Anomalous diffusion or, more generally, anomalous transport, with nonlinear dependence of the mean-squared displacement on the measurement time, is ubiquitous in nature. It has been observed in processes ranging from microscopic movement of molecules to macroscopic, large-scale paths of migrating birds. Using data from multiple empirical systems, spanning 12 orders of magnitude in length and 8 orders of magnitude in time, I employ a method to detect the individual underlying origins of anomalous diffusion and transport in the data. I show that such a decomposition of real-life data allows to infer nontrivial behavioral predictions in the fields of single particle tracking in living cells and movement ecology. Notably, I show that a similar classification of anomalous diffusion can be obtained using power spectral analysis.

Focusing on the ecological scale, I use rich datasets of movement of three avian predators to reveal unique movement characteristics. I employ the above decomposition and the study of ergodicity breaking in stochastic dynamical systems to establish that local searches are irreproducible (i.e., nonergodic) while long-range commuting is reproducible (i.e., ergodic). In ecological terms, my findings show that there are many ways to hunt within a patch but only a limited number of ways to commute between distant patches. Using a continuous-time random walk model, I identify a behavioral switch between local searches and long-range movement between patches.