连续GPS站坐标序列共性误差的提取与形变分析

EXTRACTION OF COMMON MODE ERRORS FOR CONTINUOUS GPS NETWORKS AND DEFORMATION ANALYSIS

如何去除连续GPS站点坐标时间序列中的共性误差,提高站点坐标精度,对形变分析具有重要作用。使用主成分分析(PCA)和Karhunen-Loeve展开(KLE)方法可以很好地从连续GPS网观测站的坐标时间序列中分离出共性误差。用该方法处理了南加州6个跨圣安德烈斯断层的连续GPS网观测站近5年的坐标时间序列,结果表明这些站坐标序列中存在显著的共性误差。去除共性误差后,站坐标序列的均方差有了显著提高,并且形变分析结果的精度得到改进。

How to reduce the common mode errors of the time coordinate series is important for the extraction of crustal deformation. The use of principal component analysis and Karhunen-Loeve expansion approaches can reduce the common mode errors of the time coordinate series. We apply the combination of PCA and KLE to analyze the daily time coordinate series of the six stations of Southern California Integrated GPS Network for about five years, and we demonstrate that those stations all contain the common mode errors. The root mean squares of the time coordinate series are significantly reduced by reducing the common mode errors, thereby improving the accuracy of deformation analysis.