Predicting the ecological consequences of environmental change: a review of the methods*


  • *

    The Tansley Lecture was delivered at the British Ecological Society Annual Meeting on 9 Sept 2003 at Manchester Metropolitan University.

William J. Sutherland, Centre for Ecology, Evolution and Conservation, School of Biological Sciences, University of East Anglia, Norwich NR4 7TJ, UK (fax 01603 592250; e-mail


  • 1There is a clear need to increase our ability to predict the consequences of environmental change. The seven main approaches that are currently used are: extrapolation, experiments, phenomenological models, game-theory population models, expert opinion, outcome-driven modelling and scenarios. Each approach has different strengths and weaknesses. In practice, several approaches are often combined.
  • 2Adaptive management aimed at testing hypotheses is excellent in principle and widely advocated. In reality, however, it is almost never carried out because the changes in management usually have to be severe in order to bring about detectable changes in a reasonable time, and the political risks of such management are usually considered too high.
  • 3Game-theory population models are used to determine population-level phenomena based upon the decisions individuals make in response to resource depletion, interference, territoriality or rank. This allows predictions to be made regarding responses to novel conditions. The main drawback is that for some models considerable information is required.
  • 4Much of conservation practice is not based upon evidence. Evidence-based conservation is the practice of accumulating, reviewing and disseminating evidence with the aim of formulating appropriate management strategies. Evidence-based medicine revolutionized medical practice and similar opportunities exist to improve conservation practice.
  • 5Synthesis and applications. The conventional approach of making assumptions and deriving models to make predictions about the consequences of environmental change is often unsatisfactory for complex problems, with considerable uncertainty. Tackling such problems is likely to require greater exploration of techniques such as expert opinion, output-driven modelling and scenarios.