Assessment of various processing schemes and solution strategies to improve the performance of GNSS tropospheric tomography

Abstract

In den letzten Jahren hat sich das Globale Navigationssatellitensystem (GNSS) als ein wertvolles Instrument für die Fernerkundung der Atmosphäre erwiesen. In diesem Zusammenhang hat sich die GNSS-Tomographie zu einer äußerst vielversprechenden Technik entwickelt, um die räumlich-zeitliche Struktur der Troposphäre zu rekonstruieren. Daher kann diese Methode eine ausgezeichnete Alternative zur Überwachung von Wasserdampf und feuchten Refraktionsfeldern zu geringen Kosten und einer angemessenen räumlichen Auflösung im Vergleich zu konventionellen Beobachtungen, wie Radiosonden- und Radio-Okkultationsprofilen, bieten. Es gibt jedoch noch einige Herausforderungen und offene Fragen bei der GNSS-Tomographie, welche die Qualität des rekonstruierten Feldes stark beeinflussen. Daher besteht das Hauptziel dieser Dissertation darin, verschiedene Strategien zur Lösung einer Vielzahl dieser Probleme zu untersuchen.

In recent years, the Global Navigation Satellite System (GNSS) has turned out to be a valuable tool for remotely sensing the atmosphere. In this context, GNSS tomography evolved to an extremely promising technique to reconstruct the spatio-temporal structure of the troposphere. Therefore, this method can offer a permanent monitoring service for water Vapour and wet refractivity fields at low cost and a reasonable spatial resolution compared to conventional observations, like radiosonde and radio occultation profiles. However, there are still some challenges and open questions in GNSS tomography which extremely affect the quality of the reconstructed field. Hence, the main objective of this dissertation is to investigate different strategies to solve some of them.The economic issue to deploy multi-frequency receivers with a sufficient spatial resolution of a few tens of kilometers is one of the challenges for GNSS tomography. Therefore, the feasibility of using single-frequency observations in GNSS tomography as an alternative approach is investigated. Another challenge of GNSS tomography relates to different parameterization methods for computing the design matrix. Therefore, the effect of the straight-line method versus the ray-tracing method as well as the impact of considering the topography in the tomography model is studied for computing the design matrix. Further attention is given to multi-GNSS observations in GNSS tomography due to improving observation geometry compared to a sole GPS/ GLONASS system scenario. Therefore, by focusing on GALILEO's effect, the impact of different constellations is investigated to retrieve a wet refractivity field. GNSS tomography is also suffering from the insufficient spatial coverage of GNSS signals in the voxels within the given time window. Hence, the design matrix is sparse, and the observation equation system of the tomography model is mixed-determined. Thus, physical meaningful constraints as well as external data sources should be applied. In this dissertation, the new dataset from the Geostationary Operational Environmental Satellite (GOES) sounder supplements the system of observation equations and consequently, the tomographic solution leads to an improved reconstructed wet refractivity field. Besides, this method is a kind of discrete ill-posed problem. So, all singular values of the structure matrix (A) in the tomography problem decay gradually to zero without any noticeable gap in the spectrum. Hence, slight changes in the measurements can lead to extremely unstable parameter solutions. In consequence, the regularization method should be applied to stabilize the inversion process and ensure a stable and unique solution for the tomography problem. The algebraic reconstruction techniques (ART) and the Total Variation (TV) method are examined to reconstruct a regularized solution with acceptable accuracy. Moreover, the TV method can also reconstruct a promising wet refractivity field without any initial field in a shorter time span. Thereby, retrieving the wet refractivity field using this method is also investigated. A further attempt is given to analyse the quality of the reconstructed field in GNSS tomography. To the author' best knowledge for the first time in GNSS tropospheric tomography, the spread of the resolution matrix is employed to assess the quality of the retrieved wet refractivity solution without a need to use reference observations and calculate statistical measures like RMS and Bias in this method.