In the last decades research into movement has taken flight in ecology. The development and miniaturization of tracking devices has enabled ecologists to collect and store the movement data of a large and increasing number of animals and species. Next to this increase in available data new theoretical models have been developed and discussed at length in the ecological literature. These developments together form what has been called “movement ecology”. The research in this thesis falls squarely within this new field.
In movement ecology there has been a lot of attention for the analysis of movement paths and the comparison of these paths with random walks, which were first used for describing the random movement of particles in physics. These random walks were compared with respect to their ability to encounter resources (e.g. food items); the idea being that those random walks that are more efficient will have more encounters and will through natural selection be more likely to occur in nature. In the analysis of their search efficiency there was, however, relatively little attention for the spatial resource distribution. Especially when my PhD-research started only some basic models were used to describe the spatial resource distribution and used to test the effect on the search performance of random walks. For my research I therefore set myself the aim of determining how the spatial context influences the search efficiency of the main random walks used in ecology through simulation and to test this using experiments.
The first analysis of search efficiency with varying resource density and aggregation as described in chapter 2, shows that search efficiencies are dependent on both the resource aggregation and density as well as on their interaction. These results show that any analysis of search efficiency requires specification of both density and degree of aggregation. Furthermore, the effect of changes in either density or aggregation on the availability of resources may not be straightforward (e.g. linear with density). An optimally searching animal will have to switch between different random searches, depending on the resource distribution.
The next step (chapter 3) was to determine the influence of the spatial resource distribution when it was described and approached from the “patch” framework that is ubiquitous in ecology. In this framework it is the distribution of patch sizes (number of resources) that creates variation in the spatial distribution of resources.. The results show that it is not the variance in patch sizes, but the skewness of the patch-size distribution that determines the long-term search results of random walks. This is highly relevant since such skewed distributions are often seen in the distribution of plants, animals and resources. In addition the results indicate yet again that different distributions are best exploited by different random walks, and an efficient searcher would thus change movement depending on the spatial resource distribution.
Random walks are used for particles in physics and do not use information. Many animals do, however, use additional information in their search for food or resources. The use of information in foraging has been mainly studied in foraging theory. In chapter 4 I again studied the influence of the spatial resource distribution, but now with the searcher using information on past encounters. For this model I combined a random walk model with an information-use model from foraging theory. The results show that even when a searcher uses information on recent encounters an optimally searching animal will have to change its sensitivity depending on the spatial resource distribution.
On the basis of the simulations in the preceding chapters, for chapter 5 I conducted an experiment with carabid beetles. The beetles were allowed to forage in distributions varying in density and aggregation. The main hypothesis was that their movement patterns would change to be optimal in the respective distributions. The results showed, however, that the beetles did not change their movement behaviour with changes in the offered distribution. Based on this we expect the beetles’ main resource to have a distribution that is efficiently searched using their movement pattern, which means that their resources are expected to have a random spatial distribution, with relatively high density.
Finally in the 6thchapter I review the results from the previous chapters and conclude that optimal searchers need to adapt to the resource distribution when they use random searches or information on past encounter. I argue for more emphasis on and explicit study of the spatial distribution in movement ecology, and for integration of movement ecology and foraging theory to study the how foragers should deal with aggregation which is the fundamental challenge for both foraging theory and movement ecology .
Ultimately, by including landscape elements more rigorously and integrating movement ecology with existing fields such as invasion biology and foraging theory I believe it will be possible to truly include the movement process into ecological explanations and understanding, finally enabling researchers to provide clear first principle explanations and predictions not only for ecology, but also for the benefit of epidemiologists, nature conservation and wildlife management and thus for society as a whole.