{"id":"97e41329-4589-4f8f-9254-82a3d946ab06","arxiv_id":"2505.22712","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"A simplified burst-cycle model shows the constant high-redshift MZR in FIRE-2 arises from a balance between rising inflow metallicity and falling metal production efficiency.","lead":"The authors show that in FIRE-2 simulations, the high-redshift galaxy mass-metallicity relation stays flat because two opposing processes cancel: recycled gas makes inflows more metal-rich over time, while star formation becomes less efficient at producing metals per unit of inflow. Their \"Reduced Burst Model\" explains the weak evolution seen by JWST and predicts that survey selection can hide a relation between metallicity and star formation rate.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The cancellation in Eq. (6) rests on unquantified residual ISM gas and metals at burst-cycle start; if these are non-negligible at high mass or low redshift, the claimed explanation is not yet established.","rationale":"The reader's conditional verdict identifies the burst-reset assumption as the weakest point, and my analysis agrees: the validity of Eq. (6) hinges on M_gas,i and M_Z,i being negligible at cycle start. The paper gives indirect support through agreement between the full and reduced models, the short evacuation timescale in Appendix C, and insensitivity to the minimum-gas threshold in Section 4.1. However, it does not directly report the residual gas and metal fractions at burst-cycle start, and it explicitly concedes that the model deviates at the high-mass end where gas retention is expected. That makes the reset assumption both load-bearing and under-documented. The concrete test above would settle whether the concern actually lands. If the residual fractions are small in the bins that define the MZR, the central cancellation claim would be robust; if not, the interpretation would need revision. I do not see grounds to reject the paper on the current evidence, and the existing conditional verdict already reflects the need for further testing, so the verdict should remain unchanged.","tokens_in":21994,"tokens_out":5877,"duration_ms":75577,"concrete_test":"From the particle tracking used for Section 4.1, compute at each burst-cycle start the residual fractions f_gas = M_gas,i / ∫ M_in dt and f_metal = M_Z,i / ∫ (M_Z,in + M_Z,R) dt, with integrals taken over the full burst cycle. Bin these fractions by the same stellar mass and redshift bins as Figure 5. If the median f_gas or f_metal exceeds ~10% in any bin, especially at M_* ≳ 10^9 M_⊙ or at the lowest redshifts in the sample, then Eq. (6)'s neglect of initial terms is not justified in that bin, and the cancellation explanation should be re-derived with the initial terms included.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The derivation of Eq. (6) from Eq. (2) requires that at the start of each burst cycle the residual ISM metal and gas masses, M_Z,i and M_gas,i in Eq. (2), are negligible (Section 2.2, Eq. 3). This is the load-bearing step: if significant gas survives between bursts, then Z_gas ≈ Z_in^avg + ε_Z does not follow, and the reported cancellation between Z_in^avg and ε_Z is not the actual driver of the constant MZR. The paper supports the reset assumption only indirectly: Figure 4 shows that the full and reduced models agree, and Section 4.1 notes insensitivity to the minimum-gas-mass threshold. But the full model starts from the same cycle boundaries, so agreement between the two models does not by itself prove that the initial terms are negligible; it only shows that the two approximations happen to coincide for the chosen boundaries. The paper's own Section 5.3 states that deviations at M_* ≳ 10^9 M_⊙ may be due to high-mass galaxies retaining gas between burst cycles, and that the model is expected to break down at lower redshift. Since the cancellation in Eq. (6) is presented as the explanation for the z = 5–12 constancy over the full stellar mass range, the residual ISM fraction at cycle start is the key unquantified quantity. If it is non-negligible in the high-mass or low-redshift bins, the cancellation would be at least partly an artifact of the burst-cycle partitioning rather than a robust physical explanation.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper analyzes a high-redshift (z = 5–12) suite of FIRE-2 cosmological zoom-in simulations to explain the weak evolution of the gas-phase mass–metallicity relation (MZR). The authors divide galaxy histories into 'burst cycles' bounded by feedback-driven outflows that reset the interstellar medium, and from the full gas-regulator model they derive a 'Reduced Burst Model' in which the gas-phase metallicity is approximately Z_in^avg + ε_Z, the sum of the time-averaged inflow metallicity and the stellar metal production efficiency per unit inflow (Eq. 6). They show that this reduced model reproduces the simulated MZR and the full gas-regulator predictions, and that the weak redshift evolution arises from a cancellation: as redshift decreases, Z_in^avg increases while ε_Z decreases. They additionally study a secondary dependence of metallicity on Hα-derived star formation rate, find an FMR-like signal in the mass-complete sample, and show that this signal is weakened when only rest-UV-selected (JWST-like) galaxies are considered.","tokens_in":22346,"tokens_out":8991,"duration_ms":115155,"significance":"If the central claim holds, the paper provides a concrete, physically motivated alternative to gas-fraction-based explanations of the high-redshift MZR, and it identifies inflow metallicity and star-formation efficiency as the key baryon-cycle drivers. The analysis is grounded in explicit particle and galaxy tracking, and the reduced model is checked against both the full gas-regulator expression and direct simulation measurements, with no free parameters tuned to match the MZR. The FMR selection-effect result is also timely for interpreting JWST metallicity samples. However, the derivation of Eq. (6) leans on the burst-reset assumption, and the paper does not directly quantify the residual ISM at cycle starts; the authors themselves note that the model is expected to break down at high stellar mass and at lower redshift. Those gaps make the scope of the central claim larger than the current evidence directly supports.","major_comments":[{"comment":"The reduction from Eq. (2) to Eq. (6) assumes that M_gas,i and M_Z,i are negligible at the start of each burst cycle, but the burst-cycle definition in Section 4.1 allows a new cycle to begin at a local gas-mass minimum that is as high as 50% of the previous cycle's peak gas mass. Consequently, the residual ISM can be comparable to the integrated inflow terms for a substantial fraction of cycles, especially at high stellar mass. Because the full gas-regulator model (Eq. 2) and the reduced model (Eq. 6) are evaluated over the same cycle boundaries, agreement in Fig. 4 does not by itself establish that the initial terms are negligible. Please report the binned distributions of M_gas,i / ∫ M_in dt and M_Z,i / ∫ (M_Z,in + M_Z,R) dt as functions of stellar mass and redshift, and show directly that the reduced-model predictions and the Z_in^avg–ε_Z cancellation in Fig. 5 are robust when cycles with large residual ISM are removed.","section":"Sections 3 and 4.1"},{"comment":"The paper presents the cancellation as holding the MZR approximately constant for z = 5–12 over the full stellar mass range, yet Section 5.3 states that deviations at M_star ≳ 10^9 M_sun may be due to gas retained between burst cycles, which is precisely the regime where the reset assumption underlying Eq. (6) is expected to fail. Please quantify the offsets between the reduced-model predictions and the FIRE-2 MZR in each mass bin (e.g., median offset and scatter in the highest-mass bin) and explicitly test whether the weak-evolution and cancellation conclusions hold when the analysis is restricted to M_star ≲ 10^9 M_sun. If the conclusions do not extend to the high-mass end, the scope of the central claim should be revised.","section":"Section 5.3 and Fig. 5"},{"comment":"The interpretation that the decrease in ε_Z is driven by the decrease in the star formation efficiency SFE = ∫ SFR dt / ∫ M_in dt is not directly demonstrated, because ε_Z is defined using the total stellar metal return rate M_Z,R, which includes returns from stellar populations formed in earlier burst cycles as well as from the current cycle. Please show that returns from stars formed within the current cycle dominate M_Z,R (or separate the two contributions) and verify that the ε_Z–SFE relation holds when only current-cycle returns are used. Without this check, the physical explanation for the ε_Z evolution remains suggestive rather than established.","section":"Section 5.2 and Appendix A"}],"minor_comments":[{"comment":"The phrase 'cosmic baron cycle' should read 'cosmic baryon cycle'.","section":"Section 1"},{"comment":"In the right panel of Fig. 4, the expression 'Zavg_in + Z' should read 'Zavg_in + ε_Z' for consistency with Eq. (6).","section":"Figure 4 caption"},{"comment":"There is a duplicated word: 'the near-perfect cancellation between between Z_avg_in and ε_Z' should have only one 'between'.","section":"Section 5.3"},{"comment":"The middle and right panels label the metal production efficiency as 'Z' rather than 'ε_Z'; please correct the notation to match the text.","section":"Figure 5 caption"}],"recommendation":"major_revision","confidential_remarks":"The paper is honest about its limitations and the central decomposition is clearly presented, but the burst-reset assumption is load-bearing and needs direct, quantitative support before the cancellation claim can be accepted over the full stated mass and redshift range. The requested tests are well within the scope of the existing simulation data."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The central claim here is better than the abstract makes it look. Marszewski and collaborators show that the weak evolution of the FIRE-2 MZR at z=5-12 is not a gas-fraction story but a story about two opposing trends within burst cycles: the average metallicity of inflows rises toward lower redshift (more recycled enriched gas), while the metal production efficiency falls (less star formation per inflowing mass). The cancellation is real and is demonstrated directly in Figure 5, not reverse-engineered. The Reduced Burst Model has no free parameters tuned to match the MZR; equation 6 is just the simulation's metal and gas accounting rearranged under a stated approximation. That is a legitimate and useful explanatory move, and it is clearly contrasted with the earlier gas-fraction arguments in Ma et al., Torrey et al., and Langan et al.\n\nWhat is genuinely new is the burst-cycle decomposition and the identification of cancellation as the driver. The paper also does an honest job of testing the reduced model against both the full gas-regulator expression and direct simulation measurements. The FMR section is a useful byproduct rather than the main event, but the selection-effect point is well made.\n\nThe soft spots are real but mostly in proportion. The load-bearing assumption is that at the start of each burst cycle the residual ISM gas and metal masses are negligible. The paper does not quantify the residual fraction directly; it argues from the evacuation timescale, the insensitivity to the minimum gas threshold, and the agreement between full and reduced models. The stress-test note worries that the full model uses the same cycle boundaries, so agreement between the two models does not by itself prove the initial terms are negligible. That is partly correct, but it cuts the other way too: the full model actually includes those initial terms, so sustained agreement between full and reduced predictions across mass and redshift does constrain them, at least for the median population. The paper's own Section 5.3 admits deviations at M* > 1e9 Msun and expects breakdown at lower redshift. So the claim is established for the median high-z population, not as a universal law.\n\nThe physical origin of the two trends is suggested, not derived. The recycling story for Z_in and the dense-ISM story for epsilon_Z are plausible but are not tested against alternative drivers. That is a minor weakness for a simulation analysis, not a fatal one. The absence of a public data/code release is more annoying; the analysis uses particle tracking that would be hard to reproduce independently. FIRE-2 is a community code, but the specific suite and tracking choices are not documented well enough to reconstruct without the authors' help.\n\nWho is this for? Anyone working on high-z metallicity or bursty feedback will want to read it. It deserves a serious referee; the central mechanism is testable in other simulations and with future JWST samples. I would send it to review with confidence.","headline":"A parameter-free burst-cycle decomposition that convincingly explains the flat high-redshift MZR in FIRE-2 via cancellation between inflow metallicity and metal production efficiency; the reset assumption is softer than the paper admits but not fatal.","tokens_in":22861,"tokens_out":3041,"would_cite":true,"duration_ms":34896,"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":"A simplified burst-cycle model explains the nearly constant mass-metallicity relation at z = 5–12 by a cancellation between inflow metallicity and metal production efficiency.","keywords":["mass-metallicity relation","bursty star formation","galactic outflows","wind recycling","FIRE-2 simulations","gas-phase metallicity","fundamental metallicity relation","high-redshift galaxies"],"falsifier":"Compute, for each burst cycle in a FIRE-2-like simulation, the ratio of gas mass present at cycle start to the integrated inflow over that cycle; if the median ratio is not small (say $\\lesssim 0.1$) across the z = 5–12 sample, the ISM-reset premise fails and the cancellation explanation cannot be the driver. Observationally, a mass-complete JWST sample showing more than about 0.1 dex of MZR normalization evolution from z = 12 to z = 5 would also contradict the prediction.","tokens_in":21762,"feed_emoji":"🌌","tokens_out":10420,"duration_ms":103925,"temperature":0.7,"pith_summary":"This paper asks why galaxies at z = 5–12 show a mass-metallicity relation that barely changes with cosmic time, and answers with a burst-cycle mechanism rather than gas fractions. Using the FIRE-2 simulations, the authors argue that strong stellar feedback following bursty star formation evacuates and resets each galaxy's interstellar medium, splitting its history into discrete burst cycles. Within a cycle, gas-phase metallicity is set by two quantities only: the average metallicity of inflowing gas and the metal mass returned by stars per unit inflowing gas mass. They show that, at fixed stellar mass, the first rises with decreasing redshift while the second falls, and the two trends cancel almost exactly, holding the MZR flat from z = 12 to z = 5. The paper also finds a secondary anti-correlation between metallicity and H-alpha-derived star formation rate that weakens under rest-UV selection.","feed_headline":"Galaxy burst cycles freeze the high-redshift mass-metallicity relation","feed_subtitle":"Rising inflow metallicity and falling star-formation efficiency cancel, keeping metallicity flat from z=12 to z=5.","key_machinery":"The load-bearing machinery is the burst-cycle decomposition of a galaxy's history, in which intense feedback evacuates the ISM on a 10–30 Myr timescale within 70–200 Myr cycles, followed by the Reduced Burst Model identity $Z_{\\rm gas} \\approx Z_{\\rm in}^{\\rm avg} + \\varepsilon_Z$. The identity is obtained from the full gas-regulator expression by starting integration just after an outflow-driven reset, making initial gas and metal masses negligible, and dropping outflow and astration terms that are small during most of a cycle. It isolates the two quantities whose opposite redshift trends produce the constant MZR, and the measured scaling of its numerator and denominator terms ($\\propto M_\\star^{1.12}$ versus $\\propto M_\\star^{0.75}$) yields the MZR slope.","core_discovery":"The paper's central claim is that the nearly flat high-redshift MZR in FIRE-2 is produced by a cancellation inside the Reduced Burst Model: within a burst cycle, $Z_{\\rm gas} \\approx Z_{\\rm in}^{\\rm avg} + \\varepsilon_Z$, where $Z_{\\rm in}^{\\rm avg}$ is the cycle-averaged metallicity of inflowing gas and $\\varepsilon_Z$ is the metal mass returned by stars per unit inflowing gas mass. At fixed stellar mass, as redshift falls from 12 to 5, $Z_{\\rm in}^{\\rm avg}$ rises because more previously ejected, enriched gas is recycled back into the galaxy, while $\\varepsilon_Z$ falls because less star formation occurs per unit inflow; the two trends nearly cancel, so the summed metallicity stays constant. The same model reproduces the MZR slope through the power-law scalings of the metal-inflow and metal-return integrals with stellar mass. The paper additionally claims that gas-phase metallicity at fixed stellar mass anticorrelates with H$\\alpha$-derived star formation rate, an FMR-like signal that is weakened under rest-UV selection and absent for UV-continuum SFR.","pith_inferences":["Editorial inference: if the cancellation is generic, the constancy of the high-redshift MZR is not an equilibrium but a transient balance; simulations with weaker feedback and more retained gas between bursts should show growing MZR normalization by z = 5.","Editorial inference: the burst-cycle picture predicts that abundance-ratio diagnostics such as alpha-to-iron ratios should vary with phase within a cycle (inflow, starburst, outflow), so phase-resolved JWST spectra could test the model directly.","Editorial inference: the SFR-indicator dependence of the FMR-like signal implies that apparent evolution of the FMR at high redshift may be partly a selection effect and partly a timescale effect, and surveys using H-alpha versus UV-continuum SFR estimators may find systematically different metallicity offsets."],"forward_implications":["The weak evolution of the high-redshift MZR is driven by baryon-cycle processes, not by saturated or weakly evolving gas fractions, so closed-box and leaky-box explanations are incomplete.","As redshift decreases, wind recycling enriches the gas flowing into galaxies while the star formation efficiency per inflow declines; these two trends are measurable and continue down to z = 5.","The slope of the MZR follows from the model's scaling relations: integrated metal inflow and metal return scale roughly as $M_\\star^{1.12}$ while integrated gas inflow scales as $M_\\star^{0.75}$, giving the measured slope of about 0.37.","The simulations predict an FMR-like inverse relation between gas-phase metallicity and H-alpha-derived star formation rate at fixed stellar mass, but the signal weakens when only rest-UV-selected galaxies are considered, which has direct consequences for JWST samples.","The Reduced Burst Model applies to any galaxy population whose histories separate into burst cycles, even if that population's MZR evolves; the framework can therefore be transported to other simulations and regimes."],"supporting_citations":[{"why":"Establishes the FIRE-2 MZR and its weak evolution from z = 5–12 that this paper sets out to explain.","marker":"Marszewski et al. (2024)"},{"why":"Provides the high-redshift FIRE-2 zoom-in simulation suite analyzed here.","marker":"Ma et al. (2018a,b, 2019)"},{"why":"Defines the FIRE-2 stellar feedback and metal-yield models used to compute stellar metal return rates and outflows.","marker":"Hopkins et al. (2018)"},{"why":"Supplies the gas-regulator model baseline from which the Reduced Burst Model is derived by simplification.","marker":"Lilly et al. (2013)"},{"why":"Shows in FIREbox at z = 0–3 that gas fractions do not drive MZR evolution, motivating the direct baryon-cycle decomposition used here.","marker":"Bassini et al. (2024)"},{"why":"Characterizes wind recycling of enriched gas that the paper invokes to explain rising inflow metallicities at lower redshift.","marker":"Anglés-Alcázar et al. (2017)"},{"why":"Documents rest-UV selection effects that the paper uses to interpret the weakened FMR-like signal in observable samples.","marker":"Sun et al. (2023a)"},{"why":"Provides JWST observations of a weakly evolving MZR at z = 4–10 that the FIRE-2 prediction is compared against.","marker":"Nakajima et al. (2023)"}],"fun_headline_variants":["Burst cycles cancel two trends, pinning metallicity flat from z=12 to 5","Recycled enriched gas and weaker stars freeze the high-z MZR","Inflow metallicity up, star formation down: MZR stays flat at high z","Galaxy burst cycles explain the non-evolving high-redshift MZR"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The entire cancellation explanation rests on the assumption that feedback-driven outflows empty the galaxy's ISM at the start of each burst cycle, so that gas and metals retained from earlier cycles are negligible; the authors themselves note this assumption begins to fail for the most massive galaxies and at lower redshift.","fun_headline_variants_meta":{"raw":{"variants":["Burst cycles cancel two trends, pinning metallicity flat from z=12 to 5","Recycled enriched gas and weaker stars freeze the high-z MZR","Inflow metallicity up, star formation down: MZR stays flat at high z","Galaxy burst cycles explain the non-evolving high-redshift MZR"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000239,"raw_usage":{"total_tokens":1598,"prompt_tokens":1111,"completion_tokens":487,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":727,"completion_tokens_details":{"reasoning_tokens":398}},"tokens_in":727,"tokens_out":487,"duration_ms":6015,"temperature":1.0,"reasoning_tokens":398,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T13:02:38.626617+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute, for each burst cycle in a FIRE-2-like simulation, the ratio of gas mass present at cycle start to the integrated inflow over that cycle; if the median ratio is not small (say $\\lesssim 0.1$) across the z = 5–12 sample, the ISM-reset premise fails and the cancellation explanation cannot be the driver. Observationally, a mass-complete JWST sample showing more than about 0.1 dex of MZR normalization evolution from z = 12 to z = 5 would also contradict the prediction.","supporting_citations":[],"review_version":1}