The development of toxicity tests for freshwater pollutants and their validation in stream and pond mesocosms
The number and variety of freshwater species and bioassay procedures currently recommended in European Directives for assessing the risk posed by pollutants to freshwater ecosystems is rather limited. The aim of the three consecutive European Commission-sponsored programmes described here was to modify or develop new laboratory methods that could be used for evaluating toxicity to freshwater species and which would be suitable for incorporation into the current "notification scheme" for new substances. An essential phase of each project was a validation process in which results of laboratory tests were compared with values obtained in the field using structural and functional responses of communities. The laboratory tests were developed by Fraunhofer-Institut für Umweltchemie und Ökotoxikologie (FhG) Potsdam-Rehbrücke with algae, protozoa and a variety of in vitro cellular/subcellular preparations; by the State University of Ghent with water column invertebrates, and by Cardiff University with benthic invertebrates. The laboratory tests included in vitro procedures, conventional whole organism single species tests, simple multi-species systems and microcosms. The field validation was carried out by Shell Research Sittingbourne employing artificial streams and by FhG Schmallenberg and GSF-Forschungszentrum für Umwelt und Gesundheit, Munich with artificial ponds and included single species responses but also structural and functional community responses. The same reference chemicals: lindane, copper, 3,4-dichloroaniline and atrazine were used by all laboratories. A comparison of the most sensitive toxicity response determined for each reference chemical using the laboratory toxicity tests, with that determined by the field simulation procedures revealed that for the reference chemicals lindane, copper and atrazine, adverse effects were not detected either for single organisms or for structural and functional characteristics of the community or ecosystem, at concentrations below those identified in the laboratory toxicity tests. In the case of dichloroaniline, however, effects on the population dynamics of one species in the pond mesocosm were recorded at a concentration ×16 below that reported in the laboratory. We conclude that appropriately designed, relatively simple and inexpensive laboratory toxicity tests with a selection of test species are generally adequate, with small application factors, for predicting the environmental risk of polluting chemicals to freshwater ecosystems.