Technical Basis for Revision of Regulatory Guidance on Design Ground Motions: Development of Hazard- & Risk-Consistent Seismic Spectra for Two Sites (NUREG/CR-6769)

On this page:

Download complete document

Publication Information

Manuscript Completed: March 2002
Date Published: April 2002

Prepared by:
R. K. McGuire1
W. J. Silva2
C. J. Costantino3

1Risk Engineering, Inc.,
Principal Contractor
4155 Darley Avenue,
Suite A
Boulder, CO 80305

2Pacific Engineering & Analysis
311 Pomona Avenue
El Cerrito, CA 94530

3Carl J. Costantino, Consultant
4 Rockingham Road Spring Valley, NY 10977

R. M. Kenneally, NRC Project Manager

Prepared for:
Division of Engineering Technology
Office of Nuclear Regulatory Research
U.S. Nuclear Regulatory Commission
Washington, DC 20555-0001

NRC Job Code W6248

Availability Notice


We develop recommendations for design spectra at two sites, one in the Mojave desert, California, and the second at Columbia, South Carolina. These sites were chosen because local, small earthquakes dominate the high frequencies (f≥10 Hz), but large distant events dominate the low frequencies (f≤10 Hz). Both rock and soil conditions are examined at each site.

For rock conditions, the uniform hazard spectrum (UHS) is determined at each site with a probabilistic seismic hazard analysis (PSHA). The hazard at 10 Hz and 1 Hz is deaggregated to determine the dominant magnitude M and distance R, and these values are used to generate two sets of spectral shapes. The first set comes from the recommended functions documented in McGuire et al. (2001); the second set comes from the attenuation equations used in the PSHA. In the CEUS there are separate shapes for the 1- and 2-corner seismic source model, and these are weighted using weights justified in the PSHA. The two sets of spectral shapes are scaled to the UHS amplitudes at 10 Hz and 1 Hz, as a consistency check on the shape of the UHS.

We calculate a scale factor to derive a rock uniform reliability spectrum (URS) based on the slopes of the hazard curves across the frequency range of interest at each site. The URS achieves an approximately consistent annual frequency of plant component seismic failures for all sites and across all structural frequencies. For these examples the 10-4 URS is illustrated by scaling the 10-4 UHS. The attenuation equation spectral shapes derived from the UHS are scaled to the 10 Hz and 1 Hz URS amplitudes. If the scaled spectra match the URS within a designated criterion, the scaled spectra may be used as separate design motions. This will be more accurate and realistic for sites where a broad-banded earthquake motion is not likely.

The database of strong motion records provides a source of rock motions with the correct magnitudes and distances. To develop design motions, these records are used as the starting point to develop artificial records fit to the individual scaled spectra. Matching criteria are applied to ensure compatibility between the target spectra and the artificial motions.

For soil sites, we illustrate the development of design spectra using a profile of the Meloland station in California assumed to lie at the Mojave site, and a generalized profile of the Savannah River site in South Carolina assumed to lie at the Columbia site. Soil amplification is calculated for these two sites using an equivalent-linear formulation of dynamic soil response, and using as input the rock motions calculated from the PSHA. For the Mojave site it is necessary to remove the effects of the shallow soft-rock velocity gradient to a depth corresponding to a shear-wave velocity of 4000 ft/sec, in order to provide an accurate input to the base of the soil column. We calculate soil amplification factors for rock motions corresponding to the 10-4 and 10-5 hazard, accounting for uncertainties in soil properties and documenting the uncertainty in soil response. From these calculations we can determine with sufficient accuracy the 10-4 and 10-5 UHS on soil. This accuracy is illustrated with a separate calculation of the soil hazard, using soil-specific attenuation equations developed specifically for the two profiles studied here. From the UHS on soil we derive the 10-4 URS on soil.

We scale soil spectra to the 10 Hz and 1 Hz UHS and URS, using the soil-specific amplification studies, because generic shapes for soil sites are not appropriate. The soil-specific shapes are scaled to the UHS to check consistency, and are scaled to the URS as optional design spectra. If these scaled shapes are to be used for design, they must match the URS within a stated criterion.

Artificial motions for soil sites are created in a manner similar to that for rock sites. The database of records includes soil motions for the WUS and the CEUS, and records with the appropriate magnitudes and distances are adjusted to match the target spectra (either a broad-banded spectrum or individual scaled spectra).

Overall, the procedures recommended in McGuire et al. (2001) work well in developing design spectra for the rock and soil sites examined here. Care must be taken in calculating the URS from the UHS, and in determining soil response given a rock PSHA, but sufficient consistency checks are illustrated so that one can make a determination of the validity of the final recommended spectra.


McGuire, R.K., W.J. Silva, and C. Costantino (2001). Tech. Basis for Rev. of Reg. Guidance on Dsgn. Grnd. Motions: Hazard and Risk-consistent Grnd. Motion Spectra Guidelines, US Nuc. Reg. Comm., Rept. NUREG/CR-6728, Oct.

Page Last Reviewed/Updated Thursday, March 25, 2021