| Summary: | Long-term temperature increases, higher frequencies of extreme weather events and changes in food web structures will all affect the state of Arctic tundra ecosystems at different temporal and spatial scales. Ecologists are tasked with understanding these biotic and abiotic interactions and finding methods to measure them. This thesis applies new technology and methods within the principles of adaptive monitoring to achieve four overarching goals: 1) Design a conceptual model for Svalbard’s moss tundra ecosystem and define the vegetation monitoring needs of high Arctic tundra systems in the context of climate change and herbivore management. 2) Design new monitoring approaches that help quantify habitat types and drivers of future vegetation state changes. 3) Evaluate the practical implications of using drone imagery, photogrammetry, and image classification-based approaches for monitoring. 4) Assess how the findings of the thesis can contribute to future adaptive monitoring of moss tundra. Drone images and random forest classifiers were reliably able to distinguish up to 15 different tundra ground cover classes, including those that represent disturbances such as winter damage from extreme weather events, pink-footed goose grubbing and bare ground. Snowmelt progression was mapped using drone and satellite images and combined with telemetry data to enable analysis of pink-footed goose behavior. This revealed a consistent correspondence, driven by vegetation class and snowmelt date, of habitat use and vegetation disturbance across spatial scales. Collecting ground truthing data in the field requires a good understanding of focal ecosystem components and their interactions with both abiotic and biotic factors, to not only detect visually distinctive, but also ecologically relevant ground cover classes. A close integration of detailed field-based assessments and drone images can elevate studies of causal ecological relationships into a spatial context. In addition, drone images will continue to improve the quality of information gained from satellite-based remote sensing.
|
|---|