Calibrating a standard penetration test based method for region-specific liquefaction potential assessment

Youd's method is a widely adopted standard penetration test based approach for evaluation of liquefaction potential. This method was established by pooling the post-earthquake survey data collected from different regions of the world, and its accuracy may vary when applied in different regions. This paper investigates the inter-region variability associated with the accuracy of Youd's method using a Bayesian model. It is found that Youd's method is on average biased towards conservatism, and there does exist the inter-region variability associated with the uncertainty of the model bias factor. Nevertheless, the inter-region variability associated with the model uncertainty of Youd's method is less significant than the intra-region variability. Due to the existence of inter-region variability, the same factor of safety (FOS) may imply different levels of liquefaction probabilities in different regions. If the inter-region variability is not considered, the obtained FOS-failure probability relationship may lead to underestimation or overestimation of the liquefaction potential when applied in different regions. A verification method is proposed to compare the predicted number of liquefied cases with the actual number of liquefied cases in each region. An illustrative example for assessing the liquefaction potential is used to demonstrate the application of the proposed method.