AmericanChemicalSociety.com

We study gravity-driven suspension currents sustained by continuous basal particle blowout at the head. In this regime, basal erosion behaves as a narrow, steady supply of higher-density fluid into the current's front. Large scale geophysical flows are ubiquitous in nature, including powder snow avalanches and deep sea turbidity currents. We investigate the mechanisms and conditions for basal particle blowout and determine the effect of such powerful ejection of source material on the suspension cloud dynamics. To do so we carry out laboratory experiments in a water flume, along with theory and numerical simulations for both the basal domain and the current.

Recent observations in powder snow avalanches with ground-penetrating radar (Gauer and Issler, 2004) and fixed pressure transducers (McElwaine and Turnbull, 2005) show an explosive pickup of snowpack material accompanied by a significant depression at the head. The pickup feeds particles to the flow and, in doing so, sustains the suspension cloud with net density greater than the ambient. In our theory we model this supply of particles as a narrow intense source of momentum into an oncoming flow.

This year we completed a perturbation analysis to the underlying potential flow first studied by Rankine (1863). The density perturbation involves variation of the density ratio between a uniform velocity stream of ambient density with an isotropic source of dense fluid located at the origin. The calculations revealed the magnitude of the displacement of the separatrix and a velocity jump across the interface. We also checked the separatrix displacement against numerical simulations.

Figure 1: Rankine half body schematic diagram and flume experiment with fluid body shape illuminated by sodium fluorescein in a light sheet.

In addition, we examined the role of the pressure gradient induced by the advancing powder cloud on the porous snowpack. We calculated stresses on a vertical failure surface, from which snow is siphoned up to feed the avalanche. We derived a sufficient condition for steady failure, which suggested density, internal friction and cohesion of snow packs able to sustain powder snow avalanches by frontal entrainment. Our predictions for depths of blowout failure were in good agreement with those recorded by radar measurements at avalanche test sites. Thus our work showed that blowout represents a significant particle supply to the current. We are now coupling this theory to a detailed simulation of the powder cloud.

We operated a water flume experiment reproducing particle blowout confined to a narrow line perpendicular to the main flow by injecting brine solutions of known density via an isotropic line source into the oncoming flow (figure 1). In this way, the positions of the erosion source and the front's head remained fixed in the laboratory frame. This year we recreated the Rankine half body shape and compared results with the density perturbation analysis mentioned above. Trends predicted by our potential flow theory matched measurements well.

Figure 2: Velocity vectors in the avalanche rest frame of dense particle laden fluid ejected by the snowpack and entrained into the powder cloud.

Finally we used numerical simulations to investigate the role of viscosity in creating large-scale instabilities. We also examined non-ideal aspects of the experiments, such as finite flume size and growth of boundary layers, tested the validity of assumptions in the theory, and verified accuracy of its inviscid predictions.

In the last year of this project, we plan to carry out large eddy simulations (LES) to provide insight into the interfacial mixing process. LES simulations capture the motion of the large energy carrying vortices; these are responsible for mixing and entrainment of ambient fluid into the current. Our analysis will allow us to determine the sources of drag on the current and the importance of each to the dynamics. Ultimately we aim to model complex currents driven by gravity and particle blowout by a simple dynamical system.