Application of split window algorithm to retrieve land surface temperature over northwestern china

Land surface temperature (LST) is a key variable in considerable scientific disciplines and applications, especially for earth observation fields. The goal of this paper is to investigate and validate the application of a split-window algorithm, combining with the estimation of water vapor content in atmosphere, for inversing LST over northwestern China using Moderate-resolution Imaging Spectroradiometer (MODIS) data. In comparison with the MODIS LST product (as a standard), the retrieved LST shows a good agreement with the standard: the retrieval method's LST values of max, min and mean are 304.21 K, 285.09 K and 296.69 K respectively, the LST prodcut's LST values of max, min and mean are 304.54 K, 285.43 K and 295.80 K respectively. Therefore the differences between the retrieval method's LST values of max, min and mean with the LST prodcut's are only 0.33K, 0.34K and 0.89K, which indicates that the LST retrieval method has a believable accuracy, with the mean error of 0.89 K and the RMSE value of 0.91K. ? 2009 SPIE.EI