The Hydroclimate Variability of Central Africa: seasonal cycle, mechanisms, teleconnections and impacts on neighbouring regions

Abstract Central Africa is, climatologically speaking, a poorly studied region (Clivar, 2000; Dezfuli and Nicholson, 2012; Nicholson and Dezfuli, 2012; Todd and Washington, 2004). It is considered as a knowledge gap in the understanding of the tropical climate system (Todd and Washington, 2004). Drivers of Central Africa rainfall are not well documented and deserve more attention. The aims of thesis are to enhance our fundamental understanding of Central Africa rainfall and the mechanisms involved in its seasonal and interannual variability as well as to assess how an atmospheric general circulation model forced by observed sea surface temperature (SST), the ECHAM5.3 model, does represent the main features of Central Africa hydroclimate variability. The seasonal cycle of Central Africa rainfall is primarily driven by change in the atmospheric low-pressure system of Central Africa landmass, water vapor and latent heat release rather than change of local temperature. From October to April, over Central Africa and its neighbouring regions, we highlight the existence in the mid-lower troposphere, between 1000 and 500 hPa of a dominant cyclonic and quasipermanent circulation pattern that drives the atmospheric large-scale circulation and its associated water vapor transports, namely the Central Africa Low. The Central Africa Low, with its variation strongly modulated by El Niño Southern Oscillations (ENSO), is characterized by strong convective activity due to an unstable atmosphere over central Africa, leading to high rainfall with less variance. Nevertheless, when the Central Africa Low prevails, Central Africa is a sink of water vapor, with the Indian Ocean as the main supplier. The weakening of the Central Africa Low, in May to September, is associated with the reversal of the water vapor transport at the northern boundary channel, leading Central Africa to become a source of moisture. During this season, both surrounding oceans are suppliers of moisture, with some additional contribution from the Congo basin rainforest. Central Africa rainfall variability is controlled by large-scale circulation variation, rather than variation in tropospheric water vapor. Year-round, the large-scale circulation is characterized by dominant easterly jets at middle (African easterly jets, AEJs) and upper (tropical easterly jets, TEJ) levels, owed by the Central Africa Low. At low-levels, there is a shallow zonal overturning circulation thermally direct, namely the Congo Basin Cell, driven by near-surface land-ocean thermal contrast between the warm central Africa landmass and the relatively cold Atlantic Ocean. The Congo Basin Cell, characterizes by eastward flow, persists year-round, with a maximum strength (-196.92±32.89 Sv) and width (30o degree) in August/September and minimum strength (-24.80± 17.83 Sv) and width (~6o degree) in May. The Congo Basin Cell does not play any crucial role in modulating Central Africa rainfall but it does regulate the rainfall distribution, through the seasonal position of the ITCZ. At midlevel, the atmospheric convective instability over Central Africa is controlled by the southward import of high moist static energy from the warmer Sahel associated with the AEJ over Central Africa. The saturation of the rising moist air at midlevel determines the location of high rainfall over central Africa year-round. Nevertheless, the absence of significant trend (- 0.013 mm per decade) of the Central Africa rainfall is associated with the weakening of the Central Africa Low in recent decades (1979 to 2015), consistent with Lau and Wu (2006). Further investigations on 8 physical mechanisms affecting the Central Africa hydroclimate reveals that the Central Africa Low and land-ocean thermal contrasts are the main drivers of Central Africa rainfall variability at seasonal and interannual time scale, through the control of AEJs and the Congo Basin Cell strength and width. The analysis of ECHAM5.3 experiments provide a support to these mechanisms. Finally, to unravel what are the physical mechanisms shaping the rainfall anomalies patterns associated with the interannual variability of Central Africa rainfall, we found out that the Central Africa does reflect the regional-scale response of the atmosphere to the variation of the interbasin SST anomalies gradient (ΔSST) between tropical Atlantic and Indian Oceans. Likely, the zonal contrast of central Africa rainfall is owed by the Central Africa Low, which separates central Africa in two distinct regions of opposite polarity by regulating the strength of the low-level westerly and mid-upper easterly jets and their associated water vapor transports. This east-west dipole-like pattern of Central Africa rainfall is similar to the second leading mode obtained by empirical orthogonal functions (EOF) analysis of rainfall anomalies during the long rainy season. Thus, during the positive phase of ΔSST, the Central Africa Low area change induces an anomalous clockwise zonal overturning cell over Central Africa, with ascending branch over Atlantic, indicative of deep convection leading to rainfall surplus, and sinking branch over Indian Ocean, indicative of subsistence, which suppress convection and lead to rainfall deficit, consistent with the mechanism proposed by Dezfuli et al. (2015). However, the impact of ΔSST on Central Africa rainfall variability is asymmetrical during positive and negative phases of ΔSST