An effective toolkit for the interpolation and gross error detection of GPS time series

As data gaps and gross errors inevitably exist in GPS time series, robust detection and interpolation procedures are needed to obtain a uniform time series for various geospatial studies and applications. The use of traditional methods for this purpose is usually based on some improper assumptions, which are not derived based on the real properties of the data. Moreover, different interpolation methods may need to be investigated for interpolation for various gaps with different types and amount of missing data. These make the interpolation for missing data not easy. To address the issue of mentioned above, in this study, a data-analysis method named singular spectrum analysis (SSA) for missing data is assessed for reconstructing a reliable model from unevenly sampled time series without the need for any a priori knowledge of the time series data. In this method, the interpolation and detection of gross errors are carried out in one go along with the reliable reconstructed model. Both simulation data and real GPS data testing results showed that this was an efficient method for interpolation and gross error detection.