EPA-660/2-75-009
MAY 1975
Environmental Protection Technology Series
Use of Soil Parameters for Describing
Pesticide Movement Through Soils
National Environmental Research Center
Office of Research and Development
U.S. Environmental Protection Agency
Corvallis, Oregon 97330
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development,
U.S. Environmental Protection Agency, have been grouped into
five series. These five broad categories were established to
facilitate further development and application of environmental
technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in
related fields. The five series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
This report has been assigned to the ENVIRONMENTAL PROTECTION
TECHNOLOGY STUDIES series. This series describes research
performed to develop and demonstrate instrumentation, equipment
and methodology to repair or prevent environmental degradation from
point and non-point sources of pollution. This work provides the
new or improved technology required for the control and treatment
of pollution sources to meet environmental quality standards.
EPA REVIEW NOTICE
This report has been reviewed by the Office of Research and
Development, EPA, and approved for publication. Approval does
not signify that the contents necessarily reflect the views and
policies of the Environmental Protection Agency, nor does mention
of trade names or commercial products constitute endorsement or
recommendation for use.
-------
EPA-660/2-75-009
MAY 1975
USE OF SOIL PARAMETERS FOR DESCRIBING PESTICIDE
MOVEMENT THROUGH SOILS
By
J. II. Davidson
G. H. Brusewitz
D. R. Baker
A. L. Wood
Oklahoma State University
Stillwater, Oklahoma 74074
Project Ho. R-800364
Program Element 1BB039
Roap/Task 21 AYP 13
Project Officer
Dr. George VI. Bailey
Southeast Environmental Research Laboratory
National Environmental Research Center
Athens, Georgia 30601
NATIONAL ENVIRONMENTAL RESEARCH CENTER
OFFICE OF RESEARCH AND DEVELOPMENT
U. S. ENVIRONMENTAL PROTECTION AGENCY
CORVALLIS, OREGON 97330
For Sale by the National Technical Information Service
O.S. Department of Commerce, Springfield, VA 22151
-------
ABSTRACT
Solutions of a differential equation for solute transport are described
for both transient and steady state soil water conditions. The solu-
tions use independently measured soil and soil-pesticide adsorption-
desorption characteristics to describe the movement of a pesticide in
a soil profile. Owing to the complexity of the soil-pesticide adsorp-
tion-desorption characteristics, numerical solutions of the soil water
flow and solute transport equations were used. Numerical dispersion
in the finite difference solution of the solute transport equation was
considered and a correction included in the solution.
Experimental results from laboratory and field studies were used to
test the numerical solutions for their ability to describe the move-
14
ment and distribution of pesticides in a soil profile with time. C-
labeled herbicides were used in the laboratory experiments. Equilib-
rium adsorption and desorption isotherms were not single-valued rela-
tionships in any of the soil-herbicide systems studied. Kinetic adsorp-
tion-desorption models were evaluated and shown to be inadequate for
predicting herbicide mobility at high average pore-^water velocities.
An empirical model was developed, based on experimental data, and used
to describe herbicide movement at high flow rates. The procedure
defined a fraction of the soil adsorbing sites that were in equilibrium
with the herbicide during displacement. This fraction was a function
of the average pore-water velocity. Using this empirical model, the
distribution of a herbicide in a soil during infiltration was ade-
quately described with the soil-water and solute transport equations.
ii
-------
This report was submitted in fulfillment of Project Number R-800364,
by J. M. Davidson, G. H. Brusewitz, D. R. Baker, and A. L. Wood,
Oklahoma State University, Stillwater, under the partial sponsorship of
the Environmental Protection Agency. Work was completed as of August,
1974.
111
-------
CONTENTS
Sections Page
I. Conclusions 1
II. Recommendations 4
III. Introduction 6
IV. Theory 8
Equations for Water Flow 8
Equations for Msorption-Desorption 11
Equations for Pesticide Transport 14
V. Materials and Methods 21
Laboratory Experiments 21
Steady State Soil Water Studies 21
Transient Soil Water Studies 25
Field Experiments 30
VI. Results and Discussion 34
Laboratory Experiments 36
Steady State Soil Water Studies 36
Transient Soil Water Studies 50
Field Experiments 66
VII. References 78
VIII. Glossary 82
IX. Appendices
Appendix A 85
Computer Program for Solute Transport Under Steady
State Soil Water Conditions
Input-Cutput Data Format 86
Flow Chart for Solute Transport Program 91
Solute Transport Program 99
IV
-------
Appendix B 107
Ocmputer Program for Simulating Simultaneous Movement
of Water and Solute Through Soils
Input-CXitput Data Format 108
Flow Chart for Simultaneous Transport of Water
and Solute Program 121
Simultaneous Transport of Water and Solute
Program 128
v
-------
FIGURES
NO.
1 Schematic drawing of apparatus used to study the dis-
placement and distribution of fluometuron in Norge
loam soil 22
2 Schematic of gamma-ray attenuation apparatus 26
3 Soil column and water application apparatus used in
transient soil water studies 28
4 Computed relative solution concentration distribution
using an exact and finite difference solution of
equation (13). Calculations were made using an
average pore-water velocity, dispersion coefficient,
distribution coefficient, and soil water content of
2 33
4.0 cm/hr, 0.2 cm /hr, 2.0 and 0.34 on /on , respec-
tively. 35
5 Computed solution concentration distribution for a
non-adsorbed solute during infiltration with the
numerical dispersion correction (equations [21] and
[22]) for transient soil water conditions included
and excluded. 37
6 Equilibrium adsorption and desorption isotherms
for fluometuron and Norge loam. Insert shows
VI
-------
No. Page
6 relationship between the description distribution
(Cant.) coefficient, Kp, and the maximum fluometuron
solution concentration, C^x. 38
7 Transient adsorption-desorption behavior of fluo-
meturon in Norge loam soil for 0.59 and 5.5 cm/hr
average pore water velocities. 4Q
8 Calculated and experimental relative solution con-
centration distribution for 0.59 cm/hr average
pore^water velocity. Calculated lines were obtained
from a numerical solution of equation (14) . 42
9 Calculated and experimental relative adsorbed con-
centration distribution for 0.59 cm/hr average pore-
water velocity. Calculated lines were obtained from
a numerical solution of equation (14) . 43
10 Calculated and experimental relative solution con-
centration distribution for 5.5 cm/hr average pore-
water velocity. Calculated lines were obtained from
a numerical solution of equation (14) . 45
11 Calculated and experimental relative adsorbed con-
centration distribution for 5.5 cm/hr average pore-
water velocity. Calculated lines were obtained
from a numerical solution of equation (14) . 4g
12 Calculated and experimental relative effluent con-
centration distribution from NOrge loam soil. The
calculated values were obtained using an exact
solution of equation (13) and various kp/VQ values
for the slow average pore-water velocity.
48
vn
-------
No.
13 Adsorption and desorption isotherms for fluomsturon
on Cobb sand. Solid and broken lines are best fit
for adsorption and desorption, respectively. 51
14 Fluometuron concentration distributions versus time
for selected soil depths. The soil water flux and
initial soil water content were 1.03 cm/hr and 0.13
cm-Vein . Solid lines were eye fitted to experimental
data. 53
15 Solution and adsorbed fluometuron concentration dis-
tributions immediately following infiltration.
Average soil-water flux was 29.0 cm/hr. The 6^ and 9f
are the initial and final volumetric soil water content
in the soil surface region. Solid lines were eye fitted
to experimental data. 54
16 Solution and adsorbed fluometuron concentration dis-
tributions intnediately following infiltration. Average
soil water flux was 10.0 cm/hr. The Qj_ and 8f are the
initial and final volumetric soil water content in the
soil surface region. Solid and dashed lines were eye
fitted to experimental data. 55
17 Fluometuron solution concentration and water
content distributions for the same accumulative
infiltration into an initially wet and dry soil.
The 6j_ and 6f are the initial and final volumetric
soil water content in the soil surface region.
Solid and dashed lines were eye fitted to experi-
mental data.
vxn
-------
No. Page
18 Fluoneturon solution concentration distributions
for equal accumulative infiltration into air-dry
soil. The average soil water flux at the soil
surface was 29.0 and 1.0 cm/hr. The Qj_ and 6f
are the initial and final volumetric soil water
content in the soil surface region. Solid and
dashed lines were eye fitted to experimental data. 58
19 Experimental and calculated fluometuron solution con-
centration and water content distributions after 15
and 59 minutes of infiltration. Initial soil-water
content was 0.005 cm-Van and the average soil water
flux was 29.0 on/hr. Solid lines were calculated
using equations (3) and (21). 60
20 Experimental and calculated fluoneturon solution
concentration and water content distributions after
30 and 69 minutes of infiltration. Initial soil
water content was 0.13 cmVcm3 and the average soil-
water flux was 10.1 cm/hr. Solid lines were calcu-
lated using equations (3) and (21). 61
21 Experimental and calculated fluometuron solution
concentration and water content distributions after
266 minutes of infiltration. Initial soil water
content was 0.005 cm-Van and the soil-water flux was
4.89 cm/hr. Solid lines were calculated using equa-
tions (3) and (21). 62
22 Experimental and calculated fluometuron solution
concentration and water content distributions after
IX
-------
NO.
22 30 and 69 minutes of infiltration. Initial soil
water content was 0.13 on3/cm3 and the average soil-
water flux was 10.1 on/hr. Solid lines were calcu-
lated using equations (3) and (21) for p = 0.77 g/cm3
(FREQ = 0.5). 64
23 Experimental and calculated adsorbed flucmeturon concen-
tration distribution. The 6j_ and Qf are the initial
and final volumetric soil water content in the soil sur-
face region. The average soil-water flux was 10.1 cm/hr.
Solid and dashed lines were calculated for p = 1.54 and
0.77 g/on3 (FREQ = 0.5). 65
24 Experimental and calculated total (solution and adsorbed)
fluoneturon concentration distribution in Teller sandy
loam soil following the infiltration of 5.7 cm of water
(t = 0) and 48 hours of evaporation (-0.02 cm/hr) at
the soil surface. The initial soil water content was
0.18 cm3/cm3. Vertical lines are the measured fluome-
turon concentration in the soil under natural field
conditions following 5.7 cm of water infiltration and 48
hours of evaporation and redistribution. 71
25 Simulated pesticide distribution in a soil profile
with time for steady state soil water conditions.
Average soil water content, water flux, distribution
coefficient, N and N1 were 0.20 cm3/cm , 0.128 cm/day,
0.21, 1.19 and 1.7, respectively. 77
x
-------
TABLES
NO. Page
1 Physical and Chemical Characteristics of Norge Loam Soil 21
2 Experimental Parameters for Soil Columns Studied 41
3 Parameters Associated with Desorption of Fluometuron Fran
Cobb Sand at 25+0.5 C. 50
4 The Effect of Rainfall Intensity on the Concentration and
Percent of Applied Fluometuron in Runoff Water and Sediment 67
5 Chemical and Bioassay Analysis of Fluometuron at Three Soil
Depths for Different Rainfall Intensities 69
6 The Effect of Antecedent Soil Water Content on the Concen-
trations and Percent of the Applied Fluoneturon in the
Runoff Water and Sediment 72
7 Chemical and Bioassay Analysis of Fluometuron at Three Soil
Depths for Different Rainfall Intensities 73
8 Chemical Analysis of Fluometuron in the Soil Profile Over
an Eighty-Four Day Period 74
9 Bioassay Analysis of Flucmeturon in the Soil Profile Over
an Eighty-Four Day Period 75
xi
-------
ACKNOKMXMENTS
The assistance of Dr. A. G. Hornsby (Environmental Protection Agency,
Ada, Oklahoma), Mr. M. Th. van Genuchten and Dr. P. J. Wierenga (New
Mexico State University, Las Cruces), Dr. F. L. Baldwin (University of
Arkansas, Little Rock) and Dr. P. W. Santelmann and E. W. Chin Choy
(Oklahoma State University, Stillwater) is gratefully acknowledged.
Also, the cooperation and overall project coordination by Dr. George
Bailey and his staff at the Southeast Environmental Research Laboratory
in Athens, Georgia is appreciated.
The principal investigator (James M. Davidson) wishes to express his
appreciation to the Agricultural Experiment Station at Oklahoma State
University for their financial support and recognition of agriculture's
responsibility to prevent environmental contamination. He is also
indebted to the Department of Soil Science and the Institute of Food
and Agricultural Sciences at the University of Florida, Gainesville
for the use of their facilities during his Sabbatical leave (1972-73)
from Oklahoma State University, Stillwater.
The assistance of Mrs. Mildred King for preparing quarterly financial
statements and Mrs. Neva Griffith and Mrs. Beverly Gable for typing
the final report is also appreciated.
Xll
-------
SECTION I
CONCLUSIONS
Numerous laboratory and field studies were used to evaluate the various
parameters influencing the mobility of a herbicide in the soil. These
data were then used to test existing mathematical models for their
ability to describe the movement and distribution of selected herbicides
in a soil profile with time. Based on these studies, several conclu-
sions can be drawn.
1. Analytic solutions of a transport model frequently used in chroma-
tography work were not adequate since they generally required various
simplifying assumptions that were not characteristic of the soil-water-
herbicide system. Also, the boundary conditions for which the solution
was valid could not be maintained for extended periods of time under
natural field conditions.
2. The adsorption and desorption isotherms for all herbicide soil sys-
tems studied were not single-valued functions. Excluding this nonsingu-
larity when calculating the herbicide distribution in a soil resulted
in a serious disagreement between the calculated and observed distribu-
tions. The experimental herbicide distributions were significantly
wider and the maximum concentrations lower owing to the nonsignularity.
Both laboratory and field results suggest the possibility of the
herbicide being irreversibly adsorbed to the soil.
3. Kinetic adsorption-desorption models were evaluated and found inade-
quate for predicting herbicide mobility at high average pore-water
-------
velocities. These models, however, agreed with the equilibrium
Freundlich adsorption model when describing herbicide distributions at
low pore-water velocities. The kinetic rate coefficients were difficult
to measure and unless measured independently, could be selected on the
basis of data fitting and thus erroneously include the spreading owing
to hydrodynamic dispersion.
4. The herbicides studies appeared to be in equilibrium with sane
fraction of the soil mass when displaced through the soil at high pore-
water velocities. An empirical model was used to describe the movement
of several herbicides at high flow rates. The procedure defines a
fraction of the adsorbing sites that were in equilibrium with the herbi-
cide during the displacement. The fraction was a function of the
average pore-water velocity. The procedure provides a reasonable fit
between data and calculated distributions for large flow rates. Adsorp-
tion appears to be diffusion controlled as well as a function of pore
geometry and pore size distribution.
5. Numerical solutions of the water and solute transport equations
were solved simultaneously in order to describe herbicide movement
during infiltration and soil water redistribution.. The calculated and
experimental (laboratory and field) values were in good agreement for
several infiltration and redistribution studies. Equilibrium adsorption
was assumed, but the parameters for adsorption and desorption could be
the same or different for each process. Hydrodynamic dispersion was
also considered and included in the solution.
6. Numerical dispersion in the finite difference solution of the
solute transport equation was considered and a correction included. The
solution, with a correction for numerical dispersion, was within 5% of
an analytical solution. A single-valued adsorption-desoprtion function
was used for the evaluation.
7. Fluometuron losses in the runoff under simulated rainfall conditions
-------
were less than 0.5% of the total amount of chemical applied to the dry
soil* Losses were highest when the herbicide was applied to a wet rather
than a dry soil. The highest fluometuron losses occurred in the first
runoff producing rainfall event after herbicide application.
8. The quantity of fluometuron on a gram of sediment was greater than
that in an equal quantity of water. Because of the large dilution in
the runoff, irreversible adsorption must be considered as well as the
nonsingularity between adsorption and desorption.
-------
SECTION II
IMXWMEMiATIONS
The results of this study illustrate the feasibility of predicting the
movement and distribution of herbicides under natural field conditions.
The numerical solutions, in principle, could also be used to describe
the behavior of other soluble components present in or applied to a
soil, e.g., the movement of nitrogen and phosphorus. Since the water
and solute transport equations were solved simultaneously, the suita-
bility of the model to describe the transport processes at the soil
surface (0-5 cm) should be experimentally evaluated.
Soil-pore geometry and size distribution appear to be major parameters
not presently considered in the solute transport model. The importance
of these soil physical properties to the adsorption process during
solute movement through a soil need to be studied in detail. A diffu-
sion controlled adsorption process is indicated by the present data.
Experiments which will identify the parameters and define a model to
describe the adsorption of pesticides at high flow rates should be
conducted.
The nonsingularity between adsorption and desorption of herbicides has
been well documented. However, to date these experiments have been
confined to 1:1 (water to soil) batch experiments involving only one
adsorption and desorption cycle. Under natural field conditions, the
soil to water ratio varies and several adsorption-desorption cycles may
occur within a short period of time. The adsorption-desorption process
needs to be studied under more general environmental conditions. This
-------
type of information is necessary for any continuous simulation of herbi-
cide movement and distribution in the soil.
Biological degradation submodels need to be developed and added to the
solute transport model. These processes are significant in the soil
surface region where the biomass is sensitive to temperature and water
fluctuations. These models are also necessary for a continuous simula-
tion of a biodegradable solute during the growing season. This informa-
tion can be used to describe the attenuation of a soil applied pesticide
with tine.
-------
SECTION III
INTRODUCTION
The use of pesticides in agricultural production programs has increased
over the past several years. These chemicals have contributed to
increased crop yields and in many cases, lower production costs. How-
ever, the effects of these chemicals are not always beneficial. In some
instances, crop injury or failure has occurred because of herbicide
residues from a previous cropping system. Environmental contamination
is also a possibility that must be considered with the widespread use
of pesticides. Some of these chemicals are also toxic to man and have
undesirable effects on certain ecosystems (e.g.) biological magnifica-
tion. Therefore, the agricultural producer as well as the environmen-
talist is concerned with the fate of a pesticide in a soil profile.
Compounds of particular interest are those that are nonvolatile and
biologically stable. These pesticides present a particular problem in
that they may leave the target area on sediment and in the water flow-
ing across and/or through the soil. The distribution of a pesticide
with time within a soil profile depends upon its interaction (adsorp-
tion-desorption) with the soil matrix and the rate and direction of
soil water movement.
Several mathematical models for describing the movement of an adsorbed
chemical through a soil have been proposed1 2 3 4 and tested5 6 7 8 9 10
In general, the analytical solutions of these models, subject to rigid
initial and boundary conditions, have been accepted without adequate
validation. Many of the solutions are based upon assumptions which
over-simplify the soil-water system and adsorption-desorption character-
istics of the solute. For example, Hashimoto et al.1 and Lapidus and
Amundson2 assumed pointwise equilibrium adsorption; whereas, the solu-
-------
tions of Lindstron et al.^ and Oddson et al.^ used a kinetic adsorption
model but assumed diffusion and dispersion to be negligible. Also, each
of the models required the adsorption and desorption process to be a
single-valued function; whereas, experimental data7 -^have shown a non-
singular relationship to exist for several herbicide soil systems. Thus,
it appears that the problem is too complex to obtain any analytical solu-
tion of the differential equations currently being used to describe both
the transport and adsorption-desorption processes. The above models
also require steady state soil water conditions.
The simultaneous transfer of water and solutes through a natural field
soil occur generally as a transient process. The direction and rate at
which each of the phases move are dependent upon the physical character-
istics of the soil, climatic conditions, activity of the plant and
biological system and interactions between the solute and soil matrix.
An understanding of how these processes influence the transfer of each
phase is essential in the development of management practices for mini-
mum contamination to soil and/or water supplies. Numerical techniques
to simulate the transient behavior of water and noninteracting solutes
12 13
have been described by Bresler and Hanks , Bresler , and Kirda et
al. However, a numerical solution has not been developed and tested
for the simultaneous transfer of water and adsorbed solutes in a soil.
In this study, a numerical simulation procedure for describing the
simultaneous transport of water and adsorbed and nonadsorbed solutes
was developed and evaluated. The combined effect of convection, adsorp-
tion-desorption, and dispersion (diffusion and mechanical dispersion)
were considered as well as a correction for numerical dispersion in the
finite difference solution of the solute transport equation. Experimen-
tal laboratory and field data were used to evaluate the suitability of
the two numerical solutions to describe the movement of each phase.
Adsorption and desorption was also studied for several herbicide-soil
systems. Several adsorption models were considered and evaluated in
the solute transport equation.
-------
SECTION IV
THEORY
The basic equations used to describe the movement of a pesticide in a
soil with isothermal conditions will be presented and discussed. Pesti
cide transport in the presence of both steady state and transient soil
water conditions will be considered. Computer programs (FORTRAN IV) ,
flow charts and input-output data formats for the numerical solution of
the chemical and chemical plus soil water transport equations are pre-
sented in Appendix A and B, respectively. All symbols used in this
manuscript are defined in the Glossary.
EQUATIONS FOR WATER FLOW
Darcy's equation was assumed to describe the flow of water in a satu-
rated and/or unsaturated soil. The equation for flow in one dimension
in an isotropic soil is
where q = volumetric soil-^water flux (cm/hr)
K(9) = hydraulic conductivity (cm/hr) which is a function of the
soil water content 0 (cm-Van-^)
4> = hydraulic head (on)
z = soil depth (cm) measured positive downward
The volumetric soil^water flux is smaller in magnitude than the velocity
of the water moving through the tortuous pore sequences. The average
pore-water velocity is obtained by dividing the soil-water flux by the
volumetric soil-water fraction.
-------
Equation (1) describes a steady state soil water condition. Using
equation (1) in the continuity equation (conservation of mass) gives
the basic equation used to describe the transient flow of water through
soils. The equation for one-dimensional flow is
38 3 [" 9h
3t -•§£*(6) 3^
(2)
where 6 = volumetric water fraction (cm-Van-^)
t = time (hr)
h = soil water pressure head (on)
The implicit finite difference analog of equation (2) for pressure head
is
hi - hi-1 . ,
1 1 _ l-%
At
= Ri
,
2Az -
2Az
+ h -f 2Az -
(3)
2Az
where Q. =
i = space or depth indice
j = time indice
Equation (3) is solved using a similar procedure to that described by
Hanks and Bowers . The relationships between K and 6 and h and 8 are
stored in the computer in a matrix and once h? has been calculated, 6-?
and K-? are obtained from the matrix.
The time increment, At, in equation (3) is variable and dependent upon
the water infiltration and redistribution rate. The increment At
used to calculate h. during infiltration is obtained from
j-l
At
j-l _ At°I°
(4)
where At = selected time increment for t = 0
1° = infiltration rate at t = 2At°
-------
The infiltration rate is obtained by using Darcy's equation. When evap-
oration at the soil surface and redistribution occur simultaneously
within the profile, the time increment is obtained from
max
where K = maximum hydraulic conductivity at cessation of infiltration
m
At = time increment at cessation of infiltration
m
= maximum hydraulic conductivity in soil profile at j-1 time
after the initiation of evaporation and/or redistribution
of water
The computer algorithms (FORTRAN IV) , flow chart and input-output data
format are in Appendix B as part of the Hybrid solution for the simul-
taneous movement of water and solute.
The numerical solution of equation (3) will accept any initial distribu-
tion of 6 and h with depth. The three physical cases considered in the
program are: 1) constant Q and hr at z = 0 for duration of infiltration;
o o
(2) constant water flux or any known water flux distribution which can be
approximated as a series of constant fluxs with time at x = 0; and 3)
evaporation (constant or variable with time) from the soil surface.
Hysteresis in the 6 and h relationship was not considered, but could be
included using a procedure similar to that described by Rubin16 and
Stable17.
The soil water flux and average pore^water velocity in each Az increment
of soil is calculated using equation (1) . The parameters are obtained
from equation (2) for each time increment and the matrix of h versus 6
and K versus 9. Soil water flow rate and soil water content data are
required in the pesticide transport equations described in the following
sections .
10
-------
EQUATIONS FOR ADSORPTION-DESORPTION
The distance an organic pesticide will move through a soil profile per
unit of time is significantly influenced by the intermittent capture
and release of the individual pesticide molecules by the soil matrix.
The average time required for this adsorption or desorption to occur
is dependent upon the free energy of activation required for each pro-
cess. Adsorption and desorption may also be influenced by diffusion
rate; where the capture and release occur gradually as the molecules
diffuse in and out of the region of adsorption.
Adsorption and desorption in a one- site system (one adsorption-desorp-
tion mechanism) may be represented by the following model
S-Adsorbent (6)
where k^ = adsorption rate coefficient (hr )
kj} = desorption rate coefficient (hr )
C = pesticide solution concentration (yg/on3)
S = adsorbed pesticide concentration (yg/g)
The rate of mass transfer to the adsorbed phase can be expressed as
ec1/N
dt
ec
1/N
A
k
- S
DP
(7)
where p = bulk density of the soil system (g/cm3)
N = constant that varies with pesticide and adsorbent
The other terms have been described previously and are in the Glossary.
o
This same relation was used by Lapidus and Amundson (for N = 1) to
represent nonequilibrium adsorption and is similar in form to the
adsorptic
et al.10
adsorption model used by Oddson et al. (for N = 1) and van Genuchten
Based upon the Arrhenius equation, the adsorption and desorption rate
coefficients are temperature dependent and proportional to the free
11
-------
energies of activation for adsorption and desorption
kA = k£ exp -AGA t/RT (8a)
kD = k£ exp -AGp t/RT (8b)
where kj^ = constant generally referred to as frequency factor for
adsorption
kl = constant generally referred to as frequency factor for
desorption
AG^t = free energy of activation for adsorption
AGj}t = free energy of activation for desorption
R = gas constant
T = absolute temperature
18 19
Fava and Eyring and Lindstrom et al. present a more complex adsorp-
tion model than that of equation (7) by allowing the activation energies
for adsorption and desorption to change with surface coverage. Equation
(7) assumes that the adsorption-desorption process is a single-valued
function and that a linear relationship exists between S and C at equil-
ibrium (8S/8t = 0) when N is unity. The coefficient kA/ko i-s constant
for a specific soil and pesticide and is frequently referred to as a
19
partition coefficient.
A solution to equation (7) can be obtained by using Laplace transforms.
For initial conditions of C(o) and S(o) equal to zero, the solution of
equation (7) for N = 1 is
*-• _t-
exp[-kD(t-T)]dx (9)
o
where T = variable of intergration
Note that the adsorbed concentration is an exponential function of time.
20 21 22
The Freundlich equation has been shown by several investigators
to describe the relationship between the solution and adsorbed concen-
12
-------
trations. The Freundlich equation is
S = KAC1/N (10)
where KA = adsorption distribution coefficient (yg ~ ^cm ^ gnf )
For equilibrium adsorption conditions, equations (7) and (10) are equal
with KA = kA6/kDp. Van Genuchten et al. have used equations (7) and
(10) in a solute transport equation to describe adsorption in a flowing
system.
The adsorption-desorption phenomenon is generally assumed to be a single-
valued function for isothermal conditions. The results of Davidson and
21 11
McDougal and Swanson and Dutt , however, indicate that adsorption and
desorption were not single-valued functions for all herbicide-soil sys-
tems. For equilibrium desorption, more herbicide was retained on the
soil at a given solution concentration than during adsorption. Rewrit-
ing equation (10) for desorption gives
S' = KDC'1/N (11)
where prime (') = desorption
[1-1/N'l 3/N"
KD = desorption distribution coefficient (yg Cm '
gnf1)
When the adsorption process ceases, the solution and adsorbed concentra-
tions in equations (10) and (11) are equal (C = C' and S = S1). At this
point, the solution concentration is a maximum (Cmax). Setting equation
(10) equal to equation (11) (Sjj^x) = S'max) and solving for KD
C'max) 9ives
KD = KA cmax
where C^^^ = maximum pesticide solution concentration prior to desorp-
tion (yg/Cm^)
Sjnax = maximum adsorbed pesticide concentration prior to desorp-
tion (yg/g)
Note that the desorption distribution coefficient KD is not constant,
but a function of the maximum solution concentration at the time desorp-
13
-------
tion was iniated. Also, the desorption distribution coefficient may
exceed the adsorption distribution coefficient, depending upon the
amount of chemical in solution prior to desorption. Equation (12)
assumes N1 is not a function of the maximum solution concentration; how-
ever, van Genuchten et al.^ have shown that N1 was inversely related to
Cmax, but could be assumed constant in many instances without serious
error.
EQUATIONS FOR PESTICIDE TRANSPORT
The mass transport of a pesticide through soil in one dimension is
generally assumed to be governed by convection (viscous movement of the
soil water) , diffusion (thermal motion of the pesticide within the soil
water) , mechanical dispersion (mixing owing to pore water velocity dis-
tribution) and adsorption-desorption. The partial differential equation
frequently used to describe the miscible displacement of a pesticide for
transient soil water flow conditions is
2
where D = apparent diffusion or dispersion coefficient (cm /hr)
Assuming D is independent of z and using the equation of continuity for
water (Warrick et al.23) and equations (10) and (11), equation (13)
becomes
9C _ D
+ =;
8t W „ 2 ' W
dZ
.(1/N-l)
D _90_ _ q
6 8z
where W = 1 + p^_ KACV ' ' for adsorption (15a)
0N
W = 1 + p_ KDC' "- for desorption (15b)
6N1
Based upon the work of van Genuchten et al. and Hornsby and Davidson ,
the equilibrium Freundlich adsorption model appears adequate for most
transient systems. The adsorption distribution coefficients can be
measured independently using standard batch procedures^!; whereas, the
adsorption and desorption rate coefficients are difficult to measure.
14
-------
The numerical solution of equation (14) , for steady state soil water
condition, using an explicit procedure of the finite difference scheme
s
c? = e?"1 + 4^-0
i-l
.-1
Az
(16)
Numerical solutions by finite difference are subject to serious numer-
ical dispersion (smearing of the concentration distribution) unless the
time and spatial increments, At and Az, for both 3C/3t and 3C/3z approach
zero. To correct for numerical dispersion, a procedure similar to that
24 ]"3
described by Chaudhari and Bresler for a nonadsorbed solute (W = 1)
was used. To approximate 3C/3t and 3C/3z, C, , is expanded about
vZ,t;
t + At and z - Az using a Taylor series, only the first and second order
terms fron the expansions are retained. Expanding C, .. about t using
\z i\.)
a Taylor series gives
C(t + At) = C(t) +
D 32C
LW3z2
q 3C
ew 3z
At + 9
At + at
D
W
32C q 3C
2 ew 3z
At2
2
(17)
Using the chain rule and equations (14) and (15a), equation (17) becomes
2
C(t+,
qp
w3e2
M:) c(t)
•~
D
W
n AT -M
KA C8C
N \ 3z
9
cTC
3z2
.0
t
/
6W
?
w2e2
—
3C
3z
At -
2
DJ p
_ °
? 9
3^C L q^p K „
2 J 3 N
3z WQ
K\(
NA
n />•
/EC\
U2/
At"
2
(18)
2 22
Since 1/N for most herbicide soil-systems is close to unity, (3 C/3z )
1/N(— - 1) can be assumed to be negligible and equation (18) reduced to
C(t+At) = C
(t)
2 2"
DAt q At
W
3 C _ qAt 3C
^2 we 3z
3z
(19)
2 2
The numerical dispersion associated with 3c/3z and 3 C/3z is corrected
24
for in a similar manner to that shown by Chaudhari where C. , was
15
-------
expanded in Taylor series about z-Az. The approximate correction for
numerical dispersion in solving equation (16) for a steady state soil
water condition is
(20)
An accurate approximation of equation (13) for steady state soil water
conditions may be obtained when equation (16) uses D-E^ for the disper-
sion coefficient. The computer algorithms (K3RTRAN IM, flow chart
and input-output data formats for the solution of equation (16) with
equation (20) are given in Appendix A.
The numerical solution of equation (14) for transient soil water condi-
tions was also obtained using an explicit procedure of the finite dif-
ference scheme. However, owing to the large number of terms required to
account for numerical dispersion introduced by 3C/8t (W is a function of
both C and 0 for the transient problem) in equation (14), the special
case of 1C. = 0 (W = 1) was assumed to account for the major portion of
the error from this term in the finite difference solution. The solu-
tion to equation (14) with the correction for numerical dispersion is
O _ pj-l ,
-f • *-*• »
with
X? =
G3 =
i-1 _ or^1 + pJ~3
i + c^-j
Az
2Az
xikz-(xi)2 At- G? AzAt
11 i-T-
Az
At
^ - 2±.<
Xi 2 (
(21)
(22a)
(22b)
-2Ql. +Q!
16
-------
Az'
(22c)
The G and D terms are for the special case of K = 0. The space incre-
ment, Az, is the major source of numerical dispersion (see D and D in
equations [20] and [22b]) since At is generally several orders of magni-
tude smaller. The products involving the time increment in equations
(20) and (22b) are also, in general, significantly less than Az even for
high soil water flow rates during infiltration. The computer algorithms
(FORTRAN IV) , flow chart and input-output data formats for the simultan-
eous movement of water and pesticide are given in Appendix B.
Because equilibrium adsorption-desorption is assumed, the adsorbed con-
centration, S, in each Az increment can be obtained using the calculated
solution concentration (equations [16] or [21]) at that position and the
corresponding adsorption or desorption distribution coefficient (equa-
tions [10] and [11] ) . The quantity of pesticide in each phase is then
calculated for each computer print out.
The calculation of W for equations (16) and (21) using equations (15a)
and (15b) proceeds as follows : At various locations in the soil , z = z-?
the pesticide solution concentration increases as the invading miscible
pesticide front moves through the soil, reaches a maximum (C ) and
.
begins to decrease. Equation (15a) is used at z-? when the concentration
is increasing, and equation (15b) is used when the solution concentration
at a given position is decreasing. The calculation of 1C (equation [12])
is made for each z , position when a decrease in solution concentration
is first observed at that location, and once calculated, remains con-
stant for that depth increment (see Appendix B) .
17
-------
The dispersion coefficient, D, in equation (13) is pore-water velocity
dependent and increases monotonically with an increase in pore-water
velocity. Kirda et al. and Btied and Conbarnous25 have evaluated
various relationships between the dispersion coefficient and flow veloc-
ity and molecular diffusion. The relationships studied were empirical,
but did exhibit the fact that below some pore-water velocity, the con-
tribution of molecular diffusion to mixing can be considered significant.
The following empirical relation was observed for the range of soil
water flow rates studied by Hornsby and Davidson' and vanGenuchten et
al.10
D = 0.08 + 0.02 (23)
Equation (23) was used to estimate the dispersion coefficient for equa-
tions (16) and (21).
The following initial and boundary conditions were considered in the
numerical solution of equations (16) and (21) .
C? = 0 z j> 0 t=0 (24a)
C? = C z = 0 0 T
Ci = C(z) °lzlzo t=0
Si = S(z) (24b)
c.._ = o Z>Z0 t = o
„- " ? C = ° z = 0 t>0
C-? = 0 ? > (-Q\it- +- > n
1 '-'-VAQ^1- T-^U
18
-------
where A ;= quantity of pesticide applied to the soil surface or
introduce at z = 0
(q/A9)t = position of wetting front
A6 = change in volumetric water content
Boundary condition (24a) simulates the case where a concentration CQ of
soluble pesticide is entering the soil at z = 0 (C being equal to the
maximum solubility of the pesticide in water) and continues to enter the
soil until all the material at z = 0 for t = 0 is in the soil.
For a stable solution of equation (16) or (21) the following condition
was used
(D-EUAt j,
h
(25)
Lindstrom et al. have used this criterion for steady state soil water
conditions and shown that it provides sufficient stability. However,
an additional criterion was found to be necessary in the solution of
equation (21) because of the mass transport term.
q D 96 ., . ,
The Crank-Nicholsen implicit finite difference scheme used to approxi-
mate equation (2) is unconditionally stable for any At and Az values
selected when the hydraulic conductivity is constant. The stability
conditions when K is a nonlinear function of 6 have not been determined
analytically to date.
The size of the Az increment in equations (3) , (16) , and (21) was fixed,
with Az in equation (21) fixed at an integral multiple of the Az of the
water flow solution. The At of the solute equation was allowed to
adjust itself at each time step according to equations (25) and (26)
while the At of the water flow equation varied according to equations
(4) and (5) . Since the solute time increments and the water flow time
increments were calculated independent of one another, the calculatations
were allowed to proceed in the water flow equation until the calculated
time was greater than or equal to the calculated time of the pesticide
19
-------
transport equation. When this condition occurs, the calculations of
the solute equation were carried out until the pesticide equation time
was greater than or equal to the water flow time, at which point, the
water flow equation was again used and the cycle repeated. Calculations
were carried out using an IBM 360/65 computer.
20
-------
SECTION V
MATERIALS AND METHODS
LABORATORY EXPERIMENTS
Steady State Soil Water Studies
Soil from the surface 15 on of a Norge loam profile was Ca-saturated
using 0.5 N calcium acetate and washed with 0.01 N CaSO^ until no ace-
tate was detected. The physical and chemical properties of the soil
used in this study are given in Table 1.
Table 1. PHYSICAL AND CHEMICAL CHARACTERISTTCS OF NORGE LOAM SOIL
Soil
Aggregate
size
PH
CEC
Exchangeable Organic
Ca matter
meq/100 g
Norge loam, calcium
saturated
>2.0
6.6 9.2
7.1
1.7
Mechanical analysis: 46.0% sand, 37.5% silt, 16.4% clay
The soil was air dried to 0.015 g/g water content and then passed
through a 2.0 mm sieve. The soil was packed into lucite columns 30 cm
long and 7.65 on inside diameter to an approximate bulk density of 1.5
g/cm . Fritted-glass plates were sealed to the ends of the column to
hold the soil in place and allow the inflow solution to be quickly
changed (Fig. 1) . The dry soil was then saturated with 0.01 N CaSO,
solution and inflow and outflow volumes measured until equal. The
average pore^water velocity was controlled with a constant volume pump.
21
-------
FLUSHING BOTTLES
C „
^•/'POROUS' MATERIAL v!x
TEST TUBE IN
FRACTION COLLECTOR
CONSTANT VOLUME PUMP
-20mm O.D.
MEDIUM POROSITY
FRITTED GLASS
DfSC
Figure 1. Schematic drawing of apparatus vised to study the displacement
and distribution of fluometuron in Norge loam soil.
22
-------
All studies were conducted at 25 + 0.5 C.
14
A 200 ml pulse of C-labeled 1, l-dimethyl-3(a,a,a,-trifluor-m-tolyl)
urea (fluometuron) in a 0.01 N CaS04 solution (5yc/l), equivalent to
1.11 kg/ha, was introduced at the inflow position and then displaced
through the saturate soil with .01 N CaS04 at a constant flow rate.
A total volume, including the 200 ml pulse, of 300, 600, and 900 ml of
solution was added at each flow rate to replicate soil columns.
When the desired throughput volume had been introduced, the flow process
was stopped and saturated soil samples were taken immediately at 2 cm
intervals along the soil columns. The samples were obtained from pre-
viously stoppered openings along the column (Fig. 1). These samples
were used to determine the fluometuron solution and adsorbed concentra-
tion distribution at the time of sampling. Separation of the solution
and adsorbed phases was accomplished by centrifuging the samples at
25 + 1.0 C immediately after sampling using specially constructed
fritted-glass filters (Fig. 1) at 1205 x G for 5 min. The solution
14
sample obtained by this procedure was analyzed for C activity and
assumed to represent the solution concentration (yg/cm ).
Extraction of the remaining herbicide from the soil was made using
three successive 5 ml leachings with 95% ethyl alcohol. Preliminary
studies revealed that an insignificant amount of fluometuron remained
in the soil sample after three leachings. Since the initial centrifuga-
tion did not remove all the water from the sample, the fluometuron in
the remaining solution had to be subtracted from the total concentration
measured in the alcohol. This was determined by measuring the soil
water content and knowing the solution concentration. The adsorbed con-
centration was expressed as (yg/g). It was assumed that the sampling
and centrifugation were carried out rapidly enough to provide a true
measure of the concentration of each phase.
14
Analyses of the soil solution and alcohol leachate for C activity
were made using a liquid scintillation technique. Duplicate 0.5 ml
23
-------
aliquots were pipetted into separate scintillation vials containing
15 ml of scintillation solution. The scintillation solution was com-
posed of 120 g of napthalene, 4 g of 2,5-diphenyloxazole (PPO) , and
50 mg of l,4-bis-2- (5-phenyloxazolyl)-benzene (POPOP) made to 1 liter
with p-dioxane. Preliminary studies revealed no difference in quenching
due to varying the amount of ethyl alcohol or water in the sample ali-
quot. C activity was used as a direct measure of the herbicide con-
centration.
In several studies, C-labeled herbicide was displaced through the
14
soil column and effluent samples analyzed for C activity. Five
milliters effluent water samples for herbicide analysis were collected
sequentially with a fraction collector. The relative concentration
(C/CQ) was calculated from the herbicide concentration in the effluent
(C) and the original concentration (CQ) of the herbicide at the inflow
location. At the completion of each herbicide displacement, the volume
of water (Vo) held in the soil was determined gravimetrically. The
number of pore volumes (V/VO) displaced through the soil was calculated
by dividing the volume of effluent (V) by the volumetric water capacity
(VQ) of the soil.
Equilibrium adsorption-desorption isotherms were obtained by shaking
duplicate samples of soil and herbicide solution (10 g soil to 10 ml
solution) in screw-capped glass vials for 8 hours at 25 + 0.5 C. The
vials were centrifuge d at 1250 x G for 5 min. at 25 + 1.0 C and dupli-
cate 0.5 ml aliquots of the supernatant were removed for analysis by
liquid scintillation. For the desorption isotherm, 1.0 ml of 0.01 N
CaS04 was added to replace the 1.0 ml sample removed for 14C analysis.
This dilution of the herbicide solution resulted in a desorption of the
material from the soil matrix. The procedure was repeated to achieve
successive dilutions of the herbicide solution. Blanks were run under
identical conditions (soil deleted). Differences between blanks and
samples containing soil were considered as the amount adsorbed or
desorbed. All samples were run in duplicate.
24
-------
Transient Soil Water Studies
The soil used in this study was obtained from the top 15 on of a pro-
file classified as Cobb fine sandy loam. The sampling site was located
on the Caddo Research Station near Fort Cobb, Oklahoma. Based on the
particle size distribution and for convenience, the sampled soil will
be referred to as Cobb sand in this report. The soil was air-dried and
passed through a 2.0 mm sieve. Gravimetric water content of the air-
dry soil was 0.005 g/g. The pH, organic matter content and cation
exchange capacity of the soil were 7.0, 0.5%, and 3.9 meq/100 g,
respectively. The soil had 91.8% sand, 6.0% silt and 2.2% clay.
The air-dry soil was packed into rectangular acrylic columns one meter
in length and having 13 by 13 cm inside dimensions. Soil was added to
the column in 2 on increments with each layer stirred into the top of
the previous layer. After each soil addition, each side of the column
was tapped four times with a rubber-faced mallet. This procedure was
repeated until the total depth of soil reached 95 cm. The average bulk
density of the soil in each column was 1.53 + 0.015 g/cm .
The volumetric water content and initial bulk density at various loca-
tions along the length of the soil was measured by gamma-ray attenuation.
The apparatus, Figure 2, consisted of a 250 millicurie Cesium-137 source,
thallium-activated Nal crystal scintillation detector (Harshaw Type 4S4) ,
and the following Harshaw electronic equipment: preamplifier (Model
NB-11), linear amplifier (Model NA-11) single channel pulse height ana-
lyzer (Model NC-11), sealer (Model NS-30), timer (Model NT-29), and high
voltage supply (Model NV-19). The sealer was coupled to a Hewlett
Packard 505OB digital recorder. The system was found to have a resolving
time and mass adsorption coefficients for water and soil of 3.3 micro-
seconds, 0.0855 cm /g and 0.0797 an2/gf respectively.
A fluometuron concentration of 1.83 yg/ml in absolute ethanol was
14
obtained by combining 80% wettable powder (technical grade) and C-
labeled fluometuron (100 yc/9.7 mg) in proper proportions to yield a
activity of 0.556 yc/ml of solution. Three milliliters of this
25
-------
Pb COLLIMATORS
s
0
I
L
C
0
L
U
M
HIGH VOLTAGE
SIGNAL
HIGH VOLTAGE
SUPPLY
PRE-
AMPLIFIER
LINEAR
AMPLIFIER
PULSE HEIGHT
ANALYZER
SCALER
TIMER
RECORDER
Figure 2. Schesnatic of gaititia-ray attenuation apparatus.
26
-------
solution were applied unifonnly to the soil surface prior to initiating
infiltration. This was equivalent to an application rate of 3.24 kg/ha.
The soil surface was divided into three equal areas and the solution
containing the fluometuron was added dropwise at random to each area
with a 1 ml pipette. The ethanol was allowed to evaporate prior to the
application of water.
After application of the herbicide, a 2 cm layer of 0.5-1.0 mm diameter
quartz sand was placed on the soil surface. This was done to achieve a
uniform distribution of water at the soil surface with the low infiltra-
tion rates and prevent puddling of the soil surface.
Constant infiltration rates of 1 or 5 cm/hr of 0.01 N CaS04 solution
were obtained with a constant volume pump. The pump supplied water to a
manifold with thirteen outlets. Each outlet was connected to a two-inch
length of capillary tubing mounted in a 20 cm square acrylic plate,
Figure 3. The capillary tubing had an inside diameter of 0.5 + 0.25 mm.
The plate was designed to fit on top of the acrylic columns containing
the soil. The plate was moved over the soil surface from time to time
to achieve complete coverage of the soil area. In order to monitor the
rate of water addition, the 0.01 N CaSC>4 solution source was pumped from
a flask positioned on a Mettler model P3 balance.
A one centimeter head of water was maintained on the soil surface when
the infiltration rate was not controlled. Two four-liter Erlenraeyer
flasks containing 0.01 N CaSO^ solution were used to maintain this head.
These flasks were mounted on platforms attached to the elevator of the
gamma-ray attenuation apparatus. The flasks were weighed before and
after infiltration to determine the total quantity of water that had
entered the soil.
Samples of soil water were collected at various soil depths during
infiltration through 10 mm fine-porosity fritted-glass immersion tubes.
The tubes were located in the sides of the acrylic container beginning
5 cm below the soil surface and extending to 75 cm in 5 on increments,
27
-------
SOIL
CAPILLARY TUBING
IOMMO.D. FINE POROSITY
FRITTED GLASS DISC
Figure 3. Soil column and water application apparatus used in transient
soil water studies.
28
-------
Figure 3. An additional tube was placed 2 on below the soil surface.
Rubber septums mounted on the open ends of the inmersion tubes allowed
soil solution samples to be drawn through the fritted discs. Glass
syringes were used to draw samples from the soil after the wetting
front had passed a given sampler. Herbicide concentration (*• C activity)
in each soil water sample was measured by liquid scintillation.
Immediately after cessation of infiltration, the column of soil was
removed from the elevator associated with the gamma-ray attenuation
apparatus. One side of the acrylic column was removed and the soil was
sampled at three centimeter intervals beginning at the wetted front.
The solution and adsorbed herbicide concentration in each core sample
was determined using the procedure described in the steady state soil
water studies section.
Soil moisture characteristics for Oobb sand packed to a density equal to
that used in the soil columns were determined for both wetting and drying
cycles. Soil cores 7.62 cm in diameter were placed on fritted glass
plates in Buechner funnels. The soil was saturated for 24 hours and then
allowed to drain to an equilibrium water content at a pressure of -4 cm
of water. By increasing the air pressure in the Buechner funnels in given
increments and measuring the quantity of water drained from the soil
between these increments, a soil-water content pressure head relationship
for the drainage cycle was obtained. When the soil reached equilibrium
at the last pressure increment, a constant head burette was connected to
the out flow end of the system. The pressure was then decreased by given
increments and the amount of water flowing out of the burette and into
the soil was measured. In this way, a soil-water content-pressure head
relationship was determined for a wetting cycle. The pressure at which
wetting was initiated was varied in order to obtain several soil water
characteristic scanning curves for wetting.
Soil water diffusivities were determined with the method outlined by
Bruce and Klute26. Water was applied to air-dry soil packed into a 3.1
cm diameter acrylic column. The pressure at the inflow end of the
column was maintained at -2 cm. At the end of infiltration the column
29
-------
was sectioned into one centimeter segments and the water content of each
segment was measured gravimetrically.
PTRT.n EXPERIMENTS
Runoff experiments were conducted in the field on a Teller sandy loam
near Perkins, Oklahoma. The pH, cation exchange capacity, organic matter
and clay content were 6.6, 7.3 meg/100 g, 1.2% and 20%, respectively.
The average surface slope of the experimental area was 1%. Prior to
construction of each runoff experimental plot, the area was plowed 15 cm
deep and worked to obtain a good tilth. Cotton (Gossypium hirsutum L.
"Westburn 70') was seeded in rows 1.2 m apart the day the plots were
constructed.
Individual runoff plots were 1.52 x 4.57 m in size with 1.52 m between
plots. Each plot was arranged so that a row of cotton and its corres-
ponding wheel track ran the full length of the center of the plot. Each
plot was bordered with stainless steel edging that extended 12 on below
the soil surface. A sheet metal catchment device with an opening
extending the full width of the plot was installed at the lower end of
each plot. The opening of the catchment was hooded to prevent direct
entry of rainfall. Runoff was transferred from the catchment to a
stainless steel barrel for collection and sampling. After installation
of the border and catchment, the plots were smoothed around the inside
border. The planter furrow and wheel track in the plot were left intact.
All field runoff experiments were repeated, but because of minor climatic
variations that occurred between simulated rainfall events they were not
considered replications.
Commercial fluometuron at 2.8 kg/ha was applied to the plots with a
calibrated tractor sprayer in a water carrier volume of 374 I/ha. The
spray boom extended across the plot, but the tractor was driven outside
the plot. Immediately after the herbicide application, soil samples
were taken from within the plot for initial herbicide concentration
determinations. Soil samples for fluometuron content were taken from
the 0-5 on depth.
30
-------
Rainfall was simulated with an oscillating lawn sprinkler that applied
1.2 cm of water per hour when operated continuously. This specific
sprinkler was chosen for its uniformity of coverage and output.
In all runoff experiments, water was applied until a total of 45.5 L
(approximately h acre-inch) of runoff was obtained. A representative
water sample was collected from the first 3.8 L of runoff, and another
was collected from a composite of the next 41.6 L of runoff after
thoroughly suspending the sediment. The amount of sediment in this
sample was used to compute the total sediment loss from the plot. A
sediment sample for fluometuron analysis was obtained by flocculating the
suspended material in the composite runoff sample. All runoff samples
were frozen until analysis.
The effect of rainfall intensity was studied by varying the time of
water application. Three intensities were simulated - fast, intermedi-
ate, and slow: Fast, the water was applied at 1.2 cm/hr until runoff
samples were obtained; intermediate, 0.6 cm increments of water were
added over a three-hour period until 2.5 cm was applied, and then 1.2
cm/hr of water was applied until all runoff samples were obtained; slow
0.3 cm increments of water were added over a 4-day period until a total
of 6.3 cm had been applied. No lateral runoff occurred from the slow
treatment. Two days after each water application treatment, samples
from the 0-5, 5-10, and 10-20 cm soil depths were taken from each plot
to determine the extent of vertical fluometuron movement.
The effect of antecedent soil moisture on fluometuron movement was
studied using "Vet" and "dry" plots. The plot area was dried by tillage
and then those plots designated dry were covered while those designated
wet received 4.5 cm of simulated rainfall and then were allowed to dry
for 24 hours. Fluometuron was then applied to all plots at a 2.2 kg/ha
rate. Simulated rainfall (1.2 cm/hr) was applied until runoff samples
were obtained. Soil samples were taken as described above.
31
-------
Dissipation and vertical movement of fluometuron during the growing
season was studied by sampling the soil in the rainfall intensity study
at three dates over an 84-day period. The first sampling date was
immediately subsequent to the first simulated runoff study. The sam-
pling depths were 0-5, 5-10, 10-20, and 20-30 cm.
The chemical analytical method for fluometuron in soil and water was
supplied by Ciba-Geigy, manufacturers of the herbicide, Bioassays were
also conducted on the soil samples and a few selected water samples as
an alternate analysis method. Soil samples were thawed, air dried, and
screened to pass a 1 mm screen prior to analysis. Water samples were
thawed and filtered.
For chemical assay, the fluometuron was hydrolyzed to 3-trifluoro-methyl
aniline and extracted by steam distillation with isooctane. The extrac-
ed aniline was diazolized and coupled with N-ethyl-1-napthylamine to
develop the color. The concentration of the herbicide was determined
spectrophotometrically. The procedure was modified to eliminate the use
of a clean-up column since the soil and water samples in this study did
not produce a strong background color. The background from the untreated
samples was substracted from the sample readings and the cleanup procedure
was not needed. The basic procedure was as follows: using a 25 g soil
sample or a 100 ml water sample, the fluometuron was extracted by dis-
tillation with isooctane, cooled, and the isooctane was extracted with
0.25 N HC1. Three extractions with 10 ml 1.0 N HC1 were used and the
final volume adjusted to 50 ml with.1.0 N HC1. The color was developed
as outlined, and the colored HC1 solution was extracted with a 20 ml N-
butanol. The colored butanol extract was transferred to a test tube, and
10 g of anhydrous sodium sulfate granules were added. The colored
extracts were quantified spectrophotometrically at 525 my.
Standards were prepared by spiking samples of water obtained from check
plots with analytical fluometuron. The standard curve range was from
0.01 yg/ml to 0.6 yg/ml. Water samples and standards were read in a
4 cm lightpath cuvette. Soil standards for the chemical and bioassay
32
-------
analysis were prepared in a similar manner. The herbicide was added to
soil in a volume of water sufficient to bring the soil to 10% water by
weight. The samples were mixed, air dried and screened through a 1 inn
screen. The standard curve was from 0.1 to 6.0 yg/g. Soil samples were
read in a 1 cm cuvette. To determine the precision of the method, stand-
ard deviations were determined for six replicate samples at three dif-
ferent concentrations of fluometuron in soil and water. The absorbance
values for fluometuron extracted from water or soil showed a linear
response over the range of fluometuron concentrations found in all exper-
iments. For the water analysis, the standard deviations were + 0.002 at
0.05 yg/ml, + 0.005 at 0.1 yg/ml, and + 0.03 at 0.4 yg/ml. For the
soil analysis, the standard deviations were < + 0.01 at 0.04 g/g, +
0.07 at 2 yg/g and + 0.45 at 6 yg/g. Coefficients of variation were
between 19 and 21% for all bioassay experiments.
Soil bioassays, using oats (Avena sativa L. 'Ciammaron'), were conducted
in styrofoam cups. The soils were thoroughly mixed and 225 g of soil
placed in each cup, planted, and subirrigated. The cups were placed
under lights in a randomized block design with 4 replications. After
14 days, fresh weights of the oat tops were determined. The data were
converted to yg/g using a standard curve. The standards were prepared
as previously described. Water bioassays for fluometuron utilizing
?7
cucumber leaf disks were used as described by Truelove .
33
-------
SECTION VI
RESULTS AND DISCUSSION
The numerical solution of equation (14) using the finite difference
scheme described in equation (16) is compared with an exact or analyti-
cal solution (Davidson, et al.28) in Figure 4 for steady state soil
water conditions. The initial and boundary conditions are described
in equation (24a). The adsorption-desorption process was assumed to be
linear and a single valued function, with KA and KD equal to 2.0 cnr/g.
The concentration at z = 0 was maintained at Co = 90 yg/cm for 2.0
hours after which time it was reduced to zero. All other parameters
necessary for the solution of equation (14) are given in Figure 4. The
numerical dispersion associated with equation (16) for steady state soil
water conditions was corrected for using the DJJ term from equation (20).
The agreement between the exact and numerical solution of equation (14)
is reasonable (Figure 4). The general shape and position of the two
relative concentration distributions are in good agreement; however, the
numerical solution of equation (14) resulted in approximately 5% more
area under the calculated values than under the exact solution (solid
line). The correction for numerical dispersion was better when KA was
equal to zero, and agreed with the results reported by Bresler13 for non-
interacting solutes. An adsorption distribution coefficient, KA, of
2.0 cnr/9 is large compared to that frequently measured in soil applied
pesticide systems where the mobility of the pesticide presents a major
or significant environmental problem. Thus, corrections for numerical
dispersion using the relations in equation (20) for steady state soil
water conditions should be adequate for those systems involving more
mobile solutes (KA < 2.0
34
-------
RELATIVE SOLUTION CONC. (C/C0)
0
o
OL
LU
Q
O
CO
10
0.2 0.4 0.6 0.8 KO 1.2
D
e
K,
EXACT
o NUMERICAL
= 0.2cm2/hr
= 0.34 cmVcm3
, = 2.0
Figure 4. Computed relative solution concentration distribution using an
exact and finite difference solution of equation (13).
Calculations were made using an average pore-water velocity,
dispersion coefficient, distribution coefficient and soil
water content of 4.0 cm/hr, 0.2 on2/hr, 2.0 and 0.34 crc^/cm ,
respectively.
35
-------
The use of the numerical dispersion correction relations from equation
(22) in equation (21) produces a very complex computer program. Figure
(5) shows the agreement with G included (solid line) or excluded (open
circles) from the simultaneous solution of equations (2) and (13) with
KA = 0. The Initial soil water content was 0.12 cm3/cm3 and the pressure
head at the soil surface (z = 0) was zero throughout the displacement.
If KA is not equal to zero, then the solute pulse is retarded and appears
closer to the soil surface where the 3e/9z, 326/9z2, 3q/9z and 82q/3z2
terms are more nearly equal to zero. The fact that the calculations for
G included and G excluded agree so well for KA = 0 under transient con-
ditions (Figure 5) suggests that the term could be eliminated from the
program for K^ > 0.
LABORATORY EXPERIMENTS
Steady State Soil Water Studies
Equilibrium adsorption and desorption isotherms for fluometuron and
Norge loam soil are shown in Figure 6. Both processes are described
reasonably well by the Freundlich equation (Equation 10) . Ml lines
are a least-squares fit with the correlation coefficient exceeding 0.98
for all lines. The adsorption isotherm is described by the equation
given in Figure 6. Adsorption and desorption did not give a single-
valued function. Similar results for adsorption and desorption have
been reported by Swanson and Dutt^ and van Genuchten et al.-^
Adsorption was 90% complete after 40 min of contact and equilibrium
conditions existed after 120 min. Using the parameters determined for
equilibrium adsorption, Figure 2, and the amount of herbicide adsorbed
between 10 and 40 min, the desorption rate coefficient, kD, was calcu-.
lated (Equation 7) for each initial solution concentration studied. The
average calculated rate coefficient was 3.5 hour~l and did not appear
to be concentration dependent. Desorption rate coefficients were not
calculated for times greater than 40 min owing to the difficulty of
measuring small changes in adsorbed and solution concentrations with
time.
36
-------
1.0
o>
o
z
o
o
o
o:
ID
I-
LU
^
O
ID
0.
0.01
Q
UJ
DQ
o:
o
S o.ooi
< 0.01
EQUILIBRIUM
0.8
°04
U-t
0 1.0 2.0
C max (//g/cm3)
O.I 1.0 10
SOLUTION CONG (//g/cm3)
Figure 6.
Equilibrium adsorption and desorption isotherms for fluomet-
oron and Norge loam. Insert shows relationship between the
desorption distribution coefficient, KD, and the maximum
fluometuron solution concentration,
38
-------
The dependence of the desorption distribution coefficient upon the
maximum fluometuron solution concentration is shown in Figure 6
(insert) . The line was calculated using equation (12) and a N and N1
of 1.02 and 2.30, respectively. The data points are the measured dis-
tribution coefficients and C^x from the desorption lines shown in
Figure 6. The agreement is well within experimental error and justifies
the use of a constant N1 rather than the measured value that increases
with a decrease in
F.arlier studies using effluent data have shown that equilibrium adsorp-
tion may not exist at high soil-water flow rates 5 9. Figure 7 gives
the relative solution and adsorbed fluometuron concentrations at various
locations in the soil following the introduction of 200 ml of fluomet-
uron solution plus 100 ml of 0.01 N CaSC>4 solution ( a total of
approximately 0.64 pore volume) at two flow rates. The right hand por-
tion of each "loop", for each flow rate, was obtained from the invading
fluometuron front. The left hand part of the "loop" comes from the
tailing edge or after a maximum concentration has moved past a given
soil location. The presence of a nonsingular relationship between
adsorption and desorption is shown by the fact that the adsorption and
desorption data form a loop rather than a single- valued line. Also,
the dependence of the adsorption upon the flew rate is illustrated by
the right-hand shift of the fast flow rate data as well as the lower
absorbed concentrations for a given solution concentration when compared
with the slower displacement.
The adsorption data (Figure 7) from the slow displacement is similar to
the equilibrium adsorption isotherm presented in Figure 6. Thus, for an
average pore-water velocity of 0.59 cm/hour, equilibrium adsorption
appears to exist between fluometuron and the Norge soil. For this case,
the adsorbed and solution phases move together whereas as the high flow
rate, the solution phase (peak concentration) will precede the adsorbed
phase during displacement. Therefore, a greater spreading or flatter
pQ
concentration distribution will exist .
39
-------
1.2
FLUOMETURON IN NORGE LOAM
TRANSIENT SORPTION
ISOTHERMS
/V0 — 0. 59 cm/hr
_L
_L
_L
0.2 0.4 0.6 0.8 1.0
RELATIVE SOLUTION CONCENTRATION (C/Co)
Figure 7. Transient adsorption-desorption behavior of fluoneturon in
Norge loam soil for 0.59 and 5.5 on/hr average pore water
velocity.
40
-------
The relative solution and adsorbed concentration at 2 cm increments
along the soil column for three displacement periods are given in
Figures 8 and 9. The solid lines were obtained from the numerical solu-
tion of equation (13) using equation (10) for adsorption and desorption.
The values of q/0, K_, N, N1, and D were 0.59 cm/hr, 0.4, 1.02, 2.30,
-> H
and 0.06 air/ hour, respectively. Other physical data used to calculate
the concentration distributions are given in Table 2.
Table 2. EXPERIMENTAL PARAMETERS FOR SOIL COLUMN STUDIED
Total
throughput
volume
cm3
300
600
900
300
600
900
Bulk
density
g/an (p)
1.50
1.50
1.50
1.50
1.51
1.52
Soil-
water
content
on3/cm3 (0)
0.396
0.383
0.373
0.347
0.382
0.378
Solution
flux
cm/hour
2.10
2.10
2.14
0.22
0.22
0.22
Pore^water Total
velocity time
cm/hour (v0) hour
5.30 3.15
5.49 6.30
5.73 9.30
0.63 30.1
0.57 60.7
0.58 91.0
Pore
volume
on3
539
520
507
511
520
514
The dispersion coefficient, D, for fluometuron was assumed equal to that
for chloride and was determined from chloride data (used the model pro-
posed by Hashimoto et al.-^). Reasonably good agreement was obtained
between experimental and calculated relative solution and adsorbed con-
centration distributions (Figures 8 and 9). Note that the position of
the peak concentration in Figures 8 and 9 occurs at the same location for
each throughput volume indicating that the average pore-^water velocity
was low enough to permit equilibrium adsorption and desorption. Also,
the tailing edge of the relative solution concentration is well de-
scribed by the calculated line indicating that the desorption process
was adequately defined by the model. Variations between measured and
calculated values can be attributed, partially, to experimental error.
41
-------
O
8
o
CO
1.0
0.8
0.6
0.4
0.2
LJ
QL 0.
FLUOMETURON IN NORGE LOAM
-e 0.59 cm/hr
KA = 0.4
CALCULATED LINES
5 15 25
DISTANCE ALONG COLUMN (cm)
Figure 8. Calculated and experimental relative solution concentration
distribution for 0.59 cm/hr average pore-water velocity.
Calculated lines were obtained from a numerical solution of
equation (14).
42
-------
FLUOMETURON IN NORGE LOAM
V =» 0.59 cm/hr
/ 0
K = 0.4
CALCULATED LINES
5
DISTANCE
ALONG
COLUMN
Figure 9. Calculated and experimental relative adsorbed concentration
distribution for 0.59 cm/hr average pore-water velocity.
Calculated lines were obtained from a numerical solution of
equation (14).
43
-------
Selecting a larger (2.8) or smaller (1.8) N1 value did not significantly
change the fit of the line to the data except in the maximum concentra-
tion region. Higher values of N1 moved the peak concentration to the
right and lowered the maximum concentration. The lower value of N'
produced the opposite results.
Increasing the average pore-water velocity tenfold, while holding the
throughput volumes constant, gives the solution and adsorbed concentra-
tion distributions shown in Figures 10 and 11. At this pore-water
velocity (5.5 on/hour), equilibrium adsorption does not exist at all
locations within the soil mass (Figure 7). Therefore, the kinetics of
the adsorption-desorption process would appear to be beneficial. How-
ever, the work of van Genuchten-^ with various kinetic models does not
support this conclusion. Using the same adsorption-desorption parameters
as those used for the slow flow rate (Table 2) and the numerical solu-
tion to equation (13), the relative solution and adsorbed fluometuron
concentration distributions within the soil were calculated (Figures
10 and 11).
The maximum measured relative solution concentration occurs to the
right of the measured relative adsorbed concentration peak. This
illustrated the lack of equilibrium adsorption during the fast displace-
ment of fluometuron through the soil. Also, the relative solution and
adsorbed concentration distributions are flatter than those shown in
Figures 8 and 9 for the slower average pore^water velocity. Spreading
of the concentration distribution owing to nonequilibrium adsorption
is inversely proportional to the adsorption or desorption kinetic rate
coefficient for a given displacement rate^. Spreading is inversely
proportional to the flow rate for constant rate coefficients .
The data in Figures 10 and 11 were not well described by the numerical
solution of equation (13) with equilibrium adsorption. The data were
consistently to the right of the calculated invading front. The kinetic
adsorption and desorption models assisted in describing the nonequilib-
rium adsorption condition, but were not significantly better than the
44
-------
o
8
O
o
o
h-
-J
o
CO
LJ
I
LL)
a:
i.o
0.8
0.6
0.4
0.2
FLUOMETURON IN NORGE LOAM
% = 5.50 cm/hr
KA = 0.4
CALCULATED LINES
0 5
DISTANCE
15 25
ALONG COLUMN (cm)
Figure 10. Calculated and experimental relative solution concentration
distribution for 5.5 cm/hr average pore-water velocity.
Calculated lines were obtained from a numerical solution of
equation (14).
45
-------
FLUOMETURON IN NORGE LOAM
5.50 cm/hr
KA = 0.4
CALCULATED LINES
cc
Figure 11. Calculated and experimental relative adsorbed concentration
distribution for 5.5 cm/hr average pore-^water velocity.
Calculated lines were obtained from a numerical solution of
equation (14).
-------
7 10
equilibrium model . The tailing produced by the non-singular char-
acteristics of the adsorption-desorption process was described by the
concentration dependent desorption distribution coefficient. One
explanation for the right-hand shift of the data compared to the calcu-
lated line at the high flow rate is that the total surface area avail-
able to the herbicide during displacement may be less. If the mixing
process was not complete during a herbicide's displacement, and estimate
of the adsorption characteristics for the surface area contacted would
be required. Similar results were obtained by van Genuchten et al.
for 4-amino-3,5,6-trichloropicolinic acid (picloram) effluent concentra-
tion distributions at high flow rates. They describe an empirical rela-
tionship for defining the fraction of the pore volume or soil surface
area that was actively participating in adsorption and desorption at
various flow rates. The procedure improved their ability to describe
picloram concentration distributions in the effluent over a similar
range in pore-water velocities to those used in this study.
The relative fluometuron concentration in the effluent from a water
saturated column of Nbrge loam soil flowing at two different average
pore water velocities is presented in Figure 12. These results illus-
trate the influence of flow rate on the adsorption-desorption process.
The solid lines were calculated with an exact solution of equation
(13) . A kinetic adsorption model (equation [7]) was used and D was
assumed to be negligible. The calculated lines are for the lower pore-
water velocity. The distribution coefficient (6k,/pkD) for f luometuron
and Norge loam was 0.31. The parameters associated with each calculated
curve are the ratio between the kinetic desorption rate coefficient kD
and average pore-water velocity vo. The kinetic desorption rate coef-
ficient used for the three curves in Figure 12 from left to right were:
kD = 0.10, 0.55, and 5.5 hr"1, respectively.
The shapes of the calculated curves in Figure 12 are a characteristic
of k-Vv , and not the magnitude of the individual values comprising the
ratio. For example, the same lines were calculated when the average
pore-water velocity was 5.6 cm/hr and the kD values were 1.0, 5.5 and
47
-------
NORGE LOAM
= 5.6 I cm/hr
= 0.56 cm/hr
CALCULATED
2.0
VOLUME,
V/V<
Figure 12. Calculated and experimental relative effluent concentration
distribution from Norge loam soil. The calculated values
were obtained using an exact solution of equation (13)4 and
various ki>/vo values for the slow average pore-water velocity.
48
-------
55 hr~l, respectively. Note that the calculated effluent concentration
distribution for a higher velocity (same kD) than 0.56 cm/hr would be
flatter and shifted to the left of that for a lower velocity. This is
not the case for the measured effluent concentration distribution
obtained from the high flow rate study, it has a higher maximum relative
concentration as well as being shifted to the left. Figure 12 as well
as the study by van Genuchten et al. clearly illustrates the fact that
the kinetic adsorption model does not adequately describe the mobility
of a pesticide at high pore water velocities.
It should be pointed out that the kinetic desorption rate coefficient
k^, unless measured independently, could be selected on the basis of
data fitting and thus erroneously include the spreading owing to dis-
persion. A major shortcoming of the analytical solution of equation
(13)4 is its inability to handle the nonsingular adsorption-desorption
function and hydrodynamic dispersion.
In conclusion, the distributions of the adsorbed and solution phases of
fluometuron were well described for a low flow rate using the numerical
solution of equation (13) and the specific adsorption and desorption
characteristics of the herbicide. The desorption distribution coeffi-
cient, KD, was shown to be a function of the maximum amount of herbicide
in solution prior to desorption. The movement of fluometuron at high
flow rates was only approximated with kinetic adsorption and desorption
parameters. At low flow rates where equilibrium adsorption exists, the
kinetics of the adsorption process were not important. At high flow
rates, the mobility of the herbicide is significantly influenced by a
flow-rate dependent process. An empirical model presented by van
Genuchten et al.10 described the movement of picloram at high flow
rates. The procedure defines a fraction of the adsorbing sites that
was in equilibrium with the picloram during displacement. This fraction
was a function of the average pore-water velocity. The procedure pro-
vided a reasonable fit between data and calculated effluent concentra-
tion distributions for a wide range of pore-water velocities. The dis-
persion coefficient must also be included in the model when evaluating
49
-------
solute transport at slow flow rates.
Transient Soil Water Study
The equilibrium adsorption and desorption isotherms for fluometuron and
Cobb sand given in Figure 13 were described, by the Freundlich equation
(equation [10]) . As can be seen in Figure 13, adsorption and desorp-
tion were not described by a single-valued function. For adsorption,
the values of KA and 1/N were 0.21 cm3/g and 0.84, respectively. For
desorption KD and 1/N1 were dependent upon the maximum amount of herbi-
cide adsorbed. Table 3 gives the KD and 1/N' values for each C^.^
studied. A reaction tine of 30 min was sufficient for equilibrium ad-
sorption; however, 8 hr was used to obtain all adsorption and desorption
data.
Table 3. PARAMETERS ASSOCIATED WITH DESORPTION OF
FLUOMETURON FROM COBB SAND AT 25 + 0.5 C.
'--max
yg/cm
0.42
1.26
2.12
7.82
15.8
24.1
32.3
40.4
KD
0.17
0.22
0.25
0.25
0.34
0.48
0.34
0.36
1/N1
0.60
0.64
0.63
0.63
0.66
0.57
0.68
0.69
The N' in equation (11) was a function of C^gx (Table 3) , but for
this study was assumed constant. Using equation (12), the concentration
dependent desorption distribution coefficient, KD, was easily calculated
50
-------
0.01
O.I
SOLUTION CONC
Figure 13. Adsorption and desorption isotherms for fluometuron on Cohb
sand. Solid and broken lines are best fit for adsorption
and desorption, respectively.
51
-------
in the numerical solution. From preliminary desorption isotherm data at
low herbicide concentrations, the average of the measured N" values was
1.7. This value was used to calculate the desorption distribution coef-
ficient in the numerical solution. When all the desorption data over
the concentration range given in Table 1 were collected the average N?
was 1.5.
The fluometuron solution concentration with time at various soil depths
during infiltration into Oobb sand is shown in Figure 14. The initial
soil water content was 0.13 cm /orr and the water flux at the soil sur-
face was 1.03 cm/hr. These concentrations were determined by collecting
samples of the soil-water through fritted filter discs during infiltra-
tion. The reduction in peak height and the increased spreading with
depth shown in Figure 14 are a result of mixing by velocity dispersion
and adsorption of the fluometuron in the soil. The tailing is an
indication of the nonsingular relationship between adsorption and de-
sorption.
Figures 15 and 16 show the adsorbed and solution fluometuron concentra-
tion distributions in Cobb sand. A one centimeter head of water was
maintained on the soil surface throughout the infiltration process.
These distributions were measured by taking samples from the soil col-
umns immediately after the cessation of infiltration. The maximum
adsorbed and solution concentrations generally occurred at approximately
the same depth for all treatments. However, some lagging of the
adsorbed phase is shown for the ponded infiltration into initially air-
dry soil, Figure 15. This suggests that nonequilibrium adsorption con-
ditions did not exist at the fast pore-water velocities associated with
this column. The average filtration rate was 29.0 cm/hr for the
initially dry soil and 10 cm/hr for the initially wet soil.
The initial soil-water content prior to infiltration had little effect
on the displacement of fluometuron for a given quantity of infiltrated
water. This is illustrated in Figure 17 where fluometuron concentra-
tions are compared for initial soil-water contents of 0.005 (air-dry)
52
-------
SOLUTION
.0 2
CONC.
4
(//g/cm3)
6 8
600-
Figure 14. Fluometuron concentration distributions versus time for
selected soil depths. Ihe soil water flux and initial
soil water content were 1.03 cm/hr and 0.13 cm /cm .
Solid lines were eye fitted to experimental data.
53
-------
SOLUTION CONC. (//g/cm3)
0246
= 0.005 cm3/cm3
FLUOMETURON
0246
ADSORBED CONC. (//g/g)
Figure 15. Solution and adsorbed fluometuron concentration distribu-
tions inmediately following infiltration. Average soil-
water flux was 29.0 cm/hr. The 9.^ and 6f are the initial
and final volumetric soil water content in the soil surface
region. Solid lines were eye fitted to experimental data.
54
-------
SOLUTION
0 2
r
£ 20
o
X
I-
Q- 40
LU
Q
O
CO
80
CONC.
4
6;= 0.130 cm3/cm3
FLUOMETURON
0 2
ADSORBED
4
CONC.
Figure 16. Solution and adsorbed fluometuron concentration distribu-
tions inmediately following infiltration. Average soil-
water flux was 10.0 on/hr. The 6j_ and 8^ are the initial
and final volumetric soil water content in the soil surface
region. Solid and dashed lines were eye fitted to experi-
mental data.
55
-------
SOLUTION CONC (yu.g/cin3)
2
£
o
Q_
UJ
O
O
CO
r 1
O • FLUOMETURON
A A WATER
60
80
\Qi = O.I30cm3/cm3
[0f = 0.34cm3/cm3
r
0.005 cmVcm3
0.34cm3/cm3
0 0.16 0.32 0.48
SOIL-WATER CONTENT (cmVcm3)
Figure 17. Fluomsturon solution concentration and water content distri-
butions for the same accumulative infiltration into an ini-
tially wet and dry soil. The 9^ and 0f are the initial and
final volumetric soil water content in the soil surface
region. Solid and dashed lines were eye fitted to experimen-
tal data.
56
-------
and 0.130 on /cm . One centimeter of water was maintained on the soil
surface of both columns throughout the infiltration process. The fluo-
meturon displacement depths were relatively independent of the initial
soil-water content, whereas the wetting front position was related to
the initial soil water content. The fluometuron appears to move slightly
deeper into the initially wet soil compared to the dry soil for the same
amount of water. This was only true for experimental data and not the
calculated distributions.
Figure 18 shows the influence of the water application rate and assoc-
iated boundary conditions on the displacement of fluometuron. The
cumulative infiltration was the same for both columns. The inverse
relationship between leaching efficiency and surface water content shown
here has been reported by several investigators (Keller and Alfaro30,
Warrick et al.23, and Kirda et al.14). However, this inverse relation-
ship was not necessarily valid when comparing fluometuron displacement
for ponded and 5.0 cm/hr application rates. This was a result of the
final soil-water content at the soil surface and in the transmission
zone being only slightly different for these rates. Also, the smaller
pore-water velocities during the 5.0 cm/hr application rate allowed
more time for diffusion controlled adsorption to occur. The areas under
the herbicide distribution curves in Figure 18 were not equal as a result
of some fluometuron remaining at the soil surface for the 1.0 and 5.0
cm/hr application rates. At least a portion of this residual fluometuron
appeared to be in solution. However, the validity of the measured fluo-
meturon solution concentrations at the soil surface were questionable
since some herbicide may have gone into solution as a result of the
sampling procedure.
Experimental data from this study were used to determine the usefulness
of a mathematical model for predicting herbicide displacement for tran-
sient flow conditions (Hybrid program, Appendix B). The model is similar
to the one used by Kirda et al. . The model used in this study includes
an equilibrium adsorption or sink term (equation 13). The solute trans-
port and water flow equations (equations [2] and [13]) were solved sirnul-
57
-------
(WO)
Hld3Q
II OS
Figure 18. Fluoneturon solution concentration distributions for equal
accumulative infiltration into air-dry soil. The average
soil water flux at the soil surface was 29.0 and 1,0 cm/hr-
The 0. and 0^ are the initial and final volumetric soil
water content in the soil surface region. Solid and dashed
lines were eye fitted to experimental data.
58
-------
taneously in order to predict fluometuron and water distributions during
infiltration. Equilibrium adsorption and desorption as described by
equation (10) and (11) were also used. The dispersion coefficient, D,
was assumed constant for each calculation, with a correction for numer-
ical dispersion included.
Figures 19, 20, and 21 show comparisons of calculated and experimental
distributions of fluometuron and water. The calculated fluometuron solu-
tion concentration distributions lagged behind the measured distribu-
tions in each case. However, the position of the peak concentration
and the tailing edge seem to be predicted somewhat better for the slower,
4.89 cm/hr, infiltration rate (Figure 21) than for the soil columns
on which water was ponded. These calculated distributions were obtained
using values of 0.21, 0.84, and 0.59 for KA, 1/N, and 1/N" , respectively.
The D-DN value was dependent upon the water application rate and was
0.07 cm^/hr for application rates of 1.0 and 5.0 cm/hr and 0.10 cm^hr
for columns on which water was ponded during infiltration. Equation (12)
was used to calculate values of the distribution coefficient, KD. The
velocity and soil-water content terms used in the solute transport model
were obtained from the numerical solution of the water flow equation
(equation [2]).
As shown in Figures 19, 20, and 21, the numerical solution of equation
(2) adequately described the measured soil-water content distributions.
A uniform initial soil-water content distribution was used to approxi-
mate the soil-water content distribution for each initially wet soil
column.
Van Genuchten et al.10 have suggested that at high pore-water velocities
equilibrium adsorption between all pore sizes may not exist, but that
only a fraction of the soil participates in the adsorption process. If
there was insufficient time for the fluometuron to diffuse into smaller
pores at the high pore-water velocities existing in this study, then
less adsorption than predicted would have occurred. To account for the
non-adsorbing fraction, a term similar to the FREQ term used by van
59
-------
0
SOLUTION CONC (/ig/cm3)
5 10
0 • FLUOMETURON
WATER
CALCULATED LINE
100
O.I 0.2 0.3
SOIL-WATER CONTENT (cmVcm3
Figure 19. Experimental and calculated fluometuron solution concentra-
tion and water content distributions after 15 and 59 minutes
of infiltration. Initial soil-water content was 0,005 oiP/
cm3 and the average soil water flux was 29.0 cm/hr. Solid
lines were calculated using equations (3) and (21) .
60
-------
0
SOLUTION CONC
5 10
80
100.
INITIAL
WATER
CONTENT
o • FLUOMETURON
A A WATER
CALCULATED LINE
0 O.I 0.2 0.3
SOIL-WATER CONTENT (cm3/cm3)
Figure 20. Experimental and calculated fluometuron solution concentra-
tion and water content distributions after 30 and 69 minutes
of infiltration. Initial soil water content was 0,13 on3/
on3 and the average soil-water flux was 10.1 cm/hr. Solid
lines were calculated using equations (3) and (21).
61
-------
SOLUTION CONC (yug/cm3)
.0 2.0 3.0
FLUX - 4.89 cm/hr
CALCULATED LINE
O.I 0.2 0.3
SOIL-WATER CONTENT (cmVcm3)
Figure 21. Experimental and calculated fluometuron solution concentra
tion and water content distributions after 266 minutes of
infiltration. Initial soil water content was 0.005 cm-Van
and the soil-^water flux was 4.89 cm/hr. Solid lines were
calculated using equations (3) and (21).
62
-------
Genuchten et al. was used in the calculations. The FRBQ value was
empirically determined and represented the fraction of soil participa-
tion in the adsorption process. Since the bulk density/ p, is a measure
of the mass of soil per unit volume, the FRE3Q term was multiplied by p
to give a measure of the mass of soil per unit volume which was actively
adsorbing and desorbing herbicide. The FREQ term was selected on the
basis of its ability to describe the experimental data. It should be
emphasized that a change in the p value as a result of multiplying it
by FRBQ does not indicate an actual change in the bulk density of the
soil. Rather, it is an indication of a change in the surface area which
was participating in the adsorption and desorption of fluometuron. For
i
convenience and as a first approximation, the bulk density was used as
a measure of the surface area of the soil. The model could be made more
descriptive of the physical system by the addition of a surface area
term.
Figure 22 gives the calculated fluometuron distributions for an initially
wet soil on which water was ponded during infiltration. As can be seen,
the tailing edge and the position of maximum concentration were described
very well by using a FRBQ value of 0.5, but the calculated peak concen-
tration was much larger than the measured concentration. However, the
calculated and measured maximum concentrations were approximately the
same for other columns studied3^. There also appears to be more disper-
sion at the displacing front than predicted by the model. The measured
and calculated distributions of the adsorbed fluometuron for the condi-
tions in Figure 22 are given in Figure 23. The solid line represents
the distribution predicted for a p of 1.54 g/cm3 and the broken line is
the calculated line for a p of 0.77 g/cm3 (FRBQ = 0.5). As with the
solution herbicide concentration, the model fails to describe the
measured adsorbed herbicide concentration distributions when all the
soil was assumed to be in equilibrium with the herbicide. Mien it was
assumed that only half of the soil material is actively participating in
the adsorption process (p = 0.77 g/cm3), the calculated and experimental
curves agree very well.
63
-------
SOLUTION CONC (
5
INITIAL
WATER
CONTENT
o •FLUOMETURON
A A WATER
CALCULATED LINE
O.I 0.2 0.3
SOIL-WATER CONTENT (cmVcm3)
Figure 22. Experimental and calculated flucmeturon solution concentra-
tion and water content distributions after 30 and 69 minute^
of infiltration. Initial soil water content was 0,13 on3/
on and the average soil-water flux was 10.1 cnv/hr. Solid
lines were calculated using equations (3) and (21) for p =
0.77 g/on3 (FREQ = 0.5).
64
-------
20
Q_
UJ
Q
60
ADSORBED CONC ( /j.q/g)
0.5 1.0
9i
ef
0.130 cm3/cm3
0.34 cm3/cm3
CALCULATED
LINES
Figure 23. Experimental and calculated adsorbed fluometuron concentra-
tion distribution. The Gj^ and 0^ are the initial and final
volumetric soil water content in the soil surface region.
The average soil-water flux was 10.1 cm/hr. Solid and
dashed lines were calculated for p = 1.54 and 0.77 g/cm3
(FREQ = 0.5).
65
-------
The failure of the model to describe the shape of the displacing
fluometuron front results from using a constant value for the disper-
sion coefficient and/or too low a value (equation [23]). The velocity
dependence of the dispersion coefficient has been shown by several
investigators9
In general, the mathematical model adequately described the shape and
position of the f luometuron and water distributions when a FKEQ term was
used to account for the fraction of the total surface area participating
in the adsorption process. However, further studies need to be conducted
on the influence of pore-size distribution and pore-water velocity on the
adsorption and dispersion of pesticides moving through soil. Also, the
usefulness of the mathematical model used in this study (Appendix B)
should be evaluated with additional laboratory and field data. Of par-
ticular interest would be the ability of this model to describe herbi-
cide movement for infiltration rates and associated pore-water velocities
small enough to allow greater radial diffusion of the herbicide.
FIELD EXPERIMENTS
The data for each rainfall intensity study are presented as two individ-
ual "replications". Due to differences in water amounts required to
obtain runoff, slight differences in catchment and border installation,
initial fluometuron concentrations, and a one-day interim between water
application to the two replicate plots, it was felt each plot was an
individual entity.
Rainfall intensity had no measurable effect on fluometuron concentra-
tions or total herbicide losses in the runoff (Table 4) . There was a
trend toward lower runoff concentrations and total herbicide losses
at the intermediate rainfall intensity. However, the fluometuron loss
from all plots was very low. The first 3.78 L of runoff water contained
a higher concentration than did the composite of the next 41.6 L. Appar-
ently, the fluometuron concentration in solution was decreasing with
increased volumes of water runoff. The sediment concentrations of fluo-
meturon were much higher than the concentrations in water. The fluomet-
66
-------
Table 4. THE EFFECT OF RAINFALL INTENSITY ON THE CONCENTRATION AND PERCENT OF
APPLIED FLUOMETURON IN RUNOFF WATER AND SEDIMENT
Water
Rainfall
frequency
Fast
Fast
Intermediate
Intermediate
Initial 3.78 Lb
concentration" yg/ml %
5.7 + 0.42 0.12 + 0.02
0.008
4.1 + 0.25 0.10 + 0.02
0.006
5.7 + 0.42 0.08 + 0.01
0.004
4.1 + 0.25 0.09 + 0.01
0.005
3.78-45.4 LC
yg/ml %
0.12 + 0.16
0.008
0.06 + 0.12
0.003
0.03 + 0.04
0.001
0.05 + 0.09
0.002
Sediment
yg/g %
4.1 + 0.02
0.25
2.8 + 0.01
0.13
2.9 + 0.01
0.14
2.3 + 0.01
0.009
Total
% loss
0.20
0.15
0.06
0.11
aThe concentration of fluometuron (yg/g) chemically determined to be in the top 5 cm of soil sampled
immediately after herbicide application.
concentration and percent loss of fluometuron in the first 3.78 L of runoff water from the plot.
°The concentration and percent loss of fluometuron in the next 41.6 L composite of runoff water.
-------
uron concentration ratio between the water and sediment was higher than
would be expected based upon adsorption data for the soil"?. It has
been shown by Hornsby and Davidson that adsorption and desorption of
fluometuron were not single valued. Once the soil particles were sus-
pended in the runoff, the fluoneturon evidently does not desorb back
along the adsorption isotherm. The relationship between the herbicide
and sediment, during runoff, appears to be more closely related to the
desorption than the adsorption isotherm. Although the concentration of
fluometuron was lowest in the 41.6 L composite fraction, the total loss
was greater due to the larger amount of water. Though sediment fluo-
meturon concentrations were high, very little herbicide was actually
lost in this phase since sediment losses were very low. The magnitude
of fluometuron losses were being influenced by the volume of water and
the amount of sediment leaving the target site.
Chemical and bioassay analysis data for vertical fluometuron movement
(Table 5) showed a deeper movement trend with continuous rainfall;
however, this trend was not well defined. Water was applied to the
plots until the desired runoff was obtained, and this amount varied
among plots. The deeper movement in one replication at the fast inten-
sity could be due to more water being applied to the plot or experimen-
tal error. Deeper fluometuron movement into a soil column at high
compared to low water flow rates has been reported by Hornsby and
Davidson for laboratory studies. Also, at the fast water application
rate it may have been possible to exceed equilibrium adsorption due to
a high infiltration rate; thus, producing a deeper movement of the
herbicide into the soil. At the slow rainfall intensity, rainfall was
applied over a 4-day period and some water was lost by evaporation
during this period. As a result, less water actually entered the soil
and was available for movement through the profile. For all treatments,
fluometuron moved into the 5-10 on soil depth. In general, the bioassay
analysis indicated a lower concentration of fluometuron present than did
the chemical analysis method. This was not surprising in that the chem-
ical analysis method detects fluometuron and all its metabolites down to
3—trifluoro-methyl aniline. Thus, less toxic breakdown products would
68
-------
Table 5. CHEMICAL AND BIQASSAY ANALYSIS OF FLUOMETURON AT THREE SOIL DEPTHS
FOR DIFFERENT RAINFALL INTENSITIES
Rainfall
frequency
Fast
Fast
Intermediate
Intermediate
Slow
Slow
Initial conc.a
Rainfall 0-5 cm
amount, cm Chem. Bio.
7.5 5.7 + 4.0
0.42
6.3 4.1 + 2.5
0.25
6.3 5.7 + 4.0
0.42
5.6 4.1 + 2.9
0.25
6.3 3.5 + 3.2
0.19
6.3 4.5 + 4.1
0.29
Post-rainfall cone. , yg/g
0-5 cm
Chem. Bio,
3.0 + 2.2
0.19
1.7 + 1.5
0.05
2.8 + 2.3
0.13
2.1 + 1.0
0.08
2.3 1.3
0.09
2.8 + 2.4
0.13
5-10 cm
Chem. Bio.
1.5 + 1.9
0.04
1.2 + 1.4
0.02
0.6<+ 1.0
0.01
0.9 + 1.1
0.01
0.4<+ 2.5
0.01
1.1 + 1.4
0.02
10-20 cm
Chem. Bio.
0.8 + 0.5
0.01
0 < + 0
0.01
0 < + 0
0.01
0 < + 0
0.01
0 < + 0
0.01
0 < + 0
0.01
^luoraeturon concentration (yg/g) in the top 5 cm of soil immediately after application. Chem. means
chemical assay, and bio. means bioassay results.
-------
be detected chemically but not by bioassay. It is also possible that
the herbicide in the laboratory prepared soil standards may be more
available than that in the field samples subjected to environmental
variations. In addition, the bioassay species was not sensitive to
small differences at the higher fluometuron concentrations since at these
concentrations (above 2 yg/g) the plant was killed.
Using the "Hybrid" program for transient soil water conditions (Appendix
B), the total fluometuron concentration distribution in Teller sandy
loam following infiltration (5.7 on of water) and two days of evapora-
tion (-0.02 cm/hr) and redistribution was calculated. Figure 24 shows
the agreement between the calculated and experimental fluometuron dis-
tributions. The soil water content distribution prior to infiltration,
at the cessation of infiltration and 48 hours after infiltration are
also shown in Figure 24. The adsorption distribution coefficient, N and
N? were: 0.66, 1.19 and 1.70, respectively. The agreement between the
calculated and experimental fluometuron concentration distributions was
reasonable and would be satisfactory for simulation of most field prob-
lems. These calculations also illustrate the versatility of the "Hybrid"
program for describing the movement and distribution of water and pesti-
cides in soils.
Antecedent soil moisture significantly influenced fluometuron concentra-
tions in the water and on the sediment as well as the total losses by
runoff (Table 6). A lower initial concentration of fluometuron was
applied to these plots compared to the rainfall intensity experiment.
Consistent with the results of the fast rainfall intensity experiment,
runoff from the antecedent dry plots contained very low herbicide con-
centrations in the water, but higher concentrations on the sediment.
However, when the fluometuron was applied to the wet soil and rainfall
simulated, runoff losses were much higher. The concentrations in the
first 3.78 L of water was above 0.6 yg/ml, but the concentration de-
creased substantially in the next 41.6 L. The herbicide concentrations
on the sediinent were approximately twice those from the dry plots. The
70
-------
TOTAL FLUOMETURON CONC.(/ig/g)
0 1.0 2.0 3.0 4.0 5.0
10
£
u
O. 20
LU
O
O
CO
30
FLUOMETURON
t = Ohr
t - 48 hr
e
WATER .
0 10 0.2 0.3 0.4 0.5
SOIL WATER CONTENT (cmVcm3)
Figure 24. Experimental and calculated total (solution and adsorbed)
fluometuron concentration distribution in Teller sandy loam
soil following infiltration of 5.7 cm of water (t=0) and 48
hours of evaporation (-0.02 on/hr) at the soil surface. The
initial soil water content was 0.18 on /on . The vertical
lines are the average measured fluometuron concentration in
these regions under natural field conditions following 5.7
cm of water infiltration and 48 hours of evaporation and
redistribution.
71
-------
-J
Table 6. THE EFEEX7T OF ANTECEDENT SOIL WATER CONTENT ON THE CONCENTRATION AND
PERCENT OF THE APPLIED FLUOMETUEON IN THE RUNOFF WATER AND SEDIMENT
Antecedent
soil
moisture
Dry-
Wet
Water
Initial
oonc.a
2.4 +
0.10
2.2 +
0.08
3.78 Lb
yg/nxL %
0.06 + 0.01
0.003
0.63 + 0.21
0.05
3.78-45.4 Lc
yg/ml %
0.03 + 0.07
0.001
0.21 + 0.74
0.015
Sediment
yg/g %
1.8 + 0.02
0.06
3.1 + 0.02
0.16
Total
% loss
0.10
0.96
one concentration of fluometuron (yg/g) chemically determined to be in the top 5 cm of soil sampled
immediately after herbicide application.
The concentration and percent loss of fluometuron in the first 3.78 L of runoff water from the
plot.
°The concentration and percent loss of fluometuron in the next 41.6 L composite of runoff water.
-------
losses from the individual wet plots were 0.83 and 1.09% of that applied.
These plots were wetter than would normally be the case when using sur-
face application techniques, but might be characteristic of conditions
during an aerial application.
The wet plots required approximately one-half the volume of simulated
rainfall to produce runoff compared to the dry plots. As a result, less
fluometuron was leached into the profile (Table 7). Less movement
Table 7. CHEMICAL AND BIQASSAY ANALYSIS OF FLUOMETURON AT THREE
SOIL DEPTHS FOR DIFFERENT RAINFALL INTENSITIES
Antecedent
soil
moisture
Dry
Wet
Initial
Rainfall conc.a
amount, cm Chem. Bio.
5.3 2.4 + 2.3
0.10
2.8 2.2 + 2.3
0.08
Soil depth
0-5 cm
Chem. Bio.
2.0 + 2.1b
0.07
2.2 + 1.8
0.08
5-10
Chem.
0.3 <+
0.01
0.0 <+
0.01
cm
Bio.
<0.5
0.0
fluometuron concentration (yg/g) in the top 5 cm of the profile imme-
diately after application.
Fluometuron concentrations expresses as yg/g of soil.
into the profile would leave a higher concentration of herbicide on the
surface and available for removal from the target site. The chemical
and bioassay analyses were in close agreement. No herbicide was shown
to move deeper than the 5-10 cm depth in the wet plots.
Plots from the rainfall intensity experiment were sampled at three dates
to determine fluometuron movement and dissipation (Table 8 and 9). Bio-
assay analyses were not conducted on samples from the 20-30 cm depth.
In general, the bioassays again indicated lower concentrations but both
73
-------
Table 8. CHEMICAL ANALYSIS OF FLUOME7TURON IN THE SOIL PROFILE OVER AN EIGHTY-FOUR DAY PERIOD
Initial
oonc.a
5.7
4.1
5.7
4.1
3.5
4.5
Rainfall
frequency &
amount, onb
Fast-7.5
Fast- 6. 3
Int. -6. 3
Int. -5. 6
Slow-6 . 3
Slow-6.3
Soil depths, cm
sampled 7/2/73
0-5
3.0C
1.7
2.8
2.1
2.3
2.8
5-10
1.5
1.2
0.6
0.9
0.4
1.1
10-20
0.8
0
0
0
0
0
Soil depths, cm
sampled 7/26/73
0-5
3.0
1.5
2.8
2.3
2.3
2.1
5-10
1.5
0.9
1.0
0.6
0.8
0.6
10-20
0.8
0
0
0
0
0
20-30
0.3
0
0
0
0
0
Soil depths, on
sampled 9/24/73
0-5
2.6
1.6
2.3
1.4
1.7
1.4
5-10
1.9
1.1
0.8
0.8
0.9
1.1
10-20
1.3
0.2
0.1
0.1
0.2
0.2
20-30
0.3
0
0
0
0
0
^luometuron concentration (yg/g) in the top 5 cm of the profile immediately after application.
Applied inmediately after herbicide application by simulated rainfall.
CFluometuron concentration in yg/g of soil.
-------
Table 9. BIOASSAY ANALYSIS OF FUJOMETURON IN THE SOIL PROFILE OVER AN EIGHTY-FOUR DAY PERIOD
ui
Initial
conc.a
4.0
2.5
4.0
2.9
3.2
4.1
Rainfall
frequency &
amount, cm^
Fast-7 .
Fast- 6 .
Int. -6.
Int. -5.
Slow- 6.
Slow-6 .
5
3
3
6
3
3
Soil depths, on
sampled 7/2/73
0-5
2.2C
1.5
2.3
1.0
1.3
2.4
5-10
1.9
1.4
1.0
1.1
<0.5
1.4
10-20
0.5
0
0
0
0
0
Soil depths, cm
sampled 7/26/73
0-5
1.1
0.8
2.8
1.6
1.3
2.6
5-10
0.7
0.5
1.0
0.9
0.4
1.6
10-20
0.5
0
0
0
0
0
Soil depths, on
sampled 9/24/73
0-5
0.9
0.8
1.6
1.0
0.6
1.6
5-10
0.7
0.5
1.1
1.0
0.7
0.9
10-20
0.6
0
0
0.5
0
0.4
iTluometuron concentration (yg/g) in the top 5 cm of the profile immediately after application.
Applied irrmediately after herbicide application by simulated rainfall.
^lucmeturon concentration in yg/g of soil.
-------
methods indicated similar trends. On the final sampling date, fluomet-
uron was shown to be in the 10-20 on soil layer on all plots and the
20-30 on layer on one plot by chemical analysis, but bioassay did not
indicate the presence of fluometuron at all depths. However, most con-
centrations in the 10-20 cm zone were below the limit of bioassay detec-
tion. Although attenuation was occurring, fluometuron remained at
phytotoxic concentrations to oats 84 days after application. In general,
the rainfalls between July 26, 1973 and September 24, 1973 were small
and did not produce significant runoff. Also, because of evaporation
the downward penetration of the water was not sufficient to move the
fluometuron deeper into the soil profile.
?7
The water bioassay procedure for fluometuron, described by Truelove ,
was found to accurately predict herbicide concentrations. Since the
results of selected samples analyzed by bioassay and chemical methods
were in close agreement, only the chemical assay results were presented.
The distribution of a pesticide within a soil profile with time was sim-
ulated for steady state soil water flow (Appendix A) and is shown in
Figure 25. The chemical was initially distributed uniformly in the top
15 cm of the soil profile at a rate of 2.24 kg/ha. The adsorption dis-
tribution coefficient, N and N1 were: 0.21, 1.19 and 1.7, respectively.
The soil water was assumed to be moving under unit gradient below the
plant root zone at a soil water flux of 0.128 cm/day. The soil water
o o
content was 0.20 cm /cm and represents the field water content in Cobb
sand (soil used in transient laboratory experiment) in the region below
the plant root zone. Biological degradation of the pesticide was
assumed to be absent. The maximum pesticide concentration decreased
with time and the distribution becomes more skewed as the chemical
moved deeper into the profile. Once the chemical was below the plant
root zone (150-180 cm or 5-6 feet) , the situation described in Figure 25
would be similar to a natural occurring field condition in an irrigated
region. Using this procedure and a biological degradation submodel, the
fate of a pesticide could be simulated prior to using it in an agricul-
tural program.
76
-------
HERBICIDE CONC. IN SOLUTION (/xg/cm3)
0 0.2 0.4 0.6 0.8 1.0 1.2 1.4
Figure 25. Simulated pesticide distribution in a soil profile with time
for steady state soil water conditions. Average soil water
content, water flux, distribution coefficient, N and N1 were
3 3
0.20 on /on , 0.128 cm/day, 0.21, 1.19, and 1.7, respectively.
77
-------
SECTION VII
REFERENCES
1. Hashimoto, I., K. B. Deshpande, and H. C. Thomas. Peclet Numbers
and Retardation Factors for Ion Exchange Columns. Ind. Eng. Chem.
Fundamentals 3_: 213-218, 1964.
2. Lapidus, L., and N. R. Amundson. Mathematics of Adsorption in Beds:
VI. The Effect of Longitudinal Diffusion in Ion Exchange and
Qiromatographic Columns. J. Phy. Chem. 56: 984-988, 1952.
3. Lindstrom, F. T., L. Boersma, and D. Stockard. A Theory on the Mass
Transport of Previously Distributed Chemicals in Water Saturated
Sorbing Porous Medium: Isothermal Cases. Soil Sci. 112; 291-300,
1971.
4. Oddson, J. K., J. Letey, and L. V. Weeks. Predicted Distribution of
Organic Chemicals in Solution and Adsorbed as a Function of Position
and Time for Various Chemical and Soil Properties. Soil Sci. Soc.
Amer. Proc. _34: 412-417, 1970.
5. Davidson, J. M., and R. K. Chang. Transport of Picloram in Relation
to Soil-Physical Conditions and Pore-Water Velocity. Soil Sci. Soc.
Amer. Proc. 36: 257-261, 1972.
6. Davidson, J. M., and P. W. Santelmann. Displacement of Fluometuron
and Diuron Through Saturated Glass Beads and Soil. Weed Sci. 16:
544-548, 1968.
78
-------
7. Hornsby, A. G., and J. M. Davidson. Solution and Adsorbed Fluo-
meturon Concentration Distribution in a Water-Saturated Soil:
Experimental and Predicted Evaluation. Soil Sci. Soc. Amer. Proc.
37.: 823-828, 1973.
8. Huggenberger, F., J. Letey, and W. J. Fanner. Observed and Calcu-
lated Distribution of Lindane in Soil Columns as Influenced by
Water Movement. Soil Sci. Soc. Amer. Proc. 36; 544-548, 1972.
9. Kay, B. D., and D. E. Elrick. Adsorption and Movement of Lindane
in Soils. Soil Sci. 104: 314-322, 1967-
10. van Genuchten, M. Th., J. M. Davidson, and P. J. Wierenga. An
Evaluation of Kinetic and Equilibrium Equations for the Prediction
of Pesticide Movement Through Porous Media. Soil Sci. Soc. Amer.
Proc. 38.: 29-35, 1974.
11. Swanson, R. A., and G. R. Dutt. Chemical and Physical Processes
That Affect Atrazine Movement and Distribution in Soil Systems.
Soil Sci. Soc. Amer. Proc. 37: 872-976, 1973.
12. Bresler, E., and R. J. Hanks. Numerical Method for Estimating
Simultaneous Flow of Water and Salt in Unsaturated Soils. Soil
Sci. Soc. Amer. Proc. 33: 827-832, 1969.
13. Bresler, E. Simultaneous Transport of Solutes and Water under
Transient Unsaturated Flow Conditions. Water Resour. Res. _9: 975-
986, 1973.
14. Kirda, C., D. R. Nielsen, and J. W. Biggar. Simultaneous Transport
of Chloride and Water During Infiltration. Soil Sci. Soc. Amer.
Proc. 31_: 339-345, 1973.
15. Hanks, R. J., and S. A. Bowers. Numerical Solution of the Moisture
79
-------
Flow Equation for Infiltration into Layered Soils. Soil Sci. Soc.
Alter. Proc. 26_: 530-534, 1962.
16. Rubin, J. Numerical Method for Analyzing Hysteresis-Affected Post *
Infiltration Redistribution of Moisture. Soil Sci. Soc. Atner. Proc.
31: 13-20, 1967.
17. Stable, W. J. Comparison of Computed and Measured Moisture Redis-
tribution Following Infiltration. Soil Sci. Soc. Arter. Proc. 33;
840-847, 1969.
18. Fava, A., and H. Eyring. Equilibrium and Kinetics of Detergent
Adsorption - A Generalized Equilibrium Theory. J. Phys. Chem.
60_: 890, 898, 1956
19. Lindstrom, F. T., R. Hauge, and W. R. Coshow. Adsorption from
Solution. III. A New Model for the Kinetics of Adsorption-Desorp-
tion Processes. J. Phys. Chem. 74; 495-502, 1970.
20. Bailey, G. W., and J. L. White. Factors Influencing the Adsorption,
Desorption and Movement of Pesticides in Soils. Residue Rev. 32;
29-92, 1970.
21. Davidson, J. M., and J. R. McDougal. Experimental and Predicted
Movement of Three Herbicides in a Water-Saturated Soil. J. Environ.
Qual. 2: 428-433, 1973.
22. Hance, R. J. The Speed of Attainment of Sorption Equilibria in
Some Systems Involving Herbicides. Weed Res. 1_: 29-36, 1967.
23. Warrick, A. W., J. W. Biggar, and D. R. Nielsen. Simultaneous
Solute and Water Transfer for an Uhsaturated Soil. Water Resour.
Res. I.' 1216-1225, 1971.
24. Chaudhari, N. M. An Improved Numerical Technique for Solving Mul-
80
-------
tidiroensional Miscible Displacement Equations. Soc. Petrol Eng.
J. 11: 277-284, 1971.
25. Fried, J. J., and M. A. Combarnous. Dispersion in Porous Media.
Advan. Hydrosci. 7: 169-282, 1971.
26. Bruce, R. R., and A. Klute. The Measurement of Soil Moisture
Diffusivity. Soil Sci. Soc. Amer. Proc. 20: 458-462, 1956.
27. Ttuelove, B., D. E. Davis, and Larry Jones. A New Method for
Detecting Photosynthesis Inhibitors. Weed Sci. 22; 15-17, 1974.
28. Davidson, J. M., C. E. Rieck and P. W. Santelmann. Influence of
Water Flux and Porous Material on the Movement of Selected Herbi-
cides. Soil Sci. Soc. Amer. Proc. 32: 629-633, 1968.
29. Giddings, J. C. Dynamics of Chromatography: Principles and Theory.
New York, Marcel Dekker, 1965. 323p.
30. Keller, Jack, and Jose F. Alfaro. Effect of Water Application Rate
on Leaching. Soil Sci. 102; 107-114, 1966.
31. Wood, A. L. Measured and Calculated Distributions of Fluometuron
and Water During Infiltration. M. S. Thesis. Oklahoma State Univ-
ersity, Stillwater, 1974. 107 p.
81
-------
SECTION VIII
GLOSSARY
A = quantity of pesticide applied to the soil surface or intro-
2
duced at z = 0 (yg/cm )
C = pesticide solution concentration (ug/on )
C = maximum pesticide solution concentration prior to desorption
(ug/cm )
2
D = apparent diffusion or dispersion coefficient (cm /hr)
D = correction for numerical dispersion in finite difference solu-
2
tion of one solute transport equation (cm /hr)
Ilj = correction for numerical dispersion in finite difference
solution of solute transport equation with G and K, equal to
2
zero (cm /hr)
G = correction for numerical dispersion in finite difference
solution of solute transport equation for transient soil water
conditions
AGA, AGp = free energy of activation for adsorption and desorption
I = infiltration rate (cm/hr)
82
-------
I = infiltration rate (on/hr)
K = hydraulic conductivity (cm/hr)
^ = maximum hydraulic conductivity at cessation of infiltration
(cm/hr)
= maximum hydraulic conductivity in soil profile at j-1 time
after the initiation of evaporation and/or redistribution of
water (cm/hr)
K,, K_ = adsorption and desorption partition or distribution coeffi-
cient
N = constant that varied with pesticide and adsorbent
Q = 30/3h
R = gas constant
S = adsorbed pesticide concentration (yg/g)
= maximum adsorbed pesticide concentration prior to desorption
(yg/g)
T = absolute temperature
= -
h = soil water pressure (cm)
i = space or depth indice
j = time indice
83
-------
kA, kD = adsorption and desorption rate coefficients (hr"1)
k', kl = constant generally referred to as frequency factor of adsorp-
tion and desorption (hr~l)
q = volumetric soil-water flux (on/hr)
t = tine (hr)
z = soil depth measured positive downward (cm)
9 = volumetric water fraction (an3/cm3)
p = bulk density (g/cm^)
T = variable of integration
cj> = hydraulic head (cm)
84
-------
APPENDIX A
Computer Program for Solute Transport
Under Steady State Soil Water Conditions
85
-------
INPUT-OUTPUT
DATA FORMAT
The following details the input data. Each DATA SET repre-
sents a READ statement in the program. Seme data sets are
required, others are dependent upon other input variables
read in. Do these in order unless otherwise directed.
DATA SET 1: FORMAT (515) Required only once
Read NTOTAL
NTOTAL = Total number of runs to be made, each run describ-
ing its own system.
DATA SET 2: FORMAT (515) Required NTOTAL times
Read N, NUMBER, IXXX, INTERT, NOVRID
N = For N=0, output printouts will be C vs. X. For
N=l, output printouts will be C vs. time (and
pore volume).
NUMBER = Use only for N=0; the number of different time
values for vvhich C vs. X will be printed.
IXXX = Merely an identification number distinguishing
the run.
86
-------
M For INTER3XL, program will generate chemical at
soil surface for a time T. For INTERIM), the
chemical is initially distributed within the
soil.
NOVKE1> For NOVRID=1, read in on a card an initial C/CO
distribution. For NOVRID=0, the program puts a
square pulse distribution in initially.
DATA SET 3: FORMAT (5F15.5) Required NTOTAL times
Read VEL, D, K, THETA, RHO, CD, XMAX, T, NEXP, AB, PULSE,
KTIME, H
VEL= A constant velocity.
D= Dispersion coefficient. Includes correction for
numerical dispersion associated with finite
difference calculations.
K= The absorption coefficient, S=KCT.
THETA= The water content 0.
RHO The bulk density of the soil.
00= The magnitude of input chemical concentration.
XMAX= The length, or depth in the soil which is farther
than the chemical will ever reach.
T= For INTERT=1, specify the time for which chemical
will enter at the soil surface. If INTERT=0,
T need not be specified.
87
-------
NEXP= The exponent N on S=KCT.
AB= The exponent AB on S=KdC
PULSE= For INTERT=0, specify the distance as measured
from x=0 for which there will be a non-zero
initial C distribution. For INTERT=1, PULSE
need not be specified.
KTIME= The constant time increment At, used to calculate
C(J) . Not to be confused with DELT.
H= The x increment used in the calculations.
DATA SET 4: FORMAT (5F15.5) Optional NTOTAL times
If INTERT=0 and NOVRID=1, read data set 4, or else go to data
set 5.
Read CNST
CNST= The constant initial concentration C/CO which is
initially distributed from x=0 to x=PULSE.
(NOTE: Data set 4 can easily be modified to accept as
input any initial distribution of C/CO for each x
point from x=0 to x=PULSE by making a few changes
in the program. At the time of this writing,
we had not done so in the Chemical Program.)
DATA SET 5: FORMAT (5F15.5) Optional NTOTAL times
If N=0, go to data set 6; if N=l, read data set 5.
Read DELT, TMAX, X, STARTT
88
-------
DELT= For printouts, a value of C/CO will be printed
every DELT units of time.
TMAX= For printouts, the program will halt this parti-
cular run when time=^IMAX.
X= For printouts, the C/CO value will always be
printed out for the point which is a value of X
down the soil profile.
STARTT= For printouts, values of C/CO will be printed
starting at tirne=STARTT, continuing until time=
TMAX, printed every DELT.
(NOTE: If data set 5 was used, the next data set to be
read in is data set 2, if NTOTAL>1).
DATA SET 6: FORMAT (2F15.5) Optional NTOTAL times
If N=0, read data set 6, or else go to data set 5.
Read DELX, TIMER
DELX= For printouts, at time-TIMER, C/CO values will
be printed starting at x=0 and every DELX there-
after until XMAX.
TIMER= C/00 will be printed out vs X whenever time-TIMER.
(NOTE: Data set 6 has to be executed a grand total of
NUMBER times. After NUMBER cards have been read
in, each with a DELX, TIMER pair on it, then the
next data set to be read in is data set 2, if
NTOTAL 1).
END OF DATA SETS
89
-------
An example of input data to the chemical equation is below.
data set 1
data set 2 0. 3. 1021. 1. 0.
data set 3
data set 6
data set 2
data set 3
5. .02 .4
90. 30. 2.3
0. .01 .2
.2 2.
.2 4.
.4 6.
1. 0. 1022.
2. - .13 .4
120. 45. 0.
3.2 .03 .2
.366 1.54
.88 2.5
0. 1.
.38 1.47
.9 1.0
data set 4 .478
data set 5 .1 40. 30. 12.
90
-------
FLOW CHART FOR SOLUTE TRANSPORT PROGRAM
MAIN PROGRAM
(RiCAD DATA SET
3 WRITE DATA
SET 3
CO::PUTE My\xi>fuw GRID
rOINT NUMBERS, II,
L2, M2, LM2
SET KDES(I)-DESKRO
FOR ,\LL I
91
-------
FLOW CHART FOR SOHJTE TRANSPORT PROGRAM
MAIN PROGRAM (Continued)
SOLVE (— ) + -(— ) - 1 « = o . for — value to be
used as a constant pulse Input from x « 0 to x " PULSE
at c = 0.
i For x - 0 to PULSE , set C(I) = -=- - CNST
f Co
( Set variable column len
gth = PULSE 4- 1
A - Dit/Ax
B = VEL-At/Ax
DENOM D/(D+Ax-Vel)
DC '= D/6
92
-------
FLOW CHART FOR SOLUTE TRANSPORT PROGRAM MAIN PROGRAM (Continued)
YES
NO
COMPUTE VARIABLE LENGTH OF
COLUMN AS XDSTNC+2-VEL -TIME.
IF THIS EQUALS XMAX, SET
ISTOP - 1.
93
-------
FIDW CHART FOR SOLUTE TRANSPORT PROGRAM MAIN PROGRAM (Continued)
COMPUTE VARIABLE LENGTH. IF THIS
EQUALS XMAX, SET ISTOP =• 1.
94
-------
FI£W CHART FOR SOLUTE TRANSPORT PROGRAM MAIN PROGRAM (Oontinued)
[COMPUTE VARIABLE LENGTH. IF THIS
LLQUALS_KJAX^_SET_TSTOP - 1.
WRITE OUT C VS TIME FOR
THIS TIME KTAKTT AT X |
STARTT - STAKTT + DELTI
YES
NTOT.'
NTOTAL = NTOTAL-1i
95
-------
FLOW CHART FOR SOLUTE TRANSPORT PROGRAM RUNGE SUBROUTINE
R(I)
NO
c - MVELCAX - ^t)] /c
96
-------
FLOW CHART FOR SOUUTE TRANSPORT PROGRAM RUNGE SUBROUTINE (Continued)
IFOR ALL i. SET u(i) - R(I)|
E(l) • U(2)-DENOM
(L) - U(L2)-DENOM
97
-------
FLOW CHART FOR SOLUTE TRANSPORT PROGRAM HJNGE SUBROUTINE (Continued)
SUBROUTINE SIMPSN MERELY INTEGRATES /C-dx and /S'dx USING THE TRAPEZOIDAL
INTEGRATION RULE.
/C-dx - ^
2C(2) + 2C(3) + ... + 2C(L2) + C(L)j
98
-------
SOLUTE TRANSPORT PROGRAM
c
C PROGRAM TO COMPUTE CHEMICAL CONCENTRATIONS' IN SOLUTION AND ON THE
C SOIL FOR A STEADY STATE CONSTANT FLUX SOIL-WATER SYSTEM. THIS IS
C THE MAIN PROGRAM WHICH ONLY CONTROLS BUT DUES NOT ACTUALLY
C CALCULATE C.
C
C RUNGE SUBROUTINE USED FOR CALCULATIONS OF C AND S.
C SIMPSN= SUBROUTINE USED TO CALCULATE THE AREA FOR C VS X RUNS ONLY.
S C= RELATIVE CONCENTRATION
C S = SUR3ED VALUES
C NEXP = THE CONSTANT EXPONENT ON THE TERM C**N, NEXP=N
C AB= THE CU.NbTANT IN THE EXPONENT ON THE TERM C»*U/AB) USED FOR
C DESOFPTIGN
C C0 = THE MAGNITUDE OF THE INPUT PULSE
C VEL=VELOCITY
C RHO= BJLK DENSITY OF SOIL
C THETA= VOLUMETRIC WATER CONTENT
C K= THE CONSTANT K FOR ABSORPTION
C KRHO = TH£ CONSTANT K MODIFIED SO THAT < RHO*( U**N ) =K *< C**N ) WHERE
C U = C/CO.
C KDES = THE DESORPTION K, MODIFIED AS KRHO.
; D= DIFFUSION COEFFICIENT
C *TIME= THE TIME INCREMENT USED IN THE CALCULATIONS. THIS WILL BE
C A CONSTANT FOR ALL PROGRAM:., AND IS NOT TO BE CONFUSED WITH DELT
C H = THE SPATIAL INCREMENT, SIMILAR IN USE TO KTIME, USUALLY EQUAL
C TO .2, NOT TO 3E CONFUSED rflTri DELX. CAN'T BE LESS THAN .1.
C T= PERIOD FCR WHICH CO IS TO BE APPLIED.
C TMAX= THE MAXIMUM TIME UNDER STUDY THAT WATER IS TO BE APPLIED TO
C THE SOIL SYSTEM.
C OELT= INCREMENTA. TIME FOR WHICH SAMPLES ARE TO BE TA
-------
SOLUTE TRANSPORT PROGRAM (Continued)
DOUBLE PRECISION X
REAL K.KTIME.NEXP ,KRriO,KOES
01 HENS ION CI302) , SI802) ,KDES( 802)
COMMON /8LOK1/C,S/BLOK2/KTIKE,H, A, 8,DC, DENOM, SM AX, ANT , AX
COM.J,OM /dLOK3/L2,INDt X, INJtXl ,LM2
COMMON /BLCmWKOES.KLEW.DESKRO, XPONT , KL EW 1 , VELK , ON
IE=12
REAO( 5, 10 ] NTOTAL
230 IFINTOTAL .LE. 0) STOP
C
C IF N=l, NUMBER NEED NOT BE SPECIFIED
C
READ! 5,10) N,NUMBER,IXXX,INTERT,NOVRIO
W = N
HRlTE(6f10) IXXX
C IF INTERT = 0, ENTER A PULSE VALUE, ELSE LEAVE BLANK
C
READ (5, 20) VEL,D,K,THETA,RHO, CO , XMAX , T, NEXP , AS, PULS E, KT IME, H
WR IT£( 6,30) VEL,D', K.KTI ME ,THETA ,RH 0 , T,C 0, Ad , NEXP ,H
C
C INITIALIZE CONSTANTS AND CONTROL VARIABLES
ISTOP=0
NCSTGP=0
TLD= K.TIME
DAY=