6.0 Non-woody (Herbaceous) Vegetation

Impacts on Quality of Inland Wetlands of the United States:
A Survey of Indicators, Techniques, and Applications of Community Level Biomonitoring Data
Excerpts from Report #EPA/600/3-90/073
(now out of print)

This discussion concerns communities dominated by mosses, lichens, liverworts, ferns, sedges, etc., and includes emergent, floating-leaved, and submersed forms. These taxa probably have been studied more than any other wetland taxa.

6.1 USE AS INDICATORS

Following are discussions of the community-level responses of herbaceous vegetation to various stressors.

Enrichment/eutrophication. Species richness of herbaceous plants, particularly emergent species, can increase with moderate enrichment (Graneli and Solander 1988). However, severe enrichment drastically shifts community structure, and can decrease species richness (e.g., Lachavanne 1985, Lind and Cottam 1969, Tilman 1987, Hough et al. 1989, Toivonen and Back 1989). This might be particularly true of macrophyte communities in flowing water wetlands (e.g., Pip 1987), where nutrients otherwise tend to be less limiting than in most standing water (basin) wetlands. Duarte and Kalff (1988), studying lacustrine macrophytes, similarly found that the effect of fertilization was influenced by hydrologic energy (e.g., wave action).

The greatest richness of emergent plants has been reported to occur when standing biomass of the community is less than 1000 g/m2 in British wetlands, 400-500 g/m2 in Netherlands wetlands (Vermeer and Berendse 1983), and 60-500 g/m2 in Ontario wetlands. If a goal is to maintain within-wetland species richness, the particular nutrient loadings that result in a desired biomass might be calculated from empirical data (e.g., Duarte and Kalff 1986, Duarte et al. 1986) to derive very approximate criteria for nutrient loadings, and perhaps, with further testing, for other factors that can increase plant biomass (e.g., thermal warming, hydrologic regime). However, the numeric ranges just given are probably less valid for wetlands that are grazed or subject to other significant vegetation removal processes.

Changes in composition and growth of herbaceous communities as a probable result of increased nutrients have been reported by many ecologists, including:

Guntenspergen 1984, Haslam 1982, Kullberg 1974, Jensen 1979, Kadlec and Hammer 1980, Klowsowski 1985, Kohler et al, Mahoney 1977, Pringle and van Ryswyck 1968, Schwartz and Gruendling 1985, Seddon 1972.

An important regional impact of excessive enrichment is that small, regionally rare plant species (that often characterize infertile wetlands or wetlands whose chemistry reflects weak buffering) are often out-competed by large, regionally common species (Day et al. 1988, Moore et al. 1989). Insectivorous plants, quillworts, many species that typify fen wetlands, and some orchids and mosses that typify oligotrophic wetlands, are particularly sensitive to enrichment, either airborne or waterborne (Moore et al. 1989, Roelofs 1983, 1986, Schuurkes et al. 1986). Percent cover of the dominant peat-forming mosses of bogs can probably be reduced by atmospheric nitrogen deposition rates of 4.3 g/m2/yr, but not 2.0 g/m2/yr (Ferguson et al. 1984). However, because great variability exists among tolerances of moss species, a limit of 2.0 and possibly (in oligotrophic wetlands) 1.0 g/m2/yr has been suggested (Schuurkes et al. 1987, Liljelund and Torstensson 1988).

Submersed and floating-leaved or mat-forming species usually respond more strongly to enrichment than do emergents (e.g., Ozimek 1978, Shimoda 1984), because the former obtain nutrients directly from the water column, whereas the latter obtain them from sediments. In many regions, vascular floating-leaved plants such as pondweed (Nuphar), duckweed (Lemna), and water-meal (Wolffia) become more prevalent with increasing enrichment (e.g., Bevis and Kadlec 1990, Burk et al. 1976, Ewel 1979, Kadlec et al. 1980), and in severely eutrophic lakes, emergent species may survive as floating mats (Graneli and Solander 1989). Species shifts may be less immediate or noticeable when moderate amounts of nutrients are added to wetlands that are already eutrophic, because high microbial populations that characterize such environments can be highly effective at first in competing for the nutrients (e.g., Richardson and Marshall 1986). With extreme enrichment, submersed macrophytes can eventually decline, probably as a result of being shaded out by algae (e.g., Mulligan et al. 1976, Phillips et al. 1978), and emergent species may increase.

Among emergent species, extreme eutrophication causes decreased species richness because (a) turbidity of phytoplankton blooms shades out many submersed species, and (b) as phytoplankton decays, resultant oxygen deficits in bottom sediments probably stress the most sensitive rooted species (e.g., Hartog et al. 1989). Of the emergent species, cat-tail (Typha) and common reed (Phragmites) often dominate enriched wetlands and may be the least sensitive to the initial stages of eutrophication (e.g., Kadlec 1979, Hartland-Rowe and Wright 1975, Kadlec 1990, Kadlec and Bevis 1990, Moore et al. 1989). Cat-tail biomass and production respond to annual fluctuations in nitrate, making cat-tail a successful opportunist capable of dominating wetlands that have erratic inputs of nutrients (Davis 1989). Although Phragmites can exist without any obvious sign of harm in wetlands with at least 6 mg/L phosphorus and 10 mg/L nitrogen (Ostendorp 1989), massive die-offs of this species in European wetlands have been attributed by some to excessive enrichment (Hartog et al. 1989, Ostendorp 1989). Another emergent plant--manna grass (Glyceria grandis )--increased in dominance in an Ontario wetland subjected to treated effluent (Mudroch and Capobianco 1979).

In Michigan, moderately eutrophic lakes were dominated by Ceratophyllum demersum, Utricularia vulgaris , and Cladophora fracta (Hough et al. 1989). In England, Potamogeton pectinatus, Myriophyllum spicatum, and Hippuris vulgaris dominate in highly eutrophic waters (Butcher 1946, Seddon 1971). However, Kadlec et al. (1980) and Mulligan et al. (1974) found Myriophyllum declined under increasing fertilization, along with Ceratophyllum demersum, Polygonum and Utricularia. Many wetland plant species are categorized by their nutrient-level preferences, and thus as their potential as indicators of eutrophication, in reports by Ellenberg Jeglum 1971, Moyle 1945, Pip 1979, Stewart and Kantrud 1972, Swindal and Curtis 1957, and Zoltai and Johnson 1988.

Even the submersed types of herbaceous vegetation appear a poorer indicator of eutrophication than are algal communities, which respond more quickly (Crumpton 1989). Neither macrophyte nor algal taxa are reliable indicators of moderate enrichment in naturally enriched waters, e.g., minerotrophic fens, wetlands in karst limestone regions (Hellawell 1984, Strange 1976).

Organic loading/reduced DO. Existing literature often does not adequately distinguish the effects on herbaceous plants of organic loading/reduced DO, from the effects of nutrients (discussed above) or inundation (discussed below).

At least in the short term, the biomass of herbaceous plants generally increases with moderate additions of wastewater. In acidic, oligotrophic wetlands (e.g., bogs), species richness may increase (e.g., Guntenspergen 1984). Community components with short turnover times, such as aboveground biomass and leaf area of annual plants, can respond most sensitively (e.g., Brown 1981, Odum et al. 1984).

Aggressive, introduced annuals sometimes replace native perennial species (e.g., Finlayson et al. 1986). While the occurrence of rarer, perennial species is often correlated with specific chemical conditions, the occurrence of aggressive, common species often is not (Pip 1979). Populations of such species tend to be more plastic in their response to wastewater enrichment (e.g., Guntenspergen 1984).

Over longer periods of time and/or excessive loading, wastewater additions may result in stress from low dissolved oxygen, increased hydrogen sulfide, and excessive accumulation of sediment organic matter. These conditions can selectively inhibit certain plant taxa (Barko and Smart 1983), particularly those that are unable to translocate oxygen to their roots (Brennan 1985). While cattail (Typha ) require only trace amounts of dissolved oxygen for germination (Leck and Graveline 1979), bud development is more successful in reeds (Phragmites ) if flooded soils are aerated (Haslam 1973), as is sprouting of purple nutsedge (Cyperus rotundus ) (Al-Ali et al. 1978).

Morgan and Philipp (1986) surveyed a host of New Jersey streams and listed 22 species found only in streams that, based on their location and limited chemical sampling, were assumed to be "polluted." The researchers found 18 only in "unpolluted" streams, and 21 in both types. Callitirche heterophylla, Ludwigia palustris, Polygonum punctatum, Potamogeton epihydrus, and Sparganium americanum were locally dominant only in polluted streams, and Sagittaria englemanniana, Scirpus subterminalis , and Vaccinium macrocarpon were dominant only in unpolluted streams. Polluted sites, with high nitrate and pH, had a higher percentage of non-indigenous species, vines, and herbaceous (vs. woody) plants. Vines and other low-growing species also were found by Nilsson and Grelsson (1990) to dominate riverine sites with intermediate accumulations of organic matter (i.e., 100-200 g/m2 leaf litter), whereas sites with very low or very high accumulations of organic matter were dominated by stemmed species. Emergent plant species richness also showed such a quadratic correlation with accumulated organic matter.

Contaminant Toxicity. Some herbaceous plants are quite sensitive to heavy metals and other contaminants, and as a result, contamination can alter species composition, and decrease species richness, canopy coverage, and net annual productivity of wetland communities (e.g., Cooper and Emerick 1989, Olson 1979). Based on studies of eight Colorado wetlands exposed to varying degrees of heavy metal-contaminated runoff, Cooper and Emerick (1989) noted:

"Subalpine fen wetlands in the Colorado Front Range that have less than 3 vascular plant species growing in the main part of the wetland (not the edges) and have less than 50 percent total canopy coverage and less than 100 g/m2 total annual primary production, are likely to indicate impact from heavy metal toxicity. An exception is areas that are flooded or have ponded water for much of the growing season."

Forbs (herbaceous dicots in that study) seemed particularly uncommon in polluted wetlands. The authors noted no species that occurred only at contaminated sites, but found that the sedges, Carex aquatilis, C. utricularia, and/or C. scopulorum, predominated in these areas. Species absent from areas contaminated by large concentrations of heavy metals included the following:

Swertia perennis	Cardamine cordifolia
Caltha leptosepala	Epilobium lactiflorum
Geum macrophyllum	Galium trifidum
Sedum (Clementsia) rhodantha	Juncus albescens
Bistorta bistortoides	B. vivipara
Polygonum (Bistorta) bistortoides and vivipara

Duckweed (Lemna ) is particularly sensitive to the heavy metals cadmium and nickel, and chromium concentrations of 10 mg/L are inhibitory (Huffman and Allaway 1973). Cattail (Typha latifolia ) can tolerate lead, copper, and chromium accumulations of at least 10 micrograms/g dry weight of aboveground biomass; zinc accumulations in cattail may reach 25 micrograms/g dry weight without apparent ill effects (Mudroch and Capobianco 1979). The common reed (Phragmites ) can tolerate industrial wastewater with high levels of heavy metals (e.g., up to 250 micrograms/g sediment copper concentrations), as do bulrushes (e.g., Seidel 1966).

Heavy metals and other toxicants borne in air currents and precipitation have widely been reported to alter community composition of mosses and lichens (e.g., Lee et al. 1987, Sigal and Nash 1980). Species of mosses and lichens differ considerably in their sensitivity to metals, and are prevalent in many wetland types. Thus, they may have considerable potential for use as indicators of this type of pollution.

A decline in Asclepias syriaca (milkweed) and an overall increase in species richness and equitability may have been related to contaminants associated with incinerator residue deposited in an emergent marsh in Massachusetts (Mika et al. 1985). In a major Ohio river, Stuckey and Wentz (1969) reported the following species to be rare or absent from waters contaminated by industrial effluents, but common in analogous uncontaminated habitats:

Justicia americana	Lippia lanceolata
Saururus cernuus	Helenium autumnale
Phytostegia virginiana	Eclipta alba
Rumex verticillatus	Scirpus americanus
Samolus parviflorus	Amaranthus tuberculatus
Carex frankii	Hibiscus militaris
Lycopus rubellus	Strophostyles helvola

Also, these investigators found the following plants to be common in industrially contaminated waters:

Polygonum hydropiper	Echinochloa pungens
P. persicaria	Leersia oryzoides
P. pensylvanicum	Ambrosia trifida
P. coccineum	Urtica dioica
P. lapathifolium	Arctium minus
P. punctatum	Bidens frondosa
Sagittaria latifolia

In an Ontario river, submersed species (Elodea, Ceratophyllum, and Myriophyllum ) appeared to be less tolerant of industrial wastes than floating-leaved and short, rooted aquatic plants (Potamogeton, Nuphar , and Nymphaea ), which were in turn less tolerant than cattail (Typha ) and common reed (Phragmites )(Dickman et al. 1980, 1983, Dickman 1988).

Floating-leaved herbaceous plants are sensitive to the physical effects of oil, and growth of the duckweed Spirodela oligorhiza is affected by PCB concentrations of 5 mg/L (Mahanty 1975). Cattail can tolerate petroleum oil concentrations of 1 g/L (Merezhko 1973) and, along with common reed (Phragmites ), appeared to be the most tolerant macrophyte downstream from an industrial effluent source in Ontario (Dickman 1988). The response of wetland species to an oil spill in a Massachusetts inland wetland (Burk 1977) was as follows (* = annual species):

Species not recorded after oil spill:

Bidens cernua*	H. virginicum
B. connata*	Iris versicolor
B. frondosa*	Lycopus uniflorus
Echinochloa walteri*	Mimulus ringens
Eleocharis obtusa	Polygonum punctatum*
Galium tinctorium*	P. sagittatum*
Hypericum mutilum	Sparganium americanum
Spirodela ployrhiza	Vallisneria americana
Verbena hastata

Species reduced after oil spill:

Cephalanthus occidentalis	Najas flexilis*
Eleocharis acicularis	Onoclea sinsibilis
Galium trifidum	Pilea fontana*
Leersia oryzoides	Pontederia cordata
Lindernia dubia*	Scirpus pedicellatus
Ludwigia palustris	Zizania aquatica

Species apparently unaffected or increasing after oil spill:

Alisma subcordatum	Polygonum coccineum
Carex lurida	Potamogeton crispus
Ceratophyllum demersum	P. epihydrus
Dulichium arundinaceum	Sagittaria graminea
Eleocharis palustris	S. latifolia
Elodea nuttallii	Salix nigra
Equisetum fluviatile	Scirpus cyperinus
Lemna minor	S. validus
Lysimachia terrestris	Scutellaria leteriflora
Nuphar variegatum	Sium suave
Veronica scutellataq	Vitis labrusca

Bulrushes are killed by phenol concentrations of 100 mg/L and abnormalities occur at large phenol concentrations, but new shoots form quickly (Seidel 1966). Herbicides have often been used to control some herbaceous species, notably purple loosestrife (Lythrum ) and common reed (Phragmites ), and undoubtedly affect some non-target species as well. However, herbicide effects can be species-specific, with the result being that some applications result in overall increase in algae and plant richness (although perhaps lower overall productivity), as monotypic or dominant stands are opened for invasion by less aggressive species (e.g., Murphy et al. 1981). Detergent concentrations of 15 mg/L can damage wetland macrophytes (Agami et al. 1976).

Additional toxicological information may be available through EPA's PHYTOTOX (Royce et al. 1984) and AQUIRE databases.

Acidification. Ambient pH is one of the most important factors affecting community composition of emergent and aquatic bed wetlands bordering northern lakes (Hultberg and Grahn 1975), as well as peatlands (e.g., Anderson 1986, Jeglum 1971) and perhaps other low-alkalinity, standing water wetlands. It can be a stronger influence in these systems than nutrient status or water transparency (e.g., Jackson and Charles 1988). However, its effect on overall species richness is unclear (Eilers et al. 1984, Jackson and Charles 1988, Yan et al. 1985,). Usually, fewer species of macrophytes are found in acidic lakes than in circumneutral lakes (e.g., Friday 1987, Hunter et al. 1986, Hutchinson et al. 1985), but these are often species that are regionally rare (Moore et al. 1989).

The study of Adirondack (New York) lacustrine wetlands by Jackson and Charles (1988) reported the following taxa to be relatively intolerant of acidification: Najas flexilis, Nitella flexilis, Potamogeton pusillus, P. natans, and P. amplifolius. Submersed and floating-leaved species present at pH lower than 5.5 but not in less acidic conditions included Potamogeton confervoides and Sparganium angustifolium ; species present in both acidic (pH < 5.0) and circumneutral wetlands included Nuphar, Juncus pelocarpus, Drepanocladus fluitans, Utricularia vulgaris, Isoetes muricata, Eriocaulon septangulare, Sagittaria graminea, and Myriophyllum tenellum (Jackson and Charles 1988). Emergent species present in both acidic (pH < 5.0) and circumneutral wetlands included Calla pallustris, Juncus brevicaudatus, Dulichium arundinaceum, Lysimachia terrestris, and Juncus pelocarpus (Jackson and Charles 1988). Wolffia, Lemna, and Spirodela have optimal pH's of 5.0, 6.2, and 7.0 respectively, whereas their tolerated ranges are (respectively) 4 - 10, 4 - 10, and 3 -10 (McClay 1976).

In some northern wetlands, especially those that are heavily shaded, acidification can result in increased presence of mat-forming mosses of the genus Sphagnum (e.g., Gignac 1987, Grahn 1976, Roberts et al. 1985), and these mosses can further lower the pH (e.g., Glime et al. 1982). However, under severe acidification and accompanying deposition of industrial pollutants, Sphagnum can decline and in some cases, be replaced by cottongrass (Eriophorum )(e.g., Gorham et al. 1987, Lee et al. 1987a).

Cattail, rushes, and sedges occur in sediments with a pH of at least 4.7 (Dykyjova and Ulehlova 1978), while common reed and nutsedge can tolerate a pH as low as 2.0 (Al-Ali et al. 1978, Dykyjova and Ulehlova 1978). Natural stands of sedge (Carex ) have a pH range from 4.9 to 7.4 (Baker 1971), while the range for reed canary-grass (Phalaris ) is 6.1 to 7.7 (Gross and Jung 1978, Dean and Clark 1972, Niehaus 1971, Allinson 1972). Many regional botanical texts describe approximate pH ranges of individual wetland species (e.g., Crow and Hellquist 1981), as does some literature not excerpted here (e.g., Jeglum 1971, Swanson 1988).

Reductions in plant species diversity, decreased productivity, and life cycle disruptions were among the effects attributed to high pH values downslope from a Massachusetts hazardous waste lagoon (USEPA 1989a).

Salinization. Saline inland wetlands commonly have fewer species of macrophytes (Pip 1979, Reynolds and Reynolds 1975), and may be particularly deficient in species that typically form floating mats (Lieffers 1984). Most freshwater macrophytes cannot tolerate more than 10 ppt dissolved salts (Reimold and Queen 1974). Inland wetland plants that reportedly tolerate specific conductivity of greater than about 5 mS/cm are shown in Table 7, from Kantrud et al. (1989). Other data on salinity tolerances of inland wetland plants are provided by Reimold and Queen (1974) and others listed in Table 7.

Contamination of a northern Indiana bog with road salt resulted in almost complete elimination of endemic species and replacement by non-bog species, dominated by Typha angustifolia (Wilcox 1987), which can sometimes tolerate salinities of up to 25.5 ppt (Philipp and Brown 1965, Shekov 1974), at least for short periods. As salt concentrations declined in the four years of the study, endemic plants began to recolonize the affected area; biomass and growth of Sphagnum fimbriatum was significantly reduced at NaCl concentrations greater than 900 mg/L Cl- (Wilcox 1987). The common reed (Phragmites communis ) tolerates salinities of up to 45 ppt, although seedlings may be killed by salinities of 10 ppt. Duckweed (Lemna minor ) has reduced growth at salinities above 7 ppt (Haller et al. 1974, Stanley and Madewell 1976). For many species, these values vary by genetic population, life stage, duration of exposure, temperature, and other factors. The freshwater cattail, Typha latifolia, as expected, is less salinity-tolerant than the estuarine cattail, Typha angustifolia, mentioned above (McNaughton 1966). However, a presumed hybrid, Typha gauca, appeared resistent to road salt runoff (Bayly and O'Neil 1972). Even Typha latifolia seeds appeared more tolerant of road salt in snowmelt than germinating wool-grass (Scirpus cyperinus ) and three-way sedge (Dulichium arundinaceum ); purple loosestrife seeds (Lythrum salicaria ) were similarly tolerant (Isabelle et al. 1987). The rush, Scirpus acutus , appears more salt-tolerant than its many of its congeners (Smith 1983).

• Table 7. Examples of Aquatic Macrophytes Tolerant of Saline Conditions in Inland Wetlands.Table 7. Examples of Aquatic Macrophytes Tolerant of Saline Conditions in Inland Wetlands.

These lists are reproduced from Kantrud et al. (1989), and deal primarily with prairie pothole wetlands; applicability to other wetland types is unknown. Additional salt-tolerant species may be found in lists of Haller et al. 1974, Kauskik 1963, Lieffers 1984, Mall 1969, McKee and Mendelssohn 1989, Millar 1976, Millar 1978, Moyle 1945, Nelson 1954, Pip 1979, 1987a,b, Reynolds and Reynolds 1975, and Stewart and Kantrud 1972.

Sedimentation/Burial. We found little explicit information on overall macrophyte community response to burial or sedimentation. In some cases, sedimentation creates shoals in rivers or lakes, which provide sufficient substrate within the euphotic zone for herbaceous wetlands to become established or expand, at least until a major scouring flood re-occurs (e.g., Burton and King 1983). Where sedimentation is severe, water may become too shallow for some submersed species and a shift to emergent species may occur (Edwards 1969).

Differences probably exist among herbaceous plant species with regard to their intrinsic tolerance and adaptability to excessive sedimentation. Species often noted as occurring in disturbed, sediment-laden wetlands include the common reed (Phragmites ), reed canary-grass (Phalaris )(Reed et al. 1977), and other large and robust taxa. Repeated burial by as little as 5 cm of sediment per year can be detrimental to some emergent species (van der Valk et al. 1981).

Shading/turbidity. Increased shade or turbidity (whether from suspended sediment, phytoplankton, natural staining, or other sources) generally results in a shift in community structure from submersed species to floating-leaved or emergent species (Hough and Forwall 1988). Turbidity increases and decreases in bank stability may also favor an increase in the proportion of invasive, dominating species to the exclusion of less aggressive native macrophytes (Morin et al. 1989).

A 25 NTU (nephelometric turbidity units, or about 100 mg/L suspended solids) increase in turbidity in a shallow riverine wetland can reduce production of algae and submerged aquatics by 50 percent (Lloyd et al. 1987) and a mere 5 NTU increase (about 20 mg/L) has been shown to reduce the productive area of a lake by about 80 percent (Lloyd et al. 1987). The sensitivity of submersed plants to turbidity can be expressed by the ratio of the depth maxima of species to the Secchi transparency depth, i.e., the "turbidity tolerance index" (Davis and Brinson 1980). Data on depth maxima and ranges for many submersed species are compiled in Davis and Brinson (1980). The more shade-tolerant non-emergent herbaceous species are listed in Table 8.

Vegetation removal. Harvesting of "aquatic weeds" comprises a direct impact on submersed vegetation, and can shift the community composition at least temporarily. Species richness can either increase or decrease, depending on the initial state and species that are harvested (Sheldon 1986). At the deepwater edge of lacustrine wetlands with submersed plants, milfoil (Myriophyllum spicatum) frequently becomes dominant following the catastrophic alteration of more diverse communities by dredging, herbicides, disease, storms, herbivory, or other factors (Nichols 1984).

Removal of woody overstory generally increases herbaceous vegetation biomass and diversity (Madsen and Adams 1989). In the Prairie pothole region, specific information on shifts in community composition as a result of vegetation removal from grazing, haying, and cultivation, is reported by Kantrud et al. (1989: Appendix B). Annual burning, at least of emergent wetlands of the mesic pine-wiregrass savannas of North Carolina, can increase species richness (Walker 1985).

Thermal Alteration. Changes in wetland thermal regime can cause changes in production and shifts in species composition of the herbaceous plant community (Allen and Gorham 1973, Haag and Gorham 1977). An eventual shift from perennial and woody species to annual and herbaceous species may also occur in wetlands exposed to intermediate degrees of thermal warming (Dunn and Scott 1987, Sharitz et al. 1974). Changes are due both to physiological factors and (in northern wetlands) to changes in ice cover (Geis 1984) and growing season length. Most aquatic plants are killed by temperatures warmer than 45oC for 10 minutes, and by somewhat cooler temperatures for longer periods (Christy and Sharitz 1980). Despite this fact, and the fact that macrophyte species richness may be positively correlated with temperature across broad geographic regions, temperature in itself is probably not a major factor governing the distribution of herbaceous wetland plants (Pip 1989).

• Changes in community composition as a result of thermal alteration begin with changes in the germinations, growth, and survival of individual species. For example, the introduction over one year of continuously discharged heated water into a Wisconsin marsh resulted in failed shoot emergence, spring emergence instead of fall emergence, fewer number of shoots, and greater height of shoots in the sedge, Carex lacustris (Bedford 1977). Seedling survivorship of one common floodplain species, Ludwigia leptocarpa, was reduced at 42oC. (Christy and Sharitz 1980). Seedling germination of this species did not vary significantly over the range 22-42oC. Cattail, Typha latifolia, was killed as the probable result of heat-induced depletion of non-structural carbohydrates in its underground storage organs. This cattail may grow best at a water temperature of 30oC, but survival is poor at 35oC and seed germination requires temperatures of 13-24oC (Jones et al. 1979). The common reed (Phragmites communis ) may grow best when temperatures fluctuate within the 20-30oC range (Haslam 1973), and reed canarygrass (Phalaris ) may grow best at about 25oC (McWilliam et al. 1969). For most species, these values vary by genetic population, life stage, duration of exposure, day length, light intensity, and other factors.

• Table 8. Examples of Aquatic Plants That May Indicate Reduced Light Penetration Due to Greater Turbidity or Shade.Table 8. Examples of Aquatic Plants That May Indicate Reduced Light Penetration Due to Greater Turbidity or Shade.

From Davis and Brinson (1980) and other sources. Note that these species may occur as well in wetlands that are NOT turbid, although usually in smaller proportion relative to other species.

Alisma plantago-aquatica	Nuphar lutes
Ceratophyllum demersum	Potamogeton crispus
Eichhornia crassipes	Potamogeton pectinatus
Elodea canadensis	Potamogeton perfoliatus var. bupleuroides
Heteranthera dubia	Potamogeton pusillus
Hydrilla verticillata	Potamogeton richardsoni
Lemna minor	Riccia fluitans
Myriophyllum spicatum	Ricciocarpus natans
Najas flexilis	Spirodela polyrhiza
Najas guadalupensis	Vallisneria americana
Najas minor	Zannichellia palustris

Dehydration. Deviations of seasonal and annual hydrologic cycles from their "normal" regime (including stabilization of usually fluctuating regimes) can profoundly affect structure of herbaceous wetland plant communities, perhaps even more so than the actual magnitude of the deviation (Hartog et al. 1989, Zimmerman 1988). In some cases, community changes reflect the "intermediate disturbance" hypothesis, wherein "moderate" deviations from "normal" conditions increase community diversity. For example, in Okefenokee Swamp in Georgia, Greening and Gerritsen (1987) found greater species diversity and variation in biomass at a site where drawdown was occasional and less predictable than at more predictable sites.

Many herbaceous plant communities, particularly those with rigid stems (e.g., cat-tail, common reed) can endure (and may even require) periods of a few hours or days of occasional dehydration without changing. Even a few non-rigid species can survive two or more weeks of exposure, e.g., water milfoil (Myriophyllum spicatum ), bladderwort (Utricularia gibba ), duckweed (Lemna minor ), pondweed (Potamogeton pectinatus ), and Ceratophyllum demersum (e.g., Cooke 1980).

However, if dehydrated shorelines subside (collapse) or complete water level drawdown is sustained over many days (particularly if it occurs during the growing season and results in desaturation of sediments) dehydration can trigger significant changes in wetland community structure. This is largely due to the increased availability of nutrients as sediments become desaturated and oxidized, and partly due to enhanced germination of seeds of wetland plants that have lain dormant for years in sediments.

In wetlands that are strongly influenced by ground water discharge, erect vegetation may be less vulnerable to effects of drawdown, because sediments are less likely to become totally dewatered during intentional drawdown (Cooke 1980). Effects are likely to be most severe when drawdown occurs during extremes of heat or cold.

In the short-term, complete drawdown often shifts the balance of community structure in favor of emergent and woody species, and away from submersed species. In the Southeast, aggressive aquatic plants such as alligatorweed (Alternanthera philoxeroides ) and naiad (Najas flexilis ) can increase following partial drawdown, while muskgrass (Chara vulgaris ), water lily (Nuphar spp.), and water hyacinth (Eichhornia crassipes ) can decrease (Holcomb and Wegener 1971, Lantz et al. 1964). In prairie potholes, complete water loss year after year results in reduced richness even of herbaceous plants, with Carex and Polygonum generally becoming dominant (Driver 1977). However, partial drawdown, particularly if it occurs for short periods, may greatly increase macrophyte biomass and growth, due to enhanced nutrient and light availability that otherwise limit submersed species (Wegener et al. 1974). In Minnesota peatlands, artificial drainage resulted in increased dominance of the sedge Carex lasiocarpa (Glaser et al. 1981).

In Indiana, woolgrass (Scirpus cyperinus ) was believed to indicate dessication and related disturbance of former wetlands (Wilcox et al. 1985), as was the sedge, Carex antherodes , in central Canada (Millar 1973). Woolgrass, along with reed canary-grass (Phalaris ) tolerated severe water level drawdown in a New York reservoir (Burt 1988). In temporarily drained wetlands, Mallik and Wein (1986) found that Typha (cat-tail), Calamagrostis canadensis and Brachythecium salebrosum had highest cover values. Cover and stem density of Typha increased after draining, while plant height and stem diameter decreased, compared to a flooded area. Typha may not be a good indicator of wetland dehydration, however, as the same study showed that on the flooded area, Typha, Sphagnum squarrosum (a moss) and Pellia epiphylla had the highest cover values.

In a literature review on the effects of lake drawdown for control of macrophytes in Wisconsin eutrophic lakes, Cooke (1980) surmised that only three species--Brasenia schreberi (a water shield), Hydrochloa carolinensis , and Potamogeton robbinsii (a pondweed) always decline following temporary drawdown, and Nuphar spp. and Myriophyllum spp. often decline following drawdown. Species that appear always to increase following temporary drawdown include Alternanthera philoxeroides (alligator weed), Lemna minor (duckweed), Leersia oxyzoides (cutgrass), and Najas flexilis.

In southern Florida, drainage of wet prairies and cypress domes results in increased sawgrass, broomsedge (Andropogon ), common reed (Phragmites communis ), maidencane (Amphicarpum), chainfern (Woodwardia ), and many graminoids and shrubs; in deeper waters of wetlands, cattail (Typha ) may increase (Alexander and Crook 1974, Rochow 1983, Rochow and Lopez 1984, Worth 1983, Atkins 1981).

A wealth of qualitative information about hydrologic tolerances of plants has been compiled for the U.S. Fish and Wildlife Service's "National List of Plant Species that Occur in Wetlands" (Reed 1988). This publically-available database classifies all U.S. wetland plants according to their fidelity to wet environments, i.e., obligate (nearly always in wetlands) or facultative (usually or sometimes in wetlands). As one might imagine, the obligate taxa in general tend to be less tolerant of desiccation than the facultative taxa listed in that database. Information on hydric preferences of species might be numerically summarized using the index of Michener (1983).

The U.S. Fish and Wildlife Service, through the National Ecology Research Center in Fort Collins, Colorado, has also compiled information on "moist soil management techniques." Use of proposed models will allow users to predict the effect of water level changes on herbaceous wetland plants, or perhaps conversely, what the presence of particular plants suggest about prior hydrologic regimes.

Drawdown of wetland water levels in some regions results in increased susceptibility to fires, which in turn can trigger significant changes in wetland chemistry and vegetation.

Inundation/impoundment. Effects of inundation on emergent herbaceous species are extensively compiled in Fredrickson and Taylor (1982), Knighton (1985), and Whitlow and Harris (1979). Herbaceous species of the prairie pothole region are classified according to 12 life history types, related to flooding regime, by van der Valk (1981). A few additional studies that have examined inundation effects on herbaceous plants are summarized here.

Increased water levels in aquatic bed (submersed and floating-leaved plant) wetlands appear to have little effect in some instances (Davis and Brinson 1980). However, water level increases in other instances may result in increased wave action and initially greater turbidity, which is detrimental to many aquatic plants.

Addition of permanent open water to a non-permanently flooded, emergent wetland increases the opportunity for invasion by many submersed and floating-leaved species, and generally results in an increase in on-site species richness. Normally aggressive, perennial emergents such as purple loosestrife, cat-tail, common reed, and water hyacinth may be reduced or eliminated along with less aggressove species as flooding increases.

Although species richness of an entire non-permanently flooded wetland sometimes declines for a few years after flooding becomes permanent (Sjoberg and Danell 1983), overall community richness may change only slightly in the long term, and only the position of the submersed, emergent, and meadow zones may shift (Harris and Marshall 1963, van der Valk and Davis 1976). Such zonal shifts serve as indications of long-term water level change within a wetland (Botts and Cowell 1988). They occur as dormant seeds of wetland plants, which require specific water depths for germination (Moore and Keddy 1988), germinate along the upland boundary as a result of the new flooding. At the same time, down-gradient species subjected to inundation may be lost as a result of suffocation, build-up of compounds toxic to roots, and alteration of physical conditions, e.g., erosion and scour. Floating-mat vegetation may survive.

The effects of flooding also will depend on flooding depth, frequency, duration, dominant plant species, sediment type, water velocity, and other factors. Short periods of flooding (days or weeks) were reported in Wisconsin to have no effect on wetland community composition (e.g., Nichols et al. 1989). Assemblages of herbaceous wetland plants can be a more sensitive indicator of water level change than assemblages of woody wetland plants, which respond more slowly (Paratley and Fahey 1986).

Deviations of seasonal and annual hydrologic cycles from their "normal" regime in wetlands, and particularly, the elimination of seasonal fluctuations, may reduce overall plant species richness. Native and perennial species, particularly grasses and sedges, may be replaced by taxa that are more aggressive, exotic, clonal, and/or annuals. Depending on the initial water level, these commonly include cat-tails, bulrush, pickerelweed (Sagittaria ) and pondweed (Pontederia )(e.g., Botts and Cowell 1988, McIntyre et al. 1988). In some cases, community changes reflect the "intermediate disturbance" hypothesis, wherein "moderate" deviations from "normal" annual hydrologic conditions increase community diversity.

In wetlands along Lake Erie in Ohio, diking of a marsh increased the dominance of liverwort (Riccia spp.), duckweeds (Lemna minor, Spirodela polyrhiza ), coontail (Ceratophyllum demersum ), water milfoil (Myriophyllum speciatum ), pondweeds, and bladderwort (Utricularia vulgaris ) (Farney and Bookhout 1982). Permanent inundation of other marshes has decreased the number of plant species and the dominance of Carex spp. (Farney and Bookhout 1982, Sjoberg and Danell 1983).

In Colorado, subalpine wetlands flooded for longer than about 30-45 days during the growing season had fewer emergent plant species than those inundated for shorter periods (Cooper and Emerick 1989). In Sweden, lakeshore wetlands with less than 40-60 days of flooding during the growing season had maximum richness and cover of macrophytes, in contrast to those flooded for longer periods (Nilsson and Keddy 1988).

Cattails generally tolerate deeper water than most rushes (Lathwell et al. 1973), which tolerate deeper water than most sedges (van der Valk and Davis 1976). Cattail (Typha ) can dominate wetlands with water depths generally greater than 15 cm for 6 to 12 months (Mall 1969). The common reed (Phragmites ) typically occurs in water depths of 0 to 1.5 meters (Haslam 1970, Spence 1982). For most species, these values vary by genetic population, life stage, duration of exposure, water chemistry, and other factors. The horse-tail, Equisetum fluviatile , appeared to be the most tolerant of several emergent species to modest increases in water depth (Sjoberg and Danell 1983). Depth-to-water-table preferences of many peatland species are given by Jeglum (1971).

As noted above, a wealth of qualitative information about hydrologic tolerances of plants is represented by the U.S. Fish and Wildlife Service's "National List of Plant Species that Occur in Wetlands" (Reed 1988), and in their models currently being developed for moist-soil management and in-stream flow management.

Fragmentation of Habitat. We found no explicit information on macrophyte community response to fragmentation of regional wetland resources. Biomass and cover of submersed wetland plants generally decreases with increasing lake size, while the converse is true for emergent species (Duarte et al. 1986). In England, Helliwell (1983) found greater macrophyte species richness in larger wetlands, but a large amount of the variation could be attributed to other factors. Larger wetlands also may have greater macrophyte richness because they tend to be visited more often than smaller wetlands by birds and other animals capable of introducing new plants (Pip 1987). However, regionally rarer species often occur in small wetlands with unique physical and chemical environments (Moore et al. 1989). One can surmise that species with broad environmental tolerances and that disperse easily (e.g., Godwin 1923) might be least affected as wetlands become more isolated from one another.

Other human presence. The bottoms of Sierra lakes in California with higher levels of human visitation had more coverage with rooted macrophytes (Isoetes, Anacharis, Nitella ) and bottom algae (Rhizoclonium ) than those less frequently visited; this phenomenon was evident even in lakes where use had been restricted for 10 to 20 years (Taylor and Erman 1979). Trampling and other impacts on riparian wetland vegetation are documented in Cole and Marion (1988) and in studies of wetland buffer zones in New Jersey.

In wetlands of "developed" watersheds in New Jersey, uncharacteristic herbaceous species replaced endemic ones, and herbs and vines were more prevalent than in "undeveloped" watersheds (Ehrenfeld 1983, Schneider and Ehrenfeld 1987). Common reed (Phragmites ) typically characterizes many disturbed, nutrient-poor wetlands (Haslam 1971), as do woolgrass and soft rush (Juncus effusus ) in Pennsylvania (Hepp 1987). In Ohio, increased eutrophication, warming, and turbidity were implicated in the decline of Najas gracillima and N. flexilis , and an increase in N. marina, N. minor , and N. guadalupensis over a 70-year period (Wentz and Stuckey 1971).

Channels with heavy shipping traffic connecting the Great Lakes had less dense beds of submersed macrophytes than in channels with less ship traffic. Dominance shifted from Myriophyllum spicatum, Elodea canadensis and Heteranthera dubia in the relatively undisturbed channel to Characeae, Potamogeton richardsonii, and Najas flexilis in the disturbed channel (Schloesser and Manny 1989). Similar reductions of macrophyte biomass from recreational boating have been found in Europe, as compiled by Liddle and Scorgie (1980) and Murphy and Eaton (1983).

6.2 SAMPLING METHODS AND EQUIPMENT

Factors that could be important to standardize (if possible) among wetlands when monitoring community structure of macrophytes include:

age of wetland (successional status), light penetration (particularly for submersed species), water or saturation depth, conductivity and baseline chemistry of waters and sediments, current velocity, abundance of herbivores (particularly muskrat, geese, grazing cattle, crayfish), stream order or ratio of discharge to watershed size (riverine wetlands), sediment type, existence of any prior planting programs, and the duration, frequency, and seasonal timing of regular inundation, as well as time elapsed since the last severe inundation, drought, or fire.

References that provide more detailed guidance on sampling herbaceous wetland vegetation include Fredrickson and Reid 1988a, Moore and Chapman 1986, Mueller-Dombois and Ellenberg 1974, Murkin and Murkin 1989, Phillips 1959, Schwoerbel 1970, and Woods 1975. Guidelines for collecting specimens of aquatic plants for preservation are given by Britton and Greeson (1988), and Haynes (1984). References useful in data analysis are listed in Chapter 3.

If wetlands can be sampled only once, mid-growing season is usually the recommended time. However, many plants are apparent and/or identifiable only for a few weeks of the growing season. Thus, if the aim is to quantify community composition accurately, repetitive visits that account for the diverse phenologies of wetland species should be implemented. Ideally, annual visits could be timed to coincide with year-specific weather conditions, rather than calendar date. For example, Grigal (1985), who sampled vegetation over three years, did field work at slightly different times each year. This increased the chances of finding species in flower, making identification easier. In northern bogs, early fall may be a desirable sampling time, e.g., Wilcox (1986) sampled vegetation in a bog in the first week of September because the maximum number of species was identifiable at that time in Indiana. Optimal sampling times vary geographically.

Whenever possible, plants should be identified in the field rather than collected. Trampling of herbaceous vegetation and compaction of saturated soils during even a single site visit can induce community changes detectable in subsequent visits. Thus, field crews should be as small as possible and follow the same path in and out of a wetland. In riverine and lacustrine wetlands, underwater SCUBA transects can be run (Schmid 1965).

Equipment commonly used to destructively sample herbaceous wetland vegetation (especially submersed species) includes dredges, oyster tongs, plant grappling hooks, steel garden rakes, and similar devices (Britton and Greeson 1988). Equipment designed specifically for sampling herbaceous macrophytes is described by Dromgoole and Brown 1976, Macan 1949, Satake (1987), 1977, Wood 1975, and others.

Types of commonly-collected data on herbaceous wetland communities include species per plot and percent cover. Less often, total stem count per m2, and stem count per species per plot are determined. Stem counts are usually made only of species perceived to be dominant, and may possibly include a few subdominants.

Herbaceous plant community composition is typically quantified using belt transects or replicate quadrats. Transects and quadrats can be used in all wetland types, but may give less reliable data where vegetation is submerged or otherwise difficult to access. Sampling schemes involving transects or quadrats can yield data that is particularly amenable to statistical analysis. The number, size, and spacing of transects and quadrats in a wetland depends primarily on wetland size, shape, internal heterogeneity (e.g., as perceived during an initial reconnaissance visit and/or from aerial photographs), and the statistical power one wishes to have in detecting spatial change in various community metrics. Larger wetlands require more transects or quadrats, usually spaced farther apart, to accurately characterize overall community composition. More linear wetlands (e.g., narrow fringe marshes along lakes) may require more tightly spaced sampling points, as may ecotone areas along transects. Sampling stations along transects are usually situated at even intervals, and quadrats can be placed evenly (e.g., in a grid), randomly, or clustered. Random placement of plots for the purpose of statistically characterizing a wetland is usually prohibitively expensive, due to the extreme spatial variability of most wetlands (Durham et al. 1985). Plots or transect lines are often marked for future relocation.

In most studies of herbaceous wetlands, investigators have located transects or quadrats in a manner that parallels or spans a likely stressor gradient (e.g., parallel to basin gradient, perpendicular to flow, or parallel to flow path of discharge from a chemical outfall). If the stressor is a point source, the transect should be long enough to allow complete definition of gradients in response to the stressor. Thirty meters was not far enough to show distance effects of wastewater disposal in a bog/marsh system studied by Kadlec and Hammer (1980). To avoid problems of treatment effects spilling over into control plots, Loveland and Ungar (1983) used a randomized block design of 0.25-m2 plots in each of three vegetation zones. Each zone contained five replications of each block; for controls, five plots were randomly spaced in each zone. In another study of artificial enrichment (Duarte and Kalff 1988), plots were not isolated; fertilized plots were 9 m apart and control plots were 3 m from the corresponding treated plots.

Where multiple, non-overlapping gradients are perceived, transects may be located perpendicular to, or at other appropriate angles to, each other. The number of transects and quadrats in particular cover types within the wetland may also be designed to be proportional to the overall coverage of these cover types.

Transects used for herbaceous community monitoring have ranged upwards from about 100 meters in length (depending on wetland size and shape); quadrats have ranged upwards from 0.05 m2, and may be rectangular, square, or circular. The minimum effective size can be determined statistically or by plotting of initial data, as described in section 3.3. Based on statistical analysis of dozens of published studies of submersed vegetation, Downing and Anderson (1985) suggested it is better to use small quadrats with great replication than large quadrats with little replication, especially where vegetation stands are not dense. However, they suggested cautious interpretation of this recommendation if small quadrats are being placed in dense macrophyte beds. From a study of 18 Canadian lakes, France (1988) determined that at least 21 replicate samples are required to achieve estimates within 20 percent of the mean biomass, using a sampler with an area of 45.6 cm2. A different number of replicates would probably be required if determination of richness, rather than biomass, was the objective.

Variable-sized plots also can be used, where plot size depends on life form of vegetation present in proximity to each particular point in the wetland (e.g., Mader et al. 1988). Nested frequency quadrats, in which only the number of times a species is present is recorded--have also been used (e.g., Frenkel and Franklin 1987). These have the advantage of easily data collection, objectivity, and no need to relocate plots, but interpretation depends on plot size and shape and spatial distribution of species, and this approach cannot easily be used to quantify spatial patterns, cover, or biomass.

6.3 SPATIAL AND TEMPORAL VARIABILITY, DATA GAPS

In general, the parameters most often measured in studies of herbaceous wetlands are "percent cover" and "biomass (standing crop)." Measures of community structure of submersed or emergent aquatic communities have not been uniformly collected from a series of statistically representative wetlands in any region of the country. Thus, it is currently impossible to state what are "normal" levels for descriptors of community structure such as seasonal plant density or species richness, and their temporal and spatial variability, in any type of herbaceous wetland.

Perhaps the closest approximation of a broad-scale effort is that of Duarte et al. (1986). They looked at just one parameter--lacustrine macrophyte biomass--and examined causes of local and regional variability. From their resultant equations, expected ("nominal") levels of biomass of both emergent and submersed macrophytes in lakes might be estimated. Approximate data describing lake area, depth, slope, and a few other simple parameters are needed to run the calculations.

A few, usually localized, studies of inland wetlands have published Shannon diversity index values. For example:

State	type	N	min.value	max.value	citation
SC	Pfo	?	3.58	3.78	Sharitz et al. 1974
MA	Lab	120	1.47	3.71	Burk 1977
GA	P	?	0.83	1.54	Greening & Gerritsen 1987
WA	P	>300	1.48	2.65	Meehan-Martin & Swanson 1988,1989

Species richness has been reported by many studies, but is not always standardized per unit effort or per unit area as it should. Examples include:

State	type	N	min. value	max.value	citation
SC	Pfo	650	1.2/0.25m2 SE=0.4	9.6/0.25m2 SE=0.8	Dunn and Sharitz 1987
MA	Lab	120	2.3/0.25m2 SE=0.21	8.3/0.25m2 SE=0.56	Burk 1977
IA	Pem	28	3.2/m2 SD=0.5	8.3/m2 SD=2.0	van der Valk & Davis 1976
NY	Pem	90	7.7/m2	12.2/m2	Paratley & Fahey 1986
NJ	P	18	26.8/600m SE=2.7	41.4/600m SE=5.7	Schneider & Ehrenfeld 1987

In addition, Ehrenfeld (1983) summarized her data as follows:

Mean species richness per 600m2 (N=16):
Disturbed sites = 33.9 + 2.17; range 17-47
Pristine sites = 27.8 + 2.24; range 13-44

A similar study by Morgan and Philipp (1986) reported the following values for coefficient of similarity (based on 12 plots per stream, each plot 600m2 in area):

between polluted and unpolluted streams = 16%
among polluted streams = 28%
among unpolluted streams = 26%

Many other studies, although not publishing or summarizing in a useful form their statistics on community structure, have compared herbaceous vegetation among wetlands in a region (i.e., spatial variation). Some of the more systematic or extensive quantitative comparisons include:

Albert et al. 1987, Brumfield and Evans 1982, Canfield et al. 1983, Canfield and Duarte 1988, Duarte et al. 1986, Ehrenfeld 1986, Henebry et al. 1981, Pip 1979, 1987a,b, Sheath et al. 1986, Stewart and Kantrud 1972, and Terry and Tanner 1984.

One of the more geographically extensive ongoing studies is a survey of vegetation in a large number of Great Lakes wetlands in Michigan (Albert et al. 1987). Survey locations are shown in Appendix B. Another extensive and long-term survey of wetland vegetation is being conducted as part of monitoring studies for the ELF military radiocommunications facility (Blake et al. 1987).

A significant number of studies have compared long-term change (but seldom year-to-year variation) in plant community structure in wetlands, in some cases by use of paleoecological techniques. Changes in most cases have not been quantitatively linked with particular stressors. These chronological studies include:

Baumann et al. 1974, Brown 1987, Bumby 1977, Burk 1977, Burton and King 1983, Harris et al. 1981, Hale and Miller 1978, Kadlec 1979, Niemier and Hubert 1986, Schwintzer and Williams 1974, Southwick and Pine 1975, Stuckey 1971, van der Valk and Davis 1979, and Wentz and Stuckey 1971.

Qualitative data on community structure of inland herbaceous wetlands appears to be most available for Florida, Minnesota-Michigan-Wisconsin, Louisiana, New York, and North Dakota. Apparently the least amounts of such data are for playa wetlands, and for herbaceous wetlands in the Appalachians, southern Great Plains, and Southwest. Information is most available on impacts of hydrologic alteration and nutrients, and least on impacts of partial burial, contaminant toxicity, and habitat fragmentation.

Reasonably complete, qualitative lists of "expected" wetland herbaceous plants are available for most regions through the USFWS's "National List of Wetland Plants" database, and databases of The Nature Conservancy. Quantitative data are generally most available for vascular emergent species, and less common for submersed plants and mosses. Limited qualitative information may also be available by wetland type from the "community profile" publication series of the USFWS (Appendix C).