Modelling the dynamic distribution of geochemical signatures in shallow continental magma bodies

Swan, Shona (2025) Modelling the dynamic distribution of geochemical signatures in shallow continental magma bodies. MSc(R) thesis, University of Glasgow.

Abstract

Continental arcs are critical geological settings due to their role in recycling the Earth’s crust through subduction and magma generation. These regions are characterised by intense volcanic activity and host some of the world’s largest ore deposits. Extensive studies of continental arcs have produced large geochemical databases and widely accepted conceptual models. Despite this wealth of data, there is still ongoing debate within the literature on which processes are the dominant control on the diverse range of geochemical signatures observed in continental arcs. This ongoing debate is partially due to these systems being incredibly complex and their spatially and temporally inaccessible nature. This study models the uppermost sections of volcanic plumbing systems, located just below the volcanic edifices, to explore the dynamics that govern magma mixing and fractional crystallisation within these shallow magma bodies. The aim is to understand how these processes affect the geochemical signatures within shallow systems and to determine whether these signatures can be preserved over time. A twodimensional (2D) computational numerical model was employed to simulate shallow melt-rich magma bodies, tracking fluid dynamics, thermochemical evolution, and geochemical changes. Another objective of this research was to determine if two mixing end-member compositions can be reconstructed after magma mixing and fractionation. Machine learning techniques were utilised to reconstruct the initial input compositions. The results of this study show that the volatile content is the primary control of the system dynamics. Lower volatile contents led to faster crystallisation and cooling, while higher volatile contents in the recharging magma triggered vigorous convection, mixing and homogenising of the initial geochemical signatures. The machine learning analysis revealed that a single overturn event could overwrite the original geochemistry. However, it was possible to backtrack to the original geochemical signatures in the simulations without overturn. This study highlights the importance of numerical modelling for testing hypotheses about active volcanic systems. Numerical modelling combined with machine learning could help improve field sampling strategies by identifying zones where parental geochemical signatures are most likely to be preserved within a system.

• Item Type: Thesis (MSc(R))
• Qualification Level: Masters
• Subjects: Q Science > QE Geology
• Colleges/Schools: College of Science and Engineering > School of Geographical and Earth Sciences
• Supervisor's Name: Grima, Dr. Antoniette Greta, Keller, Dr. Tobias and Neill, Dr. Iain
• Date of Award: 2025
• Depositing User: Theses Team
• Unique ID: glathesis:2025-85254
• Copyright: Copyright of this thesis is held by the author.
• Date Deposited: 24 Jun 2025 15:30
• Last Modified: 24 Jun 2025 15:38
• Thesis DOI: 10.5525/gla.thesis.85254
• URI: https://theses.gla.ac.uk/id/eprint/85254