Abstract
Improving the simulation of the West African Monsoon (WAM) system is paramount to increasing confidence in the projections of the region’s monsoon rainfall change. This work aims to thoroughly analyze the representation of the WAM system in two state-of-the-art, high-resolution (~ 25 km) regional climate models (RCMs) in order to highlight the causes of models’ biases through a process-oriented evaluation approach. Model results generally feature a north (Sahel) - south (Guinea Coast) dipole-like rainfall bias, although, sometimes, positive or negative rainfall biases are evident almost over the whole of West Africa. Our analysis shows that biases in the sea- and land-surface temperature on the one hand, and biases in the sea-level and land-surface pressure, on the other hand, lead to biases in the simulated temperature and pressure contrasts between the west African landmass and the eastern Atlantic ocean. As a consequence, biases appear in the modeled monsoon flow strength, which, in turn, lead to errors in the amount of advected moisture in the interior of the continent via southwesterlies and the West African westerly jet (WAWJ) on the one hand, and the extent of deepening of the monsoon flux inland on the other hand. In addition, the African easterly jet (AEJ) is underestimated, inducing an underestimation of the African easterly waves (AEWs) activity and a weakening of the cyclonic convective circulation resulting from the AEWs’ troughs, leading to a decrease in the southwesterly flow feeding mesoscale convective systems (MCSs) embedded within the AEJ. The modeled equatorward or northward shifting of the AEJ is likewise found to contribute to the models' wet or dry biases over the Sahel. Finally, there is no consistency between models and reanalyses on the one hand, and between RCM experiments on the other hand, in the way, the simulated atmospheric instability/stability modulates the convection, especially over the Sahel.
DOI Link to Publication