ScienceProblem: MultiWavelengthAssociations

PrimaryActor:

Research astronomer

---

ScienceGoal:

To identify observations of a set of astronomical objects in catalogues obtained in a number of different passbands.

---

DataSets:

Catalogues from a range of passbands, e.g. radio, infrared, optical, X-ray.

---

ProblemDescription:

A central concern for the VO the association of entries in different databases - a specific case of a ScienceProblem based on this issue is given in UnusualStarsInGalacticPlane, and we consider the more general problem here.

In some cases, the properties of the datasets involved (e.g. very small positional uncertainties) will mean that a source in one catalogue can readily be associated with an object in another with very high confidence, but, in many other cases, associations will have to made probabilistically, on the basis of the expected properties of the species of astronomical object under study. For example, when seeking optical identifications of sources in an X-ray catalogue, one might have a source whose X-ray spectrum was unambiguously that of a Seyfert 1 galaxy, but whose positional uncertainty was sufficiently large that its error ellipse included several optical galaxies, thereby precluding a straightforward association on the basis of proximity alone. In that case, one would choose between the possible optical counterparts within the X-ray error ellipse on the basis of how closely their known optical properties matched that of a Seyfert 1 galaxy.

---

CurrentSolution:

A mathematical framework exists (the likelihood ratio technique, as described, for example by Sutherland and Saunders 1992) for doing this, but it is typically only applied in a limited way - e.g. finding optical counterparts of radio sources on the basis of the distribution of optical magnitudes of radio sources in the appropriate flux range.

---

VOSolution:

The availability of multi-wavelength data in many areas of the sky in the era of the VO means that simultaneous associations between several catalogues can be envisaged, given an accurate description of the properties of each, and a good models for, or empirical estimate of, the joint probability distribution of the set of attributes (e.g. flux measurements in different bands) to be used.

---

Discussion:

This may seem an ambitious goal, but the VO has to do something about associations, since it cannot function without them. This is one of the areas of VO development which most needs scientific thought, in addition to technical expertise for implementation.

This ScienceProblem is clearly related to GetGalaxyRedshift and SpectralEnergyDistributions: to obtain the SED of an object you have to be able to identify its appearance in catalogues produced in different passbands, and, similarly, for estimating the redshift of a galaxy...while, conversely, the probability distributions required for the likelihood ratio method require knowledge of galaxy SEDs, plus the relative abundance of galaxies of different SED types.

---

KeyReferences:

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 refinements.

---

-- BobMann - 12 Feb 2002