{"id":"95301737-efe7-4780-9d73-d2b6a30ef031","arxiv_id":"2507.06698","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":13,"one_line_summary":"The authors extend a 30-80 MHz air-shower radio emission model to 50-200 MHz and report simulation-based electromagnetic energy resolutions below 5% for ideal arrays and below 10% for sparse, noisy arrays.","lead":"This paper adapts an existing radio-signal model for inclined cosmic-ray air showers to the 50-200 MHz band used by GRAND, tuning it with CoREAS simulations for Argentina and China. If the simulation-based performance holds in real measurements, it provides an energy-reconstruction calibration for future 50-200 MHz radio arrays.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The headline <5% intrinsic resolution is evaluated on the same CoREAS libraries used to fit the model parameters, so it is a self-consistency metric; a train/test split is needed before it can support a predictive energy-resolution claim.","rationale":"The paper does what it claims: it adapts a published 30–80 MHz model to 50–200 MHz, with explicit parametrisations and performance figures, and it includes a separate sparse-array test with noise. The strongest numerical claim, however, is the <5% intrinsic resolution. That number is produced by fitting and evaluating on the same star-shaped CoREAS libraries, so the reported scatter could be an in-sample artifact. The proposed split test is cheap, does not require new simulations, and would settle the issue quantitatively. I do not regard this as a reason to reject the paper: the sparse-array results already provide some out-of-sample evidence, and the reader's CONDITIONAL verdict appropriately asks for validation before the headline is treated as a measurement-level resolution. The external question of whether CoREAS correctly models 50–200 MHz coherence loss remains for real data, but that is a separate, slower test. My verdict is UNCHANGED because the conditional verdict already captures the concern.","tokens_in":7541,"tokens_out":10311,"duration_ms":118661,"concrete_test":"Split each of the two star-shaped CoREAS libraries (~4,000 events) into two halves, stratified by primary energy and zenith angle. Refit all model parameters from scratch on the first half only: charge-excess coefficients (Eqs. 1–2), LDF shape parameters (Table 1), and density-correction/S19/γ values (Eqs. 7–8). Apply the refitted model to the held-out half and compute resolution and bias as in Fig. 4. Repeat with the halves exchanged. If the held-out resolution remains below 5% with negligible bias, the in-sample concern is resolved. If it degrades by more than ~1–2 percentage points, the <5% headline should be reported as a cross-validated or fit-quality metric rather than a predictive resolution. A complementary check is to refit on star-shaped libraries and validate on the sparse, noisy GP300 libraries without any further tuning.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section 4 states that the first benchmark is performed on 'the simulation libraries with star-shaped antenna layout with which we tuned the parameters'. All model ingredients—the charge-excess parametrisations (Eqs. 1 and 2), the LDF shape parametrisations (Table 1), and the joint fit of the density correction and the Sgeo–Eem power law (Eqs. 7 and 8)—are tuned on those same libraries, and Fig. 4 reports the <5% resolutions on the same libraries. A multi-parameter fit to ~4,000 MC events can absorb library-specific correlations, so the quoted scatter may reflect interpolation rather than predictive accuracy. The sparse-array benchmarks in Fig. 5 use separate GP300 simulations, which is a genuine out-of-sample test of the full pipeline, but they neither isolate the <5% star-shaped claim nor remove the dependence on parameters fixed on the star-shaped libraries. The abstract's <5% is therefore a self-consistency statement until a train/test split or independent validation is shown. Parameter uncertainties are not given, so the stability of the iterative LDF parametrisation cannot be assessed from the paper.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper adapts the radio-detection signal model of Schlüter and Huege from the 30-80 MHz band to the 50-200 MHz band, for the Pierre Auger site in Argentina and the GRANDProto300 site in China. It introduces a refitted charge-excess parametrization, a modified lateral distribution function with shape parameters parametrized by the distance to shower maximum, and a density correction to account for coherence loss in the stronger Chinese magnetic field. Using CoREAS simulations, the authors report intrinsic energy resolutions below 5% for ideal star-shaped antenna layouts at both sites, and below 10% for sparse, noisy arrays representing GRANDProto300 and a 10,000 km^2 array.","tokens_in":74,"tokens_out":5705,"duration_ms":75577,"significance":"If supported, an energy resolution below 5% for inclined air showers in the 50-200 MHz band, with a separate demonstration below 10% on realistic sparse arrays, would be a valuable extension of the established 30-80 MHz method and directly relevant to GRAND and other high-frequency radio detectors. The paper's concrete two-site parameter tables, the explicit LDF functional form, and the independent simulation sets with noise, amplitude smearing, and time jitter are useful contributions; the GP300 sparse-array benchmark is a genuine out-of-sample test of the reconstruction pipeline. However, the headline <5% claim is currently evaluated on the same simulation libraries used to tune the model parameters, so it is best interpreted as an in-sample consistency measure rather than a demonstrated predictive resolution.","major_comments":[{"comment":"The <5% resolutions in Fig. 4 are obtained on 'the simulation libraries with star-shaped antenna layout with which we tuned the parameters.' All model ingredients, including the charge-excess coefficients in Eqs. (1)-(2), the LDF shape coefficients in Table 1, and the density and energy calibration parameters in Eqs. (7)-(8), are fitted to these same libraries. With roughly 4,000 events and tens of fitted parameters, the quoted scatter is not an unbiased estimate of predictive performance. Please add an explicit train/test split or K-fold cross-validation, for example by fitting on disjoint subsets in energy and zenith angle and reporting the held-out resolution; alternatively, if the in-sample 'intrinsic' resolution is the intended quantity, the claim should be reworded and supported by a stability analysis such as bootstrap uncertainties on the fitted parameters.","section":"Section 4, first paragraph and Fig. 4"},{"comment":"The sparse-array simulations are independent of the tuning libraries, which is a strength, but they are performed only for the China site and they reuse the star-shaped China calibration values shown in Fig. 4, including S19=14.16 GeV and gamma=1.9970. The sparse-array results therefore validate the LDF fit and the full reconstruction pipeline for one magnetic-field configuration only; they do not by themselves support the conclusion that the method is readily adaptable to any magnetic field configuration at sparse arrays. Please add a comparable out-of-sample sparse-array test for the Argentina site, or clearly qualify the conclusion to state that sparse-array performance has been demonstrated only for China.","section":"Section 4, Fig. 5"},{"comment":"No statistical uncertainties or per-bin event counts are reported for the resolution and bias values. The resolution is derived from bin-wise distributions, and the '<5%' and '<10%' statements may depend on the binning and on finite Monte Carlo statistics. Without uncertainties or event counts, it is difficult to judge whether differences between the two sites or between the ideal and sparse configurations are significant. Please add estimates of the statistical uncertainty on each resolution and bias point, or at least provide the number of events per bin.","section":"Section 4, Figs. 4-5"}],"minor_comments":[{"comment":"The sentence 'the atmosphere models are provided by the radiotools package [11])' contains an extra closing parenthesis after the citation; please remove it.","section":"Section 1"},{"comment":"The expression for p(r) in the r >= r0 branch is ambiguous: it should likely read p(r) = 2*(r0/r)^(b/1000) rather than the printed '2*(r0/r)^b/1000'. Please add explicit parentheses so that the mild decrease of the exponent with distance is unambiguous.","section":"Section 2, Eq. (4)"},{"comment":"The table caption states that the first four parameters use the cpar(dmax) term of Eq. (6), but it would help to state the units of dmax explicitly in the caption or in the text around Eq. (6), since the coefficients carry km and km^2 units.","section":"Section 2, Table 1"},{"comment":"The density-correction formula mixes parameters with different dimensions inside one expression; a short sentence defining the units of p2 and p3, and noting that the denominator is dimensionless, would improve readability.","section":"Section 3, Eq. (7)"},{"comment":"The paper would benefit from a data-availability or reproducibility statement describing how the CoREAS libraries and analysis scripts can be accessed, since the numerical parameter tables alone do not allow the fits to be reproduced.","section":"General"}],"recommendation":"major_revision","confidential_remarks":"The manuscript is a conference-style contribution whose central claims are plausible but currently rest on an in-sample evaluation of the headline <5% resolution. The authors should be asked either to provide a genuine train/test split or to substantially reword the claim and add stability tests. The sparse-array results are the strongest part of the paper, but they cover only one site; extending them to Argentina, or qualifying the conclusion, is needed before the 'adaptable to any magnetic field configuration' statement is fully supported."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Felix,\n\nThe quick summary: the paper extends the Schlüter–Huege 30–80 MHz signal model to 50–200 MHz, with re-fitted charge-excess parametrizations, a modified LDF with a variable Gaussian exponent, and a density correction intended to capture coherence loss. That's genuinely new work, and it's done carefully for two sites (Argentina and China) that bracket the range of magnetic-field strengths on Earth. The sparse-array benchmarks, using independent CoREAS simulations with added noise, are the strongest part: <10% energy resolution with negligible bias for both GP300 and a 10,000 km² array. That is an out-of-sample test of the full pipeline, and it is a useful result for GRAND and for higher-frequency Auger extensions.\n\nThe soft spots are exactly where the stress-test note points. The headline <5% resolution, shown in Fig. 4, is evaluated on the same simulation libraries used to tune all the model parameters — the charge-excess fits, the LDF shape parameters, and the joint fit of the density correction and the Sgeo–Eem power law. The paper says so itself in the opening sentence of Section 4. So those numbers are a self-consistency metric, not a predictive statement. The sparse-array tests avoid that circularity, but they don't isolate the <5% claim. Parameter uncertainties are absent throughout, so you can't judge how stable the iterative LDF parametrisation is. No code or data are released either, which makes it harder to reproduce the fits.\n\nNone of this is disqualifying. The authors are transparent about the tuning, the out-of-sample sparse tests do support the main practical conclusion, and the physical motivation — coherence loss matters more in China — is plausible and grounded in the recent Chiche et al. result. But a serious referee should ask for a train/test split or an explicit statement that the <5% is in-sample, and ideally some parameter stability checks.\n\nWho is this for: anyone working on radio detection of inclined cosmic-ray or neutrino showers at higher frequencies, especially GRAND, GRANDProto300, and AugerPrime. It deserves a proper peer review — this is exactly the kind of methods paper a conference proceedings can't fully vet. I'd cite it if I were doing this work, with the in-sample caveat noted.\n\nVerdict: engage, but treat the headline as Monte-Carlo self-consistency until the train/test split is shown.","headline":"Useful frequency-band extension of the Schlüter–Huege radio signal model, but the <5% resolution is an in-sample fit on the tuning libraries; the sparse-array <10% results are the stronger, out-of-sample evidence.","tokens_in":8543,"tokens_out":2155,"would_cite":true,"duration_ms":23392,"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":"Adapting a radio signal model to the 50–200 MHz band, this paper reconstructs electromagnetic air-shower energies with intrinsic resolution below 5% at both the Auger and GRANDProto300 sites, and below 10% on realistic sparse antenna…","keywords":["cosmic-ray air showers","radio detection","energy reconstruction","lateral distribution function","geomagnetic radiation","charge excess","CoREAS","GRANDProto300"],"falsifier":"Generate independent CoREAS libraries with a different hadronic interaction model (e.g., EPOS-LHC vs. Sibyll2.3) or with an independent simulation code such as ZHAireS, re-run the fixed tuning of this paper, and check whether the intrinsic resolution stays below 5% and the sparse-array resolution below 10%. If the reconstruction degrades significantly, the quoted resolutions are tuned to one simulation library rather than to the physics.","tokens_in":7263,"feed_emoji":"📡","tokens_out":8570,"duration_ms":87500,"temperature":0.7,"pith_summary":"This paper extends an existing radio-based energy reconstruction method for inclined cosmic-ray air showers from the 30–80 MHz band up to the 50–200 MHz band used by the GRAND detector family. The method isolates the geomagnetic component of the radio emission, fits a lateral distribution function to it, and applies geometry and air-density corrections to link the measured radiation energy to the electromagnetic shower energy. On CoREAS simulations, the reconstruction achieves an intrinsic energy resolution below 5% at the Pierre Auger site in Argentina and at the GRANDProto300 site in China, despite the latter having a magnetic field almost three times stronger. When tested on realistic sparse antenna layouts with added noise, the resolution stays below 10% with negligible bias, which is what matters for building practical radio arrays.","feed_headline":"Radio model hits sub-5% shower-energy resolution","feed_subtitle":"Adapted to 50–200 MHz, it stays under 10% on sparse, noisy arrays.","key_machinery":"The central object is the lateral distribution function (LDF) of the geomagnetic energy fluence, Eq. (3), a sum of a Gaussian peak and a sigmoid term whose seven shape parameters are parametrised as functions of the distance to shower maximum, $d_{\\max}$, so that the fit reduces to four degrees of freedom: the radiation energy $E_{\\mathrm{geo}}$ and the two radio-core coordinates. The LDF is fitted after an early-late correction removes geometric asymmetries and a per-site parametrisation of the charge-excess fraction isolates the geomagnetic component. The second load-bearing element is the modified density correction, Eq. (7), which compensates coherence loss before the corrected radiation energy is converted to electromagnetic energy with a power law, Eq. (8).","core_discovery":"The paper's central claim is that the geomagnetic radiation energy of an inclined air shower, reconstructed with a parametrised lateral distribution function fitted to the charge-excess-subtracted radio fluence, determines the electromagnetic shower energy to better than 5% intrinsic resolution at two benchmark sites with very different magnetic fields. The same reconstruction applied to simulated sparse arrays, including instrumental noise, smearing and timing jitter, yields a resolution better than 10% with negligible bias. To reach this, the authors re-fit the charge-excess fraction parametrisation for the higher-frequency band, express all lateral-distribution shape parameters as functions of the distance to shower maximum, and introduce a modified density correction that absorbs the coherence loss caused by shorter wavelengths and, especially in China, the strong magnetic field.","pith_inferences":["The $<5\\%$ figure is computed on the same star-shaped simulation libraries used to tune the LDF parameters; an independent test on simulations generated with a different interaction model or code is needed before treating it as a guaranteed error budget.","The realistic benchmark adds Gaussian noise, amplitude smearing and timing jitter, but does not simulate real radio-frequency interference, antenna gain errors or atmospheric uncertainties, so real-data resolution is likely to be somewhat worse.","Since the model's $S_{19}$ parameter, the reference radiation energy at 10 EeV, differs between the two sites as expected from the magnetic field scaling, a cross-check against an absolute energy scale from fluorescence or surface detectors could validate the model's normalisation."],"forward_implications":["If the simulation results carry over to data, GRANDProto300 can achieve sub-10% energy resolution for inclined showers in the $10^{17}$–$10^{20}$ eV range on its sparse antenna grid.","The same reconstruction works for a very large array of 10,000 km$^2$ with 1 km spacing, again with $<10\\%$ resolution, which is relevant for the full GRAND design.","The fit requiring only five antennas with signal implies that even partial or prototype arrays can provide per-shower energy reconstruction.","At low cosmic-ray energies, where signal approaches the noise level, the resolution degrades, setting the practical energy threshold of the method."],"supporting_citations":[{"why":"Supplies the original 30–80 MHz signal model, the charge-excess parametrisation framework, and the LDF form that this paper adapts to 50–200 MHz.","marker":"[1]"},{"why":"Provides the early-late correction used to symmetrise the inclined-shower emission pattern.","marker":"[6]"},{"why":"CoREAS simulation code used to generate the air-shower libraries that serve as both tuning data and truth for the resolution claims.","marker":"[7]"},{"why":"radiotools package supplies the atmosphere models and the theoretical Cherenkov-ring radius prediction used in the LDF parametrisation.","marker":"[11]"},{"why":"Characterises coherence loss and polarisation effects in strong magnetic fields, motivating the modified density correction for the Chinese site.","marker":"[12]"},{"why":"Explains refractive displacement of the radio core, justifying the free radio-core position in the LDF fit.","marker":"[13]"}],"fun_headline_variants":["Radio shower model nails sub-5% energy resolution","Inclined showers: radio energy to <5% accuracy","Adapted radio model: <5% shower energy error","Wideband radio model tracks shower energy under 5%","Radio reconstruction: sub-5% energy resolution"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The central calibration and the headline resolutions both come from CoREAS simulations, so the whole argument collapses if CoREAS does not correctly predict 50–200 MHz radio emission, particularly the coherence loss in strong magnetic fields.","fun_headline_variants_meta":{"raw":{"variants":["Radio shower model nails sub-5% energy resolution","Inclined showers: radio energy to <5% accuracy","Adapted radio model: <5% shower energy error","Wideband radio model tracks shower energy under 5%","Radio reconstruction: sub-5% energy resolution"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000155,"raw_usage":{"total_tokens":1200,"prompt_tokens":919,"completion_tokens":281,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":535,"completion_tokens_details":{"reasoning_tokens":201}},"tokens_in":535,"tokens_out":281,"duration_ms":3476,"temperature":1.0,"reasoning_tokens":201,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T18:56:54.178181+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Generate independent CoREAS libraries with a different hadronic interaction model (e.g., EPOS-LHC vs. Sibyll2.3) or with an independent simulation code such as ZHAireS, re-run the fixed tuning of this paper, and check whether the intrinsic resolution stays below 5% and the sparse-array resolution below 10%. If the reconstruction degrades significantly, the quoted resolutions are tuned to one simulation library rather than to the physics.","supporting_citations":[{"cited_title":"A Rotationally Symmetric Lateral Distribution Function for Radio Emission from Inclined Air Showers","cited_arxiv_id":"1808.00729","evidence_quote":"Provides the early-late correction used to symmetrise the inclined-shower emission pattern."},{"cited_title":"Glaser, A","cited_arxiv_id":null,"evidence_quote":"radiotools package supplies the atmosphere models and the theoretical Cherenkov-ring radius prediction used in the LDF parametrisation."}],"review_version":1}