In particular, I've found the following predictions (http://astro.estec.esa.nl/GAIA/Info_she ... opics.html):
- 500.000 quasars with photometric redshifts (?), 2000 expected to be lensed...
- 30% of all SN Ia within 630 Mpc: very local, but you could really look into "standard candleness" problem
- Microlensing: don't have time to go into it, but in my understanding they hope to break some degeneracies by having nost just the light curve, but also the astrometry with idiotic accuracy and thus measured the mass of the lens
- lesing: the astrometry should be accurate enough to see the light bending by the solar system, so you need a full GR to do stuff properly. Tests of GR, maybe?
- Not explicitly said there, but could you do anything useful with photometric redshifts for objects brighter than -20? Something like very shallow but full sky galaxy survey? (don't have feeling for that and need to sit down to do calculation of how this translated to redshift).
- They do seem to have a spectrograph, but I am told its useless for cosmology, because it has very limited lambda range.