Google DeepMind's WeatherNext artificial intelligence model helped forecasters provide earlier warnings during Hurricane Melissa in October 2025, according to research published in Nature.
What Happened
WeatherNext predicted that a storm system brewing over the Caribbean Sea would intensify and hit Jamaica as a Category 5 hurricane with 80 percent confidence five days before landfall. Traditional weather models at the time differed on the storm's trajectory, with some suggesting it would remain weak and affect Haiti instead. The AI model's prediction proved accurate when Hurricane Melissa caused catastrophic flooding and landslides across Jamaica. Researchers report in their paper that WeatherNext can predict cyclones with unprecedented accuracy, giving forecasters an average of one additional day of lead time compared to existing models.
Why It Matters
The extra day translates into more preparation time for communities in a hurricane's path. According to the research published in Nature, WeatherNext's three-day predictions are as accurate as previous models' two-day forecasts. That difference can mean whether residents have critical hours to evacuate, secure supplies, or make shelter arrangements. The advancement represents a measurable improvement in how AI systems can support human meteorologists during life-threatening weather events.
The Bottom Line
DeepMind's WeatherNext demonstrated real-world utility during Hurricane Melissa by providing earlier, higher-confidence forecasts than conventional models. The research shows the system gives forecasters an additional day of lead time on average for cyclone prediction, though longer-term performance across diverse storm scenarios remains under study.