{"id":"1a6d1c63-2888-41c3-97ce-8e63e1cf88e3","arxiv_id":"2504.18725","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":6,"one_line_summary":"A homogeneous 15-disk ALMA continuum analysis quantifies disk asymmetries with a new index, links strong asymmetries to inner-disk accretion and NIR excess, and shows that larger disks taper more slowly at their outer edges.","lead":"This paper maps the millimeter-wave dust structure of 15 planet-forming disks from the exoALMA survey, cataloging rings, gaps, cavities, and asymmetries using visibility-space modeling. It finds that disks with stronger asymmetries tend to have higher inner-disk accretion rates and near-infrared excess, and that larger disks have more gradually tapered outer edges.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"NAI conflates geometric model mismatch with physical asymmetry, so the NAI–accretion and NAI–NIR correlations in Sec. 5.2 may be inflated by warped/offset inner disks in the most asymmetric sources.","rationale":"Read in good faith, this is a well-executed survey paper with a publicly released pipeline, a useful substructure catalog, and independent checks: the outer-taper result is partly corroborated by the independent 12CO R90 correlation, and the disk-by-disk comparisons with published high-resolution images and gas kinematics support many individual substructure identifications. The authors also explicitly acknowledge the biased, small sample. The weakest link is the NAI metric, which is the sole quantitative basis for the headline inner–outer connection. It is defined relative to a deliberately simple axisymmetric model, carries no quoted uncertainty, and the residual maps are dominated by inner-region features for many sources, exactly where a single global geometry is least reliable. Since the highest-NAI sources include disks that the paper itself identifies as having warped or offset inner structures, the correlations in Sec. 5.2 could be inflated by geometric mismodeling rather than reflecting physical outer-disk asymmetry. This does not invalidate the catalog or the outer-taper analysis, but it makes the central correlational claim conditional on a targeted robustness test. The proposed two-geometry recomputation would settle whether the correlation survives when geometric complexity is modeled instead of absorbed into residuals. Therefore the appropriate disposition is unchanged from the reader: conditional acceptance pending that test.","tokens_in":42376,"tokens_out":6182,"duration_ms":74026,"concrete_test":"Recompute NAI for all 15 disks after augmenting the pipeline so that the inner disk/ring is fit with its own independent (i, PA, center) while the outer disk retains the galario geometry, then subtract the resulting two-geometry model before computing residuals. Re-run the Sec. 5.2 Kendall-tau tests on NAI versus Mdot/Mstar^1.8 and NAI versus NIR excess. If both correlations remain at tau > 0.4 with p < 0.05, the geometric-mismatch concern is not load-bearing; if significance drops, the claimed inner–outer connection is not robust to geometric model choice.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The nonaxisymmetry index in Eq. (6) is computed after subtracting a single frank model that assumes one global inclination, PA, and center, with geometry estimated by galario from the flux-dominant outer disk (Sects. 3.1–3.2). Any real geometric complexity in the inner disk—warps (AA Tau, HD 143006), inner-disk offsets (DM Tau, J1615, V4046 Sgr), or an eccentric/offset inner cavity (MWC 758)—is not absorbed by the model but appears in the residual image and inflates NAI. Several of the six most nonaxisymmetric disks are independently described in Sec. 5.1 as having warped, offset, or elevated inner structure, so the claimed NAI–accretion and NAI–NIR correlations in Sec. 5.2 may partly track how badly a single-geometry axisymmetric model fits the inner disk rather than only physical outer-disk asymmetry. The paper provides no uncertainty on NAI and no injection-recovery test separating geometric mismatch from true nonaxisymmetric emission. Because the same residual maps are interpreted as shadows, spirals, and offsets, the metric conflates the physical signal with the model's inability to represent it.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents the continuum component of the exoALMA Large Program for 15 protoplanetary disks, using a two-step visibility-space pipeline: galario fits for global geometry (inclination, position angle, and center offsets) and frank nonparametric fits for the azimuthally averaged radial intensity profile. From the frank residuals the authors define a NonAxisymmetry Index (NAI, Eq. 6) and report positive rank correlations between NAI and stellar-mass-normalized accretion rate (Kendall tau 0.45, p=0.02) and between NAI and NIR excess (tau 0.48, p=0.01). The paper also fits an exponential taper to the outer continuum emission beyond R90 (Eq. 7), reporting that larger dust and gas disks show shallower outer tapers. In addition, the paper catalogs rings, gaps, cavities, shadows, offsets, and candidate external substructures for each source, and compares continuum features with planet-kink locations from gas kinematics.","tokens_in":42654,"tokens_out":4338,"duration_ms":48792,"significance":"If the two central empirical claims hold, the paper provides a valuable link between outer-disk continuum morphology and inner-disk accretion state, and a new observational metric for outer-disk size evolution. The strengths of the paper are its carefully described visibility-space methodology, the public release of the analysis pipeline and value-added data products, the honest acknowledgment of the biased sample selection, and the detailed source-by-source comparison with previous work. The new morphological catalog for the homogeneous exoALMA sample is itself a useful contribution. However, the NAI currently has no uncertainty estimate and no demonstrated separation between physical nonaxisymmetry and single-geometry model mismatch, and the R90-lambda_out relation has a partly structural origin; these issues affect the two headline correlations and therefore require substantial revision rather than minor polishing.","major_comments":[{"comment":"The NAI is computed from residuals produced by subtracting a single frank model that assumes one global inclination, position angle, and center, with geometry determined by galario from the flux-dominant outer disk (Sec. 3.1). The source-by-source discussion in Sec. 5.1 identifies exactly the systems where this assumption fails: warped or misaligned inner disks (AA Tau, HD 143006), inner-disk offsets (DM Tau, J1615, V4046 Sgr), and an eccentric inner cavity (MWC 758). These geometric mismatches appear in the residual images and are summed into Eq. (6) with SNR>=5, so the NAI partly measures how poorly a single-geometry axisymmetric model represents real geometric complexity rather than only physical nonaxisymmetric emission. The paper provides no uncertainty on the NAI and no injection-recovery test that quantifies this geometric-mismatch contribution. I request an explicit test: inject synthetic disks with known warps, offsets, and eccentric cavities into the pipeline and report how much the NAI changes, or alternatively recompute the NAI after masking the inner disk or after locally refitting the geometry, and provide bootstrap uncertainties. Without this, the Sec. 5.2 correlations may be inflated by the very sources whose inner-disk geometry is most complex.","section":"Sec. 4.2, Eq. (6)"},{"comment":"The reported correlation statistics rest on only 15 targets from a sample selected for bright, extended gas emission, and the manuscript does not state how the four upper limits on NIR excess (DM Tau, J1615, J1852, V4046 Sgr in Table D.1) are incorporated into the Kendall tau calculation. In addition, the accretion rates are heterogeneous literature values with a common 0.35 dex uncertainty, and the NAI itself has no propagated uncertainty, so the quoted p-values (0.02 and 0.01) treat the most uncertain variable as exact. Because the six most asymmetric disks dominate the correlation and also include most of the geometrically complex inner disks flagged in the previous comment, the statistical significance is likely overstated. I ask the authors to state the censoring method for upper limits, to show the correlations after removing each of the high-NAI sources one at a time, and to provide a sensitivity analysis that varies the SNR threshold in Eq. (6). The qualitative claim may survive these tests, but the current significance statement is not supported.","section":"Sec. 5.2, Fig. 5 and Table D.1"},{"comment":"The reported trend that larger disks have shallower outer tapers is partly structural: lambda_out in Eq. (7) is fitted to the same azimuthally averaged CLEAN profile from which R90 is defined, and for a purely exponential outer falloff the radius enclosing a fixed fraction of the flux scales linearly with the taper length. The correlation in Fig. 8 therefore does not establish an independent physical relation between disk size and taper steepness. Moreover, three of the smallest disks (HD 143006, MWC 758, PDS 66) have sigma_fit/sigma_beam < 2 in Table 4, meaning their outer falloff is not resolved, and these same sources anchor the compact, steep-taper end of the trend. I recommend presenting R90/lambda_out as a normalized, scale-free metric, or fitting the outer falloff in a way that does not use R90 as the inner boundary, and repeating the trend with the unresolved sources removed or downweighted. The current presentation overstates the strength of the size-taper relation.","section":"Sec. 5.4, Fig. 8 and Table 4"}],"minor_comments":[{"comment":"The choice of SNR>=5 in Eq. (6) is a free parameter, and Table A.1 lists NAI values without any uncertainty; at minimum a bootstrap over the frank geometry sampling described in Sec. 3.2 should be used to attach error bars to the NAI.","section":"Sec. 4.2"},{"comment":"The text notes that the most asymmetric sources also tend to have higher stellar masses (Fig. E.2); since the accretion rate is normalized by Mstar^1.8, a residual stellar-mass dependence could contribute to the left panel of Fig. 5 and should be discussed more explicitly.","section":"Sec. 5.2"},{"comment":"The exponential fit in Eq. (7) is stated to partially fail for V4046 Sgr, yet this source is still included in Fig. 8; either exclude it from the trend or add a caveat in the text and figure.","section":"Sec. 5.4"},{"comment":"There is a typo, 'a a massive planet', and the phrase 'a a' should be corrected.","section":"Sec. 5.1, CQ Tau paragraph"},{"comment":"The paper would benefit from a short summary table that lists, for each source, the NAI, lambda_out, R90, sigma_fit/sigma_beam, and whether the inner disk is reported as warped or offset; this would make the dependence of the headline correlations on these classifications immediately transparent.","section":"General"}],"recommendation":"major_revision","confidential_remarks":"This is a strong observational paper from a major program, and the source catalog and pipeline are valuable. My recommendation of major revision is driven by the need to quantify the geometric-mismatch contribution to the NAI, to properly handle upper limits and small-sample statistics in the correlations, and to present the size-taper trend in a way that removes its structural component. These are fixable with additional analysis and revised framing; I see no reason to doubt the integrity of the data or the pipeline, and the manuscript is well within the scope of the journal."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: this is a solid, careful survey paper that introduces two useful metrics for continuum morphology and ships a catalog and clean data products. The headline correlations (NAI with accretion and NIR excess) and the outer-taper relation are plausible but should be treated as provisional: the NAI mixes inner-disk geometry with outer-disk asymmetry, and the size–taper correlation is partly structural.\n\nWhat's new and done well: the NAI for the 15 exoALMA disks, the lambda_out taper scale lengths, and the correlations themselves are new. The substructure catalog is not dramatically new since many sources were known, but the homogeneous deep dataset adds real value. The pipeline is described carefully, the authors acknowledge sample bias and unresolved falloffs, and the public data products are a plus.\n\nWhere it bends: the NAI has no uncertainties, and it is computed over residuals that, by the authors' own admission (Sec 4.2), are concentrated in the inner regions. Several of the six most asymmetric disks (AA Tau, HD 143006, MWC 758, plus DM Tau, J1615, V4046 Sgr with offsets) are exactly the ones with warped, offset, or eccentric inner structures. So the correlated signal may partly measure how badly a single-geometry flat model fits the inner disk, not a clean outer-disk asymmetry. The 'outer disk dust substructures' wording in the abstract and conclusions is stronger than the metric supports. Addressable: split NAI into inner and outer components, or run injection-recovery to calibrate the geometric contribution.\n\nThe R90–lambda_out correlation is partly built into the definitions: R90 encloses 90% of the flux, so a shallower exponential tail shifts R90 outward. The gas R90 correlation helps, but gas and dust sizes are physically linked, so it does not fully break the degeneracy. They also include three sources where the outer falloff is not resolved; a robustness check excluding them is warranted.\n\nThe Kendall tau on NIR excess includes upper limits; the paper does not say whether censored statistics were used. Minor but worth fixing.\n\nNone of this kills the paper. The correlations are suggestive, and the survey products will be reused. But the interpretive claims need tempering and a robustness pass. I'd send it to a serious referee, expect a revise-and-resubmit, and bring it to reading group.","headline":"Useful survey with new metrics, but the NAI correlations and the size–taper relation are less clean than the abstract implies.","tokens_in":43367,"tokens_out":5768,"would_cite":true,"duration_ms":60032,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["95.85.Bh"],"model":"deepseek-v4-flash","headline":"The lopsidedness of a planet-forming disk's millimeter dust tracks how fast the disk is feeding its star, linking outer substructures to inner disk activity.","keywords":["Protoplanetary disks","Dust continuum emission","Planet formation","Radio interferometry","Nonaxisymmetry index","Disk substructures","Outer disk taper","exoALMA"],"falsifier":"Recompute the NAI with spatially varying geometry, for instance giving AA Tau, DM Tau, J1615, and V4046 Sgr an inner component with its own center, inclination, and position angle before subtracting the model. If the $\\tau \\approx 0.45$ correlation with accretion rate drops to insignificance once those degrees of freedom are fitted, the asymmetry signal is substantially geometric rather than physical. A complementary check is to run the identical pipeline on a sample not selected for brightness and extent; a vanishing correlation would implicate the sample selection.","tokens_in":42192,"feed_emoji":"🪐","tokens_out":14110,"duration_ms":116028,"temperature":0.7,"pith_summary":"This paper uses the deep 0.9 mm (331.6 GHz) continuum images that the exoALMA Large Program recorded for 15 protoplanetary disks and asks what the shape of the dust emission says about how those disks are evolving. The authors subtract from each disk the best axisymmetric model built with visibility-space fitting and condense all of the leftover, nonaxisymmetric light into a single number, the nonaxisymmetry index (NAI). They find that disks with a higher NAI accrete onto their central stars more vigorously and shine with more near-infrared excess from hot inner dust, with Kendall's $\\tau = 0.45$ ($p = 0.02$) for the mass accretion rate and $\\tau = 0.48$ ($p = 0.01$) for NIR excess, and that almost all of the most asymmetric disks contain inner cavities. The same data reach the faint outermost disk, where an exponential taper $I(R) = I_0 \\exp(-R/\\lambda_{\\rm out})$ shows that larger disks fall off gradually while compact disks end sharply. If these correlations are real, a single deep continuum observation of a disk could serve as a window into its inner accretion state and its size evolution history.","feed_headline":"Lopsided dust means faster inner accretion","feed_subtitle":"Deep ALMA maps of 15 planet-forming disks tie outer asymmetry to inner-disk feeding rates.","key_machinery":"The argument runs on two newly defined scalars. The nonaxisymmetry index (NAI) is computed from CLEAN images of the data, of the axisymmetric frank model, and of their difference: $\\mathrm{NAI} = \\sum_{i,j} |I_{{\\rm res},i,j}| / \\sum_{i,j} |I_{{\\rm mod},i,j}|$ over pixels where the data exceed $5\\sigma$. It collapses spirals, crescents, shadows, warps, and off-center inner disks into one number, which is what makes the correlation search against accretion rate and NIR excess possible. The outer taper scale length $\\lambda_{\\rm out}$ comes from an exponential fit $I(R) = I_0 \\exp(-R/\\lambda_{\\rm out})$ applied between $R_{90}$ and the radius where emission drops to $5\\sigma$, with a resolution check $\\sigma_{\\rm fit}/\\sigma_{\\rm beam} > 2$ guarding the steepest slopes. Both scalars rest on the same two-step visibility pipeline: galario supplies the global inclination, position angle, and center through parametric MCMC fitting, and frank supplies the superresolution axisymmetric radial brightness profile whose residuals define the NAI.","core_discovery":"The central claim is that outer dust morphology and inner disk activity are connected, established through two correlations among the 15 exoALMA disks. After deriving each disk's geometry with the parametric visibility code galario and reconstructing the axisymmetric brightness profile with the nonparametric code frank, the authors image the residual visibilities and define the NAI as the sum of absolute residual intensities divided by the sum of absolute model intensities, restricted to pixels where the observed image exceeds $5\\sigma$. The NAI is positively correlated with the stellar-mass-normalized accretion rate $\\dot{M}/M_*^{1.8}$ (Kendall $\\tau = 0.45$, $p = 0.02$) and with the NIR excess ($\\tau = 0.48$, $p = 0.01$); the six most asymmetric disks all have NAI above 0.1, and five of them host inner cavities. The paper interprets this as evidence that a massive inner perturber could simultaneously carve the cavity, stir spirals and crescents in the outer dust, and drive accretion onto the star, while noting that the sample of bright, extended disks is biased and the trends need confirmation on a more representative population. On the outer-edge side, fitting $I(R) = I_0 \\exp(-R/\\lambda_{\\rm out})$ to the azimuthally averaged emission beyond the 90%-flux radius shows that $\\lambda_{\\rm out}$ grows with the 90%-flux radius in both dust and gas: bigger disks taper gently, smaller ones truncate steeply, a relation not yet addressed by theoretical models of disk sizes.","pith_inferences":["The NAI may partly measure geometric complexity rather than physical asymmetry, because a single global geometry is assumed for every disk; re-fitting warped or offset inner disks before computing residuals would separate the two and test whether the NAI-accretion link is carried by warps and misalignments.","Recomputing the NAI and $\\lambda_{\\rm out}$ on a sample not preselected for brightness and extent would show whether the correlations survive beyond the bright, substructure-rich sources targeted by exoALMA.","The V4046 Sgr profile, which resists a single exponential slope, warns that one scale length may conceal multiple physical components in the outer disk; decomposing such profiles could reveal separate infall and truncation signatures.","If high NAI flags active accretion, surveys that cannot measure accretion lines, such as embedded or edge-on disks, could use continuum asymmetry as a rough accretion diagnostic."],"forward_implications":["A deep continuum image becomes a proxy for inner disk state: once the NAI is calibrated, disks with strong asymmetry can be flagged as actively accreting without measuring accretion tracers directly.","The cavity-asymmetry-accretion cluster points to a specific physical picture, a massive inner companion carving the cavity and driving both outer asymmetries and accretion, that future hydrodynamical simulations can test.","The $\\lambda_{\\rm out}$-$R_{90}$ relation offers a new observable for disk evolution models: radial drift, photoevaporation, late infall, and flyby truncation all predict different outer-edge shapes that can now be compared with a measured scale length.","At exoALMA sensitivity, nonaxisymmetry may be the rule rather than the exception: 14 of 15 disks show residual structure above $5\\sigma$, with only PDS 66 appearing smooth.","Kinematic planet candidates identified in the gas sit at or beyond the dust edge in several sources, so outer gaps and truncations aligned with those kinks are consistent with planets shaping the outer disk."],"supporting_citations":[{"why":"Supplies galario, the parametric visibility-fitting code that measures each disk's inclination, position angle, and center.","marker":"Tazzari et al. 2018"},{"why":"Supplies frank, the nonparametric visibility fitter whose axisymmetric brightness profile defines the residuals and the NAI.","marker":"Jennings et al. 2020"},{"why":"Presents the exoALMA program and its sample selection, supplying the 15 disks and the deep Band 7 observations analyzed here.","marker":"Teague et al. 2025"},{"why":"Provides the stellar-mass scaling used to normalize accretion rates before correlating them with the NAI.","marker":"Manara et al. 2023"},{"why":"Supplies NIR excess values and the measurement procedure for sources not in that sample, forming the second correlated quantity.","marker":"Garufi et al. 2018"},{"why":"Establishes the ring/gap nomenclature and the width, depth, and contrast criteria used to define axisymmetric substructures.","marker":"Huang et al. 2018"},{"why":"Provides the gas kinematic velocity-kink planet candidates whose locations are compared with the continuum substructures.","marker":"Pinte et al. 2025"},{"why":"Supplies the gas (12CO) radii used in the lambda_out-size correlation.","marker":"Galloway-Sprietsma et al. 2025"}],"fun_headline_variants":["Asymmetric disks feed faster, ALMA finds","Lopsided dust hints at faster inner accretion","Outer disk lopsidedness tracks inner accretion rate","Disk asymmetry matches accretion and NIR excess","ALMA sees link: lopsided disks accrete quicker"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"Everything downstream assumes that one global inclination, position angle, and center, fitted from the brightest outer emission, deprojects the whole disk correctly; for warped or offset inner regions the geometric mismatch is absorbed into the residuals and inflates the NAI.","fun_headline_variants_meta":{"raw":{"variants":["Asymmetric disks feed faster, ALMA finds","Lopsided dust hints at faster inner accretion","Outer disk lopsidedness tracks inner accretion rate","Disk asymmetry matches accretion and NIR excess","ALMA sees link: lopsided disks accrete quicker"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00048,"raw_usage":{"total_tokens":2445,"prompt_tokens":1083,"completion_tokens":1362,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":699,"completion_tokens_details":{"reasoning_tokens":1286}},"tokens_in":699,"tokens_out":1362,"duration_ms":9968,"temperature":1.0,"reasoning_tokens":1286,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-16T10:10:50.385110+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Recompute the NAI with spatially varying geometry, for instance giving AA Tau, DM Tau, J1615, and V4046 Sgr an inner component with its own center, inclination, and position angle before subtracting the model. If the $\\tau \\approx 0.45$ correlation with accretion rate drops to insignificance once those degrees of freedom are fitted, the asymmetry signal is substantially geometric rather than physical. A complementary check is to run the identical pipeline on a sample not selected for brightness and extent; a vanishing correlation would implicate the sample selection.","supporting_citations":[],"review_version":1}