{"id":"50ddd524-fca0-4d63-80b6-a81a29b6d30e","arxiv_id":"2411.12138","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"Gas fractions in Spiderweb protocluster H-alpha emitters drop steeply at stellar mass ~10^10.5 solar masses, and AGN candidates show the lowest gas fractions, suggesting feedback and overconsumption drive quenching.","lead":"Using 475 hours of ATCA CO(1-0) observations, this paper measures cold molecular gas in 43 star-forming H-alpha emitters inside the Spiderweb protocluster at z=2.16. It finds that gas fractions drop sharply above a stellar mass of about 10^10.5 solar masses, with AGN playing a role in the depletion.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The two-regime fgas result rests on positive-only CO(1-0) upper limits that are treated as measurements, so the steep transition may be partly an artifact of the flux-recovery method.","rationale":"The reader's weakest_assumption identifies exactly the most load-bearing vulnerability: the upper-limit method in Sect. 3.1 is systematically biased toward positive flux, and those upper limits populate the low-mass, high-fgas side of the claimed two-regime transition. The paper is transparent about the caveat, and the comparison in Fig. 4 validates the method on the 10 detections, but that validation does not test the behavior on non-detections where noise dominates. The stacking analysis in Sect. 4.2 shares the same positive-only integration, so it cannot independently rescue the low-mass bin. A concrete re-analysis with a censored or Bayesian upper-limit estimator, plus a refit of Eq. 5, would determine whether the steep transition survives. This is a conditional concern rather than a fatal one: the qualitative trend is consistent with prior work and with the massive-end AGN fraction, and the authors have provided enough detail for the test to be run. Therefore the reader's CONDITIONAL verdict remains appropriate; no verdict change is needed.","tokens_in":46959,"tokens_out":2446,"duration_ms":31722,"concrete_test":"Recompute all 33 CO(1-0) upper limits with a Bayesian or full-window integration that includes negative channels (e.g., a 3-sigma likelihood integral over +/-500 km/s, or posterior sampling with a normal noise model), then re-derive the Table 2 stacked gas fractions and re-fit Eq. 5 to the PKS1138 sample. If the low-mass stacked Fgas drops by more than ~0.2 or the fitted steepness B decreases by more than ~1 sigma, the two-regime conclusion is not robust to upper-limit treatment.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim that the 43 HAEs split into two gas-fraction regimes with a steep transition at log M*/Msun ~ 10.5 depends on the low-mass end being gas-rich (fgas near unity) and on the logistic fit to Eq. 5. The low-mass end is almost entirely composed of CO(1-0) upper limits computed in Sect. 3.1 by integrating only positive flux channels within +/-500 km/s and taking the maximum over a 3x3 spaxel grid. The authors explicitly note this can overestimate fluxes, but these overestimated limits are then plotted as fgas points and included, together with the 10 detections, in the fit that defines the transition. If positive noise fluctuations are preferentially selected, the low-mass fgas values and the steepness parameter B are biased upward. The problem is not just scatter: it is a systematic floor that pushes non-detections toward fgas ~ 1, exactly the trend that anchors the claimed low-mass regime. The stacking analysis in Sect. 4.2 uses the same positive-only integration, so the stacked low-mass Fgas = 0.86 +/- 0.04 is not an independent check of the individual upper limits. A censored-data treatment or a Bayesian upper-limit estimator could change both the location of the transition and the inferred gas-richness of the low-mass population, which would directly weaken the 'replenished by inflows' and 'rapid depletion above 10.5 Msun' narrative.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper presents ATCA CO(1-0) observations of 43 spectroscopically confirmed H-alpha emitters in the Spiderweb protocluster at z=2.16, of which 10 are detected in CO(1-0) and the rest are assigned upper limits. The central claim is that the molecular gas fraction, fgas = M_mol/(M_mol + M_star), shows two distinct regimes with a steep transition near log M_star/M_sun ~ 10.5: low-mass galaxies are gas-rich (fgas near unity) while massive galaxies, especially AGN candidates, have fgas below about 0.4. The authors interpret this as evidence for rapid cold-gas depletion via AGN feedback and overconsumption, with environmental gas removal playing a secondary role. They also present stacking results, depletion times, and environmental trends in phase space and local density.","tokens_in":47272,"tokens_out":3618,"duration_ms":39485,"significance":"If the two-regime fgas pattern is robust, this is one of the most complete molecular-gas censuses of a protocluster at cosmic noon, providing an important observational benchmark for galaxy formation models. The paper also gives a clear, falsifiable picture: low-mass protocluster galaxies are replenished by cosmic-web inflows, while massive galaxies quench through gas depletion aided by AGN. The comparison with coeval field and protocluster samples and the explicit discussion of accretion-regime changes are strengths. However, the central result rests on an upper-limit method that the authors themselves note can overestimate fluxes; because those upper limits are treated as measurements in the fgas analysis, the claimed steep transition is not yet demonstrated to the standard required for such a strong physical conclusion.","major_comments":[{"comment":"The upper limits for the 33 non-detections are computed by integrating only positive flux channels within ±500 km/s and by taking the maximum over a 3x3 spaxel grid. This is a deliberately conservative upper limit, but it is also a positively biased estimator: noise fluctuations push the derived flux upward. The text acknowledges this ('can result in a flux overestimation'). These biased upper limits are then plotted as fgas points and included in the logistic fit of Eq. 5 that defines the steep transition. Because low-mass sources are mostly non-detections, this systematically raises their fgas and can create or exaggerate a gas-rich low-mass regime. The paper needs a quantitative robustness check: for example, fitting Eq. 5 using only the 10 detections, or treating upper limits with a censored-data/survival-analysis approach, and reporting how A, B, and the transition location change.","section":"Sect. 3.1 and Fig. 6"},{"comment":"The stacked low-mass Fgas = 0.86 ± 0.04 is presented as supporting the individual upper-limit trend. However, the stacking analysis uses the same positive-only, maximum-spaxel prescription of Sect. 3.1. A positive-only stack of mostly undetected sources will produce a positive flux even if the true signal is zero, so this stacking result is not an independent confirmation. The authors should either demonstrate that a symmetric integration (or one that subtracts a negative-noise control) yields the same result, or explicitly state that the stacked value inherits the same bias.","section":"Sect. 4.2, Table 2"},{"comment":"The interpretation that low-mass HAEs have gas fractions 'close to unity' and are 'replenished by inflows' is a direct consequence of the positive-biased upper limits in §3.1. The same applies to the conclusion that massive galaxies deplete their gas rapidly above log M* = 10.5. The paper should either recalibrate the upper-limit method through simulations or a control sample (e.g., using negative-flux maps to estimate the bias), or soften these physical conclusions to what is actually supported by the detections alone.","section":"Sect. 4.1 and Sect. 5.1"},{"comment":"The logistic fit is performed simultaneously on this sample and the CLJ1001 sample of Wang et al. (2018), and the text states that removing Wang et al. changes A and B by 1 sigma and 0.5 sigma, respectively. This means the steepness parameter B is not strongly constrained. The claim of a 'steep' transition should be quantified with a confidence interval on B and the location of the transition, and the fit should be repeated excluding the upper limits to show that the transition is not an artifact of the limit distribution.","section":"Sect. 4.1, Eq. 5"}],"minor_comments":[{"comment":"The text states 'we have obtained 26 CO(1-0) upper flux limits for the rest of our sample,' while Sect. 4 and Table A.1 list 33 upper limits (including the four cases with S/N>4 that are still treated as upper limits). This inconsistency should be corrected.","section":"Conclusions, item 1"},{"comment":"The description of the upper-limit extraction does not specify how the local noise rms is measured for each spaxel, which is needed to reproduce the quoted median S/N of about 2.3. A precise definition (e.g., the rms in line-free channels of that spaxel) should be added.","section":"Sect. 3.1"},{"comment":"There is a typo in the text: 'Fgas = 0.86 ± 0.0.04' should read '0.86 ± 0.04'.","section":"Sect. 4.2"},{"comment":"The caption says the red solid line is the fit 'to the PKS1138 and CLJ1001 protocluster samples,' but it does not mention that Eq. 5 with A = 10.93 and B = -2.15 is used. Adding the parameters to the caption would make the figure self-explanatory.","section":"Fig. 6 caption"},{"comment":"There is a typo 'Sect.,3.1' in the first paragraph; it should be 'Sect. 3.1'.","section":"Sect. 4.1"}],"recommendation":"major_revision","confidential_remarks":"The paper is a solid observational contribution to an important topic, and the authors are transparent about the potential flux overestimation in their upper-limit method. However, the central two-regime fgas result and the associated quenching narrative are loaded on the treatment of non-detections as measurements. If the authors can demonstrate with a censored-data fit or an explicit bias correction that the transition and the low-mass gas-rich values survive, this would be a strong paper; as written, the main claim is not yet robust enough for publication. The manuscript fits the scope of A&A and the appendices are useful."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Punchline: this is a genuinely useful dataset – the largest CO(1-0) census of Hα emitters in a z>2 protocluster – and the environmental trends are worth having. But the central two-regime gas-fraction claim leans on an upper-limit method that overestimates fluxes, so the sharp transition at log M* ≈ 10.5 is not as secure as the paper suggests.\n\nWhat's new: 43 spectroscopically confirmed HAEs with CO(1-0) measurements or upper limits, 10 detections, a stacking analysis in three bins, and environmental trends (µ_gas, SFE vs η) that haven't been published for this sample. The authors are transparent about many caveats, which is a real positive. The comparison with literature samples and the Tacconi et al. (2018) scaling relations is fair and useful. This is a solid observational contribution.\n\nSoft spots: the upper limits for 33 of 43 sources are computed by integrating only positive flux channels within ±500 km/s and taking the maximum over a 3×3 spaxel grid. The authors themselves note this can overestimate flux, yet those limits are then plotted as fgas points and included — together with the 10 detections — in the logistic fit that defines the steep transition. The stacked low-mass Fgas = 0.86 uses the same positive-only integration, so it is not an independent check of the individual upper limits. A censored-data treatment (e.g., Bayesian upper limits) could shift both the transition mass and the inferred gas-richness of the low-mass population. The qualitative trend likely survives, but the quantitative sharpness is not proven. Minor: the logistic fit is done on the combined PKS1138 + CLJ1001 sample, with only 10 detections; the result depends on treating limits as measurements.\n\nWho it's for: observers working on high-z protoclusters, cold gas, and quenching. It deserves a serious referee. The data are valuable and the analysis is mostly careful, but the upper-limit method needs revision before the central claim can be accepted. Recommendation: send to peer review with a request for a censored-data analysis or an explicit demonstration that the bias does not change the conclusion.","headline":"Valuable new CO(1-0) data for a z>2 protocluster, but the headline two-regime gas fraction is partly built on biased upper limits and needs a censored-data reanalysis before the steep transition is taken at face value.","tokens_in":47977,"tokens_out":2467,"would_cite":true,"duration_ms":28968,"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":"The Spiderweb protocluster's H-alpha emitters split sharply by gas fraction near $\\log M_*/M_\\odot\\approx10.5$, and this paper argues that AGN feedback plus gas overconsumption, not slow starvation, drives their quenching.","keywords":["molecular gas","CO(1-0)","protocluster","galaxy quenching","AGN feedback","gas fraction","H-alpha emitters","Spiderweb protocluster"],"falsifier":"Re-observe the 33 non-detected H-$\\alpha$ emitters with higher sensitivity, for example with ALMA Band 1, and re-derive their CO(1-0) fluxes using a full Bayesian upper-limit treatment that includes negative noise channels; if the true gas fractions of the $\\log M_*/M_\\odot<10.5$ galaxies fall substantially below 0.8, the reported steep transition at $\\log M_*/M_\\odot\\approx10.5$ would not be as sharp as claimed.","tokens_in":46765,"feed_emoji":"🌌","tokens_out":9247,"duration_ms":93644,"temperature":0.7,"pith_summary":"Using ATCA CO(1-0) imaging of 43 spectroscopically confirmed H-$\\alpha$ emitters in the Spiderweb protocluster at $z=2.16$, this paper measures or limits the cold molecular gas masses of a forming cluster's typical star-forming galaxies. It finds that the sample splits into two gas regimes around $\\log M_*/M_\\odot\\approx10.5$: lower-mass galaxies have gas fractions at or near unity, while more massive galaxies often drop below $f_{\\rm gas}\\approx0.4$ and are frequently AGN (12 of 20). The authors argue that these massive galaxies will exhaust their fuel and become passive by $1<z<1.6$ on typical depletion times of 1-3 Gyr, and that rapid AGN feedback combined with overconsumption, not slow starvation, drives their quenching. This matters because it ties the emergence of the red sequence in massive clusters to a sharp, mass-dependent gas-depletion process already visible while the cluster is still assembling.","feed_headline":"A sharp mass threshold splits Spiderweb protocluster galaxies by gas content","feed_subtitle":"Low-mass H-alpha emitters stay gas-rich, while massive AGN are already gas-poor and quench within a few gigayears.","key_machinery":"The load-bearing measurement is the molecular gas fraction $f_{\\rm gas}=M_{\\rm mol}/(M_*+M_{\\rm mol})$, derived from CO(1-0) fluxes using a fixed conversion factor $\\alpha_{\\rm CO}=4.36\\,M_\\odot\\,({\\rm K\\,km\\,s^{-1}\\,pc^2})^{-1}$. Ten sources have detections from the COALAS catalog; for the other 33, upper limits are obtained by integrating only positive flux channels within $\\pm500$ km/s of each H-$\\alpha$ redshift and taking the brightest spaxel in a $3\\times3$ grid around the source. Stacking the spectra in three bins (low mass, massive non-AGN, massive AGN) gives average gas fractions of $0.86\\pm0.04$, $0.55\\pm0.04$, and $0.38\\pm0.05$, and a logistic function in the $f_{\\rm gas}$--$M_*$ plane places the transition at $\\log M_*/M_\\odot=10.93\\pm0.05$. The phase-space parameter $\\eta=(R/R_{200})(|\\Delta v|/\\sigma)$ then ties gas content to position within the forming cluster.","core_discovery":"This paper reports the cold molecular gas content, traced by CO(1-0), for 43 spectroscopically confirmed H-$\\alpha$ emitters in the Spiderweb protocluster at $z=2.16$, using 10 detections and 33 upper limits. It finds that the total gas fraction $f_{\\rm gas}=M_{\\rm mol}/(M_*+M_{\\rm mol})$ drops sharply from values near unity at $\\log M_*/M_\\odot<10.5$ to values often below 0.4 above that mass. At the massive end, 12 of 20 sources show AGN signatures, and stacking gives $f_{\\rm gas}\\approx0.55$ for massive galaxies without AGN and $f_{\\rm gas}\\approx0.38$ for AGN candidates. The paper argues that this sharp transition, combined with depletion times of 1-3 Gyr, means these galaxies will become passive by $1<z<1.6$, and that the quenching mechanism is rapid cold-gas depletion driven by AGN feedback plus overconsumption rather than starvation alone.","pith_inferences":["If the threshold is universal, deep CO(1-0) surveys in other $z\\approx2$ protoclusters and in the field should reproduce the same logistic break, and tracing how that break moves with redshift would separate a mass-driven quenching clock from an environment-driven one.","A direct test of the proposed mechanism would compare black-hole-to-stellar-mass ratios among the massive HAEs: the model predicts that the most gas-poor objects host the most massive black holes even when their current AGN activity is faint.","Combining the phase-space gas-fraction gradient with the detection of a nascent intracluster medium could yield a first empirical map of where cold streams stop penetrating a forming cluster, turning the proposed starvation cutoff into a measurable radius.","If a metallicity-dependent CO conversion factor is applied, the low-mass gas fractions would shift; the steep threshold could move in mass rather than disappear, so the claim should be re-tested with $\\alpha_{\\rm CO}$ variations."],"forward_implications":["If the gas-fraction break is real, the low-mass H-alpha emitters are still being replenished by cold inflows from the cosmic web, since their gas fractions approach unity even as they sit on the main sequence.","With typical depletion times of 1-3 Gyr and no further inflows, the Spiderweb HAEs would become passive by $1<z<1.6$, matching the epoch when massive clusters build their red sequences.","The stacked gas fractions ($F_{\\rm gas}\\approx0.55$ without AGN, $F_{\\rm gas}\\approx0.38$ for AGN candidates) imply that nuclear activity removes or heats a substantial fraction of the cold gas on top of ordinary star formation at $M_*\\gtrsim10^{10.5}\\,M_\\odot$.","Because star formation stays on the main sequence while gas fractions decline toward the core, environmental effects act through reducing the fuel supply rather than by directly boosting or suppressing star formation."],"supporting_citations":[{"why":"Supplies the ATCA CO(1-0) mosaic and the blind CO detections from which the ten dual H-alpha+CO sources are drawn.","marker":"Jin et al. (2021)"},{"why":"Provides the spectroscopic H-alpha redshifts, SFRs, metallicities, and AGN line diagnostics for the parent sample.","marker":"Pérez-Martínez et al. (2023)"},{"why":"Provides the CLJ1001 protocluster CO(1-0) comparison sample showing the same gas-fraction drop.","marker":"Wang et al. (2018)"},{"why":"Supplies the logistic functional form used to fit the sharp f_gas-mass transition.","marker":"Popping et al. (2012)"},{"why":"Defines the field main-sequence scaling relations for gas fraction and star formation efficiency against which the offsets are measured.","marker":"Tacconi et al. (2018)"},{"why":"Identifies the X-ray AGN among the HAEs, supporting the elevated AGN fraction at the massive end.","marker":"Tozzi et al. (2022b)"},{"why":"Provides the narrow-band H-alpha emitter selection in the Spiderweb field that defines the parent sample.","marker":"Koyama et al. (2013)"}],"fun_headline_variants":["Gas fraction cliff at 10.5 solar masses splits protocluster","Spiderweb protocluster: low-mass gas-rich, high-mass AGN quench","Rapid cold gas drain dooms massive Spiderweb galaxies","Sharp gas transition marks quenching in Spiderweb protocluster","Mass threshold decides gas richness in Spiderweb protocluster"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The sharp transition depends on treating 33 CO(1-0) upper limits as gas measurements, but those limits were made by summing only positive noise fluctuations within a velocity window and taking the brightest of nine spatial pixels, a procedure the paper itself notes can overestimate fluxes.","fun_headline_variants_meta":{"raw":{"variants":["Gas fraction cliff at 10.5 solar masses splits protocluster","Spiderweb protocluster: low-mass gas-rich, high-mass AGN quench","Rapid cold gas drain dooms massive Spiderweb galaxies","Sharp gas transition marks quenching in Spiderweb protocluster","Mass threshold decides gas richness in Spiderweb protocluster"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00033,"raw_usage":{"total_tokens":1969,"prompt_tokens":1204,"completion_tokens":765,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":820,"completion_tokens_details":{"reasoning_tokens":674}},"tokens_in":820,"tokens_out":765,"duration_ms":7830,"temperature":1.0,"reasoning_tokens":674,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T17:51:55.786870+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Re-observe the 33 non-detected H-$\\alpha$ emitters with higher sensitivity, for example with ALMA Band 1, and re-derive their CO(1-0) fluxes using a full Bayesian upper-limit treatment that includes negative noise channels; if the true gas fractions of the $\\log M_*/M_\\odot<10.5$ galaxies fall substantially below 0.8, the reported steep transition at $\\log M_*/M_\\odot\\approx10.5$ would not be as sharp as claimed.","supporting_citations":[{"cited_title":"2018, , 867, L29","cited_arxiv_id":null,"evidence_quote":"Provides the CLJ1001 protocluster CO(1-0) comparison sample showing the same gas-fraction drop."}],"review_version":1}