Data for: Geospatial foundation models enable data-efficient tree species mapping in temperate mountain forests

This dataset accompanies the paper "Geospatial foundation models enable data-efficient tree species mapping in temperate mountain forests" published in Remote Sensing of Environment. It contains pre-distilled training datasets (NumPy archives), preprocessed forestry parcel reference data (GeoPackage), a wall-to-wall 10 m species map, experiment result files, trained model weights, and class registries for 18 tree species across the Autonomous Province of Trento (Italian Alps). See the included README.md for full documentation. Source code: https://github.com/PatBall1/trentino-trees