DATA · ARCHITECTURE

One schema.
End-to-end traceability.

Ecological signals travel auditable causal paths across five defined layers, arriving as decision-ready evidence with the full reasoning chain intact: interrogatable, validatable, and ready to act on.

05
Strata
25
Specialist engines
25
Ecological modules
Causal pathways
Unified schema

Five strata. One coherent schema.

The architecture is built on a unified schema spanning five cascading ecological layers — from soil subsurface to space — representing the continuous trace of energy flow, conversion, and utilisation.

Each stratum is independently queryable: interrogated on its own terms, in isolation, with the layers above and below undisturbed. Yet, all five are simultaneously cross-referenceable within the same schema. This means that a signal originating in the soil microbiome can be traced through its pathways via plants, biome dynamics, and climate modifiers up to satellite-derived land-cover classifications, without losing its structural identity at any stage.

Strata separation is preserved throughout. That is the architectural distinction that sets Trophic Bio apart from conventional ecological data systems.

05
SpaceEO · Satellites · Space
queryable
04
ClimateAtmosphere · Weather · Solar Flux
queryable
03
BiomesEnergy Flux · Environment · Ecosystems
queryable
02
Plant LifeEcology · Flora · Productivity
queryable
01
EarthMicrobiome Productivity · Soil Health · Subsurface
queryable

All five strata cross-referenceable within a single unified schema. Structural identity preserved at every layer.

Origin & investment

Born in Green Tundra. Built into Trophic Bio.

The unified-schema architecture began inside Green Tundra, our non-profit research arm — as a way to reason coherently across ecological strata that conventional tooling kept fragmented. From the first prototypes, the design concept proved exceptional at preserving causal traceability where other systems collapsed it into summary statistics.

We immediately recognised its value as a market-defining differentiator for ecological and environmental intelligence. To carry that promise into institutional, regulatory, and commercial contexts, we invested in productionising the architecture through Trophic Bio — hardening the schema, formalising the engine-module mirror, and engineering the quality-control spine the framework now stands on.

The lineage matters: the design concept was earned through fieldwork-aligned research, then re-engineered to meet the standards partners can build on. Green Tundra remains the upstream environment where new ideas are stress-tested; Trophic Bio is where they become dependable infrastructure.

Engine × module architecture

25 engines. 25 modules.
One mirror design.

Each of the 25 specialist engines has a dedicated module counterpart. Every ecological domain resolved in the physical environment has a corresponding module that resolves its population-, environment and ecological-level consequences. The structure is a design principle that makes up the framework of Trophic Bio.

Modules are not simple receivers. Each module is defined by the collation of its validated structural dependencies — the specific inputs confirmed to belong to that module under the unified schema. These dependencies are what give each module its shape before any authority check occurs.

The architecture runs in two modes. In the Green Tundra instance, a single shared authority and engine orchestrates all 25 modules; we utilise this for staging and first-level logic queries. In the second instance, and the core feature of the commercial prodct, each module is resolved by its own dedicated engine, which then passes through a final authority check before the output is confirmed. This two-mode architecture is what makes traceability possible at scale, where every resolved state is auditable back through its dependency structure, regardless of which instance is running.

Outputs are not fixed. What the mirror produces depends on the research question being asked and which modules are active in a given context. The architecture defines that path.

Reading the architecture
Authority engineHolds the logic for resolving an environmental condition. In the research instance, one authority engine orchestrates all modules. In the commercial instance, each module has its own engine which passes through a final authority check before resolving.
ModuleThe processing layer paired to each engine. A module is defined as the collation of its validated structural dependencies — inputs confirmed under the unified schema to belong to that module component.
DependenciesThe inputs that structurally define each module. These are not outputs — they are what the module is composed of. Cross-domain dependencies are where causal complexity and traceability emerge.
Knowledge enrichmentThe canonical rules and ecological relationships each module reasons within. Drawn from established science, known environmental and ecological equations, and in-house empirical research, this represents the closest current understanding of ecological truth. An alignment score governs each enrichment source, not a measure of the quality of the underlying work, but an empirical measure of how much can be confidently extracted and operationalised from it. This layer evolves continuously as scientific understanding advances.

The table below is a brief example of the data architecture and does not represent, in its current form, the complete framework design.

Atmospheric
  • Atmosphere couplingTemperature & radiation envelope
    solar anglealbedo indexsurface emissivitycloud cover fractionSeasonal overlays
  • SAR surfaceTerrain & surface habitat
    surface roughnesscanopy heightinundation extentsoil moisture proxy
  • Multispectral indexPrimary productivity & Photosensing
    NDVIEVINPQphotoreceptorschlorophyll reflectanceland cover classification
  • Solar irradianceLight availability & photosynthesis
    photoperiodshade durationwind dynamicscarbon assimilationNPQPAR fluxSolar angle
  • EO change detectionSeasonal state modifier & geographic area
    phenological phasedisturbance signalburn scar indexvegetation compensator
Meteorological
  • Precipitation forecastWater availability response
    rainfall intensitytemporal distributionsnow fractionantecedent moisture
  • Heat flux dynamicsHumidity & thermal envelope
    latent heat fluxsensible heat fluxvapour pressure deficitdew point
  • Drought indexChronic stress accumulation
    SPIPDSIWUEsoil water deficitevapotranspiration anomalyBiomass index
  • Convective patternPrecipitation structure & intensity
    available potential energystorm cell trackingprecipitation typefrontal passageeco-habitat modifiers
  • Anomaly detectionClimate departure signal
    baseline deviationextreme event frequencytrend breakpointarctic tundra indexseasonal phase shift
Hydrological
  • Watershed modellingCatchment resource routing
    catchment geometrysoil permeabilityprecipitation inputland cover
  • Flood riskHabitat inundation & connectivity
    return periodfloodplain extentflow velocitysoil saturation
  • Water quality indexTrophic & food web condition
    turbiditydissolved oxygennutrient loadingpHsediment flux
  • Terrain changeLandscape disturbance signal
    SARerosion rateslope instabilitychannel migrationmass movement indexSoil classification
  • Extreme weather impactPopulation shock response
    event magnitudedurationrecovery laghabitat exposureantecedent condition
Ecological systems
  • Biodiversity scoringSpecies interaction network & abiotic factors
    biodiversity indexfunctional diversityinteraction strengthconservation strategy
  • Population dynamicsGreen Tundra Population Matrix
    age structuresurvival ratesfecunditydensity dependenceimmigration flux
  • Vegetation structureFlora mechanics & habitat layer
    canopy coverunderstory densityphenological stagebiomass index
  • Land-use attributionCarrying capacity modifier
    land cover typefragmentation indexedge densitydisturbance regimehuman footprint
  • Connectivity networkDispersal & migration routing
    patch permeabilitycorridor widthbarrier indexmovement cost surface
Soil & substrate
  • Soil health scoringNutrient cycle & substrate condition
    organic matterbulk densityTOCsoil conductivitymicrobial activitypH
  • Carbon sequestrationCarbon flux & sink dynamics
    SOC stockNPP allocationdecomposition rateroot turnoverlitter input
  • Microbiome indexDisease pressure & pathogen load
    microbial diversitypathogen prevalencehost densityenvironmental persistence
  • Nutrient cycleOrganic composition & mineralisation
    nitrogen mineralisationphosphorus availabilitystoichiometric ratiosleaching rateIon typefertiliser index
  • Substrate availabilityHabitat & species establishment
    particle size distributionmoisture retentioncompaction indexnesting suitability
Causal pathways

Not 25! It's contextually infinite.

The number of cross-layer causal pathways between modules and engines is not fixed by the architecture. It is shaped by the ecological question being asked. A species population projection activates a different chain of engine-module interactions than a carbon flux query or a habitat connectivity assessment.

Concept 01

Cross-domain activation

A single ecological question may activate engine-module pairs across multiple domains simultaneously. A drought event triggers atmospheric engines, climate modules, vegetation structure, population stress, and soil nutrient cycle — all within the same trace. The pathway is the answer.

Concept 02

Historical + current resolution

The schema carries both historical baselines and current signal simultaneously. A pathway can be run against past data to validate model behaviour, or against current inputs to produce live projections — without changing the architecture. Time is a dimension, not a constraint.

Concept 03

Per-species, per-region, per-tick evaluation

Population dynamics modules are evaluated per species, per region, per simulation time step. The same engine output may resolve differently for a migratory bird versus a sedentary invertebrate in the same landscape. Resolution is not averaged — it is preserved.

Concept 04

Auditable evidence chains

Every output carries the chain of engine-module activations that produced it. For institutional partners, regulators, or research consortia, this means the reasoning behind a projection can be interrogated, challenged, and matched against independent data. No black-box inference.

Quality control

Compliant. Scalable. Reproducible.

The architecture is built to satisfy strict institutional requirements for ecological intelligence: compliance, scalability, consistent reproducibility, and FAIR alignment (Findable, Accessible, Interoperable, Reusable). These are not properties added on top — they are designed into the five mechanisms below.

Together, they form the quality-control spine that allows partners — research institutions, regulators, governments, and consortia — to depend on outputs they can interrogate, reproduce, and defend.

01
A shared data model
All five strata, all engines, and all modules conform to one canonical structure. No bespoke schemas per project, no parallel data dialects — every signal lives in a form every other component already understands.
02
A controlled vocabulary / ontology
Terms, units, taxa, and ecological concepts are defined once and referenced everywhere. Disambiguation is structural, not editorial — a species, a soil class, or an indicator means the same thing across every query.
03
An ingestion and validation layer
Every input — field signal, partner dataset, EO product — passes through schema, vocabulary, and quality validation before it enters the architecture. Non-conforming data is rejected with structured reasons, not silently absorbed.
04
Versioning
Schema, vocabulary, engines, and modules are independently versioned. Any output can be reproduced against the exact configuration that produced it — months or years later — supporting audit, peer review, and regulatory scrutiny.
05
A staging query interface that respects strata boundaries
Queries are issued against a staging interface that preserves stratum identity at every step. Cross-strata joins are explicit and traced — never implicit, never flattened — keeping every result FAIR-aligned and inspectable.
Schema logic & narrative logic

Two modes of reasoning.
One architecture.

The architecture operates simultaneously as a technical data schema and as a reasoning environment open to narrative framing. These are not in tension — they are the two registers through which the same signal chain can be read.

Schema logic

Technical strata resolution

At the schema level, the architecture is a deterministic system. Environmental inputs are resolved by specialist engines, passed through population modules, and evaluated per species, per region, per time step. Every output is reproducible.

Historical data and current signal share the same schema structure. Queries can span time without structural changes to the architecture — the schema carries the evidence, the query determines the resolution.

  • Deterministic engine-module evaluation
  • Per-species, per-region, per-tick resolution
  • Historical and current data in shared schema
  • Auditable evidence chains on every output
  • Independent strata queryability preserved
Narrative logic

Policy, research & programme framing

At the narrative level, the same architecture accepts big-picture objectives — a climate policy target, a conservation programme outcome, a consortium research question — and traces the signal chains that are relevant to that framing.

This is what makes the framework useful to institutions and governments, not just data scientists. The question can be framed in policy language; the architecture resolves it in evidence.

  • Climate policy objective alignment
  • Research programme signal mapping
  • Consortium evidence packaging
  • Regulatory and advisory output framing
  • Citizen science integration pathways

The architectural distinction is this: schema logic determines what can be traced; narrative logic determines what is worth tracing. The former is fixed by the engineering. The latter is open — to the research question, the policy objective, the consortium brief, the field observation that prompted the inquiry. Trophic Bio is the framework in which both registers operate simultaneously, without one compromising the other.