Celestia Forums

It seems that the coordinates of the stars in Celestia are exactly the coordinates specified in the Hipparcos database. Apparently their positions haven't been precessed from the epoch of the catalog (J1991.25) to J2000. It'd be nice if stars.dat could be updated with this precession applied.

I just discovered this while declaring some stars for the M16 open cluster.
Several of them were in Celestia already, but were a little distance away from the J2000 locations I was using.

I looked up one of them (HIP 89753) in the Hipparcos database.
The data page for that star says that its coordinates are for J1991.25. When I used the coordinates shown in the Hipparcos entry, the star I defined was placed exactly at the location of that star in Celestia.

Hello Selden,

coordinates of all stars in the stars.dat files (including the extended databases) are for epoch J1991.25.

The database could be quite easily precessed to J2000.0 but I don't know what epoch Chris assumes in the procedures that compute the planetary and other objects positions.

I had also a problem with that when I tried to add from other catalogs to the database.

Pascal

The planet positions are computed with respect to the J2000.0 ecliptic and equinox. So, precessing the star positions to J2000.0 would be the right thing to do.

--Chris

Ok,

I will try to find some time to make the conversion in the next days...

Hello everybody, I'm a newbie to Celestia, but I've read the docs about the coordinate system Hipparcos uses.

Saying that the Epoch of the star database is J1991.25 means that the positions of the stars in the Hipparcos database are the position they had as seen from Earth in epoch J1991.25, with coordinates given with respect to the J2000.0 ecliptic and equinox (not the J1991.25 ecliptic and equinox). So don't correct their position using the Earth's precession of equinoxes from J1991.25 to J2000.0 since that would render the stars position innacurate (or even wrong).

In fact, all Hipparcos stellar coordinates (unless otherwise noted) are given in a truly inertial (non-rotating) reference frame, know as ICRS (International Celestia Reference System). This reference frame is based on Earth's equatorial coordinate system with J2000.0 ecliptic and equinox, but was defined with very precise extragalatic reference sources. So even though Earth is precessing and the Earth's equatorial coordinate system is slowly rotating due to earth's precession, the ICRS never rotates. So, for example, the coordinate RA=0 and DEC=0 in the ICRS will be pointing to the same direction in space no matter if we are in the year 2000 or in the year 5000BC.

I've found the definition of the ICRS in a footnote in the Hipparcos database manual (it can be downloaded from the Hipparcos Astrometry Mission site in .pdf format)

Best Regards, and good Job.

P.S: I'm thinking joining the Celestia development team (it is, creating an account in Sourceforge), so I can help with some modules or improvements. But first I have to see if I'll have enough time to work in this project, even though I would love to.

Sorry, I've forgot to add my e-mail for contact:
Name: Marcio Afonso Arimura Fialho (M?rcio Afonso Arimura Fialho, if you can view accents)
e-mail: maaf1980@yahoo.com.br

I am not sure you're right Marcio.

After reading in details what is written in section 1.2 Astrometric Data and 1.5 Transformation of Astrometric Data from the documentation related to the Hipparcos and Tycho catalogs, things seem to be more complicated.

Formulaes for epoch transformation are even given in the last section (see 1.5.4)

But, what seems to have a greater effect on coordinates than precession itself is the proper motion of stars.

If you query the Hipparcos catalog yourself, you will see in the output fields, how proper motion affects the coordinates of a star between J1991.25 and J2000.0.

Thus, to get a precise position for J2000.0, proper motion should be taken in account. And this involves new problems, as some stars (how many ? I don't know) have certainly imprecise proper motions values or no values at all, in the Hipparcos or in the Tycho catalog...

Thus it could be maybe a better idea to precess back computed objects positions from J2000 to J1991.25.

If there is for now, a difference between the reference planes used for stars (J1991.25) and those used for computed objects positions (J2000.0), then a phenomena like the occultation of a star behind, let's say Jupiter or the Moon, could not be seen in Celestia (at the right time or the right position), but I didn't make any test...

It is true that it'd be nice to take into account proper motion, but I'd suggest leaving that improvement until a future major release, when Celestia actually can show star movement.

The current problem is that when one specifies the J2000 coordinates of various extremely distant stars (e.g. in the M16 cluster) they are not at the coordinates specified in the Hipparcos database.

All of the current astronomical databases specify J2000. I think it'd be best if Celestia did too.

The problem is not to take proper motion into account as a general feature of a next release of Celestia, but that proper motion of stars must be taken into account to convert their positions from the J1991.25 reference planes to the J2000.0 ones, to be accurate, and not only the precession.


Pascal

Pascal,

My impression is that the corrections due to precession are significantly larger for most stars than the changes of position due to proper motion, especially for the more distant ones. It also seems to me that doing the precession ought to be an easier correction to make. Don't the same matrix operations get applied to the coordinates of all of the stars in the existing databases? I would expect that doing a good job of accounting for proper motion would involve incorporating values obtained from quite a few different databases, not just Hipparcos and Tycho, which would be a lot more work.

Personally, I'd like to see an improvement that affects most of the stars relatively soon, followed later by the improvements due to proper motion.

Of course, if you're the one who's going to do the work, then it's your decision. The current errors aren't that large: we could live with them for a while longer.

After dealing again and again with Hipparcos documentation, and doing a few queries online on the catalog, I am now convinced that Marcio is right.

There is no need to precess coordinates given in the Hipparcos or Tycho catalogs from J1991.25 to J2000.0.

Coordinates of a star in those catalogs are given "for epoch J1991.25 (JD2448349.0625 (TT)) in the ICRS (International Celestial Reference System, consistent with J2000 reference system".

J1991.25 is just the mean epoch of observations by the Hipparcos satellite.

And this really means that the only parameter that should or could be taken into account to get very precise positions for J2000.0 is proper motion and this parameter only.

Thus, there is nothing to do with current positions in the stars databases.

I have made for example a query on HIP 89753, witch is a very distant star, with very low proper motion values. The output of VizieR don't show any differences between right ascension and declination in ICRS J1991.25 and in J2000 systems.

Tell me if I am wrong !

Pascal

Pascal,

It indeed appears that you are correct: I double checked, comparing the values from Visier with those in the Hipparcos database.

Evidently there are some numeric errors in the algorithm I was using to do the precession.

Thanks!

Hello everybody. It's me again. Marcio A. A., Fialho.
I've just created an accout in shatters.net

You can contact me through my e-mail: maaf1980@yahoo.com.br

Till next time.

selden wrote:My impression is that the corrections due to precession are significantly larger for most stars than the changes of position due to proper motion, especially for the more distant ones.

This is true.

I would expect that doing a good job of accounting for proper motion would involve incorporating values obtained from quite a few different databases, not just Hipparcos and Tycho, which would be a lot more work.

Hipparcos measured proper motions, and they are stored in the Hipparcos catalogue as columns H12,H13. The corresponding error measures are H17,H18. 10,295 stars in HIP have valid proper motions with relative errors of less than 20%, and valid parallaxes. Of these, 4930 have proper motion corrections that would be greater than the uncertainty in their distances (from parallax) over the period 1991.25 to 2000. An extra 400 stars would have proper motion corrections greater than their 3D positional uncertainties if you take 1991.25 to 2003.5. It's therefore a minor problem, but one that will only get worse.

Don't forget that Celestia is a 3D program. Radial velocities have to be taken into account in addition to the proper motion in RA and Dec.

The discussion so far seemed focussed on proper motion only. To use full 3D information would require a radial velocity catalogue, yes, but I think it's been pointed out before in these forums that such information is available for many stars. There are minimal difficulties if you're only applying the changes to the catalogue once for some correction from the observation epoch to the current epoch.