Empirical relationships are developed to predict amplification factors for 5% damped spectral acceleration (period range T = 0.01 – 5 s) as a function of site category. Amplification is evaluated by normalizing ground motion intensity measures from recordings by reference motions derived from modified attenuation relationships for active regions. The Abrahamson and Silva attenuation relationship for rock sites was used for the derivation of reference motions, with modifications to account for event terms and rupture directivity effects.
Strong motion sites are classified according to three geologic classification schemes: age-only, age + depositional environment, and age + material texture. Sites are also classified using the average shear wave velocity over the upper 30 m (Vs-30) and a recently proposed geotechnical classification scheme. Within each scheme, amplification of spectral acceleration is regressed against reference motion amplitude, and the magnitude-dependence of the residuals is evaluated.
The results of the regression indicate distinct levels of high-frequency spectral acceleration amplification across geologic age categories, but relatively modest variations between categories at long periods (T ≥ ∼ 1 s). Within the Quaternary age group, statistically significant variations in high-frequency amplification are observed between Holocene lacustrine/marine and Quaternary alluvial sediments, and also between Holocene coarse- and fine/mixed-texture sediments. Nonlinear ground response is evident in many categories from statistically significant decreases in low-period spectral amplification with increasing reference motion amplitude. Amplification of long-period ground motions is found to be less sensitive to the reference amplitude, but to increase significantly with magnitude.
We find spectral acceleration amplification functions for geologic classification schemes that incorporate information beyond age to have a smaller average residual dispersion than age- only schemes. Classification schemes based on Vs-30, and geotechnical data produced consistently higher dispersion than did detailed geologic classifications. However, the Vs-30-based classification scheme provides the clearest distinction between amplification factors in different categories. These findings have implications for the type of mapping that is most useful for regional ground motion characterizations.
Spectral acceleration amplification levels at all ground motion amplitudes and period ranges are found to be smaller than those in modern design codes (e.g., 2000 NEHRP Provisions). This iv is attributed to differences in the characteristics of “rock” sites used to develop reference motions for the studies underlying the code provisions and this study. The rock site condition associated with this study is a composite average of rock sites in active regions — which most nearly corresponds to “soft rock.” Since the same rock site condition is inherent to rock attenuation relations, the amplification functi ons derived herein provide an appropriate means by which to adjust the statistical moments (median and standard error) from rock attenuation relations (applicable to active regions) for the effects of site conditions. Most previous studies, including the work that provides the empirical basis for the current NEHRP provisions, employ reference motions from relatively firm rock sites. The relatively large reference motion amplitudes from soft rock sites lead to the smaller amplification levels.
Full List of PEER Reports: click here.
