The Chandra Data Archive (CDA) is depository as well as source of information for all Chandra-related activities (Proposal submission and management, Mission Planning, Automated Processing, Data Retrieval). The observation catalog (ObsCat), embodied in the database axafocat plays a central role in this.
axafocat, with the help of related databases:
For more information on the interfaces, see Estes et al. (2000).
The CXC Automated Processing system generates data products in three main levels (Levels 0, 1, 2) and two sublevels (Levels 0.5, 1.5) of processing. All products are archived and divided into three categories: primary, secondary, and supporting products. The primary products contain all (science) Level-2 products and some crucial science Level-1 products. The secondary products contain the remaining science Level-1 products, Level-1 aspect products, and Level-1 orbit ephemeris. The remaining products are classified as supporting. Primary and secondary products are shipped to our users automatically. Individual supporting products are available upon request.
The basic data product (file) retrieval keys for observational data are:
The vast majority of the data products that the user will encounter are FITS files; for these the metadata are extracted from the headers. In addition, the archive holds raw binary files, with accompanying log files for the metadata, and is equipped to deal with other standard formats, such as PostScript, PDF, GIF, JPEG, etc.
A FITS data product's type is identified by its CONTENT keyword. The list of recognized CONTENT values has, in principle, a flat structure, but for most querying activities a hierarchy is imposed by grouping them into instrument/subsystem and level-of-processing categories.
Starting with the HEASARC FITS Working Group guidelines and conventions, we have developed an extended set of conventions to which all FITS files produced by the CXC are supposed to adhere.
FITS files consist of one or more Header-Data Units (HDU), each containing a header and a data part. For all practical purposes, an HDU may be one of the following:
All FITS files are to be checksummed: the 32-bit one's complement checksum over each HDU (and therefore over the entire file) must be -0 (all ones). Hence, first order data product integrity may be ascertained by verifying the checksum.
The FITS headers are required to have a number of standard components:
Although we will continue the practice of sending proprietary data out to Guest Observers on physical media (4-mm DAT tape, 8mm Exabyte tape, CD-ROM), we expect that the bulk of our data transfers will happen through our web interface, for public as well as proprietary data.
The CDA Search and Retrieval interface is discussed by Estes et al. (2000). It will enable the community to download any set of CDA products, public or proprietary (provided one has the proper authorization), or stage the data for anonymous ftp transfer.
In the meantime, we have implemented a provisional web interface that allows retrieval of primary and secondary products for public observations that have been approved by the CXC Verification & Validation process.
Time is kept in FITS files as MET: seconds since 1998-01-01T00:00:00 TT. It is tied to absolute time by the keyword MJDREF (50814.0). At that epoch, the difference with UTC was: TT-UTC = 63.184 s; by launch: TT-UTC = 64.184 s.
We have adopted the FITS formats developed in the context of the RXTE
mission for orbit ephemeris files and the JPL DE-200 and DE-405 solar
and planetary ephemerides. The latter come with an ANSI C interface,
HEASARC's cfitsio, that is functionally equivalent to the JPL Fortran software. Consequently, the RXTE multi-mission software package for time conversions, barycenter corrections, and pulsar phase binning (including absolute timing and phase-resolved spectroscopy) can be used for Chandra data after only slight modifications.
Estes, A. N., DuPrie, K., Harbo, P., Pak, K., & Zografou, P. 2000, this volume, 457