Appendix B:
Categories of Contaminants
Polycyclic Aromatic Hydrocarbons
PAH assessment requires initial evaluation of study objectives, specifically whether the focus is on the 16 priority pollutant PAHs vs. the suite of 34 PAHs (including 18 parent and 16 alkyl homologs (USEPA 2003d). NOAA expanded scan protocol can also be considered if additional alklyated PAHs are sought as well as site-specific requirements for lower analytical detection limits (Lauenstein and Cantillo 1998). The decision regarding which suite of PAHs to analyze has implications for the risk assessment, as well as consideration of PAH bioavailability. PAH bioavailability has been reviewed by Burgess et al. (2003); Schwarzenbach, Gschwend, and Imboden (2003); and Meador et al. (1995).
When evaluating benthic community response to PAHs in sediment, evidence has suggested that toxicity testing and community analyses data correlate better with PAH concentrations in pore water than with concentrations in bulk sediment (USEPA 2003d). This weaker correlation between bulk chemical data and both toxicity and community structure results is likely due to the strong partitioning of PAHs on sediment OC, including recalcitrant “black carbon” (Lohman, Macfarlane, and Gschwend 2005; Ghosh et al. 2003; Accardi-Dey and Gschwend 2002). The association of PAHs with sediment organic matter may result in bulk sediment PAH concentrations that demonstrate little or no chemical bioavailability in the aqueous phase. Following EqP theory, observed variations in sediment toxicity may be ascribed to the variations in pore-water PAH concentration that are due, in turn, to factors such as sediment OC content or quality that influence PAH partitioning between pore-water and sediment solid phases (Hawthorne et al. 2007, Burgess et al. 2003, USEPA 2003d, Di Toro et al. 1991). USEPA currently recommends that EqP be employed for prediction of PAH exposure concentrations, with narcosis theory recommended for determination of whether measured PAH concentrations are sufficient for causing adverse physiological effects (USEPA 2003d, Burgess 2007).
Although PAHs can accumulate within the tissue of benthic organisms, they do not generally bioaccumulate in higher trophic organisms (Eisler 1987, Erickson et al. 2008). This lack of higher trophic level bioaccumulation is due to the rapid biotransformation of PAHs in such organisms. For example, lack of PAH bioaccumulation in finfish has been attributed to the presence of mixed-function oxidases that allows for metabolism and excretion of PAHs through transformation of these lipid-soluble compounds into more-water soluble forms (Eisler 1987). In general, invertebrates lack a similarly effective system, and invertebrates, including bivalve mollusks, are therefore more apt to accumulate higher concentrations of PAHs (Law et al. 2002).
Polychlorinated Biphenyls
PCB assessment requires initial evaluation of whether the research focus includes commercial PCB mixtures (Aroclors), total PCBs, PCB homologues, or PCB congeners. Determination of the total PCB exposure is a central component of risk assessment for sites impacted by PCB contamination because the majority of available ecotoxicity benchmarks and bioaccumulation factors for PCBs are based on a total PCB concentration. Determination of total PCB concentration from the sum of PCB Aroclors includes the sum of Aroclors 1016, 1221, 1232, 1242, 1248, 1254, and 1260. Aroclors 1262 and 1268 are not typically included in a total PCB summation. Data on toxicity associated with the total PCB concentrations is mostly derived from dosing studies employing commercial Aroclor mixtures or from determination of field effects due to Aroclor-based total PCB measurements.
Recognizing that toxicity varies between congeners, a growing body of literature has documented effects on a congener-specific basis. In particular, heightened toxicity has been ascribed to dioxin-like coplanar PCBs (Van den Berg et al. 1998, Ahlborg et al. 1994). Dioxin-like coplanar PCBs are PCB congeners characterized by either zero or one chlorine atom in the substituent position closest to the biphenyl double bond. Concentrations of these dioxin-like coplanar PCB congeners are commonly expressed as the equivalent concentration of 2,3,7,8‑TCDD, the dioxin congener with the greatest apparent toxicity (Van den Berg et al. 1998). An assessment of dioxin-like coplanar PCB risks should be considered as part of the weight of evidence in a risk assessment but should not be used as the sole measure of PCB toxicity (USEPA 2002e). The total PCB concentration calculated from the sum of PCB congeners is not directly comparable to the total PCB concentration calculated as the sum of Aroclor mixtures. This difference in summed PCB concentration is due to differences in analytical and quantitation methods between these strategies (Sather et al. 2001).
The EqP theorem is directly applicable to questions regarding PCB chemical partitioning and bioavailability. For PCBs, application of EqP theory to distinct Aroclor mixtures has demonstrated, for example, that EqP-based sediment quality benchmarks increase with increasing chlorination of PCBs, due to decreasing Aroclor bioavailability (Fuchsman et al. 2006). Likewise, as for PAHs, the OC content of the sediment impacts the bioavailability of PCBs to invertebrates species (Sun and Ghosh 2007; Moermond, Zwolsman, and Koelmans 2005; Pickard et al. 2001). For higher trophic level organisms, PCBs are not readily metabolized, and the residual tissue burden of PCBs for any higher trophic level organism is a function of the organism’s diet; PCB partitioning in the water column, pore water, and sediment; and the relative balance of the organism’s ingestion and egestion rates, as well as rates of growth, respiration, and metabolism (Campfens and Mackay 1997).
Normalization of sediment PCB concentrations to sediment OC content provides a site-specific strategy for refined assessment of PCB bioavailability (USEPA 2003d). Whereas there may be little or no relationship between the concentration of chemicals in sediment and observed biological effects, correlations between chemical concentration, bioaccumulation, and/or biological effects may be observed if chemical concentrations are presented on an OC-normalized basis (for hydrophobic chemicals) or defined in terms of a pore-water concentration (for chemicals that are only moderately hydrophobic) (USEPA 2003d). This normalization of sediment data in terms of sediment carbon content is consistent with one of the principal objectives of EqP theory (i.e., the selection of appropriate biological effects concentrations) and allows assessment of the varying bioavailability of PCBs in the sedimentary environment.
Pesticides
In general, it appears that TOC is the predominant factor influencing bioavailability of nonionic pesticides in the sediment environment. Lee and Jones-Lee (2005) state that some pesticides, such as the pyrethroid-based pesticides, tend to sorb strongly to sediments. Thus, the toxicity of pyrethroid-based pesticides depends on the TOC, with sediments with higher TOC being less toxic at a given pyrethroid concentration on a dry weight basis. Gan et al. (2005) present a summary of studies on the bioavailability of pyrethroid-based pesticides associated with aquatic sediments. These studies also support TOC playing a large role in the bioavailability of some pesticides. They also find that the amount of DOC in a water sample affects the water-column toxicity of pyrethroid-based pesticides. Lee and Jones-Lee (2005) note that the results of Gan et al. (2005) for the pyrethroid-based pesticides are similar to the results of Ankley and Collyard (1994) for the organophosphate pesticides diazinon and chlorpyrifos. As with many neutral organic chemicals, particulate TOC in sediments sorbs these pesticides, resulting in reductions in bioavailability and toxicity.
Other factors, like OC quality and quantity, and water properties, like pH and temperature, are also important in governing pesticide bioavailability. Additionally, sediment-contaminant contact time (i.e., aging) is an important determinant affecting chemical bioavailability in sediments for many organic compounds (Åkerblom 2007).
Routine chemical analyses determining bulk pesticide contaminant concentrations may not reliably predict toxicity to aquatic/sediment receptors. Standardized toxicity tests have been used to more accurately measure pesticide bioavailability/toxicity. USEPA (2008b) has recently published guidance using the EqP approach to derive site-specific sediment benchmark values for many nonionic pesticides. Advancements in determining pesticide pore-water concentrations using Tenax extraction and matrix-SPME fibers are also evolving (You, Landrum, and Lydy 2006).
Volatile Organic Compounds
VOCs tend to be weakly hydrophobic chemicals, which do not to persist in sediments due to their volatility and solubility. However, VOCs may occur in some sediments because of recent or ongoing releases. The method for assessing bioavailability of more strongly hydrophobic organic chemicals is typically the EqP approach. However, the EqP approach is ineffective for assessing VOCs in sediment because the standard EqP equation does not account for the contribution of dissolved chemical to the total chemical concentration in sediment. For chemicals with low organic Koc values, such as VOCs, a modified EqP equation is available, which accounts for the dissolved fraction of total chemical concentrations in sediment (Fuchsman 2003). Results of the standard and modified EqP equations converge with increasing Koc and are essentially identical at log Koc values exceeding approximately 3.5.
Metals
The geochemical form, or speciation, of inorganic chemicals governs their fate, toxicity, mobility, and bioavailability in contaminated sediment and water. For cationic metals (e.g., cadmium, copper, lead, nickel, silver, and zinc) oxidation/reduction conditions in the sediment frequently provide a measure of potential metal bioavailability. For example, AVS present in pore water under reducing conditions can bind with cationic metals to form insoluble sulfide complexes which have limited bioavailability. For oxyanions (e.g., arsenic, chromium, selenium) limited predictive models exist for understanding bioavailability, although understanding oxidation/reduction conditions in site sediment allows for general prediction of chemical speciation, with resultant influence on chemical bioavailability.
The bioavailability of cationic metals (typically the lack of toxicity due to the metals evaluated) in sediment can be generally predicted by measuring the AVS and SEM in sediments. If the concentration of AVS is greater than the concentration of SEM in sediment on a molar basis, no toxicity due to the applicable cationic metals (cadmium, copper, lead, nickel, silver, and zinc) is expected (USEPA 2005c). AVS minus SEM (SEM – AVS) has proven to be a useful indicator of metal bioavailability and lack of toxicity to benthic organisms (Di Toro et al. 1992, Hansen et al. 1996, Di Toro 2008). Geochemically, the SEM – AVS approach predicts that under reducing conditions in sediment, the concentration and bioavailability of these six metals in pore water will be lowered due to precipitation as (or with) insoluble sulfide phases. Even under reducing conditions, however, metal sulfide phases have demonstrated bioavailability to infaunal organisms (Lee et al. 2000). Because the SEM – AVS method does not account for dietary metal uptake from sediment or other food sources, the sometimes poor correlation observed between benthic invertebrate tissue data and SEM – AVS predictions of metal bioavailability likely results from direct metal assimilation through ingestion (Lee et al. 2000).
Although SEM – AVS appears to be generally useful tool for assessing bioavailability under reducing conditions, factors controlling metal bioavailability in oxygenated sediments are less well defined. These factors include metal sorption to iron and/or manganese oxides, clay minerals, and sediment organic matter. For oxygenated sediments, the bioavailability of cations and oxyanions has been assessed via sequential extraction assays (e.g., Romaguera et al. 2008, Schaider et al. 2007). For these assays, at least one extractant is selected that either mimics physiological conditions in the digestive tract of representative organisms (as presented by USEPA 2007b, 2008b for extraction of bioavailable lead) or directly applies extracted gut fluids to sediment samples (Lawrence et al. 1999). Di Toro et al. (2005b) have also demonstrated that the BLM can be adapted to assess speciation and toxicity in sediments with low or no AVS component. In this approach, the BLM adopts the EqP approach to relate toxicity of sediment-associated metals to dissolved metal concentrations. For oxygenated systems, this approach assumes equilibrium between the critical metal concentration on the biotic ligand and the sediment organic carbon content (Di Toro et al. 2005b).
Mercury
In contrast to most cationic metals, reducing conditions provide less of a measure of bioavailability for mercury. An understanding of mercury bioavailability is typically determined by measuring methylmercury concentrations; methylmercury is a neurotoxin and the form of mercury that bioaccumulates in aquatic organisms. Mercury bioavailability can also be better understood by comparing mercury concentrations in abiotic matrices (sediment, surface water, pore water) to mercury levels in biota. Mercury bioavailability, particularly in its methylated form, may also be predicted from sediment, pore-water, or surface-water organic matter concentrations (Gorski et al. 2008, Lamberttson and Nillson 2006, Driscoll et al. 1995) although these predictions can be highly uncertain. More complex water body– or ecosystem-scale models can also be used to predict mercury bioavailability. These models typically focus on mercury body burdens and exposure risk to higher trophic level consumers, including fish species as well as piscivorous birds and wildlife (e.g., Gandhi et al. 2007, USEPA n.d. “SERAFM”).
Radionuclides
An overview of the geochemical speciation, mobility, and bioavailability of radionuclides, as well as the significance of these data for environmental impact assessments, is presented by Salbu, Lind, and Skipperud (2005). Significant research on the bioavailability of radionuclides has also been conducted in the European Union for soils (Tamponnet et al. 2008) and sediments (IAEA 2004, 2010). For marine systems, the geochemical cycling and bioavailability of radioisotopes (including 51Cr, 60Co, and 65Zn) has been recently reviewed (Livingston 2004) and includes presentation of radioisotope bioavailability to deposit feeding and filter feeding organisms. In circumpolar marine systems, bioavailability has been sparingly assessed in benthic and pelagic organisms (e.g., Nonnis et al. 2000) although research has both highlighted the role that sediment resuspension plays in the transfer of sediment-associated radionuclides to filter feeding mollusks (Borretzen and Salbu 2009) and observed that the distribution of radionuclides within the bodies of deposit-feeding mollusks was correlated with radionuclide partitioning (Hutchins et al. 1998). That is, radionuclides present in the aqueous phase were more commonly ultimately associated with mollusk shells, whereas radionuclides present in sediment/food were more commonly ultimately associated with mollusk soft tissues.
In freshwater systems, the bioavailability of radioisotopes has been assessed through experimental radioisotope additions to mesotrophic and eutrophic lakes (Bird et al. 1998) with results highlighting radionuclide partitioning between distinct tissue compartments and minimal effect of lake trophic status on radionuclide uptake by biota.
As with other inorganic analytes, bioavailability of radionuclides has been assessed through chemical extraction methods. This approach has been somewhat more commonly applied to soils (e.g., Kennedy et al. 1997) than to sediments (e.g., Lucey et al. 2007) although factors governing radionuclide lability and bioavailability function similarly in both environments. Dominant factors likely influencing radionuclide bioavailability to flora and fauna include organic matter content of the soil or sediment and the grain size distribution and magnitude of the clay-sized fraction (Vidal and Rauret 1993).
Ordnance Compounds
Ordnance-related compounds (i.e., explosives) are typically associated with military activities at munitions production sites or training ranges (Lotufo et al. 2009). The main compounds found in freshwater sediments at military sites include 2,4,6-trinitrotouene (TNT), 1,3,5-trinitrobenzene (TNB), hexahydro-1,3,5-trinitro-1,3,5-triazine (RDX), octahydro-1,3,5,7-tetranitro-1,3,5,7-tetrazocine (HMX), 2,4,6-trinitrophenylmethylnitramine (Tetryl), and their related transformation products. As with other contaminant classes, it is important to do a careful initial evaluation of any historical site data to refine study objectives and narrow the list of compounds that should be considered for comprehensive study.
Total sediment concentrations are not reliable for predicting bioavailability (and toxicity) of ordnance compounds because only that fraction dissolved in pore water is accessible to biota and, once ingested, that fraction is further exposed to degradation by digestive fluids (Lotufo et al. 2009). Because traditional chemical extractions are designed to determine the total amount of contaminant present in sediments, these methods likely overestimate that fraction accessible for organism uptake (i.e., bioavailability). Organism body burdens therefore provide a more realistic and scientifically sound basis for numerically defining bioavailability (Lotufo et al. 2009). However, measuring organism body burdens is complicated by the fact that absorbed ordnance compounds like TNT are subject to metabolic processes, yielding transformation products that favor sequestration in tissues, thereby reducing elimination efficiency (Bowen, Conder, and La Point 2006).
BCFs reflect the potential for contaminant accumulation in organism tissues, and through experiments with Mediterranean mussels, major ordnance compounds have been reported with relatively low BCFs of 1.67, 0.87, and 0.44 for TNT, RDX, and HMX, respectively, in keeping with their expected low bioaccumulative potential based on octanol-water partitioning (Rosen and Lotufo 2007). Trophic transfer potential to higher organisms was reported to be negligible for dietary TNT exposure in feeding experiments with channel catfish (Ictalurus punctatus), and TNT biotransformation products showed greater accumulation than did parent TNT (Belden et al. 2005).
Biomimetic approaches are designed to make chemically extractable sediment contaminants more reflective of bioavailable toxicant body levels than reflected through traditional solvent extraction techniques (Hermens et al. 2001). More recently, SPME devices were used to predict bioavailable TNT and its related transformation products in sediment (Conder and La Point 2005). SPME fibers coated with polymer act to bind dissolved contaminants thought to be reflective of bioavailable compounds. As an equilibrium sampling technique, SPME is both gentle and nondepletive, lending to its utility as a biomimetic (Bowen, Conder, and LaPoint 2006). SPME offers a range of polymer-coated fiber types, is durable enough for direct burial in sediment, yields low detection limits, and exhibits linear uptake relationships across toxicologically relevant sediment explosives concentrations (Conder et al. 2003). There is still work to be done to improve and refine SPME prediction of chemical concentrations in organisms, and it is important to keep in mind that these approaches reflect partitioning of chemicals, and as such, are of limited utility when other mechanisms of uptake may be active. Nonetheless, when seeking potential bioavailability measurement techniques for ordnance-related compounds in sediment, SPME could be a worthwhile option to consider.
