Prior-free cosmological parameter estimation of Cosmicflows-4

By applying a prior-free forward-modeling approach to the Cosmicflows-4 catalog, this study reconstructs the local velocity field to derive a Hubble constant of $75.9\pm1$km/s/Mpc and identifies a significant tension between the observed bulk flow and$\Lambda$CDM predictions.

The Gist

Imagine the universe as a giant, invisible ocean. Most of the time, this ocean is expanding smoothly, like a balloon being blown up. This is the "Hubble flow." But, just like a real ocean has currents, whirlpools, and tides caused by underwater mountains, the universe has its own currents. These are caused by the gravity of massive clumps of matter (like galaxy clusters) pulling things toward them.

Astronomers call these extra movements peculiar velocities. They are the "drift" of a galaxy on top of the smooth expansion of the universe.

This paper is about a team of scientists trying to map these cosmic currents using a massive dataset called Cosmicflows-4 (CF4). Think of CF4 as a giant, slightly messy logbook containing the speed and direction of about 56,000 galaxies.

Here is the breakdown of their journey, explained simply:

1. The Problem: A Noisy, Patchy Map

The scientists wanted to answer two big questions:

1. How fast is the universe expanding right here? (The Hubble Constant, or $H_0$).
2. Is there a giant "super-current" pulling our local neighborhood of galaxies in a specific direction? (The Bulk Flow).

The problem is that the data in their logbook (CF4) is noisy and patchy.

• Noisy: Measuring the distance to a galaxy is like trying to guess the distance to a lighthouse in a foggy storm. You can see the light, but you aren't 100% sure how bright it should be, so your distance guess has errors.
• Patchy: The data isn't spread evenly. It's like looking at a map of the world where you have a high-resolution photo of Europe, but Africa is just a few blurry dots. This "patchiness" can trick you into thinking there is a strong current where there isn't one, or hiding a real one.

2. The Solution: A "Forward-Modeling" Detective

Instead of trying to clean up the messy data first (which often introduces its own biases), the team decided to work backward from the "truth."

Imagine you are trying to guess the weather in a city based on a few scattered weather reports from tourists.

• Old Method: You try to average the tourist reports, guess the errors, and then calculate the weather.
• Their New Method (Forward Modeling): They said, "Let's pretend we know the weather (the expansion rate and the currents). We will simulate what the tourists would have reported if that were true. Then, we compare our simulation to the actual tourist reports. If they match, our guess was right. If not, we tweak our guess and try again."

They used a super-computer to run millions of these "what-if" scenarios until they found the version of the universe that best matched the messy CF4 logbook. Crucially, they tried to do this without assuming the universe must behave a certain way (a "prior-free" approach). They let the data speak for itself.

3. The "Fake Universe" Test

Before trusting their method on real data, they had to test it. They created 64 fake universes (simulations) that looked exactly like the real CF4 data in terms of its messiness and patchiness.

• They knew the "true" currents in these fake universes because they built them.
• They ran their detective method on the fake data.
• The Discovery: They found that the patchiness of the data acts like a magnifying glass. It makes the currents look stronger than they actually are. If they didn't correct for this, they would have concluded there was a massive, impossible current pulling the universe.

4. The Results: What Did They Find?

A. The Expansion Rate ($H_0$)
By looking at how galaxies are flowing inward or outward, they calculated the expansion rate of the universe.

• Result: They found a value of 75.9 km/s/Mpc.
• Why it matters: This is on the "fast" side of the scale. It aligns with measurements taken from nearby stars and supernovae, but it disagrees with measurements taken from the early universe (the Cosmic Microwave Background). This disagreement is known as the "Hubble Tension," and this paper adds more weight to the idea that the local universe might be expanding faster than the early universe models predict.

B. The Cosmic Current (Bulk Flow)
They looked for a giant current pulling galaxies in a specific direction.

• Result: They found a significant "tug" in a specific direction (toward the "Great Attractor" region).
• The Tension: When they corrected for the "magnifying glass" effect of the patchy data, the current was still strong, but it was less extreme than the raw data suggested. However, it is still stronger than the standard model of cosmology (Lambda-CDM) predicts.
• The Analogy: Imagine the standard model predicts a gentle river current of 10 mph. The raw data looked like a 50 mph hurricane. After their correction, it looked like a 25 mph current. It's still faster than the model predicted, suggesting our model of the universe might be missing something about how matter is distributed.

5. The Big Picture

This paper is a triumph of honesty in data.

• They didn't force the data to fit a pretty theory.
• They admitted the data was messy and patchy.
• They built a sophisticated way to account for those messiness issues.

The Takeaway:
The universe is expanding a bit faster than some models predict, and there seems to be a stronger-than-expected cosmic current pulling our local neighborhood. While the "messiness" of our observations made the currents look wild, even after cleaning up the data, the universe still seems to be doing something a little more dramatic than our textbooks say it should.

It's like looking at a foggy window and realizing the rain streaks aren't just random; they are being pulled by a strong wind that our weather forecast didn't predict. We need to update our forecast!

Technical Summary

Here is a detailed technical summary of the paper "Prior-free cosmological parameter estimation of Cosmicflows-4":

1. Problem Statement

Galaxy peculiar velocities are crucial probes for understanding the large-scale structure (LSS) of the Universe, mapping mass distributions, and measuring the Hubble constant ($H_0$). However, catalogs of peculiar velocities, such as Cosmicflows-4 (CF4), suffer from significant challenges:

• Noise and Scarcity: Data is noisy, sparse at large distances, and subject to complex selection effects (anisotropy, magnitude limits).
• Bias in Distance Estimation: Traditional methods often rely on correcting distance moduli ($\mu$) using estimators that assume log-normal error distributions, introducing biases.
• Model Dependence: Many reconstruction methods (e.g., Wiener filters) impose strong priors based on the $\Lambda$CDM model and the matter power spectrum. This limits their ability to genuinely test the $\Lambda$CDM model, as the priors can "hide" tensions between data and theory.
• The $H_0$ Tension: There is a persistent discrepancy between local measurements of $H_0$ and those derived from the Cosmic Microwave Background (CMB).

The authors aim to reconstruct the radial flow (monopole) and bulk flow (dipole) directly from CF4 data without imposing a cosmological prior on the velocity field, while simultaneously estimating a local $H_0$.

2. Methodology

The authors developed a forward-modeling approach using a Bayesian framework to reconstruct the first two spherical modes of the velocity field.

• Forward Modeling & Likelihood:

  • Instead of correcting distances first, the model directly predicts the observed distance modulus ($\mu_{obs}$) for each galaxy based on a set of parameters: radial flow ($V_r$) and bulk flow vector ($\mathbf{V} = V_X, V_Y, V_Z$).
  • The model assumes a flat Universe and uses the redshift-distance relation to convert proposed peculiar velocities into cosmological redshifts and physical distances.
  • The likelihood function compares the predicted distance moduli with the observed CF4 data, assuming Gaussian errors on $\mu$.
  • Prior-Free: The method assumes a flat prior ($P(\pi) = 1$) on the velocity parameters, avoiding assumptions about the power spectrum of the velocity field.

• Concentric Shell Reconstruction:

  • To address the issue that data is concentrated near the observer (where errors are small) and sparse at large distances, the catalog is split into concentric shells (thickness $\Delta r = 20$ Mpc/h).
  • The Monte Carlo algorithm (Hamiltonian Monte Carlo) runs independently on each shell.
  • Results are reassembled into spherical volumes by volume-weighted averaging, allowing for a depth-dependent reconstruction of the flow.

• Simulation-Based Correction:

  • To quantify and correct for systematic biases introduced by the complex CF4 selection function (anisotropy, magnitude limits), the authors generated 64 mock catalogs.
  • These mocks were built using the Big MultiDark N-body simulation (Planck cosmology) but tailored to mimic the specific angular footprints and magnitude limits of CF4 sub-samples (SDSS-PV, 6dFGRSv, and others).
  • By comparing the forward-model output on mocks against the "ground truth" of the simulation, they derived correction factors ($\Delta$ for offsets and $R$ for amplification) to apply to the real CF4 data.

3. Key Contributions

• Prior-Free Reconstruction: The method reconstructs velocity fields without assuming a specific $\Lambda$CDM power spectrum, allowing for a more unbiased test of cosmological models.
• Direct Handling of Observables: The approach bypasses intermediate distance estimators, treating distance moduli directly within the Bayesian framework to avoid log-normal bias propagation.
• Simulation-Based Bias Correction: The development of a robust correction scheme derived from 64 realistic mocks allows for the mitigation of systematic amplification caused by the CF4 survey footprint.
• Joint Estimation: The method simultaneously estimates the radial flow, bulk flow, and the local Hubble constant ($H_0$) in a unified probabilistic framework.

4. Key Results

A. Performance on Mock Data

• The method accurately recovers radial and bulk flows within $\sim 60$ Mpc/h.
• Systematic Amplification: Due to data incompleteness and anisotropy (especially beyond 160 Mpc/h), the uncorrected method systematically amplifies the amplitude of the bulk flow (by a factor of up to 5 at the edge of the catalog).
• The simulation-based corrections successfully recover the true underlying flow, though they introduce additional variance.

B. Results on Cosmicflows-4 Data

• Hubble Constant ($H_0$):

  • By marginalizing over the radial flow and assuming the cosmological principle (radial flow converges to zero at large scales due to cosmic variance), the authors derived a local $H_0$.
  • Result: $H_0 = 75.9 \pm 1$ km/s/Mpc (statistical error).
  • This value sits at the higher end of local measurements, contributing to the existing tension with CMB-derived values ($\sim 67.4$ km/s/Mpc).

• Bulk Flow:

  • Components: The $V_Y$ and $V_Z$ components are consistent with $\Lambda$CDM predictions. However, the $V_X$ component (Supergalactic X direction) shows significant tension.
  • Magnitude: The uncorrected bulk flow amplitude rises artificially beyond 160 Mpc/h. After applying the simulation-based correction, the bulk flow amplitude peaks at $\sim 350$ km/s at $\sim 140$ Mpc/h.
  • Tension: Even after correction, there is a **$3\sigma$tension** with$\Lambda$CDM predictions regarding the magnitude of the bulk flow and the$V_X$direction at distances around 140–240 Mpc/h. The uncorrected data rules out$\Lambda$CDM entirely ($>8\sigma$), highlighting the necessity of the correction.

5. Significance and Conclusion

• Confirmation of Tensions: The study confirms the local $H_0$ tension and identifies a significant, persistent tension in the local bulk flow compared to $\Lambda$CDM predictions, even after rigorous bias correction.
• Methodological Advancement: The work demonstrates that forward modeling, when combined with realistic mock-based corrections, can extract cosmological parameters from noisy, anisotropic catalogs without relying on strong theoretical priors.
• Implications: The finding that the bulk flow amplitude rises at $\sim 140$ Mpc/h (a "Great Attractor" scale) suggests either a statistical fluctuation, a failure of the $\Lambda$CDM model on these scales, or residual systematic errors in the composite CF4 catalog that require further investigation.
• Future Work: The authors emphasize the need to better handle systematic errors inherent in composite catalogs (e.g., inter-calibration between different surveys) and to acquire more spherically distributed peculiar velocity data to reduce anisotropy.

In summary, this paper provides a robust, prior-free framework for analyzing peculiar velocities, yielding a high local $H_0$ and confirming a significant bulk flow anomaly that challenges standard cosmological expectations.