Using multiple metrics to estimate seasonal landscape connectivity for Blanding's turtles (Emydoidea blandingii) in a fragmented landscape

Landscapes and animal behavior can exhibit temporal variability and connectivity estimates should consider this phenomenon. In many species, timing of activities such as nesting, mate searching, and hibernation occurs during distinct periods in which movement events may differ, along with physical characteristics of the surrounding landscape. We estimate movement, landscape conductance, and patch importance for a turtle species across two seasonal activity periods (spring, late summer) in a fragmented agricultural region. Three connectivity approaches are compared to identify their advantages and disadvantages. A least-cost distance model, circuit-based approach, and patch-based index were used to collectively describe the potential functional connectivity of Blanding's turtle (Emydoidea blandingii) across a multi-temporal scale in an agricultural region of south western Ontario. Connectivity decreased further into the active season exhibited through lower conductance of the landscape and fewer pathways, while the importance of habitat nodes shifted due to temporal variability in the number and distribution of nodes. Models provided different yet complimentary information, with least-cost models overestimating discrete pathways yet providing a secondary measure of landscape barriers. The circuit-based model estimated corridors of least resistance providing an overall characterization of the landscape, while patch-based indices provided key information on the importance of individual habitat patches. Findings highlight the importance of including a temporal aspect in connectivity modelling as results demonstrate a change in functional connectivity over time. We also recommend employing multiple connectivity metrics to capture variation in movement behavior.