Tension in the Hubble constant
A few months ago I blogged about the apparent “tension” between different measurements of the Hubble constant. Here is an alternative view of the situation, with some recent updates. The plot has thickened a bit, but it’s still unclear to me whether there’s really a significant discrepancy.
Anyway, here’s a totally unscientific poll on the issue! Do feel free to register your vote.
There has been some hand-wringing of late about the tension between the value of the expansion rate of the universe – the famous Hubble constant, H, measured directly from observed redshifts and distances, and that obtained by multi-parameter fits to the cosmic microwave background. Direct determinations consistently give values in the low to mid-70s, like Riess et al. (2016): H = 73.24 ± 1.74 km/s/Mpc while the latest CMB fit from Planck gives H = 67.8 ± 0.9 km/s/Mpc. These are formally discrepant at a modest level: enough to be annoying, but not enough to be conclusive.
The widespread presumption is that there is a subtle systematic error somewhere. Who is to blame depends on what you work on. People who work on the CMB and appreciate its phenomenal sensitivity to cosmic geometry generally presume the problem is with galaxy measurements. To people who work on local galaxies, the CMB value is…
View original post 1,029 more words
February 28, 2017 at 11:05 am
That post is full of such nonsense! I’m afraid it is motivated reasoning from someone who doesn’t like the CMB because it conclusively rules out his favourite theory (MOND with no dark matter).
For instance, he says “The CMB data now allow only a narrow trench. I worry that it may wink out entirely. Were that to happen, it would falsify our current model of cosmology.”
The “narrow trench” referred to here is simply a consequence of the fact that the extremely well-measured location of the CMB acoustic peaks translates to an extremely well-measured value of the combination of parameters Omega_m*h^3. This looks like a “trench” in a plot of Omega_m vs h. So what? That’s just reflecting a natural degeneracy in parameter space and has zero relevance to falsifying the current model of cosmology.
February 28, 2017 at 1:47 pm
Apart from that remark, though, I think his take that the tension is probably not a show-stopper is one of the few sensible ones; others are questioning the entire paradigm. Of course, the truth is probably somewhere in-between, but we don’t know where.
On a related note, is there any good criticism of MOND? It seems that all I have come across basically cut down a straw man and/or jump through hoops to get a back-of-the-envelope number from the standard model which roughly agrees with some number in MOND, ignoring stuff like Renzo’s rule and the huge amount of MOND phenomenology, which, as far as I can tell, is actually true, whatever one thinks of the explanation for it.
February 28, 2017 at 1:53 pm
My own view, for what it’s worth, is that galaxies are inherently unreliable….
February 28, 2017 at 2:14 pm
I agree the H0 tension is probably not a show-stopper. I disagree with the speculation that the tension can be resolved by allowing non-zero curvature to shift the Planck value of H0, among various other silly things in that blog post. I tried to explain some of the problems in a comment on that page.
It’s possible there are some strange unnoticed systematics still affecting Planck CMB data, which happen to lead to a low value of H0 that is compatible with WMAP, ACT, lensing, BAO and SNe data. I think it is more likely that there are systematics affecting the distance ladder measurements. Time will tell.
As for good criticisms of MOND-without-dark-matter, the clincher as far as I am concerned is that the relative heights of odd and even acoustic peaks in the CMB necessarily require that there must be a source for the gravitational potential that is not coincident with the position of the baryons. Unless the theory of gravity becomes non-local (which MOND is not), this means some form of dark matter must necessarily exist. One could similarly point to the subdominant BAO wiggles in the matter power spectrum, as done by Dodelson.
I also agree that galaxies are inherently unreliable – another way of saying we understand linear perturbation theory better than all the stuff going on in galaxies.
February 28, 2017 at 3:24 pm
The thing is, many, perhaps even most or all, MOND supporters agree that the CMB is evidence for CDM. Their criticism is that even granted that this CDM exists, it is still difficult to explain MOND phenomenology on the scale of galaxies.
Even if MOND is completely wrong, the fact remains that its one free parameter can explain and predict stuff which the standard model can’t. Whatever the correct model is, it has to explain where this comes from.
There have been some claims that when baryons are properly included in simulations, feedback is there, and so on, then flat rotation curves occur “naturally” in the standard model. It takes millions of hours of CPU time, so this is difficult to test. The simulations are also not based on fundamental equations, so it is unclear what this means even if it is true. However, sometimes the claims are overblown, not nearly matching the observations, but just roughly near the ballpark.
February 28, 2017 at 3:52 pm
My experience of arguing with Stacy McGaugh is that he will admit the CMB is evidence for CDM one minute and then turn around and argue that dark matter doesn’t exist the next minute. If you point out his statements about the CMB are wrong, he’ll say it’s so boring that you’re telling him stuff he knows already … and then he’ll make the same incorrect statements again a few weeks later.
Disclaimer: other MOND supporters are available.
March 1, 2017 at 9:48 am
Of course, whether or not something is true is independent of the personalities of the advocates. Thus, I still think that it is worth spending more time on “MOND phenomenology”.
March 2, 2017 at 1:59 am
Sesh, Peter all : Mike Turner (a strong dark matter proponent) and who coined the name “dark energy” has argued that MOND phenomenology is a consequence of CDM. https://arxiv.org/abs/astro-ph/0107284. Do you people agree with this claim? I asked this question to N of NFW a few years ago and he disagrees. Would be interesting to hear your take on this
March 2, 2017 at 7:17 am
There are several overblown claims like this. At best they get an order of magnitude correct, ignoring detailed phenomena like Renzo’s rule. MOND people take the trouble to write a detailed critique (look at some of the citations), and this is ignored. Not good.
March 2, 2017 at 10:55 am
Philip. I am not supporting this result or claim and I agree this paper is an order of magnitude calculation. However what is interesting from a sociological point of view is that one of the leading proponent of the dark matter camp admits that the empirical observations which lead Milgrom et al to propose MOND are *correct* and the only thing they differ about is interpretation.
Whereas from most other dark matter proponents, I never even receive a straight answer as to what they think of their empirical fits to the data are correct or not.
March 2, 2017 at 12:10 pm
The main (but by far not the only) motivation for MOND are flat rotation curves of spiral galaxies. This is also one (but not the only) motivation for dark matter. I am not aware of anyone who disputes these observational data, or others such as the Tully-Fisher relation (and its lack of intrinsic scatter) which support MOND. On the other hand, I have never heard a cosmologist say “by the way, the Sun will rise tomorrow”—not because they are in denial, but because it is so obvious that everyone knows it and no-one debates it.
MOND phenomenology does not “naturally fall out of LambdaCDM” for any sensible definition of “naturally”. Not even close. MOND phenomenology is really hard to explain, even with millions of hours of CPU time. Most people don’t bother.
Cosmological simulations at large scales are so good that even experienced people can’t easily tell the simulation from reality. This is not true of galaxies. More work needs to be done here. Even in those cases where simulations produce galaxies which (usually over a limited range) some aspect of MOND phenomenology is reproduced, they don’t otherwise look like normal galaxies.
March 2, 2017 at 4:06 pm
Philip, I sort of disagree with your statement that “I am not aware of anyone who disputes these observational data”.
I once heard a seminar by Dan Hooper who said that T-F is a complete fluke. I once asked Rocky Kolb about T-F relation His reply was a facetious one ” FIshy-Tully relation” which sort of means he doesn’t care about it.
Most others use the magic word “feedback” as a panacea for all these problems.
That said Stacy’s latest paper on MADR has made multiple CDM simulators trying to explain the results through their simulations.
February 28, 2017 at 3:34 pm
Due to some blog-software blog, I can’t comment at Stacy’s blog.
A different, but related, question: Is there interesting tension between WPAP and Planck?
February 28, 2017 at 3:53 pm
In terms of the H0 value, no, none whatsoever.
March 1, 2017 at 9:51 am
Despite the tension, we shouldn’t forget how far we have come. Not that long ago, the Hubble constant was uncertain by a factor of two, with the debate between Sandage, Tammann etc on one side and de Vaucouleurs, van den Bergh etc on the other. Within the lensing community, this was mirrored in the debate over the time delay in 0957+561 (which the Hamburg group got right and Bill Press got wrong). During a debate on this at the 1993 lens conference in Liege, Paul Schechter cried out from the audience “What’s the problem? They agree at three sigma!” 🙂
March 2, 2017 at 5:35 am
Just a reminder that there are other non-MOND based modified gravity solutions to DM.
for example https://arxiv.org/abs/1401.4819 https://arxiv.org/abs/1104.0160 https://arxiv.org/abs/1308.5410
March 2, 2017 at 10:58 am
Peter : If galaxies are so complicated , how come they obey simple empirical scaling laws (Renzo’s rule, Tully-Fisher relation, fundamental plane, constant asymptotic acceleration over a wide range of scales, mass acceleration discrepancy relation)? They shouldn’t obey these simple deterministic laws.
March 2, 2017 at 11:16 am
They’re not exact deterministic relations, they are correlations. The fundamental plane, for example is neither fundamental nor a plane.
March 2, 2017 at 4:02 pm
Stacy has shown that the scatter in all these relations is very tiny. How about an explanation of Figure 3 of arXiv:1401.1146?
March 2, 2017 at 4:23 pm
It’s not “very tiny”!
Especially not when you remember that most of the plots are logarithmic.
Here is Figure 3 of that paper:
If you think that’s “tiny scatter” you really need to get out more.
March 2, 2017 at 5:00 pm
I meant tiny scatter in Baryonic T-F relation and also in the recently proposed mass acceleration discrepancy relation.
regarding figure 3 of arXiv:1401.1146, the key plot is the right panel for DM dominated systems, where the asymptotic acceleration shows a value of about 1 km^2/sec^{-2}/pc.
I find this pretty intriguing and have asked a few simulators about this, but have not get a good explanation for this.
But maybe you are right and this should not be taken seriously and will go away with better data.
March 2, 2017 at 5:02 pm
More specifically do you agree with the spirit of the 2001 paper by Turner and Kaplinghat which argues that all these (quasi)-deterministic relations can be explained within CDM and no need to modify gravity?
March 2, 2017 at 5:10 pm
I can safely say that I don’t know..this needs to be answered with simulations that include realistic star formation and hydrodynamics. I don’t think we’re close to that yet, despite what the simulators say. A huge amount of important physics happens way below the resolution scale of even the best simulations.
So I’m an agnostic. I’m not a fervent CDM supporter because of small-scale clumpiness issues which I take reasonably seriously. I’m actually quite interested in “fuzzy” DM theories, etc. I can perfectly well accept the idea that GR might not be the full story, but the evidence isn’t strong enough for me to accept that it absolutely needs modifying.
March 2, 2017 at 5:43 pm
Hi, Peter thanks for explaining your perspective. I agree with what you say regarding simulations. However as a physicist, its not obvious to me how by including realistic star formation, hydrodynamic feedback etc (which are all dissipative non-linear processes) you will magically get simple deterministic scaling laws for a diverse range of galaxes with just one free parameter , which is what Milgrom first found in 1983 and as Dave Merritt mentioned on this blog ( a few years back) all predictions made by Milgrom have turned out to be true.
March 2, 2017 at 5:47 pm
They’re not “simple deterministic scaling laws”, they’re empirical correlations.
And of course it’s “not obvious” – that’s why we need simulations. What you seem to be assuming, however, is that it is obvious that it can’t work and that I think is unjustified.
March 6, 2017 at 2:04 pm
By the way, here’s a paper from last year which says that it all works fine in LCDM…
https://arxiv.org/abs/1610.07663
March 6, 2017 at 4:03 pm
At least one paper citing it disagrees with it.
In general, I find it sad that the rhetoric on both sides is too sharp. It distracts from what should be an interesting issue. A common complaint is that not enough people work on MOND, so it is difficult to compare the state of the art of the two fields since they are so different. However, I am sure that some people don’t want to work on it not because it is not mainstream (there are many people who work on non-mainstream stuff) but rather because the rhetoric on both sides is of such low quality. One doesn’t want to fight invalid criticism from the non-MOND camp, nor be associated with similar but opposite stuff from the MOND camp.
March 6, 2017 at 5:04 pm
See my comments on March 2.
“That said Stacy’s latest paper on MADR has made multiple CDM simulators trying to explain the results through their simulations.”
The above paper is one of those.
So its good that people are now trying to explain this using conventional Lambda CDM. So now at least everyone agrees the empirical correlations are correct.
March 6, 2017 at 5:21 am
Peter and others: Check out this insightful article by Dave Merritt
in a philosophy journal
http://www.sciencedirect.com/science/article/pii/S1355219816301563
Table 1 is really disturbing in that almost no Cosmology textbook/monograph mentions about the critical acceleration scale
or Mass acceleration discrepancy relation. Peter’s book is not in this table (although I think this is restriction o books after 2005)
March 6, 2017 at 8:01 am
A long article which, at first glance, doesn’t get everything right. I should write a rebuttal, here and/or as a proper paper.
March 6, 2017 at 10:29 am
I see he’s a devotee of Karl Popper.
March 9, 2017 at 11:09 am
I’ll read it in more detail, unless at some point I decide that it is not worth doing so. At first glance, he seems to confuse improving our knowledge of the universe with some sort of paradigm shift. For example, he says that the various problems with LCDM are not seen as falsifying (though he implies that they should be) but points to the acceleration of the universe as an example where an anomaly was seen as falsifying the paradigm. I noticed this after about 20 seconds, and already so much is wrong.
First, there is a difference between measuring the parameters precisely enough that one can conclude that the universe is expanding and overturning some standard model. While it is true that many assumed cosmological parameters such that the universe was not expanding, no honest and competent person believed that acceleration was ruled out—the data just weren’t good enough to say either way. What happened? It was found that 1920s cosmology fits the data—and still fits it today, despite the huge advances in observational cosmology in the past 20 years. So, this is confirmation of a theory about a century old, not some sort of paradigm shift. If I measure something more precisely, then that is not a shift. It would be a falsification only if new measurements indicate a model which was previously ruled out, and one also found a problem with the previous measurements.
Second, LCDM, referring here to galaxy formation and so on, is not a theory in the sense that classical cosmology is a theory. That doesn’t mean that one can always move the goalposts so that it can never be falsified, but he is confusing debate within the field about details with questioning the field itself, much as young-Earth creationists latch on to some technical debate in biology (which they don’t even understand) and claim that evolution has been disproved. The jury is still out on whether various LCDM problems will be solved when simulations become more realistic or whether something really new is required. Until this is solved, and certainly not before some alternative theory explains the data just as well, one should not conclude that cosmologists are self-appointed hidebound defenders of orthodoxy.
Note also that the current standard model, including the accelerating universe, has been the standard model only for about 20 years or so. Not very long. The fact that it has not been ruled out despite all the new observations speaks for it. And it was once revolutionary in the sense that some people were sceptical and believed in the Einstein-de Sitter universe without good reason. (As I argued above, though, finding that a parameter introduced 100 years ago is non-zero is not a paradigm shift.) So, adherents of the expanding universe are hardly some fuddy duddy orthodoxy immune to observational data—quite the opposite, since the current standard model is motivated primarily by observational data.
March 9, 2017 at 11:15 am
“If Kuhn’s characterization of textbooks is correct, they are an excellent source of information about the current state of a scientific theory.”
It’s OK to be a follower of Popper. It is not OK to be a follower of Kuhn. Kuhn fundamentally misunderstood how science works, confusing non-scientific events (e.g. burning people at the stake if you don’t like their views) with genuine scientific debate. He even claims that the events due to the discovery of X-rays and radioactivity were primarily sociologically driven and not data-driven!
(Kuhn also contains the seeds of his own destruction. If his work is not scientific, then we don’t have to worry about it at all. If it is, then, if it is correct, then it is just a current paradigm to be overthrown by a new paradigm later on, and hence cannot be fundamentally correct. 🙂 )
March 9, 2017 at 11:16 am
And, of course, if it is scientific but incorrect, then, again, we don’t have to worry about it.
March 9, 2017 at 11:51 am
“Graduate-level texts on cosmology and/or galaxy formation published after 2004. Q1=“Is the failure of laboratory experiments to detect dark matter discussed?” Q2=“Is the existence of a universal acceleration scale a0 discussed?” Q3=“Is the mass discrepancy–acceleration relation discussed?”.”
Q1: If particle dark matter is mentioned at all, then of course detection isn’t mentioned, because it hasn’t happened. Does lack of detection need to be explicitly mentioned? In any case, this is a strange criticism. How many years between hypothesis and detection for neutrinos? And we knew what produced them, how many, with what energy. We don’t even know that it is “particle dark matter”. It could consist of macroscopic objects (see recent work by Starkman et al.); as Carr et al. recently pointed out, very massive primordial black holes are still not ruled out. This is a good case of “absence of evidence is not evidence of absence”. And of all the criticism of dark matter by MOND enthusiasts, this is the silliest. Both observe flat rotation curves, one group postulates dark matter, one modified gravity. Some MOND enthusiasts think that the lack of detection of dark matter in the lab is a big problem. But their modified gravity has not been detected in the lab either. So where is the beef?
Q2 and Q3 are really too technical to mention in a book which covers essentially all of cosmology. To take a more mainstream example, how many aspects (if any) of gravitational lensing are mentioned in general cosmology books? Also, there is not much point in mentioning them in a textbook as long as the situation is not clear.
March 9, 2017 at 7:06 pm
Philip or anyone else: Do you know how many books mention black hole M-sigma relation (which also is another empirical observed correlation, which we don’t understand). I haven’t done a similar survey, but am pretty sure it is more than 0. Similarly I think at least since no dispute observations 2, 3 its strange that not a single textbook mentions them.
March 9, 2017 at 11:54 am
“The third possibility, a ‘critical Universe’, would tend asymptotically to a state of zero expansion after infinite time and after reaching infinite extent.”
This refers to the Einstein-de Sitter universe. It is infinite at all times; it does not “reach” an infinite extent after any (finite or infinite) time.
March 9, 2017 at 11:56 am
“The standard model of cosmology deals with this anomaly in a different way: via an auxiliary hypothesis. It is postulated that the universe is filled with a fluid, called ‘dark energy’, that has whatever properties are needed to convert the predicted cosmological deceleration into an acceleration, and in just such a manner as to reproduce the observed dependence of galaxy redshift on distance.”
What a caricature! Give me a break!
March 9, 2017 at 11:58 am
” No attempt has been made to account for this anomaly within the framework of the standard cosmological model. Instead, the mass discrepancy–acceleration relation has been dealt with via the third of Popper’s conventionalist stratagems: It has been ignored.”
Another caricature. No attempt? Really? Of course, perhaps no successful attempt, but that is something different, and doesn’t mean that the issue is ignored.
Ignored? Really?
March 10, 2017 at 10:09 am
The Hubble constant is a thumb rule. It does not explain the process concerned. It also does not explain why the moon is accelerating away from the earth or why the earth is accelerating away from the sun or why Andromeda is accelerating towards Milky Way. I think that Hubble’s formula is incomplete.
March 10, 2017 at 10:34 am
Yes, it must be incomplete. It doesn’t explain why an Australian impersonated Justin Bieber, doesn’t explain why my beard is turning grey, and doesn’t explain why some people don’t like some kinds of cheese.
March 10, 2017 at 10:36 am
Two of the three things you mention are not true anyway.
March 10, 2017 at 11:21 am
Of course, the reply is to martenvandijk and not to me. 🙂
The Andromeda Galaxy is accelerating towards the Milky Way. The Moon is moving away from the Earth, but not accelerating away. Ditto for the Earth moving away from the Sun.
March 10, 2017 at 11:43 am
Telescoper says two out of three things, not two out of four, the first of which being the process of galaxies accelerating away from us. But the thing about the taste of some kinds of cheese is absolutely true 🙂 Anyway, the accelerating away of the moon is tue (now about 4 m per year and increasing); the earth now about 15 cm per year and increasing (recent Japanese measurements).
March 10, 2017 at 11:58 am
The moon is not accelerating away from the Earth, it is accelerating towards the Earth, which is why it is in a near circular orbit. The radius of this orbit is increasing slowly with time due to tidal dissipation but that does not mean that the Moon is accelerating away from the Earth.
March 10, 2017 at 12:14 pm
Thank you for clearing up my my misunderstanding. So is it the increase of the radius that might be responsible in the far future for the earth becoming unstable? (this is what I read some time ago)
March 10, 2017 at 1:48 pm
The Moon is receding from the Earth due to the conservation of angular momentum; tides slow down the rotation of the Earth (they have already slowed down the Moon to the point that it always keeps the same face to the Earth) and as a result the Moon recedes. There is some indication that the Moon might be responsible for preventing the tilt of the Earth’s axis from becoming too great; perhaps this is what you mean by stability. If so then, yes, this influence might be less in the future. The timescale is rather long, though. The Sun is also getting hotter; I’m not sure which will cause problems first.
March 10, 2017 at 2:24 pm
Estimated timescale is 50 billion years, far longer than the lifetime of the Sun. I’m not worried. I’ll be retired by then.
May 13, 2017 at 9:36 am
Colloquium by Adam Riess at CfA on recent measurements of the Hubble constant: https://www.youtube.com/watch?v=eSPCy-IJaPg