A mechanistic approach to weighting edge-effects in landscape connectivity assessments

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.

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.

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.

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.

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.