From jwf@ipac.caltech.eduWed Sep 10 14:43:12 1997
Date: Wed, 25 Jun 1997 16:25:58 -0700 (PDT)
Subject: WG Mtg #126 Minutes

           IPAC 2MASS Working Group Meeting #126 Minutes

Attendees: T. Evans, J. Fowler, H. McCallon, B. Wheaton, S. Wheelock, J. White


1.) Saturation Analysis
2.) Minor Planet Catalog
3.) Vacation and Observing Schedules


1.) Saturation Analysis

    B. Wheaton displayed some images pertaining to the saturation analysis he
is now beginning. An IDL plot of a dark frame was remarkable in the detail it
showed; individual pixel values were actually discernible. He also had two
images obtained from UMASS via the web that showed pixel output versus
integration time. These showed that some deviation from linear response does
indeed occur over the entire range, and the sudden onset of saturation was
also visible in each band, along with the nonmonotonic response afterwards.
    The relationship of saturation tagging to photometric requirements was
discussed and found not to be straightforward (here "saturation" is used in
the "soft saturation" sense, i.e., significant deviation from linear response
at the high end). Some time back, an error tree was developed by J. Fowler,
but the filling out of this tree with a quantitative error propagation model
was assigned as a science team activity which has apparently been awaiting the
completion of higher-priority tasks.
    Two uses for the improved saturation tagging were identified: (a.) pixels
tagged as saturated are counted and bookkept in the point-source extraction
processing, so corresponding photometry can be interpreted in the light of
this information; (b.) Read1 saturation is used to trigger first-persistence-
object photometric estimation. Neither of these seem to relate directly to the
primary photometric requirements.
    Some discussion of how to proceed in the saturation analysis took place.
The goal is to replace the single number currently used as a threshold for
tagging a pixel value as saturated with an array of individual pixel values.
It was suggested to begin the exploratory studies by computing recursive
linear fits to each Read2 pixel vs. the dark-subtracted Read1 median for the
corresponding frame in a morning twilight flat sequence (morning, so that the
pixels will be going into saturation instead of coming out of it).
    Significant deviation from linearity can occur at the low end as well as
the high end, because of noise dominating the low signal. For a given pixel,
each linear fit should reject data points beyond N sigma from the previous
iteration's fit, using the previous iteration's sigma (N = 2, TBD). Such an
algorithm is used in DARKS to compute responsivity (except that the ordinate
is a dark-subtracted Read2-Read1 pixel, and the abscissa is the corresponding
frame median). Iteration ceases when (a.) the same data points are used in two
successive iterations, or (b.) too many points have been rejected, or (c.) the
maximum number of iterations is reached (TBD 100).
    Bill will pursue these and other ideas in the next few weeks.

2.) Minor Planet Catalog

    J. Fowler reported that work was fairly advanced on the Minor Planet
Catalog (MPCAT) subsystem. At the science team meeting the first week of June,
the desire was expressed for this to include periodic comets and any planets
that might appear with high S/N in the 2MASS data. It is not clear whether
this is feasible, but some concern was expressed that planet apparitions might
get into the data base without being identified as such.

3.) Vacation and Observing Schedules

    J. Fowler reported that he will be on vacation from next Monday (June 30)
through July 15. H. McCallon reported that he will be observing at Mt. Hopkins
all of next week, probably the week after that, and possibly some of the
subsequent week as well.