Spatio-temporal Modeling And Application Of Troposphere Delay In Complex Terrain

When the GPS satellite signal is transmitted to the GPS receiver on the ground through the atmosphere outside the earth,the radio signal is bound to be affected by the neutral atmosphere.Tropospheric delay in satellite navigation and positioning generally refers to the signal delay caused by electromagnetic wave signals passing through a non-ionized neutral atmosphere below 50 km in height.In the study of signal delays,we no longer subdivide the atmosphere into the troposphere and stratosphere(as in atmospheric sciences),and we no longer take into account the difference in nature between the two.Since 80% of the delay occurs in the troposphere,I generally refer to signals that occur in this neutral atmosphere as tropospheric delay.In the actual measurement work,empirical models are mostly used to estimate tropospheric delay,but it will be affected by meteorological parameter errors and meteorological element measurement errors,making the prediction effect of tropospheric delay less than ideal,and ultimately limiting the measurement of GNSS positioning results precision.The difference method can also be used to eliminate most of the tropospheric delay errors at both ends of the baseline.However,many studies have shown that it is mainly applicable to small areas where the elevation changes are small.Large residual tropospheric delays can have a significant impact on positioning results.In order to improve the accuracy of the calculation of regional tropospheric delay,this paper uses continuous operation reference stations and related meteorological data in Sichuan to establish a regional high-precision tropospheric delay correction model to meet the needs of high-precision GNSS positioning.The main tasks as follows:1.The accuracy differences of three traditional tropospheric empirical models under different terrain conditions are compared and analyzed,and the following conclusions are drawn: The three empirical models have little difference in accuracy at sites with lower elevations,but as the site elevation increases,Hopfield model's BIAS and RMS are significantly worse than the other two models.The EGNOS model is obtained from simulated meteorological data.The overall curve is relatively smooth and the jitter is not severe.The overall accuracy is not much different from that of the Saastamonien model,but from the results of data processing,it can be seen that the fit between the results calculated by the EGNOS model in the autumn and winter seasons and the true values is obviously not as good as in other periods.Taking into consideration the selection of empirical models in practical applications to estimate the tropospheric zenith delay,the Saastamonien model should be selected.2.Based on the ECMWF reanalysis data,a grid model for the Sichuan region was established.Several CORS sites in different terrain areas were selected for comparison with the Saastamonien model.After comparison,it was found that the ZTD time series solved by the grid model was closer to the GAMIT solution,The average deviation BIAS is only6 mm,while the average deviation BIAS calculated by the Saastamonien model reaches5 cm.3.The feasibility of replacing the instantaneous decreasing rate with the average decreasing rate of the grid points is verified.Even when the height difference between the grid point and the actual station to be measured is 500 m,the error caused by naturalization is less than 2mm.At the same time,it also compares the effect of correcting the ZTD with or without considering the decreasing rate of the elevation direction,that is,the accuracy of the external ZTD solution after taking into account the decreasing rate of the elevation direction.The experimental results show that even when there is only a height difference of50 m,using a decreasing rate to correct the elevation direction can improve the accuracy by about 1.7cm to 1.8cm.Therefore,it is necessary to use the decreasing rate of grid model fitting to change the elevation direction of the station.4.Finally,the system error of the grid model is further optimized by using the actual measured ZTD data.Similar to the correction error of the broadcast error of the network RTK,the ZTD error at the interpolated grid points and the real site is used to carry out the higher-order term information of the grid points.make up.The experiment showed that the overall average accuracy improved by 1.4cm,the accuracy of Wenchuan Station,the best correction,increased by 2.2 cm,and Anyue,the worst correction,improved by 0.4 cm.5.A weighted average temperature model for the Sichuan region was constructed,combined with the tropospheric delay calculated by the ZTD grid model to achieve GNSS retrieval of atmospheric precipitation,and compared with the true PWV of the sounding station,The results show that the correlation coefficients are higher than 0.9,it proves the reliability of the result of inversion of water vapor using the atmospheric delay calculated by the grid model.