Using spaceborne surface soil moisture to constrain satellite precipitation estimates over West Africa

International audience This paper describes a methodology to use the passive microwave measurements of the 6.9 GHz bandwidth of the AMSR-E sensor which is the most sensitive to surface soil moisture, to constrain satellite-based rainfall estimates over a semi arid region in West-Africa. The paper focuses on the aptitude of AMSR-E measurements to inform if rain occurs or not. The study was conducted over a 125x100 km² region located in Niger where a dense recording raingauge network is available to build an accurate ground-based 3-hour rainfall product at the 25x20 km² resolution. A satellite-based rainfall product (EPSAT-SG), based on both infrared and microwave measurements, was compared to the ground-based rainfall product. It was shown that EPSAT-SG overestimates by about 30 % the total number of rainy events during the 2004 and 2006 rainy seasons. A simple methodology based on the AMSR-E polarization ratio variations related to the surface soil moisture leaded to suppress a large amount of the wrong rainfall events.