Linking Pyroclastic Surge Dynamics to Deposits through Scaled Laboratory Experiments

Among the deadliest of explosive volcanic hazards are pyroclastic surges – fast-moving, hot, dilute ground-hugging currents that overtop topography and leave complex deposits. Understanding the link between surge dynamics and their deposits is crucial for forecasting the impacts of future eruptions. To investigate surges, two sets of scaled laboratory experiments were conducted. Set 1 released fluid pulses into less-dense ambient water (3-m flume). Pulse fluids were saline solutions with and without particles, and alcohol-water-particle mixtures. Non-dimensional numbers are calculated using both input parameters and measured outcomes. Inputs - fluid density, particle size and concentration, and volume of fluid released - were varied to explore a range of conditions. Key output parameters obtained by video analysis are flow thickness and propagation velocity. Propagation velocity, Re, and Ri increased with increasing pulse density, while Pn decreased. Lab Re values indicate fully turbulent flows, consistent with natural flows. Lab Ri closely matched nature and flow propagation was largely controlled by negative buoyancy, with entrainment playing a minor role. All flows began as subcritical (Fr<1). Alcohol-water-particle runs exhibited buoyancy reversals caused by particle sedimentation, characterized by gradual deceleration and late-stage formation of buoyant plumes. Saline runs maintained nearly constant velocities. In the second set of experiments, alcohol-water-particle mixtures were pulsed over particle bed. Various substrate topographies were tested (flat, mound-trough sequences, steps, wedges). Deposits thickened in troughs and thinned on peaks. Progressive climbing dunes formed on the lee side of the second peak of double peaks and peak-trough combinations, and in step-up topographies. Regressive climbing dunes formed on the stoss side of the first peak of peak-trough combinations and step-down topographies, and on the stoss side of mounds. Climb angles were 16 to 36°, consistent with those documented in pyroclastic surge deposits. The occurrence of both regressive and progressive climbing dunes suggests localized transitions between subcritical and supercritical flow. No cross-beds formed on flat substrates, suggesting that complex substrate topography is required for bedforms to occur in nature. A code benchmarking effort is underway in which targeted model runs will be compared to both sets of experiments in order to develop comprehensive hazards prediction tool.