Northern Prairie Wildlife Research Center
Factors Associated with Duck Nest Success
in the Prairie Pothole Region of Canada
Mallard Production
Variation in Components
Nest success rate appeared to be the most influential factor in determining annual
production of mallards, as indexed by numbers of successful nests. The size of
the breeding population also was strongly influential, although its 2 constituents,
number of wet wetlands and number of mallard pairs per wetland, individually did
not relate strongly to production. A measure of nesting effort (estimated nests
per pair) was significantly related to production as well. Despite shortcomings
in some methods we used to estimate some values, this analysis clearly shows for
a variety of areas studied in several years that nest success rate is the most
influential component of reproduction among the components we measured. This finding
is consistent with conclusions reached in several other studies in the PPR (e.g.,
Cowardin and Johnson 1979, Cowardin et al. 1985, Johnson et al. 1986, Klett et
al. 1988, Johnson et al. 1992).
Geographic and Temporal Factors
A major finding from this analysis is the considerable variability in estimated
values of variables associated with mallard production. Geographic variation was
high, temporal variation was high, and the unexplained variation was high. This
last term is often called "error variance" because it reflects the difference
between observed values and those predicted by some model. Estimated components
of variance due to error exceeded those due to geographic and temporal variability
for all factors except temporary wetlands. In the present instance, a large error
variance suggests that the variable of interest, such as number of wet wetlands
or nest success rate does not depend solely on the study area and year, but that
it also varies in response to other factors. Some of the other factors that may
be responsible for this high error variance were discussed earlier (e.g., availability
of buffer prey, changes in habitat conditions, and habitat fragmentation).
The interaction between study areas and years for temporary and seasonal wetlands
is further evidence that drought, as it affected these 2 wetland counts, was
not manifested equally across the PPR of Canada, but affected different areas
at different times. The finding that semipermanent wetlands varied less among
years than among areas is not surprising in view of their considerable persistence
from year to year.
Numbers of mallard pairs varied less than numbers of wet wetlands of any class,
which suggests that mallard numbers are not as volatile as wetland conditions
and that mallards respond to factors other than numbers of wet wetlands (e.g.,
philopatry) when settling their breeding habitat.
Our finding that nest success was more variable among years than among areas
was somewhat surprising, because predation was the cause of most nest failures
and indices for many predator species varied considerably from area to area,
but varied little among years (Sargeant et al. 1993). Clearly predator numbers
were not the sole determinant of the nest success (Johnson et al. 1989). Other
factors, such as the abundance of buffer prey, may have contributed substantially
to temporal variation. In addition, predation rates were related to weather
variables as indexed by the wetland counts, which did vary temporally.
Previous Section -- Mortality of Adult Ducks
Return to Contents
Next Section -- Conclusions
U.S. Department of the Interior |
U.S. Geological Survey