How well can animals navigate? Estimating the circle of confusion from tracking data

State-space models have recently been shown to effectively model animal movement. In this paper we illustrate how such models can be used to improve our knowledge of animal navigation ability, something which is poorly understood. This work is of great interest when modeling the behavior of animals that are migrating, often over tremendously large distances. We use the term circle of confusion, first proposed by Kendall (1974), to describe the general inability of an animal to know its location precisely. Our modeling strategy enables us to statistically describe the circle of confusion associated with any animal movements where departure and destination points are known. For illustration, we use ARGOS satellite telemetry of leatherback turtles migrating over a distance of approximately 4000 km in the Atlantic Ocean. Robust features of the model enable one to deal with outlying observations, highly characteristic of these types of data. Although specifically designed for data obtained using satellite telemetry, our approach is generalizable to other common kinds of movement data such as archival tag data.