A Large-Scale Dataset of Three-Dimensional Solar Magnetic Fields Extrapolated by Nonlinear Force-Free Method

Area of Science:

• Solar Physics
• Heliophysics
• Computational Astrophysics

Background:

• Solar magnetic fields are crucial for understanding solar activities, particularly energetic events in the solar corona.
• Accurate reconstruction of the three-dimensional (3D) solar magnetic field from observed photospheric magnetograms is essential for solar physics research.

Purpose of the Study:

• To construct a comprehensive, large-scale dataset of 3D solar magnetic fields in active regions.
• To provide open access to data resources and source code, facilitating further research and preventing redundant efforts.
• To encourage interdisciplinary research by engaging the artificial intelligence (AI) and computer vision communities.

Main Methods:

• Utilized nonlinear force-free magnetic field (NLFFF) extrapolation techniques.
• Employed vector magnetogram data from the Helioseismic and Magnetic Imager (HMI) instrument aboard the Solar Dynamics Observatory (SDO).
• Downloaded Space-weather HMI Active Region Patches (SHARPs) data, including NOAA active region numbers, at 96-minute intervals.

Main Results:

• Generated a large-scale dataset of 3D solar magnetic fields for active regions.
• Each data sample is annotated with detailed labels for solar flare forecasting.
• The dataset offers high spatial-temporal resolution and data quality.

Conclusions:

• The open availability of this dataset and source code supports the scientific community.
• This resource is expected to foster new research avenues in AI for astronomy, particularly in analyzing large-scale solar datasets.
• The dataset serves as a valuable tool for improving solar flare prediction models.