Grazers as Sources and Sinks for Nutrients:Conclusions, Limitations, and Speculations
213
fish. Since both vertebrate and invertebrate predators are generally sizeselective, an extension of the present models to include higher trophic
levels would probably require all consumer populations to be represented
by structured population models.
The representation of perennial organisms like fish would probably also
require a closer consideration of seasonal cycles, which have been completely neglected in the stratified summer situation implicitly assumed in
the present models. A realistic representation of seasonal cycles in lakes
would then again probably also require consideration of the major physical
processes controlling stratification and vertical transport. In principle, such
a vertically structured model could also be used to model the evolutionary
games involved in vertical migration (e.g. Mangel and Clark 1988),
although the computational costs of such models would probably be very
high. Some would even argue that the whole three-dimensional flow field
should be represented in a realistic lake model, since lateral diffusion and
spatial heterogeneity are thought to deeply influence the stability properties of interacting populations (e.g. Powell and Richerson 1985).
While physically and demographically structured models (e.g. Slagstad
1981, 1982) are definitely superior in terms of realism and generality, it is
not at all certain whether the added amount of dynamic complexity and
parameter uncertainty has a proportional reward in terms of predictive
power and biological insight compared to simple, unstructured models.
Perhaps a better strategy for future research would be to use minimal
models of the kind presented here to design critical experiments and identify sensitive parameters, and then proceed to more complex, structured
models when the potential of minimal models has been definitively
exhausted.
213
fish. Since both vertebrate and invertebrate predators are generally sizeselective, an extension of the present models to include higher trophic
levels would probably require all consumer populations to be represented
by structured population models.
The representation of perennial organisms like fish would probably also
require a closer consideration of seasonal cycles, which have been completely neglected in the stratified summer situation implicitly assumed in
the present models. A realistic representation of seasonal cycles in lakes
would then again probably also require consideration of the major physical
processes controlling stratification and vertical transport. In principle, such
a vertically structured model could also be used to model the evolutionary
games involved in vertical migration (e.g. Mangel and Clark 1988),
although the computational costs of such models would probably be very
high. Some would even argue that the whole three-dimensional flow field
should be represented in a realistic lake model, since lateral diffusion and
spatial heterogeneity are thought to deeply influence the stability properties of interacting populations (e.g. Powell and Richerson 1985).
While physically and demographically structured models (e.g. Slagstad
1981, 1982) are definitely superior in terms of realism and generality, it is
not at all certain whether the added amount of dynamic complexity and
parameter uncertainty has a proportional reward in terms of predictive
power and biological insight compared to simple, unstructured models.
Perhaps a better strategy for future research would be to use minimal
models of the kind presented here to design critical experiments and identify sensitive parameters, and then proceed to more complex, structured
models when the potential of minimal models has been definitively
exhausted.