{"schema_version":1,"research_id":"eoa_inverse_innovation_exp06_external_evaluation_20260803","source_assessment_id":"representation_independent_interface_contract__environmental_climate:P2:v0","cell_id":"representation_independent_interface_contract__environmental_climate","search_queries":["Basic Model Interface BMI 2.0 paper environmental models component coupling semantics state update official","snow model modular interface land surface model SUMMA snow process model structure paper","NASA Land Information System plug and play land surface models framework official","ESMF component interface state import export official reference guide","ESM-SnowMIP snow model intercomparison divergence snow schemes GMD 2020","OGC OpenMI standard 2.0 official model component interface environmental models pdf","CCPP technical documentation physics suites interface standard host model official NOAA","snow model conservation energy mass test benchmark modular coupling interface","\"Enhancing the modularity and interoperability\" SUMMA 2025 pdf","\"When and Where Are Multiple Snow Layers Important\" pdf","site:gmd.copernicus.org/articles/11/5027/2018 ESM-SnowMIP full article","doi 10.1016/j.envsoft.2025.106668","doi 10.1029/2020WR028993","site:papers.ssrn.com/sol3/papers.cfm abstract_id 5194913 SUMMA ROO","site:ral.ucar.edu \"Enhancing the modularity\" SUMMA"],"sources":[{"source_id":"S1","title":"The Basic Model Interface 2.0: A standard interface for coupling numerical models in the geosciences","publisher":"Journal of Open Source Software","url":"https://www.theoj.org/joss-papers/joss.02317/10.21105.joss.02317.pdf","source_class":"PRIMARY_RESEARCH","publication_date":"2020-07-23","accessed_at":"2026-08-03","claims_supported":["Geoscience component modeling already uses standard interfaces for independent, reusable components.","BMI supplies standard functions for querying, modifying, and running models and has specifications for multiple programming languages.","A representation-independent snow-module interface is therefore adjacent to established geoscience component practice."]},{"source_id":"S2","title":"The Basic Model Interface: BMI 2.0 specification","publisher":"Community Surface Dynamics Modeling System","url":"https://bmi.csdms.io/en/stable/bmi.spec.html","source_class":"OFFICIAL_PRODUCT_DOCUMENTATION","publication_date":"n.d.","accessed_at":"2026-08-03","claims_supported":["BMI already defines initialize, update, update-until, finalize, variable metadata, time, grid, getter, and setter operations.","BMI includes get_value_ptr and set_value, illustrating that a standard interface can still expose or mutate model variables rather than enforce the proposal's opaque aggregate state.","The proposed contract's opacity, conservation ledger, stable errors, and checkpoint-conformance rules are narrower additions rather than a new general coupling concept."]},{"source_id":"S3","title":"OGC Open Modelling Interface Interface Standard, Version 2.0 (OGC 11-014r3)","publisher":"Open Geospatial Consortium","url":"https://docs.ogc.org/is/11-014r3/11-014r3.html","source_class":"STANDARD","publication_date":"2014-05-26","accessed_at":"2026-08-03","claims_supported":["A consensus standard already enables independently developed environmental process models to exchange data at runtime.","OpenMI defines mandatory conformance requirements and optional state-management interfaces.","Its stated goals include operational uptake, conformance, testbeds, experiments, pilots, and use by public and private bodies."]},{"source_id":"S4","title":"Land Information System","publisher":"NASA Goddard Space Flight Center","url":"https://lis.gsfc.nasa.gov/software/lis","source_class":"GOVERNMENT_OR_REGULATOR","publication_date":"n.d.","accessed_at":"2026-08-03","claims_supported":["NASA LIS is an identifiable adopter of interoperable, portable, encapsulated land-surface components.","LIS integrates multiple community land-surface models and represents snowpack and runoff states and fluxes.","NASA's operational research infrastructure demonstrates both stakeholder need and authority for component-interface work."]},{"source_id":"S5","title":"Common Community Physics Package Technical Documentation: Overview","publisher":"Developmental Testbed Center / NCAR","url":"https://ccpp-techdoc.readthedocs.io/en/latest/Overview.html","source_class":"OFFICIAL_PRODUCT_DOCUMENTATION","publication_date":"n.d.","accessed_at":"2026-08-03","claims_supported":["NOAA, NCAR, NRL, and other agencies have expressed a longstanding need to run physics suites in different host models.","CCPP checks host-provided variables against scheme requirements, communicates variables at runtime, and supports init/run/finalize lifecycle methods.","CCPP is used operationally in UFS applications, has a single-column test host, and identifies governance bodies able to authorize interface changes."]},{"source_id":"S6","title":"Structure for Unifying Multiple Modeling Alternatives (SUMMA)","publisher":"NSF National Center for Atmospheric Research","url":"https://ral.ucar.edu/model/summa","source_class":"OFFICIAL_ORGANIZATION_DATA","publication_date":"n.d.","accessed_at":"2026-08-03","claims_supported":["SUMMA already separates conservation equations from numerical solution and supports alternative process representations within a common structural core.","SUMMA supplies synthetic, laboratory, and field test cases, including snowmelt and snow-albedo cases.","SUMMA demonstrates established conservation-centered modular practice and provides a plausible incumbent or comparator, subject to GPLv3 obligations."]},{"source_id":"S7","title":"Version 3.0.2 of the Crocus snowpack model","publisher":"Geoscientific Model Development / Copernicus Publications","url":"https://gmd.copernicus.org/articles/19/6273/2026/","source_class":"PRIMARY_RESEARCH","publication_date":"2026","accessed_at":"2026-08-03","claims_supported":["A current multilayer snow model has been externalized as an independent Fortran library because of demand for coupling it to multiple land-surface schemes.","Copied implementations have diverged, and the authors advocate convergence on one maintained repository.","Variable layer counts, storage order, vectorization, numerical cost, multiple physical options, and cross-site benchmarks make representation and substitutability concerns concrete."]},{"source_id":"S8","title":"ESM-SnowMIP: assessing snow models and quantifying snow-related climate feedbacks","publisher":"Geoscientific Model Development / Copernicus Publications","url":"https://gmd.copernicus.org/articles/11/5027/2018/gmd-11-5027-2018.html","source_class":"PRIMARY_RESEARCH","publication_date":"2018-12-10","accessed_at":"2026-08-03","claims_supported":["Snow is consequential to surface energy, hydrology, climate feedbacks, and soil temperature.","Snow implementations produce materially different behavior and are evaluated through coordinated site, global, offline, and coupled experiments.","Single-column reference-site simulations offer an established low-cost benchmark format, but observational and forcing limitations remain."]}],"problem_evidence":{"support":"MODERATE","rationale":"The general problem visibly exists: BMI exposes pointers and setters; current Crocus work reports coupling demand, copied-code divergence, variable layer representations, and hardware-sensitive storage choices; SUMMA exists specifically to separate conservation equations from numerical realization. Snow behavior matters to water, energy, and climate outcomes. However, no direct dependency inventory from a named host–snow pair was found showing the exact alleged pattern of host code indexing or mutating private snow layers. The candidate's component-specific prevalence claim is therefore inferred, not verified.","source_ids":["S2","S6","S7","S8"]},"stakeholder_evidence":{"support":"STRONG","rationale":"Identifiable adopters and authorizers exist. NASA LIS explicitly encapsulates and interoperates land-surface models. NOAA/NCAR/NRL govern CCPP, whose documented goal is physics interoperability across hosts and whose schemes are used operationally. CSDMS and OGC maintain model-interface specifications. These sources establish institutional demand, although none expressly requests this exact snow-specific conservation-bound profile.","source_ids":["S1","S3","S4","S5","S7"]},"prior_art":{"proximity":"ESTABLISHED_PRACTICE","closest_analogues":[{"name":"Basic Model Interface 2.0","similarity":"Standard, language-neutral initialize/update/finalize and model-variable interface intended to make geoscience components reusable and couplable.","remaining_difference":"BMI permits direct getters, pointer access, and setters and does not itself impose the proposed opaque snow totals, conservation ledger, stable failure semantics, checkpoint reproducibility, or behavioral oracle.","source_ids":["S1","S2"]},{"name":"OGC OpenMI 2.0","similarity":"Consensus interface and conformance framework for independently developed environmental process-model components exchanging data at runtime, with optional state management.","remaining_difference":"OpenMI is a general runtime-exchange standard, not a snow-specific semantic profile with mass/energy laws and implementation-substitution tolerances.","source_ids":["S3"]},{"name":"NASA LIS and NOAA/NCAR CCPP","similarity":"Operationally relevant host frameworks encapsulate land or physics components, use documented variables and lifecycle calls, test components in single-column hosts, and have explicit governance.","remaining_difference":"The reviewed documentation does not establish one shared snow-specific black-box oracle based on aggregate state, checkpoint continuation, conservation residuals, stable error categories, and leakage audits.","source_ids":["S4","S5"]},{"name":"SUMMA and externalized Crocus","similarity":"Existing hydrologic and snow-model practice separates conservation equations from numerical choices, supports alternative physical representations, externalizes snow code for multiple host systems, and uses shared benchmarks.","remaining_difference":"The opened sources do not document a complete opaque abstract-data-type contract that forbids host reliance on layers and accepts independently implemented solvers solely through a versioned behavioral conformance suite.","source_ids":["S6","S7","S8"]}],"distinctive_claim_remaining":"Relative to a conventional BMI/OpenMI/CCPP-style variable interface, a snow-specific profile that exposes only versioned aggregate stores and boundary fluxes and gates implementations with prespecified conservation, error, leakage, and checkpoint-continuation tests will permit a change in layer/solver/device representation with fewer host-code edits and no additional contracted flux or conservation discrepancies. This is falsified if the profile cannot express necessary host behavior without layer access, does not reduce host edits, or two passing implementations diverge beyond prespecified boundary tolerances.","confidence":"HIGH"},"implementation_evidence":{"support":"STRONG","rationale":"The building blocks are implemented practice: lifecycle interfaces, language-neutral specifications, model conformance requirements, single-column hosts, interoperable land-model frameworks, independent snow libraries, common conservation structures, and reference-site benchmarks. Technical implementation is credible. Remaining challenges are semantic rather than fundamental engineering: defining energy reference states, tolerances, parameter identity, checkpoint meaning, sanctioned diagnostics, and handling implementations that are scientifically different rather than merely representationally different.","source_ids":["S1","S2","S3","S4","S5","S6","S7","S8"]},"scores":{"meaningful_impact":{"score":4,"rationale":"Snow affects hydrology, surface energy, soil temperature, and climate feedbacks, while silent coupling changes could confound scientific attribution. Realized impact is unmeasured.","source_ids":["S7","S8"]},"stakeholder_pull":{"score":4,"rationale":"NASA, NOAA, NCAR, NRL, CSDMS, and OGC visibly invest in interoperable environmental-model components, although exact demand for this contract profile was not found.","source_ids":["S3","S4","S5"]},"incremental_advantage":{"score":3,"rationale":"Opaque aggregate state plus conservation, checkpoint, and error conformance is a plausible improvement over variable-level BMI-style coupling, but its advantage over disciplined use of existing standards has not been measured.","source_ids":["S2","S3","S5","S6"]},"distinctiveness_plausibility":{"score":2,"rationale":"The broad concept substantially overlaps established interface, conformance, conservation-centered modularity, and single-column testing practice. Only the combined snow-specific semantic profile remains potentially distinctive.","source_ids":["S1","S2","S3","S5","S6","S7"]},"technical_implementability":{"score":4,"rationale":"Existing frameworks demonstrate all major technical primitives. Ambiguous scientific semantics and tolerance design, rather than unavailable technology, are the principal risks.","source_ids":["S2","S3","S5","S6","S7"]},"adoption_authority_feasibility":{"score":3,"rationale":"Named governance and integration authorities exist, but authorization is system-specific and operational replacement requires scientific and operational review beyond a software lead.","source_ids":["S4","S5"]},"evidence_readiness":{"score":3,"rationale":"Public standards, implementations, and benchmark datasets make a prototype readily testable, but there is no dependency inventory, adopter commitment, or two-implementation result for the proposed contract.","source_ids":["S5","S6","S7","S8"]},"safety_net_benefit":{"score":3,"rationale":"Shadow execution, immutable errors, checkpoint comparison, conservation gates, and rollback can reduce the chance that implementation changes contaminate canonical experiments. These controls do not establish physical validity.","source_ids":["S5","S6","S8"]},"scalability":{"score":4,"rationale":"Language-neutral interfaces and shared conformance suites can be reused across hosts and implementations, as BMI, OpenMI, LIS, CCPP, and Crocus coupling already demonstrate.","source_ids":["S1","S3","S4","S5","S7"]}},"score_confidence":"MODERATE","costs":{"first_evidence":{"band_2026_usd":"10K_TO_50K","scope":"A four-to-eight-week, non-operational single-column prototype: dependency inventory, standards crosswalk, draft semantic contract, incumbent adapter, minimal independent implementation, and frozen tests.","confidence":"LOW","assumptions":["One snow scientist at approximately 0.1–0.25 FTE for two months.","One research-software engineer at approximately 0.5–1.0 FTE for two months.","Existing open-source model, forcing series, test host, and compute allocation are usable.","No procurement, production accreditation, or new observational campaign."],"source_ids":["S5","S6","S7","S8"]},"initial_deployment_startup":{"band_2026_usd":"50K_TO_250K","scope":"Harden the contract, implement two production-quality adapters, develop property and checkpoint tests, add CI across supported compilers/devices, document versioning, and integrate one research host.","confidence":"LOW","assumptions":["Approximately 0.5–1.5 FTE-years split among software and snow-domain staff.","Existing BMI/CCPP/OpenMI concepts and benchmark data are reused.","Only one host and two snow implementations are included.","No operational forecast authorization."],"source_ids":["S1","S3","S5","S7"]},"operational_launch":{"band_2026_usd":"250K_TO_1M","scope":"Operational hardening, independent review, reproducibility and performance qualification, workflow migration, staged shadow runs, rollback tooling, documentation, and governance for one operational or publication-bound modeling system.","confidence":"LOW","assumptions":["Approximately two-to-five multidisciplinary FTE-years.","Operational regression data and compute are available.","Scientific recalibration and new field observations are excluded.","Security, software-assurance, and release requirements resemble existing institutional frameworks."],"source_ids":["S4","S5","S7","S8"]},"annual_recurring":{"band_2026_usd":"50K_TO_250K","scope":"Contract stewardship, adapter maintenance, CI and benchmark compute, compatibility review, incident triage, and periodic requalification.","confidence":"LOW","assumptions":["Approximately 0.5–1.5 continuing FTE.","A limited implementation and platform matrix.","Benchmark compute remains modest relative to full coupled production runs.","No major scientific contract revision in a typical year."],"source_ids":["S5","S7","S8"]}},"verified_pipeline_gates":{"externally_supported_problem":{"status":"UNCERTAIN","reason":"External evidence supports representation diversity, coupling demand, code divergence, and scientific consequences, but does not directly verify private-layer reach-through in a named host–snow integration.","source_ids":["S2","S6","S7","S8"]},"externally_credible_adopter_or_authorizer":{"status":"YES","reason":"NASA LIS, NOAA/NCAR/NRL CCPP governance, CSDMS, and OGC are credible adopters or authorizers for model-interface work.","source_ids":["S1","S3","S4","S5"]},"distinct_testable_incremental_claim":{"status":"YES","reason":"The remaining claim contrasts an opaque, conservation- and checkpoint-gated snow profile with variable-level interfaces and specifies observable host edits and boundary discrepancies as outcomes.","source_ids":["S2","S3","S5","S6"]},"bounded_next_evidence_step":{"status":"YES","reason":"A single non-operational column, frozen forcing series, incumbent implementation, independent simplified implementation, and prespecified comparator tests form a bounded experiment.","source_ids":["S5","S6","S7","S8"]},"no_unresolved_safety_or_authority_stop":{"status":"YES","reason":"An isolated shadow prototype is low risk if it cannot alter canonical experiments. Production substitution remains expressly outside scope and requires scientific and operational authorization.","source_ids":["S4","S5"]},"credible_cost_scope_and_range":{"status":"UNCERTAIN","reason":"Work packages can be bounded, but the eight direct sources provide no labor rates, integration effort measurements, vendor quotes, or historical project-cost data. Bands are resource-equivalent estimates only.","source_ids":[]}},"next_evidence_step":"Treat the concept as a diffusion and specialization exercise, not a new general interface invention. In one six-week, non-operational single-column study, first inventory every host read/write against one incumbent snow module and crosswalk it to BMI 2.0, OpenMI 2.0, CCPP, SUMMA, and the externalized Crocus interface. Freeze one initial state and forcing series before execution. Compare: (A) the existing integration, (B) a conventional variable-level BMI-style adapter, and (C) the proposed opaque aggregate-state adapter, each driving the incumbent and an independently coded simplified implementation. Prespecify mass and energy residual limits, flux tolerances, checkpoint continuation tolerance, stable error classes, allowed diagnostics, host-code edits, runtime, and memory. Falsify the problem if no host dependency reaches below documented totals and fluxes and a representation change needs no host edit. Falsify the intervention if necessary host behavior requires layer access, the opaque adapter does not reduce representation-dependent edits relative to B, either adapter changes the incumbent beyond tolerance, or two implementations pass the oracle yet disagree on a contracted result. Do not alter or reinterpret a canonical experiment.","blocking_evidence":["No public dependency inventory demonstrates host reliance on private snow-layer or solver state in a named target integration.","No adopter has endorsed the exact aggregate state, conservation equations, tolerances, stable error taxonomy, or checkpoint semantics.","No comparative benchmark shows lower host-edit effort or fewer semantic discrepancies than a disciplined BMI/OpenMI/CCPP implementation.","No evidence establishes that all scientifically necessary snow-host exchanges can be expressed without sanctioned layer-level diagnostics.","No implementation-independent evidence distinguishes representational equivalence from scientifically different snow formulations for the proposed state variables.","No direct cost observations support the 2026 resource bands.","World novelty, patentability, freedom to operate, market size, and realized impact were not measured."],"research_disposition":"KNOWN_PRACTICE_DIFFUSION","world_novelty_boundary":"Broad geoscience component interfaces, environmental runtime-exchange standards, conformance requirements, lifecycle methods, conservation-centered modular hydrology, independent snow libraries, single-column hosts, and shared snow benchmarks are established. The only potentially distinctive boundary is the precise combination of opaque aggregate snow state, versioned mass/energy and checkpoint semantics, stable errors, leakage auditing, and one cross-implementation acceptance oracle. This search does not establish world novelty, patentability, freedom to operate, market size, or realized impact.","arm":"COMPLETE_PROPOSAL_PORTFOLIO","candidate_version":0,"controller_recommendation":{"action":"STOP_DIFFUSION","repairable":false,"material_progress_observed":true,"progress_targets":["Reframe the proposal as a snow-specific implementation profile or extension of BMI/OpenMI/CCPP rather than a novel general component-interface architecture.","Select one named host and snow module and publish a dependency inventory establishing whether private representation coupling actually occurs.","Produce a clause-by-clause standards and prior-art crosswalk, especially for OpenMI state management, BMI pointer/setter access, CCPP lifecycle and governance, SUMMA conservation structure, and Crocus externalization.","Prespecify and run the three-way comparator study: existing integration, conventional variable adapter, and opaque conservation-bound adapter.","Obtain written approval from the relevant snow-science and model-integration authorities for abstract quantities, tolerances, checkpoint meaning, and operational exclusion boundaries.","Replace resource-equivalent cost assumptions with observed staff time and compute consumption from the bounded prototype."],"reason":"The broad intervention is established practice across BMI, OpenMI, LIS, CCPP, SUMMA, and externalized snow-model workflows. Web evidence supports diffusion and snow-specific profiling, not a sufficiently distinctive general innovation. The exact incremental benefit still requires a live two-implementation benchmark, but that does not repair the substantial prior-art overlap of the overarching proposal."},"proposal_index":2}