{"id":"be924c2f-1ace-4374-a6e1-68d5f223293f","arxiv_id":"2504.18972","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"Using new strong-line calibrators on GASP and MaNGA data, the authors report that ram-pressure stripping does not alter AGN metallicity gradients and that AGN hosts are more metal-enriched in their nuclear regions than star-forming galaxies.","lead":"This paper measures oxygen abundances across 62 AGN-hosting galaxies, 10 being stripped by ram pressure in clusters and 52 in the field, plus 83 star-forming controls, finding that ram pressure does not change AGN metallicity profiles. AGN hosts appear more metal-rich than star-forming galaxies, especially in nuclear regions, which the authors interpret as AGN-driven metal pollution.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The claimed nuclear metal enhancement depends on the mutual calibration of the SF and AGN abundance scales; independent AGN calibrators shift nuclear abundances by up to 0.4 dex, so the 1.8-2.3x R-ratio may be partly calibration-induced.","rationale":"The reader's weakest assumption identifies exactly the load-bearing condition: the P24 SF and AGN calibrators must define a mutually consistent absolute metallicity scale for the AGN-SF offset to be physical. I agree with this assessment. The concern is concrete and quantitative: the AGN-SF differences entering the R-ratio are small (roughly 0.1 dex in the nuclear annulus), while the P24 AGN calibrator disagrees with independent AGN calibrators by up to ~0.4 dex at high metallicity, and the sign of the disagreement is such that a more standard AGN SED would reduce the nuclear enhancement. The paper does perform a relevant test in Fig. 13, but it only varies Epeak in one model grid and does not propagate that variation into the headline R-ratio or compare against the SF calibrator on the same absolute scale. This is an addressable systematic, not a fatal flaw, and the authors are transparent about the assumptions; the conditional verdict and the recommendation for additional calibration checks are appropriate. I therefore do not propose changing the reader's verdict.","tokens_in":31209,"tokens_out":2550,"duration_ms":29733,"concrete_test":"Re-run the full AGN abundance measurement with NebulaBayes using the Thomas et al. (2016) NLR SED at log Epeak=-1.35 (the value adopted by Thomas et al. 2019) instead of the alpha=-2.0 power law, keeping all other choices identical; then recompute the median radial profiles of AGN and SF galaxies and the R-ratio from Sec. 4.2. If R drops substantially (e.g., below ~1.2) or the r<0.5Re AGN-SF offset becomes comparable to the annulus-to-annulus scatter, the claimed nuclear metal pollution is not robust to the assumed AGN ionizing continuum. As a complementary check, recompute R using the SB98 or C20 AGN calibrator together with the P24 SF calibrator to bound the calibration-induced range of the enhancement.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim in Sec. 4.2 is that AGN hosts are more metal-enriched than SF galaxies at all radii, with a nuclear (r<0.5Re) enhancement 1.8-2.3 times larger than the disk enhancement at r~1.25Re. This R-ratio is computed from differences between medians obtained with two different calibrators: Eq. 2 for SF spaxels and Eq. 3 for AGN spaxels. The inference is therefore only as secure as the assumed consistency of the two abundance scales. Section 5.3 shows that the P24 AGN calibrator exceeds the independent SB98 and C20 calibrators by up to ~0.4 dex, with the offset growing with metallicity. Since AGN nuclear spaxels are the highest-metallicity spaxels in the sample, a systematic overestimate of the AGN scale at high Z would preferentially inflate the nuclear AGN-SF difference and hence the R-ratio. The authors' own robustness test (Fig. 13) shows that adopting a harder AGN ionizing continuum (log Epeak=-2.0) lowers AGN abundances by 0.23±0.09 dex, more than twice the quoted 0.1 dex uncertainty; although the authors argue that 100 eV is implausibly hard, they do not re-derive the R-ratio under the T18 SED at the plausible Epeak~40-50 eV. Because the SF and AGN scales are anchored to different ionizing spectra and different N/O prescriptions, the observed nuclear enhancement could in part be a calibration artifact rather than a physical metal excess.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper measures spatially resolved gas-phase oxygen abundances in 10 ram-pressure-stripped AGN hosts from GASP and 52 field AGN hosts from MaNGA DR15, together with SF control samples, using new strong-line calibrators derived from Cloudy photoionization models and NebulaBayes. The authors report that the metallicity distributions of stripped and field AGN are statistically indistinguishable within 1.5 Re, with only two stripped AGNs (JO206 and JO171) showing lower abundances. They further report that AGN hosts are more metal-enriched than SF galaxies at all radii, with a nuclear (r<0.5 Re) enhancement larger than the disk enhancement at r~1.25 Re by a factor R~1.8-2.3, which they interpret as evidence that AGN activity enriches the nuclear gas.","tokens_in":31595,"tokens_out":4821,"duration_ms":52511,"significance":"If the central claim holds, the paper provides some of the first spatially resolved evidence that AGN activity is associated with a nuclear metal excess relative to SF galaxies, and it offers a useful null result on the role of ram-pressure stripping in shaping AGN-host metallicity. The construction of SF, AGN, and composite calibrators within the same Cloudy/NebulaBayes framework is a methodological strength, as are the explicit comparisons with the DN22, SB98, C20, and T18 calibrators. The main limitation is that the headline R ratio has no quoted uncertainty and is derived from two abundance calibrators whose mutual consistency is not fully established; this needs to be addressed before the physical interpretation can be considered secure.","major_comments":[{"comment":"The authors show that the P24 AGN calibrator (Eq. 3) lies systematically above the SB98 and C20 AGN calibrators by up to ~0.4 dex, with the offset increasing with metallicity. Because the central R statistic in Section 4.2 is a difference between AGN abundances from Eq. (3) and SF abundances from Eq. (2), and because AGN nuclear spaxels are the most metal-rich spaxels, a metallicity-dependent offset in Eq. (3) will preferentially inflate the numerator of R. The discussion in Section 5.3 demonstrates the discrepancy but does not propagate it into R or re-derive R using SB98 or C20 for the AGN spaxels. Please quantify the calibration-induced systematic on the nuclear enhancement, or show explicitly that the enhancement survives when literature AGN calibrators are used.","section":"5.3, Fig. 12, Eq. (3)"},{"comment":"The quantity R (reported as 1.77-2.24) is quoted without any uncertainty or significance test. The medians entering the numerator and denominator have substantial scatter (the 25th-75th percentiles shown in Fig. 6 are ~0.1 dex or larger), and the sample sizes are modest: 10 AGN-RPS, 52 AGN-FS, and 83 SF galaxies. A bootstrap or permutation test should be used to establish whether R is significantly greater than 1 and to attach confidence intervals to the quoted factors. Similarly, the statement that AGN hosts are more metal-enriched than SF galaxies 'at any given radius' is not accompanied by a statistical test at individual radii.","section":"4.2, definition of R"},{"comment":"JW100 is discarded from the AGN-RPS sample because its residual from the new calibrator plane is mu>0.1 dex. This is a post-hoc exclusion based on the same calibrator that is later used to measure all abundances, and the AGN-RPS sample contains only 11 objects. The exclusion therefore has large leverage on the RPS-versus-field comparison presented in Section 4.1. Please show the main RPS/field results with JW100 included, and if possible justify the exclusion using an independent diagnostic rather than the calibrator residual.","section":"3.2, exclusion of JW100"}],"minor_comments":[{"comment":"The definition of O3N2 should be log{([O iii]/Hβ)/([N ii]/Hα)}; the current notation, with the multiplicative expression, is ambiguous.","section":"Eq. (8)"},{"comment":"The text mixes Kendall tau and Spearman coefficients in the same paragraph, with symbols kτAGN and spAGN; please use consistent notation.","section":"4.1"},{"comment":"The caption states that 'the lines show the probability density distribution evaluated using KDE', but the figure panels show scatter points; please clarify what the lines represent and where they appear.","section":"Fig. 4 caption"},{"comment":"The sentence 'the main assumptions on the models are the stellar age ... the gas density ... and a slope of the AGN continuum alpha=-2.0' is a footnote; the information would be more visible if placed in the main text.","section":"3.2"},{"comment":"The conclusion says 'see Figure 4' for the nuclear-to-disk enhancement factor, but the R values are shown in Figure 6; please correct the cross-reference.","section":"Conclusion (ii)"}],"recommendation":"major_revision","confidential_remarks":"To the editor: the calibration systematic is the key issue. It is not, in my view, grounds for rejection, because the authors have the machinery in hand to re-derive R under literature calibrators and to attach uncertainties. The JW100 exclusion is a secondary but real concern for the RPS comparison. The paper would be suitable for A&A after these points are addressed."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Hi — worth a look, but not for the headline result. The genuinely new things are the P24 calibrators (a self-consistent SF/AGN/Composite set from one Cloudy/NebulaBayes grid) and the first direct comparison of metallicity gradients in RPS-stripped AGNs versus field AGNs. That comparison is clean and honestly reported: no significant difference, two outliers discussed properly, and the authors say plainly that the sample is too small to draw strong conclusions. The re-analysis of the DN22 sample, showing the sign flip in the nuclear excess, is also useful and suggests the disagreement is method-driven rather than physical.\n\nWhat the paper does well: the calibrator surfaces are shown to reproduce the spaxel-by-spaxel Bayesian estimates, the cross-checks against SB98, C20, and T18 are included rather than swept under the rug, and the appendices on the outliers are thorough. The authors are also clear about the spatial-resolution caveat. I believe them that the RPS effect, if any, is weak within 1.5 Re.\n\nThe soft spot is the central AGN enrichment claim. The R-ratio (1.8–2.3) is computed as a ratio of differences between medians derived with two different calibrators, and it is quoted without any uncertainty. Meanwhile, their own Figure 12 shows the P24 AGN calibrator running 0.1–0.4 dex above SB98 and C20 at the high metallicities that dominate nuclear spaxels. That is a systematic larger than the claimed nuclear excess. The Figure 13 test with T18 models is the right idea, but they only show how the AGN abundances shift under a hard continuum; they do not re-derive R under the plausible Epeak=40–50 eV SED, where the shift is already non-trivial. So the 1.8–2.3 factor may in part be a calibration artifact. The JW100 exclusion is also a bit post hoc—given plausible physical justification, but it removes one of only ten RPS AGNs.\n\nNone of this makes the paper unserious. It is a solid technical contribution with a central claim that needs stronger statistical machinery and a more explicit propagation of the calibration systematic. I would send it to a good referee: the calibrators deserve scrutiny, and the RPS comparison is worth publishing even if the AGN pollution argument gets softened. For my own work, I would cite it for the P24 SF and Composite calibrators; I would not yet cite the AGN pollution result as established.\n\nRecommendation: engage with it in reading group, but assign a referee who knows strong-line abundance systematics cold.","headline":"Read it for the new consistent calibrator set and the careful RPS-vs-field comparison; the case for AGN nuclear metal pollution is not yet made because the claimed offset sits within the calibration uncertainty.","tokens_in":32159,"tokens_out":2262,"would_cite":true,"duration_ms":26259,"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":"AGN host galaxies are more metal-rich at every radius than star-forming galaxies, with a nuclear excess 1.8–2.3 times the disk excess, pointing to AGN-driven metal pollution.","keywords":["active galactic nuclei","gas-phase metallicity","oxygen abundance gradients","ram-pressure stripping","strong-line calibrators","photoionization models","integral-field spectroscopy","nuclear metal enrichment"],"falsifier":"Measure direct electron-temperature abundances from auroral lines such as [O III] $\\lambda$4363 in the $r < 0.5\\,R_e$ annuli of a few dozen AGN hosts and mass-matched star-forming galaxies; if the AGN–SF nuclear offset that produces the 1.8–2.3 factor disappears or reverses on this independent scale, the AGN metal-pollution claim is falsified rather than calibration-driven.","tokens_in":31048,"feed_emoji":"🌌","tokens_out":11973,"duration_ms":107745,"temperature":0.7,"pith_summary":"Using spatially resolved spectra of nearby galaxies, this paper compares the gas-phase oxygen abundance of galaxies whose centers host an active galactic nucleus (AGN) with that of galaxies forming stars. It finds that AGN hosts are more metal-rich at every radius, and that the excess over star-forming galaxies is between 1.8 and 2.3 times larger in the nuclear region ($r < 0.5\\,R_e$) than in the disk ($r \\sim 1.25\\,R_e$). The authors take this as evidence that AGN activity itself pollutes the nuclear gas, rather than the enrichment being a by-product of the host galaxy's star-formation history. The same data show no significant difference between ram-pressure-stripped AGN hosts in clusters and undisturbed field AGN hosts within 1.5 effective radii, indicating that stripping does not alter the metal content of AGN hosts at these radii.","feed_headline":"AGN cores carry 1.8–2.3x the metal excess of disks","feed_subtitle":"Spatially resolved oxygen maps place the AGN metal excess in the nucleus, not the disk.","key_machinery":"The argument is carried by three new strong-line metallicity calibrators, one each for star-forming, AGN-ionized, and composite regions. They express $12+\\log(\\mathrm{O/H})$ as a least-squares fit in the observed line ratios $\\log([\\mathrm{N\\,II}]/[\\mathrm{S\\,II}])$ and $\\log([\\mathrm{O\\,III}]/[\\mathrm{S\\,II}])$, with coefficients calibrated against Cloudy photoionization models fitted to the data through the Bayesian code NebulaBayes. Because the same model set and inference code are used for all three ionizing regimes, the calibrators put AGN and star-forming abundances on a common scale; this is what makes the nuclear-vs-disk enhancement ratio a physical comparison rather than a cross-calibration artifact. The paper also uses the BPT diagnostic diagram, which classifies the ionizing source from optical line ratios, to assign spaxels to the three regimes.","core_discovery":"The paper's central claim is that active galactic nuclei change the metal budget of their host galaxies locally: AGN hosts show higher gas-phase oxygen abundances than star-forming galaxies at any given radius, and the steepness of the radial gradient in the nuclear annulus $r < 0.5\\,R_e$ is larger in AGN hosts. Quantitatively, the AGN–SF abundance enhancement in the nucleus is between 1.8 and 2.3 times the enhancement measured in the disk at $r \\sim 1.25\\,R_e$, in all stellar-mass bins considered. The authors interpret this as nuclear metal pollution caused by the AGN. The secondary claim is that ram-pressure stripping plays no measurable role in this budget: the metallicity distributions of 10 stripped and 52 undisturbed AGN hosts are statistically indistinguishable within $r < 1.5\\,R_e$, with only two stripped hosts showing lower abundances at every radius.","pith_inferences":["If the nuclear excess is real, the most economical physical explanations are dust destruction near the black hole or a top-heavy stellar initial mass function; these make different predictions for dust-to-gas ratio and stellar abundance patterns that the current data do not test.","The reported factor 1.8–2.3 is tied to the new AGN calibrator, which reads 0.1–0.4 dex higher than literature AGN calibrators at high metallicity; recomputing the ratio on a literature scale would show how much of the enhancement is absolute-scale dependent.","A harder AGN ionizing continuum than the assumed power-law slope would lower derived AGN abundances by up to 0.23 dex; if the true continuum sits at the hard end, the nuclear enhancement could shrink below significance, so fitting the same galaxies with harder continuum models is a natural next step.","The two stripped outliers with low abundances at all radii may be galaxies observed after the ram-pressure peak, where simulations predict both star formation and black-hole accretion to decline; targeted follow-up could test whether their metal-poor gas is accreted rather than stripped."],"forward_implications":["If AGN activity enriches nuclear gas, metallicity gradients can be used as a fossil tracer of past AGN feedback in galaxies where the AGN is no longer bright.","The new calibrators let observers measure star-forming, AGN, and composite-region abundances without switching between inconsistent literature calibrators, so AGN–SF metallicity comparisons become less dependent on method.","The ram-pressure result means that, at the stripping stages sampled, environment does not alter the metal content inside 1.5 effective radii, so AGN metallicity studies need not separate stripped from undisturbed hosts at these radii.","The lack of correlation between gradient slope and nuclear [O III] luminosity or stellar mass implies that the enrichment mechanism does not simply scale with AGN brightness or host mass."],"supporting_citations":[{"why":"Establishes the sample selection and the Cloudy+NebulaBayes model grid on which the calibrators are built, and first reported nuclear AGN enrichment.","marker":"P23"},{"why":"Provides the Cloudy photoionization code used to generate the H II and AGN model grids.","marker":"Ferland et al. 2017"},{"why":"Provides NebulaBayes, the Bayesian code that matches observed [O III]/[S II] and [N II]/[S II] ratios to model predictions.","marker":"Thomas et al. 2018a"},{"why":"Supplies the AGN photoionization-model framework and comparison standard invoked to interpret the sensitivity of the AGN calibrator to the ionizing continuum.","marker":"Thomas et al. 2019"},{"why":"Is the prior study reporting lower nuclear AGN metallicities; its linear-fit procedure is reproduced and shown to reverse sign when consistent calibrators are used.","marker":"Nascimento et al. 2022"},{"why":"Provides a literature AGN strong-line calibrator used to check the new AGN abundance scale.","marker":"Storchi-Bergmann et al. 1998"},{"why":"Provides a second independent AGN strong-line calibrator used in the same cross-check.","marker":"Carvalho et al. 2020"},{"why":"Defines the survey of ram-pressure-stripped galaxies from which the stripped AGN hosts are drawn.","marker":"Poggianti et al. 2017"},{"why":"Defines the survey and data products from which the field AGN and star-forming controls are drawn.","marker":"Bundy et al. 2015"},{"why":"Supplies the simulations linking ram-pressure intensity to black-hole accretion and star formation that motivate why RPS effects are expected mainly after the stripping peak.","marker":"Ricarte et al. 2020"}],"fun_headline_variants":["AGN metal enrichment is nuclear, not disk-wide","Ram pressure doesn't alter AGN metallicity","AGN nuclei show 1.8–2.3x metal boost","Stripped AGN keep same metallicity as field"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The star-forming and AGN calibrators define a common, mutually consistent absolute metallicity scale, so the measured AGN–SF abundance offset reflects real enrichment rather than a difference in how each calibrator was built.","fun_headline_variants_meta":{"raw":{"variants":["AGN metal enrichment is nuclear, not disk-wide","Ram pressure doesn't alter AGN metallicity","AGN nuclei show 1.8–2.3x metal boost","Stripped AGN keep same metallicity as field"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.00036,"raw_usage":{"total_tokens":2002,"prompt_tokens":1059,"completion_tokens":943,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":675,"completion_tokens_details":{"reasoning_tokens":875}},"tokens_in":675,"tokens_out":943,"duration_ms":8478,"temperature":1.0,"reasoning_tokens":875,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-16T10:05:09.334435+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure direct electron-temperature abundances from auroral lines such as [O III] $\\lambda$4363 in the $r < 0.5\\,R_e$ annuli of a few dozen AGN hosts and mass-matched star-forming galaxies; if the AGN–SF nuclear offset that produces the 1.8–2.3 factor disappears or reverses on this independent scale, the AGN metal-pollution claim is falsified rather than calibration-driven.","supporting_citations":[{"cited_title":"D., Kewley, L","cited_arxiv_id":null,"evidence_quote":"Supplies the AGN photoionization-model framework and comparison standard invoked to interpret the sensitivity of the AGN calibrator to the ionizing continuum."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Is the prior study reporting lower nuclear AGN metallicities; its linear-fit procedure is reproduced and shown to reverse sign when consistent calibrators are used."},{"cited_title":"M., Moretti, A., Gullieuszik, M., et al","cited_arxiv_id":null,"evidence_quote":"Defines the survey of ram-pressure-stripped galaxies from which the stripped AGN hosts are drawn."},{"cited_title":"2020, ApJL, 895, L8","cited_arxiv_id":null,"evidence_quote":"Supplies the simulations linking ram-pressure intensity to black-hole accretion and star formation that motivate why RPS effects are expected mainly after the stripping peak."}],"review_version":1}