🔭 astrophysics

Painting a full radio sky -- Empirical mock catalogues with multiple source populations for future radio surveys

This paper presents a modular algorithm for generating full-sky, empirical mock catalogues containing over a billion radio sources across multiple populations, designed to support the preparation and scientific exploitation of upcoming deep radio surveys like those from the SKAO.

Original authors: Tommaso Ronconi, Anna Bonaldi, Marta Spinelli, Ivano Baronchelli, Meriem Behiri, Matteo Calabrese, Carmelita Carbone, Marika Giulietti, Andrea Lapi, Marcella Massardi

Published 2026-03-27

📖 6 min read🧠 Deep dive

CC BY 4.0

Original authors: Tommaso Ronconi, Anna Bonaldi, Marta Spinelli, Ivano Baronchelli, Meriem Behiri, Matteo Calabrese, Carmelita Carbone, Marika Giulietti, Andrea Lapi, Marcella Massardi

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 an architect about to build a massive, futuristic city. Before you lay a single brick, you need to know how the city will behave: where traffic jams will form, how the population will grow, and how different neighborhoods will interact. You can't just guess; you need a perfectly detailed, virtual prototype of the entire city to test your plans.

This paper is about building that prototype for the universe, specifically for the radio sky.

The Big Picture: Why Do We Need This?

Astronomers are about to launch the SKA (Square Kilometre Array), a giant radio telescope that will be so powerful it can see billions of faint radio signals from across the universe. It's like upgrading from a pair of binoculars to a telescope that can read a newspaper from a mile away.

But before the telescope turns on, scientists need to know:

What will the data look like?
Will our computer programs be able to handle that much information?
How do we tell the difference between a real galaxy and a glitch in the software?

To answer these questions, they need Mock Catalogues. Think of these as "fake" but incredibly realistic lists of millions of radio sources (galaxies, black holes, etc.) that scientists can use to practice and test their tools.

The Problem: The Universe is a Messy Cocktail

The radio sky isn't just one thing. It's a chaotic mix of:

Active Galactic Nuclei (AGN): Super-massive black holes eating gas and shooting out powerful radio beams (like a lighthouse).
Star-Forming Galaxies (SFG): Galaxies like our Milky Way, making new stars and glowing in radio waves (like a city at night).
HI Galaxies: Galaxies filled with neutral hydrogen gas, the raw fuel for making stars (like the gas stations of the universe).

Previous simulations were like making a model city where you only had houses, or only had shops, but never both together. They also often didn't cover the whole sky or didn't look realistic enough for the new, super-sensitive telescopes.

The Solution: A Modular "Lego" Kit

The authors (led by T. Ronconi) built a new, flexible system to generate these fake universes. They call it a modular pipeline, which is like a high-tech Lego set where you can snap different pieces together.

Here is how their "recipe" works, using a simple analogy:

1. The Skeleton: The Dark Matter Light-Cone

First, they need a skeleton to hang everything on. They used a super-computer simulation called DEMNUni.

Analogy: Imagine a giant, invisible 3D spiderweb made of invisible "dark matter" that stretches across the entire universe. This web has knots (halos) where galaxies like to live. This simulation provides the map of where these knots are.

2. The Ingredients: The T-RECS Library

Next, they need the actual "stuff" to put on the web. They used a tool called T-RECS.

Analogy: This is like a massive, pre-made inventory list of different types of "radio stars." It has billions of entries for black holes, star-forming galaxies, and hydrogen clouds, with realistic rules about how bright they are and how they change over time.

3. The Glue: The SCAMPy Algorithm

This is the magic step. How do you decide which "radio star" from the inventory list goes into which "knot" on the spiderweb?

Analogy: Imagine you have a crowd of people (the galaxies) and a crowd of houses (the dark matter knots). You need to match them up.
- Step A (HOD): You decide the rules. "Only big houses can hold a black hole," or "Small houses can only hold a small star."
- Step B (SHAM): You start matching them up. You take the biggest houses and give them the most powerful black holes, the medium houses get medium stars, and so on.
- The Twist: They do this for all three types of galaxies (black holes, stars, and gas) at the same time, on the same spiderweb. This means if a black hole and a star-forming galaxy are neighbors in the fake universe, they are neighbors because of the physics of the web, not just random chance.

The Two Products: Shallow vs. Deep

The team released two versions of their fake universe:

The Shallow Catalogue (The "Current Reality" Model):
- What it is: A simulation that matches exactly what we can see right now with current telescopes.
- Use: It's like a practice run. Scientists use it to check if their current tools work correctly. It contains about 260 million sources.
The Deep Catalogue (The "Future Vision" Model):
- What it is: A simulation that pushes the limits, predicting what the SKA telescope will see in the future. It goes much deeper and fainter.
- Use: It's like a crystal ball. It helps scientists prepare for the flood of data coming from the new telescope. It contains over 1 billion sources!

Why Is This a Big Deal?

It's a "Full Sky" Map: Previous models were like looking at a single room in a house. This model covers the entire universe (4π steradians).
It's Consistent: Because they put all the different types of galaxies on the same "spiderweb," the relationships between them (like how a black hole might affect a nearby star-forming galaxy) are naturally consistent.
It's Flexible: If new data comes out tomorrow, or if we want to test a different theory of the universe, they can just swap out one "Lego piece" (the simulation or the matching rules) without rebuilding the whole thing.

The Bottom Line

This paper provides the ultimate training ground for the next generation of radio astronomers. By creating a billion-source, full-sky, realistic simulation, they are giving scientists a safe place to make mistakes, test their software, and learn how to interpret the incredible data that the SKA telescope will soon bring home. It's the difference between trying to learn to fly by jumping off a cliff, and learning to fly in a high-tech flight simulator first.

1. Problem Statement

Upcoming radio astronomy facilities, particularly the Square Kilometre Array Observatory (SKAO), will conduct surveys with unprecedented depth and sky coverage. To prepare for these surveys, researchers require realistic mock catalogues that can:

Forecast survey performance and optimize observational strategies.
Validate data analysis pipelines and quantify systematic uncertainties.
Support multi-tracer and multi-wavelength synergy studies.

Existing simulation approaches have limitations:

Hydrodynamical simulations: Physically motivated but limited in volume and sky coverage.
Semi-analytic models: Flexible but rely on specific physical prescriptions that are hard to validate across all observables.
Empirical simulations: Efficient at reproducing statistics but often focus on single populations or lack a consistent embedding within a cosmological light-cone.

There is a critical need for a modular, full-sky ( $4\pi$ sr) simulation framework that simultaneously models multiple radio populations (continuum and line emission) within a single consistent large-scale structure (LSS) realization, using minimal physical assumptions to avoid theoretical bias.

2. Methodology

The authors present a modular pipeline that combines three distinct components to generate self-consistent mock catalogues:

A. The Dark Matter Backbone (DEMNUni)

Source: The "Dark Energy and Massive Neutrino Universe" (DEMNUni) N-body simulation suite.
Configuration: Uses the high-resolution (HR) run of a flat $\Lambda$ CDM cosmology (Planck 2013 parameters) with a box size of $(500 \, \text{cMpc}/h)^3$ and $2048^3$ CDM particles.
Light-Cone Construction: A full-sky light-cone is constructed from redshift $z=0$ to $z \approx 8$ by tiling snapshots. This identifies $\sim 39$ billion haloes and $\sim 44$ billion sub-haloes using Friends-of-Friends (FoF) and SUBFIND algorithms.

B. Empirical Source Sampling (T-RECS)

Source: The Tiered Radio Extragalactic Continuum Simulations (T-RECS) framework.
Populations:
- Radio Continuum: Models Radio Loud Active Galactic Nuclei (RL AGN) and Star-Forming Galaxies (SFGs) based on evolutionary models and luminosity functions.
- HI Line Emission: Models neutral hydrogen (HI) galaxies based on the $z=0$ HI mass function and redshift evolution constraints.
Role: Generates statistical properties (fluxes, redshifts, spectral indices) of sources based on current observational constraints.

C. Galaxy-Halo Assignment (SCAMPy)

Algorithm: The Sub-halo Clustering and Abundance Matching in Python (SCAMPy) package links the T-RECS sources to the DEMNUni sub-haloes.
Two-Step Process:
1. Halo Occupation Distribution (HOD): Selects which haloes/sub-haloes host a specific population based on clustering properties (calibrated via the Halo Model). This filters the DM light-cone.
2. Sub-Halo Abundance Matching (SHAM): Probabilistically assigns specific mock sources to the selected sub-haloes by matching the cumulative distribution of sub-halo masses to the cumulative distribution of source luminosities (or HI masses).
Modularity: The pipeline allows independent modification of the DM simulation, the empirical sampler, or the assignment scheme.

3. Key Contributions

The authors release two public mock catalogues covering the full sky ( $f_{sky}=1$ ) down to $z=5$ :

Shallow Catalogue:
- Constraint: Strictly calibrated against existing observational data (VLA-COSMOS for continuum, ALFALFA for HI).
- Method: HOD parameters are tuned to maximize a likelihood function including both number density ( $n_g$ ) and the two-point correlation function ( $\xi_r$ ).
- Flux Limits: $S_{1.4\text{GHz}} \sim 8 \times 10^{-5}$ Jy (continuum) and $S_{21} \sim 2$ Jy Hz (HI).
- Size: $\sim 260$ million sources.
Deep Catalogue:
- Constraint: Extends the model via controlled extrapolation to match future SKAO sensitivities.
- Method: HOD parameters are tuned solely to reproduce the expected source number density evolution (clustering term omitted in the likelihood).
- Flux Limits: $S_{1.4\text{GHz}} \sim 4 \times 10^{-5}$ Jy (continuum) and $S_{21} \sim 0.3$ Jy Hz (HI).
- Size: $> 1$ billion sources.

Unique Feature: Both catalogues contain cross-population counterparts (AGN, SFG, and HI galaxies) residing in the same underlying dark matter structure. This allows for joint analyses and cross-correlation studies without assuming specific physical links between populations a priori.

4. Results and Validation

The paper validates the simulations by analyzing their statistical properties:

Clustering: The Angular Two-Point Correlation Function (A2PCF) of the shallow catalogue matches the injected HOD predictions for both Continuum Galaxies (CoG) and HI Galaxies (HIG) across large and small scales.
Completeness:
- The shallow catalogue achieves 90% completeness at $S_{1.4\text{GHz}} \approx 0.08$ mJy.
- The deep catalogue achieves 90% completeness at $S_{1.4\text{GHz}} \approx 0.04$ mJy.
- HI completeness is validated against line-width bins, showing a 50% completeness limit of 2 Jy Hz (shallow) and 0.3 Jy Hz (deep).
Emergent Relations (Cross-Catalogue):
- Galaxy-Halo Connection: The simulations reproduce expected trends where AGNs dominate high-mass/high-flux tails, while SFGs populate lower masses.
- HI Mass vs. Halo Mass: The derived relations align with observational data (e.g., Guo et al. 2020, Rhee et al. 2022) and hydrodynamical simulations, validating the empirical approach.
- HI Mass vs. Star Formation Rate (SFR): An emergent correlation is observed between HI mass and SFR. While consistent with literature at low redshift ( $z \lesssim 0.1$ ), deviations appear at higher redshifts, attributed to the extrapolation of the HI mass function model.
Velocity Consistency: The maximum circular velocities ( $v_{max}$ ) derived from the N-body simulation sub-haloes are consistent with those predicted by the T-RECS empirical relations, confirming the validity of linking observables to the DM backbone.

5. Significance

Survey Preparation: Provides a robust tool for forecasting SKAO and precursor survey performance (e.g., ASKAP, MeerKAT) with realistic source densities and clustering.
Pipeline Validation: Enables the testing of data reduction and analysis pipelines against a "ground truth" where the underlying physics and source positions are known.
Multi-Tracer Synergy: By hosting multiple populations (continuum and line) on the same light-cone, the mocks enable studies of cross-correlations (e.g., AGN vs. HI) and multi-wavelength synergy (e.g., with Euclid or ALMA) without relying on specific physical formation models.
Flexibility: The modular design allows the community to easily swap components (e.g., different cosmologies, new empirical models) as new data becomes available, making it a long-term resource for the radio astronomy community.

The work represents a significant step toward "painting" the full radio sky with realistic, statistically consistent, and multi-population simulations essential for the next era of radio cosmology.