註釋 Unmanned aerial vehicles (UAVs or "drones") may provide cost- and time-effective approaches for gathering information for use in ecosystem assessment and monitoring. In this scoping study focussed on the sand dune system at Kaitorete Spit, Canterbury, we assessed the ability to discriminate plant and non-plant cover type information in UAV-collected multispectral imagery, in comparison with cover types derived from field data collected using plant plot surveys. Results showed that supervised image classification generated accurate and reliable cover type information for most cover types across the study area relative to those identified from the plot surveys. A number of important species, such as pīngao (Ficinia spiralis), muehlenbeckia (pōhuehue, Muehlenbeckia complexa) and tree lupin (Lupinus arboreus) could be differentiated in the imagery; for other cover types, there was inadequate spectral differentiation among individual species, resulting in discrimination only among mixed vegetation complexes. Overall accuracies of discrimination were in the range of 85 to 95%. Further, at 59 field plot locations, the relative proportions of most dominant cover types derived from image classifications correlated well with those quantified from field surveys. These initial results lend confidence in the use of UAVs to effectively and rapidly collect spatially-explicit data on cover type distributions, as a complement to ground-based surveys and/or to provide regular ecosystem updates at broader scale than survey work allows. We recommend that further work be carried out to quantify the standards with which UAV aerial surveys can be used most effectively for ecological monitoring purposes and the ultimate cost-benefit tradeoffs. 