Testing satellite rainfall estimates for yield simulation of a rainfed cereal in West Africa

In West Africa, the spatiotemporal variability of rainfall is often considered as one of the most important causes of food insecurity. Satellite Rainfall Estimates products provide synoptic, objective and timely information on the rainfall distribution and thus are particularly relevant tools for crop monitoring over large areas. This study assesses the accuracy of the TRMM3B2 daily product for characterizing seasonal rainfall distribution and its effects when modelling crop cereals in a South West Niger site. We found that the TRMM3B42 product was not able to accurately detect rainfall temporal distribution at station level, leading to a high dispersion of yield (RMSE = 231 kg ha-1) at station level when compared to the yield simulated with measured rainfall. However the bias is low (ME = 32 kg ha-1), resulting to a very good yield estimation when aggregated at the site level.