{"schema_version":1,"research_id":"eoa_inverse_innovation_exp03_external48_20260801","source_assessment_id":"eoa_inverse_innovation_exp03_opportunity320_20260801","cell_id":"layer_decay_and_expiration_management__marine_science","selection_stratum":"DEPLOYABLE_PRIORITY","search_queries":["site:oceanobservatories.org data product version provenance superseded data OOI","site:docs.dataone.org preservation versioning provenance data lifecycle obsolete versions","site:ioos.noaa.gov data management versioning archive provenance quality control","ocean observatory data versioning superseded data repository lifecycle archival restore","NDSA Levels of Digital Preservation fixity restore backup official 2019","CoreTrustSeal requirements data integrity preservation disposal appraisal repository official","AWS S3 Glacier Flexible Retrieval pricing official restore storage 2026","BLS employer costs employee compensation professional scientific technical services 2026","site:aws.amazon.com/s3/pricing S3 Glacier Deep Archive price GB month","site:docs.aws.amazon.com Amazon S3 lifecycle transitions storage classes official","site:osf.io ndsa levels preservation v2.1 matrix fixity inventory copies 2026","site:coretrustseal.org 2026 requirements appraisal preservation plan integrity authenticity reuse"],"sources":[{"source_id":"S1","title":"Data Updates and Corrections","publisher":"Ocean Observatories Initiative","url":"https://oceanobservatories.org/data-updates-corrections/","source_class":"OFFICIAL_ORGANIZATION_DATA","publication_date":"n.d.; continuously updated","accessed_at":"2026-08-02","claims_supported":["OOI maintains a change tracker because calibration, metadata, and processing changes can affect previously downloaded data.","OOI directs users to re-download affected data and, in a documented oxygen case, discourages analysis with an affected L1 product in favor of corrected L2 data.","The documented corrections demonstrate that authoritative-use status can change after data products have been distributed."]},{"source_id":"S2","title":"The Oceans 2.0/3.0 Data Management and Archival System","publisher":"Frontiers in Marine Science / Ocean Networks Canada authors","url":"https://doi.org/10.3389/fmars.2022.806452","source_class":"PRIMARY_RESEARCH","publication_date":"2022-03-08","accessed_at":"2026-08-02","claims_supported":["Ocean Networks Canada operates a deployed observatory data-management system with versioning, provenance, DOI relationships, reprocessing, landing pages, archival replicas, and consistency checks.","The authors report that earlier traceability was insufficient to reconstruct and communicate all events for a dataset, motivating new versioning infrastructure.","Some reprocessed older versions become unavailable, but associated metadata and landing-page relationships are preserved.","The system had 1.5 PB of storage in 2021, while one replica was receiving more than 10 TB and two million files per month; rising tape-library costs prompted a cloud transition.","ONC emphasizes preserving raw unmanipulated observations and avoiding removal of valid data."]},{"source_id":"S3","title":"PO.DAAC Data Management Best Practices","publisher":"NASA Jet Propulsion Laboratory Physical Oceanography Distributed Active Archive Center","url":"https://podaac.jpl.nasa.gov/PO.DAAC_DataManagementPractices","source_class":"OFFICIAL_GUIDANCE","publication_date":"n.d.","accessed_at":"2026-08-02","claims_supported":["PO.DAAC uses a formal dataset lifecycle with acceptance, cost-impact assessment, operational review, repeated approval for new versions, and retirement plans.","Datasets deprecated by newer versions can be removed from public visibility with reduced service, and erroneous datasets can be quarantined.","Dataset-provider and repository-manager approval roles are explicit."]},{"source_id":"S4","title":"Continuous cycle of evolution of products","publisher":"Copernicus Marine Service, implemented by Mercator Ocean International for the European Commission","url":"https://help.marine.copernicus.eu/en/articles/8606090-continuous-cycle-of-evolution-of-products","source_class":"OFFICIAL_PRODUCT_DOCUMENTATION","publication_date":"2025-05-20","accessed_at":"2026-08-02","claims_supported":["Copernicus Marine provides a transition period with old and new product versions, then makes the older version inaccessible.","The service explicitly links retirement of old products to current-product reliability and infrastructure-storage costs.","The documented process lacks post-transition retrieval of retired versions, illustrating the reproducibility risk of irreversible retirement."]},{"source_id":"S5","title":"Storage and Archiving","publisher":"U.S. Integrated Ocean Observing System, NOAA","url":"https://ioos.noaa.gov/data/data-standards/storage-and-archiving/","source_class":"GOVERNMENT_OR_REGULATOR","publication_date":"n.d.","accessed_at":"2026-08-02","claims_supported":["IOOS partners must provide initial stewardship and permanent archiving of archival-quality observations, model outputs, metadata, and derived products.","Providers are expected to retain supporting documentation and algorithms, establish recovery mechanisms, and maintain off-site backups pending NCEI archiving.","NCEI is identified as the nominal archive and institutional partner for IOOS data."]},{"source_id":"S6","title":"Dataset modification and versioning","publisher":"British Oceanographic Data Centre, National Oceanography Centre","url":"https://www.bodc.ac.uk/submit_data/data_citations/modification_and_versioning/","source_class":"OFFICIAL_GUIDANCE","publication_date":"n.d.","accessed_at":"2026-08-02","claims_supported":["BODC publishes changed datasets as new versions with new DOIs and a landing page containing version history.","The immediately previous landing page routes users to its successor.","BODC retains every published version indefinitely, showing an established preservation-heavy alternative to active-tier retirement."]},{"source_id":"S7","title":"Transitioning objects using Amazon S3 Lifecycle","publisher":"Amazon Web Services","url":"https://docs.aws.amazon.com/AmazonS3/latest/userguide/lifecycle-transition-general-considerations.html","source_class":"OFFICIAL_PRODUCT_DOCUMENTATION","publication_date":"n.d.; current documentation","accessed_at":"2026-08-02","claims_supported":["Existing object-storage services support rules that transition data among active, infrequent-access, and archival tiers.","Archived objects require an asynchronous restore to a temporary copy and are not immediately accessible.","Object size, request charges, minimum storage durations, retrieval, and temporary restored copies materially affect lifecycle economics and retrieval design."]},{"source_id":"S8","title":"Employer costs per hour worked for employee compensation of private industry workers, by industry","publisher":"U.S. Bureau of Labor Statistics","url":"https://www.bls.gov/charts/employer-costs-for-employee-compensation/costs-by-industry.htm","source_class":"GOVERNMENT_OR_REGULATOR","publication_date":"2026-03","accessed_at":"2026-08-02","claims_supported":["March 2026 compensation data provide an official basis for converting specialist labor into calendar-2026 resource-equivalent cost bands.","Professional and technical services and educational services carry substantial wage-plus-benefit costs, making multi-role stewardship and engineering work the principal cost driver in a bounded pilot."]}],"problem_evidence":{"support":"MODERATE","rationale":"The underlying pattern is externally supported: OOI documents corrections that invalidate prior downloads or change recommended product levels, ONC reports reconstruction limitations and rapidly growing archival volume, and Copernicus explicitly retires old versions partly to control storage costs. However, no source quantifies how often users actually select superseded observatory layers, how much active storage is attributable specifically to redundant versions, or the resulting steward workload at the intended adopter.","source_ids":["S1","S2","S4"]},"stakeholder_evidence":{"support":"MODERATE","rationale":"OOI has built user-facing change tracking and re-download guidance, NASA PO.DAAC operates formal approval and retirement workflows, and IOOS imposes stewardship, recovery, and permanent-archive duties. These establish credible institutional users and authorizers, but no source documents demand, budget commitment, complaints, or willingness to adopt this candidate's complete mechanism composition.","source_ids":["S1","S3","S5"]},"prior_art":{"proximity":"SUBSTANTIAL_COLLISION","closest_analogues":[{"name":"NASA PO.DAAC dataset lifecycle and retirement process","similarity":"It already combines formal lifecycle stages, prioritization, cost-impact review, repeated approval for new versions, retirement of superseded datasets, reduced public visibility, retirement plans, and quarantine of erroneous data.","remaining_difference":"The public guidance does not specify citation- and derivation-dependency vetoes, reversible status simulation, persistent successor tombstones for every retired layer, sampled reconstruction tests, or periodic revalidation of exceptions as one integrated workflow.","source_ids":["S3"]},{"name":"Ocean Networks Canada Oceans 3.0 versioning and archival system","similarity":"It operates within ocean observatories and combines raw-data preservation, reprocessing, version provenance, linked version DOIs, persistent landing-page metadata, archival replicas, integrity checks, and recovery infrastructure.","remaining_difference":"The paper does not describe a unified decision policy that scores candidates for active-tier demotion while requiring authority holds, citation and derivation dependency vetoes, exception expiry, and comparative restore drills. Periodic replica-consistency checking is not the same as demonstrating end-to-end reconstruction from a demoted layer.","source_ids":["S2"]},{"name":"Copernicus Marine product transition and retirement","similarity":"It runs parallel old/new product versions during a transition, routes users toward current products, and then removes the old products to reduce infrastructure-storage cost.","remaining_difference":"Retired versions cannot be recovered after the transition, so it lacks the candidate's archive, tombstone, dependency, and restoration safety net.","source_ids":["S4"]},{"name":"BODC Published Data Library versioning","similarity":"It assigns new versions and DOIs, preserves version histories, routes a previous landing page to its successor, and retains immutable earlier versions.","remaining_difference":"It preserves all versions indefinitely and does not address governed active-tier reduction, disposition prioritization, restoration latency, or expiring exceptions.","source_ids":["S6"]}],"distinctive_claim_remaining":"The remaining testable claim is not lifecycle management itself. It is that, for a continuously processed ocean-observatory collection, adding explicit citation/derivation dependency vetoes, jointly authorized reversible tier states, expiring exceptions, persistent successor routing, and sampled end-to-end reconstruction tests will outperform authoritative-version labeling alone on obsolete-version selection and steward effort without increasing broken citations, missed dependencies, retrieval delay beyond a preregistered tolerance, or reconstruction failures.","confidence":"HIGH"},"implementation_evidence":{"support":"STRONG","rationale":"Most components are operational elsewhere: PO.DAAC implements approval, version replacement, quarantine, and retirement; ONC implements observatory-scale provenance, version-linked identifiers, raw preservation, replication, and integrity checking; BODC implements persistent successor routing; and commercial object storage implements automated tier transitions and restoration. The unverified part is reliable integration of historical citation and derivation dependencies with governance vetoes and repeated reconstruction tests in the target corpus.","source_ids":["S2","S3","S6","S7"]},"scores":{"meaningful_impact":{"score":4,"rationale":"OOI documents scientifically consequential corrections and ONC documents large, fast-growing storage. Correct authority signaling with safe tiering could therefore affect both analysis validity and stewardship resources, although the candidate's realized effect size is unmeasured.","source_ids":["S1","S2"]},"stakeholder_pull":{"score":3,"rationale":"Observatories and archives already invest in change tracking, versioning, recovery, and retirement, demonstrating a recognized need. There is no direct request or adoption commitment for this exact composition.","source_ids":["S1","S3","S5"]},"incremental_advantage":{"score":2,"rationale":"The label-only rival is too weak relative to actual practice: PO.DAAC and ONC already implement much of the lifecycle. Dependency vetoes, expiring exceptions, and reconstruction testing may add value, but their incremental benefit has not been compared with these stronger baselines.","source_ids":["S2","S3"]},"distinctiveness_plausibility":{"score":2,"rationale":"No single opened source described every proposed control, but the core composition substantially overlaps deployed ocean and Earth-science repository practice. Only the tightly integrated dependency-aware and experimentally validated governance layer remains plausibly distinctive.","source_ids":["S2","S3","S4","S6"]},"technical_implementability":{"score":4,"rationale":"Version identifiers, provenance, quarantine, tier transitions, persistent landing pages, checksums, and restoration are demonstrated capabilities. Historical identity resolution, citation discovery, and derivation graphs remain difficult integration tasks.","source_ids":["S2","S3","S6","S7"]},"adoption_authority_feasibility":{"score":3,"rationale":"PO.DAAC identifies provider and repository-manager approvals, while IOOS assigns stewardship and archive responsibilities. A real observatory can therefore identify authorizers, but cross-institutional records, governance, funder, and scientific approvals may still delay any disposition policy.","source_ids":["S3","S5"]},"evidence_readiness":{"score":4,"rationale":"A completed instrument-year, current catalog, access logs, version metadata, and archive can support a read-only comparison of baseline, label-only, and simulated governed lifecycle states. Missing lineage and citations are measurable failure outcomes rather than reasons to deploy blindly.","source_ids":["S1","S2","S3"]},"safety_net_benefit":{"score":4,"rationale":"Raw preservation, successor routing, quarantine, recovery mechanisms, integrity checks, and restoration are established safeguards. Combining them as mandatory vetoes should reduce irreversible-loss risk, although archive credentials, formats, and dependencies can still decay.","source_ids":["S2","S5","S6","S7"]},"scalability":{"score":3,"rationale":"Operational systems demonstrate observatory- and repository-scale lifecycle components, but ONC's millions of incoming files and AWS's object-count-sensitive archival costs show that heterogeneous metadata and small-file volume can limit economical automation.","source_ids":["S2","S3","S7"]}},"score_confidence":"MODERATE","costs":{"first_evidence":{"band_2026_usd":"50K_TO_250K","scope":"Read-only study of one completed instrument-year: inventory at most 500 version families; reconstruct version, citation, hold, and derivation links; replay catalog tasks under baseline, label-only, and simulated governed states; inspect access logs; and restore/checksum a stratified sample of up to 30 archived objects. Includes steward, repository engineer, instrument scientist, analyst, user-testing, coordination, software, and evaluation labor.","confidence":"MODERATE","assumptions":["Existing catalog, logs, archive credentials, and processing documentation are available without procurement.","The study uses existing infrastructure and makes no production metadata, visibility, storage-tier, or deletion changes.","Specialist labor dominates cost; the band reflects 2026 wage-plus-benefit conditions and allows several person-months across multiple roles.","No more than 500 version families and 30 restoration cases are examined."],"source_ids":["S1","S2","S7","S8"]},"initial_deployment_startup":{"band_2026_usd":"250K_TO_1M","scope":"Design and build an initial production-capable lifecycle service for one observatory collection family: state and authority model, provenance/dependency integration, holds, successor routing, audit logs, reversible tiering, restoration workflow, policy and records review, security testing, and evaluation instrumentation.","confidence":"LOW","assumptions":["The observatory already has a catalog, object/archive storage, persistent identifiers, authentication, and backup systems.","Historical dependencies require substantial cleanup but not complete reconstruction of the full legacy corpus.","Existing institutional records, ethics, sovereignty, and funder-review channels can evaluate the policy.","The work includes engineering, data stewardship, scientific validation, security, compliance, software, storage testing, coordination, and evaluation."],"source_ids":["S2","S3","S5","S6","S7","S8"]},"operational_launch":{"band_2026_usd":"250K_TO_1M","scope":"Phased launch across selected collections within one observatory, including backlog triage, dual-running with existing catalog behavior, user communication, training, archive transfers, checksum and restore validation, exception adjudication, monitoring, support, and an independent launch evaluation. It excludes automatic hard deletion and full remediation of every historical collection.","confidence":"LOW","assumptions":["The launch is phased and limited to selected collections rather than the entire legacy observatory corpus.","Unresolved identities, dependencies, citations, or authority remain held and visible to stewards.","Existing storage and identifier services can be integrated instead of replaced.","Small-file transition, retrieval, and temporary-restore costs are tested before scale-up."],"source_ids":["S2","S3","S5","S7","S8"]},"annual_recurring":{"band_2026_usd":"250K_TO_1M","scope":"Ongoing stewardship and engineering for lifecycle reviews, exception expiry, dependency and citation updates, archive storage and retrieval, checksum monitoring, scheduled restore/reconstruction drills, tombstone maintenance, incident response, user support, compliance review, software maintenance, and annual effectiveness evaluation.","confidence":"LOW","assumptions":["At least two specialist functions require sustained fractional or full-time coverage because scientific exceptions cannot be fully automated.","Routine lifecycle decisions are automated, while ambiguous or governed records receive human review.","Archive volume and object counts remain within existing infrastructure scale; major migrations or disaster recovery are excluded.","The upper portion accommodates object-count-sensitive storage operations, retrieval exercises, compliance, and external evaluation."],"source_ids":["S2","S5","S7","S8"]}},"verified_pipeline_gates":{"externally_supported_problem":{"status":"YES","reason":"Independent official and primary sources establish the general problem pattern: corrections can invalidate prior products, provenance gaps can impede reconstruction, rapidly growing archives impose stewardship pressure, and some marine services retire older products for infrastructure reasons. Target-specific prevalence remains unquantified.","source_ids":["S1","S2","S4"]},"externally_credible_adopter_or_authorizer":{"status":"YES","reason":"OOI data teams, observatory stewards, repository managers and data providers already exercise relevant responsibilities; IOOS formally assigns archive and recovery duties, and PO.DAAC documents provider-manager approval processes.","source_ids":["S1","S3","S5"]},"distinct_testable_incremental_claim":{"status":"YES","reason":"The residual claim can be compared against both current manual practice and the stronger label-plus-successor baseline: dependency-aware reversible lifecycle governance must reduce obsolete selection or review effort without degrading citation resolution, reconstruction, or retrieval.","source_ids":["S2","S3","S6"]},"bounded_next_evidence_step":{"status":"YES","reason":"A capped, read-only instrument-year study can simulate all three catalog conditions and perform sampled archive restorations without deleting, hiding, altering, or moving production data.","source_ids":["S1","S2","S7"]},"no_unresolved_safety_or_authority_stop":{"status":"YES","reason":"The next step is non-destructive. Permanent archiving and recovery obligations make unresolved identity, authority, citations, dependencies, checksum failure, or reconstruction failure explicit stop conditions rather than permissible dispositions.","source_ids":["S2","S5"]},"credible_cost_scope_and_range":{"status":"YES","reason":"The four scopes identify labor, software, data cleanup, governance, compliance, coordination, storage, retrieval, and evaluation. Official 2026 compensation data anchor labor resource equivalence, while operational observatory and storage documentation justify uncertainty about corpus and object-count costs.","source_ids":["S2","S7","S8"]}},"next_evidence_step":"Conduct a preregistered, read-only study on one completed instrument-year, capped at 500 version families and 30 stratified archive restorations. Have two stewards independently establish authoritative versions, citations, derivation dependencies, holds, and unresolved cases; then compare paired catalog tasks under (A) current practice, (B) authoritative-version labels plus successor links, and (C) the full simulated lifecycle with dependency vetoes and expiring exceptions. Measure superseded-version selection, task completion, steward review minutes, unresolved-family rate, projected active-tier bytes, broken citation or dependency routes, checksum/reconstruction success, and restoration delay. Falsify the intervention if C does not improve obsolete-selection or review effort over B by a preregistered meaningful margin, or if C causes any previously resolvable citation/dependency to fail, any checksum or reconstruction failure, or restoration delay beyond the preregistered user tolerance. Make no production status, visibility, tier, or deletion changes.","blocking_evidence":["No representative observatory audit currently quantifies superseded-version discovery, access, selection, active-tier footprint, or disposition workload.","Completeness of historical version identity, citations, derivation dependencies, legal or community holds, and disposition authority is unknown.","No controlled comparison shows that the full mechanism composition outperforms the stronger authoritative-label-plus-successor baseline.","End-to-end reconstruction and acceptable restoration latency have not been demonstrated for a stratified sample of the target corpus.","Corpus volume, object-size distribution, existing integrations, backlog quality, staffing, compliance effort, and retrieval frequency are unknown, limiting cost confidence.","The bounded search found substantial prior art but did not locate outcome evaluations of dependency-aware lifecycle governance in a continuous ocean observatory."],"research_disposition":"KNOWN_PRACTICE_DIFFUSION","world_novelty_boundary":"This bounded search found substantial collision with deployed NASA PO.DAAC, Ocean Networks Canada, Copernicus Marine, BODC, IOOS, and general object-storage lifecycle practice. It does not establish that the residual dependency-aware governance composition is new, absent elsewhere, patentable, or globally distinctive. The only defensible boundary is a local comparative claim about integrating dependency vetoes, reversible states, expiring exceptions, successor routing, and reconstruction tests in the selected observatory context."}