{"id":"124cb106-f04f-42ee-892e-06409c1be775","arxiv_id":"2411.13329","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"A status report introducing BIG-SPARC, a planned database of about 4,000 galaxies with homogeneously derived HI rotation curves, near-infrared photometry, and mass models.","lead":"This paper describes BIG-SPARC, a database of about 4,000 galaxies with hydrogen gas observations, roughly 20 times larger than the earlier SPARC sample. A larger, uniformly processed sample could sharpen tests of dark matter and modified gravity, and prepare for future SKA surveys.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The claimed 'almost 4000 galaxies' rests on unverified de-duplication: in Table 1 the per-survey unique counts sum exactly to the 3882 total, implying no cross-survey overlaps unless explicitly assigned; no such global merge is described.","rationale":"I read the paper as an honest status report: it clearly describes a work in progress, provides a plausible inventory, and shows that the 3DFIT pipeline exists via two example fits. Those positive elements deserve credit. However, the load-bearing quantitative claim is the unique-galaxy count, and the arithmetic of Table 1 exposes a specific weakness: the per-survey unique counts sum exactly to the global total, which is only consistent if every cross-survey duplicate was consciously assigned to one survey. No such global de-duplication step appears in the text. Because the APERTIF catalog is unpublished, external checks of the largest contributor are impossible. This is not an accusation of error; it is a missing validation step that the authors can easily supply. The reader's weakest assumption — that the sample is complete and correctly de-duplicated — is the same concern, so I agree with the conditional verdict. My attack does not move the verdict; it sharpens the condition: the published count should be accompanied by a global overlap analysis and, ideally, a public master catalog before the factor-of-twenty claim is treated as settled.","tokens_in":6830,"tokens_out":6603,"duration_ms":76383,"concrete_test":"Crossmatch all per-survey source lists by PGC/NED identifier supplemented by coordinates to form a global master catalog, then report the overlap matrix between surveys and the resulting distinct-galaxy count. If this distinct count falls below the arithmetic sum of 3882 by more than a few percent, the claimed 'almost 4000' sample size and the >20x factor over SPARC are overestimates that must be revised.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central quantitative claims are the sample inventory — 'a list of almost 4000 galaxies' and a factor of more than 20 over SPARC. The inventory rests entirely on SEARCH detections crossmatched to PGC/NED and visual inspection (Sec. 2), with no completeness, contamination, or duplicate tests shown. The strongest internal red flag is in Table 1: the column of unique galaxies sums exactly to the stated Total of 3882. If any galaxy appears in more than one survey, the sum exceeds the true number of distinct galaxies unless each overlap was explicitly assigned to a single survey. Cross-survey overlap is unavoidable for these samples: WHISP targets overlap heavily with later WSRT/APERTIF fields, and THINGS, HALOGAS, and PHANGS-VLA are all VLA observations of nearby disks. The text does not state that such an assignment was performed; Sec. 2 describes source detection and matching but not global de-duplication across surveys. Furthermore, the largest contribution — APERTIF, with 5954 cubes yielding 1740 unique galaxies — is drawn from Hess et al. (in prep.), so its 3.4-to-1 cube-to-galaxy ratio cannot be checked against a public catalog. If cross-survey duplications were counted twice, the 3882 total is an upper limit rather than the true sample size, and the factor-of-twenty improvement could be overstated. A secondary issue is that no resolution or signal-to-noise threshold is given for what constitutes a usable rotation curve, so the promise of homogeneous rotation curves for roughly 4000 galaxies is not yet operationally defined; Fig. 3 itself shows a 'poorly resolved' case.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This manuscript presents BIG-SPARC, a forthcoming database intended to provide homogeneously derived HI rotation curves, surface brightness profiles, and mass models for about 4000 galaxies, roughly 20 times more than the original SPARC. The paper describes the data collection from 7914 public HI data cubes (Table 1), source finding with SEARCH, crossmatching to PGC and NED, visual inspection, and the resulting claimed list of 3882 unique galaxies. It shows the all-sky distribution and systemic-velocity histogram, and illustrates the planned 3DFIT kinematic analysis with two example galaxies. The paper is explicitly a status report rather than a release of the catalog or derived products.","tokens_in":7135,"tokens_out":2819,"duration_ms":33520,"significance":"If the stated sample size and homogeneity are delivered, BIG-SPARC would be a major community resource, enabling statistically powerful tests of scaling relations, dark matter models, and galaxy evolution that are not possible with the 175-galaxy SPARC sample. The use of public archives and a single analysis pipeline (3DBarolo) is a clear strength, as is the attention to diverse data qualities exemplified by the two test galaxies. However, the paper's central quantitative claims—the number of unique galaxies and the factor-of-twenty improvement over SPARC—rest on internal inventory numbers that are not yet validated by completeness, contamination, or duplicate tests. At this stage the contribution is a plausible and useful progress report, but the quantitative claims need substantiation before the database can be relied upon as a sample definition.","major_comments":[{"comment":"The per-survey unique-galaxy counts in Table 1 sum exactly to the stated total of 3882. This arithmetic consistency can only hold if every galaxy appearing in more than one survey was explicitly assigned to a single survey, but Sec. 2 describes no such global de-duplication step. Several listed surveys target heavily overlapping populations (e.g., WHISP and later WSRT/APERTIF programs; THINGS, HALOGAS, and PHANGS-VLA all observe nearby VLA-visible disks), so cross-survey duplicates are expected. Please specify the global de-duplication procedure, report the number of duplicates found, or revise the claimed total and the factor-of-20 improvement accordingly.","section":"Table 1 and Sec. 2"},{"comment":"The central inventory is based entirely on SEARCH source finding, PGC/NED crossmatching, and visual inspection, but no completeness or contamination tests are presented. The false-detection rate after visual inspection, the completeness as a function of HI mass and distance, and the cross-match success rate all directly affect the promised sample size and homogeneity. Please provide basic validation statistics (e.g., number of SEARCH detections, number rejected as artifacts, cross-match failure rate, duplicate rate), or state explicitly which numbers remain preliminary pending the full data release.","section":"Sec. 2"},{"comment":"No resolution or signal-to-noise criterion is given for what constitutes a usable rotation curve. The two examples in Fig. 3 bracket the data-quality range, but the claim of providing homogeneously derived rotation curves requires a defined quality threshold (e.g., minimum number of independent beams across the galaxy, peak signal-to-noise ratio, or significance of the velocity gradient). Without such a criterion, the promised homogeneous sample is not yet well-defined, especially given the very heterogeneous nature of the input data cubes.","section":"Sec. 4 and Fig. 3"}],"minor_comments":[{"comment":"The term 'HI' appears in several places (e.g., 'HI datacubes' in the abstract) while elsewhere the correct 'H I' is used; please standardize the notation.","section":"Abstract and Sec. 1"},{"comment":"The text says the procedure gave 'a list of almost 4000 galaxies' while Table 1 gives exactly 3882; please use a precise number or state that the count is current as of a specific date to avoid ambiguity.","section":"Sec. 2"},{"comment":"The upper axis shows Hubble distances computed with H0 = 75 km/s/Mpc, but Sec. 3 states the final database will use Cosmicflows-4 distances where available; please clarify whether the histogram is illustrative only and whether the velocity–distance conversion will be updated in the final version.","section":"Fig. 2"},{"comment":"The text refers to 'Duey et al. 2024, in prep.' for the WISE photometry, but the reference list cites 'Duey, F., Schombert, J., McGaugh, S., & Lelli, F. 2024, AJ, 168, 19'; please confirm which status is correct and update the citation.","section":"Sec. 4"},{"comment":"The caption is very dense; labeling the columns and rows with explicit panel identifiers or adding a short legend for the sub-panels would greatly improve readability.","section":"Fig. 3"}],"recommendation":"major_revision","confidential_remarks":"This is a status report for an IAU proceedings, so the bar for a full catalog paper need not apply. Nevertheless, the de-duplication ambiguity in Table 1 is a simple, load-bearing issue: the manuscript should either describe the global merge or soften the sample-size claim. The lack of completeness/contamination tests is understandable at this stage, but the authors should clearly label the inventory as preliminary. The paper is likely acceptable after these clarifications; the central idea and pipeline are sound."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Konstantin — quick read of the BIG-SPARC status report. The pitch is right: SPARC's limits are its 175 galaxies and heterogeneous literature curves, and building a successor from 7,914 public HI cubes with a uniform 3DBarolo pipeline is the logical next step. The two example fits in Fig. 3 suggest the tools can handle both well-resolved and marginal data. That part I buy.\n\nWhat gives me pause is the inventory. Table 1 says 3,882 unique galaxies, but the column sums exactly to that. Since galaxies like NGC 3198 appear in multiple surveys, the per-survey 'unique' counts can only sum to the total if every overlap was explicitly assigned to one survey. The text doesn't say that. If cross-survey duplicates were left in, the real count is lower, possibly still in the thousands, but the 'factor of more than 20' and 'almost 4000' claims in the abstract would be overstated. The largest single contribution — 1,740 galaxies from APERTIF — rests on Hess et al. (in prep.), so the 3.4 cubes-per-galaxy ratio isn't checkable against a public source list. There are also no completeness, contamination, or resolution/S/N thresholds defining what counts as a usable rotation curve, and no catalog is released yet. For a status report that's defensible, but for a paper whose central number is the inventory, it's a gap.\n\nI'd send this to a referee, but a sharp one who will ask for the de-duplication pipeline, a public source list or a testable subset, and a statement of the resolution limits that the final RC products will meet. The project is clearly worth building, and the authors seem aware they're early in the process. As written, the paper is a useful programmatic update, but not yet something to build scaling relations on — which is fine, as long as the 'about 4000' doesn't become a citation datum before the catalog is out.","headline":"A promising, much-needed status report for a homogeneous ~4000-galaxy rotation-curve database, but the headline inventory lacks public validation and a clear de-duplication description.","tokens_in":7717,"tokens_out":3749,"would_cite":true,"duration_ms":40053,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"BIG-SPARC assembles nearly 4,000 galaxies with uniformly derived rotation curves, surface brightness profiles, and mass models, more than 20 times the size of SPARC.","keywords":["H I rotation curves","galaxy kinematics","dark matter","radio interferometry","galaxy surveys","mass models","WISE photometry","tilted-ring modeling"],"falsifier":"Examine the galaxies that appear in both SPARC and BIG-SPARC and compare the rotation curves at matched radii: if well-resolved objects show systematic offsets larger than the quoted uncertainties, or a blind rerun of the source finder on empty fields shows many false detections, the homogeneous pipeline is not delivering the promised accuracy and sample size.","tokens_in":6648,"feed_emoji":"📡","tokens_out":11895,"duration_ms":112626,"temperature":0.7,"pith_summary":"BIG-SPARC is a new database, assembled from 7,914 public H I data cubes, that will provide nearly 4,000 galaxies with homogeneously derived rotation curves, near-infrared surface brightness profiles, and mass models. The paper describes the construction: sources are found with an automated source finder, crossmatched against galaxy catalogues, visually inspected, and fitted with a uniform kinematic model. This is meant to cure the two main limitations of the earlier SPARC database: its small size (175 galaxies) and the heterogeneity of rotation curves compiled from the literature. If it works, the community gains a sample more than 20 times larger, enabling tests of dark matter models, galaxy evolution, and modified gravity theories with far greater statistical power, and preparing for the much larger H I surveys expected with the Square Kilometre Array and its pathfinders.","feed_headline":"4,000 galaxies get uniform rotation curves from archival radio data","feed_subtitle":"The sample is 20 times larger than SPARC and built with one homogeneous pipeline, ready to stress-test dark matter models.","key_machinery":"The carrying mechanism is 3DBarolo, a software package for automated source finding and kinematic modeling of H I data cubes, used in two stages: SEARCH detects sources and produces moment maps, while 3DFIT fits a tilted-ring model to derive rotation velocity, surface density, velocity dispersion, and orientation parameters at each radius. Around this sits an identification chain—crossmatching against PGC and then a general extragalactic database, followed by visual inspection—that turns raw detections into a curated list of unique galaxies; on the photometric side, Spitzer and WISE W1 imaging supply the near-infrared light needed for mass models.","core_discovery":"The paper's claim is that a single homogeneous pipeline can be applied to the archival record of H I observations to produce a database of 3,882 unique galaxies, about 4,000 in round numbers. Starting from 7,914 data cubes drawn from many surveys and individual studies, the authors run the SEARCH source finder of the 3DBarolo software, crossmatch detections with the PGC catalogue and, when that fails, a general extragalactic database, and visually inspect every detection to remove artifacts. Kinematic fits with the 3DFIT task yield rotation curves, H I surface density profiles, velocity dispersions, and geometric parameters, while near-infrared photometry from Spitzer, supplemented by all-sky WISE W1 data, provides the baryonic mass side. The result is an order-of-magnitude jump in sample size over SPARC, extending to roughly twice the maximum distance, with the promise of uniform data products that make the small intrinsic scatter of scaling relations measurable.","pith_inferences":["Inference: if the sample is as clean as claimed, the radial acceleration relation and baryonic Tully-Fisher relation can be tested at fixed stellar mass with hundreds of galaxies per bin, which would sharpen the question of whether their scatter is truly intrinsic.","Inference: because the input cubes range from well-resolved to barely resolved, the database will contain a long tail of low-quality rotation curves; future users will need per-galaxy reliability flags, which this paper does not yet specify.","Inference: the sky coverage is telescope-driven and northern-biased, so BIG-SPARC is not a volume-limited or flux-limited sample; scaling-relation work will need to model these selection effects rather than treat the catalogue as representative.","Inference: WISE W1 photometry is all-sky but not identical to Spitzer mid-infrared photometry, so combining the two without a matched zeropoint calibration would propagate a systematic into stellar mass estimates."],"forward_implications":["The galaxy sample grows from 175 to roughly 3,900, a more than 20-fold increase in the number of rotation curves.","All rotation curves, surface brightness profiles, and mass models come from one uniform fitting procedure, removing the largest source of heterogeneity in the predecessor database.","The distance baseline roughly doubles: the farthest galaxy sits at about twice the distance of the farthest in SPARC, with direct distances from Cosmicflows-4 where available and Hubble-flow distances elsewhere.","The database can support scaling-relation studies at fixed mass, environment, or gas fraction, where SPARC's 175 galaxies lacked statistical power.","The pipeline and data products establish a template for handling the next order-of-magnitude increase in H I sources expected from SKA-era surveys."],"supporting_citations":[{"why":"Establishes the original SPARC database of 175 rotation curves, the baseline BIG-SPARC is designed to surpass in size and homogeneity.","marker":"Lelli et al. 2016a"},{"why":"Supplies the SEARCH source finder and the 3DFIT kinematic fitter used on every data cube.","marker":"Di Teodoro & Fraternali 2015"},{"why":"Provides the PGC galaxy catalogue used in the first crossmatch of H I detections.","marker":"Paturel et al. 2003"},{"why":"The APERTIF survey, source of the largest share of cubes (5,954) and unique galaxies (1,740).","marker":"Adams et al. 2022"},{"why":"WALLABY early science data, contributing 592 cubes and 486 galaxies from ASKAP.","marker":"Koribalski et al. 2020"},{"why":"Supplies the H0 = 75 km/s/Mpc distance scale used to convert systemic velocities to Hubble distances.","marker":"Tully et al. 2016"},{"why":"Cosmicflows-4, the source of direct distances for the final database where available.","marker":"Tully et al. 2023"},{"why":"WISE W1 near-infrared photometry, used for the surface brightness profiles and mass models where Spitzer imaging is lacking.","marker":"Duey et al. 2024"}],"fun_headline_variants":["Homogeneous rotation curves for 4,000 galaxies from archival HI data","BIG-SPARC: 20x more galaxies with uniform mass models","4,000 galaxies get uniform rotation curves from archives","New database 20x larger for testing dark matter models"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the 3,882-galaxy list is essentially complete and correct—that the automated source finder, catalogue crossmatching, and visual inspection admit few duplicates or artifacts, and that archived cubes, many shallow or poorly resolved, still yield trustworthy rotation curves.","fun_headline_variants_meta":{"raw":{"variants":["Homogeneous rotation curves for 4,000 galaxies from archival HI data","BIG-SPARC: 20x more galaxies with uniform mass models","4,000 galaxies get uniform rotation curves from archives","New database 20x larger for testing dark matter models"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000347,"raw_usage":{"total_tokens":1900,"prompt_tokens":948,"completion_tokens":952,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":564,"completion_tokens_details":{"reasoning_tokens":879}},"tokens_in":564,"tokens_out":952,"duration_ms":8014,"temperature":1.0,"reasoning_tokens":879,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T16:32:06.879095+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Examine the galaxies that appear in both SPARC and BIG-SPARC and compare the rotation curves at matched radii: if well-resolved objects show systematic offsets larger than the quoted uncertainties, or a blind rerun of the source finder on empty fields shows many false detections, the homogeneous pipeline is not delivering the promised accuracy and sample size.","supporting_citations":[{"cited_title":"2003, A&A, 412, 45","cited_arxiv_id":null,"evidence_quote":"Provides the PGC galaxy catalogue used in the first crossmatch of H I detections."},{"cited_title":"S., Staveley-Smith, L., Westmeier, T., et al","cited_arxiv_id":null,"evidence_quote":"WALLABY early science data, contributing 592 cubes and 486 galaxies from ASKAP."},{"cited_title":"B., Kourkchi, E., Courtois, H","cited_arxiv_id":null,"evidence_quote":"Cosmicflows-4, the source of direct distances for the final database where available."},{"cited_title":"2024,AJ, 168, 19","cited_arxiv_id":null,"evidence_quote":"WISE W1 near-infrared photometry, used for the surface brightness profiles and mass models where Spitzer imaging is lacking."}],"review_version":1}