Milky Way orientation

[selden]

It seems that the orientation of the Milky Way in /data/galaxies.dat is slightly wrong.

When one goes far away and looks back at the galaxy and the stars, the plane of the galaxy is not in the same plane as the highest density of stars. I would have expected the Hipparcos survey to show most stars in the plane of the galaxy.

As a test, I generated some "fake galaxies" at increasing distances at galactic coordinates (0,0), (0,90), (0,-90) and (90,0). The (90,0) row should be in the plane of the galaxy, but is not. I then created more markers at (90,5), (90,10), (90,15) and (90,20).

As best I can tell, Celestia currently draws the Milky Way tilted by about 10 degrees. Unfortunately, I haven't been able to puzzlie out the coordinates used by the Axis and Angle specifications, so I can't provide the appropriate corrections.
[url=http://www.lns.cornell.edu/~seb/celestia/ganti1.jpg]
[/url]

This picture is taken from a distance of 100,000 ly, at (180,0)
I used the Precess utility at http://cxc.harvard.edu/ciao/ahelp/precess.html to convert galactic coordinates into (J2000) RA and Dec.

[Matt McIrvin]

I've often wondered about this myself, but it's also conceivable that the mismatch is not an error. The distribution of stars in astronomical catalogs is heavily dependent, not just on the actual distribution of stars in space, but also on the distribution of obscuring dust in the galactic neighborhood of our solar system, which hides stars from our view. There could also be local warpage in the disk which would contribute to an apparent mismatch.

These effects might not be sufficient to explain what you're seeing, but they're the reason why I never made much noise about it.

[selden]

Matt,

A 10 degree tilt is more than one would expect in this case.

A file containing the coordinates of my "calibratin blobs" is available at http://www.lns.cornell.edu/~seb/celestia/galaxies.txt

Not visible in the picture above is another row of blobs in the direction of galactic coordinates (0,0). They nicely pass through the galactic center.

[chris]

Matt,

A 10 degree tilt is more than one would expect in this case.

A file containing the coordinates of my "calibratin blobs" is available at http://www.lns.cornell.edu/~seb/celestia/galaxies.txt

Not visible in the picture above is another row of blobs in the direction of galactic coordinates (0,0). They nicely pass through the galactic center.

I oriented the Milky Way by eyeballing it . . . There's a semi-secret key in Celestia that lets you adjust the orientation of the selected galaxy and print out the axis angle rotation. However, using the galactic coordinate system is definitely preferable to my ad hoc method. You can edit galaxy orientation by pressing the ~ key and then shift+control dragging the mouse. Pressing ! echoes the orientation to standard out--you'll have to run from the command line and redirect celestia's output to a file to see it. Let me be the first to admit that this is a completely stupid way to orient galaxies, particularly without any sort of calibration objects. I'll fix the orientation of the Milky Way for 1.2.6. And presumably, there are galaxy catalogs that contain some information about orientation, so that we don't have to manually tweak every galaxy.

--Chris

[selden]

Chris,

Thanks for the background explanation.
We're all looking forward to 1.2.6!

[selden]

After playing with the galaxy edit mode for a while in v1.3.0pre1, I think I may have stumbled on a reasonable orientation.
(this links to a larger image)
topleft: in the plane of the galaxy, looking toward the sun, with the galactic center beyond it.
bottomleft: from above the galaxy, directely above the sun. This view was rotated somewhat to approximate the figure mentioned below.
bottom right: in the plane of the galaxy, galactic center toward the right. Note that the sun is slightly above the galactic plane, as it should be. This was a fluke, though: I don't know how closely it matches "reality".
top right: a "perspective" view, as seen from above and to the left.

As mentioned in the postings above, I used "calibration galaxies" positioned at appropriate locations in galactic coordinates. Here they're labelled in yellow.

Here are the corresponding coordinates for use in deepsky.dsc:

Code: Select all

Galaxy "Milky Way"
{
   Type "SBa"
   RA 17.75
   Dec -28.93
   Distance 28000
   Radius 50000
#   Axis [ 0.712 0.533 0.456 ]
   Axis [ 0.866 0.491 0.091 ]
#   Angle 149
   Angle 176
}


Unfortunately, Celestia only seems to define 2 arm spiral galaxies. As best I've been able to determine, current information seems to suggest the Milky Way has 4 arms. See for example, Figure 2.6 in the document at http://bookmarkphysics.iop.org/fullbooks/0750308370/falckech02.pdf

[chris]

Cool!

I've edited the definition of the Milky Way in deepsky.dsc and checked it into CVS. Your fixed version will appear in the next prerelease. Thanks for your efforts!

--Chris