Bayesian Modeling of Crustal Azimuthal Anisotropy Supports Partial Crust–Mantle Anisotropic Decoupling in a Continental Back‐Arc, Northeast China

How crustal deformation responds to complex mantle flow in continental back‐arc settings remains poorly constrained. Here, we develop a Bayesian working‐likelihood approach to estimate crust‐averaged azimuthal anisotropy from P‐wave receiver functions. The method separates crust‐averaged velocity anisotropy from dip‐related apparent anisotropy without requiring accurate moveout, while quantifying uncertainties under the adopted working‐likelihood framework. We apply it to Northeast China, a continental back‐arc shaped by long‐lived Pacific‐slab rollback. The inferred fast‐axis directions of crustal velocity anisotropy contain a regionally coherent E–W component, consistent with the rollback‐related extensional field. In contrast, published XKS splitting results indicate heterogeneous mantle anisotropy with locally variable fast‐axis directions. The contrast between the coherent crustal pattern and the variable mantle anisotropy supports partial crust–mantle anisotropic decoupling in Northeast China. We suggest that the crust records a long‐term, regionally integrated deformation signal, whereas mantle anisotropy likely reflects stronger contributions from localized mantle flow and possibly inherited lithospheric fabrics.