Troposphere modeling for precise GPS rapid static positioning in mountainous areas

Abstract

In global navigation satellite system precise positioning, double differencing of the observations is the common approach that allows for significant reduction of correlated atmospheric effects. However, with growing distance between the receivers, tropospheric errors decorrelate causing large residual errors affecting the carrier phase ambiguity resolution and positioning quality. This is especially true in the case of height differences between the receivers. In addition, the accuracy achieved by using standard atmosphere models is usually unsatisfactory when the tropospheric conditions at the receiver locations are significantly different from the standard atmosphere. This paper presents an evaluation of three different approaches to troposphere modeling: (a) neglecting the troposphere, (b) using a standard atmosphere model, and (c) estimating tropospheric delays at the reference station network and providing interpolated tropospheric corrections to the user. All these solutions were repeated with various constraints imposed on the tropospheric delays in the least-squares adjustment. The quality of each solution was evaluated by analyzing the residual height errors calculated by comparing the estimated results to the reference coordinates. Several permanent GPS stations of the EUPOS (European Position Determination System) active geodetic network located in the Carpathian Mountains were selected as a test reference network. The distances between the reference stations ranged from 64 to 122 km. KRAW station served as a simulated user receiver located inside the reference network. The user receiver ellipsoidal height is 267 m and the reference station heights range from 277 to 647 m. The results show that regardless of station height differences, it is recommended to model the tropospheric delays at the reference stations and interpolate them to the user receiver location. The most noticeable influence of the residual (unmodeled) tropospheric errors is observed in the station height component. In many cases, mismodeling of the troposphere disrupts ambiguity resolution and, therefore, prevents the user from obtaining an accurate position.

Variance components estimation of residual errors in GPS precise positioning

“…Satirapod and Chalermwattanachai (2005) and Wielgosz et al (2011) may be cited. Some of the publications where the authors have researched the behaviour of tropospheric residuals are Musa (2007) and Ibrahim and El-Rabbany (2007).…”

Near real-time estimation of water vapour in the troposphere using ground GNSS and the meteorological data

“…The positioning using the above methods can be supported by Ground Base Augmentation System (GBAS), which provides better results stability in the area of the GBAS network (Kee et al, 1991). One of the elements supporting precise positioning, especially for height component, is a model of the neutral atmosphere (troposphere), computed for the GBAS network area (Wielgosz et al, 2011;Grejner-Brzezinska et al, 2009). The Poland territory is covered by dense network of GNSS stations in the frame of a GBAS system called ASG-EUPOS.…”

Assessment of GPS + Galileo and multi-frequency Galileo single-epoch precise positioning with network corrections

“…As our previous research has shown for precise positioning with very short sessions, it is not recommended to estimate zenith tropospheric delay (ZTD) in the rover solution due to the slight change in satellite geometry (Wielgosz et al 2011a). For reliable estimation of ZTD, a significant distance between the stations is required (Dach et al 2007;Wielgosz et al 2011b).…”