UNIQUENESS OF ECOSYSTEMS VERSUS GENERALIZABILITY AND PREDICTABILITY IN ECOLOGY

Any part of nature is unique and has its specific characteristics. If we want to make predictions about ecological development we need to generalize. This implies an abstraction from the singular aspects of reality. The forms of abstraction determine which aspects of nature we find to be important. We use the example of art historical analysis to show that the form of abstraction that is used to understand a specific scene is subject to historical change; this provides a framework for our interpretation. Three concepts are discussed that are fundamental for building ecological models: (1) nature is understood in the form of identical units, (2) quantity and quality can be treated as entirely distinct aspects, (3) integer numbers and the idea of a continuum are available to characterize quantitative transitions. We can model ecological processes if it is possible to express them in forms such as these. Conventional models focus on a specific level of abstraction. This usually restricts the representation of variability and heterogeneity. Using an individual based model it is shown that this quite realistic representation (a low level of abstraction) can generate patterns which allow us to reconstruct the type of irregularity and singularity we observe in nature. This suggests necessary limitations of the understanding of natural processes on a predictive basis.