Atmospheric loading corrections at the observation level in VLBI analysis

• Published in: Journal of geodesy Vol. 83; no. 11; pp. 1107 - 1113
• Main Authors: Boehm, Johannes, Heinkelmann, Robert, Mendes Cerveira, Paulo Jorge, Pany, Andrea, Schuh, Harald
• Format: Journal Article
• Language: English
• Published: Berlin/Heidelberg Springer-Verlag 01.11.2009; Springer Nature B.V
• Subjects: Atmospheric models; Earth and Environmental Science; Earth Sciences; Error analysis; Geodetics; Geophysics/Geodesy; Global positioning systems; GPS; Interferometry; Measurement techniques; Original Article; Simulation; Terrestrial reference frame; VLBI; Atmospheric loading
• DOI: 10.1007/s00190-009-0329-y
• ISSN: 0949-7714

This paper investigates whether in very long baseline interferometry (VLBI) analysis atmospheric loading corrections should be applied a priori at the observation level or whether it is sufficient to correct for atmospheric loading effects a posteriori by adding constant values per session to the estimated station coordinates. Simulated observations at single stations corresponding to the precise point positioning approach of global navigation satellite systems show that the atmospheric loading effect can be fully recovered by a posteriori corrections, i.e., the height differences between both approaches stay well below 1 mm. However, real global VLBI network solutions with sessions from 1984 to 2008 reveal that the effect of neglected atmospheric loading corrections at the stations is distributed to the other stations in the network, thus resulting in station height differences between solutions with observation level and with a posteriori corrections which can be as large as 10 mm and a ‘damping’ effect of the corrections. As soon as the terrestrial reference frame and the corresponding coordinate time series are determined, it would be conceptually wrong to apply atmospheric loading corrections at the VLBI stations. We recommend the rigorous application of atmospheric loading corrections at the observation level to all stations of a VLBI network because the seven parameters for translation, rotation, and in particular the network-scale of VLBI networks are significantly affected.