Careful characterization of seismic hazard is especially important in structural engineering if critical facilities (e.g., strategic buildings, chemical deposits, energy production plants, etc.) are the object of analysis. Moreover, if the construction site is close to active faults, particular attention is required. In fact, in this condition, the fault's dynamics affect ground motion differently from site to site, resulting in systematic spatial variability, and, generally, in higher seismic demand, with respect to the far-source case. The most important of the near-source (NS) effects is known as forward rupture directivity, and can be identified by a large full-cycle pulse at the beginning of the velocity records, and containing the most of carried energy. Recent research demonstrates that, from the structural engineering point of view, hazard assessment should account for NS effects (i.e., pulse-like ground motions), but ordinary probabilistic seismic hazard analysis (PSHA) is not able to do it appropriately. On the other hand, semi-empirical models calibrated for introducing NS effects in classical PSHA are now available, and some preliminary attempts of numerical implementation of NS-PSHA exist. In the presented study, numerical applications of strike-slip (SS) fault scenarios are provided and, for a fixed return period, uniform hazard spectra are computed in order to quantify hazard increments (HIs) due to NS-PSHA with respect to ordinary-PSHA. It is shown that: i) depending on source-site geometry, these may be significant (more than a 100% increment); ii) different spectral periods are affected by significantly different values of HIs, which have their shape controlled only by the magnitude of generated earthquakes; iii) it is possible to identify a zone beyond which directivity effects are expected to become negligible, at least with a first-order approximation, independently of the characteristics of the considered SS fault.
Sensitivity analysis of directivity effects on PSHA
Abstract: