Influence of study design and movement behavior on performance of open population spatial Cormack-Jolly-Seber models: application to acoustic telemetry technology

Abstract

Spatial Cormack-Jolly-Seber (CJS) models are appealing for analyzing mark-recapture data from electronic tagging studies with fixed receiver arrays due to their ability to accommodate detections on multiple receivers for estimating apparent survival, detection probabilities, individual activity centers, and home range sizes. However, spatial CJS models assume movement and detection processes that may not be met by electronic tagging studies, especially those involving acoustic telemetry technology. To evaluate the sensitivity of spatial CJS models to underlying assumptions, we simulated acoustic telemetry detection histories in a riverine system using two movement processes, one based on movement around activity centers and another using stepwise movement with detections based on an individual’s proximity to a receiver during tag transmission. For each movement process, we evaluated a range of life history and movement parameters and four receiver spacings (configurations) to determine how study design influenced model performance. Simulated detection histories were assessed with spatial CJS models to determine how well model estimates matched known values. When the movement process matched model assumptions, the model performed well across investigated scenarios. However, under a stepwise movement process, convergence was low for all parameters, and variability between activity centers was positively biased. When individuals were undetected for several time steps, activity center estimates tended to drift to gaps in receiver arrays, which could lead to inaccurate conclusions about space use. Decreasing receiver spacing reduced drifting of activity center estimates; therefore, when a stepwise movement behavior is expected, decreasing receiver spacing may improve reliability of activity center estimates.