Development of Automated methods for the detection of solar vortex flows

Abstract

The Sun’s surface is very dynamic, even on periods of weak solar activity, producing a variety of magnetic structures, amongst them structures that can cause plasma to exhibit vortex-like motions. These so-called solar vortex flows have been widely observed on the Sun’s photosphere and chromosphere as well as on upper solar atmospheric levels. Their existence is highly connected with the creation of magnetohydrodynamic (MHD) waves, a primary candidate for channeling energy and plasma from the photosphere up to the low corona. Therefore, the automatic detection of solar vortex flows is fundamental for proper investigation of the lower solar atmosphere dynamics and the estimation of upwards energy transfer. Previous detection approaches utilized a criterion, like vorticity, based on the frequently inconsistent and often erroneous derivation of the velocity field from the solar observations. The aim of the present thesis is the implementation of an automated methodology for the detection of solar vortex flows that is based on the morphological characteristics they present.

The approach followed is the application of a series of pre-processing methods aiming to remove non-dynamic or almost linear structures not related to vortex flows from a solar data set (consisting of a sequence of frames depicting a specific solar region). Such methods include the Robust Principal Component Analysis (RPCA), in order to retrieve the dynamic parts of the solar data as well as the removal of components that relate to almost linear structures and thus are not part of a vortex flow. Subsequently, the output produced by the sequential application of the pre-processing methods is aggregated and evaluated so that the primary candidate region for vortex flow existence is detected.

The evaluation of the produced results on an actual solar data set proves that the proposed methodology succeeds in identifying the primary solar vortex flow region. More importantly, a promising analysis of the distribution of dynamic, vortex-like structures on the observational sequence is produced that can be used, by further future extensions of the current work, for the detection of multiple structures within the investigated field of view. It is concluded that the presented methodology constitutes an efficient alternative way for the identification of solar vortex flows, especially after all proposed improvements and extensions are properly implemented.