oEPA
United States
Environmental Protection
Agency
Environmental Research
Laboratory
Athens GA 30613
EPA-600 3-82-045
August 1982
Research and Development
Mathematical Model,
SERATRA, for
Sediment-Contaminant
Transport in Rivers and
Its Application to
Pesticide Transport in
Four Mile and Wolf
Creeks in Iowa
-------
EPA 600/3-82-045
August 1982
MATHEMATICAL MODEL, SERATRA, FOR SEDIMENT-
CONTAMINANT TRANSPORT IN RIVERS AND ITS
APPLICATION TO PESTICIDE TRANSPORT IN
FOUR MILE AND WOLF CREEKS IN IOWA
by
Y. Onishi and S.E. Wise
Battelle
Pacific Northwest Laboratories
Richland, Washington 99352
Contract No. 68-03-2613
Project Officer
Robert B. Ambrose
Technology Development and Applications Branch
Environmental Research Laboratory
Athens, Georgia 30613
ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
ATHENS, GEORGIA 30613
-------
NOTICE
Mention of trade names or commercial products does not
constitute endorsement or recommendation for use.
-------
FOREWORD
As environmental controls become more costly to implement and the
penalties of judgment errors become more severe, environmental quality
management requires more efficient management tools based on greater knowledge
of the environmental phenomena to be managed. As part of this Laboratory's
research on the occurrence, movement, transformation, impact, and control of
environmental contaminants, the Technology Development and Applications
Branch develops management and engineering tools to help pollution control
officials achieve water quality goals through watershed management.
Many toxic contaminants are persistent and undergo complex interactions
in the environment. As an aid to environmental decision-makers, the Chemical
Migration and Risk Assessment Methodology was developed to predict the occur-
rence and duration of pesticide concentrations in surface waters receiving
runoff from agricultural lands and to assess potential acute and chronic
damages to aquatic biota.
David W. Duttweiler
Director
Environmental Research Laboratory
Athens, Georgia
m
-------
ABSTRACT
The sediment-contaminant transport model SERATRA was used as an inte-
gral part of the Chemical Migration and Risk Assessment (CMRA) Methodology,
which simulates migration and fate of a contaminant over the land surface
and in receiving streams, to assess potential short- and long-term impact
on aquatic biota. SERATRA, an unsteady, two-dimensional (longitudinal and
vertical) finite element model, consists of three submodels coupled to in-
clude the effects of sediment-contaminant interactions—a sediment transport
submodel, a dissolved contaminant transport submodel, and a particulate
contaminant (contaminants adsorbed by sediment) transport submodel. The
sediment transport submodel simulates transport, deposition and scouring of
three sediment size fractions of cohesive and noncohesive sediments. The
dissolved contaminant transport submodel includes mechanisms of contaminant
adsorption/desorption and contaminant degradation resulting from hydrolysis,
oxidation, photolysis, volatilization, biological activities and radionuclide
decay to predict migration of dissolved contaminant. The particulate contam-
inant submodel simulates transport, deposition and scouring of contaminants
associated with three size fractions of sediments. SERATRA also predicts
changes of bed conditions for sediment and particulate contaminants.
SERATRA was applied to Four Mile and Wolf Creeks in Iowa to simulate
transport of sediment and the pesticide alachlor in these streams. Study
results demonstrate that SERATRA is a useful tool for evaluating migration
and fate of various hazardous dubstances and toxic chemicals in rivers by
including contaminant interaction and contaminant degradation.
Companion reports to this document are Methodology for Overland and
Instream Migration and Risk Assessment of Pesticides, User's Manual for the
Instream Sediment-Contaminant Transport Model SERATRA. User's Manual for
EXPLORE-I: A River Basin Water Quality Model (Hydrodynamic Module Only),
and Frequency Analysis of Pesticide Concentrations for Risk Assessment
(FRANCO Model).
This report was submitted in partial fulfillment of Contract No. 68-03-
2613 by Battelle Pacific Northwest Laboratories under the sponsorship of the
U.S. Environmental Protection Agency. This report covers the period April
1978 to January 1980, and work was completed as of January 1980.
iv
-------
CONTENTS
ABSTRACT 1v
FIGURES vii
LIST OF SYMBOLS ix
1. INTRODUCTION 1
2. CONCLUSIONS 3
3. FORMULATION OF MODEL, SERATRA 4
3.1 SEDIMENT TRANSPORT SUBMODEL 4
3.2 DISSOLVED CONTAMINANT TRANSPORT SUBMODEL. . . 8
Chemical Degradation Due to Hydrolysis ... 11
Hydrolysis 11
Chemical Degradtion Due to Oxidation ... 11
Chemical Degradation Due to Photolysis ... 12
Chemical-Degradation Due to Volatilization . . 13
Biodegradation 14
3.3 PARTICULATE CONTAMINANT TRANSPORT MODEL. . . 14
3.4 FINITE ELEMENT TECHNIQUE 16
Galerkin Weighted Residual Method .... 17
Finite-Element Equations 18
Time-Dependent Solution 20
3.5 LISTS OF INPUT DATA AND SIMULATION OUTPUT . . 21
-------
4. MODEL APPLICATION RESULTS 23
CASE 1. CALIBRATION RUN 27
CASE 2. THREE-YEAR SIMULATION RUN 30
REFERENCES 54
-------
LISTS OF FIGURES
1 Four Mile and Wolf Creeks in Iowa 24
2 Time variations of simulated concentration of sand, silt,
clay and total sediment at Four Mile Creek River
kilometer 2.5 28
3 Time variations of simulated particulate alachlor
concentrations associated with sand, silt and clay,
together with mean particulate alachlor concentrations
at Four Mile Creek River kilometer 2.5 29
4 Time variations of simulated dissolved, particulate
and total alachlor concentrations at Four Mile Creek
River kilometer 2.5 ......... 31
5 Vertical distributions of simulated sediment concentrations
at Four Mile Creek River kilometer 2.5 at 12:00 p.m.,
July 10, 1971 32
6 Vertical distributions of simulated particulate alachlor
adsorbed by sediments at Four Mile Creek River kilometer 2.5
at 12:00 p.m., July 10, 1971 33
7 Vertical distributions of simulated dissolved, particulate
and total alachlor concentrations at Four Mile Creek River
kilometer 2.5 at 12:00 p.m., July 10, 1971 .... 34
8 Time variation of predicted flow rate at Wolf Creek River
kilometer 5 during the three-year simulation period . . 35
9 Time variation of predicted total sediment concentration
at Wolf Creek River kilometer 5 during the three-year
simulation period 36
10 Time variation of predicted particulate alachlor concentration
per unit weight of sediment at Wolf Creek River kilometer 5
during the three-year simulation period 37
11 Time variation of predicted particulate alachlor concentration
per unit volume of water at Wolf Creek River kilometer 5
during the three-year simulation period 38
vii
-------
12 Time variation of predicted dissolved alachlor concentration
at Wolf Creek River kilometer 5 during the three-year
simulation period 39
13 Time variation of predicted total alachlor concentration
at Wolf Creek River kilometer 5 during the three-year
simulation period 40
14 Time variation of simulated flow rate at Wolf Creek River
kilometer 5 during May 5, 1972 to August 15, 1972 ... 41
15 Time variation of simulated total sediment concentration
at Wolf Creek River kilometer 5 during May 5, 1972
to August 15, 1972 42
16 Time variation of simulated particulate alachlor concentration
per unit weight of sediment at Wolf Creek River kilometer 5
during May 5, 1972 to August 15, 1972 43
17 Time variation of simulated total alachlor concentration
at Wolf Creek River Kilometer 5 during May 5, 1972
to August 15, 1972 44
18 Time variation of simulated flow rate at Wolf Creek River
kilometer 5 during May 26, 1973 to July 15, 1973 ... 45
19 Time variation of simulated total sediment concentration
at Wolf Creek River kilometer 5 during May 26, 1973
to July 15, 1973 46
20 Time variation of simulated particulate alachlor
concentration at Wolf Creek River kilometer 5 during
May 26, 1973 to July 15, 1973 47
21 Time variations of simulated particulate, dissolved
and total alachlor concentrations at Wolf Creek River
kilometer 5 duing May 30, 1973 to June 9, 1973 ... 48
22 Longitudinal distribution of total sediment concentration
at 6 a.m., July 4, 1973 49
23 Longitudinal distributions of simulated dissolved, particulate
and total alachlor at 6 a.m., June 4, 1973 .... 50
24 Variations of simulated particulate alachlor in the top bed
layer accumulated during the three-year simulation period . 51
vm
-------
SYMBOLS
A = the river surface area
a = coefficient in Equation (34)
B = river width
[B] = biomass per unit volume
b = coefficient in Equation (34)
C = average sediment concentration above water depth Z
Cj = concentration of sediment of jth size fraction
C-jj = sediment concentration of horizontal inflow for jth size fraction
e 7
D = —~ in Equation (53)
AZ^
d = flow depth
d0 = molecular diameters of oxygen
DJ = diameter of jth sediment
ds = molecular diameter of contaminant
«
Xei = subdomain of interest in Equation (32)
fsj = fraction of contaminant sorbed by jth sediment
fw = fraction of contaminant left in solution
= particulate contaminant concentration per unit weight of sediment on
the river bed for jth sediment size fraction
= particulate contaminant concentration per unit weight of sediment in
jth sediment size fraction
IX
-------
G-JJ = parti oil ate contaminant concentration per unit weight of sediment in
jth sediment size fraction in horizontal inflow
Gw = dissolved contaminant concentration
Gwl- = dissolved contaminant concentration in horizontal inflow
h = water depth
i = element's number counted from the river bottom
I0^ = incident light intensity of wave length A
1^ = light intensity of wave length A at water depth z
J = conversion constant in Equation (19)
KI = light attenuation coefficient for water
K£ = light attenuation coefficient clue to suspended sediment in water
KA = acid hydrolysis rate
KAB = oxidation rate of free radical oxygen of [RO-]
Kg = base hydrolysis rate
the second order rate constant for biodegradation
transfer rate of contaminant with jth non-moving bed sediment
the first order degradation rate due to hydrolysis
Kr,2 = the first order degradation rate due to oxidation
Kc3 = the first order degradation rate due to photolysis
KQ4 = the first order degradation rate due to volatilization
= the first order degradation rate due to biological activities
ra*e °^ adsorption (or desorption) between dissolved contaminant and
sediment (suspended and bed load sediments) of jth size fraction
kj = transfer rate of contaminant with jth sediment in motion
Kj = an empirical constant in Equation(9)
KN = neutrol hydrolysis rate
-------
KO = oxygen reaeration rate
Kox = oxidation rate of free radical oxygen of [R02*l
kw = [H+]-[OH-] = 10-14
RO* = free radical oxygen
R02* = free radical oxygen
£ = longitudinal distance
MJ = credibility coefficient for sediment of jth size fraction
MJ = weight of jth sediment
n = the number of subdomains
N = number of sediment size fractions considered. In this study, N = 3
(sand, silt and clay)
Qi = horizontal inflow discharge equal to U-jBAz
Q = horizontal outflow discharge equal to U0BAz
Qj = sediment transport capacity
Qja = actual sediment transport rate
Qv = vertical flow discharge equal to (WA)
R = a domain of interest in Equation (31)
= sediment deposition rate per unit area for jth sediment size fraction
= sediment erosion rate per unit area for jth sediment size fraction
t = time
U0 = horizontal outflow velocity
U-j = horizontal inflow velocity
u = mean velocity
u* = shear velocity
v = coefficient in Equation (28)
v-j = value v at i-th node
XI
-------
V = (v-j + 1/2 v-j+i) shown in Equation (54)
V = (1/2 v-j + v-j+i) shown in Equation (54)
Vw = volume of water
W = vertical flow velocity
W-j = weighting function in Equation (30)
Wj = weighting function in Equation (29)
Wsj = fall velocity of sediment particle of jth size fraction
z = vertical direction
[P] = symmetric tridiagonal matrix in Equation (50)
[P] = symmetric tridiagonal matrix in Equation (55)
[P] = symmetric tridiagonal matrix in Equation (57)
{R} = load vector in Equation (50)
{R} = load vector in Equation (55)
{R} = load vector in Equation (57)
[S] = unsymmetric tridiagonal matrix in Equation (50)
[S] = unsymmetric tridiagonal matrix in Equation (55)
[S] = unsymmetric tridiagonal matrix in Equation (57)
a = coefficient in Equation (28)
3 = coefficient in Equation (28)
Y = coefficient, i.e., probability that particle settling to the bed is
deposited
YJ = specific weight of jth sediment
Az = increment of vertical distance
ez = vertical diffusion coefficient
ez = coefficient in Equation (28)
e, = molar extinction coefficient of light with the wave length A
XI1
-------
A = decay rate of radioactive material
T[J = bed shear stress
TcDj = critical shear stress for sediment deposition for jth sediment size
fraction
TcRj = critical shear stress for sediment erosion for jth sediment size
fraction
= quantum yield
= variable in Equation (28)
= approximation of 4> shown in Equation (30)
4> = time derivative of
L[] = residual error in Equation (28)
xm
-------
SECTION 1
INTRODUCTION
The environmental impact of hazardous substances and toxic chemicals
is an increasingly important issue (EPA 1978, Onishi and Wise 1978,
Tofflamire et al. 1979). Although considerable effort is being made to
minimize the release of contaminants to receiving water bodies, the
decision makers of both government and industries must have a sound basis
for impact assessment.
Mathematical models supported by coordinated data collection programs
can be useful tools in assessing migration and ultimate fate of
contaminants in surface waters. In order to obtain accurate predictions
of contaminant transport, mathematical models must include the major
transport and fate mechanisms. These mechanisms include:
1. convection and diffusion/disperson of contaminants
2. chemical and biological degradation due to hydrolysis, oxidation,
photolysis, volatilization and biological activities
3. interaction between sediment and contaminants, such as contaminant
adsorption by sediment; contaminant desorption from sediment to
water; transport of pariculate contaminant (those adsorbed by
sediment); deposition of particulate contaminants to the river bed;
and resuspension of particulate contaminant from the bed
4. contaminant contributions from point and nonpoint sources into the
river system.
Contaminants (e.g., pesticides, radionuclides, heavy metals, and many
toxic chemicals) with high distribution coefficients or contaminants in
surface water with high suspended sediment concentration are, to a great
extent, adsorbed by river sediments. Thus, otherwise dilute contaminants
are concentrated. Contaminated sediments may be deposited on the river
bed, becoming a long-term source of pollution through desorption and
resuspension. Until recently, sediment-contaminant interaction was not
included in most models because of the complex nature of sediment
transport and contaminant adsorption/desorption mechanism (Norton et al.
1974, Leerdertse 1970, Onishi 1979).
In order to obtain more realistic predictions, the model SERATRA
(Onishi et al. 1976; Onishi 1977; Onishi et al. 1979a) was modified to
include all important mechanisms described above. The modified version of
-------
SERATRA was used as an integrated part of the Chemical Migration and Risk
Assessment (CMRA) Methodology for the overland and instream migration and
risk assessment of pesticides (Onishi et al. 1979b,c). This report
describes the formulation of SERATRA and its application results to Four
Mile and Wolf Creeks in Iowa.
-------
SECTION 2
CONCLUSIONS
The unsteady, two-dimensional (vertical longitudinal and vertical),
sediment-contaminant transport model, SERATRA, was adopted to be an
integrated part of the CMRA Methodology. SERATRA was then applied to Four
Mile and Wolf Creeks in Iowa to examine its applicability and limitations
in simulating pesticide migration in these streams. The main conclusions
are summarized as follows:
1. Application of SERATRA to Four Mile and Wolf Creeks in Iowa confirms
that SERATRA is capable of predicting migration and fate of
pesticides in rivers.
2. SERATRA is general enough to be applicable to nontidal rivers,
streams and narrow impoundments.
3. Because SERATRA continuously and simultaneously simulates sediment
and contaminant transport and fate by including all important
mechanisms, such as sediment-contaminant interaction and contaminant
degradation, it is a useful tool for evaluating both short- and
long-term migration and fate of pesticides and other contaminants
(e.g., radionuclides, heavy metals and other toxic chemicals).
-------
SECTION 3
FORMULATION OF MODEL, SERATRA
The SERATRA code utilizes the finite element computation method with
the Galerkin weighted residual technique. It consists of three submodels
coupled to include the effects of sediment-contaminant interaction. The
submodels are: 1) a sediment transport submodel, 2) a dissolved
contaminant transport submodel, and 3) a particulate contaminant
(contaminants adsorbed by sediment) transport submodel. SERATRA not only
calculates distributions of sediment and contaminant concentrations in
water, but also predicts river bed conditions, including bed elevation
change, distribution of each sediment size fraction within the bed, and
distribution of particulate contaminant concentration in the river bed.
The detailed model formulation is discussed below:
3.1 SEDIMENT TRANSPORT SUBMODEL
Since the movements and adsorption capacities of sediments vary
significantly with sediment sizes, the sediment transport submodel solves
the migration of sediment (transport, deposition and scouring) for three
size fractions of cohesive and non-cohesive sediments.
The model includes the mechanisms of:
1. convection and diffusion/dispersion of sediments
2. fall velocity and cohesiveness
3. deposition on the river bed
4. resuspension from the river bed (bed erosion and armoring)
5. sediment contributions from tributaries and point and nonpoint
sources into the system.
Sediment minerology and water quality effects are implicitly included
through the above mentioned mechanisms 2, 3, and 4.
Mass conservation of sediment passing through the control volume
leads to the following expression for the transport of sediments:
-------
rate of horizontal vertical
accumulation convection convection
a 3C1 1
' fe (£z JT*V + F J = 1> ^ •••> N
vertical sediment erosion
diffusion or deposition
where(a)
B = river width
Cj = concentration of sediment of jth size fraction
C-jj = sediment concentration of horizontal inflow for jth size fraction
h = water depth
i = longitudinal distance
N = number of sediment size fractions considered. In this study, N = 3
(i.e., sand, silt and clay)
SQJ = sediment deposition rate per unit area for jth sediment size
fraction
SRJ = sediment erosion rate per unit area for jth sediment size fraction
t = time
U-j = horizontal inflow velocity
U0 = horizontal outflow velocity
W = vertical flow velocity
WSj = fall velocity of sediment particule of jth size fraction
z = vertical direction
ez = vertical diffusion coefficient
The model neglects longitudinal diffusion and assumes lateral sediment
concentrations to be uniform. However, the model does handle vertical
variations of longitudinal velocity to cause some longitudinal dispersal
of sediment.
(a) The symbols defined above remain the same throughout the report
5
-------
Boundary conditions at the water surface (z = h) and river bed (z = 0)
are:
3C1
W - Wsj Cj - £z 9T = ° at z = h
3C.
- = •
where
Y = coefficient, i.e., probability that particle settling to the bed is
deposited.
In this study, Y is assumed to be unity, that is, for the same flow
condition all suspended sediment settling to the river bed will stay on
the river bed without going into the flow again.
Sediment erosion and deposition rates, SRJ and SQJ, are also
evaluated separately for each sediment size fraction because erosion and
deposition characteristics are significantly different for cohesive and
noncohesive sediments.
Erosion and deposition of noncohesive sediments are affected by the
amount of sediment the flow is capable of carrying. For example, if the
amount of sand being transported is less than the flow can carry for given
hydrodynamic conditions, the river will scour sediment from the stream bed
to increase the sediment transport rate. This occurs until the actual sedi-
ment transport rate becomes equal to the carrying capacity of the flow or
until the available bed sediments are all scoured, whichever occurs first.
Conversely, the river deposits sand if its actual sediment transport rate
is above the flow's capacity to carry sediment. Sediment transport capac-
ity of flow, Qj in this model, was calculated by either the Toffaleti or
the Colby formulas (Vanoni 1975). The computer program of the Colby method
developed by Mahmood and Ponce (Mahmood and Ponce 1975) was used. The
sediment transport capacity of flow, Qj, was then compared with the
actual amount of sand, Qja, being transported in a river water. Hence:
QT - QT,
QT, - Q
S =
5
A
where
A = the river bed surface area.
The availability of bed sediments to be resuspended was also examined to
determine the actual amount of sediment erosion.
-------
For sediment erosion and deposition rates of cohesive sediments (silt
and clay), the following Partheniades (1962) and Krone (1962) formulas,
respectively, were adoped in this study:
^ - vj - r (7)
where
MJ = erodibility coefficient for sediment of jth size fraction
Tb = bed shear stress
TcDj = critical shear stress for sediment deposition for jth sediment
size fraction
TcRj = critical shear stress for sediment erosion for jth sediment size
fraction
Values of MJ, ^cQj and ^cRj must be determined by field and/or
laboratory tests for a particular river regime. The model examines the
availability of cohesive sediments in the river bed to determine the
actual amount of sediment erosion.
When the fall velocity, Wsj, depends on sediment concentration and
no aggregation occurs, the fall velocity could be assumed (Krone 1962):
"sj • KI
where
KJ" = an empirical constant depending on the sediment type.
3.2 DISSOLVED CONTAMINANT TRANSPORT SUBMODEL
The dissolved contaminant transport submodel includes the mechanisms
of:
1. Convection and diffusion/dispersion of dissolved contaminants
(pesticides, radionuclides, and other toxic substances) within the
river
2. Adsorption (uptake) of dissolved contaminants by sediments (suspended
and bed sediments) or desorption from sediments into water
-------
3. Degradation of dissolved contaminants due to hydrolysis, oxidation,
volatilization, photolysis and biological activities
4. Radionuclide decay
5. Contributions of dissolved contaminants from point and nonpoint
sources into the system.
Effects of water quality (e.g., pH, water temperature, salinity, etc.) and
clay minerals are taken into account through changes in the distribution
coefficients for adsorption and desorption, K
-------
In addition to the previously defined symbols:
DJ = diameter of jth sediment
Ggj = participate contaminant concentration per unit weight of
sediment in jth sediment size fraction in river bed
TJ = specific weight of jth sediment
Gw = dissolved contaminant concentration
GW-J = dissolved contaminant concentration in horizontal inflow
GJ = part icu late contaminants concentration per unit weight of jth
sediment
KQ-J = the first order reaction rate of contaminant degradation due
to hydrolysis, oxidation, photolysis, volatilization and
biological activities
Kb-,- = transfer rate of contaminants with jth non-moving sediment in
bed
kdj> kd'j = rate °f adsorption and desorption between dissolved
contaminant and sediment (suspended and bed load sediments) of
jth size fraction, respectively
KJ, KJ = transfer rate of contaminants for adosprtion and
desorption, respectively with jth sediment in motion
A = decay rate of radioactive material.
POR = porosity of bed sediment
The distribution coefficient, KJJ and K^j, is defined by:
f/M' f
and
where
fsj = fraction of contaminant sorbed by jth sediment
fw = fraction of contaminant left in solution
Mj = weight of jth sediment
Vw = volume of water
w
-------
Hence Equation 10 may be rewritten as:
GJ • Kdj G« °r GJ • Kdj G«
The adsorption of contaminant by sediments or desorption from the sediments
is assumed to occur toward an equilibrium condition with the transfer rate,
KJ or Kj (with the unit of reciprocal of time), if the particulate con-
taminant concentration differs from its equilibrium values as expressed in
Equation 12. Longitudinal diffusion is considered to be negligible when
compared to convection. Longitudinal dispersal of contaminants due to ver-
tical variation of longitudinal velocity is, however simulated in this
study.
Boundary conditions at the water surface and river bed are:
sr
WGW . e = 0 at z = h (12)
e
z
= 0 at z = 0 (13)
One of the important mechanisms of transport and fate of contaminants
is degradation of contaminants in an aquatic environment. The contaminant
degradation includes both chemical and biological reactions. Major mecha-
nisms of chemical degradation are due to reactions of: 1) hydrolysis,
2) oxidation, and 3) photolysis (Smith et al. 1977). Due to lack of a
present knowledge on degradation and volatilization of particulate con-
taminants (Smith et al. 1977), these degradation mechanisms were considered
only for the dissolved contaminants. These degradation rates are included
in Equation 9 as the first order kinetic reaction rates, KC-J; i=l, 2, 3,
4 and 5. Actual formulations of chemical and biological degradation were
obtained from pesticide studies conducted by Smith et al. 1977, Zepp and
Cline 1977 and Falco et al. 1976. These formulations will be discussed
below:
Chemical Degradation Due to Hydrolysis:
The fundamental concept of chemical reactivity is based on the quest
to improve stability in the configuration of the outer shells. A
contaminant in solution will react with other species in solution and form
a complex if there is a large increase in stability. Hydrolysis reactions
are a specialized type of complex formation in which the [OH"] anion
acts as the ligand. They are quite sensitive to pH changes. The rate of
change of dissolved contaminant concentration due to hydrolysis is
expressed by the following equation (Smith et al 1977):
Hydrolysis
10
-------
K8-10pH-14 +
' KC1E«
where
pH = - log [H+]
KA> KB» KN = acid, base and neutral hydrolysis rates;
respectively
Rate coefficients, K/\, Kg and K^ can be determine laboratory tests
(Smith et al. 1977).
Chemical Degradation Due to Oxidation
Oxidation of contaminants by free radical processes may become
important under some environmental conditions. The rate of oxidation of
contaminant may be expressed by the second order reactions depending on
the concentration of free and radical oxygen and dissolved contaminant as
shown below (Smith et al. 1977):
jp
It is assumed that only a small concentration of dissolved
contaminants is oxidized, and that the second term in the right hand side
may be deleted (Smith et al. 1977). Hence:
dG
where
= oxidation rates of free radical oxygen of R0£ and
JRO-1, respectively
R02* = free radical exygen
RO* = free radical oxygen
n
-------
The rate constant, Kox can be obtained from laboratory tests outlined by
Smith et al. (1977).
Chemical Degradation Due to Photolysis
Pesticides can be photochemically transformed by adsorbing light,
expecially ultra violet light. The rate of contaminant concentration
change due to photolysis reactions may be expressed by (Zepp and Cline
1977, Smith et al. 1977, and Stanford Research Institute 1979):
dG,
dT
w 2.303
(17)
where
C = average sediment concentration above water depth Z
IQ\ = incident light intensity of wave length \
1^ = light intensity of wave length X at water depth Z
J = conversion constant
K-J = light attenuation coefficient for water
Kz = light attenuation coefficient due to suspended sediment in water
e^ = molar extinction coefficient of light with the wave length A
= quantam yield
Since each computational cell has a vertical finite element
thickness, the above equation was averaged over the element thickness for
each element in this study. Hence:
dGw
dT
2.303
- exp -
exp
K9C)Az
K2C)(n - i)AzJ
+ K2C) z
w
(18)
= KC3Gw
where
i = element's number counted from the river bottom
12
-------
n = total number of elements
Az = element thickness.
Various parameters and coefficients can be measured by conducting
laboratory tests and/or field measurements (Zepp and Cline 1977, Smith
et al. 1977, and Stanford Research Institute 1979).
Biodegradation:
A contaminant compound can be degraded by microbial activities in an
aquatic environment. In this study it is assumed that microbial
degradation can be expressed by the second order reaction (Falco et al.
1976, Smith et al. 1977) depending on concentrations of biomass and
contaminant in water, as shown below:
dG
- W- - KBl^Gw = KC5Gw <19)
[B] = biomass per unit volume
KBI = the second order rate constant for biodegradation
Volatilization:
The volatilization of a contaminant occurs at the air-water
interface. The change of contaminant concentration due to volatilization
may be expressed by the following first order reaction (Smith et al.
1977):
dG
- W- = KC3Gw
where
= volatilization rate of the contaminant
can be estimated by the following relationship:
(K } - (K } I
v C3' water body v o' water body \K /laboratory test condition
= (K } I— I
v o' water body yd I
where
d0, ds = molecular diameters of oxygen and contaminant, respectively
13
-------
(KC3) laboratory test conditions = volatilization rate through any
substances measure at a laboratory
(Ko) water = oxygen reaeration rate through water-air interfere
3.3 PARTICULATE CONTAMINANT TRANSPORT MODEL
The transport model of contaminants attached to sediment includes the
mechanisms of:
1. Convection and diffusion/dispersion of particulate contaminants
2. Adsorption (uptake) of dissolved contaminants by sediments or
desorption from sediments into water
3. Radionuclide decay
4. Deposition of particulate contaminants to the river bed or
resuspension from the river bed
5. Contributions of particulate contaminants from point and nonpoint
sources into the system.
As in the transport of sediments and dissolved contaminants, the
conservation of contaminants adsorbed by each sand, silt and clay sediment
may be expressed as:
j B£) + (Uo6j B - Ui6.. B) + k {(W - Ws.)Gj
rate of horizontal vertical
accumulation convection convection
(21)
) - ASj B* -H K. (Kdj6w - 6.) B£ * K. (Kdj6w - G.) B£ + £ (GBJSR - G.S0
\ /
vertical radionuclide adsorption desorption contaminated sedimei
diffusion decay erosion and deposit
where
G-JJ = particulate concentration per unit volume of water associated
with the j'th sediment size fraction in horizontal inflow (no
summation will be applied).
Longitudinal diffusion was assumed to be negligible, as compared to
longitudinal convection. However, as noted before, the longitudinal
dispersal of particulate contaminant due to nonuniform vertical
distribution of longitudinal velocity is simulated in the model. It is
also assumed that chemical and biological degradation of particulate
14
-------
contaminants, except radionuclide decay, is not significant. However, if
it is necessary to include these mechanisms, the radionuclide decay term
can include them.
The boundary conditions at the water surface and bed are:
36.
G, (W - W .) - e —i = 0 at z = h and 0 (22)
J SJ Z oZ
The finite element technique with the Galerkin weighted residual
method was used to solve transport equations of sediments, dissolved
contaminants and particulate contaminants previously described.
3.4 FINITE ELEMENT TECHNIQUE
High speed digital computers have enabled engineers to employ various
numerical discretization techniques for approximating solutions to complex
mathematical equations. The finite element method is one such technique
and has recently gained popularity for solving both linear and nonlinear
partial differential equations.
Because of its increased solution accuracy and ready accommodation to
various boundary geometries (Desai and Abel 1972, Norton et al 1973,
Onishi and Wise 1978) this method was used for this study. To apply the
finite element method to a partial differential equation, an alternate
integral equation is developed. The finite element method employing a
Galerkin weighted residual is used to solve Equations 1, 9, and 21 with
the boundary conditions of Equations 2, 3, 12, 13, and 22. Since the
governing equations of sediment and contaminant transport have similar
forms, the finite element technique is described here for the following
convection- diffusion equation of the general form:
L * =!F + !I^ -fz-Kfih^-e (23)
The coefficients v, ez, a, 3 are defined to accommodate the specific
forms of the sediment and contaminant transport equations.
Galerkin Weighted Residual Method
The governing partial differential equation can be recast in an
integral form employing the weighted residual method. This integral is
formed by taking the product of L[$] with some arbitrary set of weighting
functions, Wj, which yields:
X = / L[]W,dz (24)
R J
where R = a domain of interest.
15
-------
If $ is approximated by some polynomial:
.
(25)
where W-j = approximating functions then the quantity L[4>] represents the
residual error. For the Galerkin weighted residual method, the
approximating function W-j is chosen to be the same as the weighting
function W-j. The integral term becomes an error distribution principle
by which the nodal values 4>j can be determined so that the residual
error over the domain R is orthogonal to the polynomical selected for
weight and interpolation functions.
By expanding the equation and integrating by parts, the reduced func-
tional results:
= 0 (26)
j = 1, 2 ... n + 1
Note that Equation 26 was derived by assuming no net flux across the
boundary. In cases where the boundary has positive or negative net flux
across it, this flux was incorporated as a source or sink term in the
governing equation.
The above integral can be partitioned so that:
* - Xel + Xe2 + Xe3 + . . . + Xen = £ xei
where
n = the number of subdomains. This may be expanded to give:
(27)
X =
[ ] dz
[ ] dz
[ ] dz
(28)
The contents in brackets in Equation 28 are the same as those in
Equation 26. Equation 28 will yield a set of n+1 ordinary differential
equations in terms of the 4>n+i.
Finite-Element Equations
The contributions from any typical subregion or finite element can be
developed by substituting a particular polynomial approximation for 4> into
Equation 26, For simplicity, a linear approximation was chosen: i.e.,
16
-------
(29)
where Az =
In vector notation, one has:
$(z,t) =
where weighting functions are given by:
• If
The individual terms in the functional are approximated by:
3t
= w,,w.
3z
3*
3z
3z 3z ' 3z
l_ l_
Az' A:
(30)
(31)
(32)
(33)
(34)
Substituting these quantities into Equation 26 for a typical subdomaih, a
set of algebraic equations is obtained.
The results for individual terms are given below.
1. Time Dependent Term:
(35)
"2 r
1 2
i
dt
2. Advective Term:
JR "j 9Z dZ '
Kei lxei
ei
17
-------
|v. + iv.
3 i 6 -
3. Diffusion Term:
Rei
4. Decay Term:
Rei
5 V1
H+l
Z1
5. Sink/source Term:
2 1
1 2
*i
*i+l
(36)
"l -l"
1 -1
1
•••§
1 -1
(37)
(38)
W.3cl
W.dz
D
Rei
i
1/2
1/2.
(39)
Where 3 is assumed independent of z. However, many sink/source terms in
Equations 1, 9, and 21 are not independent of z and these terms are
integrated according to their dependency of z. Summing up and gathering
terms, one can write the element contributions to the matrix equation as
finite element:
n
E
i
where
[P]
ej
1/3
1/6
3Az I1]
~T ,
1/6
1/3
pi+l
(40)
(41)
(42)
18
-------
} = {R} (45)
where
[P] matrix = a symmetric tridiagonal
[S] matrix = an unsymmetric tridiagonal
{R} = a load vector.
This. final system of equations is approximated by a Crank-Nicholson scheme
(Varga 1965):
i ~ /f 1 n+1 f iri\ i A
M " At" -4 H (H +WH(H *1RU (46)
Solving the for n+1 volue of , this expression can be rearranged to
obtain:
r in+i r in f l
[P] t*J = [S] l*| + (R| (47)
where the new coefficient matrices are:
[P] = [P] + (^|) [S] (48)
BJ = [P] - (^)fS] (49)
= At {R] = ^ ({R}n+1 + {R}") (50)
19
-------
This type of approximation is second order correct, unconditionally stable
and easily solved by most available tridiagonal solution schemes.
A computer program has been written in the FORTRAN preprocessor
language FLECS to implement the model (Onishi and Wise 1979c). However, a
standard FORTRAN IV version of SERATRA is also available.
3.5 LISTS OF INPUT DATA AND SIMULATION OUTPUT
SERATRA, consisting of the three submodels, is applicable to nontidal
rivers and impoundments. One of the advantages of SERATRA is that it can
be applied to water bodies over large longitudinal distances arid shallow
depths. Input data requirements for SERATRA are:
• Common Data Requirements for all the Submodels:
- Channel geometry
- Discharges and flow depth of the rivers during the simulation
period
- Discharges of tributaries, overland runoff and other point and
nonpoint sources
- Vertical dispersion coefficient
• Additional Requirements for Sediment Transport Submodel
- Sediment size fraction
- Sediment density and fall velocities for sand, silt, and clay
- Critical shear stresses for erosion and deposition of cohesive
sediment (silt and clay)
- Erodibility coefficient of cohesive sediment
Initial Conditions
- Sediment concentration for each sediment size fraction
- Bottom sediment size fraction
Boundary Conditions
- Sediment concentration at the upstream end of the study
reach
- Contributions of sediments from overland, tributaries and
other point and nonpoint sources.
20
-------
• Additional Requirements for Dissolved and Participate Contaminant
Transport Submodels:
- Distribution coefficients and transfer rates of contaminant with
sediment in each sediment size fraction (i.e., sand, silt, and
clay). If values of distribution coefficients are not
available, it is necessary to know clay mineral and organic
sediment content to estimate these values.
- Degradation and decay rates of contaminants
Initial Conditions
- Dissolved contaminant concentration
- Particulate contaminant concentration for each sediment
size fraction (i.e., those attached to sand, silt, and clay)
Boundary Conditions
- Dissolved and particulate contaminant concentrations for
each sediment size fraction at the upstream end of the
study reach
- Contributions of dissolved and particulate contaminant
concentrations from tributaries, overland, and other point
and nonpoint sources.
With the input data described above, SERATRA simulates the
following:
1. Sediment simulation for any given time
- longitudinal and vertical distributions of total sediment (sum
of suspended and bed load) concentration for each sediment size
fraction
- longitudinal and vertical distributions of sediment size
fractions in the river bed
- change in bed elevation (elevation changes due to sediment
deposition and/or scour)
2. Contaminant simulation for any given time
- longitudinal and vertical distributions of dissolved contaminant
concentration
- longitudinal and vertical distributions of contaminant
concentration adsorbed by sediment for each sediment size
fraction
21
-------
longitudinal and vertical distributions of contaminant
concentrations in the bottom sediment within the bed for each
sediment size fraction.
22
-------
SECTION 4
MODEL APPLICATION RESULTS
As an integrated part of the CMRA Methodology for the overland and
instream transport simulation and risk assessment of pesticides, SERATRA
was applied to Four Mile and Wolf Creeks in Iowa to simulate migration of
sediment and the pesticide, alachlor in these streams (Onishi et al.
1979b). The purpose of the CMRA Methodology application to Four Mile
Creek watershed was to evaluate the methodology and not to perform a
complete assessment of the pesticide. Alachlor is a preemergence
herbicide used to control most annual grasses and certain broadleaf
weeds. This pesticide is the most widely used herbicide in the
predominantly agricultural Four Mile Creek watershed (Baker et al. 1979).
The study area for the instream modeling was a 67.6 km reach between river
kilometer 19.3 in Four Mile Creek and the mouth of the Wolf Creek, as
shown in Figure 1. Four Mile Creek joins Wolf Creek at river kilometer
48.3 of Wolf Creek. Simulations of sediment and alachlor migration in
these streams were performed for the three-year period between June 1971
and May 1974.
The following assumptions were made to set up the ARM model: For
each of the three catchments, dates for plowing, planting and cultivating
were randomly selected within the periods of April 1 - May 1, May 5 -
June 1, and May 30 - July 1, respectively. Soybeans were assumed to be
the major crop planted in the study area, and alachlor was applied to the
soybeans the day after planting. The only constraint on the selection of
a pesticide application date was that it would not occur on the same day
as a storm event occurred. The application rates of alachlor for Gladbook
(344 ha) and Southern (444 ha) Catchments were selected to be 0.8 kg/ha,
while the application rate for Northern Catchment (660 ha) was assumed to
be 2.3 kg/ha. Detailed description of the ARM simulation was presented in
Onishi et al. (1979). Computed sediment and pesticide loads by the ARM
model were, in turn, supplied to SERATRA as input data. The computed
runoff was used as input to EXPLORE-I (Baca et al. 1973; Onishi 1979b).
Since SERATRA requires depth and velocity distributions in Four Mile and
Wolf Creeks as input data, we applied a hydrodynamic submodel of the
general water quality model EXPLORE-I to Four Mile and Wolf Creeks to
predict depth and velocity distributions for the simulation period.
Runoff from the three catchments within the Four Mile Creek watershed was
assumed to reach Four Mile Creek at river kilometer 9.7 (see Figure 1).
Very few measured data were available for the instream pesticide
modeling. Discharge measurements of Four Mile Creek conducted at Traer
(river kilometer 4.22) during the three year simulation period indicate
23
-------
o
—I C\J
on
LU _
LTv
CD
to
O
-------
the large variation of daily discharge, ranging from 9.65 m3/sec to
0.006 nH/sec. An average discharge at Traer during this period was
0.46 m3/sec. Because the purpose of the methodology application to Four
Mile and Wolf Creeks was to evaluate the methodology and not to perform an
assessment in this area, some simplification of flow data was performed.
The base flow of Four Mile Creek without runoff contributions was
assumed to be 2 m^/sec. Detailed test conditions for sediment and
pesticide transport modeling are shown in Table 1. Based on field data
obtained by Iowa State University (Baker et al. 1979), it was assumed that
the river sediment consists of 65% sand, 15% silt and 20% clay. For
eroded soil from the overland, Baker et al. (1979) show the variation in a
particle size distribution as a function of time and sediment
concentration. Based on these particle size distribution analyses
performed by the Iowa State University on selected runoff samples,
calculated total sediment loading from these three catchments were assumed
to consist of 0% sand, 40% silt and 60% clay. Particle sizes of sand,
silt and clay were assumed to be 0.350 mm, 0.016 mm, and 0.0014 mm,
respectively. No field or laboratory studies were performed to evaluate
critical shear stresses (TCRJ and TCQJ) and the credibility
coefficient (Mj), needed to calculate erosion and deposition of cohesive
sediments. These parameters were determined through a trial and error
calibration precedure. The bed shear stresses at each location for each
time step were calculated by SERATRA code internally with a known
hydraulic condition.
The distribution coefficient of alachlor associated with the bulk
sediment was estimated to be 50 ml/g in the receiving streams (Baker
et al. 1979). In this application, adsorption/desorption processes were
assumed to be completely reversible processes. In this study,
distribution coefficients for both adsorption and desorption, Kd and Kd
values of alachlor with sand, silt and clay were rather arbitrary assumed
to be 2, 20, and 100 ml/g. The total pesticide alachlor for each storm
event was then distributed to each sediment size fraction as follows: A
ratio of the selected Kd and Kd values of sand to those of silt and clay
is 1:10:50. Hence for a equal amount of sand, silt and clay, clay
contains five times more alachlor than silt, which in turn contains
10 times more alachlor than sand. Since the sediment loading from
overland was assumed to be 40% silt and 60% clay, total particulate
alachlor contributed from the three catchments was divided into 0%, 12%
and 88% each associated with eroded overland sand, silt and clay,
respectively at the stream edge. Because of the lack of available data,
the transfer rates, Kj and Kj, for alachlor to moving sediment and
stationary bed sediment to reach to an equilibrium condition assigned by
distribution coefficients were arbitrarily assigned to be 0.36 and
0.001 per hour, respectively. Based on the sediment size fraction ratio
and the Kd and Kd values for the sediment of each size fraction, it was
determined that 0%, 12% and 88% of particulate pesticide loading from
overland at Four Mile Creek, as computed by ARM, were associated with
sand, silt and clay, respectively. Althrough SERATRA simulates various
chemical and biological degradation of dissolved pesticides individually,
there were no available data to enable the simulation of degradation by
25
-------
Table 1. Test conditions for sediment and pesticide transport modeling,
Base Flow, m3/sec
Four Mile Creek 2
Wolf Creek 7
Manning Coefficient
Four Mile Creek 0.07
Wolf Creek 0.04
Sediment Sizes, mm
Sand 0.35
Silt 0.016
Clay 0.0014
Bed Sediment Size Fraction, %
Sand 65
Silt 15
Clay 20
Size Fraction of Eroded Sediment from Overland, %
Sand 0
Silt 40
Clay 60
Erodibility Coefficient, kg/m3
Four Mile Creek
Silt 1 x 10-4
Clay 1 x 10-4
Wolf Creek
Silt 1 x 10-4
Clay 1 x 10-4
Critical Shear Stress for Erosion, kg/m2
Four Mile Creek
Silt 1.5
Clay 1.7
Wolf Creek
Silt 0.55
Clay 0.6
Critical Shear Stress for Deposition, kg/m2
Four Mile Creek
Silt 0.7
Clay 0.6
Wolf Creek
Silt 0.4
Clay 0.3
Vertical Diffusion Coefficient, m2/sec 1 x 10~6
Distribution Coefficient, m3/kg
With Sand 0.002
Silt 0.020
Clay 0.10
Transfer Rate, 1/hr
Moving Sediment 0.36
Non-moving Sediment 0.001
Total Alachlor Degradation, 1/hr 0.003
26
-------
each process. Hence in this study, all pesticide degradation was lumped
into a first order reaction having a degradation rate of 0.003 per hour.
This rate was estimated from preliminary information supplied by Monsanto
Agricultural Product Co.
Simulation of sediment and alachlor transport was conducted for each
of the following substances: 1) sand, 2) silt, 3) clay, 4) dissolved
alachlor, 5) particulate alachlor associated with sand, 6) particulate
alachlor attached to silt, and 7) particulate alachlor adsorbed by clay.
The modeling procedure for SERATRA involved simulating sediment transport
of sand, silt and clay. The results were then used to simulate dissolved
and particulate alachlor by including the effects of alachlor-sediment
interaction. Finally, changes in stream bed conditions were recorded,
including: 1) stream bed elevation change, 2) vertical and longitudinal
distributions of ratios of sand, silt and clay within the bed, and
3) vertical and longitudinal distributions of alachlor within the stream
bed.
CASE 1. CALIBRATION RUN
Calibration of EXPLORE-I to obtain proper depth and velocity
distributions in Four Mile and Wolf Creeks was performed using the largest
predicted runoff event (using the overland model, ARM) during the
three-year similation period. This event resulted in the maximum
discharge of 18.1 m^/sec in Four Mile Creek. The duration of the
calibration run was 10 days starting from July 8 to July 17, 1971. After
numerous trial runs, the proper Manning coefficients in Four Mile and Wolf
Creeks were determined to be 0.07 and 0.04, respectively. These Manning
coefficient values were judged reasonable for these small streams.
After ensuring mass conservation, EXPLORE-I and ARM simulation
results were input to SERATRA for the model calibration. The only
parameters and coefficients that require adjustment are the vertical
dispersion coefficient and the three parameters, Mj, TcRj> and Tcn.i»
used to calculate deposition and erosion rates of cohesive sediments (see
Equations 7 and 8). Transport of noncohesive sediments (sand) does not
require the model calibration. The other parameters (e.g., pesticide
distribution coefficients and degradation rate, sediment sizes, etc) are
determined by theoretical and experimental analyses and field conditions
prior to the model simulation. Thus most of the calibration effort was
oriented toward reproducing sediment distribution patterns similar to the
actual or estimated longitudinal and vertical distributions of sediment
concentrations for the 67.6-km study reach. For model calibration, Four
Mile and Wolf Creeks were divided into fourteen 4.83-km segments. A
fifteen-minute time step was used. After trial and error, the values
shown in Table 1 were selected for Mj, TCRJ, TCDj: Tne, vertical
dispersion coefficient was selected to be 1.0 x lO"6 m2/sec.
Time variations of computed sediment concentration for each sediment
size fraction and the sum of the three size fractions at the mouth of Four
Mile Creek are shown in Figure 2. This figure indicates that the majority
of sediment is clay. Figure 3 shows time variations of computed
27
-------
in
CNJ
DC
UJ
UJ
o
£
cc
UJ
i — «
cc
«
UJ
cc
UJ
— 1
»— 1
ac
cc
u_
: 1 —
L '-•J
: „
: o
: LU
: to
~~ Q ^_
• ^ "^^ ^™"
o
,1
- I,
ii
\
: 1 1 1 1 I 1 1 1 1 t 1 1 . 1 . 1 1 1 1 1 1 1 1 1 1 1 1 t !
^
1—1
ro
O
S •> •
4-> ID
•i- CVJ
$-
ir\ * **
J£j T3 4->
CO O
I
o 12
S CO J-
C 0)
o >
ro
C^ 4_> rjj
•— • c
Q o •—
OJ
•M S-
(O 3
i— O
^ JU-
r— 1 CO ro
M- -M
O C
CU
trf E
^ C -t-
I-H o -a
+J CO
(O
il '(0
rO -»->
0s > O
•i- C
OO 1— ro
ED °°
—i O)
I * • » •
in **
-------
T | 1 p—1 p
Q j_ >-
5 - 3
CO CO O
CX.
O
I. O
in
CC
UJ
UJ
o
J5
cc
UJ
UJ
u
UJ
•—I
2:
cc
o
C£ C£ CC
O O O
1
1
I
1
in*
CVJ
oo
S
J
Q
CM
(M
Q
(M
Q
I\J(n(r)(O
QSOB
in
eu
CO I
c 3 s-
o o o>
•r- •,- +J
s-
o
o:
U S- (U
(O OJ i.
r- JC O
(O +J
- s:
> 3
•»— (O O
-»-> r— U_
S- O
C^ "O (O
Q) C
1|1_> ^J Q
« >—•-
r- -r- +J
3 CO n)
CO TJ C
C 0)
«4- (d O
O CO C
O
CO JC O
|Sb
(O
«O (O r—
> -i- (O
O O)
E CO 4J
•r- CO (O
1— «•—
CO
Q.-0t X C9W/9WD NOI1UU1N30NQ3
29
-------
particulate alachlor associated with each sediment size fraction, together
with mean particulate alachlor concentrations (weighted average of three
particulate alachlor concentrations associated with sand, silt and clay)
at the mouth. Because clay has the largest Kd value, the particulate
pesticide concentration associated with clay has the highest
concentration. Sand contains the lowest pesticide concentration.
Variations of computed dissolved, particulate, and total (sum of dissolved
and particulate pesticide concentrations) alachlor concentrations with
time at the Four Mile Creek mouth are shown in Figure 4. The figure
reveals that more than 75% of alachlor near the mouth of Four Mile Creek
was transported in the dissolved form and that up to 25% of alachlor was
carried by sediment. This implies that sediment transport becomes
important when a pesticide has a large distribution coefficient, Kd, or
when a receiving water body contains very high sediment concentrations.
Vertical distributions of sediment and alachlor concentrations at
Four Mile Creek River Kilometer 2.5 are shown in Figures 5, 6 and 7. In
these figures, elevations of water and bed surfaces are 272.56 m and
271.77 m, respectively. Figure 5 indicates that all suspended sediments
have higher concentrations near the river bed, but cohesive sediments show
relatively more uniform vertical distributions. This is due to the
smaller fall velocities of silt and clay as compared to the fall velocity
of sand. Figure 6 presents vertical distributions of particulate alachlor
concentrations per unit weight of sand, silt and clay. This figure
reveals that except those attached to sand, particulate alachlor
concentrations are almost vertically uniform. Vertical distributions of
average particulate (weighted average of particulate alachlor associated
with sand, silt and clay), dissolved and total alachlor concentrations are
shown in Figure 7. In this figure, dissolved alachlor concentration is
also shown to be almost vertically uniform. However, the particulate
alachlor concentration per unit volume of water indicates higher
concentration near the river bed due to higher sediment concentrations
near the bed. The total alachlor concentration (sum of dissolved and
particulate alachlor concentrations) reflects patterns of dissolved and
particulate alachlor distributions. Because of the lack of field data,
model calibration was not performed as rigorously as it should be.
CASE 2. THREE-YEAR SIMULATION RUN
Following model calibration, EXPLORE-I and SERATRA were used to
simulate pesticide migration for the three-year duration between June 1971
and May 1974. For this case, the entire study reach was divided into
seven equal-distance segments and a 30-minute time step was used to reduce
the required computational time. Some of the simulation results are shown
in Figures 8 through 24.
Figure 8 shows the time variation of flow rate calculated near the
mouth of Wolf Creek. The figure reveals that all 10 high flows during the
three-year simulation period occurred in summer and fall. Snow melt did
not produce significant runoff to receiving rates during this study
period. Time variation of total sediment concentrations near the Wolf
Creek mouth is shown in Figure 9, which clearly indicates a small number
30
-------
CO
>-
<
o
If)
03 CM
4->
O J-
+-> O)
re
(O
I— 1-
3 QJ
o >
•^- «r-
•M o;
(O ^
CL a)
a)
•> 1-
T3 O
O)
> O)
to 2!
to
•O 3
O
TD Lu
O)
•4-> •«->
rO rd
CD
•
en
o
•
CM
CO C
§0)
o
+J O
(O O
(O O
> f—
J=
-------
«a*
I"
O
o
o
CM
CM
r«-
CM
o
•
CM
CM CM
w 'NOIIVA313
oo
•
CM
CM
CM
r^
CM
CM
K
CM
'NOI1VA3T3
OO
i—I
CM
CT>
E
CO
Pi
CM
CM
•
CM
r»-
CM
CM
OO
•
i—I
CM
'NOI1VAJ13
to
^
in o
C r-H
o
•1- >,
4J r-
H3 3
S- rO
-l->
C •>
O) .
O 1=
C •
O 0.
o
o
4J O
C •••
CU Cvl
E •-!
(O S-
•—
rt)
u
O)
S-
CU
:>
T-
Qi
LO
O)
3
01
32
-------
-
-
ill 1 1 1 1
•* CM O OO
CM CM* CM i— *
r*- r1*- r** f^
CM CM CM CM
r*-
i
CD
^ o
5: "^
2
TICULATEALAC
ITH CLAY, mg/r
^S*
'N011VA313
CD
r—i
X
5-§
tn •<
o^
T5
O)
JD
S-
o
CO *>
•o •
S- OL
O
•— O
JC O
O ••
(O CM
d) (O
-(->
(O LO
3 CM
•r- S-
+-> CU
S- •»->
(O QJ
Q- E
-O i—
QJ -r-
3 O)
E O)
l|_ Q)
O i—
4-
CM
CM
Si
CM
CM
'NOI1VA313
CM
CO S
I 3
-t-> O
3 U_
JD
•r- +J
S-
T3 C i—I
CM
CM
CM
CM*
CM
CD
CM*
CM
'NOI1VA313
<
CO >>
O) ^"s
IO
O)
3
CD
33
-------
, •
. 1 i 1 i ! i 1
•
S- C£
CO
en
CO
13 S-
co o
i— CO t-H
O <— 1^
3 O
T3 O i—I
CO U.
CO 4-> r—
i— rO 3
3 "-3
E (S)
•r- C. *
V> O •
O CO O.
S-
c c o
O CO ••
•r- O CM
4-> C r-l
3 O
-O O +->
•i- CO
J- J-
-M O LT>
•r- JC CM
CO I.
•— i— QJ
CO CO -4->
$--(->•—
CO O •!-
> -M ^.
Si
3
'NOI1VA313
34
-------
in
i
«-4
N.
CO
»
m
BC
UJ
£
UJ
£
O
1 1
— CD
• C_3
O
r^
en
«-• s-
0)
s- .x
u £
S:&
DO
CO
CO
£i
«8
CO
CM §-g
r**. >r~
cr. -e-r,
00
cn
C9 CO CO
CM ^ -H
(M S •
^4 «-i 00
CSHOD rtOU Wy3UiS
-------
cn
IT)
CD
in
cc
UJ
UJ
o
cc
UJ
cc
UJ
UJ
cc
o
CJ
O
oo
£2
cn
_i
- O
o
i— TD
O O
3 -i-
S-
-M 3
C E
QJ ••-
O t/>
c
O S-
O CO
cu
C I
cu cu
II
CU -M
I/I
+J
O 01
-O 4-
CU 3
(J
0) S-
O)
I
I
o>
E •<-
O isi
-M S-
eo CU
•r- >
S- T-
(O Qi
CU
•r- J_
1— O
cn
cn
ncDc^com=3»onc\j—<
TWO X C1/3WD NOIlUyiN3GN03 1N3WIQ3S IBIOI
36
-------
5*4
c>-
\
m
*«t
ID
i
«—i
C">-
\
«—*
•s.
CO
to
cc
UJ
\—
UJ
o
_J
I— «
b£
cc
UJ
>•
aH
vv-
Jfc.
LU
' 1 1
LLJ
cc
LJ
'JL_
1
O
^
•
5
—~~=^n *
1 1 1 1 1 1 1 1 1
O
CO ^ §
-^ & 5 n
^ ,-H 4-> in
re)
S- 5-
•4->
O) O)
L_ 0 E
"S IS
^
_ S-
O S-
i—
Z ^ >
-1 ££
i- O
m CO tO CJ -r-
1 t i 1*"^ ' — ^~
- tz en 3 <4- 0)
^ -H 0^-0.
•i- O
- £ £3§
£. (0 4-> -r-
it Q. re 4->
h- *
- r; -o -i-> •—
X OJ C 3
0 -p oj e
0 E--
_ -r- -^ 10
-o -a
OJ S-
.^. S_ 10 n3
_ ^ Q. OJ
— ^ (j ^^
i — v M — -O
^ M- 0 1
O QJ
•»-> OJ
~ C JZ f-
o cn^n
S K ^"5^
-^ en -P
(O T-
- > C CD
3 C
OJ -r-
c~i E S- S_
_^ -^ OJ 3
0 (— D.-0
•
~ 0
I— 1
1 S £
^ S 3
•"• CD
€3
m -n
5-31 X C3N/3W3
3QI3IiS3d
37
-------
en
in
CO
in
cc
UJ
cc
LU
UJ
LU
cc
CJ
ID
'~3
CD
r* O -I—
CTl T- J_
•—I -M 3
rO T3
S-
OJ i-
O O)
O E
S- r—
O -r-
-5 o s-
-5 45 gj
rO -I-
o:
oo (0
CD
D-
•(-> C
o s_ o
•r- QJ -r-
O> tO eo
S- 3 i—
Q. 3
M- O -r-
O «/)
a)
O 3 ro
•i- r—
+J O >>
ro > I
•t- o>
S- -M E .E
3 -!->
E S-
"~> ^ 5.
^^ cn
OT
CO
CO
ID
Q
CO
CM
S
«-H
S
0
I
Q
38
-------
on
10
i
CO
to
cc.
LU
LU
O
CC.
LU
:*:
LU
LU
CC.
O
CO
CO
CTi
(.n
CQ
rrj O
C S_
O 0)
•r- Q.
re C.
S- O
fjj rO
O i—
o I
O T-
o $_
r— fO
^ QJ
O >>
rO I
r— Q}
tO CO
4-
01 +->
r— OJ
o s:
to +->
•r- CD
T3 C
•r—
d) 3
4-> TO
O
•r- LO
-o
OJ S-
J- O)
OJ
o ^
•r-
4-> S-
(O O)
•r- >
s- ••-
rO 0£
-------
CT)
in
CD
10
oc
LU
oc
UJ
UJ
£
o
o
CO
o
o
cr>
i—i
CO
o
•— T3
O O
(O Q.
c c
o o
-!-> -M
3
o; -r-
O tn
O S-
O (O
O)
$- >»
O I
•— a;
c— d)
O i-
T3
a
•r- LT)
XJ
i- O)
O.-M
(U
4- e
o o
CX)
cr>
tO O)
•r- >
i- -r-
6 O)
•i- S-
I— <->
^H CO
r*** r H
CTi
s
3
CD
3aious3d
40
-------
"S- i— I
(O 3
(0 <
S-
o
2 -!->
i— CM
<+— f^«
g ^
tL ^ ^ «x
C3 ^
£ *C
•r- s:
CD
i|._ )H"*
O -r-
5-
C 3
O T3
+J LO
(O
•r"* S~
S_ CO
tO +->
> CO
CO O
H^ ^
^
CO
S-
3
CD
CM .,-
1^- U-
c— )
CSW3]
41
-------
co
CO
-CD
CM
CM
r-
in
oo
CM
f-
X.
in
\
in
in
cc
UJ
UJ
o
cc
UJ
UJ
UJ
CC
o
CM
CM
•LU
•z.
•=)
•-D
- in
CM
O LO
03 4->
i- (Si
-l-> 3
C Ol
OJ =3
O <
O O
O +J
4-> CM
c r^.
OJ CT>
o> LO
c/i
>>
i— (C
o en
4-> C
•r~
T3 i-
O) 3
-P T3
(O
i— Uf)
3
E i-
•i- O)
(/I -l->
l-
tO d)
•i- >
i- •<-
ca on
O) O)
E O)
I— O
i i r
i i
r--.
CTl
CD CO
0 Q
co in
CO CM *-i
T08'0 X C1/9WD NOIiyyiN30N03 1N3WIQ3S 1W101
42
-------
CM
c^
tn
I-H
03
i
CM
in
in
in
oc
LU
LU
o
cc
LU
LU
LU
CC
u_
o
00
-•<*
CO
'«a- IT)
rC
S- S-
O i—
O T-
J*
o t.
i— 0)
jr >
o •!-
ro o:
ro j«J
O)
O> O)
•t-> s-
(O O CM
i— r^.
3 «4- <7>
O t— r-l
•i- O
4-> 3 -
$- in
t-H
Q. fO
4J
T3 -i-> tn
CO C 3
+-> O) CD
rO E 3
r- -r- <
3 T3
E (U O
•i- V) 4->
to
M- CM
<*- o r^
o cri
C
o
01 •
•i- in
•r- rO
S- -!-> S
ro -r-
> C O)
3 C
<1J T-
E S- $-
•i- CD 3
I— Q.-O
S
in
CM
OJ
"^
CTl
S--QT x
NoiiyyiN3DN0
3i«in3ii«yd
43
-------
C\J
in
1~4
00
I
in
\
in
in
e
LU
i—
u
o
UJ
CO
oo
- CD
<
00
CO
- «* en
CM Q
LU
ZD
•"3
LO
(NJ
LT)
CM
1 T
CM
cr>
to .
c •>
o un
i-
4-> 3
OJ 3
O
1.-* ^™
-o i-
CU 3
-(-> T3
03
i— un
lei
1/5 4-3
d>
°s
O v^
-l-> S-
ns CD
•r- >
S- T-
res o;
o> oj
E O)
I— o
0)
en
• •
• •
C1/3WD NOIiyyiN30NOO
44
-------
O)
m
in
s
lu
i
£
£
_o
CO
LT>
"CXI
CSJ
-O
3 LT>
i—I
CO >,
CD rs
to
s- o
o ro
-o
O) ^
•(-> ID
ro CM
IS
en
O •!-
C 13
o -o
•-D
•i- s-
5- O)
> O)
O) O
i
i
00
CM 00
en
u_
OD (D
2 £!
CSW3]
WU3U1S
45
-------
ro
in
00
to
CM
if?
in
cc
LU
I—
UJ
o
tt:
UJ
UJ
UJ
d
C_5
u_
o j
-o
CO
ULO
ro
-M r^
n3 CTi
.—I
O ^
•r- LO
4-> i-H
fO
S- >1
4J r—
C 3
O) "~3
O
C O
O •!->
O
CO
•M r^.
C CT>
O) t— I
e
0) CM
in
O CD
•4-> C
-a s_
CU 3
•4J T3
fC
r- LO
Ife
(/I 4->
cu
•M 4-
(O <1)
•i- >
t- M-
-------
m
t*.
10
c^-
i
m
CD
CM
IO
in
cc
LU
^—
LU
o
01
UJ
LU
LU
CC
in
in
CD
Q
CSJ
"J
in
LO
.O
CO
c\j
o
CVJ
Lto
CO
vo
CM
a>
©
in
CO
r^
CT)
ro i
co
i^
en
(O i-H
J- 00
o r>«
•— a*
JC i-H
O
id c\j
OJ >»
+J «d
ns s
3 a>
u c
i- =3
ic -a
a.
in
•a
O) i.
+-> cu
(O -4->
i— QJ
i §
•r- i—
CO -r-
O i.
S-
> o
a> 14-
o
CV4
in
in
s
•
©
s--8T x C3W/3WD NouyyiN32Noo
47
-------
oo
CM
>-
<
en
e
- to s-
-(-> 3
o -a
LD
T3
c s-
fO CU
CD OJ
-P E
(O O
3 •!-
O -i S-
S- O)
fO >
T3
a;
r— S-
o c_>
CO
?~
a
•r—
T3
0)
-(->
fO
r^
3
e
•1—
in
n-
0
c
o
tf—
[ ^
rt3
s_
>
a>
•r—
"o •
3 CO
(O i— 1
1
5-
O •>
•— o
.C CO
o
fO >>
i— ro
CVJ
*
CD
00
a)
3
CT>
H/9W) NOI1VUIN30NOO 3aiDliS3d
48
-------
(O
t
s^
r**
o
g
DC
O
u_
1
t
a
S g
LU
C
_i
i 5
' 5
<:
i
s
£
i
00
3 o
0£
o: °
3 ^
^•J
§ ^
O —
CO I—
0 Q-
i i
! i
1 1
vO ^- CM
c
0
^
-
-
—
4J
c
cu
o
c
So
o
-!->
c
j— QL)
1 »r~
T3
C3 j,/" CU
LU r_
0 *
IJL, O
d ^
° J "*
0 c
-j. 0
i ^R
R u 5S
f= .,
CO (O «—
o "^
^^ ^J *^
^ -"^e
V> C (0
— O
O CS _1 10
r— 1
•
CJ
evi
0)
i.
i~\ -^
en
o •<~
CD CD CD ""
'NOIlVdiN30NODiN3Wia3S
49
-------
J
1
u
^•c
Q
c
U
J
J
i*
U
Ll
Q
C.
1
i
J
J
ELU
1 C£ i
3 0
3 '
l
1
^
» < — i o
, O —
< }r:
s - 1
1— O Q-
Ii i
!
! I
1 1
1 1 1 1 1 1 1 1
CM i"
1
-1
1
1
1
1
L.
_
n
i i
!_[
•
i
i
^^•M
1
•V
•
1
1
-
-
-
§
1
1
i;
1
i
J
f
\
i
i
• i
1 1
i
•
i *
c
CU
<0
3
0
4-
fO
fS
3 E
-XI T3
CU
\S i—
LU 0
LU CO
fV CO •
S ° ^^
i i O*»
•~j -O^H
0 ^ «
SE 45 **
LU 3 CU
0 0 .||
§ •& :
O E
^ CO •
•^^^ J^ ^}
o w '^
f\ pr "3 +j
LU -° *
> II S-
O •»-> o
CQ .^sr
o < ^ S
CM LU r-^
^^^ ^3 fO
^ -0 (0
1 — 3 +J
CO +-> 0
o s g£
1 "•* o c
_l ITS
1
tNJ
o cu
*— ' $_
' g>
50
-------
IxlO"6 r:
i
IxlO"7
UJ
a
oo
LU
a.
P 1x10
-8
1x10"
— MEAN PARTI CULATE ALACHLOR
— PARTI CULATE ALACHLOR WITH SAND
•- PARTI CULATE ALACHLOR WITH
SILT
— PARTICULATE ALACHLOR WITH
CLAY
J
1
0 10 20 30 40 50 60
DISTANCE ABOVE THE MOUTH OF WOLF CREEK, km
Figure 24. Variations of simulated particulate alachlor in the top bed
layer accumulated during the three-year simulation period.
51
-------
of sharp peaks associated with the high flows shown in Figure 8. Figures 10
and 11 indicate time variations of total particulate pesticide concentra-
tions associated with sediments per unit weight of sediment and per unit
volume of water, respectively. There are one large and several small peaks
of particulate alachlor concentrations per unit weight of sediment
(Figure 10) but only one large peak of particulate pesticide concentrations
per unit volume of water (Figure 11). This is due to the fact that
although alachlor concentrations attached to sediment per unit weight of
sediment during the summer of 1972 are high, the sediment concentration
during the same period is relatively low, so that the particulate alachlor
concentration per unit volume of water becomes low. Dissolved and total
(sum of dissolved and particulate) alachlor concentrations near the mouth
of Wolf Creek are shown in Figures 12 and 13, respectively. There are one
large and one small peaks of dissolved and total alachlor during the
three-year simulation period. These simulation results revealed that high
pesticide concentrations in the streams do not directly correlate with peak
runoff or soil erosion events, but rather with the time between pesticide
application and the first storm event after the application. In other
words, pesticide concentrations in the stream have strong seasonal patterns
(high peaks in May and June) corresponding to pesticide applications that
were closely followed by precipitation causing runoff and soil erosion.
The results imply that the amount of pesticide being transported is
controlled by the supply of pesticide on the land surface. Since alachlor
degrades very rapidly after application to farm land, significant improve-
ments in water quality can be obtained through the control and curtailment
of both runoff and erosion shortly after pesticide application.
Instream modeling demonstrates some important effects of sediment
transport on pesticide migration:
1. Through adsorption of dissolved pesticides by sediment, immediate
biological availability of the pesticide may be reduced.
2. Through deposition of contaminated sediment, pesticide concentrations
in a water column will be reduced.
3. However, the contaminated bed sediment then becomes a long-term source
of pollution through resuspension and desorption.
Figure 8 through 13 clearly indicate a small number of sharp peaks of
water discharge, sediment concentration, and concentrations of particulate,
dissolved and total alachlor occurred during the three-year period.
Comparison of Figures 11, 12 and 13 reveals that the majority of alachlor
transported during May 1972 and June 1973 near the mouth of Wolf Creek was
in a dissolved form due to the deposition of contaminated sediment before
it reached the mouth of Wolf Creek. The dissolved pesticide distribution
near the mouth of Wolf Creek shown in Figure 12 was then statistically
analyzed by the program FRANCO to obtain the frequency of occurrence and
duration of given pesticide concentrations, which were then used to
evaluate the potential acute and chronic impacts of the pesticide to
aquatic biota (Onishi et al. 1979b and Olsen and Wise 1979).
52
-------
More detailed flow rates, sediment concentrations, and concentrations
of particulate, dissolved and total alachlor near Wolf Creek mouth for the
two high flow periods during the simulation period are shown in Figures 14
through 21. The highest predicted alachlor concentrations during the
three-year period occurred on June 5, 1973. Figure 21 reveals that when
the predicted maximum alachlor concentration occurred near the mouth of
Wolf Creek on June 5, 1973, sediment carries only approximately 35% of the
total alachlor, while 65% was transported as a dissolved form.
Predicted longitudinal distributions of sediment concentration and
concentrations of dissolved, particulate and total alachlor occurring on
June 5, 1973, are presented in Figures 22 and 23, respectively. Predicted
longitudinal alachlor distribution in the top bed layer after the 3-year
simulation is shown in Figure 24. Since initially it was assumed that
there was no alachlor in the stream bed, Figure 24 shows the predicted
accumulation of alachlor in the bed occurring during the three-year
period.
Unfortunately, there are no measured data available to examine the
accuracy of model predictions.
53
-------
REFERENCES
Baca, R. G., W. W. Waddel, C. R. Cole, A. Brandstetter and 0. B. Cearlock.
"EXPLORE-I: A River Basin Water Quality Model," Battelle, Pacific
Northwest Laboratories, Richland, WA.
Baker, J. L., H. P. Johnson, M. A. Borcherding, and W. R. Payne. 1979.
"Nutrient and Pesticide Movement from Field to Stream: A Field
Study," In Best Management Practices for Agriculture and
Silviculture, Proceedings of the 1978 Connel Agricultural Waste
Management Conference, Ann Arbor Science Publishers, Inc., Ann Arbor,
MI.
Desai, C. S. and J. F. Abel. 1972. Introduction to the Finite Element
Method, A Numerical Method for Engineering Analysis, Van Norstrand
Reinhold Company, New York.
Falco, J. W., K. T. Sampson, and R. F. Carsel. 1976. "Physical Modeling
of Pesticide Degradation," Proceedings of Symposium on Model
Ecosystem Approach to Biodegradation Studies, Society for Industrial
Microbiology, held in Jekyll Island, GA, pp. 193-202.
Krone, R. B. 1962. Flume Studies of the Transport of Sediment in
Estuarial Shoaling Processes. Hydraulic Engineering Laboratory and
Sanitary Engineering Research Laboratory, University of California at
Berkeley.
Leendertse, J. A. 1970. A Water Quality Simulation Model for Well Mixed
Estuaries and Coastal Seas, Principles of Computation, Volume I,
RM-6230-RC, The Rand Corp., Santa Monica, CA.
Mahmood, K. and V. M. Ponce. 1975. Computer Programs for Sediment
Transport—Documentation and Listing, CER75-76 KM-VMP2, Engineering
Research Center, Colorada State University, Fort Collins, CO.
Norton, W. R., I. P. King, and G. T. Orlob. 1973. A Finite Element Model
for Lower Granite Reservoir, Water Resources Engineers, Inc., Walnut
Creek, CA.
Norton, W. R., L. A. Roesner, D. E. Evenson, and J. R. Monser. 1974.
Computer Program Documentation for the Stream Quality Model, Qual-II,
Water Resources Engineers, Inc., Walnut Creek, CA.
Olsen, A. R. and S. E. Wise. 1979. Frequency Analysis of Pesticide
Concentrations for Risk Assessment, Battelle, Pacific Northwest
Laboratories, Richland, WA.
Onishi, Y., P. A. Johanson, R. G. Baca, and E. L. Hilty. 1976. Studies of
Columbia River Water Quality—Development of Mathematical Models for
Sediment and Radionuclide Transport Analysis. BNWL-B-452, Battelle,
Pacific Northwest Laboratories, Richland, WA.
54
-------
Onishi, Y. 1977. Mathematical Simulation of Sediment and Radionuclide
Transport in the Columbia River.BNWL-2228, Battelle, Pacific
Northwest Laboratories, Richland, WA.
Onishi, Y. 1979a. "Evaluation of Mathematical Models of Hazardous
Substances on Surface Waters," Proceedings of the U.S. Environmental
Protection Agency's National Workshop of Verification of Water
Quality Models, West Point, NY.
Onishi, Y. 1979b. User's Manual for EXPLORE-I: A River Basin Water
Quality Model (Hydraulic Module Only).Submitted to U.S.
Environmental Protection Agency by Battelle, Pacific Northwest
Laboratory, Richland, WA.
Onishi, Y. and S. E. Wise. 1978. "Mathematical Modeling of Sediment and
Contaminant Transport in the James River Estuary," Proceedings of the
26th Annual ASCE Hydraulics Division Specialty Conference on
Verification of Mathematical and Physical Models in Hydraulic
Engineering, College Park, MD, pp. 303-310.
Onishi, Y. and S. E. Wise. 1979. User's Manual for the Instream
Sedimejit-Contaminant Transport Model, SERATRA, Battelle, Pacific
Northwest Laboratories, Richland, WA.
Onishi, Y., D. L. Schreiber, and R. B. Codell. 1979a. "Mathematical
Simulation of Sediment and Radionuclide Transport in the Clinch
River, Tennessee," Proceedings of ACS/CSJ Chemical Congress,
Honnolulu. Hawaii, April 2-7, 1979, Contaminants and Sediment, edited
by R. A. Baker, Ann Arbor Publisher, Inc.
Onishi, Y., S. M. Brown, A. R. Olsen, M. A. Parkhurst, S. E. Wise, and
W. H. Walters. 1979b. Methodology for Overland and Instream
Migration and Risk Assessment of Pesticides, Battelle, Pacific
Northwest Laboratories, Richland, WA.
Partheniades, E. 1962. A Study of Erosion and Deposition of Cohesive
Soils in Salt Water. Ph.D. Thesis, University of California at
Berkeley.
Smith, J. H., W. R. Mabey, N. Bohonos, B. R. Holt, S. S. Lee, T. W. Chon
D. C. Bomberger, and T. Mill. 1977. Environmental Pathways of
Selected Chemicals in Freshwater Systems, EPA-600/7-77-133, SRI
International, Menlo Park, CA, Submitted to EPA Environmental
Research Laboratory, Athens, GA.
Stanford Research Institute. 1979. "Fate of Selected Pollutants —
Protocall II." Photolysis in Water. U.S. Environmental Protection
Agency, Contract Number 68-03-2227.
Tofflemire, T. J., S. 0. Quinn, and P. R. Hague. 1979. PCB in the Hudson
River - Mapping, Sediment Sampling and Data Analysis. Technical Paper
No. 57, Department of Environmental Conservation, the State of New
York.
55
-------
U.S. Environmental Protection Agency. 1978. Kepone Mitigation Project
Report, Standard and Criteria Office of Water and Hazardous
Materials, Washington DC.
Vanomi, V. A., Editor. 1975. Sedimentation Engineering. Prepared by the
ASCE Task Committee for the Preparation of the Manual on
Sedimentation of the Sedimentation Committee of the Hydraulics
Division, New York, NY.
Varga, R. S. 1965. Matrix Iterative Analysis, Prentice-Hall, Inc.
Zepp, R. G. and D. M. Cline. 1977. Rates of Direct Photolysis in Aquatic
Environments. Environ. Sci. Technol. 11:359-366.
56 U.S. GOVERNMENT PRINTING, OFFICE 1982-559-092/468
------- |