Methods.

Canonical run

• Release: canonical_v6
• Study area: England, British National Grid
• Analysis unit: 1 km hexagonal cells
• Cells scored: 204,703
• Published score layer: data/interim/mvp_official_boundary_1km_v6/hex_scores.parquet

Scoring workflow

1. Standardise source layers into a common coordinate reference system.
2. Build a national 1 km analysis grid.
3. Aggregate habitat, biodiversity observation, agricultural, flood, and peat signals to each cell.
4. Transform each component so higher values mean stronger apparent restoration opportunity.
5. Apply an undersized-cell penalty so clipped coastal or boundary fragments do not dominate.
6. Generate nature-first, balanced, and lower-conflict scenario scores.
7. Export shortlists, candidate clusters, validation tables, and the interactive explorer.

Why these data were chosen

The data choices are a compromise between ecological relevance, national coverage, reproducibility, and practicality. The aim was not to build a perfect ecological model. It was to build a national screening workflow using signals people can inspect and argue with.

• Priority habitat and habitat proximity were chosen because rewilding and landscape recovery usually make more sense where existing habitat networks can be buffered, expanded, or joined up, rather than treated as isolated fragments.
• Bird and mammal observations were included as pragmatic biodiversity indicators with national availability. They are imperfect and effort-sensitive, which is why the project explicitly dampens sparse-record cells instead of treating non-recording as ecological absence.
• Agricultural Land Classification was used as a simplified tradeoff layer because restoration decisions in England are shaped by agricultural opportunity cost and conflict, even though ALC is only a proxy for real farm-level feasibility.
• Flood opportunity was included because wetland and floodplain restoration are central to many restoration strategies and because hydrological context is an important part of landscape-scale opportunity.
• Peat opportunity was included because peat systems matter for carbon, hydrology, and habitat recovery, and they often justify restoration attention even when other signals are mixed.
• Scenario scoring was used because there is no single agreed definition of the "best" rewilding landscape. Different users will care about different things.

Limitations

The model does not predict ecological outcomes, prove deliverability, assess land ownership, replace local ecological survey, or model community consent. Observation-based biodiversity signals are still sensitive to recorder effort. Agricultural opportunity is a simplified proxy, not a farm business assessment.

Literature and policy frame

• Perino et al. 2019 frame rewilding around restoring ecological processes and considering social-ecological constraints.
• Spatial conservation prioritisation provides the wider decision-support tradition for comparing conservation and restoration options spatially.
• England's Nature Recovery Network gives the policy context for creating, restoring, connecting, and expanding wildlife-rich places.
• Local Nature Recovery Strategies provide the local spatial planning context this kind of screening could support.
• Making Space for Nature remains the key England-specific reference for the case that ecological networks should be bigger, better, and more joined.
• Peatland restoration and downstream stormflow is a useful example of why peat was treated as more than a biodiversity-only variable in the scoring design.
• The 2025 IPBES Transformative Change summary provides the wider global framing for why nature recovery now has to be treated as systemic and urgent.