Practical Considerations Regarding the Probability of Liquefaction in Engineering Design

Engineering practitioners have widely adopted probabilistic methods for estimating liquefaction triggering hazards. Unfortunately, incorrect application of these methods has become commonly accepted in the United States and throughout many parts of the world. In this paper, we describe two specific incorrect applications of probabilistic liquefaction triggering prediction models that are commonly used in engineering practice: (1) improper parametric uncertainty characterization when using the models, and (2) the application of a conservative factor of safety thresholds to deterministic cyclic resistance ratio boundaries. These two practices and their implications are discussed, and illustrative examples and figures are presented. We also discuss the apparent inconsistencies in what constitutes an acceptable risk for liquefaction hazards and other significant geotechnical hazards, such as slope stability. We demonstrate the need to collectively consider the likelihoods of all aspects of liquefaction hazard problems-susceptibility, initiation, and effects-when calculating and considering liquefaction risk. Finally, we point to the concept and framework of performance-based earthquake engineering as a logical and objective path forward for dealing with the challenges raised in this study.