Landscape change and forest dynamics: multi-scale pattern analysis and biodiversity issues

This thesis was conceived as a contribution to the increasingly urgent need to have sound scientific basis and effective tools to understand, monitor and manage the landscape in order to preserve the biological diversity that characterize landscapes. In this work, concepts, theories, methodologies and tested hypothesis, are proper of Landscape Ecology. Landscape Ecology is a young branch of modern ecology that deals with interrelationship between the spatial pattern and ecological processes in a given land mosaic. In particular, this thesis work offers a double contribution: First, it provides new methodologies for the multi-temporal and multi-scale analysis of the landscape and its changes and; second, it applies Landscape Ecology theories and methods to the conservation and the management of landscape biodiversity issues. The thesis work is organized into three sections, which are in turn subdivided into different chapters. The first section attempts to answer the following question: (1) When Are Two Landscapes Significantly Different? Applications in Temperate Forest Landscapes (Chapters 1, 2). The second section focuses on the analysis of forest fragmentation over time in subtropical forests landscapes: Measuring Forest Fragmentation Over Time: Applications Subtropical Forest Landscapes (Chapters 3-5). Finally, the third section deals with some implementations of Landscape Ecology in biological conservation: Applying Landscape Ecology in Biological Conservation (Chapters 6-10). The first section of the thesis focuses on the development of new methods in order to (i) provide the statistical significance of the observed differences when comparing different landscapes or their changes over time and (ii) analyze the scale dependence of the spatial pattern change. To properly handle the first issue, a new method to test the statistical significance of pattern metric values between two maps was proposed. In particular, the method was based on the use of Neutral Landscape Models (NLMs) able to reproduce a set of real-like (in terms of composition and configuration) simulated landscapes. For each simulated landscape, pattern metrics can be computed and can used to produce null empirical distributions for each metric. These empirical distributions can be used to determine if, the observed differences in terms of pattern metrics are significant for a given confidence interval. The second goal was achieved by extending the NLMs approach into the spatial domain to analyze the scale-dependence of landscape changes. Both methods were demonstrated in temperate forest landscapes. The second section deals with the analysis of forest fragmentation in subtropical ecosystems. Also in this case new methods to quantify the spatial pattern of forests over time were proposed. First, the analysis of the spatial pattern change by means of sample-based techniques. By randomly sampling forest cover multi-temporal maps it was possible to statistically quantify and distinguish changes in forest cover from changes in the spatial pattern of forest. Second, the process of forest fragmentation was studied according to the †œlandscape context analysis†� that is, each forest location was described by its context, using indices derived from the landscape that surrounds it, across a range of scales (from local to regional). These methods each have their own set of advantages and limitations, and yet may work in a complementary manner. Moreover, in this work the usefulness of the proposed techniques was demonstrated in different environments such as temperate and subtropical forests. However, they can be easily extended to other landscapes such as urban landscapes and, in this respect, they could be an interesting field of further researches. The third section presents a series of study cases where landscape ecology was applied to biodiversity conservation. In particular, five works were presented: the first deals with the relationship between landscape structure and biodiversity in temperate forest landscapes. The second focuses on the impact of windfarms on the foraging habitat of two bat species at landscape level. The third was aimed to the study of the habitat selection by terrestrial tortoise in high conservation value protected areas. The fourth work analyzes short-term variations of high vegetation summit to the global warming. Finally, the fifth study case deals with the spatial estimation of forest age structure on the basis of remote sensing images and yield models. All the presented contributions deal with landscape ecology. Landscape change is known as one of the most important driver of the biodiversity crisis worldwide. In this respect, the present thesis attempted to give an important contribution to the quantification of the landscape spatial pattern and thus to the relationship between spatial pattern and ecological processes (e.g., fragmentation, habitat selection, etc.). Moreover, these findings potentially addressed the urgent need to have useful tools in order to define conservation measures and management plans, which are essential for the protection of natural and semi-natural habitats and their threatened species in a changing world.