顾及共模误差的西南极GPS坐标时间序列噪声分析

Noise Analysis of the GPS Coordinate Time Series in the West Antarctic with Consideration of Common Mode Error

使用Bernese 5.2软件双差定位程序处理西南极区域24个GPS基准站9 a的观测资料,获得各站单日解坐标时间序列。利用主分量分析法和独立分析法对其进行空间滤波,对比分析2种方法的滤波效果,然后采用极大似然估计法分析滤波前后各站时间序列的噪声特征,建立最优噪声模型,并分析共模误差对测站速度的影响。结果表明,2种方法均能有效提取西南极区域共模误差,主分量分析法的滤波效果略优于独立分析法;西南极区域GPS站的噪声模型具有多样性,各分量的噪声特征存在差异,主要以WN+FN、WN+ARMA(1)为主;去除共模误差后,部分测站最优噪声模型发生改变,最优噪声区域的分布特征不明显;空间滤波对测站速度场的影响小于0.1 mm/a,但可降低白噪声和有色噪声的量级,从而大幅度减小速度不确定度。

We use the Bernese GNSS software version 5.2 to analyze the data of 9 years of 24 GPS reference stations in the west Antarctic region,calculating the single-day solution under the mode of double-difference processing to acquire the coordinate time series.We employ principal component analysis(PCA)and independent component analysis(ICA)to put spatial filtering on the time series and compare and analyze the filtering effects of the two methods.We use the maximum likelihood estimation method to analyze the noise characteristics before and after filtering and establish the optimal noise model for each station.The results show that both methods can effectively extract the common mode errors in the west Antarctic region,but the results of PCA are slightly better than those of ICA.The noise models of GPS stations in the west Antarctic region are diverse,and mainly for WN+FN and WN+ARMA(1).After deducing the common mode error,the optimal noise model of some stations changes,and the regional distribution characteristics are not obvious.The influence of spatial filtering on the velocity field of station is less than 0.1 mm/a,but it can reduce the order of white noise and colored noise,thus greatly reducing the uncertainty of velocity.