Scientists at NOAA's Atlantic Oceanographic and Meteorological Laboratory have unveiled a new artificial intelligence model that significantly improves hurricane track predictions beyond the five-day forecast window. In testing against historical storms, the AI model reduced track errors by 25 percent at the seven-day range compared to current operational models.
The breakthrough leverages machine learning algorithms trained on 40 years of satellite data and ocean observations to better capture the subtle atmospheric patterns that steer tropical cyclones. The model is particularly effective at predicting the recurvature points where hurricanes change direction.
NOAA plans to integrate the AI model into its operational forecast suite beginning with the 2026 hurricane season, giving emergency managers and coastal residents more accurate extended-range guidance for evacuation and preparation decisions.
