Vesicle connectivity in pyroclasts and implications for the fluidisation of fountain-collapse pyroclastic flows, Montserrat (West Indies)

Abstract

Pyroclastic flows are noted for their highly ‘mobile’ behaviour, being able to flow on slopes as low as a few degrees, and this has been attributed to fluidisation or partial fluidisation by escaping gases. The three types of pyroclastic flows generated during the eruption of Soufrière Hills Volcano on Montserrat (dome-collapse, fountain-collapse and surge-derived) differ in their apparent ‘mobility’, suggesting different degrees of fluidisation and friction reduction in the flowing granular materials. Several possible sources of fluidising gas operate in pyroclastic flows. The purpose of this paper is to assess the feasibility of one mechanism: the release during flow of pressurised gases trapped in the vesicles of juvenile clasts. Measurements with a helium pycnometer show that all vesicles in dense lava blocks from dome-collapse pyroclastic flows are connected, so that the mechanism is probably not viable for such flows. On the other hand, pumices from fountain-collapse pyroclastic flows from Soufrière Hills and other volcanoes contain up to 10 vol% of isolated vesicles. This is confirmed by SEM imagery, which shows that Soufrière Hills pumices contain two vesicle populations, with a bimodal size distribution. The large vesicles are mostly interconnected, whereas about two thirds of the smaller population (<15 μm) are isolated. The small vesicles are located as strings, and in some cases as distinctive spheroidal clusters, in the walls between the large vesicles. Many small vesicles and spheroidal vesicle clusters inflated into the (connected) large vesicles as they grew, because they were isolated and had higher internal pressures. The populations probably record two bubble nucleation events in the ascending magma. The second nucleation event must have occurred before the pumices left the vent because the two populations are also preserved in fallout pumices ejected during the same 1997 Vulcanian explosions that generated the pyroclastic flows. One possibility supported by calculations is that the second event occurred in response to brutal decompression of the magmatic foam as it fragmented. Rupture of isolated vesicles provides a potential source of gas in fountain-collapse pyroclastic flows at Soufrière Hills and other volcanoes. Despite significant uncertainties of the relevant parameters, rough calculations show that abrasion of pumice clasts during flow transport could potentially liberate gas at a sufficient rate to fluidise or partially fluidise the material. However, this mechanism is not tenable in dome-collapse or surge-derived flows on Montserrat, the juvenile components of which lack isolated vesicles. Other gas sources, or other mechanisms of friction reduction, must be invoked for these flows.