Ambient toxicity dynamics: Assessments using Ceriodaphnia dubia and fathead minnow (Pimephales promelas) larvae in short‐term tests

Abstract

AbstractSeven‐day‐static‐renewal tests with fathead minnow larvae (Pimephales promelas) and Ceriodaphnia dubia were conducted monthly and results were used to assess spatiotemporal patterns in toxicity at 15 sites in five streams that receive discharges from the Oak Ridge National Laboratory (ORNL). The stream‐water samples were also analyzed for pH, conductivity, alkalinity, hardness and total residual chlorine (TRC). Relationships between the ambient toxicity test results and the chemical data were evaluated by (a) examining correlations between the two data sets across all site‐test combinations and (b) principal components analyses (PCA). These approaches proved to be more appropriate than procedures typically used to evaluate effluent toxicity (e.g., analysis of variance followed by Dunnett's test or Fisher's exact test).On average, P. promelas growth and C. dubia fecundity differed little from site to site, but survival of the two species varied greatly from test to test at some sites. Mean survival and its site‐specific C.V. were inversely related both for C. dubia and P. promelas (r = ‐0.857 and ‐0.810, respectively), and the slope of this relationship was greater for C. dubia than it was for P. promelas. Thus, (a) mortality patterns for both test species suggested that episodic events controlled overall patterns in ambient toxicity and (b) the C. dubia test appeared to be more sensitive than the P. promelas test.PCA identified two water quality factors (axis I, associated with hardness–conductivity, and axis II, strongly associated with TRC) that accounted for 60.5 and 17.6%, respectively, of the total variance in the chemical data. Multiple regression analyses of toxicity test response variables in relation to the two PCA factors showed that C. dubia survival was linked to axis II (p = 0.001) but not to axis I (p = 0.10). C. dubia fecundity was linked significantly to both (p = 0.011 and p = 0.019 for axes I and II, respectively). Neither survival nor growth of P. promelas was linked statistically to either axis. Occurrence‐frequency distributions for TRC showed that concentrations high enough to kill either species were exceeded, on average, twice monthly at sites where survival of C. dubia and P. promelas tended to be low. Thus, TRC, which enters ORNL streams primarily as drinking water used in once‐through cooling systems, was implicated as a toxicant governing the ambient toxicity test outcomes.