I. Overview

2MASS final data processing must be completed in less than one calendar year to allow sufficient time for final product generation and validation. The volume of data to undergo final processing, coupled with the runtime of 2MAPPS v3.0 will dictate if the job can be done in under one year.

At minimum, final processing must include at least one scan per survey Tile on the sky, or a total of 59,650 scans. There are a significant number of Tiles with multiple observations, though, for the purpose of improving quality, validation and special science observations. There are also a number of special Tiles that were observed to cover gaps between the scans of nominal Tile positions. Some or all of these extra observations will be necessary and/or useful to include in final processing.

II. Bookkeeping

The most straightforward way to select scans for final processing is to use the quality score assigned during preliminary 2MASS processing. A quality value of 5 or above is the nominal cutoff for "catalog-quality." As a reminder, only data from scans having Q>3 were loaded into the working source and image databases.

The table below lists the number of scans as a function of the quality value "Q" assigned during quality assurance review. Also shown is the cumulative number of scans having quality >Q. These statistics were derived from the Scan Information Table (SIT) dated 001109, and are slightly out of date by now. Note that the number of scans assigned quality Q=0 is a lower limit because it includes only those scans downgraded to 0 after-the-fact. Most Q=0 scans are assigned quality 0 during preprocessing triage and are not tracked in the SIT file.

Table 1 - Number of Scans as a Function of Quality

Q    N(Q) N(>=Q) 
-----------------
10  52590  52590 
9    3156  55746 
8    4193  59939 
7    2172  62111 
6     480  62591 
5    2663  65254 
4     463  65717 
3    3574  69291 
2     269  69560 
1    7364  76924 
0   [8815][85739]

All Survey Tiles have at least one scan of catalog quality (Q>5), and there are a total of 65254 such scans. This is 109% coverage of the sky. There are 69291 scans with Q>3, 116% sky coverage, loaded in the working source and image databases.

III. Proposal

Include in final processing all scans with quality Q>3 from preliminary processing quality review. This is a small increment (16%) over the minimum number, and offers maximum flexibility in selecting the scans to be included in the final catalog generations steps. It also affords some margin for quality improvement from new processing and calibration algorithms. If total runtime becomes an issue after firm timing tests are in hand for 2MAPPS v3.0 (mid-December 2000), then we can revisit the inclusion of the Q=3 and 4 scans.

IV. Issues

1. Early in preliminary 2MASS processing, the minimum allowable photometric interval was defined by 3 calibration sets. Two cal-set intervals were were downgraded to Q=0. Later, some two cal-set intervals were assigned at Q=3. Do we revive Q=0 2-cal intervals? All scans in two cal-set intervals have been reobserved with Q>5.



2. When zero points were fit for multiple calibration intervals in preliminary 2MAPPS v2.x processing, we insisted that the periods could not be contiguous. That is, any given calibration set could not serve as the endpoint for two different zero point solutions, either constant or linear fits. The quality of survey scans in at least one intercalibration period separating the endpoints of the two intervals was set to Q=0. In final processing, we are considering higher order zero point fits and/or allowing contiguous calibration intervals. Should the survey scans previously set to Q=0 in the gaps between the old calibration intervals be revived?

V. Operational Considerations:

1. If the decision is made to include in final processing scans with Q<3, those scan must be added to the working databases so that certain key pre-production tasks can be run (e.g. generating the "super *.fpos files for astrometric global chi-squared minimization).



2. Final processing requires a priori knowledge gleaned from 2MAPPS v2.x processing. If scans previously triaged before processing are to be revived for final processing, they must be run through the preliminary pipeline and then loaded into the working databases.