Seismic input modelling for relevant earthquakes in eastern Sicily

Abstract

Assessment of potential ground motions in seismically active regions is essential for purposes of seismic design and analysis. One key challenge in modern day engineering practices is to better understand the processes associated with ground shaking in order to mitigate the effects and to cope with future earthquakes. In recent years, ground motion simulations are becoming the most effective tool to study characteristics of the faulting mechanisms and regional seismic parameters. The fundamental approach in ground motion simulations is to mimic the source mechanisms, regional wave propagation properties and site conditions for estimating reliable synthetic records. A simulation technique, widely used in recent years, is based on the finite-fault stochastic method (EXSIM, Motazedian and Atkinson, 2005; Boore, 2009). It requires the specification of parameters describing source, path and site characteristics, into simple functional forms. These parameters are crucial for successful ground motion simulations, since the damage caused by earthquakes depends strictly on them. The main goal of this thesis is to understand the key parameters that locally control the ground motion, as well as to investigate the relative contributions of local site effects. The thesis includes two main parts. The first one is dedicated to the investigation of the site effects in the Etnean area and in the Hyblean Plateau. Seismic waves are greatly modified by the local geology and soil conditions. Since geometry and soil conditions of the sedimentary layer often vary over short distances, nature and levels of ground shaking can vary significantly within a small area. Records of both earthquakes and ambient noise, taken from the stations of the Permanent Seismic Network (RSP) of Eastern Sicily (INGV-Osservatorio Etneo, Catania) were analysed in detail and processed through spectral ratio techniques (Horizontal to Vertical Fourier spectral ratios; Horizontal to Vertical Response Spectra ratio) in order to compute the corresponding amplification functions and a new soil classification scheme. In particular, the soil classification was based on the estimation of the dominating period of site response. Two approaches were used, taking into account either the Fourier or the response spectra. The first approach is an indirect measurement of the local amplification since it calculates the average horizontal-to-vertical spectral ratios comparing them to the theoretical curves coming from literature information about the structure under every site. Then, amplification functions were computed and later grouped, thus finding an average amplification function for each class. The second approach performs a more direct estimate of site effects computing response spectral ratio of each station with respect to the response spectrum of a reference station, hence directly obtaining the transfer functions, that were then subdivided into groups. The 7 implemented classification scheme is similar to the one proposed by Zhao et al. (2006). In both methods, four classes were identified. The main objective of the second part is to validate the soil classification scheme with the corresponding amplification functions. Two real events, which were localized respectively at lower and higher depth beneath the Etna volcano, were simulated, using the obtained amplification functions and considering the seismological parameters regarding the source and the path reported in the literature. Simulated ground motion parameters and recorded data were finally compared. The obtained results prove the validity of the obtained site amplification functions, which can be used by scientific community as input to simulate ground motion for the studied areas (EXSIM program). They also demonstrate the validity of the proposed soil classification scheme. In the Etnean area, since it is characterized by alternating soft and rigid terrains, it is necessary to use a method that considers all the stratigraphic succession overlying the bedrock and thus different from the conventional VS30. Indeed, it is widely shown that a classification scheme of the main soil categories, founded on the average shear wave velocity of the upper 30 m alone, is not sufficient to categorise the seismic responses of geologically complex areas.