Reassessing Site Effects in Idaho National Laboratory in Light of New Data

Abstract:

The impact of local site conditions on the amplification of ground motions and associated damage to infrastructure has long been known. Modeling site effects is a fundamental step in the development of a site-specific seismic hazard curve and soil surface spectra in the seismic design of nuclear facilities. Five accelerometers have been installed near facilities at Idaho National Laboratory (INL) since 2013. A large number of ground motions from regional earthquakes (magnitudes 3.0 to 5.0) have been recorded, including the January 3, 2015, M5.0 Challis, ID, earthquake, providing a valuable dataset for studying site response at INL, especially wave propagation in a unique profile consisting of basalt that is fractured and interbedded with unconsolidated sediments. Previous site-response analyses conducted at INL have been based on one-dimensional (1D) equivalent-linear analysis using statistically generated soil profiles with shallow depths (up to 65 ft) (e.g., Payne and Costantino, 2003). While these analyses followed appropriate codes and standards for developing soil design spectra, they did not consider 1) representativeness of statistically generated soil velocity profiles, 2) the influence of strong velocity reversal below the elastic half-space boundary, and 3) two-dimensional (2D) or three-dimensional (3D) scattering effects due to sedimentary interbeds. The objective of this study is to investigate the significance of these issues using recent ground motion observations, with the ultimate goal of informing future site response and seismic hazard analyses. In this study, we first evaluate the effectiveness of the soil amplification function currently employed in seismic hazard analysis at INL by comparing it to the amplification observed from recorded ground motions. We then predict site response from alternative 1D models using measured shear wave velocity profiles with and without consideration of basalt-sediment sequence. Finally, we investigate scattering effects by using a numerical model which incorporates 3D heterogeneity of basalt and sedimentary interbeds. We observe that sedimentary interbeds can significantly reduce the amplification between 5 and 20 Hz, and the amplification function currently employed results in great conservatism.