Semi-Empirical Nonlinear Site Amplification and its Application in NEHRP Site Factors, PEER Report 2013-13


Site factors are used to modify ground motions from a reference rock site condition to reflect the influence of geologic conditions at the site of inte rest. Site factors typically have a small-strain (linear) site amplification that captures impedance and resonance effects coupl ed with nonlinear components. Site factors in current NEHRP Provisions are empirically-derived at relatively small ground motion levels and feature simulation-based nonlinearity. We show that NEHRP site factors have discrepancies with respect to the site terms in the 2008 Next Generation Attenuation (NGA) ground motion prediction equations, both in the linear site amplification (especially for Classes B, C, D, and E) and the degree of nonlinearity (Classes C and D). The misfits are towards larger linear site factors and stronger nonlinearity in the NEHRP factors. The differences in linear site factors result largely from their normalization to a reference average shear wave velocity in the upper 30 m (VS30) of about 1050 m/sec, whereas the reference velocity for current application is 760 m/sec. We show that the levels of nonlinearity in the NEHRP factors are generally stronger than simulation- and empirically-based models used in the original (2008) Next Generation Attenuation (NGA) project.

We analyze the NGA-West 2 data set to evaluate site amplification both with respect to VS30-scaling and nonlinearity. The motivation of this analysis was to support the development of a proposal for revising the NEHRP s ite factors and a site amplification model that is used in an NGA-West 2 GMPE (Boore et al., 2014; BEA14). The development of that site amplification model is described by Seyhan and Stewart (2014; SS14); this report presents supporting work that guided model development relative to regional variations in site amplification and levels of nonlinearity implied by simulations.

We investigated regional trends in VS30-scaling and found the presence of such trends to be sensitive to data selection criteria. When only data at fault distances under 80 km were selected (motivated by avoiding complications from regional variations in anelastic attenuation), we found strong regional site
amplification trends, with the Japanese data showing markedly weaker VS30-scaling than other regions at short periods. Regional variations are less significant at mid- and long-periods. When data from greater distances were considered (up to approximately 400 km) with appropriate corrections for regional variations inanelastic attenuation, regional variations in VS30-scaling are diminished to the point that the proposed site amplification model (in SS14) does not include a regional term.

We investigated the regionalization of nonlinearity in site amplification using NGA-West 2 data. While modest between-region variations are present in the data, the trends were not considered sufficiently robust to be included in the site amplification model. Levels of nonlinearity evaluated empirically were found to generally be similar to those implied by a simulation-based model (Kamai et al., 2014; KEA14), except for pseudo spectral accelerations at periods between0.5 and 3.0 sec where the data exhibits more nonlinearity than is evident from the simulations. Both the empirical- and simulation-based nonlinearities were considered in the development of the nonlinear component of the SS14 site amplification model.

The complete site amplification model (for VS30-scaling and nonlinearity) is used to derive new NEHRP site factors using a reference velo city of 760 m/sec. For relatively weak levels of shaking, the new NEHRP site factors are generally smaller than current values due to the change in reference velocity from 1050 to 760 m/sec. For stronger shaking levels and Class C and D soils, the new site factors are close to, or exceed, those used currently because of reduced levels of nonlinearity, especially at long period (i.e., in the Fv parameter). Factors for soft soil (Class E) were set conservatively, as were the orig inal NEHRP site factors, to account for larger epistemic uncertainty in the nonlinearity for this site class as compared to others. Other than Class E, the new NEHRP site factors match the BEA14 site terms nearly exactly and are generally consistent with site amplification models in other NGA-West 2 GMPEs as well. The new site factors have been approved by the Provisions Update Committee of the Building Seismic Safety Council and are expected to appear in the 2015 version of the NEHRP Provisions.

Full List of PEER Reports: click here.

Jonathan P. Stewart
Emel Seyhan
Publication date: 
June 12, 2013
Publication type: 
Technical Report
Stewart , J.P., & Seyhan, E. (2013). Semi-Empirical Nonlinear Site Amplification and its Application in NEHRP Site Factors, PEER Report 2013-13. Pacific Earthquake Engineering Research Center, University of California, Berkeley, CA.