Mapping the Optimal Sensitivity of the 21 cm Forest to Dark Matter-Baryon Scattering

2604.00025-R1 📅 13 Apr 2026 🔍 Reviewed by Skepthical View Paper GitHub

Official Review

Official Review by Skepthical 13 Apr 2026
Overall: 4.8/10
Soundness
4
Novelty
6
Significance
5
Clarity
5
Evidence Quality
4
The paper presents a timely and coherent idea—using the shape of the 21 cm forest optical-depth distribution with KL divergence to isolate dark matter–baryon scattering—and maps an optimal redshift window, but several core modeling and forecasting choices are under-specified. The Mathematical and Numerical Audits flag critical uncertainties: an unphysical sharp halo-mass cutoff, missing and non-reproducible mapping from Mcut to σ0/mχ, undefined KL normalization (critical to the shape-based claim), and optimistic Fisher assumptions that omit instrumental/systematic effects; a Poisson-scaling statement and two percent-reduction calculations also show minor inconsistencies. These gaps materially affect the headline constraint claims despite a clear narrative and reasonable qualitative trends, yielding a borderline overall assessment.
  • Paper Summary: This paper explores how the high-redshift 21 cm forest (IGM + minihalo absorption at $z\approx7$–$15$) can constrain elastic DM–baryon scattering that suppresses small-scale structure. Using the HAYASHI semi-analytic framework (Sec. 2.1), the authors implement DM–baryon effects phenomenologically through (i) a halo-mass cutoff $M_{\rm cut}$ applied to the halo mass function (Sec. 2.2) and (ii) an IGM temperature rescaling $f_{\rm cool} = T_k/T_k^{\rm ad}$ to represent non-standard cooling (Sec. 2.2). They quantify observable imprints via the optical-depth distribution $dN/d\tau$ (and cumulative $N(>\tau)$), and use the KL divergence as a shape-based “fingerprint” statistic (Sec. 2.3). A Fisher forecast over $\{M_{\rm cut}, f_{\rm cool}, \bar{x}_{\rm HI}\}$ is constructed from binned absorber counts, with Poisson uncertainties determined by an assumed redshift-dependent number of background sightlines $N_{\rm los}(z)$ (Sec. 2.4). The main conclusions are that suppressing low-mass minihalos dominates the detectable impact (removing preferentially low-$\tau$ absorbers and changing the shape of $dN/d\tau$), that varying $f_{\rm cool}$ has little effect on absorber statistics, and that the optimal redshift window is $z\approx8$–$10$ once $N_{\rm los}(z)$ is included (Sec. 3.1–3.3). The work then maps sensitivity to $M_{\rm cut}$ into projected constraints on velocity-independent ($n=0$) and Coulomb-like ($n=-4$) cross sections, claiming potential improvements over CMB limits by $4$–$5$ orders of magnitude (Sec. 3.4, Conclusions). The overall narrative is timely and the emphasis on shape information in $dN/d\tau$ is compelling, but several central modeling and forecast assumptions (sharp cutoff, mapping to $\sigma_0/m_\chi$, thermal/spin-temperature treatment, and observational/noise realism) require substantially more documentation and robustness testing for the quantitative headline claims to be convincing and reproducible.
Strengths:
Clear end-to-end narrative from DM microphysics $\rightarrow$ small-scale structure suppression $\rightarrow$ 21 cm forest observables, with a logically organized presentation (Introduction, Sec. 2–4).
Using the *shape* of $dN/d\tau$ (and information-theoretic diagnostics like KL divergence) as a discriminator is a strong and potentially powerful idea (Sec. 2.3, Sec. 3.2).
Helpful separation of two conceptually distinct DM–baryon channels—structure suppression ($M_{\rm cut}$) vs thermal effects ($f_{\rm cool}$)—and a quantitative comparison of their impacts (Sec. 2.2, Sec. 3.1).
Forecast framework includes nuisance parameters ($\bar{x}_{\rm HI}$ and $f_{\rm cool}$) and explores redshift dependence and an “optimal” observing window once source statistics are modeled (Sec. 2.4, Sec. 3.3).
Figures generally communicate the main trends clearly (especially the differential/cumulative $\tau$-statistics and the redshift dependence of forecast sensitivity).
The paper makes an explicit attempt to connect phenomenological small-scale suppression ($M_{\rm cut}$) to particle-physics parameters ($\sigma(v) = \sigma_0 v^n$), which is valuable if fully specified (Sec. 2.2, Sec. 3.4).
Major Issues (6):
  • Structure suppression is modeled as an *infinitely sharp* halo-mass cutoff $M_{\rm cut}$ (Sec. 2.2), i.e. all halos with $M<M_{\rm cut}$ are removed. For DM–baryon scattering (drag/heat exchange), realistic suppression is typically smooth and redshift/scale dependent (via a transfer function), and can also alter collapse thresholds, concentrations/formation times, and gas content. A step-function truncation risks (i) exaggerating or distorting the low-$\tau$ shape change in $dN/d\tau$ (the central “fingerprint”), and (ii) producing over-optimistic or mis-centered Fisher constraints on $M_{\rm cut}$ (Sec. 3.1–3.3), which then propagate directly to $\sigma_0/m_\chi$ (Sec. 3.4, Conclusions).
    Recommendation: In Sec. 2.2, replace the step cutoff with a physically motivated smooth suppression, or at minimum bracket the step model with a smoothed cutoff (e.g., logistic/erf in $\log M$ with tunable width) calibrated to published transfer-function/HMF results for DM–baryon scattering. In Sec. 3.1–3.2, add a robustness test showing how $dN/d\tau$, $N(>\tau)$, and $D_{\rm KL}$ change when the cutoff is smoothed and/or when a standard “half-mode mass” style mapping is used. Quantify how the inferred $\sigma(M_{\rm cut})$ (Sec. 3.3) shifts under these alternatives and temper claims that depend sensitively on the sharpness of the cutoff.
  • The mapping between the phenomenological cutoff $M_{\rm cut}$ and the microscopic scattering parameter $\sigma_0/m_\chi$ (for $n=0$ and $n=-4$) is asserted but not transparently derived or reproducible (Sec. 2.2, Sec. 3.4). The manuscript does not specify: the perturbation calculation/transfer function used, how the suppression scale is defined (e.g., 50% suppression in $T(k)$, filtering mass, half-mode), how it is translated to a halo-mass scale at relevant redshifts, dependence on $m_\chi$ and cosmology, or what baryonic target (protons/electrons/neutral H) is assumed. Since the headline result is the $\sigma_0/m_\chi$ reach and the “4–5 orders of magnitude over CMB” comparison (Sec. 3.4, Conclusions), this opacity undermines credibility and prevents verification.
    Recommendation: Add a dedicated subsection (end of Sec. 2.2 or new Sec. 2.5 / Appendix) detailing the full pipeline $\sigma_0/m_\chi \rightarrow$ transfer-function suppression $\rightarrow$ halo-mass-function modification $\rightarrow$ an effective $M_{\rm cut}$ used in HAYASHI. Explicitly state: assumed DM mass $m_\chi$, scattering targets, cosmological parameters, the definition of the cutoff scale (e.g., $k_{1/2}$ where $T(k)=1/2$), and the conversion to a characteristic mass (e.g., half-mode mass). Provide a table/figure of $M_{\rm cut}$ versus $\sigma_0/m_\chi$ for representative redshifts (e.g., $z=8,10$) and both $n=0$ and $n=-4$. In Sec. 3.4, overlay or tabulate the corresponding $M_{\rm cut}$ values for the quoted $\sigma_0/m_\chi$ sensitivities and include an estimate of systematic uncertainty from the mapping.
  • The thermal/spin-temperature and minihalo gas modeling is too compressed for the paper’s strong conclusion that even extreme $f_{\rm cool}$ has negligible impact and that the $M_{\rm cut}$-induced *shape* change robustly breaks astrophysical degeneracies (Sec. 2.1–2.2, Sec. 3.1–3.2). It is unclear how HAYASHI computes $T_s$ (collisional vs. Wouthuysen–Field coupling, radiation backgrounds), whether the regime $T_s\approx T_k$ is always realized over $z=7$–$15$, whether absorption is in the linear $\tau\ll1$ regime, and whether $f_{\rm cool}$ is applied to IGM only or also to minihalo gas (where line profiles/central densities set $\tau$). The current parameter set $\{M_{\rm cut}, f_{\rm cool}, \bar{x}_{\rm HI}\}$ may not capture astrophysical effects that also change the *shape* of $dN/d\tau$ (e.g., photoheating feedback reducing gas in low-mass halos, mass-dependent gas fractions, X-ray/Ly$\alpha$ fluctuations).
    Recommendation: Expand Sec. 2.1 with an explicit (but concise) summary of the ingredients that determine $\tau$ in the IGM and minihalos: gas density/temperature profiles, how $T_s$ is computed (collisions/Ly$\alpha$ coupling), what backgrounds are assumed, and the halo mass/redshift range driving the statistics. In Sec. 2.2 and Sec. 3.1, clarify exactly where $f_{\rm cool}$ is applied (IGM only vs also halos), whether it is redshift dependent, and report representative $T_k$ and $T_s$ ranges to support the “saturation” argument. To support degeneracy-breaking claims (Sec. 3.2), add at least one additional nuisance parameter that can preferentially suppress low-$\tau$ absorbers (e.g., a feedback/photoheating suppression mass, a low-mass gas-fraction suppression parameter, or a temperature floor) and show how it compares to changing $M_{\rm cut}$ in $dN/d\tau$ and in the Fisher constraints.
  • Forecast/statistical methodology is under-specified and likely optimistic: the Fisher analysis (Sec. 2.4) appears to assume independent Poisson counts per $\tau$-bin with ${\rm Var}(N_k)=N_k$, but the paper does not fully define the data vector (counts vs normalized PDF), $\tau$-binning and $\tau$-range, redshift binning/path length, or how $N_{\rm los}(z)$ enters $N_k$. Instrumental effects and analysis systematics that are central for low-$\tau$ features—finite spectral resolution, thermal noise, continuum fitting, line blending/confusion, completeness/detection thresholds in $\tau$—are neglected, yet much of the claimed information is in the low-$\tau$ regime (Sec. 3.1–3.3). This makes $\sigma(M_{\rm cut})$ and the optimal $z\approx8$–$10$ window potentially fragile (Sec. 3.3), and the resulting $\sigma_0/m_\chi$ reach likely too optimistic (Sec. 3.4, Conclusions).
    Recommendation: In Sec. 2.3–2.4, explicitly define the observable: whether $N_k$ are total absorber counts across all sightlines in $(z,\tau)$ bins (preferred for Poisson likelihood), or a normalized shape-only statistic; provide the $\tau$-bin edges, $\tau$-range, and any redshift binning/path length per sightline. Write $N_k(z) = N_{\rm los}(z)\times\lambda_k(z,\theta)$ (or equivalent) so the $N_{\rm los}$ scaling is explicit. Then incorporate a minimal observational realism layer: e.g., (i) a $\tau$ detection threshold/completeness curve, (ii) an effective reduction in usable $N_{\rm los}$, and/or (iii) an added fractional variance term per bin to represent non-Poisson systematics. Recompute (or at least bracket) $\sigma(M_{\rm cut})$ and the preferred redshift window under optimistic vs pessimistic thresholds, and soften conclusions where they depend on Poisson-limited low-$\tau$ performance.
  • KL divergence is used as a central “intrinsic distinguishability” metric (Sec. 2.3, Fig. 3, Sec. 3.2) but the manuscript does not define the normalized distributions entering $D_{\rm KL}$. KL divergence requires proper probability distributions $p(\tau)$, $q(\tau)$ (or discrete probabilities $p_k$, $q_k$). If computed from $dN/d\tau$ or raw $N_k$ without normalization, the interpretation changes (shape-only vs amplitude+shape), and the connection to detectability with finite counts is unclear.
    Recommendation: In Sec. 2.3, add the explicit definition of $D_{\rm KL}$ and specify precisely how $p$ and $q$ are constructed from $dN/d\tau$ or binned $N_k$ (including normalization over the chosen $\tau$-range). If the intent is “shape-only,” state that explicitly and show how amplitude information is treated separately. In Sec. 3.2, add a brief interpretive bridge: for representative $N_{\rm los}$ and total counts, what $D_{\rm KL}$ values correspond to a meaningful likelihood-ratio/$\Delta\chi^2$ separation? (A simple approximation or illustrative conversion is sufficient.)
  • The assumed redshift evolution of background radio-loud sources $N_{\rm los}(z)=10[(1+z)/8]^{-2.5}$ (Sec. 2.4) is a key driver of the forecast and the claimed optimal $z\approx8$–$10$ window (Sec. 3.3), but is currently schematic and not clearly justified by a cited luminosity-function model, flux limit, or survey strategy. Likewise, the “4–5 orders of magnitude better than CMB” statement (Sec. 3.4, Conclusions/Abstract) is presented without a like-for-like comparison (model assumptions, $m_\chi$, $n$, ionization history) and without plotting the referenced CMB limits alongside the forecast.
    Recommendation: In Sec. 2.4, cite the provenance of the $N_{\rm los}(z)$ scaling (data or simulations) and add at least one optimistic and one pessimistic alternative (varying normalization and exponent, or tying $N_{\rm los}$ to a flux limit and a luminosity function). In Sec. 3.3, show how $\sigma(M_{\rm cut})$ and the optimal redshift shift under these alternatives. In Sec. 3.4, state explicitly which CMB limits are used (citations, $m_\chi$, targets, $n$, and assumptions) and overlay them on the same $\sigma_0/m_\chi$ plot; then restate the improvement factor conditional on those assumptions, and temper the language in the Abstract/Conclusions if it only holds in optimistic source/noise scenarios.
Minor Issues (8):
  • Several plots/figures lack enough normalization/unit/definition detail to be quantitatively interpretable and reproducible (notably Fig. 1, Fig. 3, Fig. 4, Fig. 5 as referenced in the structured report). For example: units/normalization of $N(>\tau)$ and $\tau$, how derivatives were taken for $dN/d\tau$, how $D_{\rm KL}$ was computed, and what exactly the Fisher curves assume are not always explicit in captions.
    Recommendation: Strengthen figure captions to include (i) axis units and normalization (per sightline? per $\Delta z$? total survey counts?), (ii) binning and numerical differentiation method (for $dN/d\tau$), (iii) the precise definition/normalization used for KL, and (iv) Fisher assumptions ($\tau$-range, binning, included parameters). Where possible, add uncertainty bands (Poisson-only and/or systematic bracketing) rather than only showing noiseless curves.
  • Instrumental/detection accessibility is not indicated on key distributions, especially at low $\tau$ where much of the information resides (Sec. 3.1–3.3; figures showing $dN/d\tau$ or $N(>\tau)$).
    Recommendation: Overlay illustrative detectability thresholds: e.g., a $\tau_{\rm min}$ corresponding to a representative spectral resolution/channel width and S/N, or show completeness curves. Alternatively, shade regions of $\tau$ that are likely inaccessible under plausible noise/continuum-fitting performance.
  • The Fisher setup uses $\{M_{\rm cut}, f_{\rm cool}, \bar{x}_{\rm HI}\}$ (Sec. 2.4), but $\bar{x}_{\rm HI}$ as a single uniform parameter may be inadequate near the end of reionization: patchiness/topology and fluctuating ionizing/X-ray backgrounds can change absorber statistics in a nontrivial, potentially shape-changing way.
    Recommendation: Add a short discussion (Sec. 3.3 or Conclusions) of how patchy reionization and spatially varying backgrounds could impact $dN/d\tau$ and degeneracies with $M_{\rm cut}$, with citations to relevant 21 cm forest/reionization modeling literature. If feasible, include a simple bracketing scenario (e.g., two reionization histories or an added “patchiness variance” nuisance) to test stability of the main conclusions.
  • The text’s statement about Poisson scaling is somewhat inconsistent/implicit: absolute uncertainties scale as $\sqrt{N_k}$, while fractional uncertainties scale as $1/\sqrt{N_k}$, and $N_k$ should scale with $N_{\rm los}$ (Sec. 2.4).
    Recommendation: In Sec. 2.4, explicitly distinguish absolute vs fractional uncertainties, and write $N_k \propto N_{\rm los}$ so the scaling of Fisher information with $N_{\rm los}$ is transparent.
  • The focus on particular redshifts (e.g., $z=10$ examples in Sec. 3.1 and Sec. 3.4) is not fully aligned with the paper’s own statement that the optimal window depends on $N_{\rm los}(z)$ and may prefer $z\approx8$–$9$ in some cases (Sec. 3.3).
    Recommendation: Briefly justify choosing $z=10$ as a benchmark (e.g., literature convention or balancing intrinsic signal and source counts), and/or add one additional benchmark redshift within the preferred window ($z=8$ or $9$) to show that key qualitative statements about $dN/d\tau$ shape, $D_{\rm KL}$ trends, and $\sigma(M_{\rm cut})$ persist.
  • Cosmological parameters used for the HAYASHI runs and for the $M_{\rm cut}$–$\sigma_0/m_\chi$ mapping are not explicitly listed (Sec. 2.1–2.2), limiting reproducibility and making it hard to judge sensitivity to, e.g., $\sigma_8$ and $n_s$ (important for low-mass halo abundance).
    Recommendation: Add a short paragraph in Sec. 2.1 stating the fiducial cosmological parameters used. Optionally comment qualitatively (or with a quick test) on how varying $\sigma_8$/$n_s$ within current uncertainties would shift $dN/d\tau$ and inferred $M_{\rm cut}$ sensitivity.
  • A numerical consistency check flags ambiguous “percent reduction vs remaining fraction” statements (reported in the structured report as $50.2\%$/$47.0\%$ reductions not matching implied remaining fractions).
    Recommendation: Re-check the relevant text/annotations and ensure reductions vs remaining fractions are defined and computed consistently (and that figure annotations match the intended definition).
  • Some claims in Sec. 3.2 and the Conclusions that astrophysical uncertainties “typically rescale the amplitude” (and thus are easily distinguished from $M_{\rm cut}$ shape changes) are stronger than what is demonstrated with the current nuisance parameter set.
    Recommendation: Soften/qualify this language unless additional nuisance parameters/tests are added. Explicitly state the scope of the robustness claim: e.g., “for the limited set of astrophysical degrees of freedom modeled here.”
Very Minor Issues:
  • Notation inconsistency: $f_{\rm cool}$ is defined as $T_k/T_k^{\rm ad}$ (Sec. 2.2) but appears as $T_k/T_k^{\rm ud}$ in Sec. 3.1; also $f_{\rm cool}$ is called a “cooling fraction” though it is a temperature ratio and could exceed $1$ if heating is allowed.
    Recommendation: Standardize to $T_k^{\rm ad}$ everywhere and either (i) state the intended domain (e.g., $0<f_{\rm cool}\leq1$) or (ii) rename to “temperature ratio” if values $>1$ are allowed.
  • The cross-section parameterization $\sigma = \sigma_0 v^n$ does not specify whether $v$ is dimensional or normalized by $c$, which determines the units of $\sigma_0$ (Sec. 2.2).
    Recommendation: State the convention for $v$ (e.g., $v/c$) and the resulting units of $\sigma_0$; ensure this is consistent with the $\sigma_0/m_\chi$ values quoted in Sec. 3.4.
  • Minor presentation/typography inconsistencies (quotation marks around “fingerprint”, hyphenation such as “Coulomb-like”/“velocity-independent”, “21 cm” formatting, occasional formatting artifacts in headings, acronym re-definitions).
    Recommendation: Proofread for consistent typography and acronym definitions (CMB, CDM, KL) and ensure figure labels/legends are readable and consistent (including colorblind-safe styling where possible).

Mathematical Consistency Audit

Mathematics Audit by Skepthical

This section audits symbolic/analytic mathematical consistency (algebra, derivations, dimensional/unit checks, definition consistency).

Maths relevance: light

The paper’s analytical content is primarily definitional and forecasting-based: it introduces a phenomenological halo-mass cutoff $M_{\rm cut}$, a cooling parameter $f_{\rm cool}$, uses a Poisson-count Fisher matrix (Eq. 1), an empirical redshift evolution for the number of lines of sight $N_{\rm los}(z)$ (Eq. 2), and invokes KL divergence as a shape distinguishability metric. Several central mappings/definitions needed to verify key conclusions (KL definition/normalization; $M_{\rm cut}\leftrightarrow$cross-section relation) are not provided in the text shown.

Checked items

  1. Cross-section velocity power law (Sec. 2.2, p.3)

    • Claim: The DM–baryon elastic scattering cross-section can be modeled as $\sigma = \sigma_0 v^n$, focusing on $n=0$ and $n=-4$.
    • Checks: notation consistency, dimensional/units sanity
    • Verdict: PASS; confidence: medium; impact: minor
    • Assumptions/inputs: $v$ denotes the DM–baryon relative velocity under some convention (not specified)., $\sigma_0$ is a normalization constant whose units depend on the choice of $v$ units.
    • Notes: Form is internally consistent, but the units of $\sigma_0$ cannot be checked without specifying whether $v$ is dimensionless or has physical units.
  2. Cooling fraction definition (Sec. 2.2, p.3)

    • Claim: Thermal impact is parameterized by $f_{\rm cool} = T_k / T_k^{\rm ad}$.
    • Checks: definition consistency, dimensional/units sanity
    • Verdict: PASS; confidence: high; impact: minor
    • Assumptions/inputs: $T_k$ and $T_k^{\rm ad}$ are both kinetic temperatures defined at the same redshift/epoch.
    • Notes: Dimensionless ratio is consistent. Domain/interpretation (cooling-only vs possible heating) is not specified.
  3. Fisher matrix for Poisson counts (Eq. (1), Sec. 2.4, p.4)

    • Claim: $F_{ij} = \sum_k \frac{1}{N_k} \left(\frac{\partial N_k}{\partial \theta_i}\right)\left(\frac{\partial N_k}{\partial \theta_j}\right)$ for binned absorber counts $N_k$.
    • Checks: algebra/derivation sanity, assumption check, symbol consistency
    • Verdict: PASS; confidence: medium; impact: moderate
    • Assumptions/inputs: Bins $k$ are statistically independent., Counts in each bin are Poisson with ${\rm Var}(N_k)=N_k$., $N_k$ represents expected counts at fiducial parameters (typical Fisher convention).
    • Notes: Equation is standard under Poisson assumptions, but the paper does not explicitly state whether $N_k$ is expected or observed, nor how $N_{\rm los}$ enters $N_k$. These omissions reduce verifiability but do not by themselves contradict Eq. (1).
  4. Lines-of-sight evolution model (Eq. (2), Sec. 2.4, p.4)

    • Claim: Number of lines of sight evolves as $N_{\rm los}(z) = 10 ((1+z)/8)^{-2.5}$.
    • Checks: dimensional/units sanity, symbol consistency, limiting/sanity check
    • Verdict: PASS; confidence: high; impact: minor
    • Assumptions/inputs: $N_{\rm los}$ is dimensionless and intended as an effective number of usable background sources/LOS.
    • Notes: Dimensionless and monotone-decreasing with $z$ as intended. No internal inconsistencies found.
  5. Poisson scaling statement vs Poisson statistics (Sec. 2.4, p.4)

    • Claim: Statistical uncertainty on $N_k$ is Poissonian, scaling with $1/\sqrt{N_{\rm los}}$.
    • Checks: consistency with definitions, algebraic/statistical scaling sanity
    • Verdict: FAIL; confidence: high; impact: minor
    • Assumptions/inputs: $N_k$ is a count aggregated over $N_{\rm los}$ independent lines of sight., Per-LOS expected counts are approximately constant so $N_k \propto N_{\rm los}$.
    • Notes: For Poisson counts, absolute uncertainty scales as $\sqrt{N_k}$ (thus typically $\propto \sqrt{N_{\rm los}}$ if $N_k \propto N_{\rm los}$), while fractional uncertainty scales as $1/\sqrt{N_k}$ (thus $\propto 1/\sqrt{N_{\rm los}}$). The text appears to conflate absolute and fractional uncertainty.
  6. KL divergence applicability to $dN/d\tau$ (Sec. 2.3 and Sec. 3.2, pp.3–5; Fig. 3 caption p.7)

    • Claim: $D_{\rm KL}({\rm CDM} \parallel M_{\rm cut})$ quantifies distinguishability between the differential optical depth distributions.
    • Checks: definition consistency, normalization/constraints, missing-steps audit
    • Verdict: UNCERTAIN; confidence: high; impact: critical
    • Assumptions/inputs: A KL divergence is computed between two probability distributions over $\tau$ (or over bins)., The 'differential optical depth distribution' is converted into a normalized distribution (not shown).
    • Notes: KL divergence requires normalized probabilities; $dN/d\tau$ is not inherently normalized (it is a differential count/rate). The paper does not provide the formula for $D_{\rm KL}$ or the normalization procedure, so the claimed monotonicity/interpretation cannot be verified analytically.
  7. Sharp cutoff implementation in halo mass function (Sec. 2.2, p.3)

    • Claim: Structure suppression is modeled by introducing a sharp cutoff in the halo mass function below $M_{\rm cut}$ (no halos for $M < M_{\rm cut}$).
    • Checks: definition consistency, missing-steps audit
    • Verdict: UNCERTAIN; confidence: medium; impact: moderate
    • Assumptions/inputs: Baseline halo mass function exists in the HAYASHI framework., Cutoff is implemented as a Heaviside-like truncation.
    • Notes: Mathematically plausible, but the exact functional form (e.g., $dn/dM$ multiplied by $\Theta(M-M_{\rm cut})$) and any smoothing are not specified, so the analytic consequences for $dN/d\tau$ cannot be checked from the provided text.
  8. Mapping $M_{\rm cut}$ to $\sigma_0/m_\chi$ constraints (Sec. 2.2 (statement) and Sec. 3.4 (results), pp.3 and 8)

    • Claim: $M_{\rm cut}$ is directly related to $\sigma$, enabling constraints on $\sigma_0/m_\chi$ from sensitivity to $M_{\rm cut}$.
    • Checks: missing-steps audit, symbol/definition consistency
    • Verdict: UNCERTAIN; confidence: high; impact: critical
    • Assumptions/inputs: There exists a deterministic relation $M_{\rm cut}(\sigma_0/m_\chi, n, z, ...)$.
    • Notes: No equation or derivation is provided for $M_{\rm cut}(\sigma_0/m_\chi)$. Without this mapping, the translation in Sec. 3.4 is not verifiable on symbolic/analytic grounds.

Limitations

  • Only the equations and definitions explicitly present in the provided PDF text (Eq. (1) and Eq. (2) plus a few inline definitions) can be audited; key mathematical definitions (e.g., explicit KL divergence formula, normalization) are absent.
  • Figures are referenced for derivatives and trends, but no underlying analytic expressions for $N(>\tau)$, $dN/d\tau$, or their dependence on parameters are provided, preventing step-by-step derivation checks.
  • The paper’s central physical-to-phenomenological mapping (DM scattering $\rightarrow$ power-spectrum suppression $\rightarrow M_{\rm cut}$) is asserted but not written as mathematics in the provided text, so internal consistency of that chain cannot be assessed.

Numerical Results Audit

Numerics Audit by Skepthical

This section audits numerical/empirical consistency: reported metrics, experimental design, baseline comparisons, statistical evidence, leakage risks, and reproducibility.

10 candidate numerical checks were executed: 8 PASS and 2 FAIL. Passed items include a monotonicity evaluation of $N_{\rm los}(z)$ from Eq. (2), exact percent-to-fraction conversion for $0.002\%$, parameter-range sanity for $f_{\rm cool} = 0.02$, internal consistency/ordering of quoted Fisher uncertainties, scientific-notation parsing for two cross-section sensitivities, and a $17$-orders-of-magnitude comparison derived from those sensitivities. The only failures occurred in percent-reduction-to-remaining-fraction conversions for two stated reductions ($z = 7$ and $z = 10$).

Checked items

  1. C1_Nlos_powerlaw_eq2 (Page 4, Eq. (2))

    • Claim: The number of lines of sight is modeled as $N_{\rm los}(z) = 10 \times ((1+z)/8)^{-2.5}$.
    • Checks: formula_evaluation
    • Verdict: PASS
    • Notes: Computed $N_{\rm los}(7)=10.0$, $N_{\rm los}(10)=4.510692841903824$, $N_{\rm los}(15)=1.7677669529663689$; monotonic decreasing and positive.
  2. C2_absorber_reduction_Mcut1e4_z7 (Page 5, Section 3.1 (text))

    • Claim: A modest cutoff of $M_{\rm cut} = 10^4~M_\odot$ reduces the absorber count by $50.2\%$ at $z = 7$.
    • Checks: percent_to_fraction
    • Verdict: FAIL
    • Notes: Computed remaining fraction $r=1-50.2/100=0.498$; the check's reported expected_remaining_fraction was $1.0$, producing a mismatch.
  3. C3_absorber_reduction_Mcut1e4_z10 (Page 5, Section 3.1 (text))

    • Claim: A modest cutoff of $M_{\rm cut} = 10^4~M_\odot$ reduces the absorber count by $47.0\%$ at $z = 10$.
    • Checks: percent_to_fraction
    • Verdict: FAIL
    • Notes: Computed remaining fraction $r=1-47.0/100=0.53$; the check's reported expected_remaining_fraction was $1.0$, producing a mismatch.
  4. C4_cooling_fractional_change_bound (Page 5, Section 3.1 (text))

    • Claim: Across all redshifts studied, the maximum fractional change in the total number of absorbers due to cooling remains below $0.002\%$.
    • Checks: threshold_interpretation
    • Verdict: PASS
    • Notes: $0.002\%$ converts exactly to a fraction of $2\times10^{-5}$.
  5. C5_extreme_cooling_value (Page 5, Section 3.1 (text))

    • Claim: Even extreme cooling ($f_{\rm cool} \rightarrow 0.02$) produces no discernible change in $N(>\tau)$.
    • Checks: parameter_range_sanity
    • Verdict: PASS
    • Notes: Sanity check only: $f_{\rm cool}=0.02$ satisfies $0 < f_{\rm cool} \leq 1$.
  6. C6_Fisher_uncertainty_growth_ratio (Page 7, Section 3.3 (text) and Page 8, Figure 4 caption)

    • Claim: $\sigma(M_{\rm cut}) \approx 1127~M_\odot$ at $z = 7$ and $\approx 5018~M_\odot$ at $z = 15$ (uncertainty increases monotonically with redshift).
    • Checks: ratio_and_monotonicity_from_reported_points
    • Verdict: PASS
    • Notes: Positivity and ordering hold ($5018>1127$). Computed ratio $5018/1127 = 4.452528837622006$.
  7. C7_Fisher_uncertainty_z10_vs_z7 (Page 7, Section 3.3 (text))

    • Claim: $\sigma(M_{\rm cut}) \approx 1907~M_\odot$ at $z = 10$ while $\sigma(M_{\rm cut}) \approx 1127~M_\odot$ at $z = 7$.
    • Checks: ordering_check
    • Verdict: PASS
    • Notes: Ordering is consistent with monotonic increase in redshift: $1907>1127$ for $z=10>7$.
  8. C8_sigma0_over_mchi_n0_value (Page 8, Section 3.4 (text))

    • Claim: For velocity-independent ($n = 0$), a $10\%$ suppression measurement at $z = 10$ corresponds to $\sigma_0/m_\chi \approx 2.13 \times 10^{-26}~{\rm cm}^2/{\rm GeV}$.
    • Checks: scientific_notation_parse
    • Verdict: PASS
    • Notes: Parsed numeric value matches $2.13\times10^{-26}$.
  9. C9_sigma0_over_mchi_nneg4_value (Page 8, Section 3.4 (text))

    • Claim: For Coulomb-like ($n = -4$), the sensitivity reaches $\sigma_0/m_\chi \approx 2.13 \times 10^{-43}~{\rm cm}^2/{\rm GeV}$ (at $z = 10$ in the same sentence context).
    • Checks: scientific_notation_parse
    • Verdict: PASS
    • Notes: Parsed numeric value matches $2.13\times10^{-43}$.
  10. C10_orders_of_magnitude_claim_from_two_numbers (Page 8, Section 3.4 (text))

    • Claim: Compare the two quoted sensitivities: $2.13\times10^{-26}$ vs $2.13\times10^{-43}~{\rm cm}^2/{\rm GeV}$; they differ by $17$ orders of magnitude.
    • Checks: orders_of_magnitude_difference
    • Verdict: PASS
    • Notes: Computed $\log_{10}$ ratio equals $17.0$ exactly given equal mantissas.

Limitations

  • Only parsed text was available; figures are images without machine-readable numeric tables, so plot-based quantitative checks were not feasible.
  • Many central results (absorber counts, KL divergence values, Fisher derivatives) rely on model outputs not enumerated numerically in the text, preventing recomputation.
  • Feasible fast checks were limited to arithmetic conversions, inequality/monotonicity assertions from stated point estimates, and scientific-notation parsing.