🔭 astrophysics

(An)Isotropy in Pantheon+ and Type Ia supernova samples: intrinsic limits of directional tests

This paper demonstrates that current Type Ia supernova samples, including Pantheon+, lack the intrinsic statistical robustness and sky coverage necessary to reliably determine the direction of Hubble constant anisotropy using standard region-fitting or hemisphere-comparison methods, regardless of conflicting claims in previous studies.

Original authors: Antonio Quintana-Estellés, Pilar Ruiz-Lapuente

Published 2026-05-19

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Antonio Quintana-Estellés, Pilar Ruiz-Lapuente

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 Question: Is the Universe the Same Everywhere?

Imagine the Universe as a giant, expanding balloon. For decades, scientists have operated under a rule called the Cosmological Principle, which assumes that if you zoom out far enough, the Universe looks the same in every direction (isotropic) and everywhere (homogeneous). It's like a well-mixed bowl of soup: no matter which spoonful you take, the taste should be the same.

However, there is a nagging puzzle. When we measure how fast the Universe is expanding today (the Hubble Constant, or $H_0$ ) using different methods, we get different answers. Some measurements say it's expanding fast; others say it's slower. This has led some researchers to wonder: Is the Universe actually a lopsided soup? Maybe the expansion is faster in one direction and slower in another.

The Experiment: Checking the "Flavor" of the Sky

To test this, astronomers use Type Ia Supernovae. Think of these as "standard candles"—stars that explode with a known, predictable brightness. By looking at how bright they appear to us, we can calculate their distance and how fast they are moving away.

Scientists have taken huge catalogs of these supernovae (like the Pantheon+ database, which contains over 1,500 of them) and tried to map the expansion rate in different patches of the sky. They used two main methods to slice up the sky:

Hemisphere Comparison: Splitting the sky into two halves (like an orange) and comparing the expansion rate of the left side vs. the right side.
Region Fitting: Looking at smaller, specific circles or regions across the sky to find where the expansion rate is highest or lowest.

The Problem: The "Noisy" Ruler

The authors of this paper argue that previous studies claiming to find "directions of anisotropy" (lopsidedness) might be chasing ghosts.

Here is the core issue: The measuring tool itself is a bit fuzzy.

Imagine you are trying to measure the height of a crowd of people using a ruler that has a slight wobble. Even if everyone is exactly the same height, your measurements will vary slightly because of the wobble in the ruler.

In astronomy, the "wobble" comes from the intrinsic error in how we calculate distances from supernova light. This error causes the calculated expansion rate to vary by about 4 to 9 km/s/Mpc from one supernova to the next, even if the Universe is perfectly uniform.

The paper argues that when scientists slice the sky into chunks, the "noise" from this fuzzy ruler is so loud that it drowns out any real signal. It's like trying to hear a whisper in a rock concert; the whisper (a real cosmic direction) gets lost in the noise (measurement errors).

The Simulation: Shaking the Dice

To prove their point, the authors ran a massive simulation. They took the real data and added random "jitters" to the measurements, mimicking the natural fuzziness of the supernova method.

The Result: When they ran the analysis on this "jittered" data, the "directions" of the expansion rate changed completely every time.
The Analogy: Imagine you are trying to find the windiest spot in a park by measuring wind speed at different trees. But your anemometer (wind gauge) is broken and gives random numbers. If you run the test once, you might say the wind is strongest in the North. If you run it again with the broken gauge, you might say it's strongest in the South. The direction you find isn't real; it's just a result of the broken gauge.

The paper found that the "preferred directions" identified by other researchers were just random fluctuations caused by this measurement noise, not a real feature of the Universe.

The Calibration Issue: The "Local" vs. "Global" Map

The paper also highlights a specific problem with how the data is calibrated. To measure distances accurately, astronomers use Cepheid stars (a type of pulsating star) as a reference.

The Mistake: Some previous studies only used the Cepheid stars that happened to be inside the specific patch of sky they were studying. Since there are very few of these stars, this is like trying to calibrate a map of the whole world using only one or two landmarks in a single city. This led to false "directions" of anisotropy.
The Fix: The authors used all available Cepheid stars to calibrate the entire dataset. When they did this, the "lopsided" signals disappeared, and the results became consistent with a uniform, isotropic Universe.

The Conclusion: No Lopsided Soup (Yet)

The paper concludes that with the current data and current measurement tools, we cannot reliably claim that the Universe is lopsided.

The "directions" of anisotropy found in other studies are likely just statistical noise or artifacts of how the data was processed.
The variations we see in the expansion rate across the sky are consistent with what we expect from measurement errors, not from the Universe actually being different in different directions.
To truly solve this mystery, we need better data: more supernovae, a more uniform coverage of the sky, and, most importantly, more precise measuring tools that reduce that "wobble" in the ruler.

In short: The Universe might still be perfectly uniform. The apparent "lopsidedness" is likely just the result of our measuring tools being a little too fuzzy to tell the difference.

Technical Summary: (An)Isotropy in Pantheon+ and Type Ia Supernova Samples: Intrinsic Limits of Directional Tests

Problem Statement
Recent investigations into the statistical isotropy of the Universe using large Type Ia supernova (SNe Ia) databases, such as Pantheon+, have yielded contradictory claims regarding the existence of anisotropy in the Hubble constant ( $H_0$ ). While some studies report significant directional preferences (anisotropy) in different sky patches (60° or 90°), others find the Universe to be isotropic. A major source of confusion is the lack of consensus on the specific directions of anisotropy across different research works. Furthermore, previous analyses often fail to account for the intrinsic limitations of the SNe Ia lightcurve method, specifically the scatter in individual $H_0$ measurements, which can reach 8–9 km s $^{-1}$ Mpc $^{-1}$ . This paper addresses whether the current SNe Ia samples possess the statistical robustness to definitively identify directions of anisotropy or if reported signals are artifacts of methodological limitations and sample variance.

Methodology
The authors employ two widely used directional analysis techniques: the Region Fitting (RF) method and the Hemisphere Comparison (HC) method. The study utilizes the Pantheon+ dataset (1,701 light curves from 1,550 distinct SNe Ia) and the Carnegie Supernova Project (CSP) sample (342 SNe Ia).

Key methodological steps include:

Bayesian MCMC Analysis: Cosmological parameters ( $\Omega_m$ , $h$ , $M$ ) are fitted for 2,000 distinct random sky directions using a Bayesian Markov Chain Monte Carlo approach (CosmoSIS framework).
Cepheid Calibration Variations: For the Pantheon+ dataset, the authors compare two calibration scenarios:
- Using only the local Cepheid calibrators (42 total) located within each specific sky region.
- Using the full set of 42 Cepheid calibrators for all regions, regardless of location.
Cholesky Decomposition: To account for correlations between overlapping SNe Ia in different sky regions, the full covariance matrix is utilized via Cholesky decomposition to generate statistically robust significance estimates.
Intrinsic Scatter Simulations: To test the stability of inferred anisotropy directions, the authors simulate 2,000 realizations of the datasets. In these simulations, the distance modulus of each supernova is perturbed by the intrinsic scatter of the SNe Ia lightcurve method ( $\sigma_{int} \approx 0.11$ mag), which corresponds to an intrinsic variation of approximately 4 km s $^{-1}$ Mpc $^{-1}$ in $H_0$ .
CSP Analysis: The CSP sample, which lacks a full covariance matrix, is analyzed using the HC method with all 25 available Cepheid calibrators, incorporating both statistical and systematic uncertainties.

Key Results

Instability of Anisotropy Directions: When applying the RF and HC methods to the Pantheon+ dataset, the directions of maximum and minimum $h$ $h$ (dimensionless Hubble constant) are highly unstable.
- Using local Cepheids per region yields a maximum $h$ of $0.757 \pm 0.016$ at $(l, b) \approx (31.4^\circ, -15.2^\circ)$ with a significance of $2.8\sigma$ .
- However, when the full set of 42 Cepheids is used for calibration, the significance of this direction drops to $<1\sigma$ , and the preferred direction shifts to $(318.7^\circ, -11.7^\circ)$ .
Sensitivity to Intrinsic Scatter: The simulations involving intrinsic scatter ( $\sigma_{int} \approx 0.11$ mag) demonstrate that the inferred directions of anisotropy change arbitrarily. In the RF method, the directions of maximum and minimum $h$ vary irregularly across realizations. In the HC method, while the directions show some alignment, they shift beyond estimated uncertainty regions. This indicates that the identified "preferred directions" are not robust against the known intrinsic scatter of the SNe Ia method.
CSP Sample Findings: Analysis of the CSP sample initially shows a $2.5\sigma$ anisotropy signal when considering only statistical uncertainties. However, when systematic uncertainties are included (totaling a $1\sigma$ level), the signal becomes statistically insignificant. Furthermore, the inferred directions in CSP are weakly constrained and appear driven by the inhomogeneous sky coverage of the survey rather than a cosmological signal.
Magnitude of Variation: The point-to-point variation of $H_0$ in the samples is observed to be 8–9 km s $^{-1}$ Mpc $^{-1}$ . The authors note that the intrinsic error of the lightcurve method alone accounts for $\sim4$ km s $^{-1}$ Mpc $^{-1}$ . This level of scatter is sufficient to produce arbitrary directional trends when analyzed with current methods.

Significance and Conclusions
The paper concludes that the current SNe Ia samples (Pantheon+ and CSP) do not possess the statistical robustness to determine the direction of anisotropy in $H_0$ with confidence. The authors argue that the contradictory claims of anisotropy in the literature arise from:

The use of small, localized Cepheid calibration subsamples which introduce statistical fluctuations.
The failure to account for the intrinsic scatter of the SNe Ia lightcurve method, which is large enough to mimic or obscure directional signals.
The sensitivity of the RF and HC methods to the non-uniform angular distribution of the data.

The study posits that the apparent anisotropy directions found in previous works are likely artifacts of these limitations rather than genuine cosmological features. The authors state that achieving robust constraints on isotropy will require larger SNe Ia samples with improved systematics, lower individual uncertainties, and more uniform sky coverage. Until such data is available, the identification of a cosmic preferred direction using current SNe Ia datasets remains inconclusive.