Training Data

We include here two sets of training data: 1) shapefiles including polygons classified by land cover, and 2) csv files including extracted spatial data for each pixel within the polygons from 1). We used this training data to create the land cover change maps included in this FigShare collection. We include training data for our Bolivia, Brazil, and Peru case studies (see the metadata files). Land cover classes include forest, agriculture, bare soil, urban, water, desert, and wetland. These data correspond to Kinnebrew et al. 2022 (PLOS ONE) DOI: 10.1371/journal.pone.0268970