Estimation of precipitation over tropical Africa using TRMM Rain Products and Meteosat Data

The purpose of this study is to compare and validate the rainfall estimates obtained from different satellite-based methods, at various space and time scales over Tropical Africa.

Data and products of the Tropical Rainfall Measuring Mission (TRMM) satellite are used together with the Meteosat satellite data, for the period corresponding to the rainy season over West Africa (June to September 1998). A validation site near Niamey (13.5°N, 2°E), instrumented with a dense rain-gauge network, provides the reference precipitation fields, obtained by kriging the ground data, for comparison with the satellite-based estimations.

The TRMM products used in this inter-comparison study are the TRMM Microwave Imager (TMI) surface precipitation product (G2A12), mapped to a 0.5° by 0.5° longitude/latitude grid, and the TMI/PR (Precipitation Radar) combined rainfall product (G2B31), mapped to a 0.1° by 0.1° longitude/latitude grid. The multi-satellite rainfall estimations are obtained using the RACC (Rain And Cloud Classification) method. This algorithm combines the Meteosat thermal infrared and water vapour channels data and the TRMM rainfall products, in the learning phase of the classification procedure. In the application phase, it takes advantage of the good time and space resolution of the Meteosat half-hourly images, to produce cumulated rain estimations at different time scales.

The results of the inter-comparison of the different rainfall estimations are presented and the improvement of the multi-satellite rainfall retrievals using TRMM data, over a tropical African validation site, is discussed.