Abstract
The use of global navigation satellite systems (GNSS) precise point positioning (PPP) to estimate zenith tropospheric delay (ZTD) profiles in kinematic vehicular mode in mountainous areas is investigated. Car-mounted multi-constellation GNSS receivers are employed. The Natural Resources Canada Canadian Spatial Reference System PPP (CSRS-PPP) online service that currently processes dual-frequency global positioning system (GPS) and Global’naya Navigatsionnaya Sputnikovaya Sistema (GLONASS) measurements and is now capable of GPS integer ambiguity resolution is used. An offline version that can process the above and Galileo measurements simultaneously, including Galileo integer ambiguity resolution is also tested to evaluate the advantage of three constellations. A multi-day static data set observed under open sky is first tested to determine performance under ideal conditions. Two long road profile tests conducted in kinematic mode are then analyzed to assess the capability of the approach. The challenges of ZTD kinematic profiling are numerous, namely shorter data sets, signal shading due to topography and forests of conifers along roads, and frequent losses of phase lock requiring numerous but not always successful integer ambiguity re-initialization. ZTD profiles are therefore often only available with float ambiguities, reducing system observability. Occasional total interruption of measurement availability results in profile discontinuities. CSRS-PPP outputs separately the zenith hydrostatic or dry delay (ZHD) and water vapour content or zenith wet delay (ZWD). The two delays are analyzed separately, with emphasis on the more unpredictable and highly variable ZWD, especially in mountainous areas. The estimated delays are compared with the Vienna Mapping Function 1 (VMF1), which proves to be highly effective to model the large-scale profile variations in the Canadian Rockies, the main contribution of GNSS PPP being the estimation of higher frequency ZWD components. Of the many conclusions drawn from the field experiments, it is estimated that kinematic profiles are generally determined with accuracy of 10 to 20 mm, depending on the signal harshness of the environment.
Highlights
Global navigation satellite systems (GNSS) precise point positioning (PPP) is used routinely to estimate the zenith tropospheric delay (ZTD) and its components at static sites worldwide to study lower atmospheric layers and contribute to the understanding of weather patterns and prediction
The concept of using GNSS-PPP to estimate ZTD profiles and components in kinematic mode in sub-optimal signal line-of-sight conditions encountered on mountain roads has been demonstrated, with slightly lower performance than that of the static case
The Vienna Mapping Function 1 (VMF1) performance to model the variable ZTD profiles was found to be realistic, especially in its ability to account for large height variations during kinematic testing
Summary
Global navigation satellite systems (GNSS) precise point positioning (PPP) is used routinely to estimate the zenith tropospheric delay (ZTD) and its components at static sites worldwide to study lower atmospheric layers and contribute to the understanding of weather patterns and prediction. The major limitations of kinematic PPP estimation in mountains versus the static approach are (1) the presence of unpredictable multipath and changing obstructions along the trajectories due to the forestry canopy and mountainous topography, both resulting in frequent losses of phase lock and sub-optimal carrier phase integer ambiguity resolution; (2) data sets of up to several hours instead of days for static measurements at permanent stations; and (3) high correlation with epoch-by-epoch position estimates, rather than a single static position. Tests with this antenna versus a geodetic grade NovAtel (Calgary, Canada) GPS-703-GGG antenna showed more consistent results with the latter, which was used for the tests presented below It is, possible to overcome the low-cost antenna limitation by performing a relative antenna calibration with nearby precise receiver-antenna systems operating in static mode, as demonstrated by [16]; this was outside the scope of this study. During static and kinematic testing, the antennas were mounted between 0.8 m and 10 m of each other
Sign-in/Register to view highlights & summary 
Published Version (
Free)
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