Several useful points arose during the discussion of this issue. First, there appears to be no need for GALWORKS or POSPTS to output a point-source record format different from that of BANDMERGE. This allows the possibility for BANDMERGE to have a mode in which it inserts pseudo-bandfill information so that downstream processors can execute. But POSPTS does not appear to need bandfill information; it will probably just replace the position coordinates in U-Scan with refined coordinates in J2000. It is also not clear that MAPCOR needs for POSPTS to operate on the point sources for the testing of MAPCOR. Finally, it emerged that MAPCOR probably does not even need any pseudo-bandfills to be inserted by BANDMERGE.
The original problem therefore probably does not exist, but the discussion it stimulated yielded fruitful ideas for refining the details of the BANDMERGE output file defined by the SIS BMG01. The baseline approach will be for undefined fields such as PSFMag (PSF-fit magnitude from PROPHOT) for unobserved bands to have values of 999 inserted as flags. The same will be done for the uncertainties. Further details were left for the cognizant engineers to work out.
================== ACTION ITEMS FROM THE POINT SOURCE CDR ==================== Action Item/Lien List From March 1996 2MASS Science Team Meeting and Point Source CDR This is a compilation of the existing liens from SDSs, liens and recommenda- tions specified during the point source CDR and Team meeting, and action items to be undertaken during coming months during final development of the subsys- tems. Items that are considered highest priority - that can impact subsystem design and therefore should be satisfied immediately - are indicated by asterisks. Many of the listed items relate to actions that can only be taken during survey telescope/camera check-out. ------------------------------------------------------------------------------ I. All 2MAPPS Subsystems: a. Key on making end-to-end integration of all subsystems possible. This is enabled primarily through completing all Subsystem Interface Specification documents (SISs). b. Begin formulating a list of parameters that will come out of each subsystem that should be archived over life of survey for trend analysis and system health moniotring. c. Establish diagnostic output that will interface with QUALITY subsystem to monitor scan processing for each subsystem. Output can include parameters, flags, plots, etc. that indicate success and accuracy of processing. d. Summarize all "tunable" parameters in each subsystem. i. Develop plan to tune each parameters during telescope/camera check-out period. ii. List of parameters that must be available from telescope, camera, operating system, etc. Make sure that these will be available, and updated as necessary. e. Compile list of parameters, etc. that must be reset if survey camera is adjusted. (i.e. band alignment). II. PIXCAL/DFLAT: *a. Identify and correct artifacts in sky offset frames due to persistence. *b. Identify and react to incorrect sky offset correction due to confusion noise. c. Consider analytical model of bias residuals. III. FREXAS: *a. Import aperture photometry model from PROPHOT. b. Improve saturated star position and position error model. IV. POSMAN: *a. Run prototype POSFRM on repeated scans in Coma region. Examine repeatibility and stability of solutions, and extrapolate performance to 3-band case. *b. Repeat a. for a high density region (i.e. the MSX field). c. Update POSMAN SDS. d. Use absolute minimum positional error (weighting) for brightest stars (i.e. don't allow error to go to zero). e. Design plan to map/measure optical focal plane distortion during pre-survey check-out. Monitor over life of survey for changes. f. Plan to test stability and systematics of position reconstruction solutions versus a large range of parameters during telescope-check out. g. Feedback camera flexure information into POSMAN design. V. PICMAN: *a. Decide on interpolation kernel size. b. Consider cosmic ray and any other single frame artifact blanking in coadds. VI. PIXPHOT/FIND: *a. Further verify that completeness is degraded by using DAOPHOT/FIND to locate sources in R2-R1 images. *b. Implement seeing-dependent FWHM and local noise thresholds in FIND step. c. Verify that new coadd interpolation algorithm does not adversely impact point source detection efficiency. d. Finalize position error model. VII. SEEMAN: a. Make sure that seeing estimates are not driven by statistical fluctuations due to uncertainty in parameter estimation (i.e. put a "penalty" on changing the seeing estimate). VIII. PROPHOT: *a. Solidify deblending algorithm: - Examine chi-square values for real pairs. - Simulate binary/trinary fields and determine reliability of deblending under various seeing conditions. b. PSF grid: - How big should this grid be to avoids sampling error in photometry? - Develop plan to populate the grid. c. Document limits on all parameters; necessary precursor to survey check-out testing. d. Further document photometry and profile-fit algorithms. e. Make sure that all parameters needed by MAPCOR and BANDMERGE to verify source reliability are generated and passed. f. Implement cosmic ray reject in sky-annulus. g. Demonstrate that psf is uniform across focal plane during survey check-out. If it is not, then must build in distortion model. h. Develop plan to monitor stability of psf's over life of survey. IX. MAPCOR: a. Develop plan to set key parameters during survey check-out: - persistence and other artifacts. - R1 positional and photometric offsets. - "standard" aperture for aperture correction. - aperture corrections and normalizations. b. Specify the convergence criteria for measuring various corrections. Make sure that the "deltas" are small since photometric errors will be of comparable size. c. Are look-up tables of photometric corrections as function of psf FWHM sufficient in dense fields? d. Develop plan to build up statistics of photometric and artifact corrections over life of survey. Monitor for changes. e. Understand differences between measurements of internal and external reliability and completeness. f. Make provisions for handling sources that are saturated in R1. g. Clarify in SDS how intensities of artifacts that are saturated on R2-R1 but not R1 are computed. h. Generalize glint corrections. h. Generalize glint corrections. i. Make diffraction spike corrections wavelength dependent. X. BANDMERGE: *a. Examine performance of a simple positional bandmerge algorithm (i.e. closest neighbor) over a range of source densities. b. Incorporate physical model of star multiplicity into matching algorithm rather than Poisson model. c. Make sure to use blend and artifact flags provided by PROPHOT and MAPCOR.