What factors drive wildlife-vehicle collisions on highways? A case study from Western Nepal

Abstract

Wildlife-vehicle collisions (WVCs hereafter) have become a significant global challenge for biodiversity conservation and management. Understanding factors that drive collisions is crucial for devising management policies and plans. Field surveys were conducted in 2019 to collect data on the physical and ecological features of the study area in each 875 m segment, and species death records were collected from government agencies between 2015 and 2019. This study identified the most important factors that influence WVCs, and we predicted the WVC hotspot along the East-West highway of Banke and Bardia National Parks in western Nepal using the kernel density estimation function in Arc GIS 10.5. We performed Binomial regression analysis to identify the most important factors associated with WVCs. Results showed that geographic factors such as topography, presence of retaining walls, canals/drainage, land use or land cover, forest land, agricultural land, grassland, and barren land have negative influence on WVCs, whereas speed breaker and aspect exhibit a positive influence. We found that WVCs are more likely to occur near forest land than agricultural land, grasslands, and barren land. We also found that the presence of the retaining wall, cannel/drainage, and speed breakers increased the likelihood of WVCs events. Mammals like spotted deer, Axis axis (27 %) and wild boars, Sus scrofa (21 %) are the most frequently affected species in wildlife-vehicle collisions. Due to the unavailability of data from the park office, this study does not include species like insects, reptiles, amphibians, birds, and small mammals. Animal-friendly infrastructures like wildlife crossings (overpass/underpass), canopy bridges, conservation signage and reflectors near the hotspot can reduce WVCs. In addition, alternative roads outside the park area can be constructed to divert vehicle movement and reduce collisions in the national park area.