Challenges of satellite rainfall estimation over mountainous and arid parts of east Africa

Abstract

Different satellite rainfall products are used in different applications over different parts of the world. These products are particularly important over many parts of Africa, where they are used to augment the very sparse rain-gauge network. However, the quality of the different satellite products varies from one product to another and from one climatic region to another. The climate over eastern Africa varies from wet coastal and mountainous regions to dry arid regions. Significant variations could be observed within short distances. The different climates will pose different challenges to satellite rainfall retrieval over this region. This study explores the effect of mountainous and arid climates on four different satellite rainfall-estimation (RFE) algorithms. The mountainous climate is located over the Ethiopian highlands, while the arid region covers parts of Ethiopia, Djibouti and Somalia. One infrared-only product, African rainfall climatology (ARC), one passive-microwave-only product (the Climate Prediction Center morphing technique, CMORPH) and two products (the RFE algorithm and the tropical rainfall measuring mission (TRMM-3B42)), which combine both infrared and passive-microwave estimates, are used for this investigation. All the products exhibit moderate underestimation of rainfall over the highlands of Ethiopia, while the overestimation over the dry region is found to be very high. The underestimation over the mountainous region is ascribed to the warm orographic rain process, while the overestimation over the dry region may be because of sub-cloud evaporation. Local calibration of satellite algorithms and merging of satellite estimates with all locally available rain-gauge observations are some of the approaches that could be employed to alleviate these problems.