Cheng, I Kit; (2024) AI-Assisted Detection and Analysis of Magnetospheric Boundary Processes at Saturn. Doctoral thesis (Ph.D), UCL (University College London).

Preview

Text Cheng_PhD_Thesis.pdf - Accepted Version Download (34MB) | Preview

Abstract

This thesis explores the statistical properties and consequences of plasma processes, such as magnetic reconnection and mirror instability, in the vicinity of Saturn’s magnetospheric boundaries. A novel application of computer vision was developed for bow shock (BS) and magnetopause (MP) crossing detection in Cassini data. Examining electron bulk heating during MP crossings by the Cassini spacecraft, this research probed whether the observed heating was caused by reconnection heating under open or closed MP conditions. Events with energization occurred more often in the ‘reconnection possible’ parameter space, had better agreement with the predicted heating due to reconnection, and were locally ‘open’. The majority of events with no electron heating were observed at locally ‘closed’ MP and were in the ‘reconnection suppressed’ parameter space suggesting that magnetic reconnection is a key factor in electron energisation at the MP. Automating the detection of magnetospheric boundary crossings in data facilitates larger-scale statistical studies of plasma processes. A deep-learning approach, utilising a convolutional neural network classifier, outperformed a threshold-based method to detect true crossings, and obtained F1 scores of 92.1% ± 1.4% for BS and 84.7% ± 1.9% for MP (100% being perfect). This helped create Cassini’s most complete dataset of BS and MP crossings to date. Leveraging the augmented set of crossings, a global study of mirror mode (MM) waves in Saturn’s magnetosheath was performed to assess their properties and influence on Saturn’s magnetosphere. MM wave characteristics including scale size, growth rates and saturation time were found to be inconsistent with numerical predictions obtained through linear dispersion relation solvers but could be explained through non-linear growth and diffusion processes. The spatial distribution of MM in Saturn’s magnetosheath was qualitatively similar to the Jovian system. Large amplitude MM dips found near the MP could restrict magnetic reconnection to near 180º magnetic shear, thus impacting a key factor of particle energisation at the MP. The above findings contribute to the understanding of energy transfer processes in space plasmas from the lens of the Sun-Saturn system and the application of machine learning methods in post-mission spacecraft data analysis.

Download activity - last month

Export as JSONXMLCSV

Download activity - last 12 months

Export as XMLJSONCSV

Downloads by country - last 12 months

Export as XMLJSONCSV

Archive Staff Only

View Item