Changes in the Abundance and Diversity of Coastal Wetland Fauna from the Open Water/Macrophyte Edge Towards Shore

Changes in the abundance and diversity of coastal wetland fauna from the open water/macrophyte edge towards shore

B.J. Cardinale1,2, V.J. Brady3,4 & T.M. Burton3 1Department of Fisheries & Wildlife, Michigan State University, East Lansing, MI, 48824, U.S.A. 2Present address: Department of Biology, University of Maryland, College Park, MD 20782, U.S.A. 3Department of Zoology, Michigan State University, East Lansing, MI 48824, U.S.A. 4Present address: National Health & Environmental Effects Research Lab, US EPA, Duluth, MN 55804, USA

Received 15 December 1997; accepted in revised form 28 June 1998

Key words: benthic invertebrates, community structure, ﬁsh, Laurentian Great Lakes, marsh, microcrustacea, physical and chemical gradients, zooplankton

Abstract Great Lakes coastal wetlands are widely recognized as areas of concentrated biodiversity and productivity, but the factors that influence diversity and productivity within these systems are largely unknown. Several recent studies have suggested that the abundance and diversity of flora and fauna in coastal wetlands may be related to distance from the open water/macrophyte edge. We examined this possibility for three faunal groups inhabiting a coastal wetland in Saginaw Bay, Lake Huron. We sampled crustacean zooplankton and benthic macro-invertebrates at five distances from open water in the summer 1994, and fish at three distances from open water in 1994 and 1995. We found significant spatial trends in the total abundance and diversity of zooplankton and fish, as well as the diversity of benthic macro-invertebrates. Zooplankton abundance and taxa richness were highest at intermediate distances from open water in a transition zone between the well-mixed bayward portion of the wetland, and the non-circulating nearshore area. Benthic macro-invertebrate taxa richness increased linearly with distance from open water. In contrast, fish abundance and species richness declined linearly and substantially (abundance by 78%, species richness by 40%) with distance from open water. Of the 40 taxa examined in this study, 21 had significant horizontal trends in abundance. This led to notable differences in community composition throughout the wetland. Our results suggest that distance from open water may be a primary determinant of the spatial distributions of numerous organismal groups inhabiting this coastal wetland. Several possible reasons for these distributions are discussed.

Introduction degraded as a result of human influences (Krieger et al., 1992). Despite their diminishing quality and extent, the There are approximately 1,200 km2 of coastal wet- Great Lakes coastal marshes are still widely recog- lands in the United States that fringe the Laurentian nized as areas of concentrated biodiversity and pro- Great Lakes (Mitsch and Gosselink, 1993). Along ductivity (Stuckey, 1989; Smith et al., 1991; Krieger, many shorelines these wetlands are quite impressive in 1992; Randall et al., 1996; Brazner and Beals, 1997). extent; yet, they represent only a small portion of the However, our knowledge of the factors that influence coastal marshes that were present before Europeans ar- the diversity and productivity of flora and fauna within rived in the US. An estimated 60–80% of Great Lakes coastal wetlands is still rudimentary, at best. If we wetlands have been lost to agricultural, residential, or are to predict the consequences of habitat loss, or di- industrial development since settlement (Comer et al., rect efforts to conserve or restore coastal wetlands in 1995). Of those that remain, many wetlands are highly the Great Lakes, it is imperative that we uncover the

Article: wetlmitsch19 GSB: 9 Pips nr. 184695 (wetlkap:bio2fam) v.1.1 wetmit19.tex; 19/01/1999; 16:57; p.1 60

Figure 1. The position of benthic macro-invertebrate and zooplankton sampling stations relative to distance from the open water/wetland edge, and the location of fyke nets within the wetland study site in Saginaw Bay, Lake Huron, Michigan (USA). environmental factors that influence the structure of grazers and filter feeders decline, often by several or- communities inhabiting these systems. ders of magnitude, with increasing distance from open Several recent studies performed in coastal wet- water (Brady et al., 1995; Cardinale et al., 1997). lands of Saginaw Bay, Lake Huron, have described an Given these previous results, we wondered if the axis of environmental variation that may have wide- spatial distributions of other coastal wetland fauna spread biological importance. Suzuki et al. (1995) might be related to distance from the open wa- documented physical and chemical discontinuities that ter/macrophyte edge. Concurrent with our study of formed horizontally from the open water/macrophyte epiphytic invertebrates (Cardinale et al., 1997), we had edge towards the shore. As wind-induced surface two other ongoing projects at our study site. Neither waves were gradually reduced by macrophytes, turbid- was specifically designed to examine spatial distri- ity, dissolved oxygen, and pH decreased with distance butions; however, we were able to recast the data from open water while alkalinity, conductivity, and in a manner that allowed comparison of crustacean most dissolved ions increased. Changes in the abiotic zooplankton, benthic macro-invertebrate, and fish as- environment were accompanied by rather substantial semblages at several distances from open water over a changes in the biomass of phytoplanktonic and epi- five month study period. Here, we present data show- phytic algae, both of which declined by 80–97% from ing how (1) abundance, (2) taxonomic richness, and open water towards shore. (3) community composition of these three organis- We found nearly identical abiotic gradients in sev- mal groups changed with increasing distance from the eral coastal wetlands throughout Saginaw Bay (Cardi- open water/wetland edge. nale, 1996). In our most studied site, we have shown Methods that the biomass and net primary production of epiphytic algae decline substantially from open water Study site towards shore in correspondence with the abiotic gra- Our study site was located in a coastal emergent dients (Cardinale et al., 1997). We have also found that wetland typical of those extending around the south- the biomass, diversity, and survivorship of epiphytic eastern shore of Saginaw Bay, Lake Huron, USA (43◦

wetmit19.tex; 19/01/1999; 16:57; p.2 61

Table 1. The relative abundance of zooplankton taxa. Numbers Collection of invertebrates shown for the major taxonomic groups are the percentage of the ∗∗ total number collected from that station. Due to the disparity in taxonomic resolution, the relative abundance of individual clado- From June to October of 1994, we took monthly ceran taxa is based only on the total number of cladocera, not the samples of crustacean zooplankton and benthic macro- total number of zooplankton. Thus, columns do not sum to 100%. invertebrates at ﬁve stations along a transect that Dots are shown where a taxon was not present on any sampling bisected the S. americanus stand from open water to- date. DFOW = distance from open water into the macrophyte bed. wards shore (Figure 1). Zooplankton were sampled at ◦ TAXA DFOW = 20 m 100 m 180 m 280 m 360 m a random distance (1–10 m) and direction (0–360 ) from each station using a modiﬁed Gerking sampler Copepoda (21.5 cm i.d. with 250 (µm mesh, Mittelbach, 1981) Cyclopoida 19 18 15 25 41 which encircled the entire water column from surface Calanoida 2 11 6 1 < 1 to bottom along with any macrophyte stems present. Ostrocoda 6 21 32 10 11 Water depth was concurrently measured so that zoo- Cladocera** 73 50 47 64 48 plankton abundance could be reported per volume of Acroperus harpae < 12 6 1221 Alona quadrangularis 12 28 18 4 9 water. At the same location, a sediment corer (4.5 cm Bosmina longirostris 20 5 2 3 3 i.d.) was used to collect benthic macro-invertebrates Ceriodaphnia sp(p) 32 23 11 22 16 to a soil depth of 10 cm. Two cores were taken from Chydorus sphaericus 33 1 18 each location and pooled into a composite sample. All Diaphanosoma birgei < 11 1 1011 samples were immediately placed in a cooler with ice Eurycercus lamellatus < 11 2 11 and transported to the laboratory within 3–4 hours for

Ilyocryptus spinifer < 1 3 13 30 2 processing.

Leydigia sp(p) 1 < 1 In the laboratory, zooplankton and benthic macro-

Macrothrix sp(p). < 1 17 1 invertebrate samples were rinsed through a 250 µm Monospilus dispar 12<1 sieve with any plant stems rubbed by hand to dis- Pleuroxus sp(p) < 11 2 49 lodge the animals. The 250 µm mesh was suffi- Side crystallina 30 29 25 7 8 cient to capture most crustacean zooplankton in our Simocephalus sp(p). 1 < 1 < 1512 study site, but was too large for most rotifers, which were subsequently ignored by our study. Specimens were preserved in 95% ethanol with rose bengal dye. Prior to counting and identification, each sediment 370N83◦380W, Figure 1). Vegetation in this area sample was sugar floated (Anderson, 1959) and sub- was unprotected from prevailing northwesterly winds sampled. One-sixth of the total was examined under × that swept across the shallow open water of the bay. 10 magnification and benthic macro-invertebrates As a result, pelagic surface waves generally arrived were counted and identified to an operational taxo- perpendicular to the stand, and during our study were nomic unit. A minimum of 50 specimens, but typically observed to penetrate 160–200 m into the vegetation. far more, was identified from each sample. Simi- The macrophyte community was almost mono- larly, zooplankton were identified and enumerated at × typic being dominated by three-square bulrush, Scir- 40–400 from one-sixth of each sample (generally pus americanus, extending 500 m from the shore. 100 specimens). Small, isolated patches of Scirpus acutus, Typha an- Collection of fish gustifolia,andSagittaria sp. did occur in the site but were never sampled. Macrophytes grew on substrates During the summers of 1994 and 1995, paired fyke composed mostly of sand (85–97%) with lesser frac- nets with 4.7 mm mesh and 15.24 m lead lines were tions of clay and silt. Growth of the vegetation was used to collect fish at 60, 180 and 320 m from open seasonal with regeneration from rhizomes after winter water (Figure 1). On July 13 and 27 of 1994, and July ice-scour. At the peak of summer growth, the density 5–7 and August 23–24 of 1995, one paired fyke net of Scirpus americanus averaged 235 61(1 SE) stems was placed parallel to the shoreline at each of the three per square meter at our sampling stations. locations. In 1994, the nets were set between 9 and 11 AM with collections made at dusk (6–8 PM). In 1995, a nighttime sampling effort was added such that nets were emptied at dusk and collections made again at

wetmit19.tex; 19/01/1999; 16:57; p.3 62

Figure 2. Tukey box plots of the abundance (A) and taxonomic richness (B) of crustacean zooplankton, benthic macro-invertebrates, and fish relative to distance from open water. Boxes display the 25th and 75th percentiles, whiskers denote the 10th and 90th percentiles, solid lines show the medians, and dashed lines show the means. Circles indicate data points that lie outside of the 10th or 90th percentiles. Significant quadratic and linear trends are noted. NS = not significant. dawn. In all, there were 6 collections at dusk repre- richness, there were 25 samples used in the regressions senting a full day’s sampling effort and 4 collections at (5 stations × 5 months), and for fish there were 30 dawn representing a full night’s sampling effort. Each samples (3 stations × 10 sampling efforts). For regres- distance from open water received an equal amount of sions of the abundance of individual taxa, we excluded sampling effort on each date. All fish were identified dates for which a taxon was not present in at least to the lowest possible taxonomic unit on site (genus or one sample (e.g., was not found in the marsh at that species), enumerated, and then released. time). Abundances were log (x + 1) transformed, if needed, to meet the assumptions of normality and Data analyses homogeneity of variances. Finally, the relative abundance of each taxon was Linear and quadratic regressions were used to detect calculated as a percentage of the total of all specimens significant spatial trends in (a) the total abundance collected from any particular station. This allowed of each faunal group, (b) the taxonomic richness of a comparison of community composition at different each faunal group, and (c) the abundance of individ- distances from open water. ual taxa in the wetland. A significant linear regression was taken as evidence for a positive or negative relationship to distance from open water. A significant Results quadratic regression was taken as evidence for a uni- modal trend from open water towards shore (e.g., Zooplankton highest abundance in the middle of the wetland). For these analyses, each sampling effort was considered to During the five months of this study there was a sig- be independent. Thus, for regressions of zooplankton nificant spatial trend in total zooplankton abundance and benthic macro-invertebrate abundance and taxa from open water towards shore (quadratic p < 0.05,

wetmit19.tex; 19/01/1999; 16:57; p.4 63

Table 2. The relative abundance of benthic macroinvertebrate taxa as a proportion of the total collected from each station during the months of June through October, 1994. Dots are shown where a taxon was not present on any date. DFOW = distance from open water into the macrophyte bed.

TAXA DFOW = 20 m 100 m 180 m 280 m 360 m

Insects

Diptera Ceratopogonidae (biting midges) 11 2 Chironomidae (non-biting midges)

Orthocladiinae/Chironomini 12 10 12 7 10 Tanypodinae 512 Tanytarsini 7 4 9 1 12 Ephemeroptera (mayﬂies) < 12 8 1821 Trichoptera (caddisﬂies) 3 1 1 1 1

Non-Insects

Amphipoda (scuds) 711

Bivalvia (clams) 14 11

Gastropoda (snails) 12 Hydracarina (water mites) 2 1 15 2 4 Oligochaeta (segmented worms) Naididae 22 26 25 15 20

Tubiﬁcidae 20 20 12 20 7 Hirudinea (leeches) 121 Nematoda (roundworms) 20 34 7 20 9

Figure 2). The number of zooplankton per m3 in- Table 3. The relative abundance of fish species as a proportion of the total collected from each station. DFOW = distance from creased from the open water/macrophyte edge into open water into the macrophyte bed. the middle of the wetland, and then declined at the two most shoreward stations. There was also a highly TAXA DFOW = 60 m 120 m 320 m significant spatial trend in zooplankton taxa richness Alosa pseudoharengus (alewife) 72 30 74 (quadratic p < 0.01, Figure 2). Both the mean and me- Cyprinus carpio (common carp) 1 4 6 dian number of taxa increased from the macrophyte Dorosoma cepedianum (gizzard shad) 9 17 6 edge to maxima between 180 and 280 m from open Lepomis spp. (sunfishes) 1 33 6 water. Values then declined at the most shoreward Micropterus salmoides (largemouth bass) 2 5 5 station. Morone americana (white perch) 2 < 1 < 1 Significant spatial trends were noted for 65% of Notropis spp. (shiners) 6 8 1 all zooplankton taxa collected during the study (11 Perca flavescens (yellow perch) 7 2 1 of 17 taxa). The abundance of five taxa increased Other 1 < 1 < 1 linearly with distance from open water, while the abundance of two taxa decreased linearly (p < 0.05, Figure 3). Four taxa exhibited their greatest abundance in the middle of the wetland similar to that of Bosmina longirostris dominated the Cladocera. Sida zooplankton as a whole (quadratic p < 0.05, Figure and Bosmina decreased substantially in relative abun- 3). These spatial trends led to notable differences in dance towards shore. In the middle of the wetland, the community composition of zooplankton through- Alona quadrangualris and Macrothrix sp(p). became out the wetland. Cladocera were the most abundant increasingly abundant. At shoreward stations Cerio- of the four major taxa (Cladocera, Ostrocoda, Cy- daphnia sp(p)., Acroperus harpae,andSimocephalus clopoida, and Calanoida) representing 47–73% of total sp(p). were the most abundant cladocerans. Ostracod abundance throughout the stand (Table 1). At bay- zooplankton increased notably in relative abundance ward stations, Ceriodaphnia sp., Sida crystallina,and into the middle of the wetland (32%) and then declined

wetmit19.tex; 19/01/1999; 16:57; p.5 64

Figure 3. Spatial distributions of individual zooplankton taxa. Top and middle graphs show significant linear trends (p < 0.05), whereas bottom displays significant quadratic trends (p < 0.05). Each data point is the mean of five sampling dates (June–Oct., 1994). Error bars are omitted for clarity. at more shoreward stations. Cyclopoid copepods in- of total abundance at bayward stations, but also increased in relative abundance towards shore where creased linearly towards shore (p < 0.01) where it was they represented 25–41% of all zooplankton collected. the numerically dominant taxon (Table 2). Two other macro-invertebrate taxa, Tanypodinae and Hydraca- Benthic invertebrates rina, had significantly higher abundance in the middle of the wetland (quadratic p < 0.05, Figure 4). There was no significant spatial trend in the total density of benthic macro-invertebrates from open wa- Fish ter towards shore. However, there was a significant linear increase in taxa richness (p < 0.05, Figure 2). A total of 4,525 fish were collected in 116 hours Significant spatial trends were noted for four of the of sampling. Catch per unit effort declined linearly fourteen taxa considered (29%, Figure 4). Amphipods, with distance from open water (p < 0.01, Figure 2). including both Gammarus sp. and Hyallela azteca, Compared to nets placed at the bayward edge of the were never found in the outer half of the wetland wetland, catch in the most shoreward fyke net was re- (20–180 m from open water) but increased linearly duced by 78%. Taxa richness showed a similar trend, (p < 0.05) towards shore where they represented 11% declining from an average 8.4 taxa at 60 m from open of total density (Table 2). Ephemeroptera, composed water to 5.1 at the most shoreward station (linear entirely of Caenis sp., represented no more than 2% p < 0.01, Figure 2). Approximately 67% of all fish

wetmit19.tex; 19/01/1999; 16:57; p.6 65

Figure 4. Spatial distributions of individual benthic Figure 5. Significant linear trends (p < 0.05) in the spatial distri- macro-invertebrate taxa. Top graph shows significant linear bution of individual fish taxa. Each data point is the mean of ten trends (p < 0.05), whereas bottom displays significant quadratic sampling efforts (see text for details). Error bars are omitted for trends (p < 0.05). Each data point is the mean of five sampling clarity. dates (June–Oct., 1994). Error bars are omitted for clarity.

that incorporate numerous wetland complexes. While taxa collected during the five months of the study (6 such studies provide valuable insight into the factors of 9+) displayed significant linear trends in abundance that potentially structure communities between wet- (p < 0.05), with each decreasing in number from the land systems, they offer no information about the local wetland edge towards the shore (Figure 5). These factors that are important within a wetland system. An trends led to differences in the fish assemblages at dif- understanding of the latter is crucial to conservation ferent distances from open water (Table 3). Alewife or restoration efforts which are typically performed on (Alosa pseudoharengus) were the most common fish relatively small spatial scales. collected during the study, comprising 72–74% of total In coastal wetlands of Saginaw Bay, Lake Huron, catch at the shoreward and bayward edges of the wet- there is growing evidence that the spatial distributions land. In the middle of the wetland, sunfishes (Lepomis of flora and fauna are related to distance from the open spp.) were slightly more abundant than alewife. water/wetland edge. Suzuki et al. (1995) found that phytoplanktonic and epiphytic algal biomass declined by 80–97% from open water towards shore. We found Discussion similar declines in the biomass and net primary production of epiphytic algae in a nearby wetland (Cardi- Surprisingly little is known about the environmental nale et al., 1997). We also found that the biomass and factors that influence the structure of communities diversity of epiphytic invertebrates decreased dramat- in Great Lakes coastal wetlands. A number of re- ically from open water towards shore. In part, this was cent studies have made significant contributions to due to the complete absence of filter-feeding inverte- our knowledge (for example Smith et al., 1991; Jude brates beyond 180 m from open water (Brady et al., and Pappas, 1992; Randall et al., 1996; Brazner and 1995; Cardinale et al., 1997). Beals, 1997), however, these have primarily focused Data from this study have shown that several other on environmental gradients over large spatial scales groups of fauna also have spatial distributions that are

wetmit19.tex; 19/01/1999; 16:57; p.7 66 correlated with distance from open water. We docu- epiphytic scrapers (Acroperus harpae and Pleuroxus mented distinct spatial trends in the abundance and spp.). Other studies have noted similar distributions diversity of crustacean zooplankton and fish, as well as for these zooplankton (cyclopoids – Cyr and Down- the diversity of benthic macro-invertebrates. Of the 40 ing, 1988; Lalonde and Downing, 1992; D. birgei and taxa examined, 53% had significant horizontal trends Pleuroxus spp – Krieger and Klarer, 1991). Numerous in abundance over the five months of the investigation. factors may have contributed to the distributions of When these data are coupled with previous studies, these taxa in our site, however, we speculate that the it appears that distance from open water is a major most important included (1) a reduced risk of preda- determinant of community structure in these coastal tion by zooplanktivorous fish in the nearshore area of wetlands. the macrophyte bed, particularly for the large clado- The most pertinent question that ensues from our cerans and the actively swimming cyclopoids, (2) study is, “why is distance from open water important mechanical disturbance in the wave-swept, bayward to the organisms?” Unfortunately, because our study portion of the wetland that precluded colonization of was simply ‘pattern searching’ it is impossible to de- macrophytes by the epiphytic scrapers which lack any termine the causes of the distributions observed. We form of secure attachment (Dodson and Frey, 1991), do think, however, that they may be related to en- and (3) decreased turbidity in the nearshore area which vironmental gradients that were documented in this led to reduced interference of filter-feeding by larger site at the same time as this study. Using chloride cladocerans (Kirk and Gilbert, 1990). as a conservative tracer, we showed that a gradient Unlike the other epiphytic micro-crustacea, Sida in mixing between pelagic and littoral water was es- crystallina uses a strong “suction-cup” gland for at- tablished as surface waves were gradually inhibited tachment to macrophytes. This taxon was a dominant by the emergent vegetation (Cardinale et al., 1997). component of the zooplankton community in the bay- In turn, reduced mixing led to declines in turbidity, ward and middle portions of the wetland, but rep- dissolved oxygen, and pH at shoreward stations, but resented only a minor component of the nearshore increased conductivity and concentrations of associ- community – similar to the distribution shown by ated ions. In the remainder of the discussion we will Fairchild (1981). Because this epiphytic filter-feeder speculate on how these environmental changes may is at less risk of predation by visual predators than have influenced the abundance and diversity of faunal are planktonic filter-feeders (Fairchild, 1981), it is groups in this site. Our hope is that these hypothe- possible that S. crystalline distribution resulted from ses will serve as a starting point for future studies on phytoplanktonic food availability. This hypothesis is communities inhabiting Great Lakes coastal wetlands. consistent with our previous results which have shown Overall, zooplankton abundance was highest in similar spatial trends for numerous other epiphytic the middle of the wetland (180 m from open water) filter-feeders in this site (Brady et al, 1995; Cardinale in an area that corresponded to the furthest extent et al., 1997). of wave penetration and water circulation (Cardinale Many of the small, burrowing macro-invertebrates et al., 1997). Several of the most abundant taxa in (Chironomidae, Nematoda, Tubificidae, and Naidi- this area were small-bodied zooplankton that are weak dae) showed no spatial trends throughout the wetland swimmers (ostracods – Delorme, 1991; and chydorid during our study. In contrast, all of the larger, mo- Cladocera – Dodson and Frey, 1991). The distribution bile benthic invertebrates did exhibit significant spatial of these zooplankton may have resulted from limits trends. For example, the maximum abundance of to planktonic dispersal directly mediated by water cir- Tanypodinae and Hydracarina occurred in the middle culation. This possibility is consistent with our past of the macrophyte bed, the same area where ostracod results which have shown that passively distributed and small cladoceran abundance was highest. Both veligers of the zebra mussel (Dreissena polymorpha) of these taxa are highly mobile predators of zoo- reached highest densities in the middle of the marsh plankton and/or zooplankton eggs (Smith and Cook, as decreased water circulation prevented them from 1991; Berg, 1995). Because ostracods are the only colonizing more nearshore areas (Brady et al., 1995). crustacean zooplankton which deposit their eggs onto Zooplankton groups with greatest abundance in substrates (Delorme, 1991), we suspect that tanypod the nearshore portion of this coastal wetland in- chironomids and water mites may have simply distrib- cluded cyclopoid copepods, large cladocerans (Simo- uted themselves according to the availability of their cephalus spp. and Diaphanosoma birgei), and small prey.

wetmit19.tex; 19/01/1999; 16:57; p.8 67

Amphipoda and the mayfly nymph, Caenis sp., were more likely to be found in the middle and shore- are also relatively large, mobile taxa. Both of these ward portions of the marsh. It may be that reduced macro-invertebrateswere rarely found in the outer por- turbidity indirectly influences the distribution of cer- tion of the wetland, and yet each represented a major tain fish species by permitting an increase in habitat component of the benthic invertebrate community at complexity via submergent macrophyte growth. shoreward stations. The size and activity of these two In summary, we have shown that the abundance taxa make them highly susceptible to fish predation, and diversity of numerous faunal groups in this coastal and indeed, they often represent a significant portion wetland were correlated with distance from open wa- of fish diets (Crowder and Cooper, 1982; Mittlebach, ter. We have speculated these distributions may have 1988). Like many of the shoreward zooplankton, the been related to several abiotic changes that took place distribution of these taxa may have been mediated by from open water towards shore as pelagic surface differential risks of predation throughout the wetland. waves were gradually diminished into the stand. To More than 75% of Great Lakes fish species utilize the extent our speculations are correct, our results re- coastal wetlands during some portion of their life cycle emphasize the contention of past researchers that con- (Jude and Pappas, 1992). Our results showed that fish nectedness between coastal wetland and open water did not equally utilize all portions of this coastal marsh habitat is imperative to maintain community struc- habitat. Most species, including several commercially ture (Jude and Pappas, 1992; Patterson and Whillans, important taxa, displayed preferences for the outer 1985; Sager et al., 1985). This contention has sev- third of the wetland that was adjacent to the pelagic eral important implications for the conservation and zone. A number of factors may have contributed to the restoration of coastal wetlands in the Great Lakes. If disproportionate importance of this area. Turbidity is wetland biota are indeed influenced by abiotic gradi- often cited as a primary predictor of the abundance of ents established by the influx of pelagic surface waves, gizzard shad, shiners, yellow perch, and alewife (for then any land use practice that changes the propor- example Swenson, 1978; Brazner and Beals, 1997). tion of wetland habitat experiencing mixing with open Our results were consistent with these previous stud- water will likely alter the abundance and composition ies. Increased turbidity in the wave-swept portion of of organisms that inhabit the wetland. Thus, frag- this site may reduce the risk of predation by piscivo- mentation (for example, the cutting of channels for rous fish, particularly for open-water foraging species drainage or boating access) will likely increase circu- that utilize wetlands as a temporary refuge from preda- lation of lake water into nearshore wetland areas and, tors (Petering and Johnson, 1991; Jude and Pappas, in turn, alter invertebrate and fish community com- 1992). Alternatively, because turbidity was correlated position. Furthermore, any restoration effort that fails to higher primary production and secondary produc- to restore connectedness between the wetland and the tion of certain invertebrate taxa during this study open water habitat will probably fail to restore wetland (Cardinale et al., 1997), these fish species may have community structure. responded to a gradient in productivity. Macrophyte density and structural complexity are also important factors influencing the distribution of Acknowledgments certain fish species in littoral zones (Crowder and Cooper, 1982; Brazner and Beals, 1997). The den- Funding for this project was provided by the Michi- sity of emergent macrophytes was relatively uniform gan Department of Natural Resources, and the United in our site at distances beyond 40 m from open wa- States EPA in grants awarded to T. Burton. We thank ter (Cardinale, 1996). However, Suzuki et al. (1995) three anonymous reviewers for comments that sub- noted that submergent macrophyte density and di- stantially improved this manuscript. versity increased towards shore in a nearby wetland, presumably as declines in turbidity permitted growth. References Sunfishes are commonly associated with submergent macrophyte diversity (Crowder and Cooper, 1982), Anderson, R.O. 1959. A modified flotation technique for sorting and indeed, we found a tendency for higher catches bottom fauna samples. Limnol. Oceanogr. 4: 223–225. of sunfishes in the middle of the wetland. Similarly, Berg, M.B. 1995. Larval food and feeding behavior. In: Armitage, largemouth bass (which are also commonly associated P., Cranston, P.S. and L.C.V. Pinder (eds). The Chironomidae: the Biology and Ecology of the Non-Biting Midges. pp. 136-168. with habitat complexity – Brazner and Beals, 1997) Chapman and Hall, New York.

wetmit19.tex; 19/01/1999; 16:57; p.9 68

Brady, V.J., Cardinale, B.J. and Burton, T.M. 1995. Zebra mussels Lalonde, S. and Downing, J.A. 1992. Phytofauna of eleven macro- in a coastal marsh: the seasonal and spatial limits of colonization. phyte beds of differing trophic status, depth, and composition. J. Great Lakes Res. 21: 587–593. Can. J. Fish. Aquat. Sci. 49: 992–1000. Brazner, J.C. and Beals, E.W. 1997. Patterns in fish assemblages Mitsch, W.J. and Gosselink, J.G. 1993. Wetlands. 2nd edn. Van from coastal wetland and beach habitats in Green Bay, Lake Nostrand Reinhold, New York. Michigan: a multivariate analysis of abiotic and biotic forcing Mittleback, G.G. 1988. Competition among refuging sunfishes and factors. Can. J. Fish. Aquat. Sci. 54: 1743–1759. effects of fish density on littoral zone invertebrates. Ecology 63: Cardinale, B.J. 1996. The effects of a pelagic-littoral mixing gradi- 614–623. ent on an epiphytic invertebrate community. M.S. thesis, Depart- Mittleback, G.G. 1981. Patterns of invertebrates size and abundance ment of Fisheries and Wildlife, Michigan State University, East in aquatic habitats. Can. J. Fish. Aquat. Sci. 38: 896–904. Lansing, MI. Patterson, N.J. and Whillans, T.H. 1985. Human interference with Cardinale, B.J., Burton, T.M. and Brady, V.J. 1997. The com- natural water level regimes in the context of other cultural munity dynamics of epiphytic midge larvae across the pelagic- stresses on Great Lakes wetlands. In: Prince, H.H. and DíItri, littoral interface: do animals respond to changes in the abiotic F.M. (eds), Coastal Wetlands. pp. 209–239. Lewis Publishers, environment? Can. J. Fish. Aquat. Sci. 54: 1–9. Chelsea. Comer, A.J., Albert, D.A., Wells, H.A., Hart, B.L., Raab, S.B., Petering, R.W. and Johnson, D.L. 1991. Distribution of fish lar- Price, D.L., Kashian, D.M., Corner, R.A. and Shuen, D.W. 1995. vae among artificial vegetation in a diked Lake Erie wetland. Michigan’s Native Landscape, as Interpreted from General Land Wetlands 11: 123–139. Office Surveys, 1816–1856. Michigan Natural Features Inven- Randall, R.G., Minns, C.K., Cairns, V.W. and Moore, J.E. 1996. The tory Report to Water Division, U.S. EPA and Wildlife Division, relationship between an index of fish production and submerged Michigan Dept. of Nat. Res. Lansing, MI. macrophytes and other habitat features at three littoral areas in Crowder, L.B. and Cooper, W.E. 1982. Habitat structural complex- the Great Lakes. Can. J. Fish. Aquat. Sci. 53(Suppl. 1): 35–44. ity and the interaction between bluegills and their prey. Ecology Sager, P.E., Richman, S., Harris, H.J. and Fewless, G. 1985. Prelimi- 63: 1802–1813. nary observations on the seiche-induced flux of carbon, nitrogen, Cyr, H. and Downing, J.A. 1988. Empirical relationships of phy- and phosphorous in a Great Lakes coastal marsh. In: Prince, tomacrofaunal abundance to plant biomass and macrophyte bed H.H. and Díltri, F.M. (eds), Coastal Wetlands. pp. 59–66. Lewis characteristics. Can. J. Fish. Aquat. Sci. 45: 976–984. Publishers, Chelsea. Delorme, L.D. 1991. Ostracoda. In: Thorp, J.H. and A.P. Covich Smith, I. and Cook, D. 1991. Water Mites. In: Thorp, J.H. and (eds), Ecology and Classification of North American Freshwater Covich, A.P. (eds), Ecology and Classification of North Ameri- Invertebrates. pp. 691–722. Academic Press, San Diego. can Freshwater Invertebrates. pp. 587–786. Academic Press, San Dodson, S. and Frey, D.G. 1991. Cladocera and other Branchiopoda. Diego. In: Thorp, J.H. and A.P. Covich (eds), Ecology and Classifica- Smith, P.G.R., Glooschenko, V. and Hagen, D.A. 1991. Coastal tion of North American Freshwater Invertebrates. pp. 587–786. wetlands of three Canadian Great Lakes: inventory, current Academic Press, San Diego. conservation initiatives, and patterns of variation. Can. J. Fish. Fairchild, G.W. 1981. Movement and microdistribution of Sida crys- Aquat. Sci. 48: 1581–1593. tallina and other littoral microcrustacea. Ecology 62: 1341–1352. Stuckey, R.L. 1989. Western Lake Erie aquatic and wetland vascular Jude, D.J. and Pappas, J. 1992. Fish utilization of Great Lakes plant flora: its origin and change. In: Krieger, K.A. (ed.), Lake coastal wetlands. J. Great Lakes Res. 18: 651–672. Erie Estuarine Systems: Issues, Resources, Stability, and Man- Kirk, K. and Gilbert, J.J. 1990. Suspended clay and the population agement. pp. 205–256. NOAA Estuary-of-the-Month Seminar dynamics of planktonic rotifers and cladocerans. Ecology 71: Series Vol 14. U.S. Department of Commerce, Washington, D.C. 1741–1755. Suzuki, N., Endoh, S., Kawashima, M., Itakura, Y., McNabb, Krieger, K.A. 1992. The ecology of invertebrates in Great Lakes C.D., D’Itri, F.M. and Batterson, T.R. 1995. Discontinuity bar coastal wetlands: Current knowledge and research needs. J. Great in a wetland of Lake Huron’s Saginaw Bay. J. Fresh. Ecol. 10: Lakes Res. 18: 634–650. 111–123. Krieger, K.A. and Klarer, D.M. 1991. Zooplankton dynamics in a Swenson, W.A. 1978. Influence of Turbidity on Fish Abundance Great Lakes coastal marsh. J. Great Lakes Res. 17: 255–269. in Western Lake Superior. Environmental Research Laboratory- Krieger, K.A., Klarer, D.M., Heath, R.T. and Herdendorf, C.E. Duluth, Office of Research and Development Report EPA-600/3- 1992. A call for research on Great Lakes coastal wetlands. J. 78-067. Great Lakes Res. 18: 525–528.

wetmit19.tex; 19/01/1999; 16:57; p.10