Project: Globular clusters

[t00fri]

Hi all,

I wonder who of you might be interested to devote some time, ideas and know-how to address --together with me-- the long overdue implementation and rendering of globular (and also open) clusters into Celestia. Since this is a scientific-level project, some previous experience with such type of analysis would be useful.


I don't need any programming help, really since the coding is quite straightforward. However, anyone interested in coding aspects is of course welcome.

What is quite important is to search and sort out available data sources. The most accurate source for coordinates and some other data is the Januar 6th 2006 edition of the revised NGC/IC catalog by Wolfgang Steinicke. (Just as for galaxies) . These data however are lacking distances! Here I have already other information sources...
-----------------------------------
CATALOG OF PARAMETERS FOR MILKY WAY GLOBULAR CLUSTERS:
THE DATABASE

Compiled by William E. Harris, McMaster University

This revision: February 2003
-----------------------------------------------

Altogether, we talk about ~150 globular clusters with all information ACCURATELY available. These include ALL respective Messier objects, of course.


As usual, the selected information will be extracted and merged via PERL into a .dsc data file. This I can do very quickly, once the data are there.

Next comes an interesting phase of testing the rendering (via sprites), optimize the colors and design the templates, etc...

Altogether, it's not very much work, since many aspects are quite in parallel to our previous galaxy implementation.

What is particularly interesting in the context of globulars is their typical satellite locations around our MilkyWay.

Here is a respective illustration from
http://www.anzwers.org/free/universe/globular.html
A plot of the globular cluster data onto a sideways view of our MilkyWay galaxy, revealing how globular clusters form a spherical halo around it!





Let me know....

Bye Fridger

[ElChristou]

I'm out to help with the catalogues but I'm in to help on the rendering (of course ) ...

[t00fri]

I'm out to help with the catalogues but I'm in to help on the rendering (of course ) ...

Aha! The old, proven crew is coming once more on board

GREAT!

Cheers,
Fridger

[chris]

A globular cluster catalog is long overdue. I agree that most of the work is in creating the dsc files from some standard catalog. I think it makes the most sense to treat globulars as another galaxy type, since they are in some sense miniature galaxies.

--Chris

[t00fri]

A globular cluster catalog is long overdue. I agree that most of the work is in creating the dsc files from some standard catalog. I think it makes the most sense to treat globulars as another galaxy type, since they are in some sense miniature galaxies.

--Chris

Actually, all the required "infrastucture" of the DeepSkyObject base class is there already!

I would prefer to rename the existing derived OpenCluster class simply 'Cluster' with the two types 'Galactic' and 'Open'. The rest is routine. It's physics wise/conceptionally much cleaner to treat clusters in form of a separate derived class, instead of another "galaxy type". Color profiles are somewhat different from those of galaxies, the sprites must be adapted, too, and we need a few good templates. The existing data precision is VERY good. Uncertainties for the whole sample are just a few percent.

Bye Fridger

[ElChristou]

A globular cluster catalog is long overdue. I agree that most of the work is in creating the dsc files from some standard catalog. I think it makes the most sense to treat globulars as another galaxy type, since they are in some sense miniature galaxies.

--Chris

Actually, all the required "infrastucture" of the DeepSkyObject base class is there already!

I would prefer to rename the existing derived OpenCluster class simply 'Cluster' with the two types 'Galactic' and 'Open'. The rest is routine. It's physics wise/conceptionally much cleaner to treat clusters in form of a separate derived class, instead of another "galaxy type". Color profiles are somewhat different from those of galaxies, the sprites must be adapted, too, and we need a few good templates. The existing data precision is VERY good. Uncertainties for the whole sample are just a few percent.

Bye Fridger

How many templates do we need? some are very dense, others very "open"... no particular shape I suppose?

[chris]

A globular cluster catalog is long overdue. I agree that most of the work is in creating the dsc files from some standard catalog. I think it makes the most sense to treat globulars as another galaxy type, since they are in some sense miniature galaxies.

--Chris

Actually, all the required "infrastucture" of the DeepSkyObject base class is there already!

I would prefer to rename the existing derived OpenCluster class simply 'Cluster' with the two types 'Galactic' and 'Open'. The rest is routine. It's physics wise/conceptionally much cleaner to treat clusters in form of a separate derived class, instead of another "galaxy type".

It seems to me that open clusters and globular clusters should be distinct. No special rendering is required for open clusters; we need to show nothing more than the actual stars as defined in stars.dat or an .stc file. An open cluster definition is only interesting if you have labels turned on. But globular clusters should be rendered with sprites like galaxies and not as individual stars. In addition, the way that globular clusters and open clusters form is very different. So there are practical and scientific reasons for not creating a single star cluster class. Whether globular clusters should be treated as galaxies is a separate matter. From a coding standpoint, it's makes sense to treat them identically I think; but there may be a scientific argument for making them distinct classes.

--Chris

[Cham]

If we want to search among the globular clusters only, it makes sense to have them treated in a distinct class.

And I recall that we urgently need more classes of objects (custom classes too !).

[t00fri]

...
It seems to me that open clusters and globular clusters should be distinct. No special rendering is required for open clusters; we need to show nothing more than the actual stars as defined in stars.dat or an .stc file.

Here I cannot agree. Altogether in the NGC/IC catalog there are 685 open clusters. In our Galaxy we know about 1000!. For all these, we do have the detailed Trumpler classification, with 3 'richness' categories and 4 stages of concentration towards the center. While some Messier objects are indeed poor and thus booring, this is NOT to be generalized! The r category has >100 stars, for example.

Here is a familiar example: M11 in Scutum:



M11 has > 2900 stars many too faint to be included in any standard star catalog. So indeed in order to allow for a systematic inclusion of the O(1000) open clusters, special /automatic/ rendering is required! Often open clusters are associated with diffuse nebulosity which again calls for rendering.

From the physics open clusters are far more special than just another "bunch of stars". If desired I can always summarize the many properties that are common to the members of an open cluster, including notably age, distance, chemical composition.

I think that should do for now. While globular and open clusters are quite different they are nevertheless both star clusters. Just like we know various VERY different kinds of nebulae.

Bye Fridger

[t00fri]

Special rendering requirements of open clusters arise notably also from their VERY similar age hence very similar colors + diffuse nebulosities.

An example, again VERY well known, is the "Butterfly" cluster M6:

Here is an image from
http://antwrp.gsfc.nasa.gov/apod/ap990106.html



There are many such exquisite examples among the 1000 open clusters in our galaxy. They certainly deserve their own rendering!

I am currently thinking about how to achieve some typical required effects.

Bye Fridger

[chris]

...
It seems to me that open clusters and globular clusters should be distinct. No special rendering is required for open clusters; we need to show nothing more than the actual stars as defined in stars.dat or an .stc file.

Here I cannot agree. Altogether in the NGC/IC catalog there are 685 open clusters. In our Galaxy we know about 1000!. For all these, we do have the detailed Trumpler classification, with 3 'richness' categories and 4 stages of concentration towards the center. While some Messier objects are indeed poor and thus booring, this is NOT to be generalized! The r category has >100 stars, for example.

M11 has > 2900 stars many too faint to be included in any standard star catalog. So indeed in order to allow for a systematic inclusion of the O(1000) open clusters, special /automatic/ rendering is required! Often open clusters are associated with diffuse nebulosity which again calls for rendering.

From the physics open clusters are far more special than just another "bunch of stars". If desired I can always summarize the many properties that are common to the members of an open cluster, including notably age, distance, chemical composition.

I'm well aware of the properties of open clusters. But you'll agree that in some cases (e.g. the Hyades) where the stars are mostly contained in a standard catalog that it's best for Celestia to treat the cluster as nothing more than a collection of stars.

I think that should do for now. While globular and open clusters are quite different they are nevertheless both star clusters. Just like we know various VERY different kinds of nebulae.

I think that open and globular clusters are too different to be considered the same class of object. There's nothing to be gained by making them the same class. I do agree with your suggestion that we could augment the rendering of open clusters with some additional geometry for nebulosity or large numbers of uncatalogued stars.

--Chris

[buggs_moran]

If I could add my amateur voice here...

Open clusters are mostly contained within our "local" purview. From what I read in the Catalog of Optically Visible Open Clusters, if I was reading it correctly, 515 of the 539 visible open clusters (with distance data and more than 10 members) are within the 16000 LY limit of Celestia for stars. If you were to "auto" generate them, I would think you would want to use stars, not sprites. Just somehow constrain the clusters to the apparent diameter at the proper distance.

For globulars I think everyone would agree that the sprite method would be it. You could obviously have different models based on the 12 concentration classes. I imagine you would want Omega Centauri to "look" far more massive and dense than others.

Anyway you do it I am sure it will be wonderful, but I believe that, within that limit set by the program, one would want to make open clusters more "believable" with stars, as defined by Celestia, not point sprites, unless those point sprites could be given radii that would emulate stars from the user's perspective. I admit, I know very little about the point sprite limitations, so I could be way off base.....

[selden]

An advantage to representing Open Clusters with Star objects is that one can use real survey data to place cluster members in Celestia.

There are two disadvantages, though:

1. The differences in the distances to the individual stars would have to be estimated from other information (spectral type & luminosity?) or randomized.

2. Most surveys are of limited regions well within the bounds of clusters. The known members are all inside rectangular areas on the sky recorded by photographic plates or CCDs while the clusters actually extend well beyond those areas. The resulting rectangular star distributions look rather strange in Celestia.

[t00fri]

An advantage to representing Open Clusters with Star objects is that one can use real survey data to place cluster members in Celestia.

There are two disadvantages, though:

1. The differences in the distances to the individual stars would have to be estimated from other information (spectral type & luminosity?) or randomized.

2. Most surveys are of limited regions well within the bounds of clusters. The known members are all inside rectangular areas on the sky recorded by photographic plates or CCDs while the clusters actually extend well beyond those areas. The resulting rectangular star distributions look rather strange in Celestia.

Selden,

generally I agree with your pros and cons.

As I am concerned, I am NOT interested in setting up a few individual
open clusters among the O(1000) in our galaxy "by hand".


So if we go for REAL star data at all, then we should find a reliable data
source where we can find the star data for 100's of open clusters.


Since these do almost surely not exist, we need to go another way.


-- It has been my original intention to populate open clusters with
statistically generated stars (rather than sprites). The population
of the open clusters is to be conform with the respective Trumpler
classification that I can read out from the catalogs, the specified
size of the cluster and the given color data. This allows to generate
quite realistic /generic/ density distributions etc of cluster stars.


--To include the rendering option of diffuse nebulosity, since
this is a basic feature connected with open clusters and their origin.
They are believed to originate from large cosmic gas/dust clouds in the
MilkyWay. In many clouds visible as diffuse nebulae, star formation
takes still place at this moment, so that we can observe the formation
of new young open star clusters (composed of young Population I
stars).

--To account carefully of the coloration of the member stars of the
open cluster. Since both the age and chemical composition of these
stars are very similar, their color is also often very similar. Often young
(blue) stars are involved like in the Butterfly cluster above or in the
famous Pleyades, with a beautiful image being here:

http://antwrp.gsfc.nasa.gov/apod/ap021201.html


Bye Fridger

[selden]

The youngest open clusters still have nebulosities associated with them, but the older ones don't any more.

A reasonably reliable secondary source of cluster star data is http://www.univie.ac.at/webda/
I've personally used it only for individual clusters, but it should, in principle, be possible to get a listing of all the stars it has cataloged.

[t00fri]

The youngest open clusters still have nebulosities associated with them, but the older ones don't any more.

Certainly. That's why often the cases with nebulosity do involve young blue stars. In fact since globular clusters, in contrast, are VERY old, there is never any nebulosity left with globulars. The latter consist from an earlier generation of stars (Population II), which have formed from the more primordial matter present in the young galaxy just after (or even before) its formation.

A reasonably reliable secondary source of cluster star data is http://www.univie.ac.at/webda/
I've personally used it only for individual clusters, but it should, in principle, be possible to get a listing of all the stars it has cataloged.

Wow, Selden,....

that looks helpful. Looking forward to a closer examination of the data material.

Thanks,
Fridger