Seismic gravelly soil liquefaction assessment based on dynamic penetration test using expanded case history dataset

The seismic liquefaction has been observed in gravelly soils, in addition to sandy soils. Despite sandy soils, there is still a shortage of an extended case history database for developing empirical, semi-empirical, and probabilistic models to predict this phenomenon in gravelly soils. This study examines the documentations of several case histories of gravelly soil liquefaction all around the world to create a database, and then to develop probabilistic models to consider uncertainties of the models as well as the parameters for evaluating gravelly soil liquefaction triggering caused by earthquakes. The logistic regression and Bayesian mapping function, both of which are based on the maximizing likelihood estimation, were applied to present classifier curves to predict the occurrence of liquefaction. Additionally, a sensitivity analysis on the bias sampling weighting factor was performed to assess its effect on the model's prediction accuracy. The results point to the effect of extended database and sampling bias on the developed models. Meanwhile, this study highlights the importance of developing probabilistic models rather than deterministic ones to consider uncertainties.