Deep KS Imaging of the COSMOS Field
Science:
COSMOS is a multiwavelength, deep
extragalactic survey project. Its primary goal is to improve understanding of
the relationship between large scale structure, galaxy formation, and dark
matter. The angular extent of the COSMOS field samples a large linear scale of
the distant Universe for determining structure, while its depth and
multifrequency nature allow study of galaxy evolution as a function of
environment to large lookback times.
Components of COSMOS include HST
Advanced Camera observations of the defining two square degree field
(completed), a complementary deep mid-infrared survey with Spitzer, X-ray
observations with XMM, and groundbased survey and followup work with the VLA,
VLT, Subaru, and NOAO 4-m telescopes among others. IR cameras on the NOAO 4-m
telescopes are being used to survey the field in J and K, to a limiting KS
~ 21. I propose to use NEWFIRM to survey the COSMOS field to a significantly
fainter limit, KS »
22.5 and J » 24.3. These much
deeper near-IR data will provide improved photometric redshifts for less
luminous galaxies in the range 1 < z < 2. This
aids study of the formation and evolution of structure out to the epoch of peak
galaxy formation. Together with ACS morphological information, the IR fluxes
will be used to understand the properties and evolution of individual galaxies,
constrain galaxy formation scenarios, and select rare objects such as extremely
red galaxies for followup study with 8-10 m telescopes. The PI’s of the HST,
Spitzer, and NOAO observing efforts have been consulted and are very interested
in this extension of the COSMOS survey.
System performance issues:
The basic approach is to integrate
for several hours in each filter. Individual integrations will be background
limited, and the data rate will be moderate (~1 frame/minute). The number of
distinct NEWFIRM pointings is limited, so this is not challenging for
assembling a mosaic. The emphasis for system characterization and optimization
is on systematics, i.e.
non-Poissonian effects that may compromise sensitivity in very deep
integrations. These may originate in the sky (spatially or temporally variable
airglow; structure in the thermal background), in the telescope (flexure
producing very low level flatfield variations), or in the instrument (variable
electronic offset or noise level, pickup, thermal drifts of array bias level).
While the desired limiting sensitivity doesn’t break any records, it is enough
to test for very low level, perhaps intermittent effects. The data set will
enable experimentation with pipeline reduction protocols to determine optimal
ways to remove or correct these components of instrument signature in a fully
automated implementation.
Observing strategy:
The COSMOS field
is centered on RA=10 hr 00 min, DEC +2 deg 12 min.
Using SQIID
performance numbers as an estimator for NEWFIRM, the desired limiting
magnitudes can be reached (3s) with 6
hours of on-source integration per filter. Since telescope motions will be
small, overheads are dominated by readout time; this will be reduced by
coaddition of short integrations on board the array controller before readout
to the data system. I estimate 7 hours of clock time per filter, per pointing
to reach KS »
22.5 and J » 24.3.
Eight NEWFIRM pointings would almost cover the COSMOS field.
However the field position is not optimal for observing in May. It can be
observed for 1.5-2 hours
in the evening before the rate of change of airmass (at X > 2)
becomes troublesome for sky subtraction (and the seeing begins to degrade).
Observing episodically in this fashion requires ~4 nights per filter, per
pointing. Hence coverage will be limited by the number of nights available.
With a 14 night Science Verification run (presumably including a variety of
other programs) one could cover half the COSMOS field in KS or 1/4
of it in J and KS. I think the first alternative is somewhat
stronger scientifically.
This note has
been added by R. Probst subsequent to receipt of the original SV project idea.
The GOODS Survey
is similar to COSMOS in its science goals and multiwavelength nature. It goes
deeper in limiting magnitude in two significantly smaller fields, 10 x 16
arcmin centered on HDF-N and Chandra Deep Field-S. System performance issues
would be the same for deep broadband observations in either GOODS or COSMOS
fields.
The HDF-N GOODS
field is well place for SV observations in May-June, at 12:36 +62:13. It is
observable for 4-6 hours/night in May and 4 hours/night in June. A single
NEWFIRM pointing covers the field. The desired 5s limiting sensitivities (J, H, KS = 24.3, 23.4,
22.5) can be reached in 17 hours of integration per filter. Thus four scheduled
nights per filter, or 12 nights in all, would suffice to complete the project.
Part of the
motivation for selecting GOODS as a deep extragalactic field is that the IR
survey team has had very bad luck with weather and equipment at KPNO for
several years. A successful run with NEWFIRM would finish the NIR observations
for this very highly rated SIRTF Legacy Program.