Evaluating fishery impacts using metrics of community structure

Evaluation of the ecosystem effects of fishing presupposes that it is possible to measure the state of an ecosystem, detect differences among measures of its state, and determine if the differences are due to chance or impacts of fishing. Ecologists have developed or adopted four main classes of community/ecosystem metrics For such tasks: diversity and similarity indices, ordination methods, metrics of aggregate community properties. and metrics derived from ecosystem models. Each class is explained briefly, and guidance is given to the Literature. Size spectra, k-dominance curves, and multidimensional scaling emerge as promising methods for these uses. Some modelling frameworks also may be useful, but more work is needed on sensitivity testing and development of independent theory for these models. The current theoretical basis for most ecosystem models assumes that competition structures the system bring modelled. Problems with applying that framework to exploited marine systems are reviewed. Resampling methods are essential when testing for differences among community metrics. Practices for determining causality when ecosystem metrics change are reviewed, and additional knowledge and data necessary to link change to cause are highlighted. (C) 2000 International Council for the Exploration of the Sea.