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

arXiv:2603.06469 · astro-ph.CO · Submitted 2026-03-06 · Read on arXiv

Listen

Transcript

Introduction to the show: ident: Astrophysics Radio. Generated commentary on the latest astrophysics papers.

Vera: I'm Vera, and with me are Jocelyn and Subrahmanyan, guest researcher.

Jocelyn: Today's paper: "Unbiased Bayesian Inference of Peculiar Motions of Galaxies from Type Ia Supernovae Observations".

Vera: Peculiar motions of galaxies are essential cosmological probes that trace structure growth and test gravity, but their direct observation is complicated by systematic uncertainties in distance measurements.

Jocelyn: First, who's behind it and why it matters.

Paper summary: Vera: So we're looking at this paper now titled "Unbiased Bayesian Inference of Peculiar Motions of Galaxies from Type Ia Supernovae Observations." The main idea is that peculiar motions are essential for tracing how structures grow and testing gravity, but getting those velocities directly is really hard because you need super precise distance measurements. This research presents a Bayesian way to estimate the radial component of peculiar velocities using Type Ia supernovae observations, aiming to avoid biases that come from wrong cosmological assumptions or using simple local linearity approximations.

Jocelyn: I agree, Vera. It sounds like they're tackling a big systematic uncertainty in how we measure these motions. What's the core claim of this paper? What exactly are they trying to prove with this approach?

Subrahmanyan: They are proposing a method that estimates peculiar velocities from SNe Ia data by treating redshifts as free parameters and inferring their distributions through MCMC sampling. This is significant because it's designed to yield unbiased estimates even when the peculiar velocities you're measuring are comparable to the Hubble flow.

Vera: Exactly, Subrahmanyan. The paper claims this Bayesian approach offers an unbiased way to estimate these radial velocities by relying only on the background cosmological model and the precision of the SNe Ia data itself, rather than other estimators that rely on local linearity assumptions. It avoids those biases that pop up when you assume things are perfectly linear or when you use a simple Hubble residual method.

Jocelyn: That sounds very robust for observational cosmology. So, if I understand correctly, they're not just looking at the data and plugging in numbers; they're building a statistical framework to pull out the velocity information in a way that accounts for the errors inherent in both the magnitude and redshift measurements.

Subrahmanyan: That’s right. The methodology involves formulating the fitting of the magnitude–redshift relation as an "errors-in-variables model," where peculiar motions mess up both the observed magnitudes and the redshifts, which they account for in their likelihood function, generalized to include supernova magnitude covariance (MNRAS zero one–eleven (two thousand twenty-six) Preprint nine March two thousand twenty-six Compiled using MNRAS LATEX style file v3 point 3 Unbiased Bayesian Inference of Peculiar Motions of Galaxies from Type Ia Supernovae Observations) <ref:2603.06469#pg0,MNRAS 000, 1–11 (2026) Preprint 9 March 2026 Compiled using MNRAS>. They then use Bayes’ theorem to derive the posterior distribution, allowing them to estimate cosmological parameters along with the true redshifts that are corrupted by these peculiar velocities.

Vera: That sounds mathematically intense, but it makes sense given the complexity of the data they're working with. I wonder how this general MCMC method compares to what we usually do in practice when we try to measure these motions.

Paper summary: Jocelyn: The paper specifically contrasts their Bayesian approach against the standard estimator that relies on a local linear approximation of the magnitude–redshift relation, which assumes Gaussian redshift uncertainties caused by peculiar velocities. They show that this simpler linear method is only valid when the peculiar velocities are very small compared to the Hubble flow, or v p c z.

Subrahmanyan: And they point out a key issue with that standard linear approximation: it suffers from increasing bias at larger velocities, and using an incorrect cosmology on top of that introduces further systematic bias (MNRAS zero one–eleven (two thousand twenty-six) Preprint nine March two thousand twenty-six Compiled using MNRAS LATEX style file v3 point 3 Unbiased Bayesian Inference of Peculiar Motions of Galaxies from Type Ia Supernovae Observations) <ref:2603.06469#pg0,MNRAS 000, 1–11 (2026) Preprint 9 March 2026 Compiled using MNRAS>. The general MCMC method, however, remains consistent with the true peculiar velocities even when v true about c z.

Vera: So, the main advantage they're highlighting is that this general method doesn't suffer from that increasing bias at larger peculiar velocities. They validate this by running simulations using data mimicking current and upcoming survey precision, showing that their estimates are statistically consistent with the true values across all redshifts.

Jocelyn: That simulation validation is crucial for building confidence in the results when we try to apply it to real data from surveys like Pantheon+. But what about where these results actually hold up when we look at real observational constraints?

Subrahmanyan: The application of this method to the Pantheon+ sample shows that at low redshifts, specifically z < zero point zero two, the estimates are consistent with zero peculiar velocity, with a standard deviation of sigma vp about three hundred km/s. However, they also note a limitation: at higher redshifts, the method loses constraining power because of the limited precision of current data and reduced sensitivity to peculiar velocities.

Vera: That's a clear limitation they state—the method becomes effectively prior-dominated and uninformative at higher redshifts because the data just isn't precise enough there. I wonder if this means we can only really use this technique for very low-redshift samples with current technology.

Jocelyn: It suggests that while the framework is robust, its practical utility right now is limited to the low-redshift regime where our SNe Ia data has high precision. It also mentions that these estimates depend on sigma zi, which they want to measure, linking the velocity estimation back to another parameter we are trying to constrain.

Subrahmanyan: The paper concludes by comparing the two estimators: for small peculiar velocities, both methods track the ideal relation accurately, but when true velocities increase, the general method stays consistent with the true values even though its uncertainties are slightly larger than those of the linear method.

Paper summary: Vera: It really paints a picture of a trade-off they're making: sacrificing a bit of variance for freedom from bias as velocities increase. I think this is important because it gives us a more reliable way to probe structure growth when things get more dynamic.

Jocelyn: So, if we take what we've seen with the Pantheon+ sample and the simulation checks, what's the bigger picture implication of this unbiased Bayesian inference method for our understanding of dark energy and gravity?

Subrahmanyan: The paper emphasizes that peculiar velocities provide a unique probe of small-scale physics in the late universe because features on small scales can't be constrained by the CMB, which only probes the early universe and large scales (MNRAS zero one–eleven (two thousand twenty-six) Preprint nine March two thousand twenty-six Compiled using MNRAS LATEX style file v3 point 3 Unbiased Bayesian Inference of Peculiar Motions of Galaxies from Type Ia Supernovae Observations) <ref:2603.06469#pg0,MNRAS 000, 1–11 (2026) Preprint 9 March 2026 Compiled using MNRAS>. This method directly addresses those small-scale physics questions.

Vera: That connects the SNe Ia data we're seeing today to the actual gravitational dynamics happening in the late universe, which is a really concrete piece of information for us as observational astronomers. It moves us beyond just looking at distances and into mapping out how matter is moving around.

Jocelyn: It gives us a tool that accounts for the messy reality of peculiar motions without relying on potentially flawed assumptions about how those motions relate to the background cosmology, which is what this paper aims to achieve with its unbiased approach.

Subrahmanyan: The future direction they suggest is extending this applicability to larger samples from upcoming surveys like LSST and ZTF, and incorporating more realistic peculiar-velocity fields from N-body simulations, especially on non-linear scales. That’s where we can really test dark energy and gravity models with greater confidence using this kind of data.

Vera: It sounds like the path forward is to feed this robust method with the massive amounts of data coming from future surveys, which is exactly what observational astronomy needs to keep pushing these tests on cosmological scales.

Jocelyn: So, in short, this paper presents a statistically rigorous way to measure peculiar velocities from SNe Ia that minimizes systematic errors related to local linearity and cosmology assumptions while providing consistent results even at higher velocities compared to simpler estimators.

Subrahmanyan: Indeed, the title "Unbiased Bayesian Inference of Peculiar Motions of Galaxies from Type Ia Supernovae Observations" points directly toward its contribution: providing a statistically unbiased inference mechanism for these motions using SNe Ia as the primary data source.

Conclusion: Vera: So we've been diving deep into how this paper uses Type Ia supernovae to estimate peculiar velocities without those pesky biases, and now we need to wrap up by really talking about what this whole effort means for us.

Jocelyn: I think that title, "Unbiased Bayesian Inference of Peculiar Motions of Galaxies from Type Ia Supernovae Observations," really captures the core strength of the work, especially how they tackle those systematic errors we always worry about in distance measurements.

Subrahmanyan: It’s a method that moves away from simpler linear approximations and instead uses a full Bayesian framework to get a more honest picture of how galaxies are actually moving in space.

Vera: Exactly, and the authors they're talking about really put together this complex statistical machinery to handle the messy reality of these motions.

Jocelyn: When we boil it down for our listeners, it means we now have a statistically sounder way to map out the large-scale structure of the universe by looking at how galaxies are actually streaming, not just how far away they appear.

Subrahmanyan: From a theoretical standpoint, this gives us a much better handle on how matter clumps together on smaller scales, which directly impacts our models of dark energy and gravity.

Vera: It’s really about getting past the limitations of older methods so we can see what's happening in the late universe more clearly with these supernova observations.

Jocelyn: That robustness is key; they show that this approach works even when peculiar velocities get quite significant, which is where other estimators tend to start tripping up.

Subrahmanyan: The implication for cosmology is that we can finally use these distance indicators to probe the growth of structure in a way that's free from the assumptions about the distribution of velocity errors.

Vera: It’s a big step forward because it means our constraints on cosmological models won't be artificially skewed by those kinds of velocity biases anymore.

Jocelyn: And as we look ahead, this technique is going to be incredibly useful when we start using more precise data from surveys like LSST and ZTF to map out these velocities across much larger areas.

Subrahmanyan: That's what the future work points toward; extending this method to those bigger datasets will allow us to test gravity models on scales that are currently just out of reach.

Ujjwal Upadhyay, Tarun Deep Saini, Shiv K. Sethi

Department of Physics, Indian Institute of Science · Astronomy & Astrophysics Group, Raman Research Institute

astro-ph.CO

Submitted: 2026-03-06

Updated: 2026-03-06

Comments: 12 pages, 10 figures, 1 table. Comments are welcome

Journal ref: Mon Not R Astron Soc (2026)

DOI: 10.1093/mnras/stag1752

License: http://creativecommons.org/licenses/by/4.0/

Importance score: 73/100

The gist: Peculiar motions of galaxies are essential cosmological probes that trace structure growth and test gravity, but their direct observation is complicated by systematic uncertainties in distance

Key concepts

Bayesian Approach: A statistical framework that uses Bayes' theorem to update beliefs about unknown parameters (like true redshifts) as new data (supernova magnitudes and observed redshifts) becomes available. It allows for the inference of a probability distribution rather than a single point estimate.
Errors-in-Variables Model: This model treats peculiar motions as errors affecting both the measured magnitude and the redshift. Instead of assuming errors are only in one variable, this approach accounts for how galaxy motion corrupts both measurements simultaneously, leading to more accurate velocity estimates.
Linear Approximation: A simplified method where researchers assume that peculiar velocities are very small compared to the Hubble flow. This approximation is fast but fails when peculiar velocities become large, leading to increasing systematic bias in the calculated velocity estimates.
MCMC Sampling: Markov Chain Monte Carlo is a computational technique used here to explore complex parameter spaces. It generates a chain of samples that represent the posterior distribution of parameters (like true redshifts), allowing researchers to draw robust conclusions about the uncertainties and true values.

Terminology

Bayesian Inference — 15× in this paper · explained in 5 episodes
Trace — 8× in this paper · explained in 2 episodes
Regularization — 5× in this paper · explained in 3 episodes
Clustering — 3× in this paper · explained in 2 episodes
Core — 3× in this paper · explained in 2 episodes
Robustness — 3× in this paper · explained in 5 episodes
Systematic Bias — 3× in this paper · explained in 3 episodes
Calibration — 2× in this paper · explained in 7 episodes
Consistency — 2× in this paper · explained in 3 episodes
Matter Power Spectrum — 2× in this paper · explained in 2 episodes
N-body simulations — 2× in this paper · explained in 2 episodes
Bayesian Framework — 1× in this paper · explained in 3 episodes
Generalization — 1× in this paper · explained in 13 episodes

Summary

Peculiar motions of galaxies are essential cosmological probes that trace structure growth and test gravity, but their direct observation is complicated by systematic uncertainties in distance measurements. This research presents an unbiased Bayesian approach to estimating the radial component of peculiar velocities from Type Ia supernovae (SNe Ia) observations, a method that avoids biases arising from incorrect cosmological assumptions or the local linearity approximation inherent in standard estimators.

The Gist

The paper introduces a Bayesian method to estimate the line-of-sight peculiar velocities of host galaxies using SNe Ia data, treating redshifts as free parameters and inferring their posterior distributions through MCMC sampling, which yields unbiased estimates even when peculiar velocities are comparable to the Hubble flow.

Methodology and Framework

The core methodology involves treating the fitting of the magnitude–redshift relation as an errors-in-variables model where peculiar motions contribute to errors in both dependent and independent variables. The framework utilizes a general MCMC method designed for fitting such models, allowing for joint estimation of cosmological parameters (like H0 and ΩM) and true redshifts.

The likelihood function is formulated to account for uncertainties in both the observed magnitudes and the redshifts, generalized to include supernova magnitude covariance (Equation 6).
The posterior distribution is derived using Bayes’ theorem, leading to a complex expression that allows for the estimation of cosmological parameters along with the true redshifts, which are assumed to be corrupted by peculiar velocities.
The relation between redshift errors and peculiar velocity is used: Δz = (v p/c) (1 + z).

Linear Approximation vs. Exact Method

The paper compares the general Bayesian method against the standard estimator based on local linearity of the magnitude–redshift relation, which assumes Gaussian distribution for redshift errors.

. The standard linear approximation is valid only when "peculiar velocities are very small compared to the Hubble flow (v p << c z)," and it suffers from bias when this condition breaks down. For larger peculiar velocities, the linear method begins to overestimate the velocities, and using an incorrect cosmology introduces further systematic bias. 4.1.2 shows that while the linear method exhibits smaller variance, it suffers from increasing bias at larger velocities, whereas the MCMC-based general method remains consistent with the true peculiar velocities.

Validation and Results

The method is validated using simulated supernova data mimicking current and upcoming survey precision.

The simulation involves generating a random Gaussian field in real space to represent the matter density contrast, shaping it in Fourier space using the matter power spectrum, computing the peculiar velocity field, and taking an inverse Fourier transform to obtain the real-space velocity field.
In validation with simulated data, the estimates are statistically consistent with the true values at all redshifts.
The results show that the general method remains statistically unbiased within a 68% credible region, even for v true ≈ c z, contrasting sharply with the linear method which becomes significantly biased in this regime.

Application to Data and Limitations

The analysis is applied to the Pantheon+ sample of SNe Ia.

. At low redshifts (z < 0.02), the estimates are consistent with zero, with a standard deviation of σ vp ∼ 300 km/s. However, at higher redshifts, the method loses constraining power due to the limited precision of current data and the reduced sensitivity of the likelihood to peculiar velocities, leading to estimates that become effectively prior-dominated and uninformative. The paper concludes that while this approach is robust, it can only be reliably estimated at very low redshifts with present data. Furthermore, it is noted that the estimates depend on the values of σ zi, which is what we want to measure.

Comparison of Estimator Performance

The comparison between the two methods highlights a trade-off between bias and variance.

Figure 4 demonstrates that for small peculiar velocities, both methods lie along the ideal relation, indicating accurate recovery under the linear approximation.
When true velocities increase, the general method remains consistent with the true peculiar velocities, although with slightly larger uncertainties than those obtained with the linear method. This suggests that while the linear estimator has smaller variance, it suffers from increasing bias at larger velocities.
The general method is free from bias due to incorrect cosmology and avoids the assumption of Gaussianity of the peculiar velocity distribution and therefore works for both random and coherent peculiar motion.

Future Directions

The paper suggests that future work should focus on extending the applicability to larger samples expected from upcoming surveys (LSST and ZTF) and incorporating more realistic peculiar-velocity fields from N-body simulations, especially on non-linear scales, to better test dark energy and gravity models. The method is also noted as being straightforward to implement compared to galaxy survey–based methods.

**Appendix A:

Improvements for AI systems

Based on the provided scientific paper, here are specific improvements that can be made to AI systems, categorized by the capabilities they could gain:

The core improvement stems from developing a robust framework for estimating physical quantities (peculiar velocities) in the presence of complex observational noise and systematic uncertainties (like wrong cosmological models). This capability is directly transferable to any AI system requiring high-precision inference under uncertainty.

Here are the specific improvements and resulting AI capabilities:

A superior, unbiased estimator for latent variables (like peculiar velocities) in complex, non-linear systems, even when fundamental physical assumptions (like local linearity) break down.
The ability to perform joint inference between cosmological parameters and systematic observational errors simultaneously without introducing bias from incorrect assumptions about the underlying model (cosmology).
A method for self-consistent parameter estimation that avoids reliance on potentially inaccurate prior assumptions by letting the data determine the posterior distribution, especially in high-dimensional spaces where many parameters exist.

Specific AI System Improvements:

Inference Engine for High-Dimensional Observational Data (e.g., Large-Scale Surveys):
Robust Model Calibration Under Systematic Uncertainty:
Non-Linear Dynamics Simulation and Prediction:

Specific Capabilities Gained by the Improved AI System:

Inference Engine for High-Dimensional Observational Data:
Robust Model Calibration Under Systematic Uncertainty:
Non-Linear Dynamics Simulation and Prediction:

Abstract

The peculiar motions of galaxies are powerful cosmological probes that trace the growth of structures and the distribution of matter in the universe, providing a means to investigate the nature of dark energy and test gravity on cosmological scales. However, their direct observation is extremely challenging, as it requires independent and precise distance measurements to galaxies. We present a Bayesian approach to estimate the radial component of peculiar velocities of galaxies hosting Type Ia supernovae (SNe Ia), relying solely on the background cosmological model and the precision of the SNe Ia data. Unlike other peculiar velocity estimators based on Hubble residuals, our method does not assume local linearity of the magnitude-redshift relation or a fixed cosmology, making it unbiased even for large peculiar velocities and self-consistently avoiding bias due to a wrong cosmology. We validate our method using simulated supernova data with the precision of current and upcoming surveys, and further compare it with the linearized estimator to test its efficacy. We show that our estimator has lower bias than the standard estimator and remains consistent even for larger values of v p/cz. We also present a Bayesian derivation for the linearized estimator generalized to include the supernova magnitude covariance.

Sources

Reconstruction of large-scale peculiar velocity fields