Roadkill risk for capybaras in an urban environment

Abstract

Wildlife vehicle-collision is one of the most visible negative effects that roads exerts on animals and has increased dramatically across the world. Despite its conspicuousness, studies about roadkills in cities have been neglected lacking in road ecology. We developed new approach for estimating capybara roadkill hotspots in Campo Grande, a big city in Brazil. We also investigated potential driving factors correlated with roadkill occurrence, to build a predictive roadkill map for the entire city and propose mitigation measures. We monitored capybara roadkills for thirteen years and found hotspots using a graph-based kernel density estimation. We tested four predictors to identify which characteristics influence roadkill occurrence: distance from water bodies, distance from parks, vegetation cover, and traffic flow. We used a generalized linear mixed model to test for significant effects and to predict roadkill occurrences. Hotspots analysis showed four hotspots surrounding large green areas and water bodies, probably due to capybara habitat and physiological requirements. The predictive map shows latent hotspots, locations that have the characteristics necessary for a capybara to live but where we do not have observed deaths. To mitigate risk, we recommend using speed reduction tools around parks. A decrease in capybaras roadkills could positively impact human population welfare and material damage caused by these collisions.