The paper starts with a short presentation of the conceptual foundations of celestial reference systems and of the IAU decisions taken in 1991. The necessity to define and use them in the framework of General Relativity induces some difficulties that are analyzed. In the case of kinematically defined reference systems, the origins of a possible residual rotation of the frames is analyzed. The same analysis holds also for an isolated dynamical reference system, for which further developments of the metric are necessary. Finally, a few topics concerning the non-isolated case of the barycentric reference system, the transformation between reference systems, and the significance of measurements are mentioned.

After the adoption of the ICRS and its primary materialization by the ICRF in January 1998, the last three years have witnessed a burst of interest in stellar astrometry stimulated by the availability of the ICRF and its optical counterpart. This paper reports on the activities carried out by the members of the IAU appointed working group on reference frames as well on other works directly related to the main goals of this working group.

A general overview of the various factors limiting the accuracy to which astrometric observations can be made is presented. Since this is a very broad topic, it is limited primarily to radio astrometry of extragalactic objects.

At the milliarcsecond scale, most of the extragalactic radio sources exhibit spatially-extended intrinsic structures which are variable in both time and frequency. Such radio structures set limits on the accuracy of source positions determined with the Very Long Baseline Interferometry (VLBI) technique unless their effects in the astrometric data can be accounted for. We review the modeling scheme for calculating source structure corrections and discuss the magnitude and impact of these effects for the sources that are part of the International Celestial Reference Frame (ICRF). Results obtained by applying source structure corrections to actual VLBI observations on the time-varying source 4C39.25 (0923 + 392) are also presented.

The International Celestial Reference Frame (ICRF) is currently realized by the radio positions of compact extragalactic objects obtained by the technique of Very Long Baseline Interferometry (VLBI). The positions of the defining sources with the highest quality are estimated to be accurate at the 0.25 milliarcsecond (mas) level. The limiting accuracy is thought to be due mostly to tropospheric propagation effects and by the apparent motions of the sources due to variable intrinsic structure. Increasing the accuracy of VLBI astrometry beyond current levels will require an improved understanding of the variable emission structures associated with compact extragalactic sources as well as improvements in the modeling of the Earth’s troposphere.

The problem of wide-angle astrometry via interferometry is to recover the two coordinates for each star from the observed delays. In the absence of atmospheric turbulence, the geometrical delay can be defined as: dG, ij(t) ≡ dj(t) ‒ di(t) = Bij(t)·ŝ0 ‒ Cij. Here the geometrical delay is the difference between the delay line lengths di and dj that is required to equalize the effective optical paths from the star to the point of beam combination via each of two apertures i and j. Bij(t) is the baseline between the apertures, ŝ0 is the star position, and Cij is the difference between the ‘fixed’ internal optical path lengths Ci and Cj within the instrument. In principle, sufficient delay measurements would allow solution for the baseline vectors and the delay constants, as well as the positions of the stars. However, the actual situation is greatly complicated by the presence of the atmosphere and the fact that neither the delay ‘constants’ Cij nor the baseline vectors are stable over time. The design of, and the analysis of the data from, any ground-based optical interferometer must overcome all three of these effects. The design, operation, and the analysis of data from the Navy Prototype Optical Interferometer (NPOI) are presented here as examples of how to overcome the effects of the atmosphere and the instrumental instabilities in order to achieve accurate wide-angle astrometry. The status of the implementation of these techniques at the NPOI is presented.

The optical positions of 9 compact extragalactic radio sources have been determined by using the 1.56m, 1m and 60/90cm telescopes with CCDs in China and the Axial Meridian Circle at Nikolaev Astronomical Observatory since Jan. 1996 in a cooperative project between the two observatories. The instrumentation, observations and reduction are briefly described, and the preliminary results are presented. The comparison between the optical and radio positions for 9 sources are also given. More optical positions of radio sources are being processed.

Modern tasks of high precision astrometry demand optical coordinate systems including more faint stars than are available now in current conventional systems (Hipparcos). For this purpose it is suggested to use old photographic observations accumulated in the Pulkovo Observatory. Extensive observational data have been obtained at Pulkovo Observatory during the last century in the framework of several programs, e.g. Pulkovo Galaxy plan, etc. Observations have been made with the Normal Astrograph from 1894 to the present. The data are investigated with the aim of extending the Hipparcos catalog to stars fainter than 11th magnitude. All available observations are taken into account. A description of the material considered is given. The distribution of the selected plates over the celestial sphere is shown as well. Coordinates of faint stars in the Hipparcos system are calculated for selected areas. The accuracy of computed star coordinates is analyzed. Results of the investigation are presented.

Extragalactic radio source positions referred to the first USNO CCD Astrograph Catalog (UCAC1) are presented. They were derived from CCD observations taken with the 1.6-meter Cassegrain telescope at the Laboratório Nacional de Astrofísica, Brazil (LNA). The observations started in 1997 and were carried out in a joint program between USNO, Observatório do Valongo/UFRJ, Brazil, and Observatório Nacional/CNPq, Brazil. They were planned to provide, together with CTIO observations, fiducial points to link the UCAC to the ICRF. The results presented here refer to 24 radio source targets distributed mainly between −30° and −70° declinations. Reduction techniques follow standard procedures. The results are compared with VLBI and with similar precise optical positions, in order to contribute to the evaluation of the UCAC astrometry.

Since the establishment of the Hipparcos Catalog as the defining source of the optical reference frame, densification beyond its ≈ 120,000 stars has been made possible by the utilization of the Tycho-1 Catalog. The ACT, combining the old Astrographic Catalog (AC) data with the Tycho-1 positions, is the best known example of this. The Tycho-2 consortium, led by E. Høg, has performed new reductions on the Tycho data. This not only has increased the astrometric and photometric accuracies of the original 1 million Tycho-1 stars, but also has added an additional 1.5 million stars. The U.S. Naval Observatory led the effort to compute the proper motions of these 2.5 million stars. They are based not only on the AC data but also include over 140 other ground-based catalogs, all directly reduced to the Hipparcos system. The result of these efforts is the Tycho-2 Catalog, available since February 2000. Positions, proper motions, and BT and VT magnitudes are given for 2.5 million stars. The catalog is 99% complete to V=11.0, and 90% complete to V=11.5. Positional accuracies at the mean epochs vary from < 10 mas for stars V < 9 to just under 100 mas for V > 12. Proper motion accuracies are estimated to be 1.3 mas/year to 3.0 mas/year for the same magnitude ranges. Photometric accuracies range from 0.02 magnitudes for the brightest stars to 0.25 magnitudes for the faintest.

A new, highly accurate, astrometric star catalog is available from the U.S. Naval Observatory, covering ≈ 80% of the Southern Hemisphere to R ≈ 16m. The observing program is ongoing and full-sky coverage is expected by 2003. The final catalog will be used as the input catalog for the FAME space mission.

A new optical realization of the ICRS is built upon the data of the Hipparcos catalog and 60 observational and compiled astrometric ground-based catalogs of the 20th century reduced to the Hipparcos system. It is in print as two catalogs. The ‘Proper Motions of Fundamental Stars’ catalog (PMFS) contains proper motions of 4638 stars common to the Hipparcos and Basic FK5/FK5 Extension improved to the level of 0.5 mas. 1800 of the stars are the photocenters of stellar systems unresolved in the astrometric observations. The duplicity parameters of those systems were collected in the ‘Astrometric Binaries’ catalog (ABC). The photocenter nonlinear motions of several hundred of those astrometric binaries were detected and separated from their proper motions. The periods and amplitudes of several astrometric binaries are discussed in detail. This research will be continued for all Hipparcos stars observed from the ground in the 20th century. The web site is http://www.geocities.com/georgegontcharov/

The statistical analysis of the O – C and the standard errors for the astrometric RRS2 standard stars, and the analysis of the standard errors for the intermediate PIRS stars for 75 extragalactic radiosource fields are presented. This study was performed at Bucharest Observatory.

The current optical realization of the International Celestial Reference System is the Hipparcos main catalog in its entirety. This paper demonstrates that a subset of the main catalog stars — specifically those marked as double (“C” stars), acceleration solution (“G”), variability induced movers (“V”), and stochastic solution stars (“X”) — is of much lower quality than the majority of the catalog. Stars marked as suspected doubles (“S”) were investigated and found to show few problems. The authors have put forth a resolution (Resolution 2, this conference) to remove the C, G, V, and X stars, in addition to the Hipparcos orbit stars (“O stars”), from the optical realization of the reference frame.

We present an intensive analysis of the FK5 proper-motion system via the two large astrographic catalogs, the PPM and ACRS catalogs, compared with the Hipparcos proper motions. Regional, magnitude, and color-dependent systematic errors in the PPM and ACRS proper motions are found, and exhibit similar tendencies for both catalogs. The term of the global rotation between the FK5 and Hipparcos proper-motion systems cannot be explained by the constant of the FK5 precessional correction, which is given by the VLBI and LLR observations. Comparing the Hipparcos proper motions with those of the SPM 2.0 Catalog, which provides absolute proper motions of objects measured directly relative to external galaxies, we found neither strong systematic nor large regional errors between the two systems.

From Hipparcos proper motions, we have analyzed the local kinematics of the Milky Way from the young 0-B5 stars, classical Cepheids, and the late-type K-M giants. Comparing the derived results, the kinematical discrepancies between each subset of stars have been analyzed, and the possible reasons for the kinematical differences have been discussed.

Lunar occultations were analyzed to detect errors in the proper motion system of the Hipparcos frame. The analysis suggests much larger errors in the rotation of the Hipparcos reference frame than those explained by the Hipparcos team, which were determined from the direct comparison between the Hipparcos and FK5 catalogs.

Planetary and lunar ephemerides are no longer used for the determination of inertial space. Instead, the new fundamental reference frame, the ICRF, is inherently less susceptible to extraneous, non-inertial rotations than would be a dynamical reference frame determined by the ephemerides. Consequently, the ephemerides are now adjusted onto the ICRF, and they are fit to two modern, accurate observational data types: ranging (radar, lunar laser, spacecraft) and VLBI (of spacecraft near planets).

The uncertainties remaining in the inner planet ephemerides are on the order of 1 kilometer, both in relative positions between the bodies and in the orientation of the inner system as a whole. The predictive capabilities of the inner planet ephemerides are limited by the uncertainties in the masses of many asteroids. For this reason, future improvements to the ephemerides must await determinations of many asteroid masses. Until then, it will be necessary to constantly update the ephemerides with a continuous supply of observational data.