Volcanic eruptions caused by seismic shaking


Magmas are subject to dynamic changes in pressure in the Earth’s crust. These changes can be rapid, such as during earthquakes, or relatively slow, such as during magma convection or ascent in conduits. Multiphase magmas – crystal mushes and magmatic foams – can be destabilised when pressures change or oscillate rapidly. Foam collapse (in basalt systems) and mush remobilization (in silicic systems) are both thought to be key triggers for volcanic eruptions.

Magma mush is a mixture of ≳60 vol.% crystals with an interstitial melt. The high crystal content ‘locks’ the magma, making it immobile, and therefore uneruptible. However, recent work has shown that even a small perturbation can unlock a magma mush and remobilize it, for example when even a small fraction of bubbles grow in the melt, pushing the crystals apart from one another (Truby, 2016). Unlocking magma mush is thought to be a key triggering process for subsequent, large, silicic volcanic eruptions (see, for example, Cashman et al., 2017).

The student will test the extent to which crystal-bearing magmas can be remobilized when a pressure wave travels through the system. Transient oscillations of pressure are common in large seismic events, and may be of sufficient amplitude to (1) reorganise mush and trigger melt-crystal separation, and (2) rapidly nucleate and grow bubbles in the melt between crystals, unlocking the system. As yet, this process remains untested.

This project will adopt both numerical and experimental approaches to answer these fundamental questions:
1. Can a transient pressure wave of seismic characteristic frequency, amplitude and duration, result in nucleation of bubbles in super-saturated melt?
2. Is the nucleation and subsequent growth of seismically induced bubbles sufficient to unlock a magma mush?

Well known seismically triggered eruptions will be used as case studies to inform the research, and demonstrate application.


The student will use a combined numerical and scaled experimental approach. They will have access to recently developed numerical tools at Durham University for disequilibrium bubble growth in magmas, and to a host of models for the physical and chemical properties of magmas of all compositions. These models will provide a starting test bed for exploring the effects of pressure oscillation on a system in which gas exsolution and bubble growth are important. They will also apply models for mush convection and mobilization (e.g. Roper et al., 2007).

They will build, test and use a new scaled experimental apparatus for applying pressure oscillations of controlled frequency and amplitude to a body of liquid containing solid particles – an experimental analogue of a real magma mush. The properties (such as viscosity, and density ratio) of this analogue system will be carefully scaled to natural scenarios, such that the same characteristic physical processes apply. A gas phase will be dissolved in the liquid by pressurising the system prior to oscillation, such that, during each pressure cycle, the gas phase switches from exsolving to dissolving. This simulates the physical effect of a disequilibrium gas phase on mush reorganisation in natural magmatic systems and introduces a compressibility that may unlock the mush system.

The experimental materials will be designed such that they cover a range of liquid viscosities, crystal (particle) volume fractions, and potential gas contents. The oscillations will be varied so that they cover a range of plausible scaled frequencies and amplitudes of pressure.

Depending on the aptitude and preference of the student, more or less emphasis can be placed on either the numerical or the experimental campaigns. The below timeline is therefore advisory and an example possibility.

Project Timeline

Year 1

Experimental campaign: Together with the supervisory team, the student will design, build, and test a rig for applying pressure oscillations to 2-phase and 3-phase suspensions. Careful scaling analysis will be undertaken to ensure applicability to natural magma in crustal reservoirs.
Numerical modelling: The student will integrate an existing numerical code for the growth and resorption of bubbles with a model for rheology of three-phase magma. The enhanced numerical model is likely to form the basis of a first output manuscript.

Year 2

Experimental and numerical campaign: The student will conduct scaled experiments using the new rig, varying parameters across the natural range. Results will be cast as a regime map, which shows the conditions under which pressure oscillations can cause mush to mobilize. The results will be used to parameterize and validate the numerical model. The combined experimental-numerical study is expected to form the basis of a second manuscript output.

Year 3

Experimental and numerical modelling results will be applied to natural pressure-time histories for a range of pressure waveforms, which particular application to case study volcanic eruptions. The case study investigation is expected to form the basis of the third output manuscript.

Year 3.5

Results will be synthesized as a thesis.

& Skills

The student will receive detailed training in all areas of physical volcanology, but in particular in:
• Experimental design, dimensional analysis, and scaling from experiments to natural conditions. These training points will be essential for performing well-posed experiments.
• Simple manipulation of seismic datasets to extract frequencies and amplitudes of ground motion at depth.
• Simple numerical methods for using and adapting existing codes for bubble growth in super-saturated magmas.

The student will join the Durham Volcanology Group; a dynamic collaborative research environment with 7 staff and approximately 20 postdocs and postgraduate students.

References & further reading

• Manga, M. and Brodsky, E., 2006. Seismic triggering of eruptions in the far field: Volcanoes and geysers. Annu. Rev. Earth Planet. Sci, 34, pp.263-291.
• Roper, S.M., Davis, S.H. and Voorhees, P.W., 2007. Convection in a mushy zone forced by sidewall heat losses. Metallurgical and Materials Transactions A, 38(5), pp.1069-1079.
• Truby, 2016. PhD Thesis supervised by Ed Llewellin: http://etheses.dur.ac.uk/11697/
• Cashman, K., Sparks, R.S.J., Blundy, J.D., 2017. Science. Doi: 10.1126/science.aag3055.

Further Information

Apply Now