COS 140-4 - Defining null expectations for animal site fidelity

Abstract: Site fidelity is a widespread behavior in the animal kingdom: anadromous fish return to natal spawning watersheds, migratory animals to seasonal ranges, and herbivores to foraging patches. Site fidelity is broadly defined in the literature as the tendency to return to previously visited places, but this definition is purely phenomenological and does not include information about the generating mechanism, which is often assumed to be memory. We propose a new conceptual and operational definition of site fidelity as a pattern of returns that exceeds a null expectation derived from a memory-free movement model. Using an agent-based simulation approach, we show that even simple, memory-free movement models from the family of random walks can generate substantial (range = 0.04-0.92) probabilities of returns. Then, we present an application of our framework to empirical data. Using real data for one mule deer, we parameterize a null movement model that accounts for the individual’s space use in the absence of memory. The number of returns in the real track (n = 1561) exceeds what expected based on simulation from the null model (n = 679). These results demonstrate that the focal animal exhibits patterns of site fidelity that cannot be explained by movement and habitat selection alone, and can therefore be ascribed to memory. Our approach provides a framework to disentangle mechanisms underlying site fidelity and quantify it across species, systems, and scales.