The return period of soil liquefaction
Soil liquefaction is one of the most damaging effects of earthquake shaking. Current methods of predicting liquefaction have proven to be effective for a given intensity of earthquake shaking but, because they consider only a single intensity at a time, their ability to communicate total liquefaction hazard is limited.The goal of this research is to improve the assessment of seismically-induced soil liquefaction potential by including the entire range of the hazard due to ground shaking and by describing the liquefaction potential in a form that can be more easily used by other stakeholders in hazard and risk analysis. This has been accomplished by implementing two methods of liquefaction analysis into the performance-based earthquake engineering framework developed by the Pacific Earthquake Engineering Research Center. In this framework, a probabilistic liquefaction analysis is combined with a probabilistic seismic hazard analysis in a manner that accounts for the hazard from all possible earthquake scenarios with proper consideration of their likelihoods of occurrence.The first method is based on the widely-used cyclic stress ratio as the measure of ground shaking intensity and soil resistance, for which the necessary data and predictive relationships are readily available. Implementation of this method, however, required modification of the scalar PEER, framework to accommodate the magnitude-distance vector of loading information that forms the cyclic stress ratio. The second method is based on the recentlydeveloped variant of cumulative absolute velocity, CA5, for which no liquefaction analysis procedures exist. A collection of case-histories and ground motion records was assembled and modeled with logistic regression to produce a relationship that predicts liquefaction probability from CAV5 and soil penetration resistance.The research shows that current methods of liquefaction analysis produce inconsistent estimates of the liquefaction hazard across different seismic environments. This occurs because they consider only a single ground shaking intensity whereas the total liquefaction hazard is composed of contributions from all shaking intensities. These inconsistent hazard estimates result in the application of different levels of safety to different locations. Because the performance-based approach incorporates all shaking intensities, it produces consistent estimates of liquefaction hazard across different seismic environments. The future use of performance-based liquefaction analyses is discussed.The new CAV5 method offers several conceptual advantages over the cyclic stress method, but its full implementation into the performance-based framework is hampered at this time by the lack of a robust attenuation relationship needed for the development of CAV 5 seismic hazard curves.
- Civil engineering