Factors Associated with Duck Nest Success
in the Prairie Pothole Region of Canada
Nest Success and Percentage Cropland
We conducted an analysis of covariance (Milliken 1990) to examine the relation between nest success (the response variable) and percentage of Cropland (includes standing stubble, tilled stubble, and fallow land) and physiographic zone (parkland or prairie). We attempted to predict (within the observed range of nest success and habitat conditions) the threshold level of Cropland availability above which nest success would be insufficient to sustain populations of the 5 common species. Minimum threshold levels of nest success assumed to be necessary for population stability were 15% for mallards (Cowardin et al. 1985), 15% for northern pintails (Klett et al. 1988), and 20% each for gadwalls, blue-winged teals, and northern shovelers (Klett et al. 1988). Mean nest success among area-years was examined to determine if nest success depended on percentage Cropland available in parkland or prairie. Significant results (P < 0.05) were tested separately for each species to isolate which physiographic zone was responsible for rejection of the null hypothesis that regression slopes were equal to zero. After we established the relation between nest success and percentage of Cropland by zone, we predicted percentage of Cropland at suggested thresholds of nest success using the "inverse prediction" procedure described by Neter et al. (1985). Statistical comparisons were performed with GLM procedures of SAS (SAS Institute, Inc. 1989).
Previous Section -- Nest Fate
Return to Contents
Next Section -- Seasonal Variation in Nest Predation Rate
