← Back to Gallery

Comparison of State-of-the-Art Approaches Used to Account for Vs Variability in 1D Ground Response Analyses

Over the past decade, a significant amount of seismic site response research has focused on our ability to replicate recorded ground motions at borehole array sites in the U.S., Japan, and elsewhere globally, where both the input (rock) and output (surface) ground motions are known. While findings from these studies vary somewhat, on average more than 50% of borehole array sites are poorly modeled using 1D ground response analyses (GRAs) based on a single shear wave velocity (Vs) profile, presumably due to the variable subsurface conditions at most sites. Even though the engineering community is aware of these limitations, 1D GRAs continue to be overwhelmingly used in practice. As such, a crucial task is to explore whether site response predictions can be improved through modifications to traditional 1D GRAs.

We investigate five alternative approaches that can be used to account for spatial variability in 1D GRAs: (1) stochastic Vs randomization, (2) stochastic shear wave cumulative travel time randomization, (3) utilization of Vs profile suites derived from surface wave testing covering a large spatial area, (4) incorporation of pseudo-3D Vs models derived from a new horizontal-to-vertical spectral ratio (H/V) geostatistical approach, and (5) damping modifications. These approaches are investigated at two U.S. downhole array sites (Treasure Island and Delaney Park) and the results are compared with recorded small-strain ground motions. Through qualitative and quantitative comparisons, we find that using pseudo-3D Vs models derived from the H/V geostatistical approach best predicts the site response at both sites. However, all methods tend to overestimate fundamental mode amplification, apparently failing to fully account for wave scattering effects. This study suggests that while attempting to account for spatial variability in 1D GRAs is important and can in fact improve site response predictions at most sites, existing approaches work better at sites that are not very complex, and it is likely that 3D GRAs are needed at sites that have significant spatial variability.

Presenting Author: Mohamad M. Hallal

Student Presenter: Yes

Day: 4/21/2021

Time: 2:00 PM - 3:15 PM Pacific

Additional Authors