Steadiness in Dilute Pyroclastic Density Currents

Thursday, 17 December 2015: 17:00
308 (Moscone South)
Benjamin James Andrews, Smithsonian Institution, Global Volcanism Program, Washington, DC, United States
Pyroclastic density currents (PDCs) are often unsteady, as evidenced by direct observations of dilute lobes or jets emerging from the fronts of larger currents and by deposits that indicate transient transport and depositional regimes. We used scaled experiments to investigate unsteadiness in dilute PDCs. The experimental currents were run in an 8.5x6.1x2.6 m tank and comprised heated or ambient temperature 20-μm talc powder turbulently suspended in air. Experiments were scaled such that densimetric and thermal Richardson numbers, Froude number, and particle Stokes and settling numbers were dynamically similar to natural dilute PDCs. Although the experiment Reynolds numbers are substantially lower than those of natural PDCs, the experiments are fully turbulent. Experiments were observed with video and high-speed cameras and high-frequency thermocouples. Currents were generated with total eruption durations of 100 s. Unsteadiness in source conditions was produced by interrupting supply for intervals, t, with durations of 1, 2.5, 5, and 10 s in the experimental runs at 35 and 70 s. When t<2.5 s, the currents are indistinguishable from currents with steady supply. In runs with t=2.5-5 s, the individual pulses comprising each current are readily apparent near the source, but decay with distance downstream until the currents appear as single (e.g. steady) flows. In experiments with t=10 s, the 3 pulses comprising each run never merge and the currents remain unsteady. Comparison with the integral turbulent timescale, τ, and current velocity, U, show that unsteadiness is persistent when t>3<τ but currents are steady when t<τ. In currents with 3τ>t>τ, unsteadiness decays such that at a distance of ~4Ut, the currents are again steady. Applied to natural dilute PDCs, our results suggest that currents and their resulting deposits, will only show evidence of unsteadiness if they are disrupted for many seconds and those breaks may “heal” over distances of 100s of meters.