A novel process to infer the reliability of ecological information derived from passive acoustic telemetry track reconstruction

Abstract

Abstract Passive acoustic telemetry can be used within cooperative networks to track migratory species over great distances at a relatively low cost. However, the non‐uniform distribution of fixed receivers within networks often results in sporadic detection data. Here, we propose a novel combination of methods to measure the reliability of hot spot analysis results derived from track reconstructions of passive telemetry data. We use an iterative process to simulate tracks of animals, derive detection data from these tracks, and reconstruct tracks from these derived data using a movement model. We then compare quadrat count residuals from the simulated and reconstructed tracks for different grid resolutions. The methodological framework is outlined in detail and tested on the acoustic telemetry cooperative arrays off the US East Coast. Our methods are applied to a subset of blacktip shark, Carcharhinus limbatus, acoustic telemetry detection data collected off the US East Coast. We then apply the resultant quadrat count to a hot spot analysis to determine the distribution of animals derived from these track reconstruction methods. We integrate the results of our methods process with the hot spot analysis results to determine the reliability of this distribution information. The track reconstruction methods performed well in coastal regions, from Palm Beach County, FL to Long Island, NY, minimized the clustering effect of high densities of receivers, and closed the gaps in some regions that were lacking receiver coverage. This performance was primarily affected by the presence/absence of receivers, and to a lesser extent by receiver density and water depth, depending on the grid resolution. Our method combination demonstrates a means by which passive telemetry data can be regularized to determine the spatial distribution of animals across regions with non‐uniform sampling coverage. These methods also allow the user to determine the reliability of animal distribution products in a telemetry array and the factors that contribute to high accuracy and precision. Our iterative process enables managers to infer the reliability of ecological results in decision‐making processes and could be leveraged for use as a gap analysis to develop a national strategy for telemetry assets.

Keywords

• continuous time correlated random walk
• hot spot analysis
• movement model
• statistical bias
• uncertainty