Group-based modeling of ecological trajectories in restored wetlands.

Abstract

Repeated measures taken at the same restoration sites over time are used to describe restoration trajectories and identify sites that are trending toward unexpected outcomes. Analogously, social scientists use repeated measures of individuals to describe developmental trajectories of behaviors or other outcomes. Group-based trajectory modeling (GBTM) is one statistical method used in behavioral and health sciences for this purpose. I introduce the use of GBTM to identify clusters of similar restoration trajectories within a sample of sites. Data collected at 54 restored wetlands in Illinois for up to 15 years post-restoration were used to describe trajectories of six indicators: plant species richness, number of Carex (sedge) species, mean coefficient of conservatism (mean C), native plant cover, perennial plant cover, and planted species cover. For each indicator, I used GBTM to classify wetlands into three to four groups with distinct trajectories. In general, cover by native and planted species declined, while species richness and mean C increased over time or peaked then declined. Site context and management may explain trajectory group membership. Specifically, wetlands restored more recently and those restored within forested contexts were more likely to follow increasing trajectories. I show GBTM to be useful for identifying typical restoration trajectory patterns, developing hypotheses regarding factors driving those patterns and pinpointing critical times for intervention. Furthermore, GBTM might be applied more broadly in ecological research to identify common patterns of community assembly in large numbers of plots or sites.