Re: Working on Data

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

  Tom wrote:

> I have been working on the data.  That is why I have been quiet.

  I, too, have been working on the data.  I agree with Tom that
one can identify "bad" nights and portions of nights.  I also
agree that doing so does not solve all problems.

  I've been concentrating on the scatter among measurements of
bright stars.  It should be much smaller than we actually measure.
Could it be due to errors as a function of position on the chip
(which one could loosely term "flatfielding")?  

  I think that the answer is "yes" in V-band (which is where the 
scatter at the bright end is most annoying), and "no" in I-band.
Take a look at these pictures of residuals as a function of position
on the chip (for a very particular subset of all star -- details
in an upcoming Tech Note):

     http://spiff.rit.edu/richmond/temp/gamma_V.gif
     http://spiff.rit.edu/richmond/temp/gamma_I.gif

  Note the clear systematic pattern in V-band residuals, and lack
thereof in I-band residuals.

  My plan is to define a correction to V-band as a function of position
on the chip, apply it to stars, and re-calculate mean and stdev
of measured magnitudes.  If I'm right, it might reduce the scatter
at the bright end of the V-band.

  After I do _that_, I'll look into moving the Mark IV measurements
from the Tycho V,pseudo-I system onto the standard Johnson-Cousins V,I
scale.  Arne's extra calibration data should help.

  I'm thinking that _eventually_, I'll be able to sift through the
database, select a subset of all the measurements, apply a set of
corrections, and produce a "nice" catalog.  It will be analogous
to the Mark III dataset and the "tenxcat" catalog.  That's my
current goal: creating a list of stars with decent (V,I) magnitudes
on the standard scale over as large an area as I can.

  This will probably _not_ be especially relevant to those who
want to look through the Mark IV data for small-amplitude variable stars. 
Large-amplitude variables are already obvious, as a number of
people have shown -- just look at the growing lists on the Wiki.
I hope that some other people will consider ways to massage
the data for that purpose.  

                                         Michael