Feasibility of Uniformly Applicable Basin Amplification Models for the United States

Date: 4/24/2019

Time: 06:00 PM

Room: Grand Ballroom

We utilize an extended dataset of subduction and crustal events, recorded in different basins around the world to estimate empirical basin amplification factors. The analysis aims to resolve potential differences in basin amplification between different types of events (crustal and subduction) and investigate the possibility of creating a uniform basin amplification model that could be applied in seismic hazard studies. For the crustal event dataset, which includes recordings from Taiwan and Japan, we observed strongest trends in the residuals for 1, 7.5 and 10s. Weaker amplification was observed in the period range 2-5s. A comparison with the amplification function proposed for California in Boore et al. (2014) shows a similar increase in the amplification with increasing spectral period, but the magnitude of the amplification developed from our dataset is lower. Our initial analysis of basin response in Japan from subduction events (Skarlatoudis et al., 2015) indicated basin amplifications comparable to those obtained in the Seattle region by Frankel et al. (2018). However, an updated analysis that we performed for the subduction event dataset resulted in amplifications close to 1 for the whole period range (1 to 10s). While the results we obtain should not be outright discounted, we believe that the basin amplification for subduction events should not be drastically different from the crustal events and should not disagree so strongly with other empirical models. Therefore, we consider our residual analysis as inconclusive for the subduction events. These results could be attributed to the specific dataset that was used in the analysis, which may not be adequate to resolve such trends. This result could also imply that grouping together recordings from basins around the world, with different properties and geometries, is not an effective procedure for developing average, empirically estimated, basin amplification factors. As a step forward, a larger database would allow a more robust statistical analysis and better constrain the contribution of path effects.

Presenting Author: Andreas Skarlatoudis

Authors