While the spatial arrangement of individuals within plant communities has been shown to affect the outcome of competition on focal individuals and species, many theoretical studies still assume random, well-mixed communities. Of those spatially explicit studies of species interactions, most focus entirely on direct, competitive interactions, without testing for indirect or higher-order interactions (HOIs) which can mediate both competition and facilitation in communities.
Here we introduce HOIs into models of species interactions while accounting for the spatial arrangement of local neighbours. Specifically, we aim to answer the following questions: Do spatially explicit models of species interactions improve model fit in individual fitness models? What is the spatial extent of interactions in which neighbours have detectable direct and higher-order interactions? Does this distance change with the identity of the focal individual?
We address these questions by mapping the spatial arrangement and measuring fecundity as seed set for over 12,000 individual plants in the herbaceous annual understory of York gum-Jam woodlands of southwest Western Australia. The spatial and fecundity data were used to estimate the rate of decay of species interactions over space for direct and higher-order interactions between species.
Results/Conclusions
From analysis of two focal species with at least 40 different neighbourhoods, we found that including space within these individual fitness models increases model fit as well as include complexity observed in natural systems in the models. The inclusion of HOI terms allowed for facilitative interactions and the addition of spatial weighs further dampened competitive effects. We found that intraspecific interactions had a slower rate of distance decay for interaction distance compared to interspecific interactions, though the rate of decay of interspecific interactions varied depending on the neighbour species. This trend was observed for both focal species, Velleia rosea and Arctotheca calendula.
We conclude that including spatial weights increases our ability to predict coexistence within 10cm interaction neighbourhoods. Models that don’t include space are likely to underestimate species interactions at small spatial scales and overestimate them at large spatial scales. Determining the rate of decay of interactions for a specific species has broader consequences for understanding community dynamics and for species conservation. Continued study of the interactions between spatial scale and community dynamics will allow ecologists to hone their understanding of the extent to which this phenomenon is ubiquitous across plant systems.