The newest sensor technology mounted on drones (Unmanned Aerial System, UAS) provides novel opportunities for assessing how biodiversity and its drivers change at ultra-high spectral (wavelengths in nanometres), structural (dense 3D point clouds), spatial (millimeter to meters), and temporal (days to years) resolution. In particular, combining measurements of highly differentiated spectrometric signals and detecting the range of emitted laser light pulses, show great potential for simultaneously assessing the functional, structural, and taxonomic components of biodiversity, as well as its environmental drivers (e.g. hydrology, topography, nutrient status). BIOSENS is one of the first projects worldwide that combines the newest hyperspectral and LiDAR sensor technology for UAS with a detailed assessment of temporal and spatial changes in local diversity and ecological parameters. Since 2017, repeated droneflights have been conducted across a controlled grassland experiment (in Bern, Switzerland), with controlled levels of plant diversity, as well as across a natural grassland with substantial variation in plant diversity, vegetation structure, function, and ecological factors (Rewilding area, Mols, Denmark). Simultaneously, highly detailed and spatially explicit information on plant diversity and structure have been measured with traditional ecological methods and hand-held hyper-spectral sensors. Using the LiDAR data collected across the Rewilding area, Mols, Denmark, we have been able to recognize and map specific shrub species in 3D. We especially targeted the classification of Cytisus scoparius, because of the particular concern in landscape management (Fig. 1), and we successfully estimated biomass changes between autumn 2017 and spring 2018. The next steps are to understand the drivers of these changes.
With the grassland experiment in Switzerland, we have a unique opportunity to test different UAS set-ups in a controlled environment (Fig. 2). Thus, we are currently evaluating the accuracy of plot-scale biomass estimates from non-destructive LiDAR measurements, and how vegetation can be assessed using structural features. Furthermore, we will use hyperspectral measurements from 20 different grassland species to analyze the impact from manipulative treatments (nitrogen additions and pathogen exclusion) on the performance of recognizing taxonomic and functional diversity. This project contributes to theme  Fundamental Biodiversity Dynamics and theme  Ecoinformatics and New Technologies of BIOCHANGE.