GNGTS 2014 - Atti del 33° Convegno Nazionale

Directional amplification as the effect of fractured rocks. The Campo Imperatore case- study. It is well known that a typical condition that favors amplification of horizontal motions is the presence of compliant horizontal layers over stiffer bedrocks. Such cases produce increasing amplitudes of the incident seismic input and increasing ground motion duration due to seismic energy that is trapped and resonates between the free surface and the bedrock interface. Vertical discontinuities have been shown to produce variations of ground motion (e.g. Irikura and Kawanaka, 1980) that can generate trapped-waves and large motion amplification parallel to the fault zone (e.g. Ben-Zion and Aki, 1990; Li et al., 1994; Mizuno and Nishigami, 2006). On the other hand, various recent observations of ground motion in fault zone environments documented a strong directional amplification with high angle to the fault strike (e.g. Rigano et al., 2008; Di Giulio et al., 2009; Pischiutta et al., 2012). Using both volcanic tremor and local earthquakes, Falsaperla et al. (2010) found clear motion polarizations at stations in the crater area of Mt Etna, with polarization directions varying among sites but everywhere transverse to the orientation of the local fracture field. Nevertheless in non-volcanic settings too the predominant fracture orientation has been recognized to affect polarization, with an orthogonal relation between fracture directions and observed polarization (Pischiutta et al. , 2013). Pischiutta et al. (2012) concluded that the polarization of amplified motion tends to be near perpendicular to the orientation of the predominant fracture field expected from analogical and numerical modeling. Moreover, Pischiutta et al. (2014) recently found an orthogonal relation between ground motion polarization and S-wave fast direction in the Val d’Agri region. Therefore, they postulated that the existence of an anisotropic medium represented by fractured rocks causes shear wave velocity to be larger in the crack-parallel component (making S wave velocity to be higher in fracture-parallel direction) and compliance to be larger perpendicular to the crack strike (causing ground motion polarization in the fracture-perpendicular direction). In a different framework, Burjànek et al. (2010, 2012) found strongly polarized motions on unstable mountain slopes on Swiss Alps and ascribed the effect to the resonance of rock blocks separated by large open cracks (see also Moore et al., 2011). Marzorati et al. (2011) related directional ground motions on a ridge to large open fractures, polarization direction being transversal to the fractures. These are cases where directional horizontal motions occur but the amplification level is not controlled by topography. In other words, geometry of topography is not the main factor controlling amplification. We have assessed ground motion polarization across the damage zone of the Campo Imperatore fault, Abruzzo, Central Italy. This site was chosen because of the coexistence of the two factors so far recognized to have an influence on directional amplification effects: topography and fractured rocks. This study is still in progress but some preliminary results are shown in Fig. 3. The fault is oriented ENE-WSW and the footwall block of the Campo Imperatore fault zone is exposed on a pronounced crest. The relief is elongated in a nearly fault-parallel direction, up to 500 m high. The other important reason why we chose this site is that a very detailed structural geological survey has been carried out. Exactly in the same sector where structural geological features were measured, we acquired ambient noise for around 1 hour, using 25 stations along a 200 m transect crossing the fault. Stations were installed exactly in the sector where structural geological features were measured. Signals were processed to compute the horizontal-to-vertical noise spectral ratio as a function of frequency and direction of motion. Wavefield polarization was investigated in the time–frequency domain as well. In Fig. 3 we show results at four stations (CAM2, CAM 11, CAM15 and CAM20) drawing the rose diagram (cyan) obtained by the covariance matrix analysis as well as HVSR contour plots (the amplitude scale is different for each plot). We report measures of extensional fractures and faults at one location (violet star), through violet circular histograms. We find that, in spite of the high complexity of results, the observed polarization pattern is generally oriented orthogonal to the outcropping predominant fracture fields, confirming the existence of a high angle relation between ground motion polarization and fracture fields. GNGTS 2014 S essione 2.2 257