News

No “One Size Fits All”: Improving Climate Models for Africa Requires African Perspectives

A new paper, published in BAMS (the Bulletin of the American Meteorological Society), calls for collaboration between climate modellers and African scientists, to deliver a dramatic improvement in climate information over Africa.

Climate change is expected to have important consequences for African countries, many of which are already vulnerable to weather-related extremes such as drought, flooding, and – as devastatingly shown in Sierra Leone – mudslides. However, the impact of global warming on specific African regions remains quite uncertain.

Scientists can use computational models to try to predict how climate might change in future. Inspite of the fact that climate model experiments are increasingly being used by planners and risk managers in Africa, to try to prepare for a changing climate, the models have largely been developed outside of the continent. What’s more, historically, greater effort has gone into studying and improving models over the regions where they were developed. For many parts of Africa, there has been limited work to evaluate the models’ ability to capture key climate features.

This does not prevent the models from being useful for African regions: their simulations cover the entire globe and can already offer some insight into future climate change. However, African climate systems are particularly tricky to represent in climate models. There is an important role for convective processes, which occur on finer scales than the models can resolve. Rainfall in many parts of Africa is also strongly linked to remote ocean basins, and it is challenging to simulate these “teleconnections”.

“Our paper underlines the need to fill the gap in understanding climate mechanisms in many regions over the continent”, says co-author, Dr Wilfried Pokam, from University of Yaounde I, Cameroon, highlighting that Central Africa in particular has received limited attention to date.

Tailored analysis is needed for every region: there is no “one size fits all”. The Congo Basin, home to one of the largest rainforests in the world, is dominated by tropical rainfall and thunderstorms. In contrast, the semi-arid Sahel relies on one rainy season each year, brought by the West African Monsoon. In parts of East Africa, – which has been the focus of the work by co-author Dr Joseph Mutemi, from University of Nairobi – there are two rainy seasons each year, and the amount of rainfall depends strongly on interactions with the Indian and Pacific Oceans.

The paper demonstrates analysis of climate features which are important for Central, East, Southern, and West Africa, developed by the team of scientists from South Africa, Kenya, Cameroon, and the UK. The authors highlight the potential to deliver a dramatic improvement in understanding of climate models over Africa, by drawing on the wealth of local weather and climate expertise in African meteorological services, research institutes, and universities.

“Our aim is to trigger further discussion between climate modellers and African scientists about how best to evaluate climate models over Africa.” says lead author, Dr Rachel James, visiting researcher at the University of Cape Town. The authors propose the idea of a model evaluation “hub” for Africa, to coordinate further discussion, and ultimately develop diagnostics which could be embedded within model development infrastructure, and fast-track understanding of climate model behaviour over Africa.

According to the authors, it is only by understanding the models better, that scientists can really support adaptation planning. “This kind of evaluation work can provide tips to the model developers about how to fix errors. That could ultimately lead to better models, and perhaps even better predictions.” says James.

The evaluation work could also help scientists provide more useful statements to planners about how much they should trust climate model data. “Evaluating the capability of the models to simulate climate over the continent will improve our ability to assess confidence in the predictive power of the climate models, and inform application of the model products in decision making.” says co-author Dr Babatunde Abiodun, from University of Cape Town. “To ensure credible climate change information, it is crucial to ensure that models perform well for good reasons.” says Pokam.

The paper is part of the Future Climate for Africa programme, and in particular a project called IMPALA, which seeks to better understand and improve the Met Office model over Africa. This project builds on several years of focussed work on Africa by Met Office scientists and African partners to start to turn an ‘African’ lens on the model, delivering sustained improvement in the representation of precipitation over the continent.

For further information, please contact:

Rachel James (Oxford University/University of Cape Town): rachel.james@eci.ox.ac.uk,

Babatunde Abiodun (University of Cape Town): babiodun@csag.uct.ac.za

Wilfried Pokam (University of Yaounde I): wpokam@yahoo.fr

Joseph Mutemi (University of Nairobi): jnmutemi@yahoo.co.uk

Jean-Pierre Roux (Future Climate For Africa): jean-pierre@southsouthnorth.org

The paper ‘Evaluating climate models with an African lens’ has been published online by the Bulletin of the American Meteorological Society (available at: http://journals.ametsoc.org/doi/abs/10.1175/BAMS-D-16-0090.1) and will go into print later in 2017 or early 2018.