Trained on nine years of data from NASA’s Solar Dynamics Observatory (SDO), Surya can predict solar flares up to two hours in advance—improving forecast accuracy by 16 % compared to previous benchmarks. This advance offers a critical window for protecting satellites, power grids, and communication systems from potentially damaging solar storms.
Background
Solar activity, such as flares and coronal mass ejections, drives space weather, which can seriously impact modern technology. These events have the potential to disrupt satellite communications, interfere with GPS signals, cause power grid surges, and increase radiation exposure for both aviation and space missions.
Accurately predicting this kind of activity has long challenged researchers due to the Sun’s dynamic and complex behavior. NASA’s Solar Dynamics Observatory has been collecting high-resolution, multi-wavelength images of the Sun every 12 seconds for over a decade, resulting in an unprecedented long-term dataset. This consistent record of solar behavior across an entire solar cycle creates an ideal training ground for advanced AI models like Surya to identify predictive patterns in solar activity.
The Surya Model and What it Can Do
Surya, short for “Solar Understanding and Research for Yielded Advances,” represents a major step forward in using AI for space science. Unlike traditional AI systems that rely heavily on labeled datasets and are designed for narrow tasks, Surya uses a foundation model architecture that learns from vast amounts of raw, unlabeled solar data.
Trained on nine years of carefully calibrated data from the SDO, Surya incorporates ultraviolet and extreme ultraviolet imagery from the Atmospheric Imaging Assembly and magnetic field maps from the Helioseismic and Magnetic Imager. This deep and diverse dataset enables the model to detect subtle and complex patterns that shorter or less comprehensive datasets would likely miss.
One of Surya’s most valuable applications is solar flare forecasting. It can generate detailed visual predictions up to two hours in advance, outperforming existing models by 16 %. But its capabilities go further—it can track active regions on the Sun, estimate solar wind speed, and integrate data from other missions like the Solar and Heliospheric Observatory and the Parker Solar Probe.
By releasing Surya as open-source software on platforms like Hugging Face and GitHub, NASA and IBM are inviting researchers and developers around the world to contribute and expand its capabilities.
Why it Matters
Solar storms present real risks to essential infrastructure. A strong coronal mass ejection, for example, can create geomagnetic currents that overload power grids and trigger widespread outages. In aviation, solar flares can disrupt high-frequency radio and navigation systems, especially on polar routes, while exposing passengers and crew to elevated radiation levels.
The risks are even higher for human spaceflight, where astronauts beyond Earth’s magnetic shield need timely warnings to take protective action during solar particle events. Meanwhile, increasing solar activity also affects the growing number of satellites in low Earth orbit. When the upper atmosphere heats and expands, satellite drag increases, potentially pulling them out of orbit or causing them to re-enter prematurely—raising both financial and safety concerns.
Surya’s predictive capabilities offer operators a valuable head start, allowing for better satellite management and grid protection. Beyond space weather, its underlying architecture opens the door for applications across other scientific domains—from climate research to planetary science. Supported by federal initiatives like the National AI Research Resource (NAIRR) Pilot, Surya illustrates how partnerships between government and industry can drive scientific progress and enhance technological resilience.
Looking Ahead
The development of Surya by NASA and IBM marks a significant leap in space weather forecasting and heliophysics. By harnessing a decade of high-resolution solar data and cutting-edge AI architecture, the model shifts the field toward proactive prediction.
Surya not only enhances our ability to anticipate and mitigate the effects of solar activity, but also highlights the importance of long-term data collection and open-source collaboration in advancing science.
Disclaimer: The views expressed here are those of the author expressed in their private capacity and do not necessarily represent the views of AZoM.com Limited T/A AZoNetwork the owner and operator of this website. This disclaimer forms part of the Terms and conditions of use of this website.