Metzger Marsh is a 300-ha wetland in an embayment in western Lake Erie, approximately 18 km east of Toledo, Ohio, USA. It lies within the boundaries of refuges managed by the U.S. Fish and Wildlife Service (USFWS) and the Ohio Division of Wildlife (ODW). Although early attempts to dike, drain, and farm portions of the wetland failed, the site has a history of human disturbance. It once formed the mouth of an unnamed creek that was channelized directly to the lake in the late 1800s (now without hydrologic connection to the wetland). The embayment was also formerly protected from waves on the lake by a barrier beach.

Aerial photographs from 1940 show an intact barrier beach, with 58% of the embayment vegetated. Erosion during ensuing periods of high lake levels reduced the extent of the barrier beach until it was completely lost during the extremely high levels of 1973.

Progressive loss of vegetated area accompanied erosion of the protective barrier, with 19% of the wetland vegetated in 1973 and 10% vegetated in 1994. Paleoecological studies suggest that pre-European-settlement vegetation was dominated by sedges and grasses; however, in recent decades, open water area increased, and vegetation was largely restricted to islands of cattails and common reed.

Interest in restoring Metzger Marsh grew from managers' desires to provide better wetland habitat in the near term and recognition of the limits of natural restorative processes in a highly disturbed environment. Extensive armoring of the U.S. shoreline of western Lake Erie to protect human property resulted in enough loss of sediments from the littoral drift of the lake that the barrier beach would likely never return. Lake Erie had also been in an extended period of high water levels, with no intervening low levels to expose sediments and allow revegetation from the seed bank.

The potential for natural revegetation of the wetland thus hinged on waiting for water levels to drop as much as a meter for at least one growing season and the unlikely reappearance of a barrier beach to provide protection. Instead, the management agencies opted for an active restoration program, with the initial intention of long-term operation of the site as a diked wetland. However, USFWS guidelines for taking an ecosystem approach to restoration and management, coupled with the required environmental assessment process for a federal action, resulted in a philosophical change in the program. The restoration effort would attempt to return the embayment to a vegetated, hydrologically connected, coastal marsh that provided multiple wetland functions and values.

The restoration program incorporated a dike to mimic the protective function of the lost barrier beach but included a water-control structure in the dike that could be opened following restoration to allow hydrologic connection with the lake similar to the original wetland. Dike construction was completed in 1995, and the control structure was installed in 1996. The control structure remained closed in 1996 and 1997 to allow a drawdown of water levels to mimic a low lake-level period. Rotenone was also applied by ODW at the lowest water stage to kill any remaining trapped common carp.

The seed bank produced a quick response, with 73% of the wetland revegetated in 1996, 82% revegetated in 1997, and 72% revegetated in 1998 when water levels were increased. Prominent recolonizing taxa included Cyperus odoratus, Polygonum lapathifolium, Echinochloa crusgalli, Leersia oryzoides, Bidens cernua, Typha angustifolia, Scirpus validus, and Sagittaria latifolia but also considerable Phragmites australis, some Lythrum salicaria, and a number of upland species such as Abutilon theophrasti. The timing of exposure of mudflats in 1996 also allowed airborne seeds of Populus deltoides, Salix cordata, S. exigua, and S. fragilis to germinate and grow across large areas of the wetland. The invasion of trees prompted experimental management actions by ODW. Portions of the tree-dominated areas were cut during the drawdown in 1997, and portions were sprayed with 2, 4-D in 1998. Some Lythrum salicaria was sprayed with glyphosate in 1997. On USFWS property, an experimental planting of Vallisneria americana tubers was conducted in 1997 in water approximately 20 to 50 cm deep, and exclosures made by stringing mylar tape between metal posts were tested in water <10 cm deep where herbivory on Typha seedlings by Canada geese was observed.

The wetland was reflooded in 1998 without lake connection, and the control structure was opened in March 1999. The water-control structure was designed with five 2-m-wide channels that can be closed individually. The size of the openings was based on calculations of the potential flow rates between the lake and the wetland that could be driven by seiches on the lake, given the volume of the wetland, the hydrologic head created by the seiche, and the period of the seiche. Seiches with an amplitude of about 0.2 m and period of about 14 hr occur regularly, with larger seiches of about 1.5-m amplitude and 18- to 24-hr period occurring occasionally. Sudden, storm-driven seiches have also resulted in water-level changes of about 0.75 m in 4 to 5 hours. Sizing of the opening was intended to prevent major dewatering and flooding within the wetland that could affect fish-spawning or bird-nesting areas. In reality, the water-control structure mediates the seiche effect just as the natural opening in the barrier beach once did. The 5- channel design provides an option for management changes when engineering plans are tested by actual conditions.

The water-control structure also contains an experimental fish-control system that allows direct wetland access by most small fish, yet restricts access by large carp while allowing passage of other large fish.

Each of the three central channels is spanned by a grate containing vertical bars spaced 5 cm apart (with cross-bars for stability). The design was based on studies at Cootes Paradise marsh on Lake Ontario indicating that 95% of the carp seeking to enter the wetland would be excluded. The two outer channels contain fish passageways, one for fish moving into the marsh and one for fish moving to the lake. The entry side of each passageway can be fitted with experimental grates of sizes and shapes that allow larger fish, such as northern pike to pass through, while minimizing the numbers of carp that gain entry, based on differences in body size and shape. The exit side of the passageway for entering the wetland is fitted with 5-cm vertical grates to prevent carp from proceeding into the wetland.

The passageways each contain fish baskets that can be lifted with an electric hoist. The entry fish basket is operated daily during open water seasons to capture, count, and measure fish and selectively move all but carp into the wetland. The exit fish passageway is operated as necessary to monitor and move large fish out of the wetland. The ability to change grates at both ends of the passageways also provides opportunities to study other sizes and shapes of fish that move in and out of the wetland.

Pre-restoration studies were completed to measure physical attributes (such as bathymetry) and to characterize the wetland plant communities and most major groups of fish and wildlife (fish, juvenile fish, shorebirds, waterfowl, small mammals, herpetofauna, invertebrates). Limited studies continued during the drawdown and initial reflooding years. All studies are scheduled to be repeated for multiple years following hydrologic reconnection with the lake. The results will be critical in evaluating the success of the techniques employed, developing any necessary modifications, and preparing guidelines for technology transfer to other managers to assist them in opening diked wetlands, increasing wetland functions, and improving habitat for a variety of both fish and wildlife. Once completed, the Metzger Marsh restoration project is expected to serve as a model for future coastal marsh restoration projects in locations with severe sediment deficits and could guide efforts to hydrologically reconnect and manage other marshes that are currently diked.

Email Kurt Kowalski , Research Ecologist, Ecosystem Health and Restoration