Roads represent a significant threat to biodiversity. If transportation managers are to reduce the ecological effects of roads, they need large-scale data identifying where species are most likely to occur (commonly known as ‘hotspots’). In this project, ecologists and road-managers developed a comprehensive approach to identify hotspots for 10 species of amphibians and reptiles on roads in New York State, and prioritize deployment of mitigation efforts. We initially synthesized available literature to predict patterns of habitat use, with a geographic information system then used to develop spatially explicit models that integrated habitat data at both the local and regional population level. We developed two approaches for prioritizing model output: (1) overlaying the arterial classification code (a measure of traffic intensity) over model outputs, and (2) using the contiguous length of road within specified high occurrence index values. Models were evaluated using field data derived from road surveys.

Results/Conclusions

Our models showed clear differences in the predicted occurrence on roads among habitat specialists and generalists, and between life-history stages. Wide-ranging habitat generalists were predicted to have at least some probability of occurrence on most roads. Conversely, species with limited movement ranges and specific aquatic and terrestrial habitat had more limited distributions. Validation data indicated that the models were effective for predicting occurrence of species with specialized habitat requirements, but that predictions for wide-ranging generalists were less accurate. These data also demonstrated that length of continuous hotspot and traffic intensity are effective for prioritizing the deployment of mitigation for habitat specialists. Our modeling approach is a useful tool for identifying road-hotspots for herpetofauna, allowing predictions to be made over large spatial extents, and with readily available data sources. Effective mitigation for movement-limited habitat specialists may include spatially and/or temporally targeted approaches, such as road-underpasses or temporary signage, whereas for widespread generalist species broader-scale approaches such as driver education are likely to be more effective.