Abstract: Using CPT data from the Chi-Chi earthquake, CPT-based liquefaction discrimination methods proposed by Robertson and by Olsen are inspected. The prediction success ratios of the two methods are 82.61% and 80.43% for the liquefied sites, but 31.82% and 44.32% for the non-liquefied sites, respectively. CPT-based methods are reliable for liquefied soils, but not effective for non-liquefied soils. For comparison, the SPT-based liquefaction discrimination method recommended by National Center for Earthquake Engineering Research of USA is tested using data from Chi-Chi earthquake, the prediction success ratio of the procedure is 92.41% and 94.35% for liquefied and non-liquefied sites, respectively. The SPT-based method demonstrates higher prediction success ratio than that CPT-based methods. However, CPT-based soil type chart can reflect soil strength and soil type simultaneously, and can differentiate liquefied soils from non-liquefied soils in Chi-Chi earthquake. Moreover, for the preliminary discrimination of clayey soils liquefaction, CPT-based soil type chart is also better than the clay content which has been always used.

Key words: Chi-Chi earthquake, liquefaction, cone penetration test, standard penetration test