Proposed Methods For Determining
Watershed-derived Percent Crop Areas
And Considerations For Applying Crop Area Adjustments
to Surface Water Screening Models

PRESENTATION TO FIFRA SCIENCE ADVISORY PANEL

MAY 27, 1999

William R. Effland, Ph.D., Nelson C. Thurman, M.Sc., Ian Kennedy, Ph.D.
U.S. EPA Office of Pesticide Programs
Environmental Fate and Effects Division


-------
Table of Contents

Section 1: Overview 	1

Overview of the Document 	3

Goals	4

Objectives 	4

Critical Assumptions and Limitations	5

Section 2: Development of Methodology to Estimate Watershed-derived Percent Crop Areas with

Geographical Information Systems (GIS) Technology 	8

Method for Deriving a Watershed-Based Percent Crop Area 	8

Application of the PCA Adjustment to Screening Models	10

Section 3: Comparison of PC A-Adjusted Modeling Results With Monitoring Data 	17

Modeling and Monitoring For Corn-Soybean Herbicides in the Midwestern U.S	18

Monitoring Data on Corn-Soybean Herbicides in the Midwest 	19

Comparison of Monitoring Data to Modeling Results for Corn-Soybean Herbicides

in the Midwest U.S	31

Modeling and Monitoring In the Central Valley of California	35

Monitoring Data for the San Joachin River, CA 	36

Modeling of Minor-Use Pesticides in Central Valley, CA 	37

Comparison of Monitoring Data and Modeling Results in the Central Valley, CA

	37

Calculation of PC As in California	39

Section 4: Preliminary Comparison of Screening Model Results With Available Surface Water

Monitoring Data 	41

References	45

Listing of Background Documents	47

Appendix A: PRZM Model Inputs	48

Inputs for Corn-Soybean Herbicide Simulations in the Midwest U.S	48

Inputs for Simulations in the San Joachin River, CA	59

Appendix B: Output Data From PRZM/EXAMS Simulations 	64


-------
Section 1: Overview

Following passage of the Food Quality Protection Act (FQPA) in 1996, the Office of
Pesticide Programs (OPP) has developed procedures and science policies for estimating pesticide
concentrations in drinking water to support tolerance assessment. At the July 29, 1998 FIFRA
Science Advisory Panel (SAP), OPP proposed replacing the existing "edge of field pond"
estimation scenario for screening level exposure assessment of drinking water with an "index"
reservoir that would also consider the extent of crop coverage on a watershed scale. This SAP
session provides a progress report on the development of "Watershed-derived Percent Crop
Areas" (previously termed "crop area factors" in Background Document 3), and seeks scientific
advice on implementation of the watershed-derived Percent Crop Areas in the FQPA drinking
water exposure assessment using the index reservoir modeling scenario.

Science Policy 5: Estimating The Drinking Water Component of a Dietary Exposure
Assessment (Background Document 5) describes the Agency's process for including drinking
water exposure in its dietary assessments and its current use of screening models to estimate
pesticide exposure in drinking water. Briefly, the process for assessing pesticide exposure from
surface water sources of drinking water consists of:

Collecting all available laboratory and field data submitted by the registrant to assess the
fate and transport characteristics of the particular pesticide and any toxic degradation or
transformation products.

Entering pesticide-specific data from these studies in mathematical screening models to
estimate pesticide concentrations in water in pesticide use areas.

Comparing the model screening estimates to human health-based "drinking water levels of
comparison" (DWLOC), which are derived after first considering all food-related and
residential exposures for which EPA has reliable information.

If the model estimates of pesticide concentration in drinking water exceed the DWLOC,
refining its estimate by gathering available water monitoring data for analysis.

OPP generally does not base significant risk management action (e.g., revocation or denial
of a tolerance) on screening model estimates. If monitoring data are available and reliable, OPP
scientists analyze this data to derive appropriate short- and long-term exposure concentrations. If
monitoring data are not available or are not sufficient for refining the screening level estimates,
OPP makes a risk management decision as to the need for targeted surface water monitoring
and/or risk mitigation.

OPP currently uses a two-tiered mathematical screening model process to rapidly assess
whether pesticides are likely or unlikely to occur at substantial levels in drinking water derived
from surface water:

1


-------
GENEEC (GENeric Estimated Environmental Concentrations) provides an initial
screening level assessment of pesticide concentrations in surface water;

the linked Pesticide Root Zone Model (PRZM) and Exposure Analysis Model System
(EXAMS) models provide a more refined screen by considering site-specific
environmental and climatic variables.

Background Documents 1 and 3, from the December 1997 and June 1998 FIFRA SAP
presentations, provide more detail on the models. GENEEC and PRZM/EXAMS, initially used
by OPP for ecological risk assessments, are two mechanistic models available to OPP for rapidly
and cost-effectively producing high-end estimates of pesticide levels in surface water. If the
surface water estimates using GENEEC do not exceed the DWLOC, then OPP concludes that the
pesticide is not expected to pose an unacceptable risk and no further evaluation is necessary. If
the GENEEC results indicate a potential concern, then PRZM/EXAMS modeling refines the
estimates of potential pesticide levels in surface water by including more pesticide-specific
properties, simulating multiple years to reflect climatic variations, and modeling on a crop-specific
basis. In comparison to GENEEC, PRZM/EXAMS includes more site-specific information in the
scenario details regarding application method and temporal distribution with weather, and better
accommodates specific environmental fate properties of chemicals. Both models represent a
relatively small body of water (20 million liter, or 5.3 million gallons in capacity) located at the
edge of a 10-hectare (approximately 25 acres) treated field. Both models assume that the entire
area draining into the water body is planted to the crop being treated and that the entire crop is
treated with the pesticide.

The ultimate goal of the Agency is to have a validated basin-scale model to develop a
more refined estimate of pesticide levels in surface water sources of drinking water. OPP
presented a preliminary assessment of available watershed-scale surface water models at the 1998
FIFRA SAP (Background Document 3) and is continuing with model evaluation efforts. In the
interim, the existing field-scale models are being applied at a watershed-scale, with adaptations
intended to reflect watershed-scale use.

As indicated in the July 1998 FIFRA SAP presentation, OPP planned to replace the
current field pond scenario used in its screening assessments with an index reservoir based on an
actual field reservoir (see Background Documents 3, 4, and 5 for a more detailed discussion).
Further, the index reservoir model output would be adjusted for the percentage of the reservoir
area in agricultural production to more realistically reflect watershed-scale use. OPP proposed
that the watershed-based Percent Crop Area (PCA) adjustment would be applied to the model
outputs as an additional step in the screening tier process. If pesticide concentrations (peak
and/or long-term average) estimated with the first two screening model tiers (GENEEC and
PRZM/EXAMS adapted for the index reservoir) still exceeded levels of concern, then the model
results from PRZM/EXAMS would be adjusted by a factor that represents the maximum percent
crop area found for the crop or crops being evaluated. While several assumptions and limitations,
discussed below, are inherent to this process, OPP believed the incorporation of a PCA factor in

2


-------
the screening process would produce more realistic estimates of pesticide concentrations in
surface water sources of drinking water while still resulting in an appropriately conservative or
protective screening level estimate.

Overview of the Document

The SAP document contains 4 sections with Section 1 describing the introduction, goals,
objectives and critical assumptions and limitations. Section 2 presents the basic approach OPP
used to estimate watershed-derived percent crop area (PCA) factors using geographic information
system (GIS) technology and gives a brief overview of how the PCA would be applied to
screening estimates. Section 3 provides a preliminary evaluation of the PCA adjustment using the
index reservoir scenario. Model results are compared to midwestern U.S. reservoir and stream
monitoring data for selected corn and soybean herbicides, and to USGS National Water Quality
Assessment (NAWQA) data from the Central Valley, CA, for selected minor use crop pesticides.
These two groups of data are among the best available for model assessments. It is important to
know that both the index reservoir scenario and the PCA adjustment are being calibrated against
atrazine monitoring data because that is the most extensive and widespread data available. OPP
recognizes the need to evaluate these model refinements using a broader spectrum of pesticides
and uses. While a preliminary attempt has been started here, further comparisons are needed.
Section 4 describes a comparison of screening model results with available surface water
monitoring data for various drinking water assessments completed by OPP.

In developing the PCA, OPP encountered some issues that suggest that implementation of
the PCA adjustment with the reservoir scenario in screening level assessments may not provide an
appropriately conservative estimate of pesticide levels in surface water. Specifically, we are
concerned that PRZM/EXAMS may not be realistically capturing basin-scale processes for all
pesticides or for all uses. A preliminary survey of water assessments which compared screening
model estimates to readily available monitoring data suggest uneven model results (this is
discussed in further detail in Section 4). In some instances, the screening model estimates are
more than an order of magnitude greater than the highest concentrations reported in available
monitoring data; in other instances, the model estimates are less than monitoring concentrations.

OPP believes the following steps need to be taken to evaluate the effectiveness of
PRZM/EXAMS as a screening model:

1.	Conduct a sensitivity analysis of PRZM, EXAMS, and the linked models to determine
what input factors most influence the model results and identify potential conditions (site,
chemical, weather) under which the models may not work as expected.

2.	Conduct a more thorough survey of modeling and monitoring comparisons for all
pesticides in which such data is available. Identify, to the extent possible, the specific
conditions of the monitoring data (e.g., water body, date of sampling, characteristics of
the drainage area, cropping patterns and likely pesticide use areas, soils present, weather

3


-------
patterns, and other possible sources of pesticide exposure). Where more than one
modeling scenario was run, compare results of each model run to the monitoring data.

3. Through a systematic comparison of modeling and monitoring data and the ancillary data,
attempt to identify specific chemical or scenario characteristics that could be leading to
inconsistencies in the modeling results. This evaluation could then be used to determine
whether, for certain pesticides or uses, corrections are needed or whether another form of
screening is necessary.

Goals

OPP plans to develop watershed-scale models to produce refined estimates of drinking
water exposure for pesticides in surface waters. Our current approach uses tiered field-scale
modeling developed for aquatic exposure assessments to estimate potential exposure from
pesticides in surface water used for drinking water. The current approach is intended to serve as
a screening method that distinguishes between pesticides with minimal predicted impact to
drinking water from those that may pose a greater risk to human health via drinking water
exposure.

The application of percent crop area to the currently-used models for producing
screening-level estimates of drinking water exposure from pesticides is also an interim approach.
Validated basin scale models and adequate and reliable surface water monitoring data are clearly
needed to improve the Agency's ability to accurately estimate pesticide exposure from drinking
water surface water sources.

Objectives

The objectives of the Percent Crop Area project are to:

1.	Develop a method to estimate watershed-derived percent crop areas (PCA) using
Geographical Information System (GIS) technology.

2.	Evaluate the utility of the PCA adjustment as a modification to the current screening tools
by comparing PCA-adjusted PRZM-EXAMS model estimates, using the Index Reservoir,
to the best available monitoring data for a comparative evaluation:

a.	Midwestern U.S. reservoir and stream monitoring data for selected corn and
soybean herbicides; and

b.	USGS National Water Quality Assessment (NAWQA) program monitoring data in
Central Valley, CA, for selected minor crop pesticides.

3.	Identify assumptions and limitations that need to be considered in applying the PCA
adjustment to existing OPP surface water screening models.

4


-------
Our objectives in coming before the SAP are to provide an update on the approach we are
considering for developing and using PCA factors, present scientific issues and concerns we have
encountered in developing this approach, and request scientific input in resolving these issues and
proceeding in improving our screening procedures.

Critical Assumptions and Limitations

The PCA is a watershed-based modification. Implicit in its application is the assumption
that the currently-used field-scale models to which it is applied reflect basin-scale processes
consistently for all pesticides and uses. In other words, we are making the assumption that
PRZM/EXAMS is indeed modeling a watershed. This project does not attempt to validate
PRZM, EXAMS, or the linkage between PRZM-generated runoff and the fate and transport
modeled by EXAMS. These models and model scenarios have not been well validated; however,
ongoing and future efforts by the Agency and industry should increase our confidence in the
modeled predictions. If the models fail to capture pertinent basin-scale fate and transport
processes consistently for all pesticides and all uses, the application of a factor that reduces the
estimated concentrations predicted by modeling could, in some instances, result in inadvertently
passing a chemical through the screen that may actually pose a risk.

For the GIS analysis, the spatial data was derived from readily-available sources that have
inherent limitations due to their scale (1:2,000,000 for the watershed and county coverages). The
conversion of the county level data to watershed-based percent crop areas assumes the
distribution of the crops within a county is uniform and homogeneous throughout the county area.
Other limitations of the spatial analysis include the assumption that the watershed area which
contributes surface runoff is spatially uniform and the condition of adjacency between the treated
fields and the surface water bodies are not considered. These two limitations should be addressed
in future watershed-scale modeling efforts if the goal is to develop realistic predictions of the
pesticide concentrations from runoff into surface waters. The crops data were obtained from the
1992 Census of Agriculture which is the most readily available county level data for this project.
However, recent changes in the agriculture sector from farm bill legislation may significantly
impact the distribution of crops throughout the country. The methods described in this report can
rapidly be updated as more current agricultural crops data are obtained. Although yearly changes
in cropping patterns may occur, we assume this variation will cause minimal impact to the interim
goals described in this report. This project only evaluates pesticides applied to agricultural crops
so any contributions to surface waters from non-agricultural uses such as urban environments are
not considered. Furthermore, this project does not consider percent crop treated with one or
more pesticides because pesticide usage data is extremely limited at this time.

Available monitoring data for pesticides in surface waters are very limited. Sample
distribution and frequency typically are not sufficient for accurately predicting peak
concentrations and the data examined in this report normally cover a very limited time span
(frequently less than 5 years with typical ranges from 1-3 years).

5


-------
Questions for the SAP Panel

1.	Given preliminary comparisons between modeling and monitoring data, and the
information presented regarding development of a Percent Crop Area (PCA), does the
SAP think the PCA adjustment to PRZM/EXAMS modeling is a reasonable approach to
obtain more accurate and appropriately conservative estimates of pesticide concentrations
in surface water for screening evaluations of drinking water exposure? If the steps are not
considered to be appropriately conservative for screening evaluations, does the SAP have
any recommendations for how the EPA should use PC As in its drinking water assessment
process?

2.	A GIS data processing method for calculating the PCA was presented. If we are able to
resolve concerns regarding current model and monitoring inconsistencies, does the SAP
think this GIS procedure is an appropriate method to account for the portion of the
watershed planted to the crop or crops of interest?

3.	In estimating water concentrations for pesticides applied to multiple crops in a watershed,
we modeled each crop separately, applied the maximum PCA to the modeled results, and
then summed the outputs. For example, the model results for corn multiplied by the
maximum PCA of 0.46 and the results for soybeans, multiplied by 0.41, were then
summed to provide an estimate of pesticide concentrations for metolachlor use on corn
and soybeans. Limitations to this approach may occur when the pesticide is used on
multiple crops (such that the equivalent PCA is greater than 100%) or when the timing
and/or rates of application vary for different uses. Can the SAP provide any
recommendations for determining a reasonable assessment process that considers multiple
uses of one or more pesticides within a watershed?

4.	Evaluation of this project required PRZM/EXAMS modeling of selected crops with the
preliminary version of the index reservoir scenario. Furthermore, the comparison relied on
available reservoir and surface water monitoring data from limited sources (e.g., ARP,
USGS NAWQA program). This evaluation produced variable results with some cases in
which available monitoring data exceeded modeled results and other cases in which
modeled data are conservative (i.e., clearly higher than available monitoring data). In the
document overview, we have proposed steps to evaluate PRZM/EXAMS for
inconsistencies as a screening model. What suggestions does the SAP have that could
help us better understand these inconsistencies?

5.	The PRZM developer has indicated the model has scale limitations when changing from
field to watershed to basin scales. Possible sources of error when changing scales may be
caused by using a single curve number for the entire basin, the hydraulic length, and
simplifications for field scale processes that may not apply to the more complex basin
scale. In addition, the application of the PCA in this report involves differing scales:
PRZM is a field-scale model, the PCAs are generated from basin-scale areas (8-digit

6


-------
hydrologic units), and the index reservoir represents a watershed scale. Does the SAP
have any recommendations for addressing the scale limitations of this model? Can the
SAP suggest other watershed-based, mechanistic models that could estimate
environmental concentrations of pesticides in drinking water?

6. The proposed application of "percent crop areas" for watershed modeling does not
consider "percent crop treated" for one or several pesticides. Because pesticides in
drinking water sources is a localized concern, a national percent crop treated estimate is
not appropriate. At this time, data on percent crop treated at the farm are extremely
limited, available only in New York and California. Does the SAP think it would be
reasonable for the Agency to develop a method similar to the PCA approach to
incorporate "percent crop treated" into the model refinements?

7


-------
Section 2: Development of Methodology to Estimate Watershed-derived
Percent Crop Areas with Geographical Information Systems (GIS) Technology

The current PRZM-EXAMS modeling for drinking water exposure assessment assumes
the watershed is completely planted with the crop of interest (i.e., 100% cropping coverage). The
100% cropped watershed assumption is reasonable for field scale aquatic exposure such as a farm
pond adjacent to a corn field in the Midwest states. For modeling of drinking water exposure
assessment, the assumption of 100% cropped watersheds may not be reasonable. This section
discusses the development of watershed-derived percent crop areas using GIS tools as an interim
corrective approach to modeling drinking water exposure assessments in conjunction with the
Index Reservoir. The method discussed below is similar to the unpublished procedure developed
by the USGS for interpreting the NAWQA study sites (G. P. Thelin, USGS, personal
communication, 1999).

Method for Deriving a Watershed-Based Percent Crop Area

Agricultural crop data for the United States are collected at a county level by the U.S.
Department of Agriculture. The Census of Agriculture collects county-level crop data every 5
years. For this work, we examined the 1992 Agricultural Census county-level data for five major
crops (corn, soybeans, wheat, cotton and potatoes) and five minor crops (apples, citrus, grapes,
peanuts, strawberries). Additional minor crops for Central Valley, California discussed in
Section 3 of this report included almonds, apricots, alfalfa, and walnuts.

Two principal GIS coverages (County boundaries, 8-digit Hydrologic Unit Codes) were
the primary spatial data sources for this analysis. Watershed boundaries (8-digit HUCs) were
obtained from 1 to 2,000,000 scale Hydrologic Unit map of the conterminous United States
(Allord, 1992; http://water.usgs.gov/lookup/getcover7huc2m) and contains Hydrologic Unit
Boundaries and Codes for the conterminous United States. This coverage was developed for
regional and national data display rather than specific local data analysis due to the small scale of
this coverage. County boundaries were obtained from the l;2,000,000-scale map of county
boundaries for the conterminous United States (Lanfear, 1994;

http://water.usgs.gov/lookup/getcover7county2m). This coverage was derived from the
Digital Line Graph (DLG) files representing the l:2,000,000-scale map in the National Atlas of
the United States and used as the base map for the county crops information.

The watershed-derived percent crop areas for each of the 10 crops were calculated by
intersecting the HUC coverage and the County Crop coverage in Arc-View 3.1 using the
geoprocessing analysis tool. The areas for the resulting polygons within each 8-digit HUC were
updated using the Update Area feature to indicate the corrected hectares of the new polygons
(Figure 2-1). The percent county area was multiplied by the updated polygons areas to calculate
the hectares of crop for each watershed. The calculated crop area was summed for each
watershed and then divided by the total watershed area to determine the area-weighted
"watershed-based percent crop area" or PCA (expressed as a decimal). These values were ranked

8


-------
from largest to smallest value and the maximum value for each crop is listed in Table 2-1. The
ranked values were also checked by comparison to the county maps to verify that data processing
errors were not contained within the final output. For example, initial calculations of the soybean
PCA revealed a maximum value of 0.92 which resulted from an error during geoprocessing in

which the HUC also included some estuarine (non
terra firmci) area. The maximum PCAs for each
crop are applied to surface water modeling results
for the Index Reservoir scenario in Section 3 of
this report. Figures 2-2 through 2-6 illustrate the
percent crop area for each major crop (corn,
soybeans, wheat, cotton, and potatoes).

Figure 2-1. Example of Intersected HUC 18040002 with
County Boundaries, Central Valley, California.

Table 2-1. Summary of Maximum Percent Crop Areas (without Land Use coverage)

Crop

Maximum Percent Crop
Area (as a decimal)

Hydrologic Unit Code
(8-digit)

State

Corn

0.46

07090007
07100003

Illinois
Iowa

Soybeans

0.41

08020201

Missouri

Wheat

0.56

09010001

North Dakota

Cotton

0.20

08030207

Mississippi

Potatoes

0.06

02080101

North Dakota

Apples

0.06

18020117

18030008

18030009

18030010

California
(all four HUCs)

9


-------
Grapes

0.03

04140101

New York

Citrus

0.14

03080203

Florida

Peanuts

0.09

03130010

Georgia, Alabama, and





03130004

Florida

Strawberries

0.012

18060001

California

Adjustments for Land Use/Land Cover

Most PC As presented in this report were calculated without considering land use/land
cover information. Land use/land cover data can help give a more refined estimate of a PCA
because it can help to better define a crops distribution within a county. This can be important in
areas where geography limits crops to one area of a county, but is not important in areas such as
the midwest where crops are evenly distributed. We used PCAs calculated without land use/land
cover data on our corn and soybean modeling, but used land use/land cover adjusted PCAs for
modeling done for the minor crops in California. The calculation of land use/land cover adjusted
PCAs for the hydrologic unit from California is presented at the end of section 3.

Application of the PCA Adjustment to Screening Models

OPP plans to use the watershed-based PCA adjustment as an additional step in its tiered
screening process. The PCA adjustment would be applied only to those pesticides which still
exceeded levels of concern after going through the first (GENEEC) and second (PRZM/EXAMS
with the index reservoir scenario) screens. For a single crop, the output generated by the
PRZM/EXAMS model would be multiplied by the maximum PCA (expressed as a decimal
fraction) generated for the crop in question. As an example, for a pesticide used only on corn, the
PRZM/EXAMS estimated environmental concentrations would be multiplied by 0.46.

A pesticide with multiple uses raises certain questions in applying the PCA. Should the
maximum PCA for each crop be summed together or should individual PCAs for the combined
crop uses be generated for each combination? In the section that follows, the PCA for corn and
soybeans for metolachlor was generated by adding together the individual PCAs for each crop,
even though the PCAs were derived from different areas of the country. For the minor crops
evaluated against the Central Valley NAWQA data, a single PCA representing the combined crop
acreage was used. In the case of the former approach, a combined PCA of greater than 100%
could be generated by summing the maximum PCAs for each crop used. In the latter approach,
PCAs would have to be generated for each crop combination being evaluated. Other issues raised
by the multiple crop use approach include the application rate to select (whether to use the
maximum application rate from the combination of uses or split application rates) and the timing
(the modeling assumes the pesticide is applied to all fields at the same time). In each case, more
complex adjustment steps require more time and resources and introduce additional levels of
uncertainty. Some steps may be feasible in a higher tier screening approach while others may be

10


-------
better left for consideration in a validated basin scale model.

11


-------
Figure 2-2. Distribution of Watershed-derived Percent Crop Areas for Corn.

Percent Crop Area - Corn

/C4r

\

n

i

Source: 1992 Agricultural Census
800	0

800

1600 Miles

I States
Corn PCA (dec.)

0.01
0.09
0.16
0.24

| 0.31
¦ 0.39

N

ifr

0.09
0.16
0.24
0.31
0.39
0.46

Drafted by W.R. Effland, 4/7/99

12


-------
Figure 2-3. Distribution of Watershed-derived Percent Crop Areas for Cotton.

Percent Crop Area - Cotton

f

--v

-v

Source: 1992 Agricultural Census

800	0

800

1600 Miles

States
Cotton PCA (dec.)

| | 0.01 -0.05

~	0.05-0.09

~	0.09-0.12
0.12-0.16
0.16-0.20

N

4"

Drafted by W.R. Effland, 4/7/99

13


-------
Figure 2-4. Distribution of Watershed-derived Percent Crop Areas for Peanuts.

Percent Crop Area - Peanuts

P

I C/")

£J



%/
J—r-

1 	'***>>

/

¦4

Source: 1992 Agricultural Census

800

800

\\

1600 Miles

States
Peanuts PCA (dec .)
0.01 -0.03
0.03-0.05

j	I 0.05-0.07

0.07-0.09

N

W^|-E

s

Drafted byW.R Effland, 4/7(99

14


-------
Figure 2-5. Distribution of Watershed-derived Percent Crop Areas for Soybeans.

Percent Crop Area - Soybeans

f/



States
Soybeans PCA(dec.)
I 0.01

0.09

0.17
0.25
0.33

0.09
0.17
0.25
0.33
0.41

Source: 1992 Agricultural Census
800	0

800

1600 Miles

N

Drafted byW.R. Effland, 4/8/99

15


-------
Figure 2-6. Distribution of Watershed-derived Percent Crop Areas for Wheat.

Percent Crop Area - Wheat

Source: 1302 Agricultural Census

800

800

1600 Miles

States
Wheat PCA (dec.)

0.01 -0.12
	0.12-0.23

0.23-0.34
0.34 - 0.45

0.45-0.56

w

N

Drafted by W.R. EfTland, 4/7/99

16


-------
Section 3: Comparison of PC A-Adjusted Modeling Results With Monitoring

Data

When OPP initially presented the Crop Area Factor (now Percent Cropped Area, or PC A)
approach to the December 1997 FIFRA SAP, the panel recommended that OPP "validate" the
approach against monitoring data, particularly the "extensive data available for corn herbicides in
the Midwest" (FIFRA SAP, 1997; Background Document 2). The original approach proposed
applying the PC A to GENEEC (Tier 1) modeling results. In July 1998, OPP proposed an index
reservoir (IR) to replace the existing farm pond scenario used in PRZM/EXAMS (Tier 2)
modeling. While OPP is currently in the process of replacing the original index reservoir
(Shipman City Lake, IL), the results of the evaluation showed the impact of incorporating the
PCA into the modeling approach (FIFRA SAP, 1998; Background Document 3).

PRZM/EXAMS modeling with the IR resulted in higher estimated atrazine concentrations than
predicted using the standard farm pond; however, when the IR was adjusted for the PCA of 0.25
for corn, concentrations were lower (Table 3-1).

Table 3-1: Comparison of Atrazine Concentrations from Use on Corn Estimated by PRZM/EXAMS with (a)
Farm Pond Scenario, (b) IR, and (c) PCA-adjusted IR with Available Monitoring Data for Shipman City Lake
(adapted from Background Document 3).	



Peak (jig/L)

Annual Mean (jig/L)

Overall Mean (jig/L)1



Median

90 %2

Median

90 %2

Mean

UCL90 3

Model: Standard Farm Pond

9

56

4

12

6

7

Model: Index Reservoir (IR)

15

132

5

33

11

15

Model: PCA-adjusted IR 4

4

33

1

8

3

4

Monitoring: ARP, 1995 5

3



2







Monitoring: ARP, 1996 5

35



12







1	Modeling results based on 34 years of simulations, from 1950-1983.

2	Ninety percent values are greater than ninety percent of peak or annual mean values.

3	Upper 90% confidence bound on the mean

4	PCA of 0.25, based on corn acreage in Macoupin County, IL

5	Concentrations are for treated water samples taken from Shipman City Lake (Hackett, 1996, 1997)

The monitoring data for Shipman City Lake come from an acetochlor registration
partnership (ARP) surface water monitoring study (Hackett, 1996, 1997), consisting of 14
finished water samples taken in each of 1995 and 1996. Because this data is for treated water, the
concentrations in the untreated water from the reservoir is expected to be higher than those
reported. The crop adjustment factor used in the 1998 SAP presentation was based on data from
Macoupin County, IL, rather than on the watershed draining into the reservoir. The PCA-
adjusted 90% peak concentration was roughly equal to the peak concentration reported for
Shipman City Lake in 1996. The upper 90% annual mean concentration (based on 34 years of
simulations) estimated using the standard farm pond scenario was nearly identical to the mean

17


-------
concentration found in Shipman City Lake in 1996, while the PCA-adjusted model estimate was
lower. Because of the limited size of the monitoring database, both in terms of frequency of
sampling per year and number of years of sampling, a more extensive comparison of PCA-
adjusted modeling data to available monitoring data is needed.

The application of the PCA to model estimates will result in a reduction in the estimated
concentrations by the fractional area of the watershed planted to the crop or crops being
evaluated. OPP uses these models to screen out those pesticides which are not likely to be found
in surface water sources of drinking water at concentrations sufficient to pose a health concern
and to identify those pesticides for which more extensive assessments, including monitoring
studies, are needed. To be effective as an adjustment to screening model estimates, the PCA
should result in estimated concentrations that are closer to, but not less than, actual pesticide
concentrations in vulnerable (prone to pesticide-laden runoff) surface water sources.

Modeling and Monitoring For Corn-Soybean Herbicides in the Midwestern U.S.

The monitoring studies that follow provide an initial assessment of ability of PCA-adjusted
model estimates to serve as a screen for selected corn-soybean herbicides. Limitations in available
monitoring data make a rigorous assessment of the effect of the PCA-adjustments on the
screening capabilities of PRZM-EXAMS difficult. Specifically:

Few pesticide monitoring studies have been conducted for more than a few years over a
geographically broad range. More monitoring data exist for streams and rivers than for
reservoirs, which are believed to be more vulnerable to pesticide contamination.

Most reservoir studies cover a limited scope of pesticides with respect to the extent of use
and the physiochemical properties of the compounds studied. More data are available for
the corn herbicides than other pesticides. Inferences from monitoring data using these
herbicides assume that other pesticides behave similarly in the field and are handled
similarly in model simulations.

The sampling frequency in the monitoring studies is not sufficient to make meaningful
comparisons of peak concentrations in the reservoir. The farther apart the sample
intervals are spaced, the less is the likelihood that the sampling will capture the true peak
concentration of the pesticide in the reservoir. Estimates of medians or mean annual
concentrations also depend on the frequency and spacing of the sampling.

The USGS National Water Quality Assessment (NAWQA) provided non-targeted
pesticide data on a number of watersheds throughout the U.S. In some of the sub-basin
sampling sites, the sampling frequency (2-3 days during the growing season) provides the
best chance of detecting short-term peak concentrations. However, the sampling stations
represent flowing water bodies (streams, rivers) with short residence times. Longer-term
concentrations would likely be lower than in reservoir with longer residence times.

18


-------
The reservoirs and watersheds sampled in these studies are concentrated in areas identified
as having a high potential for soil, pesticide, and nitrogen runoff (Kellogg et al, 1997). The data
sets are limited in time. Rainfall, an important pesticide runoff driver, varies in frequency and
timing from year-to-year, and studies of short duration are less likely to capture the range in
pesticide concentrations possible in reservoirs. Fallon (1994) observed that timing, as well as
frequency and magnitude, of rainfall events are all critical factors controlling pesticide
concentrations in reservoirs.

Monitoring Data on Corn-Soybean Herbicides in the Midwest

ARP Reservoir Study, 1995-1997 (Hackett, 1996,1997, 1998). The ARP Reservoir
Study is part of the Acetochlor Registration Partnership agreement in which concentrations of
acetochlor, alachlor, and atrazine are being monitored in drinking water sources in 10 states, from
Iowa and Minnesota across the upper midwest and east to Delaware. The surface water portion
of the study includes sampling of treated water from 175 facilities and of untreated (raw,
infiltration gallery, and mixed water) water from 37 of those facilities. Fourteen samples were
taken in each of the sample years (1995-1997). Table 3-2 summarizes the concentrations of
atrazine in the reservoirs from which untreated water samples were taken.

Table 3-2: Atrazine Concentrations in Raw, Infiltration Gallery, and Mixed Water in Reservoirs Sampled as a
Part of the Acetochlor Registration Partnership (ARP) Study, 1995-1997.	

Site ID

City

State

Year

No.

Atrazine Concentration, jig/L











Samples

Maximum

Median

Mean1 95%UCL2

556-DA-IA

Davenport

IA

1995

14

2.3

0.2

0.4

0.6







1996

14

1.8

0.2

0.5

0.7







1997

15

1.2

0.2

0.2

0.4







All

43

2.3

0.2

0.4

0.5

574-OS-IA

Osceola

IA

1995

14

2.1

1.0

1.2

1.5







1996

14

7.7

5.5

4.3

5.7







1997

15

4.3

2.4

2.4

2.7







All

43

7.7

2.0

2.6

3.2

577-RA-IA

Centerville

IA

1995

14

2.8

2.0

2.1

2.3







1996

14

5.5

3.3

3.2

3.7







1997

15

2.5

1.7

1.8

1.9







All

43

5.5

2.0

2.3

2.6

557-DM-IA

Des Moines

IA

1995

16

1.1

0.4

0.4

0.5







1996

26

1.4

0.3

0.4

0.5







1997

28

1.0

0.2

0.2

0.3







All

70

1.4

0.2

0.3

0.4

582-WI-IA

Winterset

IA

1995

14

5.8

2.6

2.4

3.3







1996

14

5.7

1.7

2.3

3.3







1997

15

5.2

2.1

2.2

3.0







All

43

5.8

2.1

2.3

2.8

569-MI-IA

Milford

IA

1995

--

--

--

--

--







1996

14

0.3

0.2

0.2

0.3







1997

15

0.2

0.2

0.2

0.2

19


-------
Table 3-2: Atrazine Concentrations in Raw, Infiltration Gallery, and Mixed Water in Reservoirs Sampled as a
Part of the Acetochlor Registration Partnership (ARP) Study, 1995-1997.	

Site ID

City

State

Year

No.

Atrazine Concentration, jig/L











Samples

Maximum

Median

Mean1 95%UCL2







All

29

0.3

0.2

0.2

0.2

222-HI-IL

Highland

IL

1995

14

4.8

4.0

3.0

3.9







1996

14

13.4

3.7

4.9

7.0







1997

15

2.9

1.1

1.3

1.9







All

43

13.4

2.5

3.0

3.8

228-SA-IL

Salem

IL

1995

14

10.9

2.8

3.8

5.3







1996

14

74.7

2.4

15.6

26.7







1997

15

1.6

0.6

0.6

0.8







All

43

74.7

1.2

6.5

10.2

168-PA-IL

Paris

IL

1995

14

5.9

1.8

2.3

3.1







1996

14

20.8

7.3

7.3

10.3







1997

15

3.9

1.5

1.5

2.1







All

43

20.8

1.9

3.7

4.8

170-AL-IL

Altamont

IL

1995

14

1.4

0.8

0.9

1.0







1996

14

7.0

5.5

4.2

5.5







1997

15

4.1

1.8

1.9

2.2







All

43

7.0

1.5

2.3

2.9

197-EL-IL

Elgin

IL

1995

14

0.2

0.1

0.1

0.1







1996

14

1.2

0.1

0.2

0.4







1997

15

1.0

0.1

0.2

0.3







All

43

1.2

0.1

0.2

0.2

244-SP-IL

Sparta

IL

1995

14

4.3

0.9

1.2

1.9







1996

14

3.4

1.8

1.6

2.1







1997

15

9.3

1.8

3.4

5.0







All

43

9.3

1.3

2.1

2.7

245-OL-IL

Olney

IL

1995

14

5.8

3.6

3.3

4.3







1996

14

3.7

2.9

2.5

3.0







1997

15

3.0

2.4

2.1

2.4







All

43

5.8

2.4

2.6

3.0

259-SP-IL

Springfield

IL

1995

14

8.5

4.6

4.3

5.7







1996

14

14.1

2.8

4.2

6.0







1997

15

1.6

0.8

0.9

1.0







All

43

14.1

2.5

3.1

3.9

268-NA-IL

Nashville

IL

1995

14

14.4

7.4

7.4

9.7







1996

14

23.6

5.8

9.4

13.0







1997

15

7.7

3.3

3.5

4.7







All

43

23.6

4.8

6.7

8.1

603-BL-IL

Hudson

IL

1995

14

3.9

1.2

1.5

2.2







1996

14

3.0

1.8

1.5

1.9







1997

15

1.8

0.5

0.6

0.8







All

43

3.9

0.8

1.2

1.5

606-KA-IL

New Athens

IL

1995

14

15.7

3.7

4.1

5.9







1996

14

17.0

4.9

5.1

7.4







1997

15

11.6

1.2

2.3

3.6







All

43

17.0

2.9

3.8

4.8

152-BR-IL

Breese

IL

1995

--

--

--

--

--







1996

13

10.7

2.1

2.7

4.2









20










-------
Table 3-2: Atrazine Concentrations in Raw, Infiltration Gallery, and Mixed Water in Reservoirs Sampled as a

Part of the Acetochlor Registration Partnership (ARP) Study, 1995-1997.	

Site ID City State Year No.	Atrazine Concentration, jig/L
	Samples Maximum Median Mean1 95%UCL2





1997

15

8.1

0.8

1.6

2.7





All

28

10.7

0.8

2.1

3.0

225-CE-IL Centralia

IL

1995

--

--

--

--

--





1996

--

--

--

--

--





1997

15

3.8

1.7

1.7

2.3





All

—

—

—

—

—

332-MC-IN Michigan City

IN

1995

14

0.1

0.1

0.1

0.1





1996

14

0.1

0.1

0.1

0.1





1997

15

0.1

0.1

0.1

0.1





All

43

0.1

0.1

0.1

0.1

344-DU-IN Dubois

IN

1995

14

1.1

0.3

0.4

0.6





1996

14

0.9

0.4

0.4

0.5





1997

15

0.6

0.4

0.3

0.4





All

43

1.1

0.4

0.4

0.5

346-SA-IN Salem

IN

1995

14

1.3

0.9

0.7

0.9





1996

14

1.1

0.6

0.6

0.8





1997

--

--

--

--

--





All

28

1.3

0.8

0.7

0.8

328-KO-IN Kokomo

IN

1995

--

--

--

--

--





1996

14

2.8

0.7

1.0

1.4





1997

15

6.3

1.0

1.6

2.5





All

29

6.3

0.8

1.3

1.8

345-RI-IN Richmond

IN

1995

--

--

--

--

--





1996

--

--

--

--

--





1997

15

15.9

4.1

5.0

7.4





All

—

—

—

—

—

3 51 -SE-IN Seymour

IN

1995

--

--

--

--

--





1996

--

--

--

--

--





1997

15

10.0

0.7

1.5

2.7





All

—

—

—

—

—

89-MI-KS Milford

KS

1995

14

5.4

1.7

2.3

3.0





1996

14

5.1

1.3

2.1

2.7





1997

15

2.8

1.0

1.3

1.7





All

43

5.4

1.3

1.9

2.2

296-SC-MN St. Cloud

MN

1995

14

0.7

0.1

0.2

0.3





1996

14

0.5

0.1

0.1

0.2





1997

15

0.1

0.1

0.0

0.1





All

43

0.7

0.1

0.1

0.1

1069-VA-MO Vandalia

MO

1995

--

--

--

--

--





1996

14

21.5

4.8

7.4

10.9





1997

15

31.3

4.9

9.0

14.1





All

29

31.3

4.9

8.2

11.2

1070-WY-MO Wyaconda

MO

1995

--

--

--

--

--





1996

14

4.6

0.2

1.1

1.9





1997

15

7.2

3.6

3.5

4.7





All

29

7.2

0.9

2.3

3.1

1009-CO-MO Concordia

MO

1995

—

—

—

—

—


-------
Table 3-2: Atrazine Concentrations in Raw, Infiltration Gallery, and Mixed Water in Reservoirs Sampled as a
Part of the Acetochlor Registration Partnership (ARP) Study, 1995-1997.	

Site ID

City

State

Year

No.

Atrazine Concentration, jig/L











Samples

Maximum

Median

Mean1 95%UCL2







1996

13

4.4

1.6

2.4

3.0







1997

15

8.0

3.4

4.3

5.8







All

28

8.0

3.1

3.4

4.3

1016-HI-MO

Higginsville

MO

1995

--

--

--

--

--







1996

--

--

--

--

--







1997

15

5.0

3.6

3.3

4.0







All

—

—

—

—

—

593-HE-PA

Hummelston

PA

1995

14

0.6

0.1

0.1

0.2







1996

14

0.7

0.1

0.1

0.2







1997

15

1.5

0.1

0.2

0.3







All

43

1.5

0.1

0.1

0.2

737-AW-PA

Norristown

PA

1995

14

0.3

0.1

0.1

0.2







1996

14

1.7

0.1

0.2

0.4







1997

15

0.2

0.0

0.0

0.1







All

43

1.7

0.1

0.1

0.2

997-WE-PA

Mechanicsburg

PA

1995

14

0.7

0.2

0.2

0.3







1996

14

0.4

0.2

0.2

0.3







1997

15

1.0

0.1

0.2

0.3







All

43

1.0

0.2

0.2

0.3

596-DE-PA

Denver

PA

1995

14

0.3

0.0

0.0

0.1







1996

14

0.2

0.0

0.0

0.1







1997

14

0.1

0.0

0.0

0.0







All

42

0.3

0.0

0.0

0.0

13-AP-WI

Appleton

WI

1995

14

0.1

0.1

0.1

0.1







1996

14

0.4

0.1

0.1

0.2







1997

15

0.7

0.1

0.2

0.2







All

43

0.7

0.1

0.1

0.2

18-OK-WI

Oshkosh

WI

1995

14

0.3

0.1

0.1

0.1







1996

14

0.7

0.1

0.2

0.3







1997

15

0.3

0.1

0.1

0.2







All

43

0.7

0.1

0.1

0.2

1	Arithmetic mean of 14 samples; with concentrations < limit of detection (LOD) set equal to the LOD.

2	Upper 95% confidence bound on the mean

The highest concentrations of atrazine occurred in reservoirs located in Illinois. The
reservoir serving Salem, IL, had the highest peak (74.7 |ig/L in 1996) and mean annual (15.6
|ig/L with a 95% upper confidence bound of 26.7 |ig/L in 1996) concentrations. Over the entire
3-year monitoring period, the reservoir serving Nashville, IL had the highest median (4.8 |ig/L)
and mean (6.7 |ig/L with a 95% upper confidence bound of 8.1 |ig/L) concentrations of atrazine.
The reservoir serving Vandalia, MO, had a 2-year median of 4.9 |ig/L and mean of 8.2 |ig/L (95%
percent upper confidence bound of 11.2 |ig/L). Peak concentrations of greater than 20 |ig/L
occurred during at least one year in 4 of the 37 reservoirs. Ten reservoirs had peak atrazine
concentrations greater than 10 |ig/L during at least one year.

USGS Midwestern Reservoir Study, 1992-93 (Scribner et al, 1996). USGS collected

22


-------
water samples from 76 reservoirs in the midwestern United States between April 1992 and
September 1993. The reservoirs were sampled 4 times in 1992 (in early spring before herbicide
application, during the first major runoff after application, after significant flushing of the reservoir
during late summer, and in early fall) and 4 times in 1993 (in early and late winter, during
midsummer, and in September). Water samples collected from the reservoir outflow were
analyzed for 11 pre-emergent herbicides and 6 metabolites. Tables 3-3 and 3-4 summarize
atrazine and metolachlor concentrations for each of the reservoirs. The 6 reservoirs which had no
recorded detects of atrazine and 23 reservoirs with no detects of metolachlor were not included in
the analysis because no determination could be made as to whether the non-detects were due to
pesticide not reaching the reservoir or to an absence of use of the pesticide in the watershed.

Table 3-3: Summary of Atrazine Concentrations in 70 Midwestern Reservoirs Sampled by the USGS in 1992-93
(Scribner et al, 1996).	

State

Reservoir

Atrazine Concentration, jig/L







Maximum

Median

Mean1

95% UCL2

IA

Rathbun Lake

3.5

2.2

2.0

2.6



Lake Panorama

1.0

0.4

0.4

0.7



Coralville Lake

3.8

0.3

0.7

1.6



Lake Red Rock

2.8

0.6

0.8

1.4



Saylorville Lake

2.4

0.3

0.6

1.2

IL

Carlyle Lake

5.8

3.3

3.2

4.5



Rend Lake

4.3

1.0

1.4

2.3



Lake Decatur

5.5

0.7

1.4

2.7



Lake Shelbyville

2.1

1.6

1.4

1.9



Lake Vermillion

5.3

1.2

2.0

3.4



Crab Orchard Lake

1.5

0.4

0.5

0.9



Little Grassy Lake

0.4

0.1

0.2

0.2



Devils Kitchen Lake

0.4

0.1

0.2

0.3

IN

Patoka Lake

0.3

0.1

0.1

0.2



Brookville Lake

2.4

1.5

1.4

1.8



Morse Reservoir

12.1

2.4

3.4

6.2



Huntington Lake

7.7

1.1

2.3

4.4



Eagle Creek Reservoir

3.6

2.5

2.1

3.0



Mississinewa Lake

10.7

4.1

4.0

6.8



Monroe Lake

0.2

0.1

0.1

0.2



Mansfield Lake

4.2

1.7

2.0

2.9



Cataract Lake

11.0

2.2

2.9

5.4



Salamonie Lake

9.8

3.4

4.2

7.0



Lake Shafer

1.5

0.3

0.5

0.9

KS

Clinton Lake

1.6

1.2

1.2

1.4



Kanopolis Lake

1.0

0.7

0.7

0.9



Milford Lake

4.0

2.0

2.4

3.1



Perry Lake

2.9

1.9

1.9

2.5



Hillsdale Lake

3.0

2.5

2.2

2.7



Waconda Lake

2.8

2.0

2.0

2.3



Pomona Lake

2.5

2.0

1.9

2.3



Tuttle Creek Lake

4.2

1.6

1.7

2.5



Wilson Lake

0.4

0.2

0.2

0.3

MN

Sandy Lake Reservoir

0.1

0.1

0.1

0.1

23


-------
Table 3-3: Summary of Atrazine Concentrations in 70 Midwestern Reservoirs Sampled by the USGS in 1992-93
(Scribner et al, 1996).	

State

Reservoir

Atrazine Concentration, jig/L







Maximum

Median

Mean1

95% UCL2



Winnibigoshish Reservoir

0.1

0.1

0.1

0.1



Leech Lake Reservoir

0.1

0.1

0.1

0.1



Gull Lake Reservoir

0.1

0.1

0.1

0.1



Lac Qui Parle Reservoir

1.6

0.2

0.4

0.9



Cross Lake

0.4

0.0

0.1

0.2

MO

Harry S Truman Reservoir

2.7

1.2

1.3

1.8



Harrisonville Lake

3.8

3.3

2.7

3.6



Smithville Lake

3.1

2.4

2.4

2.8



Long Branch Lake

2.5

1.4

1.5

1.9



Mark Twain Lake

2.5

2.1

1.9

2.3

ND

Pipestem

0.1

0.1

0.1

0.1

NE

Enders Reservoir

0.3

0.1

0.1

0.2



Cunningham Lake

1.2

0.7

0.7

0.9



Harry Strunk Lake

1.4

0.3

0.5

0.9



Hugh Butler Lake

0.7

0.4

0.4

0.5



Harlan County Lake

1.4

1.3

1.2

1.4



Swanson Lake

0.3

0.2

0.2

0.3



Branched Oak Lake

3.0

2.5

2.5

2.8



Pawnee Lake

3.3

1.8

2.0

2.4



Willow Creek

2.6

0.4

0.9

1.5

OH

Delaware Lake

5.9

1.8

2.4

3.8



Harrisonville Lake

3.8

3.3

2.7

1.5



O'Shaughnessy Reservoir

12.4

1.3

3.0

6.1



Hoover Reservoir

4.1

1.3

1.7

2.5



Milton Reservoir

1.9

0.9

0.9

1.4



Dillon Lake

4.9

0.7

1.2

2.4



Deer Creek Lake

4.1

2.0

2.0

2.8

SD

Sand Lake

0.1

0.1

0.1

0.1

WI

Lake 7746

0.7

0.2

0.2

0.4



Lake Mendota 254

0.4

0.3

0.3

0.3



Lakes Waubesa

0.3

0.2

0.2

0.2



Lake Monona

0.3

0.3

0.3

0.3



Lake Menomin 1761

0.4

0.1

0.1

0.3



Lake Wausau 4016

0.2

0.1

0.1

0.1



Spring Valley Lake

2.0

0.3

0.7

1.3



Chippewa Flowage

0.3

0.1

0.1

0.1

1	Arithmetic mean of 14 samples; with concentrations < limit of detection (LOD) set equal to the LOD.

2	Upper 95% confidence bound on the mean

Four reservoirs - Morse, Mississinewa, and Cataract in Indiana and O'Shaughnessy in
Ohio - had peak atrazine concentrations of greater than 10 |ig/L during the sample period.
Salmonie Lake, IN, and Mississinewa Lake, IN, had the highest mean and 95% upper confidence
bound concentrations (4.2 and 7.0 |ig/L and 4.0 and 6.8 |ig/L, respectively). Mississinewa Lake
had the highest median concentration (4.1 |ig/L).

24


-------
Table 3-4: Summary of Metolachlor Concentrations in 53 Midwestern Reservoirs Sampled by the USGS in 1992-

93 (Scribner et al, 1996).	

State	Reservoir	Metolachlor Concentration, jig/L





Maximum

Median

Mean1

95% UCL2

IA

Rathbun Lake

0.6

0.2

0.2

0.3



Lake Panorama

0.5

0.2

0.3

0.4



Coralville Lake

1.5

0.2

0.4

0.7



Lake Red Rock

1.6

0.4

0.5

0.9



Saylorville Lake

1.7

0.3

0.5

1.0

IL

Carlyle Lake

1.4

0.2

0.5

0.8



Rend Lake

0.1

0.1

0.1

0.1



Lake Decatur

2.8

0.4

0.8

1.4



Lake Shelbyville

1.3

0.3

0.4

0.7



Lake Vermillion

1.3

0.4

0.5

0.8



Crab Orchard Lake

0.1

0.1

0.1

0.1

IN

Brookville Lake

0.6

0.2

0.3

0.4



Morse Reservoir

5.3

0.8

1.6

2.9



Huntington Lake

4.3

0.5

1.3

2.4



Eagle Creek Res

2.3

1.3

1.3

1.9



Mississinewa Lake

4.9

1.6

1.9

3.1



Mansfield Lake

1.9

0.6

0.7

1.1



Cataract Lake

4.6

0.7

1.1

2.2



Salamonie Lake

4.3

1.6

1.8

2.8



Lake Shafer

0.4

0.1

0.2

0.3

KS

Clinton Lake

0.3

0.1

0.2

0.2



Kanopolis Lake

0.2

0.1

0.1

0.1



Milford Lake

1.4

0.4

0.5

0.8



Perry Lake

1.9

0.4

0.7

1.1



Hillsdale Lake

0.8

0.1

0.2

0.4



Waconda Lake

0.6

0.2

0.3

0.4



Pomona Lake

0.7

0.3

0.3

0.5



Tuttle Creek Lake

2.9

0.8

0.9

1.6

MN

Lac Qui Parle Res

1.2

0.1

0.2

0.5



Cross Lake

0.1

0.1

0.1

0.1

MN

Harry S Truman Res

0.3

0.1

0.2

0.2



Harrisonville Lake

1.9

0.4

0.6

1.0



Smithville Lake

0.5

0.3

0.3

0.4



Long Branch Lake

0.1

0.1

0.1

0.1



Mark Twain Lake

0.5

0.2

0.3

0.3

NE

Harry Strunk Lake

0.2

0.1

0.1

0.1



Hugh Butler Lake

0.1

0.1

0.1

0.1



Harlan County Lake

0.2

0.1

0.1

0.1



Branched Oak Lake

0.1

0.1

0.1

0.1



Pawnee Lake

0.3

0.1

0.1

0.1



Willow Creek

0.9

0.1

0.3

0.5

OH

Delaware Lake

3.1

0.6

1.1

1.9



Harrisonville Lake

1.9

0.4

0.6

0.4



O'Shaughnessy Res

6.1

0.7

1.6

3.2



Hoover Reservoir

1.1

0.4

0.5

0.7



Milton Res

0.5

0.2

0.2

0.4



Dillon Lake

2.7

0.2

0.6

1.2


-------
Table 3-4: Summary of Metolachlor Concentrations in 53 Midwestern Reservoirs Sampled by the USGS in 1992-
93 (Scribner et al, 1996).	

State	Reservoir	Metolachlor Concentration, jig/L



Maximum

Median

Mean1

95% UCL2

Deer Creek Lake

2.4

0.4

0.8

1.3

WI Lake 7746

0.1

0.1

0.1

0.1

Lake Mendota 254

0.1

0.1

0.1

0.1

Lake Monona

0.1

0.1

0.1

0.1

Lake Menomin 1761

0.2

0.1

0.1

0.1

Chippewa Flowage

0.1

0.0

0.0

0.0

1	Arithmetic mean of 14 samples; with concentrations < limit of detection (LOD) set equal to the LOD.

2	Upper 95% confidence bound on the mean

The highest peak concentration of metolachlor detected in the study occurred in
O'Shaughnessy Reservoir in Ohio at 6.1 |ig/L. Mississinewa Lake in Indiana had the highest
median (1.6 |ig/L) and mean (1.8 |ig/L with a 95% upper confidence bound of 3.1 |ig/L)
concentrations reported in the studies.

Atrazine concentrations reported in the USGS monitoring study are less than those found
in the ARP study. Several factors may explain this difference:

(1)	Length of Study: The USGS study covered a 17-month period, while the ARP data covers
3 years. The greater the time span, the more likely the study is to capture the scope of the
year-to-year variation in pesticide concentrations.

(2)	Frequency of Sampling: The ARP study collected more samples per year (at least 14-15
per year) than did the USGS study and was thus had a greater chance of capturing high
and low pesticide concentrations. Even at this frequency, it is unlikely that the ARP study
captured the true peak concentration in the sampled reservoirs.

(3)	Sample Collection Point Within the Reservoir: The ARP study collected water samples at
the water supply intake while the USGS study collected samples downstream of the
reservoirs at the outflow. Fallon (1994) observed a pesticide concentration gradient
between the reservoir inflow and outflow. The gradient changed over the season, with the
highest reservoir concentrations occurring on the upstream end (inflow) after the runoff
flush of pesticides and the lowest concentration at that time occurring at the downstream
end (outflow). As the pesticide pulse moved down the reservoir, pesticide concentrations
were diluted by the reservoir water. Depending in the location of the water supply intake
in the reservoir, pesticide concentrations could be greater than that found at the outflow.

USGS NAWQA White River Basin Study (Crawford, 1997). The White River Basin
covers 11,350 square miles of central and southern Indiana and includes six hydrogeomorphic
regions (Figure 3-1). Agriculture accounts for 70 percent of the watershed, with corn and
soybeans accounting for 78 percent of all cropland (40 percent of the basin). Corn and soybean
production is most extensive in the northern, southwestern, and southeastern portions of the

26


-------
basin. Approximately 96 percent of all agricultural pesticide use in the basin is on corn and
soybeans, with herbicide use on corn accounting for about 70 percent of that use (Crawford,
1995, 1997).

The USGS web site (Crawford, 1997) provides data on 16 pesticides in streams at the 11
sampling stations. Atrazine, metolachlor, and simazine were the three most commonly detected
pesticides throughout the entire basin (100, >99, and 92 percent detections, respectively). Table
3-5 summarizes the concentrations of these pesticides found at each sampling station. Unlike the
ARP and USGS Midwest Reservoir studies, the data for the NAWQA White River Basin study
represent pesticide concentrations from flowing water bodies. With the short residence times,
peak pesticide concentrations in streams are likely to be lower than those found in water bodies
with longer residence times, such as reservoirs.

27


-------
T	i '—"i

1 iplOA t a if- T j Murrie .

rx— '	. .. a « k

/ l l.iinil'.i'i't J I 1	° f"XC ,

ti /Yf-^rv d^i^rmuiow f
t - Vr' Madisun' r.~"	^

r^i ' ! Vj~' "';i,,,-. . i «¦

1kwne .7	f.	.	/

uwlit.1 .a ; i	jr*- -'r	_ - ,

/7f't,Mfu

?-¦	'liwuinaDois aa C. ' i	I »

^	J/	'ImiLjnapoiB j 2 j	.1

Uf .	i

/lis	; j -:**,/

K W#; "J ^ 	iX*4

v/'r/ | O^cn. Jw5> I

/ I J—-' f ¦ —'	rrf' Y.	"T	1_	J T	\

T%. 1	, "fi LM °*fa I

V _"%¦	Okwmingtcn C>r?1	% °

> ""	t ' . 4^.'' Brown i / &	Wl>, r[4

} /srr-	5i i*\

r--. < ;:'.VIf'K:i' ...	• O - v> X

¦ . i Washington r	i-> f

/	Daviess A r	

' ¦— -<0 U ~r—' A *»*

/ k-"-a 6	' *\'j

'L ,	j p—A Orange 7 I,-'

MV;

vjidsoii

Bas* from U .S. Oeo4ofliosi Suwy d&tttf data. 1:100.000.1963
a**i* Equal-Area fwfcxaon

Standard p LOD

Max. Median Mean1 UCL952
INDICATOR SITES IN THE TILL PLAIN

1

42/
-------
Table 3-5: Atrazine, Simazine, and Metolachlor Concentrations at Stream/River Sample Stations in

the White

River Basin, IN, USGS NAQWA Study (Crawford, 1997).









Map Land Use % Sample

Pesticide

Total No.,



Concentration, |ig/L



No. ag/forest/urban Dates



n > LOD

Max.

Median

Mean1

ucl952



Simazine

112/88

1.2

0.0

0.1

0.1



Metolachlor

112/112

11.6

0.2

0.8

1.1

3 83/15/1 4/94-

Atrazine

16/16

2.3

0.4

0.9

1.3

4/95

Simazine

16/16

1.2

0.1

0.2

0.3



Metolachlor

16/16

1.0

0.2

0.4

0.5

INDICATOR SITES IN THE BEDROCK LOWLAND AND PLAIN





4 98/l/
-------
Table 3-5: Atrazine, Simazine, and Metolachlor Concentrations at Stream/River Sample Stations in the White
River Basin, IN, USGS NAQWA Study (Crawford, 1997).

Map

Land Use %

Sample

Pesticide

Total No.,



Concentration, |ig/L



No.

ag/forest/urban

Dates



n > LOD

Max.

Median

Mean1

ucl952

9

71/19/8

4/94-

Atrazine

17/17

22.0

0.9

2.9

5.3





5/95

Simazine

17/17

0.9

0.1

0.2

0.3







Metolachlor

17/17

3.4

0.3

0.9

0.3

10

69/25/5

4/94-

Atrazine

16/16

8.9

0.6

1.7

2.8





5/95

Simazine

16/16

0.4

0.1

0.1

0.2







Metolachlor

16/16

3.6

0.2

0.6

1.0

11

69/22/7

5/91-

Atrazine

183/183

13.4

0.8

2.0

2.4





9/96

Simazine

183/166

1.2

0.1

0.2

0.2







Metolachlor

183/180

5.3

0.3

0.8

0.9

1	Arithmetic mean, with concentrations < limit of detection (LOD) set equal to the LOD.

2	Upper 95% confidence bound on the mean

3	Atrazine concentration of 120 |ig/L measured at this sampling site is an estimated value, outside the range of
the method calibration. Summary statistics are calculated both including this sample (top row of numbers) and
excluding the sample (bottom row).

A peak atrazine concentration of 27 |ig/L was detected at an indicator site in the till plain.
A concentration of 120 |ig/L reported at the indicator site in the glacial lowland fell outside the
range of the method calibration and was reported as an estimated value. The highest mean
concentration for atrazine (excluding the estimated value) occurred in the bedrock lowland and
plain region (4.6 |ig/L with a 95% upper confidence bound of 6.8 |ig/L). The highest detected
concentrations for simazine were found in the karst plain (peak of 7.2 |ig/L, mean of 1.2 |ig/L,
upper confidence bound of 2.2 |ig/L) while the highest detected concentrations for metolachlor
occurred in the glacial lowland (peak of 17.0 |ig/L, mean of 1.6 |ig/L, upper confidence bound of
2.3 |ig/L), Peak concentrations for atrazine in the basin samples were within the range of peak
concentrations reported for the Illinois reservoirs sampled in the ARP study.

Perry Lake, KS, 1992-3 (Fallon, 1994). Fallon (1994) conducted five seasonal surveys
of the Perry Lake reservoir (1992 pre-application, first runoff event after herbicide application,
summer, fall, and 1993 pre-application), along with monthly sampling of the main stream feeding
the reservoir, in the reservoir, and downstream below the outfall. The Perry Lake watershed is
located in the dissected glacial till plain. Land use is primarily dry-land farming, with 30-60% of
the land in crops, principally sorghum, wheat, corn, soybeans, hay. Atrazine is used on corn,
sorghum, and winter wheat.

The upstream station showed 3 peaks of atrazine in the April - August period (7 |ig/L on

30


-------
4/24/92; 26 |ig/L on 6/12; 18 |ig/L on 7/10). Atrazine concentrations dropped to <1 |ig/L in the
stream between September and the following April. In the reservoir, 1992 pre-application survey
concentrations ranged from <1 |ig/L at the upstream end to 5 |ig/L at the downstream end. After
the first flush, the highest concentrations (5-29 |ig/L) were found in the headwaters, with mixing
occurring downstream in the reservoir. During the summer, the pulse of water with the highest
concentrations (4-5 |ig/L) moved to the middle of the reservoir. By the fall, atrazine
concentrations were consistently between 3-4 |ig/L, compared to <1 |ig/L in the headwaters. By
the time of the 1993 pre-application survey, concentrations were <1 |ig/L throughout the
reservoir, among the lowest recorded levels for Perry Lake. Below the lake, at the outfall,
concentrations varied little (2-4.5 |ig/L) between April and November and decreased slowly
between November and April.

While precipitation during the study period was 26% greater than average, most of the
excess fell during March, July, September, and November 1992. Four soaking rains in late April,
at the beginning of the growing season, were followed by little precipitation in May and June.
Most of the significant rainfall occurred after July 15th, near the end of the critical period when
atrazine is most available for runoff. Rainfall in mid- to late-July accounted for half of the atrazine
loading in the reservoir. Later runoff events diluted atrazine concentrations and mass in the
reservoir and served to flush the atrazine out of the lake. A comparison of precipitation and
runoff data during the study period to historical data suggest that runoff during the critical April-
July period was lower than normal while the higher-than-normal rainfall in later months served to
dilute the atrazine concentrations in the lake and accelerate flushing of the pesticide from the
reservoir. Fallon estimated that the atrazine concentrations measured in Perry Lake were likely
lower than average.

Comparison of Monitoring Data to Modeling Results for Corn-Soybean Herbicides

in the Midwest U.S.

We used PRZM 3.12 (dated May 7, 1998) and EXAMS 2.97.5 (dated June 11, 1997) to
simulate applications of atrazine and simazine on corn and metolachlor simulations on corn and
soybeans in the midwest. Runoff from PRZM fed into an index reservoir based on Shipman City
Lake, the same reservoir used to test the index reservoir concept in the July 1998 SAP
presentation. A description of the site- and reservoir-specific inputs for the Shipman reservoir can
be found in the July 1998 FIFRA SAP document (FIFRA SAP, 1998; Background Document 3).
Although Shipman City Lake will be replaced by another midwestern reservoir to serve as an
index reservoir in modeling for drinking water assessments, it is representative of a number of
reservoirs in the central Midwest that are known to be vulnerable to pesticide contamination.

Chemical-specific inputs for atrazine, simazine, and metolachlor can be found in Table 3-6.
Model simulations used scenarios for corn and soybeans grown in the Shipman City Lake, IL,
watershed, with weather data representing Major Land Resource Area (MLRA) Ml 14. This data
is for the period 1948 to 1983 at the Louisville, KY, meteorological station W93821. The soil
data represent a Clinton silt loam (classified as a fine smectitic, mesic Chromic Vertic Hapludalf;

31


-------
SCS Hydrologic Group B). For screening evaluations of pesticides used on corn, OPP currently
uses a scenario developed for corn grown on a Hydrologic Group C soil (Cardington silt loam,
classified as a fine, illitic mesic Aquic Hapludalf) in Ohio and weather data from MLRA Ml 11.
While the overall precipitation was slightly greater with the MLRA Ml 14 data, the Ohio soil is
more prone to runoff. As a result, estimated pesticide concentrations in water simulated with the
Ohio scenario are expected to be greater than those simulated with the Illinois scenario. For
comparison, both scenarios were run for atrazine. The specific inputs for the PRZM-EXAMS
modeling are in Appendix A.

Table 3-6. Pesticide and chemical parameters for atrazine,	simazine, and metolachlor used in the simulations.

Parameter Atrazine	Simazine	Metolachlor

Pesticide Application Rate	1 at 2.24 kg/ha	1 at 3.36 kg/ha	1 at 3.87 kg/ha

Application Method	broadcast, unincorporated	broadcast, unincorporated	broadcast, unincorporated

Kd, L/kg	0.7	1.3	4.8

Aerobic soil half-life, days	146	135	67

Molecular Weight, g/mol	216	202	284

Vapor Pressure, torr	3 x 10"7	6.1 xlO"9	3 x 10"7

Henry's Law Constant, atm-m3/mol 2.6 x 10"9	3.2 x 10"2	2.6 x 10"9

Solubility, mg/L	33	3.5	53.0

Table 3-7 compares l-in-10 year values (identified as UCL) for the peak and annual mean
concentrations simulated from the modeling with monitoring data. These model values are
currently used for drinking water assessments, with the peak estimate corresponding to acute
values and the annual mean corresponding to chronic values. The table also includes median
values for the peak and annual mean. The upper 90% confidence bound based on the standard
deviation of the annual means is provided for the overall mean. The PCA adjustment is applied to
the model outputs. The model results are compared with the highest peak, median, and 95%
upper confidence bound mean concentrations from the monitoring studies. The median value
represents the "typical" or "central tendency" exposure level and the upper confidence bound of
the mean represents a "high end" exposure level. Because the data are non-normally distributed,
the median is more robust than the mean (i.e., less sensitive to extreme values) and better
represents the "central tendency."

Table 3-7: Comparison of Selected Corn/Soybean Herbicide Concentrations Estimated by PRZM/EXAMS with
and without PCA Adjustments to Available Monitoring from Selected Midwestern Studies.

Peak (jig/L)	Annual Mean (jig/L) Overall Mean (jig/L)1

Median UCL2 Median UCL2	Mean	UCL2

Atrazine Use on Corn

32


-------
Table 3-7: Comparison of Selected Corn/Soybean Herbicide Concentrations Estimated by PRZM/EXAMS with
and without PCA Adjustments to Available Monitoring from Selected Midwestern Studies.

Peak (jig/L)

Annual Mean (jig/L) Overall Mean (jig/L)1



Median

UCL2

Median

UCL2

Mean

UCL:

IL Modeling, unadjusted

15



123

6

31

10

14

IL Model, PCA-adjusted3

7



57

3

14

5

7

OH Modeling, unadjusted

34



140

11

37

16

20

OH Model, PCA-adj.3

16



64

5

17

8

9

ARP: Salem, IL



75



2

27

7

10

USGS: White River, IN



27



1

7

-

-

USGS Midwest Reserv.4



12



4

7

-

-

USGS NAQWA5



120

















Simazine Use on Corn







Modeling, unadjusted

36



174

14

70

26

34

Modeling, PCA-adjusted3

17



80

6

32

12

16

USGS: White River



7



0

2

_

_

USGS NAQWA5

Modeling, unadjusted
Modeling, PCA-adjusted3

(a)	corn-soybean comb

(b)	corn only

USGS Midwest Reserv.4
USGS: White River
USGS NAQWA5

20

Metolachlor Use on Corn and Soybeans

49

40
22

277

233
127

6
17
70

20

17
9
2
1

87

73
40
3
2

33

27
15

39

33
18

33


-------
Table 3-7: Comparison of Selected Corn/Soybean Herbicide Concentrations Estimated by PRZM/EXAMS with
and without PCA Adjustments to Available Monitoring from Selected Midwestern Studies.

Peak (jig/L)	Annual Mean (jig/L) Overall Mean (jig/L)1

Median UCL2 Median UCL2	Mean	UCL2

1	Modeling results based on 36 years of simulations, from 1948-1983; the overall mean for the ARP reservoir is
based on 3 years of data; all other studies reflect one year of data.

2	For model data, the UCL for the peak and annual mean concentrations is the l-in-10 year value which is
greater than ninety percent of the annual values; for the overall mean, it is the upper 90% confidence bound on
the mean. For the monitoring data, the peak is the highest concentration reported; the UCL for the mean is the
upper 95% confidence bound on the mean.

3	The IL model simulates runoff from a hydrologic group B soil into the index reservoir; the OH model
simulates runoff from a hydrologic group C soil into the index reservoir. Corn PCA was 0.46; soybean PCA
was 0.41; the combined PCA was 0.87.

4	Peak concentration occurred in O'Shaughnessy Reservoir, OH; median and UCL concentrations occurred in
Mississinewa Lake, IN.

5	Summary of pesticide occurrence for all 1058 surface water sites sampled as part of the NAQWA studies,
1992-1996.

The unadjusted model estimates bounded the highest reported peak concentrations for
atrazine in the ARP data. When the PCA adjustment was applied, the model estimated dropped
below the highest reported monitoring concentrations. PCA-adjusted model estimates were still
greater than reported monitoring data for metolachlor and simazine on corn.

Another way of evaluating the effectiveness of the modeling and the PCA adjustment of
the modeling data as a screening tool is to consider the number of reservoirs for which the
pesticide concentrations are less than the predicted model values. Table 3-8 summarizes the
number of reservoirs in the ARP and USGS Midwest Reservoir studies for which measured
concentrations of atrazine and metolachlor were less than those predicted by modeling. Because
of the limited sampling frequency and range of years sampled in both studies, the peak
concentrations reported in these studies do not necessarily capture the actual peak concentrations
of the pesticides in the reservoirs. The comparisons may be more relevant for longer-term
concentrations (median, which represents the central tendency of the data, or upper confidence
bound on the mean).

Table 3-8: Number of Reservoirs in the ARP andUSGS Midwest Reservoir Studies With Reported Pesticide

Concentrations Exceeding the Model Estimates in At Least 1 Year of Monitoring.

l-in-10 Yr Peak	Median Annual Mean l-in-10 Yr Ann. Mean

Not	PCA-	Not	PCA-	Not	PCA-

adjusted adjusted1 adjusted adjusted1 adjusted adjusted1

Atrazine

Model estimate, ng/L, IL	123	57	6	3	31	14

34


-------
Model est., |ig/L. OH

140

64

11

5

37

17

ARP study vs. IL est.

0/37

1/37

2/37

12/37

0/37

2/37

ARP study vs. OH est.

0/37

1/37

0/37

3/37

0/37

1/37

USGS midwest study

0/70

0/70

0/70

5/70

0/70

0/70





Metolachlor







Model estimate, |ig/L

277

127

20

9

87

40

USGS midwest study

0/53

0/53

0/53

0/53

0/53

0/53

1 PCA of 0.46 for corn was used for both atrazine and metolachlor.

Currently, OPP uses l-in-10 year peak and annual mean model estimates as screening
values to determine whether the pesticide passes the screen or further evaluation is necessary.
Table 3-8 shows that, for the reservoirs where monitoring data are available, PCA adjustment of
the means would result in underpredictions of peak atrazine concentrations in 1 of 37 ARP study
reservoirs and of mean annual atrazine concentrations in 2 of 37 ARP study reservoirs using the
IL scenario. The OH scenario, which used a soil that is more prone to runoff, resulted in slightly
higher model estimates and covered all but one of the reservoirs in the ARP study. The PCA
adjustment resulted in no underpredictions for either atrazine or metolachlor in the USGS
midwest study reservoirs.

Modeling and Monitoring In the Central Valley of California

The Central Valley of California is one of the richest agricultural areas in the United
States. Crops grown in the area range from artichokes and alfalfa to watermelons and walnuts.
The San Joachin River, which runs north through the valley to its mouth in the San Francisco Bay,
is also one of the few areas where minor crops are grown and surface water samples have been
taken.

We chose five crops (alfalfa, almonds, grapes, apricots and walnuts) and three pesticides
(diazinon, simazine and chlorpyrifos) for our comparison. These were chosen because the crops
are relatively major crops in the area (although they still occupy less than 10% of the land area)
and relatively major pesticides and because the pesticides were analyzed for in the NAWQA
program. Of the three pesticides, chlorpyrifos is used on alfalfa, almonds and walnuts, diazinon is
used on almonds and apricots, and simazine is used on almonds, grapes and walnuts. The crops
chosen are the highest use crops for each chemicals, representing the majority of the use of the
three chemicals in the area.

35


-------
A

Figure 3-2. NAWQA sampling locations (#) in the San Joachin River valley, California. Hydrologic unit
boundaries are shown as thicker lines.

Monitoring Data for the San Joachin River, CA

Samples were taken as part of the USGS's NAWQA program and come from fourteen
sites in the San Joachin valley. They were taken mostly in 1993 and 1994, and primarily in the
winter months, which are the rainy season in California. The sampling locations are mostly in
Stanislaus County, and nearly all fall into a single hydrologic unit (HUC 18040002). A map of the
sample locations is shown in Figure 3-2. As a result, we calculated PCAs for HUC 18040002 and
used these as multipliers for PRZM/EXAMS modeling results.

The sampling sites themselves are mostly on or near the San Joachin river and the number

36


-------
of samples taken at each site range from 1 to
100, as shown in Figure 3-3.

Modeling of Minor-Use Pesticides

in Central Valley, CA

Crop and soil information used in the
model runs was based on telephone
conversations with extension agents from
Stanislaus, Merced, Kern and San Joachin
counties. Pesticide application times used in
the model runs were based on actual
application times but extended to the
maximum number of applications permitted
by the label. Similarly, application rates were
the maximum label rates. The amount of each
pesticide used in the models was therefore
higher than what farmers in the Central Valley
actually use. For the models runs, simazine
had one application in January for all three
crops for which it is applied, diazinon was
applied once (in January) on almonds, and
three times (January, May and July) on
apricots. Chlorpyrifos was allowed the most
applications (according to the label) of the three, being applied three times (march, June and July)
on alfalfa, three times (February, July and August) on walnuts and four times (January, February,
March and July) on almonds. Spray drift estimates for the models were based on a preliminary
spray drift simulation program. These numbers were somewhat lower than in farm pond
simulations because not all land in the central valley is adjacent to a water body and in the model
all spray drift directly enters the water.

Comparison of Monitoring Data and Modeling Results in the Central Valley, CA

An accurate comparison between these modeling and monitoring data is difficult. The
monitoring data was taken over a period of less than four years and had only 368 samples. The
modeling, by contrast covered a period of 36 years with over 13,000 simulated days from which
to draw the simulated peak concentrations. The comparison here, however is made with the 90th
percentile value for the peak concentration, used by EFED for estimated environmental
concentrations (EECs).

An overall summary of modeling results with a comparison to peak monitoring values, is
given in Table 3-9. For each pesticide, the results of modeling each crop were multiplied by the
crop's PC A and then summed to give an overall estimate of pesticide concentration in the

San Joachin River valley.

37


-------
hydrologic unit. From this initial look at the data, using a PCA to modify modeling results seems
to have a useful moderating effect for Chlorpyrifos and Simazine. For both of these chemicals the
PCA reduces EEC's significantly, but to a level above their highest observed concentrations by a
factor or two or more. In the case of diazinon, however, multiplying by the PCA decreased the
estimated concentrations to the same range as monitoring results, and lower than four monitoring
values.

Table 3-9. Overall estimates of the concentration (|ig/L) of three pesticides used on minor crops in Central
Valley, CA, with and without a PCA adjustment.

Diazinon	Chlorpyrifos	Simazine

Highest single crop value 53	56	98

All crops modified by PCA 2.7	2.7	10

Highest monitoring value 3.8	0.3	5.3

Time series of modeling for diazinon is shown in Figure 3-4. In order to incorporate
irrigation into the PRZM modeling, the irrigation water was added as extra summer precipitation
according to a schedule provided by Central Valley Cooperative Extension agents. In the case of

Diazinon on Almonds and Apricots

¦ PEAK A 21 DAY ~ 90 DAY
96 HOUR A 60 DAY YEARLY

Q. 5

Q.

C .

0	4

'E

1	3

o	n

o O	F	n

°	An	A a"

tta

$ $ 5 ~ , ^ ... $ ~ ~ 4 ^ ... $ ~	^ ^ ^ ^ * v ^ ^	¦

... .;. ...
0 '	1	1	1	1	1	1	1—

1950 1960	1970 1980

Year

Figure 3-4. Time series plots of PCA-adjusted estimates of diazinon concentrations from use on almonds and
apricots.

38


-------
diazinon and chlorpyrifos, which are applied to the tree foliage, it means the model simulates a
much greater amount of pesticide being washed off leaves than would occur under real flood
irrigation, where the water is applied at the base of the trees. Winter concentrations predicted by
the model are therefore likely to be somewhat lower than shown in Figure 3-4, and it is possible
that if the comparison was done without irrigation, modeling values would be lower in
comparison. Simazine is applied once per year, in the winter, on the crops considered here, so it
shows none of the problems associated with modeling irrigation.

Calculation of PCAs in California

As an example, we calculated PCA values for HUC 18040002 both using and not using
land cover data. For calculations without land cover data, a map of the hydrologic unit looks like
Figure 2-1. The calculation proceeds as follows. For each county in the hydrologic unit, we
calculate the fraction of the county's area in the unit. This fraction is then multiplied by the
acreage of the crop in the county, to get an estimate of the total area of the county's crop inside
the hydrologic unit. These areas, calculated for each county, are then summed and divided by the
total area of the hydrologic unit to get an estimate of the fraction of cropped area in the unit. A
summary of these calculations is shown in Table 3-10. To refine the calculation using land cover
data, we look only at the cropped area of each county when dividing the county's crops between
the portion inside and outside of the hydrologic unit. For the particular hydrologic unit considered
here, the western portion of each county is neither cropped nor inside the unit, making the
calculated crop percentages higher inside. A summary of the calculations using land cover data is
shown in Table 3-11.

The PCAs were used to modify modeling results by multiplying the final PRZM/EXAMS
results for each crop by the PCA for that crop and then summing over all the crops.

Table 3-10. Calculation of PCA in the San Joaquin River Hydrologic Unit (HU) without land cover
modifications.

County

Fraction of
County in HU

Almonds

Estimated acreage of crop
Apricots

in hydrologic unit
Grapes

Walnuts

Stanislaus

0.73

61,019

4,675

9,791

17,866

San Joachin

0.26

10,203

1,179

14,187

8,871

Merced

0.12

8,599

178

1,500

950

Calaveras

0.11

5

0

31

78



PCA:

0.067

0.005

0.021

0.023

39


-------
Table 3-11. Calculation of PCA in the San Joaquin River Hydrologic Unit (HU) with land cover modifications.

County	Fraction of	Estimated acreage of crop in hydrologic unit

County's

orchards in HU	Almonds	Apricots	Grapes	Walnuts

Stanislaus 1.00	83,011	8,692	39,206	24,306

SanJoachin 0.39	15,004	1,733	20,848	13,044

Merced 0.39	28,884	599	5,037	3,192

Calaveras 0.00	0	0	0	0

PCA:	0.107	0.007	0.033	0.034

40


-------
Section 4: Preliminary Comparison of Screening Model Results With
Available Surface Water Monitoring Data

In order to evaluate the effectiveness of the PCA as a screening model adjustment beyond
the preliminary comparisons in Section 3, we must first evaluate the assumption that the existing
screening models work equally for these other pesticides. In other words, the PRZM-EXAMS
model overpredicts in the same manner and fashion for other pesticides as it does for the
pesticides evaluated here. While OPP has not yet undertaken a full, rigorous comparison of
modeled estimates with available monitoring concentrations for all of the pesticides we have
evaluated, a preliminary survey suggests that the models do not consistently overpredict pesticide
concentrations found in monitoring studies.

Neither PRZM nor the linkage that empties runoff from PRZM into the water body
modeled in EXAMS have been fully validated. Our attempt here is not to validate the models per
se but to evaluate their effectiveness as a screening tool. To be effective screening tools, the
models should generate appropriately conservative estimates of pesticide concentrations in water
that do not underestimate pesticide concentrations that could actually be found in reservoirs
vulnerable to high pesticide runoff while not generating unrealistic estimates that are so high that
pesticides that do not pose an actual concern or true risk pass the screening evaluation.

For screening purposes, the model scenarios and inputs were selected to represent what
OPP believes to be high-end exposure conditions. The models use maximum pesticide application
rates, pesticide environmental fate properties that reflect highest mobility and longest persistence,
and site characteristics that are conducive to runoff. In the comparisons made below, the model
estimates reflect runoff from a 100% crop area, all treated at the same time, into the smaller (20
million liter) edge-of-field water body. The monitoring data reflect pesticide applications that
range from the minimum amount required for efficacy to maximum label rates, treatment on less
than 100% of the drainage area at varying times, with runoff into water bodies located at varying
distances from the treated fields.

The summaries below are brief capsules taken from more detailed water assessments
which go into more depth on modeling parameters, monitoring study conditions, and uncertainties
in the data. This is not a complete listing of water assessments, but only those for which Tier 2
(PRZM-EXAMS) or, in a couple of instances, Tier 1 (GENEEC) modeling estimates have been
compared with available monitoring data. The intent is to get a picture as to how often our
modeling, using the current small edge-of-field pond, overestimates pesticide concentrations
found in monitoring data and how often the modeling estimates are similar to monitoring
concentrations. In all of the water assessments listed below, the conclusion of the OPP scientist
evaluating the data was that the available monitoring data was not sufficient to base a drinking
water assessment.

Pesticides For Which Model and Monitoring Concentrations are Within the Same Order of
Magnitude:

41


-------
Aldicarb: In an initial screening assessment for Aldicarb (S. Dutta, 3/31/99), Tier 1
GENEEC model results for aldicarb were compared to STORET data. Tier 2 (PRZM/EXAMS)
modeling had not been conducted and a more extensive review of monitoring data was not
available for this project. The maximum peak modeled concentration of aldicarb estimated by
GENEEC was 88 |ig/L for aldicarb use on citrus, pecans, and sugar beets. Modeled peak
concentrations for other uses ranged from 13 to 54 |ig/L. Longer-term average model estimates
(56 days) ranged from 3 to 19 |ig/L. Detections of aldicarb and its degradates were reported in
27 surface water samples in 9 states. Monitoring concentrations ranged from 0.5 to 29 |ig/L.
The Tier 1 peak model concentrations are only 3 times greater than the limited monitoring data
used in this assessment, while the longer-term average model concentrations were less than the
highest reported monitoring level.

EPTC: The range of peak EPTC EECs generated by PRZM/EXAMS (6 to 57 ng/L)
corresponds reasonably well with the range of highest surface water concentrations of EPTC
observed (without considering data limitations or modeling short comings) in the monitoring
studies (10 to 40 |ig/L) (J. Wolf, 3/25/99) The estimated 90th percentile upper-bound value for
the annual mean ranged from 0.2 to 3.8 |ig/L. The mean of the annual means ranged from 0.16 to
3.40 |ig/L. Available monitoring from at least thirty states indicates that EPTC concentrations in
surface water are generally very low (generally less than 0.05 |ig/L, but rarely greater than 1
|ig/L), However, EPTC concentrations in surface water monitoring have occasionally been
detected at levels up to approximately 40 |ig/L.

Methidathion: Tier 2 PRZM/EXAMS modeling for California, where 90-95% of
methidathion is used, estimated peak concentrations of 3 to 6 |ig/L and annual average
concentrations of 0.2 to 0.6 |ig/L (J. Lin, 12/30/98). Methidathion was detected in 11 of 25
samples from the San Joaquin River, at concentrations ranging from less than 1 |ig/L to 15 |ig/L,
with an average of less than 3 |ig/L. The model estimates were similar to slightly lower than the
limited monitoring results. Although STORET showed no detects (limits of detection ranging
from 0.5 to 10 |ig/L) in 264 drinking water samples from California, only 5 samples represented
surface water sources.

Methomyl: Peak methomyl concentrations estimated from Tier 2 PRZM/EXAMS
screening models, which ranged from 30 |ig/L (use on lettuce) to 99 |ig/L (peaches), were
comparable to peak concentrations found in environmental monitoring studies conducted by the
registrant, which ranged from 2 |ig/L in a pond near a corn field in Illinois to 175 |ig/L in a stream
adjacent to a corn field in Georgia (N. Thurman, 8/7/97). Differences between modeled estimates
and monitoring results increased to roughly an order of magnitude or greater when 21-day
average concentrations were compared. While the water bodies sampled did not represent
drinking water sources, comparisons of the monitoring data with PRZM/EXAMS runs made by
the registrant simulating actual site and weather conditions found that the model predictions were
similar in magnitude and pattern of dissipation to the observed values.

Triallate: Tier 2 PRZM/EXAMS modeling estimated peak concentrations of 2 |ig/L for

42


-------
triallate and 2.4 |ig triallate equivalents/L for the cumulative triallate residues (triallate + TCPSA)
from fall application on winter wheat and 2.5 |ig/L and 3.5 |ig/L, respectively, from spring
application on spring wheat (A. Al-Mudallal and J. A. Hetrick, 3/17/99). The annual average
concentrations for triallate and cumulative residues, respectively, were 0.1 and 0.2 |ig/L for
winter wheat and 0.3 and 0.8 |ig triallate equivalents/L for spring wheat. The highest
concentrations of triallate detected in the USGS NAWQA monitoring program were 0.65 |ig/L in
the Central Columbia Plateau and 0.28 |ig/L in the Red River Basin. The median concentration
from 209 detects in the NAQWA program was 0.01 |ig/L. The NAWQA study units with the
highest frequency of triallate detections and highest concentrations of triallate, Red River of the
North Basin and Central Columbia Plateau, correspond to high triallate use areas (>11 lbs ai per
square mile). The reported maximum concentrations from the NAWQA are within a factor of 3
to 4 of the peak triallate concentrations estimated with PRZM/EXAMS modeling. While OPP
had sufficient monitoring data to assess the parent triallate, it did not have monitoring information
on the toxic degradate of concern, trichloropropene sulfonic acid (TCPSA). The PRZM/EXAMS
model estimates for TCPSA exceed the drinking water level of concern.

Pesticides For Which Model Estimates Are More Than An Order of Magnitude Greater
than Monitoring Concentrations:

Butylate: Tier 1 modeling with GENEEC estimated a peak concentration for butylate of
33 |ig/L and a 56-day average concentration of 30 |ig/L (J. Breithaupt, 8/18/98). These values
were compared to USGS NAQWA data, which reported 295 detections in 5,193 surface water
samples. Most detections (61 %) were in the White River in Indiana, which includes significant
corn production. The detections ranged from 0.002 to 1.4 |ig/L, with 3 detections of >1 |ig/L, 17
detections of 0.1-1 |ig/L, and 275 detections of 0.002-0.1 |ig/L. The Tier 1 modeled
concentrations were approximately 30 times greater than the maximum concentration found in the
NAQWA study. Laboratory data suggest that butylate dissipates primarily by volatility from soil
and water. The difference between model estimates and monitoring data is at least in part due to
the fact that the GENEEC model does not account for volatility of the pesticide. Also, it is
unclear if the monitoring data represent time periods with maximum usage and highest runoff.

Chlorpyrifos: Tier 2 PRZM/EXAMS model estimates for chlorpyrifos generated peak
concentrations of 41 |ig/L for chlorpyrifos use on tobacco (range of peak values from 11 to 41
|ig/L) and 90-day average concentrations of 2 to 7 |ig/L (M. Barrett, 11/20/98). The highest
concentration of chlorpyrifos detected in several stream monitoring studies was 0.4 |ig/L, an
order of magnitude lower than the longer-term model estimates and two orders lower than peak
estimates. No monitoring data were available for small reservoirs. All of the monitoring data
represent dissolved chlorpyrifos, while significant additional residues of this lipophilic pesticide
are likely to occur in sediment and suspended solids. Therefore, these estimates apply only to
drinking water exposure potential.

Halosulfuron: Tier 2 PRZM/EXAMS modeling of the application of this pesticide on
sugarcane, cotton, and fallow land generated estimated peak concentrations of 2 to 4 |ig/L and

43


-------
average annual concentrations of 0.3 to 1 |ig/L (J. Carleton, 3/30/99). Although little monitoring
data is available, a preliminary study of surface waters of in the midwest detected halosulfuron in
7 of 130 river water samples (LOD = 10 ppt, or 0.01 |ig/L) at concentrations up to 67 ppt (0.067
|ig/L) (communication with W. Battaglin, USGS, Denver, CO).

Methyl Parathion: Methyl parathion concentrations estimated from Tier 2
PRZM/EXAMS modeling was considerably greater for cotton use (peak of 214 |ig/L and long-
term mean of 7 |ig/L) than for any of the other agricultural uses modeled (peaks ranged from 4
|ig/L for alfalfa to 39 |ig/L for corn, with long-term means of <2 |ig/L) (K. Costello, 3/11/99).
The maximum concentration of methyl parathion found in agricultural streams in the USGS
NAWQA study was 0.3 |ig/L. However, the analytical recovery for methyl parathion in the
NAWQA study is 46% (SD=13%), which limits extensive quantitative interpretation of the
monitoring data. A search of other monitoring data found methyl parathion concentrations of up
to 6 |ig/L in a California DPR study of rice herbicides. These concentrations dropped in later
studies after CDPR took steps to reduce the concentration of rice pesticides in surface water.
Although the monitoring data reported for methyl parathion are lower than modeling results, they
don't necessarily reflect the use scenarios most vulnerable to contamination. For instance, the
CDPR monitoring of the Colusa Basin Drain is targeted to methyl parathion use on rice. It
includes sampling which coincides with times of application, but the maximum rate at which
methyl parathion is applied to rice is one quarter of the maximum rate applied to cotton, with
fewer applications annually. In addition, retention of water on treated fields is a mitigation
measure relevant only to rice, and not other crops to which methyl parathion is applied.

Phorate: Maximum estimated concentrations of phorate from Tier 2 PRZM/EXAMS
modeling ranged from 2 |ig/L for wheat in North Dakota to 23 |ig/L for cotton in Mississippi
(Breithaupt, 10/3/97). Parent phorate is not persistent in water and the estimated chronic
concentrations for all modeled crops was <1.0 |ig/L, except for cotton at 1.2-2.1 |ig/L. Parent
phorate was not found above 0.6 |ig/L in surface monitoring data from Colorado. However, the
amount of monitoring data is very limited. Also, the monitoring data do not assess the more
persistent and mobile sulfone and sulfoxide degradates.

Terbufos: Maximum concentrations of parent terbufos from PRZM/EXAMS modeling
were 4 |ig/L for sugar beets, 5 |ig/L for corn, and 22 |ig/L for grain sorghum (D. Farrar and J.
Breithaupt, 2/6/98). Terbufos is moderately persistent in water, and the estimated chronic
concentration for corn was 1 |ig/L. Terbufos was not found above 2.25 |ig/L in monitoring data
from the Midwest. However, the monitoring data are limited, and the data quality is unknown
for some of the information. Also, the available monitoring data do not determine the more
persistent and mobile sulfone and sulfoxide degradates of parent terbufos.

44


-------
References

Allord, G.J. 1992. 1 to 2,000,000 Hydrologic Unit map of the Conterminous United States (ed.
1.1.2), U.S. Geological Survey, Reston, Virginia, USA.
http: //water. usgs. gov/1 ookup/getcover? huc2m

Crawford, C.G. 1995. Occurrence of Pesticides in the White River, Indiana, 1991-1995. U.S.
Geol. Surv. Fact Sheet 233-95. USGS, Indianapolis, IN.

Crawford, C.G. 1997. National Water Quality Assessment Program: White River Basin Study.
Web site: http://www-dinind.er.usgs.gov/nawqa/wrnawqa.htm .

Fallon, J.D. 1994. Determining the three-dimensional distribution, transport, and relative age of
atrazine and selected metabolites in Perry Lake, Kansas. M.S. Dissertation. Kansas State
University, Manhattan, KS.

Hackett, A.G. 1996. Surface Drinking Water Monitoring Program for Acetochlor and Other Corn
Herbicides: First Year Sampling and Results. Acetochlor Registration Partnership. St.
Louis, MO. MRID 439243-01.

Hackett, A.G. 1997. Surface Drinking Water Monitoring Program for Acetochlor and Other Corn
Herbicides: Second Year Sampling and Results. Acetochlor Registration Partnership. St.
Louis, MO. MRID 442995-01.

Hackett, A.G. 1998. Surface Drinking Water Monitoring Program for Acetochlor and Other Corn
Herbicides: Third Year Sampling and Results. Acetochlor Registration Partnership. St.
Louis, MO. MRID 111.

Kellogg, R.L., S. Wallace, and K. Alt. 1997. Potential Priority Watersheds for Protection of

Water Quality from Nonpoint Sources Related to Agriculture. Poster Presentation at the
52nd Annual Soil and Water Conservation Society Conf., Toronto, Ontario, July 22-25,
1997 (Rev. October 7, 1997). http://www.nhq.nrcs.usda.gov/land/pubs/wqpost2.html .
Map also available through EPA's "Surf Your Watershed" web page, "Index of
Watershed Indicators" http://www.epa.gov/surf/iwi/dbackl2a.html.

Lanfear, K.J. 1994. Counties and county equivalents in the conterminous United States (ed.
2.3.1), U.S. Geological Survey, Reston, Virginia, USA.
http: //water. usgs. gov/1 ookup/getcover? county 2m

Scribner, E.A., D.A. Goolsby, E.M. Thurman, M.T. Meyer, and W.A. Battaglin. 1996.

Concentrations of selected herbicides, herbicide metabolites, and nutrients in outflow from
selected midwestern reservoirs, April 1992 through September 1993. U.S. Geol. Surv.
Open-File Report 96-393. Prepared as part of the Toxic Substances Hydrology Program.

45


-------
Laurence, KS.

46


-------
Listing of Background Documents

1.	Use of a CAF in Estimating Surface-Water-Source Drinking Water Exposure, presented
to the Dec. 10, 1997 FIFRA SAP Meeting, Arlington, VA.

2.	Response from the FIFRA SAP on the Dec. 10, 1997 document^ Set of Science Issues
Being Considered in Connection with Estimating Drinking Water Exp. as a Component
of the Dietary Risk Assessment.

3.	Proposed Methods for Basin-Scale Estimates of Pesticide Concentrations in Flowing
Water and Reservoirs for Tolerance Reassessment, presented at the July 29, 1998 FIFRA
SAP Meeting, Arlington, VA.

4.	Response from the FIFRA SAP on July 29, 1998 document Proposed Methods for Basin-
Scale Estimates of Pesticide Concentrations in Flowing Water and Reservoirs for
Tolerance Reassessment.

5.	Science Policy 5: Estimating The Drinking Water Component of a Dietary Exposure
Assessment.

47


-------
Appendix A: PRZM Model Inputs
Inputs for Corn-Soybean Herbicide Simulations in the Midwest U.S.

Table A-l. Selected PRZM 3.12 input parameters for the Midwestern U.S. corn/soybean scenarios

Parameter

IL Corn

IL Soybeans

OH Corn

Pan Evaporation Factor (PFAC)

Weather Factors

0.74

0.74

0.72

Erosion and Landscape Factors

USLE K Factor (USLEK), tons EI1*
* EI = 100 ft-tons * in/ acre*hr

0.42

0.42

0.37

USLE LS Factor (USLELS)

1.0

1.0

0.43

USLE P Factor (USLEP)

1.0

1.0

0.50

Field Area (AFIELD), ha

172

172

172

Land Slope (SLP)

6.0

6.0

6.0

Hydraulic Length (HLP)

464

464

354

Initial Surface Condition (ISCOND)

Maximum rainfall interception storage of
crop (CINTCP), cm

Maximum Active Root Depth (AMAXDR),
cm

Maximum Canopy Coverage (COVMAX),
%

Soil Surface Condition After Harvest
(ICNAH)

Date of Crop Emergence
(EMD, EMM, IRYEM)

Date of Crop Maturity
(MAD, MAM, IYRMAT)

Date of Crop Harvest
(HAD, HAMJYRHAR)

Maximum canopy height (HTMAX), cm

Crop Parameters

1 (fallow)
0.25

90

100

3 (residue)

May 16

Sept. 16

Oct. 1

300

1 (fallow)
0.25

90

100

3 (residue)

May 16

Sept. 16

Oct. 1

300

CN, C Factor and Manning's N sychronized with crop dates for fallow/cropped/residue

1 (fallow)
0.25

90

100

3 (residue)

May 16

Sept. 26

Oct. 11

100
covers

48


-------
Table A-l. Selected PRZM 3.12 input parameters for the Midwestern U.S. corn/soybean scenarios

Parameter

IL Corn

IL Soybeans

OH Corn

SCS Curve Number (CN)

86 / 78 / 82

86 / 78 / 82

91/85/88

USLE C Factor (USLEC)

0.50/0.50/0.50

0.50/0.50/0.50

0.50/0.25 /0.30

Mannings N (MNGN)

.04 / .04 / .04

04 / .04 / .04

0.02/0.02/0.02



Soil Parameters





Soil Series

Clinton

Clinton

Cardington

Soil Classification

fine, smectitic,
mesic Chromic
Vertic Hapludalf

fine, smectitic,
mesic Chromic
Vertic Hapludalf

fine, illitic, mesic
Aquic Hapludalf

Total Soil Depth (CORED), cm

178

178

100

Number of Horizons (NHORIZ)

4

4

2

First Sc

il Horizon (HORIZN

= 1)



Horizon Thickness (THKNS), cm

13

13

22

Bulk Density (BD), g -cm"3

1.3

1.3

1.6

Initial Water Content (THETO), cm3-H20
•cm3-soil

0.421

0.421

0.294

Compartment Thickness (DPN), cm

0.1

0.1

0.2

Field Capacity (THEFC), cm3-H20 -cm3-soil

0.421

0.421

0.294

Wilting Point, cm3-H20 -cm3-soil

0.201

0.201

0.086

Organic Carbon Content, % (w/w)

2.80

2.80

1.16

Second £

>oil Horizon (HORIZN

= 2)



Horizon Thickness (THKNS), cm

27

27

78

Bulk Density (BD), g -cm"3

1.30

1.30

1.65

Initial Water Content (THETO), cm3-H20
•cm3-soil

0.421

0.421

0.147

Compartment Thickness (DPN), cm

5

5

1

Field Capacity (THEFC), cm3-H20 -cm3-soil

0.421

0.421

0.147

Wilting Point, cm3-H20 -cm3-soil

0.201

0.201

0.087

Organic Carbon Content, %

1.067

1.067

0.174

Third Soil Horizon (HORIZN = 3)

49


-------
Table A-l. Selected PRZM 3.12 input parameters for the Midwestern U.S. corn/soybean scenarios

Parameter

IL Corn

IL Soybeans

OH Corn

Horizon Thickness (THKNS), cm

60

60



Bulk Density (BD), g -cm"3

1.35

1.35



Initial Water Content (THETO), cm3-H20
•cm3-soil

0.451

0.451



Compartment Thickness (DPN), cm

5

5



Field Capacity (THEFC), cm3-H20 -cm3-soil

0.451

0.451



Wilting Point, cm3-H20 -cm3-soil

0.251

0.251



Organic Carbon Content, %

0.40

0.40



Horizon Thickness (THKNS), cm
Bulk Density (BD), g -cm"3

Fourth Soil Horizon (HORIZN
78

Initial Water Content (THETO), cm3-H20
•cm3-soil

Compartment Thickness (DPN), cm
Field Capacity (THEFC), cm3-H20 -cm3-soil
Wilting Point, cm3-H20 -cm3-soil
Organic Carbon Content, %

1.40
0.416

5

0.416
0.216
0.167

= 4)

78
1.40
0.416

5

0.416
0.216
0.167

Table A-2. Selected PRZM 3.12 input parameters for the California scenarios

Parameter

Almonds Walnuts

Alfalfa

Apricots

Grapes



Weather Factors







Pan Evaporation Factor (PFAC)

0.85 0.85

0.85

0.85

0.85

Erosion and Landscape Factors

USLE K Factor (USLEK), tons EI1*
* EI = 100 ft-tons * in/ acre*hr

USLE LS Factor (USLELS)

USLE P Factor (USLEP)

Field Area (AFIELD), ha

Land Slope (SLP)

0.05

0.05

0.05

0.05

0.05

0.01

0.01

0.01

0.01

0.01

0.1

0.1

0.1

0.1

0.1

172

172

172

172

172

0.5

0.5

0.5

0.5

0.5

50


-------
Table A-2. Selected PRZM 3.12 input parameters for the California scenarios

Parameter

Almonds

Walnuts

Alfalfa

Apricots

Grapes

Hydraulic Length (HLP)

464

464

464

464

464

Crop Param

eters









Initial Surface Condition (ISCOND)

3 (residue)

3 (residue)

3 (residue)

3 (residue)

3 (residue)

Maximum rainfall interception storage of
crop (CINTCP), cm

0.3

0.3

0.25

0.3

0.25

Maximum Active Root Depth (AMAXDR),
cm

60

60

180

60

90

Maximum Canopy Coverage (COVMAX),
%

90

90

100

90

100

Soil Surface Condition After Harvest
(ICNAH)

1 (fallow)

1 (fallow)

1 (fallow)

1 (fallow)

3 (residue)

Date of Crop Emergence
(EMD, EMM, IRYEM)

March 15

March 15

Sept. 15

March 15

April 7

Date of Crop Maturity
(MAD, MAM, IYRMAT)

May 15

May 15

April 15

May 15

June 30

Date of Crop Harvest
(HAD, HAMJYRHAR)

Nov. 1

Nov. 1

Sept. 1

Nov. 1

Oct. 31

Maximum canopy height (HTMAX), cm

1500

1500

82

1500

150

CN, C Factor and Manning's N sychronized with crop dates for
fallow/cropped/residue covers

SCS Curve Number (CN)

86 / 59 / 82

86 / 59 / 82

86 / 59 / 82

86 / 59 / 82

85/88

USLE C Factor (USLEC)

.05/ .05/

.05/ .05/

.02/.02/.40

.05/ .05/

0.25/0.30



.05

.05



.05



Mannings N (MNGN)

.02/.02/.02

.02/.02/.02

.15/. 15/. 15

.02/.02/.02

.02/.02/.02

Soil Paramt

:ters









Soil Series

Kimberlina

Kimberlina

Kimberlina

Kimberlina

Hanford

Soil Classification

Coarse-loa

Coarse-loa

Coarse-loa

Coarse-loa

Coarse-loa



my, mixed,

my, mixed,

my, mixed,

my, mixed,

my, mixed,



superactive, superactive, superactive, superactive,

nonacid,



calcareous,

calcareous,

calcareous,

calcareous,

thermic



thermic

thermic

thermic

thermic

Typic



Typic

Typic

Typic

Typic

Xerorthents



Torriorthen

Torriorthen

Torriorthen

Torriorthen





ts

ts

ts

ts



51


-------
Table A-2. Selected PRZM 3.12 input parameters for the California scenarios

Parameter

Almonds

Walnuts

Alfalfa

Apricots

Grapes

Total Soil Depth (CORED), cm

125

125

185

125

150

Number of Horizons (NHORIZ)

3

3

3

3

3

First Soil Horizon (T

ORIZN = 1)









Horizon Thickness (THKNS), cm

5

5

5

5

30

Bulk Density (BD), g -cm"3

1.45

1.45

1.45

1.45

1.5

Initial Water Content (THETO), cm3-H20
•cm3-soil

0.212

0.212

0.212

0.212

.222

Compartment Thickness (DPN), cm

0.1

0.1

0.1

0.1

0.1

Field Capacity (THEFC), cm3-H20 -cm3-soil

0.212

0.212

0.212

0.212

0.125

Wilting Point, cm3-H20 -cm3-soil

0.097

0.097

0.097

0.097

0.075

Organic Carbon Content, % (w/w)

0.80

0.80

0.80

0.80

0.75

Second Soil Horizon (

HORIZN = 2)









Horizon Thickness (THKNS), cm

20

20

20

20

60

Bulk Density (BD), g -cm"3

1.45

1.45

1.45

1.45

1.5

Initial Water Content (THETO), cm3-H20
•cm3-soil

0.212

0.212

0.212

0.212

0.210

Compartment Thickness (DPN), cm

5.0

5.0

5.0

5.0

1.0

Field Capacity (THEFC), cm3-H20 -cm3-soil

0.212

0.212

0.212

0.212

0.12

Wilting Point, cm3-H20 -cm3-soil

0.097

0.097

0.097

0.097

0.075

Organic Carbon Content, %

0.80

0.80

0.80

0.80

0.2

Third Soil Horizon (I

IORIZN = 3)









Horizon Thickness (THKNS), cm

100

100

160

100

60

Bulk Density (BD), g -cm"3

1.45

1.45

1.45

1.45

1.5

Initial Water Content (THETO), cm3-H20
•cm3-soil

0.212

0.212

0.212

0.212

0.200

Compartment Thickness (DPN), cm

0.1

0.1

0.1

0.1

5.0

Field Capacity (THEFC), cm3-H20 -cm3-soil

0.212

0.212

0.212

0.212

0.10

Wilting Point, cm3-H20 -cm3-soil

0.097

0.097

0.097

0.097

0.075

Organic Carbon Content, %

0.80

0.80

0.80

0.80

0.125

52


-------
PRZM Input File for Atrazine on Corn, IL Scenario

PRZM 3.1 Input Data File, ILCORNB1.INP; Created June 14, 1998 ***
Modified PRZM program; Multiple Scenario/Stream-Reservoir Model ***
Location: South Western Illinois, Macoupin Cty. 150,000 acres corn
Modeled basin 427 ac.; Corn intensity: 115 ac. (27% 1997 county data)
Manning's N values for cornstalk residue fallow surface 2 ton/acre

Soil series: Clinton
Typical pedon slope:

Chemical
Location:

0.74
4

0.42
1
1
1

1609

Name:Atrazine
IL Crop
0.30

1605

50
04

1. 00

0.25
3

0110

0.50 0.50

0. 04
36

100548

100549

0. 04

210948

210949

corn;

0

silt loam; Hydrologic group B soil

3%; range 0-25%. Assume 6%

MLRA 114

16. 00

1.00 172.00

90.00 100.00

211048

211049

. 60

3 86 78

5.00

464.0

82

0.00 300.00

100583 210983
Application: 1

36

ATRAZINE; KD:
*** Record 16:

050548	0 4

050549	0 4

211083
of

1

apps ot 3.87 kg a.i./ha, BROADCAST meth.of app.,

10	0

0.73 (LOAM); ASM Tl/2 = 146 dayS; AnSM Tl/2 = 159 days
Application information; set specific to chemical ***
0.10 2.240 0.99 0.01
0.10 2.240 0.99 0.01

99 eff, 1% drift

050583 0 4 0.10 2.240 0.99 0.01
*** Record 17: Filtra., disposit. foliar pest, after harvest,

0.0	0 0.00

Soil Series: Clinton silt loam; 6

178.00	0 0 0 0

*** Record 26: Soil volatilization

and plant uptake

slope; Hydrologic Group B

0 0 0
constants

0

0. 00
4
1

0. 00

00. 00

0

WATR

1
1

7

YEAR







PRCP

TSER

0

0



RUNF

TCUM

0

0 1.

. 7El0

INFL

TSER

1

1



ESLS

TSER

0

0

1.E3

RFLX

TSER

0

0

1.E5

EFLX

TSER

0

0

1.E5

RZ FX

TSER

0

0

1.E5

0

13.000

1 . 300

0

.421

0

. 000

0.

000

0. 000

1.6E-3

1.6E-3

0

. 000











0.100

0. 421

0

.201

2

.800

00

.73



27.000

1. 300

0

.421

0

. 000

0.

000

0. 000

1.6E-3

1. 6E-3

0

. 000











5. 000

0. 421

0

.201

1

. 067

00

.73



60.000

1. 350

0

.451

0

. 000

0.

000

0. 000

1.6E-3

1. 6E-3

0

. 000











5. 000

0. 451

0

. 251

0

.400

00

.73



78 . 000

1.400

0

.416

0

. 000

0.

000

0. 000

1. 6E-3

1. 6E-3

0

. 000











5. 000

0.416

0

. 216

0

.167

00

.73



YEAR

10



PEST

YEAR



10

CONC

10

53


-------
PRZM Input File for Atrazine on Corn, OH Scenario



PRZM3 Input File, OHCORN1.inp converted 1/16/98 SWA***

OHCORN1.inp created 2/13/98***

*** Modeler: S. Abel ***

Manning's N values for cornstalk residue, fallow surface, 1 ton/acre***
Application method by broadcast incorporated 3 days before planting***
Cardington silt loam is not one of the benchmark soils ***

Benchmark soils include: blount; crosby; pewamo; miami; brookston; glynwood ***
miamian; morley; bennington; and fincastle ***

ATRAZINE



Location: OH Crop: corn

0

0.50

0.72 0.30
4

0.37
1
1
1

1605 1609 1110
0.50 0.25 0.30
0.02 0.02 0.02
36

160548	260948 111048

160549	260949 111049

0.43

0.25
3

MLRA 111
15 . 00

172.00

90.00 100.00

5 .80

91 85

6.00

464 . 0

0.00 100.00

160583 260983 111083	1

Application: 1 apps of 2.24 kg a.i./ha, BROADCAST meth.of app., 0 100 eff, 0i

36	1	0	0

ATRAZINE; KD: 0.73 (LOAM); ASM Tl/2 = 146 days; AnSM Tl/2 = 159 days
*** Record 16: Application information; set specific to chemical ***

100548	0 4 0.01 2.240 0.99 0.01

100549	0 4 0.01 2.240 0.99 0.01

drift

100583 0 4 0.01 2.240 0.99 0.01
0.	0 0.00

Soil Series: Cardington silt loam; Hydrogic Group C

100.00
0.00
2
1

0.00

22.000
1.6E-3
0.200
78.000
1.6E-3
1.000

0 0

00.00

1.	600
1.6E-3

0.294

1.	650
1.6E-3

0.147

0

0

0.294
0.000
0.086
0.147
0.000
0.087

0

YEAR



10



1









1









7

YEAR







PRCP

TCUM

0

0



RUNF

TCUM

0

0



INFL

TCUM

1

1



ESLS

TCUM

0

0

1.E3

RFLX

TCUM

0

0

1.E5

EFLX

TCUM

0

0

1.E5

RZFX

TCUM

0

0

1.E5

0

0

0.000

1.160
0.000

0.174

YEAR

0

0

0.000

0.73
0.000

0.73

10

0

0.000

0.000

YEAR

10

PRZM Input File for Simazine on Corn

*** PRZM 3.1 Input Data File, ILCRNSMZ.INP; Created June 14, 1998 ***
*** Modified 3/29/99 for use with shipman reservior and simazine ***

54


-------
***	Modified PRZM program; Multiple Scenario/Stream-Reservoir Model ***

***	Location: South Western Illinois, Macoupin Cty. 150,000 acres corn ***

***	Modeled basin 427 ac.; Corn intensity: 115 ac. (27% 1997 county data) ***

***	Manning's N values for soybean residues small grain moderate stand ***

***	Soil series: Clinton silt loam; Hydrologic group B soil ***

***	Typical pedon slope: 3%; range 0-25%. Assume 6% ***

***	this assumes continuous corn cropping ***

Chemical Name:Simazine

Location:

IL Crop:

: corn;

MLRA

. 114



0.74

A

0.30

0

16.

00

1 1

1i

0.42

1

1.00

1.00

172

. 0

6.60 3 6.00 464.0

1

1

0.25

90.00

100.

00

3 86 78 82

1

3









1605 1609

0110









0.50 0.50

0.50









0.04 0.04

0.04









36











100548

210948

211048



1



100549

210949

211049



1



100550

210950

211050



1



100551

210951

211051



1



100552

210952

211052



1



100553

210953

211053



1



100554

210954

211054



1



100583

210983

211083



1



Application Schedule: 1 ,

apps .

0 3.0

lb/acre, GROUND SPRAY 0 100%

36

1

0



0



SIMAZINE

Kd:1.2 7

(sandy

loam)

; ASM:

: Tl/2 = 135 days; AnSM: Tl/2

0.00 300.00

eff. w/0% drift
= 270 days

*** Record 16: Application information; set specific to chemical ***

050548	0 4 0.10 3.360 0.99 0.01

050549	0 4 0.10 3.360 0.99 0.01

050583
*** Record

0.0
Soil Serie

178.00
*** Record
0.00
4

0 4 0.10 3.360 0.99 0.01
17: Filtra., disposit. foliar pest, after harvest, and plant uptake ***
0 0.00

s: Clinton silt loam; 6% slope; Hydrologic Group B
0.00 000000000
26: Soil volatilization constants ***

0.00 00.00

1

13.000

1.300

0.421

0.000

0.000

0.000



5.13E-3

2.57E-3

0.000









0.100

0.421

0.201

2 .800

1.27



2

27.000

1.300

0.421

0.000

0.000

0.000



5.13E-3

2.57E-3

0.000









5 . 000

0.421

0.201

1.067

1.27



3

60.000

1.350

0.451

0.000

0.000

0.000



5.13E-3

2.57E-3

0.000









5 . 000

0.451

0.251

0.400

1.27



4

78.000

1.400

0.416

0.000

0.000

0.000



5.13E-3

2.57E-3

0.000









5 . 000

0.416

0.216

0.167

1.27



0

WATR

1
1

YEAR

10

PEST

YEAR

10

7

YEAR





PRCP

TSER

0

0

RUNF

TCUM

0

0 1.7E10

INFL

TSER

1

1

ESLS

TSER

0

0 1.E3

CONC

YEAR

10

55


-------
RFLX TSER
EFLX TSER
RZFX TSER

0

0

1.E5

0

0

1.E5

0

0

1.E5

PRZM Input File for Metolachlor on Corn

***	PRZM 3.1 Input Data File, ILCORNB1.INP; Created June 14, 1998 ***

***	Modified PRZM program; Multiple Scenario/Stream-Reservoir Model ***

***	Location: South Western Illinois, Macoupin Cty. 150,000 acres corn ***

***	Modeled basin 427 ac.; Corn intensity: 115 ac. (27% 1997 county data) ***

***	Manning's N values for cornstalk residue fallow surface 2 ton/acre ***

***	Manning's N values for soybean residues small grain moderate stand ***

***	Soil series: Clinton silt loam; Hydrologic group B soil ***

***	Typical pedon slope: 3%; range 0-25%. Assume 6% ***

Chemical Name:METOLACHLOR

Location:

0.74
4

0.42
1
1
1

1605 1609 0110
0.50 0.50 0.50
0.04 0.04
36

100548

100549

IL Crop: corn;

0.30

1.00

0.25
3

0.04

210948

210949

0

MLRA 114
16.00

1.00 172.00

90.00 100.00

211048

211049

6. 60

6.00

464 . 0

86 78 82

0.00 300.00

100583 210983 211083	1

Application: 1 apps of 3.87 kg a.i./ha, BROADCAST meth.of app., 0 100 eff, 0i

36	1	0	0

METOLACHLOR; KD: 4.81 (SANDY LOAM); ASM Tl/2 = 67 days; AnSM = 81 days
*** Record 16: Application information; set specific to chemical ***

050548	0 4 0.10 3.870 0.99 0.01

050549	0 4 0.10 3.870 0.99 0.01

drift

0.10 3.870 0.99 0.01
0 0.00

Clinton silt loam; 6% slope; Hydrologic Group B

0

050583 0 4
0.0

Soil Series:

178.00	00000000

*** Record 26: Soil volatilization constants ***
0.00
4
1

0

WATR

1
1
7

PRCP
RUNF

0.00

00.00









13.000

1.300

0.421

0.000

0.000

0.000

1. 03e-2

3.6e-03

0.000







0.100

0.421

0.201

2 .800

O
CO



27.000

1.300

0.421

0.000

0.000

0.000

1. 03e-2

3.6e-03

0.000







5 . 000

0.421

0.201

1.067

O
CO



60.000

1.350

0.451

0.000

0.000

0.000

1. 03e-2

3.6e-03

0.000







5 . 000

0.451

0.251

0.400

O
CO



78.000

1.400

0.416

0.000

0.000

0.000

1. 03e-2

3.6e-03

0.000







5 . 000

0.416

0.216

0.167

O
CO



YEAR

10

PEST

YEAR

10

CONC

YEAR
TSER
TCUM

YEAR

10

1.7E10

56


-------
INFL	TSER

ESLS	TSER

RFLX	TSER

EFLX	TSER

RZFX	TSER

1

1



0

0

1.E3

0

0

1.E5

0

0

1.E5

0

0

1.E5

PRZM Input File for Metolachlor on Soybeans

***	PRZM 3.1 Input Data File, ILSOYBB1.INP; Created June 15, 1998 ***

***	modified 23 Apr 1999 with new planting/harvesting/spraying dates ***

***	to be more representative of Illinois conditions reported by USDA***

***	Modified PRZM program; Multiple Scenario/Stream-Reservoir Model ***

***	Location: South Western Illinois, Macoupin Cty. 148,000 acres soybean ***

***	Modeled basin 427 ac.; Soybean intensity: 111 ac. (26% 1997 county data) ***

***	Manning's N values for soybean residues small grain moderate stand ***

***	Manning's N value for wheat is small grain moderate stand ***

***	Soil series: Clinton silt loam; Hydrologic group B soil ***

***	Typical pedon slope: 3%; range 0-25%. Assume 6% ***

***	Assume a three year rotation of soybean-soybean-wheat ***

***	See ILSOYBBl.wpd for scenario details (pending) ***

Chemical Name:METOLACHLOR

Location: IL Crop: soybeans MLRA 114

0.74
4

0.33
1
1
1

0.30

1.00

0.20
3

0

16.00

1.00 172.00

22.00 100.00

6. 60

6.00

464 . 0

86 78 82

0.00

90.00

1605 1609	0110

0.50 0.50	0.50

.023 .023	.023
36

250548	101048

250549	101049

250948

250949

250583 101083 250983	1

Application: 1 apps of 3.46 lbs a.i./a, BROADCAST meth.of app., 0 99 eff, It

36	1	0	0

Metolachlor; KD: 4.81 (SANDY LOAM); ASM Tl/2 = 67 days; AnSM Tl/2 = 81 days

190548	0 4 0.10 3.870 0.99 0.01

190549	0 4 0.10 3.870 0.99 0.01

drift

0.0
jil Ser:
178.00
0.00
4
1

0

WATR

1

0 4 0.10

3.870

0.99 0.01







0

0.00









.es: Clinton silt

loam; 6%

slope;

Hydrologic Group



0 0

0 0

0 0

0 0

0

0.00

00.00









13.000

1.300

0.421

0.000

0.000

0.000

1.03e-2 8

.6e-03

0.000







0.100

0.421

0.201

2 .800

o

CO



27.000

1.300

0.421

0.000

0.000

0.000

1.03e-2 8

.6e-03

0.000







5 . 000

0.421

0.201

1.067

O
CO



60.000

1.350

0.451

0.000

0.000

0.000

1.03e-2 8

.6e-03

0.000







5 . 000

0.451

0.251

0.400

O
CO



78.000

1.400

0.416

0.000

0.000

0.000

1.03e-2 8

.6e-03

0.000







5 . 000

0.416

0.216

0.167

O
CO



YEAR

10

PEST

YEAR

10

CONC

YEAR

10

57


-------
1

7

YEAR







PRCP

TSER

0

0



RUNF

TSER

0

0



INFL

TSER

1

1



ESLS

TSER

0

0

1.E3

RFLX

TSER

0

0

1.E5

EFLX

TSER

0

0

1.E5

RZFX

TSER

0

0

1.E5

58


-------
Inputs for Simulations in the San Joachin River, CA

PRZM Input File for Chlorpyrifos on Walnuts

*	* * PRZM 3.1 Input File Modified from PRZM 2 File***

*	* *CAAL.INP created 12/23/98***

***Crop: Walnuts, also applies to Pistachios and other nuts grown in region***
***Kern Co, CA; Sacramento and San Joaguin Valleys, MLRA 17***

***Based on mature trees approximately 50 feet tall with sparse grass understory** *

*	* *Chemical is SIMAZINE applied AERIAL. IRRIGATION APPLIED. MET FILE MODIFIED***
*** TALKED TO EXTENSION AGENT IN KERN CO. ***

*** cropping curve number reduced from 78 to fit the 15% of flood irrigation ***
*** water which runs off. The 15% number comes from Terry Pritchard, ***
*** San Joachin county cooperative extension, (209) 468-2085 ***

CHLORPYRIFOS

Kimberlina Sandy Loam; Hyrologic Group B

0.852
4

0.05
1
1
1

0101 0104
0.05 0.05
.023 .023
36

150348

150349

0.450

0.01

0.30
3

0112
0.05
. 023

150548

150549

0 20.000

0.1

172 . 0

60.0 90.000

011148

011149

3.!

0.5

464 . 0

86 59 82

0.0 1500.0

150383 150583 011183
Application Schedule: 3 APPS.

108

CHLORPYRIFOS :
030148 0 2

1	0

Kd: 68.7; AeSM:
0.0 4.480 0.95

AERIAL at 4.0 lbs a.i./acre,

0

Tl/2 = 77 days; AnSM: Tl/2 = 154 days
0.05

95% eff. w/ 5% drift

170748

o

N)
O

, 0

4 .480

0.95

0.05







170848

O
N)

O

, 0

4 .480

0.95

0.05







030183

0 2 0.

, 0

4 .480

0.95

0.05







180783

0 2 0.

, 0

4 .480

0.95

0.05







180883

O
N)

O

, 0

4 .480

0.95

0.05







0.00

l



0.0











0.00

0.0000



0.50











Kimberlina

Sandy

Loam; Hydrologic

Group B;





125 . 0





0 0

0

0

0 0

0 0

0

0.00

O

0.00



0.00











1

5 . 0



1.45

0.

212

0.0

0.0

0



0.009

4

.5E-03

0.

00









0.1



0.212

. 0973

0.80

68.7



2

20.0



1.45

0.

212

0.0

0.0

0



0.009

4

.5E-03

0.

00









5 . 0



.2240

. 0973

0.80

68.7



3

100.0



1. 65

0.

211

0.0

0.0

0



0.009

4

.5E-03

0.

00







n

5 . 0



0.2020

0.0962

0.80

68.7



u

YEAR



10





YEAR

10



1
1
7

PRCP
RUNF

YEAR
TCUM
TCUM

0.000

0.000

0.000

YEAR

10

59


-------
INFL

TCUM

1

1



ESLS

TCUM

0

0

1.0E3

RFLX

TCUM

0

0

1.0E5

EFLX

TCUM

0

0

1.0E5

RZFX

TCUM

0

0

1.0E5

PRZM Input File for Diazinon Use on Almonds

*	* * PRZM 3.1 Input File Modified from PRZM 2 File***

*	* *CAAL.INP created 12/23/98***

***Crop: Almonds, also applies to Pistachios and other nuts grown in region***
***Kern Co, CA; Sacramento and San Joaguin Valleys, MLRA 17***

***Based on mature trees approximately 50 feet tall with sparse grass understory** *

*	* *Chemical is SIMAZINE applied AERIAL. IRRIGATION APPLIED. MET FILE MODIFIED***
*** TALKED TO EXTENSION AGENT IN KERN CO. ***

*** cropping curve number reduced from 78 to fit the 15% of flood irrigation ***
*** water which runs off. The 15% number comes from Terry Pritchard, ***
*** San Joachin county cooperative extension, (209) 468-2085 ***

DIAZINON

Kimberlina Sandy Loam; Hyrologic Group B

0.450

0.852
4

0.05
1
1
1

0101 0104 0112
0.05 0.05 0.05
.023 .023 .023
36

150348

150349

0 20.000

0.01

0.30
3

150548

150549

0.1

172 . 0

60.0 90.000

011148

011149

3.!

0.5

464 . 0

86 59 82

0.0 1500.0

150383 150583 011183	1

Application Schedule: 1 app. 0 3.0 lb/acre, Aerial, 95%,5%

36	1	0	0

DIAZINON Kd: 4.0 (loamY SAND); ASM: Tl/2 = 38 DAYS; AnSM: Tl/2

100148	0 2 0.00 3.360 0.95 0.05

100149	0 2 0.00 3.360 0.95 0.05

34 days

100182

100183
0.00
0.00

0 2 0.00
0 2 0.00
1

0.0000

3.360
3.360
0.0
0.50

0.95 0.05
0.95 0.05

Kimberlina Sandy Loam; Hydrologic Group B;

125 . 0

1
1
7

PRCP

0

0

0

0

0

0

00

O



0.00



0.00









1



5 . 0



1.45

0.212

0.0

0.0

0.000



i.

, 80E-2

1.

80E-2

0.00











0.1



0.212

. 0973

0.80

4 . 00



2



20.0



1.45

0.212

0.0

0.0

0.000



i.

, 80E-2

1.

80E-2

0.00











5 . 0



.2240

. 0973

0.80

4 . 00



3



100.0



1. 65

0.211

0.0

0.0

0.000



i.

, 80E-2

1.

80E-2

0.00







n



5 . 0

0

.2020

0.0962

0.80

4 . 00



u



YEAR



10



YEAR

10



YEAR
TCUM

YEAR

10

60


-------
RUNF

TCUM

0

0



INFL

TCUM

1

1



ESLS

TCUM

0

0

1.0E3

RFLX

TCUM

0

0

1.0E5

EFLX

TCUM

0

0

1.0E5

RZFX

TCUM

0

0

1.0E5

PRZM Input File for Chlorpyrifos Use on Alfalfa

***	PRZM 3.1 Input Data File ***

***	CAALF.INP; created March 18, 1999 ***

***	Emergence, and maturity dates from Stanislaus country extension. ***

***	Harvest date from Extension Office Homepage for Stanislaus County ***

***	Central Valley, Stanislaus County, California ***

***	weather Metl7.met ***

***	Mannings N values for small grain, across slope, moderate stand ***

***	Modified 26 March 1999 by Ian Kennedy. Changed to 12 3-year ***

***	cropping periods. R9 changed to correct format. INCLUDES IRRIGATION ***

***	cropping curve number reduced from 72 to fit the 15% of flood irrigation ***

***	water which runs off. The 15% number comes from Terry Pritchard, ***

***	San Joachin county cooperative extension, (209) 468-2085 ***

CHLORPYRIFOS

Location:

Kimberlina samdy

loam,

MLRA

17;

Stanislaus

County, CA,

alfalfa

0.852

A

0.450

0

20.00



1

3





1i

0.05

-1

0.01

0.10

172.00

3.

80

1

0.50 464.0



1

1

0.25

180.0

100.00



1

86 59 82



0.00

1

3















0101 1009

1509















0.02 0.02

0.40















0.15 0.15

. 015















12

















150947

150448

010950

1











150950

150451

010953

1











82 . 0

150980 150481 010983
Application Schedule: 4 APPS.
drift

1	0

Kd: 68.7; AeSM:
0.0 1.120 0.99

144

CHLORPYRIFOS :
170348 0 2
020648 0 2
020748 0 2
020848 0 2

0.0
0.0
0.0

ground spray at 1.0 lbs a.i./acre, 99% eff. w/ 1£,

1.120 0.99
1.120 0.99
1.120 0.99

0

Tl/2
0.01
0.01
0.01
0.01

77 days; AnSM: Tl/2 = 154 days

170383	0	2	0.0	1.120	0.99	0.01

020683	0	2	0.0	1.120	0.99	0.01

020783	0	2	0.0	1.120	0.99	0.01

020883	0	2	0.0	1.120	0.99	0.01

0.0	3	0.0

0.00

0.000

0.5









KIMBERLINA

SAMDY

loam; Hydrologic

Group B





185 . 0



0 0

0 0

0 0

0 0

0

0.00

O

0.00

0.00









0

1

5 . 0

1.45

0.212

0.0

0.0

0.000



0.009

4.5E-03

0.00









0.1

0.212

. 0973

0.80

68.7



2

20.0

1.45

0.212

0.0

0.0

0.000



0.009

4.5E-03

0.00









5 . 0

.2240

. 0973

0.80

68.7



3

160.0

1. 65

0.211

0.0

0.0

0.000

61


-------
0.009 4.5E-03 0.00

68.7

10	YEAR	10 1

1
1

7

YEAR







PRCP

TCUM

0

0



RUNF

TCUM

0

0



INFL

TCUM

1

1



ESLS

TCUM

0

0

1.0E3

RFLX

TCUM

0

0

1.0E5

EFLX

TCUM

0

0

1.0E5

RZFX

TCUM

0

0

1.0E5

5.0 0.2020 0.0962 0.80
YEAR	10	YEAR

PRZM Input File for Diazinon Use on Apricots

*	* * PRZM 3.1 Input File ***

*	* *CAAPRICO.INP created 03/18/99***

***Crop: Apricots - information on maturity incomplete ***

***Stanislaus County, Central Valley, CA, MLRA 17, IRRIGATION INCLUDED***

***Based on mature trees approximately 15 feet tall with sparse grass understory** *

DIAZINON

Kimberlina Sandy Loam; Hyrologic Group B

0.852
4

0.05
1
1
1

0101 0104
0.05 0.05
.023 .023
36

150348

150349

0.450

0.01

0.30
3

0112
0.05
. 023

150548

150549

0 20.000

0.1

172 . 0

60.0 90.000

011148

011149

25.25

464 . 0

86 59 82

0.0 450.0

150382

150383

150582

150583

011182

011183

Application Schedule: 3 apps.

108

DIAZINON Kd: 4.
100148 0 2
100548 0 2
100748 0 2

1	0

0 (loamY SAND),
0.0 2.240 0.95
0.0 2.240 0.95
0.0 2.240 0.95

1
1

0 2.0 lb/acre, AERIAL, 95%,5%

0

ASM: Tl/2 = 38 DAYS; AnSM: Tl/2
0.05
0.05
0.05

34 days

100183

0

2 0.

,0 2.240

0.95 0.05







110583

0

2 0.

,0 2.240

0.95 0.05







110783

0

2 0.

,0 2.240

0.95 0.05







0.00



1

0.0









0.00

0

. 0000

0.50









Kimberlina

Sandy

Loam; Hydrologic

Group B;





125 . 0





0 0

0 0

0 0

0 0

0

0.00

O



0.00

0.00









1



5 . 0

1.45

0.212

0.0

0.0

0.000



1.

80E-2

1.80E-2

0.00











0.1

0.212

. 0973

0.80

4 . 0



2



20.0

1.45

0.212

0.0

0.0

0.000



1.

80E-2

1.80E-2

0.00











5 . 0

.2240

. 0973

0.80

4 . 0



3



100.0

1. 65

0.211

0.0

0.0

0.000

62


-------
1.

80E-2

CO

OE-2

0.00



5 . 0

0.

2020

0.096:

0

YEAR



10



1









1 -









7

YEAR







PRCP

TCUM

0

0



RUNF

TCUM

0

0



INFL

TCUM

1

1



ESLS

TCUM

0

0

1.0E3

RFLX

TCUM

0

0

1.0E5

EFLX

TCUM

0

0

1.0E5

RZFX

TCUM

0

0

1.0E5

0.80	4.0

YEAR	10	YEAR

63


-------
Appendix B: Output Data From PRZM/EXAMS Simulations

Table B-l: Results of 36 years of PRZM/EXAMS modeling for atrazine use on corn in the midwest, IL scenario.

Estimated Atrazine Concentrations, |ig/L

Pcak/Y early



PCA-Adj



PCA-Adj



PCA-Adj

3-Yr Avg

Probability

Peak

Peak

Yearly

Yearly

3-Yr Avg

3-Yr Avg

Probability

0.03

309

142

82

38







0.05

133

61

38

17

47

22

0.03

0.08

123

57

31

14

40

18

0.06

0.11

123

57

30

14

38

17

0.09

0.14

98

45

30

14

17

8

0.11

0.16

69

32

25

12

17

8

0.14

0.19

46

21

11

5

16

7

0.17

0.22

44

20

11

5

16

7

0.20

0.24

30

14

10

5

15

7

0.23

0.27

30

14

8

4

13

6

0.26

0.30

27

12

8

4

13

6

0.29

0.32

25

11

8

3

13

6

0.31

0.35

24

11

7

3

12

6

0.34

0.38

22

10

7

3

12

6

0.37

0.41

21

10

7

3

7

3

0.40

0.43

19

9

6

3

6

3

0.43

0.46

19

9

6

3

6

3

0.46

0.49

17

8

6

3

6

3

0.49

0.51

13

6

6

3

6

3

0.51

0.54

12

5

6

3

6

3

0.54

0.57

12

5

4

2

5

2

0.57

0.60

11

5

4

2

4

2

0.60

0.62

10

5

4

2

4

2

0.63

0.65

9

4

3

1

4

2

0.66

0.68

8

4

2

1

4

2

0.69

0.70

8

4

2

1

4

2

0.71

0.73

7

3

2

1

4

2

0.74

0.76

6

3

2

1

3

2

0.77

0.78

3

2

2

1

3

2

0.80

0.81

3

1

1

1

3

1

0.83

0.84

3

1

1

1

3

1

0.86

0.87

3

1

1

1

2

1

0.89

0.89

2

1

1

0

1

1

0.91

0.92

1

1

0

0

1

1

0.94

0.95

1

0

0

0

1

0

0.97

0.97

0

0

0

0







1 in 10 Year















Probability 1

123

57

30

14

25

13



Median

15

7

6

3







Overall Mean





10

5







(36 yr)2





14

7







1 l-in-10 year probability values are used in screening assessments.

2 Mean of the annual values and 90% upper bound on the mean, respectively.

64


-------
Table B-2: Results of 36 years of PRZM/EXAMS modeling for
atrazine use on corn in the midwest, OH scenario.	



Estimated Atrazine Concentrations, |ig/L

Pcak/Y early



PCA-adj



PCA-adj

Probability

Peak

Peak

Yearly

Yearly

0.03

231

67

106

31

0.05

175

44

80

20

0.08

162

41

75

19

0.11

129

35

59

16

0.14

121

35

56

16

0.16

110

32

51

15

0.19

106

28

49

13

0.22

103

27

47

12

0.24

102

27

47

12

0.27

82

23

38

11

0.30

69

20

32

9

0.32

63

18

29

9

0.35

50

15

23

7

0.38

45

13

21

6

0.40

43

13

20

6

0.43

41

12

19

5

0.46

39

11

18

5

0.49

38

11

17

5

0.51

30

10

14

5

0.54

29

10

13

5

0.57

25

10

11

5

0.60

23

10

10

4

0.62

22

9

10

4

0.65

22

9

10

4

0.68

21

9

10

4

0.70

17

6

8

3

0.73

17

6

8

3

0.76

16

5

8

2

0.78

14

5

6

2

0.81

14

5

6

2

0.84

14

5

6

2

0.86

14

4

6

2

0.89

12

4

6

2

0.92

10

4

5

2

0.95

7

4

3

2

0.97

6

3

3

1

1 in 10 Year









Probability 1

139

64

37

17

Median

34

16

11

5

Overall Mean





17

8

(36 yr)2





20

9

1	l-in-10 year probability values are used in screening assessments.

2	Mean of the annual values and 90% upper bound on the mean,
respectively.

65


-------
Table B-3: Results of 36 years of PRZM/EXAMS modeling for Simazine use on corn in the midwest, IL scenario.

Estimated Simazine Concentrations, |ig/L

Pcak/Y early
Probability

Peak

PCA-Adj
Peak

Yearly

PCA-Adj
Yearly

3-Yr Avg

PCA-Adj
3-Yr Avg

3-Yr Avg
Probability

0.03

460

212

173

80







0.05

208

96

82

38

105

49

0.03

0.08

184

85

80

37

92

42

0.06

0.11

170

78

66

30

78

36

0.09

0.14

147

68

61

28

41

19

0.11

0.16

136

63

55

25

40

19

0.14

0.19

103

47

35

16

39

18

0.17

0.22

75

34

31

14

39

18

0.20

0.24

64

30

28

13

36

17

0.23

0.27

58

27

25

12

34

16

0.26

0.30

57

26

23

11

33

15

0.29

0.32

51

23

22

10

32

15

0.31

0.35

47

22

21

10

30

14

0.34

0.38

46

21

20

9

30

14

0.37

0.41

43

20

19

9

20

9

0.40

0.43

39

18

19

9

20

9

0.43

0.46

38

17

18

8

18

8

0.46

0.49

37

17

14

6

18

8

0.49

0.51

35

16

13

6

17

8

0.51

0.54

33

15

13

6

16

7

0.54

0.57

31

14

13

6

16

7

0.57

0.60

31

14

12

6

14

7

0.60

0.62

26

12

12

5

14

6

0.63

0.65

23

11

11

5

12

5

0.66

0.68

23

10

11

5

12

5

0.69

0.70

22

10

9

4

12

5

0.71

0.73

20

9

9

4

11

5

0.74

0.76

19

9

8

4

10

5

0.77

0.78

12

6

6

3

9

4

0.80

0.81

12

6

6

3

9

4

0.83

0.84

12

5

6

3

9

4

0.86

0.87

8

4

4

2

8

4

0.89

0.89

7

3

4

2

6

3

0.91

0.92

6

3

3

1

6

3

0.94

0.95

6

3

3

1

4

2

0.97

0.97

1

0

0

0







1 in 10 Year
Probability 1
Median
Overall Mean
(36 yr)2

174

36

80
17

70
14
26
34

32
6
12
16







1	l-in-10 year probability values are used in screening assessments.

2	Mean of the annual values and 90% upper bound on the mean, respectively.

66


-------
Table B-4: Results of 36 years of PRZM/EXAMS modeling for metolachlor use on corn and
soybeans in the midwest, IL scenario.

Pcak/Yearly	Estimated Simazine Concentrations, |ig/L

Probability Corn	Soybean PCA-Adj Corn	Soybean PCA-Adj

Peak	Peak	Peak	Yearly	Yearly	Yearly

0.03

318

313

275

99

97

85

0.05

309

284

248

91

85

76

0.08

285

270

237

90

83

73

0.11

274

251

234

85

80

70

0.14

260

224

223

81

66

69

0.16

221

219

191

66

65

57

0.19

180

120

132

55

47

42

0.22

131

114

100

52

47

42

0.24

115

105

99

48

41

41

0.27

107

94

92

48

38

41

0.30

96

93

82

34

33

29

0.32

84

83

73

33

32

28

0.35

70

67

60

27

26

23

0.38

68

67

59

25

25

22

0.41

68

67

59

24

24

21

0.43

64

63

55

24

24

21

0.46

51

47

42

21

21

18

0.49

49

47

42

21

20

18

0.51

49

44

38

20

17

15

0.54

44

43

38

20

17

15

0.57

43

42

38

18

16

15

0.60

43

42

37

18

15

15

0.62

43

41

37

18

14

14

0.65

42

41

36

16

14

13

0.68

42

36

36

15

13

12

0.70

41

34

32

14

13

12

0.73

41

30

31

14

13

12

0.76

38

26

29

14

12

12

0.78

36

26

26

14

11

11

0.81

31

26

25

12

11

10

0.84

30

24

25

11

11

10

0.87

26

23

22

11

10

10

0.89

23

22

20

10

10

8

0.92

22

21

19

8

9

7

0.95

21

21

19

8

8

7

0.97

8

10

8

5

5

4

1 in 10 Year













Probability 1

277

257

233

87

81

73

Median

49

45

40

20

19

17

Overall Mean







32

30

27

(36 yr)2







39

36

33

1	l-in-10 year probability values are used in screening assessments.

2	Mean of the annual values and 90% upper bound on the mean, respectively.

67


-------