Individual based simulations of the direct and indirect effects of limpets on a rocky shore Fucus mosaic

Ecological models that include spatial processes can produce complex patterns in space and time. However, the ecological relevance of such behaviour remains unclear. Previous work has generally focused on systems where the dynamics of simulated species are closely linked. We present a stochastic spatial model for an open system where species are only weakly or indirectly linked. The principal interaction, that limpets reduce the local probability of Fucus recruitment, was defined empirically. Fucus abundance at large scales could be approximated without resorting to a spatially explicit model. However, the behaviour of individual limpets at intermediate densities could alter Fucus abundance and small scale spatiotemporal pattern. By altering small scale pattern, in simulations, limpets could also affect the population densities of poorly dispersing species dependent on Fucus cover. Increasing the temporal variability of limpet or Fucus populations at large scales led to correlated dynamics at different spatial scales in simulations. There was little correlation between scales in time series taken from the field. This weak correlation between scales suggests that local effects, such as those associated with individual limpet grazing, are important in the dynamics of a real system. Development and testing of the model are facilitated because predictions apply to clearly definable temporal and spatial scales.