A maximum-entropy description of animal movement

TL;DR

This paper develops a maximum-entropy framework that encompasses various stochastic models of animal movement, predicts new models with improved data, and links Langevin equations to fluctuation-dissipation principles.

Contribution

It introduces a unifying maximum-entropy approach to model animal movement, including existing and potential new stochastic processes, and establishes theoretical constraints on Langevin equations.

Findings

Unified maximum-entropy framework for animal movement models

Prediction of new models as data quality improves

Langevin equations must obey fluctuation-dissipation theorem

Abstract

We introduce a class of maximum-entropy states that naturally includes within it all of the major continuous-time stochastic processes that have been applied to animal movement, including Brownian motion, Ornstein-Uhlenbeck motion, integrated Ornstein-Uhlenbeck motion, a recently discovered hybrid of the previous models, and a new model that describes central-place foraging. We are also able to predict a further hierarchy of new models that will emerge as data quality improves to better resolve the underlying continuity of animal movement. Finally, we also show that Langevin equations must obey a fluctuation-dissipation theorem to generate processes that fall from this class of maximum-entropy distributions.