GNGTS 2016 - Atti del 35° Convegno Nazionale

GNGTS 2016 S essione 3.2 551 allowed investigation of the fracturing state of the granitic cliff at depth, which helped to better understand the geometry of the unstable sectors and to define a velocity model for the site, fundamental for the location of the detected microseismic events. Considering the complex morphological and geological context, cross-hole tomography contributed to spatially image the seismic-velocity field of the site, enabling to locate the fracture zones and the related velocity contrast. Conversely, downhole tests helped to define more punctual velocity values and to better calibrate the seismic field. Laboratory measurements of ultrasonic pulse velocities on a variety of samples collected at the site revealed to be a fast and simple method to lithologically interpret the field data, with a good agreement between the results at different scales. Density and porosity laboratory measurements on granite samples allowed also to associate the different seismic velocity ranges to different macroscopic peculiarities (e.g. weathering conditions and presence of anisotropy) of the granites. Moreover, processing of surface shot seismograms in time and frequency domain, together with surface-wave analysis, revealed the best way to constrain fracture opening in depth (e.g. in Bièvre et al. , 2012; Bergamo and Socco, 2014). All the results confirmed the presence and persistence of deep and pervasive fractures within the rock mass which isolate the prone-to-fall frontal portions of the cliff. Passive seismic monitoring. For monitoring purposes, passive seismic techniques were applied on the prone-to-fall compartment using two different approaches. The first approach consisted in detecting an increase in the number of seismic events and/or a rise in seismic energy over a given period of time, which could indicate a destabilization of the unstable mass (e.g. in Senfaute et al. , 2009; Lévy et al. , 2010). No evidence of acceleration to failure was identified during the monitored period (October 2013 - present). Proper location of microseismic events, using a calibrated 3-D velocity model, revealed to be quite challenging given the complexity of the site and of the resulting wave field. A concentration of low-energy releases close to the major fracture planes affecting the rock mass, particularly along K2 and K4, could be however depicted (Fig. 2). The second approach involved the processing of ambient seismic noise, in order to focus on the dynamic behavior of the whole unstable rock mass (see Larose et al. , 2015 for an overview of the methods and potential applications). The spectral content of seismic noise systematically highlighted clear energy peaks on the unstable sector, which were interpreted as the resonant frequencies of the investigated volume. Ground motion at these frequencies was found to be controlled by the main fractures observed at the site through numerical modeling and modal analysis. Both spectral analysis and cross-correlation of seismic noise showed short- period and seasonal reversible variations related to external air temperature fluctuations (Fig. 3). Numerical simulations revealed the essential key for the validation and interpretation of experimental results. Fig. 2 – Microseismic event hypocentral location. Sources are located predominantly near K2 and K4 fractures.