A method for calculating real-time ZTD grid data in Chinese regions based on GNSS ZTD modified ERA5 grid products

Abstract

The low spatial and temporal resolution and poor accuracy of real-time zenith tropospheric delay (ZTD) have always been one of the main constraints during the development of real-time GNSS applications. This paper presents a method to obtain real-time ZTD grid data in Chinese regions utilizing the ZTD-modified ERA5 grid products from GNSS stations. Initially, the original ZTD fitting function for vertical variation is improved. Then, the real-time ZTD grid data for the Chinese region (RZTDG) could be obtained by correcting the 7-day earlier ERA5 data through the real-time ZTD data from the GNSS stations. In this paper, we analyze the performance of the model under various weather conditions, and the results show that it still keeps a high accuracy even under variable weather conditions. ZTD data from 66 radiosonde stations in China in 2022 are used as reference values to evaluate the accuracy of the RZTDG. Experimental results show that the RMSE (2.3 cm) and MAE (1.8 cm) of the RZTDG are comparable to the ERA5 data of the day (RMSE: 2.4 cm, MAE: 1.8 cm) and significantly better than the accuracy of the HGPT2 (RMSE: 5.1 cm, MAE: 4.1 cm). Meanwhile, the Bias of RZTDG (0.1 cm) is much smaller than the Bias of HGPT2 (−0.9 cm) and the Bias of ERA5 (−0.8 cm), proving that the method in this paper has less deviation from the radiosonde ZTD data. Finally, the convergence times of the float PPP with and without tropospheric constraints are counted for 120 GNSS stations for DOY 11–13, DOY 101–103, and DOY 192–194 in 2022. The results show that the convergence time of the float PPP is reduced by 30% after constraining the ZTD using RZTDG.