England Tree Species Map Using Sentinel-2 Satellite Data

35 common tree species in England were classified using Sentinel-2 satellite data and an XGBoost machine learning model. The Forestry Commission produced this map, which achieved an overall classification accuracy of 89%. The dataset was last updated in April 2026.

Use Cases

• Analyze the predicted distribution of the 35 classified tree species for conservation planning.
• Utilize the confidence raster output to assess prediction reliability for specific geographic areas.
• Validate model performance for specific classes, such as Norway spruce, using the reported 69% precision metric.
• Investigate classification accuracy differences between well-represented species and underrepresented species grouped as 'Other broadleaf' or 'Other conifer'.

Strengths

• Model evaluation demonstrated strong performance with an overall accuracy of 89% and balanced class accuracy of 90%.
• The classifier was trained to distinguish 35 common tree species using a time-series of Sentinel-2 surface reflectance data.
• Field validation for Norway spruce within a specific demarcated area confirmed a precision of 69%, aligning with test data.

Limitations

• Minority species with fewer than 1,000 training samples were grouped into broader categories, reducing granularity for those species.
• Model accuracy was observed to be worse for underrepresented species, indicating potential class imbalance issues.
• The dataset's spatial coverage is limited to England, limiting applicability for broader regional studies.

Provenance

Source

Forestry Commission, funded by DEFRA’s Natural Capital and Ecosystem Assessment (NCEA) programme.

Collection Method

Created using Sentinel-2 satellite remote sensing data and an XGBoost machine learning model trained on labelled samples from the sub-compartment database (SCDB).

Freshness

Last updated in April 2026.

Geography

England