An Alternative to Stochastic Ground-motion Relations for Use in Seismic Hazard Analysis in Eastern North America

Abstract

The most critical and controversial input parameter to a seismic hazard analysis is generally the regional ground-motion relations, which define the peak ground motion and response spectral parameters as functions of earthquake magnitude and distance. For seismic hazard applications in California, these relations have been derived by regression analysis of recorded strong-motion data ( e.g., Boore et al., 1997; Abrahamson and Silva, 1997; Sadigh et al., 1997). For regions beyond California, such as eastern North America (ENA), regional ground-motion relations cannot be reliably derived from direct regression of strong-motion data, because these data are too sparse in the magnitude-distance range of engineering interest ( M 5 to 8 at distances less than 200 km). For this reason, alternative methods of deriving ground-motion relations have been employed. In eastern North America, nearly all recently published ground-motion relations have been derived using the stochastic method ( e.g., Atkinson and Boore, 1995; Frankel et al., 1996; Toro et al., 1997). The stochastic method is a simple and effective tool for developing ground-motion relations, whereby earthquake ground motions are treated as bandlimited Gaussian noise, whose amplitude spectrum is given by a seismological model of earthquake source and propagation processes. The model has been shown to reproduce successfully the sparse ENA database (Atkinson and Boore, 1995, 1998), and has also been demonstrated to fit the more extensive California database (Boore, 1983; Boore and Joyner, 1997; Atkinson and Silva, 2000). Nevertheless, there is significant uncertainty involved in using ground-motion relations that are based largely on a model whose accuracy for large-magnitude earthquakes can be truly known only after many more years of gathering ground-motion data in ENA. Uncertainty plays a major role in seismic hazard analysis, and the evaluation of uncertainty is an important product of such an analysis ( e.g., EPRI, 1993 …