A mechanistic approach to weighting edge-effects in landscape connectivity assessments
Matthew Dennis, Jonny Huck, C. D. Holt, E. McHenry
Abstract
Abstract Context Understanding landscape functional connectivity is critical for nature conservation in fragmented landscapes. Spatially explicit graph-theoretical approaches to assessing landscape connectivity have provided a promising framework for capturing functional components driving connectivity at the landscape scale. However, existing weighting schemes used to parameterise functional connectivity in graph theory-based methods are limited with respect to their ability to capture patch-level characteristics relevant to habitat use such as edge-effects. Objectives We set out to develop a new approach to weighting habitat connectivity as a function of edge-effects exerted by non-habitat patches through better delineation of edge-interior habitat transitions at the patch-level and parameterization of intra-patch movement cost at the landscape scale. Methods We leverage the use of raster surfaces and area-weighted exponential kernels to operationalize a mechanistic approach to computing spatially explicit edge surfaces. We integrate map algebra, graph theory and landscape resistance methods to capture connectivity for a range of species specialisms on the edge-interior spectrum. We implement our method through a set of functions in the R statistical environment. Result Through a real-world case study, we demonstrate that our approach, drawing on these behaviours, outperforms competing metrics when evaluating potential functional connectivity in a typically fragmented agricultural landscape. We highlight options for the optimal parameterization of graph-theoretical models. Conclusion Our method offers increased flexibility, being tuneable for interior-edge habitat transitions. This therefore represents a key opportunity that can help to re-align the fields of landscape ecology and conservation biology by reconciling patch-versus-landscape methodological stances.