Dynamical theories of the Earth’s rotation like SMART97 (Bretagnon et al., 1998) are to be considered in a DGRS (dynamically nonrotating geocentric reference system) (Brumberg et al., 1996). Such a theory gives the explicit expressions in terms of TCG (Geocentric Coordinate Time) of three Euler angles relating a DGRS to the ITRS (International Terrestrial Reference System). These angular quantities together with their TCG derivatives enable one to get all Earth’s rotation parameters. At the same time, the analysis of observations result in the values for slightly different angles and their TCG derivatives characterizing the relationship between the ITRS and a KGRS (kinematically nonrotating geocentric reference system). The differences between these two sets of six quantities represent kinematical relativistic corrections (due to geodesic precession, geodesic nutation and luni-planetary terms). The paper presents these differences computed by means of the VSOP87 series (Bretagnon and Francou, 1988). In particular, in analysing observations at the microarcsecond level these expressions will permit an experimental check of geodesic precession in a more direct manner than it is done nowadays (Bertotti et al., 1987).

Attaining the limit of sub-microarcsecond optical resolution will completely revolutionize fundamental astrometry by merging it with relativistic gravitational physics. Beyond the sub-microarcsecond threshold, one will meet in the sky a new population of physical phenomena caused by primordial gravitational waves from the early universe and/or different localized astronomical sources, space-time topological defects, moving gravitational lenses, time variability of gravitational fields of the solar system and binary stars, and many others. Adequate physical interpretation of these yet undetectable sub-microarcsecond phenomena cannot be achieved on the ground of the “standard” post-Newtonian approach (PNA), which is valid only in the near-zone of astronomical objects having a time-dependent gravitational field. We describe a new, post-Minkowskian relativistic approach for modeling astrometric observations having sub-microarcsecond precision and briefly discuss the light-propagation effects caused by gravitational waves and other phenomena related to time-dependent gravitational fields. The domain of applicability of the PNA in relativistic space astrometry is outlined explicitly.

It is argued that the relativistic definitions of parallax, proper motion and radial velocity consistent with an accuracy of 1 μas should be considered only within a well-defined algorithm of relativistic reduction of observational data. Such an algorithm is formulated and the corresponding definitions of astrometric parameters are discussed.

A non-perturbative general relativistic approach to global astrometry was developed by de Felice et al. (1998) to handle satellite astrometry data in a genuine relativistic framework. In this contribution, the framework above has been further exploited to account for stellar motions and parallax. Because of the relevance that accurate knowledge (to 10−5 or better) of the relativistic parameter γ has to fundamental physics, a Parametrized Post-Newtonian (PPN) model has also been implemented, which allows the direct estimation of γ along with the astrometric parameters. These models have been tested on end-to-end simulations of the mission GAIA. The results show that, within the limitation of the simulation and the assumptions of the adopted model, measurements accurate to 100 μarcsec of large arcs among stars repeated over a few years can be modelled to establish a dense reference frame with a precision of a few tens of μarcseconds. Moreover, our experiments indicate that γ can be estimated to better than 10−6.

The accuracy of astrometric observations conducted via a space-borne optical interferometer orbiting the Earth is expected to approach a few microarcseconds. Data processing of such extremely high-precision measurements requires access to a rigorous relativistic model of light ray propagation developed in the framework of General Relativity. The data-processing of the space interferometric observations must rely upon the theory of general-relativistic transformations between the spacecraft, geocentric, and solar barycentric reference systems allowing unique and unambiguous interpretation of the stellar aberration and parallax effects. On the other hand, the algorithm must also include physically adequate treatment of the relativistic effect of light deflection caused by the spherically-symmetric (monopole-dependent) part of the gravitational field of the Sun and planets as well as the quadrupole- and spin-dependent counterparts of it. In some particular cases the gravitomagnetic field induced by the translational motion of the Sun and planets should be also taken into account for unambigious prediction of the light-ray deflection angle. In the present paper we describe the corresponding software program to take into account all classical (proper motion, parallax, etc.) and relativistic (aberration, deflection of light) effects up to the microarcsecond threshold and demonstrate, using numerical simulations, how observations of stars and/or quasars conducted on board a space optical interferometer orbiting the Earth can be processed and disentangled. For numerical simulations the spacecraft orbital parameters and the telescope optical-system-characteristics have been taken to be similar to those in the Hipparcos mission. The performed numerical data analysis verifies that the relativistic algorithm chosen for data processing is convergent and can be used in practice to determine astronomical coordinates and proper motions of stars (quasars) with the required microarcsecond precision. Results shown in the paper have been obtained with a rather small number of stars (a few thousand). Simulations based on a much larger number of stars, e.g., from the Guide Star Catalogue used to model original observations will give more complete information about potential abilities of the space astrometric missions.

After recalling current performances of time and frequency standards and their use to establish time scales, their limitations in astronomical observations and studies are considered. For expressing the value of almost all quantities, the accuracy of the frequency standards is sufficient. However, an improved stability could find applications in observation techniques and the departure of realized time scales with respect to their theoretical counterparts is not negligible for some studies. The paper recalls the basic rule of quantity calculus and advocates its use in spite of its inadequacy to express difference of readings of time scales. In general, adherence to the metrological practice and the use of the SI may facilitate our work and its understanding by the non-astronomers.

Since the last presentation of SC-3 on numerical values of fundamental geodetic parameters at the IAU General Assembly at Kyoto in 1997 there were some conceptual as well as fundamental numerical changes. The four basic parameters of geodetic (ellipsoidal) reference systems (GRS) can no longer be considered as constant with time: J2, a, ω, and GM have to be replaced by clearly (±10−8 or better) specified mean values or have to be associated with a specific epoch or, in case of GM, with specific reference frames (a = semi-major axis of Earth ellipsoid, J2 = second degree zonal harmonic of geopotential, ω = spin of Earth rotation). In case of (a, J2 ....) associated tidal reductions must be specifically defined in view of particular applications and significant differences between different tidal reduction types. Or we may replace “a” by a quantity which is independent of tides like the geopotential at the geoid, W0, where, however, also temporal changes are now discussed. The official geodetic reference systems such as GRS 80 and WGS 84 (revised in 97-form) are also no longer truly representing reality; a new system GRS 2000 is desired. We are, meanwhile, able to define and determine tidal and non-tidal (secular, periodic, aperiodic) variatipns of some fundamental geodetic parameters. Others are under investigation. New precession and/or nutation formulas to be adopted by IAU in 2000 or later would imply, again, changes in geodetic parameters such as H = hydrostatic flattening. Those and related other consequences are considered.

The IAU set up the SOFA initiative at the 1994 General Assembly, to promulgate an authoritative set of fundamental-astronomy constants and algorithms. At the subsequent General Assembly, in 1997, the appointment of a Reviewing Board and the selection of a site for the SOFA Center (the outlet for SOFA products) were announced. The time since then has been spent developing the specifications for a basic collection of fundamental-astronomy algorithms. Progress has been slow, for a number of reasons: the task is difficult and requires great circumspection; a substantial toolkit of support routines must be adopted or developed before work on the astronomy routines themselves can begin; and it takes time for a geographically dispersed committee to reach a consensus on matters that involve value judgements, such as programming style. However, despite these formidable difficulties, significant progress has been made. A library of 52 vector/matrix routines has been developed, supporting a set of 25 astronomy algorithms. The Reviewing Board is about to start validating these algorithms, at which point they can be released through the SOFA Center.

The current definition of Coordinated Universal Time (UTC) is related to the unpredictable, variable rotation rate of the Earth. This is accomplished by irregular insertions of leap seconds, creating unpredictable discontinuities in UTC. With the increasing importance of a continuous, uniform time scale for users, it is appropriate to re-examine the current definition of this time scale. There are several possibilities to address this problem, and it is appropriate that the International Astronomical Union establish a working group to investigate the continuing need for leap seconds and possible changes in the definition of UTC.

The basic astrometric observable of the Space Interferometry Mission (SIM) instrument is the pathlength delay. This measurement is made by a combination of internal metrology measurements that determine the distance the starlight travels through the two arms of the interferometer and a measurement of the white light stellar fringe to find the point of equal pathlength. Because this operation requires a non-negligible integration time, the interferometer baseline vector is not stationary over this time period, as its absolute length and orientation are time-varying. This paper addresses how the time-varying baseline can be “regularized” so that it may act as a single baseline vector for multiple stars, as required for the solution of the astrometric equations.

First astrometric and photometric applications of the logarithmically scaled HDRC4 imager (developed by the Institut für Mikroelektronik Stuttgart, IMS, Germany) are presented. To test the HDRC4 imager at the telescope, a cooled camera had to be constructed by modifying a camera designed by IMS. First observations were performed with various telescopes at Hamburg Observatory, Germany. These showed the dynamic range to exceed 14 mag and the sensitivity to be comparable to that of the Kodak KAF 4200 CCD at a chip temperature of –35°C.

The Space Interferometry Mission (SIM) is a 10-m Michelson space-based optical interferometer designed for precision astrometry (4μas, 3μas/yr) with better accuracy hoped for over a narrow field of view. It is intended to search for planets and investigate a number of problems in Galactic and extra-galactic astronomy.

The accuracy and stability of SIM’s celestial reference frame is subject to degradation over the 5-year mission from the reflex motion induced by massive companions of the objects used to construct the celestial reference frame. We present the results of simulations which show the sensitivity of reference frame accuracy to companions as a function of mass and period. We assume that pre-launch ground surveys will eliminate all objects with RMS radial velocity > 10 m/s. We further assume that the standard astrometric parameters of position, parallax, and proper motion plus acceleration terms in right ascension and declination will be allowed to absorb reflex motion.

The future of space-based astrometric missions appears to be very promising. Three missions were described at IAU Colloquium 180. These are the Full-sky Astrometric Mapping Explorer (FAME), the Space Interferometry Mission (SIM) and the Global Astrometric Interferometer for Astrophysics (GAIA) missions. These missions will substantially improve the accuracy of global astrometric measurements made by the Hipparcos space mission by factors of 20 to 250. A brief summary of these projects is given.

Recent progress in the determinations of astronomical constants is reviewed. First is the latest estimation of the general relativistic scale constants, LC, LG , and LB (Irwin and Fukushima, 1999). By reestimating the uncertainty, the value of the first constant is given as LC = 1.480 826 867 4 × 10–8 ± 1.4 × 10–17. Also noted is the rigorous relation among these three, LB = LC + LG – LCLG . Based on the latest determination of the geoidal potential W 0 in the IAG 1999 Best Estimate of Geodetic Parmeters (Groten, 1999), LG and LB were reevaluated as LG = 6.969 290 09 × 10–10 ± 6 × 10–18 and LB = 1.550 519 767 3 × 10–8 ± 2.0 × 10–17. Since LG is roughly related to W 0, a proposal to fix its numerical value is presented in order to remove the geophysical ambiguity in its evaluation in the future. In that case, LG becomes a defining constant for the scale difference between the geocentric and terrestrial coordinate systems. While LC and LB remain as a primary and derived constant, respectively. Next is the correction to the current precession constant, Δp. The recent estimates of Δp based on Very Long Baseline Interferometry (VLBI) observation seem to converge to a value close to –0.30″/cy (Mathews et al., 2000; Petrov, 2000; Shirai and Fukushima, 2000; Vondrák and Ron, 2000). Unfortunately this is significantly different from –0.34″/cy, the latest value determined from the Lunar Laser Ranging (LLR) data (Chapront et al., 1999). The difference is roughly ten times larger than the sum of their formal uncertainties. Since the cause of this difference is not clear, we first arranged the best estimates based on VLBI and LLR techniques, respectively, then took a simple mean of these two best estimates, and recommend it as the current best estimate. The value derived is p = 5 028.78 ± 0.03 ″/cy. Similar estimates were given for some other quantities related to the precession formula; namely the correction to the obliquity rate of the IAU 1976 precession formula (Lieske et al., 1977), Δε 1 = (–0.024 5 ± 0.002 5) ″/cy, and the offsets of the Celestial Ephemeris Pole of the International Celestial Reference System, Δψ 0 sin ε 0 = (–17.5 ± 0.8) mas and Δε 0 = (–5.2 ± 0.4) mas. As a result, the obliquity of the ecliptic at the epoch J2000.0 was estimated as ε 0 = 23°26′21.″405 6 ± 0.″000 5. The draft IAU 2000 File of Current Best Estimates of astronomical constants, that is to replace the 1994 version (Standish, 1995) or maybe even the formal IAU 1976 System of Astronomical Constants (Duncombe et al., 1977), after discussion at the 24th General Assembly of the IAU is presented.