Most wildlife corridors, whether they are derived from simple least-cost path models or highly complex individual-based movement models, are run across a cost or resistance surface. Resistances represent an individual’s reluctance to cross a landscape feature or the physiological or survival costs they may incur.
Because of their reliance on resistance surfaces, the accuracy and effectiveness of corridor models likely depend on the accuracy of the resistance surface used. In turn, the accuracy of the resistance surface is based on decisions regarding the biological data type, the definition of the geospatial layers used to represent the landscape, and the modeling framework. Given that there have not been many comparisons among data types and modeling frameworks, researchers and conservation practitioners are often left to wonder about the best approach for their target species and study area.
To try to provide some clarity on this topic, we used 5-minute GPS collar data on pumas in southern California to look at the sensitivity of resistance surfaces and corridors to (1) data type (point, step, or path data), (2) how the landscape is defined, and (3) the resource selection function framework. Given the caveat that our analysis was on a single species/study area, our results indicate the following:
- Resistance surfaces and corridors are extremely sensitive to all three of these factors with both resistance estimates and corridor locations showing vast differences among data type, landscape definition, and analytical framework.
- Paths are preferred over steps and points for estimating resistance. However, if the GPS collar fix interval is too long to capture the true movement of the target species, points may produce less biased estimates of resistance.
- If points are used, it is preferable to identify points associated with movement events and use only those points in the analysis.
- Multiple scales should be examined with all of these data types and multi-scale models should be used to estimate resistance.
- Context-dependent models, where each used point, step, or path is paired with a biologically-relevant available area, are superior to context-independent models.
- The finest spatial grain of our geospatial layers (30m) produced the best performing models.
Zeller, K.A., K. McGarigal, P. Beier, S.A. Cushman, T.W. Vickers, W.M. Boyce. 2015. Using step and path selection functions for estimating resistance to movement: pumas as a case study. Landscape Ecology. DOI: 10.1007/s10980-015-0301-6
Zeller, K.A., K. McGarigal, P. Beier, S.A. Cushman, T.W. Vickers, W.M. Boyce. 2014. Sensitivity of landscape resistance estimates based on point selection functions to scale and behavioral state: pumas as a case study. Landscape Ecology 29: 541-557.
Zeller, K.A., K. McGarigal, S.A. Cushman, P. Beier, T.W. Vickers, W.M. Boyce. In prep. Sensitivity of resource selection and connectivity models to landscape definition.