Testing the SH1D Assumption for Geotechnical Site and Basin Response Using 3D Finite Difference Modeling

Abstract:

Current state-of-practice of geotechnical site response and soil-structure analyses generally assume a vertically propagating horizontally polarized plane wave is incident on a plane-layered (one-dimensional) soil column. Ground motions representing the wavefield incident to the bedrock base of the soil column are developed from observed and sometimes scaled time-histories or synthesized by various methods. The site-specific ground motion at the surface is then computed from the response of the soil column to the bedrock incident wavefield, possibly including non-linear response of the geotechnical near-surface. This is the so-called SH1D assumption. While this approach is widely used, it ignores important complexities of the incident wavefield. Specifically, the standard approach assumes: 1) the incident wavefield is only composed of vertically propagating body waves; 2) ignores oblique incidence; and 3) neglects the three-component nature of the wavefield that includes surface waves and rotational motions. Surface waves often carry much of the seismic energy and can excite all three components of motion. Therefore, it seems most appropriate to include the most representative characterization of the incident wavefield in site-specific analyses. We are performing parametric studies with three-dimensional (3D) elastic finite difference simulations to compare the near-surface response of sedimentary basins to horizontally polarized planes (arbitrary incident) and point source (double couple) earthquakes. Simulations involve simple, parametric representations of basin geometries and layered material properties of the sedimentary basin and surrounding hard rock. We compare the frequency-dependent site response for different excitations and attempt to quantify the differences between the plane-wave and fully 3D basin response.