Probing the Dynamics of Gaussian Dark Energy Equation of State Using DESI BAO

TL;DR

This work uses DESI DR2 BAO data, complemented by Pantheon+ and DESY5 SN samples, cosmic chronometers, and Planck distance priors to reconstruct the dark energy equation of state $w(a)$ via multiple parametrizations, including a new Gaussian-like BellDE form. A Bayesian framework with MCMC explores CPL, PADE, GEDE, GDE, and BellDE, revealing that high-redshift $w(z)$ can be phantom while late-time $w(z)$ remains near or above $-1$, with BellDE offering localized dynamics that avoid early phantom behavior. BellDE yields competitive information criteria and tighter low-redshift constraints, while DESI data sharpen $H_0$ estimates and illuminate potential deviations from $\Lambda$CDM across redshift. The results underscore the value of combining DESI BAO with SN data to probe evolving dark energy and hint at modest alleviation of the $H_0$ tension in some dynamical scenarios, motivating further analyses with upcoming surveys and full-shape information.

Abstract

We present an updated reconstruction of the DE equation of state (EoS), $w(a)$, employing the newly released DESI DR2 Baryon Acoustic Oscillation data. This analysis constrains the cosmological scenarios influenced by different models through the joint examination of a range of recently available cosmological probes, specifically the Pantheon+ sample and the DESY5 sample of Type Ia Supernovae, baryon acoustic oscillations, Hubble parameter measurements derived from cosmic chronometers, and cosmic microwave background distance priors based on the Planck 2018 data. Furthermore, we provide a concise perspective on the dynamical evolution of all models (CPL, PADE, GEDE, GDE, BellDE) and their interrelations. A Bayesian inference procedure is adopted to estimate the models parameters that yield the best fit to the data. The EoS remains within the phantom regime at higher redshifts, while favoring the quintessence regime in the current epoch. In this context, we propose a new Gaussian-like form of EoS, termed BellDE, which avoids phantom behavior (\(w \geq -1\)) at higher redshifts while remaining precisely calibrated at lower redshifts. Interestingly, BellDE exhibits a transient phantom nature (\(w < -1\)) around the transition redshift \(z \sim 0.5\), subsequently evolving into a quintessential regime (\(w > -1\)). In particular, the BellDE model provides competitive statistical preference while offering greater flexibility in the redshift regime $z \sim 0.5-1$, where DE is observationally significant.

Figures (11)

• Figure 1: The 1D and 2D marginalized posterior distributions of the model parameters for the CPL and Padé parameterizations, with the nuisance parameter $M_B$ marginalized over. The correlation matrix is constructed by combining both datasets and marginalizing over $M_B$.
• Figure 2: The 68% and 95% confidence contours in the $w_0-w_1$, $w_1-w_2$, and $\Omega_{\rm M}-w_1$ planes are shown using DESI DR2 BAO data in combination with Planck, CC, DES, and PP data. Here, CPL_1 and CPL_2 refer to the CPL parameterization applied to the BASE+PP and BASE+DES combinations, respectively; the same notation applies to the PADE model.
• Figure 3: Redshift evolution of the EoS $w(z)$. The upper panel illustrates the CPL and PADE models for the BASE+PP dataset, while the lower panel shows the results for BASE+DESY5. Shaded bands indicate the 68% ($1\sigma$) and 95% ($2\sigma$) confidence intervals derived from observational constraints. The dashed orange and blue curves represent the means of each model, and the horizontal grey dashed line corresponds to the standard $\Lambda$CDM curve.
• Figure 4: The 1D and 2D marginalized posterior distributions of the model parameters for the Generalized Emerging DE (GEDE) parameterization, with the nuisance parameter $M_B$ marginalized over. The correlation matrix is constructed by combining both datasets and marginalizing over $M_B$.
• Figure 5: Evolution of $w(z)$ as a function of redshift $z$. The upper panel depicts the behavior of the GEDE and GDE models using the dataset BASE+PP, while the lower panel corresponds to the combination BASE+DESY5. The shaded regions indicate the 68% ($1\sigma$) and 95% ($2\sigma$) confidence level constraints derived from observational data. The horizontal dashed line in both panels represents the standard $\Lambda$CDM. An inset zoom-in highlights the low-redshift region for enhanced clarity.