Speaker
Description
The parameter $\sigma_8$, which represents the root-mean-square (rms) mass fluctuations on a scale of $R_8=8h^{-1}$ Mpc (where $h$ is the reduced Hubble parameter), is commonly used to quantify the amplitude of matter fluctuations at linear cosmological scales. Derived quantities, such as $S_8=\sigma_8(\Omega_m^{0}/0.3)^{0.5}$, are also frequently employed. However, the dependence of $R_8$ on $h$ complicates direct comparisons of $\sigma_8$ values obtained under different assumptions about $H_0$, since $\sigma_8$ in such cases characterizes the amount of structure at different physical scales. This issue arises both when comparing $\sigma_8$ values from fitting analyses of cosmological models with differing $H_0$ posteriors, and when contrasting constraints from experiments that employ different priors on the Hubble parameter. As first noted by Ariel G. Sánchez in Phys. Rev. D 102, 123511 (2020), quantifying the growth tension using $\sigma_8$ or $S_8$ can introduce substantial biases and couple the growth and Hubble tensions in an intricate and uncontrolled way. To address these challenges, Sánchez proposed an alternative parameter, $\sigma_{12}$, defined as the rms mass fluctuations at a scale of $12$ Mpc, which is independent of $h$. Although Sánchez's work was published five years ago and other authors have since highlighted the limitations of $\sigma_8$, much of the cosmological community --including large collaborations -- continues to rely on this parameter rather than adopting $\sigma_{12}$, seemingly due only to historical considerations. In this work, we illustrate the biases introduced by the use of $\sigma_8$ through some clear examples, aiming to motivate the community to transition from $\sigma_8$ to $\sigma_{12}$. We show that the bias found in models with large values of $H_0$ is more prominent. This artificially complicates the search for a model that can efficiently resolve the Hubble tension without exacerbating the growth tension. We argue that the worsening of the growth tension in these models is much less pronounced than previously thought or may even be nonexistent.
