Abstract
Global navigation satellite system (GNSS) remote sensing of the troposphere, called GNSS meteorology, is already a well-established tool in post-processing applications. Real-time GNSS meteorology has been possible since 2013, when the International GNSS Service (IGS) established its real-time service. The reported accuracy of the real-time zenith total delay (ZTD) has not improved significantly over time and usually remains at the level of 5–18 mm, depending on the station and test period studied. Millimeter-level improvements are noticed due to GPS ambiguity resolution, gradient estimation, or multi-GNSS processing. However, neither are these achievements combined in a single processing strategy, nor is the impact of other processing parameters on ZTD accuracy analyzed. Therefore, we discuss these shortcomings in detail and present a comprehensive analysis of the sensitivity of real-time ZTD on processing parameters. First, we identify a so-called common strategy, which combines processing parameters that are identified to be the most popular among published papers on the topic. We question the popular elevation-dependent weighting function and introduce an alternative one. We investigate the impact of selected processing parameters, i.e., PPP functional model, GNSS selection and combination, inter-system weighting, elevation-dependent weighting function, and gradient estimation. We define an advanced strategy dedicated to real-time GNSS meteorology, which is superior to the common one. The a posteriori error of estimated ZTD is reduced by 41%. The accuracy of ZTD estimates with the proposed strategy is improved by 17% with respect to the IGS final products and varies over stations from 5.4 to 10.1 mm. Finally, we confirm the latitude dependency of ZTD accuracy, but also detect its seasonality.
Highlights
The global navigation satellite system (GNSS) signal delay depends on pressure, temperature, and water vapor content along the propagation path, which creates a link between GNSS and meteorology
With a total number of 20 different real-time solutions for a worldwide spatially distributed set of stations and two time periods, we investigate the impact of precise point positioning (PPP) processing parameters on estimated zenith total delay (ZTD)
We show that all four GNSSs can provide real-time ZTD solutions independently
Summary
The global navigation satellite system (GNSS) signal delay depends on pressure, temperature, and water vapor content along the propagation path, which creates a link between GNSS and meteorology. Troposphere delay is treated as an error source in precise GNSS positioning, there is a great potential of exploiting troposphere delays. GNSS remote sensing of the troposphere, called GNSS meteorology (Tralli and Lichten 1990; Bevis et al 1992), provides observations with spatial and temporal resolutions that are higher than any other tropospheric sensing technique and operates in all weather conditions (Bennitt and Jupp 2012). Post-processing of GNSS observations can provide results with accuracies comparable to measurements of traditional precipitable water vapor sensors (Rocken et al 1997; Haase et al 2003; Satirapod et al 2011).
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
