Bug? Obliquity and Longitude of Rotational Axis

[dtessman]

I was twiddling with Obliquity and Longitude of Rotational Axis in the solarsys.scc and I found what appear to be anomalies (pun intended).

I removed the custom orbit of Earth and changed its orbital inclination to be within the celestial equatorial plane and set its obliquity to zero. Earth's axis was still tilted. I had to set the obliquity to -23.45. to get Earth's axis 90 degrees to its orbital plane.

Obliquity should be the "The angle a planet's rotational axis makes with its orbital plane". It appears that Celestia's obliquity is based on the ecliptic in general rather than the object's actual orbital plane.

I next made sure that the longitude of the rotational axis was set to 0.0. Then I set the Obliquity of Earth to 21.55, which because of the obliquity problem makes the earth tilt 45 degrees in its new (celestial equator based) orbital plane. It did, great!

But the Earth's North Pole pointed toward 18 hours. It should point toward 0 hours. which is also supposed to be the origin for 0 longitude.

Am I missing something obvious here or is this a real problem?

[granthutchison]

In 1.2.5 the "longitude of the rotation axis" has had its name changed to reflect its real role - the equatorial ascending node. With the equator ascending node set to zero, the axis is going to point towards 18h. It's an internally consistent set of parameters (well, now the name has changed it is ... ), just not what you might expect - in Celestia, both orbital and planetary rotational parameters are defined relative to the ecliptic.

Grant

[dtessman]

And with a simple change of name... Voila! Thanks!

Are there any great arguments for aligning obliquity to the ecliptic rather than orbital plane?

[chris]

And with a simple change of name... Voila! Thanks!

Are there any great arguments for aligning obliquity to the ecliptic rather than orbital plane?

I'm aware that Celestia's definition of obliquity differs from convention. The reason that the reference plane is the ecliptic (specifically, the J2000 ecliptic) rather than the orbital plane is that the orbital plane isn't constant. I think it's much more convenient to use a fixed plane as a reference . . . at least, it works better in the Celestia code I suppose that I could use the mean orbital plane as a reference, but that seems a bit fuzzy. Also, since Celestia doesn't use Keplerian elements for the major solar system bodies, it doesn't even know what the mean orbital plane is . . . I should mention that the reference plane for obliquity and other rotation parameters is only the ecliptic for planets. For satellites, the equatorial plane of a planet is the reference.

--Chris