European Commission logo
English English
CORDIS - EU research results

Dead or Alive: Finding the Origin of Caldera Unrest using Magma Reservoir Models

Periodic Reporting for period 1 - DEFORM (Dead or Alive: Finding the Origin of Caldera Unrest using Magma Reservoir Models)

Reporting period: 2020-11-01 to 2022-10-31

Magmatic systems in the Earth’s mantle and crust contain multiple phases, including solid crystals, liquid melt and low viscosity fluids, at different proportions. However, the theories underpinning most physical models of magmatic systems describe magma as a single-phase fluid, or as two-phase solid-dominated or liquid-dominated mixtures. Connections among these flow regimes are poorly established, hindering our understanding of magma mushes at intermediate phase fractions, which is the perceived state of large crustal magma bodies. To address this knowledge gap, this project establishes a numerical model to study the mechanics of any multi-phase magma mixture. On a fundamental level, this model transcends the theoretical limitations of existing models and expands the range of possible volcanological investigations. On a broader level, unravelling the processes in magma bodies is critical for assessing the hazard they pose to society, particularly at supervolcanoes whose eruptive impacts extend globally.
We develop a MATLAB numerical implementation of a recent theory on multi-phase magmas that can be applied to any number of phases and at any phase proportion. To ensure realistic results, transport properties (e.g. viscosity, permeability, etc.) should be calibrated to real mixtures, a task that grows in complexity with more phases. Here, we develop calibrations to mafic and silicic two-phase and three-phase mixtures. After first verifying the model against well-known results in solid-dominated and liquid-dominated mixtures, we apply the model to study mush mechanics and found that melt localisation into melt-rich lenses forms an important transport process (see attached image mush2d.jpg). The stress distribution controls the orientation of melt-rich lenses, producing vertical channels in extensional tectonic environments. We show that the time taken for melts to rise to the top of a magma body and supply supereruptions is compatible with geochronological constraints.

The numerical implementation is presented in an open-access paper accepted by Geophysical Journal International on 30 Nov 2022 ( and results have been presented at multiple Geophysics and Volcanology conferences. The numerical model is publicly available on GitHub ( and is currently being used by other researchers at ETH Zurich and the University of Glasgow. Insights gained from model analysis have been used to develop a new three-phase thermo-mechanical model under liquid-rich conditions (
This project advances our understanding into magma processing from source to surface by elucidating mush mechanics, where liquid melt preferentially concentrates into narrow bands (see attached image mush2d.jpg). Mush mechanics are a key missing component in the state-of-the-art. In particular, establishing that melt segregation and diapirism can occur on a timescale compatible with the crystal record to form supereruptions addresses a knowledge gap in the volcanological community. Like the underpinning theory, this numerical model can directly apply to any number of phases, provided suitable transport properties can be defined (three-phase example containing solid crystals, liquid melt and the magmatic volatile phase in attached image threephase_convection.jpg). This broad application enables new investigations into magmas of any number of phases and at any phase fraction, providing avenues to study the formation of magmatic iron ores, porphyry copper deposits and planetesimals.
Concentration of liquid melt into liquid-rich bands in a magma mush
Buoyant rise of a fluid-rich layer in a 3-phase magma at high liquid fraction

Related documents