Creating riverine wetlands: Ecological succession, nutrient retention, and pulsing effects

Abstract

Successional patterns, water quality changes, and effects of hydrologic pulsing are documented for a whole-ecosystem experiment involving two created wetlands that have been subjected to continuous inflow of pumped river water for more than 10 years. At the beginning of the growing season in the first year of the experiment (1994), 2400 individuals representing 13 macrophyte species were introduced to one of the wetland basins. The other basin was an unplanted control. Patterns of succession are illustrated by macrophyte community diversity and net aboveground primary productivity, soil development, water quality changes, and nutrient retention for the two basins. The planted wetland continued to be more diverse in plant cover 10 years after planting and the unplanted wetland appeared to be more productive but more susceptible to stress. Soil color and organic content continued to change after wetland creation and wetlands had robust features of hydric soils within a few years of flooding. Organic matter content in surface soils in the wetlands increased by approximately 1% per 3-year period. Plant diversity and species differences led to some differences in the basins in macrophyte productivity, carbon sequestration, water quality changes and nutrient retention. The wetlands continued to retain nitrate–nitrogen and soluble reactive phosphorus 10 years after their creation. There are some signs that sediment and total phosphorus retention are diminishing after 10 years of river flow. Preliminary results from the beginnings of a flood pulsing experiment in the two basins in 2003–2004 are described for water quality, nutrient retention, aboveground productivity, and methane and nitrous oxide gaseous fluxes.

Introduction

Wetlands are being created and restored at great frequency around the world both as “mitigation” wetlands that are meant to replace or compensate for wetland habitat loss and as wetland treatment systems for improving water quality. As important as the functions of wetlands are for providing values such as habitat structure and water quality improvement, there are scarcely any long-term data from thousands of created and restored non-tidal freshwater wetlands. In contrast, there are some results of long-term (>10-year) in the literature for created salt marshes and tidal freshwater marshes (Craft et al., 2002, Craft et al., 2003, Leck, 2003). Further, few studies have investigated how macrophyte diversity and cover affect ecosystem function in created and restored wetlands, despite the frequent use of macrophyte cover and species requirements as determinants of legal and ecological success of these wetlands in mitigating wetland loss (Mitsch et al., 1998, NRC, 2001).

This paper presents the results of 10 years of study of two created riverine wetlands developed “from scratch” on non-wetland soils and maintained for those 10 years by pumped river water. These experimental wetlands allowed the simultaneous long-term study of three different aspects of newly created wetlands: (1) measuring the importance of wetland plant introduction on ecosystem function; (2) investigating the time it takes for the development of hydric soils at a site where no hydric soils previously existed; and (3) determining the long-term patterns of water quality changes of a flow-through wetland as it develops from open ponds of water to vegetated, hydric-soil, marshes.

For the plant introduction study, our hypothesis in this study was that planted and unplanted wetlands would initially diverge in structure but eventually converge in structure. Three major questions in self-design were also part of this study:

• Does human introduction of propagules have any measurable effect on ecosystem function?

• At what rate and to what degree will a diverse biological community develop in newly created hydrologically open wetland ecosystems in which little biological life initially existed?

• Does biodiversity affect ecosystem function?

We believe that introducing macrophytes in created and restored wetlands could be regarded as a human-induced switch that changes the resiliency of the wetlands and that this resiliency or lack thereof may be an important factor in allowing systems to cross thresholds that alter ecosystem function, recovery trajectory, and successional patterns. Because created and restored wetlands have rarely been monitored beyond the 5-year periods, the importance of species introduction in general and macrophyte planting in particular on ecosystem function remains poorly understood. Changes that are long-term in nature, e.g. woody plant invasion or soil carbon increases, can have significant effects on shorter-term dynamics of wetlands, e.g. algal and macrophyte vegetation dynamics. For example, Fig. 1 illustrates four possible ecosystem states possible for wetlands with two different initial conditions. In one case, a wetland is planted (high propagule introduction) and in the second case, a wetland is left to natural colonization (low propagule introduction). Achieving a healthy ecosystem state for shorter-term variables is possible in both situations (Pathways 1 and 2) but unexpected threshold shifts can occur (Pathways 3 and 4); these short-term dynamics couple with long-term ecosystem processes (Carpenter and Turner, 2001).

We have 10 years detailed data on ecosystem structure and function in two 1-ha wetland basins maintained under strict and well documented hydrologic conditions that, when analyzed in a retrospective analysis, will reveal: (1) the importance of macrophyte introduction and subsequent macrophyte community diversity on ecosystem function and (2) the effects of this diversity on ecosystem resiliency to stressors and coupling of fast and slow turnover dynamics. Our hypothesis is that in wetland ecosystems with essentially identical forcing functions, the introduction of propagules can lead to alternative ecosystem states. These alternate states may or may not provide the functions necessary to meet restoration and regulatory needs for wetlands or provide the right conditions for water quality improvement.

Studies that attempt to link wetland function with structure are often done at inappropriate spatial and temporal scales to assist those responsible for the management of wetland landscapes. While there is no single optimum scale for ecosystem experimentation, it is well know that it is easier to apply statistical methods successfully when many small replicated systems or plots are used (Carpenter, 1998, Carpenter et al., 1998b). As a result, these scales are often chosen to make inferences about wetland management. Unfortunately, when results of small-scale, short-term studies are applied to full-scale conditions, conclusions are questionable at best. For example, Engelhardt and Ritchie (2001) manipulated seventy 1.5-m diameter wading pools with one, two, and three species of the submersed pondweed (Potamogeton spp.) over one growing season and found that higher algal biomass and higher phosphorus uptake occurred in the pools with highest macrophyte species richness. They concluded that higher species richness created up to 25% higher algal biomass and caused 30% more phosphorus uptake and thus would support more wildlife and fish. They further concluded that a wetland with high richness or diversity due to disturbance might better “sustain ecosystem functioning and promote the services of those wetlands to humans.” We consider this extrapolation of short-term results from 1.5-m pools to national wetland policy in a prestigious scientific journal to be infelicitous. Extrapolation of short-term microcosm experiments across temporal and spatial scales for adaptive management of ecosystems is questionable at best.

Alternatives to the replicated small-scale mesocosms for wetland study are large-scale, long-term whole ecosystem studies that include more components of the ecosystem. Examples of such studies are described by Mitsch and Day (2004) and include systems such studies of tropical rain forests (Odum and Pigeon, 1970); forested watersheds (Likens et al., 1977); lakes (Schindler, 1977, Schindler et al., 1997, Carpenter et al., 1996, Carpenter et al., 1998a), and wetlands (Odum et al., 1977, Mitsch et al., 1995). Whole-ecosystem studies are often criticized because the size, cost, and logistics do not allow for much if any replication. Yet the lack of replication can be compensated for by the decrease in variances that has been shown to result when large-scale experiments are used. Results from large-scale experiments are less stochastic and thus more homeostatic, and often allow for the demonstration of ecosystem properties that otherwise would not appear in smaller scale experiments (Pomeroy et al., 1988, Odum, 1990, Carpenter et al., 1995, Carpenter, 1998). Kemp et al. (2001) illustrate in plankton experiments that as the scale of the experiment increases, the relative variance decreases. This may be a general principle and if it is, there is less need for great numbers of replications with increasing size of the experiment. There needs to be more understanding and acceptance of large-scale experiments and observations in the literature, if for no other reason than that they serve as a check on theories and management recommendations being published from smaller-scale studies chosen primarily because of elegant replications and statistics.