{"id":"bdb624cd-2526-4f49-9698-dfb444972411","arxiv_id":"2505.16738","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":1,"one_line_summary":"A new pipeline maps H-alpha emission across the Fornax cluster, finding 77 emitters that trace recent infall and pre-processing out to 4 Rvir.","lead":"Using narrow-band images from the S-PLUS survey, the authors mapped hydrogen-alpha emission in 77 galaxies of the Fornax cluster out to four times its virial radius. The maps reveal that most emitters are recent infallers and that many disturbed, low-mass galaxies sit beyond the cluster edge, suggesting preprocessing before cluster entry.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Detection threshold calibrated on 13 central MUSE fields is assumed uniform across S-PLUS fields varying by up to 1 mag, potentially biasing emitter identification and the derived infall and morphology fractions.","rationale":"The reader's weakest assumption matches the most load-bearing concern I could identify. The paper's central claim—detecting and mapping Hα+[NII] emission in 77 Fornax members and using the resulting sample to infer infall histories and pre-processing—rests on the reliability of the detection threshold. The validation uses 13 MUSE galaxies in central fields, while the scientific conclusions extend to 4 Rvir, where field depths vary significantly. The paper itself notes the depth variation but does not correct for it. A uniform threshold would cause a flux-dependent selection function that is spatially varying, directly undermining the quoted fractions of disturbed/infalling galaxies and their clustercentric distribution. Other potential issues (e.g., fixed redshift in the Three Filter Method, missing [NII]/dust corrections, incomplete spectroscopic parent sample) are either acknowledged by the authors as limitations or affect absolute flux calibration more than sample selection. The proposed injection test would definitively settle whether the threshold is uniform or whether per-field completeness corrections are needed. Since the reader already flagged this and assigned CONDITIONAL, my recommendation is UNCHANGED: the concern is valid and supports the conditional verdict, but does not escalate it given the paper's self-acknowledged limitations and the exploratory nature of the physical conclusions.","tokens_in":32938,"tokens_out":3519,"duration_ms":31740,"concrete_test":"For each of the 106 S-PLUS fields, measure the 1σ sky noise in the J0660 image after applying the PELE preprocessing steps (background subtraction, PSF homogenization, Voronoi binning). Then inject artificial Hα+[NII] point sources with known fluxes spanning 1e-17, 2e-17, 5e-17, and 1e-16 erg/s/cm2 at random sky positions, rerun PELE, and compute the recovery fraction as a function of injected flux and field depth. If the 2e-17 threshold achieves <90% recovery in any field shallower than the median depth of the 13 MUSE calibration fields, the uniform detection limit is invalid and the emitter sample requires per-field completeness corrections before quoting physical fractions.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central detection claim is that PELE reliably recovers Hα+[NII] flux down to ~2e-17 erg/s/cm2, based on a pixel-by-pixel comparison with 13 F3D/MUSE galaxies (Section 3.2 and Fig. 4). The paper explicitly notes that S-PLUS depth varies by up to 1 mag across fields (Section 3.2, referencing Figure 6 of Smith Castelli et al. 2024). Yet the detection threshold is applied as a single global value after a post-hoc 3σ clipping that removes 'overestimated' low-flux pixels. The 13 MUSE galaxies are all located within Rvir of Fornax, i.e., in central fields, which are not shown to be representative of the outer fields containing many of the claimed recent infallers. If the true threshold is higher in shallower fields, faint or diffuse Hα+[NII] emission—precisely the kind expected in pre-processing signatures at large clustercentric radii—would be missed. This would bias not only the counts of emitters but also the reported fractions of 91% recent infallers, 44% disturbed late-types, and the spatial distribution of the sample, since the 77 emitters are the basis for all physical conclusions. The paper acknowledges the depth variation but does not propagate it into the detection limit or completeness corrections, so the uniformity assumption is load-bearing.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper presents PELE, a semi-automated pipeline that produces Hα+[N II] emission maps from S-PLUS 12-band images using the Three Filter Method, and applies it to the Fornax cluster and its outskirts. The pipeline is validated against MUSE/F3D Hα maps for 13 galaxies and cross-checked against literature Hα emitter catalogs. The authors report 77 Hα+[N II] emitters with r<18 mag extending to about 4 Rvir, classify their optical morphologies with ASTROMORPHLIB, place them in the projected phase-space (PPS) diagram, and compare with HI/CO data and with an IllustrisTNG-50 Fornax-like system. They conclude that 91% of the emitters are recent infallers, that 44% of late-type emitters show signs of perturbation or merging, and that disturbed low-mass galaxies at large clustercentric distances indicate pre-processing before cluster entry.","tokens_in":33196,"tokens_out":6977,"duration_ms":65563,"significance":"If the results are robust, this is a valuable demonstration of wide-field narrow-band mapping of Hα+[N II] in a nearby cluster, reaching regions beyond the reach of current IFS surveys. The explicit validation against 13 MUSE galaxies, the recovery of 78% of Drinkwater et al. (2001) emitters, and the systematic morphological and phase-space analysis are strengths, and the appendix maps and table are a useful data product. The scientific conclusions about recent infall and pre-processing are plausible but depend on the uniformity of the detection threshold across fields, the completeness of the parent sample, and the adopted PPS parameters; these points need to be quantified before the headline fractions can be taken at face value.","major_comments":[{"comment":"The paper sets a single global detection limit of about 2e-17 erg/s/cm^2 from 13 F3D/MUSE galaxies, all located within the virial radius of Fornax (Section 2.3), but Section 3.1 notes that S-PLUS depth varies by up to 1 mag from field to field. The limit is implemented as a post-hoc 3-sigma clipping of low-flux pixels, so the effective threshold is not the same in shallower outer fields, where many of the claimed recent infallers and pre-processing candidates lie. This can preferentially remove diffuse, low-surface-brightness emission, which the authors themselves say is the least reliably recovered component (Section 4). Because PELE measures Hα+[N II] while MUSE maps are Hα only, part of the \"overestimation\" removed by the clipping may be genuine [N II] flux, so the calibration conflates noise rejection with an astrophysical correction. I ask for per-field detection limits derived from the S-PLUS weight/noise images, a plot of the threshold across the fields, and a completeness test as a function of field depth and clustercentric radius.","section":"3.1 and 3.2"},{"comment":"The headline 91% recent-infaller fraction depends on the adopted Vcluster=1442 km/s and sigma=318 km/s, and the paper itself shows that NGC 1380 moves from the ancient-infaller region E to the recent-infaller region C when the Iodice et al. (2019a) values V=1425 km/s and sigma=300 km/s are used. The statement that all other common galaxies occupy consistent locations is not a quantitative sensitivity test. I request the number of the 77 galaxies that change PPS region under reasonable alternative parameter values and under the parameter uncertainties, and an explicit statement of how many galaxies enter the 91% statistic. In addition, ESO 358-11 and ESO 359-25 have no radial velocities in Table A.1 and therefore cannot be placed in Fig. 8; the sample size actually used for the PPS analysis should be stated.","section":"4.3"},{"comment":"The fractions 44% for disturbed late-type emitters and 15% for early-type emitters are central results, but the C, A, Gini, and M20 measurements from ASTROMORPHLIB are presented without uncertainties or robustness tests. Many of the outer-sample galaxies are low-mass objects near the r=18 mag limit, where non-parametric morphology indices are sensitive to background subtraction, segmentation, and surface-brightness limits. I request a robustness check, for example by varying the segmentation/background parameters, excluding low-S/N galaxies, or comparing against visual classifications of disturbance, to show that the disturbed fractions are not driven by measurement systematics.","section":"4.1"},{"comment":"The parent sample of 233 galaxies with velocities is spectroscopically incomplete, and no detection-completeness function is provided for the 77 emitters as a function of radius, magnitude, or emission morphology. Because ancient or intermediate infallers are expected to have weaker Hα+[N II] emission, the conclusion that 91% of emitters are recent infallers could be partly a selection effect. Comparing the PPS distribution of emitters with that of non-emitters, or adding an upper-limit treatment, would substantially strengthen the claim that Hα+[N II] emission traces recent infall rather than simply reflecting which galaxies are detectable. The same incompleteness affects the spatial asymmetry in the northwest and the statement that the eastern region has no distinct substructure.","section":"2.1 and 4.2"}],"minor_comments":[{"comment":"The abstract says \"within the four virial radius of the Fornax cluster\"; this should be reworded to \"within four times the virial radius\" or \"out to approximately 4 Rvir\" to avoid ambiguity.","section":"Abstract and Section 2.1"},{"comment":"There is a typo in \"marked by an different in the number of objects,\" and \"Fornax AR vir\" should read \"Fornax A Rvir.\"","section":"Section 4.2"},{"comment":"The sentence \"we find an extended emission in 24 (58%) extended, knots in 20 (42%), and 4 (6%) with central concentration\" has percentages that do not sum to 100 and appears to list overlapping categories; please clarify whether the categories are exclusive or whether a galaxy can appear in more than one.","section":"Section 4.6"},{"comment":"The y-axis label \"RSFE M /yr 1\" appears corrupted; it should read \"RSFE (yr^-1)\" or similar, and the definition of RSFE in Eq. (5) should be stated consistently in the figure caption.","section":"Figure 12"},{"comment":"The radial-velocity entries for ESO 358-11 and ESO 359-25 are blank, and the text says only two of the 77 emitters lack spectroscopic confirmation. Please state explicitly how these objects are treated in the spatial and PPS analyses.","section":"Table A.1"}],"recommendation":"major_revision","confidential_remarks":"The paper is honest about its main limitations, and the self-citations to Paper I and Thainá-Batista et al. (2023) are legitimate supporting checks rather than circular inputs. The biggest risk is not the novelty but the unquantified field-to-field variation in the detection threshold and the sensitivity of the PPS classification to adopted cluster parameters; both are addressable with additional analysis and should be required before acceptance. The scope and data products fit A&A well."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this is a legitimate new data product — the first spatially resolved Hα+[NII] maps of a cluster out to ~4 Rvir, built from S-PLUS narrow-band photometry, with MUSE-based validation. The catalog of 77 emitters and the trend that disturbed, low-mass late-types dominate the outskirts are worth having.\n\nWhat's new and good: the PELE pipeline applies the established Three Filter Method to S-PLUS images and produces flux maps with a measured detection limit of ~2e-17. Validation against 13 F3D/MUSE galaxies shows good agreement at high fluxes and a clean threshold; the recovery of 78% of Drinkwater et al.'s Hα emitters is a useful external check. The morphological analysis with ASTROMORPHLIB and the PPS classification are standard but well done. The paper is honest about missing dust and [NII] corrections and about the incomplete spectroscopic parent sample.\n\nThe soft spots: the detection threshold is set by a post-hoc 3σ clip on the MUSE comparison, and applied globally even though the paper notes S-PLUS field depth varies by up to 1 mag. The 13 MUSE fields are all within ~1 Rvir, so they don't sample the shallow outer fields where many of the claimed recent infallers live. If the true limit is higher in those fields, the sample misses faint diffuse emission, and the quoted fractions (91% recent infallers, 44% disturbed LTG) could be biased. That said, the central claim — that there are star-forming, disturbed, low-mass galaxies at 2–4 Rvir — is robust to this; it takes only a few secure detections to make the pre-processing point, and the maps in the appendix look plausible. The bigger issue for reproducibility is that the pipeline isn't public, though the method description is detailed enough to reimplement.\n\nThere's no circularity problem: the threshold comes from MUSE, and the PPS regions from external simulations. The self-citations are supporting checks. The sensitivity of the PPS classification to the assumed velocity dispersion is acknowledged and the paper shows the shift explicitly.\n\nBottom line: this is a solid first result from the S+FP survey. The right referee will want a completeness analysis or at least a propagation of depth variations into the detection limit, and ideally a code release. The scientific case stands. Send it to review.","headline":"A valuable new catalog of Hα+[NII] emitters in Fornax out to 4 Rvir, with an honest but incomplete treatment of depth-dependent detection thresholds.","tokens_in":33966,"tokens_out":3502,"would_cite":true,"duration_ms":29526,"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 paper claims that the PELE pipeline, combining S-PLUS narrow- and broad-band images via the Three Filter Method, reliably maps H$\\alpha$+[N II] emission in 77 Fornax cluster members out to about four virial radii, with 91% identified…","keywords":["galaxy clusters","Fornax cluster","Hα emission","emission-line galaxies","photometric surveys","S-PLUS","pre-processing","projected phase space"],"falsifier":"If a MUSE-detected galaxy with H$\\alpha$ flux above $2\\times10^{-17}$ erg s$^{-1}$ cm$^{-2}$ is located in a shallow S-PLUS field and PELE fails to recover it, while an identical source in a deep field is recovered, the claimed uniform detection limit would be falsified.","tokens_in":32756,"feed_emoji":"🌌","tokens_out":8169,"duration_ms":57586,"temperature":0.7,"pith_summary":"This paper establishes that a new semi-automated pipeline, PELE, can recover spatially resolved H$\\alpha$+[N II] emission from wide-field S-PLUS images down to a flux of about $2\\times10^{-17}$ erg s$^{-1}$ cm$^{-2}$, calibrated against MUSE data. Applied to the Fornax cluster, it produces emission maps for 77 member galaxies with $r<18$ mag out to $\\sim4 R_{\\rm vir}$, far beyond the reach of integral-field surveys. Among these emitters, 75% are late-type and 25% are early-type; 91% occupy the recent-infall regions of the projected phase-space diagram, and 44% of the late-type emitters show morphological perturbation or merger features. Disturbed, low-mass emitters at projected radii beyond $2.5R_{\\rm vir}$ are interpreted as evidence that galaxies begin transforming before entering the cluster. The paper argues that wide-field photometric surveys can therefore trace cluster assembly history and pre-processing without requiring spectroscopy for every galaxy.","feed_headline":"Pipeline maps Hα emission in 77 Fornax galaxies out to 4 Rvir","feed_subtitle":"Recovering faint Hα+[NII] flux from S-PLUS images reveals recent infall and pre-processing in the cluster outskirts.","key_machinery":"The load-bearing tool is the PELE (Pixel-to-Pixel Emission Line Estimate) pipeline, a semi-automated code that builds 12-band S-PLUS datacubes, masks foreground stars and background sources, homogenizes the point-spread function across filters, applies a Butterworth spatial filter and Voronoi binning to reach a target signal-to-noise of 10 in the J0660 narrow-band, and then estimates the H$\\alpha$+[N II] flux per pixel using the Three Filter Method. The Three Filter Method assumes the emission line behaves as a Dirac delta at the line wavelength while the continuum is linear across the J0660 narrow-band and the $r$ and $i$ broad bands; the equation from Vilella-Rojo et al. (2015) converts the three measured fluxes into a line flux. A final 3$\\sigma$ clipping removes low-flux pixels that the method systematically overestimates, which sets the detection floor near $2\\times10^{-17}$ erg s$^{-1}$ cm$^{-2}$.","core_discovery":"The central discovery is that the PELE pipeline, using S-PLUS images and the Three Filter Method, yields reliable pixel-by-pixel H$\\alpha$+[N II] maps for 77 Fornax cluster members out to $\\sim4 R_{\\rm vir}$, with a detection limit near $2\\times10^{-17}$ erg s$^{-1}$ cm$^{-2}$ validated against MUSE observations of 13 galaxies. The recovered sample is dominated by late-type galaxies (75%), and 91% of all emitters lie in phase-space regions associated with infall within the last gigayear. A substantial fraction (38% overall, 44% of late-types) show morphological disturbance or merger signatures, and these disturbed systems preferentially appear at large clustercentric distances. The spatial distribution shows a concentration of emitters in the northwest, possibly tracing a filament connecting Fornax to Eridanus, while the Fornax A group contributes a distinct southwest component. These results are interpreted as direct observational evidence that galaxy preprocessing starts well before the galaxies cross the cluster's virial radius.","pith_inferences":["If the uniform detection threshold holds across S-PLUS fields, the same methodology could be extended to the J0378 narrow-band filter to map [O II] emission in Fornax, doubling the diagnostics available from the same datacubes.","The proposed Fornax-Eridanus filament could be tested by running PELE on S-PLUS fields between Fornax and the Eridanus cluster; a continuous chain of emitters would strengthen the filament scenario.","A direct test of the depth-uniformity assumption would be to cross-calibrate PELE on overlapping S-PLUS fields of differing depth; if the detection threshold varies, the reported 91% recent-infaller fraction may need to be re-derived as a function of field depth.","The comparison to IllustrisTNG-50 suggests a testable prediction: the resolved star-formation efficiency of infalling late-type galaxies should be skewed toward the cluster-facing side, which could be measured with IFU observations of a few nearby emitters."],"forward_implications":["The same pipeline can be reused to build H$\\alpha$+[N II] maps for other nearby clusters covered by S-PLUS, effectively turning a wide-field photometric survey into a star-formation census of cluster outskirts.","The 77 emitters provide a homogeneous target list for future H I and CO follow-up at projected radii beyond $1R_{\\rm vir}$, where radio data are currently scarce, to test the pre-processing interpretation directly.","The phase-space classification places 91% of emitters as recent infallers, implying that most of the current H$\\alpha$ emission in Fornax is tied to galaxies that entered the cluster within roughly the last gigayear.","The observed fractions of disturbed, low-mass emitters at large radii set quantitative expectations that cosmological simulations of Fornax-like systems can be compared against."],"supporting_citations":[{"why":"Supplies the Three Filter Method equation that PELE uses to convert S-PLUS narrow- and broad-band fluxes into an H$\\alpha$+[N II] line flux.","marker":"Vilella-Rojo et al. (2015)"},{"why":"Provides the complete derivation and assumptions of the Three Filter Method that PELE relies on.","marker":"Pascual et al. (2007)"},{"why":"Defines the Fornax3D MUSE survey whose H$\\alpha$ maps are used to validate PELE.","marker":"Sarzi et al. (2018)"},{"why":"Delivers the MUSE H$\\alpha$ maps for 13 galaxies that PELE is compared against to set the detection threshold and assess flux accuracy.","marker":"Iodice et al. (2019a)"},{"why":"Provides the spectroscopic membership list, mean cluster velocity, and velocity dispersion used to build the projected phase-space diagram.","marker":"Maddox et al. (2019)"},{"why":"Defines the virial radii of Fornax and Fornax A and the member catalog that the sample is drawn from.","marker":"Drinkwater et al. (2001)"},{"why":"Establishes the phase-space divisions that separate recent, intermediate, and ancient infallers, which the paper applies to its emitters.","marker":"Rhee et al. (2017)"},{"why":"Provides the MeerKAT H I pre-processing classifications for Fornax A galaxies that the paper compares with its H$\\alpha$+[N II] morphologies.","marker":"Kleiner et al. (2021)"}],"fun_headline_variants":["77 Fornax galaxies mapped in Hα out to 4x virial radius","Faint Hα maps reveal 77 Fornax galaxies' recent infall","New pipeline maps Hα in 77 Fornax members, reaching 4×Rvir","S-PLUS detects 77 Hα emitters in Fornax, hinting at pre-processing","Hα mapping of 77 Fornax galaxies shows pre-processing before cluster entry"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The detection threshold of about $2\\times10^{-17}$ erg s$^{-1}$ cm$^{-2}$, calibrated on 13 MUSE galaxies and enforced by a post-hoc 3$\\sigma$ clipping that removes overestimated low-flux pixels, is assumed to be uniform across all S-PLUS fields even though the survey's depth varies by up to one magnitude from field to field.","fun_headline_variants_meta":{"raw":{"variants":["77 Fornax galaxies mapped in Hα out to 4x virial radius","Faint Hα maps reveal 77 Fornax galaxies' recent infall","New pipeline maps Hα in 77 Fornax members, reaching 4×Rvir","S-PLUS detects 77 Hα emitters in Fornax, hinting at pre-processing","Hα mapping of 77 Fornax galaxies shows pre-processing before cluster entry"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000719,"raw_usage":{"total_tokens":3349,"prompt_tokens":1188,"completion_tokens":2161,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":804,"completion_tokens_details":{"reasoning_tokens":2046}},"tokens_in":804,"tokens_out":2161,"duration_ms":13780,"temperature":1.0,"reasoning_tokens":2046,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-07T14:55:36.260999+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"If a MUSE-detected galaxy with H$\\alpha$ flux above $2\\times10^{-17}$ erg s$^{-1}$ cm$^{-2}$ is located in a shallow S-PLUS field and PELE fails to recover it, while an identical source in a deep field is recovered, the claimed uniform detection limit would be falsified.","supporting_citations":[{"cited_title":"2015, , 580, A47","cited_arxiv_id":null,"evidence_quote":"Supplies the Three Filter Method equation that PELE uses to convert S-PLUS narrow- and broad-band fluxes into an H$\\alpha$+[N II] line flux."},{"cited_title":"2007, , 119, 30","cited_arxiv_id":null,"evidence_quote":"Provides the complete derivation and assumptions of the Three Filter Method that PELE relies on."},{"cited_title":"2018, , 616, A121","cited_arxiv_id":null,"evidence_quote":"Defines the Fornax3D MUSE survey whose H$\\alpha$ maps are used to validate PELE."},{"cited_title":"2019, , 490, 1666","cited_arxiv_id":null,"evidence_quote":"Provides the spectroscopic membership list, mean cluster velocity, and velocity dispersion used to build the projected phase-space diagram."}],"review_version":1}