Fine-scale determinants of vertebrate roadkills across a biodiversity hotspot in Southern Spain

Abstract

An increasing number of wildlife-vehicle collisions occur each year worldwide, which involves extensive economic costs and constitutes one of the main anthropogenic causes of animal mortality. Because of this, there is an urgent need to identify the factors leading to collision hotspots and thus implementing effective mitigation measures. By using a stratified random sampling survey, we investigated the fine-scale determinants of roadkill probability in small and medium-sized birds and mammals across a country-size region of Southern Spain, Andalusia (87,000 km2), located within a global biodiversity hotspot. During two consecutive seasons, we regularly surveyed 45 road transects of 10 km each and characterized the site-specific attributes of both roadkill and random points, including traffic density, road design (embankments, medians, fences, roadside vegetation and distance to curves), and adjacent landscape matrix. Based on this information, we investigated variation in collision risk according to landscape and road features, and the life history of the affected taxa. Mortality rates of mammals and birds increased with traffic density, and were also significantly affected by the distance to the nearest curve, slope of embankments, height of roadside vegetation, and land use adjacent to roads. Road mortality of both birds and mammals was related to the presence and typology of fences and center medians, so more densely vegetated medians and smaller mesh sizes reduced roadkill probability. Overall, our results indicate that roadkill risk may vary at exceedingly small spatial scales. The information provided by this extensive survey may be used to identify taxa-specific factors associated to roadkill risk and priority points for action. Our findings will therefore be relevant for the design of safer roads for both drivers and wildlife through the application of effective mitigation measures.