{"id":"3aa65d55-7d23-47f7-8822-445612603efc","arxiv_id":"2608.05281","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"At z=8-10, the strongest JWST damped Lyα absorbers require neutral gas inside and around galaxies, and the NINJA models still underproduce them unless additional unresolved birth-cloud gas is invoked.","lead":"Using the NINJA cosmological simulations, this paper builds mock Lyα absorption spectra for galaxies at redshifts 8 and 10 and argues that the strongest damped Lyα absorbers seen by JWST cannot come from the intergalactic medium alone. It quantifies how much absorption arises from the IGM, from gas around galaxies, and from gas inside galaxies, and concludes that the most extreme absorbers trace gas in and around the galaxies themselves.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"HII-region failure at >50 pkpc rests on a sharp fully-ionized boundary; residual neutral CGM within the bubble could reproduce f22 without birth clouds.","rationale":"The reader identified the flat intrinsic spectrum as the weakest assumption. That concern is legitimate: xp and the NHI mapping depend on the assumed continuum shape, and a strong Lyα emission line could fill the damping wing and shift f21/f22. However, that effect would make the simulated absorption appear stronger than reality, so it would only strengthen the conclusion that large HII regions fail to produce strong absorbers. The fully-ionized HII region assumption is more directly load-bearing: it defines the 50 pkpc threshold, and the paper's Case III only tests neutral gas within Rvir, leaving the dense CGM at 30-60 pkpc ionized. This paper's own radial profiles show overdense gas extending to ~100 pkpc, so the treatment of the r>Rvir portion of the bubble is a critical gap. A targeted numerical experiment varying x_int within RHII would settle whether the f22 shortfall truly requires sub-resolution birth clouds or can be explained by partially neutral CGM. Because the qualitative direction of the result (that galaxy-associated neutral gas is important) is likely robust, the CONDITIONAL verdict remains appropriate; the concern does not overturn the paper but sharpens the condition on the ionization structure of the HII region.","tokens_in":26204,"tokens_out":15110,"duration_ms":138704,"concrete_test":"Rerun the Case III analysis on the L50N1008 box at z=8, varying the residual neutral fraction x_int of all gas within RHII (not just within Rvir) for RHII=50, 100, and 200 pkpc, with x_int = 0.1, 0.3, and 1.0, while also testing the original Rvir-only case for comparison. Recompute f22 and compare with Mason et al. (2026) f22~18%. If f22 reaches ~18% for x_int~0.3 and RHII>50 pkpc, the central claim about HII region size is model-dependent; if f22 remains below the observed value unless x_int=1, the birth-cloud conclusion is supported.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim that HII regions larger than ~50 pkpc fail to reproduce the strongest absorbers (NHI>10^22 cm^-2) depends on the idealized assumption that all gas inside RHII is fully ionized. In Case II, the dense CGM between Rvir and RHII is forced to be ionized; in Case III, residual neutral gas is confined to within Rvir, leaving the region Rvir<r<RHII ionized. Since the simulated overdense gas extends to ~60-100 pkpc and Req is ~66-105 pkpc, the gas just outside Rvir is a plausible source of NHI>10^22 if partially neutral. The paper's ad hoc choice to assign neutral fraction xV_HI only inside Rvir, rather than to the entire HII region, may artificially suppress f22 for large RHII and bias the conclusion that unresolved stellar birth clouds are required. If this outer CGM were instead allowed a residual neutral fraction of a few tens of percent, the predicted f22 could plausibly match the observed ~18% without invoking birth clouds, undermining the stated constraint on HII region size.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper uses three NINJA cosmological hydrodynamical simulation boxes (L50N640, L50N1008, L140N1008) at z=8 and 10 to generate mock Ly-alpha absorption spectra toward simulated galaxies, with an idealized post-processing model in which each galaxy sits inside a spherical, fully ionized bubble of radius RHII (0-400 pkpc) embedded in a uniform IGM of neutral fraction xHI = 0.1, 0.8, or 1.0. Three scenarios are considered: IGM-only absorption (Case I), galaxy+IGM with fully ionized bubbles (Case II), and a variant in which gas within the virial radius retains a neutral fraction xV_HI while the region between Rvir and RHII stays fully ionized (Case III). Absorption strength is quantified by xp, the velocity offset at which the transmitted flux reaches 10%, and is mapped to an equivalent HI column density through a single-cloud calibration; the resulting statistics f21, f22, fDLA, and median NHI are compared with the JWST samples of Heintz et al. (2025), Mason et al. (2026), and Pollock et al. (2026). The central claims are that a uniformly ionized IGM or uniformly ionized HII regions larger than about 50 pkpc cannot reproduce the strongest absorbers (NHI > 10^22 cm^-2), that residual neutral gas within the virial radius is a necessary but insufficient ingredient at the resolved halo masses, implying an additional contribution from unresolved stellar birth clouds, and that the inferred NHI increases with stellar and halo mass.","tokens_in":26477,"tokens_out":14261,"duration_ms":123468,"significance":"If correct, the main conclusions are important: they imply that the strongest Ly-alpha damping wings at z=8-10 are dominated by neutral gas in the ISM/CGM of massive halos rather than by a uniformly neutral IGM, and they yield a falsifiable prediction (a positive correlation between inferred NHI and stellar/halo mass) that JWST stellar-mass measurements can test. The paper's strengths are its forward-modeling approach - the parameters xHI, RHII, and xV_HI are scanned rather than fitted to the target statistics, so the comparison with Heintz et al. (2025) and Mason et al. (2026) is not circular - and its careful cross-box convergence checks (Tables 1, 2, E.1, F.1), which show less than or about 1% agreement at z=10 and up to about 6% at z=8 between independent boxes in the overlapping halo-mass range. The direction of the mass-NHI correlation agrees with independent radiative-transfer zoom-in simulations.","major_comments":[{"comment":"The placement of the residual neutral gas boundary is an ad hoc modeling choice that is load-bearing for two headline claims: that HII regions larger than about 50 pkpc fail to produce the strongest absorbers, and that unresolved stellar birth clouds are needed to reach the observed f22. In Case III, gas between Rvir and RHII is forced to be fully ionized by fiat, even though the authors' own radial profiles (Figs. 5 and D.1; Section 4.1) show overdense gas extending to Req of about 66-105 pkpc, well beyond the virial radius of the relevant halos. No physical justification is given for choosing Rvir rather than RHII or Req as the boundary of the residual neutral gas. Quantitatively, the paper's own numbers show the sensitivity: Section 4.4 and Table 2 give f22 = 17% for Case II with RHII = 0 pkpc and xHI = 0.1, essentially matching the about 18% reported by Mason et al. (2026). A residual neutral fraction of order 10% in the overdense CGM outside Rvir would therefore plausibly bring Case III into agreement with the observed f22 for RHII = 50-100 pkpc without any contribution from unresolved birth clouds, which would weaken both the about-50-pkpc constraint and the birth-cloud inference. I recommend that the authors rerun Case III with the partial neutral fraction applied out to max(Rvir, Req), or better with a density-dependent neutral fraction following the self-shielding prescription (Rahmati et al. 2013) already used in the simulations, and report how f21, f22, and the quoted constraints shift.","section":"Section 4.5, Fig. 2, Section 6 item 4"},{"comment":"The assumed flat intrinsic spectrum with no Ly-alpha emission line is a load-bearing idealization for the quantitative comparison. The statistic xp is defined by the velocity offset at which the normalized transmitted flux reaches 10%, and the equivalent column density is read off the single-cloud xp-NHI calibration in Fig. 4; as the authors note in the Introduction, at NIRSpec PRISM resolution (R about 100) the damping wing is sampled by only a few resolution elements, so the inferred NHI and the resulting f21/f22 values are sensitive to the assumed continuum and line profile. An intrinsic Ly-alpha emission line of typical high-redshift equivalent width, or a P Cygni-type profile, would raise the flux near the wing and shift the 10% crossing point, systematically changing the inferred NHI for a given absorber. Because the absolute values of f21 and f22 are compared with Heintz et al. (2025) and Mason et al. (2026), this assumption should be stress-tested: the authors should recompute the xp distributions and f21/f22 statistics for a small grid of intrinsic profiles (for example pure absorption, Gaussian emission with EW of 20-50 Angstrom, and a broad P Cygni wing) and show that the conclusions survive, or otherwise quantify the systematic shift in the xp-NHI mapping.","section":"Section 3.2, Figs. 3-4"}],"minor_comments":[{"comment":"Please correct the typographical errors: 'Tabels' in Section 4.4, 'Summery' in Appendix E, 'and the the overall' in Section 3.3, the value '8.8' that should be '0.8' in the Fig. 2 caption and Section 3.2, and the garbled 'for ¸' in Section 4.4.","section":"Throughout"},{"comment":"The fractions f21, f22, and fDLA are quoted without confidence intervals; given the finite and partly correlated sightline samples (six axes per halo, and only 54-82 halos at z=10 per Table C.1), binomial uncertainties should be reported so that differences such as f22 = 3% versus 12% can be judged against the observed about 18%.","section":"Tables 2, E.1, F.1"},{"comment":"For column densities below about 2.6 x 10^21 cm^-2 the calibrated xp values are negative; the paper should state explicitly how the interpolation handles these cases and whether the weak-absorber end of the NHI distributions is treated as upper limits.","section":"Fig. 4, Section 3.3"},{"comment":"The estimate of about 400 pkpc for the ionized bubble assumes solar metallicity and unity escape fraction for a single halo; a brief discussion of how fesc and metallicity would reduce the maximum RHII would make the adopted 0-400 pkpc range more robust.","section":"Appendix A"},{"comment":"The column-density rescaling factors for L50N640 and L140N1008 (1.19 and 2.31) are calibrated to L50N1008, and the quoted slope uncertainties (for example B = 0.57 +/- 0.02) do not include the systematic uncertainty in this calibration; the mass-NHI slopes should be presented as indicative until the calibration is justified.","section":"Section 5.2"}],"recommendation":"major_revision","confidential_remarks":"The manuscript fits the scope of A&A and addresses a timely question given the JWST PRIMAL results. My principal concern is the Case III geometry: the Rvir boundary for the residual neutral gas is the pivot on which the 'birth clouds required' conclusion turns, and the requested variant runs (neutral fraction out to Req, or a density-dependent self-shielding prescription) are well within the scope of this paper, so a major revision is appropriate rather than rejection. The authors should also verify that the 2026-dated comparison references (Mason et al. 2026, Pollock et al. 2026, Heintz et al. 2025) correspond to the published or accepted versions, since several quoted statistics (for example f22 about 18%) are taken from those papers."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Informal read: the paper does something genuinely useful. It takes the NINJA simulations, generates mock Lyα damping-wing spectra at z=8 and 10 under a controlled grid of xHI and RHII, and produces f21/f22 and median NHI predictions that can be compared directly with JWST samples. The NHI–halo/stellar mass relation from the combined boxes is new as far as I know, and the convergence checks across L50N640, L50N1008, and L140N1008 are a real strength. The authors are also fair about their own limitations: they flag the flat continuum, the halo-center sightlines, the lack of birth clouds, and the idealized reionization treatment. That is honest work.\n\nThe central conclusion—a uniformly neutral IGM alone cannot make NHI>1e22, and galaxy-associated gas is required—survives contact with the paper. I agree with the reader's conditional verdict, but I want the soft spots named sharply.\n\nFirst and most important: the \"HII regions larger than ~50 pkpc fail\" result is tied to the toy ionization geometry. Case II sets every gas element inside RHII to zero neutral fraction. Case III allows residual neutral gas only inside Rvir, leaving the region between Rvir and RHII forced ionized. The simulations themselves show overdense gas extending to 60–100 pkpc, well beyond Rvir. Allow a modest residual neutral fraction in that outer CGM and f22 will rise; the authors' own Case II with RHII=0 and xHI=0.1 gives f22=17%, nearly the Mason et al. value. So the birth-cloud requirement is partly an artifact of where xV_HI is allowed to live, not a robust inference. The stress-test note you passed along is directionally right.\n\nSecond: the flat intrinsic spectrum and halo-center sightlines are acknowledged but still load-bearing. At R=100 the damping wing is a few resolution elements; the xp–NHI mapping depends on the assumed continuum. That sets the scale of the quantitative comparison.\n\nThird: the §5.2 rescaling factors are fit factors, and the reported f21/f22 have no error bars. The cross-box agreement is good, but I would not quote the fractions as precise.\n\nWho is this for: people working on JWST damping-wing interpretations and reionization constraints. It deserves a serious referee. Send it to review, with a request that the authors test residual neutral gas in the CGM outside Rvir and provide public data or a clearer reproducibility path. The qualitative conclusion will likely stand; the 50 pkpc threshold should be presented as model-dependent.","headline":"Solid forward-modeling study whose qualitative conclusion is right, but the 50 pkpc HII-region bound is softer than the abstract implies.","tokens_in":27004,"tokens_out":4279,"would_cite":true,"duration_ms":40176,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"At $z\\approx8$–$10$, the strongest Ly$\\alpha$ damping wings trace neutral gas in and around galaxies, not the diffuse IGM.","keywords":["Lyα damping wing","Epoch of Reionization","neutral hydrogen column density","JWST/NIRSpec PRISM","cosmological hydrodynamical simulations","circumgalactic medium","ionized bubbles","high-redshift galaxies"],"falsifier":"Measure stellar masses for the JWST galaxies that show $N_{\\rm HI}>10^{22}\\,{\\rm cm^{-2}}$: if the strongest absorbers are not preferentially hosted by the most massive halos, the paper's mass-dependent origin of these damping wings is falsified.","tokens_in":25961,"feed_emoji":"🌌","tokens_out":13976,"duration_ms":109621,"temperature":0.7,"pith_summary":"This paper asks what produces the broad Ly$\\alpha$ absorption seen redward of the Lyman-$\\alpha$ line in JWST spectra of galaxies at $z\\approx8$–$10$: neutral hydrogen spread through the intergalactic medium, or neutral gas belonging to the galaxies themselves. Using mock spectra from the NINJA cosmological hydrodynamical simulations plus an idealized model in which each galaxy sits inside an ionized bubble, the authors compare the predicted distribution of inferred H I column densities with observed absorber statistics. They find that moderate absorbers ($N_{\\rm HI}>10^{21}\\,{\\rm cm^{-2}}$) can be reproduced by several combinations of intergalactic neutral fraction and bubble size, so those systems do not uniquely measure reionization. The strongest absorbers ($N_{\\rm HI}>10^{22}\\,{\\rm cm^{-2}}$), however, require neutral gas within the virial radius of massive halos, and even then the predicted incidence falls short of the observed one, pointing to additional column density from unresolved stellar birth clouds. The conclusion determines how much of the damping wing can be read as a probe of reionization rather than of galaxy-scale gas.","feed_headline":"Neutral gas near galaxies, not the IGM, makes the deepest damping wings","feed_subtitle":"NINJA simulations show HII regions over ~50 pkpc cannot make the strongest NHI > 10^22 cm^-2 absorbers at z=8–10.","key_machinery":"The carrier of the argument is an idealized post-processing model of reionization: every galaxy sits inside a fully ionized spherical bubble of radius $R_{\\rm HII}$ embedded in an IGM with uniform neutral fraction $x_{\\rm HI}$; in Case III the gas inside the virial radius keeps a separate neutral fraction $x_{\\rm HI}^{V}$. Mock spectra are computed along halo-centered sightlines with SPH-smoothed density, temperature, and velocity fields, convolved to NIRSpec PRISM resolution ($R\\approx100$) with SNR 30 noise. The absorption strength is compressed into a single number, $x_p$, the velocity offset from the galaxy at which normalized transmitted flux reaches 10%. A precomputed Voigt-profile grid maps $x_p$ to an equivalent H I column density $N_{\\rm HI}$, and from that mapping the statistics $f_{21}$, $f_{22}$, and $f_{\\rm DLA}$ are derived, allowing direct comparison with JWST survey measurements.","core_discovery":"On the paper's own terms, the central discovery is that the Ly$\\alpha$ damping wing at $z\\approx8$–$10$ is a composite signal whose strongest part is produced by galaxy-scale neutral gas rather than by the uniformly ionized IGM. By comparing mock spectra built with three idealized models (IGM only; IGM plus a fully ionized H II region; and the same configuration with residual neutral gas inside the virial radius), the paper shows that the observed incidence of moderate absorbers ($N_{\\rm HI}>10^{21}\\,{\\rm cm^{-2}}$) can be reproduced by several combinations of IGM neutral fraction $x_{\\rm HI}$ and bubble radius $R_{\\rm HII}$, so those systems do not uniquely determine $x_{\\rm HI}$. The strongest absorbers ($N_{\\rm HI}>10^{22}\\,{\\rm cm^{-2}}$) essentially never appear in models with a uniformly ionized IGM or with $R_{\\rm HII}\\gtrsim50$ pkpc. They do appear once partially neutral gas within the virial radius is included, with the incidence increasing for more massive halos, yet the predicted $f_{22}$ still falls below the observed value, which the paper attributes to an additional unresolved neutral component associated with stellar birth clouds.","pith_inferences":["If the missing column density comes from unresolved stellar birth clouds, the damping wing becomes a partially star-formation diagnostic: inferred $N_{\\rm HI}$ should correlate with recent-star-formation tracers such as H$\\alpha$ or [O III] equivalent width, not only with stellar mass.","The authors leave implicit that the flat-continuum assumption attaches the $x_p$–$N_{\\rm HI}$ calibration to one intrinsic Ly$\\alpha$ profile; realistic asymmetric or redshifted emission would shift the inferred statistics at PRISM resolution, so the quoted $f_{21}/f_{22}$ numbers carry a systematic the paper does not propagate.","A testable selection effect follows: magnitude-limited JWST samples, which favor massive halos, should show higher $f_{22}$ than mass-matched volume-limited predictions, and this can be checked with existing stellar-mass measurements."],"forward_implications":["Because many $(x_{\\rm HI}, R_{\\rm HII})$ combinations reproduce the observed $f_{21}$, the incidence of $N_{\\rm HI}>10^{21}\\,{\\rm cm^{-2}}$ absorbers alone cannot be used to measure the IGM neutral fraction during the Epoch of Reionization.","A uniformly ionized IGM, or fully ionized bubbles larger than roughly 50 pkpc, cannot produce $N_{\\rm HI}>10^{22}\\,{\\rm cm^{-2}}$, so the strongest observed damping wings are direct evidence for neutral gas in the ISM/CGM of the galaxies themselves.","Including residual neutral gas inside the virial radius raises $f_{22}$ from below 1% to 3% for the full sample, and to about 12% for halos in the mass range $10^{10.5}$–$10^{11.5}\\,M_\\odot$ at $R_{\\rm HII}=50$ pkpc, still short of the observed ~18%.","The inferred $N_{\\rm HI}$ increases with stellar and halo mass, so magnitude-limited JWST samples that preferentially select massive halos should show systematically stronger damping wings.","At $z\\sim10$ the H II radius needed to match the observed $f_{21}$ is roughly twice as large as at $z\\sim8$, consistent with a more neutral IGM at higher redshift."],"supporting_citations":[{"why":"Provides the observed incidence of $N_{\\rm HI}>10^{21}\\,{\\rm cm^{-2}}$ in $z>8$ JWST galaxies (65–90%) that the simulated $f_{21}$ values are compared against.","marker":"Heintz et al. (2025)"},{"why":"Provides the observed incidence of $N_{\\rm HI}\\ge10^{22}\\,{\\rm cm^{-2}}$ (~18%) and the forward spectral modeling context for the $f_{22}$ comparison.","marker":"Mason et al. (2026)"},{"why":"Supplies median $N_{\\rm HI}$ upper limits and host halo mass estimates at $z\\sim9$–$12$ used in the Case III halo-mass comparison.","marker":"Pollock et al. (2026)"},{"why":"Presents the NINJA simulation suite, including the simulation boxes, resolutions, and galaxy formation physics adopted here.","marker":"Behera et al. (2026)"},{"why":"Shows that simulated damping-wing profiles around halos are stronger than IGM-only predictions, motivating the CGM contribution modeled in Case II and III.","marker":"Keating et al. (2024a)"},{"why":"Provides median ionized bubble sizes during reionization that frame the range of $R_{\\rm HII}$ explored in the toy model.","marker":"Keating et al. (2024b)"},{"why":"Gives independent JWST-based constraints on $x_{\\rm HI}$ and bubble sizes that set the reionization context for the paper's scenarios.","marker":"Umeda et al. (2026)"},{"why":"The reionization simulations used by Mason et al. (2026) to infer intrinsic absorber column densities, defining the baseline for the $f_{22}$ comparison.","marker":"Lu et al. (2024)"}],"fun_headline_variants":["Neutral gas in galaxies beats IGM for deepest damping wings","Large HII regions fail to explain the most extreme damped absorbers","Stellar birth clouds may supply the missing neutral gas for deep wings","Z~8-10 galaxies need extra neutral gas to make extreme damping wings"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the galaxy's intrinsic spectrum is flat, with no Ly$\\alpha$ emission or absorption line: because the 10%-transmission velocity used to infer column density is measured against that assumed continuum at $R\\approx100$, a systematically different intrinsic Ly$\\alpha$ profile would shift all the predicted $f_{21}$ and $f_{22}$ statistics.","fun_headline_variants_meta":{"raw":{"variants":["Neutral gas in galaxies beats IGM for deepest damping wings","Large HII regions fail to explain the most extreme damped absorbers","Stellar birth clouds may supply the missing neutral gas for deep wings","Z~8-10 galaxies need extra neutral gas to make extreme damping wings"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001632,"raw_usage":{"total_tokens":6649,"prompt_tokens":1264,"completion_tokens":5385,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":880,"completion_tokens_details":{"reasoning_tokens":5308}},"tokens_in":880,"tokens_out":5385,"duration_ms":32692,"temperature":1.0,"reasoning_tokens":5308,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-08T16:15:11.370045+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure stellar masses for the JWST galaxies that show $N_{\\rm HI}>10^{22}\\,{\\rm cm^{-2}}$: if the strongest absorbers are not preferentially hosted by the most massive halos, the paper's mass-dependent origin of these damping wings is falsified.","supporting_citations":[],"review_version":1}