{"id":"040d0c4d-e7ed-4a8b-b3f8-23fe12e1cdb1","arxiv_id":"2501.02935","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":4,"one_line_summary":"A common charge-state based solar wind classification misidentifies slow solar wind as coronal hole wind during solar minimum because its fixed oxygen ratio threshold does not track the solar cycle.","lead":"The paper compares three methods for identifying coronal hole wind in ACE solar wind data and finds that a widely used charge-state based scheme mislabels slow solar wind as coronal hole wind during solar minimum. It shows the problem comes from fixed thresholds that ignore solar cycle changes, and suggests including the solar cycle phase in future classifications.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The claim that Zhao et al. (2009) misidentifies slow solar wind in 2009 relies on the unvalidated k-means and Xu & Borovsky references as ground truth; without independent validation of these references, the disagreement alone cannot establish misidentification.","rationale":"The paper is a careful comparative study and its main observation—that the three schemes disagree sharply in 2009—is well documented with figures and data. The authors also make several good-faith gestures: they provide source code, acknowledge the absence of ground truth, and support the low-2009-fraction interpretation with external coronal-hole area observations from Fujiki et al. (2016). Their SHAP sanity check for Zhao (only the O ratio matters by construction) shows the explainability tool works as intended. However, the central inference from 'the schemes disagree' to 'Zhao is wrong' requires that the k-means/Xu & Borovsky references be more trustworthy than Zhao. That trust is not established. The k-means classifier is trained on the same ACE 2001-2010 data and k=7 is chosen from prior work without sensitivity analysis; the Xu & Borovsky scheme is itself a fixed-boundary method with a known selection bias toward recurrent high-speed streams. Because the paper's own Section 1 states no ground truth exists, the conclusion should be framed as a conditional hypothesis. The proposed test—training on one epoch and testing on another, plus k-sensitivity—would directly probe the stability of the reference. Until such a test is run, the CONDITIONAL verdict is appropriate, not ACCEPT.","tokens_in":15380,"tokens_out":4818,"duration_ms":43486,"concrete_test":"Retrain the 7-means classifier on ACE data from 2001-2004 only (solar maximum) and apply it to 2009 data, across k=5,6,7,8 and several feature subsets. If the coronal hole fraction in 2009 is not robustly low in all configurations, the reference classification itself is unstable and the conclusion that Zhao misidentifies slow wind in 2009 is not supported.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's central conclusion ('the Zhao et al. (2009) scheme likely misidentifies slow solar wind as coronal hole wind during the solar activity minimum') is reached by comparing Zhao's 2009 coronal hole fraction against the low fractions from Xu & Borovsky (2015) and a 7-means classifier. This inference is load-bearing: if the reference classifications are wrong about 2009, the claim collapses. The paper explicitly notes (Section 1) that 'a ground truth for solar wind categorization is still not available,' and the 7-means classifier was 'trained on the ACE data from 2001-2010' (Section 3), i.e., on the same data used for the comparison. The Xu & Borovsky scheme is also a fixed-threshold scheme and is admitted to be 'better adjusted for solar activity minimum than maximum conditions,' so its low 2009 fraction is not an independent standard. The SHAP analysis of Zhao in 2009 (Section 4.2) shows low feature importance for the O ratio, but this is a tautological consequence of nearly all data falling below the fixed threshold, not evidence of misclassification. The external support from Fujiki et al. (2016) is qualitative and not quantitatively connected to in-situ fractions. Thus, the central claim is currently a plausible interpretation of a three-way disagreement, not an established misidentification.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"This paper addresses the discrepancy between three solar wind categorization schemes in the fraction of solar wind identified as coronal hole wind in ACE data during 2001–2010, focusing on the deep solar minimum around 2009. The schemes are the charge-state based Zhao et al. (2009) threshold on O7+/O6+, the proton-plasma based Xu & Borovsky (2015) scheme, and a 7-means clustering trained on the same ACE data. The authors show that in 2009 the Zhao scheme classifies nearly all solar wind as coronal hole wind, while the other two schemes identify very little. They argue that the Zhao threshold is fixed while the O7+/O6+ ratio for all solar wind drops systematically during solar minimum, so that slow solar wind falls below the threshold and is misclassified. They also identify two k-means coronal hole types, CH1 and CH2, with different charge-state and speed properties, and use SHAP values to show which parameters drive each classification. They conclude that origin-oriented solar wind classification should take the solar cycle phase into account.","tokens_in":15644,"tokens_out":7487,"duration_ms":68552,"significance":"The paper identifies a concrete, physically plausible mechanism—solar-cycle modulation of the O charge-state baseline interacting with a fixed threshold—and supports it with multi-year histograms. If correct, the result has practical impact on solar wind classification and on the interpretation of composition-based solar wind statistics during solar minima. Strengths include the explicit acknowledgment that no ground truth exists (Section 1), the use of a data-driven reference in addition to two established schemes, the availability of data (Berger et al. 2023) and code (Teichmann et al. 2023), and the falsifiable recommendation to include the solar cycle phase in classification. The main weaknesses are the reliance on unvalidated reference classifications for the conclusion and the lack of quantitative uncertainties and sensitivity checks.","major_comments":[{"comment":"The central claim that the Zhao scheme 'misidentifies some slow solar wind' is supported only qualitatively. The authors show that the O7+/O6+ distribution drops below the Zhao threshold in 2008–2009 and that the Zhao coronal hole subset in 2009 contains low vp and Tp, but they do not quantify the fraction of Zhao-identified coronal hole wind with slow-wind-like properties or compare it to the fraction in years of good agreement (e.g., 2003). A quantitative decomposition, such as the fraction of Zhao coronal hole wind with vp below a slow-wind threshold or a two-dimensional density plot of vp versus O7+/O6+ in 2009, would make the misclassification claim testable. As written, 'likely misidentifies' rests on visual inspection.","section":"Section 4.1, Fig. 3"},{"comment":"The conclusion that the Zhao scheme overestimates coronal hole wind in 2009 treats the Xu & Borovsky (2015) and 7-means classifications as references, yet the paper states that no ground truth exists (Section 1) and that the 7-means classifier was trained on the same 2001–2010 ACE data used in the comparison (Section 3, 'The 7-means classification was trained on the ACE data from 2001-2010'). The Xu & Borovsky scheme is also a fixed-threshold scheme and is admitted to be 'better adjusted for solar activity minimum than maximum conditions' (abstract). The disagreement alone therefore does not identify which scheme is correct. The authors should either validate the reference classifications against an external benchmark (e.g., coronal hole source mapping or the Fujiki et al. coronal hole areas) or explicitly frame the conclusion as conditional on the data-driven reference. The current phrasing overstates the evidence.","section":"Section 3 and Section 1"},{"comment":"The fractions and counts (e.g., 153594 vs 74188 vs 33702/31070 data points) are reported without uncertainties or significance tests. The 10-Fe-ion threshold (Section 2.1) and SWICS counting statistics introduce selection effects and sampling noise; the authors even note that the threshold biases against dilute polar coronal hole wind. Without bootstrap or Poisson uncertainties on the per-Carrington-rotation fractions, the reader cannot assess whether the 2009 differences are statistically significant or dominated by sampling. This is important because the entire paper is about a difference in fractions.","section":"Figure 1 and Section 3"}],"minor_comments":[{"comment":"The phrase 'identification of the coronal wind' should be 'identification of the coronal hole wind' to match the rest of the paper.","section":"Abstract"},{"comment":"There is a typo: 'Shapeley values' should be 'Shapley values', and in the footnote 'loosing' should be 'losing'.","section":"Section 2.2"},{"comment":"The word 'collosional' should be 'collisional' in the description of the proton-proton collisional age.","section":"Section 3"},{"comment":"The description of which 7-means type dominates during solar maximum/minimum is internally inconsistent with Section 4 and the abstract. Section 4 states 'CH1 is dominant during the solar activity minimum whereas CH2 is more frequent during the solar activity maximum,' but the last paragraph of Section 4.1 states 'CH1 represents properties of smaller coronal holes dominant during the solar activity maximum and CH2 represents properties of larger coronal holes that are dominant during the solar activity minimum.' This contradiction should be corrected.","section":"Section 4.1, last paragraph"},{"comment":"The statement that 'ACE is probably connected more frequently to equatorial coronal holes in 2003' is plausible but no connectivity analysis is shown; consider adding a reference or a simple Parker-spiral connection estimate to support this interpretation.","section":"Section 4.3"},{"comment":"The sentence 'The MAG and SWEPAM data is binned to the native 12-min time resolution of SWICS' has a subject-verb agreement error; it should be 'data are binned'.","section":"Section 2.1"}],"recommendation":"major_revision","confidential_remarks":"The paper is within the scope of solar wind physics and the comparison is useful. The central result is not entirely new given that Wang & Ko (2019) already noted the high Zhao fraction in 2009, but the systematic mechanism and multi-method comparison add value. The biggest risk is overstatement of the conclusion relative to the uncertainty in the reference classification. A revision that quantifies the misclassified fraction, provides uncertainties on the fractions, and tests sensitivity to k and to the reference choice would strengthen the paper considerably."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"Short version: this paper makes a solid case that the Zhao et al. (2009) charge-state scheme badly overclassifies coronal hole wind during solar minimum, and the case is stronger than the stress-test note allows. The key evidence is not just that three classifiers disagree in 2009. Their Figure 3 shows the whole O7+/O6+ distribution dropping below the fixed threshold, and the Zhao coronal hole wind in 2009 is slow and cold in a way that no other coronal hole type is. That within-class inconsistency carries the argument, independent of the k-means reference.\n\nWhat is genuinely new here is the 7-means reference splitting coronal hole wind into CH1 and CH2 with opposite solar cycle dominance, the time-resolved SHAP analysis showing which parameters matter when, and the practical suggestion to build the solar cycle phase into future classification schemes. The authors also ship the k-means source and the merged ACE dataset, so the comparison is reproducible. And they are honest about the missing ground truth: they call k-means a data-driven reference, not the answer, and they use Fujiki et al.'s coronal hole area maps as external support for the low 2009 fraction. That is the right way to handle a messy classification problem.\n\nThe soft spots are real but not fatal. There are no uncertainties on the type fractions. The k-means reference is trained on the same 2001-2010 ACE data it is used to judge, and k=7 comes from earlier work rather than from an internal criterion. The SHAP result for Zhao in 2009, where the O ratio loses feature importance, is close to tautological once nearly everything sits below the threshold; to the authors' credit, they say exactly that in Section 4.2. Quantifying the fractions' uncertainties and validating the clusters on independent data (PSP or Ulysses) would turn a good paper into a stronger one.\n\nWho is this for: anyone who uses these solar wind labels, especially people who have relied on Zhao et al. (2009) to identify coronal hole wind near solar minimum. It is not a breakthrough, but it is a careful, reproducible critique that should change practice.\n\nRecommendation: send it to peer review. I would expect the referees to ask for error estimates and maybe one independent validation, not a rewrite.","headline":"Solid evidence that the Zhao et al. (2009) O7+/O6+ threshold overclassifies slow wind as coronal hole wind near solar minimum; worth refereeing.","tokens_in":16219,"tokens_out":3134,"would_cite":true,"duration_ms":109289,"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 paper argues that a standard charge-state classifier's fixed O7+/O6+ threshold mistakes slow solar wind for coronal hole wind during solar activity minima, especially 2009.","keywords":["solar wind classification","coronal hole wind","charge-state composition","O7+/O6+ ratio","solar activity cycle","k-means clustering","ACE data"],"falsifier":"Map the solar wind that the composition scheme labels as coronal hole wind in 2009 (O7+/O6+ below 0.145) back to its solar source using coronal hole and streamer-belt observations; if a substantial fraction traces to streamer-belt plasma, the misidentification claim is confirmed, while if it traces to faint low-latitude coronal holes, the claim fails.","tokens_in":15139,"feed_emoji":"☀️","tokens_out":8273,"duration_ms":73361,"temperature":0.7,"pith_summary":"Solar wind from coronal holes is usually identified by two established schemes, one based on the O7+/O6+ charge-state ratio and one based on proton plasma properties. This paper argues that the charge-state scheme, which applies a fixed threshold of 0.145, mistakes slow solar wind for coronal hole wind during the solar activity minimum because the O7+/O6+ ratio of all solar wind falls systematically at minimum. The evidence is that the proton-plasma scheme and an unsupervised 7-means classification both find almost no coronal hole wind at ACE in 2009, while the charge-state scheme finds almost all wind to be coronal hole wind, and that very few low-latitude coronal holes were present in 2009. If correct, this means established origin-based solar wind classifications need to be revisited and should take the phase of the solar activity cycle into account.","feed_headline":"Oxygen ratio drop mislabels slow wind as coronal hole wind","feed_subtitle":"A fixed charge-state threshold calls almost all 2009 solar wind coronal hole wind; proton and cluster schemes disagree.","key_machinery":"The central object is the O7+/O6+ charge-state ratio threshold of 0.145 used by the Zhao et al. (2009) scheme. The paper compares this with the proton-plasma decision boundaries of Xu & Borovsky (2015) and with a 7-means clustering trained on ACE data that produces two coronal hole types. SHAP values are used to show which parameters drive each classifier over time, and the O7+/O6+ ratio's systematic solar-cycle variation is the mechanism that causes the fixed threshold to fail.","core_discovery":"The paper's central claim is that the Zhao et al. (2009) composition scheme misidentifies slow solar wind as coronal hole wind in solar activity minimum, particularly 2009. It grounds this in the systematic drop of O7+/O6+ for all solar wind during the minimum, so the fixed threshold is crossed by all wind. It uses the Xu & Borovsky (2015) proton-plasma scheme and a k-means clustering with two coronal hole types as references; both give low coronal hole wind fractions in 2009, consistent with the scarcity of low-latitude coronal holes. The paper also finds that the k-means classification yields two coronal hole types, which differ in charge states and in their occurrence over the activity cycle, and that SHAP explainability values show the oxygen ratio is essentially the only parameter driving the composition scheme, so the threshold's drift explains the inflated coronal hole wind fraction.","pith_inferences":["The paper does not build an adjusted classifier, but its results imply that making the O7+/O6+ threshold track the solar cycle would shrink the composition scheme's minimum-period coronal hole wind fraction.","The two k-means coronal hole types are interpreted as polar/weak-footpoint and equatorial/strong-footpoint populations only indirectly; a direct source-mapping comparison could test that identification.","The same fixed-threshold-versus-cycle-drift tension could affect other composition-based solar wind categories, such as Fe charge-state bins, which the paper does not examine."],"forward_implications":["The widely cited conclusion that 2009 solar wind was almost entirely coronal hole wind, based on the composition scheme, should be treated as an artifact of a fixed threshold rather than a physical fact.","Studies that use the composition scheme across solar minima will have coronal hole wind samples contaminated by slow solar wind, so their statistics on coronal hole wind properties need re-examination.","Solar wind classification schemes should include the phase of the solar activity cycle explicitly, for example by using cycle-dependent decision boundaries.","The k-means result provides a data-driven way to separate two coronal hole wind populations, one associated with weaker footpoint fields and one with stronger footpoint fields, which could be used in future classifications.","The near-zero coronal hole wind fraction at ACE in 2009 is physically plausible because very few low-latitude coronal holes were observed then, not necessarily a sign that the proton-plasma scheme is broken."],"supporting_citations":[{"why":"Defines the charge-state composition scheme and the fixed O7+/O6+ threshold of 0.145 that is the central object under scrutiny.","marker":"Zhao et al. (2009)"},{"why":"Supplies the proton-plasma based scheme whose coronal hole wind fraction in 2009 is nearly zero, providing the main comparison baseline.","marker":"Xu & Borovsky (2015)"},{"why":"Provides the k-means setup and the choice k=7 used to generate the unsupervised reference classification.","marker":"Heidrich-Meisner & Wimmer-Schweingruber (2018)"},{"why":"Supports the use of unsupervised k-means clustering as a data-driven reference for solar wind categorization.","marker":"Bloch et al. (2020)"},{"why":"Previously addressed the high 2009 coronal hole wind fraction in the Zhao scheme and connects O charge states to footpoint magnetic field strength.","marker":"Wang & Ko (2019)"},{"why":"Provides coronal hole area observations showing almost no low-latitude coronal holes in 2009, making the low coronal hole wind fraction plausible.","marker":"Fujiki et al. (2016)"},{"why":"Earlier comparison reported good agreement between the schemes but only for the 2001-2004 period, which this paper contrasts with the 2009 minimum.","marker":"Neugebauer et al. (2016)"},{"why":"Documents the systematic variation of solar wind charge states with the solar activity cycle, a premise for the threshold drift argument.","marker":"Zurbuchen et al. (2002)"}],"fun_headline_variants":["Oxygen ratio drop inflates coronal hole wind fraction","Fixed charge-state cutoff misreads slow wind as coronal holes","Solar cycle phase should guide coronal hole wind ID","Composition scheme overestimates coronal holes in minimum"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The load-bearing premise is that the 7-means classification, trained on the same ACE data with k=7, is a valid reference for what coronal hole wind actually is; the paper concedes that no ground truth exists, so if the clusters are wrong the misidentification conclusion weakens.","fun_headline_variants_meta":{"raw":{"variants":["Oxygen ratio drop inflates coronal hole wind fraction","Fixed charge-state cutoff misreads slow wind as coronal holes","Solar cycle phase should guide coronal hole wind ID","Composition scheme overestimates coronal holes in minimum"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000575,"raw_usage":{"total_tokens":2765,"prompt_tokens":1050,"completion_tokens":1715,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":666,"completion_tokens_details":{"reasoning_tokens":1648}},"tokens_in":666,"tokens_out":1715,"duration_ms":13590,"temperature":1.0,"reasoning_tokens":1648,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T21:59:31.429209+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Map the solar wind that the composition scheme labels as coronal hole wind in 2009 (O7+/O6+ below 0.145) back to its solar source using coronal hole and streamer-belt observations; if a substantial fraction traces to streamer-belt plasma, the misidentification claim is confirmed, while if it traces to faint low-latitude coronal holes, the claim fails.","supporting_citations":[{"cited_title":"H., & Fisk, L","cited_arxiv_id":null,"evidence_quote":"Defines the charge-state composition scheme and the fixed O7+/O6+ threshold of 0.145 that is the central object under scrutiny."},{"cited_title":"& Wimmer-Schweingruber, R","cited_arxiv_id":null,"evidence_quote":"Provides the k-means setup and the choice k=7 used to generate the unsupervised reference classification."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supports the use of unsupervised k-means clustering as a data-driven reference for solar wind categorization."},{"cited_title":"& Ko, Y .-K","cited_arxiv_id":null,"evidence_quote":"Previously addressed the high 2009 coronal hole wind fraction in the Zhao scheme and connects O charge states to footpoint magnetic field strength."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Earlier comparison reported good agreement between the schemes but only for the 2001-2004 period, which this paper contrasts with the 2009 minimum."},{"cited_title":"2002, Geophysical Research Letters, 29, 66 13","cited_arxiv_id":null,"evidence_quote":"Documents the systematic variation of solar wind charge states with the solar activity cycle, a premise for the threshold drift argument."}],"review_version":1}