🔭 astrophysics

Towards precision cosmology with Voids x CMB correlations (I): Roman-Agora mock catalogs and pipeline validation

This paper introduces and validates a versatile "analog matching" technique to generate multi-purpose mock galaxy catalogs for the Nancy Grace Roman Space Telescope, demonstrating that matching void statistics alongside traditional clustering metrics is essential for accurately constraining galaxy-halo connections and improving precision cosmology with Void-CMB correlations.

Original authors: Mar Pérez Sar, Carlos Hernández Monteagudo, András Kovács, Alice Pisani

Published 2026-05-22

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Mar Pérez Sar, Carlos Hernández Monteagudo, András Kovács, Alice Pisani

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, three-dimensional sponge. Most of the sponge is filled with holes (these are cosmic voids), and the solid parts are where galaxies live, forming a web-like structure. Astronomers want to understand how this sponge was made and how it's stretching over time, which tells us about the invisible forces of dark matter and dark energy.

To do this, they plan to use a new space telescope called Roman (named after Nancy Grace Roman) and combine its data with maps of the Cosmic Microwave Background (CMB)—the leftover "glow" from the Big Bang. When light from the CMB passes through the cosmic sponge, it gets distorted by the gravity of the voids. By studying this distortion, scientists can learn a lot about the universe.

The Problem: The "Recipe" Gap
Before they can trust the real data from the Roman telescope, scientists need to test their methods. They need to know: If we assume the universe works a certain way, what should the data look like?

To answer this, they create mock catalogs. Think of these as "fake universes" or highly detailed video game simulations. They need a simulation that has two things:

A map of galaxies (like the Roman telescope will see).
A map of the CMB distortions caused by those galaxies.

The problem is that most existing simulations have one or the other, but not both working perfectly together. It's like having a recipe for a cake but no oven, or an oven but no recipe.

The Solution: "Analog Matching"
The authors of this paper developed a clever new method to build these fake universes, which they call "Analog Matching."

Imagine you have a photo album of real people (the Roman reference catalog) and a massive database of mannequins (the Agora simulation). Your goal is to dress the mannequins so they look exactly like the people in the photo album.

Instead of building the mannequins from scratch, you use a smart search tool (a KDTree, which is like a super-fast librarian) to find the best match.

You look at a person in the photo album.
You ask the librarian: "Find me the mannequin that has the same weight, the same height, and the same style of clothing."
Once found, you "assign" that mannequin to be the twin of the person.
You do this for everyone in the album, making sure no two people get the exact same mannequin.

By changing the criteria for the search (e.g., matching only by weight, or by weight + height + hair color), the team created different versions of these fake universes to see which matching rules work best.

What They Found
The team tested their "fake universes" to see if they looked like the real thing. They looked at two main things:

Clustering: How do the galaxies group together?
Voids: How big are the empty holes, and how are they shaped?

Here are their main discoveries, translated into everyday terms:

Weight Matters Most: If they matched the mannequins based only on their "weight" (which in the simulation is the mass of the dark matter halo holding the galaxy), the fake universe looked surprisingly similar to the real one. Adding extra details like "hair color" (galaxy type) didn't make it much better for this specific type of galaxy.
The "Environment" Trap: They tried to match based on "where the person lives" (the local environment density). This backfired. It was like trying to match a person in a crowded city to a mannequin in a forest just because they both have a "medium crowd" score. The match was wrong because the underlying maps were different. This taught them that sometimes, simpler matching rules are better.
Voids are the Truth-Tellers: This is the most important finding. You can make a fake universe that looks perfect when you count how many galaxies are in a cluster (2-point statistics). But if you look at the voids (the empty holes), the fake universe might look totally wrong.
- Analogy: Imagine two cakes that look identical from the top. But if you slice them open, one has a perfect sponge texture, and the other is full of air pockets. The "voids" are the slice that reveals the true texture. The paper shows that checking the voids is a much stricter and more sensitive test than just counting galaxies.

Why This Matters
This paper doesn't discover a new planet or a new law of physics. Instead, it builds a better testing ground.

They have created a versatile toolkit (the Roman-Agora mock catalogs) that allows scientists to:

Test their analysis pipelines before the Roman telescope even launches.
Understand how different ways of modeling galaxies affect the results.
Use "voids" as a high-precision ruler to check if their models of the universe are actually correct.

In short, they built a better "flight simulator" for cosmologists. Before they fly the real mission, they can crash the simulator a thousand times to make sure they know exactly what to expect when they finally look at the real sky.

Technical Summary: Towards Precision Cosmology with Voids × CMB Correlations (I): Roman-Agora Mock Catalogs and Pipeline Validation

Problem Statement
The Nancy Grace Roman Space Telescope, in conjunction with upcoming Cosmic Microwave Background (CMB) experiments, will enable high-precision cross-correlations, such as Void × CMB lensing. However, extracting cosmological information from these measurements requires robust modeling of the signal's dependence on cosmology and a rigorous characterization of methodological and mock-dependent systematics. A critical bottleneck is the lack of realistic mock catalogs that integrate both Large-Scale Structure (LSS) galaxy distributions and self-consistent CMB maps. Traditional methods for generating mocks, such as Halo Occupation Distribution (HOD) or Abundance Matching (SHAM), often rely on simplifying assumptions that may fail in the non-linear regime or under specific observational conditions. Furthermore, while mocks are typically validated against two-point galaxy clustering statistics, they are rarely tested against higher-order statistics like cosmic voids, which provide a sensitive probe of the matter distribution's topology and galaxy-halo connections.

Methodology
The authors propose a novel "analog matching" technique to generate multi-purpose mock galaxy catalogs designed to approximate the Roman survey characteristics using the Agora simulation. Unlike traditional methods that construct catalogs from scratch, this approach translates a reference mock catalog (specifically the Zhai et al. 2021 Roman catalog) into a new simulation (Agora) via a nearest-neighbor search.

The workflow involves:

Reference and Target Data: Using the Zhai et al. (2021) Roman mock (based on the UNIT simulation and Galacticus SAM) as the reference and the Agora simulation (based on the MDPL2 N-body simulation) as the target. Agora provides a coherent suite of secondary CMB anisotropy maps (lensing, tSZ, kSZ, CIB).
Preprocessing: The Agora halo catalog is augmented with baryonic properties (stellar mass $M_*$ , specific star formation rate sSFR) using the UniverseMachine model. The catalog is then subsampled to match the Roman survey footprint and number density. Environmental parameters (density contrast $\delta_5$ within $5 h^{-1}$ Mpc) are computed for both datasets.
Analog Matching: A KDTree-based nearest-neighbor search is performed in a multi-dimensional parameter space to assign an Agora halo counterpart to each Roman galaxy. The parameter space varies across four mock families:
- N: Simple number density matching (no analog matching).
- NM: Matching based on Halo Mass ( $M_h$ ) alone.
- NMbenv: Matching based on Halo Mass and Environment ( $\delta_5$ ).
- NMbELGs: Matching based on Halo Mass and Galaxy-type indicators ( $M_*$ , sSFR).
Void Finding: The resulting catalogs are analyzed using two void finders: a 2D density-based algorithm (Sánchez et al. 2017) and a 3D geometrical algorithm (REVOLVER/VOID).
Validation: The mocks are validated against the reference catalog using one-point statistics (void redshift distribution, size function, minimum density) and two-point statistics (galaxy angular power spectrum $C_\ell$ , void-galaxy cross-correlation function $\xi_{vg}$ ).

Key Contributions

Roman-Agora Mock Suite: The construction of a versatile set of mock catalogs that combine Roman-like emission-line galaxy (ELG) statistics with self-consistent CMB maps, specifically designed for Void × CMB studies.
Analog Matching Algorithm: A generalizable method for generating analog catalogs by mapping reference objects to simulation halos via multi-dimensional parameter space matching, avoiding the need for full hydrodynamical simulations or complex semi-analytic modeling from scratch.
Void-Based Validation: A systematic demonstration that void statistics serve as an independent and sensitive test of galaxy-halo modeling, distinct from traditional two-point clustering measures.

Results

Clustering vs. Voids: The study finds that reproducing two-dimensional galaxy clustering (via the angular power spectrum) does not guarantee the consistency of void properties. For instance, the N (simple subsampling) and NMbenv (mass + environment) families showed significant discrepancies in void statistics compared to the reference, despite varying levels of clustering agreement.
Optimal Matching Strategy: Analog matching based on halo mass alone (NM) or halo mass combined with galaxy-type indicators (NMbELGs) successfully reproduces the expected Roman ELG statistics and void properties.
Limitations of Environmental Matching: Including environmental parameters ( $\delta_5$ ) in the matching process (NMbenv) degraded performance when the underlying dark matter fields of the reference and target simulations were not perfectly aligned or when the tracer bias differed. This suggests that for ELGs, which are primarily determined by current halo mass, secondary environmental biases are less critical than for other galaxy types.
Sensitivity of Voids: 3D voids were found to be more sensitive to differences in mock construction than 2D voids. The results highlight that void measurements provide a powerful diagnostic for assessing the accuracy of mock catalogs beyond the matter power spectrum.

Significance and Claims
The paper positions this work as a foundational step toward precision cosmology with Roman and CMB data. The authors claim that their Roman-Agora mock framework offers a "versatile tool" for the community to test how modeling choices impact cosmological observables, specifically Void × CMB lensing signals.

The significance lies in the ability to:

Validate Pipelines: Provide a controlled testbed to distinguish between methodological artifacts and genuine physical signals in future Void × CMB analyses.
Assess Systematics: Systematically test how different galaxy-halo prescriptions influence large-scale structure statistics and CMB-related observables.
General Applicability: The analog matching technique is presented as fully general, applicable to any combination of simulation and reference catalog provided their parameter spaces are specified.

The authors explicitly state that this paper focuses on the construction and validation of the mocks themselves. A companion work is noted to use these data products to investigate the specific impact of these modeling choices on the forecasted detectability of the Void × CMB lensing signal with Roman and upcoming CMB surveys. The paper does not claim to resolve current cosmological tensions but rather to provide the necessary tools to ensure that future measurements are not biased by unaccounted systematics in mock generation.

More like this