Re: GSC and DS23

[aah@nofs.navy.mil]

patrick wrote:
>I have tried to match all the stars from DS23 with GSC
>1 stars.  5.74% of the stars did not have a
>counterpart within 1'.
  It is not likely that a GSC star more than a few arcsec
away from a TASS object is physically associated.  The astrometry
from Michael R.'s pipeline is really pretty good.  As you
increase such a search radius, the chances of coincidence
increase.  For example, if you average the GSC stars over the
entire sky, there are about 350 per square degree, or about
0.1 per square arcmin.  You need to do, at minimum, matching
with magnitude as well as just position.

>Would it be worthwhile to give the distribution per
>declination zone, to look for systematic errors near
>the edges of the frames ?
  What you want is error vs x,y of the CCD frame.  That
will tell you about any residuals as a function of position
on the CCD.  I would use Tycho2 stars rather than GSC for
this exercise as you want high positional accuracy of the
reference catalog; GSC has its own internal systematic
problems.  A secondary parameter is error vs. magnitude
vs x,y position.
Arne