Remote sensed-based rainfall estimations over the East and West Africa regions for disaster risk management

Abstract

Satellite precipitation products are very useful as proxies for gauged observations, especially in Africa, where there are sparse in-situ precipitation-measuring weather stations. The two most commonly used satellite precipitation products in Africa (Climate Hazards Group Infrared Precipitation with Stations (CHIRPS) and Tropical Applications of Meteorology Using Satellite and Ground-based Observations (TAMSAT)) and the relatively new Integrated Multi-Satellite Retrievals for GPM early run (IMERG-ER) of the Global Precipitation Measurement Mission (GPM) with soil moisture (GPM+SM2RAIN) are evaluated in comparison to the in-situ observations from the Trans-African Hydro-Meteorological Observatory (TAHMO) project on daily, monthly and annual scales in West (Ghana) and East (Uganda) Africa. Statistical verification methods such as continuous and categorical methods were examined. For both countries, the results showed that all the products showed weak performance in capturing daily precipitation. For the monthly and seasonal scales, all the satellite product performances were very good. Also, small precipitation thresholds (2–5 mm/day) were relatively better detected by all the satellite products. All the products failed to detect intermediate (10–20 mm/day) and large threshold precipitation (>30 mm/day). When the products were examined for disaster management applications over a select number of stations in Ghana and Uganda, the results showed inconsistent and poor performances. Using all the metrics, both CHIRPS and TAMSAT showed similar performances, confirming that their common underlying algorithms are bias-adjusted, whereas GPM+SM2RAIN was different because of its independence from in-situ observations. All the products performed better in Ghana compared to Uganda, which may be due to the latter’s complex terrain. It is possible that CHIRPS and TAMSAT may be better suited for hydrological applications like seasonal forecasting, irrigation scheduling, reservoir operations than for daily or sub-daily processes like flash floods. While all these satellite products have their strengths and weaknesses for various environmental applications in Africa, there is a need for a central platform in Africa for a better or improved precipitation analysis based on the assimilation of climate forecasts, in-situ observations, radar, and satellite observations. This will provide higher temporal (3 hr–6 hr) and spatial (~2 km) resolutions. This study is applicable for scientists, water managers, and decision-makers involved in disaster risk management studies in Africa.