The influence of road networks on brown bear spatial distribution and habitat suitability in a human-modified landscape

Abstract

Roads are human infrastructure that heavily affect wildlife, often with marked impacts on carnivores, including brown bears Ursus arctos. Here, we assessed the potential impact of road networks on the distribution of brown bears in the small, isolated and endangered Cantabrian population of north-western Spain. To ascertain whether local road networks affect brown bear spatial distribution, we first assessed potential influences on the distance of bear locations to roads using candidate models which included topographic variables, landcover types, bear age and reproductive status, traffic volume and road visibility. Then, we built two sets of habitat suitability models, both with and without roads, to discern the possible loss of habitat suitability caused by roads. The mean distance of bear locations to the nearest road was 968 +/- 804 m and the closest road was a low traffic road in 72.5% of cases. Candidate models showed little influence of our variables on bear distance to the nearest road, with the exception of elevation. Habitat suitability models revealed that road networks in our study area seem to have almost no effect on brown bear habitat suitability, except for females with yearlings during the denning season. However, this result may also be a consequence of the fact that only a small proportion (16.5%) of the cells classified as suitable bear habitats were crossed by roads, that is, most of the roads are primarily located in unsuitable bear habitats in the Cantabrian Mountains. Compared to previous studies conducted in other populations, mainly North American ones, our findings might suggest a different response of Eurasian brown bears to roads due to a longer bear-human coexistence in Europe versus North America. However, the indirect approach used in our study does not exclude other detrimental effects, for example, road mortality, increased stress and movement pattern disruption, only detectable by more direct approaches such as telemetry.