Concern about the impact of fishing on ecosystems and fisheries production is increasing (1, 2). Strategies to reduce these impacts while addressing the growing need for food security (3) include increasing selectivity (1, 2): capturing species, sexes, and sizes in proportions that differ from their occurrence in the ecosystem. Increasing evidence suggests that more selective fishing neither maximizes production nor minimizes impacts (4-7). Balanced harvesting would more effectively mitigate adverse ecological effects of fishing while supporting sustainable fisheries. This strategy, which challenges present management paradigms, distributes a moderate mortality from fishing across the widest possible range of species, stocks, and sizes in an ecosystem, in proportion to their natural productivity (8), so that the relative size and species composition is maintained.

The usefulness and relevance of size-based indicators (SBIs) to an ecosystem approach to fisheries (EAF) are assessed through a review of empirical and modelling studies. SBIs are tabulated along with their definitions, data requirements, potential biases, availability of time-series, and expected directions of change in response to fishing pressure. They include mean length in a population, mean length in a community, mean maximum length in a community, and the slope and intercept of size spectra. Most SBIs can be derived from fairly standard survey data on length frequencies, without the need for elaborate models. Possible fishing- and environment-induced effects are analysed to distinguish between the two causes, and hypothetical cases of reference directions of change are tabulated. We conclude that no single SBI can serve as an effective overall indicator of heavy fishing pressure. Rather, suites of SBI should be selected, and reference directions may be more useful than reference points. Further modelling and worldwide comparative studies are needed to provide better understanding of SBIs and the factors affecting them. The slow response to fishing pressure reflects the complexity of community interactions and ecosystem responses, and prohibits their application in the context of short-term (annual) tactical fisheries management. However, movement towards longer-term (5-10 years) strategic management in EAF should facilitate their use

Population and community descriptors that might be used as indicators of the impact of fishing are reviewed. The criteria used for the evaluation of these indicators are meaning, expected effect of fishing, exclusiveness to fishing effects, and measurability. Population indicators such as total mortality rate, exploitation rate, or average length are the most operational indicators because their meaning is clear and the expected effect of fishing on them is well understood so that reference points can be set. On the other hand, indicators based on the composition of species assemblages such as diversity indices and ordination of species abundances are difficult to interpret, and the effect of fishing on them is not easily predicted. Robust indicators describing the community functions of interest (production and transfer of biomass to large fish), such as size spectra descriptors or the proportion of piscivorous fish in the community, are more promising but are not yet well developed. New candidate indicators are proposed: the change in fishing mortality required to reverse population growth rate, the proportion of noncommercial species in the community, and the average length and weight in the community.

A new time-dependent continuous model of biomass size spectra is developed. In this model, predation is the single process governing the energy flow in the ecosystem, as it causes both growth and mortality. The ratio of predator to prey is assumed to be distributed: predators may feed on a range of prey sizes. Under these assumptions, it is shown that linear size spectra are stationary solutions of the model. Exploited fish communities are simulated by adding fishing mortality to the model: it is found that realistic fishing should affect the curvature and stability of the size spectrum rather than its slope.

No abstract