Does DESI DR2 challenge $Λ$CDM paradigm ?

TL;DR

This work investigates whether DESI DR2 BAO data challenge the ΛCDM paradigm by comparing ΛCDM to a CPL-like dynamical dark energy model, $ω_0ω_a$CDM, across multiple tracers and redshifts. Using nested sampling with uniform priors, it derives BAO distance constraints and treats the sound horizon $r_d$ as a free parameter, reporting Ω_m and $h r_d$ degeneracies and their tracer-dependent behavior. The combined DR2 analysis shows modest tensions with Planck and SNe Ia for certain tracers, and, while removing the contentious LRG1/LRG2 tracers restores concordance with ΛCDM, the full data set yields only weak or inconclusive Bayesian evidence for or against the dynamical model. The results highlight the sensitivity of inferences to tracer selection and prior volume, underscoring the need for cross-tracer and multi-probe analyses to robustly assess potential deviations from ΛCDM and to validate or refute dynamical dark energy scenarios.

Abstract

Although debate on DESI DR1 systematics remains, DESI DR2 is consistent with DR1 and strengthens its trends. In our analysis, the LRG1 point at $z_{\mathrm{eff}}=0.510$ and the LRG3+ELG1 point at $z_{\mathrm{eff}}=0.934$ are in tension with the $Λ$CDM-anchored $Ω_m$ inferred from Planck and SNe Ia (Pantheon$^{+}$, Union3, DES-SN5YR): for LRG1 the tensions are $2.42σ$, $1.91σ$, $2.19σ$, and $2.99σ$; for LRG3+ELG1 they are $2.60σ$, $2.24σ$, $2.51σ$, and $2.96σ$. Across redshift bins DR2 shows improved agreement relative to DR1, with the $Ω_m$ tension dropping from $2.20σ$ to $1.84σ$. Nevertheless, DR2 alone is not decisive against $Λ$CDM, and the apparent deviation is driven mainly by LRG1 and LRG2. In a $ω_0ω_a$CDM fit using all tracers we find a posterior mean with $w_0>-1$, consistent with dynamical dark energy and nominally challenging $Λ$CDM. Removing LRG1 and/or LRG2 restores $Λ$CDM concordance ($ω_0\to-1$); moreover, $ω_0^{\mathrm{(LRG2)}}>w_0^{\mathrm{(LRG1)}}$, indicating that LRG2 drives the trend more strongly. Model selection via the natural-log Bayes factor $\ln\mathrm{BF}\equiv\ln(Z_{Λ\mathrm{CDM}}/Z_{ω_0ω_a\mathrm{CDM}})$ yields weak evidence for $Λ$CDM when LRG1, LRG2, or both are removed, and is inconclusive for the full sample. Hence the data do not require the extra $ω_a$ freedom, and the apparent $ω_0>-1$ preference should be interpreted cautiously as a reflection of the $ω_0$$ω_a$ degeneracy with limited per-tracer information.

Figures (6)

• Figure 1: The figure shows the values of $\Omega_m$ at 68% confidence intervals for different $z_{\text{eff}}$ values. The blue band represents the Planck $\Lambda$CDM prediction for $\Omega_m$.
• Figure 2: The figure shows the posterior distributions of different tracers corresponding to different $z_{\text{eff}}$ from the DESI DR2 dataset within the $\Lambda$CDM model. These are presented at 68% (1$\sigma$) and 95% (2$\sigma$) confidence levels in the $\Omega_m - h r_d$ contour plane. The blue band represents the Planck $\Lambda$CDM prediction for $\Omega_m$.
• Figure 3: The figure shows the posterior distributions at 68% (1$\sigma$) and 95% (2$\sigma$) in the $\Omega_m - h r_d$ plane using different redshift bins from the DESI DR2 compilation. The blue band represents the Planck $\Lambda$CDM prediction for $\Omega_m$.
• Figure 4: The figure shows the normalized probability distribution of $\Omega_m$ for different tracers at various effective redshifts $z_{\text{eff}}$, independent of the $h r_d$ dependence
• Figure 5: The figure shows the confidence contours at the 68% (1$\sigma$) and 95% (2$\sigma$) levels for the $\omega_0 \omega_a$CDM model, using no LRG1, no LRG2, no LRG1 & LRG2, and the full DESI DR2 sample.