Biodiversity takes flight: mapping plant species in western rangelands with Uncrewed Aerial Vehicles (UAVs)

Authors: MALIHA, MAISHA - Boise State University; CATTAU, MEGAN - Boise State University; Copeland, Stella; DOLMAN, MEGAN - Boise State University; Olsoy, Peter; RACHMAN, RICHARD - Boise State University; TAGNEY, AMETHYST - Boise State University; SALVATIERRA, SANDRA - Boise State University; WICKERSHAM, RYAN - Boise State University; ZAIATS, ANDRII - Boise State University; CAUGHLIN, TREVOR - Boise State University

Submitted to: Rangeland Ecology and Management
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 6/13/2026
Publication Date: N/A
Citation: N/A

Interpretive Summary: What is the problem? What did you do related to the problem? What did you find out/what were your results? Why does your answer to #3 matter (i.e., what’s the impact?)? Describing and monitoring plant species in rangelands at broad scales is time and resource intensive. Uncrewed Aerial Vehicles (UAVs) or drones may be able to address this challenge, but it is unclear whether current imagery and computational methods can accurately identify different types of plants to species. This study showed that low-cost imagery from UAVs combined with computational tools can reliably identify many common sagebrush steppe species, though accuracy decreases in sites with different common plant species. This work suggests that drone imagery could be used to monitor rangeland vegetation in sites with similar plant species composition, along with field verification to confirm plant species identification.

Technical Abstract: Monitoring plant species composition at management-relevant scales remains a persistent challenge in rangeland ecosystems. High-resolution imagery form Uncrewed Aerial Vehicles (UAVs) offers a promising solution, but high sensor costs and the need for extensive field training data have hindered widespread adoption. We evaluated the performance and transferability of UAV-based machine learning models for species classification across 14 big sagebrush (Artemisia tridentata L.) steppe landscapes in the northern Great Basin. We tested models for 18 common overstory plant species in different functional groups including shrubs, Artemisia arbuscula Nutt. (low sagebrush) and Ericameria nauseosa (Pall. ex Pursh) G.L. Nesom & Baird (rubber rabbitbrush), and the perennial bunchgrass Pseudoroegneria spicata (Pursh) Á. Löve (bluebunch wheatgrass). Using structure-from-motion photogrammetry and a stacked ensemble learning approach, we achieved a mean classification accuracy of 92.1% (95% CI: 90.9–93.2%) and a weighted F1 score of 91.5% (95% CI: 90.2–92.8%), indicating strong performance despite a highly imbalanced dataset dominated by a few common species. Models trained with low-cost RGB imagery performed nearly as well as those using multispectral data, with only a ~1% difference in F1 score. However, model transferability was limited: classification accuracy declined sharply at sites where species composition differed from training data, with F1 scores ranging from <0.09 to >0.90 across test sites. These results suggest that while low-cost UAVs can produce accurate and scalable species maps, reliable application across diverse rangelands will require strategic field sampling or shared training datasets. Our findings provide practical guidance for researchers and land managers seeking to incorporate UAV technology into biodiversity monitoring, restoration planning, and invasive species management.