Pith. sign in

REVIEW 3 major objections 6 minor 1 cited by

Joint moments of weak lensing, tSZ, and X-ray maps tighten primordial non-Gaussianity constraints by a factor of two over lensing alone.

Reviewed by Pith at T0; open to challenge. T0 means a machine referee read the full paper against a public rubric. the ladder, T0–T4 →

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

T0 review · grok-4.5

2026-07-13 06:50 UTC pith:LWXCQAHN

load-bearing objection Solid multi-probe Fisher forecast showing a factor-of-two PNG gain from WL+tSZ+X-ray moments under a consistent baryon SAM; the number is real inside the model family, residual misspecification is the main caveat. the 3 major comments →

arxiv 2607.06692 v2 pith:LWXCQAHN submitted 2026-07-07 astro-ph.CO

Primordial Physics in the Nonlinear Universe: Towards particle constraints using the Weak lensing, Thermal SZ, and X-ray fields

Dhayaa Anbajagane This is my paper
classification astro-ph.CO
keywords primordial non-Gaussianityweak lensingthermal Sunyaev-ZeldovichX-raybaryon semi-analytic modelshigher-order momentsmulti-wavelength cosmologyFisher forecasts
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved

The pith

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

Primordial non-Gaussianities leave imprints on the initial density field that later shape the statistics of peaks and massive halos. This paper shows that those imprints remain measurable in the second and third moments of three late-time fields: weak-lensing convergence, the thermal Sunyaev–Zeldovich effect, and X-ray photon counts. By forward-modeling all three observables from the same N-body lightcones with a shared semi-analytic baryon prescription and correlated foregrounds, the author finds that a joint analysis recovers twice the constraining power on the amplitude of several PNG templates relative to lensing alone, even after marginalizing over cosmology, intrinsic alignments, and an extended set of baryonic nuisance parameters. The multi-wavelength combination also self-calibrates many of those nuisance parameters by breaking degeneracies that pure lensing cannot resolve. The result matters because it converts existing and near-term multi-wavelength surveys into a practical route for probing particle physics of the early Universe on scales that the CMB and galaxy clustering have not yet reached.

Core claim

Using second and third moments of the joint weak-lensing, thermal SZ, and X-ray fields yields a factor-of-two improvement in constraints on the amplitude of several primordial non-Gaussianity templates relative to a pure lensing analysis, while still marginalizing over cosmology, intrinsic alignments, and an extended baryon semi-analytic model.

What carries the argument

Simulation-based Fisher forecasts that paint consistent semi-analytic baryon profiles onto shared N-body lightcones, generate correlated multi-wavelength maps (weak lensing, Compton-y, X-ray counts), and extract second- and third-order aperture moments.

Load-bearing premise

The semi-analytic baryon model and the adopted prescriptions for CIB, radio, AGN, and metallicity fully capture the correlated astrophysical systematics at the precision needed for unbiased forecasts.

What would settle it

Apply the same multi-wavelength moment pipeline to real overlapping maps (for example DES/DECADE lensing + SPT/SO Compton-y + eROSITA counts) and test whether the recovered PNG posteriors tighten by the predicted factor of two relative to the lensing-only baseline after the same nuisance marginalization.

Watch this falsifier — get emailed when new claim-graph text bears on it.

If this is right

  • PNG constraints competitive with or better than Planck become accessible from existing multi-wavelength survey overlaps once moments of tSZ and X-ray are co-analyzed with lensing.
  • Baryonic and foreground nuisance parameters that degrade pure-lensing forecasts can be self-calibrated by the joint data vector rather than left as free systematics.
  • The same map-making pipeline can be extended to higher resolution, larger sky area, and additional summary statistics to push further gains.
  • Models whose PNG signatures peak on small scales (for example certain non-Bunch–Davies templates) gain relatively more from the multi-probe combination.

Where Pith is reading between the lines

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

  • Because the improvement is driven by degeneracy breaking around massive-halo peaks, other peak-sensitive probes such as cluster counts or kSZ may yield comparable multiplicative gains when added under the same consistent baryon model.
  • The public multi-wavelength map maker lowers the barrier for other groups to test whether higher-order statistics or machine-learned summaries can extract still more PNG information from the same simulated skies.
  • If residual model error in the baryon or foreground prescriptions re-introduces degeneracies, the claimed factor-of-two gain would shrink first for the most equilateral-like templates, which rely most heavily on intermediate and small scales.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit.

Referee Report

3 major / 6 minor

Summary. The paper forecasts constraints on several primordial non-Gaussianity (PNG) templates using second and third moments of weak-lensing convergence (LSST Y10), thermal SZ (enhanced SO + Planck), and X-ray count-rate (eROSITA DR3-like) maps. It builds multi-wavelength synthetic skies from the Ulagam N-body lightcones with Aarambam PNG initial conditions, applies a consistent baryon semi-analytic model (BaryonForge) across density, pressure, and X-ray emissivity, and correlates CIB, radio, and AGN foregrounds with the simulated density field. Using Fisher matrices with Hartlap correction and sequential marginalization over cosmology, IA, and an extended baryon SAM, the central claim is that the joint (WL+tSZ+X-ray) analysis recovers a factor-of-two improvement on f_NL relative to WL alone while self-calibrating nuisance parameters. Public code (Vaanam, BaryonForge, Aarambam) and simulation products are released.

Significance. If the forecast holds under realistic residual systematics, the work would elevate PNG constraints toward the multi-probe status already enjoyed by LCDM analyses, and would demonstrate that tSZ and X-ray moments can both add PNG information and break baryon/IA degeneracies that otherwise degrade WL-only forecasts by factors of 2–3. The public multi-wavelength map-maker and the consistent baryon SAM across three observables are concrete community assets. The numerical-convergence tests (Appendix C), Hartlap-corrected covariances, and explicit sequential marginalization over a large nuisance set are strengths relative to many Fisher forecasts in the literature. The result is therefore of clear interest for Stage-IV multi-wavelength cosmology even if the precise factor-of-two number is model-conditional.

major comments (3)
  1. Abstract and §4.2 / Fig. 3: The headline claim of a factor-of-two improvement on f_NL (joint vs WL-only, after full nuisance marginalization) is computed entirely inside the BaryonForge + transfer-function CIB (Eqs. A.9–A.10) + radio HOD/LF (Eqs. A.11–A.13) + AGN LF + two-parameter metallicity (Eq. A.19) model family. Section 4.5 and Appendices A.2–A.3 list missing physics (accretion shocks, cool-core/NCC bimodality, extended IA beyond NLA, CO, map filtering) that can re-correlate f_NL with the same nuisance directions the multi-probe combination is said to self-calibrate. The abstract and conclusions should state explicitly that the factor-of-two gain is conditional on completeness of this model family at the forecast precision, and should quantify (or bound) how residual freedom outside the varied parameters would degrade the gain, rather than only arguing O(10%) effects from amplitude
  2. §3.3 and Appendix A.3: The X-ray forward model is described as “more simplistic” than the WL and tSZ pipelines and as requiring further investigation (masking, background subtraction, cool-core bimodality) before data application. Yet the “All” and “All++” configurations in Fig. 3 and the degeneracy-breaking narrative in Fig. 4 assign non-negligible weight to X-ray auto- and cross-moments. Either (i) demonstrate that the reported joint improvement remains stable when the X-ray cluster component is down-weighted or removed, or (ii) clearly separate the tSZ-driven gain from the X-ray-driven gain so that the central claim does not rest on the least mature map model.
  3. §2.1 and §4.5: Maps are limited to NSIDE=1024 and halo catalogs to M_200c ≳ 10^14 M_⊙/h. The text correctly notes that this underestimates tSZ/X-ray signal and small-scale WL power, but the Fisher information (Eq. 4.1) and the scale-cut tests (θ > 10′, 25′) are still presented as representative of LSST Y10 / SO / eROSITA. Because the PNG signal is argued to live in peaks and one-halo regimes, a quantitative estimate of how much the factor-of-two ratio would change under higher resolution (or a statement that the ratio is conservative) is needed for the claim to be load-bearing rather than illustrative.
minor comments (6)
  1. Table 1: Several baryon SAM parameters (e.g. γ, δ, μ_θej) appear only in the extended set; a short cross-reference to A24 for their physical meaning would help readers who do not have that paper open.
  2. Fig. 2: Planck comparison lines are useful; please state in the caption whether they are 68% marginalized constraints on the same templates (and same mass parameters ν, μ) as used here.
  3. §2.3: The sparse fourth-moment subset used in “All++” is defined in prose; a compact equation or bullet list would make the datavector reproducible without re-reading the paragraph.
  4. Appendix D: The correlated Poisson sampling is a nice technical contribution; a one-sentence pointer in the main text (§3.3 or §4) would help readers who skip the appendices understand why X-ray noise does not artificially inflate Fisher information.
  5. Typos / notation: “propogating” → “propagating” (footnote 7); “vaccuum” → “vacuum” (NBD2 subsection); “derivate” → “derivative” (§4.1); “metallicty” → “metallicity” (§4.2).
  6. §4.4 / Fig. 5: Clarify whether the Planck prior is applied as a diagonal Gaussian on (σ_8, Ω_m) only, or includes the full Planck covariance; this affects the quoted 30% degradation.

Circularity Check

1 steps flagged

No significant circularity: the factor-of-two Fisher gain is measured from multi-probe moments on independent N-body lightcones, not forced by definition or by a load-bearing self-citation of the target result.

specific steps
  1. self citation load bearing [Sec. 2.1, 3.1; App. A; Data Availability]
    "Our simulations come from the Ulagam suite (Anbajagane et al. 2024c; Anbajagane & Lee 2026a,b) ... All semi-analytic modeling in this work uses the BaryonForge codebase (A24). ... Our multi-wavelength map maker can be found at github.com/DhayaaAnbajagane/Vaanam."

    The entire forward model (PNG ICs, baryon SAM, CIB transfer functions, radio/AGN sampling) is taken from the author's own prior codes and papers. This is ordinary infrastructure reuse, not a circular derivation of the factor-of-two claim: the Fisher ratio itself is recomputed from the new multi-probe moments and is not assumed or fitted in those earlier works. Flagged only as a minor, non-load-bearing self-citation pattern.

full rationale

The paper is a simulation-based Fisher forecast. PNG amplitudes enter only as shifts in the initial conditions (via Aarambam); the second- and third-order moments of the forward-modeled WL, tSZ and X-ray maps are then differentiated numerically and contracted with the covariance (Eq. 4.1). The claimed factor-of-two improvement is simply the ratio of the resulting marginalized errors on f_NL between the multi-probe and WL-only datavectors. Nothing in that chain equates the improvement to a fitted normalization or to a quantity that was defined in terms of the output. Self-citations (Papers I/II, A23, A24, BaryonForge, Vaanam) supply the IC generator, the light-cone suite and the semi-analytic baryon/foreground pipeline; they do not supply the numerical value of the information gain. Residual model incompleteness (shocks, cool-core bimodality, extended IA, etc.) is an assumption/correctness risk, not a circular reduction. Score 1 only for the ordinary presence of infrastructure self-citations that are not load-bearing for the central claim.

Axiom & Free-Parameter Ledger

7 free parameters · 5 axioms · 0 invented entities

The central forecast rests on standard cosmological assumptions (GR, LCDM background, known PNG bispectrum templates), a large set of free nuisance parameters that are marginalized, and semi-analytic prescriptions whose functional forms are taken from prior literature rather than derived here. No new physical entities are postulated; the novelty is in the joint application and the public map-making machinery.

free parameters (7)
  • f_NL (per PNG template)
    Amplitude of each primordial bispectrum shape; the target parameter whose Fisher error is reported.
  • Omega_m, sigma_8
    LCDM growth parameters marginalized in the forecasts (Table 1).
  • A_IA, eta_IA
    Amplitude and redshift slope of the NLA intrinsic-alignment model.
  • Mc, mu_beta, theta_ej, eta, theta_co, gamma, delta, mu_theta_ej, nu_Mc, nu_theta_ej
    Baryon SAM parameters controlling gas and stellar profiles (fiducial + extended sets).
  • alpha_nt
    Non-thermal pressure fraction at R_200c.
  • Z_core, Z_out
    Core and outer metallicity normalizations for X-ray emissivity.
  • f_CIB, f_AGN
    Overall rescaling amplitudes for CIB and AGN foregrounds.
axioms (5)
  • domain assumption General Relativity and a standard LCDM (or wCDM) background expansion govern structure formation.
    Stated in Section 2.1; all N-body runs assume GR.
  • domain assumption The six chosen PNG bispectrum templates (Local, Equilateral, QSF, SI, SII, NBD2) adequately sample the relevant inflationary phenomenology.
    Section 2.2; shapes taken from Chen & Wang, Sohn et al., Planck, etc.
  • domain assumption The Non-linear Linear Alignment (NLA) model captures the leading intrinsic-alignment contribution to weak lensing.
    Equation A.2 and Section 4.5; higher-order TATT terms are argued to be sub-dominant.
  • ad hoc to paper Halo-based semi-analytic baryon profiles plus transfer-function CIB and Poisson radio/AGN sampling produce sufficiently accurate multi-wavelength maps for Fisher forecasts.
    Core of Sections 3 and A; validated against Agora/WebSky at the ~50 % level but not derived from first principles.
  • domain assumption Second and third moments on 10 tophat scales (3.2'–200') capture the bulk of the accessible PNG information.
    Section 2.3; motivated by prior WL literature and computational limits on higher moments.

pith-pipeline@v1.1.0-grok45 · 49303 in / 3188 out tokens · 31477 ms · 2026-07-13T06:50:29.274082+00:00 · methodology

0 comments
read the original abstract

Primordial non-Gaussianities (PNGs) are a broad class of features in the initial density field that are connected to the particle physics of the early Universe. Measuring the amplitude of these features directly constrains fundamental physics from these earliest epochs and lends insight into energy scales that cannot be probed with terrestrial experiments. Using a new class of simulation methods, we propagate these signatures to their impact on the formation of non-linear structure and quantify the constraining power in non-Gaussian summary statistics of weak lensing, thermal Sunyaev Zeldovich (tSZ), and X-ray surveys. We use semi-analytic baryon models that consistently include astrophysical effects across all these observables, and use foreground modeling approaches that explicitly fold in correlations between the various components. We find that the tSZ and X-ray fields have significant information about PNGs, and additionally can help self-calibrate a broad set of nuisance parameters/models by breaking parameter degeneracies. Using the second and third moments of the lensing, tSZ, and X-ray fields, we find a factor of 2 improvement in PNG constraints relative to using lensing alone. Larger improvements are expected when including more scales and other complementary summary statistics. Our multi-wavelength map maker can be found at https://github.com/DhayaaAnbajagane/Vaanam. The simulations and software pipelines used in this analysis are publicly available.

Review history (2 revisions) →

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. Primordial Physics in the Nonlinear Universe: Revealing the oscillating halo bias from cosmological collider models

    astro-ph.CO 2026-07 accept novelty 7.0

    A binning-based IC method yields the first N-body measurements of oscillating halo bias from cosmological collider bispectra, with mass- and assembly-dependent phases fit by peak-background-split theory.

Reference graph

Works this paper leans on

15 extracted references · 3 linked inside Pith · cited by 1 Pith paper

  1. [1]

    Cosmology Astropart

    Abitbol M., et al., 2025, J. Cosmology Astropart. Phys., 2025, 034 Achúcarro A., et al., 2022, arXiv e-prints, p. arXiv:2203.08128 Ade P., et al., 2019a, J. Cosmology Astropart. Phys., 2019, 056 Ade P., et al., 2019b, J. Cosmology Astropart. Phys., 2019, 056 Ade P. A. R., et al., 2021, Phys. Rev. Lett., 127, 151301 Aird J., Coil A. L., Georgakakis A., Nan...

    Pith/arXiv arXiv 2025

  2. [2]

    arXiv:2404.06098 Bigwood L., et al., 2025, arXiv e-prints, p

    (arXiv:1407.2973) Bigwood L., et al., 2024, arXiv e-prints, p. arXiv:2404.06098 Bigwood L., et al., 2025, arXiv e-prints, p. arXiv:2512.04209 Blazek J. A., MacCrann N., Troxel M. A., Fang X., 2019, Phys. Rev. D, 100, 103506 Bleem L. E., et al., 2022, ApJS, 258, 36 Brandt W. N., Yang G., 2022, in Bambi C., Sangangelo A., eds, , Handbook of X-ray and Gamma-...

    Pith/arXiv arXiv doi:10.1007/978-981-16-4544-0_130-1 2024

  3. [3]

    - Virtual Workshop. p. 56 (arXiv:2107.01663), doi:10.5281/zenodo.5013757 – 32 – A Additional modeling details We now provide a more detailed description of the modeling pipeline from Section 3.3. We discuss the WL, tSZ, and X-ray observables separately in Sections A.1, A.2, and A.3, respec- tively. A.1 WL modelling As noted before, WL is the most mature o...

    Pith/arXiv arXiv doi:10.5281/zenodo.5013757 2019

  4. [4]

    linear”, or first-order, IA correction but uses the “non-linear

    = 1, andη IA is the redshift scaling. BothA IA andηIA are free parameters of the IA model. The IA signal can trivially be added to the cosmology signal asκ→κ+κIA. This IA parameterization is called the Non-linear Linear Alignment (NLA) model20. Other, more sophisticated parameterizations of the IA effect also exist (Blazek et al. 2019; Chen & Kokron 2024)...

    2019

  5. [5]

    However, in actuality, galaxy shapes trace thereduced shear,γ→γ/(1−κ)

    ( κℓm E +iκℓm B ) ,(A.3) whereX{E,B}are the E-mode and B-mode (or Q and U polarizations, in healpix notation) of the field. However, in actuality, galaxy shapes trace thereduced shear,γ→γ/(1−κ). We include this transformation in our modeling. We also explicitly include the magnification effect viaγ→γ(1+qκ), which accounts for the (observed) source galaxy ...

    2026

  6. [6]

    and therefore not highly clustered. Equation (A.5) focuses on the cosmological signal, whereas the noise field must 22We distinguish this from the apparent clustering of source galaxies due to the angular variations in survey observing conditions (e.g.,depth, seeing, etc.) across the sky. In simulation-based analyses of DES data, such effects are included...

    2022

  7. [7]

    Under the assumption of hydrostatic equilibrium, thermal gas pressure can be predicted using the gas density distribution and the total matter density distribution

    and are not necessary.23 A.2 Thermal SZ modelling The thermal SZ effect is the cosmic distribution of thermal electron pressure integrated along the line-of-sight. Under the assumption of hydrostatic equilibrium, thermal gas pressure can be predicted using the gas density distribution and the total matter density distribution. Both these components are al...

    2014

  8. [8]

    24 This conversion assumes the temperature of the electron and protons (or gas) are in equilib- rium

    We convert the gas pressure to electron pressure using the cosmic hydrogen and helium abundances,P e,th(r) = 4−2Y 8−5YPgas,th(r)whereY= 0.24is the cosmic helium abundance. 24 This conversion assumes the temperature of the electron and protons (or gas) are in equilib- rium. Recent work in tSZ data has shown hints of non-equilibrium at the cluster virial ra...

    2021

  9. [9]

    Since much remains uncertain about the nature of shocks in these outskirts, our current modeling does not include it. Future works will need to check that such features do not bias inference, as shocks will lead to correlated features across tSZ and WL.25 Using our model forP e,th, we can obtain the thermal SZ through a projection integral along the line-...

    2015

  10. [10]

    We show counts from theWebSkyreference catalog provided by Li22

    Figure 8: The radio source counts as a function of frequency (panels), split by the two source classes. We show counts from theWebSkyreference catalog provided by Li22. The luminosity function agrees well, except for some deviations at the high luminosity end, which are expected as we do not use the modified abundance-matching procedure nor the modified S...

    2023

  11. [11]

    All sampled AGN above this threshold are masked

    — our threshold is set at≈0.3 photons per second. All sampled AGN above this threshold are masked. – 43 – 10□2 10□1 100 R [Mpc] 0.4 0.6 0.8 1.0 1.2 Z/Z⊙ 10□2 10□1 100 R [Mpc] 0 2 4 6R2Np [Mpc2/s/sr] ×10□21 M200c = 10 14M⊙ z = 0.3 0.0 0.2 0.4 0.6 0.8 1.0 Zcore[Z⊙] Figure 9: The metallicity profiles (left) and X-ray counts (right) for a single halo ofM200c ...

    2026

  12. [12]

    Forexample, Mantzetal.(2017)findredshiftevolutionthatisconsistent with zero within their uncertainties, suggesting at a weaker evolution of the metallicity

    — we assume this dependence is minor relative to varying the redshift-independent normalizationitself. Forexample, Mantzetal.(2017)findredshiftevolutionthatisconsistent with zero within their uncertainties, suggesting at a weaker evolution of the metallicity. Figure 9 shows the impact of assuming a scale-independent metallicity on the X-ray photon counts....

    2017

  13. [13]

    We generate a full-sky, noiseless mock dataset from one simulation and compute the harmonic auto/cross-power spectra

    A.4 Summary To summarize, we refer to Figure 10, which shows an example of the cross-correlations between different components discussed above. We generate a full-sky, noiseless mock dataset from one simulation and compute the harmonic auto/cross-power spectra. The correlations betweenκ,y, and X-rays are broadly consistent with the best-fit model in La25 ...

    2023

  14. [14]

    Variable 1

    Figure 12: The numerical convergence of our results, estimated by the change in (marginal- ized) constraining power as we change the number of simulations used to estimate derivatives (top) and covariance (bottom) as defined in Equation (4.1). All parameters change by less than 10% if we use half the available simulations. When the noise model is complete...

    1993

  15. [15]

    Our implementation is provided as theutils.poisson_sample_fastmethod in theV aanampackage

    For a fixed set of uniform variatesu, we have confirmed our approach produces numerically identical samples as theScipymethod. Our implementation is provided as theutils.poisson_sample_fastmethod in theV aanampackage. D.2 Key-based uniform sampling The method of the above section allows us to assign random variates such that the Poisson random variate in ...

    1998