Joint Inversion for Radial Anisotropy With Multimodal Rayleigh and Love Wave Dispersion Curves

Abstract

Radial anisotropy measures the difference between horizontally and vertically polarized shear waves, which can give important constraints on crustal deformation with depth. Ambient noise surface wave tomography has become a main tool to effectively study the radial anisotropy structure in which Rayleigh and love dispersion curves extracted from ambient noise data are utilized to invert radial anisotropy parameter. However, only fundamental-mode dispersion curves are used in traditional methods, for which the inversion results could be easily affected by severe non-uniqueness of inversion. Recently, frequency-Bessel transform method has been validated that it can effectively extract multimodal Rayleigh and Love wave dispersion curves from multi-component ambient noise data. Here, we develop a joint inversion method for radial anisotropy with multimodal Rayleigh and Love wave dispersion curves, and validate its effectiveness with synthetic and realistic examples. Specifically, we modified a Markov chain Monte Carlo (McMC) Bayesian inversion tool, BayHunter, to conduct the joint inversion of multimodal Rayleigh and Love wave dispersion curves. During the inversion, the initial model is introduced and the misfit value from single objective function in each iteration process is utilized to adaptively weight the joint likelihood. We proved that the incorporation of higher-modes dispersion curves can effectively reduce the non-uniqueness of inversion with synthetic tests. Currently, we are applying this joint inversion method to realistic data from eastern South China Block and validate its effectiveness in realistic applications.