Abstract
Volcanic eruptions are driven by the growth of gas bubbles in magma. The timing and rate of bubble growth are important because they determine whether enough gas pressure can develop to fragment the melt. Bubbles grow in response to decompression and diffusive transport of dissolved volatiles (predominantly H2O) that exsolve into the bubbles. Growth is resisted by the viscosity of the melt. Both melt viscosity and H2O diffusivity have non-linear dependence on the concentration of H2O dissolved in the melt, which necessitates a numerical approach to modelling bubble growth. Several bubble growth models have previously been published and applied, but none of them has been validated against continuous, in situ experimental data or provided as a user-friendly tool. Here we present a numerical bubble growth model, implemented in MATLAB, which allows for arbitrary temperature and pressure pathways, and accounts for the impact of spatial variations in dissolved H2O concentration on viscosity and diffusivity. We validate the model against two sets of experimental data: (1) New continuous data for gas-volume fraction as a function of time, collected using optical dilatometry of vesiculating hydrous obsidian samples which were heated from 930 °C to 1000 °C at atmospheric pressure. This dataset captures isobaric, isothermal bubble growth under strongly disequilibrium conditions. (2) Discrete data from published decompression experiments at 825 °C and pressures from 200 MPa to ~5 MPa with decompression rates from 0.1 MPa s−1 to 10 MPa s−1. These experiments represent isothermal, decompression-driven bubble growth spanning equilibrium to strongly disequilibrium conditions. The numerical model closely reproduces the experimental data across all conditions, providing validation against contrasting bubble growth scenarios. The validated model has a wide range of potential volcanological applications, including forward modelling of bubble growth phenomena, and inverse modelling to reconstruct pressure–temperature–time pathways from textures and volatile contents of eruptive products.
Highlights
Bubble growth drives magma ascent in the shallow crust
Continuous data provide a strong test of the bubble growth model because: (1) they contain a much larger number of individual φ(t) datapoints than conventional discontinuous datasets from high-pressure, high-temperature (HPHT) experiments; (2) a single sample can be tracked through all stages of bubble growth; and (3) the dataset is not complicated by poorly-constrained resorption processes during quench (McIntosh et al, 2014)
Sample growth – which corresponds to bubble growth – is slow initially, accelerates as temperature increases and melt viscosity drops, before slowing again as the sample approaches equilibrium saturation at atmospheric pressure and the experimental dwell temperature
Summary
Bubble growth drives magma ascent in the shallow crust. The rate of bubble growth is an important control on whether an eruption is effusive or explosive, because rapid growth can lead to magma fragmentation (Degruyter et al, 2012; McBirney and Murase, 1970; Sparks, 1978; Verhoogen, 1951). Accurate modelling of bubble growth processes is necessary for the investigation of many physical volcanological problems. Explicit forward-modelling of bubble growth remains challenging because it involves several non-linear and coupled physico-chemical processes (Blower et al, 2001; Liu and Zhang, 2000; Lyakhovsky et al, 1996; Navon et al, 1998; Proussevitch and Sahagian, 1998; Sparks, 1978); a numerical approach is required to model bubble growth accurately. The pressure and temperature changes that drive and modify bubble growth may vary in a complex way as a packet of magma moves through the volcanic system from crustal reservoir to final emplacement (Carey et al, 2013; McIntosh et al, 2014). A model should be flexible, accurate, and numerically stable over nonlinear temperature and pressure pathways
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
