Assessing Magma Storage and Ascent Records from Volatiles in Olivine-Hosted Melt Inclusions

Magma storage depths and ascent rates are challenging quantities to constrain but are important for understanding eruption dynamics. Mineral-hosted melt inclusions (MI) can preserve a unique record of magma storage conditions and the timescales of magma ascent. Volatile contents (H₂O, CO₂) of glass in olivine-hosted MI are often used to assess magma storage depths; however, it is well established that MI can rapidly lose H₂O by diffusion during ascent. This dehydration, in combination with other post-entrapment processes (e.g., crystallization), decreases CO₂ solubility in the MI, resulting in the formation of a CO₂-rich vapor bubble (VB). These factors must be accounted for when assessing storage depths. In combination with accurately determined depths, the time dependence of diffusive hydrogen/deuterium fractionation (δD_VSMOW) associated with dehydration can be used to investigate magma ascent rates.

We determined magma storage depths and ascent rates using glassy olivine-hosted MI in tephras and hyaloclastites from five volcanic centers (Mt. Bird, Mt. Terror, Mt. Erebus, Hut Point Peninsula, Mt. Morning) in the Erebus volcanic province (Antarctica). We determined volatiles in MI glass (H₂O, CO₂, δD_VSMOW) by secondary ion mass spectrometry and the CO₂ content of VB by Raman spectroscopy (CO₂ density) and 3D x-ray tomography + traditional petrography (MI, VB volumes). Total MI (glass + VB) CO₂ contents were calculated from mass balance. Magma storage depths were derived from H₂O-CO₂ vapor solubility, using maximum H₂O and reconstructed CO₂ contents. A finite difference model for MI dehydration was used to determine magma ascent times from δD_VSMOW.

We find that: (a) volatile concentrations in MI glass from all samples cover a wide range (~0.5-2.5 wt. % H₂O, ~0.2-1.2 wt% CO₂), (b) VB contain a significant proportion of the total MI CO₂ (tens of %); total MI CO₂ contents range to significantly higher values (e.g., 1.4 wt. % in Hut Point Peninsula), (c) MI from all locations show evidence of dehydration (δD_VSMOW: ~0 to -100), (d) max storage pressures reach as deep as the seismically determined Moho beneath Ross Island (~24 km), (e) MI H₂O contents are not related to the presence/absence of amphibole, (f) ascent rates modeled from storage depths and ascent times suggest closed-system degassing and rates of 0.5-2.0 m/s.