RESUMO
Recent gas flux measurements have shown that Strombolian explosions are often followed by periods of elevated flux, or "gas codas," with durations of order a minute. Here we present UV camera data from 200 events recorded at Stromboli volcano to constrain the nature of these codas for the first time, providing estimates for combined explosion plus coda SO2 masses of ≈18-225 kg. Numerical simulations of gas slug ascent show that substantial proportions of the initial gas mass can be distributed into a train of "daughter bubbles" released from the base of the slug, which we suggest, generate the codas, on bursting at the surface. This process could also cause transitioning of slugs into cap bubbles, significantly reducing explosivity. This study is the first attempt to combine high temporal resolution gas flux data with numerical simulations of conduit gas flow to investigate volcanic degassing dynamics.
RESUMO
An amendment to this paper has been published and can be accessed via a link at the top of the paper.
RESUMO
Volcanoes with multiple summit vents present a methodological challenge for determining vent-specific gas emissions. Here, using a novel approach combining multiple ultraviolet cameras with synchronous aerial measurements, we calculate vent-specific gas compositions and fluxes for Stromboli volcano. Emissions from vent areas are spatially heterogeneous in composition and emission rate, with the central vent area dominating passive emissions, despite exhibiting the least explosive behaviour. Vents exhibiting Strombolian explosions emit low to negligible passive fluxes and are CO2-dominated, even during passive degassing. We propose a model for the conduit system based on contrasting rheological properties between vent areas. Our methodology has advantages for resolving contrasting outgassing dynamics given that measured bulk plume compositions are often intermediate between those of the distinct vent areas. We therefore emphasise the need for a vent-specific approach at multi-vent volcanoes and suggest that our approach could provide a transformative advance in volcano monitoring applications.