Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Apparent Dark-Energy Evolution from Cosmic Inhomogeneities

Yonadav Barry Ginat, Pedro G. Ferreira

TL;DR

The paper addresses whether mild cosmic inhomogeneities in a $oldsymbol{C3}$-dominated universe can masquerade as evolving dark energy. It analyzes three inhomogeneity realizations—Dyer–Roeder-like lumpy voids with $oldsymbol{ abla}(z)$, back-reaction from small-scale structure via an effective stress tensor, and large-scale curvature inhomogeneity in the LTB framework—and evaluates their impact on distance measures such as $D_A$, $D_V$, and $D_L$, mapping the results to an effective EOS $w(a)=w_0+w_a(1-a)$. All three models tend to produce a best-fit region in the $(w_0,w_a)$ plane with $w_0>-1$ and $w_a<0$ that overlaps DESI/DES, indicating that inhomogeneities can mimic some aspects of evolving dark energy at a percent level, though they do not fully replace a cosmological constant. The work highlights a potential systematic in dark-energy inference from light propagation in an inhomogeneous universe and calls for more precise, flexible modeling to robustly constrain the physics driving cosmic acceleration.

Abstract

A mildly inhomogeneous universe with a cosmological constant may look like it contains evolving dark energy. We show that could be the case by modelling the inhomogeneities and their effects in three different ways: as clumped matter surrounded by voids, as back-reaction of small-scale structure on the overall expansion of the Universe, and, finally, as a large-scale curvature inhomogeneity. In all of these cases, the propagation of light is affected, and differs from that in a homogeneous and isotropic universe. The net result is that cosmological observables, such as angular diameter and luminosity distances, become distorted. We find, in all three models, that the inclusion of these effects pushes the distance-redshift relation towards closer agreement with recent data from both supernovae Ia from the Dark Energy Survey, and from baryon acoustic oscillations from the Dark Energy Spectroscopic Instrument. The amount of inhomogeneity in these models might not be enough to explain the entirety of the deviation from a cosmological constant, but is found to be of a similar order of magnitude, hinting that these data may be consistent with a universe dominated by a cosmological constant.

Apparent Dark-Energy Evolution from Cosmic Inhomogeneities

TL;DR

The paper addresses whether mild cosmic inhomogeneities in a -dominated universe can masquerade as evolving dark energy. It analyzes three inhomogeneity realizations—Dyer–Roeder-like lumpy voids with , back-reaction from small-scale structure via an effective stress tensor, and large-scale curvature inhomogeneity in the LTB framework—and evaluates their impact on distance measures such as , , and , mapping the results to an effective EOS . All three models tend to produce a best-fit region in the plane with and that overlaps DESI/DES, indicating that inhomogeneities can mimic some aspects of evolving dark energy at a percent level, though they do not fully replace a cosmological constant. The work highlights a potential systematic in dark-energy inference from light propagation in an inhomogeneous universe and calls for more precise, flexible modeling to robustly constrain the physics driving cosmic acceleration.

Abstract

A mildly inhomogeneous universe with a cosmological constant may look like it contains evolving dark energy. We show that could be the case by modelling the inhomogeneities and their effects in three different ways: as clumped matter surrounded by voids, as back-reaction of small-scale structure on the overall expansion of the Universe, and, finally, as a large-scale curvature inhomogeneity. In all of these cases, the propagation of light is affected, and differs from that in a homogeneous and isotropic universe. The net result is that cosmological observables, such as angular diameter and luminosity distances, become distorted. We find, in all three models, that the inclusion of these effects pushes the distance-redshift relation towards closer agreement with recent data from both supernovae Ia from the Dark Energy Survey, and from baryon acoustic oscillations from the Dark Energy Spectroscopic Instrument. The amount of inhomogeneity in these models might not be enough to explain the entirety of the deviation from a cosmological constant, but is found to be of a similar order of magnitude, hinting that these data may be consistent with a universe dominated by a cosmological constant.
Paper Structure (9 sections, 14 equations, 2 figures)

This paper contains 9 sections, 14 equations, 2 figures.

Table of Contents

  1. Introduction.
  2. A Lumpy Universe.
  3. The effect of back-reaction.
  4. The Lemaître--Tolman--Bondi model
  5. Results.
  6. Discussion.
  7. Acknowledgements.
  8. A model for back-reaction
  9. Mock Likelihood

Figures (2)

  • Figure 1: Top: Three examples of $D_V(z)$ curves for the toy models explored here, with the DESI DR2 data-points DESI:2025zgx, compared with the prediction of $\Lambda$CDM with PlanckPlanck:2018vyg parameters (denoted $D_V^{\rm \Lambda CDM}$). Left: Dyer--Roeder (orange & navy blue, dash-dotted; equation \ref{['eqn:Dyer-Roeder']}), middle: a back-reaction toy-model (red; equation \ref{['eqn:back-reaction D_V']}), right: an LTB model (light blue; equation \ref{['eqn:LTB metric']}). We also show the $D_V(z)$ curve for the $w_0 w_a$CDM model preferred by DESI DR2 + DES Y5 + CMB (green, dashed) DESI:2025zgx; observe that it asymptotes to a value different from unity at high redshift because of the difference of the cosmological parameters from Planck's. Bottom: Same as above, but for distance moduli, compared with DES Y5 (re-analysed) DES-Dovekie2026. The distance-modulus curves were shifted to have the same mean (evaluated at the bin redshifts) as the data, with binning as in refs. DESI:2025zgxLietal2026.
  • Figure 2: A plot of the $(w_0,w_a)$ values for which $D_{V,w}(z)$ lies closest to the $D_V(z)$ curve, for the three toy models considered here. Left: Dyer--Roeder approximation (see text for details); varying $\alpha_0$ results in translation along the rays. The middle panel displays the back-reaction toy model, where varying $\beta_1$ moves along a ray from $\Lambda$CDM ($(w_0,w_a)=(-1,0)$), while varying $\beta_2$ results in moving from one ray to another (here $\beta_2 \in [-2,5]$). Right: the convex hull of the $(w_0,w_a)$ points for $300$ randomly-selected quadruplets of $(c_0,c_1,r_0,r_1)$ such that $\abs{c_0},\abs{c_1} \leq 0.1$. The confidence contours from ref. DESI:2025zgx for DESI + CMB and DESI + DES + CMB data are over-plotted.