基于ISE-LSC-RELSM的地壳形变分析

Crustal Deformation Analysis Based on ISE-LSC-RELSM

摘要: 将块体运动偏离整体旋转与线性应变(RELSM)的地壳形变看作“信号”，即利用“信号”对偏离RELSM的块体内部不规则形变进行补偿，建立以整体旋转与线性应变为“趋势性”的最小二乘配置模型(LSC-RELSM)，并尝试利用残差的信息扩散函数(ISE)构建方差–协方差矩阵代替以往利用高斯指数函数法确定信号的方差–协方差矩阵来解算配置模型(ISE-LSC-RELSM)。对喜马拉雅–西藏块体的实测GPS速度场数据进行拟合分析结果表明，新方法不仅优化了RELSM，并且还可以取得优于高斯指数法确定方差–协方差的拟合结果，这是因为信息扩散函数可以一定程度的抑制异常测站和粗差的影响。

Abstract: The crustal deformation of blocks deviating from global rotation and linear strain (RELSM) is regarded as “signal”, that is, the irregular deformation of blocks deviating from RELSM is compensated by “signal”, and the least square configuration model (LSC-RELSM) is established with global rotation and linear strain as “trend”. The variance-covariance matrix is constructed by using residual information diffusion function (ISE) to solve the configuration model (ISE-LSC-RELSM) instead of the variance-covariance matrix determined by Gaussian exponential function method. The results of fitting analysis on the measured GPS velocity field data of himalaya-Tibet block show that the new method not only optimizes RELSM, but also achieves better fitting results than the Gaussian exponential method to determine the variance-covariance, because the information diffusion function can suppress the influence of abnormal stations and gross errors to a certain extent.