What is the scale of the images? And the typical size of point sources (PSFs)?
-
by
JeanTate
As in, how many arcsecs wide is each? And high (I assume the pixel scale is the same in both directions)?
As the IR images are from an instrument with fairly standard optics, I guess that point sources can be fairly well approximated by a Gaussian. What is the typical FWHM of the IR PSF?
For the radio, I guess the PSF is more difficult to describe, given how the radio images are created. And the PSF may vary over the field. What would be a typical range of PSF sizes (for the radio)?
Posted
-
by
42jkb
scientist, admin
The images are 180 arcsec aside. If the radio is compact then it is around 5 arcsec. Does this help?
Posted
-
by
JeanTate
in response to 42jkb's comment.
Yes, thanks.
So the pixel scale is 0.36" (1 pixel is 0.36")? Why? Because the images are 500x500 pix.
Eyeballing an IR image, it seems a compact source (and so the PSF) is about the same size (~15 pix across).
Posted
-
by
enno.middelberg
scientist, translator
The radio PSF indeed is a beast, but it can be approximated by Gaussian, too (the fundamental reason for this is that the PSF is the Fourier transform of the electric field in the entrance pupil, and if that entrance pupil is circular, the PSF will be a sinc function. A sinc function, however, can be very well approximated by a Gaussian).
Beware that the apparent source size can be misleading because of the colour stretch in the images. A faint source will appear to have a smaller PSF, and a bright one a larger PSF. The diameter of the Gaussians at the full-width-half-maximum (FWHM) point, however, is the same.
Posted
-
by
JeanTate
in response to enno.middelberg's comment.
Thanks!
So the FWHM of radio point sources is ~5"?
On color stretches: the radio and IR images use different color schemes; do they both use something like log(flux) scaling (as the human eye does, sorta)? If so, it seems that the dynamic range in radio images is - typically - much smaller than in IR images ... or is that me over-interpreting?
Posted
-
by
enno.middelberg
scientist, translator
The IR uses log(flux), the radio doesn't. It's all a compromise between displaying as much detail as possible, but trying to suppress noise so as not to confuse users.
But you're right in that the IR in general has higher SNR than the radio.
That was well spotted! 😃
Posted
-
by
JeanTate
in response to enno.middelberg's comment.
Thanks!
What about the contours, are they linear?
Is there a fairly simple reason why extra-galactic radio sources have a much narrower range of apparent brightness than IR ones do?
Posted
-
by
enno.middelberg
scientist, translator
Apologies for the delayed reply, I was on holiday.
Contours in radio images mostly increase by some sort of power-law, for example by sqrt(2) or 2. Here, the contours scale as sqrt(3), which is a compromise to capture large dynamic ranges (this would be best done using an increase of factors of 2) while at the same time to accentuate structure (which would call for increases of sqrt(2)). When too many contours are drawn, contours begin to merge. An increase of sqrt(3) turned out to be the best compromise.
The answer to the second question is that there simply are a lot less radio sources than optical sources. If one assumes that the shape of the distribution of apparent source brightnesses is the same in the radio and IR, but there are a lot less radio sources, then it is inevitable that most radio sources seem to be fainter. Take a look at any one (average) image: how many radio sources do you see? Maybe 1 to 3. How many IR sources? Maybe 20-40 or so. Hence in most images there are a few (1-5) "bright" IR sources, while you'd have to click through many images to see a "bright" radio source.
It's a selection effect.
Posted
-
by
JeanTate
in response to enno.middelberg's comment.
Thanks!
Contours in radio images mostly increase by some sort of power-law, for example by sqrt(2) or 2. Here, the contours scale as sqrt(3), which is a compromise to capture large dynamic ranges (this would be best done using an increase of factors of 2) while at the same time to accentuate structure (which would call for increases of sqrt(2)). When too many contours are drawn, contours begin to merge. An increase of sqrt(3) turned out to be the best compromise.
So far, I think the brightest RGZ radio source I've seen (not just in images I've classified, of course) is PKS 2127+04, which is discussed in this thread (it's really cool; among other things it contains the radio equivalent of diffraction spikes!) I tried counting the contours, but it was too hard; what's the dynamic range in that image? And did the (radio) baseline get suppressed?
The answer to the second question is that there simply are a lot less radio sources than optical sources. If one assumes that the shape of the distribution of apparent source brightnesses is the same in the radio and IR, but there are a lot less radio sources, then it is inevitable that most radio sources seem to be fainter. Take a look at any one (average) image: how many radio sources do you see? Maybe 1 to 3. How many IR sources? Maybe 20-40 or so. Hence in most images there are a few (1-5) "bright" IR sources, while you'd have to click through many images to see a "bright" radio source.
It's a selection effect.
(I bolded a key part) You answered a question I'd been trying to formulate ... in both IR and optical images, even many tens of degrees away from the galactic plane, there are always both stars (in our own galaxy) and galaxies1. Yet there seem to be no radio 'stars'2, and the overwhelming majority of radio sources seem to be caused by AGNs (e.g. cores, jets, lobes), even if sometimes not directly (e.g. relics). So in the radio world - at least that of RGZ - we're studying the distribution of radio active AGNs (many of which are also DRAGNs! love the name!!), a single class of physical object (or system).
Or are we?
Last night, as I was browsing this month's astro-ph, I came across "The AGN content of deep radio surveys and radio emission in radio-quiet AGN. Why every astronomer should care about deep radio fields" (Padovani+):
We present our very recent results on the sub-mJy radio source populations at 1.4 GHz based on the Extended Chandra Deep Field South VLA survey, which reaches ~ 30 {\mu}Jy, with details on their number counts, evolution, and luminosity functions. The sub-mJy radio sky turns out to be a complex mix of star-forming galaxies and radio-quiet AGN evolving at a similar, strong rate and declining radio-loud AGN. While the well-known flattening of the radio number counts below 1 mJy is mostly due to star-forming galaxies, these sources and AGN make up an approximately equal fraction of the sub-mJy sky. Our results shed also light on a fifty-year-old issue, namely radio emission from radio-quiet AGN, and suggest that it is closely related to star formation, at least at z ~ 1.5 - 2. The implications of our findings for future, deeper radio surveys, including those with the Square Kilometre Array, are also discussed. One of the main messages, especially to non-radio astronomers, is that radio surveys are reaching such faint limits that, while previously they were mainly useful for radio quasars and radio galaxies, they are now detecting mostly star-forming galaxies and radio-quiet AGN, i.e., the bulk of the extragalactic sources studied in the infrared, optical, and X-ray bands.
Have we come across any "star-forming galaxies" in RGZ? I thought I'd found one, but it seems it contains at least one AGN ...
1 there are other objects, of course, but they are rare enough as to not spoil my point
2 well, hardly; that's another science topic I've already raised here, but haven't yet had a chance to follow up onPosted