Temperature of the Sun is 6000 K, but only from the Earth?
I see that the experts on the CMB represents the full confidence, supported by a complete ignorance.
Banned
Temperature of the Sun is 6000 K, but only from the Earth?
I see that the experts on the CMB represents the full confidence, supported by a complete ignorance.
Established Member
Intensity in terms of the power received from a given source at distance r does decline as 1/r^2. However, the surface brightness (the power you receive from a source per steradian occupied by the source) does not change. It stays constant with distance (excepting extinction and non-Euclidean/expansion effects at high z such as redshift). The amount of radiant power you receive from the galaxies in some solid angle of the sky, depends only on the intrinsic surface brightness of the galaxies (which are independent of distance) and the fraction of that solid angle which is covered by the galaxies. In this case, I estimated 4% of the sky is covered with galaxies. Does that make sense to you now?Originally Posted by Reality Check
Established Member
Yes but they only remove about 500 specific powerful point sources. There are about 10^5 galaxies in each pixel of WMAP. I haven't see any discussion of that contamination, but apparently it turns out that those galaxies only affect the measured temperature by about 1 micro Kelvin.
Established Member
On a second read through me too. Sorry about that. I was confusing the MEM bit with the point source estimation cleaning.I agree that the capitalization of Galactic refers to the Milky Way.
Edit: TooMany, read the section on ILC and tell me what that means to you.
Established Member
If there's an average of 250,000 galaxies per pixel, assuming a Poisson distribution and that all galaxies are randomly distributed across the sky, the standard deviation (pixel to pixel) is 0.2%.
This is orders of magnitude bigger than the signal variation that is being measured (approx. 0.005%). What's more, it's probably an underestimate, because galaxy size is not evenly distributed: the large galaxies will have the biggest effect, but they are far rarer than small galaxies. So their Poisson uncertainty is higher.
Then you have to consider if the distribution of galaxies is truly isotropic and randomly distributed. Do we have a systematically higher density of galaxies along some sightlines relative to others.
But the next stage is to evaluate the dependence of the overall uncertainty on this Poisson variation, and to do that is what the whole discussion on this thread is about ! If galaxies are completely transparent to the CMB, and also contribute almost zero to it at the the wavelengths under consideration, then it's a negligible source of error.
Just thinking out loud.
Order of Kilopi
As wiki says, "For a nearby object emitting a given amount of light, radiative flux decreases with the square of the distance to the object, but the physical area corresponding to a given solid angle (e. g. 1 square arcsecond) increases in the same fashion, resulting in the same surface brightness."
If you take a galaxy and move it twice as far away, its surface may have the same brightness, but the apparent extent of that surface if going to be reduced by half. So the total area of the surface appears four times less. IOW, the total amount of "power" you detect is still going to decrease with the square of the distance.
Everyone is entitled to his own opinion, but not his own facts.
Established Member
Yes it is. But you are missing the point that the extent of this surface brightness over the entire sky is conservatively about 4% of the solid angle of the sky. That's why the distance to individual galaxies does not matter. It's the coverage of the sky with galaxies that makes them significant. For very distance galaxies, I'm ignoring redshift and geometry. However out to about z =1 (which encompasses more than half the universe volume) these affects are not drastic.
Order of Kilopi
(bold added)
I must have missed it; what is "this surface brightness"? And what are the details of "conservatively about 4% of the solid angle of the sky"?
Galaxies, in some sense, may appear to stake out ~4% of the sky's real estate (or not, let's see the rationale); however, the dust within these galaxies lays claim to a far, far smaller slice. After all, an awful lot of galaxies have no dust at all (to speak of; most dwarfs, most ellipticals), and the distribution of the dust in those which do contain dust is patchy (let's say) at best. Think of our own galaxy; in the optical/visual, it tends to be concentrated within a few degrees of the galactic plane, but even there it's very unevenly distributed, as evidenced by our ability to clearly see nice, distant galaxies at very low galactic latitudes (think the Circinus galaxy, for example). And the slice of the sky carved out by (nearly) edge-on spirals is trivial, surely ... just as we have no trouble seeing galaxies at high galactic latitude, both north and south, so a distant observer would have no trouble seeing galaxies through the Milky Way disk (unless it were edge-on).
Yes, intense starburst regions may be 'choked with dust', but they are rare indeed ... very few local galaxies have extensive starburst regions (e.g. M82, Arp 220).
And so on.
Established Member
Nereid and ngc3314 did some estimates of surface brightness of galaxies (like ours) due directly to starlight:
So, let's see how that works out for the CMB spectrum using the equation for intensity ratio of two emitters at 200 GHz with temperatures T1 and T2 which I approximated from Planck's law as:To take Nereid's example in an earlier post, if we approximate the Milky Way as 100 billion sunlike stars, their collective cross-section is about 10^11 x (106 km)^2 = 1023 km^2 spread over (100,000 light-years)2 = 10^36 km2, so the collective spectrum would be an approximate 5500 K blackbody distribution, but with surface brightness lower by a factor 10^13 than predicted by the Planck blackbody function.
I2/I1 = (e^(10/T1) - 1)/(e^(10/T2) - 1)
Where T2 = 5500 K and T1 = 2.7 K
I2(star)/I1(CMB) = 40/0.0018 = 2.2*10^5
However, starlight intensity is diluted by a factor of 10^-13 due to their sparseness, so:
I2(galaxy)/I1(CMB) = 2.2*10^-8. Thus the direct contribution from starlight from a galaxy is indeed very small in comparison with the CMB. The Planck probe sensitivity to differences in CMB is at best about 1 part per million which is 50 times greater that the direct effect of galaxy starlight.
So yes, the direct affects of galactic starlight should be negligible!
I'd still like to understand how exactly the same dilution arises for dust emissions, but so far no one has replied about that. It bothers me because dust appears opaque in the optical spectrum and so should act more like a black body.
------------------------------------
Another possibility to consider:
The only way I can see that starlight from galaxies could seriously affect CMB measurements is if something happens to starlight on it's path to us over billions of light years that moves a very small part of the stellar energy into the CMB spectral range.
The ratio of total energy from a star in comparison to total energy from the CMB is about 2*10^13
E(star)/E(CMB) = T(star)^4/T(CMB)4 = 5500^4/2.7^4 =~ 2*10^13
so
E(galaxy)/E(CMB) = 2 (using the dilution by 10^13 calculated by ngc3314)
if a small fraction, say 10-5 (1 part in 100,000), of stellar energy is somehow moved into the CMB spectral range as it travels to us, it would very seriously contaminate the CMB spectrum (at the level of 40 micro Kelvin) wherever there is a galaxy on the sky. If there is some physical effect that moves such small amounts of stellar energy into the CMB spectrum without scattering it's possible that we could find it with the Planck data by looking for CMB variations where galaxies are denser than average. This is still somewhat difficult because the Planck resolution is only about 5'.
Please correct me if I've made any math errors here.
Established Member
I recently saw some paper that discussed transparency, but I think they concluded that the optical seeing through a galaxy wasn't all that useful. Of course the transparency really doesn't matter, only the surface brightness matters, including whatever radiation passes right through. If the radiation is adsorbed, then it has to be re-radiated (at lower T).
If a galaxy is tilted or on edge, then in the CMB spectrum its surface brightness may be proportionally higher per solid angle, if dust is more or less transparent at CMB frequencies.
So let's say that galaxies have some average surface brightness which only extends slightly farther than the optical disk. What is you estimate of sky coverage?
I made my guess just looking at the HUDF image (low resolution), I think it might be conservative, but one would need the actual data to either do a complete integration or a Monte Carlo analysis. I wonder if anyone has. Of course at high red shift we don't have a very good idea of surface brightness yet AFAIK, except for CIB measurements. Maybe when JWT is in place all estimates will get a lot better. I believe it will be sensitive much deeper into the infrared than HST.
By the way, I'd really appreciate if you could explain why the dust clouds are so "pale gray" that they are only 10^-6 times as emitting as a blackbody. (I'm not saying that they aren't, I'm just looking for an explanation.)
Established Member
To be more exact they remove about 500 detectable powerful point sources. If you cannot detect a point source then it is really hard to remove its contamination!
N.B. These are the point sources that are detectable by WMAP. Future missions like PLANCK will have higher sensitivity and will detect more point sources that will have to be removed.
Order of Kilopi
You might want to check your calculations, and also take a hard look at a field imaged by the Hubble and extensively studied by other instruments (e.g. GOODS, COSMOS). You may find that, in terms of coverage of the sky's real estate, the higher redshift objects ('galaxies') may, collectively, dominate.
Order of Kilopi
Galaxies are known to cluster; they form groups, and clusters of various 'sizes'. They also tend to be found near or on 'filaments'.
Those in rich clusters are also found, generally, to be in approximately spherical clouds of hot plasma; the estimated baryonic mass of these clouds exceeds the estimated total mass of the constituent galaxies by a factor of several. Per textbook physics, if the CMB is 'behind' such a cluster, the plasma will leave an imprint on the CMB; this is called the Sunyaev-Zel'dovich effectWP* (the SZE). An important test of our understanding of the CMB is whether a) the SZE can be detected in WMAP (and similar) data, from known rich clusters; and b) whether Planck can, blindly, find previously unknown rich clusters (via the SZE footprint).
Interestingly, both tests have been done, and the results are ... interesting. For WMAP, the observed SZE is not quite what was expected, from theoretical models, thus pointing to something new we learned about rich clusters of galaxies. For Planck, the blind search indeed turned up several previously unknown rich clusters; it also found some, um, unusual structures.
Of course, there are independent facilities studying the region of the spectrum where the CMB signal is strongest, and searches by these should also find, independently, both known and previously unknown rich clusters (via the SZE). At least one has (the South Pole Telescope).
Links and references? Just ask!
* strictly speaking there are actually two, closely related, effects
Order of Kilopi
Let's not.
This paper, by Tal and van Dokkum, shows that the LRGs in SDSS, collectively, extend far beyond "the optical disk" (ignoring, for a moment, that ellipticals don't have disks).
And this one, by Bakos and Trujillo (which I posted earlier in this thread), shows that spirals (or the stars in them) emit detectable light well beyond "the optical disk".
Using the distribution of stellar light as a proxy for dust is not a good idea.
I'm not sure I understand what you're asking; could you explain a bit more please?By the way, I'd really appreciate if you could explain why the dust clouds are so "pale gray" that they are only 10^-6 times as emitting as a blackbody. (I'm not saying that they aren't, I'm just looking for an explanation.)
Established Member
In the paper I link to below, the author argues that there could be about seven times as much gas (atomic and molecular hydrogen) in galaxy disks as we think. Currently it is assumed that the gas is "optically thin", when in reality, it being "optically thick" is not discounted by the available evidence. (When he says "optically" I assume he means in the radio frequencies.)
If you make this assumption (7 times as much gas), a lot of problems disappear. He sells it really well as a simple and elegant solution to a variety of issues in astronomy.
Also bear in mind the gas disk in our galaxy extends well beyond the traditional stellar disk, out to a radius of 30kpc, and this is the typical finding for other spiral galaxies also.
Question: If there is indeed this much gas in galaxy disks, and it is "optically thick", would this affect the WMAP observations, foreground subtraction etc and overall conclusions?
A Heavy Baryonic Galactic Disc
J. L. Davies (school of physics and astronomy, Cardiff Univ)
http://arxiv.org/abs/1204.4649
Order of Kilopi
Well, of course they do, with such an assumption. This is essentially why dark matter is hypothesized - the observed rotation implies there must be a lot more unseen mass in the galaxy. The reason why the "gas" or baryon hypothesis is not favored is because baryons are generally detectable, yet we're not detecting them. So, the thinking goes, it must be something not directly detectable -- dark matter. If someone can explain why we are not detecting 7 times the amount of baryons that we currently detect, then they may be on to something. But again, there are numerous ways to detect baryons. Add them all together, and we fall woefully short. That's a lot of baryons to hide.
Everyone is entitled to his own opinion, but not his own facts.
Established Member
Cougar wrote:
If someone can explain why we are not detecting 7 times the amount of baryons that we currently detect, then they may be on to something.
That's what he's trying to do -saying why we are not detecting the bulk of the gas.
However the question in this thread was, imagine IF the hypothesis is correct, that is, there is a lot of "self-opaque" gas in galaxies, what would that do to the CMB observations ? The author does not address these extra-galactic questions and I am wondering about it.
Established Member
He's asking why dust grains have absoprtion and scattering efficiencies as low as 1E-6 (i.e., only a millionth of the radiation hitting them is absorbed or scattered). However, this is heavily dependent on grain size, material, the wavelength of the radiation in question, and so on, and can vary by several orders of magnitude. The reason why lies in how atoms interact with light; I don't have time to write up a complete explanation, but any optics textbook worth its salt will go into some detail on the subject. Bruce Draine and his collaborators have written a number of high quality, in-depth papers on the subject, but they're extremely dense and technical.
Order of Kilopi
It always has been my understanding that the galactic dust clouds which are opaque to visible light are highly transparent to microwave emission.
Established Member
They are; even in the far-IR, where cold dust emits most of its light, the absorption efficiency is down around 1E-6 or 1E-5, and it gets still lower as the wavelengths get longer. Since the wavelengths involved are much, much longer than the grains are wide (their radii are in the sub-micron range and smaller for the most part), far-IR, microwaves, and radio tend to either diffract around dielectric grains or pass right on by.
The optical depth through a column of dust can be written as n*pi*a^2*Q*s, where n is the number density of dust particles, a is the radius of a dust grain, Q is the absorption/scattering efficiency, and s is the path length through the cloud (this of course assumes a uniform dust density, composition, and grain size, but this gets us within an order of magnitude of the answer or two). An optical depth of 1 means that, on average, a photon has interacted with at least one particle; higher optical depths imply more interactions per photon. For typical values of the Milky Way ISM, the path length needed to get an optical depth of 1 is on the order of gigaparsecs, but the dust is not nearly so concentrated as the MW ISM on those scales. Given this, it seems to me that dust opacity to microwaves is totally insignificant when measuring the CMB, though I've not read the WMAP papers to see what the science team thinks.
Order of Kilopi
Thanks.He's asking why dust grains have absoprtion and scattering efficiencies as low as 1E-6 (i.e., only a millionth of the radiation hitting them is absorbed or scattered). However, this is heavily dependent on grain size, material, the wavelength of the radiation in question, and so on, and can vary by several orders of magnitude. The reason why lies in how atoms interact with light; I don't have time to write up a complete explanation, but any optics textbook worth its salt will go into some detail on the subject.
Yes, I can see that that's what TooMany may have been referring to (though no one can be sure until he confirms). The thought did cross my mind that this was what he was asking about ... may I ask how you worked it out? I'm quite curious, because I've been here before, many times; i.e. I read a statement or question that I can't quite figure out, and along comes someone else, posting a response that is later confirmed to be spot on (I have had particular difficulty in this regard with TooMany's posts, sadly).
I have read quite a few of those, and yes, they do seem in-depth and very high qualityBruce Draine and his collaborators have written a number of high quality, in-depth papers on the subject, but they're extremely dense and technical.but also extremely dense and technical
Of possible pertinence to some of TooMany's questions may be one (I forget exactly which one) which shows that the estimated energy absorbed, in a spiral galaxy (the MW?) in the UV-optical-NIR part of the spectrum (and emitted by stars, mostly) is the same (to within the estimated errors) as that emitted (re-radiated) in the rest of the IR (and sub-mm) range (by dust). And why is this - possibly - directly relevant? I'd like to wait until TooMany returns, to see if he can work it out for himself (I'm a great believer in discovery learning). Of course, if anyone else would like to post an explanation ...
Order of Kilopi
This is a nice paper!
The author tackles a quite large number of questions/problems/sets of observations ('issues'); not surprisingly, as this is just a single paper (albeit one > 30 pages long), not many of the issues are covered in depth. That - necessarily - means that a reader with a critical-thinking hat on would have many questions about many of the issues that cannot be answered (because of the lack of depth). A general one might be this: Davies takes a close look at several areas, and examines/questions/challenges the consistency generally accepted in the literature; however, in doing so, he accepts other areas of consistency. There's nothing wrong with that (of course!), but a closer analysis of some of these may show 'consistency links' that give rise to various problems of their own (this is a muddled way of saying that if you pull on a thread in a densely woven fabric, you never know what you might end up unraveling). Both Cougar and Tensor (and maybe others) - both in this thread and the other where kzb introduced the paper - have made this point already (in one way or another).
Some specifics, in no particular order, and by no means comprehensive (and not necessarily well-thought through either!):
* "Galaxies were selected such that they have angular diameters (r25) greater than 1 arc min", and "The only remaining problem is how does the size given in LEDA (r25) equate to the size R0 used in our model (Note: R0 is the physical size of the disc), which we are assuming is the same as the cut-off observed at 21cm (see below)." How does Davies deal with inclination?
* a (potentially) good test of the HI optically-thick/thin issue is to study variation due to inclination (as has been done, for decades, with studies of dust in spirals); why did Davies not do this?
* "Finally, by using t > 2 we hopefully restrict our analysis to galaxies dominated by their disc." One of the interesting results to come from Galaxy Zoo is that this is not a safe assumption, and that it is possible to reliably study variation due to bulge/disk ratio (Davies probably didn't know about Lackner and Gunn (2012) when he was writing his paper, if he had, I'm sure he'd have at least referenced it)
* "This discrepancy between observed and predicted baryons is typically accounted for by invoking a huge reservoir of baryons in the warm inter-galactic medium (see the introduction of Takei et al. 2011 for a recent discussion of the issue). An alternative is that they actually reside in the disc of galaxies." Davies, in this paper, does not cover any significant aspect of "the issue", so we have no way to tell if putting all the 'cluster dark mass' into massive gas disks in the cluster galaxies has issues of its own! Most readers could likely come up with a pretty lengthy list of 'issue candidates' ...
* Section 5.5 ("Constraints imposed by quasar absorption lines") is really short, and (for me) raises at least twice as many questions as those Davies seeks to address!
* If Section 5.5 leaves you unsatisfied, Section 5.6 ("Dwarf galaxies") is even worse. One particular beef: the only three papers cited are dated 1997 and 1998; an awful lot of work has been done on dwarf galaxies since then, so why did Davies not cite any of it (Section 5.5 is even worse in this regard)?
* are there not studies of (spiral) galaxy rotation curves, out to well beyond R0? Using tracers such as distant PNe, globular clusters, satellites, and halo stars (for the MW)? If so, why did Davies not mention any of these?
Established Member
Neried:
I too noticed that he references a lot of old papers. But in some ways that is good, because you are taking a broader view of things. I've noticed in chemistry for example, that a lot of old knowledge gets lost.
are there not studies of (spiral) galaxy rotation curves, out to well beyond R0? Using tracers such as distant PNe, globular clusters, satellites, and halo stars (for the MW)? If so, why did Davies not mention any of these?
I think there are such studies, I think I have a paper on one of them. But I don't see it necessarily affects the argument, because it can't be agreed what is the distribution of the theorised non-baryonic DM halo ?
a (potentially) good test of the HI optically-thick/thin issue is to study variation due to inclination (as has been done, for decades, with studies of dust in spirals); why did Davies not do this?
Maybe the data do not exist? The paper has no new observations, it is all based on existing published data. Would he get the research grant to investigate this ?
Established Member
"10^-6 as emitting as a blackbody" immediately meant emissivity to me. I'm a grad student in astronomy, and the project I'm currently finishing up depends heavily on dust opacities and emissivities, so this was all fresh in my mind. And for a greybody, of course, emissivity and absorption efficiency are the same thing.
Order of Kilopi
For theory papers, yes, if there's been no significant development, old may be good.
For almost every aspect of observational astronomy, it's a bad thing, sometimes a really, really bad thing.
Davies cites THINGS, a recent HI observational survey, but not SDSS (he uses LEDA to pick his galaxies, not one of the value-added SDSS catalogues, for example). One advantage of more recent observational results - especially surveys - is much greater attention to systematics. Papers based on analysis of recent observations are, IMHO, better than those based on older ones (generally; there are obvious exceptions). Did Davies simply not know of the more recent work (hard to believe)?
Um, no.are there not studies of (spiral) galaxy rotation curves, out to well beyond R0? Using tracers such as distant PNe, globular clusters, satellites, and halo stars (for the MW)? If so, why did Davies not mention any of these?
I think there are such studies, I think I have a paper on one of them. But I don't see it necessarily affects the argument, because it can't be agreed what is the distribution of the theorised non-baryonic DM halo ?
If, per Davies, spiral galaxies have a massive gas disk that has a 'sharp' edge, then the observed rotation curve - beyond that edge - will reflect this; if there's a massive halo, extending far beyond R0, then then the observed rotation curve - beyond that edge - will reflect that. Either way, it (potentially) offers a good, quick way to test Davies' main hypothesis. Why he chose not to apply that test is ... odd.
No, inclination data is available in LEDA (and HI data from THINGS, etc).a (potentially) good test of the HI optically-thick/thin issue is to study variation due to inclination (as has been done, for decades, with studies of dust in spirals); why did Davies not do this?
Maybe the data do not exist? The paper has no new observations, it is all based on existing published data. Would he get the research grant to investigate this ?
After all, the astronomers (like Bill Keel/NGC3314) researching dust in spirals, by studying extinction (etc) at various wavelengths as inclination varies, have been using that data for decades.
Of course, it's entirely possible Davies 'corrected' for inclination (but simply didn't mention it); however, it's odd that he didn't, given its potential to test his main hypothesis.
Established Member
Nereid wrote:
Either way, it (potentially) offers a good, quick way to test Davies' main hypothesis. Why he chose not to apply that test is ... odd.
But I still think it could not be conclusive. There are papers supporting spherical, triaxial, doughnut rings, oblate, and probably other distributions, of non-baryonic DM. Since each of these possibilities was thought likely, at least to the author who came up with it, I don't see it could be a definitive test. I am still remembering the recent papers on the disk mass budget a few weeks back
Order of Kilopi
No, not really.Nereid wrote:
Either way, it (potentially) offers a good, quick way to test Davies' main hypothesis. Why he chose not to apply that test is ... odd.
But I still think it could not be conclusive. There are papers supporting spherical, triaxial, doughnut rings, oblate, and probably other distributions, of non-baryonic DM. Since each of these possibilities was thought likely, at least to the author who came up with it, I don't see it could be a definitive test. I am still remembering the recent papers on the disk mass budget a few weeks back
The Davies model is pretty black-and-white; spiral galaxies have a particular kind of massive disk ("Mestel disc", with a sharp outer edge, and lots of hydrogen). The rotation curve beyond that edge, of such disks, will be highly constrained (unless Davies wants to add lots of mass in a halo of some kind, that extends far beyond that edge), and if there are published observations reporting estimates of spiral galaxy rotation curves beyond that edge, Davies should have found them and at least cited them.
The question of whether a CDM halo - with any kind of distribution of the CDM - is, or is not, consistent with those 'beyond the edge' rotation curve observations is an entirely separate question (the hypothesis being tested would relate to Davies' proposal, not any CDM one).
Established Member
Shouldn't the collisionless nature of CDM together with virialization of CDM particles through gravitational interaction force a certain distribution? Baryonic matter could be discounted for a first order approximation.
I don't understand why so many different distributions are proposed.
Established Member
These galaxies (and entities) so far away (more than 10 billion light years), can only be seen in the HUDF?
Or you can see galaxies (and entities) so far away in other directions of the celestial vault?
If the universe expands isotropically is supposed to be so .. isnīt it?
Order of Kilopi
What is HUDF? Please don't just assume that all of us know every obscure acronym you might have encountered somewhere.
What "entities" other than galaxies are you referring to?
Posting Permissions
Reply With Quote