Research On Modeling For Zenith Tropospheric Delay Based On GNSS And ERA5 Data

Tropospheric delay error is currently one of the main factors affecting global navigation satellite system(GNSS)positioning accuracy,which is often corrected by empirical tropospheric delay model.In order to improve the accuracy of tropospheric delay correction over China,this paper evaluates the applicability of zenith tropospheric delay(ZTD)calculated from the latest ERA5 data,and establishes a seasonal vertical scaling model of ZTD based on ERA5 data,in addition,a ZTD interpolation algorithm based on this model is proposed.Based on the above models,three types of zenith tropospheric delay models over China were established using data of crustal movement observation network of China(CMONOC)and ERA5 meteorological data with high precision and high spatiotemporal resolution.The main research work of this paper is as follows:(1)Explain the principle of tropospheric delay generation,and summarize the research status of tropospheric delay modeling at home and abroad;(2)The measured ZTD data observed from 7 IGS stations distribution in China is used to assess the applicability of ZTD calculated with data from ERA5.The results show that ZTD bias and root mean square error(RMSE)calculated from the meteorological reanalysis data have obvious seasonal changes,which are larger in summer and smaller in winter;the average RMSE of the ERA5 dataset is 20.06 mm,which is better than the ERA-Interim dataset;The ERA5 data with high temporal resolution(1 h)can basically reflect the short-period diurnal variation of ZTD,and the ZTD bias and the RMSE calculated from ERA5 data in 1 degree spatial resolution are relatively small.Therefore,it is more appropriate to use the ERA5 data with a horizontal resolution of 1 degree and a temporal resolution of 1 h to establish a ZTD model over China;(3)Based on an optimized piecewise profile function and ERA5 monthly averaged data on pressure levels with a horizontal resolution of 1 degree over China(15°?54°N,73°?136°E),a seasonal ZTD vertical scaling model was established using a trigonometric function containing an annual cycle change term and a semiannual cycle change term,which achieves high-precision vertical scaling of the ZTD over China,and its average RMSE is 0.8 mm.(4)Based on the seasonal ZTD vertical scaling model and simplified spatial regression method,and combined with the idea of projection and extension,a high-precision ZTD interpolation algorithm is proposed.The results show that the interpolation algorithm in this paper can achieve high accuracy under different height differences,whose average RMSE is about 10 mm.What's more,it can also achieve an accuracy of 10.99 mm at the station with the largest height difference(1 771 m),which shows strong applicability.Compared with the inverse distance weighting method and the spatial regression method,the RMSE reductions of of the proposed algorithm is 86%and 41%at the station with the smallest height difference,and 97%and 70%at the station with the largest height difference,which shows that the algorithm in this paper effectively avoids the loss of ZTD interpolation accuracy caused by large height differences.(5)Based on the idea of independent modeling by period,CMONOC data and ERA5 data with high spatio-temporal resolution are independently modeled in the time domain for 24 periods.The ZTD interpolation algorithm proposed in this paper is used to implement the spatial interpolation of the reference station ZTD,and the bilinear interpolation method combined with the seasonal ZTD vertical scaling model is used to implement the spatial interpolation of the grid point ZTD.Therefore,a CTropG model and a CTropE model were established respectively,and a CTropM fusion model was established based on the law of error propagation.Taking the ZTD sequence of IGS station in 2018 as the reference value,the bias and RMSE of the three models are calculated,and compared with the GPT2w model.The results show that the three models established in this paper can better reflect the short-period diurnal variation of the ZTD.All three models have good internal and external accord accuracy,and are better than the GPT2w model.The average RMSE of the CTropG model is 46.60 mm,which is 2.4%lower than the GPT2w model,and the average RMSE of the CTropM model is 46.68 mm,which is 2.3%lower than the GPT2w model.The average RMSE of the CTropE model is 47.77 mm,slightly better than the GPT2w model.