Radio Galaxy Zoo Talk

Relative - radio/IR - positional accuracy?

  • JeanTate by JeanTate

    I am not alone in having noticed that, in some images at least, the centroid of a circular, symmetric radio object seems offset from that of its apparent IR counterpart (also a circular, symmetric object). And that, in some images, these offsets seem to be about the same for several of the radio/IR pairs in the image.

    Would an astronomer please write a bit about the distribution of relative positional accuracy (matching) of the radio and IR images? In particular, what proportion of the images do you expect to be mis-aligned by more than a tenth or so of the FWHM of the respective PSFs?

    Posted

  • enno.middelberg by enno.middelberg scientist, translator in response to JeanTate's comment.

    My understanding is that we have three image: a radio image, a IR image and a set of contours in SVG format. These are overlaid/blended in the display. I understand there have been hiccups with correctly aligning these images this, but it appears to be fixed now.

    The positional accuracy of the original data should be much better than a pixel.

    Posted

  • JeanTate by JeanTate in response to enno.middelberg's comment.

    Thanks!

    I hadn't appreciated that the contours are a separate image; they're derived from the radio data, as is the radio image, right (those two images have the same - data - source)?

    The positional accuracy of the original data should be much better than a pixel.

    That's what I expected (and hoped).

    So if the images1 are, in fact, aligned to within ~a pixel, then the apparent offsets are real! 😮 That would be very cool! 😃

    1 They're 500x500 pix, the ones we use to make classifications, right?

    Posted

  • enno.middelberg by enno.middelberg scientist, translator in response to JeanTate's comment.

    So if the images1 are, in fact, aligned to within ~a pixel, then the
    apparent offsets are real! 😮 That would be very cool! 😃

    Well - yes, and if our software manages to get the alignment correct to within that accuracy, of course...

    Posted

  • JeanTate by JeanTate in response to enno.middelberg's comment.

    Hmm, so while it's OK for us to mark objects which seem to be the same only offset (I use #RIoffset), many - perhaps all? - may turn out to be nothing but misalignments (due to your software)?

    Posted

  • JeanTate by JeanTate

    ARG00013hb is a good example of a clear offset. The host - a z_sp=0.663 QSO - is IR-bright, and definitely a radio pixel or two to the E of the (linear) jets. A clear case of misalignment, not some new class of astrophysical object ... 😦

    Posted

  • ivywong by ivywong scientist, admin

    Hmmm, just downloaded the FIRST and WISE images of ARG00013hb and find that the centre positions have an offset of approximately 4 arcseconds. Given that the resolution of WISE and FIRST are around 5-6 arcseconds, it's probable that the QSO you identified is indeed the matching host galaxy. In terms of whether the position offset is real or not, we would probably put this in the "to be followed-up" basket for higher resolution follow-up to verify the offset.

    Posted

  • DocR by DocR scientist

    I don't see an obvious offset with SDSS. see below.enter image description here

    Posted

  • JeanTate by JeanTate

    @ivywong @DocR: is there a easy way to capture Image ARG00013hb as a simple JPEG, as it appears on an ordinary zooite's monitor/screen?

    Why I'm asking is that what I see is quite inconsistent with the JPEG you posted, DocR. So maybe it's an offset between the radio contours and the actual radio emission?

    Here's a screenshot:

    enter image description here

    The 'radio midline' certainly crosses the IR source (an almost perfect point source, at least in the RGZ image), but it definitely misses the centroid of the IR source; hence 'offset'. SDSS, with its much better resolution, puts the 'radio midline' well away from the QSO SDSS J144435.11+410707.9. Move the radio contours a bit to the left, and the offset would disappear ...

    Posted

  • ivywong by ivywong scientist, admin

    Screen shot is the fastest manner to get a quick jpeg. But what I do is to go through both FIRST and WISE to download the ".fits" files using the FIRST coordinates given in Talk. Then I can use DS9 (as DocR has done) to find out roughly where the coordinate centres of both WISE and FIRST. The 4 arcsec offset that I calculated were from WISE and First without SDSS so I suspect the offset may be due to the relatively low resolution of the WISE and first images which are around 6 and 5 arcseconds respectively. That is why the 4 arcsec that you see in the jpeg you posted is not classed as a significant offset.

    Posted

  • enno.middelberg by enno.middelberg scientist, translator

    You can also use the Virtual Observatory, for example "Aladin" at http://aladin.u-strasbg.fr/java/nph-aladin.pl. Enter the coordinates of the source and add the FIRST images via File -> Load Astronomical Image -> VLA... -> FIRST. Then add an infrared or optical image. Select the radio image and add contours by clicking the contour icon near the sliders at the bottom right of the display.

    Posted

  • JeanTate by JeanTate in response to enno.middelberg's comment.

    Thanks everyone. 😃

    I've seen zooites use Aladin before, especially c_cld, and have DS9 installed on my laptop (and Linux box too; it was widely used in SpaceWarps, I think). DS9 is very good, but its lack of detailed help, tutorials, etc means it's quite a struggle to learn to use well.

    I'm going to use Aladin with all the "offset" objects in my collection, to see if any seem real; so far (provisionally!), it seems the only 'real' offsets are between the radio emission and the contours (i.e. the radio and optical/IR seem to be aligned to within ~1-2", but the contours are - at least sometimes - offset by >~5").

    No new class of radio object ... 😦

    Posted

  • JeanTate by JeanTate in response to enno.middelberg's comment.

    Thanks again, enno.middelberg! 😃

    Here's ARG00013hb/SDSS J144435.11+410707.9, from Aladin, after a bit of processing:

    enter image description here

    It's the SDSS r-band and FIRST data; the cross-hairs are centered on the coordinates of the SDSS QSO*.

    Here's my screenshot again:

    enter image description here

    I'm still getting the hang of it, but a straight line connecting the 'peaks' of each lobe passes through the 'tips' of each of the contours, and no more than ~1" from the centroid of the QSO. That's well within the ~5" FIRST resolution, so no offset here. 😃

    *more details? just ask!

    Posted

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

    I'm going to use Aladin with all the "offset" objects in my collection

    Here's ARG0003e3r:

    enter image description here

    Again, FIRST and SDSS r-band; the cross-hairs are centered on z_ph ~0.48 SDSS J091049.55+060643.3, which is also the centroid of a WISE (IR) point-source (can't add WISE data to Aladin images, as far as I can tell 😦).

    This may not be a offset; rather, there may be an invisible (IR and optical) host, at the mid-point of this otherwise very normal-looking #hourglass/doublelobe.

    Posted

  • JeanTate by JeanTate

    ARG00016al is, radio-wise, a compact source, perhaps even a point source.

    The fairly large, and rather odd-colored, SDSS J161912.12+394357.1 is the apparent host, with a spectroscopic redshift of 0.644, and a spectrum showing some activity (i.e. it's not totally red and dead):

    enter image description here

    Given that the radio source is such a classic point source, there's no doubt of an offset, of ~3" (radio contours, SDSS i-band image):

    enter image description here

    Chance coincidence, with an SDSS-invisible objecting hosting the radio source? Or a strange one-sided lobe?

    Posted

  • xDocR by xDocR

    I think chance coincidence, given how accurate other positions look. The problem we're facing is that we're looking at so many sources, that even the rare and coincidental situations are regularly popping up. Haven't figured out yet how to decide which oddities are really worth pursuing and which are just accidental. E.g., we've got a beautiful filament/jet connecting a z=0.26 galaxy (radio source) with a z=0.049 galaxy (radio source). ARG00025v9, i think. Needs some real thought.

    Posted

  • JeanTate by JeanTate

    Not an IR-radio (apparent) offset, but an optical (SDSS)-radio one:

    enter image description here

    More details in the Merger of Two galaxies with jets ? thread.

    Boilerplate: SDSS image per http://skyservice.pha.jhu.edu/DR10/ImgCutout/getjpeg.aspx, FIRST contours derived from FITS files produced using SkyView with Python code described in this RGZ Talk thread. Image center per the ARG0001iwu image (J2000).

    Posted

  • JeanTate by JeanTate

    Reviving an old thread ... I've come across several examples of apparent SDSS-FIRST offsets, and would like to get a discussion going on them. So I'll be posting them here, one post per object.

    In searching for this thread, I discovered that there are several other threads with a similar topic, or which contain examples of this kind of apparent offset; in no particular order:

    Posted

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

    Before starting to examine objects which seem to have SDSS/FIRST offsets, a quick look at an object which seems to have no offset, SDSS J142042.91+262503.2/FIRSTJ142042.9+262503, the radio source at the center of ARG0001xvf:

    enter image description here enter image description here

    The SDSS J2000.0 position (RA, Dec) is (215.17881, 26.41758), and the FIRST position (215.17884, 26.41757); the WISE position is (215.17887, 26.41761).

    The scale of the images in the first part of this post is 0.24"/pixel. The three numbers - in red - at the top left are:

    • (left) the threshold (in σ);
    • (center) the contour spacing in intensity units; and
    • (right) the Gaussian smoothing parameter for the contours.

    Varying the values of these three contour parameters make essentially no difference to the apparent offset between SDSS and FIRST (which is ~zero):

    enter image description here enter image description here enter image description here enter image description here enter image description here enter image description here enter image description here

    As the source of the radio emission seems to be the unresolved nucleus of the host galaxy, and as that nucleus is an AGN (it appears as a point source in the photometric image, and its spectrum is that of an AGN), it would be quite surprising for there to be an offset.


    If the scale is changed, there's no apparent change in the SDSS-FIRST offset (~zero, in this case). As can be seen in these 'variations on a theme':

    enter image description here enter image description here
    enter image description here enter image description here
    enter image description here enter image description here

    enter image description here enter image description here
    enter image description here enter image description here

    The images in this post were created from sources, and using methods, described in this RGZ Talk post. The object at the center of the images is SDSS J142042.91+262503.2.

    Posted

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

    Intro out of the way, time for some real examples of apparent SDSS-FIRST offsets.

    SDSS J223854.56+034255.7/FIRSTJ223854.4+034256 is the object at the center of ARG0003l57; the SDSS J2000.0 position (RA, Dec) is (339.72737, 3.71548), and the FIRST position (339.72682, 3.71559); the WISE position is (339.72724, 3.71555). The difference between the SDSS and FIRST positions, for this object, is considerably greater than for SDSS J142042.91+262503.2/FIRSTJ142042.9+262503 (above), so the images (below) are not "lying":

    enter image description here enter image description here
    enter image description here enter image description here

    However, as a zoomed-out FIRST image shows, there seems to be some kind of artifact, which may be causing at least some of the apparent offset:

    enter image description here

    The images in this post were created from sources, and using methods, described in this RGZ Talk post. The object at the center of the images is SDSS J223854.56+034255.7.

    Posted

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

    Unlike the previous example, there's no apparent artifact around; however, there is obvious - bright! - radio structure/emission nearby ... (zoomed out image later below)

    enter image description here

    Introducing SDSS J153249.44+573424.9/FIRSTJ153249.5+573425, at the center of ARG000089t.

    The SDSS J2000.0 position (RA, Dec) is (233.20602, 57.57361), and the FIRST position (233.20623, 57.57377); the WISE position is (233.20641, 57.57401), which is not very informative. The difference between the SDSS and FIRST positions, for this object, is considerably greater than for SDSS J142042.91+262503.2/FIRSTJ142042.9+262503 (above), so the images (below) are not "lying":

    enter image description here enter image description here

    The images in this post were created from sources, and using methods, described in this RGZ Talk post. The object at the center of the images is SDSS J153249.44+573424.9.

    Posted

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

    SDSS J120312.94+182322.0 is the object near the center of ARG0002gbx.

    The SDSS J2000.0 position (RA, Dec) is (180.80392, 18.38946); the WISE position is (180.80389, 18.38948). There is no FIRST catalog source near this position; FIRSTJ120312.5+182310 - at the center of ARG0002gbx - is a lobe. Perhaps the apparent offset is due to the fact that 'nuclear/central' source of this triple is so faint?

    enter image description here enter image description here

    The image in this post was created from sources, and using methods, described in this RGZ Talk post. The object at the center of the image is SDSS J120312.94+182322.0.

    Posted

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

    SDSS J123354.73+254123.0/FIRSTJ123354.8+254123 is at the center of ARG0001zgu.

    The SDSS J2000.0 position (RA, Dec) is (188.47807, 25.68973), and the FIRST position (188.47843, 25.68986); the WISE position is (188.47795, 25.69006). The offset is small; perhaps because the radio emission is extended, almost an hourglass shape?

    enter image description here enter image description here

    The images in this post were created from sources, and using methods, described in this RGZ Talk post. The object at the center of the images is SDSS J123354.73+254123.0.

    Posted

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

    SDSS J155203.47+073141.9/FIRST J155203.4+073142, above the center of ARG0003a3y.

    The SDSS J2000.0 position (RA, Dec) is (238.01446, 7.52833), and the FIRST position (238.01439, 7.52852); the WISE position is (238.01452, 7.52835). The offset is small, but non-zero.

    enter image description here
    enter image description here

    The images in this post were created from sources, and using methods, described in this RGZ Talk post. The object at the center of the images is SDSS J155203.47+073141.9.

    Posted

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

    The previous objects are from my database of candidate SDRAGNs* (see this thread for details); the next set of objects comes from my IR/radio offset Collection.

    SDSS J111321.31+032300.8/FIRST J111321.5+032258 is near the top of ARG0003m41, and is the central object of a triple.

    The SDSS J2000.0 position (RA, Dec) is (168.33880, 3.38356), and the FIRST position (168.33994, 3.38281); the WISE position is (168.33895, 3.38348). The offset is quite large, suggesting that the host is far in the background and/or too faint to be detected in SDSS. In turn, that might mean this is a giant.

    enter image description here
    enter image description here

    The images in this post were created from sources, and using methods, described in this RGZ Talk post. The object at the center of the images is SDSS J111321.31+032300.8.

    *of course, not all of them are SDRAGNs!

    Posted

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

    SDSS J073849.60+403549.7/FIRSTJ073849.7+403544 is at the center of ARG00014kv.

    The SDSS J2000.0 position (RA, Dec) is (114.70669, 40.59715), and the FIRST position (114.70719, 40.59578); the WISE position is (114.70603, 40.59576). The offset is quite large, suggesting that the host is not SDSS J073849.60+403549.7; perhaps it's one of the not-PO red blobs near the neck of the hourglass?

    enter image description here
    enter image description here

    The images in this post were created from sources, and using methods, described in this RGZ Talk post. The object at the center of the images is SDSS J073849.60+403549.7.

    Posted

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

    SDSS J161912.12+394357.1/FIRSTJ161911.9+394355 is at the center of ARG00016al.

    The SDSS J2000.0 position (RA, Dec) is (244.80050, 39.73254), and the FIRST position (244.7997, 39.73203); the WISE position is (244.80052, 39.73243). The offset is considerable, and strange; a quite bright LRG (luminous red galaxy) which seems to not be the host of a radio-bright compact source. Is there a hint of an arc? Could this be an SGL perhaps?

    enter image description here
    enter image description here

    The images in this post were created from sources, and using methods, described in this RGZ Talk post. The object at the center of the images is SDSS J161912.12+394357.1.

    Posted

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

    SDSS J150543.38+195304.0 is surely not the host of doublelobe/hourglass FIRST J150543.3+195301 (in ARG0002d5c):

    enter image description here
    enter image description here

    The images in this post were created from sources, and using methods, described in this RGZ Talk post. The object at the center of the images is SDSS J150543.38+195304.0.

    Posted

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

    Somewhat odd triple SDSS J111533.23+182156.3/FIRSTJ111532.9+182153 is near the center of ARG0002gdn.

    But there's no SDSS PO closer to the apparent position of the host (middle of the triple), so it's apparently another chance alignment, with an invisible (in SDSS and WISE) host:

    enter image description here
    enter image description here

    The images in this post were created from sources, and using methods, described in this RGZ Talk post. The object at the center of the images is SDSS J111533.23+182156.3; note that the code for the overlay image has the 'd' and 'g' switched.

    Posted

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

    SDSS J160008.75+491502.8/FIRSTJ160008.6+491456 is near the center of ARG0000mt7.

    Another chance alignment, with the host of the #bent #wat invisible in both SDSS and WISE?

    enter image description here
    enter image description here

    The images in this post were created from sources, and using methods, described in this RGZ Talk post. The object at the center of the images is SDSS J160008.75+491502.8.

    Posted

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

    SDSS J142215.96+340510.1 is yet another LRG, with a z_ph of ~0.4 (at the center of ARG0001hmw). Positionally, it seems it could the host of asymmetric doublelobe/hourglass FIRSTJ142215.6+340513, yet it's quite some way from the neck (saddlepoint).

    Yet another chance alignment, with the actual host being invisible in both SDSS and WISE?

    enter image description here
    enter image description here

    The images in this post were created from sources, and using methods, described in this RGZ Talk post. The object at the center of the images is SDSS J142215.96+340510.1.

    Posted

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

    SDSS J071933.53+322406.4/FIRSTJ071933.7+322411 is at the center of ARG0001l5p.

    But it's a STAR, and is clearly offset in position from the 'ridgeline' of the radio emission. While the WISE resolution is considerably poorer than that of SDSS, the WISE source seems to be at ~the same position as the SDSS one; the SDSS J2000.0 position (RA, Dec) is (109.88974, 32.40179), and the WISE position is (109.88983, 32.40183).

    One more chance alignment?

    enter image description here
    enter image description here

    The images in this post were created from sources, and using methods, described in this RGZ Talk post. The object at the center of the images is SDSS J071933.53+322406.4.

    Posted

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

    SDSS J091027.88+211800.3/FIRSTJ091028.2+211747, near the center of ARG0002a48, is a complex source.

    SDSS J091027.88+211800.3, with a z_sp of 0.405, looks like the BCG/cD of a cluster; pretty much every one of the yellow-ish/red-ish blobs in the SDSS image below have z_ph values consistent with 0.4. [1] Yet it is clearly not positionally placed to be the host of the radio emission. And there's no apparent SDSS - or WISE - source near the saddle point. A genuine optical+IR/radio offset? A chance alignment? Something else?

    enter image description here
    enter image description here

    The images in this post were created from sources, and using methods, described in this RGZ Talk post. The object at the center of the images is SDSS J091027.88+211800.3.

    [1] there are two obvious exceptions; the more interesting is z_ph ~0.71 SDSS J091027.87+211749.7:

    enter image description here

    Posted

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

    SDSS J135034.23+174027.7/FIRSTJ135034.3+174031 is at the center of ARG0002hyr.

    To SDSS it's a STAR, but spectroscopically it's a z=2.182 broadline QSO; the radio bright lobes are thus no surprise. However, it's not quite in the middle of the saddle; real offset, or some kind of artifact?

    enter image description here
    enter image description here

    The images in this post were created from sources, and using methods, described in this RGZ Talk post. The object at the center of the images is SDSS J135034.23+174027.7.

    Posted

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

    SDSS J120209.28+132922.8/FIRSTJ120209.4+132919 is at the center of ARG0002rzn.

    Like several others I've posted recently, the SDSS object - which is ~at the same position as the WISE one, SDSS J2000.0 position (180.53868, 13.48968) cf WISE (180.53868, 13.48956) - seems to be an LRG (it's perhaps too far away to tell if it's a BCG, or in a cluster at all). Further, it's on the 'ridgeline' of the radio emission, but not ~at the maximum; nor is there a saddle. Chance alignment? Asymmetric lobes/jets? Or perhaps this is the S lobe of an overedge doublelobe? (I'll post a zoomed-out overlay later).

    enter image description here
    enter image description here

    The images in this post were created from sources, and using methods, described in this RGZ Talk post. The object at the center of the images is SDSS J120209.28+132922.8.

    Posted

  • JeanTate by JeanTate

    SDSS J134553.11+124742.1/FIRSTJ134553.4+124745 is an apparent z_ph 0.7 compact/LRG at the center of ARG0002tos.

    The SDSS J2000.0 position (RA, Dec) is (206.47131, 12.79504), and the FIRST position (206.47279, 12.79593); the WISE position is (206.47135, 12.79496). The SDSS/WISE offset is essentially zero; the offset with FIRST considerable, and strange; a quite bright LRG which seems to not be the host of a radio-bright compact source. Perhaps the host is an LRG in a cluster which has SDSS J134553.11+124742.1 as a member, but which is too faint to be detected by either SDSS or WISE?

    enter image description here
    enter image description here

    The images in this post were created from sources, and using methods, described in this RGZ Talk post. The object at the center of the images is SDSS J134553.11+124742.1.

    Posted

  • JeanTate by JeanTate

    SDSS J144616.53+193004.6 is near the location of the host of the apparent triple whose E lobe is FIRSTJ144618.2+193001, near the center of ARG0002dx9.

    SDSS J144616.53+193004.6 is a purple STAR, the kind of SDSS object which, when associated with a FIRST compact source, often turns out to be an (unrecognized) QSO. However, in this case is seems it's truly just a field star, and that the host is close to a blob which the SDSS photometric pipeline considers to be noise. The SDSS J2000.0 position (RA, Dec) is (221.56888, 19.50128), and the WISE position is (221.56978, 19.50016); two distinct objects. The central FIRST source, at (221.56969, 19.49954), is close to neither the WISE nor SDSS positions.

    enter image description here
    enter image description here

    The images in this post were created from sources, and using methods, described in this RGZ Talk post. The object at the center of the images is SDSS J144616.53+193004.6.

    Posted

  • JeanTate by JeanTate

    SDSS J075753.17+364022.0 is the apparent host of the triple whose SW lobe is FIRSTJ075751.0+363943, in ARG0001c5f.

    However, it's not:

    enter image description here
    enter image description here

    The images in this post were created from sources, and using methods, described in this RGZ Talk post. The object at the center of the images is SDSS J075753.17+364022.0.

    Posted

  • WizardHowl by WizardHowl

    A thought about ARG000089t now I've had it to classify - could this be a hidden AGN? There does seem to be an interaction between two adjacent galaxies and the offset is such that a thick region of dust adjacent to the nearest galaxy might be obscuring the optical signs of an AGN hosting the radio core of the bent triple. Does this seem plausible?

    Posted