Scale-dependent to scale-free: daily behavioural switching and optimized searching in a marine predator

Rhythmic activity patterns are ubiquitous in animals and in the marine environment a dominant rhythmic activity is the diel vertical migration (DVM) of pelagic organisms, moving or 'migrating' from deep waters during the day to shallower waters at night. While this overall pattern of movement is well understood, the cryptic nature of the marine environment has limited the study of fine-scale movements within each phase. Active pelagic predators, such as tuna, perform consistent, predictable large-scale vertical movements; however, the fine-scale movements nested within these larger movements have not previously been investigated in detail. Further, the prey field densities are known to differ significantly between day and night, presenting an opportunity to study differences in foraging patterns between these two phases. Here, using long-term depth time series recorded from 93 bigeye tuna, Thunnus obesus, with electronic tags (18 003 days of data), fine-scale changes in vertical movement patterns between day and night time phases were investigated in the context of the Lévy foraging hypothesis, which predicts a Lévy distribution of move steps during foraging when prey is scarce, but an exponential distribution when prey is abundant and searching is not required. During the day, T. obesus were found to exhibit scale-free movements well fitted by a Lévy distribution indicating optimized searching for sparsely distributed prey. During night-time hours, however, exponentially distributed scale-dependent move step lengths were found to be dominant, supporting a simple, Brownian, movement pattern sufficient where prey is abundant. This study not only confirms the predictions of the Lévy foraging hypothesis but suggests that the identification of Lévy patterns in movement data can be a useful indicator of foraging activity in animals that are difficult to observe directly.