Mesocosms in ecotoxicology (1): Outdoor aquatic systems.

Mesocosms have been used in aquatic ecotoxicology for approximately 20 years and were sometimes claimed to be essential tools, especially for regulatory purposes. The term aquatic mesocosm currently describes indoor and outdoor artificial streams or experimental ponds and enclosures. The use of mesocosms refines the classical methods of ecotoxicological risk assessment because mesocosms provide conditions for a better understanding of environmentally relevant effects of chemicals. They make it possible to assess effects of contaminants by looking at the parts (individuals, populations, communities) and the whole (ecosystems) simultaneously. Ecotoxicological investigations in mesocosms will not entirely replace the use of laboratory animals. However, they allow tests to be performed on species that are not of major societal concern, but which play key roles in the structure and function of ecosystems. In this respect, mesocosms allow nondestructive measurements of integrated endpoints. They also appear as potent tools to predict changes at the highest levels of organization (population, community, and ecosystem) from measurements of individual endpoints. However, after a period of extensive use, regulatory studies using large-scale mesocosms were more or less abandoned at the beginning of the 1990s, mainly because their cost-effectiveness was questionable. This review covers key features of outdoor aquatic mesocosms that can be critical for their use in environmental risk assessment of chemicals and emphasizes the optimization of their use for such purpose. The originality of mesocosms is mainly based on the combination of ecological realism, achieved by introduction of the basic components of natural ecosystems, and facilitated access to a number of physicochemical, biological, and toxicological parameters that can be controlled to some extent. This characteristic determines various features of the systems such as the minimal size required, initial physicochemical and biological composition, or choice of model species for ecotoxicological investigations. Ecological maturity of mesocosms affects the degree of variability of both physicochemical and biological parameters used to investigate the impact of contaminants. Adequate time is required to establish a number of interacting functional groups. The choice of appropriate time scales must be considered in the selection of both study duration and sampling frequency. Whatever the system used, duration of experiments should be sufficient to identify both direct and indirect effects on populations and communities. The choice of the experimental design should be based on the objectives of the study rather than on theoretical considerations. In addition to classical parametric statistical methods, nonparametric approaches and multivariate analysis may significantly improve data processing. Realism, representativity, and replicability of mesocosms are critical for evaluating their usefulness in both risk and impact assessment procedures. Each natural ecosystem is unique because its structure and function mainly depend on local factors. Therefore, there is a conceptual opposition between realism and replicability when applied to mesocosms. Considering the objectives of most mesocosm studies, replicability should be preferred to realism. Replicability may be achieved, in part, by a relative simplification of the systems. Reconstituted systems do not need to exactly simulate natural conditions at all levels, but key features at both structural and functional levels should be preserved as they ensure ecological representativity. Reliability of information on ecotoxicological effects of chemicals tested in aquatic mesocosms closely depends on the representativity of biological processes or structures that are likely to be affected. Extrapolation from small experimental systems to the real world seems generally more problematic than the use of larger systems in which more complex interacti