Google DeepMind and Google Research have open sourced the WeatherNext 2 and WeatherNext Cyclones AI models, which achieved state-of-the-art accuracy in predicting tropical cyclone tracks, intensity, and wind structure. Published in Nature, the work demonstrates that these models provide forecasters with an extra day of predictive accuracy compared to prior systems.

  • The model was co-trained on nearly 20 terabytes of global atmospheric data and the IBTrACS database of 5,000 historical storms.
  • It uses Functional Generative Networks to generate ensembles of predictions, scaling from 50 to 1,000 members to capture rare events like rapid intensification.
  • WeatherNext achieves high accuracy using inputs with a resolution of 28x28km, which is 100 times coarser than traditional models.
  • A compact version, WeatherNext 2-mini, operates at an even coarser 111x111km resolution and runs on a single TPU.

The release aims to empower the research community and meteorological agencies to better predict extreme weather events and build more resilient communities.