Competition coefficients in a marine epibenthic assemblage depend on spatial structure

We investigated the importance of the spatial context of interactions in a multispecies marine epibenthic assemblage with respect to the outcomes of interspecific interactions, neighbour-specific growth rates, and the dynamics of spatial and mean-field models of the system. We compared the outcomes of interactions and overgrowth rates of pair-wise combinations of species in spatially simplified contrived interactions with the same combinations in an unmanipulated assemblage. While estimates of neighbour-specific growth rates were similar in both sets of interactions, the probability of a species winning a particular interaction was strongly dependent on whether the interaction was contrived or occurred in the unmanipulated assemblage. The dynamics of a spatial model and its mean-field equivalent parameterised from estimates of interaction outcome and neighbour-specific growth from contrived interactions were significantly different to the dynamics of models based on estimates of interaction outcome and neighbour-specific growth obtained from non-manipulated assemblages. Differences in the dynamics of models based on parameters from unmanipulated and contrived interactions are primarily due to differences in outcomes of interspecific interactions, while fluctuations in growth rates contribute to the variability around these dynamics. Our results suggest that conclusions about interspecific interactions and community dynamics examined in simplified spatial associations (e.g. in manipulative experiments) is likely to be limited to assemblages with a similarly simplified spatial structure, which is an unlikely occurrence in nature.