{"id":"6a6786c0-0a3f-4219-8617-e35d3ec4fa31","arxiv_id":"2506.08290","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":6,"one_line_summary":"Radiative transfer fits to 163 MUSE Lyα emitters at z=3-6 favor a broad, slowly expanding circumgalactic medium with halo size set by gas and source extents.","lead":"This paper fits a six-parameter expanding-halo model to the Lyα spectra and surface brightness profiles of 163 galaxies at redshifts 3 to 6. The best-fit models point to a broadly extended, nearly static cloud of gas around each galaxy.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Pile-up of best-fit rsHI at the grid maximum (0.9 rmax) makes the 'broad static CGM' claim dependent on an untested grid boundary.","rationale":"The reader's weakest assumption correctly identifies the grid-boundary problem, and I agree that the central physical conclusion is conditional on the search space being complete. However, I would refine the concern: the most load-bearing boundary is rsHI at 0.9 rmax, because the model cannot represent a more extended medium at all. The Vedge concentration at 0 km/s is not strictly at a grid boundary, since the parameter grid allows negative Vedge (e.g., Vpeak=100, DeltaV=-500 gives Vedge=-400). The pile-up at Vedge=0 could be physical (red-peak-only spectra favor outflow) or a secondary effect of the rsHI boundary, but it does not have the same direct truncation argument. The concrete test of extending rsHI and rmax would settle whether the best-fit pile-up is a real preference for a broad medium or a consequence of the grid edge. Because this concern matches the reader's overall conditional acceptance and does not move the verdict, I keep the verdict unchanged.","tokens_in":84,"tokens_out":4385,"duration_ms":66268,"concrete_test":"Refit a representative subset (e.g., the 20 highest-S/N galaxies) with an extended grid that includes rsHI in {1.0, 1.2, 1.5, 2.0} and a larger rmax (or equivalently allow the SBP fit to use a larger radial range), keeping all other parameters fixed. If the best-fit rsHI values move off the 0.9 boundary and/or the reduced chi-square improves by a meaningful amount (e.g., Delta chi^2 > 1 per object), the pile-up is an artifact and the 'broad medium' claim is not robust. Also check rsHI = 0.9 with rmax increased by 50% to separate the scale-radius effect from domain truncation.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's central conclusion that the CGM is extended and static at large radii rests on the best-fit distributions in Fig. 10: rsHI peaking at 0.9 rmax and Vedge at 0 km/s. The rsHI pile-up is at the upper edge of the adopted grid (Table 1: rsHI in [0.1, 0.9] in steps of 0.1). Because the model's density profile is exponential with scale radius rsHI and is truncated at rmax, any true best fit with rsHI > 0.9 cannot be selected; it would map onto the boundary value. The paper provides no test of rsHI > 0.9 (e.g., rsHI = 1.0, 1.5) or of a larger domain rmax. The Vedge pile-up at 0 is less obviously a boundary artifact because the derived grid allows negative Vedge (Vpeak + DeltaV with Vpeak >= 100 and DeltaV >= -500), but the rsHI boundary alone is sufficient to undermine the claim that the medium is 'broad' — the fit may simply be truncated by the search range. This is a search-space completeness issue, not a physical detection.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"Yu et al. model the Lyα spectra and surface brightness profiles (SBPs) of 163 MUSE Lyα-emitting galaxies at z=3–6 using the six-parameter expanding-halo Monte Carlo radiative transfer model of Song et al. (2020). They find that the best-fit models reproduce both observables for most galaxies, and that the simultaneous best-fit values of the HI density scale radius rsHI cluster at 0.9 rmax and the outer-edge velocity Vedge at 0 km/s, which they interpret as evidence for a broad, nearly static outer circumgalactic medium. They further analyze correlations between observables and model parameters, concluding that the spatial extent of Lyα halos is primarily set by the medium and source scale radii, while spectral peak shift and FWHM are mainly governed by optical depth with a secondary role for the velocity structure. The paper also shows that correlations derived from the full model grid and from the best-fit subset differ substantially, and it cautions against overinterpreting the full-grid correlations.","tokens_in":19432,"tokens_out":7432,"duration_ms":92213,"significance":"If the central claim holds, this paper demonstrates that a simple parametric outflow model can simultaneously reproduce Lyα spectra and SBPs for a statistically meaningful sample, extending Song et al. (2020) from 8 to 163 galaxies and placing population-level constraints on the CGM of z=3–6 LAEs. The use of direct MUSE data rather than digitized figures, the systematic MLE-versus-MAP comparison, and the explicit discussion of sample bias in correlation analysis are notable strengths that increase confidence in the fitting procedure. However, the interpretation that the outer medium is broad and static rests on best-fit parameters that pile up at the edge of the adopted parameter grid, and on the neglect of IGM transmission at redshifts where it is known to be important; both issues are load-bearing for the central conclusion and require further testing.","major_comments":[{"comment":"The best-fit rsHI values for the simultaneous fits concentrate at 0.9 rmax, which is the upper boundary of the adopted grid (rsHI ∈ {0.1, ..., 0.9}). The paper interprets this pile-up as evidence that the medium is 'largely extended', but a pile-up at a grid boundary is also what one expects if the true best fit lies beyond the searched range. No test with rsHI > 0.9 or with a larger rmax is presented, so the MLE at 0.9 cannot be taken as a reliable estimate of the preferred scale radius. Because this is one of the two pillars of the central 'broad static medium' claim, the authors should either extend the grid and show the likelihood behavior beyond the boundary, or explicitly qualify rsHI = 0.9 as a lower limit rather than a detection. The qualitative direction of the claim may survive such a test, but the current analysis does not demonstrate it.","section":"Section 4.2, Table 1, Figure 10"},{"comment":"The models ignore IGM transmission even though the sample spans z=3–6, where the mean IGM transmission declines from roughly 60% at z=3 to near zero at z=6 (Inoue et al. 2014). The model has no physical mechanism other than outflowing gas to suppress the blue Lyα peak, so IGM absorption can produce red-peak-dominated spectra even for intrinsically symmetric or weakly outflowing media. This degeneracy may bias the inferred velocity structure, notably Vedge, and possibly rsHI for the highest-redshift objects. The authors acknowledge the issue but do not quantify its impact. I request a quantitative check, such as applying a representative IGM transmission to the model spectra for a subsample and refitting, or at least a careful discussion of how the central conclusion of a static outer CGM is robust to plausible IGM attenuation.","section":"Section 3, paragraph beginning 'It is worth noting...'"}],"minor_comments":[{"comment":"The text states that 19 of 184 LAEs have double-peaked spectra and the rest have red-peak-only spectra, which would leave 165 red-peak-dominated objects, yet the modeling is performed for 163 galaxies; please clarify why two additional galaxies were excluded.","section":"Section 2"},{"comment":"The text says 'the total likelihood was computed by summing the likelihoods calculated from the spectrum and SBP,' but Equation (1) defines a log-likelihood. If the sum is over log-likelihoods, the wording should say so; if the sum is actually over likelihoods (not log-likelihoods), the statistical procedure is incorrect and needs correction.","section":"Section 3, Equation (1) and following paragraph"},{"comment":"Vedge = Vpeak + ΔV is introduced only in the Figure 10 axis label; it would be clearer to define this derived quantity explicitly in the text before discussing its best-fit distribution.","section":"Section 4.2, Figure 10"},{"comment":"The reduced chi-square values are discussed without specifying how the degrees of freedom are computed for each object; because the number of spectral and SBP data points varies and parameters are selected on a discrete grid, a brief definition of the reduced chi-square would improve interpretability.","section":"Section 4.1, Figure 4"},{"comment":"Several correlation p-values are printed as 0.000; these should be reported as <0.001 or as the actual floating-point values to avoid implying a literally zero probability.","section":"Figures 11–14"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is within the scope of a standard astrophysics journal, and the large-sample extension of Song et al. (2020) is a useful contribution. The central issue is not the model family itself but the interpretation of best-fit parameters sitting at the grid boundary without a boundary test; this is straightforward to address by extending the grid. The IGM neglect is a known limitation that the authors acknowledge, but given the redshift range up to z=6, a quantitative or semiquantitative robustness check would significantly strengthen the paper. I do not see grounds for rejection, provided the authors can show that the boundary pile-up does not dominate the conclusion."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper is a solid extension of Song et al. (2020) to 163 MUSE LAEs, but its main physical conclusion—broad, static outer medium—rests on a grid boundary that was never tested. The rsHI best-fit values pile up at the upper edge (0.9 rmax), and the paper does not explore larger scale radii. Until that is done, the 'extended medium' claim is an artifact of the search range, not a detection.\n\nWhat is genuinely new: simultaneous fitting of spectra and SBPs for two dozen times the previous sample, with real MUSE data rather than digitized figures. The inclusion of rs,cont as a free parameter and the systematic comparison of MLE vs MAP, and the full-model-versus-best-fit subset correlation analysis, are useful. The fits generally look good, with reduced chi2 near unity for most objects. The discussion of sample bias in the correlations is honest and a legitimate contribution.\n\nThe soft spots are real. First, the grid boundary issue is load-bearing. Table 1 sets rsHI in [0.1,0.9] with step 0.1; Fig. 10 shows a strong peak at 0.9. The authors interpret this as a broad medium, but it could simply be the largest value allowed. A test with rsHI = 1.0, 1.5, 2.0 is necessary. Second, IGM transmission is ignored even though the sample spans z=3–6. The authors acknowledge this but dismiss it too quickly; mean IGM transmission drops to nearly zero by z=6, and the suppression of the blue peak can mimic the red-peak spectra produced by outflowing halos. Third, there is no data release of the best-fit parameters, which makes reproduction and follow-up harder than it should be. The likelihood sum of spectrum and SBP without explicit weights is a minor concern, but the reduced chi2 values suggest it is not causing obvious problems.\n\nThe citation pattern is fine. Self-citing Song et al. (2020) is appropriate since the model comes from there, and the data are properly attributed to Leclercq et al. (2017).\n\nThis paper deserves a serious referee. For anyone working on Lyα radiative transfer or planning IFU surveys, the large sample and the honest correlation analysis are worth engaging with. The central claim is falsifiable if the grid is extended. I would ask the authors to extend the rsHI grid, add a basic IGM treatment or at least a redshift-dependent test, and publish the best-fit catalog. Without those, the headline conclusion is not yet supported.","headline":"Solid large-sample Lyα RT fitting, but the 'broad static medium' conclusion sits on an untested grid boundary and needs a grid extension before it holds.","tokens_in":19942,"tokens_out":4747,"would_cite":false,"duration_ms":49226,"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":"Six-parameter halo model reproduces 163 distant Lyα galaxies and points to broad, nearly static gas at large radii.","keywords":["Lyman-alpha emitters","radiative transfer","circumgalactic medium","MUSE","surface brightness profiles","galaxy outflows","high-redshift galaxies","Monte Carlo radiative transfer"],"falsifier":"Take a subset of the best-fitting galaxies, extend the grid to $r_{\\mathrm{s,HI}}$ up to $0.99\\,r_{\\mathrm{max}}$ and $V_{\\mathrm{edge}}$ down to −500 km/s (or allow a static outer shell rather than a linear outflow), re-run the MLE fits, and check whether the $r_{\\mathrm{s,HI}}$ and $V_{\\mathrm{edge}}$ distributions remain piled at the new boundaries; if they do, the 'extended static CGM' interpretation is a boundary effect rather than a physical detection.","tokens_in":18901,"feed_emoji":"🌌","tokens_out":4143,"duration_ms":45248,"temperature":0.7,"pith_summary":"This paper claims that a simple expanding-halo radiative transfer model, fitted simultaneously to the Lyα spectrum and surface brightness profile of each of 163 MUSE galaxies at z=3–6, reproduces the data well for most objects. The best-fit parameters pile up at the largest allowed gas scale radius and at zero outflow velocity at the halo edge, which the authors read as evidence that the circumgalactic medium around these galaxies is broadly distributed and nearly static at large radii. The paper also argues that the spatial extent of the Lyα halo is set mainly by the physical extents of the hydrogen medium and the star-forming source, while the spectral peak shift and width are governed chiefly by optical depth, with velocity structure playing a secondary role. The significance is that a single six-parameter model can capture both spectral and spatial Lyα data for a large sample, which would make halo size and outflow parameters jointly measurable from existing IFU observations.","feed_headline":"Distant galaxy halos need broad, nearly static gas","feed_subtitle":"Simultaneous spectra and surface brightness fits point to extended hydrogen with low outflow at large radii.","key_machinery":"The carrying object is the Monte Carlo Lyα radiative transfer code of Song et al. (2020), built on a spherically symmetric halo with an exponential hydrogen density profile of scale radius $r_{\\mathrm{s,HI}}$ and a piecewise-linear outflow velocity profile set by $V_{\\mathrm{peak}}$, $r_{\\mathrm{peak}}$, and $\\Delta V$ (equivalently the edge velocity $V_{\\mathrm{edge}} = V_{\\mathrm{peak}} + \\Delta V$), plus a total optical depth $\\tau_0$ and the source scale radius $r_{\\mathrm{s,cont}}$. The novelty in this paper is that for each galaxy the model spectrum and surface brightness profile are generated through the same transfer calculations and compared to the MUSE data through a summed Neyman chi-square likelihood, so the same set of photons simultaneously constrains spectral shape and spatial extent, breaking degeneracies that plague spectrum-only fits.","core_discovery":"On the paper's own terms, the central discovery is that the spectra and surface brightness profiles of 163 red-peak-dominated Lyα emitters in the MUSE Hubble Ultra Deep Field are all approximately reproduced by one six-parameter expanding-halo model, and that the best-fit values concentrate at $r_{\\mathrm{s,HI}} = 0.9\\,r_{\\mathrm{max}}$ and $V_{\\mathrm{edge}} = 0$ km/s. The authors interpret this pile-up as showing that the hydrogen medium around these galaxies is extended and essentially static at large radii, and they show that this conclusion can only be reached when spectra and surface brightness profiles are modelled together: fitting either observable alone leaves the parameters poorly determined. A correlation analysis then connects observables to physical parameters, showing that Lyα halo size tracks the medium and source scale radii, whereas spectral peak shift and FWHM track optical depth, with outflow kinematics modulating the trends and adding scatter.","pith_inferences":["If the adopted grid had allowed $r_{\\mathrm{s,HI}}$ beyond $0.9\\,r_{\\mathrm{max}}$ and $V_{\\mathrm{edge}}$ below 0 km/s, the best fits might have moved further out and further down, making the paper's physical conclusion a lower limit on gas extent rather than a measurement.","The authors neglect IGM transmission, which is strongest at z≈6, so the low-velocity edge could partly absorb blue-wing photons; a natural extension is to fold in per-object IGM transmission to see whether the $V_{\\mathrm{edge}}$ distribution shifts.","The same joint-fitting machinery could be applied to double-peaked LAEs, spectroscopically confirmed redshifts, or higher-signal-to-noise IFU data to test whether the pile-up persists with independent data."],"forward_implications":["If the claim holds, halo size measurements from Lyα surface brightness maps can be translated into physical gas and stellar extents for large samples, without needing expensive hydrodynamical simulations.","The best-fit pile-up at large $r_{\\mathrm{s,HI}}$ and low $V_{\\mathrm{edge}}$ implies that outflow deceleration zones, where gas returns to near rest at large radii, are common in the circumgalactic medium of z=3–6 star-forming galaxies.","Because spectrum-only or SBP-only fits scatter much more than joint fits, future Lyα surveys should treat spectral and spatial data as a single constraint set; conclusions drawn from either alone will be unreliable.","The peak-shift–FWHM relation commonly used to estimate systemic redshifts is shown to be sample-dependent: red-peak-only galaxies and double-peaked galaxies occupy different regions of that plane, so the relation needs velocity-profile information to be predictive."],"supporting_citations":[{"why":"Supplies the Monte Carlo Lyα radiative transfer code and the six-parameter expanding halo model that this paper applies to the larger sample; also provides the eight-galaxy baseline to which the new fits are compared.","marker":"Song et al. (2020)"},{"why":"Source of the MUSE spectra and surface brightness profiles for the 163 galaxies analysed here, including the redshift estimates and sample selection criteria.","marker":"Leclercq et al. (2017)"},{"why":"Provides the empirical peak shift–FWHM relation used as the comparison baseline for the redshift-estimation discussion, against which the paper's red-peak-only sample is shown to be offset.","marker":"Verhamme et al. (2018)"},{"why":"Earlier MUSE sample that established the detection and measurement context for extended Lyα halos, which Leclercq et al. built upon.","marker":"Wisotzki et al. (2016)"}],"fun_headline_variants":["One static-halo model fits all 163 Lyα galaxies","Lyα halos: broad, static gas from joint fits","Spectra + SBP joint fit: static gas halos","Static gas at large radii revealed by joint Lyα modeling","163 galaxies: one halo model, static gas"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The conclusion depends on the assumption that the true best fit for each galaxy lies inside the chosen parameter grid, since the best-fit values of $r_{\\mathrm{s,HI}}$ and $V_{\\mathrm{edge}}$ sit exactly at the grid boundary and the paper does not test whether extending the grid changes the result.","fun_headline_variants_meta":{"raw":{"variants":["One static-halo model fits all 163 Lyα galaxies","Lyα halos: broad, static gas from joint fits","Spectra + SBP joint fit: static gas halos","Static gas at large radii revealed by joint Lyα modeling","163 galaxies: one halo model, static gas"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000729,"raw_usage":{"total_tokens":3288,"prompt_tokens":995,"completion_tokens":2293,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":611,"completion_tokens_details":{"reasoning_tokens":2211}},"tokens_in":611,"tokens_out":2293,"duration_ms":20605,"temperature":1.0,"reasoning_tokens":2211,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T05:14:47.277452+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take a subset of the best-fitting galaxies, extend the grid to $r_{\\mathrm{s,HI}}$ up to $0.99\\,r_{\\mathrm{max}}$ and $V_{\\mathrm{edge}}$ down to −500 km/s (or allow a static outer shell rather than a linear outflow), re-run the MLE fits, and check whether the $r_{\\mathrm{s,HI}}$ and $V_{\\mathrm{edge}}$ distributions remain piled at the new boundaries; if they do, the 'extended static CGM' interpretation is a boundary effect rather than a physical detection.","supporting_citations":[],"review_version":1}