CosmoCoffee

What's New on LAMBDA: 12-Jul-06

http://lambda.gsfc.nasa.gov/outreach/whatsnew.cfm

Augmented Cosmological Parameter Table
The cosmological parameter table has been augmented with 3 additional models: non-flat CDM ("ocdm"), non-flat CDM with quintessence ("ocdm+wpert"), and flat CDM with a running spectral index and tensors ("lcdm+run+tens").

(This is an automated web monitor message)

There seems to be something wrong with the numbers in the 'ocdm' parameters using WMAP. The most obvious thing is that [tex]\Omega_m = 0.77[/tex] and [tex]\Omega_\Lambda = 0.75[/tex] don't add up to [tex]\Omega_{tot} = 1.15[/tex]. Also, with [tex]h=0.45[/tex] one doesn't get [tex]\Omega_mh^2 = 0.128[/tex].
I haven't looked for other errors, and also haven't looked carefully at any of the other parameter tables.
Can someone check/clarify this?

I checked some of the other parameter tables and they look OK to me. The "ocdm" chain I think was done more recently than the others, and maybe something caused a glitch in the automatic table generation. It should get fixed sometime in the next few days.

But I just looked at the chain itself a bit, and my guess is that the problem is that \Omega_m and \Omega_\Lambda are both poorly constrained, and the value quoted for \Omega_\Lambda appears to be an upper limit. When I take the average \Omega_\Lambda for the chain, I get 0.38, which together with the average \Omega_m matches the \Omega_k given. But the constraints on \Omega_m and \Omega_\Lambda from WMAP alone are just very poor (though the constraint on their sum \Omega_{tot} isn't quite as horrible).

The combination \Omega_m h^2 actually isn't bad at all, so I think those numbers are probably fine. It doesn't quite match the given \Omega_m and H_0, but the constraints on those two individually are so bad that it's well within the error.

I heard back, and the "official" word is that the \Omega_\Lambda number is an upper limit and the file will be fixed to denote it as such.

As for the \Omega_m, H_0, and \Omega_m h^2 not agreeing, it turns out that it's important to remember that these are all mean values over some likelihood distribution. In particular, the mean of H_0 is not equal to the sqrt of the mean of H_0^2 even for a normal distribution, unless the standard deviation is small compared to the mean. For the "ocdm" chain neither of these conditions is true. So the quoted numbers are actually all correct.