Sample size considerations for satellite telemetry and animal distributions

• Published in: The Journal of Wildlife Management
• Main Authors: Roberts, Anthony, Silverman, Emily, Gifford, Suzanne
• Format: Article in Journal/Newspaper
• Language: English
• Published: Wiley 2018
• Subjects: Arctic
• Online Access: http://dx.doi.org/10.1002/jwmg.21504; https://api.wiley.com/onlinelibrary/tdm/v1/articles/10.1002%2Fjwmg.21504; http://onlinelibrary.wiley.com/wol1/doi/10.1002/jwmg.21504/fullpdf

ABSTRACT Satellite telemetry is a powerful tool used to follow animals through their annual life cycle, informing the understanding of behavior and distribution of many species. Because boreal‐ and arctic‐nesting North American sea duck populations are challenging to survey, satellite telemetry is important for describing breeding distributions and identifying breeding population structure. Accurate knowledge of breeding distributions is needed for effective habitat and harvest management, but satellite telemetry is expensive so it is important to consider the effort necessary to accurately map breeding distributions. We construct 3 theoretical breeding distributions using existing telemetry data from 3 species of sea ducks, Barrow's goldeneye ( Bucephala islandica ), surf scoter ( Melanitta perspicillata ), and black scoter ( M. americana ), by fitting kernel densities to approximate breeding locations of individual birds. Then we determined the minimum sample size needed to approximate these theoretical breeding distributions by assessing the overlap between breeding densities simulated by sampling from the distributions and the theoretical breeding densities. Diminishing information gains with additional effort (i.e., <1–5% improvement in prediction) were reached with sample sizes ranging from 80 to 130. Sea duck mortality, transmitter failure, and exclusion of non‐breeding individuals resulted in an effective sample size smaller than the number of birds originally marked. For the cases we considered, obtaining breeding locations for 80–130 individuals would require marking 11–41% more birds than the sample size goal. Thus, although satellite telemetry provides valuable information on sea duck populations, our analysis suggests that accurately estimating the extent and relative use of breeding habitats requires substantial investment. © 2018 The Wildlife Society.