Exploring Alternative Pathways for Modelling Site Response for Seismic Risk at a Regional Scale using Insights from Strong Motion Databases

Date: 4/26/2019

Time: 06:00 PM

Room: Grand Ballroom

Characterisation of site amplification is a critical element of any assessment of seismic risk. At a regional scale the limited and variable spatial coverage of quantitative site properties mean that approaches to model site amplification may differ from those encountered at a site-specific level, and careful treatment of uncertainty is needed in this case. Recent regional seismic risk models have modelled sites in terms of a “proxy” 30m-average shearwave velocity (V_S30) inferred from topographic slope, or similar geomorphological information. Whilst this allows for V_S30 to be input into ground motion models at regional or even global scales, rarely is the inherent uncertainty of this parameter incorporated faithfully into seismic risk calculations. Seeking an approach for site response modelling that is applicable at a continental scale, we explore the correlations between observed strong motion recordings and various proxies that can be mapped at a large scale. Using strong motion data from Japan and Europe, taking observed site amplification as a period-dependent station-to-station random effect from a mixed-effects regression of a ground motion model, it is shown that amplification can be characterized as a direct function of topography (without the need to derive or infer V_S30) and is conditioned upon the geological setting. This approach can ensure that the higher uncertainty is propagated into the risk calculation, whilst still achieving the desired spatial coverage. Its feasibility for use in a pan-European seismic risk analysis is explored and comparisons made with more conventional site-specific approaches for predicting amplification at selected test sites.

Presenting Author: Sreeram Reddy Kotha

Authors