Abstract: Being used widely in eco-environment study, Land Surface Temperature (LST) is an important parameter for land-air exchange process. Traditional observations and calculations are based on a point, while remote sensing makes it possible to calculate the region LST. At present, most of LST retrieval algorithms are found on Split-Window method with thermal infrared remote sensing data, such as NOAA/AVHRR, MODIS, Landsat TM, and so on. ASTER(Advanced Spaceborne Thermal Emission and Reflection Radiometer) is a comparatively new source of remote sensing data and it is launched in 1999. ASTER has higher spatial and spectral resolution,and provides more detailed information than NOAA/AVHRR, Landsat TM, etc, which helps to improve the precision of LST retrieval. Taking an ASTER remote sensing data of the typical study region at Awati, Xinjiang Province as the data source,based on the methodology of Temperature/Emissivity Separation algorithm, LST is retrieved by using three modules of ADE (Alpha Derived Emissivity), Ratio Method and MMD (Maximum-Minimum Difference). And then, the main error sources of this model are briefly analyzed, which are from theory hypothesis of model, atmosphere effect, emissivity and characteristics of the sensor systems. According to the calculation and analysis, the conclusion is that the advanced Temperature/Emissivity Separation algorithm is effective, and using ASTER data to retrieve LST can obtain more precise results and will have a perspective future.

Key words: ASTER remote sensing data, land surface temperature