Regional Calibration of CPT Correlations for Shear Wave Velocity in Budapest Soils

Abstract

Small-strain stiffness and shear wave velocity (Vs) are key parameters in geotechnical and seismic design, yet direct in situ seismic measurements are often unavailable in routine site investigations. This study evaluates and recalibrates CPT-based Vs correlations for the geological conditions of Budapest, Hungary, using a database of 1304 paired CPT and seismic measurements from 15 urban sites. Forty-two published CPT-Vs correlations were examined and the 25 best-performing relationships retained for detailed statistical assessment. Most of them underpredicted the measured velocities. Site-level hypothesis testing confirmed a significant underprediction for the Robertson (2009) framework over the whole database, and for every best-performing correlation above about 400 m/s, where mean bias ratios fall to 0.72–0.91. The best existing relationships were those of Robertson (2009), Andrus et al. (2007) and Wolf (2018), but recalibrating the Robertson (2009-qc1N) framework improved on them. New Budapest-specific correlations are proposed for all soil types represented in the database and, separately, for Quaternary and Tertiary deposits. The all-soils correlation reduced the RMSE from 64.1 to 56.7 m/s and improved the mean bias ratio Kmean from 0.92 to 1.02. Leave-one-site-out cross-validation confirmed that the coefficients are not overfit, the out-of-sample RMSE increasing by only 2.2–6.3%. The deviation from global correlations is interpreted in terms of local soil microstructure: measured-to-estimated velocity ratios and normalized small-strain rigidity indices indicate ageing, overconsolidation and cementation. The proposed correlations reproduced the Eurocode 8 ground type at 13 of the 15 sites from CPT-derived Vs profiles, supporting their use for seismic site characterization and microzonation in Budapest.