The Distribution of Quasars and Galaxies in Radio Color-Color and Morphology Diagrams

Abstract

We positionally match the 6 cm GB6, 20 cm FIRST and NVSS, and 92 cm WENSS radio catalogs and find 16,500 matches in ~3,000 deg2 of sky. Using this unified radio database, we construct radio "color-magnitude-morphology" diagrams and find that they display a clear structure, rather than a random scatter. We propose a simple, yet powerful, method for morphological classification of radio sources based on FIRST and NVSS measurements. For a subset of matched sources, we find optical identifications using the SDSS Data Release 1 catalogs, and separate them into quasars and galaxies. Compact radio sources with flat radio spectra are dominated by quasars, while compact sources with steep spectra, and resolved radio sources, contain substantial numbers of both quasars and galaxies.

Full text

Page 1

arXiv:astro-ph/0403314v1 12 Mar 2004
February 2, 20089:53WSPC/Trim Size: 9in x 6in for Proceedingsivezic
THE DISTRIBUTION OF QUASARS AND GALAXIES IN
RADIO COLOR-COLOR AND MORPHOLOGY DIAGRAMS
ˇZ. IVEZI´C, R.J. SIVERD, W. STEINHARDT, A.S. JAGODA, G.R. KNAPP,
R.H. LUPTON, D. SCHLEGEL, P.B. HALL, G.T. RICHARDS, J.E. GUNN,
M.A. STRAUSS, M. JURI´C, AND P. WIITA
Department of Astrophysical Sciences
Princeton University, Princeton, NJ 08544, USA
M. GA´CEˇSA AND V. SMOLˇCI´C
Department of Physics
University of Zagreb, Bijeniˇ cka Cesta 4, Zagreb, Croatia
We positionally match the 6 cm GB6, 20 cm FIRST and NVSS, and 92 cm WENSS
radio catalogs and find 16,500 matches in ∼3,000 deg2of sky. Using this unified
radio database, we construct radio “color-magnitude-morphology” diagrams and
find that they display a clear structure, rather than a random scatter. We propose
a simple, yet powerful, method for morphological classification of radio sources
based on FIRST and NVSS measurements. For a subset of matched sources, we
find optical identifications using the SDSS Data Release 1 catalogs, and separate
them into quasars and galaxies. Compact radio sources with flat radio spectra
are dominated by quasars, while compact sources with steep spectra, and resolved
radio sources, contain substantial numbers of both quasars and galaxies.
1.The Era of Modern Radio Surveys
Statistical studies of the radio emission from extragalactic sources are en-
tering a new era due to the availability of large sky area high-resolution
radio surveys that are sensitive to mJy levels (e.g. FIRST, Becker, White
& Helfand 1995; GB6, Gregory et al. 1996; WENSS, Rengelink et al. 1997;
NVSS, Condon et al. 1998). The catalogs based on these surveys con-
tain millions of sources, have high completeness and low contamination,
and are available in digital form. The wide wavelength region spanned by
these surveys, from 6 cm for GB6 to 92 cm for WENSS, and the detailed
morphological information at 20 cm provided by FIRST and NVSS, allow
significant quantitative and qualitative advances in studies of radio sources.
Here we present a preliminary analysis of sources detected by the GB6,
1

Page 2

February 2, 20089:53WSPC/Trim Size: 9in x 6in for Proceedings ivezic
2
Figure 1.
(left). This bimodality essentially reflects the separation between isolated core sources
(|∆t| < 0.2) and lobe and complex sources (|∆t| > 0.2). We confirmed this conclusion
by visually inspecting about 1000 2′×2′FIRST images (Ivezi´ c et al. 2004, in prep). The
majority of lobe and complex sources, and some core sources, are resolved by FIRST on
5 arcsec scale (Θ > 0.03, right panel).
The bimodal distribution of the ∆t = tFIRST−tNV SSmagnitude differences
NVSS, FIRST and WENSS radio surveys, and of the subset also detected
by the optical Sloan Digital Sky Survey (SDSS; York et al. 2000; Stoughton
et al. 2002, and references therein). Matching FIRST and NVSS allows a
robust estimate of the source morphology at 20 cm, and the addition of
GB6 and WENSS data allows the determination of the radio spectral slope
and curvature. SDSS identifications enable the separation of sources into
quasars and galaxies, and, for some objects, also provides the redshifts.
2. The Cross-identification of Radio Sources
The cross-identification of radio surveys at different wavelengths, and with
different resolutions, is not straightforward (e.g. Becker et al. 1995; Ivezi´ c
et al. 2002). However, the accurate positions provided by FIRST allow
simple positional matching with high completeness and a low random con-
tamination rate. Based on an analysis of positional differences, we adopted
maximum positional discrepancies between FIRST and the other three sur-
veys of 15′′, 30′′and 60′′, for NVSS, WENSS and GB6, respectively. This
choice yields completeness of over 80% and a false matching rate below 1%.
The current positional overlap of the GB6, FIRST, NVSS, and WENSS
catalogs includes 16,500 sources in about 3,000 deg2. The matched sample
is flux limited by GB6 for sources with spectral slope α < 0, and by WENSS
for sources with α > 0, where Fν∝ να. This is the largest database of the
radio spectral and morphological measurements assembled to date.
For convenience, we express all fluxes on the ABνmagnitude system of
Oke & Gunn (1983), where m = −2.5log(Fν/3631Jy). The survey limits

Page 3

February 2, 20089:53WSPC/Trim Size: 9in x 6in for Proceedings ivezic
3
Figure 2.
WENSS in the 6-20-92 cm radio spectral slope diagram is shown in the two left panels
for “core” (top) and “lobe” (bottom) sources, separated using the flux difference between
FIRST and NVSS measurements. The diagonal dashed lines show the y = x locus, with
the positions for α = 0 and α = −1 marked as “flat” and “steep”, respectively. The SED
difference between these two cases is illustrated in the insert in the bottom left panel.
The two right panels compare the distributions of all matched radio sources (contours),
to those for sources detected by the SDSS (triangles for galaxies and circles for quasars).
The distribution of radio sources detected by GB6, FIRST, NVSS, and
expressed in this system are s < 13.3, tFIRST < 16.4, tNV SS< 15.4, and
n < 13.3, for GB6, FIRST, NVSS, and WENSS, respectively (magnitudes
are named by the first letter of the corresponding wavelength in cm).
3.Radio Morphology from FIRST and NVSS Data
FIRST is a high resolution (5′′) interferometric survey and thus over-
resolves large sources relative to the lower resolution (45′′) NVSS survey,
leading to underestimated fluxes. Comparing the two surveys efficiently
separates point and extended sources. The ∆t = tFIRST− tNV SS magni-
tude differences have a bimodal distribution (Figure 1) which reflects the
separation between isolated core sources (|∆t| < 0.2) and lobe and com-

Page 4

February 2, 20089:53WSPC/Trim Size: 9in x 6in for Proceedings ivezic
4
plex sources (|∆t| > 0.2). In addition to ∆t, another useful morphological
parameter is the ratio of integrated and peak FIRST fluxes – sources with
θ = log(Fint/Fpeak) < 0.03 appear unresolved by FIRST (∆t and θ are not
equivalent because they measure size on different scales – 5′′and 45′′).
4.Radio Color-Color Diagrams and Optical Identifications
The left panels in Figure 2 show the distributions of core (top) and lobe
(bottom) sources (separated using tFIRST−tNV SS) in the 6-20-92 cm radio
spectral slope (i.e. “color”-“color”) diagram. Note that there are more flat-
spectrum sources in the top panel. According to size measurements from
the FIRST survey and Figure 1, the overwhelming majority of flat-spectrum
sources are also unresolved on 5′′scales (θ < 0.03) and on 45′′scales. The
right panels in Figure 2 compare the distributions of all matched radio
sources (contours) to those for sources optically identified by SDSS. The
compact flat-spectrum sources are dominated by quasars, while compact
sources with steep spectra, and resolved radio sources, include substantial
numbers of both quasars and galaxies. It is also noteworthy that the opti-
cally identified sources are not representative of the whole radio population.
5.Conclusions
The distribution of sources in radio “color-magnitude-morphology” space
displays clear structure which encodes detailed information about the as-
trophysical processes that are responsible for the observed radio emission,
and thus provides strong constraints for the models of these processes.
Acknowledgments
Funding for the creation & distribution of the SDSS Archive (http://www.sdss.org/) has
been provided by the Alfred P. Sloan Foundation, the Participating Institutions, the Na-
tional Aeronautics and Space Administration, the National Science Foundation, the U.S.
Department of Energy, the Japanese Monbukagakusho, and the Max Planck Society.
References
1. Becker, R.H., White, R.L., & Helfand, D.J. 1995, ApJ, 450, 559
2. Condon, J. J., et al. 1998, AJ, 115, 1693
3. Gregory, P.C., et al. 1996, ApJS, 103, 427
4. Ivezi´ c,ˇZ., et al. 2002, AJ, 124, 2364
5. Rengelink, R.B., et al. 1997, A&AS, 124, 259
6. Stoughton, C., et al. 2002, AJ, 123, 485
7. York, D.G., et al. 2000, AJ, 120, 1579