Verification of correctness of using real part of complex root as Rayleigh-wave phase velocity with synthetic data

High-frequency (≥ 2 Hz) Rayleigh-wave phase velocities have been utilized to determine shear-wave velocities in near-surface geophysics since the early 1980s. One of the key steps is to calculate theoretical dispersion curves of an earth model. When the earth model contains a low-velocity half-space, however, some roots of the dispersion equation turn out to be complex numbers, which makes phase velocities disappear at some frequencies. When encountering this situation, the common practice is to append an additional high velocity layer as the half-space to the model to make the roots real or use the real parts of complex roots as Rayleigh-wave phase velocities. The correctness of the first method has been verified. The correctness of the second method, however, remains to be unproved. We use synthetic data generated by numerical modeling of the wave equation to verify the correctness of the second method. In this paper, we firstly discuss the reasons that only complex numbers of the dispersion equation exist at some frequencies when an earth model contains a low velocity half-space. Then we discuss how the nearest offset affects a synthetic model and recommend an optimal nearest offset in generating synthetic data that are close to real-world situations. Several synthetic models are used to verify correctness of using real parts of complex roots as Rayleigh-wave phase velocities when an earth model contains a low velocity layer as the half-space.

► A low-velocity half-space will bring nonexistence of real roots of dispersion equation.
► The use of real parts of complex roots as phase velocities solves this problem.
► We verified the correctness of this method using synthetic data.
► An optimal nearest offset in generating synthetic data was recommended.
► We justify an idea that substitutes a multi-layer model with a three-layer model.