Coupling a landscape-based approach and graph theory to maximize multispecific connectivity in bird communities

International audience; Environmental policies have highlighted the importance of preserving ecological networks to limit the fragmentation of natural habitats and biodiversity loss. A crucial issue for landscape managers is how to reconcile conservation measures that benefit all species and the maintenance of human activities. This study aimed to promote landscape multifunctionality, i.e., improving connectivity for several species without significantly modifying human activities. The objectives were to identify the most strategic natural landscape types to prioritize for preservation and to propose landscape management actions in highly anthropized areas that would benefit a majority of species. The analysis combined landscape types and bird species observation data to model landscape suitability for six species profiles defined by their affinity for wetlands, agricultural areas, urban areas and three types of forested landscapes. By graph modelling, we analysed the functional connectivity of the ecological networks of these species profiles. The results revealed that only ten landscape types out of 72 were core habitats for three species profiles simultaneously. These were primarily forested landscape types — either strict or associated with open areas (wasteland, forest clearing). Conversely, some anthropogenic landscapes dominated by built areas and sometimes shared with agriculture were completely unfavourable for all species profiles. The graph modelling analysis showed that the transformation of some landscape types could potentially improve connectivity for four species profiles presenting different ecological requirements. This coupling approach thus provided guidance to propose some landscape management actions that benefit the majority of species while preserving land uses.