{"id":"bd551207-a1cd-4f8a-984d-b0a9519a7b69","arxiv_id":"2507.12275","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"In 18 nearby spirals, both young and evolved stars heat the cold dust, with the dominant source varying from galaxy to galaxy and no single mechanism winning overall.","lead":"This paper maps the temperature of cold dust across 18 nearby spiral galaxies and asks whether young star-forming regions or older stellar populations are doing the heating. The answer matters because dust temperature is a key input for estimating galaxy dust masses and star formation rates from infrared surveys.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Method 1's galaxy-by-galaxy classification lacks uncertainty bounds; several r differences are within typical noise, so the 56%/44% split may be unstable.","rationale":"The reader's weakest_assumption focused on Method 2's chain of proportionality assumptions, which is indeed fragile. However, the single most load-bearing point is Method 1's lack of uncertainty quantification, because the paper's central 'varying balance' claim is directly tied to per-galaxy classifications that compare two correlation coefficients against the 1:1 line. Several galaxies are within 0.01–0.04 of that line, well inside the expected sampling uncertainty for tens of independent resolution elements. The reader's requested 'correlation-coefficient uncertainties' in their condition partially anticipates this concern, but the reader did not identify it as the primary weakness. A bootstrap analysis is a straightforward, assumption-light test that would settle whether the 56%/44% split is real. Given that the reader already recommended CONDITIONAL, my analysis strengthens the same condition rather than changing the verdict; hence UNCHANGED is appropriate. I do not see grounds for REJECT: the radial Tdust profiles, the AGN non-detection, and the basic existence of significant correlations with both tracers are descriptive results that remain valid regardless of the classification's precision.","tokens_in":25549,"tokens_out":3292,"duration_ms":37903,"concrete_test":"Bootstrap the pixel samples by resampling independent 36-arcsec pixels with replacement, preserving the S/N and Sigma_dust cuts used in the paper, and recompute the weighted Pearson r and the difference delta = r(Tdust−SigmaSFR) − r(Tdust−SigmaMstar) for each galaxy, obtaining 95% confidence intervals. Classify a galaxy as young-star- or evolved-star-dominated only if the CI on delta excludes zero, and report how many of the 18 galaxies survive. If fewer than roughly 10 galaxies remain separable, the 'no single dominant mechanism' conclusion is not robust and the paper should be revised to present the classification as unresolved for the marginal cases.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim that young and evolved stars both heat cold dust with the balance varying galaxy-by-galaxy rests heavily on Method 1's per-galaxy classification (Table 4, Fig. 7), which compares weighted Pearson r(Tdust−SigmaSFR) with r(Tdust−SigmaMstar). The paper reports no uncertainties or confidence intervals for these coefficients. With only tens of independent 36-arcsec resolution elements per galaxy, the sampling error on r is typically 0.1–0.2 (Fisher z). Several classifications sit very close to the 1:1 line: NGC 3031 (0.64 vs 0.66), NGC 4736 (0.70 vs 0.73), NGC 5236 (0.50 vs 0.52), NGC 5194 (0.67 vs 0.70), and NGC 3521 (0.79 vs 0.69). Even larger nominal splits, such as NGC 3621 (0.81 vs 0.60), may be within 2-sigma for small samples. If the differences are not statistically separable, the 10/18 vs 8/18 split loses meaning, and the claim of galaxy-to-galaxy variation is not established by Method 1. Method 2 is also conditional on the Utomo et al. (2019) proportionality chain (KS slope n=2.19, constant XCO, linear Sigma_gas−Sigma_dust), but Method 1 was intended to be the independent, assumption-light leg; without error bars on the correlation coefficients it does not carry that weight.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper examines dust heating in 18 nearby, face-on spiral galaxies from DustPedia. The authors construct T_dust maps at 36-arcsecond resolution and analyze pixel-by-pixel correlations between T_dust, Σ_SFR, and Σ_M* (Method 1), as well as the T_dust–Σ_dust relation against the Utomo et al. (2019) prediction (Method 2). They report that T_dust declines from about 24 K at the center to about 15 K at R25, that AGNs do not significantly affect T_dust on the sampled scales, and that both young and evolved stars contribute to dust heating, with the relative importance varying by galaxy. The two methods agree in 13 of 18 galaxies, which the authors interpret as supporting a picture in which young stars are not the sole heating source in most systems.","tokens_in":25697,"tokens_out":8477,"duration_ms":88156,"significance":"If confirmed, the paper's conclusion that neither young nor evolved stars alone dominate cold-dust heating in typical nearby spirals would moderate the common assumption that star-formation tracers fully determine dust temperature and would motivate per-galaxy resolved radiative-transfer and energy-balance studies. The paper's strengths are the homogeneous multi-wavelength dataset, the new T_dust maps at the SPIRE-500 resolution, and the explicit comparison with published RT results for several galaxies in the sample. However, the statistical support for the headline claim currently has two load-bearing gaps: the Method 1 classification is based on correlation coefficients without any uncertainty estimates, and the Method 2 p-value is interpreted in a way that appears reversed relative to standard chi-square goodness-of-fit usage. The paper also honestly lists the main assumptions of Method 2 in Section 4, which is commendable but does not by itself remove the need for sensitivity tests.","major_comments":[{"comment":"The classification of each galaxy as young-star or evolved-star dominated, and the resulting 56%/44% split, rests on differences between two weighted Pearson coefficients that are not accompanied by any uncertainty estimate. For the tens of independent 36-arcsec pixels per galaxy, the sampling error on r is typically 0.1–0.2 (Fisher z), yet several galaxies lie within 0.02–0.10 of the 1:1 line (e.g., NGC 3031: 0.64 vs 0.66; NGC 4736: 0.70 vs 0.73; NGC 5236: 0.50 vs 0.52; NGC 5194: 0.67 vs 0.70; NGC 3521: 0.79 vs 0.69), and even the larger separation for NGC 3621 (0.81 vs 0.60) may be within about 2σ. The authors should provide bootstrap or permutation confidence intervals for the difference, report the number of independent pixels per galaxy, and either drop or explicitly de-prioritize classifications that are not statistically separable.","section":"§3.3.1, Table 4, Fig. 7"},{"comment":"The chi-square p-value is interpreted in the opposite direction from standard usage. With the conventional definition, p<0.05 means the data are unlikely under Utomo's model (poor agreement), while p>0.05 means the data are consistent with the model (no rejection). The paper states that p<0.05 indicates that the observed relationship aligns with the trend expected by Utomo et al. and that p>0.05 implies the assumption of young-star heating does not hold. If a non-standard definition (e.g., p = P(χ2 < observed)) is intended, it must be stated explicitly; otherwise the conclusion that about 72% of the sample are not uniquely young-star heated is reversed under the standard interpretation: 13 of 18 galaxies would instead be consistent with the young-star-only model, and the agreement between Methods 1 and 2 would drop dramatically. The same inversion appears in Section 3.1, where p ≤ 0.001 is described as indicating a very good fit.","section":"§3.3.2, Eq. (7)"},{"comment":"The Method 2 test is conditional on the entire Utomo et al. (2019) assumption chain: L_IR arises entirely from reprocessed young stellar radiation, the pixel-by-pixel Kennicutt-Schmidt slope is fixed at n = 2.19 from Casasola et al. (2022) at 3.4 kpc resolution, the CO-to-H2 conversion factor is constant, Σ_gas is proportional to Σ_dust, and the dust is optically thin and in thermal equilibrium. The paper acknowledges in Section 4 that the KS slope varies galaxy by galaxy and with resolution, and it notes that X_CO and the dust-to-gas ratio vary. Under these violations, the p-value no longer isolates young-star heating; deviations from the predicted relation can be produced by any failed proportionality step. The authors should quantify how the YS/ES classifications change with n, beta, X_CO, and the Σ_dust threshold, or explicitly state that the Method 2 conclusions are conditional on these assumptions.","section":"§3.3.2 and §4"}],"minor_comments":[{"comment":"The exponent in Eq. (1) appears as a broken or misformatted expression in the manuscript; it should read T_dust = T0 U^(1/(4+β)).","section":"Eq. (1)"},{"comment":"The statement that the r coefficients range from moderate to high is inaccurate given Table 4, which includes r = 0.18 (weak by the paper's own definition) and r = 0.36 (moderate); the text should say the range is weak to high.","section":"§3.3.1"},{"comment":"The caption of Fig. B.1 says 'already displayed in Fig. 1' but should reference Fig. 6 (or the main-text figure showing NGC 3621 and NGC 5055).","section":"Fig. B.1 caption"},{"comment":"The description of the chi-square test does not specify how the normalization constant A in Eq. (7) is determined, whether it is fitted as a free parameter or fixed, and what uncertainty is propagated into the p-value; this should be stated for reproducibility.","section":"§3.3.2"},{"comment":"The table would be more informative if it reported the number of independent 36-arcsec pixels used for each galaxy, since the statistical significance of both the correlation coefficients and the p-values depends on this number.","section":"Table 4"}],"recommendation":"major_revision","confidential_remarks":"The p-value inversion in §3.3.2 is the most serious issue: if the authors are using a non-standard definition of p-value, they must define it and apply it consistently; otherwise the paper's main quantitative claim is reversed. Adding uncertainty estimates for the Method 1 correlation differences is essential for the galaxy-by-galaxy variation claim. I would recommend that the authors consult a statistician or clarify the chi-square test procedure before resubmission."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this is a competent, cleanly written resolved study of dust heating in 18 DustPedia spirals. It contributes new Tdust maps (not in C17) and applies two diagnostics. The descriptive results look solid. But there is a load-bearing statistical misreading in Method 2 that flips its classifications, and Method 1 lacks uncertainty bounds, so the headline numbers should not be taken at face value.\n\nWhat it does well: the Tdust maps are new, the radial profile (peak ~24 K, dropping to ~15 K) is useful, and the null result on AGN heating at kpc scales is a clean descriptive finding. The authors are honest about the systematic assumptions (KS slope n=2.19, constant XCO, linear Sigma_gas-Sigma_dust, resolution effects) in Section 4. The qualitative conclusion that both young and evolved stars heat cold dust, with the balance varying per galaxy, is plausible and consistent with external radiative-transfer work.\n\nThe soft spot that matters: the chi-square p-value in Section 3.3.2 is read backwards. Under any standard goodness-of-fit test, a small p (say 0.014 for NGC 3621) means the data are unlikely to arise if the Utomo relation is true; that is evidence against the model. A p near 1 (e.g., 1.000 for NGC 5055) means the data are consistent with the model. The paper says the opposite: p<0.05 is taken as 'aligned with Utomo' and p>0.05 as 'deviating.' Table 4 then has five galaxies with p<0.05 and thirteen with p>0.05; under the correct reading, the five are the ones inconsistent with the young-star-only prediction and the thirteen are consistent with it. That reverses the paper's stated 72% verdict and changes the agreement with Method 1. This is not a subtle nuance; it is the core of the second method.\n\nMethod 1's per-galaxy classification also lacks error bars. The weighted Pearson r values in Table 4 come from tens of independent 36\" elements, so the typical Fisher-z uncertainty is ~0.1-0.2. Several galaxies sit within ~0.02-0.03 of the 1:1 line (NGC 3031, NGC 4736, NGC 5236, NGC 5194), and even NGC 3621's 0.21 split may not be 2-sigma. Without bootstrap or Fisher-z intervals, the 10/18 vs 8/18 split is not established as a quantitative claim.\n\nThe fix is straightforward: re-run Method 2 with the correct p-value interpretation, report chi-square values and pixel counts, and add uncertainties to the Method 1 coefficients. Then the paper would be a useful contribution. Until then, treat the agreement fractions as illustrative rather than measured.\n\nRecommendation: send it to peer review. It deserves serious referee time, but it will need heavy revision, mainly to correct the Method 2 interpretation and to add error propagation to Method 1. I would not cite the headline numbers until that is done.","headline":"New Tdust maps and a clean presentation, but Method 2's p-values are interpreted backwards and Method 1 lacks error bars; the 72% and 56% headline numbers are not yet supported.","tokens_in":26471,"tokens_out":6602,"would_cite":false,"duration_ms":73868,"reading_group":"yes","serious_thinker":"no","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"No single mechanism heats cold dust in nearby spiral galaxies: young and evolved stars share the job, with the balance varying from galaxy to galaxy.","keywords":["dust heating","cold dust","spiral galaxies","dust temperature","star formation rate surface density","stellar mass surface density","DustPedia","far-infrared observations"],"falsifier":"Run the same two-method analysis on galaxies with resolved, per-pixel measurements of both the old and young stellar radiation fields, for instance from 3D radiative-transfer modeling or SED decomposition, and ask whether the heat absorbed from evolved stars matches the deviations from the Utomo relation; if galaxies currently classified as young-star heated show equally strong evolved heating, the conclusion fails. A cheaper version is to redo the p-value test with per-galaxy Kennicutt–Schmidt slopes and metallicity-dependent CO-to-H2 and gas-to-dust ratios: if the 13/18 deviations disappear, the claim that there is no single dominant heating mechanism would no longer be supported.","tokens_in":25193,"feed_emoji":"🌌","tokens_out":6723,"duration_ms":66450,"temperature":0.7,"pith_summary":"This paper asks what heats the cold dust that makes up most of the dust mass in nearby spiral galaxies, and it argues that there is no single answer. Comparing 18 face-on spirals from the DustPedia project, the authors find that both young, massive stars and evolved stellar populations contribute measurably, and the balance shifts from galaxy to galaxy. In 10 of 18 galaxies, dust temperature correlates more strongly with star-formation surface density than with stellar mass surface density; in the remaining 8, the evolved population takes the lead. A second test, built on the relation between dust temperature and dust mass surface density, finds that young stars alone cannot account for the observed heating in 13 of 18 galaxies. The two methods agree on the dominant heating source in 13 of 18 galaxies, supporting a picture of mixed heating rather than a universal dominant mechanism.","feed_headline":"Cold dust in spirals has no single heat source","feed_subtitle":"Two stellar populations share the job, with the balance shifting galaxy by galaxy.","key_machinery":"The quantitative backbone is Eq. (7), a scaling from Utomo et al. (2019): $$(4+\\$\\beta$)\\log T_{\\rm dust}=A+(n-1)\\log\\Sigma_{\\rm dust}.$$ It chains four proportionalities: IR luminosity from young stars ($\\Sigma_{\\rm LIR}\\propto\\Sigma_{\\rm SFR}$), the Kennicutt\\,--\\,Schmidt law $\\Sigma_{\\rm SFR}\\propto\\Sigma_{\\rm gas}^n$ with fixed $n=2.19$, a constant gas-to-dust ratio ($\\Sigma_{\\rm gas}\\propto\\Sigma_{\\rm dust}$), and optically thin modified-blackbody emission $L_{\\rm IR}\\propto M_{\\rm dust}T_{\\rm dust}^{4+\\beta}$. For each galaxy the observed log $T_{\\rm dust}$\\,--\\,log $\\Sigma_{\\rm dust}$ relation is fit with this prediction and a $\\chi^2$ $p$-value (threshold 0.05) decides whether young-star heating alone explains the data. The first method, comparing weighted Pearson coefficients $r(T_{\\rm dust},\\Sigma_{\\rm SFR})$ and $r(T_{\\rm dust},\\Sigma_{M_\\ast})$, is the less assumption-heavy cross-check.","core_discovery":"The central claim is that cold dust ($T_{\\rm dust}\\sim15$\\,--\\,$24$ K) in typical nearby spiral galaxies is heated by both young and evolved stars, with no universal dominant mechanism. This is established with two complementary diagnostics: pixel-by-pixel weighted Pearson correlations between $T_{\\rm dust}$ and $\\Sigma_{\\rm SFR}$ versus $T_{\\rm dust}$ and $\\Sigma_{M_\\ast}$, and a $\\chi^2$ comparison of the observed log $T_{\\rm dust}$\\,--\\,log $\\Sigma_{\\rm dust}$ relation against the prediction of Utomo et al. (2019) that $(4+\\beta)\\log T_{\\rm dust} = A + (n-1)\\log\\Sigma_{\\rm dust}$ under pure young-star heating. The correlation test puts 10 of 18 galaxies on the young-star side of the 1:1 line, while the $p$-value test leaves 13 of 18 galaxies above $\\alpha=0.05$, meaning the young-star-only prediction is not consistent with those data. The two methods agree on the dominant heating source in 13 of 18 galaxies, which the authors read as evidence that both stellar populations contribute, with the balance depending on the galaxy.","pith_inferences":["If the result holds, statistical SED-fitting codes that assume a single radiation-field intensity per pixel will systematically misestimate dust masses in galaxies where evolved heating is significant.","The 28% of galaxies where the two methods disagree (e.g., NGC 628 and NGC 5055) are the natural testing ground for the method's assumptions, since their deviation should vanish if per-pixel radiation fields are modeled directly.","A testable extension would be to apply the same two-method comparison to galaxies with measured metallicities, predicting that high-metallicity galaxies show stronger evolved-star heating because their CO-to-H2 conversion factors and gas-to-dust ratios break the assumed proportionality.","A future far-infrared survey with higher spatial resolution could check whether the galaxies classified as young-star heated are merely those where evolved heating is smeared out by resolution effects."],"forward_implications":["Dust temperature at a given galaxy radius is not a reliable standalone star-formation tracer, since the evolved population can maintain a substantial warm component.","Interpreting resolved far-infrared and sub-millimeter emission in terms of ongoing star formation will overestimate star formation rates in the roughly half of galaxies where evolved stars dominate the heating.","Galaxy-wide radiative-transfer models must include both stellar populations to reproduce the observed $T_{\\rm dust}$ radial gradients, not just the young stars.","Low-luminosity AGN activity can be ignored as a dust-heating agent on the 0.3\\,--\\,3 kpc scales studied, since no temperature difference is seen between Seyfert and non-AGN galaxies.","Per-galaxy Kennicutt\\,--\\,Schmidt slopes and spatially resolved metallicity will be needed to separate the young- and evolved-star contributions robustly."],"supporting_citations":[{"why":"Supplies Eq. (7), the pure young-star-heating prediction tested by the second method.","marker":"Utomo et al. (2019)"},{"why":"Provides the fixed Kennicutt–Schmidt slope n = 2.19 used in Eq. (7).","marker":"Casasola et al. (2022)"},{"why":"Justifies the constant CO-to-H2 conversion factor X_CO = 2×10^20 cm^-2 (K km/s)^-1 assumed in the second method.","marker":"Bolatto et al. (2013)"},{"why":"Defines the 18-galaxy sample and supplies the Σ_SFR, Σ_M*, and Σ_dust maps; the T_dust maps are produced with the same methodology.","marker":"Casasola et al. (2017)"},{"why":"Calibration used to derive the Σ_SFR maps from GALEX-FUV plus WISE 22 µm emission.","marker":"Bigiel et al. (2008)"},{"why":"Underpins the SED-fitting approach from which the dust temperature maps are derived via the ISRF strength parameter U.","marker":"Draine & Li (2007)"}],"fun_headline_variants":["Both young and old stars heat spiral dust","Spiral dust heating: a shared job","Two stellar generations share dust heating duty","No single star type rules dust heating","Galaxy dust warmth from young and old alike"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The second method's verdict rests on the assumption that, pixel by pixel, infrared luminosity is a clean tracer of young-star heating and that gas, dust, and star formation connect through fixed power laws: a single Kennicutt–Schmidt slope of n = 2.19, a constant CO-to-H2 conversion factor, and a constant gas-to-dust ratio.","fun_headline_variants_meta":{"raw":{"variants":["Both young and old stars heat spiral dust","Spiral dust heating: a shared job","Two stellar generations share dust heating duty","No single star type rules dust heating","Galaxy dust warmth from young and old alike"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000496,"raw_usage":{"total_tokens":2491,"prompt_tokens":1065,"completion_tokens":1426,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":681,"completion_tokens_details":{"reasoning_tokens":1361}},"tokens_in":681,"tokens_out":1426,"duration_ms":10306,"temperature":1.0,"reasoning_tokens":1361,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T16:49:48.435053+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run the same two-method analysis on galaxies with resolved, per-pixel measurements of both the old and young stellar radiation fields, for instance from 3D radiative-transfer modeling or SED decomposition, and ask whether the heat absorbed from evolved stars matches the deviations from the Utomo relation; if galaxies currently classified as young-star heated show equally strong evolved heating, the conclusion fails. A cheaper version is to redo the p-value test with per-galaxy Kennicutt–Schmidt slopes and metallicity-dependent CO-to-H2 and gas-to-dust ratios: if the 13/18 deviations disappear, the claim that there is no single dominant heating mechanism would no longer be supported.","supporting_citations":[],"review_version":1}