Rockfall localization attempts using small seismic array
Essais de localisation d’éboulements rocheux par écoute sismique
Abstract
In the framework of the French National project C2ROP, methods using seismic signals of rockfall propagation are under development in order to retrieve information on the rockfalls (frequency, localisation, propagation path, energy of impact...). The objective is to provide a low cost monitoring system which can: i) provide a better knowledge of the site activity (i.e., the number and size of rockfalls that occur during the monitoring period); ii) enable day and night monitoring in all weather conditions; iii) in the event of an emergency, allow the rapid instrumentation of a site.
However, the localization of rockfalls using seismic records remains a difficult task with classical methods of seismic source localization. These methods require identifying the arrival times of one or more types of waves (P-waves, S-waves, etc.). The picking of theses arrival times is difficult to perform as seismic signals of rockfalls: i) frequently have a low signal-to-noise ratio (the arrival of the first wave is therefore hardly perceptible); ii) are the result of superimposed arrivals of wave trains generated by successive impacts (thus, identifying the same waveform on several seismograms often proves impossible); iii) occur in fractured and heterogeneous media with strong seismic velocity variations.
Here, we show that the signal processing method proposed by Meza-Fajardo et al. (2015) allowed us to specifically isolate Rayleigh waves from the seismic signals of rockfalls. These filtered signals have characteristics which will favour the use of method of seismic source localization. In particular, these filtered signals are "simplified" and have a better signal-to-noise ratio than the original seismic signals. We used signals recorded by an array installed by ISTerre at the cliff of St Eynard near Grenoble, France. The array is composed of four stations a few hundred meters apart. A catalog of events is produced with an estimate of the initial volumes using photogrammetry.
Reference:
Meza‐Fajardo, K. C., Papageorgiou, A. S., & Semblat, J. F. (2015). Identification and Extraction of Surface Waves from Three‐Component Seismograms Based on the Normalized Inner Product. Bulletin of the Seismological Society of America, 105(1), 210-229.