Northern Prairie Wildlife Research Center
The 13 duck species reported here (Fig. 10) represent 39% of the duck species that breed in North America and 73% of the species in the genera Anas and Aythya in North America. During 1987-90, the density of the blue-winged teal (3.4 pairs/km²) was higher than those of the mallard, (2.1), gadwall (1.8), northern pintail (0.8), and redhead (0.8) (Fig. 11). Density was consistently highest on service-owned lands primarily because of the large areas of ponds.
Annual Change in Silences of Breeding Populations
The sizes of the breeding populations of the five most numerous dabbling ducks declined throughout 1987-90 as drought conditions intensified (Fig. 11 ). The declines corresponded closely to loss of ponds (Fig. 7) and pond area (Fig. 8), but the relation of pair density to area of ponds differed by species (Fig. 12). We expected species with a high degree of philopatry and possibly weak territoriality to concentrate on ponds as the number of ponds decreased because of drought. Slopes of the linear regressions of pair density on pond density (Fig. 13) were: -0.108 (gadwalls), -0.098 (mallards), -0.071 (blue-winged teals), -0.025 (northern shovelers), and -0.0004 northern pintails). The ranking is similar to published data on return rates (Anderson et al. 1992). Mallards and gadwalls exhibit strong philopatry (Lakemoen et al 1990). Our ranking of bluewinged teals is higher and of northern pintails lower than expected according to the literature. The comparison with the literature can only be approximate because of considerable spatial and temporal variations in published return rates, variation of return rates by age and sex, and rare correction of published return rates of mortality.
Our data (Fig. 12) may suggest that, where the correlation between the area of ponds and breeding population is low, the number of ducks is too low to fill the available breeding habitat. However, other explanations are possible, and we found no data that support a depression of northern shoveler populations. Johnson and Shaffer (1987) analyzed data from annual surveys by the service and concluded that estimated mallard population sizes no longer parallel estimated pond numbers. Their first possible explanation was that the number of mallards was too low to fill the habitat.
Change in the total number of pairs per km² by year varied among landownership classes (Fig. 14). The highest density was on service-owned land as expected because more wetland and more pond area are on these lands than on land in the other landownership classes. The decrease in pairs per km² during 1987-88 was less severe on service-owned lands than on other lands, probably because of the greater amount of semipermanent wetland basins on service-owned land. Birds that return to a landscape in a drought, where the less permanent ponds were dry, probably concentrated on large semipermanent wetland basins like those on service-owned land. The decline of the number of pairs was more apparent on easements than on private lands. More wetland basins are on easements than on private lands, but not as many large semipermanent and permanent wetland basins are on easements as on private lands (Fig. 8).
The number of recruits is the product of size of breeding population and the recruitment rate (Cowardin and Johnson 1979). Drought has a negative effect on both (Cowardin et al. 1985, Johnson and Grier 1988). We point out that our estimated recruitment rate was more dependent on model prediction than on observation and was highly influenced by the underlying assumptions of the model. The estimated density of the recruited ducklings (Fig. 15) followed the same general pattern as the sizes of the breeding populations among the five species (Fig. 10 and Fig. 15) and among years (Fig. 11 and Fig. 16) of the five species for which production was estimated. The estimated recruitment rates varied among species and among years (Fig. 17). The rates were highest in blue-winged teals and gadwalls and lowest in mallards and northern pintails. The annual variation in recruitment rates (Fig. 17) resulted from variation in A (a measure of nesting intensity; Table 9). Our estimates of A had a major effect on our estimates of hen success (Equation 5). The estimated clutch success in stable populations (Cowardin et al. 1985 and Klett et al. 1988) is lower than those presented in Table 10. However, the estimates presented in those papers were based on the assumption that A equaled 1. For the low A values in this study, higher clutch success is required for recruitment rates of a stable population. For mallards, a hen success of 31%, a summer survival of 0.74, and an average brood size of 4.9 (Cowardin et al. 1985) equate to a recruitment rate of 0.56 in a stable population. This is well above our estimated recruitment rate in mallards (Fig. 17).
We did not have sufficient data for estimating clutch success by year. Greenwood et al. (1995) showed that clutch success in the prairies of Canada is depressed by drought. If we had estimates of clutch success by year, the variation in our estimated recruitment rate by year (Fig. 17) would probably have been greater.
|J. Clark Salyer||0.744||0.658||0.654||0.630|