WHAT CAN WE LEARN FROM SEISMIC RECORDINGS OF DEBRIS FLOWS? PRELIMINARY FINDINGS FROM LARGE SCALE SEISMIC EXPERIMENTS AT THE USGS DEBRIS-FLOW FLUME

In June 2016 we conducted six experiments at the USGS Debris-Flow flume focusing on the seismic, infrasound, and tilt signals generated by debris flows. The flume is a 95-meter long channel sloping at 31 degrees and has been used for decades to study debris-flow dynamics using instrumentation to measure evolving flow thicknesses, flow speeds, basal stresses, and basal pore fluid pressures, as well as other variables. This year we expanded the instrumentation to include five broadband seismometers, 34 vertical and 11 three-component short-period geophones, four three-component accelerometers in the bed of and directly next to the flume, four tiltmeters, and five infrasound instruments in array configurations within 50 meters of the flume. In addition, we included five “smart rocks” in some of the experiments that measure three components of acceleration of large clasts in the flow. We measured empirical Green’s functions on the seismic instruments by using force hammer impacts of known magnitudes on the bed of the flume, thereby allowing us to account for subsurface heterogeneity in our calculations. We conducted three sets of paired replicate gate-release experiments, varying the total mass released and the grain size composition between sets. Our scientific goals include (1) understanding the physics of how debris flows generate seismic waves, (2) determining how and what information we can and cannot quantitatively extract from seismic signals, and (3) testing methods for tracking and characterizing debris flows using seismic, tilt, and infrasound arrays. We present preliminary findings from these experiments, focusing primarily on tracking and characterization methods, and discuss implications for scaling up such methods for use in natural settings.