Convergence and divergence in old field succession: Are there predictable trends in long-term dynamics?

Background/Question/Methods

A fundamental goal of ecology is to explain the consequences of interactions between species and their environments. This is of particular importance in old field succession – the process by which species communities and environmental characteristics of agricultural fields and pastures change over time following abandonment. Because these changes include large shifts in aboveground productivity, soil chemistry, and species composition, predicting specific long-term consequences of such succession is of considerable interest. Nevertheless, old field succession has proven notoriously difficult to predict, with seemingly similar fields developing along radically different trajectories. It is therefore unclear whether observed dynamics are the result of 1) a single noisy process, 2) a few distinct successional trajectories, or 3) such a large number of idiosyncratic processes that measurement error precludes accurate prediction. We address this question in a well-studied system by developing and calibrating models to explore the relative explanatory power of these three possibilities based on observed species and environmental changes in old fields at Cedar Creek, MN. To incorporate all three into a single framework, our model links stochastic estimates of dispersal and mortality with mechanistically-based tradeoffs in resource use among species.

Results/Conclusions

We find strong evidence suggesting more than one distinct successional trajectory in old fields at Cedar Creek. Among replicate fields, there are significant differences in the timing of colonization and maximal abundance of species, even for the most common species. These differences also appear to be persistent, as estimated colonization rates, mortality rates, and equilibrium abundance also differ significantly among populations of species in replicate fields. For late successional species (e.g. Schizachyrium scoparium) the best-performing model included only a few possible parameter values, suggesting a limited number of successional trajectories and long-term abundances. For early successional species (e.g. Elymus repens) the best-performing model included separate parameter values for each old field, suggesting a greater dependence on ecological context (i.e. the presence and abundance of other species). These results suggest that targeted manipulation of species pools in old fields could therefore alter the rate of successional dynamics and long-term abundance of species in old fields, perhaps shifting species and environmental trajectories to better maintain more managerially “desirable” conditions.