Scaling Remote Sensing Data for Sustainable Grassland Conservation and Management

Grasslands are degrading globally due to land use and climate change, yet conservation is hampered by a lack of scalable and cost-effective monitoring. This thesis develops multi-scale remote sensing approaches to overcome this challenge, using case studies from Schleswig-Holstein, Germany. The research makes four main contributions: (1) A novel Magenta Vegetation Index is developed using drone imagery to accurately map key indicator plant species at a fine scale, with performance comparable to field surveys. (2) A drone-based change detection framework is established to monitor indicator species populations over time, enabling data-driven conservation planning. (3) A large-scale approach using satellite time series data (Harmonized Landsat Sentinel-2) and machine learning produces annual maps that identify high-nature value grasslands in Schleswig-Holstein, guiding future surveys. (4) Model explainability is enhanced using SHapley Additive exPlanation (SHAP) values, increasing trust and lowering barriers to the adoption of remote sensing in conservation. This work demonstrates that integrating fine-scale drone and broad-scale satellite remote sensing offers a transformative monitoring solution. By building confidence through model explainability, this thesis provides tools to support more effective grassland stewardship in an era of rapid environmental change.

Rights

Use and reproduction:


CC BY 4.0

Please note that individual components of the publication may be subject to other licensing or copyright conditions.

Cite

Citation style:
Could not load citation form.