EarthCube-funded Astrophysicists Use AI for Ten Year Study Regarding our Sun

Alexander Kosovichev, a physics professor at the New Jersey Institute of Technology (NJIT), in collaboration with Egor Illarionov of Moscow University and Andrey Tlatov of the Kislovodsk Solar Station, has published results of their most recent EarthCube-funded solar corona research in The Astrophysical Journal.

Entitled Machine-learning Approach to Identification of Coronal Holes in Solar Disk Images and Synoptic Maps, the study examined and analyzed solar synoptic maps for 2010 through 2020. Because these specialized maps encompassed a daily view at a specific time, for ten years, the calculations were completed using machine learning. An emerging area of artificial intelligence, the researchers’ machine learning approach included an array of algorithms to efficiently analyze the collected data.

Thanks to machine learning, the team discovered that many coronal holes over the decade were associated with magnetic flux transport events. However, they also found that additional factors played a role in the formation and evolution of coronal holes.

“These investigations were important because larger coronal holes result in hefty solar wind speeds,” explained Kosovichev. “In turn, this causes adverse space weather conditions that negatively impact satellite and power distribution grids as well as other infrastructure in our own atmosphere.”

“The automatic real-time identification of coronal holes by the machine learning techniques will enhance the operational space weather forecasts,” continued Kosovichev.

While Kosovichev and his team shed some light on coronal holes, he said that more research is needed for further understanding. He and a team of researchers at the NJIT Center for Computational Heliophysics are now working to develop tools to further extract and analyze existing spacecraft and ground-based data for specific insight to the corona.

“Although this published study gave us a glimpse at ten years of the solar corona, we now want to reach farther for even more knowledge,” Kosovichev said. “Thanks to EarthCube funding, we are able to do just that and look forward to sharing additional insight soon.”

This research was funded by NSF (1639683).