A Split-Window Algorithm for Estimating LST From Meteosat 9 Data: Test and Comparison With In Situ Data and MODIS LSTs

The main purpose of this letter is to give an operational algorithm for retrieving the land surface temperature (LST) using the Spinning Enhanced Visible and Infrared Imager data onboard the Meteosat Second Generation (MSG2) satellite. The algorithm is a split-window method using the two thermal infrared channels IR10.8 and IR12.0. The MODTRAN 4.0 code was used to obtain numerical coefficients of the algorithm proposed. The results show that for viewing angles lower than 50^deg the algorithm is capable of producing LST with a standard deviation of 0.7 K and a root-mean-square error (rmse) of 1.3 K. The algorithm has been applied to a series of MSG2 images obtained from an MSG antenna system installed at the imaging processing laboratory (IPL) in the University of Valencia, Valencia, Spain. The LST product has been evaluated using the in situ data from the European Space Agency (ESA) field campaign named CarboEurope, FLEx and Sentinel-2 (CEFLES2) carried out in 2007 in Bordeaux (France), and using the official LST MODIS product and another algorithm to retrieve LST from the MODIS data. This evaluation has been applied over different surfaces and under different viewing angles. The results show an rmse of 1.9 K for the in situ data, 1.5 K for the official LST MODIS product, and 0.7 K compared with that of the MODIS LST algorithm.