Effects of surface topography on ground shaking prediction: implications for seismic hazard analysis and recommendations for seismic design

This study examines the role of topographic effects on the prediction of earthquake ground motion. Ground motion prediction equations (GMPEs) are mathematical models that estimate the shaking level induced by an earthquake as a function of several parameters, such as magnitude, source-to-site distance, style of faulting and ground type. However, little importance is given to the effects of topography, which, as known, may play a significant role on the level, duration and frequency content of ground motion. Ridges and crests are often lost inside the large number of sites considered in the definition of a GMPE. Hence, it is presumable that current GMPEs are unable to accurately predict the shaking level at the top of a relief. The present work, which follows the article of Massa et al. about topographic effects, aims at overcoming this limitation by amending an existing GMPE with an additional term to account for the effects of surface topography at a specific site. First, experimental ground motion values and ground motions predicted by the attenuation model of Bindi et al. for five case studies are compared and contrasted in order to quantify their discrepancy and to identify anomalous behaviours of the sites investigated. Secondly, for the site of Narni (Central Italy), amplification factors derived from experimental measurements and numerical analyses are compared and contrasted, pointing out their impact on probabilistic seismic hazard analysis and design norms. In particular, with reference to the Italian building code, our results have highlighted the inadequacy of the national provisions concerning the definition of the seismic load at top of ridges and crests, evidencing a significant underestimation of ground motion around the site resonance frequency.