⚛️ general relativity

Model-Independent Dark Energy Measurements from DESI DR2 and Planck 2015 Data

This paper presents model-independent measurements of dark energy density and equation of state using DESI DR2 and Planck 2015 data, finding consistency with a cosmological constant and suggesting that the previously reported 3.1 $\sigma$ deviation by the DESI Collaboration may be an artifact of assuming a linear parametrization for the equation of state.

Original authors: Yun Wang, Katherine Freese

Published 2026-02-17

📖 6 min read🧠 Deep dive

CC BY 4.0

Original authors: Yun Wang, Katherine Freese

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 Mystery: What is Pushing the Universe Apart?

Imagine the universe is a giant balloon. About 27 years ago, astronomers discovered something shocking: this balloon isn't just expanding; it's speeding up. Something invisible is pushing the galaxies apart faster and faster. Scientists call this mysterious force "Dark Energy."

For decades, the leading theory has been that this force is a "Cosmological Constant"—a fancy way of saying it's a steady, unchanging energy that has been the same since the beginning of time. It's like a constant wind blowing on the balloon.

But recently, the DESI Collaboration (a massive team using a powerful telescope) looked at new data and claimed: "Wait a minute! The wind isn't steady. It's changing! We found evidence that Dark Energy is evolving over time." They said this with high confidence (3.1 sigma), which in science is usually enough to say, "We might have found something new."

The New Study: "Let's Not Assume Anything"

Authors Yun Wang and Katherine Freese looked at the same data but decided to take a different approach. They asked: "What if we stop guessing the shape of the wind?"

In the DESI study, they assumed the wind followed a specific, simple pattern (a straight line). It's like trying to describe a complex, swirling storm by only drawing a straight line through it. If the storm doesn't fit the line, you might think the storm is changing, when really, you just picked the wrong tool to measure it.

Wang and Freese decided to use a model-independent approach. Instead of forcing the data into a straight line, they treated Dark Energy like a free-form sculpture. They asked the data: "What is the density of Dark Energy at different points in time?" without forcing it to fit a specific mathematical formula.

The Analogy: The Hiking Trail

Imagine you are hiking a trail (the history of the universe) and you want to know how steep the path is (the strength of Dark Energy).

The DESI Approach: They assumed the trail was a straight ramp. When they measured the height at a few points, the points didn't line up perfectly with the ramp. They concluded, "The ramp must be changing shape!"
The Wang & Freese Approach: They didn't assume it was a ramp. They just measured the height at specific checkpoints (redshifts) and connected the dots with a smooth, flexible ruler (a cubic spline). They let the data draw the shape of the trail itself.

What Did They Find?

When Wang and Freese used their flexible ruler, the picture changed completely:

It's Mostly Steady: When they measured the Dark Energy Density (how much "stuff" is pushing the balloon), it looked almost exactly like the steady Cosmological Constant theory. The only slight wobble was about 1 standard deviation (1σ). In science, this is like a coin flip; it's not strong evidence of change. It's just normal statistical noise.
The "Equation of State" is Messy: They also tried to measure the "Equation of State" (a property called w that describes how the energy behaves). This is much harder to pin down. Their results showed a 2σ deviation. This is still considered "maybe," not "definitely."
The Trap of Assumptions: They found that the DESI team's "3.1σ" result was largely an artifact of assuming the energy followed a straight line. By forcing the data into that straight line, they created an illusion of change.

The Verdict: The universe's expansion is likely still being driven by a steady, unchanging force (the Cosmological Constant). The "evidence" for a changing Dark Energy disappears when you stop forcing the data into a specific shape.

Why Measuring "Density" is Better Than Measuring "Equation of State"

The paper makes a crucial point about how we measure things.

Measuring Density (ρ): This is like measuring the amount of water in a bucket. It's direct.
Measuring Equation of State (w): This is like trying to guess the pressure of the water by looking at how the bucket deforms. It's an indirect calculation.

The authors show that trying to measure the "pressure" (w) directly from the data is like trying to hear a whisper in a hurricane. The data is too noisy. It's much more reliable to measure the "amount of water" (density) directly. When you do that, the mystery of a changing Dark Energy largely vanishes.

The "Data Gap" and Future Missions

There is a problem with our current map. We have good data for the "recent" universe (low redshift) and the "very early" universe, but there is a gap in the middle (between redshift 1.5 and 2.33).

The Current Situation: It's like trying to guess the plot of a movie by only watching the first 10 minutes and the last 10 minutes. You might think the middle is crazy, but you don't actually know.
The Future: New telescopes like Euclid and Roman are coming soon. They will fill in that missing middle section of the movie. They will give us a complete picture of the "hiking trail" so we can finally know if the wind is steady or if it's actually changing.

Summary for the Everyday Reader

The Claim: A recent study said Dark Energy is changing over time.
The Counter-Claim: This new paper says, "That conclusion comes from assuming Dark Energy follows a simple straight line. If we let the data speak for itself, it looks like Dark Energy is actually steady and constant."
The Lesson: In science, how you ask the question matters. If you assume the answer is a straight line, you might find a "wobble" that isn't really there.
The Future: We need more data from the middle of the universe's history to be 100% sure. Until then, the simplest explanation (a steady, unchanging Dark Energy) remains the best fit.

In short: The universe might not be changing its mind after all; we might just have been using the wrong ruler to measure it.

1. Problem Statement

Twenty-seven years after the discovery of cosmic acceleration, the physical nature of dark energy remains unknown. A critical challenge in cosmology is distinguishing between a true cosmological constant ( $\Lambda$ ) and dynamical dark energy (where the equation of state $w_X$ varies with time) without introducing model bias.

The DESI Collaboration recently reported a 3.1 $\sigma$ deviation from a cosmological constant using DESI DR2 Baryon Acoustic Oscillation (BAO) data combined with Planck CMB data. However, this result relied on a specific parametric assumption: the linear equation of state $w_X(z) = w_0 + w_a(1-a)$ . The authors argue that this linear parametrization may be misleading, particularly at high redshifts, and that assuming a specific functional form for $w_X(z)$ can artificially inflate the statistical significance of deviations from $\Lambda$ CDM. The paper aims to re-analyze the same datasets using a model-independent approach to determine if the evidence for time-varying dark energy holds when $w_X(z)$ is not constrained to a linear form.

2. Methodology

The authors employ a non-parametric, model-independent framework to measure dark energy properties directly from observational data.

Data Sources:
- DESI DR2: BAO distance measurements ( $D_V/r_d$ , $D_M/r_d$ , $D_H/r_d$ ) at redshifts $z = 0.295, 0.51, 0.706, 0.934, 1.321, 1.484,$ and $2.33$.
- Planck 2015/2018: Cosmic Microwave Background (CMB) distance priors (shift parameters $R$ and $l_a$ ) to break degeneracies in cosmological parameters.
- Assumptions: A flat Universe is assumed (consistent with DESI analysis). Neutrino parameters are fixed or varied within standard ranges to test robustness.
Core Approach:
Instead of fitting parameters $w_0$ and $w_a$ , the authors treat the dark energy density ratio $X(z) \equiv \rho_X(z)/\rho_X(0)$ and the equation of state $w_X(z)$ as free functions of redshift.
- Discretization: These functions are defined by their values at a set of discrete redshift nodes $\{z_i\} = \{0, 1/3, 2/3, 1, 4/3, 2.33\}$ .
- Interpolation: Cubic splines are used to smoothly interpolate between these nodes.
- Boundary Conditions: For $z > 2.33$ (where no data exists), $X(z)$ and $w_X(z)$ are held constant at their values at $z_{max} = 2.33$ . This avoids making unjustified assumptions about dark energy behavior in data voids.
- Relation: $X(z)$ and $w_X(z)$ are related via the integral:
  $X(z) = \exp\left[ \int_0^z \frac{1+w_X(z')}{1+z'} dz' \right]$
  This highlights that $X(z)$ is the more fundamental quantity, while $w_X(z)$ is derived and thus more sensitive to noise and priors.
Model Selection:
The authors use the Akaike Information Criterion (AIC) and Bayesian Information Criterion (BIC) to compare models with different redshift node choices and to determine whether measuring $X(z)$ or $w_X(z)$ is statistically preferred.

3. Key Contributions

Direct Measurement of $\rho_X(z)$ : The paper demonstrates that measuring dark energy density as a free function is more direct and robust than inferring it through the equation of state $w_X(z)$ .
Re-evaluation of DESI DR2 Tension: By removing the linear $w_0-w_a$ assumption, the authors show that the statistical significance of the deviation from a cosmological constant drops from 3.1 $\sigma$ (DESI claim) to $\sim$ 1 $\sigma$ for density and $\sim$ 2 $\sigma$ for the equation of state.
Quantification of Model Bias: The study provides concrete evidence that the linear parametrization $w_X(z) = w_0 + w_a(1-a)$ forces high-redshift behavior that is not supported by data, leading to artificial deviations.
Optimization of Redshift Binning: The authors systematically tested different redshift node configurations ( $\{z_i\}$ ) and found that the results are robust, with model selection criteria slightly favoring a configuration that omits the $z=4/3$ node due to lack of additional information.

4. Key Results

Consistency with $\Lambda$ CDM: When $X(z)$ is measured as a free function, the data is consistent with a cosmological constant ( $X(z)=1$ ) within 1 $\sigma$ at $z=2/3$ . The deviation is only 1-2 $\sigma$ in the range $0.4 \lesssim z \lesssim 0.9$ , which aligns with the specific redshift range where DESI BAO measurements themselves show a slight tension with $\Lambda$ CDM.
Superiority of Density over Equation of State:
- Model selection (AIC/BIC) strongly favors measuring $X(z)$ over $w_X(z)$ ( $\Delta AIC = 2.61$ , $\Delta BIC = 3.4$ ).
- $w_X(z)$ is significantly less constrained by data than $X(z)$ . At $z > 1$ , $w_X(z)$ becomes unconstrained (large error bars and unphysical negative values), whereas $X(z)$ remains well-behaved.
- This is because $w_X(z)$ enters the theory via an integral; errors in $w_X$ accumulate, and the derivative relationship makes $w_X$ highly sensitive to priors.
Robustness to Priors: The results for $X(z)$ are robust across different choices of CMB priors (Planck 2015 fixed neutrino mass vs. Planck 2018 varying neutrino mass) and wide priors on matter density ( $\Omega_m h^2$ ).
Comparison with DESI: The DESI best-fit linear models ( $w_0, w_a$ ) fall within the 1 $\sigma$ contours of the authors' model-independent $X(z)$ measurement. This implies that the 3.1 $\sigma$ tension reported by DESI is an artifact of the linear parametrization forcing a specific high-redshift evolution that the data does not support.

5. Significance and Implications

Caution Against Linear Parametrization: The paper concludes that assuming a linear $w_X(z)$ is "misleading" and precludes discovering the true time-variation of dark energy at higher redshifts. The linear model imposes a power-law behavior on $\rho_X(z)$ at high $z$ ( $\rho_X \propto (1+z)^{3(1+w)}$ ) without observational justification.
Urgent Goal: The most critical immediate goal in dark energy research is to definitively determine if the dark energy density varies with time, rather than focusing on the equation of state.
Future Prospects: Current data has a "gap" at $z > 1.5$ (only one Ly $\alpha$ point at $z=2.33$ ). The authors emphasize that upcoming surveys like Euclid ( $0.9 \le z \le 1.8$ ) and the Roman Space Telescope ( $1 \le z \lesssim 3$ ) are essential to fill this gap and provide the necessary constraints to distinguish between a cosmological constant and dynamical dark energy without relying on biased parametric assumptions.

In summary, this work argues that the current evidence for dynamical dark energy is weak (consistent with $\Lambda$ CDM) when analyzed without restrictive parametric assumptions, and that future progress depends on model-independent measurements of dark energy density.