Analysis of GPS Height Time Series Based On IGS Fiducially Stations

The paper mainly focuses on study GPS height time series from IGS fiducially stations, which are globally distributed. Firstly, we calculated the autocorrelation functions to judge whether there are periodic terms. We found that there were trend terms and periodic terms in the time series of most stations. Then we improved the time series that day-to-day changed in coordinates by using ARMA. The noise has decreased a lot compared to primary time series, and we can clearly see the change of coordinates. Then we calculated the linear velocities of these stations by linear fitting. We use the power spectrum to analysis time series, and found there were clear annual signals in most stations. Then we analysis GPS height time series by Symlets wavelet filter for detecting what period variance in time series. After analysis, we concluded that not only annual, but semi-annual, biennial and even long term signals in GPS height time series.