Abstract
BACKGROUND: The EU member countries are currently implementing the Water Framework Directive to promote better water quality and overview of their waters. The directive recommends the usage of bioavailability tools, such as biotic ligand models (BLM), for setting environmental quality standards (EQS) for metals. These models are mainly calibrated towards a water chemistry found in the south central parts of Europe. However, freshwater chemistry in Scandinavia often has higher levels of DOC (dissolved organic carbon), Fe and Al combined with low pH compared to the central parts of Europe. In this study, copper (Cu) toxicities derived by two different BLM software were compared to bioassay-derived toxicity for Pseudokirchneriella subcapitata, Daphnia magna and D. pulex in four Swedish soft water lakes. RESULTS: A significant under- and over prediction between measured and BLM calculated toxicity was found; for P. subcapitata in three of the four lakes and for the daphnids in two of the four lakes. The bioassay toxicity showed the strongest relationship with Fe concentrations and DOC. Furthermore, DOC was the best predictor of BLM results, manifested as positive relationships with calculated LC(50) and NOEC for P. subcapitata and D. magna, respectively. CONCLUSION: Results from this study indicate that the two investigated BLM softwares have difficulties calculating Cu toxicity, foremost concerning the algae. The analyses made suggest that there are different chemical properties affecting the calculated toxicity as compared to the measured toxicity. We recommend that tests including Al, Fe and DOC properties as BLM input parameters should be conducted. This to observe if a better consensus between calculated and measured toxicity can be established.