Atmospheric refraction and system stability investigations in short-baseline VLBI observations

Geodetic very long baseline interferometry observations of radio telescopes, which are located in an immediate neighborhood, provide an optimal experimental setup for investigations in atmospheric refraction and system stability issues. In this study, a series of dedicated VLBI sessions with very short baselines, referred to as WHISP (Wettzell HIgh SPeed) sessions, has been designed. Six sessions were observed, three on a 123-m baseline only and another three adding to the short baseline the VLBI telescope at Onsala in Sweden. New is that these sessions and their analysis benefit from the high velocities of the radio telescopes in changing from radio source to radio source providing an unprecedented number of observations on such a short baseline and consequently an extremely reliable parameter estimation. The resulting European triangle is used to compare atmospheric time series derived by two adjacent baselines. Before this could be done, the stability of the observing system, in particular the noise contributions induced by the stability of the hydrogen maser clocks and the correlation process, is investigated to separate the individual uncertainty components. We determined the uncertainty level of the observing systems to be on the order of 10 ps. We were also able to quantify the effect of applying manual phase calibration instead of scan-by-scan system calibration, which is on the order of about 20 ps in this specific example and therefore not negligible. It could be substantiated that estimating clock parameters in geodetic VLBI absorb other effects because direct H-Maser comparisons produce variations at the 5–10 ps level while clock estimates are a factor of 3–6 times larger. Atmospheric refraction has been investigated at different stages: Zenith wet delays were estimated in a differential model for one station relative to the other station and in an absolute sense using two adjacent baselines between the two Wettzell antennas and the Onsala telescope. In both cases, the variations in the estimated atmospheric parameters are found to be of the order of only 1–3 mm and the remaining variations are assigned to unmodeled random effects, particularly refractivity fluctuations in the neutral atmosphere. This was confirmed by introducing an atmospheric turbulence model yielding WRMS post-fit residuals between 7 and 20 ps when clock and correlator effects have been subtracted.