Refuge competition in the freshwater prawn, Macrobrachium Australiense : linking individual behaviours to population patterns

By utilising an individual-based model (IBM) I directly examined the link between system level properties and their structural roots, the small scale competitive behaviours of individuals. The refuging behaviours of a diurnally inactive freshwater prawn, Macrobrachium australiense, were used as the basis for the IBM which explored the role of refuge competition on exposure to predation. Physical experiments were conducted to determine the behavioural rules and parameterise the model. The first series of experiments investigated the diel refuging patterns and competitive refuging behaviours of the study animal (Chapter 2-4). These were found to be strongly anti-predatory; individuals sharply reduced diel activity and spent the vast majority of their time in available vegetation refugia (Chapter 2). The presence of conspecifics, however, affected the activity of focal individuals, and their exposure to predation, by two separate mechanisms. The first was detectability related to the Size of the conspecific and the Refuge stalk Density (SiRD mechanism) (Chapter 3-4), the second was driven by the activity of the conspecific (Motion mechanism) (Chapter 4). During the SiRD mechanism the response of the individual was also related to its position in the dominance hierarchy relative to the conspecific. For dominant individuals activity increased in all refuge densities tested with increasing conspecific detectability (Chapter 3). Subordinate response in the SiRD mechanism was also affected by the quality of the refuge (Chapter 4). When faced with increasing dominant size subordinates in low quality refugia reduced activity, this contrasted with high quality refugia where their activity increased. This was explained in terms of alternative mechanisms to minimise contact with dominants: evasion and crypsis in high and low quality refugia respectively (Chapter 4). Compared to SiRD the Motion mechanism was relatively straight forward. In all cases the activity of other individuals was found to trigger an increase in the activity of proximate conspecifics (Chapter 4). During these experiments it was noted that when introduced to a novel environment M. australiense individuals would actively explore it, showing uncharacteristically high levels of diel activity. The subsequent experiments examined whether or not they required this exploratory period to determine the location and quality of refugia (Chapter 5). The results demonstrated that the study animals were able to detect and evaluate the quality of distant novel refugia without exploring them first. However, this ability did not preclude the performance of exploratory behaviours (Chapter 5). The behaviours in the preceding chapters formed the basis for the construction and parameterisation of the IBM (described in Chapter 6). Tests of the models validity showed that it could make adequate predictions about the distribution of individuals around an experimental aquarium and was able to reproduce a majority of trends seen in a series of independent experiments (Chapter 7). Though, the SiRD mechanism was underestimating the magnitude of the size related trends. The overall results determined that the model could describe the general effects of different behaviours on the exposure of individuals to predation, but would need further parameterisation to make predictions about real world scenarios. Bearing this in mind exploratory simulation experiments were conducted to test the link between the simulated behaviours and system level patterns (Chapter 8-9). Three core effects were found to affect the populations exposure to predation. First, the cumulative Motion mechanism resulted in the emergence of a density-dependent pattern affecting predator exposure which increased with escalating population size and decreasing refuge availability (Chapter 8). Second the low movement likelihood of individuals limited their spatial range to a subsection of the simulated environment. Because of this the fragmentation of refugia only had an impact on exposure to predation when it occurred at a scale relative to this spatial range (Chapter 8). The third core effect demonstrated a mechanism by which refuge competition could affect individuals exposed to predation and subsequently the size structure of the population (Chapter 9). When individuals, which surpassed a threshold of exposure to predation, were removed from the population there was a general increase in the average body size of the population. The number of individuals removed from the population, and the rate of increase in average body size was strongly affected by the initial body size variation of the individuals. This effect was strong despite the fact that it was probably underestimated by the model, as it was driven by the SiRD mechanism (Chapter 9). Overall the behavioural rules of the model succeeded in reproducing the system level patterns, demonstrating that they captured some essential processes of competitive refuging behaviours. The main contribution of this thesis to the scientific literature is the description of an explicit link between the competitive refuging behaviours of individuals and their resultant population patterns