Empirical selection between least‐cost and current‐flow designs for establishing wildlife corridors in Gabon.

Corridors are intended to increase species survival by abating landscape fragmentation resulting from the conversion of natural habitats into human‐dominated matrices. Conservation scientists often rely on 1 type of corridor model, typically the least‐cost model or current‐flow model, to construct a linkage design, and their choice is not usually based on theory or empirical evidence. We developed a method to empirically confirm whether corridors produced by these 2 models are used by target species under current landscape conditions. We applied this method in the Gamba landscape between 2 national parks in southwestern Gabon. We collected signs of presence of African forest elephant (Loxodonta cyclotis), forest buffalo (Syncerus caffer nanus), and 2 apes, western lowland gorilla (Gorilla gorilla gorilla) and central chimpanzee (Pan troglodytes troglodytes), on transects. We used patch‐occupancy models to identify least‐cost and current‐flow corridors for these 4 species. On average, 28.7% of current‐flow corridors overlapped with least‐cost corridors, confirming that the choice of corridor model can affect the location of the resulting linkage design. We validated these corridors by monitoring signs and examining camera detections on new transects within and outside modeled corridors. Current‐flow corridors performed better than least‐cost corridors for elephants, whereas the opposite was found for buffalo and apes. Locations of the highest priority corridors for the 3 taxa did not overlap, and only 18.3% of their combined surface was common among 2 species. We used centrality metrics to calculate the average contribution of corridor pixels to landscape connectivity and derived an index that can be used to prioritize corridors. As a result, we recommend protecting at least 17.4% of the land surface area around Gamba town to preserve the preferred travel routes of the target species. Article impact statement: Choice of corridor model affects corridor locations and needs empirical validation before linkage design is approved. [ABSTRACT FROM AUTHOR]