2

Outcomes of Wetland Restoration and Creation

INTRODUCTION

Underlying wetland mitigation is the assumption that it is scientifically possible for humans to recreate the structure and functions of a wetland, either by restoring a site that had previously been a wetland or by creating an entirely new wetland. The purpose of this chapter is to discuss the ecological principles of wetland creation and restoration science and evaluate the current scientific ability of practitioners to restore or create various aspects of wetland functioning in a variety of environments. The chapter is structured to answer several questions posed in the committee's Statement of Task about the ecological basis of wetland mitigation.

Is it Possible to Restore or Create Wetland Structure?
Wetland Types That Have Been Restored and Created

The committee evaluated restored and created wetlands from around the United States, including coastal and inland projects, by reviewing numerous scientific studies (see Appendix A) and by visiting several wetland mitigation sites. The following findings are based on the committee's analysis.

Many types of herbaceous wetlands have been restored or created to a condition that appears to replicate natural wetland structure, such as freshwater emergent marshes (Lindau and Hossner 1981; Niswander and



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 22
COMPENSATING FOR WETLAND LOSSES UNDER THE CLEAN WATER ACT 2 Outcomes of Wetland Restoration and Creation INTRODUCTION Underlying wetland mitigation is the assumption that it is scientifically possible for humans to recreate the structure and functions of a wetland, either by restoring a site that had previously been a wetland or by creating an entirely new wetland. The purpose of this chapter is to discuss the ecological principles of wetland creation and restoration science and evaluate the current scientific ability of practitioners to restore or create various aspects of wetland functioning in a variety of environments. The chapter is structured to answer several questions posed in the committee's Statement of Task about the ecological basis of wetland mitigation. Is it Possible to Restore or Create Wetland Structure? Wetland Types That Have Been Restored and Created The committee evaluated restored and created wetlands from around the United States, including coastal and inland projects, by reviewing numerous scientific studies (see Appendix A) and by visiting several wetland mitigation sites. The following findings are based on the committee's analysis. Many types of herbaceous wetlands have been restored or created to a condition that appears to replicate natural wetland structure, such as freshwater emergent marshes (Lindau and Hossner 1981; Niswander and

OCR for page 22
COMPENSATING FOR WETLAND LOSSES UNDER THE CLEAN WATER ACT Mitsch 1995; Wilson and Mitsch 1996; Brown and Bedford 1997) and wet meadows (Brown and Veneman 1998). Mixed results have been reported in the restoration of wet prairies and sedge meadows (Galatowitsch and van der Valk 1996; Ashworth 1997). Certain floristic assemblages, such as sedge meadows visited by the committee in the Chicago area, require extensive planting and intensive management in order to maintain the desired species composition. Thus, the technical ability to attain a prescribed species assemblage may not be affordable or practical in the long term. Shrub swamps and forested wetlands are more difficult to create or restore because of the time needed to establish mature woody plants (Niswander and Mitsch 1995; Brown and Veneman 1998; King 2000). The committee observed examples of created wetlands where tree saplings had been planted and appeared to be viable, but forest structural characteristics (e.g., stand density, stand height, basal area per tree) were quite different from those of the mature stands they were intended to replace. Planted trees are usually small in diameter, so that basal area per tree is small in comparison to natural forested wetlands. The density (trees per unit area) of planted stands is typically higher than that of natural stands because of either permit specifications or the desire to compensate for mortality. Given sufficient time, planted trees would be expected to attain basal areas comparable to those of trees in natural stands, but densely planted stands would continue to differ from natural stands unless thinned. Seagrasses and salt marshes are sometimes described as wetlands that are relatively easy to restore or create, based on a long history of mitigation involving eelgrass (Zostera marina) and smooth cordgrass (Spartina alterniflora; Simenstad and Thom 1992; 1996). Each of these natural vegetation types has the following features that are amenable to restoration or creation: One vascular plant species dominates the vegetation; that is, it forms the matrix of the ecosystem. The species in question have been well studied for growth requirements and environmental tolerances. The matrix species is readily collected and propagated and planted because of its particular clonal growth form (ramets are produced from rhizomes that are easily subdivided). These vegetation types grow in relatively wet conditions where environmental variability is buffered by ocean water of relatively constant temperature, salinity, and pH. These species are natural colonizers of bare substrates (“early succession” species).

OCR for page 22
COMPENSATING FOR WETLAND LOSSES UNDER THE CLEAN WATER ACT Transplantation can be done with considerable assurance that environmental conditions will not change substantially after planting (relatively low risk of extreme events, high probability of establishment). Microbial and animal components of these ecosystems are readily dispersed by the widely circulating ocean waters that connect distant seagrass beds and/or salt marshes. If these are indeed the characteristics that make ecosystems easier to restore or create, one can expect greater difficulty in restoring ecosystems that have the following: Several codominants. Poorly studied species. Species that are dispersal limited or that have low reproductive capacity (few seeds, heavy seeds, no vegetative reproduction). Dominants that are “late succession” species (not ready colonizers of bare substrates). Habitats that have high environmental variability, which allows exotics an opportunity to invade and which pose risks for transplantation efforts (e.g., a marsh plain that is exposed for much of the day can become dry and hypersaline the day after planting, stressing and killing thousands of seedlings (J.Zedler, University of Wisconsin, personal observation, 2000)). No aquatic connection to other wetlands and/or no corridors for dispersal. The restorability of an ecosystem is not necessarily predictable from experiences with smooth and eelgrass cordgrass, even for closely related habitat types. For example, restoration of tall forms of Pacific cordgrass has proved to be difficult because Spartina foliosa has a high nitrogen demand and grows poorly in substrates that are coarse in texture and/or low in organic matter (Langis et al. 1991; Zedler 1998). Likewise, seagrass beds in Florida that are dominated by late-succession Thalassia testudinum are difficult to restore because this species grows slowly and not densely enough to prevent erosion (Fonseca et al. 1998). Moving upslope to mid-and high-intertidal wetlands complicates restoration and creation efforts even more, as the environment becomes highly variable in soil moisture and soil salinity. In addition, newly planted vegetation becomes susceptible to a broader range of herbivores in that both aquatic and terrestrial animals can attack delicate ramets (Zedler 2001). Wetland Types That Are Difficult to Restore or Create Wetland ecosystems that require a specific combination of plant types, soil characteristics, and water supply are difficult to impossible to create from scratch. Examples include vernal pools, fens, and bogs.

OCR for page 22
COMPENSATING FOR WETLAND LOSSES UNDER THE CLEAN WATER ACT Vernal pools The term “vernal pool” is widely applied to small shallow depressions that hold water for brief periods of time. The vegetation and invertebrates are specific to the regions where such temporary pools form, and the timing of pool formation is also variable. In the forests of the northeastern states, such pools form in spring and fall. In the arid southwestern states, they form during the winter rainy season. Attempts to restore and create vernal pools in Southern California indicate the importance of locating such efforts where the substrate can support temporary ponding. Restoring or creating vernal pools begins by re-creating the topography where there is appropriate substrate (a clay layer) that seals upon wetting. There, the hydroperiod is critical to restoration or creation of native vegetation and fauna; if the hydroperiod is too long, cattails will invade and outcompete the more ephemeral vascular plants (P.Zedler, San Diego State University, personal commun., 2000). Predators, such as fish, might invade and eliminate the macroinvertebrates (such as fairy shrimp). In a long-term study of California vernal pools that were created by excavating depressions near natural pools, the hydroperiods did not converge with those of the reference systems until year 10 (Zedler al. 1993). Seeds, spores, and resting stages of the plants and animals can be vacuumed from natural pools and used to inoculate restored and natural pools, producing new vegetation highly similar to that of the natural site if the hydroperiod is correct. Even the endangered mesa mint (Pogogyne abramsii) has been reestablished at created pools of Southern California. However, few data are available on the similarity of algal and animal components of vernal pools. Perhaps the greatest difficulty is that vernal pool landscapes cannot be replaced in areas like California, where vernal pools occur on relatively flat topography, which is prime real estate. When the remaining vernal pools are destroyed, there is no place to re-create them except by increasing the density of pools in other remnant landscapes. It is unclear how overall system function is altered by increasing the density of pools while decreasing the landscape area that vernal pools occupy. Another problem is that the uplands surrounding vernal pools also are in short supply and support rare species. Thus, as the available land decreases, vernal pool creation and restoration increasingly conflict with efforts to preserve habitat for other endangered plants and animals. Fens Fens are herbaceous wetlands that develop on calcium-rich organic soils and for which groundwater seepage is an important water source. Fens are among the most species-rich wetlands in North America. Their

OCR for page 22
COMPENSATING FOR WETLAND LOSSES UNDER THE CLEAN WATER ACT plant diversity is high per unit area, yielding long species lists; approximately 100 vascular plant species might occur in a wetland, with 15 to 30 or more species per square meter (Bedford et al. 1999). Typically, a variety of species share dominance, rather than a preponderance of any one species. Oligotrophic conditions occur because fens are fed by groundwater that is low in nutrients (nitrogen and phosphorus) and high in calcium. Calcium is important, not only for its effect on pH but also because it precipitates out phosphorus, lowers nutrient availability, and thereby reduces the chances that any one species will outcompete its associates. Drainage of fens changes both the hydrology and the soil chemistry. Exposure to surface-water runoff also changes soil chemistry. In both cases, nutrients are released to the fen, allowing opportunistic species the competitive advantage. Draining also exposes any sulfides to oxygen, forming sulfuric acid, which lowers pH and prevents the restoration of calciphilic vegetation, van Duren et al. (1998) reported slightly decreased pH for drained fens in the Netherlands, where additions of lime (calcium carbonate) were ineffective in reestablishing calciphiles. Vegetation was judged nonrestorable, even when surface-water runoff was entrained in a “treatment wetland” just upstream of fen. Likely constraints on restorability were either inflowing nutrients that escaped treatment or persistent acidic soil. Bogs Bogs occur on acidic organic soil (“peat”) that develops over millennia from the accumulation of plant decomposition remains. In eight studies summarized by Johnston (1991), natural peat accretion rates ranged from 0.1 to 3.8 millimeters (mm) per year, which indicates an extremely slow rate of development. Bog drainage exposes the organic soil to aeration, accelerating decomposition and fundamentally altering organic carbon compounds in the soil. Agricultural uses of bogs further alter soil chemistry and structure through tillage, fertilizer inputs, and subsidence as soils compact and oxidize. Peat fires can oxidize in days the organic carbon that has taken centuries to accumulate. Although vegetative cover has been reestablished on bogs that have been subjected to such extreme losses and alterations of substrate, restoration of original plant communities is extremely difficult (Mitsch and Gosselink 2000). The harvesting of live sphagnum moss from bog surfaces is a small but viable industry (Johnston 1988). Research has demonstrated that natural recovery of the moss surface following harvesting takes about 20 years (Elling and Knighton 1984). In contrast, reclamation of wetlands mined for peat has been very difficult because (1) surface mining causes major changes in local hydrology, (2) peat accumulates at a very slow rate, and

OCR for page 22
COMPENSATING FOR WETLAND LOSSES UNDER THE CLEAN WATER ACT (3) the chemistry of old peat is quite different from that of surficial peat layers that support bog vegetation (Updegraff et al. 1995). The committee concludes that some types of wetlands can be restored and/or created (e.g., freshwater emergent marshes) but that others cannot (e.g., fens and bogs). Not all emergent marshes will be easy to replace. Some types (e.g., species-rich sedge meadows), some hydrological contexts (e.g., late-summer drawdowns), and some functions (e.g., biodiversity support, especially species with narrow ranges of ecological tolerance) will likely be more challenging to reproduce than others (e.g., cattail marshes, continuous flooding, and high plant productivity, respectively). Some types of wetlands will be more difficult to restore or create in certain settings (i.e., where landscape positions, specific substrates, or adjacent land uses are inappropriate). Are Wetland Functions Replaceable? Wetlands provide a number of ecological functions. The three most commonly cited wetland functions are related to water quality, hydrology, and habitat, but other functions also exist (e.g., alteration of microclimate, carbon sequestration). Some ecological functions provide human benefits, such as improvement of downstream water quality, whereas others may benefit only nonhuman organisms (i.e., wetland flora and fauna). Knowledge of the existence of wetland functions increases with increasing scientific understanding, but the perceived importance of different wetland functions changes as human values change. For example, the carbon-sequestration function of wetlands has recently assumed increased importance with our increased understanding of the role of atmospheric trace gases in global climate change (Bridgham et al. 1995). The establishment of wetland structure does not necessarily restore all the functions of a wetland ecosystem. For example, denitrification (an ecological process that benefits water quality) requires the presence of nitrate supply, a labile carbon source, anaerobic conditions, and microbial activity. Thus, a site that has wetland structure in terms of its vegetation assemblage might not provide the function of denitrification if these four requirements are not met. FIVE WETLAND FUNCTIONS The following sections discuss five major wetland functions that warrant attention in evaluating wetland restoration and creation: hydrological functions, water-quality functions, support of vegetation, support of habitat for fauna, and soil functions. These do not represent all wetland functions but do include a number of important ones that are frequently

OCR for page 22
COMPENSATING FOR WETLAND LOSSES UNDER THE CLEAN WATER ACT overlooked in wetland functional replacement. Functional assessment of wetlands proposed for evaluating the development of mitigation sites is discussed in Chapter 7. Hydrological Function Hydrology is most often cited as the primary driving force influencing wetland development, structure, function, and persistence. Consequently, establishment of the appropriate hydrology is fundamental to wetland mitigation through either restoration or creation. Hydrological processes influence water quality through nutrient inflow and outflow from the system (Gosselink and Turner 1978; LaBaugh 1986; Carter 1986; Day et al. 1988; Novitzki 1989) and by creating an environment (soil saturation) that allows anaerobic conditions to develop and reducing chemical reactions to operate. Reducing conditions in the soil causes denitrification to occur, organic matter to accumulate, and chemical transformations of phosphorus and iron that influence their solubility (Woodwell 1956; Gosselink and Turner 1978; Richardson et al. 1978; Riekerk et al. 1979; Sharitz and Gibbons 1982; Carter 1986; LaBaugh 1986; Wilcox 1988). Hydrology also influences seed distribution (Gosselink et al. 1990; Sharitz et al. 1990), seed germination, and species establishment and composition (Christensen et al. 1981; Carter 1986; Day et al. 1988; Gunderson et al. 1988; Conner et al. 1990; Gosselink et al. 1990; Sharitz et al. 1990). For many species, seedling recruitment and establishment are a complex process that is only partially understood. Of the seed available at a microsite, germination and establishment of most species are flood/inundation sensitive (Christensen et al. 1981; Day et al. 1988; Duever 1988; Gunderson et al. 1988). Very few species germinate in standing water. Inundation cessation is climatically controlled except where structural devices are used to manage hydrology. Once established, most species tolerate a fairly wide range of hydrological and other environmental conditions (e.g., light, nutrient availability, and soil-water chemistry). In palustrine nonriverine systems, the hydrological variability that results in species diversity is caused by minor variability in the microtopography (Harper et al. 1965; Christensen et al. 1981; Daniel 1981; Hardin and Wistendahl 1983; USFWS 1983; McDonald et al. 1983; Duever 1988; Gunderson et al. 1988; Titus 1990). Vegetation, in turn, influences hydrology by retarding flow (Gosselink and Turner 1978; Gosselink et al. 1990; Sharitz et al. 1990) and influencing evapotranspiration and affecting soil and water chemistry by leaf litter, transport of oxygen, and other biological processes (Day 1982, 1983; Day et al. 1988). The difficulty of restoring wetland hydrology increases as the degree of wetland degradation increases (Long et al. 1992). Many wetland resto-

OCR for page 22
COMPENSATING FOR WETLAND LOSSES UNDER THE CLEAN WATER ACT ration efforts have not established the appropriate hydrology (Kusler and Kentula 1990; Pfeifer and Kaiser 1995; Galatowitsch and van der Valk 1996). One measure of effective restoration or creation is establishment of jurisdictional hydrology. Taking a conservative stance, the U.S. Army Corps of Engineers (Corps) 1987 Wetland Delineation Manual established the 5% criterion (see Box 2–1) as the jurisdictional threshold, a quantitative value that was reaffirmed by the NRC (1995). Since wetland hydrology is fundamental to wetland structure and function, those who implement restoration and mitigation projects have also tended to take a conservative stance, that is, to err toward the wet end of the transition zone (12.5% inundation or saturation during the growing season; see Box 2–1). The structure and character of a wetland that is inundated or saturated 5% of the growing season differ greatly from those of one that is inundated or saturated 12.5% of the growing season. The consequence of this mitigation approach has been the establishment of wetlands that are much wetter than normal for the given landscape position (Cole and Brooks 2000a) or a shift from intermittently inundated or saturated to having open water (Kentula et al. 1992a). Water-Quality Improvement Water-quality functions can be mitigated but rarely duplicated. For duplication to happen, the mitigation wetland would have to be of exactly the same wetland type with the same hydrological inputs and the BOX 2–1 Duration and Timing of Inundation or Saturation Clark and Benforado (1981) divided the hydrological continuum into six functional categories: permanently inundated (inundation >2 meters (m) 100% of the time), semipermanent (inundation <2 m 75–100%), regularly inundated or saturated (25– 75%), seasonally inundated or saturated (12.5–25%), irregularly inundated or saturated (5–12.5%) and intermittently or never inundated or saturated (<5%). They suggested that areas saturated less than 5% of the growing season clearly exhibited upland hydrological characteristics and that areas saturated more than 12.5% clearly exhibited wetland hydrological characteristics. Inundation between 5% and 12.5% of the growing season represented the transition zone, with some landscapes in this category exhibiting upland characteristics and others being more characteristic of wetlands.

OCR for page 22
COMPENSATING FOR WETLAND LOSSES UNDER THE CLEAN WATER ACT same chemical (including sediment) composition. It is entirely possible for the restoration or creation site to have water-quality functions superior to those of the impact site. If the impacted wetland is a mineral or organic soil flat (Brinson's 1995 classification), it would make only a passive contribution to water quality (Evans et al. 1993), because its only water input is rain, and the water-quality function is simply to provide an area of runoff where both the surface and the subsurface drainage waters are relatively uncontaminated with pollutants. If a mitigation site is a restored riparian wetland located between a stream and a nonpoint pollution source (either urban or agricultural), the mitigation wetland will have a water-quality function superior to the impact site. However, if the impact site is a riparian wetland while the mitigation site is on a flat, the vast majority of the water-quality function of the impact site is lost. To determine the water-quality function of either the filled area or the mitigation site, it is necessary to make some assessment of both the quality and the quantity of groundwater and surface water entering the wetland (Hill 1996; Hill and Devito 1997; Bedford 1999). Different regions of the United States would need to evaluate both the quantity and the quality of water entering a wetland in order to assess the potential for water-quality improvement. Support of Vegetation Wetlands fail to support plant biodiversity when the environment is extremely hostile (e.g., extremely contaminated or hypersaline) or when one or a few species dominate the site. Monotypic vegetation can be formed by native species or exotic species. Cattails (Typha species and hybrids) are notorious for overtaking nutrient-rich wetlands (Wilcox et al. 1984), as are giant reed grass (Phragmites australis/communis), purple loosestrife (Lythrum salicaria), and reed canary grass (Phalaris arundinacea). Invasiveness is a function of both the invader and the habitat it colonizes. Plants that invade wetlands are typically species with high seed production, high germination rates, and the ability to spread vegetatively. Seedling establishment is often the limiting factor, but once established, the clone can expand, such that a clone from a single seedling could come to dominate an entire site. An additional attribute of invasive species is their ability to take up and utilize nutrients from high concentrations in the water or soil supply. In the Everglades, for example, native Cladium jamaicence is adapted to oligotrophic waters; it absorbs and stores nutrients in leaf bases. In contrast, the invasive Typha domingensis takes up nutrients and distributes them throughout the plant, growing to greater heights and biomass, thereby outcompeting Cladium where surface waters are eutrophic (Miao and Sklar 1998).

OCR for page 22
COMPENSATING FOR WETLAND LOSSES UNDER THE CLEAN WATER ACT The habitat being colonized will be more invasible if there are microsites available for seedling colonization. Small gaps in the canopy or minor soil disturbance might be all that is needed to allow seeds to establish (Hobbs and Huenneke 1992; Lindig-Cisneros and Zedler 2001). Thus, human- or animal-caused disturbances can lead to establishment events. The combination of canopy gaps and nutrient inflows can virtually guarantee the establishment and spread of invasive species. Wetland restoration and creation sites are very susceptible to species invasions when (1) they are devoid of vegetation, (2) plant canopies have multiple gaps, and (3) their water supplies are eutrophic. All three attributes characterize many mitigation sites. The ability of a site to support biodiversity is not independent of its ability to improve water quality. It is critical that the relationship between these two functions be understood if mitigation goals are to be set that are ecologically conflicting, such as maintaining high plant biodiversity and improving eutrophic water quality. Mitigation sites that receive nutrient-rich surface-water runoff are well situated to perform water-quality-improvement functions, but biodiversity-support functions may suffer in the process. Wetlands that are designed to maximize the water-treatment function typically become monotypes of invasive species within a few years, even if they are initially planted to multiple species (Kadlec and Knight 1996). Habitat Support for Fauna None of the compensatory mitigation projects visited by the committee included design and evaluation criteria for animals. Animals are almost never manipulated or introduced into wetlands, in contrast to transplants of higher plants. Even when wetland assessments involve animals, the primary consideration is waterfowl and other birds or identifiable endangered/threatened species. Evaluations do not consider the constraint that most wetland animals are incapable of overland migration if terrestrial corridors are blocked by development, highway systems, or other situations not conducive to overland movement. Many wetland animal species are also dependent on the terrestrial habitat surrounding a prescribed wetland. The importance of considering migratory pathways and upland buffers in the design of a compensatory mitigation plan is discussed in greater detail in Chapter 3. Soil Functions Soil performs a number of important functions in a wetland that are usually overlooked in wetland restoration:

OCR for page 22
COMPENSATING FOR WETLAND LOSSES UNDER THE CLEAN WATER ACT Rooting medium. Soil serves as a rooting medium for plants, providing the physical support for above-ground plant structures. Germination medium. Seed germination requires more specialized conditions than those required to sustain mature rooted plants. Germination of annuals, for example, is often promoted by a moist, temporarily exposed soil that is free of detritus. Seed bank. Seeds and rhizomes retained in the soil remain viable for months to years. Source of water and nutrients for plants. Soil is the site of water and nutrient uptake for rooted plants, even rooted plants that are submerged. The release of plant-available forms of nitrogen from unavailable organic forms stored in the soil (i.e., nitrogen mineralization) provides a constant source of nutrition to wetland plants. Habitat for mycorrhizae and symbiotic bacteria. Roots have complex relationships with soil fungi (mycorrhizae) and bacteria that enable and enhance nutrient uptake. Examples include nitrogen-fixing bacteria living symbiotically in root nodules of legumes and Alnus and vascular arbuscular mycorrhizae that associate with Salix. Some plants require the presence of specific mycorrhizal species to survive. Water-quality functions. The soil is the locus of most of the physical, chemical, and biological processes that give wetlands the ability to improve water quality. Sediment retention takes place at the soil surface. The chemical composition of the soil, such as the presence of iron and aluminum hydroxides, affects its ability to sorb phosphorus. Denitrifying bacteria dwell in the soil and depend on soil carbon as an energy source to support denitrification. Habitat for soil macrofauna. Soil-dwelling fauna sustain wading birds that probe the sediments of mud and sandflats with their long beaks. The role of soil-dwelling fauna in other types of wetlands is less well known. Conduit for groundwater. Soil permeability affects its ability to convey water. Dense, low-permeability soils may serve as aquacludes, causing water in wetlands to be perched above the regional water table. More permeable soils have higher hydraulic conductivities, allowing wetlands to have greater interaction with groundwater. Source of contaminants. Contaminants can be released from soils, particularly where the soil is landfill or has a prior history of industrial use. Soils that are high in heavy metals may release toxic forms, such as methylmercury and selenium, when creation of a wetland induces anaerobic conditions. In wetland restoration and creation projects, soil is generally viewed as merely a rooting medium for the plants that are desired (the first function listed above). The soil that has developed in situ at a wetland creation

OCR for page 22
COMPENSATING FOR WETLAND LOSSES UNDER THE CLEAN WATER ACT FACTORS THAT CONTRIBUTE TO THE PERFORMANCE OF MITIGATION SITES Wetland type and actions taken are two variables that contribute to the performance of mitigation sites. However, regardless of wetland type, the ability of a site to be restored depends in part on the degree to which it was degraded, and the degree to which it was degraded dictates, in large part, the actions that will be required to restore it (Zedler 1999). Degradation is a function of damages to both the watershed and the immediate site. Thus, restoration of a cattail marsh or open pond will be easier than a vernal pool, fen, or bog. And the creation or restoration of a cattail marsh or pond will be easiest at a site that is not too degraded in a watershed that is not too degraded. That is, the site should have a soil substrate (not bare rock as in a gravel pit); it should not be so contaminated that vegetation will not grow (as can be true where oil or pesticides have been spilled); and it should occur in a watershed that still retains the region's natural hydrology (e.g., groundwater should not be substantially depleted), as well as space and a topographic setting appropriate for a wetland to persist in perpetuity. Finally, the actions taken to restore or create a wetland should be appropriate to the type and degree of on-site degradation. If hydrology, soils, vegetation, and fauna have all been degraded, some attention to each of these classes of factors will likely be needed. Chapter 7 outlines appropriate guidelines for restoring or creating wetlands that are ecologically self-sustaining. The committee concludes that in a degraded wetland situation, many wetland functions are difficult to restore to their pre-disturbance condition. The ability to replace wetland functions depends on the particular function, the restoration-creation approach used, and the degree of degradation at the compensation site. How Is the Likelihood of Achieving Wetland Restoration-Creation Goals Affected: By the Hydrological Regime? Wetlands are transition areas between water (wet 100% of the time) and uplands (seldom wet). In a qualitative sense, wetland hydrology has been defined by the National Research Council (NRC 1995) as recurrent, sustained inundation or saturation at or near the surface at a duration and frequency to support the development of diagnostic wetland features of hydric soils and hydrophytic vegetation. The NRC recognized that this definition is too broad to be applied directly in regulatory practice and so adopted quantitative criteria that focus on frequency, timing, and duration of inundation or soil saturation. In many cases, mitigation and resto-

OCR for page 22
COMPENSATING FOR WETLAND LOSSES UNDER THE CLEAN WATER ACT ration have required establishment of wetlands that are inundated or saturated more than 12% of the time (“seasonally inundated or saturated”; see Box 2–1). In reality, many impacted wetlands satisfy the saturation criteria closer to 5% of the time (Cole and Brooks 2000a). Thus, mitigated sites tend to be wetter than impacted sites (see Chapter 6 for more discussion on this consequence). Our ability to restore or establish a particular hydrological regime is highly dependent on the desired wetland type. Wetlands that are surfacewater-driven are, as a general rule, easier to establish than groundwater-dependent wetlands, but outcomes in both cases are highly dependent on landscape position. These wetland types are typically rainfall- or surface-runoff-driven, although some depressional areas, such as prairie potholes and the southeastern coastal plain, may rely on significant groundwater inflow. Hydrological regime is much more difficult to restore to an original state in highly modified hydrological systems (e.g., dams and levees). Hydrological restoration of prior converted cropland has been a generally successful process in part because the original conversion from wetland to agriculture resulted in relatively minor alteration of the natural hydrology. Wetland hydrology can often be restored to prior converted cropland by disrupting the artificial drainage system. Cooper et al. (1998) reported the restoration of hydrology to a Rocky Mountain fen where damage had been minor (i.e., installation of a drainage ditch). Monitoring data demonstrated that hydrology returned soon after the ditch was blocked. The return of natural water levels, however, might not restore soil chemistry or the ability to support calciphilic species, as discussed earlier. Important groundwater parameters are often inadequately characterized at reference and mitigation sites (Hunt 1996; Hunt et al. 1999). For riparian and riverine wetlands, the ability to restore hydroperiods depends on the degree to which streamflows have been modified (see Coyote Creek case study, Appendix B). Streams may have been channelized to convey storm water, and the need to protect upstream lands might preclude the restoration of natural flood pulses. In conclusion, surface-water-dominated wetlands are easier to restore or create than groundwater-dominated wetlands, but achieving a natural hydrological regime is always a challenge. Hydrological regime is much more difficult to restore to the original state in highly modified hydrological systems (e.g., with dams and levees). By Wetland Size? Large natural wetlands are rarer, and their size imparts additional value to some functions (e.g., habitat for animals with large home ranges). Conversely, small isolated wetlands play a crucial role in the biodiversity

OCR for page 22
COMPENSATING FOR WETLAND LOSSES UNDER THE CLEAN WATER ACT of other wetland-dependent fauna, such as amphibians (Semlitsch 2000). There is no question that wetland size has an impact on water quality (Chescheir et al. 1991; Daniels and Gilliam 1996; Gilliam et al. 1997). Larger wetlands have a greater capacity to assimilate constituents in the inflow; however, the larger the wetland, the smaller the pollutant removal per unit area of wetland. Pollutant removal frequently follows a first-order decay curve with regard to residence time in the wetland. For a given input of water, the larger the wetland, the more pollutant will be removed. However, since each additional increment of land removes less pollutant than the previous increment, the removal per unit area of wetland is lower. Shape may also influence the effectiveness of the wetland for pollutant removal. For functions such as water quality and nutrient retention, edge interface with stream or upland is probably more important than area. For example, pollutant removal by a 50-foot (ft)-wide wetland buffer beside a 100-ft-wide stream would remove far more contaminants than a 100-ft-wide buffer beside a 50-ft-wide stream. That would be true for water and pollutants entering the wetland from land upslope from the stream. However, that might not be true for water entering the wetland from stream bank overflow. However, for water-quality purposes, many small wetlands would be more effective than one large wetland covering the same area. The committee concludes that wetland size affects wetland functions. Thus, replacement area should be proportional to the area required to replace the functions lost. By Wetland Place in the Landscape? Riverine and slope wetlands are more difficult to restore or create than are depressional wetlands (Gwin et al. 1999; Shaffer et al. 1999). Position in the landscape affects a number of wetland functions, such as water quality (Johnston et al. 1990; Evans et al. 1993) and biodiversity (Poiani et al. 2000). Wetland place in the landscape is generally not considered a mitigation performance standard. The location of mitigation wetlands may be limited by the availability and cost of land. Recently, much attention has been focused on the need to define hydrological equivalence in the landscape (Bedford 1996). Wetlands occur in a variety of physical settings, including coastal lowlands, topographic depressions, broad flats on interstream divides, the base of slopes, and topographic highs with little slope (Winter and Woo 1990). Location in the landscape influences geological characteristics such as slope; thickness and permeability of soils; and the composition, stratigraphy, and hydraulic properties of the underlying strata, all of which influence surface and subsurface flows of water. Quite simply, hydrology is the driv-

OCR for page 22
COMPENSATING FOR WETLAND LOSSES UNDER THE CLEAN WATER ACT ing force influencing wetland development, structure, function, and persistence. In conclusion, riverine and slope wetlands are more difficult to mitigate than are depressional wetlands. By the Ecoregion in Which a Wetland Occurs? Ecoregions are quite diverse, and it was difficult for the committee to generalize about the effect of ecoregion on the likelihood of achieving wetland restoration-creation goals. Created wetlands in areas of variable precipitation, one factor that distinguishes ecoregions, may not meet the jurisdictional definition of a wetland every year. Such areas are more likely to occur in arid ecoregions, although periods of drought and temporal variations in precipitation can occur in any ecoregion. Innate temporal variability of wetlands due to climatic variation is not the same as mitigation failure. The committee concludes that created wetlands in regions of variable precipitation may not meet the jurisdictional definition of a wetland every year, increasing the risk of noncompliance with performance standards. Thus, the mitigation design should recognize and accommodate hydrological variability and extremes caused by climate. Wetland restoration and enhancement should be preferred over creation in such areas. By the Kinds of Plants Present? An important general consideration of wetland design is whether plant material is going to be allowed to develop naturally from some initial seeding and planting or whether continuous horticultural selection for desired plants will be imposed. To develop a wetland that will ultimately require low maintenance, natural successional processes need to be allowed to proceed. For forested wetlands, an initial period of invasion by undesirable species might be temporary if proper hydrological conditions are imposed and if trees shade out early invaders. One strategy is to introduce, by seeding and planting, many of the available species to allow natural processes to sort out the species and communities over time. Selective weeding may be necessary in the beginning or throughout the life of the wetland if aggressive exotic vegetation persists. Preferably, the system can sustain itself through its own successional patterns. Otherwise, labor-intensive management, which is never desirable in a compensatory mitigation wetland, will be needed. In some cases, survival of specific plantings is used to evaluate compliance. Wetlands that are readily invaded by exotics or prove to support undesirable monotypes over long periods require greater attention to planting and an establishment-phase exotic control program (see Boxes 2–2 and 2–3).

OCR for page 22
COMPENSATING FOR WETLAND LOSSES UNDER THE CLEAN WATER ACT BOX 2–2 Seeding Versus Natural Recruitment Reinartz and Warne (1993) found that early introduction of a diversity of wetland plants may enhance the long-term diversity of vegetation in created wetlands. The study examined the natural colonization of plants in 11 created wetlands in southeastern Wisconsin. The wetlands under study were small isolated, depressional wetlands. A 2-year sampling program was conducted for the created wetlands, aged 1– 3 years. Colonized wetlands were compared with five seeded wetlands where 22 species were introduced. The diversity and richness of plants in the colonized wetlands increased with age, size, and proximity to the nearest wetland source. In the colonized sites, Typha spp. comprised 15% of the vegetation for 1-year wetlands and 55% for 3-year wetlands, with the possibility of monocultures of Typha spp. developing over time in colonized wetlands. The seeded wetlands had a high species diversity and richness after 2 years. Typha cover in these seeded sites was lower than that in the colonized sites after 2 years. In conclusion, wetland mitigation designs should include plantings (e.g., sedges over cattails). Unless actively controlled at the outset, exotic and weedy plant species often dominate restoration sites. Species richness is often low in created wetlands. By the Kinds of Animals Present? Natural freshwater wetlands support among the highest levels of regional species' diversity and population densities of fauna in North BOX 2–3 Does Wetland Planting Help Self-Design? In a multiyear study of the effect of plant introduction on ecosystem function, researchers at the Olentangy River Wetland Research Park in Ohio found that a planted wetland and an unplanted wetland converged in most functions (eight biological measures; eight biophysiochemical measures) in 3 years (Mitsch et al. 1998). Continued studies showed a persistence of the planted vegetation in that basin but dominance by Typha in the naturally colonizing basin, with differences in function between the two basins. The planted wetland had more plant communities but had 50% lower net primary productivity, higher summer water temperature, and lower macroinvertebrate diversity than did the naturally occurring wetland 6 years after planting (Mitsch et al. 1999).

OCR for page 22
COMPENSATING FOR WETLAND LOSSES UNDER THE CLEAN WATER ACT America. Several biological and landscape considerations are critical to the sustainability of such species, and a variety of factors must be considered in attempting compensatory mitigation for loss of animal-support functions. Created wetlands are generally not designed to meet the needs of animals found in the impacted wetland; hence, animal species ' richness is often low in mitigation sites. Amphibians are perhaps the best-studied group of organisms from the standpoint of wetland dependency both in terms of association with a particular wetland and the landscape pattern of wetland connectivity. The importance of amphibians as a major component of wetland biodiversity, the significance and necessity of peripheral terrestrial habitat to their existence, and the requirement for wetland interconnectivity in a landscape have been documented for a variety of species and regions in North America (Berven and Gill 1983; Berven and Grudzien 1990; Berven 1990; Pechmann et al. 1991; Dodd 1992, 1993, 1995; Gibbs 1993; Semlitsch et al. 1996; Snodgrass et al. 1999; Madison 1997; Semlitsch 1998, 2000; Semlitsch and Bodie 1998; Lamoureux and Madison 1999). Biological dynamics of animal populations are discussed in Chapter 3. The committee concludes that for compensatory mitigation of a wetland to be effective for all affected fauna, the biological dynamics must be evaluated in terms of the populations present and the ecological requirements of the species, which include metapopulation aspects that are affected by the relationship of the wetland to other wetlands in the local system. By Time? After more than 2 decades of compensatory mitigation, there are now thousands of hectares of restored and created wetlands in the United States. Yet only a few studies have analyzed how various ecosystem components have changed over time, and even fewer describe ecological performance over more than 5 years. Knowing how rapidly an ecosystem matures, or fails to mature, is important for learning how to improve future projects. A 3- to 5-year monitoring period is a common permit requirement. Five years may be enough time for herbaceous plant cover to reach a peak (see Figure 2–1), but trees obviously take longer than herbaceous plants to reach peak biomass. Biomass, however, may not be equivalent between the restored or created wetland and the natural reference ecosystems. For example, Brown and Veneman (1998) analyzed 68 paired mitigation sites in Massachusetts and, using criteria for establishment for equivalency, reported that plant communities were not equivalent to reference systems after 13 years.

OCR for page 22
COMPENSATING FOR WETLAND LOSSES UNDER THE CLEAN WATER ACT FIGURE 2–1 Percent plant cover on created or restored coastal wetlands on the Atlantic and Gulf of Mexico (GOM) coasts. A second-degree polynomial fit of the data is shown. Note that 100% cover was achieved in most of the Atlantic wetlands within 5 years. SOURCE: Adapted from Matthews and Minello (1994). The animal components of communities may develop at quite different rates than do plants at mitigation sites. Mobile species may migrate in and colonize with the first floodwaters. Even with fish, however, simple analyses of population structure indicate that the trajectory from colonization to stability is not equal or smooth among population types. For one study, Rulifson (1991) surveyed a created salt marsh constructed as a mitigation site in North Carolina. Although this study was only for a single site, the findings are important. Rulifson found that the number of species captured with one piece of gear (trawl) became more like the number of species in the reference marsh within 20 months and after 42 months for fish collected with shallower-water gear (Wegener Ring). Neither the number of organisms captured with each gear type nor the number of similar species was stable after 4 years. The equilibrium position of various fisheries parameters was not reached within 4 years of the man-

OCR for page 22
COMPENSATING FOR WETLAND LOSSES UNDER THE CLEAN WATER ACT TABLE 2–1 Summary of Results from Study of a Created Salt Marsh Constructed as a Mitigation Site in North Carolina (1991) Comparison of Finfish in a Constructed Wetland Compared with Two Reference Sites Months to Reach Reference Site Conditions Trend After Equivalency First Reached Number of organisms Trawl 8 Falling Wegener Ring 28 Rising Number of species Trawl 18 Stable Wegener Ring 43 ? Species similarity Trawl 0–12 Declining Wegener Ring 0–12 Declining SOURCE: Adapted from Rulifson (1991). agement action but then continued to change. In other words, the equilibrium was obtained but not equivalency (see Table 2–1). Zedler (1990) noted that the average number of individual fish species and the average number of individuals caught at a single reference site and a wetland creation project in Southern California were reached within 5 years. However, the similarities of fish and benthic species between the reference site and mitigation sites were only 58% and 54%, respectively. Results from a variety of trajectory analyses of soil, plant, and animal communities for mitigation sites are given in Appendix C and summarized in Table 2–2. Plant cover and biomass for some marshes may reach TABLE 2–2 Time Toward Equivalency for Soil, Plant, and Animal Components in Wetland Restoration Projects Compared with That of Natural Reference Wetlands Structural Component Number of Sites Range (years) Soils 17 >3 to >30 Wetland plants Cover or biomass 7 3 to >20, or never Species composition 3 >5 to >10, or never Below-ground biomass 2 10+, or never Fish and fisheries 9 >2 to 10 Marsh fisheries 14 1 to >17, or never Birds 7 <3 to >15, or never

OCR for page 22
COMPENSATING FOR WETLAND LOSSES UNDER THE CLEAN WATER ACT an equilibrium in 3 years, as discussed above, but species composition may take 10 years or longer to stabilize, if at all. Trees, obviously, will not reach an equilibrium until at least the lifetime of the tree. The trajectories for soils, plants, and animals are not the same (e.g., Zedler and Callaway 1999; see Figure 2–2). In contrast to herbaceous vegetation, soil development may be quite slow (Craft et al. 1999), taking from 3 years to 30 years to reach equilibrium, if equilibrium is realizable. The fish and bird components may reach equivalency in as little as 2 years for some species. The trajectory may never reach an equilibrium or the equilibrium conditions may not equal those at the reference sites. However, because there has FIGURE 2–2 Long-term data for salt marshes constructed in San Diego Bay. Values are in relation to the adjacent natural salt marsh. The site was required to provide self-sustaining tall cordgrass stems for nesting by an endangered clapper rail. SOURCE: Adapted from Zedler and Callaway (1999).

OCR for page 22
COMPENSATING FOR WETLAND LOSSES UNDER THE CLEAN WATER ACT been virtually no research that compares the relative similarity of undisturbed reference sites, it is hard to make firm conclusions on this point. In their analysis of reference site selection, White and Walker (1997) found that there is so much variation on a regional scale that there may never be a perfect match for a site to be restored. They propose that a conceptual restoration model should be one of interpolation among multiple sites and sources of information, including temporally. Finally, Neiring (in Kusler and Kentula 1990) posits that the end point of the successional process cannot be predicted because it is not an orderly process; that is, we cannot predict what exact species will be at the “end” of the process. In summary, it appears that there is no general trajectory for the development of wetland ecosystems or individual components (structural or functional attributes) of a single watershed. The significance of these results is not that equivalency among reference and newly managed environments is not reached or that mitigation efforts should not be done. These results demonstrate that (1) ecological equivalency may not be reached within a few months or for several years or even decades, depending on the attribute that is of interest; (2) the ecosystem does not move smoothly to an equilibrium or at the same rate for all components; and (3) some components, including ones identified as important in permits currently being issued, may never reach equivalency with the natural reference wetland. An obvious conclusion from these results, besides the general paucity of scientific analyses, is that the generally observed 5-year limit on monitoring is insufficient when evaluating whether a site has achieved parity with a reference system. Further, the amount of mitigation required should be based on the amount needed to fully offset the permitted wetland losses. To accomplish that, mitigation ratios (area of mitigation to area lost) will often need to be increased to achieve functional equivalency, rather than simply matching wetland area. The above review of wetland restoration and creation outcomes has not differentiated projects on the basis of their starting conditions. Some of the projects discussed occurred at sites where damages were relatively minor (e.g., Cooper et al. 1998); others were at highly altered sites with artificial (dredge spoil) substrate (e.g., Zedler and Callaway 1999). As discussed earlier, it is very likely that the ability to achieve desired outcomes is a function of the degree of site degradation. Degradation, in itself, is a complex concept, involving not only the local site but also its watershed. Zedler (1999) suggests that generalizations are not easily drawn among restoration efforts involving sites with different degrees or types of degradation, suggesting instead that ecologists seek predictability of outcomes in a matrix of situations, where degree of degradation is one axis and the type of restoration effort is the other. The general pattern

OCR for page 22
COMPENSATING FOR WETLAND LOSSES UNDER THE CLEAN WATER ACT is likely to be that outcomes are more predictable where damages are minimal and restoration efforts most intensive. Badly damaged wetlands and wetland creation sites are less likely to match reference wetlands, especially if restoration efforts are minimal. In conclusion, the literature and long-term trajectories reported therein suggest that wetland restoration and creation sites do not often achieve functional equivalency with reference sites within 5 years; indeed, up to 20 years may be needed for functional attributes to be determined or assessed correctly. RECOMMENDATIONS On the basis of its evaluation of wetland structure and function, the committee makes the following recommendations for compensatory mitigation: Avoidance is strongly recommended for wetlands that are difficult or impossible to restore, such as fens or bogs. The science and technology of wetland restoration and creation need to be based on a broader range of studies, involving sites that differ in degree of degradation, restoration efforts made, and regional variations. Predictability of outcomes should then improve. All mitigation wetlands should become self-sustaining. Proper placement in the landscape to establish hydrogeological equivalence is inherent to wetland sustainability. Hydrological variability should be incorporated into wetland mitigation design and evaluation. Except for open-water wetlands, static water levels are not normal. Because of climatic variability, it should be recognized that many wetland types do not satisfy jurisdictional criteria every year. Hydrological functionality should be based on comparisons to reference sites during the same time period. Because a particular floristic assemblage might not provide the functions lost, both restoration of community structure (e.g., plant cover and composition) and restoration of wetland functions should be considered in setting goals and assessing outcomes. Relationships between structure and function should be better known. The biological dynamics should be evaluated in terms of the populations present in reference models for the region and the ecological requirements of those species. Mitigation projects should be planned with and measured by a broader set of wetland functions than are currently employed.