Multitemporal remote sensing of vegetation dynamics through seasonal spectral composites and NDVI
Research abstract
This thesis develops a multitemporal Earth-observation workflow to reconstruct vegetation dynamics across almost three decades. In Google Earth Engine, seasonal median composites of the most informative spectral bands and NDVI are being generated for each year from 2015 to 2025, complemented by historical reconstructions for 2005 and 1996. Where necessary, sensor-specific correction or harmonisation factors are considered to improve comparability among acquisition periods. In GIS, false-colour infrared composites for the benchmark years 1996, 2005, 2015 and 2025 are used to support visual interpretation of long-term changes in vegetation cover and condition. The next phase will integrate the clipped raster products with additional field or reference datasets as they become available, allowing the remotely sensed trajectories to be quantified and ecologically interpreted.