🔭 astrophysics

Present Day Cosmic Acceleration from SDSS and DESI BAO: A Call for Finer Tomography of the DESI Bright Galaxy Survey

This paper argues that the apparent preference for a non-accelerating present epoch in DESI BAO data, which contrasts with SDSS results, stems from DESI's higher effective redshift limit requiring extrapolation of dark energy models, and suggests that finer tomographic binning of the Bright Galaxy Survey to access lower redshifts is necessary to resolve this discrepancy.

Original authors: Anna Chiara Ferri, Ruchika, Alessandro Melchiorri

Published 2026-07-09

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Anna Chiara Ferri, Ruchika, Alessandro Melchiorri

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

The Big Picture: A Cosmic Speedometer Dispute

Imagine the universe as a giant car driving down a highway. For decades, astronomers have been trying to figure out if this car is speeding up (accelerating) or slowing down (decelerating). The "engine" pushing it is something called Dark Energy.

Recently, a massive new telescope survey called DESI (Dark Energy Spectroscopic Instrument) started measuring the universe's expansion with incredible precision. When scientists combined DESI's data with data from the Planck satellite (which maps the early universe), they got a strange result: the data suggested the universe might be slowing down right now, or at least, they couldn't prove it was speeding up.

This paper, written by Anna Chiara Ferri, Ruchika, and Alessandro Melchiorri, investigates why DESI's results look different from older surveys (like SDSS) and what it actually means for the fate of the universe.

The Core Conflict: Two Different Maps

The authors compared two major "maps" of the universe:

The Old Map (SDSS): This survey measured galaxies at various distances, including some very close to us (low redshift).
The New Map (DESI): This survey is much bigger and more precise, but its "closest" measurements start a bit further away than SDSS's closest measurements.

The Finding:

When you combine the Old Map (SDSS) with the Planck data, the result is clear: The universe is definitely speeding up (accelerating).
When you combine the New Map (DESI) with the Planck data, the result is confusing: The math suggests the universe might be slowing down, or at least, it's impossible to tell if it's speeding up or not.

The Analogy: The "Missing Anchor"

Why does the new, better map give a weirder result? The authors argue it's not because the physics of the universe changed, but because of where the measurements start.

Imagine you are trying to guess the speed of a car at the exact moment it passes a finish line ( $z=0$ ).

SDSS has a camera that takes a picture of the car just 10 meters before the finish line. They can see exactly how fast it is going right now.
DESI has a camera that takes a picture of the car 30 meters before the finish line. They don't have a picture closer than that.

To figure out the speed at the finish line, the DESI team has to guess (extrapolate) what happened between the 30-meter mark and the finish line. Because they are guessing, their math allows for a scenario where the car might have slowed down slightly right at the end.

The SDSS team, having a picture closer to the finish, doesn't have to guess as much. They see the car is still speeding up.

The "Equation of State" (The Engine's Setting)

In cosmology, scientists describe Dark Energy with a number called $w_0$ .

If $w_0$ is close to -1, the universe accelerates (like our current understanding).
If $w_0$ gets closer to 0, the universe might stop accelerating.

The paper shows that because DESI lacks the "close-up" data (the 10-meter mark), their math lets $w_0$ drift toward 0. This drift pushes the "deceleration parameter" ( $q_0$ ) into positive numbers, which mathematically means "slowing down."

The Twist: When the authors added Supernova data (Type Ia supernovae, which act like extra "close-up" cameras) to the DESI data, the result snapped back. The universe was confirmed to be accelerating again, and the uncertainty shrank dramatically. This proves that DESI's "slowing down" result was just a gap in the data, not a new law of physics.

The "What If" Experiment

To prove their theory, the authors did a clever experiment:

They took the SDSS data and removed the closest galaxies (the ones at the 10-meter mark).
Suddenly, SDSS's result started looking like DESI's result! The confidence in "acceleration" dropped, and the numbers drifted toward "slowing down."

This confirmed that the difference wasn't about the type of telescope or the quality of the data, but simply about how close to "now" the data reaches.

The Conclusion: A Call for Finer Slicing

The paper concludes that the apparent "non-acceleration" seen in DESI data is likely an illusion caused by a lack of data very close to us. It's not a sign of new physics; it's a sign that we need better "slices" of the data.

The Recommendation:
The authors suggest that the DESI team should re-analyze their "Bright Galaxy Survey" by splitting the data into finer bins (smaller groups). Instead of grouping all galaxies between redshift 0.1 and 0.4 together, they should split them into smaller groups (e.g., 0.1–0.25 and 0.25–0.4).

This would create a new "anchor" point closer to us (around redshift 0.18), similar to the old SDSS measurements. If they do this, they expect the DESI data to finally agree with the SDSS data and confirm that the universe is indeed speeding up.

Summary in One Sentence

The paper argues that DESI's confusing suggestion that the universe might be slowing down is just a result of missing data from the very recent past, and that by looking at the data in smaller, closer groups, we will likely confirm that the universe is still speeding up as expected.

Technical Summary: Present Day Cosmic Acceleration from SDSS and DESI BAO

Problem Statement
Recent analyses combining Dark Energy Spectroscopic Instrument (DESI) Data Release 2 (DR2) Baryon Acoustic Oscillation (BAO) measurements with Cosmic Microwave Background (CMB) data have suggested a preference for time-evolving dark energy, potentially indicating a present-day epoch that is not accelerating. This contrasts with the standard $\Lambda$ CDM model and earlier results from the Sloan Digital Sky Survey (SDSS), which consistently support an accelerating universe. The central problem addressed by this paper is the origin of the discrepancy between the DESI and SDSS inferences regarding the present-day deceleration parameter ( $q_0$ ) and the dark energy equation of state ( $w_0$ ). Specifically, the authors investigate whether this tension arises from new physics or from differences in the redshift coverage of the two surveys.

Methodology
The authors perform a comparative analysis of late-time expansion histories using four cosmological models: $w$ CDM, $\nu$ CDM ( $\Lambda$ CDM + massive neutrinos), $w$ CDM + $\sum m_\nu$ , and the Chevallier–Polarski–Linder (CPL) parametrization ( $w_0$ - $w_a$ ). The study utilizes:

DESI DR2 (Year-3) BAO chains: Combined with Planck 2018 CMB likelihoods (NPIPE CamSpec high- $\ell$ and joint Planck+ACT DR6 lensing).
SDSS DR16 BAO/RSD chains: Combined with Planck CMB likelihoods (2018 Plik lite and 2018 lensing reconstruction).
Supernova Data: The Pantheon+ compilation is added to the DESI combination to test the impact of low-redshift anchors.

The analysis focuses on the dimensionless deceleration parameter, $q_0$ , derived from the CPL parameters $w_0$ and $\Omega_{DE,0}$ via the relation $q_0 = \frac{1}{2}\Omega_{m,0} + \frac{1}{2}(1+3w_0)\Omega_{DE,0}$ . The authors explicitly compare the effective redshift ( $z_{eff}$ ) of the lowest redshift bins in each survey: $z_{eff} \approx 0.295$ for the DESI Bright Galaxy Survey (BGS) versus $z_{eff} \approx 0.15$ for the SDSS Main Galaxy Sample (MGS). To isolate the effect of redshift coverage, the authors also analyze a modified SDSS dataset with the low-redshift MGS anchor removed.

Key Results

Divergence in $w_0$ and $q_0$ :
- Planck + SDSS: Favors $w_0 = -0.71^{+0.19}_{-0.18}$ and a negative deceleration parameter $q_0 = -0.22^{+0.20}_{-0.21}$ , consistent with an accelerating universe.
- Planck + DESI: Favors a significantly higher $w_0 = -0.41^{+0.21}_{-0.22}$ and a positive median deceleration parameter $q_0 = +0.10^{+0.21}_{-0.23}$ . The median posterior lies on the decelerating side of the $q_0=0$ boundary, though the 1 $\sigma$ uncertainty straddles the transition.
- Planck + DESI + Pantheon+: The inclusion of Type Ia supernovae, which provide low-redshift distance constraints, shifts the DESI results to $w_0 = -0.84^{+0.06}_{-0.05}$ and $q_0 = -0.37 \pm 0.06$ , restoring a clear signature of acceleration and reducing tension with $\Lambda$ CDM.
Redshift Coverage as the Driver:
The authors demonstrate that the discrepancy is driven by the lowest effective redshift probed. $w_0$ and $q_0$ are present-day quantities sensitive to the lowest available data points.
- SDSS, anchored at $z_{eff} \approx 0.15$ , constrains $w_0$ and $q_0$ directly with data.
- DESI, limited to $z_{eff} \approx 0.295$ , must extrapolate the CPL parametrization down to $z=0$ . This extrapolation allows the parameters to drift toward configurations where $w_0$ is less negative and $q_0$ is positive, driven by the $\Omega_{DE}$ – $w_0$ degeneracy.
Validation Tests:
- Removing the low-redshift MGS anchor from the SDSS dataset ( $z_{eff}$ rising to $\approx 0.38$ ) causes the SDSS constraints to weaken and drift toward the DESI values ( $q_0$ shifts from $-0.22$ to $-0.10$), confirming that the loss of low-redshift leverage, rather than intrinsic survey differences, causes the shift.
- In single-parameter extensions ( $w$ CDM, $\nu$ CDM), the results from both surveys remain consistent and indicate acceleration, highlighting that the tension is specific to the dynamical dark energy (CPL) model where $w_0$ is free to vary.

Significance and Claims
The paper argues that the apparent preference of the DESI BAO+CMB combination for a non-accelerating present epoch is likely an artifact of redshift sampling rather than evidence for new physics. The authors claim that without a direct distance anchor below $z \approx 0.3$ , the CPL parametrization cannot robustly constrain present-day acceleration, leading to a "fragile" inference where the median posterior drifts into the decelerating regime due to extrapolation.

The authors conclude that the DESI preference for dynamical dark energy (specifically the $w_0$ - $w_a$ tension) is distinct from the $q_0$ inference; the former is driven by intermediate-redshift bins and survives the inclusion of supernovae, while the latter is driven by the lack of low-redshift data. Consequently, the paper calls for future DESI analyses to adopt finer tomographic binning of the Bright Galaxy Survey (BGS) to access lower mean redshifts (approaching $z_{eff} \sim 0.18$ ), thereby testing whether the inference of $q_0$ shifts back toward the accelerating $\Lambda$ CDM expectation when low-redshift leverage is restored.