An ecosystem modelling approach for deriving water quality criteria

Ecological effects of chemicals on ecosystems are the result of direct effects of the chemical, determined in single-species toxicity testing, and indirect effects due to ecological interactions between species. Current experimental methods to account for such interactions are expensive. Hence, mathematical models of ecosystems have been proposed as an alternative. The use of these models often requires extensive calibration, which hampers their use as a general tool in ecological effect assessments. Here we present a novel ecosystem modelling approach which assesses effects of chemicals on ecosystems by integrating single-species toxicity test results and ecological interactions, without the need for calibration on case-specific data. The methodology is validated by comparing predicted ecological effects of copper in a freshwater planktonic ecosystem with an experimental ecosystem data set. Two main effects reflected by this data set (a decrease of cladocerans and an increase of small phytoplankton) which were unpredictable from single-species toxicity test results alone, were predicted accurately by the developed model. Effects on populations which don't interact directly with other populations, were predicted equally well by single-species toxicity test results as by the ecosystem model. The small amount of required data and the high predictive capacity can make this ecosystem modelling approach an efficient tool in water quality criteria derivation for chemicals.Ecological effects of chemicals on ecosystems are the result of direct effects of the chemical, determined in single-species toxicity testing, and indirect effects due to ecological interactions between species. Current experimental methods to account for such interactions are expensive. Hence, mathematical models of ecosystems have been proposed as an alternative. The use of these models often requires extensive calibration, which hampers their use as a general tool in ecological effect assessments. Here we present a novel ecosystem modelling approach which assesses effects of chemicals on ecosystems by integrating single-species toxicity test results and ecological interactions, without the need for calibration on case-specific data. The methodology is validated by comparing predicted ecological effects of copper in a freshwater planktonic ecosystem with an experimental ecosystem data set. Two main effects reflected by this data set (a decrease of cladocerans and an increase of small phytoplankton) which were unpredictable from single-species toxicity test results alone, were predicted accurately by the developed model. Effects on populations which don't interact directly with other populations, were predicted equally well by single-species toxicity test results as by the ecosystem model. The small amount of required data and the high predictive capacity can make this ecosystem modelling approach an efficient tool in water quality criteria derivation for chemicals.A