Brown bear connectivity assessed through habitat models and landscape genetics

Resumen

Enhancement of landscape connectivity has become a central objective of numerous wildlife conservation initiatives around the world. This is particularly the case for the endangered brown bear (Ursus arctos) in the Cantabrian Range (NW Spain), where the remaining native populations of this species in the Iberian Peninsula are located. With a total population of about 200 individuals, Cantabrian brown bears are separated into two subpopulations by a wide unoccupied gap of about 40 km with low permeability for brown bear movements. The conservation and long-term persistence of this species depends on the maintenance of large habitat blocks and, particularly, on upholding connectivity between subpopulations, which is the cornerstone of current conservation strategies for this species. The general objective of this thesis was to provide an integrated and novel modeling and analytical framework in order to guide conservation efforts oriented to improve landscape connectivity for brown bears in the Cantabrian Range. On the other hand, we used the Cantabrian brown bears as a case study for a comparative analysis of different methodologies that are available to assess connectivity and to guide related conservation efforts. Our aim was to gain insights into the inherent assumptions, uncertainties and potential differences among existing approaches and their practical consequences for decision making. To accomplish these objectives we proposed a sequential framework based on empirical species data that combines some of the most recent developments for connectivity modelling. First, we used species occurrence data to develop a multiscale habitat suitability model, which allowed us to identify the main environmental drivers of habitat quality for brown bears and the scale at which they operate. Second, we estimated landscape resistance to brown bear movements based on genetic data and on the combination of multifactorial model-selection with reciprocal causal modeling. This analysis identified the main landscape factors (resistors) influencing the genetic structure of brown bears. Third, we compared genetic-based estimates of landscape resistance and effective distances between subpopulations with those derived from habitat suitability models. We examined the differences between the factors mediating the easiness of species movement through the landscape (gene flow) and those explaining the distribution and abundance of individuals in the study area. Fourth, we delineated regional corridor networks as the preferential movement pathways between and within subpopulations, using least cost-path analysis and circuit theory, and based both on habitat and genetic input data. Finally, we used recent methodological tools and connectivity metrics (habitat reachability, network centrality and factorial least-cost path density analyses) to provide spatially-explicit guidelines and identify critical areas for connectivity restoration and landscape matrix permeabilization for brown bears. In conclusion, this thesis provides integrated and tangible results that should be valuable to design improved restoration strategies aiming to maximize landscape permeability and connectivity for the endangered brown bears in the Cantabrian Range. In addition, the novel insights and improved understanding on the conceptual and practical differences between available approaches and metrics for connectivity analysis obtained in this thesis are of a wider scope and should be of interest for other species, management contexts and study areas in which connectivity is a major conservation concern.