EXPOSURE ASSESSMENT MODELING
FOR
ALDICARB IN FLORIDA
Anderson-Nichols/
Engineers Environmental Consultants Planners
-------
October 1984
REVIEW DRAFT
EXPOSURE ASSESSMENT MODELING
FOR
ALDICARB IN FLORIDA
by
J.D. Dean
D.F. Atwood
Anderson-Nichols & Co., Inc.
Palo Alto, CA 94303
Final Report
Contract No. 68-03-3116
Work Assignment No. 23
Project Officer
Mr. Lee A. Mulkey
U.S. Environmental Protection Agency
Environmental Research Laboratory
Athens, GA 30613
ATHENS ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
ATHENS, GA 30613
-------
ABSTRACT
A modeling study was performed to assess aldicarb
concentrations in drinking water wells in the vicinity of
citrus groves in the state of Florida. Areas in the citrus
growing region were identified, with respect to the
unsaturated and saturated zones, in which fate and transport
of aldicarb was thought to be uniquely 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
PRZM (the Pesticide Root Zone Model) and the saturated zone
using CFEST (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
from the treated band 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. The highest simulated
concentrations were approximately at the detection limit of
ii
-------
the chemical (5 to 6 ppb). Highest concentrations were
simulated in the surficial unconfined aquifer system, not in
the Floridan, overlain with "ridge" soils having a thin
unsaturated zone. Highest simulated concentrations in the
Floridan Aquifer were less than 1 ppb.
The effects of well distance from the source area were also
investigated. In the surficial aquifer, with hydrogeologic
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 upon pesticide decay
characteristics. In the Floridan Aquifer with hydrogeologic
properties most conducive to transport, a well at 300 m (1000
ft) should have from nearly the same to 10 times less
aldicarb, again, depending upon pesticide degradation rates.
Because of the regional scope of the study and model
limitations, "catastrophic" situations such as the presence of
sink holes or leaky wells which would result in much higher
concentrations were not simulated. These situations are
discussed in the report.
111
-------
CONTENTS
Page
Abstract ii
Figures vi
Tables xiii
Acknowledgments xviii
1. Executive Summary 1
1.1 Purpose of Study 1
1. 2 Technical Approach 2
1.3 Environmental and Pesticide Characteristics... 5
1.4 Summary of Results 23
1.5 Combined Results of Unsaturated and Saturated
Zone Modeling 33
1.6 Conclusions and Recommendations 38
2. Florida Citrus Growing Environment 47
2.1 Surface and Unsaturated Zone 47
2. 2 Saturated Zone 91
3 . Chemical Fate and Transport 124
3.1 Aldicarb Fate 124
3.2 Transport Considerations 130
3.3 Analysis of Pesticide Model Parameters 134
4. Model Application and Results 153
4.1 Unsaturated Zone Modeling 154
4.2 Saturated Zone Modeling 197
4.3 Combined Results of Unsaturated and
Saturated Zone 254
References 264
Appendices
A PRZM Modifications 275
iv
-------
B Plots of Soil Order Physical Properties from
Soil Characterization Analysis 291
C Derivation of First-Order Hydrolysis Rate
Equation for Saturated Zone 307
D Aldicarb TTR Data from Oviedo and Davenport
Sites for 1984 312
E Cumulative Frequency Distributions of Relative
Weil-Water Concentrations 322
-------
FIGURES
Page
1.1 Florida counties ranked by citrus acreage 6
1.2 Subareas delineated for unsaturated zone modeling. 9
1.3 Matrix showing unsaturated zone modeling
scenarios simulated 11
1.4 Principal aquifer geometries in the citrus
growing area 12
1.5 Matrix showing saturated zone scenarios simulated. 18
1.6 Schematic of aldicarb environmental chemical
pathways 20
1.7 Frequency of annual quantity of pesticide leached
to the saturated zone from thick Entisols and
Ultisols 25
1.8 Frequency of annual quantity of pesticide leached
to the unsaturated zone from thin Entisols and
Ultisols 26
1.9 Frequency of annual quantity of pesticide leached
to the saturated zone from Spodosols and Alfisols. 27
vi
-------
1.10 Overlap of unsaturated and saturated model areas.. 34
2.1 Florida counties ranked by citrus acreage 48
2 . 2 Annual precipitation in inches 51
2. 3 Annual lake evaporation in inches 52
2.4 Mean annual number of days having a minimum
temperature of 32 F (0 C) or below 55
2.5 Location and extent of soil orders on which
citrus is grown 56
2.6 Mean values of field capacity water content
vs. depth from soil characterization analysis 60
2.7 Mean values for saturated hydraulic conductivity
versus depth from soil characterization analysis.. 61
2.8 Mean values of soil pH versus depth from the
soil characterization analysis 63
2.9 Mean values of organic carbon versus depth from
soil characterization analysis 64
2.10 Typical configuration for bedded citrus in the
Flatwoods Area 71
2.11 Generalized topography of the Central Peninsula,
Florida 76
2.12 Typical cross-section through a ridge citrus area
showing relative thickness of unsaturated zone.... 77
vii
-------
2.13 Subareas delineated for unsaturated zone modeling. 79
2.14 Selected meteorologic stations for unsaturated
zone modeling 85
2.15 Fall of fre.e water level during one furrow
irrigation drying cycle in single bedded
'Ruby Red1 grapefruit groves planted on Felda and
Immokalee soil types in the Indian River area 90
2.16 Principal areas where the piezometric surface of
the Floridan Aquifer rises above the water table.. 97
2.17 Areas of the Floridan Aquifer which do not meet
drinking water standards 99
2.18 Principal aquifer geometries in the citrus
growing area 102
2.19 Potentiometric surface of the Floridan Aquifer.... 115
3.1 Schematic of aldicarb environmental chemical >.
pathways 126
4.1 Predicted and observed Aldicarb TTR in the upper
300 cm of the soil at the Lake Hamilton site, /
1984 156
4.2 Predicted movement of Aldicarb residues at the /
Lake Hamilton location 157
4.3 Location of Aldicarb TTR in the soil profile /
> 5 ppb at the Lake Hamilton site 159
VI11
-------
4.4 Predicted and observed Aldicarb TTR in the upper ,
150 cm of the soil at the Oviedo site 160
4.5 Location of Aldicarb TTR in the soil profile
> 5 ppb Oviedo 161
4.6 Comparison of Aldicarb TTR degradation rates
for 1984 Ovieda and Davenport data 166
4.7 Comparison of measured and simulated Aldicarb
TTR concentrations for two 1984 sampling dates
at Oviedo, Florida 168
4.8 Comparison of observed and simulated Aldicarb
TTR concentrations for two 1984 sampling dates
at Davenport, Florida 169
4.9 Schematic of a unit block of citrus (one tree,
not to scale) . 172
4.10 Fate of Aldicarb TTR after application to soil.... 180
4.11 Geometric mean annual quantity of pesticide
leached to the saturated zone from the treated
band 183
4.12 One way analysis of variance of the effect of
irrigation on method annual pesticide mass
leached to the saturatd zone (kg/ha) 185
4.13 Frequency of annual quantity of pesticide leached
to the saturated zone from Entisol scenarios 186
4.14 Frequency of annual quantity of pesticide leached
to the saturated zone from Ultisol scenarios 187
ix
-------
4.15 Frequency of annual quantity of pesticide
leached to the saturated zone from the Spodosol
scenarios 188
4.16 Frequency of annual quantity of pesticide leached
to the saturated zone from the Alfisol scenarios.. 189
4.17 Frequency of annual quantity of pesticide
leached to the saturated zone from thick Entisols
and Ultisols 191
4.18 Frequency of annual quantity of pesticide leached
to the unsaturated zone from thin Entisols and
Ultisols 192
4.19 Frequency of annual quantity of pesticide
leached to the saturated zone from Spodosols and
Alf isols 193
4.20 Cumulative mass curve of pesticide leaching within
the year for three representative scenarios 198
4.21 Plan view of a hypothetical citrus grove and
well configuration 202
4.22 Schematic graph of pesticide concentration through
time based on residence time under source area.... 210
4.23 Schematic diagram of pesticide accumulation
process 211
4.24 Concentration profile for eight month residence
time based on plug flow 213
x
-------
4.25 Effect of dispersion on concentration in the
well over time 214
4.26 Concentration histories at variable field
widths for the Floridan worst case scenario
with no decay 216
4.27 Concentration histories at variable field widths
for the Floridan worst case scenario with decay... 217
4.28 Relationship of field width to peak concentration
at the well for the Floridan worst case 218
4.29 Concentration histories at variable field
widths for the surficial worst case scenario
without decay 219
4.30 Concentration histories at variable field widths
for the surficial worst case scenario with decay.. 220
4.31 Concentration versus time for Floridan worst
case with a shallow well 224
4.32 Concentration versus time for the Floridan worst
case with a deep well 225
4.33 Concentration versus time for the Floridan
average case with a shallow well 227
4.34 Detail of Floridan average case with a shallow
well and decay 228
4.35 Concentration versus time for the Floridan
average case with a deep well 229
XI
-------
4.36 Detail of Floridan average case with a deep
well and decay 230
4.37 Concentration versus time for the surficial
worst case 233
4.38 Detail of the surficial worst case with the well
1000 ft from the source with decay 234
4.39 Concentration versus time for the two-aquifer
worst case with the well 300 ft from the source... 236
4.40 Comparison of the two extreme hydraulic
potential distributions 243
4.41 Effect of Aldicarb source area surrounding
the well 244
4.42 20-year simulation of average Floridan case 246
4.43 Relationship of model time step to peak
concentrations for the Floridan and surficial
worst cases 249
4.44 Unsaturated and saturated zone groupings 255
4.45 Overlap of unsaturated and saturated model areas.. 256
Xll
-------
TABLES
Page
1.1 Approximate Citrus Acreage Associated With
Each Aquifer Geometry 14
1.2 Cases Considered for Saturated Zone Modeling 17
1.3 Summary of Pesticide Loadings per Unit Citrus
Block Area for the Three Final Unsaturated Zone
Scenarios 28
1.4 Relative Concentration Values for Indicated
Saturated Scenarios at Three Exceedance
Probabilities 31
1.5 Highest Calculated Aldicarb Concentrations
(in ppb) in the Given Combined Unsaturated/
Saturated Catagories 36
2.1 Summary of Bearing and Nonbearing Citrus
(Oranges and Grapefruit) by County as of
January 1, 1982 50
2.2 Mean Monthly Rainfall and Evapotranspiration at
Two Citrus Growing Locations in Florida (cm) 53
Xlll
-------
2.3 Soil Samples Selected for Soil Characterization
Analysis 58
2.4 Citrus Irrigation Systems with Acreages for Each
of Three Areas in the State 67
2.5 Summary of Management Characteristics of
Irrigation Systems 68
2.6 Characteristics of Areas Selected for Unsaturated
Zone Modeling 80
2. 7 Unsaturated Zone Modeling Scenarios 83
2.8 Hydraulic Characteristic Data for Florida Soils
by Horizon 87
2.9 Physical Factors Influencing Pesticide Movement
in the Saturated Zone 93
2.10 Generalized Stratigraphic Units and Associated
Hydrogeologic Properties 94
2.11 Approximate Citrus Acreage Associated With Each
Aquifer Geometry 104
2.12 Cases Considered for Saturated Zone Modeling 108
2.13 Average and Worst Case Hydrogeologic Parameters
for the Three Aquifer Configurations Ill
2.14 Well Rates and Depths for Saturated Zone Modeling. 123
xiv
-------
3.1 First Order Rate Constants for Oxidation and
Hydrolysis of Aldicarb in Soil 128
3.2 First Order Hydrolysis Rate Constants for
Aldicarb, Aldicarb Sulfoxide and Aldicarb Sulfone..l29
3.3 Koc for Aldicarb and its Daughter Products 132
3.4 Multiple Linear Regression Coefficients for
Aldicarb Transformation and Degradation Rates 136
3.5 Characteristic Data for Florida Soils Used to
Estimate Pesticide Parameters by Horizon 138
3.6 Transformation and Degradation Rates (k) and
Adsorption Partition Coefficient (K) Used in
Modeling Aldicarb Fate and Transport in Florida
Soils 139
3.7 Comparison of Degradation Rates of Aldicarb
Residues Estimated by Hydrolysis With Degradation
Rates Measured in Soil and Water Degradation
Studies 144
3.8 Hydrolysis Rate Data Used to Calculate Activation
Energy Parameters 147
3.9 Distilled Water Hydrolysis Rates of Aldicarb
Sulfoxide and Aldicarb Sulfone 148
3.10 Estimated Values of Activation Energies and
Pre-Exponential Factors for Aldicarb, its Sulfoxide
and Sulfone 149
xv
-------
3.11 First-Order Decay Rates in Florida's Ground Water
Based on Activation Energy Analysis 151
4.1 Aldicarb Residues (TTR) in the Soil at Davenport,
Florida, 1984 163
4.2 Aldicarb Residues (TTR) in the Soil at Oviedo,
Florida, 1984 165
4.3 Some Water Balance Components for Twenty-Four
Unsaturated Zone Scenarios (in centimeters) 176
4.4 Fate of Aldicarb TTR after Application to Soil.... 182
4.5 Summary of Pesticide Loadings per Unit Citrus
Block Area for the three Final Unsaturated Zone
Scenarios 194
4.6 Mean Percentage of Aldicarb, Aldicarb Sulfoxide
and Aldicarb Sulfone in the Simulated Leached
Pesticide Load 196
4.7 Model Input Parameters for the Six Aquifer
Types Simulated 201
4.8 Initial Input Concentrations Used in CFEST for
Each Aquifer System Simulated (in ppb) 205
4.9 Time Required to Leach 90% of the Pesticide from
the Unsaturated Zone 207
4.10 Summary of Results for Floridan Worst Case
Simulations 223
4.11 Summary of Results for Floridan Average Case 226
xvi
-------
4.12 Summary of Results for the Surficial Worst Case... 232
4.13 Summary of Results for Two-Aquifer System,
worst case 235
4.14 Comparison of Peak Concentrations Determined by
a Simple Plug Flow Model and by Computer
Simulation 239
4.15 Results of Average Versus Actual Aldicarb Mass
Loading Simulations 253
4.16 Well Water Concentrations (ppb) for Combined
Unsaturated and Saturated Scenarios at 50 and
10 Percent Exceedance Probabilities of Each 259
4.17 Highest Calculated Aldicarb Concentrations
(in ppb) in the Given Combined Unsaturated/
Saturated Catagories 261
xvn
-------
ACKNOWLEDGEMENTS
There are many, many individuals who spent a great deal of
time and effort/ offering data and advise, both freely given
and eagerly accepted.
First of all, the authors would like to acknowledge the
support throughout the project of Mr. Lee Mulkey of the U.S.
Environmental Protection Agency in Athens, GA.
Dr. Jim Davidson of the University of Florida is thanked for
helping us to tap the rich resources of data and personnel at
the University of Florida. Dr. Daniel Spangler, Jim Jensen
and Al Quarles of the Geology Department were helpful in
delineating saturated zone scenarios and helping us find data
for aquifer properties. Mr. Dalton Harrison of the
Agricultural Engineering Department and Dr. Lawrence Parsons,
Dr. Carl Anderson, Dr. Robert Koo and Dr. Harry Woods of the
Lake Alfred Citrus Research Center are especially thanked for
advice and data on soils and irrigation practices and
hospitality in arranging field trips into groves.
Drs. R.H. Biggs and P.G. Webb of the Fruit Crops Department,
University of Florida are thanked for the information supplied
on aldicarb plant uptake.
We gratefully acknowledge the aid of Drs. Lamar Miller and
Jeff Foran (University of Florida) in supplying 1984 residue
xviii
-------
data from the Oviedo and Davenport sites.
Dr. Johan Smelt of the Institute for Pesticide Research,
Wageningen, the Netherlands and Dr. Richard Bromilow of
Rothampsted Experimental Station, U.K. are thanked for their
discussions on ferrous iron catalysis of aldicarb
transformation reactions.
Mr. Charles Tibbals of the U.S. Geological Survey Orlando
office, Dr. Tony Irwin and Dr. John Vechiolli of the U.S.
Geological Survey, Tallahassee office and Mr. Jim Frazee of
St.John's Water Management District were most helpful in their
advice and feedback on the ground-water modeling parameters.
Dr. Ann Lemley of the Department of Design and Environmental
Analysis at Cornell University is thanked for her information
on aldicarb hydrolysis rates and activation energy.
Mr. Fredrick Bond and Ms. Chris Smith of Battelle's Office of
Hazardous Waste Management provided expertise with the CFEST
model used for saturated zone modeling.
Review of the report was provided by Mr. Anthony S. Donigian,
Jr., of Anderson-Nichols. Ms. Tomi Hutchins and Ms. Lisa
Jowise prepared the graphics. Word processing was provided by
Ms. Carol McCullough and Mrs. Dorothy Inahara. Editing and
proofing was provided by Ms. Susan Reutter-Harrah.
xix
-------
SECTION 1
EXECUTIVE SUMMARY
1.1 PURPOSE OF STUDY
The pesticide Temik (aldicarb) is used extensively in Florida
citrus for the control of nematodes and other pests. It is a
systemic insecticide that is highly mobile in soils and is
also very toxic. Evidence of Temik contamination in ground
water in the state of Florida sparked public concern and
prompted a March 1983 suspension of the use of the compound in
all but three Florida counties. In September 1983, the use of
Temik was reinstated with the restrictions that it would not
be used within 90 m (300 ft) of drinking water wells and that
application would be at half the label rate.
Because of the lack of adequate data, it is difficult to make
a meaningful assessment of whether current concentrations are
high enough to support a general ban or to evaluate management
options if continued use is permitted. A coordinated data
collection program, modeling study, and risk assessment is
being conducted by the EPA Office of Research and Development
to establish current contamination levels and the health risk
associated with the continued use of Temik on citrus.
-------
The purpose of the modeling portion of the work is to provide
an assessment of the migration and fate of aldicarb in topical
Florida citrus applications 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 Temik.
This report describes the modeling portion of the study.
1.2 TECHNICAL APPROACH
The concentration of a pesticide appearing in a well is the
result of the interaction of two sets of factors; those which
affect the supply of the compound and those which affect the
transport of the compound from the location where it is
applied to the wells in which it may appear. The factors
which affect the supply and transport break down into two
major groups; those which are compound specific and those
which are site specific. The fundamental approach to this
problem is to collect and analyze information on those
compound- and site-specific factors, and then use models to
provide the interactive linkages which result in estimation of
well-water concentrations.
The first task using the above approach is to gather
information on specific chemical properties and site-specific
information about the area of interest. From the chemical
information values of constants used in modeling (e.g.,
adsorption coefficients, decay and transformation rate
coefficients) are determined. The site-specific information
is used to delineate a number of scenarios among which it is
-------
felt the fate and transport of the chemical significantly
differ. These scenarios may differ, for instance, because of
climatic, soils or hydrogeologic variables. The information
gathered is also used at later stages to develop model
parameter sets for each scenario.
A model or models are then selected which can describe the
fate and transport of the chemical in the region.
Modifications to models, which are made necessary by
idiosyncrasies of the compound or application sites, are made
at this point. If possible, models and modifications made to
models should be tested and, if necessary, calibrated in the
area of interest.
For each scenario, models are run to provide estimates of
well-water concentrations at various probability levels. A
probabilistic framework is necessary because of the stochastic
impact of climate on transport in these systems. Impact of
uncertainty in other variables (e.g., pesticide decay rate,
aquifer permeabilities) is assessed by perturbing these
variables in model simulations and observing subsequent impact
on model output.
In this case, we are specifically concerned with the compound
aldicarb as used in the citrus growing area of Florida.
Site-specific data was gathered on the Florida citrus growing
region. This information gathering was divided into two
topical areas; characteristics of the land surface and
unsaturated zone, and characteristics of the saturated zone.
Factors of particular interest in the unsaturated zone are
rainfall depths across the region, soil hydraulic
conductivities 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 which we felt would result in significantly
different fate and transport characteristics.
The next step was to summarize chemical properties for this
compound which relate to fate and transport. For aldicarb,
properties of particular interest are transformation rate
coefficients for the parent compound to aldicarb sulfoxide and
aldicarb sulfone, and hydroloysis rates and soil partition
coefficients for each species.
PRZM (the Pesticide Root Zone Model, Carsel et al., 1984) and
CFEST (Combined Fluid-Energy-Solute Transport, Gupta et al.,
1982) were chosen to represent the unsaturated and saturated
zones respectively. Modifications were necessary to PRZM to
accomodate the transformation of aldicarb to two toxic
daughter products and lateral drainage induced by the practice
of bedding citrus in some areas. No modifications to CFEST
were required. PRZM was verified using local soil residue
data to insure that it adequately represented the fate and
transport of aldicarb in the unsaturated zone. It was not
feasible to test CFEST in this manner for the saturated zone
applications due to a lack of data.
For each unsaturated 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
well-water concentrations. Within each saturated zone
scenario, sensitivity to factors such as distances from source
area to wells, size of source area, pesticide loading and
decay rates was determined.
-------
1.3 ENVIRONMENTAL AND PESTICIDE CHARACTERISTICS
This section discusses briefly the review of literature on the
pesticide aldicarb and the Florida citrus growing environment
used to develop modeling scenarios. A more detailed treatment
of this material is contained in Sections 2 and 3 of this
report.
1.3.1 Characteristics of the Citrus Growing Environment
1.3.1.1 Unsaturated Zone
Citrus is grown throughout most of central and southern
Florida. The map of Figure 1.1 shows the counties of the
state in which citrus is grown ranked according to the acreage
of grapefruit and oranges grown. A great deal of the fruit is
grown in the "ridge" area, a region of sand hills running
north-south in the center of the state. Much is also grown in
the "flatwoods" area, a region of coastal lowlands which flank
the ridge area on the east and west coasts. This entire area
was scrutinized in terms of four primary groups of factors,
differences in which, it was felt, would cause substantial
differences in the fate and transport of aldicarb in the
unsaturated zone. These groups were:
1) climate
2) soils
3) - management practices, and
4) thickness of the unsaturated zone.
Critical variables in terms of climate were preciptation and
pan evaporation depths. These are factors which determine the
-------
ALABAMA
Jacksonville
Atlantic Ocean
3-st. Lucie
4-Indian River
5-Orange
6-Hardee
7-Martin..
B-Hillsborough
9-Highlands
10-De Soto
11-Paseo-
12-Hendry
13-Osceola
14-Brevard
15-Hanatee
16-Palm Beach
17-Harion
18-Volusia
19-Collier
20-Okeechobee
21-Lee
22-Hernando
23-Seminole
24-Charlotte
25-Glades
26-Pinellas
27-Putnam
28-Broward
29-Sumter
30-Sarasota
31-Citrus
32-Flagler
33-Machua
34-St. Johns
Figure 1.1 Florida Counties Ranked by Citrus Acreage
-------
quantity of water percolating downward through the soils.
In the soil, the variables:
o field capacity water content (and saturated hydraulic
conductivity),
o soil organic matter content, and
o soil pH
were critically evaluated. While field capacity water content
(and hydraulic conductivity) determine the velocity at which
water moves through the soil, soil organic matter content
affects chemical adsorption and soil pH strongly affects
aldicarb degradation. It was determined, through data
analysis, that differences in these factors were pronounced
among four soil orders on which citrus is grown in Florida;
the entisols, ultisols, spodosols and alfisols.
Management practices were also investigated. The primary
difference in management which affects aldicarb migration is
the type of irrigation water application system utilized.
These are four dominant types of systems;
o overhead methods (i.e., permanent set or traveling
volume guns)
o low volume spray
o flood or seepage, and
o drip or trickle.
In addition, the practice of bedding citrus in the flatwoods
areas is a dominant edaphic practice. This is done to create
a greater root zone depth for the trees and to promote
drainage of the root zone.
In the flatwoods areas, water tables are typically very close
-------
to the surface and drainage of the root zone is essential to
production. In the ridge areas, on the other hand,
unsaturated zone thicknesses are greater and such practices
are unnecessary.
Based on climatic, soils, management and physiographic
considerations, the Florida citrus growing region was
subdivided into the six areas shown in Figure 1.2. Area 1 is
an area of primarily ultisols while area 2 consists primarily
of entisols. In these areas, no-irrigation, overhead methods
and low volume spray and drip methods are utilized and citrus
is not bedded. Areas 3 and 5 consist primarily of spodosols
while areas 4 and 6 contain primarily alfisols. In these
areas, citrus is bedded and the overhead, flood, low volume
spray and drip methods are used for irrigation. Areas 5 and 6
receive about 10% more rainfall on an annual basis than areas
3 and 4, however.
At the outset of the modeling exercise several decisions were
made concerning the possible scenarios which could be
simulated. First, the drip irrigation method was not
simulated for the following reasons:
1) From the geometry of the wetted area and the treated
aldicarb band it was determined that, in most cases,
the two would not intersect. There is a potential
for overlap in very young groves, however.
2) Because PRZM operates on a daily timestep, and, due to
the fact that two to three water applications are
needed per day during peak periods, this type of
system could not be adequately simulated.
Second, because PRZM is a one-dimensional model, it was felt
that meaningful simulation of flood irrigtion would not be
8
-------
Jacksonville
Atlantic Ocean
Figure 1.2 Subareas delineated for unsaturated zone modeling,
-------
possible, therefore, this irrigation method was also dropped.
Therefore, a total of twenty-four different scenarios were
actually simulated for the unsaturated zone. These are
indicated in the matrix of Figure 1.3. Ultimately, hydrologic
simulations showed that the attempted representation of the
low volume spray method was also incorrect. Therefore, these
simulations were not used in the final analysis of pesticide
loadings to the saturated zone.
Simulation of aldicarb transport under "freeze protection"
water application events using PRZM was also ruled out because
of the daily simulation timestep. Accurate representation of
this practice would require at least hourly inputs (e.g., air
temperature) and an hourly simulation timestep.
1.3.1.2 Saturated Zone
The occurrence and movement of ground water in the citrus
growing area of Florida are closely related to its geology.
The central Floridan peninsula is comprised of a thick
sequence of hydrologically connected limestone formations
which make up the principal artesian aquifer, the Floridan
Aquifer. This is overlain by younger alluvial and marine
deposits which contain the unconfined surficial aquifer and
the intermediate confined aquifers. All three are present in
the citrus growing areas and supply drinking water. For the
purposes of modeling, this very complicated system of aquifers
was simplified.
Figure 1.4 shows the delineation of the typical aquifer
geometries considered in this study. Areas 1, 2, and 3 are
areas that include the Floridan Aquifer system. Area 1 is
where the Floridan is unconfined and considered alone. Area 2
is where a surficial unconfined aquifer overlies the Floridan
10
-------
HIGH
NONE
OVERHEAD
LOW
VOLUME SPRAY
LOW
NONE
OVERHEAD
LOW
VOLUME SPRAY
THICK
ENTISOLS
ULTISOL8
SPODISOLS
THIN
THICK
THIN
THICK
THIN
THICK
ALFISOLS
THIN
Figure 1.3 Matrix showing unsaturated zone modeling scenarios
simulated.
-------
ALABAMA
Jacksonville
Atlantic Ocean
1 -
2 -
3 -
4 -
5 -
Explanation
Floridan Aquifer Alone
- Unconfined
Leaky Two-Aquifer System
- Surficial Aquifer
. - Floridan Aquifer
Leaky Three-Aquifer System
- Surficial Aquifer
- Intermediate Aquifer
- .Floridan Aquifer
Surficial Aquifer Alone
- Unconfined
Two-Aquifer System
- Surficial Aquifer
- Intermediate Aquifer
Figure 1.4 Principal aquifer geometries in the
citrus growing area.
12
-------
with a confining layer in between. Together they comprise a
leaky two-aquifer system. In area 3, a three-aquifer system
exists, consisting of the unconfined surficial aquifer and a
confined intermediate aquifer overlying the Floridan Aquifer.
In areas 4 and 5 the Floridan is not included because of its
poor water quality or because its potentiometric surface is
above the water table. Area 4 is where only the unconfined
surficial aquifer is considered. Area 5 includes both an
unconfined surficial aquifer and a confined x intermediate
aquifer.
Table 1.1 shows the approximate citrus acreage in each area.
The largest amount of citrus is grown in Area 4, where the
surficial aquifer is considered alone. This is followed by
Areas 2, 3, and 5 where a multi-aquifer system is modeled.
Area 1 has the smallest amount of citrus grown (only 3.0%),
yet it is important to consider because the risk of
contamination is high. Therefore none of the areas were
eliminated for lack of significant quantity of citrus.
These five geometries are grouped into three general aquifer
configurations for modeling: 1) the Floridan, as a single
unconfined aquifer (Area 1), 2) the unconfined surficial (Area
4), and 3) a general two-aquifer system (Areas 2, 3, and 5),
be it the surficial aquifer overlying the Floridan or an
intermediate confined aquifer. Each of these three
configurations were simulated with a worst and an average set
of hydrogeologic parameters.
In addition to the physical and chemical properties of aquifer
materials and their effects on chemical fate and transport,
the drinking water well influences the movement of the ground
water and, therefore, the transport of aldicarb. The well
position relative to the pesticide source area, the depth of
13
-------
TABLE 1.1 APPROXIMATE CITRUS ACREAGE ASSOCIATED WITH EACH
AQUIFER GEOMETRY
County
( in ranked
order)
Polk
Lake
St. Lucie
Indian River
Orange
Hardee
Martin
Hillsborough
DeSoto
Pa sco
Hendry
Osceola
Brevard
Manatee
Palm Beach
Marion
Volusia
Okeechobee
Lee
Hernando
Seminole
Charlotte
Glades
Pinellas
Sumter
Citrus
Total
%
Area
123
31,000 63,000
87,000
10,000
28,500
42,000
14,000 17,500
16,500
13,000 17,000
12,000
8,000
2,000 2,000
2,500
2,000
5,000 700
600
2,000
600 1,000
1,000
21,600 212,300 147,000
3.0 29.5 20.4
4
17,000
64,000
50,000
14,000
32,000
3,500
13,000
2,500
15,000
5,000
7,500
4,500
5,000
5,000
4,500
1,300
243,800
33.8
5
31,000
7,000
6,500
3,500
30,000
2,500
5,000
4,500
1,000
1,000
4,000
96,000
13.3
14
-------
the well and the rate at which it pumps all influence the
ground-water flow regime.
The distance of a drinking water supply well from the area of
pesticide application is crucial in the evaluation and
assessment of potential human exposure to aldicarb through
drinking water. The greater the distance necessary to travel
from the source area to the well through the ground water, the
smaller the chance of contamination. The distance of the well
from the source is an important management consideration. The
current EPA standard prohibits the use of aldicarb within 91 m
(300 ft) of a water supply well. This distance was evaluated
in all the scenarios. The EPA is considering 300 m (1000 ft)
as an alternative distance for aldicarb regulation. This
distance was also simulated in all the scenarios.
In the Florida citrus growing area, water wells vary
considerably in depth depending on the local aquifer
characteristics and water needs. Deep wells with large
capacities are usually drilled to supply municipalities.
Shallow wells have much smaller capacities and supply the
small domestic user. The depth that water is withdrawn from a
well influences the localized aquifer flow system. A shallow
well affects the near surface flow system and, therefore,
would influence the movement of a leaching contaminant sooner
than a deep well with an equivalent pumping rate.
The well rate determines the magnitude of the influence on the
regional flow system. A low pumping rate has a small
localized effect. High pumping rates can change the flow
system on a much larger scale, greatly increasing the ground
water velocities towards the well. Two well rates were
simulated with every well in each hydrologic setting.
The physical factors influencing aldicarb transport in the
15
-------
saturated zone are reviewed in Table 1.2 with the cases that
were considered for each factor. To recapitulate, three
aquifer geometries were considered, the unconfined Floridan
Aquifer, the unconfined surficial aquifer and the leaky
two-aquifer system. Each of these three geometries were
simulated with a worst and an average set of hydrogeologic
parameters. A drinking water well was modeled at both 91 m
(300 ft) and 300 m (1000 ft) distances. Two depths and two
pumping rates were simulated in each hydrogeologic setting.
All the combinations of these factors resulted in forty-eight
modeling scenarios. This number doubled to 96 potential
scenarios because of the consideration of two chemical decay
rates. Two decay rates were used; 1) the best estimate of a
representative rate, and 2) no decay.
The scenarios actually simulated are shown in the matrix in
Figure 1.5. Forty-four simulation runs were made. The
motivation for dropping most of the scenarios which were not
run was lack of significant (> 1 ppb) contamination in cases
where worse concentrations would have been expected.
1.3.2 Pesticide Characteristics
Aldicarb is a nematicide, acaricide, and a systemic
insecticide. Its environmental fate is dominated by two
factors: the fact that it forms two toxic daughter products,
and its high mobility in soils. Degradation of the toxic
residues of the compound is of intermediate duration compared
to other pesticides.
Aldicarb is a white crystalline solid which is incorporated
into soil as a granule containing either 10% or 15% active
ingredient. In order to be effective, it must dissolve in
16
-------
TABLE 1.2 CASES CONSIDERED FOR SATURATED ZONE MODELING
Influencing Factors
Cases Considered for Each Factor
Aquifer geometry
1) Unconfined Floridan Aquifer
2) Unconfined surficial aquifer
3) Leaky two-aquifer system
Aquifer properties
1) Worst case
2) Intermediate case
Well distance
1) 300 ft
2) 1000 ft case
Well depth
1) Shallow
2) Deep
Well rate
1) High rate
2) Low rate
17
-------
00
SURFICIAL
AQUIFER
UNCONFINED
UNCONFINED
FLORIDIAN
AQUIFER
TWO-AQUIFER
SYSTEM
AVERAGE
WORST
AVERAGE
WORST
AVERAGE
WORST
DECAY
NO. DECAY
DECAY
NO DECAY
DECAY
NO DECAY
DECAY
NO DECAY
DECAY
NO DECAY
DECAY
NO DECAY
HIGH
LOW
HIGH
LOW
HIGH
LOW
Figure 1.5 Matrix showing saturated zone scenarios simulated.
-------
water. Once this happens in soils, the compound begins to
transform and degrade.
The current theory is that aldicarb is fairly rapidly oxidized
to aldicarb sulfoxide which in turn is more slowly oxidized to
aldicarb sulfone. Concurrently, these three carbamates are
transformed, by hydrolysis to corresponding oximes. Hydrolysis
is a chemical reaction in which water breaks up an organic
molecule (RX), such as aldicarb, by breaking a carbon-X bond
and replacing it with OH from the water molecule.
These products of hydrolysis are far less toxic than aldicarb,
its sulfoxide or its sulfone and are of little environmental
concern. A schematic of these processes is shown in Figure
1.6.
The investigation of the rates and coefficients used to
describe fate and transport was also divided between the
unsaturated and saturated zones as described below.
1.3.2.1 Unsaturated Zone
PRZM was modified in this study so that the transformation of
aldicarb to its daughter products could be simulated in the
unsaturated zone. Therefore, the rates k^ through k5 had to
be estimated for each transformation/degradation pathway and
adsorption partition coefficients had to be estimated for
aldicarb, its sulfoxide and its sulfone. This was
accomplished by taking rates from the literature and
regressing the values on levels of environmental variables
such as pH, temperature, soil water content and organic water
at which each study was conducted. Then, by using local soil
environmental conditions for each soil order, rates were
predicted for each scenario. Adsorption partition
coefficients were determined using average values of Koc found
19
-------
Aldicarb
Aldicarb Sulfoxide
Aldicarb Sulfone
CH3
CH,S - C - CH
3 i
O
II
NOCNHCH.
O CH,
II I 3
CH-jS - C - CH
CH.,
0 *2 O CH.
II II l 3
NOCNHCH3- CH-jS - C - CH
O CH-
NOCNHCH.
(Hydrolysis)
(Hydrolysis)
to
O
Nontoxic Oximes and Nitrites
Figure 1.6 Schematic of aldicarb environmental chemical Pathways
-------
in the literature with localized values of soil organic
carbon.
The values of the rates and coefficients were then used in a
modeling verification study. Four simulations were performed
at three different sites in Florida for which residue sampling
data was available. This data was collected in 1983 and 1984.
Simulations indicated that rates and coefficients used for
spososols (i.e., flatwoods soils) were well estimated by the
multiple regression equations derived from the literature
data. Simulations of entisols (i.e., ridge soils), however,
indicated that degradation rates used in this study were
probably too fast.
1.3.2.2 Saturated Zone
For saturated zone simulations the CFEST model was used.
CFEST simulates the transport of a single constituent which is
subject to adsorption and degradation. Therefore, a single
first order decay rate had to be derived for simulation
purposes. Since the principle degradation pathway in ground
water is thought to be hydrolysis, literature values of
hydrolysis rates were analyzed. Using values of these
hydrolysis rates at various levels of pH and temperature,
activation energies for the hydrolysis reaction for aldicarb,
aldicarb sulfoxide and aldicarb sulfone were calculated using
the Arrhenius equation. Then using localized values of
aquifer pH and temperature in Florida, first-order degradation
rates were determined for specific scenarios. Because the
predicted rates for sulfone were fastest, these were used for
scenarios in which decay was simulated. As depicted
previously in Figure 1.5, scenarios were also run using
no-decay of the contaminant. Adsorption partition
coefficients were set to zero in the saturated zone
21
-------
simulations as no organic matter is thought to be present in
aquifer materials.
In conducting the literature survey on degradation rates it
was discovered that other porous media properties may affect
aldicarb degradation. Recently much higher disappearance
rates for aldicarb sulfoxide and aldicarb sulfone have been
reported when incubated under anaerobic versus aerobic
conditions. Under anaerobic conditions, disappearance of
these two compounds may be from 8 to 100 times faster than in
the same soil under aerobic conditions. The current theory is
2+
that the presence of Fe in solution catalyzes these
reactions.
Very little specific data is available on redox potential or
2+
dissolved ferrous iron (Fe ) in Florida. The most common
2+
form of iron in the ground water is ferrous iron (Fe ),
however, and ground water with a pH between 6 and 8 can carry
as much as 50 mg/1 of ferrous iron at equilibrium. The
occurrence of 1.0 to 10 mg/1 of iron in ground water is
common. Dissolved iron in the ground water in Florida varies
from as little as 0.01 mg/1 to as high as 20 mg/1. In general
the surficial aquifer has higher iron concentrations because
the aquifer materials contain more iron-bearing minerals than
the predominantly limestone Floridan Aquifer (Irwin, 1984,
personal communication).
If the transformation of aldicarb is indeed catalyzed by
ferrous iron, the process would be important in the ground
water in Florida. At this time, the process is so poorly
understood it was not included in the quantitative analysis of
the pesticide parameters for modeling.
22
-------
1.4 SUMMARY OF RESULTS
This section overviews the important results of the
unsaturated and saturated zone modeling and calculation of the
resulting well water concentrations. More detailed
information is contained in Section 4 of this report.
Important to this discussion is the fact that the unsaturated
and saturated zone modeling efforts were essentially pursued
independently. Fourteen years of meteorologic data were used
to drive the unsaturated zone model. The resulting output was
a time series of pesticide loadings at the top of the
saturated zone (i.e., ground water table). Therefore, the
unsaturated zone simulations were fully dynamic, both for flow
and transport. For PRZM, which runs very economically, this
type of simulation was feasible for a large number of
scenarios.
With CFEST, steady-flow, unsteady contaminant transport
simulations were used, for several reasons. First, a
two-dimensional and sometimes three-dimensional geometry was
needed which required far greater time and money resources
than for the one-dimensional PRZM. Long, unsteady flow
simulations were impractical, given the large number of
scenarios to be run. Second, sensitivity simulations using
CFEST showed that transient flow conditions caused by widely
fluctuating recharge rates from the unsaturated zone were
minor and short-lived. This conclusion allowed for the
convenient use of steady flow defined by the regional
hydraulic gradient and the assumption of a fixed depth from
the land surface to the saturated zone.
23
-------
1.4.1 Unsaturated Zone
The resulting annual pesticide loadings from the sixteen final
unsaturated zone scenarios were analyzed in several ways.
First, geometric mean pesticide loadings from all scenarios
were subjected to one way analysis of variance (ANOVA) to
determine if any substantive differences occured between the
mean loadings of the various scenarios. These analysis showed
that of the means of the sixteen scenarios simulated the only
significant differences were among thick ridge soils (i.e.,
entisols and ultisols), thin ridge soils, and flatwoods soils.
Visual inspection of the frequency distribution of annually
leached pesticide loads confirmed this. Thus, the output of
the sixteen scenarios run were condensed into three scenarios;
"ridge" areas with thick unsaturated zones, "ridge" areas with
thin unsaturated zones, and "flatwoods" areas.
By condensing the information in the sixteen scenarios down to
three, better probability estimates are obtained. Figures 1.7
through 1.9 show the three resulting frequency distributions
for the thick unsaturated zone ultisols and entisols, thin
unsaturated zone ultisols and entisols, and the alfisol and
spodosol groupings.
These loads, however, are leached only from the treated band.
Much of the area in the typical citrus grove is untreated,
with no resulting leached load. Therefore, the loads from
Figures 1.7 through 1.9 must be reduced by the ratio of
treated to total area in the grove. This ratio is
approximately 0.32 to 1 for ridge citrus and 0.20 to 1 for
double bedded flatwood citrus.
Table 1.3 summarizes the pesticide loads for the 90, 50 and 10
percentile exceedance probabilities for the three final
24
-------
RIDGE SOILS
THICK UNSATURATED ZONE
0
UJ
Q
UJ
ui
u
X
UJ
m
o
o
o
UJ
o
o
m
<
to
o
cc
a.
0.00
1.E
-5 1.E-4 1.E-3 1.E -2
PESTICIDE LOAD IN KG/HA
1.E-1
1.E 0
Figure 1.7
Frequency of annual quantity of pesticide
leached to the saturated zone from thick
Entisols and Ultisols.
25
-------
UJ
Q
UJ
Ul
O
X
tu
CO
Q
UJ
o
5
z
H
X
m
<
o
O
cr
Q.
RIDGE SOILS
THIN UNSATURATED ZONE
0.00
1.E
-5 1.E-4 1.E-3 1.E -2
PESTICIDE LOAD IN KG/HA
1.E-1
1.E 0
Figure 1.8 Frequency of annual quantity of pesticide
leached to the unsaturated zone from thin
Entisols and Ultisols.
26
-------
FLATWOODS SOILS
SPODOSOLS AND ALFISOLS
1.E-5 1.E-4 1.E-3 1.E -2
PESTICIDE LOAD IN KG/HA
1.E-1
1.E 0
Figure 1.9
Frequency of annual quantity of pesticide
leached to the saturated zone from Spodosols
and Alfisols.
27
-------
TABLE 1.3 SUMMARY OF PESTICIDE LOADINGS PER UNIT CITRUS BLOCK
AREA FOR THE THREE FINAL UNSATURATED ZONE SCENARIOS
Exceedance Probability
Scenario 0.90 0.5U 0.10
Alfisols and spodosols 1.8E-5 4.0E-4 2.2E-3
Thick ultisols and entisols 6.4E-6 4.2E-4 3.2E-3
Thin ultisols and entisols 1.9E-3 1.3E-2 9.6E-2
28
-------
unsaturated zone scenarios. These loads have been multiplied
by the above ratios to yield loads per unit citrus block
instead of unit treated area. A unit citrus block is defined
as the area alloted to each tree according to the spacing in
the grove. The table reveals that lowest loadings are in
general associated with alfisols and spodosols. For this
scenario there is only a 10% probability that the pesticide
load leached to ground water will exceed 0.002 kg/ha. Loads
from thick entisols and ultisols exceed those for the alfisols
and spodosols slightly. The highest loadings emanate from
thin unsaturated zone entisols and ultisols. There is a 10%
chance that loads to ground water will exceed 0.1 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 is 5.6
Kg/ha or 5 Ib/acre. Increasing or dereasing the load in any
of these scenarios by a ratio 'x', would result in an increase
or decrease in the load by the same ratio. 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 would also double, from 0.096 Kg/ha to
0.19 Kg/ha.
Also of interest is the quantity of aldicarb and that of its
two toxic metabolities 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 sulfoxide 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. The spodosols and alfisols, the ratio
is roughly the same, 0.19 to 1.
Obviously, since these transformations are kinetically
controlled the quantity of aldicarb and its toxic metabilities
29
-------
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 one, the more
aldicarb and aldicarb sulfoxide will appear in the leachate.
1.4.2 Saturated Zone
A summary of the results of the saturated zone scenarios are
shown in Table 1.4. This table shows the 90, 50, and 10%
probabilities of exceeding the given relative concentrations.
These "relative" concentrations are the simulated
concentrations at the well divided by the initial
concentration in the recharge water and are dimensionless.
Because a unit load (1 kg/ha) was used for the input in all
model simulation runs, all the well-water concentrations are
expressed in this manner.
In general the surficial aquifer cases show the highest
relative concentrations followed by the worst cases for the
Floridan Aquifer and then the average cases for the Floridan
and the two-aquifer cases. The surficial aquifer has high
relative concentrations for two reasons. The relatively thin
aquifer has less water for the pesticide to disperse into.
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 primarly due to the high ground-water
velocities which leave little time for decay or dispersion.
The two-aquifer scenarios and the average Floridan scenarios
have slower ground-water velocities and thick aquifers that
dilute the initial pesticide concentrations.
30
-------
TABLE 1.4 RELATIVE CONCENTRATION VALUES FOR
INDICATED SATURATED SCENARIOS AT THREE
EXCEEDANCE PROBABILITIES
Scenario
Floridan Aquifer - worst Case
91 meters to well
Deep well, no decay
Decay
Shallow well, no decay
Decay
300 meters to well
Deep well, no decay
Decay
Shallow well, no decay
Decay
Exceedence Probability
.90 .50 .10
6.4E-4
8.1E-5
4.0E-3
1.3E-4
1.1E-3
3.3E-5
4.0E-3
3.3E-5
4.4E-3
3.7E-4
l.OE-2
5.8E-4
5.0E-3
1.4E-4
8.5E-3
1.8E-4
8.6E-3
1.7E-3
1.5E-2
2.6E-3
8.4E-3
6.2E-4
1.2E-2
6.4E-4
Floridan Aquifer - average case
91 meters to well
Deep well, no decay 5.7E-4 3.3E-3 7.1E-3
Decay 2.5E-6 1.8E-5 4.5E-3
Shallow well, no decay 6.3E-4 3.6E-3 7.5E-3
Decay 2.5E-6 1.8E-5 4.4E-5
300 meters to well
Deep well, no decay 1.6E-4 7.3E-4 3.0E-3
Deep well, decay 9.2E-9 3.1E-8 4.6E-8
Shallow well, no decay 9.6E-5 4.1E-4 1.9E-3
Decay 2.6E-9 8.7E-9 1.3E-8
Surficial Aquifer - worst case
91 meters to well
High pumping rate, no decay 1.1E-2 5.0E-2 7.3E-2
Low pumping rate, no decay 2.8E-2 1.3E-1 1.7E-1
Both, decay 2.1E-4 1.1E-3 3.3E-3
300 meters to well
High pumping rate, no decay 5.4E-3 2.6E-2 5.9E-2
Low pumping rate, no decay 6.7E-3 4.7E-2 l.OE-1
Both, decay 4.0E-6 1.8E-5 3.3E-5
Two-aquifer System, worst case
91 meters to well
High pumping, no decay 1.5E-4 6.9E-4 1.7E-3
Low pumping, no decay 3.4E-4 1.1E-3 2.1E-3
Both, decay 1.3E-5 6.5E-5 2.7E-4
31
-------
Other trends can be observed from the results. As expected,
the cases modeled with decay resulted in relative
concentrations much lower than those modeled with no decay.
The greatest difference is in cases where the ground-water
velocity is very slow. On the other hand, in cases like the
Floridan worst case the relative concentrations are not much
lower when decay is simulated.
The pumping rate of the well has little influence. This is
especially true in cases where the regional gradient is
dominant and where decay is simulated but even in cases where
the well-induced gradient is dominant like the surficial worst
cases the relative concentrations do not differ by much. In
the Floridan worst cases the greatest difference due to
pumping rates is 3%. The surficial worst cases with no decay
show differences as much as 50% between the high and low
pumping rates, but even in this case, the trends are very
similar.
Concentrations are higher in wells only 91 m away from the
source area. In the cases with no decay, the relative
concentrations at the 300 m well are smaller because there is
more dispersion. The Floridan worst cases show only a
difference of about 20% between the 91 m and the 300 m cases.
For the average Floridan and surficial worst cases, relative
concentrations differ by as much as 70%. In the cases with
decay, the 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.
32
-------
1.5 COMBINED RESULTS OF UNSATURATED AND SATURATED ZONE
MODELING
1.5.1 Physical Overlap of Saturated and Unsaturated Zones
The results from the unsaturated zone modeling and the
saturated zone modeling efforts were combined to evaluate
expected aldicarb concentrations in drinking water wells. The
unsaturated zone scenarios cover two physiographic areas: 1)
areas where entisols and ultisols occur, and 2) areas where
spodosols and alfisols occur. Even though for the entisols
and ultisols both thin and thick unsaturated zones are
considered, thick and thin areas occur randomly and therefore
no attempt is made to separate them physically. The saturated
zone is comprised of three physical aquifer systems that were
modeled: 1) the unconfined Floridan Aquifer, 2) the
unconfined surficial aquifer and 3) the multi-aquifer system.
Figure 1.10 shows the areas of overlap of the unsaturated and
saturated zone scenarios. The entisols and ultisols (ridge
soils) overlap all three of the saturated aquifer system. The
spodosols and alfisols (flatwoods soils) overlap with the
surficial aquifer and the multi-aquifer systems but not with
the unconfined Floridan Aquifer. The results of the two
modeling studies were combined in all cases where overlap
occured.
1.5.2 Results
The complete set of results showing well-water concentrations
at the 10 and 50 percent exceedance values for all the
scenarios are shown in Appendix E. These results are
33
-------
Jacksonville
Atlantic Ocean
FLATWOODS
SAILS.,
LEGEND
FLORIDIAN AQUIFER - UNCONFINED
MULTI - AQUIFER SYSTEM
SURFICIAL AQUIFER - UNCONFINED
Figure 1.10 Overlap of unsaturated and saturated
model areas.
34
-------
summarized in Table 1.5 which reports the highest calculated
concentration for each catagory shown. In general the well-
water concentrations are very low. The highest concentration
is 6.5 parts per billion (ppb) in a scenario without decay.
This is just at the detection limit of 5 to 6 ppb for total
toxic residues (Rao, 1984, personal communication). The
highest concentrations for scenarios with decay are in the
order of 10 ppb.
The well-water concentrations show three general trends.
First, the highest concentrations are associated with thin
ultisols and entisols, followed by the thick ultisols and
entisols and the alfisols and spodosols. Second, in the
saturated zone, the worst cases for the surficial aquifer
generally have the highest well water concentrations closely
followed by the unconfined Floridan worst cases and finally
the worst cases of the two-aquifer system and average Floridan
cases. Third, within each set of saturated zone scenarios,
the highest concentrations result from the no decay simulation
with the well 91 m (300 ft) from the source area, followed by
no decay simulated with a well at 300 m (1000 ft). Next are
the decay scenarios with the well at 91 m. The smallest
concentrations result from the decay simulations where the
well is 300 m from the source.
1.5.2.1 Surficial Aquifer
Combined with the thin entisols and ultisols, the surfical
aquifer worst cases generally produce the highest
concentrations. The highest value is for worst case hydraulic
parameters, no pesticide decay, with a shallow well at 91 m
distance from the source area, pumping at a low rate, having
a well water concentration of 6.5 parts per billion (ppb).
With decay the concentrations drop one to three orders of
magnitude. When the well is 300 m from the source, the
35
-------
TABLE 1.5 HIGHEST CALCULATED ALDICARB CONCENTRATIONS (in ppb)
.IN THE GIVEN COMBINED UNSATURATED/SATURATED CATAGORIES
Floridan
Worst Cases
Floridan
Average
Surf icial
Worst Cases
Two-
Aquifer
Worst
Cases
300
ft
1000
ft
300
ft
1000
ft
300
ft
1000
ft
300
ft
No
Decay
Decay
No
Decay
Decay
No
Decay
Decay
No
Decay
Decay
No
Decay
Decay
No
Decay
Decay
No
Decay
Decay
Flatwoods
Soils
No
Overlap
No
Overlap
l.SxlCf1
2.9xlO~s
9.0xlO~2
2.7xlO~5
2.7x!0"S
3.5xlO~*
Thick
Ridge Soils
2.3xlO~2
4xlO~8
1.8xlO~2
9.8xlO~"
2.2xlO~2
1.3xlO~"
3.8xlO~3
1.4x!0"7
2.2xlO~'
4.2xlO~s
1.3X10"1
4.2xlO~5
3.9xlO~3
5.1x10'"
Thin
Ridge Soils
6.9X10'1
1.2X10'1
5.5xlO~a
2.9xlO~2
S.SxlO'1
3.9xlO~3
2.6X10"1
4. xlO~6
6.5
1.3X10"1
3.9
1.3xlO~3
1.2X10"1
1.5xlO~2
36
-------
surficial cases without decay still show significant
contamination, but when decay is simulated the concentrations
are in the order 10 ppb or less. The other unsaturated
zone scenarios show the same trends as the thin ridge soils
when combined with the surficial aquifer worst cases with even
lower concentrations. The highest concentrations are 2.17 x
10 and 1.49 x 10 for the thick ridge soils and the
flatwoods soils, respectively.
1.5.2.2 Floridan Aquifer Worst Cases
Well-water concentrations in the Floridan worst cases are not
as high as those in the surficial worst cases except when
decay is modeled with the drinking water well at 300 m. When
evaluating these results with no decay, it is necessary to
realize that these concentrations could be much higher for
wider source areas (see Section 4.2.4). The highest
concentration in the scenarios without decay is 6.9 x 10 ppb
with the well at 91 m (300 ft) and 5.5 x 10 ppb with the
well at 300 m (1000 ft). The scenarios simulated with decay
are generally one order of magnitude less than the same
scenario without decay.
1.5.2.3 The Floridan Aquifer Average Cases
The well-water concentrations in the Floridan average cases
vary considerably between scenarios modeled with decay and
without decay. The concentrations calculated for scenarios
with the well 91 m (300 ft) away from the source area and no
decay are as high as 0.65 ppb. In the cases with decay the
concentrations drop over two orders of magnitude when the well
is 91 rn from the source and four to five orders of magnitude
when the well is 300 m from source.
37
-------
1.5.2.4 Two-Aquifer System
The scenarios modeled with the two-aquifer system exhibit some
of the lowest concentrations. The highest value is .12 ppb.
With decay simulated the concentrations drop about an order of
magnitude.
1.6 CONCLUSIONS AND RECOMMENDATIONS
The results presented in the foregoing section are based upon
model representations of generalized, regional unsaturated and
saturated zone modeling scenarios. Such a study, of
necessity, overlooks specific "special situations" which may
occur within these regions. To ameliorate the effects of such
over-sights, the general approach was to attempt to look at
"worst case" values of parameters which could be selected to
describe these generalized scenarios and to adopt other
"average case" values when worst case values gave results
indicative of high contamination levels. This was not done in
all cases. High contamination levels generally were not
found, however, even under "worst case" conditions.
Inevitably, higher concentrations that those predicted here
will be found at sometime in someone's drinking water well.
The primary intent of this concluding section is to highlight
the caveats of this simulation study yielding an indication of
the types of special cases under which higher than predicted
concentrations might be found. Secondarily, recommendations
on how to avoid these high concentrations will be stated.
38
-------
1.6.1 Unsaturated Zone
In terms of the unsaturated zone simulations, several caveats
are in order. First, a limited period of meteorological
record (14 yrs) was used for the simulations. During this
period, 1965 to 1978, the mean annual rainfall was below
normal. If a longer period had been used, better estimates of
leaching probabilities would have been obtained and more
extreme years might have been included, giving better tail
probabilities at the high end of the leaching frequency
distributions. The longer period was not used because pan
evaporation data was not available and the algorithms in PRZM
used to estimate pan evaporation from mean daily temperature
consistantly underestimated the magnitude of that time series.
Mean values of state variables such as soil field capacity
water content, pH, organic carbon, etc. were used as
representative of the unsaturated zone soils. Estimates of
variability were available for these parameters and, in
retrospect, it may have been appropriate to chose values of,
say, mean plus or minus one standard deviation in the
direction of the worst case, as representative. It should be
noted that most of the soil characterization samples available
were not taken in citrus groves but in areas of natural
vegetation. Soils in citrus groves might have higher organic
carbon due to maintanence of the upper profile to promote root
growth and thus have higher water holding capacity. These
conditions would be conducive to stronger pesticide retention
and more rapid decay. Therefore, the best estimate parameters
for soils having natural vegetation might be biased toward
higher pesticide leaching then cultivated citrus soils.
The pesticide parameters, derived from regression on those
environmental state variables, were chosen based on the
39
-------
regression estimate rates and constants using mean values of
the state variables. Thus, best estimate values of soils and
pesticide parameters were used as opposed to "worst case"
values. The fact that the estimated pesticide parameters do
not represent worst-case values is demonstrated by the too
rapid simulation of pesticide degradation in 1983 and 1984 at
the Lake Hamilton and Davenport sites in the PRZM verification
runs. Pesticide leachate loadings from these "ridge" soils
would probably be greater than that represented in the model
simulations due to the relative longevity of the actual
field-monitored residues. Unfortunately, the 1984 Davenport
data was not available early enough in the study to affect the
selection of pesticide parameters for modeling scenarios.
Probably the only crop parameter that is crucial in these
simulations is the rooting depth. Representative values (60
cm for spodosols and alfisols, 150 cm for ultisols and 240 cm
for entisols) were used for each soil order.
Vertical hydraulic transmission rates for spodosols and
alfisols were based on only one data set for a representative
soil type of each soil order. The confidence to be placed in
this estimate is low and, unfortunately, the pesticide
loadings simulated for these soil types are heavily dependent
upon this transmission rate. Therefore, less confidence
should be placed in these loading estimates versus those from
the entisols and ultisols based on hydraulic considerations
alone. For the ridge soils, the "unrestricted drainage"
option was used in PRZM which does not require estimation of
this parameter.
The inability to use PRZM to simulate irrigation methods other
then the overhead types is unfortunate. Loading estimates
were not produced for low-volume spray or flood methods.
Overhead irrigation did cause an increase in pesticide
40
-------
leaching in our simulations although this increase did not
appear to be statistically significant. It is certain,
however,that irrigation water, even when properly applied,
increases the opportunity for leaching by natural rainfalls.
For those methods (i.e., low-volume spray and trickle) where
water is supplied to the tree over a relatively small area,
leaching due to irrigation will occur if the water application
rate exceeds the combined capacity of the tree roots to
extract and the soil to store the water. Concerning the
overall effect, however, this enhanced leaching may be
mitigated by the fact that it is occurring over a smaller
area.
At this point, the effects of flood irrigation on the leachate
load of pesticide are uncertain. It is evident that saturated
conditions in the profile increases downward water movement.
However, during the drainage cycle, pesticide-laden water may
move laterally, decreasing the opportunity for leaching, at
least from the field area. In addition, capillary rise of
water induced by seepage irrigation may actually cause an
upward movement of chemical and subsequent removal by plant
roots. Some two-dimensional modeling of these systems should
be done at some point, if possible.
Simulation of the application of water via irrigation systems
for freeze protection was not accomplished. Extensive changes
required PRZM and each of appropriate input data to properly
simulate this process caused it to be eliminated from
consideration. Application of large quantities of water for
this purpose would obviously move large quantitities of
aldicarb through the system. It is interesting to note that
for the Floridan worst case saturated zone scenario with decay
and a shallow well 91 m (300 ft) downgradient of the source,
if all of the 5.61 Kg/ha application of aldicarb
catastrophically entered the saturated zone, the resulting
41
-------
simulated peak well water concentration would only be 8.2 ppb.
For the surficial aquifer worst case, under the same
conditions, the simulated peak concentration would be 8.4 ppb.
For 300 m (1000 ft) downgradient distance, the peak
concentrations drop to 2.0 and 0.08 ppb for the Floridan and
surficial aquifers, respectively.
Based on results of the unsaturated zone simulations and
subsequent discussion, there are a number of recommendations
which can be made for practices which will tend to diminish
loadings from the unsaturated zone:
1) Avoid the application of aldicarb in ridge areas where
there is a thin unsaturated zone (i.e., a high ground
water table).
2) Avoid the application of irrigation water in the
treated band in quantities that would cause movement
of water and pesticide past the crop root zone.
3) Avoid the application of irrigation water for freeze
protection after the application of aldicarb to a
grove.
1.6.2 Saturated Zone
In the saturated zone modeling, emphasis was placed on
modeling the "worst" case scenarios, for the three general
aquifer configurations. Even so, the well-water
concentrations could be underestimated for several reasons.
First, assumptions were made for the simulations (such as
distribution of pesticide loads) that affect the predicted
concentrations. Second and more important are the situations
42
-------
that were not simulated due to limitations of the model or the
limitations in the regional scope of the study.
Because the unsaturated and saturated zone modeling were done
concurrently, the output of the PRZM runs was not directly
used as input for CFEST. Assumptions were made concerning the
timing of the loads from the unsaturated zone and the
distribution of that load based on preliminary results. A
standardized leaching event of four months duration was used
over which the annual pesticide load was evenly distributed.
For most of the scenarios this was a worst case estimate since
four months was the minimum time required to leach over 90% of
the pesticide on the average. Even in an anomolous year where
most of the pesticide was leached in one month, the
concentrations would go up by only a factor of four.
Likewise, assuming uniform loading throughout the four months
disregards the high spikes. But this too would not
significantly change the pesticide concentrations as shown by
the transient sensitivity simulation where the actual loadings
from a PRZM output resulted in concentrations that were higher
by only 10% in one year and lower in another year.
Sensitivity simulations demonstrated the importance of the
size of the pesticide source area on well-water
concentrations. As the time required to travel under the
source area increases, the ground-water will receive more
leached pesticide residues from the unsaturated zone. In
cases with low decay or no decay, this can result in higher
pesticide concentrations at the well. For the extreme case
where there is no decay and no dispersion, if the width of the
source area were to double, well-water concentrations would
also double.
In the scenario simulations the source area was only
considered to be up-gradient of the well. The sensitivity
43
-------
simulations demonstrated that concentrations would increase as
much as four times in situations where the well dominates the
regional gradient. Combined with the other factors discussed
above, concentrations could be as much as an order of
magnitude higher than those predicted by the model.
In this study, the worst set of the hydrogeologic parameter
values represented some of the more extreme situations in the
ground-water environment. Still the solution scheme used in
CFEST assumes a porous media where these hydrogeologic values
are averaged over the aquifer. The more likely situation,
especially in the limestone aquifers, is that solution
cavities riddle the aquifer creating a situation where
fracture flow is dominant. Large fractures or solution
cavities may have velocities many times larger than those
measured, while the surrounding rock may be more nearly
impermeable.
This is a situation that can not be simulated by CFEST and is
a current topic of active research. This type of fracture
flow could lead to concentrations much higher than those
simulated.
Another potentially dangerious situation is that of a direct
conduit to the aquifer. These can occur naturally in the form
of a sink hole or can be man-made in the form of a leaky well.
When the potentiometric surface of the lower aquifer is lower
than the water table either a sink hole or a leaky well allows
contaminated water to flow directly into the lower aquifer.
These localized situations were not modeled. A best estimate
of pesticide concentrations under these conditions would be to
use the results from scenarios simulated in an unconfined
aquifer.
In all the scenarios the hydraulic gradient was that of the
44
-------
combined regional gradient plus the drinking water well.
Florida has areas of heavy ground-water withdrawals for both
irrigation and municipal supply that depend on well fields.
The effect of these localized induced gradients due to many
wells pumping together was not considered. Much higher
gradients would make ground water velocities much higher and
therefore allow less time for decay.
In summary, all situations were not simulated due to model
limitations, limitations in the regional scope of the study,
or assumptions made for the purposes of the modeling.
Considering the effects of these assumptions in a composite
sense, well-water concentrations could increase by at least an
order of magnitude.
Based on the results from the saturated zone simulations and
the preceeding discussion, there are a number of
recommendations which can be made that will tend to reduce
expected well water concentrations.
1) Avoid application of pesticide up gradient of wells
that intersect major solution cavities in areas where
the aquifer is unconfined or where direct linkages
exist.
2) Avoid application near wells in the surficial aquifer
that have high pumping rates and are closely
surrounded by citrus groves.
3) Avoid using shallow wells where the localized induced
gradient is very large.
4) A well tapping a confined aquifer will be safer than
unconfined aquifer.
45
-------
5) Deep wells that draw water from a large section of
aquifer will have lower concentrations than similar wells
that are shallower.
46
-------
SECTION 2
FLORIDA CITRUS GROWING ENVIRONMENT
Site-specific information about the citrus growing region was
gathered to delineate a number of scenarios among which it was
felt the fate and transport of the chemical would
significantly differ. The information gathering was divided
into two topical areas: 1) the surface and unsaturated zone
and 2) the saturated zone. These two zones are discussed in
the following subsections in terms of the important physical
characteristics that distinguish unique scenarios for the
purposes of modeling.
2.1 SURFACE AND UNSATURATED ZONE
Citrus is grown throughout most of central and southern
Florida. The map in Figure 2.1 shows the counties of the
state ranked according to the acreage of bearing and
non-bearing grapefruit and oranges grown. The top ten ranked
counties are highlighted. The preponderance of fruit grown is
in the "ridge" area, a region of rolling sand hills running
approximately north-south in the center of the state, and the
Indian River area, the smaller shaded region on the east
47
-------
ALABAMA
Jacksonville
Atlantic Ocean
1-PolJc
2-Laka
3-St. Lucle
4-Zndian River
5-Orange
6-Hardee
7-Martin..
8-H111eborough
9-Highlands
10-De Soto
11-Pasco
12-Hendry
13-Osceola
14-Brevard
15-Manatee
16-Palm Beach
17-Marion
18-Volusia
19-Collier
20-Okeechobee
21-Lee
22-Hernando
23-Seminole
24-Charlotte
25-Glades
26-Pinellas
27-Putnam
28-Broward
29-Sumter
30-Sarasota
31-Citrus
32-Plagler
33-Alachua
34-St. Johns
Figure 2.1 Florida counties ranked by citrus acreage.
48
-------
coast. These are rankings as of January 1, 1982. A severe
freeze occurring in December 1983 killed substantial acreages
in the northern portion of the state. Post freeze acreages
were not available. Table 2.1 gives the acreage by county.
Over half the citrus is grown in the top five ranked counties;
over 75% is grown in the top ten ranked counties.
2.1.1 Climate
The driving force behind pesticide transport in soils is
primarily water movement downward in the profile; whether it
occurs from natural rainfall or irrigation. Rainfall is high
in the Florida citrus growing region, averaging from 132 to
162 cm/yr (52 to 64 in./yr, .Figure 2.2). Pan evaporation
ranges from 117 to 132 cm/yr (46 to 52 in./yr , Figure 2.3).
However, 65% of the rainfall on the average occurs during the
months of June, July, August and September. Evapo-
transpiration demands require that irrigation be used
extensively in March, April and May to supplement natural
rainfall and 50% of the irrigation water applied is 'applied
during these months (Harrison, 1984, personal communication).
Table 2.2 shows the monthly distributions of rainfall and
evapotranspiration for Lake Alfred (located in the ridge area)
and Fort Pierce (flatwoods area)(see Section 2.1.2). The
differences indicate that, on the average, water deficits
occur in the Lake Alfred area in fall, winter and early
spring, while at Fort Pierce, deficits generally occur in
spring.
Temperatures in the area are normally mild, however freezes
may occur. These conditions may affect chemical transport as
irrigation systems are often used to apply large quantities of
49
-------
TABLE 2.1 SUMMARY OF BEARING AND NONBEARING
CITRUS (ORANGES AND GRAPEFRUIT)
BY COUNTY AS OF JANUARY 1, 1982
County
Acreage
County
Acreage
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
Polk
Lake
St. Lucie
Indian River
Orange
Hardee
Martin
Hillsborough
Highlands
De Soto
Pasco
Hendry
Osceola
Brevard
Manatee
Palm Beach
Marion
125331
104307
71057
60387
42727
42136
38714
34807
34210
32902
31636
29962
16550
14875
12980
12275
10840
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
Volusia
Collier
Okeechobee
Lee
Hernando
Seminole
Charlotte
Glades
Pinellas
Putnam
Broward
Sumter
Sarasota
Citrus
Flagler
Alachua
St. Johns
8786
7561
. 6668
6254
5819
5790
5757
3933
2245
2129
1661
1560
1258
1144
116
96
95
TOTAL 776,568
11
Source: Florida Citrus Mutual, 1984.
Acreage for 1982-83 crop season was 776,803 with 71,053 additional
acres in specialty fruit according to the Florida Dept. of
Agriculture and Consumer Services, -1984.
50
-------
ALABAMA
Jacksonville
Atlantic Ocean
52
Figure 2.2 Annual precipitation in inches (1 inch = 2.54 cm).
Source: U.S. Dept. Commerce, 1972.
51
-------
ALABAMA
Jacksonville
Gulf of Mexico
..'*
Figure 2.3 Annual lake evaporation in inches (1 inch =
2.54 cm). Source: Kohler et al., 1959.
52
-------
TABLE 2.2 MEAN MONTHLY RAINFALL AND EVAPOTRANSPIRATION AT TWO
CITRUS GROWING LOCATIONS IN FLORIDA (cm)
LAKE ALFRED
FORT PIERCE
Month
J
F
M
A
M
J
J
A
S
0
N
D
TOTAL
Rainfall1
4.9
6.4
7.9
8.3
8.4
20.6
19.3
19.3
9.9
4.7
4.7
4.6
133.1
ET2 .
5.3
5.8
5.8
9.9
12.7
15.0
14.5
12.4
10.2
7.4
7.4
5.8
121.1
Diff
-0.4
0.4
0.0
-1.6
-4.3
5.5
4.8
6.9
-0.3
-2.7
-1.2
-1.2
Rainfall
5.4
5.0
7.0
9.8
13.2
19.0
17.1
16.9
23.4
21.2
6.8
5.2
150.0
ET2
5.3
6.6
9.1
11.4
13.5
11.2
12.4
12.2
10.2
9.1
6.8
5.3
113.3
Diff
0.1
-1.6
-2.1
-1.6
-0.3
7.8
4.7
4.7
13.2
12.1
0.0
-0.1
Rainfall Depths are approximate.
Source: U.S. Dept of Commerce, 1968
"Source: Harrison, 1984.
53
-------
water for freeze protection. The annual mean number of days
having temperatures of 32 deg. F (0 deg.C) or less are shown
in Figure 2.4. Citrus in the upper portion of the growing
area may experience close to 5 days of freezing temperatures
annually. Almost all citrus could expect at least 2 freezing
days annually.
2.1.2 Soils
There are four orders of soils in Florida on which citrus is
extensively grown. These are the entisols, ultisols, alfisols
and spodosols. The approximate locations of these groups are
shown on the accompanying map (Figure 2.5). The entisols tend
to occur in the central "ridge" area with the major occurrence
of ultisols being in the northwestern portion of this area.
The spodosols occur along the middle eastern and western
coasts of the peninsula, adjacent to the ridge, in what is
known as the "flatwoods" area. Two major areas of alfisols
occur; one running north-south which divides the eastern coast
flatwoods area and the other immediately east and south of the
ridge area. Naturally, there are isolated pockets of these
soil types occurring everywhere.
The entisols are soils which are loose, incoherent and are
developed from thick beds of marine sediments. The
characteristic of these soils is a distinct lack of profile
development, the only exception being slightly higher organic
matter in the surface soil. They are strongly acid, are well
drained to excessively drained, and their water holding
capacities are low (Anderson, 1981).
The ultisols are highly weathered soils with sandy surface
horizons but loamy to clayey subsurface horizons. In almost
54
-------
10
Atlantic Ocean
2
Figure 2.4 Mean annual number of days having a minimum
temperature of 32°F (0°C) or below.
Source: U.S. Dept of Commerce, 1968.
55
-------
ALABAMA
Mostly Ultisols
Mostly Entisols,
some ultisols
Mostly Spodosols
Mostly Alfisols,
some spodosols
Jacksonville
Atlantic Ocean
Figure 2.5 Location and extent of soil orders on which
citrus is grown.
Source: Caldwell and Johnson, 1982.
56
-------
all other aspects they are very similar to the entisols.
Their water holding capacity and CEC (cation exchange
capacity) rises substantially in the clay accumulation layer,
which is typically at a depth of 100 to 200 cm (Anderson,
1981).
The spodosols occuring along the eastern and western coasts
have an illuvial (accumulation) horizon located normally 40 to
120 cm below the surface. These soils are frequently
saturated unless artificially drained. Associated with this
may be high accumulations of organic matter in the surface
horizon with a hard pan in the eluvial (leached) horizon
(Brady, 1974).
Alfisols are moist mineral soils which are weathered but not
to the extent of the spodosols or ultisols. They are higher
in base saturation with higher pH than the spodosols or
ultisols. They have an illuvial horizon in which clays have
accumulated.
It was felt that the transport characteristics might vary
substantially among these soil orders. To investigate their
transport properties, soils which are commonly planted with
citrus (Tucker, 1978) were subdivided into their respective
orders. Then, occurrences of samples of these soils were
located in soil characterization data of Calhoun et al.
(1974), Carlisle et al. (1978) and Carlisle et al. (1981).
Samples having significant amounts of missing information in
the profile were dropped. The remaining data on organic
carbon, saturated hydraulic conductivity, pH, texture and
water holding capacities were entered into a computerized data
base. Table 2.3 shows the soils and sample numbers used for
characterization. There were a total of 19 samples of
entisols, 9 of ultisols, 20 of spodosols, and 10 of alfisols.
57
-------
TABLE 2.3 SOIL SAMPLE SELECTED FOR SOIL CHARACTERIZATION
ANALYSIS
Soil
Number of
Order Samples
Astatula Entisol
Candler Entisol
St. Lucie Entisol
Paola
Entisol
Wabasso
Pineda
Felda
Riviera
2
5
Arredondo Ultisol 6
Apopka Ultisol 3
Immokalee Spodosol 2
Myakka Spodosol 9
Oldsmar Spodosol 5
Spodosol 4
Alfisol 3
Alfisol 2
Alfisol 5
Sample Numbers
S27-22, S49-8, S64-4, S64-5,
S5-1, S5-24, S5-26
S27-6, S49-5
S50-20, S56-10, S5-27,
S49-1, S49-7
S27-7, S50-21, S64-6, S5-2,
S49-6
S27-11, S35-1, S35-2,
S42-119, Sl-6, Sl-84
S64-12, S35-3, S53-4
S50-17, S64-8
S27-14, S27-16, S49-10,
S56-6, S64-16, S5-6, S5-7,
S50-1, EXP. S50-1
S49-30, S50-33, S56-7, 556-12,
S56-2
S50-13, S64-24, S56-1, S51-2
S5-17, S50-10, S43-5
S50-2, S51-7
S49-15, S56-18, S5-13, S43-18,
S56-21
Sample numbers correspond to those used in Calhoun et al. ,
1974; Carlisle et al., 1978; and Carlisle et al., 1981.
58
-------
To analyze profile characteristics, each soil was assigned
parameter values for depths from 1 to 300 cm at 1 cm
increments. This was done by interpolating between data
points and assigning values for increments above or below the
range of measurements to the shallowest and deepest values
found, respectively. Arithmetic means and standard deviations
for each parameter were then computed at each 1 cm increment.
The results were then plotted. (For detailed plots of mean
parameter value with standard deviations see Appendix B).
Figure 2.6 shows the means of the field capacity water
contents (6 fc , % by volume) for five groupings of soils. In
PRZM, 9fc affects the rate of percolation of water, and hence,
of the chemical, through the profile. The Astatula-Candler
and St. Lucie-Paola groupings were left separate to see if
differences could be discerned between them since the latter
are entisols occuring primarily in the "flatwoods." The
curves show that 0 fc is nearly identical for all the entisols,
being roughly 5-7% at the surface, then dropping rapidly, and
asymtotically approaching an average value of 2-3% in the
lower profile. The spodosols and alfisols, on the other hand,
show a higher 9 f in the surface, 14 to 15%, which drops
rapidly to 5-6% at a depth of 20-60 cm and sharply rises again
to 25-26% at depths greater than a meter. The ultisols show a
similar trend, with slightly lower 6 fc at the surface and the
high 0£c layer occurring lower in the profile than in the
spodosols or alfisols.
Interestingly, the mean values for saturated hydraulic
conductivity (Kg, cm/hr) (Figure 2.7) show patterns which are
virtually the mirror image of the curves of 6 f c vs. depth.
Again the entisols are the deviant group for this parameter.
Of particular interest is the fact that for the soils having
the lowest mean hydraulic conductivities (alfisols), water
transmission rates are still in excess of 1 cm/hr or 24
59
-------
PARAMETER 0f(,, IN PERCENT
5.0 10.0 15.0 20.0 25.0
Explanation
0 Candler & Astatula (ridge entisols)
4 Arredondo & Apopka (ultisols)
Oldsmar, Inunokalee, Wabasso, Myakka (spodosols)
D Felda, Riviera, Pineda (alfisols)
St. Lucie, Paola (flatwood entisols)
Figure 2.6 Mean values of field capacity water content vs
depth from soil characterization analysis.
60
-------
10
PARAMETER KS/ IN cm/hr
20 30 40 50
70
Not
Explanation
O Candler & Astatula (ridge entisols)
4 Arredondo, Apopka (ultisols)
Oldsmar, Immokalee, Wabasso, Myakka (spodosols)
D Felda, Riviera, Pineda (alfisols)
shown st* Lucie & Paola entisols are 60-90 cm/hr
Figure 2.7
Mean values for .saturated hydraulic conductivity
versus depth from soil characterization analysis,
61
-------
cm/day. This has important implications for unsaturated zone
modeling in that water and fluxes may not be limited by soil
transmission rates but by water table conditions at the
interface of the unsaturated and saturated zones.
Figure 2.8 shows the mean soil pH versus depth for the five
soil groupings. The pH profiles are closely grouped within a
range of roughly 5.0 to 6.0 except for the alfisols which are
typified by pH values of 6.0 to 7.0. While these pH values
are not much higher, base-catalyzed aldicarb hydrolysis rates
may be substantially higher in these soils. Standard
deviations are generally +0.5 in the entisols and ultisols and
+1.0 in the spodosols and alfisols (see Appendix B).
Figure 2.9 shows soil organic carbon versus depth for the five
soil groupings. The quantity of organic carbon in the soil
affects the adsorption of the chemical to soil materials. For
this parameter, all soils show a nearly monotonic reduction
with depth except for the spodosols which show subsurface
accumulation, peaking at about 90 cm.
Thus we see that three of these soil orders have
distinguishing characteristics which in turn have implications
for aldicarb trknsport; the entisols with their very low water
holding capacities, the alfisols with their higher pH and the
spodosols with their accumulation layer of organic carbon.
The ultisols do not distinguish themselves in any category,
which by default categorizes them in a range of intermediate
properties.
The entisols stand out as the group for which aldicarb
transport is potentially highest, having low water holding
capacity, low pH and low organic carbon. Therefore, these
soils should show the lowest degradation rates and lowest
chemical adsorption although the low water holding capacity
62
-------
3.0
E
u
w
Q
4.0
PARAMETER pH
5.0 6.0
7.0
Explanation
0 Candler & Astatula (ridge entisols)
Arredondo & Apopka (ultisols)
Oldsmar, Immokalee, Wabasso, Myakka (spodosols)
D Feld, Riviera, Pineda (alfisols)
St. Lucie, Paola (flatwood entisols}
Figure 2.8 Mean values of soil pH versus depth from the
soil characterization analysis.
63
-------
PARAMETER ORGANIC CARBON, IN PERCENT
0.5 1.0 1.5 2.0
2.5
Explanation
0 Candler & Astatula (ridge entisols)
f Arredondo & Apopka (ultisols)
Oldsmar, Immokalee, Wabasso, Myakka (spodosols)
DFeld, Riviera, Pineda (alfisols)
St. Lucie, Paola (flatwood entisols)
Figure 2.9
Mean values of organic Carbon versus depth
from soil characterization analysis.
64
-------
may somewhat counteract the low organic carbon contents.
2.1.3 Irrigation
Irrigation is practiced in Florida citrus to boost yields, and
to protect against droughts and freezes. It is essential for
production in the coastal flatwoods. Approximately 62% of the
citrus is irrigated by some method (Stanley et al., 1980).
Irrigation water provides an additional opportunity for the
leaching of the chemical.
Although a variety of scientific methods are available to
schedule the application of water, most growers, by virtue of
intuition or other indicators, normally initiate irrigations
at 50% available water depletion (Harrison, 1984; Personal
Communication).
There are, at present, five major methods of irrigation
utilized in Florida citrus. These are 1) permanent overhead
sprinklers, 2) traveling volume guns, 3) low volume spray,
4) flood, and 5) drip. Stanley et al. (1980) subdivided the
citrus growing region into three areas; the upper ridge, the
lower ridge, and the flatwoods. The upper ridge corresponds
to the area covered by entisols lying above the southern
boundaries of Lake and Orange counties including Lake, Marion,
Orange and Pasco counties. The lower ridge corresponds to the
entisol area below this division made up primarily of Hardee,
Highlands, Hillsborough and Polk counties. The flatwoods
include areas on both the eastern and western coasts,
corresponding to the spodosols and alfisols and including
primarily Brevard, Desoto, Hendry, Indian River, Martin, Palm
Beach and St. Lucie counties.
65
-------
Table 2.4 shows the breakdown of these five and other
irrigation types by subarea. According to this survey,
traveling guns and permanent overhead guns were the favored
methods in the upper ridge, while these plus low volume spray
were favored in the flatwoods along with a high percentage of
portable guns. Harrison (1984, personal communication) has
indicated that the usage of traveling guns has diminished by
50%, portable guns by 90%, seepage by 20%, perforated pipe by
80% (there is practically no irrigation done by this method
currently), and permanent overhead by 20%. There has been a
corresponding increase in the number of drip and low volume
spray systems since 1980. The characteristics of each of
these methods and the use of irrigation for freeze protection
will be discussed individually in the following sections. A
summary of system management characteristics is given in Table
2.5.
2.1.3.1 Permanent Overhead Sprinklers
These systems consist of rotating sprinklers which are located
at intervals through the orchard on high risers which .extend
several feet above the tree canopy. The entire'area of the
grove is wetted during irrigation events. Water is applied at
rates of approximately 0.5 cm/hr (Harrison, 1984, Personal
Communication). Four to five (4-5) cm of water are applied
per event at 5 to 7 day intervals. An average of about 38 cm
of water was applied annually by permanent overhead systems
monitored in the state from 1970 to 1980 (Duerr and Trommer,
1982).
2.1.3.2 High Volume Guns
Self propelled and portable guns have the same basic
characteristics. They both discharge a large volume of water,
.0079-.076 m/s (125 to 1200 gpm) under high pressure from a
66
-------
TABLE 2.4 CITRUS IRRIGATION SYSTEMS WITH ACREAGES FOR EACH OF THREE AREAS IN
THE STATE
AREA
I.
Upper
Ridge
II.
Lower
Ridge
III.
Indian
River fi
Flat-
woods
TOTAL
Traveling
Guns
Acres
31,044
33,968
5,576
70,588
%
37.0
21.6
1.9
13.3
Portable
Guns
Acres
7,7375
105,740
113,115
%
36.6
21.4
Seepage
Acres
147.169
L47.169
t
51 JO
51. (
Perforated
Pipe
Acres
4,526
9,532
260
14,318
%
5.4
6.1
0.1
2.7
Permanent-
Overhead
Acres
33.24C
80,421
12,510
L26.171
1
39.7
51.2
4.3
23. e
Drip
Acres
4,735
9,956
7,656
22,347
%
5.6
6.3
2.6
4.2
Low Volume
Spray
Acres
2,891
23,164
9,996
\
36,051
%
3.4
14.8
3.5
6.8
Total
Irrigation
Acres
83,811
157.041
288,759
529,759
Sourcet Stanley, et. al., 1980
-------
TABLE 2.5 SUMMARY OF MANAGEMENT CHARACTERISTICS OF IRRIGATION SYSTEMS
Type
Event
Application Rate
in Wetted Area
(cm/hr)
Irrigation Frequency Wetted Area
(days) (%)
Approximate
Annual
Application
(cm)
Permanent
Overhead
High Volume
Gun
CO
Seepage
Drip
0.5
0.5
0.06-0.122
0.7 =
5-7
8-10
10-13
0.3-1
100
100
100
38
31
24
36
Microjet
0.62
2-3
171
22
Assumes a unit area of 58 m2 per tree (25x25 ft spacing)
Calculated
-------
single nozzle. The water is shot high into the air and
impacts the ground and tree canopy in much the same way that a
heavy rainfall would. The frequency and depth of irrigations
are much the same as for permanent overhead systems. The
average amount of water applied per year by guns is
approximately 31 cm (Duerr and Trommer, 1982).
2.1.3.3 Seepage (Flood)
Seepage irrigation is used only in the flatwoods areas where
there is either a shallow water table or a relatively
impermeable layer close to the soil surface. Water is
introduced into furrows adjacent to the beds. Depending on
how high the water is allowed to rise in the furrow, this type
of irrigation is either called furrow flood or crown flood.
With crown flooding, water is brought to the level of the top
of the bed, whereas with furrow flooding it is brought to
roughly one half the furrow height, relying on capillary rise
to moisten soil in the beds above this height. The amount of
irrigation water reaching the root zone is a function of the
permeability of the soil and the residence time of the water
in the furrows. Calvert et al. (1967) showed that it took 13
days on a Felda soil and 10 days on an Immokalee soil to
complete a single wetting and drying cycle using furrow
irrigation. Capillary water rise above the free water table
was measured at 15.2 cm (6 in.) in an Immokalee soil and 30.5
cm (12 in.) in a Felda soil. Based on monitoring data from
1970 through 1978, it was determined that about 24 cm of water
on the average is applied annually by furrow type systems
(Duerr and Trommer, 1982).
2.1.3.4 Drip (Trickle)
Drip irrigation is a highly efficient method of applying water
directly to the crop root zone. A low pressure water line
69
-------
running down each row of trees is tapped by emitters which
drip water in an approximately 1 sq. m area around the tree.
Emitters apply water at a rate of about 1.07 cm/s (0.017 gpm)
(there are normally two emitters per tree) and irrigations are
needed two to three times per day during peak use periods.
Duerr and Trommer (1982) monitored the average annual
application by drip systems to be 36 cm. This is on a unit
area (e.g., per acre) basis, however, and application in the
wetted area would be much higher.
2.1.3.5 Low Volume Spray
Low volume spray systems are similar to drip systems but
operate at higher pressures and wet a larger area around each
tree. The typical wetted area is approximately 10 sq. m.
These systems apply water at an average flux of roughly 15.8
cm/s (0.25 gpm). Irrigations are needed two to three times a
week. Duerr and Trommer (1982) measured the annual average
application depth of low volume spray systems to be 22 cm on a
total areal basis. Application rates are higher in the wetted
zone.
2.1.3.6 Irrigation For Freeze Protection
Most of the methods of irrigation previously mentioned can be
used for freeze protection in citrus. There are two measures
which can be taken, one which is practiced before the freeze
and one practiced during the freeze periods.
The idea behind pre-freeze irrigation is to increase the
thermal conductivity and heat capacity of the soil. Following
a freeze warning, 0.6 to 1.3 cm of water is applied. This
allows the soil to store heat which will be released into the
grove during lower temperature periods.
70
-------
w
OT
Pi
W
CO
w
ffi
o
w
00
f
03
I
00
*r
t
o
03
O
Illuvial Layer
%ssss8%s^g^ss&s^^
5-8
25
5-8
FEET
1.2 1.5-2.4
7.6
1.5-2.4 1.2 METERS
Figure 2.10 Typical configuration for Bedded citrus in the Flatwoods Area.
-------
Irrigation water applied during the freeze provides both
sensible heat contained in the water and latent heat of fusion
as that water becomes ice. Enough water must be applied so
that the sensible plus the latent heat of fusion added exceeds
the convective, radiation and evaporative heat losses from the
plant. The convective and radiation heat losses are of equal
magnitude for a wet or dry leaf. Heat consumed due to
evaporation of water, however, exceeds the heat liberated by
fusion by a factor of about 7.5. Therefore, 7.5 times as much
water must be frozen as is evaporated in order to raise the
temperature. Because high winds increase evaporative heat
losses, this method is safe to use only during radiation
freezes. In advective freeze, more harm than good can be done
(Harrison et al., 1974). Irrigation only by the drip, low
volume spray or flood methods are used during freezes. The
use of overhead systems results in severe breakage of limbs
due to ice build up and is feasible only when trees are young
(i.e., in nurseries). Where low volume sprays are used water
is applied at the sprinkler capacity until the temperature in
the grove reaches 32 deg. F (0 deg. C) (Harrison, 1984,
personal communication).
2.1.4 Cultural Practices
2.1.4.1 Bedding
The chief cultural practice which distinguishes ridge citrus
from flatwoods citrus is that of forming beds for trees in the
flatwoods areas.
A typical configuration for double bedded citrus is shown in
Figure 2.10. Other configurations may have as few as one row
per bed or as many as eight. Tree spacing is typically 7.6m
72
-------
(25 ft) between rows and varies from 3.8m (12.5 ft) to 7.6 m
(25 ft) down each row. The land is formed by excavating
furrows between each bed and using this soil to create raised
beds in the middles. This provides additional rooting depth
for trees and provides a method of readily irrigating or
draining beds. The bed heights are typically 0.3 to 0.8 m (12
to 30 inches). At times bedded citrus may also be
artificially drained by tile or corrogated plastic pipe. If
so, these drains are placed three to four feet below the top
of the bed, either in the bed centerline or in the furrows
just above the illuvial layer on a 0.001 to 0.002 m/m slope.
This practice of draining is probably used on less than 5 to
10% of the flatwoods area (Woods, 1984, personal
communication).
In the ridge areas beds are unnecessary due to the extremely
good drainage characteristics of the soils. Trees are
typically planted on 7.6 by 7.6 m (25 by 25 ft) spacing. In
the ridge areas trees normally attain greater canopy dimension
than in the flatwoods.
2.1.4.2 Control of Volunteer Vegetation
Herbicides are generally applied to 1.2 to 1.5 m (4 to 5 ft)
swaths on each side of the trees in bedded areas to eliminate
weeds. This has the advantage of reducing transpiration by
eliminating unwanted vegetation in the root zone and also
eliminating weeds which might interfere with sprinkler
patterns of low volume jets. On the remainder of the area,
natural vegetation or sod is used to reduce erosion of beds
and furrows. This is either mechanically or chemically mowed.
In the ridge areas weeds are controlled using herbicides or
discing to reduce transpiration by volunteer vegetation, clean
sprinkler patterns and to promote drainage of cold air off
hillsides during freezing periods.
73
-------
2.1.4.3 Pest Control
Common to both areas is the occurrence of nematodes. These
pests occur in the greatest number from 0.3 to 1.8 m below the
surface when there is the greatest number of feeder roots.
Usually the top 15 cm is too hot and dry to support them.
(Anderson, 1981). Mites are also problematic in the tree
canopy. Both of these pests can be controlled by the use of
aldicarb. The chemical is normally chiseled in to a depth of
5 cm, in 1 to 1.5 m bands on each side of the tree rows
centered under the drip line.
2.1.4.4 Fertilization
Groves are normally limed to maintain pH between 6.0 and 7.0
in the root zone. As mentioned previously, high soil pH may
increase aldicarb degradation due to hydrolysis.
2.1.5 Thickness of Unsaturated Zone
The thickness of the unsaturated zone is a critical factor . in
the transport of aldicarb from the soil surface to the
saturated zone. Where the unsaturated zone is thin and the
water table is close to the land surface, aldicarb will enter
the aquifer much more quickly and have less opportunity to
degrade than where the unsaturated zone is thicker. In the
citrus growing areas of Florida the thickness of the
unsaturated zone depends on the topographic region.
The central peninsula area falls into two physiographic areas;
the central highlands and the coastal lowlands (Cooke, 1945).
The central highlands is composed of a series of long narrow
74
-------
ridges 30.5 to 61 m (100 to 200+ ft above sea level) that run
parallel to the axis of the peninsula. Surrounding these
ridges is a broad upland area of lower elevation 15.3 - 30.5 ra
(50-100 ft) and much less local relief. The coastal lowlands
(also known as the flatwoods) flank the central highlands on
the Atlantic and Gulf of Mexico coasts. This area, generally
below 15 m (50 ft) in elevation, consists chiefly of old beach
ridges, terraces, lagoons, and swamps. The generalized
topographic map (Figure 2.11) clearly shows these general
features.
The thickness of the unsaturated zone closely follows the
topograpic trends. In the poorly drained lowlands, the water
table is usually within 1.5 m (5 ft) of the land surface
(Healy, 1974). In these areas the citrus groves must be
drained for successful cultivation. The water table is
controlled by a system of canals, so the thickness of the
unsaturated zone is dependent on man's activities.
In the central highlands, the depth to the water table is much
more variable. Near lakes the unsaturated zone can be very
thin. Beneath the high ridges it can be as much as 30.5 m (100
ft) thick. Generally, the unsaturated zone averages 3 to 9 m
(10 to 30 ft) in thickness. Figure 2.12 is a typical
cross-section through the ridge area near Lake Hamilton
(central highlands) showing the variability in thickness of
the unsaturated zone.
2.1.6 Delineation of Unsaturated Zone Scenarios
Based on the preceeding information on climate, soils,
irrigation methods, cultural practices, and depth to
saturation, six subareas of the citrus growing region were
75
-------
ALABAMA
Jacksonville
Explanation
Elevation in feet above sea level
0-50
50 - 100
100 - 150
Greater than 150
Atlantic Ocean
Figure 2.11 Generalized 'topography of the Central Peninsula,
Florida. Source: USGS, 1:250,000 scale topo-
graphic sheets of Florida.
76
-------
01
w
6 8 10 12 14 16 18 20
DISTANCE, IN THOUSANDS OF FEET
22 24
Figure 2.12 Typical cross-section through a ridge citrus area
showing relative thickness of unsaturated zone.
Source: ; Anderso'n and Simonds, 1982.
-------
identified. It was felt that these regions represent
distinctly unique situations with respect to aldicarb leaching
in the unsaturated zone. These areas are shown on the map in
Figure 2.13. The boundaries of these areas are approximate
and do not necessarily correspond to any political boundaries
or natural landforms found within the state.
Subareas 1 to 4 correspond to locations having entisols,
ultisols, spodosols, and alfisols, respectively, with two
subareas of wetter spodosols and alfisols broken out (subareas
5 and 6). Areas 5 and 6 receive roughly 10% or more rainfall
annually than areas 1 to 4.
Table 2.6 shows typical characteristics of these areas.
Rainfall differences are not large across the entire citrus
growing region, but higher annual depths do tend to occur in
areas 5 and 6. These differences will be mitigated in certain
cases by the addition of irrigation water. The rainfall
depths shown in Table 2.6 are the means for the period of
record (1965-1978) at Lake Alfred (areas 1-4) and Belle Glade
(areas 5 and 6) actually used in model simulations. The
potential evapotranspiration gradient runs from northeast to
southwest with upper area 3 (east coast) tending to be lowest
and area 3 (west coast) and areas 5 and 6 tending to be
highest. Actual values used in the simulation were also from
the Lake Alfred and Belle Glade meteorological records.
In each of these areas, each of four irrigation practices are
utilized. Areas 1 and 2 generally receive no irrigation or
irrigation by drip, low volume spray, or overhead systems
while areas 3, 4, 5 and 6 will tend to be irrigated by the
drip, low volume spray, overhead or flood methods. If
aldicarb is applied on treated bands centered underneath the
drip line in a mature orchard, however, the wetted area of a
drip irrigation system will not intersect the treated band.
78
-------
Jacksonville
Atlantic Ocean
Figure 2.13 Subareas delineated for unsaturated zone modeling.
79
-------
TABLE 2.6 CHARACTERISTICS OF AREAS SELECTED FOR UNSATURATED
ZONE MODELING
Annual Annual Soil1 Irrigation2 Depth to
Precipitation ETp Order Method Saturated
Area (cm)' (cm) Zone (m)
1
2
3
4
' 5
6
121
121
121
121
140
140
117-127
117-122
114-137
122-132
127-132
127-132
E
U
S
A
S
A
L./O,N
L,0,N
L,0
L ,0
L.,0
L ,0
2.7, 9.
1.8, 9.
1.2
1.2
1.2
1.2
0
0
E = Entisols, U = Ultisols, S = Spodosols, A = Alfisols
2 L = Low volume spray
0 = Overhead (Permanent Sprinklers or Guns)
N = None
80
-------
Therefore, this irrigation method was eliminated from
consideration. Neither were the flood methods considered. To
adequately model the movement of water under these irrigation
systems would require a two-dimensional model. Even though
PRZM was modified, (as discussed in Appendix A), to simulate
lateral drainage of water from bedded citrus, the simulation
of the addition of water would necessarily be overly
simplistic in a one-dimensional framework. It was decided
that the flood irrigation method would be eliminated in favor
of presenting results based on questionable simulations.
Consideration of freeze protection irrigation was also
eliminated. The duration of the application of water is
highly variable and dependent upon the duration of the
freezing period. These periods normally occur at night with
temperatures rising above freezing occurring during the day.
Therefore, mean daily temperatures are rarely below freezing.
PRZM requires mean daily temperatures as an input in order to
calculate potential evapotranspiration if pan evaporation is
missing. Based only on mean daily temperature, PRZM might not
be able to detect the occurrence of a freeze protection
irrigation event. Even if maximum and minimum daily
temperature were input, the duration of the freezing period
would still be unknown. Proper simulation of these events
would require at least hourly temperature inputs, which are
simply not available, and wind speed to determine if the
freeze is radiative or convective. In the end it was decided
that due to the complexity of the processes and data inputs
required, meaningful simulations could not be done, with the
time and available budget.
Areas 1 and 2 show the greatest variability in unsaturated
zone thickness. It was decided that at a minimum two depths
should be simulated. The depths of 180 cm and 270 cm in the
81
-------
ultisols and entisols represent approximate root zone depths
in these soil types. The 900 cm depth represents an
intermediate depth to ground water in the ridge area. Greater
depths can be much higher but occur with far less frequency.
Areas 3, 4, 5, and 6 have a shallow, relatively constant
unsaturated zone, controlled largely by drainage practices.
These combinations resulted in twenty scenarios. In addition,
a no-irrigation simulation was made in each spodosol and
alfisol area. This was done because, for the low volume spray
case, part of the treated band receives only rainfall. Thus
the pesticide load leached from the grove is the areally
weighted average of the load from the wetted area (irrigation
plus rainfall) and the nonirrigated area (rainfall only). For
ease of referring to these scenarios later in the text, they
have been assigned the codes shown in Table 2.7. The first
letter indicates the soil order; the next letter, the method
of irrigation; the following number the depth of the soil
core; and, the last letter, the low or high rainfall
condition.
2.1.7 Selection of PRZM Hydrology and Hydraulic Input
Parameters
For each unsaturated zone scenario (1-6) input data sets had
to be created for PRZM. The following sections cover the
specification of all the required hydrologic and hydraulic
input parameters. Most were selected based on guidance
provided in the PRZM user's guide (Carsel et al., 1984).
Exceptions are listed and discussed.
82
-------
TABLE 2.7 UNSATURATED ZONE MODELING SCENARIOS
Scenario
Identifier
Description
Soil
Irrigation
Core Depth(m)
EN2.7
EN9.0
EO2.7
EO9.0
EL2.7
EL9.0
UNI. 8
UN9.0
U01.8
U09.0
UL1.8
UL9.0
SN1 . 2L*
S01.2L
SL1.2L
SN1 . 2H
S01.2H
SL1.2H
AN1.2L
AO1 . 2L
AL1 . 2L
AN1 . 2H
AO1 . 2H
AL1 . 2H
Entisols
Entisols
Entisols
Entisols
Entisols
Entisols
Ultisols
Ultisols
Ultisols
Ultisols
Ultisols
Ultisols
Spodosols
Spodosols
Spodosols
Spodosols
Spodosols
Spodosols
Alfisols
Alfisols
Alfisols
Alfisols
Alfisols
Alfisols
None
None
Overhead
Overhead
Low volume spray
Low volume spray
None
None
Overhead
Overhead
Low volume spray
Low volume spray
None
Overhead
Low volume spray
None
Overhead
Low volume spray
None
Overhead
Low volume spray
None
Overhead
Low volume spray
2.7
9.0
2.7
9.0
2.7
9.0
1.8
9.0
1.8
9.0
1.8
9.0
1.2
1.2
1.2
1.2
1.2
1.2
1.2
1.2
1.2
1.2
1.2
1.2
*L or H in the last position refers to low or high rainfall
areas.
83
-------
2.1.7.1 Simulation Period
The six locations appearing on the map of Figure 2.14 show the
stations for which meteorological data was initially obtained.
The stations were primarily chosen for their coverage of the
citrus growing area and the availability of precipitation, pan
evaporation and temperature data. These data sets were put
into the proper format for PRZM.
Twenty-five years of precipitation and pan evaporation data
were available at most of these six locations. However, for
some stations pan evaporation data was missing. Therefore,
scenarios were set up to run for only 14 years. The
meteorologic data set chosen for entisols, ultisols, and low
rainfall spodosols and alfisols was the Lake Alfred Experiment
Station. The vhigh rainfall spodosol and alfisol cases were
run with Belle Glade meteorological data.
2.1.7.2 Hydrologic Parameters
The evaporation pan coefficient was set at 0.78. The ANETD
parameter is meaningless here since a crop is always present.
Percentage of daylight hours (DT(12) were input because pan
evaporation data were sometimes missing for short periods.
They were chosen based on a latitude of 28 N. The snow melt
coefficient was set to zero. The runoff curve number for AMC
II was set at a very low value of 20 so that no runoff would
occur.
2.1.7.3 Crop Parameters
The crop was assumed to be mature throughout the entire
simulation period. Trees were assumed to have an interception
potential of 0.15 cm. This was based on information found in
Viessman et al., 1977. The root zone depth of the citrus tree
84
-------
ALABAMA
Jacksonville
Gulf of Mexico
Atlantic Ocean
Lake Alfred
Vero Beach
Clewiston
Belle Glade
Figure 2. 14 Selected meteorologic stations for
jinsaturated gone modeling.
85
-------
was set at 240 cm (8 ft) and 150 cm (5 ft) in the entisols and
ultisols respectively, and at 60 cm (2 ft) in the spodosols
and alfisols (Tucker, 1978). The cover parameter (COVMAX) for
entisols and ultisols was estimated to be 75%. This is the
approximate area of a 7.6 m (25 foot) diameter circle
inscribed in a 7.6 by 7.6 m square, thus assuming that tree
canopies just touch on a 7.6 by 7.6 m spacing. For the
spodosols and alfisols, on which tree canopies are smaller
(4.6 m diameter) and plantings are closer (usually 7.6 by 4.6
m) the cover parameter was set at 50%.
2.1.7.4 Soil Hydraulic Parameters
Selection of soils parameters was primarily a function of soil
group. Therefore they are discussed by group below:
2.1.7.4.1 Entisols and UltisolsEntisols and ultisols were
divided into those having thick and thin unsaturated zones.
For the former a core depth of 900 cm (30 ft) was used and for
the latter 270 cm (9 ft) or 180 cm (6 ft) was used. Thirty
compartments were used in each core. The original PRZM
hydraulics module (HYDRl) was selected for these soils. This
method requires input values of field capacity and wilting
point water contents and assumes that water in any compartment
above field capacity drains to that level within one day.
For entisols, the profile was divided into five horizons based
on the soil characterization analysis performed as part of
this study. For ultisols, seven horizons were chosen. Table
2.8 shows the representative values of 8 fc and 9 ^ used for
each horizon.
2.1.7.4.2 Spodosols and AlfisolsFor these two soil types,
which typically occur in the flatwoods, only one core depth of
120 cm (4 ft) was used. This is because the depth to
86
-------
TABLE 2.8 HYDRAULIC CHARACTERISTIC DATA FOR FLORIDA SOILS BY
HORIZON
Soil
Group
Entisols
Entisols
Entisols
Entisols
Entisols
Horizon
1
2
3
4
5
Depth
(cm)
0-10
10-20
20-40
40-100
100+
Bfc
5.6
5.4
4.2
3.5
3.0
9wp
2.2
2.1
2.0
1.5
1.2
Ultisols
Ultisols
Ultisols
Ultisols
Ultisols
Ultisols
Ultisols
Alfisols
Alfisols
Alfisols
Alfisols
Alfisols
Spodosols
Spodosols
Spodosols
Spodosols
Spodosols
Spodosols
1
2
3
4
5
6
7
1
2
3
4
5
1
2
3
4
5
6
0-20
20-40
40-100
100-160
160-200
200-240
240+
0-20
20-40
40-80
80-120
120-240
0-10
10-40
40-60
60-100
100-200
200+
8.2
7.0
5.5
8.0
20.0
22.5
5.0
10.0
6.5
9.5
20.0
25.0
14.0
9.0
6.0
12.0
21.0
6.0
3.0
2.5
1.8
4.0
13.0
16.0
1.8
3.5
6.0
5.5
9.5
12.5
4.2
3.2
2.2
4.0
7.8
2.0
87
-------
saturation in these areas is not controlled by relief, but by
the control of the water table through drainage. Thirty
compartments were again used to simulate each core.
In the spodosols and alfisols, the presence of low
permeability soil layers or the presence of the high water
table may cause both temporary saturated conditions in the
unsaturated zone and lateral drainage into adjacent drainage
furrows. For this reason the HYDR2 option was used for soil
hydraulics with lateral drainage to more closely simulate
recharge to the saturated zone. In addition to the
specification of field capacity and wilting point water
contents for each horizon (Table 2.8), this method requires
that time constants for vertical and lateral drainage be
input.
Unfortunately, these constants are not easily estimated.
There are two ways that the movement of water vertically can
be limited; either by the presence of a low permeability layer
or the presence of a saturated zone. If the permeability is
limiting the time constant for vertical drainage can be
estimated from the saturated hydraulic conductivity. If the
presence of a saturated zone is the limiting factor then
downward vertical movement is limited by the rate of fall of
the water table.
Inspection of the saturated hydraulic conductivities in the
low permeability layers in spodosols and alfisols (see Figure
2.7) indicates that, while hydraulic conductivity is far lower
in these horizons than in the remainder of the profile, lowest
values still are on the order of 1 cm/hr (on the average).
Thus, it appears that while some profiles may exist in which
the vertical movement is restricted by permeability, for the
most part, vertical drainage is limited by the presence of a
saturated zone. Therefore the rate of movement of water and
88
-------
pesticide downward is controlled by localized manipulation of
the water table. This fact makes the time constant for
vertical drainage even harder to assess for a general case.
Similarly, the rate of lateral drainage is almost certainly
controlled by the manipulation of water depths in the furrows,
rather than soil permeabilities.
The only data located which could be used to calculate these
constants was that of Calvert et al. (1967). Figure 2.15 from
that report shows the fall of a water table in two bedded
citrus groves. On the right is an Immokalee soil (spodosol)
and on the left is a Felda soil (alfisol). Specific yields
(saturation minus field capacity water contents) for each
horizon in each soil were determined from profiles for the
respective soil types found in Calhoun et al. (1974), Carlisle
et al. (1978) and Carlisle et al. (1981). Water draining from
0-22 hrs. in the Immokalee soil and from 0-25 hrs. in the
Felda soil was assumed to all move laterally. Time constants
for lateral drainage were determined by dividing the depth of
water drained by the time of drainage and the depth of a PRZM
soil layer (in this case 4 cm). Time constants for lateral
drainage were thus determined to be 0.48 /day for the alfisol
and 0.21 for the spodosol.
Similarly, vertical time constants were estimated for the
drainage of the remainder of the water in the profile which
was assumed to all move vertically. These constants were
found to be 0.065 /day for the alfisol and 0.32 /day for the
spodosol.
In the PRZM input data set, horizons above a 60 cm (2 ft)
datum were assigned lateral drainage parameters equal to those
calculated; horizons deeper than 60 cm were assigned lateral
drainage parameters of zero. The vertical drainage parameters
calculated above were assigned to the deepest PRZM horizon;
89
-------
(0
w
g
w
Q
o-
6-
12-
18-
24-
30
36-
42-
FELDA SOIL
Ohr
0+25 hrs
I 0+241 hrsJV--^^ ^-~J
IMPERVIOUS CLAY
LAYER
I
0
IMMOKALEE
SOIL
0 hr
0+2 2_ hrs_
0+115 hrs
I I
5 10 15 10 5
DISTANCE IN FEET FROM WATER
I
0
Figure 2.15 Fall-of free water level during one furrow
irrigation drying cycle in single bedded
'Ruby.Red1 grapefruit groves planted on
Felda and Immokalee soil types in the
Indian River area. After Calvert et al.,
1967.
(1 inch =2.54 cm)
90
-------
overlying horizons were assigned high values (10./day) to
essentially provide unrestricted drainage.
2.1.7.5 Irrigation Parameters
The only irrigation parameters required by the algorithms
developed for PRZM (see Appendix A) are the rate of water
application and the percent available water content at which
events are triggered. The water application rates are not
extremely important. Since PRZM operates on a daily timestep,
the durations of the events (i.e., the amount of water
necessary to bring the root zone profile from present status
to field capacity divided by the input application rate) are
rounded to the nearest whole day. The application rates are
adjusted accordingly. Even so, the rates shown in Table 2.5
were used for the irrigation system being simulated.
The available water level at which irrigation events were
triggered were set initially, at 50% but were adjusted to
achieve an annual evapotranspiration depth of about 122 cm (48
in.) for the overhead irrigation method. Final values for
this parameter was 45% for all soil orders.
2.2 SATURATED ZONE
The goal of the saturated zone modeling study was to estimate
probable aldicarb concentrations in drinking water wells. In
general terms, there are three environmental factors
influencing the potential movement of aldicarb in the
saturated zone: 1) the geometry or configuration of the
aquifer system, 2) the physical properties of the aquifers and
confining layers, and 3) the influence of the drinking water
91
-------
well. Table 2.9 summarizes the aspects of these factors that
are most important.
2.2.1 Aquifers and Aquifer Geometries
The occurrence and movement of ground water in the citrus
growing area of Florida are closely related to its geology.
The central Floridan peninsula is comprised of a thick
sequence of hydrologically connected limestone formations
which make up the principal artesian aquifer, the Floridan
Aquifer. This is overlain by younger alluvial and marine
deposits which contain the unconfined surficial aquifer and
the intermediate confined aquifers. All three are present in
the citrus growing areas and supply drinking water. Table
2.10 summarizes the hydrologic and geologic properties
discussed below for these aquifers.
2.2.1.1 Floridan Aquifer
The Floridan Aquifer is one of the largest and most productive
aquifers in the world. It underlies the entire state of
Florida and parts of Georgia, South Carolina, and Alabama.
Parker et al. (1955) originally defined the Floridan Aquifer
as:
"...parts or all of the middle Eocene (Avon Park and
Lake city Limestone), upper Eocene (Ocala Limestone),
Oligocene (Suwannee Limestone), and Miocene (Tampa
Limestone and permeable parts of the Hawthorne
Formation that are in hydrologic contact with the rest
of the aquifer)."
92
-------
TABLE 2.9 PHYSICAL FACTORS INFLUENCING PESTICIDE MOVEMENT IN
THE SATURATED ZONE
1. Aquifer Geometries
a. Floridan Aquifer
b. Intermediate aquifer
c. Surficial aquifer
d. Physical relationship between
aquifers
2. Hydrogeologic Properties
a. Hydraulic conductivity
b. Hydraulic gradient
c. Porosity
d. Thickness of aquifer
e. Recharge to saturated zone
3. Drinking Well Specification
a. Distance between pesticide source
area and well
b. Well depth
c. Well rate
93
-------
TABLE 2.10
GENERALIZED STRATIGRAPHIC UNITS AND ASSOCIATED HYDROGEOLOGIC
PROPERTIES (Compiled from Stringfield, 1966; Klein et al.,
1964; Knochenmus and Hughes, 1976; Grain et al., 1975; Tibbals,
1981; Healy, 1982; Merritt et al., 1983).
GEOLOGIC ACE
SERIES
STRATAGRAPHIC
UNIT
RANGE LITHOLOGY
OF
THICKNESS (f t)
HATER-BEARINO
-PROPERTIES
HYDROLOGIC
UNIT
Holocene
Pleistocene
vo
Pliocene
Miocene
Unnamed alluvial, lake
and windblown deposit*
0-70 Alluvium, freshwater marl
peats and muds in stream
and lake bottoms. Also
some dunes and other wind-
blown sand
Relatively impermeable
Pamlieo formation
and marine and estua-
rine terrace deposits
0-100 Mostly marine quarts sand,
unconsolidated, and generally
well graded. Also, some
fluviatile and lacustrine
sand, clay, marl, and peat
deposits
Generally low permeabilities
Fort Thompson Formation 50-150
(contemporaneous with
Anastasia)
Alternating marine shell beds
and fresh-water marl
Generally of low permeabilitj.es
except locally where it is
solution riddled
Anastasia Formation
50-150
Sand, marl, and shell beds
Moderate to high permeability
1
Citronelle Formation
0-200 Marginal marine fine to
course grained quarts sand
containing kaolin!te matrix,
variegated red and orange
quartzite pebbles, cross-
bedded
Yields large quantities of
good quality water
Caloosahatchee Marl
0-60
Shell, sand, silt and-marl
Shell beds yield water that in
some areas is highly mineralised
Alachua and Bone
Valley Formations
0-100 Nonmarine interbedded depo-
sits of clay, sand, and sandy
clayi much of unit is phos-
phatic, base characterised by
rubble of phosphate rock and
silicified limestone residuum
in a gray and green phosphatio
clay matrix
Yields small to moderate
quantities of water to shallow
wells
H
£
Tamiami Formation
0-150
Sand, marl,
limestone
shell beds, and
Upper part of formation is very
permeable. Lower part, with the
Hawthorn, confines water in the
underlying principal artesian
aquifer
(Continued)
5
-------
TABLE 2.10 (continued)
vo
SERIES
Miocene
Oligocene
Upper
Eocene a
I
o
a
r*
a
8
Middle Eocene
BTRATIGRAPHIC
UNIT
Hawthorn Formation
Tampa Limestone
Suwanee Limestone
Crystal River
Formation
Hillleton and
Inglia Formations
Avon Park
Limestone
Lake City Limestone
RANGE LITHOLOGY
OF
THICKNESS (ft)
0-550 Marine interbedded mixture
of sand, clay, and lime-
stone i fine to course
grained phosphatic sand, .
dark green to cream mont-
morillonitlc clay, phos-
phorite pebbles, lenses of
sandy dolooitic phosphatic
limestone
0-250 Fossilferous limestone with
sands, silts and clay. Some
parts of the limestone is
silicified
0-450 Marine limestone, cream to
white, soft, hard where
silicified, porous
25-300 Gray to cream colored porous
massive limestone i often
consists of tests of foramin-
ifers, cherty, in places
25-125 Tan to cream colored granular,
porous limestone, highly
fossiliferous, lower part at
places is dolomite
400-1000 Marine limestone, light gray
to dark brown, soft to hard,
fossiliferous, dolomitic,
carbonaceous, contains gypsum
and chert
200-1000 Alternating beds of dark brown
dolomite and chalky limestone
minor amounts of gypsum and
anhydrite
HATER-BEARING
PROPERTIES
Yields small to moderate
quantities of artesian and
nonartesian water. Major
part of Hawthorn forms the
confining layer for the
underlying artesian water,
but lower part forms the
upper part of the principal
artesian aquifer
Yields large quantities of
water in west-central Florida,
but is generally lower in
permeability in other parts of
the state except locally in
discrete solution zones
Yields moderate amounts of water
but generally less than under-
lying Eocene formations.
Contains solution cavities in
recharge area
One of most productive forma-
tions of the Floridan Aquifer.
Generally yields good quantity
of water .
Hater yields generally less than
the Crystal River Formation.
Contains many solution cavities
in recharge areas
Principal source of water in
areas where overlying limestone
is thin or absent. Yields water
from porous xones highly
mineral! led in many areas
Hater yields varies, important
source of artesian water in
some areas
HYDROLOGIC
UNIT
[< Intermediate Aquifers
a \
I \
L
SM
5&
s|
r
4
t
4
t
1
\
1
4
Lower Eocene Oldsmar Limestone
400-600 Marine Limestone, light brown Contains salt water
interbedded with brown crystal-
line dolomite, fossiliferous,
contains chert and gypaum
-------
It is confined above by the low permeability clays and marls
of the Hawthorn and Tamiami Formations. The aquifer is
generally fossiliferous and porous, ranging in thickness from
100 m (300 ft) to 600 m (2000 ft). Parts are highly fractured
and riddled with solution cavities. In some areas the Avon
Park Limestone is of lower permeability and divides the
aquifer into upper and lower permeable zones (Bush, 1982).
In the northwestern portion of the study area, in Citrus,
Hernando, Sumter and Pasco counties, there is a structural
high where the limestone formations in the upper portion of
the Floridan Aquifer outcrop at the land surface (Stringfield,
1966). In this area the Floridan Aquifer is unconfined and
there is a direct connection between the surface and the
aquifer. This means water and contaminants can move much more
quickly into the aquifer than where a confining layer exists.
The limestone strata gently dip and thicken to the south and
southeast of the structural high. In these areas the Floridan
Aquifer is overlain by the marine and montmorillonitic clays
in the Hawthorn Formation and marls in the Tamiami Formations.
VThen the Floridan Aquifer is confined, water will rise above
the top of the aquifer in a well. As long as the water level
is below the level of the water table there will be potential
for downward movement of water. If, however, the water levels
in the Floridan are higher than the water levels in the upper
aquifers there will be no driving force downward and therefore
no chance for the migration of aldicarb into the Floridan
Aquifer. These areas are shown on Figure 2.16.
In some areas, especially the central highlands, sinkholes
have breached the confining layer, creating a direct
hydrologic link between the surface and Floridan Aquifer. A
sinkhole occurs when a solution cavity becomes so big the roof
can no longer support the overburden. The collapse feature
96
-------
AIABAMft
Jacksonville
Atlantic Ocean
Figure 2.16 Principal Areas where the piezometric surface of the
Floridan Aquifer rises above the water table.
Source: Healy, 1975.
97
-------
sometimes fills up with water, creating a lake. Occasionally
the low permeable lake bottom sediments plug the direct
conduit to the ground water. Often, however, the sinkholes
remain open, as a direct link to bring water and contaminants
into the aquifers.
Not all parts of the Floridan Aquifer contain potable water.
In much of the coastal area the Floridan is not potable due to
salt-water contamination. In these areas (shown in Figure
2.17) aldicarb contamination to the Floridan does not need to
be considered, since the drinking water is supplied from other
aquifers.
In summary, for the Floridan Aquifer, the highest risk areas
for pesticide contamination are in the northwestern area where
it is unconfined and in the ridge area where sink holes create
direct links with the surface. Where the Floridan Aquifer
does not meet drinking water standards and where there is no
potential for downward leakage, the Floridan Aquifer does not
have to be considered. Where the Floridan is confined, the
risk depends on how quickly water and pesticide will move
downward through the overlying aquifer or aquifers and
confining layers. The speed of contaminant transport will
depend on the hydrogeologic parameters.
2.2.1.2 Intermediate Aquifers
The intermediate confined aquifers are in general localized
and not very laterally extensive. They occur in the permeable
shell beds and limestone lenses that vary in thickness from 3
to 60+ m (10 to 200+ ft) in the Tamiami and Hawthorn
Formations of Miocene age (Healy, 1982). These aquifers are
usually confined above and below by the impermeable parts of
the same geologic formations. In Polk County an intermediate
confined aquifer occurs in the younger Pliocene deposits of
98
-------
ALABAMA
Jacksonville
Atlantic Ocean
Figure 2.17 Areas of the Floridan Aquifer which do not meet
drinking water standards.
Source: Klein, 1975.
99
-------
pebbly phosphatic sands in the Bone Valley Formation (Stewart,
1966). Some authors consider the upper limestone member of
the Hawthorn Formations an intermediate aquifer distinct from
the Floridan. Over the southern two-thirds of Polk County and
all of DeSoto and Hardee counties, this permeable limestone is
hydrologically separate from the Floridan with only a few
localized connections (Steward, 1966 and Wilson, 1977).
The intermediate aquifers exist most extensively in the
south-central part of the study area, from Polk County south.
They are tapped v/hen the Floridan Aquifer is not suitable for
drinking water and the unconfined surficial aquifer does not
yield large enough quantities of water.
2.2.1.3 Surficial Aquifer
An unconfined, surficial aquifer occurs in most areas of the
state in the uppermost, unconsolidated deposits that range in
age from Miocene to Holocene. It is extremely variable both
in composition and thickness. In most areas the aquifer
averages 6 to 9 m (20 to 30 ft) in thickness, but is as little
as .3 m (1 ft) in parts of Brevard, Highlands, Manatee, and
Palm Beach counties, and as much as 120 m (400 ft) in St.
Lucie and Indian River counties. It is composed chiefly of
sands, shells, limestone, clay and marl. The geologic units
that contain the unconfined aquifers depend on the structure
and stratigraphy of a particular area.
In the central ridge area the surficial aquifer is
discontinuous and extremely variable in thickness. . In
southeastern Polk County, for example, the unconfined aquifer
varies from 7.5 to 75 meters (25 to 250 ft) in thickness
(Wolansky et al., 1979).
In the coastal citrus area the surficial aquifer is much
100
-------
thicker and more laterally continuous. It occurs in the
extremely variable Pamlico, Fort Thompson, and Anastasia
Formations of Pleistocene age (Healy, 1982). They are
composed chiefly of marine sands and shell beds, interbedded
with clay and marl deposits.
In the ridge area the surficial aquifer supplies drinking
water for small domestic users. The surficial aquifer in the
coastal citrus areas provides the principal supply for
drinking water for both municipalities and for small single
farms.
2.2.1.4 Aquifer Geometries
For the purposes of modeling, this very complicated system of
aquifers must be simplified. The first simplification is to
only consider the Floridan Aquifer where it meets the drinking
water standards and where there is potential for downward
leakage of water (and aldicarb) to the Floridan (shown
previously in Figures 2.16 and 2.17).
Figure 2.18 shows the delineation of the typical aquifer
geometries considered in this study. Areas 1, 2, and 3 are
areas that include the Floridan Aquifer system. Area 1 is
where the Floridan is unconfined and considered alone. Area 2
is where a surficial unconfined aquifer overlies the Floridan
with a confining layer in between. Together they comprise a
leaky two-aquifer system. In area 3, a three-aquifer system
exists, consisting of the unconfined surficial aquifer, and a
confined intermediate aquifer overlying the Floridan Aquifer.
In areas 4 and 5 the Floridan is not included because of its
poor water quality or because its potentiometric surface is
above the water table. Area 4 is where only the unconfined
surfical aquifer is considered. Area 5 includes both an
101
-------
ALABAMA
Jacksonville
Gulf of Mexico
Atlantic Ocean
1 -
2 -
3 -
4 -
5 -
Explanation
Floridan Aquifer Alone
- Unconfined
Leaky Two-Aquifer System
- Surficial Aquifer
- Floridan Aquifer
Leaky Three-Aquifer System
- Surficial Aquifer
- Intermediate Aquifer
- .Floridan Aquifer
Surficial Aquifer Alone
- Unconfined
Two-Aquifer System
- Surficial Aquifer
- Intermediate Aquifer
Figure 2.18 Principal aquifer geometries in the
citrus growing area.
102
-------
unconfined surficial aquifer and a confined intermediate
aquifer.
Table 2.11 shows the approximate citrus acreage in each area.
The largest amount of citrus is grown in Area 4, where the
surficial aquifer is considered alone. This is followed by
Areas 2, 3, and 5 where a multi-aquifer system is modeled.
Area 1 has the smallest.amount of citrus grown (only 3.0%),
yet it is important to consider because the risk of
contamination is high. Remaining areas have substantial
quantities of citrus and thus should not be eliminated from
consideration.
These five geometries are grouped into three general aquifer
configurations for modeling: 1) the Floridan, as a single
unconfined aquifer (Area 1), 2) the unconfined surficial (Area
4), and 3) a general two-aquifer system (Areas 2, 3, and 5),
be it the surficial aquifer overlying the Floridan or an
intermediate confined aquifer.
The three-aquifer system would only be considered if
significant contamination is found in the lower of the
two-aquifer system.
2.2.2 Hydrogeologic Properties
The hydrogeologic properties in each of the aquifers determine
the ground-water flow regime. They describe the properties of
the porous medium that collectively affect the velocity of the
water and transport of aldicarb in the aquifers and confining
layers. The five principle hydrogeologic properties used in
the saturated zone modeling are each defined and briefly
described.
103
-------
TABLE 2.11 APPROXIMATE CITRUS ACREAGE ASSOCIATED WITH EACH
AQUIFER GEOMETRY
County
(in ranked
order)
Polk
Lake
St. Lucie
Indian River
Orange
Hardee
Martin
Hillsborough
DeSoto
Pa sco
Hendry
Osceola
Brevard
Manatee
Palm Beach
Marion
Volusia
keechobee
Lee
Hernando
Seminole
Charlotte
Glades
Pinellas
Surater
Citrus
Total
%
Area
123
31,000 63,000
87,000
10,000
28,500
42,000
14,000 17,500
16,500
13,000 17,000
12,000
8,000
2,000 2,000
2,500
2,000
5,000 700
600
2,000
600 1,000
1,000
21,600 212,300 147,000
3.0 29.5 20.4
4
17,000
64,000
50,000
14,000
32,000
3,500
13,000
2,500
15,000
5,000
7,500
4,500
5,000
5,000
4,500
1,300
243,800
33.8
5
31,000
7,000
6,500
3,500
30,000
2,500
5,000
4,500
1,000
1,000
4,000
96,000
13.3
104
-------
1. Hydraulic conductivity (L/T) is a measure of the media
properties that determine the ease with which water
is transmitted through a porous medium. High
conductivities mean it is easily transmitted, low
conductivities mean water moves slowly.
2. The hydraulic gradient (L/L) is the vertical change in
water level over a given horizontal distance. The
gradient acts as the driving force to flow. For a
given porous medium the flow will be greater the
larger (steeper) the hydraulc gradient.
3 3
3. Porosity (L/L ) is the percent void space in a repre-
sentative volume of the porous medium. The porosity
affects the ground-water velocity. High porosity
material has lower ground-water velocities than low
porosity material for the same average flux rate
(specific discharge).
4. Saturated thickness (L) of an aquifer is the thickness
in which all the pores are filled with water. When
the water supply well is in the lower aquifer, the
saturated thickness of the unconfined aquifer and
confining layer will influence the vertical travel
time of the pesticide.
5. Recharge rate (L/T) is the average annual rate at
which water is replenished to the aquifer by
percolation from the unsaturated zone.
Each of these parameters takes on a large range of values in
the aquifer systems of central Florida. Two combinations of
these property values were used for each of the three aquifer
configurations in the saturated zone modeling. A "worst case"
105
-------
was modeled using the extreme values of the hydrogeologic
properties that collectively move aldicarb through the
saturated zone most quickly. The other combination of
properties will use intermediate values that best characterize
each of the three aquifer configurations.
2.2.3 Influence of Drinking Water Well
The drinking water well influences the movement of the ground
water and therefore the transport of aldicarb. The well
position relative to the pesticide source area, the depth of
the well and the rate at which it pumps all influence the
ground-water flow regime.
2.2.3.1 Well Distance from Aldicarb Application Site
The distance of a drinking water supply well from the area of
pesticide application is crucial in the evaluation and
assessment of potential human exposure to aldicarb through
drinking water. The greater the distance necessary to travel
from the source area to the well through the ground water, the
smaller the chance of contamination. The distance of the well
from the source is an important management consideration for
EPA. The current EPA standards prohibit the use of aldicarb
within 91 m (300 ft) of a water supply well. This distance
was evaluated in all the scenarios. The EPA is considering
300 m (1000 ft) as an alternative distance in their evaluation
of the aldicarb regulations. This distance was also simulated
in all the scenarios.
2.2.3.2 Well Depth
In the Florida citrus growing area, water wells vary
106
-------
considerably in depth depending on the local aquifer
characteristics and the water needs. Deep wells with large
capacities are usually drilled to supply municipalities.
Shallow wells have much smaller capacities and supply the
small domestic user. The depth that the water is withdrawn
from a well influences the localized aquifer flow system. A
shallow well affects the near surface flow system and
therefore would influence the movement of a leaching
contaminant sooner than a deep well with an equivalent pumping
rate.
2.2.3.3 Well Pumping Rates
The well rate determines the magnitude of the influence on the
regional flow system. A low pumping rate has a small
localized effect. High pumping rates can change the flow
system on a much larger scale, greatly increasing the
ground-water velocities towards the well. Two well rates, a
high and a low rate, were simulated in each hydrologic
setting.
2.2.4 Delineation of Saturated Zone Scenarios
The factors influencing aldicarb transport are reviewed in
Table 2.12 with the cases that were considered for each
factor. All the combinations of these factors resulted in 48
modeling scenarios. This number doubled to 96 potential
scenarios because of the consideration of two chemical decay
rates. Two decay rates were used; 1) the best estimate of a
representative rate, and 2) no decay.
The following notation convention is used for identifying each
scenario.
107
-------
TABLE 2.12 CASES CONSIDERED FOR SATURATED ZONE MODELING
Influencing Factors
Cases Considered for Each Factor
Aquifer geometry
1) Unconfined Floridan Aquifer
2) Unconfined surficial aquifer
3) Leaky two-aquifer system
Aquifer properties
1) Worst case
2) Intermediate case
Well distance
1) 300 ft
2) 1000 ft case
Well depth
1) Shallow
2) Deep
Well rate
1) High rate
2) Low rate
108
-------
Aquifer Geometries
F = Floridan Unconfined
S = Surficial Unconfined
2 = Two Aquifer System
Aquifer Properties
W = Worst Case
A = Average Case
Decay Rates
D (rate in days )
Well Configurations
W (distance in feet, depth in feet,
rate in gpm)
FWD(0)W(300,50,2000), for example, would identify the scenario
that uses the unconfined Floridan Aquifer with the worst set
of hydrogeologic properties, no decay rate, a well distance of
300 ft from the aldicarb source, a well depth of 50 ft, and
pumping rate of 2000 gallons per minute (gpm).
2.2.5 CFEST Model Parameters
2.2.5.1 Hydrogeologic Parameters
The hydrogeologic parameters necessary for modeling with CFEST
detex'mine the ground-water flow regime that collectively
affect the velocity of the water and transport of aldicarb in
109
-------
the aquifers and confining layers. The equation for the
average ground-water velocity shows the relationship between
the hydrologic properties used by the model:
V = (2.1)
where Y = average ground-water velocity
K = hydraulic conductivity
I = hydraulic gradient and
e = porosity
Hydraulic conductivity, hydraulic gradient and porosity are
the most important properties affecting the speed of
ground-water flow. The saturated thickness of the aquifers
and confining layers and the annual recharge rate also
influence the hydraulics of the system as well as being
necessary modeling parameters. Each of these parameters takes
on a large range of values in the aquifer systems of central
Florida. The worst and average set of hydrogeologic values
used for modeling for each aquifer system are shown in Table
2.13. The determination of numerical values for these five
hydrogeologic parameters are discussed below.
2.2.5.1.1 Hydraulic ConductivityHydraulic conductivities
can be measured in the field or in the lab. Field
measurements are better for modeling studies because they use
several wells and test larger aquifer-scale properties. Lab
tests on the other hand are done on small samples. Therefore
the properties measured are extremely localized.
Unfortunately, field testing of hydraulic conductivities is
more expensive and not as common as lab measurements. A third
method to determine the hydraulic conductivities is to model a
particular area. When only limited data are available,
110
-------
TABLE 2.13 AVERAGE AND WORST CASE HYDROGEOLOGIC PARAMETERS FOR THE THREE
AQUIFER CONFIGURATIONS
Hydraulic Conductivity
Saturated Thickness
W Hydraulic Gradient
K Porosity
W
Recharge Rate
Floridan Aquifer
Unconfined
1000 ft/d
600 ft
2 ft/ 1 mile
.30
.20-. 35
13 in/yr
Surficial Aquifer
Unconfined
30 ft/d
75 ft
2 ft/1 mile
.20
.15-. 25
12 in/yr
TV
Surficial
23 ft/d
75 ft.
5 ft/1 mile
.20
.20-. 30
10 .in/yr
ro-Aquifer System
Confining Layer
lxlO-*ft/d
40 ft
AH = 35 ft .
.10
.05-. 10
10 in/yr
Floridan
200 ft/d
600 ft
4 ft/mile
.30
10 in/yr
Hydraulic Conductivity
Saturated Thickness
Hydraulic Gradient
Porosity
Recharge Rate
Floridan Aquifer
Unconf indd
5000 ft/d
600 ft
3 ft/1 .mile
.20
25 in/yr
Surficial Aquifer
Unoonfined
100 ft/d
75 ft
4 ft/1 mile
15
30 in/yr
TV
Surficial
40 ft/d
40 ft
25 ft/1 mile
.20 .
20 in/yr
/o-Aquifer System
Confining Layer
2. 5x10-* ft/d
10 ft
AH = 40 ft.
05
20 in/yr
Floridan
1770 ft/d
100 ft
10 ft/1 mile
.20
20 in/yr
Conversions:
1 ft = .3048 m
1 in. = 2.54 cm
-------
duplicating the regional historical hydrologic conditions with
computer simulation can often produce a good regional picture
of the hydraulic conductivites. All three of these sources of
data on hydraulic conductivity were used to determine
parameter values for this modeling study.
In the area where the Floridan is unconfined (Area 1), Ryder
(1982) has done a computer simulation model of historical
conditions in which he incorporated the existing field and lab
data on the hydrologic parameters. This provided a good
source of regionally calibrated aquifer transmissivities.
From these and the saturated thickness, hydraulic
conductivities were calculated, ranging from several hundred
4
ft/day (7.06 x 10 m/s) to over 5000 ft/day (.02 m/s). These
were broken down into an average value of 1000 ft/day (3.53 x
10 m/s) and worst case value of 5000 ft/day (0.02 m/s) as
shown in Table 2.12.
The surficial aquifer is unconfined and considered along all
the east and west coasts (Area 4). Unfortunately there is
very little technical data on the surficial aquifer. Scott
(1977), for example, reports hydraulic conductivities that
-6 -4
range from 1 to 130 ft/day (3.5 x 10 to 4.6 x 10 m/s).
These values were calculated from lab tests and a few field
tests to set the limits. Intermediate values were calculated
using lithologic and physical properties. In Martin County,
Lichtler (1960) reports field pump test results from which
hydraulic conductivities can be calculated which range from
about 20 to 150 ft/day (7.1 x 10 ~5 to 5.3 x 10" m/s) and
averaging 40 ft/day (1.4 x 10 m/s). Values reported and
calculated for St. Lucie County (Bearden, 1972), Charlotte
County (Wolansky, 1978) and Lee County (Wedderburn et al.,
1982) fall into this same range.
A two-aquifer system is modeled in the central portion of the
112
-------
peninsula (Areas 2, 3 and 5 in Figure 2.18). For this system
hydraulic conductivity is needed for the confining layer as
well as the surficial unconfined aquifer and the confined
aquifer. Data for hydraulic conductivity in both the surficial
aquifer and confining layer are limited. Lichtler (1972)
reports values ranging from 5 ft/day to 40 ft/day (1.8 x 10
to 1.4 x 10 m/s) for the surficial aquifer in the
east-central area. Two pump tests in the surficial aquifer in
the Polk County region showed hydraulic conductivities of 55
and 32 ft/day (1.9 x 10~4and 1.1 x 10 ~4 m/s) (Hutchinson,
1978). Hydraulic conductivities estimated from transmissivi-
ties in DeSoto and Hardee Counties (Wilson, 1977) and in
west-central Florida (Ryder, 1982) are in the same range.
Weighted averages based on the quality of the data and the
quantity of citrus resulted in the average and worst case
values shown in Table 2.13.
A similar process was used to determine the hydraulic
conductivity of the confining layer. Wilson and Gerhert
(1982), in their modeling study of west-central Florida,
specifically calculated vertical hydraulic conductivities.
Values from their study were compared to values estimated from
leakances in the middle gulf area (Cherry et al., 1970) and
Highland County (Bishop, 1956). The resulting worst case and
average case values are shown in Table 2.13.
The confined aquifer in the two-aquifer system could be either
an intermediate aquifer or the Floridan Aquifer. Because the
intermediate aquifers are usually localized and not very
laterally extensive, the properties of the Floridan Aquifer
are used to represent the confined aquifer. Three regional
modeling studies (Tibbals, 1982; Wilson and Gerhert, 1982; and
Ryder, 1982) were used to determine the average and worst case
hydraulic conductivities. These values compared favorably
with values estimated from transmissivities reported in the
113
-------
local county reports sited above.
2.2.5.1.2 Hydraulic Gradient The hydraulic gradient is the
vertical change in hydraulic head (water level or piezometric
surface) over a given horizontal distance. In the confining
layer the driving force is the difference in hydraulic head
between the upper and lower aquifers. The gradient is the
major driving force for ground-water flow. For a given porous
medium the flow will be greater the larger (steeper) the
hydraulic gradient. The hydraulic gradient can be measured
directly on a map of the water table or piezometric surface.
Figure 2.19 shows the piezometric surface of the Floridan
Aquifer throughout Florida. A more detailed picture of the
hydraulic gradient where the Floridan is unconfined was
obtained from a regional modeling study (Ryder, 1982).
Regional studies in the central peninsula where the Floridan
is confined (Hutchinson, 1978; Cherry et al . , 1970; and Wilson
and Gerhert, 1982) helped to estimate the extreme values of
the hydraulic gradient.
The same approach was used to determine the hydraulic gradient
for the surficial aquifer. Healy (1982) has compiled water
level maps for the surficial aquifer in the areas where it is
the principal aquifer used for drinking water (equivalent to
Area 4). Bearden (1972) showed water level maps with
gradients up to 7.58 x 10 (4 ft/mi.). Most gradients ranged
from 2.84 x 10~4(1.5 ft/mi.) to 6.63 x 10 ~4 (3.5 ft/mi.)
(Land et al., 1973; Lichter, 1960; and Grain et al . , 1975).
Where the surficial aquifer is part of the two-aquifer system
in the ridge area the gradients are much more variable due to
more variable topography. Hutchinson, 1978, reports
4
hydraulic gradients from 4.73 x 10 (2.5 ft/mi.) to 4.73 x
10 ~^ (25 ft/mi.). Most areas have gradients that average 9.47
x 10 ~4 (5 ft/mi.) (Cherry et al., 1970; Sinclair, 1974; and
Wilson and Gerhert, 1982).
114
-------
Jacksonville
Atlantic Ocean
Datum is Mean Sea Level
20 Potentiometric Contour, in feet
Figure 2.19 Potentiometric surface of the Floridan Aquifer
After: Healyr 1975.
115
-------
In the confining layer the driving force is the difference in
hydraulic head between the upper and lower aquifer. Wilson
and Gerhert (1982) present generalized hydraulic head maps for
both the upper and lower aquifers. From these maps the
difference in heads is estimated to vary between 0 and 30.5 m
(100 ft) and averaging about 10 m. Within these limits the
hydraulic gradients were estimated, checking the thickness of
the surficial aquifer and confining layer to ensure a
physically realistic system.
2.2.5.1.3 PorosityPorosity is the percent void space in a
representative volume of porous media. Below the water table
in the ground water all the void space is filled with water.
As shown in equation 2.1, the porosity affects the
ground-water velocity. Given that all other hydrologic
properties are the same, a low porosity will result in higher
ground-water velocities. This happens because there are fewer
spaces for the water to flow through so the water must flow
faster for a given specific discharge.
The limestone geology of the Floridan peninsula results in
extremely variable porosities. In areas where solution
cavities and channels are common, porosities are easily over
.50. Other areas have layers of dolomite and chert which are
more resistant to dissolution where the porosities can be as
low as .05. This makes estimation of the bulk aquifer
porosity very difficult. A lab determined porosity of a
massive limestone can not accurately represent the solution
features that are present. The modeling studies reviewed did
not consider transport so porosity was not a necessary
parameter. Stewart (1966) reported a range of porosities from
.15 to .44 calculated in the lab for Polk County. Tibbals
(1984, personal communication) suggested bulk porosities in
the Floridan Aquifer would probably be about .3 to .35
116
-------
although not at all uniformly distributed. Some U.S.
Geological Survey modeling studies have used a porosity of .2
for the Floridan with good success (Frazee, 1984, personal
communication). Solution features are most highly developed
in the northwest area where the Floridan is unconfined. Based
on data available the porosities shown on Table 2.13 were used
for modeling. Because this parameter was not very well
defined, some sensitivity runs were done varying the porosity
values. Porosities of .15 to .20 for the surficial aquifer
and .05 to .10 for the confining layer of the two-aquifer
system were determined in a similar manner (Tibbals, 1984,
personal communication; Hutchinson, 1978; Cherry et al., 1970;
and Knochenmus and Hughes, 1976). The worst case porosity is
less than the average porosity because the lower porosity will
increase the rate of ground-water movement as discussed at the
beginning of this section.
2.2.5.1.4 Saturated ThicknessWhen modeling the transport of
aldicarb through an aquifer to a well the critical depth is
the depth of the well rather than the full depth of the
aquifer. Parts of the Floridan Aquifer are over 600 m thick
(Causey and Leve, 1976). When a municipal supply well only
penetrates the upper 100 m it is not critical that the exact
aquifer thickness be known or modeled.
When modeling a multi-aquifer system, however, the thickness
of the upper unconfined aquifer and the intervening confining
layer are critical factors determining the speed of transport
to a well pumping out of the lower aquifer. The following
analysis demonstrates the importance of the saturated
thickness to the average vertical velocity of the water. The
travel time through the upper aquifer and confining layer to a
well is equal to the quotient of the vertical distance
traveled and the velocity. By substituting Darcy's Law for
velocity it is shown that:
117
-------
t = i (2.2)
KAH
where b = saturated thickness
t = travel time
e = effective porosity
K = hydraulic conductivity and
AH = difference in hydraulic heads between the upper
and lower aquifers
This analysis shows that the saturated thickness (b) is the
most important factor in determining the travel time, since it
is the only squaired term.
In Area 1, where the Floridan Aquifer is unconfined, an
average saturated thickness of 183 m (600 ft) was determined
from Ryder (1981) and Causey and Leve (1976). In Area 4,
where the unconfined surficial aquifer is considered alone,
the average saturated thickness of 23 m (75 ft) was determined
from values sited in county reports pertaining to Area 4
(Wedderburn et al., 1982; Wolansky, 1978; Lichtler, 1960;
Lichtler, 1972; and Bearden, 1972). The same values were used
for both the average and the worst cases because the saturated
thickness is not a critical value in the case of the single
aquifer as discussed above.
The same procedure was used to determine the saturated
thickness for the two-aquifer system (Areas 2, 3 and 5). The
surficial aquifer is extremely variable in thickness in the
central highlands. In Polk County it can vary from less than
15 m (50 ft) to as much as 75 m (250 ft) (Stewart, 1966 and
Hutchinson, 1978). The average saturated thickness is 23 m
(75 ft), the same as for the surficial aquifer in the coastal
118
-------
area. The vertical travel time decreases as the saturated
thickness decreases as shown in the analysis above. The worst
case value is 12 m (40 ft).
The confining layer is also extremely variable in thickness.
In areas where the clays are breached by a sink hole, the
thickness is effectively zero. Knochenmus and Hughes (1976)
report that in parts of Lake County the thickness of the
confining layer reaches as much as 30.5 m (100 ft). Buono et
al. (1979) map the thickness of the confining bed in parts of
this area as less than 7.5 m (25 ft) to over 75 m (250 ft).
The average thickness is 12 m (40 ft) and the worst case is 3
m (10 ft). Where the confining layer is actually breached the
unconfined Floridan configuration can be used to assess
pesticide exposure.
The Floridan Aquifer is estimated to be 180 m (600 ft) in the
central area. As in the case of the single aquifer, the
saturated thickness is not as critical a value as the depth of
a drinking water supply well.
2.2.5.1.5 RechargeRecharge is the average annual rate at
which water is replenished to the aquifer by percolation
through the urisaturated zone. The amount of water that is
recharged to the ground water can be roughly figured by a mass
balance analysis, where:
Recharge = Precipitation + Irrigation - Runoff
- Evapotranspiration
Other factors influence the recharge rate as well as the
supply (precipitation and irrigation). In areas where soils
are poorly drained recharge rates are close to zero. The well
drained soils typical of the ridge area (entisols and
ultisols) enhance recharge rates. Lakes and natural
119
-------
depressions such as sink holes create ideal areas for
recharge. Another factor that affects the recharge rate is
the height of the water table. If the water table is near the
land surface, extra water added to the aquifer can be
evaporated more easily.
Stewart (1980) delineates areas of high, moderate, low and no
recharge to the Floridan Aquifer. This map gives a general
picture of recharge both where the Floridan is unconfined
(Area 1) and where it is confined (Areas 2, 3, and 5). On the
average, recharge is higher where the Floridan is unconfined.
In the central portion of the peninsula where the Floridan
Aquifer is confined there are areas of high recharge along the
ridges, but much of the area has low to moderate recharge.
Estimates of recharge from the map were compared to rates
reported in the literature. Ryder (1982) reports recharge
rates that range from 0 to 50 cm/yr (20 in/yr) in the area
where the Floridan is unconfined. Where the Floridan is
confined recharge is reported anywhere from about 7 cm/yr (2.6
in/yr) (Hutchinson, 1978) to 38 cm/yr (15 in) (Knochenmus and
Hughes, 1976). The rates used in the modeling study are shown
in Table 2.13. Both the average and the worst case recharge
rates also reflect the effects of added recharge from
irrigation.
Recharge to the surficial aquifer has not been well studied
where the Floridan is not considered. A rough mass balance
approach was used to estimate the recharge rates. Visher and
Hughs (1969) have published a map showing the difference
between rainfall and potential evaporation. This was used
along with estimates of runoff (Hughes, 1976) and irrigation
to calculate the average and worst case recharge rates for the
surficial aquifer.
120
-------
2.2.5.2 Drinking Water Well Parameters
For the purposes of modeling well depths and pumping rates are
necessary. Water supply wells of different depths were
modeled in each hydrogeologic setting based on typical wells
in each area. Where the Floridan Aquifer is unconfined both a
deep well and a shallow well were simulated. The deep
municipal wells in the Floridan Aquifer typically draw water
from about 110 m (350 ft) of well which is open to the aquifer
(Tibbals, 1984, personal communication and Spangler, 1984,
personal communication) . The shallov/ wells vary from less
than a meter up to about 30 m (100 ft), with an average 12 m
(40 ft).
In the case where there is just a single unconfined surficial
aquifer, only a shallow well was modeled. Well depths vary
considerably in the surficial aquifer. An average depth of
10.7 m (35 ft) was used in the simulation.
Since the two-aquifer system is modeled to assess the
possibility of the pesticide being transported through the
confining layer to the confined aquifer, only a deep well in
the lower aquifer was simulated. Wells tapping a confined
aquifer are typically open in the highest producing zone and
have about 100 m (350 ft) open in the aquifer.
The pumping rates modeled with each hydrogeologic scenario
reflect the two most common drinking water supply situations:
a small (low yielding), domestic supply well and the large,
high volume, municipal supply well. Representative rates were
chosen based on well yields and annual water usage from the
communities and well records in the citrus growing area.
The largest capacity wells are in the confined Floridan
Aquifer. Large municipal supply wells have rates as high as
121
-------
94.6 1/s (1500 gpm). Deep wells in the confined Floridan
Aquifer also have high rates. The shallow wells in the
surficial aquifer have much lower rates ranging from 2.5 1/s
(40 pgm) to 12.6 (200 gpm). Table 2.14 shows the well depths
and rates used with each hydrogeologic setting.
122
-------
TABLE 2.14 WELL RATES AND DEPTHS FOR SATURATED ZONE MODELING
Aquifer Configuration
Surficial
Floridan unconfined
Two-aquifer system
(1 ft = .3048 ra)
(1 gpm = 0.063 1/s)
Well Depth
35 ft
40 ft
350 ft
350 ft (open)
Well Pumping Rate
40 gpm
200 gpm
40 gpm
200 gpm
500 gpm
1000 gpm
700 gpm
1500 gpm
123
-------
SECTION 3
CHEMICAL FATE AND TRANSPORT
Aldicarb is a nematicide, acaricide, and a systemic
insecticide. Its environmental fate is dominated by two
factors: the fact that it forms two toxic daughter products,
and its high mobility in soils. Degradation of the toxic
residues of the compound is of intermediate duration compared
to other pesticides.
This discussion will be broken into three primary sections;
the first dealing with factors which affect aldicarb fate
(i.e., transformation and decay), the second dealing with
factors affecting transport (i.e., adsorption partition
coefficients) and the third combining those two factors in the
analysis of pesticide parameters for the unsaturated and
saturated zone.
3.1 ALDICARB FATE
Aldicarb is a white crystalline solid which is incorporated
into soil as a granule containing either 10% or 15% active
ingredient. In order to be effective, it must dissolve in
124
-------
water. Once this happens in soils, the compound begins to
transform and degrade.
The current theory is that aldicarb, 2-methyl-2-(methylthio)
propionaldehyde 0-(methylcarbamoyl) oxime, is fairly rapidly
oxidized to aldicarb sulfoxide, 2-methyl-2-(methylsulfinyl)
propionaldehyde 0-(methylcarbamoyl) oxime which in turn is
more slowly oxidized to aldicarb sulfone, 2-methyl-2-
(methylsulfonyl) propionaldehyde 0-(methylcarbamoyl) oxime.
Concurrently, these three carbamates are transformed by
hydrolysis to corresponding oximes. Hydrolysis is a chemical
reaction in which water breaks up an organic molecule (RX),
such as aldicarb, by breaking a carbon-X bond and replacing it
with OH from the water molecule:
R-X + H0*>R-OH + X~+ H+ (3.1)
These products of hydrolysis are far less toxic than aldicarb,
its sulfoxide or its sulfone and are of little environmental
concern (Smelt et al., 1978a). A schematic of these processes
is shown in Figure 3.1.
3.1.1 Transformation and Degradation Rates in Soils
A number of studies up until 1978 were performed on the
pesticide aldicarb as reported by Smelt et al., 1978c.
Unfortunately, most of these studies only measured the residue
remaining in soil after application and therefore cannot be
used to identify rates for individual transformation and
degradation processes. Smelt et al., 1978c, identified these
field loss rates for aldicarb being, from 0.015 to 0.42/day.
125
-------
Aldicarb
Aldicarb Sulfoxide
Aldicarb Sulfone
CH.S - C - CH - NOCNHCH.
3 |
O CH0
II I 3
CH3S - C - CH
CH.,
II II I
= NOCNHCH., -CH,S - C - CH =
'II
I
CH.
O
||
NOCNHCH,
(Hydrolysis)
(Hydrolysis)
to
Nontoxic Oximes and Nitrites
Figure 3.1 Schematic of aldicarb environmental chemical Pathways
-------
They also performed independent experiments to measure
oxidation and hydrolysis rate constants for the pathways shown
in Figure 3.1. These computed rate constants are shown in
Table 3.1. Other workers (Bromilow et al., 1980 and Leistra,
et al., 1976) have performed similar experiments which are
also shown in Table 3.1.
3.1.2 Hydrolysis Rates in Water
Hansen and Spiegel (1983) measured the hydrolysis rates of
aldicarb, aldicarb sulfoxide and aldicarb sulfone in distilled
water at various levels of temperature and pH. These values
are shown in Table 3.2, taken from their report. Lemley and
Zhong (1983) measured acid and base catalyzed hydrolysis of
aldicarb, aldicarb sulfoxide and aldicarb sulfone. Second
order rate constants were calculated for data taken under
extremely acidic and basic conditions. Rate constants for
base catalyzed hydrolysis of aldicarb, aldicarb sulfoxide and
aldicarb sulfone were 0.94 liter/mole-min, 10.5 liter/mole-min
and 40.3 liter/mole-min, respectively. At pH 7.5 these
correspond to first-order rates of 0.0004, 0.0048 and
0.0184/day. These rates agree well with the data of Hansen
and Spiegel for aldicarb but are roughly 4 times greater for
the sulfoxide and 3 to 10 times greater for the sulfone.
They also reported strong temperature effects (14 fold over a
5 deg. C to 35 deg. C range) on base catalyzed hydrolysis of
the sulfone and a 'depression of hydrolysis in the presence of
neutral electrolytes. However, in the presence of 1 m NaCl,
the rates at 15 deg. C were not quite halved.
127
-------
TABLE 3.1
FIRST ORDER RATE CONSTANTS FOR OXIDATION AND HYDROLYSIS OF
ALDICARB IN SOIL
to
00
SOIL
Clay Loam
(fresh)
Clay Loam
(stored)
Greenhouse
(fresh)
Greenhouse
(stored)
Peaty Sand
(fresh)
Peaty Sand
(stored)
Sandy Loam
Clay Loam
Clay Loam
Greenhouse
Soil
Greenhouse
Soil
Sandy Loam (3)
Sandy Loam (5)
Nagele. Loam
Westmaas Loam
Ki
0.30
0.30
0.096
0.160
0.130
0.077
0.210
0.30
0.44
0.21
0.80
0.80
0.20
0.27
0.14
0.46
0.55
0.25
0.25
RATE CONSTANT (/day)
KJ K| K^
0.011
0.008
0.004
0.010
0.008
0.006
0.006
0.005
0.020
0.0036
0.027
0.015
0.033
0.034
0.035
0.025
0.011
0.030
0.013
0.031
0.031
0.015-0.03
;o. oi-o. 06
0.015
0.030
0.000
0.000
0.002
0.0007
0.005
0.05
0.07
0.06
0.07
0.07
0.04
0.065
0.07
0.07
0.07
0
0
0.013
0.018
0.0015
0.005
0.006
0.005
0.008
0.004
0.030
0.0016
0.013
0.003
0.010
0.006
0.018
0.021
0.001
0.004
0.002
0.006
0.01
.00-0.02
.00-0.05
K5
0.03
0.021
0.004
0.013
0.0045
0.0045
0.012
0.008
0.050
0.004
0.033
0.012
0.020
0.013
0.021
0.016
0.005
0.010
0.005
0.012
0.015
0.00-0.08
0.00-0.14
TEMP
c
15
15
15
15
15
15
15
6
25
6
25
5
10
15
IS
15
5
10
15
15
15
20
20
PH
(su)
7.2
7.2
6.1
6.1
5.1
5.1
7.4
7.2
7.2
6.1
6.1
7.0
7.0
7.0
7.0
7.0
6.3
6.3
6.3
6.3
6.3
7.2-7.4
7.3-7.6
ORGANIC
MATTER
(*)
4.5
4.5
17.2
17.2
9.5
9.5
1.7
4.5
4.5
12.2
17.2
1.35
1.35
1.35
1.35
1.35
5.92
5.92
5.92
5.92
5.92
2.5-4.6
1.1-4.7
HATER
CONTENT
(%)
30.0
30.0
66.0
66.0
19.0
19.0
13.5
30.0
30.0
66.0
66.0
10.0
J.0.0
5.0
10.0
15.0
10.0
10.0
5.0
10.0
15.0
REFERENCE
Smelt et al.
M mm
m mm
m mm
m mm
m mm
m mm
Smelt et al.
mm
m mm
m mm
Bromilow et al
m
m
m
m
m
"
m
m
m
Leistra et al.
mm
1978c
m
m
m
m
*
H
1978b
II
. 1980
1976
-------
TABLE 3. 2
FIRST ORDER HYDROLYSIS RATE CONSTANTS FOR ALDICARB; ALDICARB SULFOXIDE
AND ALDICARB SULFONE
to
«£>
COMPOUND
Aldicarb
Sulfoxide
Sulfone
pH
5.5
5.5
7.5
7.5
8.5
8.5
5.5
5.5
7.5
7.5
8.5
8.5
5.5
5.5
7.5
7.5
8.5
8.5
TEMP (°C)
5
15
5
15
5
15
5
15
5
15
5
15
5
15
5
15
5
15
PERIOD
(days
1-186)
1=186
1-186
1-186
1-186
1-186
1-186
1-278
1-278
1-186
1-186
1-186
1-83
1-186
1-186
1-186
1^186
1-186
1-28
RATE CONSTANT
(d-«)
0.000151
0.000214
0.000354
0.000363
0.000501
0.00407
0.000371
0.00158
0.00107
0.00191
0.0107
0.063
0.000741
0.00155
0.00110
0.00550
0.0200
0.134
HALF-LIFE
(d)
4,580
3,240
1,950
1,900
1,380
170
800
440
650
360
65
10
930
450
630
125
35
5
Source: Hansen and Spiegel, 1983
-------
3.2 TRANSPORT CONSIDERATIONS
3.2.1 Adsorption
Aldicarb and its two daughter products are very water soluble,
which indicates a high mobility in soils. Cohen et al.,
(1983) reported solubilities of 6,000, 43,000 and 7,800 ppm
for aldicarb, sulfoxide and sulfone, respectively, although
Hornsby et al., (1984) give slightly different values of
6,000, 33,000 and 8,000 ppm.
Using these solubilities and the regression equation of Kenaga
and Goring (Lyman, 1982), Koc values of 36 cm /g for aldicarb,
12 to 14 cm /g for aldicarb sulfoxide and 31 to 32 cm /g for
aldicarb sulfone are computed.
Based on the data of Bromilow et al. (1980) (reported as K, )
for a sandy loam soil and a sandy loam soil with addition of
peat, Koc for aldicarb, aldicarb sulfoxide and aldicarb
33 3
sulfone were 16-70 cm /g, 4.6 cm /g and 1 to 5 cm /g,
respectively. Koc is calculated by:
K = 0.2)
°C (OC)
where K = the adsorption partition coefficient (cm /g)
OC = the fraction organic carbon
Hornsby et al. (1984) calculated values from 0 to 47 cm /g
for aldicarb and 0 to 18 cm /g for aldicarb sulfone at a
Florida ridge citrus site and 0 to 38 cm /g for aldicarb and
0 to 11 cm /g for aldicarb sulfone at a flatwoods citrus site.
130
-------
They also reported Koc calculated for other soils which
ranged from 19
aldicarb sulfone.
ranged from 19 to 25 for aldicarb and 0 to 6 cm /g for
The various values of Koc collected from the literature are
summarized in Table 3.3.
Note that Hough et al. (1975) found adsorption coefficients
for aldicarb sulfoxide to be an order of magnitude higher in a
soil with high clay content but having the same organic matter
content as another soil with a lower coefficient.
Supak (1972), on the other hand, found aldicarb to actually be
excluded by Ca- and Al- saturated montmorillonite clays. Ca-
and Al- saturated illite and kaolinite clays showed weakly
positive adsorption isotherms and adsorption was thought to be
through interaction with water on the external clay surface.
3.2.2 Volatilization
Because aldicarb is applied in a granular formulation and
incorporated it does not seem that volatilization would be an
important process. However, it does have a moderate vapor
pressure (1x10 ~^ mm Hg @ 25 deg. C) compared to other
pesticides. Henry's Law Constant calculated using this
information (Thibodeaux, 1979) gives a value of 1.7 x 10
3 / 3
cm -water/cm -air. This value is comparable to that of other
pesticides such as DDT, dieldrin, trifluralin and EPTC.
Bull et al. (1970) determined the effects of soil moisture and
temperature on aldicarb volatilization from vials containing a
treated layer of sand capped by 5.5 cm of untreated sand. In
their experiments, volatile losses of aldicarb ranged from
131
-------
TABLE 3.3 KQC FOR ALDICARB AND ITS DAUGHTER PRODUCTS
K (cm3/g)
oc
Aldicarb Sulfoxide Sulfone
Calculated from
Solubility data 36 12-14 31-32
Bromilow et al.f 1980
Woburn (3) 70 - 1
Woburn (5) 16 5 5
Hornsby et al., 1984
Lake Hamilton site 0-47 - 0-18
Oviedo site 0-38 - 0-11
Eustis soil 19-25 - 0-6
Webster soil 21 - 6
Cecil soil 19-0
Grenada soil 19 -
Hough et al., 1975
Holtville clay - 236
Buren silt loam - 24
132
-------
16.7% to 89.0% over a 24 hr. period. Temperatures ranged from
25 deg. C to 75 deg. C. Losses were generally higher from wet
than from dry sand. First-order volatilization rate constants
computed over the 24 hr. period were 0.0036 and 0.0076 /day
for the wet and dry sand at 25 deg. C and 0.0347 and 0.0180
/day for wet and dry sand at 50 deg. C.
Other evidence presented in Supak (1972), as reported in
INTERA (1980), indicates that volatilization losses are
extremely low. Supak et al. (1977), report extremely low
losses (0.01 to 0.18%) over an 18 day period in two Texas
soils. They also found volatilization losses to be higher
from a dry soil than from a wet soil although volatilization
ceased under very dry conditions.
Richey et al.r 1977, radio-labeled aldicarb at the S-methyl,
N-methyl and tertiary carbon atoms to study molecular
fragmentation under soil degradation. As a result, they found
virtually all of the radio activity captured as volatiles was
in the form of CO and less than 1% captured as volatiles was
in the form of aldicarb, aldicarb sulfoxide or aldicarb
sulfone. Thus, based on this study, it appears that little
volatilization of the parent and major daughters would occur.
3.2.3 Plant Uptake
Some plant uptake literature was reviewed by Pacenka and
Porter (1981). Their investigations revealed that little
quantitative data is available. Based on limited information,
they estimated that 4 to 20% of the applied chemical might be
a suitable range. Biggs and Webb (1984, personal
communication) have data for the uptake of aldicarb by young
Valencia orange trees. A summary of their data indicates that
133
-------
as much as 34% of the applied chemical was taken up by plants
by 30 days after application (Biggs and Webb, 1983).
Approximately 20% of applied aldicarb was calculated to have
been taken up by 20 year old trees in the field (Biggs and
Webb, 1984, personal communication).
3.3 ANALYSIS OF PESTICIDE MODEL PARAMETERS
From the information identified from the literature, the
parameters needed for simulation of pesticide fate and
transport using PRZM and CFEST were derived. In the following
analyses the decay rates of. aldicarb and its daughter products
are estimated in the unsaturated and saturated zones based on
the appropriate physical properties. In the saturated zone
the pH and temperature of the ground water are the most
important factors influencing the pesticide degradation. In
the unsaturated zone these factors are important as well as
soil water content and amount of organic carbon.
3.3.1 Unsaturated Zone
Three different data sets (Bromilow et al., 1980; and Smelt et
al., 1978b and 1978c) were identified in which laboratory
determined rates of aldicarb conversion to the sulfoxide (k^),
the conversion of the sulfoxide to the sulfone (k ), and the
degradation of each species (k , k and k ) to nontoxics were
34 5
estimated. In these studies, measurements were also made of
levels of temperature, pH, organic matter content and water
content at which the experiments were conducted. Eight
different soils were used having a pH range from 5.1 to 7.4.
Temperatures ranged from 5 to 25 deg. C, organic matter from
. 134
-------
1.35 to 17.2% and water contents from 5 to 66%. Seventeen
meaurements of k, and k , were made versus these environmental
variables, while twenty-one measurements of k~, k. and k5 were
made.
The data were analyzed with multiple linear regression
techniques. It was felt that the major factors affecting the
constants k, and k ~ (since they involve an oxidation step)
would be temperature, water content and organic matter (if the
reaction is microbially mediated). The major factors
affecting k , k and k would be temperature, pH and perhaps
water content since these reactions are thought to be
principally hydrolysis.
The results of the regressions are shown in Table 3.4. The
percent variance in k and k , explained by the variables
temperature, OM (organic matter) and WC (water content), is
low. Stepwise regression indicates that most of the variance
is explained by organic matter (29%) followed by water content
(11%). Temperature explains only about 1%. Regressions were
also performed for k and k subsituting pH for water content.
2
These regressions had slightly lower multiple R than those
shown in Table 3.4.
Regression of k , k and k on temperature, pH and water
J T J O
content produced more favorable results. Multiple R values
were 0.50, 0.65 and 0.71, respectively. Temperature explained
the largest percentage of the variance in the k_ regression
(46%) followed by pH (21%). Water content had little impact,
explaining only about 1% of the variance.
The linear regression of these rate constants on environmental
variables provided a means of evaluating pesticide parameter
values for different unsaturated zone scenarios. The
equations are purely empirical, however. Nicholls et al.
135
-------
TABLE 3.4 MULTIPLE LINEAR REGRESSION COEFFICIENTS FOR
ALDICARB TRANSFORMATION AND DEGRADATION RATES
Regression Coefficients
R' Inter- Temp pH OM WC
cept (°C) (%) (%)
Rate
k 0.38 0.30 0.01 ~ -0.04 4.36E-3
k 0.31 0.01 7.6E-4 -1.8E-4 -2.2E-4
2
k 0.50 -0.01 -5.6E-5 0.01 -9.99E-4
k4 0.65 -0.04 9.6E-4 0.01 -3.0E-5
k 0.71 -0.06 1.4E-3 0.01 3.9E-5
136
-------
(1982) used a semi-empirical relationship to correlate the
first-order rates with environmental variables. That
relationship was:
ln(k) = a + £ln(wc) + y(l/T) (3.3)
which uses only the variables (we) water content and (T)
temperature.
They found values of a = 26.5, 3= 0.407 and y= -8571 for the
aldicarb sulfone degradation rate (i.e., kj.) in a Compton
Beauchamp soil. The present analysis using the same
regression equation for eight different soils gave a= 17.8,
g = 0.04 and y = -6349 for aldicarb sulfone. However, the
2
regression coefficient of determination was very low (R =
0.24). This was not unexpected since pH was found to be an
important regression variable for rate k5.
Lacking a better data set which is more representative of
Florida soils, or a better (more theoretical) means of
analyzing the experimental rates, the above data and multiple
linear regression approach were used to calculate first-order
transformation and degradation rates for aldicarb and its
daughter products for each unsaturated zone scenario.
Table 3.5 shows characteristic information used to estimate
model pesticide parameters by soil and by horizon. Although
soil pH and % organic carbon are not specifically inputs to
PRZM, they were used to calculate pesticide degradation rates
and adsorption partition coefficients which were input
directly. These rates and coefficients are shown in Table 3.6.
137
-------
TABLE 3.5 CHARACTERISTIC DATA FOR FLORIDA SOILS USED TO
ESTIMATE PESTICIDE PARAMETERS BY HORIZON
Soil
Group
Entisols
Entisols
Entisols
Entisols
Entisols
Ultisols
Ultisols
Ultisols
Ultisols
Ultisols
Ultisols
Ultisols
Alfisols
Alfisols
Alfisols
Alfisols
Alfisols
Spodosols
Spodosols
Spodosols
Spodosols
Spodosols
Spodosols
Horizon
1
2
3
4
5
1
2
3
4
5
6
7
1
2
3
4
5
1
2
3
4
5
6
Depth
(cm)
0-10
10-20
20-40
40-100
100+
0-20
20-40
40-100
100-160
160-200
200-240
240+
0-20
20-40
40-80
80-120
120-240
0-10
10-40
40-60
60-100
100-200
200+
efc
(%)
5.6
5.4
4.2
3.5
3.0
8.2
7.0
5.5
8.0
20.0
22.5
5.0
10.0
6.5
9.5
20.0
25.0
14.0
9.0
6.0
12.0
21.0
6.0
PH
(su)
5.2
5.3
5.5
5.6
5.7
5.8
5.8
5.7
5.4
5.2
5.1
5.0
5.8
6.1
6.7
6.2
6.5
4.5
4.7
5.2
5.2
5.5
5.6
OC
(%)
0.8
0.7
0.4
0.15
0.06
0.8
0.65
0.25
0.10
0.05
0.05
0.05
1.2
0.4
0.1
0.2
0.2
2.1
1.2
0.35
0.90
0.90
0.40
PS 3
(g/cm )
1.35
1.4
1.45
1.45
1.5
1.5
1.52
1.54
1.56
1.64
1.66
1.68
1.4
1.5
1.6
1.65
1.68
1.3
1.45
1.50
1.50
1.50
1.50
138
-------
TABLE 3.6
TRANSFORMATION AND DEGRADATION RATES (k) AND ADSORPTION
PARTITION COEFFICIENT (K) USED IN MODELING ALDICARB FATE
AND TRANSPORT IN FLORIDA SOILS
Soil Group Horizon
Depth
Entisols
Ultisols
vo
Alfisols
Spodosols
1
2
3
4
5
1
2
3
4
5
6
7
1
2
3
4
5
0-10
10-20
20-40
40-100
100+
0-20
20-40
40-100
100-160
160-200
200-240
240+
0-20
20-40
40-80
80-120
120-240
0-10
10-40
40-60
60-100
100-200
200+
0.52
0.53
0.54
0.56
0.56
0.53
0.54
0.56
0.58
0.63
0.64
0.57
0.51
0.55
0.58
a. 62
0.65
0.47
0.51
0.55
0.54
0.58
0.55
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.02
0.02
0.03
0.03
0.03
0.03
0.02
0.02
0.03
0.03
0.03
0.03
0.02
0.03
0.04
0.04
0.04
0.04
0.04
0.04
0.04
0.04
0.04
0.03
0.03
0.03
0.04
0.04
0.05
0.03
0.04
0.02
0.03
0.03
0.03
0.02
0.03
0.04
0.04
0.04
0.04
0.04
0.04
0.04
0.04
0.04
0.03
0.03
0.03
0.04
0.04
0.05
0.05
0.05
0.03
0.03
0.04
0.04'
0.04
0.04
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.03
0.04
0.04
0.04
0.04
0.02
0.02
0.03
0.03
0.03
0.03
0.16
0.14
0.08
0.03
0.01
0.16
0.13
0.05
0.02
0.01
0.01
0.01
0.24
0.08
0.02
0.04
0.04
0.42
0.24
0.07
0.18
0.18
0.08
0.11
0.10
0.06
0.02
0.01
0.11
0.09
0.04
0.01
0.01
0.01
0.01
0.17
0.06
0.01
0.03
0.03
0.29
0.17
0.05
0.13
0.13
0.06
0.07
0.06
0.04
0.01
0.01
0.07
0.06
0.02
0.01
0.004
0.004
0.004
0.11
0.04
0.01
0.02
0.02
0.19
0.11
0.03
0.08
0.08
0.04
-------
Rates k, and k« were determined by entering soil temperature,
organic matter and water content data into the regression
equations shown in Table 3.4.
An average soil temperature of 25 deg. C was used for all
layers/ in all soils. Soil temperature data from Gainesville,
Lake Alfred and Bradenton were used to draw this conclusion.
The average soil temperature at Gainesville (4 and 8" depths)
for 1980, 1981 and 1982 ranged between 23 and 24 deg. C (73
and 75 deg. F). Average temperatures at the 4" depth at Lake
Alfred and Bradenton over the same three year period were 26
and 24 deg. C, respectively. Since ground-water temperatures
seem also to average at about this level, a 25 deg. C level
seemed appropriate for the soil.
Organic matter was assumed to be 1.7 x organic carbon and the
water content of a horizon was taken to be its field capacity
value.
Rates k_, k. and k- were determined by entering temperatures,
pH and water contents into the appropriate regression
equations in Table 3.4.
The adsorption partition coefficients K , K- and K3 were
determined by assuming Koc for aldicarb, aldicarb sulfoxide
and aldicarb sulfone to be 20, 14 and 9 cm /g respectively,
and multiplying by the appropriate organic carbon content from
Table 3.5.
Dispersion/diffusion coefficients were set to zero for all
horizons. Plant uptake efficiency factors were set to one in
horizons within the root zone and zero elsewhere.
An application of 5.61 kg/ha (5 Ib/ac) was made on February 15
of each simulation year. This is the current label
140
-------
application rate for Florida citrus. The application rate
corresponds to the optimal date for pest control (Jones, 1984,
personal communication). The pesticide was assumed to be
incorporated to a five centimeter depth.
3.3.2 Saturated Zone
The fate and transport of aldicarb and its metabolites in the
saturated zone differ significantly from the unsaturated zone.
Factors such as volitalization and plant uptake, that are
important in the unsaturated zone, do not influence transport
in the saturated zone. Adsorption of aldicarb depends
primarily on the amount of organic carbon available to bind
the compound to the solid phase, although there is some
conflicting evidence of the influence of clays on this
process. Because there is an insignificant amount of organic
carbon in the saturated zone, adsorption is considered
negligible.
The fate of aldicarb seems to be dominated by hydrolysis in
the saturated zone. There are four key factors that influence
the rate of degradation in the ground water in the Florida
citrus growing area.
1) temperature of ground water
2) pH of ground water
3) ionic strength
4) physical and chemical characteristics of the aquifer
material
The effects of the ionic strength on kinetic reactions are
complex and poorly understood, and therefore also difficult to
include in a quantitative analysis of the degradation rate.
141
-------
The following discussion qualitatively reviews the current
thinking on this factor and discusses its effect on the
hydrolysis rate..
The ionic strength (I) of a solution is a measure of the
influence of the free ions:
n 2
I = L Ci^i (3.4)
where I = ionic strength
C. = concentration of ion i
Z . = valance of ion i
Because the hydrolysis reaction includes ionic species, the
hydrolysis rate is depressed in the presence of ions competing
for the same reaction sites as the OH ion of water. When
hydrolysis rates in distilled water with no competing ions
were compared to rates in well water for aldicarb sulfoxide
and aldicarb sulfone (Lemley and Zhong, 1984), the results
showed the rates in well water were almost half the rates in
distilled water. The ionic strength of the well water is
assumed to be an important factor causing the rate decrease.
Although the actual ionic strength of the well water was not
stated, an earlier study (Lemley and Zhong, 1983) tested the
effect of competing ions directly by comparing the hydrolysis
rate in a NaCl solution with an ionic strength of 1.0
equivalant/liter to hydrolysis in distilled water for aldicarb
sulfone. The rate was 36% slower in the 1 Molar Had
solution.
The ionic strength of the ground water in Florida was
142
-------
approximated from water quality data (Phelps, 1978) in the
major citrus growing counties to assess the impact of
competing ions on the hydrolysis rate. The ionic strength
ranges from .01 to .04 equivalents per liter as compared to 1
equivalent/liter in the study cited above. Based on the rough
estimates of ionic strength of the ground water in Florida and
the experimental results of Lemley and Zhong (1983), it can be
assumed that the impact of ionic strength on the hydrolysis
rate in the citrus area is negligible. Future studies under
other environmental conditions may show that ionic strength is
a more important factor.
The effect of the aquifer material on surface catalyzed
hydrolysis rate is extremely difficult to quantify. The
variability of the porous material in the citrus areas of
Florida is very great. There is little information describing
the influence of specific surface types on hydrolysis rates.
Some studies have examined hydrolysis in soil and water
systems in the laboratory. Table 3.7, from Porter et al.
(1984), compares rates in the soil system to rates estimated
using distilled water studies. The rates in soil are 3-80
times faster than the rates calculated for distilled water.
Several mechanisms have been theorized to explain this effect.
One theory suggests that with the large surface area in a
porous medium there are more surface sites available for
reactions to take place. Porter et al. (1984) suggest that
the increase in the hydrolysis rate could be due to the
buffering capacity of the soil or catalysis by colloidal soil
constituents. Higher microbial activity in soils could also
account for the higher rates. However, studies done with
sterile and non-sterile samples in Long Island, New York show
no significant differences in the degradation rates (Lemley
and Zhong, 1984). Although site specific studies have been
done with soil and ground water, it is difficult to quantify
the results for general application to the estimation of
143
-------
TABLE 3.7 COMPARISON OF DEGRADATION RATES OF ALDICARB RESIDUES ESTIMATED BY
HYDROLYSIS WITH DEGRADATION RATES MEASURED IN SOIL AND WATER
DEGRADATION STUDIES (Source: Porter et al., 1984)
Sanple location
Long Island*
Wisconsin
Oerttal Sands
Resides
Presert
Suiroxlde
Suirore
Suiroxlce
Suirore
fiiproxlmate pH
or Sanple
an) 6
and 7
Suirore Only 7
Mirth Carol! ra
Coastal Plain
Suiroxlde
Suirore
and 5
Suirore Only 5
Florida Locatlora
Lull
Lake Hamilton
Alcana .
Ovletfj
Fort Pierce
Suiroxlce
Suirore
Suiroxlee
SuKore
Suiroxitfc
Suirore
SuUoxlde
Suirore
and 6
and 6
and 6
and 7
Suirore orly 7
Stud/
Tenperature
13
2S
25
25
25
25
25
25
25
25
25
Measured Rate
Cbf start,
(days'1)1
U.00076
0.0086
0.049
0.055
0.010
0.0047
0.014
0.015
0.014
0.25
0.024
(.00051)*
(0.0020)*
(0.0020)
(0.0042)
(O.U03)*
(0.0019)*
(0.0012)
(0.00079)
(0.0017)
(O.OJ9)
(0.0037)*
Measured Hair. Estimated Hair-Ure
Ufe Based on Distilled Estimated'
(ays)2 Nater H/crolysls (dtys) Measured
916
81
14
13
67
149
49
47
49
3
29 (
(351,
(
(
(
(
(
(
(
(
(
«,
12,
10,
62,
83,
»1,
40,
38,
1
a,
) 3000 - 5000
177)* 600 - 1000
16) 80-160
16) 80
74)* 9000 - 6000
655)* 6000
61) 600 - 1000
55) 600 - 1000
63) 600 - 1000
,5) 80 - 160
47)* 80
4
10
y
6
80
40
16
17
16
40
3
1. Hjmber InparerthEsls Is the starafird Aviation or data.
2. Hjmbers Inparerthesls are the 99K corflderce Units.
3. H>n-autoclaved sanples.
Prellmlrary estimate which "111 be rerired as more data becomes available.
-------
hydrolysis rates.
The influence of ionic strength and the aquifer material on
the rate of hydrolysis of aldicarb is complex and still poorly
understood. The studies that have been done show that, in
general, these two factors have the competing effects on
hydrolysis. The presence of competing ions depresses the
rate, whereas the presence of colloids increases the rate.
The effects of temperature and pH on the hydrolysis rate have
been extensively studied, such as those studies of Hanoen and
Speigel (1983) and Lemley and Zhong (1983), sited earlier.
Recently, advances have been made in understanding how these
factors influence the hydrolysis rate theoretically as well as
empirically. In general, as both the temperature and pH
increase, the hydrolysis rate increases. At high pH
concentrations, the number of OH ions available for
hydrolysis is greater, thereby increasing the likelihood of a
reaction. At high temperatures, molecular collisions are
increased, therefore increasing the probability of a reaction.
As in the unsaturated zone, a method for estimating the
degradation rate based on the environmental factors in the
Floridan ground water is necessary to estimate model
parameters. The same approach of a regression analysis could
be used on the existing temperature and pH data, to obtain
values of hydrolysis rate constants. The analysis of Lemley
and Zhong (1983) suggests another approach. They use the
Arrhenius equation to describe a theoretically based
relationship between the second order reaction rate and
temperature. In this study, the Arrhenius relationship was
used to derive the following equation for the first-order
reaction rate based on pH and temperature (see Appendix C for
explanation and derivation):
145
-------
~EA X
log k. = + log A + pH - 14
* 2.303R T
where k = first order reaction rate (/time)
E a = activation energy (energy /mole)
o
R = universal gas constant (energy /mole K)
A = pre-exponential factor (mole/liter-time)
T = temperature ( K)
The activation energy (E ) and the pre-exponential factor (A)
A
were determined for aldicarb, sulfoxide and sulfone using the
data shown in Table 3.8. This data was taken from studies of
Hansen and Speigel (1983), Lemley and Zhong (1983), Lemley and
Zhong (1984), Porter et al . (1984) and Lemley (1984, personal
communication) . Where the second -order reaction rates were
not reported directly (Hansen and Speigel, 1983 and Porter et
al., 1984), they were calculated from the data provided in the
studies (Table 3.2 and Table 3.9). Table 3.10 shows the
values of the activation energy and ^pre-exponential factor
resulting from the data analysis. Although fairly significant
errors can result in these types of experimental analyses (see
Appendix C), there is good experimental agreement from these
2
five different sources of data as indicated by the high R
values (Table 3.10).
Temperature and pH Characteristics in Florida
The temperature in the Floridan and surficial aquifers is
fairly constant, varying from 22 deg. C - 27 deg. C (Stewart,
1966 and Grain et al . , 1975) and averaging 25 deg. C (Fayard
et al., 1983). The pH is more variable. In the surficial
aquifers, pH varies from 5.0 to 8.0. In the Floridan Aquifer
the pH is rarely below 6.5 and sometimes is as high as 9.0.
146
-------
TABLE 3.8 HYDROLYSIS RATE DATA USED TO CALCULATE ACTIVATION
ENERGY PARAMETERS
Source
Aldicarb
Aldicarb
Aldicarb
Aldicarb
Aldicarb
Sulfoxide
Sulfoxide
Sulfoxide
Sulfoxide
Sulfoxide
Sulfoxide
Sulfoxide
Sulfoxide
Sulfoxide
Sulfoxide
Sulfone
Sulfone
Sulfone
Sulfone
Sulfone
Sulfone
Sulfone
Sulfone
Sulfone
Sulfone
Sulfone
Sulfone
Sulfone
Sulfone
Sulfone
Sulfone
A
E
A
B
C
A
E
A
B
C
D
A
D
D
D
A
B
C
B
A
B
C
D
A
B
C
D
D
D
B
B
Second Order Rate
Constant, kr (liters/
mole-min)
.1637*
.435
.8946*
.94
1.15
2.38*
4.19
13.77*
10.5
11.4
4.38*
23.5
21.9*
167*
1110*
4.35*
13.6
11.3
19.1
29.41*
38.6
33.0
10.7*
94.3*
91.0
90.5
53.5*
482
2830*
145
192
Temperature
5
5
15
15
15
5
5
15
15
15
15
25
25
35
45
5
5
5
10
15
15
15
15
25
25
25
25
35
45
30
35
*Second order rate constant (kr) calculated by linear
regression from data presented in given source.
A = Hansen and Spiegel, 1983
B = Lemley and Zhong, 1983
C = Lemley and Zhong, 1984
D = Porter et al., 1984
E = Lemley, personal communication, 1984
147
-------
TABLE 3.9 DISTILLED WATER HYDROLYSIS RATES OF ALDICARB SULFOXIDE AND ALDICARB
SULFONE (Source: Porter et al., 1984)
*»
00
Tenperature
( C) 4
AldLcarb Sulfoxide
45 35 ( 34- 36)*
35 74 ( 70- 78)
25 219 (207-232)*
15 NA
Alcftcarb Sulfore
45 109 ( 99-121)
35 396 (369-426)*
25 NA
15 NA
Half-Life (days) at IndicatedpH
7 8
84 ( 74- 97)
635 (527-798)*
NA
NA
106 ( 95-12U)*
584 (384-1218)*
NA
NA
31 ( 28- 34)*
130 (122-139)*
NA
NA
15 ( 14- 16)*
64 ( 62 - 67)*
NA
NA
4.4 (4.1-4.5)
2U ( 18- 22)
16i (151-182)*
NA
1.7 (1.6-1.8)
8.2 (7.9-8.4)
82 ( 80- 83)*
420 (359-506)*
IM
2.9 (2.7-3.0)
25 ( 24- 27)
109 ( 99-122)*
N4
1.0 (0.9-1.1)
9.6 (9.4-y.B)
47 ( 45- 49)
N4
R
2.2
11 (
IM
R
0.9
4.5
(2.1-2.3)
10- 12)
(0.8-1.0)
(4.3-4.6)
Numbers inparertheses represert the 93K Corflderce Interval
N4 - rot measured
NA - eroug^ data has not yet been collected to estimate the hydrolysis rate
R - half-life Is less than one day
- preliminary value, will be refined as more data become available
-------
TABLE 3.10 ESTIMATED VALUES OF ACTIVATION ENERGIES AND
PRE-EXPONENTIAL FACTORS FOR ALDICARB, ITS
SULFOXIDE AND SULFONE
Pesticide
R
Activation
Energy, E,
(Kcal/mole)
Pre-exponential
Factor, A
(1/min)
Aldicarb 0.81
Aldicarb 0.92
Sulfoxide
Aldicarb 0.90
Sulfone
20.8
24.0
21.8
6.31 x 10
1.58 x 10
1.0 x 10
15
19
18
149
-------
The overall average for both aquifers based on the 1982 water
data (Fayard et al., 1983) is 7.2. The Floridan Aquifer alone
is somewhat higher, the surficial aquifer is lower. Based on
the average pH and temperature and the activation energy and
pre-exponential factor estimated for aldicarb and its
metabolites, the first order reaction rate was calculated
according to equation 3.5 (see Table 3.11).
In addition to the four factors described above, Smelt et al.
(1983) have recently reported much higher disappearance rates
for aldicarb sulfoxide and aldicarb sulfone when incubated
under anaerobic versus aerobic conditions. They reported that
under anaerobic conditions, disappearance of these two
compounds was from 8 to 100 times faster than in the same soil
under aerobic conditions. More recent work (Smelt, 1984,
personal communication and Bromilow, 1984, personal
communication) has shown that in anaerobic soils (redox
potentials from -60 mV to +200 mV) with relatively high
O_L
concentrations of ferrous iron (Fe ), aldicarb is transformed
into nitriles and aldehydes. Half lives for aldicarb under
these conditions were 0.5 days or less. Half lives for the
sulfoxide and sulfone in anaerobic, iron rich subsoils at 10
deg. C ranged from 2 to 131 days. The current theory is that
the presence of Fe ^+ in solution catalyzes these reactions.
Very little specific data is available on redox potential or
f\ i
dissolved ferrous iron (Fe ) in Florida. According to Hem
(1970), the most common form of iron in the ground water is
ferrous iron (Fe *+) . He states that ground water with a pH
beteen 6 and 8 can carry as much as 50 mg/1 of ferrous iron at
equilibrium and that the occurrence of 1.0 - 10 mg/1 of iron
in ground water is common. Dissolved iron in the ground water
in Florida varies from as little as 0.01 mg/1 to as high as 20
mg/1 (Fayard et al., 1983; Mycyk et al., 1983; U.S. Geological
Survey, 1982). In general the surficial aquifer has higher
150
-------
TABLE 3.11 FIRST-ORDER DECAY RATES IN FLORIDA'S GROUND WATER
BASED ON ACTIVATION ENERGY ANALYSIS
Chemical pH °C k. (days' ) Half Life (days)
Aldicarb 7,. 2 25 7.91 x 10 876
Aldicarb
sulfoxide 7.2 25 9.64 x 10 71.9
Aldicarb 7.2 25 2.34 x 10 29.7
sulfone
151
-------
iron concentrations because the aquifer materials contain more
iron-bearing minerals than the predominantly limestone
Floridan Aquifer (Irwin, 1984, personal communication).
If the transformation of aldicarb described by Smelt et al.
(1983) is indeed catalyzed by ferrous iron, the process would
be important in the ground water in Florida. At this time,
the process is so poorly understood it can not be included in
the quantitative analysis of the pesticide parameters for
modeling.
152
-------
SECTION 4
MODEL APPLICATION AND RESULTS
In this section the application of the PRZM and the CFEST
models to the unsaturated and saturated zones, respectively,
is discussed. This discussion is broken down readily into two
separate parts because the modeling of the unsaturated and
saturated zones was essentially accomplished independently.
The unsaturated zone model (PRZM) was applied under each of
the scenarios given in Section 2.1.6. Daily pesticide
loadings at the bottom of the simulated soil core were written
to an output file. These loadings were summarized and are
presented in this section. Concurrently, steady-flow/unsteady
contaminant transport simulations were made using the CFEST
model for the saturated zone. These simulations used a unit
loading of pesticide (1 kg/ha). Because of this, results are
expressed as relative concentrations in wells, that is, the
concentration simulated divided by the initial concentration
in the aquifer resulting from the input of the unit load. In
the last part of this section, results of the unsaturated and
saturated zone modeling are combined to predict actual
drinking well-water concentrations.
153
-------
4.1 UNSATURATED ZONE MODELING
Before discussing the results of the unsaturated zone
modeling, the process of applying the model, including
assumptions made, will be covered. Part of this process was
to attempt to verify the PRZM model using residue data
available for Florida citrus soils.
4.1.1 PRZM Verification
Two data sets were available for verification purposes. One
has been reported in Jones et al. (1983) and Hornsby et al.
(1983). These data were taken from two sites in central
Florida; one a typical "ridge" citrus site (Lake Hamilton) and
the other a typical "flatwoods" site (Oviedo). Another data
set was also produced as a part of this study by the
University of Florida (Foran, 1984; personal communication).
4.1.1.1 1983 Residue Data
The first data set was used in the following way. The
currently available version of PRZM (PRZM-II, Carsel et al.,
1984) was run using data inputs which were exactly (or as
close as possible, considering input changes) those used by
Jones et al. (1983) in their original PRZM model calibration
exercise. This was done because Jones et al. used an earlier
version of PRZM in their work. Model outputs were compared to
outputs produced by Jones and measured field data at the
Oviedo and Lake Hamilton sites. This done, another simulation
was performed. In this case PRZM-II, as modified for use in
this study, with parent/daughter relationships and lateral
drainage was used. Inputs used were those of the generic soil
154
-------
orders as generated by the analysis described in Section 2.
At the Lake Hamilton site, the dominant soil order is the
entisols. Therefore, input data to be used for simulating
entisols in the unsaturated zone scenarios were input.
Likewise, the input sequence to be used for spodosols was used
at the Oviedo site. Once again, results are compared to the
observed data and model results from Hornsby et al. (1983),
and Jones et al. (1983). Results for the Lake Hamilton site
are shown in Figures 4.1 through 4.3.
Figure 4.1 shows a comparison of the amount of aldicarb (total
toxic residue) remaining in the top 300 cm of the profile at
Lake Hamilton up to 200 days after application. Both the
model calibration of Jones and the model simulation with
PRZM-II agree well with the observed data. PRZM-II shows
consistently slightly less residues than the earlier version
of PRZM used by Jones. The generic input parameter set for
entisols shows more rapid decay of the TTR (total toxic
residue) than observed in the field or predicted by the other
modeling exercises. Figure 4.2 shows a comparison of the
depth of the simulated peak from the models and deepest
penetration observed in the data. The PRZM-II code with
Jones' inputs and the PRZM-II code with generic inputs under-
simulate both the data and the original PRZM up until
approximately 40 days, at which time they follow the deepest
penetration observed in the data. The reason for the
discrepancy between Jones' PRZM simulation and the simulations
in this study soon after pesticide application appears to be a
percolation event which occurred shortly thereafter. This
event caused movement of the chemical in Jones' simulation but
not in ours. In both cases, Lake Alfred Experiment Station
1983 precipitation data was used. The discrepancy may not be
in the precipitation record but in initial conditions. Our
simulations began January 1 so initial conditions were set by
the model. These particulars of Jones' simulation are not
155
-------
100
80-
UJ
ec
o
o
o
2
<
Ul
cc
Ul
o
m
la
a.
50 100
DAYS AFTER APPLICATION
Figure 4.1
Predicted and observed Aldicarb TTR in the
upper 300 cm of the soil at the Lake Hamilton
site, 1984. (After Jones et al., 1983).
156
-------
300-
0
120
DAYS AFTER APPLICATION
LEGEND
PRZM
- MODIFIED PRZM - II
PRZM - II
DEEPEST PENETRATION MEASURED
Figure 4.2 Predicted movement of Aldicarb residues at
the Lake Hamilton location (After Jones
et al., 1983).
157
-------
known. Figure 4.3 shows a comparison of the depths over which
aldicarb TTR >5 g/g (parts per billion) was observed in the
field versus our two model simulations. We note here that
aldicarb was applied on February 16, 1983. The four shaded
areas at the left of each date are samples taken in four
quadrants of the field. At the March 4 date, both model
simulations agree well with the observed data. Note that with
the PRZM-II model, as modified for this study, the TTR is
broken down into aldicarb, aldicarb sulfoxide and aldicarb
sulfone. At the April 6 date, both models and the data still
agree well. Notice that no aldicarb is simulated as being
present at this date. On May 3, PRZM-II still agrees well but
the modified PRZM-II agrees less well. This may be due either
to too rapid movement or too rapid decay. The same argument
holds for the June 15 sampling date.
Figures 4.4 and 4.5 show the results of model applications at
the Oviedo (flatwoods) site. Figure 4.4 compares the TTR
remaining in the top 150 cm of the soil profile. The measured
decay in this case does not appear to be first-order. Jones'
simulation and the modified PRZM-II simulation with generic
spodosol inputs give very close to the same results. The
PRZM-II model with Jones' inputs oversimulate the quantity
remaining in the profile.
Figure 4.5 shows the comparison of observed TTR in the field
versus PRZM-II with Jones' inputs and modified PRZM-II with
generic spodosol inputs. Agreement between the position of
residues in the field and both model simulations are good.
The modified PRZM-II model with generic spodosol inputs
appears to be slightly better. This is due to better
agreement of the decay rates in this model with those observed
in the field. Notice that only sulfone is left in the
simulated profile on the June 14 sampling date.
158
-------
3/4/83
4/6/83
5/3/83
6/15/83
100-
o.
UJ
Q
200-
300 J
OBSERVED
PRZM-II
MODIFIED PRZM-II
NO SIMULATED RESIDUE > 5ng/g
Figure 4.3 Location of Aldicarb TTR in the soil profile
> 5 ppb at the Lake Hamilton site.
159
-------
100
UJ
cr
o
u
o
z
5
UJ
oc
UJ
o
CO
UJ
Q.
OBSERVED
20-
100
DAYS AFTER APPLICATION
Figure 4.4
Predicted and observed Aldicarb TTR in the upper
150 cm of the soil at the Oviedo site. (After
Jones et al., 1983).
160
-------
3/3/83
4/5/83
6/2/83
6/14/83
150-
OBSERVED
f'.---.1 PRZM- II
MODIFIED PRZM-II
Figure 4.5 Location of Aldicarb TTR in the soil profile
> 5 ppb at Oviedo.
161
-------
These results indicate a fair agreement between the PRZM model
and the observed data from the 1983 sampling of aldicarb
residues at Lake Hamilton and a good agreement at the Oviedo
site (Hornsby et al., 1983; Jones et al., 1983). Simulations
with PRZM-II modified to simulate parent/daughter
relationships were also fair to good. The generic entisols
input data set produced higher than observed degradation at
the Lake Hamilton site, while the generic spodosol input data
produced very good agreement with observed residue data at the
Oviedo site. The agreement in both cases is suprisingly good
considering that site specific soils and pesticide parameters
were not used at either site. This somewhat justifies the
assumption that the soil characterization and pesticide
degradation and transformation rate data analysis can be used
to represent aldicarb fate and transport processes in the
citrus growing region of Florida.
4.1.1.2 1984 Residue Data
Other data were also available for 1984 from the Oviedo and
Davenport sites. The Davenport site is located near the
intersection of Interstate Highways 4 and 75 and is, again,
typical of a "ridge" citrus site. The residue samples were
collected by personnel at the University of Florida under
contract to the U.S. Environmental Protection Agency. The
samples were analyzed by the U.S. Environmental Protection
Agency in Beltsville, Maryland. The data, as received, are
shown in Appendix D. An analysis of the Davenport data is
shown in Table 4.1.
This data has two interesting features. First, the peak
concentration on 5/9/84 is deeper than the peak sampled on
6/13/84. Although this peak is shown on 5/9/84 in the 365.8
to 426.7 cm range, and on 6/13/84 in the 304.8 to 365.8 cm
range, the peaks may actually be quite close. The more
162
-------
TABLE 4.1 ALDICARB RESIDUES (TTR) IN THE SOIL AT DAVENPORT,
FLORIDA, 1984
Mean TTR (ppb)*
Depth (cm) 5/9/84 6/13/84
0-30.5 47.33 10.50
30.5-61.0 0.00 5.00
61.0-121.9 6.00 22.75
121.9-182.9 12.67 48.50
182.9-243.8 42.00 88.50
243.8-304.8 78.67 91.00
304.8-365.8 70.67 108.00
365.8-426.7 106.00 92.50
426.7-487.7 47.33 53.00
Mean TTR in profile 48.37 64.00
*Values used to compute the means were only those determined
to be within the treated band.
163
-------
troubling aspect is that the peaks are almost the same
magnitude although more than a month had elapsed between the
two sampling dates. In fact, the data shows greater than 25%
more aldicarb TTR in the profile on 6/13/84 than on 5/9/84.
These differences are not easily explained.
The data from the 1984 Oviedo sampling are very erratic.
There are many non-detectable entries in the data so that the
mean of the data from various soil cores may be somewhat
misleading. However, the mean values are presented in Table
4.2 for comparison with the Davenport data.
The TTR in the soil at the 5/22/84 sampling data at Oviedo ,
compares favorably with that in the soil and the Davenport
site on 5/9/84. The average TTR in the soil at Oviedo on the
6/3/84 sampling date of 18.4 ppb also compares favorably with
the mean TTR concentration at Davenport in the top 152 cm of
soil on 6/13/84 of 21.7 ppb. However, considering the erratic
nature of the 1984 data from Oviedo, this favorable comparison
may be fortuitous. Indeed, the pesticide was applied a full
month earlier at Davenport (2/14/84) than at Oviedo (3/14/84).
Therefore, to have roughly the same quantity of pesticide in
the soil on the sampling dates, the degradation rates have to
be higher at Oviedo.
An analysis of the degradation rates are shown in Figure 4.6.
The dashed lines on the plot correspond to half-lives,
reported by Hornsby et al. (1983), of 3 to 46 days at Oviedo
and Lake Hamilton, Florida. Assuming that little pesticide
leached past the 150 cm depth at Oviedo in 1984, the
approximate percent TTR remaining in the profile can be
determined from the data in Table 4.2 for the two sampling
dates. (A bulk density of 1.5 g/cm was used in this
calculation). The logarithms of these percentages are plotted
on Figure 4.6 along with the best fit line. This gives a
164
-------
TABLE 4.2 ALDICARB RESIDUES (TTR) IN THE SOIL AT OVIEDO,
FLORIDA, 1984
Mean TTR (ppb)
Depth (cm) 5/22/84 6/20/84
0-30-5 173.5 20.4
30.5-61.0 29.8 2.7
61.0-91.4 8.0 42.2
91.4-121.9 23.0 21.3
121.9-152.4 0.0 5.3
Mean TTR in profile 46.9 18.4
165
-------
10 20 30 40 SO 60 70 80 90 100 110 120
DAYS AFTER APPLICATION
O OBSERVED AT DAVENPORT. 1884
X OBSERVED AT OVIEDO. 1884
Figure 4.6 Comparison of Aldicarb TTR degradation rates for
1984 Oviedo and Davenport data.
166
-------
half-life of approximately 28 days for aldicarb TTR.
Similarly, the data from Davenport are plotted. The half-life
of these residues is 210 days which is five times greater than
the fastest rate from the Lake Hamilton/Oviedo 1983 studies
and roughly 10 times greater than rates determined from the
1984 Oviedo data. If, in fact, these determined rates for the
soils at Davenport are valid, they are at best three times
slower than reported values of Hornsby et al. (1983), or Rao
et al. (1984), for several Georgia coastal plains soils.
The modified PRZM-II model was applied to these two data sets
in a verification exercise. For the Oviedo site,
precipitation data from Sanford experiment station and pan
evaporation from Lisbon were used as model meteorological
inputs. The generic spodosol input parameter sequence was
used for soils, crop and pesticide parameter inputs. Results
are shown in Figure 4.7. The mean values of the measured data
are shown as vertical bars. The dashed lines indicate a 1
standard deviation range calculated from the data. The model
simulations follow the mean values fairly closely and are
certainly within a 1 standard deviation distance around the
means at each depth. The model under-simulated the total mass
in the profile by about 40% on the May 22 date and 45% on June
20. Therefore, model degradation rates seem somewhat higher
than those observed.
The modified PRZM-II model was also applied to the Davenport
site. February through June, 1984, Lake Alfred experiment
station precipitation and pan evaporation data were used as
meteorological inputs and the generic entisol data set was
used for soils, crop and pesticide parameter input. The
results, shown in Figure 4.8 indicate a radical difference in
simulated and observed pesticide concentrations. This is not
unexpected since the pesticide degradation rates in the
generic entisol parameter data set are more than an
167
-------
CO
CONCETRATION (ppb)
200 300 400
MAY 22. 1984
TI11 J MEASURED
0O SIMULATED
500
600 0
CONCENTRATION (ppb)
100 200
JUNE 20. 1964
,.!,. 1
_ J
Figure 4.7 Comparison of measured and simulated Aldicarb TTR concentrations
for two 1984 sampling dates at Oviedo, Florida.
-------
CONCENTRATION (ppb)
100 200 300
CONCENTRATION (ppb)
60
100
1-lJ
£
H"^
1 1
[ | MAY 9. 1984
J"
160 ii ;
hi-'i--
Ai 1 i
200?
^ 250
X
300
VO
350
400 6 '
450
600
-,-fr-'
1 ;
,1
IQT
MEASURED
SIMULATED
0
200
i
300
JUNE 13. 1984
Figure 4.8
Comparison of observed and simulated Aldicarb TTR concentrations
for two 1984 sampling dates at Davenport, Florida.
-------
order-of-magnitude greater than the observed TTR degradation
rates. Another perplexing aspect of this data has to do with
the depth of penetration of the peak. The data shows the peak
on 9 May 1984, at about the 60-420 cm depth in the soil. Even
when the PRZM-II degradation rates were adjusted to match
those observed in the field, the simulated position of the
peak on 9 May was approximately 30 cm. On the 13 June date,
the simulated peak was between 100 and 160 cm. It is
possible, however, that much more rainfall fell at this site
than is indicated by the Lake Alfred data.
4.1.1.3 Verification Summary
Verification of the PRZM-II model was performed using residue
data from four sampling dates at three different locations in
Florida. Generic soils, crop and pesticide parameter input
data were used as opposed to specific data for the individual
sites. Thus, the intent was to verify not only the model, but
also the generalized data sets to be used for simulation of
aldicarb fate and transport in the various unsaturated zone
scenarios.
Verification results were best at the Oviedo site were
spodosolic soils are present. The model simulations matched
observed concentration profiles well for both 1983 and 1984.
At the Lake Hamilton and Davenport sites, when "ridge" type
soils occur, simulation results were not as good. At these
locations, too rapid degradation of the pesticide appears to
be simulated. Results at the Lake Hamilton site were better,
however, than those at the Davenport site. Some aspects of
the 1984 Davenport data; the extremely low decay rates and the
deep penetration of the peak concentration were not able to be
explained.
170
-------
4.1.2 Model Application and Assumptions
The verified PRZM-II model was applied for each of the
twenty-four scenarios listed in Section 2. Simulations with
overhead irrigation were made first. The mean annual
evapotranspiration was calibrated to approximately 122 cm (48
in.) by adjusting the available water level at which
irrigations were triggered. For each soil order, this came
out to be roughly 45% (i.e., 45% of the water between field
capacity and wilting point). Other than this, no adjustments
were made to any parameter in the generic input sequences for
the no-irrigation or overhead irrigation practices.
4.1.2.1 Simulation of Low Volume Spray Irrigation
For the simulation of the low volume spray irrigation practice
application of the model was not as straightforward. The
problem lies in the fact that PRZM is a one-dimensional model
and the simulation of the low volume spray (or any irrigation
practice in which the irrigation water is not applied
approximately uniformly over the simulated area) is a spatial
problem. A schematic of a unit "ridge" citrus block is shown
in Figure 4.9. This block contains one tree. Aldicarb is
applied centered beneath the drip line on either side of the
canopy. Depending on the size of the tree and location of the
aldicarb band, the wetted area of the spray jet may or may not
intersect the treated band. If it does, there are two unique
areas with respect to aldicarb leaching;
1) the area in which only rainfall falls upon the treated
band, and
2) the area which receives rainfall plus irrigation
water.
171
-------
5.2 m
Figure 4.9 Schematic of a unit block of citrus (one
tree, not to scale).
172
-------
Notice that if spray jets are located down the line of trees
so that their wetted patterns overlap, then a wetted band as
opposed to a circle develops.
Considering these two areas together, the leached from a unit
citrus block would be:
L = W-L L! + W2 L2 (4.1)
where L, = load from the non-irrigated portion of the
treated band
band
L2 = load from the irrigated portion of the
W, , W- = fractions of the unit block made up of
areas 1 and 2
It was felt that PRZM-II could be applied to each of these
areas and the results weighted to obtain the proper load.
In order to properly simulate pesticide movement in each area,
the hydrology of each area must be properly simulated. This
means that evapotranspiration from areas which are irrigated
and not irrigated should be adjusted so that the total
evapotranspiration will equal that required on an annual basis
( 122 cm). Therefore:
ET = Vli ET-L + W2 ET2 (4.2)
The problem here is that the weights, W, and W_ are unknown.
A best guess, however, is that they correspond to the root
distribution in each area. This too is unknown. For the sake
of simplicity, the assumption was made that all the roots were
in the wetted zone (i.e., W_ = 1.0). This means that all the
173
-------
evapotranspiration demand for the entire tree must be
extracted from the wetted band. Under uniform irrigation
2
water application, the demand is 122 cm over the 7.6 x 7.6 m
unit block. Therefore, if all the demand must come from the
wetted areas (10 m2) the demand must be increased by:
7.6 x 7. 6 m2 (4.3)
10 m2
or a factor of 5.8 in the ridge areas. In the flatwoods areas
tree spacing of 4.6 m x 7.6 m was assumed, therefore, the
demand must be increased by 3.51. To accomplish this in the
simulations, the input pan evaporation factor (PFAC) was
multiplied by the above values as needed.
4.1.2.2 Simulation of Drainage in Bedded Citrus
For spodo.sols and alfisols, the lateral drainage from bedded
citrus was simulated. A discussion of the algorithms used to
accomplish this is discussed in Appendix A. Parameterization
of the algorithms was discussed in Section 2.
The water draining laterally from the soil may carry
pesticide, just as the percolating water does. To simulate
the removal of pesticide from the unit citrus block would
require that a sink term be added to the one-dimensional
advection/dispersion equation in PRZM for each soil
compartment, much like the runoff term for the surface layer.
Again, this problem is at least two-dimensional. As the
pesticide moves laterally out of the unit block, some will be
adsorbed by soil. This adsorbed chemical may be subject to
later desorption and leaching by rainfall.
174
-------
Even pesticide which enters adjacent canals may be subject to
leaching. As a "worst case" it was assumed that all of the
pesticide, whether or not advected by lateral flow, still has
the potential to move to the ground water, until it degrades.
For this reason, lateral flow was allowed to leave the unit
block without carrying pesticide. Using this assumption,
lateral drainage can still be simulated for the purpose of
simulating a proper water balance while providing a "worst
case" scenario for pesticide leaching.
4.1.2.3 Pesticide Parameter Sensitivity
No sensitivity analyses were performed on transformation or
degradation rates or adsorption partition coefficients for the
unsaturated zone. For the "flatwoods" type soils (spodosols
and alfisols) the verification results using the generic input
data sets were good and sensitivity analysis was deemed to be
unnecessary. For the entisols and ultisols, it appears that
use of slower degradation rates might be in order. This will
be discussed further later on in this section.
4.1.3 Hydrologic/Hydraulic Results
Table 4.3 shows the water balance results for the twenty-four
unsaturated zone scenarios.
Evapotranspiration (from the canopy and the soil profile)
averaged 81.1 cm (31.9 in) over the non-irrigation scenarios
for all soil orders. Evapotranspiration averaged 121.2 cm
(47.7 in) over the overhead irrigation scenarios for all soil
orders. As mentioned earlier, these depths were "calibrated"
to equal about 122 cm (48 in), which is a typical value for
irrigated Florida citrus. The irrigation by low volume spray
175
-------
TABLE 4.3 SOME WATER BALANCE COMPONENTS FOR TWENTY-FOUR
UNSATURATED ZONE SCENARIOS (in centimeters).
Values are annual means over 14 years
Scenario
EN2.7
EN9.0
E02.7
E09.0
EL2.7
EL9.0
UNI. 8
UN9.0
UO1.8
U09.0
UL1.8
UL9.0
SN1.2L
SO1.2L
SL1.2L
SN1.2H
S01.2H
SL1.2H
AN1 . 2L
A01.2L
AL1 . 2L
AN1 . 2H
A01 . 2H
AL1 . 2H
Evapotrans-
piration
82.5
82.8
124.4
124.5
518.3
525.8
86.7
86.6
125.6
125.6
567.2
570.9
80.0
124.6
355.5
79.2
113.2
333.6
75.9
121.6
343.2
75.0
110.3
322.0
Irrigation
__
60.3
59.7
436.7
444.2
58.6
58.7
480.6
486.1
61.7
259.6
45.7
227.0
61.8
240.6
48.5
212.7
Recharge
38.7
38.4
57.1
56.4
39.5
39.6
34.5
35.1
54.2
54.7
34.5
36.4
33.0
47.1
20.9
44.7
52.2
27.1
30.4
29.7
12.3
26.0
30.1
15.3
Lateral
Drainage
8.2
11.2
4.4
15.8
20.0
6.1
14.9
31.7
6.3
38.7
47.8
15.1
176
-------
yielded evapotranspiration depths in the wetted area which
ranged from 322 cm (127 in) to 570.9 cm (224 in). On a total
area basis, assuming no evapotranspiration from (i.e., no
plant roots in) the non-irrigated areas, this depth is 93 cm
(36.8 in) to 97 cm (38.2 in). It appears, then, that
evapotranspiration depths are on the order of 20 to 23% too
low for the low volume spray scenarios on a total area basis.
A possible explanation for this is that the method used to
simulate transpiration with PRZM-II created water limiting
situations in the soil profile, even with the addition of
irrigation water. That is, there is not enough available
water in the root zone at field capacity to satisify this high
demand. Therefore, not all the potential evapotranspiration
was used each day. In reality, because of the abundance of
rainfall, roots that extract water from the soil by the plant
probably occur outside the wetted area. If the assumption is
made, however, that the plant roots take water from the
non-irrigated area as well at rates taken from the
non-irrigated model simulations, then the weighted
evapotranspiration depths are too large, ranging from 150 cm
(59 in) to 169 cm (66 in).
Thus, it appears, to accurately simulate the hydrology and
hence the movement of pesticide under the low volume spray
scenario, three areas would have to be considered rather than
the two areas originally considered in Section 4.1.2.1;
1) the irrigated area including plant roots and
accompanying water and pesticide uptake
2) the non-irrigated area having plant roots and
accompanying water and pesticide uptake and
3) a non-irrigated area having no plant roots and no
transpirational or pesticide uptake.
177
-------
The weighted average evapotranspiration, ETw, for a unit
citrus block would be:
ETW = WJL ETX + W2 ET2 + W3 ET3 (4.4)
where the subscripts 1, 2, and 3 refer to the areas above and
W- are the weights for each area. A rational method of
assigning these weights would be to use the root distribution
in the three areas. However, these values are unknown, and
there is no simple way of determining these weights.
Because of this difficulty, it was felt that interpretation of
results based on one low volume spray simulation would not be
meaningful, therefore, they were not given further
consideration.
Irrigation using overhead methods resulted in an average
annual water application of 56.9 cm (22.4 in) over all soil
orders. This application rate is higher than the average
values computed from the data of Duerr and Trommer (1982) for
overhead systems in the state, of roughly 38 cm (15 in).
However, they do record data yielding annual average
application rates as high as 25.4 in. per acre (1970-1980).
Therefore the simulated depths do not seem unreasonable.
Average annual recharge rates under no irrigation ranged from
a low of 12.3 cm/yr (4.8 in/yr) for low rainfall alfisols to
44.7 cm/yr (17.6 in/yr) for high rainfall spodosols. The mean
for no-irrigation over all soil orders was 35.1 cm/yr (13.8
in/yr). These rates are in good agreement with those
mentioned in Section 2 derived from the literature. Recharge
rates increased when irrigation water was applied.
Lateral drainage, of course, only occurred in alfisols and
178
-------
spodosols. Spodosols exhibited the lowest lateral drainage
depths with average annual values being 7.9 cm (3.1 in) for
low rainfall, and 14.0 cm (5.5 in) for high rainfall
spodosols. This represents approximately 6.5% of the average
annual precipitation over the 14 year simulation period for
the low rainfall spodosols and 10% for the high rainfall
spodosols. The alfisols had still higher lateral drainage
losses, being 7.6 cm (6.9 in) for the low rainfall scenarios
and 3.9 cm (13.3 in) for the high rainfall scenarios. This
represents 14% and 24% of the low and high precipitation
depths, respectively.
Examination of the lateral drainage and recharge rates reveals
that a large percentage of the total water that exits the soil
profile due to lateral drainage and recharge exits laterally.
In the spodosols, the percentages range from 17 to 28%, while
in the alfisols the percentages range from 33 to 61%. This
indicates that the contamination of surface water is also
quite likely in areas where these soils are prevalent.
4.1.4 Pesticide Fate and Transport Results
Figure 4.10 graphically depicts the overall fate of aldicarb
TTR after application to the soil. The values presented are
annual average percentages of applied chemical over the 14
year period. Three processes are represented; uptake by
plants, degradation to non-toxic residues and leaching to the
saturated zone. By far, the most eminent fate process is
decay in the root zone. Over all scenarios, this accounted
for an average of 70.3% of the applied chemical. Plant
uptake, on the average, accounted for 29.2% of the applied
chemical. The remainder, or 0.5% of the applied chemical,
leached to the saturated zone on an annual basis over all
179
-------
PERCENT OF APPLICATION
0 10 20 SO 40 60 60 70 80 8.0 100
NO IRRIGATION
OVERHEAD
6 FOOT CORE
NO IRRIGATION
OVERHEAD
ULTISOLS
30 FOOT CORE
NO IRRIGATION
OVERHEAD
9 FOOT CORE
NO IRRIGATION
OVERHEAD
ENTISOLS
30 FOOT CORE
NO IRRIGATION
OVERHEAD
LOW RAINFALL
ALFISOLS
NO IRRIGATION
OVERHEAD
HIGH RAINFALL
NO IRRIGATION
OVERHEAD
LOW RAINFALL
SPODOSOLS
NO IRRIGATION
OVERHEAD
HIGH RAINFALL
PLANT UPTAKE
DECAY TO NONTOXICS
LEACHED TO GROUNDWATER
Figure 4.10 Fate of Aldicarb TTR after application to soil,
180
-------
scenarios. Table 4.4 gives the simulated percentages
accounted for by the three processes.
Figure 4.11 takes a more detailed look at the pesticide
leached under each scenario. Shown in the figure is the
geometric mean quantity of aldicarb TTR leached to saturated
zone over the 14 year period, expressed in Kg/ha, for each of
the sixteen unsaturated zone scenarios. The highest mean
annual loss occurs with overhead irrigation with a 270 cm (9
ft) unsaturated zone thickness (.076 Kg/ha-yr). This
represents 1.3% of the applied 5.6 Kg/ha. Next is the
ultisol, overhead irrigation, with a 180 cm (6 ft) unsaturated
zone thickness. Undoubtedly, highest loads occur on entisols
and ultisols with thin unsaturated zones, regardless of the
irrigation method used. When thick unsaturated zones are
present in combination with these soils, the loads are in
general lower than from the spodosols. Lowest leached loads
are associated with alfisols.
It is obvious from looking at Figure 4.11 that there are
differences among the mean responses from various soil orders,
unsaturated zone depths, irrigation methods and climatic
scenarios. The effect of unsaturated zone depth in the
entisols and ultisols and the effect of soil order seem to be
the most pronounced. Irrigation method seems to be less
important as does high or low rainfall in the spodosols and
alfisols. Initially, these scenarios were delineated because
it was felt that significantly different responses would
result. To test this assumption several analyses of variance
(ANOVA) tests were performed.
All ANOVA tests performed were one way. Tests were done on
all "treatments," that is, soil order, irrigation method,
unsaturated zone thickness (entisols and ultisols only) and
high and low rainfall (spodosols and alfisols only). As an
181
-------
TABLE 4.4 FATE OF ALDICARB TTR AFTER APPLICATION TO SOIL
PERCENT OF APPLICATION (in block)
Soil Type
ULTISOLS
ENTISOLS
00
to
ALFISOLS
SPODOSOLS
6 FOOT CORE
30 FOOT CORE
9 FOOT CORE
30 FOOT CORE
LOW RAINFALL
HIGH RAINFALL
LOW RAINFALL
HIGH RAINFALL
Irrigation
No
Overhead
Low Vol.
No
Overhead
Low Vol.
Spray
Spray
No
Overhead
Low Vol. Spray
No
Overhead
Low Vol. Spray
No
Overhead
Low Vol. Spray
No
Overhead
Low Vol. Spray
No
Overhead
Low Vol. Spray
No
Overhead
Low Vol. Spray
Plant
Uptake
15.57
21.70
54.97
6.91
13.57
45.46
14.36
22.44
56.51
9.38
17.61
51.17
15.33
25.28
45.31
12.55
24.74
48.31
18.13
30.09
51.35
15.54
28.76
54.46
Decay
83.70
76.95
44.17
93.08
86.40
54.53
83.80
74.49
41.54
90.55
82.26
48.79
84.66
74.70
51.68
87.40
75.21
51.68
81.80
69.82
48.64
84.19
70.97
45.49
Leached
0.73
1.35
0.86
0.01
0.03
0.01
1.84
3.07
1.95
0.07
0.13
0.04
0.01
0.03
0.01
0.05
0.05
0.01
0.07
0.09
0.01
0.27
0.27
0.05
-------
CO
to
10-°
MEAN ANNUAL PESTCIOE LOAD LEACHED TO SATURATED ZONE (KG/HA)
10'4 10-3 10-* 10-1
NO IRRIGATION
OVERHEAD
6 FOOT CORE
ULTISOLS
NO IRRIGATION
OVERHEAD
30 FOOT CORE
NO IRRIGATION
OVERHEAD
9 FOOT CORE
ENTISOLS
NO IRRIGATION
OVERHEAD
30 FOOT CORE
NO IRRIGATION
OVERHEAD
LOW RAINFALL
ALFISOLS
NO IRRIGATION
OVERHEAD
HIGH RAINFALL
NO IRRIGATION
OVERHEAD
LOW RAINFALL
SPODOSOLS
NO IRRIGATION
OVERHEAD
HIGH RAINFALL
Figure 4.11 Geometric mean annual quantity of pesticide leached to the
saturated zone from the treated band.
-------
example, the one way ANOVA of irrigation method is shown in
Figure 4.12. The data are grouped by irrigation method. The
ANOVA table shows that the computed F value of 0.91 does not
exceed the critical value of 3.1 at the 90% confidence levels.
Therefore, there is no evidence that there is a difference in
mean response due to irrigtion method.
s.
Results of the other ANOVA tests were that soil order was not
signficant, high and low rainfall levels in the alfisols was
not significant, but thickness of the unsaturated zone in
entisols and ultisols was significant. In addition, an ANOVA
in which ultisols and entisols were grouped and spodosols and
alfisols were grouped gave a significant F value. Therefore,
from these analyses, it appears that the statistically
different mean responses of the sixteen scenarios breakdown
into three groups;
1) entisols and ultisols with thin unsaturated zones
(< 270 cm),
2) entisols and ultisols with thick unsaturated zones
(> 270 cm), and
3) alfisols and spodosols.
Further justification for combining the results of scenarios
is given by the frequency plots of the annual pesticide loads
from each scenario. These are shown in Figures 4.13 through
4.16. Figures 4.13 and 4.14 show the eight plots for entisols
and ultisols, thin and thick unsaturated zones, no irrigation
and irrigation practices for each. Notice that the curves
group in both figures by depth of unsaturated zone. The thin
unsaturated zones fall virtually on top of one another. The
thick soils are also grouped but not to the extent of the thin
soils.
184
-------
DATA
Irrigation Method
N O
1.82E-2
1.02E-4
2.33E-2
1.61E-3
3.36E-3
6.65E-3
3.17E-4
4.31E-4
ANOVA Table
Source
Irrigation method
Error
Total
4.24E-2
6.92E-4
7.59E-2
3.87E-3
3.56E-3
5.59E-3
4.45E-4
4.42E-4
Sums of
Squares
0.0004
0.006
0.0064
d.f .
1
14
15
Mean
Square
0.0004
0.0004
0.91
o.io
3'102
Figure 4.12 One way analysis of variance of the effect of
irrigation on method annual pesticide mass
leached to the saturated zone (kg/ha).
185
-------
100-r
90"
U
X
»
Q
<
80"
70'
S 60'
40
ui 30
O
z
I 20 +
O
UJ
O
oc
01
Q.
10-
30 FOOT CORE
OVERHEAD IRRIGATON
NO IRRIGATION
9 FOOT CORE
OVERHEAD IRRIGATION
NO IRRIGATION
H 1 I II
1 I I I
1 1M-
1.E -5 1.E -4 1.E -3 -I.E -2
PESTICIDE LOAD IN KG/HA
1.E -1
1..E -0
Figure 4.13 Frequency of annual quantity of pesticide
leached to the saturated zone from Entisol
scenarios.
186
-------
30 FOOT CORE
OVERHEAD IRRIGATON
NO IRRIGATION
6 FOOT CORE
OVERHEAD IRRIGATION
NO IRRIGATION
1.E-5
1.E -4 1.E -3 1.E -2
PESTICIDE LOAD IN KG/HA
1.E -1
1..E -0
Figure 4.14
Frequency of annual quantity of pesticide
leached to the saturated zone from Ultisol
scenarios.
187
-------
100
Q
111
§ 80
o
X
co
80-
O .
s70
S 60
I-
2 so
40-
tu 30 -
O
z
x 20 +
O
10-
o
cc
ui
Q.
1.E-5
-tf-
HIGH RAINFALL
OVERHEAD IRRIGATION
NO IRRIGATION
LOW RAINFALL
OVERHEAD IRRIGATION
.... NO IRRIGATION
-iI I I I
1.E -4 1-E -3 -I.E -2
PESTICIDE LOAD IN KG/HA
1.E -1
1.E -0
Figure 4.15 Frequency of annual quantity of pesticide
leached to the saturated zone from the
Spodosol scenarios.
188
-------
HIGH RAINFALL
OVERHEAD IRRIGATION
NO IRRIGATION
LOW RAINFALL
OVERHEAD IRRIGATION
NO IRRIGATION
1.E -5 1-E -4 1-E -3 1.E -2 1.E -1
PESTICIDE LOAD IN KG/HA
1.E -0
Figure 4.16 Frequency of annual quantity of pesticide
leached to the saturated zone from the
Alfisol scenarios.
189
-------
Figures 4.15 and 4.16 show the same frequency curves for the
spodosols and alfisols. In these plots, all the curves are
nearly coincident. These figures reveal that highest leaching
losses were less then 1 Kg/ha over the fourteen year period.
The frequency of occurrence of any given level of pesticide
loading to the saturated zone can also be determined. For
instance, in Figure 4.13 for a no-irrigation, thick
unsaturated zone entisol scenario, there is approximately a
90% chance that the annual pesticide leach load in any given
-4
year will exceed 1.0 x 10 Kg/ha. There is only a 10%
-2
chance, however, that the annual load will exceed 1.0 x 10
Kg/ha.
By condensing the information in the sixteen scenarios down to
three, as mentioned earlier, better probability estimates are
obtained. Figures 4.17 through 4.19 show the three resulting
frequency distributions for the thick unsaturated zone
ultisols and entisols, thin unsaturated zone ultisols and
entisols, and the alfisol and spodosol groupings. Points in
the combined data sets were ordered and plotting positions
calculated using the Weibull distribution (Chow, 1964).
Of course, the loads shown in these figures are leached only
from the treated band. Much of the area in the unit citrus
block is untreated, with no resulting leached load.
Therefore, the loads from Figures 4.17 through 4.19 must be
reduced by the ratio of treated to total area in the unit
block. This ratio is approximately 0.32 to 1 for ridge citrus
as per Figure 4.9, and 0.20 to 1 for double bedded flatwood
citrus assuming a 3 m (10 ft) band down the middle on a 7.6 m
by 4.6 m (25 by 15 ft) tree spacing.
Table 4.5 summarizes the pesticide loads for the 90, 50 and 10
percentile exceedance probabilities for the three final
190
-------
RIDGE SOILS
THICK UNSATURATED ZONE
1.E-5 1.E-4 1.E-3 1.E -2
PESTICIDE LOAD IN KG/HA
1.E-1
1.E 0
Figure 4.17
Frequency of annual quantity of pesticide
leached to the saturated zone from thick
Entisols and Ultisols.
191
-------
0.00
1.E-5
RIDGE SOILS
THIN UNSATURATED ZONE
1.E-4 1.E-3 1.E -2
PESTICIDE LOAD IN KG/HA
1.E-1
1.E 0
Figure 4.18 Frequency of annual quantity of pesticide
leached to the unsaturated zone from thin
Entisols and Ultisols.
192
-------
FLATWOODS SOILS
SPODOSOLS AND ALFISOLS
1.E-5 1.E-4 1.E-3 1.E -2
PESTICIDE LOAD IN KG/HA
1.E-1
1.E 0
Figure 4.19 Frequency of annual quantity of pesticide
leached to the saturated zone from Spodosols
and Alfisols.
193
-------
TABLE 4.5 SUMMARY OF PESTICIDE LOADINGS PER UNIT CITRUS BLOCK
AREA FOR THE THREE FINAL UNSATURATED ZONE SCENARIOS
Exceedance Probability
Scenario 0.90 0.50 0.10
Alfisols and spodosols 1.8E-5 4.0E-4 2.2E-3
Thick ultisols and entisols 6.4E-6 4.2E-4 3.2E-3
Thin ultisols and entisols 1.9E-3 1.3E-2 9.6E-2
194
-------
unsaturated zone scenarios. These loads have been multiplied
by the above ratios to yield loads per unit citrus block
instead of unit treated area. The table reveals that lowest
loadings at all percentiles are associated with alfisols and
spodosols. For this scenario there is only a 10% probability
that the pesticide load leached to ground water will exceed
0.002 Kg/ha. Loads from thick entisols and ultisols exceed
those for the alfisols and spodosols slightly. The highest
loadings emanate from thin unsaturated zone entisols and
ultisols. There is a 10% chance that loads to ground water
under these soils will exceed 0.1 Kg/ha. Recall that the
thickness of the unsaturated zone in this scenario is 180 to
270 cm (6 to 9 ft) and that the input load is 5.6 Kg/ha or 5
Ib/acre. Increasing or decreasing the load in any of these
scenarios by a ratio 'x' would result in an increase or
decrease in the leached load by the same ratio. For instance,
if the application rate were doubled from 5.6 Kg/ha to 11.2
Kg/ha, the simulated load at the 10% exceedarice level in thin
entisols and ultisols would also double, from 0.096 Kg/ha to
0.19 Kg/ha.
Also of interest is the quantity of aldicarb and that of its
two toxic metabolities in the leached load. Table 4.6
summarizes the average percentage of the load accounted for by
each species. Overall, very little aldicarb parent is leached
to the saturated zone under any scenario. Under the thin
unsaturated zone ultisols and entisols, about 60% aldicarb
sulfoxide 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. The spodosols and alfisols, the ratio is roughly the
same, 0.19 to 1.
Obviously, since these transformations are kinetically
controlled, the quantity of aldicarb and its toxic
195
-------
TABLE 4.6 MEAN PERCENTAGE OF ALDICARB, ALDICARB SULFOXIDE AND
ALDICARB SULFONE IN THE SIMULATED LEACHED
PESTICIDE LOAD
Scenario
EN2.7
EN9.0
E02.7
E09.0
UNI. 8
UN9.0
U01.8
U09.0
SN1 . 2L
S01.2L
SN1 . 2H
S01.2H
AN1.2L
A01 . 2L
AN1.2H
A01 . 2H
Aldicarb
3
0
3
0
1
0
1
0
0
0
0
0
0
0
0
0
Percent Chemical
Sulfoxide
67
21
66
23
56
7
54
7
13
15
18
17
19
16
16
14
Sulfone
30
79
31
77
43
93
45
93
87
85
82
83
81
84
84
86
196
-------
metabolities 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. Figure 4.20 shows an example of the timing of
the mass load arrival at the saturated zone by month. The
quickest breakthrough and highest percentages of sulfoxide are
associated with the entisols and ultisols with thin
unsaturated zones. Over 90% of the total is leached to the
saturated zone within the first four months. For the thick
ultisols and entisols, the breakthrough curve is shifted to
the right, however, the major portion of the chemical still
leaches out in about four months. For the spodosols and
alfisols, the shape of the breakthrough curve is different.
Chemical leaches from the soil move evenly in time for these
soils. Not only does timing of the breakthrough affect the
quantities of aldicarb and its daughter products in the
leachate, but it also has other important implications for
modeling of the saturated zone. These are discussed in the
next section.
4.2 SATURATED ZONE MODELING
4.2.1 Choice of Model
The Coupled Fluid Energy and Solute Transport (CFEST) Code
(Gupta et al., 1980) was selected to predict the movement of
aldicarb in the saturated zone. CFEST is a multi-dimensional,
transient or steady state, saturated, flow, energy, and
contaminant transport model for predicting head and single
species solute concentration in a confined aquifer. The
numerical solution scheme used is the standard Galerkin
197
-------
Q
UJ
o
UJ
UJ
Q
5
tO
UJ
Q.
U.
O
UJ
U
oc
UJ
CL
UJ
P
_l
i
o
100-
90-
80-
70-
60-
50-
40-
30-
20-
10-
0
r
ULTISOLS, THIN
UNSATURATED ZONE
ULTISOLS , THICK
UNSATURATED ZONE
SPODOSOLS
( 1 1 r
J FMAMjjASOND
MONTHS
Figure 4.20 Cumulative mass curve of pesticide leaching
within the year for three representative
scenarios.
198
-------
finite-element method. The flow system may be complex,
multi-layered, heterogeneous, and anisotropic, with time
varying boundary conditions and time varying areal sources and
sinks.
Solute transport is calculated considering the processes of
dispersion, diffusion, and convection. Sorption and
contaminant degradation can also be simulated by the model.
In this study, CFEST was used to model each aquifer geometry
with and without decay and with different pumping rates, well
distances, and well depths. The model was run in two or three
dimensions, depending on the aquifer geometry and well depth.
The concentration of aldicarb in the drinking water well was
predicted by CFEST using steady state ground-water flow with
transient transport of aldicarb.
4.2.2 Model Configurations
Of the six aquifer systems delineated in Section 2.2.4, four
were simulated with variable well positions arid rates:
Floridan worst case, Floridan average cases, surficial worst
cases, and two-aquifer worst cases. The surficial average
cases and the two-aquifer average cases were not simulated due
to preliminary hydraulic results indicating very slow ground-
water velocities (0.002 m/day and 0.03 m/day in the surfical
aquifer, respectively). At this rate it would take many years
for water to travel from the edge of the source area to the
drinking water well 91 m (300 ft) away. Even with a decay
rate two or three times slower than the rate used in this
study, no significant quantities of aldicarb would ever reach
the well. Each aquifer system simulated with the saturated
model was run with two pumping rates, two distances from the
199
-------
source to the pumping well, and with and without decay. A
two-dimensional X-Y or three-dimensional X-Y-Z simulation was
used for each flow scenario. The two-dimensional X-Y
simulation was used to model the single aquifer configurations
when the drinking water well fully penetrates the aquifer.
Because the well itself integrates the aldicarb concentration
over the aquifer thickness, the two-dimensional X-Y simulation
which integrates over the vertical dimension is appropriate.
The three- dimensional X-Y-Z simulation is used when
simulating the two-aquifer system and a partially penetrating
well in an unconfined aquifer. In all, a total of 44 model
simulations were performed. The model input parameters for
all the cases are listed in Table 4.7.
A plan view of the modeled region is shown in Figure 4.21.
Although the region is 490 m (1,600 ft) wide, only half the
width was simulated in the model since the other half is the
mirror image. Figure 4.21 shows a distance of 91 m (300 ft)
between the source and pumping well. Each case was also
simulated with this distance being 300 m (1,000 ft). Due to
dimension constraints of the CFEST code, it was not always
possible to represent the field (left to right in Figure 4.21)
as being 825 m (2,700 ft) wide. Widths of 550 m and 275 m
were also simulated in some of the cases as shown in Table
4.7.
These different boundary conditions were used in the scenario
simulations as shown on the view of the model area (Figure
4.21). The boundaries parallel to the direction of flow were
no-flow boundaries. The boundary up-gradient of the well was
a constant-head or a specified flux boundary. The scenarios
which simulated partially penetrating shallow wells in the
Floridan Aquifer used constant-head up-gradient boundary. All
the other scenarios used a specified flux boundary for the
up-gradient boundary that maintained the regional gradient.
200
-------
TABLE 4.7 MODEL INPUT PARAMETERS FOR THE SIX AQUIFER TYPES SIMULATED
NJ
O
H1
Aquifer Type
Floridan Worst Case
(Shallow Well)
Floridan Worst Case
(Deep Well)
Floridan Average Case
(Shallow Well)
Floridan Average Case
(Deep Well)
Surftclal Worst Case
(Shallow Well)
Two-Aquifer Case
1) Surflclal
2) Confining Layer
3) Floridan
Field
Width
(ft)
2.700
2,700
1,800
1,800
900
900
-
Aquifer
Thickness
(ft)
600
350
600
350
35
40
10
400
Well
Screen
Interval
(ft)
10-40
0-350
10-40
0-350
0-35
.
-
50-400
Hydraulic
Conductivity
(ft/day)
5.000
5,000
1.000
1,000
100
40
0.025
1,770
Porosity
(X)
20
20
30
30
15
20
5
20
Gradient
(ft/ml)
3
3
2
2
4
25
40
10
Recharge
(in. /year)
25
25
13
13
30
20
-
Model
Dimension
3D
20
3D
2D
2D
30
-
Model
Grid
Spacing
(ft)
150
150
40
40
15
30
-
Mode
Tiim
Scei
(day
10
10
30
30
30
30
-
(1) Vertical Gradient
1 ft = 0.3048 m
1 in = 2.54 cm
-------
NO FLOW BOUNDARY
to
o
2700
NO FLOW BOUNDARY
Dimensions In Ft
Figure 4.21 Plan view of a hypothetical citrus grove and well configuration.
1 ft = 0.3048 m
-------
The boundary down-gradient of the well were all constant-head
boundaries.
-5 , j
The annual loading rate used in the model was 1x10 gm/cnr.
This loading rate was converted to an input concentration for
each aquifer case based on the average annual recharge. The
input concentrations at the source area (field) ranged from
399 to 920 ug/L (see Table 4.8).
All cases were simulated with and without decay. The
half-life used in the model was 29.7 days. Nine meters and
1.8 meters were used for longitudinal and transverse
dispersion, respectively. Typically one-tenth the critical
distance of travel is used for longitudinal dispersion and
transverse dispersion is taken to be one fifth the value of
longitudinal dispersion. The critical distance from the edge
of the field to the well is 91 m (300 ft), making longitudinal
dispersion 9 m and the transverse 1.8m. In general the model
is not very sensitive to dispersion, so when the well was
simulated at 300 m (1000 ft) from the source area, the same
dispersion values were used.
4.2.3 Input from Unsaturated Model
The output from the unsaturated zone model PRZM is the daily
water flux (cm) and daily pesticide loadings (gm/cm ). This
one-dimensional, daily pesticide output must be transformed
into meaningful input data for a single chemical component in
a two or three-dimensional ground-water model with a 10 to 30
day timestep. There are three aspects to interfacing the two
models:
203
-------
1) determining the magnitude of the pesticide load to
the saturated zone;
2) determining the temporal distribution of this load;
and
3) determining the spacial distribution of the load.
The magnitude and timing of the pesticide load to the
saturated zone would typically be prescribed directly by the
output from PRZM. Unfortunately, due to the time constraints
of the study, the two modeling efforts had to be done
simultaneously. A standard unit pesticide load of 1 x
10~5 gm/cm2 (1 Kg/ha) was used for all the scenarios.
Preliminary results from the unsaturated zone simulations
showed that this load was within the range of expected values.
Since the output of CFEST is linear with respect to input load
(see Section 4.2.6.2), the actual well water concentrations
for a given combination of saturated and unsaturated zone
scenarios can be determined by a multiplicaton factor based on
the actual output pesticide load from PRZM and the unit load.
Therefore, the output of CFEST is expressed as a "relative"
concentration, that is, the concentration simulated at the
well divided by the initial concentration calculated from the
unit pesticide load and the recharge rate shown in Table 4.8.
With pesticide applications occurring every year, there was a
possibility of aldicarb accummulating in the soil and the
loads to the saturated zone increasing. This possibility was
tested by analyzing the total annual leached load from each
year simulated. If the loads measured in each successive year
showed an increasing trend, this would indicate pesticide
accummulation. The results of analyzing the annual loads
showed that there is no upward trend of loads during the
simulation period.
204
-------
TABLE 4.8 INITIAL INPUT CONCENTRATIONS USED IN CFEST FOR EACH
AQUIFER SYSTEM SIMULATED (in ppb)
Aquifer Recharge Concentration
System (in/yr) (ppb)
Floridan, worst 25 479
Floridan, average 13 920
Surficial, worst 30 399
2-Aquifer worst 20 599
(1 inch = 2.54 cm)
205
-------
The second aspect in linking the unsaturated zone output to
the saturated zone modeling effort is determining how to
distribute the pesticide loads in time. Because of the manner
in which the timestep is handeled in CFEST, daily pesticide
loads had to be aggregated to monthly loads. Average monthly
loads were calculated to help determine the best method to
distribute the annual unit load. These monthly loads were
summed to determine the number of consecutive months required
for over 90% of the pesticide to be leached from the saturated
zone. Table 4.9 shows the results of this analysis for each
scenario.
For the ultisols and entisols, over 90% of the pesticide is
leached out within four and one-half months. In February,
pesticide would only be leached out the second half of the
month after the pesticide application in mid-February. In the
thick entisols and ultisols 5 to 6 months was usually required
to leach out over 90% of the pesticide. It is also
interesting to note that the primary leaching months over
which leaching occur start later in the year, reflecting the
additional time required for leaching through a soil four to
five times as thick.
The alfisols in areas with high rainfall leach close to 90% of
the pesticide within four months with irrigation. In areas
with lower rainfall, irrigation enhances the leaching so that
most of the pesticide is removed within four or five months.
Without irrigation, seven months is required to remove the
same amount.
The spodosols lengthen the time required to account for 90% of
the leachate to between five and eight months. The low
permeability layer in the soil slows down and prolongs
leaching. As with the alfisols in areas with high rainfall,
206
-------
TABLE 4.9 TIME REQUIRED TO LEACH 90% OF THE PESTICIDE FROM THE UNSATURATED ZONE
Consecutive
to
o
Soil Type
ULTISOLS
ENTISOLS
SPODOSOLS
ALFISOLS
Irrigation
Method
6 FT None
Overhead
30 FT None
Overhead
9 FT None
Overhead
30 FT None
Overhead
LOW RAIN None
Overhead
HIGH RAIN None
Overhead
LOW RAIN None
Overhead
HIGH RAIN None
Overhead
Months
Necessary
to Leach
90% of
Pesticide
5
4
4
5
5
4
6
6
8
7
6
6
7
5
5
4
Months
of
Leaching
Feb-June
March-June
June-Sept
May-Sept
Feb-June
Feb-May
March-Aug
March-Aug
March-Oct
March-Sept
April-Sept
March-Aug
March-Sept
April-Aug
April-Sept
April-Aug
Percent
Leached
95
93
94
90
99
97
92
94
96
93
93
93
95
90
95
90
Percent
Leached in
4 Months
86
93
94
86
88
97
78
58
67
58
78
81
78
84
89
90
-------
the pesticide travels faster requiring six months versus seven
to eight months in areas with low rainfall.
The more time the leaching of the annual load requires, the
smaller any load to the saturated zone will be. On the other
hand, if the entire annual load was leached in one or two
months, the concentrations entering the saturated zone would
be much higher.
The average of the number of months necessary to leach 90% of
the pfe-sticide is just over five. A four month period was used
in the simulations as this decision had to be made a priori
based on limited simulation results. A four month period for
loading will result in slightly higher concentrations at the
well for scenarios in which the load is distributed over a
longer period of time. The last column of Table 4.9 shows the
percentage of the pesticide that is leached in four months.
The spatial distribution of the pesticide load is a more
complex problem, which effects both the peak concentration at
the well and the time necessary for the peak to reach the
well. For some scenarios it was discovered that the width of
the source area had an effect on well-water concentrations.
Since actual field sizes in Florida vary tremendously, the
following analysis was done in an attempt to understand this
relationship so model results could be interpreted
accordingly.
4.2.4 Source Area and Well Concentration
A plug flow analysis reveals the basic relationship between
the field width and the concentration at the well. The key
variable is the residence time under the field which combines
208
-------
the ground-water velocity associated with a given scenario and
the width of the field. As the residence time under the field
increases, the water may be subjected to several pesticide
loading events. In cases with no decay or low decay, this can
result in higher concentrations at the well. Figure 4.22 is a
schematic graph of what the concentration of the water at the
well would be with and without decay based on the residence
time under the source area. At the beginning of the
simulation the concentration at the well will be zero. It
will remain zero for the length of time it takes the water to
travel from the leading edge of the field to the well. After
that, the concentration in the case of plug flow is directly
related to the residence time under the source area. The
concentration increases as more and more pesticide is leached
to the ground water. At the end of four months, no more
pesticide is leached from the unsaturated zone until the
following year. This means that even if the residence time is
over four months, the water only receives four months of
pesticide loading in one year. One way to visualize this
process is to imagine a parcel of water moving under the field
(Figure 4.23). While it is under the field and leaching is
occurring, the concentration in the parcel of water continues
to rise at a steady rate. If the water is under the field
when leaching is not occurring or when it is traveling from
the field edge to the well, the concentration remains constant
(assuming zero decay).
When there is no decay the water reaches a local maximum
concentration at a residence time of four months and then
plateaus for the eight months of no pesticide loading. When
leaching begins again the following year the concentration
begins to rise in the water below the field. For longer and
longer residence times the concentration will continue to rise
following this general pattern. In the case with decay the
pesticide reaches a maximum in four months, then begins to
209
-------
g
H
CC
J-
UJ
o
o
o
PLUG FLOW WITH NO DECAY
PLUG FLOW WITH DECAY
AND CARRY OVER
4,
PLUG FLOW WITH DECAY
AND NO CARRY OVER
TRAVEL
TIME
TO WELL
4 MONTHS 8 1 YEAR
RESIDENCE TIME UNDER FIELD
2 YEARS
Figure 4.22
Schematic graph of pesticide concentration
through time based on residence time under
source area.
210
-------
r C
DRINKING
WATER
\ WELL
-1_
'PARCEL* OF WATER
^^:^y^y^^^'::-:^^'i^f^.
Figure 4.23 Schematic diagram of pesticide accumulation
process.
211
-------
decay during the eight months of no loading. The cycle
repeats itself annually.
For a given residence time there are two general patterns: 1)
when the residence time is less than one year, and 2) the
other is when the residence time is greater than one year.
Figure 4.24 shows what the concentration pattern would look
like if the residence time of ground water under the field
were eight months. The concentration would rise to a maximum
during the four initial months of leaching, maintain that peak
for four months and then decrease for four months. The cycle
then repeats itself year after year. For residence times
greater than one year the concentration level will follow the
same pattern shown in Figure 4.22, leveling off at the
concentration corresponding to the appropriate residence time.
The decay curve has basically the same shape regardless of the
residence time. It is conceivable that if all the pesticide
does not decay in one year, the annual peak concentration
could go up over the years as shown in the curve with carry-
over in Figure 4.22.
The very real process of dispersion requires modification of
the plug flow theory. Dispersion causes the spreading of the
contaminant. Figure 4.25 shows how dispersion would change
the concentration in the well predicted by plug flow in an
example where the residence time was greater than three years.
A series of sensitivity runs were made with CFEST for two of
the aquifer cases to help understand this relationship. In
the first set of sensitivity runs the Floridan worst case
deep well, 1,000 gpm pumping) was run both with and without
decay where the field width ranged from 46 m (150 ft) up to
823 m (2,700 ft) (the maximum field width for this case). A
second set of sensitivity runs was made for the surficial
212
-------
jjji
z
o
Ul
O
z
o
o
1r
0
TRAVEL
TIME
TO WELL
PLUG FLOW. NO DECAY. RESIDENCE TIME = 8 MONTHS
2YR 3YR
TIME (AFTER FIRST PESTICIDE ARRIVAL)
4YR
Figure 4.24 Concentration profile for eight month
residence time based on plug flow.
213
-------
Z
UJ
o
z
o
o
PLUG FLOW-
WITH DISPERSION
TRAVEL
TIME
TO WELL
TIME (YEARS)
Figure 4.25 Effect of dispersion on concentration in the
well over time.
214
-------
worst case (shallow well, 200 gpm pumping rate) both with and
without decay where the field width ranged from 46 m (150 ft)
up to 275 m (900 ft) (the maximum field width for this case).
In all scenarios the pumping well was 91 m (300 ft) from the
edge of the field. Each sensitivity simulation ran for three
years with identical model parameters except for the field
width.
The results of the sensitivity analysis for the worst case of
the Floridan Aquifer are shown in Figures 4.26 to 4.28.
Figure 4.26 shows the concentration at the well versus time
for the first year of simulation with no decay. This clearly
demonstrates how the concentration at the well increases as
the field width increases. Figure 4.27 is the same scenario
with decay. There the concentration also increases with the
field width to a critical width of 366 m (1200 ft) beyond
which the concentrations essentially do not change. Figure
4.28 summarizes these trends for all the scenarios with and
without decay.
The results for the surficial worst cases, both with and
without decay, are shown in Figures 4.29 and 4.30. For the
case without decay (Figure 4.29), a peak concentration at the
well was never reached in the three year simulation. The
relationship between the maximum concentration after three
years and field width is such that the concentration gradually
increases as field width increases. The case with decay
(Figure 4.30) does reach a peak concentration at the well on
an annual cycle. This peak does not change as the field width
changes.
In all these simulations, the fields modeled all have
residence times much greater than four months. As observed in
the Floridan worst case sensitivity runs, for residence times
greater than a critical time of about three months, the peak
215
-------
\\. \ \
A 150 ft
X 300 ft
£50 «_
D 600 ft
B 1200_ft
* 1800 ft
X 2700 ft
4 MONTHS 8 MONTHS
TIME, MONTHS
12 MONTHS
Figure 4.26 Concentration histories at variable field
widths for the Floridan worst case scenario
with no decay.
216
-------
1 -I
o
LU
O
z
o
o
1200 & 1800
4 MONTHS 8 MONTHS
TIME, MONTHS
12 MONTHS
A 150 ft
X 300 ft
D 60_0_ft
H 1200ft
4 1800ft
X 2700 ft
Figure 4.27 Concentration histories at variable field widths
for the Floridan worst case scenario with decay.
217
-------
to
M
00
D)
4-
3H
CE
LU 2-
O
O
CJ
UJ
Q.
1-
Legend
A pjormxn MB
500 1000 1500 2000 2500
FIELD WIDTH, Ft
3000
Figure 4.28 Relationship of field width to peak concentration at the
well for the Floridan worst case.
-------
30-
09
i
P
tt 20-
ui
o
o
u
10-
0 1 YEAR 2 YEARS 3 YEARS
TIME, YEARS
LEGEND
A150 ft
X 300 ft
D 600 tt
B 900 ft
Figure 4.29 Concentration histories at variable field widths
for the surficial worst case scenario without
decay.
219
-------
1.5-
O
oc
UJ
u
O
O
150ft
300ft
600ft
900 ft
4 MONTHS 8 MONTHS
TIME, MONTHS
12 MONTHS
Figure 4.30
Concentration histories at variable
field widths for the surficial worst
case scenario with decay.
220
-------
concentrations are the same.
The two cases run in this sensitivity analysis range over the
spectrum from the fastest travel time (Floridan worst case) to
the slowest travel time (surficial worst case) of the cases
simulated. The relationship between peak concentration at the
well and field width for the Floridan average cases would
probably fall between the two cases studied.
Based on these results it appears that, for all cases with
decay, the peak concentration at the well is independent of
field width except for small widths (i.e., residence times
less than four months). For the cases simulated without decay
the maximum concentration is dependant on the residence time
of the water under the field. The maximum concentration will
be higher for wider fields. The exact maximum depends not
only on the field width but the ground-water velocity (i.e.,
travel time to the well) and dispersion.
It should be noted here that the regularity of the cyclical
and "rising and plateauing" well-water concentration patterns
shown in the next section is caused by the loading assumptions
made for interfacing the two models and the assumption of
steady flow. While in nature, these patterns will be highly
irregular and noisy, the same general trends discussed in the
previous section should emerge.
4.2.5 Results of Aldicarb Simulations
A summary of the results for the four simulated aquifer
geometries are shown in Tables 4.10 to 4.13. These tables
show the relative peak concentrations, the time of arrival of
the peak along with the other parameters that describe each
221
-------
specific scenario. In addition to these tables, for each
scenario, a graph is presented showing the trend of the
relative concentration at the well during the three year
simulation period (Figures 4.32 to 4.39). Four scenarios are
shown on each graph for ease of comparison. For a given well
distance (i.e., 91 m versus 300 m) and well depth (shallow or
deep), concentration histories are shown for the high and low
pumping rates, with and without decay.
The results for the worst case for the Floridan Aquifer are
shown in Figures 4.31 and 4.32 and Table 4.10. The worst case
for the Floridan Aquifer has the highest ground-water
velocities (4.3 m/day or 14.2 ft/day based on the regional
gradient). With such high velocities, the ground water has
only a six-month residence time, the shortest residence time
of all the scenarios. This short residence time result in
cyclical concentration histories for the scenarios without
decay as well as the scenarios with decay, as shown in Figures
4.31 and 4.32. Recall that with residence times less than one
year, the concentration histories is cyclical in scenarios
with no decay (see Figure 4.24). Considering the short
residence time the relative concentrations at the well are
high in comparison to most other scenarios. This is also due
to the high velocities. There is little time for dispersion
to spread out the contaminant over the distance from the grove
(source area) to the drinking, water well. Based on the
well-induced velocities, the ground water takes from 15-20
days to travel to the well 91 m from the source and 62.5-69
days to travel to the 300 m well.
The average cases for the Floridan Aquifer (Table 4.11 and
Figure 4.33-4.36) fall into two catagories, the scenarios
without decay and scenarios with decay. In the scenarios with
decay, the concentration at the well cycles annually, as shown
most clearly on Figures 4.34 and 4.36. The annual peak occurs
222
-------
TABLE 4.10 SUMMARY OF RESULTS FOR FLORIDAN WORST CASE
SIMULATIONS
Well
Depth
(ft)
10-40
10-40
10-40
10-40
10-40
10-40
10-40
10-40
0-350
0-350
0-350
0-350
0-350
0-350
0-350
0-350
Well
Distance
(ft)
300
300
300
300
1,000
1,000
1,000
1,000
300
300
300
300
1,000
1,000
1,000
1,000
Decay
No
No
Yes
Yes
No
No
Yes
Yes
No
No
Yes
Yes
No
No
Yes
Yes
Pumping
Rate
(gpm)
200
40
200
40
200
40
200
40
1,000
500
1,000
500
1,000
500
1,000
500
Peak
Relative
Concentration
.1.5 x 10"2
1.6 x 10~2
3.1 x 10"3
3.0 x 10"3
1.3 x 10~2
1.3 x 10r2
7.3 x 10"4
7.1 x 10"4
9.0 x 10"3
9.0 x 10"3
1.9 x 10"3
1.9 x 10"3
8.7 x 10"3
8.7 x 10"3
7.0 x 10"4
6.8 x 10"4
Time
to Peak
(days)
150
160
120
120
230
230
150
150
150
150
120
120
210
210
140
140
223
-------
to
N)
R
8
w
2S-
2d'
1-1
§§
53
O O
"ij
> °.-1
-> a
£
FVD(0)W(300.10-40.40)
FWD(0)W(300,10-40.200)
A
. .
w
\
FWO(29.7)W(300.10-40.200)
FWD(29.7)U(300,10-40.40)
I V
v / V
0.0 180.0 360.0 540.0 730.0
TIME IN DfflfS
a. 91 m to well
900.0
1080.0
°
Is
i.
a
§
FUD(0)U(1000.10-40.40)
I nro(0)H(1000.10-40.200)
FVD(29.7)K(1000,10-40.40)
FUD(29.7)W(1000.10-40.200)
^J* .^'x.
0.0 180.0 360.0 540.0 730.0
TIME IN DflYS
b. 300 ra to well
900.0 1080.0
Figure 4.31 Concentration versus time for Floridan worst case with a shallow
well.
-------
to
FWD(0)W(300.50-350.1000)
FUD(0)U(300.50-350.
. FUD(29.7)W(300.50-350.1000)
r FWD(29.7)W(300.50-350.500)
180.0 360.0 540.0 730.0
TIME IN DRYS
a. 91 m to well
800.0 1080.0
cu V'
»t
°
FUD(0)W(1000.50-350.1000)
FWD(0)W(1000.50-350.500)
FWD(29.7)U(1000.50-350.1000)
FMD(29.7)WMOOO. 50-350.500)
0.0 180.0 360.0 540.0 730.0
TIME IN DfttS
b. 300 m to well
800.0 1080.0
Figure 4.32 Concentration versus time for the Floridan worst case with a deep well
-------
TABLE 4.11 SUMMARY OF RESULTS FOR FLORIDAN AVERAGE CASE
Well
Depth
JftL
10-40
10-40
10-40
10-40
10-40
10-40
10-40
10-40
0-350
0-350
0-350
0-350
0-350
0-350
0-350
0-350
Well
Distance
(ft)
300
300
300
300
1,000
1,000
1,000
1,000
300
300
300
300
1,000
1,000
1,000
1,000
Decay
No
No
Yes
Yes
No
No
Yes
Yes
No
No
Yes
Yes
No
No
Yes
Yes
Pumping
Rate
(gpm)
200
40
200
40
200
40
200
40
1,000
500
1,000
500
1,000
500
1,000
500
Peak
Relative
Concentration
8.3 x 10"3
8.8 x 10"3
5.1 x 10"5
4.6 x 10"5
2.5 x 10"3
2.4 x 10"3
1.6 x 10'8
1.2 x 10"8
7.4 x 10"3
8.8 x 10"3
5.1 x 10"5
4.5 x 10"5
3.6 x 10"3
4.3 x 10"3
5.5 x 10'8
4.2 x 10"8
Time
to Peak
(days)
A
A
150
150
A
A
300
300
A
A
120
150
A
A
270
300
Aj Peak concentration was not reached in 3-year simulation.
226
-------
00
NJ
F-i 10
-------
O
a-
o
ji-
o
*
NJ
M
00
So.
Of o
FAD(29.7)W(300.10-40.200)
n
in
K-
FAD('29.7JW(300.10-40,40)
Vff i
g2
to
0.0 180.C 360.C &U.O 723.0
TIME IN DRYS
a. 91 m to well
9CC.O
108C.O
FAO(29.7)H(1000.10-40.200)
V
FAD(29.7)U(1000.10-40.40)
0.0 180.0 360.0 StO.O 720.0
TIME IN DflYS
b. 300 m to well
900.0
1C80.0
Figure 4.34 Detail of Floridan average case with a shallow well and decay.
-------
2 2
X °»
Ni
to
VO
FAD(0)W(300.50-350.500)_ /'
.50-350.1000)
^FAD(Z9.7)W( 300.50-350.1000)
XFAD(29.7)W( 300.50-350.500)
0.0 180.0 360.0 Stt.O 7X.O 900.0 1080.0
TIME IN DflYS
a. 91 m to well
oq
E*
FAD(0)H(1000.50-350,SCO)^ /
FAD(0)W(1000,50-350.1000
FAD(29.7)W(1000.50-350.1000)
FAD(29.7)W(1000.50-350.500)
0.0 180.0 360.0 640.0 730.0
TIME IN DflYS
b. 300 m to well
900.0 1080.0
Figure 4.35 Concentration versus time for the Floridan average case with a deep
well.
-------
to
u>
o
a-
a
«
aiS-l
t->
|3H
go
uKJ-
>
2-
o
FAD(29.7)U(300,50-350.1000)
T
2 2
X «°
a:
a:
to
*f
0
FAD(29.7)W(1000.50-350.1000)
FAD(29.7)W{1000.50-350.500)
0.0 180.0 360.0 5-10.0 720.0
TIME IN DflYS
a. 91 m to well
9CO.O
1C80.0
0.0 180.0 360.0 S-UJ.O 730.0
TIME IN DflYS
b. 300 m to well
900.0
iceo.o
Figure 4.36 Detail of Floridan average case with a deep well and decay.
-------
anywhere from 120 days to 300 days after the beginning of
leaching, depending on the aquifer scenario and distance to
the well. In the cases without decay, the concentrations at
the well increase throughout the three-year simulation period
with minor fluctuations (Figures 4.33 and 4.35). The ground
water in these average cases for the Floridan Aquifer has a
residence time of almost four years. With such a lengthy
residence time, the concentrations would continue to rise
according to the theory developed earlier on residence times.
The fluctuations visible in the cases with the well at 91 m
are a dispersed form of the plateaus predicted by a plug flow
analysis. At 300 m, these fluctuations have been smoothed out
by dispersion. The initial rise is also delayed, as would be
expected when the well is further from the source area.
The worst case for the surficial aquifer geometry shows trends
very similar to the Floridan average cases because of its long
residence time of five years. Table 4.12 shows the relative
peak concentrations and Figures 4.37 and 4.38 show the
concentration histories. The scenarios with decay cycle
annually. The relative concentrations are the highest of all
of the scenarios with decay. This is because the aquifer is
very shallow so the dilution of the contaminant can not be as
great as the systems with deep aquifers. In the cases without
decay, the concentrations continuously rise throughout the
three-year simulation. In these scenarios there is a large
difference between the high pumping rate and the low pumping
rate. In the surfical aquifer the well influences the
regional hydraulic gradient more than in any of the other
aquifer configurations. The well with the high pumping rate
is able to pump a higher proportion of clean water than the
well with the lower rate. The plateaus are much more defined,
to the point of decreasing before the next increase. This is
due to the slow velocities and dispersion of the contaminant
at the well.
231
-------
TABLE 4.12 SUMMARY OF RESULTS FOR THE SURFICIAL WORST CASE
Well
Depth
JftI
0-35
0-35
0-35
0-35
0-35
0-35
0-35
0-35
Well
Distance
(ft)
300
300
300
300
1,000
1,000
1,000
1,000
Decay
No
No
Yes
Yes
No
No
Yes
Yes
Pumping
Rate
(gpm)
200
40
200
40
200
40
200
40
Peak
Relative
Concentration
8.5 x 10"2
2.4 x 10"1
3.8 x 10"3
3.4 x 10"3
7.1 x 10~2
1.2 x Kf1
3.8 x 10"5
3.3 x 10"5
Time
to Peak
(days)
A
A
120
120
A
A
210
240
AJ Peak concentration was not reached in 3-year simulation.
232
-------
R
R
8
NJ
u>
u>
00
u
Ul
a:.
SWD{0)H(300.10-35.40)
SW)(0)W(300.10-35.200)
SKD(29.7)H(300.10-35.200)
'SVD(29.7)H(300.10-35.40)
0.0 190.0 360.0 Stf.O 730.0
TIME IN DflYS
800.0
1080.0
28
So
SJ
*I
Si
UO
8
//ksHD(0)U( 1000.10-35.40)
SUD(0)U(1000.10-35.200)
SWD(29.7)W(1000.10-35.200)
. SUD(29.7)W(1000.10-35.40)
0.0 180.0 360.0 540.0 730.0
TIME IN DfttS
900.0 1080.0
a. 91 m to well
b. 300 m to well
Figure 4.37 Concentration versus time for the surficial worst case.
-------
* °~\ SKD(29.7)W(1000.10-35.200)
A*
B- f\
,4
' \
is~ // \
. i i
t-« o s ' !|
(E i ' <>
?? B- I ' t>
i / / \\
O 9- / i Vi
C3 " ! > li
: i : i
S«8 // \\
1 2" / 1 \\
~5 ! i : i
=-i » i . i
ty . it : >
of : i . i
O j ' i 't
2- // yv
°. //
u> :»
° /''
ra >
A A
// \ // \
// \ // \
/ / '\ / / M
// n /; V.
; ' : t :>
/ ' Sl ; I '.'
/ \\ // \\
// \\ // \\
/ / \\ l> \\
// \\ // \\
// \\ // \
II u // i
SUD(29.7)W(1000.10-35.40)
0.0 180.0 3EO.O 540.0 720.0
TIME IN DRYS
9CO.O
1CSO.O
Figure 4.38
Detail of the surficial worst case with the
well 1000 ft from the source with decay.
234
-------
TABLE 4.13 SUMMARY OF RESULTS FOR TWO-AQUIFER SYSTEM
(WORST CASE)
Well
Depth
(ft)
50-400
50-400
50-400
50-400
Well
Distance
(ft)
300
300
300
300
Decay
No
No
Yes
Yes
Pumping
Rate
(qpm)
1,500
700
1,500
700
Peak
Relative
Concentration
1.8 x 10"3
2.2 x 10"3
3.2 x 10"4
3.0 x 10"4
Time
to Peak
(days)
150
150
120
120
235
-------
Oo
S ">'
5-5
LJ
0(3
U O'
> *
f^
i5'
r\
\*2W>{0)K(300.50-400.700)
0.0 180.0 360.0 540.0 720.0
TIME IN DflYS
800.0
1080.0
Figure 4.39 Concentration versus time for the two-aquifer
worst case with the well 300 ft from the source,
236
-------
Some of the scenarios for the two-aquifer, worst case geometry
were modeled. In these cases the well was in the lower
aquifer. Consequently the contaminant must travel through the
upper aquifer and confining layer to reach the well. The
boundary conditions for the two-aquifer cases were such that
unusually large concentrations of the contaminant accummulated
in center portion of the upper aquifer. This led to higher
than normal transport of the contaminant to the lower aquifer
and drinking water well. The results are shown, nevertheless,
in Figure 4.39 and Table 4.13. Even with the hydraulics that
enhance transport the relative concentrations are small.
Plug Flow Modeling Results
A simple plug flow model of the single aquifer systems was
also used to aid in the understanding and extrapolation of
results of the CFEST simulations. Maximum concentrations for
several scenarios were calculated using the plug flow model
developed earlier in this section. For these calculations the
contaminant load was assumed to disperse evenly in a column of
water (1 cm x 1 cm x aquifer depth x porosity). The water
-5 2
receives the annual unit load (1 x 10 gr/cm ) over four
months. In the cases with no decay this maximum concentration
does not change while traveling to the well because of the no
dispersion assumption in plug flow. In the cases with decay
the concentration was first calculated under the field
assuming four months of loading:
C = £(l - e"kt)/be (4.5)
L. K.
where C^ = concentration under the field
R = rate of pesticide loading
237
-------
k = decay rate
t = time of loading (4 months)
b = aquifer depth, and
e = porosity
The concentration at the well can then be calculated using the
following relationship:
_kx
C = C- e v (4.6)
w f
where C = concentration at well
w
Cf = concentration under field
k = decay rate
x = distance to well from field
v = average velocity of water from field to well
Table 4.14 compares those plug flow concentrations to the
concentrations determined in the computer simulations. The
degree of agreement relates directly to how much dispersion
occurs and dilutional differences between the single well used
in CFEST and a line of wells assumed for the plug flow model.
The worst cases for the Floridan Aquifer show remarkable
agreement. The average cases for the Floridan Aquifer show
fairly good agreement. The worst agreement is between the
cases for the surficial aquifers.
This technique of calculating concentrations for single
aquifers with plug flow is useful for determining upper limits
on the amount of pesticide contamination without the use of a
sophisticated computer solute transport model.
Several general trends can be noted in the scenarios simulated
with the computer. The decay scenarios versus the scenarios
238
-------
TABLE 4.14 COMPARISON OF PEAK CONCENTRATIONS DETERMINED BY A SIMPLE PLUG
FLOW MODEL AND BY COMPUTER SIMULATION
Predicted Well Water Concentrations
(ppb)
Floridan worst case (deep well)
No decay
Decay
300 ft well
1000 ft well
Plug Flow
4.69
1.05-1.11
.35-.37
Computer Simulation
4.2-4.3
.91-.93
.33-.34
to
u>
vo
Floridan average cases (deep well)
No decay
Decay
300 ft well
1000 ft well
9.36
0.13-0.28
.000072
3.3-8.1
0.041-0.047
.000039-.00005
Surficial worst case
No decay
Decay
300 ft well
1000 ft well
62.5-187.5
4.75-9.17
.02-.74
28.1-94.7
1.4-1.5
.013-.015
-------
with no decay have the greatest differences, the relative peak
concentrations dropping several orders of magnitude with
decay. The well parameters also show general trends in the
pumping rates, the depth and the distance from the source
area. These observations are enlarged upon in the following
sections.
4.2.5.2 Pumping Rates
The results from the simulations show that the pumping rates
do not affect the peak relative concentration or the general
shape of the concentration profile. In the Floridan worst
cases the greatest difference in peak concentrations is 3%.
The surficial worst cases with no decay show differences as
much as 50% between the high and low pumping rates, but even
in this case, the trends are very similar. The results show
that for all the aquifer geometries, when there is no decay,
the scenarios using the lower pumping rate have the higher
peak concentration. Conversely, in the cases with decay, the
scenarios with the higher pumping rate have the highest peak
concentration. With no decay, dilution becomes important.
The wells pumping at a higher rate bring more "clean" water
than the wells with lower rates, and consequently have lower
concentrations. This effect is especially apparent in the
worst cases for the surficial aquifer, where the pumping rate
dominates the flow field as was shown in Figure 4.31. In the
worst cases for the Floridan Aquifer the regional gradient is
not affected very much by the well.
In the scenarios with decay, the wells with higher pumping
rates have higher concentrations. In those cases the higher
pumping rates produce higher ground-water velocities near the
wells. Higher velocities mean short travel times and
therefore less time for decay to occur.
240
-------
4.2.5.3 Well Depths
Two different well depths are simulated in the worst cases and
average cases for the Floridan Aquifer. In the Floridan worst
cases the concentrations are higher in the shallow well than
the deep well, when the well is both 91 m and 300 m from the
field. In these scenarios, the ground-water velocities are
high enough that not much downward dispersion occurs before
the contaminant reaches the well. In the average cases for
the Floridan Aquifer the concentrations are about equal in the
deep well and shallow well when the well is 91 m from the
field. When the well is 300 m from the field, the deep well
shows higher concentrations than the shallow well. In the
average cases the ground-water velocities are slower than in
the worst case. This means that there is more time for
downward dispersion to occur. At 91 m from the field,
dispersion has spread the contaminant in the three-dimensional
simulation of the shallow well, so that the concentration is
about equal to the concentration in the deep well. The
difference at 300 m can only be accounted for by the
differences in aquifer depths used in the simulation. The
shallow well is modeled as a partially penetrating well in a
three-dimensional aquifer 183 m thick. The deep well is
modeled in a two-dimensional simulation of a 107 m aquifer.
The concentration in the deeper well is greater because it has
less aquifer depth in which to disperse.
4.2.5.4 Well Distances
Two different well distances are simulated with each scenario,
one at 91 m from the source area and the other 300 m from the
source area. As might be expected, the concentrations are
higher in wells 91 m away. In the cases with no decay, the
concentrations at the 300 m well are smaller because there is
more time for dispersion. The Floridan worst cases only show
241
-------
a difference of about 20% between the 91 m and the 300 m
cases. The average Floridan and surficial worst case differs
by a much greater amount (as much as 70%) due to the greater
amount of time for dispersion. In the cases with decay, the
relative concentrations are 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.
4.2.5.5 Source Surrounding the Well
In the production runs, the well was simulated as being
down-gradient from the edge of the grove. Sensitivity runs
were performed to determine the effect on peak concentration
at the well for the case where the well is surrounded by
citrus groves.
Two sensitivity runs were made using the Surficial worst case
(shallow well, 200 gpm pumping rate, decay) and the Floridan
average case (deep well, 1,000 gpm pumping rate, no decay).
These cases were chosen because the hydraulic conductivities
are low enough that pumping causes significant drawdown at the
well, thereby reversing the regional gradient. Figure 4.40
compares the potential surfaces for two scenarios, one where
the regional gradient dominates, the other where the
well-induced gradient dominates the regional gradient. The
reversed gradient allowed the well to pull contaminant from
all sides, not just from the up-gradient grove. In both cases
groves were assumed to surround the well at a distance of 91 m
(300 ft).
The results of these simulations (Figure 4.41) show that
aldicarb is pulled from the down-gradient field as well as
from the up-gradient field, and therefore, the peak
concentration at the well increases significantly. For the
242
-------
KH.E
1000
I
2000
j
350.0
349.6
349.2 349.0
348.6 s* 348.2
PUMPING WELL
a. Worst case for Floridan aquifer.
SCflLE
200 400
1 I
PUMPING WELL'
b. Worst case for surficial aquifer.
Figure 4. 40 Comparison of the two extreme hydraulic
potential distributions.
243
-------
3.
d
M
3.
Q
S
Z o.
2°
ff*
§"*
_.
z
o
£*
(-
tr
_i
"i
d
it
8.
d
t
d
o.
d
SOURCE SURROUNDING HELL
\ .-'" ^"
.'' r»
z 2-
2
,' 1-0
,' or s
£ 2'
* »^
**" -i-
/'* ^-"""" ^ §_
' -^"^ £2 d
'' / S JS
/ X * Q
/ "*^*^ ^ °"
«s^*^ FAD(0)W(300. 50-350. 1000) ^
x i
/ X
x'>^ ^-
->fti> , , , d
SOURCE SURROUNDING HELL
f '\ ,*>
, 1 / \ / \
/y / \ / \
/ \ / \ / \
/
/ %^*» '' s%** * "^»
^ »» *»»/ *»»«
/
i
;
*
f
1
1
t
1
r' SUD(29.7)V(300, 10-35.200)
A A7 A
/ \ ' ' / \
/ \ / \ /
1 \ / \ / \
. -- / . v-~,. /
0.0 180.0 360.0 540.0 720.0 900.0 1080.0
TIME IN DflYS
O.C 18C.C 350.0 54C.C 720.C 90C.O 108C.C
TIM.5 IK DflYS
a. Floridan average case without
decay.
b. Surficial worst case with decay.
Figure 4.41 Effect of Aldicarb source area surrounding the well.
-------
surficial worst case, the relative peak increases nearly a
factor of 4 over the case where the source was just
up-gradient from the well. In the Floridan average case, the
peak is not reached in 3 years, however, the concentration
after 3 years increases nearly a factor of 2 over the case
where the source was just up-gradient. The time to peak
remains the same in both cases.
The results show that peak concentrations can increase
significantly if the well is surrounded by the source area and
the drawdown at the well is large enough to locally reverse
the regional gradient.
4.2.5.6 Time to Reach a Peak Concentration
For some of the scenarios without decay (Floridan average
case, shallow and deep well; Surficial worst case, shallow
well) a peak was not reached at the pumping well during the
three-year simulation period. In order to gain an
understanding of the magnitude of the peak and time to peak
for these cases, the Floridan Average Case (shallow well, 200
gpm pumping rate, no decay, 300 ft from source to well) was
simulated for 20 years rather than the 3 years used in all
other cases.
The results of this case are shown in Figure 4.42. The peak
concentration at the well plateaus after 13 years at a value
of about 17 ug/L. This peak concentration is just over twice
the concentration at the well after 3 years. This same
relationship (i.e., actual peak concentration of about twice
the model predicted value, and a time to peak of 15 to 20
years) can probably be applied to all the Floridan average
cases where the peak is not reached in the three year
simulation period.
245
-------
cn
20-1
D)
15H
CE
uu
o
1CH
8 5-,
LLl
Q_
FAO(0)U(300.1»-40.200)
5 10 15
TIME, Yrs
20
Figure 4.42 20-year simulation of average Floridan case,
-------
Model runs were not performed to determine the time to peak
for the Surficial worst cases that do not reach a peak in
three years.
4.2.6 Sensitivity Analysis and Verification
Several sensitivity runs were made to determine the
sensitivity of the model results to assumptions made in the
modeling concerning time step and input concentration and the
validity of the assumption that the flow can be simulated as
steady state. The sensitivity runs and steady state versus
transient comparisons are discussed in this section.
4.2.6.1 Time Step
When simulating contaminant transport with a model that uses
the convective dispersion equation (such as CPEST), it is
important to select the model time step such that the
contaminant travels across one element (i.e., the finite
element used to represent the model region) in one time step.
If this condition fails, model results can change due to
numerical dispersion, or by completely passing an element in
one time step and not using the hydraulic conductivity and
gradient (i.e., ground-water velocity) associated with that
element.
The time steps for the model runs made in this study were
calculated based on the regional gradients assigned to each
aquifer case. The problem with this approach was that the
gradient changes near the well as a result of the pumping, and
therefore, the time step near the well would not be correct.
To determine if the time step selected could significantly
alter model results, a series of sensitivity runs were made
247
-------
with a range of time steps for two different aquifer cases.
Six sensitivity runs were made with the Floridan worst case
(deep well, 1,0000 gpm pumping, no decay) where the time step
ranged from 5 days to 60 days. The optimum time step for this
case, based on the regional gradient, was 10 days. All other
parameters in these simulations were identical.
The results of these runs, as shown in Figure 4.43, show that
the peak concentration is not very sensitive to changes in
time step. The peak concentrations at the well changed by
only about 3% and 6%, when the time step was halved and
doubled, respectively. The trend is such that the peak
concentration decreases as the time step increases.
The Floridan worst case had less than 0.5 m (1.5 ft) of
drawdown at the well, therefore, the change in gradient, and
the resulting change in travel time, was quite small. In
order to better test the impact of time step on peak
concentration, a series of 5 sensitivity runs were made for
the Surficial worst case (shallow well, 200 gpm pumping, no
decay) where the drawdown at the well was about 3 m (10 ft).
The optimum time step for this case was 30 days and time steps
of 5 to 60 days were tested in the sensitivity runs. All
other parameters in these simulations were identical.
The results of these simulations, also in Figure 4.43, show
that, as in the earlier case, the peak concentration is not
very sensitive to changes in the model time step. The peak
concentration at the well decreased as the time step
increased, but the rate of change was small.
Both the Floridan worst case and the surficial worst case
showed a low degree of sensitivity to change in time step, and
the percentage change in peak concentration at the well was
248
-------
to
4*
V£>
D)
13
4-
<
cc
LU
CJ
Z
S
UJ
Q.
3-
Legend
A FLOHIDAN
X SURFICIAL
10 20 30 40 50
TIME STEP, Days
60
Figure 4.43
Relationship of model time step to peak concentrations
for the Floridan and surficial worst cases.
-------
about the same for both cases. Part of the reason that time
step is not a significant fact in these simulations is that
the hydraulic conductivity is uniform within each aquifer.
Even if the distance traveled by the ground water and
contaminant is greater than one element spacing, it will still
be using the same hydraulic conductivity to calculate the
velocity. Therefore, running all cases with a single time
step based on the regional gradient, and using a uniform grid
spacing over the entire aquifer, does not impact the validity
of the model results.
4.2.6.2 Input Concentration
A sensitivity run was made to show that a direct one-to-one
relationship exists between concentration at the source and
peak concentration at the well. This relationship is
important because all the saturated model runs were made with
a unit loading at the field, and the predicted peak
concentrations at the well must be scaled to the field
loadings as predicted by the unsaturated zone analysis.
A single run was made using the Floridan worst case (deep
well, 1,000 gpm pumping rate, no decay) where the
concentration at the field was doubled every time step, and
the peak concentration at the well, were exactly twice the
values obtained from the original Floridan worst case run.
Because this direct relationship holds, peak concentrations at
the well can be scaled directly to the concentrations
predicted by the unsaturated zone model.
4.2.6.3 Transient Versus Steady State Comparison
4.2.6.3.1 Ground-water PotentialFor the purposes of this
study recharge was assumed to be evenly distributed throughout
the year, and therefore, the ground-water potential was
modeled as steady state. In actuality, rainfall and
250
-------
irrigation occur as events and are seasonal, which results in
a nonuniform distribution of water to the ground surface.
These pulses of water at the surface are smoothed as they
travel through the unsaturated zone, resulting in a fairly
uniform distribution of recharge to the water table, however,
occasional pulses do occur. If it can be shown that the
impact of these pulses is short lived (i.e., the resulting
ground-water mounds rapidly decay) then the steady state
assumption should be valid.
Two model simulations were made using the Floridan average
case (deep well, 1,000 gpm pumping rate, no decay) to
determine the time required for a ground-water mound to decay.
In both simulations the hydraulic conductivity was reduced to
152 m/day (500 ft/day) to increase the impact of the recharge.
The first simulation consisted of running the steady state
model with 33 cm (13 in) of recharge distributed uniformly
throughout the year. The second simulation consisted of
running the transient model with 33 cm of recharge evenly
distributed throughout the year, in addition to a 15 cm (6 in)
pulse on the first day. The 15 cm pulse is the maximum
probable daily pulse of water as predicted by the unsaturated
flow model over a 14 yr period. The transient model was run
until the predicted potential surface was identical to the
steady state surface.
The results of the transient simulation showed that it took 37
days for the potential surface to exactly equilibrate with the
surface predicted by the steady state model. After 10 days
the transient surface lost about 80% of its initial rise, and
95% was lost after 20 days.
The 15 cm pulse used in this simulation is an extreme maximum
that would be expected on a given day. The typical daily
251
-------
pulse, predicted by the unsaturated model, is less than one
tenth this amount, and the pulse rarely exceeds half this
amount. Given these facts, it is probably safe to assume that
the impact of most pulses will be negligible within a few
days.
Since the unsaturated zone has the effect of smoothing the
recharge distribution, and most pulses of recharge equilibrate
within a few days, the steady state assumption should be
valid.
*
4.2.6.3.2 Contaminant TransportSimulation with the
unsaturated model showed that about 90% of the aldicarb enters
the ground water within the first 4 months of each year, and
that the highest annual loading rate is about 1 x 10 gm/cm .
As a result of these findings, this annual loading rate was
used in all saturated flow model simulations, and it was
distributed uniformly over four months of each year.
To test the validity of this assumption, a simulation was made
where the actual annual loading rates as predicted by the
unsaturated model were simulated in the saturated flow model.
A two-year simulation was made using the 1970 and 1971 output
from the unsaturated flow model in the Floridan worst case
(deep well, 1,000 gpm pumping rate, no decay). The model was
run with a 5 day time step, therefore, contaminant loadings
were simulated as the total loading over each five day
interval. These results were compared to a model run where
the 1970 and 1971 annual loading rates were simulated as
uniformly entering the ground water over four months of each
year.
The results of these simulations are shown in Table 4.15. For
both years, the results (both relative peak concentration at
the well and time to peak) for the average loading simulation
252
-------
TABLE 4.15 RESULTS OF AVF.RAGE VERSUS ACTUAL
ALDICARB MASS LOADING SIMULATIONS
Relative Peak Concentration Time to Peak (days)
AverageActualAverage Actual
Year Loading Loading Loading Loading
1970 9.0 x 10"3 9.8 x 10"3 150 135
1971 8.9 x 10"3 5.4 x 10"3 150 220
253
-------
are close to those for the actual loading simulation,
indicating that the averaging assumption should be valid.
4.3 COMBINED RESULTS OF UNSATURATED AND SATURATED ZONE
MODELING
4.3.1 Physical Overlap of Saturated and Unsaturated Zones
The results from the unsaturated zone modeling and the
saturated zone modeling efforts were combined to estimate
aldicarb concentrations in drinking water wells. The
unsaturated zone scenarios were reduced to two physical areas:
1) areas in which entisols and ultisols are prevalent, and 2)
areas in which spodosols and alfisols are prevalent. For the
entisols and ultisols both a thin and a thick unsaturated zone
were considered but the physical area is the same (Figure
4.44). In the saturated zone the five aquifer systems were
combined into three physical aquifer systems that were
modeled: 1) the unconfined Floridan Aquifer, 2) the
unconfined surficial aquifer and 3) the multi-aquifer system.
These are shown in Figure 4.44. Figure 4.45 shows the areas
of overlap of the unsaturated and saturated zone scenarios.
The entisols and ultisols overlap all three of the aquifer
systems. The spodosols and alfisols overlap with the
surficial aquifer and the multiaquifer systems but not with
the unconfined Floridan Aquifer. The results of the two
modeling studies were combined in all cases where overlap
occured.
254
-------
Atlantic Octal
ENTISOLS AND ULTISOLS
SPODOSOLS AND ALFISOLS
Atlantic Ocean
LEGEND
FLORIDIAN AQUIFER - UNCONFINED
MULTI - AQUIFER SYSTEM
SURFICIAL AQUIFER - UNCONFINED
./» *
Figure 4.44 Unsaturated and saturated zone groupings.
255
-------
ALABAMA
Gulf of Mexico
Jacksonville
Atlantic Ocean
LEGEND
ENTISOLS AND ULTISOLS
FLORIDIAN AQUIFER - UNCONFINED
MULTI - AQUIFER SYSTEM
SURFICIAL AQUIFER - UNCONFINED
Figure 4.45 Overlap of unsaturated and saturated model
areas.
256
-------
4.3.2 Method of Combining the Modeling Results
Frequency distributions of pesticide loading from the
unsaturated zone and the relative pesticide concentrations in
the well water were developed from model results. Frequency
distributions for relative well-water concentrations are
located in Appendix E. A statistical approach was used so
that the risks of exposure could be accurately assessed.
Because the extreme values are of greatest interest in this
study, the ten percent exceedance value was used along with
the 50 percent exceedence value for combining the modeling
results. The unsaturated zone values were shown previously in
Table 4.5. The 10 and 50 percent exceedance values for the
saturated zone scenarios are shown in Appendix E.
The well concentrations were calculated using the following
relationship:
c = f c n (4'7)
w rel
where C = well water concentration
C r 1= relative well water concentration
f = a conversion factor equal to the
input concentration x unsaturated load
divided by the unit load
The concentration in the well is determined from the relative
concentration and a conversion factor which is the product of
the input concentration and the unsaturated load divided by
the unit load "used in the saturated zone simulations. The
input concentrations used in the saturated zone simulations
were shown in Table 4. 8 . The unit load used in saturated
zone simulations was always 1 kg/ha. Using equation 4.7, the
257
-------
resulting well water concentration for any pesticide load can
be calculated.
4.3.3 Results
The complete set of results showing the well water
concentration at the 10 and 50 percent exceedance values for
all the scenarios are shown in Table 4.16. These results are
summarized in Table 4.17 which reports the highest calculated
concentration for each catagory shown. In general the well
water concentrations are very low. The highest concentration
is 6.5 parts per billion (ppb) in a scenario without decay.
This is just at the detection limit of about 5 to 6 ppb for
total toxic residues (Rao, 1984, personal communication). The
highest concentrations for scenarios with decay are in the
order of 10" ppb.
The well water concentrations show three general trends. As
shown earlier, in the discussion of the unsaturated zone
results, the highest loads are leached from the thin ultisols
and entisols, followed by the thick ultisols and entisols and
the alfisols and spodosols. In the saturated zone, the worst
cases of the surficial aquifer generally have the highest well
water concentrations closely followed by the unconfined
Floridan worst cases and finally the worst cases of the
two-aquifer system and average Floridan cases. Within each
set of saturated zone scenarios there is a third trend that is
consistent for all the aquifers. The highest concentrations
result from the no decay simulation with the well 91 m (300
ft) from the source area, followed by no decay simulated with
a well at 300 m (1000 ft). Next are the decay scenarios with
the well at 91 m. The smallest concentrations result from the
decay simulations where the well is 300 m from the source.
258
-------
TABLE 4.16 WELL WATER CONCENTRATIONS (PPB) FOR COMBINED UNSATURATED AND SATURATED
SCENARIOS AT 50 AND 10 PERCENT EXCEEDANCE PROBABILITIES OF EACH.
to
-------
TABLE 4.16 (CONT.)
to
CTl
O
Saturated Zone
Surficial Aquifer, worst case
91 meters to well
High pumping rate, no decay 50%
High pumping rate, no decay 10%
Low pumping rate, no decay 50%
Low pumping rate, no decay 10%
Low pumping rate, decay 50%
Low pumping rate, decay 10%
300 meters to well
High pumping rate, no decay 50%
High pumping rate, no decay 10%
Low pumping rate, no decay 50%
Low pumping rate, no decay 10%
Low pumping rate, decay 50%
Low pumping rate, decay 10%
Two-aquifer System, worst case
91 meters to well
High pumping rate, no decay 50%
High pumping rate, no decay 10%
Low pumping rate, no decay 50%
Low pumping rate, no decay 10%
Low pumping rate, decay 50%
Low pumping rate, decay 10%
Alfisols and Spodosols
50% 10%
7.98x10-2
1.17x10-2
2.11x10-2
2.7x10-2
1.69x10-4
5.27x10-4
4.21x10-3
9.42x10-3
7.47x10-3
1.64x10-2
2.92x10-6
5.20x10-6
1.65x10-4
3.95x10-4
2.68x10-4
4.91x10-4
1.56x10-5
6.37x10-5
4.39x10-2
6.41x10-2
1.16x10-1
1.49x10-1
9.30x10-4
2.90x10-3
2.32x10-2
5.18x10-2
4.11x10-2
9.04x10-2
1.61x10-5
2.86x10-5
9.07x10-4
2.17x10-3
1.48x10-3
2.70x10-3
8.57x10-5
3.51x10-4
Unsaturated Zone
Thick Ultisols and Entisols
50% 10%
8.38x10-3
1.22x10-2
2.21x10-2
2.85x10-2
1.78x10-4
5.53x10-4
4.42x10-3
9.89x10-3
7.84x10-3
1.73x10-2
3.07x10-6
5.46x10-6
1.73x10-4
4.15x10-4
2.82x10-4
5.16x10-4
1.64x10-5
6.69x10-5
6.36x10-2
9.32x10-2
1.69x10-1
2.17x10-1
1.35x10-3
4.21x10-3
3.37x10-2
7.53x10-2
5.98x10-2
1.32x10-1
2.34x10-5
4.16x10-5
1.32x10-3
3.16x10-3
2.15x10-3
3.93x10-3
1.25x10-4
5.10x10-1
Thin Ultisols and Entisols
50% 10%
2.59x10-1
3.79x10-1
6.85x10-1
8.82x10-1
5.5x10-3
1.17x10-2
1.92
2.8
5.06
6.51
4.06x10-2
1.26x10-1
1.37x10-1
3.06x10-1
2.43x10-1
5.34x10-1
9.49x10-5
1.69x10-4
1.01
2.26
1.79
3.95
7.01x10-4
1.25x10-3
5.36x10-3
1.28x10-2
8.72x10-3
1.60x10-2
5.06x10-4
2.07x10-3
3.96x10-2
9.49x10-2
6.44x10-2
1.18x10-1
3.74x10-3
1.53x10-2
-------
TABLE 4.17 HIGHEST CALCULATED ALDICARB CONCENTRATIONS (in ppb)
IN THE GIVEN COMBINED UNSATURATED/SATURATED CATAGORIES
Floridan
Worst Cases
Floridan
Average
Surf icial
Worst Cases
Two-
Aquifer
Worst
Cases
300
ft
1000
ft
300
ft
1000
ft
300
ft
1000
ft
300
ft
No
Decay
Decay
No
Decay
Decay
No
Decay
Decay
No
Decay
Decay
No
Decay
Decay
No
Decay
Decay
No
Decay
Decay
Flatwoods
Soils
No
Overlap
No
Overlap
l.Sxlo'1
2.9xlO~3
9.0xlO~2
2.7x10-*
2.7xlO~3
3.5xlO~*
Thick
Ridge Soils
2.3xlO~2
4xlO~3
1.8xlO~2
9.8xlO~*
2.2xlO~2
1.3x10""
8.8xlO~3
1.4xlO"7
2.2xlO-1
4. 2xlO~ 3
1.3X10"1
4.2xlO~5
3.9xlO~3
5.1xlO~"
Thin
Ridge Soils
6.9x10'*
1.2X10"1
5.5x10"'
2.9xlO~2
6.5xlO~a
3.9xlO~3
2.6xlO-1
4. xlO~6
6.5
l.SxlO"1
3.9
1.3xlO~3
1.2X10'1
1.5xlO~2
261
-------
Surficial Aquifer
Combined with the thin entisols and ultisols, the surficial
aquifer worst cases generally produce the highest
concentrations. The highest value is for worst case hydraulic
parameters, no pesticide decay, with a shallow well at 91 m
distance from the source area, pumping at a low rate having a
well water concentration of 6.5 parts per billion (ppb). With
decay the concentrations drop one to three orders of
magnitude. When the well is 300 m from the source, the
surficial cases without decay still show significant
contamination, but when decay is simulated the concentrations
are in the order 10 ppb or less. The other unsaturated zone
scenarios show the same trends as the thin ridge soils when
combined with the surficial aquifer worst cases with even
lower concentrations. The highest concentrations are 2.17 x
-1 -1
10 and 1.49 x 10 for the thick ridge soils and the
flatwoods soils, respectively.
Floridan Aquifer Worst Cases
Well water concentrations in the Floridan worst cases are not
as high as the surficial worst cases except when decay is
modeled with the drinking water well at 300 m. When
evaluating these results with no decay, it is necessary to
recall that these concentrations could be much higher for
wider source areas. The highest concentration in the
scenarios without decay is 6.9 x 10 ppb with the well at 91
m (300 ft) and 5.5 x 10~1ppb with the well at 300 m (1000
ft). The scenarios simulated with decay are generally one
order of magnitude less than the same scenario without decay.
The Floridan Aquifer Average Cases
The well water concentrations in the Floridan average cases
vary considerably between scenarios modeled with decay and
without decay. The concentrations calculated for scenarios
262
-------
with the well 91 m (300 ft) away from the source area and no
decay are as high as 0.65 ppb. In the cases with decay the
concentrations drop over two orders of magnitude when the well
is 91 m from the source and four to five orders of magnitude
when the well is 300 m from source.
Two-Aquifer System
The scenarios modeled with the two-aquifer system have some of
the lowest concentrations. Recall that the concentrations
were suspected of being higher than they should be due to the
hydraulics of the system. Even so the concentrations are low.
The highest value is 0.12 ppb. With decay simulated the
concentrations drop over an order of magnitude.
263
-------
REFERENCES
Anderson, C.A. 1981. Fertility and Management for Citrus on
Entisols. In: Proceedings of Symposium on Fertility
Management of Florida Soils. Soil and Crop Science
Society of Florida, Vol 40.
Anderson, W. and E.P. Simonds. 1982. Hydrology of Hamilton
Lakes and Vicinity, Polk County, Florida. U.S. Geological
Survey Water-Resources Investigations Open-File Report
82-630.
Bearden, H.W. 1972. Water Available in Canals and Shallow
Sediments in St. Lucie County, Florida. Florida Bureau of
Geology Report of Investigation No. 62.
Biggs, R.H. and P.G. Webb. 1983. Fate of Aldicarb in Plants
and the Nature of Citrus Responses. In: Aldicarb
Research Task Force Report. IFAS, University of Florida,
Gainesville.
Brady, N.C. 1974. The Nature and Properties of Soils.
McMillan Publishing Co., Inc. 8th Edition.
Bromilow, R.H., R.J. Baker, M.A.H. Freeman and K. Gorog.
1980. The Degradation of Aldicarb and Oxamyl in Soil.
Pesticide Sci. 11: 371-378.
264
-------
Bull, D.L., R.A. Stokes, J.R. Coppedge and R.L. Ridgway.
1970. Further Studies of the Fate of Aldicarb in Soil.
J. of Economic Entomology. 63(4):1283-1289.
Buono, A., R.M. Spechler, G.L. Barr and R.M. Wolsansky. 1979.
Generalized Thickness of the Confining Bed Overlying the
Floridan Aquifer, Southwest Florida Water Management
District. U.S. Geological Survey Water-Resources
Investigations Open-File Report 79-1171.
Bush, P.W. 1982. Predevelopment Flow in the Tertiary
Limestone Aquifer, Southeastern United States; A Regional
Analysis from Digital Modeling. U.S. Geological Survey
Water-Resouces Investigation 82-905.
Caldwell, R.E. and R.W. Johnson. 1982. General Soil Map of
Florida. United States Dept. of Agriculture, Soil
Conservation Service.
Calhoun, F.G., V.W. Carlisle, R.E. Caldwell, L.W. Zelazny,
L.C. Hammond and H.L. Breland. 1974. Characterization
Data for Selected Florida Soils. Soil Science Research
Report No. 74-1, University of Florida, IFAS.
Calvert, D.V., R.C.J. Koo and H.W. Ford. 1967. Flood
Irrigation Studies with Citrus. Proc. Fla. State Hort.
Soc. 80:79-85
Carlisle, V.W., R.E. Caldwell, F. Sodek, III, L.C. Hammond,
F.G., Calhoun, M.A. Granger and H.C. Breland. 1978.
Characterization Data for Florida Soils. Soil Science
Research Report No. 78-1, University of Florida, IFAS.
265
-------
Carlisle, V.W., C.T. Hallmark, F. Sodek III, R.E. Caldwell,
L.C. Hammon and V.E. Berkheiser. 1981. Characterization
Data for Selected Florida Soils. Soil Science Research
Report No. 81-1, University of Florida, IFAS.
Carsel, R.F., C.N. Smith, L.A. Mulkey, J.D. Dean and P.P.
Jowise. 1984. User's Manual for the Pesticide Root Zone
Model (PRZM) Release I. U.S. Environmental Protection
Agency, Athens, GA.
Causey, L.V. and G.W. Leve. 1976. Thickness of the Potable-
Water Zone in the Floridan Aquifer. Florida Bureau of
Geology Map Series 74.
Cherry, R.N., J.W. Stewart and J.A. Mann. 1970. General
Hydrogeology of the Middle Gulf Area, Florida. Florida
Bureau of Geology Report of Investigation, No. 56.
Chow, V.T. 1964. Handbook of Applied Hydrology. McGraw-Hill
Book Co. New York, NY.
Cohen, S.Z., S.M. Creeger, R.F. Carsel and C.G. Enfield.
1983. Potential for Pesticide Contamination of Ground
Water Resulting from Agricultural Uses.
Cooke, C.W. 1945. Geology of Florida: Florida Geological
Survey Bull. 29.
Grain, L.J., G.H. Hughes and L.J. Shell. 1975. Water
Resources of Indian River County, Florida. Florida Bureau
of Geology Report of Investigation, No. 80.
266
-------
Duerr, A.D. and J.T. Tromraer. 1982. The Benchmark Farm
Program: A Method for Estimating Water Use in Southwest
Florida. USGS Water-Resource Investigation 82-17.
U.S. Geological Survey, Tallahassee, FL.
Fayard, L.D., D.M. Bonnet and R.G. Belles. 1983. Water
Resources Data, Florida, Water Year 1982, Vol. IB,
Northeast Florida Groundwater. U.S. Geological Survey
Water-Data Report FL-82-1B.
Florida Citrus Mutual. 1984. Lakeland, FL.
Florida Dept. of Agriculture and Consumer Services. 1984.
Citrus Summary 1983. Orlando, FL.
Gupta, S.K., C.T. Kincaid, P.R. Meyer, C.A. Newbill and C.R.
Cole. 1982. A Multi-Dimensional Finite Element Code for
the Analysis of Coupled Fluid, Energy and Solute Transport
(CFEST). PNL-4260, Pacific Northwest Laboratory,
Richland, WA. Prepared for the U.S. Dept. of Energy.
Hansen, J.L. and M.H. Spiegel. 1983. Hydrolysis Studies of
Aldicarb, Aldicarb Sulfoxide and Aldicarb Sulfone. Env.
Tox. and Chem. 2:147-153.
Harrison, D.S. 1984. Citrus Water Use and
Evapotranspiration (ET) Requirements for Florida
Flatwoods Soils. Florida Cooperative Extension Service
Extension Report 84-1.
Harrison, D.S., J.F. Gerbeer and R.E. Choate. 1974.
Sprinkler Irrigation for Cold Protection. Technical
Circular 348. Florida Cooperative Extension Service.
267
-------
Healy, H.G. 1974. Water Wells in the Artesian and
Nonartesian Aquifers of Florida, 1971-1972. USGS, Bureau
of Geology. Information Circular No. 85.
Healy, H.G. 1975. Piezometric Surface and Areas of Artesian
Flow of the Floridan Aquifer in Florida. Florida Bureau
of Geology Map Series No. 4, second edition.
Healy, H.G. 1982. Surficial and Intermediate Aquifers in
Principal Aquifers in Florida, B.J. Franks, ed. U.S.
Geological Survey Water-Resources Investigations Open-
File Report 82-255.
Hem, J.D. 1970. Study and Interpretation of the Chemical
Characteristics of Natural Water. U.S. Geological Survey
Water Supply Paper 1473.
Hornsby, A.G., P.S.C. Rao, W.B. Wheeler, P. Nkedi-Kizza and
R.L. Jones. 1983. In Press. Fate of Aldicarb in Florida
Citrus Soils 1. Field and Laboratory Studies.
Proceedings of the NWWA Symposium on Characterization and
Monitoring of the Unsaturated (Vadose) Zone. D. Nielsen,
ed.
Hough, A., I.J. Thomason and W.J. Farmer. 1975. Behavior of
Aldicarb in Soil Relative to Control of Heterodera
Schachtii. J. of Nematology. 7(3):214-221.
Hughes, G.H. 1977. Runoff from Hydrologic Units in Florida.
Florida Bureau of Geology Map Series No. 81.
Hutchinson, C.B. 1978. Appraisal of Shallow Ground-Water
Resources and Management Alternatives in the Upper Peace
and Eastern Alafia River Basins, Florida. U.S. Geological
Survey, Water-Resources Investigations, pp. 77-124.
268
-------
INTERA Environmental Consultants. 1980. Mathematical
Simulation of Aldicarb Behavior on Long Island:
Unsaturated Flow and Ground-Water Transport. Prepared for
the Hayzard Evaluation Division, U.S. Environmental
Protection Agency.
Jones, R.L., P.S.C. Rao and A.G. Hornsby. 1984. In Press.
Fate of Aldicarb in Florida Citrus Soil 2. Model
Evaluation. Proceedings of the NWWA Synpvision on
-j.
Characterization and Citation Monitoring of the
Unsaturated (Vadose) Zone. D. Nielsen, Ed.
Klein, H., M.C. Schroeder, W.F. Lichtler. 1964. Geology and
Ground-Water Resources of Glades and Hendry Counties,
Florida. Florida Geological Survey Report of
Investigation, No. 37.
Knochenmus, D.D. and G.H. Hughes. 1976. Hydrology of Lake
County, Florida. U.S. Geological Survey Water-Resources
Investigations, 76-72.
Kohler, M.A., T.J. Nordenson and D.R. Baker. 1959.
Evaporation Maps for the United States. U.S. Weather
Bureau Tech. Paper 37.
Land, L.F., H.G. Rodis and J.J. Schneider. 1973. Appraisal
of the Water Resources of Eastern Palm Beach County,
Florida. Florida Bureau of Geology Report of
Investigations, No. 67.
Leistra, M., T. Lexmond and J.H. Smelt. 1976. Conversion and
Leaching of Aldicarb in Soil Columns. Pesticide Sci.
7(5): 471.
269
-------
Lemley, A.T. and W.Z. Zhong. 1983. Kinetics of Aqueous Base
and Acid Hydrolysis of Aldicarb, Aldicarb Sulfoxide, and
Aldicarb Sulfone. Env. Tox and Chem. 2:147-153.
Leraley, A.T. and W.Z. Zhong. 1984. Hydrolysis of Aldicarb,
Aldicarb Sulfoxide and Aldicarb Sulfone at ppb Levels in
Aqueous Media. J. Ag. and Food Chem, July-August, 1984.
Lichtler, W.F. 1960. Geology and Ground-Water Resources of
Martin County, Florida. Florida Geological Survey Report
of Investigation, No. 23.
Lichtler, W.F. 1972. Appraisal of Water Resources in the
East Central Florida Region. Florida Bureau of Geology
Report of Investigation, No. 61.
Lyman, W.J. 1982. Adsorption Coefficient for Soils and
Sediments. Chapter 4 In: Handbook of Chemical Property
Estimation Methods. W.J. Lyman, W.F. Rheel and D.H.
Rosenblatt, eds. McGraw-Hill Book Co., New York, NY.
Merritt, M.L., F.W. Meyer, W.H. Sonntag and D.J. Fitzpatrick.
1983. Subsurface Storage of Freshwater in South Florida:
A Prospectus. U.S. Geological Survey Water-Resources
Investigation, 83-4214.
Myeyk, R.T., L.D. Fayard, W.L. Fletcher and J.K. Ogle. 1983.
Water Resources Data: Florida, Water Year 1982, Vol. 3B.
Southwest Florida Ground Water. U.S. Geological Survey
Water-Data Report FL-82-3B.
Nicholls, P.H., R.H. Bromilow and T.M. Addiscott. 1982.
Measured and Simulated Behavior of Fluometuron, Aldoxycarb
and Chloride Ion in a Fallow Structured Soil. Pestic.
Sci. 13: 475-483.
270
-------
Pacenka, S. and K.S. Porter. 1981. Preliminary Regional
Assessment of the Environmental Fate of the Potato
Pesticide, Aldicarb, Eastern Long Island, New York.
Cornell University, Center for Environmental Research.
Parker, G.G., G.E. Ferguson and S.K. Love. 1955. Water
Resources of Southeastern Florida, with Special
Reference to the Geology and Ground Water of Miami Area:
U.S. Geological Survey Water-Supply Paper 1255.
Phelps, G.G. 1978. Chemical Quality of Water used for
Municipal Supply in Florida, 1975. U.S. Geological Survey
Map Series, No. 82.
Porter, K.S., A.T. Lemley, H.B. Hughes and R.L. Jones. 1984.
Developing Information on Aldicarb Levels in Long Island
Groundwater. Second International Conference on Ground-
Water Quality Research, Proceedings. March 1984. Ada,
OK.
Rao, P.S.C., L.T. Ou, K.S.V. Edvardsson, J.T. Thomas and W.B.
Wheeler. 1984. Degradation and Sorption of Aldicarb and
Metolachlor in Dougherty Plains Soils. Progress Report
dated June 15, 1984. For the U.S. Environmental
Protection Agency, Athens, GA.
Ryder, P.D. 1982. Digital Model of Predevelopment Flow in
the Tertiary Limestone (Floridan) Aquifer System in
West-Central Florida. U.S. Geological Survey Water-
Resources Investigations 81-54.
Scott, W.B. 1977. Hydraulic Conductivities and Water Quality
of the Shallow Aquifer, Palm Beach County, Florida. U.S.
Geological Survey Water-Resources Investigation 76-119.
271
-------
Sinclair, W.C. 1974. Hydrogeologic Characteristics of the
Surficial Aquifer in Northwest Hillsborough County,
Florida. Florida Bureau of Geology Information. Circular
No. 86.
Smajstrla, A.G., D.S. Harrison, C. Tai and D. Clapp. 1982.
Water Budget of Crown Flood Irrigated Citrus.
Proceedings of the Florida State Horticultural Society,
Vol. 95.
Smelt, J.H., A. Dekker, M. Leistra and N.W.H. Houx. 1983.
Conversion of Four Carbamoyloximes in Soil Samples from
Above and Below the Water Table. Pesticide Science 14:
173-181.
Smelt, J.H., M. Leistra, N.W.H. Houx and A. Dekker. 1978a.
Conversion Rates of Aldicarb and Its Oxidation Products in
Soils. I. Aldicarb Sulfone. Pesticide Sci. 9:279-285.
Smelt, J.H., M. Leistra, N.W.H. Houx and A. Dekker. 1978b.
Conversion Rates of Aldicarb and Its Oxidation Products
in Soils. II. Aldicarb Sulfoxide. Pesticide Sci.
9:286-292.
Smelt, J.H., M. Leistra, N.W.H. Houx and A. Dekker. 1978c.
Conversion Rates of Aldicarb and Its Oxidation Products in
Soils. III. Aldicarb. Pesticide Sci. 9:293-300.
Stanley, J.M., C. Taylor, W.R. Sununerhill, Jr. and L.J.
Beaulieu. 1980. Citrus Energy Survey - Use Estimates
and Conservation. Institute of Food and Agricultural
Sciences. University of Florida, Energy Report No. 2.
272
-------
Stewart, H.G. 1966. Ground-Water Resources of Polk County.
Florida Geological Survey, Report of Investigation, No.
44.
Stewart, J.W. 1980. Areas of Natural Recharge to the
Floridan Aquifer in Florida. Florida Bureau of Geology
Map Series 98.
Stringfield, V.T. 1966. Artesian Water in Tertiary Limestone
in the Southeastern States. U.S. Geological Survey
Professional Paper 517.
Supak, J.R. 1972. The Volatilization, Degradation,
Adsorption and Desorption Characteristics of Aldicarb in
Soils and Clays. Ph.D Dissertation, Texas A&M University.
Supak, J.R., A.R. Swoboda and J.B. Dixon. 1977.
Volatilization and Degradation Losses of Aldicarb from
Soils. J. of Environ. Qual. 6(4):413-417.
Thibodeaux, L.J. 1979. Chemodynamics. John Wiley and Sons.
New York, NY.
Tibbals, C.H. 1981. Computer Simulation of the Steady-State
Flow System of the Tertiary Limestone (Floridan) Aquifer
System in East-Central Florida. U.S. Geological Survey
Water-Resources Investigations Open-File Report 81-681.
U.S. Dept. of Commerce. 1968. Climatic Atlas of the United
States.
U.S. Dept. of Commerce. 1972. Climate of the States,
Florida: Climatography of the United States, No. 60-8,
National Oceanic and Atmospheric Adm. Environmental Data
Service.
273
-------
U.S. Geological Survey. 1982. Water Resources Data:
Florida, Water Year 1981. U.S. Geological Survey
Water-Data Report, Vol. 1, 2B and 3B.
Viessman, Jr., W., J.W. Knapp, G.L. Lewis and T.E. Harbaugh.
1977. Introduction to Hydrology. 2nd edition. Harper &
Row, NY.
Visher, F.N. and G.H. Hughes. 1969. The Difference Between
Rainfall and Potential Evaporation in Florida. Florida
Bureau of Geology Map Series 32.
Wedderburn, L.A., M.S. Knapp, D.P. Waltz and W.S. Burns.
1982. Hydrogeologic Reconnaissance of Lee County,
Florida, Part 1-Text. Technical Publication 82-1. South
Florida Water Management District.
Wilson, W.E. 1977. Ground-Water Resources of DeSoto and
Hardee Counties, Florida. Florida Bureau of Geology
Report of Investigation 83.
Wilson, W.E. and J.M. Gerhert. 1982. Simulated Effects of
Ground-Water Development on the Potentiometric Surface of
the Floridan Aquifer, West-Central Florida. U.S.
Geological Survey Professional Paper 1217.
Wolansky, R.M. 1978. Feasibility of Water Supply Development
from the Unconfined Aquifer in Charolotte County, Florida.
U.S. Geological Survey Water-Resources Investigation
78-26.
Wolansky, R.W., L.R. Mills and W.M. Woodman. 1978. Water
Table in Surficial Aquifer, West Central Florida. U.S.
Geological Survey Open File Report No. 78-1045.
274
-------
APPENDIX A
PRZM MODIFICATIONS
One of the tasks required by the Florida Teraik study was to
make modifications to PRZM to cover special situations arising
from the Florida hydrologic situation and the use of aldicarb.
PRZM currently is a one-dimensional hydrologic and transport
model for pesticides in the unsaturated zone. Aldicarb forms
toxic daughter products which may have different environmental
and toxicological properties from the parent. Therefore,
algorithms to handle this situation were added. Because of
the variety of irrigation management practices used on Florida
citrus and their potential impact on transport a set of
algorithms were implemented to handle these circumstances.
PRZM currently allows for restricted vertical drainage but
does not make allowance for lateral flow under scenarios of
temporary saturated conditions caused by restricted vertical
drainage. In Florida flatwoods soils, the practice of bedding
citrus necessitates such an addition. Such an algorithm was
implemented.
In depth discussions of the algorithms are given in the
following sections.
275
-------
A.I ADDITION OF PARENT/DAUGHTER CONTAMINANT RELATIONSHIPS
The fate of pesticides in soils is a complex issue. There are
many processes (i.e., volatilization, degradation, etc.) which
must be considered in order to adequately address this issue.
One of these processes, which has been largely neglected in
pesticide leaching models, is that of the transformation of
the parent compound to various toxic daughter products. The
tendency has been to lump all the toxic family into a "total
toxic residue" and model the fate of this composite as a
single chemical. This assumption may not be acceptable,
especially if the daughters have very different decay rates or
adsorption partition coefficients from the parent or from each
other.
Aldicarb is oxidized to a sulfoxide and a sulfone as part of
the transformation/decay process. Algorithms were put in PRZM
to simulate parent/daughter relationships. An analytical
solution to the decay and transformation equation was derived
to check the numerical model.
The system which was modeled is shown in Figure A.I. The GJ
* ±
are dissolved concentrations and the C ^ are adsorbed
concentrations. The K. are adsorption partition coefficients,
the k . are decay and transformation rates in the dissolved
D *
places and the k^ are adsorbed phase decay coefficients.
Notice that only the dissolved forms nay be transformed from
one toxic form to another. A system of first order
differential equations describing this system can be written
as:
276
-------
C2* C3* ADSORBED PHASE
4
C2 _3_^ C3 DISSOLVED PHASE
T k, T
1 ' k3
Figure A.I Schematic of a system of parent
and daughter pesticides.
277
-------
(A.2)
dt
.3)
*
d C1P
-at
d C *P * *
dt
(A.6)
Making use of C^ Ki = ci the s^x equations above can be
reduced to three equations in three unknowns, namely:
d C. (A.7)
d C0
(A.8)
278
-------
in which
d C3 _ - (A.9)
dt ~ a4 C2 * "5 C3
(A-
al
(A.11)
k 0 (A.13)
a4 = 4
(A.14)
Uf^S ~ ~3 "3P
S5
These ordinary differential equations with constant
coefficients can be solved analytically for C^, 2 and 3
using the initial conditions, C^ = C'-^ when t = 0 and
C-= C-.= 0 at t = 0. The solutions are:
279
-------
C. = C.'
(A. 15)
and
' e
E^ a4
a3 ' 35)
- a^ (a^ - ag)
mm
(A.17)
In PRZM, the equations are solved numerically as part of the
general advection-dispersion equation for a solute in a porous
medium by using a backward difference implicit scheme. A new
subroutine, PSTLNK, was added to set up the transformation
(source and sink) terms for the system. In the case of
aldicarb, we have the relationship C-^-»-C2^C3, but the system
can be configured for ^\^<^ an<^ '"l^^B or
C _ and C ,, simply by selecting zero or positive values for the
appropriate transformation rate constants.
Figures A. 2 through A. 4 show the results of a series of tests
performed on the numerical model and checked by the analytical
model. In these figures, the solid line represents the "true"
280
-------
to
CO
100
80
60
40
20
NUMERICAL
ANALYTICAL
8 10 12
TIME, IN DAYS
14
16
18
20
Figure A. 2 Conversion of GI to C2 to C3 with no adsorption and no decay.
-------
100
M
00
N)
EH
2
W
w
04
2
O
M
fi
I
2
(0
§
u
1 I I I
NUMERICAL
ANALYTICAL
8 10 12
TIME, IN DAYS
Figure A. 3 Conversion of C, to C, to C3 with decay but no adsorption.
-------
100
00
H
Z
U
U
cu
U
U
z
NUMERICAL
ANALYTICAL
20 -
10
TIME, IN DAYS
Figure A. 4 Conversion of C-^ to C- to C-j with decay and adsorption,
-------
or analytical solution, while the dashed line represents the
approximate numerical solution. In Figure 2.26, there was no
decay of the dissolved phase chemicals and no adsorption of
any species. The rate of transformation from C^ to 2 was
0.2/day and that from 2 to 3, 0.5/day. After 20 days nearly
all the chemical is in form C-j. The numerical model traces
the decay and formation of each constituent closely, being
poorer in those regions where the rate of change of the
concentrations are more rapid. Figure A.3 shows the same
system with a decay rate of 0.01/day in the dissolved phase.
Figure A.4 shows the solution for the same system except that
the adsorption coefficients have been given values (Ki = 0.5,
K = 1.0 and K-. = 5.0) as have the adsorbed phase decay
^ * J
coefficients (k. = 0.01). Notice that the transformation rate
is retarded when adsorption is introduced.
Using the analytical model, the assumption of modeling the
"total toxic residue" decay as a first-order process was
tested. Adsorption coefficients for a Woburn sandy loam (K-, =
0.55, K =0.16 and K- = 0.185) and decay and transformation
rate constants (k = 0.07, k_ = 0.55, k3 = 0.01, k4 = 0.031
and k,_ = 0.0152) were taken from Bromilow et al., 1980. A
5 33
soil bulk density of 1.45, a water content of 0.27 cm /cm
and an initial aldicarb parent mass of 100 mg were also used.
The model was run for 90 days and the results are shown in
Figure A.5.
The results show that the decay of the sum of the dissolved
aldicarb concentrations does not follow first-order kinetics.
The reason for this is the conversion of aldicarb parent to
aldicarb sulfoxide. Because the sulfoxide has a lower
partition coefficient, the dissolved concentration increases
until most of this conversion is complete. Once this happens,
however, the sum of the sulfoxide and the sulfone
concentrations does follow a first-order decay curve. The
284
-------
120
oo
Cn
0 PARENT ALDICARB
A SULFOXIDE
V SULFONE
D TOTAL
40 50
TIME, IN DAYS
Figure A-5 Conversion of aldicarb to aldicarb sulfoxide to aldicarb sulfone,
-------
implication is that as long as the parent exists, the parent
and at least the sum of the daughters should be modeled
separately. Thus, in the unsaturated zone, parent/daughter
relationships should be used. However, if little of the
parent reaches the saturated zone, the sum of aldicarb
sulfoxide and sulfone can be modeled as a single solute
following first-order decay.
A.2 ADDITION OF IRRIGATION ALGORITHMS
A new subroutine, IRIGAT, was added to the PRZM code in order
to simulate the application of irrigation water.
There are three things which are crucial to the application of
water; the time that the event is triggered, the total water
applied, and the rate at which it is applied which determines
the length of the irrigation event.
The subroutine was designed to trigger irrigation water
application when the average water content in the root zone is
at a user specified percentage of the available water in the
root zone (the water content at field capacity (9 £ J minus the
water content at wilting point (9 )). The total water
applied at that time, is calculated by:
fc. -. 0 ±)A«. V-18)
This is the water required to bring the root zone profile back
to field capacity. Different types of irrigation methods
(e.g., overhead sprinklers and trickle) apply water at
different rates. The algorithms allow the user to specify
286
-------
that rate. The length of time over which the water is applied
is then calculated by dividing the total water applied by the
application rate. This is rounded to the nearest whole day
because PRZM operates on a daily timestep. The application
rate is adjusted accordingly so that the same total water
quantity is applied.
A.3 ADDITION OF LATERAL DRAINAGE ALGORITHMS
In a soil profile when drainage is restricted there is always
the possibility of lateral flow occurring if perched water
tables (i.e., saturated soil conditions) occur. In PRZM,
drainage is restricted by using a time constant which modifies
percolation from the soil layer. The equation is (Carsel
et al., 1984) :
-
__=- L fc (A. 19)
In this expression, 9 is the dimensionless hydraulic head
acting on the soil layer (cm water/cm soil) and 8 fc is the
field capacity water content. It has the solution:
*
-------
e - efj
.dt rc = - a + i ( e - 0) (A.21)
In this equation, K? is the time constant for lateral drainage
and is only positive if 6 > 9 where 9 is the saturation water
s s
content in the layer. Otherwise, K~ = 0- The equation has a
solution similar to (2.21), specifically:
Equation A.22 is used if 9 >9 and if this condition is not
o s
met over the entire time step, the algorithm computes the time
't1 when the change occurs and switches to (A.20) for the
remainder of the time step. This is, obviously, a very simple
analytical model with empirical coefficients representing a
rather complex physical process. The concept is that of a
tank with a hole in it and another hole in the side very near
the bottom. The quantity of water flowing out depends only on
the depth of water in the tank ( 9 - 9 ^ ) and the
characteristics of the holes (K^ and K2)
Equation (A.22) gives the value of at any time 't1. To
determine the quantities of water draining laterally and
percolating, it is necessary to derive expressions for these
quantities. The instantaneous rate of water percolating is:
(A.23)
288
-------
where Az is the thickness of the soil layer. One can
substitute the solution for 8 (A.22) into (A.23) and integrate
over time to obtain the total percolation (Q-^) over time 't',
thus:
- (K, + K,) t
Q, = IAS K, (e *. + (e n - e ff,> e A 2. > at
fc T l * o * fc'
o
This integral has the solution:
_ -'0^+ ^t
IT
. **Kl(0 -0 fc) _
gl ~
Similarly, for the lateral flow:
^ o ^ fe
Oj /T-r i »» \ I A ^
These equations are subject to the initial conditions Q-^ = Q2
= 0, v/hen t = 0. The time constant for lateral drainage, Kj /
can be estimated from available field data.
Although this is a simplistic representation of lateral
drainage it can be used to bracket the two-dimensional problem
of lateral drainage and percolation with adsorption and decay,
since one can either allow the lateral flow to exit the soil
289
-------
column at zero pesticide concentration or at the dissolved
concentrations in the soil column.
290
-------
APPENDIX B
Plots of soil order physical properties from soil
characterization analysis
291
-------
VO
KJ
y-s
40. O
30.0-
LU
S 25. 0+
LU
20. 0--
5. 0--
0.0
1 1 1 1
H 1 1 1 1 1 1 1-
H 1 1-
0
50
100 150
DEPTH (CM)
200
250
Figure B.I Mean and standard deviation of field capacity water content vs.
depth for Astatula and Candler entisols.
-------
to
vo
y+s
y-s
40.0-
35. 0;:
30.0-
LU
S 25. 0+
Dd
LU
Q.
20.0-
15.0-
Q 10.0"
_j
LLl
II
U_
5. 0-f"
0.0
H 1 1
H 1 1 1 1 1 1 1 1 I 1 1 1 1 > I
0
50
100 150
DEPTH (CM)
200
250
Figure B.2 Mean and standard deviation of field capacity water content vs. depth
for St. Lucie and Paola entisols.
-------
to
vo
y+S
y-s
40.0-
35-°:
30.0-
LU
S 25. 0+
o:
UJ
g 20. 0--
~ 15.0-
Q 10.0"
Q.
LU
5. 0--
0.0
/ v
N.
0
-t 1 1 1 1 1 1 I 1 1 1 I 1
I I 1 1 I 1 1
50
100
150
200
250
DEPTH (CM)
Figure B.3 Mean and standard deviation of field capacity water content vs.
depth for Apopka and Arredondo ultisols.
-------
to
vo
Ul
y+S
y-s
40. 0T
50
100
150
200
250
DEPTH (CM)
Figure B.4 Mean and standard deviation of field capacity water content vs.
depth for Iiranokalee, Oldsmar, Wabasso and Myakka spodosols.
-------
y+S
K)
vo
en
y-s
40. 0T
50
100 150
DEPTH (CM)
200
250
Figure B.5 Mean and standard deviation of field capacity water content vs. depth
for Felda, Riviera and Pineda alfisols.
-------
y+s
4. Or
3.0-
y-s
vo
O
DQ
o:
u
2. 0--
1.0-
0.0-1
0
Figure B.6
50
100 150
DEPTH (CM)
200
250
Mean and standard deviation of organic carbon vs. depth for Astatula
and Candler entisols.
-------
y+s
y-s
to
vo
00
4. 0T
3.0-
a
CO
a:
2. 0--
0
100
150
200
250
DEPTH (CM)
Figure B.7 Mean and standard deviation of organic carbon vs. depth for St. Lucie
and Paola entisols.
-------
y+S
y-s
to
vo
vo
4.0T
0
DEPTH (CM)
250
Figure B.8 Mean and standard deviation of organic carbon vs. depth for Apopka
and Arredondo ultisols.
-------
o
o
y+s
y-s
4. 0T
0
50
100
150
200
250
DEPTH (CM)
Figure B.9 Mean and standard deviation of organic carbon vs. depth for
Immakolee, Oldsmar, Wabasso, and Myakka spodosols.
-------
y+S
4. OT
.. \
\
\
3.0+ I
CD
m
ct
u 2.0+
o
CD
a:
o
1.0--
£-H nJ=q , p=H , 1 , ,H
50
100 150
DEPTH (CM)
y-s
200
250
Figure B.10 Mean and standard deviation of organic carbon vs. depth for Felda,
Riviera, and Pineda alfisols.
-------
U)
o
9. Oj
8.0--
7.0-
6.0
5.0-
4.0-
3.0-
1.0-
0.0
y+S
y-s
H 1 1 1
H 1
0
50
100 150
DEPTH (CM)
200
250
Figure B.ll Mean and standard deviation of pH vs. depth for Astatula and
Candler entisols.
-------
U)
o
U)
y+s
y-s
9. O
8.0
7.0
6. 0
5.0
3. Oy
2.0-
1.0-
0.0
\-\.IiIh-
0 50
H 1 H
100 150
DEPTH (CM)
200
250
Figure B.12 Mean and standard deviation of pH vs. depth for St. Lucie and
Paola entisols.
-------
U)
o
y+S
y-s
Q_
9. 0T
8.0
7.0
6.0
5.0
4.0'
3.0-
2.0-
1.0-
0.0
4 1 1
0
1 1 H
50
100 150
DEPTH (CM)
H - »
' - 1
200
250
Figure B.13 Mean and standard deviation of pH vs. depth for Apopka and
Arredondo ultisols.
-------
y+s
y-s
9. 0T
50
100
150
200
250
DEPTH (CM)
Figure B.14 Mean and standard deviation of pH vs. depth for Immokalee, Oldsmar,
Wabosso and Myakka spodosols.
-------
U)
o
y+S
y-s
0.
9. O
8.0
7.0
6.0-
5.0-
4.0;-
3.0-
2.0-
1.0-
0.0
H 1-
1 1 1
1 1 1 H-
H 1 1 I
0
50
100 150
DEPTH (CM)
200
250
Figure B.15 Mean and standard deviation of pH vs. depth for Felda, Riviera and
Pineda alfisols.
-------
APPENDIX C
DERIVATION OF FIRST-ORDER HYDROLYSIS RATE
EQUATION FOR SATURATED ZONE
Hydrolysis of aldicarb is strongly dependent upon pH and
temperature. This reaction is a typically first order and
that follows the equation:
(c.i)
where [A] = concentration of aldicarb (moles/liter)
k = 1st order rate constant (/time)
t = time
The second order rate constant (kj. ) can be determined by
plotting first order rate constant (kt) versus pH from several
different reactions:
k = ^- (C.2)
r [OH-]
The slope of the line is the second order rate constant (kr).
In the late 1800's Arrhenius observed that in most chemical
reactions the rate constant increases exponentially with
temperature. He described this relationship with the
following empirical equation:
307
-------
kr = A exp (-EA/RT) (c.3)
where k = rate constant
r
A = pre-expotential factor (mole/L-t)
E = activation energy (energy/mole)
o
R = universal gas constant (energy/mole K)
o
T = temperature, K
The logarithmic form of this equation is the equation for a
straight line:
kr = 2.303RT+
where log k = dependent variable
-E. 1
* (C.4)
1/T = independent variable
E = slope
A
log(A) = intercept
Although this relationship is empirically derived, a similar
relationship can be rigorously derived from thermodynamics
involving the equilibrium constant (keg) and the change in
free energy (AH°):
(c>5)
dT RT2
Substituting
keq = r and 4H° = El - E2
K2
308
-------
gives
d In(k1/k2)
dT " RT2 (C'6)
or in general
d In (JO E_
L_ = JL_ (C.7)
dT RT2
Integration leads to
kr= Ae- (c.8)
Arrhenius defined activation energy (E^ ) as the energy
threshold that molecules must overcome if they are to react
with one another. If the energy of the molecules is less than
the activation energy no reaction will occur.
The Arrhenius equation is used in this study to develop a
general equation that relates pH and temperature to the first
order reaction rate constant. Two studies (Lemley and Zhong,
1983 and Lemley and Zhong, 1984) have calculated the
activation energy and shown that the log-linear relationships
necessary for its validity hold.
The equation is rearranged so that by knowing EA , the
activation energy and A, the pre-exponential factor for a
309 :
-------
given compound, the first order rate constant k. can be
calculated for any given temperature and pH:
kr = 2.303RT
kt
Substituting k - gives
[OH J
kt \ -E 1
log ( £-) = + log A
[OH ]/ 2.303R T
or
-E 1
log k = + log A + log [OH ]
r 2.303R T
Substituting pH - 14 for log .[OH ] gives the final useful
form
-E 1
log k. = + log A + pH - 14
2.303R T
Uncertainty in Reaction Rates
A great degree of care must be taken in how the reaction rates
k are determined and used. Large uncertainty is introduced
from several different sources:
310
-------
1) temperature sensitivity
2) large extrapolations
3) experimental errors, and
4) environmental fluctuations.
Small temperature fluctuations during the hydrolysis
experiments can lead to large error propagation in k and E, .
A variation in temperature as small as .2 C leads to 2%
uncertainty in k and 5% in E . With a + 1° C change the
t A
uncertainty in k is 10% and in E , 100%. Larger variations
t A
in temperature of 10 and 25 will translate to uncertainty in
k on the order of a factor of 2.5 and 10 respectively.
Extrapolation over large temperature or pH intervals can also
lead to error. With an uncertainty in E of 5%, extrapolation
f\
over a 25°C temperature interval can result in 30% error in
the estimation of kf
Experimental error is another source of uncertainty. Most
rates are based on only three to six data points. With such a
small sample size the influence of one outlier or bad data
point can skew the entire study.
The rate constants are usually used to predict degradation in
the natural environment where there are temporal and spacial
fluctuations of both temperature and pH.
Due to all these sources of error, it is best to only consider
the estimate of k as being within an order of magnitude.
311
-------
APPENDIX D
Aldicarb TTR data from Oviedo and Davenport sites for 1984,
312
-------
Davenport 5/9/84
QUAD CORE # DEPTH(FT) TTR (ppb)
NW 2 0-16 -ND
NW 8 0-16 -ND
NE 13 0-1 ND
1-2 ND
2-4 ND
4-6 ND
6-8 60
8-10 144
10-12 72
12-14 ND
14-16 ND
NE
0-1
1-2
2-4
4-6
6-8
8-10
10-12
12-14
14-16
28
ND
ND
ND
ND
44
100
216
40
NE 12 0-1 32
1-2 ND
2-4 ND
4-6 ND
6-8 52
8-10 88
10-12 56
12-14 84
14-16 56
SW 15 0-16 -ND
313
-------
Davenport 5/9/84
QUAD CORE # DEPTH(FT) TTR (ppb)
SW 3 0-1
1-2
2-4
4-6
6-8
8-10
10-12
12-14
14-16
SE 5 0-1
1-2
2-4
4-6
6-8
8-10
10-12
12-14
14-16
SE 13 0-1
1-2
2-4
4-6
6-8
8-10
10-12
12-14
14-16
76
ND
ND
ND
ND
ND
16
52
96
84
ND
ND
32
76
104
80
72
44
64
ND
36
44
64
92
100
212
48
314
-------
Davenport 6/13/84
QUAD CORE # DEPTH(FT) TTR (ppb)
NW 3 0-16 ND
NW 8 0-16 ND
SW 11 0-1 ND
1-2 ND
2-4 44
4-6 104
6-8 296
8-10 200
10-12 328
12-14 272
14-16 88
SW
0-1
1-2
2-4
4-6
6-8
8-10
10-12
12-14
14-16
ND
ND
24
48
32
32
28
12
16
NE 6 0-1 16
1-2 ND
2-4 22
4-6 40
6-8 80
8-10 120
10-12 84
12-14 92
14-16 96
315
-------
Davenport 6/13/84
QUAD CORE # DEPTH(FT) TTR (ppb)
NE 12
NE 13
SE 4
SE 13
0-1
1-2
2-4
4-6
6-8
8-10
10-12
12-14
14-16
0-1
1-2
2-4
4-6
6-8
8-10
10-12
12-14
14-16
0-1
1-2
2-4
4-6
6-8
8-10
10-12
12-14
14-16
0-1
1-2
2-4
4-6
6-8
8-10
10-12
12-14
14-16
36
28
28
60
116
152
168
84
ND
16
ND
28
40
72
92
60
28
20
ND
ND
20
40
52
72
112
164
152
16
12
16
56
60
60
84
88
52
316
-------
Oviedo 5/22/84
ROW COL. DEPTH(FT) TTR (ppb)
11-12 38
11-12 34
11-12 13
13-14 17
13-14 52
25-26 53
21-22 50
0-1
1-2
2-3
3-4
4-5
0-1
1-2
2-3
3-4
4-5
0-1
1-2
2-3
3-4
4-5
0-1
1-2
2-3
3-4
4-5
0-1
1-2
2-3
3-4
4-5
0-1
1-2
2-3
3-4
4-5
0-1
1-2
2-3
3-4
4-5
1460
284
ND
ND
ND
ND
ND
ND
ND
ND
80
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
580
76
128
368
ND
ND
ND
ND
ND
ND
317
-------
Oviedo 5/22/84
ROW COL. DEPTH(FT) TTR(ppb)
24-25 42
22-23 31
27-28 28
28-29 29
28-29 15
27-28 12
18-19 4
0-1
1-2
2-3
3-4
4-5
0-1
1-2
2-3
3-4
4-5
0-1
1-2
2-3
3-4
4-5
0-1
1-2
2-3
3-4
4-5
0-1
1-2
2-3
3-4
4-5
0-1
1-2
2-3
3-4
4-5
0-1
1-2
2-3
3-4
4-5
32
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
ND
544
116
ND
ND
ND
ND
ND
ND
ND
ND
318
-------
Oviedo 5/22/84
ROW COL. DEPTH(FT) TTR (ppb)
40-41 48 0-1 80
1-2 ND
2-3 ND
3-4 ND
4-5 ND
32-33 29 0-1 ND
1-2 ND
2-3 ND
3-4 ND
4-5 ND
319
-------
Oviedo 6/20/84
ROW COL. DEPTH(FT) TTR (ppb)
12-13 42
13-14 6
17-18 15
17-18 15
18-19 26
34-35 - 48
30-31 33
0-1
1-2
2-3
3-4
4-5
0-1
1-2
2-3
3-4
4-5
0-1 (iced)
1-2
2-3
3-4
4-5
0-1 (no ice)
1-2
2-3
3-4
4-5
0-1
1-2
2-3
3-4
4-5
0-1
1-2
2-3
3-4
4-5
0-1
1-2
2-3
3-4
4-5
44
ND
28
ND
ND
ND
ND
ND
ND
ND
60
12
288
84
ND
80
12
40
76
ND
ND
ND
ND
32
ND
ND
ND
ND
ND
48
ND
ND
ND
ND
ND
320
-------
Oviedo 6/20/84
ROW COL. DEPTH(FT) TTR (ppb)
26-27 20(#1)
0-1
1-2
2-3
3-4
4-5
ND
ND
24
ND
ND
26-27 . 20(#2) 0-1 ND
1-2 ND
2-3 ND
3-4 ND
4-5 ND
321
-------
APPENDIX E
CUMULATIVE FREQUENCY DISTRIBUTIONS OF RELATIVE WELL-WATER
CONCENTRATIONS
The following summary table (Table E.I) and graphs show
cumulative frequency distributions of relative concentrations
in well water for the saturated zone scenarios.
The following notation convention is used for identifying each
scenario.
-------
Well Configurations
W (distance (ft), depth (ft), rates
(gpm))
FWD(0)W(300,50,2000), for example, would identify the scenario
that uses the unconfined Floridan Aquifer with the worst set
of hydrogeologic properties, no decay rate, a well distance of
300 ft from the aldicarb source, a well depth of 50 ft, and
pumping rate of 2000 gallons per minute (gpm).
323
-------
TABLE E.I RELATIVE CONCENTRATION VALUES FOR
INDICATED SATURATED SCENARIOS AT THREE
EXCEEDANCE PROBABILITIES
Scenario
Floridan Aquifer - worst Case
91 meters to well
Deep well, no decay
Decay
Shallow well, no decay
Decay
300 meters to well
Deep well, no decay
Decay
Shallow well, no decay
Decay
Exceedence Probability
.90 .50 .10
6.4E-4
8.1E-5
4.0E-3
1.3E-4
1.1E-3
3.3E-5
4.0E-3
3.3E-5
4.4E-3
3.7E-4
l.OE-2
5.8E-4
5.0E-3
1.4E-4
8.5E-3
1.8E-4
8.6E-3
1.7E-3
1.5E-2
2.6E-3
8.4E-3
6.2E-4
1.2E-2
6.4E-4
Floridan Aquifer - average case
91 meters to well
Deep well, no decay 5.7E-4 3.3E-3 7.IE-3
Decay 2.5E-6 1.8E-5 4.5E-3
Shallow well, no decay 6.3E-4 3.6E-3 7.5E-3
Decay 2.5E-6 1.8E-5 4.4E-5
300 meters to well
Deep well, no decay 1.6E-4 7.3E-4 3.0E-3
Deep well, decay 9.2E-9 3.1E-8 4.6E-8
Shallow well, no decay 9.6E-5 4.1E-4 1.9E-3
Decay 2.6E-9 8.7E-9 1.3E-8
Surficial Aquifer - worst case
91 meters to well
High pumping rate, no decay
Low pumping rate, no decay
Both, decay
300 meters to well
High pumping rate, no decay
Low pumping rate, no decay
Both, decay
1.1E-2
2.8E-2
2.1E-4
5.4E-3
6.7E-3
4.0E-6
5.0E-2
1.3E-1
1.1E-3
2.6E-2
4.7E-2
1.8E-5
7.3E-2
1.7E-1
3.3E-3
5.9E-2
l.OE-1
3.3E-5
Two-aquifer System, worst case
91 meters to well
High pumping, no decay 1.5E-4 6.9E-4 1.7E-3
Low pumping, no decay 3.4E-4 1.1E-3 2.1E-3
Both, decay 1.3E-5 6.5E-5 2.7E-4
324
-------
a o
i §
se
a:
O a
"s
LJ
£-1
5
tiJ
or c
a.
FWD(0)M(300,10-40,200)
FWD(0)H(300,10-40,40)
\J*
0.00 0.25
0.50 0.75 1.00 1.25
RELflTIVE CONCENTRRTION
1.50
1.75
*io-2
325
-------
o
LJ O
|S
LiJ
O
i« {5
o o
a:
E-< a
Sg-
O o
"o'J
in
\
.
CC O
{-
"ZL a
o 2'
a: "*
LJ
a. °
a.
\
\
\
\
v
\
\
\
\
\
\
\
i
CiJ
O ~Ns.
. x
FWD(29.7)W(300,10-40,40)
«».
i
FMD(29.7)W(300,10-40,200)
0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0
RELflTIVE CONCENTRATION
326
-------
XN ^ PWD ( 0 ) W( 300 , 50- 350 , 1000 )
FWD(0)W( 300, 50-350, 500)
0.0000.001 0.002
I I
0.003 0.004 0.005 0.006 0.007 0.008
RELflTIVE CONCENTRflTION
0.009 0.010
327
-------
LJ -
a
LJ o
LJ
o
»i 8
O o
So
O o
O
o-
in
« o
5*
cr: c
^ o J
E-
"Z. o
^2J
0
Q- °.
o.
\
\
\
\
\
\
\
\
\
\
\
\
\
\
N
N
FWD(29.7)H(300,50-350,500)
FWD(29.7)W(300,50-350,1000)
0.0 2.5 5.0 7.5 10.0 12.5 15.0 17.5 20.0
RELATIVE CONCENTRRTION
328
-------
a
a §-
a
CJ O
gg-
u
o
cn_-_
O a
E-« o
§8"
O o
*-
(C O
E-i
\.
FWD(0)W(1000,10-40,200)
FWD(0)W(1000,10-40,40)
0.000 0.002
0.004 0.006 0.008 0.010
RELflTIVE CONCENTRflTION
0.012
0.014
329
-------
s§
a
LJ o
8s
O O
a:
£- 0
§38
o
2
O o
tn
LJ
>
i i a
LJ
o
z
a: a
-
0- °.
o.
\
\
\
X
\
\
\
\
\
\
\
\
\
\
\
\
\
v
a
"X.
FHD(29.7)W(1000,10-40,40)
K
I ^
FWD(29.Z)W(1000,10-40,200) \
z. o
o2' -N
a:
\
\
0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0
RELRTIVE CONCENTRRTION
330
-------
1000,50-350,1000)
FWD(0)W(1000,50-350,500)
2.0 3.0 4.0 5.0 6.0 7.0
RELflTIVE CONCENTRflTION
8.0 9.0
HlO"
331
-------
o
LJ
cn Q.
« a
Z
O C3
o:
£-« o
s§-
O o
II o
SsH
o:
cn o
£-1
2L O
o:
\
\
\
X
\
\
\
\
\
\
\
\
\
\
\
\
\
FHD(29.7)W(1000,50-350,500)
FHD(29.7)H(1000,50-350,1000)
\
0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0
RELRTIVE CONCENTRRTION xlO~*
332
-------
3s
o
o
CJ
08
oa
QJ
O
O o
o:
f-" o
SB"'
O a
>
>-i o
_J
LJ
ct: a
a: c
T*
£-
Z a
o:
CJ
\
v
\
\\
Vx FAD(0)W(300,10-40,200)
Vv/ FAD(0)W(300,10-40,40)
\\
\
\
\
\.
\\
\\
0.000 0.001 0.002 0.003 0.004 0.005 0.006 0.007 0.008 0.009
RELRTIVE CONCENTRATION
333
-------
CxJ O
8 B-
X
LJ
a
O o
a:
E-i
8-
O o
"oJ
. if>
\
\
\
\
\
»
«
CiJ
02 ° FAD(29.7)W(300,10-40,200)
O "
en o
Z o
8 a-
o:
x FAD(29.7)H(300,10-40,40)
\
\
\
\
N
N
N
0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 40.0 45.0 50.0 55.0
RCLRTIVE CONCENTRflTION
334
-------
O
CJ O
O o
s
O o
»-« o
CO
or
B-
cc a
OR'
0.0
O.
\ \^FAD(0)W(300,50-350,500)
\ V
\
\
,^FAD(0%)W(300,50-350,1000)
\
0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0
RELRTIVE CONCENTRATION
MlO
335
-------
o
CO o
CO QV
08
CO
O
ii 8"
O a
CC
E-. o
8' . .
FAD(29.7)W(300,50-350,500)
Oo
i o
§*'
u
ce ca
CC CJ
OR'-
Z o
Ss-
a:
LJ _
a. °.
o.
\
\
\
\
\
\
\
\
\
\
\
FAD(29.7)W(300,50-350,1000)
0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 iO.O 45.0 50.0 55.0
RELRTIVE CDNCENTRRTION *10~"
336
-------
Q 5
CJ ""
a
cj o
o a
t
a:
18'
O o
-------
S §~Ts*
O
O
CO X
O
Kg- \ \
\
\\
\ \ FAD(29.7)H(1000,10-40,40)
o -
a o \
d-
tn \
\
\
\
'\
\
\
\
\
\
8 "1
33°. J
oRl
\
\
\
\
\
\
\
\
\
\
0.0 2.5 5.0 7.5 10.0 12.5 15.0 17.5
/
FAD(29.7)H(1000,10-40,200)
RELRTIVE CONCENTRRTION
338
-------
o
gS
|i
\
CD o
8
O o
ct: a
a:
CJ
i
\
\
\
N\
FAD(0)W(1000,50-350,500)
50-350,1000)
\
'^^
0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 40.0 45.0
RELRTIVE CONCENTRflTION *10"*
339
-------
0§-N.
to -
a
LiJ O
I"-
"
00
01
FH 0
a o
o 0'.
in
>
« o
a: a
f-
z o
Ss-
a:
.
\\
\
s
\
N
. S
FAD(29.7)M(1000,50-350,500)
J
r
FAD(29.7)H(1000,50-350,1000)
N
N
S
\
V
N
\
\
\
N
\
\
\
\
\
\
\
\
\
X
\
\
N
\
\
\
0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 -40.0 45.0 50.0 55.0
RELRTIVE CONCENTRRTION
340
-------
LJ
a
"J r-».
08
LJ
O
CO -I.
O 0
a:
ct:
\
\
\
*
|B
O o
°9
WS
*" 9"
a: o'
o
cc o
Z a
B°H
\
X
\
\^SHD(0)W(300, 10-35,40)
\ " --\
\ \
V-*SWD(0)W(300,10-35,200) \
\
\
\
\
\
\
\
\
\
N
\
\
\
\
\
N
\
V
\
0.000 0.025 0.050 0.075 0.100 0.125 0.150 0.175 0.200
RELRTIVE CONCENTRflTION
341
-------
tlJ
Q
U o
tJ rj.
08
O o
HH '
£s'
cc
\
O
z
O o
"o-
in
CC
QJ
a:
SHD(29.7)W(300,10-35,200)
SWD(29.7)H(300,10-35,40)
"..v......
Z o
y°J
ttT
CJ _
0.0
5.0
10.0
15.0
20.0
25.0
30.0
RELflTIVE CONCENTRRTION
35.0
40.0
342
-------
V
LJ O
wx- \v
a
O o
en
E-« o
18-
O o
S8''
«« o
e*
QJ
a:
az o
\
NX
\\
\\
*. \
\ \
\ \
X
\
SWD(0)H(1000,10-35,40)
\
\
\
B-
\ ^rSWD(0)W(1000,10-35,200)
,
\
\
\.
\
0.00 0.02 0.04 0.06 0.06 0.10 0.12
RELRTIVE CONCENTRflTION
343
-------
cj -
o
CiJ a
*
-j ""^
CO ^J \ \
\ \
O o
a:
{- a
68
2
O o
QJ
z
CC C3
2 o
8s-
\x
*-.\
\
SWD(29.7)W(1000,10-35,40)
\
\
\
\
\
\
\
\
SHD(29.7)W(1000,10-35,200) \
\
\ \
Vx
\
0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 40.0
RELRTIVE CONCENTRRTION
344
-------
a 8
a-
a
CO O
LJ
O O
.X.
CJ
»-* a
a: o
8-
(E O
5s-
. v^
\ x\
\ \ £ 2WD(0)W(300,50-400,700)
\ \
%
(^«2WD(0)W(300,50-400,1500)
»..
X V
\
\ \
\
0.0 2.5 5.0 7.5 10.0 12.5 15.0 17.5 20.0 22.5
RELRTIVE CONCENTRRTION *10~
345
-------
o
°§\
p ^ *^ *
y *
a \
o \
%+ \
CJ \
o
£g-
O a \
\
a
8
82
u1"
\\
\ \
*. N.
d \ \ 2WD(29.7)W(300,50-400,1500)
a: o ' A » x
^ I
N ./
^\ \ *
V? ' x
r\ \
\
2WD(29.7)H(300,50-400,700)
S I' " V
B \ \
CL °. \ \
O V
0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0
RELRTIVE CONCENTRRTION
346
------- |