Wet Woodland Extent Mapping in England

Overview

Wet woodland is one of the UK's most threatened and poorly mapped habitat types. This project produced the first national-scale, high-resolution map of wet woodland extent in England — a critical baseline for biodiversity monitoring, carbon accounting, and nature recovery planning under the Local Nature Recovery Strategy framework.

Delivered for Defra UK, the outputs include extent mapping at 10m resolution, density analysis at 5km hexagonal grid cells, and restoration potential scoring across agricultural land grades.

Methodology

Classification was driven by a combination of LiDAR-derived structural features and satellite embeddings. LiDAR data provided canopy height, terrain wetness indices, and vegetation structure metrics. Satellite imagery embeddings captured spectral and temporal signatures associated with wet woodland conditions across seasons.

These feature sets were fused and used to train a classification model capable of distinguishing wet woodland from adjacent dry woodland, scrub, and riparian vegetation — classes that are spectrally similar but structurally distinct.

Outputs were validated against field survey data and existing habitat inventories, and delivered as Cloud Optimized GeoTIFFs (COG) for efficient web streaming.

Interactive demo

The live demo provides four visualisation modes: extent raster, density hexgrid, LNRS regional breakdowns, and restoration potential filtering by agricultural land grade.