Drone imagery protocols to map vegetation are transferable between dryland sites across an elevational gradient

Abstract

The structure and composition of plant communities in drylands are highly variable across scales, from microsites to landscapes. Fine spatial resolution field surveys of dryland plants are essential to unravel the impact of climate change; however, traditional field data collection is challenging considering sampling efforts and costs. Unoccupied aerial systems (UAS) can alleviate this challenge by providing standardized measurements of plant community attributes with high resolution. However, given widespread heterogeneity in plant communities in drylands, and especially across environmental gradients, the transferability of UAS imagery protocols is unclear. Plant functional types (PFTs) are a classification scheme that aggregates the diversity of plant structure and function. We mapped and modeled PFTs and fractional photosynthetic cover using the same UAS imagery protocol across three dryland communities, differentiated by a landscape-scale gradient of elevation and precipitation. We compared the accuracy of the UAS products between the three dryland sites. PFT classifications and modeled photosynthetic cover had highest accuracies at higher elevations (2241 m) with denser vegetation. The lowest site (1101 m), with more bare ground, had the least agreement with the field data. Notably, shrub cover was well predicted across the gradient of elevation and precipitation ($230-1100 mm/year). UAS surveys captured the heterogeneity of plant cover across sites and presented options to measure leaf-level composition and structure at landscape levels. Our results demonstrate that some PFTs (i.e., shrubs) can readily be detected across sites using the same UAS imagery protocols, while others (i.e., grasses) may require site-specific flight protocols for best accuracy. As UAS are increasingly used to monitor dryland vegetation, developing protocols that maximize information and efficiency is a research and management priority.

Demography with drones: detecting growth and survival of shrubs with unoccupied aerial systems

“…Each flight mission produced a set of RGB images we processed using the structure from motion algorithm and georeferenced using ground control points following the protocol from Roser et al (2022). The output products included a dense point cloud and a Digital Surface Model (DSM) for each common garden.…”

Spatio-Temporal Transferability of Drone-Based Models to Predict Forage Supply in Drier Rangelands

“…Studies assessing the prediction accuracy of drone-based models beyond the calibration scenes conducted in different systems found varying results. For instance, [23] achieved similar prediction accuracy for shrubs in sagebrush steppe across different elevations but found inconsistencies in the prediction of grasses and bare ground. [24] showed how spatial heterogeneity due to diverse forest structures results in substantial differences between drone data and field measurements (more than 50%) when models were transferred to test sites for predicting forest attributes like stem volume.…”