Investigating sunspot and photospheric magnetic field properties using automated solar feature detection

QB Astronomy;QC Physics

Watson, Fraser Thomas ; Fletcher, Lyndsay

University:University of Glasgow

Department:School of Physics and Astronomy

关键词: Sun, sunspots, magnetic field, active regions, image processing, feature detection, morphology;   

Others  :  http://theses.gla.ac.uk/3429/1/2012watsonphd.pdf

来源: University of Glasgow

【 摘 要 】

The past few decades of solar observations have seen an increase in both the spatial and temporal resolution of data. The recent launch of the Solar Dynamics Observatory is the next step in a digital era and provides so much data that the satellite has its own Feature Finding Team tasked with creating automated detection algorithms to ease the burden on human analysis. This thesis will present some methods of automated solar feature recognition with the aim of finding a consistent method that can be reliably used on long term datasets (the Michelson Doppler Imager data from 1996-2010 will be used as the example in this thesis). We show methods for detecting sunspots in white light intensity data as well as a method for detecting magnetic fragments in magnetogram data.By applying these methods to a long term dataset we build a sunspot catalogue which is then used to investigate the evolution of sunspot properties over solar cycle 23. We find that the International Sunpot Number does not accurately represent the number of sunspots present on the visible solar disk although the trend does follow the number of sunspots. We also find that the umbral area of sunspots is between 20 and 40% of the total sunspot area and that this exhibits smooth variation over the solar cycle indicating there may be some change in how sunspots are formed at different points in the cycle. We then use the catalogue to investigate the Wilson depression effect and use Monte Carlo simulations along with sunspot models to show that the tau = 1 layer of the photosphere is recessed by 500-1000 km inside sunspots. Next, we examine the magnetic fields inside sunspot umbrae to investigate claims of a long term secular decrease in sunspot magnetic fields that could point to a long term solar minimum spanning many cycles. We do not see evidence of this decrease although we only analyse one cycle of data.Next, five active regions are analysed using an automated magnetic fragment detection and tracking algorithm. We also examine quiet Sun magnetic fields and note that at field strengths of 5 Gauss from the HMI/SDO instrument, the orbital motion of the satellite can be detected as a fluctuation in the measured magnetic field strength with the period of a satellite in geosynchronous orbit. We also calculate the diffusion and drift velocities of fragments in three of the observed active regions and find that our diffusion coefficients are higher than previous studies but our drift speeds are lower than those from the same studies.