Assessing landscape resistance to roe deer dispersal using fuzzy set theory and multicriteria analysis. A case study in Central Spain.

The central Iberian Peninsula has one of the highest densities of roe deer populations in Spain. A new motorway is planned to pass through the middle of the distribution of roe deer, thus making it necessary to conduct a connectivity analysis. A map of resistance to roe deer dispersal movements was obtained based on the literature and expert judgment. Three factors were selected: land use (defined by the ability to hide movements, food source, and degree of naturalness), landforms, and influence due to proximity to elements that increase (such as roads and urban areas) or decrease (water resources and proximity to optimal habitat patches) resistance at the local level. Different combinations of factors derived using the analytical hierarchy and fuzzy logic processes were analysed and compared with the real distribution of the species. More realistic resistance (cost) values were obtained for gamma values close to 0.9. This highlights the greater predominance of the fuzzy sum over the fuzzy product in modelling the cost surface. Better results were obtained in scenarios where the predominant factors were either land use and landforms or land use and proximity to humanmodified areas. This indicates that roe deer will readily range far from their optimal patches if the land use provides partial cover. These movements appear to be conditioned by steep terrain. Our case study offers a good example of building a cost resistance matrix to help locate areas where the expansion of the species may be curbed or encourage