Weather forecasting training program brings power of AI to low- and middle-income countries

Just as climate change brings extreme and unpredictable weather to communities worldwide, advances in AI are making it possible to predict these shifting conditions with greater speed, lower cost, and hyper-local precision. This breakthrough promises to extend accurate forecasting to regions that have historically lacked access to advanced technologies.

Supported by the Agricultural Innovation Mechanism for Scale (AIM for Scale), a collaboration of researchers from MBZUAI and the University of Chicago are working to ensure that governments in low- and middle-income countries (LMICs) can adopt these innovations early and build world class national services once limited to places such as the US, EU, or Japan.

“AI weather models are revolutionizing what can be done with forecasts, but because they are a new type of technology, governments often lack the training, hardware, and institutional capacity to build and use them effectively,” says Pedram Hassanzadeh, associate professor in Geophysical Sciences at the University of Chicago and member of AIM for Scale’s Technical Panel on Weather Services for Farmers.

“If we don’t act, low- and middle-income countries risk being left behind. We’re working to ensure these countries can harness these innovations – taking cutting-edge AI and climate science from the lab to the field.”

The pioneering program – funded by a grant provided by the International Affairs Office at the Presidential Court of the United Arab Emirates (UAE) – is training staff from national meteorological and hydrological services (NMHS) and ministries of agriculture across LMICs on how to use AI weather forecasting models that are tailored to their specific needs.

Their first cohort – from Bangladesh, Chile, Ethiopia, Kenya, and Nigeria – is being trained from September 22 to 26 in the UAE, hosted by MBZUAI and the National Center of Meteorology (NCM). In the years ahead, additional rounds will expand to 25 more countries, reaching a total of 30 and further broadening reach and impact to potentially millions more farmers.

Tailoring needs for greater impact

AIM for Scale is working on the ground with development partners to ensure this effort is scalable, inclusive, and delivers real impact for farmers.

“Accurate, high-quality forecasts can unlock better yields, higher incomes, and stronger livelihoods for farmers,” says Paul Winters, Executive Director of AIM for Scale and Professor of Global Affairs at the University of Notre Dame’s Keough School of Global Affairs. “By pairing AI innovation with practical agricultural decision-making, we’re creating opportunities for millions of farmers to prosper.”

Fundamental to the program, the team is engaging ministries of agriculture to ensure forecasts can be tailored to the needs of smallholder farmers who are most vulnerable to weather risks. The power of AI is not only in extending access to the types of gold-standard forecasts available in the world’s most well-resourced NMHSs, but also in enabling those forecasts to be generated directly ‘in-house’ by NMHSs at lower cost and with greater accuracy on timelines – from days to weeks – that matter most for farm management. For example, advanced knowledge of the start of the rainy season can help farmers plan what crops to plant and when.

Equally important are reliable short-term forecasts that guide day-to-day decisions about irrigation, fertilizer use, or labor. By partnering with ministries of agriculture, a feedback loop can be created: ministries help inform and guide model development to ensure forecasts meet needs on the ground, while improved forecasts in turn guide farmers to make more impactful decisions.

“From weather data to model verification and downscaling, this training covers the key dimensions of AI-powered weather forecasting that enable farmers to plan under uncertainty,” says Souhaib Ben Taieb, Associate Professor of Statistics and Data Science at MBZUAI. ”It demonstrates how frontier AI research can translate into practical solutions for global challenges like food security.”

Democratizing forecast production

The training combines technical expertise with hands-on capacity building, giving national meteorological teams the tools and autonomy to generate and deliver tailored forecasts and ensuring this information reaches farmers. Along with technical guidance and capacity building, the training solves another challenge: a lack of hardware by providing high-performance multi-GPU laptops to each participant so they can apply and sustain their training back in their home countries.

With the right training and deployment, countries that once lagged behind can now match – and even surpass – the world’s leaders in forecast production.

Alongside MBZUAI, AIM for Scale, UChicago, and NCM, the partnership is convening experts from leading global institutions, including the World Meteorological Organization (WMO), Google DeepMind and Google Research, the European Centre for Medium-Range Weather Forecasts (ECMWF), AfriClimate AI, Norwegian Meteorological Institute, and Precision Development (PxD), among others.

“This training is a powerful example of how AI can be operationalized for the global good,” says Amir Jina, assistant professor at UChicago’s Harris School of Public Policy and Chair of AIM for Scale’s Technical Panel. “For the first time in history, there is a real prospect of democratizing access to world-leading forecasts, with low- and middle-income countries able to build national services that meet the same gold standard as the most advanced countries.”