Analysis Of Coordinate Time Series Of Cmonoc GPS Fiducial Stations Using Principal Component Analysis

With the continuous development of GNSS technology,high-precision GNSS technology has become a significant measurement of monitoring the geophysical and geodynamic phenomena.The principal component analysis is performed on the coordinate time series of the GPS fiducial stations of Crustal Movement Observation Network of China(CMONOC),which is used to study the temporal and spatial distribution characteristics of coordinate time series errors.The study can understand the time series noise characteristics of CMONOC and the crustal deformation characteristics in mainland China.The characteristics of the original coordinate time series of the GPS fiducial stations of CMONOC indicate that the horizontal direction mainly exhibits linear motion caused by plate motion,while the vertical direction mainly exhibits periodic characteristics and the fluctuation amplitude is greater than the horizontal direction.Discarding the sites fail to meet the criteria of sufficient data availability as 70%,which restricts our analysis to 224 sites.An algorithm based on least square estimation is employed to remove the constant offset,trend from each time series.The outliers are also estimated and removed based on the median and interquartile range statistics.The Cubic spline interpolation method is used to fill the time gaps shorter than 2 days.For larger gaps,firstly set the values at these missing days with spatially averaged values derived from all effective sites on these days,then constructing the covariance and correlation matrices to perform principal component analysis(PCA)iteratively,until convergence to 10-6 to refill in these missing data from the previous iteration.PCA is performed on the continuous coordinate time series matrix to calculate principal components(PCs)and corresponding spatial eigenvectors.According to the PCs and the corresponding spatial eigenvectors of north,east and vertical components,the common mode error(CME),regional distribution characteristics of sites spatial response,and abnormal site impact on PCA result are analyzed.The results indicate that a single PC is no longer able to reflect the whole spatial and temporal patterns of the CME in mainland China,the first three PCs are required to be considered to analyze the CME.In addition,there are relatively uniform spatial response in the northwest region,north China and Yunnan Province,which imply water reserves variations are largely the cause of the response characteristics.After removing the site with abnormal spatial response,the first and second principal components carrying more coordinate information in each direction are greatly affected,the vertical direction is the most obvious,and the sites space response is obviously improved.