Petrologic monitoring at Volcán de Fuego, Guatemala

Abstract

Paroxysmal activity represents an end-member in the common range of activity at mafic arc volcanoes, characterised by rapid transitions across the effusive-explosive interface and thus posing significant challenges to hazard assessment. Conceptual models to explain changes in the frequency and magnitude of these paroxysmal events are based either on magma recharge or an increase in gas flux, largely framed in the context of two-phase flow. Gas- and magma-driven models are both viable mechanisms to explain the varying styles of paroxysmal behaviour observed in mafic systems; however, each has different implications for future activity. We present time series petrologic data for ash and lava samples collected at Volcán de Fuego, Guatemala, during paroxysmal eruptions between 2011 and 2018. We show that a step-change in glass composition occurred between 2015 and 2016, reflecting an increase in magma temperature and a reduction in pre-eruptive crystallisation, concurrent with an escalation in the frequency of paroxysmal activity. There was no change in the bulk or phase compositions during this period. To explain these observations, we propose that the increase in frequency of paroxysmal eruptions is modulated by the supply of exsolved volatiles from lower crustal degassing magmas, without invoking repeated transfer of new, primitive magma to a shallow reservoir. Protracted lava effusion, accompanied by more vigorous and more frequent Strombolian explosions and gas ‘chugging’, prior to the transition to sustained fountaining suggests that gas retention in crystal-rich magma may modulate the height of the magma column as gas supply increases.Slow decompression associated with effusion may determine the timing of effusive to explosive transitions in mafic arc systems more generally. A large paroxysmal eruption of Fuego on 3 June 2018, notable for the rapid escalation in eruptive intensity several hours into the eruption, produced ash with a range of textures and glass compositions consistent with magma evacuation over a range of depths and decompression rates. Given the protracted repose time between paroxysms before this event, we suggest that a shallow crystallised plug degraded, and ultimately failed, several hours into the eruption of 3 June 2018, triggering top-down decompression of magma in the conduit synchronous with the observed rapid acceleration in eruption rate. Ultimately, we propose that the frequency of paroxysms at Fuego is broadly proportional to the gas supply rate, while the range in glass compositions is related to the repose time prior to eruptive activity. Our data illustrate the potential of petrologic monitoring to distinguish between gas- and magma-driven paroxysm triggers and to anticipate future events, especially when interpreted in the context of geophysical observations and implemented within community-based ash collection initiatives.