Assessing Primary Ground-Motion Site-Response Parameters from Borehole-Array Records

Abstract:

Earthquake ground-motion site response is a major concern in engineering seismology. Currently, the time-weighted-average shear-wave velocity for the top 30 m of soils and rock, Vs30, is used as the key parameter to account for site response in engineering design and other applications. Recent studies have shown, however, that Vs30 is inadequate because it does not physically and uniquely correlate with site response. Site response depends on the shear-wave velocity structure and associated parameters such as density, damping, and nonlinearity. Site response can be uniquely quantified by two site-specific parameters—the fundamental (i.e., base mode) site frequency, f₀, or period, T_f (i.e., 1/f₀), and the corresponding amplification, A₀. We used selected earthquake recordings from two borehole arrays in the central United States, nine borehole arrays in the western United States, and six borehole arrays in Japan to compare empirical and theoretical methods for determining the fundamental site frequency and its peak amplification. The empirical methods were (1) spectral ratios (i.e., transfer function) between the surface and bedrock S-wave recordings and (2) horizontal-to-vertical spectral ratio (HVSR) of S-waves. The theoretical methods were (1) linear, (2) equivalent linear, and (3) nonlinear site-response analyses. We observed that at several deep (> 100 m) soil sites where peak amplifications occur at frequencies greater than f₀ (i.e., f_peak > f₀), f_peak migrates to fundamental frequencies under strong shaking. Using empirical measurements, we also evaluated new site-correction factors that depend on A₀ and T_f for the design response spectra using borehole data.