Kervellec, M. | bioRxiv Science

bioRxiv2023

Bringing circuit theory into spatial occupancy models to assess landscape connectivity

Connectivity shapes species distribution across fragmented landscapes. Assessing landscape resistance to dispersal is challenging because dispersal events are rare and difficult to detect especially for elusive species. To address these issues, spatial occupancy models have been developed to integrate the resistance surface concept of landscape ecology and model patch occupancy dynamics through colonization and extinction while accounting for imperfect species detection. However, the most recent approach is based on least-cost path distances which assume that individuals disperse along the optimal route. Here, we develop a new spatial occupancy model that incorporates commute distances derived from circuit theory to model dispersal across sites. Our approach allows for explicit estimation of landscape connectivity and direct measure of uncertainty from detection/non-detection data. To illustrate our approach, we study the recolonisation of two carnivores in France, and quantify the degree to which rivers facilitate Eurasian otter (Lutra lutra) dispersal and highways impede Eurasian lynx (Lynx lynx) recolonisation. Overall, spatial occupancy models provide a flexible framework to acccommodate any distance metric designed to align with species dispersal ecology. Open Research StatementData and code used in this research are available on Zenodo at https://zenodo.org/record/8376577

ecology↗

bioRxiv2022

Transnational non-invasive monitoring and spatial capture-recapture models reveal strong effects of humans on connectivity for the Pyrenean brown bear

Connectivity, in the sense of the persistence of movements between habitat patches, is key to maintain endangered populations and has to be evaluated in management plans. In practice, connectivity is difficult to quantify especially for rare and elusive species. Here, we use spatial capture-recapture (SCR) models with an ecological detection distance to identify barriers to movement. We focused on the transnational critically endangered Pyrenean brown bear (Ursus arctos) population, which is distributed over Spain, France and Andorra and is divided into two main cores areas following translocations. We integrate structured monitoring from camera traps and hair snags with opportunistic data gathered after depredation events. While structured monitoring focuses on areas of regular bear presence, the integration of opportunistic data allows us to obtain information in a wider range of habitat, which is especially important for ecological inference. By estimating a resistance parameter from encounter data, we show that the road network impedes movements, leading to smaller home ranges with increasing road density. Although the quantitative effect of roads is context-dependent (i.e. varying according to landscape configuration), our model predicts that a brown bear with a home range located in an area with relatively high road density (8.29km/km2) has a home range size reduced by 1.4-fold for males and 1.6-fold for females compared to a brown bear with a home range located in an area with low road density (1.38km/km2). When assessing connectivity, spatial capture-recapture modeling offers an alternative to the use of experts opinion when telemetry data are not available.

ecology↗