Effect of sampling method on contaminant measurement in pore-water and surface water at two uranium operations: can method affect conclusions?

This paper describes a comparison of two methods of sediment pore-water sampling and two methods of surface water sampling that were used in a broader investigation of cause(s) of adverse effects on benthic invertebrate communities at two Saskatchewan uranium operations (Key Lake and Rabbit Lake). Variables measured and compared included pH, ammonia, DOC, and trace metals. The two types of sediment pore-water samples that were compared are centrifuged and 0.45-microm filtered sediment core samples vs. 0.2-microm dialysis (peeper) samples. The two types of surface water samples that were compared are 53-microm filtered Van Dorn horizontal beta samples vs. 0.2-microm dialysis (peeper) samples. Results showed that 62% of the sediment core pore water values were higher than the corresponding peeper pore-water measurements, and that 63% of the Van Dorn surface water measurements were lower than corresponding peeper surface water measurements. Furthermore, only 24% and 14% of surface water and pore-water measurements, respectively, fell within +/-10% range of one another; 73% and 50%, respectively, fell within +/-50%. Although somewhat confounded by differences in filtering method, the observed differences are believed to primarily be related to small, vertical differences in the environment sampled. Despite observed differences in concentrations of toxicologically relevant variables generated by the different sampling methods, the weight of evidence (WOE) conclusions drawn from each set of exposure data on the possible cause(s) of in situ toxicity to Hyalella azteca from a related study were the same at each uranium operation. However, this concurrence was largely due to other dominant lines of evidence. The WOE conclusions at Key Lake were dominated by the toxicity response of H. azteca in relation to exposure chemistry, where as the WOE conclusions at Rabbit Lake were informed by exposure chemistry, the toxicological response of H. azteca, and whole-body contaminant concentrations in the test organisms. Had these multiple lines of evidence not been available, differences in exposure chemistry generated by the different sampling methods could have substantially influenced the identification of potential causes of in situ toxicity.