Application of multiple bioindicators to differentiate spatial and temporal variability from the effects of contaminant exposure on fish.

Bioindicators of fish health were measured from 1989 through 1994 in populations of redbreast sunfish (Lepomis auritus) at three sites within a stream receiving inputs of mixed contaminants from an industrial facility, and at two reference sites. Bioindicator responses differed for fish from contaminated and reference sites throughout the study period, but temporal trends at contaminated sites reflected improved water quality associated with pollution-control efforts. Temporal variability unrelated to contaminant exposure strongly influenced the response of bioindicators at both reference and contaminated sites, but spatial variability rarely influenced these responses. Temporal variability was less influential on slower responding indicators at higher levels of biological organization. The diverse response characteristics of indicators increased the ability to differentiate natural from anthropogenic sources of variability. Integrated bioindicator responses were compared among site-year groups by multivariate canonical analysis. The primary canonical variable, associated with differences in growth, exhibited trends consistent with the timing of pollution-control efforts. Two indicators of contaminant exposure (7-ethoxyresorufin O-deethylase activity and polychlorinated biphenyls in fish muscle) also exhibited trends consistent with the timing of pollution-control efforts, but the likelihood of possible mechanistic linkages cannot be assessed with available data.