{"id":"e68df32e-92f4-4317-8a10-d299d4557a65","arxiv_id":"2412.08837","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":9,"one_line_summary":"A radiative-transfer analysis of Mg II spectra from 624 z~1 galaxies finds more massive galaxies have higher column density and slower-moving cold gas.","lead":"This study fits simulated spectra to Mg II emission and absorption from 624 galaxies at z~1 and finds that more massive galaxies have more cold gas that moves slowly. It also shows that simple spherical models predict too much halo light around the most massive systems, pointing to anisotropic gas or dust.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The claimed negative N_MgII-v_exp correlation and the mass trend in v_exp are likely dominated by the absorber/emitter dichotomy; within-class correlations are not reported, so the central 'slowly moving cold gas' claim is not yet established.","rationale":"The reader's weakest assumption is the spherical, single-phase, dust-free geometry. That is a real limitation, but the more immediately testable threat to the central claim is statistical: the aggregate correlations used for the headline result are computed over a sample whose spectral classification is strongly mass-dependent. Figure 5's lower triangle separates absorbers from emitters/P-Cygni in exactly the direction claimed as a physical anti-correlation. If that correlation disappears within each spectral class, the conclusion 'higher stellar mass galaxies have slowly moving cold gas' is not supported by the data as presented. This does not impugn the observational stacking or the modeling effort; it means the claimed relation requires a class-controlled reanalysis and significance estimates. The v_exp resolution limit reinforces the concern. The paper remains a useful proof-of-concept, and a conditional acceptance with the requested check is appropriate, matching the reader's verdict.","tokens_in":22751,"tokens_out":11168,"duration_ms":130897,"concrete_test":"Using the published fitting results behind Figure 5, recompute the Pearson r between log N_MgII and v_exp separately for the absorption-only subsample and for the emission+P-Cygni subsample, and also compute partial correlations controlling for stellar mass within each class. If the within-class correlations are not significantly negative (or vanish), the claimed anti-correlation is an artifact of population mixing; if they remain negative, the concern is refuted.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The headline physical result (Section 6) is that higher stellar mass galaxies show higher N_MgII and lower v_exp, indicating an abundance of slowly moving cold gas. The evidence is the aggregate correlations in Figure 6 (r=0.25 for log N-M*, r=-0.15 for v-M*). Section 5.1's own Figure 5 shows that absorbers and emitters/P-Cygni occupy opposite regions of the N-v plane: absorbers have high N (>=10^14.5 cm^-2) and v_exp near 0, while emitters/P-Cygni have lower N (<=10^14 cm^-2) and v_exp around 200-500 km/s. Because the absorber fraction increases steeply with stellar mass (Figure 1, Table 2), the aggregate correlations can arise simply from mixing these two classes, with no continuous physical anti-correlation within either class. The paper does not report within-class Pearson coefficients or partial correlations controlling for spectral type, so the statement 'regions with higher column densities tend to contain slower-moving gas' conflates between-class differences with an intrinsic relation. Moreover, MUSE spectral resolution (~35-85 km/s Gaussian width) is comparable to the v_exp<100 km/s values fitted for high-mass absorbers, making those values effectively upper limits; the weak r=-0.15 trend may not be statistically significant. The halo overproduction factor ~3 is also post hoc (Sections 5.3-5.4), but the population-mixing issue directly undermines the central mass-kinematics claim.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This manuscript models Mg II resonance doublet spectra of 624 z~0.7-2.3 star-forming galaxies from the MAGG and MUDF MUSE programs, fitting 167 individual detections and stacked spectra in four stellar-mass bins with a five-parameter spherical, expanding, single-phase halo model in the RT-scat Monte Carlo radiative transfer code. The paper reports that fitted Mg II column density increases with stellar mass (Pearson r=0.25), fitted expansion velocity decreases with mass (r=-0.15), and random velocity increases with mass (r=0.38). It further reports a negative N_MgII-v_exp correlation and interprets these trends as evidence that more massive galaxies host an abundance of slowly moving cold gas. For the highest-mass stack, the best spherical model overproduces the 10-30 kpc halo spectrum and surface-brightness profile by a factor of about 3; the authors argue that intrinsic absorption, dust, and anisotropic gas (illustrated with a 30-degree bipolar wind plus static H I disk) can account for this suppression.","tokens_in":23147,"tokens_out":7472,"duration_ms":83661,"significance":"If the mass-dependent trends were robust, the paper would establish a z~1 connection between stellar mass and the cold CGM reservoir and kinematics, and it would demonstrate that spatially resolved Mg II spectra can be interpreted through full radiative-transfer forward modeling. The paper has notable strengths: a large and homogeneous sample, a transparent five-parameter grid of 201,600 simulated spectra, a small reported fraction of poor fits (5/167), and a first attempt to model core and halo spectra plus surface-brightness profiles together. However, the central mass-kinematics claim is currently vulnerable to the demographic composition of the sample, to spectral-resolution limits, and to the assumed model geometry; the halo-anisotropy interpretation is post hoc rather than fitted.","major_comments":[{"comment":"The central mass-kinematics claim is not established by the reported aggregate correlations. Figure 5 shows that absorbers occupy the high-N_MgII/low-v_exp region while emitters and P-Cygni sources occupy lower-N_MgII/higher-v_exp regions, and Table 2 and Figure 1 show that the absorber fraction rises steeply with stellar mass. Mixing these two populations therefore produces a positive N_MgII-M* correlation and a negative v_exp-M* correlation even if no such trend exists within either class. The paper should report Pearson or Spearman coefficients separately for absorbers and for emitters/P-Cygni, partial correlations controlling for spectral class, and p-values or confidence intervals. Without this, the Section 6 statement that higher stellar mass galaxies exhibit 'higher N_MgII and lower v_exp' conflates a demographic shift with a physical mass dependence.","section":"§5.2, Fig. 6; §5.1, Fig. 5; Table 2"},{"comment":"The decrease of v_exp with stellar mass is not resolution-safe. The grid step in v_exp is 50 km/s (Table 3) and the MUSE Gaussian resolution is 35-85 km/s (Appendix A, Eq. A1), so the fitted v_exp values below ~100 km/s for the high-mass absorbers are effectively upper limits. The weak r=-0.15 correlation may therefore reflect censoring rather than a physical trend. I request a resolution-aware test: for example, restricting the correlation to objects with v_exp well above the local resolution, using upper-limit statistics, or demonstrating that the per-mass stacked fits are inconsistent with v_exp=0 at the same confidence.","section":"§5.2, Fig. 6; Table 3; Appendix A"},{"comment":"The halo-overproduction interpretation is more an existence proof than a constraint. The spherical model fixes a constant-density, single-phase, dust-free medium with a radial velocity field (§3.1), and the anisotropic model in §5.4 is not fitted to the observed stacked spectra: the 30-degree opening angle, the fixed H I disk column density, and the disk geometry are chosen by hand, and the demonstration that this geometry suppresses the azimuthally averaged SB by a factor of about 3 is not accompanied by a goodness-of-fit comparison with the spherical model or any exploration of the opening-angle/column-density parameter space. The abstract and Section 6 statements that the fits 'indicate the presence of intrinsic Mg II absorption and strong anisotropy' therefore go beyond what the current post hoc comparison can establish.","section":"§3.1, §5.3, §5.4, Fig. 9"},{"comment":"The correlation statistics are computed from the weighted-average parameter values without propagating the weighted standard deviations, despite the known N_MgII-EW_int degeneracy shown in Figure A3. The paper should at least quote uncertainties on the regression slopes and demonstrate that the N_MgII-M* and v_exp-M* trends survive when only fits with small weighted standard deviations are used. This is particularly important because Figure 6 reports Pearson coefficients with no p-values and no error bars on the correlation itself.","section":"Appendix A, Eqs. (A3)-(A4), Fig. A3"}],"minor_comments":[{"comment":"The heading reads 'DEATILS' and should be corrected to 'DETAILS'.","section":"Appendix A heading"},{"comment":"The convolution description is dimensionally unclear: as written, 'width ~ 2.355 R_obs' is not a velocity or wavelength width, whereas the stated 35-85 km/s Gaussian widths imply the intended convolution kernel is sigma = c/(2.355 R) (or the analogous wavelength expression). Please clarify the formula and the units.","section":"Appendix A, Eq. (A1)"},{"comment":"The symbol p is used for Pearson's correlation coefficient, which can be confused with a p-value. Use r throughout and report p-values separately.","section":"Fig. 6 caption and text"},{"comment":"The text acknowledges that sigma_R measurements are inaccurate when the spectral resolution exceeds sigma_R, yet sigma_R is still presented as a mass-dependent result with r=0.38. This caveat should be carried into the conclusions or the claim should be restricted to the resolution-supported regime.","section":"§5.2.1, Fig. 6 (right panel)"},{"comment":"The notation 'H i disk' should be rendered as 'H I disk' for consistency with standard usage.","section":"§5.4"}],"recommendation":"major_revision","confidential_remarks":"This is a promising proof-of-concept paper and the modeling effort is substantial, but the headline mass-kinematics result is not yet supported by the statistics as presented. I would ask the editor to require the within-class/partial correlation analysis and a resolution-aware treatment of v_exp before acceptance. If those analyses do not support the current interpretation, the conclusions should be reframed as a demographic trend rather than a continuous cold-gas scaling. The post hoc anisotropic-wind model should be clearly labeled as illustrative unless it is fitted to the data."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nYou should know up front that this is a useful and mostly honest paper, but the headline result is not as solid as the abstract makes it sound. The new thing here is the application of a published Mg II radiative transfer code (RT-scat) to a large MUSE sample at z~1 (624 galaxies, 167 individual fits plus four mass stacks), with core and halo spectra modelled together. The mass-dependent trends in the stacked spectra – strong core absorption and extended halo emission at high stellar mass, emission-dominated cores at low mass – are credible, and the fit quality is documented carefully (only 5/167 bad fits, chi-square maps in the appendix). That is real work and worth building on.\n\nThe soft spot is the central physical claim. The paper reports a negative correlation between N_MgII and v_exp and a mass trend of decreasing v_exp with stellar mass, interpreting it as an abundance of slowly moving cold gas in massive galaxies. But Figure 5 of the paper itself shows that absorbers and emitters/P-Cygni occupy opposite regions of the N-v plane: absorbers have high N and v_exp near zero, emitters/P-Cygni have lower N and outflow velocities of a few hundred km/s. Because the absorber fraction rises steeply with stellar mass, the aggregate correlations in Figure 6 (r=0.25, -0.15, 0.38) are likely dominated by this mix. The paper does not report within-class correlations or partial correlations controlling for spectral type, so the continuous anti-correlation is not established. Also, MUSE's spectral resolution (35-85 km/s Gaussian) is comparable to the v_exp values fitted for the high-mass absorbers, so those values are effectively upper limits, and r=-0.15 is weak with no significance quoted.\n\nThe halo overproduction factor ~3 for the most massive bin is an interesting new observable, but the bipolar-wind explanation in Section 5.4 is post hoc: an opening angle of 30 degrees is chosen to reproduce exactly that factor. The text is honest about this being a demonstration, but the conclusion that 'strong anisotropy' is present goes beyond what the model fitting supports. Same for the claims about dust.\n\nNone of this is fatal. The paper is transparent about its simplified geometry (spherical, dust-free, single-phase, constant density), and it frames the work as a first step. What it establishes is a robust observational characterization: massive z~1 galaxies show more absorption and more extended Mg II emission. What it does not yet establish is the specific continuous relationships among N, v_exp, and stellar mass.\n\nI would send this to referees. The referee should ask for within-class correlations, significance estimates for the Pearson coefficients, and a recalibration of the anisotropy conclusion as an illustrative explanation rather than a detection. I'd cite this for the method and the mass-dependent spectral trends, not for the N-v_exp correlation.\n\nRecommendation: engage with it, but with the above caveats.","headline":"A solid modeling application with a believable mass-dependent spectral trend, but the central N-v_exp anti-correlation is likely a population-mixing artifact and needs within-class analysis.","tokens_in":23686,"tokens_out":3277,"would_cite":true,"duration_ms":34985,"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":"Across 167 z~1 star-forming galaxies, the fitted Mg II column density rises with stellar mass while the fitted expansion velocity falls, leading the authors to conclude that massive galaxies hold abundant, slowly moving cold gas.","keywords":["Mg II resonance doublet","circumgalactic medium (CGM)","radiative transfer modeling","galaxy haloes","cold gas at 10^4 K","stellar mass dependence","spectral stacking","z ~ 1 star-forming galaxies"],"falsifier":"A decisive calculation would refit the same MAGG and MUDF stacks with a clumpy or multiphase Mg II medium plus an intrinsic absorbing disk and random outflow orientations while keeping the same goodness-of-fit; if the $N_{\\rm MgII}$–stellar-mass correlation weakens or the $v_{\\rm exp}$–stellar-mass anti-correlation disappears, the spherical single-phase geometry was the load-bearing cause of the claimed trends. Observationally, deep MUSE maps of individual $M_*/M_\\odot > 10^{10}$ galaxies at $z\\sim1$ can discriminate between the models: the spherical halo predicts roughly circular, centrally concentrated Mg II halo emission, whereas the anisotropic model concentrates emission along outflow axes, so the axial ratio and position-angle coherence of the Mg II halo would settle which geometry describes massive galaxies.","tokens_in":22487,"feed_emoji":"🔭","tokens_out":11441,"duration_ms":105098,"temperature":0.7,"pith_summary":"The paper argues that at $z\\sim 1$ the amount and kinematics of $10^4$ K gas around star-forming galaxies are set by stellar mass. Fitting a grid of 201,600 radiative-transfer models to Mg II doublet spectra of 167 galaxies and four mass-binned stacks from the MAGG and MUDF surveys, it finds that the fitted Mg II column density $N_{\\rm MgII}$ rises with stellar mass (Pearson $r=0.25$) while the fitted expansion velocity $v_{\\rm exp}$ falls ($r=-0.15$). The authors conclude that more massive galaxies are surrounded by abundant, slowly moving cold gas. A second claim is that the spherical best-fit model overproduces the extended halo emission of the most massive stack by roughly a factor of three, which the paper attributes to intrinsic Mg II absorption and anisotropic gas distribution. If correct, Mg II spectra can directly constrain the cold circumgalactic reservoir and its kinematics in the ground-observable redshift window $0.7 < z < 2.3$.","feed_headline":"Stellar mass sets how much slow cold gas surrounds galaxies","feed_subtitle":"Fitting 167 Mg II spectra links high stellar mass to higher column density and slower outflows.","key_machinery":"The central object is the 3D Monte-Carlo radiative-transfer code RT-scat, which generates the 201,600 simulated Mg II doublet spectra used for fitting. The base model is a spherical Mg II halo with inner radius 1 kpc and outer radius $R_{\\rm H} = 100$ kpc, filled at constant number density $n_{\\rm MgII} = N_{\\rm MgII}/0.99R_{\\rm H}$, with microturbulent random motion $\\sigma_{\\rm Ran}$ and a radial velocity field $v(r) = v_{\\rm exp}\\,r/R_{\\rm H}$; because radiative transfer depends on intercepting column density and kinematics rather than physical distance, the solutions can be rescaled to any halo radius. The five free parameters are $N_{\\rm MgII}$, $v_{\\rm exp}$, $\\sigma_{\\rm Ran}$, the intrinsic emission width $\\sigma_{\\rm Src}$, and the intrinsic emission equivalent width ${\\rm EW}_{\\rm int}$. The fitting pipeline normalizes the observed spectra, convolves the simulated spectra with MUSE spatial and spectral resolution, applies a chi-square test over the velocity range $-800$ to $+1300$ km s$^{-1}$, and uses weighted averages of the five parameters to break degeneracies. A separate bipolar-wind model with an inner H I disk is used to show that asymmetric gas can suppress the azimuthally averaged Mg II surface brightness by the observed factor of about three.","core_discovery":"Using the 3D Monte-Carlo radiative-transfer code RT-scat, the paper models the Mg II $\\lambda\\lambda2796,2803$ doublet as scattering in a spherical, constant-density, dust-free halo of Mg II, with microturbulent random motions $\\sigma_{\\rm Ran}$ and a purely radial velocity field $v(r) = v_{\\rm exp}\\,r/R_{\\rm H}$, plus a central point source supplying a flat continuum and intrinsic Gaussian Mg II emission. Fitting this five-parameter model to the core spectra ($R_{\\rm p}<10$ kpc) reproduces most observed profiles, with only 5 of 167 individual fits reaching $\\chi^2_{\\min} > 10$. The fitted parameters separate cleanly by spectral type: absorbers have high $N_{\\rm MgII}$ and low $v_{\\rm exp}$, while P-Cygni and emission spectra have lower $N_{\\rm MgII}$ and higher $v_{\\rm exp}$. Across stellar mass, $N_{\\rm MgII}$ increases ($r=0.25$) and $v_{\\rm exp}$ decreases ($r=-0.15$), so the paper states that 'higher stellar mass galaxies exhibited higher $N_{\\rm MgII}$ and lower $v_{\\rm exp}$ values, indicating an abundance of slowly moving cold gas in massive galaxies.' When the best-fit core model is projected into the halo aperture (10–30 kpc), the highest-mass stack ($M_*/M_\\odot > 10^{10}$) is overproduced by a factor of about three; the paper interprets this as evidence for intrinsic Mg II absorption and strong anisotropy in the cold gas around massive haloes.","pith_inferences":["The paper does not spell this out, but the negative $N_{\\rm MgII}$–$v_{\\rm exp}$ correlation suggests a two-phase reading of the same haloes: outflowing gas is diffuse and fast, while deposited or infalling gas is dense and slow, so the mass trend may trace the buildup of a static cold reservoir in massive haloes.","The factor-of-three overproduction of the most massive halo stack can be translated into an effective clumping or covering-factor constraint: a clumpy medium lowers the fraction of scattered photons that reach the halo aperture, so matching the observed halo flux would require either less total Mg II or a narrower outflow geometry than the spherical fit assumes, which could flatten the fitted mass","The paper's own error discussion implies that $\\sigma_{\\rm Ran}$ and $\\sigma_{\\rm Src}$ are often poorly constrained when the MUSE resolution approaches the line width, so the reported random motions and intrinsic widths should be read as order-of-magnitude estimates rather than precise values.","Since the fit depends on column density and kinematics rather than physical radius, the same simulated grid can be rescaled to other halo radii and applied to future Mg II surveys; coupling the code with a hydrogen ionization model would extend the same machinery to joint Mg II and Ly$\\alpha$ predictions."],"forward_implications":["If the fitted trends are physical, stellar mass, not environment, controls the cold CGM at $z\\sim 1$: massive galaxies hold a larger reservoir of $10^4$ K gas and that gas moves more slowly.","The spectral-type separation (absorbers: high $N_{\\rm MgII}$, low $v_{\\rm exp}$; emitters and P-Cygni: low $N_{\\rm MgII}$, high $v_{\\rm exp}$) means the Mg II profile shape itself is a quick diagnostic of whether a galaxy is dominated by static cold gas or by outflowing gas.","Because only 5 of 167 individual fits have $\\chi^2_{\\min} > 10$, a five-parameter spherical model captures the dominant physics of most Mg II profiles, making Mg II a practical probe of the cold CGM at redshifts where Ly$\\alpha$ requires space-based telescopes ($0.7 < z < 2.3$).","The factor-of-three overproduction of the highest-mass halo spectrum implies that extended Mg II halos of massive galaxies cannot be interpreted as isotropic scattering alone; intrinsic absorption, anisotropy, or dust must be included in any future CGM census from emission."],"supporting_citations":[{"why":"Supplies the MAGG/MUDF galaxy catalog and the stacking procedure that produces the core and halo Mg II spectra used for fitting.","marker":"Dutta et al. 2023"},{"why":"The companion study that explores Mg II radiative-transfer parameter dependencies and underpins the fitting grid.","marker":"Chang & Gronke 2024"},{"why":"Introduces the RT-scat 3D Monte-Carlo radiative-transfer code used to generate the 201,600 simulated spectra.","marker":"Chang et al. 2023"},{"why":"Early theoretical work establishing Mg II radiative transfer as a probe of galaxy halo gas, the basis for interpreting the doublet as scattering.","marker":"Prochaska et al. 2011"},{"why":"Shows that spatially extended Mg II halos can be modeled through absorption imprinted on the observed spectrum, a direct predecessor of this fitting approach.","marker":"Zabl et al. 2021"},{"why":"Demonstrates modeling of Mg II halo emission and highlights anisotropy, the alternative interpretation the paper invokes for the overproduced halo.","marker":"Pessa et al. 2024"},{"why":"Presents observations of anisotropic Mg II halos used to support the asymmetric-gas explanation for the suppressed surface brightness.","marker":"Guo et al. 2023"},{"why":"Previous full radiative-transfer modeling of Mg II and Ly-alpha spectra at z~0.2, the closest methodological precedent.","marker":"Li et al. 2024"}],"fun_headline_variants":["Massive galaxies harbor slower, denser cold gas","Stellar mass drives cold gas column density and outflow speed","High-mass galaxies: more cold gas, slower outflows","Galaxy mass sets cold gas abundance and outflow velocity"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the scattering gas around each galaxy is a single smooth spherical shell with no dust, no clumps, and motion only along the line from the galaxy center; if the real gas is clumpy, dusty, or moves sideways, the fitted column density, outflow speed, and random speed are all biased, and the paper itself says in Sections 5.3 and 6 that asymmetric gas, dust, and multiphase structure are not captured by the fitting pipeline.","fun_headline_variants_meta":{"raw":{"variants":["Massive galaxies harbor slower, denser cold gas","Stellar mass drives cold gas column density and outflow speed","High-mass galaxies: more cold gas, slower outflows","Galaxy mass sets cold gas abundance and outflow velocity"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000425,"raw_usage":{"total_tokens":2323,"prompt_tokens":1235,"completion_tokens":1088,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":851,"completion_tokens_details":{"reasoning_tokens":1023}},"tokens_in":851,"tokens_out":1088,"duration_ms":9349,"temperature":1.0,"reasoning_tokens":1023,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T17:30:27.184541+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A decisive calculation would refit the same MAGG and MUDF stacks with a clumpy or multiphase Mg II medium plus an intrinsic absorbing disk and random outflow orientations while keeping the same goodness-of-fit; if the $N_{\\rm MgII}$–stellar-mass correlation weakens or the $v_{\\rm exp}$–stellar-mass anti-correlation disappears, the spherical single-phase geometry was the load-bearing cause of the claimed trends. Observationally, deep MUSE maps of individual $M_*/M_\\odot > 10^{10}$ galaxies at $z\\sim1$ can discriminate between the models: the spherical halo predicts roughly circular, centrally concentrated Mg II halo emission, whereas the anisotropic model concentrates emission along outflow axes, so the axial ratio and position-angle coherence of the Mg II halo would settle which geometry describes massive galaxies.","supporting_citations":[],"review_version":1}