[AstroPy] astropy.coordinates vs kapteyn.celestial Coordinate Transformations

[Thomas Robitaille]

Hi Paul,

Thanks for the input - as an aside, do you have any insight into whether
the IAU is considering proposals such as:

http://www.aanda.org/articles/aa/pdf/2011/02/aa14961-10.pdf

to re-define the Galactic coordinate system relative to ICRS? Having
Galactic coordinates defined relative to FK4 at the moment is not ideal
and introduces a lot of confusion, so re-defining galactic coordinates
relative to ICRS would be a nice solution.

Cheers,
Tom


Paul Kuin wrote:
> I asked the chair of the IAU astrometry commission. Basically the
> current coordinate system is ICRS, and the galactic frame is defined
> with reference to that with the same accuracy. I think that is in a few
> milli arcsec accuracy. The SOFA implementation is to be consulted for
> actual algorithms, and I think this has been (or will be) implemented in
> astropy as the ERFA package.  
> 
> In summary, I would think the differences found by Joseph are way too
> large. It may be that the Kapteyn algorithms are also not compliant. 
> Test should be done against the results of the ERFA/SOFA software. 
> 
> Thats all the help I can give. 
> 
> Paul
> 
> On Sat, Nov 15, 2014 at 5:32 PM, Thomas Robitaille
> <thomas.robitaille at gmail.com <mailto:thomas.robitaille at gmail.com>> wrote:
> 
>     I've played around with this a bit more, and I found that if I set:
> 
>     _ngp_B1950 = FK4NoETerms(ra=192.25*u.degree, dec=27.4*u.degree)
> 
>     (not FK4) then I explicitly set:
> 
>     Galactic._ngp_J2000 = _lon0_B1950.transform_to(FK5)
> 
>     then if I set:
> 
>     Galactic._lon0_J2000 = Angle(122.931918539, u.degree)
> 
>     Then the astropy FK5 -> Galactic and FK5 -> FK4 -> Galactic
>     transformations agree within ~0.1mas, and our results are in agreement
>     with other codes. I'll open an issue on the astropy repo and we can
>     discuss the technical details there.
> 
>     Tom
> 
>     Erik Tollerud wrote:
>     > Hi Tom and Joseph,
>     >
>     > Originally, Galactic coordinates were defined to ~arcmin precision
>     > relative to FK4 (and it's not clear one way or another if this is
>     > supposed to include E-terms or not). Since then there have been
>     > recalibrations to FK5, although there's not an "official" version
>     as far
>     > as we could find (just an appendix in a paper).
>     >
>     > So I think what's happening here is that the "shortcut" FK5-> Galactic
>     > transform is being used, which gives somewhat different results
>     because
>     > it made different arbitrary choices. We could just turn that off if we
>     > want to always be sure to use the FK4 transformation.
>     >
>     > It's not really clear which is the "right" answer, though. I think our
>     > original thinking was that no one really uses Galactic for subarcsec
>     > precision, and there FK4 transformations are a lot slower, so this
>     > "shortcut" makes sense...
>     >
>     > Hello,
>     >
>     > I've been porting pyregion to use astropy instead of kapteyn, and
>     tests
>     > with coordinate system conversions are slightly off.
>     >
>     > I think I've narrowed down the problem to my expectation that this
>     > should be nearly zero:
>     >
>     > In [21]: from astropy.coordinates import SkyCoord
>     >
>     > In [22]: from kapteyn import celestial
>     >
>     > In [23]: a = SkyCoord('292.03306305555554d 1.7592747222222223d',
>     > frame='galactic').transform_to('fk5'); print(a)
>     > <SkyCoord (FK5: equinox=J2000.000): ra=171.158093022 deg,
>     > dec=-59.2630875829 deg>
>     >
>     > In [24]: celestial.sky2sky(celestial.galactic, celestial.fk5,
>     > [292.03306305555554], [1.7592747222222223])
>     > Out[24]: matrix([[ 171.15816386,  -59.26319319]])
>     >
>     > In [25]: SkyCoord('171.15816386d -59.26319319d',
>     > frame='fk5').separation(a).to('arcsecond')
>     > Out[25]: <Angle 0.4019071919711007 arcsec>
>     >
>     >
>     > My question is: am I misunderstanding something about these coordinate
>     > transformations to make them not equivalent? A third of an
>     arcsecond is
>     > significantly big deviation, particularly for HST or interferometry.
>     > AFAIK fk5 is J2000 in both libraries and galactic coordinates have no
>     > concept of epoch or equinox time.
>     >
>     > Thanks,
>     > Joseph Booker
>     >
>     > _______________________________________________
>     > AstroPy mailing list
>     > AstroPy at scipy.org <mailto:AstroPy at scipy.org>
>     <mailto:AstroPy at scipy.org <mailto:AstroPy at scipy.org>>
>     > http://mail.scipy.org/mailman/listinfo/astropy
>     >
>     _______________________________________________
>     AstroPy mailing list
>     AstroPy at scipy.org <mailto:AstroPy at scipy.org>
>     http://mail.scipy.org/mailman/listinfo/astropy
> 
> 
> 
> 
> -- 
> 
> * * * * * * * * http://www.mssl.ucl.ac.uk/~npmk/ * * * *
> Dr. N.P.M. Kuin      (n.kuin at ucl.ac.uk <mailto:n.kuin at ucl.ac.uk>)      
> phone +44-(0)1483 (prefix) -204927 (work)
> mobile +44(0)7806985366  skype ID: npkuin
> Mullard Space Science Laboratory  – University College London  –
> Holmbury St Mary – Dorking – Surrey RH5 6NT–  U.K.
> 
> _______________________________________________
> AstroPy mailing list
> AstroPy at scipy.org
> http://mail.scipy.org/mailman/listinfo/astropy