Assessing Changes in Alewife and Rainbow Smelt Abundance in Relation to Salmonine Consumption in Lake Michigan

Since 1967, millions of Pacific salmonines, including Chinook salmon (Onchorhynchus tschawytscha), lake trout (Salvelinus namaycush), rainbow trout (Oncorhynchus mykiss), brown trout (Salmo trutta), and Coho salmon (Oncorhynchus kisutch), have been stocked in Lake Michigan every year to rehabilitate native populations, reduce abundance of exotic prey fishes and support economically important recreational fisheries.  A central management question faced by Lake Michigan fishery managers is how many fishes of each salmonine species to stock each year while maintaining adequate prey fish abundance.  Success in the management of salmonine fisheries requires an understanding of the trade-offs between the bio-energetic demands of salmonine predators and the recruitment dynamics of their prey species.  One way to improve understanding of these trade-offs is by constructing multispecies age-structured models that incorporate feedback interactions between prey and predator populations.  In this study, we reconstructed the population dynamics of alewife (Alosa pseudoharengus) and rainbow smelt (Osmerus mordax) in Lake Michigan by fitting a multispecies age-structured model to prey fish survey data (from bottom trawl and hydroacoustic surveys) and salmonine consumption estimates (from predator assessment models) from 1962 to 2008.  The prey dynamics model was constructed to track changes in abundance of alewife and rainbow smelt in relation to changes in abundance of the salmonine populations by incorporating the dynamic links between prey and predator populations through a multi-species functional response.  In fitting the prey model to time-series of “observed” data, we estimated annual recruitment for alewife and rainbow smelt and the effective search rate for the different salmonines.  Model estimates showed that alewife recruitment has declined since the late 1990s.  The decline in recruitment corresponds with an estimated increase in predation pressure during those years, suggesting that in recent years alewife have in effect been “recruitment overfished” by their predators.  By reconstructing historic alewife and rainbow smelt population numbers, the prey model also allowed us to estimate stock-recruitment parameters for both species.  Results from this model, along with outputs from the age-structured predator assessment models, are now being used as input to a simulation model evaluating alternative salmonine stocking policies in view of their effect on the prey-predator balance in the lake and other key management objectives.