COSMOMC and w(z)

Use of Healpix, camb, CLASS, cosmomc, compilers, etc.
Post Reply
Matthew Francis
Posts: 6
Joined: September 01 2005
Affiliation: University of Sydney

COSMOMC and w(z)

Post by Matthew Francis » September 13 2005

Hi, I'm attempting to modify COSMOMC to constrain parameters of various forms of evolving dark energy equations of state w(z).

For the supernovae likelyhood calculation this is not difficult, however I'm not so sure of how to treat the CMB calculation. My understanding is that dark energy does not affect the physics of the CMB greatly, as in all parametrisations the energy density is negligable in this epoch. However the form of dark energy alters the angluar diameter distance to the CMB, and affects the observed power spectrum.

My approach is to determine an effective constant w that gives the same angular diameter distance to the CMB as the w(z) model of the current proposal, and feed this value into CAMB, on the assumption that models of dark energy giving the same angular diameter distance to the CMB are degenerate. I've also changed the function AngularDiameterDistance to consider the w(z) model of the current proposal.

Is this a physically reasonable approach?

Michael Doran
Posts: 41
Joined: November 22 2004
Affiliation: ITP Heidelberg
Contact:

COSMOMC and w(z)

Post by Michael Doran » September 13 2005

Hi,

given the precision of WMAP, i'd say it's not a good approach to treat all
models with same angular diameter distance as degenerate: the ISW depends
on the detailed evolution of dark energy.

Btw: it's not entirely correct that in all parametrization, the dark energy
density is negligible at last scattering, even though you're right that most
parametrizations on the market are not particularily suited to express dark
energy during recombination. From the point of view of model building,
however, having some dark energy around in the early universe is quite
attractive, because it corresponds to an exponential potential which is well
motivated from high energy physics and yields an attractor solution early on.

So in my opinion, you will also have to run the CMB calculation. But as the code is fast, why not do it :-) ?

Post Reply