Snow cover retrieval over Italy and alpine regions using MSG data for climatologic purposes

An automated procedure that monitors the snow cover over Italy on a monthly and annual basis, based upon daily snow cover maps, is presented here. The snow retrieval has been performed using the reflectance in the 0.6 and 1.6 µm channels and brightness temperature in the 12.0 µm channel of Meteosat‐8. The pixels are divided into cloudy, snowy and snow free. A cloudy pixel is obtained when that pixel has been cloudy all day, making snow retrieval impossible. The first phase of the procedure produces temporary daily snow cover maps. In the second phase, the daily maps are analysed for cloud cover correction based on simple statistical hypotheses. In this way some cloudy pixels (about 10% of the total) are processed and labelled as snow covered or snow‐free. Finally, daily maps are further processed in order to obtain a monthly or annual map representing the daily snow cover frequency for each pixel. The procedure performance has been tested by comparison with six months Global Telecommunication System (GTS) ground observations. Contingency tables and statistical indices have been analysed. The verification procedure is also divided into three phases: the first one compares the temporary daily maps with the observations; the second one compares the daily maps, after cloud correction, with the observations, and the third one tests the monthly maps' performances, by mean differences and the root‐mean‐square differences between ground station and satellite frequencies. In the third verification phase the effect of cloud cover on the monthly maps is also assessed. Copyright © 2009 Royal Meteorological Society