A petrological model for use in thermo-chemical-mechanical models of magmatic system dynamics
Tobias Keller 1
Affiliations: 1 University of Glasgow, School of Geographical and Earth Sciences, Glasgow G12 8RZ, Scotland (Tobias.Keller@glasgow.ac.uk)
Presentation type: Poster
Presentation time: Tuesday 16:30 - 18:30, Room Poster Hall
Poster Board Number: 216
Programme No: 3.2.39
Abstract
Understanding magmatic system dynamics is challenging due to limited direct observations of subsurface processes. Numerical modelling provides a means to interpret indirect petrological and geochemical observations on igneous rocks. Central to magma dynamics is the interplay of complex multiphase fluid mechanics and multicomponent thermochemistry. Models of chemical thermodynamics require costly Gibbs free energy minimisation in systems typically comprising several dozen thermodynamic components. Existing algorithms lack the efficiency and robustness for on-the-fly tracking of petrological evolution in coupled thermos-chemical-mechanical models. Previous coupled modelling approaches have relied on simplified phase relations such as a single phase loop or precomputed lookup tables. Both approaches have strong limitations. Here, I introduce an alternative petrological model that generates multi-dimensional pseudo-phase diagrams over a set of pseudo-components in P-T-X space. This model avoids both expensive energy minimisation as well as oversimplification of phase relations and cumbersome lookup tables but preserves sufficient complexity to capture key trends in petrological evolution such as fractional crystallisation. Model calibration against experimental data and full thermodynamic model results, achieved via standard machine learning techniques, highlights the principal axes of variability, typically revealing 5-6 dominant components correlated to the appearance of a similar number of major phases on the liquidus. Consequently, the model offers significant dimensional reduction compared to full thermodynamic models while preserving critical petrological insights. The calibrated model can then be used to robustly and efficiently track the dynamic evolution of major mineral and melt phases and their major element compositions over large segments of P-T-X space in coupled models.