🔭 astrophysics

Universal distance modes from DESI BAO and Type Ia supernovae: what do cosmological rulers actually measure?

This paper uses SVD decomposition of DESI BAO and Type Ia supernova data to demonstrate that low-redshift cosmological rulers primarily measure the parameter combination $\Omega_m h^2$ (where BAO provides tighter constraints than the CMB), revealing that the observed tension with CMB predictions is concentrated in this leading direction while extensions like dynamical dark energy show no significant tension.

Original authors: Matias Zaldarriaga

Published 2026-06-30

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Matias Zaldarriaga

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

Imagine the universe as a giant, expanding balloon. For decades, cosmologists have been trying to measure exactly how fast this balloon is inflating and what is inside it. They use two main "rulers" to take these measurements:

The Cosmic Microwave Background (CMB): This is a snapshot of the universe when it was a baby (about 380,000 years old). It tells us what the universe should look like today if our standard theory of physics (called $\Lambda$ CDM) is correct.
Low-Redshift Probes (DESI and Supernovae): These are measurements of the universe as it is now (or relatively recently).
- DESI (Baryon Acoustic Oscillations): Think of this as measuring the "fossilized sound waves" left over from the early universe, now stretched out across galaxies.
- Supernovae: These are exploding stars that act like "standard candles" (they all have the same brightness), allowing us to measure distances.

The Problem: The Rulers Don't Agree

Recently, the DESI team found a discrepancy. When they measured the universe's expansion using these "now" rulers, the results didn't quite match the predictions from the "baby" snapshot. This has led some scientists to wonder: Is our theory of Dark Energy wrong? Is the universe expanding in a weird, new way?

The Paper's Solution: The "Universal Ruler"

Matias Zaldarriaga's paper uses a mathematical tool called Singular Value Decomposition (SVD). To understand this, imagine you have a complex, multi-colored knot of data. You want to know: What is the single, most important direction this knot is pulling in?

The author untangled the data from DESI and the supernovae to find the "leading direction" of the tension. Here is what he found:

1. The "One-Dimensional" Universe

The most surprising finding is that all these different measurements (DESI, Supernovae, CMB) are actually only measuring one thing effectively.

The Analogy: Imagine you are trying to describe a 3D object, but you are only allowed to look at it from one specific angle. No matter how you rotate the object, you only see a 2D shadow.
The Result: The "shadow" these low-redshift measurements cast is almost entirely a measurement of $\Omega_m h^2$ . In plain English, this is a specific combination of how much matter is in the universe and how fast it is expanding.
The Tension: The DESI data says this number is about 2.2 standard deviations lower than what the CMB predicts. The supernovae data is a bit fuzzy and doesn't have enough precision to confirm or deny this, but they don't strongly disagree either.

2. The "Dark Energy" Red Herring

Because the DESI data showed a mismatch, many scientists thought, "Ah, this must mean Dark Energy is changing over time!" (This is the $w_0$ – $w_a$ extension).

The paper argues this is a misinterpretation.

The Analogy: Imagine you are trying to tune a radio. You hear static and think the station is changing its song (Dark Energy). But actually, the static is just because the antenna (the matter density measurement) is slightly bent.
The Finding: When the author separates the "matter" signal from the "Dark Energy" signal, the "Dark Energy" part is perfectly calm. It shows no significant tension. The universe looks exactly like it should if Dark Energy is a constant (a Cosmological Constant).
Conclusion: The "preference" for evolving Dark Energy that DESI reported is mostly just the "matter density" tension dressed up in Dark Energy clothing.

3. The Curvature Exception

The paper checked if the universe might be curved (like a saddle or a sphere) rather than flat.

The Finding: There is a tiny hint that the universe might be slightly "open" (curved), but the data is very weak here. It's a marginal signal that only DESI can see, not the supernovae.

4. The Real Culprit: The "Lensing" Glitch

If the tension is real, what causes it? The paper suggests the problem might not be with the universe, but with how we read the "baby snapshot" (the CMB).

The Analogy: Imagine taking a photo of a landscape through a slightly warped window. If you don't account for the warp, your measurements of the trees will be wrong.
The Finding: The CMB measurement of matter density relies heavily on how much the universe's gravity "lenses" (bends) light. If there is a small systematic error in how we measure this "lensing" (or the optical depth $\tau$ , which is related to when the first stars turned on), it shifts the predicted value of matter density.
The Result: If you adjust for this lensing uncertainty, the tension between the "baby" snapshot and the "adult" measurements disappears (dropping from a 2.2 $\sigma$ tension to a 1.1 $\sigma$ one).

Summary in Plain English

We are mostly measuring one thing: All the new data is essentially telling us about the density of matter in the universe.
Dark Energy is fine: The "weird" behavior DESI saw isn't because Dark Energy is changing; it's because the matter density measurement is slightly off compared to the CMB prediction.
The tension is likely a measurement error: The mismatch is probably caused by a subtle uncertainty in how we interpret the early universe's "lensing" effects, not by new physics.
The "Universal Ruler": The author created a method to strip away the noise and show that, fundamentally, our current data is one-dimensional. We need better data to see if there are other dimensions (other new physics) hiding in the noise.

The Bottom Line: The universe isn't necessarily breaking the laws of physics. Instead, we might just be slightly misreading the ruler we used to measure the baby universe.

Technical Summary: Universal Distance Modes from DESI BAO and Type Ia Supernovae

Problem Statement
Recent analyses combining DESI Baryon Acoustic Oscillation (BAO) data with Type Ia supernovae (SNe Ia) and Cosmic Microwave Background (CMB) observations have reported a $\sim 3\sigma$ preference for evolving dark energy within the $w_0$ – $w_a$ parameterization. The central question addressed by this work is the origin of this preference: does it arise from a genuine signal of dynamical dark energy, or does it stem from a discrepancy in a fundamental cosmological parameter? The authors aim to identify the specific linear directions in observable space probed by low-redshift distance data and to localize the tension with $\Lambda$ CDM predictions anchored by CMB data.

Methodology
The paper employs a Singular Value Decomposition (SVD) framework applied to low-redshift distance measurements. The methodology proceeds as follows:

Data: The analysis utilizes DESI DR2 BAO measurements (13 observables across $0 < z < 2.3$ ) and three SNe Ia compilations (Union3, Pantheon+, and the recalibrated DES-Dovekie). These are compared against CMB chains from ACT DR6 and Planck 2018.
Whitening and Projection: The authors construct a matrix of whitened chain predictions (CMB posterior samples) against the observables. The data covariance is used to whiten the space, ensuring all directions have unit measurement variance.
SVD Decomposition: The matrix is decomposed to identify the leading singular vectors ( $V_\alpha$ ). The leading direction, $V_0$ , represents the mode along which the CMB posterior allows the most variation in low-redshift distance predictions.
Amplitude Measurement: The data is projected onto these predetermined directions to obtain amplitudes $c_\alpha$ . Tensions are calculated as the difference between the data amplitude and the chain mean, normalized by the combined uncertainty of the measurement error and the model variance.
Extension Analysis: The framework is extended to beyond- $\Lambda$ CDM models (dynamical dark energy $w_0$ – $w_a$ , spatial curvature $\Omega_k$ , primordial magnetic fields, early dark energy, and lensing rescaling $A_{\text{lens}}$ ) to determine if these extensions open genuinely new measurable directions orthogonal to $V_0$ .

Key Contributions and Results

Universality of the Leading Mode ( $V_0$ ): The analysis reveals that for all probes (BAO and all three SN compilations), the leading measurable direction is effectively identical. This direction, $V_0$ , is a near-perfect proxy for the parameter combination $\Omega_m h^2$ . The cross-probe inner products for this direction exceed 0.996.
- BAO Constraint: BAO constrains this direction more tightly than the CMB itself ( $\sigma_{c0} = 1.56 > 1$ ), providing genuinely new information.
- SN Constraint: The SN compilations do not constrain this direction more tightly than the CMB prior ( $\sigma_{c0} \approx 0.3 < 1$ ); thus, they add little independent information regarding $\Omega_m h^2$ .
Localization of Tension: The primary tension between low-redshift data and the CMB-anchored $\Lambda$ CDM model resides entirely in the $V_0$ ( $\Omega_m h^2$ ) direction.
- BAO: Shows a $+2.2\sigma$ tension, preferring a lower $\Omega_m h^2$ (and thus higher $H_0$ ) than the CMB.
- SNe: The three SN compilations show mild tensions in the opposite direction ( $-0.8\sigma$ to $-1.7\sigma$ ) but lack the constraining power to confirm or refute the BAO signal.
Reinterpretation of Dark Energy Preference: When extending the analysis to the $w_0$ – $w_a$ model:
- The leading measurable direction remains $V_0$ (the $\Omega_m h^2$ mode).
- A second direction, $V_1$ , becomes measurable, representing the independent dark energy degree of freedom.
- Result: The BAO data shows only a $+1.2\sigma$ tension in the dark energy direction ( $V_1$ ), which is consistent with a cosmological constant ( $w = -1$ ). The authors conclude that the reported $\sim 3\sigma$ preference for evolving dark energy in joint analyses is primarily driven by the $\Omega_m h^2$ tension ( $c_0$ ) rather than a signal in the dark energy equation of state.
Spatial Curvature: Spatial curvature ( $\Omega_k$ ) is the only other extension that opens a genuinely new measurable direction for BAO (though marginally). Both measurable directions ( $c_0$ and the residual $c_1$ ) independently prefer positive spatial curvature ( $\Omega_k \approx +0.005$ ), providing an internal consistency check, though the second direction is poorly constrained.
CMB-Side Extensions: Extensions such as primordial magnetic fields (BPMF), early dark energy (EDE), and lensing rescaling ( $A_{\text{lens}}$ ) do not open new measurable directions. However, varying $A_{\text{lens}}$ (or equivalently the reionization optical depth $\tau$ ) broadens and shifts the CMB posterior for $\Omega_m h^2$ , reducing the BAO tension from $2.2\sigma$ to $1.1\sigma$ . This suggests the tension may be resolved by systematic shifts in CMB-derived parameters rather than new low-redshift physics.

Significance and Claims
The paper claims that the "universal distance modes" analysis provides a model-independent decomposition of low-redshift data, demonstrating that current distance probes are effectively one-dimensional in their cosmological information content, dominated by $\Omega_m h^2$ .

The authors assert that the apparent evidence for dynamical dark energy in DESI results is largely an artifact of the $\Omega_m h^2$ discrepancy between CMB and BAO data, rather than a direct detection of evolving dark energy. The independent dark energy direction ( $V_1$ ) remains consistent with a cosmological constant. The work highlights that the tension is sensitive to CMB-side systematics (specifically lensing amplitude and optical depth) and that the low-redshift data alone cannot distinguish between various beyond- $\Lambda$ CDM extensions that do not alter the $\Omega_m h^2$ degeneracy.

The paper is presented as a draft utilizing AI tools for calculation and drafting, with full reproducibility provided via a public repository, and explicitly states it makes no claims of originality or novelty beyond the application of this specific SVD framework to the DESI data.