Freshwater Biology (1994) 32, 143-160 OPINION Pattern, process and prediction in aquatic ecology. A limnological view of some general ecological problems GRAHAM P. HARRIS CSIRO Institute of Natural Resources and Environment, PO Box 225, Dickson, ACT 2601, and Co-operative Research Center for Freshwater Ecology, University of Canberra, Kirinari St., Bruce, ACT 2617, Australia SUMMARY 1. New sdentific concepts such as models of chaos, complex dynamics and non-linear interactions have the potential to contribute to an improved understanding of ecological patterns and processes. This paper discusses some of the known dynamics of phytoplankton, pelagic food chains and nutrient cycles in the light of some of these new concepts. The paper brings these new conceptual models together with data from a wide range of sources in an attempt to produce a synthesis of system behaviour which allows us to understand why some things are inherently more predictable than others. In particular it looks at the limnological management tools of empirical biomass models and biomanipulation and at the need for prediction of spedes composition. 2. The structures observed in ecosystems (nutrient pools, sizes, spedes, temporal/spatial patterns) show properties at a spectrum of scales, as do the processes (fluxes, grazing, competition). Both respond to a spectrum of extemal perturbations that may be climatologically or anthropogenically induced. Empirical biomass models work because of the annual averaging of pattern and process and because of some inherent properties of the functioning of pelagic ecosystems. Many aspects of ecosystem pattern and process vary in a regular way with trophic state. Examination of empirical data sets can lead to an improved understanding of system behaviour if questions are asked about why things happen the way they do. 3. Feedbacks between pattern, process and periodidty are seen to be an inherent property of the system. Understanding the fundamental dynamics of non-linear interactions in ecosystems may make it possible to exploit the extemal spectrum of environmental perturbations and to control system function. For example, by imposing external physical perturbations on pelagic systems it may be possible to manipulate the spedes composition of the phytoplankton community. Because of the complexity of possible interactions both 'horizontally' between spedes and 'vertically' within the food chain, any prediction of spedes composition will necessarily be probabilistic. Preamble: a new ecological vocabulary? ^^^^' King^l^"^, 1985). It comes as no surprise to find, therefore, that in this so-called 'postmodern' era The conceptual basis of sdence is not free from the (Jameson, 1991) at 'the end of history' (Fukuyama, biases of sodety in general and many questions can 1992) when a belief in certainty has evaporated, sd- only be addressed in particular ways depending on , ence itself is being called into quesfion (Appleyard, the philosophical climate of the day. Concepts and 1992) and ecology is going through a period of re- vocabulary determine what can be discussed and assessment also. Peters (1991) has cdfidzed the how. Ecology is no excepfion and has a strong found- widely accepted theoretical bases of ecology and has ation in sodal and, even, economic theory (Mclntosh, cridcized ecologists for their adherence to 'elusive 143