Characterization of the Rainy Season in Central Africa by Using a Regional Climatic Model

ABSTRACT Central Africa, like most regions of the planet, is suffering the consequences of climate change, including the disruption of seasonal indices such as the Rainfall Onset Dates (RODs) and Rainfall Cessation Dates (RCDs). The inability to predict and manage these climatic events is one of the factors contributing to the slowdown in economic activities as well as famine in this region. Aware of these challenges, we used the regional climate model RegCM5 to analyze these dates in sub‐regions with homogeneous climatic characteristics in Central Africa, namely the Sahel (Sa), the Cameroon Highlands (CH), and the Congo Basin (CB). We also used observational data from the Climate Hazards Group InfraRed Precipitation with Station data (CHIRPS), along with uncorrected and corrected data from six convective schemes of RegCM5. These schemes include the uncorrected () and corrected () Kain‐Fritsch scheme, the uncorrected () and corrected () MIT‐Emanuel scheme, the uncorrected () and corrected () Grell scheme with Fritsch and Chappell closure, the uncorrected () and corrected () Grell scheme with Arakawa and Schubert closure, the uncorrected () and corrected () modified Kuo scheme, and the uncorrected () and corrected () Tiedtke scheme. We found that, except in the Cameroon Highlands (CH), the different model configurations reproduce the annual cycles of precipitation, RODs, RCDs, and the migration of the Intertropical Convergence Zone (ITCZ) from the ocean to the continent in Central Africa. However, these performances depend on the specific index being evaluated (onset or cessation), the study area, the convective scheme used, and whether the data were corrected or uncorrected. For example, the corrected data better identify the RODs ( performs best in the CH, in the Sa, and better identifies the first and second seasons in the CB). In contrast, for the cessation dates, uncorrected data perform better ( in the Sa, and and better identify the RCDs of the second and first seasons in the CB, respectively). We also note that the RCDs are much better represented by the convective schemes of RegCM5 than the RODs. This study suggests that all the factors discussed must be taken into account when using RegCM5 to determine RODs and RCDs in Central Africa. Furthermore, if the goal is solely to represent the cessation dates, the Emanuel convective scheme can be used.