Want to show a 3D model of the sun and surrounding stars

[johnpickin]

Hi
I would like to go out to say 50 million light years and look back on the sun and all the stars surrounding the sun in a 35 million light year diameter sphere.
So you would see a ball of stars with the sun in the middle and then take sections through the centre to look at the relative star densities in the various planes through the centre.
Is that possible does anyone know - or should I use a CAD package - and if so can anyone recommend one?

Thanks for any help.
Cheers

[selden]

John,

Celestia does display stars in three dimensions and some programming (in Lua) could be used to generate displays of the various distributions of stars that you mention.

Unfortunately, however, the information about the stars which is needed for your project does not yet exist. You'll have to wait another decade for the results from the Gaia astrometric satellite. It is expected to be launched in December, 2011. http://sci.esa.int/science-e/www/area/i ... fareaid=26

Good parallax distances to somewhat more than 100,000 stars have been measured, primarily by the Hipparcos astrometric satellite, to a maximum distance of only about 16,000 light years. Even so, very few have been measured at the larger distances.

Relatively inaccurate spectrographic distances have been estimated for about 2,000,000 stars.

Celestia includes the 100,000 star database, and a 2,000,000 star database is available at http://www.celestiamotherlode.net/catal ... _stars.php down near the bottom of the page.

[johnpickin]

Thanks for that.
OK if we started with just the 100,000 (or even the 2,000,000) stars, how might we approach the display side?
I know a good programmer. Is there a resource that explains how to go about it?
Thanks again.

[selden]

Documentation for CelX scripting (Lua for Celestia) is available at
http://www.celestiamotherlode.net/catal ... ation.html
http://en.wikibooks.org/wiki/Celestia/Celx_Scripting
and
http://www.lua.org/docs.html

What's the background for your project?

[johnpickin]

A scientist colleague of mine once did a manual 3D calculation (using I think the Hipparcos and Tycho catalogues) and claims that the results gave a low star density for so many light years around the sun - compared to average density throughout the "sphere". ie the conclusion would be that our star is locally in a relatively low star density region of the galaxy.

If a 3D model were built it would be nice to demonstrate or even animate that.

[t00fri]

A scientist colleague of mine once did a manual 3D calculation (using I think the Hipparcos and Tycho catalogues) and claims that the results gave a low star density for so many light years around the sun - compared to average density throughout the "sphere". ie the conclusion would be that our star is locally in a relatively low star density region of the galaxy.

If a 3D model were built it would be nice to demonstrate or even animate that.

Before drawing possibly incorrect conclusions, it is VERY important to make sure that the Hipparcos (HIP) dataset does not contain biases related to the measurement hardware or other sources. Most astronomical datasets do incorporate such biases (e.g the Sloan Digital Sky survey (SDSS) etc). So ideally one needs to first examine a display of the acceptance across the full angular region.

The HIP star data in Celestia do have a cut-off depending on errors on the parallax, since we do need sufficiently accurate distance information for 3D display...

Fridger

[johnpickin]

<<So ideally one needs to first examine a display of the acceptance across the full angular region.>>

Sorry, not sure what that means...or how to go about doing it?

[t00fri]

<<So ideally one needs to first examine a display of the acceptance across the full angular region.>>

Sorry, not sure what that means...or how to go about doing it?

An acceptance curve (e.g. in your case as function e.g. of RA and DEC coordinates) is in general a ratio

A = measured number / true number

If A<1, the number indicates quantitatively that for some reasons not all objects could be accounted for in the measurement. Hence corrections need to be applied to a naive star counting to account for unphysical depletions of stars in your case.

There may be various reasons for A<1. E.g. hardware-related ones during the original HIPPARCOS star measurements. For that you should consult the scientific HIPARCUS papers for respective results. E.g.

http://www.rssd.esa.int/SA/HIPPARCOS/docs/vol1_all.pdf

While the HIP star counts are relatively complete for apparent magnitudes <= 7.3-9 depending on galactic latitude, they are certainly incomplete for dimmer stars <=12 mag. That fact has to be corrected for etc.

Or a typical bias from within Celestia arises, if the measured star parallaxes are too small for a meaningful distance determination. Incorrect distances would obviously distort your naive star counting results across a plane.

Only if you account and correct for such NON-physical effects can you get a result that is physically meaningful.

Fridger

[johnpickin]

OK thanks.
But if the results show a star depletion over 360 degrees in all planes (as I believe his work indicated), it would indicate either that there is something unusual going on, or there is an inherent problem with the way we estimate star distances, would it not?

[t00fri]

OK thanks.
But if the results show a star depletion over 360 degrees in all planes (as I believe his work indicated), it would indicate either that there is something unusual going on, or there is an inherent problem with the way we estimate star distances, would it not?

This depends on the precise formulation of his question!

If he would ask for the star distribution across planes with apparent magnitudes < 9, then the HIP data would be about complete, while a required magnitude cut-off of say < 12 would introduce a strong artificial depletion, since so many stars in the magnitude range 9 ... 12 could not be measured in the HIP experiment.

This can be further investigated by selecting a sequence of apparent magnitude cutt-offs, say 5,6,7,8,9. Then one can evaluate in each case the results obtained with standard statistical methods and compare the magnitude dependence of a possibly observed depletion!

Similarly, if he did impose NO distance cut-off, then a strong artificial depletion would be introduced by the fact that in Celestia we have to eliminate stars, where the parallax errors are too large. This can easily be corrected as well, by only considering the stars within a finite-size sphere, the radius of which corresponds to the smallest parallax that still makes sense given the HIP errors.

Again, one would repeat the study for various radii of that sphere, corresponding to different uncertainties on the parallax data. This way one can study the systematic errors associated with distance assignments in a quantitative manner.

And so on. Clearly in such a computer experiment the most crucial aspect is to formulate the question precisely, such that biases are mostly eliminated. To do so, a careful study of the original scientific literature and the strategies applied in Celestia are required.

Finally, a good knowledge of statistical methods is needed, in order to be able to quantify the statistical relevance of the results obtained.

Fridger

[johnpickin]

Thanks Fridger, very helpful.
However........errrm.....................slight error in orders of magnitude.........
Have now spoken to the original author again. The max distance for measurement of star density around the sun was 38 light years - not 35 million (told you it was only a slight error...!)

His original paper attached in zip folder.

Any more thoughts on the validity of the data given the much smaller distances or would you still recommend repeating with various magnitude/distance cut offs etc.
Thanks again for the help.

[t00fri]

OK, I'll have a look at the paper as soon as I find some time...

Fridger