Abstract
Forest ecosystems provide essential services and have a multifaceted impact on the biosphere. However, global change poses major challenges to their future development, with mountain regions representing environments where considerable impacts are anticipated. The high environmental variability of these areas, together with a long history of land use common to many mountain regions, make it difficult to disentangle the effects of direct and indirect drivers of change. In this context, forest landscape models (FLMs) have emerged as powerful tools to explore forest dynamics under such complexity, integrating ecological processes and spatial heterogeneity. Nevertheless, given the wide range of dynamics that these models allow to explore, their use is far from straightforward, as they are often complex tools requiring large amounts of data from heterogeneous sources. In Chapter I, we presented a replicable and detailed approach for initializing a new landscape study in the Italian Eastern Alps (South Tyrol) within the iLand model framework, addressing a key methodological gap and enhancing the robustness and applicability of landscape models in the context of climate change and adaptive forest management. Setting up a landscape is a critical phase in simulation studies, yet the underlying procedures are often overlooked and rarely documented in the literature. Forest Landscape Models require considerable time for data preparation, testing, and calibration. Here, we documented the full process of establishing a new forest landscape and assessing its ability to reproduce current conditions against reference data on forest structure, composition, and natural disturbances. Overall, the digital landscape was highly consistent with the real one, accurately reflecting the forest characteristics in the reference data, thus demonstrating a strong robustness of the model in capturing key ecological processes. The initialized landscape provide a reliable framework for conducting simulation experiments on Alpine forest dynamics. By thoroughly documenting the entire process from data preparation to model validation, this work fills a key methodological gap and improves the practical use of forest landscape models. It highlights both the strengths and limitations encountered, thereby offering valuable guidance for future modeling efforts. In Chapter II, I focused on the investigation of the expansion of mountain forests and their upward limit (i.e., forest-grassland ecotone) in response to climate change and land abandonment using iLand. This chapter examined forest dynamics in an Alpine landscape in the Tyrolean Eastern Alps (Stubai valley), where I expanded the modeled landscape in iLand and simulated future developments considering land abandonment, climate change, disturbance regimes, and current management until the end of the next century. As a consequence of the increasing availability of high-elevated grasslands due to land abandonment, forests upward expansion into open areas occurred rapidly within this century already, with a net shift of the forest-grassland ecotone within the sub-alpine domain. Climate change differentially influenced species establishment and growth during expansion, particularly in relation to slope aspect (i.e., north- vs. south-facing slopes). While some species may benefit from future changes, extreme climate conditions could lead to local establishment failures for others (e.g. Stone pine). Based on my simulations, forest expansion and the upward shift of the forest-grassland ecotone can enhance climate change mitigation and ecosystem services provision. However, this shift is also projected to reduce open sub-alpine habitats, potentially threatening biodiversity. My process-based modelling approach proved effective in exploring forest succession in high-elevation abandoned areas, providing a solid framework for developing adaptive strategies to address these ecological changes. In Chapter III, I employed the landscape initialized in Chapter I to assess the naturalness of the current forest landscape, comparing it with a Potential Natural Vegetation (PNV) scenario simulated with iLand, that also incorporated natural disturbances. A spatially explicitly naturalness index, based on species dominance difference between simulated landscape versions, was then applied to quantify and compare the degree of naturalness of the current forest landscape. Successively, I simulated future forest dynamics under various climate scenarios for both the current and potential landscape versions, without forest management, assuming a re-naturalization of current forests due to the abandonment of management practices, in order to assess their respective vulnerabilities to future disturbance impacts. Overall, current forests exhibited a low naturalness compared to potential forests, largely due to historical land use practices across the Alps. However, more natural systems characterized by old growth forests, appeared to be more vulnerable to future disturbances, particularly to wind events. While past management practices favored Norway spruce stands, which are susceptible to bark beetle outbreaks, potential forests shaped solely by natural dynamics exhibited structural characteristics, such as canopy gaps and high height heterogeneity that likely increased their vulnerability to windthrows. As forests expand into increasingly unmanaged areas, my results suggest that leaving forest expansion to be driven by natural dynamics only should be carefully evaluated based on the local development context and the ecosystem services involved. For example, the development of more natural-structured forests along mountain slopes, due to their increased vulnerability to windthrows, could reduce protection against gravitational hazards, a negative effect likely to worsen under climate change. Therefore, adaptive management strategies could be designed to tackle forest vulnerability while trying to aim at higher naturalness levels. Although forest models require significant effort during the initialization phase, they offer a comprehensive framework to explore key ecological questions about future pathways of forest ecosystems. Given the growing concerns over climate change impacts on mountain forests, FLMs enable long-term simulations that can inform the development of nature-based solutions and support decision-making strategies for policy making. By accounting for regional and environmental variability, such tools allow for the evaluation of ESs provisioning and contribute to enhancing forest multifunctionality and resilience in a rapidly changing world.