{"id":"150118d5-bf1f-4d38-bfb4-bb398c74c7e8","arxiv_id":"2507.12209","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":8,"one_line_summary":"From SHMR-derived star formation histories, cold gas content, and the z=0 mass-metallicity relation, the model infers a halo-mass-dependent outflow recycle fraction of 25-75%.","lead":"This paper presents an empirical model of the galactic baryon cycle and finds that the fraction of outflowing gas recycled back into galaxies rises from roughly 25% in small halos to 75% in massive halos. A generalist might read it to see how gas recycling, a key but invisible process in galaxy formation, can be estimated from stellar masses, cold gas observations, and metallicities without a full simulation.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Eq. 10's zero-metallicity accretion assumption is the load-bearing link: if accreted gas is even mildly enriched, the inferred 25-75% recycle fraction is degenerate, and the model's own z=2 MZR tension corroborates the concern.","rationale":"The paper is transparent about its method, and several parts are independently supported: the cold gas evolution from NeutralUniverseMachine reproduces TNG50 gas fractions, the adopted FIRE-2 mass loading factor agrees better with observations than FIRE-1, and the predicted stellar metallicity and CGM metallicity trends fall within the range of published simulations. The reader's CONDITIONAL verdict is fair. The central claim, however, is not a prediction but a fit to one chosen z=0 MZR through a chain of fixed inputs: SFH from SHMR, cold gas from NeutralUniverseMachine, yield yZ, mass loading factor eta_m, recycle time t_REC, and zero-metallicity IGM accretion. Among these, the zero-metallicity accretion assumption is the most load-bearing because it is the step that makes Eq. 10 independent of the accretion rate and therefore lets the MZR isolate f_REC. The authors themselves flag this as critical in Section 2.4. The z=2 MZR discrepancy in Figure 7 is important corroborating evidence: because f_REC is assumed to depend only on halo mass, the same relation that calibrates z=0 overpredicts z=2 metallicities, which indicates that the f_REC(Mh) derived at z=0 is absorbing effects beyond physical recycling, such as enriched accretion, redshift-dependent recycling, or calibration systematics. The proposed test using FIRE-2 inflow metallicities is decisive because it directly quantifies the dropped term in Eq. 10 in exactly the halo mass and redshift range of the claim. If Z_acc is small, the concern does not land and the conditional interpretation stands; if not, the quoted 25-75% values should be reframed as degenerate with the accretion metallicity. Since the reader already identified the primordial accretion assumption as the weakest point and recommended a conditional acceptance, this stress-test does not change the verdict.","tokens_in":23155,"tokens_out":9503,"duration_ms":119177,"concrete_test":"Measure the mass-weighted metallicity of inward-accreting gas crossing the virial radius for FIRE-2 halos with Mh ~ 10^11-10^12 Msun at z<1, using the same simulation adopted for the fiducial mass loading factor. If the median Z_acc is above ~0.05 Zsun, rerun the Section 3.2 MCMC fit with Z_gas,acc(t, Mh) set to this measured profile instead of zero and compare the best-fit f_REC(Mh). If the best-fit f_REC shifts by more than ~10 percentage points, or if the posterior becomes multimodal with Z_acc, the primordial-accretion assumption is load-bearing and the 25-75% numbers must be reported as conditional on Z_acc = 0.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section 2.4 drops the term Z_gas,acc * Mdot_gas,acc from Eq. 9 on the assumption that IGM accretion is primordial, and the authors state that this is critical for the parameter constraints. Equation 10 then attributes the whole metal budget to recycling, outflow, and star formation, which is what allows the z=0 MZR fit to single out f_REC(Mh). If gas accreted into halos near z<1 is not primordial but consists partly of previously ejected, metal-enriched material that crossed the virial radius and later falls back, this dropped term is not negligible. For halos in the quoted range, the integrated accreted mass is comparable to the stellar mass, so even Z_acc ~ 0.1 Zsun could contribute a significant fraction of the total metal budget yZ * Mstar. The degeneracy is direct: increasing Z_acc lowers the f_REC required to match a given MZR, and vice versa. The claimed 25-75% range is therefore not a measurement of a physical recycle fraction unless Z_acc is independently known to be zero. The model's own independent check in Section 3.4.1 (Fig. 7) is consistent with this worry: with f_REC depending only on halo mass, the same relation that reproduces z=0 overproduces gas-phase metallicity at z=2, suggesting the fitted f_REC is absorbing missing physics or redshift dependence rather than cleanly isolating recycling. This does not invalidate the framework, but it makes the headline numbers conditional on an assumption the authors acknowledge to be fragile.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper presents an empirical 'baryon cycle' model for galaxies at z=0 with halo masses in the range roughly 10^11 to 10^12 M_sun. Star formation histories are constructed by assuming that galaxies always sit on the Girelli et al. (2020) stellar mass-halo mass relation, with halo growth taken from Fakhouri et al. (2010) and cold gas masses taken from the NeutralUniverseMachine of Guo et al. (2023). A gas-regulator chemical evolution model (Eqs. 7-10) is then used, with the z=0 gas-phase mass-metallicity relation as calibration, to fit a halo-mass-dependent recycling fraction f_REC(Mh) and an accretion efficiency epsilon(Mh). With a FIRE-2 mass loading factor the authors obtain f_REC rising from roughly 25% at Mh ~ 10^10.4 M_sun to roughly 75% at Mh ~ 10^12 M_sun, compare epsilon with NIHAO/C16 results, and make predictions for the z=2 MZR, the stellar MZR, the CGM baryon fraction and metallicity, and the metal budget. The abstract advances the 25-75% recycling relation and the CGM values (epsilon_halo ~ 0.7, X ~ 0.8) as the main results.","tokens_in":23480,"tokens_out":6444,"duration_ms":72182,"significance":"If the derived f_REC(Mh) relation is robust, it would provide a rare empirical constraint on wind recycling, a process that is usually accessible only in hydrodynamical simulations, and the preference for the FIRE-2 mass loading factor over FIRE-1 is an interesting, potentially testable conclusion. The paper is transparent in its construction, uses publicly available empirical relations, and its predicted gas fractions agree well with TNG50 (Fig. 3). The high-redshift MZR and CGM predictions are falsifiable in principle, and the authors are candid about the model's limitations. However, the central quantitative claims are currently contingent on a calibration loop, on the zero-metallicity IGM accretion assumption, and on post-hoc choices of epsilon_halo and X, so the significance of the results is conditional on whether these issues can be addressed convincingly.","major_comments":[{"comment":"The z=0 MZR agreement shown in the lower panel of Fig. 4 is a calibration check, not an independent validation. In Section 3.2 the free function f_REC(Mh) is tuned by MCMC to reproduce the Maiolino et al. (2008) MZR, so the perfect match in the figure is guaranteed by construction rather than being evidence that the model has effectively constrained the recycling process. The text should label Fig. 4 explicitly as a fit, and an out-of-sample statistic (e.g., an independent z=0 metallicity calibrator not used in the fit, or a scatter/yield sensitivity test) is needed before claiming that the MZR imprints f_REC.","section":"§3.2, Fig. 4, Eq. (10)"},{"comment":"The zero-metallicity IGM accretion assumption is the load-bearing step, and the paper itself states that the parameter constraints critically rely on it. Dropping the term Z_gas,acc * Mdot_gas,acc attributes the entire metal budget to recycling, outflow, and star formation. Since the time-integrated accreted gas mass is comparable to the stellar mass, even accreted gas with Z_acc ~ 0.1 Zsun would supply a non-negligible fraction of the total metal budget and would change the inferred f_REC(Mh). The authors should add a sensitivity run with nonzero Z_acc or an enriched-accretion model and report the corresponding f_REC(Mh); without that, the quoted 25-75% range is conditional on an assumption that the paper itself identifies as fragile.","section":"§2.4, Eq. (10)"},{"comment":"The model's main out-of-sample test is currently not passed: with f_REC(Mh) fixed by the z=0 fit, the predicted z=2 gas-phase MZR lies systematically above the Maiolino et al. (2008) z=2.2 and Sanders et al. (2021) z=2.3 measurements. The authors acknowledge this tension but defer it to future work. Because the offset is exactly the signature of f_REC(Mh) absorbing missing physics (enriched accretion, redshift-dependent f_REC, or evolving outflow metallicity), the paper should either quantify how much of the offset can be explained by known metallicity calibration systematics or present the headline f_REC(Mh) relation as provisional.","section":"§3.4.1, Fig. 7"},{"comment":"The CGM numbers are not predictions in the same sense as the rest of the paper. The choices epsilon_halo = 0.7 and X = 0.8 are selected in Section 3.4.3 specifically because they bring the model into agreement with the halo baryon fractions from Christensen et al. (2016), Tollet et al. (2019), and Hafen et al. (2019), and with the FIRE-2 CGM metallicity from Pandya et al. (2023). The agreement in Figs. 9 and 10 is therefore by construction. The abstract's statement that the model predicts that on average 70% of baryon accretion enters the halo and 80% of non-recycled outflow escapes should be reframed as consistency requirements imposed by simulation constraints, and the degeneracy between epsilon_halo and X should be explored rather than quoting a single pair.","section":"§3.4.3, Figs. 9-10, Eq. (15)"}],"minor_comments":[{"comment":"The text lists the H2 model parameters as ζ0, ζ1, ζ2, µ, and η, but the equation uses ν for the stellar-mass slope and µ is not otherwise defined; the symbol list should be corrected.","section":"§2.3, Eq. (5)"},{"comment":"There is a typo in 'redshfit' in the sentence defining the halo formation time; it should read 'redshift'.","section":"§2.3"},{"comment":"The acronym MZR is used for both the gas-phase and stellar mass-metallicity relations; after Eq. (14) it would be helpful to introduce MZR_gas and MZR_* explicitly to reduce ambiguity.","section":"§3.4.2, Eq. (14)"},{"comment":"The headline numbers (25-75% recycle fraction, 70% halo accretion, 80% escape) are quoted without uncertainties; showing the MCMC credible intervals for f_REC(Mh) and epsilon(Mh) would put these central claims on a firmer statistical footing.","section":"Abstract, §3.2, §3.4.3"},{"comment":"The claim that the epsilon_halo = 1.0 model is 'significantly higher' than the simulation results is made visually; a simple quantitative comparison (e.g., residuals or a goodness-of-fit value) would make the choice of epsilon_halo = 0.7 more transparent.","section":"Fig. 9"}],"recommendation":"major_revision","confidential_remarks":"The paper is potentially publishable after major revision. The core framework is understandable and the empirical strategy is transparent, but the four major comments concern load-bearing issues: the z=0 MZR agreement is a calibration check, the zero-metallicity accretion assumption is acknowledged by the authors as critical, the z=2 MZR prediction is in tension with observations, and the epsilon_halo/X values are selected to match the very simulation results they are then said to predict. These issues can likely be addressed within the manuscript's scope by reframing the claims and adding sensitivity tests, so I do not recommend rejection. I see no concerns about novelty disclosure or citation patterns."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Zhi, here is my read on arXiv:2507.12209. It is a clear, honest extension of the authors' own gas regulator model. The genuinely new piece is a mass-dependent recycle fraction f_REC(Mh), rising from about 25% at 10^10.4 Msun halos to about 75% at 10^12 Msun, derived by fitting the z=0 gas-phase MZR. Previous semi-analytic treatments used a constant f_REC, so this is a useful parameterization for SAMs, and the comparison with NIHAO, FIRE, and C16 simulations is a good sanity check. I also credit the authors for explicitly stating in Section 2.4 that the constraints rely critically on the assumption of primordial IGM accretion; they do not hide the load-bearing step.\n\nNow the soft spots. The z=0 MZR agreement in Figure 4 is by construction: f_REC is fitted to exactly that relation (Section 3.2). The independent test at z=2 (Section 3.4.1) is systematically high compared to the Maiolino and Sanders MZRs, so the model fails its main out-of-sample check. The authors mention possible systematics, but as it stands the fitted f_REC may be absorbing missing physics or redshift dependence rather than cleanly isolating recycling. The CGM figures in Section 3.4.3 are also parameter choices: epsilon_halo=0.7 and X=0.8 are picked to match simulation baryon fractions and FIRE-2 metallicities, then the abstract calls them predictions. That is overreach; they should be labeled as calibrated.\n\nThe deeper issue is the zero-metallicity accretion assumption. Equation 10 drops Z_acc * Mdot_acc, so the entire metal budget is attributed to star formation, outflow, and recycling. If even mildly enriched CGM gas falls back and is re-accreted, which is physically plausible at z<1, f_REC and epsilon become degenerate with Z_acc. The authors acknowledge this, but the headline 25-75% range is conditional, not a measurement. The paper should quantify how much Z_acc would shift f_REC at fixed MZR. I also note that no code or data are released, making the MCMC specifics hard to audit.\n\nWho should read this? Modelers building baryon cycle parameterizations and observers connecting the MZR to recycle physics. It deserves a serious referee, but the referee should ask for major revision: present f_REC with uncertainties, test the Maiolino calibration choice, engage the z=2 tension quantitatively, and stop calling fitted CGM parameters predictions. I would not cite the headline numbers as a constraint until those are addressed.","headline":"A clearly written empirical gas regulator model that derives a mass-dependent recycle fraction from the z=0 MZR, but the headline numbers are fit-dependent and the zero-metallicity accretion assumption is load-bearing.","tokens_in":24156,"tokens_out":3552,"would_cite":false,"duration_ms":39247,"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":"An empirical model anchored to observed scaling relations finds that galaxies recycle 25–75% of their wind ejecta, with the share rising from dwarf to Milky Way-mass halos.","keywords":["galaxy evolution","baryon cycle","gas recycling","galactic winds","circumgalactic medium","mass-metallicity relation","star formation history","empirical galaxy model"],"falsifier":"Measure the metal abundance of gas flowing into galaxies at $z=0$ across the mass range $10^{10.4}$–$10^{12}M_\\odot$, for example through ultraviolet absorption-line observations of inflowing gas around isolated galaxies; if inflow metallicities above even a few percent of solar are common, the zero-metallicity accretion assumption fails and the fitted $f_{\\rm REC}(M_h)$ curve is not unique.","tokens_in":22851,"feed_emoji":"♻️","tokens_out":10008,"duration_ms":103349,"temperature":0.7,"pith_summary":"This paper tries to establish how much of the gas a galaxy blows out in winds eventually falls back in, and how that fraction depends on dark matter halo mass. The authors build an empirical model in which a galaxy's star formation history is fixed by the observed stellar mass–halo mass relations and its cold gas content is fixed by a calibrated empirical model for atomic and molecular hydrogen, so the only free pieces are the recycling and accretion of gas. Requiring the model to reproduce the observed gas-phase mass–metallicity relation at $z=0$ yields the central result: the recycled fraction rises from about 25% in halos near $10^{10.4}\\,M_\\odot$ to about 75% in halos near $10^{12}\\,M_\\odot$, peaking near $10^{11.8}$. If correct, the same feedback law produces very different long-term fuel retention simply because massive halos recycle more of their outflows.","feed_headline":"Recycled gas climbs from 25% to 75% with halo mass","feed_subtitle":"The share of outflowing gas that falls back in grows with halo mass, reshaping how galaxies keep their fuel.","key_machinery":"The machine is the metal budget equation of the chemical evolution model (Equation 10): with fresh accretion assumed metal-free, the interstellar medium metal content is set only by star formation, outflow, and recycled outflow, so the observed mass–metallicity relation fixes the recycle fraction $f_{\\rm REC}(M_h)$. The companion gas budget (Equation 7) then fixes the IGM accretion fraction $\\epsilon(M_h)$. These two equations are fed by star formation histories from abundance-matching stellar mass–halo mass relations plus mean halo growth histories, cold gas from the NeutralUniverseMachine model, the FIRE-2 mass loading factor, and a simulation-calibrated recycle timescale $t_{\\rm REC}(M_h)$.","core_discovery":"The central claim is that gas recycling is mass-dependent and is the missing ingredient that makes the local mass–metallicity relation come out right. In the fiducial model the recycle fraction $f_{\\rm REC}(M_h)$ grows from roughly 25% at $M_h \\sim 10^{10.4}M_\\odot$ to roughly 75% at $M_h \\sim 10^{12}M_\\odot$, with a plateau or downturn above that mass. The same fit favors the FIRE-2 mass loading factor, the ratio of outflow rate to star formation rate, over the earlier FIRE-1 value. Extrapolations then yield a halo-level baryon accretion efficiency near 70% and an escape fraction near 80% for non-recycled outflow, with the prediction that circumgalactic medium metallicity rises then flattens at large halo mass.","pith_inferences":["If even mildly enriched IGM accretion is common at low redshift, the low-mass end of the inferred $f_{\\rm REC}$ should be read as an upper limit, because part of the metal budget would enter with the fresh gas rather than with recycled outflows.","The model's $z=2$ mass–metallicity overshoot, which the paper attributes to either missing physics or calibration systematics, could be turned into a test: re-fit $f_{\\rm REC}$ separately to the $z=2$ MZR. A significantly different curve would indicate redshift evolution of recycling, while no change would point to metallicity-calibration offsets.","A natural extension is to let $f_{\\rm REC}$ and $t_{\\rm REC}$ depend on redshift as well as halo mass; the current data only constrain their mass dependence, and the paper's own CGM/IGM discussion notes that the exchange rates between those reservoirs remain unconstrained."],"forward_implications":["Low-mass halos eject most of their wind gas permanently, while Milky Way-mass halos recycle about three quarters of it, so the gas retention efficiency of a galaxy is largely set by halo mass.","The local gas-phase mass–metallicity relation can be reproduced by mass-dependent recycling with the FIRE-2 mass loading factor, without needing a strongly mass-dependent outflow efficiency.","The model's quantitative predictions—roughly 70% of the universal baryon supply reaching the halo and 80% of non-recycled outflow escaping into the IGM—give semi-analytic models concrete functions $f_{\\rm REC}(M_h)$ and $\\epsilon(M_h)$ to adopt.","At $M_h(z=0)=10^{12}M_\\odot$, metals end up more than half in stars and only a few percent in gas still recycling, while the ISM keeps about 30% at all masses."],"supporting_citations":[{"why":"Establishes the parent empirical framework of SHMR-based star formation histories plus chemical evolution whose baryon-cycle equations this paper extends.","marker":"Chen et al. (2023)"},{"why":"Provides the NeutralUniverseMachine empirical relations for HI and H2 cold gas content that fix the gas budget of the model galaxies.","marker":"Guo et al. (2023)"},{"why":"Supplies the redshift-dependent stellar mass–halo mass relations whose time derivative yields the model star formation histories.","marker":"Girelli et al. (2020)"},{"why":"Provides the mean halo mass accretion rate integrated to build the halo growth histories of the model galaxies.","marker":"Fakhouri et al. (2010)"},{"why":"Supplies the $z\\sim0$ gas-phase mass–metallicity relation used as the calibration target for $f_{\\rm REC}(M_h)$.","marker":"Maiolino et al. (2008)"},{"why":"Furnishes the fiducial FIRE-2 mass loading factor, which the paper finds agrees with observational constraints and is preferred over FIRE-1.","marker":"Pandya et al. (2021)"},{"why":"Contributes NIHAO simulation recycle fraction and timescale measurements used to set $t_{\\rm REC}(M_h)$ and to compare $f_{\\rm REC}$.","marker":"Tollet et al. (2019)"},{"why":"Provides simulation measurements of recycle fraction and halo baryon content used to validate the predicted $f_{\\rm REC}$ and the halo accretion fraction.","marker":"Christensen et al. (2016)"}],"fun_headline_variants":["Gas recycling jumps from 25% to 75% with halo mass","Heavier halos recycle more of their outflowing gas","Mass-dependent recycling reshapes galaxy fuel cycles","Recycled outflow fraction doubles as halos grow"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The result stands or falls on the assumption, stated in Section 2.4, that all gas a galaxy accretes from the intergalactic medium is metal-free; if even mildly enriched gas is coming in, the metal budget no longer isolates recycling, and the inferred recycle and accretion fractions become degenerate with the metallicity of the inflow.","fun_headline_variants_meta":{"raw":{"variants":["Gas recycling jumps from 25% to 75% with halo mass","Heavier halos recycle more of their outflowing gas","Mass-dependent recycling reshapes galaxy fuel cycles","Recycled outflow fraction doubles as halos grow"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000215,"raw_usage":{"total_tokens":1473,"prompt_tokens":1032,"completion_tokens":441,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":648,"completion_tokens_details":{"reasoning_tokens":375}},"tokens_in":648,"tokens_out":441,"duration_ms":5399,"temperature":1.0,"reasoning_tokens":375,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T16:52:18.754241+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the metal abundance of gas flowing into galaxies at $z=0$ across the mass range $10^{10.4}$–$10^{12}M_\\odot$, for example through ultraviolet absorption-line observations of inflowing gas around isolated galaxies; if inflow metallicities above even a few percent of solar are common, the zero-metallicity accretion assumption fails and the fitted $f_{\\rm REC}(M_h)$ curve is not unique.","supporting_citations":[],"review_version":1}