VSOP87 and DE-405, in relation to Celestia

[JackHiggins]

DE-405 and VSOP87 are both used by JPL Horizons to compute orbits, but celestia only uses VSOP87.

So 2 questions:

• What is the difference between the two systems?
• What benefit (if any) would there be to use DE-405 in Celestia also?

Thanks

[Falck]

DE405 (these days usually in the form of Spice kernels) is used by NASA and others in the space industry to obtain cartesian positions and velocities of the planets in the Earth Mean Equator at J2000 (EME2000) coordinate frame. Units are either AU and AU/day or km and km/s. DE405, as I understand it, provides interpolating Chebyshev polynomials for the major bodies of the solar system. Data files are rather large for relatively short periods of time (10's of MB for 100 years or so). But, the advantage is, the positions of the planets are within 10's of meters of accuracy! If you think about it, thats probably well within the error to which we know the true positions of the centers of mass of the major bodies.

I've never used VSOP87, but by the looks of it, it is more geared to astronomy, rather than astrodynamics. From what I've read, it outputs helicentric latitude and longitude to within 1 arcsecond of accuracy. But while less accurate than the DE405 ephemerides, they information provided by them extends for hundreds or thousands of years.

One benefit to using DE405 in celestia would be somewhat increased accuracy in position information. I'm not sure, however, if such a difference would be noticable visually.

I would still love to see Celestia become a high quality mission visualization tool, and DE405 would probably help its acceptance in that regard. Maybe this would be a nice option in Celestia, however I think it needs improvements in other areas of mission visualization before this would become an issue. For the current time being, adding DE405 support would give hyper accurate positions, but they would be invalid after the DE405 interpolation expires. At this point, I dont think it goes beyond 2200AD (could be wrong on that though).

[HankR]

I've thought about implementing spice trajectories in Celestia experimentally. I suspect that putting it in the official version would present complications with distributing the spice libraries and kernel files.

- Hank

[Falck]

Im not sure what issues there are with incorporating SPICE capabilities into GPL software. Its publicly available, and maybe putting the capabillity in without any data files would be acceptable? (Users could obtain or create these on their own) Have you tried contacting someone at NAIF to see if there are any issues with this?

[HankR]

I wasn't thinking about legal complications. Actually, I'm pretty sure that NASA would be cooperative, given Celestia's substantial educational value. I was thinking more in terms of filesize and installation and configuration complications, considering that the precision of the SPICE datasets is probably overkill in terms of most Celestia users' needs.

- Hank

[Falck]

True, though I would think there would be those among us would would welcome the capability, at least for spacecraft ephemeris information if nothing else.

[granthutchison]

One benefit to using DE405 in celestia would be somewhat increased accuracy in position information. I'm not sure, however, if such a difference would be noticable visually.

For close flybys planets and satellites it would be noticeable, I think. There's a few hundred kilometres of error in Jupiter's position, which shows up because the Galileo sampled orbit is tied to the Sun, rather than Jupiter - the Amalthea flyby is noticeably affected, for instance.

Grant

[HankR]

Wth access to the SPICE datasets, it might be possible to create a tool that would produce improved xyz files for Celestia which would address some of these problems. For example, the Galileo SPICE trajectory data could be extracted in Jovicentric coordinates and then be converted to heliocentric coordinates using Celestia's Jovian custom orbit. Also, the sampling time interval could be adjusted based on the proximity to flyby targets to avoid collisions due to interpolation inaccuracies. But this still might not be adequate for extremely close fly-bys. (It also might be more work than just implementing SPICE trajectories directly in Celestia...)

- Hank

[Falck]

That would be a nice capabililty to have...although I think the option to use SPICE ephemerides for planets themselves (as long as the dates are available) would also be a good solution to the problem.