Limits on the predictability of phytoplankton competition

J. Huisman, University of Amsterdam, The Netherlands
F.J. Weissing, University of Groningen, The Netherlands

There are thousands of different phytoplankton species, slowly drifting in lakes and oceans. Some of these phytoplankton species are toxic, whereas others are relatively harmless. Some of these species are edible for zooplankton, whereas others are not eaten at all. Can we predict which of these species will become dominant?

Here, we argue that there are limits on the predictability of phytoplankton competition. Competition for one or two resources in a constant environment is highly predictable. However, more intricate competitive interactions may exhibit chaotic dynamics. In particular, multispecies competition models can have several alternative outcomes of competition. We show that these alternative outcomes of competition can be intermingled in the form of a fractal structure. As a result of the fractal geometry, it is fundamentally impossible to predict the winners of multispecies competition far in advance.

We therefore advocate a similar approach as in the weather forecast. Short-term predictions of changes in species composition are feasible, but long-term predictions are troublesome. Knowledge of the time scales of plankton processes may indicate proper time horizons for the predictability of phytoplankton species composition.

Reference:
Huisman & Weissing. 2001. Fundamental unpredictability in multispecies competition. American Naturalist 157: 488-494.