Multisite species-level roadkill across years reveals challenges and opportunities for monitoring and mitigation

Understanding spatiotemporal patterns of roadkill is challenging but essential for mitigation. Spatiotemporal roadkill patterns are seldom studied for single species at broad scales. We evaluated the spatiotemporal variation of roadkill of the water snake Helicops infrataeniatus (Squamata, Dipsadidae) in three Brazilian roads stretching 540 km and spanning up to 6 years of monitoring. We modelled the effect of sampling effort, evaluation scale and sample size on hotspot location over the years and evaluated hotspot persistence between roads at different evaluation scales. We also assessed the annual variability of roadkill hot moments within roads and tested if fatality was synchronous between roads. We correlated hotspots obtained with different evaluation scales and levels of sampling effort with those in a reference dataset built with fatality events from all sampling years. Coarser evaluation scales had a positive effect on correlation values, but the interaction between evaluation scale and sampling effort and the main effect of the latter differed between roads. Hotspot persistence over the years differed with road identity and increased with evaluation scales. Hot moments changed annually within roads and were concentrated in the warmest seasons of the year. Despite the ephemerality of hotspots within roads, fatality was synchronous between roads. Spatiotemporal variation of roadkills may reflect environmental, behavioural or populational changes at multiple scales. We suggest that both permanent and temporal mitigation strategies may be used according to spatiotemporal characteristics of hotspots. Our study exemplifies how synchrony studies can provide clues to understand roadkill patterns and improve mitigation planning at regional scales.

Automated summary

This study evaluated the spatiotemporal variation of roadkill of the water snake Helicops infrataeniatus in Brazil, finding that hotspots varied with evaluation scale and road identity, with annual changes in hot moments within roads. Despite hotspot persistence increasing with evaluation scales, fatality was synchronous between roads, suggesting spatiotemporal roadkill patterns may reflect changes at multiple scales.