引用格式：曹煜,方秀琴,杨露露,等.基于随机森林的西辽河流域CCI土壤湿度降尺度研究[J].地球科学信息学报,2023,25(8):1669-1681.[CaoY,FangXQ,YangLL,etal.DownscalingofCCIsoilmoistureintheXiliaoheBasinbasedonrandomforest[J].JournalofGeo-informa-tionScience,2023,25(8):1669-1681.]DOI:10.12082/dqxxkx.2023.221013基于随机森林的西辽河流域CCI土壤湿度降尺度研究曹煜，方秀琴*，杨露露，蒋心远，廖美玉，任立良河海大学水文水资源学院，南京210098DownscalingofCCISoilMoistureintheXiliaoheRiverBasinbasedonRandomForestCAOYu,FANGXiuqin*,YANGLulu,JIANGXinyuan,LIAOMeiyu,RENLiliangCollegeofHydrologyandWaterResources,HohaiUniversity,Nanjing210098,ChinaAbstract:Soilmoistureisakeyfactorintheclimatesystemandhasimportantapplicationinagriculturalmanagement,waterresourcemanagement,andecosystemmonitoringandassessment.Althoughremotesensing-derivedsoilmoistureproductscanprovidesoilmoisturedistributiononalargescale,theyusuallyhavecoarsespatialresolution,makingitdifficulttomeettherequirementsofpracticalapplications.Thus,downscalingofremotesensing-derivedsoilmoistureproductshasbecomeoneofthehottopicsrecently.Inthispaper,theESACCIdailysoilmoistureat0.25°resolutionisusedasthemaindatasource,combinedwiththeMODISunderlyingsurfacedata,topographicdata,meteorologicaldata,andotherenvironmentalfactorsat1kmresolution.ArandomforestdownscalingmodelisconstructedtogeneratedownscaledCCIdailysoilmoistureproductsat1kmresolutionintheXiliaoheRiverBasinofChinafrom2013to2020.Resultsshowthat:(1)Theanalysisoftheimportancescoreofenvironmentalfactorsshowsthatrelativehumidityanddaytimesurfacetemperaturearethetwomostimportantfactorsinfluencingsoilmoisture,followedbytopographyandlocationfactors;(2)Theperformanceoftherandomforestdownscalingmodelisverifiedbyusingthemeasureddataof22stationswithinthestudyarea,andtheresultsshowthatthedownscaledresultsconsideringmultipleenvironmentalfactors(underlyingsurface,topography,andmeteorology)aremoreaccuratethanthatconsideringonlysurfaceelements.TheRMSEofeachsiteisbelow0.0488m3/m3,theaveragecorrelation...
