v>EPA
EPA/600/R-15/289 | October 2017 | www.epa.gov/research
United States
Environmental Protection
Agency
Developing Sediment
Remediation Goals at
Superfund Sites Based on Pore
Water for the Protection of
Benthic Organisms from Direct
Toxicity to Non-ionic Organic
Contaminants
Office of Research and Development
National Human and Environmental Effects Research Laboratory
-------
Developing Sediment Remediation Goals at Superfund Sites
Based on Pore Water for the
Protection of Benthic Organisms from Direct Toxicity
to Non-ionic Organic Contaminants
Lawrence P Burkhard and David R. Mount
National Health and Environmental Effects Research Laboratory
Mid-Continent Ecology Division
Duluth, MN
Robert M. Burgess
National Health and Environmental Effects Research Laboratory
Atlantic Ecology Division
Narragansett, Rl
October 2017
1
-------
Table of Contents
Executive Summary 4
Acknowledgments 5
Glossary 6
Section 1 8
Introduction 8
1.1 Background 8
1.2 Purpose and Scope 12
Section 2 14
Measuring the Freely Dissolved Concentrations of Nonionic Organic Chemicals in Sediment Pore Water
14
2.1 Measurement of freely dissolved chemical concentrations (Cfree) in sediment pore water using
passive sampling 14
2.1.1 Passive Sampler Fouling 15
2.1.2 Passive Samplers: In-situ and Ex-situ Measurement 15
2.2 Measurement of chemical concentrations in sediment pore water using solid phase
microextraction: ATSM Method D7363-13 and EPA Method 8272 16
Section 3 18
Establishing Adverse Effects Concentrations in Sediment Pore Water for Benthic Organisms 18
3.1 Use of aquatic life criteria as an effect benchmark 18
3.2 Sensitivities of benthic and pelagic organisms 18
3.3 Derivation of EPA's AWQC FCVs 27
3.4 Sensitivities of toxicity test organisms in relation to EPA's AQWC FCVs 27
3.5 Measuring water only toxicity value for toxicant(s) 28
Section 4 29
Implementation of the pore water the Superfund Remedial Investigation/Feasibility Study (RI/FS)
Process 29
4.1 Characterization of the Nature and Extent of Contamination 29
4.2 Ecological Risk Assessment 33
4.3 Pore Water Remedial Goal Development (PWRG) 35
4.3.1 Pore Water Remedial Goal Development using Cfree Values 36
4.3.2 Remedial Goal Development by Conversion of Cfree values to concentrations in the bulk
sediment (CS;pwrg) 36
4.3.2.1 Derivation of CS;pwrg values for a sediment with one primary contaminant -dieldrin
example 37
2
-------
4.3.2.2 Derivation of Cs;pwrg values for a sediment with a PAH mixture as the primary COC 41
4.3.2.3 Derivation of Cs;pwrg values for a sediment with other contaminant mixtures 46
4.4 Suggested Methodology for Using Passive Sampler Measurements to Develop PWRGs 47
4.4.1 ESB Screening Approach 47
Section 5 50
Use of passive samplers, toxicity testing results, and pore water RGs 50
5.1 Approaches for aligning pore water RGs and sediment toxicity testing results 50
5.1.1 Exposure Response 50
5.1.2 Exposure response curves observed at sediment sites 51
5.1.3 Approaches for aligning sediment toxicity results with pore water RGs 57
5.2 Method uncertainties and confounding factors 58
Section 6 61
Appendix A: Derivation of 10- and 28-day PAH effect concentrations for Hyalella azteca for purposes of
evaluating exposure-response in sediment tests 61
Section 7 67
References 67
3
-------
Executive Summary
Evaluation of the risk posed to benthic organisms from contaminated sediments has been a long-
standing technical challenge. This document contains a methodology for developing and applying pore
water remediation goals (RGs) for nonionic organic pollutants (contaminants) in sediments for the
protection of benthic organisms. The document provides the technical approach and basis for using the
final chronic values (FCVs) from EPA's Ambient Water Quality Criteria (AWQC) for the protection of
aquatic life or secondary chronic values (SCVs) derived using EPA's Great Lakes Water Quality Initiative
(GLI) methodology to set the pore water RGs for contaminants in sediments, although other water
column values may be appropriate at any specific site or situation. Concentrations of the contaminants
in the sediment pore water are measured using passive sampling. The passive sampling measurements
directly incorporate bioavailability of the chemicals at the site into the development of site-specific
remediation goals for sediment. This document also discusses how to evaluate the consistency between
passive sampling measurements and sediment toxicity testing results. When these data are consistent,
one can be reasonably assured that the causes of toxicity to benthic organisms in the sediment have
been correctly identified and that the developed pore water RGs for the contaminants will be protective
of the benthic organisms at the site.
4
-------
Acknowledgements
The authors thank the following individuals for their detailed and through reviews of the document.
Internal Reviewers
Stephen J. Ells
Marc S. Greenberg
Karl E. Gustavson
Matthew K. Lambert
Mark D. Sprenger
US-EPA, OLEM, OSRTI, ARD, Arlington, VA
US-EPA, OLEM, OSRTI, ERT, Edison, NJ
US-EPA, OLEM, OSRTI, ARD, Arlington, VA
US-EPA, OLEM, OSRTI, ARD, Arlington, VA
US-EPA, OLEM, OSRTI, ERT, Edison, NJ
External Reviewers
Todd S. Bridges
Susan Kane-Driscoll
Upal Ghosh
Tom F. Parkerton
Jeffery A. Steevens
U.S. Army Engineer Research and Development Center (ERDC),
Vicksburg, MS
Exponent, Inc., Maynard, MA
University of Maryland, Baltimore County, MD
ExxonMobil Biomedical Sciences, Inc., Houston, TX
USGS Columbia Environmental Research Center, Columbia, MO
5
-------
Glossary
ACR Acute to chronic toxicity ratio
AWQC US-EPA Ambient Water Quality Criteria for the protection of aquatic life
Cpoiymer Concentration of chemical in polymer phase (ng/kg (dw))
Cs Concentration of chemical in sediment on a bulk basis (ng/kg (dw))
CS;pwrg Pore water RG expressed as concentration in bulk sediment (ng/kg (dw))
Csocpwrg Pore water RG expressed as concentration in sediment on an organic carbon basis (M-g/kg
(organic-carbon))
Csoc Concentration of chemical in sediment on an organic carbon basis (ng/kg (OC))
Cfree Concentration of freely dissolved chemical in pore water (|ig/L)
Cfree:PWRG Pore water RG expressed as freely dissolved concentration in water (|ig/L)
COC Contaminant of concern
CSM Conceptual site model
DQO Data Quality Objective
DOC Dissolved Organic Carbon content of water (mg/L)
EC50 Effect concentration of the toxicant that gives half-maximal response
EqP Equilibrium Partitioning
ERL Effects Range-Low
ERM Effects Range-Medium
ESB Equilibrium partitioning sediment benchmark
foe Fraction of carbon that is organic in a sediment (kg organic carbon/kg dry weight)
focss Site-specific f0c
FAV Final acute value
FCV Final chronic value
GLI Great Lakes Water Quality Initiative
K0w n-octanol/water partition coefficient for a chemical
K0c An organic carbon normalized sediment-water partition coefficient for a chemical (L/kg-
organic carbon)
Kocss Site-specific K0c
Kpoiymer Polymer/water partition coefficient for a chemical
LC50 Lethal effect concentration of the toxicant that gives half-maximal response
NAPL Nonaqueous-phase liquid
PAH Polycyclic aromatic hydrocarbon
PEC Probable effects concentration
POC Particulate organic carbon content of the water (mg/L)
PWTU Pore water toxic unit
PRC Performance reference compound
RG Remediation goal
Rl Remedial investigation
RI/FS Remedial investigation/feasibility study
SAP Sampling and analysis plan
SCV Secondary chronic values
SSD Species sensitivity distribution
6
-------
TEC Threshold effects concentration
TU Toxic unit
WOE Weight of evidence
-------
Section 1
Introduction
1.1 Background
Globally, numerous freshwater and marine ecosystems have contaminated sediments that pose
risks to the environment and/or human health. The volumes of contaminated sediments in these
ecosystems are large (e.g., in the United States quantities approaching billions of metric tons (Baker
1980; Long et al. 1996; US-EPA 2005a)), and the costs associated with managing contaminated
sediments arising from navigational dredging activities and from site remediations (i.e., dredging,
capping and post-remedy monitoring) are in the billions of dollars (US-EPA 2005a).
Because of the potential adverse ecological effects from contaminated sediments, regulatory
agencies need thresholds for determining if unacceptable risks exist for sediments from specific sites
(Mount et al. 2003; Wenning et al. 2005) and if these sites warrant cleanup. Developing contaminant
concentrations in sediment that are associated with risk thresholds has been technically challenging.
One of the first approaches developed was the sediment quality triad that combined sediment toxicity,
sediment contaminant concentrations, and benthic community data to assess the amount of risk
associated with the sediment of interest (Bay and Weisberg 2008; Chapman 1987; Chapman et al. 1987;
Long and Chapman 1985). However, the costs, in time and dollars, associated with assessing
contaminated sediment for ecological risk using approaches dependent on toxicity testing,
bioaccumulation studies, benthic community, or other data-intensive tools are very high and has fueled
the development of alternative approaches that use simpler and less expensive measures to predict
adverse effects associated with contaminated sediments.
Several approaches for developing chemical-specific sediment quality benchmarks have been
developed for classifying contaminated sediments as toxic or non-toxic. Many of the initial approaches
were developed from collections of data on the chemical concentrations in sediment and results of
laboratory sediment toxicity tests or other measures of biological effect. Examples include the Effects
Range-Low (ERL) and Effects Range-Medium (ERM) values proposed by Long and Morgan (Long and
Morgan 1991), and the Threshold Effects Concentration (TEC) and Probable Effects Concentration (PEC)
developed by McDonald and others ((MacDonald et al. 1996; MacDonald et al. 2000); see Mount et al.
(Mount et al. 2003) for more detail). Based on these approaches, guidelines were determined
empirically from large datasets by using various algorithms for evaluating concentrations of chemicals in
sediments that were or were not associated with adverse effects.
While these empirical guidelines were shown to have some ability to classify sediments into groups
with higher probability of toxicity or non-toxicity, most were based on mass-based concentrations of
sediment contaminants (e.g., M-g/kg dry weight) and did not consider additional factors that were gaining
recognition as influencing sediment toxicity. Many studies demonstrated that sediment characteristics
such as organic carbon content and sulfide (generally associated with iron) affect contaminant
bioavailability and cause widely varying toxicity among sediments with the same chemical concentration
when expressed on a mass basis. These observations drove research to develop approaches to
sediment guidelines that could account for differing contaminant bioavailability among sediments.
8
-------
For nonionic organic contaminants, early work demonstrated that sediment organic carbon
controlled the partitioning of those contaminants between sediment solids and pore water surrounding
those solids. In the late 1970s and early 1980s, Karickhoff et al. (Karickhoff et al. 1979) demonstrated
that sediment-water partitioning of hydrophobic organic contaminants was related to the
hydrophobicity of the chemical and the organic carbon content of the sediment. Predictive relationships
of the form log K0c = a + b x log K0w and K0c = b x K0w were developed where K0c is the sediment-water
partition coefficient on an organic carbon basis and K0w is the n-octanol-water partition coefficient for
the chemical of interest. Additionally, their research demonstrated that the Koc was independent of
chemical concentration and could be described as a chemical-specific equilibrium constant. This
constant, i.e., partition coefficient, is found using the equation:
Koc — {Cs/foc)/Cfree U"l)
where Cs is the concentration of chemical in the bulk sediment (ng/kg dry weight), f0c is the organic
carbon content of the sediment (kg-organic carbon/kg-dry weight), K0c is the organic carbon normalized
sediment-water partition coefficient (L/kg-dry weight), and Cfree is the freely dissolved chemical
concentration in the sediment pore water (ng/L).
Freely dissolved chemical in water is chemical held in solution by water molecules only, and is not
associated with dissolved organic carbon (DOC), particulate organic carbon (POC), or colloids in the
water phase. The freely dissolved concentrations in water can never exceed the aqueous solubility of
the chemical. For chemicals with log K0wS less than 5, the chemical's Cfree value and total concentration
of the chemical in water (determined by the extraction of bulk water phase) are nearly identical. For
chemicals with log K0wS greater than 5, the chemical's Cfree value is less than the chemical's total
concentration in the water because the chemical sorbs to DOC, colloids, and POC in the water. With
increasing K0w, the portion of the total chemical that is freely dissolved decrease significantly.
The link between partitioning of organic chemicals and sediment toxicity was demonstrated in
experiments by Adams et al. (Adams et al. 1985). In this classic study, midge larva (Chironomus dilutus,
then C. tentans) were exposed to three different sediments spiked with the pesticide Kepone. The
concentrations of Kepone in these sediments causing toxicity to midge varied by two orders of
magnitude when the pesticide concentrations in the sediment were compared on the conventional basis
of chemical mass per mass of dry sediment (Figure 1-la). However, when exposure was expressed on
the basis of Kepone concentration in the sediment pore water (chemical mass per L), the exposure-
response curves for the three sediments were very similar (Figure 1-lb). Not only were the curves
similar, but the concentration at which effects occurred in pore water was comparable to the Kepone
concentration associated with toxicity in water only exposure. This suggested that one could predict the
toxicity of a sediment by measuring (or predicting) the chemical concentration in the sediment pore
water. As discussed above, sediment organic carbon was thought to be the primary sediment phase
controlling partitioning between sediment solids and the pore water; when Adams normalized sediment
Kepone concentrations to the organic carbon content of each sediment (chemical mass per mass organic
carbon), toxicity of the three sediments were very similar, as it had been when expressed based on pore
water (Figure 1-lc).
9
-------
120 r
10D-
oO i
a
* *• - \ * '** *
¦
*
* ij'ifiC *
(?•
• *:
01 1 10 >GD
Seffiinmi Kenotw
-------
surfaces. Further, changes in redox potential can lead to the formation of new artificial particles caused
by oxidation of reduced iron. A final challenge is that concentrations of highly hydrophobic
contaminants in pore water can be very low, making analytical quantification difficult without
exceptional laboratory technique (Adams et al. 2003; Ozretich and Schults 1998; Schults et al. 1992; US-
EPA 2012a). However, as sediment assessment approaches were evolving during the 1990s,
considerable uncertainty remained as to whether the challenges of accurate isolation and analysis
would preclude reliance on pore water as a routine measurement for sediment assessment.
Rather than relying on direct analysis of pore water, focus shifted to basing guidelines on the more
easily measured bulk sediment concentrations, and predicting chemical concentration in pore water
using equilibrium partitioning relationships. EPA pursued the developing of sediment guidelines using
the physical-chemical concept of Equilibrium Partitioning (EqP) proposed by Di Toro et al. (Di Toro et al.
1991). Simply put, EqP asserts that a contaminant's bioavailability is directly proportional to its chemical
activity in sediment. EqP also asserts that a contaminant in bedded sediment is at equilibrium across all
sediment phases, and as a result, the chemical activity of the contaminant is the same in all sediment
phases. Since the freely dissolved concentration in pore water corresponds closely with chemical
activity, this rationalizes the concept that bioavailability and toxicity are proportional to concentration in
pore water as demonstrated by Adams and others. It's worth noting that despite its emphasis on
chemical activity/concentration in pore water, EqP does not assume that pore water is the only route of
exposure to organisms. Rather, the assumption is that the chemical activity (which is analogous to
fugacity which be conceptualized as "chemical pressure") is the same among all sediment
compartments (because they are in equilibrium) and therefore the intensity of organism exposure is the
same regardless of the route, i.e., via sediment ingestion, pore water, dermal contact, or any
combination of the three exposure routes.
Further analyses by Di Toro et al. (Di Toro et al. 1991) affirmed the findings of Adams (Adams et al.
1985), demonstrating that the freely dissolved concentration in pore water is not only proportional to
toxicity, but directly comparable to the concentration causing effects in water only exposures to the
same organism. Since most waters used for toxicity testing are generally low in dissolved organic carbon
and other binding phases, nonionic organic chemicals that do not exhibit extreme hydrophobicity (Log
Kow < 6) will be present in toxicity tests primarily in the freely dissolved form. Thus, it makes sense that
similar toxicity occurs in a water only exposure of the chemical and a sediment exposure with a freely
dissolved pore water concentration equaling that of the water only exposure. Thus, EqP can be used to
estimate contaminant concentrations in sediments pore water that are in equilibrium with the bulk
sediments concentrations corresponding to specific levels of toxicity (or non-toxicity) observed in water
only toxicity tests.
For many chemicals, EPA has derived water quality criteria for the protection of aquatic life, which
are chemical concentrations in water below which unacceptable effects on aquatic organisms are not
expected. Using water quality criteria as threshold values for toxicity in water, the EqP approach
translates these into bulk sediment concentrations using organic carbon normalized sediment-water
partition coefficients (K0cs) for the chemical of interest. Using this approach, EPA has developed
mechanistic based sediment quality guidelines known as Equilibrium Partitioning Sediment Benchmarks
(ESBs) for a number of common sediment contaminants, including 34 polycyclic aromatic hydrocarbons,
31 other nonionic organic chemicals, and metal mixtures (e.g., cadmium, chromium, copper, nickel, lead,
silver, and zinc) (Burgess et al. 2013; US-EPA 2003a; US-EPA 2003b; US-EPA 2003c; US-EPA 2005b; US-
EPA 2008). For the nonionic organic chemicals, the ESBs are expressed on an organic carbon normalized
concentrations in the bulk sediment (i.e., ug/g-organic carbon). For metals, the ESBs are expressed on a
11
-------
nmole/g-organic carbon basis in the bulk sediment after considering sequestration of metals by acid
volatile sulfides (AVS) and organic carbon, or on a ng/L basis when metals are measured directly in the
sediment pore water.
While the theoretical underpinnings of the ESBs approach are strong, their accuracy in application is
dependent on the robustness of their underlying assumptions. In particular, the generic formulation of
the ESBs uses a single K0c value for each chemical. This single K0c value is assumed to be appropriate for
all sediments and does not change as a function of the quantity or quality of the organic carbon in the
sediment (Burgess et al. 2000; Dewitt et al. 1992). Later research and practical experience has shown
"organic carbon" in sediments includes a variety of diagenic, petrogenic, and pyrogenic forms, and these
different forms can have different Koc values, potentially resulting in different partitioning across
various sediment types (Cornelissen et al. 2005; Hawthorne et al. 2006; Hawthorne et al. 2011; Jonker et
al. 2003). Depending on the chemical and carbon type, these differences can range from negligible to
substantial; in the particular case of PAHs, sediment-specific Koc values for a single compound have been
shown to vary as much as 100-fold. This can create substantial uncertainty in the assessment of
ecological risks posed by such sediments.
In the past decade since EPA's development of the EqP approach and resulting ESBs, much work has
been performed on developing the passive sampling technique for estimating the freely dissolved
concentrations of contaminants in the column water and sediments (Hawthorne et al. 2009; Jahnke et
al. 2012; Lydy et al. 2014; Maruya et al. 2009; Mayer et al. 2014; US-EPA 2012b). The passive sampling
technique does not require isolation of the sediment pore water from the bulk sediment but rather is
performed on the whole sediment (in the laboratory and field) or a sediment-water slurry. The
technique is nondestructive, does not change the internal partitioning of the chemical among the
sediment phases (i.e., solids, particulate, colloidal, dissolved carbon, and aqueous phases), and can be
performed on small samples of wet sediment. In this approach, an organic polymer is placed into a
sediment or sediment-water slurry, and allowed to equilibrate with the COCs. Polymers include low
density polyethylene, polyoxymethylene and polydimethylsiloxane. During the deployment time, the
contaminants diffuse from the pore water into the polymer and after their retrieval from the sediment,
the chemicals in the polymer are quantified. With the resulting data, Cfree for the chemicals of interest in
sediment pore water can be estimated with minimal artifacts, and the technique is relatively simple to
perform in the laboratory and field (Burgess et al. 2015; Fernandez et al. 2014; Gschwend et al. 2011).
The development of reliable techniques to measure chemical concentrations in pore water brings
the EqP approach full circle; rather than basing the assessment on bulk sediment concentrations and
predicting partitioning to pore water, a proxy for the chemical activity of sediment contaminants can be
measured directly via passive sampling of pore water, and Cfree concentrations can be used to predict
residues and toxicity for benthic organisms (Kraaij et al. 2002).
1.2 Purpose and Scope
In light of the improved technologies and understanding described above, EPA's Office of Superfund
Remediation and Technology Innovation requested that the Office of Research and Development
develop guidance on applying these approaches to develop pore water Remediation Goals (RGs) for the
protection of benthic organisms. Like the ESBs, pore water RGs are intended to protect organisms living
in and on the sediments (e.g., oligochaetes, annelids, amphipods, bivalves, arthropods, and other
invertebrates) from direct toxicity from sediment contaminants. This guidance is not designed to
12
-------
explicitly protect higher trophic level benthic species from effects associated with food chain
biomagnification (e.g., crab, lobster, catfish, and carp) or pelagic organisms. While the approach should
be applicable to nonionic chemicals generally, specific values are provided for polycyclic aromatic
hydrocarbons (PAHs), several pesticides, chlorobenzenes, several low molecular weight organic
compounds, and some phthalates. Although ESBs have been developed for cationic metals (Cu, Cd, Zn,
Pb, Ni, Ag), pore water RGs are not presented because passive sampling technology for these chemicals
is in a different stage of development and standardization; however, a similar conceptual approach
could be implemented using guidance contained in the ESB document for metals mixtures (US-EPA
2005b). Unless there is reason to believe that the toxicity or bioavailability would be fundamentally
different in freshwater and marine ecosystems, the guidance provided is generally applicable to both.
Applying the pore water RG approach requires two basic elements: a) a method for measuring or
inferring the freely dissolved concentration of contaminant in pore water (Cfree); and b) a toxicity
threshold chemical concentration that delineates acceptable and unacceptable exposures. These
elements are the focus of Sections 2 and 3 (respectively) of this document. Section 4 discusses how
these two measures are brought together to evaluate sediments for compliance with pore water RGs.
13
-------
Section 2
Estimating the Freely Dissolved Concentrations of Nonionic Organic
Chemicals in Sediment Pore Water
As discussed in the Introduction, centrifugation has been a common technique for isolating pore
water and measuring Cfree. With the development of the passive sampling technique for estimating Cfree
in sediment and overlying water (Hawthorne et al. 2009; Lydy et al. 2014; Maruya et al. 2009; Mayer et
al. 2014), the passive sampling technique is now the recommended approach for measuring the
concentrations of chemicals in the sediment pore water. The passive sampling technique is relatively
simple to perform in the laboratory and has lower potential for sample handling and processing artifacts
in comparison to the centrifugation technique.
2.1 Measuring Freely Dissolved Chemical Concentrations (Cfree) in Sediment Pore Water using Passive
Sampling
With the passive sampling technique, a thin sheet or fiber of an organic polymer is equilibrated with
the sediment (US-EPA 2012a; US-EPA 2012b). The target contaminant sorbs to the polymer, and after
an appropriate equilibration time (typically 28-days), the chemical achieves equilibrium between the
polymer; freely dissolved, colloidal, DOC and POC phases in the pore water; and the solids in the
sediment. With knowledge of the partition coefficient between the freely dissolved chemical and the
polymer, the freely dissolved concentration in the pore water can be determined after measurement of
the concentration of the chemical in the polymer (Cp0iVmer). In equation form, Cfree is computed:
Cfree ^Polymer / ^Polymer (2~1)
where, Kp0iVmer is the polymer-water partition coefficient for the chemical of interest. The Kp0iVmer values
are determined by equilibration studies in the laboratory, and in these studies, high purity water with
dissolved chemical is equilibrated with the passive sampler. After equilibration, both phases are
analyzed in order to compute the Kp0iVmer value. Many of these values are available in the scientific
literature for contaminants of concern like chlorinated pesticides and PAHs (US-EPA 2012a).
When a passive sampler is equilibrated with a sediment sample, equilibrium can be demonstrated
by measuring a time series of Cp0iVmer values and when these values don't change significantly over time,
equilibrium conditions have been obtained (Mayer et al. 2014). Another approach for demonstrating
equilibrium conditions is to use passive samplers with different surface to volume ratios, and when the
Cpoiymer values are the same at a single time point in the equilibration process, equilibrium has been
obtained (Mayer et al. 2014).
There will be cases where equilibrium conditions for more hydrophobic contaminants are not
attained in the experimental time frame of 28-days. Causes of non-equilibrium conditions include slow
diffusion kinetics for highly hydrophobic chemicals like dibenz[a,h]anthracene, slow desorption kinetics
from soot (e.g., black carbon phases) to the pore water, presence of oils and greases, and potentially,
biological growth on the passive sampler. To account for non-equilibrium conditions, passive sampling
is often performed using performance reference compounds (PRCs) where the PRCs are loaded into the
sampler prior to their equilibration with the sediment (Fernandez et al. 2009; Huckins et al. 2006; Reible
and Lotufo 2012). If measurements demonstrate that all of the PRCs were lost from the sampler during
14
-------
their equilibration, equilibrium conditions were obtained. If not, the loss of the PRCs enables one to
determine the extent of the equilibration of the target contaminants. Assuming the loss kinetics of the
PRCs from the polymer are similar to the uptake kinetics for the target chemicals of interest, for target
chemicals not at equilibrium, the actual freely dissolved concentrations of the target chemicals at
equilibrium can be calculated.
Passive sampling performed under actual equilibrium conditions or using PRCs to estimate
equilibrium conditions provides accurate estimates of the bioavailable (freely dissolved) chemical in the
sediment pore water. Two US-EPA documents provide details on the passive sampling approach (US-
EPA 2012a; US-EPA 2012b), and field and laboratory procedures for using passive sampling are discussed
in a US-EPA/SEDP/ESTCP (US-EPA/SERDP/ESTCP 2017) report. In addition, a series of papers from a
recent SETAC workshop titled "Guidance on Passive Sampling Methods to Improve Management of
Contaminated Sediments" published in Integrated Environmental Assessment and Management
provided further information on the passive sampling technique (Ghosh et al. 2014; Greenberg et al.
2014; Lydy et al. 2014; Mayer et al. 2014; Parkerton and Maruya 2014; Peijnenburg et al. 2014).
2.1.1 Passive Sampler Fouling
Fouling occurs when the surface of the passive sampler is coated in a biological growth,
nonaqueous-phase liquids (NAPL), or other organic material and is a well-known issue with this
methodology. PRCs can be used to correct for the effects of biological growth or the presence of
organic matter on chemical uptake by the sampler. For further information on PRCs and their use,
consult the following references (Ghosh et al. 2014; Lydy et al. 2014; US-EPA 2012a; US-EPA 2012b;
US-EPA/SERDP/ESTCP 2017).
At some sites, NAPL will be present, and samplers that come in contact with NAPL can become
fouled. PRCs cannot be used to account for the effects of NAPL fouling. If the NAPL is not properly
removed from the sampler prior to its analysis, passive sampler results will lead to overestimations
of Cfree (Heijden and Jonker 2009). Additionally, NAPL fouling may result in estimated concentrations
in water above the chemical's solubility and potentially, increased variability. As suggested by
Ghosh et al. (Ghosh et al. 2014), users should record incidents of NAPL presence and be aware of
the potential for artifacts in the resulting data.
2.1.2 Passive Samplers: In-situ and Ex-situ Measurement
Passive samplers can be deployed in the field (in-situ) and in the laboratory {ex-situ). There are
a number of reasons for performing in-situ passive sampling measurements. Primarily, the field
measurements capture all processes and conditions existing at the site that would be difficult to
replicate in the laboratory (Ghosh et al. 2014). Some of these processes and conditions include
temperature, light, bioturbation, pH, salinity, sediment resuspension, groundwater flows, organism
activity, and biodegradation of the contaminants, and all of these factors affect the target
contaminants behavior in the sediment pore water. The major challenge with in-situ passive
sampling is determining whether or not the target contaminants have achieved equilibrium with the
passive sampler. As discussed earlier, there are at least four ways to address this issue including the
use of performance reference compounds (PRCs) and temporal sampling (Lydy et al. 2014). Devices
for passive sampling in sediments in the field are readily available (e.g.,(Lydy et al. 2014; Mayer et al.
2014; US-EPA 2012b; Witt et al. 2013)). In-situ sampling requires at least two field efforts, once to
15
-------
deploy the devices and a second, to retrieve the devices. One critical issue with field sampling is
that devices can be lost, damaged, or vandalized.
There are also a number of reasons for performing ex-situ measurements. Ex-situ
measurements can be performed easily on numerous sediment samples, and ex-situ measurements
requires only one sampling trip in terms of resources (i.e., collect sediments). A principle advantage
of ex-situ sampling is that laboratory conditions can be manipulated to insure equilibrium is
achieved between the target contaminants and passive sampler resulting in greater confidence in
Ctree estimates. For example, the sediments and passive samplers can be rolled to enhance
contaminant transfer between environmental phases. One of the drawbacks of ex-situ deployments
is that the measurements are only as good as the collected sediment samples. For example,
collection of the top few centimeters of surficial sediment is difficult. Overall, the effects of sample
collection, storage, and handling in the laboratory are incorporated into the ex-situ measurements.
In addition, as discussed, unlike the in-situ deployments, ex-situ deployments do not incorporate
realistic field conditions.
Studies comparing in-situ and ex-situ measurements of contaminant concentrations in sediment
pore water are limited. Witt et al. (Witt et al. 2013) has demonstrated comparable measurements
between in-situ and ex-situ measurements for PAHs and PCBs. In addition, Fernandez et al.
(Fernandez et al. 2014) compared in-situ and ex-situ passive sampling for PCBs and DDTs at the Palos
Verdes Shelf Superfund site located off the coast of Los Angeles (CA, USA). Good levels of
agreement were observed in the calculation of DDT and PCB Cfree values when comparing ex-situ
passive samplers in rolled sediments to in-situ PRC-corrected passive samplers.
For applying the methodology in this document, in-situ and ex-situ measurements are
acceptable.
2.2 Measuring Chemical Concentrations in Sediment Pore Water using Solid Phase Microextraction,
ATSM Method D7363-13 and EPA Method 8272
Another approach to measuring freely dissolved nonionic organic chemical concentrations in
sediment pore water is ASTM Method D7363-13 (ASTM 2013) or equivalently, EPA method 8272 (US-
EPA 2007a). The method developed by Hawthorne et al. (Hawthorne et al. 2005) isolates and measures
concentrations of pore water target contaminants by absorption to a solid-phase-microextraction
(SPME) fiber. This method has been very effective for determining the concentration of several legacy
nonionic organic contaminants in contaminated sediment pore waters (Arp et al. 2011; Hawthorne et al.
2007; Hawthorne et al. 2009; Hawthorne et al. 2008) and has been adopted as US-EPA method 8272
(US-EPA 2007a) and ASTM method D7363-13 (ASTM 2013). This method is not an equilibrium-based
passive sampling method as described in 2.1 above and does not generate Cfree values. In the
Hawthorne et al. (Hawthorne et al. 2005) method, the pore water is isolated from the sediment or
sediment slurry by centrifugation and treated with alum to precipitate and remove colloidal organic
carbon. Deuterated internal standards are added to the isolated colloidal carbon-reduced pore water
and subsequently, the SPME fiber is introduced into the sample. In this application, the SPME fiber is
acting like an organic solvent in that the fiber is extracting any dissolved contaminants from the pore
water sample into the PDMS polymer coating on the fiber. The fiber is then thermally desorbed and
16
-------
analyzed for target contaminants. The dissolved concentrations are calculated based on the ratio of
analytes to corresponding internal standards. This process creates an operationally defined form of Cfree
(i.e., pore water minus colloidal and dissolved organic matter precipitated by alum).
17
-------
Section 3
Establishing Adverse Effects Concentrations in Sediment Pore Water
for Benthic Organisms
3.1 Use of Aquatic Life Criteria as an Effect Benchmark
As outlined in the introduction, implementation of the pore water RG approach requires a threshold
chemical concentration that delineates acceptable and unacceptable exposures. In the development of
EPA's ESBs, the water only effect concentration chosen was the EPA Ambient Water Quality Criterion
(AWQC) for the protection of aquatic life, and more specifically the "Final Chronic Value" (FCV). The FCV
is a derived value that is intended to estimate a concentration that would protect 95% of tested species
from chronic toxicity under long-term exposure. At contaminated sediment sites, a majority of benthic
organisms are exposed to the sediment contaminants for their entire life cycle and resultantly, chronic
exposure was selected as the appropriate time frame for exposure. In addition, the intended level of
protection of the FCV, protecting the vast majority of organisms, was deemed an appropriate protection
goal for ESBs. Pore water RGs as proposed here, are intended to provide this same level of protection,
and therefore use the FCV (or an estimate thereof) as the effect threshold.
EPA's 1985 guidelines for deriving AWQC (Stephan et al. 1985) has stringent data requirements for
developing AWQCs, and often sufficient data are not available to derive a FCV for a chemical. For some
of the common sediment contaminants that don't have AWQCs, alternative methods are available for
estimating the equivalent of a final chronic value, specifically the Great Lakes Water Quality Initiative
(GLI) methodology (US-EPA 1995; US-EPA 2008). The GLI methodology was developed from a
comprehensive distributional analysis of the relationship between the lowest available toxicity values
and FCVs derived using the 1985 AWQC guidelines. Adjustment factors were developed to account for
the uncertainties that exist when toxicity data are limited, and these factors can be applied to the
available data to provide a reasonably conservative estimate of the FCV; these estimated FCV values are
called "Secondary Chronic Values" or SCVs. FCVs and SCVs for many common sediment contaminants
are provided in Table 3-1. For reference, Table 3-1 also contains EPA's ESBs for many common
sediment contaminants.
Polycyclic aromatic hydrocarbons (PAHs) are common COCs at Superfund sediment sites, and have
several characteristics that present challenges in the development of pore water RG. First, PAHs as a
group represent a wide range of chemical structures that co-occur in the environment, and not all of
these are commonly measured in routine sediment monitoring programs. Second, depending on the
organism and the specific PAHs involved, PAHs can exert toxicity through multiple mechanisms,
including narcosis, carcinogenicity, and mutagenicity, as well as photo-enhanced toxicity (US-EPA
2003c). For benthic invertebrates, it is believed that the narcosis mechanism determines the potency of
sediment exposures to PAHs, and EPA has developed an ESB for PAH mixtures on that basis (Di Toro and
McGrath 2000; Di Toro et al. 2000; Mount et al. 2003; US-EPA 2003c). An additional feature of the
narcosis mechanism is that all PAHs contribute additively to the toxic effect, so effect concentrations in
water are based not on the single PAHs, but on the aggregate potency of all measured PAHs. To assess
the potency of individual PAHs, EPA used an approach similar to that described in the 1985 guidelines to
18
-------
derive a FCV for each individual PAH; fractional contributions of each PAH are then calculated and
summed to determine the additive potency of the mixture. Additional details on the derivation of water
column potency estimates (used here as pore water RGs) are provided in the PAH ESB document (US-
EPA 2003d). Section 4 of this document describes how pore water RG calculations for PAHs are
performed.
Table 3-1 provides pore water RG values for a variety of chemicals based on their FCV or SCV values.
Some of the chemicals in Table 3-1 that have SCV values are believed to affect benthic invertebrates
through a narcosis mechanism. Because narcotic chemicals appear to have comparatively small inter-
species differences in potency, as well as a comparatively small acute-chronic ratio, the GLI procedure
for calculating an SCV tends to be fairly conservative when applied to narcotic chemicals, particularly
those for which only limited data are available (therefore having relatively large uncertainty factors
applied). For reference, Table 3-1 also contains pore water RGs calculated based on an assumed
narcosis mechanism of action, based on the methods of DiToro et al. (Di Toro et al. 2000). These
narcosis-based pore water RGs can be used instead of the GLI SCV for chemicals expected to act through
the narcosis mechanism. However, if narcosis pore water RGs from Table 3-1 are used, it must be
remembered that all narcotics present will contribute additively to the overall potency of the chemical
mixture in the sediment, so compliance with an pore water RG must be assessed on an additive basis,
combining the fractional contributions of all narcotic chemicals present. For the detailed derivation of
the narcosis SCVs, the reader should consult EPA 2003 and 2008 (US-EPA 2003c; US-EPA 2008). Recently,
Kipka and DiToro (Kipka and Di Toro 2009) extended the target lipid model of narcotic toxicity to polar
narcotic chemicals using a polyparameter linear free-energy relationship (LFER). With the LFER model,
narcosis SCVs for polar organics can now be derived when needed.
3.2 Sensitivities of Benthic and Pelagic Organisms
The calculation methodology for FCV and SCV values combined toxicity data for benthic and pelagic
organisms, which provides a more robust and phylogenetically diverse sensitivity distribution. Applying
FCV/SCV values as pore water RGs assumes that there is no inherent bias in applying these values in
contexts where protection of benthic organisms is the explicit goal. The appropriateness of this
assumption has been evaluated in a number of analyses, asking the question, "Are benthic organisms
consistently more or less sensitive to chemical toxicants than are pelagic organisms?"
Figure 3-1 compares the acute toxicity values for the most sensitive benthic (infaunal and
epibenthic) species to the most sensitive water column species (Di Toro et al. 1991). The data are from
the 40 freshwater and 30 saltwater draft or published AWQC documents that meet minimum data base
requirements for calculation of a final acute value (FAV). Plotted in Figure 3-1 are the lowest (i.e., most
sensitive) LC50 values (lethal effect concentration of the toxicant that gives half-maximal response) for
water column and benthic species, plotted separately for freshwater and marine organisms. As can be
seen, the values are distributed closely around the unity, with no evidence of consistent bias above or
below the line. This supports the assumption of equal sensitivity between benthic and water column
organisms (Di Toro et al. 1991).
19
-------
Comparison of Most Sensitive Species
si
" i * ' i ¦ \ * y
m
3
© a
§ 3
S
5 -
0
• Freshwater +
- ® Saltwater • * -
% Jm £32)
a • X *.
" f '
J -3
^ , J .
3-11 3 8
Left© ienthlc LC50 ipglL)
Figure 3-1. A comparison of the minimum LC50 for infaunal and epibenthic species (x-axis) and water
column (y-axis) species. Each data point represents a particular chemical in either a
freshwater and saltwater exposure. The diagonal line indicates a 1:1 relationship. The data
are from AWQC or draft criteria documents (Di Toro et al. 1991). Reprinted with
permission.
Figure 3-1 combines data across chemicals, but evaluates only the most sensitive organism for each
chemical. Another way to address the benthic vs water column (pelagic) sensitivity question is to look at
the distribution of values within the species sensitivity distribution (SSD) for a single chemical. Figures 3-
2 to 3-4 show the SSDs of LC50 values for dieldrin (US-EPA 2003a), endrin (US-EPA 2003b), and PAH
mixtures (US-EPA 2003c). The symbols represent broad phylogenetic groupings, and filled and open
symbols show species that are benthic and pelagic, respectively. Examination of these figures shows
that benthic and pelagic species are well distributed across the range in organism sensitivity, and all
three plots have benthic species with sensitivities at or near the lower end of the distribution. Statistical
analysis of these distributions can be found in the referenced source documents.
The above conclusion of similar sensitivities between benthic and pelagic organisms is further
supported by a more recent analysis by Redman et al. (Redman et al. 2014). Redman et al. (2014)
demonstrated that SSDs for terrestrial (soil) and benthic (sediment) species were similar to SSDs for
aquatic species, i.e., differences less than 2 fold. Further, SSDs for acute-to-chronic toxicity ratios were
similar for aquatic and soil/sediment species.
Based upon these analyses from Di Toro et al. (Di Toro et al. 1991), Redman et al. (Redman et al.
2014), and EPA's ESBs documents (US-EPA 2003a; US-EPA 2003b; US-EPA 2003c), the uniform conclusion
is that there is no evidence that the toxicant sensitivity of benthic organisms is systematically biased
relative to water column organisms. This in turn supports the use of FCV and SCV values calculated from
toxicity data sets combining benthic and pelagic species, as is performed in the derivation of AWQC and
GLI Tier II SCVs.
20
-------
1000
Orconectes (A)
Gammarus (A,X) ir
A Arthropods
¦ Other Invertebrates
• Fish and Amphibians
Simocephalus (J,X)
Daphnia (A,J,X)
Bufo (L)
100
Psuedocris (L)
1—,
Paloemonetes (X)
Lumbriculus
Lepomis (J)
Ictalurus (X)
Ischnuro (J)
Pimephales (J)
>r Tilapia (J)
Oncorhyncus (J,X)
Corossius (J,X)
Asellus (X)
Micropterus (X)
Poecilla (J)
L Claossenio (J)
Pteronocella (J)
Pteronorcys (J,N)
20
100
Percentage Rank of Freshwater Genera
Figure 3-2. Species Sensitivity Distribution (SSD) for dieldrin of freshwater genera for acute toxicity (US-
EPA 2003a). Genus mean acute values from water-only acute toxicity tests using freshwater species
versus percentage rank of their sensitivity. Symbols representing benthic species are solid; those
representing water column species are open. A=adult; J=juvenile; N=naiads; X=unspecified life-stage.
21
-------
1000 F
100
J
as
J2
5
OJ
s
U
<
C
re
GJ
3
C
0)
(5
).l
A Arthropods
S Other Invertebrates
# Fish and Amphibians
Lumbriculus* (A) m_.
Bufo (L) •/
Hexogenio (J) /
Dophnia (L)
Simocephalus (X) gf
i Orconectes (
^ Tipulo (J)
y Rano (L)
/c.
Jordonello (J).
Tanytarsus (L)^
Gambusia (J)
Tilapio (J)
Atherix (J)
Gommorus (A)
Aseilus (A) ^ Ischnura (J)
Poecilia (X) q_a- ^ Cypridopsis (A)
Polaemonetes (A)
\ Corossius (J)
Pimephales (J).... Q ^ Baetis(J)
Brachycentrus (X) \ Pteronarcello (L)
_ Micropterus (J)q.-o * \ Oncorhynchus (J)
...a-K \ Ictolurus (J)
Cyprinus (J)
Pteronorcys (A)
Claossenia(A)
Lepomis (J)
! Acroneurio* (L)
Perco (J)
a"
20
80
100
Percentage Rank of Freshwater Genera
Figure 3-3. Species Sensitivity Distribution (SSD) for endrin of freshwater species for acute
toxicity (US-EPA 2003b). Genus mean acute values from water-only acute toxicity tests using
freshwater species versus percentage rank of their sensitivity. Symbols representing benthic
species are solid; those representing water column species are open. Asterisks indicate greater
than values. A = adult, J = juvenile, L = larvae, X = unspecified life-stage.
22
-------
400
O Water column life stages
A Benthic life stages
Tonytarsus
Crepidula
U
O
OJO
American Lobster
Aplexa. I
Cyprinodon. \ 1
Neanthes
Mudoho
100
Pimephales
0)
_3
5
0)
Chironomus
Winter Flounder
Ophiogomplius »
Lumb'iculus f\^
Physella
*-»
3
u
<
C
re
2
'Muhnia
Eurytemora
Oncorhynchus
Physa
Mya Ampelisca
Mytililus \ I
Lepomis
Daphnia
Ictalurus
'Menidio
Excirolano
Dinophilus *
Peltoperla,
Rhepoxymus , V,*
Xenopus
Eohaustorius
Corophiurr
Paqurus
Leptocheirus
Hydra
^ Shrimp
Neomysts
Hyalella
FAV = 9.32 nmoi/g
Oncorhynchus
Crangon
0 5
20
40
60
80
100
Percentage Rank of Freshwater and Saltwater Genera
Figure 3-4. Species Sensitivity Distribution (SSD) for PAH mixtures for acute toxicity (US-EPA 2003c).
Genus Mean Acute Values at a logio/Cow of 1.0 from water-only acute toxicity tests using freshwater and
saltwater genera versus percentage rank of their sensitivity.
23
-------
Table 3-1. Conventional and narcosis water-only chronic toxicity values (jag/L) (FCVs and SCVs), Equilibrium Partitioning Benchmarks (ESBs), and
narcosis equilibrium partitioning sediment benchmark (ESB) values (ng/goc) for a selection of nonionic organic chemicals (Burgess et al. 2013).
Chemical
Log
^ow
Conventional
ESB (|ig/goc)
Freshwater
Marine
Narcosis
ESB
(Hg/goc)
Conventional
FCV or SCV (jLLg/L)
Freshwater
Marine
Narcosis
SCV
(Hg/L)
Ethers
4-Bromophenyl phenyl ether 5.00 120 120
Low Molecular Weight Compounds
Benzene 2.13 16 16
Chlorobenzene 2.86 41 41
1.2-Dichlorobenzene 3.43 33 33
1.3-Dichlorobenzene 3.43 170 170
1.4-Dichlorobenzene 3.42 34 34
Ethylbenzene 3.14 8.9 8.9
1,1,2,2-Tetrachloroethane 2.39 140 140
Tetrachloroethene 2.67 41 41
Tetrachloromethane 2.73 120 120
Toluene 2.75 5.0 5.0
Tribromomethane (Bromoform) 2.35 65 65
1,1,1-Trichloroethane 2.48 3.0 3.0
Trichloroethene 2.71 22 22
m-Xylene 3.20 94 94
Pesticides
Alpha-, Beta-, Delta-BHC 3.78 11 NA
Gamma-BHC, Lindane 3.73 0.37 NA
Biphenyl 3.96 110 110
Diazinon 3.70 0.74 3.6
Dibenzofuran 4.07 37 37
Dieldrin 5.37 12 28
Endosulfan mixed isomers 4.10 0.6 0.093
Alpha-Endosulfan 3.83 0.33 0.051
1600
660
570
780
780
780
970
830
840
770
810
1200
660
650
980
1500
1700
SCV = 1.5
SCV =130
SCV = 64
SCV = 14
SCV = 71
SCV = 15
SCV = 7.3
SCV = 610
SCV = 98
SCV = 240
SCV = 9.8
SCV = 320
SCV = 11
SCV = 47
SCV = 67
SCV = 2.2
FCV = 0.08
SCV = 14
FCV = 0.1699
SCV = 3.7
FCV = 0.06589
FCV = 0.056
FCV = 0.056
SCV = 1.5
SCV =130
SCV = 64
SCV = 14
SCV = 71
SCV = 15
SCV = 7.3
SCV = 610
SCV = 98
SCV = 240
SCV = 9.8
SCV = 320
SCV = 11
SCV = 47
SCV = 67
NA
NA
SCV = 14
FCV = 0.8185
SCV = 3.7
FCV = 0.1469
FCV = 0.0087
FCV = 0.0087
19
5300
880
330
330
340
790
3700
2000
1600
1600
6000
2400
1400
700
190
170
24
-------
Beta-Endosulfan 4.52 1.6
Endrin 5.06 5.4
Hexachloroethane 4.00 100
Malathion 2.89 0.067
Methoxychlor 5.08 1.9
Pentachlorobenzene 5.26 70
Toxaphene 5.50 10
1,2,4-Trichlorobenzene 4.01 960
Phthalates
Butyl benzyl phthalate 4.84 1100
Diethyl phthalate 2.50 77
Di-n-butyl phthalate 4.61 1200
Polycyclic Aromatic Hydrocarbons'3
Naphthalene 3.356 NA
Cl-naphthalenes 3.80 NA
Acenaphthylene 3.223 NA
Acenaphthene 4.012 NA
C2-naphthalenes 4.30 NA
Fluorene 4.208 NA
C3-naphthalenes 4.80 NA
Anthracene 4.534 NA
Phenanthrene 4.571 NA
Cl-fluorenes 4.72 NA
C4-naphthalenes 5.30 NA
Cl-phenanthrene/anthracenes 5.04 NA
C2-fluorenes 5.20 NA
Pyrene 4.922 NA
Fluoranthene 5.084 NA
C2-Phenanthrene/anthracenes 5.46 NA
C3-fluorenes 5.70 NA
Cl-pyrene/fluoranthenes 5.287 NA
C3-phenanthrene/anthracenes 5.92 NA
Benz(a)anthracene 5.673 NA
0.24
0.99
100
0.11
NA
70
54
960
1400
1600
1100
FCV = 0.056
FCV = 0.05805
SCV = 12
SCV = 0.097
SCV = 0.019
SCV = 0.47
FCV = 0.039
SCV=110
FCV = 0.0087
FCV = 0.01057
SCV = 12
FCV = 0.1603
NA
SCV = 0.47
FCV = 0.2098
SCV=110
160
11
120
NA
NA
NA
SCV = 19
SCV = 210
SCV = 35
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
385
444
452
491
510
538
581
594
596
611
657
670
686
697
707
746
769
770
829
841
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
NA
193.5
81.69
306.9
55.85
30.24
39.3
11.1
20.73
19.13
13.99
4.048
7.436
5.305
10.11
7.109
3.199
1.916
4.887
1.256
2.227
25
-------
Chrysene
5.713
NA
NA
844
NA
NA
2.042
C4-Phenanthrenes/anthracenes
6.32
NA
NA
913
NA
NA
0.5594
Cl-Benzanthracene/chrysenes
6.14
NA
NA
929
NA
NA
0.8557
Benzo(a)pyrene
6.107
NA
NA
965
NA
NA
0.9573
Perylene
6.135
NA
NA
967
NA
NA
0.9008
Benzo(e)pyrene
6.135
NA
NA
967
NA
NA
0.9008
Benzo(b)fluoranthene
6.266
NA
NA
979
NA
NA
0.6774
Benzo(k)fluoranthene
6.291
NA
NA
981
NA
NA
0.6415
C2-benzanthracene/chrysenes
6.429
NA
NA
1008
NA
NA
0.4827
Benzo(ghi)perylene
6.507
NA
NA
1095
NA
NA
0.4391
C3-benzanthracene/chrysenes
6.94
NA
NA
1112
NA
NA
0.1675
lndeno(l,2,3-cd)pyrene
6.722
NA
NA
1115
NA
NA
0.275
Dibenz(a,h)anthracene
6.713
NA
NA
1123
NA
NA
0.2825
C4-benzanthracene/chrysenes
7.36
NA
NA
1214
NA
NA
0.07062
NA = Not Available. a Conventional value should be used. b For C#-PAH groups, reported log Kow values are the average log Kow values of all structures
(US-EPA 2003c). FCV = final chronic values. SCV = secondary chronic values.
26
-------
3.3 Derivation of EPA's AWQC FCVs
As discussed in Section 3.2, FCVs from EPA's AWQC should be used as the appropriate adverse
effects concentrations in the sediment pore water for the protection of benthic organisms. EPA's AWQC
(Stephan et al. 1985) are derived by assembling species sensitivity distributions (SSDs) using the genus
mean chronic toxicity values, and the FCV is the 5th percentile from the SSD for the chemical of interest.
The preferred approach for developing the FCV is to use chronic toxicity data and directly compute the
FCV from the chronic toxicity SSD. When insufficient chronic toxicity data are available, a SSD is
developed using genus mean acute toxicity data and from this SSD, the 5th percentile Final Acute Value
(FAV) is determined. Subsequently, the FAV is converted to a FCV using an appropriate acute to chronic
toxicity ratio (ACR) for the chemical of interest.
3.4 Sensitivities of Toxicity Test Organisms in Relation to EPA's AQWC FCVs
Acute and chronic sediment toxicity tests with marine amphipods (Ampelisca abdita, Eohaustorius
estuarius, Leptocheirus plumulosus, and Rhepoxynius abronius), and freshwater species (Chironomus
tentans and Hyalella azteca) provide toxicity data for these, few, select species. The acute toxicity tests
provide data on survival from a 10-day test (US-EPA 2000b, US-EPA and US-ACE 2001) while the chronic
tests provides data on survival, growth, and reproduction from a 28-day (Leptocheirus plumulosus), 42-
day (Hyalella azteca), and life-cycle (Chironomus tentans) tests (US-EPA 2000b, US-EPA and US-ACE
2001). Examination of the genus mean chronic value data for PAHs (Figure 3-5, Table 3-2) reveals that
the freshwater and marine sediment toxicity test species reside at different points along the SSD. None
of the common sediment toxicity test species have acute toxicity values at the FAV for PAHs of 9.32
nmole/g octanol (US-EPA 2003c). Because species used in sediment toxicity tests are not necessarily at
the 5th percentile in the SSD, one should not expect them to be as sensitive as the FAV. Added to this is
that the FCV is intended to protect sensitive organisms from effects on survival, growth, or reproduction
when exposed over their entire life cycle. Because many sediment toxicity test methods do not include
full life cycle exposure, further differences in sensitivity can be expected between the pore water RG
(based on the FCV or comparable effect level) and the results of sediment toxicity tests. Finally, for
chemicals whose pore water RG is based on a SCV calculated using the GLI Tier II procedures, additional
conservatism may (and may not) be introduced by the adjustment factors applied for chemicals that
have limited toxicity data availability.
To compare results of toxicity tests more directly to chemical concentrations measured in pore
water, it is possible to calculate species/chemical-specific pore water effect concentrations based on the
results of water column exposures using the same chemical, species, and endpoint. To do this,
species/chemical-specific pore water toxic units (TUs) can be estimated by replacing the FCV with the
applicable effect concentration from a water-only toxicity test and recalculating pore water TUs.
27
-------
Table 3-2. PAH mixture species sensitivity distribution genus mean acute values for marine
amphipods Ampelisca abdita, Eohaustorius estuarius, Leptocheirus plumulosus, and Rhepoxynius
abronius, and for freshwater species Chironomus tentans and Hyalella azteca.
Species
Genus Mean Acute Value
(lAmole/ g octanol)
Percentage Rank of Genera
5th Percentile distribution value
FAV = 9.32
5.0%
Hyalella azteca**
10.2%**
Leptocheirus plumulosus
19.0
22.4%
Rhepoxynius abronius
19.9
26.5%
Eohaustorius estuarius
22.1
32.6%
Ampelisca abdita
30.9
55.1%
Chironomus tentans
68.4
79.5%
** Later acute toxicity tests with H. azteca using 3 PAHs indicate sensitivity about 2.5-fold lower than
indicated here (Lee et al. 2001); this may be related to low chloride concentration in the dilution water
used in the test used for the sensitivity distribution. Recent research has suggested that low chloride
concentrations are stressful to H. azteca and can lead to greater apparent sensitivity that may not be
representative of aquatic organisms generally (Soucek et al. 2015).
3.5 Measuring Water Only Toxicity Value for Toxicant(s)
At some sites, effect concentrations for the species used by sediment toxicity tests might not be
available for the toxicants of interest. Measurement of a water-only toxicity effects concentration(s) for
the site's toxicant(s) can be conducted by performing aquatic toxicity tests with the chemical(s) of
interest in water-only exposures. Standardized acute or chronic water toxicity test methods should be
used (US-EPA 1996a; US-EPA 1996b; US-EPA 1996c; US-EPA 1996d). Proper implementation will require
measured concentrations in the water over the duration of the toxicity test. Use of passive dosing
techniques is suggested for more hydrophobic chemicals (Butler et al. 2013; Smith et al. 2010; Smith et
al. 2009) and the use of solvent carriers, e.g., acetone or DMSO, to dissolve chemicals in the tests is not
recommended. Results based upon nominal concentrations of the toxicants are unacceptable. The
testing must involve a range of chemical concentration steps in order to create dose-response curve(s)
for the toxicant(s) and enable the determination of an EC50(s) (effect concentration of the toxicant that
gives half-maximal response) for the toxicant(s). With the newly measured toxicity value(s), a FCV or
SCV can be derived for the chemical or mixture of chemicals of interest. Note, performing water-only
toxicity tests will be costly and time consuming. As such, this approach is only recommended in those
situations where the costs and time commitments warrant such efforts.
28
-------
Section 4
Implementation of the Pore Water RG Approach within the Superfund Remedial
Investigation/Feasibility Study (RI/FS) Process:
Following Superfund Guidance, the Remedial Investigation (Rl) is a three part process (US-EPA
1997): 1) characterization of the nature and extent of contamination; 2) ecological risk assessment; and
3) human health risk assessment. The investigation of the nature and extent of contamination
determines the chemicals present on site as well as their distribution and concentrations. The ecological
and human health risk assessments determine the potential for adverse effects to the environment and
human health, respectively. The focus of this document is the protection of benthic organisms living in
and on the sediments, and resultantly, only characterization of the nature and extent of contamination
and ecological risk assessment parts of Rl will be discussed. For human health risk assessment, the
reader should consult US-EPA (US-EPA 1989) for further information.
4.1 Characterization of the Nature and Extent of Contamination
Following Superfund guidance, the RI/FS process starts with Scoping Activities in which a conceptual
site model (CSM) is assembled and initial data needs and Data Quality Objectives (DQOs) are identified
for the site. In addition, a work plan for the site is developed along with a sampling and analysis plan
(SAP) for field investigations. The Rl should develop sufficient data to define (US-EPA 1988; US-EPA
1997):
• Site physical characteristics
• Physical and chemical characteristics of sources of contamination
• Volume of contamination and extent of migration
• Potential receptors and associated exposure pathways
• Baseline human health and ecological risks
In the Rl, field investigations are conducted to characterize the nature and extent of contamination
including the average contaminant concentrations. The field investigations are implemented in an
iterative fashion such that the locations and concentrations of any migrating contaminants can be
defined. The Rl should collect adequate data of sufficient quality to support risk assessment and the
analysis of remedial alternatives. As part of the baseline ecological risk assessment sediment toxicity
tests on sediment samples from the site are often performed.
To lay groundwork for later development of pore water-based remedial goals, additional data
should be collected during Rl to define the following:
• Nature and variability of the organic carbon content (f0c) of the sediments across the site
• Nature and variability of the site-specific sediment-organic carbon-water partition coefficients
(Kocs) across the site.
Rearranging Equation 1-1, illustrates why one needs to understand the nature and variability of f0c
and Koc cross the site.
Cfree = i^s/foc)/^OC = ^SOc/^OC (4"1)
29
-------
From the rearranged equation, the concentration of the freely dissolved chemical in the sediment pore
water (Cfree) is a function of both f0c and K0c parameters. The Cfree values are used in the Risk
Assessment part of the Rl (see Ecological Risk Assessment section below) to determine if concentrations
of the chemicals of concern in the sediments are at levels to cause unacceptable effects on benthic
organisms. Understanding the nature and variability of the f0c and K0c parameters across the site allows
the nature and extent of the contamination in the sediment pore water to be determined, i.e., Cfree
values in the sediments across the site. With the Cfree values, assessments of risk can be performed for
the chemicals of concern.
To provide some background on what one might see in terms of variability, f0c, data from a few sites
is provided below (Table 4-1), and the range in f0cs for a site is potentially a factor of 10 or more.
Table 4-1. The foe in sediment samples from Superfund sites
Site
Average
Standard
Deviation
Ratio of
Maximum to
Minimum Values
Minimum
Maximum
n
Reference
Fox River
0.0897
0.0655
17.5
0.0141
0.2462
10
a
Hudson River
0.0217
0.0146
CO
CO
0.0052
0.0456
10
a
Anniston
0.0152
0.0102
18.1
0.0022
0.0399
33
b
Portland Harbor
0.0220
0.0143
34.1
0.0020
0.0682
35
c
New Bedford
0.0192
0.0128
91.1
0.00057
0.0519
82
d
a (Burkhard et al. 2013) b (Ingersoll et al. 2014) c (lntegral_Consulting_and_Windward_Environmental
2006) d (Nelson and Bergen 2012)
For Koc, based upon the data from a limited number of sites (Table 4-2), the range in K0cS for PCB
congeners 118 and 153 at site can vary. For the Fox River and Anniston sites, the range in Kocs is
approximately an order of magnitude, and for the Hudson River site, range is smaller.
Table 4-2. The site-specific log Kocs for PCB-118 and PCB-153 in sediment samples from Superfund sites
Site
PCB
Average
Standard Deviation
Range
Minimum
Maximum
n
Reference
Fox River
118
7.07
0.29
1.01
6.49
7.50
10
a
Fox River
153
7.38
0.27
0.69
6.97
7.66
10
a
Hudson River
118
7.35
0.21
0.57
7.11
7.69
10
a
Hudson River
153
7.40
0.10
0.31
7.28
7.59
10
a
Anniston
118
4.59
0.30
1.29
3.91
5.20
24
b
Anniston
153
5.02
0.26
1.00
4.56
5.56
25
b
a (Burkhard et al. 2013) b (Ingersoll et al. 2014)
Another evaluation of within and between site variations in Koc values is presented by Hawthorne et
al. (Hawthorne et al. 2006), who determined K0cvalues for a range of PAHs found in 114 sediments
collected in association with 8 different sites (six manufactured gas sites and 2 aluminum smelters). As
shown in Figure 4-1, Kocvalues varied by 100 to 1000-fold across the full range of samples, illustrating
why assuming a single Koc value for a site can introduce considerable uncertainty if applied blindly.
30
-------
O)
o
-J
i »
8 *
9
C
C
V
0)
TO
CO
£
-C
jz
JZ
Q.
Q.
©
c
c
>N
JZ
Qi
E
M
CA
w
a>
a>
c
a>
c
£
0>
fiJ
Ctt
m
c
>
C
<2i
c
a>
c
£
>,
,c
(Z
a>
ft}
m
sz
o
o
-C
JZ
c
xz
u
cv
o
o
<5
TO
O
o
O
« «> ¦»
aj a) e
T, c c
& &
¦a
c.
If
O
©
Q>
(0
a>
c
a>
C
c
z
c
2
c
0?
£
>s
Q.
*
a
Q.
"S
JZ
c
s
.c
U
ro
a
&
re
3
©
N
C
c
TO
o
C=
£
c
"3
-s:
0*
a
o>
+
o
3
o
g
-Q
a
e-
a
£ &
Figure 4-1. Measured Koc values in sediments from manufactured gas and aluminum smelter sites as
reported by Hawthorne et al. (2006). Reprinted with permission
The reasons why foe and Koc parameters vary within and among sites are understood, even if they
cannot be easily predicted. In ecosystems relatively untouched by human activities, organic carbon in
the sediments arises, almost exclusively, from the diagenesis of plant materials. At Superfund sites, in
addition to naturally derived organic carbon, anthropogenic carbon from industrial activities can be
present, and could include coal, soot (black carbon), wood chips, sawdust, tars, NAPLs, oils/greases,
and/or microplastic particulates. It is well documented that anthropogenic carbon types have sorption
abilities that are different from the organic carbon resulting from the diagenesis of plant materials
(Cornelissen et al. 2005), and in general, their sorption abilities are larger. This results in the Kocs being
larger than the Kocs measured with organic carbon derived from the diagenesis of plant materials. With
larger Kocs for sediments containing anthropogenic carbon, the concentrations of chemicals in the
sediment pore water are lower in comparison to sediments with little no anthropogenic carbon (when
having the same concentrations of chemicals on a bulk dry weight basis with the same organic carbon
content).
31
-------
The foe parameter also varies at sites due differences in sedimentation rates, sediment type, and
sediment movement at individual sites. For example, locations with sediments having high sand content
generally have low organic carbon content (e.g., 0.2% or less), and depositional locations generally have
higher organic carbon contents (often 10% or more). Even where K0c is constant, this range in f0c can
lead to a 50-fold difference in the bulk sediment concentration expected to cause toxic effects. In
characterizing the f0c parameter for a site, measurement of TOC on all sediment samples is
recommended. The analysis costs for f0c is relatively inexpensive and with f0c information, one can
readily define the nature and variability of the f0c parameter across the site.
For characterizing the nature and variability of the K0c parameter for a site, the approach must
balance cost and comprehensiveness. Passive sampling on every sample more than doubles analytical
costs, because two sets of measurements are performed (i.e., measuring contaminant concentrations in
the bulk sediment and measuring contaminant concentrations in the passive samplers) and additional
effort is required to conduct the sample equilibration. However, the variability in K0c shown in the
examples above clearly shows that determining Kocfrom just a single sample would fail to inform the
assessment sufficiently.
At a minimum, a set of surface samples should be subjected to passive sampler measurements. The
number and placement of these samples is dependent in part on the size and complexity of the site.
Important characteristics to consider are how homogeneous the site sediments are (e.g., depositional
versus higher flow areas), the complexity of past operations and/or sources, and the presence of key
areas within the site that might influence remedial design (e.g., near boundaries between higher and
lower levels of contamination. Overall, the samples taken should allow the nature and variability of the
site-specific K0cS across the site to be defined; this characterization is analogous to the efforts for
defining the nature and variability of the chemical contamination at the site.
As discussed above, the Rl process is an iterative process where field investigations and analytical
techniques of increasing accuracy are employed to define the nature and extent of the contamination at
the site. Similarly, passive sampling measurements can be employed in an iterative process. For
example, one might not need to perform any passive sampling measurements on the initial sediment
samples from the site, and use only total concentrations in sediment and f0c to support a screening risk
assessment (see Ecological Risk Assessment section below) to highlight areas of potential concern.
These areas could then undergo targeted studies with passive sampling to refine risks posed by the
contaminants. If the potential risks appear wide spread based upon the organic carbon normalized
concentrations, a more extensive use of passive sampling might be warranted so that more reliable
bioavailability corrections can be incorporated into the risk assessment. If at the site, large variations in
total contaminant concentrations exist, defining how Kocs (ultimately, the contaminant bioavailability)
are a function of contaminant concentrations might be required. Successful and cost effective use of
passive sampling requires a good CSM and well defined study objectives for the measurements.
32
-------
4.2 Ecological Risk Assessment
Superfund's ecological risk assessment guidance is an eight-step process, and involves a screening
level ecological risk assessment (Steps 1 and 2); followed by problem formulation (Step 3); then, study
design and data quality objective development (Step 4); and results in a site's work plan (WP) and
sampling and analysis plan (SAP). With the WP and SAP, site investigation (Steps 5 and 6) is performed
followed by risk characterization (Step 7). The eight-step process is shown below in Figure 4-2, and
Figure 4-2 is taken directly from Superfund's "Ecological Risk Assessment Guidance for Superfund:
Process for Designing and Conducting Ecological Risk Assessments. Interim Final" guidance document
(US-EPA 1997).
For COC identification, the methodology within this document does not require any additional steps
or procedures beyond what is normally done in the Rl process.
For exposure assessment, the methodology within this document requires the development of Cfree
values for the COCs in the sediment pore water across the site or appropriate locations and/or
operational units for the site.
For toxicity assessment, the methodology within this document sets the unacceptable effects levels
to the FCV for single contaminants, or for a mixture of contaminants of the same class (e.g., PAHs; see
discussion below), the unacceptable effects level is found using the sum of the toxic units (TU; relative
to the FCV) for the mixture and setting the sum to no more than 1.0. (See examples in Section 4.3.2.2
for calculation of TUs)
For risk characterization for organic chemicals in surface sediments, we recommend that
concentrations be expressed on a total organic carbon basis, i.e., ng of chemical/kg of sediment organic
carbon. Then, these values can be compared to EPA's ESBs (with units of ng of chemical/kg of sediment
organic carbon) or similarly developed values for chemicals without published ESBs. If concentrations
are less than the ESBs, then one would conclude that there is little potential for unacceptable risks to
the benthic organisms from the COCs. This approach is based on the general experience that
bioavailability of organic contaminants in sediments is generally no greater than would be predicted by
assuming that K0c ~ K0w. Initially, evaluating contamination on an organic carbon normalized basis
before performing passive sampling measurements focuses analytical resources and efforts for passive
sampling on the locations of higher concern within the site, i.e., locations where the ESBs are exceeded.
When concentrations on an organic carbon normalized basis exceed the ESBs, unacceptable risks to
benthic organisms might exist, but that risk is dependent on the degree of the chemical's bioavailability
(partitioning). Passive sampling can address that uncertainty, by determining concentrations of COCs in
33
-------
STEP 1: SCREENING-LEVEL:
• Sltt Visit
• Problem Formulation
• Toxicity Evaluation
STEP 2: SCREENING-LEVEL:
• Exposure Estimate
• Rrsk Calculation
Risk Assessor
and Risk Manager
Agreement
i
SMDP
?
as
O
O
A3
n»
Q
STEP 3: PROBLEM FORMULATION
Tox«ity SvafuaJtort
7
Assessment
Sfldcoints
7
Conceptual Modal
Exposure Palhwa/B
i
I
Ouesiiorhs/Hycctfiews
STEP 4, STUDY DESIGN AND DOO PROCESS
• lines of Evictence
• Measurement Endpoints
Wory Plan and Sampling artO Analysts Plan
SMDP
SMDP
STEP 5: VERIFICATION OF FIELD
SAMPLING DESIGN
SMDP
STEP 6: SITE INVESTIGATION AND
DATA ANALYSIS
STEP?: RISK CHARACTERIZATION
(SMDP)
STEPS: RISK MANAGEMENT
SMDP
Figure 4-2a. Eight step ecological risk assessment process for Superfund (US-EPA1997). SMDP =
Scientific/Management Decision Point.
34
-------
FROM:
• Preliminary Assessment
• Site Inspection
• NPL Listing
RI/FS
Scoring
SREENING
ECOLOGICAL RiSK
ASSESSMENT
(STEPS 1 & 2)
PROBLEM
FORMULATION AND
STUDY DESIGN
(STEPS 3 5, 4)
Remedial investigation
Feasibility Study
Z si,°
SAp Investigation
Establish Development
Remedial and Analysis
Objectives of Alternatives
FIELD
VERIFICATION
(STEP 5}
Refine remedial
goals based on
risk assessment
ANALYSIS OF
EXPOSURE AND EFFECTS
RISK CHARACTERIZATION
(STEPS 6 & 7)
Conduct risk
evaluation d
remedial
alternatives
TO:
• Remedy Selection
• Record of Decision
• Remetial Design
• Remetfal Action
Ecological
Monitoring
Figure 4-2b. Flowchart of eight-step ecological risk assessment process in the RI/FS process {US-EPA
1997).
pore water {and calculating associated Kocs) for the areas with concentrations exceeding the ESBs.
These Cfree values are compared to the FCVsforthe COCs or when a mixture of contaminants is present,
the Cfree values are used to determine the total TUs of the mixture. When values are higher than their
FCVs or {for applicable mixtures) the total TUs exceed 1.0, unacceptable risks to benthic organisms are
anticipated. When values are less than their FCVs or the total toxic units are less than 1.0, there is low
potential risk to benthic organisms, despite the initial finding that Csoc exceeded screening values.
Depending upon data availability at your site, comparison of the COC concentrations to their ESBs might
occur in the screening level ecological risk assessment {Steps 1 and 2). However, if field measurements
are required, the comparisons would be performed later in the ecological risk assessment process, e.g.,
Steps 3, 4, or 6.
4.3 Pore Water Remedial Goal Development (PWRG)
In the ecological risk assessment, risks to benthic organisms are assessed based on Cfree
measurements and this information is used to develop PWRGs expected to be protective of benthic
organisms. Although derived using the Cfree in the pore water, remedial goals may be expressed on the
bases of Cfree itself or on conversion of Cfree to Csoc or Cs, depending on site characteristics and the needs
35
-------
of the assessment. This section discusses remedial goal development based on different expressions of
Cfree*
4.3.1 Pore Water Remedial Goal Development using Cfree Values
When developing PWRGs using Cfree values, for situations where there is only one contaminant, the
PWRG equals FCV for chemical. For situations where there is a mixture of contaminants present,
PWRGs for each chemical are determined by using the composition of the mixture and setting the total
amount of the mixture to a concentration where the total TUs for the mixture is equal to 1.0.
An example of where using Cfree directly might be advantageous is in the addition of in-situ amendments
such as activated carbon to the sediments (Beckingham and Ghosh 2011; Cho et al. 2009; Cho et al.
2012; Tomaszewski et al. 2007). The goal of these amendments is not to reduce Cs per se, but to use the
sediment amendments to reduce the Cfree values to meet the remedial goal.
4.3.2 Remedial Goal Development by Conversion of Cfree Values to Concentrations in Sediment on an
Organic Carbon Basis (Csoopwrg) or to Concentrations in the Bulk Sediment (Cs:pwrg)
Some may wonder, "Why not simply express the PWRGs using Cfree values at all sites when pore
water measurements are made?" For some sites, developing PWRGs in terms of concentrations in the
bulk sediment may lower analytical costs, simplify field collection efforts, and/or avoid issues with
passive sampler devices being lost, damaged, and/or vandalized in the field. Additionally, these issues
might apply in both RI/FS and post remedial monitoring phase for the site.
Conversion of pore water RGs expressed using Cfree values to concentration in the sediment on an
organic carbon basis (Csocpwrg (l^g/kg organic carbon) requires rearranging Equation 1-1 and dropping of
the focss term from the equation:
CsOC\PWRG = K0C:SS X Cfree\PWRG (4"2)
Conversion of pore water RGs expressed using Cfree values to concentration in the bulk sediment
(Cs:pwrg (M-g/kg dry weight)) requires rearranging Equation 1-1:
(¦S'.PWRG = K0C:SS X foC\SS X ^free\PWRG (4"3)
where K0css (L/kg-organic carbon) is the site-specific K0c, focss is the site-specific f0c (kg organic
carbon/kg dry weight), and Cfree:PWRG is the pore water RG expressed as concentration in water (ng/L).
When a single contaminant is present in the sediments, the Cfree:PWRG equals the FCV for the
contaminant. When a mixture of contaminants is present in the sediments, the Cfree:PWRG values for each
chemical are determined using the composition of the mixture in the sediment pore water and setting
the total amount of the mixture to a concentration where the total TUs for the mixture is equal to 1.0.
In the following discussion, these cases will be illustrated using RGs expressed using concentrations in
the bulk sediment (CS;pwrg )• The RGs could, just as easily, be expressed using concentration in the
sediment on an organic carbon basis (Csocpwrg) using equation 4-2.
36
-------
4.3.2.1 Derivation of CS;pwrg values for a Sediment with One Primary Contaminant - Dieldrin Example
The hypothetical example setting is a riverine Superfund site with sediments contaminated with
dieldrin (Figure 4-3). In the Rl process, concentrations of dieldrin and foes were measured in the surface
sediments in order to define their nature, variability and locations for the site. In this example,
concentrations in the surface sediments were high at the source of the chemical and then, gradually
decreased going downstream from the source (Figure 4-3: Graph B).
In the Risk Characterization phase of the Ecological Risk Assessment, concentrations of dieldrin in
the surface sediments were converted to an organic carbon basis (Figure 4-3: Graph C), and
subsequently, compared to dieldrin's ESB value of 12 ug/g-organic carbon (from Table 3-1). The surface
sediments in river miles 148 to 159 have organic carbon normalized concentrations above the dieldrin's
ESB. These data allowed the focusing of resources on the areas of the site where unacceptable effects
to benthic organisms might potentially exist. In this example, the surface sediments in river miles 148 to
159 were sampled using passive samplers along with one additional surface sediment sample at river
mile 138 for comparison purposes.
The Cfree and resulting K0cS from the passive sampling measurements for river miles 148 to 159 and
138 are shown in Graphs D and E of Figure 4-3. These data define the nature and variability across these
locations within the site for Koc and Cfree. Overall, the measured Kocs are similar across all measurement
locations and there appears to be no substantial difference between the locations downstream and the
highly contaminated areas of the river.
For this site contaminated with dieldrin, the pore water RG (Cfree:PWRG) would be set equal to the FCV
from EPA's AWQC for dieldrin, which is 0.06589 ug/L in the sediment pore water (Table 3-l)(US-EPA
2003a). As shown in Graph D of Figure 4-3, the Cfree values in the sediment pore water are greater than
dieldrin's FCV for river miles 151 to 159. The Cfree value for river mile 138 is much lower than the FCV of
dieldrin even though the measured Koc at river mile 138 is similar to those in the river miles 148 to 159.
Comparing Graphs C and D in Figure 4-3 reveals that the correction for chemical bioavailability (captured
by measuring the Cfree in the sediment pore water using passive sampling) lowers the number of
locations in the river where unacceptable effects from dieldrin potentially exists for benthic
invertebrates.
To convert the pore water RG (Cfree:PWRG) expressed on a ng/L basis to a bulk sediment basis (CS;pwrg),
Equation 4-3 is used. In order to use the equation, values for K0c:ss (L/kg-organic carbon) and f0c;ss (kg
organic carbon/kg dry weight) are needed. In Graph F of Figure 4-3, Cs;pwrg values, calculated using
Equation 4-3, are provided for river miles 151 through 159 using a variety of different calculation
methods:
A) by individual sampling location, i.e., set K0css and f0css equal to the measured K0c and f0c,
respectively, at each sampling location,
B) by setting the K0css and f0css equal to the average K0c and average f0c across the nine sampling
locations,
37
-------
C) by using the largest Koc across the nine sampling locations with foc;ss equal to the average foe
across the nine sampling locations, and
D) by using the smallest Kocacross the nine sampling locations with f0css equal to the average f0c
across the nine sampling locations.
As shown in this hypothetical example, depending upon how the K0css and f0css values are derived
for the site, slightly different CS;pwrg values are determined: 900,1900, and 300 ug/kg-dw for methods B,
C, and D, respectively.
In this hypothetical example, all surface sediment samples with concentrations of dieldrin larger
than the dieldrin's ESB value of 12 ug/g-organic carbon were passively sampled. Because of costs and
sample availability, passively sampling all surface sediment samples with concentrations greater than
the chemical's ESB might not always be feasible. The minimum number of surface sediments that
should be passively sampled is tied to the variance of the K0c values. Estimating the variance of the K0c
values for your site may be difficult initially because in the Rl, concentrations of the COCs and f0cs might
be the only data available in the initial steps of the ecological risk assessment. In these cases, using
variances for Koc values from other sites is suggested as a starting point, recognizing that with data
collection from your site, the variance of the K0c values can be determined. With the variance for the
site or sub-units (operational units or spatial locations within the site), the number of surface sediments
for passive sampling can be defined using the level of uncertainties outlined in the CSM for the K0c
values and the appropriate statistical techniques.
The overall process in this hypothetical example is illustrated in Figure 4-4. In Rl, the nature, extent,
variability, and locations of the contamination and f0c are determined for the site. In the Ecological Risk
Assessment, areas where concentrations of the contaminant in the sediment, on an organic carbon
basis, exceed the EPA's ESB for dieldrin, are subjected to passive sampling measurements. These
measurements further refine the risks at the site and incorporate corrections for chemical bioavailability
at the locations where EPA's ESB for dieldrin are exceeded. If passive sampling suggests unacceptable
risks to benthic organisms, PWRGs should be developed. In Figure 4-4, the dashed box and arrows has
been inserted in the pathway for development of the CS;pwrg. In some cases, if both K0c and f0c values
are reasonably consistent across the site, then dry weight normalization may be sufficient for developing
the Cs;pwrg values.
38
-------
Dieldrin
ESB freshwater 12 ug/goc
FCV 0.06589 ug/L
Log Kow 5.28
/
Source Area
4.0%
2.0%
0.0% I
10000
1000
^ 10
1
1000.0
y ioo.o
J 10.0
y i.o
ESB = 12 ug/g0
FCV = 0.06589 ug/L
5
J
li
J1?
8.00
7.00
6.00
5.00
4.00
3000
2000
1000
1900 ug/kg max log Kqc & avg foe
950 ug/kg avg log Koc & avg foc
300 ug/kg min log Koc & avg foc
Figure 4-3. Hypothetical riverine Superfund site with sediments contaminated with dieldrin.
Graph A-foc in sediments. Graph B - concentration of dieldrin in bulk sediment (ug/kg). Graph C-
Concentration of dieldrin in sediment on organic carbon basis (ng/kg-organic carbon). Graph D -
Concentration of freely dissolved dieldrin in sediment pore water (Cfree, ug/L), Graph E - Sediment-
water partition coefficient (K0c)- Graph F - sediment Cs:pwrgS (ug/kg dw) by individual sample, and by
using average, minimum, and maximum Koc values for river miles 151 through 159.
39
-------
ee:PWRG
Is ZPWTU > 1.0 ?
PWRG
ree:PWRG
Have
PWRGs
PWRG Development
Derive appropriate
KoC:SS
Derive Cs0C:
PWRG
Derive C$:
PWRG
Define nature and extent of:
Contamination &foc
Risk Assessment
No
unacceptable
risks present
for benthic
organisms
Determine concentrations in
sediment on organic carbon
basis (ug/kg-OC)
Use Cs.
PWRG
Derive appropriate
KqC:SS & f()C:SS
¦ Does organic carbon
normalization lower
variability?
Determine number of
passive sampling
measurements
Perform passive
sampling
measurements
Site Characterization
Unacceptable
risks present
for benthic
organisms
Figure 4-4. Components of developing PWRGs for sediment contaminants.
40
-------
4.3.2.2 Derivation of Cs;pwrg values for a Sediment with a PAH Mixture as the Primary COC
In this section, computation of the toxicity of a PAH mixture using toxic units is discussed. Unlike the
dieldrin example above, for explanation purposes, the computations are limited, to a single sediment
sample. The calculation would be conducted in the risk assessment portion of the Rl, and for every
sediment sample subjected to passive sampling analysis, this computation would be performed.
PAHs are one of the most common sediment contaminants because of their formation and release
during the use of fossil fuels by developing and industrialized societies (Burgess et al. 2003b).
Depending on the organism, the exposure setting, and the specific PAH compounds, PAHs can elicit
toxicity via several toxic mechanisms, including narcosis, carcinogenicity/mutagenicity, and photo-
enhanced toxicity (US-EPA 2003c). For benthic organisms, the primary mechanism of action for PAHs is
narcosis; photo-enhanced toxicity is possible, but is unlikely to be a factor for benthic organisms except
for sediments in very shallow water and where the water column has fairly high UV transmissivity.
Accordingly, the ESB for PAHs is derived based on narcotic toxicity (US-EPA 2003c). Table 3-1 lists the
narcosis FCVs/SCVs; readers can consult the ESB document (US-EPA 2003c) for more information on
their derivation.
An important feature of narcotic toxicants like PAHs is that their toxicity is additive; in simple terms,
if you have a pore water containing J4 the toxic concentration of PAH A, and J4 the toxic concentration of
PAH B, the combination would be toxic. In practice, PAH mixtures in sediments consist of dozens of PAH
structures, so the pore water RG calculation is more involved than the simple example above. Another
important aspect of assessing sediments contaminated with PAHs using pore water RGs is that the
common practice of measuring 13 to 16 of the common "priority pollutant" PAHs - all unsubstituted
base ring or "parent" structures - does not capture all of the PAH structures that commonly contribute
meaningfully to the toxicity of field mixtures. Measuring only the parent PAHs does not account for the
effects of alkylated PAHs (e.g., methyl-, dimethyl-, ethyl-substituted PAHs like 1-methylnapthalene or
3,6-dimethylphenanthrene) that often represent from 50% to 90% of the overall toxicity potency of
common PAH mixtures. For that reason, application of the pore water RG approach to PAH-
contaminated sediments should be performed only when passive sampling includes the suite of 34 PAH
structures described in the PAH ESB document (US-EPA 2003c) and listed in Table 4-3. Analytical
methods are available for sediments and pore water measurements of the 18 parent PAHs and 16
alkylated groups (e.g., EPA 8270) when the alkylated PAHs are included as analytes in the analytical
method, see (US-EPA 2007a; US-EPA 2014), NOAA Mussel Watch (NOAA 1998), Hawthorne et al.
(Hawthorne et al. 2005), and ASTM D7363 (ASTM 2013).
Because many historical measurements of sediment PAH contamination involved only the 13-16
priority pollutant parent PAHs, the PAH ESB included correction factors/ratios for extrapolating total
toxic units from the 16 priority parent PAHs (or some other subset of the PAHs) to the 34 PAHs (18
parent PAHs and 16 alkylated groups) required for the evaluation of toxicity via narcosis. However,
these ratios are notably variable, and this variation can result in substantial under- or over-estimation of
the total toxicity of sediment samples (US-EPA 2003c). While the costs of the more extensive PAH
analysis is higher, these costs are generally small compared to potential remedial costs, so direct
measurement of 18 parent PAHs and 16 alkylated groups in the pore water measurements is highly
41
-------
recommended. This is not to say that site-specific correction factors couldn't be developed within site
data in order to incorporate additional sediment PAH data into the overall site assessment, only that it is
best to develop site-specific relationships rather than use generic factors.
While the components of the pore water RG process are the same for PAH mixtures as they are for a
single compound, the calculations are more involved. Each of the 34 PAHs (or PAH groups) listed in
Table 4-3 will have its own measured pore water concentration, FCV/SCV, and sample-specific K0c, yet
these must be combined into a single aggregate measure. This is done using a "toxic units" concept,
wherein the fractional contribution of each specific PAH is determined, then these are summed across
all PAHs to determine if the overall PAH pore water RG is exceeded. For each individual PAH, the
measured concentration in pore water (column 3 in Table 4-3) is divided by its corresponding FCV/SCV
from Table 3-1 (column 5 in Table 4-3); the result is the fractional contribution of that PAH to the overall
sediment potency (Column 6), which is the pore water toxic units (PWTU):
PWTUi = Cfreeii/FCVt (4-4)
where Cfree,i is the freely dissolved concentration measured in the pore water using a passive sampling
technique for chemical , and FCVi (or SCVj) is from Table 3-1. The total toxicity of the mixture is
estimated by summing the PWTU of each chemical:
j
PWTUMizture = ^ PWTUi
i=1 (4-5)
where PWTUi is computed using equation 4-4 for each of the ")" chemicals in the mixture.
Using the example from Table 4-3, the measured concentration of naphthalene in this pore water
was 2.89 |ig/L, the PAH-specific FCV/SCV is 193 |ig/L, and the resulting ratio is 0.0150, which is
PWTU naphthalene- These ratios are then calculated for each of the individual PAHs, and the ratios are
added together to derive the overall toxic units (relative to the pore water RG) present. In the example
in Table 4-3, this sum (IPWTU) is 58.68, indicating that the PAH exposure in this sediment greatly
exceeds the PAH pore water RG, which is represented by a summed ratio of 1.00.
As discussed previously, site-specific K0c and f0c values are needed to convert pore water RG values
back to bulk sediment RGs. In the case of a mixture like PAHs, this calculation is complicated by the
need to base this calculation on the relative concentrations of each component of the mixture. For
illustration purposes, Table 4-3 shows the calculation based on a single sample. For this example,
IPWTU = 58.68, meaning that this mixture exceeds the pore water RG by 58.68-fold; put differently,
pore water concentrations would have to be reduced to 1/56.68 or 1.704% of their measured
concentration to meet the pore water RG. Column 7 of Table 4-3 shows the concentration in sediment
pore water that are 1.704% of the measured concentrations. Column 9 shows the sample-specific K0c
values calculated from columns 2 and 3 along with the measured f0c of 0.088 (8.8%). These values are
combined using equation 4-3 to give a PAH-specific CS;pwrg values for each PAH in M-g/kg dry weight
(column 10). For this sample, remediation to non-toxic levels would require decreasing the total PAH
concentration in the sediments from 191.27 to 3.26 ng/g (dw). Note, in this example, calculations were
42
-------
performed using two significant digits for all values to provide clarity to the readers. At your site,
appropriate significant digits should be used.
The detailed approach to handling a PAH mixture at a site would consist of:
1) Compute the total TUs (sum for all 34 PAH groups) for each sampling location and create a surface
contour plot of total TUs for the site, operable unit, or appropriate sub-area of the site
2) Derive a surface contour map for total PAHs in bulk sediment or in sediment on an organic carbon
basis that results in 1.00 TU contours for the site, operable unit, or appropriate sub-area of the site.
The map derived in 2 provides the concentrations of total PAHs in sediment protective of benthic
species, and would be the RGs for the sediment at the site or appropriate sub-area of the site.
Simplification of the above detailed approach with 34 groups of PAHs will be highly site dependent
because PAH contamination can arise from numerous sources with highly different PAH compositions
(Burgess et al. 2003a). Additionally, fate and transport processes, and biological weathering can greatly
affect the composition of the PAH mixtures in sediments. Depending on the characteristics of particular
sites, simplification, possibly using a subset of the 34 groups of PAHs, in the above process might be
possible.
43
-------
Table 4-3. Example calculation of pore water toxicity arid pore water RGs for a sediment with a PAH mixture as the known toxicants.
Measured Concentration 1.704% = 1/58.68
Sediment
Pore Water
(Qree) 3
Aqueous
Solubilityb
Narcosis
FCV/SCV
Pore Water
Toxic Unitsc
Cfree:PWRG
Pore water
RG Toxic
Units
Site-
Specific
Log K0cc
Bulk
Sediment
Cs:PWWG
PAH
Hg/g(dw)
Hg/L
Hg/L
Hg/L
Hg/L
L/kg (OC)
Hg/g (dw)
Naphthalene
3.33
2.89
30,995
193.5
0.015
0.049
0.0003
4.154
0.057
Cl-Naphthalenes
1.07
2.13
81.69
0.026
0.036
0.0004
3.794
0.018
C2-Naphthalenes
2.57
26.8 J
30.24
0.886
0.457
0.0151
3.074
0.044
C3-Naphthalenes
1.94
35.5 J
11.10
3.198
0.605
0.0545
2.830
0.033
C4-Naphthalenes
1.01
18.5 J
4.048
4.570
0.315
0.0779
2.830
0.017
Acenaphthylene
1.60
8.36
16,314
306.9
0.027
0.142
0.0005
3.375
0.027
Acenaphthene
6.69
75.1
3,800
55.85
1.345
1.280
0.0229
3.042
0.114
Fluorene
4.49
25.4
1,900
39.30
0.646
0.433
0.0110
3.340
0.077
Cl-Fluorenes
1.69
17
13.99
1.215
0.290
0.0207
3.090
0.029
C2-Fluorenes
1.38
15 U
5.305
0.707
0.128
0.0241
3.357
0.024
C3-Fluorenes
0.673
0.343 U
1.916
0.045
0.003
0.0015
4.686
0.011
Phenanthrene
19.5
60.6
1,100
19.13
3.168
1.033
0.0540
3.600
0.332
Anthracene
8.33
15.2
45.0
20.73
0.733
0.259
0.0125
3.831
0.142
Cl-Phenanthrenes/Anthracenes
7.13
37.8
7.436
5.083
0.644
0.0866
3.368
0.122
C2-Phenanthrenes/Anthracenes
3.94
33.8
3.199
10.566
0.576
0.1801
3.159
0.067
C3-Phenanthrenes/Anthracenes
1.76
15.7
1.256
12.500
0.268
0.2130
3.142
0.030
C4-Phenanthrenes/Anthracenes
0.912
1.0 U
0.5594
0.447
0.009
0.0152
4.354
0.016
Fluoranthene
20.2
19.8
239.9
7.109
2.785
0.337
0.0475
4.101
0.344
Pyrene
17.2
16.9
131.9
10.11
1.672
0.288
0.0285
4.100
0.293
Cl-Fluoranthenes/Pyrenes
10.1
11.4
4.887
2.333
0.194
0.0398
4.040
0.172
Benz[a]anthracene
9.68
1.84
11.0
2.227
0.826
0.031
0.0141
4.814
0.165
Chrysene
8.35
1.45
2.0
2.042
0.710
0.025
0.0121
4.853
0.142
Cl-Benzanthracenes/Chrysenes
4.37
1.27
0.8557
1.484
0.022
0.0253
4.629
0.074
C2-Benzanthracenes/Chrysenes
2.08
0.0138 U
0.4827
0.007
0.000
0.0002
6.572
0.035
C3-Benzanthracenes/Chrysenes
1.32
0.0174 U
0.1675
0.026
0.000
0.0009
6.274
0.022
44
-------
C4-Benzanthracenes/Chrysenes
0.527
0.0235 U
0.0706
0.083
0.000
0.0028
5.744
0.009
Benzo[b]fluoranthene
6.95 J
0.448 J
1.501
0.6774
0.661
0.008
0.0113
5.283
0.118
Benzo[k]fluoranthene
8.35
0.53
0.7999
0.6415
0.826
0.009
0.0141
5.290
0.142
Benzo[a]pyrene
10.9
0.422 J
3.810
0.9573
0.441
0.007
0.0075
5.505
0.186
Perylene
2.93
0.175
0.4012
0.9008
0.194
0.003
0.0033
5.316
0.050
Benzo[e]pyrene
5.69
0.387
4.012
0.9008
0.430
0.007
0.0073
5.260
0.097
lndeno[l,2,3-cd]pyrene
6.39
0.12 J
0.2750
0.436
0.002
0.0074
5.819
0.109
Dibenz[a,h]anthracene
1.82
0.055 J
0.6012
0.2825
0.195
0.001
0.0033
5.612
0.031
Benzo[ghi]perylene
6.40
0.173 J
0.2600
0.4391
0.394
0.003
0.0067
5.661
0.109
Total Organic Carbon
8.08%
Total
191.27
.
.
.
58.681
.
1.0
.
3.260
a U - undetected; value represents detection limit, J - estimated value. b (Mackay et al. 1992; US-EPA 2003c). c Yi detection limit used for non-detect values.
45
-------
4.3.2.3 Derivation of Cs;pwrg values for a Sediment with Other Contaminant Mixtures
With the exception of PAHs as discussed in Section 4.3.2.2, pore water RGs for other chemicals
are given for individual chemicals. However, many Superfund sites have mixtures of chemicals which
may warrant consideration of the toxicity of those mixtures as it may differ from that of the individual
compounds. As a general rule, the expectation is that chemicals with dissimilar toxicological
mechanisms will act independently, and can therefore be assessed using pore water RG values on a
chemical by chemical basis. However, those that share a toxicological mechanism, are generally
expected to show additive toxicity and are therefore likely to be more rigorously addressed using a
mixture approach. PAHs are one such group, and the PAH ESB guidance uses an additive toxicity
concept to address the potency of PAH mixtures (US-EPA 2003c). While explicit mixture guidance has
not been developed for other mixtures likely to be encountered in sediments, there are additional
chemical groups that share a common toxicological mechanism to benthic invertebrates and may co-
occur; examples include chlorobenzenes, which contribute to narcotic effects; chlorinated cyclodiene
pesticides (aldrin, dieldrin, endrin, heptachlor, chlordane, endosulfan) that act through the gamma-
aminobutyric acid (GABA) receptor-chloride channel complex; the various isomers and metabolites of
DDT (DDE, DDD); and acetylcholine esterase inhibiting organophosphate pesticides (e.g., chlorpyrifos,
malathion, diazinon, and many others). The key in deciding to apply a mixture approach lies in 1)
having reasonable scientific evidence that the group of chemicals share a toxicological mechanism; and
2) that there are enough chemicals in the group present at sufficient concentrations to warrant
evaluating them together.
Speaking again in general terms, the degree of uncertainty introduced by failing to consider mixture
effects is influenced largely by the number of chemicals involved, with the uncertainty generally
increasing with larger numbers of similarly acting chemicals in the mixture. As a simple example,
consider dieldrin and endrin, which have similar toxic action and whose toxicities would likely be
additive. If these chemicals were only evaluated separately, then a worst-case scenario might be if both
were at 0.9 PWTU, with an expected combined potency of 1.8 PWTU. While this would be an exposure
higher than intended by a pore water RG of PWTU = 1, the magnitude of this difference is small
compared to the same scenario for PAH mixtures (with 34 components) wherein the aggregate TU could
be as high as 34 chemicals x 0.9 PWTU = 30.6 TU.
In real world mixtures, it is probable that the contributions of individual chemicals to overall mixture
toxicity will vary, and would not all be right near the pore water RG. Using the dieldrin/endrin example,
if the PWTU for dieldrin was generally 20% or less of the PWTU for endrin, then the magnitude of the
resulting uncertainty would be small even if mixture effects were ignored. In the example in Table 4-3,
the highest PWTU for a single PAH was 12.5 PWTU for C3-Phenanthrenes/Anthracenes, compared to the
summed PWTU of 56.68, which would be about a 5-fold underprotection if mixture effects were ignored
and all PAHs were compared to their pore water RG values individually.
The potential importance of considering mixtures can be evaluated from the passive sampler data
by comparing the assessment conclusions if pore water RGs are applied individually or within a mixture
approach. A simple sensitivity assessment can inform the assessor as to the degree to which ignoring
mixture effects would influence the assessment. Again, the number of chemicals involved is likely to be
46
-------
a key factor. So as an example, a site contaminated with a whole suite of chlorobenzene compounds
might be a more likely candidate for a mixture approach than one contaminated with just a couple of
chlorobenzenes.
If a mixture approach is chosen, the approach is parallel to that shown in section 4.3.2.2. PWTUs are
calculated individually for each component of the mixture (Equation 4-3), then summed across the
mixture (Equation 4-4). The pore water RGs for each chemical in the mixture are computed by dividing
their Cfree values by the PWTUMi>cture value. Concentrations in the bulk sediment are found by dividing
the bulk concentrations by PWTU Mi>cture value.
4.4 Suggested Methodology for Using Passive Sampler Measurements to Develop PWRGs
4.4.1 ESB Screening Approach
The suggested approach for implementing the use of passive sampling measurements at
contaminated sediment sites to develop PWRGs is discussed in Sections 4.1 and 4.3.2.1, and is outlined
below. The approach takes bulk concentrations of the COCs in the surficial sediments developed in the
Rl and compares them to EPA's ESBs. For surface sediments with concentrations less than EPA's ESBs,
unacceptable risks to benthic organisms do not exist. For surface sediments with concentrations greater
than the ESBs, passive sampling measurements are performed in order to refine risks at those locations.
With the Cfree measurements in the sediment pore water, the total TUs in the sediments are determined,
and for samples/locations where the total TUs exceed 1.0, the potential for unacceptable risks to
benthic organisms is noted. The approach below follows Superfund's eight-step ecological risk
assessment guidance (US-EPA 1997). Depending upon data availability and when appropriate
measurements are made, comparison of bulk concentrations of COCs to ESBs might occur in the
Screening Level Ecological Risk Assessment or in the Site Investigation and Data Analysis steps of the
eight-step ecological risk assessment procedure.
Screening Level Characterization of the Nature and Extent of Contamination
1) Measure f0c and Cs for all COCs (ng/kg-dw) in surficial sediments across the site
2) Compute CSOc (|ig/kg-OC) for all COCs
Screening Level Ecological Risk Assessment
3) Compute Toxic Units (TUs) for COCs
a. For single toxicant case, TU = CSOc divided by ESB for the COC
b. For mixture of toxicants, sum TUs for all COCs where TUj = CSOc,i/ESBi for each COC
Problem Formulation
Develop CSM, exposure pathways, and assessment endpoints
Study Design and DQO Process
Develop Work Plan (WP) and Sampling and Analysis Plan (SAP) in support of CSM and data needs
Site Investigation and Data Analysis
4) Passively sample surface sediments where total TUs > 1.0
5) Derive Cfree and K0c values for surface sediments with total TUs > 1.0
Risk Characterization
47
-------
Baseline Ecological Risk Assessment
6) Compute Toxic Units (TUs) for COCs
a. For single toxicant case, PWTU = Cfree/FCV
b. For mixture of toxicants, for each COC in the mixture:
Compute pore water TU for each COC, PWTUi = Cfree,i/FCVi
Compute total mixture pore water TUs, PWTUMixture = IPWTUi
7) For locations where:
c. Total PWTUs < 1.0, little potential for risk to benthic organisms.
d. Total PWTUs > 1.0, unacceptable risks to benthic organisms indicated, proceed to
Remedial Goal Development
Remedial Goal Development
8a) If pore water RGs are expressed on Cfree basis (Cfree:PWRG M-g/L):
a. For single toxicant case, pore water RG is
Cfree:PWRG —
FCV
b. For mixture of toxicants:
Derive site-specific composition of the mixture
Compute pore water RGs for each COCs,
Cfree:PWRG,i —
FCVi X PWTUi / PWTUMixture
8b) If pore water RGs are expressed on bulk sediment basis (CS;pwrg M-g/kg dry weight):
Derive site specific f0css and K0css values for each COC
a. For single toxicant:
Pore water RG for COC:
Cs:PWRG = Kocss X foC:SS X
Cfree:PWRG
where Cfree:PWRG = FCV
e. For mixture of toxicants:
Derive site-specific composition of the mixture
Pore water RG for each COC:
Cs;PWRG,i = KoC:SS,i X foC:SS,i X
Cfree:PWRG,i
where Cfree:PWRG,i = FCVi x PWTUi / PWTUMixture
Sum CS;pwRG,i for all mixture components to provide total bulk concentration of
mixture
Cs;PWRG,Mixture = 2Cs;PWRG,i
8c) If pore water RGs are expressed on organic carbon basis (Csocpwrg M-g/kg organic-carbon):
Derive site specific Kocss values for each COC
b. For single toxicant:
Pore water RG for COC:
Csocpwrg = Kocss X Cfree:PWRG
where Cfree:PWRG = FCV
f. For mixture of toxicants:
Derive site-specific composition of the mixture
Pore water RG for each COC:
Csocpwrg,i = Kocss,i X
Cfree:PWRG,i
where Cfree:PWRG,i= FCVi x PWTUi / PWTU Mixture
48
-------
Sum Csocpwrgj for all mixture components to provide total concentration of mixture
on an organic carbon basis in the sediment
CsOCPWRG,Mixture = £CsOC:PWRG,i
49
-------
Section 5
Use of Passive Samplers, Toxicity Testing Results, and Pore Water RGs
To successfully use the methodology in Section 4 for deriving PWRGs for the protection of benthic
organisms from direct toxicity within the RI/FS process, knowledge of passive sampling, sediment
toxicity testing, and the interpretation of the results from both techniques is essential. We anticipate
that many of the users of this document will be fairly knowledgeable on the passive sampling
measurement technique and its resulting data, and less knowledgeable on the sediment toxicity testing
and its resulting data. As part of the ecological risk assessment, sediment toxicity tests are often
performed on sediment samples in order to characterize risks from the contaminants on benthic
organisms at the site.
This section provides a basic primer on the toxicity data from sediment toxicity tests, followed by
discussion of illustrative toxicity testing results that you might observe at your site. Successful
development of PWRGs allows the responses from sediment toxicity testing to be explained in relation
to (or are consistent with) the measured Cfree values of the COCs in the sediment pore water. In this
comparison, Cfree values could be expressed on toxic units or concentration basis depending your site's
COCs. Consistency means that samples non-toxic in the sediment toxicity tests are predicted to be non-
toxic based upon the measured Cfree values of the COCs and vice-versa. When your data are consistent
(i.e., no false positives and no false negatives), one can be reasonably assured that the contaminants
causing toxicity to benthic organisms have been correctly identified and that the developed pore water
RGs for the contaminants will be protective of the benthic organisms at the site. Further, you have
established a causal linkage between the CERCLA COCs at your site and the effects observed in the
sediment toxicity tests on the sediments from your site.
5.1 Approaches for Aligning Pore Water RGs and Sediment Toxicity Testing Results
5.1.1 Exposure Response
When sediment toxicity tests are performed, each test provides one or more effect endpoints (e.g.,
survival, growth, and/or reproduction), for the tested sample. If one had a set of sediment samples that
contained the complete range of contaminant concentrations, one would expect a sigmoidal shape
response curve where at lower concentrations of the contaminant, no effects are observed; at high
concentrations of the contaminant, unacceptable effects are observed on all test organisms; and at
intermediate concentrations, a graded response is observed. Plotting of the individual test results
would result in an exposure-response curve as shown in Figure 5-1 where toxic units of the contaminant
are on x-axis and endpoint responses are on the y-axis. Note, Figure 5-1 is idealized for illustrative
purposes! For a variety of reasons, the uniform spacing of the contaminant concentrations and
uniformity in biological response will be rarely observed in practice.
50
-------
100
80
c
o
Q.
"D
C
60
LU 40
I»
O
•eoooocp
10 100
0.01
0.1
1
Toxic Units
Figure 5-1. Illustrative idealized dose-response curve. Circles are test results for
individual sediment samples and a smooth line has been fitted to the individual test
results. The EC50 is shown by the dashed line.
In Figure 5-1, the midpoint in the response curve is centered on 1.0 toxic unit and for sediment
samples with concentrations of the contaminant above and below its FCV, their responses in toxic units
would reside on the line above and below the 1.0 toxic units midpoint, respectively. As discussed in the
Introduction, equilibrium partitioning theory argues that the same exposure-response curve would be
observed with a water only exposure of the same organism. In the Figure 5-1, the EC50 (effect
concentration at 50% response for the endpoint of interest) is also shown. For survival, the EC50 is
equivalent to LC50 (lethal concentration for 50% survival of the test organisms).
5.1.2 Exposure Response Curves Observed at Sediment Sites
At Superfund sites, sediment toxicity testing data typically do not imitate perfectly the illustrative
dose-response curve in Figure 5-1. Causes for deviations from the illustrative curve include:
a) Not having a set of sediment samples with a broad range of contaminant concentrations;
b) Having sediment samples whose contaminant concentrations miss important portions of the
overall exposure-response curve;
c) Presence of multiple COCs in varying proportions across the site, such that the exposure-
response curve for a single COC is confounded by the effects of another COC; and
d) Data quality issues, e.g., QC issues with toxicity tests and/or passive sampling
measurements (see Section 5.3. Method Uncertainties).
Data from 28-day sediment toxicity tests with Hyalella azteca on Hudson River sediments
contaminated with PAHs are provided in Figure 5-2 (Kreitinger et al. 2007). These data illustrate a
number of important points that might appear at your sites.
First, survival of the H. azteca when plotted against the concentration of the PAHs in the sediments on a
dry weight (Figure 5-2A) does not follow the sigmoidal shape pattern. We expect this behavior because
51
-------
the total concentration of the PAHs is poor measure of the bioavailable portion of the PAHs in the
sediments. Second, survival of the H. azteca when plotted against the concentration of the PAHs in the
sediments on an organic carbon basis (Figure 5-2B) does not follow the sigmoidal shape pattern.
Depending upon the types and consistency of the organic carbon in the sediments at your site, sigmoidal
dose-response behavior might or might not occur with organic carbon normalization. For the Hudson
River example (Figure 5-2B), organic carbon normalization does not completely account for the
bioavailability of the PAHs in the sediments. Third, when Cfree concentrations in sediment pore water are
expressed on a toxic units basis, the data follow the sigmoidal shape pattern shown in Figure 5-1, that is,
at elevated survival, there are low TUs of contaminants and at low survival, there are high TUs of
contaminants (Figure 5-2C). Kreitinger et al. (Kreitinger et al. 2007) computed the total toxic units using
EPA's narcosis FCVs (Table 3-1 (US-EPA 2003c)) and their measured concentrations in the sediment pore
water for the 18 parent PAHs and 16 alkylated PAH groups (Figure 5-2C). Fourth, in Figure 5-2C, non-
toxic and toxic samples have toxic units ranging from 0.1 to 18 and from 110 to 310, respectively. The
break point between non-toxic and toxic sediment samples occurs somewhere between 18 and 110
toxic units, and not at 1.0 toxic unit. The break between non-toxic and toxic samples at toxic units
different from 1.0 is expected and data from your site will likely have this behavior as well. The reason
for the departure from 1.0 toxic units for the break between non-toxic and toxic samples is that H.
azteca are less sensitive than the (theoretical) species driving the FCV. As discussed in Section 3.4, the
sensitivity of the test organism itself, in all likelihood, does not reside at the 5th percentile value of the
SSD, but rather at a higher percentile in the SSD. In addition, the FCV represents a very low level of
chronic effect (rather than 50% effect), and includes consideration of all endpoints, lethal and sublethal).
Coherence of the site exposure-response curve for the sediment toxicity data can be evaluated by
calculating the expected toxicity (based on other data) for the tested species and COCs. In Appendix A,
the EC50 for H. azteca has been derived from water-only toxicity testing of PAHs, and for the data in
Figure 5-2C, toxicity data for H. azteca derived from the literature agrees fairly well with measured
toxicity, i.e., the literature EC50 resides between the toxic and nontoxic samples. Finally, if one uses the
toxicity of the PAHs to H. azteca to convert the concentrations in sediment pore water to toxic units, the
28-day H. azteca chronic survival data replots as shown in Figure 5-2D where the literature EC50 for
PAHs and EC50 from the sediment toxicity tests contaminated with PAHs both reside at approximately
1.0 toxic unit. The agreement illustrated in Figures 5-2C and 5-2D is the same and the only difference is
the labeling of the x-axis.
Figures 5-2C illustrates the case where a sigmoidal dose-response is obtained; the EC50 for the
tested sediments does not reside at 1.0 TU; and the EC50 of the contaminants from toxicity tests with
the pure chemical and the test species resides between the non-toxic and toxic sediment samples.
When the EC50 of the pure chemical and test species aligns with the break between non-toxic and toxic
samples, and there is a sigmoidal dose-response, consistency between the sediment toxicity testing data
and toxicity of the COCs has been shown.
52
-------
(0
>
3
(/)
C
(0
<1>
rxy\ a a
AAA,
A^
I
:
A
B
;
— ^
N A
,—
—¦ A1 ¦ A'
.A'
10
100
1,000 10,000
Total PAH concentration
in sediment (mg/kg-dw)
0.1
1
10
100
1,000
Total PAH concentration
in sediment (mg/g-OC)
§
3
(/)
C
(0
<1>
A ArA
AZAa 1
1
a Af"A A
^ 4f
: I
1
:
1
1
j 1
1
1
i 1
C
1
1
D
: 1
—
—
1
1 1 1 1 114
¦ A'
! 1 1 1 1111
1 1 1 111111
: 1
1 LJjilil-
AN
A 1 111 hi
0.01 0.1 1.0 10 100 0.001 0.01 0.1
1
10 100
Toxic Units in sediment pore water Toxic Units in sediment pore water
EPA's ESB methodology using H. azteca toxicity value for PAHs
Figure 5-2. Hyalella azteca 28-d survival data for sediment toxicity tests with sediments from the
Hudson River at Hudson, NY (Kreitinger et al. 2007) plotted against A) concentration of PAHs in sediment
(mg/kg dw), B) concentration of PAHs in sediment on an organic carbon basis (mg/g OC), C) toxic units of
PAHs in sediment pore water, and D) toxic units adjusted to the sensitivity of the H. azteca. Toxic units
of measured Cfree values in sediment pore water were derived using EPA's ESB methodology discussed in
Section 4.3.2.2 (Figure C) and in Figure D, adjusted using the toxicity for H. azteca. Pink circle symbols
are the sediment toxicity test controls. The — and •••• lines are the mean and 95% confidence levels for
the EC50 for chronic survival derived from the water-only toxicity testing of PAHs.
53
-------
Another common occurrence with sediment toxicity testing data is the attainment of incomplete
dose-response curves. Figure 5-3 illustrates a case where all of the samples are nontoxic. The data in
Figure 5-3 are for sediments from the Hudson River at Troy, NY (USA) and most of the data are in good
agreement with the sediment test controls (considered nontoxic) (Kreitinger et al. 2007). Two field
reference sediments were include in the samples tested and these were collected near the site from
locations having low PAH concentrations and considered not affected by the PAH contamination
(Kreitinger et al. 2007). These data illustrate the following points:
First, when toxicity data for a set of sediment samples has similar results for all samples, e.g., all samples
are non-toxic or all samples are toxic, one needs to determine/obtain from the literature the toxicity of
the COCs from pure chemical testing. With the toxicity of the COCs, you should determine if the
sediment samples have the proper orientation relative to the COCs' EC50 from pure chemical testing. In
Figure 5-3, the toxicity value for PAHs derived in Appendix A is consistent with the testing data, i.e., the
TUs for the tested samples are less than the EC50 from water-only toxicity testing of PAHs. This
consistency conforms to the sigmoidal dose-response pattern and we conclude that the data in Figure 5-
3 are from the lower end of the sigmoidal dose-response curve where no effects are observed. Second,
if one has a set of sediment samples where all samples are toxic, the TUs for the tested samples should
be greater than the EC50 from water-only toxicity testing of the COCs. Although the data are consistent
with a sigmoidal dose-response curve, the consistency does not absolutely insure that contaminants
causing the toxicity in the sediment samples are the identified COCs. There may be unidentified
contaminants in the sediment causing the toxicity in the sediments.
CO
>
>
i_
3
(/)
C
CO
100
80 -[
60 -E
40
' r\n,
c A A
mA
I A
: s
- ^
I
A
I
I
I
I
I
I
;
I
I
I
0.01 0.1
1.0
10
100
Toxic Units in Sediment Pore Water
EPA's ESB methodology
Figure 5-3. Measured survival data (± one standard deviation) for Hyalella azteca in 28-d sediment
toxicity tests with sediments from the Hudson River at Troy, NY (Kreitinger et al. 2007). Toxic units of
measured Cfree values in sediment pore water were derived using EPA's ESB methodology (discussed in
Section 4.3.2.2). Downwards triangle symbols are for field reference locations for the site and pink circle
symbols are the sediment toxicity test controls. The controls are considered non-toxic, have zero toxic
units, and are arbitrarily placed on the graph. The and lines are the estimated mean and 95%
confidence levels for the EC50 for chronic survival derived from the water-only toxicity testing of PAHs
(Appendix A).
54
-------
Like the data in Figure 5-2, the data in Figure 5-3 conforms to the sigmoidal dose-response pattern,
i.e., the non-toxic sediment samples have the proper orientation relative to the EC50 of the COCs from
water-only toxicity testing of the chemicals.
A broader comparison of PAH toxicities has been performed by Hawthorne et al. (Hawthorne et al.
2007) where 97 sediment samples from six manufactured-gas plants and two aluminum smelter
sites were investigated. Each sediment sample was toxicity tested using Hyalella azteca and subjected
to passive sampling measurement. For these sediments, 28-d survival data for Hyalella azteca and the
estimated sediment toxicity, using EPA's PAH FCVs along with their measured concentrations in the
sediment pore water for the 18 parent PAHs and 16 alkylated PAH groups, are shown in Figure 5-4. Like
the data of Kreitinger et al. (Kreitinger et al. 2007) in Figure 5-2, these data follow the sigmoidal shape
pattern, i.e., at low predicted toxicity, high survival and at high predicted toxicity, low survival (Figure 5-
4). Similar to the data of Kreitinger et al. (Kreitinger et al. 2007) above (Figures 5-2 and 5-3), the
Hawthorne et al. (Hawthorne et al. 2007) toxicity data illustrates the case where the sediment test
organism is less sensitive than the 5th percentile derived from the SSD for the PAHs (US-EPA 2003c). The
toxicity value for PAHs derived in Appendix A is consistent with the testing data of Hawthorne et al.
(Hawthorne et al. 2007) (Figure 5-4).
These data illustrate some of the variability one might observe at their site. Across these eight sites,
there were only one or two potential outliers in the entire dataset. One of the samples with unusual
toxicity was almost pure sand with very low organic carbon content, and the poor survival of the test
organisms might have been related to the poor nutritional content of a sediment (Hawthorne et al.
2007). Like the data in Figures 5-2 and 5-3, the EC50 does not reside at 1.0 TU for the COCs in the
sediments tested; the non-toxic sediment samples have the proper orientation relative to the EC50 of
the PAHs derived from water-only toxicity testing of PAHs; and the data conforms to the sigmoidal dose-
response pattern. The data of Hawthorne et al. (Hawthorne et al. 2007) can be replotted using toxicities
predicted using PAH toxicity value derived for H. azteca in Appendix A instead of the EPA's FCV for PAHs.
When replotted, the predicted toxicities shift such that EC50 in the toxicity data align around 1.0 TU for
the data of Hawthorne et al. (Hawthorne et al. 2007) (Figure 5.5), and non-toxic sediment samples have
the proper orientation relative to the EC50 of the PAH mixture, and the data conform to the sigmoidal
dose-response pattern. The EC50 from Appendix A resides at ca. 20% survival endpoints in the probit
regression and the H. azteca EC50 estimate comes from entirely independent data set.
55
-------
85% Survival
120
100 -
OTO
15% Survival
80 -
<3
>
>
k.
3
CO
60 -
40 -
A Toxic
O Non-Toxic
— Probit Regression
20 -
0.0001
0.001
0.01
0.1
1
10
100
1000
Predicted Toxic Units using EPA's ESB methodology
Figure 5-4. Measured toxicity survival data for Hyalella azteca in 28-d sediment toxicity tests with 97
sediments from six manufactured-gas plants and two aluminum smelter sites, and toxicity
estimated from the concentrations of PAHs in the sediment pore water (Hawthorne et al. 2007).
The Hyalella azteca EC50 (short dash line) was derived from the water-only toxicity testing data
(Appendix A). The solid line is the probit regression fit of the data and 15% and 85% survivals
lines (long dash) are from the probit regression. Adapted with permission from Hawthorne SB,
Azzolina NA, Neuhauser EF, Kreitinger JP (2007). Predicting bioavailability of sediment polycyclic
aromatic hydrocarbons to Hyalella azteca using equilibrium partitioning, supercritical fluid
extraction, and pore water concentrations. Environmental Science & Technology 41:6297-6304.
Copyright 2007 American Chemical Society.
56
-------
85% Survival
120
15% Survival
100 -
OTO
80 -
<3
>
>
k.
3
CO
60 -
40 -
O Non-Toxic
A Toxic
Probit Regression
20 -
0.00001
0.0001
0.001
0.01
0.1
1
10
100
Predicted Toxic Units using H. azteca toxicity for PAHs
Figure 5-5. Measured toxicity survival data for Hyalella azteca in 28-d sediment toxicity tests with 97
sediments from six manufactured-gas plants and two aluminum smelter sites, and toxicity
estimated from the concentrations of PAHs in the sediment pore water (Hawthorne et al. 2007).
The Hyalella azteca EC50 ( ) was derived from the water-only toxicity testing data (Appendix
A). The solid line is the probit regression fit of the data and 15% and 85% survivals lines (long
dash) are from the probit regression. Adapted with permission from Hawthorne SB, Azzolina
NA, Neuhauser EF, Kreitinger JP (2007). Predicting bioavailability of sediment polycyclic aromatic
hydrocarbons to Hyalella azteca using equilibrium partitioning, supercritical fluid extraction, and
pore water concentrations. Environmental Science & Technology 41:6297-6304. Copyright 2007
American Chemical Society.
5.2.3 Approaches for Aligning Sediment Toxicity Results with Pore Water RGs
In general, a weight of evidence (WOE) approach should be used at your site to evaluate the
alignment of sediment toxicity testing results with the pore water RGs developed for your site. The
WOE approach will vary across sites because the amount of confidence needed in the agreement
between the toxicity testing results and pore water RGs is highly dependent upon the significance of the
decision(s) based upon the comparison. Some factors that one should consider include:
• How well does the sediment toxicity data conform to the expected sigmoidal dose-response
curve?
• Does the EC50 from water-only toxicity testing of the COCs for the species used in the
toxicity tests reasonability define the break between non-toxic and toxic sediment samples
on the sigmoidal dose-response curve?
• Are all sediment toxicity testing results explainable based on their positions on the sigmoidal
dose-response curve?
• Do outliers exists in the dose-response plot, and if so, are they explainable?
57
-------
In considering the above factors:
• Was the nature and variability of the f0c and K0c values adequately defined for the site?
Was this information used in the calculation of the toxic units for the COCs in the toxicity
tested sediment samples?
• When the COCs are a mixture, e.g., PAHs, were changes in the composition of mixture
across the site accounted for in the calculation of the toxic units for the COCs in the toxicity
tested sediment samples?
The overall approach for aligning sediment toxicity testing results with the pore water RGs
developed for your site is the comparison of the sediment toxicity testing results with the pore water
concentration measurements (or equivalently, computed toxic units). These data need to form a
sigmoidal dose-response curve and the break between non-toxic and toxic samples should be consistent
with the EC50 from water-only testing of the COCs with the toxicity testing species. When these
conditions exist, one has shown consistency between the toxicity testing results and the pore water RGs.
When consistency exists in the data, one can be reasonably assured that the contaminants causing
toxicity to benthic organisms have been correctly identified and that the developed pore water RGs for
the contaminants will be protective of the benthic organisms at the site. Clearly, if the passive sampling
measurements are for chemicals that have little or no role in the overall toxicity of the sediments,
plotting of the sediment toxicity measurements against the passive sampler measurements should
enable detection of the issue.
It is important to note that the evaluation of exposure-response is primarily intended to establish
that the inferred risk to benthic organisms (as indexed by sediment toxicity) is attributable to the COC(s)
that are the focus of the RG development. In doing so, the exposure response evaluation is often
focused on an endpoint, (e.g., survival versus reproduction), a level of effect (e.g., 50% mortality in a less
than life-cycle test) or organism that is not as sensitive as the level of protection intended by the FCV for
the chemical. The importance of these differences must be considered in characterizing the level of risk
posed by site contamination and the intended benefits of different RGs.
5.2 Method Uncertainties and Confounding Factors
Standardized operating procedures for passive sampling measurements are not currently available
although guidance is available (US-EPA/SERDP/ESTCP 2017). The technique has evolved over the past
decade (Ghosh et al. 2014; Greenberg et al. 2014; Lydy et al. 2014; Mayer et al. 2014; Parkerton and
Maruya 2014), and there are a host of issues that could arise with the passive sampling technique.
These issues including inaccurate Kp0iVmer partition coefficients, nonattainment of equilibrium conditions
when equilibration techniques are used, performance reference compounds that do not accurately
match the partitioning behavior of the toxicants, inconsistencies in polymer batches resulting in varying
partition coefficients, and detection limit issues. Some simple checks on the passive sampling
measurements could include "Are the freely dissolved concentrations less than the chemicals' aqueous
solubilities?" and "Are the freely dissolved concentrations estimated using generic Kocs greater than
58
-------
those measured by passive sampling?" Additional checks and data evaluation procedures for the quality
of passive sampling measurements will be provided in a forthcoming document, and the readers should
consult this document (US-EPA/SERDP/ESTCP 2017).
In addition to the potential uncertainties associated with passive sampling measurements, sediment
toxicity tests are performed with live organisms and these organisms are obtained from in-house
cultures and facilities specializing in culturing the test organisms. As with any living organism testing
system, occasional unusual results occur even though test controls are within performance
specifications. Different cultures of live organisms might have slightly different sensitivities to the COCs
and these sensitivity differences could slight shift the sigmoidal dose-response curve among testing
laboratories. Some simple checks on the toxicity testing results could include comparing controls with
controls from prior testing data from the testing facility and determining if organism source deviated
from the testing laboratory's normal practices. The sensitivities of the organisms to reference toxicants
and the site's COCs might also be evaluated. Clearly, the sediment toxicity test results should also meet
the criteria specified in their testing protocol, and the reader should consult these documents for these
criteria (US-EPA 1994; US-EPA 2000; US-EPA 2002a; US-EPA 2002b; US-EPA and US-ACE 2001).
There are situations where conformity with the expected sigmoidal dose-response curve might not
appear or the sigmoidal dose-response is messy. One cause is the presence of other unidentified
contaminants in the sediment samples. Unidentified contaminants could cause sediment samples to be
much more toxic than estimated from the measured concentrations in sediment pore water for the
identified contaminants. The unidentified contaminants could be additive with or exert toxicity
independent of the identified contaminants in the samples. Sediment samples with unidentified
contaminants will often appear as outliers from the expected sigmoidal dose-response curve. One
needs to understand why outliers exist in the site data and further, their influence upon the overall
conceptual model of sediment toxicity at the site, the site's CSM, and ultimately, remedy selection. Are
the outlier samples located in one portion of the site? Are the outlier samples scattered across the site?
Do the outlier samples have unusual composition, e.g., ammonia, sulfides, metals, high in oils/greases,
tars, wood chips or sand, relative to the other samples at the site? Are the outliers an artifact of the
quality of the toxicity testing and/or Cfree data? Performing sediment Toxicity Identification Evaluations
(TIEs) (US-EPA 2007b) on the outliers might be in order depending upon the site, location of the samples
within the site, and/or the cost of the remedial action. The importance of understanding why the
outliers exist cannot be under emphasized because the outliers could influence the remedy selection
and the success of the remedy.
When multiple contaminants exist at your site and the contaminants are not additive toxicologically,
plotting of toxicity testing endpoints against the predicted toxicities based upon the concentrations
measured in the pore water should be performed for each contaminant separately. Separate plots
would enable a better evaluation of the dose-response for the individual contaminants. At large sites
where different contaminants are at different locations within the site, i.e., contaminant X is in locations
A, and contaminant Y is in locations B, the data should be plotted separately to determine if a sigmoidal
dose-response exists for each contaminant at its locations within the site. Plotting of toxicity data for
both contaminants in one plot would simply confuse the interpretation of the data. Alternatively,
exposure/response data can be plotted with the exposure axis being the maximum TU from any COC (or
59
-------
additive mixture thereof). This approach may help identify cases where a sample showing toxicity at low
concentrations of one COC might be explained by a high exposure to another COC, something that can
be hard to evaluate in plots that show only one COC at a time.
At some sediment sites, PAHs reside in an oily matrix in the sediment, and the oily matrix can
contain high levels of aliphatic hydrocarbons (e.g., alkanes and cycloparaffins). Aliphatic hydrocarbons
are the major components of lubricants and greases, and are present in crude oil and numerous refined
petroleum products. A confounding issue with PAHs might occur when high levels of aliphatic
hydrocarbons are present in the sediments. For aliphatic hydrocarbons, their mechanism of toxicity for
filter feeding benthic invertebrates, such as the freshwater amphipod Hyalella azteca, can stem from a
physical effect, such as fouling of respiratory surfaces by the oil phase (Mount et al. 2015, unpublished
results). A risk assessment on mineral oils by Verbruggen (Verbruggen 2004) reported EC50s of 500
(90% CI: 460-550) mg/kg-dw and 1800 (850-4000) mg/kg-dw for mortality, and 67 (0-180) mg/kg-dw and
130 (19-240) mg/kg-dw for growth for 10-d Hyalella azteca toxicity tests with two different oil mixtures.
In most cases, the composition of mineral oil will not be highly reflective of the composition of the
aliphatic hydrocarbons at your site. Consequently, the levels reported by Verbruggen (Verbruggen
2004) should be taken as indicative. Measurements of gasoline range organics (GROs) and diesel range
organics (DROs), residual range organic (RROs) will provide some indication of the levels of aliphatic
hydrocarbons, but these measurements can include aromatic hydrocarbons depending upon the
analytical method and will complicate data interpretation. When toxicity testing data are substantially
more toxic than predicted based upon the PAH content, we suggest you consider possibility of the
aliphatic hydrocarbons as the contaminants when all other contaminant classes, e.g., pesticides, metals,
..., have been eliminated as causes of toxicity in the sediment samples of interest.
60
-------
Section 6
APPENDIX A
DERIVATION OF 10- and 28-DAY PAH EFFECT CONCENTRATIONS FOR
HYALELLAAZTECA FOR PURPOSES OF EVALUATING EXPOSURE-RESPONSE IN
SEDIMENT TESTS
Prepared by:
David R. Mount and J. Russell Hockett
U.S. EPA Office of Research and Development
Mid-Continent Ecology Division
Duluth, MN
1.0 Background and Purpose
This appendix was developed to provide an estimated response value for the amphipod, Hyalella
azteca, when exposed to polycyclic aromatic hydrocarbons (PAHs) in 28-d sediment toxicity
tests. These values are not intended to be cleanup values or otherwise used as regulatory values,
they have been derived to aid in the interpretation of 28-d sediment toxicity test data, particularly
to evaluate whether a measured PAH exposure in sediment or water is of an intensity that would
be expected to cause effects to H. azteca. These are values expected to cause 50% effect in
exposed organisms, they are not threshold effect values nor are then intended to assess effects on
any organism other than H. azteca.
2.0 Technical Basis
The conceptual framework used to derive these values is that described by Di Toro et al. (Di
Toro and McGrath 2000; Di Toro et al. 2000) and later used by U.S. EPA in deriving
Equilibrium-partitioning Sediment Benchmarks (ESBs) for PAHs (US-EPA 2003c). This
approach assumes that PAHs affect organisms like H. azteca via a narcosis mechanism of
toxicity, which in turn asserts that 1) the toxicity of individual chemicals (including PAHs)
increases with increasing octanol-water partition coefficient (Kow); and 2) that the toxicity of a
mixture of these chemicals can be predicted by summing their relative potencies (exposure
concentration/effect concentration) across all the similar chemicals in the mixture. The Di Toro
et al. (Di Toro and McGrath 2000; Di Toro et al. 2000) approach asserts that the slope of the
relationship between effect concentration and Kow has a common slope across all organisms,
such that this "universal narcosis slope" can be used to normalize the toxicity of narcotic
chemicals with different Kow values to common units and therefore compare them, even though
their absolute toxicities vary. This approach is built on the theoretical construct of a critical body
burden, and correspondingly that the relationship to Kow is tied to the partitioning of chemical
between the water and the organism (higher Kow = higher partitioning to the organism = higher
potency).
Although the approach is built around a body burden concept, the calculations are based on water
column exposure concentrations, rather than measured body burdens. This is in part because of
61
-------
the much greater availability of data for water column exposures than for residue-response
studies; it also avoids directly addressing complications of metabolites and kinetics of chemical
uptake. Finally, basing calculations on water column exposures is particularly appropriate for
purposes of evaluating interstitial water-based evaluations of sediment exposure, since
concentrations in (interstitial/pore) water are the exposure measure for such evaluations.
3.0 Data collection and analysis
3.1 Data collection and aggregation
Data relevant to this analysis were collected by searching the ECOTOX database
(www.e-pa.gov/ecolox) using the species name "Hyalella azteca" and selecting the PAH
chemicals subgroup. From there, papers containing water-only exposure data for H. azteca were
identified and the data contained therein were collated. Some papers provided toxicity data for
many time points; from these, emphasis was placed on 10-d and 28-d endpoints, as these
correspond to the durations of common sediment toxicity test procedures.
The extracted endpoint was the reported LC50 or EC50 (though some reported EC50 values
were reported on the basis of mortality). While the 10- and 28-d H. azteca sediment tests
measure sublethal endpoints (growth) in addition to survival, for most H. azteca sediment
toxicity data from PAH-contaminated sites the authors are aware of, survival is the primary
responding endpoint. That is not to say that the values derived here implicitly account for
sublethal effects, only that the derived values are expected to be applicable to many site data sets.
This search yielded ten LC50/EC50 values, 9 for 10-d exposures and one for a 28-d exposure
(Table 1). Data coverage included 7 publications and 5 different PAHs. In some cases, values
had to be visually interpolated from graphic presentations. If necessary, reported water
concentrations were converted to jig/L units using the molecular weights listed in Table 1.
As explained above, to directly compare potencies across PAHs, data must be normalized to
account for differing lipophilicity. This was performed using the method described by Di Toro
et al. (Di Toro and McGrath 2000; Di Toro et al. 2000), which estimates a theoretical critical
lipid concentration (Ce*),in units of jimol/g octanol, and calculated from the equation:
Log(C£*) = ((log EC50 in |^mol/L)+0.945*log K0w (L/kg octanol)*0.0001 kg/g (1)
Resulting values are shown in Table 1.
3.2 Selection of a 10-d H. azteca Cr|: value
The overall range of Ct* values obtained for nine 10-d exposures is about a factor of 4 overall
(range 9.6 to 34.7), but the distribution is skewed, with six of the nine values clustered in the
range of 26.2 to 34.7, and three in the range of 9.6 to 17.1. Toxicity data are often aggregated
using the geometric mean, but the observed distribution is even more skewed in log space,
arguing against that approach. Instead, the median, rather than the mean, of 27.3 jimol/g octanol
was selected as the estimated 10-d H. azteca Ce*.
62
-------
3.3 Selection of a 28-d H. azteca Ct* value
Only one applicable 28-d water only study was identified, which yielded a C«* value of 18.6
jimol/g octanol. The robustness of this value is uncertain because of having only a single value.
An alternative means to estimate the 28-d C«* value is to apply an adjustment factor to the 10-d
Ce* value to reflect the greater toxicity expected with longer exposure duration. In the case of
the single measured 28-d Ct* value, the same study (Schuler et al. 2004) had a paired 10-d
observation, which gave a 10-d Ce* value of 34.7, for a 10-d/28-d ratio of 1.86. In addition, the
Lee et al. (Lee et al. 2001) study contained daily EC50 values for up to 16 days of exposure for
four different PAHs. Plotting the Lee et al. data (Lee et al. 2001) as the log of the daily EC50
values versus the log days yielded quite linear slopes. Linear regressions were fit to those slopes
using the available data from days 7 to 16, then the resulting regressions were used to estimate
10-d EC50s and extrapolated to estimate 28-d EC50s. The resulting ratios of 10-d EC50 to 28-d
EC50 were 1.69 for naphthalene, 1.41 for fluorene, 1.22 for phenanthrene, and 1.89 for
fluoranthene. These four values were combined with the value of 1.86 for fluoranthene from the
Schuler et al. (Schuler et al. 2004) study, yielding a geometric mean of 1.59, which when applied
to the 10-d Ce* value of 27.3 yields an estimated 28-d Ct* value of 17.2 umol/g octanol.
This 28-d Ct* value is not only generally consistent with the single reported 28-d value of 18.6
(Table 1), but it is also reasonable that it is slightly lower than the Schuler et al. (2004) value,
which seems appropriate given that the 10-d Ct* value from the Schuler et al. (2004) study was
the highest among those 10-d values.
4.0 Comparison to EPA PAH ESB
While the EPA ESB document for PAHs has the same general goal of estimating potency of
PAHs to aquatic organisms, including H. azteca, and it is based on the same narcosis theory that
the current derivation is, there are differences in how the PAH ESB was calculated that create
differences between the estimated sensitivity of H. azteca in the PAH ESB document and the
analysis above. In the PAH ESB document, the ESB had only a single H. azteca value in the
acute toxicity database, a 4-d LC50 value of 44 |ig/L for fluoranthene (Spehar et al. 1999). If
one computes a 4-d Ct* value from this, the resulting value of 13.9 jimol/g octanol is
counterintuitively low compared with the higher 10-d Ct* value of 27.3 computed here, as one
would expect that Ct* should decline with length of exposure. The PAH ESB derivation did not
include any chronic H. azteca data as no life-cycle toxicity data were available for the species,
but if one applied the generic acute-chronic ratio of 4.16 (US-EPA 2003c), the species specific
chronic Ct* estimate for H. azteca would be 3.34 jimol/g octanol, which is nearly as low as the
actual ESB of 2.24 jimol/g octanol (US-EPA 2003c). Some have extended this argument to
assert that a H. azteca toxicity test is essentially as sensitive as the PAH ESB and would
therefore provide a comparable level of protection.
The data provided in Table 1 argue this is not the case, and that actual sensitivity of H. azteca to
PAHs is higher than the ESB. We believe the underlying reason for the apparent discrepancy
lies in a problem with the Spehar et al. (Spehar et al. 1999) acute value of 44 jig/L for
fluoranthene. Recent research has shown that H. azteca is sensitive to the chloride content of the
63
-------
dilution water, and that at chloride concentrations below about 15 mg/L, sensitivity of the
common strain of H. azteca used in most laboratory studies (including the Spehar study)
increases in a way not seen in other strains of H. azteca (Soucek et al. 2015). As the water
source used for the Spehar study is known to have low chloride (about 1.5 mg Cl/L), it is very
likely that the Spehar LC50 is biased low, something that was confirmed in the compilation of
data for the current analysis. Among 4-d LC50 data obtained for our analysis, computed 4-d C«*
values from five other studies ranged from 30.5 to 94.4 jimol/g octanol, far above the value of
13.9 from the Spehar study, and much more consistent with data from the 10-d and 28-d studies.
The conclusion is that the sensitivity of H. azteca that might be inferred from the PAH ESB
document is over-estimated, and is better represented by the current analysis.
Despite this difference, the potency of PAHs in sediments as calculated by the PAH ESB can be
used to estimate effect concentrations specifically for H. azteca 10-d and 28-d toxicity tests. The
PAH ESB is based on a final chronic value (expressed as Ce*) of 2.24 jimol/g octanol (U.S. EPA
2003). Accordingly, one would expect the 10-d EC50 for H. azteca to occur at 27.3/2.24 = 12.2
times the ESB concentration (or 12.2 ESBTU) and the 28-d EC50 to be 17.2/2.24 = 7.68 times
the ESB concentration (7.68 ESBTU). It is important to note that these values are not threshold
values that would be protective of H. azteca; instead they are EC50 values that would be
expected to have substantial adverse effect on H. azteca. To estimate a threshold for the H.
azteca response in 10-d or 28-d toxicity tests, the values presented here would have to be
adjusted down to go from an EC50 to an acceptably low level of effect.
5.0 PAH-specific water concentrations for evaluating 10-d and 28-d response
To predict the toxicity of a mixture of PAHs measured in interstitial water, one must calculate
the ratios of the exposure concentration to the effect concentration for each individual PAH, then
sum those ratios across all PAHs. PAH-specific values for making these calculations, based on
the 10-d and 28-d Ct* values derived above, are presented in Table 2. Fifty percent effect to H.
azteca is predicted when the summed ratios equal 1. If the summed ratios are greater than 1,
effects on H. azteca should be greater than 50%; if the summed ratios are less than 1, less than
50% effect is expected. It is critically important to understand that it is the sum of the ratios that
is evaluated, not the ratios for individual chemicals.
64
-------
Table 1 - Compiled 10-d and 28-d water column toxicity data and calculated critical lipid
concentration (Cc*) (based on Equation 1) for Hyalella azteca.
Chemical
Exposure
Duration
(day)
End point
Value
(Mfl/L)
MW
(g/mol)
log
Kow
Ce*
(jmol/g-
octanol
Source
Fluoranthene
10
LC50
110
202
5.08
34.71
(Schuler et al. 2004)
Fluoranthene
10
LC50
105
202
5.08
33.21
(Driscoll and Landrum 1997)
Naphthalene
10
LC50
3032
143
3.36
31.46
(Lee et al. 2001)
Fluorene
10
LC50
486
166
4.21
27.76
(Lee et al. 2001)
Phenanthrene
10
LC50
233
178
4.57
27.34
(Lee et al. 2001)
Fluoranthene
10
LC50
83
202
5.08
26.20
(Wilcoxen et al. 2003)
Pyrene
10
LC50
77
202
4.92
17.07
(Lee et al. 2001)
Fluoranthene
10
EC50
45
202
5.08
14.17
(Suedel et al. 1993)
Fluoranthene
10
LC50
30
202
5.08
9.56
(Suedel and Rodgers Jr 1996)
Fluoranthene
28
LC50
59
202
5.08
18.62
(Schuler et al. 2004)
65
-------
Table 2 - Estimated single chemical 10-d and 28-d EC50 values for Hyalella azteca for use in
estimating responses to PAH mixtures.
10-d Hyalella
estimated EC50
28-d Hyalella
estimated EC50
Chemical
(nq/L)
(nq/L)
Naphthalene
2355
1,482
C1 -Naphthalenes
997
627
C2-Naphthalenes
368
232
C3-Naphthalenes
135
85.2
C4-Naphthalenes
49.4
31.1
Acenaphthylene
3745
2,358
Acenaphthene
681
429
Fluorene
479
302
C1-Fluorenes
170
107
C2-Fluorenes
64.6
40.7
C3-Fluorenes
23.4
14.7
Phenanthrene
234
147
Anthracene
253
159
C1-Phenanthrenes
90.8
57.1
C2-Phenanthrenes
39.0
24.5
C3-Phenanthrenes
15.4
9.7
C4-Phenanthrenes
6.82
4.29
Fluoranthene
86.7
54.6
Pyrene
123
77.6
C1 -pyrene/fluoranthenes
59.7
37.6
Benz(a)anthracene
27.2
17.1
Chrysene
24.9
15.7
C1-Chrysenes
10.4
6.57
C2-Chrysenes
5.89
3.71
C3-Chrysenes
2.05
1.29
C4-Chrysenes
0.86
0.54
Perylene
11.0
6.92
Benzo(b)fluoranthene
8.26
5.20
Benzo(k)fluoranthene
7.83
4.93
Benzo(e)pyrene
11.0
6.92
Benzo(a)pyrene
11.7
7.35
lndeno(1,2,3-cd)pyrene
3.36
2.11
Dibenz(a,h) anthracene
3.44
2.17
Benzo(q,h,i)perylene
5.36
3.37
66
-------
Section 7
References
Adams D. 1991. Sampling sediment pore water. In: Mudroch A, MacKnight S, editors. CRC Handbook of
Techniques for Aquatic Sediments Sampling: CRC Press, Boca Raton, FL, USA. p 117-202.
Adams W, Burgess R, Gold-Bouchet G, LeBlanc L, Liber K, Williamson B. 2003. Porewater chemistry:
effects of sampling, storage, handling, and toxicity testing. In: Carr RS, Nipper MA, editors.
Porewater toxicity testing: biological, chemical, and ecological considerations: SETAC Press,
Pennsacola, FL, USA. p 95-124.
Adams W, Kimerle R, Mosher R. Aquatic safety assessment of chemicals sorbed to sediments. In:
Cardwell R, Purdy R, Baher R, editors; 1985. American Society for Testing and Materials,
Philadelphia, PA, USA. p 429-453.
Ankley G, Schubauer-Berigan M (1994) Comparison of techniques for the isolation of sediment pore
water for toxicity testing. Archives of Environmental Contamination and Toxicology 27:507-512.
Arp HPH, Azzolina NA, Cornelissen G, Hawthorne SB (2011) Predicting pore water EPA-34 PAH
concentrations and toxicity in pyrogenic-impacted sediments using pyrene content.
Environmental Science & Technology 45:5139-5146.
ASTM. 1994. Standard guide for collection, storage, characterization, and manipulation of sediment for
toxicological testing. E 1391-94. ASTM 1994 Annual Book of Standards, Conshohocken, PA, USA.
ASTM. 2013. Standard Test Method for Determination of Parent Alkyl Polycyclic Aromatics in Sediment
Pore Water using Solid Phase Microextraction and Gas Chromatography/Mass Spectrometry in
Selected Ion Monitoring Mode. D7363-13. Volume 11.02. West Conshohochen, PA, USA.
Baker RA (1980) Contaminants and Sediments. Volume 1. Fate and Transport,. Case Studies, Modeling,
Toxicity Ann Arbor Science, Ann Arbor, Michigan.
Bay SM, Weisberg SB. 2008. A framework for interpreting sediment quality triad data. In: Weisberg S,
Miller K, editors. Southern California Coastal Water Research Project Annual Report: Southern
California Coastal Water Research Project, Costa Mesa, CA, USA. p 175.
Beckingham B, Ghosh U (2011) Field-scale reduction of PCB bioavailability with activated carbon
amendment to river sediments. Environmental Science & Technology 45:10567-10574.
Bufflap SE, Allen HE (1995) Sediment pore water collection methods for trace metal analysis: A review.
Water Research 29:165-177.
Burgess R, McKinney R (1997) Effects of sediment homogenization on interstitial water PCB
geochemistry. Archives of Environmental Contamination and Toxicology 33:125-129.
Burgess RM, Ahrens MJ, Hickey CW. 2003a. Aquatic Environments: Source, Persistence and Distribution.
In: Douben P, editor. PAHs: An ecotoxicological perspective: John Wiley & Sons, London, UK. p
35-45.
Burgess RM, Ahrens MJ, Hickey CW, den Besten PJ, ten Hulscher D, van Hattum B, Meador JP, Douben
PE. 2003b. An overview of the partitioning and bioavailability of PAHs in sediments and soils. In:
Douben PE, editor. PAHs: An ecotoxicological perspective. John Wiley & Sons, London, UK. p 99-
126.
Burgess RM, Berry WJ, Mount DR, Di Toro DM (2013) Mechanistic sediment quality guidelines based on
contaminant bioavailability: Equilibrium partitioning sediment benchmarks. Environmental
Toxicology and Chemistry 32:102-114.
Burgess RM, Lohmann R, Schubauer-Berigan JP, Reitsma P, Perron MM, Lefkovitz L, Cantwell MG (2015)
Application of passive sampling for measuring dissolved concentrations of organic contaminants
67
-------
in the water column at three marine superfund sites. Environmental Toxicology and Chemistry
34:1720-1733.
Burgess RM, Ryba SA, Cantwell MG (2000) Importance of organic carbon quantity on the variation of Koc
in marine sediments. Toxicological & Environmental Chemistry 77:9-29.
Burkhard LP, Mount DR, Highland TL, Hockett JR, Norberg-King T, Billa N, Hawthorne SB, Miller DJ,
Grabanski CB (2013) Evaluation of PCB bioaccumulation by Lumbriculus variegatus in field-
collected sediments. Environmental Toxicology and Chemistry 32:1495-1503.
Butler JD, Parkerton TF, Letinski DJ, Bragin GE, Lampi MA, Cooper KR (2013) A novel passive dosing
system for determining the toxicity of phenanthrene to early life stages of zebrafish. Science of
the Total Environment 463:952-958.
Carr RS, Nipper MA (2003) Porewater toxicity testing: Biological, Chemical, and Ecological
Considerations.: SETAC Press, Pennsacola, FL, USA.
Chapman GA. 1987. Establishing sediment criteria for chemicals—regulatory perspective. In: Dickson K,
Maki A, Brungs W, editors. Fate and effects of sediment-bound chemicals in aquatic systems:
Pergamon. p 355-377.
Chapman P, Dexter R, Long E (1987) Synoptic measures of sediment contamination, toxicity and infaunal
community composition (The sediment quality triad( in San-Francisco Bay. Marine Ecology
Progress Series 37:75-96.
Cho Y-M, Ghosh U, Kennedy AJ, Grossman A, Ray G, Tomaszewski JE, Smithenry DW, Bridges TS, Luthy
RG (2009) Field application of activated carbon amendment for in-situ stabilization of
polychlorinated biphenyls in marine sediment. Environmental Science & Technology 43:3815-
3823.
Cho Y-M, Werner D, Choi Y, Luthy RG (2012) Long-term monitoring and modeling of the mass transfer of
polychlorinated biphenyls in sediment following pilot-scale in-situ amendment with activated
carbon. Journal of contaminant hydrology 129:25-37.
Cornelissen G, Gustafsson O, Bucheli TD, Jonker MT, Koelmans AA, van Noort PC (2005) Extensive
sorption of organic compounds to black carbon, coal, and kerogen in sediments and soils:
mechanisms and consequences for distribution, bioaccumulation, and biodegradation.
Environmental Science & Technology 39:6881-6895.
Dewitt TH, Jones JK, Ozretich RJ, Swartz RC, Lamberson JO, Schults DW, Ditsworth GR, Smith LM,
Hoselton L (1992) The influence of organic matter quality on the toxicity and partitioning of
sediment-associated fluoranthene. Environmental Toxicology and Chemistry 11:197-208.
Di Toro DM, McGrath JA (2000) Technical basis for narcotic chemicals and polycyclic aromatic
hydrocarbon criteria. II. Mixtures and sediments. Environmental Toxicology and Chemistry
19:1971-1982.
Di Toro DM, McGrath JA, Hansen DJ (2000) Technical basis for narcotic chemicals and polycyclic
aromatic hydrocarbon criteria. I. Water and tissue. Environmental Toxicology and Chemistry
19:1951-1970.
Di Toro DM, Zarba CS, Hansen DJ, Berry WJ, Swartz RC, Cowan CE, Pavlou SP, Allen HE, Thomas NA,
Paquin PR (1991) Technical basis for establishing sediment quality criteria for nonionic organic
chemicals using equilibrium partitioning. Environmental Toxicology and Chemistry 10:1541-
1583.
Driscoll SK, Landrum PF (1997) A comparison of equilibrium partitioning and critical body residue
approaches for predicting toxicity of sediment-associated fluoranthene to freshwater
amphipods. Environmental Toxicology and Chemistry 16:2179-2186.
Fernandez LA, Harvey CF, Gschwend PM (2009) Using performance reference compounds in
polyethylene passive samplers to deduce sediment porewater concentrations for numerous
target chemicals. Environmental science & Technology 43:8888-8894.
68
-------
Fernandez LA, Lao W, Maruya KA, Burgess RM (2014) Calculating the diffusive flux of persistent organic
pollutants between sediments and the water column on the palos verdes shelf Superfund site
using polymeric passive samplers. Environmental Science & Technology 48:3925-3934.
Ghosh U, Kane Driscoll S, Burgess RM, Jonker MTO, Reible D, Gobas F, Choi Y, Apitz SE, Maruya KA, Gala
WR and others (2014) Passive sampling methods for contaminated sediments: Practical
guidance for selection, calibration, and implementation. Integrated Environmental Assessment
and Management 10:210-223.
Greenberg MS, Chapman PM, Allan IJ, Anderson KA, Apitz SE, Beegan C, Bridges TS, Brown SS, Cargill JG,
McCulloch MC and others (2014) Passive sampling methods for contaminated sediments: Risk
assessment and management. Integrated Environmental Assessment and Management 10:224-
236.
Gschwend PM, MacFarlane JK, Reible DD, Lu X, Hawthorne SB, Nakles DV, Thompson T (2011)
Comparison of polymeric samplers for accurately assessing PCBs in pore waters. Environmental
Toxicology and Chemistry 30:1288-1296.
Hawthorne SB, Azzolina NA, Neuhauser EF, Kreitinger JP (2007) Predicting bioavailability of sediment
polycyclic aromatic hydrocarbons to Hyalella azteca using equilibrium partitioning, supercritical
fluid extraction, and pore water concentrations. Environmental Science & Technology 41:6297-
6304.
Hawthorne SB, Grabanski CB, Miller DJ (2006) Measured partitioning coefficients for parent and alkyl
polycyclic aromatic hydrocarbons in 114 historically contaminated sediments: Part 1. KOC
values. Environmental Toxicology and Chemistry 25:2901-2911.
Hawthorne SB, Grabanski CB, Miller DJ (2009) Solid-phase-microextraction measurement of 62
polychlorinated biphenyl congeners in milliliter sediment pore water samples and determination
of K DOC values. Analytical Chemistry 81:6936-6943.
Hawthorne SB, Grabanski CB, Miller DJ, Arp HPH (2011) Improving predictability of sediment-porewater
partitioning models using trends observed with PCB-contaminated field sediments.
Environmental Science & Technology 45:7365-7371.
Hawthorne SB, Grabanski CB, Miller DJ, Kreitinger JP (2005) Solid-phase microextraction measurement
of parent and alkyl polycyclic aromatic hydrocarbons in milliliter sediment pore water samples
and determination of K DOC values. Environmental Science & Technology 39:2795-2803.
Hawthorne SB, St. Germain RW, Azzolina NA (2008) Laser-induced fluorescence coupled with solid-
phase microextraction for in situ determination of PAHs in sediment pore water. Environmental
Science & Technology 42:8021-8026.
Heijden SAvd, Jonker MTO (2009) PAH Bioavailability in Field Sediments: Comparing Different Methods
for Predicting in Situ Bioaccumulation. Environmental Science & Technology 43:3757-3763.
Huckins JN, Petty J D, Booij K (2006) Monitors of organic chemicals in the environment: semipermeable
membrane devices: Springer Science & Business Media.
Ingersoll CG, Steevens JA, MacDonald DD, Brumbaugh WG, Coady MR, Farrar JD, Lotufo GR, Kemble NE,
Kunz JL, Stanley J K. 2014. Evaluation of toxicity to the amphipod, Hyalella azteca, and to the
midge, Chironomus dilutus; and bioaccumulation by the oligochaete, Lumbriculus variegatus,
with exposure to PCB-contaminated sediments from Anniston, Alabama. US Geological Survey.
Report nr 2328-0328.
lntegral_Consulting_and_Windward_Environmental. 2006. Portland Harbor RI/FS: Round 2 benthic
tissue and sediment data report. Draft Report. Lower Willamette Group, Portland, OR, USA.
Jahnke A, Mayer P, McLachlan MS (2012) Sensitive equilibrium sampling to study polychlorinated
biphenyl disposition in Baltic Sea sediment. Environmental Science & Technology 46:10114-
10122.
69
-------
Jonker MTO, Sinke AJC, Brils J M, Koelmans AA (2003) Sorption of Polycyclic Aromatic Hydrocarbons to
Oil Contaminated Sediment: Unresolved Complex? Environmental Science & Technology
37:5197-5203.
Karickhoff SW, Brown DS, Scott TA (1979) Sorption of hydrophobic pollutants on natural sediments.
Water Research 13:241-248.
Kipka U, Di Toro DM (2009) Technical basis for polar and nonpolar narcotic chemicals and polycyclic
aromatic hydrocarbon criteria. III. A polyparameter model for target lipid partitioning.
Environmental Toxicology and Chemistry 28:1429-1438.
Kraaij R, Seinen W, Tolls J, Cornelissen G, Belfroid AC (2002) Direct evidence of sequestration in
sediments affecting the bioavailability of hydrophobic organic chemicals to benthic deposit-
feeders. Environmental Science & Technology 36:3525-3529.
Kreitinger JP, Neuhauser EF, Doherty FG, Hawthorne SB (2007) Greatly reduced bioavailability and
toxicity of polycyclic aromatic hydrocarbons to Hyalella azteca in sediments from manufactured-
gas plant sites. Environmental Toxicology and Chemistry 26:1146-1157.
Lee JH, Landrum PF, Field U, Koh CH (2001) Application of a Ipolycyclic aromatic hydrocarbon model
and a logistic regression model to sediment toxicity data based on a species-specific, water-only
LC50 toxic unit for Hyalella azteca. Environmental Toxicology and Chemistry 20:2102-2113.
Long E, Morgan L (1991) The potential for biological effects of sediment-sorbed contaminants tested in
the National Status and Trends Program. NOAA Technical Memorandum NOS OMA 52. NOAA
Office of Oceanography and Marine Assessment, Ocean Assessment Division, Seattle, WA173.
Long ER, Chapman PM (1985) A sediment quality triad: measures of sediment contamination, toxicity
and infaunal community composition in Puget Sound. Marine Pollution Bulletin 16:405-415.
Long ER, Robertson A, Wolfe DA, Hameedi J, Sloane GM (1996) Estimates of the spatial extent of
sediment toxicity in major US estuaries. Environmental Science & Technology 30:3585-3592.
Lydy MJ, Landrum PF, Oen AMP, Allinson M, Smedes F, Harwood AD, Li H, Maruya KA, Liu J (2014)
Passive sampling methods for contaminated sediments: State of the science for organic
contaminants. Integrated Environmental Assessment and Management 10:167-178.
MacDonald DD, Carr RS, Calder FD, Long ER, Ingersoll CG (1996) Development and evaluation of
sediment quality guidelines for Florida coastal waters. Ecotoxicology 5:253-278.
MacDonald DD, Ingersoll C, Berger T (2000) Development and evaluation of consensus-based sediment
quality guidelines for freshwater ecosystems. Archives of Environmental Contamination and
Toxicology 39:20-31.
Mackay D, Shiu WY, Ma KC (1992) Illustrated handbook of physical-chemical properties and
environmental fate for organic chemicals. Volume II: polynuclear aromatic hydrocarbons,
polychlorinated dioxins, and dibenzofurans. Lewis Publishers, Boca Raton, FL 1992 597.
Maruya KA, Zeng EY, Tsukada D, Bay SM (2009) A passive sampler based on solid-phase microextraction
for quantifying hydrophobic organic contaminants in sediment pore water. Environmental
Toxicology and Chemistry 28:733-740.
Mayer P, Parkerton TF, Adams RG, Cargill JG, Gan J, Gouin T, Gschwend PM, Hawthorne SB, Helm P, Witt
G and others (2014) Passive sampling methods for contaminated sediments: Scientific rationale
supporting use of freely dissolved concentrations. Integrated Environmental Assessment and
Management 10:197-209.
Mount DR, Ingersoll CG, McGrath JA. 2003. Sediment Quality Guidelines for PAHs. In: Douben P, editor.
PAHs: An Ecotoxicological Perspective, p 331-355.
Nelson WG, Bergen BJ (2012) The New Bedford Harbor Superfund site long-term monitoring program
(1993-2009). Environmental Monitoring and Assessment 184:7531-7550.
70
-------
NOAA. 1998. Sampling and analytical methods of the national status and trends program mussel watch
project: 1993-1996 update. In: Lauenstein G, Cantillo AY, editors. NOAA Technical Memorandum
NOS ORCA 130. Silver Spring, Maryland, p. 233.
Ozretich RJ, Schults DW (1998) A comparison of interstitial water isolation methods demonstrates
centrifugation with aspiration yields reduced losses of organic constituents. Chemosphere
36:603-615.
Parkerton TF, Maruya KA (2014) Passive sampling in contaminated sediment assessment: Building
consensus to improve decision making. Integrated Environmental Assessment and Management
10:163-166.
Peijnenburg WJ, Teasdale PR, Reible D, Mondon J, Bennett WW, Campbell PG (2014) Passive sampling
methods for contaminated sediments: State of the science for metals. Integrated Environmental
Assessment and Management 10:179-196.
Redman AD, Parkerton TF, Paumen ML, McGrath JA, den Haan K, Di Toro DM (2014) Extension and
validation of the target lipid model for deriving predicted no-effect concentrations for soils and
sediments. Environmental Toxicology and Chemistry 33:2679-2687.
Reible D, Lotufo G. 2012. Demonstration and evaluation of solid phase microextraction for the
assessment of bioavailability and contaminant mobility. No. ESTCP-ER-200624. Texas University
at Austin, Department of Civil Engineering.
Sarda N, Burton GA (1995) Ammonia variation in sediments: Spatial, temporal and method-related
effects. Environmental Toxicology and Chemistry 14:1499-1506.
Schuler U, Landrum PF, Lydy MJ (2004) Time-dependent toxicity of fluoranthene to freshwater
invertebrates and the role of biotransformation on lethal body residues. Environmental science
& technology 38:6247-6255.
Schults DW, Ferraro SP, Smith LM, Roberts FA, Poindexter CK (1992) A comparison of methods for
collecting interstitial water for trace organic compounds and metals analyses. Water Research
26:989-995.
Smith KE, Dom N, Blust R, Mayer P (2010) Controlling and maintaining exposure of hydrophobic organic
compounds in aquatic toxicity tests by passive dosing. Aquatic Toxicology 98:15-24.
Smith KE, Oostingh GJ, Mayer P (2009) Passive dosing for producing defined and constant exposure of
hydrophobic organic compounds during in vitro toxicity tests. Chemical Research in Toxicology
23:55-65.
Soucek DJ, Mount DR, Dickinson A, Hockett JR, McEwen AR (2015) Contrasting effects of chloride on
growth, reproduction, and toxicant sensitivity in two genetically distinct strains of Hyalella
azteca. Environmental Toxicology and Chemistry 34:2354-2362.
Spehar R, Poucher S, Brooke L, Hansen D, Champlin D, Cox D (1999) Comparative toxicity of
fluoranthene to freshwater and saltwater species under fluorescent and ultraviolet light.
Archives of environmental contamination and toxicology 37:496-502.
Stephan CE, Mount DI, Hansen DJ, Gentile JR, Chapman GA, Brungs WA (1985) Guidelines for deriving
numerical national water quality criteria for the protection of aquatic organisms and their uses:
United States Environmental Protection Agency, Office of Research and Development
Washington DC.
Suedel B, Rodgers Jr J (1996) Toxicity of fluoranthene to Daphnia magna, Hyalella azteca, Chironomus
tentans, and Stylaria lacustris in water-only and whole sediment exposures. Bulletin of
environmental contamination and toxicology 57:132-138.
Suedel BC, Rodgers J H, Clifford PA (1993) Bioavailability of fluoranthene in freshwater sediment toxicity
tests. Environmental Toxicology and Chemistry 12:155-165.
71
-------
Tomaszewski JE, Werner D, Luthy RG (2007) Activated carbon amendment as a treatment for residual
DDT in sediment from a superfund site in San Francisco Bay, Richmond, California, USA.
Environmental Toxicology and Chemistry 26:2143-2150.
US-EPA. 1988. Guidance for Conducting Remedial Investigations and Feasibility Studies Under CERCLA.
EPA/540/G-89/004. Washington, DC, US EPA Office of Emergency and Remedial Response.
OSWER Directive 9355.3-01.
US-EPA. 1989. Risk Assessment Guidance for Superfund. Volume I: Human Health Evaluation Manual
(Part A). EPA/540/1-89/002.
US-EPA. 1994. Methods for assessing the toxicity of sediment-associated contaminants with estuarine
and marine amphipods. EPA 600/R-94/025. Technical Report. Narragansett, Rl.
US-EPA. 1995. Great Lakes Water Quality Initiative. Final Rule. Final water quality guidance for the Great
Lakes system. Federal Register, March 23,1995,Vol. 60, pp. 15365-15425.
US-EPA. 1996a. Ecological Effects Test Guidelines OPPTS 850.1020 Gammarid Acute Toxicity Test EPA
712-C-96-130. Office of Prevention, Pesticides, and Toxic Substances, Washington, DC.
US-EPA. 1996b. Ecological Effects Test Guidelines OPPTS 850.1035 Mysid Acute Toxicity Test EPA 712-C-
96-136. Office of Prevention, Pesticides, and Toxic Substances, Washington, DC.
US-EPA. 1996c. Ecological Effects Test Guidelines OPPTS 850.1045 Penaeid Acute Toxicity Test EPA 712-
C-96-137. Office of Prevention, Pesticides, and Toxic Substances, Washington, DC.
US-EPA. 1996d. Ecological Effects Test Guidelines OPPTS 850.1350 Mysid ChronicToxicity Test EPA 712-
C-96-120. Office of Prevention, Pesticides, and Toxic Substances, Washington, DC.
US-EPA. 1997. Ecological risk assessment guidance for Superfund: Process for designing and conducting
ecological risk assessments. Interim Final. EPA 540-R-97-006.
US-EPA. 2000. Methodology for Deriving Ambient Water Quality Criteria for the Protection of Human
Health (2000). EPA-822-B-00-004. Office of Water, Office of Science and Technology,
Washington, DC,USA.
US-EPA. 2002a. Methods for measuring the acute toxicity of effluents and receiving waters to
freshwater and marine organisms. Fifth Edition. EPA-821-R-02-012. Office of Water,
Washington, DC USA.
US-EPA. 2002b. Short-term methods for estimating the chronic toxicity of effluents and receiving waters
to freshwater organisms Fourth Edition. EPA-821-R-02-013. Office of Water, Washington, DC
USA.
US-EPA. 2003a. Procedures for the Derivation of Equilibrium Partitioning Sediment Benchmarks (ESBs)
for the Protection of Benthic Organisms: Dieldrin. EPA-600-R-02-010.: Office of Research and
Development. Washington, DC, USA.
US-EPA. 2003b. Procedures for the Derivation of Equilibrium Partitioning Sediment Benchmarks (ESBs)
for the Protection of Benthic Organisms: Endrin. EPA-600-R-02-009.. Office of Research and
Development. Washington, DC, USA.
US-EPA. 2003c. Procedures for the Derivation of Equilibrium Partitioning Sediment Benchmarks (ESBs)
for the Protection of Benthic Organisms: PAH Mixtures. EPA-600-R-02-013.: Office of Research
and Development. Washington, DC, USA.
US-EPA. 2005a. Contaminated sediment remediation guidance for hazardous waste sites. EPA 540/R-
05/012. Technical Report. In: Office of Solid Waste and Emergency Response W, DC, editor.
US-EPA. 2005b. Procedures for the Derivation of Equilibrium Partitioning Sediment Benchmarks (ESBs)
for the Protection of Benthic Organisms: Metal Mixtures. EPA-600-R-02-011.: Office of Research
and Development. Washington, DC, USA.
US-EPA. 2007a. Parent And Alkyl Polycyclic Aromatics In Sediment Pore Water By Solid-Phase
Microextraction and Gas Chromatography/Mass Spectrometry in Selected Ion Monitoring Mode.
72
-------
Method 8272. In Test Methods for Evaluating Solid Waste, Physical/Chemical Methods (SW-
846). Washington, DC.
US-EPA. 2007b. Sediment toxicity identification evaluation (TIE): Phases I, II, and III guidance document.
EPA/600/R-07/080. Office of Research and Development Washington, DC.
US-EPA. 2008. Procedures for the Derivation of Equilibrium Partitioning Sediment Benchmarks (ESBs) for
the Protection of Benthic Organisms: Compendium of Tier 2 Values for Nonionic Organics. EPA-
600-R-02-016.: Office of Research and Development. Washington, DC, USA.
US-EPA. 2012a. Equilibrium partitioning sediment benchmarks (ESBs) for the protection of benthic
organisms: Procedures for the determination of the freely dissolved interstitial water
concentrations of nonionic organics. EPA-600/R-02/012. Office of Research and Development,
Washington, DC, USA.
US-EPA. 2012b. Sediment assessment and monitoring sheet (SAMS) #3: Guidelines for using passive
samplers to monitor organic contaminants at Superfund sediment sites. EPA OSWER Directive
9200.1-110 FS. Washington, DC.
US-EPA. 2014. 8270. Semivolatile Organic Compounds by Gas Chromatography/Mass Spectrometry. SW-
846 Test Methods for Evaluating Solid Waste, Physical/Chemical Methods.
http://www.epa.gov/wastes/hazard/testmethods/sw846/online/index.htm: US-Environmental
Protection Agency, Washington, DC.
US-EPA, US-ACE. 2001. Method for Assessing the Chronic Toxicity of Marine and Estuarine Sediment-
associated Contaminants with the Amphipod Leptocheirus plumulosus. First Edition. EPA
600/R-01/020. Office of Research and Development, Newport, OR; Office of Water, Washington,
DC; and Engineer Research and Development Center, Vicksburg, MS.
US-EPA/SERDP/ESTCP. 2017. Laboratory, Field, and Analytical Procedures for Using Passive Sampling in
the Evaluation of Contaminated Sediments: User's Manual. EPA/600/R-16/357.. Office of
Research and Development, Washington, DC. https://www.epa.gov/superfund/superfund-
contaminated-sediments-guidance-documents-fact-sheets-and-policies.
Verbruggen EMJ. 2004. Environmental Risk Limits for mineral oil (Total Petroleum Hydrocarbons), RIVM
report 601501021. Bilthoven, The Netherlands: National Institute for Public Health and the
Environment,, p 79.
Wenning RJ, Batley G, Ingersoll C, Moore D (2005) Use of sediment quality guidelines and related tools
for the assessment of contaminated sediments: SETAC, Pensacola, FL, USA.
Wilcoxen SE, Meier PG, Landrum PF (2003) The toxicity of fluoranthene to Hyalella azteca in sediment
and water-only exposures under varying light spectra. Ecotoxicology and environmental safety
54:105-117.
Witt G, Lang S-C, Ullmann D, Schaffrath G, Schulz-Bull D, Mayer P (2013) Passive equilibrium sampler for
in situ measurements of freely dissolved concentrations of hydrophobic organic chemicals in
sediments. Environmental Science & Technology 47:7830-7839.
73
-------
oEPA
United States
Environmental Protection
Agency
PRESORTED
STANDARD POSTAGE
& FEES PAID EPA
PERMIT NO. G-35
Office of Research and
Development (8101R)
Washington, DC 20460
Offal Business
Penalty for Private Use
$300
------- |