VOLCANIC GAS CHEMISTRY AND THERMODYNAMIC MODELING TO DETERMINE ERUPTION TRIGGERS (Invited Presentation)

Volcanic gases compositions measured at the surface are geochemical fingerprints that track magma emplacement, differentiation, and hydrothermal interaction throughout a magma plumbing system. General relationships based largely on volatile solubility laws can serve as qualitative indications of magma movements; for example, changes in the fluxes of SO₂ and CO₂, or their relative proportions (molar CO₂/SO₂) may indicate that a volcanic system is being newly supplied by a shallower (SO₂-rich) or deeper (CO₂-rich) magma source. But, despite their wide use as qualitative diagnostic tools, no standard model exists to quantitatively interpret volcanic gas compositions with respect to related magmatic processes. The most quantitative models rely on extensive knowledge of the petrologic characteristics of the degassing magma, making them: a) difficult to apply widely; and b) less useful in forecasting eruptive behavior when the source magma is not well characterized (i.e., has not erupted yet). In this presentation I will demonstrate how the application of a general thermodynamic model to gas emission data from the 2018 eruption at Poás, Costa Rica, can be used to connect gas chemistry to subsurface magmatic process and identify a possible eruption trigger. Poás serves as an ideal example of a volcano with extensive gas monitoring but few petrological constraints. The model generates a wide range of possible gas compositions given broad constraints on magma chemistry. These synthetic gas compositions are then compared to compositions measured by Multi-GAS and UV spectrometry taken before, during, and after the eruption. The model quantitatively distinguishes magmatic and hydrothermal gases emitted at Poás and gives a range of possible volatile concentrations in undegassed Poás magma up to 2 kbar or ~6 km depth (H₂O=1-5.5 wt%; CO₂=0.05–0.29 wt%; S=0.03–0.48 wt%). The model suggests that the eruption was preceded by the formation of an S-rich hydrothermal carapace and then triggered by the rupturing of this carapace probably due to gas overpressure once the hydrothermal system became sealed. The deviation from background magmatic degassing to hydrothermally dominated degassing associated with seal formation is evident prior to the runup to eruption. This type of modeling is a first step toward a generalized model for the interpretation of volcanic gas data that could be widely used by observatories to forecast eruptive activity.