UCLA

Global sea levels may rise up to five meters by 2150, elevating coastal groundwater and intensifying cascading hazards. Little attention has been given to how changing groundwater interacts with earthquake-induced soil liquefaction. Coastal California's 25 million residents and critical infrastructure sit on loose, liquefiable soils. Because liquefaction occurs primarily in saturated, sand-like soils, groundwater elevation strongly governs the hazard. Yet the standard of practice assumes a fixed groundwater depth observed during a site investigation, an assumption that may not hold where sea level rise (SLR) and tides dynamically control the water table. The term liquefaction was first used by Hazen in 1919, and earthquake-induced liquefaction became a major focus of geotechnical research after the 1964 Niigata and Alaska earthquakes. Although probabilistic seismic hazard analysis is standard in earthquake engineering, an equivalent treatment has not been widely adopted for liquefaction because few probabilistic liquefaction hazard analysis (PLHA) tools exist. PLHA accounts for uncertainty in ground motion, earthquake load, and soil resistance. This dissertation presents a PLHA framework, implemented as the open source Python package ucla_plha, that incorporates the liquefaction term into the seismic hazard calculation. Traditional stress-based methods cannot assign a return period to a computed factor of safety; by evaluating liquefaction within the hazard integral, the framework determines the factor of safety associated with a selected return period, which may differ from the corresponding seismic hazard return period. The analysis uses the Uniform California Earthquake Rupture Forecast Version 3 fault source model, four VS30 dependent NGA-West2 ground motion models, and five liquefaction triggering models. I contributed to the UCLA Coastal Geotechnical Database, populated with cone penetration tests and borings from the California Geological Survey. Pore pressure outputs from two-dimensional groundwater models were propagated through the PLHA, and liquefaction hazard was evaluated using high performance computing along 150 coastal transects under multiple SLR scenarios. Results were synthesized into maps of Santa Monica and Moss Landing. Analyses show that SLR of up to 2 m can increase the rate of liquefaction up to a factor of ten. These findings support treating groundwater depth as a distribution and evaluating liquefaction probabilistically.