Active Region Coincidence Considerations in Solar Flare Prediction: Solar Cycle 24, Peak Years Study

In this work, active region (AR) coincidence effects on Spaceweather HMI Active Region Patch (SHARP) parameters (which may be valuable for flare prediction) are considered. Focus is on the peak years (2013 and 2014) of Solar Cycle 24. Here, coincidence refers to the occurrence of at least two solar flares within the same active region within 24 hours of one another. We term any flare that occurs in an AR, R, less than 24 hours later than another flare that also occurred in region R a coincident flare; we term any other flare a non-coincident flare.

Observations of SHARP parameter differences for coincident versus non-coincident B, C, M, and X class flares for these years are reported in this work, especially via consideration of time series values for the 24 hours prior to flare onset. Onset times were determined from a catalog of B, C, M, and X class solar flares for 2013-2014, which included their start time and a label of their X-ray class. Several SHARP parameters exhibit notable differences; in particular, the M/X class parameter values for coincident flares often diverge from the ones for non-coincident flares.

The effect of these differences on one flare prediction mode (i.e., prediction of an M/X flare based on 24-hour time series values for key SHARP parameters) is also considered here using thresholding-based predictors. Prediction success metrics are key reports.

A key finding is that methods for prediction of M/X class coincident flares are likely to be divergent from ones for prediction of M/X class non-coincident flares.