Collaborations

Accounting for the inevitable uncertainty in shear wave velocity (Vs) profiles has been recognized to be critical to seismic ground response analysis (GRA) and probabilistic seismic hazard analysis (PSHA). Vs stochastic randomization models are among the most widely used approaches to account for this uncertainty. However, the suitability of these models, as currently implemented, has been undermined by their inability to match observations at vertical seismometer arrays.

This motivated modifications, such as randomizing the cumulative shear wave travel time (tts) instead of directly Vs. While some preliminary assessments suggest that tts randomization allows for better management of Vs uncertainty, the framework has only been implemented in a few case studies. The aim of this study is to fill this gap by scrutinizing the different steps of tts randomization.

By using real experimental data and Monte Carlo simulations, we argue that some of the steps of this framework pose some challenges. For each challenge, we devise recommendations and demonstrate that they provide a clear improvement on the current framework. The findings of this study will enhance our understanding of the tts randomization framework and will enable researchers and engineers to adopt it robustly and confidently at engineering scales.

A site’s fundamental resonance frequency (f0), or its reciprocal, the fundamental period (T0), is a critical parameter in seismic studies due to its proven ability to aid in predicting local site effects. The horizontal-to-vertical spectral ratio (HVSR) method is a popular nonintrusive technique that can be used to estimate f0 in a time-efficient and cost-effective manner.

Although it is becoming more common to perform several HVSR measurements, the measurements often are irregularly spaced due to access and/or budget restrictions. This has the potential of introducing significant bias when attempting to estimate a single, representative f0 across an area of interest. To address this problem, we propose the use of Voronoi tessellations to obtain an unbiased, statistical representation of f0 or T0 from spatially distributed HVSR measurements.

Three example applications are presented to illustrate potential uses. The first application demonstrates the effectiveness of the adopted approach in correcting for bias introduced by irregular spatial sampling. The second application illustrates how better-informed seismic site classifications can be made using a statistical representation of T0. The third application compares the relative degree of spatial variability in f0 at two downhole array sites to assess the applicability of performing one-dimensional ground response analyses.