SPT-Based Evaluations of Seismic Soil Liquefaction Initiation and Severity of Its Ground Surface Manifestation

Abstract This paper introduces probabilistic predictive models for assessing the likelihood of soil liquefaction initiation and the severity of its manifestation at the ground surface of a free field site. The development of these correlations incorporates several advancements: (1) an expanded database of field performance case histories, (2) improved interpretation of standard penetration test (SPT) data, including updated corrections for overburden, energy efficiency, and short rod effects, (3) refined adjustments for site-specific earthquake ground motions and cyclic stress ratio (CSR), accounting for directivity and site effects, (4) rigorous screening of field case histories to ensure high quality data and reduced uncertainty, (5) predictive models that explicitly and jointly assess the likelihood of liquefaction initiation and its ground surface manifestation, (6) a database representativeness index quantifying the degree to which the site and seismic conditions under design considerations are well represented within the case history database, and (7) a Bayesian inference framework to refine model predictions. The resulting models address challenges in jointly modeling liquefaction triggering and its severity at the ground surface, while incorporating adjustments for variability in (1) earthquake duration, (2) vertical effective stress, and (3) amount of fines.