Spatial Curvature in $f(R)$ Gravity

TL;DR

This study constrains four $f(R)$ gravity models using cosmological data, examining the impact of spatial curvature and the Hubble constant, and finds a slight preference for an open universe when considering local $H_0$ measurements.

Contribution

It compares flat and non-flat $f(R)$ models using MCMC with cosmological data, highlighting the influence of $H_0$ on spatial curvature constraints.

Findings

Bounds on $ abla_{k,0}$ are compatible with flat geometry.

Higher $H_0$ favors an open universe at over 1$\sigma$.

Deviations in the growth rate are negligible across models.

Abstract

In this work, we consider four $f(R)$ gravity models -- the Hu-Sawicki, Starobinsky, Exponential and Tsujikawa models -- and use a range of cosmological data, together with Markov Chain Monte Carlo sampling techniques, to constrain the associated model parameters. Our main aim is to compare the results we get when $\Omega_{k,0}$ is treated as a free parameter with their counterparts in a spatially flat scenario. The bounds we obtain for $\Omega_{k,0}$ in the former case are compatible with a flat geometry. It appears, however, that a higher value of the Hubble constant $H_0$ allows for more curvature. Indeed, upon including in our analysis a Gaussian likelihood constructed from the local measurement of $H_0$, we find that the results favor an open universe at a little over $1\sigma$. This is perhaps not statistically significant, but it underlines the important implications of the Hubble tension for the assumptions commonly made about spatial curvature. We note that the late-time deviation of the Hubble parameter from its $\Lambda$CDM equivalent is comparable across all four models, especially in the non-flat case. When $\Omega_{k,0}=0$, the Hu-Sawicki model admits a smaller mean value for $\Omega_{\text{cdm},0}h^2$, which increases the said deviation at redshifts higher than unity. We also study the effect of a change in scale by evaluating the growth rate at two different wavenumbers $k_\dagger$. Any changes are, on the whole, negligible, although a smaller $k_\dagger$ does result in a slightly larger average value for the deviation parameter $b$.