Using Penrose distance to identify potential risk of wildlife–vehicle collisions

Abstract

Knowledge of the relationship between wildlife and roads is important for management of wildlife–vehicle collisions, which represent a serious threat to many wildlife populations and to human life and property. Effective reduction of these threats requires identification of variables influencing collision locations, and the use of these variables in spatially-explicit predictive models. We used Penrose distance modeling and 61 confirmed bobcat (Lynx rufus) road mortality locations in southern Illinois, USA, to demonstrate a rapid and accurate technique for the spatial mapping of wildlife–vehicle collision risk. We used the Penrose distance statistic to quantify the similarity between the mean multivariate habitat signature at bobcat-collision areas and road sections throughout the study area. Habitat variables assessed included road-related variables (e.g., traffic volume) and land cover characteristics (e.g., mean patch area of the landscape). Bobcat-collision areas were characterized by smaller, less-clustered habitat patches and more, large independent patches of grass cover than were study area roadway sections in general. As expected for a generalist carnivore in highly-suitable habitat, risk mapping indicated that large sections of the roaded landscape had high similarity to bobcat-collision areas. Unlike other modeling techniques used to identify risk of road mortality, our method requires little field data collection and relies on readily available digital spatial data. This technique can be used by wildlife managers and road planners, and may be particularly important in reduction of road mortality for species such as the Iberian lynx (Lynx pardinus), Florida panther (Puma concolor coryi), and Texas ocelot (Leopardis pardalus), which exist in small populations fragmented by development, and are limited by road mortalities.