NOAA has integrated AI-enhanced forecasting models into its hurricane prediction system for the 2026 Atlantic hurricane season, expected to improve track forecasts by 15% and intensity forecasts by 20% compared to current operational models.
The new system, called HAFS-AI (Hurricane Analysis and Forecast System with AI), combines traditional physics-based atmospheric modeling with machine learning trained on decades of hurricane observations. The AI component excels at predicting rapid intensification, historically one of the most challenging forecast problems.
Rapid intensification — defined as a 35 mph increase in maximum sustained winds within 24 hours — has become more common as ocean temperatures rise. The phenomenon has caught forecasters off guard multiple times in recent seasons, leading to dangerous situations where communities had insufficient warning.
Testing during the 2025 season showed HAFS-AI correctly predicted 70% of rapid intensification events compared to 40% for the previous operational model. Earlier and more accurate intensity forecasts enable better evacuation decisions and emergency resource positioning.
The upgrade also includes improved storm surge modeling that accounts for sea level rise and local coastal topography. The enhanced surge forecasts will be incorporated into evacuation zone maps, providing more precise guidance for emergency managers.
