R. Cutri introduced Bill Wheaton to the 2MAPPS Design Team and announced that Bill will be working on a variety of open issues, possibly including the task of generating the Point Spread Function accumulation software needed to build the library of PSFs for the survey.
R. Cutri reported that the installation of the telescope base was currently underway at the northern-hemisphere observatory.
R. Cutri reported that he is beginning work on the Operations Plan. He requested input from any team members that had suggestions. Specific guidelines are needed for governing procedures beginning with receipt of data tape through certification of the night's processing.
J. Fowler requested clarification on whether there was a team-adopted standard file naming convention for files above the scan level. Specifically, it's time to deliver the SIS for the DARKS Subsystem input to the QUALITY subsystem. Since DARKS operates on a multi-scan night level, the usual scan-number naming convention does not apply. The observation date is clearly needed as part of the file name, but whether the hemisphere ID was also involved was not clear, since the file could spend its entire life underneath the hemisphere branch point in the directory structure.
The team decided that no strict policy was needed, but when in doubt, the hemisphere ID should be included as part of a night-level file's name. The DARKS input to QUALITY will therefore use the name template qDATEHEMI.darks (e.g., q970109n.darks).
J. White has completed the task of reformatting the 3-channel lab data to make it look like it was received in accordance with the Rev. J Observatory/IPAC Interface Control Document. He also revised the existing test script for DARKS to allow DARKS to process the new data. This allowed J. Fowler to run the data through the DARKS subsystem. This is a brand-new (not yet officially delivered) subsystem, and it uses thresholds to determine whether dark scans and twilight-flat scans are acceptable relative to what is expected on the basis of previous "canonical" darks and responsivities. Since the new data come from a different camera, they appeared seriously out-of-spec relative to the protocam canonical MRD files (masks, responsivities, and darks), with the result that DARKS turned off quite a few pixels in the generated masks (almost all of them in K band). Nevertheless, new darks and responsivities were generated and can be scrutinized. DARKS required a relaxation of the requirement on the dynamic range of the twilight-flat frame median before it would accept the twilight flat scans for generating new responsivities. This is one of many currently untuned parameters in DARKS.
B. Light reported that several attempts to resolve the anomalous behavior of the aperture-magnitude curves of growth have failed to remove the problem of nonconvergence of the aperture magnitudes. Most of the original residual delta-mag/aperture-increment was removed by the switch from a mode-estimator to a median-based noise estimator, as reported in last week's minutes. But the small remaining residual is resisting removal.
One attempt to make progress was to limit the brightness of the stars included in the study; editing out the brightest stars did not solve the problem. The same result was obtained from using a smaller sky annulus and allowing the sky annulus to overlap the aperture. Suspicion has now fallen on the DAOPhot algorithm itself, since the residual is actually rather small (about 0.005 mag), prompting questioning of whether this detailed an inspection of such curves of growth has really been undertaken and whether this amount of residual is really very important. B. Light agreed to note the unresolved nature of this study in an SDS lien. In the meantime, he will have to move on to higher priority work (see next section).
B. Light reported that the implementing and testing of deblending in PROPHOT has been successfully achieved for certain restricted types of blended point sources (so far, only blending in the cross-scan direction has been simulated and tested). For flux ratios of 3-to-1 and smaller, and with only Poisson photon noise added to the simulated data frames, PROPHOT consistently deblends sources as close as 1.1 camera pixel. Tests with more noise and different blending directions will be undertaken next.
H. McCallon reported that the first scan position reconstruction employing the Tycho catalog has been performed. The results are excellent, with the standard deviation of the residual scatter being less than 0.1 arcsecond. The scan used was s022 of 950503, an M67 scan.