Magnetic Characteristics of Sunspot Groups and their Role in Producing Adverse Space Weather
Citation:
MCCLOSKEY, AOIFE ELIZABETH, Magnetic Characteristics of Sunspot Groups and their Role in Producing Adverse Space Weather, Trinity College Dublin.School of Physics, 2019Download Item:
Abstract:
The energy that powers solar flares is known to come from magnetic energy stored in sunspot groups, but the precise conditions required and processes involved in energy release remain unclear. The likelihood of flaring is directly related to the complexity of the magnetic field, such that larger, more complex sunspot groups produce larger more frequent flares than smaller, simple sunspot groups. The research in this thesis aims to investigate the evolution of sunspot complexity to gain greater understanding of how sunspot groups produce flares, with the intention of improving upon current methods of flare prediction.
Firstly, the temporal evolution of sunspot groups in terms of their white-light structural complexity (McIntosh classification) was examined during Solar Cycle 22 (SC22). It was found that over a 24-hr period the majority,??60%, of sunspot groups do not evolve in terms of the parameters described by the McIntosh system. For the first time, evolution-dependent 24-hr flaring rates were calculated and higher flaring rates were observed for practically all McIntosh classes for upward evolution with opposite behaviour observed for downward evolution (McCloskey et al, 2016). These results agree qualitatively with previous studies relating sunspot area change (i.e., flux emergence/decay) and penumbral asymmetry (i.e., higher degree of magnetic field twisting) to flare production. Following on from this, flaring rates calculated from sunspot-group evolution in the full McIntosh classification were used to develop a new forecasting method under the assumption of Poisson statistics(McCloskey et al, 2018). It was found that the new evolution-dependent method outperforms the older, point-in-time Poisson method. However, a tendency for over-forecasting was found for both of the Poisson-based methods when using SC22 training data applied to the testing period of SC23. A comparison of Cycle-to-Cycle flaring rates was carried out, with results indicating that the flaring statistics between cycles is different (i.e., SC23 shows decreased activity compared to SC22).
Machine-learning techniques were then employed on the evolution of sunspot group McIntosh classifications for the first time, to construct new flare forecasting models. A direct comparison to the evolution-based McIntosh-Poisson method was carried out and it was shown that the majority of these machine-learning models surpassed its performance. However, a similar tendency for over-forecasting was found, strengthening the findings that the flaring activity between Solar Cycles is decreasing. Finally, the relative importance of the McIntosh classification components was then examined and it was found that the most important features were the end-state compactness (i.e., a proxy for total magnetic flux), with the next most important features being either the end-state modified Zurich class or the starting compactness class. These combinations of pairings indicate that either the end-state total magnetic flux or the emergence of flux within the interior of a sunspot group (independent of group size) provides the largest contribution to predicting whether a sunspot group will be flare-productive or not.
The novel research carried out in this thesis has explored the temporal evolution of sunspot group classifications and their relation to solar flare production. The collection of results presented here has succeeded in quantifying this relationship for the first time, showing that evolution in sunspot-group structural characteristics are inherently linked to flare productivity. Future research that attempts to provide further understanding of this relationship will be crucial for the continued improvement of flare forecasting models and their utility for the space weather forecasting community as a whole.
Sponsor
Grant Number
Irish Research Council (IRC)
Author's Homepage:
https://tcdlocalportal.tcd.ie/pls/EnterApex/f?p=800:71:0::::P71_USERNAME:MCCLOSKADescription:
APPROVED
Author: MCCLOSKEY, AOIFE ELIZABETH
Publisher:
Trinity College Dublin. School of Physics. Discipline of PhysicsType of material:
ThesisAvailability:
Full text availableKeywords:
Solar Physics, Astrophysics, Space Weather, Solar Flare, Forecasting, SunspotsMetadata
Show full item recordLicences: