Galaxy Zoo Talk

Photo Z

  • LobeckDR by LobeckDR

    How close / what range do the PhotoZ values have to be to consider a galaxy or star is in close proximity to each other?

    Posted

  • Budgieye by Budgieye moderator

    PhotoZ is an guess at what the z value is based on the colours of the galaxy, when there is no spectral chart available. It can be a good guess, and it can be wrong.

    Based on the orange yellow colour, I would say without peeking, that the value of z for the central galaxy is about 0.01

    z is vaguely a measure of distance away from us, based on how the expansion of the universe is stretching the wavelengths of light. For stars in our galaxy z=0 The round white, red and some of the yellow objects are stars.

    Other galaxies are millions of light years away. So no star is ever close to a galaxy. They just overlap in the images. The exception is a supernova, and there are only a few in these images.

    enter image description here

    http://skyserver.sdss3.org/dr8/en/tools/explore/obj.asp?id=1237668757239366608 PhotoZ=0.26

    enter image description here

    http://skyserver.sdss3.org/dr8/en/tools/explore/obj.asp?id=1237668757239366606 no estimate of PhotoZ, probably it is a star in our galaxy

    enter image description here

    http://skyserver.sdss3.org/dr8/en/tools/explore/obj.asp?id=1237668757239366609 PhotoZ=0.11

    You can judge the distance of galaxies by their colours.
    Colours of Galaxies in SDSS : Redshift chart
    http://www.galaxyzooforum.org/index.php?topic=277142.0

    Zooite Guide to SDSS Spectra
    http://www.galaxyzooforum.org/index.php?topic=274815.0

    Posted

  • JeanTate by JeanTate

    To add to what Budgieye wrote, the DR8 photoz's have error bars:

    • (1st object, SDSS J202533.14-135219.8) 0.265117±0.022331
    • (3rd object, SDSS J202533.16-135224.5) 0.110225±0.037172

    And in DR10, there are two photoz's, determined using different methods (called the "KD-tree method", and the "RF method"). For these two objects, the two estimates are (KD-tree, then RF):

    • (1st) 0.265 ± 0.0223, 0.250 ± 0.0503
    • (3rd) 0.110 ± 0.0372, 0.144 ± 0.1007

    To Budgieye's point, while the DR8 photoz estimates do not overlap, the DR10 RF-method ones do. Photoz estimates for individual objects need to be treated very cautiously! 😃

    Posted