Persistent and serious challenge to the $\Lambda$CDM throne: Evidence for dynamical dark energy rising from combinations of different types of datasets

Abstract

We derive multiple constraints on dark energy and compare dynamical dark energy models with a time-varying equation of state ($w_0 w_a$CDM) versus a cosmological constant model ($\Lambda$CDM). We use Baryon Acoustic Oscillation (BAO) from DESI and DES, Cosmic Microwave Background from Planck with and without lensing from Planck and ACT (noted CMBL and CMB, respectively), supernova (SN), and cross-correlations between galaxy positions and galaxy lensing from DES. First, we use pairs or trios of datasets where we exclude one type of dataset each time and categorize them as ``NO SN", ``NO CMB" and ``NO BAO" combinations. In all cases, we find that the combinations favor the $w_0 w_a$CDM model over $\Lambda$CDM, with significance ranging from 2.3$\sigma$ to 3.3$\sigma$. For example, DESI+DESY6BAO+CMB yields 3.2$\sigma$ without SN, DESI+DESY6BAO+DESY5SN yields 3.3$\sigma$ without CMB, and CMB+DESY5SN+DES3x2pts yields 2.6$\sigma$ without BAO. The persistence of this pattern across various dataset combinations even when any of the datasets is excluded supports an overall validation of this trending result regardless of any specific dataset. Next, we use larger combinations of these datasets after verifying their mutual consistency within the $w_0 w_a$CDM model. We find combinations that give significance levels $\sim$4$\sigma$, with DESI+DESY6BAO+CMBL+DESY5SN reaching 4.4$\sigma$. In sum, while we need to remain prudent, the combination of the first step that supports a validation of the pattern of these results beyond any single type of dataset and their associated systematics, together with the second step showing high-significance results when such datasets are combined, presents a compelling overall portrait in favor of a dynamical dark energy with a time-evolving equation of state over a cosmological constant, and constitutes a serious challenge to the $\Lambda$CDM model's reign. Abridged.