UP - logo
E-viri
Recenzirano Odprti dostop
  • MegaMorph - multiwavelength...
    Vika, Marina; Bamford, Steven P; Häußler, Boris; Rojas, Alex L; Borch, Andrea; Nichol, Robert C

    Monthly notices of the Royal Astronomical Society, 10/2013, Letnik: 435, Številka: 1
    Journal Article

    We demonstrate a new multiwavelength technique for two-dimensional parametric modelling of galaxy surface-brightness profiles, which we have incorporated into the widely used software galfit. Our new method, named galfitm, extends galfit3's current single-band fitting process by simultaneously using multiple images of the same galaxy to constrain a wavelength-dependent model. Each standard profile parameter may vary as a function of wavelength, with a user-definable degree of smoothness, from constant to fully free. The performance of galfitm is evaluated by fitting elliptical Sérsic profiles to ugriz imaging data for 4026 galaxies, comprising the original Sloan Digital Sky Survey (SDSS) imaging for 163 low-redshift (v 7000 km s−1) galaxies and 3863 artificially redshifted (0.01 z 0.25) images of the same galaxies. Comparing results from single-band and multiband techniques, we show that galfitm significantly improves the extraction of information, particularly from bands with low signal-to-noise ratio (e.g. u and z SDSS bands) when combined with higher signal-to-noise images. We also study systematic trends in the recovered parameters, particularly Sérsic index, that appear when one performs measurements of the same galaxies at successively higher redshifts. We argue that it is vital that studies investigating the evolution of galaxy structure are careful to avoid or correct for these biases. The resulting multiband photometric structural parameters for our sample of 163 galaxies are provided. We demonstrate the importance of considering multiband measurements by showing that the Sérsic indices of spiral galaxies increase to redder wavelengths, as expected for composite bulge-disc systems. Finally, for the ellipticals in our sample, which should be well represented by single-Sérsic models, we compare our measured parameters to those from previous studies.