I. Two-stage data processing strategy for 2MASS motivated by quality and early data release requirements

• Preliminary processing during data acquisition



• Produce high reliability, but possibly non-uniform products


• enable science by the astronomical community as early as possible
• gauge progress of survey
• monitor health of system
• mission and observation planning for SIRTF, SOFIA, etc


• Processing pipeline designed to evolve over the life of the survey


• Final uniform processing in the year following the end of data acquisition



• Produce the 2MASS legacy data products that meet all of the Level 1 Science Requirements including uniformity



• Process all survey data with the same, stable software system



• 2MASS Production Pipeline System (2MAPPS v3.0) will:
• Fix known deficiencies in preliminary processing
• Incorporate knowledge of system performance, observatory site conditions, atmosphere and IR sky accrued over the life of the survey
• Improve pipeline functionality with external databases and new algorithms not previously available
• Employ global photometric and astrometric solutions from overlapping observations

II. Data Processing Status

• Preliminary data processing complete for >99% of collected data
• Working Source Databases contain ~10⁹ point and 4x10⁶ extended sources


• Three preliminary data releases complete
• Image and Catalog products covering 48% of the sky in the public domain


• Final processing pipeline (2MAPPS v3.0) in development now

III. Examples of Known Non-Uniformities in Preliminary Data Products

• H-band detection threshold sensitivity to background structure due to OH airglow



• Detection limit raised <1 mag in ~5% of cases

Fig III.1 - Surface density map for points sources with H<15.1 mag from Second IDR
• Noise estimation algorithm that is insensitive to low frequency BG structure delivered after several months data processed


• Photometric star network evolution over survey



• Early survey data calibrated using nightly solutions from 1 standard star per calibration field, using photometry from external source (Persson et al. 1998)
• Repeated observations of calibration fields during survey used for continual development of extensive secondary star network with internally consistent photometry (using global chi-squared minimization techniques)
• Variables identified and eliminated
• Photometric calibration will change by as much as 7% for some nights using improved secondary network


• Spatial photometric biases in high source density regions



• Fig III.2 - Mean H-K_s color for point sources in 5x5 arcmin bins (full color-range 0.2 < H-K_s < 0.7)
• Spatially-correlated biases up to 0.2 mag (full color-range in image is 0.2 < H-K_s < 0.7)
• Can lead to extinction estimation errors of ~3.0 mag (A_V=15.9E_H-K)
• Caused by confusion in aperture curve-of-growth determination
• Correct in final processing using survey-average curve-of-growth correction rather than empirical determination


• Spatial dependence of astrometric uncertainties



• Fig III.3 - Amplitude of position error ellipse major axes for 10.0 < K_s < 10.5 point sources in 5x5 arcmin spatial bins (full range 0.13"-0.5")
• ~30% poorer astrometric precision at high latitudes and at end of Survey Tiles
• Caused by astrometric reference star density and Tile-based processing strategy
• Correct in final processing using higher density Tycho 2 catalog and by globally tying astrometric solution across Tiles