Particle velocity fields and depositional processes in laboratory ash flows, with implications for the sedimentation of dense pyroclastic flows

We conducted laboratory experiments on dam-break flows of sub-250-µm volcanic ash, generated by the release of gas-fluidized and variably non-expanded to expanded (up to 35%) beds, in order to gain insights into the internal kinematics of pyroclastic flows. The flows were typically several cm thick and had frontal speeds of up to ∼2 m s−1. High-speed videos taken through the transparent sidewall of the 3-m-long channel were analyzed with a particle-tracking algorithm, providing a spatial and temporal description of transport and sedimentation. The flows deposited progressively as they traveled down the flume, being consumed by sedimentation until they ran out of volume. Deposition commenced 5–20 cm rearward of the flow front and (for a given expansion) proceeded at a rate independent of distance from the lock gate. Deposit aggradation velocities were equal to those inferred beneath quasi-static bed collapse tests of the same ash at the same initial expansions, implying that shear rates of up to ∼300 s−1 have no measurable effect on aggradation rate. The initially non-expanded (and just fluidized) flow deposited progressively at a rate indicative of an expansion of a few percent, perhaps due to shear-induced Reynolds dilation during initial slumping. The fronts of the flows slid across the flume floor on very thin basal shear layers, but once deposition commenced a no-slip condition was established at the depositional interface. Within the flows, the trajectory of the constituent particles was linear and sub-horizontal. The velocities of the particles increased with height above the depositional interface, reached a maximum, then declined slightly towards the flow surface, perhaps due to air drag. At a given location, the velocity profiles were translated upwards as the deposit aggraded. The results show that even cm-thin, poorly expanded flows of ash deposit progressively, as inferred for many pyroclastic flows. The change from (frontal) slip to (rearward) no-slip conditions at the bases of the laboratory flows are qualitatively consistent with some textural features of pyroclastic flow deposits.