Functional Analysis for Habitat Mapping in a Special Area of Conservation Using Sentinel-2 Time-Series Data

The mapping and monitoring of natural semi-natural habitats are crucial activities regulated by European policies regulations, such as the 92/43/EEC. In Mediterranean area, which is characterized high vegetational environmental diversity, particularly difficult often exclusively based on in situ observations. this scenario, it necessary to automate generation updated maps support decisions policy makers. At present, availability spatiotemporal resolution data provides new possibilities for improving habitats. work, we present a methodology that, starting from remotely sensed time-series data, generates habitat using supervised classification supported Functional Data Analysis. We constructed Sentinel-2 Special Area Conservation “Gola di Frasassi” (Code: IT5320003). particular, training set uses 308 field plots with 11 target classes (five forests, two shrubs, one grassland, mosaic, extensive crop, urban land). Starting vegetation index Principal Component Analysis was applied derive FPCA scores components. stage, considered features. obtained results out-performed previous map derived photo-interpretation domain experts. an overall accuracy 85.58% time-series, topography, lithology data. main advantages proposed approach capability efficiently compress dimensional (dense remote-sensing time series) providing compact way (e.g., mean seasonal profiles) that: (i) facilitate link between remote sensing their ecological interpretation (ii) could be complementary species-based approaches plant community ecology phytosociology.