Plenary Talk: How Pyroclastic Flows Outsmart Granular Friction During Volcanic Eruptions

Granular-fluid flow and gas-fluidization traditionally constitute key processes in the investigation and modelling of violent volcanic and non-volcanic mass flows on Earth and other planets. These include phenomena such as pyroclastic flows, and avalanches of rock and snow. Amongst these natural gas-particle flow phenomena, pyroclastic flows constitute the most lethal volcanic process known. Causing more than fifty percent of volcanic fatalities globally, these hot mixtures of volcanic particles and gas exhibit an astonishing fluidity. This allows them to transport thousands to millions of tonnes of volcanic material across the Earth’s surface over tens to hundreds of kilometres, by-passing rough and tortuous terrain with flat and upsloping surfaces. Here we give an account for the natural flow characteristics and hazard impacts of these flows to highlight a long-standing research question: how do pyroclastic flows attain their characteristic mobility? The fluidity of pyroclastic flows is attributed to an internal process that effectively counters granular friction. Aspects of this enigmatic fluidity have been variably explained by vertical gas escape, high internal gas pore pressure, acoustic fluidization, dynamic fragmentation, among others. However, the violence of real-world flows has precluded direct measurements, so that none of these processes has been quantitatively validated.

In this presentation we show, through large-scale experiments and numerical multiphase modelling, that pyroclastic flows generate their own air lubrication. This forms a near-frictionless basal region. Air lubrication develops under high basal shear when air is locally forced downwards by reversed pressure gradients and displaces particles upward. We demonstrate that air lubrication is enhanced through a positive feedback mechanism, explaining how pyroclastic flows are able to propagate over slopes much shallower than the angle of repose of any natural granular material. This discovery necessitates a re-evaluation of hazard models that aim to predict the velocity, runout and spreading of pyroclastic density currents.