Research On Tropospheric Delay Model In China

Tropospheric delay is one of the main impactson satellite navigation and positioning especially in elevation direction. Currently, tropospheric delay correction methods mainly include external correction, parameter estimation and model correction. Correction modelsare function formulas based on various assumptions and factors that reflect tropospheric delay. Since the model correction method is simple, wide application, it is the main method to calculate tropospheric delays. Correction models contain zenith tropospheric delay models and mapping function models.This paper studies the zenith tropospheric delay model in China,specifically in the following research areas:(1) Detailed analysis to the spatial and temporal variation of the tropospheric delay in China. Sum up the characteristics of temporal and spatial variation of tropospheric delays in China.(2) Research on tropospheric delay model that related to meteorological parameters in 4 IGS stations in China. BP neural network is usedto compensatethe Saastamoinen model, and we propose the ISAAS model which is based on meteorological parameters. ISAAS model has no systematic bias in 4 IGS stations of China, the average root-mean-square error of ISAAS model is ± 2.5cm,the average root-mean-square error of Saastamoinen model is ± 4.9cm.In every IGS station,model accuracy of ISAAS model increased by more than 33%,compared with. In lhaz station which is in high area, the average root-mean-square error of Saastamoinen model is ±4.8cm,the average root-mean-square error of ISAAS model is ± 1.0cm, Model accuracy of ISAAS is not impacted by hight.(3) Research on tropospheric delay model that unrelated to meteorological parameters at midand high latitudes in China. By analysis of seasonal difference of tropospheric delay at different latitudes,we found that in high latitudes regions,tropospheric delay variations violent in summer and changes slowly in winter. On this basis, we propose a model based on quadratic function. In the high latitudesregions, the average root-mean-square error of parabolic model is ±4.4cm,the average root-mean-square error of cosine function model is ± 5.3cm,model accuracy of parabolic model can improve about 0.9cm in summer, compared with cosine function model.Parabolic model can be more accurate in reflecting tropospheric delay variation in summer.(4) Research on tropospheric delay forecasting modelin China.Tropospheric delay forecasting model (Bilinear model)was established with the study of the spatial and temporal variation of the tropospheric delay in China and parabola model. With the longitude, latitude, elevation, and day of the year at the station, you can get tropospheric delay forecast values at the station directly. Bilinear model does not have systematic bias in China, and the accuracy of the model (the root-mean-square error of parabolic model is +3.9cm) increases by about 1.5cm compared to EGNOS model (the root-mean-square error of parabolic model is ± 5.4cm).