{"id":"cc1ecca6-b116-44f7-b522-57278a2466af","arxiv_id":"2608.13313","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":7,"one_line_summary":"A homogeneous catalog of mean reddening, differential reddening, and color excess ratios for 729 open clusters shows a median extinction law consistent with R_V=3.1 and large-scale variations across the Galactic disk.","lead":"Using Gaia and 2MASS data for thousands of member stars, the authors measured how much dust reddens and dims starlight for 729 galactic open clusters, including the spread of reddening across each cluster. They find that the average reddening law matches the standard diffuse interstellar medium curve, but the dust law varies across the Galaxy.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"CER scale and longitude trend may be biased by unvalidated intrinsic-color model systematics; an independent CER/R_V validation is missing.","rationale":"The paper is careful and the mean reddening scale is well supported by external comparisons. The reader's condition on the CMD-width diagnostic addresses a secondary claim; even if a null test were added, the central CER result would still lack an independent validation of the intrinsic-color model. Because both color excesses come from a single model, the CER is the quantity most sensitive to correlated model systematics. The available tests (alternative Gaia-XP-derived parameters, Teff shifts) do not break the shared-spectra correlation. A reanalysis with an independent color library is a direct, feasible check. The verdict should remain CONDITIONAL: accept once this validation (and ideally the code/data release) is provided. I agree partially with the reader: the CMD null test is worth running, but the more load-bearing gap is the unvalidated CER scale.","tokens_in":14822,"tokens_out":8755,"duration_ms":99604,"concrete_test":"Rederive cluster CERs for the same 729 clusters using an independent intrinsic-color estimator that is not trained on Gaia XP colors, e.g., PARSEC/COLIBRI isochrone colors computed from each member's SHBoost Teff, logg, and [M/H], and repeat the zero-intercept slope fit and the longitude binning of Figures 8 and 9. If the median k_oc and the quadrant/longitude pattern persist within the reported statistical plus systematic uncertainties, the CER claim is robust; if the trend changes sign or washes out, the central claim is not supported.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The CER measurement (Section 3.3) is the slope through the origin of E(BP-KS) against E(BP-RP). Both color excesses are computed from the same XGBoost intrinsic-color model applied to the same SHBoost parameters (Section 3.1). If the model has a Teff-dependent scale error in one predicted intrinsic color relative to the other, the two color-excess errors are correlated and the fitted CER is biased by an amount that depends on the stellar-parameter distribution of each cluster. The cluster sample spans different ages and distances, and the stellar content varies with Galactic longitude; the claimed quadrant dependence of k_oc (Section 4.4, Figures 8-9) could therefore be produced by the intrinsic-color model rather than by true extinction-law variations. The systematic checks in Section 3.3 (SHBoost vs GSP-Phot and ±100 K Teff shifts) use parameter sets derived from the same Gaia XP spectra, so they share common systematics and do not validate the intrinsic-color scale against an independent source. The bootstrap treats the two color-excess errors as independent although they share the same parameter errors. The median CER of 2.22 matching the R_V=3.1 bandpass value is a consistency check, not a validation; a Teff-dependent bias could still average to the correct median. In contrast to the mean reddening, which is compared with independent measurements in Section 4.2, the CER is only qualitatively compared with the Zhang & Green (2025) dust-map R_V. Thus the key new claim that the optical-to-NIR extinction behavior is not uniform across the disk is not yet secured.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents a homogeneous, member-based analysis of extinction properties for Galactic open clusters. Using Gaia DR3 and 2MASS photometry together with SHBoost stellar parameters, the authors derive per-star color excesses E(GBP-RP) and E(GBP-KS), then construct cluster-level mean reddening, differential reddening (MAD of the color excess distribution), and color excess ratios (CERs) defined as the zero-intercept slope of E(GBP-KS) versus E(GBP-RP). The catalog covers 729 clusters with at least ten members. The main results are that differential reddening increases with mean reddening, star-by-star dereddening narrows the CMD in 85% of 435 clusters with reliable CMD-width measurements, the median cluster CER is 2.22 in agreement with a standard R_V=3.1 extinction law, and the CER shows a large-scale variation with Galactic longitude, with lower values in the first and second quadrants and higher values in the third and fourth quadrants.","tokens_in":15125,"tokens_out":8312,"duration_ms":87366,"significance":"If the results are robust, this is a valuable contribution: it provides a uniform cluster-scale catalog of reddening, differential reddening, and CERs, with separate statistical and systematic uncertainties, and it connects cluster extinction measurements to large-scale extinction-law variations in the Galactic disk. The paper is notable for its large sample size, for cross-checking the mean reddening against several independent estimates (Wei et al. 2025; Cantat-Gaudin et al. 2020; Zhang and Green 2025), and for performing multiple robustness tests (quality-cut sensitivity, binary removal, Teff shifts, IR-excess removal). The central caveats are that the CMD-narrowing diagnostic is partly circular, and that the CER scale rests on color excesses derived from a single intrinsic-color model applied to Gaia-XP-based stellar parameters, so the claimed longitude dependence of the CER requires stronger validation before it can be regarded as a measurement of Galactic extinction-law variations.","major_comments":[{"comment":"The CMD-width diagnostic is not independent of the color excesses: the dereddened CMD is constructed by subtracting the very same per-star E(GBP-RP) values that were measured from those stars, so a reduction in scatter is partly guaranteed by construction even if the excesses were pure noise. The paper reports that 369 of 435 clusters (85%) show a decrease in CMD width, but it provides no null test showing that this fraction exceeds what would be obtained with random corrections of the same magnitude. Please add such a null test (for example, perturbing each star by random color excesses drawn from the measured distribution, or using a cross-validation in which the correction is estimated from an independent subset) and report the null fraction and its uncertainty. Without this, Figure 7 cannot support the claim that the measured reddening variations account for the observed CMD broadening.","section":"Section 4.3, Figure 7"},{"comment":"The two color excesses E(GBP-RP) and E(GBP-KS) are both computed from the same XGBoost intrinsic-color model evaluated on the same SHBoost stellar parameters, so their errors are correlated through the shared Teff, logg, and [M/H] uncertainties and through the intrinsic-color-model residuals. The bootstrap in Section 3.3 perturbs the two color excesses independently and therefore does not capture this covariance; this can bias the reported CER uncertainty and provides no test of how model covariance propagates into k_oc. In addition, the G_BP photometric error enters both color excesses. Please propagate the joint covariance of the stellar parameters and, if possible, of the intrinsic-color predictions into the CER fit, or at minimum report how k_oc and its uncertainty change when the two color excesses are perturbed in a fully correlated manner.","section":"Sections 3.1-3.3"},{"comment":"The systematic checks of the intrinsic-color scale compare SHBoost with GSP-Phot parameters and apply ±100 K Teff shifts, but both parameter sets are derived from Gaia DR3 XP spectra, so these tests share common systematics and do not validate the intrinsic-color scale against an independent source. Because CER is the ratio of two color excesses from the same model, a Teff-dependent scale error in one predicted intrinsic color relative to the other will bias k_oc by an amount that depends on the stellar-parameter distribution of each cluster. Since cluster age, distance, and hence stellar content vary with Galactic longitude, the claimed quadrant dependence of k_oc could in principle be produced by the intrinsic-color model rather than by true extinction-law variations. The median CER of 2.22 matching the bandpass-integrated R_V=3.1 value is a consistency check, not a validation, because a Teff-dependent bias could average to the correct median. Please validate the CER scale against independent reddening or stellar-parameter sources (e.g., spectroscopic Teff/logg from LAMOST or APOGEE, or independent color-excess measurements not based on Gaia XP), and test whether the longitude trend survives when the sample is matched or weighted by stellar-parameter distribution.","section":"Section 3.3 and Section 4.4, Figures 8-9"},{"comment":"The longitude dependence is the paper's main new physical claim, but it is currently supported only by a qualitative description and by a statement that the pattern is 'broadly consistent' with the Zhang and Green (2025) dust map. Please provide quantitative statistics: the significance of the difference between the first-second and third-fourth quadrant CER medians, the significance of the binned longitude trend, and a quantitative comparison with map-based R_V at cluster positions (e.g., a binned or per-cluster correlation). The weakening of the variation in the high-reddening subsample shown in Figure 9(b) should also be quantified, because it bears on whether part of the effect is a sample-selection effect.","section":"Section 4.4, Figure 9"},{"comment":"The observed CER distribution has a 16th-84th percentile width of about 0.23 mag, but each cluster's k_oc carries statistical and systematic uncertainties that broaden the observed distribution; the paper does not subtract them. Please report the intrinsic dispersion of the CER distribution after accounting for the measurement uncertainties. Otherwise the abstract's statement that the 'broad CER distribution' likely reflects differences in dust environments is not directly supported by Figure 8.","section":"Section 4.4, Figure 8"}],"minor_comments":[{"comment":"Please specify the exact fitting procedure for the zero-intercept CER (e.g., whether it is an errors-in-variables likelihood and how weights are defined), since the current description does not make clear how uncertainties in both variables enter the fit.","section":"Section 3.3"},{"comment":"The 435 clusters with reliable CMD-width measurements are not defined; please state the selection criteria (e.g., minimum number of stars per magnitude bin, minimum total N) and how they differ from the 729-cluster color-excess sample.","section":"Section 4.3"},{"comment":"The color scale in panel (b) is centered on the fiducial CER of 2.22; a diverging color map with explicit colorbar labels would make the spatial pattern easier to read.","section":"Figure 1"},{"comment":"The discussion of Figure 2(b) should state explicitly why the fitted slope of 2.073 is not used in preference to the literature conversion factor 2.394, given that the paper adopts the latter for converting E(GBP-RP) to AV.","section":"Section 4.2"},{"comment":"Please define the 'reddening gradient' E(GBP-KS)/d in the caption, including the adopted cluster distance source and the units.","section":"Figure 9"}],"recommendation":"major_revision","confidential_remarks":"This is a solid empirical catalog paper with a feasible revision path. The main risks are the circular CMD-narrowing diagnostic and the model-dependence of the CER scale; both are addressable with null tests and independent validation within the manuscript's scope. I do not see grounds for rejection, but the headline claims about the longitude dependence of CER should be either quantitatively supported or substantially softened. The citation of prior work appears appropriate."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear colleague,\n\nYou should know this paper: it builds a large homogeneous member-based catalog of reddening, differential reddening, and color-excess ratios for 729 open clusters. That catalog will be useful. The mean reddening scale checks out against three independent estimates, and the systematic tests are unusually thorough for this kind of work. But the headline longitude-dependent CER result is not yet on the same footing. The CER comes from a slope fit between two color excesses that both use the same XGBoost intrinsic-color model applied to the same SHBoost parameters. Correlated model errors can bias the slope in a Teff-dependent way, and the cluster sample varies in age and stellar content with longitude, so the quadrant pattern could be at least partially a model artifact rather than a real extinction-law variation. The paper's cross-checks use GSP-Phot parameters and ±100 K shifts, but those share the same Gaia XP spectral systematics; they do not validate the CER against an independent extinction-law measurement. The median CER matching R_V=3.1 is reassuring but not a validation.\n\nThe CMD-narrowing diagnostic also has a self-referential issue: you remove the derived color excesses from the same stars, so some narrowing is guaranteed. A null test with random perturbations would show whether the 85% figure exceeds chance. The paper does check that stronger differential reddening gives larger narrowing, which helps, but it is not conclusive.\n\nNone of this sinks the paper. The catalog, the reddening values, and the differential reddening measures are worth having, and the analysis is careful about uncertainties, binaries, and infrared excess. The soft spots are addressable: add a null test for the CMD diagnostic; validate the CER scale against an independent method; and release the code and processed data. I would send this to a serious referee, not desk reject it. Referees should focus on the CER bias question. I would probably cite the catalog for reddening, but I would not yet cite the longitude-dependent CER as established.","headline":"A large, well-tested open-cluster reddening catalog with an interesting but not yet validated CER longitude trend.","tokens_in":15685,"tokens_out":2076,"would_cite":true,"duration_ms":22591,"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":"This paper argues that per-star reddening corrections narrow the color–magnitude diagrams of 85% of open clusters and that the cluster color-excess ratio, while matching the standard R_V=3.1 extinction law on average, varies…","keywords":["open clusters","interstellar extinction","differential reddening","color excess ratio","extinction law","Gaia DR3","2MASS","Galactic disk"],"falsifier":"Randomly reassign the estimated color excesses among member stars within each cluster and repeat the CMD-width measurement: if the 85% narrowing fraction is reproduced under random assignment, the narrowing is an artifact of subtracting estimated color excesses, not a differential-reddening signal. A complementary check is to compare the per-star color excesses with independent sub-arcminute dust extinction maps: if the spatial pattern of the excesses does not track the map, the excesses are dominated by model or photometric noise rather than dust.","tokens_in":14592,"feed_emoji":"🌌","tokens_out":9369,"duration_ms":84809,"temperature":0.7,"pith_summary":"Using the same Gaia-based member lists that define the clusters, this paper derives star-by-star dust reddening for 729 Galactic open clusters and measures three quantities for each: mean reddening, differential reddening (how much reddening changes across the cluster), and the color excess ratio, or CER, which captures how strongly dust dims optical versus near-infrared light. It finds that differential reddening grows with total reddening, that correcting each member star individually narrows the cluster color–magnitude diagram in 369 of 435 clusters with reliable width measurements, and that the median cluster CER of 2.22 matches the value expected from the standard diffuse interstellar extinction law with R_V=3.1. The same CER, however, shifts with Galactic longitude, lower in the first and second quadrants and higher in the third and fourth, so the way dust reddens starlight is not uniform across the Milky Way's disk. If this is right, cluster ages and distances derived from broadband colors need direction-dependent extinction corrections, and member stars can map dust structure at scales that three-dimensional dust maps smooth away.","feed_headline":"Per-star dust corrections tighten 85% of cluster color diagrams","feed_subtitle":"Correcting each member star individually for dust narrows the main sequence in 369 of 435 open clusters and exposes small-scale dust…","key_machinery":"The load-bearing object is the per-star color excess, $E = (\\mathrm{observed\\ color}) - (\\mathrm{intrinsic\\ color})$, where the intrinsic color is predicted by a blue-edge-trained XGBoost model from SHBoost stellar parameters ($T_{\\mathrm{eff}}$, $\\log g$, $[\\mathrm{M/H}]$). Two excesses are used, $E(G_{BP}-G_{RP})$ and $E(G_{BP}-K_S)$. From them the cluster CER is the zero-intercept slope of $E(G_{BP}-K_S)$ versus $E(G_{BP}-G_{RP})$, a ratio that removes the absolute dust column and isolates the relative wavelength dependence of extinction; this is what lets the paper compare cluster photometry directly with bandpass-integrated extinction-curve predictions.","core_discovery":"The central claim is that a homogeneous, member-based analysis of open clusters can characterize extinction at cluster scale and reveal its spatial behavior. Per-star color excesses are computed as observed minus model-predicted intrinsic colors; the cluster CER is the zero-intercept slope of $E(G_{BP}-K_S)$ versus $E(G_{BP}-G_{RP})$, and its value across 600 quality-selected near-plane clusters has a median of 2.22, exactly matching the median 2.22 obtained by integrating a standard $R_V=3.1$ extinction curve through the Gaia and 2MASS passbands (2.15–2.27 at the 16th–84th percentiles). Differential reddening, measured as the median absolute deviation of member color excesses, increases with mean reddening according to $\\mathrm{MAD}(E_{BP-RP}) = 0.090\\,E(G_{BP}-G_{RP}) + 0.034$ mag, and star-by-star dereddening narrows CMD sequences in 85% of the 435 clusters with reliable width measurements. The CER varies systematically with Galactic longitude—lower in the first and second quadrants, higher in the third and fourth—which the paper interprets as differences in the dust environments integrated along different sight lines, not as local extinction-law variations at the clusters.","pith_inferences":["If the longitude-dependent CER pattern reflects true dust-environment differences, independent tracers such as Cepheids or red clump stars should show correlated CER signatures along the same sight lines; this is testable with existing data.","A natural check on the differential-reddening interpretation is to randomly permute the derived color excesses among member stars and re-measure the CMD-width narrowing fraction: if the 85% narrowing persists under random assignment, a large part of the signal comes from the self-consistent subtraction rather than from real dust variations.","The same per-star color excess method could be extended to bands beyond Gaia and 2MASS, such as WISE or optical narrow-band surveys, to break the degeneracy between CER and $R_V$ and to map grain-size variations across the disk.","Because the cluster CER is integrated over the full line of sight, splitting the sample into distance bins could reveal whether the quadrant contrast originates in nearby dust structures or is distributed along the Galactic disk."],"forward_implications":["Star-by-star reddening maps can recover intrinsic color–magnitude diagram morphology in clusters with strong differential reddening, improving isochrone-based ages and distances for individual clusters.","The increase of differential reddening with mean reddening implies that high-column sight lines near the Galactic plane are also the ones with the strongest small-scale extinction inhomogeneity.","Because the median CER matches the standard $R_V=3.1$ value but cannot distinguish $R_V$ values between 2.9 and 3.2, a single average CER should not be converted into a precise local $R_V$.","The longitude dependence of the CER means that extinction-law corrections for cluster and distance-scale work should be direction-dependent, and infrared-based measures are less affected than optical ones."],"supporting_citations":[{"why":"Supplies the 1481 open-cluster membership catalog on which the whole sample is built.","marker":"Cantat-Gaudin & Anders 2020"},{"why":"Provides the SHBoost stellar parameters ($T_{\\mathrm{eff}}$, $\\log g$, $[\\mathrm{M/H}]$) used to predict intrinsic colors for every member star.","marker":"Khalatyan et al. 2024"},{"why":"Trains the blue-edge XGBoost intrinsic-color model that converts stellar parameters into estimated intrinsic colors.","marker":"Zhao et al. 2024"},{"why":"Supplies the Gaia DR3 $G_{BP}$ and $G_{RP}$ optical photometry and astrometry for cluster members.","marker":"Gaia Collaboration et al. 2023"},{"why":"Supplies the 2MASS $K_S$ near-infrared photometry needed to form the $E(G_{BP}-K_S)$ color excess.","marker":"Skrutskie et al. 2006"},{"why":"Provides the reference conversion between $E(G_{BP}-G_{RP})$ and $A_V$ used in the catalog and comparison.","marker":"Wang & Chen 2019"},{"why":"Gives the three-dimensional dust map used for independent reddening comparison and for the quadrant $R_V$ pattern that supports the CER longitude interpretation.","marker":"Zhang & Green 2025"}],"fun_headline_variants":["Star-by-star dust removal tightens 85% of cluster diagrams","Cluster dust maps tighten 369 color-magnitude diagrams","Extinction law matches standard curve in open clusters","Differential reddening grows with total cluster dust","Per-star reddening fixes reveal extinction structure in clusters"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The central claim depends on the assumption that the star-to-star spread in the derived color excesses is real differential reddening, not scatter from photometric noise, intrinsic-color model errors, or the self-consistent correction that narrows color–magnitude diagrams by construction.","fun_headline_variants_meta":{"raw":{"variants":["Star-by-star dust removal tightens 85% of cluster diagrams","Cluster dust maps tighten 369 color-magnitude diagrams","Extinction law matches standard curve in open clusters","Differential reddening grows with total cluster dust","Per-star reddening fixes reveal extinction structure in clusters"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000962,"raw_usage":{"total_tokens":4122,"prompt_tokens":997,"completion_tokens":3125,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":613,"completion_tokens_details":{"reasoning_tokens":3045}},"tokens_in":613,"tokens_out":3125,"duration_ms":24624,"temperature":1.0,"reasoning_tokens":3045,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T13:54:50.982027+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Randomly reassign the estimated color excesses among member stars within each cluster and repeat the CMD-width measurement: if the 85% narrowing fraction is reproduced under random assignment, the narrowing is an artifact of subtracting estimated color excesses, not a differential-reddening signal. A complementary check is to compare the per-star color excesses with independent sub-arcminute dust extinction maps: if the spatial pattern of the excesses does not track the map, the excesses are dominated by model or photometric noise rather than dust.","supporting_citations":[],"review_version":1}