United States
Environmental Protection
Agency
Environmental Research
Laboratory
Athens GA 30613
Research and Development
EPA-600/S3-83-053 Nov. 1983
oEPA Project Summary
Fates and Biological Effects of
Polycyclic Aromatic
Hydrocarbons in Aquatic
Systems
John P. Giesy, Steven M. Bartell, Peter F. Landrum, Gordon J. Leversee, and
John W. Bowling
The hypothesis that fates of polycyclic
aromatic hydrocarbons (PAH) in eco-
systems can be predicted by mech-
anistic simulation models based on
easily measured properties of the com-
pounds in this homologous series was
tested. The research involved: 1) de-
velopment of a mechanistic, predictive
simulation model based on kinetic and
thermodynamic considerations; 2) de-
velopment of analytical and quality
assurance protocols for the extraction
and quantification of PAH associated
with biological and geological matrices;
3) determination of the vectors of and
rate constants for uptake depuration
and biotransformation of PAH by aqua-
tic organisms and sediments; and 4)
comparison of the results of simula-
tion and laboratory scale studies to a
large scale ecosystem study. Included
were studies of the effect of PAH con-
centration, temperature, and, other
exogenous factors on rate constants
and efforts to determine whether rate
constants were first order.
Laboratory studies indicated that
anthracene and benzo(a)pyrene are
rapidly biotransformed by fish and dip-
teran larvae but not by periphyton com-
munities. Biotransformation had a sig-
nificant effect on the steady state con-
centrations of parent compound and
biotransformation products. These re-
sults demonstrated that predictions of
steady state concentrations based on
14C-labeled parent compound and
the octanol-water partitioning coeffi-
cient of the parent compou nd wou Id be
in error. Thus, the octanol-water par-
titioning coefficient would not be a
good predictor of the behavior of PAH
in aquatic organisms. Uptake and de-
puration rate constants were first order
for fish but not dipteran larvae. Induc-
tion of biotransformation and changes
in biotransformation rate over time
means that predictions of long-term
disposition in the ecosystem from short
term pharmacokinetic studies, using
radio-labeled compounds, will be mis-
leading for compounds that are bio-
transformed.
Anthracene (approximately 12 jog/l)
was acutely toxic to bluegill sunfish
dosed in outdoor channel microcosms.
This mortality was not observed in
laboratory studies and was shown to
be caused by a photo-mediated toxic
mechanism. Therefore, laboratory tox-
icity studies on PAH must be con-
ducted under the same lighting con-
ditions if the results of these studies
are to be realistic representations of
field conditions.
The modeled processes that most
influence PAH transport included loss-
es to volatilization, photolytic degra-
dation, sorption to suspended particu-
late matter and sediments and net
uptake by biota. The biota in the model
included phytoplankton, periphyton,
rooted macrophytes, bacteria, zoo-
plankton, two functionally defined
benthic invertebrate components and
two functionally defined categories of
-------
fish. Model simulations were com-
pared to results of experiments con-
ducted in artificial streams. A 0.06 u.
molar solution of anthracene in ethanol
was continuously added into the head-
waters for 15 days. The model accu-
rately predicted the dissolved anthra-
cene concentration through time and
space. Uptake by periphyton was over-
estimated by the model; however, the
rate of depuration of anthracene by
periphyton was reasonably simulated.
Photolytic degradation appeared to be
the most important pathway of flux
with the channels, both experimentally
and in the simulations.
This Project Summary was developed
by EPA's Environmental Research Lab-
oratory, Athens, GA, to announce key
findings of the research project that is
fully documented in a separate report
of the same title (see Project Report
ordering information at back).
Introduction
Polycyclic(polynuclear) aromatic hydro-
carbons (PAH or PNA) are a homologous
series of compounds composed of two or
more condensed benzene rings with oc-
casional incorporations of cyclopentene
rings, such as in the fluorenes, or"hetero-
atoms" (N, 0 or S). PAH are derived from
natural and man-made sources and are
widely distributed in the environment
PAH occur as natural products from plants
and microbes and from natural pyrolytic
processes such as forest fires and volcanic
activities. Man-made sources include in-
dustrial processes but the major source is
combustion of fuels. Most of the PAH in
the environment are due to human activi-
ties and freshwater and nearshore marine
environments are enriched with PAH.
More than 230,000 metric tons of PAH
enter the oceans and surface waters each
year.
The toxic, mutagenic and carcinogenic
properties of PAH have been much studied
Many PAH are carcinogens or procarcin-
ogens in animals and maa Part of the
hazard assessment process is to deter-
mine the exposure of animals and man.
This required an understanding of the fate
and transport processes in an ecosystem.
Because of their ability to bioconcentrate
lipophylic compounds such as PAH, aqua-
tic organisms are an important vector to
man. Also, because of their flowing, sol-
vating character and organic sediments,
surface waters serve as both a transport
mechanism and a sink for PAH.
Because of the vast differences be-
tween simple laboratory tests and com-
plex field situations, scientists have formed
conceptual and operational bridges be-
tween the two systems. These include
mechanistic simulation models and micro-
cosms, which are operational models or
simplifications of environments. In this
study, both of these types of models were
employed to investigate the fate and trans-
port of PAH in aquatic systems.
Summary
This program of study tested the overall
hypothesis that a dynamic, mechanistic
simulation model of the fate of PAH in an
aquatic system can be constructed, based
on characteristics of the compound of
interest and the environment to which it is
released. The study tested the ability to
predict the fate of anthracene introduced
into a channel microcosm using a simu-
lation model, which had been param-
eterized independently. In addition, the
simulation model was used to guide the
laboratory investigations. By using a syn-
thesis of microcosm, modeling and labor-
atory studies, the authors were able to
guide their research to answer some of
the most pertinent questions related to an
understanding of the dynamics of anthra-
cene in aquatic systems.
The fate of aromatics model (FOAM)
predicts the concentrations of PAH in
water, sediment phytoplankton, periphy-
ton, macrophytes, zooplankton, two ben-
thic invertebrate components, bacteria,
suspended paniculate matter and herbiv-
orous and carnivorous fish. The overall
simulation model is organized as a main
program (MAIN) (Fig. 1), which calls sub-
routines that calculate solar radiation
(SOLAR), movement of water (HYDRO),
solubility of PAH, and resuspension and
settling of suspended components and
graphs the absolute and relative concen-
trations of PAH in each compartment
(Table 1). Other submodels calculate pho-
tolysis (PHOTO), volatilization (VOL) and
sorption to sediments (SORP). The model
is a mass balance, generalized reach model
(Fig. 2). The biological components are
simulated as a generalized production
model with the PAH inputs partitioned into
the biotic components (Fig. 3). All of the
simulated processes in FOAM are cor-
rected for temperature effects. FOAM,
which is programmed in FORTRAN, re-
quires a small number of input variables
(relative to many dynamic, mechanistic
simulation models) and is easy to use.
Modular Program Structure
Inputs
Fates of Aromatics Model (FOAM)
Figure 1. Schematic representation of modular subroutine structure of FOAM.
-------
Each subroutine calculates part of the
overall derivative for change in biomass or
PAH mass in each state variable. FOAM
simulates the dynamics of biological pro-
duction and PAH dynamics through space
and time. Each state variable value is
printed in matrix form where row elements
are values at a specific time for each reach.
Column elements are temporal values in a
particular reach. A complete listing of
Table 1. Output from FOAM. Each Variable is Simulated Hourly
Parameter
Irradiance
Allochthanous organic matter
External PAH loading
Periphyton biomass
Macrophyte biomass
Benthic invertebrate biomass
Clam Biomass
Bacterial biomass
Fish biomass
Suspended particulate organics
Dissolved organics
Bottom sediments
Suspended inorganic particulates
Settled detritus
PAH in periphyton
PAH in macrophytes
PAH in benthic insects
PAH in bacteria
PAH in fish
predaceous
herbivorous
Suspended organic particulates
PAH in dissolved organic matter
PAH in bottom sediments
PAH in suspended inorganic particulates
PAH in settled detritus
PAH dissolved in water
PAH which has been transferred
Units
langleys-h'1
g • dry weight • m "2
gPAH-m*
g- m"2
g- m'2
g- m'2
g- m'2
g- m'2
g- m'2
g- m'2
g- m'2
g- m'2
g- m'2
g- m'2
fi mol PAH- g ' ', dry weight
H mol PAH- g'1, dry weight
p. mol PAH- g ' ', dry weight
p. mol PAH- g ' ', dry weight
[I mol PAH- g'1,dry weight
fj. mol PAH- g'1,dry weight
/A mol PAH- g~',dry weight
H mol PAH- g ' ', dry weight
/i mol PAH- g ' 1, dry weight
p. mol PAH- g'1,dry weight
fi mol PAH- g ' ', dry weight
H mol PAH- T1
nmolPAH-r'
General Reach Model
Segmented
Stream
REACH,
Figure 2. Conceptualization of generalized stream reach transport model.
FOAM is available from the authors. It
should be noted that model development
is a dynamic process and FOAM has been
undergoing continuous development since
it was initially programmed
FOAM was parameterized to simulate
anthracene dynamics in the channel micro-
cosms facility located at the Department of
Energy's (DOE) Savannah River Plant
(SRP). The channel microcosms facility is
a pass-through system consisting of six
separate cinder-block-lined channels each
91.5 m long, 0.61 m wide, and 0.31 m
deep. Located at the upper end of each
channel is a pool 3.1 m long, 1.5 m wide
and 0.9 m deep. At the lower end of the
facility is a single large pool 10.2 m long,
3.1m wide and 1.0 m deep. For these
studies, the channels and headpools were
lined with a 0.05 cm thick black polyvinyl
chloride film and the bottom covered with
0.05 m of washed quartz sand.
Water was pumped from a deep well
and treated so that the inorganic water
quality was similar to that of surface
waters in the upper coastal plain. Water
flows were monitored by V-notch wiers on
each headpool. Flow rates of 75.7 l/min
were maintained by input valves. This
resulted in a current velocity of 1.0 x 10 "2
m/sec and a retention time of 2.5 h. The
water depth in the channels was main-
tained at 20 cm. The channels were natu-
rally colonized with algae, bacteria, fungi
and aquatic invertebrates. Clams and fish
were held in the channels in cages for
the validation study.
Laboratory studies demonstrated that
benzo(a) pyrene (BaP) and anthracene were
rapidly biotransformed by bluegill sunfish
and chironomid larvae but not unionid
clams or periphyton assemblages. Bio-
transformation rates were dependent on
temperature and time of exposure. The
rates of biotransformation also influenced
the bioconcentration factors predicted
from short-term pharmacokinetic studies.
Temperature and food ration also influ-
enced the rates of uptake and depuration.
Uptake and depuration were adequately
described by first-order relationships.
Depuration was generally multiphasic with
some biotransformation products bound
to tissues, such that they were very slowly
depurated.
Recovery of anthracene from sediments
was inversely proportional to the time of
exposure. Internal standards, added at the
time of extraction, did not allow accurate
determination of extraction efficiencies.
Drying of sediments resulted in reduced
recovery from sediments A benzene-aceto-
nitrile (1:2; V/V) extract resulted in the
best recovery of anthracene from sedi-
-------
Simulation of PAH Flux Through Streams
f "--~J Benthic
\ J Invertebrates
REACH^
REACH,
REACH*
Figure 3. Schematic representation of FOAM model.
merits. Capillary gas-liquid chromatogra-
phy coupled to mass spectrometry (GC-
MS) showed that anthracene could be
extracted from sediments and separated
from naturally occurring organic com-
pounds but anthraquinone (a transforma-
tion product of anthracene) could not,
Dissolved organic carbon (humic sub-
stances) interfered with the extraction of
anthracene and BaP from water by macro-
reticular resins. Humic acids also reduced
the availability of anthracene, benzo(a)py-
rene and dimethylbenzanthracene to D.
magna but increased the availability of 3-
methylcholanthrene and dibenzanthracena
The presence of suspended particulates
also reduced the availability of PAH to D.
magna.
FOAM accurately simulated and pre-
dicted the flux along physical-chemical
pathways (Table 2). The model, however,
was less accurate in predicting the ac-
cumulation and biotransformation of PAH
by aquatic biota. Future versions of pre-
dictive simulation models should be kinet-
ically based, as FOAM is, if accurate in-
formation on cycling is to be predicted.
This type of model, however, requires a
great amount of information. Because of
the scale of the input parameters, models
such as FOAM will be useful in describing
overall processes and fluxes along dif-
ferent pathways but will not be very useful
in simulating the concentrations of PAH
and PAH transformation products in in-
dividual biotic components. Also, some of
the physical processes, which are impor-
tant in determining the fates of PAH in
natural systems, are discontinuous func-
tions or catastrophic events, such as
storms, which do not lend themselves to
mathematical simulations. Therefore, mod-
els of the type developed here will always
be limited to relatively gross predictions
Accurate prediction of concentrations in
individual types of organisms may not be
attainable.
The channel microcosms were useful
for testing the reliability of FOAM to sim-
ulate the behavior of PAH in a complex
ecosystem. The greatest utility of the
channel microcosms, however, was in
allowing studies to be conducted under
more natural conditions than allowed by
laboratory systems. From these studies
we learned that exposure to sunlight
caused acute mortality due to anthracene.
Thus, laboratory studies of anthracene
Table 2. Predicted and Observed Concentrations of Dissolved Anthracene in Channels
Microcosms
Time
Dawn
observed
predicted
Reach 1
(Head)
0.066
0.056
mol anthracene "'
Reach 3
(Middle)
0.065
0.056
Reach 5
(Tail)
0.063
0.055
Noon
observed
predicted
0.067
0.056
0.045
0.045
0.028
0.036
toxicity to aquatic organisms do not ac-
curately predict toxicity observed under
more natural conditions. Although micro-
cosms may not be used as a screening tool
for individual compounds, they are useful
in studies of classes of compounds and to
verify laboratory and simulation conclu-
sions. Because of their scale, however,
microcosms will never by useful in valida-
tion of long-term or global simulation
models. The test of the benchmark hy-
pothesis was only partly supported by our
studies.
Future simulations will need to consider
more complex measures of chemical be-
havior. The proposed use of octanol/water
partitioning coefficients shows promise
for predicting uptake by organisms but not
biotransformation. Thus, the benchmark
approach will probably not give adequate
predictions of the dynamics of organic
compounds in complex environments.
A coordinated synthesis of information
on ecosystem structure and function is
required to assess environmental impacts
of any technology that must ultimately be
interfaced with the environment While
this research was on PAH compounds
(anthracene and benzo(a)pyrene specifi-
cally) it supports the hypothesis that valid
generalizations about the behavior of gen-
eral classes of organic compounds in com-
plex environments can be made from
knowledge about the compounds obtained
in relatively simple laboratory studies and
the environments to which these com-
pounds are released. Knowledge of this
type can be transferred to other organic
compounds and has broad applicability.
Recommendations
Future studies of the fates and effects of
trace contaminants should be conducted
in laboratory, field, and simulation modes
concurrently. All three modes of investi-
gation added to the overall understanding
of the dynamics of PAH in aquatic eco-
systems.
Simulation models should be developed
to predict general behavior of PAH in
aquatic systems. That is, to determine
when the greatest mass of PAH accumu-
lates and where the most sensitive eco-
system components are. Where physical
and chemical processes dominate, simu-
lation models will be able to accurately
predict the overall dynamics of PAH, but
rare events of large magnitude will reduce
the accuracy of predictions on a short-
term basis. Optimization of time step
duration relative to system level variability
needs to be further investigated to in-
crease the accuracy of predictive models.
-------
The current state of predictive simu-
lation models will not allow the accurate
prediction of concentrations of single PAH
in individual species but can be useful in
simulating overall processes in a gross
manner. Future simulation should be
kinetic in nature but the relationships used
to predict rate constants need to be based
on structure-activity relationships.
The channel microcosms used in this
study were sufficient to test some simu-
lation processes, and the authors recom-
mend the use of this type of a system in
future studies. Microcosms of this type,
however, will not be useful as screening
tools. The utility of the realism of such
systems was demonstrated by the photo-
mediated toxicity of anthracene effect ob-
served in the microcosm, which was not
observed under laboratory conditions.
This type of system needs to be used to
validate processes that are indicated by
laboratory and/or simulations because
they allow testing in a more complex and
more natural system than that of the
laboratory, while not releasing trace con-
taminants to the biosphere
J. P. Giesy is with Michigan State University, East Lansing, Ml 48824; S. M.
Bartell is with Oak Ridge National Laboratory, Oak Ridge, TN; P. F. Landrum is
with the National Oceanographic and Atmospheric Administration, Ann Arbor,
Ml 48104; G. J. Leversee is with Keene State College, Keene. NH03431, andJ.
W. Bowling is with the University of Georgia, Savannah River Ecology
Laboratory, Aiken, SC 29801.
H. W. Holm is the EPA Project Officer (see below).
The complete report, entitled "Fates and Biological Effects of Polycyclic Aromatic
Hydrocarbons in Aquatic Systems," (Order No. PB 83-250 191; Cost: $20.50,
subject to change} will be available only from:
National Technical Information Service
5285 Port Royal Road
Springfield, VA 22161
Telephone: 703-487-4650
The EPA Project Officer can be contacted at:
Environmental Research Laboratory
U.S. Environmental Protection Agency
College Station Road
Athens, GA 30613
•&U. S. GOVERNMENT PRINTING OFFICE: 1983/759-102/0795
-------
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
BULK RAT
U.S. POSTA
PAID
Cincinnati, C
Permit No. G
Official Business
Penalty for Private Use $300
------- |