Talk Watch the road: Assessing roadkill in Addo Elephant National Park

Roads are an increasing threat to biodiversity, particularly in the case of developing countries such as South Africa of which 364 131 km of roads criss-cross the country (17% paved) fragmenting much of the viable habitat available to wildlife. Roadkill is the most visual impact that roads have on wildlife populations and whilst the majority of roadkill studies have occurred in ‘western countries’, understanding of the threat from roads on biodiversity has increased in the last five years in South Africa. Globally, understanding of roadkill events is for national and regional roads, with little known of the impacts of roads in protected areas – the prime custodians of biodiversity, intended for the conservation of flora and fauna and ecosystems.

South Africa contains 23 national parks with 6.3% of the country committed to terrestrial protected areas. This study was conducted in Addo Elephant National Park (AENP). Our aim was to determine which factors of the road and roadside environment are correlated with roadkill. A priori, we created a model of roadkill risk based on characteristics of the road and roadside environment, where proximity to lodges, a paved road surface and low vegetation density were considered higher risk. We then assessed the ecological characteristics of the landscape as well as road characteristics at actual roadkill events, by driving transects in AENP on two randomly selected days/month over a six month period (May to October, 2015; winter).

In AENP, we found more invertebrate (63) roadkill than vertebrate (23) roadkill throughout the study period. The distance between roadside vegetation and the road significantly influenced the presence of roadkill: the closer the vegetation to the edge of the road the higher the likelihood of roadkill events occurring. Related to this was grass thickness on the verge of the roads: where thicker grass occurred there were fewer roadkill events. Road type and the regulations implemented on the road are correlated with roadkill events. Roadkill in AENP was related to the vegetation characteristics, the road surface and status (park road or provincial road) in the park having the largest impacts on biodiversity. Modelling can be used to determine where roadkill could potentially be highest using predictors determined in this study.

Vegetation, biodiversity, vertebrates, invertebrates, hot spots