USGS - science for a changing world

Northern Prairie Wildlife Research Center

  Home About NPWRC Our Science Staff Employment Contacts Common Questions About the Site

Factors Associated with Duck Nest Success
in the Prairie Pothole Region of Canada

Factors Related to Duck Production

Habitat Composition and Use

Habitat composition of the PPR has changed dramatically in the past 100 years. Much of the Region is now less suitable for duck production than it once was. Fewer wetlands hold water because of drainage, and large expanses of upland are tilled annually in both the parkland and prairie of Canada (Bird 1961, Lynch et al. 1963, Kiel et al. 1972, Archibold and Wilson 1980, Sugden and Beyersbergen 1984). Lynch (1984) suggested that duck production in the PPR is a boom-or-bust phenomenon, with variability greatest in prairie where dramatic annual changes in weather affect habitat conditions. Lynch et al. (1963) suggested that conditions leading to high duck production might have occurred in 4 or 5 years out of 10 under pristine conditions but in only 2 or 3 years out of 10 where habitats have now been altered by agriculture.

Large portions of all areas we studied were extensively altered through cultivation, and some areas had many drained wetlands. Most dabbling duck nests we found in extensively cultivated areas were in the scattered patchwork of vegetation that remained along roads, fences, and around wetlands. Some nests were in wetlands and hayfields, but very few were in grain stubble, except those of northern pintails. Conversely, large pastures where agricultural activities were limited to grazing were especially important to dabbling ducks. Eight study areas had large (>2.6 km²) pastures that were contiguous with additional pastureland adjacent to the study area; the largest contiguous area of pasture (92 km²) was associated with the Ceylon Study Area. Nest success on individual study areas was positively correlated with amount of pasture available on the study area (Greenwood et al. 1987).

Ducks used habitats for nesting similarly to previous reports. Many northern pintails and some mallards nested in stubble fields (Milonski 1958, Higgins 1977, Cowardin et al. 1985). Our estimates of nest initiations in Cropland were based on searches of untilled stubble and thus may be low. At more southern latitudes, growing grain provides some suitable habitat for nesting ducks (Higgins 1977, Duebbert and Kantrud 1987). At the latitude of our study, however, spring-seeded grain provided little cover for nesting until mid- to late June, and we had no fields of fall-seeded grain that Duebbert and Kantrud (1987) found to be important to dabbling ducks. Although change in tillage practices might attract more ducks to nest in cropland and reduce mechanical destruction of nests (Cowan 1982, Rodgers 1983, Duebbert and Kantrud 1987), such changes may do little to reduce predation, which we found to be the primary cause of nest failure.

Use of native grassland by nesting ducks and the importance of Brush have been well documented in the PPR (e.g., Keith 1961, Salyer 1962, Smith 1971, Stoudt 1971, Cowardin et al. 1985, Duebbert et al. 1986, Sugden and Beyersbergen 1987). Smith (1971) suggested that brush was used more for nesting in dry years (when grass was scant) than in wet years. However, we found several species used brush during both wet and dry years; Cowardin et al. (1985) observed the same with radio-equipped mallards in North Dakota.

Use of dry wetlands for nesting that we observed by blue-winged teals and northern shovelers has been reported earlier (Bellrose 1980, Klett et al. 1988), but, as we found, mallards were the only dabbling ducks to commonly nest over water (Jessen et al. 1964, Krapu et al. 1979, Bellrose 1980, Arnold et al. 1993). In 1984, the driest year of our study, 30% of all nests of the 5 common species were in Wetland compared to 3% in 1982, 13% in 1983, and 12% in 1985. Extent of nesting in dry wetlands in 1984 may reflect diminished attractiveness of upland vegetation during drought, or the attractiveness of often rank cover in dry wetland basins.

Odd area and Right-of-way contained 45-86% of the nests of the 5 common species in parkland and 18-50% in prairie. High proportions of nests in these habitats, especially in parkland, probably reflect the shortage of more preferred nesting habitats in areas where Cropland was abundant. Klett et al. (1988) reported that Odd area and Right-of-way contained 6-25% of the nests of the 5 common species during 1980-84 in North Dakota, South Dakota, and Minnesota.

We found few nests in Hayland, although alfalfa is frequently used by nesting ducks, especially mallards and blue-winged teals (Salyer 1962, Ordal 1964, Evans and Wolfe 1967, Cowardin et al. 1985). Many of our study areas contained little Hayland, especially alfalfa, and drought may have reduced its attractiveness. Cowardin et al. (1985) found alfalfa to be most used by nesting mallards when it was relatively tall.

Scant information on nesting in Woodland is available to compare to our findings. However, a recent study in Alberta showed that radio-equipped mallards regularly nest in Woodland (D. W. Howerter, Ducks Unlimited Canada, Winnipeg, Manit., pers. common.).

Proportion of Land in Cultivation

An average of 59% of the land on our study areas was Cropland, but elsewhere in the Canadian PPR the proportion of land tilled annually may be considerably higher. Lynch et al. (1963) estimated that by the mid-1950's, 72% of the grassland in Canada was farmed for grain production. Sugden and Beyersbergen (1984) reported that in 1982, 78% of the parkland in east-central Saskatchewan was tilled annually. Areas of upland and wetland continue to be converted to cropland, especially small woodlots and pastures, margins around wetlands, and road rights-of-way (Adams and Gentle 1978, Sugden and Beyersbergen 1984, Boyd 1985, Turner et al. 1987). Boyd (1985) suggested that this trend toward farming "marginal" land (much of which we classified as Odd area) poses a greater threat to ducks than changes on the best agricultural land, which is already under cultivation.

We found that nest success on study areas in prairie was negatively correlated with proportion of land cultivated annually and decreased about 4 percentage points for every 10 percentage points increase in amount of Cropland. Because few ducks (except northern pintails) nested in Cropland, nest success in other habitats apparently decreased as amount of Cropland increased. The higher preference of northern pintails for nesting in Cropland may render that species more sensitive than others to amount of Cropland that is present.

We believe that nests are at higher risk to destruction in small blocks of upland nesting habitat than are nests in large contiguous blocks that have not been fragmented by cultivation. This is supported by our finding that nest success estimates in Odd area and Right-of-way, both common habitats of fragmented landscapes, were among the lowest estimates we obtained for all habitats, except Cropland. Higgins (1977) suggested that upland-nesting ducks cannot sustain breeding populations where >=85% of the land is cultivated annually. In our study, the threshold appeared to be considerably lower. We suggest, however, that caution be exercised in directly applying our predicted values elsewhere because of the conditions of drought under which our study was conducted.


Weather events, especially spring snowstorms and flooding, had direct negative effects on nest success on some study areas. However, the more general influence of weather on duck production probably occurred indirectly through effects on habitat conditions, particularly wetland habitat. Local wetlands provide most nutrients needed by females for egg production, especially for renesting (Krapu 1974, 1979, 1981; Swanson et al. 1979; Krapu et al. 1983). Renesting is particularly important to annual production of ducks when success of initial nesting attempts is low (Gates 1962, Pospahala et al. 1974, Swanson et al. 1979).

We found that initiation date and length of nest-initiation period were related to temperature, availability of wet wetlands in May, and amount of precipitation during nesting season, but it is unclear which factors had greatest influence. Below average temperatures in early spring may delay nest initiation (Sowls 1955:85, Smith 1968, Hammond and Johnson 1984, Cowardin et al. 1985), and above-average temperatures later in the nesting season may cause ducks to terminate nest initiation (Dzubin and Gollop 1972). Abundant wet wetlands and ample precipitation both are likely to enhance the potential for renesting (Krapu 1979, Swanson et al. 1979, Hammond and Johnson 1984, Cowardin et al. 1985, Swanson et al. 1985). Crissey (1969) found a positive relationship between annual production of mallards on a continental basis and number of wet wetlands present in July in the PPR of Canada; that finding suggests that production is highest when conditions for renesting and brood survival are good. Krapu et al. (1983) and Jackson et al. (1985) noted reduced mallard nesting activity and duration during drought. Krapu et al. (1983) observed that females abandoned breeding sites and had smaller home ranges and that a high percentage of radio-marked females did not nest during drought. Duncan (1987), who studied northern pintails in southern Alberta in 1984, found virtually no renesting during that drought year.

We found that predation rates tended to decrease as the nesting season progressed, similar to Sugden and Beyersbergen (1986), who found less predation on artificial nests in late June and July than earlier in the season. Although we had no direct evidence of cause for this, we believe that seasonal changes in availability of buffer prey and in habitat conditions both may have affected nest success. Abundance of buffer prey is known to affect predation rates on birds and eggs in other ecosystems (Larson 1960, McInvaille and Keith 1974, Pehrsson 1986, Summers 1986, Beintema and Müskens 1987). As upland nesting cover becomes more dense with seasonal growth, better concealed and more dispersed nests also may reduce the foraging efficiency of some predators. Although higher nest success has been reported in dense cover (Kirsch et al. 1978; Livezey 1981; Cowardin et al. 1985; Sugden and Beyersbergen 1986,1987), a recent review by Clark and Nudds (1991) did not clearly confirm or refute the reported benefits of dense nesting cover. Clark and Nudds (1991) suggested that amount of concealment of nests was important to survival of duck nests in the PPR only when predation was predominantly by birds.

Our findings suggest that currently there is an advantage to ducks that nest later in the season because of higher nest success then; females that initiate nesting later also may be less vulnerable to predation (see Mortality of Adult Ducks). To benefit from higher hatch rates later in the nesting season, early-nesting species such as the mallard and northern pintail must delay initial nesting or be able to renest when initial nests are destroyed, whereas later-nesting species such as the gadwall, blue-winged teal, and northern shoveler may benefit more regularly. Selection might favor mallards and northern pintails that successfully nest for the first time later in the season. This selective force would be even stronger if mortality of nesting females from predation was reduced by delayed nesting (Rohwer 1992). There also are disadvantages to nesting late, however, including declines in clutch size, renesting potential, and brood survival rates that must be balanced against the apparent advantages (Rohwer 1992).

Unless success of initial nests is high, the cumulative effects of drought appear to result in an overall reduction in the average nest success rate and, ultimately, in lower hen success. Renesting, which is important to hen success, depends on quality of wetland habitat (Krapu 1979; Swanson et al. 1979, 1985; Cowardin et al. 1985). Annual estimates of nest success are derived from total nests observed throughout the nesting season. If the nesting season is truncated because of high temperatures, reduced precipitation, or deteriorating wetland conditions, the potential for renesting will be reduced. Thus during drought, contributions to annual estimates from nests initiated in the later portion of the nesting period are lost; it is then that success tends to be highest. Drought likely increased the severity of our low estimates of nest success in some area-years through effects on renesting.

Predator Community

Johnson et al. (1989) used data from our study to examine the relation of individual egg-eating predator species to nest success. They found that daily predation rates on nests of the 5 common species were positively related to activity indices of 6 mammals (coyote, red fox, raccoon, striped skunk, badger, and Franklin's ground squirrel) and 2 birds (American crow and black-billed magpie). They concluded that the red fox was the most influential predator on duck nests on our study areas. Activity indices of the red fox were positively related (P < 0.01) to predation of both early nests (initiated during first or second nest-search periods) and late nests (initiated during third nest-search period). Activity indices of American crows (P < 0.01) and badgers (P < 0.10) were positively related to predation on early, but not late nests. Activity indices of striped skunks were positively related (P < 0.10) to predation on late nests only. Neither Franklin's ground squirrels nor black-billed magpies were strongly implicated in predation on our duck nests, but Johnson et al. (1989) suggested they were locally important.

Johnson et al. (1989) also detected a strong negative relation between red fox and coyote activity indices (r = -0.51, P < 0.01), consistent with the spatial avoidance and agnostic behavior between these 2 species (Voigt and Earle 1983, Sargeant et al. 1987). A similar avoidance relationship may have influenced the distribution of coyotes and raccoons (r = -0.11, P < 0.01; Johnson et al. 1989) and red-tailed hawks and American crows (Sargeant et al. 1993). Such interactions, which appear to be only partly related to habitat, greatly complicate evaluation of relations between duck nest success and habitat because they tend to obscure habitat effects. Sovada et al. (1995) found in areas of similar habitat, for instance, that nest success of dabbling ducks averaged 15 percentage points higher in areas occupied by coyotes than in areas occupied by red foxes.

Previous Section -- Discussion
Return to Contents
Next Section -- Causes of Nest Failure

Accessibility FOIA Privacy Policies and Notices

Take Pride in America logo logo U.S. Department of the Interior | U.S. Geological Survey
Page Contact Information: Webmaster
Page Last Modified: Friday, 01-Feb-2013 19:39:43 EST
Sioux Falls, SD [sdww54]