Development of an empirical correlation for predicting shear wave velocity of Christchurch soils from cone penetration test data

• Empirical correlation between shear wave velocity and CPT data is developed.
• Correlation is applicable for soils of Christchurch, New Zealand.
• Several regression functional forms are considered.
• Multiple linear regression with nonconstant variance is used.
• Correlation is assessed with comparison of estimated and observed velocity profiles.

Following the companion study of McGann et al. [1], seismic piezocone (SCPTu) data compiled from sites in Christchurch, New Zealand area are used with multiple linear regression to develop a Christchurch-specific empirical correlation for use in predicting soil shear wave velocities, Vs, from cone penetration test (CPT) data. An appropriate regression functional form is selected through an evaluation of the residuals for regression models developed with the Christchurch SCPTu database using functional forms adopted by previous empirical correlations between Vs and CPT data. An examination of how the residuals for the chosen regression form vary with the predictor variables identifies the need for non-constant depth variance in the regression model. The performance of the model is assessed through comparisons of predicted and observed Vs profiles and through forward predictions with synthetic CPT data. The new CPT–Vs correlation provides a method to estimate Vs from CPT data that is specific to the non-gravel soils of the Christchurch region in their current state (caution should be used for western portions of the Springston Formation where SCPTu data were sparse). The correlation also enables the utilization of the large, high-density database of CPT logs (>15,000>15,000 as of 1/1/2014) in the Christchurch region for the development of both site-specific and region-wide models of surficial Vs for use in site characterization and site response analysis.