Date: Thu, 18 Jun 1998 12:12:23 -0700 (PDT)
To: 2mass
Subject: IPAC 2MASS WsubG Mtg Minutes
Cc: chas,, bgreen
Mime-Version: 1.0
Content-Transfer-Encoding: 7bit
Content-MD5: YGkWuizVAnz15cV3ckyQCA==

           IPAC 2MASS Working SubGroup (#154) Meeting Minutes

Attendees: S. Van Dyk, T. Evans, J. Fowler, L. Fullmer, D. Kirkpatrick, 
           G. Kopan, H. McCallon, B. Wheaton, S. Wheelock, T. Jarrett,
           R. Beck


1.) 971116n and updated 2MAPPS
2.) 980319s as a southern RTB night
3.) K Magnitude Anomaly
4.) 2MASS and last week's PR


1.) 971116n and updated 2MAPPS

    T. Jarrett suggested this splinter group meeting, although R. Cutri
was not available to attend, in order to get some sense from other team
members of how the newly-processed RTB night 971116n with the updated 
version of 2MAPPS was looking.  Only members of the whole group who had
a pressing interest in the proceedings attended.

    H. McCallon reported that 1) PosFrm had a few problems with that 
night, in that some ACT stars were not picked up in the processing.
The problem appears to stem from a change in the way the U-scan to
N-scan cross-scan scale factor is computed as of a recent PosFrm delivery
(2.0_980512).  Investigation showed that one of several points at which 
the change needed to be made was missed.  For the night of 971116n,
at least, the net effect on the reconstruction accuracy was small.
He is pursuing the problem and hopes to have a PosFrm redelivery ready
by tomorrow. 2) All four additional scans which dropped back to a USNOA
reconstruction due to the newly imposed more conservative requirements 
for an ACT reconstruction got worse.  Howard is considering restoring the 
more relaxed ACT requirements and counting on Quality to pick up any
resulting problems.

    T. Evans inquired of W. Wheaton about the clipping of bright star 
magnitudes; in particular, the effect of Beta Peg on scan 27.  The revised
PROPHOT gave a J magnitude of -15, rather than 99.999; this error tumbled 
through to MAPCOR.  Beta Peg is truly producing a bunch of Scheat.  This 
version of PROPHOT is therefore not clipping properly for J-band and is 
wreaking havoc downstream.

    W. Wheaton needs to check the PROPHOT version number that is in the
pipeline.  PROPHOT has had the same version number for months, although 
the actual code has been continually revised.  J. Fowler indicated that 
new code should have a new version number and date, to distinguish it 
from previous versions of code.

    W. Wheaton discussed the PROPHOT bug that is causing the program to
fail for crowded regions.  The magnitudes are being returned as 99.999
for some stars in cluster environments, etc.  Currently, the key routine
(SOLVE) in PROPHOT obtains a preliminary flux, background, and variance 
estimate by simple aperture photometry (not using APPHOT) of each star
and runs profile fit, using flux, background, and variance estimate from
the previous step.  If these quantities are not available, the profile 
fit fails.  The routine then runs APPHOT on each candidate, for 12 
apertures, writes the results, both for profile fit and APPHOT, to a .ps 
file.  The code fails in the first step in crowded regions, because the 
outer sky aperture is masked out around stars out to 6 arcsec.

    The proposed fix would be to use APPHOT only on those failures from 
the first step, obtaining alternative estimates of flux, background, and
variance.  Then, the following steps in the current code would hopefully
complete normally.  Currently, if the first step fails, there is no 
profile fit information in the output.  Then, only APPHOT information is
available, if the third step succeeds.  This can happen on 50% or maybe 
even more of the stars in a crowded environment, e.g., near the core of 
a globular cluster.  With the fix, assuming APPHOT works better than the
aperture photometry in the current first step for crowded fields, profile
fit photometry for all or most of those stars would be recovered.  Thus,
the fix would give us profile fit photometry for all or most stars in 
crowded fields, if it works.   The question remains as to how sensitive 
the output of the profile fit step is to the initial estimates of flux, 
background, and variance.

    The problem affects only one one-thousandth of the whole survey
area.  The significance here is that, since APPHOT is probably slower 
than the current first aperture photometry step, the proposed fix may 
incur a speed penalty.  Because of the small fraction of stars affected 
overall, W. Wheaton hopes that this will be a tolerable cost.  However,
very dense regions may run significantly slower than before.

    The schedule for the release of 2MASS 2.1 was reiterated.  Initial
delivery is scheduled for June 30.  Testing will occur in July, with the
final version delivery by July 31.

2.) 980319s as a southern RTB night

    T. Jarrett highly recommends this night as a RTB night for the south.
The variety of fields can provide excellent tests for the pipeline.  The
suggestions were to run this night through with new versions of POSMAN,
GALWORKS, and (at J. Fowler's suggestion) MPCAT.

3.) K Magnitude Anomaly

    S. Wheelock again reported the anomaly involving K photometry, in
which, only in a few cases, the K magnitude on one of the six 
overlapping scans was about two magnitudes fainter than on the other 
five scans.  A suggestion was made to see what aperture magnitudes, 
rather than PROPHOT magnitudes, might be in these cases.  A suggestion
that dead pixels might be causing the problem.  It is still a mystery.

4.) 2MASS and last week's PR

    S. Van Dyk reported on the excellent public relations that 2MASS
experienced the previous week, both within the astronomical community,
from the 2MASS exhibit, 2MASS-related posters, and the excellent progress
reported given by R. Cutri at the AAS meeting in San Diego, and also to 
the lay community, in the form of the press coverage on D. Kirkpatrick 
et al.'s paper on the L-type dwarfs and the Astronomy Picture of the Day
(the beautiful Orion mosaic by G. Kopan) posted on the APOD website.