Abstract: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.
WANG Zhaoyang,XING Zhe,ZHANG Feng et al. Noise Analysis of the GPS Coordinate Time Series in the West Antarctic with Consideration of Common Mode Error[J]. jgg, 2022, 42(2): 132-138.