Uncertainty in the population dynamics of alewife (Alosa pseudoharengus) and bloater (Coregonus hoyi) and its effects on salmonine stocking strategies in Lake Michigan

The dynamics of alewife (Alosa pseudoharengus) and bloater (Coregonus hoyi) in Lake Michigan were investigated and the implications of uncertainty in these dynamics on the outcomes of stocking strategies for salmonines were simulated using a stochastic model. I also analyzed the long-term trends in bloater size at age with a dynamic growth model for evidence of density-dependent growth regulation.; I fit a dynamic von Bertalanffy model and length-weight relationship with time-varying parameters to mean length and weight at ages of bloater from annual surveys. My results support a positive relationship between asymptotic length, L_∞, and the Brody growth coefficient, k, indicating that under conditions supporting larger L _∞, individuals approach L_∞ more rapidly. I explored the relationship between year-specific growth parameters and bloater abundance indices and found evidence of density-dependent growth. However, in the most recent years, L_∞ and yearling length have remained low despite low bloater abundances, suggesting a potential shift in the food web.; I reconstructed the population dynamics of alewife and bloater by fitting dynamic models to historic prey fish survey data (bottom trawl and hydroacoustic indices). These models allowed recruitment variation and accounted for mortality due to salmonine predators. Chinook salmon predation followed a Type II functional response while other predators were assumed to be consuming at a constant rate. Estimates of consumption based on existing assessments of predators were also used in model fitting. The joint posterior distribution of the stock-recruitment parameters for alewife and bloater and a key parameter of chinook salmon's functional response was approximated using Markov Chain-Monte Carlo methods. While the amount of uncertainty in the parameters of the stock-recruitment relationship for alewife and bloater was large, the uncertainty in the parameter of the functional response was moderate.; To assess the impacts of these uncertainties on the outcomes of salmonine stocking strategies in Lake Michigan, I constructed a stochastic simulation model to forecast the outcomes of stocking strategies on both alewife and chinook salmon population dynamics. Uncertainty in the stock-recruitment relationship for alewife, the functional response of chinook salmon and the response of chinook salmon mortality rates to decreases in growth rate were included in the model. I investigated both fixed and dynamic stocking strategies; for the latter, stocking rates responded to changes in the state of the system. The outcomes of all stocking strategies I considered were highly variable, with a large proportion of undesirable outcomes. The abundance of other salmonine predators had large effects on the dynamics of the system. This suggests that the salmonines require integrated management and our ability to maintain a desirable state long-term through stocking strategies alone may be limited. Removing uncertainty in stock-recruitment parameters for alewife caused the probability of undesirable outcomes to decrease suggesting that ignoring uncertainty in these parameters will cause overly optimistic predictions of future outcomes.