Unbiased Bayesian Inference of Peculiar Motions of Galaxies from Type Ia Supernovae Observations

This paper introduces an unbiased Bayesian method for estimating the radial peculiar velocities of galaxies hosting Type Ia supernovae that relies solely on cosmological models and data precision, avoiding the linearity and fixed-cosmology assumptions of standard estimators to achieve lower bias even for large velocities.

The Gist

Imagine the universe as a giant, expanding balloon. If you draw dots on it, they all move away from each other as the balloon inflates. This is the "Hubble Flow"—the smooth, predictable expansion of space itself.

However, the dots (galaxies) aren't just drifting with the wind; they are also jostling around. They bump into each other, get pulled by the gravity of massive clusters, and wiggle. These extra, chaotic movements are called Peculiar Motions.

Measuring these wiggles is like trying to hear a whisper in a hurricane. The "whisper" is the galaxy's own movement, and the "hurricane" is the massive expansion of the universe. If you don't account for the wiggle, you get the wrong idea about how fast the universe is expanding, which throws off our understanding of Dark Energy and Gravity.

Here is a simple breakdown of what this paper does to solve that problem.

The Old Way: The "Straight Line" Guess

For a long time, scientists tried to measure these wiggles using a shortcut. They assumed that if you zoom in close enough, the relationship between how far away a galaxy is and how fast it's moving looks like a straight line.

Think of it like driving on a highway. If you are driving at 60 mph and the wind blows you slightly off course, you can easily calculate the wind's push because your path is almost straight.

The Problem:
This "straight line" math only works when the wiggle is tiny compared to the expansion speed. But near us (in the "local" universe), galaxies are moving fast relative to the expansion. It's like trying to use that same straight-line math when you are driving a race car that is also spinning in circles. The math breaks, and the results become biased (wrong).

Also, the old method required scientists to guess the "rules of the road" (the exact cosmology) beforehand. If they guessed the rules wrong, the whole calculation was wrong.

The New Way: The "Smart Detective" (Bayesian Inference)

The authors of this paper, Ujjwal Upadhyay and his team, built a new, smarter tool. Instead of guessing a straight line, they use a Bayesian approach.

The Analogy: The Blurred Photo
Imagine you take a photo of a galaxy, but the camera is slightly shaky. You see a blurry dot.

• The Old Method: Assumes the blur is just a tiny, predictable wobble and tries to sharpen it using a fixed formula.
• The New Method: The detective asks, "What are all the possible ways this photo could have looked if the camera shook in different directions?"

They use a computer technique called MCMC (Markov Chain Monte Carlo). Think of this as a super-powered "trial and error" machine.

1. It looks at the data (the brightness of a Supernova and its redshift).
2. It asks: "If the universe is expanding at this speed, and the galaxy is wiggling at that speed, does the math match the photo?"
3. It tries millions of combinations of "universe rules" and "galaxy wiggles" simultaneously.
4. It finds the most likely answer without forcing the "wiggle" to be small or the "universe rules" to be fixed.

Why This is a Big Deal

1. No More "Small Wiggle" Assumption: The new method works even when the galaxy is wiggling wildly (when the peculiar velocity is as big as the expansion speed). It doesn't break down like the old straight-line math.
2. No "Guessing the Rules": It figures out the expansion speed and the wiggle speed at the same time. It doesn't need you to tell it the answer first. This makes it unbiased.
3. It Handles the "Fog": The paper shows that at very far distances, the "wiggle" is too small to see against the "hurricane" of expansion, so the method correctly says, "I can't tell for sure." But near us, where the wiggles are strong, it gives a very accurate reading.

The Results

The team tested their new "Smart Detective" on fake data (simulations) that looked exactly like what we see in the real universe.

• The Old Method: Got the answer wrong when the wiggles were big.
• The New Method: Got the answer right, even when the wiggles were huge.

They also tested it on real data from the Pantheon+ sample (a huge list of 1,700 exploding stars). They found that for nearby stars, they could measure the wiggles, but for far-away stars, the data wasn't precise enough yet. However, they are confident that with future telescopes (like LSST and ZTF) that will see thousands more stars, this method will unlock a new way to map the invisible gravitational forces of the universe.

The Bottom Line

This paper gives us a better ruler to measure the universe. By stopping the assumption that everything moves in a straight line and letting the data speak for itself, we can finally measure how galaxies are "dancing" around the cosmic expansion. This helps us understand the invisible dark energy pushing the universe apart and the gravity pulling it together.

Technical Summary

Here is a detailed technical summary of the paper "Unbiased Bayesian Inference of Peculiar Motions of Galaxies from Type Ia Supernovae Observations" by Upadhyay, Saini, and Sethi.

1. Problem Statement

The peculiar motions of galaxies (deviations from the pure Hubble flow caused by gravitational interactions with large-scale structure) are crucial cosmological probes for studying the growth of structure, dark energy, and modified gravity. However, measuring these velocities is challenging because they are entangled with the cosmological redshift.

Current standard methods for estimating peculiar velocities ($v_p$) from Type Ia Supernovae (SNe Ia) rely on Hubble residuals derived from the magnitude-redshift relation. These methods suffer from two critical limitations:

1. Linear Approximation Bias: They assume a locally linear relationship between magnitude and redshift. This approximation ($v_p \ll cz$) breaks down at low redshifts where peculiar velocities are comparable to the Hubble flow ($v_p \sim cz$), leading to significant biases.
2. Cosmology Dependence: Standard estimators require a fixed, fiducial cosmology and the true cosmological redshift ($z_{cos}$) to be known. In practice, these are approximated using observed redshifts ($z_{obs}$) and assumed parameters, introducing systematic biases if the assumed cosmology is incorrect.

2. Methodology

The authors propose a fully Bayesian framework to estimate line-of-sight peculiar velocities without relying on linear approximations or fixed cosmologies.

A. General MCMC Method (The "Exact" Approach)

The core of the proposed method treats the fitting of the magnitude-redshift relation as an errors-in-variables model.

• Model Formulation: The apparent magnitude ($m$) is the dependent variable, and the redshift ($z$) is the independent variable. Both have uncertainties.
• Peculiar Velocity as a Parameter: Instead of treating the observed redshift as the true cosmological redshift, the method treats the true redshift ($z^*$) as a free parameter for each supernova. The difference between the observed redshift ($z_{obs}$) and the inferred true redshift ($z^*$) is attributed to the peculiar velocity: $\Delta z = z_{obs} - z^* = (v_p/c)(1+z)$.
• Likelihood Construction: The posterior distribution is constructed by sampling over:
  1. Cosmological parameters ($\theta = \{H_0, \Omega_M\}$).
  2. True redshifts ($z^*_i$) for all $N$ supernovae.
  3. The likelihood accounts for the covariance of SNe Ia magnitudes ($C_m$) and the prior distribution of redshift errors (derived from expected peculiar velocity distributions).
• Regularization: A $\chi^2$ regularization factor is included to prevent overfitting, as the number of parameters ($N$ redshifts + cosmological parameters) exceeds the number of data points.
• Sampling: The authors use Metropolis-Hastings MCMC to sample the joint posterior distribution $P(\theta, z^* | D)$.

B. Generalized Linear Estimator (The "Linear" Approach)

For comparison, the authors derive a Bayesian version of the standard linear estimator.

• They assume local linearity of the magnitude-redshift relation ($m(z^*) \approx m(z) + m'(z)(z^*-z)$).
• They assume Gaussian distribution for redshift errors.
• Key Improvement: Unlike standard literature, this derivation explicitly incorporates the covariance matrix of SNe Ia magnitudes ($C_m$), providing a more rigorous analytic expression for the linear estimator.

3. Key Contributions

1. Unbiased Estimation at Low Redshift: The method removes the bias inherent in linear approximations, remaining valid even when $v_p \approx cz$ (the regime where standard estimators fail).
2. Cosmology Independence: By jointly inferring cosmological parameters and peculiar velocities, the method avoids bias introduced by assuming an incorrect fiducial cosmology.
3. Generalized Linear Derivation: Provides a rigorous Bayesian derivation of the standard linear estimator that includes magnitude covariance, serving as a robust baseline for comparison.
4. Full Posterior Inference: Unlike point-estimate methods, this approach yields the full posterior distribution of peculiar velocities, naturally capturing correlations between velocities at different redshifts.

4. Results

The authors validated their methods using simulated SNe Ia data (mimicking the Pantheon+ sample) and applied them to the actual Pantheon+ dataset.

A. Validation with Simulated Data

• Accuracy: The MCMC-based "Exact" method successfully recovered the input cosmological parameters and peculiar velocities. The estimates were statistically consistent with true values across all redshifts.
• Bias Comparison:
  • Linear Method: Showed negligible bias only when $v_p \ll cz$. As $v_p$ approached $cz$, the bias increased significantly (several standard deviations), and the estimator became unreliable.
  • Exact Method: Remained unbiased even when $v_p \approx cz$. While it exhibited slightly higher variance (uncertainty) than the linear method in the valid regime, its lack of bias made it superior overall.
• Redshift Dependence: At high redshifts ($z > 0.1$), the likelihood becomes insensitive to small redshift perturbations caused by peculiar velocities. Consequently, the estimates at high $z$ become prior-dominated and uninformative. Reliable estimation is currently limited to low redshifts ($z < 0.02$) with current data precision.
• Precision Impact: Simulations showed that future surveys with higher magnitude precision (e.g., $\sigma_m = 0.02$) would significantly reduce the uncertainty in velocity estimates.

B. Application to Pantheon+ Data

• Applied to 1701 SNe Ia from the Pantheon+ sample.
• At low redshifts ($z < 0.02$), the estimated peculiar velocities were consistent with zero, with a standard deviation of $\sigma_{v_p} \sim 300$ km/s.
• At higher redshifts, the method lacked constraining power due to current data limitations, confirming that future large-scale surveys (like LSST and ZTF) are required to extend these measurements.

5. Significance and Conclusion

• Theoretical Robustness: This work establishes a self-consistent, unbiased framework for extracting peculiar velocities from SNe Ia, addressing a major systematic limitation in current cosmological analyses.
• Cosmological Probes: By providing unbiased peculiar velocity measurements, this method enables more accurate tests of gravity on cosmological scales and constraints on dark energy models, particularly distinguishing between background expansion and perturbation growth.
• Future Outlook: While currently limited by data precision at high redshifts, the method is scalable. It is designed to handle the massive datasets expected from upcoming surveys, potentially revolutionizing the use of SNe Ia for mapping the velocity field of the universe.
• Limitations: The method is computationally expensive due to the high dimensionality of the parameter space (one redshift parameter per supernova). Additionally, it currently captures only the line-of-sight component and relies on the assumption that redshift errors are dominated by peculiar motion.

In summary, Upadhyay et al. present a rigorous Bayesian alternative to standard peculiar velocity estimators that eliminates biases associated with linear approximations and fixed cosmologies, offering a path toward more precise cosmological constraints using Type Ia supernovae.