Summary

Description

This dataset classifies the presence or absence of forest cover for the year 2020, with global coverage at 10 m spatial resolution.

Usage notes

This dataset is intended to support EUDR deforestation risk assessments (i.e. initial state on 31 December 2020). For EUDR forest degradation risk assessments, see JRC Global Forest Type dataset. The dataset is non mandatory, non exclusive, and non legally binding for EUDR compliance. A list of known issues is published here.

Licence

Open

Version

2.0

Release date

March 2025

CRS

EPSG:4326

Update schedule

Provider cost

$0.00/ha

Variables

Forest cover

Description

Forest cover indicates whether a pixel is forest (1) or non-forest (0). The data is compiled from > 29 global datasets, filtered to pixels matching the EUDR forest definition (canopy height > 5 m, no agricultural use, no other land use, no urban use, minimum mapping unit 0.5 ha), and corrected for known spatial artefacts (e.g. Landsat 7 scan lines, lava flows).

Usage notes

Reported accuracy is 91%. Dataset classifies forest (not tree) cover, so includes forest land without trees (e.g. logged areas) and excludes non-forest land with trees (e.g. urban trees). Dataset may misclassify agricultural tree crops, agroforestry, and urban forest as forest. Dataset may misclassify shifting forest cultivation and temporarily unstocked forest as non-forest. Dataset struggles to identify (non-)forest with low canopy cover and canopy height. Variable spatial resolution of input datasets can yield artificially sharp boundaries at forest edges.

Variable category

Forest cover

Variable type

Categorical

Data access

Data structure