A model of plant community dynamics based on replacement networks

Questions When a tree or a shrub dies, the space it occupied can be overtaken by plants that were recruiting beneath it or colonized by new species. Such replacement processes can drive the temporal change in species abundance in a plant community. Can we predict the dynamics of a real plant community using observational data on plant–plant recruitment interactions? What would be the relative importance of recruitment interactions vs life‐history traits in determining community dynamics? Location Forest communities dominated by Pinus halepensis and Quercus ilex in SE Spain. Methods We develop a continuous time non‐linear model that can be easily parameterized with empirical data for the interactions between adult and juvenile plants recruiting beneath them. These interactions form a complex replacement network (or matrix) that is the backbone of the model. We parameterize the model with life‐history data from the literature and from recruitment interactions observed in a successional community 12 yr after a forest fire. We explore the behaviour of the model under different intensities of chronic disturbance, and after modifications of the structure of the replacement network and the values of the parameters. Results The model predicts that the current community of the burned area will develop into a forest very similar quantitatively to the surrounding mature forests, as long as the rates of chronic disturbance remain very low. For increasingly higher levels of chronic disturbance, the community would reach stable states resembling a mixed pine–oak forest, a degraded oak forest, an oak dehesa and, finally, a steppe‐like vegetation. All these types of plant assemblages can currently be found throughout the study area. These predictions are less sensitive to variation in the estimates of species' life history (i.e. growth, death and colonization rates) than to variation in the structure of the recruitment matrix. Conclusions The model projects realistic community dynamics. The analysis of the model suggest that understanding the structure of replacement networks and how they are assembled can contribute significantly to our knowledge of the dynamics and stability of forest plant communities.