Azimuthal anisotropy retrieved from surface waves is important for constraining depth‐varying deformation patterns in the crust and upper mantle. We present a direct inversion technique for the three‐dimensional shear wave speed azimuthal anisotropy based on mixed‐path surface wave traveltime data. This new method includes two steps: (1) inversion for the 3‐D isotropic Vsv model directly from Rayleigh wave traveltimes and (2) joint inversion for both 3‐D Vsv azimuthal anisotropy and additional 3‐D isotropic Vsv perturbation. The joint inversion can significantly mitigate the trade‐off between strong heterogeneity and anisotropy. With frequency‐dependent ray tracing based on 2‐D isotropic phase speed maps, the new method takes into account the ray‐bending effect on surface wave propagation. We apply the new method to a regional array in Yunnan, southwestern China. Using Rayleigh wave phase velocity dispersion data in the period band of 5–40 s extracted from ambient noise interferometry, we obtain a 3‐D model of shear wave speed and azimuthal anisotropy in the crust and uppermost mantle in Yunnan. This model reveals that two midcrust low‐velocity zones are possible weak channels, and the azimuthal anisotropy at a depth of 5 to 30 km is mainly controlled by nearby strike‐slip faults, some of which also approximately coincide with the lateral boundaries of the crustal low‐velocity zones. Approximately south of 26°N, the upper crustal azimuthal anisotropy from our model is significantly different from the upper mantle anisotropy inferred by shear wave splitting, indicating different deformation styles between the crust and upper mantle in southern Yunnan.

The recently developed frequency‐Bessel transformation (F‐J) method is effective to extract multimode surface wave dispersion curves from ambient noise cross‐correlation functions (CCFs). However, this method is currently limited to the vertical‐vertical component CCFs, and only Rayleigh wave dispersion curves can be obtained. In this study, we first relate the F‐J spectrogram to the spatial autocorrelation coefficients; we then extend the F‐J method to the full multicomponent CCFs tensor, including the radial‐radial, transverse‐transverse, and the mixed‐component CCFs. Using the newly derived formulation, not only the signal of higher‐mode Rayleigh wave phase velocity dispersion can be enhanced, but also the multimode Love wave phase velocity dispersion curves and the higher‐mode Rayleigh wave ellipticity can be extracted. The formulation is tested in several numerical examples and is applied to field data in North America. Our derivation and formulation provide an extension to the current F‐J method and help to take usage of multicomponent CCFs. The resulting higher‐mode surface wave dispersions and Rayleigh wave ellipticity provide complementary constraint on the Earth structures.

A novel method is implemented to invert Rayleigh and Love wave dispersion curves of all paths jointly for 3‐D shear wave velocity and radial anisotropy simultaneously without intermediate steps. We use the method to derive high‐precision crustal shear wave velocity and radial anisotropy models around the eastern Himalayan syntaxis using ambient noise dispersion data (5–40 s). Results show that the crust can be divided into several subregions with different rigidity and preferred mineral alignment orientation depth‐dependently. In the middle crust, combined with other geophysical observations, 3‐D geometry of two continuous branches of eastward channelized weak zones is outlined in detail. Both branches of the weak zone are blocked by the high velocity zone with radial anisotropy of V_sh>V_sv at around 96–97°E. In the upper crust, the radial anisotropy model depicts a complex pattern, which is associated with the ongoing surface uplifting and shear strain rate distribution.

Fault zone (FZ) structure is critical for understanding earthquake rupture propagation and assessment of seismic hazard. In this study, we develop a new 2‐D dense‐array‐based inversion method to obtain high‐resolution FZ structure model using teleseismic receiver function. In our method, coherent travel times of converted waves and their multiples are first identified from teleseismic receiver functions; then the travel times are inverted for both depth extent and ratio beneath the seismic dense array. Synthetic tests with various shallow low‐velocity‐zone (LVZ) models demonstrate the effectiveness and robustness of our new method. We apply our method to the data recorded by a linear dense array across the southern segment of the Chenghai fault (CHF) in Yunnan, southwestern China. The inversion results suggest that the depth of the LVZ beneath the CHF extends to ∼1.1 km. Furthermore, we derive the ratio simultaneously and obtain the P‐wave velocity model from a S‐wave velocity model of the LVZ that were derived from ambient noise tomography. Travel times predicted by the newly obtained model are consistent with teleseismic P‐ and S‐wave arrivals. In addition, our method shows clear improvement in stability compared with the single‐station H‐ method. With dense seismic arrays deployed globally, our method would be a useful tool in shallow low‐velocity structure imaging studies, including fault zones and sedimentary basins.

The frequency–Bessel (F–J) transformation method is effective for the extraction of multimode surface wave dispersion data from ambient noise cross-correlation functions (CCFs). Recently, this method has been improved in terms of increasing resolution and reducing artifacts (or cycle skipping) in Rayleigh wave dispersion measurements. However, these advances are restricted to the ZZ-component F–J method, which is only applicable to Rayleigh waves. In contrast, they have not been extended to Love waves, which are fundamental for determining the horizontally polarized shear-wave velocity and the radial anisotropy associated with it. Furthermore, there is still a lack of a methodology for combining these advances. In this study, we propose a modified multicomponent F–J (MMFJ) method to extract high-quality dispersion data of both the Rayleigh and Love waves. To achieve high resolution, we derive formulas to optimize the MMFJ spectra. With synthetic ambient noise data and USArray data, we demonstrated the effectiveness of the MMFJ method in eliminating “crossed” artifacts and enhancing resolution. In particular, the “crossed” artifacts are greatly reduced using the MMFJ with small seismic arrays when the interstation distances in the seismic array are not dense enough. As such, the new MMFJ method has significant potential for handling seismic arrays with a limited number of receivers and the subsequent tomography of radial anisotropies at high precision.