Assessment of the Troposphere Products Derived From VMF Data Server With ERA5 and IGS Data Over China
The accurate modeling of zenith troposphere delay (ZTD) is significant for data processing of radio-space-geodetic techniques such as GNSS. The delay can be estimated by classic tropospheric models, which need the surface meteorological parameters. Considering the difficulty of obtaining meteorological data caused by various environment factors and the modeling errors in the models themselves, the above methods are often difficult to provide accurate products for GNSS data processing. Based on continuously updated NWP data from the European Centre for Medium-Range Weather Forecasts (ECMWF), the Vienna Mapping Functions (VMF) data server provides a widely used data-source for achieving ZTD with two types of models, that is, the discrete (VFM3) and empirical model (GPT3). Little or no accuracy comparison of these two models with different horizontal resolutions in GNSS community has been conducted over China. This study investigates the performance of the models in terms of calculation of ZHD, ZWD and ZTD, using the hourly values derived from the ERA5 and IGS data as references. The results show that the discrete model outperforms the empirical model and the increasing of horizontal resolution can improve the accuracy of the models. The performance of the empirical model in ZHD/ZWD calculation shows obvious influences of latitude changes and land-sea differences, while the discrete model improves this phenomenon. The empirical distribution function and boxplot well indicate the accuracy differences of the two models in different seasons. We also found that the diurnal variation can only be detected in the ZHD calculation by the discrete model. The ZTD comparison gives the details of the two types of models at GNSS stations.