WATER RESOURCES RESEARCH,
VOL. 16, NO. 6, PAGES 961-979, DECEMBER
1980
Surveyof Methodsfor Soil MoistureDetermination
T. J. SCHMUGGE
NASA GoddardLaboratoryfor AtmosphericSciences,GoddardSpaceFlight Center,Greenbelt,Maryland 20771
T. J. JACKSON
U.S.Department
ofAgriculture,
Science
andEducation
Administration,
Agricultural
Research,
Hydrolog),
Laboratory
Plant Physiology
Institute,Beltsville,
Maryland20705
H. L. McKIM
U.S. Army Corpsof Engineers,
ColdRegions
Research
Engineering
Laboratory
Hanover,New Hampshire03755
Increasedinterestin soilmoistureinformationfor applicationsin suchdisciplinesashydrology,meteorology,and agriculture
hasnecessitated
an overviewof bothexistingand proposedmethodsfor determinationof soilmoisture.This paperthusdiscusses
methodsof in situsoilmoisturedetermination
includinggravimetric,
nuclear,andelectromagnetic
techniques;
soilphysics
modelsthattrackthe behavior
of water in the soil in responseto meteorological
inputs(precipitation)and demands(evapotranspiration);andremotesensing
approaches
thatusethesolar,thermalinfrared,andmicrowave
portionsof the
electromagnetic
spectrum.
The abilitiesof theseapproaches
t6 satisfyvarioususerneedsfor soilmoisture
informationvaryfrom applicationto application.Thereforewe haveproposeda conceptual
schemefor
mergingtheseapproaches
into an integratedsystem.This systemshouldprovidesoilmoistureinformation havingthe potentialfor meetingthe requirementsof variousapplications.
INTRODUCTION
The moisturecontentin the surfacelayers of the soil is of
great importanceto the disciplinesof agriculture,hydrology,
and meteorology.In the field of agriculture,Idso et al. [1975a]
describedthe need for soil moistureinformation for many diverseapplications,includingimprovedyield forecastingand
irrigation scheduling.In hydrology,the moisturecontentof
the soil surfacelayer is importantfor partitioningrainfall 'into
its runoff and infiltration components;while in meteorology,
soil moisturedeterminesthe partitioningof net radiation into
latent and sensibleheat components.Also, recentmodel studieshave indicatedthe importanceof soil moistureto the study
of suchdiversephenomenaas desertification[Charhey,et al.,
1977;Idso, 1977; Walker and Rowntree,1977] and the central
Florida sea breeze [Gannon,1979].
The soil layer consideredin all of thesedisciplinesis that
which can interactwith the atmospherethroughevapotranspiration (ET), i.e., the soil root zone;and the moisturecontent
of this layer fluctuatesin responseto precipitationand potential evapotranspiration
(PET). The thicknessof this layer dependsupon the type and stageof the soil'splant cover,but it
is typically about 1 or 2 m. Thus we will call the moisture
ing of soilmoisturethat led to thissurvey.Severalin situmeasurementtechniqueswill be describedin this paper, gravimetric, nuclear, electromagnetic, tensiometric, and
hygrometric.
IN
SITU
METHODS
GravimetricTechniques
The oven-dryingtechniqueis probably the most widely
usedof all gravimetricmethodsfor measuringsoil moisture
and is the standard for the calibration of all other soil mois-
ture determinationtechniques.This methodconsistsof oven
dryinga soilsampleat 105øCuntil a constantweightis obtained. Usually, this weight is obtainedwithin 12 hours,but
for largesamplesthe dryingtime increases.
The wet weightof
the soil sampleis taken beforeovendrying.The amountof
waterin the samplecan be determinedand the moisturecontent calculatedand expressed
as a percentageof the dry soil
weight.If the volumetricwatercontentis required,the gravimetricvalue is multipliedby the bulk densityof the soil:
0-- w•Y•loo%
(1)
storedin thislayer,soilmoisture.
Thismoisture
isonly0.005,,%where
of the total water on the earth's surface [Nace, 1964]; but its
seasonal variation
accounts for a 1.4-cm variation
0
in sea level
[Mather, 1974].
In this paperwe presenta surveyof the generalmethodsfor
determining soil moisture. The three general approaches
which we will considerare in situ or point measurements,
soil
volumetric water content, %;
Ww weight of water, g;
Wd dry weight of soil, g;
Yd oven-drybulk density,g/cm3;
Yw densityof water,g/cmL
watermodels,
andremote
sensing.
Theremote
sensing
ap- Thereareseveral
advantages
anddisadvantages
to the
proaches
areprobably
theleast
familiar
tothereaders
ofthis oven-drying
gravimetric
procedure.
Some
advantages
are(1)
journal,
andtherefore
wewillgivethem
greater
emphasis.
It samples
canbetakenwithanauger
or tubesampler,
(2)
wasthedevelopment
ofpossible
methods
fortheremote
sens-sample
acquisition
isinexpensive,
and(3)soilmoisture
conThispaper
isnotsubject
toU.S.copyright.
Published
in 1980
by tentiseasily
calculated.
theAmerican
Geophysical
Union.
Paper number 80W0815.
Somedisadvantages
are (1) obtainingrepresentative
soil
961
962
SCHMUOOE ET AL: SOIL MOISTURE DETERMINATION
moisturevaluesin a heterogeneoussoil profile is difficult and
(2) becausesamplesare required over long periodsof time to
monitor moisturemovement or amount over time and space,
this method can be very destructiveto the site. Additional information on this procedure,as well as mostof the othersdiscussed,can be found in the works by Brakensieket al. [1979],
Black [1965], and Reynolds[1970a, hi.
Nuclear Techniques
propertiesof the soil [Wilson, 1971]. F¾svalingam
and Tandy
[1972] also found that vehicularignition noisegreatly influencedthe neutronprobereadings,probablythroughits influence on the readoutelectronics,rather than the principleof
operation.
In laboratorycalibrationthe volume of soil usedshouldbe
large enoughto be consideredeffectivelyinfinite relative to
the neutron flux. Manufacturersof probessupplya generalized calibration curve with each unit. However, if an accurate
Neutron scattering. The neutron scatteringmethod is an
indirect way of determining soil moisture content. This
method estimatesthe moisturecontentof the soil by measuring the thermal or slow neutron density.Initial development
of the neutronprobebeganin 1950[Belcheret al., 1950, 1952].
Gardner and Kirkham [1952] defined the principles that the
moisturecontentdeterminationis desired,the probe should
be calibratedfor each soil type. Procedureshave been developed for laboratory and field calibration [Douglass,1966;
King, 1967; Luebs et al., 1968; Vachaudet al., 1977; Cohen,
1964;Lal, 1979].
The moisturecontentvalue representsan averageover a
methodis actuallybasedupon.Neutronswith high energy(a volume of soil. Therefore in laboratory calibration the soil
million electronvolts or more) are emitted by a radioactive used shouldbe homogeneous
in texture, structure,density,
sourceinto the soil and are sloweddown by elasticcollisions and moisturecontent [Belcheret al., 1950; Douglass,1966;
with nuclei of atoms and becomethermalized.The average Van Bavel,1961,1962].Field calibrationof the neutronprobe
energylossis muchgreaterthan neutronscollidingwith atoms is reportedlyextremelydifficult [Lawlesset al., 1963;Stewart
of low atomicweight (in soils,this is primarily hydrogen)than and Taylor, 1957;Rawlsand Asmussen,1973]but may yield a
from collisionswith heavier atoms.As a result,hydrogencan more representativecalibrationcurve than laboratorymethslow fast neutronsmuch more effectivelythan any other ele- ods.
ment presentin the soil. The densityof the resultantcloud of
No matter what type of calibration is used, all electrical
slow neutrons is a function of the soil moisture content in the
equipment has the potential to drift. Therefore standards
liquid, solid,or vapor state.The numberof slowneutronsre- shouldbe usedto periodicallyrecalibratethe probe.Various
turning to the detectorper unit time are counted,and the soil reca!ibrationprocedureshave been reported [Churayevand
moisturecontentis determinedfrom a previouslydetermined Rode, 1966; Long and French, 1967; Marais and Smit, 1958,
calibration curve of counts versus volumetric water content.
1962; Holmes, 1966; Olgaard and Haahr, 1968; Luebs et al.,
Two typesof neutronprobeshave beendeveloped.One is a
1968;Stoneet al., 1966;Stewartand Taylor, 1957; Ursic,1967;
depth probe that is lowered into the soil through an access Bowmanand King, 1965;Bell and Eeles, 1967].
tube to the depth at which the moisture content is desired.
The sphereof influenceof the neutronprobe measurement
The other is a surfaceprobe that givesthe moisturecontentof is the volumeover which the averagemoisturecontentis calthe top few centimetersof soil.
culated and dependson the amount of moisture in the soil.
Several sourcesof high-energy neutrons have been used. Van Bayel et al. [1956] and Glasstoneand Edlund [1957] deThe americium-beryllium
(Am-Be)sourceseemsto be the one fined the sphereof influenceas that volume which contains
mostwidely used[Bell and McCulloch, 1966].Older unitsused 95% of all the thermal neutrons.This concepthas been critia radium-beryllium (Ra-Be) source. l/an Bavel and $tirk
cizedby Mortier et al. [1960]and Olgaard[1965].They suggest
[1967]found that the Am-Be sourceeliminatedgammaradia- that the sphereof importanceis the one which, if all the soil
tion, decreasedthe probeweight,increasedthe countrate, and and water outsidethe spherewere removed,would yield a
possiblyincreasedthe depth resolutionof the soil moisture neutron flux at the source that is 95% of the flux obtained in
measurement.
an infinite
The strengthof the sourcevarieswith the type and manufacturer. l/an Bavel[1962]found that 1 or 2 millicuries(mCi)
of a Ra-Be sourcewas adequate.The strengthof the sourceof
Am-Be that l/an Bavel and $tirk [1967] used was 150 mCi.
Others [Long and French,1967;Bell and McCulloch, 1966]reportedusingAm-Be sourcesof 10, 30, 50, and 300 raCi.
If subsurface
measurements
are required,the neutronprobe
must be placedin an accesstube which may be closedat the
bottom. The size and compositionof the tube can affect the
resultant neutron density [Stolzy and Cahoon, 1957]. The
The volume of soil measuredis very importantwhen measuringsoil moisturewith depth.In many studiesthe diameter
of the sphereof influencecannot be easily relatedto depth
resolutionbecauseof the heterogeneitywith soil depth. The
vertical resolutionis critical to many studies,especiallythose
dealingwith monitoringsoil moisturein time and space.
method
most used to install the tube in the field is that of
drilling a slightlyundersizedhole and tampingthe accesstube
into the drilled hole to ensurea tight fit.
The accuracyof the neutronprobecan be determinedfrom
the deviationcalculatedby regression
analysis,whereneutron
counts are converted to volumetric moisture content [Visvalingamand Tandy, 1972].The calibrationdependsupon the
strengthof the source,the nature of the detector,the geometry
of the sourceand detectorin the probe [McCauleyand Stone,
1972], the materialsused to constructthe probe, the size and
compositionof the accesstube, and the physicaland chemical
medium.
The advantages
of the neutronprobeare'(1) moisturecan
be measuredregardless
of its physicalstate,(2) averagemoisture contentscan be determinedwith depth, (3) the system
can be interfacedto accommodateautomaticrecording,(4)
temporalsoil moisturechangescan be easilymonitored,(5)
rapid changesin soil moisturecan be detected,(6) readings
are directlyrelatedto soilmoisture,and (7) measurements
can
be made repeatedlyat the samesite.
The disadvantagesare (1) inadequate depth resolution
makes measurement
of absolute moisture content difficult and
limits its usein studyingevaporation,infiltration, percolation,
and placementof the phreaticwater surface;(2) the moisture
measurementdependson many physicaland chemicalproperties of the soil which are, in themselves,difficult to measure;
(3) care must be taken to minimize health risks;and (4) the
SCHMUGGE ET AL: SOIL MOISTURE DETERMINATION
963
sphereof influenceof thedepthprobedoesnot allowaccurate due to differencesin Yd and 6tof a soil within the beam of a
dualenergysystem
causedlargemeasurement
errors.Nofziger
measurement of soil water at or near the soil surface.
Stoneet aL [1966]statedthat the accuracyof neutron probe
measurementsexceedsthat of standardtechniques,but Stewart and Taylor [1957]arguedthat it is slightlyinferior. If the
[1978]concluded
fromhisstudies
that,indeed,largeerrorin
the measurementof Yd and 6tcan occur in highly stratified
materialwhenusingthe dual gammatechnique.Generally,
neutronprobeis used,the purchasershouldlook for a stable, small errors occur if Yd and 6t changelinearly in the colliportable,durablemodel with stableelectronics
and power mated beam. He also confirmed that both the dual gamma
accuratelymeasurethe averagewaterconcomponents,
compatiblewith availableequipment[Bell, 1969; and singlesystems
Bell and McCulloch, 1966; Zuber and Cameron, 1966].
Gamma ray attenuation. The gamma ray attenuation
methodis a radioactivetechniquethat can be usedto deter-
tent in the collimated beam if the bulk density of the soil is
constant.However, the averagewater contentin the beam
maynotrepresent
thewatercontentat themiddleof thecolli-
mated beam and in the middle of the presenttime period.
From this study,graphswere preparedto estimatethe error
are relatedto the densityof matterin their path and that the due to inhomogeneityof the soil.
A major problem in many cold regionsis the inability to
specificgravityof a soil Ydin (1) remainsrelativelyconstant
as the wet densitychangeswith moisturecontentincreasesor measure in situ water conditions when the soil is freezing,
or frozen.Goitetal. [1976]conducted
studies
to evaldecreases.Changes in wet density are measured by the thawing,
mine soil moisture content within a 1- to 2-cm soil layer. This
methodassumesthat scatteringand absorptionof gammarays
gammatransmission
techniqueand the moisturecontentdetermined from this density change.
Gamma raysmay be collimatedto a narrowbeam,which
permitsa representative
readingto be obtainedat any position in the soil. Gurr [1962],Fergusonand Gardner[1962],Davidsonet al. [1963],and Dmitriyev[1966]wereinstrumentalin
developingthe theoreticalbasisand procedures
for its use.
uate attenuation of a dual gamma beam and found this a
powerful techniquefor investigatingthe swellingphenomena
associatedwith freezing soil. They found that errors resulted
when attenuationequationsdevelopedfor homogeneousmixtures were applied to stratifiedmedia. Nofziger [1978] determined that the errorsin 6tand Yddue to nonuniform soil systems must be consideredto establishthe overall accuracyof
Basicequipment
includes
a gammaray source
Surrounded gamma ray measurements.
Sinceattenuationof gamma raysis independentof the state
by a collimator,a detectorwith a collimator,and a scaler.
Gurr[1962]useda 25-mCi •37Cs
sourcewith a leadcollimator. of the water in the material tested, the measurement of attenuation is unaffectedby the transition of liquid water to ice.
Therefore the use of gamma attenuation has an advantage
lead collimatorhaving a 12.5-ramdiameterhole. Mansell et sincemeasurementsof dry bulk density and total water conaL [1973]statedthat collimatedradiationfrom 300 mCi each tent (including ice), in grams per cubic centimetercan be
of 24•Amand •37Cs
provideda high-intensity
beamcomprising made simultaneously.
The beam emergedfrom a circularhole 4.8 mm in diameter.
A scintillation counter, used as a detector,was shielded by a
60- and 662-keV photons.Count ratesmeasuredby a single
detector and a two-channel gamma spectrometerwere correctedfor coincidencelossesdue to pulse-resolving
time. They
concludedthat error in soil water contentmeasurementby the
ElectromagneticTechniques
Electromagnetic
techniquesincludethosemethodswhich
dependupontheeffectof moisture
ontheelectrical
properties
of soil.The magneticpermeabilityof soilsis very nearly that
of free space,and hencethe electromagnetic
approachesexof the soil'sdielectricproperThe gammaray attenuationtechniquehasthe sameadvan- ploit the moisturedependence
tagesasitems2, 3, and4 listedunderneutronmeters,as well ties.
Dielectricpropertiesof moistsoilmaybe characterized
by a
as the following:
1. Data can be obtainedover very small horizontalor ver- frequencydependentcomplexdielectricresponsefunction
[Bottcher,1952]:
tical distances.
dual energygammaattenuationmethodwill probablynot exceed a standard deviation of 1%.
2. The measurement
is nondestructive.
Its disadvantagesare
•(•o)= •r(•0)+ j•,(•0)
(2)
1. Large variationsin bulk densityand moisturecontent where•r(W)istherealpartof •, •,(w)istheimaginarypartof
canoccurin highlystratifiedsoilsand limit spatialresolution. j is the squareroot of- 1, and w is the (angular)frequency.
The function•r(W)is aboutconstantfrom w ----0 out to the
2. Field instrumentationis costlyand difficult to use.
3. Extreme care must be taken to ensure that the radioactive source is not a health hazard.
neighborhood
of therelaxationfrequency
wRof dipolesin the
Soane[1967]and Coreyet al. [1971]alsouseddual energy,
collimatedbeam gammaraysto simultaneously
measuredensity and moisture content of soil columns.Others who have
investigated
the techniqueincludeGardnerandRoberts[1967]
and Gardneret al. [1972].They usedtwo collimatedbeamsof
polarization,when the electricfield is removed.Beyond
the soil column from one beam to another. In their study the
error in the determinationof Ydand 0 of (1) resultedfrom the
randomnessof the emissionfrom the sources,random error in
can no longerfollow the field, and the ability of the medium
medium.The timewR
-• is thetimeconstant
for thedeca•y
of
the function •r decreasesuntil in the visibleregionof the spec-
trum,andit is equalto theindexof refraction
squared.
The
realpartof thedielectricresponse
functionisa measure
of the
energystoredby thedipolesalignedin an appliedelectromagmonoenergetic
gammaraysfrom24'Amand '•?Csbut moved neticfield.When the frequencyis greaterthan w•, the dipoles
to store electric field energy decreases.
The function•,(w) is a measureof the energydissipation
rate in the medium. Viewed as a function of frequency,and
ments,presence
of a smallhigher-energy
peakin the 241AI'nstartingfrom low •0,it risesto a peak at •0• and, thereafter,decreases.The behaviordescribedis due to permanentdipoles
spectrum,and countingdeadtime.
Goitet al. [1976]showedexperimentallythat the variability in the soil medium. In complicated heterogeneousmedia,
attenuation coefficients and soil column thickness measure-
964
SCHMUGGE ET AL: SOIL MOISTURE DETERMINATION
I
I
I
24
•
22
---o----)•
)•=21
CM (Lundien, 1971)
= 1.55 CM (Wanget.al, 1978)
20
dependson the soil texture,i.e., particle sizedistribution,0• is
smallerfor a sandand larger for a clay and has been found to
be linearly dependenton the wilting point for the soil.This effect has beendemonstratedin laboratorymeasurements
of the
dielectric constant[Lundien,1971; Newton, 1977].
This dependenceof dielectricpropertiesof soil on moisture
content can be used for either in situ or remote sensors. In this
section,in situ devices,measuringeither soil resistivityor capacitance,will be discussed;
the remotesensorapproaches
will
be presentedlater.
16
z
A variety of implantable sensors,responsiveeither to resiso
tivity (•i), polarization(•r), or both, have been constructed
14
z
o
[Wexler, 1965; Roth, 1966; Thomas, 1963; Gagne and Outwater, 1961;DePlater, 1955; Silva et al., 1974]. Traditionally,
•:
12
thesehave beendesignedfor operationat frequenciesbelow 1
MHz. Recently, however, due to a steady decreasein the
uJ
10
I
physicalsizeof high-quality,high-frequencycomponents,im1
I
plantable sensorshave becomea practicalreality [Seliget al.,
1975; Walshet al., 1979;Layman, 1979; Wobschall,1978].
The resistivityof soils dependson moisture content and
hencecan serveas the basisfor a sensor.It is possibleeither to
measurethe resistivitybetweenelectrodesin a soil or to measure the resistivityof a material in equilibrium with the soil.
Sensorsof either kind can be very compact,and an array of
them can be connectedto standarddata collectionplatforms.
o
The difficulty with resistivesensorsis that the absolutevalue
of soil resistivitydependson ion concentrationas well as on
0
10
20
30
moisture concentration [Bouyoucosand Mick, 1948]. ThereSOIL MOISTURE, WEIGHT PERCENT
fore careful calibration is required for this technique.Even
Fig. 1. Realandimaginary
partsof thedielectric
constant
for clay with carefulcalibration,the instrumentmay requirefrequent
loam soils at wavelengthsof 21 cm [Lundien,1971] and 1.55 cm recalibrationdue to changesin organicor salt concentrations.
[Wang et al., 1978].
The calibrationproblem becomeslesssevereas the operating
frequencyis increased,sincethe relative contributionof ion
REAL PART
IMAGINARY
PART
theremaybe morethan onerelaxationmechanism
and more motion decreases.
Implantable sensors,which are sensitiveto polarization,•,
than oneabsorptionpeak.Furthermore,at frequencies
above
•0R,the mediummay showfurtherdispersion
and absorption in essencemeasure capacitance [Thomas, 1963; Gagne and
Outwater, 1961;DePlater, 1955;$elig et al., 1975; Walshet al.,
regionsdueto directmolecular
excitations.
The frequency
will generallylie in the microwave
range(18 GHz in H20), 1979;Layman, 1979].This parameteris the electricalquantity
whereas the latter molecular excitations will be in the sub-
that most directly indicatesmoistureconcentration.When the
moistureheld in the soil can be regardedas free, i.e., above
transition8•, the relationshipbetween• and moistureis linear. Furthermore,evenin more complicatedmaterials,suchas
a montmorillonitic clay, where the water is relatively tightly
bound,it is possibleto determinemoisturecontentby measuring the capacitanceof an implanted sensor.Becauseof this,
several capacitive sensorshave been constructed.Most of
these have been designedfor operation below I MHz, alproperties.
Thisdependence
is shownin Figure1,whichpres- thoughmore recentlysomework hasalsobeendoneup to 100
entsthe resultsof laboratorymeasurements
at wavelengths
of MHz [Wobschall, 1978; Walsh et al., 1979; Layman, 1979;
21 and 1.55cm (frequencies
of 1.4and 19.4GHz). The wave- Silva et al., 1974].The motivationfor increasingthe operation
lengthdependence
is due to the differencein the dielectric frequencyis again to minimize the contributionof ionic conductivity,which if large, can make accuratemeasurementof
propertiesof water at thesetwo wavelengths.
At low moisturelevels there is a slow increaseof • with soil the capacitancedifficult.One other promisingtechniqueis to
moisture,but above a certain point, the slope of the curve work at 10- to 100-MHz frequencyand to use a bridge techsharplyincreases
whichis dueto the behaviorof the waterin nique that allowsa determinationof both • and •. Thesecan
the soil.When water is first addedto a soil,it is tightly bound be usedseparatelyas moistureindicators[Walshet al., 1979;
to the soilparticles.In this statethe water moleculesare not Layman, 1979].
The main advantagesof either resistoror capacitortype defree to becomealigned,and the dielectricpropertiesof this
water resemblethoseof ice for which • = 3.5. As the layer of vices are as follows.
1. With calibration,they are capable,in principle,of prowateraroundthe soil particlebecomes
larger,the bindingto
the particledecreases
and the watermolecules
behaveasthey viding absolutevaluesfor soil moisture.
2. They can be implantedat any depth;thusmoistureprodo in the liquidstate;hencethe greaterslopeat thehighersoil
moisture values. The soil moisture content 0• at this transition file data can be obtained by this method.
millimeteror infraredregionsof the spectrum[Bottcher,1952;
Hasted,1974].In soils,wnis reducedto around1 GHz, due to
the binding of the water moleculesto the soil particles
[Hoekstraand Delaney,1974].
The preceding
description
generallyappliesto all dispersive
media.In a soilthe valuesof •r are typicallybetween3 and 5;
in water the value is about 80. Hence relativelysmall amounts
of free water in a soil will greatly affect its electromagnetic
SCHMUGGE ET AL: SOIL MOISTURE
3. A wide variety of sensorconfigurationsvarying from
verysmallto largeare possible,
and hencethereis somecontrol over the sensor volume of influence.
4. The precisionof both the resistiveand capacitivesensorsis high.Resistivesensors
are alsorelativelyaccuratewhen
other parametersare adequatelycontrolled,and the capacitive sensorshave relatively high intrinsic accuracythat is
more nearlyindependentof parametersotherthan moisture.
This followslogically, in that the capacitivesensorsare directlyresponsive
to the amountof polarizedenergystoredin
the regionof the sensor,a quantitydominatedby the water
DETERMINATION
965
sion, it is well suitedfor automatic(includingsatelliterelay)
recordingsystems[McKim et aL, 1975;Elzefiawyand Mansell,
1975; Gillham et aL, 1976].
Essentialstepsin the techniqueinclude de-airing the water
or solutionin the tensiometer,placing the tensiometersystem
in the soft,and allowingit to cometo equilibriumwith the soft
water. The ceramic cup is porous to water and solute but not
to air, so that water can flow and soft water conditions or
changein moisturecontentcan be determined.As the softwater content increases, it is held at a lower tension; when the
tensiometer reads zero, the soft is saturated, and water tension
is zero. The highesttension(or suction)reading that can be
obtainedwith a tensiometeris about I bar (1 arm). In mostinstances,data cannot be obtained beyond 0.8-0.9 bar because
minimize disturbances to the soft.
the air entry value of the ceramiccup is exceeded.This means
2. There are questionsof long-termreliability and main- that the moisture content range over which the tensiometer
tenanceof the calibration,especiallyif the ionic concentration can be used is limited. Richards[1949] stated that for coarse,
sandysoilsthe range of the tensiometermay covermore than
in the softwater changes.
3. Cost of readout devices and interfaces with remote col90% of the available moisturecontentrange.Clay softsposea
different problem. For example, in soilscontainingover 42%
lectionplatformsis high.
Overall, it would seemthat the relative advantagesin some montmorilloniteclay, the tensioncan changefrom 200 to 800
applications
would warrantseriousconsideration
of the im- cm H20 with a 1% changein volumetric water content [Abele
et aL, 1979].
plantablesensors.
Soft moisture measurementproceduresusing tensiometer/
transducersystemsare ways to monitor moisturetensiongraTensiometricTechniques
dients in the field. Studiesby Bianchi [1962], Klute and Peters
The term 'tensiometer'was used by Richardsand Gardner [1962], Watson[1967],Rice [1969],)lndersonand Bun [1977]
[1936] as an unambiguousreferenceto the porouscup and have shown the advantagesof using pressuretransducersto
vacuumgagecombinationfor measuringcapillarytensionor producea fast response,low-volumedisplacementtensiomethe energywith whichwateris held by the soft.However,ten- ter system.These types of systemsare capableof monitoring
siometerswere used to measure soft water tension in unsatumoisturetensionchangesthat occur rapidly in infiltration, irrated softsas early as 1922 [Gardneret al., 1922]. Richards rigation, groundwater recharge, and evapotranspiration.
[1949]and othershavemadeextensive
developments
and im- These gradient data must be usedin flux equationswith preprovementsin the tensiometers
usedin field and laboratory viously determined conductivity values to calculate water
present.
The disadvantagesare as follows.
1. The moisture sensormust be implanted properly to
soft water studies.
movement.
The energyterm canbe expressed
aspF, whichis definedas
the commonlog of the heightof a water columnin centimetersequivalentto the softmoisturetension,or it can be expressed
asa suction(negativepressure)
or a potential(energy
per unit mass).Elrick [1967]recognized
sixcomponents
of the
total energyof softwater,of whichmatricsuctionis one. Matric suctionis the pressuredifferenceacrossa boundary permeableonly to water and solutes,whichseparates
bulk water
and soft water in hydraulic,chemical,and thermal equilib-
Tensiometershave been used for years to measure soft
moisture tension. During recent years, advancementsin system design and performancehave made possiblethe implementation of soft moisture field monitoring programs.However, care is still needed in assessingthe use of the system.
rium. Dissolved salts or chemicals in the soft water contribute
to solutesuction.Bayeret al. [1972]suggested
usingthe term
'capillarypotential'to denotethe total potential,which includesnot only surfacetensionforcesbut alsothe osmoticand
adhesion forces.
The mostwidely knownmethodfor measuringthe capillary
or moisturepotentialis baseduponthe so-calledsuctionforce
of the softfor_water[Bayeret al., 1972;Richards,1965].Ten-
Listedbeloware someof the advantages
anddisadvantages
of
using tensiometers.
Advantages.
1. Systemsare easy to designand construct.
2. Systemscostrelatively little.
3.
Information
can be obtained
under saturated and unsaturated
on moisture
distributions
conditions in near real time.
4. Tensiometerscan usually be placed in soft easily and
usually with minimal disturbance.
5. Systemscan operate over long time periodsif properly
mainta'med.
6. With the tensiometer/transducersystemthe response
time is very rapid.
7. Different typesof liquids, like ethyleneglycol solution,
uid-filled(usuallywater),porousceramiccup connectedby a
continuous
liquid columnto a manometeror vacuumgage.In can be usedto obtain data during freezingand thawing condisomedesignsthe liquid is an ethyleneglycol-watersolution tions. When the softs are frozen, the tension of the unfrozen
and the measuringgage a transducerwith electricaloutput. water is greater than 800 cm and the systemfails.
8. Systems
canbe usedwi•hpositive
or negative
reading
The transduceroutput can be interfacedto near real time data
siometersare usedto measurethe suctionand consistof a liq-
acquisitionsystems.Use of an ,ethyleneglycol-watersolution tensiometers to read both water table elevation and soft moisasa replacementfor waterin the tensiometer
allowsthe useof ture tension,dependingon softwater status.
Disadvantages.
a tensiometer/transducer
systemin cold areas[McKim et al.,
1. Tensiometersprovide direct measurementsof softwater
1976].Sincethe tensiometer/transducer
systemhasa millivolt
output and respondsrapidly to changesin softmoistureten- suction but only indirect measurementof soft moisture con-
966
SCHMUGGE ET AL: SOIL MOISTURE DETERMINATION
TRANSPIRATION
SOIL WATER
MODELS
Recent developmentsof soil water modelsbasedon column
massbalanceprovide an alternativeto directly or indirectly
measuringsoil moisturein the field. Figure 2 is a schematic
diagram of the physical system and the driving forces that
must be consideredin modeling the system.
Basedupon'conservation
of mass,the soil moisturein the
systemat any time can be determinedusingthe relationship
PREClPliATIO
N
EVAPOTRANSPIRATIO
N
RUNOFF,
EVAPORATION
INFIL
A HORIZON
ROOT
ZONE
SM,-- SM,_• + P- R- L-
E-
T + C- Q
(3)
where
J /-t•j•i/EDiSTRiBUT,O
NB
HORIZ•j•ON
SM,
LATERAL FLOW
c HORIZON
,•/,////WA/E•TJ•BL/E
J///'
PERCOLATION
Fig. 2. Schematic
diagramof soil-plant-atmosphere-water
(SPAW)
system.
soil moisture volume at time t;
SMt_! SOilmoisturevolume at previoustime;
P precipitation;
R
L
E
T
C
Q
surface runoff;
net lateral subsurfaceoutflow;
evaporationor condensation;
transpiration;
capillary risefrom lower levels;
percolation.
tent. This translationrequiresknowledgeof the moisture
characteristics for the soil.
2. Tensiometerscan be brokeneasilyduringinstallation.
3. Resultscan be determinedonly within the 0- to 800-cm
water tension range.
4. Field installations
usingpressure
transducers
drift electronically.
HygrometricTechniques
The relationshipbetweenmoisturecontentin porousmaterials and the relative humidity (RH) of the immediateatmo-
sphereis reasonably
well known.Thereforeseveralrelatively
simplesensors
for measuringRH have beendesigned.Basically, thesesensorscan be classifiedinto seventypesof hygrometers:
electricalresistance,
capacitance,
piezoelectric
sorp-
This generalizedmodel representsonly a singlecolumnthat
is horizontallyhomogeneous
at all levels.Actual systemsare
heterogeneous,
and they can be representedby spatial averagesor by linked columnsthat accountfor the spatialvariability.
Published soil moisture models vary in the level of detail
they usein representingthe physicalsystemand temporalvariations of the driving forces. Some of the important differencesbetweenmodelsare (1) method usedfor computingthe
potentialevapotranspiration,
(2) methodusedfor computing
infiltration and runoff, (3) temporal definition of evaporative
demand and precipitation,(4) considerationof saturatedand
unsaturatedlevels,(5) number of soil layersused,(6) method
used for computingsoil evaporationand plant transpiration,
and (7) consideration
of the thermalpropertiesof the soil sys-
tion, infrared absorption,and transmission,dimensionally
tem.
varyingelement,dew point, and psychrometric.
Many of the publishedmodelsthat simulatesoil moisture
Electricalresistancehygrometersutilize chemicalsaltsand
acids,aluminum oxide,electrolysis,thermal,and white hydro- were developedfor agricultural applications.A very simple
col to measure RH. The measured resistance of the resistive
model, describedby Holmes and Robertson[1959], treats the
elementis a function of RH. Bouyoucos
and Cook [1965] con- soil as a singlehomogeneouslayer. Potential evapotranspirasideredthe white hydrocolhygrometeras the bestavailable. tion is computedempirically,and the actualevaporationis set
They statedthat castingthe stainless
steelelectrodes
in white equalto this as longas moistureis available.All precipitation
hydrocol(a form of plasterof Paris)causesgreateraccuracy becomes infiltration and groundwater interactions are igbecausethe cementsetshard, is pure, hasa low solubilityand nored. All computationsare performed on a daily basis.
contains no added salts.
Jensen et al. [1971] developed an irrigation scheduling
Pheneet al. [1971, 1973]developeda heat dissipationsensor model that takes into account soil moisture.Evapotranspirathat was used to measurethe matric potential. The accuracy tion is computedon a daily basisusingone of severalalternais not affectedby
of thissensorprovedto be asgoodasor betterthan that of the tive procedures.Actual evapotranspiration
thermocouplepsychrometeror salinity measurements.
The moisture deficits. Infiltration must be computed externally.
sensor,which was highly-sensitivein the 0 to -2 bar matrix Percolation is computed using an empirical relationship.In
potentialrange,had an accuracyof +0.2 bar. The accuracy this model the soil is treated as a single layer.
Holmes and Robertson[1959] presentedand describedandecreased
progressively
to +_1 bar at a matricpotentialof- 10
bar.
other model, the modulated soil moisture model, which is
The advantagesof the hygrometertechniquesare (1) simplicity of the apparatusand (2) low cost.
The disadvantages
are (1) deteriorationof the sensingelement throughinteractionwith the soil componentsand (2)
specialcalibrationis requiredfor eachmaterialto be tested.
The main use for this technologyseemsto be in applicationswhere RH in the material is directly related to other
properties,
an examplewouldbe dryingand shrinkageof ce-
slightlymore sophisticated
than thosedescribedabove.This
model usesa two-layer soil systemand considersthe fact that
actual evaporationwill generallynot equal the potential due
to moisturedeficits.The two are set equal until the moisture
in the upper soil zone is depleted.Thereafter,moistureis extracted from the lower zone at a reducedrate proportionalto
the moisture level. This model also allows a simple runoff
ments.
computation.
Baier and Robertson[1966] improved on thesemodelswith
SCHMUGGE
ETAL:SOILMOISTURE
DETERMINATION
POT•'NTIAI_
967
ET
/Nl'ERCEPl'ION
l
$0//_
•NO• /
•
TIME
OF
YEAR
UNUSED ENERGY
ACTUAl.
PoTs..
oI
....
to
TNANSFE
,
•T•T/•L
so• O•ST.,
O-e,•V
0
m •
T••/•T/•
A
•N.
JPLANT
IMOISTURE
ISTRE•
I
•
TI• • YEAR
B
•TDIS•RI•TI•
•
o
SOIL
•PTH,
feet
%
O,YA,
SOIL MOISTURE
WITH•AWA•
/N•••/ON
FROM
SOI•
•• •
MOISTURE
• ••
REDISTRIBUTION
S•L
UOl•.,•
o
• • •
Fig. 3. Flowchartof SPAWmodel[Saxtonet al., 1974].
the versatilesoil moisturebudgetmodel.In this modelthe soil
is dividedinto severallayers,and the availablewatercapacity
of each layer is taken to be the differencebetweensoil'sfield
capacityand wilting point. Evapotranspiration
can occursimultaneouslyfrom eachlayer and dependson the soil moisture presentand the particularsoil and root distributionof the
plantsinvolved,which are represented
by coefficients.
Flow
Stuff and Dale [1978], Shaw [1963], Goldsteinet al. [1974],
Ritchie[1972],and Youkerand Edwards[1969].
All of thesemodelswere developedprimarily for agricultural applications.Hydrologistshave also developedwater
balancemodelsthat includea soilmoisturecomponent.State
of the art examplesof the approaches
usedin hydrologicmodeling can be found in the U.S. Department of Agriculture,
betweenlayersis considered;
however,the techniqueusedis Hydrograph Laboratory (USDAHL) model [Holtan et al.,
empirical,asis the procedure
usedfor computinginfiltration. 1975] and the National Weather ServiceRiver Forecast MoSaxtonet al. [1974]developeda muchmorecomprehensive del (NWSRFS) [PeCk,1976].Other modelsare reviewedby
modelthat simulatessoil-plant-atmosphere-water
systemsin Fleming[1975].Most of thesemodelscan perform a continugreater detail thart the modelsdescribedabove. A flow chart ous simulation of the volumes and rates of water movement
of the model is shownin Figure 3. As illustrated,this model occurringin each componentof the watershed.
considers
aHfactorsthatinfluencethesystem.Someprocesses, In the USDAHL model,illustratedin Figure4, the spatial
like soilmoistureredistribution,
are modeledusinga physics- variabilityof soilsand vegetationis accountedfor by using
basedapproach,whereasothers,like plant transpiration,are zoneswithin which the hydrologicparametersare averaged.
semi-empirical.
Within eachzonethe soilis subdividedinto severalhomogeMany other modelswhich resemblethosementionedabove neouslayersdeterminedfrom hydraulicproperties.Evapohave beendeveloped.Additional informationcan be found in transpirationis computeddaily usingan empiricalequation
the works by Saxton and McGuinness[1979], Feddeset al. which considersthe crop and soilcharacteristics,
aswell as the
[1978],Hildreth [1978],Singh[1971],Kanemasuet al. [1976], currentsoil moisture.Evapotranspiration
is drawn from the
968
SCHMUGGE ET AL: SOIL MOISTURE DETERMINATION
Groundwater
oped for practical applicationto such problemsas crop yield
estimation,irrigation planning, and runoff forecasting.They
were developedto use readily available meteorologicaldata
for inputs, and therefore they usually use a 1-day time step.
Goldsteinet al. [1974]pointed out that modelswhich use a 1day averagingcan produce accurate weekly averageresults;
however, the daily resultswill show some deviation on any
given day. Jensenet al. [1971] made the same point. They
noted that expecteddaily errors between 10 and 15% should
becomenegligibleover 10-20 days.
Some situations,especiallyresearch,require information of
greater time resolutionand accuracythan thesemodels can
provide. Severalcomplexmodelscapableof simulatingsoilplant-atmosphere-water
systemshave been proposed.GeneraBy, thesemodelsare more physicallybasedthan the models
describedabove. Increaseddetail is usually costly,and these
complex models usually have high data and computer requiremerits,especiallyif simulation over extendedperiodsis
u_
Recharge
ß
Zone1
Zone2
nsite
•Offsite
Alluvial
Zone
Fig. 4. The USDAHL model [Holtan and Yaramanoglu,1977].
first two layers, which are consideredto be the root zone.
These computationsare performed daily. Infiltration is also
based on soil and crop characteristicsand the current soil
moisture.A 1-hour time step is usedfor thesecomputations.
The procedureusedfor soil moistureredistributionand percolationonly considersgravity flow.
In the NWSRFS Model, illustratedin Figure 5, two zones
are used to simulate soil water storageand movement. The
upper layer respondsquickly to rainfall and controlsoverland
flow. It is usuallyvery shallow.The lower layer is the balance
of the soil columnextendingto the water table. Soil hydraulic
propertiesare averagedwithin each layer. Moisture is stored
as either tension or free water. Infiltration, percolation and
soil moisture redistribution involve the free water. They are
computedwith empirical equationsthat use as a controlling
factor the ratio of the free water presentto the field capacity
of the layer involved.Evapotranspirationis alsocomputedusing an empirical procedure.Actual evapotranspirationis set
equal to potential until all moisturein the upper layer is depleted.When this occurs,moistureis extractedfrom the lower
zone usingan equationthat considersthe moisturedeficit and
the crop characteristics.
A 6-hour time stepis usedfor simulation.
Most modelsmentioned in precedingsectionswere devel-
ttt
desired.
Much researchhas been done on the physicsof soil water
movementand storageunder bare soil conditions.Thesemodels usually involve solvingthe equationsdescribingone-dimensional vertical unsaturated flow and horizontal saturated
flow. Most of the publishedmodelsdiffer on the boundary
conditionsimposedand the numerical approximationsused
for solution.
Hillel [1977]describedseveralphysicallybasedmodelsdesignedto simulatesoil water conditionsunder bare soil conditions.,All of the modelswere programedusinga versatilesim-
ulation language called continuoussimulation modeling
programs (CSMP) [Speckhart and Green, 1976]. Other examplesof this type of model are presentedby Hanks et al.
[1969], Warrick et al. [1971], Bresler [1973], Remsonet al.
[1971], and Pikul et al. [1974].
Some progresshas been made at linking theseone-dimen-
sionalmodelsto representspatiallyvariedsystems.
Figure6 is
a schematicof one such model presentedby Hillel [1977].
Others,suchas that of Freeze [1978], have extendedthe solutions to two-dimensionalproblems.
Includingthe effectsof plantsin detailedmodelsgreatlyincreasestheir complexity.Publishedmodelsthat simulatethese
effects include those presentedby Van Bavel and Ahmed
t
E•De
mand
i
LPRECIPITATION
INPUT
t
,J
PCTIM
•[DIRECT
PX
_.rmperviøusl
area
I
ET
I
v,ous ev,ous
I^O,,e-RUNOFF
UPPER
•ZONE
/ UZTW
I
WATERUZK
•1FLOW
• FREEUZFW
•
SARVA
LOWER
TENSION
WATER ?RpE
: E!WATER
i S
SSOUT
iLZFP' LZFS
LZTW
'
.....: ........
,LZPK
SUBSURFACE
PRIMARY
BASE
FLOW•
Fig. 5. The NWSRFS model [Peck, 1976].
SCHMUGGEET AL: SOIL MOISTURE DETERMINATION
[1976], Lemon et al. [1973], Makkink and vanHeemst[1975],
Hanson [1975], Slack et al. [1977], Feddeset al. [1976], Neuman et al. [1975], and Nimah and Hanks [1973].
The reader can gather from the number of referencescited
in thissectionthat a multitudeof modelshavebeendeveloped
and documented.A comprehensive
descriptionof each, along
with their respectiveadvantagesand disadvantages,
is beyond
the scopeof this paper. In general,a model that is adaptable
to almost every problem can be found in the literature. The
principaladvantageof modelsis that they canprovidetimely
soil moistureinformation without the necessityof field visits.
A general disadvantageof models is the error of their esti-
969
POTENTIAL
EVAPORATION
PRECIPITATION
LAYER 1
...•s(1)
_._?F"-.--.•_
I,Q(1'11
LAYER
2 -•1"---....•
iofi.21•'"'T'•---,--..•
mOs(31
os(•) os(•+•l
LAYER
3
....
LAYER
N
''''j•
•(•
mates.
REMOTE
SENSING
APPROACHES
COLUMN
The remotesensingof soil moisturedependsupon the measurementof electromagneticenergythat has either been reflected or emitted
from the soil surface. The variation
in in-
1
COLUMN
2
COLUMN
M
SYMBOLS
I = LAYER
J = COLUMN
INDEX
INDEX
tensity of this radiation with soil moisturedependson either
the dielectricproperties(indexof refraction),soil temperature,
or a combinationof both. The particular property that is important dependson the wavelengthregionthat is being con-
Q(I,J) = FLOW BETWEEN TWO LAYERS.
QB(J) = FLOW RATE FOR DEEP PERCOLATION
sidered as indicated
QS(J) = SURFACERUNOFF TO COLUMN J.
in Table
1.
Reflectedsolar energyis not a very promisingtool in soil
moisturedeterminationbecausethe relationshipof the soil
spectralreflectanceto the water contentdependson several
other variables,like spectralreflectanceof the dry soil, surface
roughness,
geometryof illumination,organicmatter, and soil
texture[Jacksonet al., 1978].Thesecomplicatingfactors,plus
the fact that reflectedsolarenergyrespondsto only a thin surface layer [Idso, et al., 1975b],limit the utility of solar reflect-
OR A ZERO FLOW
COLUMN
BOUNDARY
FOR
J.
QL(.•)
= SATURATED
FLOW
TOCOLUMN
J
BELOW
THE WATER
TABLE.
T(I) = THICKNESS OF LAYER I.
Fig. 6. A distributedsoilmoisturesimulationmodel.
ration complementsthe other effectsof water in soilby reducing the amplitudeof the surfacediurnal temperaturecycle.As
a result, the day-night temperaturedifferenceis an indicator
of somecombinationof soilmoistureand surfaceevaporation.
The basicphenomenonhas been studiedin experimentsat
ance measurements for soil moisture determination
and thus
the U.S. Water ConservationLaboratoryin Phoenix,Arizona,
will not be discussed further here.
where soil temperaturesversustime were measuredwith a
Water is unique in that it is near the extremesin its thermal thermocouple,before and after irrigation [Idso et al., 1975c].
and dielectricproperties.As a result,the correspondingprop- They observeda dramaticdifferencein the maximum temperertiesin the soil are highly dependenton its moisturecontent. aturesbefore and after an irrigation.On succeedingdays the
These properties are accessibleto remote sensingthrough maximum temperatureincreasedas the field dried out.
Figure 7 is a summary of resultsfrom many such experimeasurementsat the thermal infrared (10 m) and microwave
(1-50 cm) wavelengths.The approachesare [Schmugge,1978] ments.The amplitude of the diurnal range is plotted as funcas follows:
tion of soil moisture as measured at the surface and in 0- to 1-,
1. Thermal infrared,consisting
of measurementof the diurnal rangeof surfacetemperatureor measurementof the crop
canopy-airtemperaturedifferential.
2. Active microwave,consistingof measurementof the radar backscatteredcoefficient, and passive microwave, consistingof measurementof the microwaveemissionor bright-
0- to 2-, and 0- to 4-cm layers.Correlationwith the soil moisture in the 0- to 2- and 0- to 4-cm layersof the soil is good,
and this responseis related to the thermal inertia of the soil.
Initially, when the surfaceis moist,the temperaturesare more
or lesscontrolledby evaporation.Once the surfacelayer dries
below a certain level, the temperatureis determinedby the
ness temperature.
thermal inertia of the soil. These results indicated that for this
Thermal Methods
The amplitudeof the diurnal rangeof soil surfacetemperature is a function of both internal and external factors. The
internal factorsare thermal conductivityK and heat capacity
C, whereP -- (KC) •/'-defineswhat is knownas 'thermalinertia.' The external factorsare primarily meteorological:solar
radiation,air temperature,relativehumidity, cloudiness,and
wind. The combined effect of these external factors is that of
the driving functionfor the diurnal variationof surfacetem-
perature.Thermalinertiathenis an indicationof the sowsresistanceto this drivingforce.Sinceboth the heat capacityand
thermalconductivityof a soilincreasewith an increaseof soil
moisture,the resultingthermal inertia increases.
A complicatingfactoris the effectof surfaceevaporationin
reducingthe net energyinput to the soilfrom the sun.Evapo-
particular soil, the diurnal range of surfacetemperatureis a
goodmeasureof the moisturecontentin the surface(0-4 cm)
layer.
These temperaturemeasurementswere repeatedfor different soil types that ranged from sandy or light soilsto heavy
clay soils;and it was found that for a given set of diurnal meteorologicalconditions,there can be a wide range of temperature amplitude-water content relationshipsfor these different soils[Idso et al., 1975c].However, when soil water content
was transformedinto pressurepotential,a singlerelationship
was found to be valid for all of the differentsoil typesinvestigated.This is the basisfor expressing
moisturevaluesasa percent of field capacity(FC), where field capacityis the moisture contentof the -1/3 bar pressurepotential.
It should be emphasizedthat these experimentswere all
conductedin the field, usingthermocouplesinsteadof remote
970
SCHMUGGE ET AL: SOIL MOISTURE DETERMINATION
TABLE 1. Electromagnetic
Properties
for SoilMoistureSensing
WavelengthRegion
Thistechnique
is not applicableto fieldswith a vegetative
canopy.However,thedifference
betweencanopytemperature
and ambientair temperaturehas long beenknownto be an
indicatorof cropstress[Nt'iegand
and Namken,1966;Ehrler
andvanBayel,1967;Ehrler,1973;Hiler andClark,1971;Hiler
et al., 1974].Thus if the canopytemperatureis viewedas expressing
the soilmoisturestatus,a potentialexistsfor mon-
PropertyObserved
Reflected solar
Thermal infrared
Active microwave
Soil albedo/index of refraction
Surfacetemperature
Backscattercoefficient/dielectric
properties
Microwave emission/dielectric
Passive microwave
propertiesandsoiltemperature
itoringeffective
soilmoisture
overthe rootingdepthof the
particularcrop[IdsoandEhrler,1976].Following
thisargument,Jackson
et al. [1977]established
that a runningsumof
sensors
to acquiresurfacetemperaturedata.In March 1975an dailyvaluescalled'stress
degreedays'(SDD) canpotentially
experiment
wasperformedin whichremotelysensed
thermal be usedfor irrigationscheduling;
and Millardet al: [1977,
infraredtemperatures
from an aircraftplatformwerecom- 1978]confirmed
the feasibilityof this approach
for fully
paredwith the in situthermocouple
measurements
overa 5- grownwheaton the basisof airbornedata.Similarly,stress
day period.Figure8 presentsthe resultsfrom both the field degree
daysweresuccessfully
correlated
with the yieldof
wheat[Idsoet al., 1977]andhavesubsequently
beensuccessexpe
•rnnents
(fromFigure?) andthe .aircraft
experiments
[Reg•hato
etal., 1976;$chmuggeetal.,
1978].Thefieldresults fully correlated
with yieldsof alfalfa[Reginato
et al., 1978];
areexpressed
asa percent
of fieldcapacity
sotheycanbe beans
[Walker
andHaitield,
1979];
andbarley,
soybeans,
and
compared
withtheaircraftresults
obtained
overa widerange sorghum[Idsoet al., 1980].
of soiltextures.
Agreement
betweenthe thermocouple
meaMicrowave Methods
surementsand the remotely sensedradiation measurements
madefrom the aircraftwas good,indicatingthat the conAs discussed
previously,
the dielectricproperties
of a soil
clusions
basedon thethermocouple
fieldmeasurements
would are strongfunctionsof its moisturecontent.Sincethe dielectalsobe valid for radiationtemperatureobservations.
ric properties
of a mediumdetermine
thepropagation
charac-
40
SUR]••_]
•
o
l\ ' ø
I
• •2
[
MAR.
ooo1971
o
• •
AUG.
ooo1972
SEPT ß ee 197•
Xo 0
• •8
DEC.•
• •4
øo
197•
o o
I-
,g
20
12:
0
G.
to 40
'
z
$2
•
28
0.2.
•
• •o
o•o
0
ß 2o
.•
.o8
•
z4
.•z
4o
.48
.o8
.•
. •4
.s•
VOLUMETRIC S01L WATER CONTENT (cm• cm-•)
Fig.7. Summary
of results
forthediurnaltemperature
variation
versus
soilmoisture
[Idsoet al., 1975a].
SCHMUGGE
ET AL:SOILMOISTUREDETERMINATION
I
I
I
971
surface(Rayleigh-Jeansapproximation).This productis commonly referred to as brightnesstemperature.All our results
will be expressedas brightnesstemperatures(T s). The value
of Ts observedby a radiometerat a height above the ground
I
2OO
Ts = 'r(rrskyq-(1 - r)T•ou)+ Tatm
(4)
El00
where r is the surfacereflectivityand, the atmospherictransmission.The first term is the reflectedsky brightnesstemperature, which dependson wavelengthand atmosphericconditions;the secondterm is the emissionfrom the soil (1 - r - e,
the emissivity);and the third term is the contributionfrom the
0
20
25
30
35
40
atmospherebetween the surface and the receiver. At the
15
DAILY MAXIMUM - MINIMUM •
longer
wavelengths,i.e., those best suited for soil moisture
SURFACE SOIL TEMPERATURE (øC)
sensing,the atmosphericeffectsare minimal and will be neFig. 8. Plot of T versussoilmoisturein the O- to 2-cmlayer. The glectedin this discussion.
symbolsrepresentthe differenttypesof temperaturemeasurement: Thermal microwaveemissionfrom soilsis generatedwithin
solidcircle,openbox,surfacethermocouple;
opencircle,hand-held the soil volume.The amount of energygeneratedat any point
radiometer;
opentriangle,aircraftdataovertestplot;crosses,
aircraft
dataoverthe generalagriculturalfields.Solidcircle,opencircle,open within the volumedependson the soil dielectricproperties(or
box,opentriangle,from,Reginato
et al. [1976];crosses
fromSchmugge soil moisture) and the soil temperatureat that point. As enet al., [1978].
ergy propagatesupward through the soil volume from its
point of origin, it is affectedby the dielectric(soft moisture)
teristicsfor electromagnetic
wavesin the medium,they will
gradientsalong the path of propagation.In addition, as the
affectthe emissiveand reflectivepropertiesat the surface.As
energycrosses
the surfaceboundary,it is reducedby the effeca result,the emissiveand reflectivepropertiesat the soilsurtive transmissioncoefficient(emissivity),which is determined
facedependon thesoilmoisture
content,
whichcanbe mea- by the dielectriccharacteristics
of the soil near the surface.
suredin themicrowave
regionof thespectrum
by radiometric The emissionfrom the soil surfacecan be expressedas
o•
I-- o
z
ß
(passive)
andradar(active)
techniques.
Thisphysical
relationshipbetween
themicrowave
response
andsoilmoisture,
plus
the ability of the microwave
sensors
to penetrateclouds,
makesthemveryattractive
for useassoilmoisture
sensors.
In
thefollowing
sections,
results
will bepresented
demonstratingwhere T(z) is the temperatureprofileand a(z) is the absorptithis sensitivity
to soilmoisture,
alongwith a discussion
of vity as a function of depth which dependson moisturecon-
rs=e/:r(z)a(z)
exp dz'l
dz (5)
someof the noisefactors,e.g.,vegetation
and surfacerough- tent. Resultsfrom numerical solutionsto this equation have
ness,that affectthe relationshipbetweenthe sensorresponse beenpresentedby Njoku and Kong [1977], Wilheit[1978],and
Burke et al. [1979]. These papershave included resultswhich
and soil moisture.
Passivemicrowave. A microwave radiometer measuresthe indicate that the modelsdo a goodjob of predicting Ts for a
thermal emissionfrom the surface,and at thesewavelengths smoothsurface.One of the mostsignificantresultsfrom these
the intensityof the observed
emission
is essentially
propor- modelsis that the effectivesamplingdepth is of the order of
tionalto theproductof thetemperature
andemissivity
of the only a few tenthsof a wavelength[Wilheit, 1978].Thus for a
DATA FROM TEXAS A&M UNIVERSITY
i
I
I I I I
BARE, SMOOTHFIELD
• 300
DATA
FROM
TEXAS
A&M
UNIiVERSITY
260
I
.t = 21 cm
0[] VERTICAL
POLARIZATION
HORIZONTAL
POLARIZATION
240h
i
i
•, = 21 cm
i
-
o VERTICAL
POLARIZATION,
TBv
n HORIZONTAL
POLARIZATION,
TBH
.....
« (TBv+ TBH)
-
ROUGH,
29%
• 280f
6.9%
•
220
MEDIUM
'"' 260'-
• 200
<
m
X-
A-
240
••
22C
•
200,
•
O-2crn SOIL MOISTURE
O-9crn SOIL MOISTURE
O-fern REGRESSION LINE
18o
"' 160'
m T80
o
ca
140
160
• •1-40
210
O'
120
NADIR ANGLE
(a)
.-?.::•::•
.';- ,,
NADIR ANGLE
(b)
xx
i
PERGENT S01œ MOISTURE
.•
(c)
Fig.9. Results
fromfieldmeasurements
performed
at TexasA&M University:
(a) Tsversus
anglefordifferent
moisturelevels;
(b) Tz versus
anglefordifferent
surface
roughness
at aboutthesamemoisture
level;(c)Tz versus
soilmoisture
in differentlayersfor the medium-roughfield [Newton,1977].
972
SCHMUGGEET AL: SOIL MOISTURE DETERMINATION
I
I
I
I
I
I
DAWN
290- (a)
I
I
FLIGHTS
•=
260o
ID-DAY
O 3/18/75
-
270 -•
I
OBSERVED
VALUES
A
280
I
21 CM
(a)
LIGHTS
CALCULATED
C SMOOTH
•=
OBSERVED
21 CM
VALUES
_
O 3/18/75
A 3/22/75
3/22/75
VALUES
ß
•.
/•
•---• ROUGHSURFACE
'•l•
-
M'"'•o
AO
SURFACE
C
\
ß
'•
o
•
290
280
CALCULATED
VALUES
C SMOOTH
SURFACE
•-•
ROUGH SURFACE
,'• A A
-
270
h= 0.6
260
250
.
o
o
-
230
Nx 0
•220•
•
220
Oo '... •
210
240
o
•
210
200
190
180
X
170
160
110
120
0-2CM
0
I
10
20
30
40
50
60
70
80
90
1 O0
150
0--2CM
SOIL MOISTURE, % OF FIELD CAPACITY
Fig. 10. Aircraftobservations
of Ta overagricultural
fieldsaroundPhoenix,Arizona,fromMarch 1975flightsfor both
early morningand midday flights.
21-cm wavelengthradiometerthis depth is about 2-5 cm.
The range of dielectricconstantpresentedin Figure 1 producesa changein emissivityfrom greater than 0.9 for a dry
soil to lessthan 0.6 for a wet soil, assumingan isotropicsoil
with a smoothsurface.This changein emissivityfor a soil has
been observedby truck-mountedradiometersin field experiments [Poe et al., 1971;Blinn and Quade, 1972;$chandaet al.,
1978;Newton, 1977]and by radiometersin aircraft [$chmugge
et al., 1974;Burke et al., 1979; Choudhuryet al., 1979]and satellites[Eaglemanand Lin, 1976;$chmuggeet al., 1977].In no
thus increasesits emissivity.For a dry field the reflectivityis
already small (<0.1) so that the resultingincreasein emissivity is small.As shownin Figure 10b,surfaceroughness
hasa
significanteffectfor wet fieldswherethe reflectivityis larger
DATA
FROM
BASHARINOV 8' SHUTKO (1978)
case were emissivities as low as 0.6 observed for real surfaces.
This is primarily due to surfaceroughness,which generally
has the effect of increasingthe surfaceemissivity.
As can be seenin Figure 1, there is a greaterrange of dielectric constant for soils at the 21-cm than at 1.55-cm wave-
length. This fact, combinedwith a larger soil moisturesampling depth and better ability to penetrate a vegetative
canopy,makesthe longerwavelengthsensorsbetter suitedfor
radiometric soil moisture sensing.
In Figures 9a and 9b the field measurementsof Newton
[1977] are plotted versus angle of observation for various
moisture contentsand for three levels of surfaceroughness.
The horizontal polarization is that for which the electricfield
of the water is parallel to the surfaceand the vertical polarization is perpendicular to it. These results show the effect of
moisture content on the observedvalues of Ts, as well as the
I
I
effect of surface roughness,which_increasesthe effective
3
10
20
30
emissivityat all anglesand decreasesthe differencein T8 for
WAVELENGTH
IN CM
the two polarizationsat the larger angles.
Fig. 11. Effectof vegetationon passivemicrowavesensingof soil
For the smoothfield thereis a 100K changein Ts from wet moisture.Three curvesare (1) small grains:wheat, barley, rye and
to dry soils.Clearly, this range is reducedby surfacerough- grass;(2) broad leaf cultures,like mature corn and cotton;and (3)
ness.Roughnessdecreasesthe reflectivity of the surfaceand mixed forest.
SCHMUGGE ET AL: SOIL MOISTURE
DETERMINATION
973
Their resultsare summarizedin Figure 11, where the vegetation factor (fi) is the effectivetransmissivity
of the vegetation.
Thus for small grains the sensitivityis 80-90% of that expected for bare ground at wavelengthsgreater than 10 cm.
Broad leaf cultures,like mature corn or cotton,transmit only
20-30% of the radiation from the soil at wavelengthsshorter
than 10 cm and about 60% at the 30-cm wavelength.The authors observeda fi of 30-40% for a forest at the 30-cm wavelength, although they did not mention the type or height of
trees.These resultsare encouragingfor the use of long-wavelength radiometricapproaches.
The earth resourcesexperimentpackage(EREP) on-board
Skylab contained a 21-cm radiometer. This sensorwas nonThis result will be useful if the radiometer is to be scanned to
scanning;it had a 115-km field of view between half-power
provide an image.
points.With this coarsespatialresolutionit would be difficult
When the brightnesstemperaturesfor the medium-rough to directly compare sensorresponseand soil moisture meafield are plotted versussoil moisturein the 0- to 2-cm layer, surements.However, there have been two reports of indirect
there is an approximatelinear decreaseof Ts (Figure 9c). As comparisons.McFarland [1976] showeda strongrelationship
the thicknessof the layer increases,both the slopeand the in- between the Skylab 21-cm brightnesstemperaturesand the
tercept of the linear regressionalso increase.This is because antecedentprecipitationindex (API) for data obtainedduring
the moisture values for the high TB casesincrease,whereas a passstartingover the Texas and Oklahoma panhandlesand
they remain essentiallythe samein the low Ts or wet cases. proceedingsoutheasttoward the Gulf of Mexico.
This type of b•havior was also seen in the results obtained
Eaglemanand Lin [1976]carried the analysisof the Skylab
from aircraft platformsand hasled us to concludethat the soil data a step further and comparedthe brightnesstemperature
moisturesamplingdepthis within the 2- to 5-cm rangefor the with estimates
of the soilmoisture
,overthe radiometerfoot2 l-cm wavelength.This agreeswith the predictionsof theoret- print. The soil moistureestimateswere basedon a combinaical modelsof radiative transferin soils[Wilheit, 1978;Burke tion of actual ground measurements
and calculationsof the soil
et al., 1979].
moistureusinga climaticwater balancemodel.They obtained
Resultsfrom a 1975aircraft experimentover irrigatedagri- a correlationof 0.96 with data obtained during five different
cultural fields are presentedin Figure 10 [Choudhuryet al., Skylab passesover Texas, Oklahoma, and Kansas.This result
1979].Theseresultswere obtainedover fieldswith a range of was very good consideringthe difficulty of obtaining soil
soiltexturesfrom sandyloam to heavyclays.In the analysisit moisture information over a footprint of such a size and conwas observedthat the Ts responseto soil moisturedepended sideringthe fact that the brightnesstemperaturewasaveraged
on the soil texture, that is, the slopewas greater for sandy over the wide range of cultural conditionsthat occurredover
soils,which had a narrowerrange of soil moistures(0-20%) the area.
than the clay soils(0-35%). To take this texture dependence
These results from space supportedby the more detailed
into account,the soil moisturevaluespresentedin Figure 11 aircraft and ground measurementspresentedearlier strongly
were normalizedto the field capacity(FC) value for the par- supportthe possibilityof usingmicrowaveradiometersfor soil
ticular soil which was estimated from the measured soil texmoisture sensing.A difficulty with this approach is that the
tures.
spatialresolutionis limited by the sizeof the antennathat can
The solid symbolsin Figure 10 are calculatedvaluesof Ts be flown. For example, at a wavelengthof 21 cm, a 10 x 10 m
obtained with Wilheit's [1978] model using the measured antenna is required to yield 20-km resolutionfrom a satellite
moistureand temperatureprofilesfor the fields.The solidline altitude of 800 km. It is possibleto make use of the coherent
connectsthe values determined assuminga smooth surface, nature of the signal in active microwavesystems(synthetic
and the dashedline connectsthe valuesadjustedfor surface aperture radar, or SAR) to obtain better spatial resolutions,
roughnessusing a one-parametermodel. The dashedline fit
and it is this approachthat we discussnext.
the observedvaluesquite well. The valuesof the parameter
Active microwave. The backscatteringfrom an extended
were selectedempiricallyto give a bestfit to the data. Clearly, target, suchas a soil medium,is characterizedin termsof the
the same value works well for both the dawn and midday target'sscatteringcoefficientOo.Thus Oorepresentsthe link
flights.The effectsof soiltemperatureare shownin the Ts dif- betweenthe target propertiesand the scatterometerresponses.
ferencesbetweenthe dawn and midday flights.
For a given set of sensorparameters(wavelength,polarizaThere is little changein Ts for soil moisturevalues for the tion, and incidenceangle relative to 0ø), Ooof bare soil is a
0- to 2-cm layer out to about 30% of field capacity.When the functionof the soil surfaceroughness
and dielectricproperties
data were replottedversusthe 0- to 5-cm layer, the flat region which dependson the moisturecontent.The variationsof Oo
extendedout to about 50%of FC with a steeperslopebeyond with soil moisture,surfaceroughness,incidenceangle, and
that value. In Figure I the dielectriccurve broke at a moisture observation frequency have been studied extensively in
level of about 10%, which for the soil involved would be 40ground-basedexperimentsconductedby scientistsat the Uni50% of FC. Thus these aircraft results supported the con- versity of Kansas [Ulaby, 1974; Ulaby et al., 1974; Batlivala
clusionthat the samplingdepth in the soil is about 2-5 cm.
and Ulaby, 1977]usinga truck-mounted1- to 18-GHz (30- to
A vegetativecanopyactsas an absorbinglayer whoseeffect 1.6-cmwavelengths)active microwavesystem.
will dependon the amountof vegetationand the wavelength
To understandthe effectsof incidenceangle and surface
of observation.Basharinovand Shutko [1978] and Kirdiashev roughness,considerthe plots of Ooversusangle presentedin
et al. [1979] have reportedon observationsmade in the USSR Figure 12 for five fields with essentiallythe same moisture
over the 3- to 30-cm wavelengthrange for a variety of crops. contentbut with considerablydifferentsurfaceroughness.At
(>0.4). Thus the range of TB for the rough field is reducedto
about 60 K. The smoothand rough fields representthe extremesof surfaceconditionsthat are likely to be encountered.
The smoothfield had been dragged,while the rough surface
was on a field with a heavy clay soil (clay fraction 60%) that
had been deep plowed,which producedlarge clods.Therefore
the medium rough field, with a TBrange of 80 K, is probably
more representativeof the averagesurfaceroughnesscondition that will be encountered.Another important observation
from Figures9a and 9b is that the averageof the vertical and
horizontal Ts's is essentiallyindependentof angle out to 40ø.
This indicates that the sensitivity of this quantity, (1/
2)(Tzv + TzH), to soil moisturewill be independentof angle.
974
SCHMUGGE ET AL: SOIL MOISTURE DETERMINATION
ACTIVE MICROWAVE DEPENDENCEON ROUGHNESS
DATA FROM UNIV. OF KANSAS
WETSOILS:0.34 - 0.4 g/cc IN TOPcm
RMS
HEIGHT
(cm)
4.1
2.2
3.0
1.8
1.1
,
Cl
b
Polarization:
Polarization: HH
HH
Frequency
(GHz}:L 23
•
f Frequency
(OHz):
4.2520
10
o
%
z
.
% •o,
ROUGH
•
],o
•
u.•
C) '
o -lO
0
5
0
0
ROUGH
u.O
0
0
-
••
/•x ',
. SMOOTH
E qo
'--.,
SMOOTH
/ Polarization.
HH '••
0
I
I
I
5
10
15
•
•
-20
I
I
I
I
I
I
I
I
I
25
30
0
5
10
]5
2o
25
5o
ANGLE OF INCIDENCE
'25•
o
15
25
3O
(DEGREES)
Fig. 12. Angular responseof scatteringcoefficientfor the five fieldsfor high levelsof moisturecontentat (a) L band
(1.1 GHz); (b) C band (4.25 GHz); and (c) X band(7.25 GHz) for 1975soil mositureexperiment[Batlivalaand Ulaby,
19771.
while conductingradar observationsof agricultural fields.
During a flight by the NASA aircraft over a testsitenear Garden City, Kansas, Dickey et al. [1974], using a 13.3-GHz
scatterometer,measuredseveral fields, each of which (from
aerial photographyand field crew's reports) containedsections into which irrigation water was flowing and sections
ready for irrigation but not yet wetted.For one of thesefields,
a corn field, the effect of the irrigation on the radar return
seemedto producea differenceof about 7 dB at angleswithin
40ø from nadir between the irrigated and nonirrigated sections. Since all ground conditions,except soil water content,
al., 1978].
These experimentswere performed in both 1974 and 1975. were similar over the entire field, the differencesin Oocan only
The first experimentwas performedon a field with high clay be attributed to the effect of moisture.
The presenceof a vegetationcanopy over the soil surface
content (62%), whereasfor the second,the clay contentwas
lower. Although both experimentsprovidedthe samespecifi- reducesthe sensitivityof the radar backscatterto soil moisture
cationsof the radar parametersfor soil moisturesensing,i.e., by (1) attenuatingthe signalas it travelsthroughthe canopy
frequencyaround4.75 GHz and a 7ø-17ø nadir angle,the ob- down to the soil and back and (2) contributinga backscatter
servedsensitivityof Ooto soil moisturein the 0- to 1-cm layer componentof its own. Moreover,both factorsare, in general,
was differentfor the two experiments(Figure 13b).When the a function of several canopy parameters, including plant
soil moisturecontentis expressedas a percentof field capacity shape,height and moisturecontent, and vegetationdensity.
to account for textural differences, the sensitivitiesbecame alThe effectof the vegetationcoveron the radar response
to soil
mostidentical(Figure 13a)with a correlationof 0.84. This de- moisture is to reduce the sensitivityby about 40% as comfrom baresoilwhenthe two responses
pendenceon the percentof field capacityresemblesthat ob- paredwith the response
served with the thermal inertia and passive microwave are comparedas a functionof percentof FC in the top 5 cm.
soil represents
data for
techniques.Similarly, the samplingdepth for active micro- The responsefrom vegetation-covered
wave sensors also seems to be limited to the surface few censeveralcrops,i.e., wheat,corn, soybeans,and milo, covetinga
tmeters of the soil for the wavelengthsconsideredin the Kan- wide range of growth conditions[Ulaby et al., 1977].
There are many similarities in the two microwave apsasstudy [Ulaby et al., 1978].
Although no detailed airborne investigationshave yet been proachesto soil moisturesensing,e.g., ability to penetrate
reportedon the activemicrowaveresponseto the soilmoisture cloudsand moderateamountsof vegetationand the limitation
content beneath a vegetationcanopy, we observedthe differ- to samplingonly the surface2-5 cm of the soft.The major difence betweendry soil and soil undergoingirrigation in 1971 ference is that of spatial resolution.For passivesystemsthe
the longest wavelength (1.1 GHz, Figure 12a), Oofor the
smootherfieldsis very sensitiveto incidenceangle near nadir,
whereasfor the roughfield, Oois almostindependentof angle.
At an angle of about 5ø, the effectsof roughnessare minimized. As the wavelengthdecreases(Figures 12b and 12c) all
the fieldsappear rougher,especiallythe smoothfield, and as
a resultthe five curvesintersectat larger angles.At 4.25 GHz,
they intersectat 10ø, and it wasthis combinationof angleand
frequencythat yielded the bestsensitivityto soil moistureindependentof roughness[Ulaby and Batlivala, 1976; Ulaby et
SCHMUGGE ET AL: SOIL MOISTURE DETERMINATION
975
ACTIVE MICROWAVEDEPENDENCEON SOIL MOISTURE
DATA FROM UNIV. OF KANSAS
ANGLEOF INCIDENCE:10 o
RMS HEIGHTVARIATION: 0.9 TO 4.3 cm
HORIZONTALPOLARIZATION
FREQUENCY:4.25 & 4.75 GHz
5-
ß ßßß
e,• eeS
e•
ß .;. X"_'
-
'•
o
ß • •
ß
•
I
.;/
o
ß
J
0.],
ß
I
r • 0.83
I
I
I
0.2
0.3
I
/
0.4
I
I
0.5
•g/cm
z)
SOIL MOISTURE
IN TOP CM
Fig.13. Backscattering
coefficient
plotted
asa function
ofsoilmoisture
given(a)in percent
offieldcapacity
oftop1
cmand(b)volumetrically
in topcentimeter.
The1974and1975baresoilexperiment
dataarecombined
[Batlivala
and
Ulaby, 1977].
havethreegeneralrequirements:
frequentobservations,
an estimate
of
moisture
within
the
top
1-2
m
of
the
soil,
and,
genpractical
reasons
will be limitedto 5-10 ]fin.On the other
resolution
is limitedby the sizeof the antenna,and this for
of moisture
variations
overlargestudy
hand,usingthesynthetic
apperature
techniques,
spatialreso- erally,a description
lutionsof 100m or lessarepossible
fromspace,e.g.,25-mres- areas,i.e., a countyor state.In precedingsectionswe have
anddisadvantages
of the
olution was obtained from the 18-cm wavelengthsynthetic pointedoutsomeof theadvantages
three
approaches
than
can
be
used
to
collect
soil
moisture
aperature
radarontheSeasat
satellite.
Theproblems
withthe
latterapproach
arethedifficulty
in gettingan absolute
cali- data. None of the individual techniquescompletelysatisfies
of the applications
in a costeffectivemanbrationfor the SAR, the strongsensitivityto surfacerough- the requirements
ner.
nessandlookangle,andthelargeamountof datathatwould
In situ methods can accurately estimate soil moisture
have be handledin any operationalcontext.
the profile.However,the informationis reliable
The sensitivity
to soilmoisture
of the threeremotesensing throughout
To achieve
a specified
level
approaches
discussed
herehasbeendemonstrated
in fieldor onlyat thepoi•.tof measurement.
of
accuracy
in
estimating
the
areal
average
for
most
appliaircraftexperiments
andto a certainextentfromspacecraR
are-required.
For explatforms.
Theseexperiments
havealsoindicated
someof the cations,a largenumberof pointsamples
sampled
problems
associated
witheachapproach.
Theseproblems
are ample,in a study'of a largenumberof intensively
per field),Bellet al. [1980]found
summarizedin Table 2, whichcomparesthe remotesensing fields(20 or moresamples
approaches.
Someof thelimitations
listedareof a fundamen- that for 90% of the casesthe standarddeviation o of soil mois-
and
tal nature, like cloud cover-effectsat thermal infrared, ture in the surfacesoillayerswaslessthan4%. Shedecor
[1967]presented
the followingrelationship
for estiwhereasotherscould be reducedor eliminated by more ad- Cochran
vancedtechnology,
like largerantennas.
to improve.radio- matingthe requiredsamplesize:
meterresolution
or thedevelopment
of SARcalibration
techn --4
niques'.:
A.•'i'fundamental
;limita,ti0n•.applies
to all of the
approaches'•'•they
seem
tobe.sensing
themoisture
content
ina
(6)
ß
layeronly5-10cmthickatthe-surface.
Thislimitation
implies wheren isthe samplesizeand L is the desiredlevel'of..accu-
L = 2%ando'• 4%..
that remotesensingapproaches
will not be able to satisfy racy.Usingthisequationandspe'ctfying
thoseapplications
whichrequireknowledge
of themoisture [fromBelletal., 1980],therequirednumberof samp!es
conditions in the root zone of the soil.
DISCUSSION
be 16.Thusif a largenumberof individualfield estimatesare
required,the costscan becomeprohibitive.In addition,the
currentstateof the art methodsgenerallyrequire on-siteob-
Most hydrologic,
agricultural,
and meterological
appli- servations,sincereliable remote deviceshave not been develcommitment
of manpoweris recations that could benefit from soil moisture measurements oped.Thus a significant
976
SCHMUGGE
ET AL: SOIL MOISTURE
DETERMINATION
TABLE 2. Comparisonof RemoteSensingApproaches
Approach
Thermal infrared (1012m)
Advantage
Limitation
High resolutionpossible Cloud cover,limits fre(400 m)
quencyof coverage
Large swath
Basicphysicswell
Noise Source
Local meteorological
condition
Partial vegetativecover
Surfacetopography
understood
Passive microwave
Independenceof atmos- Poorspatialresolution
phere
(5-10 km at bes0
Interference from manModerate vegetation
made radiation
penetration
Surfaceroughness
Vegetativecover
Soil temperature
sources,limits
Active microwave
operatingwavelengths
Limited swath width
Independenceof
Surfaceroughness
atmosphere
High resolutionpossible Calibration of SAR
Surfaceslope
Vegetativecover
REFERENCES
quired. Soil water modelsprovide fast answers,as well as
predictions,of the soil water regimein a field or over a region. Abele,G., H. L. McKim,andB. Brockett,Masswaterbalanceduring
These modelsrequire large amountsof meteorologicalinput
sprayirrigationwith wastewaterat Deer Creek Lake land treatment
site,Rep. 79-29,U.S. Army Corpsof Eng., Hanover,N.H., 1979.
data that can be difficult and costlyto obtain. Model parame-
ters and functions are also difficult to determine. In addition,
Anderson,M. B., and T. P. Burt,Automaticmonitoringof soilmoistureconditions
in a hillslopespurandhollow,at.Hydrol.,33, 27-36,
varioussourcesintroduceerror into the modelpredictionsthat
1977.
can lead to significantdeviations.
Baier, W., and G. W. Robertson,A new versatile soil moisture budRemote sensingmethods offer rapid data collection over
get, Can. Plant Sci., 46, 299-315, 1966.
large areason a repetitivebasis.Severalquestionsstill need to Basharinov,A. Ye, and H. M. Shutko, Determination of the moisture
contentof theearth'scoverby superhigh
frequency(microwave)
rabe answeredconcerningthe dependenceof sensorobservadiometric methods:A review, Radiotek. Electron., 23, 1778-1791,
tions on soil moisture and other parameters,like vegetation.
1978.(Radio Eng. Electron.Phys.,Engl. Transl., 23, 1-12, 1978.)
The major problemsrelated to remote sensingseemto be the Batlivala, P., and F. T. Ulaby, Estimationof soil moisturewith radar
spatialresolution,depth of penetration,and cost.For the sparemotesensing,
Proceedings
of the 1l th InternationalSymposium,
pp. 1557-1566,Remote Sensingof the Environ. and Environ.Res.
tial resolution,the passivemicrowavesensorswill measurethe
areal soil moistureover groundareasabout 100 km2 in size
Inst. of Mich., Ann Arbor, 1977.
Baver,L. D., W. H. Gardner,and W. R. Gardner,Soil Physics,
498
from satellitealtitudes.Presently,we do not know how useful
pp., John Wiley, New York, 1972.
such a measurementwould be. The limited depth of pene- Belcher,D. J., T. R. Cuykendall,and H. S. Sack,The measurementof
soil moistureand densityby neutronand gamma-rayscattering,
tration of these systemsis also a severedrawback. Even if
Tech.Der. Rep. 127, pp. 1-20, Civ. Aeron. Admin., Tech. Dev. and
suchdata were available, how could they be used?
Eval. Center, Washington,D.C., 1950.
Belcher, D. J., T. R. Cuykendall, and H. S. Sack, Nuclear metersfor
soil moisturemonitoringprogram must utilize all three of the
measuringsoil densityand moisturein thin surfacelayers,Tech.
approachesand not just one. Becauseeachhas its advantages Der. Rep. 161, pp. 1-8, Civ. Aeron. Admin., Tech. Dev. and Eval.
Center, Washington, D.C., 1952.
and disadvantages,an integratedsystemshouldbe designed
These observations lead us to conclude that a cost effective
to capitalize on the advantagesand minimize the disadvantages.In situ methods,which we considerthe most accurate,
could be usedsparinglyin sucha systemfor calibrationand
verificationof modelsand remote sensingmeasurements.
The
other two approachescould be used to interpolate between
the point measurementsfor estimating areal averages.Remotely sensedmeasurementscould provide an estimateof the
surface conditions,which could then be extrapolatedvia a
model or usedto update a model simulation.These remotely
sensedmeasurementscould also be processedrapidly to give
us a quick look at the generalconditionover large areas.
In summary,we revieweda wide variety of methodsfor estimating soil moisturein this paper. Each has its own advantages and disadvantagesrelated to large-scalesoil moisture
monitoring. If a successful
monitoring systemis to be developed,it mustincorporateall of theseapproaches.
A very brief
descriptionof suchan integratedsystemwas presented;however, considerableresearchis needed to develop an optimal
system.
Acknowledgment.The authorsare gratefulto the reviewersof this
paper for the excellentreviewthey gave.andalsofor suggesting
additional references,which we had overlooked. We believe that their ef-
forts have improvedthe readability of the manuscriptconsiderably.
Bell, J.P., A new designprinciplefor neutronsoil moisturegauges:
The 'Wallingford' neutron probe, Soil Sci., 108, 160-164, 1969.
Bell, J.P., and C. W. O. Eeles,Neutron random countingerror in
terms of soil moisture for nonlinear calibration curves,Soil $ci.,
103, 1-3, 1967.
Bell, J.P., and J. S. G. McCulloch,Soil moistureestimationby the
neutron scatteringmethod in Britain, at.Hydrol., 4, 254-263, 1966.
Bell, K. R., B. J. Blanchard,T. J. Schmugge,and M. W. Witczak,
Analysisof surfacemoisturevariationswithin largefield sites,Water Resour. Res., 16, 796, 1980.
Bianchi,W. B., Measuringsoilmoisturetensionchanges,
Agr. Eng.,
43, 398-399, 1962.
Black,C. A. (Ed.-in-Chie0,Methodsof SoilAnalyses,
part 1, Physical
and MineralogicalProperties,
IncludingStatisticsof Measurement
and Sampling,AmericanSocietyof Agronomy,Madison,Wise.,
1965.
Blinn,J. C., andJ. G. Quade,Microwaveproperties
of geological
materials:Studiesof penetrationdepthand moistureeffects,4th Annual Earth ResourceProgramReview,NASA JohnsonSpaceCenter, Houston,Tex., Jan. 17-21, 1972. (Available from Nat. Tech.
Info. Serv.,N72-29327,Springfield,Va.)
Bottcher,C. F., ElectricPolarization,Elsevier,New York, 1952.
Bouyoucos,
G. J., and R. L. Cook,Humiditysensorpermanentelectric hygrometerfor continuousmeasurementof the relative humidity of the air, 1, Soil Sci., 106, 63-67, 1965.
Bouyoucos,G. J., and A. Mick, A comparisonof electricalresistance
units for makingcontinuousmeasurements
of soil moistureunder
field conditions,Plant Phys.,23, 532, 1948.
SCHMUGGE ET AL: SOIL MOISTURE DETERMINATION
977
Bowman, D. H., and K. M. King, Determinationof evapotranspira- Gardner, W., O. W. Israelsen, N. E. Edlefsen, and H. Conrad, The
capillarypotentialfunctionand its relationto irrigationpractice,
tion usingthe neutronscatteringmethod,Can. Soil Sci., 45, 117126, 1965.
Phys.Rev., Ser. 2, 20, 196, 1922.
Brakensiek,D. L., H. B. Osborn,and W. J. Rawls, Field manual for Gardner, W., G. S. Campbell, and C. Calissendorff,Systematicand
random errorsin dual gamma energysoil bulk densityand water
researchin agriculturalhydrology,Handbook224, U.S. Dep. of
Agr., Washington,D.C., 1979.
Bresler,E., Simultaneoustransportof solutesand water under transient unsaturatedflow conditions, Water Resour.Res., 9, 975-986,
1973.
Burke,W. J., T. Schmugge,
and J. F. Paris,Comparisonof 2.8- and
2 l-cm microwave radiometer observations over soils with emission
content measurements, Soil Sci. Soc. Amer. Proc., 36, 393-398,
1972.
Gillham, R. S., A. Klute, and D. F. Heerman,Hydraulic propertiesof
a porous-medium--Measurement
and empirical representation,
Soil Sci. Soc. Amer. Proc., 40, 203-207, 1976.
Glasstone,E., and M. C. Edlund, The Elementsof Nuclear Theory,D.
Van Nostrand, Princeton, N.J., 1957.
Goit, J. B., P. H. Groenevelt,B. D. Kay, and J. G. P. Loch,The applicability of dual gammascanningto freezingsoilsand the problems
of stratification, Soil Sci. Soc. Amer. Proc., 42, 858-863, 1976.
Goldstein, R. A., J. B. Markin, and R. J. Luxmoore, Documentation
of PROSPER, A model of atmosphere-soil-plantand water flow,
Rep. EDFB-IBP73-9, Oak Ridge Nat. Lab., Oak Ridge, Tenn.,
modelcalculations,
J. Geophys.
Res.,84, 287-294, 1979.
Charney,J. W., J. Quirk, S. H. Chow, and J. Kornfield, A comparativestudyof the effectsof albedochangeon droughtin semi
arid regions,Atmos.Sci., 34, 1366-1385,1977.
Choudhury,B., T. Schmugge,
R. W. Newton,and A. Chang,Effectof
surface roughness on the microwave emission from soils, J.
Geophys.
Res.,84, 5699-5706, 1979.
1974.
Churayev,N. V., and L. G. Rode,Measuringthe moisturecontentof
organicpeatsoilsby the neutronmethod,Pochvovedenie,
1, 96-100, Gurr, C. G., Use of gamma rays in measuringwater contentand per1966.
meability in unsaturatedcolumnsof soil, Soil Sci., 94, 224-229,
1962.
Cohen,O. P., A procedurefor calibratingneutronmoistureprobesin
Hanks, R. J., A. Klute, and E. Bresler, A numeric method for estimatthe field, Israel J. Agr. Res.,14, 169-178, 1964.
ing infiltration,redistribution,drainage,and evaporationof water
Corey,J. C., S. F. Peterson,
andM. A. Wakat,Measurement
of attenfrom soil, Water Resour.Res., 5, 1064-1069, 1969.
uationof •37C'sand 2n•Amgammaraysfor soildensityand water
content determinations, Soil Sci. Soc. Amer. Proc., 35, 215-219,
1971.
Davidson,J. M., J. W. Biggar, and D. R. Nielsen, Gamma radiation
attenuationfor measuringbulk densityand transientwater flow in
porousmedia,J. Geophys.
Res.,68, 4477-4783,1963.
DePlater,C. V., A portablecapacitancetype soil moisturemeter, Soil
Sci., 80, 391, 1955.
Dickey, F. M., C. King, J. C. Holtzman, and R. K. Moore, Moisture
dependencyof radar backscatterfrom irrigatedand non-irrigated
fields at 400 MHz and 13.3 GHz, IEEE Trans. Geosci.Elect., GEl2, 19-22, 1974.
Dmitriyev, M. T., Gammascopicmeasurementof soil moisture,Pochvovedenie,2, 208-217, 1966.
Douglass,J. E., Volumetric calibrationof neutron moistureprobes,
ßSoil Sci. Soc. Amer. Proc., 30, 541-544, 1966.
Hanson, G. K., A dynamic continuoussimulation model of water
stateand transportationin the soil-plant-atmosphere
system,I, The
model and its sensitivity,Acta Agr. Scand.,25, 129-149, 1975.
Hasted, T., AqueousDielectrics.Chapman Hall, London, 1974.
Hildreth, W. W., Soil moisture modeling review, Tech. Memo. N7823646, Nat. Tech. Inform. Serv., Springfield,Va., 1978.
Hiler, E. A., and R. N. Clark, Stressday index to characterizeeffects
of water stresson crop yields, Trans.Amer. Soc.Agr. Eng., 14, 757761, 1971.
Hiler, E. A., T. A. Howell, R. B. Lewis,a•d R. P. Boos,Irrigation timing by the stressday indexmethod,Trans.Amer.Soc.Agr. Eng., 17,
393-398, 1974.
Hillel, D., ComputerSimulationof Soil-WaterDynamics,International
DevelopmentResearchCenter, Ottawa, Canada, 1977.
Hoekstra, P., and A. Delaney, Dielectric propertiesof soilsat UHF
and microwavefrequencies,J. Geophys.Res., 79, 1699-1708, 1974.
Holmes, J. W., Influenceof bulk densityof the soil on neutron mois-
Eagleman,J., and W. Lin, Remote sensingof soil moistureby a 2 lcm passiveradiometer,J. Geophys.
Res.,81, 3660, 1976.
ture meter calibration, Soil Sci., 102, 355-360, 1966.
Ehrler, W. L., Cotton leaf temperatureas relatedto soil depletionand
Holmes, R. M., and G. W. Robertson, A modulated soil moisture
meteorological
factors,Agron.J., 65, 404-409, 1973.
budget, Mon. WeatherRev., 87, 101-105, 1959.
Ehrler, W. L., and C. H. M. van Bavel, Sorghumfoliar responsesto
Holtan, H. N., and M. Yaramanoglu,A user'smanual for the Univerchangesin soil water content,Agron.J., 59, 243-246, 1967.
sity of Maryland version of the U.S. Department of Agriculture
Elrick, D. E., Soil water movement:Theory and applications,ProHydrograph Laboratory model of watershedhydrology,Rep. MP
ceedingsof the First Canadian Conferenceon Micrometeorology,
918, Agr. Exp. Sta., Univ. of Md., CollegePark, 1977.
Part II, pp. 477-487, Dep. of Trans., Toronto, Ont., Canada, 1967.
Elzeftawy, A., and R. S. Mansell, Hydraulic conductivitycalculations Holtan, H. N., G. J. Stiltnet, W. H. Henson, and N. C. Lopez, USDAHL-74 revisedmodel of watershedhydrology,Tech.Bull. 1518,
for unsaturatedsteady-stateand transient-stateflow in sand, Proc.
Soil Sci. Soc. Amer. Proc., 39, 599-603, 1975.
Agr. Res. Serv.,U.S. Dep. of Agr., Washington,D.C., 1975.
Feddes, R. A., P. Kowalik, K. Kolnska-Malinka, and H. Zaradny, Idso, S. B., A note on somerecentlyproposedmechanismof genesis
of deserts,Quart. J. Roy. Meterorol. Soc., 103, 369-370, 1977.
Simulationof field water uptakeby plantsusinga soil water depenIdso, S. B., and W. L. Ehrler, Estimatingsoil moisturein the root zone
dent root extractionfunction, J. Hydrol., 31, 13-26, 1976.
of crops:A techniqueadaptableto remote sensing,Geophys.Res.
Feddes, R. A., P. J. Kowalik, and H. Zaradny, Simulation of Field
Lett., 3, 23-25, 1976.
Water Useand Crop Yield,SimulationMonogr.,Centre for Agricultural Publishing and Documentation, Wageningen, Netherlands, Idso, S. B., R. D. Jackson,and R. J. Reginato,Detectionof soil mois1978.
ture by remotesurveillance,Amer. Sci., 63, 549-557, 1975a.
Ferguson,H., and W. Gardner, Water content measurementin soil Idso, S. B., R. D. Jackson,R. J. Reginato, B. A. Kimball, and F. S.
Nakayama,The dependenceof bare soil albedoon soil water concolumnsby gamma ray absorption,Soil Sci. Soc.Amer. Proc., 26,
,
11-18, 1962.
tent,J. Appl.Meteorol.,14, 109-113,1975b.
Idso, S. B., T. J. Schmugge,R. D. Jackson,and R. J. Reginato, The
utility of surfacetemperaturemeasurements
for the remotesensing
New York, 1975.
of soil water status,J. Geophys.Res.,80, 3044-3049, 1975c.
Freeze, R. A., MathematicalModels of HillslopeHydrology,in HillslopeHydrology,edited by N.J. Kirkby, John Wiley, New York, Idso, S. B., R. D. Jackson,and R. J. Reginato, Remote-sensingof
1978.
cropyields,Science,196, 19-25, 1977.
Gagne, G., and J. Outwater,A portableelectronicmoisturedetector Idso, S. B., R. J. Reginato,J. L. Hatfield, G. K. Walker, R. D. Jackson, and P. J. Pinter, Jr., A generalizationof the stress-degree-day
for reinforcedplastics,Rep.3219,Univ. of Vt., Burlington,1961.
conceptof yield prediction to accommodatea diversity of crops,
Gannon,P. T., Influenceof earthsurfaceandcloudproperties
on the
Agr. Meteorol., 21, 205-211, 1980.
southFlorida seabreeze,Tech.Rep.ERL 402-NHEML2, Nat. OceJackson,R. D., R. J. Reginato,and S. B. Idso,Wheat canopytemperanic and Atmos.Admin., Silver Spring,Md., 1979.
ature:A practicaltool for evaluatingwaterrequirements,Water ReGardner, R. P., and K. F. Roberts,Density and moisturecontent
Fleming, G., ComputerSimulationTechniques
in Hydrology,Elsevier,
measurement
by nuclearmethods,Natl. Coop.Highw. Res.Program Rep., 43, 1967.
Gardner,W., andD. Kirkham,Determination
of soilmoisture
by neutron scattering,Soil Sci., 73, 391-401, 1952.
sour. Res., 13, 651-656, 1977.
Jackson,R. D., J. Cihlar, J. E. Estes,J. L. Heilman, A. Kakle, E. T.
Kannemasu, J. Millard, J. C. Price, and C. Wiegand, Soil moisture
estimation using reflected solar and emitted thermal radiation,
978
SCHMUGGE ET AL: SOIL MOISTURE
chap.4, SoilMoistureWorkshop,NASA Conf PubL2073,219 pp.,
Nat. Oceanic and Atmos. Admin,, Greenbelt, Md., 1978.
Jensen,M. E., J. L. Wright, and B. J. Pratt, Estimatingsoil moisture
depletionfrom climate, crop, and soil data, Trans.ASAE, 14, 954959, 1971.
DETERMINATION
scanner,Photogramm.Eng. Remote Sensing,44, 77-85, 1978.
Mortier, P., M. DeBoodt, W. Donsercoer,and L. DeLeenheer, The
resolutionof the neutron scatteringmethod for soil moisturedetermination, Trans.Int. Congr.Soil Sci. 7th, 1, 321-329, 1960.
Nace, R. L., Water of the world, Nat. History, 73, 10-19, 1964.
Kanemasu,E. T., L. R. Stone,and W. L. Powers,Evapotranspiration Neuman,S. P., R. A. Feddes,and E. Bresler,Finite elementanalysis
model tested for soybeanand sorghum,Agron. J., 68, 569-572,
of two-dimensional
flow in soilsconsidering
wateruptakeby roots,
1976.
1, Theory, Soil Sci. Soc.Amer. Proc., 39, 224-230, 1975.
King, K. M., Soil moisture-instrumentation,
measurementand gen- Newton,R. W., Microwaveremotesensingand its applicationto soil
eral principlesof networkdesign,Soil Moisture,Proceedings
of Hymoisturedetection,Tech.Rep. RSC-81, 500 pp., RemoteSensing
drology Symposium6, Rep. 269285, Can. Dep. of Energy, Mines,
Center, Tex. A&M Univ., College Station, 1977.
and Resour., Ottawa, 1967.
Nimah, M. N., and R. J. Hanks, Model for estimatingsoil water,
Kirdiashev, K. P., A. A. Chukhlantsev, and A.M. Shutko, Microwave
plant and atmosphereinterrelations,I, Descriptionand sensitivity,
radiation of the earth'ssurfacein the presenceof vegetationcover,
Soil Sci. Soc. Amer. Proc., 37, 522-527, 1973.
Radiotek. Elektron., 24, 256-264, 1979. (NASA Tech. Trans. TMNjoku, E.G., and J. A. Kong, Theoryfor passivemicrowaveremote
75469, 1979.)
sensingof near-surfacesoil moisture,J. Geophys.Res.,82, 31083118, 1977.
Klute, A., and D. B. Peters,A recordingtensiometerwith a shortresponsetime, Soil Sci. Soc. Amer. Proc., 26, 87-88, 1962.
Nofziger, D. L., Errorsin gamma-raymeasurements
of water content
Lal, R., Concentrationand size of gravel in relation to neutron moisand bulk densityin nonuniform soils,Soil Sci. Soc.Amer. Proc., 42,
845-850, 1978.
ture and densityprobe calibration, Soil Sci., 127, 41-50, 1979.
Lawless,G. P., N. A. MacGillivray, and P. R. Nixon, Soil moisture Olgaard,P. L., On the theoryof the neutronicmethodmeasuring
the
interfaceeffectsupon readingsof neutronmoistureprobes,Soil Sci.
watercontentin soil,RisoRep.97, pp. 1-74, DanishAtomicEnergy
Soc. Amer. Proc., 27, 502, 1963.
Layman, R., Soil moisturesensor,Rep. DCPPL-SM-3, Plasma Phys.
Commission, Roskilde, Denmark, 1965.
Lab.,Da•rtmouth
Coll.,Hanover,
N.H., 1979.
Olgaard,P. L., and V. Haahr, On the sensitivityof subsurface
neutron
moisturegaugesto variationsin bulk density,Soil Sci., 105, 62-64,
perimentsin predictingevapotranspiration
by simulationwith a
soil-plant-atmosphere
model (SPAM), in Field Soil Water Regime,
editedby R. R. Bruceet al., Soil ScienceSocietyof America,Madi-
Peck,E. L., Catchmentmodelingand initial parameterestimationfor
the National Weather Serviceriver forecastsystem,NOAA Tech.
Memo. NWS Hydro-31, Nat. Weather Serv., Silver Spring, Md.,
Lemon, E. R., D. W. Stewart, R. W. Shawcroft, and S. E. Jensen,Ex-
son, Wise., 1973.
Long, I. F., and B. K. French, Measurementof soil moisturein the
field by neutron moderation,J. Soil Sci., 18, 149-166, 1967.
Luebs, R. E., M. J. Brown, and A. E. Laag, Determiningwater content of different soils by the neutron method, Soil Sci., 106, 207212, 1968.
Lundien,J. R., Terrainanalysisby electromagnetic
means,Tech.Rep.
3-693,Rep. 5, 85 pp., U.S. Army WaterwaysExp. Sta.,Vicksburg,
Miss., 1971.
Makkink, G. F., and H. D. J. vanHeemst, Simulation of the Water
Balanceof Arable Land Pastures,Centre for Agricultural Publishing and Documentation,Wageningen,Netherlands,1975.
Mansell, R. S., L. C. Hammond, and R. M. McCurdy, Coincidence
and interferencecorrectionsfor dual-energygamma ray measurementsof soil densityand water content,Soil Sci. Soc.Amer. Proc.,
37, 500-504, 1973.
Marais, P. C., and W. B. De V. Smit, Determination of soil moisture
with a neutron meter, Ind. Rev. Africa, p. 78, 1958.
Marais, P. C., and W. B. De V. Smit, Effectof bulk densityand of hy-
drogenin forms other than free water on the calibrationcurve of
the neutron moisturemeter, S. Afr. J. Agr. Sci., 5, 225-238, 1962.
Mather, J. K., ClimatologyFundamentals
and Application,pp. 151153, McGraw-Hill, New York, 1974.
McCauley, G. N., and J. F. Stone,Source-detector
geometryeffecton
neutronprobecalibration,Soil Sci. Soc.Amer. Proc., 36, 246-250,
1972.
McFarland, M. H., The correlationof skylabL-band brightnesstemperatureswith antecedentprecipitation,Proceedings
of the Conferenccon Hydrometeorology,pp. 60-65, Amer. Meteorol. Soc.,Bos-
1968.
1976.
Phene,C., G. J. Hoffman, and S. L. Rawlins, Measuringsoil mattic
potentialin situby sensingheatdissipationwithin a porousbody,I,
Theory and sensorconstruction,Soil Sci. Soc.Amer. Proc., 35, 2733, 1971.
Phene,C., G. J. Hoffman,and R. S. Austin,Controllingautomatedirrigation with soil matric potential sensor,Trans.ASAE, 16, 773776, 1973.
Pikul, M. F., R. L. Street, and I. Remson, A numerical model based
on coupledone-dimensional
Richardsand Boussinesq
equations,
Water Resour. Res., 10, 295-302, 1974.
Poe, G. A., A. Stogryn,and A. T. Edgerton, Determination of soil
moisturecontentusingmicrowaveradiometry, Final Rep. 1684 FR1, 169 pp., Aerojet-GeneralCorp., E1 Monte, Calif., 1971.
Rawls,W. J., and L. E. Asmussen,
Neutronprobefieldcalibrationfor
soilsin the Georgiacoastalplain, Soil Sci., 110, 262-265, 1973.
Reginato, R. J., S. B. Idso, J. F. Vedder, R. D. Jackson,M. B. Blanchard,and R. Goettelman,Soil watercontentand evaporationdetermined by thermalparametersobtainedfrom ground-based
and remote measurements,
J. Geophys.
Res.,81, 1617-1620,1976.
Reginato,R. J., S. B. Idso,and R. D. Jackson,Estimatingforagecrop
production:
A techniqueadaptableto remotesensing,
RemoteSensing Environ.,7, 77-80, 1978.
Remson,I., G. M. Hornberger,and F. J. Molz, NumericalMethodsin
SubsurfaceHydrology,John Wiley, New York, 1971.
Reynolds,S. G., The gravimetricmethod of soil moisturedetermination, I, A studyof equipmentand methodologicalproblems,J. Hydrol., 11, 258-273, 1970a.
Reynolds,S. G., The gravimetricmethod of soil moisturedetermination, III, An examinationof factorsinfluencingsoil moisturevariaMcKim, H. L., D. M. Anderson, R. L. Berg, and R. Tuiinstra, Near
bility, J. Hydrol., 11, 288-300, 1970b.
real time hydrologicdata acquisitionutilizing the Landsatsystem,
field tensiometersystem,Trans.Amer. Soc.
Proceedings
of the Conferenceon Soil-WaterProblemsin Cold Re- Rice, R., A fast-response,
Agr. Eng., 12, 48-50, 1969.
gions,pp. 200-211,AGU, Washington,D.C., 1975.
Richards, L. A., Methods of measuringsoil moisture tension,Soil
McKim, H. L., R. L. Bert, R. W. McGaw, R. T. Atkins, and J. IngerSci., 68, 95-112, 1949.
soll, Developmentof a remote-readingtensiometer/transducer
system for usein subfreezingtemperatures,Proceedings
of the Second Richards, L. A., and W. Gardner, Tensiometersfor measuringthe
Conferenceon Soil-Water Problems in Cold Regions, pp. 31-45,
capillary tensionof soil water, J. Arner. Soc. Agron., 28, 352-358,
ton, Mass., 1976.
Edmonton, Alberta, Canada, 1976.
Millard, J.P., R. D. Jackson,R. C. Goettelman, R. J. Reginato, S. B.
Idso, and R. L. LaPado, Airborne monitoringof crop canopytemperatures for irrigation schedulingand yield predictions, Pro-
ceedings
of the 1l th International
Symposium
on RemoteSensing
of the Environment,pp. 1453-1461, Environ. Res. of Mich., Ann
Arbor, 1977.
Millard, J.P., R. B. Jackson,R. C. Goettelman,R. J. Reginato,and S.
B. Idso, Crop water-stressassessmentusing an airborne thermal
1936.
Richards, S. J., Soil suction measurementswith tensiometersin methods of soil analysis,Agronomy,9, 153-163, 1965.
Ritchie, J. T., Model for predictingevaporationfrom a row cropwith
incompletecover, Water Resour.Res., 8, 1204-1213, 1972.
Roth, M., How to measuremoisturein solids,Chem.Eng., 73, 83-88,
1966.
Saxton,K. E., and J. L. McGuinness,Evapotranspiration,
in Hydrologic Modeling of Small Watersheds,edited by C. T. Haan et al.,
SCHMUGGEET AL: SOIL MOISTUREDETERMINATION
American
Society
of Agricultural
Engineers,
St.Joseph,
Mich.,in
Ulaby, F. T., P. P. Batlivala,and M. C. Dobson,Microwavebackscatter dependenceon surfaceroughness,soil moistureand soil texture,
press,1979.
Saxton,K. E., H. P. Johnson,
andR. H. Shaw,Modelingevapotrans=
piration and soil moisture,Trans.ASAE, 17, 673-677, 1974.
Schanda,
E., R. Hofer,D. Wyssen,A. Musy,D. Meylan,C. Morzier,
979
I, Bare soil, IEEE Trans. Geosci.Elect., GS-16, 286-295, 1978.
Ursic, S. J., Improved standardsfor neutron soil water meters, Soil
Sci., 104, 323-325, 1967.
and W. Good, Soil moisture determination and snow classification
Vachaud, G., J. M. Royer, and J. D. Cooper,Comparisonof methods
with microwaveradiomerry,Proceedings
of the 12th International
of calibrationof a neutronprobeby gravimetryon neutron-capture
Symposiumon RemoteSensingof the Environment,Environ. Res.
model, J. Hydrol., 34, 343-356, 1977.
of Mich., Ann Arbor, 1978.
van Bavel, C. H. M., Calibration and characteristics of two neutron
Schmugge,
T. J., Remotesensing
of surface
soilmoisture,
J. Appl.Me-
moistureprobes,Soil Sci. Soc.Amer. Proc., 25, 329-334, 1961.
van Bavel, C. H. M., Accuracyand sourcestrengthin soil moisture
Schmugge,
T. J., P. Gloersen,
T. Wilheit,andF. Geiger,Remotesens= neutronprobes,Soil Sci. Soc.Amer. Proc., 26, 405, 1962.
ing of soilmoisturewith microwave
radiometers,
J. Geophys.
Res., van Bavel, C. H. M., and J. Ahmed, Dynamic simulationof water deteorol., 17, 1549, 1978.
79, 317-323, 1974.
Schmugge,
T. J., J. M. Meneely,A. Rango,and R. Neff, Satellitemi=
crowave observationsof soil moisture variations, Water Resour.
Bull., 13, 265, 1977.
pletionin the rootzone,Ecol.Modeling,2, 189-212,1976.
van Bavel, C. H. M., and G. B. Stirk, Soil water measurement with an
Am24•-Beneutronsourceand an applicationto evaporimetry,
J.
Hydrol., 5, 40-60, 1967.
Schmugge,
T. J., B. Blanchard,
A. Anderson,
andJ. Wang,Soilmois= van Bavel, C. H. M., N. Underwood, and R. W. Swanson, Soil moisturesensing
with aircraftobservations
of the diurnalrangeof sur=
ture measurementby neutron moderation, Soil Sci., 82, 29-41,
face temperature,Water Resour.Bull., 14, 169-178, 1978.
Selig,E. T., D.C. Wobschall,S. Mansukhani,and A. Motiwala, Ca=
pacitance
sensor
for soilmoisture
measurement,
Record532,p. 64,
Trans. Res. Board, Washington,D.C., 1975.
Shaw,R. H., Estimatingsoil moistureundercorn,Res.Bull. 520, Iowa
StateUniv. of Agr. and Home Econ.Exp. Sta.,Ames,1963.
Silva, L. F., F. V. Schultz,and J. T. Zalusky,Electricalmethodsof
determiningsoil moisturecontent,LARS Inform. Note 112174,
Purdue Univ., Lafayette, Ind., 1974.
Singh,V. P., Soil moisturemodels(review),paperpresentedat the
Annual Meeting, Amer. Soc.of Agr. Eng., Pullman,Wash., 1971.
Slack,D.C., C. T. Haan, and L. G. Wells,Modelingsoilwater movement into plant roots, Trans.ASAE, 20, 919-927, 1977.
Shedecor,
G. W., and W. G. Cochran,Statistical
Methods,593 pp.,
6th ed., Iowa State University Press,Ames, 1967.
1956.
Visvalingam,M., and J. D. Tandy, The neutronmethodfor measuring soil moisture content•A review, J. Soil Sci., 243, 499-511,
1972.
Walker, G. K., and J. L. Hatfield, Test of the stressdegreeday concept usingmultiple plantingdatesof red kidney beans,Agron. J.,
71, 967-971, 1979.
Walker, J., and P. R. Rowntree, The effect of soil moisture on circula-
tion and rainfall in a tropicalmodel, Quart. J. Roy. Meteorol.Soc.,
103, 29-46, 1977.
Walsh,J., D. McQueeney,R. Layman,and H. McKim, Development
of a simplifiedmethod for field monitoringof soil moisture,Proceedingsof 2nd Colloquium on Planetary Water and Polar Processes,U.S. Army Cold RegionsRes. and Eng. Lab., Hanover, N.
H., 1979.
Soane,B. D., Dual energygammaray transmission
for coincidence Wang, J., T. J. Schmugge,and D. Williams, Dielectric constantsof
measurements
of watercontentanddry bulk densityof soil,Nature,
soils at microwave frequencies,II, Tech. Note 1238, Nat. Oceanic
214, 1273-1274, 1967.
and Atmos. Admin., Greenbelt, Md., 1978.
Speckhart,
F. H., and W. L. Green,A Guideto UsingCSMP The Warrick, A. W., J. W. Biggar,and D. R. Nielsen,Simultaneoussolute
Continuous
SystemModelingProgram,Prentice-Hall,Englewood
and water transfer for an unsaturatedsoil, Water Resour.Res., 7,
Cliffs, N.J., 1976.
Stewart,G. L., and S. A. Taylor, Field experimentswith the neutron
scatteringmethod of measuringsoil moisture, Soil Sci., 83, 151158, 1957.
Stolzy,L. H., and G. A. Cahoon,A fieldcalibratedportableneutron
rate meter for measuringsoil moisture in citrus orchards,Soil Scg
Soc. Arner. Proc., 21, 571-575, 1957.
1216-1225, 1971.
Watson,K. K., A recordingfield tensiometerwith rapid response
characteristics,Water Resour. Res., 5, 33-39, 1967.
Wexler, A., Humidity and Moisture,vol. 5, pp. 1-14, Reinhold, New
York, 1965.
Wiegand,C. L., and L. N. Namken, Influenceof plant moisture
stress,solarradiation,and air temperatureon cottonleaf temper-
Stone,J. F., R. H. Shaw,and D. Kirkham, Statisticalparametersand
atures,Agron. J., 58, 582-586, 1966.
reproducibilityof the neutronmethodof measuringsoil moisture, Wilheit, T. T., Radiativetransferin a plane stratifieddielectric,IEEE
Soil Scg Soc. Amer. Proc., 24, 435-438, 1966.
Trans. Geosci.Elect., GE-16, 138-143, 1978.
Stuff,R. G., and R. F. Dale, A soilmoisturebudgetmodelaccounting Wilson, R. G., Methodsof measuringsoil moisture,The Secretariat,
for shallow water table influences,Soil Sci. Soc. Arner. Proc., 42,
Can. Nat. Comm. for the Int. Hydrol. Decade, Ottawa, Canada,
637-643, 1978.
Thomas, A., In situ measurement of moisture in soil and similar sub-
stancesby fringe capacitance,J. Sci. Inst., 43, 996, 1963.
Ulaby, F. T., Radar measurementof soil moisture content, IEEE
Trans.AntennasPropagat.,AP-22, 257-265, 1974.
Ulaby,F. T., andP. P. Batlivala,Optimumradarparameters
for mapping soilmoisture,IEEE Trans.Geosci.Elect.,GS-14,81-93, 1976.
Ulaby, F. T., J. Cihlar, and R. K. Moore, Active microwavemeasure-
1971.
Wobschall,D., A frequencyshift dielectricsoil moisturesensor,IEEE
Trans.Geosci.Elect., GE-16, 112-118, 1978.
Youker, R. E., and W. M. Edwards,Simplifiedbudgetingof plow
layer moistureunder meadow,Rep.ARS 41-154, Agr. Res. Serv.,
U.S. Dep. of Agr., Washington,D.C., 1969.
Zuber,A., and J. F. Cameron,Neutronsoilmoisturegauges,Atomic
Energy Rev., 4, 143-147, 1966.
mentsof soil water content,RemoteSensingEnviron.,3, 185-203,
1974.
Ulaby, F. T., G. A. Bradley,M. C. Dobson,and J. E. Bare,Analysis
of the activemicrowaveresponse
to soilmoisture,part II, Vegetation-CoveredGround, RSL Tech. Rep. 264-19, Univ. of Kans.,
Center for Res., Inc., Lawrence, 1977.
(Received November 21, 1979;
revised March 18, 1980;
acceptedApril 30, 1980.)