{"id":"1d320238-d289-4cd5-91d7-d753a4906b93","arxiv_id":"1908.10375","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":6,"one_line_summary":"Using a combined 1,231-SN catalog and two light-curve fitters, the authors confirm that Type Ia supernovae in low-mass, star-forming hosts are about 0.06 mag fainter, a known effect that biases cosmology if uncorrected.","lead":"This paper compiles a new catalog of 1,231 Type Ia supernovae from multiple surveys and finds that those in low-mass, star-forming host galaxies are about 0.06 magnitudes fainter after standard light-curve corrections, with a larger difference of about 0.08 magnitudes when using inferred local environments. The result is an independent, larger-sample confirmation of a known environmental effect that affects cosmic distance measurements and dark energy constraints.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Local-environment result is a global-property proxy: §2.5.1 defines locally star-forming as low-mass, globally star-forming hosts and locally passive as all globally passive hosts, so the 0.08 mag 'local' step is not a direct measurement at the SN site.","rationale":"The paper's central global-host-mass result is solid: a 0.057–0.065 mag step at 4σ with two fitters, consistent with many prior studies. That part does not depend on the local proxy. The local-environment claim, however, is the paper's distinctive contribution and the basis for the age/luminosity-evolution interpretation. The reader's weakest assumption correctly identifies the proxy as the fragile link. I agree with the conditional verdict: the concern can be settled by direct local measurements on an overlapping sample, but until then the local numbers should not be read as direct evidence. I also note the abstract's global-property numbers (0.062±0.009, 0.057±0.010) do not match Table 7's mass/sSFR rows; this strengthens the case for a careful revision but is secondary to the proxy issue.","tokens_in":43179,"tokens_out":8273,"duration_ms":86646,"concrete_test":"For the subset of YONSEI SNe with direct local classifications (e.g., local specific star formation or Hα from Rigault et al. 2018, Roman et al. 2018, or the Kim et al. 2018 validation), build the 2×2 confusion matrix between the Kim et al. proxy classes and the direct classes. Weight the Hubble residuals by the inverse misclassification probabilities and recompute the local HR difference; if the corrected difference is no longer significantly larger than the global sSFR difference (0.049/0.033 mag), the claimed 0.08 mag local effect is not established.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section 2.5.1 states that local environments are not measured but inferred from Kim et al. (2018): a SN is placed in a locally star-forming environment only if its host is globally star-forming and low-mass (log Mstellar < 10), and every globally passive host is assigned to the locally passive class. Consequently, the local-environment rows of Table 7 (0.081±0.018 mag for SALT2, 0.072±0.018 for MLCS2k2) effectively compare low-mass star-forming galaxies with a passive-host sample that is dominated by high-mass galaxies; they are a re-labelled global host-property split rather than a measurement of conditions at the explosion site. Section 3 explicitly points to Kim et al. (2018) for the local analysis and presents no local star-formation indicator (Hα, local sSFR, or local U−V) here. If the proxy misclassifies, for example, SNe in high-mass star-forming hosts that explode in locally passive regions, or low-mass star-forming hosts with locally passive sites, the quoted local HR differences and the 'more direct link to the progenitor' argument do not follow. Because the abstract's final claim that the origin is 'most likely luminosity evolution' is built on interpreting the local-environment difference as a stellar-population-age difference, the local step is load-bearing despite the robust global host-mass correlation.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"Kim, Kang, and Lee construct the YONSEI Type Ia supernova catalog from 1521 light curves (1231 spectroscopically confirmed SNe Ia after cuts) fitted with both SALT2 and MLCS2k2, match 674 hosts to the stellar mass and specific star formation rate data of Kim et al. (2018), and study Hubble residuals as a function of host stellar mass, global sSFR, morphology, and an inferred local environment. They report that SNe Ia in low-mass, star-forming hosts are about 0.06 mag fainter than those in high-mass, passive hosts, and that this difference grows to about 0.08 mag when using their local-environment classification. They interpret the result as evidence that the environmental dependence has a stellar-population-age origin and therefore implies luminosity evolution of SNe Ia with redshift.","tokens_in":43442,"tokens_out":5094,"duration_ms":51487,"significance":"If the global host-mass step is taken at face value, the paper provides a useful independent confirmation of the well-known environmental dependence using a combined low-, intermediate-, and high-redshift sample with two independent light-curve fitters. The rms-scatter results (Table 7), which suggest that SNe Ia in low-mass and star-forming environments are more homogeneous standard candles, are interesting and worth reporting. The catalog itself is a community resource. However, the headline local-environment numbers are not direct measurements of the SN explosion site: they are a re-labelling of global host properties, as the authors state in Section 2.5.1. Until actual local environment indicators (e.g., H-alpha, local sSFR, or local U-V) are used, the 0.08 mag local step and the age/luminosity-evolution conclusion are not separately established beyond the global step. The analysis is traceable to public packages (SNANA and the JLA likelihood), but no machine-checked code is released with the paper.","major_comments":[{"comment":"The 'local environment' classification is an empirical proxy, not a measurement at the SN site. A SN is assigned to a locally star-forming environment only if its host is globally star-forming and low-mass (log M_stellar < 10), and every globally passive host is assigned to the locally passive class. No local star-formation indicator (H-alpha, local sSFR, or local U-V) is used in this paper. Therefore the quoted differences of 0.081 +/- 0.018 mag (SALT2) and 0.072 +/- 0.018 mag (MLCS2k2) in Table 7 are not direct measurements of the explosion-site conditions; they effectively compare a low-mass, globally star-forming sample with a high-mass-dominated passive sample. The abstract's claim that these are obtained 'when only local environments of SNe Ia are considered' is not supported by the analysis, and the Section 5.1 argument that the local result is 'more directly linked to the SN progenitor' does not follow from these data. Please either use actual local environment measurements or explicitly present the result as a global-property proxy with an associated systematic uncertainty.","section":"Section 2.5.1 and Table 7"},{"comment":"The error budget uses only statistical uncertainties, as stated in Section 2.2.3. Systematic uncertainties from photometric calibration, light-curve model training, host-galaxy SED fitting, Malmquist bias corrections, and the adopted classification boundaries (log M = 10.0 and log sSFR = -10.4) are not propagated into the reported HR differences. The 3-5 sigma significances in Tables 7 and 8 are therefore optimistic; a systematic error of even 0.02-0.03 mag on the step would substantially reduce the significance of the sSFR split. Please provide a systematic error budget for the main HR differences, or state explicitly in the abstract and conclusions that the quoted significances are statistical-only.","section":"Section 2.2.3, Eq. (2)"},{"comment":"The Hubble residuals are computed with alpha, beta, M_B, and Omega_M fitted to the same YONSEI sample, so the light-curve standardization is partly self-referential. This does not force a residual host-mass step, and the authors are careful to note that the residual dependence is not an artifact of the fit. However, the uncertainties on alpha, beta, and M_B are not propagated into the HR differences, and if an environmental step exists, the simultaneous fit of alpha and beta can partially absorb it. As a robustness test, please re-fit alpha and beta on a low-redshift subsample or with an explicit host-mass step term in the distance model, and show that the reported HR differences and significances are stable.","section":"Section 2.4.1 and Eq. (1)"},{"comment":"The LOWZ subsample shows no host-mass step for SALT2 (0.002 +/- 0.072 mag) and only a 1.8-sigma step for MLCS2k2, with a heavily mass-targeted sample (12 low-mass versus 76 high-mass hosts). The claim that the dependence is a 'global phenomenon over the whole redshift range' therefore rests almost entirely on the SDSS and SNLS panels. Please either add a joint statistical test with survey-by-survey offsets or soften the claim to state that the step is robust at intermediate and high redshift but unconstrained in this particular low-redshift sample.","section":"Table 8 and Section 3.2.1"}],"minor_comments":[{"comment":"The abstract and Section 5 quote 0.062 +/- 0.009 mag (SALT2) and 0.057 +/- 0.010 mag (MLCS2k2) for low-mass and star-forming hosts, but Table 7 gives 0.057 +/- 0.014 and 0.065 +/- 0.015 for the mass split, and 0.049 +/- 0.015 and 0.033 +/- 0.016 for the sSFR split; please clarify how the abstract values are derived.","section":"Abstract and Section 5"},{"comment":"The sample counts for the combined host sample do not match between Table 5 (657 for SALT2 mass and sSFR, 373 for local environment) and Table 7 (648 for mass, 649 for sSFR, 368 for local); please reconcile the definitions and counts.","section":"Tables 5 and 7"},{"comment":"There are several typos: 'wih' and 'establisehd' in Section 3.2.1, and 'blaek dashed lines' in the Figure 11 caption.","section":"Section 3.2.1 and Figure 11 caption"},{"comment":"The dates 'Received February 30, 2019; accepted February 31, 2019' are not valid calendar dates; please correct them.","section":"First page"},{"comment":"The MLCS2k2 fit reports H0 = 63; since the analysis uses distance moduli, the meaning and role of this H0 value should be stated explicitly or removed.","section":"Section 2.4.1"}],"recommendation":"major_revision","confidential_remarks":"The manuscript overlaps substantially with the authors' earlier Kim et al. (2018), reusing the same host-galaxy data and the same local-environment inference method. The present submission does add the combined YONSEI catalog and the two-fitter comparison, but the incremental novelty relative to Kim et al. (2018) should be clarified in the cover letter or in the text, especially because the abstract presents the local-environment result as a new measurement. The global host-mass step is robust and worth publishing, but the local-environment interpretation needs reframing or direct local data before the paper can be accepted."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Bottom line: this is a large, careful confirmation of a result the field already had. The ~0.06 mag host-mass step in SN Ia luminosity is robust here, consistent with Betoule and Sullivan, and the paper is honest about being an independent confirmation rather than a new discovery. The genuinely new pieces are the combined spectroscopically confirmed sample (1231 SNe, two fitters) and the first look at MLCS2k2 AV trends with host mass and sSFR. The catalog itself, even if currently available only on request, is a resource.\n\nThe soft spots are in the 'local environment' part. Section 2.5.1 defines locally star-forming as low-mass globally star-forming hosts, and locally passive as all globally passive hosts. That is a re-labelling of global host properties, not a measurement at the SN site. The abstract's 0.081/0.072 mag local step and the claim that this is 'most likely luminosity evolution' lean on that proxy being local. The body is more careful and points to Kim et al. (2018) for the method, but the framing in the abstract and discussion overstates what was actually measured. If the proxy misclassifies SNe in high-mass star-forming hosts with locally passive sites, or low-mass hosts with locally passive sites, the local step is not a direct measurement. This matters because the luminosity-evolution conclusion is built on interpreting the local difference as an age difference.\n\nOther issues are more minor. Only statistical errors are propagated; that is a limitation, but the step is large enough that systematics would have to be severe to erase it. The Hubble residuals use alpha, beta, and M_B fitted to the same sample; that is standard practice and does not force the residual dependence, so I do not see it as a real circularity problem. The catalog not being public is a practical annoyance for a paper whose main product is the catalog. Self-citation is not a flaw here because the host data and local method genuinely come from Kim et al. (2018), and the paper says so.\n\nWho is this for? SN cosmologists who want a consolidated confirmation and a comparison table of environmental steps across fitters, and anyone working on the origin of the mass step. It deserves peer review; the central claim holds. The revision should rename or clearly qualify the inferred environments, and the abstract should not present the local result as more direct than it is.","headline":"Careful large-sample confirmation of the known SN Ia host-mass step; the 'local-environment' result is a global-property proxy and the abstract oversells it.","tokens_in":44093,"tokens_out":3145,"would_cite":false,"duration_ms":34151,"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":"Type Ia supernovae in low-mass, star-forming host galaxies are intrinsically fainter by $0.06$-$0.08$ mag after standard light-curve corrections, and this offset most plausibly reflects progenitor age and hence luminosity evolution with…","keywords":["Type Ia supernovae","host galaxy stellar mass","specific star formation rate","Hubble residuals","luminosity evolution","SALT2","MLCS2k2","YONSEI catalog"],"falsifier":"Take a sample of low-redshift Type Ia supernovae with directly measured local star-formation activity at the explosion site (e.g., H-alpha or UV flux in a few-kiloparsec aperture) and compare the Hubble-residual offset for locally star-forming versus locally passive sites with the offset predicted by the global-host proxy; if the directly measured local offset is not larger than the global-host offset, the paper's central environmental claim loses its local-enhancement component, and if age-matched hosts show no residual offset, the age-evolution interpretation loses its basis.","tokens_in":42965,"feed_emoji":"💥","tokens_out":10400,"duration_ms":91854,"temperature":0.7,"pith_summary":"This paper builds a large, combined catalog of 1231 spectroscopically confirmed Type Ia supernovae and their host galaxies, fitted with two independent light-curve codes, and uses it to ask whether the supernovae's corrected brightness still depends on where they explode. It finds that supernovae in low-mass, star-forming host galaxies are $0.062\\pm0.009$ mag fainter (SALT2) and $0.057\\pm0.010$ mag fainter (MLCS2k2) than those in high-mass, passive hosts after standard shape and color corrections. When the comparison is restricted to the supernova's local environment rather than the whole galaxy, the offset grows to about $0.08$ mag. Because these environments differ sharply in stellar age, the paper takes this as evidence that the remaining offset is an intrinsic, age-driven luminosity evolution of Type Ia supernovae with redshift, not a correctable nuisance. The claim matters because if it is right, standard light-curve standardization leaves a systematic bias in supernova distances that must be understood before supernovae can anchor cosmology.","feed_headline":"Type Ia supernovae in young galaxies are 0.06–0.08 mag fainter","feed_subtitle":"If true, standard distance corrections miss an environment bias that shifts with cosmic time.","key_machinery":"The engine of the measurement is the Hubble residual, $\\mathrm{HR}\\equiv\\mu_{\\mathrm{SN}}-\\mu_{\\mathrm{model}}(z)$, the difference between the distance modulus a supernova implies and the distance its redshift predicts in the best-fit cosmological model; a positive residual means the supernova is fainter than the model expects. The paper bins these residuals by host stellar mass (split at $\\log M_\\star = 10.0$) and by specific star-formation rate (split at $\\log\\mathrm{sSFR}=-10.4$), with the same light-curve fits run through both SALT2 and MLCS2k2. To extend local-environment information to high redshift, it uses an empirical proxy: locally star-forming sites are selected as supernovae whose hosts are globally star-forming and low-mass, while globally passive hosts are assumed to be locally passive. The age interpretation rests on the known contrast in mean stellar population age between these bins.","core_discovery":"On its own terms, the paper's discovery is that the standardized peak luminosity of a Type Ia supernova still carries a memory of its host environment: after fitting with SALT2 and MLCS2k2, the Hubble residual (the difference between the supernova's distance and the distance its redshift predicts) is larger, meaning the supernova is fainter, in low-mass and star-forming hosts than in high-mass and passive hosts. The catalog yields a global-host difference of $0.062\\pm0.009$ mag (SALT2) and $0.057\\pm0.010$ mag (MLCS2k2); using the inferred local environment at the explosion site raises these to $0.081\\pm0.018$ mag and $0.072\\pm0.018$ mag. The authors interpret the larger local offset, together with known stellar-age differences between the environments, as evidence that progenitor age, not merely dust or calibration, drives the effect, so that the mean intrinsic luminosity of Type Ia supernovae evolves over cosmic time as host populations age.","pith_inferences":["If the age interpretation is right, high-redshift supernova samples should appear systematically fainter after standardization by roughly this $0.06$-$0.08$ mag offset, which could partially mimic dark energy; a clean test would compare Hubble residuals of supernovae at fixed redshift whose hosts have spectroscopically measured Balmer ages.","Because the paper's offset is measured on a sample spanning redshifts up to $0.85$, applying it to the low-redshift distance-ladder anchors would shift $H_0$ by a few percent, a change comparable in size to the current Hubble tension; the paper does not perform that application.","Because the local-environment proxy uses only global host properties, direct measurement of local star-formation activity at the explosion site (for example H-alpha flux in a few-kiloparsec aperture) for a subset of the sample would determine whether the $0.08$ mag local offset is a real environmental effect or a global-host selection effect."],"forward_implications":["If the offset is real, standard SALT2 and MLCS2k2 distances are systematically biased by about $0.05$-$0.08$ mag depending on host environment, a shift that is not removed by the usual shape and color corrections.","Cosmological fits that ignore environment absorb this as altered nuisance parameters; the paper shows $\\beta$ and $\\sigma_{\\mathrm{int}}$ shift between environment-binned fits, and previous work it cites translates this into a roughly 10% shift in $w$ and a 3.3% correction to $H_0$.","Supernovae in low-mass, star-forming hosts show smaller rms scatter (up to about 18% smaller) and smaller intrinsic scatter, meaning environment-selected subsamples are more homogeneous distance indicators.","If age is the origin, the mean standardized luminosity of Type Ia supernovae should drift with redshift as the universe's host population ages, adding a systematic component to dark-energy measurements that is strongest at $z>1$.","The agreement between a high-redshift-trained fitter (SALT2) and a low-redshift-trained fitter (MLCS2k2) indicates the effect is not an artifact of the fitter's training-set redshift mixing."],"supporting_citations":[{"why":"Supplies the host stellar masses, specific star-formation rates, and the global-to-local environment proxy used to define locally star-forming and locally passive samples.","marker":"Kim et al. (2018)"},{"why":"Supplies the mass and sSFR split criteria and the earlier SNLS-based measurement of the Hubble-residual offset that this paper reproduces.","marker":"Sullivan et al. (2010)"},{"why":"Provides the SALT2 light-curve fitter whose shape and color parameters define the SALT2 Hubble residuals.","marker":"Guy et al. (2007)"},{"why":"Provides the MLCS2k2 fitter and its low-redshift training, the second independent standardization used throughout.","marker":"Jha et al. (2007)"},{"why":"Supplies the bias corrections, JLA likelihood, and cosmology-cut criteria used to construct the Hubble diagram.","marker":"Betoule et al. (2014)"},{"why":"Provides Balmer-line stellar population ages of early-type hosts that ground the paper's age-based origin argument.","marker":"Kang et al. (2016)"},{"why":"Demonstrates that globally passive hosts are locally passive, a premise of the local-environment proxy.","marker":"Rigault et al. (2013)"},{"why":"Is the modern comparison showing the residual offset persists even after explicit mass and bias corrections, which the paper uses to argue for a physical origin.","marker":"Scolnic et al. (2018)"}],"fun_headline_variants":["Supernovae in low-mass star-forming hosts are 0.06 mag fainter","Host galaxy environment shifts supernova brightness by 0.06 mag","Supernova luminosity depends on host galaxy age, new catalog shows","YONSEI catalog reveals SN Ia brightness varies with host galaxy"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The inference that the enlarged local-environment offsets are real depends on the assumption that a supernova's immediate explosion-site environment can be inferred from its host galaxy's total mass and star-formation rate, with locally star-forming sites identified only in low-mass, globally star-forming hosts and globally passive hosts taken as locally passive.","fun_headline_variants_meta":{"raw":{"variants":["Supernovae in low-mass star-forming hosts are 0.06 mag fainter","Host galaxy environment shifts supernova brightness by 0.06 mag","Supernova luminosity depends on host galaxy age, new catalog shows","YONSEI catalog reveals SN Ia brightness varies with host galaxy"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00093,"raw_usage":{"total_tokens":4042,"prompt_tokens":1067,"completion_tokens":2975,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":683,"completion_tokens_details":{"reasoning_tokens":2897}},"tokens_in":683,"tokens_out":2975,"duration_ms":23212,"temperature":1.0,"reasoning_tokens":2897,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T10:46:04.354133+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take a sample of low-redshift Type Ia supernovae with directly measured local star-formation activity at the explosion site (e.g., H-alpha or UV flux in a few-kiloparsec aperture) and compare the Hubble-residual offset for locally star-forming versus locally passive sites with the offset predicted by the global-host proxy; if the directly measured local offset is not larger than the global-host offset, the paper's central environmental claim loses its local-enhancement component, and if age-matched hosts show no residual offset, the age-evolution interpretation loses its basis.","supporting_citations":[{"cited_title":"2018, Environmental Dependence of Type Ia Supernova Lumi- nosities from a Sample without a Local-Global Difference in Host Star Formation, ApJ, 854, 24","cited_arxiv_id":null,"evidence_quote":"Supplies the host stellar masses, specific star-formation rates, and the global-to-local environment proxy used to define locally star-forming and locally passive samples."},{"cited_title":"A., et al","cited_arxiv_id":null,"evidence_quote":"Supplies the mass and sSFR split criteria and the earlier SNLS-based measurement of the Hubble-residual offset that this paper reproduces."},{"cited_title":"2007, SALT2: Us- ing Distant Supernovae to Improve the Use of Type Ia Supernovae as Distance Indicators, A&A, 466, 11","cited_arxiv_id":null,"evidence_quote":"Provides the SALT2 light-curve fitter whose shape and color parameters define the SALT2 Hubble residuals."},{"cited_title":"G., & Kirshner, R","cited_arxiv_id":null,"evidence_quote":"Provides the MLCS2k2 fitter and its low-redshift training, the second independent standardization used throughout."},{"cited_title":"2016, Early-Type Host Galaxies of Type Ia Supernovae","cited_arxiv_id":null,"evidence_quote":"Provides Balmer-line stellar population ages of early-type hosts that ground the paper's age-based origin argument."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Demonstrates that globally passive hosts are locally passive, a premise of the local-environment proxy."},{"cited_title":"M., Jones, D","cited_arxiv_id":null,"evidence_quote":"Is the modern comparison showing the residual offset persists even after explicit mass and bias corrections, which the paper uses to argue for a physical origin."}],"review_version":1}