This thesis is dedicated to the development of two new tropospheric mapping functions for GNSS data processing, based on a high resolution mesoscale numerical weather model (NWM). NWMs have proven to be beneficiary in the processing of GNSS and VLBI data, both for deriving mapping functions and for providing a priori information such as zenith hydrostatic delay (ZHD). The mapping functions derived here make a greater use of the NWM information than the mapping functions currently recommended by the International GNSS Service. In addition to using a single vertical pro¯le at the site in order to derive mapping functions under the assumption of an azimuthally symmetric atmosphere, the NWM was also ray traced every thirty degrees in azimuth. This way, a complete volume of the atmosphere is sensed, and better modelling is expected if the NWM does indeed provide an accurate representation of the atmosphere, by accounting for azimuthal variations. An emphasis was put in this thesis on assessing the mathematical models used to vertically interpolate meteorological information, as they play a key role in computing the refractivities in the ray tracing algorithm. Error sources were identified and quantified. As expected, water vapour is the major source of error. However, the results showed that the model used for the total pressure induced a systematic bias. To derive an azimuth dependent mapping function, the Marini model traditionally used had to be left in favor of a cubic spline interpolation (CSI). This new approach was validated by comparing the performance of the new azimuthally symmetric mapping functions against the updated Vienna mapping functions (VMF1), the best mapping functions currently available. Similar positioning performances were obtained, therefore validating the CSI based approach. The performance of new azimuth dependent mapping functions (AMF) in handling the troposphere asymmetry were compared to those obtained when estimating horizontal tropospheric gradients with an azimuthally symmetric mapping function. Results show a good agreement in the modelling of the asymmetry, and that estimating gradients is justified. The gradient solution performed better overall, although it failed for some sites, and better inter-station consistency was obtained with the AMF. This thesis also investigated the role of the tropospheric modelling in the retrieval of the atmospheric pressure loading (APL) in GNSS data processing, which is now part of the IGS 2008 recommendations. The results show that differential height time series obtained with different tropospheric modelling can correlate with the APL signal to a level up to 0.7. In other words, the choice of tropospheric modelling strategy does greatly influence the retrieval of the APL.
| Identifer | oai:union.ndltd.org:bl.uk/oai:ethos.bl.uk:514829 |
| Date | January 2009 |
| Creators | Orliac, Etienne J. |
| Publisher | University of Nottingham |
| Source Sets | Ethos UK |
| Detected Language | English |
| Type | Electronic Thesis or Dissertation |
| Source | http://eprints.nottingham.ac.uk/10861/ |
Page generated in 0.002 seconds