Evaluating rangeland condition using fine-scale remote sensing from unoccupied aerial vehicles

This repository contains drone data from vegetation surveys conducted in 2024 in the Kalahari savanna ecosystem of Botswana. Surveys were carried out at three 50 × 50 m areas of interest (AOI) representing high, medium, and low grazing intensity. To acquire RGB images for SfM canopy height modelling, we used a DJI Phantom 4 Pro drone. To capture spectral reflectance data, we used the DJI Phantom 4 Multispectral drone. There are four main outputs for each survey: SfM point cloud—Generated in Agisoft Metashape from RGB imagery, provided in .laz format, representing the full 3D structure of the surveyed area. RGB orthomosaic—stitched from the same SfM process, using multiple overlapping RGB UAV images captured at a spatial grain of ca. 5 mm with 80 % front and side overlap. Canopy height model (CHM)—generated in R from the SfM point cloud, showing vegetation height above ground level. Multispectral orthomosaic—produced from multispectral imagery in Agisoft Metashape, including derived vegetation index layers. A CSV file contains ground-measured above-ground biomass (g/m²) from 30 circular plots per site (33 cm radius). These measurements are labelled with the corresponding plot ID codes, plot coordinates, AOI, and dominant herbaceous species. Please see the README file for further details on the data contents. For any queries on this dataset, please contact Alan Nare ([e-mail]) and Andrew Cunliffe ([e-mail]).