{"schema_version":1,"research_id":"eoa_inverse_innovation_exp06_external_evaluation_20260803","source_assessment_id":"predictive_residual_processing__environmental_climate:P1:v0","cell_id":"predictive_residual_processing__environmental_climate","search_queries":["peatland water table monitoring remote sensors energy communications telemetry peatland restoration official guidance","predictive data compression wireless sensor network environmental monitoring residual transmission research","Peatland Code monitoring water table depth requirements official","remote peatland monitoring LoRaWAN satellite telemetry energy constrained sensors research","site:nature.com peatland water table fire risk rewetting primary research","peat soil moisture temperature fire risk water table primary study peatlands","LoRaWAN specification maximum payload official LoRa Alliance data rate energy","environmental datalogger remote telemetry price water level soil moisture Campbell Scientific official","2026 price LoRaWAN water level sensor Decentlab price","remote environmental monitoring station water level rainfall soil moisture price USD telemetry","Campbell Scientific CR310 price USD 2026 cellular datalogger","Onset HOBO MicroRX water level station price official","\"Data prediction, compression, and recovery\" 2016 DOI 10.1016/j.ins.2015.10.004","site:sciencedirect.com S0020025515007124 data prediction compression recovery abstract","10.1016/j.ins.2015.10.004 full text","A Survey about Prediction-Based Data Reduction in Wireless Sensor Networks ACM Computing Surveys publication 2017 Dias Bellalta Oechsner"],"sources":[{"source_id":"S1","title":"Apply for a lowland peat water implementation grant","publisher":"UK Department for Environment, Food & Rural Affairs and Environment Agency","url":"https://www.gov.uk/government/publications/apply-for-a-lowland-peat-water-implementation-grant/apply-for-a-lowland-peat-water-implementation-grant","source_class":"GOVERNMENT_OR_REGULATOR","publication_date":"2026-06-22","accessed_at":"2026-08-03","claims_supported":["The Environment Agency is an identifiable funder and authorizer for English lowland-peat projects, with grants normally between £100,000 and £2,000,000.","Applicants must provide baseline data and future monitoring proposals; robust, high-quality monitoring is described as essential.","Water-table depth, surface-water level, and rainfall are minimum measurements, while conductivity and soil moisture may be relevant additions.","Telemetry is an eligible example of project infrastructure, and applications must itemize equipment, staff, contractor, and consultant costs."]},{"source_id":"S2","title":"Peatland Restoration for Greenhouse Gas Emission Reduction and Carbon Sequestration in the Baltic Sea Region: 1st Monitoring Report, Finland","publisher":"LIFE PeatCarbon consortium / University of Latvia","url":"https://www.peatcarbon.lu.lv/fileadmin/user_upload/lu_portal/peatcarbon.lu.lv/LIFE_PeatCarbon_1stMonitoringReport_Finland_95_.pdf","source_class":"OFFICIAL_ORGANIZATION_DATA","publication_date":"2023-12","accessed_at":"2026-08-03","claims_supported":["An active peatland-restoration program used online LoRaWAN pressure sensors in piezometers at 15-minute resolution.","The program monitored water table, peat temperature, soil moisture, greenhouse-gas variables, vegetation, and remotely sensed data before and after rewetting.","Metsähallitus was the landowner responsible for implementing the restoration plan, illustrating an identifiable operational authority."]},{"source_id":"S3","title":"CSX Peatland Carbon IoT Monitoring","publisher":"CSX Carbon","url":"https://csxcarbon.com/2026/02/12/peatland-carbon-monitoring-iot-sensors/","source_class":"COMMERCIAL_FIRST_PARTY","publication_date":"2026-02-12","accessed_at":"2026-08-03","claims_supported":["CSX conducted a six-month peatland IoT field trial using water-table and carbon-dioxide sensors.","Measurements were relayed through satellite and LoRaWAN links for near-real-time access and analysis.","CSX stated an intention to redeploy its monitoring package in 2026, showing continuing first-party interest in remote peatland telemetry."]},{"source_id":"S4","title":"Moderate drop in water table increases peatland vulnerability to post-fire regime shift","publisher":"Scientific Reports, Springer Nature","url":"https://www.nature.com/articles/srep08063","source_class":"PRIMARY_RESEARCH","publication_date":"2015-01-27","accessed_at":"2026-08-03","claims_supported":["A long-term field experiment found that moderate water-table lowering combined with wildfire changed peatland vegetation and impaired carbon accumulation.","The drained site had a substantially deeper maximum water table than the undrained site after fire.","The study supports the material ecological and carbon consequences of failing to recognize drying conditions, but does not validate this telemetry intervention."]},{"source_id":"S5","title":"Learning to Transmit: Volatility-Aware Predictive Communication for Energy-Efficient IoT Networks","publisher":"arXiv","url":"https://arxiv.org/abs/2607.19590","source_class":"PRIMARY_RESEARCH","publication_date":"2026-07-21","accessed_at":"2026-08-03","claims_supported":["A sensor-side predictor that transmits only when a normalized prediction residual exceeds a threshold is already described and evaluated.","The authors report up to 94.7% transmission reduction with quantified reconstruction error on environmental and other sensor datasets.","An online recursive-least-squares variant addresses drift, creating a close analogue to the candidate's predictor, residual transmission, reconstruction, and updating core.","The work is a recent preprint rather than evidence from a peatland field deployment."]},{"source_id":"S6","title":"A Survey About Prediction-Based Data Reduction in Wireless Sensor Networks","publisher":"ACM Computing Surveys / Universitat Pompeu Fabra repository","url":"https://repositori-api.upf.edu/api/core/bitstreams/b14043b2-dd63-485a-a69a-3db14047711f/content","source_class":"AUTHORITATIVE_SECONDARY","publication_date":"2016-11","accessed_at":"2026-08-03","claims_supported":["Prediction-based reduction of wireless-sensor transmissions is an established research area with many architectures and algorithms.","Feasibility depends on prediction accuracy, monitored phenomena, user requirements, network goals, and architecture rather than prediction alone.","The survey identifies sensor malfunction, changing dependencies, energy distribution, memory, and computation as implementation challenges."]},{"source_id":"S7","title":"LoRaWAN 1.1 Specification","publisher":"LoRa Alliance","url":"https://lora-alliance.org/wp-content/uploads/2020/11/lorawantm_specification_-v1.1.pdf","source_class":"STANDARD","publication_date":"2017-10-11","accessed_at":"2026-08-03","claims_supported":["LoRaWAN is standardized for long-range, low-power, low-data-rate applications, with stated rates from 0.3 to 50 kbps.","Data rate trades communication range against message duration, and adaptive data rate is intended to improve battery life and network capacity.","Devices must comply with region-specific duty-cycle and dwell-time restrictions and local regulation.","The standard provides device identifiers, acknowledgments, integrity checks, and cryptographic key mechanisms, but application-level model synchronization and raw-data fallback remain outside its scope."]},{"source_id":"S8","title":"MicroRX Water Level Station RX2100-WL","publisher":"Onset Computer Corporation","url":"https://www.onsetcomp.com/products/data-loggers/rx2100-wl","source_class":"OFFICIAL_PRODUCT_DOCUMENTATION","publication_date":"undated","accessed_at":"2026-08-03","claims_supported":["Commercial field hardware already supports local calculations, alarms, water-level sensing, environmental-sensor inputs, local storage, solar or battery power, and remote communications.","Listed station prices were $845 for the battery model and $929 for the solar model before required sensors, plans, installation, taxes, or engineering.","The station can log every minute, store about one million measurements, and connect as often as every ten minutes under suitable conditions.","Published battery-life figures vary materially with connection cadence and signal conditions, supporting the need to measure rather than assume an energy benefit."]}],"problem_evidence":{"support":"MODERATE","rationale":"Water-table monitoring is explicitly required or practiced in restoration programs, continuous LoRaWAN peatland telemetry visibly exists, and peatland drying can have important ecological, carbon, and fire consequences. Generic wireless-sensor literature and the LoRaWAN standard support the existence of energy, airtime, and data-rate constraints. However, no source shows that complete-vector transmission is presently a binding constraint at the proposed station, that operators are missing drying events because of it, or that the proposed multivariate stream is predictable enough to yield net savings after heartbeat, audit, synchronization, and fallback traffic.","source_ids":["S1","S2","S3","S4","S5","S6","S7","S8"]},"stakeholder_evidence":{"support":"MODERATE","rationale":"The Environment Agency is an identifiable funder requiring robust monitoring and allowing telemetry costs; LIFE PeatCarbon, Metsähallitus, and CSX are identifiable users or authorities already conducting remote peatland monitoring. This establishes credible organizations that could authorize or host a shadow trial. None expressed demand for predictive residual transmission, independent raw audits, checksum handshakes, or the complete proposed governance package.","source_ids":["S1","S2","S3"]},"prior_art":{"proximity":"SUBSTANTIAL_COLLISION","closest_analogues":[{"name":"ADAPTIVEML and ADAPTIVEML-RLS","similarity":"Maintains a lightweight sensor predictor, calculates prediction residuals, suppresses predictable observations, transmits threshold-crossing residuals, reconstructs values, and adapts under drift; this substantially overlaps the candidate's technical core.","remaining_difference":"The candidate adds peatland-specific multivariate consequence weighting, explicit model-version checks, heartbeats, random and risk-stratified raw audits, scheduled snapshots, compulsory raw fallback, safety bypasses, and human authority boundaries. The incremental value of this bundle is untested.","source_ids":["S5"]},{"name":"Prediction-based data-reduction architectures for wireless sensor networks","similarity":"A mature literature already covers dual prediction, approximate data collection, model-driven acquisition, error tolerances, and transmission suppression for environmental sensors.","remaining_difference":"The searched survey does not establish one prior system combining all of the candidate's audit independence, consequence weighting, synchronization, fallback, and non-operational peatland workflow controls.","source_ids":["S6"]},{"name":"LIFE PeatCarbon continuous LoRaWAN peatland monitoring","similarity":"Uses online LoRaWAN pressure sensors at 15-minute resolution for water-table and temperature monitoring around peatland restoration.","remaining_difference":"The report describes full monitoring and modeling, not matched sender-receiver predictors or residual-only reconstructive communication.","source_ids":["S2"]},{"name":"CSX peatland IoT monitoring package","similarity":"Uses clustered peatland sensors and satellite/LoRaWAN links for regular remote measurements and near-real-time review.","remaining_difference":"The disclosed system reports routine measurements and does not document predictive suppression, receiver reconstruction, raw audit sampling, or model-driven decompression.","source_ids":["S3"]},{"name":"Commercial remote water-level station with edge calculations and alarms","similarity":"Provides local computation, environmental sensor inputs, local storage, telemetry, and threshold alarms on deployable low-power hardware.","remaining_difference":"It does not document arbitrary matched predictive models, signed residual coding, model checksums, or the proposed independent audit and fallback logic.","source_ids":["S8"]}],"distinctive_claim_remaining":"A testable distinction remains in the integrated governance claim: compared under the same sampling cadence and link budget, a peatland-specific, reconstructive residual channel with consequence weighting, version checks, heartbeats, independent random raw audits, and mandatory raw fallback will reduce transmitted bytes and energy proxy without exceeding preregistered variable-specific reconstruction limits or reducing recall and review timeliness for consequential departures. The claim is falsified by failure of any mandatory bypass or injected fallback test, systematic structure in suppressed errors, inferiority on consequential-event recall or latency, or total operating cost not below the complete-vector comparator at required fidelity.","confidence":"HIGH"},"implementation_evidence":{"support":"MODERATE","rationale":"LoRaWAN, continuous peatland sensors, local calculations, storage, prediction-based communication, and receiver reconstruction are individually demonstrated or standardized. A shadow-only workflow avoids control-system authority and keeps raw data authoritative. Remaining gaps include edge compute and firmware compatibility, predictor synchronization under packet loss, site radio coverage and regional compliance, calibration and provenance, security configuration, audit independence, seasonal drift, safety-class definitions, false-fallback burden, data ownership, and proof that computation plus audits actually save energy. There is no peatland-specific end-to-end implementation evidence for the complete governed architecture.","source_ids":["S2","S3","S5","S6","S7","S8"]},"scores":{"meaningful_impact":{"score":4,"rationale":"Reliable, timely water-table and thermal monitoring supports consequential restoration, carbon, and fire-risk decisions, although realized ecological benefit from this communication architecture is unmeasured.","source_ids":["S1","S4"]},"stakeholder_pull":{"score":3,"rationale":"Funders and operators visibly demand robust remote peatland monitoring, but no stakeholder has requested the residual-specific solution.","source_ids":["S1","S2","S3"]},"incremental_advantage":{"score":3,"rationale":"The audit, synchronization, safety-bypass, and raw-fallback bundle could improve trustworthiness over ordinary predictive suppression, but no comparative result establishes its net advantage.","source_ids":["S5","S6"]},"distinctiveness_plausibility":{"score":2,"rationale":"Predictive residual transmission, reconstruction, adaptive thresholds, and drift adaptation substantially collide with established work; distinctiveness rests mainly on integration and governance controls.","source_ids":["S5","S6"]},"technical_implementability":{"score":4,"rationale":"Required sensing, local computation, storage, low-power telemetry, and predictive algorithms are available, while full integration and energy performance remain untested.","source_ids":["S2","S5","S7","S8"]},"adoption_authority_feasibility":{"score":4,"rationale":"Monitoring leads, landowners, grantees, and the Environment Agency can authorize a non-operational shadow trial without transferring authority over water controls or emergency response.","source_ids":["S1","S2","S3"]},"evidence_readiness":{"score":3,"rationale":"Existing continuous deployments imply that suitable raw records and stations may exist, but access, labels, link logs, energy measurements, and an adopter commitment were not verified.","source_ids":["S2","S3"]},"safety_net_benefit":{"score":4,"rationale":"Independent raw retention, audits, mandatory bypasses, and raw fallback directly address the principal risk of predictive suppression, but their detection coverage and reliability require empirical testing.","source_ids":["S4","S5","S6"]},"scalability":{"score":3,"rationale":"Prediction-based reduction and LoRaWAN can scale communications, but per-site model validation, audit storage, field maintenance, radio coverage, and human review may limit net scaling.","source_ids":["S5","S6","S7","S8"]}},"score_confidence":"MODERATE","costs":{"first_evidence":{"band_2026_usd":"10K_TO_50K","scope":"One-station walk-forward replay followed by a short, non-operational parallel shadow trial using an existing authoritative raw pipeline; includes protocol registration, model and comparator implementation, fault injection, analysis, and limited field support.","confidence":"MODERATE","assumptions":["A suitable multivariate raw record and an instrumented station are already available.","One current station costs roughly $845-$929 before sensors and services, so labor and integration dominate.","Approximately 4-10 person-weeks of engineering, ecohydrology review, and analysis are required.","No ecological-effect study or safety certification is included."],"source_ids":["S2","S5","S8"]},"initial_deployment_startup":{"band_2026_usd":"50K_TO_250K","scope":"Production-quality residual service for a small 1-5 station deployment, including edge integration, receiver software, model registry, audit storage, dashboards, security review, calibration, documentation, training, and spares.","confidence":"LOW","assumptions":["Existing sensors and primary raw-data infrastructure are reused.","Custom firmware or an auxiliary edge computer is needed because off-the-shelf support for the whole architecture is not established.","The range excludes water-control works and broader peatland-restoration construction.","Site access and communications do not require major new towers or satellite contracts."],"source_ids":["S1","S7","S8"]},"operational_launch":{"band_2026_usd":"50K_TO_250K","scope":"Validated launch across one operational peatland site or a small network, including installation, seasonal acceptance testing, incident and rollback procedures, audit governance, field commissioning, and initial support.","confidence":"LOW","assumptions":["Launch remains monitoring-only and does not control pumps, gates, or emergency systems.","The rollout is bounded to roughly 5-20 stations; a landscape-scale program could exceed this band.","Hardware costs are a minority of total integration, validation, and field-labor cost.","The Environment Agency's £100,000-£2,000,000 grant range applies to whole restoration projects and is not treated as a direct price for this subsystem."],"source_ids":["S1","S2","S8"]},"annual_recurring":{"band_2026_usd":"10K_TO_50K","scope":"Connectivity, storage, model review, calibration, audit-window review, field visits, battery or component replacement, incident response, and software maintenance for a small operational site.","confidence":"LOW","assumptions":["Approximately 0.1-0.4 FTE of combined technical and scientific oversight is required.","Routine field servicing is combined with existing monitoring visits.","No major sensor replacement campaign, satellite backhaul expansion, or regulatory study is included.","Fallback frequency remains low enough not to erase communications savings."],"source_ids":["S1","S2","S7","S8"]}},"verified_pipeline_gates":{"externally_supported_problem":{"status":"YES","reason":"Authoritative monitoring requirements, active continuous peatland telemetry, wireless-network constraints, and primary evidence of harm from drying support the general problem, although the chosen station's actual bottleneck remains unmeasured.","source_ids":["S1","S2","S3","S4","S6","S7"]},"externally_credible_adopter_or_authorizer":{"status":"YES","reason":"The Environment Agency funds and evaluates peatland monitoring, while LIFE PeatCarbon partners, Metsähallitus, and CSX operate relevant field programs capable of hosting or authorizing a shadow study.","source_ids":["S1","S2","S3"]},"distinct_testable_incremental_claim":{"status":"YES","reason":"The remaining claim is explicitly contrastive against complete-vector reporting and existing predictive suppression, with measurable reconstruction, recall, latency, traffic, energy, audit, and fallback outcomes.","source_ids":["S5","S6"]},"bounded_next_evidence_step":{"status":"YES","reason":"A one-station preregistered replay plus short parallel shadow trial is bounded in site, authority, data, duration, comparators, measures, and fault-injection falsifiers.","source_ids":["S2","S5","S8"]},"no_unresolved_safety_or_authority_stop":{"status":"YES","reason":"The next step leaves the raw pipeline authoritative, takes no operational or emergency action, and can halt on any bypass, reconstruction, synchronization, or missingness failure. Safety thresholds themselves remain unvalidated and cannot support operational use.","source_ids":["S1","S4"]},"credible_cost_scope_and_range":{"status":"YES","reason":"Current first-party station prices and an official funding framework anchor order-of-magnitude bands, while scope and labor assumptions are explicit. Confidence remains low outside the first experiment because site integration quotes are absent.","source_ids":["S1","S8"]}},"next_evidence_step":"With a willing monitoring operator, preregister a walk-forward replay on one station's authoritative raw record spanning wet, dry, rainfall-transition, maintenance, and sensor-fault conditions. Compare (A) complete-vector transmission at the existing interval, (B) fixed change/threshold reporting, (C) a published-style predictive residual method, and (D) the proposed governed residual system under the same simulated link and review budget. Freeze models and thresholds before an untouched test segment. Measure bytes and airtime, device-energy proxy, reconstruction error and suppressed-error mass by variable and regime, consequential-event recall and time to reviewer acknowledgment, raw-audit disagreements, model-maintenance effort, and total cost. Inject packet loss, reordering, checksum mismatch, stale model, missing heartbeat, sensor displacement, safety-class records, and fallback oscillation. Proceed to a short parallel field shadow only if the replay meets preregistered noninferiority limits and every compulsory fallback works; keep raw reporting authoritative and prohibit operational actions. Falsify the claim if any mandatory safety record is suppressed, any injected incompatibility is accepted, residuals retain consequential systematic structure, event recall or latency is inferior, reconstruction exceeds its variable-specific budget, or total residual-system cost is not below the complete-vector comparator.","blocking_evidence":["No direct measurement shows that communications, energy, storage, or analyst attention is binding at the proposed station under complete-vector reporting.","No stakeholder has expressed demand for this specific residual, audit, checksum, and fallback architecture.","No peatland-specific dataset has demonstrated predictability, reconstruction fidelity, rare-event recall, or net energy savings for the proposed multivariate vector.","No field test has shown that heartbeat, checksum, raw audit, and fallback mechanisms work under real packet loss, weather, sensor faults, seasonal drift, and weak connectivity.","Consequence classes and fire-safety bypass thresholds have not been approved or validated by the relevant authority.","Access to an authoritative raw record, event labels, device-energy measurements, and operator disposition data is unconfirmed.","Current site-specific integration, staffing, connectivity, calibration, and maintenance quotes are absent.","Patentability, freedom to operate, market size, world novelty, and realized ecological impact were not assessed."],"research_disposition":"PARTNERED_RESEARCH_PROGRAM","world_novelty_boundary":"The search establishes substantial prior art for prediction-based sensor transmission, residual thresholding, reconstruction, adaptive learning, LoRaWAN environmental monitoring, and peatland IoT deployments. It did not establish a direct source implementing the candidate's entire peatland-specific combination of consequence weighting, model checksums, authenticated heartbeats, independent random raw audits, compulsory decompression, safety bypasses, and bounded human authority. Absence from these eight sources is not evidence of world novelty. World novelty, patentability, freedom to operate, market size, and realized impact remain unmeasured.","arm":"COMPLETE_PROPOSAL_PORTFOLIO","candidate_version":0,"controller_recommendation":{"action":"STOP_EMPIRICAL_RESEARCH_NEEDED","repairable":false,"material_progress_observed":true,"progress_targets":["Secure a named peatland operator and written authority for access to one station's raw data and a monitoring-only shadow trial.","Obtain a raw multivariate record spanning representative regimes, with calibration, maintenance, communications, energy, and event-review metadata.","Preregister variable-specific reconstruction limits, consequential-event definitions, noninferiority margins, traffic and energy targets, comparator implementations, and stopping rules.","Demonstrate perfect invocation of mandatory bypass and raw fallback in all scripted packet-loss, checksum, staleness, heartbeat, sensor-fault, and safety-class tests.","Show lower total transmitted bytes, airtime, energy proxy, and operating cost than complete-vector and fixed-threshold comparators without inferior event recall or review latency.","Obtain independent review of raw-audit sampling, consequence weights, fire-safety boundaries, data governance, radio compliance, and rollback procedures.","Replace broad cost assumptions with site-specific engineering, field-service, connectivity, and recurring-support quotes."],"reason":"Bounded web research verifies the general monitoring need, credible authorities, substantial prior art, implementable components, and a falsifiable incremental governance claim. It cannot establish the site-specific bottleneck, predictive performance, net energy or cost advantage, rare-event safety, or fallback reliability. Those questions require proprietary raw records, operator judgments, fault-injection experiments, and a live shadow trial; therefore the next decision requires empirical research rather than further general web search."},"proposal_index":1}