Spatial self-organising as an important determinant of community dynamics in a temperate marine epibenthic community

Despite widespread acceptance of the spatial structure of ecosystems and the spatial nature of processes acting within them, critical tests of the importance of spatial phenomena to the structure and dynamics of ecosystems have not been forthcoming. In most marine epibenthic communities, the single limiting resource is space. Here I investigate the spatial dynamics of an epibenthic community in a shallow subtidal system in Tasmania and a spatial model of this community to examine how spatial process influences invasion resistance, stability, interactions among species and the growth rates of individuals. In Chapter 2 I examine the relationship between invasion resistance and species richness in the natural community. The rate of invasion increases with local species richness by two distinct mechanisms. Opportunistic colonisers invade species rich patches at higher rates because speciose patches are dominated by small colonies and mortality rates of small colonies are greater than that of large ones. Thus, mortality provides bare space for opportunists to colonise more frequently in speciose patches. However, some species avoid colonising free space but preferentially associate with established colonies of particular other species. In this case, a given preferred associate is more likely to occur in more species rich patches, and so colonisation rates are greater in more speciose patches. In Chapter 3 the importance of spatial context on the outcomes of pair-wise species interactions and neighbour-specific growth rates is examined. The outcomes of competitive (overgrowth) interactions among species and neighbour-specific growth rates in experimentally contrived pair-wise interactions are often dissimilar to their counterparts in the non-manipulated natural community. I use a spatial model and its non-spatial equivalent to demonstrate that these differences in outcomes and growth significantly affect predicted community dynamics. In Chapter 4 I develop a spatial simulation model parameterised by empirical observations of the recruitment, growth, interaction outcomes and mortality of the natural community. I compare the model dynamics to the dynamics of manipulated and non-manipulated natural communities. The model self-organises to form distinct colonies and adequately captures many features of the short-term dynamics of manipulated communities observed over a 16 month period, although some model behaviours are not reflected in the natural community. When compared to the longerterm dynamics of the non-manipulated communities (ca. 12 years), the model accurately predicts the species evenness, diversity, size structure and multivariate variance of these communities. None of these emergent features is evident from equivalent non-spatial (mean field) models. In Chapter 5 I use the model developed and tested in chapter 3 to examine the relationships between species richness, area, persistence stability of total cover, resilience stability and invasion resistance. Communities occupying small areas are less stable and less resilient than those in larger areas. While richness is positively correlated with persistence stability in small landscapes (