Can the flow dynamics of debris flows be identified from seismic data?

There is growing interest in the use of seismic and acoustic data to interpret a variety of geomorphic processes including landslides and debris flows. This measurement technique is attractive because a broad area can be monitored from a safe distance, unlike more direct methods of instrumentation, which are restricted to known flow paths and are vulnerable to damage by the flow. Previous work has shown that measurements of ground vibrations are capable of detecting the timing, speed, and location of landslides and debris flows. A remaining question is whether or not additional flow properties, such as basal stress, impact force, or flow magnitude can be inferred reliably from seismic data. This question has been difficult to answer, because detailed, independent measurements of flow dynamics are lacking.

Here, we explore characteristics of debris-flow induced ground vibrations using new data from the Chalk Cliffs monitoring site in central Colorado. Monitoring included a heavily instrumented cross-section consisting of two tri-axial geophones to record ground vibrations (at 333 Hz), a small, 225 cm² force plate to record basal impact forces (at 333 Hz), a laser distance meter to record flow stage over the plate (at 10 Hz), and a high definition camera to record flow dynamics (at 24 Hz). One geophone (A) was mounted on a boulder partially buried in colluvium; the other (B) was mounted directly to weathered bedrock typical of the site. This combination of instrumentation allowed us to compare the spectral response of different geophone installations to independently measured flow depth and basal impact force. We also compared the response of the geophones to surges that flowed over a sediment-covered bed (40-cm thick) to surges that flowed over a bare bedrock channel.

Preliminary results showed that site conditions have a large effect on recorded debris-flow vibrations. The seismic signature of debris flow was very different between the geophones, with geophone B exhibiting a broad frequency response during surges, while geophone A had energy concentrated in a narrower 30-90 Hz band. In addition, the presence of bed-sediment cover was found to damp the geophone signals. While further work is needed, these initial results highlight some of the challenges associated with extracting debris-flow dynamics from seismic data.