A new AI model called WeatherNext Cyclones (WN-C) can forecast tropical cyclone tracks and intensity up to two weeks ahead, offering the potential for an extra day of warning.

Wider public safety benefits hinge on responsible sharing and careful deployment and interpretation of AI-generated forecasts.

The story builds on the Nature article by Alet et al. detailing WN-C, with additional references from Earth Space Science, Tropical Cyclone Research and Review, and related AI/Earth system literature.

A Nature News & Views piece by Beucler and Gomez summarizes Alet et al.’s work, highlighting large-scale predictive skill for track and intensity over a two-week horizon.

The article situates WN-C within existing forecasting challenges and surveys related meteorology and AI research, including prior AI-based weather studies and validation efforts.

By extending lead times, WN-C could enhance early warning and disaster preparedness, potentially saving lives and reducing property damage when information is shared responsibly and widely.