Derivation of split-window algorithm to retrieve land surface temperature from MSG-1 thermal infrared data

Abstract

One of the most important parameters in the exchange of energy between the ground and the atmosphere is the land surface temperature (LST). In this work, a split-window algorithm for deriving LST from the first Meteosat Second Generation satellite (MSG-1) data using the two thermal infrared channels IR10.8 and IR12.0 is proposed and validated with in-situ measured temperatures. This algorithm is obtained starting from the radiance transfer equation and requires the knowledge of total atmospheric water vapor content (TAWV) and the surface emissivity (ε10.8 and ε12). First, we have simplified the Planck function to retrieve the radiance at MSG-1 from the temperature. Then, the Roberts model was used to create a relationship between atmospheric transmittance and TAWV. Next, sensitivity analysis of the algorithm indicates that the possible error of viewing angle has relatively insignificant impact on the probable LST estimation error, which is sensible to the possible error of ground emissivity and TAWV. Finally, two methods have been used to validate the proposed algorithm. On the one hand, the results show a root mean square error (RMSE) equals 0.79 K and an average accuracy equals 0.45 K for the comparison with another algorithm. On the other hand, the results show a RMSE equals 2.75 K and an average accuracy equals 1.96 K for the comparison with in-situ measurements. We conclude that the proposed algorithm is able to provide an accurate LST.