Experimental examination of the melt embayment method for determining magma decompression rate

Abstract:

The rate at which magma decompresses in transit to the Earth’s surface has a first-order control on eruption style as it dictates the time available for dissolved volatiles (e.g., H₂O, CO₂) to exsolve to a separate phase. Though decompression rate is difficult to determine, one promising technique exploits volatile concentration gradients in melt embayments as an output constraint for diffusion models. The procedure consists of finding the best-fitting diffusion profile for a measured concentration gradient by assuming parameters for the conditions of decompression and solving for the time needed to produce the measured gradient. Although this technique has significant potential and is already being applied, it has not yet been tested experimentally. The errors and limitations associated with this technique, such as the decompression rate range for which the experienced decompression rate is faithfully recorded, have not been quantified. This is critical knowledge for meaningful application of the technique.

We conducted high pressure-temperature decompression experiments with variable decompression rates to test whether a) diffusion modeling yields comparable dP/dt and b) the technique is applicable to a broad range of dP/dt. The experiments used both powdered and obsidian core starting material of Güney Daği rhyolite, Turkey (76.5 wt. % SiO₂ [1]). The obsidian core served as a location for homogeneous bubble nucleation in order to provide a second measurement of decompression rate within each capsule with bubble number density. Artificial embayments were examined using single-bore alumina tubing, which eliminates variations in geometry between experiments as a variable. Embayment diameters, as dictated by the alumina tubes, are 500 μm and 300 μm, permitting us to evaluate the effect of scaling on modeling results. Experiments were hydrated at 200 MPa and 800°C for 5 days to ensure full saturation of the obsidian core, then decompressed at variable rates (0.01-1 MPa/s) to 50 MPa and rapidly quenched. Raman mapping of experiments will provide an output constraint for 2D and 1D diffusion models, and the modeled timescale will be compared to the known imposed decompression timescale.

[1] Mourtada-Bonnefoi and Laporte (2002) J. Geophyis. Res-Sol. Ea. 107.