Particle recycling and oscillations of volcanic eruption columns

Abstract

Abstract. We study the dynamics of sedimentation and reentrainment of particles from the umbrella cloud above an axisymmetric, turbulent, particle-laden buoyant plume. We develop a model to show that the reentrainment of particulate material into the uprising plume will cause the particle flux to increase by a factor of e between the plume source and the umbrella cloud. A buoyant plume rising in an environment of uniform density may thereby become negatively buoyant if its particle loading becomes suf[iciently high. We compare the predictions of the model to a series of laboratory experiments and show that at high particle loadings the plume undergoes an oscillatory collapse. This periodic collapse generates dense gravity flows down the flanks of the plume. In the context of volcanic eruptions, such an instability, associated with particle recycling may lead to the formation of interleaved fall and flow deposits, such as have been observed near the collapse horizon in a number of pyroclastic deposits.

Particle recycling in volcanic plumes

“…For a given mass flux, the increase in the critical velocity may be as large as 5%, so that the fluctuations reported are most significant near the source. Veitch and Woods (2000) showed that, for an incompressible laboratory scale plume, the fractional increase in particle increase from source to top is e≈2.71. However, for a compressible plume, this result is modified (see the section Fallout) because the total solid mass flux entrained decreases.…”

“…However, as the mass flux reaches Q CE at time t 1 , we move into a regime where the plume can not remain stable because of the reentrainment of particles (cf. the experiments of Veitch and Woods, 2000). If the rate of increase of the mass flux is sufficiently slow then the eruption column may fluctuate between plume-forming and flow-forming behaviour.…”

“…For particles with settling speeds in the range 1-10 m s -1 , and a plume height 1-10 km, we expect the recycling time to be comparable to the particle fall time, which will be of the order 100-10,000 s. If the eruption becomes oscillatory as a result of particle recycling, we expect the period of oscillation to scale with the particle fall time, as observed in laboratory experiments (Veitch and Woods 2000). The numerical simulations of Neri and Dobran (1994) reveal oscillations on a time-scale of 50 s, similar to the smaller end of the above range.…”

“…Numerical experiments have demonstrated that unsteady flows may result from both the recycling of particles into the central jet from the collapsing eruption column and also because of complex transient dynamics of pyroclastic dispersion into the atmosphere (Neri and Dobran 1994). Veitch and Woods (2000) reported that particle recycling may have a significant effect on the dynamics of an experimental Boussinesq plume in the laboratory. At the limit when the plume height is small compared with the ambient scale height, recycling may cause a steady plume to collapse.…”

“…At the limit when the plume height is small compared with the ambient scale height, recycling may cause a steady plume to collapse. In this paper, we develop the modelling of Veitch and Woods (2000) for application to a volcanic eruption column by including the effects of heat transfer between solid and gas phases, and of pressure and density variations with height in the atmosphere.…”