{"id":"1714b64b-4fe4-4ef7-85ef-f633ae8181c0","arxiv_id":"2507.11790","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":5.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":6,"one_line_summary":"Photospheric current helicity shows a pre-eruption decrease and post-eruption increase in an MHD model and in 58% and 92% of 50 observed eruptive flares, respectively.","lead":"Using a 3D model of solar eruptions and observations of 50 large flares, this paper reports that a quantity called current helicity tends to drop before a coronal mass ejection and rise after it. The pattern is proposed as a possible early indicator of when magnetic energy is stored for an eruption.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"Pre-eruption helicity decrease is evaluated without a non-eruptive control sample, so 58% may reflect baseline fluctuations rather than a CME precursor; a same-AR control-window analysis is required.","rationale":"The reader's weakest_assumption identifies exactly the load-bearing weakness: the 58% pre-eruption decrease is not benchmarked against the rate at which the same criterion is met during non-eruptive periods. The central predictive claim depends on this comparison. A slowly varying, autocorrelated quantity like H_c can cross a 10%-over-1-hour threshold frequently by chance, especially when the analysis window is 6 hours long and the flare mask is defined after the fact. Without control windows from the same ARs, the reported 58% cannot be distinguished from the background fluctuation rate. The post-eruption increase in 92% of events is physically plausible and consistent with ribbon separation, but it is not the predictive part of the claim; it also lacks confidence intervals. The MHD simulation and the two case studies provide a credible mechanism, but they do not quantify how often the precursor appears before eruptions versus at other times. The proposed control-window test would settle this directly: if matched non-eruptive windows show a similar or only slightly lower rate of pre-decrease, the predictive claim fails; if the rate is substantially lower, the claim is supported. No additional concern rises to the same level of importance. The reader's CONDITIONAL verdict remains appropriate, pending the control analysis.","tokens_in":16834,"tokens_out":3713,"duration_ms":48977,"concrete_test":"For each of the 50 events, construct a matched control window of the same 10-h length from the same AR, starting at least 24 h away from any M/X flare or CME, using the same SHARP patch, B>300 G threshold, and, if possible, the same QSL/flare-mask geometry derived at the control time. Apply the identical Section 3.1 criteria: does |H_c| show a continuous >10% decrease over >1 h ending at the window midpoint (pseudo-eruption time), and a >10% increase over >36 min afterward? Compute the control pre-decrease fraction and its binomial 95% confidence interval, and compare with 29/50 via a paired McNemar test or bootstrap. If the control fraction is >~40% or not significantly below 58%, the pre-eruption decrease cannot be attributed to CME onset. An additional permutation test randomizing the flare time within each observed 10-h window should yield a null distribution for the 58% statistic.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section 3.1 classifies a pre-eruption decrease as any continuous >10% drop in |H_c| lasting >1 h within 6 h before the flare, and Section 3.2 reports 58% (29/50) as evidence for a precursor. The paper never measures how often this criterion is met in non-eruptive periods for the same ARs. Because H_c is a slowly varying, autocorrelated integrated quantity over a flare mask, a 10%/1h threshold could be crossed routinely during normal AR evolution, flux emergence, or unrelated reconnection. Without a baseline rate, 58% is not statistically interpretable: with 50 events, a 95% confidence interval for the true rate is roughly [43%,72%], and a chance rate near 50% cannot be excluded. The predictive claim in Section 3.4 and the Conclusion therefore rests on an uncontrolled classification. The post-eruption 92% is better grounded physically but is not the predictive part; it also lacks error bars and could be influenced by the flare-mask definition. This is the single load-bearing weakness: the central claim that current helicity 'can predict CMEs to some extent' is only as strong as the excluded-baseline comparison.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper investigates the evolution of photospheric current helicity, H_c = ∫ j_z B_z dS, around coronal mass ejections. Using a 3D MHD simulation based on the tether-cutting scenario, the authors identify a pre-eruption decrease and post-eruption increase in H_c, which they attribute to the redistribution of electric currents: concentration toward the polarity inversion line before eruption and separation afterward. They then analyze 50 eruptive flares ≥ M5.0 from SDO/HMI and report that 58% show a pre-eruption decrease and 92% show a post-eruption increase in unsigned current helicity. Detailed case studies of two events are presented to support the current-redistribution mechanism, and the paper suggests that current helicity may serve as a precursor indicator for CMEs.","tokens_in":17058,"tokens_out":9199,"duration_ms":106822,"significance":"If the reported pattern is real, this would connect a routinely measurable photospheric quantity to the buildup and release of free magnetic energy, and the MHD mechanism offers a physical explanation. The post-eruption increase in current helicity is a plausible photospheric response to flare ribbon separation, and the agreement between the simulation and the two case studies is suggestive. However, the statistical claim rests on an uncontrolled classification with no baseline, and the sign convention between the model and observations is not clearly reconciled. The paper therefore has the potential to be a useful contribution, but the current observational validation is not yet convincing.","major_comments":[{"comment":"The sign convention for H_c is inconsistent between the model and the observations. In the model, H_c is defined as a signed integral (Eq. 5), and the helicity density is stated to be 'dominated by negative helicity' (§2.2, Fig. 2c). A decrease in this signed H_c is therefore an increase in |H_c|. However, the observational classification in §3.1 and §3.2 uses the unsigned |H_c|, and a 'pre-eruption decrease' means a decrease in |H_c|. Unless the model curves in Fig. 1(b) and Fig. 3(d) are explicitly plotting |H_c| and |j_z|, the model's 'pre-eruption decrease' has the opposite sign to the observational diagnostic. The authors must state whether the model's H_c is positive or negative and clarify which quantity is plotted; if the model's signed H_c is negative, then the model predicts an increase in |H_c| before eruption, contradicting the 58% observational claim.","section":"§2.2 and §3.2"},{"comment":"The pre-eruption decrease is evaluated without a control sample or a baseline rate. The criterion (a continuous >10% drop in |H_c| lasting >1 h within 6 h before the flare) is applied only to eruptive events, so the 58% (29/50) figure is not statistically interpretable. Because H_c is a slowly varying, autocorrelated integrated quantity over a flare mask, the same criterion could be met during normal active-region evolution, flux emergence, or non-eruptive reconnection. With 50 events, the binomial 95% confidence interval for 29/50 is roughly [43%, 72%], which does not exclude a chance rate near 50%. The predictive claim in §3.4 and the Conclusion is therefore unsupported without a control analysis. I request that the authors apply the same classification to non-eruptive time windows in the same active regions, or to confined flares, and report the false-positive rate.","section":"§3.1 and §3.2"},{"comment":"The measurement uncertainties in the H_c time series are not characterized. HMI vector magnetograms have noise in the transverse field component, which propagates into j_z and H_c; the 10% threshold used for classification may be comparable to the noise level. No error bars are shown on any H_c evolution curve, and no noise estimate is provided. The thresholds (10% magnitude, 1 h duration, 36 min duration, 300 G field cutoff) are arbitrary, and no sensitivity analysis is presented to show the classification is robust to these choices. The authors should estimate the uncertainty in |H_c| from the HMI data and show that the reported decreases/increases exceed the noise.","section":"§3.1"},{"comment":"The long-term evolution analysis is limited to two cases that were selected precisely because they showed the reversal pattern. This is a selection-biased sample and does not provide a predictive skill measurement. The statement that 'current helicity may have predictive potential to some extent' (abstract and conclusion) is not supported by a forecast metric such as a contingency table, a true-positive versus false-positive comparison, or any application of the same trend to a non-eruptive control interval. Please either add such an analysis or temper the predictive claim to match what the data actually show.","section":"§3.4"}],"minor_comments":[{"comment":"Typographical errors: 'enengy' should be 'energy' in the abstract and in the first sentence of the Introduction; 'Moreoever' in the Conclusion should be 'Moreover'; 'occuring' in §3.3.2 should be 'occurring'; 'indicting' in the caption of Figure 3 should be 'indicating'.","section":"Abstract and Introduction"},{"comment":"The description of the flare mask construction would benefit from a figure showing an example mask, because the choice of high-Q contour determines the pixels included in the H_c integral and could influence the results. The statement in the Discussion that the results are robust to mask shape and size is not documented; please show the test or provide a reference.","section":"§3.1"},{"comment":"The notation 'j_z' is used ambiguously in the text and figures; in some places a 'decrease in j_z' seems to mean a decrease in |j_z|, while in others it could mean a decrease in the signed value. Please define whether the shown curves and figures refer to the signed j_z or its absolute value, and use a consistent notation (e.g., |j_z|) throughout.","section":"Figure 3 and Figure 6"},{"comment":"The table is densely formatted and the caption could be clearer about the '≥6' and '≥4' entries. In addition, the flare start times are given in different formats; please standardize them to one time zone (UT) with a consistent format.","section":"Table 1"},{"comment":"The sentence 'Our analysis shows that the trend of H_c/Φ_z^2 is very similar to that of H_c in almost every case (not shown in the paper)' is a verification claim that cannot be checked because the results are not shown. Please include these plots in an appendix or in the electronic supplementary material, or state the quantitative agreement.","section":"§4"},{"comment":"The statement that 'the specific threshold used to define strong B_z does not affect the physical interpretation' is an assertion without a supporting test. Since the classification of 'strong-B_z regions' is central to the mechanism, please demonstrate this insensitivity explicitly.","section":"§3.3.1"}],"recommendation":"major_revision","confidential_remarks":"The paper addresses a timely topic and the MHD part is internally consistent. The main concern for the journal is the observational validation: the lack of a control sample is a standard but essential requirement for precursor claims in solar physics, and the sign-convention issue between signed model H_c and unsigned observed |H_c| needs to be carefully resolved. If the sign issue turns out to be a genuine mismatch, the central model-observation comparison would be reversed; I recommend that the editor request a clear statement of the sign conventions and a control-window analysis before acceptance."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The paper has a solid MHD result and a useful observational extension, but the headline claim that current helicity can predict CMEs is not supported by the statistics as presented. The post-eruption increase at 92% is robust and physically grounded; the pre-eruption decrease at 58% lacks the control baseline needed to interpret it.\n\nWhat the paper does well: the MHD analysis is internally consistent, showing current concentrating toward the PIL before eruption and moving away after, which drives the helicity reversal. The two case studies illustrate the same current redistribution in observations. The authors also correctly connect to Liu et al. (2023), noting that alpha_weighted equals H_c normalized by flux, so the two measures should behave similarly; they add raw H_c, a 50-event sample, and a mechanism. The robustness checks on flux emergence and mask size are welcome.\n\nThe soft spot is exactly where the stress test lands. Section 3.1 defines a pre-eruption decrease as any continuous >10% drop in |H_c| lasting >1h within the 6h pre-flare window, and Section 3.2 reports 58% (29/50). The paper never measures how often that criterion is met in non-eruptive windows of the same ARs. Because H_c is an integrated, autocorrelated quantity over a flare mask, a 10%/1h threshold could be crossed routinely during normal AR evolution or flux emergence. With 50 events, the 95% confidence interval for a 58% rate is roughly 43-72%, so a chance rate near 50% is not excluded. The predictive claim in Section 3.4 and the Conclusion outruns the evidence. The post-eruption 92% is better grounded physically and less likely to be pure noise, but it also lacks error bars and could be influenced by the flare-mask definition. The thresholds feel arbitrary, and selecting only eruptive flares removes any non-eruptive baseline.\n\nWho this is for: people working on photospheric precursors and flare/CME forecasting will find the MHD mechanism and the 92% post-eruption increase worth knowing. The paper deserves a serious referee, but the statistical analysis needs a real revision before the prediction claim can stand. A control-window analysis on the same ARs, plus noise-aware thresholds and a significance test, would fix the main flaw. I would send it to peer review with that request; the observational pattern and the mechanism are worth preserving even if the predictive language needs to be dialed back.","headline":"Solid MHD mechanism and a useful observational extension, but the CME-prediction claim needs a control sample before it can be taken seriously.","tokens_in":17623,"tokens_out":1971,"would_cite":true,"duration_ms":22026,"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":"Photospheric current helicity falls before coronal mass ejections and rises after them, and the paper traces this reversal to electric currents concentrating at the polarity inversion line before eruption and separating afterward.","keywords":["current helicity","coronal mass ejections","solar flares","solar eruptions","photospheric magnetic field","polarity inversion line","MHD simulation","flare prediction"],"falsifier":"Compute $H_c$ time series for a sample of active regions that never produce an M5.0 or larger flare, matched to the 50 events in field strength and area, and count how often a monotonic >10% decrease lasting over one hour occurs; if that rate approaches 58%, the pattern is not specific to eruptive events.","tokens_in":16610,"feed_emoji":"☀️","tokens_out":11716,"duration_ms":120119,"temperature":0.7,"pith_summary":"Photospheric current helicity is the area integral of vertical current density times vertical magnetic field, $H_c=\\int j_zB_z\\,dS$. This paper tries to establish that this quantity shows a characteristic reversal around coronal mass ejections: a pre-eruption decrease and a post-eruption increase, driven by electric currents concentrating toward the polarity inversion line before eruption and separating away from it afterward. The authors reproduce the pattern in a 3D MHD simulation of the tether-cutting scenario and find it in observations of 50 eruptive flares, with 58% showing the decrease and 92% the increase. Two case studies show the same current redistribution as the model, and long-term helicity curves show the decrease many hours before the eruption, suggesting that $H_c$ could indicate when an active region has stored enough current to erupt and may help predict CMEs.","feed_headline":"Solar current helicity dips before CMEs, then rebounds","feed_subtitle":"Pattern tracks current buildup near the polarity inversion line, which may forecast CMEs.","key_machinery":"The central object is the photospheric current helicity $H_c=\\int j_zB_z\\,dS$, computed from magnetograms as the product of vertical electric current density and vertical magnetic field. The mechanism that carries the argument is the redistribution of $j_z$: before an eruption, electric currents that were spread over the active region concentrate toward the polarity inversion line, and after the eruption they move apart in a double-ribbon pattern that follows the flare ribbon separation. Because $B_z$ remains roughly stable, the reversal in $H_c$ is essentially the reversal of $j_z$ in the strong-$B_z$ regions, and the paper quantifies this through difference maps of $j_z$ and time series of the mean $j_z$ in the pole and PIL regions.","core_discovery":"The central discovery is that the photospheric current helicity reversal is not caused by a change in the overall magnitude of $j_z$ or $B_z$ but by a redistribution of electric current. In the pre-eruption phase, the double-J current structure contracts toward the polarity inversion line, reducing $j_z$ in the strong-$B_z$ regions that dominate $H_c$ and thus lowering the helicity. After the eruption, the current ribbons separate back toward the magnetic poles, increasing $j_z$ in the strong-field regions and raising the helicity, a pattern consistent with flare ribbon separation caused by the upward progression of the reconnection site. The same spatial pattern is seen in two observed eruptive flares and in the tether-cutting MHD simulation, and the pre-eruption decrease appears in the long-term helicity evolution of both cases, which the authors interpret as evidence that $H_c$ tracks the build-up of electric currents that power a CME.","pith_inferences":["Editorial inference: the 58% pre-eruption detection rate needs calibration against a control sample of non-erupting active regions to establish whether the decrease is specific to eruptive events or reflects the normal fluctuation level of $H_c$.","Editorial inference: the authors' mechanism predicts that maps of $j_z$ alone, without $B_z$ weighting, should show the same contraction and separation pattern; testing this directly would separate the geometric redistribution from any concurrent change in field strength.","Editorial inference: the 10% and 1-hour thresholds were fixed globally, so an operational application would likely need region-specific tuning, and the average decrease magnitude of 27% with durations sometimes exceeding the 6-hour window suggests the pattern may be stronger than the threshold-based statistics imply.","Editorial inference: the paper's caveat that total 3D coronal helicity behaves differently from photospheric $H_c$ implies that extending the analysis to time-dependent coronal extrapolations is an open test of whether the photospheric signal is a trustworthy proxy for the coronal energy buildup."],"forward_implications":["A sustained drop of more than 10% in $H_c$ lasting over an hour can flag an active region as approaching eruption conditions, complementing existing flare-prediction parameters.","The post-eruption increase appears in 92% of events, so $H_c$ may serve as a reliable marker for confirming that an eruption has taken place in magnetogram time series.","Because confined flares reportedly lack the clear pre-eruption decrease and post-eruption increase, the reversal pattern may help distinguish eruptive from confined flares in space-weather forecasting.","A single snapshot of $H_c$ or $\\alpha_{\\mathrm{weighted}}$ is not a reliable measure of eruptive potential, since a decrease in $H_c$ can accompany the current concentration that makes eruption more likely.","Monitoring $H_c$ over timescales of 10-15 hours, as in the two case studies, could extend warning times beyond those of flare-onset criteria."],"supporting_citations":[{"why":"Provides the tether-cutting MHD simulation in which shearing alone forms a current sheet near the PIL and initiates eruption, the model analyzed here.","marker":"Jiang et al. (2021)"},{"why":"The specific DARE-MHD simulation run used for the photospheric parameter analysis, from which this paper takes the first of its three consecutive eruptions.","marker":"Bian et al. (2022)"},{"why":"Observational study of 15 eruptive flares that found the same pre-eruption decrease and post-eruption increase in the $B_z$-weighted force-free parameter, which the present paper reinterprets in terms of current redistribution.","marker":"Liu et al. (2023)"},{"why":"Supplies the squashing-factor contour method used to build the flare mask that isolates the erupting flux region in the observational analysis.","marker":"Liu et al. (2017)"},{"why":"Explains the post-eruption increase in average current density via downward compression from reconnection, used as a secondary contributor to the helicity increase.","marker":"Bian & Jiang (2023)"},{"why":"Provides the noise level for HMI vector magnetograms that sets the 300 G field-strength threshold in the pixel selection.","marker":"Hoeksema et al. (2014)"}],"fun_headline_variants":["Helicity dip precedes CMEs, rebound follows","Current helicity reversal tracks CME cycle","Solar helicity flips before and after CMEs","Watch helicity: it predicts CME build-up","Pre-eruption helicity drop hints at CME risk"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The claim that the pre-eruption decrease is a genuine precursor rests on the untested assumption that current helicity in non-eruptive active regions does not also drop by more than 10% for over an hour at a comparable rate.","fun_headline_variants_meta":{"raw":{"variants":["Helicity dip precedes CMEs, rebound follows","Current helicity reversal tracks CME cycle","Solar helicity flips before and after CMEs","Watch helicity: it predicts CME build-up","Pre-eruption helicity drop hints at CME risk"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000228,"raw_usage":{"total_tokens":1516,"prompt_tokens":1026,"completion_tokens":490,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":642,"completion_tokens_details":{"reasoning_tokens":413}},"tokens_in":642,"tokens_out":490,"duration_ms":6277,"temperature":1.0,"reasoning_tokens":413,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-06T17:01:05.774559+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute $H_c$ time series for a sample of active regions that never produce an M5.0 or larger flare, matched to the 50 events in field strength and area, and count how often a monotonic >10% decrease lasting over one hour occurs; if that rate approaches 58%, the pattern is not specific to eruptive events.","supporting_citations":[{"cited_title":"S., & Leake , J","cited_arxiv_id":null,"evidence_quote":"Supplies the squashing-factor contour method used to build the flare mask that isolates the erupting flux region in the observational analysis."}],"review_version":1}