Advancements in 1D and 2D near-surface seismic site characterization using surface waves and full waveform inversion
Seismic site characterization is a critical part of understanding earthquake hazards in geotechnical engineering. This is often accomplished through various invasive and non-invasive methods for measuring shear-wave velocity (Vs) in-situ, as it is directly related to small-strain shear modulus. For civil engineering applications, the seismic conditions of the near surface (top 30 m) are of particular interest. Surface wave testing has become the tool of choice for many engineers due to its flexibility, efficiency, and ability to characterize a wide variety of subsurface conditions. Surface wave testing is also particularly well suited to near-surface imaging due to the prevalence of surface waves within the elastic wavefield at shallow depths. Surface wave testing, however, is not without limitations. Inversion of surface wave dispersion data is ill-posed and non-unique, meaning that when it is performed rigorously with full consideration of epistemic uncertainty, a potentially large number of reasonable and different 1D Vs profiles are produced. This presents a challenge of evaluating which profiles should be used for further analysis or design. Additionally, engineers often desire information about the lateral variability of seismic parameters in the subsurface, but the inherently 1D nature of the processing and inversion techniques used in surface wave testing make acquiring this information challenging. Evaluation of lateral variability is generally accomplished through multiple individual 1D surface wave analyses across the site, providing only pseudo-2D information. This also introduces a new challenge: how to collect the large amount of experimental data required for multiple analyses as the efficiency of traditional surface wave acquisition is limited by the need to physically move geophone arrays with limited numbers of sensors. This dissertation discusses these challenges and presents potential solutions through the application of the DeltaVs method, distributed acoustic sensing, and full waveform inversion.