Northern Prairie Wildlife Research Center
Devastating alterations have taken place in wetland habitats in and around refuges across the U.S. In the northern Great Plains, both private and public wetlands still are very productive and contribute to the continent's waterfowl population. They also provide necessary migration and breeding habitats for numerous nongame birds. However, drainage and the destruction of associated uplands have led to regional extirpations of birds (Delphey and Dinsmore 1993). Refuge wetlands, consequently, have become more valuable to faunal communities.
Many nongame vertebrates rely on a variety of wetlands for some part of their life history needs (Duebbert 1981); others are totally reliant on wetland habitats (Gibbs 1993). Considerably less information is available for birds in wetlands than in upland habitats.
Survey methods for large semipermanent wetland basins with tall, monotypic emergent vegetation have not been effectively standardized (Reynolds et al. 1980, Edwards et al. 1981). Often, multiple survey techniques are used; Manci (1985), for example, compared the three techniques of airboat transects, 1-ha semicircular plot counts, and road-side transects at Horicon Marsh, Wisconsin.
Edwards et al. (1981) found that a variable-circular plot method could effectively sample large, continuous habitats. This was a technique that estimated the distance of each detected bird in 10-m concentric bands extending away from the observer, and it accounted for a greater number of species and greater occurrences of uncommon species than other methods, they reported.
We used a semicircular plot with a fixed radius of 75 m. Semicircular plots were easier to use than centering an observer in a homogeneous stand of emergent Typha spp which often were over 2 meters above water surface.
Nongame birds were surveyed between 15 May and 4 July 1996. Survey plots (n = 116) were placed in representative stands of emergent vegetation.
In some instances, edges of habitats were surveyed to determine species occurrence and abundance in these areas. However, when homogeneous stands of habitat were surveyed, edges were avoided.
Wetland habitats surveyed included semipermanently flooded emergent wetland vegetation (PEMF, Typha spp and Phragmites australis) and seasonally and temporarily flooded forested vegetation (PFOC, cottonwood, willow [Salix spp], and Russian olive [Eleagnus angustifolia]) according to the classification of Cowardin et al. (1979).
National Wetland Inventory maps showed 11 different types of wetlands within the refuge. However, continuous high water for 5 years (1992-1996) changed the characteristics of SLNWR wetlands. During the study, they basically functioned as a semipermanent wetland. Cattails and common reed accounted for the vast majority of the tall emergent vegetation.
Therefore, primary habitat types of the wetland area on SLNWR were grouped as emergent vegetation (n = 106, Fig 5) and flooded forested vegetation (n = 10, Fig 6).
|Figure 5. Location of semi-circular bird survey plots in semipermanently flooded emergent wetlands at Sand Lake National Wildlife Refuge, Brown County, South Dakota.||Figure 6. Location of semi-circular bird surveys in seasonally flooded forested wetlands at Sand Lake National Wildlife Refuge, Brown County, South Dakota.|
Plots were surveyed for birds while wading or from a 16-foot boat. The boat also became an observation vantage point for surveys in vegetation which often exceeded 1.5 to 2.0 meters above the water surface.
An attempt was made to survey all wetland area at SLNWR. However, the large Franklin's gull colony south of SD Highway 10 was avoided to minimize disturbances. Nor were transect surveys via airboats conducted, reducing the probability of auditory disturbance (Manci 1985).
After a 3-minute "waiting period" upon arrival at a site for birds to become accustomed to the presence of a human (Bollinger et al. 1988), birds were surveyed for 10 minutes at each plot location. This time period enabled an observer to account for all birds present and to detect a greater number of uncommon birds. All birds heard singing or seen perched, flushed, or flying over the plot were recorded by species and sex (Reynolds et al. 1980). The first 3 minutes of a survey period were spent counting all birds in the plot area. During the next 3 minutes, recorded continuous loop tapes (Library of Natural Sounds, Cornell Laboratory of Ornithology, Ithaca, NY 14850) of territorial male calls were played to elicit responses from secretive birds (Marion et al. 1981) including American bitterns, least bitterns, soras, and Virginia rails. The remaining 4 minutes were spent listening and watching for any previously undetected birds.
Surveys were not conducted on days with heavy precipitation, low temperatures, or excessive winds (≥20 km/hr) (Mikol 1980). On suitable days, plots were surveyed for birds from one-half hour before sunrise until 1000 hours or from 1800 hours until sunset if morning counts could not be made.
Presence/absence data were compiled for every sample plot. Percent frequency of occurrence (number of plots in which a species was detected divided by the total number of plots surveyed × 100) was calculated for each avian species within primary habitat types.
Thirty-two bird species were surveyed in the flooded woodland wetlands (Table 6). Species occurring in ≥50% of the plots included the yellow-headed blackbird, marsh wren, red-winged blackbird, American coot, tree swallow, and Franklin's gull.
|Table 6. Frequency of occurrence (%) of nongame breeding birds in flooded forested wetlands (PFOC) (n = 10) and emergent wetlands (PEMF) (n = 106) at Sand Lake National Wildlife Refuge National Wildlife Refuge, Brown County, South Dakota, 4 June - 4 July 1996.|
|Species||Wetland habitat class|
|Great blue heron||10||2|
|American white pelican||0||4|
Species surveyed in flooded woodland habitat but common to terrestrial woodlands included the tree swallow, eastern kingbird, mourning dove, and yellow warbler. Many of the species surveyed in flooded woodlands were "users." They did not nest in the habitat but used it for foraging. However, tree swallows and eastern kingbirds were found nesting in flooded trees.
Surveys in semipermanently flooded emergent habitats revealed 41 bird species (Table 6). Yellow-headed blackbirds occurred in 99.1% of the plots and yellow-headed blackbirds, marsh wrens, American coots, Franklin's gulls, and black terns occurred on ≥50% of the plots. Virginia rails and black-crowned night-herons occurred in 25% and 29% of the plots, respectively.
Playback recordings to elicit responses from secretive species such as rails and bitterns were effective. In forested wetlands, soras and least bitterns responded in 10% of the plots and Virginia rails responded in 20% of the plots. In emergent vegetation types, soras, Virginia rails, American bitterns, and least bitterns responded to calls in 17.0%, 24.5%, 16.0%, and 3.8% of the plots, respectively. The absence of American bittern responses in the flooded forested wetlands may not necessarily represent an avoidance of these habitats but rather may be an artifact of a low sample size (n = 10) for this habitat type.
A third of North American birds use wetland habitats (Gibbs et al. 1991), and of these, 75% are nongame birds. SLNWR is located along a major migration corridor for waterfowl and numerous species of nongame bird species in the prairie pothole region of North America (Schneider 1978, Sayler et al. 1988). SLNWR, with its interspersion of wetlands and uplands, represents a dynamic prairie marsh ecosystem (SLNWR narrative 1990).
SLNWR supports habitat for 61% of the birds (n = 46) listed as rare species by the South Dakota Natural Heritage Program. Among these, 21 species (28%) have nested on the refuge in wetland or associated habitats (Table 7). The first nesting of a common moorhen in South Dakota was reported in 1995 (Meeks and Higgins 1997).
|Table 7. Rare bird species monitored by the South Dakota Natural Heritage Program which have occurred at Sand Lake National Wildlife Refuge wetlands or associated habitats.|
|Horned grebe*||Long-eared owl|
|Clark's grebe*||Northern saw-whet owl|
|Least bittern*||Ruby-throated hummingbird|
|Great blue heron*||Olive-sided flycatcher|
|Great egret*||Brown creeper|
|Snowy egret*||Eastern bluebird*|
|Little blue heron*||Veery|
|Black-crowned night-heron*||Loggerhead shrike*|
|Yellow-crowned night-heron*||Yellow-throated vireo|
|White-faced ibis*||Black-and-white warbler|
|Hooded merganser||LeConte's sparrow*|
|Common merganser||Sharp-tailed sparrow*|
|Bald eagle**||Peregrine falcon|
|Sharp-shinned hawk*||Prairie falcon|
|Cooper's hawk*||Long-billed curlew|
|Northern goshawk||California gull|
|Broad-winged hawk||Least tern|
|Swainson's hawk*||Common tern*|
|Ferruginous hawk||Black tern*|
|Golden eagle||Burrowing owl|
|*species which have
nested on SLNWR in wetlands and associated habitats
**Bald eagles have attempted to nest on the refuge
Secretive species such as rails and bitterns were detected by using playback recordings. At SLNWR this group of species, except American bitterns, was found throughout all available habitat types, which included many deep-water cattail sites. Sora and Virginia rails were found in emergent stands of vegetation in water up to 1.0 m deep. This differs from Capen and Low (1980) who classified rails as shallow-water marsh dwellers. Johnson and Dinsmore (1986) reported that soras preferred shoreward sites whereas Virginia rails preferred deeper-water sites with robust emergent vegetation. They found no ecological niche segregation between the two species.
Deep water may require use of morphological or behavioral characteristics for ecological separation of birds (Reid 1993). For example, Weller (1961) reported that least bitterns grasped emergent vegetation or the vegetation in feeding platforms to facilitate their foraging in deep water sites.
Franklin's gulls occurred in 70% of the plots in flooded forested wetlands and in 72% of the plots in semipermanently flooded emergent wetland plots. These large frequencies can be attributed to the large gull colony on SLNWR. This colony of 100,000 to 150,000 pairs of Franklin's gulls is located in homogeneous stands of cattail, with the majority on the south side of Houghton Grade. The habitat supporting the Franklin's gull colony also supports white-faced ibises, Forster's terns, common terns, black terns, black-crowned night-herons, cattle egrets, great egrets, little blue herons, and snowy egrets.