{"id":"2191f529-ac50-44de-bab8-370570cc1497","arxiv_id":"2607.14207","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"A large UNIONS weak-lensing sample finds an increasing subhalo-to-stellar-mass ratio with cluster-centric radius in redMaPPer clusters, qualitatively confirming tidal stripping of satellite dark matter halos.","lead":"This paper uses weak lensing from 4.5 million UNIONS background galaxies to measure the dark matter halos of 330,000 satellite galaxies in redMaPPer clusters. It finds that satellites closer to the cluster centre hold significantly less dark matter per star, supporting the tidal-stripping picture — but the quoted masses come from a model that fails to fit its own data.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Poor model fit and degeneracy between M_sub, host amplitude A, and mis-centering σ make the reported HSMR trend model-dependent; the positive trend may be inflated by the free Rayleigh σ in the outer bins.","rationale":"The reader's weakest_assumption identifies the mis-centering model and two-halo term as the central worry. My analysis agrees and adds a specific mechanism: the large free σ in the outer bin, together with the omitted two-halo term, can bias M_sub upward through degeneracy with the host-halo subtraction. The high reduced χ² (≈3) shows the model is statistically rejected, so the reported uncertainties are not reliable and the significance of the trend is overstated. The Appendix B admixture model, while not physically preferred, demonstrates that changing the host treatment changes the HSMR values substantially (e.g., outer bin from 53 to 32). However, the trend is not entirely an artifact—it persists in the admixture model—so the verdict remains CONDITIONAL rather than REJECT. The proposed test directly probes the sensitivity of the outer-bin mass to the two most questionable modeling choices; if the shift is small, the trend is robust, and if not, the paper's central claim weakens. I therefore recommend no change to the reader's conditional verdict, but the test is necessary to settle the issue.","tokens_in":16921,"tokens_out":5085,"duration_ms":53569,"concrete_test":"Re-fit the outer cluster-centric bin (0.86–1.29 Mpc) with a model that includes a two-halo term and fixes the Rayleigh σ to the Zhang et al. (2019) value (~100 kpc) rather than leaving it free. Compare the resulting M_sub posterior to the reported 12.58±0.12. If the median shifts by more than the reported 1σ uncertainty, or if the HSMR trend across bins flattens, the claimed trend is not robust to host-subtraction modeling.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim is the positive HSMR trend in Table 1 (11.99, 33.68, 53.02). The load-bearing assumption is that the offset host-halo term is correctly described by a single Rayleigh mis-centering distribution (Eq. 6) with free σ, and that the two-halo term is negligible. This is insecure. First, the best-fit model is formally rejected: reduced χ² = 3.59, 2.77, 3.39 for 11 dof (Table 1), so the model does not reproduce the data. Second, the fitted σ rises from 83 to 348 kpc across bins, and the outer two bins disagree strongly with the independent Zhang et al. (2019) estimates (Table 2). A large σ smears the negative host-halo signal at intermediate R; the fit can compensate by increasing M_sub, inflating HSMR in the outer bin. The omitted two-halo term (acknowledged in §2 and §6) has the same direction: a missing positive large-scale contribution can bias A and M_sub. Since A and σ are free per bin, M_sub is degenerate with the host subtraction. Appendix B's admixture model fits better (χ²_red 1.7–1.8) and yields lower HSMR values (4.0, 13.4, 32.1), demonstrating strong model dependence, although a trend persists. The claim 'confirm tidal stripping' thus rests on a model that fails the data.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper measures the galaxy-galaxy lensing signal (excess surface density) around ~330,000 redMaPPer satellite galaxies in three cluster-centric radius bins, using UNIONS/ShapePipe source shapes. The lensing signal is modelled as the sum of a point-mass stellar term, an NFW subhalo term with free subhalo mass M_sub, and an offset NFW host-halo term with a free amplitude A and a Rayleigh-distributed BCG mis-centring scale sigma. The key result is a claimed strong positive trend in subhalo-to-stellar mass ratio (HSMR) with cluster-centric radius: M_sub/M* = 11.99, 33.68, 53.02 in the three bins, interpreted as observational confirmation of tidal stripping of satellite dark matter halos. The paper also presents an alternative 'admixture' host-halo model in Appendix B that fits the data better and yields lower HSMR values (4.0, 13.4, 32.1), while preserving the trend.","tokens_in":17334,"tokens_out":2829,"duration_ms":33794,"significance":"If the result holds, the paper would provide a high-S/N observational confirmation of tidal stripping of satellite subhalos in clusters, using a large and valuable data combination (UNIONS sources, redMaPPer lenses). The data volume is a genuine strength: ~4.5 million source galaxies and ~330,000 satellite lenses. The authors are transparent about their modelling choices and include an appendix with an alternative host-halo model. The trend with cluster-centric radius is plausible and broadly consistent with earlier work. However, the significance of the paper is undermined by the fact that the adopted model is statistically rejected by the data (reduced chi-square 3.59, 2.77, 3.39 for 11 dof in Table 1), and by the strong model-dependence of the HSMR values (Appendix B). The conclusion 'confirm tidal stripping' is therefore not yet supported at the level claimed.","major_comments":[{"comment":"The adopted model is statistically rejected by the data. The reduced chi-square values are 3.59, 2.77, and 3.39 for 11 degrees of freedom in the three bins, corresponding to very small p-values. This contradicts the abstract's claim that the model 'effectively reproduces the observed lensing signal'. The paper itself documents large residual features, including an unexplained 'hump' near 0.5 Mpc in the outermost bin (Section 6, Figure 4). Since the central claim is based on parameters derived from this model, the poor fit is a load-bearing problem: the quoted M_sub and HSMR values are derived from a model that does not describe the data.","section":"§5, Table 1"},{"comment":"The mis-centring model is a single Rayleigh distribution with a free scale sigma, fitted independently in each bin. The best-fit sigma increases from 83.5 to 347.7 kpc across the three bins, and the outer two bins are inconsistent with the external Zhang et al. (2019) estimates by factors of ~1.6 and ~4 (Table 2). A large sigma smooths the negative host-halo trough and can be partially degenerate with a larger M_sub, especially in the outer bin. The authors acknowledge the divergence but do not quantify how much of the HSMR trend is driven by this freedom. A concrete test would be to impose a prior on sigma from Zhang et al. (2019) or a more flexible mis-centring distribution and show the effect on M_sub.","section":"§2.2, Eq. (6), Table 2"},{"comment":"The paper's own alternative 'admixture' model fits the data substantially better (reduced chi-square 1.72, 1.17, 1.76) and yields HSMR values (4.0, 13.4, 32.1) that are a factor of ~1.6–3 lower than the primary model (11.99, 33.68, 53.02) in the corresponding bins. This demonstrates that the reported HSMR amplitudes—and by extension the quantitative strength of the 'strong positive trend'—are strongly model-dependent. The trend itself persists in the admixture model, but the paper does not provide a model-comparison or a formal test of whether the trend is significant under both models. The central claim of a 'strong' trend is therefore not robust to reasonable changes in the host-halo treatment.","section":"Appendix B, Table B1"},{"comment":"The two-halo term is omitted but is expected to be non-negligible beyond ~0.6 Mpc, as the authors state in Section 2 and discuss in Section 6. The outermost cluster-centric bin (0.86–1.29 Mpc) is exactly where such a term matters. An omitted positive large-scale contribution can bias the fitted host-halo amplitude A and, through degeneracy with M_sub, bias the outer-bin HSMR. Since the two-halo term is a known, physical contribution rather than a nuisance, the analysis should either include it or restrict the fitting range to scales where it is negligible. At minimum, the paper should quantify the systematic shift in M_sub and HSMR when a two-halo term is added.","section":"§2, §6"}],"minor_comments":[{"comment":"The phrase 'confirm tidal stripping' is too strong given the modelling caveats; the analysis shows consistency with tidal stripping, not a direct confirmation. Consider softening to 'are consistent with'.","section":"Abstract"},{"comment":"Typo: 'partially loosing mass' should be 'losing mass'.","section":"§1"},{"comment":"Typo: 'the former is is easier'.","section":"§2.1"},{"comment":"In the discussion of the common-halo test, the sigma values are reported with asymmetric errors that are not all in the same convention; ensure consistency and state the median and 16th/84th percentiles throughout.","section":"§6"},{"comment":"The column header 'Pmem' is described in the caption as 'median cluster membership probability', but the text in Section 3.1 refers to a threshold of 0.8. Clarify the definition and whether the values in the table are medians of the lens sample.","section":"Table 1"}],"recommendation":"major_revision","confidential_remarks":"The paper has a valuable data set and an important question, but the current version overclaims. The primary model is rejected by the data, and the paper's own appendix shows that reasonable changes to the host-halo model change the HSMR values by factors of 2–3. The trend with radius may survive, but the analysis currently does not establish the quantitative 'strong positive trend' claimed. I would ask for a revised version that (i) includes the two-halo term or justifies its omission quantitatively, (ii) tests sensitivity to the mis-centring model, and (iii) reframes the conclusions to match the model-dependence. If the authors prefer to keep the simple model, they need to demonstrate that the HSMR trend is not driven by the sigma freedom."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The thing to know: this is the largest satellite-lensing sample ever used for this measurement (~330k lenses, 4.5M sources), and the trend they report — HSMR rising with cluster-centric radius — is probably real. But the model they use to extract the masses fails the data, so the headline numbers should not be taken at face value.\n\nWhat's new: the UNIONS data set and the statistical power. The method is basically Li et al. (2014/2016) with a free Rayleigh mis-centering scale, so the novelty is the dataset, not the technique. The trend is consistent with prior work, which is good confirmation.\n\nThe soft spots are real and central. The best-fit model has reduced chi-square of 3.6, 2.8, and 3.4 for 11 dof, which is a clear statistical rejection. The authors acknowledge under- and over-predicted regions, including an unexplained hump near 0.5 Mpc in the outer bin. The mis-centering scale sigma rises from 83 to 348 kpc and disagrees strongly with Zhang et al. (2019) in the outer two bins. A large sigma smears the host-halo term and can trade against Msub, inflating the outer-bin HSMR. The omitted two-halo term, which they mention, goes in the same direction. So the masses are degenerate with the host-subtraction model.\n\nTo their credit, they include Appendix B, an admixture model that fits much better (chi2_red 1.2–1.8) and gives lower HSMR values (4, 13, 32 vs. 12, 34, 53) — the trend survives but the amplitude shifts by a factor of 2–3. They also compare their sigma to external X-ray estimates and flag the discrepancy. That is honest and useful.\n\nWho this is for: anyone working on satellite lensing or cluster subhalo physics. The dataset is a resource, and the trend is a solid confirmation, but the quoted masses are not yet trustworthy. A revision with a two-halo term, a more physical mis-centering prior, and systematic-error accounting would make this a strong paper.\n\nRecommendation: send it to peer review, not desk reject. The sample is unique, the analysis is careful, and the authors are transparent. But expect major revision before the masses can be used as measurements.","headline":"A large-sample UNIONS measurement confirms the tidal-stripping trend in direction, but the fitting model is formally rejected by the data, so the quoted HSMR amplitudes are model-dependent.","tokens_in":17888,"tokens_out":2621,"would_cite":false,"duration_ms":30120,"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":"This paper shows that satellite galaxies in clusters lose dark matter as they fall inward: the subhalo-to-stellar mass ratio rises from about 12 near the core to 53 in the outskirts.","keywords":["galaxy clusters","satellite galaxies","subhalo masses","tidal stripping","weak gravitational lensing","galaxy-galaxy lensing","excess surface density","dark matter"],"falsifier":"Re-fit the same lensing signal for the outer bins using cluster centres from X-ray observations or including a two-halo term; if the subhalo-to-stellar mass ratios in those bins drop to the inner-bin value, the reported tidal-stripping trend is an artifact of the offset-host model. The paper's own alternative two-centre host model gives ratios of 4, 13, and 32, so the trend's amplitude is sensitive to host-offset modelling.","tokens_in":16829,"feed_emoji":"🔭","tokens_out":8272,"duration_ms":77462,"temperature":0.7,"pith_summary":"The paper tries to establish that satellite galaxies lose dark matter, but not stars, as they fall into galaxy clusters. It stacks weak-lensing measurements around roughly 330,000 cluster satellites and fits a model with a subhalo, a mis-centred host halo, and a stellar point mass. The resulting subhalo-to-stellar mass ratio is 11.99, 33.68, and 53.02 in the three cluster-centric bins, a strong positive radial trend. If correct, this is direct observational confirmation of tidal stripping and supports the hierarchical picture in which dark halos are shredded while the galaxies' stellar bodies survive.","feed_headline":"Satellite dark-matter halos shrink toward cluster centres","feed_subtitle":"Stacked weak lensing of 330,000 cluster satellites sees subhalo-to-stellar mass rise from 12 to 53.","key_machinery":"The load-bearing instrument is the combined excess surface density model ΔΣ(R) = A·ΔΣ_host + ΔΣ_sub + ΔΣ_*, applied to stacked lensing. The host-halo term is generated by Monte-Carlo point masses following an NFW profile centred on a brightest cluster galaxy whose offset from the catalogue position is drawn from a Rayleigh distribution with a fitted scale σ; the subhalo term is a projected NFW profile with free mass M_sub; the stellar term is a point mass. This decomposition isolates the subhalo's own lensing contribution from the dominant cluster halo, which is what allows a mass to be assigned to the satellite halos.","core_discovery":"The paper's central claim is that the dark-matter halos of cluster satellites are progressively stripped toward the cluster centre, and that this can be measured cleanly in stacked galaxy-galaxy lensing. From the excess surface density signal it infers median subhalo masses log M_sub/M_sun = 12.02, 12.45, and 12.58 in the 0.143-0.43, 0.43-0.86, and 0.86-1.29 Mpc bins, giving subhalo-to-stellar mass ratios of 11.99, 33.68, and 53.02. The monotonic increase is the result the paper wants to be judged on; it is presented as confirmation that tidal stripping or similar environmental processes remove dark matter from infalling galaxies while their stellar mass remains intact.","pith_inferences":["Inference: a natural next split is by satellite stellar mass or host richness; if stripping efficiency depends on mass, low-mass satellites should show a steeper HSMR trend, testable with the same lens-source pairs.","Inference: the fitted Rayleigh σ growing with cluster-centric radius may be absorbing a missing two-halo term or an anisotropic mis-centring distribution rather than a true physical offset; comparing with X-ray-centred cluster samples would discriminate.","Inference: if the trend survives such tests, stellar-mass-based halo estimates for cluster members are environmentally biased, and cluster-centric radius should enter abundance-matching calibrations for satellites."],"forward_implications":["If the trend is real, an inner satellite carries roughly a quarter of the dark-matter-to-stellar mass ratio of an outer satellite, quantifying how much dark matter cluster tides remove.","The result confirms the standard hierarchical picture: satellite galaxies' dark halos are stripped preferentially near the cluster centre while their stellar mass stays intact.","With roughly 330,000 lens satellites and 4.5 million source galaxies, the inner-bin measurement is far tighter than earlier work and anchors the radial subhalo-to-stellar mass relation at low radius."],"fun_headline_variants":["Satellite halos lose dark matter toward cluster cores","Subhalo-to-stellar mass ratio climbs steeply with radius","Cluster satellites shed dark matter toward core","Weak lensing reveals satellite halo stripping in clusters","Satellite dark halos shrink toward cluster centres"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The host-halo subtraction assumes the true cluster centre is offset from the catalogued brightest cluster galaxy by a single Rayleigh distribution with a fitted scale per bin; the paper itself notes (Section 6) that this scale disagrees with X-ray-based estimates in the outer bins and that a two-halo term is omitted beyond roughly 0.6 Mpc. If the offset model is wrong, the inferred subhalo masses and the radial trend are biased.","fun_headline_variants_meta":{"raw":{"variants":["Satellite halos lose dark matter toward cluster cores","Subhalo-to-stellar mass ratio climbs steeply with radius","Cluster satellites shed dark matter toward core","Weak lensing reveals satellite halo stripping in clusters","Satellite dark halos shrink toward cluster centres"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001531,"raw_usage":{"total_tokens":5996,"prompt_tokens":806,"completion_tokens":5190,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":550,"completion_tokens_details":{"reasoning_tokens":5116}},"tokens_in":550,"tokens_out":5190,"duration_ms":32858,"temperature":1.0,"reasoning_tokens":5116,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-02T02:45:22.222813+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Re-fit the same lensing signal for the outer bins using cluster centres from X-ray observations or including a two-halo term; if the subhalo-to-stellar mass ratios in those bins drop to the inner-bin value, the reported tidal-stripping trend is an artifact of the offset-host model. The paper's own alternative two-centre host model gives ratios of 4, 13, and 32, so the trend's amplitude is sensitive to host-offset modelling.","supporting_citations":[],"review_version":1}