Does DESI 2024 Confirm $Λ$CDM?

TL;DR

This work probes whether DESI 2024 BAO constraints align with flat $Λ$CDM or imply late-time evolution of cosmological parameters. It shows that the $z_{ extrm{eff}}=0.51$ LRG bin favors a high $Ω_m$ around $0.65$–$0.67$, conflicting with Planck’s $Ω_m obreak ightarrow 0.315$ and driving a DESI-alone hint for $w_0 > -1$ in the $w_0w_a$CDM framework. An independent, prior-free ratio method using $(D_M/r_d)/(D_H/r_d)$ confirms the anomaly across tracers with a ~2.2σ tension relative to Planck. The authors argue the signal may reflect statistical fluctuation or systematics and stress the importance of multi-probe consistency; subsequent DESI DR1/DR2 results are more consistent with constant $Ω_m$, highlighting the need for continued cross-checks before invoking new late-Universe physics.

Abstract

We demonstrate that a $\sim 2 σ$ discrepancy with the Planck-$Λ$CDM cosmology in DESI Luminous Red Galaxy (LRG) data in the redshift range $0.4 < z < 0.6$ with effective redshift $z_{\textrm{eff}} = 0.51$ translates into an unexpectedly large $Ω_m$ value, $Ω_m = 0.67^{+0.18}_{-0.17}$. We independently confirm that this anomaly drives the preference for $w_0 > -1$ in DESI data \textit{alone} confronted to the $w_0 w_a$CDM model. Given that LRG data at $z_{\textrm{eff}} = 0.51$ is at odds with Type Ia supernovae in overlapping redshifts, we expect that this anomaly will decrease in statistical significance with future DESI data releases leaving an increasing $Ω_m$ trend with effective redshift at higher redshifts. We estimate the current significance of the latter in DESI data at $\sim 1.8 σ$ and comment on how it dovetails with independent observations. It is imperative to understand what makes DESI LRG data at $z_{\textrm{eff}} = 0.51$ an outlier when it comes to $Ω_m$ determinations.

Figures (3)

• Figure 1: $68 \%$ credible intervals for $\Omega_m$ versus effective redshift $z_{\textrm{eff}}$. Neglecting the DESI LRG constraint at $z_{\textrm{eff}} = 0.51$, the data shows good agreement with the corresponding Planck-$\Lambda$CDM confidence interval in red. The lower $\Omega_m$ in the full DESI dataset is primarily driven by $z_{\textrm{eff}} = 0.71$ LRG.
• Figure 2: Distribution of 50,000 $\Omega_m$ values at $z_{\textrm{eff}}=0.51$ alongside Planck constraints $\Omega_m = 0.315 \pm 0.007$ in red. The probability of finding an $\Omega_m$ value within the Planck $1 \sigma$ confidence interval or lower is $p = 0.015$ corresponding to $2.2 \sigma.$
• Figure 3: $w_0w_a$CDM model with and without LRG data at $z_{\textrm{eff}} = 0.51$. Any hint of dynamical DE, in particular $w_0 > -1$, in DESI data is driven by LRG data. We acknowledge use of GetDistLewis:2019xzd.