Abstract ID: 3.12488 | Accepted as Poster | Poster | TBA | TBA

Mountain treeline ecotones are transition zones between the closed forest and the upper treeless vegetation and are considered sentinels of global change effects on terrestrial ecosystems. Treeline ecotones are constrained by multiple factors acting at different spatial and temporal scales. For this reason, we present OLYMPUS, a project based on a multiscale approach to monitor treelines in the Italian Alps and Apennines, the two main peninsular mountain ranges. The specific goals of the research are: 1) to detect the current forestline and to assess its drivers and dynamics through LANDSAT time series; 2) to assess human pressure and land abandonment effects on the treeline ecotone by adopting a diachronic approach with historical (70 years) aerial images; 3) to evaluate the role of biotic and abiotic factors on tree establishment and survival by integrating centimeter UAV images with field data collection and manipulation experiments. Landsat images of the last 4 decades were used to infer long-term trends in vegetation dynamics (greening and wetness) at the regional scale. At the landscape scale (10-20 km2) we measured land-use change dynamics with landscape metrics and treeline ecotonal shift through a deep learning land-cover classification of historical (1950s) and current (2020s) aerial photographs. A fine-scale assessment of treeline ecotones included a selection of 10 treeline ecotones where a UAV-based structure-from-motion (SfM) photogrammetric approach was used to derive local land cover and seedling position maps. We analyzed biotic and abiotic drivers of seedling establishment at the treeline through spatially-explicit statistical analysis and tree growth performance and patterns were assessed by tree-ring analysis. The overall contribution of the project is its spatially explicit approach crossing different disciplines (remote sensing, landscape ecology, forest ecology, dendroecology, etc.) and different spatial scales (region-landscape-tree).