多种面波层析成像方法及其在青藏高原的应用与对比

Multiple surface wave tomography methods and their applications to the Tibetan Plateau

面波层析成像是一种广泛应用的获取地壳与上地幔地震波速度与各向异性结构的地球物理方法.本论文综述了面波层析成像方法的简要历史,阐述了多种常用的面波层析成像方法(双台法、双平面波法、程函方程法、Helmholtz方程法、背景噪声成像法与直接面波层析成像法)的基本原理及其优缺点.双台法的理论与实际使用简单,但该方法要求震源与两个台站需近似位于同一大圆弧路径,在台站较少且记录时间较短的区域,获取的相速度水平分辨率偏低.双平面波法能一定程度克服地震波多路径传播与散射对频散的影响,但其对面波波形数据要求较为严格,且通常适用于区域地震台网.程函方程法和Helmholtz方程法可直接从地震记录同时获取相速度与方位各向异性,计算方便快速,无需经过正演与反演过程,但这两种方法要求台站分布密度要高,不适用于台站间距大且分布稀疏的区域.与程函方程法相比,Helmholtz方程法不仅考虑了波形的相位,还利用了其振幅,能获取更准确的相速度与方位各向异性信息.背景噪声成像法的优点是无需震源就可获取高分辨率地壳尺度的成像结果,但通常缺乏长周期面波的信息,难以约束岩石圈深部与软流圈结构.直接面波层析成像法能直接从台站间的面波频散数据通过反演获取三维剪切波(S波)速度结构与方位各向异性信息,省去了反演相速度图的中间步骤.此外,我们对比了多种方法在青藏高原获取的相速度结果.结果表明,多种面波层析成像方法获取的同一周期相速度结果高度吻合,主要特征表现在:在中长周期的相速度图中,青藏高原内部主要以低速为主,而周缘区域(如,柴达木盆地、四川盆地等)以相对高速为主,这表明青藏高原中下地壳与上地幔的流变强度均比其周缘区域要弱,青藏高原的岩石圈变形受控于周缘块体的阻挡.在青藏高原东南缘,短中周期(20~40 s)的相速度图像指示受强度较大的川滇地块阻挡,青藏高原中下地壳物质以地壳流的方式沿地壳薄弱带(即红河断裂带与鲜水河断裂带)向南挤出逃逸.此外,祁连山在短中周期(20~40 s)的相速度图中都表现为低速特征,可能与局部地幔物质上涌造成地壳的高温度异常有关.地震面波层析成像方法(双台法、双平面波法、程函方程法、Helmholtz方程法)与背景噪声层析成像法结合可获取短长周期范围(如,4~200 s)的瑞利波与勒夫波相速度,用于同时构建壳幔速度与各向异性结构.本文还提出开展地震高阶面波、伴随成像与联合反演等方法综合研究可望获取精度更高与更为可靠的壳幔结构.

Surface wave tomography is a widely used geophysical method to measuring seismic velocity and anisotropy in the crust and upper mantle.This paper briefly reviewed the history of surface wave tomography,and summarized the principles and advantages of multiple surface wave tomography methods(i.e.,two-station method,two-plane-wave method,Eikonal and Helmholtz surface wave tomography,ambient noise tomography,and direct surface wave tomography).The theory and application of the two-station method are straightforward,but this method requires the earthquake and seismic stations to be located in the same great circle.This restriction results in a low lateral resolution phase velocity map for regions where stations are sparse and deployment times short.The twoplane-wave method can overcome the effect of scattering and multipathing of seismic wave propagation on phase velocity dispersions,but this method requires high-quality surface wave waveforms and is usually suitable for regional seismic networks.The Eikonal and Helmholtz surface wave tomography can directly compute the phase velocities and azimuthal anisotropy straightforwardly without any processes of forward modeling and inversion,but this method is limitedly suitable for high-density and orderly seismic arrays.In comparison to the Eikonal surface wave tomography method,the Helmholtz surface wave tomography approach analyzes both the phase and amplitude of the waveform and can produce better results.Ambient noise tomography can obtain high-resolution crustal structure without seismic events.However,this method is difficult to obtain long-period phase velocities,resulting in a lack of constraints on the mantle and lithosphere structure.Direct surface wave tomography can obtain the shear-wave velocity and azimuthal anisotropy from dispersive curves without the inversion process of phase velocity maps.We conducted detailed comparisons of the phase velocity maps at the short-intermediate periods(20~40 s)previously obtained from these surface wave tomography methods in the central-northern,northeastern and southern Tibetan Plateau.The comparisons demonstrated that the Rayleigh and Love wave phase velocity maps from different tomography methods show highly consistency in velocity variation patterns at same periods.The results show that the plateau’s interior by characterized with low-velocities in phase velocity maps at intermediate and long periods,whereas relatively high velocities dominate adjacent regions,such as the Qaidam Basin and Sichuan Basin.These features indicate that the Tibetan Plateau’s mid-lower crust and upper mantle tend to be weaker and easily deform during the northward movement of the Indian plateau and the barrier of the strong blocks(e.g.,the Qaidam Basin and the Alashan Platform).The short periods phase velocity maps(i.e.,20~40 s)reveal that the midlower crustal flow of the Tibetan Plateau escapes southward surrounding the weakening zones(i.e.,the Red River Fault and the Xianshui River Fault)due to the barrier of the strong Chuandian tectonic block.The Qilian Orogen also shows low velocities in the phase velocity maps at the short-intermediate periods,which is likely due to thermal anomaly with localized mantle upwelling.It is noteworthy that the Rayleigh-wave and Love-wave phase velocities at different period ranges(e.g.,4~200 s)can be obtained via incorporating earthquake surface wave tomography and ambient noise tomography.Those phase velocities can be used as input to inverting three-dimensional shear wave velocities and radial anisotropy models simultaneously.This paper envisions that integrating higher-mode earthquake surface wave tomography,adjoint tomography,and joint inversion could obtain higher resolution and more reliable crustal and upper mantle structures.