Abstract
Precise Point Positioning-Real Time Kinematic (PPP-RTK), a synthesis of PPP and RTK techniques, is a highly precise positioning technique that has been extensively studied in the GNSS community. The Zenith Wet Delay (ZWD), an important estimated parameter in PPP-RTK, significantly impacts the user positioning performance. It is crucial to determine the optimal ZWD processing methods for PPP-RTK. The present study contributes to this research trend by designing and contrasting four ZWD processing strategies for PPP-RTK, confirming the benefits of Global Forecast System (GFS) products (a global numerical weather prediction (NWP) model) in both the network and user models of PPP-RTK. In this study, the full-rank ionosphere-float undifferenced and uncombined (UDUC) PPP-RTK network and user models are first derived. To determine the optimal ZWD processing method in PPP-RTK, four ZWD processing strategies are designed and a comprehensive positioning performance evaluation is conducted for the four different ZWD processing strategies using 10-day GPS observation data from 55 GPS stations in the USA. The results show that a priori GFS ZWD information into the PPP-RTK network model's ZWD constraint significantly improves the user positioning performance. With the GFS ZWD constraint, for user stations, the average convergence time after network filter restart is reduced by 12.7%, from 55 to 48 min, the average vertical positioning RMS of the float ambiguity solution (including convergence time) is reduced by 18.8%, from 8.5 to 6.9 cm. Additionally, a priori GFS ZWD information can also offer users an absolute ZWD constraint option, which does not require communication with the network, achieving performance comparable to PPP-RTK network ZWD products.