[Paper Review] Cosmological constraints on dark energy models using DESI BAO 2024

Jie Zheng, Da-Chun Qiang+1 more|arXiv (Cornell University)|Dec 6, 2024

Solar and Space Plasma Dynamics4 citations

TL;DR

This study investigates whether the observed tension in DESI BAO data with the $Λ$CDM model arises from the parameterization of dark energy's equation of state. Using four dynamical dark energy models—CPL, BA, JBP, and FSLL—combined with SNe Ia and quasar data, it finds that the LRG1 and LRG2 data points drive the deviation from $Λ$CDM, and that the BA and FSLL models are statistically favored by AIC and BIC, suggesting potential dynamic dark energy properties.

ABSTRACT

Recently, the measurements of baryon acoustic oscillations (BAO) by the Dark Energy Spectroscopic Instrument (DESI) indicate a potential deviation from the standard $Λ$CDM model. Some studies suggest that the data points from the luminous red galaxies (LRG) survey in DESI BAO data may contribute to this discrepancy. In this work, our main goal is to investigate whether this deviation is caused by the parameterization of the equation of state (EoS) of dark energy (DE). Hence, we have examined four popular parameterized dark energy models in our analysis: the Chevallier-Polarski-Linder (CPL), Barboza-Alcaniz (BA), Jassal-Bagla-Padmanabhan (JBP), and Feng-Shen-Li-Li (FSLL) parameterizations. Considering that LRG1 and LRG2 data points may lead to deviation from the $Λ$CDM model, we use two versions of DESI BAO data, differing in whether these data points are included. Additionally, to break the parameter degeneracies and obtain robust constraint results, we introduce Type Ia supernovae (SNe Ia) and quasars (QSO) in our analysis. Our findings indicate that in these parameterizations, the deviation from ($w_0$,$w_{1}$)=(-1,0) becomes more pronounced when using the combined data from DESI BAO, SNe Ia, and QSO compilations. Here, $w_{0}$ and $w_{1}$ represent the EoS of DE. It suggests that the parameterizations of the EoS of DE have little impact on the deviation from the $Λ$CDM model. Besides, our analysis potentially hints that dark energy may have dynamic properties. In addition, the results obtained from different BAO datasets demonstrate that the LRG1 and LRG2 data points do indeed contribute to a deviation from the $Λ$CDM model. Finally, according to the statistical criteria, the Akaike Information Criterion (AIC) and the Bayesian Information Criterion (BIC), the joint constraints provide substantial observational support to the BA and FSLL models.

Motivation & Objective

• To determine whether the parameterization of dark energy's equation of state (EoS) contributes to the observed tension between DESI BAO data and the $Λ$CDM model.
• To assess the impact of LRG1 and LRG2 data points on cosmological constraints in dynamical dark energy models.
• To evaluate the performance of four EoS parameterizations—CPL, BA, JBP, and FSLL—using joint constraints from DESI BAO, SNe Ia, and quasars.
• To test whether the inclusion of late-time probes (SNe Ia and QSO) breaks parameter degeneracies and improves cosmological constraints.

Proposed method

• Four dark energy EoS parameterizations—CPL, BA, JBP, and FSLL—are used to model time-varying dark energy.
• Two versions of DESI BAO data are analyzed: one including LRG1 and LRG2, and one excluding them, to isolate their impact.
• Late-universe probes—Type Ia supernovae (SNe Ia) and quasars (QSO)—are combined with BAO data to break parameter degeneracies and extend redshift coverage.
• Cosmological constraints are derived using Markov Chain Monte Carlo (MCMC) methods, with model comparison via AIC and BIC.
• The analysis is performed under both flat and non-flat spatial curvature assumptions to test robustness.
• Statistical criteria (AIC and BIC) are used to rank competing dark energy models against $Λ$CDM.

Experimental results

Research questions

• RQ1 Does the choice of dark energy equation of state parameterization affect the tension between DESI BAO data and the $Λ$CDM model?
• RQ2 To what extent do the LRG1 and LRG2 data points in DESI BAO contribute to the observed deviation from $Λ$CDM?
• RQ3 How do joint constraints from DESI BAO, SNe Ia, and quasars affect the statistical preference for dynamical dark energy models?
• RQ4 Which dark energy model—CPL, BA, JBP, or FSLL—is most favored by current observational data, as measured by AIC and BIC?

Key findings

• The inclusion of LRG1 and LRG2 data points significantly increases the deviation from the $Λ$CDM model, with all models showing 1–2$σ$ tension except the flat JBP model.
• When combining DESI BAO, SNe Ia, and QSO data, all models show 1$σ$ inconsistency with $Λ$CDM, except non-flat BA and non-flat FSLL models.
• The LRG1 and LRG2 data points are identified as the primary drivers of the observed tension with $Λ$CDM, as excluding them reduces the tension significantly.
• The Akaike Information Criterion (AIC) and Bayesian Information Criterion (BIC) favor the Barboza-Alcaniz (BA) and Feng-Shen-Li-Li (FSLL) models over $Λ$CDM and other parameterizations, especially when LRG1 and LRG2 are included.
• The flat JBP model is substantially penalized by the data, indicating it is less consistent with current observations than other models.
• The joint use of SNe Ia and QSO data improves parameter degeneracy breaking and enhances $H_0$ precision to approximately 2%, though a tension with SH0ES remains while agreeing with Planck CMB results.