UC Davis

The performance of Çark Canal in the 1999 M=7.5 Kocaeli earthquake was evaluated using common liquefaction vulnerability index (LVI) methods, a nonlinear dynamic analysis (NDA) method, and a Newmark sliding block method to examine possible factors contributing to why different analysis approaches often overestimate liquefaction effects in interbedded deposits. The characterization of the interbedded fluvial stratum based on cone penetration test (CPT) data utilized an inverse filtering procedure to correct CPT data for thin-layer and transition zone effects. Common LVIs computed using the measured and inverse-filtered CPT data with a site-specific fines content calibration showed that the combination of these two steps reduced the LVIs by 30%-50% for this site and seismic loading. Two-dimensional NDAs were performed using stochastic realizations for the interbedded stratum and the PM4Sand and PM4Silt constitutive models for the sandlike and claylike portions, respectively. Computed deformations were evaluated for their sensitivity to stochastic model parameters, the cyclic strength assigned to the sandlike soils, the undrained shear strengths assigned to the claylike soils, the level of shaking, and other input parameters. Newmark sliding block analyses were performed with different allowances for the influence of interbedding on the composite strength of the interbedded stratum. The differences between results obtained with these analysis methods, along with those presented by others, provided insights into how the various factors can contribute to an overestimation of ground deformations in interbedded deposits of sands, silts, and clays.