Comparison Of Five Methods For Measuring Sediment Toxicity Of Hydrophobic Contaminants YIPING XU, †,‡ FRANK SPURLOCK,* ,§ ZIJIAN WANG, ‡ AND JAY GAN † Department of Environmental Sciences, University of California, Riverside, California 92521, Research Center for Eco-Environmental Sciences, Chinese Academy of Sciences, Beijing, China, and California Department of Pesticide Regulation, Sacramento, California Received July 31, 2007. Revised manuscript received October 16, 2007. Accepted October 22, 2007. Sediment toxicity from hydrophobic organic compounds (HOCs) is complicated by chemical partitioning among multiple phases and sediment-specific bioavailability. In this study, we used three hydrophobic pyrethroid insecticides as test compounds and derived 10-d median lethal concentrations (LC50s) for Chironomus tentans in three different sediments. The LC50s were expressed using HOC concentrations on a bulk sediment basis ( C S ), organic carbon (OC)-normalized sediment basis ( C S-OC ), porewater basis (C PW ), dissolved organic carbon (DOC)-normalized porewater basis (C PW-DOC ), and freely dissolved porewater basis (C free ). The bulk phase C S and C PW yielded highly variable LC50s across sediment types, whereas the use of normalized concentrations C S-OC and C PW-DOC generally reduced variability due to sediment type but not that due to aging. In contrast, LC50s based on C free were essentially independent of sediment conditions. The sediment porewater samples contained approximately 20-90 mg L -1 DOC, and the C free expressed as a percentage of the total bulk pore water concentration ranged from 9 to 28% for fenpropathrin (mean ) 19%), 8 to 18% for bifenthrin (mean ) 13%), and 3 to 8% for cyfluthrin (mean ) 6%) across the different sediments. These results indicate that the use of C free reduces uncertainties caused by sediment variables such as OC properties and aging effects. Introduction Hydrophobic organic contaminants (HOCs) have a strong affinity for sediments, and sediment toxicity from HOC contamination has been the focus of numerous studies and regulatory guidelines (1–8). It is generally recognized that the bulk sediment concentration CS is a poor indicator for benthic organism exposure due to variation in contaminant bioavailability, with the result that C S -based HOC median lethal concentrations (LC50s) are variable among sediments (9–11). Di Toro et al. (10) introduced the concept of equilibrium partitioning (EqP) to account for the effect of phase distribution on bioavailability. At equilibrium, an organism’s exposure is related to the contaminant activity in any of the equilibrated phases, regardless of the con- taminant uptake route. Therefore, if sediment effect con- centrations are expressed in concentration units that are proportional to chemical activity, those effect concentrations should be generally applicable to other sediments. That explains, for example, why median lethal concentrations (LC50s) expressed in units of organic carbon (OC)-normalized sediment concentration CS-OC are often less variable across different sediments as compared to LC50s expressed in terms of C S (10). Alternative units for expressing sediment LC50s also include porewater concentration C PW (10). However, most recent studies express effect concentrations in terms of C S- OC (2–5). This preference is attributable to the generally easier analytical determination of HOCs in bulk sediment as opposed to porewater (12, 13). The use of C S-OC implicitly assumes a constant K OC for a given HOC across different sediments and conditions (10). However, various properties of sediment organic matter (e.g., black carbon content) and other factors (e.g., contaminant residence time) affect HOC sorption, so that K OC s for a single HOC may be quite variable (11, 14, 15). From a practical standpoint, this means that C S-OC -based effect concentrations determined in one sedi- ment may be only approximate for other sediments and/or conditions. For instance, Amweg et al. (5) and Maund et al. (6) observed substantial differences in Hyalella azteca and Chironomous tentans C S -based LC50s for pyrethroid com- pounds among different sediments. For four of six pyre- throids, OC normalization reduced LC50 variability in the study of Amweg et al. (5), but substantial variation still occurred in Maund et al. (6) where high OC sediments were included. These and other studies demonstrate that other factors beyond OC content influence the sediment toxicity of HOCs. Consequently other approaches for predicting sediment toxicities should be explored. For porewater-based effect concentrations, the direct use of C PW may also result in high variability, especially for strongly hydrophobic HOCs. This occurs because the freely dissolved concentration (C free ) can deviate greatly from C PW due to complexation with dissolved organic carbon (DOC) in porewater (16, 17). The U.S. Environmental Protection Agency has recognized the importance of accounting for porewater DOC in their draft methods for deriving site-specific equi- librium partitioning sediment guidelines (8). Those guidelines recommend either direct measurement of C free , or use of K DOC to estimate C free . In practice five phase concentrations may be used to describe sediment exposure and toxicity: C S , C S-OC , C PW , C PW-DOC (C PW normalized by DOC), and C free . Although finding the most suitable phase concentration has been an issue of discussion for several years (8, 10, 18), to our knowledge no study has experimentally compared all five methods for describing median lethal concentrations of highly hydro- phobic organic chemicals in sediment. Therefore, the merits and limitations of the different approaches are somewhat unclear. In this study we compared the use of the five phase concentrations for expressing Chironomus tentans LC50s in different sediments using pyrethroid insecticides as test compounds. Pyrethroids are acutely toxic to benthic inver- tebrates at sediment concentrations in the ppb range (5), and they are strongly hydrophobic with log Kow ranging from 4.53 to 7.0 (19). The overall goal was to identify the most reliable and practical approach to obtaining LC50s for HOCs that are independent of sediment conditions. * Corresponding author e-mail: fcspurlock@cdpr.ca.gov. † University of California, Riverside. ‡ Chinese Academy of Sciences. § California Department of Pesticide Regulation. Environ. Sci. Technol. 2007, 41, 8394–8399 8394 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 24, 2007 10.1021/es071911c CCC: $37.00  2007 American Chemical Society Published on Web 11/13/2007