{"id":"94f62fd0-bd3c-4ba0-92b3-bcaec7d2a797","arxiv_id":"2608.12012","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":8,"one_line_summary":"A large, faint gamma-ray glow in the southern Galactic sky is identified as a plausible counterpart to the southern eROSITA bubble, with lower luminosity and a softer spectrum than the north.","lead":"Using 17 years of Fermi telescope data, researchers found a large, faint glow of gamma rays in the southern sky that lines up with the southern eROSITA bubble, a giant X-ray structure around our galaxy's center. The glow is dimmer and softer than its northern counterpart, hinting at how material and energy flow out of the Milky Way's core.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The claimed preference for the filled eB template over Loop I may be a template-shape artifact: the comparison tests a shell-like Loop I geometry but not a volume-filled Loop I model, so the morphological evidence for an eB counterpart is not yet secure.","rationale":"The detection of a southern large-scale excess that persists across 64 GALPROP diffuse-emission models is a meaningful and fairly robust observational result; I do not dispute that. The reader's weakest-assumption focus on the DGE ensemble is reasonable, but it is less decisive than the template-morphology issue: all 64 models live within one GALPROP framework and share broadly similar large-scale inverse-Compton morphologies, so they may not bracket an arbitrarily shaped foreground. The more load-bearing step for the paper's headline interpretation is the claim that a filled eB-like template is preferred over the Loop I shell geometry. That comparison is asymmetric: a filled template is compared with a limb-brightened shell template, even though gamma-ray emission from a Loop-I-like structure could plausibly be volume-filled. If a filled Loop I template fits as well as the eB template, then the 'plausible gamma-ray counterpart of the southern eB' conclusion loses its main morphological support, leaving only a spatial coincidence. The paper already acknowledges that a localized foreground cannot be ruled out, so the appropriate verdict remains conditional; the new point is that a specific, cheaply testable template degeneracy should be resolved before the eB-morphology preference is stated as robust.","tokens_in":12732,"tokens_out":4969,"duration_ms":50806,"concrete_test":"Construct a volume-filled Loop I template by assigning uniform emissivity inside the Wolleben (2007) Loop I shells on the same HEALPix grid, convolve with the LAT PSF, and repeat the Section 3 template likelihood comparison against the eB template using the benchmark model and the 64-model ensemble. If the AIC difference between filled Loop I and eB becomes small or favors filled Loop I, the claimed eB-like morphology is not established; if filled eB still wins by a large margin (e.g., delta_AIC > 10), the morphological claim survives. A complementary check is to include both eB and filled Loop I in a single fit and test whether the eB template adds significant TS after the filled Loop I component is included.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central astrophysical claim is not merely that a southern large-scale gamma-ray excess exists, but that it is better described by a filled eROSITA-bubble-like template than by the Loop I shell geometry, and is therefore a plausible eB counterpart. The comparison in Section 3 uses the Wolleben (2007) radio synchrotron shell geometry as the only Loop I hypothesis. However, gamma-ray emission from a local Loop I / NPS-type structure need not follow the limb-brightened radio shell: inverse-Compton emission from electrons distributed through the bubble interior would produce a filled or partially filled morphology. If the true foreground gamma-ray morphology is filled, the eB template will win the AIC comparison even if the excess has nothing to do with the eROSITA bubbles. The paper's caveat that Loop-I-related foreground cannot be excluded does not remove the problem, because the key discriminator (filled eB vs. filled Loop I) was never tested. The TS improvement over baseline is large, but that only establishes that some additional large-scale component is needed; it does not identify its geometry. Thus the load-bearing step is the morphological discrimination, not the DGE ensemble, and it is underdetermined by the current template set.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper analyzes 17 years of Fermi-LAT data with an energy-resolved template likelihood method, testing 64 GALPROP diffuse Galactic emission (DGE) models as backgrounds. The authors report a large-scale diffuse gamma-ray excess in the Southern Galactic Hemisphere that persists across the DGE ensemble, spans tens of degrees, and is spatially consistent with the southern eROSITA bubble (eB). They compare a filled eB-like spatial template with the Wolleben Loop I shell geometry and find the eB template preferred (AIC difference -78.47 in the benchmark model). Under the eB assumption, the southern component is softer and fainter than the northern counterpart, with an integrated luminosity ratio of about 7.2. The paper also presents leptonic and hadronic spectral fits and discusses energetic implications, concluding that the structure is a plausible gamma-ray counterpart of the southern eB while acknowledging that Loop I or other foreground emission cannot be excluded.","tokens_in":12998,"tokens_out":5477,"duration_ms":53534,"significance":"If the morphological association with the southern eROSITA bubble holds, this would be the first quantified gamma-ray counterpart of the southern eB, providing a new observable for Galactic outflow models and cosmic-ray transport in the halo. The paper is methodical: it uses a long dataset, tests an ensemble of DGE models, applies masks, and is explicit about many limitations. The recovery of the well-known northern NPS/Loop I feature is a useful validation benchmark. The spectral and energetic analyses are careful in distinguishing leptonic and hadronic scenarios. However, the central morphological interpretation is underdetermined by the template set actually tested, and the strength of some quantitative claims (persistence across DGE models, luminosity ratio) is not fully supported by the reported statistics.","major_comments":[{"comment":"The morphological preference for the eB template over the Wolleben Loop I shell geometry does not test a volume-filled or diffuse Loop I foreground. Gamma-ray emission from a local Loop I / NPS-type structure need not follow the limb-brightened radio shell; inverse-Compton emission from electrons distributed through the bubble interior would naturally produce a filled or partially filled morphology. The AIC difference of -78.47 therefore supports only the statement that the excess is not well described by the specific Wolleben shell geometry, not that it is eB-like rather than foreground-like. The authors should add a filled (or broadened) Loop I template to the comparison before claiming that the morphology favors an eB association. This is load-bearing for the central interpretation, and the manuscript's own caveat in the abstract (\"Loop-I-related or other localized foreground emission cannot be excluded\") does not substitute for the missing test.","section":"Section 3, Figure 3"},{"comment":"The claim that the southern excess is \"detected with high statistical significance across all 64 models\" is not supported by the reported ensemble statistics. Only the mean template TS values are given (⟨TS_Loop I⟩=763.0 and ⟨TS_eB⟩=932.2); no minimum, maximum, quartiles, or per-model distribution is shown. If some DGE models yield much lower TS values for the southern template, the persistence claim would be materially weakened. Please provide the full distribution of the 64 TS values (e.g., a histogram or table) or at least the minimum and quartiles, and similarly characterize the spread of the AIC difference between the two templates.","section":"Section 3 and Appendix A"},{"comment":"The north-south luminosity ratio of ~7.2 is quoted without an uncertainty, even though it is derived from template fits and inherits systematic errors from the DGE model choice, the spatial template, and the spectral model. The comparison to the eROSITA intensity ratio (quoted as a 3-6 range) would be much more meaningful if the gamma-ray ratio were accompanied by a systematic error estimate. Since this ratio appears in the abstract and is used to support the eB-asymmetry interpretation, an error budget should be provided.","section":"Section 3, Appendix B"}],"minor_comments":[{"comment":"The sentence \"The northern and southern hemispheres are fitted separately and then merged for visualization\" is ambiguous: it is unclear whether this applies only to the residual map construction or also to the baseline likelihood fits. Please clarify whether the baseline model parameters are fitted independently in each hemisphere, as this affects the interpretation of the residual map.","section":"Section 2.2"},{"comment":"The description of the 4FGL source component as \"included as a template and fitted simultaneously\" does not state whether its normalization is fitted independently in each energy bin or as a single energy-independent scale. Please specify the treatment of the source template normalizations.","section":"Section 2.2"},{"comment":"The template TS values (TS=881.46 for Loop I and TS=959.93 for eB) are quoted for the benchmark model, but the number of additional degrees of freedom introduced by the large-scale template (one normalization per energy bin, i.e., 25 parameters) is not stated. This information is useful for interpreting the magnitude of the TS.","section":"Section 3"},{"comment":"The selection of the benchmark DGE template as \"the template set that most closely aligns with the ensemble mean\" is vague; please specify the metric used (e.g., minimum chi-square, maximum likelihood, or sum of squared deviations) so the choice is reproducible.","section":"Appendix A"},{"comment":"The log-parabola parameters α and β are given only in the figure legend. They should be reported in a table with their 1σ uncertainties for both the northern and southern components, since the spectral softening is a key qualitative result.","section":"Figure 2 caption and text"},{"comment":"The hadronic eB local scenario is labeled \"PL/ECPL\" without specifying which model is used; the text says the ECPL is used for the eB model. Please make the table consistent with the text.","section":"Table 1"},{"comment":"Several references list arXiv URLs directly in the reference list (e.g., Lallement 2022, Scheel-Platz et al. 2023, Yeung et al. 2026, Zhang et al. 2026). These should be formatted consistently with the journal style, including journal names and DOIs where available.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"The paper presents a careful analysis and a potentially important detection, but the central morphological claim is currently supported only against a specific shell geometry. The missing filled-Loop I test is a correctable gap rather than a fundamental flaw. The persistence claim across 64 DGE models would be much stronger if the ensemble distribution of TS values were shown. The authors should also provide uncertainties for the luminosity ratio. I recommend major revision rather than rejection because the excess itself appears robust and the interpretation, while not yet secure, is a reasonable hypothesis that can be tested with the requested additional template fits."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper reports a large-scale diffuse gamma-ray structure in the southern Galactic hemisphere using 17 years of Fermi-LAT data. The main result: an extended residual excess that persists across 64 different GALPROP diffuse-emission templates, and a template comparison that favors a filled eROSITA-bubble-like morphology over the Wolleben Loop I shell geometry. That first part is solid. The excess is clearly there, and the paper does a careful job with masking, energy-dependent fitting, and a broad systematic ensemble.\n\nWhat's genuinely new is the dedicated characterization: the energy spectrum of the southern structure, the north-south luminosity ratio (about 7:1), and the explicit eB-vs-Loop I template comparison. Previous work, like Scheel-Platz et al., showed extended southern gamma-ray emission but didn't quantify the morphology and spectrum this way. The spectral asymmetry and the softer southern component are interesting and potentially important for outflow models.\n\nThe soft spot, and it's a real one, is the morphological discrimination. The paper compares a filled eB template against a shell-like Loop I template. But gamma-ray emission from a local Loop I structure need not be limb-brightened: inverse-Compton from electrons in the bubble interior would produce a filled morphology. If the true foreground is filled, the eB template will win the AIC comparison even if the excess is not from the eROSITA bubbles. The paper acknowledges that Loop I foreground can't be excluded, but it never tests a filled Loop I geometry, so the key discriminator is missing. This makes the eB interpretation plausible but not secure.\n\nOther nits: the residual significance map is approximate, the quoted TS and AIC differences come without uncertainties, and no code or template maps are released. These are fixable. The luminosity ratio is derived from the same template fits that define the structure, so it's circular in a sense, but that's typical for this kind of analysis.\n\nOn the whole, the detection of a southern excess is likely robust, and the paper is honest about its ambiguities. The physical interpretation as an eB counterpart is reasonable but underdetermined. I'd send it to peer review—a serious referee will likely push on the filled-Loop I issue, but the core observational result deserves to be out there.","headline":"Convincing southern gamma-ray excess; the eB-vs-Loop I morphology claim is plausible but under-tested because no filled Loop I template is considered.","tokens_in":13542,"tokens_out":2277,"would_cite":true,"duration_ms":20065,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":["95.85.Pw"],"model":"deepseek-v4-flash","headline":"A large gamma-ray structure in the southern sky matches the eROSITA bubble.","keywords":["Fermi-LAT","gamma-ray astronomy","eROSITA bubbles","Galactic diffuse emission","Loop I","Galactic outflow","inverse-Compton emission","north-south asymmetry"],"falsifier":"Repeat the template analysis with a diffuse background built from a different code or from a data-driven model not contained in the 64 GALPROP templates; if the southern excess vanishes or its template TS drops below roughly 100, the structure is an artifact of the tested background family rather than a genuine astrophysical component.","tokens_in":12507,"feed_emoji":"🌌","tokens_out":8596,"duration_ms":74322,"temperature":0.7,"pith_summary":"This paper claims that 17 years of Fermi-LAT data contain a previously unrecognized, tens-of-degrees-wide gamma-ray emission structure in the Southern Galactic Hemisphere. The excess persists when the fitted Galactic diffuse background is varied over an ensemble of 64 physically distinct models, and its spatial shape is better matched by a filled region following the southern eROSITA bubble than by the shell geometry of Loop I. If the structure is real, it would supply the long-missing southern gamma-ray counterpart to the eROSITA bubbles and to the northern large-scale gamma-ray emission, and it would imply that Galactic-center-driven outflows distribute nonthermal particles far beyond the Fermi bubbles. The paper also reports that the southern structure is fainter and softer than the northern one by a factor of about 7 in luminosity, mirroring the asymmetry seen in X-rays.","feed_headline":"Faint gamma-ray counterpart of the southern eROSITA bubble found","feed_subtitle":"Seventeen years of Fermi-LAT data show a soft, faint southern structure that fits the X-ray bubble better than Loop I.","key_machinery":"The machinery is a spatially binned, energy-dependent template likelihood analysis in which every emission component—Galactic diffuse emission, Fermi bubbles, isotropic background, and resolved 4FGL sources—is fitted simultaneously per energy bin, and then two non-nested large-scale templates are added and compared: a filled uniform-intensity region bounded by the eROSITA X-ray contours (the eB template) and the two intersecting synchrotron shells of the Wolleben Loop I model. The load-bearing comparison is the template TS relative to the same baseline and the AIC difference between the two shapes, repeated over 64 GALPROP diffuse-emission templates to show persistence. This isolates the southern excess from background uncertainties and lets the authors assign a morphology and spectrum to it.","core_discovery":"Using an energy-resolved maximum-likelihood decomposition of 17 years of Fermi-LAT CLEAN-class events from 200 MeV to 500 GeV, the authors subtract a baseline model built from GALPROP diffuse templates, the Fermi bubbles, isotropic emission, and 4FGL sources, and find a coherent residual excess stretching from mid-latitudes toward the South Galactic Pole. The excess is detected across all 64 alternative diffuse-emission templates; template fits give TS = 959.9 for a filled eROSITA-bubble-like template versus TS = 881.5 for the Wolleben Loop I shell geometry, with an AIC difference of -78.5 favoring the filled morphology on the benchmark background. Under the eB interpretation, the southern component's integrated 0.2-500 GeV luminosity is about 7.2 times lower than the corresponding northern structure and its spectrum is softer, consistent with an aged or less efficiently reaccelerated electron population; a hadronic interpretation would require roughly $10^{56}$ erg of protons and is judged energetically demanding. The authors present the structure as a plausible gamma-ray counterpart of the southern eROSITA bubble while explicitly allowing that Loop-I-related or other foreground emission cannot be excluded.","pith_inferences":["A testable extension: a deep radio-synchrotron survey of the same region should show non-thermal emission with the same filled-eB boundaries if freshly accelerated electrons are present, or match the Loop I radio shells if the foreground interpretation is correct.","Because the 64-template family is not exhaustive, the structure's reality will ultimately rest on future data-driven diffuse models; if such a model suppresses the excess, the eB association would have to be abandoned.","The reported soft spectrum and the CMB-only inverse-Compton fits imply a steep electron population above the break, so GHz-to-mm synchrotron observations could independently constrain the magnetic field and low-energy electron content in a way the gamma-ray data alone cannot.","A sharper test of the eB association is spatial: with more statistics, the gamma-ray boundary should track the eROSITA X-ray edge more tightly; any systematic divergence would point toward a separate foreground shell."],"forward_implications":["The southern gamma-ray structure, if confirmed, gives the eROSITA bubble system a measurable high-energy counterpart in both hemispheres, so the Milky Way's large-scale outflow must place nonthermal particles at radii of several kiloparsecs from the Galactic center.","The fitted luminosity ratio of about 7:1 (north to south) provides a gamma-ray-side measure of the north-south asymmetry that can be compared directly with the eROSITA X-ray brightness asymmetry.","The softer southern spectrum favors inverse-Compton emission from aged and/or reaccelerated electrons, so future models of Galactic-center feedback must include transport and in-situ reacceleration rather than only the Fermi-bubble electron population.","The energetics argument implies that a hadronic origin would require roughly $10^{56}$ erg in protons and around $10^{57}$ erg of mechanical energy, making that channel testable and, as modeled here, disfavored."],"supporting_citations":[{"why":"It defines the eROSITA bubble X-ray morphology used for the filled eB spatial template and supplies the roughly $10^{56}$ erg energy scale for comparison.","marker":"P. Predehl et al. 2020"},{"why":"It is the discovery paper of the Fermi bubbles and provides the Fermi-bubble template that is fitted as a baseline component.","marker":"M. Su et al. 2010"},{"why":"It provides the two-shell Loop I geometry used as the competing spatial template.","marker":"M. Wolleben 2007"},{"why":"It is the GALPROP code used to generate the 64 diffuse-emission templates that bracket the background.","marker":"T. A. Porter et al. 2022"},{"why":"It identified the northern gamma-ray counterpart to the NPS/Loop I region that the southern structure is compared with.","marker":"M. Ackermann et al. 2014"},{"why":"It predicted faint southern X-ray and gamma-ray emission from the eB outflow, motivating the search.","marker":"K. C. Sarkar 2019"},{"why":"It gave an earlier indication of extended southern gamma-ray emission near the Fermi bubbles that this analysis quantifies.","marker":"L. Scheel-Platz et al. 2023"},{"why":"It provides the eROSITA north-south intensity ratio used to compare with the gamma-ray luminosity asymmetry.","marker":"G. Mou et al. 2023"},{"why":"It provides the halo gas density estimate adopted for the hadronic energetics calculation.","marker":"M. J. Miller & J. N. Bregman 2015"},{"why":"It supplies a second halo-gas density estimate used in the same energetics argument.","marker":"Y. Faerman et al. 2022"}],"fun_headline_variants":["17 years of Fermi-LAT reveal soft southern gamma-ray bubble","Southern gamma-ray excess favors eROSITA bubble over Loop I shell","Fainter and softer southern bubble hints at aged particles","Gamma-ray excess persists across 64 background models","Southern gamma-ray bubble fainter and softer than northern"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The whole detection rests on the assumption that the 64 GALPROP templates bracket the true diffuse Galactic gamma-ray emission in the southern high-latitude sky; if the real background has structure outside that family, the residual excess could be a modeling artifact rather than a real astrophysical structure.","fun_headline_variants_meta":{"raw":{"variants":["17 years of Fermi-LAT reveal soft southern gamma-ray bubble","Southern gamma-ray excess favors eROSITA bubble over Loop I shell","Fainter and softer southern bubble hints at aged particles","Gamma-ray excess persists across 64 background models","Southern gamma-ray bubble fainter and softer than northern"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001541,"raw_usage":{"total_tokens":6205,"prompt_tokens":1027,"completion_tokens":5178,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":643,"completion_tokens_details":{"reasoning_tokens":5096}},"tokens_in":643,"tokens_out":5178,"duration_ms":37114,"temperature":1.0,"reasoning_tokens":5096,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-16T00:18:41.172154+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Repeat the template analysis with a diffuse background built from a different code or from a data-driven model not contained in the 64 GALPROP templates; if the southern excess vanishes or its template TS drops below roughly 100, the structure is an artifact of the tested background family rather than a genuine astrophysical component.","supporting_citations":[{"cited_title":"2020, Nature, 588, 227","cited_arxiv_id":null,"evidence_quote":"It defines the eROSITA bubble X-ray morphology used for the filled eB spatial template and supplies the roughly $10^{56}$ erg energy scale for comparison."},{"cited_title":"2007, The Astrophysical Journal, 664, 349","cited_arxiv_id":null,"evidence_quote":"It provides the two-shell Loop I geometry used as the competing spatial template."},{"cited_title":"2014, The Astrophysical Journal, 793, 64","cited_arxiv_id":null,"evidence_quote":"It identified the northern gamma-ray counterpart to the NPS/Loop I region that the southern structure is compared with."},{"cited_title":"2023, Astronomy & Astrophysics, 680, A2","cited_arxiv_id":null,"evidence_quote":"It gave an earlier indication of extended southern gamma-ray emission near the Fermi bubbles that this analysis quantifies."},{"cited_title":"J., & Bregman, J","cited_arxiv_id":null,"evidence_quote":"It provides the halo gas density estimate adopted for the hadronic energetics calculation."},{"cited_title":"S., & Sternberg, A","cited_arxiv_id":null,"evidence_quote":"It supplies a second halo-gas density estimate used in the same energetics argument."}],"review_version":1}