Estimating of total atmospheric water vapor content from MSG1-SEVIRI observations

Abstract

In this work, we proposed a methodology to estimate total atmospheric water vapor content (TAWV) from observations of the Spinning Enhanced Visible and Infrared Imager (SEVIRI) on board the first Meteosat Second Generation satellite (MSG1). The method used is called the split-window technique which requires only the data from the channels IR10.8 and IR12, therefore this method not requires any ancillary data. This method is based on the MSG1 observations of the same geographic location over land at two slightly different times during a period when the ground temperature is changing rapidly. The main contribution of the present work is to consider that the relationship between TAWV and the ratio of the two split-window channel transmittances (τ12/τ10.8) is a quadratic formula, this assumption is based on the "Roberts" approach simulations using MSG1-SEVIRI filter response functions for a 2311 atmospheric situations from the TIGR dataset. For validation, we have examined the accuracy of the TAWV estimated in this work by comparison with the data obtained from radiosonde and from aerosol robotic network (AERONET). On the one hand, the comparison with the radiosonde data show that the root mean square error (RMSE) equals 0.66 g cm−2, the standard deviation (SD) equals 0.59 g cm−2 and the correlation coefficient (R) equals 0.79. On the other hand, the comparison with the AERONET data show that the RMSE equals 0.42 g cm−2, the SD equals 0.29 g cm−2 and the R equals 0.82. Also, the comparison with another method demonstrates that the spatial variation of TAWV here is reasonable. We have concluded in this study that the TAWV can be determined from the MSG1-SEVIRI observations with accuracy acceptable which can be used for climate change research.