Colorado State University Improves Hail Forecasting with Artificial Intelligence

Atmospheric and radar scientists at Colorado State University are developing a new artificial intelligence-powered forecasting system to detect severe hailstorms that damage crops, roofs, and vehicles in advance. Funded by the National Science Foundation, the project aims to provide warnings hours or days ahead of time in the Colorado and Wyoming regions.

While current forecasting models typically cannot provide hail warnings more than 30 minutes in advance, CSU researchers aim to extend this timeframe to hours or even days. The project combines a short-term nowcasting model developed with NVIDIA with a CSU machine learning model used for long-term storm forecasting.

Project funding is provided through the Advanced Environmental Decision-making Intelligence and Computation (ASCEND) Engine in Colorado and Wyoming. This engine is operated with 10 years of federal support totaling up to $160 million. Researchers also plan to install RM60 sensors manufactured by Vaisala in the spring of 2027 to validate the predictions of the new artificial intelligence model.