简介：Moreaccurateestimationofcropevapotranspiration(ETc)inaregionalscalehasalwaysbeenoneofthemostimportantchallenges.TemporalandspatialmonitoringofETcusingsatelliteimagescanhelptoenhanceaccuracyofestimations.Inthisstudy,the(ETc)ricemapswereproducedbyusingstatistical/experimentalmethodsbasedoncropcoefficient(Kc)mapsderivedfromvegetationindex(VI).Kcwasestimatedusingfourmethods,includinglinearrelationshipbetweenKcandVI(Kc-VI),calibratedmodelofKc-VI,linearrelationshipbetweenKcb(thebasalcropcoefficient)andVI(Kcb-VI),andcalibratedmodelofKcb-VI.TheresultsshowedthatcalibratedmodelofKc-VIhadabetterperformancecomparedtotheothermethods,withnormalizedrootmeansquareerrors(NRMSE),meanabsoluteerrorandrootmeansquareerrorbeing5.7%,0.05mm/dand0.06mm/d,respectively.(ETc)ricemapswereproducedbyusingcalibratedmodelofKc-VIandreferenceevapotranspiration(ET0)fromFAOPenman-Monteithmethod.TheNRMSEwas21.3%forusingFAOPenman-Monteithmethod.Therefore,calibratedKc-VImodelincombiningwithET0basedontheLandsat7ETM+imagescouldbeprovidedagoodestimationof(ETc)riceinregionalscale,andcanbeappliedtoestimatewaterrequirementduetothefreeandfacilitateaccess.

简介：Remote-sensingdataforprotectedareasinnorthernTogo,obtainedinthreedifferentyears(2007,2000,and1987),wereusedtoassessandmapchangesinlandcoverandlanduseforthisdroughtpronezone.Thenormalizeddifferencevegetationindex(NDVI)wasappliedtotheimagestomapchangesinvegetation.Anunsupervisedclassification,followedbyclassesrecoding,filtering,identifications,areacomputingandpost-classificationprocesswereappliedtothecompositeofthethreeyearsofNDVIimages.Maximumlikelihoodclassificationwasappliedtothe2007image(ETM+2007)usingasupervisedclassificationprocess.Sevenvegetationclassesweredefinedfromtrainingdatasets.Thesevenclassesincludedthefollowingbiomes:riparianforest,dryforest,floodedvegetation,woodedsavanna,fallows,parkland,andwater.Fortheseclasses,theoverallaccuracyandtheoverallkappastatisticfortheclassifiedmapwere72.5%and0.67,respectively.Dataanalysesindicatedagreatchangeinlandresources;especiallybetween1987and2000probablyduetotheimpactofdemocratizationprocesssocial,economic,andpoliticaldisorderfrom1990.Wide-scalelossofvegetationoccurredduringthisperiod.However,areasofvegetationclearingandregrowthweremorevisiblebetween2000and2007.Themainsourceofconfusioninthecontingencymatrixwasduetoheterogeneitywithincertainclasses.Itcouldalsobeduetospectralhomogeneityamongtheclasses.Thisresearchprovidesabaselineforfutureecologicallandscaperesearchandforthenextmanagementprograminthearea.

简介：