
CosmoCoffee

[1007.3725]
Precision Cosmology Defeats Void Models for Acceleration

Authors:  Adam Moss, James P. Zibin, Douglas Scott 
Abstract:  The suggestion that we occupy a privileged position near the centre of a
large, nonlinear, and nearly spherical void has recently attracted much
attention as an alternative to dark energy. Putting aside the philosophical
problems with this scenario, we perform the most complete and uptodate
comparison with cosmological data. We use supernovae and the full cosmic
microwave background spectrum as the basis of our analysis. We also include
constraints from radial baryonic acoustic oscillations, the local Hubble rate,
age, big bang nucleosynthesis, the Compton ydistortion, and for the first time
include the local amplitude of matter fluctuations, \sigma_8. These all paint a
consistent picture in which voids are in severe tension with the data. In
particular, void models predict a very low local Hubble rate, suffer from an
"old age problem", and predict much less local structure than is observed. 

[PDF]
[PS] [BibTex] [Bookmark]

View previous topic :: View next topic 
Author 
Message 
Maciej Bilicki
Joined: 12 May 2010 Posts: 19 Affiliation: University of Cape Town

Posted: July 22 2010 


The authors have done a thorough analysis of different observational constraints (e.g. CMB, SNe, BAO, σ_{8}, ...) and seem to rule out inhomogeneous models with a large void. Any opinions on this work? 

Back to top 


Michael Schneider
Joined: 19 Nov 2006 Posts: 9 Affiliation: Lawrence Livermore National Laboratory

Posted: July 23 2010 


This indeed looks very thorough, but I was confused by the different claims of this other recent paper: 1007.3065 that says they can fit a similar collection of data (albeit neglecting sigma_8) by allowing nonzero spatial curvature combined with a varying void profile to match larger H0 values. It looks to me like Moss et al. include such scenarios in their analysis, including quite general void profiles that fail to alleviate the low H0 prediction.
Can anyone explain the apparently discrepant claims between these two papers? 

Back to top 


Timothy Clifton
Joined: 26 Jul 2010 Posts: 1 Affiliation: QMUL

Posted: July 26 2010 


Embedding these voids in a nonflat background helps fit the CMB peaks as the shift parameter is mostly sensitive to spatial curvature at high z, rather than low (see 0902.1313). This also helps the problem with low H0 a bit, but not by enough.
It looks like Biswas et al. (1007.3065) also have a low H0 problem in a similar way to Moss et al. They seems to find that very very wide voids help alleviate this a bit, but even in this case it looks like H0 is a little low (h=0.62 to 2 sigma, is this too low?).
Possible ways of getting more acceptable H0 are to include inhomogeneous radiation fields, as in Clarkson and Regis (1007.3443), or allowing a nonsimultaneous big bang, as in 0902.1313. 

Back to top 


Wessel Valkenburg
Joined: 14 Mar 2008 Posts: 6 Affiliation: RWTH Aachen

Posted: July 26 2010 


Hi there,
About the confusion: in both papers the finding is that voids embedded in an EdS metric are ruled out. So the technical results from both papers are really in agreement. It is just that since in 1007.3065 we also consider global curvature (as Timothy explains), and find good overall fits (that is, simultaneous CMB+BAO+HST+SN, where H0 being only one datapoint can easily be off by 2 σ), we draw a completely different overall conclusion. But yes, h~60 may not be good enough, although this depends on your choice of datapoints as well. If you take for example H_{0}=62 from astroph/0603647, you'll see that the void can gives almost exactly as good a fit as LCDM to CMB+BAO+HST+SN.. Anyway, to remain conservative, we provide fits to other H_{0}values as well.
One significant difference is for the BAO though. In 1007.3725 the authors fit the radial BAO only, which seems to favour LCDM over voids. In 1007.3065 we fit the 'complete' BAO (so radial times angular scales), which seems to favour voids over LCDM. It is not clear to me which of the two approaches is more fair. Does anyone have an idea?
Having said that, 1007.3065, 1007.3443 and 1007.3725 came out practically simultaneously, agree in the principle results, but apart from that addressed different issues. So one should really read them all.... (that should help taking away the confusion.) 

Back to top 


Jim Zibin
Joined: 26 Jul 2010 Posts: 4 Affiliation: University of British Columbia

Posted: July 26 2010 


About the apparent discrepancy between our paper and Biswas et al, 1007.3065, I'd note that for their higher local Hubble rate cases, Biswas et al find very large local curvature parameter, Ω_{k,in}, as their Table XIII shows. According to that table, their higher local H_{0} profiles require what we would consider unrealistically high local Ω_{k,in}, eg. Ω_{k,in} = 0.936 or 0.984. In our study we imposed what we considered a conservative prior of Ω_{k} < 0.9 at the void centre today. We suspect this is the source of the discrepancy. Even so, it seems their highest H_{0} value (57 km/s/Mpc) is still low compared with most local estimates.
About the BAO, we argued in our paper that the angular scale should be a weaker discriminator from LCDM because it depends on the angular diameter distance, which is basically tuned in void models to fit the LCDM angular distance. The radial scale is a very strong discriminator, and the only question is about the statistical significance of the radial BAO data.
Having said this, I agree with Wessel that it looks as though our two papers are largely in agreement. 

Back to top 


Wessel Valkenburg
Joined: 14 Mar 2008 Posts: 6 Affiliation: RWTH Aachen

Posted: July 27 2010 


Yes, it is true that we allow for any value for Ω_{k,in} (the curvature at the center of the void) in 1007.3065. But as H_{0,in} (the observed local Hubble rate) mostly depends on Ω_{k,in}, the 5% difference in local curvature (Ω_{k,in}=0.935 in 1007.3065 in stead of 0.9 in 1007.3725) does not explain the 25% difference in H_{0,in} (57 vs 45). I think the effect is twofold, one is the higher local curvature, but the other (and that is the main finding in 1007.3065) is the global curvature which has a significant effect on the distance to last scattering, and hence allows for playing with the expansion rate.
For example, the model in 1007.3065 with the highest H_{0,in} also has a CMB temperature today of 4.54 K (where the observed temperature is still 2.726K, as it is redshifted due to the void). A universe so young yet with the correct distance to last scattering, is only possible with a global curvature term.
By playing more with curvature profiles, I think it is possible to get an even higher H0. The more solid observation that can rule out even these very empty voids that nevertheless could fit HST, seems to me to be the comptony scattering and σ_{8} that you have addressed in 1007.3725 for the first time.
Would you agree? 

Back to top 


Jim Zibin
Joined: 26 Jul 2010 Posts: 4 Affiliation: University of British Columbia

Posted: July 27 2010 


It's much more than a 5% difference. 0.9 was our cutoff. Our best fit Ω_{k,in} is roughly 0.8, as our Fig 5 shows. I agree that the details of the void profile and whether there's spatial curvature outside the void or not will have an effect on how high we can get H_{0}, but it will be a subdominant effect to the local curvature.
And I encourage you to try to get higher H_{0}! 

Back to top 


Wessel Valkenburg
Joined: 14 Mar 2008 Posts: 6 Affiliation: RWTH Aachen

Posted: July 27 2010 


Ah, yes, I see your point.
Maybe it is important to point out that we are fitting different observations in the papers. In 1007.3725 you always fit the local σ_{8}, which weighs in on the allowed values for Ω_{matter,in}, as you pointed out (private). In the fits where we included the matter power spectrum in 1007.3065, we also found much lower Ω_{k,in} and hence low H_{0,in}. On the other hand, the profiles that led to a high H_{0,in}, change so significantly around the redshifts relevant for the large scale structure, that we didn't want to fit them with an effective FLRW. So I think the source of different Ω_{k,in} is not so much the prior, as it is the data that one fits.
About the global curvature being subdominant: yes, you're absolutely right, the main ingredient is Ω_{k,in}. But, the two values you quoted from our paper for Ω_{k,in}, 0.98 and 0.93, are for models that do predict the same H_{0,in}. The difference is the shape of profile and the global curvature..
It would be good to be able to calculate some LSS for the rapidly varying profiles too... Work to do! 

Back to top 


Jim Zibin
Joined: 26 Jul 2010 Posts: 4 Affiliation: University of British Columbia

Posted: July 27 2010 


Just one more clarification: our allowed values of Ω_{k,in} (our Fig 5) are determined solely by the CMB + SN data (mostly the SNe, actually). But I agree that pushing the profiles towards larger σ_{8} would mean pushing them towards smaller Ω_{k,in}, since models with smaller Ω_{k} will experience less suppression of perturbation growth. 

Back to top 




You cannot post new topics in this forum You cannot reply to topics in this forum You cannot edit your posts in this forum You cannot delete your posts in this forum You cannot vote in polls in this forum

