XrayGalaxyClusterSurvey < VO

VO Web>XrayGalaxyClusterSurvey (2002-08-07, AnitaRichards)EditAttach

ScienceProblem: XrayGalaxyClusterSurvey

ScienceGoal:

To produce a large redshift survey of galaxy clusters, selected using serendipitous detections in archival X-ray data and confirmed using optical/near-IR imaging.

The primary dataset here is the archive from ESA's XMM-Newton X-ray mission, while the optical/near-IR imaging archives to be used include those from the SuperCOSMOS and Sloan sky surveys, as well as, in the future, UKIDSS and VISTA.

ProblemDescription:

Clusters of galaxies are of interest both because they are intriguing objects themselves, and also because they are useful probes of cosmology and drivers of galaxy evolution. A major new galaxy cluster survey would therefore be an important new astronomical resource, which would provide the starting point for many research programmes. For a variety of reasons, it is preferable to select galaxy clusters in the X-ray passband, rather than in the optical, and, as argued recently by Romer et al. (2001), an ideal opportunity to produce such a survey is provided by the advent of XMM-Newton, and the archive that will store all the data taken during its ~10 year lifetime.

In essence, this ScienceProblem is a description of the survey (called XCS), proposed by Romer et al. (2001). This entails searching for clusters in every image taken by XMM's EPIC cameras in a field suitable for serendipitous cluster detection, obtaining optical/near-IR imaging data for cluster candidates with and then getting spectra for a sufficient number of galaxies which are members of a confirmed cluster to yield a reliable redshift for that cluster. The resultant cluster survey database can then be used as the basis for many science programmes, some of which are outlined by Romer et al. (2001).

CurrentSolution:

The XCS has only just started operations, and the only XMM data publicly available to date are those from Calibration/Performance Verification observations, which may be downloaded from an ESA WWW site. Soon, public data will start to become available from the XMM Science Archive (XSA) at ESA's Vilspa data centre in Spain: the prototype XSA system is based on the ISO Data Archive (also located at Vilspa), which can be interrogated using a Java-based query tool.

In both cases, the selection of useful observations has to be made interactively, and the subsequent download of either raw data or pipeline-generated data products must also be triggered manually. Once the data files have been downloaded, the user must then either search the catalogues of sources extracted by the pipeline processing for extended sources (which are candidates for being clusters) or run his/her own analysis pipeline to generate a list of cluster candidates.

For many purposes - such as the derivation of constraints on cosmological parameters from the evolution with redshift of the abundance of clusters - the most interesting clusters are very massive ones lying at high redshift (z>1), so one focus of the XCS will be to filter its list of cluster candidates to prioritise those which are likely to be massive high-z clusters for follow-up observations. In principle, a good way to determine which cluster candidates are likely to be the most massive is to look at the low resolution X-ray spectra that can be constructed from the same photons that led to their detection in an EPIC image. In practice, however, it is very difficult to distinguish between a high-temperature (and, hence, massive) cluster at high redshift and a local low-temperature (low-mass) cluster or group of galaxies: in the absence of significant detections of iron lines in the X-ray spectrum, its continuum shape only constrains the cluster's apparent temperature, defined by T/(1+z), where T is the true temperature of its intracluster gas (assumed to be isothemal) and z is its redshift.

So, an important stage in the XCS data analysis chain will be to try to distinguish which objects of a given apparent temperature are local groups and which are massive high-z clusters. This can be done using optical data, since local groups should be detectable as (albeit fairly low amplitude) overdensities in the distribution of galaxies on the sky in shallow optical imaging data, such as those derived from scanning Schmidt plates, while high-z clusters will not be detectable to the limiting depth of the plates. So, one way to weed out the groups is to cross-correlate the positions of cluster candidates with a map of the projected galaxy distribution in the same area of sky, smoothed on an appropriate scale, and flag as likely groups those cluster candidates that do match overdensities in the smoothed galaxy distribution. Currently, this can only be done interactively, and rather indirectly, by obtaining an image of the appropriate region of sky from an image server, such as Aladin, or use GAIA to display a SuperCOSMOS image, and looking to see whether there seems to more galaxies than usual populating the field surrounding the position of the cluster candidate.

VOSolution:

In principle, the VO can simplify a number of the steps in the creation of a cluster catalogue like XCS - indeed, it will have to do so, if it is to be possible to make the catalogue production sufficiently automated that more attention can be directed towards the scientific exploitation of the growing catalogue.

Firstly, the VO should be able to automate the selection and download of archival datasets that satisfy the criteria for consideration by the XCS. Selection criteria can be expressed in terms of the metadata describing the obseravation - e.g. how long was the integration?; was it taken in a field with sufficiently low Galactic absorption (required for a good spectrum)?; was the target of the original XMM observation a known cluster (in which case the observation is not appropriate for inclusion in a serendipitous survey)? - and it should be possible either to run a cron job that queries the XSA using these criteria once a week, say, or to post a query to the archive such that the user is notified (and/or the desired set of files prepared for delivery to the user) whenever a dataset matching the selection criteria becomes publicly available in the archive.

Secondly, if the serendipitous detection and characterisation of cluster candidates requires processing beyond that performed by the standard pipeline, it should be possible for the user to upload code onto a computer system hosting a copy of the XSA so that it could be run automatically on data files for each suitable observation that enters the archive. It is likely, of course, that this level of persistent processing capability would only be granted to a limited number of projects, and would probably have to be restricted using some sort of resource quota system, but it should certainly be made available for a small set of important archival research projects.

Thirdly, it should be possible to automate the cross-correlation of cluster candidates with catalogues of groups and/or clusters of galaxies selected from optical/near-IR databases. These might be made available as prepared datasets for the major databases (e.g. the SDSS) or might be generated on-the-fly (although that would be at considerable computational expense). In this way, in an era when relatively deep galaxy catalogues exist for a significant fraction of the sky, it will be possible to use archival imaging data to confirm cluster candidates, and to produce more accurate positions for their members, to use as the basis for determining the cluster's redshift - either by searching spectroscopic databases for existing spectra for cluster members, or drawing up target lists for follow-up spectroscopy, or deriving photometric redshift estimates from the galaxy catalogue data themselves.

KeyReferences:

Romer A.K., Viana P.T.P., Liddle A.R., Mann R.G., 2001,ApJ, 547, 594

XCS Home Page

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 - 09 Feb 2002

Associated UseCases