Radio Galaxy Zoo Talk

J024936.66+055656.8 QSO?

  • Ptd by Ptd

    Is there any chance that J024936.66+055656.8 is a QSO rather than a star? Do the catalogs ever get such things wrong?

    Posted

  • ivywong by ivywong scientist, admin

    Unfortunately there isn't a spectrum available but yes, often the automatic classifier for SDSS mistakes a QSO for a star...

    Posted

  • JeanTate by JeanTate in response to Ptd's comment.

    Do the catalogs ever get such things wrong?

    It's not so much that the catalog - SDSS in this case - is "wrong" as that it's important to understand how catalogs reach conclusions (if we may call them that).

    In the case of SDSS:

    • there are two automatic pipelines, both of which produce automated classifications
    • the spectroscopic one tries to fit a single kind of source to the spectrum, and determine a redshift
    • the photometric pipeline can generate only two classifications, STAR or GALAXY (it's a bit more complicated, but that's good enough)

    The spectroscopic pipeline has a library of templates (~30?), with several types of star, and GALAXY, QSO, AGN, and a few others. For stars, the fitted redshift is very small, but I don't know if there are range limits on the redshifts of galaxies etc. Obviously, this pipeline can make 'mistakes', especially for featureless and/or noisy spectra (a type of QSO called a BL Lac, or blazar, is particularly difficult to classify using SDSS spectra; they are pretty rare, but very interesting).

    Historically, QSOs (quasi-stellar objects) were so called because on the best images of the time - photographic plates - they seemed no different from stars. In SDSS many QSOs can be resolved, i.e. you can see that they are bright, star-like objects surrounded by some fuzz. Hubble images of QSOs nearly always show at least some hint of the galaxy the QSO is in. A story - not sure how true it is - which illustrates the ingenuity of astronomers: not long after QSOs were discovered, one enterprising astronomer put a tiny 'occulting spot' on some part of a telescope, then aligned it so that the QSO being observed fell exactly on this spot. The resulting image - photographic plate? - showed some faint 'fuzz', the galaxy! Being a good scientist, I'm sure the astronomer also used this set-up to image several stars ... in images they would have had no such 'fuzz'.

    It's not hard to see where each pipeline can "go wrong":

    • classifications of faint or noisy objects will always need to be taken with a grain of salt
    • chance alignments: these certainly happen, but both pipelines try to fit just a single type of object
    • moving objects: the photometric pipeline does try to identify moving objects (especially asteroids), but it's not always successful
    • transient objects: fortunately, few astronomical objects vary dramatically in intensity over the ~few minutes it takes for SDSS to record; however some of the most interesting do (e.g. GRBs)

    Perhaps the most dramatic of chance alignments are symbiotic stars: these are binaries, with one an M-dwarf and the other a white dwarf. Photometrically they are - correctly - identified as single stars (the components are separated by a distance, on the sky, far less than what SDSS can resolve). Spectroscopically? All sorts of mistakes! A white dwarf spectrum is blue, and pretty featureless; an M-dwarf's is full of features and is red; no wonder the spectroscopic pipeline gets a headache trying to fit a single template!

    Hope this helps, and happy hunting! 😃

    Posted