Fault Directivity and Seismic Hazard

Abstract

In planning the design of structures in a region of potential seismic activity, a specification of the “strength” of the earthquake ground motion, or the most likelihood ground motion level, is needed. The earthquake occurrence, and its effects, is described as a stochastic process. Thus its realization is linked to state variables defined over a a known space through a continuous function. The Ground Motion Predictive Equation (GMPE) realize this function and, despite its shortcoming as an effective design tool to control damage (Priestly, 2003), it is still the most widely used representation of earthquake ground motion employed in engineering practice. As a consequence the majority of hazard estimations are based on the GMPE providing a ground motion specification as a function of a certain number of variables. In fact in many situation there are not enough data to allow a direct empirical specification of ground motion. Only few regions, i.e. Japan, have strong-motion network and data-banks sufficient to carry out seismic hazard assessment without the benefit of regionally-derived ground motion predictive model. The central role they hold in the hazard assessment motivates the recent efforts in better synthesize all available regional informations and general knowledge about earthquakes. The representation of the ground motion through the GMPE is simple compared to the complexity of the physical process involved. If only the magnitude and distance are taken into account, the GMPEs predict isoseismal curves that are expected to be isotropic around the hypocenter and uniform if no other effects are considered (i.e. site effects). Instead, the presence of a fault plane, across which a process of failure in shear develops, make this general formulation divert from the observations on a specific case. In fact the dynamic propagation of rupture results in anisotropy effects not included in the predictions although back-analyses of ground motions from past earthquakes have shown that such effects have a strong influence on the spatial distribution of ground motion.Although the anisotropy effects resulting from the propagation of rupture have been generally recognized and finally incorporated in predictions, its effect has not been tested yet in an hazard context. On the contrary, all the aforementioned issues motivate an in depth analysis of its contribution on the present tools of seismic hazard assessment. This work is mainly addressed to conduct such analysis. One guidance is provided answering to the following questions: Does directivity improves the performance of ground motion prediction in real time applications? Is directivity still effective in a PSHA framework? What deterministic hazard model can tell about directivity ?