Extract Common-Mode Error in Middle-Scale GPS Network Using Principal Component Analysis

摘要

Based on the correlation coefficient between the time series of the GPS network, extracting the common model error (CME) using principal component analysis (PCA) in Crustal Movement Observation Network of China (CMONOC). This method does not need the assumption of spatially uniform distribution like stacking. Compared with the traditional regional stacking (RS), correlation weighted regional stacking (CWRS), and traditional PCA, this method extracts CME more accurately and reduces the influence of local effects during extracting the CME.