Ground Motion Variability From Finite-Fault Simulations

Abstract eng:
The standard deviation (sigma) of the Ground Motion Prediction Equations (GMPEs) has a crucial impact on the results of seismic hazard analysis, especially for long return periods. Recently, great efforts have been made to find strategies to reduce sigma values, focusing on the different components that contribute to the ground-motion variability. This is an important task, since if the sources of ground motion variability are recognized, they could be accounted for the epistemic components of the uncertainties, thus reducing the aleatory component. To explore contributions related to site and propagations effects, specific dataset composed of records at the same site from different earthquakes or at multiple sites from events restricted in a given source-region can be used. Strong-motion data can be hardly used to investigate the variability from a single fault, since different events on the same source have been recorded very rarely. To overcome this limitation, synthetic seismograms can represent a valid alternative to build a fault-specific dataset. In recent years, the use of numerical simulations is increasing and a number of initiatives worldwide promote their application for hazard assessment purposes. When numerical simulations are used to predict future ground motions, a large number of possible earthquake scenarios on the same fault can be considered, varying the model parameters of the source-rupture process. Through the massive calculation of synthetic seismograms, the expected ground motion and associated variability at the site of interest can then be evaluated. However, to date, there is no standard approach to quantify and treat the generated ground-motion variability. In this paper, we explore the use of numerical simulations based on kinematic rupture models in order to treat the parametric variability of expected ground motions. Following the strategy adopted in the GMPE community, we generate synthetic dataset for single and multiple sites in the proximity of a single fault, considering numerous rupture scenarios. We establish a framework for treating the different components of the variability related to the synthetic datasets and quantify the contribution related to rupture scenarios and to the spatial distribution of sites with respect to the source. Since we use three simulation methods, we also evaluate how these components depend on the numerical approach.

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