Investigating the Efficacy of Topologically Derived Time-Series for Flare Forecasting. I. Dataset Preparation

TL;DR

This paper introduces a new dataset of topological flux time-series derived from magnetogram data, showing potential for improved solar flare prediction by analyzing spatial and temporal correlations with eruptive activity.

Contribution

It develops and publicly releases a comprehensive dataset of helicity and winding flux time-series from 144 active regions, highlighting their predictive potential for solar flares.

Findings

Time-series signals often peak 1-8 hours before flares.

Spatial correlations link flux signals with eruptive activity.

The dataset enables new flare prediction approaches.

Abstract

The accurate forecasting of solar flares is considered a key goal within the solar physics and space weather communities. There is significant potential for flare prediction to be improved by incorporating topological fluxes of magnetogram datasets, without the need to invoke three-dimensional magnetic field extrapolations. Topological quantities such as magnetic helicity and magnetic winding have shown significant potential towards this aim, and provide spatio-temporal information about the complexity of active region magnetic fields. This study develops time-series that are derived from the spatial fluxes of helicity and winding that show significant potential for solar flare prediction. It is demonstrated that time-series signals, which correlate with flare onset times, also exhibit clear spatial correlations with eruptive activity; establishing a potential causal relationship. A significant database of helicity and winding fluxes and associated time series across 144 active regions is generated using SHARP data processed with the ARTop code that forms the basis of the time-series and spatial investigations conducted here. We find that a number of time-series in this dataset often exhibit extremal signals that occur 1-8 hours before a flare. This, publicly available, living dataset will allow users to incorporate these data into their own flare prediction algorithms.