{"id":"5743b5e7-cb91-409e-88ec-6d6727fc8b0e","arxiv_id":"1908.10148","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"IceCube observations from 2010 to 2017 show the TeV cosmic-ray Sun shadow weakens as sunspot number rises, favoring solar magnetic field model predictions over a simple disk.","lead":"Using seven years of IceCube data, this analysis finds that the cosmic-ray shadow cast by the Sun changes over time and tracks the 11-year solar activity cycle. It also shows that solar magnetic field models describe the shadow better than a simple opaque disk, though tensions remain.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Off-source windows may be Sun-magnetically modulated at 50–60 TeV, biasing the sunspot-number trend.","rationale":"The reader identified the off-source background estimate as the weakest premise, and I agree that it is load-bearing. My concern sharpens the reader’s general worry about anisotropy/acceptance into a specific physical mechanism: at the median energy of 50–60 TeV, the solar magnetic field’s influence can extend to the off-source windows, making the background itself solar-activity dependent. This is more dangerous than a static detector asymmetry because it could directly mimic the claimed sunspot-number correlation. The Moon shadow is an excellent stability check, but it cannot validate the Sun-specific off-source windows because the Moon has no magnetic field. The analysis as presented does not demonstrate that the off-source windows are unaffected, and the quoted significances are purely statistical. This does not force rejection: the observed trend is large, and the same simulation machinery used for the model predictions could be repurposed as the decisive test. Therefore the CONDITIONAL verdict stands, pending this specific check.","tokens_in":6506,"tokens_out":16151,"duration_ms":181185,"concrete_test":"Use the same simulation pipeline (CORSIKA plus back-tracking through PFSS/CSSS) to compute the relative deficit in each of the eight off-source windows with respect to a no-Sun isotropic background for each of the 7 observing seasons. If the average off-source deficit is nonzero or varies with sunspot number, the background definition in Eq. (2.2) is contaminated. Then re-derive the annual RD(1°) using a background from a direction far from the Sun (e.g., the same declination band but opposite right ascension) and recompute the 6.4σ significance. A data-only cross-check: compute each off-source window’s deficit using the other seven off-source windows as background; a nonzero or time-varying result would indicate the off-source region is not neutral.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The strongest evidence for the central claim is the 6.4σ preference for a decreasing linear relation between the measured Sun shadow deficit and average sunspot number (Fig. 4). That result assumes the eight off-source windows used in Eq. (2.2) are a clean, Sun-free background. At 50–60 TeV, the solar magnetic field (Parker spiral plus coronal field) deflects cosmic rays by angles of order the Sun’s angular radius over 1 AU, and the resulting shadow/deflection pattern can extend over several degrees around the Sun. If this angular scale is comparable to the offsets of the off-source windows, then ⟨Noff⟩ is not a no-Sun background; its own solar-cycle dependence would directly bias every annual RD(1°) in Fig. 3, potentially creating or inflating the 6.4σ trend. The Moon control (Fig. 2) does not rule this out because the Moon has no magnetic field. Section 2.2 defines the off-source windows but gives no test that they are free of Sun-induced modulation, and Section 4 quotes Poisson-only significances without propagating this possible systematic.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"This proceedings paper (ICRC 2019) reports a seven-year (2010-2017) measurement of the cosmic-ray Moon and Sun shadows with IceCube at a median primary energy around 50-60 TeV. The Moon shadow is used as a control and is consistent with the expectation from a geometric lunar disk (p = 0.32). For the Sun, the relative deficit inside a 1-degree-radius circle around the center of gravity of the shadow, Eq. (2.3), falls from roughly 6% in the 2010/11 season to about 2% in 2016/17. The authors report a 7.3-sigma rejection of a simple solar-disk model, a 6.4-sigma preference for a decreasing linear relation between the relative deficit and the average sunspot number, and agreement with PFSS/CSSS solar magnetic field models that is much better than the disk expectation, with residual tensions of 3.0 sigma (PFSS) and 2.8 sigma (CSSS). The paper concludes that the TeV Sun shadow varies over the 11-year solar cycle, consistent with expectations from the solar magnetic field cycle.","tokens_in":6725,"tokens_out":21985,"duration_ms":208548,"significance":"If correct, the central result—a solar-cycle-correlated variation of the cosmic-ray Sun shadow at 50-60 TeV, an energy regime above the Tibet and ARGO measurements—is a noteworthy probe of the coronal and interplanetary magnetic field. The analysis has genuine strengths: the correlation is established against external WDC-SILSO sunspot numbers and is not fitted; the PFSS/CSSS predictions are driven by SOLIS magnetograms rather than by the shadow data; the stable Moon shadow (Fig. 2) controls for detector drift, pointing, and acceptance over the same seven years; and the concluding section explicitly acknowledges the factor-of-two magnetogram scale uncertainty and the neglect of CMEs. The weaknesses are statistical rather than conceptual: the off-source background windows are not demonstrated to be free of solar modulation, the quoted significances are Poisson-only with no propagated detector or model systematics, and the test statistics are not fully specified. These issues are identifiable and fixable within the manuscript's scope, and none of them constitutes an internal inconsistency in the analysis.","major_comments":[{"comment":"The 6.4-sigma trend in Fig. 4 and every annual RD(1°) point in Fig. 3 rest on the assumption that <Noff> in Eq. (2.2), the average of the eight off-source windows, is an unbiased, Sun-free background. The manuscript provides no test of this assumption: the angular separation of the off-source windows from the Sun's nominal position is not stated, and the §3 simulations (which produce angular structure over the 3-degree by 3-degree maps shown in §4) are not used to estimate the solar modulation at the off-source positions. Because the magnetic deflection pattern and its magnitude vary with the solar cycle, any solar-cycle-dependent component in <Noff> would bias the yearly points in a way that correlates with sunspot number, and the Poisson-only significance would not capture it. The Moon control in Fig. 2 cannot rule this out because the Moon has no magnetic field. The contamination may well be negligible at the actual offsets, but the paper should demonstrate that, for example by reporting the year-by-year stability of the eight individual off-source windows, by estimating the fractional modulation at the off-source positions with the back-tracking simulations, or by recomputing RD(1°) with alternative off-source geometries.","section":"§2.2, Eq. (2.3), Figs. 3-4"},{"comment":"The headline significances are not reproducible as stated. The paper does not give the test statistic (chi-square, likelihood ratio, or other), the number of degrees of freedom, or the way the error bars on the annual RD(1°) points are computed for (i) the 7.3-sigma rejection of the solar-disk model, (ii) the 6.4-sigma preference for a decreasing linear relation versus a constant, and (iii) the 3.0-sigma (PFSS) and 2.8-sigma (CSSS) model deviations. All quoted significances are Poisson-only; no systematic uncertainty is propagated. This matters most for the model comparison, where the factor-of-two magnetogram scale uncertainty acknowledged in §5 via ref [21] is of the same order as the 3.0/2.8-sigma tensions; the claim that the models agree with the data better than the disk should be re-quantified with a systematic band on the model predictions. In addition, if any of the p-values are computed from the smoothed 0.1-degree-binned maps, the 0.7-degree boxcar smoothing makes neighboring bins strongly correlated, and treating them as independent would overstate the significance; the scale uncertainty in ref [21] needs to be applied to the model predictions, not only cited.","section":"§4 (Figs. 3-4), §5"},{"comment":"The center-of-gravity correction is a selection on the same data that are then used to measure RD(1°): the shadow center is defined by the bins with smoothed deficit greater than 3%, and the deficit is then measured around that center. Self-centering biases the recovered deficit upward, and the bias is expected to be largest for the shallow, noisier shadows of the solar-maximum years, which are precisely the points that set the slope in Fig. 4. Please quantify the bias with pseudo-experiments (for example, injecting shadow templates of known depth and center shift into the event sample and repeating the COG correction), or demonstrate through a comparable cross-check that the solar-cycle trend in Figs. 3-4 is unchanged. The paper should also state whether there are any years in which no bins pass the 3% threshold and how such years are handled.","section":"§2.2"}],"minor_comments":[{"comment":"As defined, the 'relative deficit' is negative for a shadow, while Figs. 2-4 plot positive values; please state explicitly that the plotted quantity is the absolute value of RD(1°).","section":"§2.2, Eqs. (2.2)-(2.3)"},{"comment":"Please state the angular separation between the on-source window and the nearest off-source windows; the figure suggests a 3x3 arrangement of touching 6-degree by 6-degree boxes, implying roughly 3 degrees between the on-source center and the nearest off-source edges.","section":"Fig. 1, §2.2"},{"comment":"The text says the simulations are processed like the data; please confirm explicitly that the simulated model predictions in Fig. 3 undergo the identical center-of-gravity correction and 1-degree-circle RD definition, since the comparison in that figure depends on it.","section":"§3, §4"},{"comment":"Specify the time interval over which the average sunspot number is computed for each season, given that the Sun filter collects data for roughly 90 days per season (§2.1).","section":"Fig. 4"},{"comment":"'Down-going muons, with an energy of roughly more than 400 GeV' should be rephrased; also, give the median energy of the Sun sample rather than only the range quoted.","section":"§1.2"},{"comment":"State how many statistically independent simulated primaries remain after the resampling (duplicating each event by factors of 20-100 does not add statistical power), and note that the Delta-delta distribution is preserved, not resampled.","section":"§3"},{"comment":"'These numbers, independent on the exact solar field model prove a variation' should read 'independent of' and a verb such as 'indicate' or 'support' is more appropriate than 'prove' for a Poisson-only analysis; the abstract statement that the results correlate 'with theoretical models of the solar magnetic field' is also stronger than the 3.0/2.8-sigma tensions justify.","section":"§5, abstract"},{"comment":"Ref. [15] is cited in its arXiv preprint form although it provides the back-tracking framework used here; please update to the published version if available.","section":"References"}],"recommendation":"major_revision","confidential_remarks":"This is a conference proceedings contribution, and several of the requested items (an off-source window stability test, pseudo-experiments for the COG bias, systematic bands on the model predictions) would naturally be reported in a full journal paper; I am nonetheless recommending major revision because the central significance claims are not self-contained as written. I see no evidence of a flawed or circular result, and the requested checks are well within the collaboration's means given the existing back-tracking and detector-simulation framework. The self-citation to ref [15] is appropriate, and the use of external sunspot and magnetogram data is properly disclosed. If the venue cannot accommodate the extended material, the collaboration should at least include concise versions of the checks in the arXiv version."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Hi Fred, quick take on 1908.10148.\n\nThe genuinely new result is the 6.4σ preference for a decreasing linear relation between the IceCube Sun-shadow deficit and average sunspot number over seven seasons at a median energy of 50-60 TeV. That is a higher-energy handle than the Tibet (~10 TeV) and ARGO (~1 TeV) measurements. The PFSS/CSSS comparison is a real test of solar magnetic field models, and the Moon control is the right check: the Moon shadow is stable over the same period, so detector drift is not an obvious explanation. They cite the earlier work properly and acknowledge the magnetogram scale uncertainty and the neglect of CMEs.\n\nThe softest spot is the background estimate. The significances in Eq. (2.2) all rest on the eight off-source windows being a clean, Sun-free background. The paper never explicitly tests that. The stress-test worry is plausible in principle: during solar maximum the deflection pattern around the Sun is larger, and if the off-source windows pick up any of it, the relative deficit would be biased. At 50-60 TeV the deflection should be sub-degree, so the risk is likely small if the windows are as separated as the 6°x6° layout suggests, but the text does not demonstrate that. A displaced-window or fake-event check would settle it.\n\nThe second issue is that the significances are Poisson-only. No systematic bands are propagated into the PFSS/CSSS comparison, and the residual 2.8-3.0σ tensions are exactly where a systematic envelope matters. The paper mentions the factor-of-two magnetogram scale uncertainty but does not fold it into the model curves. That is a fixable omission.\n\nThese are not deal-breakers. The existence of a solar-cycle variation is consistent with Tibet and ARGO, and the Moon control supports detector stability. I agree with the reader's conditional verdict: the result is probably right, but the quoted confidence overstates what is known.\n\nWho benefits? Anyone working on Sun-shadow physics, heliospheric magnetic fields, or cosmic-ray transport in the inner heliosphere. It is a conference proceedings, so the depth is limited, but the result deserves a serious referee. I would accept it for peer review, and if I were the referee, the main request would be a clean demonstration that the off-source windows are Sun-free, plus a systematic error budget for the model comparison. Reading group: maybe. Cite: yes. Serious thinker: yes.","headline":"The IceCube Sun shadow paper is a credible incremental result—a 7-year, 50-60 TeV shadow variation tracking the solar cycle with a stable Moon control—but the off-source background needs a cleaner test before the nominal significances are taken at face value.","tokens_in":7258,"tokens_out":7100,"would_cite":true,"duration_ms":71622,"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":"The cosmic-ray shadow of the Sun, measured by IceCube at median energies around 50-60 TeV, varies over time and tracks the 11-year solar activity cycle as traced by sunspot number and solar magnetic field models.","keywords":["cosmic-ray Sun shadow","IceCube","solar magnetic field","solar activity cycle","sunspot number","cosmic-ray Moon shadow","Monte Carlo back-tracking","coronal magnetic field models"],"falsifier":"Re-run the full pipeline with event directions scrambled in right ascension (or with a different set of off-source windows) and check whether a sunspot-correlated 'shadow' still appears; if it does, the background estimate is biased and the claim fails. A simpler falsifier is a full solar cycle in which the Sun shadow stays constant while sunspot number changes, which would contradict the reported 6.4 sigma linear trend.","tokens_in":6315,"feed_emoji":"☀️","tokens_out":9839,"duration_ms":90413,"temperature":0.7,"pith_summary":"This paper aims to show that the cosmic-ray shadow of the Sun, recorded by IceCube at a median primary energy of roughly 50-60 TeV, is not a static feature but changes over time in step with the 11-year solar activity cycle. The key evidence is that the shadow's relative deficit decreases with the average sunspot number, with a linear decline favored over a constant at 6.4 sigma, and that a simple opaque solar disk is rejected at 7.3 sigma. If true, the Sun shadow becomes a remote, time-resolved probe of the solar magnetic field near the Sun, a region that cannot be probed directly. The Moon shadow, measured with the same pipeline, stays constant, which argues that the Sun shadow variation is physical rather than detector drift.","feed_headline":"Sun's cosmic-ray shadow tracks the 11-year solar cycle","feed_subtitle":"At 50-60 TeV the shadow weakens as sunspots rise, giving a remote probe of the solar magnetic field.","key_machinery":"The quantity carrying the analysis is the relative deficit $RD(1^\\circ) = 100\\% \\cdot \\frac{N_{\\rm on} - \\langle N_{\\rm off}\\rangle}{\\langle N_{\\rm off}\\rangle}$, computed inside a 1-degree radius circle around the shadow's center of gravity, with the background $\\langle N_{\\rm off}\\rangle$ estimated from eight nearby off-source windows. Year by year this deficit is compared against three expectations: a simple opaque solar disk, the predictions obtained by back-tracking simulated cosmic-ray primaries through time-evolving PFSS and CSSS coronal magnetic field models with a per-particle passing probability, and the average sunspot number. The Moon shadow, processed identically, provides the detector-stability control.","core_discovery":"Using seven years of IceCube data, from May 2010 to May 2017, the paper reports that the relative deficit of cosmic rays within one degree of the Sun's shadow center is significantly lower when the Sun is active: the data favor a decreasing linear relation between the deficit and the average sunspot number over a constant relation at 6.4 sigma, and the constant solar-disk expectation is excluded at 7.3 sigma. Simulations in which primary cosmic rays are back-tracked through two time-dependent models of the coronal magnetic field (the PFSS and CSSS models) reproduce the time variation better than a disk, though with residual tensions near 3 sigma. Over the same interval the Moon shadow matches a simple lunar disk with a p-value of 0.32, indicating that the detector's angular reconstruction and event selection were stable enough for the solar variation to be taken at face value.","pith_inferences":["A natural next test, not performed in the paper, is to apply the same background-subtracted deficit to individual solar rotations to search for transient shadow changes associated with coronal mass ejections.","The reported anti-correlation suggests the solar magnetic field acts as an energy-dependent scattering opacity; applying the same analysis at lower primary energies could reveal whether the deficit saturates at solar maximum.","If the off-source background estimate is independently validated with synthetic or scrambled events, the deficit-sunspot relation could become a ground-based constraint on heliospheric magnetic field models, complementing in-situ spacecraft data.","The ~3 sigma residual disagreement with both PFSS and CSSS models points toward missing magnetic structure such as non-potential fields; a time-dependent non-potential simulation would be the direct test."],"forward_implications":["The Sun shadow can serve as an indirect, energy-resolved measurement of the solar magnetic field close to the Sun, where direct measurements are not available.","A longer IceCube data span covering more of the solar cycle can test whether the linear sunspot-deficit relation persists through solar minimum and into the next maximum.","The residual ~3 sigma tension with both field models motivates including coronal mass ejections, a more realistic solar wind profile, and corrected magnetogram scales in future simulations.","Because the Moon shadow is stable over the same seven years, the Sun shadow variation cannot be attributed to detector acceptance or reconstruction drift.","At these energies the shadow responds to solar activity, so combining with lower-energy and higher-rigidity cosmic-ray observatories can map the rigidity dependence of the effect."],"supporting_citations":[{"why":"Supplies the theoretical back-tracking framework that predicts a time-dependent Sun shadow from solar magnetic field models; the data are compared against these predictions.","marker":"[15]"},{"why":"First IceCube detection of the Sun shadow above 40 TeV; this analysis updates that work with seven seasons of data.","marker":"[11]"},{"why":"Prior measurement demonstrating that the Sun shadow varies with the 11-year solar cycle at ~10 TeV, the higher-energy claim this paper extends.","marker":"[1]"},{"why":"Provides the average sunspot numbers used to test the correlation between shadow deficit and solar activity.","marker":"[20]"},{"why":"Generates the detector-specific air-shower simulations that set the median primary energy and the expected event weights.","marker":"[13]"},{"why":"Supplies the primary cosmic-ray flux model used to weight the simulated events.","marker":"[14]"},{"why":"Define the PFSS potential-field source-surface model of the coronal magnetic field used in the simulation.","marker":"[16, 17]"},{"why":"Define the CSSS model including current sheets, the second coronal field model used in the simulation.","marker":"[18, 19]"},{"why":"Documents factor-of-two systematic errors in magnetogram field-strength scales, invoked to interpret the residual ~3 sigma disagreement with the field models.","marker":"[21]"}],"fun_headline_variants":["Sun shadow weakens as solar activity rises, IceCube finds","IceCube sees Sun's cosmic-ray shadow follow 11-year cycle","Sunspot counts linked to Sun's cosmic-ray shadow variation","Solar cycle modulates Sun's shadow in TeV cosmic rays","IceCube: Sun's shadow dips with rising solar activity"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The analysis assumes that the eight off-source windows provide an unbiased estimate of the background in the on-source window, so the relative deficit is purely the shadow signal; if cosmic-ray anisotropy, seasonal muon-rate changes, or detector acceptance differ between the windows, every deficit and significance is biased.","fun_headline_variants_meta":{"raw":{"variants":["Sun shadow weakens as solar activity rises, IceCube finds","IceCube sees Sun's cosmic-ray shadow follow 11-year cycle","Sunspot counts linked to Sun's cosmic-ray shadow variation","Solar cycle modulates Sun's shadow in TeV cosmic rays","IceCube: Sun's shadow dips with rising solar activity"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000539,"raw_usage":{"total_tokens":2575,"prompt_tokens":921,"completion_tokens":1654,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":537,"completion_tokens_details":{"reasoning_tokens":1570}},"tokens_in":537,"tokens_out":1654,"duration_ms":13097,"temperature":1.0,"reasoning_tokens":1570,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T10:51:24.135868+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Re-run the full pipeline with event directions scrambled in right ascension (or with a different set of off-source windows) and check whether a sunspot-correlated 'shadow' still appears; if it does, the background estimate is biased and the claim fails. A simpler falsifier is a full solar cycle in which the Sun shadow stays constant while sunspot number changes, which would contradict the reported 6.4 sigma linear trend.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"First IceCube detection of the Sun shadow above 40 TeV; this analysis updates that work with seven seasons of data."},{"cited_title":"Amenomori et al., Phys","cited_arxiv_id":null,"evidence_quote":"Prior measurement demonstrating that the Sun shadow varies with the 11-year solar cycle at ~10 TeV, the higher-energy claim this paper extends."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides the average sunspot numbers used to test the correlation between shadow deficit and solar activity."},{"cited_title":"Heck et al., F orschungszentrum Karlsruhe Report FZKA6019 (1998) 90","cited_arxiv_id":null,"evidence_quote":"Generates the detector-specific air-shower simulations that set the median primary energy and the expected event weights."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the primary cosmic-ray flux model used to weight the simulated events."},{"cited_title":"Riley et al., Sol","cited_arxiv_id":null,"evidence_quote":"Documents factor-of-two systematic errors in magnetogram field-strength scales, invoked to interpret the residual ~3 sigma disagreement with the field models."}],"review_version":1}