{"id":"fa06ecde-c123-4fc8-8b26-798675142fed","arxiv_id":"2508.09376","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"high","formal_verification":"none","parameter_count":0,"one_line_summary":"Proposes that small, well-timed perturbations, amplified by atmospheric chaos, could steer extreme weather events to reduce their impacts.","lead":"This paper proposes 'Weather Jiu-Jitsu': using small, precisely timed interventions to exploit atmospheric chaos and redirect destructive weather. It is a call for a global research agenda, not a demonstration that the approach works.","discovery_kind":"review","skeptic_critique":{"model":"deepseek-v4-flash","headline":"No quantitative bridge from chaos to controllability: the key feasibility point rests on an admitted open delivery problem and an in-review self-citation; a GFS-based perturbation experiment would settle it.","rationale":"I read the paper as a perspective/research agenda, not a proof of weather control. The title and abstract use hedged language ('could develop'), but the strongest claim still implies feasibility: that small, well-timed perturbations can steer real extremes. The reader's weakest assumption—transfer from low-order models to the real atmosphere—is exactly the point I find most load-bearing. I agree with the conditional verdict: the paper frames a legitimate research program, but the scientific case as written is incomplete. My stress-test adds specificity: the missing piece is not just 'model realism' but a quantitative estimate of required perturbation energy and a demonstration that chaotic amplification can be directed, not merely that it exists. The paper's own admission that delivery is open and its reliance on an in-review self-citation (ref. 79) for L84 control strengthen the concern. A targeted experiment using the paper's own illustrative GFS AR case is the most direct way to test the central premise without requiring full-scale field trials. If the test showed the advertised footprint change is indistinguishable from internal variability, or that required energies are orders of magnitude beyond feasible delivery, the central claim would be unsupported. If it succeeded, the paper's core idea would have at least one concrete quantitative anchor. Since the reader already assigned CONDITIONAL with high correctness risk, my assessment does not move the verdict; it sharpens the condition under which the claim would be accepted.","tokens_in":13135,"tokens_out":4510,"duration_ms":52877,"concrete_test":"Use the NOAA GFS case behind Fig. 1: initialize an ensemble at t=0; apply the three cyan-dot nudges at t=1.0, 2.25, and 3.5 at amplitudes consistent with the figure's 'small perturbations'; also run a control and a same-amplitude 'wrong direction' ensemble. Compare the t=4 precipitable-water footprint and landfall distribution. If the advertised redistribution is not separable from ensemble spread, or if the amplitude required to move the AR landfall by, say, 100 km exceeds a physically plausible delivery energy (e.g., >0.1% of the storm's latent-heat release), the chaos-amplification premise is unsupported. This test does not depend on the in-review L84 result.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The central claim is that small, well-timed perturbations, amplified by atmospheric chaos, can steer extreme weather trajectories. The paper's evidence is (i) controllability of L63/L84 toy models and (ii) the fact that the atmosphere is chaotic. Neither supports the required controllability claim for a real, high-dimensional, moist, forced-dissipative atmosphere. Chaos gives sensitivity to initial conditions, not controllability in a desired direction; a small nudge may grow but project onto many scales, and the minimum energy needed to redirect a synoptic-scale event is never estimated. The paper's own open questions (a)-(c) state that whether nudges can control trajectories is open, and the delivery section explicitly says 'This is an open challenge.' A key supporting result, adaptive chaos control of L84, is ref. 79, 'Submitted to Geophys. Res. Lett. (in review, 2025)', so it is not independently checkable. Figure 1 is labeled 'Conceptual illustration' and is not a simulation result. The passage 'Direct impacts will be restricted to the event's time scale' is an unsupported overclaim, since chaos can amplify remote teleconnections. The concern is not that the idea is impossible; it is that the manuscript's central inference from sensitivity to steering is a logical leap, with no quantitative, real-atmosphere demonstration. This is the load-bearing condition: if a realistic model shows required perturbations are too large or uncontrollable, the proposed infrastructure collapses to speculation.","agreement_with_reader":"agree"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper argues that conventional climate risk management—physical infrastructure, insurance, and early warning—cannot scale to handle catastrophic weather extremes, and proposes a new adaptive strategy called \"Weather Jiu-Jitsu.\" The idea is to exploit atmospheric chaos: small, precisely timed and placed energy perturbations could be amplified by intrinsic instabilities to steer specific weather systems (hurricanes, atmospheric rivers, blocking patterns) away from destructive trajectories. The argument is built on (i) the well-known sensitivity of the atmosphere to initial conditions, (ii) chaos-control results in the Lorenz 63 and Lorenz 84 toy models, (iii) recent ML-based weather prediction, and (iv) a conceptual GFS-based illustration of an atmospheric river. The authors explicitly identify the delivery mechanism as an open challenge and list open technical questions, but nevertheless assert that sensitivity to initial conditions \"implies\" steerability toward desirable regimes. The manuscript is written as a perspective/agenda piece rather than a quantitative feasibility study.","tokens_in":13426,"tokens_out":3003,"duration_ms":33555,"significance":"If the central claim were established, the paper would outline a transformative approach to disaster risk reduction: a low-energy, nature-assisted infrastructure for steering weather extremes. The manuscript usefully synthesizes the limitations of existing infrastructure, insurance, and geoengineering approaches, and it points to a concrete research program with identifiable milestones. It also honestly acknowledges several open problems. However, the load-bearing inference—from chaotic sensitivity to practical controllability of synoptic-scale weather—is not supported by quantitative evidence or a mechanistic argument. The value of the paper at present is primarily agenda-setting: it poses an important question and sketches a plausible route, but it does not yet demonstrate that the route is physically viable. This should be reflected in the framing and in the strength of the claims made.","major_comments":[{"comment":"The central claim is the sentence \"sensitivity to initial conditions also implies that small, well-timed perturbations may be able to steer the atmosphere toward desirable regimes.\" This is a logical leap: chaos implies divergence of nearby trajectories, not controllability to a specified desirable regime in a high-dimensional, moist, forced-dissipative atmosphere. The paper itself lists this as open question (a) and later states the delivery problem is \"an open challenge.\" The manuscript should either replace \"implies\" with a clearly labeled hypothesis or provide a quantitative plausibility argument (e.g., scaling of required perturbation energy versus available instability growth, or results from an intermediate-complexity atmospheric model). This is the load-bearing point for the entire proposal.","section":"\"Toward Weather Jiu-Jitsu\" (pp. 13–18)"},{"comment":"The key supporting result for L84 adaptive chaos control is reference 79, described as \"Submitted to Geophys. Res. Lett. (in review, 2025).\" Readers cannot verify this result, and the paper's argument depends on it. The authors should either cite published, peer-reviewed chaos-control studies for L84/L63, or include the essential numerical results in the manuscript (e.g., in supplementary material) so that the claim is independently checkable.","section":"Ref. 79 (p. 28)"},{"comment":"Figure 1 is labeled a \"Conceptual illustration\" and is described as \"Adapted from NOAA Global Forecast System (GFS) model output.\" No details are given about the nudges, their amplitude, the model configuration, or how the perturbed trajectory was produced. If this figure is intended as evidence of feasibility, the experimental setup must be described (including perturbation energies relative to natural variability, control runs, and a comparison with an unperturbed forecast). If it is only illustrative, the caption should state explicitly that it does not constitute a simulation result.","section":"Figure 1 (pp. 14–15)"},{"comment":"The statement \"Direct impacts will be restricted to the event's time scale\" is an unsupported overclaim. Because the atmosphere is chaotic and teleconnections are widespread, a localized nudge of a jet-stream wave or blocking pattern can have nonlocal and delayed effects. The manuscript acknowledges the possibility of adverse impacts only vaguely. This statement should be tempered or supported by an analysis of remote response and unintended consequences, especially in a paper whose central mechanism is chaos amplification.","section":"Final paragraph before References (p. 19)"}],"minor_comments":[{"comment":"The spelling of the proposed term is inconsistent: \"Weather Jiu-Jitsu\" (title, abstract) and \"Weather Jiu Jitsu\" (main text, e.g., pp. 3, 6). Please standardize.","section":"General"},{"comment":"Several references are in-press or preprint (e.g., refs. 51, 52, 79). For a perspective piece this is acceptable, but the authors should mark them clearly and, where possible, update to published versions.","section":"References"},{"comment":"Supplementary Figures 1–5 and Supplementary Video 1 are cited in the text but were not included in the manuscript provided for review. Please ensure these are available and that their content is summarized in the caption or main text, since some are used to support factual claims about exposure and circulation patterns.","section":"Supplementary material"}],"recommendation":"major_revision","confidential_remarks":"This is best evaluated as a perspective/research-agenda paper, not as a demonstration of feasibility. The central claim is currently overstated relative to the evidence presented, and the manuscript relies on an in-review self-citation for a key supporting result. I would recommend major revision: reframe the \"implies\" statement as an explicitly open hypothesis, add a quantitative plausibility estimate or an intermediate-model demonstration, and clarify the status of Figure 1. The paper's scope and proposal are appropriate for a perspectives venue, but the current framing risks overstating the state of the science."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"This is a perspective piece, not a research paper. The core idea—small, well-timed perturbations amplified by chaos to steer extreme weather—is old in toy models, but this is a coherent attempt to translate it into a research agenda for mid-latitude extremes. The paper does a solid job reviewing why traditional adaptation is stretched thin and connecting to real threads: Lorenz 63/84 control, data-assimilation MPC, Moonshot 8. The authors are also candid about the two biggest gaps: they list open questions (a)-(c) and explicitly call the delivery mechanism \"an open challenge.\"\n\nThe problem is the same one the stress test identifies: the paper never bridges \"chaos is sensitive to initial conditions\" to \"we can steer the atmosphere toward desirable regimes.\" That is a logical leap, not an argument. Sensitivity cuts both ways—a nudge can grow, but it can also grow into the wrong pattern. The Figure 1 example is labeled conceptual and the key L84 control result is ref. 79, in review, so no one can check it. And the claim that \"direct impacts will be restricted to the event's time scale\" is unsupported and probably wrong, since teleconnections are a thing.\n\nNone of that is fatal for what the paper actually is. It's a call for research, and the authors say so. The soft spot is the rhetorical slide from \"worth exploring\" to \"humanity could develop a global infrastructure.\" The evidence so far is toy models and a sketch.\n\nWho should read it: anyone thinking about radical adaptation, or about whether controllability of chaotic systems can ever scale to weather. It deserves a serious referee if the journal publishes perspective/agenda pieces. Reviewers should push for a sharper separation of speculation from agenda, and for a realistic-model numerical demonstration before the strong claims are made.","headline":"A coherent and honest research agenda for weather control via chaos, but the central feasibility inference is a leap the paper doesn't try to bridge.","tokens_in":13918,"tokens_out":2967,"would_cite":false,"duration_ms":30687,"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":"Weather Jiu-Jitsu proposes using atmospheric chaos to steer extreme weather away from disasters.","keywords":["weather jiu-jitsu","chaos control","extreme weather","atmospheric circulation","climate adaptation","adaptive control","weather modification","atmospheric rivers"],"falsifier":"A decisive test would be a systematic hindcast search over many past extremes (e.g., two dozen Atlantic hurricane landfalls and atmospheric-river flood weeks) using a high-resolution weather model or deep-learning emulator: compute the smallest initial-condition perturbation that changes the event's impact. If no perturbation below a fixed energy budget (say, less than 0.1% of the event's domain-integrated kinetic energy) succeeds in a substantial fraction of cases, the premise fails. The search would also be falsified if successful redirections consistently create comparable or larger extreme","tokens_in":12998,"feed_emoji":"🌪️","tokens_out":7160,"duration_ms":74082,"temperature":0.7,"pith_summary":"This paper argues that the same property of the atmosphere that makes weather hard to predict—sensitivity to tiny disturbances—can be turned into a strategy for reducing disasters. 'Weather Jiu-Jitsu' would apply small, precisely timed, low-energy nudges to mid-latitude circulation so that natural dynamics amplify the effect and steer hurricanes, atmospheric rivers, or blocking patterns away from harmful outcomes. The paper builds the case by showing that chaotic control is already feasible in the low-order models that capture these phenomena, and that the traditional trio of dams, insurance, and early warnings is structurally overwhelmed by catastrophic events. The authors present the idea as an urgent research agenda, with the physical delivery of the perturbation explicitly left as an open problem.","feed_headline":"Tiny, well-timed nudges could steer extreme weather","feed_subtitle":"The same chaos that limits forecasts could be harnessed to redirect hurricanes, floods, and heat waves.","key_machinery":"The mechanism is adaptive chaos control: repeatedly estimate the atmospheric state through data assimilation, identify where disturbances grow fastest using Lyapunov exponents, and apply small nudges via model predictive control to shift the trajectory toward a desired basin of attraction. The paper uses the L84 model—a low-order idealization of jet-stream and eddy interaction—as the conceptual bridge between toy-model control and real mid-latitude weather extremes, and treats deep-learning emulators of the full atmosphere as the operational path to real-time control.","core_discovery":"The paper's central claim is that the chaos that caps weather predictability is also a control handle. Because mid-latitude circulation possesses multiple coexisting regimes and is exponentially sensitive to small disturbances, well-timed, low-energy nudges can be amplified by the atmosphere's own dynamics and push a storm track, an atmospheric river, or a blocking pattern into a less destructive state. The authors propose two control settings: targeted nudges while an extreme event is developing, and regular sub-seasonal nudges that reduce the probability of undesirable circulation regimes. They argue the most tractable targets are jet-stream-mediated phenomena—hurricane steering, atmospher","pith_inferences":["Editorial extension: a low-cost first test does not require physically perturbing the atmosphere: in reanalysis or ensemble hindcasts of past extreme events, one can search for initial-condition perturbations below a chosen energy threshold that change the outcome, using existing deep-learning weather emulators.","Editorial extension: the paper leaves the delivery problem open, but the control problem and the delivery problem can be separated; optimal perturbation locations from adjoint or sensitivity methods could be validated in regional models years before any laser, seeding, or heating technology is deployed.","Editorial extension: if block entry/exit transitions are indeed the high-instability windows, Weather Jiu-Jitsu becomes a circulation-regime management tool, changing the probability of a whole class of extremes rather than single storms—which would raise transboundary governance questions the paper only flags.","Editorial extension: a natural next step is to compute, for historical catastrophes, the minimal perturbation energy needed to shift the outcome; those numbers would indicate whether the approach is energetically plausible or confined to idealizations."],"forward_implications":["If the thesis holds, the costliest tail of disaster risk—hurricane landfalls, atmospheric-river floods, and persistent blocking-related heat waves—could be reduced with far smaller energy inputs than building or retrofitting hard infrastructure everywhere.","The same rapidly improving machine-learning weather models that now rival physics-based forecasts could serve as the control emulators, making an operational prototype testable before physical delivery systems exist.","Adaptive nudging would give insurance systems a way to shrink the catastrophic losses that currently bankrupt pools like the U.S. flood program, since the rare cluster events would be the primary targets.","Success would create a second, independent rationale for investing in dense atmospheric observation and data assimilation: not just prediction, but control.","Because extremes will persist even after deep decarbonization, this approach would be useful regardless of how the greenhouse-gas trajectory unfolds."],"supporting_citations":[{"why":"Supplies the classic low-order chaotic model of atmospheric convection that is the proof-of-concept arena for chaos control.","marker":"[32]"},{"why":"Supplies the low-order model of jet-stream/eddy interaction whose multiple regimes are the conceptual bridge to blocking and storm-track extremes.","marker":"[33]"},{"why":"One of the cited demonstrations that chaos control can stabilize or redirect trajectories in the L63 system.","marker":"[71]"},{"why":"Shows data assimilation can be paired with control to steer the butterfly attractor, grounding the adaptive-control architecture.","marker":"[76]"},{"why":"Demonstrates model predictive control with data assimilation driving the L63 system to a prescribed regime.","marker":"[77]"},{"why":"Extends the data-assimilation-plus-control result to ensemble methods, supporting real-time state estimation for nudging.","marker":"[78]"},{"why":"The authors' own companion study applying adaptive chaos control to idealized weather models, the direct evidence behind the proposal.","marker":"[79]"},{"why":"Shows blocking events are marked by high instability during regime transitions, identifying the timing window for effective nudges.","marker":"[92]"},{"why":"Provides a concrete low-energy delivery mechanism (laser-guided lightning) that could realize the required perturbations.","marker":"[98]"}],"fun_headline_variants":["Use chaos to redirect hurricanes and floods","Weather Jiu-Jitsu: small nudges, big storm shifts","Chaos gives us a lever to steer extreme weather","Tiny nudges, timed right, can reroute storms","Harnessing chaos to tame extreme weather"],"cache_read_input_tokens":2816,"weakest_assumption_plain":"The load-bearing premise is that controllability shown in idealized low-order models transfers to the real, high-dimensional, moist atmosphere—i.e., that a small, well-timed energy input can reliably redirect a specific weather system without causing comparable adverse effects elsewhere.","fun_headline_variants_meta":{"raw":{"variants":["Use chaos to redirect hurricanes and floods","Weather Jiu-Jitsu: small nudges, big storm shifts","Chaos gives us a lever to steer extreme weather","Tiny nudges, timed right, can reroute storms","Harnessing chaos to tame extreme weather"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000415,"raw_usage":{"total_tokens":1942,"prompt_tokens":672,"completion_tokens":1270,"prompt_tokens_details":{"cached_tokens":256},"prompt_cache_hit_tokens":256,"prompt_cache_miss_tokens":416,"completion_tokens_details":{"reasoning_tokens":1208}},"tokens_in":416,"tokens_out":1270,"duration_ms":8374,"temperature":1.0,"reasoning_tokens":1208,"cache_read_input_tokens":256,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-05T21:05:11.082466+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"A decisive test would be a systematic hindcast search over many past extremes (e.g., two dozen Atlantic hurricane landfalls and atmospheric-river flood weeks) using a high-resolution weather model or deep-learning emulator: compute the smallest initial-condition perturbation that changes the event's impact. If no perturbation below a fixed energy budget (say, less than 0.1% of the event's domain-integrated kinetic energy) succeeds in a substantial fraction of cases, the premise fails. The search would also be falsified if successful redirections consistently create comparable or larger extreme","supporting_citations":[{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the classic low-order chaotic model of atmospheric convection that is the proof-of-concept arena for chaos control."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Supplies the low-order model of jet-stream/eddy interaction whose multiple regimes are the conceptual bridge to blocking and storm-track extremes."},{"cited_title":"& Yau, H.-T","cited_arxiv_id":null,"evidence_quote":"One of the cited demonstrations that chaos control can stabilize or redirect trajectories in the L63 system."},{"cited_title":"& Sun, Q","cited_arxiv_id":null,"evidence_quote":"Shows data assimilation can be paired with control to steer the butterfly attractor, grounding the adaptive-control architecture."},{"cited_title":"& Kotsuki, S","cited_arxiv_id":null,"evidence_quote":"Demonstrates model predictive control with data assimilation driving the L63 system to a prescribed regime."},{"cited_title":"Ensemble Kalman filter meets model predictive control in chaotic systems","cited_arxiv_id":null,"evidence_quote":"Extends the data-assimilation-plus-control result to ensemble methods, supporting real-time state estimation for nudging."},{"cited_title":"Weather Jiu-Jitsu","cited_arxiv_id":null,"evidence_quote":"The authors' own companion study applying adaptive chaos control to idealized weather models, the direct evidence behind the proposal."},{"cited_title":"& Lucarini, V","cited_arxiv_id":null,"evidence_quote":"Shows blocking events are marked by high instability during regime transitions, identifying the timing window for effective nudges."},{"cited_title":null,"cited_arxiv_id":null,"evidence_quote":"Provides a concrete low-energy delivery mechanism (laser-guided lightning) that could realize the required perturbations."}],"review_version":1}