Volcanic SO2 emissions and the abundance and distribution of exsolved gas in magma bodies

Abstract

The mass of sulfur dioxide (SO2) released by most explosive volcanic eruptions greatly exceeds the amount originally dissolved in the erupted volume of silicate melt. Several lines of evidence suggest that this discrepancy is due to the presence of SO2-bearing gas bubbles in the magma before the time of eruption. Comparison of remote sensing data for SO2 emissions with conventional petrologic estimates suggests that the discrepancy can be resolved if andesitic and dacitic magmas contain about 1–6wt% exsolved gas, equivalent to 5–30vol%, prior to eruption. Such large mass fractions of exsolved gas in pre-eruptive magma are consistent with previously published physical models in which crystallization-induced gas exsolution gradually increases overpressuring of a magma reservoir, eventually triggering an eruption. Simple mass balance models for SO2 emissions from magma bodies in which there is an upwards increasing gradient in exsolved gas mass fraction (i.e. gas-rich at the top) yield SO2 vs. eruptive volume trends that are similar to those observed for eruptions ranging in size from 0.01 to 10km3 of magma. Despite the many uncertainties involved, these patterns are consistent with the hypothesis that subvolcanic magma reservoirs are separated from volcanic vents by cupolas of gas-rich magma that supply much of the gas released in explosive eruptions. The volume fraction of exsolved gas inferred from SO2 data for the upper regions of many pre-eruptive magma bodies (∼30vol%) is similar to the percolation threshold at which gas bubbles become sufficiently interconnected to allow permeable gas flow through a bubble network. Thus this value may reflect a physical limitation on the maximum exsolved gas volume fraction that can occur at the roof zone of a magma reservoir because any additional gas would be lost by advective flow through the permeable bubble network. Andesitic and dacitic magma in crustal reservoirs are probably gas saturated due to recharge and underplating by basaltic magma that is saturated with H2O–CO2–S gas. Comparison of repose times, eruptive volumes, and basaltic magma supply rates for a spectrum of volcanic systems suggests a relatively steady state flux of S from Earth's mantle to atmosphere through many andesitic and dacitic magma systems.