Time-frequency Analysis On Coordinate Frame Points Of Shandong CORS

Shandong Continuously Operating Reference Station System(SDCORS)build a technology and service platform in different application areas,and it's one of the infrastructure of “digital Shandong” and also the spatial data infrastructure project of Shandong province.This paper focuses on the analysis and research of SDCORS frame points coordinate time series in order to establish the suitable model of coordinate time series,correct the error models,improve the accuracy of SDCORS and provide a certain constraint for Shandong regional coordinate reference frame.Due to various reasons,there are singular values,varying sampling and the offset in the coordinate time series.Therefore,before the analysis of the coordinate time series of frame points,pretreatment should be performed so that provide true and correct data for the following research.Base stations are affected by a variety of geophysical phenomena,the coordinate time series show certain periodicity.Study on the law of the periodic change can provide certain constraints for various physical phenomena and the seasonal variation of vertical crustal movement so that improve the measurement precision of the frame points of SDCORS.This paper presents the wavelet spectral are used to detect periodic terms in order to gets the frequency characteristics of coordinate time series.By the wavelet analysis the periodic terms are extracted.In time domain,we get the amplitude and phase characteristics of each periodic terms.The results show that there are annual period,half year period,seasonal period and monthly periodic signal in the coordinate time series.As the existence of common mode error in the regional GPS network coordinate time series,the deformation characteristics of the station itself might be covered.In this paper,A simulation example verifies the feasibility of spatial filtering based on principal component analysis and then the coordinate time series of 15 frame points are analyzed.By principal component analysis the common mode errors of the regional GPS network was extracted in three coordinate components,and then the spatial filtering of the regional network was carried out.After spatial filtering the correlation coefficients relative to ANQI were decreased by average of 74%,85% and 75% in E,U and N,respectively.By wavelet spectral analysis,the frequency characteristic of the common mode error series has been got and we also analyzed the generation mechanism of the common mode errors.The results show that there are annual period,half year period and monthly periodic signal in the common mode error series,and each period has different strength and emergence time.The errors caused by the ionosphere,the troposphere,and the tide are the main mechanism of common mode errors.Take the coordinate time series of 26 frame points in Shandong as an example,By using spectral index method and maximum likelihood estimation,the type of noise and the components of noise are determined and then the noise combination model was built.The results shows that the best noise combination model of Shandong frame points is white noise plus flicker noise,the mean value of white noise is 1.25 mm and the flicker noise is 5.46 mm.The establishment of noise combination model of frame points have important reference value for improve it's accuracy.The multi-scale correlation analysis of frame points on both sides of the yishu fault zone in Shandong province was carried out to figure out correlation coefficient in terms of different periods.Thus,provide a new approach to study the relationship between two columns of geodetic signals and tectonic movement of the fault zone.