CosmoCoffee

Hi Antony, Several questions,

First of all for WMAP combination with other CMB observations, on regions where the ls overlap Spergel et al claimed there could be some correlation
between WMAP and for example, Boomerang. So what have you done regarding the overlap places and what is you comment on that?

Secondly you also mentioned that Fig.14 of Spergel et al astro-ph/0603449 is not correct. Actually we could not reproduce their picture with their available chains online. Anyway in more details, basing on what did you claim their Fig. 14 is wrong?:)

Finally I think there are cases where both lensing and SZ effects should be included, so when will the updated CAMB and COSMOMC appear including the option of SZ? Thanks!

Bo Feng

The usual procedure with overlapping l ranges is to use only the bandpowers on smaller scales where WMAP is noise dominated - I just used the CosmoMC default high l bandpowers. (however the effect of including overlap is a rather small error).

The problem with Fig. 14 is precisely that it doesn't seem to correspond with the WMAP chains - one of those inevitable glitches that I believe it will be fixed when they update the paper. (I only mention it in the footnote because I've already seen this figure used in four people's talks!)

It is fairly trivial to modify the current CosmoMC to include SZ if you want to: I set the WMAP SZ_amp parameter in subroutine ParamsToCMBParams in params_CMB.f90, using the unused parameter 13 for the SZ amplitude as in the WMAP chains. Maybe it will be in the next CosmoMC update.

PS. The updated Spergel et al. Fig 14 is now available from Hiranya's Irvine talk:

http://www.physics.uci.edu/CMB/Peiris.pdf

Hi Antony,

One technical question: In your fittings to SDSS/Viel Lyman alpha forest,
did you consider additional bias parameters? Thanks in advance!

Bo Feng

May I ask a general and more radical question about the interpretation of the WMAP data?

There are many parameters (The Spergel paper lists 23), which are made to fit the data refining the mainstream LambdaCDM model, and the value of one parameter often depends on the value of more basic ones, the "priors".

One of the more basic parameters is obtained from the first peak of the power spectrum, which is interpreted as being consistent with a flat, or nearly flat, spatial geometry. (This yields the total density Omega = 1.010 +0.009/-0.016)

However, as the data is angular in nature and conformal transformations are angle preserving, the data is consistent with a conformally flat spatial geometry as well.

A finite and closed conformally flat geometry would also be consistent with a low-l power deficiency. The interpretations of the other peaks, such as the baryon density determined by the third peak, are also degenerate with respect to conformal transformations.

Whereas such conformally flat geometries are not predicted by GR, the mainstream model is so predicted. Hence is not a statement such as in the Spergel paper: "The standard model of cosmology has survived another rigorous set of tests." a circular argument? For the confirmation of the GR model has been obtained using data interpreted under the GR cosmological theory! Certainly by fitting the many parameters and invoking Inflation, non-baryonic DM and DE, all undiscovered in laboratory physics, the process and data have been made self-consistent. However, the possibility exists that the data is also consistent with a non GR gravitational theory conformally related to it.

The possibility that GR is not the last word in gravitational theory may be indicated by the problem in developing a quantum gravity theory together with such conundrums as the Pioneer anomaly.

Garth

Bo Feng wrote: One technical question: In your fittings to SDSS/Viel Lyman alpha forest,
did you consider additional bias parameters? Thanks in advance!
Bo Feng

This paper now only had the SDSS data in it - for a much fuller analysis see astro-ph/0604310 and astro-ph/0604335. I did not include any bias parameters, but the uncertainty from many Lyman-alpha parameters is already encoded in the SDSS likelihood code of astro-ph/0407377 which I used.

However, as the data is angular in nature and conformal transformations are angle preserving, the data is consistent with a conformally flat spatial geometry as well.

WMAP alone does have a geometrical degeneracy, that is clear. However if you add almost any other data you will find the combined constraint is close to the flat universe model. e.g. if you believe the HST Hubble constraint, you can see from the figure 20 of astro-ph/0603449 that a nearly flat universe is consistent with this, a significantly non-flat universe is not. In this sense the flat prior is a prior from non-CMB data. The flat prior also correponds closely to many people's beliefs about the early universe (e.g. inflation).

Antony Lewis wrote:

However, as the data is angular in nature and conformal transformations are angle preserving, the data is consistent with a conformally flat spatial geometry as well.

WMAP alone does have a geometrical degeneracy, that is clear. However if you add almost any other data you will find the combined constraint is close to the flat universe model. e.g. if you believe the HST Hubble constraint, you can see from the figure 20 of astro-ph/0603449 that a nearly flat universe is consistent with this, a significantly non-flat universe is not. In this sense the flat prior is a prior from non-CMB data. The flat prior also correponds closely to many people's beliefs about the early universe (e.g. inflation).

Yes but that Fig. 20 was for "The models in the figure are all power-law CDM models with dark energy and dark matter, but without the constraint that [tex]\Omega_m + \Omega_{\Lambda} = 1[/tex] (model M10 in Table 3)." The assumed prior here is of course GR.

I was thinking of a more radical, conformally flat, modified GR model.

Garth

Antony Lewis wrote:The usual procedure with overlapping l ranges is to use only the bandpowers on smaller scales where WMAP is noise dominated - I just used the CosmoMC default high l bandpowers. (however the effect of including overlap is a rather small error).

Hi Antony, I have another question regarding small scale CMB (CMBsmall), in the paper by WMAP team astro-ph/0603449 they have considered the rescaling and calibration uncertainties. On the other hand I suppose typically one does not include such detailed technial aspects, such as in astro-ph/0604335 and astro-ph/0608277. So what is your comment on this? Thanks!

Bo Feng

I think you generally need to include the calibration and beam uncertainties. These are modelled in CosmoMC, and as far as I know both papers you mention included them.

I see, thanks!

Antony Lewis wrote:I think you generally need to include the calibration and beam uncertainties. These are modelled in CosmoMC, and as far as I know both papers you mention included them.