Microsoft Word 4751 doc Evaluation of Water Quality Translators for Mercury Regulatory Analysis and Scientific Affairs PUBLICATION 4751 DECEMBER 2005 Evaluation of Water Quality Translators for Mercur[.]
Methods for Evaluating Total Mercury in Fish Tissue and Water
Method 1631
EPA Method 1631 is utilized for determining total mercury in both filtered and unfiltered water The method involves several steps: first, mercury species are oxidized to Hg(II) using bromine monochloride (BrCl); second, mercury is reduced to elemental mercury (Hg(0)) with hydroxylamine hydrochloride (NH2OH·HCl) and stannous chloride (SnCl2); third, mercury is purged onto gold traps for pre-concentration, which are then heated to release the elemental mercury for detection via cold-vapor atomic fluorescence spectroscopy (CVAFS) This method has a detection limit of 0.2 ng/L and is performance-based, with criteria data established before and during analysis Precision is assessed through replicate analysis of actual samples, maintaining a limit of 21% relative standard deviation (RSD), while accuracy is verified using National Institute of Standards and Technology (NIST) standards and ongoing precision and recovery (OPR) samples.
77 – 123% and MS/MSD samples between 71 – 125% Many laboratories perform this analysis; costs range from
$55 – 200 per sample with a four week reporting time.
Method 7471B/SW-846
EPA Method 7471B/SW-846 is applicable for solid, semi-solid, and aqueous samples, with the analysis portion covered by EPA Method 245.5 This method involves digesting samples with potassium permanganate and acid to convert all mercury forms into a water-soluble Hg(II) state Detection is achieved through cold-vapor atomic adsorption spectroscopy (CVAAS), where stannous chloride converts mercury to elemental form, which is then purged with nitrogen to an atomic adsorption detector The method has a typical detection limit of 0.2 µg/L and is performance-based, with precision assessed through replicate analyses Accuracy is verified using NIST standards, with ongoing precision and recovery (OPR) samples required to be within 90-110% and matrix spike duplicates (MS/MSD) between 85-115% Analysis costs range from $51 to $200 per sample, with a reporting time of four weeks.
Methods for Evaluating Methylmercury in Fish Tissue and Water
Method 1630
EPA Method 1630 is the primary technique for measuring methylmercury in fish tissue and water samples, although it has not yet been officially adopted The method involves three key steps: first, distillation to extract methylmercury from the sample; second, ethylation to convert it into methylethylmercury, which is then collected on a graphitic carbon adsorbent trap; and third, thermal desorption through a gas chromatography column to separate the mercury species before converting it to elemental mercury in a decomposition furnace The detection of elemental mercury is performed using Cold Vapor Atomic Fluorescence Spectroscopy (CVAFS) Method 1630 has a method detection limit (MDL) of 0.02 ng/L, with the capability to detect levels as low as 0.009 ng/L.
Appendix A outlines the method for determining dimethylmercury, which differs from the original Method 1630 by omitting the distillation step and purging the entire sample onto a graphitic carbon absorbent trap Elemental mercury is separated from dimethylmercury using gas chromatography, allowing for the detection of two distinct mercury peaks through CVAFS The method provides a minimum detection limit (MDL) for dimethylmercury.
The performance-based method operates at a concentration of 2 pg/L, with criteria data established prior to and during the analysis Precision is assessed through replicate analysis of actual samples, maintaining a limit of 31% relative standard deviation (RSD) Accuracy is verified using NIST standards for ongoing performance verification (OPR) and method validation (MS/MSD), with OPR samples required to fall within 67% to 133% and MS/MSD samples adhering to specified accuracy ranges.
Several studies have referenced the Appendix A method, which may eventually be utilized for validation, although it is currently still in draft form Unlike EPA Method 1630, the method for analyzing dimethylmercury does not provide specific guidelines for testing solid samples, such as fish tissue.
UW-Madison SOP
The UW-Madison SOP, established at the University of Wisconsin-Madison in 1988, closely resembles Method 1630, differing primarily in its use of a larger sample volume for distillation, which lowers the method detection limit By incorporating copper sulfate during distillation to bind sulfates, this method achieves a detection limit of 30 pg/L for methylmercury Laboratories that conduct Method 1630 can also perform this analysis for around $200 per sample, with results available in approximately four weeks.
USGS Method
The USGS method closely resembles the UW-Madison SOP and Method 1630, utilizing a larger sample volume during distillation (DeWild et al 2001) Additionally, copper sulfate is incorporated into the distillation process to effectively bind sulfides This method achieves a method detection limit (MDL) of 40 pg/L for methylmercury, and the same laboratories that conduct Method 1630 are responsible for this analysis.
1630 could perform this analysis and the cost is approximately $200 per sample with a four week reporting time.
Clean Hands Sampling
EPA Method 1669 outlines the clean hands techniques for sampling and preserving water samples, as established in 1996 This method involves a two-person team: one individual acts as the "clean hands" sampler, while the other serves as the "dirty hands" sampler, responsible for operating equipment and handling potentially contaminated tasks The method provides detailed procedures for preparing blanks, duplicates, spikes, sample filtration, and preservation to ensure sample integrity.
Considerations in the Selection of an Analytical Method
When selecting a method for testing total mercury or methylmercury, it is essential to consider factors such as method detection limit, method validity, and quality control procedures, as these influence the quality and usability of the data Without stringent quality assurance (QA) criteria, comparing samples and addressing matrix interference can be challenging Conversely, overly broad QA criteria may lead to significant variability in results Implementing tighter QA criteria enhances the generation of high-quality data.
The Method Detection Limit (MDL) for mercury analysis varies based on the analytical technique employed Atomic fluorescence methods achieve a lower MDL compared to atomic adsorption techniques for total mercury determination Notably, atomic fluorescence offers a wider linear dynamic range and enhanced sensitivity over atomic adsorption.
Methylmercury analysis methods remain experimental and lack full validation by the EPA, unlike the more established total mercury methods, which have been the focus of extensive research due to higher interest from both the EPA and industry Typically, methylmercury is analyzed in fish tissue to gain insights into mercury bioaccumulation and to assess the relationship between total mercury and methylmercury levels.
Many laboratories offer competitive services for total mercury analysis using EPA Methods 1631 and 7471B, supported by robust quality assurance and quality control (QA/QC) procedures However, only a limited number of laboratories are equipped to conduct methylmercury analysis.
Most analytical laboratories do not offer routine methods for methylmercury analysis Before conducting methylmercury sample analysis, it is essential to confirm that the laboratory can produce results that comply with QA/QC standards.
This section evaluates the EPA's draft guidance on establishing site-specific translators for fish tissue criteria Typically, translating the EPA's fish tissue criterion into water column concentrations is essential for developing water quality-based effluent limits However, translation may not be necessary if elevated mercury levels in fish result mainly from atmospheric or non-point sources, allowing point source dischargers to maintain mercury levels in effluent at or below current levels.
Overview of Translations Within and Between Media
The translation of methylmercury in fish to total mercury in water can be approached as either a direct method or a multi-step process The direct method equates total mercury in fish directly to total mercury in water, while the multi-step approach accounts for the higher bioaccumulation and bioavailability of methylmercury in water Additionally, translating methylmercury in fish to total mercury in fish is often more feasible due to the ease and lower cost of measurement However, incorporating various intermediate steps oversimplifies the complex processes of mercury bioaccumulation Practically, it may be more effective to simplify the translation into a ratio between the easily measurable total mercury in fish and water.
Figure 1—Conceptual Overview of Mercury Translations
Developing mercury translators faces significant challenges due to the vast variability in site-specific mercury methylation and bioaccumulation potential This variability is highlighted by data from the U.S Geological Survey's National Water Quality Assessment (NAWQA) program, which shows a clear trend of increasing methylmercury concentrations with total mercury levels However, total mercury concentrations can correspond to methylmercury levels that vary by two orders of magnitude, leading to uncertainty in predicting fish tissue concentrations from aqueous methylmercury levels Key factors influencing mercury bioavailability and bioaccumulation include pH, dissolved organic carbon (DOC), salinity, water flow, temperature, redox potential, sulfide and sulfate levels, suspended solids, nutrient loading, fish age and size, and the prevalence of wetlands and forested land cover in the watershed.
Figure 2—Total Mercury versus Methylmercury in Stream Water Samples Collected Throughout the United States as Part of the NAWQA Program
The EPA (2000a,b) gathered data on mercury ratios across different media, as detailed in Table 1, offering insights into the central tendency and variability of these ratios Comprehensive data sets were established for freshwater systems, encompassing both lentic (lake and reservoir) and lotic (stream and river) environments However, the data for estuarine systems were found to be scarce, and no marine data were included.
Table 1—Summary of Key Mercury Ratios within and Between Media, as Compiled by EPA (2000a,b)
Geometric Mean b Range b Number of
Trophic level 4 fish methylmercury : aqueous dissolved methylmercury (BAF)
Trophic level 3 fish methylmercury : aqueous dissolved methylmercury (BAF)
Dissolved methylmercury as a fraction of total mercury in water (fd)
Estuaries exhibit bioaccumulation factors (BAFs) ranging from 0.185 to 0.195, although specific BAFs for estuarine systems were not established due to insufficient data The reported values are derived from a combination of direct measurements and conversions as outlined by the EPA (2000a,b), with the ranges reflecting the 5th and 95th percentiles Each value represents an average for a species at a specific site, and in some instances, data from multiple water bodies or species within a single study were combined Notably, the higher BAF values for estuaries may be linked to the limited data set available, as elevated salinity has been shown to reduce both methylation and mercury bioaccumulation.
The ratios summarized above were developed based on a literature review performed during the late 1990s
Recent studies have provided extensive data on freshwater, estuarine, and marine systems, with selected relevant studies listed in Appendix A While the new data for freshwater systems generally aligns with the EPA's findings from 2000, a detailed comparison was not performed in this report The significant amount of available data warrants further examination by the EPA or other organizations to enhance the understanding of bioaccumulation factors (BAF) and fraction dissolved (fd) values in estuarine and marine systems, as well as to refine the current knowledge of these parameters in freshwater systems.
The ratio of methylmercury to total mercury in fish is crucial for developing mercury translators, as indicated by the draft EPA guidance, which assumes that nearly all mercury in edible fish tissue is methylmercury (EPA 2004a) This assumption stems from Bloom's (1992) influential study, which suggested that previous lower estimates of methylmercury in fish muscle were due to analytical artifacts Identified sources of these artifacts include contamination with inorganic mercury, incomplete sample homogenization, and discrepancies in recovery rates between methylmercury and total mercury methods Subsequent research has largely supported this conclusion for higher trophic level carnivorous fish in relatively uncontaminated waters (Lasorsa and Allen-Gil 1995; Wagemann et al 1997) Although some studies have reported slightly lower methylmercury percentages in fish muscle (e.g., Kannan et al 1998; Neff 2002), the potential impact of incomplete analytical recovery of methylmercury remains a challenge to fully dismiss.
Baeyens et al (2003) present a significant counterexample, revealing that methylmercury constituted only 58% of the total mercury found in commercial fish from the Scheldt estuary in Belgium In contrast, fish of the same species from the Belgian coastal zone and the Greater North Sea predominantly contained methylmercury, aligning with expectations.
The authors suggest that the Scheldt fish may consume food at a lower trophic level due to significant untreated sewage presence (Baeyens et al 2003) Although this finding is surprising, the uncertainty linked to assuming 100% methylmercury in edible fish tissue is relatively low compared to other uncertainties involved in translating criteria between fish and water.
Mercury in fish eaten by humans is primarily in the form of methylmercury; however, this is not always true for fish that wildlife consume Unlike humans, wildlife often eat whole fish and generally prefer smaller species.
Methylmercury represents a significant portion of total mercury found in lower trophic level fish, particularly in their muscle tissue (Bloom 1992; EPA 2000a) Research by the EPA indicates that piscivorous wildlife may be more vulnerable to mercury exposure than humans (EPA 1995) Therefore, if water quality standards are established to protect wildlife, it is essential to further assess the ratio of methylmercury to total mercury in whole fish.
EPA Proposed Translation Methods
Bioaccumulation Models for Mercury
The EPA (2004a) does not endorse specific bioaccumulation models for mercury, favoring modeling approaches over default translators Despite extensive research, accurate, nationally applicable models for predicting mercury bioaccumulation remain elusive Existing models for hydrophobic organic chemicals are not suitable for mercury, although modifications have been proposed (Toose and Mackay 2004) The Mercury Cycling Model (Tetra Tech 1999) does not predict bioaccumulation but uses bioaccumulation factors (BAF) as inputs Hope (2003) created a probabilistic food web model for mercury related to TMDL development in Oregon's Willamette River, relying on site-specific empirical data for critical components like the methyl- to total mercury ratio and BAFs Consequently, this model is not transferable to other waterways without extensive data collection and assumes independence of site-specific factors from total mercury loading, which may not hold true (Marvin-DiPasquale et al 2000; Schaefer et al 2004) Therefore, without further validation, this model cannot replace ongoing direct sampling of fish tissue.
The draft EPA guidance proposes using regression models that include variables like pH, dissolved organic carbon (DOC), and fish age as a viable alternative to traditional process-based or mechanistic mathematical models for translating criteria This approach aligns with the EPA's Methodology for Deriving Ambient Water Quality Criteria aimed at protecting aquatic ecosystems.
Models predicting bioaccumulation should focus on the most biologically relevant chemical forms, such as dissolved methylmercury, and facilitate the translation between bioavailable forms and total water concentrations Although recent regression models have not achieved high accuracy, they are likely more reliable than national default bioaccumulation factors (BAFs) Brumbaugh et al (2001) developed a model using NAWQA data to predict total mercury in fish based on unfiltered methylmercury concentrations in water, identifying pH, percent wetlands in the watershed, and acid volatile sulfide (AVS) in sediment as key parameters However, this model only accounted for 45% of the variability in fish tissue concentrations Ongoing research through NAWQA and other initiatives may enhance the modeling of mercury bioaccumulation in stream ecosystems.
Regression models developed using regional data sets from lake sampling programs have provided greater accuracy than the NAWQA example cited above Several examples are discussed below
Moore et al (2003) identified that pH, dissolved organic carbon (DOC), and methylmercury levels in water accounted for 80% of the variability in mercury bioaccumulation across 38 Canadian lakes However, this model only explained 32% of the variation when applied to NAWQA stream data, highlighting the limited applicability of the findings due to differences between lake and stream ecosystems and regional geological factors Additionally, mercury methylation in streams exhibits greater temporal variability compared to lakes Kelly et al (1995) further demonstrated that significant correlations between total mercury and methylmercury in water were observed only in lakes, not in streams.
Hakanson (2000) evaluated mercury levels in fish from 39 Swedish lakes, finding that his regression model accounted for 85% of the variation in mercury concentrations based on factors such as lake morphometry, sediment mercury, pH, phosphorus, and lake depth He suggests that this model could aid in creating lake remediation strategies, but warns against its application to reservoirs or lakes in varying climatological zones.
A regional model developed by Kamman et al (2004) for lakes in New Hampshire and Vermont predicted mercury concentrations in yellow perch fillets without directly measuring mercury levels in water Instead, it utilized data on acid neutralizing capacity, dissolved organic carbon (DOC), pH, conductivity, and lake flushing rate to make binary predictions regarding mercury content With an impressive error rate of only 13%, the model demonstrated significant accuracy, although it did not fulfill the Bioaccumulation Factor (BAF) model requirements established by the EPA in 2000.
On a cautionary note, Sonesten (2003a,b) argues that multiple linear regression, the method used by Moore et al
Multiple linear regression is deemed mathematically inappropriate for assessing mercury bioaccumulation due to the intercorrelation among environmental variables, which may exaggerate the explained variation and misidentify key factors Sonesten (2003a,b) suggests using partial least squares regression; however, this method fails to generate a predictive equation suitable for mercury translator development.
Regionally applicable models for lake systems can effectively aid in translating mercury criteria, provided they are available for the specific lake in question Additionally, there are models for reservoirs that consider factors such as flooded area, reservoir age, and the ratio of catchment area to runoff However, there is a lack of accurate models for river systems, and mercury bioaccumulation modeling has not yet been explored in estuarine or marine environments Consequently, further research is necessary to develop bioaccumulation models that can be utilized for NPDES or TMDL regulations in these systems.
National Default Translators
The draft EPA guidance presents default values to facilitate the conversion of fish tissue methylmercury criteria into water quality criteria for total mercury It includes two sets of values: bioaccumulation factors (BAFs) that illustrate the correlation between dissolved methylmercury in water and methylmercury in fish, and fraction dissolved (fd) values that depict the relationship between total mercury and dissolved methylmercury in water Additionally, the guidance discusses partition coefficients (KD values) that consider the impact of total suspended solids (TSS) on dissolved mercury concentrations, although specific default values for these coefficients are not provided.
The default Bioaccumulation Factors (BAFs) are based on central tendency values derived from the data in Table 1 Distinct BAFs are available for trophic levels 3, such as yellow perch and bluegill, and level 4, including bass and walleye However, the national default BAFs do not differentiate between lentic, lotic, and estuarine systems.
Separate values for lentic and lotic fd are provided, reflecting the central tendency of the available data, similar to the BAFs These BAF and fd values can be utilized to derive a total mercury water quality criterion from a methylmercury fish tissue criterion.
Fish tissue criterion Water quality criterion BAF x f d
To establish a weighted-average bioaccumulation factor (BAF), it is essential to determine the proportion of fish consumed from various trophic levels However, the EPA has not provided default values for the fraction of fish consumed across these different levels in its draft guidance.
The Middle and Lower Savannah River Mercury TMDL, although withdrawn, is still referenced by the EPA as a notable example of a mercury TMDL In this case, the fish tissue criterion was set at 0.4 mg/kg Site-specific data informed the selection of a bioaccumulation factor (BAF) of 4,000,000 and a dilution factor (fd) of 0.0353, with fish sampling conducted in the basin during late summer.
In 2000, research focused on largemouth bass, a species at trophic level 4, revealed bioaccumulation factors (BAFs) ranging from less than 1,000,000 to over 8,000,000 The BAF used for criterion derivation was based on the central tendency of data for fish measuring 315 mm, deemed most representative of the size typically consumed The fd value reflected the median data across the watershed, leading to a water quality criterion of 2.8 ng/L.
0.4 ppm Water quality criterion 4,000,000 x 0.0353 = 0.0000028 ppm
The EPA draft guidance discusses KD values, which aim to minimize uncertainty by considering mercury partitioning to total suspended solids (TSS) However, it lacks recommended default values, and the equations for KD application are inaccurately presented Specifically, Equations 6 and 7 in the draft guidance (EPA 2004a) require correction.
KD C MeHg d x TSS (Equation 6) and fd= 1
KD = partition coefficient for dissolved methylmercury to particulate mercury (L/mg);
CHg,p = concentration of particulate total mercury (mg/L);
CMeHg,d = concentration of dissolved methylmercury (mg/L); fd = fraction of total mercury present as dissolved methylmercury (unitless); and
TSS = total suspended solids (mg/L)
The term KD has various definitions based on the specific form of mercury being examined The KD discussed here is known as a "pseudo" KD, as it connects the total and dissolved concentrations of different mercury forms Utilizing this "pseudo" KD method, the EPA (2000b) established default log KD values of 6.83 for lentic systems and 6.44 for lotic systems.
The KD approach aims to reduce uncertainty by considering TSS concentrations; however, KD values are not constant and fluctuate with TSS levels due to mercury partitioning to colloids Research indicates that factors such as iron concentrations and the ratio of drainage area to lake area also influence KD values Similarly, the fd approach exhibits significant variability, being inversely related to total mercury concentrations due to concentration-dependent demethylation processes Seasonal variability further complicates this relationship The EPA has not indicated a preference for either the fd or KD approach, suggesting that the choice should rely on the strength of the data available To enhance the usability of guidance, the EPA should assess the data to determine if the KD approach effectively reduces variability compared to the fd approach, as no such evaluation has been performed to date.
The EPA's current methodology may introduce unnecessary uncertainty by relying on two layers of default values for dissolved methylmercury bioaccumulation factors (BAFs) and fd values This is significant as multiple steps in criteria translation can amplify uncertainty, potentially leading to a two-fold error in the final criterion Furthermore, evidence suggests that methylmercury BAFs and fd values might be inversely related, contrary to common assumptions Analysis of NAWQA sampling sites indicates that land use types with the highest mercury levels in fish and water also exhibit the lowest methylmercury BAFs Therefore, it is crucial to thoroughly investigate this relationship before applying national default translators at various sites, which could be achieved by calculating total mercury BAFs for comparison with the existing two-step translation method.
The authors of the draft guidance mistakenly referenced a "short-cut" equation for back-calculation instead of the correct definition of the KD value as outlined in Equation 6 This equation, taken from the original source document (EPA 2000b, p 5), contained a typographical error, as noted in Appendix A of EPA 2000b, which contributed to confusion surrounding the application of the KD value in Equation 7.
National default values for mercury levels are often significantly inaccurate for specific sites, as the translation from mercury in fish to total mercury in water can vary greatly The existing bioaccumulation factors (BAFs) do not differentiate between lentic (still water) and lotic (flowing water) systems, leading to potential over-protection for streams and rivers while under-protecting lakes and reservoirs Acidic lakes and recently flooded reservoirs are particularly vulnerable to mercury bioaccumulation and may not be sufficiently safeguarded by these default values In contrast, estuarine and marine systems tend to experience lower mercury bioaccumulation, making the default values overly conservative in these environments The draft EPA guidance emphasizes that these national default values should be viewed with caution.
First-phase estimates of water column targets should be tailored with site-specific information when available (EPA 2004a) If national default values are utilized for site targeting in further investigations, the efficiency of these programs could be enhanced by the EPA adopting more specific default values for various aquatic systems Utilizing recent data, as referenced in Appendix A, would support the attainment of this objective.
Site-Specific Translators
The EPA recommends using site-specific translators for criteria translation, particularly for mercury, provided the study design anticipates high variability While Kelly et al (1995) questioned the feasibility of identifying these translators in lotic systems due to a lack of correlation between total and methylmercury concentrations, their findings indicated that it is still possible to differentiate sites based on average mercury methylation levels Although similar studies on within-site variability in methylmercury bioaccumulation factors (BAFs) are lacking, NAWQA monitoring studies suggest that nationwide variation among multiple sites is relatively low when accounting for fish size and species differences (Brumbaugh et al 2001), indicating that within-site variation is likely to be even lower.
To ensure accurate assessments of mercury bioaccumulation, ultra-clean sampling and analytical techniques are essential to prevent trace-level contamination It is important to measure fish length consistently, while determining fish weight and age is also beneficial Normalizing fish tissue concentrations to a constant length aids in the development of bioaccumulation factors (BAF) Individual fish should only be composited for analysis if they are nearly identical in size Additionally, collecting multiple water and fish tissue samples over time is crucial to capture seasonal and interannual variations in mercury concentrations Fish tissue concentrations reflect long-term exposure with minimal seasonal variation, while water samples only provide a temporary view of fluctuating mercury levels, particularly in flowing water systems.
2002, 2004) Therefore, water samples may be collected with greater frequency than fish tissue samples, in order to provide data for estimating average aqueous exposure levels
To develop a site-specific study design, the initial step is to establish the spatial boundaries for the application of the site-specific water quality criterion Following this, it is essential to address several key questions related to the study.
To determine the fish species targeted for analysis and their sampling locations, it is essential to identify where people catch fish and which species they consume While a formal creel survey would yield the most accurate data, a less rigorous approach can still provide valuable insights Understanding the local fish community's species composition, access points, and regional consumption preferences can help estimate local consumption patterns and select a reasonable number of target species Recent data from the EPA (2004b) on average mercury concentrations among species can ensure that the selected target species are representative of other potentially consumed species regarding bioaccumulation Additionally, considering the average size of consumed fish will guide sample collections, while site-specific fish consumption rates may be relevant for adjusting fish tissue criteria, although such adjustments are outside the scope of this report.
Habitat differences within the study area may significantly influence mercury bioaccumulation, highlighting the need for separate evaluations of various subareas For example, bioaccumulation levels can vary between tributary streams, mainstem rivers, impoundments, and wetlands Additionally, sub-basins of larger lakes may also show distinct patterns in mercury accumulation (AMEC and ENVIRON 2003).
• Is there a need or motivation to understand the site-specific mechanisms underlying any variations in bioaccumulation potential over space and time? Although the draft EPA guidance focuses on methylmercury
Site-specific bioaccumulation factors (BAFs) can be developed to correlate total mercury concentrations in water and fish tissue While focusing solely on total mercury reduces sampling costs, it may hinder the interpretation of results and the identification of mercury sources Southworth et al (2002) highlight the importance of measuring methylmercury in a post-remediation monitoring program to clarify unexpected findings However, in some cases, a comprehensive understanding may not be essential When resources are limited, a straightforward total mercury assessment can be more justifiable than relying on national default values.
The development of site-specific translators varies significantly in scope and cost across different sites The decision on the number and types of samples to collect is a collaborative judgment between risk managers and dischargers For instance, in a scenario with a homogeneous study area, risk managers may decide against developing a mechanistic understanding of bioaccumulation Conversely, in a scenario with diverse habitat types, a more detailed understanding of bioaccumulation mechanisms is necessary due to the potential for future remedial actions.
In this study, two fish species, largemouth bass and bluegill sunfish, are chosen to represent high and medium bioaccumulation of mercury, respectively, based on EPA data Three public access sampling stations are established, and a two-year study is conducted to assess interannual variability Fish tissue samples are collected biannually in spring and fall, totaling 24 samples (2 species x 3 locations x 4 events) Surface water samples are taken every other month, with results averaged over six-month periods for comparison with fish tissue data Additionally, water samples are collected upstream, adjacent to, and downstream of each fish sampling station, resulting in 108 water samples (9 locations x 12 events) All samples are analyzed for total mercury, with analytical costs detailed in Table 2.
Table 2—Estimated Laboratory Costs for Mercury Sampling—Scenario 1
Number of samples Total cost
*Will vary depending on the specific laboratory selected to perform the analysis
This cost estimate offers a general overview of the investment required to create a basic site-specific mercury translator It does not account for site-specific sample collection costs or the expenses related to analyzing quality control samples, such as duplicates and blanks.
In the second scenario, separate translator development is conducted for the river upstream, the impoundment, and the river downstream, maintaining the same number of target species and sampling frequency as the first scenario However, fish sampling stations are reduced to two per subarea, with three associated water sampling locations The analytical program is significantly expanded to capture various environmental variables crucial for understanding bioaccumulation differences This includes several surface water parameters and a limited sediment sampling component, as sediment organic carbon and sulfide content have been linked to mercury bioaccumulation Sediment samples are collected at the same locations and frequency as fish tissue samples, totaling 72 samples While fish tissue is analyzed only for total mercury, water samples are tested for dissolved methylmercury and other correlated parameters A field probe measures pH, temperature, and conductivity during surface water sampling, with approximate analytical costs detailed in Table 3 Sample collection and QC analysis costs are not included, similar to the first scenario.
Table 3—Estimated Laboratory Costs for Mercury Sampling—Scenario 2
Analysis Cost per sample* Number of samples
*Will vary depending on the specific laboratory selected to perform the analysis
A viable hybrid approach would involve analyzing a selected subset of surface water samples for the constituents listed in Table 3 This method could significantly lower analytical costs by nearly 50%, while still allowing for an understanding of the factors influencing variations in mercury bioaccumulation.
Both scenarios offer adequate data to determine total mercury bioaccumulation factors (BAFs), facilitating a direct approach to translating site-specific mercury criteria The key benefit of the second scenario is that it establishes a baseline data set for evaluating future fish tissue data This baseline is crucial for analyzing potential causes if efforts to lower mercury levels in fish do not yield consistent results.
4 Use of Translators in TMDL and Permit Limit Calculations
Section 4 outlines the process for calculating Total Maximum Daily Loads (TMDLs) and permit limits based on the methylmercury water quality criterion A key factor in determining these limits is assessing compliance with mercury concentration levels in either fish tissue or water Consequently, many issues addressed in Section 3 are relevant to TMDLs and permit limits The section also presents specific methodologies for applying the methylmercury criterion in TMDL and permit limit determinations, along with recommendations for establishing realistic mercury permit limits.
Introduction to WQBELs and TMDLs
This section offers a concise summary of the essential concepts and terminology required to comprehend the application of mercury translators in calculating permit limits, drawing from diverse sources of information.
The Clean Water Act (CWA) of 1972 introduced the National Pollutant Discharge Elimination System (NPDES) to regulate point source discharges of pollutants into surface waters NPDES permits include effluent limitations that are determined by the technology available for pollutant control.
Technology-based effluent limits (TBELs) must align with the water quality standards of the receiving water, known as water quality-based effluent limits (WQBELs) When TBELs fall short of ensuring compliance with these standards, the Clean Water Act (CWA) and National Pollutant Discharge Elimination System (NPDES) regulations mandate the implementation of stricter WQBELs to protect water quality.
WQBELs represent a comprehensive strategy aimed at safeguarding aquatic ecosystems and human health from the harmful impacts of toxic pollutants These limits are grounded in water quality standards mandated by Section 303(c) of the Clean Water Act (CWA) Legally enforceable water quality criteria, including those for methylmercury, must be integrated into a state's water quality standards to be applicable.
The Water Quality Standards (WQS) outline the designated use of water bodies and include antidegradation provisions This is crucial, as any change in the designated use may render the existing criteria inapplicable.
To assess the necessity of Water Quality-Based Effluent Limitations (WQBELs) for protecting water quality, permit writers must evaluate if the discharge has the potential to exceed water quality standards as outlined in 40 CFR §122.44(d)(1) This evaluation typically relies on effluent monitoring data, where a statistically determined worst-case concentration is used to estimate the projected in-stream concentration under critical conditions If this projected concentration surpasses the established water quality criterion, WQBELs become mandatory While reasonable potential can also be assessed without specific monitoring data, WQBELs derived in this way are more susceptible to challenges.
The permit limits for a facility are based on the wasteload allocation (WLA) for that specific point source, ensuring that effluent limitations prevent the facility from exceeding its daily mass allocation The WLA is established through water quality modeling, which can vary in complexity and may include a mixing zone allowance depending on state regulations and the pollutant type Simple steady-state models utilize a mass balance equation with constant inputs for effluent flow and concentration, while dynamic models consider multiple variables affecting pollutant transport, providing a probability distribution for receiving water concentrations Ultimately, the WLA for each pollutant is converted into maximum daily and average monthly permit limits through statistical methods.
The WLA is essential in TMDLs for allocating the allowable pollutant load to specific point sources A TMDL estimates the maximum pollutant amount a water body can handle without breaching quality standards, factoring in contributions from point and nonpoint sources, natural background, safety margins, seasonal variability, and potential future increases While the TMDL process effectively integrates various pollutant sources, managing point sources is significantly simpler than addressing nonpoint sources.
Listing Issues that Trigger TMDLs
Under Section 303(d) of the Clean Water Act (CWA), states must provide the EPA with a list of water bodies that fail to meet water quality standards, indicating they are impaired A water body is deemed impaired when it does not support one or more designated uses, such as aquatic life, recreation, public water supply, and fish consumption Notably, fish consumption is often the primary category leading to impairments due to mercury and other bioaccumulative chemicals Total Maximum Daily Loads (TMDLs) are necessary for each specific combination of impaired water and the pollutant responsible for the impairment.
When assessing impairment, states must consider all relevant data consistent with their assessment methodology (EPA 2003a) To evaluate whether fish and shellfish consumption-based water quality standards are met, states can utilize various data types, including chemical data, fish consumption advisories, shellfish growing area classifications, and bacteria criteria However, only chemical data and fish consumption advisories are relevant for mercury and other bioaccumulative chemicals The three types of chemical data include concentrations in fish tissue, water column, and sediment While most states monitor tissue concentrations directly, some also analyze water column and sediment samples Although water sample analysis is more cost-effective, tissue data accounts for fluctuations in water column concentrations over time and is essential for calculating human health screening values Additionally, sediment concentration analysis can yield more accurate information for chemicals metabolized in fish tissues, such as polycyclic aromatic hydrocarbons (PAHs).
States can utilize fish consumption advisories created by health or environmental agencies to determine compliance with water quality standards These advisories, while not regulations, provide essential information to the public regarding the recommended limits on consuming specific fish species from designated water bodies.
In October 2000, the EPA clarified that fish or shellfish advisories indicate impairment under the Clean Water Act (CWA), as the fishable use is not achieved, regardless of numeric water quality criteria exceedances Furthermore, the presence of these advisories typically signifies a failure to meet the narrative water quality standards (WQS) aimed at protecting public health.
Reporting Requirements provides specific criteria for determining impairment:
• The advisory is based on fish and shellfish tissue data;
• A classification below “Approved” (under the National Shellfish Sanitation Program) is based on water column and/or shellfish data;
• The data are collected from the specific segment in question; and
The risk assessment parameters, including toxicity, risk level, exposure duration, and consumption rate, must collectively be equal to or less protective than the state's water quality standards.
Water bodies with fish consumption advisories do not need to be included on the 303(d) list unless specific data for those bodies is available Statewide warnings about mercury-contaminated fish do not imply that all water bodies in the state should be listed, and stakeholders may want to contest such generalizations The EPA permits states to utilize various monitoring and modeling tools to make extrapolations from limited sampling data According to EPA guidance, human health criteria for fish consumption are considered met if the mean concentrations in tissue and water do not exceed established limits, with additional statistical guidance provided on sampling and error rates.
States present documentation of listings through "Integrated Reports," which address the requirements of CWA Sections 303(d) and 305(b) for assessing water quality When a draft Integrated Report is released, permittees should evaluate the assessment of the water body where they discharge If the report indicates any category other than "all designated uses are being met," it suggests that the state requires more data or that the water body is impaired or threatened Permittees can enhance the final decision's technical basis by commenting on the assessment methodology, results, or by submitting additional data Once a water body is classified as impaired, the process for delisting becomes challenging.
Evaluation and Comparison of Mercury TMDL Targets
Mercury contamination has led to fish consumption advisories in at least 45 states, with over 1,000 water bodies identified as impaired Various methods are employed to develop Total Maximum Daily Loads (TMDLs) for mercury, as illustrated in Appendix B These examples show that TMDL targets may focus on different concentrations, including fish tissue levels, as seen in the Mermentau-Vermilion-Teche River Basin; water column concentrations, as demonstrated in the Savannah River; or suspended sediment concentrations, as observed in San Francisco Bay.
Using fish tissue concentration as the Total Maximum Daily Load (TMDL) target is the most effective method for protecting human health, as it directly measures the desired endpoint This concentration reflects the influence of various exposure media, including water, sediment, and food, along with physical, chemical, and biological factors It serves as an integrated measure, eliminating the assumptions required for calculating water and sediment concentrations Additionally, the target tissue concentration can be informed by local fish consumption rates However, a significant drawback is the challenge of applying this target to wastewater dischargers, as NPDES permit limits must be defined in aqueous concentrations or mass loadings Deriving the appropriate water concentration from tissue levels involves complex assumptions due to mercury cycling and bioaccumulation In the case of the Mermentau and Vermilion-Teche River basins, the TMDL was based on fish tissue concentrations without necessitating conversions, and it did not include WLAs for NPDES sources or examples of mercury permit limit derivation.
The EPA Office of Water cites the withdrawn Savannah River TMDL as an example of a mercury TMDL, targeting a water column concentration of mercury to prevent unacceptable bioaccumulation in fish tissue While this approach aligns with current methods for WLAs and NPDES permits, it oversimplifies the relationship between total mercury loadings and fish concentrations, potentially leading to overly restrictive permit limits that fail to meet objectives Evidence suggests that the interactions affecting mercury concentrations are complex and variable, indicating that load reductions should not be mandated until a clearer understanding of mercury cycling and bioaccumulation is established Despite these knowledge gaps, states and the EPA are expected to continue developing TMDLs and permit limits for mercury.
The San Francisco Bay Total Maximum Daily Load (TMDL), developed by the State of California, focuses on mercury concentrations in suspended sediment as a key target It also sets numeric targets for human health, specifically a fish tissue concentration target, and for wildlife protection, with a bird egg concentration target The choice of suspended sediment concentration is due to its stability compared to water column concentrations, making it a more accurate indicator of mercury levels in the Bay Notably, the mercury present in San Francisco Bay primarily originates from "legacy mercury" linked to sediments from historical mining activities, rather than from point sources, non-point sources, or atmospheric deposition.
Allocation Approaches
Allocating loadings among point and nonpoint sources is a crucial aspect of a Total Maximum Daily Load (TMDL) One effective approach is to maintain the same relative percentage contribution from each source after necessary reductions, as demonstrated in the Mermentau Vermilion-Teche and Savannah River TMDLs Factors such as cost-effectiveness, feasibility, past commitments, and implementation likelihood are also important considerations Notably, current EPA guidance requires "reasonable assurance" that nonpoint source controls will achieve expected reductions if both point and nonpoint sources contribute to impairment In the Savannah River and Mermentau Vermilion-Teche TMDLs, planned controls on air emissions under the Clean Air Act were deemed sufficient for achieving necessary nonpoint source load reductions.
The EPA's draft Guidance for implementing the January 2001 Methylmercury Water Quality Criterion outlines three approaches for allocating loadings based on the contributions of point and nonpoint sources In cases where point source loadings are predominant, the Total Maximum Daily Load (TMDL) will mandate reductions in these loadings to meet water quality standards (WQS), as seen in mercury TMDLs for water bodies affected by mining Conversely, when point source loadings are minimal, reductions in nonpoint sources are anticipated to achieve the TMDL, exemplified by the Savannah River TMDL, which identified a necessary 44% reduction in mercury loadings by 2010 under existing air regulations Lastly, in scenarios with minor point source contributions where nonpoint source reductions alone are insufficient, such as the mercury TMDL for the Ochlockonee basin in Georgia, a significant 76% reduction is required, yet only a 31-41% reduction in air loadings is projected In this case, Waste Load Allocations (WLAs) were established for major mercury dischargers, with the remaining allocation distributed among other point sources.
The San Francisco Bay TMDL adopted a strategy similar to that of the Savannah River TMDL, basing allocations for NPDES discharges on existing loadings This TMDL took into account a wider range of mercury sources and system losses Allocations for direct atmospheric deposition, non-urban stormwater runoff, and municipal and industrial wastewater were established at current loadings To meet the TMDL requirements, reductions in loadings were necessary from urban stormwater runoff sources, two upstream watersheds, and natural bed erosion processes.
Implementation Approaches
Reasonable Potential
Under the Clean Water Act (CWA), specifically 40 CFR 122.44(d)(1), the permitting authority must assess whether a discharger causes or has the potential to cause violations of water quality standards This assessment follows state-specific procedures that consider existing controls on point and nonpoint sources, variations in chemical concentrations from permitted effluents, species sensitivity in toxicity testing, and the dilution of effluents in receiving waters where mixing is permitted.
The assessment of reasonable potential for water quality violations is a key aspect of water quality standard implementation policies by permitting authorities These authorities may utilize various methodologies for this evaluation or may choose not to conduct it at all When performed, a statistical approach is applied using existing effluent data to estimate the upper range of effluent concentrations for comparison with water quality standards (WQS) In certain limited situations, permitting authorities can predict these upper concentrations without actual effluent data, but they must justify the imposition of water quality limits in the absence of supporting data.
The reasonable potential analysis (RPA) is a conservative assessment performed by permitting authorities to establish water quality-based permit limits based on fish tissue concentrations Due to uncertainties in translators, this analysis may expand the range of regulated discharges It is essential for permittees to accurately characterize their effluent discharges using statistically robust data, minimizing interference from sampling and analytical artifacts This characterization should follow the methodologies outlined in Section 2 and avoid elevated detection levels that could misrepresent the need for water quality-based effluent limits (WQBELs) Additionally, permittees should actively participate in the RPA development process, particularly during revisions of state implementation documents or the Continuing Planning Process, when permitting authorities will explore alternative methodologies for establishing WQBELs and evaluating permitted discharges.
Historically, RPA assumed that mercury was absent in discharges when concentrations were at or below the MDL However, advancements in analytical methods now allow RPA to be performed at concentrations that align with or are even lower than water column WQS Permittees discharging into water bodies identified or potentially identified as impaired by mercury should start collecting data immediately to accurately characterize their discharges.
Permitting authorities will update the RPA procedure following the EPA's final guidance on converting fish tissue criteria to water quality criteria In the interim, they are unlikely to implement new RPA procedures; instead, they will rely on current methods to compare with existing water column-based criteria or adjust the existing RPA to adopt a more conservative approach.
NPDES Implementation Procedures
The EPA's latest draft guidance outlines three methods for implementing fish tissue-based criteria in NPDES permits The first method involves calculating Reasonable Potential Analysis (RPA) and numeric effluent limits using traditional approaches, assuming the state has derived a fish tissue-based water column concentration from a Bioaccumulation Factor (BAF) The second method allows states to use a fish tissue value as the water quality standard, which is then converted into a water column value for RPA and Water Quality-Based Effluent Limits (WQBELs), also requiring a BAF.
The state may consider adopting a fish tissue-based standard for permits, as recommended by the EPA (2004a), which could lower environmental monitoring costs and eliminate the need for site-specific bioaccumulation factors (BAFs) However, this approach presents challenges for permittees, as it relies on the assumption that reasonable potential exists when mercury is detected in effluent and fish tissue exceeds quality criteria This assumption implies a direct cause-and-effect relationship between the discharger and fish tissue contamination, which is debatable Furthermore, this method contradicts established reasonable potential assessment (RPA) principles, which state that the mere presence of a pollutant in effluent does not confirm reasonable potential The permitting authority would have flexibility in determining the relevance of fish tissue data collection, which could involve monitoring by the discharger Nonetheless, it remains uncertain whether the methylmercury found in fish can be directly linked to the discharger's activities.
The three draft approaches, not yet officially released, are summarized in the table below, highlighting their potential advantages and disadvantages.
Table 4—Proposed Approaches to Implementation of Fish-Tissue Based Criterion in Permits
Form of fish tissue- based water quality standard
Form of permit limits BAF needed? Comments
Yes, to develop the water quality standard
Likely to be applied consistently to dischargers statewide; predictable
Yes, to develop the permit limit
Could be site-specific for each permittee
No Could require tissue and ambient water column sampling by permittee Flexible but not predictable Cannot link tissue concentrations to discharge loads
According to the EPA's recommended approach, the implementation of fish tissue-based criteria in permits will start with effluent monitoring using EPA Method 1631 If mercury levels are detected, permits will prohibit increased loadings unless permitted under anti-degradation provisions, and a mercury minimization plan must be developed if mercury is used or if wastewater containing mercury is accepted Even if non-point source controls are expected to meet water quality standards, the EPA advises including a non-numeric limit in permits to prevent increased mercury mass loadings and requiring a mercury minimization plan A key action within this plan for publicly-owned treatment works (POTWs) could involve creating a public education program aimed at reducing mercury contributions from sources such as dental offices and thermometers.
The EPA advises that permitting authorities take into account the contributions of non-point sources of mercury when establishing Water Quality-Based Effluent Limitations (WQBELs) The Mercury Maps tool is recommended for identifying mercury sources within watersheds If data on non-point sources is insufficient, the EPA suggests either converting the fish tissue criterion to a water column value using traditional methods or initiating the Total Maximum Daily Load (TMDL) development process to determine WQBELs based on the TMDL.
Development of WQBELs
When translating mercury water quality standards from fish tissue concentrations to water concentrations, direct adoption into permit effluent limits is not guaranteed Permitting authorities have the discretion to create implementation procedures that integrate water quality considerations into permits This section outlines practices that permitting authorities may adopt in developing water quality-based effluent limitations, such as mixing zones, tiered approaches, and other site-specific criteria.
Mixing zones are a regulatory approach used by permitting authorities to ensure compliance with water quality standards In these zones, permit holders are not required to meet water quality criteria at the point of discharge; instead, higher ambient concentrations of specific pollutants are permitted in the immediate area of effluent release However, some permitting authorities may prohibit mixing zones for certain pollutants or altogether.
Mixing zones are defined by their size and the proportion of flow from the receiving stream An initial dilution zone is created to ensure that acute aquatic criteria are met, reducing lethal effects on organisms within the zone At the boundary of the mixing zone, chronic aquatic criteria must also be satisfied The permitting authority determines the mixing zones based on the flow ratios between peak effluent discharge and critical low flow in the receiving stream Additionally, the authority may mandate validation of mixing zone assumptions through stream sampling to confirm compliance with size limitations.
Tiered approaches are essential for permitting authorities to address impairments in receiving streams through point source loading reductions and the implementation of water quality standards in permits A common example is a permit reopener, which occurs when new information, such as the establishment of a Total Maximum Daily Load (TMDL) or identification of stream impacts, becomes available Additionally, a permit compliance schedule allows permittees to conduct studies and projects aimed at improving effluent discharge quality, such as mass-balance inventories and the design of new treatment systems This schedule can also help in assessing the receiving stream for site-specific criteria development Permit holders may propose a tiered approach when pollutant impairments are identified but not fully quantified, as failing to do so risks the establishment of a Water Quality-Based Effluent Limitation (WQBEL) that cannot be relaxed later due to anti-backsliding prohibitions in the Clean Water Act (CWA).
The EPA acknowledges that the nationwide ambient water quality criteria may not be suitable for all locations, as they can be either overly protective or insufficiently protective based on specific site conditions Consequently, states have the authority to establish site-specific water quality criteria, supported by tools developed by the EPA However, the existing tools, including the Recalculation Procedure, the Water Effect Ratio Procedure, and the Resident Species Procedure, are not applicable to the methylmercury criterion, as they focus on aquatic life effects rather than human health impacts due to bioaccumulation.
States have the flexibility to implement site-specific adjustments to criteria based on local conditions and exposure patterns For methylmercury, instead of adhering to the national value of 0.3 mg/kg, potential modifications may include changes to the relative source contribution factor and the default fish consumption rate These adjustments to the default fish consumption rate were instrumental in setting the targets for the mercury Total Maximum Daily Loads (TMDLs) outlined in Appendix B.
Other Options for Implementation
Options for implementation of a mercury TMDL, other than through traditional permit limits, are available within water bodies and on a regional basis
A phased approach for implementing a mercury Total Maximum Daily Load (TMDL) is suitable when both point and nonpoint source reductions are necessary, particularly when point sources rely on the reductions from nonpoint sources This method is termed "phased" due to the need for revisions based on new data, including evidence of reductions For mercury, such a phased TMDL may involve monitoring point source discharges with updated analytical methods to gather essential data for reasonable potential determinations.
The article discusses various approaches to managing pollution prevention programs, including monitoring and evaluating technology enhancements for plant performance The San Francisco Bay TMDL initially included these strategies instead of Waste Load Allocations (WLAs) and Water Quality-Based Effluent Limitations (WQBELs) for permitted dischargers However, following objections from EPA Region 9 regarding the absence of WLAs, California incorporated them into the Basin Plan, along with a commitment to include WQBELs in permits In contrast, EPA Region 6 did not establish WLAs for the Mermentau Vermilion-Teche TMDL Meanwhile, EPA Region 4's Savannah River TMDL presented two options for NPDES permittees: one with conservative WLAs and another deferring WLA development until after mercury minimization methods are implemented This highlights the inconsistency in mercury TMDL implementation across different EPA regions, with some requiring specific WLAs for permittees while others do not.
The EPA is advocating for watershed-based permitting and trading to tackle water quality issues on a larger scale, rather than focusing solely on individual discharges This approach includes synchronizing permit cycles within a basin and developing water quality-based effluent limits (WQBELs) for multiple facilities By facilitating water quality trading, this method allows point and nonpoint sources to exchange their allocations, leading to cost-effective overall load reductions for pollutants like phosphorus, nitrogen, and sediment However, the EPA does not generally endorse trading for persistent bioaccumulative toxic pollutants (PBTs), although it may consider limited pilot projects under specific conditions, such as ensuring that the main PBT load does not originate from point sources and that trading does not exceed health criteria while achieving significant net reductions in PBT levels.
Atmospheric deposition is the primary source of mercury in watersheds, necessitating regional management strategies to effectively reduce mercury levels in fish tissue Given that airsheds encompass larger areas than watersheds, significant reductions in mercury contamination require coordinated efforts Notable examples of such regional approaches include the EPA Region 5's draft proposal for a mercury phase-down and the initiatives put forth by the Quicksilver Caucus, a coalition of state environmental association leaders established in May 2001.
The EPA Region 5 proposal aims to phase down mercury by focusing on its reduction, remediation, and prevention across various media During this initiative, mercury Total Maximum Daily Loads (TMDLs) in Region 5 states may be postponed while alternative actions are implemented If these measures fail to show significant progress towards meeting Water Quality Standards (WQS), the EPA will require the implementation of TMDLs In July 2004, Region 5 introduced the Draft Mercury Pollutant Minimization Program Guidance, which outlines how publicly owned treatment works (POTWs) can identify, reduce, or eliminate mercury discharges as part of their variance application from mercury WQS.
The Quicksilver Caucus is advocating for a national mercury reduction strategy, arguing that creating Total Maximum Daily Loads (TMDLs) for specific water bodies is inadequate due to the transboundary nature of airborne mercury pollution This approach recognizes that mercury emissions originate from regional, national, and global sources, necessitating a more comprehensive solution rather than localized measures.
Regional strategies for mercury reduction fail to establish a direct or quantitative connection between mercury reductions and the achievement of water quality standards in specific watersheds or water bodies While these alternatives may demand less effort from states, they can impose significant costs on individual point source dischargers, often without guaranteeing compliance with standards In some instances, these strategies may necessitate mercury reductions that are far below what is required to meet established standards.
The 2002 draft implementation guidance from the EPA included both initiatives, but they were removed in the 2004 version Despite the perceived stagnation of these initiatives, states continue to express concerns regarding the resources needed to conduct complex Total Maximum Daily Loads (TMDLs), particularly in relation to the cross-border movement of pollutants.
The unique environmental properties of mercury make the issues discussed in this report largely inapplicable to most other chemicals or metals The EPA has only considered developing a tissue-based water quality criterion for selenium, the only metal besides mercury, with a revised draft issued for public comment in November 2004 While selenium does not biomagnify as significantly as mercury, there have been documented cases of fish and wildlife poisoning linked to selenium from coal-burning power plants and agricultural drainage from seleniferous soils.
The EPA's Draft Aquatic Life Water Quality Criteria for Selenium—2004 establishes a Final Acute Value (FAV) for selenite in both freshwater and saltwater, along with a freshwater FAV for selenate and a Final Chronic Value (FCV) for selenium based on whole body tissue concentration Due to insufficient data, a saltwater FAV for selenate could not be calculated Following the Guidelines for Deriving Numerical National Water Quality Criteria, the FAVs and FCVs are determined from acute or chronic studies involving a minimum of eight families for acute assessments and three taxa for chronic evaluations The FAV aims to protect 95 percent of all evaluated taxa, while the FCV is derived either from an acute to chronic ratio or by ensuring protection for a critically important species, specifically focusing on a whole body tissue concentration that safeguards against adverse effects in fish.
The freshwater final acute value (FAV) for selenite, derived from data on 28 genera, indicates that the most sensitive 5th percentile genus has a calculated value of 514.9 àg/L, with the Criteria Maximum Concentration (CMC) set at 257.5 àg/L (EPA 2004d) In saltwater, data from 18 species shows that the four most sensitive genera differ by a factor of 4.7, leading to a saltwater FAV of 253.4 àg/L and a CMC of 126.7 àg/L (EPA 2004d).
The EPA (2004d) developed a FAV for selenate using a sulfate correction, highlighting that selenate toxicity diminishes in the presence of sulfate due to competition for uptake by aquatic organisms Acute toxicity data for selenate were available for 12 invertebrate and 11 fish species, with species mean acute values ranging from 593 àg/L for Daphnia pulicaria to 1,515,616 àg/L for Nephelopsis obscura The four most sensitive genera exhibited a difference of a factor of 3.0 At a sulfate concentration of 100 mg/L, the FAV for selenate for the most sensitive 5th percentile genus was determined to be 834.4 àg/L Additionally, the CMC for selenate is expressed as e(0.5182[ln sulfate]+3.357), equating to 417.2 àg/L at a sulfate level of 100 mg/L (EPA 2004d).
Research indicates that, like mercury, water quality concentrations are not reliable indicators of selenium uptake Instead, diet plays a crucial role as the primary pathway for chronic toxicity in fish (Canton and Van Derveer).
In 2004, the EPA established a tissue-based criterion for selenium, recognizing that selenium concentration in the ovary may be a strong predictor of adverse effects However, due to limited data, whole body tissue concentrations were chosen for assessment While this criterion accounts for site-specific conditions influencing selenium uptake and toxicity, it also presents implementation challenges similar to those encountered with mercury As of now, the EPA has not provided guidance on applying the tissue-based criterion for compliance with water quality standards.
The tissue-based selenium criterion was established using data from an aquatic invertebrate, eight fish species, and a mix of Centrarchidae fish Evaluations employed logistic regression to identify a 20 percent response level or relied on scientific judgment to determine the geometric mean of no observable adverse effect concentration and low observable adverse effect concentration The EPA (2004d) identified a lowest genus mean chronic value of 9.5 µg Se/g dw whole body from Coyle et al (1993) on bluegill, which was later adjusted to 7.91 µg Se/g dw based on Lemly (1993), who noted adverse effects on bluegill at winter low temperatures (4ºC) At 20ºC, bluegill in Lemly's study accumulated only 5.74 µg Se/g dw Consequently, the FCV for selenium was set at 7.91 µg Se/g dw for winter The EPA advises that if selenium concentrations in whole body fish tissues reach 5.85 µg Se/g dw during summer or fall, additional winter sampling is recommended to ensure the FCV of 7.91 µg Se/g dw is not exceeded.