Acoustic monitoring of an avalanche release zone

Fig.4: Our field site in winter. The poles mark the sensor positions.

Dry-snow slab avalanches release when a crack in an unstable weak layer within the snowpack reaches a sufficiently large size to propagate. As the formation of cracks is accompanied by acoustic emissions (AE), monitoring AE is a promising method for evaluating the stability of avalanche prone snow slopes. According to the theory of critical phenomena, the amplitude distribution of acoustic signals generated during crack formation changes as the system approaches failure. Such behavior has been observed in other heterogeneous, natural materials such as limestone, wood, or ice.

Laboratory experiments

In order to adapt AE monitoring for field use within snow, we performed preliminary laboratory fracture experiments with snow samples containing a weak snow layer. Our results showed that the acoustic signals originate from within the weak layer (Fig. 1). Analyzing the distribution of the amplitude squared in running time windows, we observed changes in the distribution’s exponent β, a parameter used to describe amplitude distribution, before, during, and after fracture (Fig. 2). The exponent β could therefore serve as in indicator of fractures within the snowpack and thus also as a precursor to avalanche release in the field.

Fig.1: Localization of acoustic emission events (red dots) recorded by the acoustic sensors (green dots) within a snow sample containing a weak layer (yellow area). The left plot shows acoustic energy with height averaged over sample width.

Fig. 2: AE measurements during a fracture experiment with a snow samples containing a weak layer. Top: Event count (blue) and cumulative count (black line) over time. Middle: Signal amplitude squared over time. Bottom: Exponent β over time. The vertical red line marks the fracture of the snow sample.

Field study

The aim of our ongoing study at our Totalp field site above Davos is to measure acoustic precursory patterns (changes in exponent β) to avalanche release. We placed acoustic sensors in the snowpack in an alpine avalanche release zone ( Figs. 3 and 4). We measured acoustic emissions throughout the winter and compared their amplitude distribution with observed slope stability (assessed by the result of avalanche control work using explosives). Preliminary results are encouraging. The development of a reliable avalanche warning system based on (acoustic) precursors is certainly of great interest for practitioners. Further research in this direction is underway.

Fig. 3: Schematic of our measurement setup at the avalanche release zone of a big avalanche at Totalp near Davos.

• Ingrid Reiweger