A bioavailability model for chronic silver toxicity to the freshwater alga Raphidocelis subcapitata

Silver (Ag) is widely used across industries and in society, leading to releases into fresh waters. To implement bioavailability-based environmental quality standards, robust tools are needed to account for site-specific water chemistry. While an acute biotic ligand model (BLM) exists for Ag, a chronic model is lacking. Here, we investigated the influence of key water chemistry variables-pH, calcium, magnesium, sodium, chloride, and natural organic matter (NOM) concentration-on chronic silver toxicity to Raphidocelis subcapitata, across conditions representing 90% of European rivers, and developed a chronic bioavailability model. Two NOM sources and NOM fractions (humic and fulvic) were compared. NOM exerted the dominant protective effect on toxicity. Default Ag-NOM binding parameters in WHAM7 and default assumptions (that NOM consists of 65% active fulvic acid) could not explain NOM effects on toxicity, because computed free Ag+ activities at EC10 level were not equal across NOM concentrations. Increasing the binding strength parameters improved consistency in Ag+ activities at EC10, suggesting that default parameters underestimated Ag-NOM binding. Increasing chloride did not reduce toxicity of dissolved Ag, indicating AgCl0aq complex bioavailability, resulting in incorporation of AgCl0aq as toxicologically active in the model. Toxicity decreased with decreasing pH, and magnesium enhanced toxicity. Both effects could not be explained by classic BLM mechanisms and were incorporated empirically in the model. Model performance was evaluated against the development data set and a validation data set in which magnesium and chloride varied simultaneously. Predicted EC10s deviated up to twofold from observations, indicating good model performance. This algae model may represent a first step toward incorporating chronic bioavailability models into risk assessment frameworks for silver.