Community-level analysis of spatiotemporal plant dynamics

Abstract

We present an approach for the analysis of spatiotemporal patterns and scale-dependence in the dynamics of plant communities that combines well-known methods to reduce complexity arising from the high-dimensionality of ecological data. Our approach takes into account both correlations and autocorrelations in the structure and dynamics of the community. In application to a sand dune annual plant community, we find answers to questions regarding the nature of community response in space and time, and identify significant spatiotemporal interactions and spatiotemporal scales. Community-level spatial correlograms reveal a fixed pattern in time that is not apparent from species-level dynamics. In this sense, the community is shown to be more stable than the sum of its parts. This form of stability is not a simple artifact of averaging species-specific responses, and points to some consistency in species interactions. Temporal correlograms are highly spatially-specific. In this sense, the inclusion of the spatial information is shown to change our view of temporal dynamics. The relative importance of biological processes versus an environmental gradient is examined in the temporal community response. We find that community response to rainfall is also spatially-specific and that the temporal autocorrelation dynamics shows similar patterns even when the effect of rainfall is removed. Our findings are relevant to both empirical and theoretical work aimed at understanding the interaction between space and time in ecological communities.