I’ve been reading a number of papers lately about GAIA, a future spaced-based mission planned for launch in 2012. The aim of GAIA is to make astrometric measurements of nearly one percent of the Milky Way’s stellar population, amounting to some 1 billion stars down to a visual magnitude of m=20. Following up on the success of Hipparcos, the GAIA mission will achieve an extraordinary degree of precision, with stellar positions measured to an accuracy of about 10-20 µ arc seconds in conjunction with radial velocity and parallax measurements. Consider, the moon with a diameter of about 1700 km subtends an apparent angle of 30 arc minutes, so 10 μ asec is like pinpointing the position of a marble-sized object (1-2 cm) on the surface of the moon! In comparison to Hipparcos, GAIA will be about 50 times more accurate and survey some 10,000 times more objects.
Artist Rendition of GAIA. Source:ESA
Astrometry has gotten short shrift over the years. After all, it’s not going to produce the kinds of mind-blowing images that Hubble pumps out, and that have made the HST a cultural icon. And efforts to catalog huge swaths of the Milky Way to such precision would seem to be an exercise in obsessive compulsion. But by precisely measuring the position, motion, and distance of billions of stars, the GAIA mission will enable fundamentally new discoveries including:
- New insights into the structure, dynamics, and evolution of the Milky Way.
- A veritable factory for the discovery of tens of thousands of new exoplanets (at least!) and a half-million quasars.
- An improved understanding of the nature and distribution of dark matter in our galaxy
- Cataloging pulsating variables as the result of repeated measurements over it’s five year mission, enabling new statistical studies of variable stars.
The opportunities for data mining and in silico discovery abound. In a status update from a couple of years back, S. Jordan of the Astronomisches Rechen-Institut, Zentrum für Astronomie, Heidelberg, German (see references below) writes of the computational challenges:
The total raw data volume is estimated to be of the order of 100TB, the total amount of processed and archived data is of the order of 1 PB. Current estimated [sic] for the computational volume is 1.5.1021 floating-point operations, but this may actually be a lower limit and will probably increase due to the correction tasks for the radiation-induced CCD damages. If the processing of one star (with typically 1000 measurement per star) would need 1 second, 30 years of the data analysis would be needed. However, with massively distributed computation and the faster computers of the time of Gaia’s data analysis, this task will be feasible.
And it is the end-product of this data analysis that promises to be stunning – producing in effect a kind of movie of the Milky Way. So while GAIA may not immediately return the kinds of images that HST has become famous for, it’s data will be no less beautiful, and its value to astronomical research no less important.
References:
P.T. de Zeeuw (2002). Into the future with GAIA.
S. Jordan (2008). The Gaia Project – technique, performance, and status.
A. Brown (2008). Getting ready for the micro-arcsecond era.
M. A. Strauss et al. (2009). Wide-Field Astronomical Surveys in the Next Decade: A White Paper for Astro2010.
Posted by John Rachlin 
