Advances in the tracking of marine species: using GPS locations to evaluate satellite track data and a continuous-time movement model

Abstract

Argos satellite tracking provides information about the large-scale movements of marine species, but the limitations in position accuracy and frequency make it difficult to interpret fine-scale behaviour. With Fastloc global positioning system (GPS) technology, it is now possible to overcome these limitations when tracking diving marine species. We compared differences among archived GPS (GPS), transmitted GPS (GPS-t) and Argos satellite (PTT) tracks acquired simultaneously on 30 northern fur seals Callorhinus ursinus. We examined times and distances between locations, as well as overall trip characteristics (e.g. distance traveled and transit rate). The GPS data were also used to test the accuracy of a continuous-time correlated random walk model created to cope with the spatial error and gap times associated with PTT locations. Significantly more GPS locations per day were acquired than PTT locations (31.6 ± 1.9 vs. 12.0 ± 0.3, respectively), and the GPS locations were more evenly distributed along the track. The influence of data type (GPS, GPS-t, PTT) varied based on the parameter measured, ranging from different among all (e.g. average transit rate) to no significant difference (e.g. maximum distance traveled). Modeling of both PTT and GPS-t data resulted in tracks with over 79% of predicted locations less than 5 km from the GPS location (average location error: 3.2 ± 0.1 and 1.7 ± 0.1 km, respectively). This study demonstrates the added benefit of using GPS to track marine species, as well as how and when modeled PTT data may be sufficient to address study questions.