区域GPS网共模误差的提取与成因分析

Extraction and Causal Analysis of Common Mode Error in Regional GPS Network

以位于美国南加州区域的41个GPS站点为例,利用主成分分析法对该区域共模误差进行了有效提取,结果表明共模误差在该区域空间响应具有较好的一致性。滤波后,均方根值(RMS)在E方向上平均降低了42.9%,在N方向上平均降低了39.1%,在U方向上平均降低了28.8%,表明空间滤波能够有效提高GPS坐标精度。此外,滤波后各站点相对于LOWS站点的相关系数在E方向上平均下降了89.3%,在N方向上平均下降了92.3%,在U方向上平均下降了86.5%,结果证实了共模误差是区域GPS网中各站点之间空间相关的主要来源。最后,皮尔逊相关性分析和小波相干分析证实了质量负载并不能有效地解释共模误差,共模误差最有可能来源于空间传播延迟误差。

Taking the 41 GPS sites located in Southern California as an example,principal component analysis is used to effectively extract the common mode error in this region.The results show that the common mode error has good consistency in spatial response in this region.After filtering,the RMS value decreases by an average of 42.9%in the E direction,39.1%in the N direction,and 28.8%in the U direction,indicating that spatial filtering can effectively improve the accuracy of the GPS coordinates.In addition,after filtering the correlation coefficient of each station relative to the LOWS station decreased by 89.3%in the E direction,92.3%in the N direction,and 86.5%in the U direction.The results confirm that the common mode error in regional GPS network is the main source of spatial correlation between different sites.Finally,Pearson correlation analysis and wavelet coherence analysis prove that the mass load can not effectively explain the common mode error,and the common mode error is most likely to result from space propagation delay error.