.svg){kind=spacer}
Details
- Google DeepMind and Google Research have introduced Weather Lab, an experimental AI-powered platform for predicting tropical cyclones, collaborating closely with the U.S. National Hurricane Center (NHC) ahead of the 2025 Atlantic hurricane season.
- The Weather Lab leverages Google’s stochastic neural networks and integrates data from NHC, ECMWF, and NOAA to predict cyclone track, intensity, size, and formation.
- The system produces 50 scenarios extending up to 15 days ahead—three times longer than traditional models—and simultaneously forecasts both cyclone track and intensity.
- Internal results show the model’s 5-day cyclone track forecasts are 140 km more accurate than the ECMWF’s ENS model, and it surpasses NOAA’s HAFS model for intensity predictions.
- FOX Weather notes active evaluation of the AI model during forecasts for the projected above-average 2025 Atlantic hurricane season, with meteorologist Bryan Norcross involved in the rollout.
Impact
This AI innovation addresses longstanding challenges in both track and intensity hurricane forecasting, granting emergency managers potentially 1.5 additional days for critical decision-making. The partnership with leading agencies demonstrates rising confidence in AI forecasts, and global collaborations hint at far-reaching improvements in climate response and disaster preparedness strategies.
