The risk allocation hypothesis (RAH) predicts that temporal variation in predation risk can influence how animals allocate foraging activities among situations that differ in predation pressure. The model predicts that in situations of relative safety, prey can meet their energetic demands and exhibit strong anti-predator responses when rare, dangerous events occur. Conversely, if high-risk situations predominate, an animal must take greater risks to gather enough of its energetic requirements to survive. The RAH is based on three underlying assumptions related to prey energetic state, the energetic cost of predator avoidance relative to starvation, and the predictability of risk. Despite the elegance of the model, most real-life systems show only partial support. This study seeks to develop a more comprehensive theory of optimal risk allocation by constructing an agent-based model to more accurately reflect different prey foraging strategies. We first reproduce the basic RAH model and the patterns of prey activities expected under different risk intensities (attack ratios) and frequencies (duration). The model then examines how different prey energetic reserves, life histories, and prey methods in risk assessment can violate the assumptions of the RAH and result in deviations from the predicted activity patterns.
Results/Conclusions
Simple rules minimizing predation and starvation risk in prey were successful in reproducing activity patterns predicted by the RAH. However, energy-rich prey, prey that minimized predation risk versus growth, and prey with different cognitive abilities (information processing) were differentially affected by temporal variations in their risk regime. Energy-rich prey and predation-sensitive foragers decreased their activity levels under constant high predation, whereas ‘growth-maximizers’ undervalued and ignored predation risk when constant and intense to minimize starvation. Even with small imperfections in current knowledge, prey both over- and under-estimated risk when it varied from high to low, respectively, even though costs of anti-predator responses had changed accordingly. These findings can be used to explain how deviations from the RAH can occur, and can stimulate the re-analyses of existing data, and/or the design of new field and experimental studies. Our agent-based model can furthermore be used to facilitate tests of the predictions and expectations of foraging-prey risk allocation under more realistic simulated natural systems and with added environmental complexity. As predation is one of the most important selective pressures acting on prey behaviour, these modelling approaches are of general significance across many biological systems.