Effect of atmospheric absorption and surface emissivity on the determination of land surface temperature from infrared satellite data

Abstract Infrared satellite data can be used to determine land surface temperatures but they have to be corrected from atmospheric absorption and soil emissivity. To correct for atmospheric absorption, we present a method that involves the LOWTRAN-6 radiative transfer model and Chedin and Scott's Improved Initialization Inversion procedure to retrieve atmospheric profiles of temperature and moisture. To correct for soil emissivity, no method is proposed, but the emissivity effect is thoroughly investigated. We show, by comparing surface temperatures derived in the two NOAA-9 Advanced Very High Resolution Radiometer thermal infrared channels, that the soil emissivity presents definite spectral variations. Consequently, the split-window techniques currently applied to retrieve surface temperature over the ocean cannot be readily applied over land.