Evaluation and projected changes in daily rainfall characteristics over Central Africa based on a multi-model ensemble mean of CMIP5 simulations

Abstract

This study uses daily rainfall data from 20 global climate models (GCMs) simulations, participating in the phase 5 of the Coupled Model Intercomparison Project (CMIP5) and eight daily rainfall indices defined by the Expert Team on Climate Change Detection and Indices (ETCCDI), to investigate the changes in extreme weather conditions over Central Africa under the representative concentration pathway 8.5. The performance of the multi-model ensemble (MME) mean which in fact refers to the best performing models selected through the Taylor diagram analysis was evaluated by comparing with two gridded daily observation datasets during the historical period (1998–2005). Results show that although some uncertainties may exist between the gridded observation datasets, MME consistently outperform individual models and reasonably reproduced the observed pattern of daily rainfall indices over the region, except in the case of consecutive wet day (CWD) where the high variability of individual members has resulted in the degradation of the overall skill of the MME. The assessment of the climate change signal in the eight daily rainfall indices was done for the mid and late twenty-first century (2026–2056 and 2066–2095 respectively), relative to the baseline historical time period (1976–2005). We found a significant increase in the total wet day rainfall amount (PRCPTOT) over southern (northern) Central Africa from December to February (from September to November). This is mainly due to the increase of high intense rainfall events rather than their frequency. The results also reveal that the increase in PRCPTOT was coupled with increase in the maximum consecutive 5-day rainfall amount (RX5DAY), the 95th percentile (R95), and the total wet day rainfall amount above the 95th percentile (R95PTOT), with more robust patterns of change at the late twenty-first century. The increase in extreme rainfall events (RX5DAY, R95, and R95PTOT) is likely to increase flood risks over Cameroon, Central African Republic, Gabon, Congo, Angola, Zambia, and Democratic Republic of Congo. On the other hand, changes in CWD and PRCPTOT are projected to significantly decrease over Angola, Zambia, and Democratic Republic of Congo from September to November. This is due to a substantial increase of zonal moisture divergence fluxes in upper atmospheric layers. The analysis has also shown that areas where CWD and PRCPTOT decreases coincides with those where consecutive dry days (CDD) increase. The decrease in CWD and PRCPTOT coupled with the increase in CDD could worsen drought risk and significantly disrupt priority socio-economic sectors for development such as rain-fed agriculture, hydroelectric power generation, and water resource availability. The results thus underline the importance for decision-makers to seriously consider adaptation and mitigation measures, in order to limit the risks of natural disasters such as severe droughts and floods that Central African countries may suffer in the future.