An algorithm to retrieve land surface temperature from ASTER thermal band data for agricultural drought monitoring

High spatial resolution ASTER data have 5 thermal bands, of which band 13 and 14 are especially suitable for land surface temperature (LST) estimation. Generally, LST retrieval from two thermal bands is done through so-called split window technique. In the past two decades above 17 split window algorithms have been proposed. However, such algorithm for ASTER data has not been reported, probably due to the new availability of the data for environmental application. In the study, a new split window algorithm has been developed for LST retrieval from ASTER data. Our algorithm only involves two essential parameters for LST retrieval while keeping the same accuracy as those having more parameters. Detailed derivation of the split window algorithm has been given in the paper, which including formulation of thermal radiation transfer equation, determination of algorithm constants, and estimation of the essential parameters. Comparison of our algorithm with the existing ones for validation of its accuracy and applicability in the real world indicates that our algorithm has an average root mean square (RMS) error of 0.67°C when transmittance has an error of 0.05 and emissivity has an error of 0.01. Thus we can conclude that our algorithm is a very good alternative for accurate LST retrieval from ASTER data. Application of the algorithm to Wuxi-Suchou region in Yangtze River Delta produces a very reseasonable LST image of the region, hence confirms the applicability of the algorithm.