Observation strategies for current and future geodetic very long baseline interferometry

Doctoral thesis, 2018

Geodetic Very Long Baseline Interferometry (VLBI) is an essential technique for space geodesy. It is uniquely capable of simultaneously observing all Earth Orientation Parameters (EOP) and directly giving access to the Earth’s rotation angle (related to Universal Time, UT1). The EOP provide the link between the terrestrial and celestial reference frame. The latter is defined by positions of extra-galactic radio sources observed with geodetic VLBI whereas the terrestrial frame is realised through station positions and velocities. Ongoing phenomena such as the sea-level rise caused by global warming have magnitudes in the millimetre per year range. An accurate global reference frame is therefore crucial for reliably measuring these changes. The Global Geodetic Observing System (GGOS) and its VLBI component, the VLBI Global Observing System (VGOS), are designed to meet these challenges. The transition to the VGOS era brings challenges for all aspects of geodetic VLBI: telescope design, receiver development, recording, data transfer, correlation, observation planning, and analysis. The transition to VGOS involves gradually phasing out the legacy dual- frequency S/X telescopes, while delivering all IVS geodetic products and ensuring the continuity of the time series of geodetic parameters. The VGOS targets include continuous observations and delivery of initial geodetic products in less than 24 hours. This will require a fully automated VLBI analysis chain to make results available in near-real time.
This thesis aims at contributing to the improvement of current geodetic VLBI products and supporting the transition from the legacy S/X systems to observations in the VGOS era. Broadband VGOS observations necessitate upgrades for the receiver chain and the data recording devices. The Onsala Space Observatory (OSO) operated its analogue and a new digital back-end in parallel for almost two years. We present the results from a comparison, in which the new system was found to have no biases w.r.t. the old setup. We also investigate ways to improve the current IVS Intensive sessions. This involves using approaches that have relevance for the upcoming VGOS observations. We present fully automated analysis of INT1 sessions between 2001 and 2015 to investigate different analysis strategies and the impact of mapping functions, the use of auxiliary data, and lack of recent a priori EOP on the UT1-UTC accuracy. Up-to-date a priori polar motion was recognized as a key factor for the accuracy of UT1 estimates. Results from implementation and testing of fully automated robust L1-norm based ambiguity estimation are presented. We find that the L1-norm outperforms least-squares for ambiguity estimation. Lastly, optimal locations for a third station in tag-along mode for INT sessions are determined. We conclude that UT1-UTC WRMS can be reduced to 61 % (INT1) and 67 % (INT2) of the WRMS without the tag-along station. The UT1-UTC was improved significantly even without optimised schedules.