Richard A. Erickson, John M. Brausch, Kathryn A. Brausch, Jennifer M. Cole, Adcharee Karnjanapiboonwong, Rodica Gelca, Stephen B. Cox, Céline A. J. Godard-Codding, and Jonathan D. Maul. Texas Tech University
Background/Question/Methods Organic contaminants are ubiquitous and consistent stressors that influence most aquatic ecological relationships and processes. Many ecologists have considered contaminants in their studies, indicating an increased awareness of the importance of these stressors on basic ecological processes. Thus, it is generally acknowledged that contaminants can affect ecological processes at multiple levels of biological organization. During a preliminary review of the ecological literature, we observed several commonly used design and study approaches that may potentially affect the accurate assessment of contaminant impacts. Many of these were related to exposure frequencies and behavior of the chemical studied (i.e., fate). Our objective was to review ecotoxicological studies published in ecology-based journals, and characterize approaches and designs used to study the impacts of contaminants. Our goal was to share this information with the ecological community and provide suggestions and tools that should lead to improved assessment of organic contaminants on freshwater ecological processes. We used ISI Web of Knowledge to search the top 50 ecology journals (ranked by journal impact factor) for articles that evaluated the effects of organic contaminants on freshwater vertebrate and invertebrate organisms.
Results/Conclusions Overall, most studies were well designed. For example, discussion of contaminant bioavailability and relevant exposure concentrations were found in about 66% of studies and analytical chemistry techniques were frequently used to quantify exposure concentrations. However, some experimental design issues were detected that could be improved. For example, environmentally realistic exposure frequencies were not used in about 50% of experiments. Furthermore, point estimates (e.g., EC50s), while often mentioned, were rarely modeled even when experimental designs may have allowed for their estimation.