The response of wildlife populations to mortality resulting from collisions with vehicles is shaped by demographic factors, available habitat and the spatial patterns of road networks. Here, we introduce an innovative spatially explicit individual-based model to predict the extinction risk faced by populations due to road mortality. We present two case studies from India, one on the leopard (Panthera pardus, IUCN Category: Vulnerable), and the other on the leopard cat (Prionailurus bengalensis, Least Concern). We modelled the risk of local extinction for each species, at a 10x10km resolution, in India for 2030, 2050, 2080 and 2100, considering the average, minimum and maximum roadkill rates observed.
Our findings indicate that for the leopard, by 2100, under the most favorable roadkill scenario, approximately 22.26% of its range faces a 50% or higher likelihood of local extinction, whereas under the worst-case scenario, it escalates to 64.33%. Even by 2030, 5.32% of its range has a >50% chance of extinction under the best-case scenario, and 40.96% under the worst-scenario, with most of the loss expected to occur south, where the road network is denser. In contrast, the leopard cat exhibits significantly different outcomes, with roadkill exerting minimal impact on its distribution. In the worst-case scenario, the expected decrease is 0.16%. These results are consistent with existing expectations, underscoring the leopard's vulnerability to road mortality. By providing a more precise prediction of extinction risk, our model can be a critical tool to guide road managers and decision makers on road planning and developing mitigation measures.

Wildlife roadkill databases often compile valuable information regarding species, timing, and locations of roadkill incidents, with some databases also including details such as gender and age classes. As part of the “LIFE LINES - Linear Infrastructure Networks with Ecological Solutions” project, we aggregated roadkill data from various sources, including road operators, traffic police, and academic institutions, into the Portuguese National Roadkill Database. This database encompasses over 130,000 records spanning most terrestrial vertebrates found in the country, ranging from small (< 1 kg) to medium (1-10 kg) and large (> 10 kg) species.
Different sources contribute varying data, with academia accounting for all species and sizes, road operators primarily documenting medium and large vertebrates, and traffic police focusing mostly on large vertebrates. Notably, different data sources yield distinct roadkill hotspot patterns. By analysing a specific subset of the database consisting of daily monitored roadkill incidents, we investigated how different monitoring intervals affect the frequency and location of hotspots across various vertebrate classes, orders, species, and size categories. Our findings reveal that for smaller species like bats and amphibians, sampling intervals exceeding one day result in contrasting patterns compared to daily sampling, while for most medium and larger species, such discrepancies occur primarily with sampling intervals exceeding one week. We discuss the implications of these results for the development of effective roadkill monitoring and mitigation initiatives.

Animal Vehicle Collisions (AVCs) are identified as a major cause of mortality in wild animal populations, in different parts of the world. This topic has become an important issue in road ecology, due to increasing human and economic consequences (namely, road safety) as well as a pervasive effect on wildlife populations.
Due to this relevance, we developed a survey methodology, in used since 2015, that allowed for the daily monitoring of our roads, that implies the use of a mobile app which allows a large data collection, with direct upload to Ascendi server and real time access to the recollected data.
Wildlife road mortality tends to aggregate spatially at locations commonly referred as road mortality hotspots. Mortality hotspots are frequently used to identify and prioritise locations for implementing mitigation, representing the basis of the biodiversity action plan developed by Ascendi.
Hotspot identification depends on predictive models and their fitting to our road type and environment. KDE+, presented by Bíl and colleagues, is centred on the principles of Kernel Density Estimation. In this case study, we used the KDE Toolbox in our Geographic Information System to integrate the KDE+ method to Ascendi's Network that encompasse fenced highways north to south of Portugal. We estimate the most important clusters for several species, and with a case by case study we demonstrate the variation of hotspots for each, according with the seasonal biological cycle. This allows for better understanding of the ecological variables at play as well as the planning of seasonal mitigation actions

Roadkill constitutes one of the manifestations of human-animal conflict insufficiently studied on island environments. We use the Mediterranean island of Cyprus as a case study to understand its causes and effects. This research aims to reveal the species most affected and use temporal and spatial covariates to understand when and where roadkill is most likely to occur based on road characteristics (e.g. speed limit, surrounding land use, distant from freshwater bodies).

Roadkill data were recorded by citizens between 2013 and 2023 using a structural protocol developed by the Cyprus Roadkill Observation System (CyROS). Spatial roadkill rates (indicating roadkill hotspots) were quantified as the number of individuals killed per 500m road segment. The volume of available information was analysed using ArcGIS and R statistics, while Generalised Additive Models (GAMs) were used to identify the drivers of roadkill occurrence throughout the island’ s road network (∼19,500 km).

Statistical analysis revealed relations between roadkill rates and road characteristics. In absence of large wild animals (besides mouflons), the most affected groups were mammals (hedgehogs and foxes), followed by reptiles (snakes) and birds (owls). A clear increase in incidents can be correlated with the rising temperatures of spring and summer. Highways and high speed roads (>50 Km/h) are responsible for the majority of fox roadkill, while other groups (reptiles and birds) are largely observed in roads with lower speed limits.

Identifying the different impacts of roadkill, on biodiversity can facilitate mitigation and conservation actions along roads and provide a planning model for other Mediterranean islands.

To establish long-term monitoring of wildlife-vehicle collisions (WVC) in Taiwan and identify roadkill hotspots, estimate roadkill density and species most threatened by vehicle collisions, we implement a systematic roadkill monitoring system utilizing citizen science platforms. This allows us to track changes in WVC patterns over time, enabling adaptive management strategies. The systematic survey program initiated in July 2017. To ensure systematic sampling, Taiwan has been divided into 1,440 5x5km grids across seven eco-regions, three levels of road density, and four categories of road types. From these, a total of 420 sampling grids were selected using stratified random sampling methods with proportional allocations from each eco-region and road density level. Transects were established along four kind of road types and must be longer than 3km. Volunteers were recruited to conduct fixed-sample-area surveys in January, April, July, and October each year. By the end of 2023, an average of 196±13 volunteers participated in the surveys each quarter, covering a total road length of 1,690±108 km across 243±10 sample areas, representing approximately 4% of Taiwan's total road length. County roads and country roads exhibited the highest roadkill density, with a cumulative record of 27,949 roadkill individuals. The results indicate that autumn is the peak season for roadkill incidents in Taiwan, with amphibians being the most affected taxonomic group (58%), followed by reptiles (23%), birds (11%), and mammals (8%). It is estimated from this systematic roadkill survey that about 15 to 20 million animal roadkill occur annually in Taiwan.

Wildlife-vehicle collisions have become increasingly common worldwide, being one of the threats to wildlife. With the knowledge of accidents related to animals, more effective and better targeted mitigation can be designed to reduce the impact on both animals and traffic. We aimed to analyse temporal and spatial patterns of wildlife-vehicle collisions registered in 2017 in Poland, based on data from the General Directorate for National Roads and Motorways (GDDKiA). We analysed over 23,900 collisions related to wildlife, including roe deer (5,902), red deer (318), wild boar (2003), moose (245), European badgers (1,327), red foxes (4,979), martens (794), feral dogs (3,161) and cats (5,151). We investigated whether the number of collisions changed over months and seasons. The most of accidents related to roe deer were registered in May, with red deer and moose in September, and with wild boar in November. The most of accidents related to badger were registered in August, with fox in Autumn (without clear peak), martens in May and June. Cats were mostly involved in collision in March and August, while dogs in December. According to spatial patterns of wildlife-vehicle collisions, we compared vehicle traffic (number of vehicles/day) and maximum allowed speed, as well as land-use and habitat characteristics (i.e. share of forests, open areas, build-up areas) at collision and random sites distributed along the roads network. Our results showed that habitat variables influenced the number of wildlife-vehicle collisions, while maximum allowed speed and traffic on the roads affected accidents in a non-linear way.

Roadkills are significant negative factor and threat for a lot of species globally addressing the need of maintenance and management of the transport networks in the most biodiversity friendly way. Monitoring of roadkill is essential to define the spatial distribution of the hot spots and how the ecological corridors and local routine paths for wildlife intersect the road networks. In the Region of Central Macedonia in Greece, a roadkill monitoring project implemented around the lakes Koroneia and Volvi under the responsibility of the Management Unit of Protected Areas of Central Macedonia of the Natural Environment & Climate Change Agency during 2021 and 2022.
The initial aim of the monitoring was to record the protected species of the area (Canis lupus, Canis aureus, Felis sylvestris, Spermophilus citellus and Lutra lutra) and Capreolus capreolus but the final survey included the fox as the most common roadkill victim and other species such as smaller mammals and reptiles. Additionally, several culverts and bridges as well as specific interesting spots on wildlife paths were monitored using trap-cameras while the traffic load measurement was recorded in 12 months base.
The results show 26 roadkills of the protected species, 30 foxes and 137 other species as small mammals and reptiles. The analysis of the results followed by the mapping of the critical hot spots and the proposal of the appropriate measures related with the improvement of the existing culverts in order to serve wildlife as crossing structures, the vegetation management and the wildlife collision preventive measures.