{"id":"25aa555b-4194-4c09-8752-9fcc04273bea","arxiv_id":"2411.12574","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"Spline and template fitting velocity methods for Ic-BL supernovae have similar precision once smoothing and phase errors are included, with a proposed red-continuum bias explaining most discrepancies.","lead":"This paper tests two common ways of measuring how fast supernova debris expands, and finds both have hidden error sources. It shows that the two methods often agree, and proposes when one method overestimates the expansion speed.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The red-continuum bias conclusion rests entirely on comparing new spline fits to published Modjaz et al. template velocities for two SNe; without recomputing those template fits, a systematic offset in the published values could produce or erase the claimed 12000 km/s bias.","rationale":"The reader's weakest assumption is the right target. The paper is genuinely useful: it quantifies smoothing and phase errors, provides reproducible code, and correctly notes that spline uncertainties are underestimated. But the headline physical conclusion, that red continua bias spline velocities toward the bluest triplet line by 10000-15000 km/s, is inferred rather than demonstrated. The offset for SN1998bw and the null offset for SN2013dx are consistent with the continuum story, but consistency with two data points is weak evidence, especially when the template-fitting half of the comparison was not recomputed. A systematic offset in the Modjaz et al. (2016) values, from phase, smoothing, or template choices, would directly mimic the claimed effect. Recomputing template fits on the same spectra is the minimal decisive check; a synthetic-spectrum test with controlled continuum slopes would be a complementary validation. If the check does not reproduce the claimed offset, the paper should be revised to present the continuum mechanism as a hypothesis rather than an identified cause. Because the paper already acknowledges template errors as unquantifiable and the current verdict is conditional, no change in verdict is needed; the condition should be that the comparison is re-run before the mechanism is cited as established.","tokens_in":23321,"tokens_out":5443,"duration_ms":56877,"concrete_test":"Recompute template-fitting Fe II velocities for the same SN1998bw and SN2013dx spectra used in the spline fits, using the original Modjaz et al. code or an independently reimplemented version, with matched smoothing, phase, and redshift conventions. Compare the recomputed template velocities with both the published values and the spline velocities. If the SN1998bw offset changes by more than the template-fitting errors (2000-5000 km/s) or the SN2013dx agreement is lost, the red-continuum mechanism is not established by the current data.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section 5 builds the central mechanism from a two-event comparison. The spline velocities are computed here, but the template velocities are taken from Modjaz et al. (2016, Figs. 10-14) without re-fitting the same spectra with the same pipeline. The two sets therefore differ in data reduction, smoothing (Fourier k versus Savitzky-Golay), phase assignment, binning, and possibly template implementation. The entire evidence for the red-continuum bias is that SN1998bw shows a ~12000 km/s offset while SN2013dx agrees; if the published template velocities for SN1998bw are systematically low for any of these reasons, the inferred mechanism is not established. The paper itself flags that template fitting has unquantifiable errors (Section 6.3), and the conclusion only states that the spline method overestimates 'assuming that the template fitting velocities are correct.' With one discrepant event and no independent quantitative measurement of the continuum slope, the continuum explanation is the least secure link in the central claim.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper compares two widely used methods for measuring Fe II expansion velocities in broad-line Ic supernovae: the template-fitting method of Modjaz et al. (2016) and a spline-fitting method. It quantifies the sensitivity of both methods to pre-smoothing choices, identifies an additional epoch/phase-shift error for template fitting (~500-2000 km/s), and an additional scatter-type error for spline fitting (~1000 km/s). Using two GRB-SNe (SN1998bw and SN2013dx), the paper argues that the two methods agree when the Fe II feature sits on a blue continuum but disagree by ~12000 km/s when the feature sits on a red continuum, because the blended minimum is biased toward the bluest triplet line at 4924 Å. The paper also provides best-practice recommendations and concludes that the velocity-evolution morphology is the same for both methods.","tokens_in":23543,"tokens_out":2389,"duration_ms":25390,"significance":"If the central claims hold, the paper is a useful methodological contribution: it quantifies several previously neglected error terms, gives concrete heuristics for smoothing parameters, presents a falsifiable explanation for method discrepancies, and provides the first publicly available Python implementation of the Fourier smoothing algorithm used in the template-fitting literature. The empirical demonstration that the two methods can agree (SN2013dx) and disagree (SN1998bw) in a way that tracks the local continuum is a conceptually simple and testable idea. The paper is also honest about the limits of its evidence, explicitly, for example, stating that the red-continuum conclusion assumes the template-fitting velocities are correct.","major_comments":[{"comment":"The central claim that the spline method overestimates Fe II velocities on a red continuum rests entirely on comparing new spline fits to template-fitting velocities taken from Modjaz et al. (2016) without re-fitting the same spectra with the same pipeline. The two velocity sets differ in data reduction, smoothing (Fourier k vs. Savitzky-Golay), phase assignment, binning, and possibly template implementation. A systematic offset in the published template velocities for SN1998bw, unrelated to the methods themselves, could produce or erase the claimed 12000 km/s difference. To make the red-continuum mechanism load-bearing, the authors should recompute template-fitting velocities for the same reduced spectra used for the spline fits, or otherwise demonstrate that the published values are directly comparable.","section":"§5, Figs. 10 and 14"},{"comment":"The proposed red-continuum mechanism is inferred from visual inspection of two events and a single literature temperature for SN2013dx (16000 K at 9.3 days), with no quantitative measurement of the continuum slope under the Fe II feature for either event. The predicted 10000-15000 km/s offset, taken from the separation of the 4924, 5018, and 5169 Å lines, is said to 'match very well' the 12000 km/s difference, but with two events this is a two-point coincidence unless the continuum slope is actually measured. The authors should quantify the local continuum slope at the compared epochs (e.g., by fitting a pseudo-continuum or blackbody to the same spectra) or perform a simple simulation of the blended triplet on red and blue continua to demonstrate that the mechanism produces the claimed magnitude of bias.","section":"§5 and §6.1"},{"comment":"The claim that spline fitting underestimates uncertainties by around 1000 km/s is derived by fitting power-law and broken power-law functions to the velocity evolution and measuring the residual scatter. This procedure assumes that the fitted model is the true velocity evolution; any model misspecification is automatically absorbed as an 'additional error' of the method. The paper does not report the residuals separately from the fit uncertainty, nor does it test whether the 1000 km/s scatter is consistent with the reported Monte-Carlo errors. At minimum, the authors should report the RMS of residuals and the typical Monte-Carlo error per epoch, so that the reader can see how much of the scatter is genuinely unexplained.","section":"§4, Fig. 9"},{"comment":"The recommendation of ~10% spline density and ~2.5% filter width as optimum parameters is based mainly on one high-resolution, high-S/N spectrum (SN1998bw, Fig. 5). The appendix shows that the optimum varies by event and spectral quality: for SN2020bvc (low-resolution, high-S/N) the suggested optimum is 5% filter width and 5% knots, while for SN2016P (low-resolution, low-S/N) there is 'no optimal smoothing level'. The stated 'optimum parameters' therefore appear to be case-dependent. The paper does note that parameters should be determined case by case, but the abstract and Section 7 present the 10%/2.5% combination as a general recommendation, which overstates the evidence from four test spectra.","section":"§2.2 and Appendix A"}],"minor_comments":[{"comment":"The axis label in Fig. 2 reads 'vFe vFe(k=100)' with a missing subscript, and the left panel of Fig. 7 is labeled 'GRB980424-SN1998bw' while the text and table use 'GRB980425-SN1998bw'. Fig. 14's caption says 'GRB130207A-SN2013dx' but the paper consistently refers to 'GRB130702A-SN2013dx'. These typos should be corrected.","section":"Figures 2, 7, and 14"},{"comment":"The description of the spline-fitting method, including the definition of 'spline density' and the Monte-Carlo error estimation, references Finneran et al. (2024b) but does not describe the method in this paper. For a self-contained methodological comparison, the essential algorithmic steps (how the minimum is located, how errors are propagated) should be summarized here or in an appendix.","section":"§2.2"},{"comment":"The paragraph on 'Unquantifiable errors' correctly notes that template construction and poor χ² fits may contribute to the large template-fitting errors, but it does not cite a specific figure or analysis showing how large these effects may be. A brief quantitative statement, even an order-of-magnitude estimate, would make this section more useful.","section":"§6.3"},{"comment":"There are many instances of the typographical spacing 'di fferent' and 'o ffers' (e.g., in the abstract, Section 2, and elsewhere). A thorough proofread would eliminate these artifacts.","section":"Throughout"}],"recommendation":"major_revision","confidential_remarks":"The paper is likely a good fit for A&A. The central methodological comparison is useful, but the evidence for the headline red-continuum mechanism is thin: it rests on two objects and published template velocities that were not recomputed here. I would be willing to accept after the authors either re-run the template-fitting on the same spectra or provide a quantitative measurement/simulation of the continuum bias. The paper's self-assessment is refreshing; it explicitly flags that the conclusion assumes template-fitting velocities are correct. I would encourage the editor to ask for a modest additional analysis rather than rejecting, since the practical guidance on smoothing and phase errors is valuable regardless of the red-continuum story."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"First thing you should know: this is a genuinely useful methods paper, and the central empirical result—that the two velocity methods can agree perfectly for one GRB-SN (SN2013dx) while disagreeing by ~12000 km/s for another (SN1998bw)—is new and well demonstrated. The paper also quantifies two previously neglected error terms: the smoothing parameter choice in the spline method contributes ~500–1000 km/s (up to 3000–4000 km/s in low-S/N spectra), and the phase/template choice in the template method contributes ~500–2000 km/s. Those numbers will be handy. The code for the Fourier smoothing is publicly available, which is a real plus.\n\nThe proposed explanation for the discrepancy, that a red continuum biases the spline minimum toward the bluest triplet line (4924 Å), is plausible, and the spectral plots in Figures 11–13 support the basic picture: the spline tracks the blue sub-feature, the template tracks the red one. The paper is careful to say the template velocities are assumed correct, and it flags in Section 6.3 that template fitting has unquantifiable errors. It also debunks the older claim that template fitting always handles blending better, since SN2013dx shows the two methods tracing the same feature.\n\nWhere it's softer: the magnitude of the claimed bias (10,000–15,000 km/s) rests entirely on comparing new spline fits to published Modjaz et al. template velocities for two SNe, without re-fitting the template side with the same pipeline. A systematic offset in those published values, for any reason, would change the inferred bias. The paper acknowledges this dependency but does not test it. The spline error estimate from power-law residual scatter also assumes the functional form, so it is a floor on the true scatter. Those are real limitations, but they are addressable, and the paper is honest about most of them.\n\nBottom line: it deserves peer review. The methods comparison and the negative result (blending does not always hurt spline fitting) are solid, and the red-continuum mechanism is a testable hypothesis the community should be aware of. I would want the authors to recompute template fits on the same spectra before hanging a strong claim on the 12000 km/s number, but the paper stands on its own as a useful error-budget reference.","headline":"A solid, honest methods paper that quantifies neglected error sources and shows the spline/template discrepancy is not universal, though the new red-continuum mechanism rests on a two-event comparison with published template velocities.","tokens_in":24082,"tokens_out":2634,"would_cite":true,"duration_ms":26563,"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":"Spline fitting inflates supernova iron velocities on red continua","keywords":["supernovae: general","gamma-ray burst: general","methods: data analysis","Fe II velocities","spline fitting","template fitting","line blending"],"falsifier":"Measure the local continuum slope at the Fe II feature for a sample of Ic-BL supernovae and compare it with the difference between spline and template fitting velocities: the red-continuum mechanism predicts a strong correlation, with spline velocities exceeding template velocities by roughly 10,000–15,000 km/s when the continuum is red, and near-zero differences on blue continua. Alternatively, apply both methods to synthetic Ic-BL spectra with a known injected velocity and a controlled continuum slope; if the spline method fails to show the predicted bias on red continua, the proposed mechanism is wrong.","tokens_in":23097,"feed_emoji":"🔭","tokens_out":5824,"duration_ms":47396,"temperature":0.7,"pith_summary":"This paper tries to establish which of the two standard ways of measuring iron velocities in broad-line Ic supernovae is trustworthy, and why the two sometimes disagree. It quantifies previously ignored errors in both methods: template fitting suffers phase-choice and smoothing effects of roughly 500–2000 km/s, while spline fitting routinely underestimates its uncertainties by about 1000 km/s and can be shifted by fine-tuning of smoothing parameters. The central claim is that the velocity discrepancies seen for some supernovae are not a general superiority of template fitting over spline fitting. Instead, the paper argues spline fitting overestimates the Fe II velocity only when the iron feature sits on a red continuum, which biases the blended minimum toward the bluest line of the iron triplet; on a blue continuum both methods agree. If true, the bias can reach 10,000–15,000 km/s, and the choice of method matters only for a subset of objects.","feed_headline":"Spline fitting inflates supernova velocities on red continua","feed_subtitle":"Both velocity methods agree on blue continua; only red continua bias spline fitting, by up to 15,000 km/s.","key_machinery":"The central object is the Fe II feature near 5000 Å, a blended triplet with lines at 4924 Å, 5018 Å, and 5169 Å, whose minimum wavelength is converted to an expansion velocity by the Doppler formula. The spline fitting method locates the minimum of the blended feature after smoothing with a Savitzky–Golay filter and spline interpolation, while the template fitting method blueshifts and broadens an average Ic spectrum's iron lines to match the input spectrum, combining the fitted blueshift with the template's known velocity. The load-bearing mechanism is the slope of the local continuum under the triplet: assuming roughly equal line strengths, a blue continuum makes the reddest line (5169 Å) closest to the feature minimum, while a red continuum pulls the minimum toward the bluest line (4924 Å), biasing spline fitting to higher velocities. This mechanism explains why the discrepancy appears for GRB980425-SN1998bw (cooler, red continuum) and is absent for GRB130702A-SN2013dx (hotter, blue continuum), and why template fitting, which flattens the spectrum before fitting, does not show the same bias.","core_discovery":"On the paper's own terms, the discovery is that the two popular velocity-measurement methods for the Fe II feature near 5000 Å do not always disagree, and the template fitting method is not always better at handling blended lines. Using direct comparisons of the well-sampled supernovae GRB980425-SN1998bw and GRB130702A-SN2013dx, the authors show that spline fitting produced velocities more than 12,000 km/s higher than template fitting for the first object, while the two methods agreed within errors for the second, despite similar apparent blending. They attribute this to the slope of the local continuum: when a red continuum underlies the iron triplet, the minimum of the blended feature shifts toward the 4924 Å line, yielding an artificially high velocity from spline fitting; when the continuum is blue, the 5169 Å line sits closest to the minimum and both methods trace the same feature. The paper further quantifies additional errors of roughly 500–2000 km/s from template phase choice and smoothing in both methods, and shows that spline fitting underestimates its uncertainties by about 1000 km/s. The conclusion is that no single method is always optimal, but the velocity evolution shape appears identical regardless of method.","pith_inferences":["A testable extension would be to compute the continuum slope locally at the Fe II feature for a larger sample and check whether the spline-minus-template velocity offset correlates with that slope as the mechanism predicts.","The mechanism implies that population-level velocity comparisons between GRB-associated and non-GRB Ic-BLs are unlikely to be distorted by this bias, since the paper argues the red-continuum configuration is relatively rare based on typical temperature evolution; a broader sample could confirm this directly.","If the red-continuum bias is common in low-temperature or dusty objects, multi-feature analyses that rely on spline fitting for Si II and Ca II may need the same continuum-slope check, since the same minimum-shift argument could apply to other blended features.","The paper's suggestion to flatten spectra before spline fitting, or to trace the 5169 Å line backward from late times, could be validated with synthetic spectra where the true velocity is known."],"forward_implications":["For Ic-BL supernovae with a blue continuum at the iron feature, spline and template fitting give consistent velocities, so large published samples using either method can be compared for those objects.","For objects with a red continuum, spline fitting velocities can be inflated by 10,000–15,000 km/s, so absolute velocity comparisons (e.g., between Ic-BLs and ordinary Ic supernovae) could be systematically biased if such objects are common.","Velocity evolution studies for the iron feature are robust to method choice, because the same line is traced throughout if one avoids switching lines after de-blending.","The template fitting method's additional errors from phase shifts (~1000 km/s) and smoothing (~500–1000 km/s) should be added in quadrature to its quoted uncertainties, especially at late times.","Spline fitting uncertainties should be increased by about 1000 km/s to account for unmodelled scatter, and smoothing parameters should be chosen per spectrum, with optimum filter widths near 2.5% of the spectrum length for high-quality data."],"supporting_citations":[{"why":"Supplies the template fitting method, the template spectra, and the published Fe II velocities for GRB980425-SN1998bw and GRB130702A-SN2013dx used in the direct comparison.","marker":"Modjaz et al. (2016)"},{"why":"Provides the Fourier smoothing algorithm that the template fitting method relies on, and whose smoothing parameter k the paper varies to quantify smoothing errors.","marker":"Liu et al. (2016)"},{"why":"Provides the GRB980425-SN1998bw spectra used in the spline tuning violin plots and the direct comparison, with host contamination removed.","marker":"Patat et al. (2001)"},{"why":"Provides the GRB130702A-SN2013dx spectra used in the direct comparison, with afterglow and host contribution removed.","marker":"D'Elia et al. (2015)"},{"why":"Describes the spline fitting implementation and parameters that this paper tests, and whose large-sample velocities motivated the discrepancy question.","marker":"Finneran et al. (2024b)"},{"why":"Previous demonstration that blending in the Fe II region can bias spline-fitting velocities, which the paper re-examines and qualifies.","marker":"Prentice et al. (2018)"},{"why":"Supplies the blackbody temperature evolution for Ic-BLs used to argue that blue continua are typical and red-continuum bias may be relatively rare.","marker":"Taddia et al. (2019)"}],"fun_headline_variants":["Red continua skew spline fitting, inflating SN velocities","Template fitting not always better for SN velocities","Spline vs template: velocity gap traced to red continua","Spline fitting overestimates velocities on red continua","SN velocity errors: template phase and smoothing matter"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the velocities from the previously published template fitting results of Modjaz et al. (2016) are directly comparable to the spline fitting velocities computed in this paper, even though the two sets were produced with different data reductions, smoothing choices, and template implementations; if the published values are offset for unrelated reasons, the red-continuum mechanism and its 10,000–15,000 km/s magnitude would not be established.","fun_headline_variants_meta":{"raw":{"variants":["Red continua skew spline fitting, inflating SN velocities","Template fitting not always better for SN velocities","Spline vs template: velocity gap traced to red continua","Spline fitting overestimates velocities on red continua","SN velocity errors: template phase and smoothing matter"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000442,"raw_usage":{"total_tokens":2337,"prompt_tokens":1144,"completion_tokens":1193,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":760,"completion_tokens_details":{"reasoning_tokens":1118}},"tokens_in":760,"tokens_out":1193,"duration_ms":10173,"temperature":1.0,"reasoning_tokens":1118,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T17:21:58.975910+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the local continuum slope at the Fe II feature for a sample of Ic-BL supernovae and compare it with the difference between spline and template fitting velocities: the red-continuum mechanism predicts a strong correlation, with spline velocities exceeding template velocities by roughly 10,000–15,000 km/s when the continuum is red, and near-zero differences on blue continua. Alternatively, apply both methods to synthetic Ic-BL spectra with a known injected velocity and a controlled continuum slope; if the spline method fails to show the predicted bias on red continua, the proposed mechanism is wrong.","supporting_citations":[{"cited_title":"B., & Graur, O","cited_arxiv_id":null,"evidence_quote":"Provides the Fourier smoothing algorithm that the template fitting method relies on, and whose smoothing parameter k the paper varies to quantify smoothing errors."},{"cited_title":"J., Ashall, C., Mazzali, P","cited_arxiv_id":null,"evidence_quote":"Previous demonstration that blending in the Fe II region can bias spline-fitting velocities, which the paper re-examines and qualifies."}],"review_version":1}