IPAC 2MASS Working Group Meeting #59 Minutes

IPAC 2MASS Working Group Meeting #59 Minutes, 4/04/95

ATTENDEES

C.Beichman, R.Cutri, T.Evans, J.Fowler, G.Kopan, B.Light, S.Wheelock, J.White

AGENDA

  1. Prototype Operations Software (POS)
  2. 2MAPPS Schedule Review
  3. Time-Variable PSF Status
  4. Alternate Flattening Software

DISCUSSION

  1. Prototype Operations Software (POS) -- R. Cutri reported that the version of the protocam pipeline needed for the April '95 data is supposed to be ready for a test run by 17 April (changed from 15 April, which is a Saturday). This pipeline is denoted POS (Prototype Operations Software) to distinguish it from previous versions. It integrates the Read2 and Read1/Merge capabilities, employs a modified directory tree structure, and includes a modified front end for the latest scan header input. It will soon include G. Kopan's new code for computing a measure of the seeing for use in KAMPhot (see attachment), a modified version of KAMPhot that uses this to select a PSF model, and a pipeline version of the galaxy processor. The first good night's data from the April 20 run will be shipped to IPAC via Federal Express for immediate processing. This should suffice for refining POS parameters (e.g., in StufIt).

    The new machine on which POS will run, lugosi, had arrived on the loading dock as of meeting time (at least 728 pounds worth). J. White reported that the current version of POS (known as bigpipe) has been running scan 057 of 940601 (a Ks scan of M92).

    R. Cutri reported that the recent observing run on Mount Lemmon had several nights of sub-arcsec seeing, and good photometry and images of more than 70 galaxies were obtained. These were galaxies previously detected by the galaxy processor. Approximately half of the galaxies were observed with an image resolution of about 0.9 arcsec/pixel, with the remainder at about 0.3 arcsec/pixel. These observations will allow good truth tables to be generated for use in the development and testing of the galaxy processor.

    R. Cutri also reported that source lists produced by the POS runs up to the point just prior to band merging will be made available in ASCII format for limited access via the web. New source lists will be added as each night's processing is completed. Analysis tasks will begin as soon as the data they need are available. By 1 August, the first pass through the new data should be complete. In addition, some TBD restricted access to raw data and images will be provided. This is in response to science team requests. The raw data will be copied to tape during initial processing and shipped to the recipients, who should be identified soon. Images will be available via ftp with password control.

  2. 2MAPPS Schedule Review -- The 2MAPPS development schedule was reviewed briefly. J. Fowler reported that the first meeting of each month will include a quick check of the schedule, with requests for notification by any team members who are aware of any milestones that may be in significant jeopardy. Satisfactory progress will be assumed toward any milestones not reported in jeopardy. Such reports need not be held until the first meeting of each month; it is simply at that time that a check will be made to remind the team members to examine their schedule status. The earliest possible warning for jeopardized milestones is desirable and may be submitted to J. Fowler or G. Kopan.

  3. Time-Variable PSF Status -- G. Kopan reported on the new program that computes a measure of the seeing for use by KAMPhot. This program is extracted from the code for the 2MAPPS STATS subsystem. After the meeting, G. Kopan issued an email description of this program, and an edited version of that is attached below (note that some S/N values have changed).

    B. Light reported that association of the new seeing parameter with corresponding KAMPhot PSFs will be extended to the new PSFs computed by J. Fowler by Moffat-broadening of the real 940601 PSF, and KAMPhot changes to use the seeing parameter to access the appropriate Moffat-broadened PSF have not yet been made but appear to be easy to implement.

  4. Alternate Flattening Software -- The flattening method used in the first run of POS will be the same as before. The same version of CFlat will be used. The variations of POS tested on karloff will include the new C#Flat program that will use a different flattening method. This involves an independent calibration of the responsivity, probably via analysis of twilight flats but still somewhat TBD. The responsivity and dark images will be used by C#Flat to estimate offset errors in each pixel (i.e., an additive error instead of the multiplicative error for which CFlat computes corrections). The new method will use an algorithm worked out a few weeks ago by G. Kopan and J. Fowler, with a modification suggested recently by R. Cutri.


The following is the attachment referred to in item #3, above:
=====================| Attachment Supplied by G. Kopan |=====================

    A measure of seeing has been incorporated into the prototype for the
2MAPPS STATS program (bumps). This measure (along with other diagnostic
measures) is currently being written into a single-line file with a name 
of the form:

         BfSSS.pfrac

         where:
                  B   - band (j,h, or k)
                  SSS - scan (i.e. 001)

    This file can be generated immediately after CFLAT is run, and copied
to the sSSS/sdata directory, to be used by KAMPHOT and other downstream
programs. The 'bumps' program runs in ~60 seconds on a typical scan on
karloff.

    The file contains the flattened frame file root (e.g., kf001), the
S/N threshold, and the 20, 40, 60, and 80 percentile values of the measure
(described below). These files have been generated for simulated data(*) for
a range of input 'seeing' values as summarized in the table below:

scan           >S/N     20%     40%     60%     80%      (seeing)
--------------------------------------------------------------------

kf001          20.00  0.5834  0.6480  0.6917  0.7358      (1.5")
kf001          20.00  0.5273  0.5871  0.6300  0.6698      (2.0")
kf001          20.00  0.5091  0.5682  0.6075  0.6471      (2.2")
kf001          20.00  0.4906  0.5457  0.5872  0.6243      (2.4")
kf001          20.00  0.4884  0.5396  0.5771  0.6140      (2.5")
kf001          20.00  0.4777  0.5327  0.5678  0.6057      (2.6")
kf001          20.00  0.4663  0.5165  0.5530  0.5898      (2.8")
kf001          20.00  0.4511  0.4981  0.5337  0.5704      (3.0")

    The 80 percentile parameter is the best indicator of the seeing; the
lower percentiles can be used as diagnostics, as they are more sensitive to
'false broadening' than the 80 percentile value.

    These parameters are estimators of the 'enpixelated energy' of a
point source centered on a pixel. The parameters are percentiles of a 
histogram of a parameter computed as follows:

       - the parameter is that fraction of the total energy (above the 
       median background) of a source which is contained in the peak
       pixel

       - only sources whose centroids are within 0.3 pixels of the center
       of the peak pixel and are over the S/N threshold are histogrammed

    This measure is a sensitive measure of the broadness of the optical
psf for those point sources centered on a pixel. The 80 percentile is 
used as a robust measure due to the following confounding factors:

       - Confused objects (confused point sources, saturated point sources
       and associated diffraction spikes, galaxies and other extended 
       structure) have lowered 'enpixelated energies' and pump up the
       lower part of the histogram

       - Lower S/N objects smear the histogram

    The current estimator is a prototype, and subject to errors of 2% or
more due to confusion (M92, 3rd mag star), which can translate into a 5-10%
error in 'seeing'. Galaxies are usually at a low enough density to have
little effect on the 80th percentile. Techniques to reject confused
measurements are under investigation and will be incorporated.

    This measure is relatively robust, has the advantage of simplicity
and can be evaluated early in the pipeline. Suggestions for improvements or 
better techniques are welcomed.


(*) - simulated data provided by Bob Light & Sherry Wheelock