Hannah Hess, Columbia University

Brown bears (Ursus arctos) persist at high densities in the Romanian Carpathians, where they coexist within an intact mammalian community alongside expanding human infrastructure such as logging and agriculture. Effective bear management requires understanding both bear habitat selection and interactions with the broader mammalian community across space and time. Using a two-year camera-trap dataset of 13 mammal species, we applied piecewise structural equation modeling to quantify direct and indirect seasonal pathways shaping the mammalian network (including bears, large carnivores, mesocarnivores, and prey species) while integrating human activities and habitat characteristics. Preliminary results indicate that interspecific interactions are the primary drivers of community structure in the cold season, while environmental and human factors play a stronger role in the warm season. Bears exhibit distinct roles within the network between seasons. During the cold season, bear relative abundance (e.g., number of camera trap detections) predicted badger presence, suggesting a facilitative relationship, potentially through carrion provisioning or soil disturbance during foraging. In both cold and warm seasons, wild boar relative abundance consistently predicted bear density, indicating that bears may predate wild boar as well as track boar activity to locate concentrated food resources (e.g., ground-dwelling insects, tubers, etc.).

To translate these findings into a management-relevant spatial context, we used linear mixed models and GIS to map bear relative abundance based on wild boar occurrences across environmental gradients and human disturbance regimes. By linking these community-level processes to the landscape, we aim to anticipate where bear space use, mediated by facilitative interactions and human pressures, may elevate the risk of human-bear encounters, supporting targeted monitoring and proactive conflict mitigation strategies.