Accuracy of geolocation estimates for flying seabirds

Abstract

Geolocation (Global Location Sensing or GLS logging) using archival light-recording tags offers considerable potential for tracking animal movements, yet few studies of flying seabirds have exploited this technology. Our study evaluated its effectiveness for determining foraging ranges of black-browed albatrosses Thalassarche melanophrys fitted simultaneously with GLS loggers and satellite-transmitters (Platform Terminal Transmitters, PTTs). After some preliminary validation, the position of an albatross could be determined by geolocation with a mean error +/- SD of 186 +/- 114 km (SDs of 1.66degrees and 1.82degrees of latitude and longitude, respectively). Errors from identical static loggers were lower (mean +/- SD of 85 +/- 47 km, with overall SDs of 0.61degrees and 0.99degrees of latitude and longitude, respectively) and less variable, with the difference attributable to variation in sensor orientation, intermittent shading by plumage, and the difficulty of correcting for extensive, potentially non-linear movements of flying birds. Iterative smoothing reduced both the mean error and the inflation of kernel ranges derived from GLS data, but over-smoothing contracted the extremes of the range. This reduced the overlap with radial cores apparent in the control data, and should be avoided for multinuclear GLS fix distributions. The accuracy of GLS tags is more than adequate for tracking migration and breeding-season foraging ranges of pelagic species, and for identifying broad-scale habitat preferences, overlap and potential conflict with commercial fisheries.