{"id":"222f45e5-7a62-46a1-93b0-d3cdbd17cd10","arxiv_id":"2508.21646","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"Updated KASCADE-Grande cosmic ray energy spectra, obtained with QGSJET-II-04, EPOS-LHC, SIBYLL 2.3d and the first test of EPOS-LHC-R, confirm a heavy knee near 100 PeV and a light-component flattening.","lead":"KASCADE-Grande data from 17 million air showers are re-analyzed with four modern hadronic interaction models, giving updated cosmic ray energy spectra from 10 PeV to 1 EeV. The results confirm a heavy-component knee near 100 PeV and a light-component hardening, and provide the first KASCADE-Grande test of the new EPOS-LHC-R model.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"EPOS-LHC-R confirmation is compared on unequal footing: Fig. 5/Table 1 mix an unfolded spectrum for three models with a raw EPOS-LHC-R spectrum, so the claimed persistence with EPOS-LHC-R is not yet demonstrated.","rationale":"The reader's weakest assumption identifies a broader circularity (EPOS-LHC-R is calibrated with EPOS-LHC-R), which is real. I focus instead on a sharper, immediately testable flaw in the same conclusion: the EPOS-LHC-R comparison in Fig. 5 and Table 1 is between an unfolded spectrum (three models) and a raw spectrum (EPOS-LHC-R). The paper transparently flags this, but the transparency does not make the comparison valid. Since the central feature persistence is already established by the three unfolded models, this concern does not overturn the paper; it only means the new model's confirmation is preliminary. That is exactly a conditional verdict. I therefore leave the reader's verdict unchanged. Agreement is partial because my concern is not the calibration circularity but the unequal analysis state, though both target the strength of the EPOS-LHC-R consistency claim.","tokens_in":6217,"tokens_out":6961,"duration_ms":83962,"concrete_test":"Apply the same Bayesian unfolding used for the other models to the EPOS-LHC-R spectrum: construct/use the EPOS-LHC-R response matrix R_ij from Section 2, unfold the raw heavy and light spectra, and refit the heavy component with the same broken power law as Table 1. Compare the new break position, gamma1, gamma2, and the light hardening to the raw EPOS-LHC-R row and to the other three models. If the break moves by more than ~0.05 in log10(E/GeV) or Delta gamma changes by more than ~0.1, the EPOS-LHC-R confirmation claim is not established; if stable, the claim survives.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim that the heavy knee and light hardening persist with EPOS-LHC-R relies on Fig. 5 and Table 1. Section 3 explicitly states that the EPOS-LHC-R spectrum is still raw (shower-to-shower fluctuations not yet corrected), whereas QGSJET-II-04, EPOS-LHC, and SIBYLL 2.3d spectra have been unfolded. Table 1 even notes that the EPOS-LHC-R fit is based on the raw spectrum. The paper's estimate that unfolding changes fluxes by less than 10% does not settle the comparison: a 10% bin-to-bin shift can move a fitted break position and slopes, especially if the migration correction is non-smooth near 10^17 eV, and for the light component the paper itself says the flux 'probably will slightly increase' after unfolding. Thus the conclusion that 'EPOS-LHC-R ... shows spectral behavior consistent with other models and confirms previous findings' is not supported by an apples-to-apples comparison. The three-model persistence claim is unaffected; only the EPOS-LHC-R confirmation leg is at risk.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This ICRC contribution uses roughly 17 million KASCADE-Grande events to reconstruct the energy spectra of heavy (Si+Fe) and light (H+He+CNO) primary cosmic rays with three post-LHC hadronic interaction models: QGSJET-II-04, EPOS-LHC, and SIBYLL 2.3d. The reconstruction uses the y_CIC electron/muon correlation for mass separation and a Bayesian unfolding procedure to correct shower-to-shower fluctuations. The authors report a knee-like break in the heavy component near 10^17 eV and a hardening of the light component above about 10^17 eV for all three models, with systematic uncertainties quoted as about 13% (light) and 10% (heavy) at 10^17 eV. A first comparison with the newly released EPOS-LHC-R model is also presented, including new y_CIC separation threshold and energy calibration coefficients, but the EPOS-LHC-R spectrum is explicitly stated to be raw (not unfolded). All-particle spectra are compared with ESA-TOP, IceTop, Pierre Auger, and Telescope Array.","tokens_in":6514,"tokens_out":5560,"duration_ms":63611,"significance":"If the three-model result holds, it provides a useful multi-model confirmation of the heavy knee and the light-component hardening around 10^17 eV, and the quantified systematics and public KCDC data release are strengths. The EPOS-LHC-R section is a valuable preliminary test but is not yet on equal footing with the other models because the comparison uses a raw spectrum. The central three-model spectral features are not endangered by this issue; only the strength of the EPOS-LHC-R confirmation claim needs revision.","major_comments":[{"comment":"The statement that EPOS-LHC-R 'shows spectral behavior consistent with other models and confirms previous findings' is not supported by an apples-to-apples comparison. Figure 5 and Table 1 mix unfolded spectra (QGSJET-II-04, EPOS-LHC, SIBYLL 2.3d) with the raw EPOS-LHC-R spectrum, as acknowledged in the text and Table 1 note. The <10% unfolding estimate does not by itself justify that fitted slopes and the break position are unaffected, especially because the paper notes the light flux 'probably will slightly increase' after unfolding. Either unfold the EPOS-LHC-R spectrum before quantitative comparison, or restrict the conclusion to a preliminary raw-spectrum comparison and label this clearly in the figure and conclusion.","section":"Sec. 3, Fig. 5, Table 1, Conclusion"}],"minor_comments":[{"comment":"Typo: 'which is is of the order' should be 'which is of the order'.","section":"Sec. 2"},{"comment":"Energy notation is inconsistent: the text switches between '10^17 eV', '10 17 eV', and '1017 eV'. Please use a uniform superscript notation.","section":"Sec. 2, Sec. 3"},{"comment":"The figure caption should state explicitly that the EPOS-LHC-R spectrum is raw, as the text does. This would prevent readers from misinterpreting the comparison.","section":"Fig. 5"},{"comment":"Because the y_CIC separation threshold (0.8496) and the energy calibration coefficients (a=0.897, b=1.705 heavy; a=0.953, b=1.109 light) are fitted to EPOS-LHC-R simulations, the subsequent comparison is a consistency check within a model-dependent reconstruction rather than a fully independent test. A sentence acknowledging this would be appropriate.","section":"Sec. 3"},{"comment":"The abbreviation y_CIC is used without definition. Please define 'CIC' at first use (presumably 'charged-particle / muon correlation' or similar).","section":"Sec. 2"}],"recommendation":"minor_revision","confidential_remarks":"This is a conference proceedings contribution and the three-model spectral analysis is sound. The main issue is an overclaim about EPOS-LHC-R confirmation based on a raw spectrum; a careful wording change or a preliminary-statement restriction should suffice. I do not consider the preliminary nature of the EPOS-LHC-R result grounds for rejection."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"KASCADE-Grande keeps producing. The genuinely new content here is (1) updated light and heavy energy spectra from the full dataset using QGSJET-II-04, EPOS-LHC and SIBYLL 2.3d with shower-to-shower fluctuation corrections, and (2) a first look at EPOS-LHC-R with KASCADE-Grande data. The central result—a knee-like steepening in the heavy component near 10^17 eV and a hardening of the light component above that—is consistent across the three unfolded models and holds up.\n\nWhat the paper does well: it is transparent. The y_CIC separation threshold and the energy calibration coefficients are stated explicitly (0.8496; a=0.897, b=1.705 heavy; a=0.953, b=1.109 light). The overall systematic uncertainty is quoted (13% light, 10% heavy at 10^17 eV). The comparison with Auger, TA, IceTop and ESA-TOP gives useful context, and the data go to KCDC.\n\nThe soft spots are real but not fatal. The stress-test note is right: the EPOS-LHC-R comparison in Fig. 5 and Table 1 mixes a raw spectrum with three unfolded ones. The paper's own '<10% effect' estimate does not settle it, because a 10% bin-to-bin migration can shift a fitted break position and slopes; and for the light component the authors themselves say the flux 'probably will slightly increase' after unfolding. So the conclusion that EPOS-LHC-R 'confirms previous findings' is too strong as stated. It should be 'consistent within current (uncorrected) systematic uncertainties' or they should show the unfolded version. This is easily addressable. The calibration circularity is also worth flagging: the EPOS-LHC-R energy scale and mass-separation threshold are derived from EPOS-LHC-R Monte Carlo, so the 'test' of that model is not independent. For the three-model persistence claim this matters less, since the features survive across models with different tunes.\n\nThe paper defers some systematic detail to companion papers—normal for ICRC proceedings, but it means the standalone value is limited.\n\nVerdict: this deserves a serious referee. The measurement is useful, the analysis is honest, and the EPOS-LHC-R comparison needs a revision, not a rejection. I'd bring it to a reading group if the topic is cosmic-ray composition; otherwise a skim suffices.","headline":"Solid KASCADE-Grande update: the three-model unfolded spectra confirm the heavy knee and light hardening; the EPOS-LHC-R leg is suggestive but preliminary because it compares raw to unfolded.","tokens_in":7163,"tokens_out":2423,"would_cite":true,"duration_ms":26101,"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 shows that the heavy cosmic-ray knee near 10^17 eV and the hardening of the light component above it persist when KASCADE-Grande data are reconstructed with four post-LHC hadronic interaction models, including the first test of E","keywords":["cosmic ray energy spectrum","KASCADE-Grande","hadronic interaction models","EPOS-LHC-R","heavy knee","light hardening","mass composition","air showers"],"falsifier":"Compare the observed y_CIC = log10(N_mu)_CIC_CF / log10(N_ch)_CIC distribution for events near log10(E/GeV) = 7.8 with the EPOS-LHC-R-simulated distribution used to set the 0.8496 threshold: if the measured fraction of electron-rich events in that energy bin differs from the model expectation by more than the quoted roughly 13% systematic uncertainty, the reconstructed heavy-knee position and light-hardening energy would shift, weakening the central claim.","tokens_in":6144,"feed_emoji":"☄️","tokens_out":9799,"duration_ms":99556,"temperature":0.7,"pith_summary":"The paper tries to establish that two key features of the cosmic-ray energy spectrum—a knee-like break in the heavy component around 10^17 eV and a hardening of the light component just above that energy—do not depend on which modern hadronic interaction model is used to interpret the data. The same features appear with QGSJET-II-04, EPOS-LHC, SIBYLL 2.3d, and, for the first time with this data set, the newly released EPOS-LHC-R model. If true, these spectral features are real astrophysical boundaries rather than artifacts of a particular simulation model, and they can serve as fixed targets for theories of the transition from galactic to extragalactic cosmic rays. The paper also provides updated spectra with explicit shower-to-shower fluctuation corrections and an early test of EPOS-LHC-R against a ground-based air-shower array.","feed_headline":"Four shower models agree: heavy knee at 10^17 eV, light hardening","feed_subtitle":"KASCADE-Grande tests the new LHC-tuned model EPOS-LHC-R and confirms the same heavy knee and light hardening.","key_machinery":"The analysis is carried by the y_CIC parameter, defined as log10(N_mu)_CIC_CF / log10(N_ch)_CIC, the ratio of the attenuation-corrected muon number to the charged-particle number, which sorts each event into an electron-rich light group (proton, helium, CNO) or an electron-poor heavy group (silicon, iron). For each hadronic model, energy calibration is a linear fit in log-log space, log10(E_true/GeV) = a * log10(N_ch) + b, with separate coefficients for the two mass groups; for EPOS-LHC-R the mass-separation threshold is 0.8496. A response matrix built from Monte Carlo showers is unfolded with an iterative Bayesian algorithm to correct bin-to-bin migrations from shower-to-shower fluctuations","core_discovery":"The paper asserts that KASCADE-Grande's shower-size measurements, analyzed with the y_CIC mass-separation technique and per-model energy calibrations, yield a heavy-component spectrum with a knee-like break at about 10^17 eV and a light-component spectrum that hardens above that same energy, in all three established post-LHC models. It then applies the same procedure to EPOS-LHC-R, a new model with deeper shower maxima and altered muon production, and finds that EPOS-LHC-R reproduces the same spectral features: a heavy knee at log10(Ek/GeV) around 7.79 and a light-component hardening, with the all-particle spectrum differing only mildly from the other models. The paper concludes that the hea","pith_inferences":["The EPOS-LHC-R comparison in the paper is not fully symmetric because its spectrum has not yet been unfolded while the other three have; applying the same response-matrix unfolding to EPOS-LHC-R is the direct next step and would likely shift its light-component flux slightly upward.","Because the EPOS-LHC-R energy calibration and the 0.8496 mass-separation threshold are themselves derived from EPOS-LHC-R simulations, the 'test' of the model is partly self-referential; a stronger test would measure the same spectral features with an energy estimator independent of hadronic models, such as radio or Cherenkov emission.","If both the heavy knee and the light hardening are composition-driven, their near-coincidence around 10^17 eV suggests a single rigidity-dependent transition mechanism, which could be tested by checking whether the break energy scales with nuclear charge across the light and heavy groups.","EPOS-LHC-R's relative muon content now sits closer to QGSJET-II-04 and SIBYLL 2.3d than to the older EPOS-LHC, which may point toward a partial resolution of earlier muon-deficit discrepancies in air-shower simulations."],"forward_implications":["The heavy-component break is stable across models at log10(Ek/GeV) about 7.77 to 7.79 with a spectral-index change of about 0.5, so any model of the galactic-to-extragalactic transition must produce a steepening in the silicon-plus-iron flux at that energy.","The light-component hardening above 10^17 eV is also model-independent, strengthening the interpretation that an extragalactic contribution starts to dominate below 10^17 eV.","EPOS-LHC-R, with its deeper shower maximum but smaller relative muon content, produces spectra consistent with the other models, making it a viable model for future air-shower analyses of KASCADE-Grande-type data.","The model-averaged KASCADE-Grande all-particle flux is within about 10% of other experiments and statistically consistent with the Pierre Auger Observatory near 10^18 eV, allowing cross-experiment fits to use these spectra.","The unfolded spectra for QGSJET-II-04, EPOS-LHC, and SIBYLL 2.3d, together with the raw EPOS-LHC-R spectrum, give a concrete set of break positions and slopes for testing upcoming hadronic interaction models."],"supporting_citations":[{"why":"Establishes the KASCADE-Grande detector and data set whose shower-size measurements are the basis of all reconstructed spectra.","marker":"[2]"},{"why":"Provides the response-matrix and unfolding method used to correct the reconstructed spectra for shower-to-shower fluctuations.","marker":"[4]"},{"why":"Previously reported the iron-like break near 10^17 eV that this paper re-derives consistently across all four interaction models.","marker":"[5]"},{"why":"Previously observed the ankle-like flattening of the light component that this paper confirms with updated data and models.","marker":"[6]"},{"why":"Supplies the CORSIKA simulation program used to generate the model-dependent shower and detector responses.","marker":"[7]"},{"why":"Defines the QGSJET-II-04 hadronic interaction model, one of the three baseline reconstructions.","marker":"[8]"},{"why":"Defines the EPOS-LHC hadronic interaction model, the second baseline reconstruction and predecessor of EPOS-LHC-R.","marker":"[9]"},{"why":"Defines the SIBYLL 2.3d hadronic interaction model, the third baseline reconstruction whose relative muon content EPOS-LHC-R now approaches.","marker":"[10]"},{"why":"Supplies the new EPOS-LHC-R model and its predicted changes in cross-sections, nuclear fragmentation, muon production, and shower maximum.","marker":"[11]"},{"why":"Gives the per-mass-group energy calibration coefficients and the systematic uncertainty budget used in the spectral reconstruction.","marker":"[13]"}],"fun_headline_variants":["EPOS-LHC-R first run on KASCADE-Grande: same heavy knee, light hardening","One new model, three old ones: all agree on cosmic ray knee and hardening","Heavy knee at 100 PeV and light hardening hold in every shower model tested","First cosmic ray test of EPOS-LHC-R matches older model results","KASCADE-Grande: new model reproduces heavy knee and light hardening"],"cache_read_input_tokens":2688,"weakest_assumption_plain":"The reconstruction assumes that the Monte Carlo showers produced by each hadronic model—especially the new EPOS-LHC-R—give the correct relation between measured charged-particle and muon numbers and the true primary energy; if a model gets the muon content wrong, the inferred knee position and light-hardening energy shift.","fun_headline_variants_meta":{"raw":{"variants":["EPOS-LHC-R first run on KASCADE-Grande: same heavy knee, light hardening","One new model, three old ones: all agree on cosmic ray knee and hardening","Heavy knee at 100 PeV and light hardening hold in every shower model tested","First cosmic ray test of EPOS-LHC-R matches older model results","KASCADE-Grande: new model reproduces heavy knee and light hardening"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.001016,"raw_usage":{"total_tokens":4096,"prompt_tokens":681,"completion_tokens":3415,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":425,"completion_tokens_details":{"reasoning_tokens":3309}},"tokens_in":425,"tokens_out":3415,"duration_ms":24595,"temperature":1.0,"reasoning_tokens":3309,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-05T14:04:58.625142+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compare the observed y_CIC = log10(N_mu)_CIC_CF / log10(N_ch)_CIC distribution for events near log10(E/GeV) = 7.8 with the EPOS-LHC-R-simulated distribution used to set the 0.8496 threshold: if the measured fraction of electron-rich events in that energy bin differs from the model expectation by more than the quoted roughly 13% systematic uncertainty, the reconstructed heavy-knee position and light-hardening energy would shift, weakening the central claim.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Establishes the KASCADE-Grande detector and data set whose shower-size measurements are the basis of all reconstructed spectra."},{"cited_title":"Apel et al., KASCADE-Grande Collaboration, Astrop","cited_arxiv_id":null,"evidence_quote":"Provides the response-matrix and unfolding method used to correct the reconstructed spectra for shower-to-shower fluctuations."},{"cited_title":"Apel et al., KASCADE-Grande Collaboration, Phys","cited_arxiv_id":null,"evidence_quote":"Previously reported the iron-like break near 10^17 eV that this paper re-derives consistently across all four interaction models."},{"cited_title":"Apel et al., KASCADE-Grande Collaboration, Phys","cited_arxiv_id":null,"evidence_quote":"Previously observed the ankle-like flattening of the light component that this paper confirms with updated data and models."},{"cited_title":"Heck et al., Report Forschungszentrum Karlsruhe, FZKA 6019 (1998)","cited_arxiv_id":null,"evidence_quote":"Supplies the CORSIKA simulation program used to generate the model-dependent shower and detector responses."},{"cited_title":"Ostapchenko, Phys","cited_arxiv_id":null,"evidence_quote":"Defines the QGSJET-II-04 hadronic interaction model, one of the three baseline reconstructions."},{"cited_title":"Pierog et al., Phys","cited_arxiv_id":null,"evidence_quote":"Defines the EPOS-LHC hadronic interaction model, the second baseline reconstruction and predecessor of EPOS-LHC-R."},{"cited_title":"Riehn et al., Phys","cited_arxiv_id":null,"evidence_quote":"Defines the SIBYLL 2.3d hadronic interaction model, the third baseline reconstruction whose relative muon content EPOS-LHC-R now approaches."},{"cited_title":"Pierog and K","cited_arxiv_id":null,"evidence_quote":"Supplies the new EPOS-LHC-R model and its predicted changes in cross-sections, nuclear fragmentation, muon production, and shower maximum."},{"cited_title":"Kang et al., KASCADE-Grande Collaboration, PoS(ICRC2023)307","cited_arxiv_id":null,"evidence_quote":"Gives the per-mass-group energy calibration coefficients and the systematic uncertainty budget used in the spectral reconstruction."}],"review_version":1}