Pshae (Probabilistic Seismic Hazard Enhanced): the Case of Santiago in Chile and Istanbul

Abstract eng:
The Probabilistic Seismic Hazard Analysis (PSHA) only relying on ground motion prediction equations (GMPEs) tends to be insufficiently constrained at short distances and data only partially account for the rupture process, seismic wave propagation and three-dimensional (3D) complex configurations. Given a large and representative set of ground shaking scenarios from 3D physics-based numerical simulations, analyzing the resulting database from a statistical point of view and implementing the results as a generalized attenuation function (GAF) into the classical PSHA might be an appealing way to deal with this problem [1]. Nonetheless, the limited amount of computational resources or time available tend to pose substantial constraints to a broad application of the previous method. Furthermore, the method is only partially suitable for taking into account the spatial correlation of ground motion as modelled by each forward physics-based simulation (PBS). In this context, we envision a streamlined and novel approach for enhanced PSHA (“PSHAe”) in large urban areas, alternative to the previous method, based on the integration into the probabilistic framework of a limited number of deterministic scenarios, being the latter wisely chosen and associated to a probability of occurrence. The experience gathered in the past year regarding 3D modelling of seismic wave propagation in complex alluvial basin (see e.g.: [2], [3] and [4]) allowed us to enhance the choice of simulated scenarios in order to explore the variability of ground motion, preserving the full spatial correlation necessary for risk modelling. 3D numerical modelling of scenarios occurring along the North Anatolian Fault in the proximity of Istanbul are carried out through the spectral element code SPEED (http://speed.mox.polimi.it). The results are introduced in a PSHA based code, exploiting the capabilities of this methodology.

• Export as: BibTeX | MARC | MARCXML | DC | EndNote | NLM | RefWorks
• Add to your basket