Future precipitation projections over central and southern Africa and the adjacent Indian Ocean: what causes the changes and the uncertainty?

Future projections of precipitation at regional scales are vital to inform climate change adaptation activities. Therefore, is it important to quantify projected changes and associated uncertainty, and understand model processes responsible. This paper addresses these challenges for Southern Africa and adjacent Indian Ocean focusing on the local wet season. Precipitation projections for the end of the 21st century indicate a pronounced dipole pattern in the CMIP5 multi-model mean. The dipole indicates future wetting (drying) to the north (south) of the climatological axis of maximum rainfall, implying a northward shift of the ITCZ and South Indian Ocean Convergence Zone, and therefore not consistent with a simple ‘wet-get-wetter’ pattern. This pattern is most pronounced in early Austral summer suggesting a later and shorter wet season over much of southern Africa. Using a decomposition method we determine physical mechanisms underlying this dipole pattern of projected change, and the associated inter-model uncertainty. The projected dipole pattern is largely associated with the dynamical component of change indicative of shifts in the location of convection. Over the Indian Ocean, this apparent northward shift in the ITCZ may reflect the response to changes in the north-south SST gradient over the Indian Ocean, consistent with a ‘warmest-get-wetter’ mechanism. Over land subtropical drying is relatively robust, particularly in the early wet season. This has contributions from dynamical shifts in location of convection, which may be related to regional SST structures in the Southern Indian Ocean, and the thermodynamic decline in relative humidity. Implications for understanding and potentially constraining uncertainty in projections are discussed