A Minimalist Model of Salt-Marsh Vegetation Dynamics Driven by Species Competition and Dispersal

We present a new bidimensional, spatially-explicit ecological model describing the dynamics of halophytic vegetation in tidal saline wetlands. Existing vegetation models employ relatively simple deterministic or stochastic mechanisms, and are driven by local environmental conditions. In the proposed model, in contrast, vegetation dynamics depend not only on the marsh local habitat, but also on spatially-explicit mechanisms of dispersal and competition among multiple interacting species. The role of habitat quality, here determined by the local elevation relative to the mean sea level as a proxy for environmental conditions, is mathematically modeled by a logistic function that represents the fundamental (theoretical) niche of each halophytic species. Hence, the model does not artificially impose any constraints to the ability of a species to colonize elevated areas where it is usually not observed: such limitations naturally arise through competition with fitter species across marsh topographic gradients. We qualitatively test our model against field data based on a suitable assemblage of focus species, and perform a sensitivity analysis aimed at determining how dynamic equilibria in vegetation distributions are affected by changes in model input parameters. Results indicate that the model is robust and can predict realistic vegetation distributions and species-richness patterns. More importantly, the model is also able to effectively reproduce the outcomes of classical ecological experiments, wherein a species is transplanted to an area outside its realized niche. A direct comparison shows that previous models not accounting for dispersal and interspecific competitions are unable to reproduce such dynamics. Our model can be easily integrated into virtually any existing morphodynamic model, thereby strengthening our ability to simulate the coupled biotic and abiotic evolution of salt marshes under changing climate forcings.