AfriClimate AI has officially launched its flagship project, Forecast4Africa.
Funded by Google.org, the initiative looks to tackle one of Africa’s most pressing challenges: the lack of accessible, accurate climate data and localised weather forecasts.
According to Rendani Mbuvha, co-founder of AfriClimate AI, over 70% of Africans depend on rain-fed agriculture, yet too many still lack timely and localised weather information.
“Climate change is making rainfall patterns more unpredictable, and without accurate forecasts, planning becomes a risky guess. Forecast4Africa aims to change that,” he says.
Forecast4Africa will partner with key organisations to deploy automatic weather stations, gather high-quality local climate data, and build AI-powered medium-range forecasting models that are fine-tuned to local conditions.
While AI offers the potential to make weather forecasts more accurate and widely accessible, forecasts only deliver real-world impact when they actively guide real-world decisions – from farm planning, to energy operations, to disaster risk management, notes Mbuvha.
“Through our Forecast4Africa workshops, we aim to bring key user groups together, ensuring the forecasts we develop are practical, relevant and more likely to drive meaningful outcomes.”
For millions across the continent, knowing when the rains will or will not come, for example, can be the difference between a good harvest and a failed one, he says.
“The funding from Google.org marks an important milestone in realising our mission of harnessing AI for a climate-resilient Africa.
“It enables us to co-develop open source, AI-based weather forecasting tools and close critical climate and weather data gaps. The funds also enable us to build local capacity in this fast-changing field. Most importantly, it ensures African expertise and data are at the heart of model development and evaluation, thus laying the foundation for the long-term sustainability of the project.”
Forecast4Africa’s first pilot will take place in South Africa and the first automatic weather stations will be in rural areas where subsistence farmers primarily operate.
In South Africa, this includes regions identified by the South African Weather Service as having coverage gaps, such as rural Limpopo, rural coastal zones of the Eastern Cape, and the flood-prone interior highlands of KwaZulu-Natal.
“We also plan to expand coverage into other parts of the SADC region,” he notes.
Additionally, AfriClimate, through the Forecast4Africa project, hopes to release an open source toolkit to make it easier for others to improve forecasts in their own regions.
“Too many African farmers and communities operate at the margins of viability, due to increased vulnerability to variability in weather and trends in climate. By combining world-class AI with local data, expertise and infrastructure, we can close that gap and help safeguard lives and livelihoods,” says Mbuvha.
Forecast4Africa will run from August 2025 through to December 2027, and invites weather forecast developers, users, researchers and related communities to join their journey of Africa-led climate innovation.
To learn more or get involved, visit https://africlimate.ai/ or reach out to research@africlimate.ai.
Share