Search for a command to run...
Corridors are popular conservation tools because they are thought to allow animals to safely move between habitat fragments, thereby maintaining landscape connectivity. Nonetheless, few studies show that mammals actually use corridors as predicted. Further, the assumptions underlying corridor models are rarely validated with field data. We categorized corridor use as a behavior, to identify animal-defined corridors, using movement data from fishers (Martes pennanti) tracked near Albany, New York, USA. We then used least-cost path analysis and circuit theory to predict fisher corridors and validated the performance of all three corridor models with data from camera traps. Six of eight fishers tracked used corridors to connect the forest patches that constitute their home ranges, however the locations of these corridors were not well predicted by the two cost-based models, which together identified only 5 of the 23 used corridors. Further, camera trap data suggest the cost-based corridor models performed poorly, often detecting fewer fishers and mammals than nearby habitat cores, whereas camera traps within animal-defined corridors recorded more passes made by fishers, carnivores, and all other non-target mammal groups. Our results suggest that (1) fishers use corridors to connect disjunct habitat fragments, (2) animal movement data can be used to identify corridors at local scales, (3) camera traps are useful tools for testing corridor model predictions, and (4) that corridor models can be improved by incorporating animal behavior data. Given the conservation importance and monetary costs of corridors, improving and validating corridor model predictions is vital.