Granular flows were generated by the release of beds of particles in various fluidized states, which then deaerated in a horizontal channel. We describe characteristics of the flows and their deposits. Morphological similarities between deposits in experiments and in the field suggest that pyroclastic flow deposits form from a fluidized mixture. The experiments show that slightly expanded, fluidized flows are more mobile than non‐fluidized flows of equivalent volume and material composition. They travel to a fixed distance from their source, which depends only weakly on their initial degree of fluidization. Flows of fine particles (<100 μm) deaerate slowly and are highly mobile. Pyroclastic flows commonly have large amounts of fine ash, which may have a controlling influence on their high mobility.
Heating of water under hot pyroclastic flow deposits can drive hydroeruptions, forming craters and aprons of secondary deposits. According to the established conceptual model, steam pressure builds until failure of the pyroclastic overburden, and a relatively low permeability (fine-grained) cap promotes secondary explosions. We explore a complementary model where the stress from drag related to gas flow up through the particle interstices, is comparable in magnitude to the static pressure difference between the base and the top of the pyroclastic flow deposit. The drag force supports (part of) the weight of the particles and so reduces inter-particle friction; in a mono-sized bed this friction is effectively eliminated at the 'minimum fluidisation velocity', which depends on the size and density of the particles. Through analogue experiments we show that violent outbursts can be generated when there are vertical variations in the minimum fluidisation velocities of granular materials. We ran experiments with layers of particles with different sizes or size distributions (bi-modal with different proportions of fine and coarse particles) in a tank with a porous base allowing a distributed upward airflow through them. A finer-grained layer capping a coarser layer does not generate jets of particles or craters; rather, increased gas flux leads to fluidisation of first the fine, and then the coarse (lower), layer. However, when the upper layer is coarser, for some combinations of layer thicknesses and grain sizes, the bed domes upward as a gas pocket grows within the finer layer; when the gas pocket penetrates the top of the bed, it forms a crater and erupts particles. The gas velocity when doming initiates is greater than that calculated for the weight of the top layer to be balanced by drag and the pressure difference across that layer. This discrepancy is explained by the layers having a strength (from inter-particle friction), consistent with the observed dependence of the initiation velocity on the absolute thickness of the layer. Using data from Mt St Helens 1980 deposits, we show that the drag-related trigger observed in the laboratory is a feasible mechanism for secondary hydroeruptions through pyroclastic flow deposits.
Many volcanoes show continuous but variable deformation over timescales of years to decades. Variations in uplift rate are typically interpreted as changes in magma supply rate and/or a viscoelastic response of the host rock. Here we conduct analogue experiments in the laboratory to represent the inflation of a silicic magma body at a constant volumetric flux, and measure the chamber pressure and resulting surface displacement field. We observe that dyke intrusions radiating from the magma body cause a decrease in the peak uplift rate, but do not significantly affect the spatial pattern of deformation or spatially averaged uplift rate. We identify 4 distinct phases: 1) elastic inflation of the chamber, 2) a gradual decrease in the rate of uplift and pressurisation, associated with the formation of visible cracks 3) propagation of a dyke by mode 1 failure at the crack tip and 4) a pressure decrease within the chamber. Phase 2 can be explained by either a) crack damage, which reduces the elastic moduli of the surrounding rock or b) magma filling pre-existing cracks. Thus these experiments provide alternative mechanisms to explain observed variations in uplift rate, with important implications for the interpretation of deformation patterns at volcanoes around the world.
Abstract The fluid mechanics of laminated sheet manufacture was studied by pouring a liquid filler from a point source onto an inclined moving sheet and then passing it under a narrow gap below a pinch bar that spreads the liquid under the influence of gravity. The liquid filler must pass under the pinch bar before it starts to solidify. The spreading of a viscous liquid as it undergoes this process was studied theoretically and experimentally. Steady similarity solutions were obtained based on a vertically averaged formulation of the mass and momentum equations for viscous flows. Experimental measurements of the spread of the fluid and the time taken for fluid to travel from the source to the pinch bar agreed well with theoretical predictions. The new models are applied to the operation of the manufacturing process and the implications are discussed.
The influence of an external laminar flow on the spreading of a viscous gravity current moving over a horizontal floor is studied theoretically and experimentally. The viscous stress exerted by the ambient flow drives the viscous gravity current streamwise with a velocity proportional to the local height of the current. The one-way coupling between the ambient flow and the spread of the current is examined. Similarity and numerical solutions are developed to describe viscous gravity currents spreading from line and point sources. An experimental study of the spreading of viscous gravity currents issuing from a point source in a channel flow, for both constant-flux and instantaneous releases, confirms the essential character of this description.
'%3e%3cpath fill='%23012169' d='M0 0v30h60V0z'/%3e%3cpath stroke='%23fff' stroke-width='6' d='m0 0 60 30m0-30L0 30'/%3e%3cpath stroke='%23C8102E' stroke-width='4' d='m0 0 60 30m0-30L0 30' clip-path='url(%23b)'/%3e%3cpath stroke='%23fff' stroke-width='10' d='M30 0v30M0 15h60'/%3e%3cpath stroke='%23C8102E' stroke-width='6' d='M30 0v30M0 15h60'/%3e%3c/g%3e%3c/svg%3e)
