Liquefaction in gravelly soils has often been overlooked in the past; recent case histories have shown that it can cause significant damage during major earthquakes. The dynamic cone penetration test (DPT) is a practical tool for evaluating liquefaction potential of gravelly soils. In this study, new DPT-based probabilistic triggering curves for gravelly soils are developed based on a global database reevaluated under a unified framework. The models are formulated within a Bayesian framework, explicitly accounting for input parameter uncertainties, model errors, gravel content (GC), earthquake magnitude (Mw), and effective overburden stress (σvo′). A new magnitude scaling factor (MSF), overburden-pressure correction factor (Kσ), and GC-related correction to the DPT blow counts are derived from the models. Comparative analyses demonstrate that the proposed models provide reliable performance, with Model 1 showing advantages for high GC cases and yielding physically interpretable parameters. These developments enhance the applicability of liquefaction assessments in gravelly soils and support seismic design.
Zhao et al. (2026) studied this question.