R. W. ALLMENDINGER 2015-16

CROSS SECTIONS

CHAPTER 10

Usually, we draw sections so that they are oriented perpendicular

to the structures and the strike of bedding.

In general, we try to avoid drawing sections through very complicated areas and also avoid drawing them parallel to structure, although there are good reasons for doing both.

We usually want to locate sections where we have a lot of orientation data and in areas where the outcrop allows us to place contacts with some certainty

Step-by-Step Section Construction

1. Select the line of section perpendicular to structure and where there
are plentiful high quality data, avoiding local complications.
2. Construct a topographic profile 1:1 (i.e., without vertical exaggeration) along the line of section (Fig. 10.2). Profile should be the same
scale as the map. Ink in the topographic profile or, if you are constructing the section in a graphics program, lock the profile or put it
in a separate layer along with the elevation scale.
3. Plot apparent dip tick marks on the topographic profile (remember
that if the strike is not perpendicular to the line of section, then you
have to calculate the apparent dip!). For dips that lie close to, but not
Tc
Kd

Kd
Kd

Kt

Te

Tc

Kt

Te
Jp

Tc

Tc
Kd

Kd
Kt

Jr
Jm

Figure 10.2 Plotting of data in preparation for drawing the cross section.
We construct the topographic profile along the line of section and then
project apparent dips, stratigraphic or structural contacts, and any subsurface information such as wells.

MODERN STRUCTURAL PRACTICE

"204

R. W. ALLMENDINGER 2015-16

CROSS SECTIONS

CHAPTER 10

on the line of section, remember that you project them parallel to the
fold axis and not parallel to the strike. Do not project strikes and dips
across faults onto the line of section. You can plot the apparent dip
tick marks permanently. Some people prefer to use a bar and ball
symbol, that is a small circle a that location of the projected orientation with a tick mark indicating the apparent dip in the plane of the
section (Fig. 10.2).
4. Plot the geologic contacts on the topographic profile along the line of
section. These should be marked lightly and in pen or pencil and labeled.
5. Plot any subsurface constraints you may have such as seismic reflection data or borehole data as shown in Figure 10.2.
6. If the region is folded, you need to decide on a folding model as well
as the relationship between folds and faults. This will depend on your
knowledge of the field area, field observations (e.g., slickensides on
bedding planes, axial planar cleavage, etc.). In general, for unmetamorphosed sedimentary rocks, a parallel fold model is most appropriate whereas similar folds are more likely in cleaved rocks which have
suffered some metamorphism. Trishear or parallel folds are most likely to be associated with tip lines of faults, etc. If you choose a parallel
fold model, you must further decide whether the folds are kink (sharp
hinges and planar limbs), concentric (limbs not planar, hinges broad),
etc. Usually, the strikes and dips and field observations will guide you.
7. Once you decide on a fold model and the relationship between folding and faulting, you can begin to use the rules of the fold model (e.g.,
constant bedding thickness, dip isogons, kink axes, orientations of axial surfaces, etc.) to project your structures to depth.
8. For faults, be sure to show the sense of slip in cross section.
9. In general, do not draw your section deeper than the known stratigraphic section. Do not be afraid to use question marks where you
dont know how a structure behaves.

MODERN STRUCTURAL PRACTICE

"205

R. W. ALLMENDINGER 2015-16

CROSS SECTIONS

CHAPTER 10

A formal cross section is not a sketch! All features are plotted as precisely as possible
following the rules that you have established for the region based on characteristics
of the region that you have actually observed.

Fold-Fault Models
Understanding the relations between coeval faults and folds is important for
constructing the most plausible cross sections, especially in stratified rocks. We will
start out with a simple two-dimensional assumption: that all of the folds are (approximately) cylindrical. As we saw in our -diagram of the Big Elk anticline
(Chapter 3, Exercise 3.2), with natural folds this is, at best, a crude approximation,
but it does make our lives easier.

Fold Kinematics
In this chapter, we deal exclusively with cylindrical folds for the simple reason that they represent plane (i.e., 2D) strain: a cross section perpendicular to the
fold axis should capture, ideally, all of the strain. Within this general category, there

angular shear,

90

kin k

fold

es
d hing
curve

60

30

30

60

90

120

150

180

bedding dip,

Figure 10.3 Graphs of Equations 10.6 showing angular shear as a function of bedding
dip for curved and kink hinge parallel folds. Note that bedding dipping 180 is completely
overturned.

MODERN STRUCTURAL PRACTICE

"206

R. W. ALLMENDINGER 2015-16

CROSS SECTIONS

CHAPTER 10

are three basic kinematic fold models and how we draw the cross section depends
on which model we choose.
Parallel folds have Class 1 dip isogons and preserve bedding thickness
throughout the fold. More importantly for section construction, the folds are flexural slip, meaning that shear occurs on the bedding surfaces between the beds. We
can calculate the amount of angular shear, , parallel to bedding as a function of
the dip, , of bedding (Fig. 10.3):

" kink folds: = tan1 2 tan ;

curved hinges: = tan 1 ( 0.0175 )

(10.6)

Because the shear is layer parallel, the bed surfaces themselves are lines of no finite elongation. As
we will see below, this is the critical assumption that enables us to do line length
balancing. A surprisingly larger number of natural folds have relatively narrow
hinges and straight, rather than curved limbs; these are known as kink folds. In
order to preserve bedding thickness, the axial planes of kink folds must bisect the
interlimb angle; otherwise one limb will be thicker than the other. Kink folds are
common where faults change dip over a very small distance and in stratified rock
sequences with multiple thin layers with weak bedding planes. A second important
category is the concentric fold (Fig. 10.4) where all of the bedding surfaces have

Figure 10.4 Illustration of the Busk method of cross section construction. Centers of
curvature are determined from the intersection lines drawn perpendicular to adjacent
dips. Two sets of concentric arcs are highlighted in yellow and light blue. In concentric
folding, synclines become very broad down section whereas anticlines terminate in cusps.

MODERN STRUCTURAL PRACTICE

"207

R. W. ALLMENDINGER 2015-16

CROSS SECTIONS

CHAPTER 10

the same center of curvature but different radii of curvature. Concentric folding is
the basis for the Busk method of cross section construction (Fig. 10.4).
Similar folds have Class 2 dip isogons and kinematically can result from
parallel shear planes oriented obliquely to bedding. Because of the oblique shear,
these folds preserve neither bedding thickness nor bed length. They do, however,
maintain area. In areas with curved fault planes, similar folding provides and easy
and straightforward way of modeling how beds move over a bend or ramp in the
fault plane. One should not mistake easy for correct, however!
Trishear folds constitute a third kinematic model of cylindrical folding. All
shear planes emanate from a blind fault tip and spread out upward into a triangular
zone of simple shear. We discuss these folds in some detail, below.

Fold Type and the Propagation/Slip Ratio

The relationship between faulting and folding, and the type of fold that results, depends strongly on the ratio between the rate of propagation of the fault tip
line and the slip rate of the fault, or propagation-to-slip ratio, P/S (Williams and
Chapman, 1983; Hardy and Allmendinger, 2011) For example, in the idealized
fault-bend fold (Fig. 10.5a), the entire fault plane exists prior to any displacement
on the fault, a P/S = (infinity) because the propagation is a finite number while
the slip is zero. In fault-propagation folds (Fig. 10.5b), the rate of propagation is
only somewhat greater than, or can even be slightly less than, the rate of slip. The
loss of slip which is necessary near the tip line is diffused out into the folded rocks
in a way that depends on the specific model assumed. Detachment or lift-off
folds experience slip with essentially no propagation of the flat fault with which
they are associated, which is known as a decollement (the French word for unsticking or ungluing). Think of pushing a carpet across a room: when the edge of
the carpet runs into a wall the decollement between the carpet and the floor ceases
to propagate and a bulge or ruck forms at the stationary tip. Tip-line folds, commonly referred to as forced folds, associated with normal faults have negative
propagation to slip ratios because the tip-line propagates towards the surface (as in
thrust faults) whereas the slip of the hanging wall is downward relative to the foot-

MODERN STRUCTURAL PRACTICE

"208

R. W. ALLMENDINGER 2015-16

CROSS SECTIONS

CHAPTER 10

(a) Fault-bend Folds

(b) Fault-propagation Folds

0 < P/S 5

tip-line

P/S 5
Trishear

(c) Detachment Folds

P/S = 0
(d) "Forced" (drape) Folds
P/S < 0

(e) Wedge-fault Folds

Figure 10.5 Spectrum of fault fold relationships, emphasizing the importance of the propagation to slip ratio, P/S. See text for discussion.

wall. Finally, listric normal faults produce roll-over anticlines produced by similar
folding during shear oblique to the upper plate
The kinematic models for many of these fold types fall into two general categories: instantaneous limb rotation and progressive limb rotation. Faults
with sharp bends produce kink folds where the limb changes from flat to its final
dip without going through any intermediate stages. In contrast, the limbs of folds
associated with curved faults tend to rotate progressively to its final dip. The differences between these two models are particularly striking when one sees the geometry of the growth strata, sediment that accumulate in the vicinity of the fold during its growth. We discuss each of these type of folds in the following sections.

MODERN STRUCTURAL PRACTICE

"209

R. W. ALLMENDINGER 2015-16

CROSS SECTIONS

CHAPTER 10

Figure 10.6 the ramp and

flat geometry of fault-bend
folds, illustrated with kink
geometry (top) and curved
fault geometry (bottom).

hanging wall flat

Fault-bend Folds
Fault-bend folds, or ramp or hanging-wall anticlines, are so named because
the upper plate moves over pre-existing bend in the fault surface (Fig. 10.5a, 10.6).
They were first described by J. L. Rich (1934) in his mapping of the Pine Mountain
block in the southern Appalachians. Bends in faults occur because thrust faults in
stratified rocks with differing mechanical properties tend to have a stair step trajectory producing a ramp and flat geometry. Ramps occur where the fault cuts
across bedding whereas flats are where the fault is parallel to bedding. For every
step in a fault, once motion has occurred there are two ramps: where the fault cuts
across bedding in the hanging wall and where it cuts across bedding in the footwall
(Fig. 10.6). The key to restoring these folds is matching up the hanging wall ramp
with the footwall ramp.
Many structural geologists prefer the kink fault-bend fold (Fig. 10.6, top) because its geometry can be reproduced very precisely using the equations of Suppe
(1983). Suppe showed that:

MODERN STRUCTURAL PRACTICE

sin ( ) sin ( 2 ) sin

cos ( ) sin ( 2 ) sin sin

" = tan-1

" = + (180 2 ) = +
"210

(10.7a)
(10.7b)

R. W. ALLMENDINGER 2015-16

CROSS SECTIONS

CHAPTER 10

60

20

20

= 10

b = 10

10

30

40

100

70

50

80

60

90

30
20

90
90

Suppe's (1983) Fault-bend fold Gr

30
40
=

Axial Angle,

60

120

50

60
70
80

30

90
140

anticlines
0
90

60

30

Initial cutoff Angles,

Figure 10.7 Graph of the solutions to Equation 10.6 for kink fault-bend folds modified
from Suppe (1983). Knowing any two angles, the other two can be solved and a precise
cross section constructed. The green curve shows the special case of = . For any ramp
angle, , there are generally two possible solutions, mode 1 with large axial angles, , and
small hanging wall cutoffs, , and mode 2 (gray shaded region) where the reverse is true.

Where the angles are those shown in the inset diagram of Figure 10.7. Note that,
as used here, is the kink angle and not the shear strain! There are four unknown
angles and two equations; Thus one need only determine any two of the four angle
and the other two can be calculated. In practice, Equation (10.7) is usually solved

MODERN STRUCTURAL PRACTICE

"211

R. W. ALLMENDINGER 2015-16

CROSS SECTIONS

CHAPTER 10

using Newtons method of successive approximations but a graph of solutions enables anyone can construct kink fault-bend folds (Fig. 10.7). Equation (10.7a) simplifies considerably when =, which represents a simple ramp up from a decollement. For that special case, the green line in Figure 10.7 shows a maximum at =
30. In other words, the steepest ramp up from a decollement that one can have
and still preserve bedding thickness throughout the structure is 30. One of the important consequences of the basic fault-bend fold theory is that, for parallel kink
folding with no global shear in thrust plates (i.e., there is no layer parallel shear in
horizontal beds), the hanging wall cutoff is larger than the footwall cutoff ( > )
and the fault slip must decrease across the hanging wall ramp.
It is entirely possible to construct fault-bend folds using other approaches.
For example, Figure 10.6 (bottom) was constructed using layer oblique shear and
similar folding. We shall see the equations and approach for doing that when we
talk about listric normal faults, below. Increasingly, structural geologists are applying full mechanical modeling to the problem, though that is beyond the scope of
this book. In unmetamorphosed parts of thrust belts, however, the prevalence of
slickenside bedding surfaces indicates that, to a first order, a flexural slip fold model
is an adequate approximation.

Figure 10.8 Two flavors of

fault-propagation folding,
both with a propagation to
slip ratio of 2. Top, parallel
kink fold model (Suppe and
Medwedeff, 1990). Bottom, a
trishear fault propagation
fold. The dashed line is the
triangular shear zone.

MODERN STRUCTURAL PRACTICE

"212

R. W. ALLMENDINGER 2015-16

CROSS SECTIONS

CHAPTER 10

Fault-propagation Folds
Fault-propagation folds (Fig. 10.8) are produced by the propagation of the tip
line at a rate that is somewhat faster than the slip rate of the fault. In drawing cross
sections, the key question is: how is the loss of slip on the fault accommodated by
folding? Two popular kinematic models exist in the literature: The first employs
parallel kink folding where the loss of slip is compensated by enhanced growth of

90

50

40

max ramp angle for ( =) is 60

30

5
10

20

20

10

=
5

30
40
50

Axial Angle, * or *

60

= 45

30

10

a ramp < 25 produces

an overturned fold
0
90

60

Cutoff Angle, 2

30

Figure 10.9 Graphs of fault-propagation fold geometry from equation 10.7, modified from Suppe and Medwedeff (1990).

MODERN STRUCTURAL PRACTICE

"213

R. W. ALLMENDINGER 2015-16

CROSS SECTIONS

CHAPTER 10

the crest of the fold (Fig. 10.8 top) and the second, trishear, results in deformation
being resolved in a triangular zone of shear planes that radiate from the tip line of
the fault (Fig. 10.8 bottom). The former model produces no deformation in the
footwall of the fault whereas the latter yields a deformed footwall and bedding
thickness changes for that part of the beds in the triangular zone.
The kink fault-propagation fold model of Suppe and Medwedeff (1990) can
be quantified much the same way as kink fault-bend folds can. The pertinent,
messy looking equation (ignoring shear in horizontal layers) is:

sin * sin ( * 1 )

" sin 2 =

*

sin

sin

(
)
1
1
*
sin +
) sin ( 2 * )
( 1
1

Where

(10.8)

" = 90 + * 1; 2 = 180 2 * + 1; and b = 2 ( * ) .

The parallel kink (constant thickness) fault-propagation fold theory yields

several important results:
The maximum ramp up from a decollement is 60 (unlike faultbend folds where the maximum step up angle is 30).
The fold changes geometry at the stratigraphic level of the tip line:
the fold in beds above that level has a flat crest but below that level
the fold is a sharp kink.
For ramps off of a decollement, the geometry fixes P/S = 2.
The fore- and back-limb dips are quite diagnostic in fault propagation folds (Fig. 10.10). This is very useful because the dips are easy
to determine from field data.
Trishear fault propagation folds (Erslev, 1991) have more independent parameters (Fig. 10.11) and are thus not so amenable to geometric analysis. Several fine
computer programs are available to calculate forward models of these structures
and because the models can be run backwards or forwards, there is even a grid
search modeling strategy for application to real structures. The derivation following
MODERN STRUCTURAL PRACTICE

"214

R. W. ALLMENDINGER 2015-16

CROSS SECTIONS

CHAPTER 10

120
Suppe & Medwedeff (1990), Fig. 22

10

20

30

90

=5

Front Dip, f

40

10

0
=5
1

60

=
2

20
30

40

Constant thickness Theory

30
0

30
Back Dip, b

60

Figure 10.10 Fault-propagation fold geometry as a function of the forelimb and

back limb dips (modified from Suppe and Medwedeff, 1990).

Zehnder and Allmendinger (2000) of the most basic form of trishear is instructive
for its modern modeling approach.
We assume that the footwall is fixed and the hanging wall moves at a constant velocity, vo. The only particles that experience deformation are located inside
the triangular shear zone (delimited by the dashed lines in Fig. 10.11) and we need
to determine a velocity field of a point inside the triangular zone. The velocity field
will be written in vector form as

MODERN STRUCTURAL PRACTICE

" v(x, y) = vx ( x, y ) i + vy ( x, y ) j

"215

R. W. ALLMENDINGER 2015-16

CROSS SECTIONS

CHAPTER 10

Figure 10.11 The

key parameters of
the trishear kinematic model, showing
the velocities, angles,
and coordinate systems. All of the relations below are given
in the XY coordinate
system, which moves
with the tip line of
the fault. After
Zehnder and Allmendinger (2000).

where the x and y axes are parallel and perpendicular to the fault line, and " i , " j are
the usual unit vectors in the x and y directions. The origin of the coordinates is attached to the fault tip (Fig. 10.11).
!

In the hanging wall sector, " v = voi . In the footwall sector " v = 0 . We now seek
to construct a velocity field in the trishear zone that conserves area, is continuous
and matches the hanging wall and footwall sector velocities on the top and bottom
boundaries of the zone. The boundary conditions are

"

x = 0 , y = 0

on

y = x tan 1

x = 0,

on

y = x tan 2

y = 0

(10.9)

The condition that area is conserved, or that the flow is incompressible, is

that the divergence of the velocity field is zero,

" div

x y
+
= 0
x y

(10.10)

The approach will be to choose a " x field consistent with Eq (10.9), then determine the " y field from the above equations. Other than satisfying the boundary
conditions and continuity, any reasonable field and any combination of trishear
angles can be chosen. Note that equation (10.10) can be stated in three dimensions
for analyzing 3-D deformation.

MODERN STRUCTURAL PRACTICE

"216

R. W. ALLMENDINGER 2015-16

CROSS SECTIONS

CHAPTER 10

The simplest of an infinite number of conditions we could assume, is that

the velocity field is symmetric, so 1 = 2 = , and that vx varies linearly in y. To simplify writing the equations, let " m = tan . One choice for " x is

" x =

0
y
sgn(y)
+1 ,

xm
2

x > 0,

-xm y xm ,

(10.11)

where sgn(y) denotes the sign of y. For s = 1, the velocity distribution is linear in x,
producing a strain rate that is nearly uniform with respect to y. It can easily be seen
that the above field satisfies the " x boundary conditions in Eqs. (10.9). To find " y ,
we differentiate Eq. (10.11) with respect to x, invoke incompressibility (Eqn. 10.10),

"

= x ,
y
x

and integrate with respect to y yielding

2
0m y
"y =
+C
2(1+ s) mx

The constant of integration, C, is found by using the boundary conditions given in

Eqs. (10.9). The resulting velocity field in the trishear zone is

!
" (x, y) = 0
2

2
y

m y

+
1
i
+

mx
1+ s mx

(10.12)

Although Equation (10.12) represents just one of an infinite number of possible velocity fields that fits the boundary conditions, it does a surprisingly good job
of producing realistic appearing geometries and strain distributions.

Extensional Drape Folds

We commonly think of folding in connection with horizontal shortening
though folds can also form in regions of horizontal extension. In the trishear model, where P/S < 0 normal faulting results. In these cases, the tip line propagates to-

MODERN STRUCTURAL PRACTICE

"217

R. W. ALLMENDINGER 2015-16

CROSS SECTIONS

CHAPTER 10

Figure 10.12 Green layer and

black lines: tracing of a photograph of a clay slab deformed
over a faulted rigid substrate.
Blue and red s lines of no
finite elongation as proxies for
shear planes in trishear model
of structure. Light gray and red
shaded areas: region of reverse
faulting in trishear model and
analog clay experiment, respectively. From Allmendinger
(1998).
75

wards the surface while the hanging wall moves downward with respect to the
footwall. The fold that forms in front of the advancing tip line has been called
drape fold or a forced fold because, qualitatively, sediments are being draped over
faulted rocks beneath (Fig. 10.5d). The exact same equation (10.12) can be used to
model the extensional folds. It is important to realize that the deformation that accrues in the hanging wall and footwall when a particle was located within the triangular zone remains after tip line migrates past the particle. An interesting result of
both numerical trishear forward models and analog experiments of extensional
forced folds is, when the underlying fault is steep enough, some of the shear zones
within the triangular zone will be oriented as high angle reverse faults (Fig. 10.12)!

Roll-over Anticlines and Listric Normal Faults

In listric normal faults, the hanging wall must deform to remain in contact with the footwall. Although it is possible to model such deformation using parallel folding, the standard approach is to assume shear oblique to the layers, which
is borne out by the common observation of populations of normal faults deforming the hanging wall.
The constant heave method, which uses simple vector addition, is generally
applied (Fig. 10.13). We assume that, points overlying each constant dip segment
of the fault are displaced horizontally by the same amount, h. So that the actual slip
MODERN STRUCTURAL PRACTICE

"218

R. W. ALLMENDINGER 2015-16

CROSS SECTIONS

CHAPTER 10

Figure 10.13 The constant heave

method of shear oblique to layers for deformation above a normal fault with two
different dip segments. The fault is
shown in red and the upper plate antithetic shear plane in blue dashed line.
The black triangles show the vector addition necessary to ensure that the slip
vector is parallel to the fault segment.

s1

h
s2

(a) Initial known geometry

ro l l -o

ve r a

nticlin

(b) Reconstruction of fault trace

h

reconstructed fault trace

from roll-over and shear
angle

actual fault trace used to

generate the roll-over

Figure 10.14 Graphical reconstruction of the fault trace knowing the shear angle and the shape of the roll-over anticline. (a) the initial, know geometry; (b) the
reconstruction of subsequent segments of the fault. Green dashed lines show upper plate shear planes. In (b), the fault originally used to generate the roll-over is
show as a light gray line.

vector, s, over that segment is parallel to the fault, the point must be sheared downwards by a vector that is parallel to the upper plate shear planes. Those planes are
defined by their deviation from vertical, angle in Figure 10.13. Thus, the slip vector and the amount of shear parallel to the upper shear planes changes over each
fault segment. The upper plate shear can be antithetic, vertical, or synthetic with
MODERN STRUCTURAL PRACTICE

"219

R. W. ALLMENDINGER 2015-16

CROSS SECTIONS

CHAPTER 10

respect to the shear on the fault plane. Equation (10.13) shows how the slip magnitude varies with respect to the dip of the fault segment (), the heave (h), and the
upper plate shear angle ().

" sn =

h

cos n (1+ tan n tan )

(10.13)

where n refers to the n-th fault segment. The equation works equally well for antithetic, synthetic, or vertical shear and in either extensional or shortening environments.
The same approach just describe can be used to reconstruct the trajectory of
listric faults in the subsurface if the shape of the roll-over anticline is known (Fig.
10.14a). In this method, the heave, h, is set by projecting the hanging wall shear direction from the known intersection of the fault plane and the top of the roll-over
(in pre-growth strata) to a level equal to the top of the fault (Fig. 10.14a). One then
uses the shape of the roll-over, itself, to constrain the subsequent vector additions as
shown in Figure 10.14b. The method depicted in Figure 10.14 can be used for synthetic or antithetic shear planes.

Balanced Cross-sections
The fault-fold relations described in the last section allow us to construct rigorous cross-sections that obey geometric and kinematic rules. An admissible
section is a geologic cross-section which represents the known structures of the region. A viable section is one that can be restored to an undeformed state without
gaps or overlaps that cannot be explained by geologic observation. The without
gaps or overlaps bit is an acknowledgement of strain compatibility. A cross-section which is both admissible and viable is known as a balanced cross-section.

The Fundamental Bases for Balanced Cross-sections

1. Conservation of mass At the most basic level, matter is neither created nor destroyed and little if any is converted into energy.
Thus, mass is conserved. In continuum mechanics, this condition
is specified with the continuity equation:
MODERN STRUCTURAL PRACTICE

"220

R. W. ALLMENDINGER 2015-16

CROSS SECTIONS

CHAPTER 10

"

( vi )

+
= 0
t
xi

(10.14)

This equation basically says that the change in density () with respect to
time (t) of a volume, plus the flux of mass in and out of the volume (given by the second term of the equation, v = velocity), must be equal to
zero. This equation can be expanded as:

"
where

d
+ div v = 0
dt

(10.15)

v1 v2 v3
+
+

x1 x2 x3

" div v =

(10.16)

div v is the divergence of the velocity field.

2. Conservation of Volume (no change in density) For subaerial thrust belts, compaction during deformation is not generally
important so densities do not change during the deformation.
Thus, volume is conserved. This means that:

"

d
=0
dt

Because density itself is finite, this condition, known as incompressibility, means that the divergence of the velocity field in
equation must be equal to zero:
v1 v2 v3
+
+
= 0
x1 x2 x3

" div v =

(10.17)

3. Conservation of Area (Plate Strain) Because the wavelength of individual structures in a thrust belt is short compared to
their strike parallel dimension, we can assume plane strain. This
means that there is no velocity parallel to strike (the X3 axis):

MODERN STRUCTURAL PRACTICE

"221

R. W. ALLMENDINGER 2015-16

CROSS SECTIONS

CHAPTER 10

"

v3
=0
x3

Thus, cross-sectional area perpendicular to strike must be conserved and we can write the two dimensional form of the incompressibility criterion:
v1 v2
+
= 0
x1 x2

" div v =

(10.18)

This is the necessary condition for area balancing.

4. Conservation of Bedding thickness Finally, if bedding
thickness is preserved, then we can assume parallel folding with
shear parallel to bedding. Thus, line lengths parallel to bedding is
preserved. Bedding marks the orientation of one of the two lines
of no finite elongation.
In summary:
Table 10.1
Type of Balance Dimension Assumptions

Folding Model

Volume

3D

density of rocks constant during deformation,

no compaction, pressure solution, or growth
strata

non-cylindrical or
cylindrical

Area

2D

Plane strain with no strain perpendicular to the

line of section, area preserved

Cylindrical folding (parallel,

similar, trishear, etc.)

Line Length

1D

Linear strain, no bedding thickness changes

during deformation, shear parallel to layers so
beds are lines of no finite elongation

Parallel folding (kink,

concentric, etc.)

To date, only area and line length balancing are really practical although
there is a lot of noise about true 3-D balancing. The general sorts of geologic features to watch out for which violate the assumptions of line length and area balancing are:

MODERN STRUCTURAL PRACTICE

"222

R. W. ALLMENDINGER 2015-16

CROSS SECTIONS

CHAPTER 10

Volume changes:

compaction during deformation (particularly important in accretionary prisms and foreland basin strata)

Sediment accumulation during faulting

Pressure solution (especially in carbonates but also in silicic rocks

like shales and siltstones)

intrusions, diapirs, etc.

Non-plane strain:

strike-slip faults

lateral ramps

Line Length Balancing

Most of the cross sections in the literature, especially those drawn by hand
and not with a computer program such as 2DMove, are one dimensional line
length balanced sections. There are actually two cross sections that are constructed
simultaneously: The present day, deformed section and the restored, or retrodeformed, stratigraphic section. In the case of a thrust belt, the retrodeformed section will be much longer than the present day section and the difference in length
will be the horizontal shortening that occurred due to thrusting. The basic procedure is to work from the external part of the deformed belt to the internal part.
The steps to constructing a line length balanced section are illustrated in
Figure 10.15:
1. To begin, one constructs the stratigraphic sequence in the undeformed footwall of the most external thrust fault. This will become
our ultimate frame of reference, which is signified in line length
sections as the regional pin line.
2. The deformed section of the first thrust plate, back to the next
thrust at the trailing margin of the plate, is drawn using parallel
folding constrained by the surface geology and any subsurface information available.
MODERN STRUCTURAL PRACTICE

"223

R. W. ALLMENDINGER 2015-16

CROSS SECTIONS

CHAPTER 10

(a) Deformed section

n
le

loose
line

un

u n it s

its

15

corrected section that

would produce a more
realistic fault trajectory

length

shortening = 2.33 units

(b) Restored
section

gt

3
.11
=5

deformed
loose line

.
=4

regional
pin line
length = 5.113 units

length = 4.415 units

unrealistic restored fault trajectory

Figure 10.15 A simple line length balanced section, illustrating the restoration of
the most external thrust plate, some problems with the initial restoration, and a
modification (thin dashed lines) to the deformed section to correct the unrealistic
restored fault (although more subtle problems remain).

3. Next, measure the lengths of the deformed stratigraphic horizons

and draw those as straight lines of the same length, measured from
the right hand thrust in the restored footwall.
4. The left ends of the restored stratigraphic horizons (Fig. 10.15b)
define the restored geometry of the trailing thrust fault (on the left
side of the section). This restored geometry appears unrealistic because it bends back on itself. Some explanations for this result are:
(a) The left-hand thrust fault is younger than the right hand
thrust, a so-called out-of-sequence relationship as the normal
progression of thrust in a thrust belt is from internal to external.
(b) The deformed section is in error, lacking bed length in either
the footwall of the right thrust or the footwall of the left thrust.
The suggested correction shown in Figure 10.15 adds additional bed length to the footwall of the second thrust.
MODERN STRUCTURAL PRACTICE

"224

R. W. ALLMENDINGER 2015-16

CROSS SECTIONS

CHAPTER 10

(c) Although the correction shown improves the initial trajectory

of the left thrust, it does not correct another problem which is
identified by the loose line a line drawn perpendicular to
horizontal strata in the deformed state, which is retrodeformed
with the stratigraphy in the restored state. The loose line shows
that the fold style and hanging wall ramp geometry implies
that there must have been very significant angular shear in horizontal beds. This might actually be the case, but can be tested
by looking for slickensides on horizontal bedding surfaces in
the field.
5. Correcting the bed length in the hanging wall of the right-hand
thrust (effectively lowering the ramp angle of the fault) would reduce, but not eliminate the issues raised by the loose line. To remove completely the effect would require a reevaluation of the
hanging wall ramp and the style of folding depicted in the section.
6. Note that we correct all of these issues before drawing the structure in the upper plate of the left thrust! Otherwise, all of these errors will propagate through the entire section.
The simple example in Figure 10.15 that we just went through is a special
case where the hanging wall cutoffs at the hanging wall ramp are preserved. A
much more common case is depicted in Figure 10.16a: the hanging wall cutoffs are
eroded away. In these cases, we dont know exactly where the hanging wall was relative to the footwall, so we need a local reference frame that is specific to that individual thrust plate. That is the role of local pin lines. The location of the local
pin lines is a key issue (Fig. 10.16b, c). Ideally, it should be placed in horizontal strata. If horizontal strata are not available, you can calculate the amount of angular
shear in the layers based on their dip from Equations 10.6 or Figure 10.3 and adjust the local pin line accordingly so that, when restored to vertical, there is no
shear in horizontal beds (Fig. 10.16d).
The restored thrust plate is placed as close to the footwall as possible without
overlapping it (Fig. 10.16b, c); this is the reason that many structural geologists will
claim that their balanced section is a minimum estimate (more on that, below). The

MODERN STRUCTURAL PRACTICE

"225

R. W. ALLMENDINGER 2015-16

CROSS SECTIONS

CHAPTER 10

(a)

Figure 10.16 The use of

local pin lines to reconstruct a thrust plate with
eroded hanging wall cutoffs. (a) The deformed
section. (b) restoration
using a local pin line in
horizontal rocks in section (a). (c) Restoration
using pin line B, which
introduces a layer-parallel shear in the undeformed state. In (d), we
use a local pin line C that
has been drawn to take
into account the angular
shear in dipping beds
(Eqn. 10.6).

(b)

30

(c)
note shallower ramp because
pin line B assumes no
shear in dipping
beds
24

(d)

30

initial shape of the thrust plate itself is reconstructed by measuring the lengths of
the stratigraphic horizons in both directions from the local pin line.

Area Balancing
Line length balancing is restricted to parallel folds and there are many cases
where shear oblique to layering invalidates the preservation of line length assumption. In these cases, as long as the folds are cylindrical, horizontal shortening can
still be estimated by carrying out an area balance. In its most general form, area
balancing assumes nothing more than that the deformed and initial areas must be
equal (Fig. 10.17). The deformed area can be calculated by drawing a polygon
around the region of interest. The area, A, of a polygon of n vertices can be calculated as (Judge and Allmendinger, 2011):

MODERN STRUCTURAL PRACTICE

"A=

1 n1
( xi yi+1 xi+1yi )
2 i=0

"226

(10.18)

R. W. ALLMENDINGER 2015-16

CROSS SECTIONS

CHAPTER 10

(a) Pre-deformation

(b) Deformed section

Ax
At

At

future decollement

Ax
decollement

Figure 10.17 In a simple area balance, the total area (At) of the undeformed and deformed packages (shown in blue) must be equal. Eon. 10.18 can be used to calculate those
areas. The excess area (Ax) above a regional datum of a surface can be used to calculate
the displacement, d, or the depth to the decollement, h.

where the first vertex (x0, y0) and the last vertex (xn, yn) are the same. Determination
of the initial undeformed area requires knowledge of the initial unit thicknesses on
either side of the deformed zone.
In cross sections where you can identify the local excess area (Ax in Fig.
10.17b) above a regional datum of the unit, you can extract quite a lot more information. The concept has its origins with Chamberlin (1910, 1919) who used it, incorrectly as it turns out, to differentiate thin- (Appalachians) and thick-shelled (Colorado Rockies) mountain belts. Today, we know these terms as thin- and thickskinned. As you can see in Figure 10.17b,

" Ax = dh

(10.19)

If you know the excess area, Ax, and the displacement, d, it is trivial to calculate the
depth to the decollement, h or vice versa. Using the approach of Epard and
Groshong (1993), you can use the excess area of several beds to estimate both the
displacement and the depth to the decollement. The method is illustrated in Figure
10.18:
1. Determine the excess area of each resolvable bedding surface
above its regional level. The excess area of three of the seven beds
are highlighted in different colors in Figure 10.18a.
2. Measure the height of each surface above (or below) and arbitrary
reference horizon or datum. the height is measured outside of the

MODERN STRUCTURAL PRACTICE

"227

R. W. ALLMENDINGER 2015-16

CROSS SECTIONS

CHAPTER 10

Figure 10.18 The use of excess area of successive beds to determine displacement,
depth to decollement, and average layer-parallel extension following Epard and
Groshongs (1993) method. (a) The original model to be analyzed was generated in FaultFoldForward as two propagating trishear structures with opposite vergence. The program
calculates and colors the strain ellipses as shown. (b) A line-length restoration of each
bed. Comparison of the restored length and actual original length yields the average bed
length extensional strain. (c) Table of height, excess area, and bed length strain measurements that were used to generate the graphic in (d) where the horizontal displacement and
depth to decollement are determined. The colored excess areas for three beds in (a) are
show in cells of the same color in the table in (c).

region of excess area. The reference horizon does not have to be

the decollement level.
3. Plot the excess area against the height for each bedding surface
(Fig. 10.18d). For simple structures, these data points should lie on
a straight line. The slope of the line, Ax/h, is equal to the displacement, d.

MODERN STRUCTURAL PRACTICE

"228

R. W. ALLMENDINGER 2015-16

CROSS SECTIONS

CHAPTER 10

4. The decollement is located where the excess area goes to zero.

This can be found on the graph by projecting the best fit line until
it incepts the height axis (Fig. 10.18d).
5. The displacement, d, can be compared to the restored bed lengths
(Fig. 10.18b) to determine the average extension of each bed (Fig.
10.18c). Because this example was generated from a complex forward model, we already know the strain distribution. You can see
that the actual strain is much more heterogeneous than implied by
the average bedding parallel extension (Fig. 10.18a), but in the real
world you are seldom lucky enough to have a whole bunch of
strain ellipses at your disposal!
For the purposes of evaluating the horizontal shortening in a region, area
balancing holds several advantages over line length balancing: it is independent of
fold fault kinematic model and because it can be calculated analytically, one can
formally propagate the errors associated with the calculation. It fails, however,
when one is interested in predicting the structural geometry of a single structure,
for reasons of resource extraction, seismic hazard evaluation, etc. In those cases,
one should employ a variety of kinematic and mechanical models to evaluate the
range of possibilities.

A Final Word About Balancing

Thrust belts, with their imbricated stack of individual faults, are essentially
sub-horizontal zones of simple shear. Thus, the horizontal shortening calculated
from balanced cross sections is not a principal axis of shortening but instead is just
a horizontal chord in a finite strain ellipse that is elongate and inclined towards the
hinterland of the belt.
Although the minimum estimate referred to above is commonly cited in the
literature, there are many other sources of error inherent in the construction and
calculation of shortening in balanced sections. Some of those errors include:
Appropriateness of structural model
Quality of the geologic maps used

MODERN STRUCTURAL PRACTICE

"229

R. W. ALLMENDINGER 2015-16

CROSS SECTIONS

CHAPTER 10

Eroded hanging wall cutoffs

Position of the decollement
Location of other subsurface points
Stratigraphic uncertainty (thickness & shape of the wedge)
It may come as a surprise that the uncertainty in the thickness and shape of the
stratigraphic wedge is the single biggest source of error in section balancing (Allmendinger and Judge, 2013). Even the best constrained balanced sections have total uncertainty equivalent to ~20% of the calculated horizontal shortening value
and this is for well-behaved thrust belts which deform a simple sedimentary package above basement. Although many people have attempted crustal scale balanced
sections (including yours truly), the uncertainties on those are unquantifiable to
the extent that one might question whether whole crustal balancing is a worthwhile
exercise!

MODERN STRUCTURAL PRACTICE

"230

R. W. ALLMENDINGER 2015-16

CROSS SECTIONS

CHAPTER 10

ExercisesChapter 10
Cross sections are not sketches! They are precisely drafted documents. All angles
should be measured and plotted with a protractor. There should be no vertical exaggeration, i.e., V:H = 1:1. Sloppy or incompletely labeled cross sections will have
points taken off. Many of these exercises can be done in a vector graphics program
as long as angles and line lengths can be measured accurately.
1. Youll be given a large sheet with a geologic map and space to draw a cross section. You have seen this geology before: The NE corner of the map is where
calculated dips from three point problems and determined stratigraphic thicknesses of several units. The NW corner contains the Big Elk anticline for which
you programmed a down plunge projection.

Construct a cross section along line AA on the map in the space provided below the map. A topographic profile has been provided for you. The rocks are
unmetamorphosed strata so you may assume a parallel fold model. Be sure to
follow the basic guidelines in the Drawing Cross Sections part of this chapter.
Be sure to work in pencil as you will undoubtedly be erasing a lot.
2. Derive Equation (10.6) for kink folds using an area balance. The geometries
and key angles are given for you in the following figure. The area, A, to the left
of the kink axis in the undeformed (left) and deformed (right) sections must be
the same. is the dip, is the angular shear, the kink axial angle, and h is the
bed thickness.

x1
A

x1
A

x2

x2

MODERN STRUCTURAL PRACTICE

"231

R. W. ALLMENDINGER 2015-16

CROSS SECTIONS

CHAPTER 10

3. Given the fault bend fold templates on the following page, construct (a) a mode
1, and (b) a mode 2 fault bend fold, using the relations in Figure 10.5. In each
case, determine how much slip changes from the left side of the structure to the
right side.
4. Construct a fault-propagation fold from the given the stratigraphic template on
page 234 and the angles from Figure 10.7.
5. The diagram on page 235 shows two identical rollover anticlines associated
with a listric normal fault. From the shape of the rollover, construct the listric
fault trajectory. As indicated, use antithetic simple shear in the first case and
vertical simple shear in the second case
6. Derive Equation 10.12 in this chapter. Make sure to label and explain your calculations adequately.

MODERN STRUCTURAL PRACTICE

"232

R. W. ALLMENDINGER 2015-16

CROSS SECTIONS

CHAPTER 10

MODERN STRUCTURAL PRACTICE

"233

slip in

2b. Mode 2

slip in

2a. Mode 1

Name: ______________________________

Exercise 3: Fault-bend fold exercise

R. W. ALLMENDINGER 2015-16

CROSS SECTIONS

CHAPTER 10

slip in

= 30

tip line

Exercise 4: Fault-propagation fold exercise

MODERN STRUCTURAL PRACTICE

"234

R. W. ALLMENDINGER 2015-16

CROSS SECTIONS

CHAPTER 10

MODERN STRUCTURAL PRACTICE

fault trace

rollover

5b. Construction using vertical shear

fault trace

5a. Construction using 30 antithetic shear

rollover

Exercise 5: Roll-over anticline exercise

"235

R. W. ALLMENDINGER 2015-16

CROSS SECTIONS

CHAPTER 10

MODERN STRUCTURAL PRACTICE

"236

R. W. ALLMENDINGER 2015-16

Chapter 11
Structural Interpretation of Seismic Reflection Data
Introduction
Geology presents us with a basic problem. Because rocks are opaque, it is
very difficult to see through them and thus it is difficult to know what is the threedimensional geometry of structures. This issue is particularly obvious when constructing cross sections as in the previous chapters. It is too expensive to drill closeC
D
ly-spaced holes in order to constrain the geometry of rocks in the subsurface and,

commonly, there are some structures that have no surface expression and thus cannot be projected to depth from surface outcrop.
To the rescue comes a geophysical technique for remotely sensing the subsurface using sound waves. Seismic reflection profiling has been standard practice
in the oil and gas industry for more than 50 years and is the most commonly used
technique for mapping the subsurface. Some of the most profound structural observations about our planet thrust belts have decollements, low-angle normal
faults exist are best demonstrated with seismic reflection data. In this chapter, we
will give you the bare-minimum background needed in order to begin using this
uniquely useful type of data as structural geologists.

SEISMIC REFLECTION DATA

CHAPTER 11

Echo Sounding
Seismic reflection profiling is exactly analogous to echo sounding (Fig. 11.1).
Lets examine the simple case of making an echo first to see what the important parameters are. Why do you get a reflection or an echo? You get an echo because the
densities and sound velocities of air and rock are very different. If they had the
same density and velocity, there would be no echo. More specifically, the P-wave
velocity is:

" velocity = V =

(11.1)

where E = Youngs Modulus and is the density. We tend to think of velocity increasing with density but you can see that in Equation (11.1) density is in the denominator. There are a few rock types that have high velocity but low density; the
most common one is salt.
The acoustic impedance of a material is its density times the velocity of
sound in the material, V. The reflection coefficient is:

" reflection coefficient = R =

amplitude of reflected wave

V 1V1
= 2 2

amplitude of the incident wave 2V2 + 1V1

(11.2)

This is what tells us how strong the reflection will be. If you were in Yosemite Valley making echo by shouting at the granite walls of the valley, the reflection coefficient, R 0.999944. In other words, almost all of the sound is reflected back at you
air, vair

rock , vrock

a very small
amount of
sound
continues
into the rock
most sound is reflected
back to the listener

Figure 11.1 Anatomy of

an echo. It is the density
and velocity contrast
across the air-rock interface that produces the
echo.

rock wall

MODERN STRUCTURAL PRACTICE

"238

R. W. ALLMENDINGER 2015-16

SEISMIC REFLECTION DATA

CHAPTER 11

(a) in the Earth

(b) what we measure

ground!
surface

time sound was made

time

time to go down to the

1st layer and return

depth

st

1 subsurface
interface

time to go down to the

2nd layer and return

nd

2 subsurface
interface

Figure 11.2 (a) We make a sound (red star) on the surface of the Earth, the sound then
goes down to different interfaces within the earth and some of that sound is bounced
back and recorded on the surface. (b) At the surface we can only measure the time that
the sound was made and the time that it takes for the sound to go down to each interface
and come back to the surface

from the interface, but a very small proportion actually continues into the rock (Fig.
11.1).
In seismic reflection profiling, what do you actually measure? If you think
about the Yosemite example again, we could measure the time that we made the
sound and the time that we recorded the echo. The time difference is a function of
the velocity of sound in the air and twice the distance between us and the wall because the sound has to go from us to the wall and come back again. When you
make an echo, the source of the sound (your mouth) and the receiver of the
sound (your ears) are essentially in the same place. As we will see below, in seismic
reflection profiling, the source of the sound (an explosion, a vibrating truck, etc.)
and receiver (the geophones) are offset from each other but we process them as if
they were in the same place.
The above example highlights three important things about seismic reflection profiling (Fig. 11.2):

Measure time, not depth,

The time recorded is round trip or two-way time, and

To get the depth, we must know the velocity of the rocks.

MODERN STRUCTURAL PRACTICE

"239

R. W. ALLMENDINGER 2015-16

SEISMIC REFLECTION DATA

depth

CHAPTER 11

3 km/s
3 km

6 km/s

6 km

time

1s
2s
6 km horizontal
reflector

3s

Figure 11.3 Top: geologic section showing a slow velocity sedimentary basin
on the right hand side and continuous high velocity material on the left side.
Bottom: a time section showing the distortion produced by the laterally varying
velocity.

Velocities of rocks in the crust range between about 2.5 km/s and 6.8 km/s.
Most sedimentary rocks have velocities of less than 6 km/s. These are velocities of
P-waves or compressional waves, not shear waves and most seismic reflection surveys measure P- not S-waves.
Because we measure time and not depth, although seismic reflection profiles
resemble geologic cross-sections, they are not. They are a spatially distorted picture
of the earth because rock velocities vary, both laterally and vertically. To illustrate
impact of laterally varying velocities, consider the case depicted in Figure 11.3.
This case is commonly encountered in rift provinces where sedimentary basins alternate with older, higher velocity rocks in mountain ranges. The horizontal interface at 6 km depth looks like it has a step in it on the time section at the bottom because the sound waves travel more slowly to the 6 km interface on the right hand
side than they do on the left hand side. This is just one of many types of artifacts
for which the structural geologist/seismic interpreter needs to be aware!

MODERN STRUCTURAL PRACTICE

"240

R. W. ALLMENDINGER 2015-16

SEISMIC REFLECTION DATA

CHAPTER 11

Common Mid-Point (CMP) Method

In the real earth, the reflectivity at most interfaces is very small, R 0.01,
and the reflected energy is proportional to R2. Thus, at most interfaces ~99.99% of
the energy is transmitted and 0.01% is reflected. This means that your recording
system has to be able to detect very faint signals coming back from the subsurface.
An additional complication is that, because the source usually involves a lot of energy, it must be offset from the receivers.

Data Redundancy and Signal to Noise Ratio

The standard strategy for dealing with very weak signals is to increase the
signal-to-noise ratio. If you measure something many times, the signal in which
we are interested should add together constructively (because it is the same every
time) whereas the random noise should add together destructively (because it is different every time).
st

(a) 1 Shot
source

receivers (geophones)

reflecting interface
one ray through point

(b) 2

nd

Shot

reflecting interface
two rays through point
rd

(c) 3 Shot

Figure 11.4 Cross

sectional geometry of a
standard common midpoint seismic reflection
survey where the sound
is emitted (a shot,
reflecting the fact that
dynamite used to be
the most common
source) at three different stations.

reflecting interface
three rays through point

MODERN STRUCTURAL PRACTICE

"241

R. W. ALLMENDINGER 2015-16

SEISMIC REFLECTION DATA

CHAPTER 11

The geometry by which this is achieved is shown in Figure 11.4. Each of the
three panels corresponds to one shot (i.e., one episode of making noise at a station on the surface). The black dot, and each point on the reflector with a ray going
through it, is a common midpoint (CMP), sometimes referred to as a common
depth point (CDP). Notice that there are twice as many CMPs as there are stations on the ground (where the geophones are). That is, there is a CMP directly
underneath each station and a CMP half way between each station (hence the
name common midpoint). As the source is advanced in the direction of the profile, each midpoint on the reflector of interest gets sampled multiple times.
In a complete survey, the number of traces through each midpoint will be
equal to one half the total number of active stations at any one time (not including
the ends of the lines where there are fewer traces and assuming that the source
moves up only one station at a time). The number of channels in the recording
system determines the number of active stations. Most modern seismic reflection
surveys use at least 96 (and sometimes as many as 1024 channels or more), so that
the number of traces through any one CMP will be at least 48.
This number is the data redundancy, of the fold of the data. For example,
24 fold or 2400% means that each depth point was sampled 24 times. Sampling
fold in a seismic line is the same thing as the over-sampling which you see advertised in compact disk players. In general, the higher the fold of the data, the better
the profile.

Correction for Offset from the Source

The first basic step in processing the data is to collect, or gather together all
of the traces that go through each midpoint (Fig. 11.5). Even after that book-keeping step, although each CMP has been sampled multiple times (Fig. 11.4), we cant
add them together because each ray through the midpoint is a different length and
thus the travel time will also be different (Fig. 11.5). Thus, the processor has to determine a set of velocities, known as stacking or normal move out (NMO) velocities, which will correct for the fact that each ray through a CMP has a path of a
different length. These velocities should line up all of the individual blips corre-

MODERN STRUCTURAL PRACTICE

"242

R. W. ALLMENDINGER 2015-16

SEISMIC REFLECTION DATA

CHAPTER 11

source

(a) CDP Gather

(b) CDP Gather with

NMO Correction

distance from source, x

distance from source, x

time

time

to
tNMO

tx
near
offset

far offset
the NMO velocity is whatever velocity that
lines up all the traces in a CDP gather. It
is not the same as the rock velocity

[in practice, there is no geophone at

the source because it is too noisy]

Figure 11.5 (a) Gathering together all of the traces that go through the same midpoint.
There is a trace for each recording station. The time that the reflection is recorded at each
station increases with offset from the source. (b) Before the traces can be added together,
they must be lined up by applying a stacking velocity to correct for the moveout time.

sponding to a single reflector on adjacent traces. The relation between the horizontal offset, x, and the time at which a reflector appears at that offset, tx, is:

"t = t +
2
x

2
o

x2
2
Vstacking

(11.3a)

Or

" t NMO = t x t o = t +
2
o

x2
2
Vstacking

to

(11.3b)

If you have a very simple situation in which all of your reflections are flat and there
are only vertical velocity variations (i.e. velocities do not change laterally), then you
can calculate the rock interval velocities from the stacking velocities using the Dix
Equation (11.4):

MODERN STRUCTURAL PRACTICE

"243

R. W. ALLMENDINGER 2015-16

SEISMIC REFLECTION DATA

CHAPTER 11

" Vi(1,2) =

2
Vst22t 2 Vst1
t1

t 2 t1

(11.4)

where Vi12 is the interval velocity of the layer between reflections 1 and 2, Vst1 is the
stacking velocity of reflection 1, t1 is the two way time of reflection 1, etc. The interval velocity is important because, to convert from two-way time to depth, we
must know the interval, not the stacking, velocity. Once the correction for normal
move out is made, we can add all of the traces together, or stack them. This is
what produces the familiar seismic reflection profiles.
Processing seismic data like this is simple enough, but there are huge
amounts of data involved. For example a typical academic deep seismic reflection
profile is 20 s long, has a 4 ms digital sampling rate (the time interval between
numbers recorded), and is 48 fold. In a hundred station long line, then, we have

( 200 CDPs )( 48 sums )( 20 s ) = 48 10 6

0.004 s
data sample

data samples

If you are a field geologist by training or nature, which most structural geologists
are, that seems like an intimidating amount of data! Even with modern digital
recording systems like smart phone compass apps, it is pretty tough to collect even a
couple of hundred measurements per day. But, dont let the numbers fool you: the
vast majority of the numbers collected in a seismic reflection profile are worthless,
they are noise rather than data. The geophones in a seismic reflection survey record
everything wind rustling the bushes, traffic, thunder, electrical noise not just
the reflections from subsurface interfaces in which we are interested. Much of the
data processing, and we have ignored here most of the advanced processing that is
done, is meant to filter out or subdue the noise. Therein lies the difference between
the field geologist and the geophysicist (Fig. 11.6)! Much of the process of learning
field geology is learning how to filter information when you see it in the field and
deciding small part of that data is worth writing down.

MODERN STRUCTURAL PRACTICE

"244

R. W. ALLMENDINGER 2015-16

SEISMIC REFLECTION DATA

CHAPTER 11

grass

grass

s
g ra s
tick
breez
UGH breeze
e
grass
ane
grass
airpl
grass
s
bu s h
g ra s
flow
uito
er :)
grass
mosq
!
grass
&^%#
grass
p
grass
sla
035/4
7 b dg
s
breez
g ra s
Ag h!!
e
SNAK
E!!! gra
s
ss
grass
g ra s
bird
e in
grass
stont
bo o
grass
tree
s
g ra s
grass
grass

s
g ra s
s
g ra s
buzz
fly

s
g ra s
s
g ra s

rfly
butte
s
s
a
r
g
s
g ra s
y
bunn
s
g ra s
g ra s

If a field geologist collected data

like a seismic reflection profile
Figure 11.6

Geophysicists record all the data and then filter it in the computer whereas field
geologists filter out what they consider to be noise before ever writing anything
down!

Migration
The effect of CMP processing is to make it look like the source and receiver
coincide (e.g., having 48 vertical traces directly beneath the station). Thus, all
common midpoints are plotted as if they were vertically beneath the surface. This
assumption is fine for flat layers, but produces an additional distortion for dipping
layers, because of course reflection from a dipping layer isnt beneath us (Fig. 11.7).
The effect of this distortion is that all dipping reflections are displaced down-dip
and have a shallower dip than the reflector that produced them. The magnitude of
this distortion is a function of the dip of the reflector and the velocity of the rocks.
The process of migration corrects this distortion, but it depends on well-determined velocities and on the assumption that all reflections are in the plane of the
section (see sideswipe, below). A migrated section can commonly be identified
MODERN STRUCTURAL PRACTICE

"245

R. W. ALLMENDINGER 2015-16

SEISMIC REFLECTION DATA

CHAPTER 11

geophones at surface

itio

e
tru

s
po

no

fl
f re

Figure 11.7 A dipping reflector in the subsurface will not

appear in its proper position
on a seismic reflection profile
because all CMPs are assumed
to be vertically beneath their
corresponding station at the
surface. Migration is a process
that corrects this artifact.

tor

ec

because it has broad migration smiles at the bottom and edges (Fig. 11.8). Smiles
within the main body of the section probably mean that it has been over-migrated. As we shall see, migration also removes diffractions.
a

Figure 11.8 Same portion of a seismic section, shown both unmigrated (left) and migrated (right). In the bottom section, note the arcuate, sweeping features at the bottom of
the section which are classic migration smiles.

Resolution of Seismic Reflection Data

The ability of a seismic reflection survey to resolve features in both horizontal and vertical directions is a function of wavelength:

MODERN STRUCTURAL PRACTICE

"246

R. W. ALLMENDINGER 2015-16

SEISMIC REFLECTION DATA

CHAPTER 11

At low frequencies (long

wavelengths), these three
beds will be "smeared out"
into one long waveform

layered sequence
in the Earth

Figure 11.9 The dependence of vertical resolution on frequency.

At higher frequencies (shorter

wavelengths) the three beds will be
distinguishable on the seismic section

Figure 11.10 Taughannock

Falls near Ithaca, New York is
65 m high. The snippet of
seismic section on the right
side is approximately scaled
to the photograph for a frequency of 60 Hz and velocity
of 4500 m/s.

= velocity / frequency.

Wavelength increases with depth in the Earth because velocity increases and frequency decreases. Thus, seismic reflection surveys lose resolution with increasing
depth in the Earth.

MODERN STRUCTURAL PRACTICE

"247

R. W. ALLMENDINGER 2015-16

SEISMIC REFLECTION DATA

CHAPTER 11

Vertical Resolution
Generally, the smallest (thinnest) resolvable features are 1/4 to 1/8 the dominant wavelength. At a higher frequency, finer scale layering can be resolved but at
a low frequency, detail is lost (Fig. 11.9). A shallow, high resolution seismic reflection
survey may have dominant wave lengths in the range of 60 Hz. Assuming our sedimentary rock sequence has a velocity of 4500 m/s, we can calculate that one
wavelength is about 18.75 m. Thus, you can see that even a high resolution seismic
survey is imaging layering which is much coarser in scale (tens of meters) than the
bedding or layering that the field geologist sees. This point is emphasized in Figure
11.10.

Horizontal Resolution
The horizontal resolution of seismic reflection data depends on the Fresnel
Zone, a concept which should be familiar to those who have taken optics. The minimum resolvable horizontal dimensions are equal to the first Fresnel zone (Fig.
11.11). Because frequency decreases with depth in the crust, seismic reflection profiles will have greater horizontal resolution at shallower levels. At 1.5 km depth with
typical frequencies, the first Fresnel Zone is ~300 m. At 30 km depth, it is about 3
km in width.
Consider a discontinuous sandstone body (Fig. 11.12). The segments which
are longer than the first Fresnel Zone will appear as reflections, whereas those
which are shorter will act like point sources. Point sources and breaks in the sandstone will generate diffractions, which have parabolic curvature: Because frequen-

(a) higher frequency

(b) lower frequency

first Fresnel Zone

first Fresnel Zone

MODERN STRUCTURAL PRACTICE

"248

Figure 11.11 Illustration of the first

Fresnel Zone dependence on wavelength,
. Two wavefronts, a
quarter wavelength
apart, impinge on a
horizontal surface.

R. W. ALLMENDINGER 2015-16

SEISMIC REFLECTION DATA

CHAPTER 11

(a) discontinuous sandstone bed

Fresnel zone

(a) seismic reflection profile

reflection

diffraction

diffractions
Figure 11.12 (a) cross section of a discontinuous sandstone bed. (b) holes in the bed, or
segments of the bed that are smaller than the first Fresnel zone will generate diffractions
rather than reflections.

cy decreases with depth in the crust (and wavelength increases), seismic reflection
profiles will have lower horizontal resolution at deeper levels.

Diffractions
Diffractions may look superficially like an anticline but they are not. They
can be extremely useful, especially because seismic reflection techniques are
biased toward gently dipping layers and do not image directly steeply dipping or vertical features. Diffractions help you to identify such features. For example, a vertical dike would not show up directly as a reflection but you could determine its presence by correctly identifying and interpreting the diffractions from
it. High-angle faults are seldom imaged directly on seismic reflection profiles, but
they, too, can be located by finding the diffractions from the truncated beds (Fig.
11.13). The shape and curvature of a diffraction is dependent on the velocity. At
faster velocities, diffractions become broader and more open. Thus at great depths
in the crust, diffractions may be very hard to distinguish from gently dipping reflections. Note that a well-done migration should remove all of the diffractions from
the seismic reflection profile. For that reason, it is often advantageous to see both
the unmigrated and migrated versions of a profile.

MODERN STRUCTURAL PRACTICE

"249

R. W. ALLMENDINGER 2015-16

SEISMIC REFLECTION DATA

CHAPTER 11

(a) Geologic section

(a) Seismic reflection section

Figure 11.13 (a) geologic section of a high angle normal fault; (b) schematic seismic reflection profile. The fault plane would not be imaged seismically because it is
too steep but the truncations of the offset bed would generate diffractions, allowing
one to interpret at fault (dashed line).

Artifacts
The seismic reflection technique produces a number of artifacts misleading features that are easily misinterpreted as real geology which can fool a
novice interpreter. A few of the more common pitfalls are briefly listed below.

Velocity Pullup/pulldown
We have already talked about this artifact when we discussed the distortion
due to the fact that seismic profiles are plotted with the vertical dimension in time,
not depth (Fig. 11.3). When you have laterally varying velocities, deep horizontal
reflectors will be pulled up where they are overlain locally by a high velocity body
and will be pushed down by a low velocity body (as in the example in Figure 11.3).
If unrecognized, this distortion could look like folding of the deeper layer.

Multiples
Where there are very reflective interfaces, you can get multiple reflections, or
multiples, from those interfaces. When the acoustic energy bounces off of a reflector at depth and comes back to the surface, an additional bounce can be generated by the air-rock interface (Fig. 11.14). The effective reflectivity of multiples is

MODERN STRUCTURAL PRACTICE

"250

R. W. ALLMENDINGER 2015-16

SEISMIC REFLECTION DATA

CHAPTER 11

(b) Seismic reflection section

(a) Geologic Section

simple
raypath

time

depth

primary reflection
multiple
raypath
multiple at twice the
travel time of the primary

Figure 11.14 (a) Geologic section and (b) a simple multiple on a seismic section
generated by the ray path bouncing off of the underside of the surface of the Earth.

(b) Seismic reflection section

depth

primary re

simple
rathpath

flection

mu
trav ltiple at
el tim
tw
e of ice the
the
prim
ary

multiple
rathpath

time

(a) Geologic Section

time

depth

primary reflections
simple
raypaths

pegleg
raypath

pegleg multiple

Figure 11.15 More complicated multiples can result from dipping horizons (top)
and from reverberation between two layers (bottom).

the product of the reflectivity of each reflecting interface. For simple multiples (see
below) then,

MODERN STRUCTURAL PRACTICE

Rmultiple = R2primary.

"251

R. W. ALLMENDINGER 2015-16

SEISMIC REFLECTION DATA

CHAPTER 11

If the primary reflector has a reflection coefficient of 0.01 then the first multiple
will have an effective reflection coefficient of 0.0001. In other words, multiples are
generally only a problem for highly reflective interfaces, such as the water bottom
in the case of a marine survey or particularly prominent reflectors in sedimentary
basins (e.g. the sediment-basement interface).
In addition to the simple multiple shown in Figure 11.14, more complicated
reverberations can exist. For example, multiples from dipping layer have twice the
dip of the primary reflector and you can have peg-leg multiples that result from the
ray path bouncing around between two layers in the subsurface (Fig. 11.15)

Sideswipe
In seismic reflection profiling, we assume that all the energy that returns to
the geophones comes from within the vertical plane directly beneath the line of the
profile. Geology is inherently three-dimensional so this need not be true. Even
though geophones record only vertical motions, a strong reflecting interface which
is out-of-the-plane can produce a reflection on a profile, as in the case illustrated in
Figure 11.16. Reflections from out of the plane is called sideswipe. Such reflections will cross other reflections and will not migrate out of the way. (Furthermore
they will migrate incorrectly because in migration, we assume that there has been
no sideswipe!) The main way of detecting sideswipe is by running a sufficient
number of cross-lines and tying reflections from line to line. Sideswipe is particularly severe where seismic lines run parallel to the structural grain.

out of plane ray path

(sideswipe) from dike
dire
c
pro tion of
file

dike

e
ton

ds
san

Figure 11.16 Illustration of a

geometry where you could get
out of plane reflections known
as sideswipe.

in plane ray path

from sandstone

MODERN STRUCTURAL PRACTICE

"252

R. W. ALLMENDINGER 2015-16

SEISMIC REFLECTION DATA

CHAPTER 11

Buried Focus
Tight synclines at depth can act like concave mirrors to produce an inverted image
quite unlike the actual structure. Although the geological structure is a syncline, on
the seismic profile it looks like an anticline. At the dawn of the seismic reflection
age, many an unhappy petroleum geologist has drilled a buried focus hoping to find
an anticlinal trap! The likelihood of observing a buried focus increases with depth
because more and more open structures will produce the focus. A good migration
will correct for buried focus.
(b) Seismic reflection section

(a) Geologic Section

d
c

f
e

time

depth

Figure 11.17 Tight synclines such as that depicted in (a) produce multiple
CMPs on a single layer per station, an effect known as a buried focus. (b)
When we use standard CMP processing, we plot all of those points from different parts of the structure directly beneath the station resulting in a bowtie
pattern and what appears to be an anticline where there is really a syncline.

Structural Interpretation of Seismic Profiles

Modern seismic interpretation software packages are extremely powerful and
provide excellent tools for interpretation of both two- and three-dimensional data.
In fact, it is virtually impossible to interpret 3D data cubes without the benefit of
such packages. Here, we have much more modest aims: how to extract useful structural data from traditional 2D data. Because every seismic profile is different, there
is no cookbook approach to interpretation. Perhaps one of the most important
things to remember is that seismic reflection data is biased towards horizontal reflectors and steeply dipping interfaces cannot be imaged at all. In ad-

MODERN STRUCTURAL PRACTICE

"253

R. W. ALLMENDINGER 2015-16

SEISMIC REFLECTION DATA

CHAPTER 11

dition, there are some basic things that all interpreters look for: truncations, discontinuities, and kink axes.

Truncations and Discontinuities

Truncations imply an abrupt geological boundary, though there are many
potential explanations. Among the most common are: truncation at a fault, intrusive or diapiric contact, an angular unconformity, or various types of downlap or
onlap. Which of these you choose will depend on your knowledge of the area.

Figure 11.18 Migrated seismic reflection profile showing the truncations (red arrows) of
well-defined reflections corresponding to a Cenozoic sedimentary basin sequence against
a much less reflective zone. a and b show the base of the basin sequence in the hanging wall and footwall. Truncations are interpreted as a thrust fault, even though there are
no fault plane reflections.

MODERN STRUCTURAL PRACTICE

"254

R. W. ALLMENDINGER 2015-16

SEISMIC REFLECTION DATA

CHAPTER 11

Figure 11.19 Migrated seismic reflection profile showing a discontinuity with

dipping reflectors above the relatively flat lying reflector indicated by the red arrows.

Truncations with layered reflections that represent strata on both sides are most
likely either fault truncations, unconformities, or onlap or downlap. Truncations at
a high angle with layered reflections on one side and a featureless area on the other
could be fault, intrusive, or diapiric contacts.
Figure 11.18 shows numerous truncations, indicated by the red arrows. This
section comes from an area of known thrust-faulting, so it is highly likely that the
main set of truncations define a curved thrust fault inclined towards the right. The
upper plate of this thrust has a zone of poorly resolved reflections that probably
correspond to basement rocks or more highly folded and deformed Paleozoic strata. On top of this relatively transparent zone, above horizon a, lies a perched
Cenozoic sedimentary basin. The interpreted thrust fault parallels the base of the
Cenozoic basin suggesting that the fault is in a position of a hanging wall flat over a
footwall ramp. There are also more subtle truncations below horizon b at the
base of the well-defined Cenozoic basin in the footwall. These could represent an
angular unconformity but with the evidence available, one could not rule out the
possibility of a sub-basin fault that parallels the Cenozoic basin fill.
Figure 11.19 shows an example of a discontinuity with dipping reflectors
above a relatively flat reflection horizon indicated by the red arrows. This geometry
is commonly indicative of a decollement or detachment in the crust above which
the rocks have been either shortened or extended relative to the rocks below (you
will get to decide which in the exercises!).
MODERN STRUCTURAL PRACTICE

"255

R. W. ALLMENDINGER 2015-16

SEISMIC REFLECTION DATA

CHAPTER 11

Kink Axes
Even folds with curved hinges
commonly have long, relatively
straight limbs so it is possible to define
kink axial surfaces on seismic reflection data. Recall that, in fault-related
C
D
folds, kink axes are systematically related to the hanging wall and footwall
A
ramps and thus, drawing kink axes on
the seismic data can aid in the construction of the underlying fault
geometry. In parallel folds, kink axes
B
should bisect the angle between the
two limbs, but on time sections, this
will only be true if the section is displayed at a 1:1 scale for the rocks involved. Figure 11.20 depicts a small
Figure 11.20 Kink band on a migrated
kink band imaged on a migrated
seismic reflection profile. Sequence D
seismic reflection profile. Despite
changes thickness across the kink band producing the growth axial surface, C.
the curved nature of the hinge regions, the axial surfaces A and B
are relatively straightforward to position. They are parallel to each other and are reasonably close to being interlimb
angle bisectors, suggesting that this part of the seismic section has little vertical exaggeration.

Kink Axes and Growth Strata

The axis C in figure 11.20 is poorly defined but is necessary because axis B
disappears up-section before A does. The reason for this becomes clear upon careful inspection of stratigraphic section labeled D: this section is thicker to the left of
B/C than it is to the right of A. Thus, we can interpret that the kink band began to
grow at a time equivalent to the age of the strata at the change from B to C. The

MODERN STRUCTURAL PRACTICE

"256

R. W. ALLMENDINGER 2015-16

SEISMIC REFLECTION DATA

CHAPTER 11

end of growth would correspond to the age of the strata where C meets A; the exact position of this juncture is not well imaged on this section because of the lack
of resolution higher in the profile. Axis C, which is known as a growth axial surface, does not bisect the angle between the limbs because the thickness changes
across this boundary. Sequence D is known as growth strata because they accumulate on and around a structure during its growth. In general, axes that produce
folding of the entire section, such as A, are know as active axial surfaces and are
commonly fixed to the footwall ramp of the underlying structure. B is known as a
fixed axial surface and moves with the hanging wall ramp of the fault (Suppe et
al., 1992). There is quite a bit more to interpreting growth strata geometries than
what we have touched on here but, to review that, would make an already long
chapter even longer!
A convenient plot for identifying growth strata is the vertical separation
diagram. Vertical separation is simply the vertical difference between the depth to
a horizon in one vertical section (or drill hole) on top of the growing (or subsiding)
structure and that in another vertical section lying outside of the structure. In pregrowth strata, the vertical separation is constant whereas in the growth strata, the
vertical separation diminishes up-section towards zero at the depositional surface
(Fig. 11.21). One great advantage of the vertical separation diagram is that it does
(a) Geologic section

(b) Vertical separation diagram

Vertical separation

C
B

Depth in Well 1

Figure 11.21 Construction of a vertical separation diagram over a growth fault-bend

fold. This diagram has no dependence on fault-fold kinematic models

MODERN STRUCTURAL PRACTICE

"257

R. W. ALLMENDINGER 2015-16

SEISMIC REFLECTION DATA

CHAPTER 11

not depend at all on the fault-fold kinematic model; one simply needs two vertical
sections. It works equally as well for growth normal faulting on passive margins as it
does for thrust faulting. The vertical separation diagram is also quite useful when
studying fault reactivation (Mitra, 1993).

MODERN STRUCTURAL PRACTICE

"258

R. W. ALLMENDINGER 2015-16

SEISMIC REFLECTION DATA

CHAPTER 11

Exercises Chapter 11
1. The diagram below shows a very simple geological cross section where basement rock with a velocity of 6 km/s has been thrust over a sedimentary basin
with a velocity of 3 km/s. Accurately construct what an unmigrated time section would look like, assuming both the base of the basin and the thrust fault
plane are reflective.
0
30
6 km/s
3 km/s

Depth (km)

4
6 km/s

Two-way Time (s)

MODERN STRUCTURAL PRACTICE

"259

R. W. ALLMENDINGER 2015-16

SEISMIC REFLECTION DATA

CHAPTER 11

2. Interpret the migrated time section on the following page by plotting and inter preting kink axes. Finalize your interpretation by constructing a depth section
in the box below the seismic section. Use the stacking velocities and times at
points A, B, and C, given in the tables, below, to calculate the depths. Recall
that the time is the two-way travel time.

Velocity at Point A
Time (s)

Velocity at Point B

Stacking Velocity
(km/s)

Time (s)

Stacking Velocity
(km/s)

0.6

3.64

0.6

4.27

4.4

0.88

4.8

1.25

4.8

1.54

5.52

1.5

5.0

1.9

5.6

1.8

5.3

2.7

5.8

2.6

5.8

3.0

5.9

3.0

5.8

4.0

6.3

Velocity at Point C
Time (s)

Stacking Velocity
(km/s)
0.6

2.94

0.8

3.27

1.1

3.8

1.3

4.1

2.2

5.0

2.8

5.4

3.0

5.6

MODERN STRUCTURAL PRACTICE

"260

R. W. ALLMENDINGER 2015-16

SEISMIC REFLECTION DATA

CHAPTER 11

B
A
MODERN STRUCTURAL PRACTICE

"261

R. W. ALLMENDINGER 2015-16

3.0

2.0

Depth (km)

1.0

Two-way time (s)

SEISMIC REFLECTION DATA

CHAPTER 11

MODERN STRUCTURAL PRACTICE

1.5

1.0

0.5

ground!
surface

2 km

3. Interpret the section, below, using kink axial surfaces and, if you identify any
growth strata, indicate at what level growth initiated. Is the structure still active?

"262

R. W. ALLMENDINGER 2015-16

Prob 08-3

Name: _____________________________
SEISMIC REFLECTION DATA

CHAPTER 11

Growth History Exercise

4. The
table, below, shows the tops of dated formations in two well, A and B.
The table, below, shows the tops of dated formations in two wells, A & B. Make a vertical separation vs.
depth
diagram
and determine
the following:diagram and identify (a) which well is on the
Construct
a vertical
separation
growing
thegrowing
time(s)
of growth; (c) the rate of vertical displacement
1. Whichstructure;
well is located(b)
on the
structure(s),
of the structure (assuming uniform compaction), and (d) the growth history of
2. The time(s) of growth,
the region.
3. The rate(s) of vertical displacement of the structures assuming uniform compaction, and
4. The growth history of the region.

Formation Age

Depth in Well A
(m)

Depth in Well B
(m)

Plio-Quaternary

250

250

Pliocene

500

500

8 Ma

800

738

upper Miocene

1355

1180

10 Ma

1950

1650

Miocene

2280

1975

middle Miocene

2730

2435

14.3 Ma

3198

2900

lower Miocene

3540

3160

lower Miocene

3875

3295

lower Miocene

4162

3405

27.5 Ma

4480

3543

Eocene

4700

3678

Paleocene

5090

4065

Upper
Cretaceous

5510

4485

Upper
Cretaceous

5855

4830

MODERN STRUCTURAL PRACTICE

"263

R. W. ALLMENDINGER 2015-16

SEISMIC REFLECTION DATA

CHAPTER 11

MODERN STRUCTURAL PRACTICE

"264

R. W. ALLMENDINGER 2015-16

Chapter 12
Solid Mechanics and Structural Geology
Introduction
Up to this point, we have focused mostly on geometry and kinematics, borrowing a smattering of concepts from the much broader realm of solid mechanics.
The concepts that we have developed to this point vectors, tensors, stress, strain,
some basic material models, etc. cover about the first 25-30% of a standard continuum mechanics textbook. This background provides a splendid point of depar11
12

21
a

Pf

22

ture for beginning to explore the rich world of solid mechanics. One must become
conversant with this world if you want to explore why structures form and behave
as they do. Solid mechanics is a broad field and there are entire geology-oriented
books devoted to this subject (Johnson, 1970; Turcotte and Schubert, 1982; Jaeger
and Cook, 1976; Middleton and Wilcock, 1994; Pollard and Fletcher, 2005).
The purpose of this chapter is to give you a glimpse of the basic approach
used in a more complete analysis provided by mechanics, as well as review some
fundamental results that are particularly germane to structural geology. Bear in
mind that this is only a taste, a teaser for the real thing. Hopefully, this will give you

SIMPLE MECHANICS

CHAPTER 12

the motivation to dive in deeper, either on your own or in subsequent classes, to explore this world.

The Mechanical Approach

A mechanical approach involves a clear definition of the components necessary to solve the problem of interest. The starting point is commonly the simplifying assumption that the distribution of properties in a material is continuous; i.e.,
that the material is a continuum. This is the origin of the term, continuum mechanics. Because properties vary smoothly in time or space, we describe them in
terms of gradients, which mathematically are derivatives. Because the Earth is a
three dimensional place, the gradients in which we are interested vary in all directions and are specified along the three axes of our Cartesian coordinate system, so
we will define how things vary in terms of partial differential equations. We
have already had a taste of this in Chapters 7 and 8 where the equations governing
three dimensional strain are expressed as partial derivatives of displacement (or po-

Physical Principles!
!
Conservation of mass!
!
Conservation of linear momentum!
!
Conservation of angular momentum!
!
Strain compatability

Always hold in classical

mechanics, regardless of the
problem, setting, or materials

Constitutive Equations!
!
Elasticity!
!
Viscosity!
!
Plasticity!
!
Combinations of the above

Depend on the material being

analyzed and the environmental
conditions (pressure,
temperature, time span, etc.)

Limiting Conditions!
!
Boundary conditions!
!
Initial Values

Depend on the specific problem

being analyzed and what we
know, a priori, about it

Figure 12.1 Hierarchy of components of a full mechanical analysis.

MODERN STRUCTURAL PRACTICE

"266

R. W. ALLMENDINGER 2015-16

SIMPLE MECHANICS

CHAPTER 12

sition) with respect to position. The basic approach relies on three levels of abstraction (Fig. 12.1; from most general to most specific): (1) physical principles, (2)
constitutive equations, and (3) boundary conditions and initial values.

Physical Principles
Physical principles are those which apply to any body or substance. Any deformation of a continuous medium that we wish to analyze must conform to these
principles which form the fundamental basis of classical mechanics. The first of
these is conservation of mass. As we have already seen in the introduction of
balanced cross-sections and the trishear fault-fold model, the principle of conservation of mass is defined by the continuity equation:

"

( vi )
d
+
= 0
dt
xi

(12.1)

where the first term is the material derivative of density with respect to time (sometimes written using capital D). This equation states that the change in density
with time plus the flux of material in or out of the system must be equal to zero. It
is from this general equation that we derive the specialized condition of incompressibility for volume constant deformation:

v1 v2 v3
+
+
= 0
x1 x2 x3

" div v =

(12.2)

The cornerstone of physical principles follows from Newtons Second Law,

which deals with the conservation of momentum. The momentum of a body is
equal to its mass, m, times its velocity, v. Newton stated that the rate of change of
momentum is proportional to and in the direction of the impressed force, F:

"F=

d ( mv )

dt

(12.3a)

The term, impressed force means the vector sum of all of the forces acting on a
body. If the mass does not vary with time, then we can write Newtons second law
in a more familiar format:
MODERN STRUCTURAL PRACTICE

"267

R. W. ALLMENDINGER 2015-16

SIMPLE MECHANICS

CHAPTER 12

"F= m

dv
= ma
dt

(12.3b)

where a is the acceleration. When the force equals zero, momentum must be constant.
From the condition of Newtons second law, one can derive the equations
of motion in various formats (see Malvern, 1969 or Pollard and Fletcher, 2005 for
details of the derivation). Perhaps the most general and insightful is given by
Cauchys First Law of Motion:

"

dvi ji
=
+ gi
dt
x j

(12.4)

As with Equation (12.1), d/dt is the material time derivative and gi is acceleration
due to gravity. The terms in this equation have units of force per unit volume. This
equation then says that the total force per unit volume is equal to the gradient of
stress with distance [with units of N m2 m1 = (kg m s2) m2 m1 = (kg m s2) m3]
or surface forces per volume plus the body forces per volume.
Torque, or moment, is the force multiplied by the distance from a pivot
point or fulcrum in a material. In a way exactly analogous to what we have just
seen, the conservation of angular momentum says that the sum of all torques is
equal to the rate of change of total angular momentum. Cauchys Second Law
of Motion is an expression of this and its result is a simple and elegant proof that
the stress tensor must be symmetric.

" ij = ji

for

i, j = 1 to 3

For bodies in equilibrium, the change of linear and angular momentum with
respect to time must equal zero. This condition yields two fundamental relationships: the balance of forces and the balance of moments (i.e., torques):

" F = Fsurface + Fbody = 0

(12.5)

"M = 0

(12.6)

MODERN STRUCTURAL PRACTICE

"268

R. W. ALLMENDINGER 2015-16

SIMPLE MECHANICS

CHAPTER 12

These are the starting conditions for analyses involving mechanics of static equilibrium. When you draw a free body diagram, it should depict all of the forces
and torques on a body and, if it is a problem in static equilibrium as many problems in geology are, those should all sum to zero.
Finally, for some problems, we wish to ensure that the displacement field associated with a particular strain field is single valued and continuous. That is, the
strains imposed produce no gaps or overlaps. This condition of strain compatibility which is specified by St.-Venants equations:

2 ij
2 jl
2 kl
2 ik
"
+

= 0
X k Xl Xi X j X j Xl Xi X k

(12.7)

where ij is the infinitesimal strain tensor (Chapter 7). Recall that for infinitesimal
strain the material and spatial coordinates are the same. Equation (12.7) represents
six equations that must be satisfied if the displacement field is to be smooth and
continuous. This equation finds important applications in elasticity theory and is,
for example, one of the underlying tenants of the construction of the world strain
map (Holt et al., 2000). Nonetheless, in the pantheon of physical laws it is a lesser
god and there are a number of perfectly physically plausible geological processes
that do not comply.

Constitutive Equations
So far in this chapter, we havent said anything about materials, yet, and how
they respond to applied forces (or body forces). Geological materials are extremely
complex and different processes may be active at different scales and even in adjacent mineral grains. Nonetheless, there are a small number of models that successfully describe the macroscopic behavior of many natural materials under different
conditions. We have already reviewed the three basic models in Chapter 9: elastic,
plastic, and viscous.
Elasticity
Because rocks in the upper crust deform by fracturing at even modest strains,
elasticity theory is intimately related to the concepts of infinitesimal strain (Chapter
MODERN STRUCTURAL PRACTICE

"269

R. W. ALLMENDINGER 2015-16

SIMPLE MECHANICS

CHAPTER 12

7). We have already seen some of the basic equations of elasticity in Chapter 9. In
elastic deformation, by Hookes Law stress is linearly related to strain by a variety
of elastic moduli depending on the type of deformation:
11 = E11 where G = Youngs Modulus (for axial deformations)

" ij = G ( 2 ij ) where G = Shear Modulus (for i j; i.e., simple shear deformations)

( 11 + 22 + 33 )

(12.8)

3 = B ii where B = Bulk Modulus (for volume change deformations)

An important additional parameter, Poissons Ratio, , defines the relation of the

axial to the transverse strain (Fig. 9.1):

w f wi
w
e
i
" = t =

e
f i

i

(12.9)

For incompressible deformation, = 0.5, but rocks have 0.10 0.33. Poissons
ratio can be used to relate the elastic moduli of equation (12.8):

"G =

3B (1 2 )
E
=

2 (1+ )
2 (1+ )

(12.10)

Thus, for a linear isotropic material, only two elastic moduli are necessary to the
elastic deformation. The constitutive equations for linear elasticity are usually written using Lams constants:

" ij =

(1+ )
E

ij

ij kk
E

" ij = 2G ij + ij I

(12.11a)
(12.11b)

Where I is the first invariant of the infinitesimal strain tensor and ij is the Kronecker delta. The shear modulus, G, is sometimes written using the Greek letter, .
The Lam constant, , is related to the other elastic moduli by:

MODERN STRUCTURAL PRACTICE

"270

R. W. ALLMENDINGER 2015-16

SIMPLE MECHANICS

CHAPTER 12

" =

E
2
and = K G
3
(1+ )(1 2 )

(12.12)

Viscosity
Over long time spans, even seemingly solid materials creep viscously. In the
simplest form of viscosity, a Newtonian fluid, shear stress is linearly related to the
deformation rate (Fig. 9.3) via the viscosity, :

" = !

(12.13)

In a manner similar to elasticity, the constitutive equation for linear viscosity can be
written:

" ij = p ij + ij

v v
vk
+ i + j
xk
x j xi

(12.14)

Where p is the thermostatic pressure, is the second coefficient of viscosity

and v is the velocity. You may recall that the mean pressure is equal to the first invariant of the stress tensor divided by three (Chapter 5):

" mean =

11 + 22 + 33

3

(12.15)

The thermostatic pressure can be related to the mean stress by:

" mean = p + +

2 vk

3 xk

(12.16)

Plasticity
Unlike the simple, linear forms of elasticity and viscosity theory, plasticity is
inherently nonlinear and requires the use of hyperbolic partial differential equations. So, we will not pursue plasticity any more here except to point you towards
the classic reference in the field: Hill (1950).

MODERN STRUCTURAL PRACTICE

"271

R. W. ALLMENDINGER 2015-16

SIMPLE MECHANICS

CHAPTER 12

Boundary Conditions and Initial Values

The general procedure for carrying out a mechanical analysis is to solve a set
of differential equations that result from manipulation of the physical principles
and appropriate constitutive equations by integration. This results in constants of
integration that must be evaluated. We have seen an example of this already (although not in the context of a full mechanical analysis): In Chapter 10, the velocity
field for the trishear model was derived from the condition of incompressibility and
the and an arbitrary choice for the vx component of the velocity field. Integrating to
solve for the vy component resulted in a constant of integration, which we evaluated
based on the boundary conditions on the two borders of the trishear zone (Eqn.
10.8).
In general, to solve the integrated differential equations, you must specify either the boundary conditions or the initial conditions. boundary conditions are
limiting values or conditions on the dependent variables at the edges of your model. If you are analyzing the flow of material in a channel, a boundary condition
might be that the velocity of the flow must go to zero at the edge of the channel, or
in the case of the trishear model just discussed, that the velocity in the triangular
shear zone must go to zero on the footwall boundary of the zone. In many problems, one might assume that an important boundary condition is that the surface
of the Earth is a traction free surface and thus must be perpendicular to a principal
stress. Initial conditions are the values of the time dependent variables at time zero
of your analysis. In the case of the flow in the channel, you might specify the velocity of the fluid entering the channel.
Commonly, one specifies either the boundary conditions and solves for that
dependent variable in the interior of the body or the initial conditions and solves
for the values of that dependent variable at some later time. Imagine that you are
studying the formation of a laccolith (Johnson, 1970; Pollard and Fletcher, 2005).
You would specify where vertical displacements go to zero (boundary conditions)
and, via elasticity theory, solve for the displacement of the bending layer. A problem where the boundary conditions are set is known as a boundary value problem whereas in analyses where the initial conditions are set it is known as an initial value problem. Needless to say, which type of analysis you do is dependent

MODERN STRUCTURAL PRACTICE

"272

R. W. ALLMENDINGER 2015-16

SIMPLE MECHANICS

CHAPTER 12

on what you know already and what your objectives are in the analysis. Take the
classic physics problem of a projectile (Middleton and Wilcock, 1994): if you know
the mass, angle, and the velocity of the projectile, you know the initial conditions
and can calculate how far the projectile should travel and where it will land. From a
more practical standpoint, however, you know where the target is that you want to
hit (a boundary condition) and you want to calculate the initial velocity and angle
that is necessary to hit the target.

Some Simple Geological Examples

There is a rich geological literature of mechanical analysis of structures at
various scales and complexity. In this section, however, we will limit ourselves to
some simple, yet powerful results, first involving rigid bodies and then from linear
elastic fracture mechanics. These results are germane to topics that we have already
discussed: thrust belts, hydraulic fracturing and flexure.

Mechanics of Thrust Belts

Hubbert and Rubeys (1959) Force Balance for Thrust Plates
One of the most famous papers in structural geology was entitled Role of
fluid pressure in mechanics of overthrust faulting (Hubbert and Rubey, 1959).
This paper build on earlier work by M. K. Hubbert (1951) and analyzed the case
of a block of geological dimensions that was pushed over a pre-existing surface.
Their analysis is a particularly nice example of static equilibrium balance of forces.
In their paper, they only balance forces in the X1 direction although Hubbert was
clearly aware that this did not constitute a total force balance (Fig. 12.2). Consult
Pollard and Fletcher (2005, p. 255-260) if you wish to see the entire three dimensional force balance. The free body diagram (Fig. 12.2) is posed in terms of tractions (stress vectors); we will have to covert these to forces by integrating the traction along the area of interest (i.e., the side or bottom of the block) in order to do
the force balance. In the one-dimensional force balance, the normal force on the
left side of the block should be equal to the frictional shear force on the base of the
block:

MODERN STRUCTURAL PRACTICE

"273

R. W. ALLMENDINGER 2015-16

SIMPLE MECHANICS

CHAPTER 12

X3
x
13

13

11

gz
31

X1
33

Figure 12.2 Free body diagram for pushing a block across a horizontal surface,
the basic problem that Hubbert and Rubey (1959) set up to explore the importance
of frictional resistance to sliding. The tractions in red are the ones that they explicitly analyzed. The black tractions would be necessary for a complete two-dimensional force balance. Fortunately, the lithostatic component of the normal tractions
and the shear tractions on the ends of the block cancel each other out in the full
analysis.

" 0 11 dx3 = 0 31 dx1

(12.17)

Evaluating the right hand side of the equation, recall that the frictional resistance is
a function of the normal stress, 33 = gz, times the coefficient of static friction
(from Byerlees law):

" 0 31 dx1 = 0 s 33 dx1 = 0 s gz dx1 = s gzx

(12.18)

Evaluating the left side of Equation (12.17) requires a slight explanation:

There are two possible outcomes as the value of 11 is increased: (a) the push will
exceed the frictional resistance to sliding that we have just calculated and the entire
block will slide coherently, or (b) 11 will exceed the fracture strength of the material
and the block will break up rather than slide as a rigid block. The problem we are
going to solve is actually the latter. To do so, we need an expression for the
Coulomb Failure criteria in terms of the principal stresses (Eqn. 6.8, repeated
here):

MODERN STRUCTURAL PRACTICE

"274

R. W. ALLMENDINGER 2015-16

SIMPLE MECHANICS

CHAPTER 12

" 1 = Co + K 3

where

K=

1+ sin
1 sin

and

Co = 2So K

(12.19)

where So is the cohesion and is the angle of internal friction. Under these conditions, 11 = 1 and 33 = 3. Now, we can expand the left side of Equation (12.17):

" 0 11 dx3 = 0 ( Co + K 3 ) dx3 = 0 ( Co + K gz ) dx3 = Co z +

z

K gz 2

2

(12.20)

With both sides of Equation (12.17) evaluated, we have:

" Co z +

K gz 2
= s gzx
2

(12.21)

Rearranging the result of Equation (12.21) we get an expression for the length of
the block, x, in terms of its thickness, z, the friction along the base, etc.:

" xmax =

Co
Kz
+

s g 2 s

(12.22)

For z = 5 km, = 2750 kg m3, = 30, and s = 0.85, we calculate the maximum
length of the block is 11.8 km, which is much less than the dimension of large
thrust sheets, such as the Lewis overthrust in Glacier National Park that can be
tracked 80 km or more down dip and hundreds of kilometers along strike.
The basic problem is that, at these dimensions, rocks are fundamentally weak
as Hubbert demonstrated in an earlier paper when he posed the thought experiment of whether a crane large enough could lift the entire state of Texas! This
simple analysis captures the so-called paradox of low-angle thrust faults that structural geologists have been debating since the early 1900s. Hubbert and Rubey went
on to propose that pore fluid pressures, combined with a thrust decollement that
dipped gently towards the foreland, could explain large thrust plates. Their work on
pore fluid pressures was pioneering but, alas, they were wrong about the dip of the
decollement as well as the shape of the thrust block. Our modern understanding of
the mechanics of thrust belts as critically tapered wedges is summarized by Dahlen
(1990).

MODERN STRUCTURAL PRACTICE

"275

R. W. ALLMENDINGER 2015-16

SIMPLE MECHANICS

CHAPTER 12

water

1
g

Figure 12.3 Diagram showing the coordinate system, key angles, and principal
stresses for the general critically tapered wedge model, after Dahlen (1990).

Critically-Tapered Wedges (Dahlen, 1990)

In a remarkable series of papers, beginning in 1983 (Davis et al., 1983) and
culminating with Dahlen (1990), the Princeton group laid out the modern mechanical basis for understanding thrust belts. Their analysis built on earlier work by Elliot (1976) and Chapple (1978), both of whom recognized that thrust belts in cross
section had the form of a finely tapered wedge rather than an rectangular block. In
this section, we summarize Dahlen's (1990) general two dimensional force balance
in a non-cohesive wedge.
The equations of static equilibrium (force balances) in terms of partial differential equations take into account the z as well as the x direction (Fig. 12.3).
Summing in the x direction first, we get:

"

xx xz
+
gz sin = 0
x
z

(12.23a)

xz zz
+
+ gz cos = 0
x
z

(12.23b)

and in the z direction:

MODERN STRUCTURAL PRACTICE

"276

R. W. ALLMENDINGER 2015-16

SIMPLE MECHANICS

CHAPTER 12

At the upper surface of the wedge, the boundary conditions are: z = 0; xz = 0 (i.e.,
no shear stress on the surface of the wedge); and zz = fgD (the weight of the overlying water, or 0 in the case of subaerial wedges). The Hubbert and Rubey pore
fluid pressure ratio in the interior of the wedge is given by:

" =

p f f gD
zz f gD

(12.24a)

and along the base by

" b =

p fbasal f gD
zz f gD

(12.24b)

Assuming constant, , , porosity, and coefficient of internal friction (), the

components of the stress tensor at any point within the wedge are:

xz = f gz sin

" zz = f gD gz cos

(12.25)

csc sec 2 0 2 + 1
xx = f gD gz cos

csc sec 2 0 1

The angles that the principal stresses make with the upper and lower surfaces of
the wedge are:

sin
0 = 0.5 sin 1

sin

"

1 sin b

b = 0.5 sin
b

sin

(12.26a)

The primed and are the surface slope and the basal friction angle, modified by
the influence of pore fluid pressure:

MODERN STRUCTURAL PRACTICE

"277

R. W. ALLMENDINGER 2015-16

SIMPLE MECHANICS

CHAPTER 12

1 f

= tan 1
tan

"

1 b
1
b = tan b

(12.26b)

Out of all this comes an exact and stunningly simple relationship for the critical taper of the wedge (Dahlen, 1990):
+ = b 0

(12.27)

The taper depend on no length parameters and is therefore self-similar. The angles
of the principal stresses, and thus the taper, are dependent only on the material
properties and the pore fluid pressure and thus do not vary throughout the wedge.
Another assumption that we have made is that the entire wedge is on the verge of
failure everywhere.
By making a number of small angle assumptions that , , 0, and b are
all assumed to be much less than 1 we can recover the initial result of Davis et
al. (1983). For subaerial wedges, the approximate expression for the critical taper is:

" +

+ b (1 b )

sin
2 (1 )
1 sin

(12.28)

Step-up Angle of Thrusts

Now that we know the angles that the principal stresses make with the basal
decollement, it is a simple matter to calculate the angle that faults within the wedge
will make with respect to the decollement (Dahlen, 1990). From simple MohrCoulomb theory, the poles to newly formed faults should form at (45 + /2) with
respect to the 1 principal stress direction. Given the above calculations, there are
two possible orientations faults (here given as the angle between the fault plane and
the basal decollement):

MODERN STRUCTURAL PRACTICE

"278

R. W. ALLMENDINGER 2015-16

SIMPLE MECHANICS

CHAPTER 12

" b =

90
b
2

and

90
b =
+ b
2

(12.28)

This gives rise to lower angle synthetic thrust faults (forward verging thrusts) and
higher angle antithetic thrust faults, i.e., steep back-thrusts (Fig. 12.4). The failure
stress on the thrusts within the wedge is:
cos

sin 2 0

(12.29)

sin 2 b

sin 2 0

(12.30)

" = ( f ) gz sin

and on the decollement

" b = ( f ) gz sin

The ratio of these two stresses:

"0

b sin 2 b
=
1

cos

(12.31)

In other words, the decollement must always be weaker than the wedge. This must
obviously be the case or the thrust belt would not move but would break up internally. In fact, if you watch a pile of snow or sand in front of a plow blade, you can
observe an alternation between failure of the wedge and sliding on the base.

45 + /2

Figure 12.4 Step up angles of thrust

faults within the wedge relative to the
basal decollement. Faults within the
wedge are assumed to be Coulomb
shears.

b
b

MODERN STRUCTURAL PRACTICE

"279

R. W. ALLMENDINGER 2015-16

SIMPLE MECHANICS

CHAPTER 12

Holes and Cracks: Some Important Results from Linear Elastic Fracture Mechanics
Linear elastic fracture mechanics has provided some deep insights into the
deformation in the upper part of the Earths crust. The derivations of some of the
fundamental equations involve imaginary numbers, complex variable theory,
and the Cauchy-Riemann equations. These are beyond the scope of this manual
but the interested student may check out the development in Jaeger and Cook
(1976) or McGinty (2015). As you will see, the results of this section are especially
germane to subsurface exploration and drilling for hydrocarbons, geothermal, or
fluid injection.
Circular Holes
The problem of the stresses around a circular hole in a material has been
called the most important single problem in rock mechanics (Jaeger and Cook,
1976, p. 249). Given that we drill circular holes in rocks for a variety of reasons, it is
not hard to see why this is the case! Kirsch (1898) gave the fundamental solution for

11
12

21
a
Pf

MODERN STRUCTURAL PRACTICE

22

"280

Figure 12.5 General two dimensional load on a plate with a circular hole.
We use a Cartesian coordinate system
with X3 parallel to the axis of the hold
and pointing into the page. Stresses
around the hole are specified in polar
coordinate system with the angle
measured up from the horizontal (inset view). The hole has a diameter of a
and the distance from the center of the
hole is specified by r.

R. W. ALLMENDINGER 2015-16

SIMPLE MECHANICS

CHAPTER 12

the case of uniaxial loading but here we will go straight to the general two dimensional loading case.
We assume a far field coordinate system parallel and perpendicular to the
axis of the hole and the far field stresses are defined in that coordinate system (Fig.
12.5). The stresses around the hole are defined in a polar coordinate system as
shown in the inset diagram. The hole has a radius, a, and the stresses are calculated
at a distance, r, from the center of the hole. The radial and tangential normal and
shear stresses are:
" rr =

( 11 + 22 ) 1 a 2 + 1 4 a 2 + 3 a 4 ( 22 11 ) cos 2 +
2

12

sin 2

(12.32a)

2
4

a ( 22 11 )
11 + 22 ) a
(
" =
cos 2 + 12 sin 2 (12.32b)
1+ 1+ 3

2
4

a
a ( 22 11 )
" r = 1+ 2 3
sin 2 + 12 cos 2
r
r
2

(12.32c)

For the special case where 11 = 1 and 22 = 2 and there is fluid pressure, Pf, in the
hole, the 12 term goes zero and the pertinent equations become:
1
2

2
a 1
r 2

a
r

4
a
r

a
r

" rr = ( 1 + 2 )1 + ( 2 1 )1 4 + 3 cos 2 + Pf

"

4
2
a 2 1

a
a
1
= ( 1 + 2 ) 1+ ( 2 1 ) 1+ 3 cos 2 Pf
r
r
2
r 2

(12.33a)

(12.33b)

is commonly referred to as a hoop stress.

Figure 12.6 shows how the stresses vary around the hole for the case of uniaxial loading with 1 = 50 MPa and 2 = 0 MPa. You can see that the tangential or
hoop stress at 0 and 180 is three times greater than the far field 1 value! Furthermore, at 90 and 270 the stress is tensional. This value of 3 (in uniaxial load
conditions) is known as the stress concentration factor. Interestingly enough,

MODERN STRUCTURAL PRACTICE

"281

R. W. ALLMENDINGER 2015-16

SIMPLE MECHANICS

CHAPTER 12

Tangential Stress Variation

Stress at distance r/a

150.0

150.0

100.0
75.0

Stress at = 0

() at a/r = 1

125.0

50.0
25.0

0.0
-25.0
-50.0

90

180

270

360

125.0

rr

100.0
75.0
50.0
25.0
0.0

10

r/a

Figure 12.6 Left: Variation of the hoop stresses () around the edge of a borehole for
the case of uniaxial loading. Right: radial and tangential stress variation with distance
from a borehole as a function of the radius of the borehole.

this factor is independent of the size of the hole: a small hole produces just as much
stress concentration as a large hole. You can see that the stress concentration is very
localized near the hole (Fig. 12.6, right). Within a distance of five borehole radii,
the hoop stress has dropped to within 2% of the regional value.

Hmax
borehole
breakouts

Hmin

Figure 12.7 The location of breakouts and potential tension cracks

around a vertical borehole. The breakouts should parallel the minimum
principal horizontal stress direction.

tension
cracks

MODERN STRUCTURAL PRACTICE

"282

R. W. ALLMENDINGER 2015-16

SIMPLE MECHANICS

CHAPTER 12

122

120

118

Seattle
Monterey
Salt Lake City
San Francisco
3600'

3600'

Bakersfield

3430'

Figure 12.8 The orientation of 1 from

borehole breakouts
near the San Andreas
fault (thin black line) in
central California from
the World Stress Map
(Heidbach et al., 2008).

3430'

Santa Barbara

122

120

118

Projection: Mercator

World Stress Map Rel. 2008

Helmholtz Centre Potsdam
GFZ German Research Centre for Geosciences

Now, lets say you are an oil company systematically drilling holes in a producing area. Every one of those boreholes will have these very large stress concentrations and in some cases the stresses will be high enough to cause the well bore to
deform by spalling off of pieces in the areas of high stress concentration: these are
known as borehole breakouts (Fig. 12.7). For traditional vertical boreholes, breakouts should form by compressive failure in the direction of the least principal horizontal stress and perpendicular to the maximum horizontal principal stress. This
turns out to be one of the best ways to determine the orientations of the stress field
in the plane perpendicular to the borehole. Figure 12.8 shows the orientations of 1
in the vicinity of the San Andreas fault in central California.
Cracks
After Kirschs circular hole solution, the theory evolved to elliptical holes
which in the extreme case become cracks. Cracks are important because, not only
do they determine the ultimate strength of the material, but they also are the site
where subsequent failure via fault tip migration occurs. Some complicated math

MODERN STRUCTURAL PRACTICE

"283

R. W. ALLMENDINGER 2015-16

SIMPLE MECHANICS

CHAPTER 12

X2

Figure 12.9 Stress around a crack tip

for a crack with a half length of a. The
coordinate system is for Irwins (1957)
approximate solution. is the far field
stress.

X1

2a

ensued with each succeeding step forward, but the results are surprisingly simple
and powerful. We skip here this important early development and go straight to Irwins (1957) approximate solution for stresses near a crack tip, that is r a/10 (Fig.
12.9). He calculated that the stresses are:

" 11

a

3
cos 1 sin sin
2
2 2
2 r

(12.34a)

" 22

a

3
cos 1+ sin sin
2
2 2
2 r

(12.34b)

a

3
cos sin cos
2 2
2
2 r

(12.34c)

" 12

The quantity " a is know as the stress intensity factor:

" KI = a

(12.35)

The above analysis is appropriate for Mode I (opening) cracks (Fig. 6.1). For a
Mode II (sliding) crack, the equations are (Pollard and Fletcher, 2005):

MODERN STRUCTURAL PRACTICE

"284

R. W. ALLMENDINGER 2015-16

SIMPLE MECHANICS

CHAPTER 12

" 11 =

3
sin 2 2 + cos 2 cos 2

(12.36a)

K II


3
sin cos cos

2
2
2
2 r

(12.36b)

K II

3
cos 1 sin sin

2 2
2
2 r

(12.36c)

K II
2 r

" 22 =

" 12 =

And, for a Mode III crack:

" 13 =

K III
2 r

sin 2

(12.37a)

" 23 =

K III
2 r

cos 2

(12.37b)

05
0.

04
0.

03
0.

01

02
0.

-0

X
-0.05
-0.04
-0.03
-0.02
-0.01
0.00
0.01
0.02
0.03
0.04
A
0.05
240
230
220
210
200
190
180
170
160
150
140
130
120
110
100
90
80
70
60
50
40
30
20
10
Y

0.

.0
5

0.

0.

0.

0.

0.

0.

-0

-0

-0

-0

-0

X0
-0.05
-0.04
-0.03
-0.02
-0.01
0.00
0.01
0.02
0.03
0.04
A
0.05
1
0
-10
-20
-30
-40
-50
-60
-70
-80
-90
-100
-110
-120
-130
-140
-150
Y
150
140
130
120
110
100
90
80
70
60
50
40
30
20
10

05

-0.05

04

-0.05

03

-0.04

02

-0.04

01

-0.03

00

-0.03

.0
1

-0.02

.0
2

-0.02

.0
3

-0.01

.0
4

-0.01

00

0.00

0.

0.00

.0
1

0.01

.0
2

0.01

-0

0.02

-0

0.02

.0
3

0.03

-0

0.03

-0

0.04

0.04

.0
4

Mode II
0.05

.0
5

Mode I
0.05

Figure 12.10 Two dimensional maps of the 12 = 21 magnitude at the tip of a

horizontal crack (white line), located at (0,0). Width of each diagram is 0.05a.
Color map is different in both but the contour interval is the same. In the Mode I
case, blues and greens are opposite signs.

MODERN STRUCTURAL PRACTICE

"285

R. W. ALLMENDINGER 2015-16

SIMPLE MECHANICS

CHAPTER 12

In all of these cases, we find the square root of distance from the crack tip, r, in the
denominator. Thus, at the crack tip itself, these equations all suggest infinite stress.
Of course, there cant really be infinite stress there, it is just that linear elasticity
doesnt apply near r = 0. Figure 12.10 shows plots of the magnitude of shear stress
on planes perpendicular or parallel to the crack.

Final Thoughts: Simulation vs. Illumination

There are many types of models constructed for different purposes. The
fault-related folding models that we saw in Chapter 10, for example, have the goal
of simulating the geometry, and occasionally the sequence, of deformation. This is
a perfectly reasonable objective and there are many practical reasons why we want
to project the geometry to depth using predominantly kinematic rules: to define the
geometry of a structural trap in a hydrocarbon reservoir, calculate the amount and
distribution of shortening for a palinspastic restoration, evaluate the goodness of
fit, or define the likely fault geometry to assess seismic hazard. The goal in simulation is to reproduce, as faithfully as possible, those parts of the structure that we
cannot see and what they might look like. Some simulations (e.g., the trishear model) can be carried out extremely rapidly, allowing us to test many possible geometries and find a best fit to the data. The problem comes, however, when we assume that the kinematic model explains the structure, because we have not, in
fact, tested whether the model conforms to the well known physical principles described in this chapter, nor whether the boundary or initial conditions are reasonable.
In this Chapter, we have gotten a glimpse of the suite of physical principles,
constitutive equations, and boundary conditions that can be used to illuminate a
structural problem of interest. Full mechanical models commonly do not have, as
an objective, the simulation of an overall structural geometry. We are not trying to
draw a more accurate cross section; instead we are trying to understand why something formed the way that it did. To answer that question, we don't need to reproduce all aspects of the geometry of a structure. What we try to do is distill the
problem until all that remains are its most fundamental elements. By making a
model simpler, we are more likely to be able to isolate, and illuminate, the key features. When we make models more complex, the number of free variables increases
MODERN STRUCTURAL PRACTICE

"286

R. W. ALLMENDINGER 2015-16

SIMPLE MECHANICS

CHAPTER 12

to a point that we can no longer say what is most likely or important. Powerful
computer packages to carry out large scale numerical models especially finite
element and discrete element models run the risk of being so complicated that
one can no longer isolate, and get insight into, the key parameters.
So, both simulation and illumination have their place in structural geology.
The students goal is to learn the wisdom to decide what type of model is likely to
answer the question of interest.

MODERN STRUCTURAL PRACTICE

"287

R. W. ALLMENDINGER 2015-16

SIMPLE MECHANICS

CHAPTER 12

Exercises Chapter 12

1. A simple, well known experiment

in structural geology that relates
directly to the Hubbert and
Rubey analysis was proposed by
the great French fluid mechanics
expert, M. A. Biot, and was
called the beer can experiment4.
(a) Experimental procedure
(i) Take the glass plate provided and place the empty can on one end. Tilt
the plate by raising the end on which the can rests. Record the angle of
the plate at which the can begins to slide down the tilted glass.
(ii) Cover the plate with a film of water (if the plate is dirty, you will have to
clean it with mild detergent first). Repeat step one. Again, record the angle at which the can slides down the plate.
(iii) Now chill the can by placing it in the cooler with dry ice [Safety note: Do
not handle the dry ice with your bare hands or you run a serious risk of
rapid frostbite. Use gloves to place or remove the can from the cooler.
Dry ice or frozen carbon dioxide is much colder than ice made
from water!].
(iv) After two to three minutes, remove the can from the cooler and place the
can on the wetted glass plate with the open end facing up. Tilt until the
can slides and record the angle as before.
(v) Finally, cool the can again, briefly, and place it on the wetted plate, open
end down. Tilt until the can slides and record the angle
(b) Draw a two dimensional free body diagram for the beer can experiment.
(c) Use a force balance to calculate the coefficient of static friction between the
can and the glass for steps (i) to (v) in part (a). How do you explain the result

Editorial note: Biot proposed this experiment back in the 1950's, when beer cans were sturdy affairs. Today's beer
cans, while being ecologically much more acceptable (though much less welcome on college campuses), have wimpy
thin sides that make them unusable for this experiment, so we are reduced to the ignominious fate of having to use a
different type of can!

MODERN STRUCTURAL PRACTICE

"288

R. W. ALLMENDINGER 2015-16

SIMPLE MECHANICS

CHAPTER 12

in step (v)? Derive the appropriate equations that demonstrate what is actually going on in step (v) relative to the other steps.
2. The following questions relate to Figure 12.6, which shows the orientations of
1 in central California.
(a) Make a sketch showing the average orientation of breakouts in the boreholes that were used to calculate the map of stresses. Be sure to include geographic axes and the orientations of the principal stresses.
(b) Hydraulic fracturing experiments at a depth of 1.3 km in the region show
that 1 is 49 MPa with an azimuth of 036 and 3 = 25 MPa on an azimuth of 126. Calculate the hoop and radial stresses around the borehole.
(c) Assume that the breakouts formed by small Coulomb shear fractures that
extend 1.2 borehole radii into the rock (measured from the center of the
borehole). What are reasonable values of cohesion and internal friction for
the rock mass that would explain the formation of the breakouts?
(d) The breakouts reported are commonly from depths of around 3700 m. Assuming that the principal horizontal stresses remain the same and = 2600
kg m3, calculate the values and orientations of 1, 2, and 3. How would
these new values change your answer to part (c)?
3. Figure 12.8 shows the magnitudes of stresses 12 (i.e., xy) with distance from the
tip of a Mode I and Mode II crack. What exactly does this stress mean? Explain how you would go about calculating the maximum shear stress around
the crack tip for the two cases.
4. The following questions apply to the critically tapered wedge theory described
earlier in the Chapter.
(a) Using Equations 12.26 and 12.27, describe what happens as either the static
friction on the base of the wedge, b 0 or the pore fluid pressure ratio on
the base, b 1. The former applies where thrust belt decollements are located in salt horizons; the latter is commonly observed in submarine accretionary prisms.
(b) The figure on the following page shows two cross sections and topographic
profiles across the Subandean fold and thrust belt in Bolivia. The rainfall in
the region of the northern profile is 1600 to 2400 mm/yr whereas along the
southern profile it is 600-800 mm/yr. Discuss the contrasts between these

MODERN STRUCTURAL PRACTICE

"289

R. W. ALLMENDINGER 2015-16

SIMPLE MECHANICS

CHAPTER 12

two sections/profiles in the context of the critically tapered wedge theory

and, where the possible explanations are non-unique, describe what type of
data you would like to collect to resolve any ambiguities (assume that money
is no obstacle!). You may wish to read Dahlen (1990) first.

12

68

64

60

2000 m

1000 m

16

alpha = 0.73

Bolivia

20

beta = 6.3

Roeder 1988

25 km
2000 m

alpha = 0.56
1000 m

beta = 2.2

MODERN STRUCTURAL PRACTICE

"290

Herrail 1989

R. W. ALLMENDINGER 2015-16

Postface
What youve accomplished and where to go next
If you have made it this far, to the end of this lab manual, congratulations!
You have tackled material that takes serious work to learn. If you come from a typical undergraduate geology major background, this manual has introduced you to
some tremendously important concepts and methods that are not typically, or adequately, taught to undergraduates in their first structure course. These concepts include powerful linear algebra methods such as transformations of coordinates, vectors, and tensors; the concepts of principal axes and invariants of tensors; the significant differences between stress, infinitesimal strain and finite strain. We have
learned about material versus spatial coordinates and how important it is in the
analysis of strain as well as in mechanics. We have seen that even the most basic
structural geology calculation has significant uncertainty associated with it and gotten a glimpse of how to determine those uncertainties. Finally, you have seen that
even the simple spreadsheet program is a powerful computing environment that
enables you to solve problems that cannot be solved with traditional graphical
methods. And now, a confession on my part: I hate spreadsheet programs! That is
because I know how much more powerful, faster, and more capable traditional
computer programming is. Hopefully this manual has given you the motivation to
acquire those skills as well.
If you come from an engineering or physics background, you are probably
wondering what all of the fuss is about! Most of the topics that we have touched on
in this manual are treated in the first two years of a typical undergraduate curriculum in those disciplines. Structural geology is, fundamentally, solid mechanics applied to earth materials. Though the methods are the same, there are some significant differences in approach: an engineer might want to know the maximum load
that a beam can bear and how much it will flex so that they can design a bridge to
a particular specification. The structural geologist is more like a forensic scientist:
they come upon the scene when the bridge has already collapsed and is lying in a
heap on the valley floor. Our job is to extract from the chaos of a deformation that

WHERE TO GO NEXT

POSTFACE

has already happened commonly many millions of years ago what the key
factors were and thereby what the Earth is capable of.
As forensic scientists, we have to go to the field which is the scene of the
crime. There is a misconception that field geologists observe and map whereas
quantitative or theoretical structural geologists stay at home and program their
computers. The best structural geologists throughout history have done both: they
go to the field (or lab) and make careful quantitative observations and they know
how to analyze and probe their data quantitatively to extract meaningful conclusions. This is an iterative process. Initial observations may stimulate deeper mechanical analysis which provides a set of observables for testing with further observation, resulting in further refinements of the theory.
Lets say you come from that typical undergraduate geology major background. After reading this lab manual, I flatter myself to think that a few of you
might now be inspired to join the ranks of the best structural geologists. Where
should you turn next on this journey? In no particular order, I suggest the following
studies:
Take at least four semesters of college math, which will take you
through linear algebra and partial differential equations. The
earth is a multivariate system with gradients of properties in all directions. Inverse methods, a branch of linear algebra, are extremely powerful. This is not just the minimum level of math required for structural
geology but in fact the minimum level for anyone wanting to become a
physical scientist or engineer in just about any field.
Learn a real computing environment, whether it be Matlab,
Python, Fortran, Basic, C++, etc. If you cannot program, you have to
wait until someone else writes the program for you and it is unlikely
that, by the time that happens, the problem you wanted to solve will still
be cutting edge. Like the previous bullet, computing is a skill that all scientists and engineers should have.
Take geomechanics or engineering mechanics courses. Because
these courses require dedication and perseverance, they will be more
approachable once you have some idea what the problems are that a
MODERN STRUCTURAL PRACTICE

"292

R. W. ALLMENDINGER 2015-16

WHERE TO GO NEXT

POSTFACE

structural geologist wants to solve. This is a necessary step if, as mentioned in Chapter 12, you want to illuminate and not just simulate a
structure of interest.
Most of the problems that we wish to address in structural geology have
a high degree of uncertainty due to incomplete data, subjective hypotheses and observation, and lack of knowledge about appropriate
boundary conditions, initial values, and constitutive relations. Thus, you
will need to learn some statistics; fortunately much of the mathematical background needed for statistics is the same as that needed for
structural geology itself. Once you have armed yourself with statistical
methods, remember that statistics can only help you eliminate models
or hypotheses; they cannot prove a hypothesis. The best model is
simply one that has not yet been eliminated.
Structural geology provides the basis for a tremendously rewarding career:
we study deformation features in some of the most beautiful and remote places on
Earth. However, we also are expert on problems of extreme societal relevance and
interest: earthquakes, induced seismicity, hydraulic fracturing, surface instabilities,
and so on. The background that I have described above will enable you to pursue
either or both of these branches of our science at the highest level. Good luck!

MODERN STRUCTURAL PRACTICE

"293

R. W. ALLMENDINGER 2015-16

WHERE TO GO NEXT

POSTFACE

MODERN STRUCTURAL PRACTICE

"294

R. W. ALLMENDINGER 2015-16

REFERENCES CITED

References Cited
Albee, H. F., and Cullins, H. L., 1975, Geologic map of the Poker Peak Quadrangle, Bonneville County, Idaho: U.S. Geological Survey Geologic Quadrangle
Map GQ 1260.
Allmendinger, R. W., 1998, Inverse and forward numerical modeling of trishear
fault-propagation folds: Tectonics, v. 17, no. 4, p. 640-656.
Allmendinger, R. W., and Judge, P. A., 2013, Stratigraphic uncertainty and errors
in shortening from balanced sections in the North American Cordillera: Geological Society of America Bulletin, v. 125, no. 9/10, p. 1569-1579, doi:
10.1130/B30871.1.
Allmendinger, R. W., Cardozo, N. C., and Fisher, D., 2012, Structural Geology
Algorithms: Vectors & Tensors: Cambridge, England, Cambridge University
Press, 289 pp.
Berth, D., Choukroune, P., and Jgouzo, P., 1979, Orthogneiss, mylonite and non
coaxial deformation of granites: the example of the South Armorican Shear
Zone: Journal of Structural Geology, v. 1, no. 1, p. 31-42.
Bevington, P. R., and Robinson, D. K., 2003, Data reduction and error analysis for
the physical sciences: New York, McGraw-Hill, 320 pp.
Byerlee, J., 1978, Friction of rocks: Pure and Applied Geophysics, v. 116, p.
615-626.
Cardozo, N., and Allmendinger, R. W., 2013, Spherical projections with
OSXStereonet: Computers & Geosciences, v. 51, no. 0, p. 193 - 205, doi:
10.1016/j.cageo.2012.07.021.
Chamberlin, R.T., 1910, The Appalachian folds of central Pennsylvania.: Journal
of Geology, v. 18, p. 228251.
Chamberlin, R.T., 1919, The building of the Colorado Rockies: Journal of Geology, v. 27, p. 145164.
Chapple, W. M., 1978, Mechanics of thin-skinned fold and thrust belts: Geological
Society of America Bulletin, v. 89, p. 1189-1198.

MODERN STRUCTURAL PRACTICE

"295

R. W. ALLMENDINGER 2015-16

REFERENCES CITED

Charlesworth, H. A. K., Langenberg, C. W., and Ramsden, J., 1976, Determining

axes, axial planes, and sections of macroscopic folds using computer-based
methods: Canadian Journal of Earth Science, v. 13, p. 54-65.
Dahlen, F. A., 1990, Critical taper model of fold-and-thrust belts and accretionary
wedges: Annual Reviews of Earth and Planetary Science, v. 18, p. 55-99.
Davis, D., Suppe, J., and Dahlen, F. A., 1983, Mechanics of fold-and-thrust belts
and accretionary wedges: Journal of Geophysical Research, v. 88, p.
1153-1172.
Elliott, D., 1976, The energy balance and deformation mechanisms of thrust
sheets: Philosophical Transactions Royal Astronomical Society Philosophical
Transactions Royal Astronomical Society, v. A-283, p. 289-312.
Epard, J.L., and Groshong, R.H., Jr, 1993, Excess area and depth to detachment:
American Association of Petroleum Geologists Bulletin, v. 77, p. 12911302.
Erslev, E. A., 1991, Trishear fault-propagation folding: Geology, v. 19, no. 6, p. 617620.
Fossen, H., and Tikoff, B., 1993, The deformation matrix for simultaneous simple
shearing, pure shearing, and volume change, and its application to transpression-transtension tectonics: Journal of Structural Geology, v. 15, no. Nos.
3-5, p. 413-422.
Fry, N., 1979, Random point distributions and strain measurement in rocks:
Tectonophysics, v. 60, no. 1, p. 89-105.
Hardy, S., and Allmendinger, R. W., 2011, Trishear: A review of kinematics, mechanics, and applications, in McClay, K. et al., eds., Thrust fault related folding: Tulsa, Oklahoma, American Association of Petroleum Geologists,
Memoir 94, p. 95-119.
Heidbach, O., Tingay, M., Barth, A., Reinecker, J., Kurfe, D. and Mller, B.,
2008, The World Stress Map database release 2008 doi:10.1594/GFZ.WSM.Rel2008.
Hill, R., 1950, The mathematical theory of plasticity: London, Oxford University
Press, 355 pp.

MODERN STRUCTURAL PRACTICE

"296

R. W. ALLMENDINGER 2015-16

REFERENCES CITED

Holt, W. E., Shen-Tu, B., Haines, J., and Jackson, J., 2000, On the determination
of self-consistent strain rate fields within zones of distribution continental deformation, in Richards, M. A. et al., eds., The history and dynamics of global plate motions: Washington, D.C., American Geophysical Union, Geophysical Monograph, p. 113-141.
Hubbert, M. K., 1951, Mechanical basis for certain familiar geologic structures:
Geological Society of America Bulletin, v. 62, p. 355-372.
Hubbert, M. K., and Rubey, W. W., 1959, Role of fluid pressure in mechanics of
overthrust faulting, I. Mechanics of fluid-filled porous solids and its applications to overthrust faulting: Geological Society of America Bulletin, v.
70, p. 115-166.
Irwin, G.R., 1957, Analysis of stresses and strains near the end of a crack
traversing a plate: Journal of Applied Mechanics, v. 24, p. 361-364.
Jackson, M. D., and Pollard, D. D., 1988, The laccolith-stock controversy: New
results from the southern Henry Mountains, Utah: Geological Society of
America bulletin, v. 100, no. 1, p. 117-139
Jaeger, J. C., and Cook, N. G. W., 1976, Fundamentals of Rock Mechanics: London, Chapman and Hall, 593 pp.
Johnson, A. M., 1970, Physical processes in geology; a method for interpretation of
natural phenomena; intrusions in igneous rocks, fractures, and folds, flow of
debris and ice: San Francisco, Freeman, Cooper.
Judge, P. A., and Allmendinger, R. W., 2011, Assessing uncertainties in balanced
cross sections: Journal of Structural Geology, v. 33, p. 458-467, doi:
10.1016/j.jsg.2011.01.006.
Kirsch, E.G.,1898, Die Theorie der Elastizitt und die Bedrfnisse der Festigkeitslehre: Zeitschrift des Vereines deutscher Ingenieure, v. 42, p. 797-807.
Klotz, J., Angermann, D., Michel, G., Porth, R., Reigber, C., Reinking, J.,
Viramonte, J., Perdomo, R., Rios, V., Barrientos, S., Barriga, R., and
Cifuentes, O., 1999, GPS-derived deformation of the Central Andes
including the 1995 Antofagasta Mw = 8.0 Earthquake: Pure and Applied
Geophysics, v. 154, p. 709-730.
MODERN STRUCTURAL PRACTICE

"297

R. W. ALLMENDINGER 2015-16

REFERENCES CITED

Kostrov, V. V., 1974, Seismic moment and energy of earthquakes, and seismic flow
of rock: Izv. Acad. Sci. USSR Phys. Solid Earth Izv. Acad. Sci. USSR Phys.
Solid Earth, v. 1, p. 23-44.
Malvern, L. E., 1969, Introduction to the mechanics of a continuous medium: Englewood Cliffs, N.J., Prentice-Hall, Inc., 713 pp.
McGinty, B., 2015, Fracture mechanics: http://www.fracturemechanics.org/fm/
index.html, (2015).
Menke, W., 1984, Geophysical Data Analysis: Discrete Inverse Theory: Orlando,
Academic Press, 260 pp.
Middleton, G. V., and Wilcock, P. R., 1994, Mechanics in the earth and environmental sciences: Cambridge, UK, Cambridge University Press, 459 pp.
Mitra, S., 1993, Geometry and kinematic evolution of inversion structures:
American Association of Petroleum Geologists Bulletin, v. 77, no. 7, p.
1159-1191.
Molnar, P., 1983, Average regional strain due to slip on numerous faults of different
orientations: Journal of Geophysical Research, v. 88, p. 6430-6432.
Nye, J. F., 1985, Physical properties of crystals; their representation by tensors and
matrices: Oxford, England, Oxford University Press, 329 pp.
Petit, J.-P., 1987, Criteria for the sense of movement on fault surfaces in brittle
rocks: Journal of Structural Geology, v. 9, p. 597-608.
Pollard, D. D., and Fletcher, R. C., 2005, Fundamentals of Structural Geology:
Cambridge, England, Cambridge University Press, 500 pp.
Ramsay, J. G., 1967, Folding and fracturing of rocks: New York, McGraw-Hill, 568
pp.
Ramsay, J. G., and Graham, R. H., 1970, Strain variation in shear belts: Canadian
Journal of Earth Sciences, v. 7, p. 786-813.
Ramsay, J. G., and Huber, M. I., 1983, The techniques of modern structural
geology, volume 1: Strain analysis: London, Academic Press, 307 pp.

MODERN STRUCTURAL PRACTICE

"298

R. W. ALLMENDINGER 2015-16

REFERENCES CITED

Rich, J. L., 1934, Mechanics of low-angle overthrust faulting as illustrated by

Cumberland thrust block, Virginia, Kentucky, and Tennessee: American Association of Petroleum Geologists Bulletin, v. 18, p. 1584-1596.
Sibson, R. H., 1974, Frictional constraints on thrust, wrench and normal faults:
Nature, v. 249, p. 542-544.
Sibson, R. H., 1985, A note on fault reactivation: Journal of Structural Geology, v.
7, p. 751-754.
Suppe, J., 1983, Geometry and kinematics of fault-bend folding: American Journal
of Science, v. 283, no. 7, p. 684-721.
Suppe, J., Chou, G. T., and Hook, S. C., 1992, Rates of folding and faulting determined from growth strata, in McClay, K. R., ed., Thrust tectonics: London, Chapman & Hall, p. 105-121.
Suppe, J., and Medwedeff, D., 1990, Geometry and kinematics of fault-propagation folding: Eclogae Geologicae Helvetiae, v. 83, no. 3, p. 409-454.
Taylor, J. R., 1997, An introduction to error analysis: The study of uncertainties in
physical measurements: Sausalito, California, University Science Books, 327
pp.
Tikoff, B., and Fossen, H., 1993, Simultaneous pure and simple shear: the unifying
deformation matrix: Tectonophysics, v. 217, no. 3, p. 267-283.
Turcotte, D. L., and Schubert, G., 1982, Geodynamics: Applications of Continuum
Physics to Geological Problems: New York, John Wiley & Sons, 450 pp.
Vollmer, F., 2015, EllipseFit 3.2.0: http://www.frederickvollmer.com/ellipsefit/.
Williams, G., and Chapman, T., 1983, Strains developed in the hangingwalls of
thrusts due to their slip/propagation rate; a dislocation model: Journal of
Structural Geology, v. 5, no. 6, p. 563-571.
Zehnder, A. T., and Allmendinger, R. W., 2000, Velocity field for the trishear
model: Journal of Structural Geology, v. 22, p. 1009-1014.

MODERN STRUCTURAL PRACTICE

"299

R. W. ALLMENDINGER 2015-16

REFERENCES CITED

MODERN STRUCTURAL PRACTICE

"300

R. W. ALLMENDINGER 2015-16

QUICK REFERENCE

Quick Reference
This section contains a listing of formulae that you are likely to use (a lot). They are
listed in the order that they occur in the text with their original equation numbers.

"

180
= 57.2958;

3
180 = ; 270= ; 360=2
2

1 radian =

90 = ;
2

(1.1)

Vector magnitude:

" v = v12 + v22 + v32

(2.2)

" v = cos cos cos

(2.3)

" u i v = v i u = u v cos = u1v1 + u2 v2 + u3v3

(2.6)

" = cos1 ( v1u1 + v2u2 + v3u3 )

(2.7)

Unit vector and direction cosines:

Dot product:

Cross product:

"

v u = u v = v u sin
= ( v2u3 v3u2 ) ,

( v3u1 v1u3 ), ( v1u2 v2u1 )

(2.8)

Direction cosines of a line or pole from trend and plunge or strike and dip in a
NED coordinate system:

MODERN STRUCTURAL PRACTICE

"301

R. W. ALLMENDINGER 2015-16

QUICK REFERENCE

Table 2.1-1
Axis

Direction Cosine

Lines

Poles to Planes
(strike & dip using RHR)

North

cos

cos(trend)cos(plunge)

sin(strike)sin(dip)

East

cos

sin(trend)cos(plunge)

cos(strike)sin(dip)

Down

cos

sin(plunge)

cos(dip)

Trend and plunge from direction cosines:

" plunge = sin 1 (cos )

cos
cos

" trend = tan 1

" trend = 180 + tan 1

if

cos
cos

cos > 0

if

(2.11a)

cos < 0

(2.11b)

Special case of cos = 0:

" trend = 90 if

( cos = 0

and cos 0 )

(2.11c)

" trend = 270 if

( cos = 0

and cos < 0 )

(2.11c)

Mean vector standard deviation:

1

N R 1 N 1
" cos p = 1
1
R 1 p

(2.12)

Transformation matrix:

a11 a12

" aij = a21 a22

a31 a32

MODERN STRUCTURAL PRACTICE

a13 cos11 cos12

a23 = cos 21 cos 22
a33 cos 31 cos 32

"302

cos13

cos 23
cos 33

(4.2)

R. W. ALLMENDINGER 2015-16

QUICK REFERENCE

Vector transformation, new coordinates in terms of old:

v1 = a11v1 + a12 v2 + a13v3
" v2 = a21v1 + a22 v2 + a23v3
v3 = a31v1 + a32 v2 + a33v3

(4.4)

Vector transformation, old coordinates in terms of new:

v1 = a11v1 + a21v2 + a31v3
" v2 = a12 v1 + a22 v2 + a32 v3
v3 = a13v1 + a23v2 + a33v3

(4.5)

Stratigraphic thickness:
" t 3 b3 = ( sin ( strike) sin ( dip ))( t N bN ) ( cos ( strike) sin ( dip ))( t E bE ) + cos ( dip )( bU tU ) (4.10)

Down plunge projection transformation matrix (X2 is the fold axis):

sin(TFA)
cos(TFA)
0

" bij = cos(TFA)cos(PFA) sin(TFA)cos(PFA) sin(PFA)

cos(TFA)sin(PFA) sin(TFA)sin(PFA) cos(PFA)

(4.12)

Rotation transformation matrix:

a11 = cos + cos 2 (1 cos )

a12 = cos sin + cos cos (1 cos )

a13 = cos sin + cos cos (1 cos )

a21 = cos sin + cos cos (1 cos )

" a22 = cos + cos2 (1 cos )

(4.14)

a23 = cos sin + cos cos (1 cos )

a31 = cos sin + cos cos (1 cos )
a32 = cos sin + cos cos (1 cos )
a33 = cos + cos 2 (1 cos )

MODERN STRUCTURAL PRACTICE

"303

R. W. ALLMENDINGER 2015-16

QUICK REFERENCE

Summation convention (i is the free suffix and j the dummy suffix):

3

" vi = aij v j = aij v j = ai1v1 + ai 2 v2 + ai 3v3

(5.5)

j=1

Tensor (dyad) product:

" T = u v = uT v

uv uv
1 2
11
Tij = ui v j = u2 v1 u2 v2

u3v1 u3v2

or

u1v3

u 2 v3

u 3 v3

(5.7)

Invariants of a tensor:

" 3 I 2 II III = 0

(5.9)

I = T11 + T22 + T33 = T1 + T2 + T3

(T T
" II =

ij ij

I2

) = (T T

1 2

+ T2T3 + T3T1 )

(5.10)

III = det T = Tij = T1T2T3

Tensor transformation, new in terms of old:

" Tij = aik a jlTkl

(5.11)

Orientation tensor:

"T=

cos
cos
cos

[i ]

[i ]

cos [i ]

[i ]

cos [i ]

cos

[i ]

cos
cos

cos [i ]

[i ]

[i ]

cos [i ]

cos [i ]

cos

cos

[i ] [i ]
cos2 [i ]

cos

[i ]

(5.22)

Cauchys Law:

MODERN STRUCTURAL PRACTICE

"304

R. W. ALLMENDINGER 2015-16

QUICK REFERENCE

" pi = i j j

(5.23a)

p1 = 111 + 12 2 + 13 3
" p2 = 211 + 22 2 + 23 3
p3 = 311 + 32 2 + 33 3

(5.23b)

Mohrs Circle for stress:

1 + 3 1 3
+
cos 2
2 2

" n =

(5.25a)

1 3
sin 2
2

" = s =

(5.25b)

Mean and deviatoric stress:

" m =

11 + 22 + 33 1 + 2 + 3
=

3
3

m 0
0

" ij = 0 m 0
0
0 m

(5.26)

11 m
12
13

21
22 m
23
+

31
32
33 m

(5.27)

Slip from piercing points, where p1 is the position of the line, v is the direction
cosines of the line, p2 is a point on the plane, n the direction cosines of the pole to
the plane:

"u =

n i [ p2 p1] n1 ( p21 p11 ) + n2 ( p2 2 p12 ) + n3 ( p2 3 p13 )

=

n i v
n1v1 + n2 v2 + n3v3

" p1 = p11 + uv1;

" slip =

p2 = p12 + uv2 ;

p3 = p13 + uv3

( p(hw)1 p( fw)1 )2 + ( p(hw)2 p( fw)2 )2 ( p(hw)3 p( fw)3 )2

(6.4)
(6.5)
(6.6)

Coulomb failure envelope:

MODERN STRUCTURAL PRACTICE

" = s = So + n

"305

(6.7)

R. W. ALLMENDINGER 2015-16

QUICK REFERENCE

In terms of principal stresses:

" 1 = Co + K 3

K=

where

1+ sin
1 sin

and

Co = 2So K

(6.8)

Failure envelope for frictional slip on pre-existing planes:

" = s n

(6.9)

Coulomb failure with pore fluid pressure:

" = s = So + ( n Pf ) = So + n*

(6.12)

Pore fluid pressure ratio:

" =

Pf

where

Plithostatic

Plithostatic = gz

(6.13)

Principal stress ratio:

"R=

2 1 c13c23
=

3 1 c12 c22

(6.22)

Stretch, S, and inverse stretch, s:

"S =

f x
=

i X

(7.2a)

"s =

i X
=

f x

(7.2b)

Extension, E or e, can be defined as the change in length over the initial length (an
initial state frame of reference) or the final length (a final state frame of reference):

MODERN STRUCTURAL PRACTICE

"E =

f i u
=
=

i
i X

"306

(7.5a)

R. W. ALLMENDINGER 2015-16

QUICK REFERENCE

"e=

f i u
=
=

f
f x

(7.5b)

Shear strain, , and the angular shear, , as:

" =

u
Y

and

u
= tan 1

Y

(7.7)

Displacement gradient tensor (asymmetric):

" ui =

ui
X j = Eij X j
X j

or

" dui = Eij dX j

dui =

ui
dX j = Eij dX j
X j

ui = ti + Eij X j

(7.10)
(7.11)

Deformation gradient tensor (asymmetric):

" xi =

xi
X j = Dij X j
X j

" dxi = Dij dX j

(7.12)

xi = ci + Dij X j

(7.13)

Infinitesimal strain and rotation from the displacement gradient tensor:

" Eij = ij + ij

" ij =

1
Eij + E ji
2

and

(7.19)

" ij =

1
Eij E ji
2

(7.20)

Rotation vector, ri, as follows:

" r1 =

( 23 32 )
,
2

r2 =

( 13 + 31 )
,
2

and

r3 =

( 12 21 )

2

(7.21)

Mohrs Circle for infinitesimal strain:

MODERN STRUCTURAL PRACTICE

"307

R. W. ALLMENDINGER 2015-16

QUICK REFERENCE

11 =

"

(1 + 3 ) + (1 3 ) cos 2
2

( )
13 = = 1 3 sin 2
2
2

(7.28)

P- and T-axes for faults and earthquakes, n is the pole, s the slip unit vectors. k = 1
for normal faults and k = 1 for reverse faults:

" Pi =

ni + ksi
2

and

Ti =

ni ksi

2

(7.34)

Seismic moment, Mo, and geometric moment, Mg:

" M o = As

and

M g = As

(7.35)

Asymmetric moment tensor as the dyad product of the slip and normal vectors:

" M ij =

n faults

M g ui n j

(7.36)

Kostrovs symmetric moment tensor, ij:

" Eij = ij + ij =

M (u n
g

+ u j ni

2V

) + M (u n
g

u j ni

2V

(7.38)

Lagrangian strain tensor (initial or material state):

1 u

u u

" Lij = i + j + k k = Eij + E ji + Eki Ekj

2 X j Xi Xi X j 2

(8.1)

Eulerian finite strain tensor, " Lij , is the same as the Lagrangian strain tensor but referenced to the final (spatial) state:

1 u

u u

" Lij = i + j + k k = eij + e ji + eki ekj

2 x j xi xi x j 2

(8.5)

Green deformation tensor (initial or material state):

MODERN STRUCTURAL PRACTICE

"308

R. W. ALLMENDINGER 2015-16

QUICK REFERENCE

xk xk
= Dki Dkj
Xi X j

" Cij =

(8.6)

Cauchy deformation tensor, " Cij , is similar to Equation (8.6), but referenced to the
final or spatial state:

" Cij =

Xk Xk
= dki dkj
xi x j

(8.8)

Mohr circle for strain in the deformed (final or spatial) state:

" C11 =

" =

(C

+ C3 )
2

(C

C3 )
2

( 1 + 3 ) + ( 1 3 ) cos 2
2

" C13 =

and

cos 2

and

" =

(C

C3 )
2

sin 2

( 1 3 )
=
sin 2

(8.15)
(8.16)

Calculating the Cauchy deformation tensor from three deformed lines:

" b =
c

dx1

f

dx1

f

dx1

f

dx1dx2
2 2

dx1dx2
2 2

dx1dx2
2 2

dx2

f

dx2

f

dx2

f

C11

C12

C22

(8.22)

Poissons ratio:

w f wi
w
e
i
" = t =

e
f i

i

(9.2)

Elastic moduli (G = shear modulus, E = Youngs modulus, K = bulk modulus):

MODERN STRUCTURAL PRACTICE

"309

R. W. ALLMENDINGER 2015-16

QUICK REFERENCE

"G =

3K (1 2 )
E
=

2 (1+ )
2 (1+ )

(9.3)

Lithostatic load or pressure:

z

" Plith = g dz gz

(9.4)

Minimum stress ratio for reactivation:

1*
=
*
3 min

" min =

(1+ 2 ) +

(9.8)

Table 9.1: Frictional Strength of the Crust-1

Vertical Stress

Tectonic Environment

Minimum Differential Stress

thrust faulting

" 1

3* ( min 1) gz (1 )

normal faulting

" 1

3*

( min 1) gz (1 )

1* 3*

strike-slip faulting
"

min

( min 1) gz (1 )
( min + 1) + 1

where =

2 3
1 3

Power law creep:

" e! = Co ( 1 3 ) exp
RT
n

(9.10)

Shear zone displacement from foliation orientations ( = shear strain):

MODERN STRUCTURAL PRACTICE

" tan 2 =

"310

(9.12)

R. W. ALLMENDINGER 2015-16

QUICK REFERENCE

" d = dy =

2
dy
tan 2

(9.13)

General 2D shear:
( S1 S3 )
S
ln 1 S
3

S1
" Dij =

( )
S3

(9.16)

Fault-bend fold equations (Suppe, 1983):

sin ( ) sin ( 2 ) sin

cos ( ) sin ( 2 ) sin sin

" = tan-1

" = + (180 2 ) = +

(10.6a)
(10.6b)

Fault-propagation fold equations:

sin * sin ( * 1 )

" sin 2 =

*
sin * + sin 1 sin ( 1 )
) sin ( 2 * )
( 1
1

" = 90 + * 1; 2 = 180 2 * + 1; and b = 2 ( * ) .

(10.7)

Trishear fault-propagation folding (linear velocity field):

!
" (x, y) = 0
2

2
y

m y

+
1
i
+

j


mx
mx
1+
s

(10.11)

Simple shear deformation of the hanging wall in similar folding:

MODERN STRUCTURAL PRACTICE

" sn =

h

cos n (1+ tan n tan )

"311

(10.12)

R. W. ALLMENDINGER 2015-16

QUICK REFERENCE

Incompressibility:

v1 v2 v3
+
+
= 0
x1 x2 x3

(10.16)

1 n1
( xi yi+1 xi+1yi )
2 i=0

(10.18)

" div v =

Area of a polygon with n vertices:

"A=

P-wave velocity:

" velocity = V =

(11.1)

Stacking velocity and normal move out time:

" t NMO = t x t o = t +
2
o

x2
2
Vstacking

to

(11.3b)

Dix Equation:

" Vi(1,2) =

2
Vst22t 2 Vst1
t1

t 2 t1

(11.4)

The continuity equation:

"

( vi )
d
+
= 0
dt
xi

(12.1)

Law of conservation of momentum (mv):

"F=

d ( mv )

dt

(12.3a)

Cauchys First Law of Motion:

MODERN STRUCTURAL PRACTICE

"312

R. W. ALLMENDINGER 2015-16

QUICK REFERENCE

"

dvi ji
=
+ gi
dt
x j

(12.4)

Cauchys Second Law of Motion:

" ij = ji

i, j = 1 to 3

for

Strain compatibility, specified by St.-Venants equations:

2 ij
2 jl
2 kl
2 ik
"
+

= 0
X k Xl Xi X j X j Xl Xi X k

(12.7)

Lams constants:

" ij =

(1+ )
E

ij

ij kk
E

(12.11a)

" ij = 2G ij + ij I

(12.11b)

Where I is the first invariant of the infinitesimal strain tensor and ij is the Kronecker delta. The Lam constant, , is related to the other elastic moduli by:

E
2
and = K G
3
(1+ )(1 2 )

" =

(12.12)

Kirschs (1898) solution for stresses around a circular hole:

" rr =

( 11 + 22 ) 1 a 2 + 1 4 a 2 + 3 a 4 ( 22 11 ) cos 2 +
2

12

sin 2

(12.32a)

2
4

a ( 22 11 )
11 + 22 ) a
(
" =
cos 2 + 12 sin 2 (12.32b)
1+ 1+ 3

( 22 11 )

" r = 1+ 2 3
r
r

MODERN STRUCTURAL PRACTICE

"313

sin 2 + 12 cos 2

(12.32c)

R. W. ALLMENDINGER 2015-16

QUICK REFERENCE

For the special case where 11 = 1 and 22 = 2 and there is fluid pressure, Pf, in the
hole:
2
4
2
a 2 1

a
a
a
1
" rr = ( 1 + 2 )1 + ( 2 1 )1 4 + 3 cos 2 + Pf
r
r
r
2
r 2

1
2

2
a 1
r 2

4
a
r

a
r

" = ( 1 + 2 )1+ ( 2 1 )1+ 3 cos 2 Pf

(12.33a)

(12.33b)

Irwins (1959) solutions for stresses near a crack tip:

" 11

a

3
cos 1 sin sin
2
2 2
2 r

(12.34a)

" 22

a

3
cos 1+ sin sin
2
2 2
2 r

(12.34b)

a

3
cos sin cos

2
2
2
2 r

(12.34c)

" 12

Stress intensity factor:

MODERN STRUCTURAL PRACTICE

" KI = a

"314

(12.35)

R. W. ALLMENDINGER 2015-16