Assessment of potential strong ground motions in the city of Rome

Abstract

A methodology is used which combines stochastic generation of random series with a finite-difference technique to estimate the expected horizontal ground motion for the city of Rome as induced by a large earthquake in the Central Apennines. In this approach, source properties and long-path propagation are modelled through observed spectra of ground motion in the region, while the effects of the near-surface geology in the city are simulated by means of a finite-difference technique applied to 2-D models including elastic and anelastic properties of geologic materials and topographic variations. The parameters commonly used for earthquake engineering purposes are estimated from the simulated time histories of horizontal ground motion. We focus our attention on peak ground acceleration and velocity, and on the integral of the squared acceleration and velocity (that are proportional to the Arias intensity and seismic energy flux, respectively). Response spectra are analyzed as well. Parameter variations along 2-D profiles visualize the effects of the small-scale geological heterogeneities and topography irregularities on ground motion in the case of a strong earthquake. Interestingly, the largest amplification of peak ground acceleration and Arias intensity does not necessarily occur at the same sites where peak ground velocity and flux of seismic energy reach their highest values, depending on the frequency band of amplification. A magnitude 7 earthquake at a distance of 100 km results in peak ground accelerations ranging from 30 to 70 gals while peak ground velocities are estimated to vary from 5 to 7 cm/s; moreover, simulated time histories of horizontal ground motion yield amplitudes of 5% damped pseudovelocity response spectra as large as 15-20 cm/s for frequencies from 1to 3 Hz. In this frequency band, the mean value is 7 cm/s for firm sites and ranges from 10 to 13 cm/s for soil sites. All these results are in good agreement with predictions based on regressions of Italian and Western North American data.