Northern Prairie Wildlife Research Center
Mallard Recruitment in the Agricultural Environment of North Dakota
Models and Tests of Assumptions
Models have received increased use in waterfowl research. They are useful for
making generalizations (Bailey 1981), setting priorities for research (Ringelman
and Longcore 1980), and evaluating management options (Cowardin and Johnson
We found models useful for a variety of purposes in the conduct of this
study and analysis of results. Cowardin and Johnson (1979) presented a model
where size of population in year t + 1 was defined by the following
Nt + 1 = NtS + NtRS*
Nt = the number of hens in the spring of year t,
S = the annual survival rate of adult hens,
S* = the survival rate of immature hens from fall to spring, and
R = the number of hens recruited to the fall population per hen
in that year's spring population.
The parameter recruitment (R) was separated into subcomponents, which
R = Pe(1-P)2ZB ÷ 2,
P = nest success,
Z = survival of at least 1 brood member to approximately
age IIc, and
B = average brood size at fledging.
That model was the basis for selecting the attributes of recruitment to be studied,
and a slightly modified model will be used for predicting population change.
A stochastic model developed by statisticians and biologists at the Northern
Prairie Wildlife Research Center, Jamestown, North Dakota (Cowardin et al. 1983),
will be used to furnish a simulated population of nests for testing some of
the basic assumptions discussed later and to evaluate potential biases that
may have affected our study.
The basic structure of the model is illustrated in a flowchart (Fig. 11).
Modifications and adaptations will be described in conjunction with some of
our specific uses.
Previous Section -- Brood Survival
Return to Contents
Next Section -- Evaluation of Assumptions