Predictors of brown bear predation events on livestock in the Romanian Carpathians

Abstract Livestock depredation by brown bears is one of the main source of human–wildlife conflict in rural Eastern Europe. Thus, identifying environmental and anthropogenic drivers of human–bear conflict, and developing spatial predictions for predation intensity are critical to mitigate such conflicts. We used 756 records of bear‐caused livestock predation collected between 2008 and 2016 in the Romanian Carpathians and evaluated predictors and spatial distribution of bear livestock predation events (BPEs) using separate binomial generalized linear mixed models for cows, sheep, and other livestock. Despite differences in the direction and magnitude of the effect, the prevalence of BPE for all livestock was driven by the interaction between environmental drivers along with relative bear abundance. Distance from forest was a strong negative predictor for cows and sheep, while distance to villages was a strong negative predictor for cows. Landscape heterogeneity was positively associated with cow and other livestock predation and negatively associated with sheep. Relative bear abundance data collected by wildlife managers was a positive predictor for predation on all livestock. Livestock damage was more prevalent near villages, showcasing plasticity of food resources sought by bears. Our work informs brown bear and livestock management strategies to develop awareness and implement damage prevention measures.