Across the globe, roads fragment wildlife habitat and impede movement, and road density is increasing. Drivers in North America are probably familiar with deer or moose crossing signs, but vehicle collisions affect many species across the world. Among the victims are grizzly bears, elephants, and monarch butterflies.
Wildlife crossings for roadkill mitigation are receiving increasing conservation attention. Due to limited resources, however, careful site selection for these measures is crucial. Roadkill risk models seek to predict where, when, and why vehicle collisions are most likely to occur.
Valerio, F., Godinho, S., Salgueiro, P., Medinas, D., Manghi, G., Mira, A., Pedroso, N.M., Ferreira, E.M., Craveiro, J., Costa, P. and Santos, S.M. 2023. Integrating remote sensing data on habitat suitability and functional connectivity to inform multitaxa roadkill mitigation plans. Landscape Ecology. DOI: 10.1007/s10980-022-01587-6.
In an effort to broadly mitigate risk, many models take a species-agnostic approach, studying connectivity without a focal species. However, there is debate about whether a species-agnostic approach to roadkill mitigation leaves some animals behind.
A new study comes at roadkill risk from all directions– using high resolution data, a long-term study period, and multiple taxa. Researchers used fifteen years of animal occurrence data in Southern Portugal from linear transects, camera traps, point counts, and roost surveys. The data included nineteen animals from three taxa: land mammals, birds, and bats.
The study also used remote sensing data to map ecosystem features. They compared these data with species distribution models to identify the best predictors of roadkill risk.
Metrics which described vegetation characteristics like biomass and senescence were the most useful in predicting risk for non-flying mammals. Radar data, which penetrated vegetation to describe vertical structure, were the best predictor for birds. Thermal metrics proved to be the best predictors for bat occurrence. This fusion approach – mixing species distribution modeling with remote sensing data – helped identify patterns in productivity that explained species occurrence, like animals avoiding drought and heat, seeking habitat with adequate water resources.
The study also identified road planning units with high risk and mortality scores across all taxonomic groups, which could be prioritized for mitigation structures. Mortality prediction was generally similar for birds and mammals, both between and within those groups. However, bats had more variation within the taxa (mostly between specialist and generalist species) and different areas of risk than the other two groups. Though the study recommends prioritizing locations with the highest overall risk due to limited resources, bats might benefit from their own infrastructure, if possible.
While species distribution modeling and remote sensing analysis are independently useful to predict roadkill risk, using them in combination, as in this study, provides a better understanding of the ecological factors underlying animal movement. Doing so also reveals where a one-size-fits-all approach underserves some animals.
Methods like those used in this study could be useful in environmental impact assessment, and might be applied to other types of infrastructure (and other taxa) as well. Considering the many dimensions of animal movement and habitat will help planners make the most of limited resources to reduce collision deaths.
Related Posts
Appalachian Mountains: policies to keep iconic species connected
The Appalachian Mountains are a connected landscape that supports both wildlife and communities, and recent policies can help keep it connected for safer roads, healthier ecosystems, and a more resilient future.
U.S. Policy Preview: Wildlife Road Crossings Program Reauthorization Act
Three bipartisan bills aim to make the Wildlife Crossings Pilot Program permanent and reauthorize funding through 2031.
Connectivity influences wildlife-vehicle collision hotspots
A new study examines how habitat connectivity, land cover, and traffic patterns influence wildlife-vehicle collision hotspots in New Hampshire, revealing that more connected areas may actually experience fewer collisions depending on species and landscape context.
