Previous assessments of wildlife road mortality have not used directly comparable methods and, at present, there is no standardized protocol for the collection of such data. Consequently, there are no internationally comparative statistics documenting roadkill rates. In this study, we used a combination of experimental trials and road transects to design a standardized protocol to assess roadkill rates on both paved and unpaved roads. Simulated roadkill were positioned over a 1 km distance, and trials were conducted at eight different speeds (20–100 km·h−1). The recommended protocol was then tested on a 100-km transect, driven daily over a 40-day period. This recorded 413 vertebrate roadkill, comprising 106 species. We recommend the protocol be adopted for future road ecology studies to enable robust statistical comparisons between studies.

No abstract

Roads impact wildlife through a range of mechanisms from habitat loss and decreased landscape connectivity to direct mortality through wildlife-vehicle collisions (roadkill). These collisions have been rated amongst the highest modern risks to wildlife. With the development of "citizen science" projects, in which members of the public participate in data collection, it is now possible to monitor the impacts of roads over scales far beyond the limit of traditional studies. However, the reliability of data provided by citizen scientists for roadkill studies remains largely untested. This study used a dataset of 2,666 roadkill reports on national and regional roads in South Africa (total length ∼170,000 km) over 3 years. We first compared roadkill data collected from trained road patrols operating on a major highway with data submitted by citizen scientists on the same road section (431 km). We found that despite minor differences, the broad spatial and taxonomic patterns were similar between trained reporters and untrained citizen scientists. We then compared data provided by two groups of citizen scientists across South Africa: (1) those working in the zoology/conservation sector (that we have termed "regular observers," whose reports were considered to be more accurate due to their knowledge and experience), and (2) occasional observers, whose reports required verification by an expert. Again, there were few differences between the type of roadkill report provided by regular and occasional reporters; both types identified the same area (or cluster) where roadkill was reported most frequently. However, occasional observers tended to report charismatic and easily identifiable species more often than road patrols or regular observers. We conclude that citizen scientists can provide reliable data for roadkill studies when it comes to identifying general patterns and high-risk areas. Thus, citizen science has the potential to be a valuable tool for identifying potential roadkill hotspots and at-risk species across large spatial and temporal scales that are otherwise impractical and expensive when using standard data collection methodologies. This tool allows researchers to extract data and focus their efforts on potential areas and species of concern, with the ultimate goal of implementing effective roadkill-reduction measures.

No abstract

Roads are a major cause of wildlife mortality by animal-vehicle-collisions (AVCs). We monitored the patterns and frequency of AVCs on two sections of a major highway in Northern Tanzania and compared these patterns to the knowledge and perceptions of drivers who frequently use the roads. While actual field survey showed that more birds were killed by AVCs, mammals were perceived by the drivers to be the most common AVC. Drivers were indifferent to whether AVCs were a major problem on the road, and 67% strongly felt that AVCs were mainly accidental, either due to high vehicle speed or poor visibility at night. There was a negative correlation between the likelihood of a species being hit by vehicles and its average body mass. Only 35% of drivers said they had attended an educational program related to the impact of roads on wildlife. This study highlights a need for collaborative efforts between the wildlife conservation and road departments to educate road users on the importance of driving responsibly and exercising due care for wildlife and human safety. This should be coupled with effective mitigation measures to reduce the extent of AVCs.

Across Africa, transport infrastructure, including roads, is being built in over 30 planned development corridors that are likely to have major impacts on remaining natural habitats and associated biodiversity. Linked to this is the projected increase in human population size, which is predicted to grow by 1.3 billion people by 2050. Road ecology is the study of the ecological effects (both positive and negative) of roads and traffic but is perceived to be under-researched in Africa. In this context, we undertook a systematic review of road ecology research in Africa to understand the geographic and taxonomic scope of work undertaken to date, as well as recommendations for reducing the impacts of roads. We discovered 210 road ecology publications from Africa (published between 1954 and 2016), with most publications from the more affluent and politically stable regions (e.g., southern and East Africa). We found more publications than expected, with relevant research concealed within studies whose primary focus was on other topics. Most publications (1) focused on single species, and in particular on mammals (where chimpanzees and forest elephants were most studied); (2) were from southern Africa; and 3) were conducted in the grassland and savannah biome and the tropical and subtropical forest biome. Most publications examined the direct impacts of roads, in particular wildlife-vehicle collisions. Only one-third of the publications provided some form of recommendation for intervention to reduce or mitigate the impacts of roads, based on evidence from the publication. Most recommended interventions related to ecosystem or natural process recreation, as well as site and area stewardship. Gaps and future directions for research include rigorous testing of measures to mitigate the impacts of roads, inclusion of traffic monitoring in studies, understanding the impacts of upgrading roads, and exploring livelihood, economic and moral incentives and education and training as potential interventions for reducing the impacts of roads. Our review has highlighted the need for accelerating the study of the impacts of roads on natural habitats and biodiversity, in light of planned large-scale infrastructure development, and especially the study of appropriate mitigation measures that can be rigorously assessed and implemented before and during construction and upgrading of roads in Africa.

Using a standard protocol, we conducted vertebrate roadkill surveys in the Greater Mapungubwe Transfrontier Conservation Area (GMTFCA), South Africa, which is a World Heritage Site. A total of 991 roadkill were recorded on the paved roads and 36 roadkill on the unpaved roads. Identifiable roadkill comprised 162 species from 24 orders and 65 families. Ninety-three roadkill could not be identified to species level. Roadkill counts were strongly influenced by road type and season. More roadkill was recorded on the paved than the unpaved roads. Irrespective of road type, the proportion of roadkill was greatest in the hot/wet season (4.3 paved roadkill/km/day paved and 1.3 roadkill/km/day unpaved) and lowest in the cold/dry season (2.0 roadkill/km/day paved and 0.1 roadkill/km/day unpaved). The high numbers of vertebrates identified as roadkill suggests that road traffic has the potential to directly and negatively affect biodiversity conservation in this part of South Africa. We recommend continued roadkill data collection across South Africa to assist with creating an inventory of species most likely to be at risk from roads. This will, in turn, better inform the implementation of potential mitigation measures.Key words: roadkill, protocol, vertebrates, ecological season. INTRODUCTIONGrowing concern about the ecological effects of roads has led to the emergence of road ecology as a scientific discipline (Forman, 2003; Fahrig & Rytwinski, 2009). Roads impact ecosystems in two ways; indirect impacts through fragmentation of habitat (Hels & Buchwald, 2001) and direct impacts via mortality (i.e. roadkill; Clevenger, Chruszcz & Gunson, 2003). Roads therefore pose a threat to the survival of individual animals and entire populations.Here, we use a uniform protocol for roadkill data collection (Collinson, Parker, Bernard, Reilly & Davies-Mostert,2014), to document the species diversity of roadkill in the Greater Mapungubwe Transfrontier Conservation Area (GMTFCA), Limpopo province, South Africa. In addition, we assess the influence of season, road surface type and animal activity patterns on roadkill rates. METHODSOur study area was the GMTFCA (Fig. 1) which falls within the sub-tropical region of South Africa and is in a dry Savanna subregion of the Mopane Bioregion (Schulze & McGee, 1978). The study area is characterized by hot (17-27°C) summers and mild (4-20°C) winters with occasional frost (Nel & Nel, 2009). The mean annual rainfall is 278 mm but can be as low as 154 mm during dry years and as high as 451 mm per annum during wetter years (Nel & Nel, 2009). The rainy season is predominantly from November to March (summer) when the province receives 90% of its total annual rainfall (M'Marete, 2003).The area contains a high species richness of reptiles (~120 species; Branch, 1998) We drove transects 1.5 hours after sunrise, at a speed of 40-50 km/h, using a single, trained observer (Collinson, Parker, Bernard, Reilly & Davies-Mostert, 2014). A 100 km section of paved road and a 20 km section of unpaved road were dri...

Social media discussions highlight public concern for wildlife‐vehicle collisions (WVCs) inside protected areas. Using a quasi‐experimental field trial, we investigated factors affecting the likelihood of WVCs within Pilanesberg National Park, South Africa, and assessed the comparative effectiveness of wildlife‐warning signage (WWS) for altering driver behaviour. We laid a dummy snake crosswise on roads across four combinations of habitat and road shape and recorded 10 driver‐related variables for 1454 vehicles that passed the dummy snake, including whether there was a collision. An interaction between speeding and driver occupation (staff/visitor) was the best indicator for WVC. When driving below the speed limit, visitors were almost three times more likely than staff to hit the dummy snake. Collision probabilities increased when speeding and became more similar between visitors and staff, although still significantly higher for visitors. We then investigated the effectiveness of roadside signage in modifying driver behaviour by erecting four variations of WWS, depicting a snake or a cheetah, and in photographic or silhouette form. We positioned the dummy snake 100 m or 1 km after the signage and recorded our 10 variables (n = 6400 vehicles). Sixty‐one per cent of drivers who passed a WWS changed their behaviour when they saw the dummy snake, compared to 37% with no sign present. Further, this behaviour change significantly reduced collisions, where 98% of drivers who changed their behaviour avoided a collision. Finally, an interaction between the animal depicted and distance before the dummy snake affected collisions. A WWS depicting a snake, and placed 100 m before the dummy snake, was most effective at reducing collisions. Our results suggest that drivers adapt their behaviour to signage that portrays smaller animals and awareness retention is low. Ultimately, to reduce WVCs within protected areas, we suggest steeper penalties for speeding and WWS placed in WVC hotpot areas.