How WeatherNext helped the National Hurricane Center better predict Hurricane Melissa’s historic landfall in Jamaica

2026-05-16 · Source: Google DeepMind News · Field: Science & Research — Environmental Science & Earth Systems, Research Methodology & Innovation · Depth: Intermediate, medium

Summary

The Google DeepMind and Google Research AI model, WeatherNext, significantly enhanced the National Hurricane Center's (NHC) ability to forecast Hurricane Melissa's historic Category 5 landfall in Jamaica in October 2025. Melissa, the strongest hurricane on record for Jamaica and tied for the strongest in the Atlantic, was predicted by the NHC to reach Category 5 intensity from an initial Category 1 wind speed—a first. WeatherNext provided this critical prediction five days in advance with 80% confidence, increasing to nearly 100% three days out, by excelling at both track and intensity forecasting. This capability, achieved by training on decades of global weather patterns and specialized tropical cyclone datasets, addresses a historical trade-off in meteorological models. WeatherNext also generates 50 different "what-if" scenarios and was the top-performing individual model in the NHC's 2025 verification report, enabling earlier warnings and effective community preparation.

Key takeaway

For emergency management officials and meteorologists focused on hurricane preparedness, integrating advanced AI models like WeatherNext into your operational toolkit is crucial. This technology provides significantly earlier and more accurate predictions of rapid intensification and landfall, as demonstrated by Hurricane Melissa's forecast. Utilizing these high-confidence signals allows you to issue urgent warnings days in advance, enabling communities to mobilize resources and coordinate evacuations more effectively, ultimately reducing human and economic toll.

Key insights

AI model WeatherNext bridges the historical gap in predicting both tropical cyclone track and intensity.

Principles

• Rapid intensification is exceedingly dangerous and difficult to predict.
• Traditional models trade track accuracy for intensity resolution.
• Ensemble forecasting informs decision-making with broader possibilities.

Method

WeatherNext trains on decades of global weather patterns and specialized tropical cyclone datasets, then runs ensembles of 50 "what-if" scenarios.

In practice

• Integrate AI forecasts with physics-based models and real-time data.
• Utilize ensemble data for expert decision-making.
• Expand AI forecasting to other critically affected regions.

Topics

• WeatherNext
• Hurricane Forecasting
• Rapid Intensification
• AI Weather Models
• Disaster Preparedness
• Ensemble Forecasting

Best for: Executive, AI Scientist, Research Scientist, Director of AI/ML, Domain Expert

Related on AIssential

• Predicting a historic storm earlier with WeatherNext — AI models can significantly improve the accuracy and lead time of hurricane intensity and trajectory forecasts.
• Towards a world where no one is surprised by a natural disaster — AI-powered forecasting and real-time alerts significantly enhance global preparedness and response to natural disasters.
• Data-Driven Climate Research: Insights from the Thai Meteorological Department (หลังฟัง) — Data-driven climate research and expert interpretation are crucial for mitigating extreme weather impacts.
• Hybrid NARX-LLM for Greenland Iceberg Discharge: Prompt-Driven Residual Correction — A hybrid NARX-LLM framework uses physics-informed prompts to correct iceberg discharge predictions, enhancing accuracy for complex climate dynamics.
• Enhancing AI and Dynamical Subseasonal Forecasts with Probabilistic Bias Correction — Probabilistic Bias Correction (PBC) significantly improves subseasonal weather forecast accuracy and extreme event prediction.
• Benchmarking Physics-Informed Time-Series Models for Operational Global Station Weather Forecasting — The WEATHER-5K dataset provides a comprehensive, large-scale benchmark for global station weather and general time-series forecasting, addressing prior data limitations.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Google DeepMind News.