Is vertebrate mortality correlated to potential permeability by underpasses along low-traffic roads?

Road permeability to animal movements depends among several factors on structures which, integrated in the road design, operate as safe conducts to mitigate vehicle collision and barrier effects. There is abundant evidence that wildlife makes use of such structures as safe passages to cross roads. We analyzed the spatial relationship between road drainage elements (N = 253; mostly culverts) as potential faunal underpasses, and mortality due to vehicle collisions in two seasons and on four relatively low-traffic roads (<5000 cars/day) traversing oak rangelands of western Andalusia (S Spain). Focusing on amphibians, reptiles and mammals, we recorded and located casualties (N = 238 individuals, 35 species) along these roads, identifying and characterizing all potential underpasses. Overall frequencies of casualties and spatial distribution were highly variable both within and among these roads. We obtained an estimation of potential permeability for the different roads. We detected, located and described a wide supply and a very variable pattern of drainage culverts and other underpasses, with differences among roads in passage attributes potentially affecting permeability for wildlife, such as spatial arrangement, number, density (frequency or concentration of passages) and dimensions. We used Mantel tests to assess spatial congruence of passages and road-killed animals. We applied generalized linear mixed models fitted by maximum likelihood through Akaike Information Criterion to explain the variation in the distance of the 238 casualties to the nearest underpasses, with road transect and season as random factors, and traffic intensity, speed and vertebrate class as fixed effects. Both road-killed animals and underpass distribution followed aggregated patterns, and casualties were not significantly related to underpasses along any of the 4 roads. There were no differences in distance of casualties to the nearest underpass for the three vertebrate classes. Although existing underpasses were abundant, we could not correlate potential permeability with reduced mortality along these roads, and other factors potentially affecting roadkill aggregations should be evaluated along with permeability assessment. Mitigation of road-caused mortality can still be greatly improved for these roads, through measures of reconditioning and proper management of existing underpasses, aiming to maximize road permeability and reducing major impacts upon animal populations of Andalusian rangelands.