---

---

---

Introduction

Habitats of the Northern Great Plains are vital to many species of birds. Stewart (1975) related numerous early observations of prairie wildlife populations that included impressive assemblages of breeding waterfowl and an abundance of nesting passerines. These populations remained bountiful until the mid-1800's, when the first pioneers began to convert native grasslands to cropland and wetlands were drained to facilitate expanding farming operations.

Native wetland and upland habitats in the Northern Great Plains are still being altered or destroyed each year through agricultural practices. Destruction of natural basin wetlands in North Dakota was estimated at over 8,100 ha annually from 1967 to 1980 (U.S. Fish and Wildlife Service, unpublished data). Weller (1979) reviewed the impact of this type of habitat loss on waterfowl and other marsh birds in Iowa and found that populations of many marsh birds were reduced or extirpated by wetland drainage. Conversion of native grasslands to cropland also has an impact on breeding bird populations.

The Garrison Diversion Unit, a large public works project in North Dakota, under construction by the U.S. Water and Power Resources Service, will have substantial impacts on both upland and wetland habitats (International Garrison Diversion Study Board 1976). As planned, this project would supply agricultural irrigation water to nearly 101,214 ha of cropland in central and southeastern regions of North Dakota. Construction of the project, however, will destroy 30,539 ha of wetlands and 26,979 ha of upland native prairie. One strategic project feature of Garrison Diversion is the Lonetree Reservoir on the Sheyenne River in Sheridan and Wells counties (Fig. 1).

Lonetree Reservoir will serve as the principal water-regulation reservoir for the Garrison Diversion Unit (U.S. Bureau of Reclamation 1974). At full capacity, the reservoir will occupy 8,128 ha that are now wetlands, upland native prairie, woodlands, and cropland. About 40 km of the upper reaches of the Sheyenne River, and the Sheyenne National Wildlife Refuge, will be inundated.

The value of the study area to wildlife has long been recognized, but specific investigations to determine the importance of the area to migrant and nesting birds have not been conducted. Knowledge of numbers and species of birds using Sheyenne Lake and adjacent habitats is of importance to adequately determine impacts caused by construction of Lonetree Reservoir.

From April to June 1980, I conducted investigations of habitats and bird populations in the proposed site of the Lonetree Reservoir. The objectives of the study were to (1) determine the species composition of birds using Sheyenne Lake and adjacent habitats during spring migration and nesting periods, (2) determine the utilization of these habitats by migrant and nesting birds, and (3) describe the diversity of breeding birds in various habitat types.

Study Area

Lonetree Dam on the Sheyenne River, Wintering Dam on a tributary to the Wintering River, James River Dike, and seven smaller dikes will be constructed to impound water from the Missouri River in Lonetree Reservoir. Water in the reservoir will have a maximum depth of 24 m and will fluctuate up to 6 m during the annual operating season (U.S. Bureau of Reclamation 1974). The dam will be constructed in the valley of the Sheyenne River about 8 km southwest of Harvey, North Dakota. The reservoir will extend west from the dam for about 40 km.

The Sheyenne Lake area occurs at the extreme southeastern corner of the Northwestern Drift Plain physiographic region of North Dakota (Stewart and Kantrud 1972). A glacial meltwater channel that contains the Sheyenne River is the prominent topographic feature of the study area. Bedrock consists of Cretaceous-age Pierre Shale and Dakota Sandstone, overlain by undifferentiated Pleistocene glacial till (Simpson 1929). The valley has a thin mantle of alluvial soils and channel outwash deposits (U.S. Bureau of Reclamation 1974). Predominant soils in the region are silt loam and clay.

The climate is continental, with hot, dry summers and cold winters. Mean annual precipitation at Drake, North Dakota, 11 km north of the study area, is 55.8 cm and the average July temperature is 21.8°C (U.S. Department of Commerce 1979). The spring and summer of 1980 were dry and hot compared with normal precipitation and temperature (Table 1).

Habitats

The five major habitat types in the study area were wetlands, prairie thicket, upland native prairie, shelterbelt, and cropland. A general description of these habitats is provided in Stewart (1975). Cropland was the predominant habitat type, followed by wetlands, and upland native prairie (Table 2). A description of the vegetative characteristics of these habitats follows.

Wetlands

Wetlands made up over 27 % of the study area. Six major wetland classes (Stewart and Kantrud 1971) occurred in the study area: ephemeral (Class I), temporary (Class II), seasonal (Class III), semipermanent (Class IV), permanent (Class V), and alkali (Class VI). Because of the hot and dry spring of 1980, all ephemeral wetlands were dry, and variation in wetness of seasonal wetlands was considerable. Emergent vegetation in seasonal wetlands included tall mannagrass (scientific names of plants mentioned in the text are provided in Appendix A), slough grass, whitetop, slough sedge, western waterplantain, and marsh smartweed. Emergent vegetation of semipermanent wetlands was primarily common cattail, hardstem bulrush, slender bulrush, and river bulrush. Emergent vegetation of permanent wetlands occurred in narrow bands around the periphery of the wetland and in bays. Common cattail, river bulrush, and hardstem bulrush were the predominant emergent species in permanent wetlands. Submerged aquatic vegetation in semipermanent and permanent wetlands was characterized by claspingleaf pondweed, sago pondweed, coontail, and water milfoil. Little vegetation occurred in alkali wetlands. No emergent vegetation occurred on these wetlands in 1980 due to low water levels, and submerged vegetation was dominated by saltwater widgeongrass. The dry, shallow marsh zone was dominated by saltgrass and alkaligrass. Slightly brackish wet meadows occurred adjacent to brackish or saline basin wetlands, or intermittently along reaches of the Sheyenne River. Vegetation in wet meadows was regulated primarily by the salinity of groundwater. Important species at Sheyenne Lake were fowl mannagrass, phragmites, water sedge, northern reedgrass, prairie cordgrass, wooly sedge, slender sedge, and Baltic rush.

Prairie Thickets

Prairie thickets occurred to a limited extent, but I found them to be the most diverse habitat type in the study area, both in terms of vegetation and associated birds. Prairie thickets occupied the lower reaches of coulees and the soils were water-soaked. A combination of shrubs, forbs, grasses, and sedges contributed to the diversity of prairie thickets. Shrub-layer vegetation included hawthorn, wild plum, chokecherry, red raspberry, western rose, and wolfberry.

Native Mixed-grass Prairie

Vegetation typical of native mixed-grass prairie occupied about 24% of the study area. This habitat type at Sheyenne Lake conforms with the eastern mixed-grass prairie of Stewart (1975). Predominant species were blue grama, green needlegrass, needle-and-thread, prairie junegrass, and switchgrass. Numerous patches of wolfberry occurred on moist sites throughout the native prairie and dense stands of silverberry occurred on some tracts.

Shelterbelts

Shelterbelts were the least extensive of the five major habitat types in the study area. These man-established habitats were usually associated with agricultural fields, or were adjacent to farmsteads. The vegetative characteristics were variable, but box elder, common caragana, Russian olive, green ash, Rocky Mountain cedar, and Chinese elm were most prevalent. Two types of tree plantings were present on the study area: single-row windbreaks occurred in rows between fields, and multi-row shelterbelts, which were more numerous, were typically associated with farmsteads.

Cropland

Agricultural fields, producing a variety of crops, occupied 47% of the study area. Spring wheat, sunflowers, and alfalfa were the main agricultural crops.

Methods

Selection of Census Plots

Census plots (Fig. 2) were selected randomly from strata based on habitat types. I sampled a minimum of 10% of the area of each habitat except semipermanent wetlands and shelterbelts. Area of major habitats was determined from aerial photographs (scale: 1:15840), and then field-checked. I sampled upland habitats and wet meadows using 16.2-ha (0.4 x 0.4 km) census plots. Shelterbelts were sampled as they occurred on random plots, rather than selecting individual shelterbelts at random. Basin wetlands were sampled by selecting a minimum of 10 % of the area based on wetland class. During each census, the amount of water in each basin was visually estimated. Because ephemeral wetlands were dry throughout the study, censusing was not conducted in these wetlands.

I chose 16.2-ha plots for censusing upland habitats and wet meadows instead of the 64.7-ha plots used by Stewart and Kantrud (1972). Stewart and Kantrud selected the 64.7-ha plot size because two observers were involved. They believed that complete counts of birds could be made in a 2-h census period on those plots. Census plots of this size have considerable merit for investigating large, wideranging groups such as waterfowl or raptors but are probably too large for small passerines with weak voices (e.g., Sprague's pipit [scientific names of birds mentioned in the text are presented in the Annotated Species Accounts] and sharp-tailed sparrow). Another reason I chose smaller sample units was time. One observer was responsible for all upland and wet-meadow sampling and I chose to use the smaller plot size because it was more manageable. Also, by censusing smaller plots more replicates of the habitat type could be examined.

Census Methods

Birds were censused in upland habitats and wet meadows by following a zig-zag course within each plot. I followed the recommendation of Bond (1957), and walked slowly for 2-3 min, then stopped for 5 min to observe territorial behavior and to note if an individual bird was in or out of the census plot. All territorial males were counted as they flushed in front of me, or as their singing locations were encountered. The numbers of females were noted for brown-headed cowbirds and Wilson's phalaropes. The numbers of indicated pairs of sexually monomorphic species (black tern, barn swallow) were derived by halving the total of individuals counted on each plot.

Breeding birds were censused only when wind speeds were < 15 km/h, and only on sunny or partly cloudy days. Censusing in open habitats and wetlands was conducted during the period of sunrise to one-half hour before sunset. Wooded habitats, including prairie thickets, were censused only between sunrise and 1100. I limited by activities in wooded habitats to early hours because song output was reduced and movements were restricted during later hours (Robbins and Van Velzen 1967; Stewart and Kantrud 1972).

Populations of breeding birds were determined by the number of indicated pairs encountered during normal nesting periods. Determination of waterfowl populations and interpretations of pairs and lone birds followed the methods of Hammond (1969). Pairs and lone males were counted as indicated pairs. Densities of all species are expressed in pairs per square kilometer.

The census period extended from 17 April to 27 June. Basin wetlands were censused for early-nesting waterfowl 17 to 23 April, mid-nesting waterfowl and large shorebirds 20 to 23 May, and late-nesting waterfowl, other marsh birds and passerines, 23 to 27 June. Refer to Stewart and Kantrud (1973) for a listing of early-, mid-, and late-nesting waterfowl species.

Census methods in wetlands varied according to wetland class (Stewart and Kantrud 1974). In seasonal wetlands, counts were made by slowly walking through the wetland and making noise to flush birds. Semipermanent and permanent wetlands were often heavily vegetated and a different method was required. All ducks in the open area were counted when we (the census team) first arrived at the wetland. Once numbers of these had been determined, we moved slowly through the vegetation and recorded additional birds that were observed. Upon completion of a census, we compared observations to determine if any birds were counted twice. Alkali wetlands had little vegetation and we censused these from a high vantage point with the use of a spotting scope. We then moved slowly around the wetland to record additional birds that may have been missed.

Breeding birds in wet meadows and upland habitats were censused during 23 to 27 June. I sampled wet meadows only during June for two reasons. First, Stewart and Kantrud (1973) reported that <0.5% of the breeding waterfowl in North Dakota occupy wet meadows during the nesting season. Observations during April-May confirmed that few waterfowl were using wet meadows. Second, wet meadows usually occurred as broad expanses associated with intermittent portions of the Sheyenne River. Therefore, censusing breeding birds in wet meadows was more practical through the examination of 16.2-ha blocks.

In addition to intensive surveys of individual sample plots, I also conducted a 39.4-km Breeding Bird Survey (BBS) route (Robbins and Van Velzen 1967). Results obtained from this route provided another indication of relative abundance of breeding birds. Data from the BBS have been incorporated into the Species Accounts. I kept records of all breeding birds observed while moving from one census area to another, and made several visits to specific parts of the study area not sampled to search for additional birds. During all visits to the study area I investigated several sites to determine their use by migrants or breeding species. These data are presented in Annotated Species Accounts.

Data Analysis

Population estimates for breeding birds followed the methods described by Stewart and Kantrud (1972), including the calculation of highest probability density (HPD) intervals Johnson 1977). Bayesian confidence intervals were used because the number of breeding pairs is known to be non-negative, and because HPD intervals are more precise.

Species diversity indices, H' (Tramer 1969), and equitability, J' (Kricher 1972), were calculated for each habitat type. Species diversity is measured by the Shannon-Weaver formula

where P_i is the proportion of individuals of speciesⁱ. This index is influenced by the number of species and the numher of individuals within each species. Equitability is calculated by using

where H' = the species diversity index, and H' max is the maximum diversity possible in the sample. Equitability is highest when all species in the sample are as nearly equal in population as possible, and is smaller as single species become more predominant in the population.

Results and Discussion

Avian Use

During investigations of intensive but short duration, 175 bird species were recorded in the study area in 1980. The most recent compilation of North Dakota birds indicates that 353 species have been recorded in the State (C. A. Faanes, unpublished data). Thus, 49.6% of the total avifauna of the State was recorded in the Sheyenne Lake area during April-June 1980.

The avifauna included 68 species that occurred exclusively as migrants, 92 species that nested, 8 that occurred as summer residents, and 7 that occurred as summer visitors. Summer residents included birds for which the Sheyenne Lake area was within their known breeding range in the State, but no positive evidence of nesting was found. In all likelihood, these species nested in the study area. Summer visitors included species which nest elsewhere in the State, but occurred on the study area as feeding adults or as non-breeding vagrants. This group included white pelican, double-crested cormorant, and Franklin's, California, and ring-billed gulls.

Stewart (1975) listed 196 species that nest, have nested, or probably nest in North Dakota. The 92 breeding species recorded in my study area represented 46.9% of the known breeding avifauna of the State. Assuming that the eight summer residents nested, the total breeding component at Sheyenne Lake was 100 species, or 51% of the State total.

Breeding Populations

Seventy-five breeding bird species were recorded on the 83 census plots in the study area. Records were obtained of an additional 25 presumed breeding species. Records of active nests (those with at least one egg or young) or dependent young, were obtained for 55 species. Most of the 25 species not recorded on census plots were either wideranging birds which maintain very large territories (e.g., red- tailed and Swainson's hawk) or were rare, local, or highly secretive species that are difficult to census (e.g., piping plover, yellow rail).

Among taxonomic groups, four orders made up about 95 % of the total population. Passeriformes were the most numerous, contributing nearly 68% of the total, Within the Passeriformes, over 54% of the total breeding pairs were wetland or wetland-related species. Marsh nesting blackbirds were among the more numerous members of the Passeriformes. Anseriformes were second in numbers (16.1 %), followed by Columbiformes (6.3 %) and Charadriiformes (4.6%). Percentage composition of three of these orders was considerably different than the statewide mean. Stewart and Kantrud (1972) reported that Passedformes made up 79.5% of the State population, Anseriformes; made up 8.7%, and Columbiformes 2.4%. Only Charadriiformes were similar (3.9%). Among other orders, populations of breeding birds in the Sheyenne Lake study area were usually above the statewide means (Table 3).

The large number and the numerical importance of the Charadriiformes nesting in the study area were considerably higher than in surrounding areas. This group includes both shorebirds and terns; the status of many species in this order has been of continent-wide concern recently (see Arbib 1978, 1979). Among the species of greatest numerical importance was the black tern. A recent analysis of North American BBS data (D. H. Johnson and C. A. Faanes, unpublished data) indicates that the black tern is declining significantly at the periphery of its breeding range, and maintaining itself only marginally in its center--the prairie pothole region of the United States and south central Canada.

Based on data collected on census plots, the total breeding population was estimated at 21,798 pairs. Thirty-six species made up about 90% of the total breeding population in 1980 and represented 19,556 breeding pairs (Table 4). Ten species made up over 52% of the population.

The red-winged blackbird was the most abundant breeding bird, making up over 13 % of the total breeding population. Its ubiquitous nature is reflected in its occurrence in seven of the eight major habitats censused. Redwinged blackbirds were the most abundant breeding species in prairie thickets, seasonal wetlands, and wet meadows.

The second most abundant breeding species was the yellow-headed blackbird, which made up about 8.8 % of the total population. The yellow-headed blackbird was restricted exclusively to wetlands, reaching greatest density in semipermanent wetlands. The mean density in that wetland class was more than 78% greater than on the seasonal wetlands, which supported the next highest density. The mourning dove occurred in four major habitats; the largest densities were recorded in shelterbelts.

The species composition of breeding waterfowl in the study area was similar to that found throughout the Prairie Pothole Region. The projected population of waterfowl in 1980 was 3,501 pairs, which included 3,046 (87.0%) dabbling duck pairs, 446 (12.7%) diving duck pairs, and nine (0.3 %) Canada goose pairs (Table 5).

Breeding waterfowl populations at the Sheyenne Lake Study Area were considerably greater than the means for both the physiographic region and the State. I attribute these differences to the presence of high-quality semipermanent and seasonal wetlands and the climatic conditions in 1980. During April and May the proportion of seasonal wetlands containing surface water declined from about 35 to 18 %, but rebounded to about 30 % following June rains. Semipermanent and permanent wetlands contained water throughout the breeding season and were used extensively by waterfowl. In areas adjacent to Sheyenne Lake, water conditions were similar.

Although based on only 1 year of data, available records suggest that during dry years such as 1980, natural wetlands in the Sheyenne Lake Study Area may be serving as an important "refuge." The high-quality wetlands present in the study area may be serving to attract and maintain breeding pairs from the surrounding area during poor water years.

Habitat Distribution of Breeding Birds

Among the 36 major breeding species recorded, the breakdown of overall habitat affinities includes 18 species (50%) which are primarily wetland related, 6 (17%) which occurred primarily in shelterbelts, 5 (14%) were prairie species, 4 (11%) occurred primarily in prairie thickets, 2 (5%) were rather ubiquitous in habitat selection, and 1 (3%) was a cropland species.

Seasonal Wetlands

The second highest mean breeding bird density on wetlands occurred in seasonal basins, but these wetlands supported the lowest total population of all freshwater habitats. Red-winged and yellow-headed blackbirds were the two most abundant species in seasonal wetlands, making up about 46.6 % of the total population. Six species accounted for 65% of the species abundance: the two blackbirds, blue-winged teal, northern shoveler, pintail, and brown-headed cowbird. Waterfowl were represented by six species which made up about 24 % of the population.

Stewart and Kantrud (1973) found that the best environmental conditions for breeding dabbling ducks ling ducks occurred when large numbers of seasonal wetlands contained surface water. Because of drought conditions prevalent in 1980, seasonal wetlands were not sufficiently wet to attract most breeding bird species. During late May when most waterfowl were nesting, only 18% of the seasonal wetlands contained surface water. The proportion of waterfowl occupying seasonal and permanent wetlands (Table 6) was almost exactly the reverse of the pattern described by Stewart and Kantrud (1973).

Stoudt (1971) reported that when water conditions were optimal, most breeding waterfowl occupied wetlands that were not permanent. However, Jessen et al. (1964) found higher percentages of breeding waterfowl in more permanent wetlands regardless of wetland conditions in the prairie pothole region in Minnesota.

Semipermanent Wetlands

Semipermanent wetlands supported the second highest density of breeding birds and the third highest total breeding population (Table 7). The yellow-headed blackbird was the dominant breeding species, accounting for nearly one-third of the population. The yellow-headed and red- winged blackbirds made up about 48% of the breeding birds occupying semipermanent wetlands. Other species represented by 100 or more breeding pairs, in decreasing order of abundance, were blue-winged teal, gadwall, mallard, long-billed marsh wren, sora, common yellowthroat, and American coot.

Most investigations of breeding birds in wetlands have dealt with waterfowl and few data are available for non-anatid species. Weber (1978) found that highest breeding densities of both yellow-headed and red-winged blackbirds occurred in semipermanent and permanent wetlands. He reported that for both species, highest breeding densities were positively correlated with area of surface water and type of vegetative cover. Thus, large water bodies with an interspersion of emergent aquatic plant species were heavily used by both species. Weller and Spatcher (1965) found that the greatest density and diversity of marsh birds were associated with seasonal and semipermanent wetlands which provided a variety of vegetative life forms.

The total density of breeding birds occupying semipermanent wetlands in my study area was 1,358 pairs/km.² Knodel (1979) censused four semipermanent and permanent fresh wetlands in McLean County, North Dakota, and found a maximum density of 390 pairs/km²; on one wetland, the yellow-headed blackbird made up 47.5 % of the total density. Krapu and Green (1978) found 13 species on one semipermanent wetland; maximum density was 2,168 pairs/km.² Of the four wetlands they censused, the yellow-headed blackbird was the most abundant species in each.

Permanent Wetlands

Permanent wetlands occupied the largest area among freshwater basin wetlands, and had the highest number of species, but supported the lowest density of breeding birds in fresh wetlands. The blue-winged teal was most abundant, accounting for 13.7% of the population. Eleven species made up 75 % of the breeding population in permanent wetlands. These species, in decreasing order were blue-winged teal, red-winged blackbird, yellow-headed blackbird, mallard, black tern, gadwall, longbilled marsh wren, killdeer, eared grebe, pintail, and redhead. Waterfowl were represented by 13 species and made up about 42% of the population. Permanent wetlands were also occupied extensively by colonial and semicolonial water birds, particularly western, horned, and eared grebes, black-crowned night heron, and black tern.

Avian use of permanent wetlands on the study area, particularly by waterfowl, was considerably different from previous observations. Dwyer et al. (1979) reported that breeding female mallards used permanent wetlands the least of the five main wetland classes available. Kantrud and Stewart (1977) found that permanent wetlands were used by all major breeding waterfowl species in North Dakota, but that only the lesser scaup reached its maximum breeding density on these wetlands. H. A. Kantrud (unpublished data) recorded 13 of 22 non-anatid breeding bird species on permanent wetlands, but only the western grebe attained its maximum density on this wetland class.

Available data suggest that permanent wetlands showed a high degree of use by breeding birds but their importance for avian production, particularly by waterfowl, can be argued. Because of the relatively constant water depth and low amount of nutrient cycling on permanent wetlands, food productivity is limited. G. A. Swanson (personal communication) has noted that invertebrate productivity and use by breeding birds in permanent wetlands occur only when mudflats are exposed or water levels are reduced during periods of extreme drought. Krapu and Duebbert (1974) reported that periodic drawdowns of surface water and periodic drying periods were important in contributing to the high degree of productivity in a large marsh in southeastern North Dakota. These conditions do not regularly occur on permanent wetlands; therefore, the high densities and diversity of breeding birds I observed are probably masking the actual lesser importance of these large wetlands to avian production.

Alkali Wetlands

Alkali wetlands supported the lowest total breeding population of all habitats (Table 7). Species richness was second lowest; only cropland supported fewer species. The gadwall was the most abundant breeding species, contributing about 24 % of the total. Six species made up 76 % of the breeding population on alkali wetlands. These species, in decreasing order of abundance, were gadwall, northern shoveler, pintail, Wilson's phalarope, mallard, and blue-winged teal. Waterfowl were represented by 10 species and made up of about 69.9% of total abundance. Shorebirds were the second most numerous group of breeding birds, accounting for 27% of the population. Most abundant among these were Wilson's phalarope, American avocet, killdeer, and spotted sandpiper.

I found nine species that reached their maximum breeding densities in the Sheyenne Lake area on alkali wetlands. The American avocet and spotted sandpiper were restricted exclusively to these wetlands. H. A. Kantrud (unpublished data) found nine non-anatid species occupying alkali wetlands throughout North Dakota; the American avocet and piping plover reached their maximum density on these wetlands. Although of lesser importance to most breeding birds, alkali lakes were extremely important as feeding and resting areas for both migrant shorebirds and for waterfowl that nested on nearby wetlands.

Wet Meadows

Breeding bird populations in wet meadows were the second lowest among all wetland classes (Table 7). The red-winged blackbird was the most abundant species, making up 31.9 % of the total population. Five species made up about 75% of the species abundance in wet meadows; in order of decreasing abundance, they were the red-winged blackbird, savannah sparrow, bobolink, yellow-headed blackbird, and brown-headed cowbird. Blackbirds and sparrows were the most numerous groups of breeding birds; each was represented by five species. About 62.5 % of total abundance was attributed to blackbirds and 29.6 % to sparrows. Le Conte's sparrow was restricted to this habitat and seven other species reached their highest density in wet meadows.

Because of its relative unimportance to breeding waterfowl, wet meadow habitat has received very little attention in the Northern Great Plains. However, its use by breeding songbirds in the Sheyenne Lake area suggests that additional study should be placed on this wetland class. I found 19 species occupying wet meadows in 1980, and 8 of these reached their maximum densities in these wetlands. H. A. Kantrud (unpublished data) found that of 22 selected non-anatid species, 4 (killdeer, marbled godwit, red-winged blackbird, savannah sparrow) reached their maximum densities in these wetlands. Stewart (1975) listed 11 species (red-winged blackbird, savannah sparrow, marsh hawk, short-eared owl, Wilson's phalarope, short-billed marsh wren, Sprague's pipit, common yellowthroat, bobolink, western meadowlark, and Le Conte's sparrow) as primary and secondary indicators of wet meadow habitat.

Use of these wetlands by Le Conte's sparrow is well known (Southern 1962; Murray 1969; Faanes 1981). The density of Le Conte's sparrow that I encountered is probably among the greatest known for this species. The mean density in wet meadows throughout the study area was 15 pairs/km², and on one plot I recorded 43 pairs/km.² Bobolink and savannah sparrow occupied other habitats, primarily the drier portions of seasonal wetlands, and low, wet areas in upland native prairie. The density of bobolinks in wet meadows was 19 % greater than in seasonal wetlands and 98 % greater than in native prairie. Savannah sparrow densities in wet meadows were 95% greater than in seasonal wetlands and 94% greater than in native prairie.

The extensive use of wet meadows by "typical" upland native prairie species such as Baird's sparrow and Sprague's pipit was related to drought conditions. Stewart (1975) listed Baird's sparrow as a primary indicator of the eastern mixed-grass prairie in North Dakota and Sprague's pipit as a secondary indicator. In 1980, the density of Baird's sparrow in wet meadows was about 55% greater than in upland native prairie. Baird's sparrow is known to shift habitat use from native prairie to wetland habitats during dry periods (Kantrud and Faanes 1979), and Sprague's pipit is suspected of doing so. I believe that both species shifted habitat use to wet meadows in 1980 because of extremely dry conditions in the native upland habitat. R. E. Stewart (personal communication) reported that both Sprague's pipit and sharp-tailed sparrow reach maximum breeding density in North Dakota during drought periods when their primary habitat use is wet meadows.

Prairie Thickets

Prairie thickets were rather limited, but supported the third highest mean breeding density of all habitats (Table 7). The red-winged blackbird was the most abundant species, accounting for 22.9% of total abundance. Seven species made up about 75% of the total population: in order of decreasing abundance, the red-winged blackbird, yellow warbler, clay-colored sparrow, common yellow-throat, willow flycatcher, song sparrow, and brown-headed cowbird. Blackbirds were the most numerous family, contributing 31.7% of the total abundance, and warblers were second at 21.1 %. Six species reached their maximum density in prairie thickets, and two (American kestrel, willow flycatcher) were recorded only in this habitat.

Prairie thickets are largely restricted to grassland areas in the Prairie Pothole and in the Southwestern Slope regions of North Dakota (Stewart 1975). Those which are best developed in central North Dakota occur primarily along the eastern edge of the Missouri Coteau in isolated coulees that frequently serve as the watershed for intermittent streams. Because of the geologic origin of soils underlying prairie thickets, resultant vegetative growth can take characteristics of wet meadow, native prairie, and woodland communities. The predominant vegetative feature is the presence of a well-developed stand of various shrubs or small trees.

The density of breeding birds, species richness, and species diversity in prairie thickets were probably related to the presence of several microhabitats. Because the equitability component is highest when all species in the sample are as nearly equal in the population as possible (Kricher 1972), the equitability statistic obtained would indicate that no individual species was particularly dominant in prairie thickets. I found that nearly 75% of the population was accounted for by 37 % of the species. An additional seven species made up the next 22% of the total. In most other major habitats, a smaller proportion of the species accounted for a larger proportion of the population. With the clumped arrangement of similar seral stages, nesting birds that are primarily components of each microhabitat in the prairie thicket are probably able to reach maximum densities without excluding other species, as may occur in simpler habitats such as grasslands. Therefore, this component of evenness in the prairie thicket is operating to allow the maximum number of species to use prairie thickets at the expense of very few.

Native Prairie

Native prairie supported the second highest breeding density of birds in upland habitats, but the second lowest density of all habitats combined (Table 7). About 12.8% of all breeding birds in the study area occupied native prairie. The chestnut-collared longspur was the most abundant species, contributing 21.7 % of total abundance. Six species made up about 78% of the population including the longspur, grasshopper sparrow, clay-colored sparrow, western meadowlark, lark bunting, and horned lark. Sparrows were the most numerous family, making up 68 % of total abundance. The Icterids were second in abundance; four species contributed 20.4 % of the total population. Five species reached maximum density in native prairie, and three (sharp-tailed grouse, chestnut-collared longspur, lark bunting) were restricted exclusively to this habitat.

Species diversity in native prairie (Table 7) was among the lowest in the study area, and followed the general trend described by Tramer (1969). The diversity value was related to the rather simple physiognomic form of prairie grasslands, i.e., the lack of vegetative "layers" as is evident in structurally more complex habitats such as woodlands. The equitability index was in the middle of the nine major habitats, and probably reflected the rather uniform breeding populations of most prairie nesting species in the area.

Stewart (1975) listed 11 breeding bird species which he considered primary indicators of the eastern mixed-grass prairie. Of these, 10 were recorded on native prairie census plots and made up nearly 80 % of total abundance. Native grasslands in the study area maintained an interspersion of species that are indicative of several grassland types. Species such as bobolink and savannah sparrow are typical primarily of eastern tallgrass prairie. Baird's sparrow, Sprague's pipit, and clay-colored sparrow are indicators of the central mixed-grass prairie, and the horned lark and chestnut-collared longspur are indicative of drier western shortgrass prairie. The predominance of the chestnut-collared longspur on the study area is similar to general conditions on native grasslands across most of the Northern Great Plains. Stewart and Kantrud (1972) found that the chestnut-collared longspur was the second most abundant breeding bird in North Dakota and made up about 9.5 % of the State's breeding avifauna.

Breeding birds occupying native prairie included species that reached maximum densities under different grazing intensities and moisture conditions. Among the six most numerous species occupying native prairie, two (chestnut-collared longspur, horned lark) reached maximum densities in moderately to heavily grazed prairie, two (grasshopper sparrow, lark bunting) reached maximum densities under moderate to light grazing conditions, the western meadowlark occurred on grasslands under several grazing conditions, and the clay-colored sparrow inhabited primarily clumps of low shrubs, usually wolfberry.

The drought that prevailed in 1980 appeared to affect the annual growth of prairie grasses. Vegetation on many grassland areas that had previously been overgrazed grew very slowly during the growing season. Several species which are considered characteristic of vegetation that is lightly to moderately grazed (e.g., Baird's sparrow, Sprague's pipit), contributed very little to total abundance. Those which are more tolerant of shorter herbage characteristic of heavily grazed areas, reached high levels.

Shelterbelts

The highest density of breeding birds and the largest population occurred in shelterbelts. Twenty-four species were recorded on the breeding bird census plots in shelterbelts. Eight species made up 74.6% of the population: mourning dove, western kingbird, eastern kingbird, brown thrasher, common grackle, brown-headed cowbird, American goldfinch, and least flycatcher. The mourning dove, family Columbidae, was the most abundant species, accounting for 27.1 % of total abundance. Tyrarmidae (three species) was the second most abundant family, and accounted for 24.6 % of the population. Icteridae (five species) was third in abundance with 19.0 %. Eighteen species reached maximum densities in shelterbelts. Seven species (common flicker, house wren, loggerhead shrike, warbling vireo, house sparrow, orchard oriole, northern oriole) were recorded only in this habitat type.

Total breeding density in shelterbelts on the study area was 2,659 pairs/km.² Yahner (1980a, 1980b, 1980c, 1980d) reported slightly higher densities in four mature Minnesota shelterbelts, ranging from 3,437 to 9,793 pairs/km.² Common grackles and mourning doves were the most abundant breeding birds in this area; largest densities were 8,412 and 709 pairs/km² Ressell (1973a, 1973b, 1973c, 1973d) reported slightly lower densities ranging from 699 to 2,136 pairs/km² in mature eastern North Dakota shelterbelts; mourning dove, yellow warbler, and American robin were the three most abundant species. Species richness in Ressell's study areas ranged from 4 to 14 species. H. A. Kantrud (unpublished data) found that breeding densities in several central North Dakota shelterbelts and singlerow windbreaks averaged 803 pairs/km.² Species composition in Kantrud's area included 33 species.

The number of breeding bird species occupying shelterbelts was the largest of all upland habitats. Species diversity and equitability in shelterbelts were among the largest of all major habitats. The diversity index is closely related to species richness, but equitability is regulated by the numbers of individuals of one or more species that are the dominant components of the habitat. Although a large number of species occupied shelterbelts, only two species accounted for nearly 40 % of the total density. This dominance was reflected in the high equitability index.

Cropland

Agricultural fields in various types of crop production made up over 47 % of the study area, but only 5.9 % of the breeding bird population occupied croplands. There were 13 species recorded in this habitat during the nesting season, and 10 of these probably nested there. The horned lark, which nested in wheat stubble fields and fallow fields, was the most abundant breeding species, contributing 51.7 % of total abundance. Three species made up nearly 78 % of the breeding population in cropland. Those in second and third place were red-winged blackbird, which nested in alfalfa fields, and vesper sparrow, which occupied the edges of various fields. Three species reached their maximum density in cropland, and the gray partridge occurred only in this habitat type.

Breeding bird populations in cropland were the lowest of all major habitats studied in 1980. Wheat, sunflower, and fallow fields constituted most of the land use. Little alfalfa was grown in the study area, and only one alfalfa field was included in the random selection of census plots. Of all crop fields investigated, bird populations were highest in alfalfa. The breeding density in croplands in the study area was 34 pairs/km.² Carlisle (1975) reported that five species with a density of 316 pairs/km² occupied a Dunn County, North Dakota, alfalfa field. In Illinois, Graber and Graber (1963) reported that mean breeding bird densities in alfalfa ranged from 543 to 1,153 birds/km² in 1957. Red-winged blackbirds were the most numerous breeding bird in Illinois alfalfa fields, making up almost 36% of total density.

The depauperate nature of breeding bird populations in cropland is demonstrated by the low species richness, mean densities, species diversity, and the equitability index that I observed (Table 7). In all instances the figures for cropland were the lowest in the study area.

Conclusion

The habitats proposed for inundation by the Lonetree Reservoir are among the most unique and diverse in North Dakota. Densities of many breeding bird species in this area equal or exceed statewide means. This finding, I believe, was related to the close interspersion of habitat types, the floristic composition, and the area's geographic setting. Because of its location, the breeding avifauna demonstrates a mixture of eastern, central, and western prairie species. About 24 % of the study area was upland native prairie, which supported about 13 % of the breeding birds. General use of lands surrounding the study area was predominantly for cereal grain or sunflower production. Thus, the existing grasslands appeared to be islands of native vegetation surrounded by altered habitats. Such diverse wetland-prairie complexes are rapidly disappearing from the landscape of North Dakota as agricultural production intensifies.

Breeding waterfowl populations generally exceeded statewide means in response to the numerous high-quality wetlands of several classes. If overall wetland conditions had been more favorable and seasonal wetlands had held water longer, breeding densities of dabbling ducks would probably have been higher. Nonetheless, the quality and density of wetlands was apparently sufficient during a drought to attract and maintain an exceptionally high density of waterfowl.

The Sheyenne River is an international resource because it is a tributary of the Red River of the North which flows into Canada. The study area was at the headwaters of the river, one of the largest and most important prairie streams in the State. Extensive areas of wet meadows have formed in the seepage areas associated with the River. In North Dakota, wet meadows form as isolated zones on the periphery of larger wetlands. Large expansive wet meadows, such as those at Sheyenne Lake, are rare and localized across the State. Breeding birds associated with wet meadows, particularly yellow rail, Le Conte's sparrow, and sharp-tailed sparrow, form a unique component of the State's avifauna.

The alkali lakes in the study area were not used extensively by breeding birds. The value of these wetlands was best demonstrated from their use by migrants. During late April, flocks of 30-50 whistling swans were frequently observed feeding and resting on alkali wetlands; their use of other wetlands was virtually non-existent. During late May, 29 shorebird species were observed in the study area. Of these, 17 species either reached maximum numbers or were restricted exclusively to alkali wetlands. Over 3,000 shorebirds were observed on one alkali wetland on 22 May, of which over 2,200 were northern phalaropes. Large mixed-species flocks of waterfowl were frequently observed feeding on alkali wetlands throughout May and June.

Among the other unique habitats of the area are the prairie thickets which formed along intermittent streams at the base of the Missouri Coteau escarpment. Breeding bird diversity and densities in prairie thickets were among the highest of all natural habitats studied. I found Lonetree Lake and the extensive marshes on the upper Sheyenne River to be very important for nesting and brood rearing among most marsh birds; the areas were also used extensively by molting waterfowl.

The single most diverse area investigated was a 31.6-ha semipermanent wetland in Secs. 17 and 20, T. 149 N., R. 75 W. The avifaunal diversity and abundance rivals the number of bird species using wetland vegetation at Kraft Slough in Sargent County, North Dakota (Krapu and Duebbert 1974). I recorded 29 species on this wetland during the breeding season, which represented nearly one-third of the total breeding avifauna of the study area. Present on the wetland were large numbers of breeding black and Forster's terns; both species are declining over most of their range. Numerous colonial water birds bred on the wetland or in its surrounding vegetation. Most important among these were four grebe species and black-crowned night heron. The wetland was used extensively by waterfowl and American coots. Over 200 male canvasbacks were observed there on 25 June, as well as numerous broods of most duck species.

---

---

NPWRC Home | Site Map | About Us | Staff | Search | Contact | Web Help | Copyright