A dynamic ecosystem simulation model of Mille Lacs Lake, Minnesota: Ecosystem-based management learns from the past to shape the future

Background/Question/Methods

Mille Lacs is a large 537km², shallow glacial lake in northern Minnesota that supports important native (Ojibwe) and recreational fisheries. As part of an initiative in ecosystem-based management, the Minnesota Department of Natural Resources provided extensive field data  that allowed the construction of a whole-ecosystem model. The lake food web is modelled with 50 functional groups of organisms, from primary producers like macrophytes, periphyton and phytoplankton through zooplankton, zooplanktivores like cisco, benthic organisms, predatory fish and top carnivores such as walleye, northern pike, muskellunge and birds. The model is constructed first as a static mass-balance model in which each functional group is parameterized using life history, production, consumption and diet items. The mass-balance model is then used to initiate a dynamic simulation that tracks biomass changes in the food web with time: the final ecosystem model results when model biomass changes are fitted to time series survey and fisheries data from the lake over a 22-year period from 1985 -2006.

Results/Conclusions

First, the ecosystem model is used to make estimates of the maximum sustainable yields (MSY) for the commercially exploited fish species in Mille Lacs and these values are compared to MSY values from conventional single species methods. In about half the cases the two MSY values are similar, but in the other half the ecosystem-based estimates point to lower recommended fisheries catches when food web interactions are considered, showing the value of an ecosystem-based approach in lake fishery management. Secondly, the dynamic ecosystem model is used to explore the consequences of scenarios of unexpected increases or decreases in key ecosystem organisms such as walleye, pike and fish eating birds (cormorants) and the fisheries themselves. When there is no fishing, the model suggests increases in predatory species, but reductions in the abundance of some planktonic organisms due to trophic cascades. Similar outcomes ensue for increases and decreases of predatory organisms. We discuss the likely realism of these scenario predictions, show how parameter uncertainty may be incorporated through Monte Carlo simulations, and finally suggest how the ecosystem model may be used to address invading species such as zebra mussels and crayfish