Do wildlife crossing structures mitigate the barrier effect of roads on animal movement? A global assessment

The widespread impacts of roads on animal movement have led to the search for innovative mitigation tools. Wildlife crossing structures (tunnels or bridges) are a common approach; however, their effectiveness remains unclear beyond isolated case studies. We conduct an extensive literature review and synthesis to address the question: What is the evidence that wildlife crossing structures mitigate the barrier effect of roads on wildlife movement? In particular, we investigated whether wildlife crossing structures prevented an expected decline in cross-road movement, restored movement to pre-construction conditions, or improved movement relative to taking no action. In an analysis of 313 studies, only 14% evaluated whether wildlife crossing structures resulted in a change in animal movement across roads. We identified critical problems in existing studies, especially the lack of benchmarks (e.g. pre-road, pre-mitigation, or control data) and the use of biased comparisons. Wildlife crossing structures allowed cross-road movement in 98% of data sets and improved movement in ~60%. In contrast, the decline of wildlife movement was prevented in fewer than 40% of cases. For most structure types and species groups there was insufficient evidence to draw generalisable conclusions. Synthesis and Applications: The evidence to date suggests that wildlife crossing structures can mitigate the barrier effect of roads on wildlife movement, but in many cases have been poorly implemented or evaluated. The most supported measures were the addition of ledges and vegetation cover to increase movement for small mammals; underpasses to prevent the decline in movement of ungulates following road construction; and improving road-crossing for arboreal mammals using canopy bridges and vegetated medians. We strongly recommend that future use of crossing structures closely adheres to species-specific, best-practice guidelines to improve implementation and be paired with a thorough evaluation tha