Detecting and characterising small-scale brightenings in solar imaging data

TL;DR

This paper introduces a novel method for detecting and analyzing small-scale brightenings in solar EUV imagery, enabling detailed statistical studies of these events to better understand solar atmospheric heating mechanisms.

Contribution

The paper presents a new detection and analysis technique for small-scale solar brightenings, including event tracking and characterization, applied to IRIS data, with validation on synthetic datasets.

Findings

Power-law distributions for brightness, area, and duration.

Maximum brightness detection with 6% error.

Event detection rate of approximately 3.96×10^{-4} arcsec^{-2}s^{-1}.

Abstract

Observations of small-scale brightenings in the low solar atmosphere can provide valuable constraints on possible heating/heat-transport mechanisms. We present a method for the detection and analysis of brightenings and demonstrate its application to IRIS EUV time-series imagery. The method uses band-pass filtering, adaptive thresholding and centroid tracking, and records an event's position, duration, and total/maximum brightness. Area, brightness, and position are also recorded as functions of time throughout their lifetime. Detected brightenings can fragment or merge over time, thus the number of distinct regions constituting a brightening event is recorded over time and the maximum number of regions are recorded as a simple measure of an event's coherence/complexity. A test is made on a synthetic datacube composed of a static background based on IRIS data, Poisson noise and $\approx10^4$ randomly-distributed, moving, small-scale Gaussian brightenings. Maximum brightness, total brightness, area, and duration follow power-law distributions and the results show the range over which the method can extract information. The recorded maximum brightness is a reliable measure for the brightest and most accurately detected events with an error of 6%. Area, duration, and speed are generally underestimated by 15% with an uncertainty of 20-30%. Total brightness is underestimated by 30% with an uncertainty of 30%. Applying this method to real IRIS QS data spanning 19 minutes over a 54.40"$\times$55.23" FOV yields 2997 detections. 1340 of these either remain un-fragmented or fragment to two distinct regions at least once during their lifetime equating to an event density of $3.96\times10^{-4}$arcsec$^{-2}$s$^{-1}$. The method will be used for a future large-scale statistical analysis of several QS data sets from IRIS, other EUV imagers, as well as H-$\alpha$ and visible photospheric imagery.