An Examination of Observed Coronal Structures through Automated Techniques Across Solar Cycle 24

Abstract

The solar corona is an intricate and magnetically dominated plasma environment consisting of closed (eg. active region loops) and open (eg. polar plumes) magnetic structures. Although it is a highly studied portion of the solar atmosphere, little research has been conducted to investigate how the basic nature of coronal structures could vary across an entire solar cycle. While there are numerous studies of specific targeted coronal events, the research outlined in this thesis seeks to broaden the body of evidence by xystematically capturing and then analysing a range of coronal structural parameters throughout solar maximum and minimums. The theoretical framework for understanding the parameter distributions of coronal structure was a self-organized criticality (SOC) approach, which predicts certain power law distributions of events given the presence of a non-linear, stochastic, critical process which can be described by the sum of interactions of small-scale elements (such as is expected to be the case for D.C. heating scenarios in the corona). The thesis work involved the development of a modular and scientifically robust algorithm for determining structural details that can be employed upon legacy, current, and future datasets. In this case, COSDA (COronal Structural Detection and Analysis) has been used on four Extreme Ultraviolet (EUV) wavelengths (171, 193, 211 and 304 ˚A) of the Solar Dynamics Observatory (SDO) Atmospheric Imaging Assembly (AIA) from across Solar Cycle 24 and four EUV wavelengths (171, 195, 285 and 304 ˚A) of the Solar and Heliospheric Observatory’s (SOHO) Extreme ultraviolet Imaging Telescope (EIT) across Solar Cycle 23. Closed magnetic regions (loops) were examined for both cycles 23 and 24 while open, polar structures (plumes) were also examined using the SDO/AIA datasets. From these investigations, width and positional information was obtained from tens of thousands of coronal structures from within a specially constructed annulus used to control for the effects of differential rotation and image noise.

Width analysis demonstrated the existence of a power law of coronal structural width distributions, whose gradient varied depending on the wavelength, as well as the phase of Cycle 24 in which they were observed. These power law gradients, which ranged from -1.6 to -5.5 in the closed loops and -2.2 to -4.9 in open plumes demonstrated both a changing value depending on the period throughout the solar cycle and the observed wavelength, with the most well populated displaying a resemblance to expected SOC behaviour (-1.5 gradient). These variations may be indicative of change in observational structure formation which are analogous to SOC statistical modelling, such as the ability for reconnection to perturb neighbouring strands being similar to quenching, or varying energy addition mechanisms being more or less stochastic. These results could aid in models of coronal heating and loops/structures by constraining the variables such as width distributions and power law slopes that models would need to be able to replicate for their simulations to be physically accurate.

Coronal structure latitude distributions across Cycle 24 reveal a complex North-South asymmetry. There are more structures detected in the northern hemisphere compared to the southern hemisphere (between 63 to 45 percent of total for closed active region and 68 to 11 percent of open polar regions) while asymmetry indices (defined as the number of northern minus southern structures divided by the total number of structures) lies between 0.26 to 0.45 (closed) and 0.36 to -0.09 (open). Latitude distributions show evidence of two distinct functions of structure latitude versus time for active region loops, as well as latitudinal stratification for plume populations by wavelength. Across Cycle 23 as seen in EIT, the latitudes of locally normalized coronal intensity peaks generally follow distributions seen in Cycle 24 in closed active and open polar regions, though these profiles were more difficult to discern due to noise, and width distributions were unable to be recovered due to resolution limitations. This indicates that certain latitudinal trends of structure
distribution may be characteristic of coronal structure distributions throughout at least two solar cycles.

Future work detailing improvements to the COSDA algorithm, machine learning approaches to data processing, and integrating upcoming datasets such as the EUI of the Solar Orbiter into the body of research are also outlined.