COM
               BIOACCUMULATION:



             ASSESSMENT AND REGULATION
         Camp Dresser & McKee

-------
 U.S. Environmental Protection Agency
       Contract No. 68-01-6403
           BIOACCUMULATION:

      ASSESSMENT AND REGULATION
           Gary T. Hickman
             Ming P. Wang
           Laura L. Bowers
         Camp Dresser & McKee
7630 Little River Turnpike,  Suite 500
      Annandale, Virginia 22003
            September 1984

-------
                                  CONTENTS

Section                                                              Page
   1     INTRODUCTION                                                1-1
   2     BIOACCUMULATION THEORY AND MODELING                         2-1
             Background                                              2-1
             Empirical Steady-State Method                           2-5
             Kinetic Methods                                         2-7
                 Uptake from Water                                   2-7
                 Uptake from Sediment                                2-15
             Thermodynamic Equilibrium Methods                       2-17
                 Uptake from Water                                   2-17
                 Uptake from Sediment                                2-25
             Other Bioaccumulation Methods                           2-32
             Discussion                                              2-34
   3     REGULATORY APPLICATION OF BIOACCUMULATION DATA              3-1
             Implementation Manual                                   3-1
             Decision Guidelines                                     3-4
                 Background                                          3-4
                 Application                                         3-5
                 Discussion                                          3-5
                 Review of Decision Guideline Levels                 3-10
             Sediment Quality Criteria                               3-12
                 Background Levels                                   3-13
                 Water Quality Criteria                              3-13
                 Biological Response                                 3-14
                 Equilibrium Partitioning                            3-15
REFERENCES
APPENDIX A.  ANALYSIS OF MEAN CONTAMINANT RESIDUES IN AQUATIC
             INVERTEBRATES OF THE NEW YORK BIGHT

-------
                              ACKNOWLEDGEMENT
The authors wish to thank Richard Petti cord and Victor McFarland of the
U.S. Army Waterways Experiment Station and Norm Rubinstein of the U.S.
Environmental Protection Agency for their valuable input.

-------
                                 SECTION 1

                                INTRODUCTION
Aquatic organisms are capable of accumulating concentrations  of toxic
chemicals in their tissues that may be detrimental  to themselves or to
animals which feed upon them, including man.   Sources of contaminants for
biological uptake include water, sediment,  and biota.  A variety of methods
have been developed for estimating the extent of biological uptake. These
include empirical steady-state exposure, kinetic modeling,  and methods
based on thermodynamic equilibrium partitioning.

Predicting biological uptake from contaminated sediment is  complicated
because only a fraction of the total (bulk) concentration of  contaminants
in sediment may be bioavailable.  The bioavailability of a  sediment-
associated contaminant can be defined as the degree to which  the
contaminant is biologically active or accumulated by aquatic  biota (Adams,
1984).  Sediment contamination has become a matter of concern in part
because negative biological effects have been observed in aquatic
ecosystems where water column contaminant concentrations do not exceed  EPA
water quality criteria.  It has been suggested that the occurrence of
adverse effects in the absence of water quality criteria violations may be
due to sediment contamination, but that is difficult to prove.

An example of such a situation is the New York Bight Apex.  Average water
column concentrations of toxic chemicals in Apex waters are substantially
lower than the average value saltwater quality criteria, while the benthic
community is severely degraded in some areas of the Bight (Wang et al.,
1984).  Pavlou and Weston (1983) have suggested a similar scenario for
Puget Sound.  Dredged material and sewage sludge are discharged into the
New York Bight Apex, representing two sources of contaminated sediment
whose biological consequences need to be evaluated.  The regulatory process
requires that bioavailability be evaluated for dredged material  proposed
for discharge into ocean waters.
                                    1-1

-------
The objectives of this report are:
    (1)  to review the currently available methods for assessing biological
         uptake, with specific emphasis on uptake from sediment, and

    (2)  to review the regulatory application of bioaccumulation data,  with
         specific emphasis on dredged material  disposal  in the marine
         environment.
                                    1-2

-------
                                 SECTION  2


                    BIOACCUMULATION THEORY  AND MODELING


BACKGROUND


Aquatic organisms can accumulate chemicals  from their environment via  at

least three different pathways (Swartz and  Lee, 1980):


    o    direct adsorption to the body wall  or exoskeleton;

    o    direct absorption through the Integument,  gills,  or other
         respiratory surfaces; or

    o    ingestion of contaminated sediment, food,  or water  followed by
         absorption through the gut.


Three terms are frequently used in describing biological  uptake of
chemicals:  bioconcentration, bioaccumulation and biomagnification.  These

terms are used inconsistently in the literature.   Their meaning in this
paper is consistent with the terminology  of Brungs and Mount (1978):


    o  Bioconcentration is the process by which toxic substances present in
       solution enter aquatic organisms directly through  the gills or
       epithelial tissue.

    o  Bioaccumulation includes bioconcentration and the  uptake of toxic
       substances from dietary sources such as contaminated  particles  or
       prey organisms.

    o  Biomagnification is the process by which the tissue concentration of
       a bioaccumulated toxic substance Increases as the  material passes up
       through two or more trophic levels.


In general, aqueous-phase contaminants are  bioavailable and  are readily

bioconcentrated.  For most chemicals, direct uptake from  water is the

primary route of exposure, while dietary  uptake is relatively insignificant

(Macek et al., 1979; Bruggeman et al., 1981).  However, uptake from dietary

sources may be an important route of exposure for very lipophilic chem-

icals.  Macek et al. (1979) reported that dietary uptake  by  fish was
                                    2-1

-------
 insignificant compared to bioconcentration for seven of eight chemicals
 studied of  varying lipophilicity, but that dietary uptake of DDT con-
 tributed significantly to the total body burden (Jarvinen et al., 1977).
 DDT was the most lipophilic chemical in the data reviewed by Macek et al.
 and, consequently, exhibited the slowest depuration rates of residues from
 fish.  Likewise, the models of Weininger (1978) and Thomann and Connolly
 (1984) have predicted that food chain transfer of PCBs in Lake Michigan
 accounts for nearly 100% of the total PCB residues in adult lake trout.
 Bruggeman et al. (1981) stated that blomagnification factors (ratio between
 concentration in fish and in food) greater than one can be expected only
 for very lipophilic chemicals having octanol/water partition coefficients
 greater than 5, and that under natural conditions a biomagnification factor
 of one or greater may indicate a significant contribution of dietary uptake
 to the total body burden.

 Chapman (1984) presented a simple model  of the relative importance of food
 and water as the route of contaminant uptake by fish.  Respiratory exposure
 was computed as the product of ventilation rate and water concentration,
 and dietary exposure was computed as the product of feeding rate and
 contaminant concentration in food.  Assuming that the relative efficiency
 of uptake via the two routes of exposure is equivalent, the relative
 Importance of each route was estimated as a function of relative con-
 taminant concentrations in food and water as shown in Table 1.   Chapman's
 (1984) calculations (Table 1) indicate that dietary uptake is insignificant
 until  contaminant concentrations in food exceed concentrations  in water by
 more than 1,000-fold, and that dietary uptake is the dominant route of
 exposure when food concentrations are greater than 10,000 times aqueous
 concentrations.

 It is clear that contaminated sediments  can be an important source of
contaminants for biological  uptake.   Uptake from sediment can occur via two
 potential  avenues:   desorption to interstitial and interfacial  water
followed by bioconcentration, or bioaccumulatlon through ingestion of
contaminated sediment particles.   The relative contribution of  those two
mechanisms has not been elucidated,  but  it will  depend upon the feeding and
 respiratory strategies of the organism.   Deposit-feeding Infauna and
epibenthos are of primary interest with  respect to uptake from  sediment.
                                    2-2

-------
                                  TABLE 1
             RELATIVE IMPORTANCE OF FOOD AND WATER AS THE  ROUTE
              OF UPTAKE FOR FISH AS A FUNCTION OF THE RATIO  OF
                    CONTAMINATION BETWEEN FOOD AND WATER
                 Contaminant in                 Percent Uptake3
            Food (mg/kg):Water (mg/1)              from Food
                       1:1                           < 1

                      10:1                           < 1

                     100:1                             1

                   1,000:1                             9

                  10,000:1                            50

                 100,000:1                            91

               1,000,000:1                            99
 Calculations based on respiratory volume of 200 1/kg/day,  food consumption
 of 0.02 kg/kg/day, and uptake/depuration rates independent of route of
 exposure.

Source:  Chapman, 1984.
                                    2-3

-------
Hydrophobia organic compounds and some toxic metals have a strong affinity
for participate matter; consequently, those contaminants are commonly
associated with the bottom sediments in aquatic systems.  Toxic metals
associated with sediments generally are not readily bioavaiTable.  Thus,
the water column appears to be the predominant source for biological  uptake
of many toxic metals (Fowler et al., 1978).  In contrast, contaminated
sediments can be an important source of hydrophobic organic compounds for
biological uptake (Halter and Johnson, 1977; Fowler et al., 1978; Courtney
and Langston, 1978; Rubinstein et al., 1983; Rubinstein et al., 1984),
Although hydrophobic organic substances are readily taken up from water,
they have low solubilities and are highly associated with particulate
material. Consequently, relatively little of those substances is available
for aqueous uptake compared to the amount associated with sediments.   For
example, Fowler et al. (1978) measured concentration factors for PCBs in  a
polychaete worm (Nereis diversicolor) of 800-fold from water and 3.5-fold
from sediment, but because sediment concentrations of PCBs are typically  so
much greater than aqueous concentrations, the authors concluded that  85-99
percent of the worm's body burden could be attributed to uptake from
sediment.

A number of recent studies have shown dietary uptake to be a major source
of biological residues of hydrophobic organic compounds (Thomann, 1981;
Jensen et al., 1982; Pizza and O'Connor,  1983;  Thomann and Connolly,  1984;
Rubinstein et al., 1984).  Those substances are predominantly associated
with particulate organic material in sediments, which can serve as a  food
source for infaunal  and epibenthic organisms.  Such species can accumulate
synthetic organic compounds from contaminated sediments (Roesijadi et al.,
1978;  Fowler et al., 1978; Courtney and Langston,  1978; McLeese et al.,
1980;  Wyman and O'Connors, 1980; Lynch and Johnson, 1982; Rubinstein  et
al., 1983; Adams et al., 1983; Rubinstein et al.,  1984), and many of  those
organisms are important prey species for  organisms at higher trophic
levels.  Rubinstein et al. (1984) demonstrated  that contaminated sediment
can serve as a source of PCBs for uptake  and trophic transfer in marine
systems.  Fish (Leiostomus xanthurus) exposed to PCB-contaminated sediments
and fed a daily diet of polychaetes (Nereis virens) from the same sediment
accumulated more than twice the PCB whole-body  residues than fish exposed

                                    2-4

-------
to PCB-contaminated sediment but fed uncontaminated polychaetes  (Rubinstein
et al., 1984).

EMPIRICAL STEADY-STATE EXPOSURE METHOD

Chemicals can be taken up from the environment by aquatic  organisms.   In
turn, animals may be capable of eliminating contaminants from  their bodies.
When the environmental concentration of a contaminant is constant, a
theoretical equilibrium level  exists where the amount of chemical entering
an animal is equal  to the amount leaving the animal.   The  condition under
which a contaminant concentration is an animal is constant is  defined  as
steady state, and is often recognized as a plateau in the  observed tissue
concentration (Figure 1).  In laboratory tests aquatic organisms commonly
require days to months of exposure to acquire steady  state tissue con-
centrations of toxic organic compounds or toxic metals.

Equilibrium residue concentrations can be determined  directly  by exposing a
large number of animals to a constant concentration of a chemical,
periodically sampling the chemical  concentration in the animals, until
steady-state conditions are reached.  This method can be used  to measure
uptake from water,  sediment, food,  or combinations of these sources.

The aqueous exposure system permits the most accurate measurement of
biconcentration factor (BCF), which is defined as the ratio of the
equilibrium concentration of a chemical in aquatic organisms to the
concentration in water:
                     C ss
               BCF = Lb                                                 (1)
                      w
    where
              Cbss = steady-state concentration  of  the chemical  in biota
                Cw = concentration of the  chemical  in water  (which is kept
                     constant during the test)
                                    2-5

-------
p
c
p
n
y
T
U
E
e
H
   e.64-1
6.63-
   6.6i-
a  6.
    66-
              ae
                  n	>	[—
                  46      60
                    DAYS
                                    se
    e.s-
    e.4-
    e.a-
    ....
                   i    •   i
                  46      66
                    DAVS
                                     86
                                  MS-*
    §
    i »*>
                                            166
                                                       6.01
                                                          10  10   90  40  SO  AO  70  10

                                                                 DAVI  EXPOSURE
                                                                         35   42 |.j .1 .10 -14  -Jl  -28
                        I*
                       O*r
        Figure 1.  Typical  uptake curves.  Uptake of PCBs from contaminated sediments
                   by  (a) grass  shrimp, (b) polychaete worm, and  (c)  Asiatic clam.
                   Uptake from contaminated water of (d) mercury  by eastern oyster
                   and  (e)  isopropalin by bluegill.   Sources:  Rubinstein et al.,
                   1983b; McFarland et al., 1984; Kopfler, 1974;  Hamelink, 1977.
                                              2-6

-------
Uptake  from the water column has been measured for many chemicals and many
species.  For example, in the Ambient Water Quality Criteria Documents
(EPA, 1980, 1983), BCFs are reported for a variety of freshwater and
saltwater organisms for cadmium, mercury, chlordane, DDT, dieldrin,
heptachlor, PCBs, toxaphene, and other toxic chemicals.

Since BCFs consider only uptake from water and ingnore ingestion of
contaminated food and/or sediment, the term bioaccumulation factor (BAF)
has been used to describe uptake from sediment or uptake from all sources.

The empirical steady-state exposure approach has the advantages of
conceptual simplicity and a high degree of biological certainty (Hamelink,
1977), but the disadvantages of long test durations and the associated
difficulty of maintaining constant test conditions.  For example,
Rubinstein et al. (1983) found that 30 to 40 days of exposure were required
for Nereis worms to accumulate steady-state PCB concentrations.  Another
disadvantage is the inherent variability of tissue concentrations about the
mean which makes it inappropriate to represent the equilibrium concentra-
tion as the observed body burden at the end of the test period.  To resolve
that problem, Rubinstein et al. (1983) used a nonlinear regression program
to fit their data to the 3-parameter uptake model of Banner and Oglesby
(1981) to produce smooth uptake curves and yield an estimate of the
steady-state concentration (Figure 2).

KINETIC METHODS

UPTAKE FROM WATER

The process of bioconcentration frequently has been described as the
balance between two competing kinetic processes:   uptake and depuration
(e.g., Branson et al.,  1975;  Hamelink, 1977;  Ernst, 1977; Veith et al.,
1979;  McLeese et al.,  1980;  Mackay, 1982).  The simpliest kinetic model of
biological  uptake consists of a medium (water) compartment (CJ and a
                                                             n
biotic compartment (Cb>:
                                    2-7

-------
6-
*~* 5"
•o
O
CO
CL
*-* 3-
c
z 2-
r (
z
UJ
CJ
S 6-
LU
oo 5-
*— i
l—
CQ
0


•
J* i • 1
f '
I , 5°~
1 y= 5 16/11+ 601"7)
i **
3 20 40 eb ab iob c
I • 60-
^ ' •
i 55-
1
I y=600/(l*52'"6) :
o 45-
20 40 60 80 100 C
DAYS
Bahner and Oglesby
Y = PI
(t-P3)
1 + P2
(£) • ! •
• V^fl 	 8
• CF
x i.i m
y* 6I6/(I» 91 l l0')
> 20 40 60 80 100
(p) s . •
"x^ ^*
• f
«y • o jO/ 1 1 ^ 92 )
* ^™
20 40 60 80 100
Model
Where Y = In Cg at time t
PI = In Cn» maximum
P2 = rising slope
P3 = t when Y = Pl/2
Figure 2.   Bioaccumulation of PCBs to steady-state by N.  Virens exposed to
           four different sediments.   The uptake curves were constructed
           using the non-linear regression model of Bahner and Oglesby
           (1981).   Source:  Rubinstein et al., 1983.
                                     2-8

-------
                        "l
where kj and k2 are the rate constants  for  movement of the chemical into
(uptake) and out of (depuration)  the animal,  respectively.  Assuming first
order kinetics for uptake and depuration, the following  relationship has
been derived:

            dCb/dt = kLCw - k2Cb                                        (3)

Integration of this differential  equation,  when  Cw is constant and for
initial conditions of t and C.  equal to zero, gives:

                Cb = (k^kg) Cw (l-e'V)                               (4)
As time approaches infinity, equilibrium Is reached,  i.e.,  dCu/dt a 0, and
the above equation becomes:

              CbSS = (kl/k2) Cwss                                       (5)

or

         CbSS/cw" = kl/k2 = BCF
Thus, BCF can be estimated as the ratio between the uptake and depuration
rate constants, which can be determined experimentally.   Test conditions
employed to determine the uptake rate constant are essentially the  same as
those used in the empirical steady-state exposure approach except that
animals are sampled more frequently during a brief exposure period.   After
the uptake period, animals are transferred to aquaria containing clean
water (and sediment) and periodically sampled until the  depuration  rate
     is established.  Then, k  can be calculated.  Branson et al. (1975)
                                    2-9

-------
used this kinetic model to calculate a BCF for 2,2',4,4'-tetrachlorobi-
phenyl in rainbow trout muscle, and found that a 5-day uptake period plus  a
clearance time of one half-life (i.e., the time required to eliminate 50
percent of the accumulated contaminant residue) was sufficient to
characterize uptake and depuration.  The validity of the method was
demonstrated when the BCF thus obtained compared favorably to that deter-
mined experimentally during 42 days of exposure.

McFarland et al. (1984) described an alternative method for estimating
steady-state body burdens of hydrophobic organic compounds from residues
measured after a short exposure period in which steady-state was not
reached.  Combining Equations 4 and 5 and rearranging the resulting
expression yields:
                ss
                   =
Equation 7 implies that under conditions of constant exposure a steady-
state tissue concentration (Cbss) can be estimated from a tissue concen-
tration (Cb) at any time (t), provided the elimination rate constant (k2)
is known or can be estimated for that chemical.   The actual  exposure
concentration need not be known, but must be assumed to be constant
(McFarland et al., 1984).

Spacie and Hamelink (1982) demonstrated that the kinetic elimination rate
constant,  k2» for a particular chemical can be predicted from its
octanol-water partition coefficient (KQW).  Using the data of Meely et al.
(1974) and Konemann and van Leeuwen (1980), Spacie and Hamelink derived the
regression:

            log k2 = 1.47 - 0.414 log KQW                              (8)

                 r = 0.95

McFarland et al .  (1984) used k? values estimated in this way along  with
residues of di- and tri-chlorob1phenyl in asiatic clams and fathead minnows

                                    2-10

-------
measured after seven days of exposure to estimate steady-state body  burdens
using Equation 7.  Steady-state residues predicted from single-time-point
observations did not differ significantly from tissue concentrations
measured at a time corresponding to 99 percent of steady-state (McFarland
et al., 1984).

A number of more complex kinetic models have been proposed for evaluating
residue accumulation.  Figure 3A schematically represents the simple two-
compartment model described above.  Figure 3B and 3C are three-compartment
models (one water compartment and two biotic compartments) presented by
Blau et al. (1975); the only difference being that Figure 3C includes a
rate constant for elimination from compartment 3 back to the water.   Blau
and co-workers suggested that a three-compartment model may be appropriate
in cases where the chemical is preferentially accumulated in a particular
tissue or where the metabolization rate constant (k*?) is substantially
different than the rate of uptake from the medium (k12K  Notice that in
the model shown as Figure 3B when k32 is much greater than k23 the model
reduces to the two-compartment model shown as Figure 3A.  Krzeminski et al.
(1977) presented the model shown diagramatically in Figure 3D and
mathematically below, which includes a rapid exchange compartment (viscera)
and a slow exchange compartment (tissue) for both the parent compound and
metabolites.
                   • kxW + k4T - (k2 + k3 + ks) V                       (9)

               dVM
               dT ' k5V + k7TM - (k6 + V VM

                £ = k3V - k4T

               dTM
               dF = k6VM " k7TM
where
                 W = the nominal exposure level during the exposure phase,
                     and zero during the withdrawal phase
                                    2-11

-------
     a
                   Water
                Compartment
                                                     Fish
                                                 Compartment
  Constant

Chemical Addition
                               Constant Rate
                                    Of
                              Chemical Removal
                                                   Fiih
                                                Compartments
water

water
k' .
KI

* k'

Viscera
(parent
compound)
k,
\
f
Viscera
(metabolites
+ natural
products)
* ^

•x
m

k.

" H,
Tissue
(parent
compound)

Tissue
(metabolites
+ natural
products)
                                                                 Level of
                                                                 Complexity
Figure  3.   Kinetic compartment models of  biological  uptake.
            Blau  et al.,  1975; Krzeminski  et al.,  1977.
                                                     Sources:
                                         2-12

-------
                 V = the concentration of the parent compound in the
                     viscera at time t
                VM = the concentration of the metabolites in the viscera at
                 "   time t
                 T = the concentration of the parent compound in the tissue
                     at time t
                TM = the concentration of the metabolites in the tissue at
                     time t
                 k = the rate constants for the reactions shown in Figure
                     3D above
For refractory substances that resist metabollzation such as PCBs,  DDT,  or
toxic metals, this model would reduce to level one complexity,  that is,  the
three-compartment model shown as Figure 3B.

Although first order kinetics are usually assumed in kinetic studies of
bioaccumulation, they are rarely verified.  While simple first  order
kinetics are apparently appropriate for describing initial uptake rates
regardless of the mechanism or number of compartments used to model
subsequent distribution in the animal, Spacie and Hamelink (1982) suggest
that models with biphasic, second order, or Michaelis-Menten kinetics may
better describe depuration, depending on exposure level  and mode of
elimination.  For example, often an initial rapid loss phase is followed by
a slower elimination rate, forming a biphasic pattern that can  be resolved
into two linear components (Figure 4).  The initial  rapid phase represents
elimination from a fast or central compartment, while the slower phase
represents redistribution from a peripheral compartment within  the  animal
to the central one before elimination occurs.  Figure 3B represents this
type of model, and the rate constants can be determined using graphical
constants as shown in Figure 4.

The major advantage of using kinetic models for assessing bioaccumulation
is that shorter test durations are required for some chemicals  compared  to
the empirical steady-state exposure method.  Kinetic models can also
provide insight into the routes and rates functioning in the bioaccumu-
lation process.   Use of kinetic models has the disadvantage of  requiring
                                    2-13

-------
      A + B
                         Time
                   '32
= Ab + Ba
   A + B
                   '23 = a  + b -  k,,  -  k
                       = k
                         K32
                                  32
                 21
Figure 4.   Graphical  representation  of results  obtained  from
           a depuration experiment for which  the  first-order,
           three-compartment model (e.g.,  Figure  3b) would
           apply.   Source:   Spacie and Hamelink,  1982.

                            2-14

-------
nonlinear regression parameter estimation techniques  to estimate  rate
constants, and multi compartment models may require separate  analysis of
multiple anatomical  compartments and/or analysis of parent compounds and
metabolites.

UPTAKE FROM SEDIMENT

McLeese et al. (1980) used the first order,  two compartment  kinetic model
described by Equation 4 to estimate steady-state PCB  concentrations
accumulated by Neveri s virens exposed to contaminated sediment.   The
authors modified Equation 4 by substituting the ambient sediment
concentration, GS, for the aqueous concentration term, Cw:
C  = (kk) C  (l-e')                                        (4)
         b =      g   s
where:  Cs = concentration of PCB in sediment (ug/g dry weight)
The uptake rate constant (kj) was determined from whole-body residue data
measured over 32 days of exposure to contaminated sediment during which
steady-state was not approached, and the depuration rate constant was
determined from residue measurements after the worms were transferred to
aquaria containing clean sand.  Bioaccumulation factors, defined as the
ratio of wet weight tissue concentration to the dry weight sediment
concentration at steady state, were calculated mathematically as k1/k2.  A
schematic representation of the bioaccumulation model used by McLeese et
al. (1980) is shown as Figure 5a.

McFarland et al . (1984) also applied the first order kinetic model to
biological uptake of PCBs from contaminated sediment, but their exposure
system was designed to ensure that the aqueous phase was an intermediate
step between sediment and biota (Figure 5b).  Indeed, it has been suggested
that water is the probable medium of exchange for all pathways of bioaccum-
ulation (Rubinstein et al . 1984).  Fathead minnows (Pimephales promelas)
and Asiatic clams (Corbicula fluminea) were, except for one series of
exposures in which clams were allowed to burrow into sediment, restrained
from contacting the contaminated sediments in the exposure tanks.  Equation
                                    2-15

-------
          a
                SEDIMENT
                                 uptake
                                depuration
               BIOTA
  SEDIMENT
                 deaorptlon
                  sorptlon
WATER
               uptake
              depuration
BIOTA
Figure 5.  Schematic of  kinetic uptake models used by:  (a) McLeese et al
          (1980) and (b) McFarland et al. (1984).
                                  2-16

-------
4 (without modification) was used to predict steady-state body  burdens  and
BCFs.

THERHOPYMAMIC EQUILIBRIUM METHODS

The thermodynamic equilibrium approach allows the prediction  of equilibrium
tissue concentrations of contaminants in biota based on physico-chemical
equilibrium properties of the contaminant.  The method is invalid for
chemicals that are significantly transformed in the biotic phase {Mackay,
1982).  However, many toxic compounds which are of greater environmental
concern are fairly persistent, and this method shows great promise  in
predicting equilibrium contaminant concentrations in biota.

The theoretical basis behind the thermodynamic equilibrium approach was
discussed by Mackay (1979).  Mackay introduced the concept of fugacity  by
considering the environment as several compartments—including  atmosphere,
water, sediment, soil, and biota—some of which are contiguous  (e.g., water
and sediment) while others are not (e.g., atmosphere and aquatic sediment).
If it is assumed that each compartment is completely mixed and  at thermo-
dynamic equilibrium, then a contaminant in  the environment will  be
distributed such that the chemical potential (or fugacity) of the substance
in each phase is equal (Mackay, 1979).  The distribution of a substance
between source and sink at equilibrium is independent of pathways of
chemical exchange and can be described by a distribution or partition
coefficient (K):

          sink/'' source = ^source, sink

where Ce is the concentration of a substance in a particular  phase  at
equilibrium (equal chemical potential).

UPTAKE FROM WATER

Simple relationships have been established  between equilibrium  properties
of organic chemicals and their ability to accumulate In aquatic biota.
Several structure-activity correlations have been presented by  which

                                    2-17

-------
bioconcentration factors of hydrophobia  organic  chemicals can be predicted
from n-octanol/water partition  coefficient  (KQW)  or  aqueous solubility  (S).

BCF - K0n Relationship

Bioconcentration can be viewed  simply as the  partitioning of a chemical
between biota and water.  n-Octanol  is a good medium for simulating natural
fatty tissues of plants and animals;  therefore,  the  octanol/water partition
coefficient,
                     Ce
               K   =   n-octanol                                        (11)
                ow   ~e
                       water
is a good predictor of the partitioning of lipophilic  organic  chemicals
between biota and water (BCF).  Table 2 lists BCF-KQW  regression  equations
published in the literature.  Those regressions were developed from fish
{rainbow trout, fathead minnow, bluegill,  guppy, mosquitofish) biocon-
centration data.

Even though all researchers obtained a good linear relationship between log
BCF and log KQW, the values of the coefficients in the regression equations
vary a great deal.  Figure 6 is a comparison of the nine regressions listed
in Table 2.  The difference between the lowest and highest BCF predicted
from a given KQW by the lines in Figure 6  ranges between 30 to 103.  The
disparity among the regression equations may be due to at least three
sources of variation:  the chemical compounds used in  developing  the
regression, the test species (and/or their lipid content) used in deriving
BCFs, and the method used to measure or estimate K  '$.  The variation
                                                  ow
among the regression lines calls for a closer examination of the
biological- chemical meaning of the parameter values.

Applying the fugacity approach and assuming that the chemicals are not
appreciably transformed, Mackay (1982) proposed that BCF is proportional  to
                                    2-18

-------
                      TABLE 2
PUBLISHED REGRESSION EQUATIONS BETWEEN BCF AND K
                                                ow

1
2
3
4
5
6
7
8
9

log
log
log
log
log
log
log
log
log

BCF
BCF
BCF
BCF
BCF
BCF
BCF
BCF
BCF
Number
of
Equation Chemicals
(n)
= 0.124 + 0.542 log KQW 8
= -0.7504 + 1.1587 log KQW 9
= 0.7285 + 0.635 log KQW 11
= -1.495 + 0.935 log KQW 26
= -0.70 + 0.85 log KntaJ 55
uw
= -0.23 + 0.76 log KQW 84
= -1.32 + log KQW 44
= -0.063 + 0.980 log KQW 6
= -0.075 + 0.74 log KQW 36
Correlation
Coefficient Reference
(r)
0.97 Neely et
1974
al.,
0.98 Metcalf et al.,
1975
0.79 Kenaga &
1980
0.87 Kenaga &
1980
0.95 Veith et
1979
0.91 Veith et
1980
Goring,
Goring,
al.,
al.,
0.97 Mackay 1982
0.99 Konemann
Leeuwen,
0.88 Pavlou &
i noi
& van
1980
Vleston,
                       2-19

-------
       -n
       _j.

       c
       T


       en





       o
       «-«•

       o
       CO
       o
      n
ro
i
ro
o
       -s
       ro
       -s
       (D
       O
       3
       (D
o
ID
 O)
       in
       r+

       a.
       cu
       cr
       n>
                         2  -
                          1  -
                                                                                                                              8
                                                                                logK
                                                                      ow

-------
           BCF/KOW = A                                                 (12)

where A is a constant.  In other words,  in the general  linear  relationship
between log BCF and log KQW:
           log BCF = a + b log K
                                QW
the coefficient b should be 1 and the value of a  is  a  function of lipid
content (yL) and the relative activity. coefficient (TJ)  and molar volume
( "j) in lipfd (L) or octanol  (0):
                 a - log A = log  Lv ° "°                             (13)
Mackay (1982) tested his hypothesis using the data  of Veith et  al.  (1979).
Some of those data were eliminated because of suspected  error in measure-
ment or discrepancies between KQW values used by  Veith et  al. and those
reported elsewhere in the literature.   Data for chemicals  with  log  Kow
greater than 6 and for compounds which are surfactants or  may ionize were
also discarded.  The equation that resulted is:

           log BCF = - 1.32 + log KQW                                  (14)

                   r = 0.97

This approach makes fuller use of the  available theory and provides insight
Into the variables that affect the slope and Intercept of  BCF-KOW rela-
tionships.  For example, the Hp1d content of the test organisms is an
important variable which is often neglected by researchers.

Mackay (1982) suggested that other BCF-KQW correlations  have often
exhibited b values less than unity primarily due  to the  tendency to
overestimate KQW for high molecular weight compounds.  Mackay further
cautioned the use of the correlations  when membrane permeability limits
bioconcentration and for chemicals that bind to proteins causing an
underestimation of the biotic concentration.

                                    2-21

-------
Published BCF-KQW regressions have been developed from BCF data for a very
limited group of organisms, primarily nektonic fish.  For example,  90
percent of the data points included in Mackay's (1982) analysis (Equation
14) incorporated BCF values obtained using fathead minnows.   The ability of
such a regression to accurately predict bioconcentration in  other species
is uncertain.  Veith et al. (1979) investigated the effect of species
difference on bioconcentration and found that fathead minnow and green
sunfish accumulated three times more PCB residues than rainbow trout.

It is possible that variability in lipid content is the primary cause of
the different bioconcentration potential among species.  McFarland et al.
(1984) reported that bioconcentration is activity specific rather than
species specific.  McFarland and co-workers found that, when normalized by
the lipid content of the test species, BCFs of PCBs were similar for
Asiatic clams (2.03 percent lipid) and fathead minnnows (5.03 percent
lipid).  Based on their results, McFarland et al. (1984) suggested that
expression of residue data on a lipid basis may enable comparisons among
unrelated organisms.  If this is true, it could greatly increase the
utility of BCF-KQW regression equations, and, therefore, further
investigation is warranted to verify that suggestion.

BCF - Water Solubility Relationship

The aqueous solubility of organic chemicals can be used to predict
bioconcentration in a manner similar to KOW.  Several authors have shown
that solubility is related to KQW (Hansch et al., 1967; Chiou et al., 1977;
Tulp and Hutzinger, 1978; Kenaga and Goring, 1980; Mackay et al., 1980).
Consequently, solubility correlates satisfactorily with BCF, as shown by
the regression equations listed in Table 3 and plotted in Figure 7.  Figure
7 shows that, for a given solubility, a BCF predicted by the plotted
equations will be within a difference of 10 to 50.  This variation is
considerably less than that associated with BCF-KQW regressions.

The BCF-Kftb. and BCF-S correlations are useful because they allow prediction
         UVf
of contaminant residues from aqueous concentrations and physio-chemical
properties that are more readily available or more easily measureable than
                                    2-22

-------
                                       TABLE 3


          PUBLISHED REGRESSION EQUATIONS BETWEEN  BCF AND  WATER  SOLUBILITY(S)
                                               Number
                                     Units       of       Correlation
          Equation                   of S     Chemicals   Coefficient     Reference
                                                 (n)           (r)


1 log BCF = 3.9950 - 0.3891 log S     ppb         11          0.92    Lu  & Metcalf,
                                                                    1975

2 log BCF = 3.41 - 0.508 log S      umole/1         8          0.96    Chiou  et  al.,
                                                                    1977

3 log BCF = 2.791 - 0.564 log S       ppm         36          0.72    Kenaga &  Goring,
                                                                    1980

4 log BCF = 2.183 - 0.629 log S       ppm         50          0.66    Kenaga &  Goring,
                                                                    1980

5 log BCF = 3.71 - 0.316 log S        ppb         25          0.57    Davies &  Dobbs,
                                                                    1984
                                                                    (Recalculated
                                                                    from Veith et
                                                                    al., 1980)

6 log BCF = 5.09 - 0.85 log S         ppb         11          0.87    Metcalf et al.,
                                                                    1973

7 log BCF = 2.83 - 0.55 log S         ppm         42            -     Davies &  Dobbs,
                                                                    1984
                                                                    (Extrapolation
                                                                    from Kobayashi,
                                                                    1981)
                                         2-23

-------
ro
        n>
      -I -o
      01 — •
      o- o
      — • c+
      n>
         o
      ro -h
       it  i

      T3  Q>
      T3  r+
o

c
o-
ro


ro

CO

o
3

n>


o>
         o

         10
         rt-
         0)
         O.
                       6
                u_
                O
                m
                o>
                o     3
                        -5
                          -4
-3
-2        -1         0


               log S

-------
BCFs.  However,  they must be used discriminately.  They may not accurately
predict BCF for:  compounds with log KQW outside  the  range of 1 to 6;
compounds with short half-Hfes {high k2); compounds  that differ in
chemical nature from those used in the regression development; or compounds
with steric hinderances to accumulation (Tulp  and Hutzinger, 1978; Mackay,
1982).  Furthermore, these bioconcentration  models do not account for
changes in chemical  forms resulting from the presence of suspended solids
or sediments or for the effect of ingestion  of contaminated food organisms
or particles on total body burdens.

Bruggeman et al. (1981) developed an equilibrium  partitioning model for
directly estimating bioconcentration of lipophilic chemicals.  Since
lipophilic chemicals accumulate primarily in the  lipid tissue of biota, the
BCF can be estimated as:

                     kl     "Vw
               DfT —  •*—   " V  — V   V
                   " F"  ""  T  l "  wl  1
where
                 T = activity coefficient of the chemical  in water  (w) or
                     lipid (1)
                X] = lipid content of the organism (g lipid/g  organism
                     weight)
               Kwl = Hpid/water partition coefficient

Bruggeman et al.  (1981) reported that Kwl is relatively organism
independent.

UPTAKE FROM SEDIMENT
Recently, biological uptake from contaminated sediments has generated much
concern, especially since negative biological effects,  including bio-
accumulation, have been observed in areas where water column contaminant
concentrations do not exceed ambient water quality criteria.  Thermodynamic
                                    2-25

-------
methods have been proposed for predicting residue accumulation by  benthic
organisms from contaminated sediments.

Karlckhoff et al. (1979),  Means et al.  (1980),  and KaricJchoff (1981)
studied the equilibrium sorption behavior of hydrophobic  organic chemicals
and found that at low pollutant concentrations  (Cw <  0.5  S)  sorption
isotherms were linear, reversible, and  characterized  by a partition
coefficient,

               C% = Kws Cew                                           (16)

where Ces and Cew are the  equilibrium concentrations  of the  substance  in
the sediment and water, respectively, and KWS is  the  sediment/water
partition coefficient.  Kws is directly related to the organic carbon
content of sediments;  consequently,  KW$ fs  frequently normalfzed  for
organic carbon:
                     CX
            Kws/Xoc=
where XQC is the fractional  mass of organic  carbon  in  the  sediment  and KQC
is the sediment-organic-carbon/water partition coefficient,  commonly called
the soil sorption coefficient.   KOC'S are relatively consistent over a wide
variety of sediments (Means  et al., 1980; Karickhoff,  1981).  Regression
equations have been developed for estimating KQC from  octanol/water par-
titioning or aqueous solubility, and are presented  in  Tables 4 and  5 and
Figures 8 and 9.

Once the partition coefficients between sediment and water (Kws)  and
between biota and water (BCF) have been determined  or  estimated by  methods
described above, the contaminant residues that are  expected  to occur in
benthic biota exposed to interstitial or interfaclal (water  at the
sedient/water interface) waters can be approximated as follows:
                                    2-26

-------
                     TABLE 4
PUBLISHED REGRESSION EQUATIONS BETWEEN K._ AND Kft
                                        OC      On

'
2


3
4
5


6



7
8

9
Equation
log K._ = 0.885 + 0.524 log Kftu
\J\f \Jrl
log KQC = - 0.21 + 1.00 log KQW
K-_ = 0.63 K-., .
oc ow
logKoc = 1.377 + 0.544 log KQW
log Koc B - °'317 + 1Q9 Kow
log KA_ = - 0.346 + 0.989 log KniJ
oc ow
Kor = 0.411 K
oc ow
log KQC = -0.006 + 0.937 log KQW



log KQC = 0.02 + 0.94 log KQW
log K.. = - 0.18 + 1.029 1ogKou
\f\f wW

log K_ = 0.158 + 0.843 log KrtlJ
Number
of
Chemicals
(n)
30
10
10

45
22
5
5

19



9
13

19
Correlation
Coefficient Reference
(r)
0.96
1.00
0.98

0.86
0.99
1.00
1.00

0.97



-
0.95

0.96
Briggs, 1973
Karickhoff
et al., 1979

Kenaga &
Goring, 1980
Means et al . ,
1980
Karickhoff,
1981

Pavlou &
Weston, 1984
(From Brown
et al . , in
prep.)
Lyman et al.,
1982
Rao &
Davidson, 1980
Pavlou &
                                                    Weston,  1984
                       2-27

-------
                                       TABLE 5
          PUBLISHED REGRESSION EQUATIONS BETWEEN  K    AND  WATER  SOLUBILITY(S)
                                                 oc
          Equation
                             Units
                             of  S
 Number
   of
Chemicals
   (n)
Correlation
Coefficient
    (r)
Reference
1 log Kor = 0.44 - 0.54 log S       mole
                                  fraction
xoc
2 log Koc = 4.040 - 0.557 log S   umole/1
3 log Kor = 3.64 - 0.55 log S
       oc
5 log Kor = 4.070 - 0.82 log S
'oc
                             ppm
4 log K   • 4.273 - 0.686 log S     ppm
                             ppm
6 log Koc = 3.933 - 0.642 log S     ppm
    10


    15


   106


    22


     4


     5
   0.97     Karickhoff et
            al.,  1979

   0.99     Chiou et al.,
            1977

   0.84     Kenaga & Goring,
            1980

   0.97     Hassett et al.,
            1980

   1.00     Means et al.,
            1980

   0.97     Calculated from
            data  given in
            Karickhoff, 1981
                                         2-28

-------
                       -s
                       fD
                       00
                       O
                       c-t-
ro
i
ro
                     O
                     O
                     O




                       fD
                      Id

                       n>
O)
-Q
C
Q)
                       O
                       3
                       fD
                       O.
                       OJ
                       cr
                       fD

                       OJ
                                          6   -
                    5   -
                    4  -
                                   o
                                   O
O)
o
o      3  -
                                          2  -
                    1   r.
                                                                                                    log

-------
          n>
        -HT3
        O> — '
        o- o
        — ' r+
        n>
        I/O
         r>
         ii  i
ro

u>
O
o

£Z
CT
n>
          n>
           O
           3

           0>

           c
           Ol



           o
           O>
           0.
                       6
                  o
                  O
                 O)
                 O     3
                       0
                        -4
                         -3
-2
-1
0          1


   log  S

-------
                   = K  u                                               (18)
                „;   Nsb                                               l   '

and

           C% ' Ksb c%                                               (19)

where Ksb is the biota/sediment partition  coefficient.  Using this
relationship, Connor (1984)  derived an  equation  for estimating K$b by
combining two regression  equations of Kenaga  and Goring (1980) along with
the relationship between  KQC and Kws (Equation 17):

           log BCF = -  1.495 + 0.935 log Kow
           log K   =   1.377 + 0.544 log K"
               K   = K   /X
 	NQC   WS7 PC	
log Ksb = log (BCF/KWS) = -  2.872 + 0.391  log KQW - log XQC             (20)

Karickhoff (1984) has proposed a more direct  method of determining the
potential for bioaccumulation of hydrophibic  organics  from contaminated
sediments.  This method is based on the assumption that the predominant
pools for hydrophobic organic chemicals are lipid tissue  in biota and
organic carbon in sediments.  At a state of equl-chemical  potential, the
distribution of a substance should be equal to the ratio  of the chemical's
affinity for lipid and  organic carbon,  which  should be highly chemical
independent.  Thus, the thermodynamic potential  for biological uptake of
hydrophobic organics from a sediment source should depend only on the
organic carbon content  of the sediment, the lipid content of the organism,
and the lipld/organic carbon partition  coefficient (KQC ^):
The value of KQC ^ must be determined experimentally but should  be  on  the
order of unity, I.e.,
                                    2-31

-------
                     x,
                                                                       {22)
The pollutant concentration in an aquatic organism at a state  of equi-
chemlcal potential with the source defines the maximal chemical  uptake  from
purely thermodynanric considerations (Karlckhoff,  1984).  Unfortunately,  no
analogous method currently exists for determining the potential  for
bloaccumulation of metals or water soluble organics.

OTHER BIOACCUMULATION MODELS

Several other models capable of predicting bioaccumulation have  been
developed which do not fit precisely into the above sections.  These are
discussed briefly below.

Norstrom et al. (1976) developed a kinetic bioconcentration model based on
fish bioenergetics.  In general, their model can be described  by (Neely,
1979):
                     EC R
                          -                                             (23)
where
                 E = efficiency of chemical transfer across the gill
                     membrane
                 C = chemical concentration in fish (f) or in water (w)
                Rv = volume of water flowing past the gills per unit time
                     (ventilation rate)
                 F = weight of Hsh

Bruggeman et al. (1981) also used bioenergetics terms in a kinetic model of
bioaccumulation from contaminated food.  Their model of uptake due to
constant dietary exposure is given by:
                                    2-32

-------
             Cf(t) = f^ Cfd (Ue    ]                                  (24)
where
                 C = chemical concentration in fish (f)  and in food (fd)
                 e = adsorption efficiency for ingested  chemical
                 f = feeding rate (food weight per fish  weight per time)
                k2 = depuration rate
                 t = time

It follows from Equation 24 that the steady-state fish/food partition
coefficient (Kfd f), termed the biomagnification factor  by the authors, can
be determined by:
                     css
             Vf=7I? = £                                          (25)
                     cfd    z

The above equation was further modified by substituting an approximation
for the depuration rate constant

                     efFK ,X,
             K
              fd,f
where all variables are as previously defined.   The above expression  for
uptake from contaminated food incorporates a combination of kinetic-
bioenergetics terms and an equilibrium term, Kwl.   Total  bioaccumulation
(uptake from both food and water) can be determined by:

                Cf ' Kfd,f cfd + BCF cw

More complex food chain models have been developed (Weininger,  1978;
Thomann, 1981; Thomann and Connolly, 1984).   Although  these models are more
complex due to the number of auxiliary terms, Lake et  al.  (1984)  reported

                                    2-33

-------
that, In general, these models predict contaminant residues in  fish  by
applying the following differential  equation:


               -^- = klCw + kfdCfd efd " k2Cf                           (28)
               dt

where

               kfd = rate constant for uptake  from food
                 f = uptake efficiency term from food

Fugacity models have been developed to predict the distribution of
chemicals in idealized environments (Mackay, 1979; Mackay and Peterson,
1981, 1982).  At thermodynamic equilibrium, the fugacity (f), or escaping
tendency of a chemical substance from a phase, in all environmental
compartments is equal, e.g.:

              fair = fwater = fsediment = fbiota                       (29)

Model assumptions are that:  environmental  compartments are completely
mixed (i.e., homogeneous); there is no inflow  or outflow from the system;
and sufficient time has elapsed so that all compartments are in
equilibrium.

DISCUSSION

Two basic types of models have been proposed to describe the concentration
in the biota:  kinetic and equilibrium.  The first model considers the
biota concentration as a balance between the kinetic processes  of uptake
and depuration.  The second model considers the biotic phase as an in-
animate volume of material that is approaching thermodynamic equilibrium.
When first order kinetics are used to describe the uptake and depuration
rates of the first model and the rate to approach equilibrium of the second
model, the biota concentration (Cb) at any time (t) can be expressed
as (Mackay, 1982):
                                    2-34

-------
    Model I:    Cb = (kj/kg) Cw [l-exp(-k2t)]                          (30)
    Model II:   Cb = (BCF) Cw [1-exp [-   _ t)]                        (31)
                                        BCF
It Is apparent that Equations 30 and 31 are of the same form, only their
coefficients are different.  Thus, uptake measurements are inherently
Incapable of differentiating between the two models.  However, an Important
conceptual difference Is that Model I Is dependent on the kinetic rates and
therefore the steady state concentration may not be the thermodynamic
equilibrium concentration, while Model  II Is based on thermodynamic
equilibrium partitioning and does not consider kinetic processes (routes
and rates).

Thermodynamic equilibrium concentrations in the biotic phase will seldom be
reached by steady- state concentrations in the environment (Peddicord and
McFarland, personal coiranunciation)..  However, steady-state concentrations
estimated by kinetic modeling generally have not been compared in the
literature to thermodynamic equilibrium concentrations.  Comparisons
between steady state concentrations (estimated from kinetic rates) and
equilibrium concentrations (estimated from structure-activity relation-
ships) might be misleading due to the large variation in the published
thermodynamic equilibrium regression equations (e.g., see Figures 6 and 7).
BCFs predicted using those relationships may vary as much as 103.; conse-
quently, the outcome of such a comparison greatly depends on the equation
chosen to estimate the thermodynamic equilibrium concentration.

In the absence of biological transformation of bioaccumulated contaminants,
the steady state concentration in an organism should be equal to the
thermodynamic equilibrium concentration for the biotic phase.  If chemical
transformation occurs, however, the steady-state body burden will deviate
from the thermodynamic equilibrium concentration.  An important objective
of a kinetic study should be to investigate and quantify this rate of
biotransformation.  However, depuration rates are commonly estimated by
transferring test organisms to clean water.  This measured depuration rate
may be the cumulative result of exchange processes at the gill, excretion
                                    2-35

-------
via the kidney or bile, and biotransformation.   If exchange and  excretion
processes are solely responsible for movement of a chemical  from biological
tissue to water and the chemical is not metabolized,  the  kinetic steady-
state residue should be equivalent to the thermodynamic equilibrium
concentration.  Thus, the net depuration rate,  k2, provides little
information on the deviation from equilibrium concentrations.  In some
instances, it is suspected that the metabolism rate is much smaller than
the rate of elimination due to exchange or excretion,  or  there is a
significant lag period for the metabolism rate;  in such cases, the
cumulative depuration rate is truly representative of the rate of chemical
exchange between the biotic compartment and the environment.  For example,
the ko rates reported by Neely et al. (1974) have been found to  correlate
well with the thermodynamic equilibrium property KQW (Spade and Hamelink,
1982).
                                    2-36

-------
                                 SECTION 3


               REGULATORY APPLICATION OF BIOACCUMULATION  DATA
IMPLEMENTATION MANUAL


The "Implementation Manual"  (EPA and COE,  1977)  specifies  evaluations  to be

performed on dredged material  proposed for discharge into  marine waters.

The bioavailability of sediment-associated contaminants  is assessed  via a
10-day solid phase bioassay  followed by bioaccumulation  analyses  (whole-

body residues) on surviving  organisms.  The procedure for  assessing
bioaccumulation potential  described in the Implementation  Manual  is

summarized in Figure 10 and  the accompanying text below.


    1.  The Manual (EPA and  COE, 1977) expresses a preference for  assessing
        bioaccumulation potential in the field,  when possible,  rather  than
        using laboratory bioassays, because the  field method integrates
        influencing factors  such as mixing zone, sediment  transport, and
        long exposure times.  Caged animals may  also be  used, although for
        those animals with no history of exposure long periods  of  time jji
        situ will be necessary.

    2.  Conditions required  for field assessment.  The use of the  field
        approach for assessing bioaccumulation potential  is "technically
        valid only where there exists a true historical  precedent  for  the
        proposed operation being evaluated.  That is, it can be used only
        in the case of maintenance dredging where the quality of the sedi-
        ment to be dredged is considered not to  have deteriorated or become
        more contaminated since the last dredging and disposal  operation.
        In addition, the disposal must be proposed for the site at which
        the dredged material in question has been previously disposed  or
        for a site of similar sediment type supporting a similar biological
        community."

    3a. Species selection for field assessment.   Organisms must occur  in
        sufficient numbers at all stations for collection  of an adequate
        sample (several grams of tissue) to permit measurement of  chemical
        concentrations.  It is desirable to collect large  organisms  but
        they must be relatively immobile--bivalves, some gastropods, and
        large polychaetes are recommended.

    3b. Species selection for laboratory assessment.  Section 227.27(d) of
        the Federal Register (42 PR 2481) defines "appropriate sensitive
        benthic marine organisms" as at least three species consisting of a
        filter-feeding, a deposit-feeding, and a burrowing species.  It is


                                    3-1

-------
                       1.  Goal:  to evaluate the
                        bioaccumulation potential
                         of organisms exposed  to
                        dredged  material  compared
                        to organisms not exposed
                          to  dredged material.
     2a.   Conditions  necessary
       for field assessment of
      bioaccumulation  satisfied.
                 I
    J3a.   Select  species
                  I
     4a.   Collect  organisms
       exposed  and not  exposed
        to dredged material
                 I
                                   I
                                                     1
2b.  Conditions necessary
  for field assessment of
      bioaccumulation
       not satisfied.
3b.  Select Species
4b.  Conduct 10-day solid
      phase bioassay
                                   1
                       5.  Select chemical
                         constituents of concern.
            J
                       6.  Analyze surviving
                             organisms for
                         tissue concentrations
                       7.   Statistically analyze
                          data to determine if
                          organisms  exposed to
                            dredged  material
                       accumulated higher  tissue
                         levels than those not
                       exposed to dredged material
                       8.   Data  interpretation
Figure 10.   Flow chart of bioaccumulation potential  assessment prescribed
            in the Implementation Manual.
                                  3-2

-------
    recommended to include a crustacean,  an  infaunal  bivalve, and an
    infaunal polychaete.   The Manual  lists recommended  species  for use
    in solid phase bioassays in Table Fl.  The  guidance manual  for the
    New York District (COE and EPA,  1982)  specifies  the use of
    Palaemonetes sp., Mercenaria mercenaria,  and Nereis sp. or  Neanthes
    sp. In bioaccumulation tests conducted on dredged materials pro-
    posed for disposal  in the New York Bight.

4a. Field method design and conduct.   Organisms are  collected from at
    least three stations  within disposal  site boundaries  and from at
    least six stations outside the disposal  site with substrate that is
    sedimentologically similar to that within the disposal site.  Of
    the six outside-site  stations, three  should be "downstream" from
    the site (in the  direction of net bottom transport) and three or
    more must be located in an uncontamlnated sediment  in a direction
    opposite that of  the  net bottom transport (data  from  the latter
    will provide a reference level of tissue concentrations).   After
    collection, organisms are held in clean  water in the  lab to allow
    voiding of digestive tracts (2-3 days).   Also, shells are not
    included in the chemical analyses.

4b. Lab method design and conduct.  Organisms are collected  (preferably
    from the reference stations) and used in the 10-day solid phase
    bioassay described in the Manual  (Appendix  F).  Reference aquaria
    contain a 45 mm layer of reference sediment.  The New York  District
    COE has identified specific locations for the collection of
    reference and control sediments for use  in  laboratory solid phase
    bioaccumulation tests (COE and EPA, 1982).   Test organisms  are
    allowed to establish  themselves in a  30  mm  layer of reference
    sediment in exposure aquaria and then a  15  mm layer of dredged
    material is added.   Seawater may be added as a continuous flow-thru
    system or as a static system with periodic  replacement.  After ten
    days of exposure, the surviving organisms are removed and held in
    clean seawater to void digestive tracts  as  in the field method.

5.  Selection of chemical constituents for analysis. Chemical
    constituents to be assessed in the tissues  of test  organisms are
    those deemed critical by the District Engineer and  Regional
    Administrator after consideration of  known  inputs to  the dredged
    material.  Section 227.6 of the Federal  Register cites the
    following constituents of particular  concern:

    a.  Organohalogen compounds,
    b.  Mercury and its compounds,
    c.  Cadmium and its compounds,
    d.  Petroleum hydrocarbons,
    e.  Known or suspected carcinogens, mutagens, or teratogens.

    In the New York District, bioaccumulation analyses  are to be per-
    formed for cadmium, mercury, PCBs, and petroleum hydrocarbons, or
    other material as determined necessary under the above categories
    (COE and EPA, 1982).
                                3-3

-------
    6.  Analysis of tissue concentrations.   Tissue  concentrations of  the
        selected constituents are analyzed  according to  standard methods
        described in references cited in the Manual  (Table  Gl).

    7.  Statistical analysis of data.  Data is  tabulated for each con-
        stituent and species.Tissue concentrations in  exposed animals are
        compared to those in reference animals  to determine if the  former
        show statistically significant higher levels—indicating bio-
        accumulation potential  of the dredged material.   Statistical
        methods include:

        •   Cochrane test—to see if variances  of the data  sets are
            homogeneous and if data transformation  is necessary.

        t   Analysis of variance

        •   F-test—to see if there is any  statistical difference between
            exposed and reference tissue concentrations.

        •   Student-Newman-Keuls Multiple range test—to determine  which
            exposed tissue concentration means  are  significantly different
            (higher) than reference tissue  concentration means.

    8.  Data Interpretation.  The Implementation Manual  recommends  the
        environmentally protective approach of  assuming  that any statis-
        tically significant differences in  tissue concentrations between
        reference and exposed organisms are a potential  cause for concern.
        The "Decision Guidelines" are an attempt at interpreting laboratory
        bioaccumulation data for regulatory purposes.


DECISION GUIDELINES


BACKGROUND


The Decision Guideline documents (COE and EPA,  1980, undated a-c) provide a

method for interpreting contaminant residues acquired by organisms  exposed

to dredged material in accordance with the  Implementation Manual (EPA and

COE, 1977) laboratory procedure.  In developing the Decision Guidelines,

average tissue levels in biota of the New York  Bight were estimated from

the available data for PCBs, cadmium, mercury,  total naphthalenes,  and DDT

and its metabolites.  Those levels were adopted as  maximum  values for each

of three test species which, if not exceeded, should prevent significant

undesirable effects from occurring in areas contiguous with the dredged

material disposal site in the New York Bight Apex.   The  Decision Guidelines
                                    3-4

-------
were developed on an interim basis under the philosophy  of maintaining  the
status quo and thereby preventing further degradation of the  Bight.

The Decision Guidelines are regulatory tools, based on the available
bioaccumulation data and the short-term administrative goal of no-further-
degradation; they are not scientifically defensible because they  do not
consider the environmental significance or biological  effects of  con-
taminant residues.  The long-range goals of bioaccumulation interpretation
are effects-based regulation and mitigation of the environmental  stress on
the New York Bight due to dredged material  disposal.

APPLICATION

The Decision Guidelines for the New York Bight and their method of
derivation are summarized in Table 6.   Figure 11 illustrates  the  use of the
Decision Guidelines to interpret bioaccumulation data generated by the
10-day solid phase bioaccumulation test (EPA and COE,  1977).   Figure 11 is
actually a continuation of Figure 10 and together they represent  the
overall procedure for assessing the bioavailability of contaminants in
dredged material  proposed for disposal in the New York Bight  Apex.

DISCUSSION

A major criticism of the Decision Guidelines interpretive approach is that,
for some chemicals, non-steady-state residues acquired over ten days of
laboratory exposure are compared to steady-state residues obtained from
field data.  The House Committee on Merchant Marine and  Fisheries (1980)
voiced that criticism in a question posed to the Corps of Engineers at  a
Congressional hearing on dredge spoil  disposal and PCB contamination (p.
685):

    QUESTION 5;  The Corps stated at the May 21 hearings that "it is
    equally clear that different species have different  uptake rates
    for different contaminants and that within this ten-day exposure
    period, it is not possible to predict right now what the  ultimate
    level in a particular organism might be."  If this is the case,
    what validity do the matrix values have if the test  data  do not
                                    3-5

-------
                                  TABLE 6
                 DECISION GUIDELINES FOR THE NEW YORK BIGHT
Chemical
Polychlorinated
blphenyls
Cadmium
Mercury
Total naphthalenes
DDT + metabolites
Decision Guideline Level
Wet Weight Tissue Derivation
Test Organism Concentration (mg/kg) Method
Mercenaria
mercenaria
Palaemonetes sp.
Nereis sp.
all 3 species*
all 3 species
all 3 species
all 3 species
0.1
0.1
0.4
0.3
0.2
0.02
0.04
1
1
2
3,2
3,2
3
2
1.  Red Book (EPA, 1976;  AFS,  1979) criteria level  in fish  for  protection
    of freshwater and marine life and consumers thereof including  fish-
    eating birds and mammals.

2.  (NYB representative water column concentration)  x (selected BCF)  =
    (estimated average tissue concentration in NYB  biota)

3.  Grand mean of available tissue concentration data for invertebrates
    from the NYB.
aMercenaria, Palaemonetes,  and Nereis
                                    3-6

-------
                             9.  Goal:   Interpretation
                                of laboratory derived
                              levels  of bioaccumulation,
                                to  prevent  significant
                              undesirable  effects from
                                  occuring  in areas
                             contiguous with the Dredged
                              Material Disposal  Site in
                               the New York Bight Apex
                             10.  Applicant submits
                              bioaccumulation test data
                                  for  review  by  the
                                permitting authority.
                       1
                                         1
            lla.  Animals exposed to
              test sediment show no
           significant bioaccumulation
               compared to animals
                   exposed to
               reference  sediment.
                               1
                   lib.  Animals exposed to
                       test  sediment show
                   statistically significant
                   bioaccumulation compared
                      to reference animals.
           12a.  Material may be
                  ocean dumped
               without containment
            ±
                  12b.  Bioaccumulated levels
                        are compared to
                   Decision Guideline Limits.
13a.  Statistically
significant bioaccumulation
in 1 or 2 species; level(s)

-------
    represent the ultimate concentration level  to be reached in the
    organism?


The Corps' response, while answering the question, did not fully address

the issue of data incomparability.
    RESPONSE 5:  The matrix considers the variation in uptake by
    different species by considering the uptake of three species of
    very different ecological  and taxonomic groupings.  It also is
    based on actual uptake demonstrated upon exposure to the intact
    mixture of all contaminants in the dredged material.

    The matrix is clearly viewed by all involved parties as nothing
    more than an interpretive tool which will  be revised on a regular
    basis as Interpretive abilities improve.  The use of such Interim
    tools is necessary since Congress mandated that bioaccumulation be
    evaluated at a time when the state-of-the-art for the environ-
    mental interpretation of such data is far from fully developed.
    Without the use of some sort of interim tool, the only two
    approaches to interpretation are to consider any indication of
    potential bioaccumulation unacceptable or to consider all bio-
    accumulation acceptable, unless it is at a grossly high level
    where damage is obvious.  Neither of these extremes represents  a
    socially or environmentally acceptable view.  The Ocean Dumping
    Implementation Manual gives preference to bioaccumulation studies
    done in the field over laboratory studies.  In this manner, it  is
    posible to assess bioaccumulation by animals that have spent major
    portions of their lives in and on sediments very similar to the
    dredged material in question.  These organisms exist under the
    physical and chemical conditions actually occurring at the dis-
    posal site.  The tissue concentrations in such animals are in
    equilibrium with their environment and can be compared directly to
    tissue levels in organisms from areas unaffected by the dumping.
    This approach is the most desirable, but until appropriate field
    studies designed specifically for the New York Bight have devel-
    oped an adequate data base for intrepretation, laboratory studies
    must be used.  However, the interpretive difficulties posed by
    laboratory bioaccumulation investigations require some sort of
    interpretive tool and the matrix is the best currently available
    attempt in devising such a tool.  ...


The 10-day bioaccumulation test (EPA and COE,  1977) was designed to

indicate whether there is a potential for contaminant uptake from dredged
material (Peddicord, 1980, p.  587).  Rubinstein et al. (1983) found that
ten days of exposure is sufficient to demonstrate the potential for PCB
bioaccumulation, but that there is no clear relationship between 10-day

whole-body concentrations and steady-state concentrations.  Steady-state
                                    3-8

-------
concentrations are clearly more meaningful than levels reported without
reference to kinetic processes (McFarland et al., 1984).  From their
experimental data, Rubinstein et al. (1983) concluded that since 10-day
residues do not reflect steady-state concentrations they cannot be compared
directly to a value intended to represent concentrations in indigenous
biota at the disposal site (i.e., the Decision Guideline values).

If a 10-day exposure period is to have predictive value, its relationship
to long-term exposure must be determined.  McFarland et al. (1984) have
described a method by which a steady-state contaminant residue can be
estimated from a single time-point measurement of the contaminant  con-
centration in tissue and an elimination rate constant (k2)  estimated from
the octanol/water partition coefficient of the chemical.  This procedure,
which is outlined in the Section 2, allows estimation of steady-state body
burdens from 10-day bioaccumulation data.  Conceptually, the Decision
Guideline levels are more appropriately compared to steady-state residues
than to 10-day, non-steady-state concentrations.

The Decision Guidelines can also be criticized over the inconsistency of
methods used in their development.  In spite of that inconsistency,
however, the concentrations adopted to represent average contaminant
residues in New York Bight macroinvertebrates appear to be fairly  well
supported by the currently-available data (as illustrated in the next
section).  Some arguable points pertaining to the derivation of the
Guideline values are outlined below:
    •  The EPA Red Book value of 0.1 mg/kg adopted as the PCB  Decision
       Guideline limit in Mercenaria and Palaemonetes is not based on
       recent or site-specific data.
    •  The approach wherein a water column concentration is  multiplied  by  a
       BCF to estimate an average tissue concentration (Method 2  in Table
       6) assumes that bioaccumulation of contaminants is via  the water
       column.  That assumption may not be valid for infaunal  worms or
       deposit-feeding species which may accumulate contaminants  primarily
       from sediment.
    •  Using the method described above in the  derivation of the  PCB
       Decision level  for worms, the highest value of the range of reported
       water column concentrations was selected (44 ng/L), representing the
       least conservative choice.

                                    3-9

-------
    •  The Decision Guideline levels for cadmium and mercury  were
       established based on the grand mean of body burdens  measured  in
       invertebrates from the New York Bight.   The accuracy of those values
       was "tested" using Method 2 (Table 6).   However,  the range  of BCFs
       reported for cadmium and mercury are wide, and whether or not the
       grand mean value fell within the range of values  calculated by
       Method 2 depended on the arbitrary selection of BCF.   For example,
       "the lowest bivalue BCF" (750 reported for quahog) was selected  as
       the most conservative BCF for cadmium,  whereas for mercury  a
       conservative BCF of 1,000 was selected which is the  value
       "intermediate between that of lobsters and oysters exposed  to
       inorganic mercury"; while for PCBs the BCF chosen for  setting the
       Guideline limit for worms was an approximate median  value (104
       selected from a range of 10 -10s).

REVIEW OF DECISION GUIDELINE LEVELS

The Decision Guidelines were viewed by their authors as  "a  dynamic tool
which will be frequently reviewed and modified as additional  data  and more
detailed analyses become available" (COE and EPA, 1980). According  to
Murphy (1980, p. 46), EPA expected the PCB Interim Decision Matrix to be
usable for only about a year.  Nevertheless, the Decision Guideline  values
have not been revised since their development, and the New  York District
Corps of Engineers continues to use the Decision Guidelines in their
original  form.

The currently-available data on residues of cadmium, mercury, PCBs,  and DDT
and its metabolites (quantitative data for total naphthalenes were scarce)
in New York Bight biota were assembled for the purpose of comparing  the
Decision Guideline levels to our best estimate of actual average con-
taminant body burdens (Appendix A).  To be consistent with  the Decision
Guidelines, only invertebrate species were included in the  analysis. The
grand mean of existing invertebrate body burdens was calculated for  each
contaminant, and the mean values are recorded in Table 7 along with  the
Decision Guideline levels.  Separate mean PCBs concentrations were cal-
culated for polychaete worms and for all other invertebrates  for comparison
to the two corresponding Decision Guideline concentrations.   Also, since
DDT residues measured in blue mussels (Mytil us edulis) were anomolously
high compared to levels reported in other species, a mean DDTs concen-
tration was calculated for all invertebrates except blue mussel.
                                    3-10

-------
                                      TABLE  7

                      COMPARISON OF DECISION GUIDELINE  LEVELS
                         WITH AVERAGE CONTAMINANT  RESIDUES
                          IN NEW YORK BIGHT  INVERTEBRATES
    Chemical
Decision Guideline
  Concentration
  (mg/kg, wet)
Mean Concentration
   in all NYB
  Invertebrates
  (mg/kg, wet)
Mean Concentration
   in other NYB
Invertebrate Groups
   (mg/kg, wet)
Total PCBs
O.la
0.4b
0.193 (n=325)
0.17d (n=304)
0.528e'c (n=21)
Cadmium
     0.3
  0.337C (n=281)
Mercury
     0.2
  0.099 (n=124)
Total Naphthalenes
     0.02
DDT + Metabolites
     0.04
  0.073 (n=106)
   0.033r (n=87)
Decision Guideline level  for PCBs in Palaemonetes  sp.  or Mercenaria mercenaria.

Decision Guideline level  for PCBs in Nereis  sp.  or Neanthes  sp.  The Decision
 Guideline concentrations  for contaminants  other  than PCBs  are applicable for all
 test species.
£
 No statistical  difference («= 0.05) between the mean  residue and the Decision
 Guideline level.  All  other computed mean  residues are statistically different
 from the Decision Guideline concentrations.

 Mean concentration of PCBs in NYB invertebrates  other  than polychaete worms.
P
 Mean concentration of PCBs in NYB polychaete worms.

fMean concentration of DDTs in NYB invertebrates  excluding  blue mussel.
                                       3-11

-------
The computed mean contaminant concentrations in New York  Bight  biota were
statistically compared to the Decision Guideline levels using the  t-test.
At the 5 percent significance level,  the residue means were statistically
different from the Decision Guideline concentrations in every case except
two:  the mean cadmium concentration  for all invertebrates  and  the mean
PCBs concentration in polychaete worms only.  However, the  differences
between calculated mean tissue concentrations of contaminants in New York
Bight biota and the Decision Guideline levels are not large,  and,  in
general, the Decision Guidelines are  well  supported by the  present analysis
of the literature data.

SEDIMENT QUALITY CRITERIA

The Decision Guidelines were developed with the interim goal  of preventing
further degradation of the New York Bight, but the ultimate goal of
regulation is to effect improvement of the Bight ecosystem  by mitigating
the environmental stress caused by dredged material disposal.  Preferable
to the Decision Guidelines approach is to base regulation of contaminated
sediments on biological effects.

EPA has published water quality criteria for many toxic chemicals, based on
an extensive data base that relates biological effects to aqueous  con-
taminant concentrations (EPA, 1980, 1983).  Analogous sediment  quality
criteria have not been established.

A number of problems hinder the development of sediment quality criteria.
Because a significant portion of sediment-associated contaminants  may not
be bioavailable and, therefore, not contribute to biological  effects, it is
inappropriate to base sediment quality criteria on total  (bulk) concentra-
tions of contaminants in sediments.  Consequently, the bioavailability of
sediment-associated contaminants needs to be evaluated.  Bioaccumulation is
currently the only reliable measure of the bioavailabillty  of contaminants
in sediments.  Unfortunately, at the  present, scientists  do not know the
precise relationship between tissue concentrations of contaminants and
their biological consequences, although the relationships between  contam-
inant body burdens and biological  effects are beginning to  be investigated.

                                   3-12

-------
The lack of residue/effects data  makes  It  difficult to identify threshold
concentrations which discriminate between  acceptable  and unacceptable
levels of sediment contamination.

A number of alternatives for establishing  sediment quality criteria have
been proposed.  Pavlou and Weston (1983) summarized several methods for the
establishment of sediment criteria,  including  options based on:  (1) back-
ground levels, (2) water quality  criteria,  (3)  biological response, and (4)
equilibrium partitioning.

BACKGROUND LEVELS

This approach considers any increase in contaminant concentrations above
background levels undesirable.  Background levels are determined from
surficial sediments from areas isolated from known pollutional sources or
from pre-anthropogenically contaminated strata of deep sediment cores.
Implementation of sediment criteria  based  on background levels would
require either (1) prohibiting any increase in contaminant concentrations
over background levels or (2) allowing  some increment of contamination
above background levels.  The former alternative would be unnecessarily
restrictive for some chemicals and the  latter  alternative would require
identification of threshold levels of sediment contamination—i.e.,
require the existence of some other type of sediment  criterion, making the
use of background levels superfluous (Chapman, 1984).

WATER QUALITY CRITERIA

This approach applies EPA ambient water quality criteria to  interstitial
and elutriate waters.  The approach 1s  appealing because it  makes  use of
the extensive data base on biological effects  underlying the water quality
criteria.  The application of water quality criteria  to  interstitial/
interfaclal waters is supported by the  findings of Adams et  al.  (1983) who
concluded that "[t]oxic effects can be expected to occur in  benthic
invertebrates only if the chemical concentration is high enough  in the
sediments such that the equilibrium interstitial water concentration
                                    3-13

-------
reached by desorption is equal  to or higher  than  the concentration
demonstrated to cause an effect in a water exposure test."

However, this approach assumes  that the  predominant route of exposure is
via the interstitial  water and  water at  the  sediment/water  interface, and
neglects exposure following ingestion of sediment particles as an important
exposure route.  Consequently,  direct application of water  quality criteria
to interstitial/interfacial waters is not appropriate  for any chemical  for
which dietary uptake from sediment is a  significant avenue  of exposure.
Chapman (1984) suggested that a simple model  similar to  the one  described
in Section 2 (Table 1) can be developed  for  sediment-dwelling organisms.
Such a model could qualitatively indicate the relative significance  of
uptake from ingested sediment particles.

In addition, the water quality  criteria  approach  is hampered by  the
difficulties of sampling and analyzing contaminants in interstitial  water
and by the unavailability of water quality criteria for  a myriad of  organic
chemicals.

BIOLOGICAL RESPONSE

This approach involves the use  of laboratory bioassays and  biological field
surveys to assess the biological response to contaminated sediments.  Solid
phase bioassays (e.g., EPA and  COE, 1977) are valuable in assessing  the
toxicity of a particular sediment, but they  do not indicate the  contami-
nants) responsible for toxicity and, therefore,  provide no guidance for
setting regulatory criteria.  Bioassays  conducted with otherwise clean
sediments spiked with a known contaminant concentration  could be used to
develop sediment quality criteria just as aqueous bioassay  data  were used
to develop water quality criteria.  However, because  of  the general  paucity
of this type of data and the wide variety of possible  sediment character-
istics, development of sediment criteria using this approach  is  unlikely to
be possible in the near future.  The utility of biological  field surveys in
sediment criteria development is limited because:  they  are site specific,
natural biological variability can obscure  pollutional effects,  and
negative environmental effects are observable only after the  fact.

                                    3-14

-------
EQUILIBRIUM PARTITIONING


The equilibrium approaches discussed by Pavlou  and  Weston  (1963) use
equilibrium partitioning constants,  which can be derived theoretically or

empirically, to determine contaminant concentrations  in  various environ-

mental compartments (water, sediment, and biota).   Water quality criteria

or FDA action levels are used to identify threshold levels of  contaminants
1n those environmental  compartments.  Pavlou  and Weston  (1983) discussed

four potential approaches for establishing sediment criteria based  on

equilibrium partitioning (Figure 12):


    Approach #1. uses KD and the existing EPA water quality criteria
    (applied to interfacial water) to compute a sediment threshold
    concentration.  This calculated value is  then compared to  the
    actual measured contaminant concentration in the  ambient sediment
    of a designated site to estimate the extent of  violation.  Con-
    versely, using the ambient sediment burden  and  KD the  interfacial
    water concentration can be computed and compared  directly  to the
    appropriate EPA water criteria value.

    Approach #2, is based on the application  of water quality  criteria
    to Interfacial waters and the use of the  BCF to compute a  biota
    burden.  The sediment threshold value can then  be calculated by
    using the ARS constant.  Indirectly, the  computed biota burden  may
    be compared to an existing body burden level known to  induce a
    toxic effect.

    Approach #3, uses the ARS quantity with either  a  measured  body
    burden which induces a toxic effect (e.g.,  pathologic, behavioral
    or metabolic effect) or a federally established tissue concentra-
    tion (if public health risks are the prime  consideration)  to
    compute a sediment threshold level.

    Approach #4, is a combination of the above.  It establishes a
    biological threshold concentration and determines a corresponding
    sediment threshold value via contaminant  transfer through  the
    aqueous phase.


Pavlou and Weston (1983) noted several limitations  to the  establishment  of

sediment criteria based on the equilibrium partitioning approaches
described:
       Threshold levels based on water quality criteria do not consider
       direct transfer of contaminants from sediment to biota (e.g.,
       following ingestion of sediment particles);
                                    3-15

-------
wactual


Wcrlterla
            I
Sactual


Scrlterla
                                                                           B
                                                                            threshold
                                                         W
                                                           criteria
                                                                    BCF
                                                       I
                                                     B
                                                      criteria
                                                                ARS
                                                     'criteria
OJ
I
(ft
      B
       actual
                   APPROACH
                       B
                         criteria
   I
B
 threshold
                                  AR
'threshold
                   APPROACH +3
                                                 APPROACH +2
                                                          B
                                                           actual
                           B
                                                                       threshold
                                                                       BCF
                                                            'threshold
                                                                                      Kr^W"
                                                                                                'threshold
                                                APPROACH
                                  KD <=KW,S) = Sediment/Water Partition Coefficient
                                ARS (=K s'b) = Accumulation Relative to Sediment
                                BCF (=Kw',b) = Bioconcentration Factor
          Figure 12.  Schematic representation of sediment criteria development using equilibrium  partitioning
                    constants.  Source:  Pavlou and Weston, 1983.

-------
    •  The use of multiple partition coefficients increases  the  potential
       for error;

    •  The sediment/water partition coefficients (KQ)  determined empirical-
       ly from partitioning between suspended participate matter and water
       (and tabulated by Pavlou and Weston,  1983) are  probably not  applic-
       able to partitioning between sediment and interstitial water because
       of the different physico-chemical  conditions in sediments; and

    •  The biota/sediment partition coefficient (ARS)  does not address the
       bioaccumulation mechanism.
Also, appropriate biota criteria are generally unavailable.   Insufficient
data exists to relate toxic effects to residue concentrations.  FDA action

levels are one type of biota criteria, but they have been  established  for

only a few chemicals and they are designed to protect only human  consumers
of fish and shellfish, not the health of the aquatic community.


Pavlou and Weston (1984) used sediment/water partitioning  (Approach #1)  to

develop sediment criteria for toxic contaminants in marine waters (Puget
Sound).  A sediment criterion was defined as the concentration of a con-

taminant in a sediment which insures that its concentration in interstitial
water does not exceed the EPA water quality criterion.   The expression for

the sediment/water partition coefficient normalized for organic carbon
                                                                       (32)
where
              = contaminant concentration expressed in units  of mass
                of contaminant/mass of organic carbon,

          CjW - contaminant concentration in interstitial  water,
           KD = sediment/water partition coefficient (Kws).
          TOC - total  organic carbon content of the sediment
                expressed as fractional  mass on a dry weight  basis.
                                    3-17

-------
was rearranged and modified by substituting the EPA water quality criterion
 (Cw/cr) for CIM.  Sediment criteria (Cs/cr/oc) were derived by the result-
 ing equation:

                Cs/cr/oc = Koc cw/cr                                   (33)

The general sediment criteria are applied to a specific site by multiplying
Cs/cr/oc ky tne organic carbon content of the sediment of concern, to
obtain criteria that can be directly compared to contaminant concentrations
in the sediment (Cs/cr):

                Cs/cr ' cs/cr/oc TOC                                   <34>

The method described above can be used to develop sediment criteria for
both organic and inorganic contaminants, providing that a water quality
criterion is available.  Pavlou and Weston (1984) estimated KQC values  for
organic contaminants theoretically using a KOC-KOW regression equation
developed for organic contaminants of concern in Puget Sound.  KQC values
for trace metals were derived empirically by taking the mean of KQC values
calculated for a variety of sediments from measurements of trace metal
concentrations measured in bulk sediments and interstitial  water reported
in the literature (Brannon et al., 1980).

The use of KQC instead of the unnormalized sediment/water partition
coefficient is based on the assumption that the organic carbon content  of
sediment is a major environmental  variable affecting the sediment/water
partitioning of contaminants.  While that is clearly the case for synthetic
organic chemicals, for trace metals it is not so evident.  Pavlou and
Weston (1984) evaluated the relationship between the sediment/water
partition coefficient (KD) and sediment organic carbon content for six
trace metals.  Three of the metals evaluated (Cu, Cd, Pb) exhibited
statistically significant (albeit weakly so) correlations between KD and
sediment organic carbon, while three metals (Zn, As, Hg) did not show
significant relationships.
                                    3-18

-------
Pavlou and Weston (1984)  derived sediment criteria  from  both  acute and
chronic water quality criteria,  where available.  However, acute sediment
criteria are generally considered to be inappropriate  because of the
reasons cited by the authors:

    •  Sediment contaminant concentrations reflect  long-term  conditions and
       do not demonstrate the  extreme temporal  variability of water column
       concentrations.
    •  Benthic organisms  often lack the mobility  required to  escape a
       contaminated environment and therefore  are susceptible to impacts
       resulting from long-term chronic exposure.

Thermodynamic estimation  of the movement of hydrophobic  orgam'cs from
sediment to water may create an unnecessarily  high  potential  for error due
to environmental variability (Peddicord and McFarland, personal communi-
cation).  Hydrophobic organic  compounds are predominantly associated with
organic carbon in sediments.  The solubility of hydrophobic orgam'cs in
sediment organic carbon is much greater (orders of  magnitude) than their
solubility in water.  Chemical solubility in any  environmental compartment
is affected by a variety  of environmental factors,  including  pH, redox
potential, temperature, salinity, suspended particulates, organic matter
content, etc.  Consequently, the effect of environmental variability on the
solubility of hydrophobic orgam'cs is low compared  to  their solubility in
organic carbon, but is very high compared to their  aqueous solubility.
Thus, there is a relatively high probability of significant error
propagation associated with partitioning between  sediment and water.

In contrast, the solubility of hydrophobic organic  compounds  in organic
carbon is comparable to their solubility in lipid tissue.  Hence, the
probability of error associated with thermodynamic  estimation of the
movement of chemicals between  sediment and tissue is relatively low.
Karickhoff (1984) described a method by which  the potential for bioaccum-
ulation of hydrophobic orgam'cs can be estimated  directly from sediment
concentrations without using the aqueous phase as an intermediate step (see
Section 2).  Since hydrophobic organics in the sediment  are primarily
associated with organic carbon and those chemicals  in  biota are primarily
                                    3-19

-------
associated with lipid tissue, the thermodynamic  potential  for bioaccumu-
lation can be estimated from the organic carbon  content  of the  sediment
(XQC) the lipid content of the animal  (X-j)  and the  lipid/organic carbon
partition coefficient (Kocj):
EPA and the Corps of Engineers are working toward the development of  sedi-
ment quality criteria, using the theory behind Equation  21.

Workers at the U.S. EPA Environmental  Research Laboratory— Narragansett are
developing a thermodynamic equilibrium model  for predicting  maximum
residues of organic compounds accumulated by  aquatic organisms  exposed to
dissolved and suspended parti culate-bound contaminants (Lake et al.,  1984).
Based on the fugacity model  described by Mackay (1979),  Lake et al.  (1984)
suggest that the bioaccumulation factor (BAF), at thermodynamic equilib-
rium, may be defined as:
          BAF = C/(C  + Cc)                                          (35)
where
           Ce = equilibrium contaminant concentration in  lipid  (1), water
                (w), or organic carbon (oc)

Or, since the fugacity (escaping tendency) of a chemical  in any phase  is
given by:

       fphase = "phase Yphase fR                                       (36)
                                    3-20

-------
where

            f = fugacity
            x = mole fraction
            y = activity coefficient
           fR = reference fugacity

the concentration terms in Equation 35 can be converted to mole  fractions
and the expression for BAF can be rewritten as:

          BAF = (y  + O/Y-I                                          (37)
                  W    OC    I

For conditions in which the aqueous phase is not an important storage
compartment, as with hydrophobic oranic chemicals, the BAF will  depend only
on the contaminant concentration in the organic carbon of suspended
particulate material and in the Hpids of biota.  Thus, the thermodynamic
maximum whole-body tissue concentration can be predicted by (Lake et al.,
1984):

           re   Xl re                                                  (38)
            b  " 1T~ LSPM
 or,  if the  lipid/organic carbon partition coefficient (KQCj) were deter-
 mined rather than  being assumed approximately equal to one (Karickhoff,
 1984):
            re - J.  K      ce                                           (39)
             b  *oc  oc>1   SPM
 where
            cj = contaminant concentration  in  the  organism at equilibrium

          CSPM = contaminant concentration  in  the  SPM  at equilibrium
            X1 = lipid weight of the organism

                                     3-21

-------
          X   = organic carbon weight of the  suspended  solids
        K     = lipid/organic carbon partition  coefficient
         OC 9 I

Karickhoff (1984) suggested that this model can be referenced  to  the
sediment, for hydrophobic organics, by:
                «             r®
           re - -1 K         .  .s                                       (40)
           Cb - XocKoc,l 1
where
           C* = contaminant concentration in sediment at equilibrium
         aSPM = susPended solids concentration
          KOC = sediment-organic-carbon/water partition coefficient

The simple partitioning model described by Lake et al. (Equation 38) is
designed to estimate maximum concentrations of organic chemicals in biota.
After the model is tested, the authors propose to use the model as a first
level screen to assess the advisability of ocean disposal of a given waste
(e.g., dredged material, sewage sludge).  If the theory is correct, pre-
dicted maximum concentrations should be greater than, or at least equal to,
residues accumulated in the environment, thereby providing a conservative
estimate for environmental protection regulatory purposes.

Workers at the U.S. Army Engineer Waterways Experiment Station similarly
are developing a method for establishing sediment quality criteria for
hydrophobic and neutral organic compounds based on thermodynamic equi-
librium  (maximum)  residues  (Peddicord, 1984).  Thermodynamic equilibrium
concentrations are limited by the basic laws of thermodynamics and
therefore cannot be exceeded in the environment.  Thus,  sediment quality
criteria set as concentrations of contaminants in sediment which, at
thermodynamic equilibirum, will not result  in unacceptable residues in
biota will  insure  that any  sediment with chemical concentrations below the
criteria is environmentally  acceptable.  Sediments with  chemical concen-
trations above such criteria are potentially unacceptable, but because

                                    3-22

-------
thermodynamic equilibrium concentrations are seldom reached in the  envi-
ronment, more detailed Investigation of bioaccumulatlon Is necessary  to
assess their impact potential.  Sediment quality criteria based on
equilibrium partitioning can be expressed mathematically as (Peddicord,
1984):

      log SQC = (log CT - 2.28) + log TOC                              (41)

where

         SQC =  sediment quality criterion for the particular compound  on
                total sediment, dry weight basis, in same units
                as CT,
          Cj =  acceptable tissue concentration for the particular
                compound, expressed on a lipid basis,
         TOC =  percent (not decimal fraction) total organic carbon in
                the sediment in question,
        2.28 =  factor accounting for the relative activities of
                hydrophobia or neutral compounds in TOC and in lipid, and
                for expressing sediment concentrations  on a % TQC basis.

As noted earlier,  appropriate, acceptable tissue concentrations have  not
been established,  and insufficient data are presently available to  relate
biological effects to contaminant residues.  In the absense of an adequate
data base on residue/effects, the COE has elected to use the product  of an
EPA chronic water quality criterion and an estimated BCF to calculate an
acceptable tissue  concentration (CT) for a given chemical.   Peddicord
(1984) used the BCF - K-w regression equation of Konemann and van Leeuwen
(1980) to estimate BCF:

      log BCF = 0,980 log KQW - Q.063                                  (42)

Thus, CT can be estimated as:

       log Cy = 0.980 log KQW - 0.063 + log WQC - 3.0                  (43)
                                    3-23

-------
 where
           Cy = acceptable tissue concentration for the compound
                expressed on a lipid basis in units of ug/kg,
          KOW = octanol-water partition coefficient for the compound of
                concern,
          WQC = chronic water quality criterion for the compound in
                units of ug/1,
          3.0 = factor to convert the equation to units of ug/1.

Combining Equations 41 and 43, a sediment quality criterion for a contami-
nant in a particular sediment can be calculated from the knowledge of the
chronic water quality criterion and the octanol/water partition coefficient
for the given chemical and the organic carbon content of the given sediment
as follows:

      log SQC = (log WQC + 0.980 log KQW - 5.346) + log TOC            (44)

where

          SQC = sediment quality criterion for the compound of concern,
                dry weight basis, units of ug/kg,
          WQC = water quality criterion for the compound of concern,
                units of ug/1,
          KOW = octanol/water partition coefficient for the compound of
                concern
          TOC = total organic carbon content of the sediment,  dry weight
                basis, units of % (not decimal  fraction).

This procedure is part of a tiered approach to sediment assessment recom-
mended by the COE (Peddicord and McFarland, personal  communication)  in
which the thermodynamic equilibrium method expressed in Equation 44  is
applied as the first tier screening level.  If the contaminant concentra-
tion in the sediment of interest is lower than the concentration computed
by Equation 44,  the sediment is considered environmentally acceptable;  if
the ambient concentration is higher than the computed level, the sediment
may or may not be acceptable.  When the latter case exists,  the sediment
                                    3-24

-------
must undergo tier 2 testing in which bioaccumulation potential  is  assessed
via empirical steady-state exposure or kinetic modeling.

Peddicord (1984) noted that the weakest scientific point  of the COE
approach is the present need to use water quality criteria to estimate  Cp
based on the assumption that chronic water quality criteria are comprehen-
sively protective of all  aspects of aquatic ecology.  In  addition, chronic
criteria have only been established for a few hydrophobic or neutral
organic chemicals.

It is important to note that the COE thermodynamic equilibrium approach can
be used to develop sediment quality criteria only for hydrophobic  or
neutral organic compounds and cannot be used for inorganics, water soluble
organics, or compounds which associate with sediment primarily via elec-
trostatic interactions.
                                    3-25

-------
                                 REFERENCES
Adams, W.J.  1984.  Bloava11 ability and Safety Assessment of Lipophilic
Organics Sorbed to Sediments.  Draft paper presented at the Workshop  on  the
Role of Suspended and Settled Sediments in Regulating the Fate and Effects
of Chemicals in the Aquatic Environment, 13-17 August 1984, Florissant,  CO.

Adams, W.J., R.A. Kimerle, and R.G. Mosher.  1983.   Aquatic Safety
Assessment of Chemicals Sorbed to Sediments.   Paper presented at ASTM
Seventh Annual Aquatic Toxicology Symposium,  Milwaukee, MI, 17-19 April
1983.

Banner, L.H. and J.L. Oglesby.  1982.   Models for Predicting Bioaccum-
ulation and Ecosystem Effects of Kepone and Other Materials.  In:
Environmental Risk Analysis for Chemicals, R.A. Conway (ed.), Van Nostrand
Reinhold Co., New York, N.Y.

Blau, G.E., VI.B. Neely, and D.R. Branson.   1975.  Ecokinetics:   A Study  of
the Fate and Distribution of Chemicals in  Laboratory Ecosystems.  Amer.
Inst. Chem. Engin. J., 21:854-861.

Boehm, P.D.  1982.  Organic Pollutant Transforms and Bioaccumulation  of
Pollutants in the Benthos from Waste Disposal-Associated Sediments.   Report
prepared by Energy Resources Co. Inc.  for  NOAA.  Rockville, MD.

Boehm, P.D. and P. Hirtzer.  1982.  Gulf and Atlantic Survey for Selected
Organic Pollutants in Finfish.  NOAA Tech. Memo.  NMFS-F/NEC-13.  NOAA,
Woods Hole, MA.

Brannon, J.M., R.H. Plumb, Jr., and I. Smith, Jr.  1980.  Long-Term Release
of Heavy Metals from Sediments.  In:  Contaminants  and Sediments, Vol. 2,
R.A. Baker (ed.), Ann Arbor Science, Ann Arbor, MI.

Branson, D.R., G.E. Blau, H.C. Alexander,  and W.B.  Neely.  1975.
Bioconcentration of 2, 2', 4, 4' - Tetrachlorobiphenyl  in Rainbow Trout  as
Measured by an Accelerated Test.  Trans. Am.  Fish.  Soc., 104:785-792.

Briggs, G.G.  1973.  A Simple Relationship Between  Soil Sorption of Organic
Chemicals and Their Octanol/Water Partition Coefficients.  In:   Proc. 7th
British Insecticide and Fungicide Conf. Vol.  1, The Boots Company,  Ltd.,
Nottingham, G.B.

Brown, D.S., S.W. Karickhoff, and E.W. Flagg.  In preparation.   Empirical
Prediction of Organic Pollutant Sorption in Natural  Sediments.

Bruggeman, W., L. Martron, D. Kooiman, and 0. Hutzinger.  1981.   Accumu-
lation and Elimination Kinetics of Di-, Tri-  and Tetra Chlorobiphenyls by
Goldfish after Dietary and Aqueous Exposure.   Chemosphere, 10:811-832.

-------
Brungs, W.A. and D.I. Mount.  1978.  Introduction to a Discussion  of the
Use of Aquatic Toxicity Tests for Evaluation of the Effects of Toxic
Substances.  In:  Estimating the Hazard of Chemical Substances to  Aquatic
Life, J. Cairns, Jr., et al. (eds.) ASTM STP 657.  Philadelphia, PA.

Chapman, G.A.  1984.  A Review of Procedures for Evaluating the Signifi-
cance of Sediment Contamination and Establishing Sediment Criteria.   Draft
paper presented at the Workshop on the Role of Suspended and Settled
Sediments in Regulating the Fate and Effect of Chemicals in the Aquatic
Environment, 13-17 August 1984, Florissant, CO.

Chiou, C.T., V.H. Freed, D.W. Schmedding, and R.L. Kohnert.  1977.
Partition Coefficient and Bioaccumulation of Selected Organic Chemicals.
Environ. Sci. Technol., 11:475-478.

Chiou, C.T., L.J. Peters, and V.H. Freed.  1979.  A Physical  Concept of
Soil-Water Equilibria for Nonionic Organic Compounds.  Science,
206:831-832.

COE and EPA.  1980.  Interim Guidance Matrix (Decision Guidelines  for
Polychlorinated Biphenyls in the New York Bight).

COE and EPA.  1982.  Guidance for Performing Tests on Dredged Material to
be Disposed of in Ocean Waters.  U.S. Army Corps of Engineers, New York
District, NY.

COE and EPA.  Undated  a. Decision Guidelines for Cadmium and Mercury in
the New York Bight.

COE and EPA.  Undated  b. Decision Guidelines for DDT and Metabolites in
the New York Bight.

COE and EPA.  Undated  c. Decision Guidelines for Petroleum Hydrocarbons  in
the New York Bight.

Connor, M.S.  1984.  Fish/Sediment Concentration Ratios for Organic
Compounds.  Environ. Sci. Technol., 18:31-35.

Courtney, W.A.M. and W.J. Langston.  1978.  Uptake of Polychlorinated
Biphenyl (Aroclor 1254) from Sediment and from Seawater in Two Intertidal
Polychaetes.  Environ. Poll., 15:303-309.

Davies, R.P. and A.J. Dobbs.  1984.  The Prediction of Bioconcentration in
Fish.  Water Res., 18:1253-1262.

EPA.  1980, 1983.  Ambient Water Quality Criteria Documents.   U.S. EPA,
Washington, D.C.

EPA and COE.  1977.  Ecological Evaluation of Proposed Discharge of  Dredged
Material into Ocean Waters.  Implementation Manual for Section 103 of
Public Law 92-532 (Marine Protection, Research, and Sanctuaries Act  of
1972), U.S. Army Waterways Experiment Station, Vicksburg, MS.

-------
Ernst, W.  1977.  Determination of the Bioconcentration Potential  of  Marine
Organisms - A Steady State Approach.  I.   Bioconcentration Data  for Seven
Chlorinated Pesticides In Mussels (Mytil us edulls)  and Their Relation to
Solubility Data.  Chemosphere, 11:731-740.

Fowler, S.W., G.G. Polikarpov, D.L.  Elder, P.  Parsi,  andJ.P.  Villeneuve.
1978.  Polychlorinated Biphenyls:  Accumulation from  Contaminated  Sediments
and Water by the Polychaete Nereis dlversicolor.  Marine Biology,
48:303-309.

Hall, R.A., E.G. Zook, and G.M. Meaburn.   1978.  National  Marine Fisheries
Service Survey of Trace Elements in the Fishery Resource.   NOAA  Technical
Report NMFS SSRF-721.

Halter, M.T. and H.E. Johnson.  1977.  A Model System to Study the
Desorption and Biological Availability of PCB  in Hydrosoils.  In:   Aquatic
Toxicology and Hazard Evaluation, F.L. Mayer and J.L. Hamelink (eds.), ASTM
STP 634, Philadelphia, PA.

Hamelink, J.L.  1977.  Current Bioconcentration Test  Methods and Theory.
In:  Aquatic Toxicology and Hazard Evaluation, F.L. Mayer and J.L. Hamelink
(eds.), ASTM STP 634, Philadelphia,  PA.

Hansch, C., J.E. Quinlan, and G.L. Lawrence.  1967.  J. Org. Chem. 33:347.

Hassett, J.J., J.C. Means, VI.L. Banwart,  and S.G. Wood.  1980.  EPA
600/3-80/041, Athens, GA.

House of Representatives Committee on Merchant Marine and Fisheries.   1980.
Questions posed to the COE at the Hearings before the Committee  on Merchant
Marine and Fisheries, House of Representatives, Ninety-Sixth Congress,
Second Session on Exploring the Various Aspects Related to the Dumping of
Dredged Spoil Material in the Ocean and the PCB Contamination Issue.
Serial No. 96-43, 14 March, 21 May 1980.

Jarvinen, A.W., M.J. Hoffman, and T.W. Thorslund.  1977.  J. Fish. Res. Bd.
Can. 34:2089-2103.

Jensen, A.L., S.A. Spigarelli, and M.M. Thoiranes.   1982.  PCB Uptake by Five
Species of Fish in Lake Michigan, Green Bay of Lake Michigan,  and  Cayuga
Lake, New York.  Can. J. Fish. Aquat. Scl., 39:700-709.

Karickhoff, S.W.  1981.  Semi-Empirical Estimation  of Sorption of
Hydrophobia Pollutants on Natural Sediments and Soils.  Chemosphere,
10:833-846.

Karickhoff, S.W.  EPA-ERL Athens, GA.  1984.  Bioconcentration Potential:
Sediment Sources.  Letter to V. McFarland, COE-WES, Vicksburg, MS.

Karickhoff, S.W., D.S. Brown, and T.A. Scott.   1979.   Sorption of
Hydrophobic Pollutants on Natural Sediments.  Water Research,  13:241-248.

-------
Kenaga, E.E. and C.A.I. Goring.  1980.  Relationship Between Water
Solubility, Soil Sorptlon, Octanol-Water Partitioning, and Concentration of
Chemicals in Biota.  In:  Aquatic Toxicology, J.G. Eaton, P.R.  Parrish,  and
A.C. Hendricks  (eds.), ASTM STP 707, Philadelphia, PA.

Kobayashi, K.   1981.  Proceedings of OECD Workshop on the Control  of
Existing Chemicals, pp. 141-163, Unwelltrandesant, Berlin.

Konemann, H. and K. van Leeuwen.  1980.  Toxicokinetlcs in Fish:   Accum-
ulation and Elimination of Six Chlorobenzenes by Guppies.  Chemosphere,
9:3-19.

Kopfler, F.C.   1974.  The Accumulation of Organic and Inorganic Mercury
Compounds by the Eastern Oyster (Crassostrea virginica).   Bull. Environ.
Contain. Toxicol., 11:275-280.

Krzeminski, S.F., J.T. Gilbert, and J.A. Ritts.  1977.  A Phamacokinetic
Model for Predicting Pesticide Residues in Fish.  Arch. Environ.  Contarn.
Toxicol., 5:157-166.

Lake, J.L., N. Rubinstein, and S. Pavignano.  1984.  Predicting Bio-
accumulation:  Development of a Simple Partitioning Model for Use as a
Screening Tool for Regulating Ocean Disposal of Wastes.  Draft  paper
presented at the Workshop on the Role of Suspended and Settled  Sediments in
Regulating the Fate and Effects of Chemicals in the Aquatic Environment,
13-17 August 1984, Florissant, CO.

Lee, G.F. and R.A. Jones.  1977.  An Assessment of the Environmental
Significance of Chemical Contaminants Present in Dredged Sediments Dumped
in the New York Bight.  A report to the New York District Corps of
Engineers.  Piano, TX.

Lu, P.Y. and R.L. Metealf.  1975.  Environmental Fate and Biodegradability
of Benzene Derivatives as Studied in a Model Aquatic Ecosystem.  Environ.
Health. Persp., 10:269-284.

Lyman, W.J., W.F. Reehl, and D.H. Rosenblatt.  1982.  Handbook  of Chemical
Property Estimation Methods.  Environmental Behavior of Organic Compounds.
McGraw-Hill, New York, NY.

Lynch, T.R. and H.E. Johnson.  1982.  Availability of a Hexachlorobiphenyl
Isomer to Benthic Amphipods from Experimentally Contaminated Natural
Sediments.  In:  Aquatic Toxicology and Hazard Assessment:  Fifth
Conference, J.G. Pearson, R.B. Foster, and W.E. Bishop (eds.),  ASTM STP
766, Philadelphia, PA.

Macek, K.J., S.R. Petrocelli, and B.H. Sleight, III.  1979.   Considerations
in Assessing the Potential for, and Significance of, Biomagm'fication of
Chemical Residues in Aquatic Food Chains.   In:   Aquatic Toxicology, L.L.
Marking and R.A. Kimerle (eds.), ASTM STP  667,  Philadelphia,  PA.

Mackay, D.  1979 .  Finding Fugacity Feasible.   Environ.  Sci. Technol.,
13:1218-1223.

-------
Mackay, D.  1982.  Correlation of Bioconcentration Factors.   Environ.  Sci.
Technol., 16:274-278.

Mackay, D., A. Bobra, and VI.Y. Shiu.  1980.   Relationships Between  Aqueous
Solubility and Octanol-Water Partition Coefficients.   Chemosphere,
9:701-711.

Mackay, D. and S. Paterson.  1981.  Calculating Fugacity.   Environ.  Sci.
Technol., 15:1006-1014.

Mackey, D. and S. Paterson.  1982.  Fugacity Revisited.   Environ. Sci.
Technol., 16:654A-660A.

MacLeod, W.D., Jr., L.S. Ramos, A.J. Friedman,  D.G. Burrows,  P.G. Prohaska,
D.L. Fisher, and D.W. Brown.  1981.  Analysis of Residual  Chlorinated
Hydrocarbons, Aromatic Hydrocarbons and Related Compounds  in  Selected
Sources, Sinks, and Biota of the New York Bight.  NOAA Tech.  Memo.  OMPA-6.
NOAA, Boulder, CO.

McFarland, V.A., A.B. Gibson, and I.E. Meade.  1984.   Application of
Physicochemical Estimation Methods to Bioaccumulation from Contaminated
Sediments.  II.  Steady-State from Single Time-Point Observations.   In:
Applications in Water Quality Control—Proceedings, U.S.  Army Hydrologic
Engineering Center, Davis, CA.

McLeese, D.W., C.D. Metcalfe, and D.S. Pezzack.  1980. Uptake of PCBs  from
Sediment by Nereis virens and Crangon septimosa.  Arch. Environ.  Contam.
Toxicol., 9:bU/-bl8.

Means, J.C., S.G. Wood, J.J. Hassett, and W.L.  Banwart.   1980.  Sorption  of
Polynuclear Aromatic Hydrocarbons by Sediments  and Soils.   Environ.  Sci.
Technol., 14:1524-1528.

Metcalf, R.L., I.P. Kapoor, P.Y. Lu, C.K. Schuth, and P.  Sherman.  1973.
Model Ecosystem Study of the Environmental Fate of Six Organochlorines.
Environ. Health Persp., June, 33-44.

Metcalf, R.L., J.R. Sanborn, P.Y. Lu, and D. Nye.  1975.   Laboratory Model
Ecosystem Studies of the Degradation and Fate of Radio!abeled Tri-,  Tetra-,
and Pentachlorobiphenyl Compared with DDE.  Arch. Environ. Contam.
Toxicol., 3:151-165.

Murphy, J.H.  1980.  Comments at the Hearings before  the Committee  on
Merchant Marine and Fisheries, House of Representatives,  Ninety-Sixth
Congress, Second Session on Exploring the Various Aspects  Related to the
Dumping of Dredged Spoil Material in the Ocean  and the PCB Contamination
Issue.  Serial No. 96-43, 14 March, 21 May 1980.

Neely, W.B.  1979.  Estimating Rate Constants for the Uptake  and Clearance
of Chemicals by Fish.  Environ. Sci. Technol.,  13:1506-1510.

Neely, W.B., D.R. Branson, and G.E. Blau.  1974.  Partition Coefficient to
Measure Bioconcentration Potential of Organic Chemicals in Fish.  Environ.
Sc1. Technol., 13:1113-1115.

-------
Norstrom, R.J., A.E.  McKinnon,  and A.S.W.  DeFreitas.   1976.  A
Bioenergetics-Based Model  for Pollutant Accumulation  by  Fish.   Simulation
of PCB and Methylmercury Residue Levels in Ottawa  River  Yellow  Perch  (Perca
flavescens).  J. Fish.  Res.  Bd. Can.,  33:248-267.

O'Brien & Gere Engineers,  Inc.   1979.   Bioaccumulation Study on Homarus
americanus.  U.S.  Army  Corps of Engineers, New York District, Syracuse, NY.

O'Connor, J.M., J.B.  Klotz,  and T.J. Kneip.   1982.  Sources, Sinks, and
Distribution of Organic Contaminants  in the New York  Bight Ecosystem.  In:
Ecological Stress and the New York Bight:   Science and Management, G.F.
Mayer (ed.), Estuarine  Research Federation,  Columbia, SC.

Pavlou, S.P. and D.P. Weston.  1983.   Initial  Evaluation of Alternatives
for Development of Sediment Related Criteria for Toxic Contaminants in
Marine Waters (Puget Sound).  Phase I:   Development of Conceptual
Framework.  Report prepared by JRB Associates for  EPA, Seattle, WA.

Pavlou, S.P. and D.P. Weston.  1984.   Initial  Evaluation of Alternatives
for Development of Sediment Related Criteria for Toxic Contaminants in
Marine Waters (Puget Sound).  Phase II: Development  and Testing of the
Sediment-Water Equilibrium Partitioning Approach.  Prepared by  JRB
Associates for EPA, Washington, D.C.

Peddicord, R.K.  1980.   Comments at the Hearings before  the Committee on
Merchant Marine and Fisheries,  House  of Representatives, Ninety-Sixth
Congress, Second Session on Exploring  the  Various  Aspects Related to  the
Dumping of Dredged Spoil Material in  the Ocean and the PCB Contamination
Issue.  Serial No. 96-43,  14 March, 21  May 1980.

Peddicord, R.K.  1984.   Scientifically Defensible  Development and
Implementation of Sediment Quality Criteria (Working  Draft).

Peddicord, R.K. and V.A. McFarland, Environmental  Laboratory, U.S. Army
Engineer Waterways Experiment Station,  Vicksburg,  MS. Personal
communication with CDM, July, 1984.

Pizza, J.C. and J.J.  O'Connor.   1983.   PCB Dynamics in Hudson River Striped
Bass.  II.  Accumulation from Dietary  Sources.  Aquatic  Toxicology,
3:313-327.

Raltech Scientific Services.  1981. Analysis of Marine  Worms for
Chlorinated Insecticides, PCBs, and Heavy  Metals.  Report prepared for the
U.S. Fish and Wildlife  Service, Madison, WI.

Rao, P.S.C. and J.M.  Davidson.   1980.   Estimation  of  Pesticide  Retention
and Transformation Parameters Required in  Nonpoint Source Pollution Models.
In:  Environmental Impact of Nonpoint  Source Pollution,  M.R. Overcash and
J.M. Davidson (eds.), Ann Arbor Science, Ann Arbor, MI.

Roesijadi, G., J.W. Anderson, and J.W.  81 aylock.  1978.  Uptake of
Hydrocarbons from Marine Sediments Contaminated with  Prudhoe Bay Crude Oil:
Influence of Feeding Type of Test Species  and Availability of Polycyclic
Aromatic Hydrocarbons.   J. Fish. Res.  Bd.  Can., 35:608-614.

-------
 Rubinstein, N.I., W.T. Gillian, and N.R. Gregory.  1984.  Dietary
 Accumulation of PCBs from a Contaminated Sediment Source by a Demersal  Fish
 (Lelostomus xanthurus).  Submitted to Aquatic Toxicology.

 Rubinstein, N.I., E. Lores, and N.R. Gregory.  1983.  Accumulation of PCBs,
 Mercury, and Cadmium by Nereis vlrens. Mercenaria mercenarla, and
 Palaemonetes puglo from Contaminated Harbor Sediments.  Aquatic Toxicology,
 3:249-260.

 Spacie, A. and J.L. Hamellnk.  1982.  Alternative Models for Describing the
 Bloconcentration of Organlcs In Fish. Environ. Toxicol. Chem., 1:309-320.

 Spagnoll, J.J. and L.C. Skinner.  1977.  PCBs In Fish from Selected Waters
 of New York State.  Pesticides Monitoring Journal, 11:69-87.

 Swartz, R.C. and H. Lee, II.  1980.  Biological  Processes Affecting the
 Distribution of Pollutants in Marine Sediments.   Part I.  Accumulation,
 Trophic Transfer, Biodegradation and Migration.   In:  Contaminants and
 Sediments.  Vol. 2.  Analysis, Chemistry, Biology, R.A. Baker (ed.), Ann
 Arbor Science, Ann Arbor, MI.

 Thomann, R.V.  1981.  Equilibrium Model of Fate  of Microcontaminants in
 Diverse Aquatic Food Chains.  Can. J. Fish. Aquat. Sc1., 38:280-296.

 Thomann, R.V. and J.P. Connolly.  1984.  Model of PCB in the Lake Michigan
 Lake Trout Food Chain.  Environ. Scl. Technol.,  18:65-71.

 Tulp, M.T.M. and 0. Hutzinger.  1978.  Some Thoughts on Aqueous
 Solubilities and Partition Coefficients of PCB,  and the Mathematical
 Correlation Between Bioaccumulation and Physico-Chemical Properties.
 Chemosphere, 10:849-860.

 Veith, G.D., D.L. DeFoe, and B.V. Bergstedt.  1979.  Measuring and
 Estimating the Bloconcentration Factor of Chemicals in F1sh.  J.  Fish.  Res.
 Bd. Can., 36:1040-1048.

 Veith, G.D., K.J. Macek, S.R. PetrocelH, andJ.  Carroll.  1980.   An
 Evaluation of Using Partition Coefficients and Water Solubility to Estimate
 Bioconcentration Factors for Organic Chemicals in Fish.  In:  Aquatic
 Toxicology, J.G. Eaton, P.R. Parrlsh, and A.C. Hendricks (eds.),  ASTM STP
 701, Philadelphia, PA.

Wang, M.P., G.T. Hickman, and S.A. Dendrou.  1984.  Technical  Review of the
 l2-M1le Sewage Sludge Disposal Site.  Report prepared by Camp Dresser &
McKee for EPA, Washington, D.C.

Weininger,  D.   1978.   Accumulation of PCBs by Lake Trout in Lake  Michigan.
Ph.D. thesis,  University of Wisconsin,  Madison,  WI.

Wenzloff, D.R., R.A.  Grieg,  A.S. Merrill  and J.W. Ropes.  1979.   A Survey
of Heavy Metals in the Surf Clam, Spisula solidissima.  and the Ocean
Quahog, Arctlca Islandica. of the Mid-Atlantic Coast of the United States.
Fishery Bulletin, /7:Z80-285.

-------
Wyman, K.D. and H.B. O'Connors, Jr.  1980.   Implications of Short-term PCB
Uptake by Small Estuarine Copepods (genus Acartia) from PCB-Contaminated
Water, Inorganic Sediments and PhytoplanktorTFst.  Mar. Coast.  Sci.,
11:121-131.

Zdanowicz, V.  Printed data received from Robert Reid,  Sandy Hook
Laboratory, NMFS, Highlands, NJ.
                                     8

-------
              APPENDIX A






ANALYSIS OF MEAN CONTAMINANT RESIDUES



   IN AQUATIC INVERTEBRATES OF THE



            NEW YORK BIGHT

-------
                 TABLE A-l




CONTAMINANT RESIDUES IN NEW YORK BIGHT BIOTA
SPECIES
Clan, Butter (Saxidonus gigantea)
Clan, Hard [Quahog] (Mercenana mercenana)


Clan, Surf (Spisula solidissina)





























Crab (Cancer sp.)
Crab, Blue ( Call i nee tes sapidus)
Crab, Dungeness (Cancer nagister)
Crab, King (Paralithodes camschatica)
Crab, Rock (Cancer irroratus)









REF
5
5
11
4
7
7
7
7
7
7
7
7
7
7
7
7
12
12
12
11
3
3
3
3
3
3
3
4
4
9
9
13
13
13
2
11
5
5
5
5
5
7
7
7
7
7
7
7
CADMIUM
mean n
0.178
0.567

0.800












0.130
0.130
0.150








0.100
0.130


0.040
0.050
0.060


0.175
0.242
0.140
0.210
0.550







6
6

1












6
11
11








1
1


1
1
1


2
9
1
1
1







MERCURY PCBs
mean n mean
0.036 6
0.052 6
0.230

0.020
0.010
0.010
0.020
0.040
0.020
0.040
0.040
0.020
0.030
0.070
0.070



0.360
0.004
0.033
0.030
0.025
0.049
0.030
0.070







0.002
0.320
0.375 2
0.128 9
0.330 1
0.240 1
0.140 1
0.030
0.070
0.060
0.040
0.040
0.060
0.040
n


36

1
1
1
1
1
1
1
1
1
1
1
1



6
1
3
2
2
1
6
2







4
14





1
1
1
1
1
1
1
DOTs
mean




0.018
0.008
0.015
0.026
0.055
0.069
0.090
0.015
0.025
0.018
0.053
0.036













0.012
0.004



0.000






0.094
0.051
0.051
0.047
0.039
0.073
0.039
n




1
1
1
1
1
1
1
1
1
1
1
1













1
1



4






1
1
1
1
1
1
1
                       -1-

-------
TABLE A-l  (Continued)
SPECIES REF
7
7
2
6
6
6
6
6
10
10
3
3
3
3
3
3
3
3
3
3
3
3
4
4
4
13
13
13
Lobster, American (Homarus amencanus) 7
7
7
7
7
7
2
2
11
11
8
8
8
8
8
8
8
8
8
8
CADMIUM
mean n



0.120
0.130
0.080
0.130
0.010














0.640


0.080
0.100
0.080










0.014
0.024
0.018
0.015
0.017
0.015
0.023
0.024
0.011
0.017



1
1
1
1
1














1


1
1
1










1
1
1
1
1
1
1
1
1
1
MERCURY
mean





0.100
0.190
















0.190
0.030













0.190
0.120
0.180
0.130
0.080
0.080
0.170
0.310
0.130
0.140
n





1
1
















1
1













1
1
1
1
1
1
1
1
1
1
PCBs
mean
0.020
0.000
0.043
0.203
0.229
0.385
0.287
0.546
0.410
0.500
0.340
0.003
1.813
0.793
0.050
0.418
0.041
0.030
0.460
0.003
0.019
0.047






0.040
G.Q40
O.G40
0.040
0.001
0.020
0.095
0.150
0.130
0.150
0.070
0.120
0.200
0.230
0.120
0.160
0.130
0.070
0.094
0.083
n
1
1
12
1
1
1
1
1
1
1
1
1
1
1
2
4
1
1
1
1
1
1






1
1
A
1
1
1
1
6
3
1
1








1
1
DOTs
mean
0.017
0.009
0.025





0.041
0.047


















0.050
0.100
0.060
0.035
0.002
0.044
0.024
0.050












n
1
1
12





1
1


















1
1
1
1
1
1
6
3












             -2-

-------
TABLE A-l  (Continued)
SPECIES REF
8
8
8
8
8
8
8
8
8
8
3
3
3
4
4
4
4
9
9
13
13
13
13
13
13
13
13
13
13
.obster, Spiny [Alt.] (Panulirus argus) 5
lussel, Blue (Mytilus edulis) 7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
7
CAOMIIM
wan n
0.018
0.017
0.018
0.021
0.023
0.016
0.017
0.012
0.012
0.020



0.140
1.080




0.160
0.140
0.070
0.070
0.120
0.090
0.080
0.150
0.080
0.070
0.058


















1
1
1
1
1
1
1
1
1
1



1
1




1
1
1
1
1
1
1
1
1
1
10


















MERCURY PCBs
mean n mean n
0.200 1 0.100
0.130 1 0.079
0.340 1 0.220
0.160 1 0.095
0.150 1 0.260
0.260 1 0.210
0.230 1 0.140
0.260 1 0.150
0.150 1 0.062
0.370 1 0.190
0.040
0.151
0.145


0.034 1
0.320 1
0.030











0.061 10
0.050
0.200
0.200
0.090
0.200
0.400
0.030
0.200
0.100
0.070
0.800
0.300
0.400
0.400
0.400
0.100
0.200
0.200
1
1
1
1
1
1
1
1
1
1
1
10
10




1












1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
DDTs
mean n


















0.007











0.232
0.240
0.510
0.150
0.000
0.070
0.170
0.260
0.290
0.270
0.600
0.230
0.340
0.370
0.310
0.211
0.190
0.110


















1











1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
           -3-

-------
TABLE A-l  (Continued)
SPECIES









Oyster, American (Crassostrea uirginica)

Oyster, Giant [Pac.] (Crassostrea 91935)
Polychaete worn
Polychaete worn (Ceriantheopsis aoencanus)
Polychaete worn (Glycera dibranchiata)
Polychaete worn (Haplosoloplos robustus)
Polychaete worn (Lunfannereis fragilis)



Polychaete worn (Lunbnnereis tenuis)
Polychaete worn (Nepthys bucera)

Polychaete worn (Nepthys incisa)






Polychaete worn (Nepthys picta)
Polychaete worn (Nereis virens)
Polychaete worm (Ninoe nigripes)

Polychaete worn (Ophioglycera gigantea)
Polychaete worn (Pherusa affinis)






Polvchaete worm (Travisia carnea)
Polychaete worn(Neothys sp.+Pherusa affinis)




REF
7
11
3
3
3
3
3
3
3
11
4
5
7
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
10
1
1
1
1
1
CAOrlllH
nean n










2.900
2.812


0.260
0.070



0.130

0.290
0.100
0.290
0.180




0.130
0.100
0.710
0.060


0.060
0.190
0.060
0.250
0.180
0.380
0.300
0.070















1
10


1
1



1

1
1
1
1




1
1
1
1


1
1
1
1
1
1
1
1





MERCURY
nean n











0.038



0.142



0.077

0.073
0.052
0.073





0.058
0.052
0.021
0.660




0.032
0.086

0.154
0.070

















10



1



1

1
1
1





1
1
1
1




1
1

1
1






PCBs
nean n
0.320
0.000
0.130
0.210
0.050
0.210
0.140
0.450
0.447
0.340


0.200
0.300


0.201
0.710
0.188

0.188




0.830
0.760
1.970
1.020




1.230
1.270








0.099
0.214
0.134
0.216
0.185
1
36
1
1
2
1
2
1
3
36


1
1


1
1
1

1




1
1
1
1




1
1








1
1
1
1
1
DDTs
nean
0.246











0.032
0.041


0.029
0.083
0.007

0.007




0.093
0.024
0.074
0.112




0.143
0.039








0.018
0.019
0.005
0.009
0.009
n
1











1
1


1
1
1

1




1
1
1
1




1
1








1
1
1
1
1

-------
TABLE A-l  (Continued)
SPECIES REF
1
1
1
1
Quahogs, Ocean (Artica islandica) 12
12
12
12
4
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
13
CADMIUM
mean




0.540
n




8
MERCURY PCBs DDTs
•ean n mean n mean n
0.254 1 0.016 1
0.263 1 0.013 1
0.712 1 0.016 1
0.145 1

0.420 IS
0.420
0.390
0.540
0.250
0.1(0
0.470
li.220
0.230
0.170
0.320
0.210
0.220
0.160
0.210
0.300
0.220
0.240
0.160
0.200
0.090
0.040
0.380
0.240
0.400
0.230
0.140
0.580
0.080
0.060
0.470
0.820
0.340
0.350
0.490
0.370
0.460
0.450
0.480
0.280
0.610
0.290
0.800
9
9
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1










































             -5-

-------
                                     TABLE  A-l   (Continued)
SPECIES











Scallop, Sea (Palctopecten nagillanicus)













Shrimp, Alaskan (Pandalopsis dispar)
Shrimp, Grass (Palaenonetes pugio)
Squid, Longfinned [Alt.] (Loligo pealu)



Squid, Shortfinned (Illex illecebrosus)

REF
13
13
13
13
13
13
13
13
13
13
13
7
7
7
7
2
13
13
13
13
13
13
13
13
13
5
7
5
S
5
5
5
5
CADMIUM
aean n
0.500
0.260
0.710
0.620
0.620
0.450
0.220
0.180
0.440
0.330
0.310





0.320
0.360
0.070
0.180
0.130
0.090
0.210
0.190
0.080
0.085

0.145
0.155
0.158
0.168
0.150
0.179
1
1
1
1
1
1
1
1
1
1
1





1
1
1
1
1
1
1
1
1
10

2
3
13
5
2
4
MERCURY PCBs DDTs
mean n mean n nean n











0.030 1 0.060 1
0.020 1 0.050 1
0.010 1 0.004 1
0.020 1 0.050 1
0.001 5 0.000 5









0.040 10
0.190 1 0.094 1
0.000 2
0.030 3
0.048 13
0.023 5
0.120 2
0.135 4
Average Concentration                               0.337          0.099          0.193         0.073

-------
             REFERENCES
 1 Boehm, 1982
 2 Boehm and Hurtzer. 1982
 3 COE and EPA. 1980
 4 COE and EPA, undated a
 5 Hall et al., 1978
 6 Lee and Jones. 1977
 7 MacLeod et al.,  1981
 8 O'Brien and Gere, 1979
 9 O'Connor et al.,  1982
10 Raltech Scientific Services, 1981
11 Spagnoli and Skinner, 1977
12 Wenzloff et al.,  1979
13 Zdanowicz, 1984

-------