Summary

Models are analogies representing important features of a system. By constructing and then studying a model we hope to explain how the system functions. However, mathematical models are made using particular rules and procedures of construction. These determine both the types of simplification made and how the model can be assessed. Rules and procedures for constructing three types of ecological model are reviewed and their use in ecological research is discussed.

Dynamic systems models have a defined mathematical form, usually differential or difference equations. They have a long history of describing idealized ecological interactions and speculating about general properties of ecological systems. Their use has been criticized because of diffculties in assessing how effectively they can represent particular ecological systems. Statistical models have a mathematical structure specifically designed for fitting to data and calculating error. Parsimony is an important procedure in construction – using no more complicated a model than is needed to describe patterns in measured data. An example of a model of a stochastic process is described. Statistical models have been criticized because of diffculties in interpreting ecological significance of parameters and lack of uniqueness, i.e., that different models may fit a data set equally well. General explanations have to be built up by comparing models constructed in different instances.

Systems simulation models take advantage of increased computing power and flexibility in computer languages to make more comprehensive representations of ecological processes than either dynamic systems models or statistical models. However, the potential for complexity can lead to difficulty in deciding a model's bounds, what should, and should not be included, and how to assess the model.