Comparative Analysis of Seismic Velocity Models in Urban vs. Rural Seismic Zones

Thesis Abstract

This study investigates how seismic velocity models differ between urban and rural seismic zones, addressing the problem that conventional one-size-fits-all velocity models may mischaracterize site effects and misinform hazard assessment in heterogeneous landscapes. The aim is to develop and compare context-specific seismic velocity models that enhance ground-motion prediction and seismic hazard estimation for urban and rural settings. Specific objectives include (1) compiling and harmonizing a dataset of seismic refraction and surface-wave measurements from representative urban and rural sites, (2) deriving 1D and 3D velocity models using multi-method inversion techniques, (3) assessing the influence of anthropogenic and geological factors on velocity distributions through statistical modeling, (4) evaluating the transferability of velocity models between urban and rural contexts, and (5) formulating recommendations for regionally adaptive ground-motion prediction equations. The study adopts a comparative cross-sectional design anchored in theories of site response and geotechnical seismology, drawing on the Earth’s structural heterogeneity framework and the equivalence principle in seismology to interpret velocity contrasts. The population comprises seismic survey data from five metropolitan areas and five undeveloped or sparsely populated regions within a tectonically active country, spanning shallow to intermediate crustal depths (0–30 km). A stratified sampling approach selects 60 urban sites and 60 rural sites to ensure balanced representation of soil types, depth to bedrock, and urbanization intensity. Data collection instruments include portable refraction seismographs, ambient-noise tomography arrays, and multi-channel analysis of surface waves (MASW) for dispersion measurements, complemented by borehole logs, soil classification, and anthropogenic activity indicators. Instrument calibration, quality control, and cross-site standardization procedures are implemented to ensure measurement consistency. Validity and reliability are addressed through inter-site calibration, replicate measurements, and benchmarking against well-characterized reference sites. The analytical framework integrates multiple methods (i) joint inversion of travel-time and dispersion data to derive 1D Vs profiles and subsequent 3D velocity tomographies; (ii) Bayesian inversion to quantify uncertainties and compare model credibility between urban and rural sets; (iii) regression and generalized linear models to relate velocity parameters to proxies of urbanization (building density, traffic) and geological factors (lithology, depth to discontinuities); (iv) analysis of variance (ANOVA) to test differences in velocity statistics across settings; and (v) sensitivity analyses to identify dominant drivers of velocity contrasts. A conceptual model linking urbanization intensity, near-surface geotechnical properties, and seismic velocity distributions is proposed, with results interpreted in the context of site amplification and ground-motion variability. Expected findings include systematic higher shallow velocities in rural zones due to well-compacted soils and bedrock exposure, contrasted with more complex urban velocity structures caused by anthropogenic modifications and heterogeneous fill, leading to distinct dispersion signatures and shallower low-velocity zones beneath cities. The study anticipates that urban models will exhibit greater uncertainty in shallow-depth estimates owing to noise and heterogeneous materials, while rural models will show clearer stratification aligned with geologic layering. The contribution to knowledge lies in providing empirically derived, context-specific seismic velocity frameworks for urban and rural regions, informing regionally calibrated ground-motion prediction equations, and improving seismic hazard models through explicit incorporation of urbanization-driven velocity heterogeneity. The main conclusion posits that urban-rural dichotomies in seismic velocity are pronounced at shallow depths and persist in deeper structures, but differences diminish where lithology dominates. Recommendations include adopting regionally adaptive velocity models in hazard assessments, integrating urban-specific priors in Bayesian inversions, and expanding datasets to include emergent urban growth areas and peri-urban zones to refine transferability and uncertainty quantification.