Galaxy Zoo Starburst Talk

BPT diagrams from the SDSS spectroscopic pipeline

  • mlpeck by mlpeck

    I'm trying to forge ahead here with some substantive analysis even though the basic data tables aren't in final shape yet. I extracted emission line fluxes for both the quench and control samples estimated by the SDSS "idlspec2d" routine. Emission line fluxes are included in the FITS spectra files for DR8+, and flux errors are also tabulated. The SDSS pipeline also has a convention for indicating missing or unusable data, which seems to be a bit of a problem with the MPA-JHU pipeline. I'll say a little more about the comparison in another thread.

    For these BPT diagrams I followed standard practice of excluding flux estimates with estimated S/N < 3, and as I'll comment below this has a significant effect on the results. Here are the BPT diagrams, first for the quench sample and next for the controls. In these graphs the diagonal line in the upper right was proposed by Schawinski et al. (2007 - http://adsabs.harvard.edu/abs/2007MNRAS.382.1415S) to separate AGNs from LINERs.

    enter image description here

    enter image description here

    One thing you can probably tell visually is that there are many more AGNs and "transitional objects" in the quench sample. The percentage of star forming galaxies on the other hand is nearly the same (26.8% QS vs. 30.3% QC).

    What the graph can't show is that there were also more unclassifiable objects in the control sample - 54.9% vs. 42.6% in the quench sample. This actually isn't too surprising because the control sample has a substantial number of "red and dead" passively evolving systems.

    Sorry for the cut and paste job, but here is the detailed breakdown. In the output 0 is unclassified, 1 starforming, 2 "transitional", 3 AGN, and 4 LINERs.

     table(kclass.quench)/length(kclass.quench)
kclass.quench
         0          1          2          3          4 
0.42600000 0.26833333 0.21233333 0.06133333 0.03200000 
 table(kclass.control)/length(kclass.quench)
kclass.control
         0          1          2          3          4 
0.54900000 0.30333333 0.10966667 0.02266667 0.01033333

    Just for comparison I performed the same exercise with the MPA-JHU data, only without culling based on S/N since we still don't have error estimates. Here are the results:

    table(kclass.mpa)/length(kclass.mpa)
kclass.mpa
        0         1         2         3         4 
0.0670000 0.3916667 0.2300000 0.1183333 0.1930000

    Notice there's a huge shift out of the unclassifiable group and into both the starforming and non-starforming (AGN and LINER) classifications.

    Posted

  • mlpeck by mlpeck

    I'm still trying to forge ahead with some substantive analysis. Here are BPT diagrams for the quench and control samples with line fluxes from the MPA-JHU pipeline and non-detections by the usual S/N < 3 rule culled:

    enter image description here

    enter image description here

    Here is a breakdown of the classifications with the same interpretation as in the message above:

    table(kclass.quench)/length(kclass.quench)kclass.quench
        0         1         2         3         4 
0.2806667 0.3383333 0.1996667 0.0750000 0.1063333 
table(kclass.control)/length(kclass.control)kclass.control
         0          1          2          3          4 
0.49665328 0.34337349 0.10542169 0.02443106 0.03012048

    Notice that compared to the table above, which was based on SDSS pipeline fluxes, the quench sample has many fewer unclassified objects. I haven't checked for sure yet but this is most likely because the MPA pipeline has more firm H-beta detections. Most of those newly classified objects were placed in the LINER area of the BPT diagram, with some additional in the starforming region as well.

    There was much less change in the control sample, although the LINER region did get some additional objects (3% vs. 1% previously).

    Posted

  • JeanTate by JeanTate

    Reviving this old thread ...

    The heavily cited Brinchmann et al. (2004) has two classes of object where at least one of the four emission lines has a S/N < 3:

    Low S/N AGNs A minimum classification for AGN galaxies is
    that they have [N II]6584/Hα > 0.6 (and S/N> 3 in both lines)
    (e.g. Kauffmann et al. 2003c). It is therefore possible to classify
    these even if [O III]5007 and/or Hβ have too low S/N to be useful
    (cf. Figure 3). A similar approach is taken by Miller et al. (2003).
    In general, we will include these galaxies together with the AGN
    class.

    Low S/N SF After we have separated out the AGN, the composites
    and the low S/N AGNs, we have thrown out most galaxies with
    a possible AGN contribution to their spectra. The remaining galaxies
    with S/N> 2 in Hα are considered low S/N star formers.We can
    still estimate the SFR of these galaxies from their line strengths,
    even though we cannot use the full modelling apparatus described
    in the following section.

    Unclassifiable Those remaining galaxies that are impossible to
    classify using the BPT diagram. This class is mostly made up of
    galaxies with no or very weak emission lines.

    Making no cuts - i.e. looking at the full QS catalog - and using the emission line parameter values (and errors) given in that catalog, I get the following numbers of each kind of object, per their BPT diagram location:

    • AGN: 546
    • SF: 1046
    • Composite: 604
    • low S/N AGN: 460
    • low S/N SF: 80
    • Unclassifiable: 266

    The criteria I used are:

    • AGN: S/N > 3 for all lines, AND log(O[III]/Hβ) > (0.61/(log([NII]/Hα)-0.47)+1.19
    • SF: S/N > 3 for all lines, AND log(O[III]/Hβ) < (0.61/(log([NII]/Hα)-0.05)+1.3
    • Composite: S/N > 3 for all lines, AND NOT AGN, AND NOT SF

    Aside from the 'nuclear only spectrum' problem - which is explicitly discussed in Brinchmann et al. (2004), see below - the last three categories may include objects with 'bad spectra': if the wavelength region including one or more of the four emission lines is missing, then the classification may be erroneous. Several examples of objects with just such bad spectra have already been noted, in the Outliers - collect them here! thread (e.g. on page 6).

    Aperture effect, per Brinchmann et al. (2004):

    Note that this is merely a ‘’nuclear” classification — it does
    not make any statement about the properties of the parts of the
    galaxy outside the region sampled by the fibre. In particular, we
    expect the Unclassifiable category to include a substantial number
    of galaxies where we only sample the central bulge — for such
    galaxies there may be considerable amounts of star formation outside
    the fibre.

    Posted

  • mlpeck by mlpeck

    Cid Fernandes et al. (2010, 2011) proposed an alternative classification system based only on [N II] and Hα flux ratios and equivalent widths. Most but far from all quench sample objects that aren't classifiable in the classic BPT diagram fall into the "retired galaxy" bin in their scheme.

    I'd post more details but really, enough is enough.

    Posted