WeatherNext 2 AI model advances cyclone prediction

Predicting the development of hurricanes and cyclones has long been a difficult task, with each hour of advance warning being crucial. Tropical cyclones are some of the most destructive weather occurrences on the planet, making it imperative for meteorologists to provide accurate and timely warnings.

In a recent publication in Nature, researchers at Google have demonstrated that their WeatherNext AI model has achieved cutting-edge precision in forecasting the path, intensity, and wind structure of cyclones. This advancement signifies a significant leap forward, equating to approximately ten years of progress in meteorology encapsulated in a single model.

In an effort to support global climate resilience, Google is now making the WeatherNext 2 model available to the worldwide research community. Further details can be found on the Google DeepMind blog.