Experimental analysis of bubble-driven magma motion in the conduit, for persistently active, open-vent volcanoes

Abstract

This study uses analogue experiments to understand the role of bubbles in inducing conduit convection, for persistently degassing lava lake systems. To do so, air was fluxed through an initially stagnant column of liquid and the resulting return flow was measured. The dynamics suggested by the experimental results is compared to that of quiescent, persistently active volcanoes, with a focus on eight volcanoes that exhibit summit lava lakes. We find that magma flux is a function of combined gas flux, conduit size and magma rheology. Experiments with high gas fluxes through low viscosity liquid took on turbulent characteristics, which correspond to high degree of return flow, whereas lower gas fluxes through high viscosity liquids yielded slug flow, which corresponded to less return flow. We model the magma flux due to bubble ascent and find that gas-driven liquid flow can yield faster flow rates than other mechanisms at work in volcanic conduits. This can explain the discrepancy between previous estimates of magma flow in conduits and relatively fast, lava lake surface velocity observations. We show how bubble-driven convective flow can work alongside density-driven convection and discuss the depths in the conduit where each are likely to dominate the system.

An interpretative view of open-vent volcanoes

“…The consequence of having a separated gas flow is very important for the magma viscosity because the gas volume fraction is significantly reduced when compared to that of a homogeneous gas flow (Fig 5b and 6 in Vergniolle and Jaupart 1986). The first estimates of the gas volume fraction had assumed a homogeneous gas flow (Burton et al 2007b;Allard et al 2016b), whereas the framework to calculate the gas volume fraction for a separated gas flow had now been developed (Pioli et al 2012;Pansino et al 2019). Three mechanisms can lead to a separated gas flow.…”

“…Laboratory experiments show that a global liquid motion can result from the bubble motion (Pioli et al 2012;Pansino et al 2019). When these conditions are met the small crystals will passively follow the magma motion, thereby decreasing the local crystal volume fraction.…”

“…The bubbles present in the rising degassing magma may be sufficiently large to behave as a separated gas flow, or drift flow, in regards to that of the liquid (Vergniolle and Jaupart 1986). In this case, the gas volume fraction should be calculated from the superficial velocity of the gas phase and the drift velocity of a bubble, which depends on both the bubble diameter and the liquid viscosity (Vergniolle and Jaupart 1986;Pioli et al 2012;Pansino et al 2019). These calculations rely on a good knowledge of the magma viscosity, estimated without the bubbles present in the separated gas flow, and of the conduit radius, a second key parameter (Pansino et al 2019).…”

“…The magma supply rate derives from the SO 2 flux and the melt sulphur content, which implies full magma degassing from depth to surface. The total gas flux used by Pansino et al (2019) combines the ratios of di-atomic gases such as CO 2 /SO 2 and H 2 O/ SO 2 , with the SO 2 flux at several volcanoes. However, the total gas flux, i.e.…”

“…However, the total gas flux, i.e. obtained by adding CO 2 , H 2 O, SO 2 fluxes, and the magma supply rate are not independent, invalidating the use of the estimates of gas volume fractions obtained by Pansino et al (2019). Nevertheless, a proper estimate of the gas volume fraction should consider whether the small bubbles are flowing in a homogenous flow, i.e.…”