中国陆态网GPS基准站坐标时间序列主成分分析

发布时间：2018-03-23 11:25

本文选题：中国陆态网　切入点：GPS坐标时间序列　出处：《西南交通大学》2017年硕士论文

【摘要】：随着GNSS技术不断发展,高精度的GNSS技术已成为地球物理及地球动力学现象的重要监测手段。对中国大陆构造环境监测网络(简称"中国陆态网")GPS基准站坐标时间序列进行主成分分析,并以此研究中国大陆地区坐标时间序列误差时空分布特征,能够提高对中国陆态网坐标时间序列噪声特性和中国大陆地区地壳形变特征的认识。陆态网站点原始坐标时间序列特征表明基准站水平方向主要以板块运动引起的线性运动为主,垂直方向主要呈现周期性变化特征且波动幅度大于水平方向。剔除掉观测数据小于70%的部分站点,共保留了 224个站点坐标时间序列用于本次研究。基于最小二乘准则去除站点原始坐标序列中存在的趋势项、突变项,基于四分位数粗差探测法去除粗差,并且采用三次样条插值法补齐站点中小于三天的连续缺失坐标序列数据。大于等于三天的缺失坐标序列,先计算出该观测时段内有效站点的坐标序列的平均值来补齐缺失数据,然后对坐标时间序列进行迭代主成分分析,取10-6为迭代过程中缺失数据前后两次差的阈值来获取连续的坐标时间序列。针对预处理后连续的坐标分量时间序列,分别组建坐标时间序列矩阵进行主成分分析。根据N、E、U(南北、东西、垂直)三个方向主成分以及对应的空间特征向量分析了共模误差,站点响应区域分布特征以及异常站点对主成分分析结果的影响。分析结果表明仅通过第一主成分已不能体现公共模式的特征,本文可以将前三个主成分纳入共模误差分析。此外西北地区、华北地区以及云南地区各主成分的站点空间响应显示出了相对的一致性分布特征,水储量变化很大程度上是引发该响应特征的原因。在去除了区域空间响应异常的站点后,各方向携带较多坐标序列信息的第一、二主成分受到较大影响,垂直方向表现最为明显,且站点空间响应都明显提高。
[Abstract]:With the development of GNSS technology, High precision GNSS technology has become an important monitoring method for geophysical and geodynamic phenomena. The temporal and spatial distribution characteristics of coordinate time series errors in mainland China are also studied. It can improve the understanding of the noise characteristics of the coordinate time series of the land network in China and the crustal deformation characteristics in the mainland of China. The characteristics of the original coordinate time series of the land site point indicate that the horizontal direction of the reference station is mainly caused by the plate movement. The linear motion of the. The vertical direction is mainly characterized by periodic variation and the fluctuation is larger than that in the horizontal direction. Some stations with observation data less than 70% are excluded. A total of 224 site coordinate time series were retained for this study. Based on the least square criterion, the trend term in the original coordinate sequence of the station was removed, the mutation term was removed, and the gross error was removed based on the quartile gross error detection method. Using cubic spline interpolation method to correct the missing coordinate sequence data for less than three days, the missing coordinate sequence greater than or equal to three days, the average value of the coordinate sequence of the valid site in the observation period is calculated first to correct the missing data. Then the coordinate time series is analyzed by iterative principal component analysis, and 10-6 is taken as the threshold value of the two differences before and after the missing data in the iterative process to obtain the continuous coordinate time series. The coordinate time series matrix is set up for principal component analysis, and the common mode error is analyzed according to the three principal components in the three directions (north, south, east and west, vertical) and the corresponding spatial eigenvector. The characteristics of site response area distribution and the influence of abnormal sites on the results of principal component analysis (PCA). The results show that only the first principal component can no longer reflect the characteristics of the common pattern. In this paper, the first three principal components can be incorporated into the common mode error analysis. In addition, the spatial responses of the principal components in Northwest China, North China and Yunnan show relatively consistent distribution characteristics. The change of water reserves is the reason for this response feature to a great extent. After removing the stations with abnormal regional spatial response, the first and second principal components of each direction carrying more coordinate sequence information are greatly affected. Vertical performance is the most obvious, and site space response are significantly improved.
【学位授予单位】：西南交通大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：P228.4

【相似文献】