DHI - Tree cover map (2020-2022)

The DHI – Tree cover map (2020–2022) provides a spatial assessment of the probability that an area is covered by trees. The primary purpose of the dataset is to represent tree presence/ absence consistently across both forested and non‑forested landscapes at Danish national scale. The dataset provides a continuous spatio-temporal probability measure based on random forest machine‑learning models that estimate tree probability at pixel level.

The model is based on multi‑temporal Copernicus Sentinel-1, Sentinel‑2[1] satellite imagery and national digital elevation data[2] covering the period 2020–2022, and uses training data derived from the Danish National Forest Inventory (NFI)[3].

The spatial resolution of the dataset is 10 meters, which provides an unprecidented level of detail and accuracy mapping of dynamic tree-populated and forested areas. 

The output expresses the likelihood of tree presence rather than a binary classification, allowing forest dynamics and transition zones to be captured. The dataset supports applications such as forest monitoring, land‑use screening, environmental assessment, and baseline mapping. This is of particular interest when monitoring the development of young forests, either planted or naturally regenerated, but when monitoring deforestation events, either through management and storm felling. Due to its national coverage, the dataset is well suited for large‑scale analyses and as input to downstream ecological spatial modelling.

The work is based on a research grand from INNO-CCUS, the Danish government’s CCUS (carbon capture, utilisation and storage) research partnership, supported by Innovation Fond Denmark.

Governance around use of the dataset refers to Creative Commons BY-NC-ND 4.0 (Deed - Attribution-NonCommercial-NoDerivatives 4.0 International - Creative Commons), which allows reusers to copy and distribute the material in any medium or format in unadapted form and for noncommercial purposes only. Credit must be given to DHI. Only noncommercial use of dataset is permitted. Noncommercial means not primarily intended for or directed towards commercial advantage or monetary compensation. No derivatives or adaptations of the dataset is permitted. 

Access to the full dataset can be obtained by emailing Lotte Nyborg, ln@dhigroup.com or Mikkel Høegh Bojesen, mihb@dhigroup.com.