Modeling Surface Roughness to Estimate Surface Moisture Using Radarsat-2 Quad Polarimetric SAR Data

摘要

Microwavebackscatteringfromtheearth'ssurfacedependsonseveralparameterssuchassurfaceroughnessanddielectricconstantofsurfacematerials.Thetwoparametersrelatedtowatercontentandporosityarecrucialforestimatingsoilmoisture.Thesoilmoistureisanimportantparameterforecologicalstudyandalsoafactortomaintainenergybalanceoflandsurfaceandatmosphere.Directroughnessmeasurementstoalargearearequireextratimeandcost.Heterogeneityroughnessscaleforsomeapplicationssuchashydrology,climate,andecologyisaproblemwhichcouldleadtoinaccuraciesofmodeling.Inthisstudy,wemodeledsurfaceroughnessusingRadasat-2quadPolarimetricSyntheticApertureRadar(PolSAR)data.Thestatisticalapproachestofieldroughnessmeasurementswereusedtogenerateanappropriateroughnessmodel.ThismodelingusesaphysicalSARapproachtopredictsradarbackscatteringcoefficientintheparameterofradarconfiguration(wavelength,polarization,andincidenceangle)andsoilparameters(surfaceroughnessanddielectricconstant).SurfaceroughnessvalueiscalculatedusingamodifiedCampbellandShepardmodelin1996.Themodificationwasappliedbyincorporatingthebackscatteringcoefficient(σ°)ofquadpolarizationHH,HVandVV.ToobtainempiricalsurfaceroughnessmodelfromSARbackscatteringintensity,weusedforty-fivesamplepointsfromfieldroughnessmeasurements.WeselectedpaddyfieldinIndramayudistrict,WestJava,Indonesiaasthestudyarea.ThisareawasselectedduetointensivedecreasingofriceproductivityintheNorthernCoastregionofWestJava.ThirddegreepolynomialisthemostsuitabledatafittingwithcoefficientofdeterminationR2andRMSEareabout0.82and1.18cm,respectively.Therefore,thismodelisusedasbasistogeneratethemapofsurfaceroughness...