{"id":"0772adf2-493e-4c8a-917f-80ef54f745e2","arxiv_id":"2411.10981","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":6,"one_line_summary":"Stellar mass-to-light ratios for nearby galaxies are most stable near 1.6 microns, and correcting for star formation rate reduces the scatter to about 0.02 dex.","lead":"This paper maps how accurately a galaxy's stellar mass can be inferred from near-infrared light, using 2,853 nearby galaxies and simulated SPHEREx-like spectra. It finds the smallest scatter near 1.6 microns, and that knowing a galaxy's star formation rate can shrink the scatter to about 0.02 dex.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 0.02 dex claim is an in-sample scatter around a relation fitted to the same CIGALE double-exponential SFH outputs; M*, SFR, and M*/L are not independent, so it overstates external accuracy.","rationale":"The reader's weakest-assumption analysis identifies the central issue: treating CIGALE-derived stellar masses, SFRs, and stellar-only luminosities as ground truth converts the post-correction scatter into an accuracy statement. My stress-test agrees with that reading and sharpens it: in this paper, the 0.02 dex scatter is not merely 'measured against a model,' it is the residual of a relation fitted to one output of the same SED fit that produced both axes (M*/L and sSFR). The double-exponential SFH with a fixed 12 Gyr old component and a young component forces sSFR to be a near-monotonic tracer of the old-to-young stellar mass ratio, which is exactly the physical quantity that determines the 1.6 micron M*/L. Removing that dependence by fitting Eq. 1 is therefore close to removing the only degree of freedom the model allows; the tiny residual reflects the smoothness of the broken power law and photometric noise, not the performance of the method on real galaxies. The manuscript itself provides supporting evidence for this concern: the explicit caveat in footnote 1 that zero-point offsets of 0.3–0.4 dex are not included; Appendix A's demonstration that adopting a different SFH/IMF shifts stellar masses by ~0.12 dex scatter; Appendix B's NIR residual systematics; and the Section 4.2 warning that observable SFR indicators will not equal CIGALE instantaneous SFR. All of these are genuine error terms that are not represented in the headline 0.02 dex. A single decisive test would be to repeat the analysis with a non-parametric or otherwise independent SFH parameterization; if the residual scatter at 1.6 micron rises well above 0.02 dex, the strong abstract claim fails. If it remains near 0.02 dex, then the result is more robust than the concern suggests. Either way, the paper's model-internal scatter maps are useful forecasts, but the wording 'accuracy of ~0.02 dex' should be conditioned on the model family and on externally validated SFR and mass calibrations. The reader's CONDITIONAL verdict is therefore appropriate and no change is needed from this stress-test pass.","tokens_in":23127,"tokens_out":3700,"duration_ms":39896,"concrete_test":"Refit the same 2853 galaxies with a non-parametric SFH (e.g., Prospector or CIGALE with 5–7 time bins and a continuity prior), keeping the same photometric bands, stellar library, and IMF. Recompute the 1.6 micron M*/L from the best-fit stellar SEDs and fit Eq. 1 to the new M*/L–sSFR pairs. If the residual scatter is significantly larger than 0.02 dex (e.g., >0.05 dex), the claim is an artifact of the double-exponential SFH; if it remains ~0.02 dex, the concern is disconfirmed.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's headline result—~0.02 dex stellar-mass 'accuracy' at 1.6 micron given SFR prior—is not supported as an accuracy claim. In Section 2.2, stellar mass, SFR, and the stellar-only SED all come from the same CIGALE fit with a two-component (12 Gyr old plus young 10–5000 Myr) double-exponential SFH. The M*/L used in Figures 11–12 is computed from that same best-fit model, and sSFR is SFR/M* from the same fit. The paper concedes in Section 4.2 that the tight M*/L–sSFR relation is 'rather expected' because sSFR tracks the old-to-young mass ratio in that SFH. Eq. 1 therefore absorbs the model's internal covariance; the 0.02 dex residual measures fitting noise within one model family, not the accuracy of a real spectrum plus an independent SFR estimate. The paper itself flags the omitted systematics: footnote 1 in Section 3.1 notes a zero-point offset of up to 0.3–0.4 dex from parameterized SFH is not considered; Appendix A shows that switching from a flexible delayed SFH and Salpeter IMF to the adopted setup changes M* with ~0.12 dex scatter; Appendix B reports the model SEDs underestimate JHKs fluxes by up to ~0.04 dex; Section 4.2 warns that common SFR indicators (H-alpha, PAH, etc.) differ from CIGALE instantaneous SFR. None of these enter the 0.02 dex number. Thus the abstract's 'accuracy of ~0.02 dex' should be reframed as precision around the model's own M*/L–sSFR relation, conditional on the assumed double-exponential SFH.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper uses CIGALE spectral energy distribution (SED) fits to 2853 nearby galaxies (435 from DustPedia, 2418 from Stripe 82) to compute stellar mass-to-light ratios M*/L over 0.75-5.0 microns, motivated by SPHEREx-like all-sky near-infrared spectroscopy. It reports that the scatter in M*/L is minimized near 1.6 microns (~0.10 dex for stellar-only luminosity), that M*/L is weakly correlated with stellar mass and strongly correlated with specific star formation rate (sSFR), and that after correcting the sSFR dependence with a broken power law the scatter drops to ~0.02 dex. The abstract concludes that stellar masses can be estimated with ~0.02 dex accuracy given a prior knowledge of SFR from future infrared spectra.","tokens_in":23577,"tokens_out":6618,"duration_ms":69297,"significance":"The wavelength-resolved M*/L scatter maps (Figs. 4, 5; Table 1) are useful products for planning SPHEREx-type surveys, and the comparisons against Meidt et al. (2014), Querejeta et al. (2015), and Jarrett et al. (2023) provide valuable external anchors. The demonstration that 3.3 micron PAH emission complicates simple NIR color corrections (Figs. 6-8) is an interesting and likely robust result. If the headline 0.02 dex number were a genuine external accuracy, it would be a major advance; as presented, however, it is an in-sample scatter around a relation fitted to the same SED-fitting outputs that produced M*, SFR, and M*/L. The paper is commendably honest in its footnotes and appendices about several missing systematics, but the abstract and conclusions do not carry those caveats, and the forecast for future surveys is therefore overstated.","major_comments":[{"comment":"The central claim that stellar mass can be estimated with ~0.02 dex accuracy is not supported by the analysis as presented. In Section 2.2, the stellar mass M*, the SFR, and the stellar-only luminosity used to compute M*/L all come from the same CIGALE fit with the adopted double-exponential SFH (old 12 Gyr plus young 10-5000 Myr), and sSFR is defined as SFR/M* from that same fit. Section 4.2 itself states that the tight M*/L-sSFR relation is 'rather expected' because sSFR approximately traces the old-to-young mass ratio in that SFH. Equation (1) therefore fits the internal covariance of a single model family, and the ~0.02 dex residual in Figure 12 measures how well the broken power law reproduces CIGALE's own outputs, not the accuracy with which a real spectrum plus an independent SFR estimate recovers true stellar mass. This statement should be reframed as model-conditional precision, and the abstract should not call it 'accuracy.'","section":"Abstract; Section 4.2; Eq. (1)"},{"comment":"The systematics that the 0.02 dex number omits are substantial and should be folded into any accuracy statement. Footnote 1 in Section 3.1 concedes that parameterized star formation histories can produce zero-point offsets up to 0.3-0.4 dex. Appendix A shows that changing from the flexible delayed SFH plus Salpeter IMF to the adopted setup changes M* with a scatter of ~0.12 dex. Appendix B reports that the best-fit models underestimate the JHKs fluxes by up to ~0.04 dex, which the paper notes could bias NIR-based M*/L low by a similar amount. None of these terms enters the 0.02 dex residual. At minimum, the paper should report internal precision and external systematics separately, and it should avoid quoting 0.02 dex as the expected error budget for future surveys.","section":"Footnote 1; Appendix A; Appendix B"},{"comment":"The practical recipe implied by the abstract requires an SFR prior, but the paper does not propagate the uncertainty of that prior. Equation (1) uses sSFR, so any application requires a stellar mass estimate before the correction can be applied; Section 4.3 acknowledges this iterativeness but does not quantify the resulting error. More importantly, the SFR from CIGALE is an instantaneous SFR, whereas the SFR indicators proposed for SPHEREx-type data (hydrogen recombination lines, 3.3 micron PAH) trace averages over ~10 Myr and differ from CIGALE's definition, as the paper notes in Section 4.2. The comparison quoted in Section 2.2 already shows a large dispersion in SFR relative to GSWLC-2 (-0.12 +/- 0.56 dex). With a fitted slope alpha ~0.2 in Equation (1), a 0.3-0.5 dex uncertainty in sSFR translates into roughly 0.06-0.1 dex in M*/L, dwarfing the 0.02 dex residual. This propagation needs to be quantified before the forecast is made.","section":"Section 4.2; Section 2.2; Section 4.3"}],"minor_comments":[{"comment":"The broken-power-law fit parameters show discontinuous jumps around 3.4 and 4.4 microns (for example, beta changes from ~0.6 to 0.007 and back, while the residual scatter sigma increases from ~0.02 to ~0.05 dex). The authors should explain whether these are separate fit branches or degeneracies and should report fit-parameter uncertainties; the current presentation makes the wavelength dependence of the correction hard to interpret.","section":"Table 3"},{"comment":"Please state explicitly in the captions that the reported ~0.02 dex is the residual scatter after the Equation (1) fit is subtracted, not the raw scatter of M*/L; this distinction is central to the paper's claims.","section":"Figure 12; Table 3"},{"comment":"The Gaussian approximation for SPHEREx channel transmission is used for all numerical results; please add a short justification or sensitivity test showing that the exact channel shape does not change the scatter estimates.","section":"Section 3.2"},{"comment":"There are several formatting and typographical issues: in the abstract, 'theinfraredspectraldatafacilitatethepreciseestimation' lacks word spaces; 'Hershel' should be 'Herschel'; and the wavelength range is given inconsistently as '0.75-5 um' and '0.75-5.0 microns.'","section":"Abstract; Section 1"},{"comment":"No data or code availability statement is provided; consider making Table 1 and the per-wavelength fit results available in machine-readable form to support reproducibility.","section":"Data availability"}],"recommendation":"major_revision","confidential_remarks":null},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper's real product is the first wavelength-resolved map of stellar M*/L scatter from 0.75 to 5 microns at R~40, built from 2,853 galaxies with consistent CIGALE fits across DustPedia and Stripe 82. That is genuinely new and useful for SPHEREx survey design. The finding that ~1.6 micron is the sweet spot, and that the scatter is roughly constant around 0.10 dex once dust is removed, is solid model-internal statistics. The comparison with Meidt, Jarrett, and Querejeta gives useful empirical anchor points, and the appendices are honest about known systematics.\n\nThe soft spot is the headline. The abstract says stellar mass can be estimated with \"~0.02 dex accuracy\" given SFR, but that number is the residual scatter around a broken power-law relation between M*/L and sSFR, where M*, SFR, and M*/L all come from the same CIGALE double-exponential SFH fit. The paper itself concedes (Section 4.2) that the tight M*/L–sSFR relation is \"rather expected\" because sSFR tracks the old-to-young mass ratio in that SFH. So 0.02 dex is precision within one model family, not accuracy against real galaxies. Footnote 1 (zero-point up to 0.3-0.4 dex), Appendix A (~0.12 dex scatter from SFH/IMF choice), and Appendix B (~0.04 dex NIR flux underestimation) are all left out of the 0.02 number. That's not a fatal flaw—the maps still stand—but the abstract overstates it, and the practical claim should be reframed as \"precision around the model's own M*/L–sSFR relation.\"\n\nOne minor issue: the analysis leans on an unpublished companion paper (Lee et al. submitted) for the comparison at 1.6 micron. That should be cited as in prep or the result condensed here.\n\nWho this is for: anyone planning stellar mass measurements with SPHEREx or other low-resolution NIR surveys, and people who need a quantitative sense of where M*/L scatter lives across 0.75-5 micron. The central practical recommendation—use ~1.6 micron, correct for sSFR—holds up. It deserves a serious referee; the request should be to reframe the headline claim and add an external check (e.g., dynamical masses or simulations) if space allows.\n\nRecommendation: send to peer review with a request for revision, not rejection.","headline":"A useful wavelength-resolved map of M*/L scatter for SPHEREx, but the headline 0.02 dex 'accuracy' is in-sample precision around the same fitted model, not external accuracy.","tokens_in":24198,"tokens_out":2178,"would_cite":true,"duration_ms":21950,"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":"Stellar masses of nearby galaxies can be pinned down to 0.02 dex using 1.6 micron light plus a star-formation-rate correction.","keywords":["stellar mass-to-light ratio","near-infrared galaxy spectra","specific star formation rate","spectral energy distribution fitting","dust and PAH emission","galaxy stellar masses","nearby galaxies"],"falsifier":"Take a sample of nearby galaxies with independent dynamical stellar masses and H-$\\alpha$ star-formation rates, compute their $\\sim 1.6\\,\\mu$m $M_*/L$ from stellar-only light, subtract the paper's sSFR-corrected relation, and measure the residual scatter; if it is appreciably larger than $0.02$ dex for galaxies with complex star-formation histories, the claimed accuracy is an artifact of the assumed two-component star-formation model.","tokens_in":22903,"feed_emoji":"🔭","tokens_out":13513,"duration_ms":133579,"temperature":0.7,"pith_summary":"This paper asks how accurately a galaxy's stellar mass can be read off from its near-infrared light in the 0.75–5.0 $\\mu$m window that future all-sky spectral surveys will cover. Using synthetic spectra from multi-wavelength spectral energy distribution (SED) fits of 2,853 nearby galaxies, the authors compute the scatter in the stellar mass-to-light ratio, $M_*/L$, at 68 spectral channels. They find a U-shaped scatter curve: young stellar populations inflate the scatter at short wavelengths and dust emission inflates it at long wavelengths, with a minimum near $\\sim 1.6\\,\\mu$m. After removing the strong dependence of $M_*/L$ on the specific star-formation rate (sSFR) with a broken power-law correction, the scatter at $\\sim 1.6\\,\\mu$m falls to $\\sim 0.02$ dex, about 5 percent. If this holds, a near-infrared spectrum plus a star-formation-rate estimate would deliver stellar masses of nearby galaxies to roughly 5 percent precision for very large samples.","feed_headline":"Stellar masses shrink to 0.02-dex scatter at 1.6 microns","feed_subtitle":"A star-formation-rate correction drops the mass-to-light scatter from 0.10 to 0.02 dex.","key_machinery":"The analysis rests on synthetic SEDs built by fitting multi-wavelength photometry with a two-component (old plus young) star-formation history, a fixed initial mass function, a dust attenuation curve, and nebular emission. From those SEDs the paper computes luminosity and $M_*/L$ in 68 spectral elements of resolution $\\sim 40$ spanning $0.75$–$5.0\\,\\mu$m. At each wavelength it measures the scatter in $\\log(M_*/L)$, regresses out the weak stellar-mass dependence, and then fits the sSFR dependence with a smoothly broken power law, Eq. (1): $\\log(M_*/L)=a\\left(\\frac{1}{2}\\left[1+\\left(\\frac{\\log \\mathrm{sSFR}}{\\delta}\\right)^{1/\\beta}\\right]\\right)^{-\\alpha\\beta}$. This relation is the load-bearing identity of the paper: removing it from the data collapses the scatter at $\\sim 1.6\\,\\mu$m from $\\sim 0.10$ dex to $\\sim 0.02$ dex.","core_discovery":"The paper's central claim is that the near-infrared stellar mass-to-light ratio is almost entirely set by the specific star-formation rate once dust contribution is removed from the light. At $\\sim 1.6\\,\\mu$m, using only stellar light, the scatter in $\\log(M_*/L)$ across the combined sample is $\\sim 0.10$ dex; after correcting for the smoothly broken power-law dependence on sSFR, that scatter drops to $\\sim 0.02$ dex. The authors conclude that stellar masses of nearby galaxies can be estimated to an accuracy of $\\sim 0.02$ dex from $\\sim 1.6\\,\\mu$m spectral data when an SFR estimate is available. They also show that total-luminosity $M_*/L$ values are systematically contaminated by dust continuum and the 3.3 $\\mu$m polycyclic aromatic hydrocarbon (PAH) feature, and that the standard near-infrared colors (W1$-$W2 or IRAC1$-$IRAC2) cannot reliably remove that contamination, whereas a spectral color between 3 and 4 $\\mu$m traces the dust fraction well.","pith_inferences":["If real galaxies follow the same sSFR-$M_*/L$ relation, the paper's table of broken power-law parameters becomes a ready-made calibration for future surveys; a straightforward check is to apply it to galaxies with independent dynamical masses.","The very small post-correction scatter suggests that, within the assumed two-component star-formation history, sSFR almost fully determines the recent-to-total stellar mass ratio, so the claimed 0.02 dex may partly reflect the structure of the model rather than an intrinsic property of galaxies.","The paper's spectral color-dust relation between 3 and 4 $\\mu$m can be tested against high-spatial-resolution near-infrared spectroscopy of nearby star-forming galaxies; if dust fraction and $m_{3\\mu m}-m_{4\\mu m}$ decouple at low metallicity, the dust model would need revision.","Because the sSFR correction requires an initial stellar mass estimate, routine implementation will need the iterative scheme the paper sketches; a practical extension is to test how the iteration converges when the SFR and the spectrum are noisy."],"forward_implications":["All-sky near-infrared spectral surveys can, in principle, produce stellar masses for huge numbers of nearby galaxies at $\\sim 0.02$ dex scatter, provided each galaxy has a star-formation-rate estimate.","Wavelengths beyond roughly $2\\,\\mu$m require spectral dust removal; two-band colors like W1$-$W2 are not enough because the 3.3 $\\mu$m PAH feature and warm dust push the color in opposite directions.","The $\\sim 1.6\\,\\mu$m region is the natural choice for mass estimation because stellar-light $M_*/L$ is least scattered there and dust contamination is minimal.","A spectrum that also provides an SFR indicator, such as hydrogen recombination lines or PAH emission, can supply both the correction and the mass, so no external SFR catalog is needed.","The $0.02$ dex figure is internal to the adopted stellar population models; systematic offsets from the initial mass function, star-formation history, and stellar-population modeling are not included and can be larger."],"supporting_citations":[{"why":"Supplies the SED-fitting engine that yields the stellar masses, SFRs, and synthetic spectra used throughout.","marker":"Boquien et al. 2019"},{"why":"Provides the stellar population synthesis models from which the stellar SEDs are computed.","marker":"Bruzual & Charlot (2003)"},{"why":"Supplies the massive-galaxy subsample's matched-aperture multi-wavelength photometry and SED-fit stellar and dust masses.","marker":"Li et al. (2023)"},{"why":"Provides the earlier SED fits for the local-galaxy sample whose systematic differences are quantified in Appendix A.","marker":"Nersesian et al. (2019)"},{"why":"Sets the initial mass function adopted when converting light into stellar mass.","marker":"Chabrier (2003)"},{"why":"Provides the dust attenuation curve used in the SED fitting.","marker":"Calzetti et al. (2000)"},{"why":"Gives the IRAC-band mass-to-light reference that the paper's broadband values are checked against.","marker":"Meidt et al. (2014)"},{"why":"Offers the empirical W1-based mass-to-light relation whose offset from the present measurements is explained by the photometric system.","marker":"Jarrett et al. (2023)"},{"why":"Supports the proposal that star-formation rates can be obtained from near-infrared recombination lines for the sSFR correction.","marker":"Kennicutt & Evans (2012)"}],"fun_headline_variants":["SFR correction cuts galaxy mass scatter to 0.02 dex","Near-IR mass estimates hit 0.02-dex accuracy with SFR","1.6 micron light plus SFR yields 0.02-dex stellar masses","Star-formation rate key to 0.02-dex mass-to-light accuracy","How to get stellar masses to 0.02 dex: know the SFR"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the SED-fitting model's estimates of stellar mass, star-formation rate, and starlight are close enough to the truth that the remaining 0.02 dex scatter measures estimation accuracy, rather than just the internal consistency of one assumed star-formation history.","fun_headline_variants_meta":{"raw":{"variants":["SFR correction cuts galaxy mass scatter to 0.02 dex","Near-IR mass estimates hit 0.02-dex accuracy with SFR","1.6 micron light plus SFR yields 0.02-dex stellar masses","Star-formation rate key to 0.02-dex mass-to-light accuracy","How to get stellar masses to 0.02 dex: know the SFR"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00023,"raw_usage":{"total_tokens":1568,"prompt_tokens":1117,"completion_tokens":451,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":733,"completion_tokens_details":{"reasoning_tokens":348}},"tokens_in":733,"tokens_out":451,"duration_ms":5480,"temperature":1.0,"reasoning_tokens":348,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T19:04:32.304297+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take a sample of nearby galaxies with independent dynamical stellar masses and H-$\\alpha$ star-formation rates, compute their $\\sim 1.6\\,\\mu$m $M_*/L$ from stellar-only light, subtract the paper's sSFR-corrected relation, and measure the residual scatter; if it is appreciably larger than $0.02$ dex for galaxies with complex star-formation histories, the claimed accuracy is an artifact of the assumed two-component star-formation model.","supporting_citations":[{"cited_title":"2003, MNRAS, 344, 1000 12 Kim et al","cited_arxiv_id":null,"evidence_quote":"Provides the stellar population synthesis models from which the stellar SEDs are computed."},{"cited_title":"H., Cluver, M","cited_arxiv_id":null,"evidence_quote":"Offers the empirical W1-based mass-to-light relation whose offset from the present measurements is explained by the photometric system."}],"review_version":1}