Date: Thu, 10 Sep 1998 09:34:53 -0700 (PDT)
To: 2mass
Subject: IPAC 2MASS WG Meeting #164 Minutes
Cc: chas, stiening, bgreen

           IPAC 2MASS Working Group Meeting #164 Minutes

Attendees: R. Cutri, S. Van Dyk, R. Beck, T. Evans, R. Hurt, J. White,
           J. Fowler, H. McCallon, S. Wheelock, R. Tam, D. Kirkpatrick,
           T. Jarrett, W. Wheaton


1.) Project Update 
2.) Biases and Corrections
3.) A Long, but Relatively Fruitless Discussion
4.) New Action Items   <-- A New Feature in WG Minutes


1.) Project Update 

     Northern OPS at Mt Hopkins should resume this week.  Nothing new from the
southern facility, except the weather has not been great.

     Southern processing has begun, but not without a small hiccup.  980319s
was run with the old PSFs, 980321s was run with the new.  980714s, a long night
with particularly bad seeing, is being run tonight with the new production
PSFs, and will allow derivation of the MAPCOR normalizations with the new PSFs. When the MAPCOR namelist values are derived, the processing will resume.

2.) Biases and Corrections

     S. Wheelock discussed the analysis of the cross-scan photometric bias/
calibration correction.  The algorithm appears to work well for most scans; all
of these seem to have the bias removed.  Other scans have the bias 
"straightened out," but not quite right.  A small bias offset still exists for
these scans, which typically are high-stellar-density scans.  A physical reason
for the bias may be close to being understood: it appears to be something in the
telescope optics.  M. Skrutskie analyzed a focus test and found that the pupil
image is slightly off-center.  It may be possible to nudge the focus in the
south some to correct for the photometric bias at the telescope.  In the north
the astigmatism is worse, so it may not be possible to do a telescope 
correction, but in the north, the photometric bias is a much smaller effect 
that probably does not require correction.

     The correction algorithms continue to be tested and will need more tuning
for the southern processing.  
     H. McCallon presented data showing significant improvements for the 971116n
position reconstruction obtained via a "Martinizing" of contiguous survey scans.
Each scan was broken down into 12 segments and average in-scan/x-scan position 
differences with respect to adjacent scans were computed.  Through an iterative
scheme all scan segments were adjusted to improve the overlap differences, and a
table of in-scan/x-scan position corrections was built. The resulting 
corrections were applied to all survey extractions using linear interpolation 
between segment centers.  The overlap position differences were regenerated, as 
well as the ACT residuals.  Sigmas using all survey overlap data dropped from 
0".186 and 0".159, to 0".133 and 0".128 for in-scan and x-scan, respectively.  
The resulting increase in  ACT residuals was quite small, going from 0".115 and
0".094 to 0".122 and 0".106 for in-scan and x-scan, respectively.  Improvements 
for problem scans were even more dramatic than indicated by the overall stats, 
especially for scan 113, which has no ACT's.

It was decided not to include the position corrections in the Sampler release
for several reasons.  Consistency would require changing the extended point
source positions as well, along with coadd headers, galaxy stamp headers,
USNOA association distances, and asteroid association distances and match
confidences. H. McCallon was tasked to scope out a plan for possible inclusion
in the March release of an expanded approach using all available overlaps
from the database.  "The minimum requirement for the Spring 1999 data release is
that positional uncertainties reflect reality," said R. Cutri.

3.) A Long, but Relatively Fruitless Discussion

     The positional uncertainties and positional reconstructions performed
during the inevitable reprocessing of data in the future will subsequently have
its effect on object names, which are derived from their (accurate) positions.
The proposed 2MASS naming convention is accurate to one-hundredth of a time second
and one-tenth of an arcsecond.  Positions for sources, particularly point 
sources, may change by some fraction of 1", and so the adopted name must 
undoubtedly change.  There will be two catalogs, due to reprocessing, the
initial one and the final one, plus a cross-reference between the two.  An
overly long discussion ensued, with copious suggestions from J. Fowler, on how
to avert this inevitable problem.  Nothing was resolved regarding this important
issue, for now.  But, then, a solution will not be necessary until catalogs are
re-released, although it would be nice to have a working plan in place to deal
with this.

     Finally, T. Evans has sent around a URL containing information about her
analysis with R. Tam on the MAPCOR error and its effects.  R. Cutri encouraged
all members of the 2MASS team to please look at the webpage.

4.) New Action Items

%                                                      %
&                 ACTION ITEMS                         %
%                                                      %

     This will be a new feature of the WG minutes, and its format will likely
evolve.  Here, the new action items for the ensuing week will be listed, often
directed at particular personnel.  It is meant as a reminder, to be referenced
as such.

     1. For this next week, R. Cutri directed the few who had not yet tested the
CatScan and Survey Visualizer web interface to do so.  It is imperative that all
2MASS team members test the site and provide their input before the Sampler

     S. Van Dyk directed some general comments regarding the CatScan interface.
He will direct pointed comments to J. Good and J. Mazzarella soon.  Comments 
and suggestions from other 2MASS team members are still welcome and should
be directed to Good and Mazz, being sure to cc R. Cutri in all communications.