🔭 astrophysics

Wasserstein Distance in Cosmological Structure Formation: An Optimal Transport Perspective

This paper proposes a novel framework for modeling cosmological structure formation by applying optimal transport theory to derive a Wasserstein distance that unifies gravitational mass transport, galaxy formation bias, and shot noise into a single statistical measure linking initial density fields to observed galaxy catalogs.

Original authors: Tsutomu T. Takeuchi (Nagoya University)

Published 2026-03-17

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Tsutomu T. Takeuchi (Nagoya University)

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

Imagine the universe as a giant, invisible ocean of dark matter. In the beginning, this ocean was almost perfectly smooth, like a calm lake. But over billions of years, gravity acted like a giant hand, pulling the water into swirling currents, deep whirlpools, and massive waves. This is how galaxies and galaxy clusters formed.

Traditionally, astronomers have studied this process by taking "snapshots" of the water's height at different points. They measure the amplitude of the waves (how high or low the water is) and calculate statistics like the "Power Spectrum." It's like measuring the average height of the waves in a storm.

But there's a problem: Knowing the height of the waves doesn't tell you how the water moved to get there. Did a gentle breeze push it? Did a massive tsunami crash through? The traditional methods tell us about the state of the water, but not the journey it took to get there.

This paper, by Tsutomu Takeuchi, proposes a new way to look at the universe. Instead of measuring wave heights, he suggests we measure the cost of moving the water. He uses a mathematical tool called Wasserstein Distance, which comes from a field called "Optimal Transport."

Here is the simple breakdown of his idea:

1. The "Moving Furniture" Analogy

Imagine you have a room full of furniture (the initial smooth universe) and you want to rearrange it into a specific messy pattern (the current universe with galaxies).

Old Way (Power Spectrum): You take a photo of the messy room and count how many chairs are in the corner, how many tables are in the middle, etc. You describe the result.
New Way (Wasserstein Distance): You ask, "What is the minimum amount of energy required to move every single piece of furniture from the starting room to the final room?"

In cosmology, "moving furniture" means moving mass (dark matter) from where it started to where it is now. The Wasserstein distance calculates the total "effort" or "cost" of this cosmic rearrangement.

2. The Three Ingredients of the Cosmic Cost

The paper argues that the total "cost" of rearranging the universe isn't just one thing. It's actually a sum of three distinct bills, like a receipt for a complex delivery service:

A. The Gravity Bill (Gravitational Transport)

This is the cost of moving the mass due to gravity. As the universe expands and clumps together, gravity pulls matter around.

The Metaphor: This is the cost of the truck driver actually driving the truck from point A to point B.
The Math: This part is calculated using the "Power Spectrum" (the traditional tool), but interpreted as a measure of how far matter had to travel.

B. The "Galaxy Bias" Bill (Formation Bias)

Here is the tricky part. We don't see the dark matter directly; we see galaxies. Galaxies are like "tracers" or "beacons" that form in the densest parts of the dark matter. However, they don't form perfectly everywhere. They might be "biased" to form only in the biggest clumps.

The Metaphor: Imagine you are trying to track a delivery truck by looking at the mailboxes it stops at. But the mailboxes are only built in the richest neighborhoods. If you only look at the mailboxes, you might think the truck took a different route than it actually did. This "bias" adds a correction cost to our calculation.
The Innovation: The paper shows that this bias creates a specific geometric cost that looks different from standard statistics. It weighs the "clumpiness" of the universe in a new way.

C. The "Pixelation" Bill (Shot Noise)

Finally, we have a limited number of galaxies to look at. We can't see every single star; we only see a finite sample.

The Metaphor: Imagine trying to draw a smooth curve using only a few dots. The more dots you have, the smoother the curve looks. If you have very few dots, the curve looks jagged and "noisy." This "graininess" of the galaxy catalog adds a final cost to the calculation.
The Math: This is called "Shot Noise." It's the error introduced because we are sampling a continuous ocean with a finite number of buckets.

3. The Big Discovery: A New Ruler

The most exciting part of the paper is a specific mathematical finding about how we measure "clumpiness."

Old Ruler: Traditional statistics measure clumpiness by weighing the distance between galaxies with a factor of $r^2$ (volume). It's like measuring the size of a room by its volume.
New Ruler (Wasserstein): This new method weighs the distance with a factor of $r$ (linear distance). It's like measuring the length of a hallway.

Why does this matter?
Because the universe is a dynamic system where things move, measuring the "distance" (linear) is often more physically meaningful than measuring the "volume" (cubic). The Wasserstein distance is sensitive to the geometry of the flow. It tells us not just how clumpy the universe is, but how far the matter had to travel to get there.

Summary

Think of the universe as a giant, evolving puzzle.

Traditional Cosmology counts the pieces and describes the picture on the box.
This Paper calculates the effort it took to move the pieces from the box to the table.

By using Optimal Transport, the author unifies three different physical processes (gravity, galaxy formation, and observation limits) into a single, elegant geometric number. This new "Wasserstein Distance" could help astronomers better understand the history of the universe, separating the effects of gravity from the quirks of how galaxies form, all by measuring the "cost" of the cosmic journey.

1. Problem Statement

Traditional cosmological statistics, such as the power spectrum ( $P(k)$ ) and the two-point correlation function ( $\xi(r)$ ), characterize the amplitude and statistical structure of density fluctuations. However, they do not directly quantify the spatial redistribution of matter that occurs during the formation of large-scale structure.

The paper addresses the need for a statistical framework that treats cosmological structure formation as a transport problem. Specifically, it seeks to quantify the geometric cost of rearranging mass from the initial linear density field to the final observed galaxy catalog, which involves:

Gravitational evolution (mass transport).
Galaxy formation bias (mapping dark matter to galaxies).
Discrete sampling (shot noise from finite galaxy counts).

2. Methodology

The author employs Optimal Transport (OT) theory, specifically the Wasserstein distance (also known as the Earth Mover's Distance), to model the evolution of the universe.

Generative Process Framework: The formation of the observed universe is modeled as a hierarchical generative process:
$\rho_{\text{lin}} \xrightarrow{\text{Gravity}} \rho_{\text{NL}} \xrightarrow{\text{Bias}} \lambda_{\text{gal}} \xrightarrow{\text{Sampling}} \mu_N$
Where $\rho_{\text{lin}}$ is the initial linear density, $\rho_{\text{NL}}$ is the nonlinear density, $\lambda_{\text{gal}}$ is the galaxy formation intensity, and $\mu_N$ is the observed point process (galaxy catalog).
Mathematical Tools:
- Quadratic Wasserstein Distance ( $W_2^2$ ): Defined as the minimal cost to transport mass between two distributions with a squared-distance cost function $c(x,y) = |x-y|^2$ .
- Dynamic Formulation: Utilizing the Benamou-Brenier formulation, the distance is linked to the kinetic energy of a pressureless fluid flow, connecting OT to the Euler-Poisson system governing cosmological evolution.
- Zel'dovich Approximation: In the weakly nonlinear regime, the displacement field is treated as a gradient map, which corresponds to a geodesic in the Wasserstein space.
- Small-Fluctuation Approximation: The density field is linearized ( $\rho = \bar{\rho}(1+\delta)$ ) to derive analytical expressions for the distance.

3. Key Contributions and Results

The paper derives an approximate expression for the squared Wasserstein distance between the initial linear density field and the observed galaxy catalog. The central result is that this distance naturally decomposes into three distinct physical terms:

$\langle W_2^2 \rangle \simeq W_{\text{grav}}^2 + W_{\text{bias}}^2 + W_{\text{samp}}^2$

A. Gravitational Transport Term ( $W_{\text{grav}}^2$ )

Physical Meaning: Represents the mass rearrangement driven by gravitational instability.
Mathematical Form: Under the Zel'dovich approximation, this term equals the mean squared displacement of particles. In Fourier space, it is the integral of the matter power spectrum:
$\langle W_{\text{grav}}^2 \rangle = \frac{1}{2\pi^2} \int_0^\infty P_m(k) \, dk$
Significance: It directly links the Wasserstein distance to the cumulative power spectrum, measuring the total "effort" required to move matter from its initial to final position.

B. Galaxy Formation Bias Term ( $W_{\text{bias}}^2$ )

Physical Meaning: Accounts for the non-linear mapping between the dark matter density and the galaxy distribution (bias).
Mathematical Form: This term appears as a weighted integral of the galaxy correlation function:
$\Delta W_{\text{bias}}^2 \sim \int_0^\infty r \, \xi_{\text{gal}}(r) \, dr$
Key Distinction: Unlike standard statistics where the correlation function is weighted by the volume element ( $r^2$ ), the Wasserstein distance weights it by $r$ . This implies a fundamentally different sensitivity to spatial scales, making it a direct probe of transport distances rather than volume-averaged clustering.

C. Sampling/Shot Noise Term ( $W_{\text{samp}}^2$ )

Physical Meaning: Arises from the discreteness of the observed galaxy catalog (finite number of galaxies).
Mathematical Form: Modeled as a Poisson point process, this term scales inversely with the mean number density ( $\bar{n}$ ) and the smoothing scale ( $R$ ):
$\langle W_{\text{samp}}^2 \rangle \simeq \frac{1}{4\pi^{3/2} \bar{n} R}$
Significance: It quantifies the geometric approximation error introduced by observing a continuous field through a discrete sample.

D. Finite Volume Effects

The paper also analyzes the impact of finite observational volumes (window functions). It shows that finite volumes suppress long-wavelength modes (bulk flows), leading to an underestimation of the true Wasserstein distance. The distance in a finite volume depends on the minimum wave number $k_{\text{min}} = 2\pi/L$ .

4. Significance and Implications

Unified Geometric Framework: The paper provides a unified statistical quantity that simultaneously captures gravitational evolution, galaxy bias, and observational sampling effects. This offers a holistic view of structure formation that traditional two-point statistics cannot provide.
New Sensitivity to Scales: The unique weighting ( $r$ vs. $r^2$ ) in the correlation integral suggests that the Wasserstein distance probes the geometry of mass redistribution differently than the power spectrum. It is more sensitive to the "transport cost" of moving matter over specific distances.
Bridging Theory and Observation: By treating the galaxy catalog as a point process derived from a continuous field, the framework allows for a rigorous comparison between theoretical density fields and observed discrete data using a single geometric metric.
Future Applications: The authors suggest that measuring $W_2$ in N-body simulations and real galaxy surveys could help disentangle the contributions of gravity, bias, and noise, potentially leading to new constraints on cosmological parameters and galaxy formation physics.

In conclusion, this work establishes the Wasserstein distance as a powerful new tool in cosmology, shifting the perspective from analyzing density fluctuations to analyzing the transport geometry of the cosmic web.