Validation of a two-layer depth-averaged model by comparison with an experimental dilute stratified pyroclastic density current

Abstract

Numerical results of a two-layer depth-averaged model of pyroclastic density currents (PDCs) were compared with an experimental PDC generated at the international eruption simulator facility (the Pyroclastic flow Eruption Large-scale Experiment; PELE) to establish a minimal dynamical model of PDCs with stratification of particle concentrations. In the present two-layer model, the stratification in PDCs is modeled as a voluminous dilute turbulent suspension layer with low particle volume fractions (<10-2) and a thin basal bedload layer with high particle volume fractions (~10-2) on the basis of the source condition in the experiment. Numerical results for the dilute layer quantitatively reproduce the time evolutions of the front position and body thickness of the dilute part in the experimental PDC. The numerical results of the bedload thickness and deposit mass depend on an assumed value of mean deposition speed at the bottom of the bedload (D). We show that the thicknesses of bedload and deposit in the simulations agree well with the experimental data, when D is set to about 3.5 x 10-4 m/s. This value of the deposition speed is two orders of magnitude smaller than that predicted by a hindered-settling model. The small value of D suggests that the erosion process accompanied by saltating/rolling of particles plays a role in the sedimentation in the bedload.