How to attenuate the barrier effect of linear infrastructures A method for prioritizing existing crossings to improve wildlife regional connectivity

Linear transportation infrastructures traverse and separate wildlife populations, potentially leading to their decline at local and regional scales. Wildlife crossings are built to mitigate this barrier effect. Their location should be considered before road building begins so as to limit the impact on connectivity. But in practice, biodiversity seldom takes precedence over economics and wildlife crossings are often developed as an afterthought for already existing infrastructures. In such cases, re-designing existing crossings to improve their permeability to wildlife is often the cheaper and preferred option. Limiting the economic costs also necessitate to find the best crossing location optimizing the connectivity for the larger number of species.

We propose to set up a standardized protocol using graph theory for prioritizing these existing crossings so as to improve the connectivity of a set of species with varying degrees of mobility and living in different habitats. The study area is located in the Gresivaudan valley in the French Alps. The method was based on five successive steps: (1) defining the virtual-species groups, (2) land-cover mapping, (3) constructing graphs to model the ecological networks of virtual species, (4) prioritizing wildlife crossing locations depending on the connectivity gains they provide, and (5) combining the results in a multispecies diagnosis. To integrate the nesting of ecological processes and explore connectivity at regional scale and over decades, we proposed to construct graphs based exclusively on habitat areas that were sufficiently large and/or interconnected to support a viable population over time. The prioritization method consists in computing a global index quantifying the initial connectivity of the network and then evaluating the potential contribution of a crossing along the highway corridor by quantifying the increase in the connectivity index provided by this crossing.

The results show that the connectivity gain varied greatly among the eight virtual species. Two species groups can be distinguished: the forest species that had most of the habitat area in the study region and the mountain and open habitat species. The connectivity of the first group increased from 4 to 10% with the re-designing of the best crossing. By contrast, the connectivity gain for the second group was small, under 1%. A set of several crossing locations on the eastern branch of the highway optimized the connectivity of the five forest virtual species. This area also held several locations improving connectivity for mountain species. Three crossing locations can be identified in this area as an interesting compromise for these two groups. The locations of the best crossings for the open habitat species were too different from the locations for the other species to reach a compromise.

This method could guide planners in identifying crossing locations to increase the connectivity of different species at regional scales over the long term.

Barrier effect, multispecies, connectivity, highway, dispersal, daily distance, wildlife crossing structures