Item: TOWARDS AUTOMATIC AVALANCHE DETECTION AND CHARACTERIZATION USING SEISMIC AND INFRASOUND ARRAY DATA
-
-
Title: TOWARDS AUTOMATIC AVALANCHE DETECTION AND CHARACTERIZATION USING SEISMIC AND INFRASOUND ARRAY DATA
Proceedings: Proceedings, International Snow Science Workshop, Whistler, BC, Canada, 2026
Authors:
- Christine Seupel [ WSL Institute for Snow and Avalanche Research SLF, Davos, Switzerland ]
- Patrick Düggelin [ WSL Institute for Snow and Avalanche Research SLF, Davos, Switzerland ]
- Andri Simeon [ WSL Institute for Snow and Avalanche Research SLF, Davos, Switzerland ]
- Cristina Pérez-Guillén [ WSL Institute for Snow and Avalanche Research SLF, Davos, Switzerland ]
- Andreas Fichtner [ Department of Earth and Planetary Sciences, ETH Zürich, Institute of Geophysics, Zürich, Switzerland ]
- Alec van Herwijnen [ WSL Institute for Snow and Avalanche Research SLF, Davos, Switzerland ]
Date: 2026-09-28
Abstract: Reliable avalanche detection and localization are essential for effective hazard mitigation. They support avalanche forecasting, risk management, and the development of new, data-driven forecasting tools. Both seismic and infrasound sensors can be used to record avalanche signals. Generated by different parts of the avalanche flow, these signals contain complementary information. When deployed as arrays, seismic and infrasound sensors enable advanced signal processing to estimate source parameters, including apparent velocity and backazimuth, which can be used to locate avalanches. Although both types of systems are well suited for avalanche detection and localization, each possesses specific strengths and limitations that can be exploited in a complementary manner. To investigate this potential, we installed a co-located seismo-acoustic array of five seismic and five infrasound sensors in the Dischma Valley near Davos, Switzerland. We have recorded continuous data since the 2020-2021 winter season. While infrasound array processing yields stable and coherent source parameters even when signal amplitudes are close to the noise level, beamforming results from seismic data are more variable but still robust. In addition, cross-correlation between co-located seismic and infrasound sensors reveals distinct patterns of seismo-acoustic coupling that reflect differences in avalanche types. Based on these findings, we are developing a combined seismo–acoustic detection framework that exploits array-derived source parameters to automatically detect avalanches and provide additional information on avalanche type and flow regime. Specifically, the framework identifies candidate avalanche signals using threshold-based rules applied to source parameters, signal amplitudes, and durations, while cross-correlations between seismic and infrasound signals are used to estimate the avalanche type. We evaluate the detection algorithm using a dataset of avalanche signals recorded over five winter seasons. Preliminary results are promising, indicating that the combination of seismic and infrasound data improves detection robustness, while also enabling basic avalanche characterization.
Object ID: ISSW2026_O5.3.pdf
DOI: https://doi.org/10.15788/1790098611
Language of Article: English
Presenter(s): Christine Seupel
Keywords: Avalanche Detection, Seismic, Infrasound, Array Processing, Instrumentation and Remote Sensing
Page Number(s): 1174 - 1180
-