Page 6 of 6 FirstFirst ... 456
Results 151 to 179 of 179

Thread: Bias effects in galaxy detection

  1. #151
    Join Date
    Sep 2003
    Posts
    10,399
    Thanks for the explanation, parejkoj.
    We know time flies, we just can't see its wings.

  2. #152
    Join Date
    Mar 2004
    Posts
    13,441
    Introducing T10.

    Like T8, and all the other toy galaxies to date (except T9), it is a uniformly luminous disk; unlike them, its radius is 3.2 kpc.

    Like T8, it emits only at the (restframe) wavelengths 265, 350, 460, 558.3, and 666.5 nm.

    If we regard each line emission as a separate galaxy, the Mbol's would be -17.2, -19.4, -21.1, -21.7, and -22.2, respectively. Combined, this gives T10 an Mbol of -23.

    By my criteria, T10 becomes indistinguishable from a point source somewhere between z = 0.25 and 0.5.

    In terms of detectability, it would be pretty obvious locally - in each of the n2t, ut, gt, rt, and it bands - locally and at z = 0.1. It would have 'sunk' in the ut band by z = 0.25, and the gt band by z = 0.5 (i.e. it'd be a point source in only three bands, it, zt, and IR1t). This would become just two bands at z ~1, IR1t, and IR2t. Somewhere between z = 1 and z = 1.5, T10 would sink entirely.

    If, somehow, we could observe T10 across all 10 toy bands in an integrated fashion, it would be barely detectable at z = 1.5, and have sunk at higher redshifts.

  3. #153
    Join Date
    Mar 2004
    Posts
    13,441
    Quote Originally Posted by parejkoj View Post
    Can I answer?
    Of course you can!

    And I see that you did, thanks.

    I don't know the cost of the telescope used in the NDTSS, but most surveys that get good signal-to-noise at ~20th magnitude and 1.4" seeing have costs in the tens of millions of dollars. "Changing the magnification" means building another telescope, in a location with better seeing.

    Sorry.

    Nereid: I haven't seen a proper cost writeup for the NDTSS? What kind of construction costs are you looking at?
    Fair question.

    Let's see now ... PC, ~€1,500; software, free; internet connection, ~€50/month; time ... I'd rather not say.

    Building telescopes, with snazzy filters and automated operations (e.g. observation scheduling, data reduction pipelines), in my toy universe is astonishingly cheap. However, I'm so happy that someone else has invested hundreds of millions (billions?) of $/£/€ in real telescopes, etc, and millions of hours of effort in writing up and publishing results. This has enabled me to build my toy survey - so far, just the NDTSS - in a way that rather nicely approximates a very real survey (SDSS).

  4. #154
    Join Date
    Mar 2004
    Posts
    13,441
    Introducing T11.

    Like T8, and all the other toy galaxies to date (except T9), it is a uniformly luminous disk; unlike them, its radius is 32 kpc.

    Like T8 and T10, it emits only at the (restframe) wavelengths 265, 350, 460, 558.3, and 666.5 nm.

    If we regard each line emission as a separate galaxy, the Mbol's would be -17.7, -19.7, -21.2, -21.7, and -22.1, respectively. Combined, this gives T11 an Mbol of -23.

    Being much bigger than T8 and T10, T11 remains an extended source at all redshifts.

    However, its large size means that its SB is low; so low in fact that it would not be detectable, even locally, in the n2t and ut bands. And in the gt band, it'd be an LSB (low surface brightness) galaxy even locally; by z=0.25 it'd be marginally undetectable in the rt band, but clearly detectable in both the it and zt bands.

    By z = 0.5 it would be marginally detectable in only the IR1t band.

    If we could observe it across all bands, it would be marginally detectable at z = 0.75, but would have sunk by z = 1.

    Now suppose T8, T10, and T11 were, in fact, the same galaxy. In other words, an axisymmetric disk (in terms of uniformity of emission), with three sharp jumps in surface brightness, at radii of 3.2 kpc, 10 kpc, and 32 kpc. Stay tuned!

  5. #155
    Join Date
    Mar 2004
    Posts
    13,441
    OK, I'm a tease.

    In this post, instead of examining T12, the superposition - in some sense - of T8, T10, and T11, I'm going to look at three variants of T1, a red (T13), white (T14), and blue (T15) galaxy [1].

    T14, the white galaxy.

    Remember, it's a uniformly luminous disk, 10 kpc in radius, with an Mbol of -24. T14 emits light at ten discrete wavelengths: 153, 259.5, 366.3, 469.3, 604.3, 748.7, 903.7, 1085.5, 1318.8, and 1615 nm (restframe). Each Mtoyband is thus -21.5. As with T1, T14 is always an extended object in NDTSS, no matter what the redshift. By z=1, it has sunk, if observed by/in the NDTSS, in terms of per-band integrated magnitude ... and even if it hadn't, it's a dropout in the ut, rt, and IR1t bands (i.e. none of the lines gets redshifted into these bands). However, it doesn't disappear until a somewhat higher redshift (but below 1.5), in terms of falling below the SB limit.

    T13, the red galaxy.

    This galaxy is also a uniformly luminous disk, 10 kpc in radius, with an Mbol of -24. Like T14, it emits light at the same ten discrete wavelengths. However, the ratio of flux at any two such is 1.2, with the longer wavelength having the greater flux. For this galaxy, the 'disappears in integrated magnitude in a given band before disappearing, in the same band, because the SB drops too low' (whew) is the same, as for T1, T2, and T14. However, the redshift at which it sinks is different: the shortest wavelength emission (restframe 153 nm) - which by then is observed in the n2t band - has sunk before z=1. It's still detectable in the next (259.5 nm), in the gt band, at this redshift, but is gone by z=1.5, It's the same for all the other emission lines, except the last four. However, for them, redshift has taken them beyond the IR3t band, so they won't be detected.

    T15, the blue galaxy.

    Ditto, except that the 1.2 ratio is for the flux at the shorter wavelength of each pair. Applying the same test (i.e. looking only at when the SB falls too low), T15 sinks only at a redshift a tad above z=2, at the shortest wavelength (where it's detectable in the gt band). Ditto for the next longest emission line (iz band). At longer wavelengths, the galaxy sinks well before the SB limit is reached, because these wavelengths are redshifted beyond the IR3t band. In other words, despite having the exact same Mbol, T15 remains detectable until (somewhat after) z=2, as a galaxy, whereas T14 (the white galaxy) and T13 (the red galaxy) have sunk by z=1.5.

    But how realistic are these red, white, and blue galaxies? Not particularly. However, if you don't know what the restframe SED of a galaxy is, outside the tiny sliver of the electromagnetic spectrum astronomers doing surveys in the 1960s were limited to (UBVR, to summarise), the Disney&Lang analysis is severely limited.

    In terms of flux, a galaxy will usually have an unambiguous - if sometimes rather broad - maximum in its SED. After all, two of the three main components of real galaxies - stars, gas/plasma, and dust (AGNs are ignored in this analysis) - can be crudely approximated by Plankian SEDs (gas/plasma is the exception). However, those maxima can certainly be outside the 'optical' (i.e. UBVR); a galaxy with lots of young, bright blue stars may have its SED's maximum in the UV, and a dust-choked one, well into the MIR.

    In fact, each of T8, T10, and T11 have SEDs that are still rising at the red end.

    OK, can we look at T12 now? Stay tuned!

    [1] No, I'm not necessarily feeling love for the USofA; do you know how many of the world's flags contain red, white, and blue (and only red, white, and blue)? Sticking to only major ones, and going alphabetically, Australia, Cambodia, Chile, Cuba, Czech Republic, ...

  6. #156
    Join Date
    Mar 2004
    Posts
    13,441
    OK, the tease continues.

    In this post, instead of examining T12, the superposition - in some sense - of T8, T10, and T11, I'm going to introduce you to a marvelous camera, the TACS (toy ACS, get it?), which is mounted on an awesome platform, the THST.

    The TACS is equipped with ten filters, as earlier, and the TACS+THST combo has wonderful optics. And, being out of this world (i.e. above the Earth's atmosphere) its PSF is a mere 0.05" (FWHM)! Using this facility, galaxies become undetectable - as point sources - at 26.5 mag, in all bands (vs 22.2 in the NDTSS). In terms of SB, this facility is better than the NDTSS too; the cutoff is 27 mag per arcsec squared, in all bands (vs 25 in the NDTSS).

    How do T8, T10, and T11 fare, when observed by TACS+THST? What about T13, T14, and T15?

    Well, T10 never becomes confused with a point source; all galaxies are easily resolved as extended sources.

    The white galaxy (T14) sinks around (a bit before) z=3, and the blue (T15) a tad beyond z=4; in both cases the SB falls below 27. And the red (T13)? Somewhere between z=2 and z=3. As with being observed in the NDTSS, the problem is that the wavelengths in which T13 is brightest have been redshifted beyond the IR3t band, so the galaxy sinks below the SB limit more quickly. In terms of integrated magnitude, within a band, the red and white galaxies disappear somewhere between z=5 and 6; the blue galaxy remains easily detectable at z=6.

    T11. Its low SB means that it will sink at quite modest redshift, cosmologically speaking. Even locally, it would be an LSB-too-farfaint in the n2t band (though its integrated magnitude would make it quite impressive). In whatever toyband the emission lines were redshifted to, T11 would be detectable at higher z than in the NDTSS, out to z=1.5 (well, not quite), in the IR3t (or perhaps IR2t) band.

    T8: Easily detectable to z=2; sunk by z=3, mostly because the brightest emission lines have redshifted beyond the IR3t band.

    T10: This galaxy's greater SB, in all toybands, means that it's much more easily detectable. Even in the shortest wavelength emission line - redshifted to the gt band - it's easily detectable at z=0.75 (but has sunk at z=1). At the next line (350 nm), its SB does not lead it to sink until z=3 (although its integrated magnitude is below the 26.5 limit). Beyond this it disappears because the brighter lines have been redshifted beyond the IR3t band; if the TACS were equipped with a suitable MIR (?) photometer, T10 would be easily detectable at z=6.

    In summary, the TACS+THST is certainly able to detect these toy galaxies at cosmological distances, some of them almost to z=5 or 6.

    Quote Originally Posted by Disney&Lang
    The Visibility Window depicted in fig 3 is immutable, mathematical and pinned in local coordinates because it shows the contrast to ones local sky, be it on the ground or in space. What we need to calculate next are the properties, in particular the sizes and intrinsic SBs, of the kinds of galaxies, seen at different redshifts, which will make it through that narrow window, particularly near its peak, taking into account the Tolman e ects described above, which both dim a galaxy and increase its apparent size.

    [...]

    Even at z = 0.5 many of the most Visible galaxies that were in region A (Fig 1) at low redshift would be translated into region C and be far too dim to see. They have Sunk.

    [..]

    More than 50 per cent of the light from a galaxy that would be at the peak nearby, has already been lost at redshift 0.5, 82 per cent at redshift 1, and all by 1.2. These figures alone are enough to query the feasibility of trying to study galaxy evolution by using deep fields.

    [..]

    If now we remove the Local population to redshift 1, virtually all the previously prominent galaxies will sink below the sky thanks to Eqn. (43).
    At least in the toy universe in which I've been doing my observing, what Disney & Lang write is, not to put too fine a point on it, hogwash.

    So, you must be thinking, how closely does my toy universe resemble the real one?

    But first, can we look at T12 now, please? OK, sure thing, in my next post ...

  7. #157
    Join Date
    Mar 2004
    Posts
    13,441
    I've promised to look at T12, the superposition - in some sense - of T8 (a red, uniform, 10 kpc disk, Mbol -22.9), T10 (a red, uniform, 3.2 kpc disk, Mbol -23), and T11 (a red, uniform, 32 kpc disk, Mbol also -23).

    First, though, I'd like to point out some things mildly interesting about these three galaxies. Despite their very different sizes, they have pretty much the same Mbol, each a full magnitude fainter than T1. Their colours are similar, but not the same. Let's observe them at z=0.25, so each of the five emission lines falls in just one toy band, and those bands somewhat resemble the SDSS ones (for brevity I've shortened the band names; the first colour should be ut-gt).


    Tn. u-g g-r r-i i-z
    T08 2.0 1.6 0.6 0.4
    T10 2.2 1.7 0.6 0.5
    T11 2.0 1.5 0.5 0.4

    OK, has a lightbulb gone off in anyone's head yet?

    T12 is built from T8, T10, and T11 as follows: the core is T10. Cut out a disk of radius 3.2 kpc from T8, so that it's now an annulus. Plonk this down on top of T10, so the centres coincide (all disks are in the same plane). Cut out a disk of radius 10 kpc from T11, so it too is now an annulus. Plonk it down on top of T10+T8' (i.e. the modified T8). Anyone want to have a go at saying how visible T12 would be, in the NDTSS? In a survey taken with the TACS+THST?

    Now I had planned to repeat this exercise, with modifications/extensions of T1 (or T13, T14, and T15), building up to my grand finale, ... T31 (get it?). Along the way I'd have done a slow reveal for T12, and any extensions etc necessary to make T31 seem like an amazing coincidence.

    However something funny happened on the way to the forum before I could get very far.

    What? Well, a dog failed to bark in the night.

    Specifically, an interesting paper/preprint did not get mentioned.

    But that story is going to have to wait, until another day ...

  8. #158
    Join Date
    Mar 2004
    Posts
    13,441
    Just in case any reader is interested in the how of my toy universe observing (are you such a reader? please speak up if you are!), here are some details.

    I particularly like ngc3314's approach, as I've already said.

    First, it starts with the total energy flux, across the entire electromagnetic spectrum. This is what the Tolman dimming is keyed to, so the physics is sound.

    Next, the uniformly luminous circular disk, normal to our line of sight, is both an easily understood thing (element), and one that can be easily and robustly manipulated. I call it an element because it acts like a Lego block; you can readily add lots together to make realistic models. For example, as I started to do with T12, a series of nested disks can approximate an exponential SB profile, or a de Vaucouleurs one (β=4), or any Sérsic profile, or ... and this can be done to essentially an arbitrary degree of detail.

    Independently, the SEDs of the disk elements can be approximated by discrete emission lines, chosen to fall just where you want them in terms of the photometry you want, or plan, to do. Because the disks' total emission - energy flux - is what you start with, allocating this across bands also gives you colours, for free as it were. Sure, it's a real pain to have to be constantly switching back and forth between energy/power (e.g. joules/watts) and magnitudes (AB system? STMAG system? have I messed up my zero points? ...), and the chances that you'll mess up very high, but at least you can always see what's going on, in a (relatively) straight-forward and intuitive way.

    The work I've done to date has been done in a spreadsheet. I'm sure it could be taken a lot further, especially with macros; however, I think it'd be a lot easier to switch to code. For one thing, it's surely a lot easier to go from three nested disks to 100 in code than it is in a spreadsheet; ditto ten emission lines to 1,000. And I'd planned on doing so, until I stumbled upon the dog which did not bark ...

    Anyone yet worked out what real universe galaxy T12 approximates?

  9. #159
    Join Date
    Mar 2004
    Posts
    13,441
    Here are some (edited) extracts from the preprint; I'll give full details in a later post.

    Our sample of {galaxies} is drawn from the HST imaging with Advanced Camera for Surveys (ACS) and WFC3, which was obtained as part of the [...] program. Near-UV and near-IR observations were acquired as part of the WFC3 [...] program [...] a 104 orbit medium-depth survey using the HST UVIS and IR cameras. A general introduction to the performance and calibration of the WFC3 is provided in Windhorst et al. (2011). The [...] program observed approximately 50 square arcminutes in the GOODS-S field with the HST WFC3 UVIS in three filters: F225W and F275W for 2 orbits, and F336W for 1 orbit, per pointing, respectively. The program observed approximately 40 square arcminutes in the same field with the WFC3 IR in three filters: F098M, F125W, and F160W, each for 2 orbits per pointing. The 5σ 50% point-source completeness limits are: F225W=26.3, F275W=26.4, F336W=26.1, F098M=27.2, F125W=27.5, and F160W=27.2 mag (see Windhorst et al. 2011). The analysis presented here was completed using mosaicked images produced for each of the UVIS and IR band tilings, and each image mosaic was drizzled to a pixel scale equal to 0.090" pixel-1.

    [...]

    The WFC3 mosaics roughly cover the northern one-third of the GOODS-S field (Giavalisco et al. 2004), and we incorporate the pre-existing ACS dataset (F435W, F606W, F775W, and F850LP) with the WFC3 observations. We produced mosaicked images of the GOODS-S ACS data, which were binned to match the pixel scale of the WFC3 UVIS/IR mosaics.
    Selection Criteria

    We require our galaxies to have: (1) been imaged in all UV and IR bands, to uniform depth; (2) a spectroscopically-confirmed redshift in the range 0.35 ~< z < ~1.5; and (3) an {galaxy} morphology.

    There are many techniques for identifying {galaxies} at intermediate redshift. [...] However, the robustness of each of these classifiers can be dramatically affected by a variety of systematics, such as the image signal-to-noise ratio (Conselice et al. 2003; Lisker 2008) and the bandpass in which the technique is applied (Taylor-Mager et al. 2007; Conselice et al. 2008). In lieu of these techniques, we select our sample by visual classification. This technique is subjective, and as such can introduce new biases, but it has been successfully applied to the identification of both low redshift (z ∼0.1; [...]) and intermediate redshift (z ~< 1.3; [...]) {galaxies}. We will demonstrate in §{Y} that the spectroscopic redshift requirement, and not the morphological selection technique, is the most significant source of bias.

    [...]

    UV imaging can provide unique insight into {something interesting}. Thus, we require our sample {galaxies} to be observed in each of the UV filter mosaics. To ensure that all galaxies were observed to a similar depth, we also require each {galaxy} in the sample to be observed in the UV and IR image mosaics for at least the mean exposure time measured for each filter as given by Windhorst et al. (2011). Since we are interested in {something interesting}, and the WFC3 UVIS channel is only sensitive to UV emission at  ∼ 1500 Å for objects at redshift z ~> 0.35, we define this redshift as low-redshift cutoff of the sample. The high-redshift cutoff was selected to ensure that the visual inspection and classification of the {galaxy} — in the filter set outlined above – considers the rest-frame V-band morphology. We are sensitive to at least the UV-optical SED of every {galaxy} in our catalog.

    [...]

    We find 102 {galaxies} that satisfy these selection criteria.
    Photometry

    We measured object fluxes using SExtractor in dual-image mode (Bertin & Arnouts 1996), with the WFC3 F160W image as the detection band. For source detection, we required sources to be detected in minimally four connected pixels, each at ≥ 0.75σ above the local computed sky-background. For deblending, we adopted a contrast parameter of 10−3 with 32 sub-thresholds. Object photometry was determined with MAG AUTO parameters Kron factor equal to 2.5 and minimum radius equal to 3.5 pixels.

    We adopted gains for each filter using the mean exposure time calculated for each mosaic as follows: F225W and F275W equal to 5688 sec; F336W equal to 2778 sec and F098M, F125W, and F160W equal to 5017 sec (see Windhorst et al. 2011). From Kalirai et al. (2009a,b) we assumed zeropoints for the filter set F225W, F275W, F336W, F098M, F125W, F160W equal to 24.06, 24.14, 24.64, 25.68, 26.25, 25.96 mag, respectively. We assumed zeropoints for the filter set F435W, F606W, F775W, and F850LP equal to 25.673, 26.486, 25.654, and 24.862 mag, respectively.

    In Table {X} we present the measured photometry for the {galaxies}. SExtractor non-detections are designated " · · · " (23 galaxies) and {galaxy} fluxes with detections fainter than the recovery limits (discussed below) are designated "—" (52 galaxies), as explained in the footnotes of Table {X}.

    The combination of the stable WFC3 UV-optical-IR PSF and high spatial resolution allows many compact or low surface brightness (SB) {galaxy} candidates to be detected and measured. These candidates may meet the morphological selection criteria in the "detection" image, but in dual-image mode SExtractor returns flux measurements for these {galaxies} which are significantly below the formal completeness limits in the "measurement" image. Their formal flux uncertainties are larger than ∼1 mag (implying a signal-to-noise ratio ~< 1). To ascertain the reliability of these faint flux measurements in the UV bandpasses, we inserted simulated galaxies into the images, and performed an object recovery test to measure the flux level where the signal-to-noise typically approaches ∼1. To derive 90% confidence limits, we inserted ∼60,000 simulated galaxy images representing a range of total magnitudes (24 mag < m < 30 mag) and half-light radii (0.8" < rhl < 2.25") into each of the UVIS mosaics, and measured the fraction of simulated galaxies which were recovered by SExtractor, using the same SExtractor configuration as discussed above. The simulated galaxies were defined with an r1/4 ("bulge") or exponential SB profile ("disk"). From these simulations, we estimated the 90% recovery limits for simulated [...] profiles with half-light radius of 1.0" equal to F225W=26.5, F275W=26.6, F336W=26.4, and F435W=26.7 mag, respectively. We interpret {galaxies} with magnitudes fainter than these recovery limits as 1-σ upper limits.
    Let's remind ourselves of one of Disney & Lang's most eye-catching lines: "If now we remove the Local population to redshift 1, virtually all the previously prominent galaxies will sink below the sky thanks to Eqn. (43)."

    Hmm.

    You'll be able to see for yourself, later, from Table {X}, just how many of the 102 galaxies examined - in considerable detail - have a redshift of >=1. Too, once I tell you what {galaxies} this (as yet unnamed) paper sought to study, you can decide if any are prominent in the "Local population" (Disney & Lang - deliberately? - do not provide a definition of what they mean by this term).

    In any case, I would guess that Disney & Lang would be rather astonished by this paper; in quite a few ways it would seem - at face value - to contradict several of the main conclusions of their own. And, as in all science, experiment (or observation) always trumps theory ...

  10. #160
    Join Date
    Mar 2004
    Location
    Ocean Shores, Wa
    Posts
    5,257
    I always get to wonder how much the real sinking of galaxies below Neried's horizon distorts our over-all perceptions. Consider the assumption that most ultra-bright uv events are more-or-less universally shrouded and highly directional. We would see very few of these locally, but the odds of finding the precise veiwing angle will increase with distance, and be doubled and redoubled by lensing effect. Neried's toy universe predicts very different fall-out patterns when complex geometries come into play.

  11. #161
    Join Date
    Mar 2004
    Posts
    13,441
    Quote Originally Posted by Jerry View Post
    I always get to wonder how much the real sinking of galaxies below Neried's horizon distorts our over-all perceptions. Consider the assumption that most ultra-bright uv events are more-or-less universally shrouded and highly directional. We would see very few of these locally, but the odds of finding the precise veiwing angle will increase with distance, and be doubled and redoubled by lensing effect. Neried's toy universe predicts very different fall-out patterns when complex geometries come into play.
    Interesting comment Jrery.

    As it's about "events", not galaxies, perhaps we could discuss it further, in a separate thread?

    Let me return to a key theme in the D&L paper, prominence; e.g. "If now we remove the Local population to redshift 1, virtually all the previously prominent galaxies will sink below the sky thanks to Eqn. (43)."

    Now, to me, what make local galaxies prominent isn't only their surface brightness in the B (e.g. ~386 to 485 nm) and V (e.g. ~490 to 584 nm) bands; it's also their size. Moving any local galaxy - that we know of - to redshift 1 will make it small. Take one that we can approximate as a disk of radius 100 kpc [1]; from at least z = 0.5 to z = 6, this galaxy will appear as a circle (if we view it face-on) of not even 25 'SDSS seeing-pixels' (in radius). Hardly 'prominent', eh?

    It's pretty obvious that D&L do mean to imply that there are, locally, galaxies we are not aware of, with effective radii > 100kpc!

    But if you take the (strongly implied) 'high SB, in the B and V bands' prominence criterion, it's easy to see why their paper is looking at the wrong thing. Consider the redshift at which the very bluest of detectable photons from a local galaxy sinks from detectability, by moving beyond the red end of the V band. What is that redshift? Why it's only 1.51. In other words, no galaxy we detect, in B and V, at redshifts > 1.5, is prominent locally (no matter how high its local B and/or V SB is)!

    Oh, and Jeryr, it's Nereid.

    [1] do you, dear reader, know any real galaxy, of any redshift, that is, in some sense, this big?
    Last edited by Nereid; 2012-Feb-11 at 02:00 PM. Reason: typos

  12. #162
    Join Date
    Mar 2004
    Posts
    13,441
    Continuing with some comments on prominence.

    Quote Originally Posted by Disney&Lang (abstract)
    Or could they be representatives of quite a different dynasty whose descendents are no longer prominent today? We explore this latter hypothesis and argue that Surface Brightness Selection E ects naturally bring into focus quite di erent dynasties from different redshifts. Thus the HST z = 7 galaxies could be examples of galaxies whose ...
    For starters, as has already been noted earlier in this thread, "the HST z = 7 galaxies", observed in the (observer) UBV frame would not only not be prominent locally (i.e. if magically transported as they are/were to z = 0), but they'd be completely invisible (or nearly so)!

    Why?

    Because what we see of these galaxies, now, at z~7, in the UBV bands, is/was emitted in the EUV, blue-ward of the Lyman limit ... and we see precisely zero such galaxies locally (thanks ngc3314).

    And even at lower redshifts, what are the objects which are prominent locally that we would expect to see, in UBV (at a redshift of, say, 1 - 3)?

    Well, they would certainly include the Markarian galaxies.

    The most prominent local galaxies (other than the MW), in the UBV bands, are, of course, the Magellanic Clouds, M31, and M33. The first two are dwarfs, and are only prominent because they are so close. And M33 isn't much bigger, or brighter. M31 and the MW? Well, they're not Markarian galaxies, no matter how much you stretch the definition of such galaxies. If we widen our definition - of locally prominent galaxies - to the Messier objects galaxies, how many of those are (northern) Markarian galaxies?

    By (deliberately?) ignoring this 'prominent in one narrow slice of the spectrum' != 'prominent in another, quite different, slice', D&L have thus rendered most of their paper relevant only to a toy universe which is dramatically different from the real one we do our astronomical observations in.

  13. #163
    Join Date
    Mar 2004
    Posts
    13,441
    As is clear from the various posts in this Q&A thread, prominence - as in, of galaxies - is a somewhat slippery concept.

    Of interest, from that thread, to a broader question (than the ones D&L asked): there are lots of prominent galaxies in clusters (among the Messier galaxies, >40% are in the Virgo cluster ... and the non-Virgo cluster galaxies include four in the Local Group), and quite a few in tight groups.

    If all the Messier galaxies were at the same distance from us - 10 Mpc, say - how would they be ranked, in terms of prominence?

    I'll have a go at answering this later (unless someone beats me to it), but my guess would be that the giant ellipticals/lenticulars - which go by the moniker ETG (early type galaxies) I believe - would all be at, or near, the top; e.g. M84, M85, M86, and M87.

  14. #164
    Join Date
    Mar 2004
    Posts
    13,441
    Well, the Messier galaxies are quite an interesting lot, when studied in terms of prominence!

    First, the source I used: SEDS.

    If you plot the listed magnitude against distance, and also against size, fit a trend line (taking logs as appropriate), then identify the extreme galaxies (i.e. those which are farthest from the respective trend lines), what do you find?

    Well, would you believe M31 is clearly the most extreme galaxy?

    It is a full 3 mags brighter than it 'should' be, for its distance (the next most extreme, in this sense, is a mere 1.9 mags brighter), and 1.2 mags brighter than it 'should' be for its area - in square arcmins - which puts it equal second (the most extreme galaxy is 1.4 mags off the trend line).

    Three other spirals are also significantly brighter by both criteria: M104 (the real outlier), M81 (still >1σ off in both), and M83 (<1σ in both).

    In the faintness stakes, the local dwarfs (M32 and M110) take the 'distance' prize (i.e. their brightnesses are the furthest from the trend line), with M110 also being a wimp in the area race (M32 is well above that trend line). Several spirals are near, or over, the 1σ mark in both distance and area relationships: M74, M98, M91.

    So which Messier galaxies are likely to stand out, if moved to greater and greater distances, ignoring any SED effects? If all you had to go on were the two relationships above, then they'd be the big, bright spirals, M31, M104, M81, and M83. Next would come some of the Virgo cluster ETGs, M49, M87, M86, and M60, and one of the Virgo spirals, M77. Spirals M33 and M51 would also be selected. Notably different is M101: ~1σ bright in the distance race, ~1σ faint in the area one.

    And which galaxies best represent the two trends - as in, they are close to both trend lines? One spiral, M96, and one Virgo cluster ETG, M59.

    But isn't it rather odd that so many of the closest galaxies are near, or at, the top of the class? Surely we don't, here in the MW, happen to be especially blessed with having the most prominent galaxy by far (M31) - out to a distance of ~60 Mly - right on our very doorstep? No, of course not. So, how to get a better handle on 'prominence'? Well, what would M31 (etc) look like if it were at a distance of 60 Mly?

    Stay tuned!

  15. #165
    Join Date
    Sep 2006
    Posts
    1,429
    Could you show us your graphs, please?

  16. #166
    Join Date
    Mar 2004
    Posts
    13,441
    Quote Originally Posted by StupendousMan View Post
    Could you show us your graphs, please?
    I'd love to, but I'm not sure I can, easily.

    However, here is the data (from SEDS):

    31 S 2.9 3.4 178x63
    32 E 2.9 8.1 8x6
    33 S 3 5.7 73x45
    49 E 60 8.4 9x7.5
    51 S 37 8.4 11x7
    58 S 60 9.7 5.5x4.5
    59 E 60 9.6 5x3.5
    60 E 60 8.9 7x6
    61 S 60 9.7 6x5.5
    63 S 37 8.6 10x6
    64 S 19 8.5 9.3x5.4
    65 S 35 9.3 8x1.5
    66 S 35 8.9 8x2.5
    74 S 35 9.4 10.2x9.5
    77 S 60 8.9 7x6
    81 S 12 6.9 21x10
    82 I 12 8.4 9x4
    83 S 15 7.6 11x10
    84 E 60 9.1 5x5
    85 E 60 9.1 7.1x5.2
    86 E 60 8.9 7.5x5.5
    87 E 60 8.7 7x7
    88 S 60 9.6 7x4
    89 E 60 9.8 4x4
    90 S 60 9.5 9.5x4.5
    91 S 60 10.2 5.4x4.4
    94 S 14.5 8.2 7x3
    95 S 38 9.7 4.4x3.3
    96 S 38 9.2 6x4
    98 S 60 10.1 9.5x3.2
    99 S 60 9.9 5.4x4.8
    100 S 60 9.3 7x6
    101 S 27 7.9 22x22
    102 E 45 9.9 5.2x2.3
    104 S 50 8 9x4
    105 E 38 9.3 2x2
    106 S 25 8.4 19x8
    108 S 45 10 8x1.5
    109 S 55 9.8 7x4
    110 E 2.9 8.5 17x10

    Columns:
    1: Messier number
    2: galaxy type, S=spiral, E=ETG (i.e. elliptical or S0/lenticular), I=irregular
    3: distance, in Mly
    4: brightness, in mags
    5: area, in arcmins

    The fitted trend lines are:
    Distance trend line: mag = 1.006806 * ln(dist) + 5.313
    Area trend line: mag = -0.76188 * ln (area) + 11.7323

    I think it should be easy to plot both graphs, from this data. If you do, and are able to post here, would you (or any other reader) please do so?

  17. #167
    Join Date
    Sep 2006
    Posts
    1,429
    Using data from the list Nereid provided in the previous post, I computed the distance modulus for each galaxy, and the absolute magnitude of each galaxy.

    mess.png

    I do not see the trend that Nereid mentions. What I do see is a selection effect: we can and do see intrinsically dim galaxies nearby, but we don't see similar galaxies at large distances (because they don't reach our detection thresholds).

    Nereid, if I've misinterpreted your post, please correct me.

  18. #168
    Join Date
    Mar 2004
    Posts
    13,441
    Quote Originally Posted by StupendousMan View Post
    Using data from the list Nereid provided in the previous post, I computed the distance modulus for each galaxy, and the absolute magnitude of each galaxy.

    mess.png

    I do not see the trend that Nereid mentions. What I do see is a selection effect: we can and do see intrinsically dim galaxies nearby, but we don't see similar galaxies at large distances (because they don't reach our detection thresholds).

    Nereid, if I've misinterpreted your post, please correct me.
    Yes, sorry; I wasn't as clear as I should have been (obviously).

    What I am (was) doing is setting the scene for something similar to what you have, in fact, just done.

    The context is (still, after all this time) the D&L paper. Specifically, I'm exploring one aspect of this statement therein: "If now we remove the Local population to redshift 1, virtually all the previously prominent galaxies will sink below the sky ...". Now D&L did not define what they meant by "the Local population", so I'm attempting to do so (note that I've already shown that their "remove ... to redshift 1" thought experiment to be seriously flawed, at least in terms of the conclusion they draw).

    The Messier galaxies appear prominent. No argument with that, right? Further, since we (now) know they are all closer than ~20 Mpc, we can treat them as "the Local population" (or part thereof), at least as an exercise.

    All I did was plot, on the y-axis, the observed mags (per SEDS), and on the x-axis distance (again, per SEDS). To get a nice-looking linear trend, I made the scale on the x-axis logarithmic*; add a trend line, and voilà! The mag-area relation is the same, with the extra step of calculating the area first.

    Two points on the mag-distance plot/graph/chart (all these words all the same, to all my readers?) will pin down that trend line: (2.9,6.4), and (60,9.4).

    By 'moving' all the Messier galaxies to 60 Mly (say), it will be possible to examine the question of what (sorts of) galaxies are (locally) prominent in an unbiased way^ ...

    I'm pretty sure this sort of thing is so familiar to you, and other professional astronomers, as to be downright boring (so, many thanks for continuing to even read my posts in this thread); however, I suspect it's poorly understood by many BAUTians. In any case, by moving only in baby steps, I hope the general reader will not lose the plot (so to speak) in my demolition discussion of the D&L paper.

    * the software I'm using enables me to do with by simply ticking a box; it can also produce a trend line - several actually, of different kinds - with the click of my mouse. Of course, I can do all this myself, successfully (modulo mistakes), but it takes me a lot longer ...

    ^ or at least allow examination of possible biases other than distance!

  19. #169
    Join Date
    Mar 2004
    Posts
    13,441
    Repeating Stupendous Man's calculations, and plotting, I get the same as he does [1]. To recap: what this calculation of absolute magnitude, of each galaxy, does is put them all at the same distance (in this case 10 pc, so the numbers are ridiculous, but the conclusions are robust).

    And what a difference removing that selection bias makes! In the top quintile (i.e. the top eight galaxies), there are now six ETGs and only two spirals ... that's half the ETGs of the entire set, in just the top quintile! The two spiral stars are M104 and M77; only 0.7 mags separates #1 (M104) from #8 (M85). The galaxy which is so dominant in the local sky - M31 - is now merely average (it's ranked #20). The difference between #1 and #20 is only ~1.6 mags.

    So now we know something about what the locally prominent galaxies are, in terms of absolute brightness.

    What about in terms of surface brightness?

    Stay tuned!

    [1] correcting for the fact that the SEDS distances are quoted in Mly, and SM took them to be Mpc.

  20. #170
    Join Date
    Mar 2004
    Posts
    13,441
    The average surface brightness (SB)* of the Messier galaxies doesn't vary all that much; 19.7 (M105) to 23.5 (M101) mags per arcsec squared. The median SB is ~21.85, with most of the ETGs above (i.e. higher/brighter average SB) this (nine, out of 12). The one irregular (M82) makes it into the top quintile. There's no trend worth reporting, with distance.

    However, all but one of the Virgo cluster spirals (M77) has a lower (fainter) average SB than any ETG, and only two of the ETGs have lower average SBs than M77 (M49 and M85).

    It's looking like ETGs are, as a class, the winners in the prominence stakes, although an occasional spiral or two may win the prize of most prominent individual galaxy.

    What of M104, the brightest Messier galaxy? It's in the top quintile, so it's certainly prominent (only M94 has a higher SB, among the spirals; a high SB dwarf isn't going to figure in questions about D&L's paper). And M31? It's quite a wimp; it ranks #32, just missing out on being in the bottom quintile!

    But, as M105 demonstrates so well, it's not a high SB per se that prevents a galaxy from becoming Sunk (ignoring the main factor, SED, for now). After all, a small, high SB galaxy will quickly become indistinguishable from a star as its redshift increases (if all we do/have is rather poor broadband photometry; i.e. we cannot reliably estimate the likelihood that a point source is a galaxy and not a star or quasar, simply on its broadband colours).

    So what about area? What are the most prominent Messier galaxies in terms of area? Well, we have to move them all so they're at the same distance, to get a fair comparison. When we do that, what do we find?

    Stay tuned!

    * i.e. taking the reported magnitudes and spreading the flux evenly over the reported area, of each galaxy. For this exercise I assumed a nice, Euclidean universe; one in which SB is independent of distance.

  21. #171
    Join Date
    Sep 2006
    Posts
    1,429
    Quote Originally Posted by Nereid View Post
    After all, a small, high SB galaxy will quickly become indistinguishable from a star as its redshift increases (if all we do/have is rather poor broadband photometry; i.e. we cannot reliably estimate the likelihood that a point source is a galaxy and not a star or quasar, simply on its broadband colours).
    ORLY?

    Please scan the figures in http://arxiv.org/abs/astro-ph/0010052 or any of a number of other papers which show how one can (usually) distinguish stars from galaxies using broad-band colors.

  22. #172
    Join Date
    Mar 2004
    Posts
    2,693
    Quote Originally Posted by StupendousMan View Post
    ORLY?

    Please scan the figures in http://arxiv.org/abs/astro-ph/0010052 or any of a number of other papers which show how one can (usually) distinguish stars from galaxies using broad-band colors.
    Well, Nereid did specify the case where the photometry is poor (ellipticals could masquerade as K stars easily enough if close to the flux detection threshold).

  23. #173
    Join Date
    Mar 2004
    Posts
    13,441
    Quote Originally Posted by StupendousMan View Post
    ORLY?

    Please scan the figures in http://arxiv.org/abs/astro-ph/0010052 or any of a number of other papers which show how one can (usually) distinguish stars from galaxies using broad-band colors.
    Yep, and, in this case, I should have not posted at all (or not without extensive re-writing). But hindsight is always 20/20, isn't it?

    In fact, I hadn't given much thought at all to this part of D&L's thesis (i.e. that small - in area - local galaxies with high SB will become indistinguishable from stars if they are moved to high z). Clearly, if I had, I'd have found something like what you posted. But then D&L's observational astronomy is conducted in a single band (by inference, I don't they state anything much about passbands, etc).

    Some interesting questions: how far out can UCDs (ultra-compact dwarf galaxies) be confidently separated - by broadband photometry alson - from stars? quasars?

    Anyway, time to move along.

    You may be wondering what I used as a model for T12. Well, here's the answer:

    Quote Originally Posted by Nereid View Post
    Tal and van Dokkum (2011): "The faint stellar halos of massive red galaxies from stacks of more than 42000 SDSS LRG images".
    T12 is a composite of LRG (luminous red galaxies) at a z of ~0.34. These galaxies are (mostly?) BCGs, and are all ETGs.

    Recall this post?
    Quote Originally Posted by Nereid View Post
    Here are some (edited) extracts from the preprint; I'll give full details in a later post.

    Our sample of {galaxies} is drawn from the HST imaging with Advanced Camera for Surveys (ACS) and WFC3, which was obtained as part of the [...] program. Near-UV and near-IR observations were acquired as part of the WFC3 [...] program [...] a 104 orbit medium-depth survey using the HST UVIS and IR cameras. A general introduction to the performance and calibration of the WFC3 is provided in Windhorst et al. (2011). The [...] program observed approximately 50 square arcminutes in the GOODS-S field with the HST WFC3 UVIS in three filters: F225W and F275W for 2 orbits, and F336W for 1 orbit, per pointing, respectively. The program observed approximately 40 square arcminutes in the same field with the WFC3 IR in three filters: F098M, F125W, and F160W, each for 2 orbits per pointing. The 5σ 50% point-source completeness limits are: F225W=26.3, F275W=26.4, F336W=26.1, F098M=27.2, F125W=27.5, and F160W=27.2 mag (see Windhorst et al. 2011). The analysis presented here was completed using mosaicked images produced for each of the UVIS and IR band tilings, and each image mosaic was drizzled to a pixel scale equal to 0.090" pixel-1.

    [...]

    The WFC3 mosaics roughly cover the northern one-third of the GOODS-S field (Giavalisco et al. 2004), and we incorporate the pre-existing ACS dataset (F435W, F606W, F775W, and F850LP) with the WFC3 observations. We produced mosaicked images of the GOODS-S ACS data, which were binned to match the pixel scale of the WFC3 UVIS/IR mosaics.
    Selection Criteria

    We require our galaxies to have: (1) been imaged in all UV and IR bands, to uniform depth; (2) a spectroscopically-confirmed redshift in the range 0.35 ~< z < ~1.5; and (3) an {galaxy} morphology.

    There are many techniques for identifying {galaxies} at intermediate redshift. [...] However, the robustness of each of these classifiers can be dramatically affected by a variety of systematics, such as the image signal-to-noise ratio (Conselice et al. 2003; Lisker 2008) and the bandpass in which the technique is applied (Taylor-Mager et al. 2007; Conselice et al. 2008). In lieu of these techniques, we select our sample by visual classification. This technique is subjective, and as such can introduce new biases, but it has been successfully applied to the identification of both low redshift (z ∼0.1; [...]) and intermediate redshift (z ~< 1.3; [...]) {galaxies}. We will demonstrate in §{Y} that the spectroscopic redshift requirement, and not the morphological selection technique, is the most significant source of bias.

    [...]

    UV imaging can provide unique insight into {something interesting}. Thus, we require our sample {galaxies} to be observed in each of the UV filter mosaics. To ensure that all galaxies were observed to a similar depth, we also require each {galaxy} in the sample to be observed in the UV and IR image mosaics for at least the mean exposure time measured for each filter as given by Windhorst et al. (2011). Since we are interested in {something interesting}, and the WFC3 UVIS channel is only sensitive to UV emission at ∼ 1500 Å for objects at redshift z ~> 0.35, we define this redshift as low-redshift cutoff of the sample. The high-redshift cutoff was selected to ensure that the visual inspection and classification of the {galaxy} — in the filter set outlined above – considers the rest-frame V-band morphology. We are sensitive to at least the UV-optical SED of every {galaxy} in our catalog.

    [...]

    We find 102 {galaxies} that satisfy these selection criteria.
    Photometry

    We measured object fluxes using SExtractor in dual-image mode (Bertin & Arnouts 1996), with the WFC3 F160W image as the detection band. For source detection, we required sources to be detected in minimally four connected pixels, each at ≥ 0.75σ above the local computed sky-background. For deblending, we adopted a contrast parameter of 10−3 with 32 sub-thresholds. Object photometry was determined with MAG AUTO parameters Kron factor equal to 2.5 and minimum radius equal to 3.5 pixels.

    We adopted gains for each filter using the mean exposure time calculated for each mosaic as follows: F225W and F275W equal to 5688 sec; F336W equal to 2778 sec and F098M, F125W, and F160W equal to 5017 sec (see Windhorst et al. 2011). From Kalirai et al. (2009a,b) we assumed zeropoints for the filter set F225W, F275W, F336W, F098M, F125W, F160W equal to 24.06, 24.14, 24.64, 25.68, 26.25, 25.96 mag, respectively. We assumed zeropoints for the filter set F435W, F606W, F775W, and F850LP equal to 25.673, 26.486, 25.654, and 24.862 mag, respectively.

    In Table {X} we present the measured photometry for the {galaxies}. SExtractor non-detections are designated " · · · " (23 galaxies) and {galaxy} fluxes with detections fainter than the recovery limits (discussed below) are designated "—" (52 galaxies), as explained in the footnotes of Table {X}.

    The combination of the stable WFC3 UV-optical-IR PSF and high spatial resolution allows many compact or low surface brightness (SB) {galaxy} candidates to be detected and measured. These candidates may meet the morphological selection criteria in the "detection" image, but in dual-image mode SExtractor returns flux measurements for these {galaxies} which are significantly below the formal completeness limits in the "measurement" image. Their formal flux uncertainties are larger than ∼1 mag (implying a signal-to-noise ratio ~< 1). To ascertain the reliability of these faint flux measurements in the UV bandpasses, we inserted simulated galaxies into the images, and performed an object recovery test to measure the flux level where the signal-to-noise typically approaches ∼1. To derive 90% confidence limits, we inserted ∼60,000 simulated galaxy images representing a range of total magnitudes (24 mag < m < 30 mag) and half-light radii (0.8" < rhl < 2.25") into each of the UVIS mosaics, and measured the fraction of simulated galaxies which were recovered by SExtractor, using the same SExtractor configuration as discussed above. The simulated galaxies were defined with an r1/4 ("bulge") or exponential SB profile ("disk"). From these simulations, we estimated the 90% recovery limits for simulated [...] profiles with half-light radius of 1.0" equal to F225W=26.5, F275W=26.6, F336W=26.4, and F435W=26.7 mag, respectively. We interpret {galaxies} with magnitudes fainter than these recovery limits as 1-σ upper limits.
    Let's remind ourselves of one of Disney & Lang's most eye-catching lines: "If now we remove the Local population to redshift 1, virtually all the previously prominent galaxies will sink below the sky thanks to Eqn. (43)."

    Hmm.

    You'll be able to see for yourself, later, from Table {X}, just how many of the 102 galaxies examined - in considerable detail - have a redshift of >=1. Too, once I tell you what {galaxies} this (as yet unnamed) paper sought to study, you can decide if any are prominent in the "Local population" (Disney & Lang - deliberately? - do not provide a definition of what they mean by this term).

    In any case, I would guess that Disney & Lang would be rather astonished by this paper; in quite a few ways it would seem - at face value - to contradict several of the main conclusions of their own. And, as in all science, experiment (or observation) always trumps theory ...
    Well, the paper in question is Rutkowski et al. "A Panchromatic Catalog of Early-Type Galaxies at Intermediate Redshift in the Hubble Space Telescope Wide Field Camera 3 Early Release Science Field".

    Yep, once again the galaxies under study are ETGs.

    But what's perhaps the most remarkable, even ironic, thing about this paper?

    Stay tuned?

  24. #174
    Join Date
    Mar 2004
    Posts
    13,441
    Quote Originally Posted by Nereid View Post

    But what's perhaps the most remarkable, even ironic, thing about this paper?

    Stay tuned?
    No one?

    Well, think about this: here is some text from the abstract of the Disney&Lang paper:

    Quote Originally Posted by Disney&Lang
    Conversely the ancestors of the Milky Way and its obvious neighbours will have completely sunk below the sky at z > 1:2,
    [...]

    This Succeeding Prominent Dynasties Hypothesis (SPDH) fits the existing observations both naturally and well, including the bizarre distributions of galaxy surface brightness found in deep fields, the angular size  (1+z)-1 law, 'downsizing' which turns out to be an 'illusion' in the sense that it is does not imply evolution, 'Infant Mortality', i.e. the discrepancy between stars born and stars seen, and fi nally the recently discovered and unexpected excess of QSOAL DLAs at high redshift. If the SPDH is true then a large proportion of galaxies remain sunk from sight, probably at all redshifts. We show that fi shing them out of the sky by their optical emissions alone will be practically impossible, even when they are nearby. More ingenious methods will be needed to detect them. It follows that disentangling galaxy evolution through studying ever higher redshift galaxies may be a forlorn hope because one will be comparing young oranges with old apples, not ancestors with their true descendants.
    Sure that's pretty general, but how reasonable is it, in light of Rutkowski et al.'s paper? And, if Disney and/or Lang were to read Rutkowski et al.'s paper, which parts of the above - or their own paper - would they wish to at least seriously re-write?

    Or, bluntly, would Rutkowski et al.'s paper come as a big surprise to Disney and Lang?

  25. #175
    Join Date
    Mar 2004
    Posts
    13,441
    Quote Originally Posted by Nereid View Post
    Or, bluntly, would Rutkowski et al.'s paper come as a big surprise to Disney and Lang?
    No one willing to hazard a guess, eh?

    Well, consider who the "et al." are: S. H. Cohen, S. Kaviraj, R. W. O'Connell, N. P. Hathi, R. A. Windhorst, R. E. Ryan Jr., R. M. Crockett, H. Yan, R. A. Kimble, J. Silk, P.J. McCarthy, A. Koekemoer, B. Balick, H. E. Bond, D. Calzetti, M. J. Disney, M. A. Dopita, J. A. Frogel, D. N. B. Hall, J. A. Holtzman, F. Paresce, A. Saha, J. T. Trauger, A. R. Walker, B. C. Whitmore, and E. T. Young.

    There's some interesting names there, eh? Joe Silk, for one; oh, and look, M. J. Disney is an author too!

    Could you have guessed that?

  26. #176
    Join Date
    Mar 2004
    Posts
    13,441
    A cool astro-ph, just two days' old: "Ultra deep sub-kpc view of nearby massive compact galaxies", by Trujillo et al.:

    Using Gemini North telescope ultra deep and high resolution (sub-kpc) K-band adaptive optics imaging of a sample of 4 nearby (z~0.15) massive (~10^{11}M_{sun}) compact (R<1.5 kpc) galaxies, we have explored the structural properties of these rare objects with an unprecedented detail. Our surface brightness profiles expand over 12 magnitudes in range, allowing us to explore the presence of any faint extended envelope on these objects down to stellar mass densities ~10^{6} M_{sun}/kpc^{2} at radial distances of ~15 kpc. We find no evidence for any extended faint tail altering the compactness of these galaxies. Our objects are elongated, resembling visually S0 galaxies and have a central stellar mass density well above the stellar mass densities of objects with similar stellar mass but normal size in the present universe. If these massive compact objects will eventually transform into normal size galaxies, the processes driving this size growth will have to migrate around 2-3x10^{10}M_{sun} stellar mass from their inner (R<1.7 kpc) region towards their outskirts. Nearby massive compact galaxies share with high-z compact massive galaxies not only their stellar mass, size and velocity dispersion but also the shape of their profiles and age of their stellar populations. This makes these singular galaxies unique laboratories to explore the early stages of the formation of massive galaxies.
    "Our surface brightness profiles expand over 12 magnitudes in range ..." - I thought Disney&Lang showed this was simply not possible; did I miss something?

  27. #177
    Join Date
    Sep 2006
    Posts
    1,429
    Quote Originally Posted by Nereid View Post
    "Our surface brightness profiles expand over 12 magnitudes in range ..." - I thought Disney&Lang showed this was simply not possible; did I miss something?
    Well, the reason that this new paper by Trujillo et al. can claim such a large range in surface brightness profiles is simply that the angular resolution of the images is high. As you'll see if you examine their figure 2, much of this 12-magnitude range comes from measurements within a radius of 1 arcsecond. Remember, surface brightness measures "magnitudes per square arcsecond". If you can define a very small patch of sky, with an area of very few square arcseconds, then you can compute a very large surface brightness.

    Disney and Lang claimed that this wasn't possible? Could you please provide a guide to the location of such a statement in their paper?

  28. #178
    Join Date
    Mar 2004
    Posts
    13,441
    Quote Originally Posted by StupendousMan View Post
    Well, the reason that this new paper by Trujillo et al. can claim such a large range in surface brightness profiles is simply that the angular resolution of the images is high. As you'll see if you examine their figure 2, much of this 12-magnitude range comes from measurements within a radius of 1 arcsecond. Remember, surface brightness measures "magnitudes per square arcsecond". If you can define a very small patch of sky, with an area of very few square arcseconds, then you can compute a very large surface brightness.

    Disney and Lang claimed that this wasn't possible? Could you please provide a guide to the location of such a statement in their paper?
    D&L's paper is based on a toy universe in which the (azimuthly averaged?) radial SB profile of all galaxies is wavelength/band independent (crudely, a galaxy which is an exponential in the V band, will be an exponential in every band, even bands to the blue of the Lyman limit), and which is populated by only two kinds of galaxies - those with an exponential SB profile (i.e. Sérsic index 1) and giants with a de Vaucouleurs profile ('Giant Ellipticals', i.e. Sérsic index 4).

    The former, per Fig. 3, cannot have an SB contrast of 12 mags; the latter, per Fig. 7, certainly can.

    Exponential galaxies are, nearly always (?), disks; their Hubble type ranges from S0 (lenticulars) to Sm (borderline irregulars); their bulges - if they have them - may have a de Vaucouleurs profile (with an effective radius much smaller than that of the disk), or an exponential profile (and are called 'pseudo-bulges?).

    The four galaxies covered in the Trujillo et al. paper have, per Table 2, Sérsic indices ranging from 2.18±0.24 to 3.55±0.39; the authors refer to their apparent morphology as 'disky'*. And we know, from the excellent Lackner and Gunn paper that I briefly mentioned earlier, that modelling real galaxies with Sérsic profiles requires a very significant proportion of them to have indices significantly different from 1 or 4.

    So, it's not so much that D&L showed that an SB contrast of 12 mags is not possible as that they did not consider anything like the full range of Sérsic profiles (local) galaxies are known to closely approximate.

    On angular resolution: D&L's paper is essentially blind to this; their logic applies equally well to a local galaxy examined at typical 'ground' resolution as to a more distant one examined at 'space' resolution. The exception they explicitly cite is "The maximum heights of the two curves [...] assume a sample for which [...], typical of all Exponential galaxies, save those hundreds of pixels across", to quote one version (from Section IV). In fact, much of Section IV is about just how truly trapped we are in "our lighted cell", in terms of angular resolution (and several other factors). Of course, this section is also (mostly) about exponential galaxies only ...

    * "A visual inspection of the nearby compact massive galaxies shown in Fig. 1 indicates that the most common morphology of our objects is disky. In fact, two objects (SDSS J103050.53+625859.8 and SDSS J120032.46+032554.1) visually resemble S0 galaxies viewed in edge-on projection. The other two galaxies (SDSS J153934.07+441752.2 and SDSS J212052.74+110713.1) have a more distorted morphology but still are compatible with being S0 galaxies with a lower inclination (see also Valentinuzzi et al. 2010)."

  29. #179
    Join Date
    Mar 2004
    Posts
    13,441
    Here's an interesting paper: Disc scalelengths out to redshift 5.8:

    Quote Originally Posted by Fathi et al.
    We compute the exponential disc scalelength for 686 disc galaxies with spectroscopic redshifts out to redshift 5.8 based on Hubble Space Telescope archival data. We compare the results with our previous measurements based on 30000 nearby galaxies from the Sloan Digital Sky Survey. Our results confirm the presence of a dominating exponential component in galaxies out to this redshift. At the highest redshifts, the disc scalelength for the brightest galaxies with absolute magnitude between -24 and -22 is up to a factor 8 smaller compared to that in the local Universe. This observed scalelength decrease is significantly greater than the value predicted by a cosmological picture in which baryonic disc scalelength scales with the virial radius of the dark matter halo.
    It didn't make the Fun Papers In Arxiv cut.

    I wonder what Disney and Lang thought of it (assuming they read it)? And did Fathi et al. include the Tolman surface brightness effect in their calculations, I wonder.

Similar Threads

  1. Arp 291 A "wind effects" galaxy
    By RickJ in forum Astrophotography
    Replies: 1
    Last Post: 2010-Nov-28, 03:23 AM
  2. observational bias
    By toothdust in forum Against the Mainstream
    Replies: 23
    Last Post: 2008-Jul-18, 10:29 PM
  3. Replies: 2
    Last Post: 2007-May-09, 06:47 PM
  4. Bias in Papers
    By RussT in forum Against the Mainstream
    Replies: 13
    Last Post: 2006-Oct-24, 08:31 AM
  5. The artful bias of the BBC
    By Glom in forum Off-Topic Babbling
    Replies: 33
    Last Post: 2006-Feb-07, 11:08 AM

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •  
here
The forum is sponsored in-part by: