A novel high-resolution eco-functional vegetation mapping of coastal dunes

Abstract

Coastal dune landscapes are dynamic systems shaped by complex biotic-abiotic interactions and impacted by human disturbances. This study focuses on Mediterranean coastal dunes, analyzing spatial distribution and ecological characteristics of their plant communities through an innovative approach that integrates vegetation functional traits with UAV-based remote sensing. Using synchronized vegetation sampling and UAV surveys along the central Adriatic coast, we identified and mapped distinct eco-functional types within the dune landscape. Field sampling recorded 46 taxa, reflecting diverse habitat types according to the EUNIS classification system. Clustering based on community traits revealed four eco-functional classes with notable differences in seed mass, taxonomic diversity, and species composition, as well as variations in remotely sensed environmental parameters. The first class, situated near the drift-line, included Thinopyrum junceum, a species adapted to wind and sand movement. The second class bridged dynamic foredune and more stable areas, featuring ruderal species like Scabiosa atropurpurea. The third class, impacted by moderate environmental stresses, included Silene colorata and Bromus madritensis, with a scattered, transitional distribution. Finally, the fourth class, characterized by dense, stable vegetation, was typified by Festuca fasciculata, indicating mature dune grasslands. Our findings underscore the potential of integrating UAV remote sensing with functional trait analysis to enhance vegetation classification accuracy and enable fine-scale spatial mapping of coastal dunes. This approach offers a valuable tool for ecosystem monitoring, supporting conservation management amid rising environmental pressures.

Keywords

• Mediterranean dunes
• Plant traits
• UAVs
• LTER
• Machine learning