{"id":"1a1f99b7-93fe-4fef-9097-c220ec682ea6","arxiv_id":"2602.00691","paper_version":2,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":3,"one_line_summary":"A hybrid SWGO-like water-Cherenkov array with SST-1M Cherenkov telescopes is simulated to improve gamma/hadron separation and flux sensitivity above 10 TeV by ~60%/30%.","lead":"This paper simulates adding two SST-1M Cherenkov telescopes into a water-Cherenkov detector array like SWGO, and reports that the hybrid setup improves gamma-ray flux sensitivity above 10 TeV by about 60% (monocular) and 30% (stereoscopic). A generalist should read it because it quantifies whether combining two established ground-based gamma-ray techniques is worth the engineering cost.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Load-bearing risk: idealized uniform 12.5% FF WCD array and simplified detector response may inflate the claimed 60%/30% sensitivity improvement; a realistic SWGO layout/response simulation is needed.","rationale":"The paper's central claim is internally consistent: the Monte Carlo pipeline is standard, the RF classifiers are tested on an independent sample, and the sensitivities follow the usual prescription. In that sense the reader's CONDITIONAL verdict is appropriate. The single most load-bearing premise is that the simplified WCD framework used to define LCm and P_tail reproduces the real SWGO detector response well enough that these proxies retain their separation power in the hybrid geometry. This premise is fragile because the array is idealized (uniform 12.5% FF, deliberately large) and the detector response is parametrized; the authors acknowledge that the real array will differ but do not quantify the effect. The concern is not internal inconsistency but an unsupported extrapolation from simulation to the proposed real observatory. The concrete test proposed—rerunning with the SWGO reference layout and a fuller response—would directly quantify that extrapolation. No other critique appears as decisive: the energy reconstruction propagates realistic resolution, the independent test sample limits classifier overfitting, and the use of a true-muon oracle is explicitly illustrative. Therefore the reader's weakest assumption matches ours, and the verdict should remain CONDITIONAL until the robustness check is performed.","tokens_in":11736,"tokens_out":7234,"duration_ms":84815,"concrete_test":"Repeat the full Monte Carlo analysis using the current SWGO reference array design (e.g., the graded-fill layout and detector response model from Ref. [6]) while keeping the identical SST-1M simulation, reconstruction pipeline, and sensitivity calculation. Compare the ROC AUC of LCm+P_tail and the resulting flux sensitivity above 10 TeV. If the monocular improvement drops below ~30% and the stereo below ~15%, the headline 60%/30% numbers are not robust to the WCD array idealization.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim rests on the WCD-derived features LCm and P_tail contributing substantial gamma/hadron discrimination beyond the SST-1M alone. In Sec. 2, the WCD array is modeled with a simplified parametrized detector response (Ref. [12]) and a deliberately optimistic uniform grid of 9997 tanks at 12.5% fill factor over 1 km^2. The authors explicitly state this FF is chosen to minimize border effects and that a real array will use an optimized size and possibly a graded fill factor, which is 'beyond the scope' of this paper. LCm and P_tail are characterized in this same simplified framework (Refs. [12,16,18]) and are not cross-checked against a full detector response simulation or real data. If the actual SWGO array is sparser or has non-uniform coverage, or if the parametrized detector response overestimates the muon-pulse information available from the photomultiplier signals, the WCD features will separate less cleanly. A substantial part of the claimed ~60% (monocular) and ~30% (stereo) sensitivity improvement above 10 TeV could then be an artifact of the idealized setup. The paper provides no systematic uncertainty on this load-bearing assumption.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper studies a hybrid observatory concept in which two SST-1M imaging atmospheric Cherenkov telescopes are embedded in a SWGO-like water-Cherenkov detector (WCD) array at 4700 m altitude. Using CORSIKA/sim_telarray for the IACT response and a simplified parametrized WCD framework from Ref. [12], the authors add two WCD-based gamma/hadron discriminators (LCm and P_tail) to the Random-Forest gammaness classification in sst1mpipe. They report an improvement of about 60% (monocular) and 30% (stereoscopic) in flux sensitivity above 10 TeV for a point-like source at 20° zenith, benchmarked against the true muon number as an idealized reference. The paper also discusses cross-calibration, fast follow-up, and technical operational aspects of operating SST-1M at the SWGO site.","tokens_in":12101,"tokens_out":6626,"duration_ms":81455,"significance":"If the reported gain is robust, the result provides a concrete quantitative case for embedding SST-1M-type telescopes inside SWGO as a cost-effective sensitivity upgrade above 10 TeV. The IACT simulation chain is standard and detailed, the analysis uses an independent test MC sample, and the use of a true-muon oracle is a useful diagnostic. The main strength is that the claimed gain is a falsifiable prediction derived from a concrete simulation pipeline. However, the central quantitative claim rests on an idealized WCD array and a simplified detector response that are not validated against a full detector simulation or real data, so the headline numbers should currently be regarded as an upper-limit-style estimate until robustness is demonstrated.","major_comments":[{"comment":"The load-bearing assumption is the idealized WCD array: 9997 tanks of 12.6 m^2 on a uniform triangular grid with 12.5% fill factor, 'deliberately chosen to minimize border effects', while a realistic optimized/graded array is explicitly declared 'beyond the scope of the present paper'. The WCD signals are computed with the simplified parametrized framework of Ref. [12], and the same framework is used to define LCm and P_tail. No full detector response simulation or real-data cross-check is provided. Since the claimed 60%/30% sensitivity gain is driven by these WCD observables, the idealized geometry and parameterization could inflate the separation power. I ask the authors to add a robustness study, e.g. repeating the analysis with a realistic SWGO layout (lower or graded fill factor) and/or with a full WCD response simulation, or at least quantifying how the sensitivity gain varies with","section":"Section 2"},{"comment":"The headline claim 'improvement ... by about 60% for monocular and 30% for stereoscopic ... above 10 TeV' is never defined quantitatively. Are these ratios of energy-integrated flux sensitivities above 10 TeV, or ratios at a particular energy? Figures 3 and 6 show energy-dependent improvement, so a single number requires an explicit integration range and definition. In addition, the abstract says the improvement comes from 'additional parameters from the WCD array', while Section 7 attributes it to 'the LCm parameter' alone, and the figures show separate curves for LCm, P_tail, and LCm+P_tail. The manuscript must state which configuration yields the quoted 60%/30% numbers. This is not a cosmetic issue because the abstract's central quantitative claim is otherwise untestable.","section":"Section 4 / abstract and Section 7"},{"comment":"The sensitivity curves are presented without statistical uncertainties. The high-energy bins (E > 100 TeV) contain very few Monte Carlo events after cuts, and the RF-based gammaness and theta-squared cuts are optimized on finite samples. The ROC AUC values are likewise point estimates. Without at least bootstrap or Poisson uncertainties on the sensitivity ratios, the precision implied by the abstract ('about 60%', 'about 30%') is not justified. Please add error bands or confidence intervals to the sensitivity curves, or explicitly state that the quoted percentages are central values from a single MC realization.","section":"Section 4, Figs. 3/6"}],"minor_comments":[{"comment":"The label 'N/uni03BC' appears as a literal string; it should read N_mu (N_μ). The same issue affects the legend entries.","section":"Figures 3, 5, 6"},{"comment":"The variable is written as P^alpha_tail in Section 3 and as P_tail in most figures and text. Please use one consistent notation throughout, including in the abstract and summary.","section":"Notation"},{"comment":"The true muon number N_mu is an oracle variable and is correctly labeled 'not usable in real data'. However, in Figs. 1, 3, 4, and 6 it is plotted on the same footing as the physical observables, with near-perfect AUC values. The authors should explicitly state in the figure captions and text that the N_mu curves are idealized upper bounds, not achievable performance, to prevent overinterpretation.","section":"Section 3"},{"comment":"The text states 'sst1mpipe v0.7.4' but the Zenodo citation [14] lists version v0.7.3. Please harmonize the version number and date.","section":"Section 2 / Ref. [14]"},{"comment":"The technical discussion of the SST-1M infrastructure (power consumption, data volume, cooling, wind limits) is useful context but is not connected to the simulation results. Consider moving it to an appendix or tightening it to the aspects that affect the hybrid sensitivity claim.","section":"Section 6"}],"recommendation":"major_revision","confidential_remarks":"The stress-test concern about the idealized WCD array is the decisive issue; it is load-bearing for the quantitative claim but appears fixable in revision. The IACT simulation and analysis are sound, the code provenance is clear, and the paper is honest about many limitations. I do not see grounds for rejection, but the abstract's 60%/30% numbers should not be published until the WCD-array model is either validated or its sensitivity gain is shown to be robust under realistic array configurations and signal-fluctuation assumptions. Please also ensure the reported improvement is defined and attributed to a specific configuration (LCm alone vs. LCm+P_tail)."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Two things to know. First, this is a careful simulation study that gives the first quantitative sensitivity projection for SST-1M telescopes embedded in a SWGO-like water-Cherenkov array. The ~60%/30% improvement above 10 TeV is a real, testable prediction from their MC chain. Second, the number is only as good as the assumed WCD array, and that array is ideal: 9997 tanks on a uniform grid at 12.5% fill factor, with a parametrized detector response. The authors say the real array will be optimized and graded, but they don't test that.\n\nWhat they do well: the pipeline is standard CORSIKA + sim_telarray + sst1mpipe, and they take care to use an independent test sample, propagate reconstructed energy into the WCD parameter computation, and include a true-muon oracle to show how close LCm and P_tail get to the ideal. That is honest and reproducible. The LCm variable comes from their own prior work, but the gain here is measured in simulation, not derived from the definition, so the self-citation is not circular.\n\nThe soft spots are in the detector model. The WCD response is a simplified framework from Ref. [12] with no cross-check against a full detector simulation or real data. The uniform 12.5% FF is chosen to avoid border effects and is acknowledged to be optimistic. If the real SWGO layout is sparser or has a graded fill factor, the discrimination power of LCm and P_tail could drop, and the 60%/30% would shrink. The paper gives no systematic uncertainty on this. That said, the authors are upfront about the simplification, and for a feasibility projection this is acceptable as long as it is read as an upper bound. I would like to see at least a test of sensitivity to FF or a realistic layout before anyone uses these numbers for an observatory design.\n\nWho is this for? The VHE instrumentation community, specifically people planning SWGO or thinking about hybrid IACT+WCD arrays. It deserves a serious referee. The methodology is clear, the result is novel, and the weak point is identifiable and fixable.","headline":"Careful MC study that gives the first quantitative look at SST-1M inside a SWGO-like array; the 60%/30% gain above 10 TeV is plausible but rests on an idealized WCD layout and a simplified detector response.","tokens_in":12994,"tokens_out":2237,"would_cite":true,"duration_ms":27040,"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":"Simulations show that placing Cherenkov telescopes inside a water-Cherenkov array improves gamma-ray sensitivity above 10 TeV by up to 60%.","keywords":["gamma-ray astronomy","imaging atmospheric Cherenkov telescope","water-Cherenkov detector","hybrid observatory","gamma/hadron separation","flux sensitivity","monte carlo simulation","SST-1M"],"falsifier":"Place an SST-1M telescope inside an operating water-Cherenkov array (or run a full detector simulation of the WCDs) and measure the gamma/hadron separation above 10 TeV on proton and gamma-ray showers; if the actual background rejection at fixed gamma efficiency falls short of the simulated area-under-curve of about 0.997, the quoted sensitivity gain of 60% would not be realized.","tokens_in":11689,"feed_emoji":"🔭","tokens_out":3763,"duration_ms":38608,"temperature":0.7,"pith_summary":"The paper argues that placing small Cherenkov telescopes inside a high-altitude array of water-Cherenkov detectors turns the ground array's sensitivity to muons into a powerful gamma/hadron separator for the telescopes. Using detailed Monte Carlo simulations, it shows that adding two ground-array parameters—the azimuthal fluctuation of the ground signal (LCm) and a tail parameter correlated with muon count (P_tail)—improves the telescopes' flux sensitivity above 10 TeV by about 60% in monocular mode and 30% in stereoscopic mode. If true, hybrid observatories of this kind could detect faint gamma-ray sources in the multi-TeV to PeV range that are inaccessible to either technique alone. The gain comes from rejecting the hadronic background while retaining nearly all gamma-ray events.","feed_headline":"Cherenkov telescopes gain 60% sensitivity inside water-tank array","feed_subtitle":"Simulated ground-array muon measurements sharpen gamma/hadron separation above 10 TeV","key_machinery":"The two ground-array parameters LCm and P_tail. LCm quantifies how much the lateral signal distribution of the air shower fluctuates around the azimuthal direction; gamma-ray showers are smooth while hadronic showers are clumpy. P_tail measures the fraction of detectors whose signal exceeds the average at comparable core distance, which tracks the total muon number. In the analysis, these are added as features to a machine-learning gamma/hadron classifier (the 'gammaness' parameter), allowing the telescope's own imaging variables to be supplemented by muon-sensitive information. The true muon count is used in the same classifier as an upper-bound reference.","core_discovery":"The central claim is that a hybrid observatory—SST-1M imaging Cherenkov telescopes operated inside a dense water-Cherenkov array at 4700 m altitude—achieves significantly better gamma-ray flux sensitivity than the telescopes alone, because the ground array measures the muon content of air showers. For showers above 10 TeV, the parameters LCm (azimuthal fluctuations of the ground signal) and P_tail (a measure of detectors with signals far above the local average) are fed into the telescope's gamma/hadron classifier. In the simulations, LCm alone nearly matches the performance of the true (unmeasurable) muon count, and even surpasses it at the highest energies. The monocular sensitivity improv","pith_inferences":["The quoted gains rely on a simplified, parameterized model of the water-Cherenkov detectors; a full detector simulation or real prototype data could shrink or enlarge the gains, especially if the real array is sparser than the simulated 12.5% fill factor.","The improvement is demonstrated only for events with reconstructed energy above 10 TeV; adapting muon-sensitive variables to lower energies, as suggested in the paper, might extend the gain down to about 1 TeV, but that is not shown here.","The same hybrid logic could be tested today by placing a small IACT near an existing water-Cherenkov array and measuring the gamma/hadron rejection on a known source such as the Crab Nebula.","The cost of the gain is operational overhead: the telescope can only observe on clear nights, so the improved sensitivity applies to a fraction of the duty cycle unless bright-moon operations are validated."],"forward_implications":["Point-like gamma-ray sources above 10 TeV become detectable at roughly half the flux required by an SST-1M telescope alone in monocular mode.","The hybrid concept turns a wide-field survey array into a precision instrument: the same event is simultaneously seen by both detectors, enabling cross-calibration of energy scale and angular resolution.","Because LCm performs as well as or better than the true muon count at ultra-high energies, the ground-array-only discrimination could be exploited in standalone arrays at PeV energies.","Stereoscopic IACT operation leaves less room for improvement, but still gains about 30% sensitivity, so the benefit is largest when only one telescope is available.","Follow-up observations of transient sources can start within seconds, since the same site hosts both instruments."],"fun_headline_variants":["Simulated water tank array boosts Cherenkov gamma sensitivity 60%","Ground muon data improve gamma sensitivity 60%","Hybrid observatory lifts gamma sensitivity 60%","Water-tank muon readings sharpen gamma detection 60%","Gamma sensitivity up 60% with Cherenkov plus water tanks"],"cache_read_input_tokens":2304,"weakest_assumption_plain":"The load-bearing premise is that the parameterized water-Cherenkov signal model reproduces the real detector response closely enough that LCm and P_tail keep their separation power in the hybrid geometry; the paper simulates an idealized dense array and contains no real-data validation.","fun_headline_variants_meta":{"raw":{"variants":["Simulated water tank array boosts Cherenkov gamma sensitivity 60%","Ground muon data improve gamma sensitivity 60%","Hybrid observatory lifts gamma sensitivity 60%","Water-tank muon readings sharpen gamma detection 60%","Gamma sensitivity up 60% with Cherenkov plus water tanks"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.002163,"raw_usage":{"total_tokens":8183,"prompt_tokens":668,"completion_tokens":7515,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":412,"completion_tokens_details":{"reasoning_tokens":7430}},"tokens_in":412,"tokens_out":7515,"duration_ms":58311,"temperature":1.0,"reasoning_tokens":7430,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-03T05:54:40.849950+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Place an SST-1M telescope inside an operating water-Cherenkov array (or run a full detector simulation of the WCDs) and measure the gamma/hadron separation above 10 TeV on proton and gamma-ray showers; if the actual background rejection at fixed gamma efficiency falls short of the simulated area-under-curve of about 0.997, the quoted sensitivity gain of 60% would not be realized.","supporting_citations":[],"review_version":1}