采用剪切波分裂方向约束的板块绝对运动模型研究

板块绝对运动(即岩石圈板块相对于深部地幔的运动)导致软流圈深度存在较强的剪切作用,为软流圈呈现较强的地震波速各向异性提供了一种物理解释.相应地,软流圈地震各向异性的实测数据为反演板块绝对运动提供了一种定量的约束.本文利用前人发表的由474个剪切波分裂数据组成的全球软流圈地震各向异性方向数据集,结合板块相对运动模型MORVEL,通过加权最小二乘法反演板块绝对运动.计算结果表明,由实测数据约束的板块绝对运动模型不能将观测数据拟合到原始数据测量误差要求的统计水平上,反映出以地震各向异性方向指示板块绝对运动方向存在原始数据测量误差之外的系统性误差.该误差或可归因于板块绝对运动控制软流圈地震各向异性的物理机制复杂性以及小尺度地幔流动的区域复杂性.因此,地震各向异性方向只能从统计平均的意义上约束板块绝对运动方向,两者间的差别除了测量误差外还应包含复杂物理成因的模型误差.在原始数据测量误差的基础上增加20°的模型误差并且剔除一个离群数据后,本文得到了最优拟合剪切波分裂数据集的SKS473模型,并与由热点数据反演得到的板块绝对运动模型进行对比.由于约束板块绝对运动的地震各向异性数据和热点数据都存在误差较大、地理分布不均的局限性,结合两类数据的联合反演或可成为未来建立更高精度板块绝对运动模型的有效途径.

Study on absolute plate motion model constrainted by shear wave spliting orientation

The absolute plate motions, i.e., the motions of the lithospheric plates relative to the deep mantle, cause strong shear forces in the asthenosphere, which provides a physical explanation to the strong seismic anisotropy observed at the asthenospheric depths. Accordingly, the observed asthenospheric seismic anisotropy data provide quantitative constraints to invert the absolute plate motions. In this paper, a published global data set of asthenospheric seismic anisotropy including 474 shear wave splitting orientations is adopted together with the MORVEL model for relative plate motions to invert absolute plate motions based on a weighted least square method. The computational results indicate that the obtained absolute plate motion model constrained by the observed data cannot fit the data to a level statistically required by the original measurement errors of the data, implying that there exist systematical errors in addition to the measurement errors for seismic anisotropy orientations to represent the directions of the absolute plate motions. Such additional errors may be attributed to the complicated physical mechanism for the seismic anisotropy being controlled by the absolute plate motions and the complicated small-scale regional mantle flows. Therefore, the orientations of seismic anisotropy can only constrain the directions of absolute plate motions in the sense of statistical average, and the difference between the two should include not only measurement errors, but also model errors of complex physical causes. By adding a model error of 20 degrees to the measurement errors and eliminating one outlier datum, the SKS473 model that best fits the shear wave splitting data set is obtained and compared with the absolute plate motion models constrained by the hot spot data. Due to the limitations of large errors and uneven geographical distribution in the seismic anisotropy data and hot spot data for constraining the absolute plate motions, joint inversions of the two types of data may be an effective approach in the future to establish higher accuracy absolute plate motion models.