{"id":"eb150f73-49fe-4c96-ad4e-3838af9b39f2","arxiv_id":"2505.03158","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"FAST observations reveal an extended HI halo around the edge-on galaxy NGC2683, containing roughly 13% of its neutral hydrogen and likely fed by accreting gas from dwarf companions and clouds.","lead":"Using the FAST radio telescope, astronomers mapped neutral hydrogen around the edge-on galaxy NGC2683 and found a much taller, fainter gas layer than previous VLA maps showed, along with gas clouds that appear to be falling in. The result suggests the galaxy is accreting gas from its surroundings, providing a new case study in how galaxies gather fuel.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Halo mass and 4x vertical extent may largely inherit VLA missing short spacings and single-dish artifacts; the paper's own excess map is its only quantitative check.","rationale":"The paper's direct detection of extended HI toward NGC 2683 is plausible and independently supported by the channel maps and the consistent small residuals after the halo model subtraction. However, the specific quantitative claims—the 4x vertical extent, the 3e8 M_sun halo mass, and the 13% fraction—depend on measuring a faint diffuse component in a single-dish image and on subtracting a reprocessed VLA map. The paper itself concedes that negative residual regions exist and attributes them to pointing uncertainties without quantifying the effect. It also states that the excess is predominantly ~3e19 cm^-2, only a factor ~10 above the FAST sensitivity and a factor ~2 above the VLA sensitivity, so small systematic errors in either map can create or remove the apparent halo. The reader's weakest assumption captures this correctly, and the requested checks (uncertainty analysis, null tests, explicit systematic-error accounting) are the appropriate remedy. These are not fatal objections; they are the conditions under which the halo claim can be accepted. Hence a CONDITIONAL verdict, unchanged from the reader, with the condition being a quantitative systematic-error analysis.","tokens_in":18713,"tokens_out":1713,"duration_ms":16617,"concrete_test":"Recompute the FAST-VLA excess map after (a) masking the VLA image to its valid uv-support and applying the same 3x3 smoothing, and (b) adding an explicit null test: shift the VLA map by ±12 arcsec in RA/Dec and compute the resulting FAST-VLA residuals. If the positive vertical excess persists at >3 sigma after these shifts and after excluding VLA regions where the VLA map is below its own 1-sigma noise, the excess is not a pointing or short-spacing artifact. Also re-derive the halo mass using a 3-sigma blanking threshold rather than a two-times-rms mask and report the resulting mass range.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The central halo claim has the following form: FAST sees faint, vertically extended HI at roughly 3e19 cm^-2 that VLA (especially after uv-tapering to a 3.1'x2.8' beam) does not. The FAST excess is then attributed to a halo component in the Vollmer model, with chi=0.1 and a 13% halo mass fraction. The load-bearing operation is subtraction of the reprocessed VLA from FAST, but the paper's only quantitative checks are (i) a few 'negative regions likely attributable to pointing uncertainties' and (ii) the statement that the excess column density is 'predominantly distributed around ~3e19 cm^-2.' These are not quantitative or null-tested. Specifically, the excess map is shown after unrestricted subtraction without filtering the VLA map; if the VLA map has residual negative bowls from missing short spacings, then FAST-(VLA) will show positive residuals at large radii regardless of any true halo. Similarly, single-dish baseline errors, stray radiation, or a 12 arcsec pointing offset against a 2.9' beam can produce broad, faint, apparently extended residuals. Because the halo mass (3e8 M_sun, 13% of total HI) is measured by applying a mask built from the Vollmer disk+flare model to the FAST cube, if the mask is inaccurate or if the FAST data contain baseline/stray contamination at the 1e19 cm^-2 level, the inferred mass and the inferred 4x vertical extent are both unquantified. The strongest claim therefore rests on an under-tested subtraction.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents FAST 19-beam HI observations of the edge-on galaxy NGC 2683 and its surroundings. Compared with the VLA data of Vollmer et al. (2016), the FAST maps reach about four times larger vertical extent and a slightly higher total flux (112.1 Jy km/s, M_HI = 2.32e9 Msun). The authors also report the detection of the dwarf galaxies KK69, dw1, and dw3?, and three compact high-velocity complexes A, B, and C. Interpreting the diffuse vertical excess as an HI halo, they add a halo component to the Vollmer et al. flare model, with normalization chi = 0.1 tuned to the FAST data, and claim the halo model matches the FAST data better than the pure flare model. The derived halo mass is about 3e8 Msun, roughly 13% of the total HI mass, which they attribute to external gas accretion from the dwarf companions and HVCs.","tokens_in":19003,"tokens_out":3537,"duration_ms":36069,"significance":"If the extended faint HI distribution around NGC 2683 is real, the paper provides a rare single-dish detection of a candidate neutral gas halo in an edge-on spiral, with a mass fraction comparable to HALOGAS extraplanar gas fractions. The detection of the dwarf galaxy dw3? and the HI tail of KK69, along with the three velocity complexes, adds useful information on gas accretion in a low-mass group. The paper also demonstrates the sensitivity of FAST for faint extended HI emission, which is a constructive result. The direct reprocessing of the VLA data to a matched beam is a step in the right direction. However, as explained in the major comments, the central 'halo not flare' conclusion and the quoted halo mass rest on model normalization, subtraction, and masking choices that are not yet quantitatively tested.","major_comments":[{"comment":"The comparison between the V-model-flare and V-model+halo is not a fair test of the presence of a halo. The halo model is normalized to the FAST data via the parameter chi, which is explicitly 'derived from the comparison of the flux obtained from the model with the FAST observations' (chi = 0.1), whereas the V-model-flare is the original Vollmer et al. model not renormalized to FAST. A model fitted to the data is naturally expected to yield smaller residuals. To support the 'halo not flare' claim, the authors should either fix chi a priori (e.g., from the expected halo mass fraction or from fitting the halo model simultaneously to both VLA and FAST data) or present a comparison where both models are normalized in the same way.","section":"Sec. 4.2, Figs. 8, 10, 11"},{"comment":"The central evidence for the extended halo is the FAST minus VLA excess map, but the paper provides no quantitative uncertainty analysis or null tests for this subtraction. The few negative regions are dismissed as 'likely attributable to pointing uncertainties' without a quantitative estimate of the magnitude of the pointing offset (the pointing accuracy is ~12 arcsec against a 2.9' beam) or of the effect of baseline subtraction and stray radiation. Single-dish baseline errors and residual negative bowls from the VLA's missing short spacings can produce broad, faint positive residuals at the level of ~3e19 cm-2 claimed here. The authors should quantify these systematic effects, e.g., by jackknifing the data, by comparing two independent halves of the FAST observations, or by injecting and recovering a model disk into the FAST data to show that the excess is not an artifact.","section":"Sec. 3, Fig. 3"},{"comment":"The halo mass of 3e8 Msun is derived by masking the FAST data with a model cube built from the VLA-based disk+flare model, using an arbitrary threshold of twice the rms noise. This mass therefore inherits the model's vertical structure; if the flare law is even slightly misestimated, the mask will misclassify disk or flare emission as halo, or vice versa. In addition, the scale-height power law used in the halo model is fitted to VLA data (Figure 9) that, by the authors' own statement, do not show the extended halo, creating a circularity: the halo's vertical structure is derived from the same data that supposedly lack the halo. The paper gives no error bars, no variation of the threshold, and no test of the sensitivity of the derived mass to the assumed disk/flare subtraction. A direct measurement of the halo column density and a proper propagation of uncertainties are needed.","section":"Sec. 4.2, halo mass and scale-height fit"},{"comment":"The claim that FAST detects HI about four times farther than the VLA in the vertical direction is not an apples-to-apples comparison. The two data sets have very different sensitivities: the reprocessed VLA has 5-sigma column-density sensitivity of 1.4e19 cm-2, while FAST reaches 3.7e18 cm-2. A more sensitive observation will naturally detect more extended faint emission. The paper should either match the sensitivity and then compare the extent, or model what vertical extent the VLA would have detected if the same extended HI distribution were present. Without this, the factor-of-four extent is not a robust property of the galaxy.","section":"Sec. 3, Fig. 2 and vertical extent statement"}],"minor_comments":[{"comment":"\"We not only detected the H i distribution of three dwarf galaxies\" is ungrammatical; it should read \"We not only detected the H i distribution of three dwarf galaxies, but also discovered three HI complexes.\"","section":"Sec. 1"},{"comment":"The sentence \"The antenna temperature was convert to flux density\" is a typo, should be \"was converted\".","section":"Sec. 2"},{"comment":"The 5-sigma column-density sensitivity for NGC 2683 is listed as 3.7e18 cm-2, while in Sec. 3 it is also quoted as 5-sigma = 3.7e18 cm-2; please ensure the factor-5 convention is consistent throughout the text and table.","section":"Sec. 2, Table 1"},{"comment":"The contour list in the caption of Figure 4 is long and appears to include values not visible in the panel; consider trimming the description or referencing the velocity ranges in the text.","section":"Fig. 4 caption"},{"comment":"The symbol 'chi' should be defined as a dimensionless normalization factor in the text before Eq. (1). Also, 'we adopted chi = 0.1' would benefit from an explicit statement of the range of chi that gives a reasonable fit, if explored.","section":"Sec. 4.2"},{"comment":"In conclusion (2), 'We also determined the mass of the three complex clouds by assuming them as the HVCs' is awkward; 'by assuming them to be HVCs' would be clearer.","section":"Sec. 5"}],"recommendation":"major_revision","confidential_remarks":"The paper reports a potentially interesting detection, but the main claim of an HI halo and the halo mass are not yet robust to the model-normalization and subtraction issues detailed in the major comments. I do not see evidence of intentional misrepresentation; the concerns are methodological. The paper would benefit from a more careful systematic-error analysis and a model comparison that treats both models symmetrically. Given the current state, I would not accept the paper in its present form, but with substantial revisions the central result may become publishable."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Here's the short version: this is a genuinely new FAST dataset and the extended, faint HI around NGC 2683 is worth knowing about. The paper's halo mass and its 'halo, not flare' conclusion, though, are not yet backed by the same standard of evidence as the raw detection.\n\nWhat's new: this is the first FAST HI mapping of NGC 2683 and its surroundings, with column-density sensitivity about five times better than the earlier VLA work. The authors detect a vertically extended component roughly four times larger than the VLA's, along with dwarf galaxy detections and high-velocity clouds. The sensitivity gain and the reprocessing of VLA data to the same resolution are both sensible, and the channel maps show plainly that the VLA does not see much of the faint emission. That direct observational comparison is the paper's strongest asset.\n\nWhere it gets softer: the interpretation depends on a subtraction/mask/model chain. The VLA reprocessing may retain negative bowls from missing short spacings, and single-dish FAST data can carry baseline or stray-radiation contamination. The only checks are a few negative residual patches attributed to pointing and a statement that the residuals sit around 3e19 cm^-2. That is not a real null test. The halo model uses Eq. 8 from Vollmer et al. 2016, a scale-height power law fitted to VLA data, and a normalization chi=0.1 chosen by matching the FAST observations. Since chi is fitted to the very data the model is judged against, the visual claim that 'V-model+halo matches better' is weaker than it looks. The 3e8 Msun halo mass, 13% of the total HI, is derived by masking the FAST cube with the disk+flare model, so it is not simply the fitted chi, but it inherits all the same assumptions about where the flare is and it carries no error bars. Missing are uncertainties on all central numbers, a quantified comparison against a model without a halo, and a test of whether baseline or stray-radiation errors can produce the apparent vertical extent.\n\nThe citation pattern is fair; the paper engages with Vollmer et al. 2016 and the HALOGAS sample. This deserves a serious referee, because the observation is new and the subfield needs more single-dish versus interferometer comparisons, but it will need major revision before the halo claim is credible. The referee should ask for quantified residual statistics, jackknife or baseline-error tests, and error bars on the halo fraction, or at least an honest statement of which assumptions dominate that number.","headline":"Genuinely new FAST data reveal a much more extended HI distribution around NGC 2683, but the halo mass and the 'halo not flare' conclusion outrun the quantitative evidence.","tokens_in":19581,"tokens_out":2416,"would_cite":true,"duration_ms":26011,"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":"Using high-sensitivity FAST observations, this paper claims that the edge-on galaxy NGC 2683 has a neutral hydrogen halo four times taller than earlier VLA maps did, holding about 300 million solar masses of gas.","keywords":["galaxy accretion","galaxy interactions","interstellar medium","neutral hydrogen clouds","HI halo","edge-on galaxies","dwarf galaxies","high-velocity clouds"],"falsifier":"Reprocess the FAST maps with an independent stray-radiation and baseline correction, or re-observe with an interferometric array that has complete short-spacing coverage, and check whether the vertical excess at column densities near $3\\times10^{19}\\,\\mathrm{cm^{-2}}$ persists. If the excess vanishes or is traced to sidelobes or Galactic HI confusion, the halo claim would collapse; if the short-spacing-complete map recovers the full $112.1\\,\\mathrm{Jy\\,km\\,s^{-1}}$ flux with only a flaring disk, no halo would be needed.","tokens_in":18481,"feed_emoji":"📡","tokens_out":10625,"duration_ms":100935,"temperature":0.7,"pith_summary":"The paper claims that the edge-on spiral galaxy NGC 2683 is surrounded by a thick neutral hydrogen halo that earlier interferometric observations missed. Using high-sensitivity mapping with FAST, it detects HI emission extending about four times farther from the galactic plane than the VLA saw, corresponding to a halo of roughly $3\\times 10^8$ solar masses, about 13% of the galaxy's atomic hydrogen. The vertical distribution cannot be explained by a flaring gas disk alone, so the authors add a halo component to the existing disk/flare model and find that it matches the FAST data. They interpret the halo as evidence of ongoing external gas accretion from dwarf companions such as KK 69 and infalling high-velocity clouds, rather than star-formation-driven galactic fountains. If correct, this would mean that single-dish HI surveys recover substantial extraplanar gas that interferometers systematically undercount.","feed_headline":"FAST finds an HI halo four times taller than the VLA saw","feed_subtitle":"About 13 percent of the galaxy's neutral hydrogen sits in a thick halo that earlier VLA maps missed.","key_machinery":"The central comparison is between FAST total-power data and VLA interferometric data reprocessed, tapered, and smoothed to a matched $3.1'$ beam. Because FAST measures total power, it retains the short-spacing information that the VLA misses, and its $5\\sigma$ column-density sensitivity of $3.7\\times10^{18}\\,\\mathrm{cm^{-2}}$ is roughly four times deeper than the reprocessed VLA limit. The model machinery is the earlier tilted thin-disk, flare, warp, and outer-ring model, augmented by an empirical vertical halo profile $\\rho_{\\rm halo}=\\chi\\rho_{\\rm disk}\\,\\sinh(z/z_0)/\\cosh^2(z/z_0)$, with scale heights fitted as a power law in radius from the VLA data and $\\chi=0.1$ set by matching the model flux to FAST. The decisive evidence is the set of residual channel maps and vertical flux profiles: the flare-only model cannot fill the vertical extent that FAST sees, while the flare-plus-halo model leaves residuals of order $0.5\\times10^{19}\\,\\mathrm{cm^{-2}}$.","core_discovery":"The paper claims that NGC 2683 hosts an extended neutral hydrogen halo that earlier VLA observations did not reveal. FAST detects HI emission out to about 18.7 arcminutes perpendicular to the disk, roughly four times the vertical extent of the reprocessed VLA map, at a total flux $F_{\\rm HI}=112.1\\,\\mathrm{Jy\\,km\\,s^{-1}}$, corresponding to $M_{\\rm HI}=2.32\\times10^9\\,M_\\odot$. Reproducing the FAST channel maps requires adding a halo component with the empirical vertical profile $\\rho_{\\rm halo}\\propto \\rho_{\\rm disk}\\,\\sinh(z/z_0)/\\cosh^2(z/z_0)$ to a flaring disk model; the flare-only model leaves large systematic residuals. The inferred halo mass is $3\\times10^8\\,M_\\odot$, about 13% of the total HI mass, and the authors conclude that external gas accretion from dwarf companions and high-velocity clouds, rather than a galactic fountain, is the likely origin of this halo.","pith_inferences":["Editorial inference: if the same FAST-versus-VLA comparison were applied to other edge-on galaxies, some previous interpretations of vertical HI extent as purely flaring disks might need revisiting, because missing short spacings could hide similar halos.","Editorial inference: the HI tail connecting KK 69 to dw3? may be a directly observable cold-accretion stream; mapping it at higher angular resolution and measuring its kinematics could put the accretion rate onto NGC 2683 on a quantitative footing.","Editorial inference: a straightforward testable extension is to apply the same mask-and-compare halo-mass estimate to a small sample of isolated edge-on spirals, to see whether a roughly 10-15% halo fraction is generic or specific to this interacting group."],"forward_implications":["The total HI mass of NGC 2683 is about 10% larger than the earlier VLA-based estimate, so interferometric surveys can miss a substantial diffuse component.","A flaring disk alone cannot reproduce the vertical extent; realistic models of this galaxy must include an extraplanar HI halo.","The proximity and velocity structure of dw3? relative to the HI tail of KK 69 indicate that a dwarf companion is currently donating gas to NGC 2683.","Complex B appears to be in the process of being accreted, making infalling high-velocity clouds a plausible continuing supply for the halo.","Roughly 13% of the galaxy's atomic hydrogen resides outside the thin disk, pointing to external gas accretion rather than star formation as the mechanism building the reservoir."],"supporting_citations":[{"why":"Supplies the VLA data and the disk/flare/warp/ring model that the FAST maps are compared against, plus the halo density formula adopted here.","marker":"Vollmer et al. (2016)"},{"why":"Provides the GMRT and VLA observations of KK 69, dw1, and dw3? that establish the dwarf companions and their HI offsets and tails.","marker":"Saponara et al. (2020)"},{"why":"Supplies the Gaussian-mask method used to separate disk and flare emission from the extraplanar halo and the comparison sample of extraplanar gas fractions.","marker":"Marasco et al. (2019)"},{"why":"Source of the empirical vertical halo profile and the extraplanar gas morphology of NGC 891 that the model draws on.","marker":"Oosterloo et al. (2007)"},{"why":"Catalog of ultra-compact high-velocity clouds used to classify complexes A, B, and C and to argue they are unlikely to be chance line-of-sight objects.","marker":"Adams et al. (2013)"},{"why":"Galactic fountain models used to rule out a star-formation origin for the thick halo.","marker":"Fraternali & Binney (2006)"},{"why":"Provides the low star formation rate estimate for NGC 2683 that weakens the fountain interpretation.","marker":"Wiegert et al. (2015)"},{"why":"Earlier VLA observation that established extended HI beyond the optical disk, providing the baseline that the new vertical extent is measured against.","marker":"Casertano & Van Gorkom (1991)"}],"fun_headline_variants":["FAST spots HI halo four times taller than VLA","FAST reveals NGC2683's hidden HI halo","FAST catches NGC2683 accreting gas from dwarfs","FAST sees HI four times farther from NGC2683's disk"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The vertical excess seen by FAST consists of real, diffuse neutral hydrogen attached to NGC 2683, not artifacts of single-dish calibration, baseline subtraction, stray radiation, or Galactic HI confusion.","fun_headline_variants_meta":{"raw":{"variants":["FAST spots HI halo four times taller than VLA","FAST reveals NGC2683's hidden HI halo","FAST catches NGC2683 accreting gas from dwarfs","FAST sees HI four times farther from NGC2683's disk"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000928,"raw_usage":{"total_tokens":4068,"prompt_tokens":1133,"completion_tokens":2935,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":749,"completion_tokens_details":{"reasoning_tokens":2865}},"tokens_in":749,"tokens_out":2935,"duration_ms":22007,"temperature":1.0,"reasoning_tokens":2865,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-15T23:57:38.353299+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Reprocess the FAST maps with an independent stray-radiation and baseline correction, or re-observe with an interferometric array that has complete short-spacing coverage, and check whether the vertical excess at column densities near $3\\times10^{19}\\,\\mathrm{cm^{-2}}$ persists. If the excess vanishes or is traced to sidelobes or Galactic HI confusion, the halo claim would collapse; if the short-spacing-complete map recovers the full $112.1\\,\\mathrm{Jy\\,km\\,s^{-1}}$ flux with only a flaring disk, no halo would be needed.","supporting_citations":[{"cited_title":"2016, Astronomy & Astrophysics, 586, A98","cited_arxiv_id":null,"evidence_quote":"Supplies the VLA data and the disk/flare/warp/ring model that the FAST maps are compared against, plus the halo density formula adopted here."},{"cited_title":"S., Patra, N","cited_arxiv_id":null,"evidence_quote":"Provides the GMRT and VLA observations of KK 69, dw1, and dw3? that establish the dwarf companions and their HI offsets and tails."},{"cited_title":"2019, Astronomy & Astrophysics, 631, A50","cited_arxiv_id":null,"evidence_quote":"Supplies the Gaussian-mask method used to separate disk and flare emission from the extraplanar halo and the comparison sample of extraplanar gas fractions."},{"cited_title":"2007, The Astronomical Journal, 134, 1019","cited_arxiv_id":null,"evidence_quote":"Source of the empirical vertical halo profile and the extraplanar gas morphology of NGC 891 that the model draws on."},{"cited_title":"A., Giovanelli, R., & Haynes, M","cited_arxiv_id":null,"evidence_quote":"Catalog of ultra-compact high-velocity clouds used to classify complexes A, B, and C and to argue they are unlikely to be chance line-of-sight objects."},{"cited_title":"2006, Monthly Notices of the Royal Astronomical Society, 366, 449","cited_arxiv_id":null,"evidence_quote":"Galactic fountain models used to rule out a star-formation origin for the thick halo."},{"cited_title":"2015, The Astronomical Journal, 150, 81","cited_arxiv_id":null,"evidence_quote":"Provides the low star formation rate estimate for NGC 2683 that weakens the fountain interpretation."},{"cited_title":"1991, Astronomical Journal (ISSN 0004-6256), vol","cited_arxiv_id":null,"evidence_quote":"Earlier VLA observation that established extended HI beyond the optical disk, providing the baseline that the new vertical extent is measured against."}],"review_version":1}