Updates on dipolar anisotropy in local measurements of the Hubble constant from Cosmicflows-4

Using the Cosmicflows-4 catalogue and a logarithmic distance-modulus formulation, this study finds that while uncorrected local Hubble constant measurements exhibit a significant dipolar anisotropy, this signal is largely driven by peculiar velocity flows and survey selection effects rather than a fundamental breakdown of isotropic expansion, implying a limited impact on the global Hubble tension.

The Gist

The Big Picture: The "Speed Limit" of the Universe

Imagine the universe is a giant, expanding balloon. As it inflates, every point on the surface moves away from every other point. The rate at which it expands is called the Hubble Constant ($H_0$). Think of this as the "speed limit" of the universe.

For a long time, scientists have been arguing about what that speed limit actually is.

• Team CMB (The Baby Picture): Looking at the oldest light in the universe (the Cosmic Microwave Background), they say the speed limit is about 67 km/s per Megaparsec.
• Team Local (The Neighborhood Watch): Looking at nearby stars and galaxies, they say the speed limit is about 73 km/s per Megaparsec.

This difference is the famous "Hubble Tension." It's like two speed cameras on the same highway giving you two different speed limits. One says 60, the other says 70. Someone has to be wrong, or something weird is happening.

The New Investigation: Is the Speed Limit Different in Different Directions?

This paper asks a specific question: Is the speed limit the same in every direction?

Maybe the universe isn't expanding perfectly evenly. Maybe it's expanding faster to the North than to the South, or faster toward the Shapley Supercluster (a giant cluster of galaxies) than away from it. If the speed limit changes depending on which way you look, that could explain why we are getting confused measurements.

The authors used a massive database called Cosmicflows-4 (CF4), which is like a giant GPS directory for 55,000 galaxies, telling us how far away they are and how fast they are moving.

The Three Big Mistakes They Fixed

The authors noticed that many previous studies made three subtle but important mistakes. They fixed these to get a clearer picture:

1. The "Ruler" Problem (Math): Most studies tried to measure distance in "miles" (Luminosity Distance). But the data they had was actually in "miles per hour" (Distance Modulus). It's like trying to weigh a bag of flour by looking at a speedometer. The authors realized that if you do the math using the raw "speedometer" numbers (logarithms), the errors behave much more nicely. It's like using a ruler that doesn't stretch.
2. The "Bad Neighborhood" Problem (Data Selection): The galaxy database isn't perfect. It has gaps (like the Milky Way blocking our view) and it's crowded in some areas (like the Northern Hemisphere sky). The authors acted like a quality control inspector. They threw out the "bad apples"—data that was too close (too messy) or too far (too biased)—and only kept the "sweet spot" of data where the measurements were reliable.
3. The "Wind" Problem (Peculiar Velocities): This is the most important part. Galaxies aren't just moving because the universe is expanding; they are also moving because of gravity, like leaves blowing in a wind. If a galaxy is being pulled toward a giant cluster, it looks like it's moving faster than it really is.
   • The Analogy: Imagine you are on a moving walkway at an airport (the expansion of the universe). But you are also running forward because you are late for a flight (gravity pulling you). If you just measure your total speed, you think the walkway is faster than it is. You have to subtract your running speed to know the true speed of the walkway.

What They Found

The authors ran their tests in two ways:

1. The "Raw Data" Test (Ignoring the Wind)
When they looked at the raw numbers without correcting for the "wind" (gravity), they found a strong signal.

• The Result: The universe did look like it was expanding faster in one direction (a "dipole") and slower in the opposite direction.
• The Direction: This direction pointed roughly toward the Shapley Supercluster, a massive gravitational attractor.
• The Meaning: This looked like the universe was anisotropic (different in different directions).

2. The "Corrected Data" Test (Subtracting the Wind)
Then, they used a sophisticated model to subtract the "wind" (the peculiar velocities caused by gravity).

• The Result: The strong signal disappeared. The "speed limit" became much more uniform. The remaining signal was very weak and likely just random noise.
• The Meaning: The "anisotropy" (the difference in speed) wasn't because the universe is expanding unevenly. It was because the galaxies were being pulled by local gravity.

The Conclusion: It's Not the Universe, It's the Traffic

The paper concludes that the "Hubble Tension" is not caused by the universe expanding differently in different directions.

• The Analogy: Imagine you are trying to measure the speed of a river. If you stand on a boat that is also being pushed by a strong wind, you might think the river is flowing faster than it is. Once you account for the wind, you realize the river is flowing at a steady, normal speed.
• The Takeaway: The "weird" results we see in local measurements are mostly due to local traffic jams (galaxies being pulled by gravity) and survey biases (where we looked), not a fundamental flaw in the laws of physics or the expansion of the universe.

Does This Solve the Hubble Tension?

Not exactly.

The authors checked if this "local wind" could explain why the local speed limit (73) is higher than the baby picture speed limit (67). They found that while the local wind creates a dipole, the galaxies used to calibrate the local speed limit (Supernovae) are spread out in a way that doesn't perfectly align with this wind.

So, while the "wind" explains why the data looks messy, it doesn't fully explain the 7 km/s difference between the two teams. The mystery of the Hubble Tension remains, but this paper tells us that we don't need to invent new physics about the universe expanding unevenly to explain the messiness of our local data. The messiness is just local traffic.

Technical Summary

1. Problem Statement

The paper addresses the ongoing "Hubble tension," the significant discrepancy ($\sim 7\sigma$) between the Hubble constant ($H_0$) inferred from the Cosmic Microwave Background (CMB) under the $\Lambda$CDM model ($H_0 \approx 67.4$ km s$^{-1}$ Mpc$^{-1}$) and local measurements using the distance ladder ($H_0 \approx 73\text{--}74$ km s$^{-1}$ Mpc$^{-1}$).

A leading hypothesis to resolve this tension involves local anisotropies or bulk flows that might cause $H_0$ to vary across the sky. Previous studies using the Cosmicflows-3 and Pantheon+ catalogs have reported dipolar anisotropies in $H_0$. However, these studies often suffer from:

• Methodological inconsistencies: Using luminosity distances ($d_L$) which have log-normal error distributions, while applying statistical tools assuming Gaussian errors.
• Sample selection bias: Lack of rigorous internal consistency checks regarding the depth dependence of $H_0$ and residual skewness/kurtosis.
• Ambiguity in corrections: Unclear separation between observed anisotropies driven by peculiar velocities (local kinematics) and those potentially arising from a breakdown of isotropic expansion.

This study aims to re-evaluate dipolar anisotropy using the Cosmicflows-4 (CF4) catalog with a more rigorous statistical formulation, internal consistency checks, and a clear distinction between raw and peculiar-velocity-corrected data.

2. Methodology

A. Data Source

The authors utilize the Cosmicflows-4 (CF4) catalog, which contains distance estimates for 55,877 galaxies in 38,065 groups.

• Key Inputs: Distance moduli ($\mu$), galactic coordinates, and observed systemic radial velocities in the CMB frame ($v_{CMB}$).
• Peculiar Velocity Correction: The authors also use CF4pec, a subset where peculiar velocities ($v_{pec}$) are inferred via a Bayesian forward-modeling approach (based on [60]) within a $\Lambda$CDM framework. This allows for the subtraction of local flows: $v_{corr} = v_{CMB} - v_{pec}$.

B. Statistical Formulation (Key Innovation)

The authors argue that standard analyses using $H_0 = v/d_L$ introduce biases because distance moduli ($\mu$) have Gaussian errors, whereas luminosity distances ($d_L$) are log-normally distributed.

• Logarithmic Hubble-Lemaître Law: They reformulate the law directly in terms of distance moduli to preserve Gaussian error properties:
  $$\frac{\mu}{5} - 5 = \log(v_{corr}) - \log(H_0) + f_\mu(z, q_0, j_0)$$
  where $f_\mu$ is the cosmographic expansion of the distance modulus, distinct from the expansion of $d_L$.
• Correction Terms: They explicitly account for the difference between the cosmographic expansion of $d_L$ and $\mu$, noting that previous studies often conflated the two.

C. Sample Selection and Consistency Checks

Rather than using the full catalog, the authors perform internal consistency tests to define "healthy" subsamples free from Malmquist bias and selection effects:

• Depth Dependence: They analyze the variation of $\langle \log H_0 \rangle$ with distance modulus ($\mu$) and redshift ($z$).
• Residual Statistics: They examine the skewness and kurtosis of residuals across radial shells.
• Selected Ranges: They identify conservative subsamples where the Hubble flow is stable and residuals are Gaussian:
  • Distance Modulus: $\mu \in [31, 36]$
  • Redshift: $z \in [0.03, 0.06]$
  • They note that $\mu < 31$ suffers from low statistics, while $\mu > 36.5$ shows signs of selection bias (increasing skewness/kurtosis).

D. Anisotropy Analysis

• Spherical Harmonic Expansion: They fit the angular distribution of $\log H_0$ using a spherical harmonic expansion up to the octupole order ($\ell_{max}=3$):
  $$H_0(\theta, \phi) = \sum_{\ell=0}^{3} \sum_{m=-\ell}^{\ell} a_{\ell m} Y_{\ell m}(\theta, \phi)$$
• Model Selection: They use Bayesian evidence (Bayes Factor) to compare models (Monopole vs. Dipole vs. Quadrupole vs. Octupole) against a constant $H_0$ (Monopole-only) hypothesis.
• Radial Evolution: They test for a monotonic decrease in dipole amplitude with distance, a signature predicted by certain differential expansion models (e.g., LTB solutions).

3. Key Contributions

1. Statistical Rigor: The paper establishes that using the logarithmic formulation based on distance moduli is methodologically superior for CF4 data, avoiding the log-normal bias inherent in $d_L$-based fits.
2. Internal Consistency: It provides a robust protocol for defining subsamples by analyzing the depth dependence of $H_0$ and the statistical moments (skewness/kurtosis) of residuals, effectively filtering out Malmquist bias.
3. Separation of Kinematics vs. Cosmology: By comparing raw CF4 data with the peculiar-velocity-corrected CF4pec data, the study isolates the contribution of local bulk flows from potential cosmological anisotropies.
4. Re-evaluation of Dipole Direction: It offers updated constraints on the dipole direction and amplitude, comparing them with CMB dipoles and previous literature.

4. Results

A. Uncorrected Data (CF4)

• Significant Anisotropy: In the selected subsamples ($\mu \in [31, 36]$ and $z \in [0.03, 0.06]$), the uncorrected CF4 data shows a statistically significant anisotropic signal.
• Dipole Dominance: The signal is dominated by a dipole term. The Bayes Factor strongly favors a dipole+quadrupole model over a monopole-only model (decisive evidence).
• Direction: The dipole points toward galactic coordinates roughly $(l_g \approx 299^\circ, b_g \approx 10^\circ)$ for the $\mu$-binned sample and $(l_g \approx 296^\circ, b_g \approx -10^\circ)$ for the $z$-binned sample. This direction is consistent with previous findings using CF3/Pantheon but does not align with the CMB dipole (angular distance $\sim 52^\circ\text{--}65^\circ$).
• Bulk Flow: The inferred bulk flow associated with this dipole is $\sim 298$ km s$^{-1}$, aligning with the Shapley Supercluster attractor.

B. Corrected Data (CF4pec)

• Suppression of Signal: When peculiar velocities are subtracted, the anisotropy amplitude is strongly reduced.
• Residuals: At lower depths, the anisotropy becomes statistically insignificant. At larger distances, a weak residual signal remains, but the Bayes Factor no longer provides decisive evidence for anisotropy over a monopole.
• Conclusion: The dominant anisotropy in the raw CF4 data is driven by local kinematic flows (peculiar velocities) rather than a fundamental breakdown of isotropic expansion.

C. Radial Evolution

• No Monotonic Trend: The study tested for a distance-dependent decrease in dipole amplitude (as predicted by some LTB/differential expansion models).
• Result: No robust evidence for a monotonic trend was found. The dipole amplitude fluctuates across shells without a clear decreasing pattern, weakening the case for large-scale anisotropic expansion.

D. Impact on Hubble Tension

• Limited Impact: The authors map the positions of SNe Ia calibrators (CCHP and SH0ES) and Hubble-flow SNe Ia onto the anisotropy map.
• Finding: The sky distribution of these calibrators does not show a strong alignment with the detected dipole pattern.
• Implication: While local anisotropy affects local $H_0$ measurements, it is unlikely to be the primary driver of the global Hubble tension. In fact, if anything, the "cold" regions of the anisotropy map where many SNe Ia reside might suggest $H_0$ is slightly underestimated locally, potentially exacerbating the tension rather than solving it.

5. Significance and Conclusion

This paper provides a critical "reality check" on claims of Hubble constant anisotropy. Its primary conclusion is that the dipolar anisotropy observed in local galaxy surveys is largely an artifact of local peculiar velocities and survey selection effects, not a sign of new physics or a violation of the Cosmological Principle on large scales.

• Methodological Impact: It highlights the necessity of using logarithmic formulations for distance moduli to avoid statistical biases in cosmographic analyses.
• Physical Insight: It confirms that local bulk flows (e.g., toward Shapley) are the primary source of apparent $H_0$ variations in the local universe.
• Future Outlook: The authors conclude that resolving the Hubble tension will likely require new physics or better understanding of systematics, rather than invoking large-scale anisotropic expansion. Future surveys with full covariance matrices and homogeneous selection functions are needed to detect any residual, genuine cosmological anisotropy beyond the local kinematic regime.