AI Boosts Climate Resilience in West Africa

Project Cumulus, funded by the Gates Foundation and the Foreign, Commonwealth & Development Office, will see researchers from the Alan Turing Institute, the University of Cambridge and the University of Leeds link up with universities and meteorological agencies in Ghana and Senegal to co-design more accurate forecasting systems which could help farmers improve crop yields and reduce economic losses.

These systems will be affordable and adaptable, enabling West African partners to produce their own forecasts, build expertise and drive local innovation.

Weather forecasting in West Africa currently presents a unique set of challenges. As a region increasingly vulnerable to the effects of climate change, unpredictable weather patterns have a direct impact on food security and economic stability.

Farmers often lack access to adequate weather forecasts, forcing them to make critical decisions on planting or harvesting crops without the information they need.

"Forecasting and agricultural production are deeply interlinked; from planning, crop management and harvesting through to storage, transportation and marketing," said Professor Leonard K Amekudzi, from Kwame Nkrumah University of Science and Technology in Ghana.

There is an urgent need for weather prediction tools that perform better under African conditions. Global forecasting systems such as the Integrated Forecasting System (IFS) have achieved remarkable advances in overall accuracy, however in sub-Saharan Africa, where weather systems behave differently and local observations are limited, forecasts can still be less reliable.

"Forecasting rainfall in the tropics is a unique challenge, further complicated by climate change and a historical lack of localised data," said Dr Scott Hosking from the Alan Turing Institute. "To protect lives and livelihoods in these regions, we cannot rely on off-the-shelf AI solutions."

The way the Earth rotates, and solar heating drives the atmosphere, means that weather in Africa behaves differently from the mid-latitudes such as Europe and the USA, where most forecasting methods were designed. This means that we need new models which are optimised for African conditions. Challenges are compounded by the reduced availability in West Africa of in-situ measurements (such as weather stations) used to inform and build forecasts.

Because traditional physics-based forecasting approaches used in the Global North are less effective in sub-Saharan Africa, the Cumulus initiative will draw on the Cumulus project team's work on emerging technologies like Aardvark Weather and the Aurora Earth System Foundation Model, to develop new AI-based forecasting methods tailored to African conditions which therefore deliver more accurate and locally relevant insights.

"AI is both revolutionising and democratising weather prediction," said Professor Richard Turner from Cambridge's Department of Engineering. "Tasks that once required a supercomputer can now run on a laptop, producing accurate forecasts in a fraction of the time and cost. It's exciting to combine West African and international expertise and put new advanced forecasting tools directly in the hands of local experts, enabling them to fine-tune models for their own local conditions for the first time."

These technologies show great promise to inform this project as unlike well-established physics-based weather prediction systems Aardvark is fully driven by AI, combining satellite imagery, ground observations and existing forecast data to create a clearer picture of the atmosphere. It can draw on both remote-sensing and local measurements, learning from data-rich regions to improve predictions where data is more scarce such as in sub-Saharan Africa.

The Aurora Earth System Foundation model has showed how a single AI model could be adapted for a wide range of forecasting tasks.

The agility of models like Aardvark and Aurora will allow the Cumulus project to create systems tuned to local weather patterns as well as extending forecasts to sub-seasonal timescales most useful for farmers and planners.

"The emerging science of AI weather prediction, paired with local insights into the physics and statistics of the region's climate, is poised to amplify the benefits we're already seeing in our universities and weather services," said Professor Amadou Gaye, from University Cheikh Anta Diop in Senegal. "This international partnership acts as a catalyst for strengthening climate resilience and food security in the region."

The Alan Turing Institute is the lead institution on the Cumulus project and the partner institutions include the University of Cambridge and the University of Leeds, with Kwame Nkrumah University of Science and Technology (KNUST) in Ghana, University Cheikh Anta Diop of Dakar (UCAD) in Senegal, the Senegalese metrological agency ANACIM and the Ghanaian met agency GMet. The project is made possible due to funding from the Gates Foundation and UK International Development from the UK government.

Adapted from a piece published on the Alan Turing Institute website.