Welch-Stetson technique, Mark IV plans, and software

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

  Greetings to Doug Welch!  We've discussed your method of finding
variable stars quite a bit in the past, actually.  Check out some
of the messages from January and February, 1998, in the E-mail archive.
I tried to summarize the method very briefly in 

     http://a188-l009.rit.edu/tass/mailarchive/1998-02/msg00070.html

but I hope you'll correct me if I made a mistake.

  As for the "target zones", Tom has already stated that each Mark IV
user will have his own plan.  As far as I know, we don't have an
overall strategy.  My own plan is to do two programs:

       a) Repeated coverage (twice an hour?) of a small number of fields,
              spread out in RA and probably north of the celestial equator.
              I want to measure each field a few hundred times in order
              to check for systematics in the photometry, test the limits
              of the Mark IV, look for variable stars, _maybe_ look for
              objects which move through the fields (but not follow them up).
              I haven't thought about how to choose the particular fields;
              anyone have ideas?
            
       b) Follow-up photometry of ephemeral sources -- mostly cataclysmic
              variables in outburst.   I'll watch VSNET for announcements
              of targets, pick a few, and follow them for several weeks 
              each.  I figure this will be an easy way to bring Mark IV
              data to the astronomical community; it may lead to a few
              brief data papers in IBVS or (with a lot more work) PASP.
              
  I will be watching Arne's progress very closely.  I believe that his plan
is to do a photometric survey of large portions of the northern sky
down to around magnitude 13 (?) in V and I.  As soon as he signs off on
photometry in some area, I'll use his work to calibrate my data.  
Until then, I plan to calibrate my V and I against Tycho Vt and
a pseudo passband: "Tycho It".  The TASS Mark III paper in PASP describes
a method for estimating an I-band-like magnitude from a subset of 
Tycho stars.  Since the Tycho catalog provides such a good match to
the Mark IV data, and since I'll be using it for astrometric calibration,
it seems to me reasonable to use it as a photometric reference, too.

  Another possibility is to schedule enough observations of fields with
Landolt (or other) calibrations in the Johnson-Cousins system --
over a wide enough range of airmass -- to perform a photometric calibration
for each night.  The extra constraints this scheme places on 
field selection and scheduling, plus the much smaller number of reference
stars, lead me to prefer the Tycho scheme.  But I welcome calm,
reasoned discussion on this point :-)

  Finally, a word about software:  I'm putting the finishing touches
on a first cut of a pipeline to reduce Mark IV data.  It's a bit
of a chimera, consisting of several pieces:

       - the XVista package of image reduction programs -- C
                 (create darks, flats, subtract darks, divide by flats,
                  find stars, measure positions and instrumental mags)

       - the "match" program -- C
                 (match detected stars to reference catalog, 
                  calculate astrometric transformations)

       - the "collate" and "photom" programs -- C
                 (match detections in simultaneous V and I exposures,
                  perform photometric solution, apply to raw magnitudes)

       - a pipeline and associated parameter files -- TCL
                 (run raw image files through the various steps of 
                  reduction in the proper order, allow user to fiddle
                  with parameters without recompiling code)

  All of the C programs work properly, though some may require fiddling
with input parameters to succeed on different nights; I'm very slowly
improving the XVista package, but the current version does enough for
a start.  The pipeline is not quite finished -- it runs data through
everything up to photometric calibration, but needs several days of work
in that area.  I am well aware that the most important components
are missing:

         - external documentation
         - tutorial
         - installation instructions

  I have used the pipeline to run through Mark IV data on Disk 15.
As a rough guide, I find that a night's worth of data (240 images,
for purposes of discussion) will take roughly four hours to run through
the pipeline on my current computer (Dell Optiplex GX110, with 
a 667-MHz P3 and 512 MB of RAM).  

  The end result of the pipeline is a set of plain ASCII text files,
one per field, containing columns of information like this:

    id RA Dec JulianDate  airmass    V vmag verr vflag  I imag ierr iflag

where the "id" is some meaningless internal ID number, the RA and Dec
are J2000 coordinates, the JulianDate is the time of midexposure,
airmass is airmass, and each passband of data contains a calibrated
magnitude, an uncertainty, and a quality flag.  Only stars which
are detected in both V-band and I-band images appear in the pipeline output.

  It will be another big job to stuff these data files into a database.
Under some circumstances, one might consider stuffing some other information
into the database; for example, one might want to know about stars which
appear only in V-band or only in I-band.  Or one might want to stuff
some of the intermediate files, with information about the FWHM or
peak pixel value for each star in the corrected images.  

  After I finish the last bits of the TCL pipeline, I'll run all 
of the data on Disk 15 through it and make the results available 
on the TASS home page.  I'll then try to document the code and wrap
it up into a small number of packages for distribution.  Everyone 
is free to use it ... or ignore it, of course :-)

                                              Michael Richmond