Individual-level variation and higher-level interpretations of space use in wide-ranging species: An albatross case study of sampling effects

Abstract

Marine ecologists and managers need to know the spatial extent of at-sea areas most frequented by the groups of wildlife they study or manage. Defining group-specific ranges and distributions (i.e., space use at the level of species, population, age-class, etc.) can help to identify the source or severity of common or distinct threats among different at-risk groups. In biologging studies, this is accomplished by estimating the space use of a group based on a sample of tracked individuals. A major assumption of these studies is consistency in individual movements among members of a group. The implications of scaling up individual-level tracking data to infer higher-level spatial patterns for groups (i.e., size and extent of areas used, overlap or segregation among groups) is not well documented for wide-ranging pelagic species with high potential for individual variation in space use. We present a case study exploring the effects of sampling (i.e., number and identity of individuals contributing to an analysis) on defining group-specific space use with year-round multi-colony tracking data from two highly vagile species, Laysan (Phoebastria immutabilis) and black-footed (P. nigripes) albatrosses. The results clearly demonstrate that caution is warranted when defining space use for a specific species-colony-period group based on datasets of small, intermediate, or relatively large sample sizes (ranging from n = 3-42 tracked individuals) due to a high degree of individual-level variation in movements. Overall, we provide further support to the recommendation that biologging studies aiming to define higher-level patterns in space use exercise restraint in the scope of inference, particularly when pooled Kernel Density Estimation (KDE) techniques are applied to small datasets for wide-ranging species. Transparent reporting in respect to the potential limitations of the data can in turn better inform both biological interpretations and science-based management decisions.