{"id":"9cb0819c-8b5a-4502-b646-48c147c71a14","arxiv_id":"1908.09154","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":0,"one_line_summary":"Eta Carinae's mid-infrared brightness has been stable from 1968 to 2018, implying its true luminosity did not change despite the star's dramatic brightening at optical and ultraviolet wavelengths.","lead":"Astronomers measured the heat glow of Eta Carinae's surrounding dust cloud across 50 years, using new 2018 images and archived data back to 1968. They find the star's total brightness has been steady, so the recent brightening we see is likely caused by dust clearing along our line of sight, not by the star itself.","discovery_kind":"replication","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The no-decline conclusion rests on excluding the ISO 1996 flux densities from the averages; without a quantitative recalibration, the central null is partly circular.","rationale":"The paper's strongest claim is a null result: no mid-IR decline over 50 years, so eta Car's luminosity is stable and the optical/UV brightening is due to line-of-sight extinction. The new VISIR data and the 2005-2018 image comparison are solid and independently useful, but they cannot by themselves establish the 50-year stability claim. The historical compilation is the load-bearing part, and within it the decisive analytic choice is the exclusion of the ISO flux densities. Those are precisely the points that contradict the null, and the paper provides only qualitative reasons for distrusting them. Excluding them without a quantified calibration analysis risks circularity: the averages are computed after removing the low values, so the conclusion of no long-term change is partly built into the method. The reader flagged the broader issue of heterogeneous historical photometry and mentioned ISO exclusion in the rationale, but did not make it the central weakest assumption; my concern is more specific and more directly actionable. A mixed-effects refit including ISO with instrument-level systematics is a feasible check that would settle whether the exclusion is justified. The reader's CONDITIONAL verdict remains appropriate; the condition should explicitly include this sensitivity analysis, so no change to the verdict is needed.","tokens_in":14472,"tokens_out":5301,"duration_ms":64223,"concrete_test":"Recompute Fig. 3 including the ISO/Morris et al. (2017) flux densities at 10.5 and 18 um, using systematic uncertainty estimates from Van Malderen et al. (2004) and Morris et al. (2017), and fit the full 1968-2018 time series with a model that treats instrument/epoch as random intercepts. If the best-fit slope becomes significantly negative at >=95% confidence, the stability claim is not robust; if it remains consistent with zero under a range of plausible ISO calibration offsets, the exclusion is justified.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim of no long-term mid-IR decline is derived after explicitly removing the ISO 1996 flux densities from the averages in Fig. 3, whose caption states 'averages exclude the lower ISO flux values.' Those are the same measurements that Morris et al. (2017) used to claim a 25% decline. The paper's only justification is a general statement that ISO absolute spectrophotometric calibration was hampered by detector nonlinearities and memory effects (Sec. 3.1); no quantitative correction, no error inflation, and no contemporaneous cross-check is supplied. If the ISO points are even partly valid, the remaining 13–22% scatter in the historical photometry cannot rule out a decline of that size, and the no-change result is partly produced by the selection itself. This is not a minor calibration caveat: the central inference that eta Car's luminosity has been stable, and that the UV/optical brightening is due only to line-of-sight extinction, depends on the ISO exclusion being correct. A sensitivity analysis that includes these data with defensible systematic uncertainties is therefore necessary before the null result can be accepted.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents new 2018 VISIR mid-infrared images and photometry of η Car, together with archival VISIR 2005 and TIMMI2 2003 data, and combines these with published photometry from 1968–2005 to reconstruct the 8–20 μm spectral energy distribution of the integrated Homunculus nebula. The central claim is that the mid-IR flux has shown no long-term decline over five decades, implying a stable bolometric luminosity of about 4.6×10^6 Lsun, and that the well-documented UV/optical brightening since the late 1990s must instead be caused by a decrease in circumstellar extinction along the line of sight only. The authors also report expansion of the Butterfly loops and no brightness change in the inner loops between 2005 and 2018. The paper is transparent about data-quality limitations, explicitly noting the poor TIMMI2 data, ghost-affected 2005 VISIR data, partial saturation of the central core, and the heterogeneity of the historical measurements; however, the no-decline conclusion is built on averages that deliberately exclude the ISO 1996 flux densities, and the robustness of that exclusion is not quantified.","tokens_in":14670,"tokens_out":5792,"duration_ms":56211,"significance":"If the conclusion holds, it rules out large changes in η Car's bolometric luminosity over 50 years, strengthens the interpretation that the observed brightening is a line-of-sight extinction effect, and has direct consequences for Eddington-limit arguments and stellar evolution models. The paper contributes genuinely new, high-angular-resolution mid-IR photometry from 2018 and a careful re-reduction of archival data, and the authors are commendably candid about calibration and detector artifacts. The main weakness is that the central null result depends on the exclusion of the ISO 1996 points and on the comparability of very heterogeneous historical photometry, but neither is treated with a quantitative sensitivity analysis. If the requested statistical and systematic tests are added, the paper would provide a solid and important observational constraint on one of the most luminous stars in the Galaxy.","major_comments":[{"comment":"The central null result is constructed from averages that \"exclude the lower ISO flux values\" (Fig. 3 caption), and the only justification offered in the text is a general statement that the ISO absolute spectro-photometric calibration was hampered by detector non-linearities and memory effects (Sec. 3.1). This is the same ISO 1996 dataset used by Morris et al. (2017) to claim a 25% decline, so excluding it without a quantitative re-calibration makes the no-decline conclusion partly circular. Please provide either a re-reduction of the ISO spectra with a revised absolute calibration, or a sensitivity analysis that includes the ISO points with explicitly inflated systematic errors (e.g., 20–30%) and demonstrates that the inferred secular trend remains consistent with zero. Without this, the paper cannot rule out a real decline of the size previously claimed.","section":"Sec. 3.1 and Fig. 3 caption"},{"comment":"The historical comparison combines measurements obtained with different apertures, instruments, calibration standards, extraction methods, and even values read by hand from isophotal contour maps (open symbols in Fig. 2). The Fig. 2 caption itself warns that \"Calibration uncertainties may be larger than reported.\" The stated 13–22% scatter is therefore a lower bound on the true epoch-to-epoch uncertainty, and the claim that there is \"no evidence for a long-term change\" is based on visual inspection rather than a formal statistical test. Please add a trend analysis (e.g., fitting a constant plus a linear term in each wavelength region, adding a per-epoch systematic floor in quadrature) and report the 95% upper limit on a fractional flux decline per decade that the data can exclude. This would convert the null result from a statement about plotted scatter into a quantitative, falsifiable claim.","section":"Sec. 3.1 and Fig. 2"},{"comment":"The luminosity of about 4.6×10^6 Lsun is derived by combining 1973 near-IR photometry, 2005 and 2018 VISIR photometry, 1978 35–175 μm photometry, and 2010 450/850 μm data, but the SED integration method, the assumed spectral shape between bands, the adopted distance, and the propagated uncertainties are not described. Since the conclusion that \"Eta Car's luminosity has thus probably been stable\" is a central scientific claim, please specify how the SED was integrated or modeled, and quantify the systematic error budget including the distance dependence. This would also clarify how a 25% change in luminosity would manifest in the integrated flux and whether the data are sensitive to it.","section":"Sec. 3.1 and Sec. 4"}],"minor_comments":[{"comment":"Table 1 lists formal errors such as 9648 ± 30 Jy, which are far smaller than the stated 10% (N-band) and 20% (Q-band) systematic calibration limits; please quote errors with one significant figure and add the systematic floor in quadrature or state it separately.","section":"Table 1"},{"comment":"The orbital phases of the individual measurements are not given, although the paper discusses possible orbital-phase variations; please add the orbital phase of each epoch or include a supplementary table.","section":"Fig. 2 and Fig. 3"},{"comment":"The claim of \"the highest angular resolution mid-IR images of η Car to date\" is not supported by a quantitative comparison with earlier resolved imaging (e.g., the 2005 VISIR images achieve ~0.35'' resolution); please add a brief comparison with previously published angular resolutions.","section":"Abstract and Sec. 2"},{"comment":"The procedure used to \"calculate the areas in isophotal contour maps\" with SketchAndCalc is not described; please state how the contour integration was performed and quantify the resulting uncertainty, as these open-symbol points enter the light curve.","section":"Acknowledgements"},{"comment":"There are typographical errors in the reference list, e.g., \"V ol.\" in the entries for Conti (1997), Nota & Lamers (1997), and Rigaut & Gehring (1995); please proofread the reference formatting.","section":"References"},{"comment":"The sentence \"The 4 − 8 μm flux also shows no long-term variation (Russell et al. 1987; Figure 3)\" is not fully supported by Figure 3, whose M-band panel has very sparse coverage; please qualify the claim to reflect the limited number of epochs.","section":"Sec. 3.2"}],"recommendation":"major_revision","confidential_remarks":null},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The key thing to know: this paper claims Eta Car's mid-IR flux has been stable for 50 years, contradicting earlier claims of a 25% decline. The new 2018 VISIR images are genuine, high-quality data, and the paper is transparent about its own data problems. But the stability claim depends on excluding the ISO 1996 measurements, and that exclusion is argued only qualitatively. That's the load-bearing point a referee has to probe.\n\nWhat is actually new: first published VISIR images of Eta Car at 0.22-0.3 arcsec, plus a careful reconstruction of the 8-20 micron SED from 1968 to 2018. The paper also shows the 'Butterfly' loops expanded between 2005 and 2018 at about 100 km/s, with no brightness change. The authors are unusually honest about data quality: TIMMI2 data are flagged as bad, the 2005 VISIR data are ghost-affected, and the core is partially saturated. They also catch a transcription error in Morris et al. (2017). Good practice.\n\nThe soft spots are real. The most important is the handling of the ISO 1996 flux densities. These are the data that Morris et al. used to claim a 25% decline. The caption of Fig. 3 says 'averages exclude the lower ISO flux values,' and the text justifies this by saying ISO calibration was hampered by detector nonlinearities and memory effects. That may be correct, but no quantitative correction or error inflation is supplied. If ISO is partly right, the remaining scatter of 13-22% cannot rule out a decline of that size. So the null result is partly produced by the selection. A sensitivity analysis that includes ISO with a defensible systematic floor is required.\n\nA second, minor issue is the reliance on heterogeneous historical photometry, including values read by hand from contour maps. The authors are upfront about this, and they note that calibration uncertainties may be larger than reported. Still, a linear fit to the data with standard errors would make the claim more formal.\n\nOverall, the paper is a useful and important re-evaluation. The authors don't overclaim: they explicitly say variations within 25% could be present, and they conclude only that there is no decline larger than that. With the sensitivity analysis, the central claim would be much stronger. As is, it deserves peer review, and an editor should route it to a referee who knows the IR calibration literature.\n\nWho is this for? Massive-star and Eta Car specialists, and anyone concerned with dust extinction changes in LBVs. I'd bring it to reading group and would cite it if I worked in the area.","headline":"Transparent new VISIR data and a 50-year SED reconstruction argue for mid-IR stability in Eta Car, but the no-decline conclusion leans heavily on excluding ISO 1996 data without quantitative justification.","tokens_in":15249,"tokens_out":3391,"would_cite":true,"duration_ms":32182,"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":"Eta Car's mid-infrared flux stayed steady from 1968 to 2018, so its bolometric luminosity has probably remained constant while the UV-optical brightening reflects declining dust along our line of sight.","keywords":["eta Carinae","Homunculus nebula","mid-infrared photometry","luminous blue variables","circumstellar dust","dust extinction","spectral energy distribution","massive stars"],"falsifier":"Re-reduce the archived space-based spectra and the published 1970s ground-based photometry with a single, homogeneous calibration and aperture prescription, and compare the resulting fluxes with the 2018 VISIR values; if the re-reduced historical fluxes fall systematically more than about 25 percent below the modern values at 8–20 $\\mu\\mathrm{m}$, the claimed fifty-year stability is disproved.","tokens_in":14274,"feed_emoji":"⭐","tokens_out":16776,"duration_ms":126120,"temperature":0.7,"pith_summary":"Eta Carinae, the brightest mid-infrared source outside the Solar System, has brightened at ultraviolet and optical wavelengths since the late 1990s. This paper assembles new 2018 images, 2003-2005 archival data, and published photometry from 1968 onward into a fifty-year mid-infrared light curve of the surrounding Homunculus nebula. The paper finds that the $8\\text{--}20\\,\\mu\\mathrm{m}$ flux has not declined over that period and concludes that Eta Car's bolometric luminosity has probably stayed near $4.6 \\times 10^6\\,L_\\odot$ (at a distance of 2.3 kpc). The result overturns earlier reports of a 25 percent mid-infrared decline, implying that the ultraviolet and optical brightening comes from circumstellar dust clearing in our line of sight only, not from an increase in the star's luminosity.","feed_headline":"Eta Car's mid-infrared brightness held steady for 50 years","feed_subtitle":"The star's UV-optical brightening is likely dust clearing along our line of sight, not a rise in luminosity.","key_machinery":"The central object is the spectral energy distribution (SED) of the Homunculus nebula, the bipolar shell of dust and gas ejected during Eta Car's nineteenth-century eruption. Because most of Eta Car's optical and ultraviolet light is absorbed by circumstellar dust and re-radiated at infrared wavelengths, the integrated mid-infrared SED acts as a bolometer for the star's luminosity. The argument is carried by comparing epoch-by-epoch photometry at matching wavelengths (4.6, 8–13, and 17–20 $\\mu\\mathrm{m}$) and by anchoring the comparison to new diffraction-limited VISIR images from 2018 at $0.22^{\\prime\\prime}$–$0.5^{\\prime\\prime}$ resolution, together with 2003 TIMMI2 and 2005 VISIR archival data. The paper also corrects a transcription error in an earlier published figure that had overstated the historical 11 $\\mu\\mathrm{m}$ flux values.","core_discovery":"On the paper's own terms, the central discovery is that the mid-infrared flux densities of Eta Car's Homunculus nebula from 8 to 20 $\\mu\\mathrm{m}$ show no long-term decline since the first available mid-infrared photometry in 1968. The paper reconstructs the spectral energy distribution of the spatially integrated nebula over fifty years and finds that, within the 13–22 percent scatter of the heterogeneous measurements, the flux is stable. It therefore concludes that Eta Car's bolometric luminosity has probably been stable over the last five decades, at about $4.6 \\times 10^6\\,L_\\odot$, and that the ultraviolet and optical brightening since the late 1990s is unrelated to the integrated mid-infrared flux. The paper states that circumstellar dust must be declining in our line of sight only, with short-term flux variations within about 25 percent of the mean levels still possible.","pith_inferences":["If line-of-sight dust clearing is the cause, the ultraviolet and optical brightening should eventually slow or stop once the sightline is largely cleared; continued photometry could set a dissipation timescale for the dusty clump.","The same epoch-comparison approach could separate luminosity changes from extinction changes in other luminous blue variables and supernova impostors with decades of infrared photometry.","A homogeneous space-based mid-infrared monitoring campaign across several orbital cycles would test whether the residual roughly 25 percent scatter hides a periodic signal tied to periastron, which the current sparse ground-based data cannot resolve.","If the integrated mid-infrared flux is a reliable luminosity tracer, a future eruptive episode should produce a mid-infrared brightening, giving observers a cleaner trigger for a true luminosity increase than the UV-optical behavior."],"forward_implications":["Eta Car's ultraviolet and optical brightening since the 1990s is a line-of-sight extinction change, not a change in bolometric luminosity.","Mass and Eddington-limit estimates for Eta Car remain anchored near $4.6 \\times 10^6\\,L_\\odot$, so earlier claims of a 25 percent luminosity decline over recent decades are not supported.","The absence of a mid-infrared decline weakens the idea that dust destruction has been reducing circumstellar extinction globally; the clearing is localized to our sightline.","Short-term variations of up to about 25 percent, possibly tied to the 5.5-year orbital cycle, remain consistent with the data and would require phase-resolved monitoring to confirm.","The two equatorial dust loops of the 'Butterfly' nebula have expanded at up to $0.01^{\\prime\\prime}\\,\\mathrm{yr}^{-1}$ (about 100 km s$^{-1}$) without brightness changes, suggesting an optically thin structure in simple expansion."],"supporting_citations":[{"why":"supplies the first 1968 mid-infrared photometry that anchors the start of the fifty-year light curve.","marker":"Westphal & Neugebauer 1969"},{"why":"provides 1970s near- and mid-infrared photometry, including the 2.3–4.9 $\\mu\\mathrm{m}$ values used to derive the luminosity, and contains the 11 $\\mu\\mathrm{m}$ table-versus-figure inconsistency the paper relies on when re-reading historical fluxes.","marker":"Gehrz et al. 1973"},{"why":"provides 1971–1972 mid-infrared measurements showing no variation larger than the uncertainties, a key early epoch in the stability claim.","marker":"Robinson et al. 1973"},{"why":"supplies the 35–175 $\\mu\\mathrm{m}$ flux values used to reconstruct the full spectral energy distribution and derive the total luminosity.","marker":"Harvey et al. 1978"},{"why":"adds 1980s photometry, some of which the paper re-estimates from isophotal contour maps, extending the coverage of the light curve.","marker":"Hackwell et al. 1986"},{"why":"reported a 30 percent decrease in the 10.5 $\\mu\\mathrm{m}$ silicate feature versus the 1970s; this is an earlier claim the paper's stable-flux result directly contradicts.","marker":"Russell et al. 1987"},{"why":"reported a significant 1990s mid-infrared decrease compared with the 1970s; another earlier claim tested against the reconstructed light curve.","marker":"Smith et al. 1995"},{"why":"claimed a 25 percent mid-infrared decline over 20 years from archived space-based spectra; the paper re-evaluates this claim and identifies a transcription error in its figure.","marker":"Morris et al. 2017"},{"why":"provides 1997–1998 N-band and 18 $\\mu\\mathrm{m}$ measurements used as the 1990s epoch and documents periastron-related brightening.","marker":"Polomski et al. 1999"},{"why":"supplies the calibration standard used for the new VISIR photometry, on which the modern anchor of the light curve rests.","marker":"Cohen et al. 1999"}],"fun_headline_variants":["Eta Car's mid-IR flux steady for 50 years","No long-term mid-IR change in Eta Car","Eta Car's luminosity stable over five decades","Eta Car's glow holds steady, dust clears"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The conclusion depends on treating fifty years of heterogeneous photometry — different telescopes, apertures, calibration standards, and analysis methods, including values read by hand from isophotal maps — as directly comparable within the quoted 13–22 percent uncertainties; if hidden systematic offsets between epochs are larger than that, a real decline of about 25 percent could be missed.","fun_headline_variants_meta":{"raw":{"variants":["Eta Car's mid-IR flux steady for 50 years","No long-term mid-IR change in Eta Car","Eta Car's luminosity stable over five decades","Eta Car's glow holds steady, dust clears"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000209,"raw_usage":{"total_tokens":1431,"prompt_tokens":993,"completion_tokens":438,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":609,"completion_tokens_details":{"reasoning_tokens":374}},"tokens_in":609,"tokens_out":438,"duration_ms":5068,"temperature":1.0,"reasoning_tokens":374,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T11:19:58.128443+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Re-reduce the archived space-based spectra and the published 1970s ground-based photometry with a single, homogeneous calibration and aperture prescription, and compare the resulting fluxes with the 2018 VISIR values; if the re-reduced historical fluxes fall systematically more than about 25 percent below the modern values at 8–20 $\\mu\\mathrm{m}$, the claimed fifty-year stability is disproved.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"supplies the first 1968 mid-infrared photometry that anchors the start of the fifty-year light curve."},{"cited_title":"D., Ney, E","cited_arxiv_id":null,"evidence_quote":"provides 1970s near- and mid-infrared photometry, including the 2.3–4.9 $\\mu\\mathrm{m}$ values used to derive the luminosity, and contains the 11 $\\mu\\mathrm{m}$ table-versus-figure inconsistency the paper relies on when re-reading historical fluxes."},{"cited_title":"R., & Thomas, J","cited_arxiv_id":null,"evidence_quote":"provides 1971–1972 mid-infrared measurements showing no variation larger than the uncertainties, a key early epoch in the stability claim."},{"cited_title":"M., Hoffmann, W","cited_arxiv_id":null,"evidence_quote":"supplies the 35–175 $\\mu\\mathrm{m}$ flux values used to reconstruct the full spectral energy distribution and derive the total luminosity."},{"cited_title":"A., Gehrz, R","cited_arxiv_id":null,"evidence_quote":"adds 1980s photometry, some of which the paper re-estimates from isophotal contour maps, extending the coverage of the light curve."},{"cited_title":"W., Lynch, D","cited_arxiv_id":null,"evidence_quote":"reported a 30 percent decrease in the 10.5 $\\mu\\mathrm{m}$ silicate feature versus the 1970s; this is an earlier claim the paper's stable-flux result directly contradicts."},{"cited_title":"H., Aitken, D","cited_arxiv_id":null,"evidence_quote":"reported a significant 1990s mid-infrared decrease compared with the 1970s; another earlier claim tested against the reconstructed light curve."},{"cited_title":"W., Gull, T","cited_arxiv_id":null,"evidence_quote":"claimed a 25 percent mid-infrared decline over 20 years from archived space-based spectra; the paper re-evaluates this claim and identifies a transcription error in its figure."},{"cited_title":"F., Telesco, C","cited_arxiv_id":null,"evidence_quote":"provides 1997–1998 N-band and 18 $\\mu\\mathrm{m}$ measurements used as the 1990s epoch and documents periastron-related brightening."}],"review_version":1}