25 February 2025
Experts who have previously worked on studies investigating the use of AI in weather forecasting at the ºÚ¹Ï³ÔÁÏÍø comment on that ECMWF’s Artificial Intelligence Forecasting System (AIFS) is now active.
Professor Andrew Charlton-Perez said: “The first AI-based forecasting model to enter operations at a major weather prediction centre is a huge milestone, perhaps on the same level as the first operational computational weather forecasts in the 1950s and 1960s or the first operational ensemble forecasts in the 1990s.
“The speed of development of these techniques has surprised everyone in the field. As , both they and other organisations see huge potential in further developing AI-based models to complement and enhance warnings of extreme weather events on a range of different timescales and at different resolutions.
“There is a huge amount of new science to do to interrogate and challenge these models to produce forecasts consistent with the physical phenomena that cause extreme weather. It’s also a fantastic opportunity for talented mathematicians and computer scientists to bring their skills to the weather forecasting problem, a key adaptive tool for climate resilience.”
Dr Simon Driscoll said: "ECMWF's new AI weather system is an exciting development. A 2024 ºÚ¹Ï³ÔÁÏÍø University study showed AI can match traditional forecasts while using far less computing power. Our research testing AI ability to forecasting certain storms showed their remarkable performance in some aspects but their need to improve in the forecast intensity for instance. The AIFS takes this further - it's 20% more accurate for predicting tropical cyclones and uses 1,000 times less energy. This advances technology that protects people while slashing energy costs - exactly what's needed as extreme weather becomes more common."
We are in the process of finalising our postgraduate taught courses for 2026/27 entry. In the meantime, you can view our 2025/26 courses.
