Event-Based Analysis of Wildlife-Vehicle Collisions Under Temperature Extremes and Weather Variability.

Wildlife-vehicle collisions (WVCs) represent an important ecological and road-safety problem, yet the influence of short-term meteorological variability on their occurrence remains insufficiently understood. This study investigated the associations between temperature extremes, abrupt temperature changes, and broader hydro-meteorological conditions and reported WVC frequency in Serbia over a 10-year period (2016-2025). A municipality-day dataset comprising 6281 police-reported WVCs was analyzed using an event-based methodology. Exploratory analyses were complemented by Poisson regression models to evaluate prolonged temperature episodes, abrupt temperature shocks, and broader hydro-meteorological conditions, while additional regional analyses examined the consistency of the observed associations across Serbia. The results demonstrated pronounced seasonal variation, with the highest reported WVC frequencies occurring during spring and late autumn. At the national level, prolonged cold episodes were associated with significantly lower reported WVC frequencies during their onset and middle phases, whereas heat episodes showed no significant associations. Abrupt cool-down shocks were associated with a short-term increase in reported WVC frequency on the day of the temperature decrease, while warm-up shocks showed no significant effects. Moderate and heavy precipitation, snow-day conditions, and prolonged dry spells were associated with reduced reported WVC frequency, whereas daily mean temperature and atmospheric pressure were not independently associated with reported WVC frequency. Regional analyses generally supported the national findings, although no regional associations remained statistically significant after false discovery rate correction. These findings demonstrate that short-term meteorological variability is associated with reported WVC frequency in a temporally dependent manner and highlight the importance of considering both environmental events and regional variability when investigating wildlife-vehicle collisions.