A new NASA tool can predict which parts of the Sun are likely to produce solar storms before those regions even appear. The model, called COFFIES, uses artificial intelligence to scan the Sun's surface and flag areas that may become active and unleash space weather toward Earth. This gives forecasters a head start of up to a few days, a major leap from the current ability to warn only hours ahead.
A Forecast That Looks Beneath the Surface
COFFIES, short for Convective Onset Forecasting from Features in Imaging and Emission Spectroscopy, does not just watch the Sun's visible surface. It analyzes data from NASA's Solar Dynamics Observatory, including magnetic maps and extreme ultraviolet images, to detect subtle changes deep in the Sun's interior. When those changes suggest a new active region is about to emerge, the model flags it as a potential storm source.
The system was developed by a team led by researchers at the University of Colorado Boulder and tested against years of historical solar data. In trials, COFFIES correctly predicted the emergence of active regions that later produced solar flares and coronal mass ejections. The model also reduced false alarms, which have been a persistent problem in space weather forecasting.
Why Early Warning Matters on Earth
Solar storms can disrupt power grids, GPS signals, and satellite operations, and they pose risks to astronauts. In the United States, agencies like NOAA rely on timely forecasts to protect infrastructure. With COFFIES, forecasters could know days in advance that a storm-causing region is forming, giving utilities and satellite operators more time to prepare.
The tool is still experimental, but NASA plans to integrate it into operational forecasting. For now, COFFIES offers a promising glimpse of how machine learning can turn vast streams of solar data into practical, life-protecting predictions. The Sun may still be unpredictable, but humanity's ability to see its next move is growing sharper.
