Testing gravity with CAMB and CosmoMC

TL;DR

The paper addresses the need for model-independent tests of gravity on cosmological scales by introducing MGCAMB, a patch to CAMB and CosmoMC that implements a generalized modification of linear perturbation growth via $\mu(k,a)$ and $\gamma(k,a)$. It provides a consistent framework usable across all redshifts by evolving all species and maintaining covariant conservation, with a single set of equations that avoids explicit super-/sub-horizon separation. The authors explore specific parametrizations (notably the Brans-Dicke–like form) and alternative representations (such as $\Sigma$, $Q$, and $R$), and implement these in CAMB/CosmoMC, including CMB source modifications and ISW cross-correlation handling. In a worked example, they constrain MG parameters for $f(R)$-like and Yukawa-type models alongside the neutrino mass, finding $B_0<0.4$ (95% C.L.) for the former and $0.7<\beta_1<1.7$ (95% C.L.) for the latter, with $f_{\nu} \lesssim 0.05$ and little degeneracy with MG parameters, supporting GR as a good fit to current data. The patch is publicly released to enable broader, model-independent gravity tests in cosmology.

Abstract

We introduce a patch to the commonly used public codes CAMB and CosmoMC that allows the user to implement a general modification of the equations describing the growth of cosmological perturbations, while preserving the covariant conservation of the energy-momentum. This patch replaces the previously publicly released code MGCAMB, while also extending it in several ways. The new version removes the limitation of late-time-only modifications to the perturbed Einstein equations, and includes several parametrization introduced in the literature. To demonstrate the use of the patch, we obtain joint constraints on the neutrino mass and parameters of a scalar-tensor gravity model from CMB, SNe and ISW data as measured from the correlation of CMB with large scale structure.

Figures (3)

• Figure 1: CMB TT power spectrum for $\Lambda$CDM (black solid), $f(R)$ model with $B_0 = 0.5$ (red dashed) and $\Lambda$CDM+massive neutrino models with $f_{\nu} = 0.05$ (green dash-dotted).
• Figure 2: Marginalized posterior distribution for the $f(R)$ model parameters and 2-D contour plots showing the ranges of and correlations between parameters of interest and the $68$ and $95$% confidence limits, left: with massive neutrinos, right: without massive neutrinos.
• Figure 3: Same as in Fig \ref{['fig-fRMC']} but for the Chameleon-type model considered in the paper.