{"id":"99687440-7760-420b-9b64-9b7bbaef6463","arxiv_id":"1908.06997","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Interferometric velocity maps of R Dor show a nearly static atmosphere below 1.5 stellar radii and outward motion of 7 to 15 km/s in CO layers near 1.8 stellar radii.","lead":"Using the VLTI/AMBER interferometer, astronomers mapped gas motions across the surface and atmosphere of the nearby dying star R Dor and found its outer layers streaming outward at 7 to 15 km/s while deeper layers stay nearly still. This is the first spatially resolved look at where the wind of an old, Sun-like star may be accelerated.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The claimed acceleration between 1.5 and 1.8 R* rests on assumed stratified line-formation heights for Mg, H2O, and CO; a radiative-transfer and moving-atmosphere simulation is needed to validate both the heights and the velocity extraction.","rationale":"The paper's velocity maps are likely robust: the CO blueshifts are visually evident in the spatially resolved spectra (Figure 6), and the spread across 18 reconstruction set-ups is only 1 km/s (Appendix B). The fragile step is the conversion of image extensions into line-formation heights and the assumption that Mg, H2O, and CO trace a single stratified flow. The paper explicitly assigns heights from reconstructed image extents and literature, and it states the CO lines are optically thick, so the observed emission at any pixel is a column average. H2O and CO also probe different chemical species whose spatial distributions can differ, so the velocity contrast could reflect different gas components rather than radial acceleration. The paper's own discussion lists alternative explanations (ballistic motion, pulsation) and cites ALMA observations showing no outflow at a different epoch. These caveats are appropriate, but without a radiative-transfer model of R Dor showing the claimed formation radii and a moving-atmosphere simulation of the velocity-extraction pipeline, the acceleration inference remains conditional. The reader's CONDITIONAL verdict is therefore appropriate: the maps are valuable and the acceleration is plausible, but it is not established beyond the stated assumptions. Hence UNCHANGED.","tokens_in":23975,"tokens_out":8882,"duration_ms":97233,"concrete_test":"Construct a dynamical model atmosphere of R Dor (using the parameters of Table 1 and a prescribed accelerating velocity profile), compute the line-formation contribution functions for the Mg 2.28164, H2O 2.28478, and CO 2.30150 um lines, and generate synthetic AMBER observables at the observed uv coverage. Then run the same image-reconstruction and cross-correlation pipeline on the simulated data to verify that (a) the peak formation radii are below 1.13 R*, about 1.5 R*, and about 1.8 R*, respectively, and (b) the pipeline recovers the input layer-to-layer velocity difference without sign or amplitude bias. If either condition fails, the acceleration claim is unsupported.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's central physical claim, that material accelerates between approximately 1.5 and 1.8 R*, depends on the assumption that the Mg, H2O, and CO lines form in stratified shells at the heights inferred from reconstructed image extensions (Section 4.1: Mg <1.13 R*, H2O about 1.5 R*, CO about 1.8 R*). These heights are not derived from a radiative-transfer model of R Dor but are assigned from the apparent size of the line images and literature arguments. Because the CO lines are optically thick to the extreme limb (Section 4.3), the spatially resolved CO spectrum at any sky position is an opacity-weighted average over a range of radii along the line of sight, not a sample of gas at a single geometric height. Furthermore, H2O and CO are different molecules with different excitation and opacity conditions; their velocity maps could trace spatially distinct gas components or clumps rather than a single accelerating outflow. The paper itself notes that the outward motion may be intermittent or ballistic (Section 5), and ALMA observations at a different epoch show no outward motion within 4 R* (Vlemmings et al. 2018). Without a model confirming both the layer-to-layer stratification and that the same flow is traced at each height, the acceleration inference is not established.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents VLTI/AMBER spectro-interferometric observations of the AGB star R Dor, reconstructs images at 309 wavelength channels, and derives line-of-sight velocity maps from Mg, H2O, and CO lines. The authors report that the Mg and H2O lines, formed below ~1.5 R*, show little systematic motion, while the CO first-overtone lines show outward motion of 7-15 km/s at ~1.8 R*. They interpret this as strong acceleration of material between ~1.5 and ~1.8 R*, possibly driven by radiation pressure on dust, while acknowledging that the motion could be intermittent or ballistic. This would be the first spatially resolved, multi-height velocity diagnosis of an AGB star.","tokens_in":24225,"tokens_out":3296,"duration_ms":33631,"significance":"If the acceleration claim holds, the paper provides a qualitatively new constraint on the wind-acceleration region of an AGB star, directly complementing dust-formation radii inferred from polarimetry and the lack of outflow seen in ALMA at a later epoch. The observational work is careful: the uv coverage is good, image reconstruction is tested against simulated data with 18 different reconstruction setups, uncertainty maps are derived from the reconstruction scatter plus calibration terms, and the interferometric data are made public in OIFITS format. The main weakness is that the layer-to-layer velocity comparison rests on assumed line-formation heights and on the assumption that the Mg, H2O, and CO lines trace the same accelerating flow; these assumptions are not validated with a radiative-transfer or moving-atmosphere model.","major_comments":[{"comment":"The assignment of line-formation heights (Mg ≤1.13 R*, H2O ~1.5 R*, CO ~1.8 R*) is based on the spatial extension of the reconstructed images and on the extent of a hydrostatic MARCS model, not on a radiative-transfer model of R Dor. Because the paper itself states in Section 4.3 that the CO lines are optically thick to the extreme limb, the spatially resolved CO spectrum at a given sky position is an opacity-weighted average over a range of radii along the line of sight, not a sample of gas at a single geometric height. Consequently, the layer-to-layer comparison of velocities does not by itself establish acceleration of a single flow between ~1.5 and ~1.8 R*; the Mg, H2O, and CO lines could trace distinct gas components or a broad height range. This is load-bearing for the central claim and requires either a radiative-transfer model of R Dor or a conservative rephrasing of the conclusion.","section":"Sections 4.1 and 4.3"},{"comment":"The simulated-data tests validate the image reconstruction for a static, known source, but they do not validate the velocity-extraction procedure. The line-of-sight velocities in Section 4.3 are derived by cross-correlating spatially resolved spectra, including off-limb emission spectra, with a hydrostatic MARCS absorption spectrum. No test demonstrates that this cross-correlation recovers a known input velocity field for emission lines or for optically thick CO lines. I request such a test, for example using synthetic spectra from a moving model atmosphere with a known velocity law, to confirm that the measured 7-15 km/s blueshifts are not biased by the choice of the reference spectrum or by the absorption-to-emission transition across the limb.","section":"Appendix B"},{"comment":"The ALMA observations of Vlemmings et al. (2018), cited by the authors, show no outward motion within 4 R* at a different epoch, and the paper itself notes that the outward motion may be intermittent or ballistic. The abstract nevertheless presents a 'strong acceleration of material between ~1.5 and 1.8 R*' as a main result. Given the height-assignment issue raised above and the epoch discrepancy, the abstract and conclusion should more prominently qualify the acceleration claim as one possible interpretation rather than an established measurement.","section":"Section 5 and Abstract"}],"minor_comments":[{"comment":"There are two typos: 'Adoped' in Table 1 should be 'Adopted', and the phrase 'limb-darkened disk diameter xsand distance' in Section 3.1 should read 'limb-darkened disk diameter and distance'.","section":"Section 3.1 and Table 1"},{"comment":"The uncertainty maps in Figure 7d-7f show the total 1σ including the reconstruction scatter and calibration terms, but the color bars are labeled only with the standard deviation; it would help to state explicitly in the caption that these panels include the systematic terms quoted in the text.","section":"Figures 7d-7f"},{"comment":"In the caption of Figure 5, the reference to 'Figure 4f' in the sentence about blueshifts at position 4 appears to be a typo; the relevant panel is Figure 5f, not Figure 4f.","section":"Section 4.2, Figure 5 caption"}],"recommendation":"major_revision","confidential_remarks":"The paper reports a novel and potentially important dataset, and the data treatment is thoughtful. The central issue is that the acceleration claim depends on assumed line-formation heights and on the assumption that the same flow is traced at all three heights; the requested radiative-transfer or moving-atmosphere validation is feasible and would considerably strengthen the paper. I therefore recommend major revision rather than rejection."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague —\n\nThis is a careful, well-caveated observational paper that gives the first spatially resolved, multi-height velocity maps for an AGB star. The authors take a method they developed for Antares and apply it to R Dor with Mg, H2O, and CO lines. The CO maps reveal large-scale blueshift of 7–15 km/s at about 1.8 R*, while Mg and H2O show little systematic motion below 1.5 R*. If that acceleration gradient holds, it is a genuine new result and the first direct look at wind acceleration in an AGB star's outer atmosphere.\n\nWhat earns credit: the uv coverage is good, the image reconstruction is tested on simulated data with 18 different regularization setups, the uncertainty maps are explicit, and the OIFITS files are deposited. They compare reconstructed images to the observed visibilities and phases. The paper also says plainly that the outward motion could be ballistic or intermittent, and it cites the ALMA result that shows no outflow within 4 R* at a different epoch. That is honest handling of the evidence.\n\nThe soft spots are real but not fatal. The layer-to-layer acceleration inference depends on line-formation heights that are assigned from image extension and literature, not from a radiative transfer model of R Dor. CO is optically thick to the extreme limb, so each sky position samples a range of radii; Mg and H2O may trace different gas components. Also, the velocity extraction for off-limb emission lines uses cross-correlation with a hydrostatic template and the sign handling is not fully documented. A simulated moving-atmosphere test would remove a lot of doubt about the velocity pipeline. I would not call the central result wrong, but I would not call the acceleration demonstration complete either.\n\nBottom line: this paper deserves serious refereeing. For stellar astrophysics, it is a significant measurement and the data are reproducible. The referee should ask for either a moving-atmosphere simulation or a careful sensitivity analysis of the assumed formation heights before the wind-acceleration interpretation is taken as established. I'd bring it to reading group and cite it once the caveats are stated.","headline":"First multi-height velocity maps for an AGB star; the measurement is likely right, but the acceleration claim outruns the line-formation assumptions.","tokens_in":24772,"tokens_out":1887,"would_cite":true,"duration_ms":20196,"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":"This paper maps gas velocities at three atmospheric heights of the AGB star R Dor, finding deep layers nearly still while the outer CO layer streams outward at 7–15 km/s, the first resolved view of where an AGB wind accelerates.","keywords":["AGB stars","stellar mass loss","infrared interferometry","VLTI/AMBER","velocity-field maps","R Dor","dust-driven stellar winds","stellar atmospheres"],"falsifier":"A velocity-resolved radiative-transfer model of R Dor's 2.3 μm spectrum in which the CO lines form over a broad range of heights reaching into the near-static photosphere would break the height stratification: if such a model reproduces the observed 7–15 km/s blueshifts with no accelerating shell at 1.8 R*, the acceleration claim collapses. Observationally, an epoch of the same interferometric velocity-mapping at a different pulsation phase that shows the outward motion absent while dust at 1.5 R* remains present would demonstrate the motion is intermittent ballistic motion rather than steady wind acceleration.","tokens_in":2385,"feed_emoji":"⭐","tokens_out":2870,"duration_ms":126730,"temperature":0.7,"pith_summary":"This paper tries to establish where and how the stellar wind of an asymptotic giant branch (AGB) star is launched, using the closest such star, R Dor, as the test case. The authors combine very high spectral resolution (8000) with 6.8 milliarcsecond spatial resolution — seven times finer than the star's 51.2 mas disk — to reconstruct images at 309 wavelengths and extract a velocity map over both the stellar surface and the extended atmosphere. Magnesium lines from the deep layers (below about 1.13 stellar radii) and water lines from about 1.5 stellar radii show no motion beyond the 1.7 km/s measurement uncertainty, while the carbon monoxide first-overtone lines from about 1.8 stellar radii show systematic outward motion of 7–15 km/s over a large fraction of the star. The juxtaposition implies strong acceleration of material between about 1.5 and 1.8 stellar radii, the region where dust is known to form, so the authors propose radiation pressure on dust grains as the driver while leaving intermittent ballistic motion from convection or pulsation as an open alternative. If correct, this is the first three-dimensional (two spatial dimensions plus line-of-sight velocity) dynamical picture of an AGB star's atmosphere.","feed_headline":"R Dor's gas accelerates between 1.5 and 1.8 stellar radii","feed_subtitle":"Interferometric velocity maps reveal a 7-15 km/s outflow in the outer atmosphere while deeper layers stand still.","key_machinery":"The load-bearing device is a height ladder of spectral lines observed simultaneously with VLTI/AMBER between 2.278 and 2.308 μm at spectral resolution 12,000 (binned to 8,000): the Mg line forms below ~1.13 R*, the H2O lines near ~1.5 R*, and the CO first-overtone lines out to ~1.8 R*, so each line tags a different geometrical height in the same snapshot. At each of the 309 wavelength channels, images are reconstructed with the MiRA algorithm using Fourier phases restored from differential-phase measurements and are convolved to the 6.8 mas beam; spatially resolved spectra are then cross-correlated against a hydrostatic MARCS synthetic spectrum to assign a line-of-sight velocity to every position. Comparing the three velocity-field maps turns the line-formation height into a radius-velocity measurement, converting unresolved Doppler information into a resolved acceleration profile.","core_discovery":"The paper's central claim is that R Dor's atmosphere has a sharp kinematic transition with height: layers probed by the Mg line at 2.28164 μm (below ~1.13 R*) and by three H2O lines near 2.28 μm (~1.5 R*) are quiet to within the 1.7 km/s measurement uncertainty, whereas the CO first-overtone lines reveal blueshifted, outward motion at 7–15 km/s over a substantial fraction of the surface and out to ~1.8 R*. Because dust is detected at ~1.5 R*, the authors interpret the velocity jump as strong acceleration in the shell between 1.5 and 1.8 stellar radii, most plausibly caused by radiation pressure on dust grains — though they stress that ballistic motion driven by convection or pulsation, and hence intermittency, cannot yet be excluded. The paper also presents the first resolved images of the extended atmosphere of a non-Mira AGB star and finds a bright surface region with about 25% intensity contrast, qualitatively consistent with three-dimensional convection simulations.","pith_inferences":["If dust radiation pressure drives the acceleration, the fastest outward motion should be spatially matched to the clumpy dust clouds seen in polarimetric imaging; overlaying the CO velocity map on the dust-scattering maps would test this, with the southern 15 km/s region as the natural first target.","Filling in the ladder with lines forming near 1.6 R* or including fundamental CO bands could distinguish a sharp acceleration kick between 1.5 and 1.8 R* from a gradual outward rise in velocity — a testable extension of the same method.","Because the unresolved spectrum hides the outflow, some AGB stars previously classified as having static outer atmospheres on spectroscopic evidence may harbour undetected accelerating layers; a small survey of nearby AGB stars with the same technique would measure how common such hidden outflows are.","Applying the same three-height velocity-mapping to stars spanning different pulsation amplitudes and mass-loss rates would reveal whether dust-driven acceleration is universal or one of several wind-launching modes in evolved cool stars."],"forward_implications":["If the acceleration is real, the wind-launch zone of R Dor sits between about 1.5 and 1.8 stellar radii, coinciding with the dust-formation radius, which supports dust-driven wind models in which radiation pressure — possibly scattering on composite grains — accelerates the outflow.","The measured outward velocities of 7–15 km/s exceed the wind's terminal velocity of about 5.5 km/s reached beyond 20 R*, so the same accelerating flow can explain the high-velocity components seen in submillimeter molecular lines.","Spatially unresolved spectroscopy cannot see this outflow at all, because blueshifted CO absorption over the disk is filled in by blueshifted CO emission outside the limb; spatially resolved spectro-interferometry is therefore necessary to diagnose wind acceleration in such stars.","The dynamics of R Dor's extended atmosphere (systematic outflow) differ from those of the red supergiant Antares (turbulent clump motion, no systematic outflow), a first observational hint that wind-launching may differ between AGB stars and red supergiants.","Since the hydrostatic photosphere extends only to about 1.24 R*, the CO and H2O atmospheres out to 1.5–1.8 R* require a non-hydrostatic mechanism, and finding one in a low-amplitude semiregular variable widens the class of stars that can sustain extended dynamic atmospheres."],"supporting_citations":[{"why":"the Antares velocity-field map this work extends to three dimensions and to an AGB star","marker":"Ohnaka et al. 2017b"},{"why":"detected ~0.3 μm transparent grains at ~1.5 R* in R Dor, fixing the dust formation radius","marker":"Norris et al. 2012"},{"why":"polarimetric imaging locating clumpy dust formation at ~1.5 R* toward R Dor, the basis of the dust-pressure interpretation","marker":"Khouri et al. 2016"},{"why":"3D convection-plus-dust models predicting Al2O3 cores near 1.4 R* and silicate mantles at 1.7–1.8 R*, matching the H2O and CO heights","marker":"Höfner & Freytag 2019"},{"why":"adopted systemic velocity (7.5 km/s) and the ~5.5 km/s terminal wind velocity the measured 7–15 km/s is compared with","marker":"Van de Sande et al. 2018"},{"why":"3D convection simulations showing ballistic gas motion near 2 R*, the alternative explanation the paper must exclude","marker":"Freytag et al. 2017"},{"why":"the MARCS model atmosphere grid used to compute the redshift-free reference spectrum for velocity cross-correlation","marker":"Gustafsson et al. 2008"},{"why":"the technique of restoring Fourier phases from wavelength-differential phases, essential to the image reconstruction","marker":"Ohnaka et al. 2011"},{"why":"ALMA imaging finding rotation but no outward motion within 100 mas (4 R*), the evidence that the outflow may be intermittent","marker":"Vlemmings et al. 2018"}],"fun_headline_variants":["R Dor: gas abruptly accelerates from 1.5 to 1.8 stellar radii","Dust-driven wind? R Dor's gas jumps to 7-15 km/s at 1.8 R*","First resolved 3D velocity map of an AGB star shows wind kick","R Dor's atmosphere: inner layers static, outer shell races outward"],"cache_read_input_tokens":26880,"weakest_assumption_plain":"The argument assumes the Mg, H2O, and CO lines form in thin, nested shells at the claimed heights (below about 1.13, about 1.5, and about 1.8 stellar radii) and that the three velocity maps trace the same outward-accelerating gas; those heights come from the spatial extent of the reconstructed images and earlier literature, not from a radiative-transfer model of R Dor's own atmosphere.","fun_headline_variants_meta":{"raw":{"variants":["R Dor: gas abruptly accelerates from 1.5 to 1.8 stellar radii","Dust-driven wind? R Dor's gas jumps to 7-15 km/s at 1.8 R*","First resolved 3D velocity map of an AGB star shows wind kick","R Dor's atmosphere: inner layers static, outer shell races outward"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000246,"raw_usage":{"total_tokens":1605,"prompt_tokens":1078,"completion_tokens":527,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":694,"completion_tokens_details":{"reasoning_tokens":434}},"tokens_in":694,"tokens_out":527,"duration_ms":5374,"temperature":1.0,"reasoning_tokens":434,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T12:30:23.067582+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A velocity-resolved radiative-transfer model of R Dor's 2.3 μm spectrum in which the CO lines form over a broad range of heights reaching into the near-static photosphere would break the height stratification: if such a model reproduces the observed 7–15 km/s blueshifts with no accelerating shell at 1.8 R*, the acceleration claim collapses. Observationally, an epoch of the same interferometric velocity-mapping at a different pulsation phase that shows the outward motion absent while dust at 1.5 R* remains present would demonstrate the motion is intermittent ballistic motion rather than steady wind acceleration.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"detected ~0.3 μm transparent grains at ~1.5 R* in R Dor, fixing the dust formation radius"},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"polarimetric imaging locating clumpy dust formation at ~1.5 R* toward R Dor, the basis of the dust-pressure interpretation"},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"adopted systemic velocity (7.5 km/s) and the ~5.5 km/s terminal wind velocity the measured 7–15 km/s is compared with"},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"3D convection simulations showing ballistic gas motion near 2 R*, the alternative explanation the paper must exclude"},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"the technique of restoring Fourier phases from wavelength-differential phases, essential to the image reconstruction"},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"ALMA imaging finding rotation but no outward motion within 100 mas (4 R*), the evidence that the outflow may be intermittent"}],"review_version":1}