Factors Associated with Duck Nest Success
in the Prairie Pothole Region of Canada
Selection Criteria
We selected study areas on a nonrandom basis from surveyed transects that had >=8 pairs/km² of breeding mallards on the air-ground segment during 1977-81 (Off. of Migr. Bird Manage., USFWS, Laurel, Md., unpubl. data). We used surveyed transects as our sampling universe because they have wide geographic distribution and our study supported other research associated with them (Brace et al. 1987); recent aerial photography was available for air-ground segments of these transects, as were annual counts of ducks and wetlands. We selected transects that had relatively high densities of breeding mallards to increase our chances of finding adequate numbers of nests to estimate mallard nest success. We apportioned numbers of study areas similarly in parkland and prairie. We recognize the limitations on inferences that can be made due to our nonrandom sampling, but selection of high density areas was necessary to obtain adequate sample sizes of nests.
In 1982, to evaluate logistics, study procedures, and the amount of nest searching that field personnel could accomplish, we selected 1 area in parkland and 2 in prairie of central Saskatchewan; CWS biologists provided input regarding 1982 selections. We found that a crew of 5 persons could work on 2 areas <= 250 km apart. Resources limited us to 4 such crews in 1983 and 5 each in 1984-85; each crew annually collected data on 2 areas.
We continued data collection for 3 more years beyond 1982 on 2 of the areas. To acquire information on both geographic and temporal variation in nest success with limited resources, we added new areas each year (1983-85) on which we obtained data for 1 or 2 years. We thus obtained data from 7 areas for 1 year, 8 areas for 2 years, and 2 areas for 4 years—a total of 31 area-years (Table 1). The use of the same study area in successive years resulted in some loss of statistical independence, because habitat conditions and predator populations were likely similar in both years.
We selected study areas with the intent of obtaining wide geographic distribution within the PPR of Canada, excluding the southern PPR of Alberta where we did not work because drought there since 1979 had severely reduced numbers of wet wetlands and breeding ducks to the extent that we doubted we could find numbers of nests needed. In 1983, crews worked in the northwestern PPR of Alberta, southern and central PPR of Saskatchewan, and southwestern Manitoba. In 1984, we continued work in those localities and added another crew in the north-central PPR of Saskatchewan.
The decision to work on an area for 1 or 2 years was made annually for logistical convenience. New study areas were selected to obtain maximum distance (<=250 km) from study areas that were retained. We had no personal knowledge of any area before it was selected in 1983-85, except its location and density of mallard breeding pairs. Study areas were retained for a second year without regard to current mallard breeding population, upland or wetland habitat conditions, or number of nests found previously.
Previous Section -- Definition
Return to Contents
Next Section -- Habitat
