If you build it, will it work?

Following the recent article, Evaluating the success of wildlife crossing structures using genetic approaches and an experimental design: Lessons from a gliding mammal by Kylie Soanes et al., Associate Editor, Yolanda Wiersma explores the world of wildlife crossing structures.

Large-scale restoration projects represent human optimism in the face of anthropogenic change. In response to the negative effects of human activities on habitat loss and fragmentation, humans have engineered some impressive structures to attempt to restore habitat connectivity. Aerial images of some of the larger highway crossing structures that have been built to facilitate wildlife movement are impressive. Less dramatic, but still important, are smaller crossing structures, such as toad underpasses, modified culverts for fish passage or overhead canopy bridges or glider poles for volant mammals. Installation of all such structures carries a cost, both in terms of research to identify how to optimally design and locate them, and to construct them. Costs can range up to $5 million for the more elaborate highway overpasses. But is this money well spent for conservation? It seems like evaluating whether a crossing structure is worth the cost should be pretty straightforward: if animals start using it, then it works. The reality is that evaluating success of efforts to restore connectivity is not quite so straightforward.

In a recent paper, Soanes et al. evaluated the success of wildlife crossing structures (glider poles and canopy bridges) for a squirrel glider (Petaurus norfolcensis) in eastern Australia. They used radio tracking and genetic measures of connectivity of squirrel gliders caught on both sides of a 4-lane freeway (with over 10,000 cars/day) to assess the success of the installation of these crossing structures at four sites. They used three controls for comparison – a section of unmitigated freeway with natural canopy, a section of unmitigated freeway with no canopy connectivity, and a non-freeway site. Their experimental design fits the criteria of a BACI, with the Before-After representing before-after installation of the crossing structures and control-impact representing the comparison to sections of freeway without crossing structures, but with and without potential natural crossing via narrow canopy gaps. Interestingly, they found that the freeway did not pose as much of a connectivity barrier as they had previously thought; the spatial genetic structure of the squirrel gliders showed that individuals on both sides of the presumed barrier belonged to the same genetic cluster or had mixed membership in both genetic clusters on either side of the barrier at all but one site. This ran counter to the radio tracking data which showed a very low rate of road-crossing. The installation of the crossing structures did show a positive response at the one site where genetic structure consistent with the freeway as a barrier was observed. There was also an increase in cross-freeway parent pairs (i.e., mother and father of an individual came from opposite sides of the freeway) after the installation of the crossing structures.

The analysis by Soanes et al. illustrates the challenge of carrying out robust evaluations of restoration actions. Had their data been restricted to radio tracking data, they would have concluded that the freeway was a significant barrier to movement and the impact of the crossing structures might have appeared more dramatic with increased movements of individuals on the structures. The addition of genetic data, however, showed that the freeway might not have been as large a barrier as previously thought, thus lessening the impact of the crossing structures. This effect might not have been detected without a full BACI (Before-After-Control-Impact) analysis, something that is often difficult to carry out in restoration evaluations. Many wildlife crossing structures are evaluated on “if you built it- did they come?” principle, and if animals are detected using the crossing (either directly or indirectly via track plates or camera traps), the crossing structure is deemed to have been a success.

Fish crossing structures are the aquatic analogue to highway crossing, and are engineered culverts or ladders that help fish move upstream across movement barriers created by dams and culverts that alter flow regimes. Many of the post-restoration evaluations of culverts compare fish movement in a restored culvert to a reference area of a river or stream and/or to an unmitigated culvert (e.g., Favaro et al. 2014) which represents a Control-Impact design. Others compare fish movement before and after a culvert is restored (e.g., Evans et al. 2015; Myers and Nieraeth 2016), which represents a Before-After design. Inferences of restoration success with a B-A or C-I design will be different from a full BACI. Had Soanes et al. only looked at movement and genetic data before and after the installation of the crossing structures, they would have seen some change, but they would have missed the inference that the genetic structure of the squirrel glider populations on either side of the freeway was not dissimilar from sections of highway where structures were absent. Thus, they would have missed the ability to realize that the freeway was not a complete barrier in the absence of crossing structures. Thus, crossing structures may be an expensive solution that is not really needed.

All of this is not to suggest that we avoid building crossing structures to mitigate human-caused habitat fragmentation. However, studies such as the one by Soanes et al. suggest that wildlife managers plan their approach carefully to be as cost-effective as possible and to rigorously evaluate the impact of mitigation measures. A full cost-accounting should include the cost of not doing anything, which in the case of large wildlife includes the cost of wildlife-vehicle collisions (see Huijser et al. 2009) for an example.

The article, Evaluating the success of wildlife crossing structures using genetic approaches and an experimental design: Lessons from a gliding mammal by Soanes et al. is available in Journal of Applied Ecology.