arxiv: 2512.09891 · v2 · submitted 2025-12-10 · 🌌 astro-ph.CO · astro-ph.GA

Recognition: 1 theorem link

· Lean Theorem

Self-consistent secondary cosmic microwave background anisotropies and extragalactic foregrounds in the FLAMINGO simulations

Tianyi Yang , Ian G. McCarthy , Fiona McCarthy , Boris Bolliet , Jens Chluba , William Coulton , John C. Helly , Matthieu Schaller

show 1 more author

Joop Schaye

Authors on Pith no claims yet

Pith reviewed 2026-05-16 23:14 UTC · model grok-4.3

classification 🌌 astro-ph.CO astro-ph.GA

keywords secondary CMB anisotropieshydrodynamical simulationsSunyaev-Zeldovich effectscosmic infrared backgroundCMB lensingmock mapsextragalactic foregroundsFLAMINGO

0 comments

The pith

Hydrodynamical simulations produce self-consistent mock maps of CMB secondary anisotropies and foregrounds.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper presents a new set of mock CMB maps constructed self-consistently from the FLAMINGO suite of cosmological hydrodynamical simulations. These maps incorporate CMB lensing, thermal and kinetic Sunyaev-Zeldovich effects, cosmic infrared background, radio point sources and anisotropic screening. The simulations reproduce a wide range of observational constraints. This approach enables direct comparison to earlier dark-matter-only models and shows how the signals vary with cosmology and feedback choices.

Core claim

Starting from lightcone-based HEALPix maps and catalogues, the authors create mock CMB maps from the FLAMINGO hydrodynamical simulations that include CMB lensing, thermal and kinetic Sunyaev-Zeldovich effects, cosmic infrared background, radio point sources and anisotropic screening. The simulations reproduce a wide range of observational constraints and match observations at least as well as previous independent models while preserving self-consistency across all components.

What carries the argument

Lightcone-based HEALPix maps extracted from the FLAMINGO hydrodynamical simulations, which supply consistent gas and matter distributions for multiple secondary CMB effects at once.

If this is right

The mock maps provide a resource for exploring correlations between different secondary anisotropies and other large-scale structure tracers.
Cross-correlations between signals differ significantly from those produced by previous independent mocks.
The signals depend on cosmology and feedback modelling, allowing tests of these aspects.
The maps support forecasts for upcoming surveys.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

These mocks could be combined with real survey data to tighten constraints on galaxy formation feedback models.
Discrepancies at small scales might highlight needed refinements in simulation resolution or included physics.
Self-consistency may change how baryonic effects are marginalised in cosmological analyses from future CMB data.
The model variations offer a direct way to propagate baryonic uncertainties into forecasts.

Load-bearing premise

The FLAMINGO hydrodynamical simulations accurately capture the baryonic physics, feedback processes and gas distributions required to model the secondary anisotropies and foregrounds at the relevant scales and redshifts.

What would settle it

Comparison of the simulated power spectra or cross-correlations against measurements from next-generation CMB surveys such as Simons Observatory; statistically significant mismatches at relevant angular scales would indicate the models do not capture the true signals.

Figures

Figures reproduced from arXiv: 2512.09891 by Boris Bolliet, Fiona McCarthy, Ian G. McCarthy, Jens Chluba, John C. Helly, Joop Schaye, Matthieu Schaller, Tianyi Yang, William Coulton.

**Figure 1.** Figure 1: Constraints on the 𝛽𝑑, 𝑇0, and 𝛼 parameters of the SED of the CIB model, obtained by fitting to the 353/545/857 GHz auto-power spectrum measurements from Lenz et al. (2019). are used to generate the mock maps and compute the power spectrum. We find 𝛽𝑑 = 1.65 ± 0.02 and 𝑇0 = 35.14 ± 0.18 K for the fiducial FLAMINGO model (we discuss how these parameters may vary as a function of feedback and cosmology in S… view at source ↗

**Figure 2.** Figure 2: The 150 MHz radio luminosity function (RLF) reconstructed from the black hole particle lightcone in the fiducial (1 Gpc) 3 run. This is an example showing the case for a black hole selection of 𝜆Edd < 10−2 over the redshift range 0.5 < 𝑧 < 1.0. The observed RLF from the LOFAR survey within the same redshift interval (black points), as well as its best-fitting parametric model (red thick dashed line), are o… view at source ↗

**Figure 3.** Figure 3: Comparison of the differential source number counts between observations and simulations. The blue, orange, and green solid curves are the best-fitting source count models for the three black hole selections considered in this study (corresponding to different Eddington rate cuts), with the best-fitting 𝛼radio values obtained by fitting to the measured SPT data at three frequencies (Everett et al. 2020) (b… view at source ↗

**Figure 4.** Figure 4: shows an example of the impact of this lensing effect on the resulting power spectrum analysis, where we show the tSZ, kSZ and 217 GHz CIB auto-power spectrum as an illustration. In general, the lensing-induced modifications are small, where a suppression of power by around 1-2% is expected at small scales. The lensing effect on the CIB map at 217 GHz is slightly stronger than that on the SZ effect, due to… view at source ↗

**Figure 5.** Figure 5: Full-sky mock intensity maps of the thermal Sunyaev–Zel’dovich (tSZ) effect and its relativistic correction. These maps are generated from the lightcone outputs of the fiducial (2.8 Gpc) 3 run integrated up to 𝑧 = 4.5. Left: tSZ intensity map at 857 GHz, computed using Equation 3 (i.e., non-relativistic tSZ). Middle: Difference map between the relativistically corrected and non-relativistic tSZ intensity m… view at source ↗

**Figure 6.** Figure 6: Comparison of the tSZ angular power spectrum from different simulations. Thick dash-dot lines in different colors show results from the FLAMINGO (2.8 Gpc) 3 fiducial run and a subset of its (1 Gpc) 3 model variants, integrated up to 𝑧 = 3.0. The shaded region shows the spread obtained by averaging over 8 independent lightcones. The orange line shows the spectrum from the WebSky simulation. Solid lines are … view at source ↗

**Figure 7.** Figure 7: Comparison of the kSZ auto-power spectra from different simulations. The thick dark line shows the result from the FLAMINGO fiducial (2.8 Gpc) 3 run, integrated up to 𝑧 = 4.5. The thick red line shows the AGORA result up to 𝑧 = 3.0, and the thick orange line shows the WebSky result up to 𝑧 = 4.5. For reference, the primary CMB power spectrum (thin blue line) and the CMB instrumental noise curves for some … view at source ↗

**Figure 8.** Figure 8: The auto- (left) and cross- (right) power spectra measured from the CIB maps generated from the FLAMINGO fiducial (2.8 Gpc) 3 run at the Planck HFI frequencies. All curves are obtained by averaging over 8 independent lightcones. Results from the AGORA (red dashed) and the WebSky (orange dashed) simulations are displayed for comparison. Observational data points from Planck Collaboration et al. (2014) and L… view at source ↗

**Figure 9.** Figure 9: CIB monopole computed using the best-fitting CIB SED from the FLAMINGO fiducial (2.8 Gpc) 3 run, integrated up to different maximum redshifts. For comparison, the measurements provided by Chiang et al. (2025) are overplotted as a black dashed line. For reference, the vertical dashed lines are the Planck HFI frequencies — 353 GHz, 545 GHz, and 857 GHz — which are the frequencies used in the SED fitting to t… view at source ↗

**Figure 10.** Figure 10: CIB-LSS tracers cross-power spectra at different Planck frequencies. Left: Comparison of the CIB-tSZ cross-power spectrum between simulations (dark solid: fiducial (2.8 Gpc) 3 FLAMINGO run, with shaded regions estimated by averaging the results from eight different lightcones; red dashed: AGORA; orange dashed: WebSky) and halo model predictions from Planck Collaboration et al. (2016c) (black dashed). Righ… view at source ↗

**Figure 11.** Figure 11: Auto-power spectrum of radio point sources for the three 𝜆Edd cuts considered in our mock catalogue, where the radio luminosities are assigned by abundance matching to the observed radio luminosity function of low-excitation radio galaxies, measured at 150 MHz from the LOFAR survey (see Section 3.6 for details). Left: power spectrum computed from the full sample without masking bright sources. For compari… view at source ↗

**Figure 12.** Figure 12: Cross-power spectrum between radio flux density (after bright source removal) and other LSS tracers, where the radio luminosities are assigned by abundance matching to the observed radio luminosity function of low-excitation radio galaxies, measured at 150 MHz from the LOFAR survey (see Section 3.6 for details). Left: cross-correlation with the Compton 𝑦 field for three different 𝜆Edd cuts. Right: cross-c… view at source ↗

**Figure 13.** Figure 13: Feedback and cosmology dependencies of the CIB auto-power spectra for the case when SED parameters are fixed to the values from fitting the fiducial (2.8 Gpc) 3 run to the Lenz et al. (2019) data (see Section 3.5.1, left), and for the case when SEDs are refitted to the Lenz et al. (2019) data for different models (right). Green data points are the measurements from Lenz et al. (2019). The ratio curves bet… view at source ↗

**Figure 14.** Figure 14: Feedback and cosmology dependencies of the CIB-tSZ (left) and CIB-𝜅 (right) cross-power spectra for the case when CIB SED parameters are fixed to the values obtained from fitting the fiducial (2.8 Gpc) 3 run to the data from Lenz et al. (2019) (see Section 3.5.1). Ratios between different models and the fiducial curves for 857 GHz are shown on the bottom panel. other frequencies display the same general t… view at source ↗

**Figure 15.** Figure 15: Left: The best-fitting rest-frame dust temperature (𝑇0) and spectral index (𝛽d) for all FLAMINGO model variants considered in this study, with best-fitting values obtained by refitting the modelled CIB power spectra to Lenz et al. (2019) data. Right: Cosmic star formation rate density as a function of redshift for the different model variants. 10−2 10−1 100 CCIB − y ` [10 − 6 Jy /sr] 857GHz Halo model P16… view at source ↗

**Figure 16.** Figure 16: Feedback and cosmology dependencies of the CIB-LSS cross-power spectrum at 857 GHz for the case when SEDs are refitted to the Lenz et al. (2019) data for different models. Ratios between different models and the fiducial curves are shown in the bottom subpanels. The left panel shows the results for the CIB-tSZ cross-power spectrum, and the right panel shows the CIB-𝜅 curves. 6 CONCLUSION Secondary anisotr… view at source ↗

**Figure 17.** Figure 17: Feedback and cosmology dependencies of the CIB-LSS correlation coefficient at 545 GHz for the case when SEDs are refitted to the Lenz et al. (2019) data for different models, with shaded regions estimated by averaging the results from eight different lightcones. Ratios between different models and the fiducial curves are shown in the bottom subpanels. The left panel shows the results for the CIB-tSZ corre… view at source ↗

**Figure 18.** Figure 18: The kSZ auto-power spectrum for different FLAMINGO model variants. amplitudes differ due to varying treatments of subgrid physics and intergalactic medium (IGM) modelling (Figures 6 and 7). • Using a simplified three-parameter model, which includes a linear SFR−𝐿bol,IR conversion law, a modified blackbody spectral energy distribution (SED) template for infrared sources, and a power law redshift evolution … view at source ↗

read the original abstract

Secondary anisotropies in the cosmic microwave background (CMB) contain information that can be used to test both cosmological models and models of galaxy formation. Starting from lightcone-based HEALPix maps and catalogues, we present a new set of mock CMB maps constructed in a self-consistent manner from the FLAMINGO suite of cosmological hydrodynamical simulations, including CMB lensing, thermal and kinetic Sunyaev-Zeldovich effects, cosmic infrared background, radio point source and anisotropic screening maps. We show that these simulations reproduce a wide range of observational constraints. We also compare our simulations with previous predictions based on dark matter-only simulations which generally model the secondary anisotropies independently from one another, concluding that our hydrodynamical simulation mocks perform at least as well as previous mocks in matching the observations whilst retaining self-consistency in the predictions of the different components. Using the model variations in FLAMINGO, we further explore how the signals depend on cosmology and feedback modelling, and we predict cross-correlations between some of the signals that differ significantly from those in previous mocks. The mock CMB maps should provide a valuable resource for exploring correlations between different secondary anisotropies and other large-scale structure tracers, and can be applied to forecasts for upcoming surveys.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

FLAMINGO hydro sims deliver self-consistent mocks for multiple secondary CMB anisotropies that match observations at least as well as prior work while producing different cross-correlations.

read the letter

The main point is that this paper builds mock maps for CMB lensing, tSZ, kSZ, CIB, radio sources, and anisotropic screening all from the same FLAMINGO hydrodynamical lightcones instead of combining separate dark-matter runs. That self-consistency is the practical advance. The authors show the resulting maps reproduce a range of observational constraints comparably to earlier mocks and use the FLAMINGO variations to test how feedback and cosmology change the signals. They also note that some cross-correlations come out differently than in the independent-modeling case, which matters for anyone planning joint analyses with upcoming surveys. The maps are presented as a public resource for cross-correlation work and forecasts. The comparisons to data and to previous mocks are the parts that give the work its value. The soft spots are limited. The abstract summarizes agreement with observations without quoting the exact statistics or tension levels, so the strength of the match depends on the quantitative figures in the full text. If those are solid and include proper error budgets, the claim holds; if they stay mostly qualitative, it weakens a bit. The underlying assumption that FLAMINGO captures the relevant gas physics at the right scales is standard for this field and not a new flaw here. This paper is aimed at the CMB cosmology and large-scale structure community, particularly people who need consistent mocks for cross-correlation studies or survey forecasts. It deserves a serious referee because it supplies a usable, internally consistent dataset with clear advantages over piecemeal approaches, even though it is mainly a methods and data-release contribution rather than a new physical result. I would send it to peer review.

Referee Report

2 major / 3 minor

Summary. The paper presents a new set of mock CMB maps constructed self-consistently from the FLAMINGO suite of cosmological hydrodynamical simulations. Starting from lightcone-based HEALPix maps and catalogues, the mocks include CMB lensing, thermal and kinetic Sunyaev-Zeldovich effects, cosmic infrared background, radio point sources, and anisotropic screening. The authors show that these reproduce a wide range of observational constraints, compare them to previous dark-matter-only mocks (finding comparable or better performance while retaining self-consistency), explore dependence on cosmology and feedback variations, and predict cross-correlations that differ from prior independent-component mocks. The maps are positioned as a public resource for correlation studies and survey forecasts.

Significance. If the central claims hold, the work supplies a timely, publicly available set of self-consistent mocks that improve on the common practice of modeling secondary anisotropies independently. The hydrodynamical origin allows consistent treatment of baryonic effects across components, which is particularly valuable for cross-correlation forecasts with upcoming surveys. The inclusion of model variations in cosmology and feedback further enables sensitivity tests that are difficult to perform with post-processed dark-matter-only mocks.

major comments (2)

[§4] §4 (light-cone construction): the description of how the various components are added to the same light-cone volume is insufficient to confirm absence of internal inconsistencies (e.g., whether the same gas particles contribute to both tSZ/kSZ and lensing without double-counting or resolution mismatches). A short explicit consistency check or flowchart would strengthen the central self-consistency claim.
[§5.3] §5.3 and Table 2: the statement that the hydro mocks 'perform at least as well' as prior DM-only mocks is supported only by qualitative visual agreement in power spectra; quantitative metrics (e.g., reduced chi-squared or fractional residuals integrated over multipoles) are not reported, making it difficult to assess whether the improvement in cross-correlations is statistically significant.

minor comments (3)

[Figure 3] Figure 3 caption: the redshift range and angular resolution of the displayed maps should be stated explicitly so readers can judge the scales being compared to observations.
[§6] §6: the public data-release section should include a brief description of file formats, HEALPix N_side values, and any required post-processing steps for users.
[References] References: several key papers on tSZ/kSZ cross-correlations with DESI or Euclid are cited only in passing; adding one or two sentences on how the FLAMINGO predictions differ from those forecasts would improve context.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for their positive assessment and constructive comments on our manuscript. We address each major comment point by point below, providing clarifications and indicating where revisions will be made.

read point-by-point responses

Referee: [§4] §4 (light-cone construction): the description of how the various components are added to the same light-cone volume is insufficient to confirm absence of internal inconsistencies (e.g., whether the same gas particles contribute to both tSZ/kSZ and lensing without double-counting or resolution mismatches). A short explicit consistency check or flowchart would strengthen the central self-consistency claim.

Authors: All maps are generated from the identical light-cone particle data and halo catalogues produced by the FLAMINGO simulations. Lensing convergence is computed from the total matter surface density (dark matter + gas + stars + black holes) of every particle. The tSZ and kSZ signals are computed exclusively from the electron pressure and line-of-sight momentum of the gas particles only; no other components enter those maps. Because each observable uses a distinct physical quantity of the same particles, there is no double-counting. All maps are projected at the native simulation resolution before HEALPix binning, so resolution is uniform. We will insert a concise paragraph plus a simple flowchart in the revised §4 that explicitly traces the particle-to-map pipeline for each component. revision: yes
Referee: [§5.3] §5.3 and Table 2: the statement that the hydro mocks 'perform at least as well' as prior DM-only mocks is supported only by qualitative visual agreement in power spectra; quantitative metrics (e.g., reduced chi-squared or fractional residuals integrated over multipoles) are not reported, making it difficult to assess whether the improvement in cross-correlations is statistically significant.

Authors: The central claim in §5.3 is that the hydrodynamical mocks reproduce the observed auto-power spectra at a level comparable to previous DM-only mocks while additionally providing self-consistent cross-correlations. The figures demonstrate that our spectra lie within the observational error bars across the plotted multipole range in a manner visually indistinguishable from the earlier works. We agree that a quantitative summary would strengthen the presentation. In the revision we will add a supplementary table listing the mean fractional residual (and its rms) between each mock and the observational data, integrated over the multipole bins shown in the figures. This will allow readers to judge the agreement more rigorously without altering the conclusion that the hydro mocks perform at least as well. revision: partial

Circularity Check

0 steps flagged

No significant circularity in derivation chain

full rationale

The paper constructs and releases self-consistent mock CMB maps (lensing, tSZ, kSZ, CIB, radio sources, screening) directly from the existing FLAMINGO hydrodynamical simulation light-cones and catalogues. All load-bearing steps are (1) post-processing of simulation outputs to generate maps and (2) direct comparison of those maps to external observational constraints. No equations or claims reduce a prediction to a fitted parameter by construction, no uniqueness theorem is invoked via self-citation, and no ansatz is smuggled in. The self-consistency is a methodological feature (components drawn from the same baryonic physics run) rather than a definitional loop. The work is therefore a data-release and validation exercise whose central claims remain externally falsifiable against independent observations.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

The central claim rests on the accuracy of the FLAMINGO hydrodynamical model for baryonic effects and standard cosmological initial conditions; no free parameters or invented entities are mentioned in the abstract.

axioms (2)

standard math Standard Lambda-CDM cosmology with parameters used to initialize the simulations
Basis for the lightcone construction and structure formation
domain assumption Hydrodynamical treatment of gas, cooling, star formation and AGN/stellar feedback accurately reproduces the relevant observables
Required for self-consistent modeling of SZ, CIB and radio signals

pith-pipeline@v0.9.0 · 5565 in / 1352 out tokens · 36437 ms · 2026-05-16T23:14:56.732636+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

4 extracted references · 4 canonical work pages

[1]

Abbott T. M. C., et al., 2022, Phys. Rev. D, 105, 023520 Ade P., et al., 2019, J. Cosmology Astropart. Phys., 2019, 056 Aihara H., et al., 2018, PASJ, 70, S8 AlonsoD.,SanchezJ.,SlosarA.,LSSTDarkEnergyScienceCollaboration 2019, MNRAS, 484, 4127 Amon A., et al., 2023, MNRAS, 518, 477 Bagla J. S., Ray S., 2003, New Astron., 8, 665 Bahé Y. M., et al., 2022, M...

work page doi:10.1051/epjconf/202429300008 2022
[2]

For the logarithmic-scale versions, please see Figures

MNRAS000, 1–??(2025) FLAMINGO secondary CMB anisotropies25 0.5 1.0 1.5 ℓCℓ [Jy/sr]2 ×106 353x353 L19 L2p8 m9 (three-params) L2p8 m9 (four-params) AGORA WebSky 2 3 4 5ℓCℓ [Jy/sr]2 ×106 353x545 0.75 1.00 1.25 1.50 ×107 545x545 1000 2000 ℓ 0.50 0.75 1.00 ℓCℓ [Jy/sr]2 ×107 353x857 1000 2000 ℓ 1 2 3 4 ×107 545x857 1000 2000 ℓ 2 4 6 8 ×107 857x857 FigureB1.𝐶 CI...

work page 2025
[3]

Green data with error bars are measurements from Lenz et al

Shaded regions are the cosmic variance estimated by averaging the results from eight different lightcones. Green data with error bars are measurements from Lenz et al. (2019). 0 10 20 353GHz Simulations L2p8 m9 AGORA WebSky 25 50 75ℓCCIB□y ℓ [10□6 Jy/sr] 545GHz 500 1000 1500 2000 2500 3000 ℓ 50 100 150 857GHz FigureB2.AsFigureB1,butfor𝐶 CIB-y ℓ .Forthelog...

work page 2019
[4]

Yang et al

MNRAS000, 1–??(2025) 26T. Yang et al. 103 ℓ 106 107 108 109 1010 Cℓ ℓ(ℓ + 1)/(2π) [ Jy/sr]2 217GHz 353GHz 545GHz 857GHz Simulations L2p8 m9 AGORA WebSky Simulations L2p8 m9 AGORA WebSky 103 ℓ 108 109 1010 Cℓ ℓ(ℓ + 1)/(2π) [ Jy/sr]2 353GHz×545GHz 353GHz×857GHz 545GHz×857GHz Simulations L2p8 m9 AGORA WebSky Simulations L2p8 m9 AGORA WebSky Figure C1.The CIB...

work page 2025