Split-window algorithm for land surface temperature estimation from MSG1-SEVIRI data

This letter addresses the land surface temperature (LST) estimation from the data acquired by the spinning enhanced visible and infra-red imager (SEVIRI) on board the first geostationary satellite meteosat second generation (MSG1) using the generalized split-window algorithm proposed by Wan and Dozier (1996). The generalized split-window algorithm was developed for eight view zenith angles (VZAs) by dividing the LST, the average emissivity () and the column water vapour (W) into several sub-ranges to improve the LST estimating accuracy. The simulated results show that the root mean square errors (RMSEs) increase with VZAs and W, and they are less than 1.0K for all sub-ranges with the VZA less than 45, or for the sub-ranges with VZA less than 60 and W less than 3.5cm. The land surface emissivities (LSEs) and W used in the generalized split-window algorithm were estimated from MSG1-SEVIRI data by the method developed by us in previous studies. The results at the four specific locations show that the LSEs were well derived, and the LSTs estimated from MSG1-SEVIRI data are basically consistent with the ones extracted from MODIS/Terra LST products