An artificial intelligence can detect certain active regions on the Sun up to 12 hours before any visible manifestation appears on its surface.
These regions correspond to strong concentrations of magnetic fields. When they reach the solar surface, they can form dark sunspots. They are also responsible for more violent phenomena, including flares and coronal mass ejections that may affect Earth's space environment.
The Sun observed on August 14, 2026, at 335 angstroms by the AIA instrument aboard the Solar Dynamics Observatory.
Credit: NASA/SDO, AIA, EVE and HMI science teams.
The problem is that an active region begins forming beneath the surface. Visible signs therefore appear relatively late. To gain time, a team brought together around NASA's COFFIES research center looked for subtler traces of this upward movement that occur before sunspots appear.
The researchers relied on data from the Solar Dynamics Observatory, a NASA satellite that continuously observes our star. They studied small variations in acoustic waves traveling through the Sun. Magnetic material rising from the interior disrupts these vibrations before becoming clearly visible at the surface.
An artificial intelligence system was trained to recognize how these signals evolve over time. It identifies a combination of changes indicating that an active region is rising. According to NASA, the method can provide up to 12 hours of advance warning.
This lead time could have practical value for space weather forecasting. Solar flares can send energetic radiation and charged particles into space. Some events disrupt radio communications, threaten satellites or increase astronauts' exposure. A few extra hours could therefore help the operators concerned prepare.
However, detecting the emergence of an active region is not the same as precisely predicting a flare. Not all regions produce a dangerous event, and their behavior evolves after they emerge. The new tool therefore adds early information, but it does not replace magnetic monitoring and the other methods used after sunspots appear.
The team is now working to improve forecasts and better connect the signals detected beneath the surface with the events observed afterward. This approach could also provide scientists with new data on how magnetic fields rise through the Sun's outer layers.