Author Topic: Paper with overlapping-galaxy catalog submitted!  (Read 11164 times)

NGC3314

  • Global Moderator
  • Hero Member
  • ****
  • Posts: 1590
    • View Profile
Re: Paper with overlapping-galaxy catalog submitted!
« Reply #30 on: July 31, 2012, 02:21:45 pm »
I wonder, what is given in the catalog as the smallest redshift difference between 2 overlapping galaxies?

The current catalog version includes 494 pairs for which both redshifts are known. Of these, 19 (3.8%) have redshift difference <0.0001 (the precision to which the SDSS redshifts are listed on the spectrum plots, and generally the error is a bit larger than this). This corresponds to a Doppler shift difference from line-of-sight relative motion less than 30 km/s. Of course, these are not our favorite overlapping pairs, because such a small redshift difference makes it very probable that the galaxies are physically associated, and possibly interacting and therefore asymmetric . A long-term problem in catalogs of galaxy pairs has been the fact that the internal velocities within small groups of galaxies are the same as the of interacting pairs, so that from some directions we would see galaxies in a group appearing very close together and potentially interacting (Andromeda and the Milky Way would be an example, since they are not yet distorted by interaction with each other). Seeing structural signs of interaction (tails, bridges, asymmetric spiral structure) is much safer, but could miss galaxies which are just making their first close approach; tidal distortions take time to grow, so galaxies don't look very distorted until about their first close passage by each other.

In the overlap catalog, we cut a subsample with |z2-z1| > 0.008 as being nonassociated with each other - this is a very conservative cut, since the value is really much smaller except in rich clusters (the redshifts in the catalog show a break in the distribution at about 700 km/s or |z2-z1|=0.0023). For the dual-AGN project, for example, the desirable division goes in the opposite direction - redshift difference smaller than this would give a very high probability of physical association.

zutopian

  • Hero Member
  • *****
  • Posts: 2852
  • REOPEN this forum!!!
    • View Profile
Re: Paper with overlapping-galaxy catalog submitted!
« Reply #31 on: November 29, 2012, 10:28:40 am »
Warning: this topic has not been posted in for at least 120 days.

"Galaxy Zoo: A Catalog of Overlapping Galaxy Pairs for Dust Studies"
William C. Keel, Anna Manning, Benne W. Holwerda, Massimo Mezzoprete, Chris J. Lintott, Kevin Schawinski, Pamela Gay, Karen L. Masters
(Submitted on 28 Nov 2012)
http://arxiv.org/abs/1211.6723

NGC3314

  • Global Moderator
  • Hero Member
  • ****
  • Posts: 1590
    • View Profile
Re: Paper with overlapping-galaxy catalog submitted!
« Reply #32 on: November 29, 2012, 03:25:36 pm »
Zutopian beat me to this announcement! Anyway, more detail on the blog:

Overlaps and backlights and silhouettes. Oh, my!

Thanks to all of you who have made this forum-based project possible. And, as usual, keep an eye out for those background systems. We're still covering more of the sky with better data all the time.

zutopian

  • Hero Member
  • *****
  • Posts: 2852
  • REOPEN this forum!!!
    • View Profile
Re: Paper with overlapping-galaxy catalog submitted!
« Reply #33 on: November 29, 2012, 04:40:23 pm »
Thanks for the update in the GZ blog post about this topic!


« Last Edit: December 02, 2012, 06:47:41 pm by zutopian »

NGC3314

  • Global Moderator
  • Hero Member
  • ****
  • Posts: 1590
    • View Profile
Re: Paper with overlapping-galaxy catalog submitted!
« Reply #34 on: August 16, 2013, 03:23:52 pm »
I just saw this from a listing in the NASA ADS. The overlap catalog itself, as accompanied the PASP paper and hosted on data.galaxyzoo.org,  is now also hosted by SIMBAD and may be queried in many ways through its VizieR service here. They have added positional searches for SDSS DR9 objects near each entry. This interface allows queries filtered on any of the data columns.

zutopian

  • Hero Member
  • *****
  • Posts: 2852
  • REOPEN this forum!!!
    • View Profile
Re: Paper with overlapping-galaxy catalog submitted!
« Reply #35 on: May 20, 2014, 03:34:52 am »
I just saw this from a listing in the NASA ADS. The overlap catalog itself, as accompanied the PASP paper and hosted on data.galaxyzoo.org,  is now also hosted by SIMBAD and may be queried in many ways through its VizieR service here. They have added positional searches for SDSS DR9 objects near each entry. This interface allows queries filtered on any of the data columns.

It is also hosted by NED, but there are just 1603 objects listed, though the GZ overlap catalog contains 1990 galaxy pairs !
Why is the listing in NED incomplete ? http://ned.ipac.caltech.edu/cgi-bin/ex_refcode?refcode=2013PASP..125....2K
e.g. 587743960499880043 is in the overlap catalog, but it hasn't the overlap paper as Ref in NED !
« Last Edit: May 20, 2014, 05:57:45 am by zutopian »

NGC3314

  • Global Moderator
  • Hero Member
  • ****
  • Posts: 1590
    • View Profile
Re: Paper with overlapping-galaxy catalog submitted!
« Reply #36 on: May 29, 2014, 02:15:26 pm »
My first thought, that these might be the objects not already listed by NED, turned out to be incorrect. I sent a query to NED and got the following reply from Marion Schmitz (quoted with his kind permission). Executive summary: it's complicated. Longer form: things that didn't have an unambiguous automatic match with DR6 ObjIDs have to get manual attention (there was a change in SDSS addressing after DR6 which NED hasn't caught up with).

Quote
Hi Bill,

Thanks for your question about the NED links between objects in your paper and objects in NED. There is an explanation, but I'm not sure how "simple" it is.

A few years ago we started a new procedure of doing a first pass processing of journal articles in which named objects in a paper have the links generated just based on the name. Confirming information, such as coordinates, redshifts, nuclear activity, are NOT checked at this stage. This allows us to rapidly make these linkages available to the public. More detailed processing, which incorporates the confirming information, is done at a future time. This latter step allows us to identify inconsistencies, improvements, and errors. As you know, not every piece of information in a paper is accurate.

While processing your 2013PASP..125....2K article last fall, the object names we used for the first pass was the SDSS name listed in the first column realizing that some would not be found in NED yet because you used SDSS DR7/8 and we had loaded DR6 into NED. The "new" names from your paper would be added as cross-identifications in the second pass processing. This method seems to have done a fair job for, as you noticed, 1,603 out of 2,028 (79%) links were quickly made. (The 2,028 differs from your 1,990 because we expanded many of the 26 entries without SDSS names into individual components of pairs as indicated by the comments.)

You may wonder why we used the SDSS name in preference to the "Cross-ID" column. The answer is one of efficiency. The "Cross-ID" column is a mix of multiple names (e.g. NGC 935/IC1801; Arp 276), mix of components and whole systems (e.g. HCG 99, UGC 12899 where the HCG is the whole group), typos (e.g. UGZA 295 which should be UGCA 295; ESO 0320-51 which should be ESO 032012-5150.1), and comments (e.g. "HST pair, in Abell 3558" which appears on three different entries). The SDSS name just required the least amount of "clean-up". The second pass processing would use all of this information, but takes longer to analyze.

When I did look more carefully into the second pass processing, I saw that there were several cases of data which seemed to be swapped between the different components. For example, the velocity given in your table for the eastern object of "HCG 99, UGC 12899" seems to be swapped with that of the western object when I compare it with NED. Other cases showed large differences in velocity compared to what NED has (e.g. IC 1669 has z=0.0907 in your table while NED has z=0.01907 as per the RC3).

Investigation of these differences was taking much longer than I expected and I had to put it off for a while due to other priorities.

Sorry for the detailed explanations, but I thought you would be interested in a short glimpse of the behind-the-scenes procedures we use in NED in order to provide the most accurate information possible. This article takes a bit longer than most.

Cheers,
Marion