Probing Generalized Emergent Dark Energy with DESI DR2

TL;DR

GEDE offers a minimal extension to ΛCDM by introducing a free parameter $Δ$ that controls dark-energy evolution, with ΛCDM recovered at $Δ=0$ and PEDE at $Δ=1$, enabling a test of emergent dark-energy scenarios. The authors perform a Bayesian analysis using nested sampling (PyPolyChord) on DESI DR2 BAO, three SNe Ia compilations, and CMB distance priors, constraining $h$, $Ω_{m0}$, and $Δ$ via the likelihood $\mathcal{L}_{tot}=\mathcal{L}_{BAO}\times\mathcal{L}_{SNe Ia}\times\mathcal{L}_{CMB}$. The GEDE model yields $ω(z)$ that approaches $-1$ without crossing, with the best-fit $Δ$ shifting sign from positive (CMB+DESI DR2) to negative upon inclusion of SN calibrations, and a modest improvement in fit around $z\sim1$ over ΛCDM; Bayesian evidence shows weak-to-moderate support depending on the SN dataset, indicating consistency but only a mild preference for GEDE. Overall, GEDE remains a viable extension of ΛCDM that warrants future high-precision measurements, particularly near $z\sim1$, to decisively determine the presence of emergent dark-energy behavior.

Abstract

As an update on the initial findings of DESI, the new results provide the first hint of potential deviations from a cosmological constant ($ω=-1$), which, if confirmed with significance $>(2-4)σ$, would challenge the validity of $Λ$ within the $Λ$CDM model. We explore the Generalized Emergent Dark Energy (GEDE) model using recent BAO measurements from DESI DR2, Type Ia supernova compilations, and CMB distance priors. Employing nested sampling, we constrain the parameter $Δ$, which characterizes deviations from $Λ$CDM. Our analysis shows that with CMB+DESI DR2 alone, GEDE tends to prefer positive values of $Δ$. However, when different SNe Ia calibrations are included, the model favors negative values of $Δ$, corresponding to an earlier injection of dark energy. The Marginalized constraints on $ω(z)$ further shows that GEDE sharply emerges but then asymptotes to $ω=-1$ without crossing it. At $z \sim 1$ data, GEDE provides a better fit than $Λ$CDM, while at $z \lesssim 0.5$ the data favor $ω> -1$, bringing the model deviate from $Λ$CDM. Bayesian model comparison shows weak support for GEDE with CMB+DESI DR2 ($\ln BF=1.96$), moderate with PP ($\ln BF=2.65$), weak-to-moderate with Union3 ($\ln BF=2.34$), and weak with DES-SN5Y ($\ln BF=1.44$). Overall, GEDE is consistent with current data and mildly favored when SNe Ia are included, making it a viable extension of $Λ$CDM that merits further investigation with future high precision measurements.

Figures (4)

• Figure 1: The figure shows the contour plot of the GEDE model showing the 68% (1$\sigma$) and 95% (2$\sigma$) confidence levels, using DESI DR2, CMB, and Type Ia Supernova datasets (PP, DES-SN5Y, and Union3).
• Figure 2: The figure shows the marginalized posterior distributions in the $\Delta$–$\Omega_{m0}$ plane for the GEDE model at 68% (1$\sigma$) and 95% (2$\sigma$) confidence levels, using DESI DR2, CMB, and Type Ia Supernova datasets (PP, DES-SN5Y, and Union3). The vertical dashed gray line at $\Delta = 0$ represents the standard $\Lambda$CDM limit, while the vertical dashed gray line at $\Delta = 1$ corresponds to the PEDE scenario.
• Figure 3: The figure shows the evolution of the angle-averaged distance, $D_V(z)/r_d$, and the ratio of transverse to line-of-sight comoving distances, $D_M(z)/D_H(z)$, scaled by $z^{-2/3}$ and $z^{-1}$, respectively, in the upper panel. The bottom panel displays the corresponding residuals.
• Figure 4: Marginalized constraints on the dark energy equation of state, $\omega(z)$, and normalized energy density, $f_{\mathrm{DE}}(z)$, within the framework of the GEDE model.