Study On The Statistical Correction Model And The Mechanisms For Nonlinear Variations Of Gps Stations' Coordinates

With the rapid development of all kinds of space geodetic technologies and the higher requirement of location accuracy for human production and life,the millimeterscale terrestrial reference frame is being gradually established and applied to geodesy.Because of the nonlinear motion is the reflection of real movement of the stations,weakening the nonlinear motion by establishing statistical correction models,and researching the nonlinear motion mechanisms of stations,are the important means to maintain and refine the millimeter-scale terrestrial reference frame.This thesis mainly focuses on establishing the statistical correction models and researching the nonlinear motion mechanisms of the global GPS stations' coordinates.The main works and conclusions of the thesis are as follows.1.From the perspective of time series interpolation and fitting,this thesis introduces the data preprocessing method which has good treatment effect at present.Secondly,it explains the most accurate and efficient method of detecting periodic terms--spectrum analysis.Finally,according to the thought of harmonic analysis,the method of extracting parameters of periodic terms based on least square fitting is derived.2.The spectral analysis is used to detect and count the periodic terms of the three-dimensional directions of 461 stations in the world.The results show that the annual terms common exist in the coordinates time series of three-dimensional directions of all stations.It is found that amplitudes of three-dimensional annual terms increase first and then decrease with the increase of latitude,the three-dimensional annual initial phases between the stations located in north and south hemisphere show the opposite rule with the increase of latitude,the differences between the initial phase of mid-latitude stations located in northern and southern hemisphere are about 180 degrees,and the functions of global distribution law of the annual amplitudes and the annual initial phases of the GPS station are fitted.3.For the first time,the three-dimensional statistical correction models for vertical annual variations of global GPS stations based on distribution laws are constructed,and the correction effect of models on the stations' coordinates residuals is evaluated by introducing the two parameters of model,correction ratio and residual ratio.The results show that the three-dimensional statistical correction model for annual variations can weaken 30%~50% coordinates residuals of most GPS stations in northern,eastern and vertical directions.4.The influence mechanisms of the seasonal temperature effect,the higher-order ionosphere delay and antenna phase center deviation which is included in the GPS systematic error on the nonlinear variations of the GPS stations are studied.The results are as follows.(1)Seasonal temperature change will lead to the threedimensional annual and semi-annual variations of the stations,which can explain 10%~20% three-dimensional annual variations.(2)The higher-order ionosphere delay will lead to the annual motion in north and semi-annual motion in zenith of the lowlatitude GPS stations,the annual motion in zenith and the semi-annual motion in north of the middle-latitude GPS stations,and the 1.4-year motion in east and the annual motion in zenith,the higher-order ionosphere can explain 30%~35% threedimensional annual variations,and the 23%~27% three-dimensional semi-annual variations.(3)antenna phase center deviation can explain 20%~30% GPS annual systematic errors in three-dimensional directions,which is probably one of reasons that causes northern,eastern and vertical annual motion and vertical semi-annual motion of global GPS stations,the northern semi-annual motion of middle-latitude and low-latitude GPS stations.