{"id":"696b1406-7951-4fd1-89da-f4109b9b1c65","arxiv_id":"1908.04474","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":7.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"Modeling the well-known size asymmetry of sunspot pairs in surface flux transport simulations reduces the error in reconstructed solar axial dipole strength by 30-40 percent and changes the predicted timing of polar field reversals.","lead":"This paper shows that the size difference between the leading and following spots in a sunspot pair changes how magnetic field is carried to the Sun's poles, which matters for predicting the next solar cycle. Including this asymmetry in standard solar surface simulations reduces the mismatch with observed magnetic dipole measurements by roughly a third.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The 30–40% improvement is measured in-sample: the same WSO axial dipole series is used to fit the Table 2 parameters and to compute the error, so the headline reduction may reflect parameter fitting rather than the asymmetry mechanism.","rationale":"The physics in §3 is credible: a single Gaussian patch's contribution to the polar field depends strongly on its width (Fig. 2), and a wider following patch visibly reduces or reverses the cross-equatorial flux (Fig. 3). My objection is not to the mechanism but to the quantitative headline. The full-cycle comparison in Table 2 has the structure of a model-calibration exercise: BT, c−cmin, and T21–T24 are chosen by minimizing the L2 deviation from WSO axial dipole, and the same WSO series is then used as the yardstick for the 30–40% figure. Because both runs have the same number of fitted parameters but the asymmetric run happens to adopt a different effective c and different tilt factors, the error gap could be a byproduct of the optimization rather than evidence for the asymmetry. The fact that c−cmin is optimized to its lower boundary in both cases (Table 2) further signals that the comparison is being made at a constraint edge, where the mapping from observed area asymmetry to model size asymmetry (Eqs. 19–20, 26–27) is most sensitive and least tested. The paper honestly states the polar-field discrepancy at the end of Cycle 21 and the fspot=0.4 vs 0.8 observational tension, but these caveats do not by themselves secure the quantitative claim. An out-of-sample protocol is the minimal check that would tell whether the asymmetry genuinely improves hindcasts or just improves the fit to the calibration data. Since the reader's CONDITIONAL verdict already asks for out-of-sample validation, my stress-test confirms rather than changes that verdict.","tokens_in":14775,"tokens_out":9141,"duration_ms":91767,"concrete_test":"Leave-one-cycle-out cross-validation: for each held-out cycle in Cycles 21–24, optimize BT, c, and the remaining tilt factors on the other three cycles, separately for fspot=0.4 and fspot=1.0, then compute the RMS axial-dipole error on the held-out cycle. If the fspot=0.4 model does not beat fspot=1.0 in mean held-out error, the claimed 30–40% improvement is an in-sample artifact rather than a robust effect of morphological asymmetry.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's central quantitative claim is the 30–40% reduction in RMS axial-dipole error from including spot asymmetry (Abstract; §5). Table 2 produces this number by optimizing BT, c−cmin, and the cycle tilt factors T21–T24 against the WSO axial-dipole time series and then computing the residual error on that same series. This is an in-sample fit, not a predictive test. The asymmetric case also sits at the lower boundary of the allowed c range (c−cmin=0, Table 2), so c is forced to 3.88 rather than the symmetric case's 2.72, and the cycle-dependent tilts differ; the error reduction could therefore be purchased by these parameter shifts rather than by the asymmetry physics isolated in §3. No cross-validation, bootstrap, or sensitivity analysis is reported, and §6 concedes that the asymmetric model deviates from the observed polar field near the end of Cycle 21. Consequently the abstract's numerical improvement claim is not yet established as a property of the asymmetry rather than of the fitting procedure.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper investigates whether the well-known morphological asymmetry between leading and following sunspots affects polar-field formation and solar-cycle prediction. Using a one-dimensional surface flux transport model, the authors first isolate the effect of a single bipolar magnetic region (BMR) and show that a larger, more diffuse following polarity reduces or even reverses the net cross-equatorial flux that builds the polar field, especially for large and high-latitude BMRs (Section 3). They then convert an observed sunspot-area asymmetry fspot into the model's patch-size asymmetry fdelta via a flux-area scaling with a parameter c (Section 4), and run simulations for Cycles 21-24, optimizing free parameters against the Wilcox Solar Observatory axial dipole series (Section 5). The paper reports a 30--40 percent reduction in RMS axial-dipole error when the asymmetry is included, while also acknowledging uncertainties in fspot and a remaining discrepancy in the polar field near the end of Cycle 21 (Section 6).","tokens_in":14948,"tokens_out":5686,"duration_ms":53488,"significance":"If the quantitative claim were established, the paper would be important: standard SFT models treat leading and following polarities as symmetric, and the paper's Section 3 experiments show a clean, parameter-free mechanism by which size asymmetry changes cross-equatorial flux and the timing of polar-field reversal. This mechanism is genuinely new in the SFT context and is supported by the isolated single-BMR simulations. The manuscript also makes a falsifiable qualitative prediction that asymmetric BMRs contribute less to the polar field, which deserves credit. However, the headline 30--40 percent RMS improvement is not yet established as a property of the asymmetry itself: it comes from an in-sample optimization of several free parameters against the same axial dipole series used to compute the error, and the solution sits on the boundary of the allowed parameter range. The qualitative conclusion is sound; the quantitative conclusion needs additional validation.","major_comments":[{"comment":"The 30--40 percent improvement reported in the Abstract and Section 5 is an in-sample fit residual, not a predictive test. The same WSO axial dipole time series is used both to optimize BT, c-cmin, and the cycle-dependent tilt factors T21-T24 and to compute the RMS error; no cross-validation, holdout, bootstrap, or sensitivity analysis is presented. Because the asymmetric case also has different optimized parameters and sits at c-cmin = 0, the reduction from 3.50 G to 2.18 G cannot be uniquely attributed to the asymmetry mechanism rather than to the parameter freedom absorbed by the fit. The authors should provide an out-of-sample or cross-validated assessment, or at least a parameter-sensitivity analysis showing that the improvement is stable when the fitted parameters are perturbed.","section":"§5, Table 2 and Fig. 7"},{"comment":"The total-flux factor c is optimized to the lower boundary cmin in both the symmetric and asymmetric cases in Table 2, meaning the reported improvement relies on the flux-area relation Phi_tot = c BT Aspot at exactly the boundary of the Lambert-W existence range. This is a delicate point: the boundary value cmin = 2e/(1+fspot) forces the peak field Bmax to take a specific value, and the asymmetric case therefore uses a different total flux per spot area than the symmetric case. The paper does not justify why the physical system should sit at this boundary, nor does it explore the sensitivity of the 30--40 percent improvement to the assumed flux-area scaling. Showing results for c above cmin, or for an independently calibrated relation between spot area and flux, would substantially strengthen the quantitative claim.","section":"§4.2, Eq. (27), and Table 2"},{"comment":"The quantitative conclusion is heavily contingent on the adopted value fspot = 0.4, which is not optimized but fixed using Tlatov et al. (2014). The paper itself notes that Muraközy et al. (2014) implies fspot ~ 0.8, which would produce a much weaker asymmetry, and that the asymmetric simulation deviates from the observed polar field near the end of Cycle 21. These concessions mean that the reported 30--40 percent error reduction may shrink or disappear under a reasonable alternative observational input. The authors should either quantify how the RMS error changes with fspot (e.g., a sweep over fspot = 0.4, 0.6, 0.8, 1.0) or substantially soften the abstract's numerical claim.","section":"§6"}],"minor_comments":[{"comment":"The title uses 'sun spots' as two words, while the abstract and body use 'sunspot'; the spelling should be made consistent.","section":"Title"},{"comment":"The sentence 'In the weakly asymmetric case (fdelta = 1.5; Fig. 3b), the amount of the cross-equatorial flux becomes smaller than the asymmetric case' should read 'smaller than in the symmetric case', since the comparison is with fdelta = 1.0.","section":"§3.2, Fig. 3 caption"},{"comment":"The text 'We used delta_spot = 0.4 as the typical ratio between the leading and following sunspot areas' appears to refer to fspot, the area ratio, not delta_spot, and should be corrected.","section":"§6"},{"comment":"The definition Phi_tot = c BT Aspot is a strong modeling assumption; the paper would benefit from a brief discussion of how this form relates to previous SFT parameterizations and what range of c is physically plausible from sunspot magnetic-field observations.","section":"§4.2, Eq. (26)"}],"recommendation":"major_revision","confidential_remarks":"The Section 3 demonstration is solid and likely publishable. The main risk is that the abstract's 30--40 percent claim is presented as a finding about asymmetry when it currently measures an in-sample fit under boundary-condition parameter choices. I would be comfortable with acceptance after the authors either provide a robustness analysis (cross-validation or parameter sweeps) or rephrase the quantitative claim as conditional on model assumptions."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The asymmetry effect is real and previously underappreciated, but the 30–40% improvement is an in-sample fit, not a predictive result. The single-BMR experiments in Section 3 are the core of the paper and they're solid. A wider following spot dumps more flux across the equator, and for large, high-latitude BMRs it can reverse the polar-field contribution. That is a genuine new result, and the paper is the first to quantify it in an SFT context. The derivation in Section 4 connecting area asymmetry fspot to patch-size asymmetry fδ via the Lambert W function is neat, and Appendix B gives a simple diffusion argument for why larger BMRs keep the asymmetry longer.\n\nThe soft spot is the quantitative claim in Section 5. The 30–40% RMS reduction is computed after a grid search over BT, c, and T21–T24 that minimizes the L2 difference to the WSO axial dipole on the same data used for the error. That is in-sample fitting, not a predictive test. Both optimized cases sit at the lower boundary of c, and the cycle-dependent tilts shift, so the improvement could be purchased by parameter changes rather than by the asymmetry physics isolated in Section 3. No cross-validation, bootstrap, or sensitivity analysis is provided, and the paper itself concedes the asymmetric model deviates from the observed polar field near the end of Cycle 21. The fspot choice is also a single value (0.4 from Tlatov et al.); using Muraközy et al.'s 0.8 would weaken the effect, as the paper notes.\n\nThat said, the limitations section is honest, the data calibration is careful, and the citation pattern is appropriate. This is a paper that a serious referee should engage with, but the verdict should be conditional: the mechanism is plausible, the quantitative improvement is not yet established. For a revision, I'd ask for an out-of-sample test (e.g., fit on cycles 21–23, score on 24) and a sensitivity scan over fspot and c. The paper is worth bringing to a reading group and citing for the effect, but not as evidence for a 30–40% prediction gain.\n\nRecommendation: send to peer review, but insist on the sensitivity analysis before publishing the headline number.","headline":"The asymmetry mechanism is credible and the single-BMR experiments are clean, but the headline 30–40% improvement is an in-sample fit, not a validated prediction.","tokens_in":15523,"tokens_out":3014,"would_cite":true,"duration_ms":38084,"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":"Sunspot pairs' size asymmetry changes polar field formation enough that including it cuts model error against observed axial dipole strength by 30–40 percent.","keywords":["solar cycle prediction","surface flux transport model","bipolar magnetic region","morphological asymmetry","sunspot area asymmetry","polar magnetic field","axial dipole strength","Joy's law"],"falsifier":"Run the same surface flux transport hindcasts for cycles 21–24 using a following-to-leading sunspot area ratio of 0.8 (the alternative observational estimate) or with a flux-area scaling parameter above its minimum; if the axial dipole error no longer improves by 30–40 percent relative to the symmetric model, the central claim is refuted. A direct test would measure the magnetic-flux-weighted width ratio of leading versus following polarity patches in synoptic magnetograms and compare it with the model's inferred $f_\\delta$.","tokens_in":1739,"feed_emoji":"🌞","tokens_out":3333,"duration_ms":85063,"temperature":0.7,"pith_summary":"This paper argues that a long-ignored trait of sunspot pairs—the following spot being smaller in area but more spread out in magnetic field—shapes the Sun's polar magnetic field enough to matter for solar cycle prediction. Standard surface flux transport models treat each bipolar region as two identical magnetic patches; the paper shows that letting the following patch be spatially wider changes how much flux crosses the equator and reaches the poles. In hindcasts of solar cycles 21–24, adding this asymmetry reduces the root-mean-square error from the observed axial dipole strength by 30–40 percent. If true, cycle forecasts built on polar field precursors should account for spot asymmetry, and the asymmetry may help explain the anomalously long minimum between cycles 23 and 24.","feed_headline":"Asymmetric sunspots improve solar cycle hindcasts by 30–40%","feed_subtitle":"A long-ignored spot asymmetry reshapes polar field timing and cuts hindcast error on four solar cycles.","key_machinery":"The central object is the bipolar magnetic region (BMR) modeled as two Gaussian magnetic patches of widths $\\delta_L$ and $\\delta_F$ with flux balance $B_{\\max}^L(\\delta_L)^2 \\sim B_{\\max}^F(\\delta_F)^2$, so the following patch is wider and weaker when its sunspot area is smaller. The asymmetry parameter is $f_\\delta = (\\delta_F/\\delta_L)^2$, tied to the observed sunspot area ratio $f_{\\rm spot}$ through a Lambert-W inversion that yields the Gaussian width and peak field from area and flux. The diagnostic that carries the argument is the southern net hemispheric flux after five years: it measures how much cross-equatorial magnetic flux a BMR delivers to the polar field. The paper also uses the one-dimensional azimuthally averaged surface flux transport equation, whose linearity lets the authors calibrate the initial condition and threshold field by least squares and grid-search the remaining parameters.","core_discovery":"The central claim is that the morphological asymmetry between leading and following sunspots has a significant, previously neglected effect on the evolution of the large-scale surface magnetic field. Because magnetic flux is balanced between the two polarities, a smaller following sunspot area implies a wider, more diffuse following magnetic patch; that wider patch is disproportionately carried across the equator by diffusion and meridional flow, canceling the leading-polarity flux that normally builds the polar field. For large, high-latitude bipolar regions the effect is strong enough that strongly asymmetric BMRs reverse the regular polar field formation. In simulations driven by the observed sunspot record for cycles 21–24, the asymmetric model reduces the RMS difference of the axial dipole strength from WSO observations by 30–40 percent compared with the symmetric model, and shifts the timing of polar field reversal closer to observations.","pith_inferences":["If this effect is as large as reported, the magnetic-flux-weighted width of following-polarity patches in synoptic magnetograms should systematically exceed that of leading patches; this is directly testable without relying on sunspot area catalogs.","The 30–40 percent improvement may partially absorb other missing physics, such as tilt-angle scatter or active region inflow; a model that includes both effects together could attribute the gain more precisely.","Because the asymmetry decays on a timescale of roughly a month for typical spots, and more slowly for large BMRs, cycle-to-cycle variation in the population of large active regions could drive part of the observed variability in polar field timing, independent of tilt angle.","A natural next test is to build cycle 25 hindcasts with the asymmetric source term and compare the predicted polar field and next-cycle amplitude against observations, using the inferred asymmetry as a free parameter per hemisphere."],"forward_implications":["Standard surface flux transport models that assume symmetric spots overestimate polar field buildup and underestimate how much following-polarity flux reaches the poles, especially in cycles with large, high-latitude BMRs.","The timing of polar field reversal shifts: with asymmetry, the axial dipole begins to decline before the end of the cycle rather than amplifying to the end, matching observed profiles.","Solar cycle predictions based on polar field precursors should include spot size asymmetry, otherwise the predicted amplitude of the next cycle can be systematically off.","The long flat minimum between cycles 23 and 24 can be reproduced by the asymmetric model, suggesting asymmetry as a candidate explanation for weak, prolonged minima.","The asymmetry increases the grainy structure in the activity belts of the butterfly diagram, offering a partial explanation for observed low-latitude flux concentrations without raising high-latitude net flux."],"supporting_citations":[{"why":"It supplies the surface flux transport equation, meridional flow profile, and Gaussian magnetic patch shape that the BMR model extends.","marker":"van Ballegooijen et al. (1998)"},{"why":"It provides the BMR parameterization, including tilt angle scaling and polarity separation, that this paper augments with size asymmetry.","marker":"Cameron et al. (2010)"},{"why":"It gives the observed following-to-leading sunspot area ratio fspot ≈ 0.4 used to set the asymmetry in the fiducial runs.","marker":"Tlatov et al. (2014)"},{"why":"It offers a milder area asymmetry estimate (≈0.8) that defines the uncertainty range discussed in Section 6.","marker":"Muraközy et al. (2014)"},{"why":"It justifies the one-dimensional azimuthally averaged flux transport equation used for the simulations.","marker":"Cameron & Schüssler (2007)"},{"why":"It is one of the prediction baselines and the source of the turbulent magnetic diffusivity value (250 km²/s) used in all simulations.","marker":"Cameron et al. (2016)"},{"why":"It documents the small-group area bias in the SOON data that motivates excluding active regions under 100 msh.","marker":"Muñoz-Jaramillo et al. (2015)"}],"fun_headline_variants":["Sunspot asymmetry cuts solar cycle forecast error by 40%","Asymmetric sunspots: a previously ignored factor in solar cycle prediction","Sunspot asymmetry can reverse polar field formation, improving forecasts","Previously ignored sunspot asymmetry reshapes polar field and cycle forecasts","Asymmetric sunspots reduce hindcast error by 30–40%"],"cache_read_input_tokens":17664,"weakest_assumption_plain":"The reported 30–40 percent improvement rests on assuming the true sunspot area asymmetry is about 0.4 (following spot 40 percent as large as leading spot) and on a particular conversion from spot area to magnetic patch width whose free parameter the optimization pins to its lower bound; if the real asymmetry is milder, as one competing study suggests, the improvement could shrink or vanish.","fun_headline_variants_meta":{"raw":{"variants":["Sunspot asymmetry cuts solar cycle forecast error by 40%","Asymmetric sunspots: a previously ignored factor in solar cycle prediction","Sunspot asymmetry can reverse polar field formation, improving forecasts","Previously ignored sunspot asymmetry reshapes polar field and cycle forecasts","Asymmetric sunspots reduce hindcast error by 30–40%"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000766,"raw_usage":{"total_tokens":3374,"prompt_tokens":898,"completion_tokens":2476,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":514,"completion_tokens_details":{"reasoning_tokens":2388}},"tokens_in":514,"tokens_out":2476,"duration_ms":18553,"temperature":1.0,"reasoning_tokens":2388,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T13:42:02.949643+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Run the same surface flux transport hindcasts for cycles 21–24 using a following-to-leading sunspot area ratio of 0.8 (the alternative observational estimate) or with a flux-area scaling parameter above its minimum; if the axial dipole error no longer improves by 30–40 percent relative to the symmetric model, the central claim is refuted. A direct test would measure the magnetic-flux-weighted width ratio of leading versus following polarity patches in synoptic magnetograms and compare it with the model's inferred $f_\\delta$.","supporting_citations":[{"cited_title":"H., Jiang, J., Schmitt, D., & Sch¨ ussler, M","cited_arxiv_id":null,"evidence_quote":"It provides the BMR parameterization, including tilt angle scaling and polarity separation, that this paper augments with size asymmetry."},{"cited_title":"G., Vasil’eva, V","cited_arxiv_id":null,"evidence_quote":"It gives the observed following-to-leading sunspot area ratio fspot ≈ 0.4 used to set the asymmetry in the fiducial runs."},{"cited_title":"2007, ApJ, 659, 801","cited_arxiv_id":null,"evidence_quote":"It justifies the one-dimensional azimuthally averaged flux transport equation used for the simulations."},{"cited_title":"H., Jiang, J., & Sch¨ ussler, M","cited_arxiv_id":null,"evidence_quote":"It is one of the prediction baselines and the source of the turbulent magnetic diffusivity value (250 km²/s) used in all simulations."}],"review_version":1}