Theory And Methodology For China Geocenter Coordinate Reference Frame Establishment And Maintainence Using GPS

As the realization of Terrestrial Reference System (TRS), Terrestrial Reference Frame (TRF) is the basis of the establishment of the globle geocentric coordinate system. The dynamic process of TRF can reflect important information on globle changes such as plate motion, change of the sea level, change of underground water and so on. Research on establishment of TRF has significant meanings on the development of the economy and the society no matter in the field of science or reality.This article summarizes the definition and basis theories of coordinate system, analyses the main factors that have impact on the realization of coordinate system, researches the feature of plate motion of Chinese mainland, acquires the horizontal and vertical velocity fields of the national GNSS Continuously Operating Reference Stations(known as CORS), and does some analysis on time series of GNSS station coordinates. The main achievement and contribution of this paper is:The article introduces the establishment methods of International Terrestrial Reference System (ITRS), and the development of International Terrestrial Reference Frame(ITRF), mainly discusses the realization and development actuality of regional reference frame EUREF and SNARF. On basis of this, it analyses the main impact factors of coordinate fame realzation.The article give brief of4kinds of plate motion models, and acquires the velocity fileds in Chinese mainland by using Euler Model, GPS actual measurement and its fitting model. The realted research was carried out, for example NUVEL-1A model in China, comparison the consistency of similarity transformation and Euler vector model methods. It turns out that the two methods can reflect the motion feature of Chinese mainland very well.In this paper, brief introdution of various GPS networks, which were established in1980s in China individually which have good relationship with ITRF. Currently, there are continuous and abounant observation data in last11years. This paper give more detail about data resources, selection principle of base stations, strategies of data processing, adjustment method and statistics. The final results invole34national base stations and1630GNSS geodetic control points.Based on the resut of time series of national GNSS base stations, this paper discusses the stability, instantaneity and dynamism of both frame in long and short periods, and indicate accurate quantified indicators of TRF in China. Through analysis of time series of national GNSS base stantion’s weekly solutions, it obtains linear and nonlinear motion feature of the fiducial stations. The result shows that time series of horizontal vector have linear changes generally, and the quantity and direction of these changes is related to plate movement and relate movement between different plates; vertical motion shows result of a sinusoidal function with annual period. The power of periodic motion in vertical direction is bigger than it in horizontal direction and it is one of the main factors that impact on vertical accuracy.This paper introduce Principle Componant Analysis (PCA) method to eliminate the common error in the national CORS network, which effectively reduce the impact of the large scale factors on time series and improve the signal ratio. Result shows that after PCA filtering, the average mean square error in the N, E, and U reduced by26%,22%and26%respectively compared to the value before the filtering.The paper give the result of different noise character on10years time series before and after the filtering using the method of power spectrum and maximum likelihood estimation. Finally, it shows that noise characteristic are different in the orientation of the horizon and the height of the national CORS network, and the white noise, flicker noise and random walk noise are the basic characteristics in the coordinate time series, and the colored noise in the vertical component is bigger than the horizontal component obviously. In this paper, the analysis of colored noise source shows that higher-order ionosphere errors are the main source. Introducing different noise model to velocity and its precision estimation, the comparison result between colored noise and white noise shows that elocity evaluation bias is generally around5%～10%, the error is amplified2～4times. So the conclusion is that variable white noise plus color noise stretagy are more suitable with actual motion characteristics of national base stations.