NASA has developed a machine learning model called COFFIES that analyzes magnetic field and acoustic wave measurements at the sun's surface to predict where active regions (sunspots) will form, up to 12 hours before they become visible. The model works indirectly, inferring subsurface magnetohydrodynamic activity rather than observing it directly, and functions as something of a black box. Researchers hope its predictions could give extra advance warning for geomagnetic storms, such as a future Carrington Event-class solar flare, though it remains a research tool rather than a deployed early-warning system.
Questions this post answers
How does NASA's COFFIES model predict sunspots before they're visible?
COFFIES (Consequence Of Fields and Flows in the Interior and Exterior of the Sun) is a machine learning model that analyzes magnetic field and acoustic wave measurements at the sun's surface to infer material flows and magnetic activity deep inside the sun, allowing it to predict active regions, or sunspots, up to 12 hours before they visibly form. daily.dev surfaces how machine learning pattern recognition gets applied to real-world forecasting problems like this.