A Taxonomy for Site Complexity Using the HVSR: Application to the KiK-Net Database

Session: Near-Surface Effects: Advances in Site Response Estimation and Its Applications
Type: Oral
Date: 4/28/2020
Time: 09:00 AM
Room: 110 + 140
Description: 

Earthquake ground motions are amplified at varying intensities and frequencies as a function of site geology, an effect known as site response. Some common properties that cause site effects are (1) high impedance contrasts, which cause amplification due to energy conservation across an interface; (2) basin geometry, which causes 3D basin effects from surface wave energy compounding on a basin edge; (3) and geologic complexity, which causes wave scattering and dissipation of energy. Partitioning these different sources of site response uncertainty is essential to characterizing a basin and ultimately predicting future site amplifications from different, potentially unrecorded, earthquake sources. Using data from Japan’s Kiban-Kyoshin (KiK-net) network of vertical seismometer arrays, Thompson et al. (2012) developed a taxonomy for identifying sites that are well predicted by 1D assumptions and thus can be used to calibrate and validate 1D constitutive models. They attribute sites that are poorly modeled by 1D assumptions to “site complexity,” which can be attributed to errors in assumed soil properties, limitations in the constitutive model and/or wave propagation assumptions like non-vertical incidence. Because surface-borehole seismic stations such as those in the KiK-net seismic array are not the norm for seismic characterization, there is a need to translate the site complexity taxonomy to single station (i.e. surface-only) site characterization. In this study, we compare horizontal to vertical spectral ratio (HVSR) quasi-transfer functions (Nakamura 1989; 2019) from 114 stations in the KiK-net seismic array to the surface-downhole empirical transfer functions used by Thompson et al. 2012 to establish a single surface station taxonomy for complex site response. Using the same sites and ground motions as Kaklamanos and Bradley (2018), we adapt the Thompson et al. 2012 rules for interevent variability and correlation to the 1D transfer function and propose additional site complexity metrics that are appropriate for the HVSR quasi-transfer functions.

Presenting Author: Marshall A. Pontrelli

Authors