Simulating extreme temperatures over Central Africa by RegCM4.4 regional climate model

Nowadays, extreme temperatures are associated with global warming and cause great concerns worldwide, particularly in Africa because of low resilience, adaptation and mitigation strategies. The present study investigates the performance of RegCM4 climate model in reproducing extreme temperatures over Central Africa (CA) during 2002-2006 period. Apart of daily minimum and maximum temperatures, the study uses six extreme temperature indices based on the recommendations of Expert Team on Climate Change Detection and Indices. These include four intensity indices and two duration indices. Focus has been done in two sub-regions embedded over the study area for more specific analysis. The first sub-region named zone 1 is semi-arid while the second one named zone 2 is humid. We found that whatever the region or season over CA, RegCM4 displays fairly well the intensity indices associated to daily minimum temperatures, namely TNN and TNX, with underestimations not exceeding 4 degrees C. For intensity indices based on daily maximum temperatures, namely TXX and TXN, the warm bias can reach up to 10 degrees C. This model behavior is associated to its inability to depict surface downward solar and thermal fluxes as well as surface sensible heat flux. Regarding the duration indices, RegCM depicts the hot sequences with high biases, unlike the cold sequences. Statistical parameters used for evaluation show that despite the differences between model and ERAS, RegCM4 can be used to investigate extreme temperatures over the study region mainly if indices of interest are based on daily minimum temperatures and cold sequences.