USGS - science for a changing world

Northern Prairie Wildlife Research Center

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

Habitat Use and Movements of Canvasback Broods in Southwestern Manitoba


The breeding biology of canvasbacks has received intensive study in a number of areas of the Canadian parklands (Smith 1971; Stoudt 1971, 1982; Anderson 1984). However, few studies have investigated canvasback brood ecology, and most of these relied on observations of unmarked broods. Evans et al. (1952) observed extensive movements of broods among wetlands. Studies of other species, mostly dabbling ducks, documented extensive brood movements among wetland and wetland complexes (Talent et al. 1982, Duncan 1983, Smith and Flake 1985). Studies of habitat use or movements of diving ducks are generally lacking.

We documented habitat use and mobility of canvasback broods accompanying marked females in the aspen parklands of southwestern Manitoba. We also evaluated characteristics of ponds used by canvasback broods of different ages.

Study Areas and Methods

Investigations were conducted during 1974-80 in a 181.3-km² area near Minnedosa, Manitoba (50° 15' N, 99° 50' W). This region is among the most productive breeding areas for canvasbacks (Bellrose 1980) but has been significantly impacted by agriculture and wetland drainage (Adams and Gentle 1978, Rakowski and Chabot 1983, Turner et al. 1987). Wetlands, topography, vegetation, and breeding waterfowl of the area were described by Kiel et al. (1972), Adams and Gentle (1978), and Stoudt (1982). Brood observations collected in 1975 were excluded because of small sample sizes. We observed no canvasback broods in 1977, when drought conditions resulted in low nest success for canvasbacks (Serie, unpubl. data).

We conducted thorough searches of a 15.5-km² study block during a brood count in mid- or late July each year. Observations of canvasback broods were also recorded during intensive, almost daily searches on foot of portions of the study area from mid-June through early August. We identified individual broods by the accompanying brood hens, which were uniquely marked with nasal saddles (Doty and Greenwood 1974), using binoculars and spotting scopes. Data recorded for each observation included date, pond location, brood hen identity, and the size, age, and species composition (number of canvasbacks and parasitic redheads) of broods. Nest location (by pond number), hatch dates, and number of ducklings hatched were known from a concurrent study (Serie, unpubl. data).

For each location of a brood, we determined pond class according to Stewart and Kantrud (1971) (temporary, seasonal, semipermanent, and permanent) and size (small [≤0.2 ha] or large [>0.2 ha]).

We plotted locations of individual broods on a digitized map of the study area. Because brood locations indicated only which pond they were on, we determined distances moved among ponds by measuring the distance between the centers of ponds to the nearest meter. We included distances moved from the nesting pond to the first observation of the brood.

Data analysis

Broods were classified into three age classes (≤14 days old, 15-28 days old, and >28 days old), based on known hatch dates. We excluded brood observations where age, number of ducklings, or species composition were inconsistent with hatch information and previous observations. For all years combined, we used contingency table analyses to analyze pond use (class and size) relative to brood age class. Because of small sample sizes, we pooled temporary and seasonal ponds (hereafter referred to as seasonal; 85% of these were seasonal). Thus, we compared use among seasonal, semipermanent, and permanent ponds and between small and large ponds. To minimize bias of irregular observations of broods, we summarized data as average distance moved per day, using only movements that occurred within an age class. We examined average distances moved per day by broods relative to age class, year, and hen using a repeated measures model.

Results and Discussion

Over five years, we collected 202 observations on 69 canvasback broods. Frequency of sightings varied somewhat by brood age (≤14 days old = 2.3 sightings/brood, 15-28 days old = 1.8, and >28 days old = 1.6). Brood age at last observation averaged 30 days (range 3-53 days).

Broods were observed on an average of three different ponds (range 1-6). Over all brood age classes, 7% of sightings occurred on seasonal ponds, 81% on semipermanent ponds, and 12% on permanent ponds (Table 1). We were able to directly compare use (expressed as percentage of brood observations) and availability of ponds by class in 1978, when a thorough survey of the 15.5-km² area was conducted in July. Broods selected semipermanent (94% of brood observations vs. 43% of ponds available) and permanent ponds (4% vs. 1%) but exhibited minimal use of seasonal ponds (2% vs. 56%).

Table 1.  Use of different sizes and classes of wetlands by canvasback broods near Minnedosa, Manitoba, 1974-80
Duckling age N a Seasonal Semipermanent Permanent
≤ 0.2 ha > 0.2 ha ≤ 0.2 ha > 0.2 ha > 0.2 ha
≤14 days old 91 3 (3)b 3 (3) 26 (29) 52 (57) 7 (8)
15-28 days old 87 2 (2) 4 (5) 10 (11) 59 (68) 12 (14)
>28 days old 64 0 5 (8) 6 (9) 44 (69) 9 (14)
All age classes 242 5 (2) 12 (5) 42 (17) 155 (64) 28 (12)
a Number of sightings
b Number (percentage in parentheses) of broods observed

Use of ponds varied by pond class and size and by brood age class (X² = 17.25, 8 df, P = 0.028) (Table 1). When differences in use by pond class and size were examined separately, we found no significant differences relative to pond class (X² = 2.32, 4 df, P = 0.677). However, pond size was significantly related to the age class of the brood (X² = 14.90, 2 df, P = 0.001). Canvasback broods less than 15 days old used smaller ponds more often than older broods (32% vs. 12% of sightings), and broods greater than 28 days old used mostly large ponds (88% of sightings). This relationship between pond size and brood age may result in part from declining availability of small ponds late in the season.

Studies of prairie- or parkland-breeding mallards (Anas platyrhynchos), northern pintails (Anas acuta), and wood ducks (Aix sponsa) have documented use of up to ten different wetlands and movements of up to 9 km (Dzubin and Gollop 1972, Talent et al. 1982, Duncan 1983, Smith and Flake 1985). Erickson (1948) reported movements of up to 6.4 km by canvasback broods in Oregon. In our study, canvasback broods moved an average of 195 m (SE = 28 m) between observations and were observed to move as far as 4.4 km over the observation period. Our estimates of movements between discrete wetlands are probably conservative because we did not observe broods daily, and observations did not include distances moved within discrete ponds. Broods observed on two consecutive days moved up to 985 m between ponds. Eighty-three percent of broods observed within seven days of hatch (n = 46) had moved from the nesting pond. Three broods (one in 1974 and two in 1979) hatched on large, semipermanent, or permanent ponds, were never observed off the nesting pond during 31-36 days of brood-rearing. Broods occasionally revisited ponds; four broods returned to the nesting pond after moving to other ponds, and five broods returned to ponds they had previously used. Of 11 brood females observed more than one year, two used some of the same ponds for brood-rearing each year; one of these was sighted on the same ponds over three years.

Brood movements appear to be most extensive within two weeks of hatching (Talent et al. 1982, Smith and Flake 1985, Leonard 1990). We found the average distance moved per day was highest in broods less than 15 days old (Table 2). However, the repeated measures model, incorporating age class, year, and hen identity, was nonsignificant (F = 1.32; 108, 86 df; P = 0.090). Leonard (1990) also found no differences in frequency or distance of moves by canvasback broods by age; weekly distance traveled by broods was greatest during the first two weeks, but the difference was not significant.

Table 2.  Mean distance moved per day by canvasback broods of different age classes, Minnedosa, Manitoba, 1974-80.
Year Age Class a No. broods No. obs. Mean distance moved per day (m)
x SE
1974 1 5 12 231 144
2 4 11 110 85
3 4 6 166 105
1976 1 9 14 366 98
2 5 10 252 97
3 6 12 356 123
1978 1 11 21 272 74
2 3 5 128 52
3 6 9 360 151
1979 1 28 49 304 68
2 17 25 172 53
3 4 5 0 0
1980 1 4 12 600 97
2 2 3 366 366
3 1 1 0 0
Mean 1 48 108 330 42
2 30 54 181 39
3 20 33 258 66
a 1 = ≤14 days old; 2 = 15-28 days old; 3 = >28 days old

Brood studies have often sought to link overland or among-wetland movements with duckling survival (Talent et al. 1982, Smith and Flake 1985, Leonard 1990). Our data on broods observed on the first day after hatch or on two consecutive days were inconclusive. Of 12 broods observed the day following hatching, six did not move from their nesting pond, and two of these broods lost 1-2 ducklings (as determined from number hatched at the nest). Of the six broods that moved from the nesting pond, three lost 1-2 ducklings. No ducklings were lost in the longest movements from nesting ponds (400 m, n = 2). We observed 12 broods on two consecutive days. Only one brood of seven that did not move lost a duckling, whereas three of five broods that did move between the ponds lost a duckling each. One brood that moved three times between two ponds over two days lost two ducklings. Whether the brood lost each duckling during the overland movement is unknown.

Efforts to conserve or restore wetland habitat for breeding waterfowl must consider the varying habitat needs and extensive movements of waterfowl during the entire breeding season. Breeding canvasbacks have relatively large home ranges (up to 525 ha, Dzubin 1955; Serie, unpubl. data) and use a variety of wetlands during the nesting and brood-rearing periods. Large semipermanent wetlands are used by canvasbacks during both nesting and brood-rearing, whereas seasonal wetlands are used primarily during nesting (Sugden 1978, Stoudt 1982). Temporary wetlands appear relatively unimportant to breeding canvasbacks, although they may be used for nesting (Sugden 1978, Stoudt 1982). Temporary or seasonal ponds may also provide corridors for movement among semipermanent wetlands within a wetland complex (Leonard 1990). Females may select wetlands with relatively high invertebrate abundance (Talent et al. 1982, Smith and Flake 1985) and move their broods to take advantage of insect emergences (Swanson and Sargeant 1972) or to adapt to changing food needs of the growing ducklings (Sugden 1973). Thus brood habitat should include several adjacent wetlands which include semipermanent wetlands. Further studies are needed to relate wetland conditions to changing food availability and dietary requirements of ducklings.

Return to Contents
Next Section -- Literature Cited

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 18:37:08 EST
Menlo Park, CA [caww54]