Analyzing and modeling animal movements in heterogeneous landscapes

Date of Completion

January 2004


Biology, Ecology|Biology, Zoology




As humans modify, destroy and create habitats, there is a pressing need for understanding the effects of landscape heterogeneity on population dynamics. In the first two chapters, I used simple movement models to show that redistribution kernels (functions that describe the probability that an individual moves a certain distance) converge to Gaussians with diffusion parameters that depend on boundary permeability and landscape composition but not on the details of spatial structure. Before reaching convergence, the shape of kernels were dynamical and showed long tails. Individual based models showed that population rate of increase were slower when the landscape was composed mostly of slow habitat. Equilibrium population density decreased with increasing spatial autocorrelation, probably due to individuals aggregating in slow-habitat patches and thus increasing local competition. Redistribution kernels are usually composed of mixtures of different movement behaviors. Landscape properties, together with boundary behavior determine the relative amount of each movement in the mixture. In the third chapter I present Bayesian methods that allow for identification of different movement states using several properties of observed movement paths. Analysis of relocation data from elk released in east-central Ontario suggests a bi-phasic movement behavior: elk are either in an ‘encamped’ state in which step lengths are small, or, in an ‘exploratory’ state, in which daily step lengths are several kilometers. Animals encamp in open habitat, but the exploratory state is not associated with any particular habitat type. ^ In the fourth chapter I asked what movement strategies were likely to be selected under different landscape contexts. Individual decision-making was modeled with neural networks that received as input those variables suspected to be important in determining movement efficiency. Energetic gains and losses were tracked using known physiological characteristics of ruminants. Genetic algorithms were used to improve the performance of the decision processes in different landscapes. Emergent properties of movement paths were formation of home ranges and an alternation between small, localized movement with larger, exploratory movements. I found similarities in several aspects of their movement patterns such as in the distributions of distance moved and turning angles, a tendency to return to previously visited areas, and avoidance of steep slopes. ^