Predicting Coronal Mass Ejections Using Machine Learning Methods

TL;DR

This study employs machine learning to identify features that distinguish solar flares associated with coronal mass ejections from those that are not, using magnetic field and X-ray data to improve CME prediction accuracy.

Contribution

It introduces a novel approach using physically meaningful features and machine learning to forecast CMEs associated with solar flares, achieving high predictive skill.

Findings

True Skill Statistic of ~0.8 indicates strong predictive performance

Six key parameters effectively capture relevant magnetic field information

Machine learning can distinguish CME-producing flares from non-CME flares

Abstract

Of all the activity observed on the Sun, two of the most energetic events are flares and Coronal Mass Ejections (CMEs). Usually, solar active regions that produce large flares will also produce a CME, but this is not always true (Yashiro et al., 2005). Despite advances in numerical modeling, it is still unclear which circumstances will produce a CME (Webb & Howard, 2012). Therefore, it is worthwhile to empirically determine which features distinguish flares associated with CMEs from flares that are not. At this time, no extensive study has used physically meaningful features of active regions to distinguish between these two populations. As such, we attempt to do so by using features derived from [1] photospheric vector magnetic field data taken by the Solar Dynamics Observatory's Helioseismic and Magnetic Imager instrument and [2] X-ray flux data from the Geostationary Operational Environmental Satellite's X-ray Flux instrument. We build a catalog of active regions that either produced both a flare and a CME (the positive class) or simply a flare (the negative class). We then use machine-learning algorithms to [1] determine which features distinguish these two populations, and [2] forecast whether an active region that produces an M- or X-class flare will also produce a CME. We compute the True Skill Statistic, a forecast verification metric, and find that it is a relatively high value of approximately 0.8 plus or minus 0.2. We conclude that a combination of six parameters, which are all intensive in nature, will capture most of the relevant information contained in the photospheric magnetic field.