dealing with obscuration

[Stupendous Man <richmond@a188-l009.rit.edu>]

>  ... if I try to look at too large a declination, that the V camera 
>  sees less light in the NE corner of the image.  (It is looking at the roof) 
>  As I understand it, Michael is doing aperture photometry.  That is, he is 
>  comparing the brightness of a star to a dark ring around it.  Now in this 
>  case, both the star and the ring get fainter.  But if it is then compared 
>  against the brightness of all the catalog stars in the whole field, it will 
>  seem too dim.  If it is only compared to the catalog stars that are nearby, 
>  then one might get a better measurement.  

  Back in the Old Days, astronomers using photographic plates suffered
from non-uniform response across the plate (due to variations in emulsion
thickness, I think).  One solution was to break (conceptually) the 
plate up into sub-plates; say, in a 5x5 grid, or 10x10 grid.  With a
sufficient density of catalog stars, one can perform independent
photometric solutions in each sub-grid.  That way, target stars in 
an obscured region are compared to catalog stars in the same
obscured region.

  One can improve this method by keeping track of the differences
between the zero-points in different sub-plates, and then forcing
the zero-point differences between each sub-plate to be the same
over the course of a night.  In other words, average together
the results from 50 images over a night, and notice that the upper-left
corner always shows stars to be about 0.10 mag fainter than the center ...
THEN, go back and re-reduce the data all at once, using the fact
that stars in the upper-left corner are 0.10 mag fainter than those
in the center.  One can iterate, combine data from different nights,
etc.

  Another way to deal with this problem is to make a map of the
differences between magnitudes of stars in a catalog, and raw magnitudes
of stars in a frame, as a function of position in the frame.  
After averaging the results from tens or hundreds of images,
the map can be used to correct for the obscuration.

  ANOTHER way to do the (almost) same thing is to look at a particular
field of stars at a number of different pointings: move the 
telescopes by 1/4 or 1/5 of a field between images, so that
stars appear all over the image.  One can use the repeated
measurements of the stars to make a map of sensitivity across
the field.  The trouble with this particular method, in Tom's
case, is that if he moves the telescope, he may cause the
obscuring roof to move its position in the field.

  So there are plenty of ways to deal with the problem.
It just takes some extra work ...  Some of this can be done
post-facto, without having to re-reduce all the raw image data.
  
                                  Michael