WMAP 2006 - Mainstream model parameters

[Garth]

WMAP2006: Cosmological Parameters and Large-scale Structure of the Universe

Hence, the data of observational cosmology spreading over scales from 1Mpc to 10000Mpc indicate that \LambdaCDM model with parameters \Omega_{\Lambda} = 0.736, \Omega_m = 0.278, \Omega_b = 0.05, h = 0.68,
\sigma_8 = 0.73 and n_s = 0.96 is the best-fit for whole data set. So this model can be considered as the closest to the true model of Universe within the class of 6-parameter cosmological models.

So that makes \Omega_{Total} = 1.014 - a just closed universe with spherical spatial geometry.

You can get the range of best fit values from Table 3 of that paper, they do not seem to include the flat space model \Omega_{Total} = 1.

Garth

 

[Wallace]

Look closely at their method. They use the same data that has been analyzed previously by other groups however they use a far less rigorous method to fit the data to the same model.

They don't use a CMB code to get the anisotropy power spectrum for each proposal, instead they pre-compute this on a grid with the CMBFAST and then interpolate. This interpolation will lose accuracy. Other studies have not made this approximation.

Secondly they use a simple Gaussian model for the marginalized uncertainties for each parameter. Again this is an approximation not made previously and indeed previous studies have shown asymmetric uncertainty regions that would be poorly fit by a Gaussian profile.

Thirdly they do not use a adequate fitting procedure, using a least squares minimization downhill solver which is inappropriate given the degeneracies between parameters in the LCDM model. Previous work has used a much more rigorous Markov chain Monte Carlo approach that searches the full parameter space.

In short this work offers no new innovations but instead does what has been done before less accurately. I realize that's a pretty blunt assessment but I can't see any redeeming features and given the significance that it appears has been suggested these results have in the OP I think it needs to be made clear.

Don't expect this to be accepted to any journal, note that it is only an online pre-print and destined to remain so.

 

[Garth]

Thank you Wallace for that concise critical analysis of their methods.

The authors are not endorsers on the physics ArXiv, but most have published other papers in refereed journals on similar subjects, so it will be interesting to see what, if anything, happens to this paper.

Garth

 

[marcus]

I concur with Garth, it's good to have a clear concise critique of their methods.
Thanks Wallace!

About peer review publication, I think that Kinematics and Physics of Celestial Bodies (the Ukranian journal) is listed as peer-reviewed in some people's publication lists.
The Naval Observatory library regularly translates every issue and has a complete file of the journal from 1985 to present.

I don't have any direct familiarity with the journal and for all I know it is second-rate.
But it could be respectable journal, so couldn't one say they have already gotten the paper accepted?
It seems possible, at least if their reference "Kinematika i Fizika Nebesnykh Tel (in Ukrainian), V23, N2 (2007)"
indicates hardcopy publication.

In any case, I recall that Spergel et al, WMAP Third Year Results, which they cite, also had an errorbar for Omega_k which was like [-0.041, -0.010]

So these Ukranian guys are not pretending to say anything new, are they? Their paper is just going along with Spergel et al, with maybe (as Wallace says) sloppier methods. Perhaps it is best to see it as a review paper, bringing home to the Ukraine a bunch of international results that have been around for almost two years. Correct me if I have the wrong impression.

What I get from Spergel et al, is that their 68 percent errorbar for Omega is [1.010, 1,041], at least in the case where the EOS w is allowed to vary along with Omega.