dielectric properties of soils as a function of …€¦ · 3 figure 25. the permittivities of...

CRES .REMOTE SENSING LABORATORY

HIM

DIELECTRIC PROPERTIES OF SOILS AS A FUNCTIONOF MOISTURE CONTENT

REMOTE SENSING LABORATORY

, RSL Technical Report 177-47

"(NlsT-CB-Tuf868)~ "blELECTRfc PROPERTIES OFSOILS AS A FUNCTION OF MOISTURE CONTSNT(Kansas Univ. Center for Research, Inc.)61 p HC $4.25 CSCL 083

G3/46

N75-2&555

Unclas266U4

Josef CihlorFowwox T. Ulaby

November, 1974

Supported by:

NATIONAL AERONAUTICS AND SPACE ADM1NISTRAT1OLyndon B. Johnson Space Center f '̂- \ ""

Houston, Texas 77058 /C^

CONTRACT NAS 9-10261 /C? ^ *•

•^7

THE UNIVERSITY OF KANSAS CENTER FOR RESEARCH, INC.2291 Irving Hill Drive—Campus West Lawrence, Kansas 66045

THE UNIVERSITY OF KANSAS SPACE TECHNOLOGY CENTERRaymond Nichols Hall CENTER FOR RESEARCH , INC .

2291 Irving Hill Drive—Campus West Lawrence, Kansas 66045

Telephone:

DIELECTRIC PROPERTIES OF SOILS AS A FUNCTION

OF MOISTURE CONTENT

CRES Technical Report 177-47

Josef Cihlar

Fawwaz T. Ulaby

November, 1974

Supported by:

NATIONAL AERONAUTICS AND SPACE ADMINISTRATIONLyndon B. Johnson Space Center

Houston, Texas 77058

CONTRACT NAS 9-10261

.REMOTE SENSING LABORATORY

TABLE OF CONTENTS

Page

ABSTRACT v

1.0 INTRODUCTION 1

2.0 TARGET-SENSOR INTERACTION 1

2.1 Reflection Coefficient 22.2 Skin Depth 3

3.0 DIELECTRIC PROPERTIES OF SOILS 4

3.1 Moisture 53.2 Bulk Density ". 343.3 Soil Type 383.4 Temperature 423.5 Frequency 42

4.0 REPRESENTATIVE DIELECTRIC VALUES 43

APPENDIX A. BIBLIOGRAPHY 57

REFERENCES 60

LIST OF FIGURES

Figure 1. The dielectric spectrum of water at two temperatures.(FromHoekstra and Delaney (1974)). Rage 7

Figure 2. Relative dielectric constant values of soil as a function ofgravimetric water content: Sand; frequency 0.13 GHz - 3.0 GHz;real part. Poge g

Figure 3. Relative dielectric constant values of soil as a function ofgravimetric water content: Sand; frequency 0.13 GHz - 3.0 GHz;imaginary part. Page 9

Figure 4. Relative dielectric constant values of soil as a function ofgravimetric water content: Loam; frequency 0.13 GHz - 3.0 GHz;real part. page 10

Figure 5. Relative dielectric constant values of soil as a function ofgravimetric water content: Loam; frequency 0.13 GHz - 3.0 GHz;imaginary part. Page 11

Figure 6. Relative dielectric constant values of soil as a function ofgravimetric water content: Clay; frequency 1.1 GHz - 1.5 GHz;real part. Page 12

Figure 7. Relative dielectric constant values of soil as a function ofgravimetric water content: Clay; frequency 1.1 GHz -1.5 GHz;imaginary part. Page 13

Figure 8. Relative dielectric constant values of soil as a function ofgravimetric water content: Sand; frequency 9.5 GHz - 19.4 GHz;real part. Page 14

Figure 9. Relative dielectric constant values of soil as a function ofgravimetric water content: Sand; frequency 9.5 GHz - 19.4 GHz;imaginary part. page 15

Figure 10. Relative dielectric constant values of soil as a function ofgravimetric water content: Loam; frequency 9.4 GHz - 19.4 GHz;real part. Page 16

Figure 11. Relative dielectric constant values of soil as a function ofgravimetric water content: Loam; frequency 9.5 GHz - 19.4 GHz;imaginary part. Page ]j

it

Figure 12. Relative dielectric constant values of soil as a function ofgravimetric water content: Clay; frequency 9.4 GHz - 13.7 GHz;real part. Page 18

Figure 13. Relative dielectric constant values of soil as a function ofgravimetric water content: Clay; frequency 10.0 GHz - 13.7 GHz;imaginary part. Page 19

Figure 14. Relative dielectric constant values of soil as a function ofgravimetric water content: Sand, loam, clay; frequency 34.5 GHz -37 GHz; real part. Page 20

Figure 15. Relative dielectric constant values of soil as a function ofgravimetric water content: Sand, loam, clay; frequency 34.5 GHz -37 GHz; imaginary part. Page 21

Figure 16. Measured dielectric constant data of loamy soils as a function ofgravimetric water content around 10 GHz. Solid curves weredrawn to fit the data points and the broken curves were extrapolated. Page 23

Figure 17. Relative dielectric constant values of soil as a function of volumetricwater content: Loam, clay; frequency 0.03 GHz - 0.5 GHz; real part. Page 24

Figure 18. Relative dielectric constant values of soil as a function of volumetricwater content: Loam, clay; frequency 0.03 GHz - 0.5 GHz;imaginary part. Page 25

Figure 19. Relative dielectric constant values of soil as a function of volumetricwater content: Sand, loam, clay; frequency 1.1 GHz - 1.5 GHz;real part. Page 26

Figure 20. Relative dielectric constant values of soil as a function of volumetricwater content: Sand, loam, clay; frequency 1.1 GHz -1.5 GHz;imaginary part. Page 27


Figure 22. Relative dielectric constant values of soil as a function of volumetricwater content: Sand, loam, clay; frequency 3.8 GHz - 4.0 GHz;imaginary part. Page 29


Figure 24. Relative dielectric constant values of soil as a function of volumetricwater content: Sand, loam, clay; frequency 10.0 GHz - 26.0 GHz;imaginary part. Page 31

• • •MI

3Figure 25. The permittivities of powdered rocks at a density of 1.0 g/cm .

showing the mean, and maximum and minimum values as derived fromsolid rocks. (From Campbell and Ulrichs (1969)). Page 35

Figure 26. The permittivities of powdered rocks at 40% porosity. (FromCampbell and Ulrichs (1969)). Page 36

Figure 27. Effect of bulk density on the relative dielectric constant. Page 37

Figure 28. Change in relative dielectric constant due to soil moistureunits used. Page 39

Figure 29. Bulk density variation due to compaction as a function ofgravimetric water content. Page 40

Figure 30. Soil textural triangle. Page 41

Figure 31. The permittivity of rocks at 35 GHz as a function of temperature.(From Campbell and Ulrichs (1969)). Page 44

Figure 32. A comparison of the permittivities of solid rocks at 450 MHz and35 GHz.(From Campbell and Ulrichs (1969)). Page 45

Figure 33. Real and imaginary part of the relative dielectric constant as afunction of frequency for a silty clay, moisture content 15% byweight. (From Hoekstra and Delaney (1974)). Page 46

Figure 34. Representative dielectric constant values as a function ofvolumetric water content for sand, loam, and clay af 1.3 GHz. Page 51

Figure 35. Representative dielectric constant values as a function ofvolumetric water content for sand, loam, and clay at 4.0 GHz. Page 52

Figure 36. Representative dielectric constant values as a function ofvolumetric water content for sand, loam, and clay at 10.0 GHz. Page 53

Figure 37. Skin depth as a function of volumetric water content, frequency,and soil type. Page 54

Figure 38. Power reflection coefficient and emissivity as a function of volumetricwater content, frequency, and soil type. Page 56

IV

ABSTRACT

In studying applications of microwave remote sensing techniques to

agricultural, hydrological, and other related problems, the soil dielectric

properties are of considerable importance. In recent years, numerous soil

dielectric constant measurements have been made. Due to the complexity of

the data acquisition procedure, however, the number of measurements in a

single experiment is usually too small to permit a systematic analysis of the

effects of the various parameters that influence soil dielectric properties.

The objective of this report is twofold: 1) to present soil dielectric

constant measurements obtained by various researchers in one publication, thereby

assisting in analysis and utilization of microwave remote sensing data and 2) based

on these measurements, to determine the dependence of the dielectric constant on

various soil parameters. Moisture content is given special attention because of its

practical significance in remote sensing and because it represents the single most

influential parameter as far as soil dielectric properties are concerned. From the

experimental measurements collected in this report, relative complex dielectric

constant curves are derived as a function of volumetric soil water content at three

frequencies (1.3 GHz, 4.0 GHz, and 10.0 GHz) for each of three soil textures

(sand, loam, and clay). These curves, presented in both tabular and graphical form,

were chosen as representative of the reported experimental data. Calculations based

on these curves showed that the power reflection coefficient and emissivity, unlike

skin depth, vary only slightly as a function of frequency and soil texture.

1.0 INTRODUCTION

The target-sensor interaction process in remote sensing is governed by the

target geometry and dielectric properties relative to the sensor parameters. At

microwave frequencies, the dielectric properties of soils are particularly important

because 1) they are very susceptible to moisture content, and 2) at incidence angles

close to nadir, the target response of vegetation-covered surfaces can be influenced

by contribution from the underlying soil. These characteristics have recently stimulated

considerable research efforts to ascertain the operational feasibility of microwave sensors

in remote soil water content determination (Appendix A).

The relationship between dielectric properties and soil characteristics has

not been dealt with systematically, partly because experimental data are presented

in a variety of sources. The purpose of this report is to analyze the behavior of

soil dielectric properties at microwave frequencies as a function of several soil

variables.

2.0 TARGET-SENSOR INTERACTION

The target response in the microwave region of the electromagnetic

spectrum can be characterized in terms of its backscattering coefficient for active

sensors (radar) and in terms of its emissivity for passive radiometers. Both quantities

are functions of the following sensor and target parameters:

Sensor Parameters

1o Frequency

2. Incidence angle

3. Polarization

Target Parameters

4. Dielectric properties

5. Roughness of the surface and of subsurface from which a

measurable radiation is reflected or emitted (relative to

wavelength)

6. Thermometric temperature of the surface and subsurface

material.

The effect* of the first five parameters are explicit for either active or

passive sensors. The rmome trie temperature affects the surface scattering and

emission character through its influence on the dielectric properties. Moreover,

the radiation measured by a radiometer is directly proportional to the target

rhermometric temperature.

Dielectric properties of a medium influence the radar scattering

coefficient and the radiometer emissivity in two ways: a) through the Fresnel

reflection coefficient and b) by defining the effective depth in the medium

responsible for the backscattered or the emitted energy. Hence, before discussing

the dielectric properties of soils, it might be helpful to define these wave propagation

parameters in mathematical form.

2.1 Reflection Coefficient

The Fresnel reflection coefficient r is defined for a perfectly smooth

interface between two media. It relates the magnitude and phase of the reflected

electric and magnetic fields to those of the incident fields of a plane wave. In terms

of the subject of this report, the parameter of interest is the power reflection coefficient,2

R = r . If the incident wave is in air (assumed to have the same electromagnetic

properties as vacuum) and the reflecting medium is homogeneous, then R is given by

(Hidy, et al., 1972):

R, =h [(p+y)

[(yk -p)2 + ( yk 2 -q )2 ]

• 7 - 9 (2)

where h and v are subscripts defining horizontal and vertical polarization, respectively,

k, and V* are rne real anc' imaginary parts of the relative complex dielectrfc constant,

respectively, and y = cos 6 where 6 is the incidence angle relative to nadir. The

parameters p and q are given by:

r-t -

1/9o '/ *

+ - +v 2

For a smooth surface, the emissivity is defined as:

e. = 1 - R. ; i = h or v (3)i i ' x '

Although most natural earth surfaces (excluding calm water) are not electromagnetically

smooth at microwave frequencies, the dependence of R on soil moisture (through k, and

k^) can be used as a guide in the study of the response of microwave sensors to soil

moisture .

2.2 Skin Depth

The energy received by a microwave radiometer viewing a given target is

composed of contributions from the target surface as well as sub-surface layers.

Similarly, the backscattered energy (in the case of active microwave sensors) is also

depth dependent. The penetration of microwaves into a medium is defined by the

attenuation coefficient of the medium, which in turn is a function of its dielectric

properties:

k 2 1/2 1/2

where

a = attenuation coefficient, nepers/m ;

X = wavelength, m;

k. , k«= real and imaginary parts of the relative dielectric constant of

the medium.

In the above expression, the relative magnetic permeability was assumed to be unity,

which is a good approximation at microwave frequencies for most natural surfaces

including water.

At a depth h beneath the surface, the power is related to the surface value

P(o) by:

At a depth 6 such that ad= 1, -p-Ĵ U- =0.13. 5 is defined as the skin depth.

Reflections from a plane 6 deep by a perfect reflector will undergo additional

attenuation by the same amount, thereby arriving at the surface having a magnitude

that has been reduced to ~ 0.017. If the medium depth profile of k, and kj is not

a constant, then 8 can be defined by:

6i

adh = 1 (6)

3.0 DIELECTRIC PROPERTIES OF SOILS

For naturally occurring substances, dielectric constant measurements in the

microwave region are usually not repeatable by different investigators (Edgerton, et

al., 1971, Ward et al., 1969). Dielectric measurements of soils require high pre-

cision and are time-consuming because of the number of factors that affect soil

dielectric behavior (Poe et al., 1971). Furthermore, since adequate techniques

are available only in the laboratory, preparation of samples for measurements in-

volves a profound disturbance of the mutual arrangements of soil aggregates. It is

not certain to what extent the laboratory-prepared samples are representative of

soils as they exist in the field. However, since no better data are available, these

measurements are continually used for interpretation of remote sensing data,,

The dependence of the dielectric constant has been studied as a function of

several parameters: moisture, bulk density, soil type, temperature, and frequency.

These effects are discussed in the following sections.

3.1 Moisture

For soil dielectric constant measurements, soil moisture is usually expressed

on a weight basis (i.e., gravimetrically):

Wm = yrr^ x 100 (%) (7)

W j

where

m = moisture content by dry weight (%);wW = weight of water in the sample (g);wW , = weight of dry soil in the sample (g).

Lundien (1971) suggested that moisture content expressed on a volume basis may be

preferrable because data plotted against the volumetric water content m should

have little sensitivity to changes in bulk density or degree of soil compaction. The

value of m may be calculated using the bulk density of soil sample Ps and the

density of water pw

Wpw

3(grams/cm) (8)

When soil moisture is expressed on a weight basis, it is possible for two soils to contain

the same water content in percent by weight while the actual amounts of water differ.

For example, a 30% moisture content represents both 0.36 g of water for bulk density ofq q o

1.2 g/cm and 0.54 g for bulk density of 1.8g/cm in 1 cm of soil. It has been

argued that the number of water molecules (i.e ., dipole moments) per unit volume

rather than the weight proportion determines the dielectric contribution of water

(Hoekstra and Delaney, 1974). Formulas for calculating dielectric constants of

mixtures are also based on volume fractions (Poe et al., 1971; Birchak et al., 1974).3 3Consequently, soil water should be expressed in cm or grams per 1 cm of soil when

related to soil dielectric properties. However, many soil dielectric constant

measurements have been presented on a weight basis only. In this report, both data

have been included.

Available studies show that the relative (with respect to vacuum) dielectric

constant of soils increases with increasing moisture content. This is to be expected

since the relative permittivity of dry soil is less than 5 while that of water can be

more than an order of magnitude larger (Figure 1). Figures 2 through 15 contain

measurements on the weight basis, and Figures 17 through 24 show dielectric constant3as a function of water expressed in grams per cm . The data are divided according to

frequency and soil texture (i.e., relative proportions of soil particles of different

sizes). The following observations can be made.

(1) Few of the measurements were taken for moisture contents near 0 per

cent. Results of Matzler (1970, Figure T2) and Geiger and Williams

(1972,Figure 14) show a tendency toward sharp increase in the real

part of the dielectric constant k^ for this moisture content range. Matzler

(1970) suggested that such increase may be due to chemical changes

resulting from addition of water, such as hydration. From the remote

sensing standpoint, these changes are not very important since under

natural conditions, dry soil always contains some water. For example,

the amount of hygroscopic water (i.e., soil water held mostly by soil

.colloids under tensions above approximately 31 atmospheres so that it

is unavailable to plants) was found to be 3.4% for a sandy soil and16.1% for a silty clay (Russell, 1939).

(2) At low moisture contents, k , increases slowly (Figures 2, 4,

8, 10) or remains constant (Figures 6, 12). Lundien (1971)

and Wiebe (1971) explained this phenomenon by the adsorption

of water at the surface of fine particles; the adsorbed water

has a limited mobility (similar to water molecules bound in

ice) and consequently its dielectric constant can be as low

as 0.1 that of free water (Grim, 1968). The possibility that

80

60

40

20

0

40

20

0

10'

\\\

10' 1010

Frequency, Hz10n 10 12

Figure 1. The dielectric spectrum of water at two temperatures. (FHoekstra and Del one y (1974)).

A YUMA SAND, 1.1 GHz (LUNDIEN, 1971)* YUMA SAND, 1.5 GHz (LUNDIEN, 1971)• SAND, 0.13GHz(LESCHANSKII etal., 1971)o SAND, 0 .3GHzLESCHANSKIIe ta l . , 1971)x SAND, 1.0GHz(LESCHANSKIIetaL, 1971)° SAND, 3.0GHz(LESCHANSKII etal.,1971)

24 r

22 -

20 -

£ 18

2 16

^ 14 -C£L 1H '

£* 12 -

£ 10 -

3 8ho

oLU—ILU

Q

6

4

0

A€>•

QBR

xs&

- a

0 5 10 15 20SOIL MOISTURE

(PERCENT BY WEIGHT)

25

Figure 2. RelaMve dielectric const-ant values of soil as a functionof gravimetric water content: Sand; frequency 0.13 GHz3.0 GHz; real part.

8

ro — —.̂ «j no b. •as TOCO

CO

£DO)

38

36

34

32

30

28

26i —

1 24

^ 22

g 20

30

28

26

H- 24CtL

3. 22

20

is16

A OPENWOOD SILT LOAM, 1.1 GHz (LUNDIEN, 1971)A OPENWOOD SILT LOAM, 1.5 GHz (LUNDIEN, 1971)• RICHFIELD SILT LOAM, 0.3 GHz (LUNDIEN, 1966)a LOAM, 0.13 GHz (LESCHANSK11 et al., 1971)f LOAM, 0.3 GHz (LESCHANSK 11 etal., 1971)o LOAM, 1.0 GHz (LESCHANSK 11 etal., 1971)V LOAM, 3.0 GHz (LESCHANSK 11 etal., 1971)

a

£ 1400

o 12o 10cc0 oLU O

LU

0 6

4

2

0^

-D •

-

f•

"

f ° ff V- ° Vo A" 1 ? * ^ ^ ^. \7,A ,

0 5 10 15 20 25 30 35SOIL MOISTURE (PERCENT BY WEIGHT)

40

Figure 5. Relative dielectric constant values of soil as a function of gravimetricwater content: Loam; frequency 0.13 GHz - 3.0 GHz; imaginary part.

11

;>»

36

1 A34

32

30

28

5 2 6a.

-o LONG LAKE CLAY, 1. 1 GHz

(LUNDIEN, 1971)• LONG LAKE CLAY, 1.5 GHz

(LUNDIEN, 1971)-

0

-

LU24

22

20

I 18oa 16oa 14

10

8

6

41-

2

00 5 15 10 20 25 30 35 40 45

SOIL MOISTURE (PERCENT BY WEIGHT)50

Figure 6. Relative dielectric constant values of soil as a function of gravimetricwater content: Clay; frequency 1.1 GHz - 1.5 GHz; real part.

12

LU LU

O O

Mnio

0 0

O

O •

0 •

O •

LTN

Jo

oLU

QQ

o O^~* CO

UTN

CNJ o oo CNJ

AdVNIOVWI'iNVlSN033ldl3313ia

jl)o

§ .° Ii- a.01 _

c —

**• hTO X

Joo —

Ifi|O 1)

Jr _o

=5^« i*> c

*Z V^ c« 8

IO)

13

30

28

26

24

| 22

^ 20

A SAND, 9.5 GHz(LESCHANSKII et al., 1971)

a SAND, 15.8 GHz(LESCHANSKII etal . , 1971)

o.LOAMY FINE SAND, 19.4 GHz(GEIGER AND WILLIAMS,TO BE PUBLISHED)

24

22

20

£18

£ 14

sg 12o-10r

* ABILENE CLAY LOAM,10.6GHz(WIEBE, 1971)

x HOBAN SANDY LOAM,10.6GHz(WIEBE, 1971)

o GILA SANDY LOAM,10.6GHz(WlEBE, 1971)

D AMARILLO FINE SANDY LOAM,10.6 GHz (WIEBE, 1971)

* RICHFIELD SILT LOAM,9.4 GHz (LUND I EN, 1966)

+ LOAM, 9.5 GHz(LESCHANSKII etal.,1971)

f LOAM, 15.8 GHz(LESCHANSKII etal., 1971)

\7 FINE SANDY LOAM, 19.4 GHz(GEIGER AND WILLIAMS,TO BE PUBLISHED)

* SANDY CLAY LOAM, 19.4 GHz(GEIGER AND WILLIAMS,TO BE PUBLISHED)

v

OLU

8 h

6

V

* *

00 5 10 15 20 25 30 35 40 45

SOIL MOISTURE (PERCENT BY WEIGHT)

Figure 10, Relative dielectric constant values of soil as a function of gravimetric water content:Loam; frequency 9.4 GHz - 19.4 GHz; real part.

16

r ~

! î r>-i 0"

— 'LU"

; co^ ^r "j-1i ^^ ^ "̂i -•••*. ^^^

i i*-- r— •—•--,a^ &* r*- î— i •— i O^ • — •

.. -^ NLU LU • 31CO CO LU O

i±|!±l SvO

î2M M ^~31 1C M — •.0 0 31 2|sosoOg

o o *>. "^ -* o >-

ssî< < ̂ "<O O *= co^^gLU>- >- _J Z^gîzo < § oLUOO ^-}

^

^_

- C

o MILLER CLAY, 10.0 GHz(S. L. LEE , TEXAS A &M UN IVPERSONALCOMMUNICATION)

A TARRANT STONY CLAY, SIEVED,10.6 GHz (WIEBE,1971)

o TARRANT STONY CLAY, UNSIEVED,10.6 GHz (WIEBE,1971)

• HOUSTON BLACK CLAY, 10.6 GHz(WIEBE,1971)

• MILLER CLAY, 10.6 GHz(WIEBE.1971)

A SHARKLEY CLAY, 9.4 GHz(LUNDIEN,1966)

a SOIL NOT SPECIFIED, 13.7 GHz(POE et al., 1971)

* CLAY, 10.0GHz(MATZLER, 1970)0 •

*

•

A• ^\

•OA

DO

A A

fg]

1 1 1 1 1 ! 1

5 10 15 20 25 30 35

o

o

a

A

'

1 i I

40 45 50SOIL MOISTURE (PERCENT BY WEIGHT)

Figure 12. Relative dielectric constant values of soil as a function of gravimetricwater content: Clay; frequency 9.4 GHz - 13.7 GHz; real part.

18

0°

M

O

0

or— I

Oa:LU— 1

i0

•>

^

CO

X

u_TUJ

_I

-ZO

CO

i—

<Qi0£

F— <

I (W

IEBE

.197

• ^

o>oo'

>-"

o3CQ

0

CO)̂

o310

1

. — .

CQ

§

MIO

of— «

M

0NO

0•-H . .

^JT —

O LU"

25?i ^iB

i

Mi

O

CO

OLU .̂— r— 1u_ r*.

O ^LUo_ _r1/1 TO

o "*>

^ Q_0 "CO

D

O

i °-' -D

o . -

* 3

oft-

*±. -•1 1 1 1 1 " ^ f

o >_^^ CQ

SO LLJCS4 oW>X

a:LU

0 °-ou —LUCXL

^^

OC3 Si— 1 ,

OCO

•e— »

OUJ

CVI C3 OO vO "̂ CVJ CD

clNViSNOOOIHL0313IQ

c£ou

o

u

2

o J

Itv5 c

8 I

'S O

c N2 x

o .o o

8 II I^> •*-

^ ^«S u

CO

V

O)

!

i— •r —^^bV*-,

1 9t

•

"ro

~o£

~_

îCO2:-5O^_**^CJ

^^^i— 1

*m-̂,LU

Q~^ID^_^

M^^^n_oLT\

r̂m

>-~

ô>-LU

^/

0£<^CO

^

,

^

i "^

+-J j1— 1r —o^

^^^TOi .O>III*r -*-J

oQ_

Mmor—ro

••,QLU

U_

OLUQ.CO

1—O~^

— i

OCO

•

~

**

o

*^Fs-U^*

o

0

•< D> ^ -

. D •

Dr .^ •^ X*o • ^ -

O „ 0D °

0 D«~oD)

'o

_̂o"uc3 .0 &8 o.~ o2 2

'B N

° î« o1-^_ 1

l£t/t * r\c O8 «o^ •

o -«r•̂ CO

ti X

1 20) 4)

=5 £A\a) ?|> ^

iavd 'iNviSNOO ouuoaiaia 0) —Of U

O)

20

M

ÔooNO"ro•

Q

-

S

Ô

0

CNJI —O^^

•*CO

^^^^

^̂J

iQ

<

ryL£~LU0LU0

M

Of^^^p*»m

•

S<0

>-O

- < Q. r3n^ ^^0 Lu t̂ •£^. « 5 , ,>- Z LUQ LU ^ O2 0 — Q_< — o —CO CO

• •

•N k̂ I~̂ > —ŝ J 1**̂

I

•

*t>

D •

«"*• "

soil dielectric constant is affected mostly by the loosely held

water is of potentially great practical importance since only

the water held with sufficiently small strength is available to

plants. These relations should affect data in Figures 2 to 14

in the following ways:

(i) The range in which k. changes slowly with in-

creasing moisture content should expand with

increasing proportion of clay particles. Data

in Figure 8, 10 and 12 (sand to clay in texture)

indicate that this probably occurs but the mea-

surements are not sufficiently detailed to estab-

lish the trend with certainty. Considering the

range of the hygroscopic water contents between

sand and clay mentioned above, one would

expect the differences to be more pronounced.

It should be noted that careful measurements by

Geiger and Williams (1972, Figure 14) showed

such a trend within a somewhat narrow textural

range ( loamy fine sand to sandy clay loam ).

Lundien (1971) also demonstrated this dependence

using measurements at five frequencies between

1.074 GHz and 1.499 GHz.

(ii) As the proportion of fine particles increases, the k,

curves should be shifted in the direction of higher

moisture contents and should have less steep slopes.

This is partly illustrated in Figures 2, 4, 6 and Figures

8, 10, 12, but a considerable overlap exists among

various soils.

In the measurements presented on a volume basis (Figures 17 through

24), the above trends are not as apparent. For example, k values

of Yuma sand and San Antonio clay loam near 1.0 GHz (Figure 19)

are very similar. In addition, the values of kj appear to increase

22

REAL PARTA ABILENE CLAY LOAM, 10.6 GHz (WIEBE, 1971)a AMARILLO FINE SANDY LOAM. 10.6 GHz (WIEBE, 1971)o RICHFIELD SILT LOAM, 9.4 GHz (LUNDIEN, 1966)\7 LOAM, 9.5 GHz (LESCHANSKII et al., 1971)• WATER, 9.3 GHz, 20° C (PARIS, 1969)

IMAGINARY PARTABILENE CLAY LOAM,

10.6 GHz (WIEBE, 1971)AMARILLO FINE SANDY

LOAM, 10.6 GHz(WIEBE, 1971)

WATER, 9.3 GHz, 20° C(PARIS, 1969)

IMAGINARY PART,K

0 20 30 40 50 60 70 80SOIL MOISTURE (PERCENT BY WEIGHT)

Figure 16. Measured dielectric constant data of loamy soils as a function ofgravimetric water content around 10 GHz. Solid curvet were drawnto fit the data points and the broken curves were extrapolated.

38

36

22

§ 20Ou

< 18

£ 16

£ 14Q

Id

10 h

k^_* _^

S 8

6 -

0

*o

• FAIRBANKS SILT, 0.5 GHz(HOEKSTRA AND DELANEY, 1974)

¥ SAN ANTON 10 CLAY LOAM,0.03 GHz (HIPP, 1974)

o PUERTO RICO CLAY LOAM,0.03 GHz (HIPP, 1974)

T GOODRICH CLAY, 0.5 GHz(HOEKSTRA AND DELANEY, 1974)

? SUFFIELDSILTYCLAY, 0.5 GHz(HOEKSTRA AND DELANEY, 1974)

0.0 0.1 0.2 0.3

SOIL MOISTURE (GRAMS PER CM3)

0.4

Figure 17. Relative dielectric constant values of soil as a function of volumetricwater content: Loam, clay; frequency 0.03 GHz ~ 0.5 GHz;imaginary part.

24

56

54

36

34*

24

£ 22a.

£ 2°

- 18o

i 16

I WH-CO

§12O

510I—oa 8LU

Q 6

4

2

>*

2- *

0*-^0.0

o? o f


* SAN ANTON 10 CLAY LOAM,0.03 GHz (HIPP, 1974)



? SUFFIELDSILTYCLAY, 0.5 GHz(HOEKSTRA AND DELANEY, 1974)

0.1 0.2 0.3 0.4SOIL MOISTURE (GRAMS PER CMJ)

Figure 18. Relative dielectric constant- values of soil as a functionof volumetric water content: Loam, clay; frequency 0.03GHz - 0.5 GHz; imaginary part.

25

26

24

22

LU0£

on

18

: 16

14o 10o 12

10 -

£ 8o

6 -

0

o YUMASAND, 1.1 GHz(LUNDIEN, 1971)

• YUMASAND, 1.5 GHz(LUNDIEN, 1971)

D OPENWOOD STREET SILT,1.1 GHz (LUNDIEN, 1971)

•OPENWOOD STREET SILT,1.5 GHz (LUNDIEN, 1971)

* SAN ANTON 10 CLAY LOAM,1.0 GHz (HIPP,1974)


f LONG LAKE CLAY,1.1 GHz (LUNDIEN, 1971)

V LONG LAKE CLAY,1.5 GHz (LUNDIEN, 1971) !

o *

BO T

o

D

O

m

o

0.0 0.1 0.2 0.3 0.4 0.5SOIL MOISTURE (GRAMS PER CM3)

Figure 19. RelaMve dielecrric constant values of soil as a function of volumetricwater content: Sand, loam, clay; frequency 1.1 GHz - 1.5 GHz;real part.

26

20

18Q_

>-16eel

O

112

110

o 8oo

o

o YUMASAND, 1.1 GHz(LUNDIEN, 1971)

• YUMA SAND, 1.5 GHz(LUNDIEN, 1971)

n OPENWOOD STREET SILT,1.1 GHz (LUNDIEN, 1971)

• OPENWOOD STREET SILT,1.5 GHz (LUNDIEN, 1971)

• SAN ANTON 10 CLAY LOAM,1.0 GHz (HIPP, 1974)


? LONG LAKE CLAY,1.1 GHz (LUNDIEN, 1971)

7 LONG LAKE CLAY,1.5 GHz (LUNDIEN, 1971)

SOIL MOISTURE (GRAMS PER CM)

Figure 20. Relative dielectric constant values of soil as a function of volumetricwater content: Sand, loam, clay; frequency 1.1 GHz - 1.5 GHz;imaginary part.

27

• MANCHESTER FINE SAND, 4.0 GHz(HOEKSTRA AND DELANEY, 1974)


• SAN ANTONIO CLAY LOAM,3.8 GHz (HIPP, 1974)

0 PUERTO RICO CLAY LOAM,3.8 GHz (HIPP, 1974)

^ GOODRICH CLAY, 4.0 GHz(HOEKSTRA AND DELANEY, 1974)

1 SUFFIELDSILTYCLAY, 4.0 GHz(HOEKSTRA AND DELANEY, 1974)

20 r


Figure 21. Relative dielectric constant values of soil as a function

of volumetric water content: Sand, loam, clay; frequency3.8 GHz - 4.0 GHz; real part.

28

C£.

14

O< 10

LO

Oooo:\—o

8



• SAN ANTON 10 CLAY LOAM,3.8 GHz (HIPP, 1974)



7 SUFFIELDSILTYCLAY, 4.0 GHz(HOEKSTRA AND DELANEY, 1974)

v?

0.0 0.1 0.2 0.3SOIL MOISTURE (GRAMS PER CM3)

0.4

Figure 22. Relative dielectric constant values of soil as a functionof volumetric water content: Sand, Ioam7 clay; frequency3.8 GHz - 4.0 GHz; imaginary part.

29

h-

Q_

25

h- "

H-CO

OO

oQCL\—Oa

16

14

12

10

8

6

H

* MANCHESTER FINE SAND, 10.0 GHz(HOEKSTRA AND DELANEY, 1974)

o AAANCHESTER FINE SAND, 26.0 GHz(HOEKSTRA AND DELANEY, 1974)

•FAIRBANKS SILT, 10.0 GHz(HOEKSTRA AND DELANEY, 1974)

o FAIRBANKS SILT, 26.0 GHz(HOEKSTRA AND DELANEY, 1974)

* CLAY, 10.0 GHz (MATZLER, 1970)* GOODRICH CLAY, 10.0 GHz

(HOEKSTRA AND DELANEY, 1974)v GOODRICH CLAY, 26.0 GHz

(HOEKSTRA AND DELANEY, 1974)? SUFFIELDSILTYCLAY, 10.0 GHz

(HOEKSTRA AND DELANEY, 1974)V SUFFIELDSILTYCLAY, 26.0 GHz

(HOEKSTRA AND DELANEY. 1974)

-* V f

0.1 0.2 0.3SOIL MOISTURE (GRAMS PER CMJ)

0.4

Figure 23. Relative dielectric constant values of soil as a function ofvolumetric water content: Sand, loam, clay; frequency10.0 GHz - 26.0 GHz; real part.

30

10 r

o^ 6o c^o g

- 2


a MANCHESTER FINE SAND, 26.0GHZ(HOEKSTRA AND DELANEY, 1974)


o FAIRBANKS SILT, 26.0 GHz(HOEKSTRA AND DELANEY, 1974)

• CLAY, 10.0 GHz (MATZLER. 1970)T GOODRICH CLAY, 10.0 GHz

(HOEKSTRA AND DELANEY. 1974)v GOODRICH CLAY, 26.0 GHz

(HOEKSTRA AND DELANEY, 1974)f SUFFIELDSILTYCLAY, 10.0 GHz

(HOEKSTRA AND DELANEY, 1974)V SUFFIELDSILTYCLAY, 26.0 GHz

(HOEKSTRA AND DELANEY. 1974)

I

0.0 0.1 0.2 0.3 0.4


Figure 24. Relative dielectric constant values of soil as a function ofvolumetric water content: Sand, loam, clay; frequency10.0 GHz - 26.0 GHz; imaginary part.

31

faster at low water contents than in Figures 2 to 14, and the effect

of adsorbed water is not shown. Hoekstra and Delaney (1974)

concluded that within the experimental error, the relaxation of

water in sandy soils was identical to that in clay soils; this is

also evident from data from Manchester fine sand, Fairbanks silt,

and Goodrich c|ay (Figure 21).

To summarize, evidence regarding the effect of adsorbed

water on the soil dielectric properties is not conclusive. Results

of Lundien (1971), Geiger and Williams (1972), and Hipp (1974)

suggests that within a given set of experimental conditions, such an

effect may be shown if "proper" soils are used; the latter condition

seems necessary because the dielectric constant has been found to

differ for two soils of the same texture (San Antonio clay loam and

Puerto Rico clay loam, 1.0 GHz, see Figure 19). The difference

between soils of various textures appears smaller if the dielectric

constant is expressed as a function of water content by volume.

Consequently, before more conclusive results become available,

the effect of adsorbed water on soil dielectric constant may be

considered negligible in the 1 GHz to 35 GHz frequency range insofar

as the interpretation of microwave remote sensing data is concerned.

(3) Values of k should increase for higher moisture contents and

ultimately should reach k^ values of water at the measured

frequency (Figure 1). Although soil dielectric constants are

rarely measured at moisture contents above 40 per cent, the

available data indicate trends toward the k, values of water.

An exception is measurements by Geiger and Williams

(Figures 8, 10, 14) which appear to level off at high moisture

contents.

Figure 16 shows average curves for real and imaginary

parts of the dielectric constant based on several measurements

of loamy soils at frequencies near 10 GHz. The broken lines

were drawn by extrapolating between experimental measure-

ments and the dielectric constants of water. Assuming that

the extrapolation is realistic, it can be seen that the real

32

part of the dielectric constant should increase nearly linearly

at higher soil moisture content. Lundien's (1966, 1971) and

Wiebe's (1971) measurements showed a similar relationship.

However, k . values of Geiger and Williams, Leschanskii

et al., (1971) and others tended to level off at higher moisture

contents. The discrepancy has not yet been resolved; it is

conceivable, however, that the soil/water mixing properties

and the tendency for separation at higher moisture content

could cause these differences. Anomer factor affecting the

shape of these curves is bulk density. As discussed in Section

3.2 , the upper portions of the curves would shift to the right

if moisture were measured in the volumetric units.

(5) The imaginary part of the dielectric constant kj was smaller

at most frequencies rhan the real part and also increased monotonically

with the increasing moisture content. Two exceptions to this

trend were found. First, measurements of Geiger and

Williams (Figures 9, 11, 15) showed an inflection point and

were of the same order of magnitude as k, curves. Secondly,

data of Leschanskii et. al., (1971) obtained at 0.13 GHz and

0.3 GHz exhibited relatively high V~ values for loam (Figure

5); in the extreme case (0.3 GHz, 20% moisture)^2 exceeded

k-.. In contrast,k« values for sand were ten times smaller

under otherwise identical conditions (Figure 3). This differ-

ence was explained by attenuating effect of ions present in

the soil exchange complex; the amount of ions retained in

loam is larger than in sand, and therefore attenuation is

higher (Leschanskii et. al., 1971). The difference in attenu-

ation was not detected at higher frequencies (centimeter

wavelengths) because the attenuation by molecules of water

predominates there and consequently the effect of soil mater-

ial is obscured (Leschanskii et0 al., 1971). Lundien (1971) also

demonstrated that conductivity (which is directly proportional

to k? at a given frequency) increased as the proportion of

fine soil particles increased from sand to clay for frequencies between

33

1.074 GHz and 1.499 GHz . Figure 18 shows that k~ was

much higher at 0.03 GHz than at 0.5 GHz; no such difference

was detected at higher frequencies for the same soils (Figure 22). .

Thus the ion exchange characteristics of soil material appear to

affect soil dielectric properties at low frequencies. However,

the magnitude of the effect at 0.3 GHz is more pronounced in

the data reported by Leschanskii et al., (1971) (Figure 5) compared

with Lundien's (1966, 1971). It should be noted that Hoekstra and

Delaney (1974) did not detect any ionic interaction at 0.5 GHz.

The relationship between the relative dielectric constant and soil moisture

has been quantified by Lundien (1971) for five frequencies (between 1.074 GHz and

1.499 GHz) in the following form:

m = 1 - Oz2.6- + 0 . 1 1 (grams per cm3) (9)

OUV

3.2 Bulk Density

Campbell and Ulrichs (1969) showed that the real part of the dielectric

constant k, of rocks occupied a narrow range of values when the bulk density was

constant (Figure 25). This range expanded considerably for the same rocks when

samples with different bulk densities (constant porosity) were prepared (Figure 26).

This difference could be due to either inherent dielectric properties of the various rocks

or change in bulk density PS . Figure 27 shows that k, increases with bulk density for

a given soil type; consequently, bulk density of a given soil or rock material affects

its dielectric constant. The curve in Figure 27 was calculated from Krotikov's

(1962, in : Peake et al., 1966) formula:

k,= 1 + -V (io)

Equation (10) was also found to hold for pumice at 10 GHz (Peake et al., 1966).

The trend shown in Figure 27is supported by results of Edgerton et al., (1971) who

noticed a decrease in dielectric constant with decreasing bulk density for low moisture

contents.34

co

9i 55

i§o ,_~5 ouO ^

o> .t:« E=* »o a.a: ^

If)

o•o!n 6

•= >

0CM

a - o * o

sslif^0 .0"

ne

a. >> vo t -g 50. w £ c

-•o ? Pm *

Oliv

Ser

pen Ano

e

oinoO

;««'•= an i

Lak

e C

ount

y , O

Gab

bro

, B

ytow

ni • •5 cT.-J--£« o i«-0 >0

1258^3^M o S I - -w ^ ^

Se

Bas

alt ,

chan

ging

S. .. _ ^

» 9-

>. y

11

° -s s

.t: §O u.j_

_o_ooo

>soo:

it

o o> tt> — co o>•^ *• ^- o o -^

i *

I

I -

0

•O *

•o

.t c TJ . i:J go

-oco

0)_QQ.

OU

*

8

8.vOtf^C

Oo

ff

-

§ 25° o .

s

--̂ a>o -S g

* o™ •»c JS0> 00

coy'M

oQ. O

2

wQ.

0)

O)

36

SAN ANTON 10 CLAY LOAM• 0.03 GHz• 1.0 GHzT 3.8 GHz

PUERTO RICO CLAY LOAMo 0.03 GHza 1.0 GHzA 3.8 GHz

MOISTURE CONTENT: 0%SOURCE: HIPP (1974)

CO I—z oco<O 0.

Oa

3

2

1

01.2 1.3 1.4 1.5 1.6 1.7

SOIL BULK DENSITY ps(GRAMS PER CM3)

Figure 27. Effect of bulk density on the relative dielectric constant.

37

Bulk density has an important indirect effect on the relationship between

dielectric constant and soil moisture content. Figure 28 presents some data for

Openwood Street silt (Lundien, 1971); it is apparent that the k, values measured

for this soil can be represented adequately by a single curve regardless of the degree

of compaction if expressed on the volume basis. When moisture contents are calculated

on the weight basis, then according to Eqs. (8) and (9), individual data points should2

shift to the left (usually, p > 1.0 g/cm ). This is demonstrated in Figure 28 for bulk^ 3

densities of 1.2 and 1.6 g/cm . As a result of the shift an artificial uncertainty is

introduced into the k, versus moisture relationship. For example, k, lies between3

approximately 12.0 and 14.0 for moisture content of 0.25 g/cm when the scatter

around the mean curve is taken into account. It contrast, k, varies between 14.0

and 21.5 for a moisture content of 25% if bulk density of the measured sample rangeso

from 1.2 to 1.6 g/cm .

The above uncertainty is included in data presented in Figures 2 through 15.

It cannot be removed unless bulk density values are also given. The intensity of

compaction added in some cases (e.g., Wiebe, 1971) is not sufficient because for a

constant compaction value, bulk density varies as a function of soil type and moisture

content; this is indicated in Figure 29 which was prepared from data by Lundien (1971).

Unfortunately, a large proportion of soil dielectric constant measurements available lacks

the bulk density information. The artificial uncertainty is introduced in a similar

manner into the kj versus moisture relationship.

3.3 Soil Type

The term soil type was previously used in soil survey to designate a subdivision

of soil series based on the particle size distribution of the surface soil (Soil Survey

Staff, 1951). In the Comprehensive Soil Classification System employed by the U.S.D.A.

since 1965, the designation soil type has been replaced by soil phase (Soil Survey Staff,

1960). The term soil type is still being used in the remote sensing literature. It should

be noted that the textural classes on the basis of which soil rypes have been defined

have not changed (Figure 30).

Results of various experiments indicate that at higher frequencies, the dielectric

constant is almost unaffected by the chemical and mineralogical composition of rocks

(Campbell and Ulrichs, 1969) or soils (Lundien, 1971) except where significant amounts

of metallic or magnetic minerals are present (Edgerton et al., 1971). Differences

among soil types observed at these frequencies (Lundien 1966, 1971; Wiebe, 1971) have

been attributed to soil water interactions which depend on the particle size distribution.

38

5 30o.

a;-20

"10

Oa!±i 0

SOIL: OPENWOOD STREET SILTFREQUENCY: 1.074 GHzCOMPACTION MOISTURE BY

(N/cm2)5. 7411. 8318. 47

VOLUME WEIGHToA

1.6g/cm3

ps = 1.2 g/cm3

0.0 0.1 0.2 0.3 0.4SOIL MOISTURE (GRAMS PER CM3)

l i i i i i i i i

0 10 20 30 40SOIL MOISTURE (PERCENT BY WEIGHT)

Figure 28. Change in relative dielectric constant due to soil moisture

units used.

39

o

UJQ_

1.8

1.6

1.4

YUMA SAND

g 1.6

1 M

5l.2

CO

LU1.0

5. 74 N/cm2 |11.83 N/cm2

18.47 N/cm2

SOURCE:LUNDIEN (1971)

OPENWOOD STREET SILT

1.6

S 1.4

1.2

1.0

0.8

LONG LAKE CLAY

I I

0 5 10 15 20 25 30 35 40 45 50SOIL MOISTURE (PERCENT BY WEIGHT)

Figure 29. Bulk density variation due to compaction as a function ofgravimetric water content.

40

A

Cco

CO

CCDUJ-

In some studies (e.g., Hoekstra and Delaney, 1974), differences between dielectric

properties of various soil types have not been observed. This topic was discussed

in some detail in Section 3.1 .

At low frequencies, soil type modifies the dielectric properties by its ionic

complex characteristics. In particular, the dielectric loss increases with increasing

proportion of.clay particles in the soil, due to presence of larger amounts of ions at

the exchange sites. From the measurements available, this effect appears to become

important below 0.3 GHz.

3.4 Temperature

The dielectric constant of solid soil particles is relatively independent of

temperature. For example, Compbell and Ulrichs (1969) found small differences at

35 GHz ( < 1 .0) between real parts of the dielectric constant of five rocks (rholeiitic

basalt, olivine basalt, quartz, aplite granite, dunite) measured at various temperature.s

(Figure 31). On the other hand, dielectric constant of water may be calculated

accurately as a function of temperature (Paris, 1969). Lundien (1971) developed such

an expression based on experimental data measured at L-band using both distilled and

tap water:

k1 = 88.6- 0.368T

where T is temperature in C.

Since soil almost always contains some water (unless oven-dried), its

dielectric constant depends on temperature. This has been confirmed experimentally

by Poe et al. (1971) who observed that the real part of the dielectric constant of

dry soil exhibited no change when temperature was raised from 20°C to 60 C but

such change occurred when a small amount of water was present.

3.5 frequency

Data presented in Figures 2 to 15 indicate little difference between the

dielectric constant values of dry soils at frequencies between 0.13 GHz and 37

GHz. Lundien's (1971) measurements at five frequencies between 0.01 GHz and

1 .499 GHz showed that the dielectric constant was approximately constant for

all soils and frequencies above 1 .0 GHz; the values increased below approxi-

mately 0.05 GHz, 0.15 GHz and 0.45 GHz for sand, silt, and clay, respectively.

42

These data are in agreement with the results measured for rocks: Campbell and

Ulrichs (1969) found that the dielectric constant values of a variety of terrestrial

rocks and minerals exhibited only small differences between two frequencies,

450 MHz and 35 GHz; materials that displayed differences tended to be the in-

homogeneous ones (Figure 32).

It should be noted again that the above results hold for dry materials; the

dielectric constant becomes frequency dependent when water is added. Hoekstra and

Delaney (1974) presented kj, k2, and loss tangent (kj/ k,) values for a silty clay

at 15% moisture content (Figure 33). Values above 0.1 GHz represent Suffield

silty clay (Figures 17 through 24) while data at lower frequencies were obtained from

a study by Smith-Rose (1933). The dielectric loss can be seen to be small around 0.1 GHz

but increases in both directions. At higher frequencies, the presence of water causes

increase in dielectric loss. The shape of the curves at lower frequencies will depend on

soil type.

4.0 REPRESENTATIVE DIELECTRIC VALUES

As evident from the previous section, a large number of dielectric constant

measurements by various investigators have been made. The data exhibit a scatter

which is due to both inherent soil variability and experimental error. To use these

measurements for analyzing microwave signal variations with moisture content,

however, only one representative data set is needed. Results of section 3.0 suggest

that the dielectric constant vs. soil moisture relationship should be given (i) in terms

of volumetric water content, and (ii) for various frequency/soil type combinations;

lower frequencies are more relevant because of the higher penetration depth.

Using data shown in Figures 17 through 24, average curves were drawn through

the points for both k, and k2 for three frequencies (1.3 GHz, 4.0 GHz and 10.0 GHz)

and three soil types (sand, loam, and clay). Values read off from the average curves

(Figures 34-36) are given in Table 1. The last column of Table 1 gives the primary

data source for the average curves; it shows that the three frequencies and soil types

represent a mean condition only.

43

10

8

€'

o Tholeiitic Basalt ; OreA Olivine Basalt ; Colo.O Quartz; N.H.V Aplite Granite ; Colo.a Olivine Peridotite(Dunite); N.C.

300 400 500 600 700 800Temperature , °K

900

Figure 31 . The Permittivity of rocks at 35 GHz as a function of temperature(From Campbell and Ulrichs (1969)).

44

1 X2 Oo mm 10

o

N

O

-oaNX

_*OO

o -

oI/I0)

'> ^

.•t •

10 5

10 4

103

10 2

101

10°

ID'1

10 -2

I I I • I I I

\ Imaginary Part

Loss Tangent \

I I I I

io3 io4 io7 ~ io8 io9 io'° 10"Frequency, Hz

Figure 33. Real and imaginary part of the relative dielectric constant as a functionof frequency for a silty clay, moisture content 15% by weight. (FromHoekstra and Delaney (1974)).

46

W

.2'o0)

*J*£u_«2

_ct—

o1*.in0)

"5

•c

couL.

0)

Qai*w

•*— i/i

^~ Q.

—

O fl.C (U® i:in -G

Q.-D0) C

/v O

_•

~O

1—

0)uL.

oin

1 £co u

2 So 2-

Q §T>» i

™ *2C^ *-— -

ô

"5) ^_

P O CM« Q. -^

"5 t:0) O —

Q^ Cu ^

£^^ ^

• - " S oO ° \

0)~ Q.O X

«/> t—

UC N4) T*

a- 0£i

^

2^>— 'c

cD

,_T^J

oo

>

00

co -̂ r

8S00

ôto

CO

,_r

^^r[

^|1|i ^ ...

^^ f\ ^^••• ^^

î ^ ^^ ̂ ™

^^ *^^« ^^• ^^

I- E g-

S ^0*i °~3

ill^ G —

Q_ C 01

CJ Ô ^^

CO ^0 *̂ ^5 CO C^ CO *O '̂ CM Ô GO CO C^

O O O •— -— O O O O -— — — CM CM

C ĵ QO ̂ 3 Ô — *" *"^ ^^) ^^ Ô OO Ô ^^^ ^^^ ^^1

r̂t GO ^^ l̂ ^ O^ CO ^^ Ô ^^ ^^ C^J *O tF~lj ^^^ 12̂ :;:. ^ ^.^^(^

O*OOiT)O OiOOiOO»OOtr)Oi— •— CM CM CO O O — •— CM CM CO CO •̂ ~

ooooo ooooooooo

Eoo

47

^^

IIIQ- o

0)ocooQ

oL.

|u.a.

o

*o>o'""'

"o

"oCO

*oco

uQ)3

£u.

'sruDOM

^XUc0)

a-£J;

^""5>!2.

L.

O CMa. _*

0 f—O- -S£

£

o ,w

s.

NI0

gl— -

>»_^

^^-^C P^ hjo o^ gjQ) ' — '^,

O. °-

§ X

E«>

III°" o" >-••• ° ~o£.2 o« c _o"c •§ ̂O ^f Oo ^^ ^~

7= c aO D 3u_ t/> a.

O O O - — CNCNCOCO-*

coco-^-'OOC)'— inocof— t— i— CM

^3 l^^ ^3 Ô *iJ Lrt ^^ IjO ^^

^3 ^Tj3 ^^ *^™ ^^J C^4 ^5 CO ^J

oooooo'ooo

j

j?^~E^fc^C- a)cX O

o Q ^f KÔ T3 oJ ^

o ^-Il.llo -S in ^^ Q *«ku> O rr^^'

•^ ̂ *co

X ^t c'' M"

^^ ^^ O*^ ci O ^^0 iL"U °-5^0.2

-c '« .iJ gÔ-o Q£ ii

L» ™J" Q J™^D *» ^^

o t a) c,X 3 => 0x^ I/) Û (/)

•— inOincNooinco-— ooo-— •— CMCMco-^- in in

coo«ooo-— co-— o^tvincoînôo-CNôo-coiv .

— — •— CM CM

oinoinoinoinoinOO-— •— C M C N C O C O - ^ t T Toooooooooo

0

u

o

A ^k

c>:^0)c_o0)

-oc00̂

I/I

1X

c?0

• *»

coa>c

1£a>

-Cun1

•~ tx ^ co co m coo o — CM co-î- m

•* >o «— o o o •—CN co in K co •— co

o in o in o m oo o •— •— CM CN co0 0 O 0 OOO

•D

0CO

o0

ĉ}:J™"

^ÎS)c_oVo

oo1-

0)o

6;o•^

«̂n

co.tou_

CM «O •— OO CO O- O •— CM

o o •— •— CM co in o^

\

•^••^J-f^-»tCOCN CMCMCMC M C O - ^ - O C O O CN^M3

Oino inom o m oO O -- — CM CM COCO-*

O O O O O O O O O

0o

48

J)u0>

£o

0)

o1oU

uJ)0)

Oo

y D

ata

Base

frequency

, so

urc

4)

0) TJ

O

11

^ ^^

..— ^

^ v»

-? 8.(0 —

0°°2 4)C Q)

Q.-00) C

Q£ O

-OO

Since the representative dielectric constant curves shown in Figure 34-36

wore related to volumetric moisture, the large number of measurements shown in

Figure 2 through 15 could not be used. It was of interest to determine, however,

whether the two types of dielectric values are similar for equivalent conditions.

To this end, average k, and k« values were extracted from Figures 2 through 15 for

two frequencies (1.3.GHz and 10.0 GHz) and three soil types (sand, loam, and

clay). The average curves were based primarily on the data by Leschanskii, et al.

(1971), Lundien (1971), Wiebe (1971) and Matzler (1970). Secondly, skin depth

was computed as a function of moisture for every frequency/soil type combination;

skin depth was used for the comparison because it reflects the effect of both k and

\

26

24

22

20i—

118

8 160

LU

512

8

6

4

2

FREQUENCY: 1.3 GHzSOIL TYPE:

SANDLOAMCLAY

28

26^ \j

24

22

20

r

-

-

FREQUENCY: 4.0 GHz

SOIL TYPE:SANDLOAM

_ _ _ _ _ _ P I A VLLAY

ti§////

////

////

///r

S18I 16

£ 14oUJ

^ 12Q

jgio

S 8

6

4

0

REAL PART

IMAGINARYPART

0.0 0.1 0.2 0.3 0.4


0.5

Figure 35. Representative dielectric constant values ess a function of volumetricwater content for sand, loam, and clay at 4.0 GHz.

52

20

18

z 16

£ 14ooo 12

5 10—iUJ

o 8LU

;= aLU

_____

_ _ _ _ _

SANDLOAMCLAY

1.3 GHz

0.001

SOIL MOISTURE (GRAMS PER CM)

Figure 37. Skin depth as a function of volumetric water content, frequency,and soil type.

54

coefficient changed from 0.08 (-11.0 dB) for a dry soil to 0.45 (-3.5 dB) at3 30.45 g/cm , an average increase of .0082 for every 0.01 g/cm . Emissivity

changed in the opposite direction (Eq. 3). Figure 38 thus suggests that the

proportion of microwave energy normally incident on a smooth, homogeneous

uniformly moist soil which is reflected from the soil is practically independent

of frequency and soil type.

55

^0.6LU

^0.5u_

SO-"

So.3oLU

gO.2

II. Iit 1 1

§; i

O *CL.0.0

0

-

• 1.3 GHz,• 4.0 GHz,* 10.0 GHz0 1.3 GHz,° 4.0 GHz,A 10.0 GHz

A 1.3 GHz,V 4.0 GHz,* 10.0 GHz

• 1• ¥i

fc *k

.0 0.1 0.2

SANDSAND, SAND

LOAMLOAM

, LOAM

CLAYCLAY

, CLAY

* I_ f X• 1 4

*

-

-

0. 3 0. 4 0.

0.4

0.5

0.6^

0.7 ^to

0.8S

0.9

1.0


Figure 38. Power reflection coefficient and emissivity as a function of volumetricwater content, frequency, and soil type.

56

APPENDIX A.

BIBLIOGRAPHY

Armand, N. A., A. E. Basharinov, L. F. Borodin, and A. M. Shutko, "MicrowaveRadiation Properties of Thermal and Moist Land Areas," Proceedings of theURSI Specialist Meeting, September, 1974, Bern, Switzerland.

Basharinov, A. E., A. S. Gurvich, S. T. Yegorov, A. A. Kurskaya, D. T. Matvyevand A.M. Shutko, "The Results of Microwave Sounding of the Earth's SurfaceAccording to Experimental Data from the Satellite Cosmos 243," Space Research,vol. 11, Akademie-Verlag, Berlin, 1971.

Basharinov, A. E., L. F. Borodin, and A. M. Shutko, "Passive Microwave Sensingof Moist Soils," Proceedings Ninth International Symposium on Remote Sensingof Environment, University of Michigan, Ann Arbor, April, 1974.

Blinn, John C., III. and Jack G. Quade, "Dependence of Microwave Emission onMoisture Content for Three Soils," Proceedings 1973 International IEEE"GAP& USNC/URSI Meeting, Boulder, Colorado, August, 1973.

Cihlar, J., "Soil Moisture Determination by Microwave Measurements," RSL TechnicalMemorandum 177-38, (Internal Working Paper Only), University of KansasCenter for Research, Inc., Lawrence, Kansas, March, 1973.

Cihlar, J., "Ground Data Acquisition Procedure for Microwave (MAPS) Measurements,"RSL Technical Memorandum 177-42, (Internal Working Paper Only), Universityof Kansas Center for Research, Inc., Lawrence, Kansas, July, 1973.

Cihlar ,J., "Ground Data Acquisition for 1973 Microwave (MAPS) Measurements:Results," RSL Technical Memorandum 177-44, (Internal Working Paper Only),University of Kansas Center for Research, Inc., Lawrence, Kansas, December,1973.

de Loor, G. P., "Radar Ground Returns Part III: Further Measurements on the RadarBackscatter of Vegetation and Soils," Physics Laboratory TNO, Report No.PHL-974-05, The Hague, The Netherlands, March, 1974.

Dickey, F., C. King, J. Holtzman and R. K. Moore, "Moisture Dependency ofRadar Backscatter from Irrigated and Non-Irrigated Fields at 400 MHz and13.3 GHz," IEEE Transactions on Geoscience Electronics, vol. GE-12,no. 1, pp. 19-22, January, 1974.

Eagleman, J. R., "Moisture Detection from Skylab," Proceedings Ninth InternationalSymposium on Remote Sensing of Environment, University of Michigan,Ann Arbor, April, 1974.

57

Eaglemon, J. and F. T. Ulaby, "Radiometer-Scatterometer Soil Moisture Detection,"Proceedings of the A me ri con Astronomicol Society Meeting, August, 1974,Los Angeles, California.

Edgerton, A. T., "Engineering Applications of Microwave Radiometry," ProceedingsFifth International Symposium on Remote Sensing of Environment, Universityof Michigan, Ann Arbor, 1968.

Edgerton, A. T., R. M. Mandl, G. A. Poe, J. E.Jenkins, F. Soltis and S. Sakamoto,"Passive Microwave Measurements of Snow, Soils, and Snow-Ice-Water Systems,"Technical Report No. 4, Aerojet General Corporation, El Monte, California.

Edgerton, A. T., F. Ruskey, D0 Williams, A. Stogryn, G. Poe, D. Meeks and O.Russell, "Microwave Emission Characteristics of Natural Materials and theEnvironment," Final Technical Report 9016R-8, Aerojet General Corporation,El Monte, California.

Jean, B. R., C. L. Kroll, J. A. Richerson, J. W. Rouse, Jr., T. G. Sibley andM. L. Wiebe, "Microwave Radiometer Measurements of Soil Moisture," TechnicalReport RSC-32, Remote Sensing Center, Texas A&M University, College Station,Texas, 41 p.

Kondratyev, K. Ya., V.V. Melentyev, Yn. I. Rabinovich, and E. M. Shulgina,Doklady, AN SSSR, vol. 208, no. 2, 1973.

Kondratyev, K. Ya., E. M. Shulgina, O. M. Pokrovskiy, and Yn. M. Timofeev.Trudy, GGO, 295, pp. 86-97, 1973.

Kroll, C. L., "Remote Monitoring of Soil Moisture Using Microwave Radiometers,"Texas A&M University, Remote Sensing Center, College Station. Texas,1973, 127p.

Poe, G. A. and A. T. Edgerton, "Determination of Soil Moisture Content withAirborne Microwave Radiometry," Final Technical Report 4006R-1, AerojetGeneral Corporation, El Monte, California, 1971.

Poe , G. A., "Remote Sensing of the Near-Surface Moisture Profile of SpecularSoils with Multi-Frequency Microwave Radiometry," Proceedings SPIE,vol. 27, November, 1971. B

Poe, G., A. Stpgryn, and A. T. Edgerton, "Determination of Soil Moisture ContentUsing Microwave Radiometry," Final Technical Report 1684-1, AerojetGeneral Corporation, El Monte, California.

Popov, A. Ye., E. A. Sharkov, and V. S. Etkin, "Characteristics of radiationfrom wet soils in the microwave region," Meteorol. and Gidrol. (USSR),no. 10, pp. 49-57, 1974.

Rouse, J. W., Jr., "On the Effect of Moisture Variations on Radar Backscatrerfrom Rough Surfaces," Technical Memorandum RSC-52, Remote SensingCenter, Texas A&M University, College Station, Texas, 10p., 1972.

58

Rouse, J. W.. Jr., R. W. Newton and S. L. Lee, "On the Feasibility of MonitoringSoil Moisture with Microwave Sensors," Proceedings Ninth InternationalSymposium on Remote Sensing of Environment, University of Michigan,Ann Arbor, April, IV/4.

Schmugge, T., P. Gloersen and T. Wilheit, "Remote Sensing of Soil Moisture withMicrowave Radiometers," Goddard Space Flight Center, NASA, Greenbelt,Maryland, 32p., Preprint X-652-72-305, August, 1972.

Schrnugge, T., P. Gloersen, T. Wilheitand F. Geiger, "Remote Sensing of SoilMoisture v/ith Microwave Radiometers," Journal of Geophysical Research,vol. 79, no. 2, pp. 317-323, 1974.

Stiles, W. H., "The Effect of Surface Roughness on the Microwave Measurementof Soil Moisture Content," M. S. Thesis, University of Arkansas, Fayetteville,Arkansas, 1973.

Ulaby, F. T. and R. K. Moore, "Radar Sensing of Soil Moisture," Proceedings 1973International IEEE-GAP & USNC/URSI Meeting, Boulder, Colorado, August, 1973.

Ulaby, F. T.r "Radar Measurement of Soil Moisture Content, "IEEE Transactions onon Antennas and Propagation, vol. AP-22, no. 2, March, 1974.

Ulaby, F. T., J. Cihlar and R. K. Moore, "Active Microwave Measurement of SoilWater Content," Remote Sensing of Environment, vol. 3, pp. 185-203, 1975.

Ulaby, F. T., "Vegetation and Soil Backscatter Over the 4-18 GHz Region," Proceedingsof the URSI Specialist Meeting, September, 1974, Bern, Switzerland.

Ulaby, F. T. , "Soil Moisture Detection by Skylab's Microwave Sensors," Proceedings ofthe URSI Specialist Meeting, September, 1974, Bern, Switzerland"!

Ulaby,F., T., T. F. Bush, P. P. Batlivala and J . Cihlar, "The Effects of Soil Moistureand Plant Morphology on the Radar Backscatter from Vegetation," RSL TechnicalReport 177-51, University of Kansas Center for Research, Inc., Lawrence,Kansas, July, 1974.

59

REFERENCES

Birchak, J. R., C. G. Gardner, J. E. Hipp and J. M.. Victor, 1974, "HighDielectric Constant Microwave Probes for Sensing Soil Moisture,"Proceedings of the IEEE, vol. 62, no. 1, pp. 93-98.

Campbell, M. J. and J. Ulrichs, 1969, "Electrical Properties of Rocks and TheirSignificance for Lunar Radar Observations," Journal of GeophysicalResearch, vol. 74, pp. 5867-5881.

Edgerton, A. T., R. M. Mandl, G. A. Poe, J. E. Jenkins, F. Soltis, andS. Sakamoto, 1968, "Passive Microwave Measurements of Snow, Soils,and Snow-Ice-Water Systems," Technical Report No. 4, Aerojet GeneralCorporation, El Monte, California.

Geiger, F. E. and D. Williams, 1972, "Dielectric Constants of Soils at MicrowaveFrequencies," Goddard Space Flight Center, Greenbelt, Maryland, 30p.

Grim, R. E., 1968, Cloy Mineralogy, McGraw-Hill Book Co., New York.

Hidy, G. M., et al., 1972, "Development of a Satellite Microwave Radiometer toSense the Surface Temperature of the World Oceans," NASA CR-1960,North American Rockwell Corporation, Downey, California.

Hipp, J. E., 1974, "Soil Electromagnetic Parameters as a Function of Frequency,Soil Density, and Soil Moisture," Proceedings of the IEEE, vol. 62, no. 1,pp. 98-103.

Hoekstra, P. and A. Delaney, 1974, "Dielectric Properties of Soils at UHF andMicrowave Frequencies," Journal of Geophysical Research, vol.79, no. 11,pp. 1699-1708.

Leschanskii, Yu. I., G. N. Lebedeva and V. D. Schumilin, 1971, ("ElectricalParameters of Sandy and Loamy Soils in the Range of Centimeter, Decimeterand Meter Wavel engths). Izvestiya Vysschich Uchebnich Zavedenii,Radiofizika, vol. 14, no. 4, pp. 562-569.

Lundien, J. R., 1966, "Terrain Analysis by Electromagnetic Means," TechnicalReport No. 3-693, Report 2, U. S. Army Engineer Waterways ExperimentStation, Vicksburg, Mississippi, 55p.

Lundien, J. R., 1971, "Terrain Analysis by Electromagnetic Means," TechnicalReport No. 3-693, Report 5, U. S. Army Engineer Waterways ExperimentStation, Vicksburg, Mississippi, 33 p.

60

Marzler, C., 1970, "Messung der Wormeslrahlung der Erdoberflache imMikrowellengebiet," Institute fur Angewandre Physik, Universirat Bern,76P.

Paris, J. f.r 1969, "Microwave Radiomelry and Its Application to Marine Meteorologyand Oceanography," Department or Oceanography, Texas A&M University,210p.

Peake, W. H., et al., 1966, "The Mutual Interpretation of Active and PassiveMicrowave Sensor Outputs," Proceedings Fourth Symposium on Remote Sensingof Environment, University of Michigan, Ann Arbor, pp. 7/1-777.

Poe, G., A. Stogryn and A. T. Edgerton, 1971, "Determination of Soil MoistureContent Using Microwave Radiometry," Final Technical Report 1684-1, AerojetGeneral Corporation, El Monte, California.

Russell, M. B., 1939, "Soil Moisture Sorption Curves for Four Iowa Soils," Soil ScienceSociety of America Proceedings, vol. 4, pp. 51-54.

Smith-Rose, R. L., 1935., "The Electrical Properties of Soil at Frequencies Up to 100 MHz;with a Note on the Sensitivity of Ground in the United Kingdom," ProceedingsPhys. Soc., London, vol. 47, p. 923.

Ward, S. H., G. R. Jiracek and W. I. Linlor, 1969, "Some Factors Affecting Electro-magnetic Detection of Lunar Subsurface Water," IEEE Transactions on GeoscienceElectronics, vol. GE-7, pp. 19-27.

Wiebe, M. L., 1971, "Laboratory Measurements of the Complex Dielectric Constant ofSoils," Technical Report RSC-23, Texas A&M University, College Station,19p.

61

dielectric properties of soils as a function of …€¦ · 3 figure 25. the permittivities of...

Documents