Geospatial Data for TNFD: Coordinates, Maps, Data Quality and Confidentiality

Educational practitioner guidance. Not legal, scientific-validation or assurance advice. Apply TNFD to the organisation's facts, stated materiality approach, ecological context and reporting basis.

Location analysis supports finding interfaces with nature, applying sensitive-location criteria, selecting assessment locations and disclosing priority locations. If the team saves only a screenshot, it loses the source geometry, query, data versions and rationale that made the result reproducible.

Separate the source record, cleaned geometry, analysis layer, reviewed conclusion and public output. This allows a coordinate or dataset to be corrected without rewriting the whole assessment.

Record the original address, standardised address, geocoder/service date, returned coordinates, match type and confidence. A postcode, street or locality match must not be relabelled as a surveyed coordinate simply because it has many decimal places.

Data can arrive in different coordinate reference systems and formats. Identify the input CRS, transform to the analysis CRS, preserve original geometry and test for swapped latitude/longitude, invalid polygons, duplicates, implausible locations and records in the wrong country or sea.

Raw layer — immutable source coordinates or geometry plus source metadata.

Standardised layer — cleaned geometry in the selected CRS with documented transformations.

Analysis layer — buffers, basins, intersections and classifications generated by a versioned query.

Release layer — reviewed and possibly aggregated geometry approved for disclosure.

Geospatial lineage connects source locations to geocoding, standardisation, spatial joins, review and controlled output. Metadata prevents proxies from being mistaken for surveyed precision.

A spatial join includes geometry, dataset/version, CRS, topological relationship, distance or intersection rule, edge-case treatment, output fields and review. “Within five kilometres” is different from “shares a catchment” or “route intersects a corridor”. Save the script, GIS model or documented manual workflow.

Record a downloadable edition, API version, endpoint, extraction date or archived result where permitted. Separate ecological observation date, dataset publication/version and export date. Changes in a layer should trigger review of affected locations rather than silently rewriting prior conclusions.

Vessels, projects, construction teams and fleets interact across routes and time windows. A registered-office coordinate is rarely meaningful. Use route segments, operating zones, ports, licence areas or time-stamped events and link them to the assessment period.

Define the relevant temporal window and season.

Separate planned, permitted and actual routes or areas.

Use aggregated tracks where raw telemetry is restricted, retaining the internal aggregation logic.

Identify hotspots and short-duration exposure rather than averaging them away.

Map release requires technical, ecological, security/rights and publication QA. Internal precise and public layers can differ, but confidentiality is not a blanket reason to remove location information.

Use role-based access for precise coordinates, raw supplier data and sensitive ecological information. Public aggregation may use basin, region, grid or cluster, but it must not group materially different locations or make a key issue disappear. State the level of specificity and rationale.

1. Define the data model. Set stable IDs, geometry classes, precision grades, confidence and access fields.

2. Ingest without overwriting. Preserve addresses, surveys, supplier declarations, polygons and routes.

3. Geocode and grade. Record service, match level, date and confidence.

4. Standardise and validate. Transform CRS, repair geometry, deduplicate and test plausibility.

5. Link to business data. Connect every geometry to activity, product, commodity, supplier and period.

6. Version nature datasets. Record publisher, edition, resolution, licence and extraction date.

7. Run reproducible joins. Document relationship, threshold, query and exceptions.

8. Perform technical and ecological QA. Sample edge cases and challenge false precision.

9. Approve access and public granularity. Separate precise internal and public layers.

10. Archive and trigger updates. Retain inputs, outputs, scripts and decisions.

The register keeps each evidence class visible. Verified polygons support local screening; declared districts support subnational screening; country proxies support portfolio prioritisation only; unknown volume stays in the improvement queue. The public map reports sourcing regions and confidence categories while the internal layer retains permitted precision. The report does not claim full traceability.