United States
Environmental Protection
Agency
Environmental Research
Laboratory
Athens GA 30613
Research and Development
EPA/600/S3-85/051 July 1987
&EPA Project Summary
Exposure Assessment
Modeling for Aldicarb in Florida
J. D. Dean and D. F. Atwood
A modeling study was performed to
assess aldicarb concentrations in drink-
ing water wells in the vicinity of citrus
groves in Florida. Areas in the citrus
growing region were identified, with
respect to the unsaturated and satu-
rated zones, in which transport and
transformation of aldicarb was thought
to be different. In addition, an extensive
literature search was conducted to
determine degradation rates and
adsorption coefficients for aldicarb.
These regional and chemical data were
used to define various simulation
scenarios. The fate and migration of
aldicarb was then simulated for the
unsaturated zone using the Pesticide
Root Zone Model and for the saturated
zone using the Combined Fluid-Energy
Solute Transport Model.
Results of the unsaturated zone
modeling showed that there were three
statistically distinct scenarios with
regard to pesticide leaching: "ridge"
soils with thick unsaturated zones;
' 'ridge" soils with thin unsaturated
zones; and ' "flatwoods" soils. The
highest loads leached to ground water
were approximately 1 kg/ha, occurring
in areas of ' 'ridge" soils with thin
unsaturated zones.
Combined results of the unsaturated
and saturated zone modeling showed
that, in general, drinking water well
concentrations should be low but may
exceed 10 ppb for aldicarb. The highest
simulated concentrations were in the
range of 12 to 20 ppb, or three to four
times the detection limit of the chem-
ical (5 to 6 ppb). Highest simulated
concentrations were in the surficial
unconfined aquifer system, overlain
with ''ridge" soils having a thin
unsaturated zone.
The effects of well distance from the
source area also were investigated. In
the surficial aquifer, with hydrogeo-
logic properties most conducive to
aldicarb transport, a well at 300 m
(1000 ft) versus a well at 91 m (300
ft) should have from 2 to 100 times
less aldicarb, depending on pesticide
decay characteristics.
Because of the regional scope of the
study and model limitations, ' 'cata-
strophic" situations such as the pres-
ence of sink holes or leaky wells that
would result in much higher concen-
trations were not simulated. These
situations are not discussed in the
report.
This Project Summary was devel-
oped by EPA's Environmental Re-
search Laboratory. Athens. GA. to
announce key findings of the research
project that is fully documented in a
separate report of the same title (see
Project Report ordering information at
back).
Introduction
The pesticide aldicarb is used exten-
sively in Florida for the control of
nematodes and other pests in citrus
orchards. A systemic insecticide, it is
highly mobile in soils and very toxic.
Evidence of aldicarb contamination in
ground water in Florida prompted a
March 1983 suspension of its use in all
but three Florida counties. In September
1983, the use of the compound was
reinstated with the restrictions that it
would not be used within 90 m (300 ft)
of drinking water wells and that appli-
cation would be at half the label rate.
To provide an assessment of the
migration and fate of aldicarb in typical
Florida citrus applications, a modeling
-------
study was conducted with emphasis on
the potential for leaching and ground-
water contamination. This information
can be used subsequently to perform
exposure and risk assessments and to
evaluate the effects of management
alternatives and current use restrictions
on the compound.
The approach used was to collect and
analyze information on compound- and
site-specific factors, and then to use
models to provide the interactive linkages
that would result in estimation of well-
water concentrations. From the chemical
information, values of constants used in
modeling (adsorption coefficients, decay
and transformation rate coefficients, etc.)
were determined. The site-specific infor-
mation was used to delineate a number
of scenarios that represent the likely
conditions for use of the chemical and
where the fate and transport behavior of
the chemical can vary significantly.
These scenarios may differ, for instance.
because of climatic, soils or hydrogeo-
logic variables in addition to chemical
application and agronomic practices. The
information also was used at later stages
to develop model parameter sets for each
scenario.
Site-specific data were gathered on the
Florida citrus growing region. This
information collection effort was divided
into two phases: characteristics of the
land surface and unsaturated zone; and
characteristics of the saturated zone.
Factors of particular interest in the
unsaturated zone were rainfall depths
across region, soil hydraulic conductiv-
ities and pH, and irrigation and other
management practices. In the saturated
zone, recharge rates, aquifer hydraulic
conductivity, water table gradients,
aquifer configuration, ground-water
temperature and pH are of primary
interest.
This information was used to delineate
both unsaturated and saturated zone
modeling scenarios. These delineations
were based on observed variations in the
important factors that can result in
significantly different fate and transport
behavior. The chemical properties of
particular interest for aldicarb are trans-
formation rate coefficients for the parent
compound to aldicarb sulfoxide and
aldicarb sulfone, and hydrolysis rates and
soil partition coefficients for each
species.
The Pesticide Root Zone Model (PRZM)
and the Combined Fluid-Energy-Solute
Transport (CFEST) model were chosen to
represent the unsaturated and saturated
zones, respectively. For each unsatu-
rated zone scenario, PRZM was run to
produce a time series and frequency
distribution of pesticide loadings. These
loadings were subsequently used by
CFEST to produce estimates of wellwater
concentrations. Within each saturated
zone scenario, sensitivity to factors such
as distance from source area to wells,
size of source area, pesticide loading and
decay rates was determined.
Unsaturated Zone Results
The annual pesticide loadings from 16
final unsaturated zone scenarios were
analyzed in several ways. First, geomet-
ric mean loadings from all scenarios
were subjected to one way analysis of
variance (ANOVA) to determine whether
any substantive differences occurred
between the mean loadings of the
various scenarios. These analyses
showed that, of the means of the 16
scenarios simulated, the only significant
differences were among thick ridge soils
(i.e., entisols and ultisols), thin ridge
soils, and flatwood soils. Visual inspec-
tion of the frequency distribution of
annually leached pesticide loads con-
firmed this. Thus, the outputs of the 16
scenarios were condensed into three
scenarios: "ridge" areas with thick
unsaturated zones; "ridge" areas with
thin unsaturated zones; and "flatwoods"
areas.
The lowest loadings were associated,
in general, with thick entisols and
ultisols. For this scenario, there is only
a 10% probability that the pesticide load
leached to ground water will exceed 0.01
kg/ha. Loads from alfisols and spodosols
slightly exceeded those for the thick
entisols and ultisols. The highest load-
ings occurred in thin unsaturated zone
entisols and ultisols. There is a 10%
chance that loads to ground water will
exceed 0.3 kg/ha. The thickness of the
unsaturated zone in this scenario is 180
cm for ultisols and 270 cm for entisols
and the input load (application rate) is
5.6 kg/ha (5 Ib/acre). Increasing or
decreasing the load for any of these
scenarios by a factor of "x" would result
in an increase or decrease in the load
by the same factor. For instance, if the
application rate were doubled from 5.6
kg/ha to 11.2 kg/ha, the simulated load
at the 10% exceedance level in thin
entisols and ultisols also would double,
from 0.3 kg/ha to 0.6 kg/ha.
Also of interest is the quantity of
aldicarb and that of its two toxic metab-
olites in the leached load. Simulations
revealed that usually less than one
percent of the aldicarb parent is leached
to the saturated zone under any scenario.
Under the thin, unsaturated zone ultisols
and entisols, about 60% aldicarb sulfox-
ide and 40% aldicarb sulfone makes up
the leached load (a 1.5 to 1 ratio). In the
thick, unsaturated zone entisols and
ultisols, the ratio is closer to 0.17 to 1,
sulfoxide to sulfone. For spodosols and
alfisols, the ratio is roughly the same,
0.19 to 1.
Obviously, because these transforma-
tions are kinetically controlled, the
quantity of aldicarb and its toxic metab-
olites appearing in the leachate is a
function of the residence time of the
chemical in the profile. The sooner after
application the pesticide is leached to the
saturated zone, the more aldicarb and
aldicarb sulfoxide will appear in the
leachate.
Saturated Zone Results
In general, the surficial aquifer cases
show the highest relative concentrations
followed by the worst cases for the
Floridan Aquifer and then the average
casesforthe Floridan andthetwo aquifer
cases. The surficial aquifer has high
relative concentrations for two reasons.
The relatively thin aquifer has less water
for dilution of the pesticide load from the
unsaturated zone. Also, the well-induced
gradients and permeabilities are high
enough to significantly increase the
ground-water velocity to the well. This
allows less time for decay. The high
concentrations determined for the worst
case scenarios in the Floridan Aquifer are
primarily due to the high ground-water
velocities that leave little time for decay
or dispersion. The two-aquifer scenarios
and the average Floridan scenarios have
slower ground-water velocities and
thicker aquifers that dilute the initial
pesticide concentrations.
Concentrations are higher in wells 91
m away from the source area. Even in
cases with no decay, the relative con-
centrations at the 300 m well are smaller
because there is more dispersion. The
Floridan worst cases show only a differ-
ence of about 20% between the 91 m
and the 300 m cases. For the average
Floridan and surficial worst cases,
relative concentrations are usually at
least an order of magnitude smaller at
the 300 m well than at the 91 m well.
Again, the greater amount of time
-------
required to travel to the 300 m well
allows more time for decay to occur.
Conclusions
This modeling study was intended to
provide estimates of aldicarb concentra-
tions for a human health risk assess-
ment. The judgment as to whether these
study results represent favorable or
unfavorable situations in that respect
goes beyond the scope of this report.
Considering the worst-case simulation
scenarios and the potential uncertainties
in the technical approach, this study has
shown that the use of aldicarb on citrus
in Florida could potentially lead to well-
water concentrations exceeding 10 ppb
although most estimates were much
lower.
The results are based on model repres-
entations of generalized, regional unsat-
urated and saturated zone modeling
scenarios. Such a study, of necessity,
overlooks specific "special situations"
that may occur within these regions. To
ameliorate the effects of such oversights,
the general approach was to attempt to
look at "worst case" values of param-
eters that could be selected to describe
these generalized scenarios and to adopt
other "average case" values when worst
cases values gave results indicative of
high contamination levels. This was not
done in all cases.
The highest concentration simulated
(in the surficial aquifer) of 20.3 ppb is
within a factor of three to four of the
detection limit for aldicarb and exceeds
the current drinking water standard. This
was for a scenario in which no pesticide
decay was simulated in the saturated
zone. The more realistic scenarios were
those that assumed decay using the best
available estimates for the degradation
coefficient. The highest simulated con-
centration among the decay scenarios
was 0.4 ppb. This concentration was
simulated in the Floridan aquifer with a
shallow well 91 m from the source and
in the surficial aquifer with a shallow
well 91 m from the source. In both cases,
the overlying soils were thin "ridge"
soils.
Based on results of the unsaturated
zone simulations, a number of recom-
mendations can be made for practices
that will tend to diminish loadings from
the unsaturated zone:
• Avoid the application of aldicarb in
ridge areas where there is a thin
unsaturated zone (i.e., a high ground-
water table).
• Avoid the application of irrigation
water in the treated soil bands' in
quantities that would cause move-
ment of water and pesticide past the
crop root zone.
• Avoid the application of irrigation
water for freeze protection after the
application of aldicarb to a grove.
Based on the results from the satu-
rated zone simulations, a number of
recommendations can be made that will
tend to reduce expected well-water
concentrations:
• Avoid application of pesticide up-
gradient of wells that interest major
solution cavities in areas where the
aquifer is unconfined or where direct
linkages exist. (This can sometimes be
discovered from drilling logs.)
• Avoid application near wells in the
surficial aquifer that have high pump-
ing rates and are closely surrounded
by citrus groves.
• Avoid using shallow wells where the
localized induced gradient is very
large.
• Use of a well tapping a confined
aquifer will be safer than one in an
unconfined aquifer.
• Use of deep wells that draw water
from a large section of the aquifer will
result in lower concentrations than
similar wells that are shallower.
Overall Project
The modeling study described in this
summary is part of a risk assessment
performed by EPA to evaluate the poten-
tial risk posed by aldicarb contamination
of drinking water in Florida. Other
components of the study (reported
elsewhere) include a detailed monitoring
of drinking water supplies and a risk
assessment for populations using public
and private water supply wells.
-------
J. D. Dean and D. F. Atwood are with Anderson-Nichols & Co., Inc., Palo Alto,
CA 94303.
L. A. Mulkey is the EPA Project Officer (see below).
The complete report, entitled "Exposure Assessment Modeling for Aldicarb in
Florida." (Order No. PB 87-188 801/AS; Cost: $30.95) will be available only
from:
National Technical Information Service
5285 Port Royal Road
Springfield, VA 22161
Telephone: 703-487-4650
The EPA Project Officer can be contacted at:
Environmental Research Laboratory
U.S. Environmental Protection Agency
Athens, GA 30613
U.S. OFFICIAL MAIL'
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
Official Business
Penalty for Private Use $300
EPA/600/S3-85/051
000052»
------- |