Abstract:In two recently spilt window algorithms for land surface temperature (LST) inversion,on the basis of Sobrino et al (2003) for LST inversion and the estimation of surface emissivity by NDVI (Normalized Differential Vegetation Index) threshold method,combined with the algorithm for calculating atmosphere water vapor content in method of Mao et al (2005),the LST of Shanxi province was retrieved on three dates,05/28/2006,10/09/2006 and 11/14/2006.The LST distribution maps indicated that the retrieved results were similar to LST products,which were used as a standard to validate this method.According to the validating result of one temporal data on 10/09/2006,the errors of maximum value,minimum value and average value between retrieved LST and LST product were 1.08℃、0.16℃ and 0.75℃,respectively,and a very good agreement between retrieved LST and LST product was observed (the Root Mean Square Error was 1.30℃),which all demonstrated the high accuracy of this retrieval algorithm.Therefore,by using this simplified method,the LST in large scale could be obtained quickly and accurately,without depending on the ground truth data,which provide a convenient path to monitor agricultural environment,especially for drought monitoring.

Application of MODIS data to retrieve land surface temperature