Large-scale individual tree species mapping using AVIRIS-NG and LiDAR data fusion

Overview This dataset provides a continuous, high-resolution tree species map for the Gruyère Pays-d’Enhaut Regional Nature Park (Swiss Pre-Alps), generated using AVIRIS-NG hyperspectral imagery, airborne LiDAR data, and a Random Forest classifier. The map covers five dominant species (Picea abies, Fagus sylvatica, Abies alba, Acer pseudoplatanus, Fraxinus excelsior).To address gaps caused by incomplete hyperspectral coverage, we applied chessQS (Gerber et al., 2025), a tile-based Multiple-Point Statistics gap-filling algorithm, ensuring spatial continuity while preserving realistic species distribution patterns. Data description This dataset contains the two final maps (gap-filled) derived from hyperspectral-only (treemap_hsi.tif) and hyperspectral-Lidar fusion (treemap_hsi_lidar.tif) models. Each map has a resolution of 2m. A detailed description of the methodology used to generate these products can be found in the associated publication. We gratefully acknowledge financial support through the Action Plan of the Swiss Biodiversity Strategy by the Federal Office for the Environment (FOEN) for financing the ValPar.CH project. All the institutions and contributors who made the input data freely available are gratefully acknowledged. Reference Lambiel, A., Gerber, L., Schweiger, A.-K., Kneubühler, M., Mariethoz, G., Lehmann, A., & Külling, N. (2026). Large-scale individual tree species mapping using AVIRIS-NG and LiDAR data fusion (preprint). Zenodo. https://doi.org/10.5281/zenodo.18803977