The ability of landscapes to impede species’ movement or gene flow may be quantified by resistance models. These models form the basis of many connectivity analyses such as designing linkage networks, predicting impacts of future landscape change, and siting mitigation projects. Because empirical data is often unavailable or difficult to acquire, many resistance models are parameterized by expert opinion. Importantly, there has been little exploration of how expert parameterization of resistance models affects their ability to predict rates of movement and gene flow. Additionally, resistance models may also vary in their spatial and/or thematic resolution as well as their focus on the ecology of a particular species or more generally on human landscape modifications. The impact of these resistance model design decisions is also poorly understood. In this report, we used empirical data collected from Greater Sage-Grouse (Centrocercus urophasianus) in Washington State, USA to evaluate the ability of species-specific expert opinion models (both fine and coarse resolution) and a landscape integrity model to predict rates of movement, gene flow, and lek persistence. We systematically varied parameterization of the species-specific expert models to potentially identify an alternative model with greater predictive power. We found that species-specific, fine-scale models performed better than coarse-scale or landscape integrity models, and alternative parameterizations markedly improved the fit of the expert models. Our study offers guidance that could improve predictive power of resistance models and the reliability of connectivity analyses that depend upon them.
This report, submitted by the Washington Connected Landscapes Project, is the Final Report for Greater Sage-Grouse model validation deliverables outlined in Agreement numbers 60181, 60181AG501, and F10AP0028 with the United States Fish and Wildlife Service.
