Spatial and temporal patterns of mammalian roadkill across subtropical protected areas in Nepal

Abstract

Wildlife–vehicle collisions are a major source of anthropogenic wildlife mortality, making it essential to understand their spatial and temporal patterns for effective mitigation. In this study, we conducted a multisite, multiyear analysis of wildlife roadkill incidents recorded across six subtropical protected areas in Nepal over a five-year period (2017–2022). A total of 618 individual animals from 37 vertebrate species were killed in vehicle collisions, averaging five roadkills per 10 km annually. Roadkill accounted for approximately 27% of all recorded wildlife mortalities, highlighting it as a significant source of mortality and a growing threat to wildlife in protected areas. We found a clear taxonomic bias in roadkill, with ungulates and primates accounting for ~65% of all incidents. However, during the COVID-19 lockdown, carnivores were disproportionately affected, suggesting a potential shift in risk linked to altered traffic conditions. Multiscale regression modeling showed that roadkill risk increased with Normalized Difference Vegetation Index (NDVI), annual rainfall, and canopy cover, with consistent effects across scales of analysis (100-, 250-, and 500-m buffers). Seasonal differences were evident, with peak roadkill in winter, followed by autumn. Roadkill incidents increased modestly (3%) over the five-year period, but trends varied, with three of six protected areas showing an increasing trend. Importantly, the three key metrics (total magnitude, kill rate, and temporal trend) did not align; the protected area with the highest number of total roadkill did not exhibit the highest kill rate per km or increasing trend, but the one with the shortest road length passing through its area experienced an increasing trend. This mismatch likely reflects a complex interplay of factors, including road length and site-specific factors. This variation suggests that protected areas vary in their vulnerability to roadkill, underscoring the need for context-specific mitigation strategies tailored to each metric and site condition. As Nepal's road network continues to expand, targeted research and monitoring will be essential for identifying roadkill hotspots, assessing mitigation effectiveness, and informing conservation planning.