{"id":"d76beffc-0945-4d77-b71e-106ab1c7b904","arxiv_id":"1908.02226","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":2,"one_line_summary":"Using daily counts of the Mount Wilson classes, the authors find that complex active regions (βγ, βγδ) peak roughly two years later than simple ones (α, β) in both solar cycles 23 and 24, and that the complex-region peak count barely changes between cycles.","lead":"Sunspot groups with tangled magnetic fields appear on the Sun roughly two years after the simpler groups peak, in both solar cycles studied between 1996 and 2018. The result adds a timing constraint that could sharpen space weather forecasts and test dynamo models, though the proposed mechanism is not directly measured.","discovery_kind":"new_method","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The two-year SAR/CAR peak lag may result from lifetime-weighted daily counts; per-AR classification is not checked.","rationale":"The reader's weakest assumption already identifies the lifetime-weighting issue, and I agree it is the most load-bearing because the paper's main claim—a roughly two-year offset between SAR and CAR peaks—is computed from a time series in which each active region contributes one count per day rather than one count per emergence. The paper never compares this daily-count series to a per-AR series, so it cannot distinguish a genuine delay in complex-region formation from a delay introduced by the counting method. The proposed recomputation is a direct, low-cost check using the same public SRS data. If the lag and the cycle-to-cycle ratios survive the per-AR count, the central observational result stands; if not, the SSD/LSD conclusion is unsupported. This does not change the reader's conditional verdict: the paper should be revised to include this robustness test (and the other listed conditions) before full acceptance. I do not find a more serious concern beyond this, and the Table 2 arithmetic and autocorrelation issues are secondary.","tokens_in":11473,"tokens_out":10500,"duration_ms":116303,"concrete_test":"Recompute the monthly SAR and CAR series from the same SRS data using exactly one Mount Wilson class per AR per disk passage, e.g., the class reported on the day of that AR's maximum sunspot area (following Jaeggli & Norton 2016), and also using the most complex class reached during the disk passage. Apply the same 7-month running average, locate the SC 23 and SC 24 maxima, and compare the SAR-CAR lag and the SC23-to-SC24 peak ratios (Table 3). If the ~2-year lag and the 15% vs 51% drop persist, lifetime weighting is not responsible; if they weaken or disappear, the central result is an artifact of the daily count.","verdict_should_be":"UNCHANGED","load_bearing_attack":"Section 2's daily approach counts each active region on every day it appears in the SRS list, so the monthly CAR series is the sum over regions of the number of days each region is classified as βγ or βγδ. The paper itself notes this 'will emphasize long-lived ARs more than short-lived ones,' and §1 states that an AR's magnetic complexity 'might change from one class to another during a region's lifetime.' Because the Mount Wilson class can evolve across a disk passage, the daily-count series conflates the emergence rate of complex regions with the time each region spends in a complex class. If complex classes tend to be assigned late in a region's disk passage, or if the lifetime or classification duration of complex regions varies with cycle phase, the smoothed CAR curve can peak later than the underlying emergence of complex ARs, and the cycle-to-cycle amplitude can be distorted. The paper's central claims—the ~2-year SAR/CAR peak lag and the near-equal CAR peak counts (Table 3)—are read directly from this lifetime-weighted series, with no check against a per-AR count. This is the load-bearing assumption because the LSD/SSD interpretation depends on the lag and on CAR counts being cycle-independent. If a per-AR count removes the lag or the cycle-independence, the paper's main conclusions collapse.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper analyzes NOAA/SRS daily active region (AR) data from January 1996 to December 2018, assigning each AR a Mount Wilson complexity class per day. It introduces a 'daily magnetic complexity approach' in which each region contributes one count for every day it appears in a given class, and then groups \\u03b1 and \\u03b2 classes as simple active regions (SARs) and \\u03b2\\u03b3 and \\u03b2\\u03b3\\u03b4 as complex active regions (CARs). The main findings are: (i) monthly SAR numbers track the NOAA sunspot number (NSN) closely (r=0.99) and peak during the first NSN peak in each cycle (May 2000 in SC 23, December 2011 in SC 24); (ii) monthly CAR numbers peak about two years later, during the second NSN peak (November 2001 and January 2014); (iii) the total number of CARs in a four-year window around the CAR peak drops only 15% from SC 23 to SC 24, versus 51% for SARs (Table 3); (iv) CARs peak before the latitudinal width of the activity band starts to decrease (Fig. 8). The authors interpret these results as evidence for competition between a cyclic large-scale dynamo (LSD) and a cycle-independent small-scale dynamo (SSD), with the SSD gaining relative influence in the declining phase and producing the late, cycle-independent CAR peak. No model fitting is performed; all results are direct data reads.","tokens_in":2028,"tokens_out":2566,"duration_ms":64739,"significance":"If the reported peak lag and the near-cycle-independence of CAR counts hold, the paper would provide a new observational constraint on active region formation, suggesting that complex regions are a late-cycle phenomenon whose total rate is nearly independent of cycle strength. This would challenge the earlier Jaeggli & Norton (2016) interpretation that complex regions appear because the latitudinal activity band narrows, and it would support a surface small-scale dynamo contribution. The paper is laudably simple, uses a public dataset exclusively, fits no free parameters, and reports quantitative KS tests for the latitudinal distributions. Its central interpretation, however, rests on the unvalidated daily-count weighting: the two-year lag and the Table 3 cycle-to-cycle comparison are read from a lifetime-weighted series, and the 'peak before width decrease' claim in Fig. 8 is not statistically tested. These gaps do not invalidate the direct reading of the smoothed curves, but they materially weaken the causal interpretation.","major_comments":[{"comment":"The daily magnetic complexity approach counts each active region once for every day it is classified as α, β, βγ, or βγδ. As the paper acknowledges, this 'will emphasize long-lived ARs more than short-lived ones,' and §1 notes that an AR's complexity can change during its lifetime. The monthly CAR series is therefore not a measure of the emergence rate of complex regions but a lifetime-weighted sum of the days regions spend in complex classes. If complex classifications tend to be assigned late in a region's disk passage, or if the lifetime or classification duration of complex regions changes with cycle phase, the smoothed CAR curve can peak later than the underlying emergence and the cycle-to-cycle amplitude can be distorted. The central claims (the ~2-year SAR/CAR lag and the near-equal CAR peak counts in Table 3) are read directly from this series, with no per-AR control. A robustness check that counts each region once (e.g., at first appearance, at maximum area, or at maximum complexity) is necessary to support the LSD/SSD interpretation.","section":"§2 and §3.2"},{"comment":"The analysis treats the Mount Wilson classifications in the NOAA/SRS list as a stable, unbiased measure of magnetic complexity over 22 years. The paper does not discuss instrument or observer changes in the SOON network, nor does it test whether classification practice drifted over 1996-2018. A time-dependent classification threshold could produce an apparent phase lag or alter cycle-to-cycle amplitudes, directly affecting the main claims. Please provide at least a qualitative discussion of known data caveats, and preferably a consistency check against an independent classification source for a subset of the period.","section":"§2"},{"comment":"The Pearson correlations r=0.99 (SARs vs NSN) and r=0.87 (CARs vs NSN) are computed on seven-month moving averages without correcting for autocorrelation. With heavy smoothing, the effective number of independent samples is far below the number of monthly points, so the significance of these coefficients is overstated. In addition, r=0.99 for SARs is near-tautological because SARs comprise about 88% of all ARs and sunspot numbers are strongly driven by AR counts. The paper should report confidence intervals based on effective degrees of freedom, or correlations on unsmoothed or detrended data, and should avoid implying that 0.99 vs 0.87 is evidence of a different cycle relationship.","section":"§3.1 and Fig. 5"},{"comment":"The statement that CARs 'peaked before the latitudinal width starts to decrease' is an assessment of the figure with no quantitative test. There are no error bars, no peak-timing statistic, and no comparison against a null hypothesis (e.g., simultaneous peaking). Since this result is used to reject the Jaeggli & Norton (2016) hypothesis that complex regions increase because the latitude band narrows, it should be supported by a formal analysis, such as a cross-correlation of the two time series or a segmented regression on the width curve.","section":"§3.4 and Fig. 8"},{"comment":"The 'rate of change' from SC 23 to SC 24 is computed from four-year windows centered on the respective CAR and SAR peaks, but the windows are not symmetric in cycle phase and the two cycles have different shapes. The 15% drop in CAR counts versus 51% for SARs is presented without uncertainty. Given that monthly CAR peaks are around 50 (Fig. 5), a 15% difference may be within the noise, especially after smoothing. The paper should provide an estimate of the uncertainty on the peak-window sums (e.g., bootstrap on the unsmoothed monthly counts) before interpreting the near-equality as evidence for a cycle-independent SSD.","section":"Table 3"}],"minor_comments":[{"comment":"The abstract contains small language errors: 'active region' should be 'active regions' in the Aims sentence, and 'decease' should be 'decrease' in the Results sentence.","section":"Abstract"},{"comment":"The phrase 'the period of two years before and after their maximum values' describes a four-year period; please clarify whether the maximum month is included in the window and whether the window is the same for both groups in each cycle.","section":"§3.2 and Table 3"},{"comment":"The shaded bars in Fig. 6 are described in the text as 'the periods when the largest difference between the time variation of SARs and the NSN happens,' but the caption does not explain how these periods are identified. Please specify the criterion used to draw the bars.","section":"Fig. 6"},{"comment":"The 'double peak' behavior is described qualitatively. It would be helpful to state how the two peaks are identified, for example whether they are local maxima in the smoothed curve with a minimum between them.","section":"§3.1 and Fig. 5"},{"comment":"The HELIO website is cited by URL in a footnote; please add a formal reference or access date for the dataset version used.","section":"§2"},{"comment":"The hemispheric curves in Fig. 7 may be difficult to distinguish in grayscale; using different line styles (solid vs dashed) or labels inside the panels would improve clarity.","section":"§3.3 and Fig. 7"}],"recommendation":"major_revision","confidential_remarks":"The paper is a straightforward observational A&A article with a clear central result that appears to be genuine as a direct reading of the data. The main weakness is not the observational claim itself but the gap between the claim and the interpretation: the daily-count weighting and the classification stability assumptions are load-bearing for the LSD/SSD conclusion, and the paper does not provide per-AR or consistency checks. The Fig. 8 claim against Jaeggli & Norton is also under-quantified. These issues are fixable with additional analysis within the scope of the manuscript, so I recommend major revision rather than rejection. I would also ask the editor to consider whether the novelty of the 'daily approach' is sufficiently highlighted relative to the Jaeggli & Norton per-AR method; the difference matters methodologically but is only discussed briefly."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Dear Colleague,\n\nThe headline result—complex active regions (CARs) peak roughly two years after simple ones (SARs)—is a genuine pattern in the daily Mount Wilson counts from the SRS list. It survives the choice of smoothing window, and it is new: Jaeggli & Norton (2016) used one class per region at maximum area and did not see a lag. The paper also reports that the CAR peak is nearly the same in cycles 23 and 24 (15% drop vs 51% for SARs), and that CARs peak before the activity band narrows. Those are useful observational constraints for dynamo modeling.\n\nWhat the paper does well: it uses public NOAA/SRS data, describes the daily counting method clearly, and is upfront that this weights long-lived regions more heavily. The comparison with earlier work is fair.\n\nThe soft spots are real, and one of them is load-bearing. The daily count is the sum over regions of the number of days each region is classified as complex. Since a region's class can evolve—often from simple to complex as it develops—the CAR time series mixes emergence with the duration of the complex phase. If complex classes tend to be assigned late in a region's disk passage, or if that duration varies over the cycle, the two-year lag and the near-constant CAR peak could be artifacts. The paper does not test this with a per-AR count, say assigning each region once at its maximum complexity or at first complex classification. That is the first thing I would ask for.\n\nOther issues are minor. The r=0.99 correlation between SARs and sunspot number is partly tautological since SARs are ~88% of all ARs, and smoothing without autocorrelation makes the Pearson values overconfident. The Table 2 total (33,496 in the text vs 33,671 in the table) is inconsistent. The Fig. 8 timing claim is read off the plot with no quantitative test. And the dynamo interpretation is presented as a conclusion in the abstract; it is a hypothesis and should be framed that way.\n\nNone of this kills the central observation as a property of the daily-count series. But the physical interpretation hinges on whether the lag reflects the emergence of complex regions or simply the timing of their classification. That is a substantive, testable revision, not a desk reject.\n\nI would send this to a serious referee and ask for the per-AR check. If that check preserves the lag, the paper becomes a solid contribution.\n\nBest,","headline":"The two-year SAR/CAR lag is a real pattern in the daily SRS counts, but lifetime-weighted counting could produce it; the paper deserves revision and peer review.","tokens_in":12261,"tokens_out":3864,"would_cite":false,"duration_ms":40824,"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":"Complex active regions lag simple ones by about two years in each solar cycle, and their peak abundance is nearly identical across the two cycles.","keywords":["solar active regions","Mount Wilson classification","magnetic complexity","solar cycle","sunspot number","small-scale dynamo","large-scale dynamo","solar activity cycles"],"falsifier":"Compute monthly simple- and complex-region counts using a single complexity class per region, taken from the day of its maximum sunspot area; if the complex-region peak then coincides with the sunspot-number peak rather than lagging by two years, the lag is a product of the daily weighting. Splitting the record by observing era would further expose a classification-drift artifact if the two-year lag disappears in one sub-interval.","tokens_in":11240,"feed_emoji":"☀️","tokens_out":12179,"duration_ms":117546,"temperature":0.7,"pith_summary":"Using more than two decades of daily Mount Wilson classifications of sunspot groups, the paper finds a clean separation in how simple and complex active regions behave over the solar cycle. Simple regions ($\\alpha$ and $\\beta$) track the sunspot number almost perfectly and peak at the first maximum of each cycle, while complex regions ($\\beta\\gamma$ and $\\beta\\gamma\\delta$) peak roughly two years later in both cycles studied. The peak number of complex regions barely changes from cycle 23 to cycle 24, while the simple-region peak falls by about half. The authors read this as evidence that active-region formation is a competition between a cyclic large-scale dynamo, which favors simple regions at solar maximum, and a cycle-independent small-scale dynamo, whose relative influence grows in the declining phase and favors complex regions. The result matters because it links a routine sunspot classification to the internal dynamo balance and identifies the late maximum as the time of greatest complex-region abundance.","feed_headline":"Two-year lag: complex sunspot regions peak after simple ones","feed_subtitle":"Their peak counts barely change between cycles, pointing to a cycle-independent dynamo effect.","key_machinery":"The central object is the daily magnetic-complexity count: instead of assigning one complexity class to each active region at a single moment, the method adds one count of the region's current Mount Wilson class ($\\alpha$, $\\beta$, $\\beta\\gamma$, or $\\beta\\gamma\\delta$) for every day the region appears on the disk. This lifetime-weighted daily counting is what separates the SAR and CAR time series and reveals the two-year lag, because a region that spends part of its life simple and part complex contributes to both groups. The comparison series is the monthly sunspot number, smoothed with a seven-month running average, and the load-bearing statistics are the peak timings, the correlation coefficients, and the cycle-to-cycle ratios of peak counts.","core_discovery":"Using a newly developed daily counting method, in which each active region contributes one count per day and may change complexity class during its lifetime, the paper shows that monthly counts of simple active regions (SARs, classes $\\alpha$ and $\\beta$) track the reference sunspot number with correlation $r=0.99$ and peak in May 2000 during cycle 23 and December 2011 during cycle 24. Complex active regions (CARs, classes $\\beta\\gamma$ and $\\beta\\gamma\\delta$) instead peak in November 2001 and January 2014, about two years later in each cycle, with a weaker correlation of $r=0.87$. From the two-year window around each peak, the CAR count drops only 15% between cycles 23 and 24, while the SAR count drops 51%; the same pattern holds over the full cycles, with drops of 32% and 45%. In both hemispheres the CAR peak occurs before the latitudinal width of the activity band starts to decrease, which the paper takes as evidence against the idea that complex regions appear because the activity band narrows. The paper concludes that the two dynamo processes compete during flux emergence: the cyclically varying large-scale dynamo dominates at solar maximum and produces mostly simple regions, while the cycle-independent small-scale dynamo becomes relatively more important in the declining phase and makes an increasing share of emerging flux complex.","pith_inferences":["A natural test of the two-year lag is the next solar cycle: if the competition story is right, the complex-region peak of cycle 25 should again trail the sunspot-number peak by roughly two years.","Recomputing the counts with one classification per region, taken from the day of maximum sunspot area, would isolate whether the daily lifetime-weighting method itself creates the lag; if the lag disappears, the dynamo interpretation would rest on a counting artifact.","The near-constant complex-region peak suggests a complexity floor that is nearly independent of cycle strength; if real, the ratio of simple to complex peak counts is a more sensitive dynamo diagnostic than either count alone."],"forward_implications":["Monthly counts of simple active regions become a near-perfect proxy for the sunspot cycle, so intervals where that correlation breaks down mark times when the complex-region contribution is changing.","The two-year lag means each cycle's late maximum and early declining phase, not its peak, is when complex active regions are most abundant.","The 15% versus 51% decline in peak counts implies that complex-region production is far more stable across cycles than bulk sunspot activity.","Since complex regions are the more flare- and CME-productive class, the late-cycle CAR peak implies elevated eruptive activity roughly two years after sunspot maximum.","The CAR peak occurring before the activity band narrows argues against the previously suggested mechanism that latitudinal compression creates complex regions."],"supporting_citations":[{"why":"Baseline complexity-cycle statistics and the latitudinal-width hypothesis that this paper tests and rejects.","marker":"Jaeggli & Norton 2016"},{"why":"Introduces the Mount Wilson complexity classes from which the simple/complex grouping is built.","marker":"Hale et al. 1919"},{"why":"Supplies the current class definitions and naming table used to categorize regions.","marker":"van Driel-Gesztelyi & Green 2015"},{"why":"Describes the observing network whose daily sunspot reports feed the active-region summary list.","marker":"Balasubramaniam & Henry 2016"},{"why":"Documents the operational set of complexity classes the network assigns each day.","marker":"Andrus 2013"},{"why":"Provides the sunspot-number reference series used for all cycle comparisons.","marker":"Hathaway 2015"},{"why":"Foundational theory of the small-scale turbulent dynamo invoked as the cycle-independent process.","marker":"Batchelor 1950"},{"why":"Foundational theory of the large-scale cyclic dynamo invoked as the cycle-varying process.","marker":"Parker 1955"},{"why":"Support for the claim that small-scale dynamo fields do not vary with the solar cycle.","marker":"Petrovay & Szakaly 1993"},{"why":"Observational evidence that small-scale surface magnetic fields are cycle-independent, supporting the SSD interpretation.","marker":"Buehler et al. 2013"}],"fun_headline_variants":["Complex sunspots peak two years after simple ones","Two-year lag in peak of complex active regions","Complex active regions peak later and change less","Sunspot complexity peaks late, resists cycle decline","Cycle-independent complex regions peak with delay"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The analysis assumes that daily Mount Wilson complexity labels in the 22-year record are assigned consistently over time, and that counting each region once per day does not systematically shift complex regions to later in their lifetimes; if either fails, the two-year lag and the cycle-to-cycle stability of complex counts are artifacts.","fun_headline_variants_meta":{"raw":{"variants":["Complex sunspots peak two years after simple ones","Two-year lag in peak of complex active regions","Complex active regions peak later and change less","Sunspot complexity peaks late, resists cycle decline","Cycle-independent complex regions peak with delay"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000713,"raw_usage":{"total_tokens":3298,"prompt_tokens":1124,"completion_tokens":2174,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":740,"completion_tokens_details":{"reasoning_tokens":2105}},"tokens_in":740,"tokens_out":2174,"duration_ms":19233,"temperature":1.0,"reasoning_tokens":2105,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-14T14:52:00.594605+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Compute monthly simple- and complex-region counts using a single complexity class per region, taken from the day of its maximum sunspot area; if the complex-region peak then coincides with the sunspot-number peak rather than lagging by two years, the lag is a product of the daily weighting. Splitting the record by observing era would further expose a classification-drift artifact if the two-year lag disappears in one sub-interval.","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Baseline complexity-cycle statistics and the latitudinal-width hypothesis that this paper tests and rejects."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Describes the observing network whose daily sunspot reports feed the active-region summary list."},{"cited_title":"2013, Air Force Weather Agency Manual, http://static","cited_arxiv_id":null,"evidence_quote":"Documents the operational set of complexity classes the network assigns each day."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Foundational theory of the small-scale turbulent dynamo invoked as the cycle-independent process."},{"cited_title":"& Szakaly, G","cited_arxiv_id":null,"evidence_quote":"Support for the claim that small-scale dynamo fields do not vary with the solar cycle."}],"review_version":1}