Seabed shear-wave velocity estimation by interface-wave dispersion using different inversion methods

This paper applies both linearized and nonlinear inversion methods to estimation of shear-wave velocity profiles using interface-wave dispersion curves extracted from multi-component ocean bottom seismic data. For the nonlinear inversion, model selection based on the Bayesian Information Criterion is carried out to determine the most appropriate parameterization for seabed shear-wave velocity profile supported by the data. Five different parameterizations for the shear-wave velocity profile are studied including models consisting of 3 to 6 uniform-velocity layers over a halfspace and a dynamic power-law model over a halfspace. The parameterization with the 4-layer over a halfspace is selected as the preferred model. The inversion results by different model parameterizations and different inversion methods are compared and discussed. The penetration depth by the nonlinear inversion with the preferred model parameterization is about 180 m down to the sediments which is larger than that by linearized inversion method.