Appendix 3. Long Exposure (6x) Scan Databases, Catalogs and Images

5. Data Processing

a. Modifications Relative to Final Survey Processing

Reduction of the 2MASS Long Exposure (6x) scan data was carried out using a minimally modified version of the 2MASS Automated Production Pipeline (2MAPPS v3.0) that was used for final processing of all of the primary survey data (see IV.1). The basic processing steps shown in Figure 1 of IV.1 remained unchanged from the final survey processing, and the pipeline output products from 6x processing were very similar in form to those from the survey data processing.

The modifications to 2MAPPS for the 6x processing fell into three categories and priorities. Highest priority was given to the fundamental changes that were necessary so that the 6x data would successfully process without software failures, such as corrections dealing with the longer exposure times and resulting higher sky background count levels and structure. Second priority was given to changes that fixed software bugs that were known to exist 2MAPPS v3.0, or that were discovered since the release of the All-Sky Catalogs. Finally, modifications to the software and/or the quality assurance process that further improved data quality and brought output data formats closer to those of the release products were made as time and resources permitted.

Brief summaries of the major modifications made to each of the data reduction subsystems are given below. A link to the pertinent section of the main survey supplement is given for each subsystem. Further descriptions of the changes made to the astrometric and photometric calibration procedures, extended source processing and artifact identification for the 6x data are given in Sections A3.5.b, A3.5.c, A3.5.d, and A3.5.e, respectively.

As with the main survey, 6x data processing was done in two steps. Preliminary processing was carried out shortly after the 6x observations were made using a 2MAPPS version that was modified only to accommodate the longer exposure times of the 6x data. The results of the preliminary reduction were used to assess the basic 6x data quality for observatory support and to prepare for the final processing. Final 6x data processing was carried out as part of the 2MASS Extended Mission, and incorporated best practices learned from the survey itself, as well as analysis of the preliminary 6x processing results. Because final 6x processing followed the release of the final survey data products, it was possible to incorporate Quality Assurance steps that compared the 6x products to the well-characterized All-Sky Catalogs and Images.

i. Synopsis of Reduction Pipeline Changes By Subsystem

Instrumental Calibration (IV.2)

• Generate new bias/dark exposure frames using longer R2 exposure data. 6x processing used the same responsivity (flat-field) images and saturation threshold levels as were used for the main survey.

Atlas Image Generation (IV.3)

• Blank frames containing very bright stars to avoid large scale background matching errors all along the scan. This blanking was applied to only a small number of scans containing extremely bright stars, such as that covering R Doradus in the 6x LMC scans. The result of this blanking is a slightly lower depth-of-coverage in the region of the Atlas Image immediately surrounding the star.

Point Source Photometry (IV.4.b)

• Adjust variance maps of existing point spread function templates to produce profile-fit photometric uncertainties consistent with repeatability analysis. This modification removed the need for the corrections to the uncertainties that were used during processing of the main survey and calibration data. Point source photometric uncertainties are called [jhk]_msig in the 6x point source database, and [jhk]_cmsig in the survey and calibration databases to reflect this difference.

Position Reconstruction (IV.6)

• Use the 2MASS All-Sky Point Source Catalog as the primary astrometric reference for 6x position reconstruction. This modification is discussed in A3.5b.
• Correct the software bug (I.6.b.xiii) that passed on READ1 source detections for bandmerging before frame distortion corrections could be applied during main survey and calibration scan data processing (see A3.5.b). Bright source astrometric residuals should be slightly smaller for 6x point sources than for the PSC.
• Account for frame position errors in absolute position uncertainties by computing residuals with respect to the 2MASS PSC positions in 0.5° long segments within 6x scans, and adding in quadrature to the relative position errors generated during the bandmerging process. For survey data processing, this was accomplished using position offsets computed in the overlap region between adjacent scans. Overlap scan information was not always available for the 6x data, and the higher density of PSC reference stars provides a more robust estimate of the absolute errors with respect to the reference frame.

Optical Associations (IV.4.f)

• Correct software bugs (I.6.b.xvi that resulted in missed associations between 2MASS point sources and USNO-A and Tycho 2 sources, and incorrectly setting the associated catalog identifier (a) to "0" when the separation between 2MASS and optical source was larger than 5".

Extended Source Processing (IV.5)

• Modifications to the extended source processing for the 6x data, and examples of its performance are discussed in A3.5.d.
• Modify magnitude-dependent and SNR-dependent thresholds to enable measurements of fainter objects.
• Adjust confusion noise and stellar number density thresholds to accommodate deeper exposures.
• Tune background fitting to compensate for higher background levels and increased contrast in apparent structure.
• Increase the effective masking radius size around bright stars to compensate for brighter instrumental magnitudes.
• Replace bright galaxy and star seedlists with 2MASS All-Sky XSC and bright star list drawn from the PSC.

Artifact Identification (IV.7)

• Adjust magnitude thresholds and geometric artifact identification parameters to account for longer exposure times and resulting larger instrumental source brightnesses. A summary of the revised parameters is given in A3.5.e.

Photometric Calibration (IV.8)

• Modify Read_2 instrumental zero point magnitudes to compensate for longer exposure times.
• Develop new secondary standard star network in the special LMC and SMC calibration fields used to support the 6x observations of those regions. The 6x calibration strategy is discussed further in A3.5.c.

Quality Assurance (IV.10)

• Implement photometric and astrometric comparisons between the 6x source lists and the 2MASS All-Sky Release PSC and XSC to validate photometric accuracy, position reconstruction performance, completeness and reliability of the 6x processing.

[Last Updated: 2007 January 8; by R. Cutri]