United States
Environmental Protection
Agency
Environmental Research
Laboratory
Athens GA 30613
Research and Development
EPA-600/S3-83-007 Apr. 1983
Project Summary
Verification of a Toxic Organic
Substance Transport and
Bioaccumulation Model
Jerald L Schnoor, Narasinga Rao, Kathryn J. Cartwright, Richard M. Noll, and
Carlos E. Ruiz-Calzada
A field verification of the Toxic Or-
ganic Substance Transport and Bioac-
cumulation Model (TOXIC) was con-
ducted using the insecticide dieldrin
and the herbicides alachlor and atrazine
as the test compounds. The test sites
were two Iowa reservoirs. The verifica-
tion procedure included both steady-
state analyses and quasi-dynamic simu-
lations using time-variable flows and
pollutant loadings along with model
coefficients derived from laboratory
and literature data.
Laboratory measurements were ap-
plied in simulations using alachlor,
atrazine and dieldrin, and model predic-
tions were well within an order-of-
magnitude of field observations. For the
herbicide, alachlor, for example, labora-
tory protocol measurements were used
directly in model simulations with ex-
cellent agreement between model
predictions and measured concentra-
tions.
The TOXIC model, therefore, was
considered to be field verified. More-
over, the successful field verification
supports the validity of EPA's Exposure
Analysis Modeling System (EXAMS),
which handles pollutant transport and
transformation kinetics in an almost
identical manner.
This Project Summary was developed
by EPA's Environmental Research Labo-
ratory, 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 order-
ing information at back).
Introduction
A number of mathematical models have
been developed in recent years to assess
the fate and transport of agricultural
chemicals in the aquatic environment.
Such models contain kinetic formulations
that describe the chemical, physical and
biological reactions that a pesticide can
undergo once it reaches a water body.
Chemical hydrolysis, volatilization, photol-
ysis, and biological degradation comprise
the reactions considered in these models.
The relative importance of the various
pathways of pesticide fate (hydrolysis,
volatilization, photolysis, biological degra-
dation, oxidation, and sorption) may be
assessed in the laboratory in conjunction
with field evaluations.
These models generally have not com-
bined fate and transport modeling with
biological impacts such as bioconcentra-
tion and, heretofore, have not been
extensively verified by comparing model
predictions with site-specific field meas-
urements. In this research, the Toxic
Organic Substances Transport and Bio-
accumulation Model (TOXIC) was devel-
oped and verified in an application on the
Coralville and Rathbun Reservoirs in
Iowa.
The reader should realize that the
kinetics handled by TOXIC are very similar
to those in EPA's Exposure Analysis
Modeling System (EXAMS). Therefore,
many of the conclusions drawn from this
study will be somewhat applicable to
EXAMS as well.
-------
Model Description
A schematic of the kinetics and trans-
port processes that comprise TOXIC,
which is an extension of a model devel-
oped by Stanford Research Institute (SRI),
is shown in Figure 1. The solid lines in
Figure 1 are in accordance with the
original SRI model formulation. Modifica-
tions in both the kinetics and transport
algorithms made in the development of
TOXIC and intended to add realism to the
model are represented by dashed lines. A
two-compartment or "pond" representa-
tion of the test reservoir sites was initially
assumed because there existed little in-
reservoir data with which to calibrate a
multi-compartment model. Figure 2 de-
picts the physical configurations of the
two-compartment pond model and a more
complex eight-compartment lake model
also utilized in the analysis. Coralville
Reservoir dimensions were used in the
pond configuration for model calibration,
and simulations were later performed
using the lake configuration as well.
In addition to chemical and biological
reaction pathways, uptake and depuration
(excretion and metabolism) in fish were
included in the model. The bioaccumula-
tion model formulation is depicted in
Figure 3. Biouptake was assumed to be
proportional to the product of the fish
biomass and the dissolved pesticide con-
centration. Pesticide was assumed to be
removed by the fish as water passed the
gill membrane. Biouptake from sediment
and/or food (prey) was also included in
this portion of the model. For the applica-
tion portion of this study, it was assumed
that the pesticide residue was metabo-
lized within the fish, but in some cases it
might be necessary to recycle the depu-
rated pesticide as a separate dissolved
input, as shown in Figure 3.
Field Validation Tests with
Dieldrin in Coralville Reservoir
Agricultural use of pesticides is wide-
spread in Iowa, particularly the grass and
broadleaf herbicides and row crop soil
insecticides. One insecticide widely used
from 1960 to 1975 to control the corn
rootworm and cutworm was the chlori-
nated hydrocarbon, aldrin, which metab-
olizes to a persistent epoxide, dieldrin.
Both a two-compartment pond and multi-
compartment lake version of TOXIC were
applied to the Coralville Reservoir to test
the ability of the model to predict the fate
and effects of dieldrin in the ecosystem.
Although the historical field data avail-
able reflected individual storm events.
SKI Model
Modifications
,
Wetfa/l
Dryfall
Absorption
Volatilization '
1 r\irt*t**
f- 'v-—»
U^^
6 *
t
' / \
i_)
Microlayer
Formation?
Photolysis
Photolysis
Dissolved
Pesticide
/sis 1
Biolysis
i
1
Adsorption
Desorntinn
Paniculate
Pesticide
Biouptake j
i
Hydrolysis j
Biolysis
Sed
T
Diffusion
Biota
Death and
Excretion
Biouptake
Scour
Dissolved
Hydrolysis
Biolysis
(Anaerobic)
Paniculate
Hydrolysis
Biolysis
(Anaerobic)
Figure 1. Chemical kinetics and transport processes modeled in TOXIC.
the goal here was to represent annual
average concentrations and mass flows.
Therefore, constant annual average in-
flow and outflow rates were assumed,
together with an average annual volume
for the reservoir. Coralville Reservoir does
not thermally stratify to any great extent,
so the failure to include a hypolimnion
compartment in the pond configuration of
TOXIC was not viewed as a serious
problem.
The first test exercise used the two-
compartment configuration of TOXIC.
Because the objective of this simulation
was to analyze annual average concen-
trations, transport properties (i.e., flow,
volume, and sedimentation rate) were
averaged over the period of simulation.
-------
The model simulation results versus the
field data are shown in Figure 4. Dieldrin
concentrations in water, shown in Figure
4, represent monthly grab sample field
measurements. The peak concentrations
during the runoff events recorded in the
field data were not matched by the model
results, since average inflow and a
smooth loading rate function were as-
sumed.
Figure 4 also shows the model results
and field data for dieldrin residues in
sediment and the edible tissue of bottom
feeding fish in Coralville Reservoir. A
sensitivity analysis was performed to
assess the importance of the partition
coefficient. The results indicated that a
several-fold increase or decrease of the
partition coefficient from 0.5 to 2.0 to
0.10 changed the total concentration by
approximately minus and plus 50%, re-
spectively. Results also indicated that the
bioconcentration of hydrophobic pesti-
cides such as dieldrin in Coralville Reser-
voir fish was directly proportional to the
oil or lip'td content of the catch. Although
the data were sparse, it appeared that
such a simple bioconcentration model
has validity.
The model predictions estimated that
40% of the dieldrin inflow to Coralville
Reservoir was lost through sedimentation
and 50% was released through the dam
gates of this relatively short detention-
time reservoir. Uptake by fish accounted
for the remaining 10% of the dieldrin
input due to the extremely large fish
biomass (1000 Ib/acre). The partitioning
of dieldrin in the water column was 64%
in the fish, 24% dissolved in the water,
a nd less than 12% adsorbed to suspended
solids. Mean residue levels in the edible
tissue of bottom feeding fish eventually
declined below the FDA guideline of 300
Aig/kg. Fish and sediment concentrations
were found to be essentially in equilib-
rium with the mean dissolved dieldrin
concentrations. Under low flow condi-
tions, the sediment became a net source
for pesticide in the reservoir through
desorption and pore water diffusion.
Bottom feeding fish bioaccumulated
dieldrin in proportion to the oil content
(determined by petroleum ether extrac-
tion) of the fish sample. Therefore, pollu-
tant averages or composites for very oily
fish tended to be higher than model
predictions. Model projections indicated
that by 1986, dieldrin residues in bottom
feeding fish flesh should steadily decline
to an average of less than 100 /^g/kg.
To investigate the sensitivity of the
model to compartmentization, an eight-
— -|
0.3m —»
5000m K-
2 -Compartment Pond
1—
t
E
CM
10
[• — 7800m — •
/
2
/ t 1
5
/
1
/J
8-Compartment Li
p — 7800m —
/
3
/t A
6
ft
lAe
Outflow
o
\ \
\ \ Water Compartment
Sediment Compartment
Figure 2. Physical configurations of the TOXIC model.
Input
Input
Food Items
Rapid
Sorptive
Equilibrium
Sediment
Fish
Worms
Metabolites
Sedimentation
y Hydrolysis
Biolysis
Photolysis
Volatilization
Oxidation
Figure 3. Bioaccumulation model schematic.
-------
L
I
I!
0.06
0.05
0.04
0.03
0.02
0.01
0.0
12
10
a
6
4
2
0
1500
1250
1000
750
500
250
Coralville Outflow
68 * €9 "70^ 71 ^72^ 73 ' 74 ' 75 r 76 T 77 T 78 ^ 79
Sediment Compartment
Z Field data, X±1s
71 Model Results
' 65 ^70 ^77 ' 72 ^73 ' 74 ' 75 T 75 ' 77 ' 78 * 79
Bottom Feeding Fish
Bioconcentration Factor
BCF= 70,000
FDA Action Level
Figure 4.
68 J 69 ' 70 ' 71 ' 72 ' 73 ' 74 r 75 ' 76 ' 77 ' 78 ' 79 '
/ear
/Ve/
-------
220
200-
180 -
160
140 -
Aquatic Exposure Model
Game Fish Food Items
/f, = 0.0135/day
= 0.0415/day
I
1}
ti
r
Gill+Sediment+Predation+Benthos
successful field verification of TOXIC also
supports the use of EPA's Exposure
Analysis Modeling System (EXAMS),
which incorporates similar modeling pro-
cedures. As an example of the model's
utility, the Iowa Conservation Commis-
sion based its decision to lift its ban on
commercial fishing in Coralville Reservoir
in 1979 largely on a series of TOXIC
simulations.
1968 70 72 74 76 78 80 82 84 86
Time (years)
Figure S. Simulated dieldrin residues in Coralville Reservoir game fish.
tory-derived rate constants directly in a
field verification effort. The relatively
soluble, biodegradable chemicals were
chosen to complement the dieldri n study.
The resulting modeling efforts on Lake
Rathbun for the highly soluble herbicides
alachlor and atrazine were much different
than for hydrophobic pollutants such as
dieldrin, which tend to sorb on sediment.
In order to assess herbicide dynamics, a
three-step effort was undertaken includ-
ing biotransformation tests in natural
waters, microcosm testing, and field data
collection.
In summary, the laboratory biotrans-
formation and microcosm studies were
helpful in predicting the dynamics of
alachlor and atrazine in Lake Rathbun.
Model predictions were within a factor of
five for atrazine, a factor .of two for
dieldrin, and a factor of one for alachlor.
For most purposes, these would be accep-
table simulations and the model should
be considered as field verified. In these
simulations, rate constants were not
adjusted to fit the field data, that is, there
was no model tuning. Laboratory and
literature rate constants were used di-
rectly in a comparison with field obser-
vations.
Conclusions
The field verification exercise demon-
strated the acceptable use of laboratory
process rate measurements in a modeling
study. Laboratory measurements were
used in simulations of alachlor, atrazine,
and dieldrin, and model predictions were
well within an order-of-magnitude (maxi-
mum of a factor of five) of field data. The
-------
Jerald L Schnoor, Narasinga Rao, Kathryn J. Cartwright, Richard M. Noll, and
Carlos E. Ruiz-Calzada are with the University of Iowa, Iowa City, I A 52242.
T. O. Barnwell is the EPA Project Officer (see below).
The complete report, entitled "Verification of a Toxic Organic Substance Trans-
port andBioaccumulation Model." (Order No. PB 83-170 563; Cost: $17.50,
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: 1983/659-095/1923
-------
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
Postage and
Fees Paid
Environmental
Protection
Agency
EPA 335
Official Business
Penalty for Private Use $300
Pb 0000329 , _ „
U S ENVIR PROTECT ION AGENCf
REGION 5 LIBRARY
230 S DEARBORN STRtET
CHICAGO IL 606U4
------- |