Identification of stochastic process in GNSS time series

In this study, determination of noise types in GNSS stations time series for the purpose of daily coordinates time series of N(North), E(East), and U(Up) of TUSAGA-Active (CORS-TR) stations as AKHR, AKSI, AKSR, ANRK, BEYS, CIHA, HYMN, KAMN, KAPN, KIRS, KLUU, NEVS ve NIGD were used. For analysis, the outlier measurements and data gaps in the GNSS time series were eliminated and the series were ready state for modeling. Noise data were obtained after removing the trend (velocity values) and periodic component effects determined by a linear and trigonometric function applied to the series. In statistical terms, 1/f (colored) noise process was investigated by one of the recommended frequency domain methods for the parameter estimation, Periodogram Based Method and Wavelet Transformation Based Method. Accordance with to the noise component (color noise index) values determined by these investigations, the time series of GNSS stations were found to contain white noise and power law noise. Keywords: GNSS time series, Trend component, Periodic component, Periodogram, Wavelet transform, Colored noise analysis.