Accelerating multi-GNSS and multi-frequency PPP-AR convergence using global operational NWM forecasts

Abstract Precise point positioning with ambiguity resolution (PPP-AR) provides globally available centimeter-level positioning, but its convergence time remains a limiting factor for time-critical applications. One important contributor to slow convergence is short-term mismodeling of the zenith wet delay (ZWD), which affects the float ambiguity estimates and delays integer fixing. Numerical weather models (NWMs) offer physically realistic, globally consistent tropospheric predictions that can mitigate this issue. This study investigates the use of operational Global Forecast System (GFS) forecasts to augment real-time PPP-AR through adaptive ZWD accuracy constraints. Using ZTD data from nearly 400 IGS stations in 2024, we analyze the performance of GFS-derived delays and identify a clear latitudinal dependence, with RMS differences reaching about 2 cm in tropical regions and dropping to below 1 cm at high latitudes. Based on these characteristics, a latitude-dependent accuracy model is constructed and applied to dual- and triple-frequency multi-GNSS PPP-AR for 223 MGEX stations across four representative months. The results demonstrate that GFS-derived tropospheric augmentation reduces PPP-AR convergence times by 20–40% globally, with larger improvements in summer and at mid- to high latitudes. These seasonal and latitudinal patterns arise from atmospheric variability: summer months exhibit stronger short-term water-vapor fluctuations that amplify the benefit of external ZWD constraints, while mid- and high-latitude regions generally show more stable tropospheric conditions, allowing NWM-derived delays to be more accurate than in the tropics. In addition to convergence, the augmentation also improves the time to first fix (TTFF), defined as the first epoch at which ambiguities remain fixed for ten consecutive epochs, with typical reductions of about one minute and relative improvements of roughly 15–25%, particularly in mid- and high-latitude regions. After convergence, the augmentation has minimal impact on positioning accuracy. These findings demonstrate that operational GFS forecasts can serve as an effective global source of tropospheric information for accelerating PPP-AR, offering a practical pathway toward improved performance in time-sensitive applications.