Benjamin D. Barst

Proposed development of a mine within Alaska's Bristol Bay watershed (USA) has raised concerns about the potential impact of copper (Cu) on Pacific salmon ( Oncorhynchus spp.). We conducted 96‐h flow‐through bioassays using low‐hardness and low dissolved organic carbon water to determine the acute lethal toxicity of Cu to sockeye ( Oncorhynchus nerka ), Chinook ( Oncorhynchus tshawytscha ), and coho salmon ( Oncorhynchus kisutch ) fry. We aimed to determine Cu toxicity under field‐relevant water quality conditions and to assess three methods of calculating ambient Cu criteria: the biotic ligand model (BLM), a multiple linear regression model endorsed by the US Environmental Protection Agency, and the hardness‐based model currently used by the State of Alaska. The criteria generated by all models were below 20% lethal Cu concentrations by factors ranging from 2.2 to 54.3, indicating that all criteria would be protective against mortality. The multiple linear regression‐based criteria were the most conservative and were comparable to BLM‐based criteria. The median lethal concentrations (LC50s) for sockeye, Chinook, and coho were 35.2, 23.9, and 6.3 µg Cu/L, respectively. We also used the BLM to predict LC50s for each species. Model predictions differed from empirical LC50s by factors of 0.7 for sockeye and Chinook salmon, and 1.1 for coho salmon. These differences fell within the acceptable range of ±2, indicating the model's accuracy. We calculated critical lethal Cu accumulation values for each species to account for differing water chemistry in each bioassay; the present study revealed that coho salmon were most sensitive to Cu, followed by sockeye and Chinook salmon. Our findings underscore the importance of considering site‐ and species‐specific factors when modeling Cu toxicity. The empirical data we present may enhance Cu risk assessments for Pacific salmon. Environ Toxicol Chem 2023;42:2440–2452. © 2023 SETAC