Subject: Re: Summer plans for visit to Garching
Date: Tue, 02 Feb 1999 09:54:50 -0500
From: Peter Stockman <stockman@stsci.edu>
To: rfosbury@eso.org

Bob

Seeing the solar eclipse from Munich sounds cool.

In regards to the integral field versus MOS
trades, I've just revisited the issue of sky coverage -- apparently
prompted by your study!. By looking at a spreadsheet that I made
from the HDF statistics (in the I band), I get about the same answer
as you do in that ~10% of the sky is covered at 28mag/arcsecsec squared
(Iab I think). Ferguson gets about 10% using sextractor on the H band image
and thinks that much of that is due to the NICMOS PSF (e.g. 2.4 m diameter)
My spreadsheet model suggests that much/most of that coverage is due to
~100 relatively bright galaxies with sizes ranging from 1-5 arcsec radius
at that isophote. As a result, I'm not so sure that it logically calls for
a integral field spectrograph since most of this coverage may be
handled from the VLT (and other facilities). My next step -- still needing
some planning work -- is to look at the following.

1/ Assume a pixel scale, 4-6 pixels per resolution element.
2/ Assume a noise characteristic
3/ For various spectral resolutions, calculate the visibility and
useful aperture size for various galaxies using HDF number counts and a max
integration time of ~ 10^5 s and 10:1 signal to noise.

This is a multi-dimensional study but I think the results will be something
like:
o There will be an optimum pixel scale to observe the faintest galaxies. This
would set the scale of both the MOS apertures and slicer widths. How does
the European team come down on resolving the sources? Two slicer widths per
object or one? Since they land randomly on the sky, 2 sounds more realistic for
both the MOS and the IFS. (Remember, the WFC scale basically sucks).
o As one increases the spectral resolution, fainter objects will drop out and
the optimum pixel sizes will INCREASE (maybe just slightly).
o There may not be many faint galaxies with surface brightness features
bright enough for R ~ 3000 and resolutions ~ 40 mas.
o The overlap with the ground will be hard to estimate since there is not
a strong redshift vs brightness relationship in the number counts. But th
ground MAY be very competitive until Halpha moves beyond the K band (say
z < 3 ! ) for high resolution spectroscopy and excellent, natural seeing.
If the seeing is "natural", the ground can have a larger field of view than
NGST. Thus
one is tempted to think that at "high" resolution, NGST should move to
high angular resolution, integral field science. My concern is whether there
is enough science there with the detectors that we are likely to have. (On
the other hand, the high res, MOS spectrograph in the NIR will be quite a
challenge for the ground.)

Sooo. I don't know what the answer is!
Pete

H.S. (Peter) Stockman                   NGST Study Scientist
stockman@stsci.edu                      Space Telescope Science Institute
410-338-5007                            3700 San Martin Drive
410-338-1592(fax)                       Baltimore, MD 21218