Provenance-backed MRV for biodiversity credits and TNFD reporting
The technical infrastructure that makes biodiversity data creditable: SHA-256 content provenance, two-person review, Darwin Core standardisation, DwC-A export, taxonomic confidence modelling, and an integrity-hashed audit trail.
Capabilities
Every capability is built, deployed to production, and verified against the live database. What is listed here is not a roadmap — it is the current state of the platform.
SHA-256 content provenance
Every photograph is SHA-256 hashed at the moment of capture. The hash is stored in the Darwin Core multimedia record and verifiable by any third party — auditor, verifier, or credit buyer.
Hash stored in mrv_record_photos.sha256, mapped to DwC-A Audubon Core multimedia extension
Two-person review with no self-approval
A different person reviews and approves each record. The constraint is enforced at the database level — a user cannot approve their own submission. Reviewer identity and timestamp are preserved on every record.
CHECK constraint: reviewed_by IS DISTINCT FROM created_by
Darwin Core standardisation
Records are mapped to Darwin Core terms: event, occurrence, and multimedia. Geodetic datum is WGS84 (EPSG:4326). basisOfRecord is always HumanObservation. GBIF taxon keys are resolved where available.
Biodiversity Information Standards (TDWG) Darwin Core
Darwin Core Archive (DwC-A) export
Approved and locked records export as a standards-compliant DwC-A ZIP: event.txt, occurrence.txt, multimedia.txt, meta.xml, and eml.xml with CC-BY 4.0 license. Live and deployed to production.
GET /api/mrv/sites/{siteId}/dwc-archive — reviewer/admin role required
Immutable locked records
Once approved, a record is locked. It cannot be edited. Taxonomic identifications are versioned — new identifications supersede previous ones via a database trigger, preserving the full identification history.
is_current flag + supersede trigger on mrv_taxon_identifications
Taxonomic confidence grades
Every species identification carries a confidence grade: ai_suggested, observer, expert_verified, or disputed. This lets auditors weight evidence appropriately and lets credit verifiers distinguish AI suggestions from expert confirmation.
identification_grade enum on mrv_taxon_identifications
Sensitive-species coordinate generalisation
Records flagged as sensitive species have their coordinates generalised to a 1km grid in the DwC-A export. The generalisation is documented in the record metadata, not silently applied.
is_sensitive flag + 1km grid generalisation in DwC-A projection layer
GPS accuracy mapped to coordinateUncertaintyInMeters
GPS accuracy is mapped verbatim to Darwin Core coordinateUncertaintyInMeters — not inflated, not estimated. The raw accuracy_m value is preserved alongside the Darwin Core term.
accuracy_m mapped to dwc:coordinateUncertaintyInMeters, verbatim
Versioned identifications with supersede trigger
When a new taxonomic identification is recorded for a site, the previous identification is automatically marked as superseded (is_current = false). The full identification history is preserved for audit.
Supersede trigger on mrv_taxon_identifications, current-per-record index
Audit trail with integrity hashing
Every state change on a record is logged in an audit table with an integrity hash chain. Each audit entry links to the previous entry via prev_integrity_hash, creating a tamper-evident trail.
mrv_audit_entries with integrity_hash + prev_integrity_hash chain
Standards and frameworks
The platform aligns with international biodiversity data standards and the regulatory frameworks driving the nature finance market.
Taxonomic identifications and sensitive-species records are QA fixtures inserted for a stakeholder demo — schema and local-harness validation confirmed
Full live end-to-end API run (DwC-A ZIP fetch + protocol 400 guardrails) pending reviewer session
Want to see the DwC-A export on your data? A fifteen minute call costs nothing. Book a call.