Empirical relationships are developed to predict amplification factors for 5% damped response spectral acceleration as a function of site condition. Amplification factors are evaluated as residuals between ground motion recordings and predictions from modified rock attenuation relationships.
Both shallow and deep characteristics of site condition are considered to identify those parameters that are most effective from the standpoint of bias and dispersion reduction. The parameterization of shallow site condition is based on (1) surface geology, (2) NEHRP classification, (3) geotechnical site categories, and (4) average shear wave velocity in upper 30 m of site (Vs-30). Also considered are parameters that reflect the relatively deep sedimentary structure at many of the strong motion sites, including depth to the 1.5 km/s shear wave isosurface (z1.5) as well as the location of the source inside or outside of the basin in which the site is located. Sites located in a basin overlaying the source are denoted as having coincident source and site basin locations (CBL) and are differentiated from distinct source and site basin locations (DBL).
It is found that standard devi ation is minimized with the use of detailed surface geology or Vs-30 as the site parameter for shallow site condition. The
Vs-30-based amplification model has several innovative features, including Vs-30-dependent nonlinearity and standard deviation and quantification of the reference velocity for a number of rock attenuation relationships. For all site categories, standard deviation was found to increase with period, being as much as 0.3 larger at long periods than short periods.
The work on basin parameters utilized residuals calculated with respect to ground motion predictions derived using rock attenuation relations coupled with amplification factors for shallow site condition. Models relating amplification to z1.5 were developed for the CBL and DBL data groups. The results indicate that the use of basin models is generally worthwhile for periods T ≥ 0.75 s. At those long periods, residuals are significantly sensitive to z 1.5 for CBL but not for DBL when shallow site condition is parameterized based on Vs-30. Standard deviation is also reduced at long periods, such that the increase with period is significantly reduced.
Full List of PEER Reports: click here.