Old growth forests preservation, protection, and promotion: a framework for Pollino National Park monitoring

Old-growth forests (OGFs) are natural ecosystems characterised by minimal human activity and exceptional structural and ecological complexity. Although rare, fragmented and often confined to remote areas, these forests play a crucial role in conserving biodiversity and maintaining ecosystem function. However, their structural complexity and long-term functional dynamics are not fully explored, particularly in Mediterranean regions, where climate change poses significant challenges. Forest structural complexity strongly influences biodiversity patterns, ecosystem functioning and resilience; however, quantitative analysis across spatial scales and over time still needs to be developed. This doctoral thesis aims to: (i) analyse forest structural heterogeneity in Mediterranean OGFs; (ii) upscale field-based structural indicators through advanced remote sensing modelling; and (iii) evaluate long-term canopy dynamics to test whether structurally complex forests exhibit enhanced functional stability. To achieve these aims, an integrated, scalable methodological framework combining field-based structural assessments, multi-source remote sensing, and long-term satellite time series was applied to old-growth forest (OGF) stands in Pollino National Park. In the first chapter, forest structural complexity was assessed at the plot level using the Structural Heterogeneity Index, which was derived from detailed field data. Multivariate ordination techniques were then employed to examine the relationships between structural attributes, environmental variables and plant species diversity. Results revealed the central role of structural heterogeneity in shaping plant community composition, interacting with topographic and site-related factors. In the second chapter, a multi-source remote sensing approach was implemented to spatially model forest structural complexity at a local scale. Random Forest model was applied to a range of predictors, including LiDAR metrics, multispectral indices, topographic variables and long-term vegetation trends, for classifying forest structural heterogeneity according to SHI classes. This approach enabled the production of high-resolution and spatially explicit maps of relative old-growthness, bridging ground-based ecological assessment and landscape-scale monitoring. In the third chapter, long-term forest canopy dynamics (1985-2024) were analysed using multi-decadal time series. Vegetation greenness trends, estimated at the pixel level, were assessed by combining Theil–Sen slope estimators and modified Mann–Kendall tests. Finally, block bootstrap resampling and effect-size metrics were implemented to compare trends across forest types and classes. The results revealed canopy greenness stability across the past four decades, while significant differences in trend magnitude were observed between old-growth stands and comparable managed areas. By integrating field-derived structural indices, remote sensing-based spatial modelling and multi-decadal time series analysis, this thesis advances a novel technology-driven framework for assessing forest structural heterogeneity and functional dynamics. The proposed approach moves beyond binary OGF classification, characterising old-growthness as a continuous and spatially variable property, and provides a robust decision-support tool for adaptive forest monitoring and conservation strategies under climate change.