Listening to debris flows: What can ground vibrations tell us about debris-flow entrainment and flow density?

Abstract:

Debris flows generate seismic waves as they travel downslope and can grow in size and destructive potential by entraining sediment along their paths. Recent observations from the Chalk Cliffs monitoring site in central Colorado show there is a systematic relation between the magnitude of seismic waves and both (1) the amount of erodible sediment beneath the flow, and (2) the density of the flow. Specifically, we observed that the spectral power of debris-flow induced ground motion increased by two orders of magnitude after a 34-cm layer of bed sediment was eroded from a bedrock channel. We also observed that high-density (sediment-rich) debris-flow surges generate about two orders of magnitude greater spectral power than low-density (water-rich) surges of similar thickness. These observations lead us to the hypothesis that the recorded ground motions are generated primarily by the impacts of grains on bedrock sections of the channel. This hypothesis is supported by ball drop tests which showed that impacts on deformable loose bed sediment in the channel (if present) generate negligibly small surface waves compared to impacts on bedrock. We thus expect debris-flow induced ground motion to increase as sediment entrainment exposes bedrock in channel, and as the flow density (and number of grains) increase. We explored the connection between ground motions and debris-flow entrainment/density by adapting a model from fluvial seismology [Tsai et al., GRL, 2012]. We used the adapted model to estimate rates of sediment entrainment and the density of flows over bare bedrock channels. Our estimates of sediment entrainment compared favorably with previous direct measurements of entrainment rates at the site. Estimates of flow density are sufficiently accurate to distinguish between three density levels: low (<1200 kg/m³), medium (1200-1600 kg/m³), and high (<1600 kg/m³). Although more testing is needed, these initial results suggest the approach may be a new indirect way to obtain rare field constraints of debris-flow entrainment rate and flow density.