PESTICIDE LEACHING ANALYTICAL MODEL AND GIS-BASED APPLICATION IN
AGRICULTURAL WATERSHEDS
Zhonglong Zhang 1, and Mohamed M. Hantush 2, Associate Member, ASCE
ABSTRACT: Ground-water contamination by pesticides and other organic pollutants has been
detected across agricultural areas and is on the increase. Because ground-water monitoring is too
costly to define the geographic extent of contamination at such large scales, indirect methods are
needed to predict transport and fate of pesticides in the subsurface and assess the potential for
ground-water contamination that may result from applications of pesticides. In this paper, two
analytical solutions that describe one-dimensional transport and fate of pesticides in the soil,
were integrated with a geographic information system (GIS) to estimate spatial solute leaching
distributions and to create ground-water vulnerability assessment maps. The GIS-based models
use process-based equations that consider advection and complete mixing to simulate solute
transport and to predict pesticide transport in both space and time. To demonstrate its
capabilities, the modeling approach is applied to an agricultural watershed in the mid-Atlantic
coastal plain using data from county soil survey and monitoring study in this region. The impact
of processes, such as net infiltration, biochemical degradation, adsorption and advection on the
spatial and temporal distribution of pollutant emissions to ground water are investigated. The
applicability of GIS-based models for vulnerability assessment and determining the risk of
ground-water contamination by pesticides is demonstrated in susceptible agricultural areas.
INTRODUCTION
Fertilizers and pesticides are extensively used in the U.S. to increase agricultural production.
These agrichemicals are significant sources of nonpoint pollution, and long-term exposure of
land areas to these agrichemicals provides the potential for ground-water contamination. Across
the country, pesticides and other organic pollutants have been detected in ground water
underlying agricultural regions (USEPA, 1992). The possible deleterious effects of pesticides on
humans and ecosystems have been documented (Biradar and liayburn, 1995), and careful
management of pesticides on agricultural areas in order to avoid ground-water contamination is
desired by both farmers and the general public.
Ground water monitoring is too costly to adequately define the geographic extent of
contamination at a large scale. Therefore, indirect methods are needed to assess the potential for
ground-water contamination by pesticides over large areas. Agricultural practices that provide
sources or enhance the potential for pesticide leaching to ground water need to be identified and
their geographic extent defined. Examination of the spatial distribution of contaminant sources
may help identify areas susceptible to pesticide contamination. Vulnerability assessments can
provide essential screening tools to understand the relative mobility of pesticides and predict
potential ground-water contamination.
1 ManTech Environmental, U.S. EPA NRMRL Subsurface Protection and Remediation Division, P.O. Box 1198,
Ada, OK 74820 (E-mail: zhang.zhonglong@epa.gov)
2 U.S. EPA National Risk Management Research Laboratory, Subsurface Protection and Remediation Division,
P.O. Box 1198, Ada, OK 74820
1
-------
Ground-water vulnerability assessment methods can be classified into three general types:
statistical methods, overlay and index methods, and process-based transport models (Water
Science and Technology Board, 1993). In this paper, a process-based model that describes one-
dimensional solute transport at subsurface is considered. Solute transport models have become a
useful tool in understanding ground-water contamination caused by the migration of pesticides
through soil into ground water. This method may be distinguished from the others in its potential
to predict solute transport in both space and time. Several models that use process-based
equations are capable of simulating the transport and transformation of pesticides in the
subsurface. Models used to evaluate the ground-water contamination potential by pesticides
range from one-dimensional to multi-dimensional models. Comprehensive transport models such
as GLEAM (Knisel et al., 1993), and PRZM-2 and 3 (Carsel et al., 1997) are usually data
intensive and require knowledge of soil, land use, environment, and pesticide parameters {Rao et
al., 1985). In most cases, such parameters are neither available nor likely to be obtained for a
large number of soil-land-pesticide combinations. For these reasons, the modeling simulations
are not always satisfactory to address potential ground-water contamination problems on a large
scale.
In order to evaluate ground-water contamination potential on a large scale, several vadose
zone solute transport models have been integrated with a geographic information system (GIS).
A comprehensive review of GIS-based nonpoint source pollutant modeling in the vadose zone
has been presented by Corwin, et al. (1997). Wesseling and van Duersen (1995) programmed
dynamic mass transport models into a GIS. Tim (1996) coupled a one-dimensional advective
dispersion-reaction equation with Arclnfo GIS to demonstrate the capabilities of embedding a
vadose zone model into a GIS. Lasserre et al. (1999) linked an unsaturated zone transport model
with IDRISI GIS to simulate nitrate fluxes into ground water. Hantush et al. (1999) coupled
multi-phase mass fraction models with a GIS and applied the framework to the assessment of
uncertainty of ground water vulnerability to some pesticides in agricultural watersheds in the
mid-Atlantic coastal plains. The framework and the application to the Locust Grove study area
demonstrated the existing uncertainty for predicted results when fixed input parameters are
considered in fate and transport models. Shukla et al. (2000) coupled the attenuation factor (AF)
model with a GIS on a county-scale to screen more than 70 pesticides. Based on the literature,
the trend in coupling GIS and environmental models is leading toward more sophisticated
process-based models.
In this study, two solutions of a one-dimensional solute transport model were integrated with
GIS for predicting the fate of pesticides in the subsurface. Pesticide transport mechanisms
including advection and complete mixing were investigated for a case study with limited
watershed field data. The purpose of this work was to develop a simple procedure which can be
employed to estimate the relative behavior of pesticides on a large scale (e.g., watershed) and to
evaluate the pesticide contamination potential. As compared to comprehensive transport models,
a GIS-based transport model relies on georeferenced data which allows visualization of spatially
distributed data on a true map scale. The modeling outputs give the spatial solute leaching
distributions and the vulnerability assessment maps, which can be used to identify areas
susceptible to potential ground-water contamination. To illustrate the modeling procedure, the
ground-water vulnerability maps of three pesticides: atrazine, dicamba, and cyanazine, are
presented for the case study. These results demonstrate the potential for GIS-based transport
models to be used as pollution assessment screening tools for managing a large number of
pesticides on an agricultural area.
2
-------
TRANSPORT AND FATE MODELS
Problems of solute transport involving first-order decay reactions frequently occur in soil and
ground-water systems. When the volumetric moisture content and the pore-water velocity remain
constant in time and space (steady-state flow), the governing solute transport equation in a one-
dimensional system including linear equilibrium adsorption and first-order decay is given by
Equation (1):
where C = the solute concentration in water (M/L3); t = the time (T); z = the distance along the
flow path (L); D = the dispersion coefficient (L2/T); v = the Darcy velocity (L/T); 0= the
volumetric water content of the soil (L3/L3); k = the first-order decay coefficient in water (1/T); R
= the soil retardation factor; KH = the dimensionless Henry's constant; k = volumetric air
content, Kd = the linear distribution coefficient; Koc = the organic carbon partition coefficient;
and foc = the fraction of organic content of the soil by weight.
In general, analytical solutions offer substantial advantages over other solution methods for
processing transport models. Two conceptual solutions which include advection and complete
mixing are considered in this study. The solutions mentioned here do not represent all of the
modeling options available for pesticide leaching simulation, but they may be worth of
investigation. Approximation errors associated with complete mixing and advective solutions are
relatively small for large Peclet number {Hantush et al, 2000; and Luckner and Schestakow,
1991).
Complete Mixing Solution
In this approach, the soil is considered to be composed of the crop-root zone and the vadose
zone, each with uniform properties. The model simulates movement of the contaminant through
both zones and complete mixing of the pesticide is assumed. The solution takes into account
processes such as advection, adsorption, degradation, and volatilization from soil surface.
Pesticide mass per unit area of soil is assumed to be instantaneously mobilized by infiltrating
water into the root zone.
The solution of Equation (1) under the complete mixing condition has been given and
described in detail by Hantush and Marino (1996). It is described separately for the root zone
and vadose zone.
The root zone extends from the ground surface down through the crop roots. Pesticide
leaching occurs in this zone due to irrigation and rainfall. The analytical solution of the
concentration with time below the root zone can be obtained by:
(1)
R = l + (Kdpb + kKh / 6)
(la)
Kd = Kocfoc
(lb)
Root zone
3
-------
C,(0 = [C,(0 + Co]exp
1 + (T / /l)(ln2 + ju)
T
(t-t o)
(2)
M =
(FS + cr/h);I
ln(2)ftft
(2a)
Tr —
kRrQr
o = DgKJd
(2b)
(2c)
C„ =Mo/9hR
(2d)
where h = the root zone depth (L); 6r = the moisture content; R, = the retardation factor in the
root zone; v* = the net ground-water recharge (L/T); Dg = the soil-gas diffusion coefficient (L2/T); d
= the air boundary layer thickness above the surface (L) and assumed to be equal to 5 cm; F =
the transpiration stream concentration factor; S = the transpiration rate (1/T); A = the half-life
(T); and Mo = the total pollutant mass applied perunit area(M/L2).
Vadose zone
The vadose zone extends from the bottom of the root zone to the water table. Almost all the
water that enters this zone percolates downward to the water table. An analytical solution of
concentration with time to the water table from the vadose zone can be obtained by:
Cu (t) = exp
l + ln2(T IX)
T
exp
l + (r //t)(ln2 + //)
T
c-o
c-o
C,('.) + [C('.) + C.]r
exp
1 + In 2(Tu / X)
T
(t~K)
T [1 + In 2 (T / A)] -T[ 1 + (T / A)(ln 2 + //)]
(3)
T„ =
HRudu
(3 a)
where H = the ground-water depth below the root zone (L); du = the moisture content; and Ru =
the retardation factor in the vadose zone.
Advective-Reactive Solution below Root Zone
For solute transport in the subsurface, advection transport may be dominant. In this case,
transport due to solute diffusion or dispersion may sometimes be neglected. When only
advection is considered, the concentration in the liquid phase is provided by the equation (1) by
4
-------
setting dispersion, D, to zero. The use of advective model allows for accounting for delayed
breakthroughs, which are not accounted for by the complete mixing model. We consider the case
of advection in the vadose zone. Given the following initial and boundary conditions:
C(z, 0) = C (4a)
C(0,0 = C,(0 (4b)
where Q = the initial uniform concentration in the vadose [M/L3], Cr(t) = the concentration at
the bottom of the root zone [M/L3] and is given by the above complete mixing model, and t = the
time period [T], With these boundary and initial conditions, a solution of Equation (1), in which
only advection is addressed, is obtained using Laplace transformation:
C(z,t) = Ge-
C(z,t) = Ge-kt +C[t-
R(z - h)
vie
kR(z-h)
]e
vie
t R(z-h)
v/e
R(z-h)
t>
v/e
(5a)
(5b)
C[/ _«(£*) j = [C (I) + C^
v!0
1+(Tr/A)(ln2+/i) R(z-h)
—z V-hrr-to)
(5c)
GIS-BASED MODEL APPLICATION
The described transport model solutions were integrated with ArcView GIS to perform
regional assessments of ground-water contamination potential by selected pesticides commonly
used in the Delmarva Peninsula (Koterba et al., 1993). GIS-based transport model simulations
can be used to rank pesticides according to their relative leaching potential. With the model
solutions programmed within GIS, an interactive and fully integrated contaminant transport
modeling system is developed. To demonstrate its capabilities, the GIS based model was applied
to a case study. Computation of the concentration for each soil type was performed for each
pesticide and predictive spatial ground-water vulnerability maps were created. All the data layers
required by the model were generated and organized within the GIS.
The study area is an agricultural watershed drained by the Morgan Creek located in Kent
County, Maryland (see, also, Hantush et al., 1999). The soil coverage is shown in Fig. 1 and is
predominantly loamy. Due to extensive agricultural activities at the site, the potential for ground-
water contamination by pesticides exists. The GIS-based transport models are applied to this
region to predict the spatial distribution of the pesticide leaching potential. Ground-water
recharge estimates are based on the monthly water balance between the infiltration water from
precipitation and irrigation, and the outflow of water from surface runoff and evapotranspiration.
The SCS curve number method was used to estimate surface runoff. Monthly evapotranspiration
(ET) is estimated using the method described by Thornthwaite and Mather (1957). All of
calculations are performed within a GIS.
For the case study, the pesticides listed in the Table 1 are chosen for assessing their relative
contamination potential using GIS based models. These pesticides were selected based on
assumptions of their potential use, which may differ from their actual use. A single application of
5
-------
2.5 lb/acre is assumed to produce the predictive ground-water vulnerability GIS maps. Figures 2
displays the results for the three pesticides (atrazine, dicamba, and cyanazine) at a depth of 1.5
m, different time period after application. The spatial trend and magnitude of pesticide leaching
concentration predicted by the GIS-based models indicated that the majority of regions with high
pesticide concentrations are located in areas of high ground-water recharge for complete mixing
model. Subsequent effect of dilution and decay cause the low concentration at the high recharge
area.
Pesticides with relatively short half-life and low Koc such as dicamba showed low leaching
concentration levels in the vadose zone. So it is considered to have low mobility among the
selected pesticides. Atrazine showed high leaching concentrations due to relatively large half-life
and high Koc. Cyanazine, which appears to be as mobile as atrazine, showed greater
concentration as a consequence of its relatively large half-life and low Koc.
The modeling results also indicate that advection plays a significant role in the pesticide
leaching. As compared to the advection model, the complete mixing model may underestimate
the leaching concentration at subsurface. For the complete mixing assumption, the solution tends
to underestimate peak concentrations as shown in Fig. 2 because it ignores diffusion and does not
explicitly treat advection. Pesticide movement through the vadose zone proceeds by both
advection and dispersion. The vulnerability GIS maps show that the predicted concentration will
decrease with the application time. These results provide an initial estimate of contaminant
potential by these pesticides but the limited field data in this study preclude a detailed evaluation
of the simulation results.
SUMMARY
Ground-water vulnerability assessments are useful for pesticide management in order to
prevent future ground-water contamination problems. This paper presents two GIS-based solute
transport models that can be used to estimate the relative behavior of pesticides in the subsurface
at a watershed scale. The assessment of the impact of contaminant sources on ground water on
this scale is greatly enhanced by the use of georeferenced data in a GIS-based model, and the
resulting vulnerability maps show the spatial distribution of a pesticide through the soil leaching
to ground water. Relationships between the spatially distributed data and results can be
established through the use of GIS.
This paper demonstrates that GIS-based transport models that incorporate the physical
processes influencing ground-water vulnerability can be used to obtain estimates of solute
leaching to ground water. The analytical solutions of solute transport used in the GIS-based
models are based on many simplifying assumptions and their use may lead to serious errors if
improperly applied. The complete mixing model may underestimate the pesticide contamination
potential whereas the advective model appears to provide more accurate predictions for pesticide
contamination potential.
The simulated results presented here are preliminary. Before simulated results are used for
making regulatory and remediation decisions, assessments of past, current, and future pesticide
use should be conducted. Future work in this area will concentrate on expanding both the soil
and pesticide databases as well as increasing the range of the analysis to the soil series and
parameters. The correlation between soil properties will also be considered. Additionally, future
research on this subject will include the use of the analytical solutions in conjunction with
additional program modifications to further enhance the model's applicability.
6
-------
Table 1. Selected pesticide properties
No.
Pesticide
S (mg/L)
Koc (m3/kg) Kh
X (day)
k (day"1)
1
Atrazine3
32
1.60E-01
2.50E-07
71
0.00976
2
Dicambab
830
2.00E-03
8.90E-8
14
0.0495
3
Cyanazine3
750
1.68E-01
1.20E-04
108
0.00260
a. From Rao et al. 1985.
b. From Shukla at al., 1998.
c. S = solubility, Koc = organic carbon partition coefficient, KH = Henry's constant, X = half-life, and k = decay
rate coefficient.
C &v Hi a m
Flies a idytoam
G raue I rjraid)
G rail* ihifciam
Hem to
Loam
Loam "fa id
Sai dyc&Ytoam
Sa i d>i¦ toa m
S lltyc tjn- to am
Slrtv"toam
mJat r
(b) (c)
Figure 1. Soil distribution (a), watershed (b) and ground-water recharge (c) at Locust Grove site
7
-------
dteamba (ppb)
~.~36-Dili
¦¦ D.Di - DlliS
11.0*5-01151
0.051 - DATS
¦¦ mat r
dicamba ifpb)
atailre (ppb)
S1. + - 57.1
57.1 - SS.1
H es.1- 7 3.1
73.1 - 1 13.7
M.afcr
Figure 2. Predicted pesticide concentration: (a) dicamba (6 mnh), (c) atrazine (6 mnh), (e)
cyanazine (6 mnh) and (g) atrazine (1 yr.) from complete mixing model; (b) dicamba (6 mnh),
(d) atrazine (6 mnhs), (f) cyanazine (6 mnhs) and (h) atrazine (1 yr.) from advective model.
8
-------
REFERENCES
Biradar, D.P., and Rayburn, A.L. (1995). "Chromosomal damage induced by herbicide
contamination at concentrations observed in public water supplies." J. Envir. Qual., 24,
1222-1225.
Carsel, R. F., Imhoff, J.C., Hummel, P.R., Cheplick, J.M., and Donigian, A.S. (1999). PRZM-3,
A model for predicting pesticide and nitrogen fate in the crop root and unsaturated soil
zones: Users manual for release 3.0. U.S. Environmental Protection Agency, Athens, GA.
Corwin, D.L., Vaughan, P. J., and Loague, K. (1997). "Modeling nonpoint source pollutants in
the vadose zone with GIS." Envir. Sci. Tech., 31(8), 2157-2175.
Hantush, M.M., and Marino, M.A. (1996). "An analytical model for the assessment of pesticides
exposure levels in soils and groundwater." J. Envir. Mod. &_Assess.t, 1(4), 263-276.
Hantush, M.M., Zhang, Z., Marino, M., and Hoque, E. 1999. "Uncertainty in leaching potential
of nonpoint source pollutants with application to a GIS." Proceeding Water Lesson of World
Development. Universities Council on Water Resources, pp. 21-37.
Hantush, M.M., Zhang, Z., Murray, V., and Marino, M. 1999. "Leaching models for subsurface
pollution assessment in agroecosystems " Environmental Engineering' 99, G.C. Schafran, ed.
American Society of Civil Engineers, pp. 177-186.
Hantush, M.M., Marino, M.A., and Islam, M.R. (2000). "Model for soil-aquifer vulnerability
assessment and pollution prevention." J. Hydro., 227(2000), 66-80.
Knisel, W.G., Davis, F.M., and Leonard, R.A. (1993). GLEAMS Version 2.1, Part III: User
Manual, U.S. Department of Agriculture, Agricultural Research Service.
Koterba, M.T., Banks, W.S.L., and Shedlock, R.J. (1993). "Pesticides in Shallow Groundwater
in the Delmarva Peninsula." J. Environ. Qual. 22, 500-518.
Lasserre, F., Razack, M., and Banton, O. (1999). "A GIS-linked model for the assessment of
nitrate contamination in groundwater." J. Hydro., 224(1999), 81-90.
Luckner, L. and Schestakow, W.M. (1991). Migration Processes in the Soil and Groundwater
Zone, Lewis Publishers.
Rao, P.S.C., Hornsby, A.G., and Jessup, R.E. (1985). "Indices for ranking the potential for
pesticide contamination of groundwater." Soil Crop Sci. Soc. Fla. Proc., 44, 1-8.
Shukla, S., Mostaghimi, S., Shanholt, V.O., Collins, M.C. and Ross, B.B. (1998). "A GIS-based
modeling approach for evaluating ground water vulnerability to pesticides." J. Am. Wat.
Peso. Ass. 34(6), 1275-1294.
Shukla, S., Mostaghimi, S., Shanholt, V.O., Collins, M.C. and Ross, B.B. (2000). "A county-
level assessment of ground water contamination by pesticides." Ground water Monitoring &
Remediation, 20(1), 104-119.
Thornthwaite, C.W., and Mather, J.R. (1957). "Instructions and tables for computing potential
evapotranspiration and the water balance." Publications in Climatology, Vol. X, No. 3,
Centerton, New Jersey.
Tim, U.S. (1996). "Coupling vadose zone model with GIS: emerging trends and potential
bottlenecks." J. Envir. Qual., 25, 535-544.
U.S. Environmental Protection Agency (USEPA). (1992). "Final comprehensive state ground
water protection program guidance." EPA 100-R-93-001.
Van Genuchten, M.Th. (1981). "Analytical solutions for chemical transport with simultaneous
adsorption, zero-order production and first-order decay." J. Hydro., 49, 213-233.
9
-------
Wagenet, R.J. and Huston, J.L. (1986). "Predicting the fate of nonvolatile pesticides in the
unsaturated zone." J. Envir. Glial., 15(4), 315-322.
Water and Science Technology Board. (1993). Ground Water Assessment: Contamination
Potential Under Conditions of Uncertainty. National Research Council, National Academy
Press, Washington, DC. pp. 204.
Wesseling, C.G., and van Deursen, W.P.A. (1995). "A spatially modeling language for
integrating dynamic environmental simulation in GIS." Proceedings ofJEC-GI, 368-373.
ACKNOWLEDGEMENTS
This paper has been reviewed in accordance with the U.S. Environmental Protection
Agency's peer and administrative review policies and approved for presentation and publication.
The authors would like to acknowledge the contributions of the Environmental Protection
Agency (EPA) of the Subsurface Protection and Remediation Division, National Risk
Management Research Laboratory and the Center for Subsurface Modeling Support personnel.
10
------- |