🔭 astrophysics

The Manticore-Local Cluster Catalogue: A Posterior Map of Massive Structures in the Nearby Universe

This paper introduces the Manticore-Local Cluster Catalogue, a probabilistic map of 401 massive structures in the nearby Universe derived from Bayesian field-level reconstruction of 2M++ galaxies, which is independently validated by Planck thermal SZ measurements and cross-referenced with X-ray data to reveal systematic mass biases and constrain individual assembly histories.

Original authors: Stuart McAlpine

Published 2026-03-17

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Stuart McAlpine

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 you are trying to solve a massive, three-dimensional jigsaw puzzle of the entire universe, but you only have a few scattered pieces from the "nearby" neighborhood (within a few hundred million light-years). You know the rules of how the pieces fit together (gravity, dark matter, the Big Bang), but you don't have the picture on the box.

This paper, "The Manticore-Local Cluster Catalogue," is like a team of super-sleuths who have used a powerful new detective tool to reconstruct that missing picture. They haven't just guessed; they have created a "Digital Twin" of our cosmic neighborhood.

Here is the breakdown of what they did, using simple analogies:

1. The Detective Tool: "Manticore-Local"

Think of the universe as a giant, invisible web made of dark matter. Galaxies are like fireflies sitting on this web. We can see the fireflies (the galaxies), but we can't see the web itself.

The scientists used a method called Bayesian Field-Level Inference.

The Analogy: Imagine you are in a dark room and you can only see the shadows of furniture cast on the wall. You know the rules of how light works. By looking at the shadows (the galaxies), the computer works backward to figure out exactly what the furniture (the dark matter) looks like, where it is, and how heavy it is.
The Twist: Instead of giving you just one guess, the computer runs the simulation 80 different times. It creates 80 slightly different versions of the universe that all look exactly the same based on the data we have. This is called a "posterior ensemble." It's like asking 80 different detectives to solve the same crime; if they all agree on who the culprit is, you can be very sure they are right.

2. The "Halo Associations": Finding the Clusters

In these 80 different universe versions, the scientists looked for massive clumps of dark matter (called "halos") that hold galaxy clusters together.

The Problem: In one version of the universe, a giant cluster might be here. In another version, it might be slightly to the left.
The Solution: They grouped these 80 versions together. If a cluster appears in the same spot in most of the 80 versions, they call it an "Association."
The Result: They found 401 massive structures (like the Virgo or Coma clusters) that are real, robust, and statistically significant. They didn't just say "there's a cluster here"; they said, "There is a cluster here, and we are 95% sure of its exact location, mass, and speed."

3. The "Time Machine" Effect: Rewinding History

One of the coolest parts of this paper is what they learned about the past.

The Analogy: Imagine you see a finished house. Usually, you can't tell exactly how the bricks were laid or which way the workers walked to build it. But because this "Digital Twin" is built on the laws of physics, the scientists can hit "rewind" and watch the house being built.
The Discovery: When they rewound time to see how these clusters formed, they found that the "Digital Twin" clusters didn't just form randomly. They all started from very specific, tight little patches of space and flowed together in a coordinated way.
The Insight: Even though they only looked at the present day (today's galaxies), the laws of physics forced the universe to have a very specific history. It's like looking at a finished cake and being able to say, "The baker must have mixed the batter in this specific bowl, in this specific order," because any other way would have resulted in a different cake.

4. The "Sniff Test": Did They Get It Right?

How do you know your digital reconstruction is real? You have to check it against reality.

The Hot Gas Test: Galaxy clusters are filled with super-hot gas that emits a specific signal (the Sunyaev-Zel'dovich effect) that distorts the Cosmic Microwave Background (the afterglow of the Big Bang). It's like a heat signature.
The Result: The scientists took their 401 digital clusters and looked at the real sky data from the Planck satellite. Bingo. Where their digital clusters were, the real sky showed a heat signature. This proved their digital twins are sitting in the right places and have the right amount of mass.

5. The "Weighing Scale" Discrepancy

Finally, they compared their digital weights to the weights astronomers usually calculate using X-ray telescopes.

The Finding: Their digital weights were about 1.7 times heavier than the old X-ray estimates.
Why? It turns out the old X-ray methods often underestimate the weight because they assume the gas is calm and still (like a glass of water). But these clusters are often turbulent and chaotic (like a washing machine on the spin cycle). The new method accounts for this chaos better.
The Takeaway: This catalogue acts as a "truth detector." If an astronomer sees a cluster that looks weird compared to this catalogue, they know to investigate that specific object to see if there's something special going on, rather than just assuming it's a statistical fluke.

Why Does This Matter?

This paper gives us a map of the nearby universe that is:

Probabilistic: It admits what it doesn't know (giving error bars).
Dynamic: It shows how things move and merge over time.
Validated: It has been checked against real heat signatures from space.

It's like moving from looking at a blurry, static photo of a city to having a high-definition, 3D simulation where you can zoom in, rewind time, and see exactly how the traffic flows, all while knowing exactly how confident you can be in every detail. This allows scientists to study the universe not just as a collection of statistics, but as a collection of specific, individual objects with unique histories.

1. Problem Statement

Traditional large-scale structure analyses rely on summary statistics (e.g., power spectra, correlation functions) which discard non-Gaussian information and phase data. Consequently, they cannot link specific observed galaxy clusters to their unique simulated counterparts within a forward model. While constrained simulations (e.g., CLUES, ELUCID) have produced observationally anchored catalogues, they often lack a fully quantified Bayesian posterior ensemble, making it difficult to rigorously account for reconstruction uncertainties or perform object-by-object comparisons. There is a need for a high-fidelity, probabilistic map of massive structures in the local Universe ( $z \lesssim 0.05$ ) that provides robust posterior distributions for mass, position, velocity, and formation history.

2. Methodology

Data Source & Inference Framework:

Input: The study utilizes the Manticore-Local posterior ensemble, a state-of-the-art Bayesian field-level reconstruction of the nearby Universe ( $R < 200$ Mpc).
Algorithm: It employs the BORG (Bayesian Origin Reconstruction from Galaxies) algorithm combined with the Manticore model.
Observational Constraint: The inference is conditioned on the 2M++ galaxy catalogue.
Cosmology: The reconstruction assumes a $\Lambda$ CDM model with parameters consistent with DES Year 3 ( $h=0.681, \Omega_m=0.306$ ).
Ensemble: The suite consists of 80 independent posterior resimulations (expanded from the original 50) using the SWIFT $N$ -body code. Each realization contains $1024^3$ dark matter particles.

Catalogue Construction (Halo Associations):

Definition: An "association" is defined as a group of central haloes ( $z=0$ ) from different posterior realizations that cluster in Eulerian space, representing the same inferred galaxy cluster.
Clustering Algorithm: The DBSCAN (Density-Based Spatial Clustering of Applications with Noise) algorithm is applied in 3D comoving coordinates.
- Parameters: Neighborhood radius $\epsilon = 2.75$ Mpc; Minimum neighbors $N_{min} = 10$ .
- Constraint: Each posterior realization contributes at most one member halo to an association to prevent over-merging.
Selection Criteria:
- Mean mass $\langle M_{200} \rangle \geq 10^{14} M_\odot$ .
- Ambiguity rate (frequency of multiple haloes from one realization) $< 5\%$ .
- Minimum membership: At least 20 member haloes across the 80 realizations.
Output: A fiducial catalogue of 401 associations and a strict subset of 162 associations.

Validation Techniques:

Thermal Sunyaev–Zel'dovich (tSZ) Stacking: Using the Planck PR4 NILC Compton- $y$ maps. The team performs aperture photometry (inner radius $1.5 R_{500}$ , outer $5.0 R_{500}$ ) on the association positions to detect coherent tSZ decrements.
Cross-Matching: Comparing catalogue positions and masses against external X-ray catalogues:
- MCXC-II: ROSAT-based, hydrostatic mass estimates.
- eROSITA eRASS1: Weak-lensing calibrated mass estimates.
Evolutionary Pathway Analysis: Tracing progenitor haloes back to $z=5$ in merger trees and comparing the spatial localization of constrained associations against "control" haloes from unconstrained $\Lambda$ CDM simulations.

3. Key Contributions

First Probabilistic Cluster Catalogue: Delivers a publicly available catalogue where every massive structure is represented not as a single point estimate, but as a posterior ensemble with quantified uncertainties in position, mass, velocity, and shape.
Object-Level Digital Twins: Provides specific "digital counterparts" for observed clusters, enabling direct cross-probe comparisons (e.g., matching a specific observed cluster to its simulated merger history) rather than relying on ensemble-averaged scaling relations.
Information Gain Quantification: Demonstrates that $z=0$ observational constraints alone significantly restrict the possible assembly pathways of massive structures, reducing the configuration space of progenitors.
Systematic Bias Flagging: Establishes a framework to identify individual systems with significant mass discrepancies between reconstruction and observation, distinguishing between reconstruction failures and observational systematics.

4. Key Results

Catalogue Properties:

The fiducial catalogue contains 401 associations with $\langle M_{200} \rangle \geq 10^{14} M_\odot$ .
Completeness: Recovery is mass- and distance-dependent. It recovers $\sim20\%$ of all X-ray clusters but $>90\%$ of the most massive systems ( $M_{500} > 5 \times 10^{14} M_\odot$ ). Completeness degrades beyond 150 Mpc due to weakening 2M++ constraints.
Positional Accuracy: Median angular offsets are $\sim0.33^\circ$ (MCXC) and $0.35^\circ$ (eROSITA), corresponding to $\approx 0.4\sigma$ of the posterior uncertainty, indicating well-calibrated errors.

tSZ Validation:

Stacked Planck tSZ measurements at association positions yield significant detections ( $>3\sigma$ ).
The signal follows the expected $M_{500} - Y$ scaling relation (higher mass = stronger decrement), confirming that the associations trace genuine hot gas structures.
A null test using random sky positions yielded zero signal, ruling out map artifacts.

Evolutionary Pathways:

Lagrangian Localization: Progenitor haloes of the associations occupy compact, anisotropic regions of Lagrangian space at $z=5$ , exhibiting coherent infall along preferred directions.
Information Gain: Compared to mass-matched controls in unconstrained simulations, the constrained progenitors occupy 2–5 times smaller initial volumes.
Quantification: The median information gain is $I \approx 1.5$ bits, representing a factor of $\sim3$ reduction in configuration-space volume. This proves that $z=0$ galaxy data alone restricts the range of plausible formation histories.

Mass Scale Comparisons:

MCXC-II Comparison: Manticore-Local masses are systematically higher by a factor of $\sim1.7$ . This aligns with known hydrostatic bias and scaling relation systematics in ROSAT-based masses.
eROSITA Comparison: Manticore-Local masses agree with eROSITA (weak-lensing calibrated) with a median ratio close to unity.
Outliers: A subset of systems shows large mass discrepancies (ratios > 3 or < 0.33). Analysis suggests these are driven by catalogue-specific X-ray systematics (e.g., dynamical state, projection effects) rather than reconstruction failures, as outliers do not show broader posterior uncertainties.

5. Significance

This work represents a paradigm shift in cosmological analysis by moving from statistical population studies to individual object-level validation using Bayesian field-level inference.

Hybrid Analyses: It enables researchers to combine observed sky positions (for stacking) with simulated quantities (merger trees, velocities, dynamical states) for the same physical system.
Bias Diagnosis: By providing a "ground truth" simulation for specific observed clusters, the catalogue allows for the systematic investigation of biases in observational mass proxies (e.g., hydrostatic bias in X-ray vs. weak lensing).
Future Applications: The catalogue serves as a foundation for "Digital Twins" of the local Universe, facilitating tests of galaxy formation physics, hydrostatic equilibrium, and environmental quenching with unprecedented realism.
Community Resource: All data products (positions, masses, velocities, full posterior distributions) are publicly available via cosmictwin.org, fostering community-driven research and cross-correlation studies.

In summary, the Manticore-Local Cluster Catalogue provides a rigorously validated, uncertainty-aware map of the local massive structure, bridging the gap between observational data and theoretical cosmology through a fully probabilistic framework.