Celestial coordinate system

[WaynePage]

Is there a coordinate system that may be used with ArcGIS to do a sky survey.

Looking for a projection that is spehrical with longitude expressed as Right Asention in Hours, Minutes, Seconds.  I might be trying to do this in the wrong progam and might have to look elsewhere.  Every projection I have looked at with Arc doesn't seem to support such functionality or at least allow the coordinates to display as R.A and Dec. as typically used in astronomy will be using the 2000 epoch as datum, if that is how you would define it in this context.

Thanks.

[DanDuriscoe]

I have begun to explore this problem with ArcGIS and it is possible to create your own custom GCS for equatorial, ecliptic, and galactic coordinates.  The transformations between the three are fixed if you stick with J2000 coordinates.  However:

--I cannot find a way to execute the transformations on the fly, I have found a 3 step approach which must be scripted.  See

http://webhelp.esri.com/arcgiSDEsktop/9.3/index.cfm?TopicName=Equation-based%20methods

the seven parameter method will work, but I have not been able to figure out how to make a transformation, say from equatorial to galactic, with one transformation.  Instead I have to run three different ones sequentially.  Essentially, there are three rotations involved, first around the z axis, then the y axis, then the z axis again.  If you or anyone you know can decipher the matrix algebra involved and apply it to these transformations, I would appreciate any feedback.

--I have succeeded with vector data (point features representing stars).  Raster data presents significant problems which I have yet to completely figure out.  Basically, transforming and projecting grids covering the whole earth (or sky) are not what arcGIS does best.  Difficulties with the "international date line" of 180 east/west present themselves when rotating a whole sky grid around the z axis.  Setting the environment variables and trying different GCS sphere models provide no help.  I also find that spatial references on grids are ephemeral things, and only certain GCS coordinates seem to "stick".  I found a workaround, but it is very cumbersome and must be scripted.

--Celestial coordinates are drawn the same as longitude and latitude on earth in that longitude (right ascension) increases eastward (to the right on a north up map) as long as you are looking at the OUTSIDE of the sphere.  With star maps you are looking at the INSIDE, so resulting maps must the mirrored left to right to match the real sky.  Hours, minutes, and seconds of Right Ascension are best converted to degrees by multiplying decimal hours by 15.  If you want to avoid the mirroring problem, you can calculate "reverse right ascencion" by using 360-(decimal hours * 15) as a transformation to a custom coordinate system.  The maps will then be right reading when compared with the sky.  Again, easy for vector data by merely adding fields in the attribute table, not so easy for raster data.

Any further thoughts are appreciated.

Dan