Understanding the details of the Galaxy in which we live is one of the great intellectual challenges embraced by modern science. Our Galaxy contains a complex mix of stars, planets, interstellar gas and dust, radiation, and the ubiquitous dark matter. These components are widely distributed in age (reflecting their birth rate), in space (reflecting their birth places and subsequent motions), on orbits (determined by the gravitational force generated by their own mass), and with complex distributions of chemical element abundances (determined by the past history of star formation and gas accretion).
Astrophysics has now developed the tools to measure these distributions in space, kinematics, and chemical abundance, and to interpret the distribution functions to map, and to understand, the formation, structure, evolution, and future of our entire Galaxy. This potential understanding is also of profound significance for quantitative studies of the high-redshift Universe: a well-studied nearby template galaxy would underpin the analysis of unresolved galaxies with other facilities, and at other wavelengths.
Understanding the structure and evolution of the Galaxy requires three complementary observational approaches: (i) a census of the contents of a large, representative, part of the Galaxy; (ii) quantification of the present spatial structure, from distances; (iii) knowledge of the three-dimensional space motions, to determine the gravitational field and the stellar orbits. Astrometric measurements uniquely provide model-independent distances and transverse kinematics, and form the basis of the cosmic distance scale. Complementary radial velocity and photometric information are required to complete the kinematic and astrophysical information about the individual objects observed.
Photometry, with appropriate astrometric and astrophysical
calibration, gives a knowledge of extinction, and hence, combined with
astrometry, provides intrinsic luminosities, spatial distribution
functions, and stellar chemical abundance and age information. Radial
velocities complete the kinematic triad, allowing determination of
dynamical motions, gravitational forces, and the distribution of
invisible mass. The GAIA mission will provide all this information.
Even before the end of the Hipparcos mission, a proposal for an
ambitious follow-on space astrometry experiment was submitted to ESA
(Roemer: Høg 1993; Lindegren et al. 1993a; Høg & Lindegren 1994). The idea of using
CCDs as a modulation detector behind a grid (Høg & Lindegren 1993), similar to
Hipparcos, was replaced by the more powerful option adopted for
Roemer (Høg 1993) where CCDs measure the direct stellar images in
time-delayed integration (TDI) mode in the scanning satellite. A more
ambitious interferometric mission, GAIA, was proposed and subsequently
recommended as a cornerstone mission of the ESA science programme by
the Horizon 2000+ Survey Committee in 1994 (Battrick 1994). The GAIA
proposal demonstrated that accuracies of 10 
as at 15 mag were
achievable using a small interferometer (Lindegren et al. 1993b; Lindegren & Perryman 1996).
The European scientific community and ESA have now completed a detailed study of the science case and instrument design, identifying a number of further improvements, including reverting to full-aperture telescopes (Høg 1995a; Høg 1995b). The results demonstrate that unique and fundamental advances in astrophysics are technically achievable on the proposed time-scales, and within a budget profile consistent with the current ESA cornerstone mission financial envelope.
GAIA will be a continuously scanning spacecraft, accurately measuring one-dimensional coordinates along great circles in two simultaneous fields of view, separated by a well-known angle. The payload utilises a large but feasible CCD focal plane assembly, passive thermal control, natural short-term instrument stability due to the Sun shield and the selected orbit, and a robust payload design. The telescopes are of moderate size, with no specific manufacturing complexity. The system fits within a dual-launch Ariane 5 configuration, without deployment of any payload elements. The study identifies a "Lissajous'' orbit at L2 as the preferred operational orbit, from where about 1 Mbit of data per second is returned to the single ground station throughout the 5-year mission. A comprehensive accuracy assessment has validated the proposed payload and the subsequent data reduction.
This paper provides a summary of the key features of the improved GAIA design, and the resulting scientific case, evaluated during the recent study phase (ESA 2000; see also http://astro.estec.esa.nl/GAIA). A comparison between the scientific goals of GAIA, and other post-Hipparcos space astrometry missions, is given in Sect. 8.
Copyright ESO 2001