Comparison of dynamical dark energy with ΛCDM in light of DESI DR2

In this study presented in 2503.17342, we delve into the evolving landscape of dark energy research, building upon our prior work (2503.08658) which explored the potential for dynamical dark energy using a flexknot reconstruction method. This approach utilizes a flexible linear spline with free-moving nodes to model the dark energy equation-of-state parameter, w(a), offering a non-parametric lens for examining deviations from the standard cosmological constant model (ΛCDM). This paper examines the implications of the newly released DESI DR2 Baryon Acoustic Oscillation (BAO) data, combined with Pantheon+ (2022ApJ…938..110B) and DES5Y (2024ApJ…973L..14D) Type Ia supernovae measurements, on our understanding of dark energy dynamics.

Our lead author, A. N. Ormondroyd, along with W. J. Handley, M. P. Hobson, and A. N. Lasenby, revisit the central questions raised in our previous analysis. Does the evidence for dynamical dark energy, hinted at by a W-shaped w(a) reconstruction in our earlier study, hold up against the increased precision of DESI DR2? And do standard parameterizations like wCDM and CPL adequately capture the complexities of dark energy behavior, or do they fall short, as suggested by the nuanced features observed in our initial reconstructions?

The analysis incorporates the improved DESI DR2 BAO measurements, offering tighter constraints on cosmic expansion due to a significantly larger dataset of galaxies and quasars compared to DR1. A key advancement in DR2 is the separate reporting of D_M(z) and D_H(z) measurements for the Quasar Sample (QSO), enabled by the enhanced signal-to-noise ratio, whereas DR1 only provided a volume-averaged D_V(z) value (2025arXiv250314738D, 2024arXiv240403000D). This granularity is incorporated into our flexknot reconstruction framework (2012JCAP…06..006V, 2012JCAP…09..020V) to assess the potential of dynamical dark energy models, allowing for comparisons with the simpler wCDM (n=1 knot) and CPL (n=2 knots) scenarios. The use of the nested sampling algorithm, PolyChord (2015MNRAS.450L..61H), facilitated robust posterior sampling and evidence calculation, providing a statistical framework for model comparison.

Our findings reveal a fascinating interplay between data and model complexity. While the overall form of w(a) retains a qualitative similarity to our previous results, the statistical support for highly dynamical models (with numerous knots) diminishes when relying solely on DESI DR2. Intriguingly, the simpler wCDM and CPL models gain favor relative to ΛCDM with DESI DR2 alone, echoing findings from other DESI analyses (2025arXiv250314738D). The combination of DESI DR2 BAO data with Pantheon+ largely confirms our prior work’s conclusions. However, the inclusion of DES5Y supernovae data shifts the balance, increasing the preference for more complex flexknot models, placing wCDM and CPL on par with ΛCDM. Further analysis, including tension quantification using the log R statistic (2019PhRvD.100d3504H), sheds light on the consistency between different datasets and the potential for dynamical dark energy models to resolve discrepancies between DESI and DES5Y.

Content generated by gemini-1.5-pro using this prompt.

Image generated by imagen-3.0-generate-002 using this prompt.