Ecological Archives A019-087-A4

Jeffrey W. Matthews, Greg Spyreas, and Anton G. Endress. 2009. Trajectories of vegetation-based indicators used to assess wetland restoration progress. Ecological Applications 19:2093–2107.

Appendix D. Restoration trajectories in individual wetlands.

This Appendix provides a more in-depth examination of restoration trajectories from individual wetlands. Restoration trajectories of eight vegetation-based indicators of restoration progress from four sites, including two forested wetlands and two herbaceous wetlands, are shown in Fig. D1. Site 28 is a created marsh in northern Illinois (see Appendix A for site information). The site was excavated and planted with several marsh and wet prairie plant species. Relative to reference wetlands from northern Illinois, Site 28 achieved high richness of native species (Fig. D1A), Carex species (Fig. D1U) and conservative species (Fig. D1Q), as well as a high Floristic Quality Index (FQI; Fig. D1M) and mean Coefficient of Conservatism (mean C, Fig. D1I). Even in this relatively successful restoration, however, proportion native species (Fig. D1E) and total importance value of native species (Fig. D1AC) remained close to the median level in reference wetlands, suggesting that the representation of native relative to exotic plant species may be one the most difficult characteristics of the plant community to restore.

Site 12, in contrast, was less successful. This created marsh was constructed within a highway interchange in central Illinois, and was rapidly invaded by the non-native Typha angustifolia, which after five years made up 22% of the plant cover at the site. Nevertheless, indicators based on species richness, including native richness (Fig. D1B), FQI (Fig. D1N) and conservative richness (Fig. D1R), generally increased over time and by the end of five years had achieved levels higher than a majority of reference wetlands. Mean C in this site leveled off below the median level in reference sites (Fig. D1J). Proportion and importance of native species initially increased, but then declined over time (Fig. D1F, D1AD).

Site 5 was a restored floodplain forest on former agricultural land in southern Illinois. As in Site 28, most indicators increased over time, and indicators based on species richness achieved levels exceeding those in a majority of reference, southern Illinois, forested wetlands (Fig. D1C, D1S, D1W). Proportion native species and Mean C, however, remained low relative to reference sites (Fig. D1G, D1K).

Site 20 was also a restored floodplain forest on former agricultural land, but was located in northwestern Illinois. This site was heavily invaded by the aggressive, exotic grass Phalaris arundinacea, which comprised 42% of the herbaceous cover after eight years. In contrast to the other sites, most vegetation-based indicators declined over time or showed no discernable trends (Fig. D1). Even in this site, however, native richness (Fig. D1D), conservative richness (Fig. D1T) and FQI (Fig. D1P) were high relative to at least one set of reference wetlands.

Across all four sites, proportion perennial species, where it increased over time, increased only slowly relative to reference sites (Fig. D1Y, D1Z, D1AA, D1AB). This was especially the case for the restored forested wetlands, suggesting that this characteristic of natural forested wetlands recovers very slowly.

 
   FIG. D1. Restoration trajectories of eight vegetation-based indicators (rows), expressed as percentiles relative to reference site distributions, in four restored wetlands (columns). Filled symbols represent percentiles relative to Illinois Department of Transportation (IDOT) reference wetlands and open symbols represent percentiles relative to Critical Trends Assessment Program (CTAP) reference wetlands. Solid curves are best-fit trajectories relative to IDOT references and dashed curves are best-fit trajectories relative to CTAP references. Trajectory curves are not shown where R2 < 0.5. Site numbers refer to Appendix A.

 

The oldest forested wetland among our sampled wetlands was Site 16, a restored floodplain forest on former agricultural land. Even after 14 years of site development, proportion perennial species remained below the median level in reference forests (Fig. D2D). Carex richness, on average, also tended to increase slowly relative to reference wetlands (see Fig. 3), but in many sites, including Site 16 (Fig. D2C), it eventually reached levels higher than in most reference sites. After 14 years, the importance of native species remained low relative to reference forests (Fig. D2B).

 
   FIG. D2. Restoration trajectories of four vegetation-based indicators, expressed as percentiles relative to reference site distributions, in Site 16, a restored floodplain forest on the LaMoine River in western Illinois. Filled symbols represent percentiles relative to IDOT reference wetlands and open symbols represent percentiles relative to CTAP reference wetlands. Solid curves are best-fit trajectories relative to IDOT references and dashed curves are best-fit trajectories relative to CTAP references.

 

Indicators in some sites failed to follow trajectories that were adequately described by a negative exponential increase or a peaked double exponential function. For example, in Site 17, a restored floodplain forest that became heavily invaded by Phalaris arundinacea, native richness fluctuated widely over 9 years (Fig. D3A). This site was frequently disturbed by high energy flood events, which led to scouring, debris accumulation, sediment deposition, and damage to planted trees. Other indicators in this site increased slowly over time and eventually declined, but trajectories were not well described by a double exponential function (Fig. D3).

 
   FIG. D3. Restoration trajectories of four vegetation-based indicators, expressed as percentiles relative to reference site distributions, in Site 17, a restored floodplain forest on the Mackinaw River in central Illinois. Filled symbols represent percentiles relative to IDOT reference wetlands and open symbols represent percentiles relative to CTAP reference wetlands. Solid curves are best-fit trajectories relative to IDOT references and dashed curves are best-fit trajectories relative to CTAP references.

 

Although general patterns emerge when viewing the temporal trajectories of vegetation-based indicators across a large number of wetlands (see Fig. 2, Table 3) individual site trajectories reveal that not all sites follow the general patterns. Some sites follow unique trajectories attributable to external factors such as climatic events. In Site 14, a restored floodplain forest, a flood event breached a nearby levee during the second year after site construction, trapping water on the site for an extended duration and killing most vegetation. Species richness declined, but quickly recovered (Fig. D4A). Although mean C (Fig. D4B) and proportion perennials (Fig. D4D) remained low relative to reference wetlands, several water-dispersed Carex species were observed on the site in the years following the flood event, and Carex richness was higher than all reference wetlands by the fifth year (Fig. D4C).

 
   FIG. D4. Restoration trajectories of four vegetation-based indicators, expressed as percentiles relative to reference site distributions, in Site 14, a restored floodplain forest on the Sangamon River floodplain near Springfield, Illinois. Filled symbols represent percentiles relative to IDOT reference wetlands and open symbols represent percentiles relative to CTAP reference wetlands.

[Back to A019-087]