Assessment of liquefaction potential using simplified method and one-dimensional effective stress ground response analysis during 2017 Pohang earthquake in South Korea: A case study

Abstract

We assessed the liquefaction potentials at six profiles where sand boils were observed during the 2017 Pohang earthquake in South Korea, which had a moment magnitude (M) of 5.5. Two of the sites are located within 2 km from the epicenter, whereas the third one is located 8 km away. To predict the onset of liquefaction, we used both the simplified cyclic stress-based method and one-dimensional (1D) effective stress (ES) ground response analysis (GRA). A major source of uncertainty in applying the simplified method to M < 7.5 earthquakes is determining the magnitude scaling factor (MSF). We tested four empirical MSF relationships. All MSF equations produced similar predictions for profiles where peak ground acceleration of input motion (a(max)) > 0.15g and cyclic stress ratio (CSR) > 0.2. However, at profiles with a(max) < 0.15g and CSR < 0.2, the intensity-dependent MSF provided most reliable predictions of the liquefaction potential, whereas the other three equations overestimated the cyclic resistance ratio. The ES GRAs were conducted using accumulated stress- and strain-based pore pressure models implemented in a 1D GRA program. One key advantage of the ES GRA over the cyclic stress-based method is that it does not require an empirical MSF. The stress-based model produced higher pore pressure estimates than the strain-based model, yielding correct predictions for five out of six profiles. The strain-based model, highly sensitive to the shear wave velocity (V-S) profile, tended to underestimate pore pressure for a(max) < 0.15g, suggesting caution when using this model for moderate-intensity motions. Among the two sets of input parameters applied to the strain-based model, the set conditioned on V-S yielded the lowest pore pressure predictions and is therefore not recommended.