This dataset maps the dominant driver of tree cover loss globally from 2001 to 2025 globally. Produced by the World Resources Institute (WRI) and Google Deepmind. Using satellite imagery and a neural network to classify seven categories of drivers.
Source & citation: Zenodo record 14163025; also on the WRI Data Explorer.
This publication is a mirror of Global drivers of forest loss at 1 km resolution (Sims et al. 2025), version 1.3, coverage 2001 to 2025. We added a virtual icechunk store that mirrors the .tif file.
drivers_forest_loss_1km_2001_2025_v1_3.tif.
Version DOI https://doi.org/10.5281/zenodo.19485190.
Concept DOI https://doi.org/10.5281/zenodo.14162799.CC-BY 4.0. Attribution is required. Cite the paper and the Zenodo record. The mirror does not imply endorsement by WRI or Google DeepMind.
Nothing. The original file is a COG (Cloud-optimized GeoTIFF). We mirror it unchanged.
The data has 8 bands. Band 1 gives the classification. Bands 2 to 8 give the quantized pseudo-probabilities.
Classes:
Example of an anonymous read of the COG through the source.coop proxy:
Example of a GDAL-free read with icechunk and xarray. Import virtual_tiff first. This import registers the codec that the chunks use: