Workflow and end products. The basic data consists of Sentinel-2 images and LiDAR data, with two additional phenology-based bands in the year-map workflow. Ground truth is used for training and testing sets for the Random Forest classifier.