Re-calibrating the thresholds for the classification of liquefaction potential index based on the 2004 Niigata-ken Chuetsu earthquake

• The LPI scale depends on different analysis methods for soil liquefaction.
• We describe re-calibrating the LPI and its application to seismic hazard mapping.
• For Chinese criteria, 50% of liquefied sites show LPI > 15.
• For Bray–Sancio criteria, 50% of liquefied sites show LPI > 21.
• Our LPI scale and mapping method were able to predict approx. 70% of damaged areas.

The liquefaction potential index (LPI) has been applied with increasing frequency to assess the potential for liquefaction-induced ground failures. This paper describes strategies used to re-calibrate the LPI using accepted methods for screening the liquefaction potential and evaluates its application to the seismic hazard mapping of political tracts. In this study soil profiles using standard penetration tests were evaluated using two methods, the Chinese criteria and the Bray–Sancio criteria. The LPI values were then calculated from these profiles at non-liquefied and liquefied sites in Nagaoka City during the 2004 Niigata-ken Chuetsu earthquake in Japan. Our results suggest that 50% of the liquefied sites exhibited LPI values > 15, and 75% of the non-liquefied sites exhibited LPI values < 11 when applying the Chinese criteria. For the Bray–Sancio criteria, 50% of the liquefied sites and 75% of the non-liquefied sites exhibited LPI values > 21 and < 14, respectively. We then attempted to map liquefaction hazards using the alternative LPI thresholds (low hazard if LPI < 14; moderate if 14 ≤ LPI < 21; high if 21 ≤ LPI) and applied regression analyses between the LPI and the controlling variables (i.e., groundwater depth and peak ground acceleration). Based on the comparison of the predictions with the liquefaction inventory for the 2004 earthquake, our LPI thresholds and mapping method were able to predict ~ 70% of damaged areas.