{"id":"b24d839a-8469-4848-bfc1-da6f32cd50b6","arxiv_id":"1908.02766","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"Massive post-starburst galaxies at redshift 1 to 1.4 show a blue-shifted Mg II absorption component, revealing typical outflows of about 1150 km/s in their interstellar medium.","lead":"Using stacked optical spectra of about 40 distant post-starburst galaxies, astronomers found fast gas outflows moving near 1150 km/s. The result supports the idea that powerful feedback, from a starburst or a brief active nucleus, abruptly shut down star formation in massive galaxies.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Outflow detection hinges on an assumed, partly unconstrained systemic MgII model; a wrong doublet ratio or the Appendix B F-star feature could erase the blue excess.","rationale":"The reader's conditional verdict is well aligned with the load-bearing weakness I find. The paper's claim of high-velocity outflows is supported by several consistency checks: the two-component F-test (p<0.003), the boxcar estimate (809±104 km/s), and the UVBLUE/Coelho-based Models B and C in Appendix B. These checks make a purely statistical artifact unlikely. However, all of them ultimately depend on how the systemic component is represented. The fixed 1.2:1 doublet ratio and the single-Gaussian treatment of the stellar+ISM systemic absorption are parametric assumptions, and the authors explicitly note that the MgII feature is not independently constrained by the rest of the spectrum and that an unidentified F-star feature blue-ward of MgII may contribute to the asymmetry. Because the entire detection is a residual from this model, a systematic error in the doublet ratio or the stellar template could reduce the outflow significance or change the inferred velocity. The boxcar method is less model-dependent in its definition of asymmetry, but it still interprets any blue asymmetry as outflow, so it does not settle the stellar-feature question. The concrete test would settle the matter by varying the systemic model and using empirical stellar templates; until that is done, the appropriate verdict remains conditional rather than accept. I do not see grounds for rejection: the evidence is substantial, the caveats are openly discussed, and the proposed test is a modest additional robustness check rather than a demonstrated failure. Therefore the reader's CONDITIONAL verdict should stand unchanged.","tokens_in":26236,"tokens_out":6936,"duration_ms":91199,"concrete_test":"Re-fit the high-z PSB blue-optimised stack (Fig. 4, upper-left) with the systemic doublet treated as a free parameter: intensity ratios 1:1, 1.2:1, 1.5:1 and 2:1, free intrinsic width, and with/without an additional blue-ward Gaussian at the wavelength indicated by the F-star feature found in Appendix B (5000 < T_eff < 7500 K). For every configuration, run the same F-test and bootstrap and record whether the blue outflow component is required at p<0.05 and what Δv results. Also redo the decomposition using an empirical UV stellar library (e.g., NGSL/IUE stellar spectra) instead of synthetic UVBLUE. If the outflow disappears or shifts by more than ~500 km/s across the grid, the detection is an artifact of the assumed systemic profile rather than a kinematic wind.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central detection rests entirely on the decomposition of one unresolved stacked line. Section 3.2(i) fixes the systemic MgII component to a single Gaussian-convolved doublet with intensity ratio 1.2:1, centred on the template zspec, and calibrates its width from the red side only. The authors themselves flag in Appendix B that (a) a weak, unidentified absorption feature blue-ward of MgII is visible in the UVBLUE stellar fits and is important in F-star atmospheres, and (b) the stellar MgII strength is 'essentially unconstrained' by fits over 2550–4350 Å, since it changes with the mask used. Synthetic UVBLUE/Coelho templates are also known to reproduce the UV MgII region imperfectly. If the true systemic (stellar+ISM) profile has a different doublet ratio (saturated systems approach 1:1 rather than 1.2:1) or contains the F-star feature, then the 25% blue excess attributed to an outflow may be partially or wholly non-kinematic. The Appendix B Models B/C and the boxcar method provide supporting evidence, but Model B forces all systemic absorption into the stellar template and masks the blue side, so it cannot independently validate the assumed stellar profile. The abstract's 'clear evidence' and 'typical in massive PSBs' therefore ride on the least-secure component of the analysis.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper stacks the rest-frame optical spectra of 41 spectroscopically selected post-starburst galaxies at 1 < z < 1.4 in the UDS field and analyzes the Mg ii λ2800 Å absorption profile. The authors fit the stacked continuum to obtain a typical stellar velocity dispersion (σ* ~ 200 km/s for PSBs vs ~ 140 km/s for passive galaxies), and then model the Mg ii feature as a Gaussian-convolved doublet fixed at the systemic redshift, optionally adding a blueshifted outflow component. They report a statistically significant blue excess in PSBs (F-test p < 0.003; Δv ~ 1500 ± 150 km/s; vout ~ 1150 ± 160 km/s after a simulation-based calibration), whereas passive galaxies show no outflow. They use a boxcar method as a cross-check (mean vout ~ 809 ± 104 km/s), explore alternative stellar continuum models in Appendix B, and interpret the outflows as residual feedback from a recent compaction event that quenched star formation.","tokens_in":26490,"tokens_out":8097,"duration_ms":90778,"significance":"If correct, the detection of vout ~ 1000 km/s ISM outflows in typical massive z > 1 PSBs is an important observational result. It would connect rapid quenching at high redshift to energetic feedback and strengthen the case that fast winds are a generic feature of the PSB phase, not only of rare luminous systems. The paper is commendable for including bootstrap uncertainties, an F-test, a boxcar cross-check, and an appendix with synthetic stellar libraries; these give the central detection a reasonable empirical foundation. However, because the result hinges on the decomposition of an unresolved doublet and on a calibration that assumes a specific outflow fraction and velocity distribution, the significance statement in the abstract is stronger than the systematic robustness demonstrated.","major_comments":[{"comment":"The systemic Mg ii component is modelled with a fixed doublet intensity ratio of 1.2:1, motivated by Weiner et al. (2009). At the stack's effective resolution (FWHM ~ 5.8 Å, doublet separation ~ 7.2 Å), the ratio and width jointly determine the blue-side profile of the systemic component. If the true systemic ratio is closer to the optically thin 2:1 value, the model's blue line is too shallow and part of the fitted 'outflow' could simply be the missing doublet strength; conversely a saturated ratio closer to 1:1 would work in the opposite direction. The paper does not fit or marginalize over the doublet ratio, nor does it report the maximum ratio change allowed by the red-side fit. I request a test with the ratio as a free parameter (or a grid spanning 1:1 to 2:1) in the one- and two-component models, reporting how Δv, the F-test p-value, and the ~ 25% blue excess change. Without this, the quoted > 3σ significance is conditional on a single assumed ratio.","section":"Section 3.2(i), Fig. 4"},{"comment":"The models used to test the stellar Mg ii contribution do not fully close the systematic. The authors state that the stellar Mg ii strength is 'essentially unconstrained' by the fits and that a weak feature blue-ward of Mg ii, important in F-star atmospheres, is also uncertain. Model B masks the blue side of the profile, so it cannot reveal whether the blue excess is partly stellar in origin; it simply forces all systemic absorption into the stellar template. Model C masks the whole doublet, but because the stellar Mg ii strength is degenerate with the systemic ISM component, the partition is not identifiable. I request a quantitative test in which an additional absorption feature is placed at the F-star line wavelength with a range of plausible strengths (scaled down from the synthetic libraries, since those lines are known to be too strong), and the maximum strength that still leaves an outflow component required at > 3σ is reported. The current Appendix B demonstrates consistency under alternative decompositions but does not bound the possible contamination.","section":"Appendix B"},{"comment":"The conversion from Δv to vout relies on simulations in which exactly 50% of input spectra contain an outflow and the outflow velocities are drawn uniformly from 0 to vmax. The 350 km/s offset is therefore a function of these assumptions, and the paper's D ~ 0.5 is itself inferred using a local covering fraction (Cf = 0.4–0.5) that may not hold for z > 1 PSBs. The authors should show the sensitivity of the calibration curve and of vout to D = 0.3 and 0.7 and to non-uniform velocity distributions (e.g., Gaussian or centrally concentrated). In addition, the boxcar estimate (809 ± 104 km/s) and the calibrated decomposition estimate (1150 ± 160 km/s) differ by ~ 340 km/s; the paper should state explicitly whether this difference is within the systematic uncertainty of the calibration and discuss the implications for the claim that vout exceeds the escape velocity (~ 950 km/s).","section":"Section 3.2 / Appendix A"}],"minor_comments":[{"comment":"The sentence 'This F-test yields a p-value for accepting the null hypothesis ... and rejects the two-component model if p > 0.05' is confusing; p is not the probability of the null, and the decision rule should be phrased as 'we include the outflow component when p < 0.05'.","section":"Section 3.2"},{"comment":"The phrase 'clear evidence' is too strong given the systematic caveats stated in Appendix B; consider 'strong evidence under our assumed doublet model' or equivalent.","section":"Abstract / Section 4"},{"comment":"For Models B and C, the quoted Δv uncertainties (~ 400 km/s) are much larger than Model A's (~ 143 km/s); this should be noted in the text so that the reader sees the loss of constraining power.","section":"Figure B1"},{"comment":"The Mg ii symbol appears as 'Mg /i.sc/i.sc' throughout the text due to a LaTeX conversion issue; this should be fixed.","section":"Throughout"}],"recommendation":"major_revision","confidential_remarks":"The paper is likely acceptable after the requested tests. The main claim may survive, but the current wording in the abstract overstates the robustness. I see no evidence of duplication or scope problems; this fits MNRAS well."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThe headline is simple: this is the first stacking detection of high-velocity outflows in spectroscopically confirmed post-starburst galaxies at z>1, and it is probably right. If you work on quenching and the build-up of the red sequence, this is worth a careful read.\n\nWhat the paper does well is mostly method. Maltby et al. take ~40 PSBs from the UDS, stack their spectra, and model the MgII lambda2800 absorption. They find a blue excess that requires a second component moving at Delta v~1500 km/s, and after a simulation-based calibration they quote v_out~1150+/-160 km/s. The significance is >3-sigma by an F-test. They cross-check with a boxcar method that gives a slightly lower but still high velocity (809+/-104 km/s), and they spend an entire appendix exploring the stellar contribution to MgII using synthetic libraries. The bootstrap uncertainties and the passive-galaxy control sample are handled carefully. The measurement of sigma*~200 km/s and the dynamical-mass argument add useful context.\n\nThe soft spots are real but not fatal. All of the outflow signal hangs on one unresolved stacked line, and the decomposition assumes a systemic doublet ratio of 1.2:1, fixed from the red side only. If the true ratio is closer to 1:1, or if the weak F-star feature blueward of MgII noted in Appendix B is present, part of the blue excess could be non-kinematic. The authors are honest that the stellar MgII strength is 'essentially unconstrained' by their fits. The boxcar method avoids the doublet-ratio assumption but is still sensitive to any extra blue absorption. So the 'clear evidence' in the abstract is a notch stronger than the data support. The 'typical' in massive PSBs also sits on a detection fraction D~0.5 that is inferred from an assumed covering fraction, not measured directly. And the sample selection removes AGN-emission PSBs by construction, so the wind could be starburst-driven or AGN-driven in a phase that no longer shows optical AGN signatures.\n\nOn balance, I think the detection is genuine. The consistency across three different modeling approaches (two-component, boxcar, stellar-template subtraction) is persuasive, and the passive galaxies show no asymmetry. The paper deserves a serious referee. I would recommend minor revision: soften the abstract, add a caveat about the doublet ratio and the F-star feature in the conclusions, and perhaps show the boxcar result more prominently. I would send it to peer review.\n\nBest.","headline":"First stacking detection of outflows in z>1 post-starbursts, probably real but with a model-dependent line decomposition that warrants a caveated abstract.","tokens_in":27337,"tokens_out":5802,"would_cite":true,"duration_ms":64624,"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":"Massive post-starburst galaxies at z > 1 carry ~1150 km/s winds in their interstellar medium, as seen in stacked Mg II absorption.","keywords":["post-starburst galaxies","galactic outflows","Mg II absorption","high redshift","galaxy quenching","stacking analysis","stellar velocity dispersion","UKIDSS Ultra Deep Survey"],"falsifier":"Take high-resolution ($R > 3000$) spectra of individual massive PSBs at $1 < z < 1.4$ that resolve the Mg II doublet. If the blue-shifted excess disappears once stellar templates (including the suspected weak F-star feature) are subtracted, or if no matching blue-shifted component appears in other low-ionisation tracers such as Fe II or Si II, the outflow interpretation would be refuted.","tokens_in":26036,"feed_emoji":"💨","tokens_out":11231,"duration_ms":102421,"temperature":0.7,"pith_summary":"This paper claims that massive post-starburst galaxies — systems that recently experienced a burst of star formation and then shut it off — at redshift 1 to 1.4 regularly host fast outflows in their interstellar medium. Stacking roughly forty optical spectra from the UDS survey, the authors find that about a quarter of the Mg II $\\lambda 2800$ Å absorption is shifted blueward of the systemic velocity, which a two-component model attributes to an outflow moving at roughly $v_{\\rm out}\\sim1150\\pm160\\rm\\,km\\,s^{-1}$. The same analysis finds no outflow in older passive galaxies at the same redshift, and the outflow speed exceeds the galaxies' typical escape velocity, so the wind can drive gas out of the system. These galaxies are compact and spheroidal, and the authors propose the winds were launched during a recent compaction event and represent the residual feedback that turned star formation off.","feed_headline":"Post-starburst galaxies at z>1 host 1150 km/s outflows","feed_subtitle":"Stacked Mg II spectra show a blue-shifted wind that may be the leftover feedback which shut down star formation.","key_machinery":"The central object is the stacked Mg II $\\lambda\\lambda 2796,2803$ Å absorption doublet, a tracer of low-ionisation interstellar gas. The argument is carried by a two-component spectral model in which the systemic absorption is a single Gaussian-convolved doublet with fixed intensity ratio $1.2:1$ and the outflow is a second doublet with a free centroid; an F-test decides whether the second component is statistically required, and a simulation calibrates the fitted velocity offset ($\\Delta v\\sim1500\\rm\\,km\\,s^{-1}$) down to a median outflow velocity ($v_{\\rm out}\\sim1150\\rm\\,km\\,s^{-1}$) by accounting for the fact that only a fraction of the galaxies contribute the outflow. A boxcar method based on equivalent widths of the blue and red sides provides an independent check.","core_discovery":"Massive ($M_\\ast > 10^{10}\\,M_\\odot$) post-starburst galaxies at $1 < z < 1.4$ show a statistically significant, strongly blue-shifted component in their stacked Mg II $\\lambda\\lambda 2796,2803$ Å absorption profile, which the authors interpret as high-velocity outflows in the interstellar medium with a typical velocity of $v_{\\rm out}\\sim1150\\pm160\\rm\\,km\\,s^{-1}$. The detection is significant at $>3\\sigma$ by an F-test and is recovered by two independent methods — a two-component Gaussian-convolved doublet fit and a boxcar equivalent-width measurement — while passive galaxies at the same redshifts show no such component. The galaxies also have a stellar velocity dispersion of $\\sigma_\\ast \\sim 200\\rm\\,km\\,s^{-1}$ (dynamical mass $M_{\\rm d}\\sim10^{11}\\,M_\\odot$) and are compact ($r_{\\rm e}\\sim1$--$2$ kpc) and spheroidal (Sérsic index $n\\sim3$), consistent with the idea that the outflows were launched during a recent compaction event such as a major merger or disc collapse.","pith_inferences":["If the inferred detection fraction D ~ 0.5 is correct, only about half of these PSBs are caught while their wind is still active; combining D with the D4000-based age distribution would give the first direct measurement of the wind duty cycle after quenching.","A monotonic decline of outflow velocity with D4000 in a larger sample would strengthen the case that the wind is the quenching agent rather than a longer-lived by-product; the current hint (younger PSBs, D4000 < 1.24, show more significant outflows) points this way but is not conclusive.","If the outflow is starburst-driven, its momentum flux should scale with the star-formation rate surface density of the preceding burst; stacking by SFR or by compactness (r_e, n) would test this and could distinguish starburst from AGN driving.","X-ray stacking of these PSBs could reveal hidden AGN that leave no optical trace; a correlation between X-ray luminosity and outflow velocity would implicate the AGN as the driver, whereas a null result would support the starburst-feedback scenario."],"forward_implications":["If the central claim holds, high-velocity outflows are typical, not exceptional, in massive post-starburst galaxies at z > 1, making the PSB phase a direct window onto quenching feedback.","Because the measured outflow velocity (~1150 km/s) exceeds the typical escape velocity (~950 km/s at 1 kpc), the outflowing gas can escape the galaxy or be driven into the circum-galactic medium, where it may suppress later gas accretion.","The absence of outflow signatures in passive galaxies, and the tentative trend with the D4000 index, imply the wind fades as the burst ages, coupling the outflow to the quenching event rather than to the passive phase.","The lack of optical AGN signatures means the wind is either powered by stellar feedback from the starburst itself, or by an AGN episode that has already faded by the post-starburst phase."],"supporting_citations":[{"why":"Supplies the 1.2:1 Mg II doublet intensity ratio used in the systemic component and the z~1.4 star-forming comparison sample with absorption depth Ad~0.55.","marker":"Weiner et al. 2009"},{"why":"Provides the z~0.6 luminous PSB outflows (>1000 km/s) that this work connects to, suggesting a common quenching mechanism.","marker":"Tremonti et al. 2007"},{"why":"Defines the boxcar method (mean outflow velocity from equivalent widths) used as an independent check on the two-component fit.","marker":"Rubin et al. 2010"},{"why":"Provides the standard application of the boxcar method and typical outflow velocities in star-forming galaxies used for comparison.","marker":"Bordoloi et al. 2014"},{"why":"Supplies the penalized pixel-fitting (ppxf) method used for full spectral fitting to measure stellar velocity dispersion and model the stellar continuum.","marker":"Cappellari & Emsellem 2004"},{"why":"Provides the UVBLUE synthetic stellar library used in Appendix B to test whether a stellar component can explain the blue-shifted Mg II absorption.","marker":"Rodríguez-Merino et al. 2005"},{"why":"Establishes the spectroscopic sample of high-z PSBs in the UDS that this study extends and stacks.","marker":"Maltby et al. 2016"},{"why":"Shows a large fraction of massive galaxies at z>1 pass through a PSB phase, motivating why these outflows matter for quenching.","marker":"Wild et al. 2016"},{"why":"Gives the typical covering fraction of local starburst outflows (Cf~0.4-0.5) used to estimate the detection fraction D~0.5 in the PSB stack.","marker":"Rupke et al. 2005"},{"why":"Provides the virial coefficient kd(n) used to convert sigma* and r_e into the dynamical masses that set the escape velocity comparison.","marker":"Bertin et al. 2002"}],"fun_headline_variants":["Dead galaxies at z>1 still exhale 1150 km/s winds","Massive post-starbursts at z>1 blow 1150 km/s outflows","Quenched galaxies host leftover 1150 km/s winds","Post-starburst winds hint at quenching feedback at z>1","Compact, dead, and windy: z>1 galaxies show 1150 km/s flows"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The claim collapses if the blue-sided asymmetry in the stacked spectrum can be produced by something other than outflowing gas, such as an unrecognized stellar absorption feature or a different ratio of the two Mg II lines.","fun_headline_variants_meta":{"raw":{"variants":["Dead galaxies at z>1 still exhale 1150 km/s winds","Massive post-starbursts at z>1 blow 1150 km/s outflows","Quenched galaxies host leftover 1150 km/s winds","Post-starburst winds hint at quenching feedback at z>1","Compact, dead, and windy: z>1 galaxies show 1150 km/s flows"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000624,"raw_usage":{"total_tokens":3024,"prompt_tokens":1214,"completion_tokens":1810,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":830,"completion_tokens_details":{"reasoning_tokens":1708}},"tokens_in":830,"tokens_out":1810,"duration_ms":14158,"temperature":1.0,"reasoning_tokens":1708,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T14:34:51.580663+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take high-resolution ($R > 3000$) spectra of individual massive PSBs at $1 < z < 1.4$ that resolve the Mg II doublet. If the blue-shifted excess disappears once stellar templates (including the suspected weak F-star feature) are subtracted, or if no matching blue-shifted component appears in other low-ionisation tracers such as Fe II or Si II, the outflow interpretation would be refuted.","supporting_citations":[],"review_version":1}