Automated vegetation monitoring for forest restoration
ABSTRACT: We discuss the potential of automating vegetation monitoring, to aid forest restoration. We propose that automated monitoring focuses on estimating forest biomass and tree diversity, because these are relevant to many ecosystem services, and they can be assessed with existing automated technologies, to some extent. We discuss the importance of setting baselines and realistic goals that take into account site history and landscape context. We review relevant technologies, including unmanned aerial vehicles (UAVs), lidar, multi-spectral and hyperspectral sensors, visible-light cameras and data-processing software. We discuss advantages and disadvantages of below- versus above-canopy surveys. We identify technological obstacles to automated monitoring, including tree-species identification in diverse forests, and the assessment of forest structure in high-density forests. These obstacles are particularly relevant to tropical forests, which are typically dense and diverse. We also identify battery lifetime as a limitation to large-scale surveys, and one that is unlikely to be alleviated soon. Despite these caveats, available technology is adequate for automating small-scale assessments of some forest variables that are relevant to restoration, particularly in less dense, less diverse temperate and boreal forests. A fruitful approach may be to use intensive ground-level and low-altitude automated surveys, to calibrate data from satellite imagery that is subsequently applied to monitor restoration over larger areas.