Previous Article
Next Article
Table of Contents

Astrometry for Hydra (1Dec92)
(from KPNO, NOAO Newsletter No. 32, 1 December 1992)

For years we optical spectroscopists have been able to get by just fine
with crummy coordinates, due, in part, to the ease of identification of
objects on a slit viewing TV, combined with the fact that most
telescopes did not point all that well anyway. However, as our
telescopes have begun to perform better and better, and as we have
continued to observe fainter and fainter objects, most of us have found
that our observing efficiency is considerably enhanced if we arrive at
the telescope with coordinates good to an arcsec or two. Now, with the
advent of multi-fiber spectrometers, we suddenly find ourselves in a
different regime, where an error of "an arcsec or two" means that
virtually no light will go down a fiber.

Users of Hydra must come prepared with coordinates good to <  0.5 arcsec
over the entire 45 arcmin field if they are to have any hope of
success, and furthermore the astrometry of their program objects must
be on the same "system" as the brighter stars used for tweaking up the
telescope alignment (the so-called "FOPS" stars). In good seeing an
error of 0.5 arcsec will result in negligible light-losses; an error of
1.0 arcsec, however, results in losing 50-75% of the light with our 2.0
arcsec fibers, depending upon the seeing. Exact details can be
extracted from Figures 1 and 2 in the paper by Donnelly et al. (1989,
PASP, 101, 1046). The physical positioning of the Hydra fibers is
believed to be accurate to of order 20 um (0.15 arcsec).

There are a number of resources available through NOAO that can help
Hydra users determine good coordinates.  We list these below.

1) Guide Star Catalog. If the objects you plan to observe with Hydra
are bright (V <  15-16), stellar, and relatively uncrowded, chances are
good that they will be in the Space Telescope Guide Star Catalog (GSC).
If you have the approximate positions, then the Fortran routine FINDER
can be used to search the GSC, (which we keep on-line on two CD-ROMS).
Care must be taken to select coordinates that all come from a single
"plate," as it is well known that the positions of GSC stars that occur
on multiple plates have coordinates that are typically offset by 1-2
arcsec. Aside from this concern, we have found that the GSC provides
coordinates which have excellent internal consistency (< = 0.2 arcsec). In
addition, since their coordinates were determined from recent (circa
1985) plates, corrections for proper motion are likely negligible.

2) POSS plates and the Grant Machine.  In the downtown Kitt Peak plate
vault, we have glass copies of the old (circa 1952) POSS. Positions of
objects down to a stellar magnitude of 21 can be readily measured on
these using the 2-axis Grant machine and reduced using FINDER/ASTRO
routines. However, because the epoch of the plate material is 40 years
old, great care must be taken to assure that proper motion for your
FOPS stars and/or program objects is either explicitly accounted for,
or is demonstrably negligible. Assuming that one is measuring "faint,
far away" things, one could use stars from the SAO catalog as the
reference system; to use these same stars as FOPS stars, however, you
will need to explicitly correct the catalogued positions using the
proper motions listed in the SAO catalog when you construct the Hydra
coordinate file. Unfortunately, the proper motions listed in the SAO
catalog are of variable quality, and care must be taken to select stars
whose listed errors in proper motion are small.  Alternatively, one can
use the GSC as the reference standards, and simply keep only those
stars whose residuals are small in the solution; these stars must have
low proper motion (19851952). Again, care must be taken to select only
GSC stars whose coordinates come from a single plate.

3) CCD frames and IRAF's "finder/tfinder" routines. If you have
selected your objects from wide-field CCD frames, then you can use this
material directly for determining excellent coordinates. To aid in
this, Rob Seaman has provided a set of routines in the "nlocal" package
"finder." These routines will allow you to search the Guide Star
Catalog for stars that are on your CCD frames, and display your image
overlaying the predicted location of the GSC stars it finds.
Interactive cursor options allow you to shift and find "astrometric
quality" x and y centers for these reference stars on your frame. Good
x and y centers for your program objects can be found using any of a
variety of routines within IRAF, that range in complexity from
positioning a cursor on a star and striking a key to obtaining centers
with psf-fitting in "daophot." Once you have good x and y centers for
your reference stars and program objects, the AAT/STARLINK "astrom"
routine is then used to find the six-coefficient plate solution, and
the coordinates of the program stars are then output directly in a
format that is needed for the Hydra assignment program. As long as one
restricts oneself to GSC coordinates determined for a single plate,
solutions are typically good to < = 0.2 arcsec RMS.  (The software is
designed to make this easy.) Because the "tfinder" routines are
considered a prototype, and because they require access to the GSC
CD-ROMS, these routines are not generally exported, although Rob Seaman
has successfully transported them elsewhere; first time users should
plan on using them in Tucson. Potential Hydra users are reminded that
the wide fields covered by 2048 x 2048 CCDs on the KPNO 0.9-m and
Burrell Schmidt telescopes are very useful for isolating samples of
objects and performing astrometry.

4) Digitized sky survey images. In a trial agreement between NOAO and
Space Telescope Science Institute, STScI has agreed to provide the
digitized scans of the "Quick V" (1985) Palomar Schmidt survey used in
producing the Guide Star Catalog. These scans contain stars as faint as
V = 19 (i.e., several magnitudes fainter than the GSC itself), and come
with an accurate "plate solution" as part of the header information.
Routines in STSDAS (usually distributed with IRAF) can then be used to
take x and y positions and output accurate celestial coordinates.
Measurements on two test fields provided by STScI have yielded good
results. The advantage to using this material is that the astronomer
can perform his/her astrometry at home, rather than traveling to NOAO,
and since the "Quick V" plate material is of recent vintage, proper
motions are usually immaterial. In addition, STScI has agreed to
include the southern SRC survey and the old POSS E plates, if these are
needed instead. The field size is limited to 1 degree per side. Tod
Lauer has agreed to coordinate requests to Space Telescope for this
material in support of Hydra runs; contact him directly
(tlauer@noao.edu). Requests should specify field center coordinates
(including equinox), field size (limited to 1 degree per side) and what
plate material (Quick V, POSS E, or SRC).  Please contact Tod by 15
January 1993 with this information if you wish to obtain fields for any
time in the spring semester.

5) Mix and match. If you have a small-field CCD image for which you
need accurate positions, it may be necessary to measure "secondary
reference stars" using either (1) or (4), and then use these as the
basis for computing the "plate solution" for your frame. This can be
done using either "astrom" or "astro", but will doubtless require a
good deal of hand editing.

In order to make use of any of the NOAO facilities significantly in
advance of an observing run, you should write to David De Young.
Additional advice can be obtained from the Astrometry Subdivision of
the Hydra crew (pmassey@noao.edu, tarmandroff@noao.edu).

Phil Massey

Previous Article
Next Article
Table of Contents