[News] USM Plans/AIs for the coming Q3 and Q4

[Dr. Mark Neeser]

Hi astro-wisers,

here is a summary of current USM plans and AI's for
the coming two quarters:

Cheers,

Mark





Ongoing/New Astro-wise Activities at USM for Q3 and Q4 2004:
------------------------------------------------------------
------------------------------------------------------------



Continuing WP2 activities:
--------------------------

For the spectral energy distribution and photometric redshift fitting tool
the parameter definitions, the required database inputs, and the python
interface necessary to query the SED fitting routines have been completed
(Jan Snigula).

Three main problems remain to be solved:

1/  The convolution between the spectral energy distributions used in
    our photometric redshift routines and the filter curves of OmegaCam
    are still missing.  For demonstration purposes we will integrate the
    SED's with WFI filter curves.

2/  The crucial issue of PSF homogenization is still unresolved.  Since
    SED matching and photometric redshifts (among many other scientific
    analysis routines!) requires that we compare data from many nights,
    under varied seeing conditions, and in a number of filters, some
    form of PSF correction is needed in order to obtain meaningful
    results.  See our proposal below for a possible solution.

3/  In the SAssociate code, a function to match a master catalogue to a list of
    multiple catalogues, in a single data base process, is missing.  Currently,
    we are doing this in python, but this is unacceptably slow (several minutes
    for a few 1000 objects).  (AI: Kor).  Additionally we have a problem
    with the precise definition of a search area.



Additional Projects in Quality Control
--------------------------------------

As was outlined by Ronald, astro-wise foresees three quality control issues
for its photometry:  Catalogue creation, atmospheric extinction, and zeropoints.
We propose creating a fourth quality control parameter for photometry regarding
the apertures.  In lieu of homogenizing the PSF's for all images in the astro-wise
data base (not really feasible since it would require that we either a priori
define a "seeing" standard and convolve all frames to this level (degrade in the
case of better seeing), or selectively choose only frames near to one another in
seeing for our scientific analysis), we propose creating a scaling algorithm, at
the source catalogue level, that accounts for the differences in seeing.

USM proposes to attempt to create an algorithm that uses the global
psf of a given frame (say, using PSFeX, allstar, or daophot), to determine
the most appropriate aperture size to minimize the intrinsic photometric error.
This correction would be determined on a source by source basis and would be
computed from the source lists; a distinct advantage over having to modify
the image frames (as required by psf homogenization).  Such a magnitude
"correction" (re-scaling) will be akin to a "curve of growth" analysis, but
will have to be expanded to also include sources that are resolved (ie. non-stellar).
A possible solution would be to run SExtractor photometry with a set of at least
10 apertures of incrementally increasing size.  Based on the sources' morphology
(including information on axis ratio, colour, etc.?) the optimum aperture is computed
and a magnitude is determined based on an interpolation of the available apertures.
We will extensively test this by attempting to recover the known photometry of the
FORS Deep Field from various psf degradations, and by using a number of simulated
data fields.

Such a tool is required since some level of homogeneity of integrated fluxes across
images of varying seeing and different filters is a necessity for any SED fitting
routines, photometric redshifts, and colour-colour plots, to have scientific
validity.  It is a requisite for quality control tools, such as colour-magnitude plots,
and photometric colour corrections.  This is particularly relevant for the continuously
observed polar photometric fields.


Timescales:
-----------

With the resolution of a SAssociate match between a master source list and multiple
lists (point 3/ above), a working photometric redshift routine could be delivered
by the beginning of Q4.  For the complete testing of the feasibility of an algorithmic
correction to aperture disparities, we will need until the end of Q4.