{"id":"2e73d318-cdda-417c-a8a2-2a48f375249d","arxiv_id":"2411.11986","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Type Ia supernovae in redder, brighter host environments have lower SALT2 stretch, and the slope of this relation differs between elliptical and disk hosts.","lead":"This study decomposes the host galaxies of 728 ZTF Type Ia supernovae into elliptical, disk, bulge, and bar components using 2D image fitting. It reports that supernovae in redder and denser stellar environments have narrower light curves, with the strongest relation in elliptical galaxies.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 16.8σ and 5.1σ colour–x1 relations may be inflated by the x1-dependent selection into the 728-galaxy sample; no completeness correction is presented.","rationale":"The reader's weakest-assumption analysis identifies the same load-bearing concern: the heavy selection from 3628 to 728 host galaxies is x1-dependent (Fig. 8) and correlated with galaxy properties that enter the environmental tracers, yet no selection correction is applied. The paper's defense that trends are rate-independent is incomplete because a correlation slope can be biased by differential selection even when rates are not the target. This is the single most load-bearing issue because all three headline results (elliptical 16.8σ, disk 5.1σ, and local surface brightness 6.1σ) are measured on this selected sample, and the physical interpretation ('age/metallicity effect') depends on the slopes being unbiased. The paper does provide useful supporting evidence: the code is public, Galaxy Zoo agreement is 90–92%, and the sample covers a broad x1 range. However, the absence of any completeness simulation or inverse-probability weighting leaves the central quantitative claims conditional. The arbitrary c=0.035 split for the 3.3σ dust trend is a secondary concern and does not change the verdict. Therefore the reader's CONDITIONAL verdict is appropriate and no adjustment is needed.","tokens_in":35724,"tokens_out":2321,"duration_ms":25087,"concrete_test":"Use the full ZTF DR2 sample to estimate the probability of entering the 728-galaxy final sample as a function of x1, SN-component colour, host stellar mass, and redshift (e.g., binned rates or logistic regression). Apply inverse-probability weights to the 728 retained SNe and recompute the weighted linear regressions and significances for the elliptical and disk colour–x1 relations. If the slopes and the 16.8σ/5.1σ significances survive reweighting, the selection-bias concern is retired; if they shrink materially, the headline claim is not established.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The final sample retains only 728 of 3628 SNe Ia (20%), and selection is not random with respect to the variables being correlated. Fig. 8 shows an excess of low-stretch (negative x1) SNe Ia and a deficit of high-stretch SNe Ia in the final sample, and Section 5.1 shows that fit success depends on host mass, redshift, and morphology (ellipticals are over-represented at 51% versus an expected ~37%). The paper argues in Section 5.1 that trends can still be studied because they do not depend on absolute or relative rates, but this is not sufficient: a slope within a subpopulation is biased if the selection probability depends jointly on x1 and the environmental tracer. For ellipticals, redder and more massive galaxies are easier to fit (higher surface brightness, more pixels), and low-x1 SNe Ia preferentially reside in redder ellipticals. The retained sample may therefore preferentially include low-x1 SNe in red galaxies and high-x1 SNe in blue galaxies, steepening the observed 16.8σ relation and the disk 5.1σ relation. No simulation, weighting, or selection correction is provided to show the slopes and significances survive. The c=0.035 split for the dust trend is separately arbitrary, but the selection issue is the load-bearing threat to the headline claim.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper uses 2D image decomposition of DESI-LS images to model 728 host galaxies of SNe Ia from ZTF DR2, separating elliptical, disk, bulge/bar, and halo environments. The central results are linear correlations between SALT2 stretch x1 and the model-derived galaxy component colour: a 16.8σ relation for SNe Ia in ellipticals and a 5.1σ relation for SNe Ia in disks, with lower-stretch SNe Ia in redder environments, plus a 6.1σ trend between x1 and local r-band surface brightness in disk hosts. The authors interpret these as age/metallicity effects and argue that the slopes differ fundamentally between old and young stellar environments. They also report a 3.3σ possible dust effect in SALT2 c versus local surface brightness for disk SNe, based on a c>0.035 split, and confirm that 91bg-like SNe reside in the reddest old environments.","tokens_in":36006,"tokens_out":4734,"duration_ms":52468,"significance":"If the correlations are genuine, this is an important step in connecting SN Ia light-curve stretch to progenitor stellar populations, with implications for SN Ia standardisation and progenitor models. The paper benefits from a large homogeneous sample, an automated decomposition pipeline with public code, and a 90–92% morphological agreement with Galaxy Zoo classifications for a subset. The environmental variables are derived from independent galaxy image models rather than from the SN light curves, so the headline correlations are not circular by construction. The main significance risk is the heavy, x1-dependent selection into the final 728-galaxy sample, which is acknowledged but not corrected or bounded; this threatens the quantitative claims (16.8σ, 5.1σ) and the comparison between elliptical and disk slopes.","major_comments":[{"comment":"The final sample retains only 728 of 3628 SNe Ia (20%), and Fig. 8 demonstrates that selection into the final sample is x1-dependent: negative-x1 SNe are over-represented and positive-x1 SNe are under-represented. Fig. 7 further shows that fit success varies with redshift and host stellar mass. Because redder, more massive ellipticals and brighter disks are easier to fit, the selection probability is likely correlated jointly with x1 and the environmental tracer used in the headline regressions. The statement in §5.1 that trends can still be investigated because they do not depend on absolute or relative rates is not sufficient: a regression slope within a subpopulation is biased when selection depends on both x1 and the tracer. Please provide a selection simulation, inverse-probability weighting, or a bounded-bias estimate demonstrating that the 16.8σ and 5.1σ slopes and their significances survive.","section":"§5.1, Fig. 8"},{"comment":"The final morphological mix is heavily biased, with 51% ellipticals versus the expected ~37% based on Li et al. (2011) and Kelvin et al. (2014). The abstract and §5.1 state that the results are 'robust,' but no quantitative sensitivity test is given for how this over-representation affects the elliptical versus disk comparison. A reweighting of the sample to the expected morphological mix, or a repeat of the regressions on the Galaxy Zoo-confirmed subset (92%/90% agreement, §4.7), would strengthen the claim that the observed differences are physical rather than an artefact of which galaxies are fit successfully.","section":"§5.2, Table 5"},{"comment":"The paper claims in the abstract and §6.2 that the elliptical and disk colour–x1 relations 'correlate fundamentally differently,' and §5.3 describes the disk slope as 'much steeper' than the elliptical one, but no formal test of slope difference is reported. Because the two slopes are estimated on different subpopulations with different selection functions and systematics, a quantitative comparison (e.g., a bootstrap or a likelihood-ratio test on a combined fit) is needed before the 'fundamentally different' conclusion can be assessed.","section":"§5.3, Fig. 10"}],"minor_comments":[{"comment":"The removal of nine host galaxies after manual inspection (reducing the sample from 728 to 719) is described without a quantitative outlier criterion; please state the threshold or procedure used and show the main regressions with and without these objects.","section":"§5.2"},{"comment":"The 3.3σ dust trend is based on splitting the sample at c=0.035, a choice the text itself calls arbitrary, and the trend is identified after a visual separation of the innermost SNe. This is a post-hoc analysis; please present it with an explicit multiple-testing caveat or a sensitivity scan over the split value.","section":"§5.4, Fig. 12"},{"comment":"The text says 'Using Sterling's approximation' in §4.1; this should be 'Stirling's approximation.' There are also several typographical spacing issues (e.g., 'di fferent') throughout the manuscript.","section":"§4.6"},{"comment":"The statement 'No apparent bias is seen in the SALT2 c distribution' is qualitative; a two-sample test (e.g., KS or Anderson-Darling) between the full sample and the final sample would be a more robust way to support that claim.","section":"§5.2"}],"recommendation":"major_revision","confidential_remarks":"The selection-function issue is the main blocker; if the authors can provide a convincing selection correction, simulation, or bias bound, the paper could be a valuable contribution. I do not see grounds for rejection beyond that, but the current manuscript does not yet support the quantitative significance claims as stated."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Colleague,\n\nThis is a real step forward. Senzel et al. take the ZTF DR2 sample and apply automated 2D bulge/disk/bar decomposition to the host galaxies, then split the sample by morphological component instead of using global host mass or colour. The headline results—the 16.8σ correlation between SALT2 x1 and SN-component-colour in ellipticals, the 5.1σ version in disks, and the 6.1σ local surface-brightness trend in disks—are the kind of thing that could sharpen the mass-step conversation. The paper also shows that the component colour explains more of the x1 variance (R²=0.37 in ellipticals) than host mass (R²=0.13), which is a genuinely useful observation.\n\nThe pipeline is careful. The code is public, the morphological classification checks against Galaxy Zoo with 90–92% agreement, and they are transparent about the cuts, the outlier removal, and the degeneracies between models. I believe the central physical picture: low-stretch SNe Ia come from older, redder stellar populations, and the slope differs between ellipticals and disks. That is consistent with prior global-colour results and now gets component-level support.\n\nThe soft spot is the selection function, and it is the right thing to worry about. They go from 3628 SNe Ia to 728 modelled hosts—an 80% loss—and the loss is not random in x1 or in host properties. Their Fig. 8 shows an excess of low-x1 SNe in the final sample, and ellipticals are over-represented relative to the expected rate. The paper argues that trends can still be studied because they don't depend on absolute rates. That is true for rate-independent correlations only if selection does not depend jointly on x1 and the environmental tracer. Redder, more massive ellipticals are easier to fit, and low-x1 SNe preferentially occur in those redder ellipticals. So the observed slope could be steeper than the true relation. No simulation, weighting, or completeness correction is presented to show the 16.8σ and 5.1σ significances survive. I don't think this kills the result—there's enough external evidence that environment-colour tracks x1—but it is a load-bearing uncertainty, and the abstract's claim that the results are \"robust\" is not established by the paper.\n\nMinor points: the c=0.035 split for the dust trend is arbitrary, and the paper admits it; that result is speculative anyway. The K-correction template choice depends on the fitted bulge-disk colour difference, which is a mild circularity but contained. The 91bg-like colour offset (3.5σ) comes from small numbers.\n\nWho is this for? Anyone working on SN Ia standardization or host environment. It's a useful data product and worth citing. A serious referee should engage with it; the main request should be a selection-corrected or at least simulated robustness check of the slopes. My own verdict would be conditional: the qualitative picture is likely right, but the quoted significances should be treated with caution.","headline":"A real step forward in SN Ia environmental tracers, with headline significances that are probably inflated until the heavy sample selection is accounted for.","tokens_in":36642,"tokens_out":3268,"would_cite":true,"duration_ms":35067,"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":"16.8-sigma link ties supernova stretch to host galaxy colour","keywords":["Type Ia supernovae","host galaxy decomposition","SALT2 stretch x1","SN-component-colour","galaxy colour","stellar population age","ZTF DR2","galaxy morphology"],"falsifier":"Reconstruct the ZTF DR2 selection by simulating the full pipeline from host images to final galaxy model for SNe Ia injected with known $x_1$ values, and check whether the recovered $x_1$-colour slopes in ellipticals and disks survive; a cheaper version is to re-weight the 728-event sample using the $x_1$-dependent entry rate in Fig. 8 together with the redshift and mass success rates in Fig. 7, and see whether the $16.8\\sigma$ and $5.1\\sigma$ slopes remain.","tokens_in":35545,"feed_emoji":"💥","tokens_out":8192,"duration_ms":70981,"temperature":0.7,"pith_summary":"This paper sets out to establish that the stretch of a Type Ia supernova light curve ($x_1$) is set by the stellar population of the galaxy component in which its progenitor formed. Modelling 728 host galaxies from the ZTF DR2 sample with an automated two-dimensional decomposition into elliptical, disk, bulge, and bar components, the authors find a strong linear relation between $x_1$ and the model-derived $g-r$ colour of the host component in both elliptical galaxies ($16.8\\sigma$, $R^2=0.37$) and disk galaxies ($5.1\\sigma$, $R^2=0.15$), with lower-stretch supernovae in redder environments. In disk hosts they also find a $6.1\\sigma$ trend between lower stretch and brighter local $r$-band surface brightness, consistent with age and metallicity gradients toward galaxy centres. If these trends are real, supernova stretch is a direct environmental age and metallicity indicator, and the differing slopes in old versus young stellar populations could sharpen the corrections applied to Type Ia supernovae used as cosmological distance indicators.","feed_headline":"16.8-sigma link: SN Ia stretch tracks host galaxy colour","feed_subtitle":"In 728 ZTF hosts, lower-stretch SNe Ia sit in redder, older stellar environments, in ellipticals and disks alike.","key_machinery":"The load-bearing object is the 'SN-component-colour': the intrinsic $g-r$ colour of the galaxy model component (elliptical, bulge/bar, or disk) that hosts the supernova, obtained from a custom fully automated two-dimensional decomposition of DESI-LS $g$- and $r$-band images. The code fits S\\'ersic profiles for ellipticals, bulges, and bars plus an exponential disk, with PSF convolution, superellipse isophotes, K-corrections, and morphological selection assisted by external deep-learning morphology classifications, then assigns each SN to a component by the flux ratio at its position. Because the galaxy shape is common to both bands, each component has a single colour; this average component colour is argued to track the birth environment of the progenitor better than the local pixel colour or host stellar mass. The machinery's key output is that this colour, and in disks the local $r$-band surface brightness, predicts SALT2 $x_1$ at high significance, with host stellar mass as a weaker tracer.","core_discovery":"The central discovery is that SN Ia light-curve stretch correlates separately, and significantly, with the intrinsic colour of the specific galaxy component that hosts the explosion. For normal SNe Ia in ellipticals the relation between SN-component-colour and SALT2 stretch has a significance of $16.8\\sigma$ (Pearson $r=-0.61$, $R^2=0.37$); for SNe Ia in disks the relation is $5.1\\sigma$ ($r=-0.39$, $R^2=0.15$), with a steeper slope that is driven mainly by red 'Green Valley' disks. The authors interpret redder components as older and/or more metal-rich stellar populations, so lower stretch is an age and metallicity effect. They also report a $6.1\\sigma$ correlation between $x_1$ and local $r$-band surface brightness in disk hosts, a possible $3.3\\sigma$ dust-related trend between SN colour $c$ and surface brightness for $c>0.035$, and a $3.5\\sigma$ offset whereby 91bg-like SNe Ia live in redder ellipticals than normal SNe Ia at fixed host mass. The intersection of the elliptical and disk stretch relations gives a low/high-stretch crossover of $x_1=-0.39\\pm0.06$, consistent with companion ZTF papers.","pith_inferences":["The heavy selection from 3628 to 728 hosts could amplify the reported slopes; a weighting or simulation built from the paper's own x1-dependent entry rate would test whether the $16.8\\sigma$ and $5.1\\sigma$ significances survive.","The SN-component-colour definition invites a direct follow-up: compare the same 728 hosts against integral-field-unit spectroscopic age and metallicity maps to separate the age and metallicity contributions that the paper cannot disentangle.","The same automated pipeline, applied to larger samples from next-generation wide-field surveys, could turn $x_1$ into a cheap environmental age indicator for SNe Ia, provided the selection function is modelled forward.","The dust interpretation for $c>0.035$ could be tested by matching reddened disk SNe Ia to spatially resolved extinction maps and checking that their line of sight through the disk is longer than that of blue SNe Ia at the same surface brightness."],"forward_implications":["If stretch is set by progenitor age and metallicity, environmental corrections for SN Ia cosmology should use component-level galaxy colour or local surface brightness rather than global host mass alone.","The elliptical-only $16.8\\sigma$ relation suggests that, under a double-degenerate progenitor picture, normal SNe Ia in ellipticals may come from a single progenitor channel whose properties are set by stellar age.","Bulge and bar SNe Ia share the low-stretch, red-colour behaviour of ellipticals, so mixing them with disk SNe Ia dilutes environmental correlations.","The $x_1=-0.39\\pm0.06$ crossover between the elliptical and disk trends co-locates with the broken-\\alpha stretch split seen in companion ZTF analyses, supporting two distinct stretch regimes.","Intrinsic SN colour $c$ is mostly environment-independent, with red $c>0.035$ SNe Ia in disk hosts plausibly tracing dust toward galaxy centres."],"supporting_citations":[{"why":"Defines the ZTF Cosmo DR2 sample, its light-curve quality cuts, redshift sources, and host stellar masses that the analysis builds on.","marker":"Rigault et al. (2024)"},{"why":"Provides the DESI-LS $g$- and $r$-band images from which isophote maps and galaxy models are built.","marker":"Dey et al. (2019)"},{"why":"Supplies the SALT2 light-curve model whose $x_1$ stretch and $c$ colour are the dependent variables of the study.","marker":"Guy et al. (2007)"},{"why":"Companion ZTF paper establishing the broken-$\\alpha$ nonlinearity in $x_1$ standardisation, used as the comparison for the stretch crossover.","marker":"Ginolin et al. (2024)"},{"why":"Binary population synthesis linking faster declining SNe Ia to older progenitor populations, used to interpret the colour-stretch trends as an age effect.","marker":"Shen et al. (2017)"},{"why":"Supplies SN Ia rates per Hubble type used to quantify the bias toward elliptical hosts in the final sample.","marker":"Li et al. (2011)"},{"why":"Deep-learning morphology probabilities used to choose between competing galaxy models and to benchmark the classifications.","marker":"Walmsley et al. (2023)"},{"why":"Earlier evidence that galaxy colour is a stronger environmental tracer than host mass, which the colour-based results extend.","marker":"Kelsey et al. (2021)"}],"fun_headline_variants":["SN Ia stretch: 16.8σ host colour link in ellipticals","ZTF: SN Ia stretch tracks host colour, age","728 galaxies: SN Ia stretch-environment relation robust","Elliptical vs disk hosts: SN Ia stretch follows colour","Age and metallicity set SN Ia stretch, ZTF finds"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The analysis assumes that the heavy selection from 3628 SNe Ia down to 728 modelled hosts does not bias the relative $x_1$-environment correlations; no simulation, weighting, or selection correction is presented, and Fig. 8 shows an $x_1$-dependent retention with an excess of low-stretch events.","fun_headline_variants_meta":{"raw":{"variants":["SN Ia stretch: 16.8σ host colour link in ellipticals","ZTF: SN Ia stretch tracks host colour, age","728 galaxies: SN Ia stretch-environment relation robust","Elliptical vs disk hosts: SN Ia stretch follows colour","Age and metallicity set SN Ia stretch, ZTF finds"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000411,"raw_usage":{"total_tokens":2250,"prompt_tokens":1187,"completion_tokens":1063,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":803,"completion_tokens_details":{"reasoning_tokens":978}},"tokens_in":803,"tokens_out":1063,"duration_ms":10686,"temperature":1.0,"reasoning_tokens":978,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T18:02:29.911080+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Reconstruct the ZTF DR2 selection by simulating the full pipeline from host images to final galaxy model for SNe Ia injected with known $x_1$ values, and check whether the recovered $x_1$-colour slopes in ellipticals and disks survive; a cheaper version is to re-weight the 728-event sample using the $x_1$-dependent entry rate in Fig. 8 together with the redshift and mass success rates in Fig. 7, and see whether the $16.8\\sigma$ and $5.1\\sigma$ slopes remain.","supporting_citations":[{"cited_title":"J., Toonen, S., & Graur, O","cited_arxiv_id":null,"evidence_quote":"Binary population synthesis linking faster declining SNe Ia to older progenitor populations, used to interpret the colour-stretch trends as an age effect."},{"cited_title":"2023, MNRAS, 526, 4768","cited_arxiv_id":null,"evidence_quote":"Deep-learning morphology probabilities used to choose between competing galaxy models and to benchmark the classifications."}],"review_version":1}