An iterative weighted least square fitting method for crustal anisotropy using receiver functions

The harmonic variation of the P-to-S converted phases (i.e. Pms) observed from receiver functions (RFs) includes information on crustal azimuthal anisotropy. However, this harmonic analysis is easily influenced by low-quality RF traces, and the measurements may be misleading. Here, we propose an improved method, named the Iterative Weighted Least-Square method (IWLS), to extract the splitting parameters of the crust and simultaneously retrieve the two-lobed and four-lobed components of back-azimuthal variation. The quality and weights of different RF traces are estimated properly in the IWLS method. The weight function is related to the sharpness of the Pms phase and the smearing of other signals. We conduct many synthetic tests, and the IWLS method provides stable measurements for poor back-azimuthal coverage, strong noise, weak P-wave azimuthal anisotropy, and multiple anisotropic layers. We apply the IWLS method to observational data from two temporary stations on the southeastern Tibetan Plateau and North China Craton, respectively. The measurements are comparable to previous results and provide insight into crustal deformation.