As are most ecosystems, lakes and oceans are heterogeneous in space and time. This heterogeneity presents opportunities for species coexistence. We apply adaptive dynamics and related trait-based modeling techniques to understand the emergence of community structure in heterogeneous environments, with particular emphasis on planktonic communities. The primary axis of spatial variation in aquatic systems is the vertical axis, due to steep gradients in the essential resources of light and nutrients. Here we use reaction-diffusion models for phytoplankton competition. A major source of temporal variability is due to the annual variation in light, temperature, and mixing, which leads to seasonal succession of species. Here we use externally forced systems of differential equations.
Results/Conclusions
Spatial structure: Under weakly mixed conditions, phytoplankton can play a game against conspecifics where to live within the water column, resulting in an aggregation at depth. In turn, this intraspecific aggregation can enable coexistence of multiple species that show a trade-off between light and nutrient competitive abilities. Community structure varies along gradients of nutrient supply, mixing, and light attenuation.
Temporal dynamics: We consider two coexistence mechanisms in variable environments: relative nonlinearity due to varying resource levels and the storage effect due to varying temperature. In both cases we examine the evolutionary robustness of coexistence. Relative nonlinearity easily leads to evolutionarily stable coexistence of two species, but higher diversity is unlikely. The storage effect allows more diverse communities provided the environmental response functions are sufficiently narrow relative to the environmental forcing. Heritable intraspecific trait variation leads to rapid evolution and can undermine species coexistence.
Conclusions: Trait-based models provide a way to understand the emergence of community structure from a universe of possible species. Mathematical and computational advances are making it possible to apply these techniques to more complex community models that include the key features that define particular ecosystems, leading to empirically relevant predictions.