{"id":"641eb587-917a-461a-aaee-f35fa14e7615","arxiv_id":"2412.19414","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"HETDEX's fourth data release presents 15,940 spectroscopically identified AGN over 62.9 square degrees, with a raw density of 253.4 per square degree and 10,499 secure redshifts.","lead":"HETDEX releases a new catalog of 15,940 active galactic nuclei (AGN) found by untargeted spectroscopy across 62.9 square degrees, including faint and red AGN that imaging-selected quasar surveys miss. The catalog roughly triples the previous HETDEX AGN sample and provides a resource for studying how supermassive black holes grow and shape their host galaxies.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The headline density and reach claims rest on SDSS-DR16Q-anchored completeness that the paper itself notes is sparse at g > 22.5, exactly the faint, narrow-line, red regime where HETDEX is claimed to add new AGN.","rationale":"The reader's primary weakest assumption matches the concern I identify: the SDSS DR16Q cross-match anchors completeness only where the catalog is not novel, and the faint end is uncalibrated. I mark agreement as partial rather than full because the reader also lists the FoF linking length as a second fragile premise. I do not find that premise equally load-bearing: Figure 2 shows the unique-AGN count decreasing steeply below δr about 5 arcsec but reaching roughly a constant near 16,000 for δr ≳ 5 arcsec, so the adopted 5 arcsec value is near the plateau and small changes would not threaten the headline count. The central claim, that untargeted IFU spectroscopy reveals a substantially larger and fainter AGN population than imaging-selected quasar surveys, depends on the faint-end completeness and contamination rates, and those are precisely the quantities the paper cannot calibrate with SDSS DR16Q. The paper is transparent about this limitation, which is why the issue supports a conditional verdict rather than rejection: the catalog and flags are released, but the quantitative density comparisons and the g ~ 25 reach should be interpreted with the uncalibrated faint-end caveat until an independent faint-end test is performed.","tokens_in":17096,"tokens_out":4913,"duration_ms":48019,"concrete_test":"Run an end-to-end injection-recovery experiment on real HDR4 datacubes. Inject synthetic AGN spectra spanning g = 22.5-25, with rest-frame EWs and FWHMs drawn from the catalog's faint sIBL and narrow-line parameter space, including FWHM near 1200 km/s and narrower, plus red continua, into noise-realized HETDEX datacubes at realistic sky positions. Run the full chain, including line search, LP, sBL, the SNR/chi-squared quality cuts used before visual inspection, and FoF grouping, and measure recovery and false-positive rates per g, EW, and FWHM bin. If the recovered fraction at g ~ 24-25 remains near the bright-end 80%, the faint-end density claims survive; if it drops materially, the 166.9 deg^-2 secure density and the g ~ 25 reach require downward revision or explicit corrections.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The most load-bearing premise is that the completeness and contamination curves in Figure 4 characterize the faint AGN population that motivates the catalog. The curves are built by cross-matching 5,872 SDSS DR16Q quasars covered by good HETDEX fibers, and the paper states 'the sample size is limited at g ≳ 22.5' (Section 5). The quoted recovery fractions, 93% for the pipeline and 80% for LP/sBL AGN selection, are therefore measured on brighter, bluer, mostly broad-line quasars. The abstract's new population, low-luminosity, narrow-line, and/or red AGN down to g ~ 25, is not represented in that benchmark. If faint AGN have lower line-to-continuum ratios, narrower lines, or redder continua, the LP 8σ/5σ and sBL FWHM > 1200 km/s thresholds will not recover them at the same rate; the faint-bin completeness at g > 22.5 is unknown. The reported raw densities, 253.4 deg^-2 and 166.9 deg^-2, and the comparison to SDSS 56 deg^-2, are uncalibrated in exactly the regime that makes them interesting. The contamination story is equally unanchored: at g > 22.5, 28% of redshifts are unconfirmed and 15% of AGN are unconfirmed, and the g ~ 24 peak is described as suspicious. This is a completeness/contamination risk, not a circularity: the SDSS benchmark is external, but it under-samples the discovery space. The FoF linking length is less load-bearing because Figure 2 shows the unique-AGN count plateaus for δr ≳ 5 arcsec, so modest changes around 5 arcsec would not change the headline 15,940 number.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper presents the AGN catalog from the fourth HETDEX data release (HDR4), an untargeted IFU spectroscopic survey covering 62.9 deg^2. The catalog contains 15,940 unique AGN, of which 10,499 have redshifts confirmed by line pairs or SDSS DR16Q matches; the remainder are single-line candidates, with 3,358 flagged as intermediate-broad-line AGN candidates. The raw AGN density is 253.4 deg^-2 (166.9 deg^-2 for the secure-redshift subset), which the authors compare to ~56 deg^-2 for SDSS quasars. The paper also reports that 39% of redshift-confirmed AGN have spatially extended emission-line regions. Completeness is estimated by cross-matching SDSS DR16Q quasars with HETDEX fibers (5,872 matches) and is stated to be limited at g > 22.5.","tokens_in":17359,"tokens_out":8307,"duration_ms":70344,"significance":"If the catalog is robust, it provides an important untargeted spectroscopic AGN sample that can probe lower-luminosity, narrow-line, and red AGN missed by imaging-selected quasar surveys, and the publicly released FITS products with detailed flags will be a valuable resource. The paper is generally transparent: flags for confirmed vs estimated redshifts and secure vs candidate AGN are provided, the duplicate-removal method is tested with friend-of-friend linking-length experiments, and the completeness is benchmarked against an external catalog. The main risk to significance is that the completeness calibration is sparsely sampled in the faint regime where the catalog's novelty lies, so the headline densities and faint-end statistics are not yet firmly established.","major_comments":[{"comment":"The completeness and contamination curves are anchored to a cross-match with SDSS DR16Q, with 5,872 quasars covered by good HETDEX fibers; the paper itself notes 'the sample size is limited at g > 22.5.' The recovery fractions (93% pipeline, 80% AGN selection) are therefore measured on a brighter, bluer, mostly broad-line quasar population than the low-luminosity, narrow-line, red AGN the catalog is claimed to add. Because the LP thresholds (8sigma/5sigma) and the sBL FWHM > 1200 km/s criterion depend on line properties that may differ in the faint population, the completeness curves in Figure 4 cannot be extrapolated to g > 22.5, exactly the regime where the densities 253.4 deg^-2 and 166.9 deg^-2 and the 'down to g~25' claim are made. The authors should provide a completeness model as a function of line flux, equivalent width, and line width (as in Liu et al. 2022c) or cross-match to deeper surveys (e.g., COSMOS, CANDELS) to calibrate the faint end, or restrict the quantitative density claims to the calibrated bright regime.","section":"Section 5, Figure 4"},{"comment":"The comparison 'HETDEX secure AGN density 166.9 deg^-2 ... significantly higher than that of the SDSS quasars (56 deg^-2)' is not a like-for-like comparison. The SDSS value is for color-selected, spectroscopically confirmed quasars over a different effective area and depth, and the HETDEX number is a raw detection density without completeness correction. Incompleteness at the faint end (which the paper acknowledges) and differing selection functions mean that this comparison does not by itself demonstrate a higher space density of AGN. The text should either present completeness-corrected density estimates or explicitly label the comparison as raw detection rates with caveats.","section":"Table 1; Section 5"},{"comment":"The 39% extended-emission-line fraction (4,060/10,499) is a headline result but is based on a threshold criterion ('member detections >5sigma more extended than their relevant PSF models (FWHM_virus + 3 FWHM_virus_err)/2') whose false-positive rate from PSF errors, source blending, or multiple-fiber detections is not quantified. The interpretation as outflows or ionized IGM is speculative and needs support from 2D surface-brightness fitting (as for the example agnid=407) or a control sample of stars or compact sources. As written, the abstract presents this number without the required validation.","section":"Section 4"}],"minor_comments":[{"comment":"The text states 'we release the 15,877 AGN identified from the fourth HETDEX Data Release (HDR4),' but Table 1 and the abstract give 15,940; this numerical inconsistency should be corrected.","section":"Section 1"},{"comment":"The word 'diamter' is a typo for 'diameter.'","section":"Section 2"},{"comment":"The text says 'all 500 million HETDEX HDR4 fibers' were cross-matched, but Section 2 implies 345,874 IFU observations with 34,944 fibers each; please clarify what 'fibers' means here.","section":"Section 5"},{"comment":"The gray SDSS DR16Q histogram is described as 'scaled ... to generally match the number of HETDEX AGN at z~1.5'; the scaling factor should be stated for reproducibility.","section":"Figure 5"},{"comment":"The format string 'yyyymmddsss multi bbb ccc ddd aa fff' for fiberid is cryptic; an explicit example would aid users of the catalog.","section":"Section 7, Column 18"}],"recommendation":"major_revision","confidential_remarks":"The completeness issue in the faint regime (g > 22.5) is the main correctness risk and directly affects the abstract's density and reach claims. The authors are aware of the limitation, but the presentation in the abstract and Section 5 goes beyond what the SDSS-DR16Q-anchored calibration supports. The density comparison with SDSS should be tempered or made completeness-corrected. The extended-emission fraction also needs validation before it is used as a science result. Otherwise, the catalog construction is careful and the public data products will be useful."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Bottom line: this is a solid data-release paper for a genuinely bigger AGN catalog, and the authors are honest about its main weakness: the completeness that would justify the headline density claims is only measured on bright, blue SDSS quasars, not on the faint, narrow-line, red AGN that make HETDEX interesting.\n\nWhat's actually new: the HDR4 footprint (62.9 deg^2) roughly triples the earlier HDR2 sample to 15,940 AGN, with the SDSS match updated from DR14Q to DR16Q. The catalog construction is transparent: flags separate confirmed versus estimated redshifts and secure versus candidate AGN, the friend-of-friend duplicate grouping is tested against linking length with a clear plateau at ~5\", and visual inspection is described. The completeness test anchored to SDSS DR16Q is external, so there's no circularity. The paper even flags the suspicious g~24 peak and the low completeness at g>22.5.\n\nThe soft spot is real and load-bearing for the \"253.4 deg^-2 versus SDSS 56 deg^-2\" comparison. Only 5,872 SDSS quasars are covered by good HETDEX fibers, and nearly all are brighter than g=22.5. The 93% pipeline and 80% AGN-selection recovery fractions therefore don't tell you how the LP 8-sigma/5-sigma and sBL FWHM>1200 km/s thresholds behave for the faint population. If those AGN have weaker lines relative to continuum or narrower or redder continua, the recovery rate could drop, and the density comparison would not hold. The 39% extended-emission fraction also lacks an uncertainty and depends on the PSF-extension threshold. These don't invalidate the catalog as a resource, but they do mean the headline numbers should be read as upper limits or as brightness-limited rather than as true faint-end densities.\n\nWho this is for: anyone working on AGN luminosity functions, clustering, or narrow-line AGN populations. They should use the secure subset (agn_flag=1, zflag=1) and treat the faint end cautiously. The paper deserves a serious referee; I'd send it to review with a request to tone down the density comparison and either add a faint-end completeness estimate (e.g., injecting simulated faint AGN into the data) or state explicitly that the completeness is uncalibrated beyond g~22.5. No fatal flaw, just an overreach in the current wording.","headline":"Bigger HETDEX AGN catalog with transparent methods, but the faint-end completeness is unmeasured exactly where it claims to find new AGN.","tokens_in":18114,"tokens_out":2612,"would_cite":true,"duration_ms":22566,"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":"This paper claims that an untargeted integral-field spectroscopic survey, HETDEX, has produced a catalog of 15,940 active galactic nuclei whose redshift-secure density is about three times that of SDSS quasars, revealing a large…","keywords":["active galactic nuclei","AGN catalog","integral field spectroscopy","untargeted survey","quasar density","survey completeness","emission-line regions","HETDEX HDR4"],"falsifier":"Take the 5,441 AGN with estimated redshifts and obtain follow-up spectra for a random sample: if a large fraction show only narrow lines or star-forming signatures, then the raw density claim (253.4 $deg^{-2}$) is inflated, while the secure density (166.9 $deg^{-2}$) stands. Alternatively, compute the recovery fraction of HETDEX AGN against a deep X-ray-selected sample in the COSMOS or GOODS-N fields: if recovery drops below the SDSS-based curve at g>22.5, the completeness model fails in exactly the regime where the survey claims novelty.","tokens_in":16774,"feed_emoji":"🔭","tokens_out":8827,"duration_ms":70345,"temperature":0.7,"pith_summary":"Sitting inside the fourth HETDEX data release, this paper claims that an untargeted integral-field spectroscopic survey—one that picks no targets by color or image—finds 15,940 active galactic nuclei over 62.9 square degrees, with 10,499 secure redshifts. That gives a raw AGN density of 253.4 $deg^{-2}$ and a secure density of 166.9 $deg^{-2}$, about three times the 56 $deg^{-2}$ of SDSS quasars. The paper's point is that imaging-selected quasar surveys miss a large population of low-luminosity, narrow-line, and red AGN down to g~25 that only blank-sky spectroscopy can reveal. It also finds that 39% of the redshift-confirmed AGN have emission-line regions more extended than the point-spread function, plausibly outflows or ionized halo gas around the black hole.","feed_headline":"Untargeted survey finds 15,940 AGN, triple SDSS density","feed_subtitle":"HETDEX's blind spectroscopy reaches g~25 and a secure 166.9 AGN per square degree.","key_machinery":"The identification machinery is a two-track emission-line search on the HETDEX line and continuum detections. The line-pair method finds AGN by requiring two characteristic lines (e.g., C IV 1549 plus a companion) at >8 sigma and >5 sigma; the single broad-line method fits multi-Gaussian profiles and flags lines broader than FWHM 1200 km/s at >5 sigma, assigning a redshift guess from the line's equivalent width. A cross-match with SDSS DR16Q confirms or corrects redshifts and recovers a few AGN whose lines fall outside the 3500-5500 Å window. To build the unique catalog, friends-of-friends grouping with linking length (5.0\", 0.1 in redshift) merges the 63,827 detections into 15,940 AGN; the choice is justified by FoF experiments showing the count plateaus near 16,000 for δr >= 5\". The pipeline's completeness and contamination curves are constructed by cross-matching all fibers with SDSS DR16Q.","core_discovery":"The central discovery is that the AGN population at these depths is substantially larger and fainter than imaging-selected catalogs suggest, and that an IFU survey can recover it without selection cuts. From 345,874 IFU observations the paper builds a catalog of 15,940 AGN at z=0.1-4.6; 10,499 have redshifts confirmed by line pairs or SDSS matches, and the secure subset alone has a surface density of 166.9 $deg^{-2}$ versus 56 $deg^{-2}$ for SDSS quasars. The remaining 5,441 are single broad-line candidates, of which 3,358 have intermediate widths (FWHM ~ 1200 km/s) and are explicitly flagged as possible non-AGN. Additionally, 4,060 of the 10,499 secure AGN show extended emission-line regions at >3 $\\sigma$ beyond image quality, which the paper interprets as strong outflows or ionized circum/intergalactic gas.","pith_inferences":["A testable consequence the paper leaves open: if the SDSS-anchored completeness curves do not transfer to faint narrow-line AGN, the true recoverable density could be higher than 253.4 deg^-2 (if HETDEX detects line-emitting AGN that SDSS colors reject) or lower (if faint AGN have weak lines); this can be checked by comparing to X-ray- or radio-selected AGN in the same fields.","The friend-of-friend choice directly sets the 15,940 count and the 39% extended fraction; re-deriving the extended-fraction statistic from per-object radial profiles rather than group membership would make it independent of that choice.","The second magnitude peak at g~24, dominated by intermediate-broad-line candidates, is a natural target for narrow-band follow-up: if those objects are mostly star-forming LAEs, the raw density should be quoted with the agn flag=0 removed.","Extending the same blind selection to the completed 540 deg^2 survey would test whether the spring/fall field densities (268 and 233 deg^-2) are cosmic variance or a real large-scale AGN enhancement."],"forward_implications":["If the density comparison is correct, luminosity functions and black-hole accretion rates built from imaging-selected quasars are missing the majority of the AGN population at g < 25.","The 39% extended-emission fraction means typical, not just rare, AGN can be used to map outflows and gas halos with spatially resolved spectra.","Users can isolate a 'secure AGN' sample of 12,582 objects using the agn flag, removing the 3,358 intermediate-broad-line candidates that may be star-forming contaminants.","The catalog's reach to z=0.1-4.6 (via SDSS matches beyond the nominal 0.25-4.32 line window) makes it a resource for AGN demographics across cosmic time."],"supporting_citations":[{"why":"Defines the line-pair and single broad-line AGN identification methods and the visual inspection flow that this paper applies to HDR4.","marker":"Liu et al. 2022a"},{"why":"Supplies the HETDEX survey design, data reduction, and emission-line detection algorithms that produce the parent line and continuum catalogs.","marker":"Gebhardt et al. 2021"},{"why":"Provides the SDSS DR16Q quasar catalog used for redshift confirmation, missed-AGN recovery, and completeness cross-matching.","marker":"Lyke et al. 2020"},{"why":"Source of the SDSS DR14Q quasar density (56 deg^-2) used as the comparison benchmark for the density claim.","marker":"Pâris et al. 2018"},{"why":"Supplies the completeness modeling formalism for HETDEX AGN against line wavelength, flux, width, and position, referenced for the survey's completeness framework.","marker":"Liu et al. 2022c"},{"why":"Provides the ELiXer software used to estimate HETDEX g-band magnitudes and fit line parameters in the catalog.","marker":"Davis et al. 2023"},{"why":"Describes the VIRUS instrument whose IFU design enables the untargeted multi-fiber spectroscopy central to the survey.","marker":"Hill et al. 2021"}],"fun_headline_variants":["No imaging, triple density: 15,940 AGN","HETDEX triples AGN density without any cuts","Down to g~25: blind survey reveals 15,940 AGN","4,060 AGN show extended emission: outflows or intergalactic gas?","Blind spectra reveal hidden AGN population at z~4.6"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The headline density advantage assumes that the AGN recovery fractions measured against bright SDSS quasars transfer to the faint, narrow-line AGN this survey uniquely finds, and that the chosen friend-of-friend linking length of 5 arcseconds separates real AGN from extended emission rather than merging or splitting them.","fun_headline_variants_meta":{"raw":{"variants":["No imaging, triple density: 15,940 AGN","HETDEX triples AGN density without any cuts","Down to g~25: blind survey reveals 15,940 AGN","4,060 AGN show extended emission: outflows or intergalactic gas?","Blind spectra reveal hidden AGN population at z~4.6"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001189,"raw_usage":{"total_tokens":4956,"prompt_tokens":1043,"completion_tokens":3913,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":659,"completion_tokens_details":{"reasoning_tokens":3817}},"tokens_in":659,"tokens_out":3913,"duration_ms":27515,"temperature":1.0,"reasoning_tokens":3817,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-11T00:37:06.483063+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Take the 5,441 AGN with estimated redshifts and obtain follow-up spectra for a random sample: if a large fraction show only narrow lines or star-forming signatures, then the raw density claim (253.4 $deg^{-2}$) is inflated, while the secure density (166.9 $deg^{-2}$) stands. Alternatively, compute the recovery fraction of HETDEX AGN against a deep X-ray-selected sample in the COSMOS or GOODS-N fields: if recovery drops below the SDSS-based curve at g>22.5, the completeness model fails in exactly the regime where the survey claims novelty.","supporting_citations":[],"review_version":1}