ScienceProblem: UnusualStarsInGalacticPlane

PrimaryActor:

Research astronomer

---

ScienceGoal:

To discover more examples of unusual types of star in the Galactic Plane.

---

DataSets:

A range of (geographically-distributed) survey datasets, which combine to yield a multi-colour space in which interesting sources can be detected as outliers: MSX, 2MASS, SuperCOSMOS Sky Survey, H alpha survey, plus possibly radio data (from the Canadian Galactic Plane Survey or HIPASS, and, in a few years, near-infrared data from the UKIDSS Galactic Plane Survey.

---

ProblemDescription:

Infrared wavebands, where the extinction by Galactic dust is low, enable one to survey through the Galactic Plane and discover interesting stars on the far side of the Galaxy. This can be useful for studies of Galactic structure, where one wishes to trace the distribution of a certain class of star throughout the Galaxy, and also for the construction of samples of intrinsically rare classes of star, for which one must survey a large volume of space.

Unusual stars can be detected as outliers in multi-dimensional colour spaces, and there now exist the survey datasets over the range of passbands required to do this properly. The advent of MSX is a great boon here, since it provides a great improvement over IRAS, both as regards sensitivity and angular resolution: the latter is especially important here, since it means that, unlike IRAS, MSX is not confused in the Plane.

What is necessary is to associate point sources in the catalogues generated by these various sky surveys and look for outliers in the multi-colour space so created.

---

CurrentSolution:

This can already be done now, in principle, since the datasets are not particularly large. But there are two areas where the VO is expected to deliver the enhanced functionality that will make this a much easier task. Firstly, the VO is expected to provide easier means of visualising and manipulating multi-dimensional datasets than is commonly available yet.

Secondly, the VO is expected to provide a more sophisticated system for associating sources in different catalogues than is typically applied in astronomy now. This is the key to this ScienceProblem, since the survey datasets included here vary significantly in positional accuracy, so it is crucial that the correct sources in each catalogue are associated or else, in such crowded fields, many spurious identifications will be made, with the result that many objects with apparently unusual colours will be ediscovered erroneously.

---

VOSolution:

The VO must provide a sophisticated mechanism for the association of sources in catalogues in different wavebands. Some techniques already exist for doing this (e.g. Downes et al. 1986, Sutherland & Saunders 1992), but they commonly operate between only two catalogues, whereas it may be preferable (if difficult) to attempt associations on the basis of a larger number of measurements. Furthermore, the VO must provide with its association results (likely to be expressed as probabilities) enough information on the association procedure used that the astronomer can judge whether the associations made are likely to be reliable for his/her particular analysis, and to offer a means of readily re-running them, with uploaded code, should that not be the case.

---

KeyReferences:

Peredur Williams (IfA, Edinburgh), private communication

Downes A.J.B., Peacock J.A., Savage A., Carrie D.R., 1986, MNRAS, 218, 31

Sutherland W., Saunders W., 1992, MNRAS, 259, 413

---

---

GoodStyle: Please add comments below. This area should be used for refinement of the above document. If you want to ask questions or start a dialogue with the author, please use (or create) a topic in the Science Problems Forum. For other ScienceProblems, refer to the ScienceProblemList.
Author: Once the refinements here and comments in the forum die down, perhaps you could rewrite the problem, incorporating the comments and refined

---

Presumably the ability to flag outliers in a distribution in an arbitary space is useful when re-performing the analysis using a different space?

-- NicholasWalton - 08 Apr 2002

This is similar to some other general outlier type problems. Uses lots of data we don't currently have so leave for the VO.

-- NicholasWalton - 17 Apr 2002

---

-- BobMann - 15 Feb 2002