Sensitivity Analyses of the Life Cycle of Midcontinent Mallards

Relationships between vital rates and population growth rate (λ) are critical to understanding and managing population dynamics. Considerable study of the midcontinent mallard (Anas platyrhynchos) population has been directed to understanding how vital rates respond to environmental fluctuations and management, but inference to the relative importance of specific vital rates to λ remains weak. We used analytic and simulation-based sensitivity analyses of a stage-based matrix model of female midcontinent mallards to compare the relative importance of vital rates to λ. For each vital rate, we estimated mean values and process variation (biological variation across space and time) for females breeding on sites of approximately 70 km2in the Prairie Pothole Region (PPR) of the United States (Montana, North Dakota, South Dakota, Minnesota) and Canada (Saskatchewan, Manitoba, Alberta). We conducted perturbation analyses (i.e., analytic sensitivities and elasticities) to predict the relative influence of changes in vital rates on λ. We conducted variance decomposition analyses to assess the proportion of spatial and temporal variation in λ explained by process variation in each vital rate. At mean values of vital rates, analytic sensitivity of λ was highest to nest success and survival of adult females during the breeding season and non-breeding season; hence, equal absolute changes in these vital rates would be predicted to result in the largest Δλ, relative to other vital rates. Variation in sensitivities and elasticities across process variation in vital rates was primarily explained by variation in nest success and survival of ducklings. Process variation in breeding parameters was driving variation in λ: vital rates explaining the most variation were nest success (43%), survival of adult females during the breeding season (19%), and survival of ducklings (14%). Survival of adult females outside the breeding season accounted for only 9% of variation in λ. Our analyses suggested that predation processes on the breeding grounds were the primary proximate factors limiting population growth.