NSF Funds AI Research to Improve Solar Storm Forecasting
The National Science Foundation awarded a three-year, $600,000 grant to Soukaina Filali Boubrahimi, an associate professor at the School of Computing, to develop machine learning models that predict dangerous solar radiation storms.
Solar energetic particle events occur when the sun releases bursts of high-energy particles into space. These events can disable satellites, disrupt GPS and radio communications, damage power grids, and threaten astronauts on space missions. Current forecasting methods struggle with accuracy because the underlying physics remains poorly understood.
Filali Boubrahimi's approach combines data from multiple NASA missions and observatories-including the Solar Dynamics Observatory, Solar and Heliospheric Observatory, Wind, Geostationary Operational Environmental Satellites, and Advanced Composition Explorer-into a single database. The project will analyze solar images, magnetic field measurements, solar wind conditions, and particle observations together to identify patterns that precede major events.
How the Models Will Work
The research will build machine learning systems that operate at two levels. Some models will examine all available data simultaneously to detect broad patterns. Others will focus on specific observations to catch subtle warning signs. Together, they will analyze solar activity from multiple angles.
The models will also investigate how solar particles accelerate and travel through space after eruptions. Machine learning will help group similar events and clarify the physical mechanisms that drive the most severe radiation storms.
Public Access and Training
The final forecasting system will be released through the Community Coordinated Modeling Center, making it available to researchers, government agencies, and space weather forecasters.
The project will also integrate space weather data and machine learning methods into data science courses. Students will work with large scientific datasets and analytical tools increasingly demanded across science and technology fields.
Filali Boubrahimi has secured more than $2.3 million in NSF funding over the past four years as principal or co-principal investigator. The SHINE program distributes roughly $3 million annually to research focused on how the sun produces and sends energy and particles toward Earth.
For professionals working in research and data analysis, understanding these approaches to pattern recognition in complex datasets offers insights applicable across scientific fields. AI Research Courses and AI Data Analysis Courses cover similar methodologies for extracting patterns from multimodal data sources.
Your membership also unlocks:
