Understanding climate change in African rainfall via a new generation of high-resolution models

Africa is home to some of the populations most vulnerable to climate change. Regional climate projections of rainfall for Africa, and in fact elsewhere in the tropics, are highly uncertain and in many regions we do not know if it will get wetter or drier. The representation of the convective clouds that deliver rainfall in the tropics is known to be a major source of error in global climate models, but the new generation of high-resolution “convection-permitting” models provide a route for addressing this, and a new opportunity to answer fundamental questions on the role of upscale impacts of convection on regional climate change. This project will use the first ever ensemble simulations of future African climate run at a convection-permitting scale to generate new understanding of the processes controlling climate change in Africa, how this is represented in state-of-the-art models and the implications for climate projections.

Organised atmospheric convection pover Africa (credit: NASA)


Africa is the continent with the largest tropical land mass, and its rapidly growing population includes many of the people who are most vulnerable to climate change. Rainfall in the tropics and sub-tropics is dominated by convective storms. It is these storms, or their absence, that provides most high-impact weather. Furthermore, these storms form a key component of tropical atmospheric circulation. Rainfall corresponds to the net condensation of water vapour, and it is this condensation that provides most heat to the tropical atmosphere. This heating is communicated to the wider atmosphere and affects atmospheric circulations on scales far greater than the convection itself. Similarly, winds and shading from storms affect sea surface temperatures, and rainfall affects soil moisture: these then feedback onto the atmosphere on time-scales longer than the convective storms. Global climate models are run at spatial resolutions that do not capture convective storms, and so the storms are parameterised as sub-grid processes in such models. These parameterisations are known to introduce errors in not only local rainfall, but also large scale weather systems, and indeed entire monsoon systems.

Convection-permitting simulations allow an explicit representation of convection and a step-change in performance for not only modelled rainfall, but also couplings between the convection and larger scales. They are however limited by computational power. Leeds played a leading role in the £20 million Future Climate for Africa (FCFA) programme, which ran the first convection-permitting future climate simulations for Africa. Referred to as “CP4A”, these had a small enough grid-spacing (approximately 4km) to explicitly capture convective storms, but the simulations were limited by the fact that they were only for a single ten-year future period, with the boundary-conditions provided to the regional model over Africa provided by a single realisation of global climate change from a single global model. The Met Office is now running the first-ever ensemble of such simulations driven my different parent global models. This provides a unique new opportunity for understanding how climate change will affect weather in Africa, how convection affects larger scales, and the limitations of the current generation of global climate models,



The project will address the following scientific objectives:

  • How will climate change affect regional rainfall projections in Africa, and how does this depend on the couplings between convection and larger scale systems, from sea breezes to the entire Hadley circulation?
  • How does this depend on internal atmospheric processes versus coupled processes such as land and ocean change?
  • How do changes seen in models relate to those observed in reality, and can this be used to constrain the diverse predictions from models?

Depending on the interests of the candidate, further objectives could include: (i) identification of the role of non-greenhouse gas forcings such as aerosols and land-use change; or (ii) or more focus on the changing rainfall extremes.

These objectives will be tackled using  the unique new ensemble of convection-permitting Unified Model simulations run by the Met Office, alongside satellite datasets and reanalyses (a blend of observations and models). The student will need a strong background in physical science and computational approaches, a willingness to work with and visit the Met Office. There may also be opportunities for travel to or fieldwork in Africa. Other opportunities include collaboration with scientific partners overseas.