By imaging in 3 near-IR bands it will be posible to construct a 3-D Atlas of the galaxy. Distances will be determined by dereddening in a JHK two-colour diagram and kinematically via follow up in the radio waveband. It will not be possible to reliably determine the distance to every star, since many will not have J band detections. However O, B and A type stars all have similar and well known IR colours (Planckian spectra) so it will be possible to deredden these sources to a colour-magnitude main sequence even with just H and K band detections. Uncertainty in derived luminosity may be ~10% due to the width of (or calibration errors in) the main sequence but this leads to only a 5% uncertainty in distance. We emphasise that the near-IR dereddening law is effectively uniform throughout the galaxy (Clayton, Cardelli & Mathis 1989) despite the spatial variation seen in the optical; this is because dust grain radii are much smaller than the wavelength of observation for wavelengths > 1 micron, except in extremely dense environments such as accretion disks.
By mapping out all the bright stars throughout a large fraction of the Milky Way, the GPS Atlas will provide an unprecedented tool for studying the three-dimensional structure of the Galactic disk. 21cm studies have shown that the outer parts of the Milky Way's gas disk are strongly warped away from the Galactic plane (Burton & te Lintel Hekkert 1986), and it would be very interesting to ascertain whether the stellar distribution is similarly warped. If there turns out to be a difference between the two, then it will be hard to explain the warp as a purely gravitational phenomenon, and gas dynamics must be viewed as an important factor.
Carney & Seitzer (1993) found tentative evidence for a stellar warp using optical data, but were hampered by reddening, and were unable to quantify its morphology. Porcel et al. (1997) analyzed the diffuse K-band emission in the Galactic plane (as recorded by the COBE satellite), and concluded that these data were not consistent with the warp in the gas unless the stellar distribution is truncated at ~15 kpc, or the light is dominated by local features in the stellar distribution. However, lacking distance information, they were not able to discriminate between these interesting hypotheses.
With the GPS observations providing positions and distances to stars throughout a large part of the Milky Way, it will be possible to see directly where (if at all) the disk is truncated, and whether the stars's locations deviate from the Galactic Plane in the same manner as the gas. The sky coverage of the survey will sample the warp structure very effectively. In the integrated K-band flux, Porcel et al.'s (1997) model for a non-truncated warped disk shows a maximum deviation in the peak surface brightness of ~2o above the plane at l ~ 100o, while the GPS observing window 150o < l < 230o would trace such a warp from approximately the Galactic mid-plane down to the maximum deviation below the plane of ~ -2o. By introducing the third dimension of approximate distances, the stellar distribution will be fully mapped out in these regions, finally giving a definitive picture of the stellar warp and its origins.