Animal movements: an optimal foraging approach

Foraging animals, in addition to deciding where and when to forage, what to feed on, and how long to spend in one area before departing for another, must also decide how to get from one location to another or what movement strategy to adopt. If you observe foraging animals you will generally see them moving or changing location, exhibiting patterns in terms of distances and directions as they go and patterns in terms of where they end up spending their time. In attempting to understand such patterns, Optimal Foraging Theory (OFT) may help.

Early attempts, during the 1970s and early 1980s, to use OFT to understand movements of foraging animals generally took a Cognitive Forager Approach (CFA) in which animals were assumed to be aware of their internal state, have a sense of direction and thus able to maintain directionality to their movements, an ability to sense potential food items at a significant distance, and memory regarding previous foraging. They also often assumed that food is patchily distributed, with multiple items tending to occur relatively close to one another (e.g., seeds on the ground) or nearby food locations (e.g., flowers) tending to have similar amounts of food.

More recently, some attempts to use OFT to understand movements of foraging animals have adopted the Lévy foraging hypothesis (LFH), which is the antithesis of the CFA. According to the LFH, as originally conceived in the late 1990s, animals employ a simple random walk (i.e., no directionality) to search for randomly distributed food items in a featureless environment and in a completely uninformed manner, with no sense of direction, no ability to perceive food items unless they ‘bump’ into them, and no memory regarding previous circumstances. In other words, animal foragers are assumed to be clueless, senseless and uninformed.

The LFH should be abandoned in favour of the CFA, despite the former receiving considerable acclaim from its proponents. The assumptions behind the Lévy foraging hypothesis bear little resemblance to biological reality, as no foraging animal is clueless, senseless and uninformed, and thus its predictions lack validity. The LFH also unrealistically omits directionality of movements, where successive movement segments tend to be in the same direction, and Area Restricted Search, where food encounter or encounter with a relatively high amount of food is followed by increased turning and shorter movement segments which tend to keep an animal in the vicinity of encountered food. In addition, Lévy-like behaviour is hardly surprising, arising as an emergent property of many natural processes, and hence tells us little to nothing about how and why animals forage in the ways they do.

Since about 2005, the CFA has been reborn as a revamped ‘movement ecology’ with models labelled as agent-based models (ABM) or individual-based models (IBM). However, these models have been primarily descriptive rather than theoretical, and not so far resulted in the development and testing of predictions.

Future research on movements by foraging animals should therefore focus on situations where foraging can be differentiated from other behaviour, what the animals can perceive and remember is reasonably clear, food encounter can be recorded, the abilities of animals to collect and consume encountered food are reasonably clear, the spatial and temporal distribution of food can be determined, and everything can be expressed as a mathematical model. Model assumptions can then be refined through comparing emergent patterns with observed patterns of foraging movement and optimality predictions can be compared with observations.

In the end, we should better understand why foraging animals move the way they do.