Joint analysis of seasonal oscillations derived from GPS observations and hydrological loading for mainland China

Abstract Hydrological loading effects are one of the principal sources of the seasonal oscillations in GPS position time series, and they should be taken into account for improving GPS system accuracy. In this study, the daily vertical position time series of 23 GPS stations derived from the Crustal Movement Observation Network of China (CMONOC) are used to investigate the hydrological loading contributors of seasonal oscillations in GPS observations. The hydrological loading deformations at each GPS station are estimated by the Global Land Surface Discharge Model (LSDM). The result of period analysis suggests that the hydrological loading primarily results in the annual oscillation in GPS observations. Therefore, Multichannel Singular Spectrum Analysis (MSSA) is utilized to derive the annual signal from GPS observations and LSDM-derived deformations simultaneously for each GPS station. Compared with Singular Spectrum Analysis (SSA), the percentage of the variance of the annual signal estimated by MSSA and SSA differs by 6% on average, indicating that MSSA can effectively separate annual signal from other signals and clear noise. Meanwhile, Cross Wavelet Transform (XWT) is performed to measure the correlation and phase relationship between GPS observations and LSDM-derived deformations in the time-frequency space. The result demonstrates that the hydrological loading can only explain the annual oscillation in GPS observations at 5 stations (namely LHAS, LUZH, KMIN, QION, and XIAG) well. For the most GPS stations, other factors (e.g., other geophysical factors, hydrological modeling errors, and systematic errors) and hydrological loading jointly contribute to the annual oscillation in GPS observations. After hydrological loading correction, the Root Mean Square (RMS) values of GPS observations are reduced at 15 stations, especially for the GPS stations located in regions with significant water storage variations (up to 2.46 mm at KMIN). Moreover, the percentage of the variance of the annual signal estimated by MSSA correlates well with the RMS reduction, implying that the RMS reduction may be related to the annual signal derived from hydrological loading deformations.