Evaluation of SEVIRI Rainfall Retrievals over West-Africa using TRMM-PR

In tropical regions, precipitation is a large component of the hydrological cycle. In some regions, such as in the West-African Sahel zone, local economy and ecology are very susceptible to changes in rainfall patterns. For example, less rainfall decreases the soil moisture content, which can finally result in gradual desertification. Despite the large sensitivity to changes in rainfall patterns in these areas, the coverage by rain gauges is rather sparse, thereby hindering accurate rainfall monitoring. The lack of sufficient rain gauges can at least partly be overcome with rainfall estimates from geostationary satellites.

The Spinning Enhanced Visible and Infrared Imager (SEVIRI) onboard the Meteosat-8 and -9 satellites, with its unprecedented 15-minute temporal and 3x3 km2 nadir spatial sampling resolution, may be useful to monitor precipitation over western Africa. Roebeling and Holleman (2009) developed an algorithm to estimate rainfall from retrieved cloud water path, cloud phase and cloud particle effective radius, and showed that the algorithm is capable of retrieving the occurrence and intensity of rainfall accurately over Europe.

In this paper, the accuracy of the SEVIRI rainfall retrieval algorithm over West-Africa is investigated through a comparison with the Tropical Rainfall Measurement Mission Precipitation Radar (TRMM-PR) for the monsoon season of 2006 (May through September). The accuracy is assessed in terms of the occurrence of rainfall using various rain rate threshold values, the intensity of rainfall and the daylight cycle of occurrence and intensity of rainfall using TRMM-PR over West-African land and ocean areas.

The results indicate that SEVIRI rainfall retrievals, taking benefit of their high temporal and spatial resolution, are accurate and very useful for monitoring tropical daytime convection in the tropics. It will be shown that with the progression of the monsoon season, both the observed time of convection onset and maximum convective activity shifts from the late afternoon to the early morning.