Stochastic modelling of animal movement using intermittent search patterns

• Main Author: Bläßle, Alexander Josef
• Language: English
• Published: University of British Columbia 2013
• Online Access: http://hdl.handle.net/2429/44564

In a world that is changing rapidly due to anthropogenic disturbances, an understanding of animal behaviour is key to the protection and restoration of endangered populations. Animal movement is strongly correlated to changes in the environment and constrained by physiology. Intermittent search patterns are compositions of different movement modes enabling a forager to switch from smaller intensive search steps to large extensive search steps mostly used for relocation. In this thesis, we extend previous modelling approaches for intermittent search patterns and test these extensions on optimal foraging strategies with respect to various definitions of search efficiency incorporating costs of metabolism and locomotion. We show that intermittent search patterns appear to be optimal mainly if there is either a penalty on larger steps such as a decrease in perceptual radius or predation success rate, if the global prey density is low enough to force the predator to relocate between patches or individuals, or if the prey distribution is sparse. We also show that optimal foraging strategies are especially efficient if one movement mode is more advantageous with respect to the defined search efficiency than another, resulting in a larger difference between the optimal and worst strategies. Relocation phases are the heart of intermittent search patterns. We thus also show how predators might optimize their encounter rate during relocation (extensive search mode) phases moving along more erratic trajectories. In the second part of this thesis, we focus on how we can measure the spatial heterogeneity of prey and what this tells us about encounter rates and the chances that tagged prey items are caught. We present a novel metric that enables relative prediction of encounter rates and hitting probabilities for sparse environments by taking the initial location of prey into account. However, since measuring the location of all prey may be difficult in real world experiments, we also provide a method that uses tools developed in our optimal foraging study to estimate possible ranges for encounter rates and hitting probabilities when a set of movement and environment parameters are unknown.