GNSS PPP with different troposphere models during severe weather conditions

Abstract

Global navigation satellite systems (GNSS)-derived zenith wet delays must be estimated precisely for monitoring weather variations and rain passages in the troposphere. We processed a set of International GNSS Service (IGS) stations within the area affected by the central European Flooding 2013 and assessed the performance of post-processed precise point positioning (PPP) during severe weather by applying different troposphere models: the Vienna Mapping Function (VMF1) together with the European Centre for Medium-Range Weather Forecasts grids, the global mapping function with the Global Pressure and Temperature 2, the Niell mapping function with the University of New Brunswick (UNB), and the VMF1 with the UNB/VMF1 from the National Centers for Environmental Prediction numerical weather model (NWM) data. Wet delay estimates from each PPP session have been verified through the IGS final troposphere products, local surface measurements and double-difference (DD) GNSS solutions performed at the same sites. All the PPP solutions agree well with the IGS. The mean residuals are all below 2.0 mm, and the PPP VMF1 performs better with RMS of 4.0 mm. The PPP solutions applying an NWM offer better agreement with the PPP solutions using real surface measurements to model the troposphere. Both PPP and DD VMF1 solutions agree with RMS of 5.4 mm, which allowed PPP to offer a strong alternative to post-processed DD solutions, even during severe weathers. With respect to the daily ITRF08 coordinate solutions, the gridded VMF1-based PPP height repeatability is better for most stations before and after the observations are corrected for atmospheric non-tidal loading effect.