WeatherNext: AI model achieves breakthrough in forecasting cyclones

Tropical cyclones are among the most destructive weather phenomena, responsible for over 700,000 deaths and $1.4 trillion in economic losses globally in the past 50 years. Issuing timely, accurate warnings is a constant race against time, and predicting how dangerous cyclones develop has been a longstanding challenge. The work brought together AI researchers and engineers at Google DeepMind and Google Research with expert forecasters at the National Hurricane Center (NHC), the Cooperative Institute for Research in the Atmosphere (CIRA), the UK Met Office, and weather agencies worldwide.

Predicting cyclones has typically forced a trade-off: a cyclone's track is steered by massive global atmospheric currents (best modeled by coarse global models), while its intensity is driven by highly localized processes (requiring high-resolution models). WeatherNext bridges this gap by iteratively predicting both global weather patterns and fine-scale cyclone tracks up to 15 days ahead, starting from global atmospheric conditions such as those during Hurricane Milton (October 2024). Running a 1,000-member ensemble generates localized probability maps of tropical storm to hurricane-force winds.

The model achieved state-of-the-art accuracy in predicting cyclone track, intensity, and wind structure in a paper published in Nature. On average, WeatherNext's three-day forecasts are as good as what prior models provided for two days — an extra day of predictive accuracy, corresponding roughly to a decade of meteorological progress. During the 2025 hurricane season, it helped the NHC make a historic forecast for Hurricane Melissa by predicting rapid intensification and landfall in Jamaica, enabling advance warnings and giving teams on the ground critical preparation time. The team is now predicting 1,000 possible scenarios for each cyclone to support forecaster decision-making.

Given the broad impact, Google DeepMind is open sourcing the WeatherNext 2 and WeatherNext Cyclones models used during the hurricane season. By making the technology openly available, the team hopes to empower researchers, local forecasters, renewable energy projects, and extreme-weather preparedness worldwide.