Northern Prairie Wildlife Research Center
There is still a lack of information regarding habitat use and nongame bird responses to traditional management practices. More data on neotropical migrants are also needed, because many populations are declining on a local and national scale (Robbins et al. 1985).
Nongame birds were surveyed in terrestrial habitats between 15 May and 14 July 1995 using fixed-width belt transects (Mikol 1980, Wakeley 1987). Belt transects were 40 m wide (20 m on each side of the survey line) and varied in length depending on field size. If a small wetland or other obstacle was encountered while walking the transect line, sampling was resumed along the line on the opposite side of the obstacle.
Habitats surveyed included deciduous woodlands and planted shelterbelts (cottonwood and green ash), tame grasslands (dominated by smooth brome, and including intermediate wheatgrass [Agropyron inter-medium]), reseeded native grasslands (big and little bluestem), and alfalfa (Medicago sativa) (Table 1).
|Table 1. Field type and size (ha) of habitats surveyed for bird species at Sand Lake National Wildlife Refuge, Brown County, South Dakota.|
A total of 38 transects were randomly distributed among habitat types (Fig 3). Birds were counted twice within each belt transect during a field season, except in alfalfa plantings which were surveyed only once before cutting, after which they became unacceptable for a second survey. All transects occurred in homogeneous habitats, and the habitat edges were avoided (≥50 m, except in woodland shelterbelts) to minimize potential bias (Arnold and Higgins 1986).
|Figure 3. Location of nongame birds transects at Sand Lake National Wildlife Refuge, Brown County, South Dakota.|
All birds seen or heard within each belt transect were recorded by species, sex, and age (adult/juvenile) according to visual or audio cues. Bird counts were conducted from one-half hour before sunrise to 1000 hours. Counts were made by two observers walking at 1.0-1.5 km/hr (Mikol 1980), one surveying birds, the other recording data. Sampling did not take place during substantial precipitation, low temperatures, or excessive winds (≥20 km/hr). These can hamper an observer's ability to detect birds or may alter bird behavior (Mikol 1980).
Bird species names (Appendix A) and abbreviations followed AOU (1983) accepted standards.
Estimates of breeding pair densities (total number of perched and/or singing males divided by the belt transect area) and total bird densities (regardless of sex) were determined for each habitat type. Percent species composition (number of birds of a given species divided by the total number of birds of all species in an area × 100), percent frequency of occurrence (number of plots in which a species occurred divided by the total number of plots surveyed × 100), and species richness (total number of bird species present in a given habitat type) were calculated for each habitat type.
Breeding bird density (number of males per belt transect area) between habitats was tested using PROC GLM (General Linear Models Procedure) (SAS 1990) to determine normality of the data and to determine any relationships between breeding bird densities and habitat type. Analysis of variance (ANOVA) was conducted to test for significant differences, using habitat type as the error term. The same procedures were performed to test for differences between overall bird densities, regardless of sex and habitat type.
Breeding bird densities (males regardless of species)/100 ha) did not differ among habitat types (P = 0.8917), nor did overall bird densities (birds/100 ha) differ among habitat types (P = 0.6449). This was due in part to variation in size of fields of the same habitat type.
Bird species differed in diversity among habitat types. Species richness in woodland habitats was greater than in alfalfa, tame grassland, or native reseeded habitats (Fig 4).
|Figure 4. Species richness in four primary habitat types at Sand Lake National Wildlife Refuge, Brown County, South Dakota.|
Ten woodland habitats contained 35 species (628 total detections), with 76.6% of all detections made up of 13 species-house wrens, song sparrows, American robins, mourning doves, eastern kingbirds, brown-headed cowbirds, orchard orioles, yellow warblers, blue jays, western kingbirds, northern orioles, common yellowthroats, and downy woodpeckers. House wrens and song sparrows accounted for 37.6% of all detections in woodland habitats (Table 2).
Eighteen bird species were found in tame grassland habitats (n = 20 and 505 detections, Table 3). Bird species richness (n = 18) in tame grassland habitats was about half that in woodland habitats (n = 35). Sedge wrens, common yellowthroats, red-winged blackbirds, bobolinks, and clay-colored sparrows accounted for 81.0% of all detections in grasslands.
Twelve bird species were seen in native grassland habitats (n = 4, 111 observations, Table 4). Sedge wrens, common yellowthroats, song sparrows, red-winged blackbirds, and brown-headed cowbirds accounted for 67.6% of all bird detections in native grassland habitat. LeConte's sparrows made up 2.7% of the species composition.
Alfalfa habitats (n = 4, and only surveyed once) had the lowest overall species richness (n = 7, Table 5). Yellow-headed blackbirds, common yellowthroats, and red-winged blackbirds made up 76.2% of total detections.
The diverse terrestrial avifauna of SLNWR can be attributed to the variety of different habitat types. An interspersion of grasslands, both native and tame, and woodlands on SLNWR supports many generalist and edge species.
Woodlands are of considerable importance to birds and other wildlife (Yahner 1983), by possessing both horizontal area and vertical heterogeneity which permits co-occurrence of more bird species (Blake and Karr 1987). Woodlands and shrubby areas provide foods for many granivorous and insectivorous birds and areas for breeding, nesting, and loafing (Robel and Browning 1981).
Although they made up only 1% of the total land area on the SLNWR, woodlands supported the highest bird species richness. Faanes (1982) found a similar correlation of high species numbers in woodland habitats that were a small percentage of total land area at Sheyenne Lake in North Dakota. Vertical heterogeneity may increase habitat for many generalist species and may attract a higher number of species than might be expected in a small area.
Only two species, the house wren and the song sparrow, were very common in woodland habitats, accounting for 38% of total species abundance. The remaining 32 species were less common but occurred in similar frequencies.
Virtually all woodland habitats at SLNWR are artificial. Before refuge establishment and though in a riverine setting, the area had been mostly devoid of woodland vegetation. During refuge establishment, several thousand trees were planted in multi-row tree belts, and they were carefully maintained as windbreaks and wildlife cover (SLNWR narratives 1936-1939). The interspersion of trees among grassland habitats increased overall avian diversity on the refuge.
However, subdividing large contiguous habitats and increasing species diversity results in losses for habitat-size-dependent species (Temple et al. 1979). Grassland habitat fragmentation may have contributed to the elimination of the greater prairie chicken (Tympanuchus cupido pinnatus), which once was present on the refuge.
While the number of bird species in woodlands (or any habitat) is influenced by area, small habitats also may possess only edge species (Galli et al. 1976, Ambuel and Temple 1983). Smaller woodlands such as shelterbelts are dominated by ecological generalists that frequently travel to other nearby habitats to forage (Blake and Karr 1987).
SLNWR woodland habitats are all generally small, ranging from 1.1 to 7.6 ha in size. Excessive edge associated with these smaller woodlands may lead to smaller populations of species dependent on large blocks of habitat, and many birds detected in these woodlands may be area-independent or edge species. Habitat-size-independent species (Samson 1980) detected on the refuge included the European starling, the common grackle, and the American robin.
Although grassland habitats in the northern Great Plains are vital to many species of birds (Faanes 1982), nongame bird affinities to these habitats are poorly understood, in comparison to those of waterfowl and upland game birds (Kantrud and Higgins 1992).
This lack of understanding becomes even more critical as grassland habitat continues to be destroyed. During the last 25 years, grasslands have been lost to modern row crop and small grains agriculture (Graul 1980, Renken and Dinsmore 1987), resulting in steeper and more consistent population declines than experienced by other taxa (Knopf 1994).
Consequently, public grasslands have become islands of habitat for native avifauna. They must play a growing part in conservation of these avian communities (Graul 1980), and public land managers will need information on the habitat requirements of nongame bird species (Renken and Dinsmore 1987).
Nongame grassland birds, like woodland species, are greatly affected by habitat size and distance between habitat islands (Samson 1980, Anderson and Robbins 1981). Although grassland bird diversity is generally lower than woodland bird diversity, when large blocks of habitats are available, area-dependent bird communities will increase. Species diversity will increase only minimally, for it is achieved best by maximizing habitat diversity (Anderson and Robbins 1981). However, landscapes which retain natural habitats support higher relative abundances of nongame endemic birds than do agricultural and urban areas (Flather and Sauer 1996).
Minimum habitat size requirements for nongame breeding birds in riparian grasslands and forests are not completely known (Samson 1980). Bird species positively associated with increasing habitat area include the grasshopper sparrow, the bobolink, and the savannah sparrow (Herkert 1994). In native grassland areas on SLNWR, grasshopper and savannah sparrow percent composition was 3.6% and 0.9%, whereas in tame grasslands they were only 0.6% and 0.2% of the total bird community, respectively. Bobolinks were abundant on the refuge and, as reported by others (George et al. 1979, Ryan 1986), were often associated with tall residual cover which was abundant in most tame grassland habitats.
Bird species not dependent on increasing habitat size included the song sparrow, red-winged blackbird, and American goldfinch (Herkert 1994). Regardless of patch size, in native grasslands on the refuge, red-winged blackbirds made up 6.3% and song sparrows 11.7% of the bird community. In tame grasslands these two species were approximately 21% of the bird community.
Bird species that were habitat (not area) dependent included the sedge wren, the upland sandpiper, and the common yellowthroat. Sedge wrens were abundant in either tame or native grasslands. In tame grasslands ( = 12.3 ha, range 4.9 - 31.0 ha) they occurred as 28.9%, and in native grassland ( = 14.1 ha, range 5.4 - 21.8 ha) they were 31.5% of total detections.
Upland sandpipers (not surveyed during terrestrial surveys, but detected in 1996) and western meadowlarks (not seen in the plots) are categorized as area-dependent species (Samson 1980), but neither species occurred in any fields examined during this study.
SLNWR participates in a cooperative agricultural program in which cooperators plant soybeans, corn, small grains, and alfalfa. Cropland makes up 12% (1,045 ha) of the refuge uplands.
This program, in existence since early refuge days (SLNWR narratives 1936-39), currently provides alternative food sources that help decrease white-tailed deer (Odocoileus virginianus) depredation on adjacent landowner crops and provides food plots for other wildlife (Wm. Schultze, pers comm, USFWS, Columbia, S.D.).
Alfalfa fields with good legume cover may be attractive to dickcissels, savannah sparrows, and bobolinks (Ryan 1986). We found only seven bird species in refuge alfalfa fields, the lowest overall species richness of all habitats surveyed on SLNWR. Alfalfa habitats had about half as many species as grasslands and about a fifth as many as woodlands. Three bird species (yellow-headed blackbirds, common yellowthroats, and red-winged blackbirds) were responsible for 75% of the total bird sightings in alfalfa fields.