Thermomechanical modelling of the shallow magmatic body at Krafla

The intersection of molten rhyolitic magma at 2.1 km depth by the IDDP-1 well in 2009 revealed the presence of an active shallow magmatic system beneath the Krafla caldera (NE Iceland). This unexpected finding challenged the preexistent belief of rapid solidification of such bodies at shallow depths. The present thesis develops and applies a novel thermomechanical numerical framework to investigate the internal dynamics, longevity, and thermal evolution of this magmatic pocket. The numerical algorithm couples the steady, partially compressible Stokes flow equations with the heat transport equation, explicitly incorporating latent heat release during change of phase. Magma is treated as a multicomponent, single-phase mixture of melt, crystals, and gas, where the bulk thermomechanical properties—density, viscosity, compressibility, heat capacity, and thermal conductivity—evolve self-consistently with local temperature, pressure, and composition. The magma–host-rock coupling is realized through conductive heat flux continuity at the chamber boundaries. The algorithm is implemented within the finite-element software GALES, enabling parallelized two-dimensional simulations in complex geometries. Validation against classical convection and heat transfer benchmarks demonstrates the robustness and numerical stability of the scheme. Applications to the Krafla system are based on the rhyolitic composition of the retrieved glass cuttings and physical parameters constrained by IDDP-1 glass cuttings and in situ conditions (initial temperature ≈ 900 °C, pressure ≈ 45 MPa). Simulations explore three chamber sizes and a range of volatile contents (H₂O–CO₂ mixtures from 1.7 to 2.5 wt% water and 85 ppm to 1.1 wt% carbon dioxide). The modeled systems initially undergo a smooth conductive–convective cooling regime, followed—after approximately 50 years—by the emergence of a quasi-stationary, oscillatory behavior characterized by alternating cooling and heating episodes. These thermal oscillations arise from the latent heat released during crystallization, which partially remelts existing crystals and reabsorbs exsolved volatiles. This feedback delays solidification and sustains a largely molten state for several hundred years. Convection plays a fundamental role in redistributing heat and material within the chamber. Vigorous internal circulation detaches protomush accumulations from the margins and transports them toward the hotter core, where they are remobilized and partially reabsorbed. The combined effect of latent heat buffering and convective remixing prevents the development of a stable crystal mush at the magma–rock interface. This mechanism provides a natural explanation for the absence of a gradational crystallinity transition in the IDDP-1 borehole and the direct encounter with molten magma. Parametric variations in chamber size, pressure, and volatile content alter cooling rates but do not suppress the stepped cooling regime, indicating that the feedback between crystallization and latent heat release is the principal control on chamber longevity. The results support the hypothesis that small, shallow rhyolitic magma pockets can remain molten over century- to millennial-long timescales, consistent with either 18th-century or 20th-century emplacement of the Krafla magma. The model thus reconciles geophysical observations, petrological constraints, and drilling results within a unified thermomechanical framework. Ultimately, this work establishes a robust numerical tool for investigating the coupled thermal, mechanical, and compositional evolution of shallow magmatic systems, contributing to the preparation of forthcoming projects such as the Krafla Magma Testbed, the world’s first in-situ magma observatory.