Comparison of tropospheric delay modeling approaches in VLBI analysis

Abstract

In VLBI analysis, the tropospheric signal delay is modeled using so-called mapping functions and gradients. They relate a delay in zenith direction to the delay at the elevation and azimuth at which an observation takes place. The objective of this thesis is to investigate how different settings for modeling of the tropospheric delays and for the estimation of tropospheric parameters in the Vienna VLBI and Satellite Software (VieVS) affect the results of VLBI analysis. To do this, 3296 legacy S/X sessions from 2000 until 2020 and 36 VGOS sessions from 2017, 2019 and 2020 are analyzed. A main focus of investigation are baseline length repeatabilities (BLRs), which serve as a good measure of the accuracy achievable with a specific setting.The first part of the thesis deals with testing different choices of mapping functions for the hydrostatic and wet delays. It is found that the differences between any combination of the Vienna Mapping Functions (VMF1 and VMF3) are negligible when comparing baseline length repeatabilities. However, it is further revealed that using VMF3 for the hydrostatic and VMF1 for the wet delay component will introduce a bias to the estimated stations heights of about one to two millimeters height decrease on average, compared to either using VMF1 or VMF3 consistently for both delay components. This result is important because this inconsistent approach was used as default by the Vienna Analysis Center at the start of this investigation. A section of the thesis is then dedicated to the differences between VMF1 and VMF3 which give rise to this bias in a mixed application.In the second part, the influence of the estimation intervals for zenith wet delays and gradients on baseline length repeatabilities is investigated. A small influence is observed. Here, shorter estimation intervals improved average baseline length repeatabilities for S/X sessions, while longer intervals did so for VGOS sessions.