Interpreting trajectories of restored wetland plant communities
Each year approximately 20,000 hectares of wetlands are created or restored to mitigate for wetland losses permitted under the U.S. Clean Water Act. The underlying assumption is that compensatory wetland restoration can adequately and rapidly replace destroyed natural ecosystems, implying a high degree of predictability in ecological succession. Restoration progress in these restored wetlands has been conceptualized as a trajectory—the change over time in a measurable indicator of restoration progress, which ideally increases until it reaches a predefined goal or some set of legally mandated performance standards. However, restoration outcomes have been difficult to predict due to the erratic and context-dependent nature of restoration trajectories. As a consequence, restored wetlands often fail to meet expectations. For restoration ecology to advance beyond descriptions of case studies and toward predictive and mechanistic science, we need to uncover commonalities among trajectories. To this end, we have conducted long-term monitoring and comparative studies of plant communities at several restored wetlands in Illinois.
We have found that analyzing patterns of convergence and divergence among restorations is useful way to characterize restoration trajectories. Immediately after restoration, wetlands trend toward target plant species composition, becoming progressively more similar to reference wetlands. However, through time, many restored wetlands deviate from this trajectory to converge upon a species composition indicative of low quality, degraded wetlands. Starting from an analogy that trajectories of ecological indicators are similar in many ways to trajectories of health indicators in humans, we have also introduced a statistical method developed within the behavioral sciences, called group-based trajectory modeling, to study restoration trajectories. Group-based trajectory modeling is a procedure for identifying clusters of unique trajectories and classifying subjects, in this case sites, into discrete “trajectory groups”. By identifying common trajectory shapes this method makes it possible to describe patterns of divergence and convergence among restorations. For example, in wetlands that were initially dissimilar, trajectories of plant community metrics such as species richness and cover by perennial and planted species converged on common outcomes. Comparative analyses of restorations can be used to distinguish between achievable and unachievable restoration goals, develop hypotheses regarding the factors which constrain restoration trajectories, and pinpoint critical times for management intervention.