[daip] Coordinate Systems question

[Eric Greisen]

Shami Chatterjee writes:

 > I'm working on an archival VLA data set (A array, C band, 0.3 asec
 > resolution) which was taken in B1950 coordinates. And the aips eposwtch
 > verb warns that it is "of adequate accuracy for normal imaging but should
 > not be used in astrometric circumstances."
 > 
 > In order to convert coordinates of detected point sources back to J2000, I
 > need to know: is the implied reference frame FK4 for B1950, FK5 for J2000?
 > 
 > And what do you recommend for the conversion?

The ordinary imaging in AIPS is not done with the precision implied
by your question.  The u,v,w are measured at the moment of observation
but are given to AIPS by the on-line system rotated to a specified
equinox.  This implies that the only affect of precession is a field
center shift and rotation.  EPOSW continues that assumption and simply
computes that shift and rotation to go from 1950 <-> 2000.  They are
B1950 (implication FK4 but I would not be certain) and J2000
(implication and I'm fairly certain FK5).  Real astrometric work would
require you to undo the rotation done by the on-line system and then
do imaging in coordinates of date.  Coordinates of sources in that
image would then be separately converted by proper astrometric
formulae to standard dates.  EPOSW is implimented in $AIPSUB/AU7.FOR
and calls routines STEPCH, CRDSET, CRDTRN, and B2JPOS (all in
$APLSUB).  You have these routines and can examine them.  No one in
the AIPS group is an astromotrist and we were simply given these
routines.

Eric