Final Report Summary - SAFER (Detecting slow deformation signals preceding dynamic failure: A new strategy for the mitigation of natural hazards) The proposal has aimed to identify ‘in situ’ characteristic slow deformation signals and to carry out rock mechanics/deformation tests to quantitatively determine the thermo hydro-mechanical processes leading to instability. To this purpose the monitoring of unstable rock masses can provide a better knowledge of the active processes and help to forecast their evolution to failure. Passive seismic monitoring can be potentially helpful for this purpose. Detection, classification and localization of microseismic events within the prone-to-fall rock mass can provide information about the incipient failure of internal rock bridges. We have first carried out a cross-hole seismic tomography survey coupled with laboratory ultrasonic velocity measurements and physical properties determination on rock samples to characterize the damaged and potentially unstable granitic cliff of Madonna del Sasso (NW, Italy). Results have allowed to obtain: i) a lithological interpretation of the velocity field obtained at the site, ii) a systematic correlation of the measured velocities with physical properties (density and porosity) and macroscopic features of the granite (weathering and anisotropy) of the cliff. The multi-scale approach adopted within this study revealed to be crucial for the imaging at depth of the main fractures affecting the cliff (site-scale seismic tests) and for the understanding of the variations in the seismic velocity between altered and intact rock (laboratory scale tests); similar approaches can be potentially used in further microseismic monitoring studies.On the basis of these indications a microseismic network has been installed during the summer of 2014 on the Madonna del Sasso cliff. The network consists of four triaxial geophones (4.5 Hz, 1-4KHz) connected to a multichannel acquisition system (Granite - Kinemetrics). During the monitored period, we recorded thousands of events with different waveforms, duration and frequency content. The signals were analysed and classified to extract only the ones with a possible relation with fracture processes. No acceleration in the event rate or significant correlation with variations of external factors (air temperature, rainfalls) was detected in the monitored period.Finally a 3-D numerical modelling of the cliff response was performed using the continuous-medium finite-element software Comsol Multiphysics. The two tested models gave fully comparable results in term of motion orientation, with a slight difference in the obtained first three frequency values. Numerical simulations allowed to semi-quantitatively interpret the observations.
