Assessment of Ground Motion Variability in Urban Seismic Hazard Zones

Thesis Abstract

In recent years, the increasing frequency and severity of seismic events in urban areas have underscored the critical need to understand ground motion variability and its implications for seismic hazard assessment and urban resilience. Despite substantial advances in seismology, the heterogeneity of ground motion across seismic hazard zones remains inadequately characterized, particularly concerning how local soil conditions, geological heterogeneity, and site-specific factors influence ground motion amplitudes and frequencies. This study aims to comprehensively assess the variability of ground motion in urban seismic hazard zones, with specific objectives including quantifying spatial differences in ground motion levels, identifying key factors influencing variability, and developing predictive models for site-specific ground response. The research adopts a mixed-methods approach, integrating empirical field data collection with quantitative analytical techniques. The study sample comprises 150 seismographic stations strategically distributed within identified urban seismic hazard zones, selected through stratified random sampling to capture variability across different geological and soil conditions. Ground motion data are collected using broadband seismometers with a sampling rate of 100 Hz, installed at designated sites over a period of 12 months to capture multiple seismic events, including microseisms, small-magnitude earthquakes, and larger regional quakes. Metadata such as soil type, depth to bedrock, seismic source characteristics, and local site effects are collected through geotechnical surveys and geophysical logs. Data analysis involves a combination of descriptive statistics, spatial analysis, and advanced modeling techniques. Descriptive statistics are computed to characterize the extent of ground motion variability, while spatial variance is analyzed using Geographic Information Systems (GIS) combined with variogram models. Regression analysis, specifically multiple linear regression and principal component analysis (PCA), is employed to identify and quantify the influence of site-specific factors on ground motion amplitudes. In addition, the study employs spectral analysis and wavelet transforms to assess frequency-dependent characteristics of ground motion variability. The empirical findings are interpreted in the context of well-established seismological theories, notably the Site Response Theory and the Convolution Model of ground motion, to explain observed heterogeneities and develop site-specific ground motion models. The expected outcomes of this research include a detailed characterization of ground motion variability within urban seismic zones, identification of predominant controlling factors, and the development of spatially explicit predictive models for local ground motion amplitudes. These models are anticipated to enhance current seismic hazard assessments by integrating site-specific variability, thus providing more accurate risk estimates for urban planning and seismic engineering. The study is expected to reveal significant heterogeneity in ground motion amplitudes driven by geological and soil conditions, with variations exceeding 30% across different urban sites. The results will demonstrate that local site effects account for a substantial portion of the observed variability, as evidenced by regression models with an adjusted R-squared exceeding 0.75. The research contributes to the existing body of knowledge by providing a refined understanding of ground motion heterogeneity at a granular spatial scale, thereby addressing existing gaps related to site-specific seismic risk evaluation. Main conclusions highlight the importance of incorporating local ground motion variability into seismic hazard models to mitigate urban seismic risk effectively. Recommendations include integrating site-specific ground motion models in regional seismic zoning, enhancing geotechnical site investigations prior to urban development, and adopting canonical ground motion spectra tailored for different soil classes. Future research should focus on extending the models to include nonlinear soil behavior and exploring the influence of seismic source characteristics on ground motion variability. Overall, this study aims to provide vital empirical insights that can inform seismic risk mitigation strategies and improve resilience planning for urban populations vulnerable to seismic hazards.