Computer scientists at the Argonne National Laboratory and leading technology companies are using artificial intelligence to transform weather forecasting — and, if successful, to give power companies a better shot at avoiding catastrophe from future storms.
“AI is already making revolutionary changes in weather forecasting,” said Rao Kotamarthi, the Department of Energy lab’s chief scientist.
Scientists in the U.S., Europe and China have built self-learning AI models based on historical weather data assembled by the European Centre for Medium-Range Weather Forecasts (ECMRF). Compared with conventional weather forecasts, the AI models have trained themselves to produce much faster 12-day weather forecasts with improved accuracy.
Hurricanes Helene and Milton in the past month caused losses to life and infrastructure in parts of the Southeast. Scientists say AI technology could be a game-changer as utility companies and grid operators make difficult real-time decisions during storms big enough to knock offline power generation or flood substations.
