News facts
How can artificial intelligence help forecasters anticipate floods and farmers decide when to harvest? Today at the UN Climate Change Conference (COP31), a masterclass is exploring how developing countries can use AI to make water, energy and food systems more resilient to climate change and reduce emissions.
Martial Bernoux, Senior Natural Resources Officer at the Food and Agriculture Organization of the United Nations (FAO), is leading the Masterclass on AI for Water-Energy-Food Systems. A fireside discussion and audience Q&A is examining where AI is already proving useful, what skills and resources countries need to move beyond pilot projects, and how to weigh the benefits of AI against its own demand for water and energy.
The Paris Committee on Capacity-building, the Technology Executive Committee and the UN Climate Technology Centre and Network (CTCN) are co-organizing the session in partnership with the Korea International Cooperation Agency. It forms part of the UNFCCC Technology Mechanism Initiative on AI for Climate Action.
FAO estimates that around 1.2 billion people live in agricultural areas facing severe water constraints. Agriculture currently accounts for 72 per cent of global freshwater withdrawals, with irrigation the main driver. Where farmers rely on pumps, deciding when and how much to irrigate affects both the water and the energy needed to grow food.
How Malawi uses AI to forecast floods
A global study published in Nature found that AI can forecast extreme river floods up to five days in advance. In river basins without gauges, those forecasts were as reliable or more reliable than same-day estimates from the Copernicus Global Flood Awareness System.
In Malawi, a project supported by UN CTCN through the Adaptation Fund Climate Innovation Accelerator (AFCIA) is laying the groundwork for AI-assisted flood warnings. River-level and rainfall readings that once reached government offices at the end of the month are now available within minutes. More than 100 observers send readings by free text message to Blue IQ, a national platform that forecasters use to track conditions as they change.
AI can then be used to analyse these readings alongside satellite data and forecast where flooding is likely. Earlier, more localised warnings could give farmers time to harvest before floodwaters arrive and help emergency teams decide where to position supplies.
The USD-60 million AFCIA, funded by the Adaptation Fund, provides grants and technical assistance to test and scale these kinds of innovative climate solutions. Here are a few other examples of how UN CTCN is supporting AI-driven solutions through AFCIA:
- Georgia is using AI to detect early signs of smoke through a network of solar-powered monitoring towers with satellite connectivity. Drones and digital maps help responders locate fires. Following the pilot, the Agency of Protected Areas co-financed an expansion, bringing approximately 12,000 hectares under monitoring.
- Nepal is using AI and more than 30 years of local weather records to advise farmers in Rainas, Bhojpur and Sarlahi on when to sow, irrigate or delay a harvest. Developed with CTCN, RMSI and The Small Earth Nepal, the service turns forecasts into farming advice and sends updates every three days. Most farmers choose to receive them by text.
- Mongolia piloted a digital insurance product that lets herders submit photographs by phone, speeding up claims assessments and payouts. The US$200,000 pilot attracted US$7.5 million in follow-on funding from the Government of Canada. A planned crop insurance tool would combine farmers’ photographs with satellite data and AI to assess risks and process claims.
"From Malawi to Mongolia, we can already see powerful examples of how AI is giving communities more time to prepare for climate impacts while helping planners identify where protection is most urgent," says Ariesta Ningrum, Director or UN CTCN.
"But accessibility is still a pressing issue and we have to work harder to ensure these solutions are reaching the people who need them most."