Tropical Dry Forest Classification in Northern Colombia

Dataset for the paper: Configuration Patterns of Tropical Dry Forest in Northern Colombia: A 10m Resolution Map for Highly Fragmented Ecosystem

The tropical dry forest (TDF) is considered one of the world's major biomes, but it is also recognized as the most threatened ecosystem with less than 10% of its original extent remaining globally. In Colombia, only about 8% is preserved; however, in some regions, TDF estimates are imprecise or have large information gaps. Here, we present a detailed map of the TDF in the Department of Córdoba, northern Colombia. The map was developed using four supervised classification algorithms based on Sentinel-2 imagery. To enhance prediction quality, we applied TDF_algorithm classification in R, a multi-model ensemble approach that integrates precision and uncertainty from the four algorithms for improved land cover classification. The code is available at GitHub TDF_algorithm-classification

TDF_algorithm classification is a collaborative application of the SC-R: a supervised classification methodology at SC-R: a supervised classification methodology

This R TDF_algorithm classification code contains the functions for land cover classification for the identification of tropical dry forest. For the use of the code, a shape file containing the control points and a raster file with the bands to be used for the supervised classification is required. In this case we used Sentinel 2 bands B2, B3, B4, NIR and NDVI with 10m resolution. For an example of the application of this code, a smaller area of the PDA TDF clipped in the "raster_ROI.tif" file is applied, and we use as control points the “shape_points” file containing 534 points.

TDF10m-map is the Tropical Dry forest Map for the department of Cordoba in northern Colombia in raster and vector format.