Continue to document the evolution of Metzger Marsh's fish assemblage
Compare Metzger Marsh's fish assemblage to other regional marshes, sampled on previous projects

Metzger Marsh is a 367-ha wetland on Lake Erie’s southwest shore. Historically, the site was protected by a barrier beach that was destroyed by anthropogenic reduction in sediment supply and fluctuating lake levels; the original beach was replaced by an experimental dike designed to be open to the lake’s hydrology while excluding adult common carp (Cyprinus carpio). Construction was completed in 1995; the marsh was reflooded in 1998 and reconnected to Lake Erie in 1999. We sampled with larval pushnets, trapnets, and electrofishers, both pre- and post-restoration. We collected 46 species (15 families) as well as 18 larval taxa. Overall catch rates have not changed substantially, but the fish assemblage’s structure has changed dramatically. Non-metric Multidimensional Scaling separated pre- from post-restoration samples without overlap in ordination space. After restoration, the marsh’s adult fish assemblage shifted from one of low diversity dominated by tolerant, lake-associated fishes (especially clupeids) to one of high diversity with good representation of phytophilic and game species (especially centrarchids). Ordination of ichthyoplankton catches did not reveal obvious patterns. In general, lake-associated larvae maintained consistent numbers; phytophilic taxa increased. The scarcity of well-vegetated wetlands accessible by Lake Erie’s fishes makes Metzger Marsh a potentially important resource.

Three-fourths of the nation's fish production depends on wetlands, thus efforts to restore and reconnect wetlands for the purpose of increasing fish habitat are critical. Clearing and draining has reduced wetland area in the United States by more than 400,000 ha.m, and wetland loss has negatively impacted wetland fishes in many locations.

Great Lakes wetlands have been affected by land-use practices in the watershed, reduced water quality, invasion by alien species, and nutrient enrichment. Shoreline stabilization in service to agriculture and development in general has resulted in the draining and destruction of a large proportion of the vast original Great Lakes wetland system. Herdendorf (1992) estimates that of the over 4,000 km² of original coastal marshes and swamps on western Lake Erie, only 100 km² remain.

Largely in an effort to protect wildlife habitat from further reductions, both private and governmental organizations have assumed control of many of the remaining coastal marshes on western Lake Erie (Johnson 1989; Herdendorf 1992). These marshes have often been diked to protect against the action of wind and waves, water level fluctuations, seasonal seiches, and allow for greater control of marsh water levels to enhance wildlife habitat (Herdendorf 1987; Johnson 1989; Herdendorf 1992; Mitsch 1992; Mitsch et al. 1994); 90% of Ohio's remaining marshes are so controlled (Bookhout et al. 1989). While protecting Lake Erie's coastal marshes from complete loss, diking also dramatically changes the morphology and function of these marshes.

Current management practices rely heavily on dike construction and continuous control of water levels. Wetland managers usually decide to remove water in the spring and add it in the fall. This practice alters the plant assemblage's in these wetlands in favor of facultative wetland plants and terrestrial vegetation to the detriment of more obligate wetland species (Johnson et al. in revision). In addition, diking compromises these wetlands as potential spawning, nursery, or feeding areas for many Lake Erie fishes. Diking also effectively isolates a wetland from the landscape and has the potential to eliminate many other wetland functions, such as pollutant control, because no functional hydrologic connection exists between the wetland and adjacent water bodies.

Diked marshes are demonstrably isolated from the main lake as indicated by differences in water quality (Mitsch 1992; Mitsch et al. 1994) and by slower growth of the same species when compared to fish immediately outside the dike (Markham et al. 1997). Lake Erie's traditional diked wetlands typically have fish assemblages dominated by tolerant species such as common carp (Cyprinus carpio), goldfish (Carassius auratus), bullheads (Ameiurusspp.), white crappie (Pomoxis annularis), and gizzard shad (Dorosoma cepedianum) while sustaining few predatory fishes or those of recreational or commercial value (Johnson et al. 1997b). We have shown fish exclusion imposed by dikes reduces fish numbers, diversity, and condition in spite of increasing habitat diversity (Johnson 1989; Johnson et al. 1997a).

Great Lakes wetlands have been historically dynamic, shifting inland and lakeward depending on water level. In the last several decades, diking to protect the mainland as well as to preserve wetland habitat for wildlife has reduced the shifting nature of these wetlands. In addition to the long-term fluctuations in water levels, shorter period seiches also occur on the Great Lakes. Coastal wetlands along the lakes were historically subjected to water exchanges from seiches in much the same way that coastal salt marshes are subjected to tides although seiches are not as periodic or predictable as coastal tides. The importance of short-term lake water level fluctuation to the biotic communities of Great Lakes wetlands is not well known although it may be analogous to the role of tides in saltwater wetlands.

Metzger Marsh represents a unique restoration effort. The remarkable aspects of the marsh's fish assemblage may be attributable to the hydrologic connection with the lake and the exclusion of adult common carp during part of the year, attributes not usually associated with traditionally diked wetlands. The scarcity of well-vegetated coastal wetlands accessible by Lake Erie's fish populations makes Metzger Marsh a potentially important resource, one worthy of continued study and evaluation.

Data collection
We collected monthly samples at Metzger Marsh from May to September2003 at six stations around the marsh with both fyke nets and electrofishing gear. Fyke nets with 30-m leads and 1.3-cm mesh were set for approximately 24 hr at each station: catch represented as fish per hour. We fished each station with our boat-mounted electrofisher running 250-275 V adjusted to maintain ca. 4-6 A. Electrofishing runs were 15 min long; we measured transect length with a GPS unit for conversion to catch per meter. Catch from both gears were identified to species, enumerated, measured to the nearest mm, and released.

Data analysis
We calculated diversity, both Simpson's (1/λ) and Shannon-Wiener's (e^H) indices, as well as conducting basic guild analysis of the catch for comparison to previous years of data collection. In addition, data will be combined with that from previous years in an ordination plot of nonmetric multidimensional scaling (NMS) using PC-ORD to describe the progression of the fish assemblage through time.

Dissemination
Metzger Marsh represents a unique and important restoration effort. Because success impacts both recreationists and professional ecologists/managers in informing future restoration efforts, the outcome of this restoration needs to be investigated and communicated to both professional and lay audiences. Results will be disseminated through professional journals, regional and national meetings, and Sea Grant publications. Our findings will contribute to OSU Extension fact sheets on the importance of coastal wetlands to lake productivity and especially fish recruitment.

A wetland fish assemblage: Wetland definitions depend upon three criteria: 1) the presence of water, 2) the presence of hydric soil, and 3) the presence of wetland biota (Mitsch & Gosselink 2000).  The association of macrophyte species to wet soils is defined by the USFWS (Reed 1988).  Fishes are not so easily categorized, and there is less of a consensus regarding what constitutes a wetland species of fish.  It is a fairly safe argument that a fish species’ status as wetland-dependent largely depends upon the presence of macrophytic vegetation or complex habitat structure.  Fish are mobile and their association with wetlands may be seasonal and/or brief, but this association may still play a critical role in their life cycles.  For example, esocids—some of which are large and important game fishes—require last season’s vegetation to receive their spawn for successful reproduction.

One of the goals in restoring Metzger Marsh was to create a more “wetland-like” community, and the marsh’s fish assemblage seems to indicate that such a thing is evolving there.  When we first sampled Metzger Marsh in 1994 it was an open embayment subject to storm surges, wave action, and fluctuating water levels.  Its sparse vegetation was limited to the most protected areas and the areas most open to the lake exhibited sand and gravel substrates typical of a region subject to high-energy wave action.  The fish assemblage reflected these conditions with most being lake-associated species. After diking, the combination of protection from wave action and the development of a significant vegetation assemblage had a strong and anticipated impact on the fish assemblage.  Only a short period of time was required for the dominant open-lake species present before dike construction to be mostly replaced by phytophilic and centrarchid species (centrarchid presence is used as an indicator of good habitat structure by the Ohio Environmental Protection Agency (1990)).  This shift in assemblage structure is so complete that all pre- and post-restoration restoration samples are separated in ordination space with no overlap (fig. 1).  The fact that open-lake fishes such as the emerald shiner have persisted as important species is unique in diked wetlands, as is the presence and importance of major unstocked piscivores, in particular northern pike and largemouth bass. 

A unique system: In most respects, the Metzger Marsh fish assemblage shows development atypical of traditional diked marshes.  Most dramatic has been the highly significant increase in fish diversity over the years of the study; in no instance of our sampling of dozens of diked wetlands since 1984 have we measured diversity of this magnitude.  Diversity indices respond to increases in species richness (number of species) as well as evenness of distribution, but dramatic increases in diversity have occurred within Metzger Marsh almost entirely through the new assemblage becoming more evenly distributed across species.  It is likely that we are underestimating both species richness and diversity since most fish-sampling gear is not very effective in dense vegetation.  The vegetation is dense enough that some of our sampling stations at the back of the marsh have had to be abandoned.

One of our objectives in planning the open-but-guarded wetland was to increase species diversity, and that goal has certainly been met thus far.  While diversity is not necessarily essential to a quality wetland, it is usually understood that diversity of organisms indicates a diversity of functions being met (Zedler 2001).  In that sense, diversity of fishes is one measure of the complexity of wetland function.

The hydrologic connection with the lake is the feature most likely to have allowed Metzger Marsh to develop not only a strong phytophilic fish base, but also to maintain a good representation of lake species.  In particular, the high proportion of gizzard shad, alewife, and emerald shiner is unusual in a diked wetland.  These fishes are usually found around the opening in the wetland and along the dike’s inside edge.  It is likely that they serve as an important food base for predators within the wetland.  In contrast, most diked wetlands start with a mixed assemblage of lake and wetland species, but the species composition shifts toward tolerant species such as carp, goldfish, bullhead, gizzard shad, and white crappie.  Less tolerant lake associated species are not usually sampled within diked marshes (Johnson et al. 1997).

It appears that Metzger Marsh’s connection to Lake Erie results in a larger piscivore population than is common in most Lake Erie wetlands.  Largemouth bass are quite uncommon in most lake-connected (Kleber 2004) and almost absent in isolated Lake Erie wetlands (Johnson et al.  1997a).  In fact, predators make up a higher proportion (0.13-0.20) in post-restoration Metzger Marsh than would be expected in most natural systems.  In stocking programs predators comprise only about 10% of a fish assemblage (Dillard and Novinger 1975). At some stations within the marsh, we caught as many as 204 largemouth bass per hour of electrofishing.  A catch of 100 per hour is considered a dense largemouth bass population (Flickenger et al. 1999).

Personal observations and casual discussions with anglers at Metzger Marsh indicate that a substantial largemouth bass fishery has developed at the marsh.  Examination of the ratio of largemouth bass above 200 mm as compared to fish above 300 mm (Proportional Stock Density) shows 35% of the largemouth measured in August of 2000 to be above 300 mm, while only 7% were in that category in 2001.  While the reason for this mortality of larger adults is unknown, our experience indicates that harvest is a likely cause.

The appearance of northern pike in Metzger Marsh is unique in our 15-year sampling of wetlands.  Northern pike are dependent on senesced wetland vegetation for early spring spawning in combination with access to the Lake, and we have not sampled them in any other diked wetlands except those in which they have been stocked.  The northern pike population has increased each year after dike completion and appears to be self-sustaining to this point.  The combination of the northern pike feeding, which peaks in the cooler temperatures of spring and fall with largemouth bass, which are most active in the summer, creates an exceptional predation pressure on the prey base.  With the excellent diversity of prey in the system, sustained growth of this large predator population is possible.  The strong predator presence in conjunction with the protective grates may help control soft-rayed fishes such as gizzard shad, common carp, and goldfish, improving wetland water quality through reduction in turbidity.

Metzger Marsh management

The diversity of fish species in combination with extensive wetland vegetation, relatively clear water, and constant lake influence may make Metzger Marsh an ideal location for the introduction of uncommon, rare, or extirpated fish species if managed as an open system. Many species that have been seriously reduced are dependent on clear, vegetated water: a resource in short supply in Ohio.  The combination of federal and state supervision of this facility as well as its flexibility in regard to water manipulation, protective grates to exclude common carp, and management of public access made it an excellent research site.

After four years of mandated Lake Erie connection, the gates of Metzger Marsh were closed in the fall of 2003.  The closure was for the management of invasive Phragmites sp. through water level manipulation.  Hopefully that closure will be short-term, since we would expect fish diversity to decrease over time and the export of fish larvae will be non-existent.  The closed wetland fish assemblage will follow one of two courses based on our research of closed dike systems: either 1) the predators trapped in the wetland will be able to control the common carp and goldfish, creating a centrarchid-based fish assemblage with few representatives of lake fishes or 2) the carp and goldfish will start to create a steady-state turbid system which will favor their continued dominance.  The vegetation will start to decline, allowing early spring turbidity and a further reduction in vegetation.  The resultant fish assemblage will be composed of tolerant fishes including common carp, goldfish, gizzard shad, white crappie, and bullhead.  In either case, the assemblage’s diversity could be expected to decline.

Metzger Marsh represents a new management concept for front-line Lake Erie wetlands.  It represents an optimum use concept rather than the more traditional maximum use style.  Certainly, not all wetlands can be opened to the lake.  What we need are new wetlands, not conflict over the few we have.  Metzger Marsh has shown in its initial development that it can provide fishing, hunting, excellent fish diversity, spawning and nursery habitat for both wetland and open lake fish species, and a hydrologic connection that has become too rare.  If front-line wetlands on the lake can use seiche currents to provide pulse benefits through the open-but-guarded system, then valuable resources needed to pump water could be reserved for the interior wetlands, thus providing a mosaic of wetland types useful to more mobile resources.