Five metal mixture dose‐response models are used to predict toxicity of porewater to young sturgeon at areas of interest in the Upper Columbia River and to evaluate these models as tools for risk assessments. Dose components of metal mixture models include exposure to free metal ion activities or metal accumulation by biotic ligands or humic acid, whereas links of dose to response use logistic equations, independent joint action equations, or additive toxicity functions. Laboratory bioassay studies of single metal exposures to juvenile sturgeon, porewater collected in‐situ in the fast‐flowing Upper Columbia River, and metal mixture models are used to evaluate toxicity.The five metal mixture models are very similar in their predictions of adverse response of juvenile sturgeon and in identifying copper (Cu) as the metal responsible for the most toxic conditions. Although the modes of toxic action and EC20 values are different among the dose models, predictions of adverse response are consistent among models because all doses are tied to the same biological responses. All models indicate that 56+5% of 122 porewater samples are predicted to have <20% adverse response, 25+5% of samples are predicted to have 20‐80% adverse response, and 20+4% are predicted to have >80% adverse response of juvenile sturgeon.The approach of combining bioassay toxicity data, compositions of field porewater, and metal mixture models to predict lack of growth and survival of aquatic organisms due to metal toxicity is an important tool that can be integrated with other information (e.g., survey studies of organism populations, lifecycle and behavior characteristics, sediment geochemistry, and food sources) to assess risks to aquatic organisms in metal‐enriched ecosystems.