Modeling method of GNSS continuous station coordinate time series considering random characteristics

GNSS technology provides important observations for the study of crustal physical phenomena such as plate motion, co-seismic and post-seismic deformation, and slow slip without earthquake, among which the key step is to accurately extract the tectonic and non-tectonic movement information of time series fully considering the random characteristics of time series. In this paper, we proposed a modeling method of GNSS continuous station coordinate time series based on Bayesian framework to accurately extract the optimal solution and error of model parameters. The optimal solution and error of the model parameter were obtained from the parameters were obtained from the parameter calculation method based on Bayesian framework.yWhile, the influence of the random characteristics of the model parameters and observation on the model solution results is systematically analyzed. Then, we applied the method to the northeastern corner of the Qinghai-Tibet Plateau, the extraction of the constructive and non-constructive deformations of the GNSS data was effective. Compared with the traditional method of GNSS continuous station coordinate time series modeling based on least squares fitting, this paper introduces GWMCMC algorithm, a parameter optimal solution method based on Bayesian framework, which takes into account the random characteristics to precisely and efficiently extract the tectonic movement trend and non-tectonic movement, and can more accurately obtain the optimal solution of the parameter and their error. Meanwhile, compared with the traditional MCMC algorithm, the GWMCMC algorithm improves its performance and computational efficiency through parallel computing. The research results of this paper provide data support for the subsequent use of tectonic deformation to carry out the kinematic characteristics and dynamic mechanism of crustal deformation.