Re: Astrometry errors

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

  Doug Welch asked:

> What is believed to be the main source of astrometry errors
> in the TASS scans (besides quality of focus)? Is there a 
> secular drift in a single coordinate or is it a focal length
> scale change (or both?)

  If you are asking about the Mark III scans, my answer would
be "probably poor focus".  Wait -- you want the NEXT largest
source of errors?  Gosh, I really don't know.  I haven't studied
the astrometry of individual Mark III scans, and I don't know
if anyone else has, either.

  We took the individual scans, measured (row, col) positions,
compared to Tycho positions, derived an coordinate transform,
and applied it to calculate (RA, Dec) values for each star on each
image ... BUT the next step is the killer: we (basically) threw
away the (RA, Dec) from each image.  You see, our primary desire
was to merge together measurements of the same star at different
times, so all we did was compare an individual (RA, Dec) 
measurement against a catalog of stars; if it was within some
critical distance, we said "Ah, this is the same star", and
merged its photometry with the other photometry for that star.

  If you're interested in errors in astrometry, you can still
retrieve the (RA, Dec) values for each detection.  But it will be ugly -- 
you will have to connect to the database directly, and query the 
appropriate tables:

            observ_3v01            for individual V-band measurements
            observ_3r01             "      "      R   "       "
            observ_3i01             "      "      I   "       "

  For example, if I connect to the "tassdb" database on stupendous.rit.edu
(as described in Tech Note 56) and make the query

        select * from observ_3v01 where (tass_id = 6);

I get the following:

tass_id|sourceid|julian_date|  ra|      dec|mag_inst|mag_erro|ccd_no|site_id
-------+--------+-----------+----+---------+--------+--------+------+-------
      6|    7186|  742664969|3600|-13868639|   14317|     445|     1|B      
      6|    6260|  713737420|6840|-13859639|   13997|     379|     1|B      
      6|    6557|  717726369|6480|-13857119|   14225|     355|     1|B      
      6|    6764| 1083717459|9720|-13858919|   13997|     538|     1|B      
      6|    3708| 1018898149|1440|-13866479|   13545|     398|     1|B      
(5 rows)

  So, there are some actual (RA, Dec) measurements.  Note that the 
units are bizarre (as described in Tech Note 56): RA and Dec are
both in milliarcsec, away from RA = 0 and Dec = 0, respectively.
The Julian Dates have wierd units, too.

  Doug, your question was quite specific:

> Is there a 
> secular drift in a single coordinate or is it a focal length
> scale change (or both?)

  Is there some reason you believe one or the other of these factors
to a factor in the Mark III data?

  Now, on the other hand, if you are talking about the Mark IV images --
which are stare-mode frames, not scans -- then the main source of
astrometry errors is distortion in the PSF and in plate scale
near the corners of the chip.  I'm not sure which is more important;
perhaps Andrew Bennett or Jure Skvarc could give us an informed opinion.


                                              Michael Richmond