Cooperative Fish and Wildlife Research Units Program: New York
Education, Research and Technical Assistance for Managing Our Natural Resources

Sutherland, C., A. K. Fuller, and J.A. Royle. 2014. Modeling non-Euclidean movement and landscape connectivity in highly structured ecological networks. Methods in Ecology and Evolution 6: 169–177. doi: 10.1111/2041-210X.12316


Movement is influenced by landscape structure, configuration and geometry, but measuring distance as perceived by animals poses technical and logistical challenges. Instead, movement is typically measured using Euclidean distance, irrespective of location or landscape structure, or is based on arbitrary cost surfaces. A recently proposed extension of spatial capture-recapture (SCR) models resolves this issue using spatial encounter histories of individuals to calculate least-cost paths (ecological distance: Royle et al. 2013) thereby relaxing the Euclidean assumption. We evaluate the consequences of violating the Euclidean distance assumption when estimating abundance in highly structured landscapes, and demonstrate the value of this approach estimating biologically realistic space-use patterns and landscape connectivity. We simulated SCR data in a riparian habitat network, using the ecological distance model under an range of scenarios where space-use in and around the landscape was increasingly associated with water (i.e. increasingly less Euclidean). To assess the influence of miscalculating distance on estimates of population size, we compared the results from the ecological and Euclidean distance models. We then demonstrate that the ecological distance model can be used to estimate home range geometry when space use is not symmetrical. Finally, we provide a method for calculating landscape connectivity based on modelled species-landscape interactions generated from capture-recapture data. Using ecological distance always produced unbiased estimates of abundance. Explicitly modelling the strength of the species-landscape interaction provided a direct measure of landscape connectivity and better characterised true home range geometry. Abundance under the Euclidean distance model was increasingly (negatively) biased as movement was 30 more strongly associated with water and, because home ranges are assumed to be symmetrical, produced poor characterizations of home range geometry and no information about connectivity. The ecological distance SCR model uses spatially indexed capture-recapture data to estimate how activity patters are be influenced by landscape structure. The approach produces 35 unbiased estimates of abundance, but perhaps more importantly, provides biologically realistic representations of home range geometry and direct information about species-landscape interactions. The incorporation of both structural (landscape) and functional (movement) 38 components of connectivity provides a direct measure species-specific landscape connectivity.