1 Realizing a Celestial Reference Frame from the VGOS
We have analyzed delay observations from 177 VGOS sessions from December 2017 to June 2024 and made both session-wise and global solutions. Based on these data, a VGOS CRF comprising positions for 377 sources was produced (Xu & Charlot, 2025).
The important findings for the VGOS CRF can be summarized as follows:
(1) For sources with declination above 20 degrees, the residuals of the source coordinates are on the order of 0.1 – 0.2 mas (in the sense of standard deviation), and the dominant cause is source structure.
(2) Based on the impact on geodesy and astrometry, source structure may be divided into two parts: invisible/in-beam structure (i.e. within the beam size) and visible structure (on larger scales). The latter mainly causes closure delays leading to large post-fit delay residuals in geodetic solutions whereas the former causes source position changes.
2 Astronomical calibration and imaging of VGOS experiments
In order to perform the structure correction during the geodetic analysis, we need to calibrate the visibilities and image the interferometric data as is commonly done in astronomical analysis. The goal is to correct for various instrumental and propagation effects which affect the visibilities. We have made significant progress in the reduction of geodetic data, and we are routinely carrying out the calibration of VGOS sessions (Chamani & Xu, 2025).
There are about 200 24-hour VGOS sessions as of May 2025, and these 24-hour sessions are currently observed with a rate of once per week. In the future, VGOS is planning to observe 24/7, and produce large amounts of data for geodetic analysis, making the automation of the data reduction necessary. We are currently developing a user-friendly calibration pipeline for AIPS in its Python interface, ParselTongue.
With the development of novel imaging algorithms, we are able to form so-called closure quantities, which can be used to correct for station-based errors. We are currently using the ehtim algorithm (Xu et al., 2021) to image our data.
3 Kinematic analysis of MOJAVE AGN
While we are working on the calibration and imaging of geodetic data, we are implementing the structure correction in the geodetic analysis software pSolve. As a test for both the methodology and the accuracy of the geodetic products with source structure correction, we are using high-quality astronomical data, which are suitable for geodetic analysis as well. For this purpose, we are using the data of the Monitoring Of Jets in Active Galactic Nuclei with VLBA Experiments (MOJAVE) program, which has been observing hundreds of AGN for over quarter of a century at 15 GHz with the Very Long Baseline Array (Lister at el., 2023). A group member is in charge of model fitting of MOJAVE observations since 2022.
4 Modeling source structure for the MOJAVE observations
MOJAVE provides an excellent testbed for modeling source structure. So far, we have processed 147 24-hour MOJAVE sessions in geodetic mode to obtain the time series of geodetic products. With a close cooperation with the MOJAVE team, we were provided with the structure models for 500 MOJAVE sources over the decades. The group delay correction for source structure is calculated from these aligned image models for each observation and is added to the theoretic model in geodetic solutions. Promising results have been obtained from modeling source structure in MOJAVE observations.
5 Mitigating the effects of source structure
With modeling source structure as the final goal, there can be two strategies to mitigate these effects as alternatives: (1) weighting the geodetic observations based on the information of source structure and (2) actively avoiding the observations with the jet angles and baseline vectors in parallel in scheduling.
For the first strategy, we add source structure-dependent noise terms into the stochastic model of the least-squares estimation process in the geodetic solutions (Niko et al., 2024).
For the second strategy, we have analyzed a scheduling method in intensive sessions (single baseline observations between OE/OW and IS for a duration of one hour) that constrains observations based on the angle between the source jet direction angle and the projected baseline orientation angle (Wolfs et al., in preparation).
6 improving the imaging capability for VGOS
Six dedicated 24-h VGOS–R&D sessions were used to investigate the impact of SNR-based scheduling approach, which can lead to significant improvement in the VGOS imaging capability, by greatly increase the number of observations and scans per source. This was achieved by establishing an SNR-based scheduling approach with shorter observation times of around 10s and reduced overhead times (Schartner et al., 2024; 2025).
7 Investigating the impact of atmospheric turbulence
Another major systematic error source in VGOS observations is the atmospheric effect. A detailed study has been carried out in the project to investigate the impact of atmospheric turbulence on VGOS observations. Our findings suggest that such coarse temporal sampling is inadequate for fully modeling the stochastic effects induced by turbulent refractivity changes (Albentosa-Ruiz & Xu, in preparation).
