{"id":"51e82299-369a-4dd3-9c7f-6bcac71150c5","arxiv_id":"1908.04813","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"Observed passive fractions of local central galaxies track stellar and bulge mass, but the L-GALAXIES and SAGE simulations predict passive fractions driven by halo, black hole, and bulge properties, leaving clear tensions with SDSS data.","lead":"Massive galaxies that have stopped forming stars in the local universe line up with stellar and bulge mass, while two leading computer models tie the shutdown to halo, black hole, or bulge properties and miss key observed galaxies. The study tests whether supermassive black hole feedback in simulations can explain why big galaxies go red and quiet.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The adaptive central-galaxy selection is validated only on simulations; observed purity is assumed, so the SDSS/SAM passive-fraction comparison could be partially driven by selection systematics.","rationale":"The paper's central comparison rests on selecting 'central' galaxies identically in SDSS and in two SAMs. Appendix A tests the adaptive aperture against simulation truth and shows reasonable purity and completeness in the models, but there is no equivalent validation for observed galaxies. Because selection purity can vary with stellar mass and environment, and because the same aperture is used to define both the sample and the environment density, a systematic difference between observed and model central selection would directly shift the passive-fraction maps and the inferred correlations with M*, Mhalo, MBH, and MBulge. This is the weakest load-bearing link. I do not see an internal inconsistency or a red flag that would reject the paper; the qualitative trends are plausible and the authors are candid about tensions and about SAGE's calibration to the MBH-MBulge relation. However, the abstract's strong phrasing ('show a good correlation with stellar mass and bulge mass') somewhat overstates what is demonstrated, particularly because the observed bulge-mass correlation is largely read off model contours rather than measured SDSS bulge masses. This supports the existing CONDITIONAL verdict rather than changing it.","tokens_in":19611,"tokens_out":4558,"duration_ms":48424,"concrete_test":"Construct an SDSS-like mock light cone from one of the SAMs (or an observed group catalogue such as Yang et al. 2007) and compute the purity and completeness of the mass-rank central selection against known true centrals as a function of stellar mass and neighbour density. Then recompute the observed passive-fraction maps using only group-defined centrals and compare them with the mass-rank maps. If fpass differs by more than ~0.1 in the bins that drive the claimed correlations (high M*, high density, and massive-isolated systems), the central-galaxy selection is not safely comparable across SDSS and the models.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim depends on comparing passive fractions for 'central' galaxies in SDSS and in two SAMs, selected with the same adaptive aperture (Eq. A1 with n=8, rmax=2.5 Mpc, vdepth=2000 km/s). Appendix A tests the purity and completeness of this selection only against simulation truth; the observed SDSS sample has no equivalent validation. If the mass-rank aperture misclassifies satellites as centrals (or misses true centrals) with a different efficiency as a function of stellar mass and neighbour density than in the models, the passive-fraction maps in the M*–Sigma plane could be biased in a way that mimics or hides model-data tension. The paper itself notes in Appendix A that contamination can affect the trends and lead to inaccurate conclusions, but no observed-purity check is provided. This is especially relevant in the high-density/low-mass and massive-isolated regions where the claimed discrepancies are most visible. Without a demonstration that observed and model central selections have comparable purity across the plotted parameter space, the central comparison is not fully secured.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper compares passive fractions of local central galaxies in SDSS DR7 with predictions from two semi-analytic models, L-GALAXIES (Henriques et al. 2015) and SAGE (Croton et al. 2016), using the same adaptive cylindrical aperture to select centrals in both data and models. The observed passive fractions are studied in the stellar mass versus neighbour density plane and in relation to model halo mass, black hole mass, and bulge mass. The authors report that observed passive fractions correlate best with stellar mass and bulge mass; L-GALAXIES passive fractions correlate with halo and black hole mass; SAGE passive fractions correlate with black hole and bulge mass; and that neither model fully reproduces the observed quenched population, especially massive isolated galaxies. They also compare the black hole to bulge mass relation in the models with recent observational fits, finding SAGE in better agreement, while noting tensions remain.","tokens_in":19855,"tokens_out":3262,"duration_ms":38898,"significance":"If the central claim holds, the paper provides a useful observational constraint on AGN feedback prescriptions in semi-analytic models, with a concrete technique for selecting central galaxies uniformly in observations and simulations. The use of the same adaptive aperture in SDSS and in both SAMs is a genuine methodological contribution, and the paper is honest about residual tensions. However, the headline conclusions are largely qualitative: passive fractions are shown without uncertainties, the claimed correlations are not quantified, and part of the model-data agreement is explicitly attributable to model calibration. The significance is therefore moderate: the paper is a useful test case for radio-mode AGN feedback, but its evidential weight is weaker than its conclusions imply.","major_comments":[{"comment":"The adaptive central-galaxy selection is validated only against simulation truth, so the observed SDSS sample is assumed to have the same purity and completeness as the model samples. The central comparison depends on matching central selection in data and models; differential contamination as a function of stellar mass and neighbour density could bias the passive-fraction maps and mimic or hide model-data tension. The paper itself states in Appendix A that contamination 'can affect the trends seen in passive fractions and leads to inaccurate conclusions.' Please add an observed-side validation, e.g., using group catalogues or varying r(n, rmax, vdepth) and showing that the maps and conclusions are stable, or explicitly quantify the expected contamination in the SDSS sample.","section":"Appendix A"},{"comment":"Passive fractions are presented without error bars, bootstrap uncertainties, or significance tests. Equation (8) defines a weighted observed fraction, but no statistical uncertainty is propagated, and claims such as 'good correlation with stellar mass and bulge mass' are not supported by any quantitative correlation measure. Without this, it is not possible to tell whether the qualitative differences between L-GALAXIES, SAGE, and SDSS are statistically meaningful. Please add uncertainties and report appropriate significance tests (e.g., rank correlations or binomial errors per bin) for the central claims.","section":"Section 5, Figs 2-8"},{"comment":"The better agreement of the SAGE black hole-bulge mass relation with observations is explicitly attributed in Section 6 to the model having been calibrated to reproduce that relation, and L-GALAXIES/H15 was tuned to match observed passive fractions (Section 2.1). The paper should therefore not present these agreements as independent support for the physical prescriptions. The calibration-dependent parts of the comparison need to be clearly separated from genuinely predictive statements, such as the distribution of quenched galaxies in the M*-Sigma plane and the behaviour of massive isolated galaxies, which are the more convincing elements of the analysis.","section":"Section 6 and Section 7"},{"comment":"Bins with fewer than 30 objects are dropped post hoc and are not marked in the figures. This affects the appearance of the maps, particularly in the low-mass, high-density region, and the threshold choice is not justified. The paper should either report the excluded bins in the figures, show that the qualitative results are robust to the threshold value, or justify the threshold a priori.","section":"Section 5.1"},{"comment":"The passive definition is specified for the observed sample (sSFR < 0.3 t_hubble^-1 ~ 10^-11 yr^-1) but not for the model galaxies. Since the central claim is a comparison of passive fractions, it is essential to state explicitly whether the same sSFR threshold is applied to the SAM outputs or whether the models use a different quiescent/star-forming classification. If different definitions are used, the comparison is not uniform and the reported tensions may be partly definitional.","section":"Section 3 and Section 5"}],"minor_comments":[{"comment":"The abstract contains subject-verb agreement errors ('passive fractions in L-GALAXIES correlate...', 'For SAGE, the passive fraction correlate...'); these should be corrected.","section":"Abstract"},{"comment":"The sentence 'The datasets provides the number of neighbours...' has a subject-verb agreement error; also, the catalogue construction from Wilman et al. (2010) could be described more precisely.","section":"Section 3"},{"comment":"The phrase 'A central galaxies population' is ungrammatical and should read 'A central galaxy population' or 'The population of central galaxies'.","section":"Section 5.3"},{"comment":"The caption refers to left and right panels ('The left panel shows... the right panel shows...'), but the figure appears to be a single panel; please correct the caption or the figure layout.","section":"Figure 9"},{"comment":"The parameters alpha and beta in Eq. A1 are introduced but their values are never given or justified; please state the adopted values for the adaptive aperture used in the analysis.","section":"Appendix A, Eq. A1"},{"comment":"There are several typographical issues with accented characters (e.g., 'sersic', 'Beifiori') and inconsistent spacing around equations; a thorough proofreading pass is recommended.","section":"Throughout"}],"recommendation":"major_revision","confidential_remarks":"The paper is likely publishable after substantial revision. The main concerns are statistical rigour and calibration circularity rather than a fundamental error in the central idea. I would not reject, but the authors need to either strengthen the quantitative analysis or soften the conclusions accordingly."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThis paper is worth a look if you work on SAMs and quenching: it applies a uniform adaptive-aperture central selection to SDSS DR7 and two semi-analytic models (L-GALAXIES and SAGE) and asks which galaxy property drives passive fractions. The answer is clear and honestly reported: observed passive fractions track stellar and bulge mass, L-GALAXIES tracks halo and black hole mass, SAGE tracks black hole and bulge mass. Neither model fully matches the observed quenched population. That comparative result is real and not present in the earlier literature.\n\nWhat the paper does well is the construction of the comparison. The Fossati et al. central selection is applied identically to observations and models, which is a step up from comparing centrally-averaged properties, and the authors are transparent about where the models fail. The SAGE versus L-GALAXIES contrast is useful because the two radio-mode feedback schemes are genuinely different.\n\nThe soft spots are, in order of importance. First, the headline SAGE agreement in the MBH-MBulge slope is a calibration output: Section 6 explicitly says SAGE was tuned to reproduce that relation. The paper is honest about this, but the abstract's 'agrees better with observations' framing overstates the independent predictive power. Second, the observed central selection has no purity check. Appendix A validates the aperture against simulation truth only, and the paper itself notes contamination can alter the trends; whether the observed sample has comparable purity across the M*-Sigma plane is assumed. That could bias the model-data comparison in the exact regions where the discrepancies appear. Third, the passive fractions are shown without error bars and correlations are judged by eye, with bins below 30 objects dropped post hoc. A few scatter plots or a simple significance test would firm up the central claims.\n\nNone of these are fatal. The central qualitative statement - that neither SAM captures the observed dependence of quenching on stellar/bulge mass - is supported well enough to survive. But the specific quantitative agreements and the claimed correlations should be read as suggestive rather than measured.\n\nWho benefits? People comparing SAMs to SDSS, and anyone thinking about how to define central galaxies observationally. I would send this to a competent referee, yes. It is not a breakthrough, but it is a clean comparative study with honest caveats, and it deserves proper review rather than a desk rejection.","headline":"Useful and honest model-data comparison, though the headline agreements are partly calibration and the observed central selection lacks a purity check.","tokens_in":20387,"tokens_out":2425,"would_cite":true,"duration_ms":26299,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"The paper shows that two semi-analytic models with different AGN feedback prescriptions fail to reproduce the passive fractions of local central galaxies, and concludes that radio-mode black hole feedback alone does not explain quenching.","keywords":["galaxy quenching","AGN feedback","semi-analytic models","active galactic nuclei","passive galaxies","black hole-bulge relation","SDSS central galaxies","radio mode feedback"],"falsifier":"A definitive test would be to build a large, complete sample of local massive central galaxies with reliable group halo masses from redshift surveys or X-ray groups and measure passive fraction at fixed stellar mass: if f_pass rises steeply with halo mass at fixed stellar mass, the paper's claim that observed quenching is driven by stellar and bulge mass rather than by halo mass would be overturned. Alternatively, if direct dynamical black hole masses show that the observed passive fraction is actually set by black hole mass at fixed bulge mass, then the AGN-feedback picture the models embody would be supported.","tokens_in":1989,"feed_emoji":"🌌","tokens_out":2454,"duration_ms":91039,"temperature":0.7,"pith_summary":"The paper asks whether energy injected by a supermassive black hole in 'radio mode' feedback is what shuts down star formation in massive central galaxies in the local Universe. Using a uniform, observationally motivated way to pick central galaxies from both the SDSS and two semi-analytic models, the authors measure the passive fraction as a function of stellar mass, neighbour density, halo mass, black hole mass, and bulge mass. They find that observed passive fractions track stellar mass and bulge mass, while the L-GALAXIES model links quenching to halo and black hole mass and the SAGE model links it to black hole and bulge mass. Because neither model fully reproduces the observed passive population, and both keep massive isolated galaxies forming stars, the paper concludes that current radio-mode AGN feedback prescriptions do not fully explain quenching in local central galaxies.","feed_headline":"Black hole feedback doesn't fully explain quiet galaxies","feed_subtitle":"SDSS comparison shows two galaxy formation models miss the quenched centrals seen in the local universe.","key_machinery":"The load-bearing device is the adaptive cylindrical aperture used to identify central galaxies by stellar-mass rank in both data and models. For each galaxy, a cylinder of radius $r(n, r_{\\rm max}, v_{\\rm depth}) = \\min(r_{\\rm max}, n\\,10^{\\alpha \\log M_* + \\beta})$ and velocity depth $v_{\\rm depth}$ is built, and the galaxy is called central if it has the highest stellar mass among galaxies inside. This gives an observationally measurable definition of 'central' that applies uniformly to SDSS and to both simulations, turning the model-data comparison into a controlled experiment. The second ingredient is the passive fraction $f_{\\rm pass} = \\Sigma w_{\\rm pass}/\\Sigma w_{\\rm all}$, computed in bins of stellar mass and neighbour density, with the neighbour density acting as a halo-mass proxy that breaks the stellar mass-halo mass degeneracy.","core_discovery":"The paper's central discovery is a mismatch between where quenching happens in observations and where the models place it. In the SDSS, the passive fraction of central galaxies rises with stellar mass and bulge mass across environments, with massive field galaxies overwhelmingly passive. In L-GALAXIES, the quenched fraction instead correlates most strongly with halo mass and black hole mass, with radio-mode feedback suppressing cooling for essentially all centrals above $\\log_{10}(M_{\\rm BH}/M_\\odot) \\approx 6$. In SAGE, passive fractions correlate with black hole and bulge mass, but even at high black hole mass only about 60% of central galaxies have their cooling completely suppressed. The authors also show SAGE produces a tighter $M_{\\rm BH}{-}M_{\\rm bulge}$ relation with a slope closer to local observations, yet this improved black hole calibration does not remove the model-data tension. The paper concludes that neither the simple phenomenological feedback of L-GALAXIES nor the more physical coupled cooling-heating scheme of SAGE fully explains the observed quenched central population.","pith_inferences":["A natural extension is to add a bulge-mass-dependent quenching channel, such as morphological quenching or gas starvation, to the models; the prediction would be that passive fractions in massive isolated centrals rise without invoking stronger AGN heating.","The adaptive-aperture central selection could be applied to hydrodynamical simulations or to higher-redshift surveys; if its purity and completeness vary with redshift, part of the model-data tension could be a selection artefact rather than a physics gap.","The paper's 'non-resolved substructures' hypothesis is testable: re-running SAGE with a higher-resolution simulation or with orphan galaxies surviving longer should lower the cold-gas supply in massive centrals and raise their passive fractions.","A direct observational check of the merger-fed cold gas idea is to measure molecular gas in a sample of massive isolated passive centrals; a significant gas reservoir would support the model tension, while gas-poor galaxies would instead point to an over-efficient quenching mechanism."],"forward_implications":["Observed quenching in local central galaxies is tied to baryonic structure, especially stellar mass and bulge mass, more than to environment, so any successful model needs a quenching channel that tracks the bulge.","In L-GALAXIES, radio-mode feedback acts as a sharp switch: nearly every central with $\\log_{10}(M_{\\rm BH}/M_\\odot) \\geq 6$ has its gas cooling suppressed, making black hole mass the effective determinant of quiescence.","In SAGE, the same feedback is less efficient, suppressing cooling in only about 60% of high-black-hole-mass centrals, which explains why SAGE leaves more massive centrals star-forming.","The two models' failure to reproduce massive isolated passive galaxies points to missing physics, likely merger-driven cold gas supply or the treatment of orphan satellites, rather than to the absence of AGN feedback.","SAGE's closer agreement with the observed black hole-bulge mass relation does not by itself solve the quenching problem, showing that a good black hole scaling relation is not sufficient for a good passive fraction."],"supporting_citations":[{"why":"Supplies the L-GALAXIES model: its galaxy formation prescription, radio-mode AGN feedback in Eq. 1, and the Millennium/Millennium-II based catalogues.","marker":"H15"},{"why":"Supplies the SAGE model with its coupled cooling-heating radio mode feedback in Eqs. 4-6 and the treatment of orphan and disrupted galaxies.","marker":"C16"},{"why":"Provides the adaptive aperture mass-rank method used to define central galaxies in both data and models.","marker":"Fossati et al. 2015"},{"why":"Provides the modified SDSS DR7 catalogue with neighbour counts in cylindrical apertures that defines the environment parameter.","marker":"Wilman et al. 2010"},{"why":"Provides the MPA-JHU stellar masses and star formation rates used to classify galaxies as passive or star-forming.","marker":"Brinchmann et al. 2004"},{"why":"Establishes the two-mode quasar/radio AGN feedback framework that both models extend.","marker":"Croton et al. 2006"},{"why":"Supplies one of the observed black hole-bulge mass relations against which the models are compared.","marker":"Saglia et al. 2016"},{"why":"Establishes the observed dependence of passive fraction on bulge mass that motivates the baryonic quenching comparison.","marker":"Bluck et al. 2014b"}],"fun_headline_variants":["Galaxy models get quenching wrong for massive centrals","Black hole feedback fails to fully explain galaxy shutdown","Observed quenched centrals challenge AGN feedback models","SDSS shows models miss quiescent massive galaxies","Where do galaxies stop forming stars? Models disagree"],"cache_read_input_tokens":22528,"weakest_assumption_plain":"The whole comparison rests on the assumption that the adaptive aperture with n = 8, r_max = 2.5 Mpc and v_depth = 2000 km/s selects the same central population in the SDSS as in the two simulations, since the observed sample's purity cannot be checked against real halo membership and small selection differences could masquerade as model-data tension.","fun_headline_variants_meta":{"raw":{"variants":["Galaxy models get quenching wrong for massive centrals","Black hole feedback fails to fully explain galaxy shutdown","Observed quenched centrals challenge AGN feedback models","SDSS shows models miss quiescent massive galaxies","Where do galaxies stop forming stars? Models disagree"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000383,"raw_usage":{"total_tokens":2061,"prompt_tokens":1013,"completion_tokens":1048,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":629,"completion_tokens_details":{"reasoning_tokens":973}},"tokens_in":629,"tokens_out":1048,"duration_ms":8132,"temperature":1.0,"reasoning_tokens":973,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T13:32:41.313629+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A definitive test would be to build a large, complete sample of local massive central galaxies with reliable group halo masses from redshift surveys or X-ray groups and measure passive fraction at fixed stellar mass: if f_pass rises steeply with halo mass at fixed stellar mass, the paper's claim that observed quenching is driven by stellar and bulge mass rather than by halo mass would be overturned. Alternatively, if direct dynamical black hole masses show that the observed passive fraction is actually set by black hole mass at fixed bulge mass, then the AGN-feedback picture the models embody would be supported.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the adaptive aperture mass-rank method used to define central galaxies in both data and models."}],"review_version":1}