Towards robust corridors: a validation framework to improve corridor modeling

Context: Ecological corridors are one of the most recommended ways to mitigate biodiversity loss. With growing recognition of corridor importance, corridor modeling lags others in the development of robust, quantitative validation methods. Objective: We propose a post-hoc corridor validation framework, considering the range of methods across data needs and statistical intensity. We demonstrate the importance of post-hoc corridor validation by testing several validation methods on different corridor model outputs. Methods: We used three different transformations on a Florida black bear (Ursus americanus floidanus) habitat suitability model to create different resistance grids, independent GPS collar data from a case study population, and Circuitscape to create corridor models. We used several validation methods, including a novel method, to compare resulting corridors. Results: Transformed resistance grids were all correlated, yet differing validation and resistance grids resulted in different recommended corridors. The use of one resistance surface and one validation type can result in the selection of inefficient or ineffective corridors. At a minimum, modelers should determine what proportion of an independent population falls within resulting corridors and should move towards more robust, documented methods as resources allow. The use of multiple validation methods can ensure greater confidence of modeling results. Conclusions: We encourage the use and further development of the framework presented here to drive the corridor modeling field towards more effective corridor creation and improved conservation outcomes. If validation methods are not improved, the ecological and economic cost of poor corridor science will continue to increase with increasing biodiversity loss.