The GFZ VLBI Contribution to ITRF2020 Employing PORT - Impact of Non-tidal Loading on Sub-secular Scale

Abstract:

The subject of this contribution is the network scale estimated from the analysis of very long baseline interferometry (VLBI) data. Our objective is to evaluate the scale distortion at sub-secular time scales induced by: (i) non-tidal loading; and (ii) network geometry. Mass transport within Earth’s fluid envelope gives rise to non-tidal loading, which in turn displaces stations at the cm-level from the seasonal (hydrological and atmospheric) to the synoptic (atmospheric and oceanic) band. Neglecting to model this effect at the observation equations level in VLBI analysis manifests into seasonal scale variations up to 2 mm and in scale scatter up to 10 mm. Compared to other space geodetic techniques the VLBI scatter is larger, a fact we partly attribute to the observation geometry which features fewer stations often observing in non-overlapping networks. We have performed global solutions, stacking normal equations at intervals ranging from one day to one quarter year, to slightly mitigate the network effect thus facilitating a more appropriate inter-technique comparison. While the scale scatter decreases by increasing the span of the data accumulated for each global solution and the network size, the annual amplitude remains close to 1.0 mm. The latter differs across the techniques that contribute to the ITRF2020 - that is, VLBI, GNSS, SLR, and DORIS - owing to diverse locations of the individual network elements. Moreover, we evaluate the loading impact on station positioning at sites where more than one VLBI station is co-located with stations of other techniques. The single-session analysis was carried out employing the least-squares module of the Potsdam Open-source Radio interferometry Tool (PORT), a software package forked from VieVS, dedicated to the analysis of S-/X-Band and VGOS VLBI observations. To contribute to ITRF2020, the VLBI group at the GFZ analyzed over 18 million multi-band group delays distributed in 6265 sessions, spanning the period 1979-2020, featuring 184 stations and 5237 radio sources.