I.

Katz

RADAR REFLECTIVITY

T here are many places in which a knowledge of

the electrical reflectivity of natural surfaces is
used-radio communications, missile guidance, and
have been involved in research on scattering for
applications in radar map-matching and in missile
guidance. A need is now developing to expand
radar detection and tracking of targets, among our emphasis to encompass some broader aspects
others. The new emphasis on space research brings of the problem.
with it a need for additional data on the reflec- It is our intent to present a simplified treatment
tivity of planetary surfaces. of many of the experimental results available to
There are distinct similarities between reflection date and to suggest some of the newer trends. To
phenomena in the microwave region and at optical, understand the scattering phenomenon more
infrared, and the longer radio wavelengths. In completely, one needs reflectivity as a function of
addition, certain common features are exhibited in the angle made by the field vector with reference
the scattering of sound and the scattering of radio to some fixed angle for both the transmitter and
waves from irregular surfaces. While the subject of receiver over the, entire possible range of angles.
scattering from rough surfaces has been studied
Sm.ooth Surface Scattering
since the publication of Rayleigh's classical work
on reflections of acoustic waves, l there remains Scattering from a smooth, partially conducting
much to be done before a complete understanding surface has been worked out to a high degree of
is available. There are current efforts to solve the approximation for linear and circular polariza-
problem theoretically, and with the mathematical tions; these derivations may be found in the litera-
tools developed recently in noise theory large ture. 2 Specular reflection takes place, with the
strides can be expected soon. At the same time, in angle of incidence equal to the angle of reflection.
order to answer certain practical questions, much If the complex index of refraction of the surface is
experimental work on scattering is being done. We known, the amplitude reflection coefficient p and

2 H . R. Reed a nd C . M. Russell , Ultra Hi gh Frequency P ropagation,

1J . W. S. Rayleigh , The Theory oj S ound, Vol. 2, D over P ublications,
New York, 1945. J ohn Wiley a nd Sons, New Y ork , 1953.

10 APL Technical Digest

 Radar reflectivity of rough surfaces is being studied
at AP L with a view toward improved radar and missile
guidance performance. This paper reviews the present
state of our experimental knowledge of the reflectivity
of smooth and rough land and water surfaces.

of the
EAR TH'S SURFA CE

the change of phase on reflection cJ> can be com- For propagation above a smooth surface, the
puted for any incident wavelength and polariza- signal at any point in space can be computed in
tion. As an illustration, we show in Fig. 1 the the following way. The total signal strength T for
reflection coefficients for horizontal and vertical linear horizontal or vertical polarization is given by
polarization, Pv and Ph respectively, for C-band
(5 cm) radiation incident upon a smooth sea
T = D + Rei 0, (1 )
surface. where D is the strength of the direct ray, that of
When considering circular polariza tion we the reflected ray is R, and D is the phase-angle
must take the sense of receiver polarization into difference between the direct and reflected rays
account. A right-hand circularly polarized wave because of path-length differences. The amplitudes
normally incident on a flat conducting plate be- of the reflected ray for vertical and horizontal
comes left-hand · circularly polarized on reflection. polarization, respectively, are given by
This same wave incident on the same surface at
small grazing angles retains its sense of polariza- (2)
tion; the angle at which this transition occurs is and
called the Brewster angle. Thus, when we speak
of the reflectivity of a circularly polarized wave we where Vr and Hr are the voltages of the incident
must state the sense of polarization; this is done wave obtained from the transmitter antenna pat-
by subscript, Pcs meaning reflection coefficient tern in the two cases. Since a circularly polarized
"circular, same," and Pco meaning "circular, wave is composed of a vertical and a horizontal
opposite." To detect a circularly polarized wave component, we write for the total signal resulting
in free space we use identical antennas at the from circularly-polarized illumination,
transmitter and receiver. If, on the other hand,
we wish to detect a circularly polarized signal T -- Vd +2 JHd + 2" [V.r pv eicJ>" +·H
1 io
J r Ph eicJ>h] e, (3)
reflected from a flat conducting surface, we use
antennas polarized with opposite sense. where Vd and Hd are the vertical and horizontal

January - February 1963 11

 I .o --------~-------r------_.------_,--------~------Ir------~------~------~

0.8 ~--t---t---±:::::::;;:;;;;;;;;+----_t-=::t===~:::::=-..,--.,

I-
Z
ill

EE 0.6
oo
z LI NEAR POLARIZATION:
Q
I- VERTICAL - p,
~ 0.4 ~-I-.'-~~---I--+----_____1f_---___t----__r___ HORIZONTAL - Ph
Li
ill CIRCULAR POLARIZATION:
~
OPPOSITE - p co
SAM E- p c s

0.2 ~-l---J~------~~----+-------4-------~-------r-------t-------t------~

--- ---
°0L--------1~0--------2~0---------
30L--------4~0--------~
50L-------~60~------~~-~-~-~~~0~------~
90
GRAZING ANGLE (d egrees)

Fig. I-SJDooth·sea reflection coefficient for linear and circular polarizations.

signal voltages that would be received if the surface and

were absent, Vr and Hr are again the corresponding
. - 1 {pv (sm(j>v
. Ph Sm(j>h
. ) }•
voltages incident on the surface, and j and (j>co = sm 2 PCB - Pv (5c)
i are the space and time quadrature vectors,
respectively. One function of the circularly Values of Pes and Peo are plotted in Fig. 1. The re-
polarized antenna is to remove the space quadra- flectivity of a circularly polarized wave is unity for
ture relationship between the vertical and hori- low grazing angles and is zero for normal inci-
zontal components. In Eq. (3) the transmitting dence when the receiver is polarized in the same
and receiving antennas have the same sense of sense as the transmitter. For a receiver polarized
circular polarization. Since, for circular polariza- opposite to the transmitter, the reverse is true; i.e.,
tion, Vr = H r , and since the space vector disap- at small depression angles the reflection coefficient
pears from the output, we have is zero and it increases to 0.8 at if; = 90°.
(4a) Thus, the problem of forward scattering for
where radio waves from smooth surfaces may be con-
sidered solved. If we know the surface electrical
PCB = UP v2 + Ph2 + 2P vPh cos ((j>h - (j> v)] 1/2, (4b)
properties, we can predict the" signal strength at
and any point in space to the degree of precision re-

</>" ~ Sin-l{~: (Sin</>. + ~Sin</>k)}. (4c)

quired for our purposes.

Forward Scattering Over Rough

If the receIVmg antenna is polarized to receive Water Surfaces
polarization of opposite sense to that transmitted,
Real surfaces are, in general, not smooth. Of
D = 0, course, how a given surface appears to the radio
and wave depends on the relationship of the size of the
T=R ![VrP vei ~ v - jllrPhe i ~h] surface irregularities to the electromagnetic wave-
(5a) length. Rayleigh proposed a quantitative criterion
v:.pcoei~co,
of roughness, suggesting that if the phase difference
where between two adjacent rays is small, then the sur-
Peo = UP v2 + Ph2 - 2P vPh cos ((j>h - (j> v) ]1/2, (5b) face seems smooth. Quantitatively, the surface is

12 APL Technical Digest

 1.0 ~--"'----r------,------,
considered to be smooth if h sin1f < }./ 8, where h
is a measure of the peak-to-trough irregularity
height, 1f is the grazing angle, and}. is the wave- s: 0.8
length. Q)

.Eo
More recently a model was ro osed 3 that n:: 0.6
predicts certain features of forward scattering 3
over water. The si nal is considered to consist 2
w O.4 Ll-----~--~--------~~1_----------_1
of a constant direct si nal, a coherentl reflected >
i=
signal (in the sense of the smooth surface discussed :s
a ove w ose am litude and p'hase are deter- ~ 0 .2 1--\-------~~--~~------+--~~----__1
mined b geometry and sea conditions, and a
fluctuating reflected signal, random in amplitude 0 L---~~~~ ____~~__~__________~
and hase. The amplitudes of the coherent and o 1.0 . 2.0 3.0
FREQU ENCY (cps)
incoherent com onents de end on the roughness
parameter h1f/ }., where h is now the root-tnean- Fig.3- Spectra of radio waves, showing broaden-
square wave height. In Fig. 2 are shown values of ing with increasing surface roughness.
t e co erent an incoherent parts of the reflected
In addition to the mean signal value and its
signal as found by experiments at the Golden Gate
variance, we can predict the time spectrum of the
and verified later in the Gulf of Mexico;4 in this
amplitude variation of the received signal in point-
figure 1f is in milliradians. Both C and (1, the co-
to-point transmission in the following way. It was
erent and incoherent terms respectively, are
found in experiments involving small grazing
normalized to the direct ray D and the smooth-
angles 5 that the signal received at a fixed point in
sea reflection coefficient p . Notice that the coherent
space from a fixed transmitter varied in its fre-
term drops gradually from unity to near zero in
quency content as a function of the vertical motion
the range of roughness values between zero and
of the surface and the surface roughness. As illus-
about 300, while the incoherent term rises much
tration, Fig. 3 shows how the frequency spectrum
more sharply to a value of about 0.3, then flattens
of the amplitude variation of the received radio
and tends to drop somewhat at states of higher
signal broadens with increased roughne~s. The
roughness. Whether the drop in (11Dp is real and
signal at an hi/; I}. = 25 has most of its energy in
continues for increasing roughness above that
frequencies well below 0.5 cps, whereas for h1f/ }. =
shown in the graph is a subject for future experi-
590, there is still considerable energy above 2 cps.
ment.
Curves are available that show a quantitative
By using the above with curves of smooth sea
relationship between signal-strength spectra, the
reflection coefficient, we can predict signal strength
frequency (in cps) of the maximum of the height-
and its fluctuation (variance) at points in space
versus-time spectrum of the ocean waves, and
above the surface if the sea state is known.
ocean roughness. If we form the ratio of the fre-
_~ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _1 0.5
1.0 quency at which the radio spectrum has dropped
Yt
to, say, 0 power to the frequency of the peak in
c5 0.8 the ocean spectrum prevailing at the same
U :2
time, and plot it against ocean roughness, we fi~d
:2
S5 0.6 0.3 S5
f-
a linear relationship. That is to say, the ratIo
f- t- f d ' If increases linearly with h1fI }.· If this is
f-

~ 0.4 0.2
Z
w
S5 d~~~ f;;eSother power-ratio points in the signal
S5 I spectrum, we obtain a set of curves, typical ones
I o
o U being shown in Fig. 4 for power points from 0.1
U 0.2 0.1 ~
to 0.9. From these curves, if one knows the ocean
spectrum and wave height, the spectral content of
~~-L~~L-L-~-L~-~±;~~~--~3~0~ the radio signal can now be predicted.
100
ROUGHNESS PARAM ETER hl/;/ >" (mils)
Fig. 2.- Coherent and incoherent reflection com-
Forward Scattering Over Land
ponents for rough ocean surfaces. As is true with forward scattering over water,
there is a strong dependence of field strength on
3 C. 1. Beard,!. Katz, an d L. M. Spetner, " P henomenological Model
the character of the land surface. Quantitatively,
of Microwave Reflections from the Ocean ," Trans . I RE, AP-4, April
1956, 162-1 67.
5 C. 1. Bear d a n d 1. Katz, " T he Depen dence of Microwave R a d io Signa l
4 C. 1. Beard , "Coheren t a n d Incoherent Scattering of Microwaves from Spectra on Ocean R oughness a n d Wave Spect ra ," T rans. IR E , AP-5,
t he Ocean ," Trans. I RE, AP-9, Sept. 1961, 470-483. April 1957, 183-191.

January - February 1963 13

 2 0 ~----~-----r-----,~----.-~---r-----'
In natural terrain the irregularities can be
large and receivers may be situated in shadow
regions. In such locations we are concerned with
18 ~----4-----~-----+------~------r-----1
multipath propagation; reflections from hills or
large objects give rise to strong signals from direc-
tions other than the transmitter direction, these
reflections being independent of frequency or
polarization. "Ghosts" in television reception and
cross-talk in communications systems are good
examples of multipath conditions.
Propagation over ground is also highly de-
1 2 ~----4-----~~---~------~J~--~--~ pendent on the season of the year because of
changes in the dielectric constant of the soil.
, Signal strength changes of as much as 14 d b have
~ 10 ~----4-----~--~-~---~~+----~~---.~ been found over smooth ground when soil moisture
content varied. Trees, shrubs, and buildings have
a decided effect on transmission of radio waves.
A forest may be either opaque or transparent
depending on the tree density. Attenuations
through wooded areas have been measured,
varying between 0.02 and 0.35 db per meter 6 de-
pending on the density, polarization, season, and
4 ~----+""-I
frequency.
By comparing the foregoing with the earlier
treatment of over-water propagation, one can see
easily that the over-water case is better understood.
Clear-cut quantitative statements can be made
concerning the over-water situation, and as a
OL-____ ~ _____ L_ _ _ _ ~ ______ ~ ____ ~ ____ ~

result we may expect dependable predictions for

o 200 400 600
given sea-state conditions. In the over-land case,
RO UGHN ESS hofr/ ).. (mils)
on-site tests must be made to determine the propa-
gation expected to prevail at any distance from a
Fig.4-Variation of the ratio of the radio signal
spectral width to the frequency of the maximum transmitter.
of the wave spectrum, with ocean roughness.
Backsca tter Over Water
however, we are in a much less favorable position We now turn our attention to the backscatter or
regarding forward scatter over terrain as compared "clutter" problem. A smooth conducting surface
with propagation over water. If the earth is smooth, backscatters only at normal incidence. For all
the reflection coefficient can be computed just angles other than 90° there is no backscattered
from a knowledge of the electrical properties of energy. A rough surface, however, scatters energy
the surface. Measurements made over rough in other than the specular direction. Radars
surfaces indicate that reflection coefficients are aboard ships and aircraft are frequently limited in
about 0.2 to 0.4. However, small differences in their operation because of such clutter. If the
receiver position can cause significant changes in clutter is larger than the return from a target, the
signal strength; hence, signal strength prediction target will be obscured. While the physical nature
is difficult. Measurements almost always fall below of the backscattering problem is not yet well under-
predicted values, sometimes as much as 30 db. stood, there are certain features of the problem
Most research over land has been done with broad- which are gradually becoming clear.
cast frequencies. The deviation from plane-earth Sea clutter is generally described in terms of the
predicted values varies inversely with frequency in radar cross section per unit area of the sea surface,
the range of frequency between 40 and 1000 me; 0'0. This is a dimensionless measure of the reflec-
at 40 me measured and predicted median values tivity of the surface, and it depends on the grazing
are in agreement, while at 1000 me the discrepancy angle, polarization, sea roughness, and orienta-
is about 27 db; the variation is linear on a semi-log tion.
plot of db below plane-earth value versus fre-
6M. P. Bachynski, "Microwave Propagation over R ough Surfaces,"
quency. RCA Review, 20. June 1959, 308-335.

14 APL Technical Digest

 20
As yet we do not have quantitative measures of
~
~
« roughness for backscattering. If the roughness
ID
1- ::0- measured for forward scattering holds for clutter,
z~
=> .
'-.. b
then at 10° and for X-band (3 cm), for example,
Zu.i it would take ripples of about 2 in. root-me an-
OU
f= « square to make the sea effectively rough. Thus,
~~ - IO
Vl~ one would expect to pass from the "smooth" curve
Vl
O~ - 20
to the "rough" curve with only the slightest wind
~ Vl
ULL
blowing over the sea. "Smooth" conditions for
:(0 - 30 microwaves would indeed be rare.
o A better understanding of the clutter problem
«
~
- 40 is obtained by studying results of experiments
o 10 20 30 40 50 60 70 80 90
DEPR ESS ION ANGLE (degree s)
performed expressly to determine CTO. Curves of
Fig. 5-Qualitative variation of radar cross section
CTo shown in Fig. 6 were obtained by the Naval
of the sea with depression angle for smooth, inter- Research Laboratory. The clutter is greater for
mediate, and rough surface conditions. vertical polarization than f<;>r horizontal at the
smaller angles, but the difference tends to disap-
Flying over sea surfaces varying from "smooth" pear at higher angles. Note also that the curves
to "rough" to determine the ocean cross section seem composed of two curves, one rather sharply
versus grazing angle, we would find the following. peaked around the vertical, and a second, much
Under calm water conditions, measurements would flatter curve. One interpretation of such composite
indicate a peak in CTO at 90°, with rapidly decreas- curves states that the peak around vertical in-
ing values at angles off the vertical as illustrated cidence results from scattering from large facets
in Fig. 5. If the wind waves and/ or swell increased, on the surface, while the less steep portion is
the measured curve would have less of a peak at caused by more isotropic scattering from smaller
90°, but the magnitude of CTO would increase at the facets.7 For high sea states, there would be no dif-
smaller angles and would perhaps follow the curve ference in return on the two polarizations.
marked "intermediate." For a very rough sea, the 7L. M. Spetner and I. Katz, " Two Sta t istical Models for Radar Terrain
CTo-versus-grazing-angle curve would be quite flat. Return," Trans. IRE, AP-8, May 1960, 242- 246.

20

X· BAN D

10 o = HOR IZ . POLA RI ZATION

I:J. = VERT . POLARIZATION
:0
~ (from Naval Research Laboratory data)
~

~
<
I-
Z
=>
--- ---
'-
z
0
f=

~
- 10

..,"'"
.",""'"
.",.,.- --
NEW J ERS EY WOOD S _ _ - --

Vl

0oc
U

<
0
«
- 20 r--

- 30 ~

-40
o 10 20 30 40 50 60 70 80 90
DEPR ESSION ANG LE (d eg rees)

Fig. 6-Variation of the radar cross section of a rough surface with depression angle,
in water- covered and forest-covered surfaces.

Jan uary - February 1963 15

 There is a clear dependence of clutter on wind In the paper discussing the theory of radar
direction; the upwind direction gives greater terrain return/ a model was formulated that applies
return than the downwind. Again, as with two to both land and water surfaces and deals with the
polarizations, the difference tends to disappear wavelength dependence of various types of sur-
at the larger depression angles. faces . That approach considered two types of
Many curves have been obtained experimentally surface, a distribution of random scatterers and a
that illustrate the change of cross section with surface consisting of specular reflectors. The nor-
radar frequency and with sea roughness. These malized radar cross section of the surfaces was
data bear out the qualitative statements made computed as a function of depression angle and of
earlier that clutter increases with increasing wavelength. The isolated scatterer model led to no
roughness and with increasing frequency. For a dependence on depression angle but to a wave-
given wave-height condition, as one increases the length dependence which varied from ~- 6 to ~-2,
radar frequency (decreases wavelength), the ap- depending on density of scatterers and on the rela-
parent roughness increases. tionship between ~ and the geometrical areas of
Of prime importance for search radar functions the scatterers. On the other hand, the specular
is the shape of the clutter curve at very low angles. reflection surface showed a depression-angle de-
Generally, clutter on vertical polarization is pendence of
greater than on horizontal; however, there is some exp { -cOt(~/ 2CT 82) } or exp { -cot2(~/ 2CT 82)},
indication that, at angles below perhaps 3°, on
horizontal polarization the clutter is greater than where ~ is the depression angle and CT 8 is the
on vertical. variance of the slope of the reflecting facets. The
wavelength dependence was ~ -2 for this surface.
Backscatter Over Land It appears that both models represent true surfaces.
A somewhat typical curve for CTo-versus-grazing- But whatever the model, it seems certain that
angle over land is shown in Fig. 6. This one was there is a fairly strong wavelength dependence. It
obtained in flights over a forested area. Unques- has been proposed ll that this property be used in a
tionably, trees 60 ft high present a "rough" surface radar to distinguish various types of surfaces. A
to 3-cm radiation. As a result, it can be expected "polychromatic" (multifrequency) radar could be
that the curve would be quite flat. Actually, used in mapping the ground for automatic navi-
curves similar to the one shown are flat even for gation. Whether such a radar is feasible depends
grassy areas with vertical irregularities just a few on further research in terrain reflectivity. If
inches high, i.e., even grass is rough at ~ = 3 cm. various natural surfaces turn out to have pre-
Not until we approach something like a concrete dictable and different wavelength dependencies,
road does the radiation approach the smooth-earth then a polychromatic radar is a real possibility.
value.
Literally hundreds of curves of terrain clutter Future Research
now exist for different terrains, polarizations, and From the preceding discussion on forward- and
frequencies. 8 The results obtained are surprisingly backscattering, one may reasonably conclude that,
repeatable- better than ± I db. although progress is being made in our under-
No unified theory has yet been advanced that standing of the total reflectivity process, many
can be used to predict backscatter for any given gaps are still present. Under certain geometries
surface, over land or over water. Some limited and surface conditions we can predict the field
success has been achieved by Peake 9 using the strength .in space. In general, however, this is not
method of Rice. l o His success has been limited to true; the exact causal mechanisms are as yet not
surfaces with roughness and electrical properties firm.
like those of asphalt and concrete roads. For more Two major problem areas are clear. First, a
irregular grassy surfaces, the present theoretical complete solution of the rough-surface reflectivity
approach seems to fail. problem will likely come only with the bistatic *
approach since forward- and backscattering are
8R . L. Cosgriff, W. H . P eake, a nd R. C . T aylor , "Terra in Scattering
Propert ies for Sensor System Design ," T errain Handbook, Ohio State special cases of the general scattering phenomenon.
University , May 1960.

9 W. H. P eake, " T he Interaction of Elect romagnetic R a diation wi t h 111. Kat z and L. M. Spetner, " A P olychro matic Ra da r," The J ohns
Some Natural Surfaces, " Ohio State U niversi t y Report No. 898-2, Ma y H opkins University, Applied Physics Labora tory , C F 2898, Oct. 24,
30, 1958. 1960.

10S. O. Rice, "Refl ection of Electromagnetic Waves by Slightly R ough * Where t he tra nsmi tter a nd receiver are in t he same pla ce we designate
Surfaces, " Th e Theory of E lectromagnetic Waves, In terscience P ublishing the m easured reflectiv ity as "monostatic"; if t hey a re separated, t he
Co., New Y ork , 1951. scattering is called " bist a t ic."

16 APL T echnical Digest

 Secondly, an adequate description of the surface be accurately known over a wide angular region.
is usually lacking. New research is being con- Areas on the surface at different depression angles
ducted in both of these areas; some concepts and Y;R contribute different doppler components to
trends of this research are discussed below. the received signal. If the reflectivity of the target
airplane or calibrator is known and the received
Bistatic Reflectivity signal spectrum is measured, the reflectivity of
The bistatic geometry is pictured in Fig. 7. each of the areas on the sea may be determined . A
Here we see two aircraft, one containing the trans- program to obtain bistatic reflectivity is contin-
mitter and the other the receiver. Ideally, these uing at APL, and it is hoped that a dependable
aircraft are equipped with narrow-beam antennas set of bistatic curves may be obtained. Clearly,
in order to illuminate or receive from a narrow with the transmitter mounted on shipboard, the
angular region. The indicated depression angles maximum Y;T will be restricted to less than several
Y;T and Y;R are not usually the same. For a degrees.
given surface condition, all possible combinations
of angular values are scanned. In practice, for Sea Surface Measurelllents
practical reasons, it is likely that a wide-beam It was mentioned earlier that before a complete
antenna model would be used with the transmitter solution of rough-surface scattering is obtained,
and a narrow-beam with the receiver. Experi- we must find an adequate description of the sur-
ments using this procedure are visualized for the face. Present methods for measuring ocean waves
future. are not applicable to the present problem. We can,
at any point in water and to any desired sensitivity,
determine the height of water as a function of
time. Although from height-time recordings a
spectrum can be computed, this spectrum is not
the one required. The spectrum obtained from a
measurement of height versus distance along the
surface is the one needed. If all waves moved with

, constant speed and direction, we could transform

the time spectrum into the desired space spectrum.
Unfortunately, waves of different lengths move
with different speeds and not necessarily in the
same direction; thus, the transformation from the
time to the space spectrum cannot be made.
To obtain the description of the surface, a joint
program sponsored by the Office of Naval Re-
search and APL is in progress to develop a stereo-
photogrammetric method to measure surface
structure. A special pair of stereo cameras is being
mounted aboard a ship to determine the feasibility
of making wave measurements with the desired
Fig.7-Geometry of bistatic reflectivity problem;
V;T = transmitter depression angle, and V;R = re- sensitivity. The present goal is to measure wave
ceiver depression angle. heights with a sensitivity of ±O.l in. From these
photographs, three-dimensional space spectra are
A program is now being conducted to obtain to be obtained for times during which radar re-
data from missile flights by using the doppler- flectivity measurements are made.
shifted frequency of the returned signal. With the
transmitter aboard ship and the receiver on an Conclusions
aircraft or missile, the bistatic reflectivity for a Progress is being made in our understanding of
somewhat limited, but useful, range of transmitter reflectivity of electromagnetic waves from rough
depression angles can be obtained. The ship surfaces. Ultimately, we hope, radar observation of
transmits a continuous-wave signal that illuminates surfaces will permit computations of satisfactory
a target airplane and the sea surface. Signals re- descriptions of the surface states. Over land
flected from the target and the sea contain doppler it will provide an all-weather navigation system.
frequencies that depend on aircraft speeds and Over water we will be able to measure wave
geometry. The receiver airplane or missile should height with a remote probe, and our ability to
have a wide-beam antenna, and its pattern must detect targets amid sea clutter will be improved.

Ja nuary - February 196) 17