• Proceedings of the National Institute of Ecology of the Republic of Korea
• /
• v.2 no.4
• /
• pp.274-278
• /
• 2021

Road structures play an important role in collisions involving vehicles and wildlife. Our study aimed to determine the effect of various types of road structures on the risk associated with roadkill. We surveyed 50 previously identified roadkill hotspots, ranked from one to five according to roadkill density. We collected nine types of road structure data on each hotspot road section. Structures with similar characteristics were grouped together, resulting in five categories, namely, median barrier, high edge barrier, low edge barrier, speed, and visibility. We examined the existence of each road structure category at each hotspot rank. The cumulative link model showed that the absence of bottom blocked median barrier increased the roadkill hotspot rank. Our study concluded that a visual obstacle in the middle of roads by the median barrier decreases wildlife road crossing attempts and roadkill risk. We suggest that future roadkill mitigation plans should be established considering these characteristics.

• Journal of Environmental Impact Assessment
• /
• v.28 no.6
• /
• pp.580-591
• /
• 2019

This study analyzed roadkill hotspots in Yeongju, Mungyeong-si Andong-si and Cheongsong-gun to compare the method of searching the area of the spatial cluster for selecting the roadkill hotspots. The local spatial autocorrelation index Getis-Ord Gi* statistics were calculated by different units of analysis, drawing hotspot areas of 9% from 300 m and 14% from 1 km on the basis of the total road area. The rating of Z-score in the 1km hotspot area showed the highest Z-score in the 28th National Road section on the border between Yecheon-gun and Yeongj-si. The kernel density method performed general kernel density estimation and network kernel density estimation analysis, both of which made it easier to visualize roadkill hotspots than district unit analysis, but there were limitations that it was difficult to determine statistically significant priority. As a result, local hotspot areas were found to be different according to the cluster analysis method, and areas that are in common need of reduction measures were found to be the hotspot of 28th National Road through Yeongju-si and Yecheon-gun. It is deemed that the results of this study can be used as basic data when identifying roadkill hotspots and establishing measures to reduce roadkill.

• Korean Journal of Environment and Ecology
• /
• v.36 no.5
• /
• pp.460-467
• /
• 2022

In this study, we collected snake roadkill data from 2006 to 2017 and developed a species distribution model to identify the pattern of snake roadkill and predict the potential hotspot of snake roadkill in the Odaesan National Park of South Korea. During the study period, snake roadkills occurred most frequently on the road, which passes through between forest and stream at an altitude of about 600 m. The modeling result showed that the occurrence probability of snake roadkill was high on a road with a gentle slope at a distance of 25 m from the stream and an altitude of 600 m. The most susceptible regions for snake roadkill in the Odaesan National Park were located on National Route 6, about 2.2 km and 11.7 km away from the southern border of the park, and on Local Road 446, 3.44 km away from the southern border of the park. The results of this study suggest that providing alternative basking places and eco-corridors and installing protection fences that block the inflow of snakes into roads, preferentially around roads and streams at an altitude lower than 700 m would be an effective way of reducing snake roadkill in the Odaesan National Park.

• Journal of the Korean Society of Environmental Restoration Technology
• /
• v.26 no.5
• /
• pp.19-32
• /
• 2023

The fragmentation of habitats resulting from human activities leads to the isolation of wildlife and it also causes wildlife-vehicle collisions (i.e. Road-kill). In that sense, it is important to predict potential habitats of specific wildlife that causes wildlife-vehicle collisions by considering geographic, environmental and transportation variables. Road-kill, especially by large mammals, threatens human safety as well as financial losses. Therefore, we conducted this study on roe deer (Capreolus pygargus tianschanicus), a large mammal that causes frequently Road-kill in Jeju Island. So, to predict potential wildlife habitats by considering geographic, environmental, and transportation variables for a specific species this study was conducted to identify high-priority restoration sites with both characteristics of potential habitats and road-kill hotspot. we identified high-priority restoration sites that is likely to be potential habitats, and also identified the known location of a Road-kill records. For this purpose, first, we defined the environmental variables and collect the occurrence records of roe deer. After that, the potential habitat map was generated by using Random Forest model. Second, to analyze roadkill hotspots, a kernel density estimation was used to generate a hotspot map. Third, to define high-priority restoration sites, each map was normalized and overlaid. As a result, three northern regions roads and two southern regions roads of Jeju Island were defined as high-priority restoration sites. Regarding Random Forest modeling, in the case of environmental variables, The importace was found to be a lot in the order of distance from the Oreum, elevation, distance from forest edge(outside) and distance from waterbody. The AUC(Area under the curve) value, which means discrimination capacity, was found to be 0.973 and support the statistical accuracy of prediction result. As a result of predicting the habitat of C. pygargus, it was found to be mainly distributed in forests, agricultural lands, and grasslands, indicating that it supported the results of previous studies.