Type of presentation: both

 

Site response for generic profiles in the Mississippi Embayment

Salome Romero, salome.romero@ce.gatech.edu, Glenn J. Rix, School of Civil and Environmental Engineering, Georgia Institute of Technology, Atlanta, GA 30332-0355, WALTER J. SILVA, Pacific Engineering and Analysis, 311 Pomona Avenue, El Cerrito, CA 94530

The thick deposits of unconsolidated, low-velocity sediments within the Mississippi embayment have a significant effect on site response in the Central United States. In particular, the deep soil column is susceptible to ground motion amplification. Current site amplification factors recommended by the 1997 NEHRP provisions are based primarily on the geology of the Western U.S. and may not adequately reflect the influence of the geology of the Central U.S. on site response. Therefore, generic shear wave velocity (Vs) profiles were developed for Holocene age alluvial deposits (Lowlands) and Pleistocene age deposits (Uplands) in the Mississippi embayment. The generic Vs profiles are based on measured profiles for near-surface deposits (0 to 70 meters) and on estimated and modeled profiles for deeper deposits (70 to 1000 meters). Additionally, the effect on site response of a near-surface, high-velocity, cemented layer often encountered in the Memphis, Tennessee area was analyzed for the Uplands profile.

The generic profiles were randomized to account for the inherent spatial variability in Vs and used to compute site amplification factors for different geologic environments in the Mississippi embayment. Input rock motions were generated by the stochastic approach described by Boore (1983) based on the Brune source model and scaled using random vibration theory to estimate the Fourier amplitude spectrum. The k-filter proposed by Anderson and Hough (1984) was used to shape of the Fourier amplitude spectra at high frequencies. Parameters such as the stress drop (Ds), corner frequency (fc), and the spectral decay parameter (k) were selected from published studies of the seismicity of the Central U.S. Site response was calculated using an equivalent-linear approach encoded in the computer program RASCALS (Silva and Lee, 1987; Schneider et al., 1993). Since RASCALS functions in the frequency domain, it is computationally efficient incorporating probability without the need for conducting Monte Carlo simulations in the time domain. Site amplification factors for response spectra were developed for the generic profiles and compared with those recommended by NEHRP.

This is a student paper