青藏高原的形成、演化和隆升机制一直都是国际地学研究的热点问题。地壳上地幔S波速度和各向异性结构可以为研究这些问题提供直接的地震学依据。青藏高原地区地震台站相对较少，且分布不均匀，采用传统层析成像技术中的均匀网格划分策略以及相同的正则化约束条件，可能会造成在射线密集地区结果过于平滑，丢失了一些模型中的细节信息，而在射线稀疏地区出现一些虚假速度异常。本项目将使用基于背景噪声和基于地震事件的面波层析成像，建立青藏高原地区高分辨率地壳上地幔S波速度和各向异性结构。反演中将使用近年来发展的可变维度贝叶斯层析成像理论。该方法根据射线分布对研究区进行自适应多尺度网格划分，可以获取射线密集分布区更多的细节信息，同时避免射线稀疏地区由于网格划分带来的虚假信息。在此基础上，对青藏高原地区的相关动力学问题进行讨论。

The formation, evolution and uplift mechanism of the Tibetan Plateau have always been hot issues in geoscience community. The S-wave velocity and anisotropic structure of the crust and upper-most mantle can provide direct seismological basis for these issues. Due to the uneven distribution of seismic stations in this region, the use of the uniform meshing strategy and the same regularization constraints in the traditional tomography techniques may cause the false anomalies in areas with sparse ray-density and loss of details in areas with intensive ray-density. In this project, we intend to obtain the high-resolution crustal and upper-most mantle S-wave velocity and anisotropic structure by combining ambient noise based and earthquake based surface wave tomography. We will adopt the transdimensional Bayesian inversion method, in which the model space can be parameterized adaptively according to the resolving ability. The velocity and anisotropy model will have more details in areas with intensive ray-density and avoids false anomalies in areas with sparse ray-density. The deformation mechanism of the Tibetan Plateau will also be inferred according the models.