{"id":"cc1d6cbb-2d0c-4ad5-a40a-8a7dd47d3786","arxiv_id":"2608.03799","paper_version":1,"verdict":"CONDITIONAL","confidence":"HIGH","novelty_score":6.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"During El Niño, a shoaling 26°C isotherm in the tropical central-north Pacific reduces subsurface heat and offsets the SST-driven increase in tropical cyclone genesis.","lead":"During El Niño, the ocean surface warms across the central-north Pacific, but the warm layer becomes thinner as the 26°C isotherm shoals, and this subsurface cooling limits tropical cyclone formation. The paper explains why El Niño does not raise storm counts as much as surface warmth alone would suggest, and argues global warming projections must account for both surface and subsurface ocean changes.","discovery_kind":"new_application","skeptic_critique":{"model":"deepseek-v4-flash","headline":"GPI-based counterfactual is a redescription of a statistical fit, not a physical prediction; 0.19 vs 0.15 may reflect D26's collinearity with other predictors.","rationale":"The reader's weakest assumption identifies the same core issue: the quantitative counterfactual depends on the empirical GPIocean fitted to the very TC data being explained. My stress-test refines this by noting that the decomposition's independence assumption is likely violated because D26 changes are dynamically coupled to the atmospheric variables (η1000, F) through the same wind-stress-curl forcing. This strengthens the concern that the 0.19 vs 0.15 numbers are not an independent physical prediction. However, the qualitative mechanism—anti-correlation between SST and D26 during El Niño—is supported by multiple data products and a plausible Ekman-pumping explanation, so the paper's overall conclusion is tenable as a conditional claim. The verdict remains CONDITIONAL, and no change to the reader's verdict is needed.","tokens_in":11286,"tokens_out":9416,"duration_ms":108225,"concrete_test":"Re-fit a Poisson regression of monthly TC genesis counts in the solid box (5°N–20°N, 160°E–170°W) on concurrent η1000, T-, F, and 1-month-lagged D26, using the same JTWC months; test the lagged D26 coefficient for significance and compute the out-of-sample El Niño–minus–La Niña count change when D26 is set to climatology. If the lagged D26 coefficient is insignificant (p>0.05) or the counterfactual change is not >0.04 TCs/month, the subsurface limitation is not causally identified.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The paper's central quantitative result—that shoaling D26 limits El Niño TC genesis in the central-north Pacific—is derived from a partial-derivative decomposition of GPIocean (Eq. 2) and the counterfactual 'if D26 was held to its climatology' (Results). GPIocean (Eq. 1) is an empirical index: exponents f,g,h,i and scaling p were fitted to the same JTWC TC counts used to evaluate the claim (ref. 30 and Methods 'GPI'). The decomposition treats D26 as an independent channel while holding all other variables fixed, but the El Niño D26 anomaly is forced by the same anomalous wind-stress curl that also directly alters η1000 and F (Fig. 5), and D26 may proxy for omitted atmospheric drivers (vertical wind shear was tested but dropped). Thus the 0.19 vs 0.15 comparison may be a mathematical consequence of the fitted function rather than an independent physical limitation. The observed negative T–D26 correlation (Fig. 4c) is consistent with a common cause, not necessarily a causal thermodynamic effect.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper documents that during El Niño, in the tropical central-north Pacific (5°N–20°N, 160°E–170°W), the 26°C isotherm depth (D26) shoals while sea surface temperature warms, producing a negative correlation between mixed-layer temperature (T̃) and D26. It attributes the shoaling to anomalous Ekman suction from cyclonic wind-stress curl associated with a weakened North Pacific Subtropical High. Using the empirical GPIocean index and a partial-derivative decomposition, it concludes that D26 shoaling offsets the favorable SST effect, such that GPI-inferred TC genesis increases by 0.15 TCs/month rather than 0.19 TCs/month if D26 were climatological. The qualitative pattern is verified across GODAS, SODA, ERSST, BOA_Argo, and HighResMIP.","tokens_in":11606,"tokens_out":4704,"duration_ms":52643,"significance":"If the causal interpretation holds, the result is significant: it identifies a subsurface mechanism that counteracts the conventional SST-based expectation of enhanced TC genesis during El Niño, with implications for projecting TC activity under El Niño-like warming. The paper's strengths include multi-dataset consistency, explicit significance testing of the T̃–D26 anti-correlation, and robustness checks with alternative GPIs and HighResMIP. However, the quantitative counterfactual is only as strong as the empirical GPI, and the causal language in the abstract and Results goes beyond what the statistical decomposition can support.","major_comments":[{"comment":"The central quantitative claim—0.19 vs 0.15 TCs per month—is obtained by holding D26 at climatology in GPIocean (Eq. 1). Because GPIocean's coefficients were fitted to the same JTWC counts used to define the ENSO TC differences, this partial-derivative exercise is a property of the fitted surface rather than an independent physical counterfactual. The problem is not the empirical nature of GPI per se, but that D26 is treated as an isolated channel: Fig. 5 shows the same wind-stress-curl anomalies that force ΔD26 also directly affect η1000 and F, and vertical wind shear was tested but dropped, leaving D26 potentially collinear with omitted atmospheric drivers. Please either (a) reframe the 0.19/0.15 numbers as a diagnostic decomposition of GPIocean (e.g., 'the D26 term reduces the GPI-inferred increase by 0.04 TCs/month'), or (b) add a causal check, e.g., partial correlations controlling","section":"Results, Eq. (2) and counterfactual"},{"comment":"The values 0.15 and 0.19 TCs per month are described as 'statistically significant at the 99% confidence level', but the Methods do not describe the test: null hypothesis, resampling scheme (block bootstrap? Monte Carlo?), handling of ENSO autocorrelation, or degrees of freedom. Without this, readers cannot assess whether the 0.15 and 0.19 differences are distinguishable from sampling noise, or whether the 0.04 difference between them is itself significant. Please add the exact procedure and report the uncertainty on the D26-induced increment.","section":"Results (0.15 vs 0.19 per month)"},{"comment":"The Discussion appropriately states that 'the main results have a quantitative dependence on the form of GPI (Supplementary Figs. S7 and S8)'. This is consistent with the GPI being a statistical proxy, but the abstract and Results nevertheless present 0.19 vs 0.15 as a robust physical finding. The manuscript should ensure that the abstract and Results use diagnostic language (e.g., 'according to GPIocean') rather than implying a controlled counterfactual, so that the caveat is not confined to the Discussion.","section":"Discussion, caveat paragraph"}],"minor_comments":[{"comment":"The typesetting of Eq. (2) is corrupted in the received text: '4GPI' should be 'ΔGPI', and the multiplication and partial-derivative symbols are missing. Please ensure the equation is properly rendered.","section":"Eq. (2)"},{"comment":"The caption is garbled: 'T--', 'Niño and La Niña. d , where ΔD26...' appears truncated. Please rewrite the caption so that panels a–d are clearly described.","section":"Fig. 3 caption"},{"comment":"T̃ is defined in the text but the symbol is visually ambiguous in the equation; consider a clearer notation (e.g., T_ML) and define it at first use in Eq. (1).","section":"Notation, Eq. (1)"},{"comment":"The code is 'available from the corresponding author on request'; for reproducibility, consider depositing analysis code in a permanent public repository (e.g., Zenodo) with a DOI.","section":"Data/code availability"},{"comment":"There is a typo: 'devasting' should be 'devastating'.","section":"Discussion, p. 4"},{"comment":"The fitted values of p, f, g, h, i in Eq. (1) are not stated; please either list them in the Methods or explicitly refer to Table/equation in ref. 30 where they appear.","section":"Methods, GPI"}],"recommendation":"major_revision","confidential_remarks":null},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"The genuinely new thing here is the observation that in the tropical central-north Pacific, SST and the 26°C isotherm depth are negatively correlated during ENSO, opposite to the canonical thermocline picture. The paper shows that this subsurface shoaling, driven by anomalous wind stress curl, offsets the favorable SST and limits TC genesis. That qualitative result holds up across GODAS, SODA, ERSST, BOA_Argo, and HighResMIP, with significance tests. The authors also give the mechanism (Ekman suction) and check CP/EP El Niño subtypes. Credit where earned: this is a useful, well-documented empirical finding.\n\nThe soft spot is the quantitative counterfactual. The claim that 0.15 vs 0.19 TCs per month would occur if D26 were held at climatology is computed by partial-differentiating GPIocean, an index whose coefficients were fitted to the same JTWC counts being explained. So the \"0.19\" is a property of the fitted function, not an independent physical prediction. The stress-test note adds a fair point: the D26 anomaly is forced by the same wind-stress-curl anomaly that alters other GPI variables, and D26 may proxy for omitted drivers like vertical wind shear. That said, the paper itself includes a caveat about quantitative dependence on the GPI form, and the qualitative conclusion survives that dependence. The negative correlation is an observational fact; only the exact number is index-bound.\n\nAlso worth saying: the paper is honest about its limitations and does not overclaim. The \"unexpected limitation\" framing is accurate for the community, which often assumes SST and upper-ocean heat content move together. The main weakness is that the central number is descriptive, and a reader should not treat it as a prediction.\n\nI'd send this to peer review if it crossed my desk. The observational finding is important enough for the TC/ENSO subfield, and the methods are standard. The counterfactual should be labeled as a sensitivity of the fitted index, not a physical experiment, but that is fixable in revision. Worth citing for the negative SST–D26 correlation; I'd also bring it to a reading group to discuss what a GPI decomposition can and cannot tell you.","headline":"A solid observational finding (negative SST–D26 correlation in the central-north Pacific during ENSO) buried under a headline counterfactual that is really just a derivative of an empirical index fitted to the same TC data.","tokens_in":12109,"tokens_out":1679,"would_cite":true,"duration_ms":20126,"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":"During El Niño, a shoaling 26°C layer in the central-north Pacific limits tropical cyclone genesis, counteracting warm sea surface temperatures.","keywords":["tropical cyclone genesis","El Niño","26°C isotherm depth","subsurface ocean heat content","genesis potential index","wind stress curl","Ekman pumping","western North Pacific"],"falsifier":"Look at the 5°N–20°N, 160°E–170°W box during future El Niño events with in-situ Argo profiles and observed genesis records: if TC genesis rises by roughly the SST-predicted 27% while D26 shoals, or if D26 deepens when wind stress curl is cyclonic, the subsurface limitation claim would fail. A complementary check: recompute the El Niño–La Niña TC difference with a genesis index that omits D26; the result should show a larger, not equal, difference.","tokens_in":11188,"feed_emoji":"🌪️","tokens_out":6140,"duration_ms":62324,"temperature":0.7,"pith_summary":"This paper takes on a longstanding puzzle: El Niño pushes the western Pacific warm pool eastward, yet the total number of tropical cyclones in the western North Pacific barely changes between El Niño and La Niña. The authors argue that the missing increase is explained by the ocean subsurface, not the atmosphere. During El Niño, anomalous winds over the tropical central-north Pacific drive Ekman suction that shoals the 26°C isotherm, cutting upper-ocean heat content even where sea surface temperatures are warm. Using a genesis potential index that includes the 26°C isotherm depth, they decompose the El Niño–La Niña difference in TC genesis and find that the subsurface shoaling offsets part of the SST-driven increase. A reader should care because SST-only indices would overstate how many cyclones an El Niño-like warming pattern would produce.","feed_headline":"Subsurface cooling puts a brake on El Niño typhoon gains","feed_subtitle":"Warm surface water hides a shoaling 26°C layer that limits cyclone genesis in the central-north Pacific.","key_machinery":"The load-bearing object is GPIocean, a genesis potential index for western North Pacific TCs built from four variables: 1000-hPa absolute vorticity, mean upper-mixed-layer temperature, net surface longwave radiation, and D26, the depth of the 26°C isotherm. The paper's decomposition (Eq. 2) varies one variable at a time while holding the others at climatology, following the method of ref. 19, to attribute the El Niño minus La Niña change in GPI to each factor. The physical mechanism that carries the argument is Ekman pumping: the anomalous cyclonic wind stress curl during El Niño drives upward Ekman velocity (Eq. 3), shoaling D26 in the central-north Pacific even as SST warms. D26 is the sta","core_discovery":"The central claim is that in the tropical central-north Pacific (roughly 5°N–20°N, 160°E–170°W), sea surface temperature and the depth of the 26°C isotherm are anticorrelated during ENSO, contrary to the canonical positive correlation elsewhere. The mechanism is a cyclonic wind stress curl anomaly over the region during El Niño, which produces Ekman suction and shoals D26 while westerly wind anomalies simultaneously warm the surface. The shoaling reduces upper-ocean heat content enough to counteract the otherwise favorable SSTs. Quantitatively, the GPIocean index yields 0.15 more TCs per month in this box during El Niño relative to La Niña; if D26 were held at climatology, the increase would","pith_inferences":["If the mechanism is correct, seasonal forecasts of typhoon counts in the western North Pacific should skillfully use upper-ocean heat content (D26 or similar) as a predictor, not only SST.","A testable extension: in climate-model runs under greenhouse warming, the El Niño-like mean-state change should produce a similar negative SST–D26 correlation in the central-north Pacific; models that fail to reproduce the Ekman shoaling would over-project TC frequency.","The same surface-subsurface cancellation may be at work in other basins where warming is accompanied by changes in wind stress curl, so the effect may not be Pacific-specific.","Because central-north-Pacific-origin storms have more ocean to cross, a modest reduction in count could coincide with an increase in accumulated cyclone energy—consistent with the paper's note that intense typhoons may still increase."],"forward_implications":["SST-only genesis potential indices overestimate the El Niño increase in central-north Pacific TC genesis, because they miss the negative D26 anomaly.","The sign of the SST–D26 correlation is not uniform: in the central-north Pacific it is negative (−0.40), opposite to the positive correlation in the northwestern and eastern Pacific.","Projections of TC activity under an El Niño-like warming pattern must include subsurface heat content changes, not just surface temperature, to avoid overestimating frequency.","The counteracting effect persists even when El Niño is split into central-Pacific and eastern-Pacific types.","TCs that do form in the central Pacific during El Niño travel farther before landfall, so the same number of storms may pose a different threat."],"supporting_citations":[{"why":"Supplies the GPIocean index with the D26 term that the paper decomposes.","marker":"30"},{"why":"Provides the one-variable-at-a-time GPI decomposition method used for Eq. (2).","marker":"19"},{"why":"Lists ocean thermal energy as a key parameter for TC formation, motivating D26.","marker":"26"},{"why":"Shows El Niño delivery of subsurface ocean heat affects eastern Pacific TCs, the subsurface-energy precedent.","marker":"23"},{"why":"Links changes in ocean subsurface environment to TC intensification under warming.","marker":"24"},{"why":"HighResMIP simulations reproduce the GPI response, supporting the observational conclusion.","marker":"39"},{"why":"SODA ocean reanalysis is used to verify the D26 and heat-content anomalies.","marker":"43"},{"why":"BOA_Argo observations verify the negative T-–D26 correlation in the central-north Pacific.","marker":"45"}],"fun_headline_variants":["Warm surface, cool subsurface: El Niño’s hidden cyclone brake","Shoaling 26°C layer curbs Pacific cyclone births during El Niño","El Niño’s subsurface chill limits typhoons in central-north Pacific","Surface heat hides a cooling layer that blocks tropical cyclones"],"cache_read_input_tokens":2688,"weakest_assumption_plain":"The argument assumes that the D26 term in GPIocean, whose coefficient was fitted to the same observed TC counts the paper explains, measures a causal physical influence of subsurface heat on TC genesis; if it is only a statistical match, the 0.15 versus 0.19 counterfactual restates the fit rather than proving a mechanism.","fun_headline_variants_meta":{"raw":{"variants":["Warm surface, cool subsurface: El Niño’s hidden cyclone brake","Shoaling 26°C layer curbs Pacific cyclone births during El Niño","El Niño’s subsurface chill limits typhoons in central-north Pacific","Surface heat hides a cooling layer that blocks tropical cyclones"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000191,"raw_usage":{"total_tokens":1199,"prompt_tokens":784,"completion_tokens":415,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":528,"completion_tokens_details":{"reasoning_tokens":337}},"tokens_in":528,"tokens_out":415,"duration_ms":5065,"temperature":1.0,"reasoning_tokens":337,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-05T12:12:10.497278+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Look at the 5°N–20°N, 160°E–170°W box during future El Niño events with in-situ Argo profiles and observed genesis records: if TC genesis rises by roughly the SST-predicted 27% while D26 shoals, or if D26 deepens when wind stress curl is cyclonic, the subsurface limitation claim would fail. A complementary check: recompute the El Niño–La Niña TC difference with a genesis index that omits D26; the result should show a larger, not equal, difference.","supporting_citations":[{"cited_title":"& Wang, C","cited_arxiv_id":null,"evidence_quote":"Supplies the GPIocean index with the D26 term that the paper decomposes."},{"cited_title":"J., Emanuel, K","cited_arxiv_id":null,"evidence_quote":"Provides the one-variable-at-a-time GPI decomposition method used for Eq. (2)."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Lists ocean thermal energy as a key parameter for TC formation, motivating D26."},{"cited_title":"I., Chou, C","cited_arxiv_id":null,"evidence_quote":"Links changes in ocean subsurface environment to TC intensification under warming."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"HighResMIP simulations reproduce the GPI response, supporting the observational conclusion."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"SODA ocean reanalysis is used to verify the D26 and heat-content anomalies."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"BOA_Argo observations verify the negative T-–D26 correlation in the central-north Pacific."}],"review_version":1}