Long-term data reveals increase in vehicle collisions of endangered birds in Hokkaido, Japan

Wildlife-vehicle collisions have significant consequences for both humans and animals, including injuries, deaths, and vehicle damage. Therefore, analysis of accident data is important for planning countermeasures and appropriate wildlife management. In this research field, roadkill incidents have been extensively studied in many taxa, while railway accidents have received less attention despite their obvious impact on wildlife. Here we applied a Bayesian state-space model to 31 years of collision data, both on railways and on roads, collected by the Ministry of the Environment in Hokkaido prefecture, Japan, to reveal the spatiotemporal dynamics of accidents for white-tailed eagles, Steller's sea eagles, and red-crowned cranes, for which over hundred accidents were reported in the data. Our analysis suggested that the mean annual number of individuals collected per collision site across Hokkaido increased 47,377-fold in the white-tailed eagle, 40,277-fold in the Steller's sea eagle, and 50,584-fold in the red-crowned crane between 1991 and 2021. There have been concerns about the impact of traffic accidents on the population dynamics of these endangered birds, but no formal analyses have been conducted. Our analysis showed numerically that the negative impact has been increasing annually. These results suggest that long-term data accumulation over large spatial scales allows us to understand the dynamics of accidents and predict potential factors underlying collision risks.