Verification of Operational Numerical Weather Prediction Model Forecasts of Precipitation Using Satellite Rainfall Estimates over Africa

Rainfall is an important variable to be able to monitor and forecast across Africa, due to its impact on agriculture, food security and a number of climate related diseases. Numerical weather models are an important component of this work, due to their complete spatial coverage, high resolution and their link to forecast products. In this study, the European Centre for Medium-Range Weather Forecasts (ECMWF) and the National Centers for Environmental Prediction-Global Forecast System (NCEP-GFS) are verified by Rainfall Estimates 2.0 (RFE2). We found that the two forecast rainfall products show good consistency with the satellite products in spatial distribution over Africa on the seasonal timescale. However, the metrics (i.e. Hit rate, Threat score, Probability of Detection, False Alarm Ratio) show a high spatial difference over Africa for daily predictions. Both NWP models tend to underestimate the rainfall amount in western and central Africa, while they overestimate the rainfall amount in eastern and southern Africa. However, the ECMWF is better than the GFS to predict both rainfall amount and rainfall events. Due to complex rainfall features and topography, the largest bias occurs over and around western Africa. The predictability of daily rainfall using global NWP is also discussed.