<%BANNER%>

Integration of Active and Passive Microwave Signatures for Characterization of Soil Properties

Permanent Link: http://ufdc.ufl.edu/UFE0045310/00001

Material Information

Title: Integration of Active and Passive Microwave Signatures for Characterization of Soil Properties
Physical Description: 1 online resource (124 p.)
Language: english
Creator: Liu, Pang-Wei
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2013

Subjects

Subjects / Keywords: microwave -- moisture -- radar -- radiometer -- remote -- sensing -- soil
Agricultural and Biological Engineering -- Dissertations, Academic -- UF
Genre: Agricultural and Biological Engineering thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Soil water is one of the key factors governing water and energy fluxes at the land surface. Microwave remote sensing at C- and L-bands provides soil moisture estimates, due to its high sensitivity to water content changes in the soil. The performances of the soil moisture estimates from retrieval and assimilation algorithms using microwave observations rely on realistic estimates of microwave signatures, backscattering (s0) or brightness temperature (TB) from microwave backscattering or emission models. This dissertation explores the impacts of soil moisture distribution within the near-surface and the soil characteristics on the state-of-the-art forward models that fail to reliably relate the near-surface soil moisture to observed s0 and TB. Concurrent passive microwave observations at C- and L-band were obtained from a previous conducted experiment in 2006. In addition, a field experiment was conducted to obtain simultaneous active and passive (AP) observations at L-band in 2012. These dataset were obtained at unprecedented high temporal resolution, every 15 minutes, from sandy, bare soils during highly dynamic periods. Procedures for AP sensor calibration were developed during the experiments. A methodology was implemented using dual-polarized C-band observation to estimate physically consistent soil parameters, for an irrigation event and subsequent drydown. These derived parameters were used in conjunction with the in situ moisture at deeper layers and different moisture profiles within the moisture sensing depth to obtain estimates of H-pol TB at L-band, that improve the RMSDs of TB estimates by 15 K during drydown periods. Furthermore, the complementarity of AP signatures was investigated by evaluating the sensitivity of s0 and emissivity (ep) to observed soil moisture and roughness measurements. It was found that the ep is consistently sensitive to the soil moisture on smooth and rough soil, but largely insensitive to surface roughness, in contrast to s0. Such complementarity of AP was utilized to estimate the soil moisture within moisture sensing depth using TB, while surface roughness was estimated from s0. These derived soil parameters provided physically consistent estimations of s0HH, s0VV and TB with RMSDs of 1.47 and 1.24 dB, and 4.55 K, respectively, with respect to the observations during rough surface period.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Pang-Wei Liu.
Thesis: Thesis (Ph.D.)--University of Florida, 2013.
Local: Adviser: Judge, Jasmeet.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2013
System ID: UFE0045310:00001

Permanent Link: http://ufdc.ufl.edu/UFE0045310/00001

Material Information

Title: Integration of Active and Passive Microwave Signatures for Characterization of Soil Properties
Physical Description: 1 online resource (124 p.)
Language: english
Creator: Liu, Pang-Wei
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2013

Subjects

Subjects / Keywords: microwave -- moisture -- radar -- radiometer -- remote -- sensing -- soil
Agricultural and Biological Engineering -- Dissertations, Academic -- UF
Genre: Agricultural and Biological Engineering thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Soil water is one of the key factors governing water and energy fluxes at the land surface. Microwave remote sensing at C- and L-bands provides soil moisture estimates, due to its high sensitivity to water content changes in the soil. The performances of the soil moisture estimates from retrieval and assimilation algorithms using microwave observations rely on realistic estimates of microwave signatures, backscattering (s0) or brightness temperature (TB) from microwave backscattering or emission models. This dissertation explores the impacts of soil moisture distribution within the near-surface and the soil characteristics on the state-of-the-art forward models that fail to reliably relate the near-surface soil moisture to observed s0 and TB. Concurrent passive microwave observations at C- and L-band were obtained from a previous conducted experiment in 2006. In addition, a field experiment was conducted to obtain simultaneous active and passive (AP) observations at L-band in 2012. These dataset were obtained at unprecedented high temporal resolution, every 15 minutes, from sandy, bare soils during highly dynamic periods. Procedures for AP sensor calibration were developed during the experiments. A methodology was implemented using dual-polarized C-band observation to estimate physically consistent soil parameters, for an irrigation event and subsequent drydown. These derived parameters were used in conjunction with the in situ moisture at deeper layers and different moisture profiles within the moisture sensing depth to obtain estimates of H-pol TB at L-band, that improve the RMSDs of TB estimates by 15 K during drydown periods. Furthermore, the complementarity of AP signatures was investigated by evaluating the sensitivity of s0 and emissivity (ep) to observed soil moisture and roughness measurements. It was found that the ep is consistently sensitive to the soil moisture on smooth and rough soil, but largely insensitive to surface roughness, in contrast to s0. Such complementarity of AP was utilized to estimate the soil moisture within moisture sensing depth using TB, while surface roughness was estimated from s0. These derived soil parameters provided physically consistent estimations of s0HH, s0VV and TB with RMSDs of 1.47 and 1.24 dB, and 4.55 K, respectively, with respect to the observations during rough surface period.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Pang-Wei Liu.
Thesis: Thesis (Ph.D.)--University of Florida, 2013.
Local: Adviser: Judge, Jasmeet.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2013
System ID: UFE0045310:00001


This item has the following downloads:


Full Text

PAGE 1

INTEGRATIONOFACTIVEANDPASSIVEMICROWAVESIGNATURESFORCHARACTERIZATIONOFSOILPROPERTIESByPANG-WEILIUADISSERTATIONPRESENTEDTOTHEGRADUATESCHOOLOFTHEUNIVERSITYOFFLORIDAINPARTIALFULFILLMENTOFTHEREQUIREMENTSFORTHEDEGREEOFDOCTOROFPHILOSOPHYUNIVERSITYOFFLORIDA2013

PAGE 2

c2013Pang-WeiLiu 2

PAGE 3

Tomyparents,Shing-MouLiuandShu-JhenWu,andmywife,Hisu-WenTsai 3

PAGE 4

ACKNOWLEDGMENTS Thisdissertationistheresultofhardworkandsupportofmanypeople.Firstandforemost,Iwouldliketothankmygraduateadvisor,chairofsupervisorycommittee,Dr.JasmeetJudge,whohasbeenanenthusiasticeducatorandlifecoachthroughoutmygraduatestudy.Itiswithherinvaluableteaching,guidance,andinspiration,couldIhaveaccomplishedvariousgoalsinmygraduatestudy.Mysinceregratitudeisgiventomysupervisorycommitteemembers,Dr.WendyGrahamandDr.RobertMooreoftheUniversityofFlorida,Dr.RogerDeRoooftheUniversityofMichigan,andDr.AlejandroMonsivais-HuerteroofNationalinPolytechnicInstituteofMexico,fortheirsupportandinputtothisdissertation.IwasparticularlyprivilegedtohaveworkedwithDr.Anthony(Tony)EnglandandDr.RogerDeRoooftheUniversityofMichigan.TheweeklyteleconswereveryhelpfulandIthankthemfortheirvaluablesuggestionsandadvicesthatmademyresearchandthisdissertationmoremeaningfulandcomplete.Inaddition,IwouldliketoexpressmygratitudetoallmycolleaguesinCenterforRemoteSensingandDepartmentofAgriculturalandBiologicalEngineering,Mr.DanielPreston,Mr.PatRush,Mr.StevenFeagle,Ms.TaraBongiovanni,andDr.KarthikNagarajan,forallthetechnicalandnon-technicalsupportsandassistancesofeldworks.IalsothanktoJimBoyerandhisteaminthePlantScienceResearchandEducationUnitforeldpreparationandcropmanagement.Lastlybutnottheleast,Iwouldliketothankmyparents,Mr.Shing-MouLiuandMrs.Shu-JhenWu,andmywife,Hsiu-WenTsaifortheirendlesscareandsupportthroughoutthisprocess. 4

PAGE 5

TABLEOFCONTENTS page ACKNOWLEDGMENTS .................................. 4 LISTOFTABLES ...................................... 7 LISTOFFIGURES ..................................... 8 ABSTRACT ......................................... 10 CHAPTER 1INTRODUCTION ................................... 12 1.1MicrowaveRemoteSensingforSoilMoisture ................ 13 1.1.1PassiveMicrowaveTechnique ..................... 14 1.1.2ActiveMicrowaveTechnique ...................... 16 1.1.3IntegrationofActiveandPassiveMicrowaveTechniques ...... 18 1.1.4KnowledgeGapsinActiveandPassiveMicrowaveRemoteSensingforSoilMoisture ............................ 19 1.2ScienceQuestionsandDissertationObjectives ............... 20 1.3DissertationFormat .............................. 21 2MODELINGMICROWAVESIGNATURESFROMBARESOILS ......... 24 2.1MicrowaveBackscatteringModel ....................... 24 2.2DielectricConstantModel ........................... 25 2.3MicrowaveEmissionModel .......................... 28 2.4RoughSurfaceEmissivity ........................... 30 3MICROWAVE,WATERANDENERGYBALANCEEXPERIMENTS ....... 34 3.1FifthMicrowaveWaterandEnergyBalanceExperiment .......... 34 3.1.1UniversityofFloridaL-andC-bandMicrowaveRadiometers .... 35 3.1.2RadiometerCalibration ......................... 37 3.1.2.1Externalcalibration ..................... 37 3.1.2.2Internalcalibration ...................... 38 3.1.3SurfaceRoughnessMeasurements .................. 38 3.2EleventhMicrowaveWaterandEnergyBalanceExperiment ........ 39 3.2.1UniversityofFloridaL-bandMicrowaveRadiometer ......... 41 3.2.2UniversityofFloridaL-bandAutomatedRadarSystem ....... 42 3.2.2.1SubsystemsofUF-LARS .................. 43 3.2.2.2UniversityofFloridaLARSsignalprocessing ....... 44 3.2.2.3UniversityofFloridaLARSexternalcalibration ...... 47 3.2.2.4Fadingreduction ....................... 50 3.2.3RelationshipBetweenActiveandPassiveMicrowaveObservations 51 5

PAGE 6

4INTEGRATIONOFPASSIVEANDACTIVESIGNATURESFORMOISTUREDISTRIBUTIONINSANDYSOILS ......................... 72 4.1SoilMoistureDistributionUsingPassiveObservations ........... 72 4.1.1ComparisonofSoilEmissionModels ................. 72 4.1.2ExtractingSoilParametersUsingC-bandSignatures ........ 73 4.1.3EstimatingTBatL-band ........................ 75 4.1.4ResultsandDiscussions ........................ 76 4.1.4.1EstimatingsurfaceparametersfromC-bandsignatures 76 4.1.4.2ImpactsofsoilmoisturedistributiononpassivemicrowavesignaturesatL-band ..................... 77 4.2SoilMoistureSensitivityofActiveandPassiveMicrowaveSignatures ... 81 4.2.1Sensitivityof0toSoilMoisture .................... 82 4.2.2SensitivityofeptoSoilMoisture .................... 84 4.3IntegrationofActiveandPassiveMicrowaveSignatures .......... 85 4.3.1ActiveandPassiveMicrowaveSimulations .............. 86 4.3.2ResultsandDiscussion ........................ 87 4.4Summary .................................... 90 5CONCLUSIONS,CONTRIBUTIONS,ANDRECOMMENDATIONS ....... 109 5.1Conclusions ................................... 109 5.2Contributions .................................. 111 5.3RecommendationsforFutureResearch ................... 112 REFERENCES ....................................... 114 BIOGRAPHICALSKETCH ................................ 124 6

PAGE 7

LISTOFTABLES Table page 3-1MeasurementsofsoilpropertiesinPSREU .................... 52 3-2SpecicationsofUFLMRandUFCMR ....................... 52 3-3SoilsurfaceroughnessmeasurementsduringMicroWEX-5 ........... 52 3-4SoilsurfaceroughnessmeasurementsduringMicroWEX-11 .......... 52 3-5SpecicationsoftheUF-LARS ........................... 53 4-1ThesoilpenetrationdepthofC-andL-bandandthesensingdepthratio .... 92 4-2RMSDandSDbetweentheobservedandmodeledH-polTBatL-bandduringtwodrydownperiodsinMicroWEX-5 ........................ 93 4-3Measurementsofwaterinputduringthedrydowns ................ 94 4-4ofobserved0atHH,VV,andcross-poltoobservedsoilmoistureobserved 94 4-5RegressionResultsof0toVSM .......................... 94 4-6ofobservedepatH-poltoobservedsoilmoistureobserved .......... 94 4-7RegressionResultsofeptoVSM .......................... 94 4-8RMSDandSDbetweentheobservedandmodeledH-polTBandco-pol0atL-bandduringsmoothperiodinMicroWEX-11 .................. 95 4-9RMSDandSDbetweentheobservedandmodeledH-polTBandco-pol0atL-bandduringroughperiodinMicroWEX-11 ................... 96 7

PAGE 8

LISTOFFIGURES Figure page 1-1ComparisonofobservedTBfromMicroWEX-5withthoseestimatedfromcurrentforwardmodels .................................... 22 1-2Comparisonofobserved0atco-polfromMicroWEX-11withthoseestimatedfromcurrentforwardmodels ............................. 23 2-1Dielectricconstantsimulationsat1.25GHzintermsofVSM ........... 32 2-2Dielectricconstantsimulationsat1.25GHzintermsofsoilporosity ....... 32 2-3Theillustrationsoftheincoherentsolution,rst-order,andzero-orderapproximationsforestimatingemissionfromalayeredsoil ............ 33 3-1GeophysicallocationofeldandsensorslayoutofMicroWEX-5 ........ 54 3-2TheUFLMRsystem ................................. 55 3-3TheUFCMRsystem ................................. 56 3-4SoilmoistureandtemperatureobservationsduringMicroWEX-5 ........ 57 3-5Randomheightvariationonsurfaces ........................ 57 3-6Surfaceroughnessmeasurements ......................... 58 3-7BlockdiagramoftheUFLMR ............................ 59 3-8BlockdiagramoftheUFCMR ............................ 60 3-9TBobservationsduringMicroWEX-5 ........................ 61 3-10SensorslayoutofthebaresoilexperimentduringMicroWEX-11 ........ 62 3-11ObservedsoilmoistureandtemperatureduringbaresoilofMicroWEX-11 ... 63 3-12Surfaceroughnessmeasurementsusinggrid-board ............... 64 3-13TBobservationsduringbaresoilofMicroWEX-11 ................. 64 3-14TheUF-LARS .................................... 65 3-15BlockdiagramofUFLARS .............................. 65 3-16BlockdiagramoftheARFSub-system ....................... 66 3-17PostprocessingstepsinUF-LARS ......................... 67 3-18StatetransitiondiagramregulatingtheSDAsub-systemofUF-LARS ...... 67 8

PAGE 9

3-19Theowchartofstepstoestimate0 ....................... 68 3-20Denitionofgeometryoftrihedralcalibrationtarget ................ 68 3-21Convergencetestfortheeffectiveantennabeamwidthusingnormalizedantennagainpattern ...................................... 69 3-22Geometryofradarfootprints ............................ 70 3-23ObservationsofbackscatteringcoefcientsduringbaresoilofMicroWEX-11 70 3-24ScatterplotsofTBand0duringthebaresoilperiodinMicroWEX-11 ..... 71 3-25Scatterplotsof0atcross-andco-polduringthebaresoilperiodinMicroWEX-11 ..................................... 71 4-1ComparisonofemissionmodelsatL-band .................... 97 4-2Distributionofmoistureandtemperatureinamodiedsoillayersforthemicrowaveemissionmodel ............................. 98 4-3DielectricconstantsimulationsatC-andL-bandintermsofsoilporosity .... 98 4-4ComparisonofobservedTBfromMicroWEX-5withmodeledTBduringthedriestandthewettestperiodsatC-band ...................... 99 4-5ThecomparisonofobservedTBfromMicroWEX-5andbestestimatedTBatC-band ........................................ 100 4-6Thecomparisonofsoilmoistureestimatedinthetop0-2mmandobservationsat2cmand4cmduringMicroWEX-5 ....................... 100 4-7TheowchartofstepstoestimateH-polTBsatL-bandusingdifferentproles 101 4-8Thesoilmoistureprolesduringrstandseconddrydowns ........... 102 4-9ComparisonTBobservedfromMicroWEX-5tothoseestimatedusingestimatedsoilproles ...................................... 103 4-10Thescatterplotsofobserved0toVSMat2cmandsensitivityduringsmoothandroughperiods .................................. 104 4-11ThescatterplotsofobservedeptoVSMat2cmandsensitivityduringsmoothandroughperiods .................................. 105 4-12EstimatedVSMin0-2mmandtheVSMprolesduringsmoothperiod ..... 105 4-13ComparisonofobservedandestimatedTBand0duringsmoothperiod .... 106 4-14EstimatedVSMin0-2mmandtheVSMprolesduringroughperiod ...... 107 4-15ComparisonofobservedandestimatedTBand0duringroughperiod ..... 108 9

PAGE 10

AbstractofDissertationPresentedtotheGraduateSchooloftheUniversityofFloridainPartialFulllmentoftheRequirementsfortheDegreeofDoctorofPhilosophyINTEGRATIONOFACTIVEANDPASSIVEMICROWAVESIGNATURESFORCHARACTERIZATIONOFSOILPROPERTIESByPang-WeiLiuMay2013Chair:JasmeetJudgeMajor:AgriculturalandBiologicalEngineeringSoilwaterisoneofthekeyfactorsgoverningwaterandenergyuxesatthelandsurface.MicrowaveremotesensingatC-andL-bandsprovidessoilmoistureestimates,duetoitshighsensitivitytowatercontentchangesinthesoil.Theperformancesofthesoilmoistureestimatesfromretrievalandassimilationalgorithmsusingmicrowaveobservationsrelyonrealisticestimatesofmicrowavesignatures,backscattering(0)orbrightnesstemperature(TB)frommicrowavebackscatteringoremissionmodels.Thisdissertationexplorestheimpactsofsoilmoisturedistributionwithinthenear-surfaceandthesoilcharacteristicsonthestate-of-the-artforwardmodelsthatfailtoreliablyrelatethenear-surfacesoilmoisturetoobserved0andTB.ConcurrentpassivemicrowaveobservationsatC-andL-bandwereobtainedfromapreviousconductedexperimentin2006.Inaddition,aeldexperimentwasconductedtoobtainsimultaneousactiveandpassive(AP)observationsatL-bandin2012.Thesedatasetwereobtainedatunprecedentedhightemporalresolution,every15minutes,fromsandy,baresoilsduringhighlydynamicperiods.ProceduresforAPsensorcalibrationweredevelopedduringtheexperiments.Amethodologywasimplementedusingdual-polarizedC-bandobservationtoestimatephysicallyconsistentsoilparameters,foranirrigationeventandsubsequentdrydown.Thesederivedparameterswereusedinconjunctionwiththeinsitumoistureatdeeperlayersanddifferentmoistureproleswithinthemoisturesensingdepthtoobtainestimatesof 10

PAGE 11

H-polTBatL-band,thatimprovetheRMSDsofTBestimatesby15Kduringdrydownperiods.Furthermore,thecomplementarityofAPsignatureswasinvestigatedbyevaluatingthesensitivityof0andemissivity(ep)toobservedsoilmoistureandroughnessmeasurements.Itwasfoundthattheepisconsistentlysensitivetothesoilmoistureonsmoothandroughsoil,butlargelyinsensitivetosurfaceroughness,incontrastto0.SuchcomplementarityofAPwasutilizedtoestimatethesoilmoisturewithinmoisturesensingdepthusingTB,whilesurfaceroughnesswasestimatedfrom0.Thesederivedsoilparametersprovidedphysicallyconsistentestimationsof0HH,0VVandTBwithRMSDsof1.47and1.24dB,and4.55K,respectively,withrespecttotheobservationsduringroughsurfaceperiod. 11

PAGE 12

CHAPTER1INTRODUCTIONSoilwatercontentisacriticaldriverforhydrologicalprocessesandagoverningfactorincropgrowth.Accurateestimationofthesoilmoistureisessentialforpredictingmoistureuxessuchasevapotranspiration,surfacerunoff,inltration,andrecharge,aswellascropgrowthandyield[ 23 52 81 115 ].BecausetheDebyerelaxationfrequencyofwaterliesinthemicrowaveregion,thedielectricconstantexhibitslargedifferencesbetweenwetanddrysoils.Forexample,thedielectricconstantofwaterat1.4GHzisapproximately81,whilethatofdrysoilisabout4.Thus,observationsatmicrowavefrequenciesaresensitivetochangesinthewatercontentinthetoplayersofsoil.Particularly,observationsatthelowerfrequencies,<10GHz,suchasatL-(1-2GHz)andC-band(4-8GHz),areoptimalforsoilmoisturestudiesbecauseofnegligibleatmosphericattenuation,betterpenetrationthroughvegetation,andgreatersensitivitytomoisturethanathigherfrequencies.Satellite-basedactiveandpassivemicrowavesystemsallowfrequentobservationsofsoilmoisturewithglobalcoverage.Ingeneral,theactivesystemssuchastheradarsorscatterometers,provideobservationsatnerspatialresolutions.Forexample,aSyntheticApertureRadar(SAR)system,RADARSAT-2,developedbyCanadianSpaceAgencyprovidesquadpolarized(quad-pol)backscatteringobservationsatafrequencyof5.3GHz(C-band)ataspatialresolution3-50m[ 11 64 ].IndianSpaceResearchOrganizationlaunchedaSARsystem,RadarImagingSatellite1,whichwillmonitorglobalsoilmoistureusingquad-polobservationsatafrequencyof5.35GHz(C-band)ataspatialresolutionof3-50m[ 49 ].TheAdvancedSCATterometer(ASCAT)byEuropeanSpaceAgency(ESA)providesVV-polbackscatteringobservationsatafrequencyof5.26GHz(C-band)toretrievesoilmoistureatthespatialresolutionsof25and50km[ 6 ].InadditiontothetheseactiveobservationsatC-band,JapanAerospaceEXplorationAgency(JAXA)mountedanAdvancedMicrowaveScanningRadiometer-2(AMSR-2) 12

PAGE 13

onaGlobalChangeObservationMission-Water(GCOM-W)satellite,launchedinMay2012,toprovidesevenchannelsofpassivemicrowaveobservationsatfrequenciesfrom6.95to89GHz,includingtwochannelsofobservationsatC-band,6.95and7.3GHz,ataspatialresolutionof56km[ 51 ].AlthoughsoilmoisturestudieshavebeenconductedwithsuchC-bandsatellitesystemsandhavebeenreported[ 61 ],theL-bandsystems,atlongerwavelengths,arepreferableforthesoilmoisturestudiesparticularlyintheagriculturalregionsduetoitsbetterpenetrationthroughvegetation[ 4 10 68 ].TheSoilMoistureandOceanSalinity(SMOS)missionbyESA[ 54 ]providesdual-polpassivemicrowaveemissionobservationsat1.4GHzwithaspatialresolutionof40-50kmandarepeatcoverageof2-3days.TherecentlylaunchedAquariussensorbyNationalAeronauticsandSpaceAdministration(NASA)isequippedwithaquad-polscatterometerat1.26GHzandadual-polradiometerat1.41GHzprimarilyformonitoringthevariationsinthesurfacesalinityintheopenocean,aswellassoilmoistureataspatialresolution100km,witha7-dayrepeatcoverage[ 56 ].TheNASASoilMoistureActivePassive(SMAP)mission[ 34 ],scheduledforlaunchin2015,willcombineactiveandpassive(AP)observationsat1.26and1.41GHz,respectively,toprovideglobalsoilmoistureataspatialresolutionof9km,witharepeatcoverageof2-3days.AlthoughbothmicrowavebackscatterandemissionmeasuredbytheAPtechniques,respectively,aresensitivetosoilmoisture,theactivetechniqueishighlysensitivetosoilroughnessandvegetation[ 4 30 44 60 78 ],effectivelymaskingmuchofsignalchangesduetomoisturevariationsinthesoil.UsinganAPsystemsuchasNASA-SMAP,thesensitivitytosoilmoisturemaybeobtainedfromthepassivetechniqueatcoarserspatialresolutionsoftheradiometer,whileradarmaybeusedtoestimatethesubpixelheterogeneityofthelandsurface. 1.1MicrowaveRemoteSensingforSoilMoistureSoilmoistureestimatesfromAPmicrowaveobservationsusingeitherretrievalorassimilationalgorithmsrelyonaccuratemicrowavebackscatteringoremissionmodels. 13

PAGE 14

Overthelastthreedecades,manystudieshavebeenconductedtodevelopmicrowaveemissionmodelsforsoilmoisturestudiesfrompassiveobservations,whilelessprogresshasbeenmadetotheactivemicrowavetechnique[ 5 ].Thisisprimarilyduetoreducedsensitivityofactivesignaturestosoilmoisturecomparedtotheirsensitivitytosurfaceroughnessandtovegetatedlandsurfaces.Thissectionreviewsthecurrentstateofpassiveandactivemicrowavetechniquesforsoilmoisturestudies. 1.1.1PassiveMicrowaveTechniqueIngeneral,spectralBrightnessofablackbody,Bf,inWm)]TJ /F5 7.97 Tf 6.59 0 Td[(2Sr)]TJ /F5 7.97 Tf 6.58 0 Td[(1Hz)]TJ /F5 7.97 Tf 6.59 0 Td[(1ataspecicfrequency,f,isgivenbythePlank'sradiationlaw[ 100 ]: Bf=2hf3 c21 ehf=KT)]TJ /F10 11.955 Tf 11.96 0 Td[(1(1)where,hisPlank'sconstantinJoules,cisthespeedoflightinmsec)]TJ /F5 7.97 Tf 6.59 0 Td[(1,TisthephysicaltemperatureoftheblackbodyinKelvin(K),andKisBoltzmann'sconstantinJouleK)]TJ /F5 7.97 Tf 6.58 0 Td[(1,andInthemicrowaveregion,theBrightnessisdirectlyproportionaltoT,inaccordancewithRayleigh-Jeansapproximation,as: Bf=2f2KT c2=2KT 2(1)Thus,thebrightnessofablackbody,Bbb,inWm)]TJ /F5 7.97 Tf 6.59 0 Td[(2Sr)]TJ /F5 7.97 Tf 6.58 0 Td[(1atanarrowbandwidthoffisdened: Bbb=Bff=2KT 2f(1a)Similarly,theBrightnessofanon-blackbodytarget,B,isalsodirectlyproportionaltoablackbodyequivalentradiometrictemperature,calledbrightnesstemperature(TB),as: B=2KTB 2f(1b)Thus,TBistypicallyusedtorepresenttheradiancefromtargetinmicrowaveremotesensing. 14

PAGE 15

TheTB,pobservedbyradiometeratpolarization,p,fromanunvegetatedlandsurfaceconsistsofcontributionsfromsoil(TBsoil,p)andsky(TBsky,p),givenas: TB,p=TBsoil,p+TBsky,p=epTe+rpTsky (1) where,pisthepolarizationeitherH-orV-pol,Tskyisdownwellingskybrightnesstemperature,usuallysetto5-8KformicrowaveatL-andC-band[ 102 ],TeistheeffectivetemperatureofbaresoilinK,episemissivityofthesoil,andrpisreectivityofsoilequalto(1)]TJ /F6 11.955 Tf 12.11 0 Td[(ep).Manystudieshavebeenconductedtounderstandtherelationshipbetweenthepassiveobservationsandthesoilmoisture,forexamplein[ 63 67 74 75 80 85 88 92 106 108 110 ],andprovidevaluableinformationfordevelopingnumericaloranalyticalmodelsofTeandep.Teisdenedasanintegralofradiativetemperaturesoversoilmedium[ 85 102 ],givenas: Te=Z0T(z)s(z)exp)]TJ /F11 11.955 Tf 11.29 16.28 Td[(Z0zs(z0)dz0dz(1)where,zisthedepthofsoil,T(z)isphysicalsoiltemperatureinK,sisabsorptioncoefcientofsoildenedass=4 Imp ,whereisthewavelengthandisthedielectricconstantofsoil.Thesoilisassumedasanon-isothermal,semi-innitelayereddielectricmedium,thezero-andrst-orderapproximationstotheintegralareusedtoestimatetheTe[ 9 85 ].However,undermostconditionsthesoilmoistureandtemperatureprolesarenotavailable.Thus,mostimplementationsuseempiricalformulationstoestimateTeasalinearfunctionofsoiltemperaturesatsurfaceandatadeeperlayer,usually50cm[ 15 20 ].Wigneronetal.(2007)[ 109 ]modiedsuchformulationtoincludesoilmoistureinthetop3cm,alsoknownasnear-surfacesoilmoisture.Thesoilemissivity,ep,isafunctionofsurfaceroughnessanddielectricconstant,,ofmoistsoil.Itcanbeestimatedusingeitherempirical[ 107 ],semi-empirical 15

PAGE 16

modelssuchastheQ-hmodel[ 106 ]anditsmodications[ 63 85 106 109 ],ormorephysically-basedmodelssuchasIntegralEquationModel(IEM)anditsmodications[ 18 39 ].Theisafunctionofsoilmoistureandphysicalproperties.Typically,themoistsoilismodeledasacompositionofsoilsolids,water,andair.Dobsonetal.(1985)[ 25 ]estimatedtheofwetsoilforthefrequenciesfrom1.4to18GHzwithasemi-empiricalmodelwhichconsideredthemoistsoilasacompositionofsoilsolids,boundwater,freewater,andair.Peplinskietal.(1995)[ 76 ]extendedtheapplicabilityoftheDobsonetal.(1985)modelforfrequencyrangeof0.3-1.3GHz.Mironovetal.(2009)[ 62 ]proposedamineralogically-basedmodelwithrefractivemixingequationstoestimatetheforthefrequencyrangeof45MHz-26.5GHzapplicableformorevarietiesofsoilsuchasthosewithsandcontentfrom2to98%byvol. 1.1.2ActiveMicrowaveTechniqueThebasicradarequationdescribestherelationshipbetweenthepowerreceived(Pr)andtransmitted(Pt)as[ 101 ]: Pr=PtG22 (4)3R4(1)where,GistheGainoftheantenna,Ristherangebetweenantennatothetargetinm,isthewavelengthofsignalinm,andistheradarcrosssection(RCS)ofthetargetinm2.TheRCSisdeterminedbydielectricproperty,size,andshapeoftarget.Foradistributedtargetwithsmallscattererssuchasthelandsurface,thetotalreceivedsignal(Pr)couldbeconsideredasasummationofsignal-returnsfrommultipleindividualscatterers,givenas: Pr=2 (4)3NXi=1PtiG2ii R4i(1) 16

PAGE 17

where,iistheithscattereronthelandsurface.Adifferentialscatteringcoefcient,alsoknownasnormalizedbackscatterorbackscatteringcoefcient(0),isdenedby: 0=1 NNXi=1i Ai(1)where,iisRCSofithscattereronthelandsurface,andAiistheilluminationareaontheithscatterer.Theradarorscatterometersystemsobservethepowerratioofpowerreceivedatpolarization,p,tothepowertransmittedatpolarization,q,andthe0pqisobtainedbycomparingthepowerratioobtainedusingacalibrationtargetwithknownRCS.The0pqisafunctionofincidenceangle,frequencyofthesensor,targetproperties,surfaceroughness,anddielectricconstant,.Earlystudies,forexample[ 24 94 97 ],foundthatco-polarized,0HHand0VV,isalinearfunctionofvolumetricsoilmoisture(VSM)foragivensoilcondition,wheretheslopeandinterceptaredependentuponthesoiltextureandroughness,respectively.Suchsimplisticregression-basedmodelsresultininconsistentsoilmoistureestimatesunderdifferentsoilconditions[ 90 ],duetolackofinsitueldmeasurementsfordevelopingacomprehensivelook-uptable.Moresophisticatedempiricalandsemi-empiricalmodels[ 31 72 ]areproposedbyrelating0HHand0VVto,estimatedfrommodels,ortoreectivityestimatedfrom,withthesurfaceroughness,rootmeansquareheight,s.Modicationsofsemi-empiricalmodelsweredeveloped(forexamplein[ 69 71 ]),involvingbothsoilroughness,s,andcorrelationlength(cl)inmodels,andextendingforcross-pol,0HVand0VH.Morephysically-basedbackscatteringmodelsbaseduponelectromagnetictheoryandboundaryconditions,suchastheKirchhoffApproximation(KA),SmallPerturbationMethod(SPM)[ 93 ],SmallSlopeApproximation(SSA)[ 105 ],andIEM[ 39 42 ]havebeendevelopedforsoilmoisturestudies.Inthesemodels,the0pqtypicallyconsistsofasinglescattering(Spq)andamultiplescattering(Mpq)terms: 0pq=Spq+Mpq(1) 17

PAGE 18

TheMpqhasbeenfoundinsignicant,<1dB,whenterrainshavesmallrmsslopewhichistrueformostnaturalterrains[ 18 19 42 ].ThecurrentmodicationsofIEMfocusonimprovingtheSpq[ 18 40 111 112 ]tobetterestimate0pq. 1.1.3IntegrationofActiveandPassiveMicrowaveTechniquesManysoilmoistureproductsareavailablefromeitheractiveonlyorpassiveonlymicrowaveobservationsusingsatellite,airborne,andground-basedsystems.Studies,comparingsensitivitiesofAPmicrowaveobservationstosoilmoisture[ 30 68 79 ]orsoilmoistureestimationstoinsitumeasurements[ 2 8 43 83 ],suggestthatAPsignaturesprovidecomplementaryinformationregardingsoilandvegetation.Typically,thepassivetechniqueprovidesbettersoilmoistureestimateswithcoarserspatialresolution,whilethesoilmoisturederivedfromactiveobservationshashigheruncertaintyandhighlysensitivetothevegetationandlandsurfacestructurewithnnerspatialresolution[ 34 ].UtilizingsuchcomplementaryinformationandintegratingAPobservationshavebeendemonstratedtoimprovesoilmoistureestimatesparticularlyduringthevegetationgrowingseason[ 22 58 73 79 ].AfewstudieshaveintegratedAPalgorithmsinvegetatedlandsurfacesbyusingabackscatteringmodellinkedwithanemissionmodelforsoilmoistureestimates[ 16 17 37 ].StudiesarebeingconductedtodevelopandtestalgorithmstocombinetheAPmicrowavesignaturesatdifferentspatialresolutionstoobtainsoilmoistureestimatesfromtheupcomingNASA-SMAPmission,providingsimultaneousAPmicrowaveobservationsatL-bandwithdifferentspatialresolutionsof40and3km,respectively[ 21 58 77 ].However,theprogressintegrationAPtechniquesisstillcomparativelylessmaturethanusingactiveonlyorpassiveonlyduetolackoflong-termconcurrentAPobservationsathightemporalfrequencytocalibrateandtesttheretrievalalgorithmsunderdynamicmoistureconditions. 18

PAGE 19

1.1.4KnowledgeGapsinActiveandPassiveMicrowaveRemoteSensingforSoilMoistureThegapsstillremaininthecurrentdevelopmentofAPmodelseventhoughmanypreviouseffortshavesignicantlyadvancedourunderstandingforsoilmoisturestudies.First,thequalityofsoilmoistureestimatesusingempiricalandsemi-empiricalmodelsareusuallysitedependentbecauseoftheregressionparametersusedtodevelopthesemodels.Althoughtheseempiricalandsemi-empiricalmodelsprovidesimpleralgorithmsforretrievalofsoilmoistureandhavebeenusedforthecurrentsatellitemissions[ 28 29 31 47 72 90 109 ],theuncertaintyofmodelparametersintheseapproachesarehighlydependentuponconcurrentmicrowaveandsoilmoistureobservationsusedduringtheparameterregressions.Second,thecurrentstudiesusingthesoilmoistureaveragedovertop0-2or0-5cm,callednear-surfacemoisture,forepand0pqarenotadequateandmayresultinunrealisticTBand0pqestimation.Thesemoisturevaluesinthe0-2or0-5cmareobtainedeitherfrominsitumeasurementsorfromlandsurfacemodels(LSMs)[ 12 32 82 ].IncreasingtheverticalresolutionintheLSMsinvariablyincreasescomputationaldemandsforoperationaluse.However,thecurrentinsitusensorsarenotabletomeasuremoisturewithhighverticalresolutionfromsurfacetoabout2cmofthesoil.Suchmeasurementscannotrepresentthehighlydynamicproleofsoilmoistureintheupperfewcentimetersforsandysoils,particularlyduringandimmediatelyfollowinghydrologicevents.Forexample,Figures 1-1 A-Cshowthatcurrentstate-of-the-artformulationsareunabletoproviderealisticestimationsofTBatL-andC-bandwhencomparedwiththeobservedTBvaluesforsandysoils.TheTBwereestimatedusingthefrom[ 62 ],Tefrom[ 109 ],epfrom[ 107 108 ],and[ 39 ],andusingeldobservationsofsoilmoistureat0-2cm,temperatureat0-2and64cm,andsurfaceroughnessinsandysoilsofNorthCentralFloridafromthe2006experiment,mentionedinChapter3[ 14 ].DuringthesimulationperiodfromMarch28,DayofYear(DoY) 19

PAGE 20

87.5,toApril4(DoY94.5)in2006,theeldwasirrigatedonDoY88.03andonDoY90.83.Similarly,0estimateswerealsounrealisticusingthecurrentstate-of-the-artbackscatteringmodels,asshowninFigures 1-2 AandB.Theco-pol0wereestimatedusingsemi-empiricalmodelfrom[ 69 ]andphysically-basedmodelfrom[ 18 ],thefrom[ 62 ],andeldobservationsofsoilmoistureat0-2cmandsurfaceroughnessmeasuredduring2012experiment,mentionedinChapter3.DuringthesimulationperiodfromMay27(DoY148)toJune5(DoY157)in2012,precipitationoccurredonDoY148.95,151.39,and153.18.Inaddition,theimpactofsoilpropertieswithinthe0-2or0-5cmonTBand0pqisnotwellunderstood.Severalstudieshaveuseddifferentapproachestoobtainrealisticestimates[ 27 86 ].Forexample,Schneebergeretal.(2004)[ 86 ]addedatransitionlayeronthetopsoilandobtainedplausibledielectricconstantbychangingsoilporosityinthetransitionlayer.SuchapproachimprovedtheconsistencybetweenthemodeledandobservedTB.However,thedielectricheterogeneitiesinthesoilarenotonlyimpactedbyporosity,butalsoimpactedbymoisturedistributionandanisotropiesinthetopsoil[ 87 ],andimprovingonlytheporosityatthesurfacemaynotbesufcientforaccurateestimationofsoilemission.Third,asignicantgapstillexistsinthedevelopmentandevaluationofphysically-basedmethodologiesthattakeadvantageoftheAPobservationsinregionswithhighdynamicchangeofsoilmoisture,primarilyduetolackofconcurrenteldobservationsofAPmicrowavewithhightemporalfrequency. 1.2ScienceQuestionsandDissertationObjectivesThisdissertationfocusesonexploringtheimpactsofsoilmoisturedistributioninthenearsurface,porosity,andsurfaceroughnessontheAPmicrowavesignatures.Thesciencequestionsaddressedinthisdissertationare: 1. Whatarethesourcesthatmaycauseunrealisticestimatesofmicrowavesignaturesusingcurrentmicrowaveforwardmodelsfornear-surfacesoilmoistureestimates? 20

PAGE 21

2. WhatisthesensitivityoftheAPmicrowavesignaturestothesoilmoistureandsurfaceroughness? 3. HowcanthecomplementarityofAPmicrowavesignaturesbeusedforprovidingsoilmoistureestimatesfortheupcomingNASA-SMAP?Thisdissertationhasvemainobjectives:1)conductingintensiveeldexperimentstoobtainAPmicrowaveobservationsatL-bandfromanagriculturaleld,withconcurrentsoilmoistureandtemperatureobservationsathightemporalresolutionofevery15minutes;2)developingdataprocessingandcalibrationproceduresfortheAPmicrowaveobservations;3)evaluatingtheimpactsofmoisturedistributionunderbaresoilconditionsonmicrowaveemissionusingdual-polC-bandandH-polL-bandradiometricobservations;4)evaluatingthesensitivityofAPobservationstoobservedsoilmoistureandsurfaceroughness;and5)utilizingthecomplementarityofAPmicrowaveobservationsforimprovementofsoilmoistureestimates. 1.3DissertationFormatTheChapter2discussesthephysically-basedmicrowavebackscatteringandemissionforwardmodelsusedinthisdissertation.Chapter3describestheeldobservationsduringtheMicrowaveWaterandEnergyBalanceExperiments(MicroWEXs)anddescriptionsofthesensorcalibrationalgorithmsfortheAPmicrowavesystems.Inchapter4,thecurrentmicrowavemodelsareevaluatedbyeldobservationstounderstandthepotentialsourcesresultinginunrealisticmicrowavesignatureestimations.ImprovementsforL-bandTBestimatesbyusingsoilmoistureprolewithinmicrowavesensingdepth(MSD)estimatedfromC-bandsignaturesareproposed.Inaddition,thesensitivityanalysesofAPsignaturestosoilmoistureandroughnessareconducted,andcomplementarityofAPmicrowavesignatureatL-bandareinvestigatedandutilizedforimprovementofsoilmoistureestimates.Chapter5providesconclusionsofthisdissertationandrecommendationsforthefutureresearch. 21

PAGE 22

Figure1-1. ComparisonofobservedTBfromMicroWEX-5withthoseestimatedfromcurrentforwardmodels.A)L-bandH-pol.B)C-bandH-pol.C)C-bandV-pol.Forthissimulation,thes=0.55cm,cl=6.83cm,andsoilporosity=0.37. 22

PAGE 23

Figure1-2. Comparisonofobserved0atco-polfromMicroWEX-11withthoseestimatedfromcurrentforwardmodels.A)HH-pol.B)VV-pol.Forthissimulation,thes=1.71cm,cl=9.45cm,andsoilporosity=0.37. 23

PAGE 24

CHAPTER2MODELINGMICROWAVESIGNATURESFROMBARESOILSThischapterprovidesdescriptionsofmicrowavebackscatteringandemissionmodelsforbaresoilusedinthisdissertation. 2.1MicrowaveBackscatteringModelInthisdissertation,anadvancedIEM(AIEM)[ 18 ]thatincludesamorecompleteexpressionoftheSpqthantraditionalIEM[ 42 ]isusedtomodelthebackscatteringfromaroughagriculturaleld,givenas: Spq=k21 2exp)]TJ /F6 11.955 Tf 9.3 0 Td[(s2(k2iz+k2sz)1Xn=1s2n n!Inpq2W(n)(ksx)]TJ /F6 11.955 Tf 11.95 0 Td[(kix,ksy)]TJ /F6 11.955 Tf 11.95 0 Td[(kiy)(2a)where,k1isthewavenumberintheair,sisthermsheightofterrainsurface,andW(n)istheFouriertransformofthenthpowerofthenormalizedsurfacecorrelationfunction.ThisstudyusedexponentialcorrelationfunctionbecauseitismorerepresentativetothenaturalterrainthanGaussianfunction[ 41 72 ].TheInpqisdenedas:Inpq=(ksz+kiz)nfpqexp()]TJ /F6 11.955 Tf 9.29 0 Td[(s2Kizksz)+1 4nF(+)pq1(ksz)]TJ /F6 11.955 Tf 11.96 0 Td[(q1)nexp)]TJ /F6 11.955 Tf 9.3 0 Td[(s2(q21)]TJ /F6 11.955 Tf 11.96 0 Td[(q1(ksz)]TJ /F6 11.955 Tf 11.96 0 Td[(kiz))+F(+)pq2(ksz)]TJ /F6 11.955 Tf 11.96 0 Td[(q2)nexp)]TJ /F6 11.955 Tf 9.3 0 Td[(s2(q22)]TJ /F6 11.955 Tf 11.96 0 Td[(q2(ksz)]TJ /F6 11.955 Tf 11.96 0 Td[(kiz))+F()]TJ /F5 7.97 Tf 6.59 0 Td[()pq1(ksz+q1)nexp)]TJ /F6 11.955 Tf 9.3 0 Td[(s2(q21+q1(ksz)]TJ /F6 11.955 Tf 11.96 0 Td[(kiz))+F()]TJ /F5 7.97 Tf 6.58 0 Td[()pq2(ksz+q2)nexp)]TJ /F6 11.955 Tf 9.3 0 Td[(s2(q22+q2(ksz)]TJ /F6 11.955 Tf 11.95 0 Td[(kiz))ou=)]TJ /F7 7.97 Tf 6.58 0 Td[(kix,v=)]TJ /F7 7.97 Tf 6.59 0 Td[(kiy+1 4nF(+)pq1(kiz+q1)nexp)]TJ /F6 11.955 Tf 9.3 0 Td[(s2(q21)]TJ /F6 11.955 Tf 11.96 0 Td[(q1(ksz)]TJ /F6 11.955 Tf 11.96 0 Td[(kiz))+F(+)pq2(kiz+q2)nexp)]TJ /F6 11.955 Tf 9.3 0 Td[(s2(q22)]TJ /F6 11.955 Tf 11.96 0 Td[(q2(ksz)]TJ /F6 11.955 Tf 11.96 0 Td[(kiz))+F()]TJ /F5 7.97 Tf 6.59 0 Td[()pq1(ksz)]TJ /F6 11.955 Tf 11.96 0 Td[(q1)nexp)]TJ /F6 11.955 Tf 9.3 0 Td[(s2(q21+q1(ksz)]TJ /F6 11.955 Tf 11.96 0 Td[(kiz))+F()]TJ /F5 7.97 Tf 6.58 0 Td[()pq2(kiz)]TJ /F6 11.955 Tf 11.95 0 Td[(q2)nexp)]TJ /F6 11.955 Tf 9.3 0 Td[(s2(q22+q2(ksz)]TJ /F6 11.955 Tf 11.95 0 Td[(kiz))ou=)]TJ /F7 7.97 Tf 6.58 0 Td[(ksx,v=)]TJ /F7 7.97 Tf 6.59 0 Td[(ksy (2b)where,fpqandFpqareKirchhoffandcomplementaryeldcoefcients,respectively,andtheyarefunctionsofFresnelreectioncoefcientsdenedin[ 18 ].Kix,Kiy,Kiz,Ksx,Ksx, 24

PAGE 25

Ksxaregivenas:kix=k1sinicosi;ksx=k1sinscosskiy=k1sinisini;ksy=k1sinssinskiz=k1cosi;ksz=k1cossqm=(k2m)]TJ /F6 11.955 Tf 11.96 0 Td[(u2)]TJ /F6 11.955 Tf 11.95 0 Td[(v2)1 2 (2c)where,themiseither1or2presentingmediaairandsoil,respectively,kisthewavenumber,andareelevationandazimuthangles,respectively,andthesubscripts,iands,representincidenceandscatteringdirectionsofwavepropagation,respectively.Incomputation,the0pqinAIEMisafunctionofincidenceangle,frequency,soilroughnesssandcl,andofsoil.Themodelswillbedescribedindetailinsection 2.2 .Inaddition,ithasbeenfoundthatthecurrentAIEMmodelestimating0atcross-polstillneedstobeimproved.Forexample,avalidationofAIEMusingRADARSAT-2wasconductedin[ 27 ]andfoundthattheRMSEwithrespectto0HVmeasurementisupto30dB,whichismuchhigherthanthoseof0HHand0VVof2.5and3.1dB,respectively.Therefore,only0atco-polwillbeincludedforfurthermodelinganalysesinthisdissertation. 2.2DielectricConstantModelIngeneral,dielectricconstant,,ofmoistsoilscanbedescribedasafunctionofitsconstituentsincludingsoilsolids,freeandboundwater,andair,given[ 101 ]: soil=ssss+fwfw+bwbw+aa(2)where,thesubscriptssoil,ss,fw,bw,andarepresentmoistsoil,soilsolids,freewater,boundwater,andair,respectively,isaconstantshapefactor,andanddenotevolumefractionsandthedielectricconstantofeachcomponent,respectively.Semi-empiricalmodels,developedbyDobsonetal.(1985)andPeplinskietal.,(1995)[ 25 76 ],arewidelyusedtoestimatetheofmoistsoilforfrequenciesfrom0.3to18 25

PAGE 26

GHz.Recentstudieshavefoundsuchmodelssignicantlyoverestimateofsandysoils[ 62 109 ],duetotheirlackofwiderangeofvarietiesofsoiltypesusedforparametercalibrations.Mironovetal.(2009)[ 62 ]proposedarefractivemixingmodeltoestimateofmoistsoilsfrequenciesfrom45MHzto26.5GHz,shownas: nsoil=8><>:nd+(nbw)]TJ /F10 11.955 Tf 11.95 0 Td[(1)mvmvmvtnd+(nbw)]TJ /F10 11.955 Tf 11.95 0 Td[(1)mvt+(nfw)]TJ /F10 11.955 Tf 11.95 0 Td[(1)(mv)]TJ /F6 11.955 Tf 11.96 0 Td[(mvt)mv>mvt(2a) soil=8><>:d+bwmvmvmvtd+kbwmvt+fw(mv)]TJ /F6 11.955 Tf 11.96 0 Td[(mvt)mv>mvt(2b) soil=)]TJ /F6 11.955 Tf 5.48 -9.68 Td[(n2soil)]TJ /F6 11.955 Tf 11.95 0 Td[(k2soil+i(2nsoilsoil)(2c)where,nistherefractiveindex(RI),isthenormalizedattenuationcoefcient(NAC),thesubscriptssoil,d,bw,andfwrefertothemoistsoil,drysoil,boundwater,andfreewater,mvisthesoilwatercontentinm3=m3,andmvtisthefractionofthemaximumboundwaterinm3=m3.TheboundandfreewatercomponentscanbeacquiredbyDebyerelaxationequations,withrelaxationtimesestimatedbaseduponsoiltype.Mineralogicallybasedequationsfor,nd,d,andmvt,wereformedbaseduponawidevarietiesofsoiltypes[ 62 ],givenas: nd=1.634)]TJ /F10 11.955 Tf 11.95 0 Td[(0.53910)]TJ /F5 7.97 Tf 6.59 0 Td[(2C+0.274810)]TJ /F5 7.97 Tf 6.59 0 Td[(4C2(2a) d=0.03952)]TJ /F10 11.955 Tf 11.96 0 Td[(0.0403810)]TJ /F5 7.97 Tf 6.59 0 Td[(2C(2b) mvt=0.02863+0.3067310)]TJ /F5 7.97 Tf 6.59 0 Td[(2C(2c)whereCistheclaycontentofsoilin%byvol.However,ndanddarethecombinationsofsolidsoilandair.ThemineralogicallybasedEquations 2a and 2b ,dependingonlyuponthesoiltype,cannotrepresentthechangeofvolumefractionofthesoilsolidsunderdifferentsoilporosity.Therefore, 26

PAGE 27

thedrysoilcomponentndanddcanbewrittenintermsofsoilbulkdensity(b)andspecicdensityofsolidsoilparticles(s),givenas:nd=(nss)]TJ /F10 11.955 Tf 11.95 0 Td[(1)b s+1d=ssb s (2)ThendanddarereplacedandEquations 2a and 2b canbewrittenas: nsoil=8><>:nssb s+nbwmv+amvmvtnssb s+nbwmvt+nfw(mv)]TJ /F6 11.955 Tf 11.96 0 Td[(mvt)+amv>mvt(2a) soil=8><>:ssb s+bwmvmvmvtssb s+bwmvt+fW(mv)]TJ /F6 11.955 Tf 11.96 0 Td[(mvt)mv>mvt(2b)where,aisthevolumefractionofairfrom(1)]TJ /F6 11.955 Tf 12.6 0 Td[(mv)]TJ /F12 7.97 Tf 13.8 5.26 Td[(b s),bisacquiredfromthesoiltextureanalysising=cm3,sissetto2.65g=cm3inthisstudy,andnssandssarethereal(RI)andimaginary(NAC)partsofp ss,givenas[ 25 ]: ss=(1.01+0.44s)2)]TJ /F10 11.955 Tf 11.96 0 Td[(0.062(2)mvtisobtainedfromEquation 2c ,andnb,b,nfandfcanbecalculatedfromtheDebyerelaxationequations.Figure 2-1 AshowsthecomparisonofestimatedusingPeplinskietal.(1995),Mironovetal.(2009),andthemodiedmodelatfrequencyof1.25GHz.Typically,therealpartofismuchlargerthantheimaginarypart,andisdominantinthecomputationofFresnelreectivityforsurfacebackscatteringandemissionmodels.TheestimationfromPeplinskietal.(1995)issignicantlyhigherthanothertwomodelsbyapproximate2-17inrealpartofasVSMfrom0.04to0.5m3=m3.0HHand0VVweresimulatedbyusingAIEMwithroughness,s=0.55andcl=6.83cm,similartoasmoothagriculturaleld.Figure 2-1 Bshowsthatfrom[ 76 ]overestimated0tomodiedmodelbyasmuchas4dBatHH-andVV-pol,respectively.Thedifferencesof0estimatesfromMironovetal.(2009)andmodiedmodelareonly 27

PAGE 28

within1dBatHH-andVV-pol,respectively.Figure 2-2 AshowsthatthedecreaseswiththesoilporosityincreasingwhenVSM=0.1,0.2,0.3,and0.4m3=m3.Assamesoilconditionasabove,Figure 2-2 Bpresentsthechangeofsoilporosityfrom0.1to0.6resultingindifferenceof0HHand0VVof3dBand2dBfordryandwetsoil,respectively. 2.3MicrowaveEmissionModelAsmentionedinEquation 1 ,TBfromanunvegetatedlandsurfaceisestimatedasthesumofcontributionsfromsoilandsky.ThemicrowaveemissionmodelsforthesoilrequireknowledgeofmoistureandtemperaturedistributioninthesoiltoestimateTBsoil.Typically,thesoilisassumedasanon-isothermal,semi-innitelayereddielectricmedium,witharoughsurfaceattheupperboundary.TheincoherentsolutionforTBaccountsforreectionsatlayer-interfaces,andthepropagationoftheradiancethrougheachlayer,asshowninFigure 2-3 .Thebrightnesstemperatureabovealayerj,TBj,isgivenas: TBj=8><>:T+BjT)]TJ /F7 7.97 Tf -1.59 -8.28 Td[(Bj9>=>;=Bj,j+1TBj+1+Sj+1(2a)where,T+BjandT)]TJ /F7 7.97 Tf -1.59 -8.28 Td[(Bjareupwellinganddownwellingbrightnesstemperaturesimmediatelyabovethejthinterface,respectively.Bj,J+1,thebackwardpropagationmatrixdenedinEquation 2b ,representsreectionfromthej+1layerreachingthejthlayerandattenuationduetoabsorptionwithinlayerj+1.Thesecondterm,Sj+1,istheemissioncontributionfromthej+1thlayer,denedinEquation 2c Bj,j+1=1 1)]TJ /F6 11.955 Tf 11.96 0 Td[(Rj8><>:(1)]TJ /F5 7.97 Tf 6.59 0 Td[(2Rj) Lj+1RjLj+1)]TJ /F7 7.97 Tf 6.59 0 Td[(Rj Lj+1Lj+19>=>;(2b) Sj+1=Tj+11)]TJ /F5 7.97 Tf 18.82 4.71 Td[(1 Lj+1 1)]TJ /F6 11.955 Tf 11.96 0 Td[(Rj8><>:1)]TJ /F6 11.955 Tf 11.96 0 Td[(Rj(2+Lj+1))]TJ /F10 11.955 Tf 9.3 0 Td[((Rj+Lj+1)9>=>;(2c)where,RjistheamplitudeofFresnelreectionattheinterfacebetweenthejthandthej+1thlayers,Tj+1isthephysicaltemperatureofthej+1thlayer,andLj+1istheloss 28

PAGE 29

factorofthej+1thlayer,expressedasLj+1=exp(sj+14zj+1secj+1),wheresj+1istheabsorptioncoefcientofsoilatthej+1thlayer,4zj+1isthethicknessofthej+1thlayer,andj+1istherefractiveanglebetweentheinterfacesofthejthandj+1th.Theupwellinganddownwellingbrightnesstemperaturesofthebottom,semi-innitelayer,N,asshowninFigure 2-3 ,isaspecialcase,wheretheupwellingbrightnesstemperature,T+Bh,isequaltothephysicaltemperatureoftheNthlayer,TN,andthedownwellingbrightnesstemperatures,T)]TJ /F7 7.97 Tf -1.59 -8.28 Td[(Bh,isunknown.Sothat,TBN=BNTBh,where: BN,h=1 1)]TJ /F6 11.955 Tf 11.96 0 Td[(RN8><>:(1)]TJ /F10 11.955 Tf 11.95 0 Td[(2RN)RN)]TJ /F6 11.955 Tf 9.3 0 Td[(RN19>=>;(2d) TBh=8><>:TNT)]TJ /F7 7.97 Tf -1.59 -8.28 Td[(Bh9>=>;(2e)Settingthedownwellingskybrightnesstemperature,T)]TJ /F7 7.97 Tf -1.59 -8.27 Td[(B0,tozero,T+B0becomesTBsoil,pfromaspecularsoilsurface.Theroughsurfacebrightnesstemperaturecanbeexpressed: TBsoil,p=epT+B0 1)]TJ /F6 11.955 Tf 11.95 0 Td[(R0(2)where,epistheroughsurfaceemissivity,asmentionedpreviouslyinEquation 1 .First-orderorzero-orderapproximationstothepropagationEquation 2a areoftenusedinwhicheithersinglereectionsatinterfacesareconsideredorreectionsareignored,respectively.Therst-orderapproximationisgivenas[ 9 ]: TBsoil,p=epNXj=1Tj(1)]TJ /F10 11.955 Tf 15.11 8.09 Td[(1 Lj)(1+Rj Lj)jYi=2((1)]TJ /F6 11.955 Tf 11.95 0 Td[(Ri)]TJ /F5 7.97 Tf 6.59 0 Td[(1) Li)]TJ /F5 7.97 Tf 6.58 0 Td[(1)(2)Thezero-orderapproximationthatdoesnotaccountforanyreectionsbetweenthesoillayersisgivenas[ 85 ]: TBsoil,p=epNXj=1Tj(1)]TJ /F10 11.955 Tf 15.11 8.09 Td[(1 Lj)jYi=21 Li)]TJ /F5 7.97 Tf 6.59 0 Td[(1(2) 29

PAGE 30

AllofthesemodelsusecompletesoilmoistureandtemperatureprolesfortheTBsoil,pestimation,whensuchmeasurementsareavailable,andallowforuser-denedthicknessofsoillayers.Inthisdissertation,thebackwardpropagationsolutionanditstwoapproximationsarecomparedanddiscussedtounderstandtheapplicabilityoftheapproximationsinestimatingrealisticL-bandTBofthesandysoilsinChapter4. 2.4RoughSurfaceEmissivityAlthoughthesemi-empiricalapproachesforroughsurfaceemissivities[ 63 74 75 80 108 110 ]arewidelyused,suchasfortheESA-SMOSmission,themodelparametersintheseapproachesarehighlydependentupontheobservationsusedduringtheregressionprocess.Inthisdissertation,theAIEM[ 18 39 ]isusedtoestimatetheroughsurfaceemissivityinordertoavoidsuchuncertainties,andismoreapplicableforawiderangeofroughnessandfrequencies[ 18 ].TheemissivityinanIEM-basedmodelisdenedas: ep=1)]TJ /F10 11.955 Tf 11.96 0 Td[((rinp(i)+rcohp(i))(2)where,rinp(i)andrcohp(i)areincoherentandcoherentreectivities,respectively,attheincidenceangle,i,andcanbewrittenas: rinp(i)=1 4cosiZ20Z 20[0pp(i,i;s,s)+0pq(i,i;s,s)]sinsdsds(2a) rcohp(i)=Rp0(i)exp[)]TJ /F10 11.955 Tf 9.3 0 Td[((2kscosi)2](2b)where,0pqisthescatteringcoefcientintroducedinsection 2.1 ,(i,i)and(s,s)areelevationandazimuthanglesfortheincidenceandthescatteredradiation,respectively,Rp0istheFresnelreectivityofasmoothsurface,kisthewavenumber(2 ),andsistherootmeansquareheight.Notethatallthecurrentapproachesestimateepusingasinglesoilmoisturevaluewithinthemoisturesensingdepth(MSD),alsoreferredtoasthesurfacelayer,wherethesoilisassumedhomogeneous.TheMSDisdenedasthedepthatwhichtheemission 30

PAGE 31

is1/edescribedtop=1=s.Suchadenitiondoesnotaccountforthereectionsbetweensoillayers,sotheactualMSDmaybedifferentthanthe=1=s. 31

PAGE 32

Figure2-1. Dielectricconstantsimulationsat1.25GHzintermsofVSM.A)Realandimaginarypartsofdielectricconstantsat1.25GHzsimulatedfromPeplinskietal.(1995),Mironovetal.(2009),andmodiedmodels.B)EstimationofHH-andVV-pol0usingdielectricconstantsat1.25GHz. Figure2-2. Dielectricconstantsimulationsat1.25GHzintermsofsoilporosity.A)Realandimaginarypartsofdielectricconstantsat1.25GHzsimulatedfrommodiedmodelsusingVSM=0.1,0.2,0.3,and0.4m3=m3.B)EstimationofHH-andVV-pol0usingdielectricconstantsat1.25GHz. 32

PAGE 33

Figure2-3. Theillustrationsoftheincoherentsolutionoftheradiativetransferequation,rst-order,andzero-orderapproximationsforestimatingemissionfromalayeredsoilwithinnitelowerboundary. 33

PAGE 34

CHAPTER3MICROWAVE,WATERANDENERGYBALANCEEXPERIMENTSTheperformanceofsoilmoistureretrievalandassimilationalgorithmsusingAPmicrowaveobservationsrelyonaccurateestimatesof0andTB.Asmentionedearlier,thedevelopmentandvalidationofbackscatterandemissionmodelsandtheintegrativeuseoftheAPestimatesrequireconcurrenteldobservationsofAPsignatureswithhightemporalresolutiontocapturethehighlyvariablesoilmoisturedynamics.TheMicrowave,WaterandEnergyBalanceExperiments(MicroWEXs)areaseriesofseason-longexperimentsconductedinnorthcentralFlorida,(seeFigure 3-1 A),tomonitorthemicrowavesignaturesofsoilandvegetationduringdifferentstagesofgrowthwithtemporalresolutionofevery15minutes[ 7 13 14 53 57 ].TheexperimentalsitewassandywithsoilphysicalpropertiesobservedatthesitearegiveninTable 3-1 .ObservationsunderbaresoilconditionsduringthefthandeleventhMicroWEXs(MicroWEX-5and-11),respectively,wereusedinthisdissertation. 3.1FifthMicrowaveWaterandEnergyBalanceExperimentTheMicroWEX-5wasconductedduringagrowingseasonofsweetcornfromMarch9(DoY68)toMay26(DoY150),2006[ 14 ].Inthisdissertation,observationsfromMarch28(DoY87)toApril4(DoY94)wereused,duringwhichtheLeafAreaIndex(LAI)waslow,at<0.3,representingbaresoilconditions.Figure 3-1 BshowstheeldsiteandsensorlayoutduringMicroWEX-5.DuringMicroWEX-5,TBswereobservedat1.4and6.7GHz(=21.0and4.48cm),respectively,every15minutes,usingthetower-mountedUFL-bandandC-bandMicrowaveRadiometers(UFLMRandUFCMR),respectively,asshowninFigures 3-2 and 3-3 .ThedetailsofmicrowaveradiometersareprovidedinSection 3.1.1 .Volumetricsoilmoisture(VSM)andtemperatureprolesweremeasuredusingCampbellScienticTimeDomainReectometer(TDR)andthermistorprobes,respectively,atdepthsof2,4,8,16,32,64,and120cminthesoil,concurrentwiththemicrowaveobservations.Figures3-4AandBshow 34

PAGE 35

theobservedsoilmoistureandtemperatureprolesduringthebaresoilperiod.Twodrydownperiodsareconsideredinthisdissertation,withirrigationeventsonDoY88.03and90.83,providingtotalwaterinputof7.5and5.0mmduringthe30-minirrigation,respectively,recordedbytipping-bucketraingauges.Soilroughness,includingtherootmeansquareheight(s)andcorrelationlength(cl),weremeasuredneartheradiometerfootprintsusingthetraditionalgridboardmethod[ 50 ]andusingaground-basedLightDetectionAndRanging(LiDAR)[ 36 ]atthebeginningoftheexperiment,onDoY69.Thesandclrepresenttherandomsurfacecharacteristicandtheperiodicityofthesurface,respectively[ 101 ],andFigures 3-5 AandBillustratetherandomvariationsforaperiodicandatsurfaces,respectively.Thedetailofsurfaceroughnessmeasurementisdescribedinsection 3.1.3 .DetailsofotherobservationsthatwerenotusedinthisdissertationduringtheMicroWEX-5experimentareprovidedin[ 14 ]. 3.1.1UniversityofFloridaL-andC-bandMicrowaveRadiometersTheground-basedUFLMRandUFCMRweredesignedandbuiltbytheMicrowaveGeophysicsGroupattheUniversityofMichigan(UM-MGG),asshowninFigures 3-2 and 3-3 ,respectively.TheUFLMRconsistedofanH-polarizedtotalpowerradiometeroperatingattheof1.4GHz,matchingthefrequencyoftheESA-SMOSandtheupcomingNASA-SMAPmissions.Thesystemwashousedatopa10mtowerinstalledona16'trailerbed,andarotarysystemwasusedtorotatetheelevationangleoftheUFLMRforeldandskymeasurements.TheUFCMRconsistedofadualpolarizedtotalpowerradiometeroperatingatthecenterfrequencyof6.7GHz,similartotheJAXA-AMSR-2aboardtheGCOM-Wsatellite.Thesystemwashousedatopa10mtowerinstalledonatrailerbedwitharotarysystemforelevationanglechangesduringtheeldandskymeasurements.Table 3-2 listsspecicationsofthetworadiometersystemsandthenextdiscussesthetheoriesofoperationindetail.DuringtheMicroWEX-5,theUFLMRandtheUFCMR 35

PAGE 36

observedtheTBatheightsof6.81and5.90m,respectively,attheincidenceangleof50o.Basedupontheincidenceangleandthebeamwidthofradiometers,theobservedfootprintsofUFLMRandUFCMRwere4.99m6.95mand5.03m6.17m.Theradiometerswerecalibratedweeklywithamicrowaveabsorberandaninternalmatchedloadashotreferenceloadsandmeasurementsofskyasacoldload.ThecalibrationforboththeradiometersarediscussedindetailintheSection 3.1.2 .TheUFLMRandUFCMRweredesignedsimilarlyinmanyrespects,andtheblockdiagramsofUFLMRandUFCMRareshowninFigures 3-7 and 3-8 .Itisessentialtomaintainthermalcontrolinsidetheradiometertoensureminimaldriftingainsandprovidethemeasurementconsistency.Bothradiometersuseathermoelectriccooler(TEC)builtbyMcShaneInc.,forthermalcontroloftheRadioFrequency(RF)electronics.TheTECisusedtocoolorheattheelectronicsusingaProportionalIntegral-Derivative(PID)algorithmwithaprecisionof0.01oC.Twoelectromechanicallatches,SubMiniatureVersionA(SMA)connectors,drivenbytheZ-Worldcontrolboardareusedasswitchesineachradiometer.TherstSMAlatchswitchesbetweentheV-andH-polsequentially,andthesecondswitchesbetweentheinputsignalfromtheantennaandthematchedloadsignal.Thesignalsreachtheisolator,wherethecenterfrequenciesof1.4GHzand6.7GHzarepickedupbyabandpasslter(BF)inUFLMRandUFCMR,respectively.AfteramplifyingandlteringbyaseriesofLow-NoiseAmpliers(LNA)andBFs,thesignalsreachtoaSquareLawDetectorandPost-DetectionAmplier(PDA).BecausetheUFLMR'sseptumhornantennaissingle-polarized,onlyH-polsignalareguidedfromantennatocoaxtotheRFblock,andtheV-polinputtotheRFblockisanopencircuit.Bothradiometerswereequippedwithamicrocontrollerthathascapableofconductingmeasurements,monitoringandcontrollingthethermalenvironmentthroughTEC,andstoringdatauntiladownloadisrequested.AlaptopcomputerisusedforrunningtheuserinterfacestocommunicatewiththeUFLMRandUFCMR,respectively,throughRadiometerControlLanguage 36

PAGE 37

(RadiCL).TheoutputvoltagesfromtheradiometerareinVandcalibration,describedinSection 3.1.2 ,wasusedtoconvertthemeasurementsofVintoTBinK. 3.1.2RadiometerCalibrationRadiometersarehighlysensitivetothethermalelectromagneticemissionofsoil.Thesquare-lawdetectorsintheUFLMRandtheUFCMRprovidelinearitybetweentheantennaapparenttemperature(T'B)andtheoutputvoltage(Vout)[ 100 ],as: T'B=slVout+I(3)where,slandIaretheslopeandinterceptofthecalibrationcurve,respectively.TwocalibrationtargetsofknownemissivitiesatdifferentphysicaltemperaturescanbeusedtoestimateslandI.DuringMicroWEX-5,UFLMRandUFCMRwerecalibratedusingtheskyasacoldloadandamicrowaveabsorber,withemissivity=1,atambienttemperatureasthehotloadforexternalcalibration(EC)oraninternalreferenceloadofknownbrightnessasthehotloadforinternalcalibration(IC).Theexternal(EC)andinternal(IC)calibrationtechniquesin[ 91 ]wereusedtoobtaintheobservedTB. 3.1.2.1ExternalcalibrationTheECusedthemicrowaveabsorberatambientairtemperatureandtheskymeasurementatzenithangleof20oand40o,respectively,forC-andL-bandradiometers.TheslandI0arederivedas: sl=(TB,sky)]TJ /F6 11.955 Tf 11.95 0 Td[(Tabs)+(Tant,sky)]TJ /F6 11.955 Tf 11.96 0 Td[(Tant,abs)(1)]TJ /F4 11.955 Tf 11.95 0 Td[() Vout,sky)]TJ /F6 11.955 Tf 11.96 0 Td[(Vout,abs(3) I=TB,sky+Tant,sky(1)]TJ /F4 11.955 Tf 11.96 0 Td[())]TJ /F6 11.955 Tf 11.95 0 Td[(slVout,sky(3)where,TabsisthephysicaltemperatureofabsorberinKelvin,istheantennaefciency,equalto0.90and0.86forUFLMRandUFCMR,respectively,Tant,skyandTant,absarethephysicaltemperaturesofantennaduringtheskyandabsorbermeasurementsinKelvin,respectively,andVout,skyandVout,absaretheoutputvoltagesduringtheskyandabsorbermeasurementsinVolts,respectively.TheTB,skycanbeestimatedbydownwelling 37

PAGE 38

atmosphericradiation[ 100 ],givenas: TB,sky()=secZ10a(z0)T(z0)exp()]TJ /F4 11.955 Tf 9.3 0 Td[((0,z0)sec)dz0(3)where,isthezenithangleindegree,a(z0)istheatmosphericabsorptioncoefcientattheheightofz0inNpm)]TJ /F5 7.97 Tf 6.59 0 Td[(1,T(z0)isthephysicaltemperatureatheightz0inKelvin,and(0,z0)istheopticalthicknessoftheverticallayerbetweenthesurfaceandheightz0.Thea(z0)areprimarilyduetoabsorptionbyoxygenandwatervapor,and(0,z0)=Rz00a(z0)dz0.Giventheatmospherictemperatureandairpressure,a(z0)canbeestimatedassumingtheU.S.StandardAtmosphere[ 100 ]. 3.1.2.2InternalcalibrationICusesaninternalmatchedloadinsidetheradiometerasthehottarget.Thecoldtargetmeasurementswereobtainedbyskymeasurementsat20oand40oforUFCMRandUFLMR,respectively.TheslandIwerecalculatedusing: sl=TB,sky+Tant,sky(1)]TJ /F4 11.955 Tf 11.95 0 Td[())]TJ /F6 11.955 Tf 11.95 0 Td[(Tcal Vout,sky)]TJ /F6 11.955 Tf 11.96 0 Td[(Vout,cal(3) I=Tcal)]TJ /F6 11.955 Tf 11.95 0 Td[(slVout,cal(3)whereTcalisthephysicaltemperatureofthematchedloadinKelvin,andVout,calistheoutputvoltagewhenswitchedtothematchedloadatthephysicaltemperatureofTcalinvolts.TheTB,skyisestimatedasmentionedinSection 3.1.2.1 .TheFigures 3-9 A,B,andCshowthecalibratedTBofL-bandH-pol,C-bandH-pol,andC-bandV-pol,respectively,usingECandIC,duringMicroWEX-5.BecausetheICprovidedmoreconsistentcalibrationresultsunderdifferentenvironmentalconditions[ 91 ],onlyICwasusedinthisdissertation. 3.1.3SurfaceRoughnessMeasurementsA2m-longgridboardwasusedtomeasurethe2Dsurfaceprolesofthesoil,asshowninFigure 3-6 A,neartheL-andC-bandradiometerfootprints,respectively,perpendiculartotherowstructureintheeld.Thepicturesofthesurfaceproleswere 38

PAGE 39

digitizedusingaprogramdevelopedinMathWorkMATLAB,andtheheightofprole(zi)atahorizontalpositionxiwasmeasured.Themeansurfaceheightzwasacquiredusingz=1 NPNi=1zi.Theswasestimatedas[ 50 ]: s="1 N NXi=1(zi)2)]TJ /F6 11.955 Tf 11.96 0 Td[(N(z)2!#(3)where,Nistotalnumberofdigitizingpoints.Theclwasdeterminedusingasurfaceautocorrelationcurve.Foradiscretecase,(x0)isgiven: (x0)=PN+1)]TJ /F7 7.97 Tf 6.59 0 Td[(ji=1zizj+i+1 PNi=1z2i(3)where,x0isthehorizontaldisplacement,equalto(j)]TJ /F10 11.955 Tf 12.25 0 Td[(1)x,x=xi+1)]TJ /F6 11.955 Tf 12.25 0 Td[(xi,andjisaninteger1.When(x0)=1 e,thehorizontaldisplacementx0=cl.Inaddition,theMobileTerrestrialLaserMapping(M-TLS)system,integratedbyNationalCenterforAirborneLaserMapping(NCALM),UF(nowattheUniversityofHouston),wasusedtoobtainthe3Dpointcloudsofthesoilsurface,asshowninFigure 3-6 B.ThecoreoftheM-TLSisacommercial2-axisground-basedlaserscanner,ILRIS3D,producedbyOptechInc.,andwasinstalledonamobiletelescoping,rotating,andtiltingplatform,whichcanbeelevatedashighas10mabovethesurfaceandismountedonthebedofaheavy-duty4x4truck.DuringMicroWEX-5,theheightoftheLiDARsensorwassetapproximatelyat6m,withasamplespacingof2cm,toobtain3Dsurfacepointcloudstocoveranareaof18m2.These3Dpointcloudswererectiedtotareferenceplane,andtransformedintoadigitalelevationmodel(DEM).Theprolesalongthedirectionperpendiculartotherowstructurewereextracted,andtheheightandmeanheightoftheprolesweremeasuredforcalculatingsandcl[ 36 ].Table 3-3 liststhesandclnearL-andC-bandradiometerfootprints,respectively. 3.2EleventhMicrowaveWaterandEnergyBalanceExperimentMicroWEX-11wasconductedtoobtainconcurrentAPmicrowavesignaturesduringagrowingseasonofsweetcornandelephantgrassfromApril25(DoY116) 39

PAGE 40

toDecember6(DoY341)in2012.AbaresoilexperimentwasconductedduringMicroWEX-11fromMay9(DoY122)toJune8(DoY160),priortoplantingofcorn,andthebaresoilobservationsareusedinthisdissertation.DuringthebaresoilcomponentofMicroWEX-11,concurrentL-bandAPmicrowaveobservationswereconductedusingtheUFL-bandAutomatedRadarSystem(UF-LARS)andUFLMR,respectively.Theseobservationswereconductedevery15minutesonanunvegetatedagriculturaleldatanincidenceangleof40o.TheUF-LARSoperatedatthefrequencyof1.25GHz(=24.0cm)andobservedradarbackscatteratfourpolarizationcombinations.DetailsofelectromechanicalandcontrolsystemsinUF-LARSareprovidedinSection 3.2.1 .ConcurrentobservationsofsoilmoistureandtemperatureproleswereobtainedusingCampbellScienticTDRandthermistorprobes,respectively,atdepthsof2,4,8,32,and64cmconcurrenttothemicrowaveobservations.Fourraingaugeswereusedtorecordtheamountofwaterinputduringtheirrigation/precipitationevents.Figure 3-10 showsthesensorlayoutofthebaresoilcomponentduringMicroWEX-11,andsoilmoistureandtemperatureobservationsareshowninFigures 3-11 AandB.AtthebeginningoftheexperimentonDoY122,theeldwastilledanddiscedastypicallypreparedpriortoplanting.After23days,onDoY145,fakeplantingwasconductedwithoutseedstosimulatetheroughnessofatypicalagriculturaleldduringplanting.Soilroughnessmeasurements,sandcl,wereconductedusingtraditionalgridboardmethod,asdescribedinsection 3.1 ,duringthesmoothperiodonDoY132and138,andduringtheroughperiodonDoY146and163.Six2Dsurfaceproleseach,indirectionsparallelandperpendiculartotherowstructureweremeasuredusinga2m-longgridboard,asshowninFigure 3-12 .ThesurfaceprolefromeachgridboardmeasurementwasdigitizedandthesandclalongeachdirectionwereacquiredbyaveragingsixsandclcalculatedfromheightandmeanheightofsurfaceprolesusingEquations 3 and 3 .Table 3-4 liststhesandclmeasurementsparallelandperpendiculartotherowstructure. 40

PAGE 41

3.2.1UniversityofFloridaL-bandMicrowaveRadiometerTheUFLMRusedduringMicroWEX-11isthesameradiometerusedduringMicroWEX-5describedinSection 3.1.1 ,excepttheinclusionofacoldinternalloadforcalibration,ColdFet.TheV-polinputtotheRFblockwasmodiedtoincludetheColdFet,insteadofanopencircuitusedinMicroWEX-5.DuringthebaresoilcomponentoftheMicroWEX-11,UFLMRobservedtheTBsattheincidenceangleof40o,matchingtheangleoftheupcomingNASA-SMAPobservations.TheobservedfootprintofUFLMRwas5.02m5.51mfromthesensorheightof7.90m.Theradiometerwascalibratedweeklywithaninternalmatchedloadaswarmreferenceloadandmeasurementofskyatazenithangleof40oandcoldFetasthecoldloads.InadditiontotraditionalIC,mentionedinSection 3.1.2.2 ,thecalibrationincludingtheColdFetwasdeveloped.ThecalibrationschemeusingtheColdFetreliesuponweeklyskymeasurementstodeterminetheColdFetbrightness.AssumingtheColdFetbrightnesstemperature,TB,cf,doesnotchangebetweenthetwoskymeasurements,theTB,cfcanbeusedasacoldreferencealongwiththeinternalhotloadtocalibrateeach15minutemeasurement.TheTB,cfiscalibratedusingtheskymeasurementatthetimeoftc,givenas: TB,cf(tc)=TB,sky+Vout,cf)]TJ /F6 11.955 Tf 11.95 0 Td[(Vout,sky Vout,cal)]TJ /F6 11.955 Tf 11.95 0 Td[(Vout,sky(Tcal)]TJ /F6 11.955 Tf 11.95 0 Td[(TB,sky)(3)where,Vout,cfistheoutputvoltageoftheColdFetduringtheskymeasurementattimeoftc.TheTB,sky,Tcal,Vout,cal,andVout,skyhavebeendenedpreviouslyinSection 3.1.2.2 .BecausethesensitivityoftheColdFetbrightnesstophysicaltemperatureofmatchedloadispresumedtobeconsistentatthetimeoftargetmeasurements,tm,every15minutes,itimpliesthat: TB,cf(tc) Tcal(tc)=TB,cf(tm) Tcal(tm)(3) 41

PAGE 42

Thus,theTB,cfattimetmcanbederivedas: TB,cf(tm)=Tcal(tm) Tcal(tc)TB,cf(tc)+1)]TJ /F4 11.955 Tf 11.96 0 Td[( Tcal(tm) Tcal(tc)Tant,sky(tc))]TJ /F6 11.955 Tf 11.95 0 Td[(Tant,tar(tm)(3)where,Tant,taristhephysicaltemperaturesofantennaduringthetargetmeasurementsinKelvin.Finally,theTBofthetargetattimetmcanbeobtainedby: TB(tm)=TB,cf(tm)+Vout(tm))]TJ /F6 11.955 Tf 11.95 0 Td[(Vout,cf(tm) Vout,cal(tm))]TJ /F6 11.955 Tf 11.96 0 Td[(Vout,cf(tm)(Tcal(tm))]TJ /F6 11.955 Tf 11.95 0 Td[(TB,cf(tm))(3)Figure 3-13 .showsthecalibratedTBusingICwithandwithoutColdFet.ByincorporatingColdFetinthecalibration,thevariationsintheconstructiveanddestructiveinterferencesofbrightnesscanbeminimized.Inthisdissertation,theresultsfromICwithColdFetwereused. 3.2.2UniversityofFloridaL-bandAutomatedRadarSystemTheUF-LARSwasmountedona25mGeniemanlift,asshowninFigure 3-14 ,andconsistofthreemajorsubsystems,AntennaandRF(ARF),Electro-MechanicalPositioning(EMP),andSoftwareControlandDataAcquisition(SDA)subsystems,designedandbuildbytheUM-MGGandUF-CRS.Figure 3-15 showsblockdiagramofUF-LARS[ 65 ]andTable 3-5 listsspecicationsoftheradarsystem.ThesubsystemsofUF-LARSaredescribedindetailinSection 3.2.2.1 .DuringMicroWEX-11,theUF-LARSoperatedatthefrequencyof1.25GHztoobserveradarbackscattersatfourpolarizationcombinations,HH,VV,HV,andVH,andincidenceangleof40o,matchingthatoftheupcomingNASA-SMAP,andanazimuthrangeof0oto180o,withanazimuthalresolutionof9oevery15minutes,resultingin21samples,asshowninFigure 3-10 ,withanareaof1033.77m2fromsensorheightof16.2m,baseduponthe6dBbeamwidth.Theradarsystemwascalibratedweeklywithaaluminumtrihedralcorner-reectorasareferencetarget.ThecalibrationforUF-LARSisdiscussedindetailintheSection 3.2.2.3 42

PAGE 43

3.2.2.1SubsystemsofUF-LARSAntennaandRF(ARF)sub-system:TheARFsub-systemisbasedupontheestablisheddesignsforground-basedscatterometersemployingavectornetworkanalyzer(NWA),andisshowninblockdiagramforminFigure 3-16 .ThemagnitudeandphasemeasuringcapabilityofthevectorNWAenablesfull-polarimetricmeasurementswiththeUF-LARS.TheuseoftheAgilent8753seriesofNWA[ 48 ]permitsdirectaccesstotheNWA'sthreereceiverchannels:R,A,andB.TheRchannelisusedtocorrectfordriftsintheNWAandcablingbetweentheNWAplacedontheplatformoftheliftandtheradarelectronicspackagelocatedbehindtheantenna,asshowninFigure 3-17 .TheAandBchannelspermitthesimultaneousacquisitionofV-andH-polarizedreturnsfromthetarget.Asinglemulti-throwPINswitchselectsthetransmitpolarization,eitherVorH,oraninternalpathtobothreceive-channelsformonitoringgaindriftsintheamplier.Theantennaisacustom-designeddual-polarizationhornantennawithacustom-designed-turnstileorthmodetransducerbaseduponthedesignby[ 66 ].Thesecomponentsoperateatacenterfrequencyof1250MHz,withabandwidthof300MHzandtheone-way3dBbeamwidthof14.7ointheE-planeand19.7ointheH-plane.Thepolarizationisolationatthecenterfrequencywas>37dBforallprincipalplanesanddecreasestoabout23dBatthebandedgeof1400MHz.Electro-mechanicalpositioning(EMP)sub-system:TheEMPsub-systememploysacomputer[ 1 ]andiscomposedofthreedigitalinclinometers[ 103 ],anelevation-over-azimuthcontroller,linearactuators,andalaserrangender.OneinclinometerisinstalledontheantennaandtwoontheX-YandY-Zplanesoftheliftplatform,asshowninFigure 3-15 .Theinclinometersprovideincidenceanglesduringscans.Theelevation-over-azimuthcontroller[ 113 114 ]iscomprisedofanazimuthrotationunitinstalledontheGenieplatformtableandtheelevationcontrollermountedonmovablebracketsattachedtotheazimuthrotationunit.Thesystemoffersanangularmovementof360oinazimuthand180oinelevationtherebyprovidingthecapabilityto 43

PAGE 44

positiontheantennatowardsanydirection.Platformlevelingisenabledbycontrollingthreelinearactuators[ 104 ]installedonthethreecornersoftheplatform.Alaserrangender[ 59 ],capableofmeasuringdistancesupto50mataresolutionof1mm,isinstalledontheplatform.Thesystemiscompletelyautomatedenablingcontinuousdatacollection,dayandnight,inallweatherconditions,exceptduringexternalcalibrationmeasurements.Althoughthemeasurementsweremadeataconstantincidenceangle,thesystemiscapableofcontinuouselevationscans.Softwarecontrolanddataacquisition(SDA)sub-system:TheSDAsub-systemwasdevelopedinVisualC++andthebehaviorisregulatedthroughthestatetransitiondiagramthatcomprisedofvedistinctstates:INITIALIZE,CHECK,RECORD,POSITION,andIDLEasshowninFigure 3-18 .Thetransitionsbetweenthestatesareenabledbythreeags:Status,representingthestateofplatformleveling,Scans,thenumberofcumulativemeasurementsobtainedinthecurrentscan,andt,representingtheelapsedtimesincelastscan. 3.2.2.2UniversityofFloridaLARSsignalprocessingTheNetworkAnalyzer(NWA)providesmagnitudeandphaseofthepowerratiosforthefourpolarizationcombinations,VV,HH,HV,andVH,assteppedfrequencydomainmeasurements.Asignalprocessingandcalibrationprocedure,consistingstepsofInternalcalibration,Gatingprocessing,andExternalcalibration,asshowninFigure 3-19 ,isproposedinthissectiontoconvertthereceivedsignalinto0.Internalcalibration:TheinternalcalibrationwasconductedtoeliminatethegainsandlossesfrombothactiveandpassiveelementsintheNWA,whichalteredinmagnitudesandphasesofreturnmeasurements,bysuinganinternalcalloopequippedinUF-LARS.Thesignalreturn,jSptj2,byNWAis: jSptj2=LampLantPr Ptpt(3) 44

PAGE 45

where,LampisthegainsandlossesofcomponentswithinNWA,thecablingbetweenNWAandtheradar,andthepolarizationswitch,theLantincludeslossesfromthepolarizationswitchtotheapertureandbackfromtheaperturetotheinternalcalsignalcoupler,andPtandPrarethetransmittedandreceivedpowerofradar.Theinternalcalibrationmeasurement,jSicj2,coincidentwiththetargetmeasurementisgivenas: jSptj2ic=LampLic(3)where,LampissameasinEquation 3 ,andLicisthelossesassociatedwiththeinternalcalibrationloop.ThesignalreturnmeasurementisdividedbytheinternalcalibrationmeasurementtoeliminatetheeffectbyLampbeforethefurtherprocess.MptrepresentsthepowerratioincludingthegainsandlossesfromLantandLic,givenas: Mpt=Lant LicPr Ptpt(3)BecauseLantandLicdonotchangeovertime,theywillbeeliminatedbyexternalcalibrationusingcalibrationtarget,withknowngeometryandRCS.ThepowerratiomentionedinthefollowingprocessingrepresentsMpt.Gatingprocessing:Thepowerratiocorrespondingtobackscatterfromthetargetisextractedbygatingtheobservationswithrespecttotime,usingtherangetothetarget.Becausetherawdatainthefrequencydomainisband-limited,theovershootandringingeffectscouldoccurinthetimedomainresponseafterdatatransformation.Inordertomitigatetheseeffects,thefrequency-domaindataislteredwithaband-passlter.Suchlteringofthedatainfrequency-domainproducesatimedomainsignalwithlowersidelobeaftertransforming.Inthisstudy,theKaiserBesselwindowwasappliedasaband-passlter,givenas: !n=I0q 1)]TJ /F10 11.955 Tf 11.95 0 Td[((2(n)]TJ /F5 7.97 Tf 6.59 0 Td[(1) N)]TJ /F5 7.97 Tf 6.59 0 Td[(1)]TJ /F10 11.955 Tf 11.95 0 Td[(1)2 I0(),0nN)]TJ /F10 11.955 Tf 11.96 0 Td[(1(3) 45

PAGE 46

Where,I0isthezero-ordermodiedBasselfunctionofrstkind,isarealnumberthatdeterminestheshapeofthewindowandwassetto1.93,andnandNaretheorderandthetotallengthofdatasequence,respectively.ThelteredfrequencydomaindataweretransformedintotimedomainbyapplyingtheInverseChirp-ZTransform(ICZT)[ 3 38 ].TheICZTimprovesthesamplingresolutiononatarget'sprolebyusingparametersthatdenethestartingpositionofatarget'sproleandthesamplingresolution.TheChirp-Ztransform(CZT)isageneralizationoftheZ-transform,withadifferencethatCZTisnotrestrictedtotheunitcircleonthez-plane.TheCZTisdenedinEquation 3 ,whereA0isthestartingradius,0isthestartingangle,0istheangularstepsize,W0isaparameterthatdeterminesifthecontourspiralsinoroutonthez-plane,andNandMarethenumberofsamplesinthetimeandfrequencydomains,respectively.IfW0isgreaterthan1,thecontourspiralsin,andifitislessthan1,itspiralsout.ForW0=1,theradiusisconstant.TheCZTisequaltotheZ-transformwhenA0=W0=1,0=0,and0=2=N.Inthisstudy,wherebothA0andW0weresetto1,theICZTwasobtainedbytakingthecomplexconjugateoftheCZTofthecomplexconjugateofthefrequencydomaindata:x(n)=ICZT(X(k))=[CZT(X(k))]. X(z)=N)]TJ /F5 7.97 Tf 6.59 0 Td[(1Xi=0x(n)z)]TJ /F7 7.97 Tf 6.59 0 Td[(nz=AW)]TJ /F7 7.97 Tf 6.58 0 Td[(kW=W0ei20A=A0ei20 (3) AKasiserwindow,asEquation 3 ,with=0.16wasappliedtothetimedomaindatatoextractthebackscattererfromthesurfacetarget.Then,theextractedsignalwastransformedbacktothefrequencydomainbyapplyingCZTin( 3 ).Finally,thesignalinthefrequencydomainwasdividedbyfrequencylterusedbeforetorecoverthepowerscaleateachstepfrequency. 46

PAGE 47

3.2.2.3UniversityofFloridaLARSexternalcalibrationThegoaloftheexternalcalibrationprocessistoeliminatetheLantandLicmentionedinsection 3.2.2.2 andconvertthepowerratio,afterthegatingprocessing,intobackscatteringcoefcient,0.Measurementsof0forasurfacetargetwasestimatedbycalibratingthepowerratiosobtainedfromthesurfacetargetwiththoseobtainedfromacalibrationtargetwithknowngeometryandRCS,.Atrihedralcorner-reector,withanedgeoftheisoscelestriangularsections(l)of0.9m,wasusedforthecalibrationoftheradar,showninFigure 3-20 .Thereectorconsistedofthreesidesoftriangularpanels,withporestodrainrainwater.Thelongestsideofeachtriangularpanelwasreinforcedwithanangularbartominimizesurfacecurvature.Thepanelswereattachedperpendiculartoeachotherusingnutsandbolts.Duringcalibration,thetrihedralcorner-reectorwaslocatedonaatsurface,adjacenttotheeldsite.Threemeasurementsofcalibrationtargetwereconductedatthecal=45oandcal=40o,asshowninFigure 3-20 .Thecalwassetto40otomatchthatoftheupcomingNASA-SMAPmission.Thecorner-reectorwasremovedandthreemeasurementsofthebackgroundwereconductedtoremovetheinterferencefromground.ThedenitionofthedifferentialRCSfortransmitpolarizationqandreceivepolarizationp,i.e.backscatteringcoefcient(0pq)isgiveninEquation 1 .Itsapproximationdescribedbackscatterfromaneffectiveilluminatedareaisgivenas: 0pq= A=4jSpqj2 A(3)Where,istheRCSwithinilluminationareaofAfromanunknowndistributedtarget,andSisthescatteringamplitudeoftheunknowndistributedtarget.Inthisdissertation,aSingleTargetCalibrationTechnique(STCT)wasusedforradarcalibrationandthescatteringamplitudeforfourpolarizationsaredenedin[ 84 ]. suhh=1 (1)]TJ /F6 11.955 Tf 11.95 0 Td[(C2)2)]TJ /F10 11.955 Tf 9.3 0 Td[(2C2muvh m0vh+muhv m0hv+(1+C2)muhh m0hh+C2muvv m0vvs0ru r02e)]TJ /F5 7.97 Tf 6.58 0 Td[(2ik(r0)]TJ /F7 7.97 Tf 6.59 0 Td[(ru) 47

PAGE 48

suvv=1 (1)]TJ /F6 11.955 Tf 11.95 0 Td[(C2)2)]TJ /F10 11.955 Tf 9.3 0 Td[(2C2muvh m0vh+muhv m0hv+(1+C2)muvv m0vv+C2muhh m0hhs0ru r02e)]TJ /F5 7.97 Tf 6.58 0 Td[(2ik(r0)]TJ /F7 7.97 Tf 6.59 0 Td[(ru)suvh=C (1)]TJ /F6 11.955 Tf 11.95 0 Td[(C2)22muvh m0vh+2C2muhv m0hv)]TJ /F10 11.955 Tf 11.95 0 Td[((1+C2)muvv m0vv+muhh m0hhs0ru r02e)]TJ /F5 7.97 Tf 6.59 0 Td[(2ik(r0)]TJ /F7 7.97 Tf 6.58 0 Td[(ru)suhv=C (1)]TJ /F6 11.955 Tf 11.95 0 Td[(C2)22muhv m0hv+2C2muvh m0vh)]TJ /F10 11.955 Tf 11.95 0 Td[((1+C2)muvv m0vv+muhh m0hhs0ru r02e)]TJ /F5 7.97 Tf 6.59 0 Td[(2ik(r0)]TJ /F7 7.97 Tf 6.58 0 Td[(ru) (3) Thenormalizedbackscattersofco-polarized,0HHand0VV,andcross-polarized,0VHand0HV,canbederivedbyusingEquation 3 as: 0HH= A r4u1 j1)]TJ /F6 11.955 Tf 11.95 0 Td[(C2j4)]TJ /F10 11.955 Tf 9.3 0 Td[(2C2muvh m0vh+muhv m0hv+(1+C2)muhh m0hh+C2muvv m0vv21 r040VV= A r4u1 j1)]TJ /F6 11.955 Tf 11.95 0 Td[(C2j4)]TJ /F10 11.955 Tf 9.3 0 Td[(2C2muvh m0vh+muhv m0hv+(1+C2)muvv m0vv+C2muhh m0hh21 r040VH= A r4u1 j1)]TJ /F6 11.955 Tf 11.95 0 Td[(C2j42muvh m0vh+2C2muhv m0hv)]TJ /F10 11.955 Tf 11.96 0 Td[((1+C2)muvv m0vv+muhh m0hh21 r040HV= A r4u1 j1)]TJ /F6 11.955 Tf 11.95 0 Td[(C2j42muhv m0hv+2C2muvh m0vh)]TJ /F10 11.955 Tf 11.96 0 Td[((1+C2)muvv m0vv+muhh m0hh21 r04 (3) where,thesuperscriptsuand0representtheunknowntargetsurfaceandcalibrationtarget,respectively,ristherangebetweentheantennaandthemeasuredtarget,misthevoltageratioofradarmeasurementacquiredbytakingsquarerootofpowerratio,Cistheantennacross-talkfactordenedasC=1 a(1)]TJ /F11 11.955 Tf 12.1 10.39 Td[(p (1)]TJ /F6 11.955 Tf 11.96 0 Td[(a))anda=m0vhm0hv m0hhm0vv,andisthetheoreticalradarcrosssectionofthecalibrationtargetwithknowngeometry.Thewascalculatedusinggeometricalopticsmodel[ 26 ],given: (cal,cal)=8><>:4 2l44c1c2 c1+c2+c32c1+c2<=c34 2l4c1+c2+c3)]TJ /F5 7.97 Tf 29.16 4.71 Td[(2 c1+c2+c32c1+c2>c3(3)where: c1=cos(cal)c2=sin(cal)sin(cal)c3=sin(cal)cos(cal)(3) 48

PAGE 49

and,calandcalrepresenttheazimuthandzenithangleoftheradarimpingingdirectionasshownintheFigure 3-20 .Thecalandcalweresetto45oand40o,respectively,asthesettingforthecalibrationtargetmeasurements.Becausetheradargainpatternisnotuniformlydistributedinelevationandazimuth,normalizedradargainpatternisusedtoacorrecttheilluminationintegral.ThetermofA r40intheEquation 3 ,knownastheilluminationintegral(Iill),canbedescribedinitscontinuousanddiscreteformsas: Iill=Zg2 r4dtdA'Xg2i R4dtiAiA PAi(3)where,thegiisthenormalizedradargainpatterntothedirectionofinterest(i,i),Rdtiistherangebetweentheantennaandthetargetsurfaceassociatedtothedirection(i,i),Aiisthelocalilluminationareaoftheassociateddirection(i,i),andAistheactualilluminationareaoftheradarpulsewithinaeffectivebeamwidth.Inthisstudy,3DnormalizedradargainpatternwassimulatedfromsoftwareAnSoftHFSSwithangularresolutionof0.5oandwasvalidatedbythelaboratorymeasurementsattheE-andH-planatcentralfrequenciesof1.1,1.15,1.2,1.25,1.3,1.35,and1.4GHz,conductedinananechoicchamberattheUniversityofMichigan.Overall,therootmeansquaredifferences(RMSD)betweenthesimulatedandmeasuredgainpatterniswithin0.05dBforthe6dBbeamwidth.Inreality,theactualilluminationareaoftheradarpulsecannotbecalculatedanalytically,socomputationsofAiandAareusuallybasedonassumptionsofuniformdistributionwithindiscreteintervals.Inthisstudy,theregionilluminatedbyradarwasgriddedinto0.01m2areaasAiwithinwhichtheradarsignalwasassumeduniformlydistributed.Thus,theA PAiinEquation 3 isequalto1.Theassociatedgiwasacquiredbyprojectingthedirection(i,i)fromthecenterofthelocalilluminationareabacktotheantenna.Aconvergencetestwasconductedatantennaheightof16.2m,withincidenceangleof40o,sameassetupofUF-LARSduringMicroWEX-11,asshowninFigure 3-21 .TheIillincreaseswiththebeamwidth 49

PAGE 50

increasingandconvergesatabout6dBbeamwidthwithincrementwithin0.1dB.Thus,the6dBwastheoptimalvaluefortheeffectiveantennabeamwidth. 3.2.2.4FadingreductionAdistributedtarget,suchasthebaresoil,canbeconsideredasconsistingofrandompointscatterers.Interferenceofbackscattersfromthesepointscatterersatdifferentrangesresultsinfading.Statistically,thefadingsignalsareconsideredasaddnoise.Ifonlyonemeasurementofpowerratioisused,thefadingeffectcouldresultinapowerratiorangeof17.7dB,withastandarddeviationof5.57dB.Averagingindependentsamples,eitherinspaceorinfrequency,toobtainameanamplitudeofradarreturnisnecessarytoreducetheuncertaintyofaradarmeasurementcausedbyfading.Theseindependentsamplescanbeobtainedbyobservingdifferentilluminatedareasofthetarget,i.e.spatiallyindependentsamplesorobservingatdifferentfrequenciesclosetothecenterfrequency,i.e.spectrallyindependentsamples.Thecriteriafortheindependentsamplesinclude[ 98 ]: Spatiallyindependentsamples:thespacingbetweentwoadjacentfootprints,dasshowninFigure 3-22 A,mustbelargerthanthecorrelationlengthofthesurface. Spectrallyindependentsamples:f150 rMHz,wherefisthedifferencebetweentwoadjacentfrequencies,andtheristhedifferencebetweenthemaximumandminimumranges,r=rmax)]TJ /F6 11.955 Tf 12.22 0 Td[(rmin,asshowninFigure 3-22 B,fromtheantennatotheeffectiveilluminatedareaofthetarget.Inthisdissertation,theUF-LARSprovidedmeasurementsat9oincrementsinazimuthandspacingbetweentwoadjacentwas3.33mwhichisgreaterthanthemeasuredcorrelationlengthsduringthesmoothandroughperiods,aslistedintheTable 3-4 .Becauseradarsignaturesoverrow-structuredagriculturaleldsarethemostsensitivetoazimuthalviewinggeometry,theangleswithinabout15ooforthogonaltotherowdirectionprovideconsistentandthehighestbackscatter[ 24 ].Inthisdissertation,vemeasurementsobtainedwithintheazimuthalscanfrom-18oto+18owereused.Therwere9.54and6.99mforHHandVV,respectively,at6 50

PAGE 51

dBbeamwidth,associatedwiththeeffectivespacing15.73and21.46MHz.Ninefrequencysampleswith30MHzincrementswereselectedbetween1130to1370MHz.Inall,45sampleswereaveragedforeachbackscatteringmeasurementandthestandarddeviationoffadingwasdecreasedto0.662dB[ 46 ].Figure 3-23 showstheobserved0atthefourpolarizationcombinationsandassociatedvolumetricsoilmoisture(VSM)observedat2,4,8,16,32,and64cmduringthebaresoilexperimentofMicroWEX-11.Theamplitudes,S,atVHandHVareidentical,i.e.0VH=0HV[ 99 ],asalsoobservedinFigure 3-23 B,whereRMSDbetweentheobserved0VHand0HVis1.08dB.Infurtheranalyses,the0VHand0HVaremergedtogetherforthecross-polobservations,0cr. 3.2.3RelationshipBetweenActiveandPassiveMicrowaveObservationsConceptually,TBand0arenegativelycorrelated.TheseAPobservationsfromMicroWEX-11canbeusedtoprovideinformationregardingtherelationshipsbetweenTBand0forbaresoil,thataredirectrelevancetothecurrentbaselinesoilmoistureretrievalalgorithmusingcombinedAPobservationsforexampleinEntekhabietal.,2012[ 33 ]InFigures 3-24 A,theslope,0,changeswithsoilwetness.BothTBand0increasewithincreasingsoilmoisturewhenVSMislow,duetovolumescattering,resultinginpositiveslopeforHH-pol.However,duringtheroughperiodFigure 3-24 B,theslopesofbothwetanddryperiodsareverysimilar,withlessscatterinthedatacomparedtothesmoothperiod.Theseguresareforallthediurnalobservationsandarenotlimitedtothe6am/6pmoverpasstime.InFigure 3-25 A,nosignicantrelationshipisobservedbetweenthecross-polandco-polbackscatterduringthesmoothperiod,duetolowreturnsignalfromsmoothsurfaceparticularlyatcross-pol.InFigure 3-25 B,thecross-polandco-polbackscatterarehighlylinearduringroughperiod,withverysimilarslopesbetweenthetwopolarizationcombinations,exceptduringverylowvaluesofsoilmoistureintheroughperiod. 51

PAGE 52

Table3-1. MeasurementsofsoilpropertiesinPSREU SoilMeasurements PhysicalProperties Texture: Sand=89.4%byvol Clay=7.17%byvol Silt=3.5%byvol Bulkdensity:1.67g=cm3 Table3-2. SpecicationsoftheUFL-bandmicrowaveradiometer(UFLMR)andtheUFC-bandmicrowaveradiometer(UFCMR) ParameterQualierUFLMRUFCMR Frequency(GHz)Center1.46.7Bandwidth(MHz)3dB2020Beamwidth(deg)3dB22.523PolarizationSequentialHV/HNoiseFigure(dB)FromTrec3.993.99RFGain(dB)7985 Table3-3. Soilsurfaceroughnessmeasurementsofrootmeansquareheight(s)andcorrelationlength(cl)usinggridboardandground-basedLiDARonDoY69duringMicroWEX-5 GridBoardLiDAR s(cm)cl(cm)s(cm)cl(cm) L-band0.78.10.79.1C-band0.45.30.43.0 Table3-4. Soilsurfaceroughnessmeasurementsofrootmeansquareheight(s)andcorrelationlength(cl)usinggridboardduringbaresoilofMicroWEX-11 DoYParallelPerpendicular (2012)s(cm)cl(cm)s(cm)cl(cm) 1320.415.640.689.801380.4212.350.709.331460.451.781.719.451630.3612.871.3012.22 52

PAGE 53

Table3-5. ThespecicationsoftheUF-LARS ParameterQualierUF-LARS Frequency(GHz)Center1.4Bandwidth(GHz)0.3Beamwidth(deg)3dB14.7&19.7atE-&H-PlaneBeamwidth(deg)6dB21.4&28.1atE-&H-PlanePolarizationIsolation(dB)Center/Edge>37/23PolarizationHH,VV,VH,andVH 53

PAGE 54

Figure3-1. GeophysicallocationofeldandsensorslayoutofMicroWEX-5.A)Geophysicallocationofexperimentaleld.B)Thesensorslayout. 54

PAGE 55

Figure3-2. TheUFLMRsystem.A)ThefrontviewofwholeUFLMRsystem.B)ThefrontviewoftheUFLMRshowingthereceiverantenna.C)ThesideviewoftheUFLMRshowingtherotarysystem.PhotoscourtesyofJ.Casanova. 55

PAGE 56

Figure3-3. TheUFCMRsystem.A)ThefrontviewofwholeUFCMRsystem.B)ThefrontviewoftheUFCMRshowingthereceiverantenna.C)ThesideviewoftheUFCMRshowingtherotarysystem.PhotoscourtesyofJ.Casanova. 56

PAGE 57

Figure3-4. SoilmoistureandtemperatureobservationsduringMicroWEX-5.A)Volumetricsoilmoistureatdepthsof2,4,8,16,32,64,and120cm.B)Soiltemperatureatdepthsof2,4,8,16,32,64,and120cm. Figure3-5. Randomheightvariationonsurfaces.A)Aperiodicsurface.B)Aatsurface[ 101 ]. 57

PAGE 58

Figure3-6. Surfaceroughnessmeasurements.A)Using2-mlonggridboard.B)Usingground-basedLiDAR.PhotoscourtesyofJ.Casanova. 58

PAGE 59

Figure3-7. BlockdiagramoftheUFLMR[ 14 ]. 59

PAGE 60

Figure3-8. BlockdiagramoftheUFCMR[ 14 ]. 60

PAGE 61

Figure3-9. TBobservationsduringMicroWEX-5.A)CalibratedbyECandICinUFLMRatH-pol.B)CalibratedbyECandICinUFCMRatH-pol.C)CalibratedbyECandICinUFCMRatV-pol. 61

PAGE 62

Figure3-10. SensorslayoutofthebaresoilexperimentduringMicroWEX-11. 62

PAGE 63

Figure3-11. ObservedsoilmoistureandtemperatureduringbaresoilofMicroWEX-11.A)Volumetricsoilmoistureatdepthsof2,4,8,16,32,and64cm.B)Soiltemperatureatdepthsof2,4,8,16,32,and64cm. 63

PAGE 64

Figure3-12. Surfaceroughnessmeasurementsusinggrid-board.A)ParalleltoagriculturalrowstructureonDoY138.B)PerpendiculartoagriculturalrowstructureonDoY138.C)ParalleltoagriculturalrowstructureonDoY146.D)PerpendiculartoagriculturalrowstructureonDoY146.PhotoscourtesyofP.-W.Liu Figure3-13. TBcalibratedbyICwithoutandwithcoldFetinUFLMRatH-polduringbaresoilofMicroWEX-11. 64

PAGE 65

Figure3-14. TheUF-LARS.A)ThefrontviewofthewholeUF-LARS.B)ThefrontviewoftheUF-LARSshowingthereceiverantenna.C)ThesideviewoftheUF-LARSshowingtherotarysystem.PhotoscourtesyofD.Preston. Figure3-15. SubsystemswithintheUF-LARS.PhotocourtesyofD.Preston. 65

PAGE 66

Figure3-16. BlockdiagramoftheARFSub-system. 66

PAGE 67

Figure3-17. PostprocessingstepsinUF-LARS. Figure3-18. StatetransitiondiagramregulatingtheSDAsub-systemofUF-LARS. 67

PAGE 68

Figure3-19. Theowchartofstepstoestimate0HH,0VV,0VH,and0HV. Figure3-20. Denitionofgeometryoftrihedralcalibrationtarget.PhotocourtesyofK.Nagarajan. 68

PAGE 69

Figure3-21. Convergencetestfortheeffectiveantennabeamwidthusingnormalizedantennagainpattern.A)AtHH-pol.B)AtVV-pol.C)Atcross-pol. 69

PAGE 70

Figure3-22. Geometryofradarfootprints.A)Fronfrontview.B)Fromsideview. Figure3-23. ObservationsofbackscatteringcoefcientsduringbaresoilofMicroWEX-11.A)Atco-pol.B)Atcross-pol. 70

PAGE 71

Figure3-24. ScatterplotsofTBand0duringthebaresoilperiodinMicroWEX-11.A)Duringsmoothperiod.B)Duringroughperiod.The0and0aretheinterceptandslopes,similar(note:theyaremathematicallydifferent)totheandconceptinEntekhabietal.,2012. Figure3-25. Scatterplotsof0atcross-andco-polduringthebaresoilperiodinMicroWEX-11.A)Duringsmoothperiod.B)Duringroughperiod. 71

PAGE 72

CHAPTER4INTEGRATIONOFPASSIVEANDACTIVESIGNATURESFORMOISTUREDISTRIBUTIONINSANDYSOILSThischapterusesthebackscatteringandemissionmodelsdiscussedinChapter2alongwiththeMicroWEXobservationsdescribedinChapter3tounderstandtheimpactsofsoilconditionsonAPsignatures;theirsensitivitiestosoilmoisture;andtheircomplementarityinretrievingsoilcharacterization. 4.1SoilMoistureDistributionUsingPassiveObservationsTwodrydownperiodsduringMicroWEX-5,DoY90.5-94.5andDoY87.5-90.5,2006,withvaryingirrigationamounts,wereusedtoestimatesoilmoisturedistributionusingpassiveobservationsatC-andL-band.Duringtherstdrydownperiod5mmofwaterandduringthesecondperiod7.5mmofwaterwereusedforirrigation.Thedatasetconsistedof480observationsofH-andV-polTBatC-bandand375observationsofH-polTBatL-bandduringtherstdrydownperiod,andconsistedof281observationsofTBatC-bandand282observationsofTBatL-bandduringtheseconddrydownperiod. 4.1.1ComparisonofSoilEmissionModelsFigs. 4-1 AandBshowtheestimatedL-bandTBatH-andV-pol,respectively,usingtheincoherentsolutionfromEquations2-8and2-9,therst-orderfromEquation 2 ,andthezero-orderapproximationsfromEquation 2 .TheseestimateswereobtainedwithsoilmoistureandtemperatureobservedduringMicroWEX-5.Soilmoistureobservedat2cmwasusedtocomputeep.TheTBestimatesfromthethreeapproachesareverysimilar,withthemaximumdifferenceof<1.2Kamongthem.ThisindicatesthatthereectionsbetweenthesoillayersmaynotbesignicantinthesoilcolumnofthesandysoilsattheMicroWEX-5site.Inthisdissertation,weusetherst-orderapproximationtoaccountforthesinglereectionthatmaybecomesignicantattheinterfacesofthesandysoil,wherethemoistureinthetopsurfaceishighlydynamic. 72

PAGE 73

4.1.2ExtractingSoilParametersUsingC-bandSignaturesAsshowninFigure 1-1 anddiscussedinsection 1.1.4 ,thesoilmoistureobservedat2cmforepisnotadequateforrealisticTBestimationsatbothC-andL-band.Forexample,themodeloverestimatesH-polTBatL-bandbyabout28KduringtheirrigationeventsandunderestimatesTBbyabout30Kwhenthesoilisdry,2-3daysfollowingtheevents.Thisisbecausemostcurrentmodelsarecalibratedtoexpectednorms,thesealgorithmsaremostlikelytofailduringextremesofwetanddryhydrologicalconditions.Wetextremesincludethe0-24hoursofdrydownimmediatelyfollowingahydrologicevent,anddryextremesoccurwhennormallymoistnear-surfacesoilshavebecomedesiccatedoftheirfree-waterduringnear-droughtconditions.Typically,thetop-mostlayerofsandysoilsconsistsofloosesandparticles,withhighporositycomparedtothelowersoilmedium.TheC-bandradiometricobservationsarehighlysensitivetothepropertiesatthesurface,suchastheroughness,porosity,andwatercontentintheshallowersoil.Thesoilroughness,porosity,andmoisturewithintheMSDwerederivedfromthedualpolTBobservationsattheC-band,asdescribedinthesectionbelow.AsshowninFigure 4-2 ,thethicknessofthesoillayersinthetop2.5cmwassetto1mmtocapturethestrongdynamicsofthemoisture.Thesoilbelow2.5cmwasdividedinto1cm-thicklayersupto64.5cmandasemi-innitelayerbelow64.5cm.Thesoilmoistureobservationsatthedepthsof2,4,8,16,32,64,and120cmfromMicroWEX-5representedthevaluesforthemodeledlayersat1.9-2.1,3.5-4.5,7.5-8.5,15.5-16.5,31.5-32.5,63.5-64.5,andbelow64.5cm,respectively,asshowninFigure 4-2 .Soilmoisturevaluesintheotherlayerswereobtainedbylinearlyinterpolatingthevaluesbetweenthelayers.Thesoiltemperatureobservedat2cmduringMicroWEX-5wasusedasthetemperatureforthemodeledlayersfrom0to2.5cm.Thetemperaturesfordeeperlayers,>2.5cmwereassignedinasimilarmannertothesoilmoisturevalues. 73

PAGE 74

A.EstimationofsurfaceroughnessandsoilporositywithinMSDatC-band:TheinitialMSDofC-bandwassetto2mmtoestimatesoilpropertiesandthemoistureconsistentwithdual-polobservationsofC-bandduringthedrydownperiod.Thesensitivityofmicrowaveemissiontosurfaceroughnessincreasesasthesoilgetswetter[ 18 ].Therefore,theroughnessproperties,sandcl,wereestimatedduringthedriestperiodoftherstdrydown,aroundnoononApril3(DoY93.5),andwerefurtherrenedduringthewettestperiod,aroundmidnightonApril1(DoY91),toensurethattheestimatedvaluesprovidedrealisticTBestimatesduringtheentiredrydownperiod.Duringthedriestperiod,thesoilmoisturewithintheMSDwassetto0.01m3=m3becausethesurfaceofthesandysoilisextremelydryduetoinsolation.ThemeansofsandclobservationsusinggridboardandLiDARduringMicroWEX-5were0.55and6.38cmwithstandarddeviationsof0.17and2.77cm,respectively.Theinitialvaluesofsurfaceroughness,sandcl,andporosityweresetto0.55,6.38cm,and0.37,respectively,asobservedduringMicroWEX-5.Becausethemicrowaveemissionismoresensitivetormsheight,s,andsoilporositythanthecorrelationlength,cl,onlysandporositywereadjustedtoprovidethelowestrootmeansquaredifference(RMSD)betweentheobservedandthemodeledTBduringdriestperiod.Figures 4-3 AandBshowthatTBsatC-andL-bandweresimulatedwithsoilporositychangingatVSMof0.1,0.2,0.3,and0.4m3=m3andusingTeof290o.Thedifferencesbetweenusingsoilporosity0.1and0.6are19and11KatbothC-andL-bandH-polduringdryandwetsoil,respectively.ThesensitivityatV-polforbothC-andL-bandisrelativelylowerthanatH-polwithdifferences,betweenusingsoilporosity0.1and0.6,of15and11Kduringdryandwetsoil,respectively.Duringthewetperiod,thevalueofswasfurtheradjustedalongwithsoilmoisturevalue.Porositywasxedatitsvalueduringthedryperiod.B.EstimationofsoilmoisturewithintheMSDatC-bandduringthedrydownperiod:Thesebestestimatesofs,cl,andporosity,foundduringthedriestperiod,were 74

PAGE 75

implementedduringthewettestperiodofthedrydown,immediatelyaftertheirrigationevent.ThesoilmoistureandsvaluesthatprovidedtheTBestimatesclosesttotheobservedTBwerechosenduringthiswettestperiod,constrainedbytheporosityvaluesestimatedforthedrysoil.Thesoilmoisturewithinthemoisturesensingdepthwasestimatedduringthedrydownbydividingthedrydownintoseveralsmallintervals,withbreakingpointsbasedupontemporalchangesintheobservedTBatC-band.ThebestestimateofthesoilmoisturewithinMSDofC-bandwasdeterminedbycomparingthemodeledandtheobservedTBvaluesatbothV-andH-pol.Thesoilmoisturevaluesestimatedatthebreakingpointswerelinearlyinterpolatedtoobtaincontinuousvaluesduringtheentiredrydownperiod. 4.1.3EstimatingTBatL-bandSinglesoilmoisturevalue,mv f,withintheMSD,isusedtoestimateepforTBestimates,givenas: mv f=Z10mv(z)w(z)dz(4)where,frepresentsthefrequency,mv(z)isthesoilmoistureasafunctionofsoildepthz,andw(z)istheweightingfunction,withR10w(z)dz=1.Inthisstudy,Iexploreuniformandexponentiallydecayingweightingfunctions,givenas: wu(z)=8><>:1 zf0zzf0zf
PAGE 76

assumedatC-band,theMSDatL-bandis2.5cm.TheweightingfunctionscanbeappliedtothelayeredsoilmediumindiscreteformandtheEquation 4 becomes: mv L=PNj=1mvj N(4)usinguniformweightingfunction,and mv L=PNj=1mvjWej PNj=1wej(4)usingexponentiallydecayingfunction.Where,LrepresentsfrequencyatL-band,jisthejthlayerofsoil,NisthesoillayerattheMSDofL-band,Wejisthesoilmoistureinthejthlayerofsoil,andwejistheweightgiventojthlayerofsoil.ThemvLisusedtoestimateepfromIEMmodelandtheTBatL-band. 4.1.4ResultsandDiscussions 4.1.4.1EstimatingsurfaceparametersfromC-bandsignaturesA.Estimationofsurfaceroughnessandthesoilporosityinthetop0-2mm:IntheIEMmodel,theTBatH-polincreaseswithroughness,whilethatatV-poldecreases[ 18 ].Inaddition,theTBincreaseswithporosityanddecreaseswithincreasingsoilmoisture.Increasingthesurfaceporosityresultsinanincreaseddynamicrangeofsoilmoistureatthesurface.TheserelationshipsintheIEMmodelwereusedtodeterminethesandclthatwerewithinobservedrangesduringtheMicroWEX-5.Thesandclof0.53and6.38cmwerefoundtoprovidethebestTBestimationsduringthedriestperiod,fromDoY92.5-94.5,buttheTBwerestillunderestimatedbyupto13KatH-polandbyupto5KatV-pol,whencomparedwiththeobservedvalues.However,whensoilporositywasincreasedto0.55inthetop2mm,consistentwithtypicalloosetoplayerinsandysoils,thebestestimateofTBwasobtainedduringthedriestperiod.B.Estimationofmoistureinthetop0-2mmofsoilduringthedrydown:Theparameters,s=0.53cm,cl=6.38cm,andsoilporosity=0.55foundduringthedryperiod,wereappliedduringthewetperiod,fromDoY90.82to91.3,tondthewatercontentin 76

PAGE 77

thetop0-2mmofsoil.Thesoilmoisturevalueof0.3m3=m3inthetop2mmduringthewetperiodprovidedthebestTBestimations,eventhoughtheywereoverestimatedbyupto12KatH-pol,andthoseattheV-polwereunderestimatedbyupto10K.Afurtherdecreaseinsto0.41cmresultedinabetterTBestimateduringboththewetandthedryperiods,asshowninFigure 4-4 .TheestimatedvalueofsisclosetoonestandarddeviationoftheobservedsvaluesduringMicroWEX-5.Tenbreakingpointswereselectedduringthedrydownperiodtoobtainthewatercontentinthetop0-2mmofsoilthatresultedinthebestH-andV-polTBestimatesatC-bandatthosepoints.Figure 4-5 showsthemoisturevaluesfoundatthebreakingpoints,andthebestestimatedH-andV-polTBatC-bandwiththenewsoilmoisturevaluesinthetop0-2mm.TheRMSDoftheH-andV-polTBwere2.67Kand2.43K,andstandarddeviationswere2.63Kand1.70K,respectively,whencomparedwiththeTBobservedduringtheMicroWEX-5.InFigure 4-6 ,thesoilmoistureinthetop0-2mmiscomparedtothatobservedat2and4cmduringMicroWEX-5.Asexpected,thedynamicrangenearthesurfaceismuchlargerthantherangeatthedeeperlayers.Inaddition,theTBestimatesatL-bandusingtheIEMmodel,shownintheFigure 1-1 ,suggestthatawettersoilsurfaceduringthewetperiodsandadriersoilsurfaceduringthedryperiodswereobservedthanpredictedbymodels. 4.1.4.2ImpactsofsoilmoisturedistributiononpassivemicrowavesignaturesatL-bandThesoilpropertiesofs,cl,porosity,andsoilmoistureinthesurfacelayer,obtainedusingC-bandobservationsinthesection 4.1.4.1 wereusedtoestimateH-polTBatL-bandusingdifferentmoistureproleswithintheMSDatL-band,asshowninFigure 4-7 .ThesoilmoistureforthelayersbelowtheMSDatC-bandwasthesameasgiveninsection 4.1 .TheestimatesofTBwerecomparedtothoseobservedduringMicroWEX-5togaininsightsintotheimpactsofthedistributionofsoilmoistureonthemicrowavesignaturesandunderstandtheassumptionsnecessarytomatchtheobservations. 77

PAGE 78

InFigure 4-9 ,theobservedTBatL-bandduringMicroWEX-5iscomparedwiththeestimatesofTBusingdifferentmoistureproles.Therstmoistureproleconsistedofsoilmoisture0-2cmfromobservedvalueduringMicroWEX-5,asshowninFigure 4-8 A,withthenewlyestimatedsoilphysicalpropertiestounderstandtheeffectofonlyphysicalproperties.TheestimatedTB,shownasthegreendashedlineinFigure 4-9 A,showthatthedifferenceofoverallestimatesofTBiswithin2KtotheresultshownintheFigure 1-1 usingoriginalsoilphysicalproperties(seeTable 4-2 ).TheresultimpliesthattheH-polTBestimatesatL-bandremainlargelyunaffectedwhenonlysoilphysicalpropertiesarechanged,andthesoilmoistureobservedat2cmisinsufcienttorepresentthehighlydynamicofsoilmoisture.Thesecondmoistureprolewasobtainedbylinearlyinterpolatingthesoilmoisturevaluesin0-2mmestimatedfromC-band,andthoseobservedat2and4cmduringMicroWEX-5,asshowninFigure 4-8 B.Usingthisprole,theaveragesoilmoisturewithinthe0-2.5cm,i.e.12.6timestheMSDatC-band,wasusedtoestimateroughsurfaceemissivity.ThusthismethodweighseachlayerwithintheMSDequally,usingauniformweightingfunction.TheoverallTB,shownasbluedashedlineinFigure 4-9 A,matchtheobservationswellwithanRMSDof4.53KandSDof4.52K.ThisindicatesthattheoveralldynamicsTBduringthedrydowncouldbecapturedusingthelinearmoistureprolewithauniformweightingfunction.However,TBwasoverestimatedduringthewetperiod,withhigherRMSDof10.37K,asshowninTable 4-2 .BecausethesurfacesoilmoistureisconstrainedbytheC-bandobservations,theresultmayindicateaneedforusingnon-uniformweightingfunctionsuchasanexponentiallydecayingfunctionthatwouldincreasethecontributionfromthesurfacelayer.ItmayalsoindicatethattheMSDatC-bandneedtobeshallower.Inthenexttwoproles,wetestedthesehypotheses.Thethirdmoistureprolewasthesameassecondmoistureprole,butwiththesoilmoistureusedforepobtainedusinganexponentiallydecayingweightingfunction. 78

PAGE 79

Overall,theTBestimates,shownasthegraydashedlineinFigure 4-9 A,areveryclosetothoseobtainedusingtheuniformweightingfunction,withadifferenceofRMSDwithin1K.However,duringthewetperiod,immediatelyfollowingtheirrigationevent,theexponentiallydecayingweightingimprovestheRMSDofTBatL-bandby3K,asshownintheTable 4-2 .Duringthedryperiod,theTBusingtheexponentialfunctionhasahigherRMSDthantheTBsobtainedusingtheuniformweightingby1K,becausethedriersurfacelayercontributesmorethanthedeeperwetlayers.EventhoughusingtheexponentiallydecayingfunctionimprovedtheTBestimatesduringthewetperiod,theTBsatthetimeoftheeventandwithin12hourshaveRMSDsarestillhighatabout7.3K.ThisindicatesthattheMSDatC-andL-bandmaybeshallowerthan0-2mmand0-2.5cm,respectively.Thefourthmoistureproleusedashallowersurfacelayer,0-1mm,withitsmoistureestimatedfromtheC-band.The0-1mmsoilmoisturewaslinearlyinterpolatedwiththemoistureobservedat2cmduringMicroWEX-5,asshowninFigure 4-8 C.Theaveragesoilmoisturevalueusedforepobtainedusingtheexponentiallydecayingweightingfunctioninthetop1.3cm,theMSDatL-band.TheTBestimates,shownasthebrowndashedlineinFigure 4-9 A,areclosertothoseobservedwithin12hoursafterirrigationthanthethirdmoistureprole,withRMSDsreducedto4.11K,asshowninTable 4-2 .DecreasingtheMSDatC-bandto1mmandaveragingthesoilmoistureusingexponentiallydecayingweightingfunctioninthetop1.3cmforestimatingTBatL-bandproducedthebestTBestimatesduringthewetperiod.TherealisticestimatesofTBatL-bandabove,dependeduponfourassumptions:1)theMSDofC-bandis1mmand2mmduringthewetanddryperiods,respectively,withthesoilmoistureaslowas0.01m3=m3duringthedriestperiod;2)theMSDatL-bandis12.6timesthatatC-band,usingMironovetal.(2009);3)thesoilmoistureprolewithintheMSDatL-bandislinear;and4)theroughsurfaceemissivity(ep)atL-bandcanbeadequatelyobtainedusingeitherauniformoranexponentiallydecaying 79

PAGE 80

weightingfunctionforaneffectivesoilmoisturevaluewithintheMSD.However,theseassumptionsmaynotbevalidfordifferentsoiltypesorduringdifferenthydrologicalconditions.Forexample,thesameapproachwasappliedtotheseconddrydownperiod,fromDoY87.5-90.5,inwhichthesoilmoisturewithinMSDatC-bandof0-1mmforwetperiod,beforeDoY88.3,and0-2mmfordryperiodwasusedalongwiththeexponentialweightingfunctiontoestimateep.TheRMSDsof2.34and2.60KforH-andV-polTBatC-band,respectively,wereobtained.TheRMSDswerehigherfortheH-polTBatL-bandcomparedtothoseobtainedduringtherstdrydownperiod,seeFigure 4-9 Band 4-2 .BecausethemoisturedistributionwithintheMSDishighlydynamicduringthewetperiod,particularlyduringthe12hoursfollowinganirrigationevent,adynamichydrologicalmodelsuchasLSM,NOAH,andLSP,isneededtoprovidedetailedmoistureproleinthesoilcolumn,withaverticalresolutionsufcientforTBestimationatL-band.Escorihuelaetal.(2010)[ 35 ]usedahydrologymodeltoobtaindynamic0-2cmsoilmoistureandfoundthattheTBestimatesatL-bandmatchedtheobservations.However,asdiscussedearlier,theverticalresolutionof0-2cmwasfoundtobetoocoarseforthesandysoilsinthisstudy.Inaddition,boththeMSDandtheweightingfunctionsdependuponthesoilmoisture,andtheuseofonesoilmoisturevalue,assuminghomogeneoussoilsurface,toestimateemissivityisunrealistic,particularlyduringhydrologicalextremes.Currentmicrowavealgorithmsworkfairlywellduringlessdynamicstagesofthedrydownperiodsunderhydrologicequilibrium.However,theyprovideunrealisticTBestimatesduringhydrologicextremes.Itismoreimportanttoknowthehydrologicstorageanduxesduringtheseperiodsbecausetheresponseofthesystemismoresignicantthanduringtheequilibriumperiods.Animprovementofabout20-30KinTBisneeded,resultinginanimprovementof0.05-0.10m3=m3involumetricsoilmoistureduringandimmediatelyfollowingahydrologicaleventtoprovideaccurateestimatesofthestorageanduxesusingassimilationandretrievalalgorithmsfromcurrentspaceborneobservations.Inthisstudy,themethodsdevelopedforbare 80

PAGE 81

soilusingC-andL-bandobservationsprovideinsightsintotheerrorsourcesinthecurrentmicrowavealgorithmsduringthehighlydynamicperiods.TheobservationsfromsensorssuchasAMSR-EandASCATatC-bandaresensitivetomoisture,porosity,androughnessatthesurface,whilethosefromSMOSandSMAPatL-bandaresensitivetothesoilpropertiesinthelowerlayers.Thisstudyprovidestworecommendationstoimprovecurrentpassivemicrowavealgorithms.First,usingadynamichydrologicalmodeltoprovidedetailedmoistureprolewithintheMSDfortheC-andL-bandsensors,andsecond,developingabetterapproachtoestimateepthatutilizesthedetailedmoistureprolefromthehydrologicalmodel. 4.2SoilMoistureSensitivityofActiveandPassiveMicrowaveSignaturesSinceactiveobservationswereunavailableduringMicroWEX-5,concurrentAPmicrowaveobservationsduringMicroWEX-11wereusedtoevaluatethesensitivityof0andeptosoilmoisture.Inthisdissertation,threedrydownperiodswereselected.TwoperiodswereduringthesmoothperiodfromDoY130.5to134.5andDoY139.7to142.7,andoneduringtheroughperiodfromDoY148.8to157,whenfakeplantingwasconducted.Thesedrydownperiodsprovidedagoodrangeofsoilmoisturevaluestounderstandthesensitivityofbackscatterobservations.Table 4-3 providesinformationregardingwaterinputduringthesedrydownperiods,andPeriod3gainedthemostwaterduetoatropicalstorm.Correlationcoefcient()anddifferentialanalyseswereusedtoquantifythesensitivitybetweentwovariables[ 45 ].Thecanbeusedtoobtaintherelationshipbetweentwofunctionsquantitatively,aswellastoanalyzetherelativelysensitivitybyrankingcoefcients.Here,valuesbetweentheobserved0HH,0VV,0cr,andepandVSMatdepthsof2,4,8,16,32,and64cmwereobtained,respectively.Themostsensitivelayerofsoilmoisturewasdetermined.While,thedifferentialanalysis,alsoreferredtothedirectmethod,computestheratioofthechangeinonevariabletothechangeoftheotherwhichprovidesacomparableresult.Forthedifferentialanalysis,Equations 4a and 4b areusedtocomputethesensitivityof0andepto 81

PAGE 82

soilmoistureobservedatthemostsensitivelayer,respectively. Spq=@0pq @V(4a) Sp=@ep @V(4b)where,pandqrepresentthepolarizationsofthereceivedandtransmittedsignal,0isthebackscatteringcoefcientindB,andVistheVSMin%byvol. 4.2.1Sensitivityof0toSoilMoistureTable 4-4 liststhe-valuesofsoilmoistureobservedat2,4,8,16,32,and64cmto0HH,0VV,and0crduringthethreedrydownperiods.Asexpected,the0atco-polaremoresensitivetoVSMat2cm,accordingtothevalues,thantotheVSMatdeeperlayers,duringthreedrydownperiods.Particularly,the0cristoolowforthesmoothsurfacetoexhibitthedependencyonsoilmoisture.Duringtheroughperiod,the0crvaluesalsoshowhighersensitivitytoVSMat2cm,similartoco-polobservations,whichindicatesthatthechangeofobserved0isbettercorrelatedtothewaterdistributedintheroughterrainthaninthesmoothterrain.Theseresultssuggestedthattheactivemicrowavesystemissensitivetothenearsurfacesoilmoistureatabout2cmwithsurfaceroughness.Thescatterplotsofobserved0HH,0VV,and0crtotheVSMat2cmareshownintheFigures 4-10 A-C.0HHand0VVdecreasewithVSMfordryconditioninthesmoothperiodwhenVSM<5%byvol,duetothevolumescattering.Suchphenomenonoccursatlargerwavelengthforsmooth,drysoils[ 55 89 ].AthigherVSMvalues,the0HHand0VVincreasewithVSM,assurfacescatteringbecomesincreasinglydominant.Thisindicatesatransitionfromvolumescatteringasadominantprocesstosurfacescatteringasthesoilmoistureincreasesduringsmoothsoil.The0crduringsmoothperiodexhibitaconstantorveryslightpositivetrendwithrespecttoVSM,primarilyduetothelowreturnsignal. 82

PAGE 83

Becausethesurfaceroughnessisstableduringsmoothperiod,thesimilarrelationshipbetween0andVSMareobservedduringthetwodrydownperiods.Thus,0HH,0VV,0crobservationsduringthetwodrydownperiodswithsmoothsurfacearegroupedtogetherforlinearregressionusingEquation 4a .Theco-andcross-pol0valuesduringtheroughperiodfollowapowercurvewithrespecttoVSM,givenasEquation 4b : 0pq=apqV+bpq(4a) 0pq=cpqVdpq(4b)Where,VistheVSMvaluein%byvol,aandbarecoefcientsfromthelinearregression,andcanddarecoefcientsfromthepowerregression.Table 4-5 liststheregressionresultsoflinearandpowerfunctions,includingaandbforsmoothperiod,canddfortheroughperiod,andR2forallthreeperiods.Overall,theroughperiodprovidesthemostreliableestimationsatco-andcross-polaccordingtotheR2values.ThelowerR2valuesduringthedry,smoothperiodindicatethatthesoilmoistureisnotadominantfactoraffectingvolumescattering.Thesensitivityof0isobtainedasinEquation 4a .Figure 4-10 DshowsthesensitivityintermsofVSM.Duringthewetandsmoothperiod,the0VVismoresensitivetoVSMthan0HHby0.2dB/%byvol.Acontrastconditionwasobservedduringtheroughperiod,duringwhenthesensitivityvalueof0VVishigherby1to0.1dB/%byvolthanof0HHand0crassoilwasdry,while0HH,0VV,ands0crarealmostequallysensitivetoVSMwiththesensitivityvaluesconvergingtoabout0.45dB/%byvolassoilgotwet.Thisisprimarilybecausethebackscattersatthesepolarizationsapproachedtotheirsaturatedpointsunderthiscondition.Thesensitivityvaluesofco-polarecomparedbetweenthesmoothandroughperiodsintheVSMrangeof5-8%byvol,theeffectiveVSMrangefortheobservationfromsmoothsurface.Thesensitivityvalueissignicanthigherbyupto2.0dB/%byvolduringthesmoothperiod,becausethesensitivitywasmaskedbythesurfaceroughness.Althoughhighersensitivityisobservedinthe 83

PAGE 84

smoothsurfacesoil,theseresultssuggestthattheactivemicrowaveobservationsfromroughsurfacesoilaremoreapplicableforthesoilmoistureestimation,particularlywithoutvolumescattering.Theseresultssuggestthatproperroughnessisnecessarytoproduceadequatebackscatterwhensoilisdry,andthesensitivityof0totheVSMdecreasesontheroughsoilwhengettingwet. 4.2.2SensitivityofeptoSoilMoistureCorrelationcoefcients,,betweenepandobservedVSMatdifferentlayerswerecalculatedusingsimilarmethodologyasinsection 4.2.1 tondthemostsensitivelayeramongtheobservationsandareshowninTable 4-6 .TheepwasestimatedasobservedTBdividedbyTe,whereTeistheeffectivetemperatureestimatedbytherstorderapproximationinEquation 2 usingobservedsoilmoistureandtemperatureproles.Asexpected,theepatH-polismoresensitivetoVSMat2cm,accordingtothevalues,thantotheVSMatdeeperlayers,duringthreedrydownperiods.ThevaluesexhibitalmostequallysensitivetotheVSMduringthreedrydownperiods.Asexpected,thescatterplot,inFigure 4-11 A,showsthattheepdecreasesmonotonicallywithVSMincreasing.Novolumescatteringisobserved.Theobservationsduringtwodrydownperiodswithsmoothsurfacearegroupedtogetherforthedifferentialanalysisbecausetheyexhibitsimilarlyinthescatterplot.WhenVSM>6%,theseparationbetweenepsduringsmoothandroughperiodsbecomessignicantly,becausethesensitivityofmicrowaveemissiontosurfaceroughnessincreasesasthesoilgetswetter.Thedifferentialanalysisisconductedassection 4.2 forep.Thesecond-orderpolynomialfunctionareusedfordataregressionduringsmoothandroughperiods,givenas: ep=fV2+gV+h(4) 84

PAGE 85

where,f,g,andhareequationcoefcients,VistheVSMvaluein%byvol.Table 4-7 liststhecoefcientsandR2ofregression.BothsmoothandroughperiodsprovidereliableestimationsaccordingtotheR2values.ThesensitivityofepisobtainedasinEquation 4b ,andisshowninFigure 4-11 B.Overall,theepduringsmoothperiodismoresensitivetoVSMthanthatduringroughperiod.WhenVSM<6%,thedifferenceofsensitivitybetweensmoothandroughperiodiswithin0.051/%byvol,andthentheyseparatessignicantlybecausethemicrowaveemissionismoresensitivetosurfaceroughnessduringwetsoil.Thesensitivityconvergesto-0.011/%byvolwhenVSM>10%byvolduringtheroughperiod.ThevaluesareusedtocomparethesensitivityofAPsignaturestoVSM.Overall,thesensitivityofeptoVSMishigherthanthatof0givenbyhighercorrelationcoefcient,particularlyduringthesmoothperiod.However,epislargelyinsensitivetoroughness,particularlyduringthedryperiod,asshowninFigure 4-11 A,aswellasinFigure 3-23 ,wheresimilarTBvalueswereobservedattransitionsbetweensmoothandroughperiods,incontrastto0,asshowninFigures 4-10 (a-c)and 3-13 .ThisstudysuggestsapotentialincorporationofAPmicrowavesignaturesforimprovementofsoilmoistureandsurfaceroughnessestimates. 4.3IntegrationofActiveandPassiveMicrowaveSignaturesBasedupontheresultsinsection 4.2 ,theactivemicrowaveobservationswereusedforsurfaceroughnessestimatesduetotheirhighersensitivitytosurfaceroughness.Thesesurfaceroughnessestimationswereusedintheemissionmodeltoretrievesoilmoistureinthenear-surfaceusingpassiveobservations.Inaddition,theabilityofthecurrentemissionandbackscattermodelswasinvestigatedtoprovideaphysicallyrealisticmoisturedistributionconsistentwithbothAPobservations.Thesimulationswereconductedfordrydownperiods1and3forthesmoothandroughsurfaces,respectively. 85

PAGE 86

4.3.1ActiveandPassiveMicrowaveSimulationsAsmentionedinthesection 4.1 ,thetop-mostlayerofsandysoilsconsistsofloosesandparticleswithhighporosity.Similarly,thethicknessofsoillayersinthetop2.5cmwas1mmtocapturethestrongdynamicsofthemoisture.Soilbelow2.5cmwasdividedinto1cmlayersupto63.5cm,andmodeledasasemi-innitelayerbelow63.5cm.SMobservationsatthedepthsof2,4,8,16,32,and64cmfromMicroWEX-11representedthevaluesforthemodeledlayersat1.9-2.1,3.5-4.5,7.5-8.5,15.5-16.5,31.5-32.5,andbelow63.5cm,respectively.SMvaluesintheotherlayerswereobtainedbylinearlyinterpolatingvaluesbetweenlayers.SoilsurfacetemperatureobservedbyTIRwasusedassoiltemperatureat0-2mm.TemperaturesforotherlayerswereassignedinasimilarmannertotheSMvalues.Insection 4.1.4.2 ,ithasbeenfoundthattheVSMat2cmisnotadequatetoobtainrealisticTBestimatesatL-bandusingcurrentforwardmodels,andmoisturedistributioninthenear-surfaceisneeded,particularlyduringdynamicconditions.ToachievetherealisticestimatesofTBand0consistently,thecomplementarityofAPsignatureswereused.A.EstimationofVSMinthetop0-2mmusingthepassiveobservations:Thedrydownperiodsweredividedintosmallintervals,withbreakingpointsbaseduponhydrologicaleventsandtemporalchangesinobservedTB.Sixandsixteenbreakingpointswerefoundduringsmoothandroughperiods,respectively.ThemeansofsandclmeasuredduringMicroWEX-11wereusedasinitialvalues.Forthesmoothperiod,theinitialsandclwere0.68and9.8cm,respectively,whiletheinitialvaluesduringroughperiodwere1.71and9.45cm.Thesoilporosityinthetopsoilwassetto0.55,asestimatedinsection 4.1.4.1 .AbestestimateofVSMin0-2mmwasobtainedatthebreakingpointsthatprovidedtheclosestmatchtotheobservedTBafterusinganexponentiallydecayingweightingfunctionasinEquation 4 tointegratesoilmoisture 86

PAGE 87

valueswithinthetop0-2.5cmforIEM.The0-2mmSMvaluesbetweenthebreakingpointswerelinearlyinterpolated.B.Estimationofsurfaceroughnessusingtheactiveobservations:Thesoilmoistureprolesestimatedbyusingpassiveobservationswereappliedforthebackscattermodel.TheeffectiveVSMforbackscattermodelwasestimatedbyanexponentiallydecayingweightingfunctiontointegratesoilmoisturevalueswithinthetop0-2.5cm.Roughnessparameters,sandclweremodiedbyobtainingmodeled0thatprovidetheclosestmatchtoobserved0.StepsAandBwereconductediterativelytondtheVSMandroughnessparametersthatprovidetheclosestmatchtobothTBand0duringdrydownperiods. 4.3.2ResultsandDiscussionThebestestimateofVSMin0-2mmandnear-surfacesoilmoistureproleduringthesmoothperiodareshowninFigures 4-12 AandB.Asexpected,VSMnearthesoilsurfaceismoredynamicthanatdeeperlayers.Themodiedvaluesofsandclareestimatedat0.68and9.79cm.Theprole3mentionedinsection 4.1.3 wasusedtoacquirethebestestimatesofTBatH-poland0atHHandVV-polduringthesmoothperiod,andresultedintheRMSDsof9.86K,and4.24and1.88dBforTBatH-poland0atHHandVV-pol,respectively,asshowninFigure 4-13 andTable 4-8 .Overall,usingdetailedsoilmoistureprole(Prole3)providedbetterRMSDsonTBatH-poland0atHH-andVV-pol,respectively,thanusingprole1,duringthewholedrydown,wet(DoY130.9-131.5),anddryperiods(131.5-134.5),asshowninTable 4-8 .However,thevolumescatteringduringdryperiodcontributeshigherRMSDsby6.28K,and3.4and0.9dBthanthoseduringwetperiod.Thissuggeststheneedforavolumescatteringmodel,particularlyfor0atHH-pol,duringdryandsmoothsoils.Thesoilmoistureinthenear-surfaceandroughnesswereestimatedusingthesamemethodologyduringtheroughperiod.ThebestestimateVSMin0-2mmandnear-surfacesoilmoistureproleduringtheroughsmoothinFigures 4-14 AandB. 87

PAGE 88

Similartothesmoothperiod,thesoilmoisturenearsurfaceismoredynamicthanthoseobservedatdeeperlayers.Thesandclareestimatedas2.08and6.58cmusingactivesignaturesduringtheroughperiod.TheseestimationsofTBatH-poland0atHH-andVV-polusingprole3resultinlowerRMSDsby21.2K,2.7,and2.6dB,respectively,thanthoseestimatedusingprole1,respectively,asshowninFigure 4-15 andTable 4-8 .Therefore,itisnoteworthythataphysicallyrealistic,plausiblesoilmoisturedistributionandroughnessparametervalueswerepossibleusingconcurrentAPobservationsathightemporalresolutionandcurrentbackscatterandemissionmodels.ThewaterinputmeasuredbyraingaugewascomparedwiththewatergainsestimatedfrommoistureproleusingsoilmoisturemeasurementswiththeTDRduringMicroWEX-11andmoistureproleobtainedusingmicrowaveobservations,i.e.Prole3.Thiscomparisonprovidedfurtherinsightsintotheplausibilityofthemicrowave-derivedsoilmoistureprole.Twoprecipitationevents,onDoY148.95andDoY157.17,andoneirrigationeventonDoY130.89wereusedforthecomparisons.Itwasassumedthatthewaterapplicationduringtheeventsandthewaterstoredineachhomogeneouslayerofsoilarebothuniform.ThewatergainbysoilduringtheeventwasobtainedbyEquation 4 : Wg=Wa A=XVSMdi idi(4)where,Wgisthedepthofwatergainduringeventinmillimeter,Waisthetotalamountofwatergaininm3withinthefootprint,Aistheareaoffootprintinm2,VSMdi iisthedifferenceofVSMatithlayerimmediatelypriorandafterevent,anddiisthethicknessofithlayerofsoilinmeter.Figure4-2showsthelayerdiscretizationforthemoistureprolesfromtheTDRandthemicrowaveobservations.ForsoillayersofAP-derivedproleconsistedof1mmthicklayersinthetop2cmandthickerlayersbelow2cm,asshowningure4-2.ForthemoistureproleobtainedfromtheTDR,theobservationsat 88

PAGE 89

2cmwereusedtorepresenttheVSMinthetop2cmofsoil,andthelayersbelow2cmweresetasAP-derivedmoistureprole.Duringtherstprecipitationevent,onDoY148.95,itrainedfor1.25hoursandprovided5.6mmofwater,basedupontheraingaugeobservations.ThewatergaininsoilestimatedbytheTDRmeasurementswas2.6mm,suggestingadecitof3mm.TheestimatedwatergaininsoilusingProle3obtainedfromtheAPobservationswas6.2mm,suggestingaslightoverestimationof0.6mm.Thestandarddeviationoferrorsintheraingaugemeasurementsistypically12%[ 14 ],i.e.0.67mm,indicatingtheplausibilityofProle3.Thesecondprecipitationevent,onDoY157.17lastedfor3.5hours,with4.8mmofwater,asmeasuredbyraingauge.TheestimatedwatergaininsoilusingtheTDRmeasurementswas2.8mm,suggestingadecitof2mmwater.TheestimatedwatergaininsoilusingderivedmoistureprolebyAPobservationswas5.2mm,thatwaswithintheerrorrangeofraingauge.ThisalsoindicatedtheplausibilityofProle3duringthesecondevent.TheirrigationeventonDoY130.89thatlastedfor30minutes,with3.8mmofwatermeasuredbyraingauge.TheestimatedwatergainusingTDRobservationwas1.1mm,suggestingadecitof2.7mmofwater.WhiletheestimatedwatergaininthesoilusingProle3wasoverestimatedby2.4mm.Thisisgreaterthantheerrorofraingaugemeasurementsandseemsunrealistic.Thisoverestimationmaybeduetothestandingwaterontopofthesoilresultedfromhighintensityoftheirrigationevent.AsshownintheFigures 4-13 A,itisobservedthattheTBdecreasedimmediatelyduringtheirrigationevent,andincreasedatafasterrateforthenexttwohours,indicatingadrydownduetoeitherevaporationorinltrationofstandingwater.However,duringtheprecipitationevents,asshownin 4-15 A,theTBdecreasedduringeventsandremainedconstantforcouplehours.Nosignicantstandingwaterwasobserved. 89

PAGE 90

Theseresultsconrmthatthesoilmoistureinthenear-surfacemeasuredusingTDRisnotadequatetopresenttherealisticsoilmoistureimmediatelyafterthehydrologicevents.ItisalsosuggeststhatderivedsoilmoistureprolesusingAPobservationsarephysicallyrealistic,particularlyduringtheprecipitationevents. 4.4SummaryCurrentpassivemicrowavealgorithmsuseasinglemoisturevaluetorepresenttheeffectivemoistureinthemoisturesensingdepththatresultsinunrealisticTBestimatesforthebareandvegetatedsoilatL-band,particularlyduringextremehydrologicconditions.C-bandradiometricobservationswereusedtoestimatephysicallyconsistentsurfaceparametersofsandysoilssuchass,porosity,andmoisturewithintheMSDatC-band,withthesoilmoistureof0.01m3=m3duringthedriestperiod.AnexponentiallydecayingweightingfunctionwasusedtoobtaineffectivesoilmoisturevalueswithintheMSDatL-bandfortheroughsurfaceemissivity(ep)estimatesofH-polTBatL-band.ReasonableestimatesofTBcanbeobtainedduringhighlydynamicperiodsinthesandysoils.TheresultsindicatethattheimpactofmoisturedistributionwithintheMSDontheTBatL-bandinsandysoilsismoresignicantthanthatofphysicalproperties,suchass,cl,andporosity.Similarly,currentphysically-based,AIEM,resultsinunrealisticestimatesof0atHH-andVV-polusingonlysoilmoistureobservedat2cmduringMicroWEX-11.ThischapterprovidesamethodologytointegratetheAPmicrowavesignaturetoimprovethesoilmoistureestimatesaswellasthesurfaceroughness,sandcl.ThecomplementarityofAPmicrowaveobservationswasinvestigatedbyanalyzingtheirsensitivitiestosoilmoistureandsurfaceroughnessusingunprecedentedhightemporalresolutionobservations,duringdynamicmoistureconditionsinsandy,agriculturalelds.Overall,passiveobservationsweremoresensitivetosoilmoistureandlargelyinsensitivetothesurfaceroughness,incontrastwithactiveobservations.Theimprovedestimatesofsoilmoisturedistributionwithinthenear-surface,aswellasthesurfaceroughness 90

PAGE 91

parameters,arefeasiblebyintegratingactivemicrowaveobservations.Theseestimatedsoilmoistureproles,sandclprovidedconsistentAPestimationsbyusingcurrentemissionandbackscattermodels,particularlyduringroughsurface.ComparisonofmeasuredwaterinputandestimatedwatergainwereconductedandresultsindicatedthattheestimationofsoilmoistureproleusingAPobservationsisphysicallyrealistic,particularlyduringprecipitationevents. 91

PAGE 92

Table4-1. Thesoilpenetrationdepth(p)ofC-andL-bandandthesensingdepthratio SoilmoisturepofC-band(cm)pofL-band(cm)Sensingdepthratio 0.0121.22267.3512.600.054.2453.4712.610.12.1226.7412.610.21.0613.3712.610.30.718.9112.55 92

PAGE 93

Table4-2. RootMeanSquareDifferences(RMSDs)andStandardDeviations(SD)betweentheobservedandmodeledH-polTBatL-bandusingdifferentmoistureprolesandsoilpropertiesofrootmeansquareheight(s),correlationlength(cl)andporosityduringtwodrydownperiods FordrydownduringDoY90.5-94.5;wetperiodDoY90.8-91.3;anddryperiodDoY92.5-94.5 Moistureprolestos(cm)cl(cm)PorosityRMSDinKSDinK estimateTBatL-bandOverallWetDryOverallWetDry Currentapproach(asinFigure 1-1 )0.556.380.3719.1920.9922.4215.673.553.63Prole10.416.380.5519.1819.5623.6515.673.513.62Prole20.416.380.554.5310.372.644.523.962.46Prole30.416.380.554.557.313.654.504.022.67Prole40.416.380.559.144.1110.278.604.143.59 FordrydownduringDoY87.5-90.5;wetperiodDoY88.0-88.3;anddryperiodDoY89.5-90.5 Prole3&40.416.380.559.366.405.806.634.444.22 93

PAGE 94

Table4-3. Measurementsofwaterinputduringthedrydowns DrydownEventDescriptionWaterInput(mm)RangeofVSM Period1IrrigationonDoY130.84for45mins3.82-8%Period2IrrigationonDoY139.84for15mins4.42-8%Period3PrecipitationonDoY148.95for1.5days59.61-15%PrecipitationonDoY151.39for45mins1.6PrecipitationonDoY153.18for4hrs5.0 Table4-4. ofobserved0atHH,VV,andcross-poltosoilmoistureobservedat2,4,8,16,32,and64cm DepthPeriod1Period2Period3 HHVVCrossHHVVCrossHHVVCross 2cm0.380.610.610.630.730.130.870.900.874cm0.280.520.620.610.710.100.790.830.788cm0.070.310.620.370.500.100.590.640.5816cm-0.020.220.600.110.270.180.390.430.3732cm-0.130.110.60-0.49-0.540.090.220.240.1664cm0.01-0.14-0.47-0.28-0.40-0.080.060.04-0.01 Table4-5. EquationcoefcientsandR2ofregressionresultsatHH,VV,andcross-pol PolarizationSmoothRough 2%VSM<5%5%VSM8%1%VSM15% abR2abR2cdR2 HH-0.92-16.940.222.70-35.450.68-22.82-0.340.88VV-0.60-19.020.112.96-37.260.79-23.52-0.470.91Cross0.35-39.460.130.35-39.460.13-41.24-0.160.78 Table4-6. ofobservedepatH-poltosoilmoistureobservedat2,4,8,16,32,and64cm DepthPeriod1Period2Period3 2cm-0.92-0.97-0.924cm-0.85-0.95-0.898cm-0.69-0.80-0.7716cm-0.61-0.64-0.6332cm-0.480.70-0.4864cm0.430.75-0.34 Table4-7. EquationcoefcientsandR2ofregressionresultsofepatH-pol PolarizationSmoothRough 2%VSM8%1%VSM15% fghR2fghR2 H-0.02290.13700.75200.920.0021-0.00561.05470.88 94

PAGE 95

Table4-8. RootMeanSquareDifferences(RMSDs)andStandardDeviations(SD)betweentheobservedandmodeledH-polTBandco-pol0atL-bandusingdifferentmoistureprolesandsoilpropertiesofrootmeansquareheight(s),correlationlength(cl)andporosityduringsmoothperiod. FordrydownduringDoY130.5-134.5 Moistureprolestos(cm)cl(cm)PorosityTBHinK0HHindB0VVindB estimateTB@L-bandRMSDSDRMSDSDRMSDSD Prole10.819.790.5522.4821.185.381.52.561.48Prole30.819.790.559.866.184.241.961.881.74 ForwetperiodduringDoY130.9-131.5 Prole10.819.790.5540.7314.846.611.784.231.69Prole30.819.790.554.413.871.331.141.660.65 FordryperiodduringDoY131.5-134.5 Prole10.819.790.5516.866.115.211.362.171.15Prole30.819.790.5510.685.794.731.462.581.25 95

PAGE 96

Table4-9. RootMeanSquareDifferences(RMSDs)andStandardDeviations(SD)betweentheobservedandmodeledH-polTBandco-pol0atL-bandusingdifferentmoistureprolesandsoilpropertiesofrootmeansquareheight(s),correlationlength(cl)andporosityduringroughperiods FordrydownduringDoY148.8-157 Moistureprolestos(cm)cl(cm)PorosityTBHinK0HHindB0VVindB estimateTB@L-bandRMSDSDRMSDSDRMSDSD Prole12.089.450.5525.7118.824.151.393.841.50Prole32.089.450.554.554.711.471.451.241.23 96

PAGE 97

Figure4-1. ComparisonofemissionmodelsatL-band.A)AtH-pol.B)AtV-pol.SimulationusedthesoilmoistureandtemperatureprolesobservedduringMicroWEX-5. 97

PAGE 98

Figure4-2. Distributionofmoistureandtemperatureinmodiedsoillayersforthemicrowaveemissionmodelandcorrespondingobservationsatthedepthsof2,4,8,16,32,64,and120cmduringMicroWEX-5. Figure4-3. DielectricconstantsimulationsatC-andL-bandintermsofsoilporosity.A)RealandimaginarypartsofdielectricconstantsatC-bandsimulatedfrommodiedmodelsusingVSM=0.1,0.2,0.3,and0.4m3=m3.B)RealandimaginarypartsofdielectricconstantsatL-bandsimulatedfrommodiedmodelsusingVSM=0.1,0.2,0.3,and0.4m3=m3.. 98

PAGE 99

Figure4-4. ComparisonofobservedTBfromMicroWEX-5withmodeledTBduringthedriestandthewettestperiodsatC-band.A)ForH-pol.B)ForV-pol.Thesimulationusedthes,cl,andporosityestimatedduringthedriestperiodandmoistureestimatedduringthewettestperiod. 99

PAGE 100

Figure4-5. ThecomparisonofobservedTBfromMicroWEX-5andbestestimatedTBatC-band.A)ForH-pol.B)ForV-pol.Tenbreakingpoints,shownasverticaldashedlines,withtheirassociatedsoilmoisturevalues(inm3/m3)inthetop2mm. Figure4-6. Thecomparisonofsoilmoistureestimatedinthetop0-2mmandobservationsat2cmand4cmduringMicroWEX-5. 100

PAGE 101

Figure4-7. TheowchartofstepstoestimateH-polTBsatL-bandusingdifferentproles,wheres,cl,P,SMandMSDrepresentrootmeansquareheight,correlationlength,porosity,soilmoisture,andmoisturesensingdepth,respectively. 101

PAGE 102

Figure4-8. Thesoilmoistureprolesduringrstandseconddrydowns.A)Constantin0-2cmusingobservedsoilmoisture(SM)at2cmduringtherstdrydown.B)SMin0-2cmobtainedbylinearlyinterpolatingtheSMin0-2mmestimatedfromC-bandandobservedSMat2cmduringtherstdrydown.C)SMin0-2cmobtainedbylinearlyinterpolatingtheSMin0-1mmestimatedfromC-bandandobservedSMat2cmduringtherstdrydown.D)SMproleduringtheseconddrydown,SMin0-2cmobtainedbylinearlyinterpolatingtheSMin0-1(wetperiod)or0-2mm(dryperiod)estimatedfromC-bandandobservedsoilmoistureat2cm.SMrepresentssoilmoistureandthecolorbarshowstherangeofSMfrom0.01-0.32inm3=m3. 102

PAGE 103

Figure4-9. ComparisonTBobservedfromMicroWEX-5tothoseestimatedusingestimatedsoilproles.A)Usingobservedsoilmoistureat2cmforep;moisturewithinMSDof2.5cmatL-bandwithuniformweightingfunction(wu)forep;moisturewithinMSDof2.5cmatL-bandwithexponentiallydecayingweightingfunction(we)forep;andmoisturewithinMSDof1.3cmatL-bandwithweforep,duringDoY90.5-94.5.B)UsingmoisturewithinMSDof1.3and2.5cmatL-bandduringwetanddryperiods,respectively,withwe(dashedline),duringDoY87.5-90.5. 103

PAGE 104

Figure4-10. Thescatterplotsofobserved0toVSMat2cmandsensitivityduringsmoothandroughperiods.A)Scatterplotof0HHtoVSMobservedat2cmandtheirregressioncurves.B)Scatterplotof0VVtoVSMobservedat2cmandtheirregressioncurves.C)Scatterplotof0crtoVSMobservedat2cmandtheirregressioncurves.D)Thesensitivityof0atthreepolarizationcombinationstoVSMobservedat2cmintermsofVSMduringsmooth(S)andrough(R)periods. 104

PAGE 105

Figure4-11. ThescatterplotsofobservedeptoVSMat2cmandsensitivityduringsmoothandroughperiods.A)ScatterplotofepatH-poltoVSMobservedat2cmandtheirregressioncurves.B)ThesensitivityofepatH-poltoVSMobservedat2cmintermsofVSMduringsmooth(S)andrough(R)periods. Figure4-12. EstimatedVSMin0-2mmandtheVSMprolesduringsmoothperiod.A)EstimatedVSMin0-2mmandthoseobservedat2and4cm.B)VSMprolesinthetop5cm. 105

PAGE 106

Figure4-13. ComparisonofobservedandestimatedTBand0duringsmoothperiod.A)TBatH-pol.B)0atHH-pol.C)0atVV-pol.SimulatiionsusedVSMobservedat2cmandmodiedVSMproles,s,andcl. 106

PAGE 107

Figure4-14. EstimatedVSMin0-2mmandtheVSMprolesduringroughperiod.A)EstimatedVSMin0-2mmandthoseobservedat2and4cm.B)VSMprolesinthetop5cm. 107

PAGE 108

Figure4-15. ComparisonofobservedandestimatedTBand0duringroughperiod.A)TBatH-pol.B)0atHH-pol.C)0atVV-pol.SimulationsusedVSMobservedat2cmandmodiedVSMproles,s,andcl. 108

PAGE 109

CHAPTER5CONCLUSIONS,CONTRIBUTIONS,ANDRECOMMENDATIONSThischaptersummarizessignicanceandcontributionsofthisdissertation,anddiscussestheimplicationsofthendingsforsoilmoistureestimatesusingactiveandpassivemicrowaveremotesensing.Therecommendationsofthedirectionsforthefuturestudiesarealsoprovidedinthefollowingsection. 5.1ConclusionsThisdissertationprovidesinsightsintotheimpactsofsoilmoisturedistributionwithintheMSD,soilporosity,andsurfaceroughnessonAPsignatures,andintomicrowaveemissionandbackscattermodelingforrealisticestimatesofAPsignaturesforbaresandysoils.Threesciencequestionswereaddressedinthisdissertation.Whatarethesourcesthatmaycauseunrealisticestimatesofmicrowavesignaturesusingcurrentmicrowaveforwardmodelsfornear-surfacesoilmoistureestimates?Forthepassivesignature,TBfromanunvegetatedlandsurfaceisestimatedasthesumofcontributionsfromsoilandsky,asmentionedinEquation 1 .Becausethecontributionfromskyisinsignicantat3-5K,thepotentialsourcesthatcauseunrealisticestimatesofTBareprimarilyinmodelingthesoilcontribution.Insection4.1.1,itwasfoundthatthereectionsbetweenthesoillayersmaynotbesignicantinthesoilcolumnofthesandysoilsattheMicroWEX-5site,irrespectiveofwhetherincoherentsolution,zeroorderorrstorderapproximationtotheradiativetransferequationswasused.Inaddition,theimpactofthesephysically-basedmodelsonTBisverysmall,about3K.Therefore,soilemissivityhasthemostsignicantimpactonTBsignatures.Theroughsurfaceemissivityismodeledasafunctionofandsurfaceroughness,sandcl.AlthoughtheTBatC-bandismoresensitivetosandclthanthoseatL-band,theimpactsofsurfaceroughnessontheestimatedTBatC-andL-bandarenotsignicantenoughtoaccountfortheobserveddiscrepanciesofupto35KintheFigures 1-1 .Thus,isthemostdominantfactorinthesoilemissionandcanberelatedtothesoil 109

PAGE 110

moistureandsoilporosity.SoilporosityismoreimpactfulontheTBduringthedryperiod,butwasstillnotadequatetoimprovetheTBestimationduringextremehydrologicconditions.Itwasfoundthatthesoilmoistureisthemostimpactfulvariabletothemodel.However,thecurrentinsitusoilmoisturemeasurementsat2cmusingTDRprobearenotabletomeasuremoisturewithhighverticalresolutioninthenear-surfaceofthesoil.Suchmeasurementsarenotadequatetorepresentthehighlydynamicproleofsoilmoistureintheupperfewcentimetersforsandysoils,particularlyduringandimmediatelyfollowinghydrologicevents,and,therefore,causestheunrealisticestimatesofmicrowavesignaturesusingcurrentmicrowavemodels.Fortheactivesignature,thebackscatteringmodelisafunctionofofandsurfaceroughness,sandcl.Duringhydrologicevents,thesoilmoisturedistributedinthenear-surfaceisthemostimpactfulfactorforthehighlydynamicsoilconditions.AsshowninFigures4-13BandCand4-15BandC,theabruptincreasingoftheobserved0cannotbeobtainedassoilmoisturemeasurementsat2cmwereused,evenwithmodiedsurfaceroughnessestimations.WhatisthesensitivityoftheAPmicrowavesignaturestothesoilmoistureandsurfaceroughness?Sensitivityof0andeptoobservedVSMandsurfaceroughnesswereevaluatedtounderstandthecomplementarityofAP.ThismaybetherststudythatusedtheconcurrentAPobservationsatunprecedentedhightemporalfrequencyof15minutesandsoilmoisturemeasurementsforthesensitivityanalysesandprovedthattheepismoresensitivetoVSMthan0,particularlyduringthesmoothperiod,butislargelyinsensitivetoroughness.Duringdry,smoothperiod,signicantvolumescatteringwasobservedontheco-pol0showinganegativelinearrelationshiptoVSM,andinsignicantrelationshipbetweencross-pol0toVSMwasalsoobserved.Inaddition,thisstudyalsoindicatesthaterrorsusingactive-onlyobservationsforsoilmoistureestimationsmayoccurbecauseofweakreturnsignalbackscatteredunderthedryandsmootheld. 110

PAGE 111

HowcanthecomplementarityofAPmicrowavesignaturesbeusedforprovidingsoilmoistureestimatesfortheupcomingNASA-SMAP?Baseduponpreviousndings,amethodologytointegratetheAPobservationsatL-bandforsoilmoistureandsurfaceroughnessestimateswasimplementedinthisdissertation.TheTBswereusedtoestimateasoilmoistureprole,duetoitshighersensitivitytoVSM,whilethe0swereusedtoestimatesandcl.Theestimatedsoilmoistureproles,s,andclimprovedRMSDsbetweentheobservedandmodeledTB,0HHand0VVvaluesby21K,and2.68and2.60dB,respectively,duringtheroughsurface.Lesssignicantimprovementwasfoundduringthesmoothperiod,particularfortheactiveestimations,becauseofvolumescattering.Itwasdemonstratedthatimprovedestimatesofsoilmoisturedistributioninthenear-surface,aswellasthesurfaceroughnessparameters,arefeasiblebyintegratingAPmicrowaveobservations.Theseestimatedsoilmoistureproles,sandclprovidedconsistentAPestimationsbyusingcurrentemissionandbackscattermodels,particularlyduringroughsurface. 5.2ContributionsTheresearchinthisdissertationresultedinvesignicantcontributionsdescribedbelow: ImpactsofmoisturedistributionwithintheMSDandthesoilcharacteristicsonbothpassiveC-andL-bandsignaturesisexploredandamethodologythatincorporatepassivedual-polC-bandandH-polL-bandobservationstoestimatethesoilmoistureandporosityinthenear-surfaceandsurfaceroughnessisimplemented.ThisresultedinphysicallyconsistentmicrowaveemissionestimationsatbothC-andL-bandforsandysoils. TheconcurrentobservationsofAPmicrowavesignaturesattheL-bandandsoilmoistureandtemperatureproleswereconductedduringMicroWEX-11forabareagriculturaleldoverapproximate30days,whichincludeeldpreparationforplantingandafterfakeplanting.TogetherwiththeobservationsduringFall,2012,forthesweetcorn,wehaveadatasetcoveracompletesweetcorngrowingseasonforthefurthermodeldevelopmentandvalidations. DataprocessingandcalibrationproceduresarealsoprovidedinthisdissertationfortheAPmicrowavesensorstoconverttherawobservationsintoTBand0. 111

PAGE 112

ThecomplementarityofAPsignatureswasinvestigatebyevaluatingthesensitivityof0andeptoobservedVSMandtosurfaceroughnessusingobservationsfromMicroWEX-11.AmethodologywasimplementedtointegratetheAPobservationsforsoilmoistureandsurfaceroughnessestimatesforhighlydynamicsandysoils. TherelationshipbetweenAPsignaturewasexplored,andthesensitivitiesofAPsignaturestosoilmoistureandsurfaceroughnesswereevaluated.TheseinformationfromsuchahightemporalfrequencydatasetmaybeprovidedfordevelopingAPcombinedalgorithmsfortheupcomingNASA-SMAPmission. 5.3RecommendationsforFutureResearchInthissection,sixrecommendations,ontheperspectivesofmodellingandeldwork,areprovidedforfurtherimprovementsforsoilmoistureestimatesbyusingAPmicrowaveremotesensing. Animprovedapproachthatcanusethedetailedproleratherthanoneeffectivesoilmoisturevaluetoestimate0andTBisessential.AlthoughthesuccessfulimplementationofusingweightingfunctionstointegratesoilmoisturedistributedwithinMSDforL-band0andTBestimates,theMSDandweightingfunctionsdependuponthesoilmoistureandtextures,whichstillcannotbewelldenedphysically. AccuratedynamichydrologicalmodelsuchasLSM,NOAH,andLSP,maybedevelopedandusedtoprovidedetailedmoistureproleinthesoilcolumn,withaverticalresolutionsufcientforTBand0estimationatL-band,becauseitisalmostimpossibletoaccuratelyacquiresoilmoisturedistributionwithin0-2cmusingcurrentinsitusensors.Althoughsuchamethodwillincreasethecomputationaldemands,currentcapabilitiesinhighperformancecomputingmayallowsuchinclusions. Theoperationaluseofproposedmethodologiesinthisdissertationneedfurtherupscalingfromeldtosatellitescalesforthecurrentandupcomingspacemissions.Developingdataup-anddown-scallingalgorithmsarespeciallyessentialforapplyingthesemethodologiesforsoilmoistureestimatesuniversally. Toincorporateaphysically-basedvolumescatteringmodelwithsurfacebackscatteroremissionmodelisrequiredforbetterestimatesof0andTB,particularlyduringdry,smoothsandysoils. Becausestate-of-the-artsoilmoisturesensorsarenotabletomeasurethesoilmoisturechangeinthetop2cm,suchmeasurementsresultinunrealisticmicrowavesignatureestimatesusingmicrowaveforwardmodelsandcouldcausepoorperformanceofthesoilmoistureestimates.Thegravimetricsoilmoisture(GSM)measurementsusingsoilsamplesaresuggestedformodeldevelopment 112

PAGE 113

andvalidation.The0-1,0-2,and0-5cmareparticularlysuggestedwithfourmeasurementsatabout12am,6am,1pm,and6pmbaseduponthediurnalvariationofTBobservedintheeldmeasurementsandthesatellitessuchasESA-SMOSandNASA-SMAPrevisitedtime.Moremeasurementsareneedduringthehydrologicevents,ateveryhourforabout5hoursfollowingbytheevents.Althoughthediurnalvariationisalsoaffectedbysoiltemperature,themoisturechangeonthetop-surfaceisfoundtobetheprimarycontributor.Inaddition,bycalibratingmoistureprolemodelusinggravimetricsoilmoisturemeasurements,relationshipofmoisturebetweentop-surface(fewmm)fromGSMandatdeeperlayer(2cm)observedbyTDRcouldbebuilt. ThecurrentcalibrationerrorsofUF-LARSwasestimatedusingtheradarequationwithinputsofuncertaintiesofheightandattitudemeasurementsfromrangelaserandinclinometer.Becausethegeometricrelationshipbetweenthecenterofsmall-sizetrihedralcorner-reectorandtheantennaboresightishardlydetermined,theerrorcontributedfromcalibrationtargetisnotwelldenedyet.Thus,thecalibrationerrorofradarmeasurementsneedtobebettercharacterizedbyusingapointinglasertoassistdeterminingthelocationofboresightandanglebetweenboresightandthecenterofcalibrationtarget.Inaddition,thesphericaltargetisworthwhiletobeusedandtocomparewiththecurrenttrihedralcorner-reector.Althoughthetrihedralcorner-reectorprovidedreasonablecalibratedresults,thetargetmeasurementisverysensitivetotheangletowardtheboresightofantennaatcross-polmeasurements.Thesphericaltargetisnotsensitivetotheangleandwillprovidemoreaccuratecalibratedco-pol0. 113

PAGE 114

REFERENCES [1] (2012)AAEON6915RevBEmbeddedController.[Online].Available: http://embeddedpc.aaeon.com.tw [2] C.Albergel,P.deRosnay,G.C,J.Munoz-Sabater,S.Hasenauer,L.Isaksen,Y.Kerr,andW.Wagner,Evaluationofremotelysensedandmodelledsoilmoistureproductsusingglobalground-basedinsituobservations,RemoteSens.Env.,vol.118,pp.215,2012. [3] L.BabinerandR.Schafer,Thechirpz-transformalgorithm,IEEETrans.AudioElectroacoust.,vol.AU-17,no.2,pp.86,1969. [4] A.Balenzano,F.Mattia,G.Satalino,andM.Davidson,DensetemporalseriesofC-andL-bandSARdataforsoilmoistureretrievaloveragriculturalcrops,IEEEJ.Sel.TopicsAppl.EarthObserv.RemoteSens.(JSTARS),vol.4,no.2,pp.439,2011. [5] B.Barrett,E.Dwyer,andP.Whelan,Soilmoistureretrievalfromactivespacebornemicrowaveobservations:Anevaluationofcurrenttechniques,RemoteSens.,vol.1,no.3,pp.210,2009. [6] Z.Bartalis,V.Naeimi,S.Hasenauer,K.Scipal,H.Bonekamo,J.Figa,andC.Anderson,InitialsoilmoistureretrievalsfromtheMETOP-AAdvancedScatterometer(ASCAT),Geophys.Res.Letters,vol.34,no.L20401,p.doi:10.1029/2007GL031088,2007. [7] T.Bongiovanni,H.Enos,A.Monsivais-Huertero,B.Colvin,K.Nagarajan,J.Judge,P.Liu,J.Fernandez-Diaz,R.D.R.Y.Goykhman,X.Duan,D.Preston,R.Shrestha,C.Slatton,M.Moghaddam,andA.England,Fieldobservationsduringtheeighthmicrowave,water,andenergybalanceexperiment(MicroWEX-8):fromJune16throughAugust24,2009,http://edis.ifas.u.edu/AE476,CenterforRemoteSensing,UniversityofFlorida,Tech.Rep.,2009. [8] L.Brocca,S.Hasenauer,T.Lacava,F.Melone,T.Moramarco,W.Wagner,W.Dorigo,J.Martnez-Fernadez,P.Llorens,J.Latron,C.Martin,andM.Bittelli,SoilmoistureestimationthroughASCATandAMSR-Esensors:AnintercomparisonandvalidationstudycrossEurope,RemoteSens.Env.,vol.115,pp.3390,2011. [9] W.Burke,T.Schmugge,andJ.Paris,Comparisonof2.8-and21-cmmicrowaveradiometerobservationsoversoilswithemissionmodelcalculations,J.Geophys.Res.,vol.84,no.C1,pp.287,1979. [10] J.-C.Calvet,J.-P.Wigneron,J.Walker,F.Karbou,A.Chanzy,andC.Albergel,Sensitivityofpassivemicrowaveobservationstosoilmoistureandvegetationwatercontent:L-bandtoW-band,IEEETrans.Geosci.RemoteSensing,vol.49,pp.1190,2011. 114

PAGE 115

[11] (2008)TheRADARSAT-2ASARinstrument.[Online].Available: http://www.asc-csa.gc.ca/eng/satellites/radarsat2/ [12] J.CasanovaandJ.Judge,EstimationofenergyandmoistureuxesfordynamicvegetationusingcoupledSVATandcrop-growthmodels,WaterRes.Research,vol.44,no.W07415,p.doi:10.1029/2007WR006503,2008. [13] J.Casanova,T.Lin,M.Jang,K.Tien,J.Judge,O.Lanni,andL.Miller,Fieldobservationsduringthefourthmicrowave,water,andenergybalanceexperiment(MicroWEX-4):fromMarch10throughJune14,2005.Circularno.1482,http://edis.ifas.u.edu/AE362,CenterforRemoteSensing,UniversityofFlorida,Tech.Rep.,2005. [14] J.Casanova,F.Yan,M.Jang,J.Fernandez,J.Judge,C.Slatton,K.Tien,T.Lin,O.Lanni,andL.W.Miller,Fieldobservationsduringthefthmicrowave,water,andenergybalanceexperiment(MicroWEX-5):fromMarch9throughMay26,2006.Circularno.1514,http://edis.ifas.u.edu/AE407,CenterforRemoteSensing,UniversityofFlorida,Tech.Rep.,2006. [15] A.Chanzy,S.Raju,andJ.-P.Wigneron,EstimationofsoilmicrowaveeffectivetemperatureatL-andC-bands,IEEETrans.Geosci.RemoteSensing,vol.35,pp.570,1997. [16] N.Chauhan,Soilmoistureestimationunderavegetationcover:combinedactivepassivemicrowaveremotesensingapproach,Int.J.RemoteSensing,vol.18,no.5,pp.1079,1997. [17] N.Chauhan,D.LeVine,andR.Lang,Discretescattermodelformicrowaveradarandradiometerresponsetocorn:comparisonoftheoryanddata,IEEETrans.Geosci.RemoteSensing,vol.32,no.2,pp.416,March1994. [18] K.Chen,T.Wu,L.Tsang,Q.Li,J.Shi,andA.Fung,Emissionofroughsurfacescalculatedbytheintegralequationmethodwithcomparisontothree-dimensionalmomentmethodsimulations,IEEETrans.Geosci.RemoteSensing,vol.41,no.1,pp.90,2003. [19] K.Chen,T.Wu,M.Tsay,andA.Fung,AnoteonthemultiplescatteringinanIEMmodel,IEEETrans.Geosci.RemoteSensing,vol.38,no.1,pp.249,2000. [20] B.Choudhury,T.Schmugge,andT.Mo,Aparameterizationofeffectivesoiltemperatureformicrowaveemission,J.Geophys.Res.,vol.87,no.C2,pp.1301,1982. [21] N.Das,D.Entekhabi,andE.Njoku,AnalgorithmofmergingSMAPradiometerandradarforhigh-resolutionsoil-moistureretrieval,IEEEJ.Sel.TopicsAppl.EarthObserv.RemoteSens.(JSTARS),vol.49,no.5,pp.1504,2011. 115

PAGE 116

[22] R.deJeu,W.Wagner,T.Holmes,A.Dolman,N.vandeGiesen,andJ.Friesen,Globalsoilmoisturepatternsobservedbyspacebornemicrowaveradiometersandscatterometers,Surv.Geophys,vol.29,pp.399,2008. [23] O.DenmeadandR.Shaw,Availabilityofsoilwatertoplantsasaffectedbysoilmoisturecontentandmeteorologicalconditions,Agronomyjournal,vol.54,no.5,pp.385,1962. [24] M.DobsonandF.Ulaby,Activemicrowavesoilmoistureresearch,IEEETrans.Geosci.Electron.,vol.GE-24,no.1,pp.23,January1986. [25] M.Dobson,F.Ulaby,M.Hallikainen,andM.El-Rayes,Microwavedielectricbehaviorofwetsoil-PartII:DielectricMixingModels,IEEETrans.Geosci.RemoteSensing,vol.GE-23,no.1,pp.35,January1985. [26] A.Doerry,ReectorsforSARPerformanceTesting,SandiaNationalLaboratories,Albuquerque,NewMexicoandLivermore,California,USA,Tech.Rep.,2008.[Online].Available: http://prod.sandia.gov/techlib/access-control.cgi/2008/080396.pdf [27] L.Dong,N.Baghdadi,andR.Ludwig,Validationoftheaiemthroughcorrelationlengthparameterizationateldscaleusingradarimageryinasemi-aridenvironment,IEEEGeosci.RemoteSens.Lett.,vol.10,no.3,pp.461,May2013. [28] M.Drusch,T.Holmes,P.deRosnay,andG.Balsamo,ComparingERA-40basedL-bandbrightnesstemperatureswithSkylabobservations:Acalibration/validationstudyusingtheCommunityMicrowaveEmissionModel,J.Hydrometeorology,vol.10,pp.213,2009. [29] M.Drusch,E.Wood,andT.Jackson,Vegetativeandatmosphericcorrectionsforthesoilmoistureretrievalfrompassivemicrowaveremotesensingdata:Resultsfromthesoutherngreatplainshydrologyexperiment1997,J.Hydrometeorology,vol.2,pp.181,2001. [30] Y.Du,F.Ulaby,andM.Dobson,Sensitivitytosoilmoisturebyactiveandpassivemicrowavesensors,IEEETrans.Geosci.RemoteSensing,vol.38,no.1,pp.105,January2000. [31] P.Dubois,J.vanZyl,andT.Engman,Measuringsoilmoisturewithimagingradars,IEEETrans.Geosci.RemoteSensing,vol.33,no.4,pp.915,July1995. [32] M.Ek,K.Mitchell,Y.Lin,E.Rogers,P.Grunmann,V.Koren,G.Gayno,andJ.D.Tarpley,ImplementationofNoahlandsurfacemodeladvancesintheNationalCentersforEnvironmentalPredictionoperationalmesoscaleEtamodel,J.Geophys.Res.,vol.108,no.D22,p.doi:10.1029/2002JD003296,2003. 116

PAGE 117

[33] D.Entekhabi,N.Das,E.Njoku,J.Johnson,andJ.Shi,AlgorithmtheoreticalbasisdocumentL2&L3radar/radiometersoilmoisture(Active/Passive)dataproducts,http://smap.jpl.nasa.gov/les/smap2/L2&3 SM AP InitRel v11.pdf,JetPropulsionLaboratory,CaliforniaInstituteofTechnology,Tech.Rep.,2012. [34] D.Entekhabi,E.Njoku,P.O'Neill,K.Kellogg,W.Crow,W.Edelstein,J.Entin,S.Goodman,T.Jackson,J.Johnson,J.Kimball,J.Piepmeier,R.Koster,N.Martin,K.McDonald,M.Moghaddam,S.Moran,R.Reichle,J.Shi,M.Spencer,S.Thurman,L.Tsang,andJ.V.Zyl,Thesoilmoistureactivepassive(SMAP)mission,Proc.oftheIEEE,vol.98,pp.704,2010. [35] M.Escorihuela,A.Chanzy,J.-P.Wigneron,andY.Kerr,EffectivesoilmoisturesamplingdepthofL-bandradiometry:Acasestudy,RemoteSens.Env.,vol.114,pp.995,2010. [36] J.Fernandes-Diaz,Characterizationofsurfaceroughnessofbareagriculturalsoilsusinglidar,Ph.D.dissertation,Univ.ofFlorida,Gainesville,Florida,USA,December2010. [37] P.FerrazzoliandL.Guerriero,PassiveMicrowaveRemoteSensingofForests:AModelInvestigation,IEEETrans.Geosci.RemoteSensing,vol.34,no.2,pp.433,1996. [38] D.Frickey,UsingtheInverseChirp-ZTransformforTime-DomainAnalysisofSimulatedRadarSignals,IdahoNationalEngineeringLaboratory,IdahoFalls,ID,USA,Tech.Rep.,1995.[Online].Available: http://www.osti.gov/bridge/servlets/purl/10110067-xpHJWC/webviewable/ [39] A.Fung,MicrowaveScatteringandEmissionModelsandTheirApplications.Norwood,MA:ArtechHouse,1994. [40] A.FungandK.Chen,AnupdateontheIEMsurfacebackscatteringmodel,IEEEGeosci.RemoteSens.Lett.,vol.1,no.2,pp.75,April2004. [41] A.FungandN.Kuo,Backscatteringfrommulti-scaleandexponentiallycorrelatedsurfaces,J.Electro.WavesandApp.,vol.20,no.1,pp.3,2006. [42] A.Fung,Z.Li,andK.Chen,Backscatteringfromarandomlyroughdielectricsurface,IEEETrans.Geosci.RemoteSensing,vol.30,no.2,pp.356,March1992. [43] C.Gruhier,P.deRosnay,S.Hasenauer,T.Holmes,R.deJeu,Y.Kerr,E.Mougin,E.Njoku,F.Timouk,W.Wagner,andM.,Soilmoistureactiveandpassivemicrowaveproducts:intercomparisonandevaluationoveraSaheliansite,Hydro.EarthSyst.Sc.,vol.14,pp.141,2010. 117

PAGE 118

[44] F.Hall,J.Townshend,andE.Engman,Statusofremotesensingalgorithmsforestimationoflandsurfacestateparameters,RemoteSens.Env.,vol.51,pp.138,1995. [45] D.Hamby,Areviewoftechniquesforparametersensitivityanalysisofenvironmentalmodels,Environ.Monit.Assess.,vol.32,no.2,pp.135,1994. [46] D.Hoekman,Speckleensemblestatisticsoflogarithmicallyscaleddata,IEEETrans.Geosci.RemoteSensing,vol.29,no.1,pp.180,1991. [47] T.Holmes,M.Drusch,J.-P.Wigneron,andR.deJeu,Aglobalsimulationofmicrowaveemission:ErrorstructuresbasedonoutputfromECMWFsoperationalintegratedforecastsystem,IEEETrans.Geosci.RemoteSensing,vol.46,no.3,pp.846,2008. [48] (2012)HP8753DNetworkAnalyzer.[Online].Available: http://www.teknetelectronics.com/DataSheet/HP AGILENT/HP 8753.pdf [49] (2011)TheRISAT-1instrument.[Online].Available: http://www.isro.org/satellites/RISAT-1.aspx [50] M.Jang,K.Tien,J.Casanova,andJ.Judge,MeasurementsofsoilsurfaceroughnessduringtheFourthMicrowaveWaterandEnergyBalanceExperiment:April18throughJune13,2005,http://edis.ifas.u.edu/AE363,CenterforRemoteSensing,UniversityofFlorida,Tech.Rep.,2005. [51] (2012)Globalchangeobservationmission:Fourthresearchannouncement.[Online].Available: http://suzaku.eorc.jaxa.jp/GCOM/materials/ra/GCOM RA4 guide E.pdf [52] J.Judge,Microwaveremotesensingofsoilwater:Recentadvancesandissues,Trans.oftheASABE,vol.50,no.5,pp.1645,2007. [53] J.Judge,J.Casanova,T.Lin,K.-J.Tien,M.Jang,O.Lanni,andL.Miller,Fieldobservationsduringthesecondmicrowave,water,andenergybalanceexperiment(MicroWEX-2):fromMarch17throughJune3,2004.Circularno.1480,http://edis.ifas.u.edu/AE360,CenterforRemoteSensing,UniversityofFlorida,Tech.Rep.,2004. [54] Y.Kerr,P.Waldteufel,J.Wigneron,,S.Delwart,F.Cabot,J.Boutin,M.Escorihuela,J.Font,N.Reul,C.Gruhier,S.Juglea,M.Drinkwater,A.Hahne,M.Marti'n-Neira,andS.Mecklenburg,TheSMOSmission:Newtoolformonitoringkeyelementsoftheglobalwatercycle,Proc.oftheIEEE,vol.98,no.5,pp.666,2010. [55] F.Koudogbo,H.Mametsa,andP.Combes,Surfaceandvolumescatteringfromnaturalandmanmaderoughsurfacesintheprocessofsettingupdatabasecoefcients,GeoscienceandRemoteSensingSymposium,2003.IGARSS'03.Proceedings.2003IEEEInternational,vol.7,July2003. 118

PAGE 119

[56] D.LeVine,G.Lagerloef,F.Colomb,S.Yueh,andF.Pellerano,Aquarius:Aninstrumenttomonitorseasurfacesalinityfromspace,IEEETrans.Geosci.RemoteSensing,vol.45,no.7,pp.2040,July2007. [57] T.Lin,J.Judge,K.Tien,J.Casanova,M.Jang,O.Lanni,L.Miller,andF.Yan,Fieldobservationsduringthethirdmicrowave,water,andenergybalanceexperiment(MicroWEX-3):fromJune16throughDecember21,2004.Circularno.1481,http://edis.ifas.u.edu/AE361,CenterforRemoteSensing,UniversityofFlorida,Tech.Rep.,2004. [58] Y.Liu,R.Rarinussa,W.Dorigo,R.deJeu,W.Wagner,A.vanDijk,F.McCabe,andJ.Evans,Developinganimprovedsoilmoisturedatasetbyblendingpassiveandactivemicrowavesatellite-basedretrievals,Hydro.EarthSyst.Sc.,vol.15,no.2,pp.425,2011. [59] (2012)LokeEngineeringRangeLaser.[Online].Available: http://loke.de/index.php?option=com contentn&view=articlen&id=26n&Itemid=33n&lang=en [60] R.Lucas,J.Armston,R.Fairfax,R.Fensham,A.Accad,J.Carreiras,J.Kelley,P.Bunting,D.Clewley,S.Bray,D.Metcalfe,J.Dwyer,M.Bowen,T.Eyre,M.Laidlaw,andM.Shimada,AnevaluationoftheALOSPALSARL-bandbackscatterabovegroundbiomassrelationshipQueensland,Australia:Impactsofsurfacemoistureconditionandvegetationstructure,IEEEJ.Sel.TopicsAppl.EarthObserv.RemoteSens.(JSTARS),vol.3,no.4-2,pp.576,2010. [61] A.Merzouki,H.McNairn,andA.Pacheco,MappingsoilmoistureusingRADARSAT-2dataandlocalautocorrelationstatistics,IEEEJ.Sel.TopicsAppl.EarthObserv.RemoteSens.(JSTARS),vol.4,no.1,pp.128,March2011. [62] V.Mironov,L.Kosolapova,andS.Fomin,Physicallyandmineralogicallybasedspectroscopicdielectricmodelformoistsoils,IEEETrans.Geosci.RemoteSensing,vol.47,no.7,pp.2059,July2009. [63] T.MoandT.Schmugge,Aparameterizationoftheeffectofsurfaceroughnessonmicrowaveemission,IEEETrans.Geosci.RemoteSensing,vol.25,no.4,pp.47,1987. [64] L.Morena,K.James,andJ.Beck,AnintroductiontotheRADARSAT-2mission,Can.J.RemoteSensing,vol.30,no.3,pp.221,2004. [65] K.Nagarajan,P.-W.Liu,R.DeRoo,J.Judge,R.Akbar,P.Rush,S.Feagle,andR.Terwilleger,AutomatedL-bandradarsystemforsensingsoilmoistureathightemporalresolution,grsl,vol.Submitted,2012. [66] A.NavarinniandT.Pisanu,L-bandorthomodetransducerforthesardiniaradiotelescope,vol.7014.Proc.SPIE,2008. 119

PAGE 120

[67] E.NjokuandD.Entekhabi,Passivemicrowaveremotesensingofsoilmoisture,J.Hydrology,vol.184,no.1-2,pp.101,October1996. [68] E.Njoku,W.Wilson,S.Yueh,S.Dinardo,F.Li,T.Jackson,V.Lakshmi,andJ.Bolten,Observationsofsoilmoistureusingapassiveandactivelow-frequencymicrowaveairbornesensorduringSGP99,IEEETrans.Geosci.RemoteSensing,vol.40,no.12,pp.2659,December2002. [69] Y.Oh,Asemi-empiricalmodelformicrowavepolarimetricradarbackscatteringfrombaresoilsurfaces,JKoreanSocRemoteSensing,vol.10,no.2,pp.17,1994. [70] ,Quantitativeretrievalofsoilmoisturecontentandsurfaceroughnessfrommultipolarizedradarobservationsofbaresoilsurfaces,IEEETrans.Geosci.RemoteSensing,vol.42,no.3,pp.596,March2004. [71] Y.Oh,K.Sarabandi,andF.Ulaby,Asemi-empiricalmodeloftheensemble-averageddifferentialMuellermatrixformicrowavebackscatteringfrombaresoilsurfaces,IEEETrans.Geosci.RemoteSensing,vol.40,no.6,pp.1348,June2002. [72] ,Anempiricalmodelandaninversiontechniqueforradarscatteringfrombaresoilsurfaces,IEEETrans.Geosci.RemoteSensing,vol.30,no.2,pp.370,March1992. [73] P.O'Neill,N.Chauhan,andT.Jackson,Useofactiveandpassivemicrowaveremotesensingforsoilmoistureestimationthroughcorn,Int.J.RemoteSensing,vol.17,no.10,pp.1851,July1996. [74] R.Panciera,J.Walker,andO.Merlin,ImprovedunderstandingofsoilsurfaceroughnessparameterizationforL-bandpassivemicrowavesoilmoistureretrieval,IEEEGeoscienceandRemoteSensingLetters,vol.6,no.4,pp.625,2009. [75] T.Pellarin,Y.Kerr,andJ.-P.Wigneron,Globalsimulationofbrightnesstemperatureat6.6and10.7GHzoverlandbasedonSMMRdatasetanalysis,IEEETrans.Geosci.RemoteSensing,vol.44,no.9,pp.2492,2006. [76] N.Peplinski,F.Ulaby,andM.Dobson,Dielectricpropertiesofsoilsinthe0.3-1.3GHzrange,IEEETrans.Geosci.RemoteSensing,vol.33,no.3,pp.803,May1995. [77] M.Piles,D.Entekhabi,andA.Camps,Achangedetectionalgorithmforretrievinghigh-resolutionsoilmoisturefromSMAPradarandradiometerobservations,IEEEJ.Sel.TopicsAppl.EarthObserv.RemoteSens.(JSTARS),vol.47,no.12,pp.4125,2012. 120

PAGE 121

[78] R.Prakash,D.Singh,andN.Pathak,AfusionapproachtoretrievesoilmoisturewithSARandopticaldata,IEEEJ.Sel.TopicsAppl.EarthObserv.RemoteSens.(JSTARS),vol.5,no.1,pp.196,2012. [79] C.Prigent,F.Aires,W.Rossow,andA.Robock,Sensitivityofsatellitemicrowaveandinfraredobservationstosoilmoistureataglobalscale:Relationshipofsatelliteobservationstoinsitusoilmoisturemeasurements,J.Geophys.Res.,vol.110,no.D07110,p.doi:10.1029/2004JD005087,2005. [80] C.Prigent,J.-P.Wigneron,W.Rossow,andJ.Pardo-Carrion,Frequencyandangularvariationsoflandsurfacemicrowaveemissivities:CanweestimateSSM/TandAMSUemissivitiesfromSSM/Iemissivities,IEEETrans.Geosci.RemoteSensing,vol.38,no.5,pp.2373,2000. [81] P.Rao,L.Venkataratnam,P.Rao,K.Ramana,andM.Singarao,Relationbetweenrootzonesoilmoistureandnormalizeddifferentcevegetationindexofvegetatedeld,Int.J.RemoteSensing,vol.14,no.3,pp.441,1993. [82] R.Reichle,W.Crow,andC.Kepenne,AnadaptiveensembleKalmanlterforsoilmoisturedataassimilation,WaterRes.Research,vol.44,no.W03423,p.doi:10.1029/2007WR006357,2008. [83] C.Rudiger,T.Holmes,J.-C.Calvet,R.deJeu,andW.Wagner,AnintercomparisonofERS-ScatandAMSR-EsoilmoistureobservationswithmodelsimulationsoverFrance,J.Hydrometeorology,vol.10,no.2,pp.431,2009. [84] K.SarabandiandF.Ulaby,Aconvenienttechniqueforpolarimetriccalibrationofsingle-antennaradarsystems,IEEETrans.Geosci.RemoteSensing,vol.28,no.6,pp.1022,November1990. [85] T.SchmuggeandB.Choudhury,Acomparisonofradiativetransfermodelsforpredictingthemicrowaveemissionfromsoil,RadioScience,vol.16,no.5,pp.927,1981. [86] K.Schneeberger,M.Schwank,C.Stamm,P.deRosnay,C.Matzler,andH.Fluhler,Topsoilstructureinuencingsoilwaterretrievalbymicrowaveradiometry,VadoseZoneJournal,vol.3,pp.1169,2004. [87] M.Schwank,I.Volksch,J.-P.Wigneron,Y.Kerr,A.Mialon,P.deRosnay,andC.Matzler,Comparisonoftwobare-soilreectivitymodelsandvalidationwithL-bandradiometermeasurements,IEEETrans.Geosci.RemoteSensing,vol.48,no.1,pp.325,2010. [88] J.Shi,K.Chen,L.Qin,T.Jackson,P.O'Neill,andL.Tsang,Aparameterizedsurfacereectivitymodelandestimationofbare-surfacesoilmoisturewithL-bandradiometer,IEEETrans.Geosci.RemoteSensing,vol.40,no.12,pp.2674,2002. 121

PAGE 122

[89] J.Shi,J.vanZyl,J.Soares,andE.Engman,Retrievalbaer-soilmoistureusingL-bandSAR,InternationalArchivesofPhotogrammetryandRemoteSensing,vol.29,PartB7,pp.595,1992. [90] D.Thoma,M.Moran,R.Bryant,M.Rahman,C.Holield-Collins,S.Skirvin,E.Sano,andK.Slocum,ComparisonoffourmodelstodeterminesurfacesoilmoisturefromC-bandradarimageryinasparselyvegetatedsemiaridlandscape,WaterRes.Research,vol.42,no.W01418,p.doi:10.1029/2004WR003905,2006. [91] K.-J.Tien,R.D.roo,J.Judge,andH.Pham,Comparisonofcalibrationtechniquesforground-basedC-Bandradiometers,IEEEGeoscienceandRemoteSensingLetters,vol.4,no.1,pp.83,2007. [92] J.T.J,P.ONeill,andC.Swift,Passivemicrowaveobservationofdiurnalsurfacesoilmoisture,IEEETrans.Geosci.RemoteSensing,vol.35,no.5,pp.1210,1997. [93] L.Tsang,J.Kong,andK.-H.Ding,ScatteringofElectromagneticWaves,TheoriesandApplications.JohnWileyandSons,Inc,2001. [94] F.Ulaby,Radarmeasurementsofsoilmoisturecontent,IEEETrans.Ant.Prop.,vol.AP-22,pp.257,1974. [95] F.UlabyandP.Batlivala,Optimumradarparametersformappingsoilmoisture,IEEETrans.Geosci.Electron.,vol.GE-14,pp.286,1976. [96] F.Ulaby,P.Batlivala,andM.Dobson,MicrowaveBackscatterDependenceonSurfaceRoughness,SoilMoisture,andSoilTexture:Part1-Baresoil,IEEETrans.Geosci.Electron.,vol.GE-16,no.4,pp.286,October1978. [97] F.Ulaby,J.Cihlar,andR.Moore,Activemicrowavemeasurementsofsoilwatercontent,RemoteSens.Env.,vol.3,pp.185,1974. [98] F.UlabyandM.Dobson,HandbookofRadarScatteringStatisticsforTerrain.Norwood,MA:ArtechHouse,1989. [99] F.UlabyandC.Elachi,RadarPolarimetryforGeoscienceApplications.Norwood,MA:ArtechHouse,1990. [100] F.Ulaby,R.Moore,andA.Fung,MicrowaveRemoteSensing:ActiveandPassive.Vol.I.Boston,MA:ArtechHouse,1981. [101] F.Ulaby,R.More,andA.Fung,MicrowaveRemoteSensing:ActiveandPassive.Vol.II.Boston,MA:ArtechHouse,1982. [102] ,MicrowaveRemoteSensing:ActiveandPassive.Vol.III.Boston,MA:ArtechHouse,1986. 122

PAGE 123

[103] (2012)USDigitalInclinometer.[Online].Available: http://usdigital.com/products/inclinometers/absolute/T7 [104] (2012)VentureManufacturingLinearActuators.[Online].Available: http://www.venturemfgco.com/pdf/RVSO-Techincal-Data.pdf [105] A.Voronovich,Small-slopeapproximationforelectromagneticwavescatteringataroughinterfaceoftwodielectrichalf-spaces,WaveRandomMedia,vol.4,no.3,pp.337,1994. [106] J.WangandB.Choudhury,Remotesensingofsoilmoisturecontentoverbareeldat1.4GHzfrequency,J.Geophys.Res.,vol.86,pp.5277,1981. [107] U.WegmullerandC.Matzler,Roughbaresoilreectivitymodel,IEEETrans.Geosci.RemoteSensing,vol.37,no.3,pp.1391,May1999. [108] J.-P.Wigneron,A.Chanzy,Y.Kerr,H.Lawrence,J.Shi,M.Escorihuela,V.Mironov,A.Mialon,F.Demontoux,P.deRosnay,andK.Saleh-Contell,EvaluatinganimprovedparameterizationofthesoilemissioninL-MEB,IEEETrans.Geosci.RemoteSensing,vol.49,no.4,pp.1177,2011. [109] J.-P.Wigneron,Y.Kerr,P.Waldteufel,K.Saleh,M.-J.Escorihuela,P.Richaume,P.Ferrazzoli,P.deRosnay,R.Gurney,J.-C.Calvet,J.Grant,M.Guglielmetti,B.Hornbuckle,C.Matzler,T.Pellarin,andM.Schwank,L-bandmicrowaveemissionofthebiosphere(L-MEB)model:Descriptionandcalibrationagainstexperimentaldatasetsovercropelds,RemoteSens.Env.,vol.107,no.4,pp.639,2007. [110] J.-P.Wigneron,L.Laguerre,andY.Kerr,SimplemodelingoftheL-bandmicrowaveemissionfromroughagriculturalsoils,IEEETrans.Geosci.RemoteSensing,vol.39,no.8,pp.1697,2001. [111] T.Wu,K.C.J.Shi,andA.Fung,Atransitionmodelforthereectioncoefcientinsurfacescattering,IEEETrans.Geosci.RemoteSensing,vol.39,no.9,pp.2040,2001. [112] T.Wu,K.C.J.Shi,H.Lee,andA.Fung,AstudyofanAIEMmodelforbistaticscatteringfromrandomlyroughsurfaces,IEEETrans.Geosci.RemoteSensing,vol.46,no.9,pp.2584,2008. [113] (2012)YaesuG5500Elevation-over-AzimuthController.[Online].Available: http://www.yaesu.com/ [114] (2012)YaesuGS232AComputerInterface.[Online].Available: http://www.yaesu.com/ [115] H.ZhangandT.Oweis,Water-yieldrelationsandoptimalirrigationschedulingofwheatintheMediterraneanregion,AgriculturalWaterManagement,vol.38,pp.195,1999. 123

PAGE 124

BIOGRAPHICALSKETCH Pang-WeiLiuwasborninHengchun,Taiwan.HereceivedhisB.S.andM.S.fromDepartmentofGeomatics,NationalCheng-KungUniversity,Taiwan,in2001and2003,andaM.S.fromDepartmentofCivilandCoastalEngineering,UniversityofFloridain2009.HismasterresearchinvolvedfeatureextractionfromSARimagesforchangedetectionandevaluationofmicrowavesignalattenuationinforestareasusingLiDARdatasets.Pang-WeistartedhisPhDstudiesintheCenterforRemoteSensing,DepartmentofAgriculturalandBiologicalEngineering,UniversityofFloridafromSpring2010.Hisresearchwasfocusedondevelopingnovelapproachestomodelactiveandpassivemicrowavesignaturesforsoilmoistureretrievalandassimilationunderbaresoilanddynamicvegetation.InMay2013,Pang-WeigraduatedfromUniversityofFloridawithamajorinagriculturalengineeringandaminorinelectricalengineering. 124