Wednesday, October 13, 2010
Comet 103P in Stellarium, a cautionary tale
Left image, 103P as imaged with the GRAS 14 telescope in New Mexico; Right Image, predicted position of 103P as visualized in Stellarium at the same time.
Stellarium is a pretty handy program. Realistic planetary visualization, customizable, I can run as a planetarium as well as control telescopes, and above all, it's free! You can also add objects to it. However, a reader (Wojtec) has posted a problem with Stellarium's implementation of 103P Hartley.
UPDATE: Reader Dagger stab writes:
If you are using the MPC's Minor Planet & Comet Ephemeris Service as a source, there's an error in your Stellarium data: it should be "orbit_PericenterDistance = 1.058686". (There should be an 8 after the 5.) It's correct in the Celestia script. :)(Thanks RD. Strangely, I used both the JPL horizons data as well a the MPC data, why SkyMap gets it right with the MPC data is a mystery, I will check back)
UPDATED UPDATE: I ended up completely rebuilding the Hartley add-on with fresh MPES data to match SkyMap, and with the rebuilt add-on 103P is within 5' of the CCD position, good performance and about as accurate as SkyMap. I've posted the revised figures on the original post and at the end of this post. The JPL horizons data was older than the MPES data, so it gave wacky results. Makes sense that it was the add-on elements rather than Stellarium being weird.
Wojtec found the Stellarium position was off by over 20 minutes of arc, I've confirmed this, actually, I found a 37 minute of arc discrepancy, and that was using a 103P definition file with the latest MPC parameters. Now this isn't much, for visual location or using binoculars and low power telescopes it will be just fine. But if you are using CCD imaging, then this is a significant discrepancy, if say, you were using the GRAS G14 imager the comet would be significantly displaced (see above images, and the image below which makes it clearer), and with the GRAS G5 imager the comet could be out of frame.
Stellarium location of 103P compared to the SkyMap prediction, and the actual measured position from the G14 CCD image (click to embiggen)
I've tired two separate definition files for 103P, and there is still the same discrepancy, in the image to the left I've used the same parameters for SkyMap and Stellarium. So it's not the values. I've also been careful to check that the times were correct (imaging run vs Stellarium prediction) so the apparent differences were not due to timing errors.
Curiously, it only is happening with 103p Hartley (and to some extent 2010 TD54). For Comet Enke, Comet 10P Temple, and Vesta the Stellarium prediction is within 5 minutes of arc of the actual object, which is quite tolerable. It can't be a topocentric vs geocentric problem, the discrepancy for 2010 TD54, which has a huge parallax, was only 18' compare to 103P's 37' difference (and with only a short observation time you would expect a large uncertainty in the 2010 TD54 position).
So, I'm baffled, I'll put up a query at the Stellarium forums [answered rapidly, see above, thanks DS], but for now, just be aware that if you are using Stellarium to guide your telescopic CCD imaging, 103P's position will be off. [Now Fixed, see below]
name = 103P/Hartley
parent = Sun
radius = 24
oblateness = 0.0
halo = true
color = 1.0,1.0,1.0
tex_halo = star16x16.png
tex_map = nomap.png
coord_func = comet_orbit
orbit_TimeAtPericenter = 2455497.76729
orbit_PericenterDistance = 1.058678
orbit_Eccentricity = 0.695113
orbit_ArgOfPericenter = 181.1969
orbit_AscendingNode = 219.7661
orbit_Inclination = 13.6168
lighting = false
albedo = 0.5
orbit_visualization_period = 365.25