资源描述:
《净初级生产力遥感估算模型空间尺度转换-王莉雯》由会员上传分享,免费在线阅读,更多相关内容在行业资料-天天文库。
1、1007-4619(2010)06-1074-16JournalofRemoteSensing遥感学报Spatialscalingofnetprimaryproductivitymodelbasedonremotesensing112WANGLiwen,WEIYaxing,NIUZheng1.CollegeofUrbanandEnvironmentScience,LiaoningNormalUniversity,LiaoningDalian116029,China;2.TheStateKeyLaboratoryofRemoteSensingScience,InstituteofR
2、emoteSensingApplications,ChineseAcademyofSciences,Beijing100101,ChinaAbstract:Spatialscalingfornetprimaryproductivity(NPP)referstothetransferringprocessofestablishingquantitativecorrelationbetweensimulatedNPPderivedfromdataatdifferentspatialresolutions.HowtotransferNPPatonescalebythealgorithm
3、withsmallererrortoatanotheristheurgentproblem.NonlinearityandeffectsfromlandcovertypearetwomainproblemsinNPPscaling.Inthispaper,thecontexturalapproachbasedonmixedpixelsandsupportvectormachine(SVM)algorithmareusedtomakethescalingmodelfromthefineresolution(TM)tothecoarseresolution(MODIS).Spatia
4、lscalingfromNPPretrievedfromfineresolutiondatatoNPPderivedfromcoarseresolutionimagesisperformed,andthecorrectionofscaleeffecttoNPPretrievedfromcoarseresolutiondataofMODISisaccomplished.TheresultshowsthatthecorrelationbetweenRj_correctedofthecorrectionfactorforscaleeffectand1-Fmiddledensitygra
5、sslandestimatedbySVMregressionmodelishigher22(R=0.81).Beforethecorrectionforscaleeffect,thecorrelationbetweenNPPMODISandNPPTMislower(R=0.69;RMSE=3.47),2whilethecorrelationbetweenNPPTMandcorrectedNPPMODIS_correctedishigher(R=0.84;RMSE=1.87).Therefore,NPPcorrectedforscaleeffecthasbeengreatlyimp
6、rovedinbothcorrelationanderror.Keywords:netprimaryproductivity,lightuseefficiencymodel,remotesensing,scaling,supportvectormachineCLCnumber:TP761Documentcode:ACitationformat:WangLW,WeiYXandNiuZ.2010.Spatialscalingofnetprimaryproductivitymodelbasedonremotesensing.JournalofRemoteSensing.14(6):10
7、74—10891INTRODUCTIONofvariouskindsofspectralsignalsfromlandsurfaceobjectsreceivedbysensors.Inaddition,thequantitativecorrelationbetweenmixedspectralsignalsandlandsurfaceparametersisTheadvantageofremotesensingistoprovidelandsurfacedifficulttob