This page is for the CSC 2.0 release. Please see the CSC home page for the latest release.

This version has been superseded by the September 2017 release

Please see the September 2017 preliminary detections list (CSC 2.0 pd2).

The initial edition of the preliminary detections list (designated CSC 2.0 pd0) was made available on 2015 August 03, and includes 241,074 preliminary detections covering the range 0h-18h RA. After CSC 2.0 pd0 was made available, an issue was identified in the catalog source detection pipeline that may have resulted in incorrect parameters being determined for a few percent of the detections (and which may have affected positions for a very small number of detections). Users should carefully review the caveats listed below, and consider verifying critical detection properties independently if CSC 2.0 pd0 data are utilized for detection-specific studies. Statistical studies should not be affected significantly. All the impacted detections have since been reprocessed.

Preliminary detections

[Version: full-size]

Preliminary detections

The locations of the detections provided in the August 2015 CSC 2.0 pd0 Preliminary Detections List. Note that, as described in the caveats, the coverage of the Chandra archive is incomplete, with only a subset of ACIS and HRC-I fields included, and the display shows detections rather than sources.

Contents

• Data Access
• Column definitions
• Caveats

Data Access

The August 2015 release of the data is available at:

• preliminary_detlist.2015-08.fits.gz (37 Mb compressed and 83 Mb uncompressed)

It contains the results of Maximum Likelihood Estimator (MLE) fits to the candidate source detections. The column definitions and Caveats for this file can be found below. The file contains 241074 detections, with 163869 labelled as SRC_QUALITY = ‘TRUE’ and 77205 (32%) with SRC_QUALITY = ‘MARGINAL’. When split by EXTSRC_CLASS, there are 237892 detections with a value of ‘POINT’, 2716 (1%) with ‘EXTENDED’, and 466 (0.2%) with ‘POSSIBLE’. The breakdown by EBAND - the energy band correponding to the measured values - there are 186238 broad band ('b'), 195 ultra-soft ('u'), 17200 soft ('s'), 16876 medium ('m'), 18895 hard ('h'), and 1670 wide (HRC) band ('w') rows; the band definitions are the same as in release 1. There are 200 detections with STREAK_SRC_FLAG = TRUE.

Column Definitions

unix% dmlist preliminary_detlist.2015-08.fits blocks
 
--------------------------------------------------------------------------------
Dataset: preliminary_detlist.2015-08.fits
--------------------------------------------------------------------------------
 
     Block Name                          Type         Dimensions
--------------------------------------------------------------------------------
Block    1:                                Null        
Block    2: DETLIST                        Table        48 cols x 241074   rows

DETLIST HDU: PRELIMINARY DETECTIONS LIST

Caveats

1. This version of the preliminary detections table is constructed from a partial run of Chandra observations that will be included in release 2.0 of the CSC. In particular:

   • The table is very incomplete. Only a subset of ACIS and HRC-I fields in the range 0–18 hr RA has been processed to date.

   • Many fields have not yet competed processing or are awaiting manual quality assurance review (this is especially true for more complex or crowded fields).

2. The detections have not undergone complete quality assurance review, and may include detections that will be rejected by the final catalog quality assurance analyses, or that will be reprocessed to resolve the identified issues. In particular:

   • Detections identified near the edges of (stacked observation) fields, near edges of individual observations within observation stacks, near areas of extended emission, and on readout streaks, may have a higher than nominal false detection rate. If this is a concern, we recommend that users review images of the individual observations from the Chandra Data Archive (e.g. using the Chandra Footprint Service).

   • Although the concern is primarily for “faint” detections (those with SRC_QUALITY = ‘MARGINAL’), very bright (saturated or “piled-up”) detections will also be compromised. Note that detections that are identified as extremely piled-up (cratered) will have LIKELIHOOD = +Inf.

   • Some compact detections identified as definitely or possibly extended (EXTSRC_CLASS = ‘EXTENDED’ or ‘POSSIBLE’) may actually consist of a close association (typically a pair or triple) of point detections. These will be resolved in the final catalog release.

3. The preliminary detections table includes detections and not sources. In particular:

   • Sources that are detected in multiple observation stacks are not combined in any way. Each detection is reported separately.

   • Off-axis detections from one observation stack may correspond to multiple detections in another observation stack where the detections are on-axis. These matches will be resolved in second major catalog processing phase

   • As a consequence of the above source names have not yet been assigned. Instead, an identifier (given in the DETECT_ID column) has been provided for each detection.

4. Detections are categorized as ‘MARGINAL’ or ‘TRUE’ (recorded in the SRC_QUALITY column) depending on their likelihood evaluated by the new maximum likelihood estimator tool. These categorizations are preliminary and will change in the final catalog. In particular:

   • In the preliminary detections table the thresholds for determining whether detections are ‘TRUE’ or ‘MARGINAL’ (or false) are set somewhat conservatively, particularly for off-axis angles θ ≲ 5–8′. The final catalog is expected to include considerably fainter detections, especially for θ ≲ 5–8′.

   • The likelihood values reported in the preliminary detections table (the LIKELIHOOD column) are not normalized in any way.

5. Position errors are reported in the form of an error ellipse that records the statistical confidence limits derived from a Markov chain Monte Carlo analysis of the fitted source model. The position error ellipse does not include the absolute uncertainty in the astrometry relative to the reference frame. The latter will not be determined until the second major catalog processing phase. However, the systematic error derived from release 1 of the catalog (0.16 arcseconds radial) should be a reasonable approximation to the final value.

6. Fit amplitudes are provided as proxies for the full Bayesian aperture photometry analysis that will be included in the final catalog (see the photometry-comparison plot below). While these are expected to be reasonably reliable for relative photometry on a statistical basis, systematic uncertainties for a single detection basis may be large. Detailed Bayesian X-ray aperture photometry for each source will be calculated as part of the second major catalog processing phase.

7. Miscellaneous notes:

   • Detections near the edges of the field of each observation, or near the gap between adjacent back-illuminated and front-illuminated ACIS-S CCDs, or on bright ACIS readout streaks, are not included. This is intentional because these detector areas had an excessively high false source rate in release 1 of the catalog. In the final catalog release, the detection sensitivity will be set to zero in these areas. However, detection sensitivity information will not be available until completion of the second major catalog processing phase.

   • Only point and compact detections are included. Extended (“convex hull”) detections are not included in the preliminary detections table, but will appear in the final catalog release.

   • The maximum likelihood estimator fits are incorrect for a small percentage (~0.4%) of detections. The properties for these detections will be poorly determined and the actual errors will likely exceed the reported confidence intervals. Unfortunately, there is no simple way for the user to determine which detections in the preliminary detections table are affected at this time.

   • A (very) small number of (bright) detections on ACIS readout streaks may be missing in the preliminary detections table.

   • Calibration of confidence intervals is preliminary. The reported confidence intervals may not exactly match the specified limits (typically 68% or 95% confidence limits).

Comparison of source photometry

[Version: full-size, PDF]

Comparison of source photometry

Correlation between the fitted source amplitude and detailed Bayesian X-ray aperture photometry analysis for 14,360 detections, in the ACIS broad energy band. Note that the correlation is somewhat worse for faint detections categorized as ‘MARGINAL’. The correlation has not been demonstrated for energy bands other than the ACIS broad band; however, we expect that the correlations in the other bands should be similar.