{"id":"a0605059-2f35-4ff7-9d39-56141a46169b","arxiv_id":"2509.07744","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Orbit-based clustering of LAMOST DR9 M giants recovers the Milky Way's known merger relics, finds two unclassified groups, and supports the scenario that the Gaia-Enceladus merger heated the primordial high-alpha disk into the Splash.","lead":"Clustering 3,343 distant M giant stars from the LAMOST DR9 survey by the shape of their orbits recovers the Milky Way's known merger remnants: the Sagittarius stream, the anticenter structure GASS, the Gaia-Enceladus-Sausage, the Splash, and the high-alpha disk. The result bolsters the established picture in which the ancient Enceladus merger stirred part of the primordial alpha-enhanced disk into the Splash population.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Distance-error propagation into IoM clustering is never quantified; a 25% distance uncertainty can move M giants across the adopted linking lengths, so the group assignments and the metal-rich GES claim rest on unverified orbital parameters.","rationale":"The reader's weakest_assumption is exactly the same load-bearing concern I identify: the Qiu et al. (2023) distances with ~25% relative uncertainty are used to compute all orbital parameters, and the paper never propagates the uncertainty. The reader also correctly notes the strongest claim — metal-rich GES M giants confirming a merger-driven formation scenario — depends on the membership assignments. I strengthen the case by pinning it to the FoF linking length: distance errors of the quoted size move orbital parameters by an amount comparable to the range of linking lengths used (0.138-0.487), so the clustering itself is at risk, not just the chemical labeling. I also give credit where it is due: the paper recovers five known substructures, which is a real and nontrivial benchmark, and it honestly discloses the Sgr distance correction and the small APOGEE sample sizes. But the unquantified distance-error budget is a load-bearing gap because the paper's main interpretive claims (Sections 4.3 and 5) are phrased as confirmations of an evolutionary scenario for the early Milky Way, not just as a catalog paper. A Monte Carlo distance-resampling test would settle the concern without requiring new data. I therefore keep the reader's CONDITIONAL verdict rather than moving to REJECT, because the concern is concrete but potentially addressable with a focused numerical experiment.","tokens_in":20320,"tokens_out":3086,"duration_ms":25737,"concrete_test":"Re-run the full FoF pipeline with Monte Carlo resampled distances: for each of the 3,343 stars, draw 100 realizations from the Qiu et al. (2023) distance error distribution, recompute all IoM parameters and delta_ij, and re-run the FoF clustering with the same linking lengths. If the GES, Splash, or high-alpha disk memberships change by more than ~20% in realized samples, or if fewer than ~80% of realizations preserve the reported group structure, the clustering and the metal-rich GES claim are not robust. Additionally, re-run once with the Li et al. (2016a) C-M relation applied to all stars (not just Sgr) to test whether the reported group memberships are an artifact of the Sgr-specific distance correction.","verdict_should_be":"CONDITIONAL","load_bearing_attack":"The central claim that metal-rich M giants in the GES region confirm a specific merger-driven chemical-evolution scenario (Section 4.3, Abstract, Section 5) depends entirely on the FoF group assignments in IoM space. Those assignments are computed from orbital parameters (Section 3) derived from Qiu et al. (2023) photometric distances with a stated relative uncertainty of ~25% (Section 2). The paper never propagates these distance uncertainties into energy, angular momentum, eccentricity, semi-major axis, or the orbit-likelihood distance delta_ij of Eq. (1). With typical Qiu distances of 2.5-5 kpc, a 25% error shifts a star by 0.6-1.3 kpc, which changes orbital energy by tens of percent and can easily move a star across the adopted linking-length thresholds (0.138-0.487 in normalized units). The failure mode is not merely statistical scatter: the paper itself demonstrates a large systematic distance error for Sgr stream members (Section 4.1) and replaces their distances with a different calibration from Li et al. (2016a), 'noticeably underestimated' in the authors' own words. If the same Sgr-specific bias (old, metal-poor populations with different star formation histories) affects the GES, Splash, or high-alpha disk populations, then the 115 GES and 121 Splash memberships, the bimodal low-alpha/high-alpha split among only 11 APOGEE-matched GES stars, and the 'expand the sample' claims for GASS/high-alpha disk could all be artifacts of distance systematics. The paper also does not report the Milky Way potential model used to compute orbits, the exact normalized weights in Eq. (1), or a robustness test against different potentials or distance recalibrations, so the reader cannot check whether the recovered substructures are stable. The external benchmarks (recovering Sgr, GASS, GES etc.) show the method works in broad strokes, but the paper's most novel assertion — that its M giants trace the metal-rich, bimodal-chemistry GES population formed from GES-delivered gas","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper uses the LAMOST DR9 M-giant catalog of Li et al. (2023), distances from Qiu et al. (2023), Gaia DR3 proper motions, and APOGEE DR17 abundances to identify Milky Way substructures through a Friends-of-Friends clustering algorithm applied in integrals-of-motion space. The authors report members of the Sagittarius stream, GASS, GES, Splash, and the high-α disk, plus two groups without clear literature counterparts. They further claim that metal-rich M giants in the GES region show bimodal α-abundances, supporting the Zhao & Chen (2021) scenario in which the GES progenitor delivered metal-enriched gas, and that the identified Splash and high-α disk populations confirm a specific early-Milky-Way evolutionary picture.","tokens_in":20702,"tokens_out":2342,"duration_ms":26441,"significance":"If the structure identifications are robust, the paper provides a useful demonstration that M giants are viable tracers of known Galactic substructures in IoM space, and the metal-rich GES M giants with low- and high-α sequences add an independent observational datapoint to the debate about the GES progenitor's gas content and star formation. The external benchmarks against Belokurov et al. (2014), Hernitschek et al. (2017), Yang et al. (2019b), Li et al. (2021), and Tang et al. (2024) are a strength, and the machine-readable tables of member parameters are a useful community resource. However, the central quantitative claim—the association of individual M giants with specific substructures—rests on orbital parameters derived from photometric distances with ~25% relative uncertainty, and the paper does not demonstrate that the clustering is stable under realistic distance errors. The strongest wording of the conclusions, especially 'confirm the existence of metal-rich constituents within the GES' and 'confirm the evolutionary scenario,' goes beyond what the data and analysis can support.","major_comments":[{"comment":"The paper never quantifies how the ~25% relative distance uncertainty of Qiu et al. (2023) propagates into the orbital parameters used in Eq. (1) and into the FoF group assignments. Since the linking lengths are as small as 0.138 in normalized units, a star displaced by a 25% distance error (0.6–1.3 kpc at the relevant distances) can easily cross the boundaries between groups. The problem is not hypothetical: §4.1 states that the Qiu et al. distances for Sgr members are 'noticeably underestimated' and the authors replace them with a different calibration. That correction was applied after clustering for Sgr only, and the same systematic bias could affect the GASS, GES, Splash, or high-α memberships. A robustness test—e.g., re-running the FoF with distance realizations drawn from the quoted uncertainties, or a bootstrap over distance/velocity errors—is needed to show that the 115 GES, 121","section":"§2, §3, §4.1"},{"comment":"The claim that metal-rich M giants 'confirm' the existence of metal-rich constituents within GES and the evolutionary scenario is not supported by the presented evidence. The GES sample has only 11 stars with reliable APOGEE [M/H] and [α/M] measurements, and the bimodal low-α/high-α separation is based on a visual inspection of Figure 6. No statistical test of bimodality, no significance estimate, and no contamination assessment from the background disk/halo are provided. With only 11 stars, the bimodal appearance could arise from small-number statistics or from overlap with Splash and thick-disk populations. The abstract and §5 should be tempered to 'consistent with' or 'suggest' rather than 'confirm'.","section":"§4.3, §5"},{"comment":"The Sgr distance correction introduces an inconsistency in the analysis pipeline. Clustering in §3 used the original Qiu et al. (2023) distances, while §4.1 replaces the distances for Sgr members with the Li et al. (2016a) color-magnitude relation and then compares the corrected Sgr sample with literature data. It is not stated whether the orbital parameters and group membership were recomputed after the correction, nor how this correction affects the other substructures. This inconsistency weakens the reliability of the Sgr leading-arm decomposition and raises questions about the uniform application of the distance model across the full sample.","section":"§4.1"}],"minor_comments":[{"comment":"The wording 'confirm the existence' and 'confirms the evolutionary scenario' is too strong for the data presented; consider using 'suggests' or 'supports'.","section":"Abstract"},{"comment":"The linking length is selected individually for each group (Appendix A). The paper should state how many groups used each of the two extreme values and whether the results are stable to small changes in the linking length.","section":"§3, Table 2"},{"comment":"The authors acknowledge low-α disk contamination in the high-α sample. It would be useful to quantify the contamination fraction, e.g., by comparing the [α/M] distribution of members with the field sample.","section":"§4.5, Figure 9"},{"comment":"Group 1 and Group 2 have only 15 and 13 members, respectively, with essentially no chemical data. The claim that Group 2 is a metal-rich extension of Thamnos is highly speculative and should be labeled as such.","section":"§4.6"}],"recommendation":"major_revision","confidential_remarks":null},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Quick take: this is a useful catalog paper, not a discovery paper. The clustering is a competent rerun of Wang et al. (2022) on the LAMOST DR9 M giant sample, and the memberships it ships (1,597 stars across 47 groups) are a real new data product. The external benchmarking is the right way to validate it: Sgr, GASS, GES, Splash and high-alpha groups land where they should in Lz-E, Vr-Vphi and eccentricity space, and the Sgr members match external streams. Credit also for being explicit that the Qiu et al. distances are 'noticeably underestimated' for Sgr and for switching to the Li et al. (2016a) CMR.\n\nThe soft spots are real but mostly standard for this literature. The ~25% distance uncertainty is never propagated into orbital parameters or clustering; a 25% error at 2.5-5 kpc moves a star by 0.6-1.3 kpc and can change energy enough to cross the linking lengths. The paper reports linking-length ranges but not the Milky Way potential or the weights in Eq. (1), so readers cannot rerun or test robustness. The sharpest overstatement is the abstract and Section 5: 'confirm the existence of metal-rich constituents within GES...' rests on 11 APOGEE-matched stars, and the bimodal alpha split is a handful of points. That is evidence worth reporting, not confirmation. The two 'unclassified groups' are small (15 and 13 stars) and the paper itself notes Group 1 has no chemistry; calling them candidate new substructures is fine, but the title's 'Unveiling' oversells.\n\nNone of this sinks the paper. The method is established, the recovery of known structures is a genuine check, and the artifacts are honestly noted -- the Sgr distance issue, low-alpha contamination in the high-alpha disk, and the missing chemistry for Group 1 are all in the text. The central identification is externally benchmarked, not circular. I would send this to a serious referee, with a request to add distance-error propagation (even a Monte Carlo), report the potential model and weights, and soften the 'confirm' language to 'support' or 'are consistent with.'","headline":"Solid catalog paper with honest limitations; overstates its conclusions in the abstract.","tokens_in":21415,"tokens_out":1775,"would_cite":true,"duration_ms":20545,"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":"M giant stars recover the Milky Way's known substructures and reveal metal-rich stars formed from gas delivered by the ancient Gaia-Enceladus merger.","keywords":["M giants","Galactic substructures","integrals of motion","Friends-of-Friends clustering","Gaia-Enceladus-Sausage","Splash","Milky Way formation","stellar populations"],"falsifier":"Recompute the clustering on the same 3,343 M giants using distances from an independent calibration, for example asteroseismic or spectrophotometric distances, and repeat the friends-of-friends linkage. Then check whether the 115 Gaia-Enceladus-Sausage members—especially the metal-rich ones with two alpha sequences—still occupy the high-eccentricity, low-angular-momentum locus. If the metal-rich group dissolves, shifts toward the thick-disk sequence, or changes its energy-angular momentum shape, the claim that these stars formed from merger-delivered gas is unsupported. A cheaper check: apply","tokens_in":20146,"feed_emoji":"🌌","tokens_out":7415,"duration_ms":79030,"temperature":0.7,"pith_summary":"The paper tries to establish that M giant stars—luminous, low-temperature red giants that resist extinction—are practical tracers for finding the Milky Way's ancient building blocks. Applying friends-of-friends clustering to 3,343 M giants selected from an updated spectroscopic catalog, with astrometry and chemical abundances, it recovers five known substructures: the Sagittarius stream, the Galactic Anticenter Substructure, the Gaia-Enceladus-Sausage (GES), Splash, and the high-alpha disk, plus two groups that do not match anything previously catalogued. Its central claim is chemical: some M giants inside the GES are metal-rich and split into two alpha-abundance sequences, which the authors read as stars formed from metal-enriched gas delivered by that ancient merger. If true, this adds an independent datapoint for the scenario in which the early Milky Way's high-alpha disk was later heated by the merger into the Splash population while the undisturbed remainder became today's high-alpha disk.","feed_headline":"M giants recover five Milky Way substructures","feed_subtitle":"A friends-of-friends scan of 3,343 bright stars finds the Sgr stream, GASS, GES, Splash, and the high-alpha disk.","key_machinery":"The central machinery is the integrals-of-motion (IoM) space built from five orbital parameters—eccentricity, semi-major axis, orbital pole direction, and apocenter direction—computed from each star's full six-dimensional phase-space coordinates. In this space, the paper applies a friends-of-friends algorithm with an orbit-likelihood distance, so stars sharing similar orbits link into groups even when dynamical mixing has scrambled their current positions. The load-bearing tracer is the M giant itself: luminous and less affected by extinction, it reaches low Galactic latitudes and large distances, which is why the recovered groups extend earlier samples built from K giants, RR Lyrae stars, o","core_discovery":"The paper's central claim is that an M-giant sample, clustered in integrals-of-motion space, can both recover known Milky Way substructures and reveal the chemical signature of star formation fed by an ancient merger. Using five orbital parameters—eccentricity, semi-major axis, orbital pole direction, and apocenter direction—the authors compute an orbit-likelihood distance and apply friends-of-friends clustering to 3,343 M giants. They identify 47 groups totaling 1,597 stars: 182 in the Sagittarius stream, 594 in the Galactic Anticenter Substructure, 115 in the Gaia-Enceladus-Sausage, 121 in Splash, and 557 in the high-alpha disk. The GES members show the characteristic low-angular-momentum,","pith_inferences":["If the distance systematics that forced a correction for Sagittarius stream members also affect GES members, the metal-rich GES claim could be tested by re-running the clustering with an independent distance scale; the paper does not propagate these uncertainties.","The two unclassified groups might be fragments of known structures—Group 2 could be a metal-rich tail of the low-energy retrograde substructure Thamnos—or genuinely new groups; the paper only offers these as hypotheses.","Jointly using chemistry inside the clustering likelihood could suppress the low-alpha contamination the paper notes within the high-alpha disk group.","The M giants with |Z|<2 kpc, deliberately excluded here, could extend the same analysis into the disk plane where extinction matters most and where M giants are especially valuable."],"forward_implications":["M giants can serve as standard tracers for low-latitude and distant Galactic substructures, complementing K giants, RR Lyrae stars, and BHB stars.","The GES sample is extended toward metal-rich M giants with two alpha sequences, supporting the idea that the GES progenitor delivered gas that formed new stars.","The GASS sample grows to 594 M giants, giving more leverage on the outer disk's structure and possible ripple patterns.","Splash and high-alpha disk samples provide an independent view of the early disk heating event: the same thick-disk chemistry appears in two different orbital states.","Two previously unknown groups with no clear counterpart suggest that M giants may reveal further structures not seen with other tracers."],"supporting_citations":[{"why":"Supplies the updated M giant catalog and radial velocities from LAMOST DR9, the starting sample for the analysis.","marker":"Li et al. (2023)"},{"why":"Provides the photometric distances with roughly 25% relative uncertainty used to build six-dimensional phase space.","marker":"Qiu et al. (2023)"},{"why":"Defines the friends-of-friends algorithm and orbit-likelihood distance used for clustering in IoM space.","marker":"Wang et al. (2022)"},{"why":"Gives the metal-rich Sausage-kinematic stars interpretation that the paper's GES chemistry corroborates.","marker":"Zhao & Chen (2021)"},{"why":"Defines the Splash population and the early disk heating scenario the paper confirms.","marker":"Belokurov et al. (2020)"},{"why":"Defines the Galactic Anticenter Substructure and provides the previous M giant GASS sample being expanded.","marker":"Li et al. (2021)"},{"why":"Provides comparison M-giant, K-giant, and BHB samples for the Sagittarius stream and the disk-star chemical baseline.","marker":"Yang et al. (2019b)"},{"why":"Provides the GES and high-alpha disk samples used for kinematic and chemical comparison.","marker":"Tang et al. (2024)"},{"why":"Gives the color-absolute magnitude relation used to correct Sagittarius stream distances, highlighting the distance caveat.","marker":"Li et al. (2016a)"}],"fun_headline_variants":["M giants map 5 hidden Milky Way structures","3343 M giants uncover 5 Galactic substructures","M giants expose ancient merger's chemical traces in 5 structures","Metal-rich M giants reveal Milky Way's merger-built disk"],"cache_read_input_tokens":2688,"weakest_assumption_plain":"The load-bearing premise is that the catalog's photometric distances—with roughly 25% relative uncertainty—are accurate enough that the computed orbital parameters separate true stellar groups rather than distance-error artifacts; the authors themselves replace those distances for Sagittarius stream members because they are noticeably underestimated.","fun_headline_variants_meta":{"raw":{"variants":["M giants map 5 hidden Milky Way structures","3343 M giants uncover 5 Galactic substructures","M giants expose ancient merger's chemical traces in 5 structures","Metal-rich M giants reveal Milky Way's merger-built disk"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001177,"raw_usage":{"total_tokens":4746,"prompt_tokens":834,"completion_tokens":3912,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":578,"completion_tokens_details":{"reasoning_tokens":3847}},"tokens_in":578,"tokens_out":3912,"duration_ms":28645,"temperature":1.0,"reasoning_tokens":3847,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-04T21:42:03.236147+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Recompute the clustering on the same 3,343 M giants using distances from an independent calibration, for example asteroseismic or spectrophotometric distances, and repeat the friends-of-friends linkage. Then check whether the 115 Gaia-Enceladus-Sausage members—especially the metal-rich ones with two alpha sequences—still occupy the high-eccentricity, low-angular-momentum locus. If the metal-rich group dissolves, shifts toward the thick-disk sequence, or changes its energy-angular momentum shape, the claim that these stars formed from merger-delivered gas is unsupported. A cheaper check: apply","supporting_citations":[{"cited_title":"2023, Research in Astronomy and Astrophysics, 23, 055008, doi:10.1088/ 1674-4527/acc1533, 7, 16","cited_arxiv_id":null,"evidence_quote":"Provides the photometric distances with roughly 25% relative uncertainty used to build six-dimensional phase space."},{"cited_title":"2021, Science China Physics, Mechanics, and Astronomy, 64, 239562, doi:10","cited_arxiv_id":null,"evidence_quote":"Gives the metal-rich Sausage-kinematic stars interpretation that the paper's GES chemistry corroborates."},{"cited_title":"2024, ApJ, 965, 62, doi:10.3847/1538-4357/ad32402, 5, 7, 10, 11, 12, 13, 14","cited_arxiv_id":null,"evidence_quote":"Provides the GES and high-alpha disk samples used for kinematic and chemical comparison."}],"review_version":1}