Landscape patterns of ocelot–vehicle collision sites

Context Road networks can negatively impact wildlife populations through habitat fragmentation, decreased landscape connectivity, and wildlife-vehicle collisions, thereby influencing the spatial ecology and population dynamics of imperiled species. The ocelot ( Leopardus pardalis ) is a federally endangered wild felid in South Texas, with a high mortality rate linked to vehicle collisions. Objectives Using a multi-scale approach, we quantified and examined landscape spatial structure at ocelot roadkill locations, and between roadkill locations of male and female ocelots. Methods We quantified the spatial distribution of land cover types at 26 ocelot–vehicle collision sites in South Texas that occurred from 1984–2017. We compared landscape metrics of woody, herbaceous, and bare ground cover types across multiple spatial scales at roadkill locations to those from random road locations, and between male and female ocelots. Results Roadkill sites consisted of 13–20% more woody cover than random locations. Woody patches at roadkill sites were 7.1–11% larger (2.4 ha) closer to roads and spaced 10–16 m closer together farther away from roads compared to random locations. Percent woody cover was the best indicator of ocelot–vehicle collision sites; there were no differences in woody cover between male and female road mortality locations. Conclusion These findings suggest that ocelots are likely struck by vehicles while crossing between habitat patches. Roads that bisect areas of woody cover have negative impacts on ocelots by increasing habitat fragmentation and vulnerability to vehicle collisions. Crossing structures should be placed in areas with ≥ 30% woody cover and 3.5 ha woody patches.