partially explain why trout densities are sustained im- 

 mediately downstream of the Warm Springs Ponds. 



It is difficult in a natural environment such as the 

 Clark Fork to gauge the conditions that fish and other 

 aquatic organisms are exposed to because metals concentra- 

 tions may fluctuate greatly over short intervals of time. 

 For example, even frequent sampling such as that conducted by 

 Phillips and Hill (three times per week) may not describe 

 conditions that are present during short, intense thunder- 

 storms. In any case, various authors have documented that 

 Silver Bow Creek and Clark Fork waters are sometimes toxic 

 to some invertebrates and early life stages of fish. Toxic 

 responses have been observed when metals concentrations were 

 as low as 20-50 ug Cu/1 and 30-80 ug Zn/1 and when water 

 hardnesses ranged from approximately 100-200 mg/1 (as CaC03) . 

 At water hardnesses of 100 and 200 mg/1, federal criteria 

 documents recommend that metals concentrations should not 

 exceed 12 and 21 ug Cu/1 and 37 and 66 ug Zn/1 to prevent 

 occurrence of chronic toxicity. Toxicity information for 

 the Clark Fork indicates that these criteria are not overly 

 protective and at times may provide only a very narrow 

 margin of safety. 



Some preliminary metal speciation work has been done by 

 researchers at the University of Montana. They hypothesized 

 that because trout populations in the upper Clark Fork 

 decline downstream from the sources of metals near Butte and 

 Anaconda, water chemistry changes might result in a greater 

 prevalence of toxic metal species downstream. To test this 

 hypothesis, the researchers applied several approaches to 

 modelling metal speciation to water chemistry data from the 

 upper Clark Fork (Caciari and Watson, in review) . These 

 approaches ranged from a simple model developed by the EPA's 

 Western Fish Toxicology Station (WFTS) at Corvallis, Oregon, 

 which predicts the percentage of free copper from alkalinity 

 and pH, to a complex multiequation equilibrium model 

 (MINTEQ) . 



The different modeling approaches predicted a similar 

 amount of free copper, but the MINTEQ model predicted much 

 higher levels of copper hydroxide and lower amounts of copper 

 carbonate than the other approaches. None of the approaches 

 showed any substantial downstream trend in metal speciation, 

 largely because there was no substantial downstream trend in 

 pH or alkalinity. Seasonal trends cannot be discerned based 

 on one year of data. Because the simple WFTS model is in 

 good agreement with more complex models, it should be applied 

 to the Clark Fork's past and future data sets to determine if 

 seasonal or longer-term trends emerge. 



3-117 



