Local and regional evaluation of liquefaction potential index and liquefaction severity number for liquefaction-induced sand boils in Pohang, South Korea

Abstract

An earthquake of moment magnitude 5.4 occurred on November 15, 2017 in Pohang, South Korea. This earthquake is the second largest instrumented earthquake ever recorded in South Korea and caused extensive damage to the ground and infrastructures. Hundreds of liquefaction-induced sand boils were observed in the rice fields near the epicenter. We calculate the liquefaction potential index (LPI) and liquefaction severity number (LSN) using the existing borehole data from the area, which includes 166 soil borings, to predict local and regional sand boils. Three of the soil borings considered in the dataset are within 47 m from the sand boils, and one is within 174 m from the sand boil. We provide detailed soil profiles at 20 sites (four of which are considered to be at liquefied sites) and soil properties at three sites (two of which are at liquefied sites). The calculated LPI and LSN values are in good agreement with the actual sand boil locations at 19 out of the 20 sites. Then, we use the existing borehole data to produce a regional liquefaction map using interpolation methods. As is common in regional seismic hazard maps, geotechnical and ground motion data are limited; therefore, we assess the uncertainty of important input parameters, such as the peak ground acceleration, water table, and N-value. For the regional maps, the maximum accuracy of the LPI is 69% at an LPI threshold of 2, and that of the LSN is 71% at an LSN threshold of 10. We report some ground settlements in the areas with high LPI and LSN values. Furthermore, we propose models for predicting the sizes and occurrence probabilities of sand boils.