UC Berkeley

Liquefaction damage from the 2010-2011 Canterbury earthquake sequence devastated parts of Christchurch, New Zealand. There were many sites where state-of-practice liquefaction assessment procedures indicated liquefaction would be expected to occur, and surface manifestations of liquefaction were observed. However, there were also numerous sites, which were predominantly silty soil sites, where state-of-practice liquefaction assessment procedures indicated that liquefaction would be expected to occur, but no surface manifestations of liquefaction were observed. This discrepancy between state-of-practice liquefaction assessments and post-earthquake liquefaction observations led to the development of the research program presented in this dissertation. Several silty soil sites were selected for investigation to further our understanding of fine-grained soil liquefaction response and to evaluate potential limitations in the current state-of-practice liquefaction assessment procedures, which are based primarily on case histories and laboratory testing of sands. This dissertation investigates the liquefaction response of silty soil sites through no-liquefaction case histories from the Canterbury earthquake sequence, evaluating depositional environment effects on observed liquefaction performance, site characterization of silty soil deposits, and laboratory testing to characterize element-scale cyclic response.

Depositional environment effects are evaluated through regional CPT-based analyses and site-specific comparisons. Stratified silty soil swamp deposits are shown to have mitigating effects on the manifestation of liquefaction beyond what can be captured by simplified liquefaction assessment procedures in Christchurch. Differing surficial geology and depositional environments are found through examining historical documents to explain in part the limitations of current liquefaction assessment procedures in the swamps of southwest Christchurch, which contain stratified silt/sand deposits or thick silt layers. Consideration of depositional environment distinguishes between liquefaction performances that could not be differentiated through the CPT-based assessment alone. CPT resolution is not sufficient to capture the thin layering at these stratified sites, and the simplified liquefaction assessment methods do not take into account the effects of the stratification on pore water pressure movement within a soil profile. Continuous sampling and careful logging of high-quality samples provides important insights on in-situ stratification at these silty soil swamp sites, discerning differences in stratigraphy resulting from differences in depositional environment.

Site investigation techniques are evaluated at the silty soil case history sites to discern their capability to characterize thin layers and groundwater table fluctuation, two potential causes for the discrepancies between state-of-practice liquefaction assessments and post-earthquake liquefaction observations. CPT, mini-CPT, sonic borings, and high-quality sampling are critiqued in terms of their ability to capture thin layer stratigraphy, which is of importance for liquefaction assessment. Piezometers, sonic borings, high-quality sampling, crosshole testing, and regional groundwater maps are evaluated to assess their ability to capture groundwater table fluctuation. CPT, mini-CPT, and conventional sonic borings offer important information for site characterization, but they do not capture full details of thin layering at silty soil sites. Detailed logging of high-quality samples captures the actual in-situ layering that helps explain limitations of simplified liquefaction assessment procedures. Use of multiple groundwater measurement methods more fully illuminate fluctuating groundwater conditions. Subsurface investigation programs should utilize tools that characterize features impacting liquefaction potential in adequate detail for the intended engineering purpose. Use of multiple, complementary investigation techniques provides the most robust assessment.

A field investigation and advanced laboratory testing program was conducted in Christchurch. High-quality samples were obtained using a Dames & Moore hydraulic fixed-piston thin-walled sampler for cyclic triaxial testing to characterize the liquefaction response of silty soils at the no-liquefaction sites in southwest Christchurch. These natural silty soil specimens contained heterogeneity and variability that should be considered and is difficult, if not impossible, to replicated with laboratory-prepared specimens. Test results for stress-strain response and axial strain accumulation indicate a nuanced range of transitional responses for these intermediate soils. Post-liquefaction reconsolidation testing shows clear differences in specimen response, ranging from "sand-like" immediate reconsolidation to time-dependent reconsolidation. Simplified liquefaction assessment procedures estimate significant liquefaction at these case history sites and yet no liquefaction manifestations were observed during the Canterbury earthquake sequence. Laboratory estimates of cyclic resistance (CRR) are consistent with estimates from the simplified procedures, and both estimates of CRR are well below simplified procedure estimates of seismic demand (CSR). Depositional characteristics such as thin-layering of fine sand and silt may be why manifestations of liquefaction were not observed at these sites. Post-liquefaction reconsolidation testing provides insight that water and ejecta may not accumulate in these stratified silty soils as they would accumulate in thick deposits of liquefiable clean sands. Additional mitigating factors may also contribute to the discrepancy between simplified procedure estimates of liquefaction and the lack of liquefaction observed at these sites. The interaction of several factors contributing to observed liquefaction response at these silty soil sites indicates that in-situ “system” response should be considered and that further research on silty soils is warranted.