Geostatistical models for the assessment of the influence of shear wave velocity uncertainty and variability on ground response analyses

A robust assessment of the influence of epistemic uncertainties and aleatory variabilities on ground response analyses is a fundamental requirement for modern probabilistic seismic hazard analyses. The study of epistemic uncertainties and aleatory variabilities should follow three sequential steps: identification, quantification, and management (IQM procedure). A fundamental input for ground response analyses is the shear-wave velocity (VS) profile. The VS profile can be obtained via various types of geophysical tests. Each test has specific characteristics. For these reasons, a consistent IQM procedure must be established to assess the type and amount of uncertainties involved in the measurements, also addressing the peculiarities of the investigated site. We propose a geostatistical model for the management of the uncertainties included in the VS profiles obtained with surface wave tests. The model is based on the separate randomization of travel times and layer thicknesses. It generates profiles that are consistent with the measured “site signatures” such as the Rayleigh wave dispersion curve and the predominant site period. The model can be easily generalized for other geophysical tests.