Comparison of Dynamic Models for Edible Phytoplankton

Alternative models for the dynamics of edible phytoplankton were compared using long-term data from mesotrophic Paul Lake and eutrophic Lake Mendota. Alternative models fit to the data contrasted linear versus logistic algal growth, type I versus type II functional response, and prey-dependent versus ratio-dependent predation. In both lakes, the model with lowest prediction error had logistic algal growth and a type I, prey-dependent functional response. Under these models, the spring bloom and clear-water phase of productive lakes can be explained as an incomplete predator–prey cycle: the spring pulse of edible algae is followed by a peak of Zooplankton; then, edible algae are overgrazed leading to the clear-water phase and collapse of the grazer biomass. This study demonstrates the use of time series data at the ecosystem scale to identify process-based models, contrast alternative models on a probabilistic basis, and estimate parameters. This approach avoids the assumptions involved in extrapolating ecosystem models from smaller scale studies.