Prioritizing road defragmentation using graph-based tools: Fid-Bau Portal

Prioritizing road defragmentation using graph-based tools

Ascensão, Fernando / Mestre, Frederico / Barbosa, A. Márcia

Highlights • Roads are one of the main causes of barrier effects for wildlife. • Mitigation planning needs to highlight the most problematic roads. • We used spatial graphs to prioritize roads for mitigation in the Iberian Peninsula. • We exemplify using native carnivore mammals in the Iberian Peninsula. • This approach can provide a first broad-scale evaluation of road networks.

Abstract Roads are a main cause of habitat fragmentation but mitigating the full road network is unfeasible. A key goal in the road mitigation planning process is to highlight, at the transportation network level, the most problematic roads, i.e. where mitigation measures are most required in order to maximize the benefits for biodiversity while keeping implementation costs as low as possible. Grounded on the concepts of habitat amount and accessible habitat, we prioritized roads for mitigation based on dual spatial graphs, where the land polygons delimited by roads are the nodes and the roads themselves are the links. The rationale was to identify those links (roads) that connect the nodes with higher potential biodiversity (as a proxy for quality habitat). We applied this approach to prioritize the defragmentation of the major road network of the Iberian Peninsula, targeting all native mammalian carnivores inhabiting this region. Our goal was to identify those roads that, by dividing areas with the best habitat quality and/or are major potential barriers for connectivity, should be prioritized in the mitigation process. We used two complementary metrics: Area Weighted Metric and the Integral Index of Connectivity. Highlighted roads bisect regions of high potential biodiversity for carnivores in northern Spain and along the Portugal-Spain border. Thirty-five roads were scored as high-priority by both metrics, suggesting that they have particular impact both in the amount of quality habitat and in overall landscape functional connectivity. This approach is completely scalable, allowing a fast assessment from local to continental scales.