# Marginal evidence for cosmic acceleration from Type Ia SNe

1. Jun 4, 2015

### Garth

A paper was published on today's ArXiv that questions the empirical basis of the acceleration of the expansion of the universe: Marginal evidence for cosmic acceleration from Type Ia supernovae . The authors are: Jeppe Trst Nielsen1, Alberto Guanti1, and Subir Sarkar1;2 (1Niels Bohr International Academy and Discovery Center, Niels Bohr Institute, University of Copenhagen, Denmark and 2Rudolf Peierls Centre for Theoretical Physics, Oxford, UK).

On the thread Standard candle - in question - affects distance estimates , which has been closed for Moderation until a Moderator "who knows this stuff look into this", (BTW Has there been a decision?) I made the suggestion that an alternative linearly expanding model might also fit the data (https://www.physicsforums.com/threa...fects-distance-estimates.808071/#post-5079683).

Even treating the SNe 1a as standard candles, when a larger data base, in which due allowance is made for the varying shape of the light curve and extinction by dust, then the standard model may be brought into question. In the present paper we read:
(emphasis mine)

Furthemore
Garth

Last edited: Jun 11, 2015
2. Jun 4, 2015

### wabbit

Interesting. As I read it, their analysis clearly favours an accelerating model (best fit with a pretty decent likelihood ratio over alternatives) (fig 2 and table I), but due to larger error ellipses than in the standard analysis, they conclude that the "cleaned" data is not sufficient to conclusively reject the possibility of a linear expansion.

Last edited: Jun 4, 2015
3. Jun 4, 2015

### ruarimac

All they seem to have done is relax the errors a bit and given they aren't doing anything significantly new like a Baysian approach their analysis isn't particular interesting. It was always the case that SN-1a was only really useful when combined with the CMB and the BAO data which have different probability distributions of the parameters. The difference between LambdaCDM and non-acclerating is small. This is what rules out non-acceleration.

This doesn't call the standard model into question, what their claiming is that in their analysis showed constant expansion isn't ruled out by one particular test. For this to have impact you would need to make the same claim in the face of the other 2 major tests of cosmology. To call the standard model into question you would have to show LambdaCDM was ruled out or in tension.

4. Jun 4, 2015

### Garth

But people have made just such a claim, as I said in https://www.physicsforums.com/threa...fects-distance-estimates.808071/#post-5079683
1st other major test:
2nd other major test:
Such as the tension between the standard model and a possible age problem?

Garth

5. Jun 4, 2015

### ruarimac

But you're talking about this paper. Where is the joint statistical analysis considering the other data sets as well? I see your hypothesis but where has that been tested in statistical rigour?

The "age problem" is not a cosmological test. It's a calculation involving both cosmology and galaxy formation, for that reason people don't use results like that for cosmology. It's convoluted and impossible to tell where the issue lies.

Last edited: Jun 4, 2015
6. Jun 4, 2015

### Garth

I was originally talking about the OP paper, however in my comment "But people have made just such a claim" I was responding to your post #3, "For this to have impact you would need to make the same claim in the face of the other 2 major tests of cosmology."

As I said in my post referred to in the OP "One alternative model is the linearly expanding model proposed by various authors under different guises such as: http://arxiv.org/abs/astro-ph/0306448%5B/URL[/URL],[URL='http://arxiv.org/PS_cache/arxiv/pdf/1110/1110.3054v1.pdf'] Introducing the Dirac-Milne universe[/URL], [URL='http://www.aanda.org/articles/aa/pdf/2013/05/aa20447-12.pdf']The Rh = ct universe without inflation[/URL]

Such a model expands as the Milne empty universe and requires either an EoS of ω=−1/3 or replusive antimatter as in the Dirac-Milne theory in order to produce the Milne model without it being empty."

Now I agree that as a heterodox alternative to the $\Lambda$CDM model there has been relatively little work done on the linearly expanding model, and therefore it has not been tested to the same degree of rigour, however authors that have written about it have found a surprising concordance with various data sets.

(viz: [URL='http://arxiv.org/abs/astro-ph/0306448'] A Concordant “Freely Coasting” Cosmology[/URL], [url=http://www.worldscinet.com/ijmpd/09/0906/S0218271800000682.html]Nucleosynthesis in a universe with a linearly evolving scale factor[/url], [URL='http://arxiv.org/abs/astro-ph/0502370']A case for nucleosynthesis in slowly evolving models [/URL], and the OP link.)

[quote]The "age problem" is not a cosmological test. It's a calculation involving both cosmology and galaxy formation, for that reason people don't use results like that for cosmology. It's convoluted and impossible to tell where the issue lies.[/QUOTE] Agreed, however with such objects as the ultraluminous quasar, SDSS J010013.02+280225.8, a BH ~1.2 × 10[SUP]10[/SUP]M[SUB]⊙[/SUB] seen at z=6.30 - 900 Myrs after BB - (alongside about 40 other quasars at z>6) there does appear to be at least a tension with the standard model. Note the problem in getting such an behemoth to form with Super Eddington accretion, say by direct collapse, is that such a process would not be bright - you need an accretion disc.

(cf. the [I]Nature[/I] letter [URL]http://www.nature.com/nature/journal/v518/n7540/full/nature14241.html#close[/URL]which can be read on the physics arXiv [URL='http://arxiv.org/pdf/1502.07418.pdf']here .[/URL])

Garth

Last edited by a moderator: May 7, 2017
7. Jun 4, 2015

### Chronos

<<Mentor note: Moved from separate thread>>

This paper - http://arxiv.org/abs/1506.01354, Marginal evidence for cosmic acceleration from Type Ia supernovae, calls into question the original SNe-Ia data based conclusion favoring accelerated expansion of the universe. Utilizing additional SNe-Ia data, the authors instead contend the luminosity data does not rule out a constant expansion universe. This appears to apply the brakes to the LCDM model and appears it will make its way into the popular media in rather quickly.

Last edited by a moderator: Jun 4, 2015
8. Jun 4, 2015

9. Jun 4, 2015

### Chalnoth

It might fit the supernova data, as the error bars are pretty large. But there's no way it's going to fit the supernova data, the CMB data, and the baryon acoustic oscillation data at the same time.

10. Jun 5, 2015

### Chronos

There are varying opinions on this point, Chalnoth, like http://arxiv.org/abs/1304.1802, Cosmic Chronometers in the R_h=ct Universe. I still view the OP paper as a serious challenge to LCDM.

11. Jun 5, 2015

### Chalnoth

Yeah, I don't buy that for an instant. His paper that discusses the CMB, for instance, is here:
http://arxiv.org/pdf/1207.0015.pdf

This paper is extremely thin motivation. The fundamental problem is that the $R_h = ct$ universe cannot produce a nearly scale-invariant power spectrum. You can see this complete and utter failure in fig. 6 of the paper, where he plots the first few multipoles of the CMB power spectrum as measured by WMAP. This plot is extremely revealing because he doesn't even bother to try and plot the higher multipoles as measured by WMAP. The richest information in the CMB starts at around $\ell = 200$, and WMAP measured the power spectrum pretty well out to about $\ell = 800$ or so. He only plotted up to about $\ell=50$, barely using the data at all. He completely ignores this figure in the text, probably because this figure completely destroys his analysis: the $R_h=ct$ universe diverges wildly from the WMAP observations by $\ell=20$, which is barely scratching the surface of the richness of available data.

So no, the $R_h = ct$ model cannot possibly explain our universe. Its theoretical motivation is nonexistent and it cannot fit the data.

Last edited: Oct 24, 2016
12. Jun 5, 2015

### ruarimac

What I said was "This doesn't call the standard model into question, what their claiming is that in their analysis showed constant expansion isn't ruled out by one particular test. For this to have impact you would need to make the same claim in the face of the other 2 major tests of cosmology." Nobody has done that. Claiming you believe they will fit the data is irreverent, that's a hypothesis. You wanted to talk about the paper talk about the paper, not 20 others which do not address the point.

On the question of age, no there is no tension with Lambda CDM. There is tension with the product of some models of supermassive black hole formation and LambdaCDM. You can't ignore the assumptions made. Now, if everyone was agreed on how you form supermassive black holes that could be an issue but there are a good few mechanism. Hierarchical assembly for example has no limit to growth. It's a false assumption to say a body which grew like this couldn't be bright, it doesn't have to get bright via the mechanism that formed it.

Last edited by a moderator: May 7, 2017
13. Jun 5, 2015

### Garth

Well, I started by just talking about the OP paper, and mentioned the closed PF thread on the question over SNe 1a as standard candles because it was relevant. You raised the point over two other major tests and I responded to that. Are you objecting to the fact that your assumptions may be challenged by reference to other papers?
Well there is a tension, i.e. a problem, in explaining how these behemoths formed apparently so early and the tension keeps reappearing such as here: - http://arxiv.org/abs/1506.01377[/PLAIN] [Broken] -The Impossibly Early Galaxy Problem (thank you Chronos for that link by pm).
Garth

Last edited by a moderator: May 7, 2017
14. Jun 5, 2015

### ruarimac

I made no assumption, their paper is totally lacking in important information. An objective fact. No one else has done this and for some reason you felt obliged to spam unrelated papers. I came here to talk about a paper, not for you to push some cosmology which only exists by ignoring easily available data.

The objective reader will also notice the "Impossibly Early Galaxy Problem" contains only one set of stellar mass functions. That's just complete nonsense, you cannot pick one model of galaxy formation and declare job done. That's simply indefensible.

15. Jun 5, 2015

### wabbit

I find it strange to claim that data which clearly supports the hypothesis of accelerating expansion over others, albeit not as strongly as other data, should be taken as a challenge of that hypothesis or as support for an alternative it describes as about 150 times less likely, on a par with "empty universe" (cf. table I in the paper).

It only says "contrary to other studies finding accelerated expansion more than 1000 times more likely than constant expansion, we find the ratio to be only 150".

16. Jun 5, 2015

### Garth

Firstly may I belatedly welcome you to these Forums ruarimac, your contributions will be much appreciated. I didn't notice that you were a new member when you first posted.

I referred to the linearly expanding model just because it was mentioned in the OP paper as being "rather surprisingly" consistent with the data and we had already discussed it in the earlier thread.

Otherwise the OP paper does not disprove the standard model, just that the evidence for acceleration is more marginal than previously thought.
So would other sets of stellar mass functions explain how the reported number of massive halos could have formed at such high red shifts?

Garth

Last edited: Jun 5, 2015
17. Jun 5, 2015

### marcus

Enjoying the discussion. thanks all, appreciative welcome to Nuari.
I hope without distracting active participants to insert a pedagogical footnote for any reader unfamiliar with the way positive cosm. curvature const. is deduced from the Type!A Sne data.
The dimness of standard candle supernovae indicates their distance and the redshift luminosity data boils down to redshift-distance data.
In this graph I use the variable S = z+1. The distance to a source is proportional to the area under the curve from 1 to S.
Each curve corresponds to a different estimate of the curvature constant Lambda. that is, to a different asymptotic (long term) expansion rate H. Different possible H are denoted by their corresponding Hubble radii expressed in billions of years. The possibilities are 16.3, 17.3 (the current estimate), 18.3, 40, 1000. The Λ corresponding to a Hubble radius of 1000 billion LY is effectively zero, a negligible curvature constant. There is a big gap between the neighborhood of the current estimate of Λ curve (17.3), and the zero Λ. That gap is the reason people can confidently talk about zero Λ being "150 times less likely." there is a big difference between the areas under the curves.

Last edited: Jun 5, 2015
18. Jun 5, 2015

### Garth

Hi wabbit, the paper is not suggesting the standard $\Lambda$CDM model is to be replaced by the original 'vanilla' zero $\Lambda$ model, as I said above, it finds that the 'constant rate of expansion' (linearly expanding) model (with EoS $p = -1/3 \rho$) is 'rather surprisingly, still quite consistent' with the data.

Garth

19. Jun 5, 2015

### Chalnoth

It's only consistent with a small, cherry-picked fraction of the data.

20. Jun 5, 2015

### wabbit

Right, but "quite consistent" still means "only 150 (closer to 250 actually) times less likely than their best fit showing an accelerated expansion with $\Omega_\lambda\simeq0.6$ " (using the term likely in a loose way, just translating their LLH figures here) - their fig 2 shows that the unaccelerated line just barely escapes being excluded.

As I read it, what they show is that their model may still have a fighting chance if all errors happen to favor the standard model by chance.

LCDM may very well be incorrect, but their model seems a rather unlikely cure in that case, based on their data and their analysis.

Just to add, I have no expertise in these models, just questionning their formulation given the numerical results they show. To me, their statistics simply do not support their conclusions - at best they show that their model is not completely ruled out.

Last edited: Jun 5, 2015
Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook