For each detected X-ray source, the Chandra Source Catalog will list the source position and a detailed set of source properties, including commonly used quantities such as source dimensions, multi-band fluxes, and hardness ratios. In addition to the traditional static catalog elements, the Chandra Source Catalog will incorporate "data objects". These are dynamic quantities that can be manipulated interactively by the catalog user (for example, by applying spatial or spectral filters), and include source images, event lists, light curves, and source spectra. Relevant calibration information will also be available.

The catalog and interface will provide simple and quick access to X-ray data for individual sources with good scientific fidelity. They will directly support medium sophistication scientific analyses. In addition, easy searches and analyses of a wide range of statistical properties of classes of X-ray sources should be facilitated. The catalog should prove invaluable for identifying areas of interest and future study, as well as preparing observing proposals and planning observations.

Chandra Source Catalog users will be able to query the static catalog data directly, and retrieve data objects for the selected sources for further analysis in CIAO. Additional queries on the data objects themselves can be performed using the CIAO scripting language, enabling a degree of query sophistication that is not possible with traditional catalogs containing purely static table data.

As an example, one could query the static catalog data to search for ACIS observations of all non-point sources with an extent less than, say, 10 arcsec diameter and a hardness ratio higher than some threshold, and retrieve the event list data objects for all of the matching sources. The investigator could then write a CIAO script to extract for each source all X-ray events falling outside of a region proportional to the size of the local point spread function, derive statistical properties for the extended Xray emission, and search for specific signatures in the data.

The Chandra Source Catalog will be subject to continual improvement. As observations become public, they will be processed and the detected sources and source properties will be added to the catalog. This will ensure that the most up-to-date information on any observed source is always available. Because of the scientific advantage of having a homogeneous catalog, we may from time to time also release snapshots of the catalog that have been processed in a uniform manner. The timing of such releases will depend on the trade-off between the improvements in the processing quality and the resources required to redo the catalog.

The Chandra Source Catalog will be realized by processing all science observations through a set of "Level 3" standard data processing pipelines. Unlike the Level 1 and 2 pipelines that create the science data products distributed to the observers shortly after an observation is completed, the Level 3 pipelines will be run only after the proprietary period has elapsed and the data have become public. The outputs of the Level 3 pipelines will be used to populate the Chandra Source Catalog.

A key difference between Level 1 and 2 data products and Level 3 data objects is that the former are constructed for the entire field of view on a per-observation basis, whereas the latter are constructed on a per-source basis and (depending on the specific data object) may incorporate information from multiple observations.

Level 3 processing will be split into a pipeline that populates observation interval-specific static data and data objects in a per-observation interval source list, and a pipeline that merges data from all of the observation intervals in which a source is detected into a master source list.

In the initial implementation, the merge pipeline will populate the master source list by selecting the best information available from all the individual observation intervals, using some yet to be determined method. Ultimately, however, the source properties in the master source list will be determined by simultaneously fitting data from all of the available observations.

Also in the initial implementation, the Level 3 pipelines will focus on deriving detailed source properties for point source and compact (spatial extents of a few arcseconds at most) objects. Accurate characterizations of extended sources will be added in a future software release.

At the time of writing, we have developed a prototype Level 3 pipeline that is being used to evaluate algorithms and techniques. A first version of the production Level 3 pipelines will be incorporated in our Automated Processing system on the Spring 2004 timescale. We expect to begin production of the Chandra Source Catalog in the Fall, with full support for extended sources and the final merge pipeline about a year later.

Ian Evans for the Chandra Source Catalog project team