Bridging applied ecology and network theory to improve landscape management for conservation

Agricultural intensification is widely considered a major threat to biodiversity and ecosystem services. To better protect biological communities, it is necessary to gain a more detailed understanding of the way species use habitats and move across increasingly simplified landscapes. Traditional landscape ecology approaches are mainly based on the dichotomy between focal semi-natural habitat patches and the surrounding agricultural matrix. While the advances made possible by the landscape mosaic model are undeniable, this approach fails to account for real-world complexity, as many species are known to use multiple habitat types (natural and disturbed) during their life cycle. In addition to landscape changes, many local factors (including management such as soil disturbance or pest control) can impact biodiversity, often interacting with each other. In order to inform efficient biodiversity management actions in the future, it is crucial to increase our knowledge on the way local and landscape factors can impact biodiversity at multiple spatial scales. The general aim of this thesis was to develop a novel approach to the study of species-habitat interactions, and to apply the approach to answer some pressing questions about the way landscape simplification influences important arthropod functional groups. Additionally, we studied the effects of local factors on an arthropodmediated ecosystem service (weed seed predation) in the same area. Arthropods were chosen as they are among the most abundant and ecologically relevant organisms in agroecosystems, providing a wide variety of pivotal services. The new approach, based on network theory, showed that landscape simplification reduces habitat specialization in low-mobility insect groups, and allowed us to pinpoint the most important habitat types for the facilitation of arthropod movement through the landscape mosaics. Our local-level study, on the other hand, highlighted how multiple factors can interact in a complex way in shaping ecosystem services. This study demonstrates the potential of the novel species-habitat network approach as a complementary tool for investigating landscape-biodiversity interactions, while simultaneously unveiling new information on the way landscape changes and local factors influence key arthropod groups. This enabled us to provide a series of recommendations for biodiversity management actions, varying depending on the target group. Overall, our research is a reminder of the importance of taking into account multiple potentially interacting factors at different spatial scales to correctly understand and manage biodiversity-related processes