{"id":"71e057c5-5dd7-4705-9537-f7891460f5a3","arxiv_id":"2511.00229","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":6,"one_line_summary":"Cold-gas survival in the circumgalactic medium is set by t_cool/t_mix; small-scale cold gas requires dense, quiescent regions, not turbulent cascade patches.","lead":"This paper simulates turbulent gas patches in the halos around galaxies to see when cold, dense clumps survive. It finds that the ratio of cooling time to mixing time, not box size, controls survival; cold gas lives in dense, quiet regions rather than uniformly turbulent ones.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Tmix scaling claim is suspect: FidMHDLR vs 8TmixMHDLR have similar t_cool/t_mix (2.7 vs 2.9) but cold fractions 0.42 vs 0.18; the control parameter may be an input, not a measured steady-state quantity.","rationale":"I evaluated the central claim that t_cool/t_mix, not box size or turbulent heating, controls cold gas survival. The most load-bearing risk is that the evidence for this scaling is weaker than stated. The Tmix runs are intended to hold t_cool/t_mix fixed, but column 5 reports initial input values, and the actual realized t_mix—based on measured velocity dispersions—deviates, especially for 0.125Tmix. More importantly, the two same-resolution runs with nearly identical reported t_cool/t_mix (FidMHDLR 2.7 vs 8TmixMHDLR 2.9) end up with cold mass fractions 0.42 vs 0.18. This factor-2.3 discrepancy suggests that a global integral-scale t_cool/t_mix does not fully determine cold gas survival; cloud-scale mixing times or other box-size-dependent effects intervene. The thermostat is a real secondary concern, as the authors themselves show Qvw changes the cold fraction from 0.20 to 0.33, but it affects the absolute normalization more than the scaling. The concrete test—computing the measured t_cool/t_mix from the data and checking whether the correlation collapses—would settle whether the scaling is genuine or an artifact of input parameters. Until then, conditional acceptance is appropriate, with the additional test requested.","tokens_in":23305,"tokens_out":16375,"duration_ms":144512,"concrete_test":"Using the simulation snapshots, compute the time-averaged t_mix at the integral scale from the measured velocity dispersion and eddy scale, and the corresponding t_cool/t_mix, for each run in Table 1 (especially the LR runs). Plot cold gas mass fraction vs measured t_cool/t_mix. If FidMHDLR and 8TmixMHDLR have measured ratios within ~20% yet cold fractions 0.42 and 0.18, the claimed collapse onto t_cool/t_mix fails. If the measured ratios differ by more than a factor of 1.5, the reported correlation is an artifact of using initial input values rather than realized physics.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Table 1 reveals that the Tmix series, intended to hold t_cool/t_mix fixed, does not. In hydro, 8Tmix (t_cool/t_mix=3.1) and FidLR (3.0) have comparable cold fractions (0.14 vs 0.20), but the MHD counterpart fails: FidMHDLR (2.7) retains 42% cold mass while 8TmixMHDLR (2.9) retains only 18%. The 0.125Tmix runs also deviate (t_cool/t_mix=1.9–2.2, cold fractions 0.30–0.40), showing that the 'matched' runs actually span a factor ~1.6 in the control parameter. The column-5 t_cool/t_mix values are initial inputs (per note), not measured from the simulations. Because t_mix is set by the target driving velocity, and actual σ_v deviates (e.g., 0.125TmixHydroLR reaches 1.4 km/s vs ~2.5 km/s target), the reported scaling may reflect design intent rather than the realized physics. Moreover, even at the same resolution (LR), similar integral-scale t_cool/t_mix fails to predict cold gas fraction in MHD by more than a factor of two, suggesting that cloud-scale mixing times—which depend on cloud size, density contrast, and the cascade below the integral scale—are not captured by a single global t_cool/t_mix. This directly weakens the central 'key scaling relation' and the inference that t_cool/t_mix, not box size, controls cold gas survival.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents a suite of 0.125–8 kpc magnetohydrodynamic and hydrodynamic turbulence simulations of CGM-like gas, with parameters motivated by quasar absorption-line observations. The simulations resolve the minimum cooling length by at least 30 cells for gas above ~10^4 K. The central claim is that the ratio t_cool/t_mix, rather than box size or turbulent heating rate, controls cold gas survival: low-density runs (n=3e-4 cm^-3) cannot sustain cold gas below 10^4 K, while higher-density Fiducial runs (n=3e-3 cm^-3) reach a multiphase steady state with cold mass fractions up to ~50% in MHD and ~20% in hydro. By varying box size under two scalings—fixed energy injection rate (Dedt) and fixed mixing time (Tmix)—the authors argue that matched t_cool/t_mix yields comparable cold fractions across scales, implying that small-scale cold gas arises in dense, quiescent CGM regions rather than in uniformly turbulent cascades. The paper also examines magnetic field geometry, compressive vs. solenoidal forcing, volume- vs. mass-weighted heating, and ICM-like density contrasts.","tokens_in":23810,"tokens_out":4364,"duration_ms":42104,"significance":"If the central scaling claim holds, the paper provides a useful organizing principle for connecting pc-scale cloud simulations, kpc-scale turbulent box simulations, and cosmological CGM simulations. The simulations are well-resolved in the sense that c_s t_cool is resolved by at least 30 cells, and the use of an open-source GPU-capable code (AthenaK) is a strength. The comparisons to observed non-thermal line broadening and the predicted flat density power spectra in multiphase gas are interesting and potentially relevant for FRB scattering and absorption-line diagnostics. However, the central claim that t_cool/t_mix is a robust control parameter is currently under-supported by the presented data, because the reported t_cool/t_mix values are initial inputs rather than measured simulation quantities, and because the matched runs in Table 1 do not actually show comparable cold fractions in the MHD case. The compensatory heating prescription also directly shapes the quoted steady-state cold fractions and needs more scrutiny. The paper's significance would be substantially strengthened by reporting measured mixing times and by quantifying the sensitivity of the cold fraction to the hea","major_comments":[{"comment":"The key scaling claim—simulations with similar t_cool/t_mix exhibit comparable cold gas mass fractions—is not convincingly supported by Table 1. Column 5 lists t_cool/t_mix as an 'initial ratio' (see table note), not a measured steady-state quantity. The actual velocity dispersion deviates from the assumed target; for example, 0.125TmixHydroLR has σ_v=1.4 km/s while the stated target in §2.5.2 for a 0.125 kpc box with Tmix scaling would be ~2.5 km/s. More importantly, the MHD matched runs fail: FidMHDLR has t_cool/t_mix=2.7 and M_cold/Mtot=0.42, while 8TmixMHDLR has t_cool/t_mix=2.9 and M_cold/Mtot=0.18. These are similar control parameters but cold fractions differ by more than a factor of two. The 0.125Tmix runs also have t_cool/t_mix=1.9–2.2 and cold fractions 0.30–0.40, spanning a factor ~1.6 in the control parameter. Since each configuration is a single realization without ensemble","section":"§4, Table 1, Fig. 8"},{"comment":"The compensatory heating term Q(t) is an artificial thermostat that offsets all net cooling and turbulent work to maintain global thermal balance. With the default mass-weighted heating (α_heat=1), Q is preferentially deposited in denser gas, and as cold gas condenses the net cooling rate rises, which increases Q and pumps heat back into the hot phase. This feedback loop directly regulates the steady-state cold gas mass fraction. The paper's headline result of up to 50% cold mass in the Fiducial MHD run therefore rests on this prescription. The sensitivity run with volume-weighted heating (QvwHydroLR, α_heat=0) already shows a material difference: M_cold/Mtot=0.33 versus 0.20 for FidHydroLR (Table 1 and Fig. 13). The authors should quantify how much of the steady-state cold fraction is controlled by the thermostat rather than by the t_cool/t_mix balance, justify the α_heat=1 choice physi","section":"§2.2, Eq. (2b); §5.2, Fig. 13"}],"minor_comments":[{"comment":"The text says 'second-order Runge–Kutta (RK3)' but RK3 is third-order. Please correct the order or the acronym.","section":"§2.3"},{"comment":"The caption mentions 'green solid lines in the upper panels' representing initial density perturbations, but the figure as rendered appears to show only colored time-series lines. Please check the color/legend and update the caption or figure.","section":"Fig. 5 caption"},{"comment":"The column header says t_cool/t_mix, but the note says this is the initial ratio. To avoid confusion, either rename the column to 'initial t_cool/t_mix' or provide a second column with measured steady-state values.","section":"Table 1"},{"comment":"The description of the two box-size scalings is clear, but Eq. (5a) fixes t_mix as an input while the realized t_mix depends on the actual velocity dispersion. Please clarify in the text that the Tmix labels denote target conditions, and that the realized values must be measured.","section":"§2.5.2"},{"comment":"The two Afruni et al. 2023a,b entries appear to have identical author lists and page numbers (A680, A112). If they are the same paper, merge them; if different, correct the bibliographic details.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The paper tackles an important and timely problem, and the simulation campaign is substantial. However, the central scaling claim is currently under-supported by the data in Table 1: the matched-t_mix runs do not actually match in realized t_cool/t_mix, and the MHD comparison shows a factor >2 difference in cold fraction despite similar input ratios. The heating-prescription sensitivity is also a concern for the quantitative conclusions. Both issues are addressable by reanalyzing existing output (measuring t_mix from the velocity field, computing multi-seed statistics) and by reframing the results as conditional on the heating model. I therefore recommend major revision rather than rejection."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Take a look at this one, but read Table 1 before you trust the abstract. The paper does a real service by showing that sub-sampling a uniformly-driven turbulent CGM box to smaller scales (constant energy injection) artificially shortens t_mix and kills cold gas, while runs that keep t_mix fixed across box sizes do not. That is a genuinely useful warning for anyone who post-processes cosmological simulations by carving out kpc patches. The resolution is a genuine step up: they resolve c_s t_cool by 30+ cells across the relevant temperature range, and the hydro vs MHD contrast—magnetic fields suppress mixing and keep cold gas filamentary—is clean and well-illustrated. The VSF comparison to absorption-line data (Chen et al. 2023) is a nice touch.\n\nThe soft spots are concentrated where the paper makes its strongest claim. The key scaling relation—that simulations with similar t_cool/t_mix have similar cold mass fractions—does not hold cleanly in Table 1. The t_cool/t_mix values in column 5 are initial inputs, not measured steady-state quantities, and the realized velocity dispersions deviate from the targets, especially in the 0.125 kpc boxes. More alarming, FidMHDLR (t_cool/t_mix=2.7) has cold fraction 0.42 while 8TmixMHDLR (t_cool/t_mix=2.9) has 0.18. That is a factor of two-plus scatter at near-identical control parameter. The Tmix series is partly a consistency check by construction, and the MHD counterpart fails that check. The authors need to measure the actual t_cool/t_mix in steady state and show the scaling with that quantity, not with the design input.\n\nThe compensatory heating term (eq. 2b) is another soft spot. With mass-weighted heating, the thermostat can pump heat into the hot phase exactly when cold gas condenses, and the one volume-weighted run changes the cold fraction by about 60%. That is a load-bearing uncertainty for the 'up to 50%' number, though not for the basic direction of the result. Single realizations per setup and no released code/data are minor but worth a referee note.\n\nThis is a careful, honest paper that deserves a serious referee. I would send it to review, but ask for a measured t_cool/t_mix analysis, a heating-prescription sensitivity discussion, and ideally code/data release. It is a good reading-group paper too: the gap between stated and realized control parameter is an instructive failure mode.","headline":"Important sub-sampling warning and clean high-res MHD results, but the central t_cool/t_mix scaling is shakier than the abstract claims — Table 1 shows the parameter is not actually held fixed across the Tmix series.","tokens_in":24317,"tokens_out":4687,"would_cite":true,"duration_ms":42536,"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":"Cold gas survives in the circumgalactic medium only where cooling beats turbulent mixing.","keywords":["circumgalactic medium","multiphase gas","MHD turbulence","cold gas survival","cooling time","mixing time","t_cool/tmix","baryon cycle"],"falsifier":"If observations find cold gas below 10^4 K persisting for ≳100 Myr in low-density, highly turbulent CGM gas with t_cool/t_mix ≳ 10 (n ≈ 3×10^-4 cm^-3), or if a simulation run without the global compensating heating term fails to reproduce the 20–50% cold fractions at matched t_cool/t_mix, the central claim would be falsified.","tokens_in":23163,"feed_emoji":"🌌","tokens_out":9339,"duration_ms":97760,"temperature":0.7,"pith_summary":"This paper tries to establish that the survival of cold gas (T < 10^4 K) in the circumgalactic medium is controlled by a single dimensionless ratio: the cooling time divided by the turbulent mixing time, t_cool/t_mix. In magnetohydrodynamic simulations of 0.125–8 kpc CGM patches that resolve the cooling length, low-density gas (n ≈ 3×10^-4 cm^-3) cannot sustain cold gas because t_cool/t_mix ≫ 1, while denser environments (n ≈ 3×10^-3 cm^-3) reach a multiphase steady state with up to 50% of the mass in the cold phase. The key scaling result is that simulations with similar t_cool/t_mix produce comparable cold gas mass fractions and lifetimes across box sizes, making this ratio a predictive control parameter for the CGM. If correct, this means the small-scale (≲10 kpc) cold gas seen in absorption arises in relatively dense, quiescent cloud complexes, not in turbulence cascading from large scales — a concrete prediction for where and how cold gas is distributed in halos.","feed_headline":"One ratio decides if cold gas survives in a galaxy's halo","feed_subtitle":"Cooling time vs mixing time, not box size, sets cold gas survival — and it favors dense quiet regions.","key_machinery":"The operating object is the dimensionless ratio t_cool/t_mix, where t_cool is the isobaric cooling time and t_mix = l_int / v_l is the turbulent mixing time at the integral scale; for multiphase gas, t_mix carries a √χ factor for the hot/cold density contrast. The paper varies box size (0.125–8 kpc) under two scalings — constant energy injection rate (so t_mix ∝ l^{2/3}, mimicking a turbulent cascade) and constant t_mix — to isolate this ratio's role. The identity that carries the argument is scale-invariance: runs with matched t_cool/t_mix reproduce each other's cold gas mass fraction and lifetime, making the ratio a transferable control parameter between local patch simulations and cosmolo","core_discovery":"The central claim is that t_cool/t_mix — the ratio of isobaric cooling time to turbulent mixing time at the integral scale — is the control parameter for multiphase gas in the circumgalactic medium, not the box size or turbulent heating rate. Simulations with matched t_cool/t_mix show comparable cold gas mass fractions and lifetimes even when the box size differs by a factor of 64, whereas holding the energy injection rate fixed shortens t_mix in smaller boxes and erases the cold phase. Low-density gas (3×10^-4 cm^-3) has t_cool/t_mix ≥ 1 at all temperatures and cannot sustain gas below 10^4 K, while denser gas (3×10^-3 cm^-3) reaches a multiphase steady state with 20–50% of mass cold occupy","pith_inferences":["If t_cool/t_mix is a universal control parameter, cold-gas survival could be estimated directly from observed density, temperature, and velocity dispersion, without running simulations.","The quantitative cold-gas fractions (20–50%) likely depend on the paper's responsive compensating heating term; a less responsive heating prescription could shift the budget even if the t_cool/t_mix scaling holds.","The flat density power spectra imply multiphase gas adds density fluctuations down to AU scales, predicting enhanced FRB scattering that current FRB observations can test.","The ICM-like runs suggest matching t_cool/t_mix at the integral scale already absorbs the density-contrast correction via smaller cloud sizes; a testable extension is that cold gas mass fraction is roughly independent of χ at matched t_cool/t_mix."],"forward_implications":["Cold gas at ≲10 kpc scales in the CGM is expected to live in relatively dense, quiescent cloud complexes; large-scale turbulent cascade tends to destroy it.","Low-density CGM gas (n ≈ 3×10^-4 cm^-3) is largely single-phase: any cold gas that forms evaporates within a few mixing times.","Cosmological and meso-scale simulations cannot be compared by simple volume sub-sampling; they should be compared at matched t_cool/t_mix.","Dense CGM environments can sustain a steady multiphase state with ~20–50% cold mass fraction (hydro vs MHD), ~1% volume fraction, and area covering fractions near 80%.","Magnetic fields suppress mixing and fragmentation, extending cold gas lifetime and producing filamentary cold structures."],"fun_headline_variants":["Why cold gas lives: it's a cooling-to-mixing ratio","Galaxy halo cold gas: density rules, not turbulence","One number predicts cold gas survival in galaxy halos","Cold gas in halos needs dense calm spots, not turbulence","Turbulence isn't the key: cooling vs mixing decides"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The simulations maintain global thermal balance with an artificial heating term (eq. 2b) that returns all net cooling and turbulent work as heat; the simulated cold-gas fractions, including the up-to-50% steady state, depend on that heating being as responsive as assumed.","fun_headline_variants_meta":{"raw":{"variants":["Why cold gas lives: it's a cooling-to-mixing ratio","Galaxy halo cold gas: density rules, not turbulence","One number predicts cold gas survival in galaxy halos","Cold gas in halos needs dense calm spots, not turbulence","Turbulence isn't the key: cooling vs mixing decides"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000191,"raw_usage":{"total_tokens":1295,"prompt_tokens":975,"completion_tokens":320,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":719,"completion_tokens_details":{"reasoning_tokens":250}},"tokens_in":719,"tokens_out":320,"duration_ms":3795,"temperature":1.0,"reasoning_tokens":250,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-04T06:51:34.704439+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"If observations find cold gas below 10^4 K persisting for ≳100 Myr in low-density, highly turbulent CGM gas with t_cool/t_mix ≳ 10 (n ≈ 3×10^-4 cm^-3), or if a simulation run without the global compensating heating term fails to reproduce the 20–50% cold fractions at matched t_cool/t_mix, the central claim would be falsified.","supporting_citations":[],"review_version":1}