{"id":"cfed2b7d-8c03-4512-82d9-d63c98cdba9a","arxiv_id":"2507.21277","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"Sparse local Galactic magnetic field measurements can backtrack 10^20 eV cosmic rays to M82 when the field is near 1 µG, but not when the field is ten times stronger.","lead":"This paper simulates whether sparse measurements of the Milky Way's magnetic field in interstellar clouds could still be used to trace ultra-high-energy cosmic rays back to their source. Using mock data, the authors find that very sparse sampling works when the magnetic field is weak, but fails when the field is about ten times stronger than current models predict.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Quantitative sparsity thresholds rest on an unrealistically smooth ground truth: the simulation uses only JF12's ordered component, omitting small-scale/turbulent structure that sparse measurements cannot resolve, so the 'satisfactory at ~1600 pc' claim is untested for real fields.","rationale":"The reader's weakest assumption and my load-bearing concern match: the simulation's ground truth is a smooth JF12 ordered field without small-scale/turbulent structure. This is the most central assumption because the paper's quantitative claims concern the degree to which sparsity degrades backtracking; if the real field has unresolved small-scale structure, the reported degradation is underestimated. The concern is supported internally: Section 1 explicitly mentions that events are dispersed by the turbulent GMF component, yet the simulation uses only the ordered component (Section 2) and the mock-event scatter (Section 4) accounts only for experimental and intergalactic deflections, not Galactic turbulence. Other concerns (lack of error bars, unspecified integration step, abstract overstatement for f=10) are secondary and addressable without invalidating the qualitative result. The simulation is otherwise well-designed: it includes control runs, uses nearest-neighbor sampling as a deliberately simple reconstruction, and provides a useful diagnostic (Figure 4) connecting deflection to inverse energy. The proposed test, adding a turbulent component to the ground truth, would directly settle whether the quantitative sparsity thresholds are robust for realistic fields; if the thresholds shift, the paper's central feasibility claim would need to be rephrased as an upper bound on achievable accuracy rather than a direct statement about real Galactic measurements.","tokens_in":1500,"tokens_out":879,"duration_ms":133308,"concrete_test":"Repeat the full workflow with a ground truth that augments the JF12 ordered field (f=1 and f=10) with Gaussian-random turbulent realizations (coherence length ~100 pc, B_rms ~1-3 microgauss, e.g. from JF12's random component or a standard Kolmogorov spectrum). Recompute the mean and standard deviation of angular distance vs slice length L (Figures 5-6). If the f=1 curve at L=1600 pc rises by more than ~2 degrees above the no-turbulence control, or if the f=10 degradation begins at L<400 pc, the quantitative sparsity thresholds in the abstract do not hold for realistic fields.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central quantitative claims (abstract, Section 5) are that for ~1 microgauss ordered fields, backtracking with measurement spacing L~1600 pc is 'satisfactory', while f=10 breaks down at L~400 pc. These thresholds are obtained by backtracking through a ground truth that is the smooth, analytic JF12 ordered field only, rescaled by f (Section 2). No small-scale ordered structure or turbulent/random component is present in the ground truth; the mock-event generation (Section 4) injects only a 3-degree source scatter (to mimic experimental and intergalactic deflections) rather than Galactic turbulent deflections. In the real Galaxy, the GMF contains a random/turbulent component with coherence scales of order 10-100 pc and strengths comparable to the ordered field. For protons at 40-100 EeV traversing several kpc toward M82, typical estimates of turbulent deflections are several degrees, comparable to or larger than the reported mean angular distances (Figure 5, control ~2 degrees). Because sparse measurements at L>=400 pc cannot sample the turbulent field, the backtracking would be unable to correct these deflections; the final angular errors would include a turbulence floor and additional aliasing from unresolved small-scale ordered structure. The reported mean angular distances are therefore optimistic lower bounds, and the specific thresholds (1600 pc / 400 pc) are not validated against realistic field structure. The paper itself notes (Section 6) that results are lower limits without sophisticated reconstruction, but the abstract states the thresholds as feasibility statements.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper simulates UHECR backtracking through sparse local measurements of the Galactic magnetic field, using the JF12 ordered-field model as ground truth with its strength rescaled by a factor f = 1, 3, or 10. Mock proton events are generated with energies drawn from a shifted exponential fit to 74 Telescope Array hotspot events and with arrival directions that backtrack to within 3 degrees of M82 under the exact ground-truth field. The authors then backtrack these events through a cubic grid of field measurements of spacing L, using the nearest-neighbor value and adding 0%, 25%, or 50% Gaussian errors on field strength. The main results, shown in Figures 5 and 6, are that mean angular distance from M82 remains small for f=1 and f=3 even at L ~ 1 kpc, while f=10 degrades sharply for L of a few hundred parsecs. The authors conclude that sparse local measurements are not automatically a deal-breaker for backtracking in low-field directions, and that their results should be read as a lower limit on achievable quality.","tokens_in":10782,"tokens_out":2857,"duration_ms":37373,"significance":"If the quantitative thresholds in the abstract are correct, the paper provides a useful, concrete motivation for local GMF surveys such as PASIPHAE and SOUTH POL: even very sparse sampling could support UHECR source localization in regions of modest ordered-field strength, while strong-field regions require denser sampling. The study has several genuine strengths: the nearest-neighbor sampling and the explicit refusal to interpolate make the reported accuracies a conservative lower limit relative to any reconstruction method; the control run at L=1 pc cleanly isolates the residual 3-degree source scatter; and the energy distribution is taken from external Telescope Array data rather than fitted to the simulation output. The closed-loop design is a standard and appropriate test of reconstruction accuracy, not a circular derivation. The significance is, however, limited by the smoothness of the ground truth, as discussed below: the paper does not actually demonstrate feasibility in a field containing the small-scale and turbulent structure known to exist in the real interstellar medium.","major_comments":[{"comment":"The quantitative thresholds in the abstract (satisfactory at L ~ 1600 pc for 1 microgauss field; breakdown at L ~ 400 pc for f=10) are derived from a ground truth that contains only the smooth, analytic JF12 ordered component, rescaled by f. The mock event generation in Section 4 adds only a 3-degree scatter meant to represent experimental and intergalactic deflections; no turbulent or small-scale ordered Galactic field is present. In the real Galaxy, the random component of the GMF has coherence scales of order 10-100 pc and strength comparable to the ordered component, and it would be completely unresolved by measurements spaced 400-1600 pc apart. The reported mean angular distances are therefore optimistic lower bounds, and the specific 1600 pc / 400 pc thresholds are not validated against realistic field structure. I request either an explicit turbulence-inclusive simulation (e.g., adding a random component with 10-100 pc coherence to the ground truth) or a substantial rewording of the abstract and conclusions so that the quantitative thresholds are presented as conditional on a smooth ordered field only.","section":"Section 2, Section 4, Section 5"},{"comment":"The central claims of 'satisfactory' backtracking and 'breakdown' are not quantitatively defined, and no statistical uncertainties accompany the Monte Carlo averages. Figure 5 reports the mean angular distance for N=1000 events, but there are no error bars, confidence intervals, or a criterion such as 'the fraction of events within X degrees of M82'. Without such a criterion, the abstract statement that L ~ 1600 pc yields 'satisfactory' results cannot be independently evaluated, and it is unclear whether the difference between f=1 and f=3 curves is significant given finite sample noise. Please define quantitative success metrics (e.g., median angular error, fraction of events within 3 or 5 degrees, barycenter offset) and provide corresponding uncertainties, at least for the headline configurations.","section":"Section 5, Figures 5 and 6"},{"comment":"The discussion in Section 6 states that sparsity has 'only a modest result' on accuracy for f=1 and f=3, but this is based on visual inspection of Figure 5 rather than any formal comparison. The mean angular distance at L ~ 1600 pc appears to be several times larger than the control value of about 2 degrees, even for f=1 with 25% error. Whether a degradation from ~2 degrees to, say, ~4-5 degrees is 'satisfactory' depends entirely on the scientific goal (source identification versus low-energy counterpart localization), which is never specified. Please tie the success criterion to the intended downstream use and report results against that criterion.","section":"Section 5, Section 6"}],"minor_comments":[{"comment":"The abstract says 'charge-particle astronomy'; this should be 'charged-particle astronomy'.","section":"Abstract"},{"comment":"The text contains the typo 'sparce' in the paragraph after Figure 5; it should be 'sparse'.","section":"Section 5"},{"comment":"The phrase 'Glactic coordinates' should be 'Galactic coordinates'.","section":"Equation (7)"},{"comment":"The assumption that all three magnetic-field components are sampled is stated clearly, but the two justifications given are qualitative. Since the claim that line-of-sight sensitivity is small is central to this assumption, a quantitative estimate (or a reference to the explicit demonstration in Tsouros et al. 2024b) would strengthen the argument.","section":"Section 2"},{"comment":"The figure captions do not state the number of Monte Carlo events or the definition of the plotted quantities beyond the axis labels; adding N=1000 and a reference to Equations (9) and (10) in the captions would improve clarity.","section":"Figures 5 and 6"}],"recommendation":"major_revision","confidential_remarks":"The paper is a clean, well-scoped simulation study, and the nearest-neighbor design is a sensible conservative choice. My main concern is scope: the central quantitative claims are presented without the caveat that the ground truth contains no turbulent or small-scale structure, and the success criteria are not defined. These issues are fixable within the manuscript's scope by adding a turbulence-inclusive experiment or substantially qualifying the abstract, and by defining and applying explicit success metrics. The paper is likely to be of interest to the astroparticle and ISM communities, though its novelty relative to the authors' own related work (Tsouros et al. 2024a,b,c) should be clarified in the introduction."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"You should know two things about this paper before reading it. First, it gives a genuinely new quantitative map: backtracking through a nearest-neighbor-sampled GMF keeps mean angular errors near the ~2-degree floor even at 1600 pc spacing when the ordered field is ~1 µG, but at 10× that strength errors climb to ~10 degrees already at 400 pc. That is exactly the kind of number the PASIPHAE and SOUTH POL communities need when planning survey strategies. Second, the thresholds are computed in a smooth, ordered-only GMF. There is no turbulent component in the ground truth, so the results are optimistic for the real Galaxy, not feasibility statements.\n\nWhat the paper does well: it isolates sparsity from reconstruction by deliberately avoiding any interpolation and using the nearest measurement at each step. That makes the results a conservative lower bound on what a simple method can achieve. The mock-event generation is careful—energies from a shifted-exponential fit to the Telescope Array hotspot events, events selected only if they backtrack within 3 degrees of M82, 1000 events per field-strength case, and a control run at L=1 pc. The scan over f=1,3,10 and 25/50% strength errors is sensible, and the qualitative conclusion that ordered-field strength matters more than sparsity is likely robust.\n\nThe soft spots are real but not fatal. The biggest is the missing turbulence. The real GMF has a random component with coherence scales of tens of parsecs and strength comparable to the ordered field. Sparse measurements at L≥400 pc cannot sample that structure, and the backtracking code therefore cannot correct turbulent deflections. For protons at 40–100 EeV crossing several kpc, turbulent deflections are typically several degrees—more than the reported mean errors. So the reported angular distances are lower bounds, and the \"satisfactory at 1600 pc\" claim is untested for realistic fields. The paper does note in Section 6 that results are lower limits without sophisticated reconstruction, but the abstract states the thresholds as feasibility statements, which oversells it. Also, \"satisfactory\" and \"breaks down\" are never given quantitative definitions, the Monte Carlo averages are shown without error bars (minor with 1000 events, but still), and the abstract's \"breaks down\" at f=10 is softer in the discussion, which says \"more challenging.\" The all-three-components-sampled assumption is stated and defended, so that one is fine.\n\nWho it is for: anyone planning sparse local GMF surveys or interpreting their data for UHECR backtracking. The qualitative message—that sparsity is not automatically a deal-breaker if the ordered field is modest—is valuable. The specific thresholds need a follow-up with a turbulent field before being used for design decisions. It deserves a serious referee, but a referee who will push for a turbulence-inclusive version or a clearly framed \"ordered-field-only\" scope. I would read the revised version, but I would not cite these thresholds as physical limits.","headline":"A clean closed-loop simulation showing that sparsity tolerance for UHECR backtracking depends strongly on ordered-field strength, but the headline thresholds are optimistic because the ground truth omits turbulence.","tokens_in":11332,"tokens_out":3263,"would_cite":false,"duration_ms":43426,"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":"Sparse local measurements of the Galactic magnetic field can backtrack ultra-high-energy cosmic rays to a suspected source, until the ordered field is roughly ten times stronger than models predict.","keywords":["ultra-high-energy cosmic rays","Galactic magnetic field","cosmic-ray backtracking","charged-particle astronomy","JF12 model","M82","magnetic-field sparsity","Telescope Array hotspot"],"falsifier":"Measure the local magnetic field along the M82 line of sight with mean spacing around 1.6 kpc, backtrack the actual Telescope Array hotspot events through that measured field, and compare the mean angular distance and spread to the near-2-degree level predicted for $f=1$; if residual structure on scales below 1.6 kpc produces larger deflections, or if the measured field strength is closer to $f=10$, the predicted tolerance to sparsity fails.","tokens_in":10278,"feed_emoji":"🔭","tokens_out":9991,"duration_ms":102975,"temperature":0.7,"pith_summary":"The paper asks whether the sparse, cloud-by-cloud magnetic-field measurements now becoming available can actually be used to backtrack ultra-high-energy cosmic rays (UHECRs) to their source. Using the JF12 model as a stand-in for the true Galactic magnetic field and mock proton events aimed within 3 degrees of M82, it finds that when the ordered field has its nominal $\\sim$1 $\\mu$G average strength, backtracking stays accurate even with measurements spaced about 1600 pc apart. But when the ordered field is rescaled upward by a factor of 10, accuracy breaks down already at about 400 pc spacing. The authors conclude that sparsity is not automatically fatal for charged-particle astronomy, while high-strength field regions remain a serious obstacle, and they frame their no-interpolation results as a lower limit on what reconstruction techniques could achieve.","feed_headline":"Sparse magnetic data can backtrack 10^20 eV cosmic rays","feed_subtitle":"Simulation shows 1.6-kpc-spaced measurements trace M82 events, but a 10x stronger field breaks them apart.","key_machinery":"The argument is carried by a mock-observation pipeline. The ground-truth field is the JF12 model with its ordered component multiplied by a factor $f$; the region toward M82 is divided into cubes of side length $L$, each cube assigned a constant field equal to the model value at its center, with Gaussian noise added to the field strength at the 25% or 50% level. Backtracking integrates the discretized Lorentz-force equation $\\hat{v}_{\\mathrm{prev}} = \\hat{v}_{\\mathrm{now}} - (Z e c^2 / E)(\\hat{v} \\times B)\\,\\delta t$, using at every step the field of the nearest cube. The control parameter is the slice length $L$—the spacing between measurements—and the rescalings $f=1,3,10$ set the field-strength regime; the output metrics are the mean and standard deviation of the angular distance of the 1000 backtracked events from M82.","core_discovery":"The central claim is that the feasibility of UHECR backtracking through sparsely sampled local GMF measurements hinges more on the strength of the ordered Galactic field in the direction of the source than on the density of measurements. For mock proton events at $10^{20}$ eV traced from M82, a uniform grid of magnetic-field measurements with linear spacing up to about 1600 pc yields mean backtracked angular distances from M82 near the control-run level of about 2 degrees when the ordered component of the JF12 field is at its nominal strength (scaling factor $f=1$). Raising the ordered field by a factor of 10 ($f=10$) makes the mean angular distance grow to about 10 degrees once the spacing reaches a few hundred parsecs, with a comparable spread between events; $f=3$ gives intermediate behavior. Because the backtracking uses no interpolation—each step simply adopts the nearest measurement—the paper presents these numbers as a lower bound on achievable accuracy, with reconstruction methods expected to improve on them.","pith_inferences":["The paper's $f=1$ versus $f=10$ asymmetry suggests a staged strategy that the authors only gesture at: first use sparse local measurements to determine which field-strength regime the Galaxy is in, then decide whether to invest in dense, high-accuracy sampling; that explicit decision rule is our inference.","The relevant tolerance parameter is the path length a particle spends in the Galactic field, so sources at higher Galactic latitude or with shorter in-galaxy path lengths should tolerate sparser sampling than M82; testing this by repeating the experiment for different sky directions is a natural extension.","Because the mock data contain no small-scale ordered structure beyond JF12's parameterization, injecting coherent field patches below the grid scale $L$ would quantify how much tolerable spacing shrinks when the real field is clumpy—a testable extension the paper does not run.","The thresholds are derived for protons; if the M82 events include heavier nuclei, their lower rigidity at fixed energy means the same backtracking accuracy would require denser measurements, so composition uncertainty should widen the quoted error bars."],"forward_implications":["Sparse local GMF measurements with about 1.6 kpc spacing are enough to backtrack $10^{20}$ eV protons to M82 with mean angular errors near the 2-degree control level when the ordered field is at JF12 strength, so sparsity alone does not rule out charged-particle astronomy.","If the ordered field toward M82 is ten times stronger than JF12 predicts, mean backtracked angles reach about 10 degrees at few-hundred-parsec spacings, so source localization in strong-field directions will need dense, accurate measurements.","Measurement error in the field strength is tolerable (up to 50%) when the field is weak, but error tolerance shrinks sharply at high field strength, so the observational priority shifts from quantity to accuracy in strong-field regions.","These results are obtained with no interpolation or field reconstruction—just nearest-neighbor assignment—so they set a lower limit; reconstruction techniques should push the achievable accuracy beyond what is reported.","A successful backtracking erases the linear correlation between the angular distance from M82 and $E^{-1}$ seen in the mock arrival directions, providing a data-driven check of whether the magnetic field model used for backtracking is adequate."],"supporting_citations":[{"why":"Provides the JF12 Galactic magnetic field model used as the ground truth for both generating mock events and sampling sparse measurements.","marker":"Jansson & Farrar 2012"},{"why":"Supplies the 74 Telescope Array events attributed to M82 whose fitted energy distribution seeds the mock event set, and the M82-source hypothesis.","marker":"Abbasi et al. 2020"},{"why":"Motivates the f=3 and f=10 field-strength rescalings by indicating that GMF models may underestimate the field toward M82.","marker":"Tritsis et al. 2019"},{"why":"Provides the dispersion-based method by which local field-strength uncertainties of 25–50% are calibrated for the mock measurements.","marker":"Skalidis & Tassis 2021"},{"why":"Describes the PASIPHAE optopolarimetric survey that motivates the premise that abundant local cloud measurements are on the way.","marker":"Tassis et al. 2018"},{"why":"Shows that approximate line-of-sight field information suffices to supplement plane-of-sky measurements, supporting the paper's assumption that all three field components are sampled.","marker":"Tsouros et al. 2024b"},{"why":"Demonstrates that sophisticated reconstruction improves backtracking, framing the paper's nearest-neighbor results as a lower limit.","marker":"Tsouros et al. 2024c"}],"fun_headline_variants":["1600-pc magnetic gaps still trace 10^20 eV rays","Field strength, not gap size, decides cosmic-ray backtracking","UHECR backtracking survives sparse fields unless 10x stronger","Where cosmic rays track: sparse data work, strong fields don't","Magnetic sparsity OK for 10^20 eV unless field is 10x strong"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The simulation assumes the smooth JF12 model is the true Galactic magnetic field, so the only field structure it contains is JF12's large-scale ordered component; if the real field toward M82 has strong small-scale structure, sparse measurements spaced about 1600 pc apart would miss it, and backtracking errors could be larger than the paper reports.","fun_headline_variants_meta":{"raw":{"variants":["1600-pc magnetic gaps still trace 10^20 eV rays","Field strength, not gap size, decides cosmic-ray backtracking","UHECR backtracking survives sparse fields unless 10x stronger","Where cosmic rays track: sparse data work, strong fields don't","Magnetic sparsity OK for 10^20 eV unless field is 10x strong"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000624,"raw_usage":{"total_tokens":2978,"prompt_tokens":1122,"completion_tokens":1856,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":738,"completion_tokens_details":{"reasoning_tokens":1760}},"tokens_in":738,"tokens_out":1856,"duration_ms":15061,"temperature":1.0,"reasoning_tokens":1760,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T12:56:08.506507+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Measure the local magnetic field along the M82 line of sight with mean spacing around 1.6 kpc, backtrack the actual Telescope Array hotspot events through that measured field, and compare the mean angular distance and spread to the near-2-degree level predicted for $f=1$; if residual structure on scales below 1.6 kpc produces larger deflections, or if the measured field strength is closer to $f=10$, the predicted tolerance to sparsity fails.","supporting_citations":[{"cited_title":"& Farrar, G","cited_arxiv_id":null,"evidence_quote":"Provides the JF12 Galactic magnetic field model used as the ground truth for both generating mock events and sampling sparse measurements."},{"cited_title":"U., Abe, M., Abu-Zayyad, T., et al","cited_arxiv_id":null,"evidence_quote":"Supplies the 74 Telescope Array events attributed to M82 whose fitted energy distribution seeds the mock event set, and the M82-source hypothesis."},{"cited_title":"2019, ApJ, 873, 38","cited_arxiv_id":null,"evidence_quote":"Motivates the f=3 and f=10 field-strength rescalings by indicating that GMF models may underestimate the field toward M82."}],"review_version":1}