were capable of predicting toxicity by within a factor of two, and that the modified BLM produced 17 the most accurate toxicity forecasts. However, this was generally dependent on the inclusion of 18 effluent specific water chemistry characteristics in the speciation modelling as well as optimising 19 the dissolved organic carbon 'active' fraction. Only the refined D. magna BLM predicted all 20 EC50 values by within a factor of two using default model parameters and standard model 21 inputs. The results also suggested that the biotic ligand stability constant for sodium may be a 22 poor approximation of the mechanisms governing the influence of sodium where concentrations 23 exceed the range within which the biotic ligand stability constant value had been determined.
24These findings support the use of BLMs for the establishment of site-specific water quality 25 standards in waters that contain a substantial amount of wastewater effluent, but re-enforces 26 the need for regulators to scrutinize the composition of models, their thermodynamic and biotic 27 ligand parameters, and the limitations of those parameters. 28 29