Abstract

Abstract: This paper describes analyses concerning annual signals in GPS-derived coordinates. The data was processed in the Military University of Technology Local Analysis Centre with Bernese 5.0 software. We used observations from 129 permanent GPS stations which belong to the Polish Active Geodetic Network (ASG-EUPOS), for the period of GPS weeks 1465-1729, corresponding to about 5 years. The annual signals have been estimated using the Least Squares Estimation (LSE) with assumption of stationarity of amplitude and phase, and uncertainties calculated with the coloured noise assumption using the First Order Gauss Markov model (FOGM). The phase was defined as the time span from the beginning of the year to the maximum of the best-fitting sinusoid. The data was analysed using the TSView software (Herring, 2003). Amplitudes of best-fitting sinusoids into the North components range between 0.1 and 3.5 mm, the East component: 0.1-4.1 mm, the Up component: 0.1-4.0 mm with similar uncertainties of 0.1-0.3 mm. The phase shifts are unevenly distributed over the considered area, no clearly visible spatial dependencies were discovered. The North and East components reveal clearly maximums concentrated over winter- (January and February) and summer-time (June-July). They could be explained by the hydrological and atmospheric (including thermal) influences. However, this theory does not seem to be supported by the Up distribution with no clearly visible extreme values. We noticed that the velocity bias due to annual oscillation could range from –0.6 to +0.5 mm/y, but when we look at the issue relatively we can mismodel the North and East velocity by 3.4 % (0.5 of 14.8 mm/y), while the Up velocity at the tectonically stable areas considered in this publication could be mismodelled by even 600 % (0.6 of 0.1 mm/y). But the annual signal in the GPS time series has to be understood twofold:

Citations

Authors (2)