The goals of this project are two-fold. First, we have extended the scope of the original 2G01 project to include the preparation of a final written project report documenting considerations associated with the selection of calibration sites and comparison of ground response predictions to other ground motion prediction techniques. Included in this report is discussion of a technique for the interpretation of ground response results suggested by Norman Abrahamson that involves the estimation of response spectra as the product of the input target spectrum and the median RRS (ratio of response spectra from ground response analysis). Second, we are validating Walt Silva’s surface-geology based amplification factors against strong motion data.

The project is helping develop statistical methods for evaluating ground response effects on spectral acceleration, which is an important IM within the PEER methodology.

The scope associated with the first objective consisted of completing the project web page and final report.

The scope associated with the second objective involves (a) estimating the depth to the 1 km/s isosurface for southern California and Bay Area sites (referred to subsequently as z1), (b) estimating the approximate uncertainty on z1, (c) using these parameters and available surface geology classifications to predict amplification factors (per Silva’s model) and associated response spectral ordinates, and (d) identifying any statistical bias in the results within surface geology categories (by comparing the predicted spectra to data).

The 2G01 project was successfully completed within the last funding cycle. An abstract of this work follows:

When geotechnical 1-D ground response analyses are performed to evaluate site effects in lieu of attenuation relations, it is with the expectation that the attenuation model’s standard deviation, and any bias in the median, would be reduced. In the 2G01 project, we evaluated the degree to which these and other benefits are realized and developed recommendations for implementing the results of equivalent-linear, 1-D ground response analyses into hazard calculations. This is accomplished by comparing Sa from recordings to predictions derived using ground response analysis procedures as well as attenuation relationships with and without amplification factors. The results are compiled for 134 motions from 68 sites, and prediction residuals are interpreted to assess the models’ relative bias and dispersion. We find that ground response analyses are unbiased for T < ~1 s, but underestimate longer period Sa in deep basins. For soft soils, ground response analyses reduce dispersion for T < 1 s relative to alternative models. This dispersion reduction is not observed for other site categories nor at longer periods. These results suggest that ground response analyses are beneficial for Sa predictions at soft soil sites, but generally provide no identifiable benefit for stiff site conditions.

The expected Year 6 accomplishment is the validation of the Silva factors. This will be presented in a comprehensive report.

There is some interest within SCEC in the basin parameter uncertainty issue, although no formal project is underway to investigate this. A SCEC workshop is planned later this year to investigate this issue.

Project 2G02 is in the planning stages, and will investigate the use of nonlinear ground response procedures for estimation of site effects. The scope will otherwise be similar to 2G01.

I believe the 2G01 results have been circulated within Caltrans and PG&E and used to assist engineers in decision making regarding when ground response analyses are worthwhile.

1. Preparation of main project report for 2G01 (complete)
2. Documentation of basin depth parameters and surface geology classifications for strong motion accelerograph sites in the PEER database that are located in the San Francisco Bay Area and Los Angeles. This classifications will be provided to Walt Silva and other interested parties for review when available (complete).
3. Preparation of short report on Silva model validation.

Deliverables for the work would include the main project report (i.e., the first “goal” item listed above) and a separate short report on the Silva model validation.