Stepwise joint inversion for 1-D crustal Vp/Vs structure using surface wave dispersion, Rayleigh wave ZH ratio, and receiver function data

Abstract:

The joint inversion method, which uses a variety of data with different sensitivities to structures, has greatly improved the resolution of the crust velocity structure. We develop a stepwise joint inversion method using three datasets, i.e., receiver functions, surface wave dispersion, and the Rayleigh wave ZH amplitude ratios, to obtain the 1-D depth-dependent Vs and V_p/V_s model. Rayleigh wave ZH ratio and surface wave dispersion data are more sensitive to absolute shear wave velocity variations at depths, but their sensitivity kernels are different and complimentary. And receiver function data are more sensitive to sharp velocity contrast. The three datasets also have some sensitivities to V_p structures in the crust. Taking advantages of the complementary sensitivities of each dataset, our joint inversion method can better constrain the V_p/V_s model in the crust.

In this method, we use a linearized iterative algorithm to constrain the fine 1-D crustal V_p/V_s structure. In order to make full use of the different sensitivities of these three datasets to the crustal structure, we propose a three-step strategy to perform inversion, which gradually restores the fine crustal structure by adjusting the weights of different datasets. First, the surface wave dispersion and the ZH amplitude ratios are jointly used to constrain a smooth absolute shear wave velocity model, and then the dataset of receiver functions is added to further constrain a more refined shear wave velocity model with much better interface information. For the first two steps of inversion, the P wave velocity and density are calculated from the shear wave velocity model using an empirical formula. Finally, based on the obtained shear wave velocity model, these three datasets are jointly used to further invert for the Vs and V_p/V_s model.

To testify the proposed inversion method, we have performed a series of synthetic and real data tests, as well as Monte Carlo error tests to access the uncertainty of the inversion method. The results show that the stepwise joint inversion method can not only obtain a stable and reliable shear wave velocity model, but also has good constraints on the crustal V_p/V_s structure, which is important for understanding the lithology and physical state of the crust at depths.