Ground motion amplification at sites with pronounced topography: the controversial role of local geology

Abstract

The topographic amplification of seismic waves has received an increasing interest in the last four decades following observations of large amplification on mountain tops (e.g. Davis and West, 1973; Griffiths and Bollinger, 1979; Çelebi, 1987; Umeda et al., 1987; Kawase and Aki, 1990; Ponti and Wells, 1991; Hartzell et al., 1994; Pedersen et al., 1994a; Chavez-Garcia et al., 1996). The recurrence and consistency of these observations has motivated much work both in terms of theoretical investigations and numerical simulations of the diffraction of seismic waves caused by the topography (e.g. Bouchon, 1973: Bard and Tucker, 1985; Géli et al., 1988; Anooshehpoor and Brune, 1989; Gaffet and Bouchon, 1989; Sanchez-Sesma and Campillo, 1991; Pedersen et al., 1994b; Le Brun et al., 1999; Paolucci, 2002). The simulations and the observations are often in qualitative agreement with the amplification at the topography top and for wavelengths comparable to the mountain width. The disagreement concerns the calculated amplification level that tends to underestimates observations. This discrepancy has suggested that other effects could be responsible of the amplification effect, as the geological setting, complicated incident wave field, more complex topography, etc (Bard and Chaljub, 2009). Beside strong amplification, topographic irregularities have been recognized to produce directional effect of resonance; the scattered wave field is polarized in a site-characteristic direction. Spudich et al. (1996) found that directional amplification occurs transversally to the hill major axis, as subsequently assessed by several other authors (e.g., Del Gaudio and Wasowski, 2007; Massa et al., 2010; Pischiutta et al., 2010). In the framework of a statistical study performed using stations of the Italian seismic network to check the recurrence of directional amplification effects, Pischiutta et al. (2010 and 2011) and Rovelli et al. (2011) investigated the relation between the direction of maximum amplification and the hill elongation at around 40 selected stations of the Italian seismic network. They found that only the 25% of stations showed an angular relation between directional amplification and the hill elongation ranging from 80 and 90 degrees. The same conclusions were reached by Burjanek et al. (2014a and 2014b) who investigated the relation between the S-wave velocity profiles and the amplification occurrence at 25-instrumented sites with complex topography in Switzerland and Japan. They stressed that the amplification was controlled primarily by the sub-surface velocity structure and they did not identify any link between the surface topography and the observed response at the studied 25 sites. Thus recent findings have suggested that large systematic amplifications at topographic sites cannot be explained by surface geometry only, and that although the effect of geometry is present, it cannot be simply decoupled from the site response. We think that directional amplification observed at sites with pronounced topography are often correlated with rock fractures. This feature has not considered adequately so far. Here we propose a model that could explain directional amplification. Similar effects have been recently observed (Marzorati et al., 2011) and associated to gravitational instabilities (Burjanek et al., 2010) as well as to fault damage zones (Falsaperla et al., 2010; Pischiutta et al., 2012 and 2013; Di Giulio et al., 2013). Pischiutta et al. (2014 and 2014) interpreted the strong polarization in terms of fracture fields that make the rock more compliant in the strike-transverse direction (Pischiutta et al., 2012 and 2014).