Combining in situ surveys with remote sensing techniques for vegetation mapping and monitoring in the Natura 2000 site “Duna e Lago di Lesina-Foce del Fortore” (Puglia, S Italy)

In many parts of the world, and especially in the Mediterranean area, coastal dune systems are under severe pressures leading to different forms of degradation. In this scenario, continuative and consistent monitoring programs are fundamental to implement effective conservation strategies (1). The combination of in situ surveys with remote sensing methods and techniques has proven to be a robust approach for the monitoring of coastal ecosystems on large spatial and temporal scales (2). Specifically, we addressed our study to a Natura 2000 site in southern Italy (Puglia), namely IT9110015 - “Duna e lago di Lesina-Foce del Fortore”. The site is characterized by extensive dune systems hosting several vegetation types, many of which are habitat types under the 92/43 EEC Directive. The activities included field surveys and the analysis of remote sensing products (i.e., satellite images and orthophotos, by both aircraft and drone). The first ones were mainly aimed at characterizing the morphology and the structure of plant communities; the second ones at achieving an automatic habitat-tailored mapping system. Original vegetation data (about 120 relevés) were sampled according to the phytosociological approach between 2022 and 2023 and then were subjected to numerical analysis in order to define dune plant communities. The connection of plant communities to habitat types and vegetation transects according to the Italian Manual for the Monitoring of species and habitats of community interest (3) was performed to assess dune standard zonation; all this information was used to process indexes for the evaluation of the conservation status of the dune system habitats and vegetation types (4). The remote sensing approach followed different strategies: the simple visual interpretation of aerial multi-temporal orthophotos; the classification of Very High Resolution (VHR) satellite images by both Data-Driven Object-Based (DDOB) and Knowledge-Driven Object-based (KDOB) approach supported by the expert knowledge of the site; the combined use of machine learning techniques and ground truth data obtained through both field campaigns and Unmanned Aircraft System (UAS) acquisitions. Both UAS acquisitions and KDOB procedures have produced very promising results with a view to automatic mapping of the vegetation types in terms of coverage (distribution area) and occupied volume (biomass). Finally, a set of specific Landscape Metrics (LMs) (5) was applied to vegetation maps, across multiple spatial and temporal scales, to evaluate spatial patterns and assess their changes over time; results were consistent with those obtained from the field surveys. 1) Van der Biest K, De Nocker L, Provoost S, Boerema A, Staes J, Meire P, 2017. Dune dynamics safeguard ecosystem services. Ocean & Coastal Management, 149: 148-158. 2) Adamo M, Tarantino C, Tomaselli V, Veronico G, Nagendra H, Blonda P, 2016. Habitat mapping of coastal wetlands using expert knowledge and Earth observation data. J Appl Ecol, 53:1521-1532. 3) Angelini P, Casella L, Grignetti A, Genovesi P (eds), 2016. Manuali per il monitoraggio di specie e habitat di interesse comunitario (Direttiva 92/43/CEE) in Italia: habitat. ISPRA, Serie Manuali e linee guida 142/2016. 4) Helm A, Zobel M, Moles AT, Szava-Kovats R, Pärtel M, Rejmanek M, 2015. Characteristic and derived diversity: implementing the species pool concept to quantify conservation condition of habitats. Divers Distrib, 21: 711-721. 5) Tomaselli V, Mantino F, Tarantino C, Albanese G, Adamo P, 2023. Changing landscapes: habitat monitoring and land transformation in a long‑time used Mediterranean coastal wetland. Wetlands Ecol Manage, 31: 31-58.