Evolution of the radio source position uncertainties in radio astrometric catalogs over the past three decades

TL;DR

This work quantifies the evolution of extragalactic radio source position uncertainties in VLBI-derived catalogs from 1997 to 2025. Using 30 catalogs processed with a consistent Calc/Solve framework, it analyzes the semi-major axis of the error ellipse (SMA) for all sources and a common subset of 467 sources, revealing that median SMA drops by about a factor of 10 over the period and follows a power-law trend over time. The total number of observations grows from roughly 1.3 million to 14.4 million, and SMA exhibits a strong dependence on the number of observations $N_{obs}$ (roughly quadratic in $\log N_{obs}$) compared to the number of sessions $N_{sess}$ (roughly linear in $\log N_{sess}$). More than 95% of sources achieve SMA below 1 mas, with ~40% below 0.1 mas, indicating substantial improvement in the radio reference frame; the results emphasize that maximizing observation counts per source is the key driver of accuracy enhancements in VLBI catalogs.

Abstract

In this paper, progress in improving the coordinates uncertainty of extragalactic radio sources derived from astrometric and geodetic VLBI observations is investigated. For this purpose, 30 catalogs of radio source positions computed in 1997-2025 were analyzed. Over these years, the median source position uncertainty represented by the semi-major axis of the error ellipse for 467 sources in common to all catalogs improved by one order of magnitude: from 0.156 mas to 0.015 mas. It was also found that the improvement in the position uncertainty over time follows a power law with high accuracy. The number of observations used for determination of the coordinates of the common sources in the input catalogs increased over time from 1.3 to 14.4 million, also following a power law. A discussion of the results led us to the conclusion that the number of observations is the primary factor in improving the source positions uncertainty compared to the number of sessions in which the source was observed.

Figures (5)

• Figure 1: Observation statistics for 30 catalogs used in this work: the number of sources in each catalog (top panel) and the number of observations (bottom panel) for all sources and for 467 sources in common for all catalogs. The line in the bottom panel corresponds to a power law approximation of the number of observations of common sources.
• Figure 2: Cumulative number of observations depending on the number of sources for the usn2025a catalog. Summations starts from sources with the maximum number of observations.
• Figure 3: Median position uncertainties in $\alpha^{\ast}$ and $\delta$, and the semi-major axis of the error ellipse (SMA) for 30 catalogs used in this work for all sources (top panel) and 467 common sources (bottom panel). The line in the bottom panel corresponds to a power law approximation of the SMA data.
• Figure 4: Dependence of the source position uncertainty represented by the semi-major axis of the error ellipse on the number of observations (top panel) and the number of sessions (bottom panel) for the usn2025a catalog. One source with the position uncertainty of 141 mas is not shown.
• Figure 5: Cumulative number of sources with the position uncertainty represented by the semi-major axis of the error ellipse is less than the value given on the $x$ axis (usn2025a catalog). One source with the position uncertainty of 141 mas is not shown.