Testing, improving and validating models of bedload-induced seismic noise from steep flume experiments

Abstract:

Recent work in fluvial seismology indicates that seismic noise near rivers can be used to monitor bedload transport, a process that remains very challenging to observe through other means. Theoretical studies have shown that bedload-induced noise can be separated from water-flow-induced noise, and suggest that bedload transport physics and fluxes can be inferred from river-induced noise. However, prior to using these models in geomorphological contexts, they need to be tested and validated from targeted experiments with independent measurements: this is the main goal of this study.

We conduct laboratory experiments with controlled bedload transport rates in a relatively large flume with typical mountain-river geometries (2 to 8% slopes, 9 to 25 cm flow depths, 3 to 10 cm grain sizes). In these experiments, we simultaneously monitor the micro-mechanics and the ambient seismic noise caused by bedload transport from accelerometers embedded in moving grains (smart rocks) and a seismometer placed nearby the flume. The smart rock measurements allow us to provide new constraints on the dependence upon hydraulic parameters of bedload transport characteristics (grain downstream velocities, hop sizes) and grain impact statistics (probability distributions of impact amplitudes and rates). By incorporating these new constraints into a modified mechanical model of bedload-induced seismic noise, we show that bedload transport fluxes can be accurately inverted from ambient seismic noise. These results validate the application of our theoretical frameworks to estimate realistic bedload sediment fluxes from seismic noise, which is particularly encouraging for future field applications.