WeatherNext: AI model achieves breakthrough in forecasting cyclones

Google DeepMind, in collaboration with Google Research and meteorological agencies like the National Hurricane Center, the Cooperative Institute for Research in the Atmosphere, and the UK Met Office, has developed WeatherNext, an AI model that predicts tropical cyclone track, intensity, and wind structure with state-of-the-art accuracy. The model, detailed in a Nature paper, gives forecasters an extra day of predictive accuracy, equivalent to a decade of meteorological progress. It uses Functional Generative Networks to produce ensembles of predictions, scaling to 1,000 scenarios per cyclone. The model was co-trained on global weather data and the IBTrACS historical cyclone database. It requires only 28x28km resolution data, 100x coarser than traditional models, and a mini version works at 111x111km. The technology has been open-sourced, including WeatherNext 2 and WeatherNext Cyclones models, and is accessible via Weather Lab, part of Google Earth AI.