共模误差是GPS台站位置时间序列中的主要误差源之一，分离和消除共模误差可以探测到原先淹没在噪音中的微弱信号。本文对中国260个卫星导航定位基准站2011～2018年的数据进行处理，获得了位置时间序列，采用主成分分析（PCA）方法对其扣除构造运动等形变信号后的残差序列进行区域共模误差分析。结果表明，PCA方法有效地提取了中国区域共模误差，经过滤波后时间序列的信噪比明显提高，第一主分量空间分布均一，第二主分量空间分布呈现明显的梯度变化，表明第二主分量不是少数台站的局部效应，是对第一主分量的补充和修正，中国区域的共模误差分析不能忽略二阶分量。对振幅调制的影响进行了对比分析，发现顾及振幅调制的共模误差不仅明显减弱了年周期的特征，还显著降低了闪烁噪声，表明由于参数过于简化而未被估计的“信号”会演变成共模误差，未被正确模型化的“信号”会影响区域滤波的准确性。

We use the 2011～2018 data of 260 continuous Global Positioning System(GPS)stations to study CME of China. The PCA method is used to conduct regional common-mode error analysis on the obtained residuals time series which is after deduction of deformation signals such as tectonic movements. The PCA method has proved to be an effective tool to estimate the CME at continental scale. Space-time filtering of position time series with the results of the PCA significantly improves the signal-to-noise ratio. The CME in the study region contained within the first two PCs,the spatial distribution of the first principal component is uniform,and the spatial distribution of the second principal component presents an obvious gradient change pattern,which suggests that the second principal component is not a local effect of a few stations,but is a supplement and correction to the first principal component. The second PCs should not be ignored for the CME analyses of our country. By comparing the common mode errors of the residual time series of the constant amplitude seasonal term and the modulated amplitude seasonal term,we found that CME which consider amplitude modulated significantly weakened the characteristics of the annual cycle,indicating that the un-estimated “signals” which are ignored when time series analyses will be evolved into CME,revealing that common mode errors,including signals and errors,are not simply errors.

P315

国家自然科学基金（41304008）资助