{"id":"0217d839-9748-4906-9b13-8b2ba079af9d","arxiv_id":"2504.17024","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"A revised GeV Lgamma-SFR correlation and SFR-scaled spectra suggest about a dozen star-forming galaxies, including M83, M33, NGC 1068, NGC 4945, and Circinus, may be detectable by CTAO, LHAASO, or SWGO.","lead":"This paper predicts which star-forming galaxies could be detected by next-generation TeV gamma-ray telescopes, and finds about a dozen candidates including several not yet seen by Fermi-LAT. It updates the gamma-ray luminosity versus star-formation-rate correlation and compares predicted TeV fluxes to realistic CTAO, LHAASO, and SWGO sensitivities.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 'almost a dozen' count may hinge on a biased Lγ–SFR normalization: the fit drops five outliers and ignores all upper limits, and a ~2× normalization shift would push M83/M33 below CTAO sensitivity.","rationale":"The reader's CONDITIONAL verdict is appropriate, and the paper is unusually transparent about its limitations. The Fermi-LAT extrapolations for the five top candidates are anchored by actual GeV spectra, and for NGC 253, M82, and NGC 1068 by existing VHE data, so those detections are robust to the concern raised here. The soft spot is the empirical model used for non-4FGL candidates and for the borderline 'o' class: Equation (1) inherits a normalization from a 9-point ODR fit in Fig. 1 that excludes five detected outliers and ignores all Table 1 upper limits. The authors cite Ambrosone et al. as estimating a factor-of-two normalization shift from including non-detections; since M83 and M33 sit at the CTAO threshold, such a shift removes the two non-4FGL members of the 'almost a dozen' count, and the Arp 220/Arp 299 possibility classifications also weaken. Internal absorption, the reader's chosen weakest assumption, is explicitly confined by the authors to energies beyond roughly 10 TeV and mostly affects LHAASO/SWGO rather than the CTAO-band claim, so I rate it valid but secondary. A censored Bayesian regression including all galaxies is the decisive check; until it is run, CONDITIONAL remains the right verdict, so no change to the reader's recommendation is needed.","tokens_in":25748,"tokens_out":10331,"duration_ms":109250,"concrete_test":"Using the Fermi-LAT TS profiles for the 13 non-4FGL galaxies in Table 1, perform a Bayesian regression of Lγ,2GeV on SFR with censored likelihoods for the upper limits, and include all 14 4FGL associations in the same fit with a mixture component for AGN-contaminated sources. Re-derive Table B.1's CTAO classifications from the posterior predictive distribution. If the posterior normalization log C shifts downward by more than about 0.3 dex (the factor of two cited from Ambrosone et al.) or the slope changes substantially, the M83/M33 and Arp 220/Arp 299 entries must be reclassified, and the 'almost a dozen' headline count should be revised.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The empirical model behind Equation (1) inherits its 2 GeV normalization from the ODR fit in Section 2.3 (Fig. 1), which is performed on only 9 of the 14 4FGL-detected galaxies after excluding NGC 3424, NGC 4945, Circinus, NGC 2403, and NGC 7059, and which ignores all 13 upper limits from Table 1. The paper itself notes in the Section 2.3 footnote that Ambrosone et al. estimate the normalization could be impacted by a factor of two if non-detected galaxies are properly accounted for. That shift is not a rounding error: M83 and M33, the two non-4FGL candidates whose empirical-model spectra sit at the CTAO 50-hour threshold (Section 3.2, Fig. 3), would fall below the sensitivity curve under a factor-of-two downward normalization, and the Arp 220/Arp 299 'possibility' classifications in Table B.1 would also weaken. Excluding the five outliers is defensible given the disputed associations of NGC 2403 and NGC 7059, but it is a selection on the dependent variable; a mixture or censored regression is required to know whether the resulting normalization is unbiased. Internal absorption, the reader's chosen caveat, is a real but secondary issue because the authors explicitly defer it to energies beyond about 10 TeV, whereas the CTAO-band fluxes for M83 and M33 are controlled by the sub-TeV normalization of the SFR scaling relation.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper compiles a sample of 27 star-forming galaxies (14 detected by Fermi-LAT in 4FGL-DR4 and 13 not detected) and assesses their detectability at TeV energies with CTAO, LHAASO, and SWGO. The GeV analysis uses standard Fermipy procedures on roughly 15 years of Fermi-LAT data, yielding upper limits for the non-detected galaxies. The authors update the L_gamma-SFR correlation at 2 GeV using the 4FGL-detected galaxies, then predict TeV fluxes with two models: an empirical SFR-scaled power law of photon index 2.2 (Eq. 1) and a straight extrapolation of the 4FGL power laws, both attenuated by the EBL. These predictions are compared with public CTAO prod5 IRFs, LHAASO, and SWGO sensitivities. The main result is that almost a dozen star-forming galaxies may be detectable by upcoming gamma-ray telescopes, with NGC 253, M82, NGC 1068, NGC 4945, and Circinus above the CTAO 50-hour sensitivity, M83 and M33 at the threshold, and Arp 220, Arp 299, NGC 2403, and NGC 3424 as possible targets with longer exposures.","tokens_in":1644,"tokens_out":2004,"duration_ms":72054,"significance":"If the predictions hold, the paper provides a useful and timely target list for CTAO, LHAASO, and SWGO, extending earlier work by including non-4FGL galaxies and using updated instrument response functions. Strengths include the use of public IRFs, a standard and reproducible Fermi-LAT analysis pipeline, the validation that spectra above the differential sensitivity curves correspond to combined significance above 5 sigma, and the transparent case-by-case discussion of candidates. The main value is in identifying which galaxies deserve dedicated VHE observations and in framing the L_gamma-SFR correlation as a tool for population predictions. However, the central quantitative claim depends on a correlation normalization that is not robust to the treatment of upper limits and excluded outliers, and the 'almost a dozen' count is not tied to a crisp detectability criterion.","major_comments":[{"comment":"The empirical model that places M83 and M33 at the CTAO detection threshold is calibrated with an ODR fit to only 9 of the 14 4FGL-detected galaxies, after excluding NGC 3424, NGC 4945, Circinus, NGC 2403, and NGC 7059, and the fit ignores all 13 upper limits from Table 1. The footnote in Section 2.3 acknowledges that Ambrosone et al. estimate the normalization could shift by a factor of two if non-detected galaxies are properly accounted for, but this systematic uncertainty is not propagated into the detectability claims. A factor-of-two downward shift in the normalization is roughly four times the quoted statistical uncertainty on log C (38.19 +/- 0.08) and would push M83 and M33 below the CTAO sensitivity curve and weaken the 'o' classifications for Arp 220 and Arp 299. Please either perform a censored or mixture regression that includes the upper limits, or explicitly present the candidate list as conditional on the optimistic normalization and show how the list changes under a factor-of-two systematic shift.","section":"Section 2.3, Eq. (1), Fig. 3, Table B.1"},{"comment":"The headline claim that 'almost a dozen' star-forming galaxies may be detectable is not backed by a well-defined criterion. The paper states that a spectrum above the differential sensitivity curve corresponds to a combined significance above 5 sigma, yet the 'almost a dozen' count appears to include galaxies classified as 'o' (possibility) and '?' (detailed study needed), whose spectra lie below the 50-hour sensitivity curve. Please provide a quantitative definition of 'detectable' (e.g., predicted combined significance, or flux relative to the sensitivity curve integrated over the CTAO band) and state explicitly how many galaxies satisfy that criterion, and how that number changes under the normalization uncertainty discussed above.","section":"Abstract, Section 3.2, Table B.1"},{"comment":"The paper correctly cautions that internal gamma-ray absorption is not modeled and that results are optimistic beyond 10 TeV, but for the borderline candidates M83 and M33 the CTAO-band fluxes are controlled by sub-TeV emission, so internal absorption is not the dominant uncertainty for those objects. The dominant uncertainty is the normalization of the SFR scaling relation. The manuscript would be strengthened by separating these two caveats and by showing the predicted CTAO-band integrated flux (or significance) with a band that includes both the fit parameter uncertainties and the factor-of-two systematic, rather than only the shaded region based on the best-fit line.","section":"Section 3.1, Eq. (1)"}],"minor_comments":[{"comment":"Several figures contain typographical spacing issues in the energy axis label, e.g., 'T eV' instead of 'TeV'; please correct these in the final version.","section":"Throughout"},{"comment":"The sentence 'The figure show that NGC 3424, NGC 4945, Circinus, NGC 2403 and NGC 7059 stand out...' has a subject-verb agreement error; please revise.","section":"Section 2.3"},{"comment":"The notation L_2GeV(SFR) in Eq. (1) would be clearer if it matched the L_gamma,2GeV notation used in Section 2.3, and if the units of each quantity in Eq. (1) were stated explicitly.","section":"Eq. (1) and Table 2"},{"comment":"The last column header 'LHAASO, SWGO' is ambiguous because each row has a single symbol that may apply to either instrument or both; please clarify the correspondence, for example by splitting the column or defining the convention in the caption.","section":"Table B.1"},{"comment":"The statement that a minimum of 5 signal counts per bin and 3 sigma per-bin significance are imposed would benefit from a reference or justification, since these thresholds affect the derived sensitivity curves.","section":"Section 3.1"}],"recommendation":"major_revision","confidential_remarks":"The paper is well-structured and the Fermi-LAT and sensitivity calculations are reproducible in principle, but the central 'almost a dozen' claim is more fragile than the abstract suggests. The factor-of-two normalization uncertainty acknowledged in the footnote should be addressed quantitatively before publication; otherwise the M83/M33 threshold claim could mislead observers planning CTAO time. A censored regression or an explicit systematic band around the empirical model predictions would resolve this."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Best read as a target-selection study, not as a set of firm predictions. It is a genuinely useful update: the 2 GeV L_gamma–SFR correlation from 4FGL-DR4, fresh Fermi-LAT upper limits for 13 non-detected galaxies, and a careful comparison with public CTAO prod5, LHAASO, and SWGO response functions. The case-by-case discussion is clear, and the authors deserve credit for flagging associations that are shaky (NGC2403, NGC7059) and for noting where their model is optimistic. The census demotes IC342 and NGC6946 relative to Shimono 2021 and promotes NGC1068, NGC4945, Circinus, M83, and M33 as CTAO targets. That is a real service.\n\nThe soft spot is the normalization of the empirical model. Equation (1) plugs SFR into a relation fitted to the same galaxies whose detectability is being predicted. The fit excludes five outliers and ignores all 13 upper limits. The authors mention Ambrosone's estimate that accounting for non-detections could move the normalization by a factor of two. A factor of two matters here: M83 and M33 sit right at the CTAO threshold, so a downward shift removes them. The 'almost a dozen' in the abstract is a count that includes objects at threshold and mere 'possibility' entries. I would not call that misleading—they qualify it in the text—but the headline is optimistic.\n\nThe Fermi-LAT extrapolations are somewhat more robust for the five 4FGL-detected candidates, but they inherit the usual power-law-extrapolation uncertainty into TeV. Internal absorption is a real effect but mostly above ~10 TeV; for M83/M33, the sub-TeV normalization is the issue, so the stress-test is right to put the normalization bias first.\n\nOverall: the pipeline is standard, the data products are public, and the paper states its limitations openly. I'd send it to a competent referee. The revision should add a sensitivity test on the normalization (e.g., predictions with log C shifted by ±0.3) and tone down 'almost a dozen' to something like 'up to about a dozen, several of which are at the threshold.' The paper is worth citing as the current best target list for next-generation TeV observatories.","headline":"Useful target-selection census for next-gen TeV observatories, but the 'almost a dozen' count rests on a calibration that could be off by a factor of two.","tokens_in":26673,"tokens_out":3137,"would_cite":true,"duration_ms":31297,"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 almost a dozen star-forming galaxies, led by NGC 253, M82, NGC 1068, NGC 4945, and Circinus, may be detectable in the TeV band by the upcoming Cherenkov Telescope Array Observatory.","keywords":["gamma-ray astronomy","star-forming galaxies","starburst galaxies","TeV gamma rays","Cherenkov Telescope Array Observatory","Fermi-LAT","star formation rate","cosmic-ray transport"],"falsifier":"A CTAO campaign reaching the 50-hour sensitivity at the predicted flux on NGC 1068, NGC 4945, and Circinus that finds no source would contradict the paper's central detection claim; likewise, measuring a spectral break in NGC 253 or M82 above roughly 10 TeV would show the unbroken power-law assumption fails and would lower the expected detectability of the ULIRGs.","tokens_in":25519,"feed_emoji":"🔭","tokens_out":6070,"duration_ms":50521,"temperature":0.7,"pith_summary":"The paper asks which star-forming galaxies the next generation of ground-based gamma-ray telescopes, above all the Cherenkov Telescope Array Observatory, could finally detect in the TeV band. Using the measured relation between gamma-ray luminosity at 2 GeV and star formation rate, together with an assumed photon index of 2.2 extending from GeV to TeV energies and attenuation by the extragalactic background light, the authors predict that almost a dozen galaxies may rise above the 50-hour point-source sensitivity of CTAO. The strongest candidates are the starbursts NGC 253 and M82 plus NGC 1068, NGC 4945, and Circinus, with M83 and M33 sitting at the threshold. Such detections matter because TeV emission would probe cosmic-ray acceleration and transport in extragalactic star-forming environments, a regime currently sampled by only two galaxies.","feed_headline":"Nearly a dozen galaxies could become TeV gamma-ray sources","feed_subtitle":"CTAO forecasts put NGC 253, M82, NGC 1068, NGC 4945, and Circinus above 50-hour sensitivity.","key_machinery":"The load-bearing relation is the observationally calibrated correlation between gamma-ray luminosity at 2 GeV and star formation rate, $L_{\\gamma,2\\,\\mathrm{GeV}} = C\\,\\dot{M}_*^m$, converted into a flux prediction by $F(E) = L_{2\\,\\mathrm{GeV}}(\\mathrm{SFR})/(4\\pi d^2) \\times (E/2\\,\\mathrm{GeV})^{2-\\alpha} e^{-\\tau(E,z)}$ with $\\alpha = 2.2$ and EBL optical depth $\\tau$. A second channel uses the straight power-law extrapolation of the 4FGL-DR4 spectra. Both are compared with the differential point-source sensitivity of CTAO, LHAASO, and SWGO, computed with the published instrument response functions.","core_discovery":"The paper claims that the GeV band already contains the raw material for TeV discoveries: extrapolating the Fermi-LAT spectra of the 14 GeV-detected galaxies, or scaling a 2 GeV luminosity to star formation rate, places NGC 253, M82, NGC 1068, NGC 4945, and Circinus above the CTAO 50-hour sensitivity curve, and M83 and M33 at its threshold. A revised correlation $L_{\\gamma,2\\,\\mathrm{GeV}} = C\\,\\dot{M}_*^m$ with $m = 1.27 \\pm 0.10$ (or $m = 1.31 \\pm 0.08$ after removing five outliers) provides the normalization for the empirical model. The paper therefore concludes that almost a dozen star-forming galaxies may be detectable by upcoming gamma-ray telescopes, a significant increase over the two known TeV starbursts.","pith_inferences":["A natural extension is to stack the predicted sub-threshold TeV fluxes of the many non-detected galaxies; a stacked signal could test the SFR-gamma-ray correlation in a regime where individual detections remain out of reach.","If internal absorption is mild, the same scaling law suggests that deeper exposures (e.g., 100 hours) would bring Arp 220, Arp 299, NGC 2403, and NGC 3424 into detectable range, making them interesting targets for a follow-up program.","The predicted TeV luminosities can be combined with neutrino expectations for star-forming galaxies; a CTAO detection of NGC 1068 would strengthen the case that its reported neutrino signal is hadronic and connected to star formation."],"forward_implications":["If the predictions hold, CTAO will go from two TeV star-forming galaxies (NGC 253 and M82) to roughly a dozen, enabling the first population-level study of TeV emission from star-forming galaxies.","TeV spectra of these galaxies would constrain cosmic-ray acceleration up to multi-TeV energies in extragalactic environments and test whether hadronic emission dominates outside the GeV band.","Observations beyond 10 TeV by LHAASO or SWGO would probe internal gamma-ray absorption inside galaxies, turning the predicted cutoff into a diagnostic of the radiation field.","The outlier galaxies NGC 3424, NGC 4945, Circinus, NGC 2403, and NGC 7059 are candidates for hidden AGN or misassociation, so TeV follow-up would help sort out the origin of their gamma-ray excesses."],"supporting_citations":[{"why":"Established the SFR-gamma-ray luminosity correlation that anchors the empirical model.","marker":"Ackermann et al. 2012"},{"why":"Provides the TeV detection of NGC 253 that fixes the comparison flux and demonstrates the GeV-TeV extension.","marker":"H.E.S.S. Collaboration et al. 2018"},{"why":"Provides the TeV detection of M82 used as the other known starburst benchmark.","marker":"VERITAS Collaboration et al. 2009"},{"why":"Supplies the 4FGL-DR4 catalog whose power-law spectra are extrapolated for the GeV-detected galaxies.","marker":"Ballet et al. 2023"},{"why":"Provides the extragalactic background light model used to attenuate the predicted TeV spectra.","marker":"Domínguez et al. 2011"},{"why":"Supplies the Revised MANGROVE sample from which the non-GeV-detected candidates are drawn.","marker":"Biteau 2021"},{"why":"Provides the multi-wavelength SFR estimation method and earlier correlation study on which the updated values rely.","marker":"Kornecki et al. 2020"},{"why":"Quantifies internal gamma-ray absorption inside starbursts, the main effect that could spoil the optimistic extrapolations.","marker":"Peretti et al. 2020"}],"fun_headline_variants":["TeV starbursts: nearly a dozen galaxies may be detectable","CTAO could spot a dozen star-forming galaxies in gamma rays","From GeV to TeV: new starburst candidates for CTAO","Almost a dozen starburst galaxies poised for TeV discovery"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The predictions assume the GeV power law continues unbroken to TeV energies with photon index 2.2, attenuated only by the extragalactic background light, while gamma-ray absorption inside the host galaxy, potentially strong in starbursts and ultraluminous infrared galaxies, is not modeled.","fun_headline_variants_meta":{"raw":{"variants":["TeV starbursts: nearly a dozen galaxies may be detectable","CTAO could spot a dozen star-forming galaxies in gamma rays","From GeV to TeV: new starburst candidates for CTAO","Almost a dozen starburst galaxies poised for TeV discovery"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000288,"raw_usage":{"total_tokens":1746,"prompt_tokens":1056,"completion_tokens":690,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":672,"completion_tokens_details":{"reasoning_tokens":614}},"tokens_in":672,"tokens_out":690,"duration_ms":6251,"temperature":1.0,"reasoning_tokens":614,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-16T10:50:08.393070+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A CTAO campaign reaching the 50-hour sensitivity at the predicted flux on NGC 1068, NGC 4945, and Circinus that finds no source would contradict the paper's central detection claim; likewise, measuring a spectral break in NGC 253 or M82 above roughly 10 TeV would show the unbroken power-law assumption fails and would lower the expected detectability of the ULIRGs.","supporting_citations":[{"cited_title":"A., Aliu , E., et al","cited_arxiv_id":null,"evidence_quote":"Provides the TeV detection of M82 used as the other known starburst benchmark."},{"cited_title":"J., del Palacio , S., et al","cited_arxiv_id":null,"evidence_quote":"Provides the multi-wavelength SFR estimation method and earlier correlation study on which the updated values rely."},{"cited_title":"2020, , 493, 5880","cited_arxiv_id":null,"evidence_quote":"Quantifies internal gamma-ray absorption inside starbursts, the main effect that could spoil the optimistic extrapolations."}],"review_version":1}