WMAP3 analysis

Post Reply
Antony Lewis
Posts: 1941
Joined: September 23 2004
Affiliation: University of Sussex
Contact:

WMAP3 analysis

Post by Antony Lewis » March 21 2006

Firstly, congratulations to all involved involved in WMAP3. The TT results are really great, and the polarization now looks fairly believable.

My main TT question is: why is SZ included in the analysis but not CMB lensing? If you are going to model secondaries, it would seem sensible to me to include all the secondaries which have comparable signals. CMB lensing in fact pushes n_s the opposite direction from SZ, so including it would (slightly) decrease the significance of the n_s <1 detection. In fact I suspect that not including either effect would be rather accurate, since the effects nearly cancel.

The main polarization question is: what is the systematic error on \tau from the foreground subtraction? As far as I can see there is no attempt made to marginalize over the templates (as they do for the temperature) to include foreground subtraction uncertainties in the error budget.

Some relevant plots are in my Moriond talk, at

http://cosmologist.info/notes/Moriond2006.ppt

where I also show WMAP TT-only parameters, that using a \tau prior \tau = 0.1\pm 0.03 with WMAP TT gives essentially identical basic parameters to using the full dataset, and the constraints including Lyman alpha (LUQAS and SDSS)

Mike Hudson
Posts: 5
Joined: September 09 2005
Affiliation: U Waterloo
Contact:

WMAP3 analysis

Post by Mike Hudson » March 22 2006

Antony

What is the effect of CMB lensing on Omega_m and sigma_8? Its a bit difficult to tell from your Moriond PPT. It looks like the normalization increases, and so does Omega_m and n_s. All of these effects will presumably increase sigma_8 ...?

Antony Lewis
Posts: 1941
Joined: September 23 2004
Affiliation: University of Sussex
Contact:

Re: WMAP3 analysis

Post by Antony Lewis » March 26 2006

From my chains (slightly different priors to WMAP3 analysis)

No SZ, no lens:

\sigma_8 = 0.756 \pm 0.05
\Omega_m = 0.238 \pm 0.035

No SZ, lens:

\sigma_8 = 0.762 \pm 0.05
\Omega_m = 0.243 \pm 0.036

SZ, no lens

\sigma_8 = 0.741 \pm 0.05
\Omega_m = 0.232 \pm 0.036

Not a very big deal.

Mike Hudson
Posts: 5
Joined: September 09 2005
Affiliation: U Waterloo
Contact:

WMAP3 analysis

Post by Mike Hudson » March 29 2006

Thanks ...

My analysis of several peculiar velocities data sets give [tex]\sigma_8 (\Omega_m/0.3)^{0.6}=0.85\pm0.05[/tex] (Pike and Hudson, 2005, ApJ 635, 11)

This result is similar in its degeneracies and best-fit values to weak lensing results, so presumably overlaps the WMAP error contours, as in Fig 7 of Spergel et al.

Hans Kristian Eriksen
Posts: 60
Joined: September 25 2004
Affiliation: ITA, University of Oslo
Contact:

Re: WMAP3 analysis

Post by Hans Kristian Eriksen » July 11 2006

Antony Lewis wrote: My main TT question is: why is SZ included in the analysis but not CMB lensing? If you are going to model secondaries, it would seem sensible to me to include all the secondaries which have comparable signals. CMB lensing in fact pushes n_s the opposite direction from SZ, so including it would (slightly) decrease the significance of the n_s <1 detection. In fact I suspect that not including either effect would be rather accurate, since the effects nearly cancel.
So, what is the latest "consensus" in the community about this issue? Do people think that SZ should be included, or is it better to leave it out?

Post Reply