Fracture Network Imaging on the Preonzo rock slope instability using resonance mode analysis

Abstract

Coseismic landslides and rockfalls are among the most devastating secondary effects of earthquakes. In Switzerland, a country of moderate seismicity, such effects occured, for example, following the 1946 Mw 5.8 Sierre earthquake at Rawilhorn. Combining geospatial susceptibility proxies, such as topography, with models for peak ground acceleration allows for estimating the likelihood of earthquake-induced mass movements on a regional scale (Cauzzi et al., 2018). However, to evaluate the coseismic landslide hazard on a specific slope, detailed investigations on site are required. Therefore, it is of crucial importance to understand the dynamic response of a slope and its dynamic behavior during strong ground shaking. Describing the dynamic response of rock slopes can be achieved by measuring ambient seismic vibrations. It is generally observed that the seismic wavefield polarizes perpendicular to open fractures and that unstable slopes exhibit strong wavefield amplifications. Kleinbrod et al. (2019) established a classification scheme for ambient seismic recordings on rock slope instabilities with two end members: depth-controlled and volume-controlled sites. Depth-controlled sites are characterized by the presence of propagating surface waves and a broad ramp of increasing amplification towards higher frequecies. In contrast, volume-controlled sites exhibit normal mode behavior due to standing wave phenomena within compartments clearly separated by well-defined fracture sets. Normal mode analysis is a well established technique in civil engineering to assess the structural integrity and the dynamic response of the object studied. We performed frequency domain decomposition (FDD) modal analysis on ambient vibration data acquired on an unstable rock site with a volume larger than 150'000 m3 near Preonzo, Canton of Ticino, Switzerland (Häusler et al, 2019, see Fig. 1). We show that the high ground motion amplification and the clear polarization pattern identified by FDD compare well to previous studies that are based on site-to-reference spectral ratios and time-frequency polarization analysis (Burjánek et al., 2018). In addition, FDD allows for a better detection of higher modes, which can be used to efficently map dominant fracture sets. This is of special interest on rock slope instabilities where little or no surface expressions of fractures are developed and where geodetic monitoring systems are not installed yet. Enhanced FDD additionaly provides the damping parameters (energy loss) and an improved estimate of the resonance frequency. These parameters are especially relevant for long term monitoring since they are expected to change with increasing damage, either rapidly after strong ground shaking or other external loading or gradually due to progressive degradation of the rock mass over time (e.g. Michel et al., 2011).