Investigation of false-positive liquefaction case history sites in Christchurch, New Zealand

Abstract

Over 25,000 cone penetration test (CPT) soundings collected in Christchurch, New Zealand are archived on the New Zealand Geotechnical Database (NZGD) and have allowed for detailed comparisons between post-earthquake land performance during the 2010-2011 Canterbury Earthquake Sequence and CPT-based retrospective predictions of liquefaction severity. Initial comparisons resulted in a number of false-positive assessments in which liquefaction triggering was predicted by the simplified CPT-based method, but no liquefaction damage was observed. A detailed in-situ site characterization study was initiated in August 2015 in order to investigate this initial observation. Thirty one sites were selected and subjected to a testing program including: (1) seismic CPT (SCPT), (2) high-resolution, direct-push crosshole (DPCH) testing, and (3) continuous soil sampling via sonic drilling. Results from the in-situ site characterization program, as well as the CPT data archived on the NZGD, allowed for a comprehensive investigation of both CPT- and V[subscript S]-based liquefaction triggering analyses at each site. After performing the standard, simplified, deterministic CPT- and VS-based procedures, refinements were made to address several additional factors that may have contributed to the over-prediction of liquefaction severity, including: (1) site-specific soil plasticity and fines content, (2) partially saturated soils below the water table (as indicated by compression wave velocity), (3) coarse-to-fine-grained soil interlayering, (4) non-liquefiable crust thickness, and (5) soil microstructure. Observed liquefaction severity at each site was compared to the retrospective predictions from the CPT- and V[subscript S]-based liquefaction triggering procedures, both before and after the refinements for additional factors. It was found that the refinements to the standard CPT-based liquefaction triggering procedure, primarily due to adjustments for site-specific fines contents, significantly increase the number of correct predictions of liquefaction severity. However, they also slightly increase the number of under-predictions. Several case history sites are addressed in detail in this thesis.