A Fusion Modeling Method for Multi-Source Zenith Tropospheric Delay Based on Adaptive Weighting

High-resolution zenith tropospheric delay (ZTD) is of great significance for regional navigation and positioning. To address the challenge of obtaining high spatial resolution ZTD with a single technique, this paper proposes an adaptive weighted multi-source ZTD fusion modeling (AZF) method. This method uses GTrop and polynomial models to express the linear features and bias information in the ZTD time series, and respectively and adaptively determines the weights of multi-source data through the improved variance component estimation method to obtain more reasonable and accurate high-resolution ZTD model coefficients. The experiment selects the ZTD data of the Global Navigation Satellite System (GNSS) and the fifth generation of ECMWF reanalysis data set (ERA) in the western region of South America, generating high-resolution ZTD fusion images with hourly resolution in 2020. With the independent GNSS and ERA5 data as references for verification, the experimental results show that the average RMSE of ZTD obtained by the AZF method proposed in this paper is 8.7 mm and 8.8 mm, respectively, corroborating the high quality and high resolution of AZF-obtained ZTD.