{"id":"18616bdc-48c0-4894-a566-c48849f0979f","arxiv_id":"2411.09681","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"Archival quasar time-lags are consistent with X-ray reverberation from a standard accretion disc when the X-ray corona is above about 40 gravitational radii.","lead":"Using published quasar monitoring data, the authors averaged ultraviolet and optical time-delay measurements and fitted them with a detailed X-ray heating model. The fit supports the idea that delayed variations in quasars are caused by X-rays from a hot region above the accretion disc, and it places that hot region at least about 40 black-hole radii above the disc.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Claimed h>40 Rg lower limit is not supported by the paper's own confidence intervals: M01.7 best fit is 42^{+34}_{-23} Rg, so h≈19 Rg is within 1σ.","rationale":"The reader flagged extrapolation of the K21b analytic relations beyond their stated mass and accretion-rate range as the weakest assumption. That concern is real but partially mitigated by the paper's robustness tests in Sect. 2.2, where stricter cuts give 'almost identical' results. I find a more decisive, internal problem: the headline lower limit h>40 Rg conflicts with the paper's own Table 1 confidence intervals. For the non-spinning best fit, the 1σ range extends down to about 19 Rg, so the data are consistent with substantially lower corona heights. This directly undermines the central quantitative claim rather than depending on model accuracy. The marginal fit quality (pnull=0.04) reinforces the need for caution, but does not by itself invalidate the analysis. The appropriate response is to keep the CONDITIONAL verdict and require the authors to report the χ²(h) profile and correct or soften the h>40 Rg statement, rather than to reject the work outright.","tokens_in":15941,"tokens_out":8568,"duration_ms":84052,"concrete_test":"Produce the Δχ²(h) profile for M01.7 and M12.4 over h=5–80 Rg, and state explicitly the confidence-level convention for the Table 1 errors. Specifically evaluate χ² at h=20, 30, and 40: if χ²(20)−χ²(42) < 4, or if the 68% interval for h includes 20, then the claim that h must exceed 40 Rg is contradicted, and the abstract and Sect. 5.2 would need revision.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's central quantitative conclusion—'corona height should be equal or larger than 40 Rg' (Sect. 5.2) and the abstract's 'as long as the corona height is larger than ~40 Rg'—is not supported by the best-fit errors reported in Table 1. For the non-spinning, fcol=1.7 model (M01.7), the best fit is h=42^{+34}_{-23} Rg; for the maximally spinning, fcol=2.4 model (M12.4), h=50^{+15}_{-12} Rg. If these are standard 1σ intervals, then h≈19 Rg is within 1σ for M01.7 and h≈38 Rg for M12.4. A height of 20 Rg would increase χ² by about 1 relative to the M01.7 minimum (χ²_min=11.4/5, pnull=0.04), giving pnull≈0.03—still an acceptable fit at the 1% level. Thus the data do not require h>40 Rg, and the abstract's lower limit is an overstatement. The paper also calls pnull=0.04 'fit well', although conventional 5% threshold would reject; this makes the h>40 claim even less secure. The issue is internal to the reported statistics, not dependent on model extrapolation.","agreement_with_reader":"disagree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper compiles archival UV/optical continuum time-lag measurements for 128 quasars, computes a rest-frame mean time-lag spectrum in six wavelength bins, and fits it with the analytical X-ray reverberation models of Kammoun et al. (2021a,b), extended to different colour-correction factors by Kammoun et al. (2023). For each of six model variants (spin a*=0 or 1 with fcol = 1, 1.7, 2.4), the only free parameter is the X-ray corona height h. The authors report that only the models with (a*=0, fcol=1.7) and (a*=1, fcol=2.4) can fit the data, with best-fit heights h=42^{+34}_{-23} R_g and h=50^{+15}_{-12} R_g respectively, both with chi^2_min/dof = 11.4/5 and p_null = 0.04. The central claim is that the observed quasar time lags are fully consistent with the X-ray reverberation hypothesis, provided the corona height is larger than about 40 R_g, and that this is consistent with microlensing disc-size constraints.","tokens_in":16169,"tokens_out":3975,"duration_ms":37361,"significance":"If substantiated, the result would be a notable step: it would show that standard Novikov-Thorne discs illuminated by a compact X-ray corona can reproduce the UV/optical time lags in luminous quasars, potentially resolving the apparent discrepancy between observed and predicted disc sizes without invoking non-standard disc physics. The study uses a large archival sample and a physically motivated model that includes relativistic effects and radial-dependent external heating, and it explicitly connects to independent microlensing constraints. However, the statistical support is marginal (p_null = 0.04, below the conventional 0.05 threshold), and the headline lower limit on the corona height is not supported by the confidence intervals reported in Table 1. The analysis also relies on an arbitrary error-inflation factor and on the applicability of K21b relations outside their calibrated parameter range.","major_comments":[{"comment":"The claim that 'the corona height is larger than ~40 Rg' is not supported by the best-fit uncertainties reported in Table 1. For model M01.7 the best fit is h=42^{+34}_{-23} Rg, so h≈19 Rg is within the 1σ lower interval; for model M12.4 h=50^{+15}_{-12} Rg, so h≈38 Rg is within 1σ. Thus the data are consistent with heights well below 40 Rg, and the statements in the abstract and in Sect. 5.2 ('the corona height should be equal or larger than 40 Rg') are overstatements. Please quote the confidence intervals and rephrase the conclusion as a best-fit range rather than a firm lower limit.","section":"Abstract; Sect. 5.2; Table 1"},{"comment":"The reported p_null = 0.04 for the two acceptable models is below the conventional 0.05 threshold for rejecting the null hypothesis. The text nevertheless describes these fits as fitting 'well' and later concludes that the data are 'fully consistent' with the model. This is an internal inconsistency in the statistical interpretation. The authors should acknowledge that the fits are marginal at the 5% level, report the exact p-values, and discuss how the conclusion depends on the choice of significance threshold.","section":"Sect. 4; Table 1"},{"comment":"The error-inflation factor of 1.1 applied to the observed mean time-lag errors is ad hoc and is not a propagation of the stated uncertainties on MBH, Lbol, and L2-10keV. Since the chi^2 values are near the rejection threshold, the choice of this factor directly affects whether a model is deemed acceptable. The authors should either propagate the input parameter uncertainties through the model to obtain model time-lag uncertainties, or at minimum perform a sensitivity analysis varying the inflation factor and reporting how chi^2_min and p_null change.","section":"Sect. 4"},{"comment":"The K21b analytic time-lag relations were developed for AGN with MBH <~ 1e8 Msun and accretion rate <~ 0.5, but the final sample extends to MBH < 1e9 Msun and lambda_Edd <= 1. The text states that the analysis was repeated on smaller subsamples with MBH < 5e8 and 2e8 Msun and that the results were 'almost identical', but no numerical results are provided. This robustness test is load-bearing for the extrapolation claim; please present the best-fit heights, chi^2 values, and p-values for these restricted samples, or justify why the K21b relations remain valid in the extrapolated regime.","section":"Sect. 2.2"}],"minor_comments":[{"comment":"The abstract states that the model assumes 'the measured BH mass, accretion rate and X-ray luminosity', but the 2-10 keV luminosity is not measured; it is estimated from the Lusso et al. (2012) relation. Please clarify this in the text.","section":"Abstract"},{"comment":"The 'Conclusions' heading is followed by no text. A concise conclusions paragraph summarizing the main findings and their caveats should be added.","section":"Section 5"},{"comment":"For models M01 and M11 the best-fit height is listed as 80 Rg, which is the upper boundary of the searched range. Please state explicitly whether the chi^2 minimum is at the boundary and whether the fit was attempted beyond 80 Rg.","section":"Table 1"},{"comment":"The description 'we increased the error of \\bar t_lag,obs by a factor of 1.1' is ambiguous: it should state whether the factor is applied to the standard error of the mean or to the individual time-lag errors before averaging.","section":"Sect. 4"},{"comment":"The Kolmogorov–Smirnov test result is quoted as 'pnull = 0.15'. Please define this quantity explicitly and ensure the notation is consistent with the p_null used for the chi^2 fits.","section":"Sect. 4; Fig. 7"}],"recommendation":"major_revision","confidential_remarks":"The paper's central quantitative claim (h > 40 Rg) is contradicted by the paper's own Table 1 uncertainties, and the reported p_null = 0.04 does not support the language of a 'fully consistent' fit. These issues are reparable with revised statistical reporting and a more cautious interpretation, so I do not recommend rejection. The authors should also be encouraged to make the restricted-sample robustness results available, since the model's validity beyond the calibrated parameter range is a key assumption. The heavy reliance on the authors' own K21a/b models is not inappropriate given the model's physical sophistication, but the novelty claim ('first time') should be carefully verified against the literature."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"First thing you should know: this is the first time the K21a/b relativistic reverberation model has been fitted to stacked quasar continuum lags, and it broadly works—but the 'h > 40 Rg' lower limit in the abstract and Section 5.2 is not supported by the paper's own error bars. For the non-spinning, fcol=1.7 model, the best fit is 42^{+34}_{-23} Rg; for the maximally spinning fcol=2.4 model, 50^{+15}_{-12} Rg. A height of ~19 Rg is within 1σ of the first fit and ~38 Rg of the second. So the data do not require h>40; they are consistent with a range that extends below it. The stress-test note holds up.\n\nWhat's genuinely new: nobody has applied the K21 model to this sample of quasars before, and the mean time-lag spectrum from 128 AGN is a useful compilation. The model fits two of six variants (M01.7 and M12.4) with χ²_min/dof around 11.4/5. The residual analysis is honest—the distribution of individual residuals is consistent with Gaussian scatter. The authors also check smaller subsamples to guard against extrapolation of the analytic relations, and they acknowledge the flat-disc assumption could lower the height. That's good practice.\n\nThe soft spots are in proportion. The pnull=0.04 is below the standard 0.05 threshold, so calling it 'fit well' is generous; it fits marginally. The error inflation factor of 1.1 is ad hoc, though not likely to change the qualitative picture. The model is used beyond its original MBH and accretion-rate range; the subsample check mitigates this, but you can't fully remove the concern without recomputing the response functions. None of this kills the central claim that X-ray reverberation is a viable explanation for quasar lags. It does mean the height lower limit should be softened to 'consistent with h ~ 40 Rg, with lower values allowed.'\n\nWho this is for: AGN reverberation people, especially those working on disc sizes and corona geometry. It deserves a serious referee—the analysis is reproducible in principle and the result matters. But the referee should ask for a revision that fixes the abstract and the height claims, and rephrases 'fit well.'","headline":"First real application of the K21 model to quasar continuum lags, but the 'h > 40 Rg' lower limit is not supported by the paper's own 1σ error bars.","tokens_in":16783,"tokens_out":3372,"would_cite":true,"duration_ms":32318,"reading_group":"yes","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"The UV/optical time lags of 128 quasars are fully consistent with X-ray reverberation, provided the X-ray corona sits above about 40 gravitational radii.","keywords":["X-ray reverberation","accretion disc","quasar time lags","corona height","black hole spin","Novikov-Thorne disc","UV/optical continuum","active galactic nuclei"],"falsifier":"A direct measurement of the X-ray corona height in one of these quasars — for example from an iron-K X-ray reverberation lag that locates the corona, or from microlensing size measurements — that placed the corona below about 20 gravitational radii would contradict the best-fit models, which require about 40-50 Rg. Alternatively, if future time-lag data for quasars above $10^9\\,M_\\odot$ (where the analytic relations are extrapolated) require heights below 40 Rg to fit, the paper's central claim would be falsified.","tokens_in":15703,"feed_emoji":"🕳️","tokens_out":11971,"duration_ms":95996,"temperature":0.7,"pith_summary":"The paper tests whether the well-known wavelength-dependent delays between UV and optical variations in quasars can be produced by X-ray reverberation, the heating of the accretion disc by X-rays from a compact corona. Using archival time-lag measurements for 128 quasars from four monitoring surveys, it builds a rest-frame mean time-lag spectrum and fits it with analytic X-ray reverberation models. The data are fully consistent with the model when the corona height is larger than roughly 40 gravitational radii, for both a non-spinning black hole (best fit 42 Rg with colour correction 1.7) and a maximally spinning black hole (best fit 50 Rg with colour correction 2.4). This matters because earlier work reported quasar disc sizes that appear two to three times too large; the paper's models remove that tension without invoking enlarged or exotic discs.","feed_headline":"Quasar light delays match a high X-ray corona, not a giant disc","feed_subtitle":"Standard accretion discs explain quasar UV/optical lags if the corona sits about 40 gravitational radii up.","key_machinery":"The load-bearing object is the K21b analytic time-lag model, built on the K21a disc response functions. These functions compute, at each disc radius, the extra time-variable UV/optical flux produced when X-rays from a point-like corona are absorbed by the disc, including relativistic light travel and bending, the disc ionisation state, and a radial-dependent ratio of external (X-ray) to internal (accretion) heating. Given a black hole mass, accretion rate, 2-10 keV luminosity, spin, colour-correction factor (the standard adjustment for electron scattering in the disc atmosphere), and corona height $h$, the model predicts the lag between two wavelengths; the authors fix all parameters from the literature and leave $h$ as the only free parameter, scanning 5 to 80 $R_{\\rm g}$ (gravitational radii). A later colour-correction prescription is used to adapt the model to $f_{\\rm col}$ values of 1 and 1.7 as well as the original 2.4.","core_discovery":"On the paper's own terms, the discovery is that the continuum UV/optical time lags of luminous quasars can be explained by thermal reverberation of a standard Novikov-Thorne accretion disc, with no need for anomalously large disc radii. The mean rest-frame lag spectrum of 128 quasars is well fitted by the K21b analytic time-lag relations for both zero and maximal black hole spin, as long as the X-ray corona height is above about 40 gravitational radii. For a non-spinning black hole with colour correction 1.7, the best-fit height is 42 Rg ($\\chi^2=11.4$ for 5 degrees of freedom); for a maximally spinning black hole with colour correction 2.4, it is 50 Rg ($\\chi^2=11.3$). The residuals of individual quasars are consistent with a zero-mean Gaussian with the expected scatter, and the paper notes the same corona-height range independently explains the half-light radii of microlensed quasars as supporting evidence. The paper states this is the first time quasar continuum UV/optical time lags have been shown to be consistent with X-ray reverberation of standard discs.","pith_inferences":["A natural extension the paper does not pursue is to split the sample by black hole mass or Eddington ratio and re-fit the corona height in each bin; a trend would reveal that height is not universal.","If the best-fit heights are taken at face value, the X-ray corona sits well above the disc plane in luminous quasars, which bears on models of the disc-corona interface and on jet-launching geometry; the paper does not discuss this connection.","Future wide-field time-domain surveys with longer, denser UV/optical light curves could apply the same stacking method in several redshift bins, and simultaneous far-UV monitoring could help break the spin-colour degeneracy the paper identifies.","Because the flat-disc assumption tends to push the inferred height upward, a finite disc scale height modelled with radiative transfer could bring the best-fit heights down to around 20 Rg; this is a testable extension of the same model rather than a claim of the paper."],"forward_implications":["The observed UV/optical quasar lags do not require disc radii larger than standard thin-disc predictions; X-ray heating of a Novikov-Thorne disc accounts for them.","The X-ray corona in quasars is typically located at or above about 40 gravitational radii, matching the height independently inferred from microlensed quasar half-light radii.","Continuum time-lag measurements can constrain corona height, and the best-fitting height is insensitive to whether the black hole is non-spinning or maximally spinning.","Earlier models that reported factor-of-2-3 lag discrepancies likely failed because they assumed a constant ratio of X-ray to internal heating across the disc, rather than computing it radius by radius.","If the flat-disc assumption is relaxed, smaller corona heights may fit equally well, so the quoted heights are upper-side estimates."],"supporting_citations":[{"why":"Supplies the disc response functions and initial analytic time-lag prescription, including relativistic effects and ionisation-dependent reflection.","marker":"K21a"},{"why":"Provides the analytic two-wavelength time-lag relations for spin zero and spin one that the paper fits to the data.","marker":"K21b"},{"why":"Adds the colour-correction prescription that lets the models run at fcol values of 1 and 1.7.","marker":"Kammoun et al. (2023)"},{"why":"Supplies the SDSS-RM subsample of 95 quasars with ICCF time lags, BH masses, and 3000 Å luminosities.","marker":"Homayouni et al. (2019)"},{"why":"Supplies the ZTF subsample of 38 quasars with g-r and g-i time lags, masses, and luminosities.","marker":"Guo et al. (2022)"},{"why":"Supplies the Pan-STARRS subsample of 39 quasars with g-r, g-i, and g-z time lags.","marker":"Jiang et al. (2017)"},{"why":"Supplies the ZTF subsample of 19 quasars with g-r and g-i time lags.","marker":"Jha et al. (2022)"},{"why":"Provides the microlensing half-light radius result that independently favours corona heights above about 40 Rg.","marker":"Papadakis et al. (2022)"},{"why":"Provides the bolometric-to-2-10 keV luminosity relation used to estimate X-ray luminosities for each source.","marker":"Lusso et al. (2012)"}],"fun_headline_variants":["Quasar lags point to high corona, not giant disc","Standard disc plus high corona matches quasar lags","X-ray reverberation solves quasar UV/optical lag mystery","Quasar delays fit standard disc if corona sits high","Corona height over 40 Rg explains quasar time lags"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the analytic lag relations, originally calibrated for black holes at or below about $10^8\\,M_\\odot$ with accretion rate below 0.5, stay accurate for the quasars in this sample, which reach $10^9\\,M_\\odot$ and Eddington ratios up to 1; if they lose accuracy there, the inferred corona heights and the good fits would both change.","fun_headline_variants_meta":{"raw":{"variants":["Quasar lags point to high corona, not giant disc","Standard disc plus high corona matches quasar lags","X-ray reverberation solves quasar UV/optical lag mystery","Quasar delays fit standard disc if corona sits high","Corona height over 40 Rg explains quasar time lags"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000281,"raw_usage":{"total_tokens":1754,"prompt_tokens":1127,"completion_tokens":627,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":743,"completion_tokens_details":{"reasoning_tokens":553}},"tokens_in":743,"tokens_out":627,"duration_ms":5673,"temperature":1.0,"reasoning_tokens":553,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-12T20:23:26.167419+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A direct measurement of the X-ray corona height in one of these quasars — for example from an iron-K X-ray reverberation lag that locates the corona, or from microlensing size measurements — that placed the corona below about 20 gravitational radii would contradict the best-fit models, which require about 40-50 Rg. Alternatively, if future time-lag data for quasars above $10^9\\,M_\\odot$ (where the analytic relations are extrapolated) require heights below 40 Rg to fit, the paper's central claim would be falsified.","supporting_citations":[{"cited_title":"J., & Wang, S","cited_arxiv_id":null,"evidence_quote":"Supplies the ZTF subsample of 38 quasars with g-r and g-i time lags, masses, and luminosities."},{"cited_title":"E., Dov ˇciak, M., & Kammoun, E","cited_arxiv_id":null,"evidence_quote":"Provides the microlensing half-light radius result that independently favours corona heights above about 40 Rg."}],"review_version":1}