Scientists Deploy AI to Forecast Solar Storms Weeks in Advance
Researchers at Southwest Research Institute and NSF-NCAR have developed PINNBARDS, a physics-informed neural network that predicts dangerous solar eruptions up to three weeks before they strike—a dramatic leap beyond the hours-long warning windows that currently protect Earth's satellites, power grids, and astronauts.
The tool marks the first operationally viable system capable of linking surface solar observations to the Sun's deeper magnetic architecture, transforming space weather forecasting from a reactive discipline into a proactive one. A research team led by Dr. Mausumi Dikpati published their findings in the Astrophysical Journal, presenting PINNBARDS as a potential game-changer for infrastructure protection and human spaceflight safety.
The Long-Standing Problem
For decades, heliophysicists have struggled with a fundamental blindness: they cannot reliably predict where or when the Sun's tangled magnetic field lines will erupt into solar flares and coronal mass ejections (CMEs)—explosive phenomena that hurl radiation and energetic particles toward Earth at speeds exceeding 1,000 kilometers per second. A single major CME can cripple satellites, destroy communications networks, disrupt GPS systems, and knock power grids offline across entire continents. The 1859 Carrington Event—a solar superstorm that occurred before electricity dominated society—would cost a modern economy trillions in damages if it repeated today.
The challenge stems from the Sun's chaotic subsurface dynamics. Active regions form when magnetic field lines become twisted and tangled beneath the photosphere, eventually breaking through to create visible sunspots and flare-prone structures. Traditional forecasting methods observe only surface signatures—the visible aftermath. They offer perhaps 12 to 24 hours of warning before a flare erupts, leaving little time for meaningful mitigation.
"Understanding where and when large, flare-producing active regions on the sun would emerge is a long-standing problem in heliophysics," explained Dr. Subhamoy Chatterjee, an early-career scientist at SwRI and co-author of the research.
How PINNBARDS Works
Instead of focusing solely on surface features, PINNBARDS reconstructs the subsurface magnetic state using data from NASA's Solar Dynamics Observatory (SDO) and its Helioseismic and Magnetic Imager (HMI). The system ingests real-time magnetograph data—measurements of the Sun's magnetic field—and uses a physics-informed neural network (a hybrid AI-physics approach) to infer the 3D magnetic structure deep beneath the photosphere.
Think of it as X-raying the Sun's interior. Once PINNBARDS maps the subsurface field geometry, it feeds those reconstructed initial conditions into forward simulations of solar magnetic evolution. The result: predictive maps showing where active regions are likely to emerge weeks in advance.
"The reconstructed subsurface states from PINNBARDS provide initial conditions for forward simulations of solar magnetic evolution, opening the door to predicting where and when large, flare-producing active regions are likely to emerge weeks in advance," Dikpati said.
Implications for Space Infrastructure and Exploration
The 14-to-21-day forecast horizon transforms operational responses. Satellite operators can adjust orbits, power grid managers can stage defensive protocols, and space agencies can schedule extravehicular activities around forecasted danger windows. For crewed missions to Mars or long-duration lunar bases, the ability to predict radiation events weeks ahead becomes critical to crew safety—potentially allowing trajectory adjustments or shelter-in-place protocols before hazardous particles arrive.
What Comes Next
The research team is working to integrate PINNBARDS into operational forecasting systems used by NOAA, the U.S. Space Force, and international space agencies. Early validation studies are underway. If the tool maintains its predictive accuracy across a full solar cycle, it could fundamentally reshape how humanity prepares for space weather—moving from damage control to prevention.
