Noise correlation in the time series of GPS stations in the North China Plain

A total of 20 stations in the North China Plain were selected. The noise variances of each station were estimated by the least squares variance component estimation method under the combination of white noise and flicker noise. The noise variances are composed of the white noise and flicker noise vectors of different components. The correlations among the different component noise vectors were then analyzed by univariate linear regression, and the regression equations were established. Results show that a moderate correlation exists among the white noise vectors in each station direction, especially a high correlation degree exists among the horizontal components. The flicker noise vector only has a moderate correlation among the horizontal directions. A change in one noise direction for the two correlated noise vectors can explain the more than 60% noise change in the other direction.