DeepMind’s hurricane breakthrough has surprised weather scientists
Open source WeatherNext model can make accurate predictions with lower-resolution weather data.
In October 2025, Hurricane Melissa wreaked havoc across Jamaica, causing widespread flooding and landslides. Weather models had differing predictions about the storm's path, with some suggesting it would end in Haiti while others believed it would intensify and head toward Jamaica. Artificial intelligence model WeatherNext, created by Google’s DeepMind and Google Research, correctly forecasted the storm's landfall in Jamaica five days ahead of time with an 80 percent confidence. This early warning allowed communities to better prepare for the catastrophic event.
Researchers from DeepMind published their findings in Nature, showcasing that WeatherNext can predict cyclones with unprecedented accuracy. On average, the AI model provides forecasters with a full day more lead time compared to existing models. This additional time translates to more accurate three-day predictions than previous models' two-day predictions. In practical terms, that extra day can make a significant difference in response efforts and overall safety.
Written by urgent.news from Ars Technica Science's reporting — not their text. Machine-written; read the original for the full account.
