2006 Articles

Panchromatic properties of 99 000 galaxies detected by SDSS, and (some by) ROSAT, GALEX, 2MASS, IRAS, GB6, FIRST, NVSS and WENSS surveys

Obric, M.; Ivezic, Z.; Best, P. N.; Lupton, R. H.; Tremonti, C.; Brinchmann, J.; Agüeros, Marcel Andre; Knapp, G. R.; Gunn, J. E.; Rockosi, C. M.; Schlegel, D.; Finkbeiner, D.; Gacesa, M.; Smolcic, V.; Anderson, S. F.; Voges, W.; Juric, M.; Siverd, R. J.; Steinhardt, W.; Jagoda, A. S.; Blanton, M. R.; Schneider, D. P.

We discuss the panchromatic properties of 99 088 galaxies selected from the Sloan Digital Sky Survey (SDSS) Data Release 1 ‘main’ spectroscopic sample (a flux-limited sample for 1360 deg^2). These galaxies are positionally matched to sources detected by ROSAT, Galaxy Evolution Explorer (GALEX), two-Micron All-Sky Survey (2MASS), Infrared Astronomical Satellite (IRAS), Green Bank GB6 survey (GB6), Faint Images of the Radio Sky at Twenty-centimetres (FIRST), NRAO VLA Sky Survey (NVSS) and Westerbork Northern Sky Survey (WENSS). The matching fraction varies from less than 1 per cent for ROSAT and GB6 to ∼40 per cent for GALEX and 2MASS. In addition to its size, the advantages of this sample are well-controlled selection effects, faint flux limits and the wealth of measured parameters, including accurate X-ray to radio photometry, angular sizes and optical spectra. We find strong correlations between the detection fraction at other wavelengths and optical properties such as flux, colours and emission-line strengths. For example, ∼2/3 of SDSS ‘main’ galaxies classified as active galactic nucleus (AGN) using emission-line strengths are detected by 2MASS, while the corresponding fraction for star-forming galaxies (SFs) is only ∼1/10. Similarly, over 90 per cent of galaxies detected by IRAS display strong emission lines in their optical spectra, compared to ∼50 per cent for the whole SDSS sample. Using GALEX, SDSS and 2MASS data, we construct the ultraviolet–infrared (UV–IR) broad-band spectral energy distributions for various types of galaxies, and find that they form a nearly one-parameter family. For example, the SDSS u- and r-band data, supplemented with redshift, can be used to ‘predict’ K-band magnitudes measured by 2MASS with an rms scatter of only 0.2 mag. When a dust content estimate determined from SDSS spectra with the aid of models is also utilized, this scatter decreases to 0.1 mag and can be fully accounted for by measurement uncertainties. We demonstrate that this interstellar dust content, inferred from optical SDSS spectra by Kauffmann et al., is indeed higher for galaxies detected by IRAS and that it can be used to ‘predict’ measured IRAS 60 μm flux density within a factor of 2 using only SDSS data. We also show that the position of a galaxy in the emission-line-based Baldwin–Phillips–Terlevich diagram is correlated with the optical light concentration index and u−r colour determined from the SDSS broad-band imaging data, and discuss changes in the morphology of this diagram induced by requiring detections at other wavelengths. Notably, we find that SDSS ‘main’ galaxies detected by GALEX include a non-negligible fraction (10–30 per cent) of AGNs, and hence do not represent a clean sample of starburst galaxies. We study the IR–radio correlation and find evidence that its slope may be different for AGN and SFs and related to the Hα/Hβ line-strength ratio.