Steps in Quality Assurance:

24hr QA (cal scans):

Done when tape reaches IPAC, to judge photometricity of night. Scans falling in non-photometric intervals are immediately fed back to the observatory for re-observation.

Nightly Science QA (cal and sci scans):

Done when night is fully processed. Each scan receives a quality score that reflects the likelihood that those data meet the level-1 science requirements.

Quality scoring for scans

Scoring is turned around on a rapid timescale, <1 week after tape receipt.

Within this one-week time frame, we know which scans are of catalog quality and which need to be scanned again.

The scoring incorporates our acquired knowledge of how sky, telescope, and instrument conditions affect the data. In-depth details of the scoring process are given in the Scoring Rules Document. In a nutshell, there are five factors that are checked for every scan, and these are ---

(1) Photometricity

(2) Sensitivity

(3) Seeing and Focus

(4) Untracked Seeing

(5) Airglow Variations

(1) Photometricity

Are calibrator zero-points stable?: Plots at J, H, and Ks bands can show stable, photometric periods or unstable, non-photometric periods.

Does photometry in overlap regions agree?: Plots of overlapping sci or cal pairs can show good repeatability indicative of stable photometry or non-repeatability due to clouds.

Are background plots cloud free?: Plots of the mean frame backgrounds show normal three-band-correlated variations indicative of clear skies or strong, primarily Ks-band variations due to clouds.

Stability of zero-points and repeatability of overlaps are judged on a graduated scale, as outlined in the Scoring Rules Document, and used to score each scan. Quality=10 means the scan is as good as it gets and quality=0 means it clearly does not meet level-1 specs. Clear indications of clouds also downgrade scans to quality=0.

(2) Sensitivity

Is scan depth adequate for level-1 specs? The derived sensitivity score is detailed in Roc's memo. The sensitivity parameter has been empirically derived and is a product of the seeing shape times the background level to the 0.29 power.

Sensitivity is also judged on a graduated scale and converted to a score: quality=10 means the data have a 100% probability of meeting level-1 specs, quality=5 means they have a 50% probability of meeting specs, and quality=0 means they have no chance of meeting specs.

(3) Seeing and Focus

Are images round? Is the seeing below the maximum threshold for good star/galaxy discrimination?

A statistical study of seeing variations has been given elsewhere, but we now downgrade to quality=1 any scan with an average J-band seeing shape of 1.25 or above (FWHM > 3.5 arcseconds) or maximum J-band seeing shape of 1.30 or above. Furthermore, scans with images that are distinctly non-round (average image moment ratios smaller than 0.81) are given quality=1.

(4) Untracked Seeing

Is the software tracking the trend of seeing with time adequately so that point and extended sources can be distinguished? Most of the time, scans have well-tracked seeing, but occasionally the rapidity or extent of the seeing variations causes the scan to have untracked seeing.

If the seeing has not been accurately tracked over lengths exceeding 900 arcseconds, the scan receives quality=1.

(5) Airglow Variations

Are variations so large or erratic that they cause problems with extended source identifications? Extreme airglow variations can cause problems with extended source reliability.

A background noise statistic, which is the difference between the measured coadd background noise (after modelling large-scale gradients and structure) and the theoretical noise expected from the overall background level, has been developed to flag the worst cases of airglow. Such cases are downgraded to quality=1.

Assessment of these factors results in a score for each scan:

The lowest of the above five quality scores is the final score for the scan --

Those with scores of 5 and above are release quality.

Scores of 4 and below set the priority for rescanning.

Quality assessments are archived in the Scan Information Table

The Scan Information Table contains a record of every science scan observed by either of the 2MASS observing facilities. Each entry contains basic information on the scan (tile number, date, hemisphere, scan number, RA and Dec centers, etc.) and each of the quality assessments that went into deriving the scan's final quality score. Also given for each scan are the quality factors derived from each of the quality assessment values above, flags to indicate when those factors were overridden, and the final quality score of the scan.

This table is updated weekly to include the latest nights run through OPS processing. As such, it serves as the cumulative history of scan scoring over the course of the survey.

More detailed information on the content of this file can be found in its subsystem interface specification.