How much has DESI dark energy evolved since DR1?

Abstract

DESI has reported a dynamical dark energy (DE) signal based on the $w_0 w_a$CDM model that is in conflict with Hubble tension. Recalling that the combination of DESI DR1 BAO and DR1 full-shape (FS) modeling are consistent with $Λ$CDM, in this letter we comment on the status of fluctuations in DR1 BAO documented in \cite{DESI:2024mwx, Colgain:2024xqj} in the DR2 update. In particular, we note that neither DR1 BAO nor DR2 BAO nor DR2 BAO+CMB confronted to the $w_0 w_a$CDM model with relaxed model parameter priors confirm late-time accelerated expansion today. Translating DESI BAO constraints into flat $Λ$CDM constraints, we observe that the LRG1 constraint remains the most prominent outlier, a distinction now held jointly with ELG1, LRG2 switches from smaller to larger $Ω_m$ values relative to Planck-$Λ$CDM, and ELG data drive the relatively low $Ω_m$ in the full DR2 BAO. We observe that one cannot restore $w_0 = -1$ within one $1 σ$ by removing either LRG1 or ELG1 or LRG2, but LRG2 in DR2, in contrast to LRG1 in DR1, now has the greatest bearing on $w_0 > -1$. We conclude that BAO has yet to stabilise, but the general trend is towards greater consistency with DESI DR1 FS modeling results, where there may be no dynamical DE signal in DESI data alone.

Figures (5)

• Figure 1: CPL model posteriors for DR2 BAO data subject to the DESI priors $w_0 \in [ -3, 1]$, $w_a \in [-3, 2]$ and $w_0 + w_a < 0$. The skewness in $(H_0 r_d, \Omega_m, w_0)$ posteriors comes from the bound $w_a \geq -3$.
• Figure 2: Posteriors for $w_0 w_a$CDM models and DESI BAO data subject to the priors $w_0 \in [ -10, 5]$, $w_a \in [-20, 10]$ and $w_0 + w_a < 0$.
• Figure 3: $q_0$ posteriors for the CPL posteriors in Fig. \ref{['fig:dr1_dr2_w0']}.
• Figure 4: Differences in the $\Lambda$CDM $\Omega_m$ constraints from individual tracers between DR1 in faded blue and DR2 in blue. The red and green bands denote Planck and full DESI DR2 BAO sample constraints on $\Omega_m$. In magenta we separate LRG3 and ELG1 constraints. We have displaced redshifts for visually purposes.
• Figure 5: DESI DR1 BAO constraints on the $\Lambda$CDM parameter $\Omega_m$ in faded blue relative to the DESI DR1 BAO+FS modeling constraints in blue. The green strip denotes the constraint on the full sample from BAO+FS and it can be confirmed that all constraints show excellent agreement. Ideally, BAO and FS modeling should agree on $\Omega_m$.