Abstract: The water body of Ebinur Lake is extracted by using normalized difference water index (NDWI),modified normalized difference water index (MNDWI),automatic water extraction index (AWEI) and modified automated water extraction index (MAWEI) with Landsat 5/7/8 series remote sensing images as the data source.The extraction results of the four indexes are compared and analyzed by establishing confusion matrix.MAWEI index is proved to be of the highest precision (over 95%) with the Kappa coefficient greater than 0.94.Furthermore,the area of Ebinur Lake is extracted by using the MAWEI index,and the change of lake area is analyzed through the dynamic degree.From 2000 to 2003,the area of Ebinur Lake increased from 768.93 km² to 982.27 km²,up by 27.74%,and the dynamic degree was 9.24%;from 2003 to 2014,the lake area shrank to 447.08 km²,down by 54.48% with the dynamic degree of -4.95%;from 2014 to 2018,the lake expanded again to 852.77 km²,rising by 90.74% with the dynamic degree reaching 22.68%;and from 2018 to 2020,the lake area reduced again to 593.79 km²,falling by 30.36% with the dynamic degree of -15.18%.In the past two decades,the area of Ebinur Lake displayed a trend of expansion first and then retreat in general.The change of lake area is positively correlated with annual precipitation and the runoff of Bortala River and Jinghe River,while negatively with evaporation.The reduction of water resources as well as the surge of socio-economic water use caused by the steady growth of population,irrigation area,and GDP also have impacts on the change of lake area.

Key words: lake area, water index, MAWEI, driving factors, confusion matrix, ecological environment, Ebinur Lake