{"schema_version":1,"experiment_id":"eoa_inverse_innovation_exp11_mechanism_context_external20_20260804","research_id":"eoa_inverse_innovation_exp11_external_scrutiny_20260804","cell_id":"invariant_mode_decomposition_design__biology_ecology","opaque_id":"invariant_mode_decomposition_design__biology_ecology__A","search_lanes":{"direct_problem":{"queries":["shallow lake restoration dynamic mode decomposition eigenmode mesocosm control","lake regime shift early warning multivariate leading eigenvalue community indicators","shallow lake mesocosm factorial fish removal macrophyte nutrient reduction clear water"],"source_ids":["S1","S5","S8"],"no_result_note":"No direct empirical study of the complete proposed modal-control experiment was found; lake evidence instead establishes unreliable conventional warnings and interacting trophic responses."},"closest_prior_art":{"queries":["ecological community time series multivariate autoregressive leading eigenvalue stability intervention","lake food web Jacobian eigenvalue early warning community matrix management","ecological network control theory eigenmodes intervention ecosystem management","Koopman control ecology ecosystem management dynamic mode decomposition"],"source_ids":["S2","S3","S4"],"no_result_note":"Close methods estimate multivariate stability modes or input-output dynamics, but the retained sources do not demonstrate a randomized shallow-lake policy chosen to damp an empirically identified risky mode."},"historical_terminology":{"queries":["multispecies time series return rate reactivity community matrix vector autoregressive ecology","generalized modeling ecosystem early warning dominant eigenvalue lake restoration","biomanipulation top-down bottom-up shallow lake restoration"],"source_ids":["S2","S3","S5","S6"],"no_result_note":null},"products_practices_standards":{"queries":["site:gov lake restoration biomanipulation fish removal macrophyte nutrient official guidance shallow lakes","PCLake official model shallow lake restoration management dynamic model","EU Water Framework Directive lake restoration biomanipulation guidance macrophytes fish removal"],"source_ids":["S6","S7"],"no_result_note":"Official and first-party materials establish permits, integrated biomanipulation, and scenario modeling, but not a standard for modal intervention selection."},"non_english_regional":{"queries":["浅水湖泊 生态修复 生物操纵 鱼类移除 沉水植物 营养盐 模型 预警 特征值","Flachsee Restaurierung Biomanipulation Fischentnahme Makrophyten Nährstoffreduktion Frühwarnung Eigenwert","ondiepe meren herstel biomanipulatie visverwijdering waterplanten nutriënten PCLake","restauración lagos someros biomanipulación peces macrófitas nutrientes modelo alerta temprana"],"source_ids":["S7","S8"],"no_result_note":"Dutch and Chinese terminology produced regional modeling and warning-method literature, but no exact modal-policy implementation."},"composition_subproblems":{"queries":["shallow lake factorial mesocosm top-down bottom-up macrophyte phytoplankton","dynamic mode decomposition with control ecological system input output","multivariate autoregressive freshwater plankton stability perturbation response","shallow lake restoration nutrient reduction fish manipulation macrophyte combined management"],"source_ids":["S2","S4","S5","S6","S7"],"no_result_note":"The component subproblems are individually well represented; their proposed randomized, gate-controlled composition was not located."}},"sources":[{"source_id":"S1","title":"Early warning signals have limited applicability to empirical lake data","url":"https://www.nature.com/articles/s41467-023-43744-8","publisher":"Nature Communications","date_or_year":"2023","source_type":"PRIMARY_RESEARCH","language":"English","claims_supported":["Across nine long-term lake datasets, many apparent regime shifts were not critical transitions.","Most tested early-warning indicators performed poorly; multivariate indicators were only weakly superior to univariate indicators.","Different trophic levels can exhibit different transition dynamics, supporting concern that separate indicators may be misleading while also challenging any assumption of one stable warning mode."]},{"source_id":"S2","title":"Quantifying effects of abiotic and biotic drivers on community dynamics with multivariate autoregressive (MAR) models","url":"https://faculty.washington.edu/eeholmes/Files/Hamptonetal2013.pdf","publisher":"Ecological Society of America, Ecology","date_or_year":"2013","source_type":"SECONDARY_RESEARCH","language":"English","claims_supported":["MAR models estimate species-interaction and environmental-driver matrices from multivariate ecological time series.","The fitted matrices support perturbation-response and community-stability calculations, including eigenvalue-based return rate and reactivity.","Freshwater plankton communities were already a major ecological application area, and observation error, irregular sampling, openness, model uncertainty, and natural-history interpretation are important limitations."]},{"source_id":"S3","title":"A closed form for Jacobian reconstruction from time series and its application as an early warning signal in network dynamics","url":"https://epic.awi.de/id/eprint/54516/1/rspa20200742.pdf","publisher":"Proceedings of the Royal Society A","date_or_year":"2021","source_type":"PRIMARY_RESEARCH","language":"English","claims_supported":["A leading Jacobian eigenvalue can be reconstructed from multivariate time-series covariance when sufficient structural constraints are available.","The method was demonstrated on nonlinear ecological meta-foodweb models as an early-warning approach.","The authors found the required data volume still prohibitive for ecological application, directly supporting stringent residual, identity, and sampling gates."]},{"source_id":"S4","title":"Dynamic Mode Decomposition with Control","url":"https://epubs.siam.org/doi/10.1137/15M1013857","publisher":"Society for Industrial and Applied Mathematics","date_or_year":"2016","source_type":"PRIMARY_RESEARCH","language":"English","claims_supported":["DMD with control separates estimated unforced dynamics from effects of measured actuation and produces data-driven input-output models.","The method uses snapshots of observables and actuation without requiring governing equations.","This establishes the generic causal machinery behind estimating a transition operator and mapping randomized controls into modal coordinates, but not its validity in shallow lakes."]},{"source_id":"S5","title":"The relative importance of top-down and bottom-up control of phytoplankton in a shallow macrophyte-dominated lake","url":"https://hero.epa.gov/reference/6979847/","publisher":"Freshwater Biology; indexed by U.S. EPA HERO","date_or_year":"1998","source_type":"PRIMARY_RESEARCH","language":"English","claims_supported":["Replicated factorial shallow-lake mesocosms tested zooplankton grazing, nutrient supply, fish, and macrophyte mechanisms affecting clear water.","Responses were context dependent: macrophytes did not reduce nitrate in the experiment, and fish density altered their refuge function.","The work establishes feasibility of controlled trophic perturbations while showing why causal effects cannot safely be inferred from generic ecological expectations alone."]},{"source_id":"S6","title":"Fish and Fisheries Management in Lakes and Reservoirs: Technical Supplement to the Lake and Reservoir Restoration Guidance Manual","url":"https://nepis.epa.gov/Exe/ZyPURL.cgi?Dockey=20004RPR.TXT","publisher":"U.S. Environmental Protection Agency and Terrene Institute","date_or_year":"1993","source_type":"OFFICIAL_GUIDANCE","language":"English","claims_supported":["Fish-based biomanipulation for water clarity is established lake-management practice and often works with other restoration techniques.","The manual identifies removal or manipulation of bottom-feeding, predatory, and planktivorous fish as management levers while emphasizing complex interactions, uncertainty, controls, and caution.","Management plans have identifiable governing or funding authorizers; many techniques require permits and some may be performed only by state agencies."]},{"source_id":"S7","title":"Effecten van beheersmaatregelen in ondiepe meren inzichtelijk met online rekenmodel PCLake","url":"https://www.pbl.nl/publicaties/effecten-van-beheersmaatregelen-in-ondiepe-meren-inzichtelijk-met-online-rekenmodel-pclake","publisher":"Netherlands Environmental Assessment Agency (PBL)","date_or_year":"2007","source_type":"FIRST_PARTY_PRODUCT","language":"Dutch","claims_supported":["PCLake is an operational water-quality model developed for estimating effects of proposed management measures in shallow lakes.","Dutch lake managers can use it to assess critical phosphorus loading and restoration options.","The source states that phosphorus improvement alone may not restore clarity because interacting ecological processes determine clear versus turbid states."]},{"source_id":"S8","title":"浅水湖泊稳态转换预警识别方法局限与展望 (Limitations and prospects of early-warning identification methods for regime shifts in shallow lakes)","url":"https://www.ecologica.cn/stxb/ch/html/2017/11/stxb201604010594.htm","publisher":"Acta Ecologica Sinica","date_or_year":"2017","source_type":"SECONDARY_RESEARCH","language":"Chinese","claims_supported":["Shallow-lake warning methods include experiments, dynamical models, and statistical critical-slowing indicators.","Reliable experimental warning can require controls, substantial data, appropriate perturbation speed and magnitude, adequate sampling frequency, and mechanistic understanding.","Multiple related trophic transitions, strong forcing, nonlinear processes, and low sampling frequency can suppress, distort, or entirely miss warning signals."]}],"problem_evidence":{"status":"PARTLY_SUPPORTED","finding":"The broader problem is real: shallow lakes contain interacting trophic and physicochemical processes, conventional single-series warnings often perform poorly, and multivariate models can expose community-level stability directions. However, no retained source establishes that the specified seasonal mesocosm regime contains a reproducible, bootstrap-stable coupled direction with useful lead time; empirical lake evidence also shows that trophic levels can transition differently and that most warning methods perform poorly.","source_ids":["S1","S2","S3","S5","S8"],"uncertainty":"The exact problem statement remains an experimentally falsifiable hypothesis. Weekly sampling over one season may provide too few transitions for an 8–12-variable operator unless replication, shrinkage, and cross-unit pooling are carefully validated."},"adopter_evidence":{"status":"SUPPORTED","finding":"Mesocosm investigators, lake-model users, lake managers, governing bodies, and permit-holding agencies are identifiable adopter or authorizer classes. The immediate adopter is a principal investigator or facility/permit holder; any later lake deployment would require the relevant fisheries and ecological authorities.","source_ids":["S5","S6","S7"],"uncertainty":"The exact jurisdiction is unspecified, so the particular animal-care, collection, stocking, and plant-transfer permissions cannot be named in advance."},"implementation_evidence":{"status":"PARTLY_SUPPORTED","finding":"Every major technical ingredient is feasible in isolation: factorial shallow-lake mesocosms, multivariate transition and stability models, eigenvalue-based warning, DMD with measured controls, scenario modeling, and established restoration levers. The exact composition—randomized identification of control-to-mode effects followed by a matched-budget policy that dampens a validated local mode—was not demonstrated in a retained lake study.","source_ids":["S2","S3","S4","S5","S6","S7"],"uncertainty":"Major unresolved issues are effective time-series sample size, cross-mesocosm heterogeneity, seasonal nonstationarity, non-normal dynamics, post-treatment mode rotation, omitted forcing, and whether estimated modal leverage transfers from Stage A to Stage B."},"prior_art":{"disposition":"ADJACENT_PRIOR_ART","closest_analogues":[{"name":"Freshwater-community MAR stability analysis","source_ids":["S2"],"same_problem":true,"same_causal_lever":false,"overlap":"Fits multivariate ecological transition matrices, separates biotic and abiotic drivers, and derives eigenvalue-based return and perturbation metrics from freshwater plankton time series.","remaining_difference":"It does not identify randomized restoration-input effects in modal coordinates or prospectively choose and test a lowest-disturbance mode-damping policy."},{"name":"Leading-Jacobian-eigenvalue ecological early warning","source_ids":["S3"],"same_problem":true,"same_causal_lever":false,"overlap":"Reconstructs a coupled ecological stability direction and leading eigenvalue from multivariate time series to warn of instability.","remaining_difference":"It is a simulated meta-foodweb warning method with high data demand, not a randomized intervention map or a policy trial in shallow-lake mesocosms."},{"name":"Dynamic Mode Decomposition with Control","source_ids":["S4"],"same_problem":false,"same_causal_lever":true,"overlap":"Provides the generic data-driven operator-and-actuation decomposition needed to map measured controls into dynamical modes.","remaining_difference":"It does not establish ecological validity, a shallow-lake state vector, biological safety gates, or superiority over direct nonlinear ecological optimization."},{"name":"Factorial top-down/bottom-up shallow-lake mesocosms","source_ids":["S5"],"same_problem":true,"same_causal_lever":false,"overlap":"Randomizes coupled trophic and nutrient-related manipulations in replicated lake enclosures and measures clear-water mechanisms.","remaining_difference":"Treatments test ecological factors directly; they are not selected to oppose a preregistered transition eigenmode and are not compared with matched-budget modal and nonmodal policies."},{"name":"Integrated biomanipulation and PCLake scenario management","source_ids":["S6","S7"],"same_problem":true,"same_causal_lever":false,"overlap":"Uses combinations of nutrient, fish, food-web, and model-based measures to pursue clear water under management authority.","remaining_difference":"These practices optimize or prescribe measures in ecological coordinates and scenarios, without the proposed randomized control-to-mode identification, spectral identity gates, or confirmatory modal-policy contrast."}],"contrastive_claim_remaining":"Within the declared clear-water neighborhood, a mode chosen only from preregistered Stage-A randomized data can remain stable enough that the lowest-disturbance admissible treatment combination predicted to oppose it produces at least 25% lower risky-mode-amplitude AUC and 0.5 SD lower ecological loss than both an intensity-matched coordinate policy and a directly optimized nonlinear state-space policy in independent mesocosms.","contrastive_claim_falsifier":"The contrastive prior-art claim would be falsified by a pre-existing shallow-lake or closely equivalent community experiment that already combines randomized control-input identification, a locally fitted transition operator, validated eigen/singular-mode selection, treatment chosen to damp that mode, and an independent matched-budget comparison against coordinate-wise and direct nonmodal control. The empirical claim itself is falsified by failure of any preregistered model gate or either simultaneous superiority contrast.","confidence":"MODERATE","search_limitations":"The review was deliberately bounded to exactly eight opened direct sources. It covered six adversarial lanes and English, Dutch, and Chinese terminology, but cannot exhaust paywalled literature, patents, theses, regulatory jurisdictions, non-indexed products, or every language. Similar work may use system-identification, resilience, community-matrix, adaptive-management, Koopman, or biomanipulation terminology without mentioning modes."},"researchability_gates":{"externally_supported_problem":{"status":"PASS","rationale":"Independent lake evidence shows that separate early-warning indicators frequently fail, trophic components can change differently, and multivariate community dynamics contain interaction and stability information worth testing.","source_ids":["S1","S2","S8"]},"identifiable_adopter_or_authorizer":{"status":"PASS","rationale":"Mesocosm investigators and facility or permit holders are concrete first adopters, while governing bodies and fisheries/ecological agencies are identifiable authorizers for any broader management action.","source_ids":["S5","S6","S7"]},"distinct_testable_incremental_claim":{"status":"PASS","rationale":"Adjacent work covers sensing, system identification, experiments, and restoration separately, but the preregistered claim of superior matched-budget control by damping an independently validated local community mode remains distinct and quantitatively falsifiable.","source_ids":["S2","S3","S4","S5","S7"]},"bounded_next_evidence_step":{"status":"PASS","rationale":"A preregistered Stage-A randomized mesocosm study, followed only after locked residual, gap, identity, capacity, and power gates by an independent four-arm Stage B, is a bounded evidence sequence.","source_ids":["S3","S4","S5","S8"]},"no_unresolved_safety_or_authority_stop":{"status":"PASS","rationale":"The first step is confined to enclosed mesocosms, site-native organisms, bounded nonlethal manipulations, explicit oxygen and mortality stopping rules, and prior approvals. Official guidance confirms that permits and agency authority must be resolved before applicable fish manipulations; no inherently prohibited action is required.","source_ids":["S5","S6"]},"adequate_search_evidence":{"status":"PASS","rationale":"The bounded search covered the direct problem, closest methods, historical terminology, products and official practice, non-English regional terminology, and component combinations using eight sources from seven publication venues, including five primary, official, or first-party sources.","source_ids":["S1","S2","S3","S4","S5","S6","S7","S8"]}},"strict_success":true,"screen_survival":true,"remaining_research_value":"HIGH","recommended_next_step":"Before committing all units, preregister and simulate the exact Stage-A estimator under the planned weekly sampling count, missingness, block structure, and 8–12-variable state dimension. Proceed with the randomized Stage-A mesocosms only if that locked simulation shows acceptable mode-recovery and false-selection rates; then apply the stated residual, spectral-gap, identity, capacity, safety, and power gates without relaxing them.","world_novelty_boundary":"This bounded search supports only the contrastive conclusion that no close implementation appeared in the eight retained sources. It does not establish world novelty, patentability, freedom to operate, market size, prevalence, realized impact, or absence of relevant work under other terminology or outside the searched indexes."}