On Sahelian-Sudan rainfall and its moisture sources

Sammanfattning: The African Sahel is one of the most vulnerable regions to climate variability at different time scales. It is an arid to semi-arid region with limited water resources. The summer rainfall is one of these sources, but it exhibits pronounced interannual variability. This thesis presents several aspects of Sahelian Sudan rainfall. Sudan is located at the eastern fringe of the Sahel and its least studied part. We have examined the impact of tropical deforestation on the rainfall, the moisture sources of the region and the temporal characteristics of the observed and modeled rainfall. In a sensitivity study we performed three simulations, one control simulation and then setting the surface condition of South Sudan to either grass or desert conditions. The rainfall was reduced by 0.1 − 0.9 in the grass scenario and by 0.1 − 2.1 mm day−1 (hereafter mm d−1) in the desert scenario. These changes also propagated northward into Sahelian Sudan, indicating a remote impact. The total moisture convergence into Sahelian Sudan was reduced by 11.5% and 21.9% for grass and desert conditions, respectively. The change in moisture convergence into the region motivated a comprehensive analysis of the moisture sources for the region. Two different modeling approaches, Lagrangian and Eulerian, were applied to identify the moisture sources and quantify their contributions to the total annual rainfall budget. The analysis shows that atmospheric flows associated with the Inter-Tropical Convergence Zone (ITCZ), e.g. from Guinea Coast, Central African and Western Sahel, brings about 40% − 50% of the annual moisture supply, while local evaporation adds about 20%. The rest of the moisture comes from the Mediterranean, Arabian Peninsula and the Southern part of the Indian Ocean. While there were differences in the details between the results from the two modeling approaches, they agree on the larger scale results. In an attempt to characterize the temporal character of the rainfall, observed and modeled daily rainfall from different regional climate models was classified into five categories: weak (0.1 −1.0), moderate (>1.0 − 10.0), moderately strong (>10.0 − 20.0), strong (>20.0 − 30.0), and very strong (>30.0) mm d−1. We found that most rain-days were in the weak to moderate rainfall categories, accounting for 60% − 75%. Days that have strong rainfall represent about 6% of the total rain-days, yet they represent about 28% − 48% of the total amount of the annual rainfall. Regional climate models fail to produce the strong rainfall, instead most of the modeled rain-days are in the moderate category and consequently the models overestimated the number of rain-days per year.