New research by Callum Munday, DPhil student at Oxford University’s School of Geography and Environment, sheds light on complex southern African climate dynamics, offering hope for improved regional climate prediction.  Understanding the 21st century changes in the patterns and characteristics of Southern African rainfall is critically important. The region’s socioeconomic activity is closely tied to rainfall and its variability: agriculture production is predominately rain-fed rather than irrigated, while electricity production is largely derived from hydropower e.g. in Zambia hydro accounts for almost all power generated. Climate models are the best tools we have to predict future changes in rainfall. Unfortunately, in Southern Africa, the latest generation of more than 20 climate models vary greatly in their estimation of historical rainfall (the wettest model is 1.5 times wetter than the driest model). Most of the climate models also tend to overestimate rainfall compared to observational records. This poses a problem: which models can we trust for projecting future climate change? The first step in addressing this problem is to understand why models differ in their simulation of rainfall in the present day. Our recent work, as part of the UMFULA climate project, attempts to address this question by looking at how well the range of models simulate the Angola Low; an important regional circulation feature. The Angola Low is a low pressure system which is anchored over Angola and northern Namibia. In the summer months, December to February, it acts like a broker between the wet, tropical African region and the drier subtropical region. When active, it spins clockwise and diverts moisture in low levels of the atmosphere southward from the wet tropical area to the dry subtropical area. Our hypothesis was that wetter models would have a stronger Angola Low than drier models because more moisture would be transported into the subtropics in these models. Our results supported this hypothesis. In the summer season, between 40 and 60% of the variability in model rainfall is associated with the strength of their simulated Angola Low. Wet models also simulate enhanced moisture transport from tropical to subtropical Southern Africa. Interestingly, the observationally-based reference datasets (“reanalyses”) simulate a weaker Angola Low than most models and are closer in their rainfall climatology to the drier models in the ensemble. This might suggest that drier models are producing a more realistic estimate of Southern African climate, although better observations of the Angola Low are needed to be confident in this assertion. We are continuing this work by investigating in greater detail how the Angola Low affects rainfall in Southern Africa both in the present day and in the future. This work will help to identify credible models for projecting future climate change. These projections can be used to inform adaptation strategies ahead of climate change in the coming century. This article first appeared on Water. This research is part of Callum’s DPhil supervised by Richard Washington and funded by the National Environmental Research Council (NERC). The full paper can be found here.