Abstract: Evapotranspiration is estimated based on NDVI,surface net radiation,air temperature by using China monthly meteorological data and MODIS remote sensing data which include evapotranspiration data(MOD16) and normalized difference vegetation index data(MOD13).Simplify and improve the parameters of mixed linear dual source remote sensing evapotranspiration model to estimate the surface evapotranspiration in oasis region of southern Xinjiang,China.Analysis of spatial and temporal variation characteristics of the actual-evapotranspiration (ET) and potentialevapotranspiration(PET) from 2001 to 2014.Besides,based on the correlation analysis of measured evaporation data from meteorological stations in southern Xinjiang,the applicability of MOD16 products in oasis region is already credible by means of statistical test.According to the evapotranspiration drought index (EDI),the distribution characteristics of drought in oasis of southern Xinjiang is analyzed,and the degree of drought is determined by the EDI anomaly.The results reflect that the annual mean evapotranspiration from 2001 to 2014 shows little change,and these are significant difference among the four seasons.The highest evapotranspiration is in summer and the lowest in winter.The spatial trend of ET and PET is in opposite situation.In addition,there is a huge disparity between the annual value of ET and PET which means the surface water shortage is serious in oasis.Furthermore,the average annual value of EDI index in oasis area is fairly large,the response of EDI anomaly to drought and the judgment of drought degree are reliable.From the perspective of energy and waterbalance,this research simplifies the model parameters and enhances the applicability of the model.It has certain guiding significance for large scale areas evapotranspiration estimation and drought monitoring.

Key words: southern Xinjiang, oasis, surface evapotranspiration, evapotranspiration drought index(EDI), spatial and temporal variation characteristics