CSU, Wyoming teams test AI tools for longer hail warnings

Colorado and Wyoming scientists are testing artificial intelligence models to predict hailstorms earlier.

The project combines existing machine learning tools with radar data to forecast weather events up to eight days ahead. Kristen Rasmussen, an associate professor at Colorado State University, leads a team of five professors and four research scientists. The group includes three CSU faculty members and two from the University of Wyoming. Rasmussen said the system feeds atmospheric data into a hail prediction model to identify conditions that lead to ice formation. The team is currently developing and testing the technology. Additional field research is planned for spring 2027. The researchers hope to work with the National Weather Service to put the system into practical use.

Farmers in the region stand to benefit from the longer lead time. Dave Petrocco, a farmer in LaSalle, said severe weather remains a major challenge for his operation. Hailstorms hit his crops in June, damaging some fields. The ice arrived early enough in the growing season for the plants to recover. Petrocco said the storm was hard to predict. He noted that workers often have tractors and equipment in the field during planting and harvesting. He described the hail as stones from the sky. Petrocco said better warnings would save the farm money and reduce stress. He called the potential for earlier notice a real gift.

Hailstorms cause billions of dollars in damage across Colorado. The weather event is a major factor in homeowners insurance costs in the state. Rasmussen said the new technology could give homeowners and drivers more time to prepare. The system would allow people to move vehicles and protect property before a storm arrives. The research aims to improve accuracy and timing for destructive weather events in the region.