Äîêóìåíò âçÿò èç êýøà ïîèñêîâîé ìàøèíû. Àäðåñ îðèãèíàëüíîãî äîêóìåíòà : http://www.stsci.edu/ftp/meetings/galaxy-morphology/keel.ps Äàòà èçìåíåíèÿ: Wed Sep 7 18:19:51 1994 Äàòà èíäåêñèðîâàíèÿ: Sat Dec 22 14:18:03 2007 Êîäèðîâêà: Ïîèñêîâûå ñëîâà: âîçäóøíûå ìàññû

Structures and Photometric Redshifts of Faint Spiral Galaxies
William C. Keel, Jordan M. Burkey, and Wentao Wu
(University of Alabama)
We analyze WF/PC images of a sample of faint spiral galaxies with known spectroscopic
redshifts, deriving separate bulge and disk (V­I) colors by decomposing the images into
r 1=4 and exponential components. The color behavior of bulges with redshift in the range
z = 0:2 \Gamma 0:8 is very systematic, tracking models for elliptical­galaxy populations with star
formation confined to an initial burst. The relation is tight enough that even with the pre­
STS­61 PSF, redshifts can be constrained to \Sigma0:1 from the bulge color alone. Disk colors
show a wide range, scattering between the colors of bulges and the constant­SFR case.
Presented at Quantifying Galaxy Morphology at High Redshift, a workshop held at
the Space Telescope Science Institute, Baltimore MD, April 27--29 1994
1 Introduction
HST imaging makes possible detailed morphological studies for vast numbers of faint galaxies
at significant redshifts, far too many for detailed spectroscopic followup. Some investigations
on faint­galaxy properties from such data (i.e. the Burkey et al. 1994 study of galaxy pairing
at faint levels) require at least statistical redshift information for their interpretation. We
consider here whether the detailed structural information implicit in WF/PC images can
provide enough additional constraints to allow useful redshift estimates from colors of the
individual morphological components of such galaxies. Since ellipticals are well­behaved in
this regard and have only a single dominant component to deal with, we consider here the
case of spiral galaxies.
1

Keel, Burkey, & Wu Structures and Photometric Redshifts of Faint Spiral Galaxies
2 Decomposing Bulge and Disk Spectra
We separated disk and bulge properties for sample nearby galaxies, as a first step in decom­
posing distant systems from colors and morphological information. For galaxies with known
integrated spectra and photometric decompositions, we subtract a trial bulge spectrum (in
practice, for an intermediate­luminosity elliptical such as NGC 3379) normalized to match
the measured bulge­to­disk ratio at the appropriate wavelength. The residuals comprise the
estimated disk spectrum, generally close to that expected from a population with constant
star­formation rate.
We illustrate using NGC 3627 (fig. ??), with the integrated spectrum from the Kennicutt
(1992) atlas and photometry from Kent (1985). The strong wavelength dependence of disk­
to­bulge ratio is sobering for doing Hubble classification at constant observed wavelength.
The few galaxies for which data are readily available for such an analysis show an overall
good fit for the elliptical + constant­SFR model, though variations are seen at a given Hubble
type.
3 HST Data: Two­Dimensional Image Modelling
To measure disk and bulge colors from WF/PC images of galaxies at redshifts z = 0:2 \Gamma 0:6,
we fitted combinations of r 1=4 and exponential profiles, convolved with the local PSF, to
the original aberrated data. Deconvolved images were used to speed convergence of the fits
by using them to give initial values for the geometric parameters ­ orientation and axial
ratios ­ of the individual components. We took empirical PSFs from stars observed in the
same program where available, and TinyTIM PSF models otherwise. This procedure gives
estimates of galaxy structural parameters free of any biases introduced by deconvolution,
to the extent that these two components accurately describe a galaxy's morphology. The fits
are made in the two­dimensional image plane, using a ü 2 ­like criterion. It differs from an
exact ü 2 approach in normalization, since the amount of blank sky or nearly­blank sky on
the galaxy outskirts included in the fit is a matter of judgment and is often affected by
companion galaxies or other neighboring objects. For galaxies with well­determined local
PSF and symmetric structure, Keel & Windhorst (1993) show that the agreement among
structural parameters determined by completely disparate approaches is quite reasonable.
For this initial comparison, we use a set of galaxies in the 53W field (covered by the Leiden­
Berkeley Deep Survey, for example) in Hercules, which have deep WF/PC wide­field images
(Windhorst et al. 1994) and MMT redshifts. The WFC data are in F555W (``V'') and

Keel, Burkey, & Wu Structures and Photometric Redshifts of Faint Spiral Galaxies
4000 5000 6000 7000
0
0.5
1
1.5
2
Figure 1: Derived spectra of morphological components of NGC 3627. The bulge spectrum
is approximated by an elliptical galaxy scaled to match the photometric bulge:total intensity
ratio at the V band, and the disk spectrum is as the residual from this subtraction.

Keel, Burkey, & Wu Structures and Photometric Redshifts of Faint Spiral Galaxies
F785LP (''I'') filters; modelled quantities use the sensitivity curves for these same passbands
as well.
4 Results
Figure ?? shows the behavior of bulge and disk colors in the WF/PC V­I system with
redshift. Bulges are photometric well­behaved (particularly if errors due to the photometric
decomposition are taken into account), while disks exhibit a wide color range at a given
redshift. We superimpose models from Bruzual's GISSEL programs, for elliptical­like and
constant­SFR cases. Bulge colors from these data predict redshifts to \Deltaz = \Sigma0:1, while disk
colors are essentially constant with redshift (and widely scattered) over the relevant range.
The particulars of the galaxy models shown are as follows: assumed Hubble constant was 50
km s \Gamma1 Mpc \Gamma1 and q 0
= 1=2; the elliptical model had an initial burst 13.75 Gyr ago, and the
spiral disk model began star formation at a constant rate 10 Gyr ago. For z ! 0:8, the color
changes are dominated by passband effects (K­corrections) rather than genuine population
evolution.
5 Summary
We have compared colors of disks and bulges for intermediate­redshift galaxies (z = 0:2\Gamma0:8),
derived from WF/PC imagery, and redshifts. Bulge colors can be used as redshift indicators
(mostly because of strong K­corrections) when spectroscopic redshifts are unfeasible, to an
accuracy about 0.1 in z. Disk colors are very scattered, and indicate a wide range of star­
forming histories from initial bursts to constant­SFR (which forms the blue envelope of disk
color versus redshift).
We thank G. Bruzual for providing his GISSEL software, and the NSSDC for providing the
Kennicutt spectral atlas in digital form. This work was supported in part by STScI grant
GO­3545.02­91A.
Burkey, J.M., Keel, W.C., Windhorst, R.A. & Franklin, B. 1994, ApJL, 429, L13
Keel, W.C. & Windhorst, R.A. 1993, AJ, 106, 455
Kennicutt, R.C., Jr. 1992, ApJS, 79, 255
Kent, S.M. 1985, ApJS, 59, 115
Windhorst, R.A. et al. 1994, ApJ 435, in press

Keel, Burkey, & Wu Structures and Photometric Redshifts of Faint Spiral Galaxies
Redshift
0.2 0.4 0.6 0.8
2.5
2.0
1.5
1.0
(V-I)
(WF/PC
system)
Disks
Bulges
Figure 2: Colors of galaxy components in the F555W­F785LP (V­I) system derived from
bulge­disk decompositions versus spectroscopic redshift. The upper curve is the elliptical­
galaxy model, and the lower is a constant­SFR model. Filled circles are bulges and squares
indicate observations of disks.