IMM sms MARCH 2005
EPA/600/R-05/029
GIS-Based Risk Assessment of
Pesticide Drift Case Study:
Fresno County, California
By
E. Henry Lee
Connie A. Burdick
David M. Olszyk
Office of Research and Development
National Health and Environmental Effects Research Laboratory
Western Ecology Division
U.S. Environmental Protection Agency
Corvallis, OR 97333
Prepared for
Office of Pesticide Programs
Office of Prevention, Pesticides, and Toxic Substances
U.S. Environmental Protection Agency
Washington, DC 20460
-------
-------
Table of Contents
1. Executive Summary 1
2. Introduction 5
3. Overview of Ecological Risk Assessment 8
3.1. GIS-based Risk Assessment 10
4. High Potential Risk Herbicides Considered 11
4.1. Federal and State Restricted Pesticides 12
4.2. Drift Complaints 15
5. Study Area 16
6. Data 19
6.1 Pesticide Usage Reporting Data 20
6.2 Field Boundary Data 22
6.3 Wind Model Data 25
6.4 Meteorological Data 25
7. Assumptions and Calculations in Risk Assessment of Spray Drift on
Nontarget Crops 28
7.1. AgDRIFT Model Estimates of Spray Drift Deposition 29
7.2. Estimated Total Acreage at Risk from Ground and Aerial
Applications 30
7.3. Crop Management Dates and Time of Maximum Plant
Herbicide Sensitivity 32
8. Trends in Agricultural Use of High Potential Risk Herbicides in the
United States 33
8.1. Corn 34
8.2. Soybeans 36
8.3. Wheat 38
8.4. Upland Cotton 41
9. Results - Timing and Method of Pesticide Applications for High
Potential Risk Herbicides 43
-------
9.1. 2,4-D and 2,4-DB 43
9.2. Bromoxynil 44
9.3. Clopyralid 46
9.4. Dicamba 46
9.5. Glyphosate 47
9.6. Imazethapyr 50
9.7. MCPA 50
9.8. Paraquat 50
9.9. Sethoxydim 53
10. Results - GIS-based Risk Analysis of Selected High Potential Risk
Herbicides 53
10.1. Bromoxynil Risk 60
10.2. Dicamba Risk 60
10.3. Glyphosate Risk 61
10.4. Sethoxydim Risk 61
10.5. Summary of Risk Assessment for Fresno County 67
11. Conclusions 68
Acknowledgements 71
References 71
Appendix A: Abbreviations and Acronyms 87
Appendix B: Colloquial and Latin Names 90
Appendix C: List of the High Potential Risk Herbicides and Associated
Sensitive Crop Species Reported in the Published Literature ... 93
Appendix D: Background Information on the High Potential Risk
Herbicides 123
D.I. 2,4-D 123
D.2. 2,4-DB 124
D.3. Bromoxynil 125
D.4. Chlorsulfuron 127
ii
-------
D.5. Clomazone 128
D.6. Clopyralid methyl 130
D.7. Dicamba 131
D.8. Glufosinate 132
D.9. Glyphosate 133
D.10. Imazapyr 136
D.ll. Imazethapyr 136
D.12. MCPA 137
D.13. Metribuzin 138
D.14. Metsulfuron-methyl 139
D.15. MSMA 140
D.16. Nicosulfuron and primisulfuron-methyl 141
D.17. Paraquat dichloride 143
D.18. Picloram 145
D.19. Propanil 146
D.20. Prosulfuron 147
D.21. Quinclorac 148
D.22. Rimsulfuron 148
D.23. Sethoxydim 150
D.24. Sulfometuron-methyl 152
D.25. Thifensulfuron-methyl and Tribenuron-methyl 152
D.26. Triclopyr 154
Appendix E: Crop Planting and Harvest Dates in Fresno County, CA 156
Figures
Figure 1. Estimated spray deposition for aerial and ground applications
produced by the AgDRIFT model 10
Figure 2. Agricultural counties in the Central Valley of California. 17
in
-------
Figure 3. Top agricultural counties in California based on agricultural
production 17
Figure 4. Top agricultural crops by acreage in Fresno County in 2000 18
Figure 5. Townships 13S 13E and 16S 15E within the agricultural region in
Fresno County, CA 24
Figure 6. Crop field boundary for townships A) 13S 13E and B) 16S 15E
in Fresno county and crop grown in 2001 24
Figure 7. Wind trajectories after A) 1 hour and B) 2 hours based on NOAA's
HYSPLIT model for a central location in Fresno Co, CA between May
and October 2000 26
Figure 8. First order meteorological stations in California 27
Figure 9. Wind rose for Fresno County, California in April 27
Figure 10. AgDRIFT estimates of deposition rates for aerial application of
glyphosate using four spray droplet distributions and four boom
heights of A) 3.0m, B) 3.7 m, C) 4.3 m and D) 4.9 m 31
Figure 11. Drift zone for a square field of size 40 ha (100 acres) assuming a
drift range of 300 m for an aerial application 31
Figure 12. Crop planting dates for major field crops for California 32
Figure 13. Percent cotton planted in California for years 1996-2001 based on
weekly crop reports by the USDA NASS 33
Figure 14. Time of increased plant sensitivity for crops in Fresno County, CA ... 34
Figure 15. Trends in herbicide use on corn in the United States in the years
1992-2003 based on the annual USDA Agricultural Chemical Usage
reports 36
Figure 16. Trends in herbicide use on corn in California in the years
1992-2002 based on the CPUR database. 37
Figure 17. Trends in herbicide use on soybeans in the United States in the years
1992-2002 based on the annual USDA Agricultural Chemical Usage
reports 38
IV
-------
Figure 18. Trends in herbicide use on spring wheat in the United States in the
years 1992-2002 based on the annual USDA Agricultural Chemical
Usage reports 39
Figure 19. Trends in herbicide use on winter wheat in the United States in the
years 1992-2002 based on the annual USDA Agricultural Chemical
Usage reports 40
Figure 20. Trends in herbicide use on wheat in California in the years 1992-2002
based on the California Pesticide Use database 40
Figure 21. Trends in herbicide use on Upland cotton in the United States in the
years 1992-2003 based on the annual USDA Agricultural Chemical
Usage reports 42
Figure 22. Trends in herbicide use on cotton in California in the years 1992-2000
based on the California Pesticide Use database 42
Figure 23. 2,4-D usage in Fresno County, CA in 2000 44
Figure 24. Bromoxynil usage in Fresno County, CA in 2000 45
Figure 25. Bromoxynil usage on BXN® cotton in Fresno County, CA in 2000 45
Figure 26. Clopyralid usage in Fresno County, CA in 2000 46
Figure 27. Dicamba usage in Fresno County, CA in 2000 47
Figure 28. Total kg a.i. of glyphosate applied by crop in Fresno County in 2000. 48
Figure 29. Glyphosate usage in Fresno County, CA in 2000 48
Figure 30. Glyphosate usage on Roundup Ready™ cotton in Fresno County, CA
in 2000 49
Figure 31. Glyphosate usage on A) grapes, B) almonds, C) oranges and
D) processing tomatoes in Fresno County, CA in 2000 51
Figure 32. Imazethapyr usage in Fresno County, CA in 2000 52
Figure 33. MCPA usage in Fresno County, CA in 2000 52
Figure 34. Paraquat usage in Fresno County, CA in 2000 53
Figure 35. Paraquat usage on A) cotton, B) almonds and C) table, raisin and
wine grapes in Fresno County, CA in 2000 54
-------
Figure 36. Sethoxydim usage in Fresno County, CA in 2000 55
Figure 37. Sethoxydim usage on A) alfalfa and B) cotton in Fresno County, CA
in 2000 55
Figure 38. Sections in Fresno County where non-BXN® cotton was planted in
2000 and bromoxynil was or was not applied by air (A) and/or ground
(G) spray equipment between May 23 and July 29 63
Figure 39. Sections in Fresno County where A) grapes and B) wine grapes were
grown in 2000 and bromoxynil was or was not applied by air (A) and/or
ground (G) spray equipment between March 23 and June 6 63
Figure 40. Sections in Fresno County where A) cotton, B) alfalfa, C) table and
raisin grapes, D) wine grapes and E) tomato fields were grown in
2000 and dicamba was or was not applied by air (A) and/or ground (G)
spray equipment at time of increased plant sensitivity 64
Figure 41. Sections in Fresno County where non-RR cotton was planted in 2000
and glyphosate was or was not applied by air (A) and or ground (G)
spray equipment between May 23 and July 29 65
Figure 42. Sections in Fresno County where A) alfalfa and B) tomato were
planted in 2000 and glyphosate was or was not applied at time of
increased plant sensitivity. 65
Figure 43. Sections in Fresno County where A) corn, B) onion, C) table
and raisin grapes, D) wine grapes, E) rice and F) sorghum were
grown in 2000 and glyphosate was or was not applied at time of
increased plant sensitivity 66
Figure 44. Sections in Fresno County where A) corn and B) sorghum were
grown in 2000 and sethoxydim was or was not applied at time of
increased plant sensitivity. 67
Figure Dl. Agricultural use of 2,4-D in the United States in 1997 123
Figure D2. Agricultural use of 2,4-DB in the United States in 1997 125
Figure D3. Agricultural use of bromoxynil in the United States in 1997 126
vi
-------
Figure D4. Agricultural use of chlorsulfuron in the United States in 1997 128
Figure D5. Agricultural use of clomazone in the United States in 1997 129
Figure D6. Agricultural use of clopyralid in the United States in 1997 130
Figure D7. Agricultural use of dicamba in the United States in 1997 131
Figure D8. Glufosinate use in kg a.i. applied in California in 1997-2002 133
Figure D9. Agricultural use of glyphosate in the United States in 1997 134
Figure D10. Agricultural use of imazethapyr in the United States in 1997 137
Figure Dll. Agricultural use of MCPA in the United States in 1997 138
Figure D12. Agricultural use of metribuzin in the United States in 1997 139
Figure D13. Agricultural use of metsulfuron in the United States in 1997 140
Figure D14. Agricultural use of MSMA in the United States in 1997 141
Figure D15. MSMA usage in kg a.i. applied in California in 1991-2002 142
Figure D16. Agricultural use of nicosulfuron in the United States in 1997 143
Figure D17. Agricultural use of paraquat in the United States in 1997 144
Figure D18. Agricultural use of picloram in the United States in 1997 145
Figure D19. Agricultural use of propanil in the United States in 1997 146
Figure D20. Agricultural use of prosulfuron in the United States in 1997 148
Figure D21. Agricultural use of quinclorac in the United States in 1997 149
Figure D22. Agricultural use of rimsulfuron in the United States in 1997 150
Figure D23. Agricultural use of sethoxydim in the United States in 1997 151
Figure D24. Sethoxydim use in kg a.i. applied in California in 1991-2002 151
Figure D25. Agricultural use of thifensulfuron in the United States in 1997 153
Figure D26. Agricultural use of triclopyr in the United States in 1997 155
Tables
Table 1. Herbicide usage in California and Fresno County in 2000 and in the
United States in 1997 13
Table 2. Factors influencing near-field spray drift 19
Table 3. Basic Tier III model parameters used in AgDRIFT simulations 30
vii
-------
Table 4. Percent acreage of corn, Upland cotton, soybeans, rice, spring and
winter wheat treated with the 28 high potential risk herbicides in the
United States in 2000 35
Table 5. Total hectares planted of various crops subject to offsite spray drift
from ground and aerial applications within the same section for Fresno
County, CA in 2000 57
Table 6. Estimated total crop acreage in drift range from ground and aerial
applications within the same section for Fresno County, CA in 2000 62
Table Cl. High potential risk herbicides that have been shown to cause yield
reductions of 10% or greater in susceptible plants when exposed to
10% of typical application rate 93
Table C2. High potential risk herbicides and the associated susceptible crops
ranked by reported EC 10 values (given in parentheses in g/m2) in the
published literature 118
Table El. Relevant dates of importance for crops in Fresno County, CA 156
Vlll
-------
1. Executive Summary
A deterministic GIS-based risk assessment model was developed to evaluate the
ecological risk of pesticide spray drift on nontarget fields of sensitive crops based on pesticide
use reporting data at the section level. USEPA defines pesticide spray drift as the wind-driven
movement of pesticide particles in the air to unintended sites at the time of application. In order
for a pesticide application to cause damage to nontarget crops growing in fields adjacent to the
target field, the pesticide must be applied at the time of increased plant sensitivity.
Consequently, the study of pesticide spray drift effects on sensitive crops is data-intensive and
requires spatially- and temporally-explicit information on pesticide applications in agricultural
production. Very few states with the exception of California have full pesticide use reporting
databases at sufficient temporal and spatial resolution to infer the effects of pesticide spray drift
on nontarget crops with any degree of accuracy and precision. In this report, the California
Pesticide Use Reporting database for Fresno County in 2000 was analyzed using a deterministic
GIS-based risk assessment model to quantify the potential crop acreage at risk due to unintended
pesticide exposure from spray drift. The results for Fresno County and other counties in
California may be applicable to neighboring states which are likely to have similar weed
management practices and agricultural production systems. However, in other states where full
pesticide use data are not available, a probabilistic risk assessment model is required to simulate
agricultural applications of pesticides across time and space in addition to spray drift events and
ensuing impacts to unintended sensitive crops. The lack of resolution of pesticide use data is the
primary source of uncertainty and variability in a spatially-explicit risk assessment model for
pesticide drift.
We identified 28 federally-approved herbicides as having high risk potential to sensitive
crops based on scientific evidence of damage at sublethal levels in the peer-reviewed literature
as well as frequency of drift complaints reported to state and county agencies. A high potential
risk herbicide was reported to have caused a 10% or greater reduction in crop yield at 10% of the
application rate in at least one published study. Domestic corn producers used 15 of the 28 high
potential risk herbicides (2,4-D, bromoxynil, clopyralid, dicamba, glufosinate, glyphosate,
imazapyr, imazethapyr, metribuzin, nicosulfuron, paraquat, primisulfuron, prosulfuron,
1
-------
rimsulfuron and thifensulfuron) to control weeds based on the USDA's Agricultural Chemical
Usage reports (USDA, 1993-2004). Ten of the 28 high potential risk herbicides (2,4-D, 2,4-DB,
clomazone, glyphosate, imazethapyr, metribuzin, paraquat, sethoxydim, thifensulfuron and
tribenuron) were used on soybeans. Wheat producers used 13 of these herbicides. Several of
these herbicides (propanil, quinclorac, triclopyr) are used almost exclusively on rice.
The crop acreage at potential risk due to unintended exposure to these 28 high potential
risk herbicides varies from state to state because pesticide use depends upon the weed problem
and types of agronomic cropping systems which vary from state to state. Paraquat, triclopyr and
picloram are federally restricted pesticides (denoted Restricted Use Pesticides) that require the
purchaser to be a certified private applicator or a pesticide licensee. A state pesticide permit is
not required to purchase or use a Restricted Use Pesticide. Each state can impose stricter
standards and prohibit or severely restrict the use of federally approved pesticides but few states
do. California, Texas, Arkansas and Mississippi are among the few states that regulate the
purchase and use of federally approved pesticides because of their potential to cause adverse
effects to unintended, commercially-important, susceptible crops. For example, the herbicides,
2,4,5-T, CGA-152005 (or prosulfuron), metsulfuron-methyl, primisulfuron, thifensulfuron and
tribenuron, are not registered for use in California. The California Code of Regulations Title 3
Section 6400 lists pesticide products containing the active ingredients, 2,4-D, 2,4-DB, dicamba,
MCPA, paraquat and propanil, as Restricted Materials which allows county agricultural
commissioners to restrict the sale and use of these products. Dicamba, propanil and several
phenoxy herbicides such as 2,4-D, 2,4-DB and MCPA are prohibited from use below 305 m
elevation in the Central Valley (which includes the agricultural section of Fresno County) from
March 16 to October 15 due to the high sensitivity of major cash crops to these herbicides.
Consequently, the crop acreage at potential risk from these state-restricted-use herbicides is
negligible.
The introduction of herbicide-resistant crops has had a dramatic effect on weed
management systems for cotton and soybeans and, to a lesser extent, corn in the United States.
This is evidenced by the increasing trends in glyphosate use on cotton and soybeans and the
decreasing trends in the use of other higher potential risk herbicides in California and the United
-------
States. Glyphosate is the most commonly used herbicide on corn and soybeans which account
for about 80% of all herbicide use in domestic agriculture. Glyphosate use on conventional
crops such as wheat has also increased while use of the more potentially damaging herbicides
such as 2,4-D have remained steady or decreased. In California, glyphosate use of 2,202,358 kg
active ingredient applied in 2000 was greater than the total use of the other 27 high potential risk
herbicides combined. Similarly, glyphosate is the top-ranked herbicide in Fresno County,
California based on total amount of pesticide applied or hectares treated. Glyphosate is not as
damaging to sensitive crops as 2,4-D and dicamba and other high potential risk herbicides but
has the greater potential to damage sensitive crops due to spray drift because it is applied
throughout the year in large quantities.
In this report, a worst-case scenario was used to estimate the total acreage at risk due to
pesticide spray drift in Fresno County in 2000 based on the California Pesticide Use Reporting
database. Crop planting and harvest dates reported by the USDA annual crop reports and Fresno
County weekly crop reports in conjunction with reported times of increased plant sensitivities in
the published literature were used to determine the time periods when the sensitive crops were
susceptible to pesticide spray drift. We assumed that all pesticide applications within these
times of increased plant sensitivity resulted in spray drift damage to all nontarget fields within
the same section as the target field. For each of the 28 high potential risk herbicides and the
corresponding crops susceptible to spray drift damage, spatially-explicit maps were generated
using the GIS-based risk assessment model to indicate the sections where the pesticide was
applied at the time of increased plant sensitivity in the vicinity of a susceptible crop field. Based
on the AgDRIFT model (Teske et al, 1997, 2002; Bird et al, 2002) given a4.5 m/s wind, the
estimated size of the drift zone in a nontarget field of size 100 acres (0.40 km2) is 47% of the
total acreage for an aerial application and 4.7% for a ground application. The total acreage at
risk to spray drift exposure was calculated as the total hectares planted of the sensitive crop in
the sections where the pesticide was applied multiplied by either 47% for an aerial application or
4.7% for a ground application.
In Fresno County in 2000, the estimated total acreage at risk was highest for nontarget
fields of alfalfa due to drift from aerial applications of glyphosate, followed by non-Roundup-
-------
Ready™ cotton, tomatoes and grape acreage potentially damaged due to drift from aerial and
ground applications of glyphosate. About 3,900 ha of alfalfa or 11% of alfalfa acreage in Fresno
County was estimated to be at risk from offsite pesticide spray drift of glyphosate. This, in part,
was due to the fact that alfalfa had the widest period of increased plant sensitivity from March 4
to July 29. In comparison, about 2,200 ha of non-Roundup-Ready™ cotton or 2.3% of
bromoxynil-tolerant (BXN®) and conventional cotton acreage were estimated to be potentially at
risk due to wind-blown pesticide spray drift from glyphosate ground and aerial applications to
adjacent fields. Herbicides other than glyphosate posed little or no risk to susceptible crops in
Fresno County with the possible exception of sethoxydim drift on corn and bromoxynil drift on
non-BXN® cotton. About 2.6% of corn acreage and 0.2% of cotton acreage were estimated at
risk due to pesticide spray drift from sethoxydim and bromoxynil ground and aerial applications,
respectively. In general, more crop acreage was at risk due to exposure from aerial applications
than ground applications because spray from an aerial application was assumed to drift 10 times
farther than that from a ground application.
These findings illustrate the risk assessment approach based primarily on pesticide use
data at the section level. A more refined GIS-based approach that incorporates other factors
such as wind speed and direction is possible when pesticide use is reported at the field level.
Spatially-explicit maps at the field level provide useful information on the relative locations and
shapes of the target and nontarget fields that would result in greater accuracy and precision in
estimating total acreage at risk due to pesticide spray drift Conversely, pesticide use data
reported at the county level would significantly degrade the ability to estimate the total acreage
at risk due to drift, as would the loss of temporal resolution. Unfortunately, the only available
pesticide use data for states other than California lack both temporal and spatial resolution.
Other data layers such as cropland use from GAP, the United States Department of Agriculture
or other sources could be used to improve the spatial resolution but more research is needed to
quantify the uncertainty and variability in the risk assessment calculations when pesticide use
data are lacking.
-------
2. Introduction
In support of the EPA's Office of Pesticide Programs' regulation of the use of pesticides,
this report addresses the need to estimate the amounts of off-target spray drift and the associated
potential risks to the environment. It focuses on pesticide use by the agriculture sector which
accounts for about 70% of the total use in the United States (Donaldson et al, 2002; USEPA,
2004). EPA defines pesticide spray drift as the wind-driven movement of pesticide particles in
the air to unintended sites at the time of application. When pesticide solutions are sprayed by
ground spray equipment or aircraft, droplets formed by the nozzles of the equipment may move
laterally beyond the boundary of the intended field and pose potential ecological risks. Other
routes of pesticide exposure including movement of volatilized droplets, soil particles or
groundwater which may contain pesticide residues are not considered in this report.
Spray drift is the primary airborne transport pathway by which pesticides can move
offsite and potentially damage nontarget crops and native vegetation. This exposure is the result
of near-field drift (< 1 km) of relatively large spray droplets and far-field drift (-1-10 km) of
very small droplets (< 100 m) at the time of application. Particle movement is largely
determined by wind speed and direction, atmospheric conditions, spray equipment, droplet size,
and method of application. In order to minimize spray drift, many pesticides have label
recommendations for specific droplet sizes such as "apply as a MEDIUM spray" as defined by
the ASAE Standard S572 (ASAE, 2000). Aerial applicators are utilizing computer models
(Kirk, 2001, 2002) to select a nozzle and its orientation, spray pressure, airspeed and other
application parameters that correspond to label recommendations for a desired spray droplet
spectra under specified conditions. As droplet size increases, the downwind drift is reduced
(Teske and Thistle, 1999). Applicators generally consider the volume median diameter (Dv05)
and spray volume in droplets smaller than 100 m (V<100 m) and 200 m (V<200 m) when
selecting their application equipment. The V<200 m represents the portion of spray that is most
likely to drift and V<100 m represents the portion of spray that is most likely to drift beyond 1 km.
The droplet spectra for a reference nozzle classified as medium based on the ASAE Standard
S572 has Dv0 5 = 294 m, V<100 m 5%, and V<200 m 25%. In comparison, a fine droplet
spectra has Dv0 5 = 180 m, V<100 m 20%, andV<200 m 60%. Fine droplet spectra are
-------
generally not used for pesticide applications in California due to drift risk (pers. comm. Barry,
2005). Consequently, this report assesses the potential risk of near-field spray drift for medium
and coarse sprays to unintended crops within 1 km of the site of application.
Spray transport models such as AgDRIFT indicate that near-field pesticide drift from
aerial application is most likely to occur within 300 m of the field boundary and considerably
less from ground application (Teske et al., 2002; Bird et al., 2002). Other estimates of pesticide
drift from aerial application range from 100 m to 600 m before deposition rates of 1% or less
were reported, depending upon wind speed and direction, relative humidity, atmospheric
stability, spray equipment, type of aircraft, airplane speed, droplet size, height of application and
number of swaths (Yates et al., 1978; Chester and Ward, 1984; Wilson et al., 1986; Barnes et
al., 1987; Riley and Wiesner, 1989; Huitnik et al., 1990; Payne et al., 1990; Ernst et al., 1991;
Marrs et al., 1992; Payne, 1993; Bird et al., 1995, 1996; Woods et al., 2001). Consequently, the
potential effects of near-field pesticide drift are limited to adjacent fields from where the
pesticide was applied.
Environmental risk assessments often report risks as single values or a table of values
and do not include specific spatial information. Because pesticide spray drift is limited to within
about 300 m of the boundary of the target field, spatially-explicit information on pesticide
applications, cropland use, weather and other influencing factors are needed to assess the
potential risk of spray drift on nontarget crops. When pesticide use reporting (PUR) data are
available at high temporal and high spatial resolutions, a deterministic GIS-based approach can
be used to estimate the crop acreage at risk based on the time, location and method of the
reported pesticide applications. When high temporal and spatial resolution PUR data are not
available, a probabilistic approach is required to simulate the pesticide applications and drift
events over time and space in order to estimate the potential effects of pesticide spray drift on
nontarget crops.
Geographic information system (GIS) mapping is an important component in risk
assessment of pesticide drift on the environment because of the spatially explicit nature of the
problem. This is best performed by storing the spatial information in a GIS database as inputs
for the risk assessment model. These GIS-based risk assessment models can run both
-------
deterministic as well as probabilistic risk assessments. Using statistical functions available in
GIS, the total acreage of various crop and non-crop land in the drift zone can be calculated and
stored with their corresponding geographic locations in the GIS database. More importantly, the
GIS approach provides spatially-explicit maps of the locations of the pesticide applications in
relation to the nontarget fields at the resolution of the PUR data. These maps are the primary
GIS outputs for assessing the potential effects of pesticide spray drift on nontarget crops.
The long-term goal of this research is to develop a conceptual model that describes the
qualitative relationships between pesticide application, patterns of exposure, and ecological
effects. The model identifies factors that determine exposure and responses of terrestrial
ecosystems adjacent to treated fields, including physical and chemical properties of the
pesticide, application and agronomic practices, soil characteristics, weather, characteristics of
biological receptors, and ecological processes. Climate, physical properties of the pesticide and
application practices influence exposures by controlling the movement of pesticides in the air.
Each of these groups of factors includes several specific components. During the problem
formulation stage of the risk assessment, sources of uncertainty and variability associated with
each component are identified, and decisions are made as to how these uncertainties and
variabilities will be addressed.
This report illustrates the use of a deterministic GIS-based risk assessment model for
spray drift on agronomic fields based on the best available data for pesticide applications at high
temporal and high spatial resolutions. Because drift occurs at small spatial scales, anational risk
assessment of pesticide drift was not possible due to inadequate data to develop GIS database
coverages for most states. Initially, a detailed literature review on exposure and effects of
herbicides to agricultural crops and other plant species was conducted to identify the herbicides
of national interest and the associated crops that are susceptible to drift damage from agricultural
application of these herbicides. We examine trends in pesticide use on the major cash crops in
the United States and California to determine whether pesticide use patterns in California mirror
those for other states. Based on the California Pesticide Use Reporting (CPUR) database at the
section level, a worst-case scenario was used to examine the crop acreage at potential risk due to
pesticide spray drift from agricultural application of a given herbicide. In this report, the GIS-
-------
based approach was applied to a county in California to generate a series of spatially-explicit
maps of pesticide applications in the vicinity of susceptible crops at the time of increased plant
sensitivity. And finally, we discuss future research to develop an alternative GIS-basedrisk
assessment approach to assess the risk of pesticide drift when PUR data are lacking in spatial
resolution.
3. Overview of Ecological Risk Assessment
Ecological risk assessment is defined by the Society of Environmental Toxicology and
Chemistry (SETAC) as "A formal set of scientific methods for estimating the probabilities and
magnitudes of undesired effects on plants, animals, and ecosystems resulting from events in the
environment including the release of pollutants, physical modification of the environment, and
national disasters." This report follows the general framework for ecological risk assessment as
practiced by the U.S. EPA Office of Pesticides and Toxic Substances (OPTS) for regulatory
decisions with respect to pesticides and other toxic chemicals (USEPA, 1992, 1996). Initially,
an ecological risk characterization phase involving ecotoxicological hazard and environmental
exposure is undertaken to determine whether the herbicide level that clearly produces effects in
controlled field studies are equal to or less than the estimated amount of off-site spray drift.
This first phase uses the familiar "Quotient Method" to identify the high potential risk
herbicides and sensitive crop species of interest for further analysis in a GIS-based risk
assessment model. The risk characterization phase focuses on general risk levels but does not
account for the temporal and spatial aspects of exposure. The second phase of risk assessment
focuses on potential effects to the environment associated with wind-driven pesticide spray drift
at the section level for a given time and method of application. Spatially-explicit pesticide use
data are incorporated in a GIS mapping system to identify the crop species that are most likely
affected by the application of the herbicides at the time of maximum plant sensitivity.
In a traditional ecological risk assessment, an ecotoxicological hazard assessment and an
environmental exposure assessment are first carried out to determine qualitatively the likelihood
of the risk. Ecotoxicological hazard is the intrinsic quality of a chemical to cause adverse
effects under a specified set of conditions. Much research effort has been devoted to
measurements and models of plant response when exposed to sublethal levels of herbicides, in
8
-------
particular for the economically important crops and the most popular herbicides including 2,4-D,
glyphosate, dicamba and the sulfonylureas. Initially, a comprehensive literature review was
undertaken to determine the herbicide levels required to cause an adverse effect on crop yield.
Based on readily available information in the published literature, the ecotoxicological hazard of
a herbicide was defined as the effective concentration associated with 10% yield loss (i.e.,
EC 10). Depending upon the experimental design, number of replicates and the experimental
variation, a 10% yield loss may or may not have been statistically significant at the 0.05 level
but reflects the amount of damage to susceptible crops when exposed to sublethal levels of a
herbicide resulting from offsite pesticide spray drift.
In regards to pesticide spray drift to unintended fields, environmental exposure is the
amount or concentration of spray deposition to which nontarget crops may be exposed. Because
off-site drift is more or less a physical process that depends on droplet size and meteorological
conditions rather than the specific properties of the herbicide, environmental exposure was
based on the AgDRIFT model (Teske et al., 1997,2002; Bird et al, 2002). AgDRIFT is a
model developed as a joint effort by the EPA Office of Research and Development and the
Spray Drift Task Force (SDTF), a coalition of pesticide registrants. AgDRIFT is based on the
algorithms in FSCBG (Teske and Curbishley. 1990; Teske et al., 1993), a drift model previously
used by USD A Further details of AgDRIFT are available at www.AgDRIFT.com. For aerial
applications, AgDRIFT allows very detailed modeling of drift based on the physical properties
of the applied product, the configuration of the aircraft, as well as wind speed and temperature.
For ground applications, AgDRIFT provides estimates of drift based solely on distance
downwind as well as the configuration of the spray equipment. AgDRIFT estimates the fraction
of a.i. deposited as a function of distance downwind (Figure 1). For aerial application with a
4.5 m/s wind and a boom height of 4.9 m, the estimated spray deposition range was 300 m
before the fraction of a.i. dropped below 1%. For ground application with a 4.5 m/s wind, the
spray drifted out to about 30 m from the edge of the field before the fraction of a.i. dropped
below 1%.
-------
Deposition Estimates For Aerial and Ground Applications
Produced By The AgpROT Spray Drift Model
a H
D.1
U.W
O.W1
Figure 1. Estimated spray deposition for aerial and ground applications
produced by the AgDRIFT model.
High potential risk pesticides have characteristics that increase the likelihood of adverse
impacts when they are released to the environment or nontarget organisms are exposed to them.
High potential risk pesticides ideally would be identified using an index that considers a wide
variety of risk attributes. Such a framework allows various strategies to be evaluated on the basis
of something that is readily available in the published literature. Research has shown that off-
site pesticide spray drift ranges between 1% and 10% of the applied rate (Al-Khatib and
Peterson, 1999; Bailey and Kapusta, 1993; Snipes et al, 1991, 1992). In this report, we used
10% of the label rate for a postemergent, postplant application of the herbicide to characterize
the expected spray drift concentration (ESDC) in comparison to the EC 10 values reported in the
literature. A high potential risk herbicide was defined as having an EC10 ESDC, i.e., the
likelihood of risk is indicated by the quotient, EC 10 / ESDC > 1. The problem formulation
phase identifies the herbicides of interest and the sensitive species that will be most affected by
spray drift but does not provide any information on the frequency of applications of these
herbicides in the proximity of the myriad of crop species at the time of increased sensitivity.
3.1. GIS-based Risk Assessment
10
-------
Ecological risk assessment to determine if off-site drift of pesticides applied to
agronomic fields poses an unreasonable risk to the environment can best be addressed in a GIS
framework because off-site spray drift is a spatially-explicit problem determined by the
boundary of the target field and the crop species in the adjacent fields. In a GIS, field boundary
coverages can be used to identify the agronomic fields where pesticides were applied and the
adjacent nontarget fields in the direction of the prevailing winds. Of special concern are the
agronomic fields where the high potential risk herbicides were applied aerially or by ground in
relation to the agronomic fields of the most sensitive crop species. The data requirements to
evaluate the risk of pesticide spray drift are substantial. In order to quantify the likelihood of
risk of off-site pesticide spray drift in a GIS framework, pesticide use data at high spatial and
temporal resolution and field boundary coverages are needed. These data requirements are not
met for most states except California and possibly Arizona where full or partial commercial
PUR is required by law. The current assessment reflects the most up-to-date field boundary
coverages and the most complete PUR data available.
4. High Potential Risk Herbicides Considered
High potential risk herbicides and corresponding susceptible crops of interest were
selected based on current knowledge of herbicide effects on crop yield. A literature search was
conducted to obtain all relevant information on crop yield response at maturity subsequent to
sublethal doses of herbicide applied directly to the foliage under field conditions. The
screening process excluded studies that showed no crop damage at sublethal levels of the
herbicide and focused on studies that reported a yield reduction of 10% or greater at less than
label rate. For studies replicated over time and space and/or having multiple cultivars, we chose
to report the EC 10 for the most sensitive cultivar and the site/year with the greatest yield
reductions. Studies in the published literature identified 26 high potential risk herbicides
including 2,4-D, dicamba and glyphosate based on the quotient, EC10/ESDC, exceeding one in
at least one published study (Table Cl in Appendix C). Some of these pesticides are used on
only a few crops (e.g., primisulfuron, propanil, triclopyr), while others are used on a great
variety of crops (e.g., glyphosate, 2,4-D). For example, propanil, quinclorac and triclopyr are
used almost exclusively on rice. See Appendix D for a description of the properties, use and
11
-------
risks of these high potential risk herbicides. Some of these herbicides are heavily used on
California agriculture (glyphosate, propanil use exceeded 0.5 million kg a.i.), while use of others
is relatively rare (Table 1).
4.1. Federal and State Restricted Pesticides
Paraquat, triclopyr and picloram are federally restricted pesticides (denoted Restricted
Use Pesticides or RUPs) that require the purchaser to be a certified private applicator or a
pesticide licensee. A state pesticide permit is not required to purchase or use a RUP. Each state
can impose stricter standards and prohibit or severely restrict the use of federally approved
pesticides but few states do. California, Texas, Arkansas and Mississippi are among the few
states that regulate the purchase and use of federally approved pesticides because of their
potential to cause adverse effects to unintended, commercially-important, susceptible crops. For
example, the federally approved herbicides, 2,4,5-T, CGA-152005 (or prosulfuron),
metsulfuron-methyl, primisulfuron, thifensulfuron and tribenuron, are not registered for use in
California. The California Code of Regulations Title 3 Section 6400 lists pesticide products
containing the active ingredients, 2,4-D, 2,4-DB, dicamba, MCPA, paraquat and propanil, as
Restricted Materials which allow county agricultural commissioners to restrict the sale and use
of these products. Propanil and several phenoxy herbicides such as 2,4-D, 2,4-DB and MCPA
are prohibited from use below 305 m elevation in parts of the Central Valley (which includes the
agricultural section of Fresno County) from March 16 to October 15. Aerial applications of
paraquat for preplant or preemergence weed control are restricted to a maximum boom height of
3 m in winds of less than 4.5 m/s. The Texas Department of Agriculture severely restricts the
use of 2,4-D, dicamba, MCPA, quinclorac and propanil by requiring a state permit for
application and/or prohibiting aerial applications of these chemicals at varying times of the year
depending upon the county. The State of Arkansas designated all products containing 2,4-D,
MCPA or quinclorac as Class F which severely restricts the purchase and usage of these
compounds; a key restriction is specification of a 6.4 km buffer zone for aerial applications (1.6
km for ground applications) between the target field and susceptible crops from April 15 to
September 15. In addition, Class F products cannot be applied when wind speed exceeds 3.6
m/s and temperature exceeds 32.2°C in Arkansas. In 19 counties in Mississippi, a special permit
12
-------
Table 1. Herbicide usage in California and Fresno County in 2000 (CPUR) and in the United
States in 1997 (Gianessi and Marcelli, 2000). The high potential risk herbicides, CGA-152005
(or prosulfuron), metsulfuron-methyl, primisulfuron, thifensulfuron and tribenuron, are not
registered for use in California. Clomazone and quinclorac were registered for use in CA in
2002.
Chemical (CA Chem code)
2,4-DB (mainly 838)
2,4-D (mainly 806)
Bromoxynil (mainly 834)
Chlorsulfuron(2143)
Clomazone (3537)
Clopyralid (mainly 5050)
Dicamba (mainly 849)
Glufosinate (3946)
Glyphosate (mainly 1855)
Imazapyr (2257)
Imazethapyr (mainly 2340)
MCPA (mainly 786)
Metribuzin(1692)
MSMA (34)
Nicosulfuron (3829)
Paraquat dichloride (1601)
Picloram (mainly 593)
Primisulfuron-methyl (5103)
Propanil (503)
Metsulfuron-methyl (2222)
Kg a.i. used in
CA in 2000
(Fresno County)
17,097 (228)
201,778(7,002)
68,521 (10,666)
1,179(14)
0(0)
5,978 (179)
18,031 (939)
0(0)
2,202,358
(284,585)
6,086 (0)
3,375 (358)
93,397 (5,236)
12,801 (102)
45,407 (5,847)
719 (5)
444,268 (102,390)
2(0)
0(0)
615,970 (239)
0(0)
Total ha treated
CA in 2000 (Fresno
County)
19,031 (208)
280,238 (15,140)
180,571 (30,829)
12,879 (0)
0(0)
14,445 (979)
69,630 (5,469)
0(0)
1,578,545 (269,506)
11,393(0)
41,292(4,057)
118,883(6,215)
31,743(138)
21,990(4,798)
18,577(174)
699,538 (158,786)
16(0)
0(0)
135,187(59)
0(0)
Kg a.i. used
in U.S. in
1997
273,958
18,411,295
1,324,591
27,100
1,148,115
404,451
4,738,880
NA
273,958
18,411,295
1,324,591
404,451
27,100
4,738,880
NA
15,793,016
NA
568,372
1,506,032
1,148,115
13
-------
Table 1 (cont).
Chemical (CA Chem code)
Prosulfuron (5115)
Quinclorac (5104)
Rimsulfuron (3835)
Sethoxydim(2177)
Sulfometuron-methyl (2149)
Thifensulfuron-methyl (2237)
Tribenuron-methyl (2338)
Triclopyr (mainly 2131)
Kg a.i. used in
CA in 2000
(Fresno Co)
0(0)
0(0)
529 (230)
14,984(1,305)
3,756 (79)
0(0)
0(0)
84,063(315)
Total ha treated
CA in 2000 (Fresno
Co)
0(0)
0(0)
49,327 (22,926)
51,823 (4,849)
2,399 (0)
0(0)
0(0)
116,649(51)
Kg a.i. used
in U.S. in
1997
33,147
130,500
8,969
778,941
NA
47,693
28,628
267,786
is required for aerial application of 71 commonly used preplant bunidown herbicides mostly
including the a.i. glyphosate, sulfosate and/or paraquat from the middle or end of March through
the end of April, depending on which county you farm in; the new label reduced drift complaints
to about a dozen annually except for 2003 when spring applications of pesticides were delayed
by wet weather. In 2003, the Mississippi Bureau of Plant Industry received 46 off-site spray
drift complaints primarily related to glyphosate and aerial applications (Delta Farm Press April
9, 2004).
Chlorsulfuron, metsulfuron-methyl, nicosulfuron, primisulfuron, prosulfuron,
rimsulfuron, thifensulfuron-methyl, tribenuron-methyl are sulfonylurea (SU) herbicides used at
very low application rates for non-selective control of annual and perennial grasses and
broadleaf weeds in non-crop and industrial areas. Application rates of SU herbicides are
generally over 100 times lower than those for older conventional herbicides, but SU herbicides
are remarkable for their high potency, and therefore low application rates. SUs have very low
acute and chronic toxicities to mammalian and other animal species. Because of their low
application rates in the field, the total kg of a.i. applied is typically several orders of magnitude
smaller for the SU herbicides than for conventional herbicides. Consequently, gross kg a.i. does
not reflect the rapid rise in the use of this class of herbicides in the U.S. and worldwide.
14
-------
4.2. Drift Complaints
Increased use of herbicide-resistant technology by producers creates the possibility of
off-site movement onto adjacent conventional crops. The role of total postemergence programs
to control grass and broadleaf weeds has expanded with the development of herbicide-resistant
crops. Because of the diversity of cropping systems in the United States, it is not uncommon for
herbicide-resistant crops to be planted near susceptible conventional crops. Postemergence
application of a herbicide to a genetically-modified (GM) crop often occurs when non-GM
plants are in the early reproductive growth stage and are most susceptible to damage from
herbicide drift (Ghosheh et al, 1994; Hurst, 1982; Snipes et al., 1991, 1992). Consequently,
most drift complaints occur in spring and summer as the use of postemergence herbicide
applications increase. For example, the wet spring in 2003 in Mississippi delayed the second
application of glyphosate on four-leaf Roundup Ready™ (RR) cotton and the first glyphosate
application made on RR soybeans. Consequently, more drift complaints were filed in 2003 in
Mississippi than in 2001-2002 combined due to in-season glyphosate treatments on RR crops
drifting onto susceptible crops, including conventional cotton and rice varieties (Delta Farm
Press, June 20, 2003). Similarly, several weeks of wet weather followed by several weeks of
high winds in Arkansas in 2004 when crops needed spraying resulted in increased drift
complaints. The most common postemergence herbicides associated with drift are 2,4-D,
glyphosate, halosulfuron, oxyfiuorfen, paraquat and sulfometuron (Lanini, 2003). A recent
national pesticide drift survey conducted by the American Association of Pesticide Control
Officials (AAPCO) in 1999 found the most commonly reported herbicides were 2,4-D, atrazine,
clomazone, dicamba, glyphosate, paraquat, picloram and propanil. Consequently, we included
clomazone and paraquat in our list of high potential risk herbicides. Further, most drift
complaints involved commercial applications of agricultural pesticides in rural areas and ground
applications accounted for 2/3 of the complaints. The top herbicides in agricultural drift
complaints vary from state to state because pesticides are used quite differently on different
crops and in different states.
We did not include atrazine in our list of high potential risk herbicides because potential
risk of atrazine to agricultural plants due to spray drift is low. Cucumbers, soybeans and
15
-------
cabbage are among the more sensitive crops to atrazine (USEPA, 2002) but there were no
studies that reported yield losses of 10% or greater at sublethal levels of atrazine. The main
concern is the human carcinogenic potential of atrazine through dietary or occupation exposure
to the pesticide. Based on the available data on human health effects, the USEPA changed the
classification of atrazine from category C to "likely human carcinogen" in the 1990s. Atrazine
poses a drinking water dietary risk because it is the most commonly detected pesticide in ground
and surface water (USEPA, 2002).
5. Study Area
In this report, we focused on the state of California because: (1) California has full
pesticide use reporting data since 1990 at the section level; (2) California ranked number one
among the 50 states in terms of pesticide use and accounted for 22% of all agricultural pesticide
use in the United States in 1994; (3) California is the leading agricultural state in the country and
has been for over 50 years, e.g., the state's farmers produced more than $26 billion in farm value
in 2002 (California Agricultural Statistics Service; CASS, 2002); (4) California has a wide
diversity of agriculture with more than 250 commodities in production and leads the nation in
production in 75 commodities (California Farm Bureau Federation Facts and Stats about
California Agriculture); and (5) cropland and ranches account for about one-third of the state's
total area. Agriculture accounts for about 30% of the total economy in the Central Valley of
California which extends north to south nearly 724 km from the Klamath/Cascades to the
Tehachapis Mts. between the Coast Range and the Sierra Nevada (Figure 2). This alluvial plain
contains the largest irrigated agricultural area west of the Rocky Mountains. The majority of
California's top producing agricultural counties are in the Central Valley, and of the top six, only
one (Monterey) is not in the Central Valley (Figure 3). This area is not only the most productive
in California, it is widely considered the most productive in the world. Agricultural production
in the Central Valley was estimated to be $14 billion in 1995 (California Dept. of Food and
Agriculture, 1996).
Within California, we focused on Fresno County because: (1) Fresno County is the most
agriculturally productive county in the state (Figure 3) and in the nation (CASS, 2002); (2)
Fresno County ranked number one among all 58 counties in California in terms of pesticide
16
-------
Figure 2. Agricultural counties in the Central Valley
of California.
Eleven of the top 20 agricultural counties in California
are in the Central Valley
(all 8 San Joaquin Valley Counties are among them)
10
2 3,500,000 j
» tr 3,000,000-
5 | 2,500,000-
| ~ 2,000,000 -
o | 1,500,000-
o | 1,000,000-
ol, 500,000-
Centra/ Valley counties shown by
pattern-filled bars, others are outlined only.
,RRRRR
Oil >^C C"O(l)tftO rOnjWOn! C0n'1~i(')^f^
S ^ ? I i S1 1 S .« « = =2 lBEl= » * » »
| .E m
= P I ^ 11 | i § I | • * | 1 i 11 * | g s '
Source: 1996 Calif. Agricultrual Rresources Directory (CA Dept. of Food and Ag ) c:\clata\centval\newer\ag.xlc
Figure 3. Top agricultural counties in California based on agricultural
production (California Dept of Food and Agriculture).
17
-------
sage; and (3) Fresno County is one of a few counties in which a complete field boundary
coverage had been developed that could be merged with the California Pesticide Use Reporting
(CPUR) database. Thus, Fresno County provides a unique situation where the amount of
potential pesticide drift can be quantified at the section level.
Primary agricultural products in Fresno County include cotton, grapes, poultry, tomatoes,
milk, head lettuce, almonds, cattle and calves, nectarines, and oranges (Figure 4). Grapes and
cotton are the leading agricultural crops in Fresno County and represent about 50% of the
agricultural acres and 29% of the agricultural value.
Seven counties including Fresno in the San Joaquin Valley (SJV) account for 56% of the
state's total reported pesticide use. In 2000, there were 16 million kg of active ingredients (a.i.)
applied in Fresno County versus 85 million kg a.i. for the state (CPUR). Glyphosate is among
the top herbicides used in Fresno County and is heavily used on cotton, grape and almond to
control weeds (Table 1). Glyphosate use of 39 to 41 million kg a.i. ranked first among
pesticides in the U.S. agricultural market sector in 2001 (USEPA, 2004).
n Cotton
• Gropes
• Tomatoes
DA If a If a
D Almonds
• Wine Grape
• Other
Figure 4. Top agricultural crops by acreage in Fresno
County in 2000(California Department of Pesticide
Regulation).
18
-------
6. Data
In this report, data for the most important factors affecting herbicide drift including
method, amount, time and location of pesticide application, wind speed and direction are used to
assess the agricultural risk of off-site pesticide drift in Fresno County, CA. Other factors
influencing near-field spray drift are set to values corresponding to a worst-case scenario for
lack of data (Table 2). Several factors including spray droplet size, spray pressure, nozzle type
and wind speed (4.5 m/s) were used to estimate off-site spray deposition using the AgDRIFT
model but were not included as data layers in the GIS. The pesticide use data at high temporal
and spatial resolution were the primary data used in a geospatial risk assessment framework to
quantify the likelihood of crop damage in terms of total hectares affected when exposed to spray
drift.
Table 2. Factors influencing near-field spray drift (Dexter, 1993).
Factor
Spray droplet size
Release (or boom) height
Wind speed
Spray pressure
Nozzle size
Nozzle orientation (aircraft)
Nozzle location (aircraft)
Nozzle type
Air temperature
Air stability
Herbicide volatility
Relative humidity
More drift
smaller
higher
higher
higher
smaller
forward
beyond 3/4 wing span
smaller droplets
higher
inversion
volatile
lower
Less drift
larger
lower
lower
lower
larger
backward
3/4 or less wing span
larger droplets
lower
lapse
nonvolatile
higher
19
-------
6.1. Pesticide Use Reporting Data
Both USEPA (Aspelin, 1997; Aspelin and Grube, 1999; Donaldson et al, 2002) and the
USDA National Agricultural Statistics Service (NASS) collect annual data on the sale and use
of pesticides at the state level. USDA pesticide use surveys conducted by NASS since 1993
provide on-farm chemical use data for eleven crops: corn, cotton, soybeans, wheat, rice, grain
sorghum, peanuts, fall potatoes, other vegetables, citrus, and apples (USDA, 1994-2004). Under
the sponsorship of USEPA, USDA, and the Water Resources Division of USGS, the National
Center for Food and Agricultural Policy (NCFAP) assembled a comprehensive database of
pesticide use in American agriculture (Gianessi and Silvers, 2000; Gianessi and Marcelli,
2000). The NCFAP database is a summary compilation of studies conducted by public agencies
for the years 1992 and 1997, including:
1. NASS surveys of pesticide use in field crops, vegetable crops, and fruit and nut crops,
Reports funded by the USDA Cooperative Extension Service (CES),
Pesticide benefit assessments from the USDA National Agricultural Pesticide Impact
Assessment Program (NAPIAP), and State of California compilations of farmers
pesticide use records, supplemented by NCFAP surveys of Extension Service specialists,
and, where necessary, imputations developed from the assumption that neighboring
States pesticide use profiles are similar.
The NCFAP database covering 220 a.i. and 87 crops are state-level point estimates
focused on two use coefficients: (1) the percent of a crop's acreage in a state treated with an
individual a.i., and (2) the average annual application rate of the active ingredient per treated
acre. Because these data represent the average application and treatment rates by state, they do
not yield precise estimates of use at the sub-state level. The state use coefficients represent an
average for the entire state and consequently do not reflect the local variability of cropping and
management practices. The USDA NASS have developed county coefficients based on
agricultural landuse acreages to apportion the state-level pesticide use values to the counties
(Thelin and Gianessi, 2000). However, the low spatial (and temporal) resolution of the USDA
and EPA pesticide use databases do not meet the data requirements for the GIS-based risk
assessment models for pesticide drift.
20
-------
Pesticide use data at high spatial and temporal resolution is a critical component of this
project. In the western United States, local geographic information such as agricultural fields are
reported using the Public Land Survey System (PLSS) with one square mile as the smallest unit
(i.e., a section). California has required full commercial PUR since 1990 and Arizona has
required full reporting by custom agricultural applicators since the 1960s. Data from the CPUR
system are reported at the section level with unique field identification numbers that are
sufficient to examine the risk of pesticide drift on nontarget crops. In the Arizona 1080 reports,
pesticide use data are not as complete and are reported at the section level so that individual
fields within a section cannot be identified. Arizona requires the reporting of only certain kinds
of applications, namely applications by commercial applicators, applications of pesticides
registered under section 18 registrations and certain applications of pesticides on the Arizona
Dept of Environmental Quality Groundwater Protection List. Pesticide use reporting is not
required or is at much lower spatial resolution for states other than California and Arizona.
Connecticut, New Hampshire, Massachusetts and New Mexico require commercial applicators
to report the total amount of pesticide use but do not require the applicators to report where the
pesticides were applied. Louisiana passed a bill in 1995 mandating school districts to report the
amounts of pesticides used. Several other states (New York, Oregon and Wisconsin) have
passed bills mandating development of a PUR program but have not yet begun data collection.
Each state's PUR system is unique.
The CPUR database was designed specifically to assess pesticide risk, that is to estimate
the likelihood of an adverse health or ecological effect from pesticide exposure(s) due to spray
drift. According to the California Code of Regulations, the pesticide user must report:
1. date and hour of application;
2. name or identify of the person(s) who made and supervised the application;
3. location of property treated, by county, section, township, range, base and meridian;
4. crop commodity, or site treated;
5. total acreage or units treated and amount of chemical used at the site;
6. total acreage (planted) or units at the site;
7. name of pesticide, including the US Environmental Protection Agency (USEP A) or state
21
-------
registration number, and amount used;
8. hour the treatment was completed (aerial vs ground application);
9. the operator identification number issued to the operator of the property treated;
10. the site identification number issued to the operator of the property treated.
Extensive validity checks of pesticide use data are performed by the county at the time the data
are entered and by the California Department of Pesticide Regulation (CDPR) prior to transfer
of the data into the main production database (CDPR, 2000). The data are checked for data
entry errors in the data fields for site ID, commodity code, planted acreage and location and for
statistical outliers in the data fields for acres treated and pounds of pesticide used. Typically,
less than one percent of the CPUR records are removed in the data processing. The CPUR
database for 2000 was purchased from the CDPR for use in this study.
6.2. Field Boundary Data
Geospatial digital crop field boundary and cropland use data can be used, alone or in
conjunction with pesticide use data, to better quantify the likelihood of risk of off-site pesticide
drift at the field level. However, in nearly all areas of the country, crop field boundary data are
not available. The USDA Farm Service Agency (FSA) Remote Sensing Section in Washington,
DC and the USDA Aerial Photography Field Office (APFO) in Salt Lake City, UT are in the
process of implementing GIS and Global Positioning Systems (GPS) technology to develop a
national database of digital field boundaries by 2007. In the future, USDA field offices will be
able to query the GIS database to obtain information on field and land boundaries, soil types,
crop type, producer information, place names and populations for efficient data management.
The spatial scale chosen by the USDA is the Common Land Unit (CLU) which is the smallest
unit with a permanent contiguous boundary and land cover, i.e., afield. The CLU layer will
ultimately include all farm fields, rangeland and pastureland in the United States. Consequently,
the national field boundary GIS database in conjunction with state- and county-level estimates of
pesticide usage by crop can be used to assess the potential effects of pesticide drift at the field
level for the whole country. This undertaking by the USDA represents a major breakthrough to
better assess the risk of pesticide drift and overcomes the problem of low spatial resolution in
the existing pesticide usage databases.
22
-------
In California, the CPUR database identifies each agricultural field by acreage planted,
crop commodity, identification number, and by county, section, township, range, base and
meridian. The spatial resolution of the CPUR database is insufficient to determine the shape
and location of the field within the section. Various state and federal agencies in California
involved in assessing the risk of pesticide exposure on human health and the environment have
recently developed GIS field boundary coverages for several counties in California. The U.S.
Bureau of Reclamation Mid-Pacific Region, in cooperation with the County of Yolo Department
of Agriculture (CYDA), California Department of Water Resources (CDWR) and CDPR, has
developed a draft field-boundary database coverage for Yolo County that defines all discrete
agricultural fields. The field boundaries were created by on-the-fiy digitizing in ArcEdit over 4
m pixel resolution aerial imagery supplied by CDWR in State Plane projection, and were
appended to CDWRs existing 1997 Yolo County land use database. The coverage was then
converted to a shapefile for CYDA to modify an attribute based on county agricultural permit
information.
Different methodologies including aerial photogrammetry, database technology and GIS
were used to develop the crop field boundary data for Fresno County under the San Joaquin
Crop, Water and Land Use project. The crop field boundary shapefile for Fresno County was
obtained from the Interdisciplinary Spatial Information Systems (ISIS) Center at California State
University, Fresno. A total of 36,817 fields were identified and mapped as field bonder
polygons. Total area mapped was 504,530 ha that covered 2,608 sections of land in the eight
agricultural commission districts of Fresno County. Attributes of the shapefile included the
pesticide permit identification number which can be used to link to the CPUR database.
Initially, two townships in Fresno Co (13S 13E and 16S 15E) were selected based on pesticide
usage and total farmland acreage to develop methodology to merge the field boundary and
CPUR databases (Figure 5).
However, efforts to merge the crop field boundary data for Fresno County with the
CPUR data has proven difficult due to inconsistencies in the pesticide permit identification
number. About 70% of the fields in the these two townships merged correctly and the remaining
23
-------
13S.13E
16S.15E
Ag Districts
Figure 5. Townships 13S 13E and 16S 15E within the agricultural region in
Fresno County, CA.
A. Township 13S 13E
B. Township 16S 15E
Crop
Cotton, NonRR
| Tomsto
[ Uncultivated
J Apricot
J Bean
J Cotton, NonR
I Cotton, RR
I Onion,dry
1 Soil Fum/pre
^^| Tomato ',-.'::' • ':^:>HMfc™
I P.-
Figure 6. Crop field boundary for townships A) 13S 13E and B) 16S 15E in Fresno County and
crop grown in 2001 (ISIS CSU Fresno).
24
-------
30% of the fields were manually assigned to similar-sized fields in the CPUR database and field
boundary coverage by section (Figures 6A&B). Once the CPUR database and field boundary
coverage were merged, the attributes of a crop field (i.e., pesticide use, crop type, amount of
chemical used, hectares planted, shape and size of the target and adjacent fields, method of
pesticide application, and date of application) could be obtained by querying the GIS database.
6.3. Wind Model Data
The HYSPLIT (HYbrid Single-Particle Lagrangian Integrated Trajectory) transport and
dispersion model, developed by the National Oceanic and Atmospheric Administration (NOAA)
and Australia's Bureau of Meteorology, was used to generate the atmospheric wind trajectories
of airborne particles from a specified start point and time (Draxler and Hess, 1997). The model
is available for download or can be run interactively from the NOAA website
www.arl.noaa.gov/readv/hysplit4.html The model uses previously gridded meteorological data
from the NOAA National Centers for Environmental Prediction (NCEP) weather prediction
models to calculate the dispersion rate for puffs or particles released at a given point in time and
space. The HYSPLIT model was run for 1 and 2 hours beginning at 12 PM each day for a
central location in township 13S 13E of Fresno County. Based on the HYSPLIT model, the
wind pattern was predominantly from the northwest between May and October 2000 (Figure 7).
Similar results were found at a second location within the agricultural section of Fresno County
(not shown).
6.4. Meteorological Data
The National Weather Service (NWS) collects hourly surface weather data (temperature,
precipitation, wind speed and direction, humidity, pressure) from over 500 first order stations in
the United States. These data are compiled and distributed by the National Climate Data Center
(NCDC) as the Surface Airways database. We obtained the NCDC Surface Airways database
from Earthinfo Inc. The first order weather stations represent the major airports in the United
States for the years 193 8-present; see Figure 8 for the weather stations in California. Thirty
years of hourly wind data for two weather stations in Fresno County were used to generate
historic wind patterns.
The April wind rose for Fresno Yosemite International Airport (FYIA) shows that the
25
-------
winds at Fresno blow from the northwest much of the time and rarely blow from the northeast or
the southwest (Figure 9). Similar results were obtained for Lemoore Reeves National Airport
(not shown) and the HYSPLIT model (Figure 7). In fact, the 3 spokes around the northwest
direction (WNW, W and NNW) comprise 50% of all hourly wind directions. These wind roses
also provide details on speeds from different directions. Examining winds from the northwest
(the longest spoke) one can determine that approximately 8% of the time in April at Fresno the
wind blows from the northwest at speeds between 1.8 and 3.34 meters per second. Similarly, on
this spoke it can be calculated that winds blow from the northwest at speeds between 3.34 and
5.4 m/s about 10% of the time, at speeds between 5.4 and 8.49 m/s about 6% of the time,
between 8.49 and 11.06 m/sec about 1% of the time, and less than 0.5% of the time at speeds
greater than 11.06 m/s.
A) 1-hour trajectory
B) 2-hour trajectory
Figure 7. Wind trajectories after A) 1 hour and B) 2 hours based on NOAA's HYSPLIT model
for a specified location within the agricultural district in Fresno County, CA between May and
October 2000.
26
-------
Figure 8. First order meteorological stations in
California.
(WEST
WIND SPEED
(m/s)
>- 11 1
• 8.5-11.1
• 3.6- 5.7
• 2.1 - 3.6
• 0.5- 2.1
Calms: 17.42%
Figure 9. Wind rose for Fresno
County, California in April.
27
-------
7. Assumptions and Calculations in Risk Assessment of Spray Drift on Nontarget Crops
In this report, we considered a worst-case scenario for pesticide drift due to limitations in
the spatial resolution of the CPUR database. The field boundary shapefile for Fresno County
was obtained from the Interdisciplinary Spatial Information Systems (ISIS) Center at California
State University, Fresno which we intended to merge with the CPUR database to increase the
spatial resolution to the field level. However, merging the two datafiles was problematic due to
inconsistencies in the site location identification fields. This will be addressed in a future report.
The following assumptions were made in the risk assessment of pesticide drift based on the
CPUR database:
1. We assumed that herbicides were applied to 100% of the planted acreage for each crop.
The USDA reported 95% or more of the crop acreage was treated with herbicides for the
reporting states in 2000 (USDA, 2001). The risk assessment is based on agricultural
fields with at least one pesticide application reported in the CPUR database.
Agricultural fields which were not sprayed were not reported in the CPUR database.
2. The worst-case scenario assumes that each reported agricultural pesticide application to a
target field results in offsite spray drift to all other crop fields within the same section.
Without the field boundary information, the relative locations of the target and nontarget
fields are unknown.
3. We assumed a drift range of 300 m for an aerial application and 30 m for a ground
application based on AgDRIFT model simulations and a 4.5 m/s wind.
4. We assumed that each field was a square of size equal to the square root of the acres
planted.
5. As a worst-case scenario, we assumed that crops were uniformly sensitive to herbicide
exposure within the time of maximum plant sensitivity. For some crops, the time of
increased sensitivity was determined by growth stages given in terms of days from
emergence which, in turn, was linked to planting dates. The earliest time of increased
plant sensitivity corresponded to the earliest plantings and the latest time of increased
plant sensitivity corresponded to the latest plantings of the crop. Consequently, a portion
of the crops would be sensitive at any given time with the greatest proportion of sensitive
28
-------
crops occurring in the middle of the time period of increased sensitivity.
6. We assumed that fields in a different section than where a pesticide was applied were not
subject to spray drift. This excludes the possibility of two adjacent fields, a target and
nontarget field, located in different sections.
7. We assumed that wind is omnidirectional so that drift occurs in all directions from the
target field. Wind speed and direction data are available from NOAA's HYSPLIT model
and Earthinfo Surface Airways database but cannot be incorporated in the risk
assessment without the field boundary data.
The most important factors affecting herbicide drift are droplet size, type of application (aerial
vs ground), release height and weather conditions (e.g., wind speed and direction). The type of
application is reported in the CPUR database and is used in the initial risk assessment in this
report. Wind speed and direction will be included in a more refined risk assessment when
problems with merging the field boundary shapefile and the CPUR database are resolved.
7.1. AgDRIFT Model Estimates of Spray Drift Deposition
The potential for near-field spray drift has been extensively studied in a series of field
studies conducted by the EPA Office of Research and Development and the SDTF, a coalition of
pesticide registrants (Teske et al. 2001; Bird et al, 2002). We used the AgDRIFT model
developed by EPA and the SDTF to estimate the near-field spray deposition rate from low-flight
applications of pesticides using a variety of spray droplet distributions and boom heights and a
fixed wind speed of 4.5 m/s. Table 3 lists the AgDRIFT V2.0.05 model parameters that were
used in the model simulations for aerial applications. Based on the AgDRIFT model
simulations for a medium droplet distribution and boom height of 3 m, the estimated deposition
rates were 4.6%, 1.6%, 0.5% and 0.3% of nominal application rates at distances of 50, 100, 200
and 300 m downwind (Figure 10A). For a medium droplet distribution and boom height of 4.9
m, the estimated deposition rates were 8.7%, 3.4%, 1.1% and 0.6% of nominal application rates
at distances of 50, 100, 200 and 300 m downwind (Figure 10D). Off-site spray deposition rates
increased dramatically from medium to fine to very fine droplet distribution, less so for
increasing boom height (Figure 10). Off-site deposition may reach 1% of the nominal
application rate at distances of 1600 m downwind, depending on the topography and weather
29
-------
conditions (Chester and Ward, 1984). In comparison, the estimated spray deposition rate was
1% of the application rate at a distance of about 30 m downwind for a ground application
(Figure 1).
Table 3. Basic Tier HI model parameters used in AgDRIFT simulations.
Parameter
Value
Flight lines
Swath width defn
Swath width (m)
Swath displacement defn
Fraction
Spray material
Wind speed (m/s)
Wind dir
Temperature (C)
RH (%)
Transport flux plane (m)
Surface roughness (m)
Nozzle dist (%)
Aircraft type
20
fixed
18.3
fraction swath width
0.0313
glyphosate
4.5
90
21
75
0
0.015
76.32
AT-401
7.2. Estimated Total Acreage at Risk from Ground and Aerial Applications
The proportion of the nontarget field that would be drifted upon depends upon the size,
shape and location of the field in relation to that of the target field. Average acreage planted in a
crop in Fresno County, California is under 40 ha (CPUR). For a square field of size 40 ha (0.40
km2) and a 300 m drift range for an aerial application in an adjacent target field, the estimated
size of the drift zone in the nontarget field is 0.19 km2 or 47% of the total acreage (Figure 11).
For a ground application, the estimated drift zone for a nontarget field of size 40 ha is 0.019
square km or 4.7%. A drift range of 300 m for an aerial application and 30 m for a ground
application represents a worst-case scenario based on the AgDRIFT model for a medium spray
droplet distribution and typical boom height.
30
-------
o
o
o
o
o
o
100
200
Distance- m
300
400
100
200
Distance- m
300
400
-i o
d Tr~
*- o
g
— o
100
200
Distance - m
300
400
100
200
Distance - m
300
400
Figure 10. AgDRIFT estimates of deposition rates for aerial application of glyphosale
using four spray droplet distributions and four boom heights of A) 3.0 m, B) 3.7 m, C)
4.3 m and D) 4.9m.
• 0.3 km •
-0.64 km-
Figure 11. Drift zone for a square field of size 40 ha
(100 acres) assuming a drift range of 300 m for an aerial
application.
31
-------
7.3. Crop Management Dates and Time of Maximum Plant Herbicide Sensitivity
The likelihood of drift damage to sensitive crops depends upon when, where and how the
herbicide was applied in relation to the time of increased plant sensitivity and proximity of the
nontarget crop. For most crop species, the time of increased sensitivity occurs at or prior to
anthesis. Typical crop planting dates for the major field crops in California based on
information assembled by the NASS are given in Figure 12. Statewide estimates of percent crop
planted based on NASS weekly crop reports were also used to determine planting dates for a
given year. There is considerable variability in the crop planting dates over time due to yearly
differences in meteorological conditions. For example, cotton in California was planted at a
much earlier date in 2000 than in 1998 (Figure 13). Cotton planting dates are dependent upon
temperature, i.e., the heating degree days (HDD) with a 15.6°C base must exceed 13.9 degree-h
the following five days is the general guideline for cotton planting prior to April 29 (day=120).
Many different sources of information including the USDA crop planting and harvesting
dates, weekly crop reports for Fresno County, CA, the CPUR database and studies reporting the
Wheat, winter
Wheat, durum
Sugarbeets, spring
Sugarbeets, fall
Rice
Oats, spring
Oats, fall
Hay, other-
Hay, alfalfa
Cotton
Corn, silage
Corn, grain
Beans, dry
Barley, spring
Barley, fall
150 200
Day
Figure 12. Crop planting dates for major field crops for California
(USDA, 1997).
32
-------
1 896
1997
1 99E!
1999
2000
2001
Day
Figure 13. Percent cotton planted in California for years 1996-
2001 based on weekly crop reports by the US DA NASS.
phenological stages of increased damage at sublethal doses were used to determine the time of
maximum plant sensitivity. Table El in Appendix E lists the typical planting dates and time of
maximum sensitivity for sensitive crops that are grown in Fresno County, CA. In this report, we
chose a broad period of increased sensitivity because the data were not of sufficient resolution to
refine the growth stage when damage was likely to occur. Consequently, the results of drift
effects correspond to a worst-case scenario. In this report, the time of maximum plant
sensitivities (Figure 14) were used in the risk assessment to determine the acreages at risk from
offsite drift exposure.
8. Trends in Agricultural Use of High Potential Risk Herbicides in the United States
Herbicide use on corn and soybeans accounts for about 80% of all herbicide use in
domestic agriculture. In 2000, corn growers applied 69 million kg of herbicides on corn,
followed by 31 million kg of herbicide on soybeans. Corn producers used 15 of the 28 high
potential risk herbicides to control weeds based on the USDA's Agricultural Chemical Usage
reports (USDA, 1993-2004) (Table 4). Ten of the 28 high potential risk herbicides were used on
soybeans. Wheat producers used 13 of the high potential risk herbicides. The introduction of
33
-------
o
o
CM
O
o
~
o
o
o
o
o
f
4 \
~~f AITalta
^ Dried bean, sorghum
Grapes
* Onion, Peppers
' Rice
> Sugarbeets
^ Tomatoes
1 5 9 13 17 21 25 29 33 37 41 45 49
Week
Figure 14. Time of increased plant sensitivity for crops in Fresno County,
CA.
herbicide-resistant crops has had a dramatic effect on weed management systems for cotton and
soybean and, to a lesser extent, corn in the United States. This is evidenced by the increasing
trends in glyphosate usage on cotton and soybean and the decreasing trends in the other high
potential risk herbicides in California and the United States.
8.1. Corn
Herbicides were applied to about 97% of the corn planted in 2000 (USDA, 2001).
Atrazine was the most commonly used herbicide on corn with 68% of the reported acreage being
treated in 2000. Of the high potential risk herbicides, dicamba was the most widely used
herbicide being applied to 21% of the reported acreage in 2000, followed bynicosulfuron (15%),
glyphosate (9%) and rimsulfuron (9%) (Table 4). Based on total kg a.i. applied, dicamba and
glyphosate were the top high potential risk herbicides used to control weeds on corn. The use of
the dominant corn herbicides that are not among the high potential risk herbicides such as
atrazine and acetochlor have changed very little but other corn herbicide use patterns have
changed in recent years. In particular, corn producers
34
-------
Table 4. Percent acreage of corn, Upland cotton, soybeans, rice, spring and winter wheat treated
with the 28 high potential risk herbicides in the United States in 2000 (USDA, 2001).
Chemical
2,4-D
2,4-DB
Bromoxynil
Chlorsulfuron
Clomazone
Clopyralid
Dicamba
Glufosinate
Glyphosate
Imazapyr
Imazethapyr
MCPA
Metribuzin
Met sul fur on- methyl
MSMA
Nicosulfuron
Paraquat dichloride
Picloram
Primisulfuron-m ethyl
Propanil
Prosulfuron
Quinclorac
Rimsulfuron
Sethoxydim
Sul fom eturon -me thyl
Thi fens ulf uro n-m ethyl
Tribenuro n-m ethyl
Triclopyr
Corn
8
4
9
21
2
9
2
3
2
15
1
9
4
9
<1
Cotton
1
6
2
<1
56
14
<1
<1
Soybean
5
<1
<1
62
12
4
<1
2
6
<1
Rice
17
32
12
1
<1
62
25
18
Spring
Wheat
45
26
14
25
20
44
3
2
4
15
Winter
Wheat
13
1
9
4
7
3
<1
<1
<1
6
6
35
-------
have increased their use of glyphosate and decreased their use of dicamba and 2,4-D (Figure
15). Corn growers in California also applied more glyphosate and less dicamba and 2,4-D
(Figure 16). Glyphosate is currently the most widely used high potential risk herbicide on corn
with 19% of the reported acreage being treated in 2003 (USDA, 2004). Corn producers are
favoring the use of glyphosate in lieu of the more damaging high potential risk herbicides in
order to avoid drift problems. Agricultural usage of the SU corn herbicides has remained steady
over time with nicosulfuron and rimsulfuron being applied to 11% and 10%, respectively, of the
reported acreage in 2003. With respect to glufosinate-resistant corn, growers have not adopted
this technology as widely as glyphosate-resistant corn. Liberty Link® corn represents less than
5% of the corn planted in the United States based on only 2-3% of the total acreage being treated
with glufosinate in the years 2000-2003. Roundup Ready™ corn represents a much larger market
share than Liberty Link® corn because glyphosate provides more effective control of weeds than
is achieved by glufosinate.
8.2. Soybeans
Pesticide use on soybeans in terms of total kg a.i. applied is second only to that on corn.
8_
ZJ
E
8_
2,4-D
Bromoxynil
Clopyralid
Dicamba
Glufosinate
Glyphosate
Nicosulfuron
1992
1994
1996
1998
2000
2002
Figure 15. Trends in herbicide use on corn in the United States in
the years 1992-2003 based on the annual USDA Agricultural
Chemical Usage reports.
36
-------
2,4-D
Bromo>ynil
Paraquat
Dicamba
Glyphosate
Nicosulfuron
1992
1994
1996
I
1998
2000
2002
Year
Figure 16. Trends in herbicide use on corn in California in the years
1992-2002 based on the CPUR database.
Herbicides were applied to 97% of the soybean acreage. There were at least 70 individual
herbicides or mixtures registered for use on soybeans in the early 1990's. In 1992, soybean
producers in the 16 major producing states applied trifiuralin to 35% of the 21.5 million ha
planted, imazethapyr to 29%, pendimethalin to 21%, metribuzin to 14% and glyphosate to 7% of
the total acreage. In 1996, the most widely used herbicide was imazethapyr applied to 38% of
the soybean acreage, followed by glyphosate (28%), pendimethalin (25%), and trifiuralin (21%).
In 2002, the most widely used herbicide was glyphosate applied to 78% of the soybean acreage,
followed by imazethapyr (9%), pendimethalin (9%), and trifiuralin (7%). The dominant
soybean herbicides changed dramatically since the introduction of glyphosate-resistant soybean
in 1996. Soybean farmers have adopted RR soybean technology at a rapid rate because it greatly
simplifies weed management. The new technology offers greater flexibility in the control of
weeds using a single postemergent herbicide without crop injury or crop rotation restrictions. In
addition, off-site drift of glyphosate is less damaging to nontarget susceptible crops than that of
other soybean herbicides such as 2,4-D and imazethapyr. Agricultural usage of the other high
potential risk herbicides on soybean peaked in 1997, followed by a sharp decline (with the
37
-------
possible exception of 2,4-D) due largely to the commercial success of RR soybeans (Figure 17).
8.3. Wheat
The most widely used herbicide on spring and winter wheat in area applied and quantity
was 2,4-D with 61% of reported spring wheat acreage and 24% of reported winter wheat acreage
treated in 1997 (USDA, 1998). MCPA, dicamba and tribenuron were also widely used on
spring wheat with 42%, 27% and 20%, respectively, of reported acreage treated in 1997.
Metsulfuron, chlorsulfuron and tribenuron were widely used on winter wheat with 14%, 11%
and 8%, respectively, of reported acreage treated in 1997. Herbicides were applied to about
37% of the winter wheat acreage and to 97% of durum wheat acreage in ND in 2000 (USDA,
2001). The reliance of these high potential risk herbicides to control weeds in wheat cropping
systems in the U.S. has been fairly consistent over time because there is no herbicide-tolerant
wheat. 2,4-D, MCPA, dicamba and tribenuron are still heavily used on spring wheat but 2,4-D
use decreased from 61% of the reported acreage in 1997 to 45% in 2000 (Table 4). Bromoxynil
and glyphosate on spring wheat had increased to 26% and 20%, respectively of reported acreage
treated in 2000 (USDA, 2001). These trends in herbicide use were also evident in the quantity
used on spring wheat (Figure 18). 2,4-D remained the most widely used herbicide on winter
2,4-D
Clomazone
Glyphosate (/10)
Irnazethapyr
Metribuzin
Sethoxydirn
Figure 17. Trends in herbicide use on soybeans in the
United States in the years 1992-2002 based on the annual
USDA Agricultural Chemical Usage reports.
38
-------
wheat in 2000 but was applied to only 13% of the reported acreage (USDA, 2001). Glyphosate
use in terms of total kg a.i. applied on winter wheat increased from 1998 to 2000 in concert with
a drop in 2,4-D usage (Figure 19). These results indicate a shift in herbicide use on spring and
winter wheat away from the more potentially damaging herbicides (i.e., 2,4-D) towards safer
herbicides (i.e., glyphosate and bromoxynil) in order to avoid drift damage to susceptible crops.
In 2002, wheat producers applied MCPA to 44%, 2,4-D to 36%, bromoxynil octanoate to 19%,
dicamba to 15% and glyphosate to 6% of the total ha planted (CPUR). The top wheat herbicides
in California did not change over time. Total acreage planted in wheat in California peaked at
268,000 ha in 1996, followed by a steady decline to 124,000 ha in 2002. Herbicide use on
wheat in California showed a decline from 1998-2002 in concert with the decline in acreage
planted (Figure 20) and was similar to domestic herbicide use on spring wheat. In 2000, hard
red spring wheat accounted for 78% of the total ha planted, durum spring wheat accounted for
16% and white wheat accounted for the remaining 6%. Wheat acreage in Fresno County,
California accounted for less than 10% of the state acreage and declined from a peak of 19,000
ha in 1998 to 12,000 ha in 2002.
8 .
s.
2,4-D
Bromoxynil
Metribuzin
Dicamba
MCPA
Glyphosate
Tribenuron
Figure 18. Trends in herbicide use on spring wheat in the
United States in the years 1992-2002 based on the annual
USDA Agricultural Chemical Usage reports.
39
-------
£ s.
§-
2,4-D
Bromoxynil
Metribuzin
Dicamba
MCPA
Glyphosate
Tribenuron
Figure 19. Trends in herbicide use on winter wheat in the
United States in the years 1992-2002 based on the annual
USDA Agricultural Chemical Usage reports.
2,4-D
Bromoxynil
Dicamba
MCPA
Glyphosate
Figure 20. Trends in herbicide use on wheat in California in
the years 1992-2002 based on the California Pesticide Use
database.
40
-------
8.4. Upland Cotton
The most widely used herbicides on Upland cotton were trifluralin, fluometuron,
MSMA, pendimethalin and pyrithiobac-sodium with 55%, 44%, 29%, 28% and 23%,
respectively, of the reported acreage treated in 1997. Herbicides were used on 97% of the
reported acreage. The weed management program developed in the 1970s was based on the
herbicide trifluralin as preplant incorporated for residual control of grasses and several broadleaf
weeds, followed by a banded preemergence application of fluometuron. This was followed by
multiple post-directed applications of DSMA/MSMA mixed with cyanazine for control of
emerged grasses and nutsedge as well as germinating weeds. More than 80% of the total
acreage of U.S. cotton received this treatment in the 1980s (McWhorter et al, 1992). The
disadvantage of this system is that trifluralin must be mechanically incorporated into the soil
because of its high volatility and potential for photodegradation. Consequently, each tillage pass
over the field results in a loss of water in the soil between 1 and 4 cm. Pendimethalin and
trifluralin are very similar products. Pyrithiobac is a newer herbicide that can be applied
preemergence, postemergence, or both and can be used safely over the top of cotton without
injury. However, pyrithiobac does not control cocklebur, ragweed, lambsquarters and sicklepod
which are important weedy species in cotton.
The introduction of herbicide-tolerant transgenic cotton revolutionized the weed
management systems for cotton by offering cotton growers postemergent herbicides that kill
many grasses and/or broadleaf weeds with no crop injury. BXN® cotton was introduced in 1995
and Roundup Ready™ cotton was introduced in 1997. Bromoxynil controls many broadleaf
weeds but does not control grasses, pigweeds or sicklepod. On the other hand, glyphosate has a
broad spectrum of activity which includes most of the problematic annual and perennial grass
and broadleaf weeds infesting cotton fields. Consequently, cotton producers have eagerly
adopted RR cotton technology because of its simplicity, flexibility and broad spectrum of
activity. Glyphosate use on Upland cotton increased from 700,000 kg a.i. applied to 14% of the
reported acreage in 1997 to 5,731,000 kg a.i. applied to 69% of the reported acreage in 2003
(Figure 21). Planting of RR cotton was highest in South Carolina due to its effectiveness on
sicklepod and palmer amaranth. California was slow to adopt RR cotton because of the
41
-------
8 _
2.4-D
Bromoxynil
Clomazone
Dicamba
MSMA
Glyphosate
Setho»/dim
A
H
•
D
1992
1994
1996
1998
2000
2002
Figure 21. Trends in herbicide use on Upland cotton in the United
States in the years 1992-2003 based on the annual USDA
Agricultural Chemical Usage reports.
-"= H
o 5
Figure 22. Trends in herbicide use in cotton in California in the
years 1992-2002 based on California Pesticide Use database.
42
-------
requirement of a three-year testing period by the SJV Cotton Board. Approximately 76% of the
400,000 ha of cotton were planted in Acala and Upland cotton in California. Approximately 40-
50% of the Upland cotton grown in California is herbicide-tolerant, the majority being RR
cotton. Glyphosate is the top cotton herbicide in California (Figure 22).
9. Results - Timing and Method of Pestidde Applications for High Potential Risk
Herbicides Herbicide drift to nontarget terrestrial plants is a problem in many areas of the
United States, especially when producers apply herbicides under windy conditions. Unintended
spray drift will result in adverse effects to some plant species. In assessing the risk of drift
damage to nontarget crops in Fresno County in 2000, the frequency and amount of herbicide
applications during the times of increased plant sensitivity were used to identify the herbicides
that pose the greatest potential risk to sensitive crops. Damage from drift is assumed to occur
only during the times of increased plant sensitivity which correspond to weeks 10-38 for the
more susceptible crops (Figure 14). This period corresponds roughly to the March 16 to
October 15 period when the most damaging herbicides such as 2,4-D, 2,4-DB and MCPA are
banned from use in the SJV of California. We examined the temporal patterns of herbicide
application for each of the high potential risk herbicides in Fresno County in 2000 based on the
CPUR database. The high potential risk herbicides, CGA-152005 (or prosulfuron), clomazone,
metsulfuron-methyl, primisulfuron, quinclorac, thifensulfuron and tribenuron, were not
registered for use in California in 2000. Clomazone and quinclorac were registered for use in
CA in 2002. No glufosinate applications were reported in Fresno County in 2000.
9.1. 2,4-D and 2,4-DB
The top three major crops in Fresno County, California (grapes, cotton and tomatoes) are
particularly sensitive to 2,4-D and 2,4-DB and significant damage to these crops can occur at
drift levels of these compounds (Table C2). By restricting their use in the SJV between March
16 and October 15, drift damage due to 2,4-D and 2,4-DB is minimal in Fresno County. The
herbicide 2,4-D is applied primarily by ground application as a dimethylamine salt formulation
to wheat, almond, barley and oats in Fresno County when plants are less prone to damage
(Figure 23). In 2000, less than 12 applications of 2,4-D and 2,4-DB occurred at a time when
43
-------
sensitive crops were most susceptible to damage from offsite drift exposure. These results
indicated that applications of 2,4-D and 2,4-DB pose minimal risk to crops in Fresno County
due to offsite drift. Consequently, these herbicides were not included in the GIS spatial analysis
to determine whether any fields planted in the sensitive crops were within drift distances from
the edge of the target fields.
9.2. Bromoxynil
Bromoxynil is used as a postemergent herbicide on BXN® cotton, wheat, oats, alfalfa,
garlic and onions in Fresno County. The majority of the aerial applications of bromoxynil were
used to control weeds on wheat fields in the first quarter and occurred before grapes became
sensitive to pesticide exposure (Figure 24). Of the approximately 600 reported applications of
the compound primarily as bromoxynil octanoate, there were a number of ground applications
and a few aerial applications over the top of BXN8 cotton that may have drifted offsite onto
non-BXN® cotton and grapes (Figure 25). In addition, a number of bromoxynil applications on
garlic, onion and alfalfa in weeks 10-16 occurred at a time when grapes were sensitive to
herbicide exposure.
o
o
-------
-'-
'I'
-------
9.3. Clopyralid
Less than 200 kg a.i. of clopyralid was applied in Fresno County in 2000 to control
weeds in landscape maintenance, rights-of-way, pastures and sugarbeets. Several applications of
the compound occurred when grapes, beans, cotton and tomatoes were highly sensitive to
herbicide exposure (Figure 26). Due to minimal use of the compound, we did not include
clopyralid in the GIS spatial analysis to estimate the acreage of the sensitive crops within drift
range of the target fields. About 950 kg a.i. of dicamba was used in Fresno County in 2000 to
control weeds in landscape maintenance, corn and wheat. Most of the ground and aerial
applications of dicamba in the first 10 weeks were used to control weeds in wheat fields (Figure
27). Weekly dicamba use was typically less than 115 kg a.i. in the spring and summer (weeks
10-28) when susceptible crops were most sensitive to herbicide exposure.
9.4. Dicamba
About 950 kg a.i. of dicamba was used in Fresno County in 2000 to control weeds in
landscape maintenance, corn and wheat. Most of the ground and aerial applications of dicamba
in the first 10 weeks were used to control weeds in wheat fields (Figure 27). Weekly dicamba
o
IT! '
"D
-------
'-
o
-------
Cotton
Fruits
Mills
Vegetables
Crap Type
Figure 28. Total kg a.i. of glyphosate applied by crop in Fresno
County in 2000 (CDPR).
o
o
o
o
O
o
CD
O H
cs
Aerial
Ground
Cotton
Grapes
Onion, Peppers
1 5 9 13 17 21 25 29 33 37 41 45 49
Week
Figure 29. Glyphosate usage in Fresno County, CA in 2000
(CPUR).
48
-------
which is 34% of the total cotton acreage planted. BXN® cotton was planted on about 13,300
acres or 4% of the total cotton acreage and the remaining 62% was conventional cotton. Most of
the glyphosate used in cotton production in Fresno County was applied on RR cotton rather than
on conventional or BXN® cotton. Typically, two over-the-top applications of glyphosate at least
10 days apart are made through the fourth true-leaf stage of cotton to provide control of most
weeds in a RR cotton weed management system (Cotton Farming, 2000). More than two
postemergent glyphosate applications maybe required in heavily infested fields in the absence
of soil-applied herbicides. The Roundup® label prohibits over-the-top application on cotton
larger than the four-leaf stage. Spray drift from these postemergent glyphosate applications is a
concern because unintended nontransgenic crops are in the early stage of growth development
and are highly sensitive to herbicide exposure. Weekly use of glyphosate on RR cotton often
exceeded 2,300 kg a.i. at times of increased plant sensitivity for most of the sensitive crop
species (Figure 30).
In Fresno County, glyphosate is also heavily used in the production of grapes, almonds,
oranges, and tomatoes. Glyphosate is typically applied as a postemergent herbicide with a
controlled application or with low pressure fiat fan nozzles in vineyards. Young vineyards were
o
o
-
Is §
-------
very sensitive to herbicides and glyphosate exposure at 1% of the label rate resulted in fewer and
smaller berries in newly planted Lemberger wine grapes as well as established 1-year-old and 2-
year-old plants (Al-Khatib et al., 1993). Glyphosate is commonly applied to the major crops
such as grapes, almonds and oranges at the time of increased plant sensitivity (Figures 31A-C).
We assumed that the postemergent glyphosate applications on grapes and wine grapes in weeks
12-23 were on mature vineyards which are less sensitive to herbicide exposure. Note that
reported glyphosate applications on tomatoes were low at the time of increased plant sensitivity
(Figure 3 ID). The majority of glyphosate applications on spring and early summer were ground
applications which limit the range of offsite spray drift to about 30 m from the edge of the target
field. Consequently, drift damage is likely low for an individual application but may be
appreciable when accumulated across all applications during the time of increased plant
sensitivity.
9.6. Imazethapyr
Less than 200 applications and 363 kg a.i. of imazethapyr were reported in Fresno
County in 2000, primarily to control weeds in alfalfa. A number of ground and aerial
applications of imazethapyr occurred at the time of increased plant sensitivity (Figure 32).
Weekly use of the compound peaked at about 57 kg a.i. at about the time grapes and corn were
sensitive to herbicide exposure and was typically less than 18 kg a.i. later in the spring and
summer. Drift damage is expected to be minimal due to the low use of imazethapyr in Fresno
County.
9.7. MCPA
About 5,200 kg a.i. of the dimethlyamine salt formulation of the herbicide MCPA was
used in Fresno County in 2000 to control weeds in small grain crops. All applications of MCPA
occurred in the first 10 weeks when susceptible crops such as grapes, cotton and sugarbeets were
less likely to be damaged (Figure 33). The California Department of Pesticide Regulation
prohibits the use of MCPA between March 16 and October 15.
9.8. Paraquat
According to the California Code of Regulations Title 3 Section 6, aerial applications of
paraquat are restricted to a maximum discharge height of 3 m in winds of less than 4.5 m/s while
50
-------
ground applications are unrestricted. About 50% of the reported usage of 105,000 kg a.i. of
paraquat was applied to 52% of the cotton acreage in Fresno County, followed by almond and
grapes with 12% and 11%, respectively, of the total a.i. applied. The compound was most
heavily applied as a post-emergence herbicide on winter fallow beds prior to week 13 and after
week 38, typically when corn was not sensitive to herbicide exposure (Figure 34). Agricultural
use in weeks 9-13 and after week 38 was mostly on cotton (Figure 35A). Paraquat was typically
applied to almonds at a time of increased plant sensitivity for corn and cotton (Figure 35B) and
earlier in the season to grapes (Figure 35C).
B. Almonds
3-
C. Oranges
D. Tomatoes, processing
Aerial
Ground
Alfalfa
Sorghum
-* Corn
Cotton
Grapes
^ Onion, Peppers
Rice
8-
Groi ind
Figure 31. Glyphosateuse on A) grapes, B) almonds, C) oranges and D) processing tomatoes in
Fresno County in 2000 (CPUR).
51
-------
Aerial
Ground
o
LO ^
Sorghum
Corn
Grapes
1,
1 5 9 13 17 21 25 29 33 37 41 45 49
Week
Figure 32. Imazethapyr usage in Fresno County, CA in 2000
(CPUR).
Aerial
Ground
Cotton
Grapes
Sugarbeets
1 5 9 13 17 21 25 29 33 37 41 45 49
Week
Figure 33. MCPA usage in Fresno County, CA in 2000 (CPUR).
52
-------
-~
'I'
Cor
Cotton
1 5 9 13 17 21 25 29 33 37 41 45 49
Week
Figure 34. Paraquat usage in Fresno County, CA in 2000 (CPUR).
9.9. Sethoxydim
Most of the agricultural usage of 1,300 kg a.i. of sethoxydim occurred at the time of
increased plant sensitivity in Fresno County in 2000 (Figure 36). About 45% of the compound
was used on alfalfa, followed by cotton with 24% of the total kg a.i. applied. Sethoxydim is a
postemergence herbicide applied on alfalfa when grass weeds have emerged and are small,
typically when corn and sorghum are most sensitive to herbicide exposure (Figure 37A).
Similarly, one application of sethoxydim is typically applied postplant over the top or
postemergence directed on cotton at time of increased plant sensitivity fornontarget crops
(Figure 37B).
10. Results - GIS-based Risk Analysis of Selected High Potential Risk Herbicides
Of the 28 high potential risk herbicides, bromoxynil, dicamba, glyphosate and
sethoxydim represent the greatest risk to sensitive crops in Fresno County, California due to
offsite drift exposure because these herbicides are used in quantity during times of increased
plant sensitivity. Several of the most important agricultural crops such as cotton, table grapes
and wine grapes, which account for 42% of the 11 million ha of cropland in Fresno County, are
susceptible to damage at sublethal levels of these herbicides (Table C2). Glyphosate drift poses
53
-------
A. Cotton
i-
g:
B. Almonds
S ~> Aerial
^ Ground
E
CD
6
Cotton
1 5 9 13 17 21 25 29 33 37 41 45 49
Week
1 5
13 17 21 25 29 33 37 41 45 49
Week
C. Grapes
g
Cotton
1 5 9 13 17 21 25 29 33 37 41 45 49
Week
Figure 35. Paraquat use on A) cotton, B) almonds and C) table, raisin and wine grapes in
Fresno County, CA in 2000 (CPUR).
the highest risk because of its broad spectrum of activity and its use far exceeds that of the other
three herbicides combined. Of the four high potential risk herbicides, sethoxydim is of least
concern in Fresno County because the sensitive crops at risk, i.e., corn and sorghum, represent
less than 13,000 ha or < 0.1% of total cropland in the county. A GK-based spatial analysis of
these four high potential risk herbicides was undertaken using the CPUR database to determine
the magnitude and location of the cropland acreage of sensitive crops that were potentially at
risk from offsite spray drift of these herbicides during the time of increased sensitivity.
54
-------
o
00
Corn
1 5 9 13 17 21 25 29 33 37 41 45 49
Week
Figure 36. Sethoxydim use in Fresno County, CA in 2000
(CPUR).
A. Alfalfa
13 17 21 25 29 33 37 41 45 49
Week
B. Cotton
° -i Aerial
Ground
Sorghum
> Corn
1 5 9 13 17 21 25 29 33 37 41 45
Week
Figure 37. Sethoxydim usage on A) alfalfa and B) cotton in Fresno County, CA in 2000
(CPUR).
55
-------
Based on the CPUR database for Fresno County in 2000, the risk of pesticide spray drift
to unintended agricultural fields was quantified by the total ha planted of susceptible crops that
were in sections where at least one application of a high potential risk herbicide was made at the
time of increased plant sensitivity. For bromoxynil, we assumed that postemergence
applications of the herbicide on cotton were made to BXN® cotton which resulted in damage to
adjacent fields of non-BXN® cotton, grapes and wine grapes. Similarly, postemergence
applications of glyphosate on cotton were made to RR cotton which resulted in potential damage
to non-RR cotton (including BXN® cotton), alfalfa, corn, grapes, wine grapes, onion, rice,
sorghum and tomatoes. We assumed that applications of glyphosate on grapes and wine grapes
at the time of increased plant sensitivity were made to mature plants so the target fields were not
at risk due to glyphosate exposure.
Because glyphosate use greatly exceeds that of bromoxynil, dicamba and sethoxydim
combined, glyphosate poses the greatest risk to agricultural crops due to offsite spray drift in
Fresno County (Table 5). In terms of total cropland at risk, dicamba and sethoxydim posed the
least risk of the four high potential risk herbicides and were estimated to damage less than 2,000
ha planted in crops sensitive to these herbicides. About 4,000 ha of cotton and 2,000 ha of
grapes and wine grapes were at risk from potential drift of bromoxynil applications on BXN®
cotton, wheat and oats. Considerably more cropland of grapes, non-RR cotton, alfalfa,
tomatoes, wine grapes, corn and onions were at risk from potential drift of glyphosate
applications. About 46% of alfalfa acreage, 39% of corn acreage, 84% of rice acreage and 28%
of tomato acreage were potentially at risk from unintended exposure to glyphosate at the time of
increased plant sensitivity.
These results in Table 5 suggest that damage to grapes from offsite drift of glyphosate
represents the greatest potential crop yield loss. However, only newly established and young
plantings of grapes and wine grapes are sensitive to glyphosate at drift levels (Al-Khatib et al.,
1993). Many preemergent and contact herbicides pose a risk of damage to young vineyards
because young vine roots are shallow and foliage is close to the ground. Bearing grape acreage
in Fresno County in 2000 was estimated at 73,397 ha for table and raisin grapes and 22,851 ha
for wine grapes (CASS, 2001); nonbearing grape acreage (acreage due to come on line in the
56
-------
Table 5. Total hectares planted of various crops subject to offsite spray drift from ground and
aerial applications within the same section for Fresno County, CA in 2000.
Crop (sensitive period)
Alfalfe (03/04-07/29)
Ground
Aerial
Ground & Aerial
Beans (04/15-07/29)
Ground
Aerial
Ground & Aerial
Corn (04/02-09/21)
Ground
Aerial
Ground & Aerial
Cotton (05/23-07/29)
Ground
Aerial
Ground & Aerial
Grapes (03/23-06/06)
Ground
Aerial
Ground & Aerial
Grapes, wine (03/23-06/06)
Ground
Aerial
Ground & Aerial
Bromo xynil
3,713
3,564
149
0
1,362
1,240
97
25
621
493
128
0
Dicamba
443
410
0
33
0
0
0
0
580
571
0
9
207
207
0
0
132
99
0
33
Glyphosate
16,903
9,566
3,332
4,005
3,955
3,106
410
440
23,693
21,137
1,272
1,284
44,538C
35,841
251
8,446
6,770e
5,998
36
736
Sethoxydim
1,770
1,334
436
0
Total Crop Acreage
36,561
4,935
10,254
51,378aRR
5,366bBXN®
91,729 Nontransgen
71,488 CAPUR
73, 397d bearing
l,633dnonbearing
22,851 CAPUR
17, 6 13d bearing
l,653dnonbearing
57
-------
Table 5 (continued).
Onion (03/1 8-07/29)
Ground
Aerial
Ground & Aerial
Rice (05/27-07/30)
Ground
Aerial
Ground & Aerial
Sorghum (04/1 5-07/29)
Ground
Aerial
Ground & Aerial
Tomatoes (04/29-07/29)
Ground
Aerial
Ground & Aerial
81
81
0
0
2,688
2,400
0
288
1,902
1,902
0
0
57
57
0
0
15,215
13,709
728
778
57
57
0
0
6,729
2,272
66
54,319
a Cotton acreage with reported glyphosate applications between April 21 and September 23 and little or no reported
applications of other postemergent over-the-top herbicides such as pyrithiobac-sodium, clethodim, MSMA and
bromoxynil were assumed to be RR cotton.
b Cotton acreage with reported bromoxynil applications were assumed to be BXN cotton.
c There were 18,050 ha of grapes that were treated with glyphosate between March 23 and June 6; these grape
acreage were assumed to be mature plants and were not included in the total acreage at risk. Older, more
established plantings are less sensitive to the herbicide. Only the 1,653 ha of nonbearing grapes are sensitive to
glyphosate exposure.
d California Agricultural Statistics Service.
e There were 7,817 ha of wine grapes that were treated with glyphosate between March23 and June 6; these grape
acreage were assumed to be mature plants and were not included in the total acreage at risk. Only the 1,653 ha of
nonbearing wine grapes are sensitive to glyphosate exposure.
58
-------
next three years) was estimated at 1,633 ha for table and raisin grapes and 1,653 ha for wine
grapes. We assumed that only the 3,286 ha of nonbearing table, raisin and wine grapes were
potentially at risk due to offsite pesticide drift. In comparison with the grape acreage estimated
by CASS, it appears that about 3,600 of the 22,850 ha of wine grapes in the CPUR database
were miscoded and should be table and raisin grapes.
About 12,100 of the 17,612 ha of bearing wine grape acreage and 68,400 of the 73,400
ha of bearing table and raisin grape acreage in Fresno County in 2000 were planted in 1992 or
earlier (CASS, 2001). Since the major replanting in the early 1990s, the acreage of new
plantings of grapes have declined between 1993 and 2000. In 2000, only 146 haof wine grapes,
274 ha of raisin grapes and 174 ha of table grapes were planted in Fresno County.
Consequently, less than 4% of total grape acreage in Fresno County in 2000 were nonbearing.
These findings indicate that most of the grape acreage in Fresno County were older, more
established plantings that were much less sensitive to unintended pesticide exposures than the
nonbearing grapes planted within the last three years.
When the multiplier of either 47% for an aerial application or 4.7% for a ground
application was included in the calculation (Section 7.2), the estimated total acreage at risk was
highest for unintended fields of alfalfa due to drift from aerial applications of glyphosate,
followed by non-RR cotton, tomatoes and grape acreage potentially damaged due to drift from
aerial and ground applications of glyphosate (Table 6). For sections where there were both
ground and aerial applications of a pesticide at the time of maximum plant sensitivity, the higher
multiplier of 47% was used in the calculation. About 3,900 ha of alfalfa or 11% of alfalfa
acreage in Fresno County was estimated to be at risk from offsite pesticide spray drift of
glyphosate. This, in part, was due to the fact that alfalfa had the widest period of increased plant
sensitivity from March 4 to July 29. In comparison, about 2,200 ha of non-RR cotton or 2.3% of
BXN® and conventional cotton acreage were estimated to be potentially damaged due to
unintended exposure of glyphosate applied to adjacent fields by ground and aerial spray
equipment in Fresno County. Cotton was less at risk from glyposate drift than alfalfa because
there were more aerial applications of glyphosate prior to May 23 than at the time when cotton
plants were more sensitive to pesticide exposure. Similarly, about 1,300 ha of processing
59
-------
tomatoes or 2.5% of tomato acreage in Fresno County was estimated to be potentially damaged
from unintended exposure of glyphosate applied to adjacent fields. Herbicides other than
glyphosate posed little or no risk to susceptible crops in Fresno County with the possible
exception of sethoxydim drift on corn and bromoxynil drift on non-BXN® cotton. About 2.6%
of corn acreage and 0.2% of cotton acreage were estimated at risk due to spray drift from
sethoxydim and bromoxynil applications, respectively. In general, more crop acreage was at risk
due to exposure from aerial applications than ground applications because spray from an aerial
application was assumed to drift 10 times farther than that from a ground application.
10.1. Bromoxynil Risk
The shift towards transgenic crops and postplant postemergent applications on these
crops increases the likelihood of drift damage to conventional crops which may have emerged in
adjacent areas. For bromoxynil, the most likely drift problem is from a ground application on
BXN® cotton that moves offsite to an adjacent non-BXN® cotton field (Figure 25 and Table 5).
The sections where non-BXN® cotton were planted and bromoxynil was applied at the time of
increased plant sensitivity were scattered throughout the cotton-growing area in Fresno County
(Figure 38). Similar to glyphosate, weight loss in wine grapes exposed to bromoxynil at early
spring and midsummer was greatest for new plantings and least for 2-year-old plants (Al-Khatib
et al., 1993). Consequently, only the 3,286 ha of nonbearing grapes were potentially at risk from
offsite spray drift of bromoxynil. The grape acreage at risk was actually less because only 1,362
ha of table and raisin grapes and 621 ha of wine grapes were grown in sections where at least
one bromoxynil application was made between March 23 and June 6 (Table 6). The sections
where grapes were potentially at risk from bromoxynil drift are shown in Figure 39. Most
likely, only a portion of the grape acreage in these sections were nonbearing grapes but the
CPUR database does not distinguish between bearing and nonbearing grapes.
10.2. Dicamba Risk
Because dicamba usage was minimal in Fresno County, there were only a few sections
where crop acreage of sensitive crops and dicamba applications were reported (Figure 40). Crop
acreage in these sections where dicamba applications were made at time of increased plant
sensitivity ranged from 81 ha for tomatoes to 580 ha for cotton (Table 5). Most of these
60
-------
potential offsite pesticide spray drift events were from ground applications on corn and wheat
which represent less than 0.3% of total cropland in Fresno County.
10.3. Glyphosate Risk
Glyphosate was widely used in large quantities at the time of increased plant sensitivity
as a postplant postemergent herbicide on many crops such as RR cotton, almonds, grapes,
oranges and tomatoes (Figures 30, 31). In Fresno County, the most likely drift problem is from
a postemergent ground application of glyphosate on RR cotton that moves offsite to an adjacent
non-RR cotton field (Figure 30). About 35% of total cotton acreage in Fresno County is RR
cotton which is often grown beside non-RR cotton (Figure 6). Consequently, there were many
sections where non-RR cotton acreage were at risk due to glyphosate drift at the time of
increased plant sensitivity (Figure 41). The next most likely drift problem from glyphosate
application is to alfalfa and tomatoes which are grown in similar parts of Fresno County as
cotton (Figure 42), followed by corn (Figure 43A) and onions (Figure 43B). Note that about
20% of the alfalfa acreage in Fresno County were at potential risk due to offsite spray drift from
glyphosate applied at the time of increased plants sensitivity (Table 6). There were many more
sections with grape acreage where glyphosate was applied between March 23 and June 6
(Figures 43C-D), but only a small number of these sections included the nonbearing grape
acreage of < 8,200 acres. In a worst-case scenario, 4,034 acres of nonbearing table and raisin
grapes and 1,908 acres of nonbearing wine grapes were at risk due to glyphosate drift from aerial
applications between March 23 and June 6 (Table 6). The least likely drift problem from
glyphosate is to sorghum and rice which were grown on 162 and 5,615 acres, respectively, in
Fresno County in sections where glyphosate was applied by ground (Figures 43E-F).
10.4. Sethoxydim Risk
There were only a few sections where crop acreage of sensitive crops and sethoxydim
applications were reported for Fresno County in 2000 (Figure 44). Crop acreage in these
sections where sethoxydim applications were made at time of increased plant sensitivity was
1,770 ha for cotton and 57 ha for sorghum (Table 5). Most of these potential offsite pesticide
spray drift events were from ground applications on alfalfa and cotton (Figure 37).
61
-------
Table 6. Estimated total crop acreage (ha) in drift range from ground and aerial applications
within the same section for Fresno County, CA in 2000.
Crop
Alfalfa
Beans
Corn
Cotton
Grapes
Grapes, wine
Onion
Rice
Sorghum
Tomatoes
Total
Appl. Method
Ground
Aerial
Ground
Aerial
Ground
Aerial
Ground
Aerial
Ground
Aerial
Ground
Aerial
Ground
Aerial
Ground
Aerial
Ground
Aerial
Ground
Aerial
Bromoxyni
1
168
70
58
57
23
60
436
Dicamba
19
13
0
0
27
4
10
0
5
15
4
0
97
Glyphosate
450
3,448
146
399
994
1,202
77a
767a
78b
363C
113
136
89
0
3
0
644
708
9,617
Sethoxydim
63
205
3
0
271
Total Crop Acreage
36,561
4,935
10,254
51,378RR
5,366BXN®
91,729 Nontransgen
71,488 CAPUR
73, 397 bearing
1,633 nonbearing
22,851 CAPUR
17,613 bearing
1,653 nonbearing
6,729
2,272
66
54,319
452,244
a Assuming only nonbearing grapes are at risk due to pesticide exposure, only 1,633 ha of grape were at risk of
unintended exposure from either ground or aerial applications.
b Assuming only nonbearing grapes are at risk due to pesticide exposure, only 1,653 ha of wine grape were at risk
of unintended expo sure from ground ap plications.
c Assuming only nonbearing grapes are at risk due to pesticide exposure, only 772 ha of the 1,653 ha of wine grape
were at risk of unintended expo sure from aerial applications.
62
-------
Figure 38. Sections in Fresno County where non-BXN® cotton
was planted in 2000 and bromoxynil was or was not applied by air
(A) and/or ground (G) spray equipment between May 23 and July
29(CPUR).
A. Table and raisin grapes
B. Wine grapes
Figure 39. Sections in Fresno County where A) grapes and B) wine grapes were grown in 2000
and bromoxynil was or was not applied by air (A) and/or ground (G) spray equipment between
March 23 and June 6 (CPUR).
63
-------
A. Cotton
B. Alfalfa
C. Table and raisin grapes
D. Wine grapes
Figure 40. Sections in Fresno County where A) cotton, B) alfalfa, C) table and raisin grapes, D)
wine grapes and E) tomato fields were grown in 2000 and dicamba was or was not applied by air
(A) and/or ground (G) spray equipment at time of increased plant sensitivity (CPUR).
64
-------
Figure 41. Sections in Fresno County where non-RR cotton was planted in
2000 and glyphosate was or was not applied by air (A) and or ground (G)
spray equipment between May 23 and July 29 (CPUR)
A. Alfalfa B. Tomatoes
Figure 42. Sections in Fresno County where A) alfalfa and B) tomato were planted in 2000 and
glyphosate was or was not applied at time of increased plant sensitivity (CPUR).
65
-------
A. Corn
B. Onions
C. Table and raisin grapes
E. Rice
D. Wine grapes
Figure 43. Sections in Fresno County where A) corn, B) onion, C) table and raisin grapes, D)
wine grapes, E) rice and F) sorghum were grown in 2000 and glyphosate was or was not applied
at time of increased plant sensitivity (CPUR).
66
-------
A. Corn
B. Sorghum
Figure 44. Sections in Fresno County where A) corn and B) sorghum were grown in 2000 and
sethoxydim was or was not applied at time of increased plant sensitivity (CPUR).
10.5. Summary of Risk Assessment for Fresno County
The NCFAP's 1997 pesticide usage database tracks the use of 220 active ingredients in
U.S. crop production, of which 96 are classified as herbicides. A risk characterization using
information in the published literature identified 28 of the 96 herbicides as high potential risk
herbicides, i.e., they can damage crops at sublethal levels. Spatial and temporal trends in
pesticide use domestically and specifically in California showed shifts in agronomic weed
management systems towards increasing usage of less harmful chemicals such as glyphosate and
static or decreasing usage of the more damaging chemicals such as 2,4-D, 2,4-DB, dicamba, and
MCPA, in order to minimize the potential risk of drift damage to sensitive crops. California, as
well as several other states, imposed state-limited restrictions on the use of the most damaging
of the federally approved high potential risk herbicides by permit holders during times of
maximum plant sensitivity. As a result, glyphosate usage of 284,000 kg a.i. in Fresno County in
2000 exceeded that of the other 27 high potential risk herbicides combined. Based on the CPUR
database, none of the 28 high potential risk herbicides pose an unacceptably high environmental
risk to sensitive crops due to offsite pesticide drift in Fresno County. We estimated 9,617 ha or
2.1% of total crop acreage planted in the sensitive crops at potential risk due to offsite spray drift
from glyphosate applications (Table 6). The majority of the crop acreage at risk was grown in
67
-------
cotton, grapes, tomatoes and alfalfa which represent over 50% of the cropland in Fresno County.
The total crop acreage at potential risk due to near-field spray drift from applications of the other
27 high potential risk herbicides was estimated at about 800 ha or 0.2% of total acreage planted
in the sensitive crops in Fresno County in 2000.
The CPUR database reports all agricultural applications of pesticides at the section level
and is the most critical input for the GIS-based risk assessment. We were unable to resolve the
inconsistencies between the field boundary coverage and the CPUR database in order to increase
the spatial resolution of the pesticide application data to the field level. Consequently, it was not
possible to use wind data to predict the offsite movement of a pesticide application from a target
field to unintended fields because the exact locations of the crop fields were unknown. Due to
limitations in the spatial resolution of the CPUR database, we assumed that every crop field
within a sector was subject to offsite spray drift from pesticide applications in the same sector.
Our results represent a worst-case scenario and provide an upper bound of the total crop acreage
at risk due to pesticide spray drift from herbicide applications at the time of increased plant
sensitivity.
11. Conclusions
A deterministic GIS-based risk assessment model can be used to evaluate the ecological
risk of pesticide spray drift on unintended fields of sensitive crops based primarily on pesticide
use reporting data at high spatial resolution. When PUR data are available at the parcel or
section level, GIS is used to generate spatially-explicit maps showing the locations of the
nontarget crop fields at potential risk from spray drift deposition. As demonstrated for Fresno
County, an upper bound for the potential crop acreage at risk due to unintended pesticide
exposure from near-field spray drift can be estimated with reasonable accuracy and precision.
However, very few states with the exception of California have full pesticide use reporting
databases at sufficient temporal and spatial resolution to infer the effects of pesticide spray drift
on nontarget crops with any degree of accuracy and precision. Inferences based on the CPUR
database may be applicable to neighboring states which are likely to have similar weed
management practices and agricultural production systems. However, in other states where full
pesticide use reporting data are not available, a probabilistic risk assessment model is required to
68
-------
simulate agricultural applications of pesticides across time and space in addition to spray drift
events and ensuing impacts to unintended sensitive crops. The lack of resolution of pesticide
use data is the primary source of uncertainty and variability in a spatially-explicit risk
assessment model for pesticide drift.
Far-field drift (> 1 km) of very small droplets can also potentially damage sensitive crops
growing in fields or orchards several miles from the site of application. Only the most sensitive
crops (Table C2) would be potentially damaged due to far-field drift from aerial applications of
very bioactive herbicides because deposition rates are expected to be much less than 1% of the
application rate. For Fresno County, far-field drift is not a concern because there were very few
aerial applications of bioactive herbicides at the time of maximum plant sensitivity for the most
sensitive crops within the same section (Table 5). Fine droplet spectra are generally not used for
herbicide applications in California due to drift risk. However, in the rice area of California
north of Sacramento, far-field drift of very bioactive herbicides such as propanil and 2,4-D
under low wind conditions is an ongoing problem. Our GIS-based approach can be extended to
consider both near-field and far-field drift. A tilted Gaussian Plume model such as FSCBG can
be used to simulate far-field drift from an aerial application. However, as drift distances
increase, the uncertainty and variability in assessing the potential risks of spray drift on sensitive
crops increase. Consequently, an estimate for the upper bound for the potential crop acreage at
risk due to far-field drift will be less accurate and precise than that for near-field drift.
We identified 28 federally-approved herbicides as having high risk potential to sensitive
crops based on scientific evidence of damage at sublethal levels in the peer-reviewed literature
as well as frequency of drift complaints reported to state and county agencies. A high potential
risk herbicide was reported to have caused a 10% or greater reduction in crop yield at 10% of the
application rate in at least one published study. For each of the high potential risk herbicides,
the sensitive crops were rank ordered as low, medium or high sensitivity based on known or
potential effects reported in the published literature. Therefore, it is possible to identify the
counties where these more sensitive crops are grown and use of the high potential risk
herbicides is high based on pesticide use and cropland use data from the USDA NASS.
For example, glyphosate drift is a major concern in the central and southern states where
glyphosate usage is high and some of the top crops such as corn, rice and soybean are
69
-------
susceptible to damage due to pesticide exposure at low levels. Rice in the reproductive stage is
particularly sensitive to glyphosate. For example, glyphosate drift from soybean fields is
problematic in Arkansas which has 2.9 million acres of RR soybean planted beside 1.4 million
acres of rice. In return, soybeans and vegetables are highly sensitive to propanil and other rice
herbicides which are applied multiple times in the season and often are aerially applied for
postemergence control of grasses and broadleaf weeds. Soybeans at early- to mid-bloom stage
are also highly sensitive to 2,4-D and dicamba which are widely used herbicides in the
agricultural production of corn and wheat in the Corn Belt and wheat-growing states. In states
that do not have restricted materials regulations, one can expect more drift problems than was
found in Fresno County, California involving a greater variety of herbicides and the major crops
such as corn, rice and soybeans.
Unfortunately, the only available pesticide use data for states other than California lack
both temporal and spatial resolution. For states other than California, the NCFAP's 1997
pesticide usage database summarizes the national use of 220 active ingredients on 87 crops
based on state-level usage data from USDA crop surveys and other sources. In addition, the
USDA publishes annual state-level usage data on select crops using data obtained from the
Agricultural Resources Management Study and the Objective Yield Survey to augment the
national Census of Agriculture surveys conducted by the USDA every five years. The 1997
Census of Agriculture is the most comprehensive data source for American agricultural
production at the county level and can be combined with the 1997 NCFAP pesticide use data to
estimate usage by county (Thelin and Gianessi, 2000). Additional assumptions and
methodology are needed to address the lack of information on the method and time of pesticide
applications and the locations of the target and nontarget fields. Future research will utilize the
USDA county-level pesticide databases to estimate the total crop acreage at risk due to pesticide
drift for other states in a GIS-based probabilistic risk assessment model that captures the
variability and uncertainty in the various sources of information. Other data layers such as
cropland use from GAP, the United States Department of Agriculture or other sources could be
used to improve the spatial resolution but more research is needed to quantify the uncertainty
and variability in the risk assessment calculations when pesticide use data are lacking.
70
-------
Acknowledgements
The information in this document has been funded wholly by the US Environmental Protection
Agency. It has been subjected to the Agency's peer review and administrative review, and it has
been approved for publication as an EPA report. Reference herein to any specific commercial
products, process, or service by trade name, trademark, manufacturer, or otherwise, does not
necessarily constitute or imply its endorsement, recommendation, or favoring by the United
States Government. The views and opinions of authors expressed herein do not necessarily state
or reflect those of the United States Government and shall not be used for advertising or product
endorsement purpose.
The authors thank Sandra Bird, USEPA, for providing the AgDPJFT® V.2.0.05 computer model
and its companion drop size distribution model Sidekick® for estimating downwind deposition
rates of pesticides applied by aerial or ground equipment.
References
Al-Khatib, K, R. Parker, E.P. Fuerst. (1992a). Alfalfa (Medicago sativd) response to simulated
herbicide spray drift. Weed Technol., 6, 956-960.
Al-Khatib, K, R. Parker, E.P. Fuerst. (1992b). Sweet cherry (Primus avium) response to
simulated drift from selected herbicides. Weed Technol., 6, 975-979.
Al-Khatib, K.R., R. Parker, E.P. Fuerst. (1993). Wine grape (Vitis vunifera L.) response to
simulated herbicide drift. Weed Technol., 7, 97-102.
Al-Khatib, K.R, D. Peterson. (1999). Soybean (Glycine max) Response to simulated drift from
selected sulfonylurea herbicides, dicamba, glyphosate, and glufosinate. Weed Technol.,
13, 264-270.
Al-Khatib, K., A. Tamhane. (1999). Dry pea (Pisum sativum) response to low rates of selected
foliar- and soil-applied sulfonylurea and growth regulator herbicides. Weed Technol.,
13, 753-758.
Al-Khatib, K.R., R.S. Currie, L.D. Maddux, C.R. Thompson, T.M. Price. (2000). Corn
response to simulated herbicide drift. Proc. N. Cent. Weed Sci. Soc., 55, 5-6.
71
-------
Al-Khatib, K.R, M.M. Classen, P.W. Stahlman, P.W. Geier, D.L. Regehr, S.R. Duncan, W.F.
Heer. (2003). Grain sorghum response to simulated drift from glufosinate, glyphosate,
imazethapyr, and sethoxydim. Weed Technol., 17, 261-265.
Aspelin, A.L. (1997). Pesticides industry sales and usage 1994 and 1995 market estimates. U.S.
Environmental Protection Agency, Office of Prevention, Pesticides and Toxic
Substances, Biological and Economic Analysis Division, Washington D.C. EPA report
no. EPA-733-R-97-002.
Aspelin, A.L., A.H. Grube. (1999). Pesticides industry sales and usage 1996 and 1997 market
estimates. U.S. Environmental Protection Agency, Office of Prevention, Pesticides and
Toxic Substances, Biological and Economic Analysis Division, Washington D.C. EPA
report no. EPA-733-R-99-001.
Auch, D.E., W.E. Arnold. (1978). Dicamba use and injury on soybeans (Glycine max) in South
Dakota. Weed Sci., 26, 471-475.
Bacon, C.G., S.M. Zedaker, R.E. Kreh, M.C. Vodak. (1983). Efficacy of sulfometuron methyl
(sulfometuron) for Christmas tree production. Proc. 37th Ann. Meeting Southern Weed
Sci. Soc., 37, 200-208.
Bailey, J.A., G. Kapusta. (1993). Soybean {Glycine max) tolerance to simulated drift of
nicosulfuron and primisulfuron. Weed Technol., 7, 740-745.
Baker, R.S., Barrentine, W.L. (1986). Response of cotton to simulated drift of two soybean
herbicides. Proc. South. Weed Sci. Soc., 39, 476.
Baldwin, F.L. (1995). Weed control in Roundup tolerant soybeans. Proc. South. Weed Sci.
Soc., 48, 46.
Baldwin, F.L. (2001). No off-target complaints after Command applied by aircraft. Arkansas
Land and Life, 6, 26.
Ball, D. (1994) Simulated thifensulfuron:tribenuron drift on fall canola. Columbia Basin Ag.
Res. Cent. Rep. 60-61.
Ball, D.A., R. Parker, J. Colquhoun, I. Dami. (2004). Preventing herbicide drift and injury to
grapes. Oregon State University, Extension Service, Report No. EM 8860, March 2004.
72
-------
Banks, P.A., J. Schroeder. (2002). Carrier volume affects herbicide activity in simulated spray
drift studies. Weed Techno!., 16, 833-837.
Barker, M.A., L. Thompson Jr., R.P. Patterson. (1984). Effect of 2,4-DB on soybeans (Glycine
max). Weed Sci., 32,299-303.
Barnes, C.J., T.L. Lavy, J.D. Mattice. (1987). Exposure of non-applicator personnel and
adjacent acres to aerially applied propanil. Bull. Environ. Contam. Toxicol., 39, 126-
133.
Benbrook, C.M. (2004). Genetically engineered crops and pesticide use in the United States:
The first nine years. BioTech InfoNet Technical Paper Number 7. October 2004.
Available at http:/www.biotech-info.net/highlights.html#technical_papers.
Bhatti, M.A., K. Al-Khatib, A.S. Felsot, R. Parker, S. Kadir. (1995). Effects of simulated
chlorsulfuron drift on fruit yield and quality of sweet cherries (Prunus avium L.).
Environ. Toxicol. Chem., 14, 537-544.
Bird, S.L. (1995). A compilation of aerial spray drift field study data for low flight agricultural
applications. In M.L. Leng et al. (Ed.) Agrochemical environmental fate studies: State of
the art. Lewis Publ, Chelsea, MI. pp. 195-207.
Bird, S.L., D.M. Esterly, S.G. Perry. (1996). Off-target deposition of pesticides from
agricultural aerial spray applications. J. Environ. Qual., 25, 1095-1104.
Bird, S.L., S.G. Perry, S.L. Ray, M.E. Teske. (2002). Evaluation of the AgDISP aerial spray
algorithms in the AgDRIFT model. Environ. Toxicol. Chem., 21, 672-681.
Bode, L.E., C.B. McWhorter. (1977). Toxicity of MSMA, fiuoremeturon, and propanil to
soybeans. Weed Sci., 25, 101-105.
Bond, J.A., J.L. Griffin, D.A. Peters, J.M. Ellis, S.D. Linscombe, J.L Godley. (2000) Corn and
rice response to simulated drift rates of lighting. Proc South. Weed Sci Soc, 53, 22.
Boutin, C., H.-B. Lee, T.E. Peart, S.P. Batchelor, R.J. Maguire. (1999). Responses of five plant
species sprayed with sub lethal doses of metsulfuron methyl. Proc. 1999 Brighton
Conference - Weeds.
Bradshaw, L.D., S.R. Padgette, S.L. Kimball, B.H. Wells. (1997). Perspective on glyphosate
resistance. Weed Techno!., 11, 189-198.
73
-------
Brammall, R.A., Higgins, V.J. (1988). The effect of glyphosate on resistance of tomato to
Fusarium crown and root rot diseases and on the formation of host structural defensive
barriers. Can. J. Hot., 66, 1547-1555.
Brammer, T.A.,Ball, D.A., Thill, B.C. (1996) Dose response of spring-planted canolato
thifensulfuron-tribenuron. Western Weed Sci. Soc. Res. Progr. Rep. 70-71.
Burnside, O.C., R.L. Lavy. (1966). Dissipation of dicamba. Weeds, 14, 211-214.
Byford, W,F., J. Prince. (1976). Experiments on the effect of hormone weed-killers on sugar
beet. J. Agric. Soc., 86, 135-139.
California Agricultural Statistics Service. (2001). California grape acreage 2000. Sacramento,
CA. June 2001. Available at http://www.nass.usda.gov/ca/bul/acreage/indexgab.htm.
California Department of Pesticide Regulation. (2000). An overview of California's unique full
reporting system. California Department of Pesticide Regulation, Sacramento, CA. May
2000. Available online at http://www.cdpr.ca.gov/docs/pur/purovrvw/ovr52000.pdf.
Cameron, J., N. Turvey. (1977). Weed control and nutrition of eucalypt plantations. Proc. (Part
1) Australian Forest Development Institute Forestry Conference, Traralgon, Victoria.
Carey, V.F., R.E. Hoagland, R.E. Talbert. (1995). Verification and distribution of propanil-
resistant barnyardgrass (Echinochloa cruss-galli), in Arkansas. Weed Technol., 9, 366-
372.
Carter, P.C. (1996). Modification of clomazone application methods and formulation for
reducing volatilization from cotton soils. M.S. Thesis, University of Arkansas,
Fayetteville.
Chester, G., R.J. Ward. (1984). Occupational exposure and drift hazard during aerial
application of paraquat to cotton. Arch. Environ. Contam. Toxicol., 13, 551-563.
Clay, P.A., J.L. Griffin, D.L. Jordan. (1995). Sicklepod (Senna obtusifolia) control programs in
Roundup Ready soybeans. Proc. South. Weed Sci. Soc., 48, 49-50.
Constantini, A. (1986). Tolerance ofPinus caribaea var. hondurensis transplants to eleven
herbicides. Aust. For., 49, 241-245.
Cotton Farming. (2000). Weed control systems and strategies. June issue available at
http ://www. cottonfarming. com.
74
-------
Coyne, D.P., O.C. Burnside. (1968). Differential plant injury and yield responses of tomato
varieties to 2,4-D. Neb. Agr. Expt. Stn. Res. Bull, 226, 21 pp.
Cremlyn, R.J. (1991). Agrochemicals: Preparation and mode of action. Chichester, UK: John
Wiley & Sons, pp.219,226,227.
Culpepper, AS., A.C. York (1998). Weed management in glyphosate-tolerant cotton. J.
Cotton Sci., 4, 174-185.
Derksen, D.A. (1989). Dicamba, chlorsulfuron, and clopyralid as sprayer contaminants on
sunflower (Helianthus annuus), mustard (Brassicajuncea), and lentil (Lens culinaris),
respectively. Weed Sci., 37, 616-621.
Dexter, A.G. (1993). Herbicide spray drift. North Dakota State University Extension Service
report A-657, revised August 1993. Available at
http://www.ext.nodak.edu/extpubs/plantsci/weeds/a657w.htm.
DiTomaso, J.M., J.R. Miller, G.B. Kyser, A.W. Hazebrook, J. Trumbo, D. Valcore, V.F.
Carrithers. (2004). Aerial application of clopyralid demonstrates little drift potential and
low toxicity to toads. Calif. Agric., 58, 154-158.
Donaldson, D., T. Kiely, A. Grube. (2002). Pesticides industry sales and usage 1998 and 1999
market estimates. U.S. Environmental Protection Agency, Office of Prevention,
Pesticides and Toxic Substances, Biological and Economic Analysis Division,
Washington D.C. EPA report no. EPA-733-R-02-001. Available at
http://www.epa.gov/oppbeadl/pestsales/99pestsales/market_estimates 1999.pdf.
Draxler, R.R., G.D. Hess. (1997). Description of the HYSPLIT_4 modeling system. National
Oceanic and Atmospheric Administration, Air Resources Laboratory, Silver Springs,
MD. November 1997. NOAA technical memorandum ERL ARL-224. Available at
www.arl.noaa.gov.
E.I. du Pont de Nemours and Company. (1984). Technical Manual: Solve your vegetation
management problems with DuPont Products. 58 p.
E.I. du Pont de Nemours and Company. (2000-2002). DuPont™ Oust® XP herbicide (specimen
label). Available at www.cdms.net.
75
-------
Eberlein, C.V., M.J. Guttieri (1994). Potato (Solarum tuberosum) response to simulated drift of
imidazolinone herbicides. Weed Sci., 42, 70-75.
Ellis, J.M., J.L. Griffin. (2002). Soybean (Glycine max) and cotton (Gossypium hirsutum)
response to simulated drift of glyphosate and glufosinate. Weed Technol., 16, 580-586.
Ellis, J.M., J.J. Griffin, C.A. Jones. (2002). Effect of carrier volume on corn (Zea mays) and
soybean (Glycine max) response to simulated drift of glyphosate and glufosinate. Weed
Technol., 16, 587-592.
Ellis, J.M., J.L. Griffin, S.D. Linscombe, E.P. Webster. (2000). Crop response to simulated
drift of Roundup Ultra and Liberty herbicides. La. Agric., 43, 18-19.
Ellis, J.M., J.L. Griffin, S.D. Linscombe, E.P. Webster. (2003). Rice (Oryza sativa) and corn
(Zea mays) response to simulated drift of glyphosate and glufosinate. Weed Technol.,
17, 452-460.
Ernst, W.R., P. Jonah, K. Doe, G. Julien, P. Hennigar. (1991). Toxicity to aquatic organisms of
off-target deposition of endosulfan applied by aircraft. Environ. Toxicol. Chem., 10,
103-114.
Fletcher, J.S., T.G. Pfleeger, H.C. Ratsch, R. Hayes. (1996). Potential impact of low levels of
chlorsulfuron and other herbicides on growth and yield of nontarget plants. Environ.
Tox. Chem., 15, 1189-1196.
Franz, J.E., M.K. Mayo, J.A. Sikorski. (1997). Toxicology and environmental properties of
glyphosate. P. 103-137 in Glyphosate: A Unique Global Herbicide. Washington, D.C.,
American Chemical Society Monograph 189.
Garcia, A.M., F.G. Benavides, T. Fletcher, E. Orts. (1998). Paternal exposure to pesticides and
congenital malformations. Scandinavian J. Work and Environ. Health, 24, 473-480.
Gealy, D.R., C.M. Boerboom, A.G. Ogg. (1995). Growth and yield of pea (Pisum sativum L.)
and lentil (Lens culinaris L.) sprayed with low rates of sulfonylurea and phenoxy
herbicides. Weed Sci., 43, 640-647.
Ghosheh, H.Z., J.M. Chandler, R.H. Bierman. (1994). Impact of DPX-PE350 drift on corn and
grain sorghum. Proc. South. Weed Sci. Soc., 47, 24.
76
-------
Gianessi, L.P., M.B. Marcelli. (2000). Pesticide use in U.S. crop production: 1997. National
summary report. National Center for Food and Agricultural Policy, Washington, B.C.
November 2000. Available at www.ncfap.org.
Gianessi, L.P., C.S. Silvers. (2000). Trends in crop pesticide use: Comparing 1992 and 1997.
National Center for Food and Agricultural Policy, Washington, B.C. November 2000.
Available at www.ncfap.org.
Giesy, J.P., Bobson, S., Solomon, K.R. (2000). Ecotoxicological risk assessment for Roundup
herbicide. Rev. Environ. Contamin. Toxicol., 167, 35-120.
Gilreath, J.P., C.A. Chase, S.J. Locascio. (2000). Phytotoxic effects of glyphosate on pepper
(Capsicum annuum). Weed Technol.,14, 488-494.
Gilreath, J.P., C.A. Chase, S.L. Locascio. (2001a). Crop injury from sublethal rates of
herbicide I. Tomato. Hortscience, 38, 669-673.
Gilreath, J.P., C.A. Chase, S.L. Locascio. (200Ib). Crop injury from sublethal rates of
herbicide III. Pepper. Hortscience, 38, 677-681.
Gjerstad, B.H., L.R. Nelson. (1983). New herbicides for herbaceous weed control in pine
plantations. Proc. 37th Annual Meeting Southern Weed Sci. Soc., 36, 245.
Grossmann, K. (1998). Quinclorac belongs to a new class of highly selective auxin herbicides.
Weed Sci., 46, 707-716.
Haderlie, L.C., P.J. Petersen, P.W. Leino. (1986). Potato seed vigor and yield potential
following herbicide drift or carryover. Res. Prog. Rep. West. Soc. Weed Sci., 324-327.
Hamilton, K.C., H.F. Arle. (1979). Response of cotton to dicamba. Weed Sci., 27, 604-607.
Hartzler, B. (2001). Effect of glyphosate drift on corn yields. ISU Weed Science Online Report
available at http://www.weeds.iastate.edu/mgmt/2001/glydrift.htm.
Hemphill, BB. Jr,, M.L. Montgomery. (1981). Response of vegetable crops to sublethal
application of 2,4-B. Weed Sci., 29, 632-635.
Huitnik, G., J.T. Walker, T.L. Lavy. (1990). Bownwind deposition of 2,4-
dichlorophenoxyacetic acid herbicide (2,4-B) in invert emulsion. Trans. ASAE 3, 1051-
1056.
77
-------
Hurst, H.R. (1977). Cotton tolerance to over-the-top and directed applications of bentazon.
Proc. Beltwide Cotton Prod. Res. Conf., Atlanta, GA, 181-182.
Hurst, H.R. (1982). Cotton (Gossypium hirsutum) response to simulated drift from selected
herbicides. Weed Sci., 30, 311-315.
Idaho Dept. of Agriculture. (2002). Press release and publications: Idaho State Department of
Agriculture completes Oust investigation. Jan. 18.
Jacoby, P.W., C.H. Meadors, L.E. Clark. (1990). Effects of triclopyr, clopyralid, and picloram
on growth and production of cotton. J. Prod. Agric., 3, 297-301.
Jeffery, L.S., T. McCutchen, P.E. Hoskinson. (1972). Effects of DSMA and MSMA on cotton.
Tennessee Farm and Home Sci. Prog. Rep., 84, 140-141.
Johnson, B.J. (1997). Sequential application of preemergence and postemergence herbicides for
large crabgrass control in tall fescue turf. Weed Technol., 11, 693-697.
Jordan, T.N., R.R. Romanowski. (1974). Comparison of dicamba and 2,4-D injuryto field
grown tomatoes. Hortscience, 9, 74-75.
Knowe, S.A. (1984). Herbaceous weed control in slash pine and cottonwood plantations. Proc.
37th Annual Meeting Southern Weed Sci. Soc., 37, 214-215.
Kurtz, M.E., J.E. Street. (2003). Response of rice (Oryza sativa) to glyphosate applied to
simulate drift. Weed Technol., 17, 234-238.
Lanclos, D.Y., E.P. Webster, W. Zhang, S.D. Linscombe. (2003). Response of grufosinate-
resistant rice (Oryza sativa) to glufosinate application timings. Weed Technol., 17, 157-
160.
Lanini, W.T. (1999). Simulated drift of transline (clopyralid) on tomaotes, cotton, saffiower and
sunflower. Proc. California Weed Sci. Soc., 51, 201-204.
Lanini, W.T. (2003). Affects of spray drift on trees and vines. Proc. California Weed Sci. Soc.,
55,63.
Lawson, H.M., J.S. Wiseman. (1991). The effects of contamination of a potato crop with a
thifensulfuron-methyl/metsulfuron methyl herbicide. Proc. 1991 Br. Crop Prot. Conf,
239-246.
78
-------
Lawson, H.M., J.S. Wiseman, G.M. Wright. (1992). Seasonal and rational implications of
accidental contamination of potato crops with herbicides. Asp. Appl. Biol., 33, 173-180.
Leino, P.W., L.C. Haderlie. (1985). Simulated herbicide drift injury in potatoes. Proc. West.
Soc. Weed Sci., 38, 93-99.
Lobb, P.O. (1989). Response of potatoes, onions and tomatoes to herbicides used for brushweed
control. Proc. 42nd New Zealand Weed and Pest Control Conf, 251-253.
Lobb, P.O., G.W. Woon. (1983). Response of potatoes, grapes and kiwifruit to low rates of six
commercially available pasture and brushweed herbicides. Proc. 36th New Zealand
Weed and Pest Control Conf, 148-150.
Locke, D. Landivar, J.A., Moseley, D. (1995). The effects of rate and timing of glyphosate
applications of defoliation efficiency, regrowth inhibition, lint yield, fiber quality and
seed quality. Proc. Beltwide Cotton Conf, National Cotton Council of America, 1088-
1090.
Lyon, D.J., R.G. Wilson. (1986) Sensitivity of fieldbeans (Phaseolus vulgaris) to reduced rates
of 2,4-D and dicamba. Weed Sci., 34, 953-956.
Ma, L.Q., W. Harris, J. Sartain. (2000). Environmental impacts of lead pellets at shooting
ranges & arsenical herbicides on golf courses in Florida. University of Florida, Florida
Center for Solid and Hazardous Waste Management, Gainesville, FL Report #00-03.
Available at www.floridacenter.org.
Malik, M.S., R.E. Talbert, E.F. Scherder, M.L. Lovelace, B.V. Ottis. (2002). Weed control
programs involving clomazone applied postemergence and response of pecans to drift
rates of clomazone. In: R.J. Norman and J.-F. Meullenet (eds). B.R. Wells Rice
Research Studies.
Mallory-Smith, C.A., B.C. Thill. (1993). Simulated thifensulfuron-tribenuron drift injury to
spring peas. Proc. West. Soc. Weed Sci., 46, 11.
Marrs, R.H., A. Frost, R. Plant, P. Lunnis. (1992). Aerial applications of asulam: A bioassay
technique for assessing buffer zones to protect sensitive sites in upland Britain. Biol.
Conser., 59, 19-23.
79
-------
Matsumura, N., C. Takeuchi, K. Hishikawa, T. Fujii, T.Nakaki. (2001). Glufosinate
ammonium induces convulsion throughNmethylDaspartate receptors in mice. Nerosci.
Lett, 304, 123-125.
McWhorter, C.G., C.T. Bryson. (1992). Herbicide Use Trends in Cotton. Weeds of Cotton:
Characterization and Control, Cotton Foundation.
Miller, O.K., R.G. Downer, B.R. Leonard, E.M. Holman, S.T. Kelly. (2004). Response of
nonglyphosate-resistant cotton to reduced rates of glyphosate. Weed Sci., 52, 178-182.
Miller, J.H., H.M. Kempen, J.A. Wilkerson, C.L. Foy. (1963). Response of cotton to 2,4-D and
related phenoxy herbicides. USDA Tech. Bull. No. 1289. 28 pp.
Obrigawitch, T.T., G. Cook, J. Wetherington. (1998). Assessment of effects on non-target
plants from sulfonylurea herbicides using field approaches. Pestic. Sci.,. 52, 199-217.
Ogg, A.G. Jr., M.A. Ahmedullah, G.M. Wright. (1991). Influence of repeated applications of
2,4-D on yield and juice quality of concord grape. Weed Sci., 39, 284-295.
Payne, N.J. (1993). Spray dispersal from aerial silvicultural glyphosate applications. Crop
Protec. 12, 463-469.
Payne, N.J., J.C. Feng, P.E. Reynolds. (1990). Off-target deposits and buffer zones required
around water for aerial glyphosate applications. Pestic. Sci., 30, 183-198.
Ray, J.W., A.L. Vanner, B. Richardson, G. Coker. (1996). Determination of the No Observable
Effect Level (NOEL) of four commonly used forestry herbicides on tomatoes. Proc. 49th
New Zealand Plant Protection Conf, 188-191.
Riley, C.M., C.J. Wiesner. (1989). Off-target deposition and drift of aerially applied
agricultural sprays. Pestic. Sci., 26, 159-166.
Robbins, W.A., W.S. Taylor. (1957). Injury to canning tomatoes caused by 2,4-D. Proc. Amer.
Soc. Hort. Sci., 70, 373-378.
Romanowski, R.R. (1980). Simulated drift studies with herbicides on field-grown tomatoes.
Hortscience, 15, 793-794.
Rowland, C.D. Jr, D.B. Reynolds, R.H. Blackley Jr. (1999). Corn and cotton response to drift
rates of non-desired herbicide applications. Proc. South. Weed Sci. Soc., 52, 30.
80
-------
Scherder, E.F., R.E. Talbert, M.L. Lovelace, N.W. Buerhing, F.L. Baldwin. (2001). Control of
propanil resistant and susceptible barnyardgrass. In: R.J. Norman and J.-F. Meullenet
(eds). B.R. Wells Rice Research Studies 2000. University of Arkansas Agricultural
Experiment Station Research Series, 485, 86-91. Fayetteville.
Schweizer, E.E. (1978). Response of sugarbeets (Beta vulgaris) to sublethal rates of 2,4-D.
WeedSci., 26, 629-631.
Slife, F.W. (1956). The effect of 2,4-D and several other herbicides on weeds and soybeans
when applied as post-emergence sprays. Weeds, 4, 61-68.
Smith, D.T., A.E. Wiese. (1972). Cotton response to low rates of 2,4-D and other herbicides.
Texas Agricultural Experiment Station Bulletin B-l 120, College Station, TX, Texas
A&m University.
Smith, L.L., J. Geronimo. (1984). Response of seven crops to foliar applications of six auxin-
like herbicides. Down Earth, 40, 25-32.
Smith, R.J. Jr., W.C. Shaw. (1966). Weeds and their control in rice production. U.S.Dept.
Agr. Handbook 292. U.S. Gov. Printing Office, Washington, D.C. 64 p.
Snipes, C.E., J.E. Street, T.C. Mueller. (1991). Cotton (Gossypium hirsutum) response to
simulated triclopyr drift. Weed Techno!., 5, 493-498.
Snipes, C.E., J.E. Street, T.C. Mueller. (1992). Cotton (Gossypium hirsutum) injury from
simulated quinclorac drift. Weed Sci., 40, 106-109.
Swanton, C., J.K.Chandler , M.J. Elmes, S.D. Murphy, G.W. Anderson (1996) Postemergence
control of annual grasses and corn (Zea mays) tolerance using DPX-79406. Weed
Technol. 10, 288-294.
Teske, M.E., S.L. Bird, D.M. Esterly, T.B. Curbishley, S.L. Ray, S.G. Perry. (2002).
AgDRIFT®: A model for estimating near-field spray drift from aerial applications.
Environ. Toxicol. Chem., 21, 659-671.
Teske, M.E., S.L. Bird, D.M. Esterly, S.L. Ray, S.G. Perry. (1997). A user's guide for
AgDRIFT 1.0: a tiered approach for the assessment of spray drift of pesticides. Technical
Note Number 95-10. Continuum Dynamics, Inc., Princeton, NJ.
81
-------
Teske, M.E., J.F. Bowers, J.E. Rafferty, J.W. Barry. (1993). FSCBG: An aerial spray dispersion
model for predicting the fate of released material behind aircraft. Environ. Toxicol.
Chem., 12, 453-464.
Teske, M.E., T.B. Curbishley. (1990). Forest Service Aerial Spray Computer Model FSCBG
4.0 User Manual. Technical Report Number 90-06. Continuum Dynamics, Inc.,
Princeton, NJ.
Thelin, G.P., L.P. Gianessi. (2000). Method for estimating pesticide use for county areas of the
conterminous United States. U.S. Geological Survey Open File Report 00-250, USGS,
Sacramento.
Turner, S.A. (1987). Post-application movement of sulfometuron methyl from treated rights of
way areas via wind (soil) erosion. Proc. Fourth Symposium on Environmental Concerns
in Rights-of_way Management. Oct. 25-28, 1987, Indianapolis, IN.
U.S. Department of Agriculture, Agricultural chemical usage: Fruits summary
1991/1993/1995/1997. National Agricultural Statistics Service. Available at
http://usda.mannlib.cornell.edu/reports/nassr/.
U.S. Department of Agriculture, Agricultural chemical usage: 1993 field crops summary.
(1994). National Agricultural Statistics Service. Available at
http://usda.mannlib.cornell.edu/reports/nassr/.
U.S. Department of Agriculture, Agricultural chemical usage: 1994 field crops summary.
(1995). National Agricultural Statistics Service. Available at
http://usda.mannlib.cornell.edu/reports/nassr/.
U.S. Department of Agriculture, Agricultural chemical usage: 1995 field crops summary.
(1996). National Agricultural Statistics Service. Available at
http://usda.mannlib.cornell.edu/reports/nassr/.
U.S. Department of Agriculture, Agricultural chemical usage: 1996 field crops summary.
(1997). National Agricultural Statistics Service. Report Ag Ch 1(97). September.
Available at http://usda.mannlib.cornell.edu/reports/nassr/.
82
-------
U.S. Department of Agriculture, Agricultural chemical usage: 1997 field crops summary.
(1998). National Agricultural Statistics Service. Report Ag Ch 1(98). May. Available
at http://usda.mannlib.cornell.edu/reports/nassr/.
U.S. Department of Agriculture, Agricultural chemical usage: 1998 field crops summary.
(1999). National Agricultural Statistics Service. Report Ag Ch 1(99). May. Available
at http://usda.mannlib.cornell.edu/reports/nassr/.
U.S. Department of Agriculture, Agricultural chemical usage: 1999 field crops summary.
(2000). National Agricultural Statistics Service. Report Ag Ch 1(00). May. Available
at http://usda.mannlib.cornell.edu/reports/nassr/.
U.S. Department of Agriculture, Agricultural chemical usage: 2000 field crops summary.
(2001). National Agricultural Statistics Service. Report Ag Ch 1(01 )b. May. Available
at http://usda.mannlib.cornell.edu/reports/nassr/.
U.S. Department of Agriculture, Agricultural chemical usage: 2001 field crops summary.
(2002). National Agricultural Statistics Service. Report Ag Ch l(02)a May. Available
at http://usda.mannlib.cornell.edu/reports/nassr/.
U.S. Department of Agriculture, Agricultural chemical usage: 2002 field crops summary.
(2003). National Agricultural Statistics Service. Report Ag Ch 1(03). May. Available
at http://usda.mannlib.cornell.edu/reports/nassr/.
U.S. Department of Agriculture, Agricultural chemical usage: 2003 field crops summary.
(2004). National Agricultural Statistics Service. Report Ag Ch l(04)a May. Available
at http://usda.mannlib.cornell.edu/reports/nassr/.
U.S. Environmental Protection Agency Office of Pesticide Programs. (1983). Pesticide fact
sheet: Dicamba. Washington, D.C., Oct 17.
U.S. Environmental Protection Agency Office of Pesticides and Toxic Substances (1986).
Guidance for the reregistration of pesticide products containing glyphosate as the active
ingredient. Washington DC, June.
U.S. Environmental Protection Agency, Office of Pesticide Programs. (1988a). Pesticide Fact
Sheet: 2,4-Dichlorophenoxyacetic Acid. Washington, D.C. EPA Publication No.
540/FS-88-114.
83
-------
U.S. Environmental Protection Agency, Office of Pesticides and Toxic Substances. (1988b).
Guidance for the Reregistration of Pesticide Products Containing 2,4-
Dichlorophenoxyacetic Acid (2,4-D) as the Active Ingredient. Washington, B.C. EPA
Publication No.540/RS-88-l 15.
U.S. Environmental Protection Agency. (1990a). Mollusc 48-hour embryo-larvae study. HOE
039866 technical. Data evaluation report. Reviewed by K.D. Rhodes, KBN Eng. Appl.
Sci. Inc., June 1.
U.S. Environmental Protection Agency. (1990b). Mollusc 48-hour embryo-larvae study. HOE
039866 200 gl soluble concentrate. Data evaluation record. Reviewed by K.D. Rhodes,
KBN Eng. Appl. Sci. Inc., June 1.
U.S. Environmental Protection Agency. (1990c). Estuarine invertebrate toxicity test. HOE
039866 technical. Data evaluation record. Reviewed by K.D. Rhodes, KBN Eng. Appl.
Sci. Inc., June 1.
U.S. Environmental Protection Agency. (1992). Framework for Ecological Risk Assessment.
EPA/630/R-92/001. Risk Assessment Forum, Washington, DC.
U.S. Environmental Protection Agency, Office of Pesticides and Toxic Substances. (1993).
Pesticides fact sheet: Glufosinate ammonium. Washington, DC, June.
U.S. Environmental Protection Agency. (1996). Proposed Guidelines for Ecological Risk
Assessment. Federal Register 61:47552-47631. Available at: http://www.epa.gov/
ORD/WebPubs/ecorisk. U.S. Environmental Protection Agency, Health Effects Division
(1997). Carcinogenicity Peer Review Committee report on 2,4-D, Jan 29.
U.S. Environmental Protection Agency, Office of Pesticides Programs. (1998). R.E.D. Facts
Bromoxynil. Washington, D.C. Sept. 1998. EPA-738-F-98-011.
U.S. Environmental Protection Agency, Office of Prevention, Pesticides and Toxic Substances.
(2002). Overview of atrazine risk assessment. May 2, 2002. Available at
http://www.epa.gov/oppsrrd 1 /reregistration/atrazine/srrd overview may02.pdf.
U.S. Environmental Protection Agency, Office of Prevention, Pesticides and Toxic Substances.
(2004). Pesticide Industry Sales and Usage. Washington, D.C. May 2004. EPA-733-R-
04-001. Available at http://www.epa.gov/pesticides.
84
-------
Wall, D.A . (1994a). Potato (Solarum tuberosum) response to simulated drift of dicamba,
clopyralid, and tribenuron. Weed Sci., 42, 110-114.
Wall, D.A. (1994b). Tolerance of five annual broadleaf crops to simulated
thifensulfuron:tribenuron (2:1) spray drift. Weed Technol., 8, 785-793.
Wall, D.A. (1996a). Effect of sub lethal dosages of 2,4-D on annual broadleaf crops. Can J.
Plant Sci., 76, 179-185.
Wall, D.A. (1996b). Sunflower (Helianthus annuus L.) tolerance to sublethal doses of
imazathapyr . Can. J. Plant Sci., 76, 937-940.
Wall, D.A. (1997). Effect of crop growth stage on tolerance to low doses of
thifensulfuron:tribenuron. Weed Sci., 45, 538-545.
Wall, D.A., D.A. Derksen, L.F. Friesen. (1995). Canola (Brassica napus) response to simulated
sprayer contamination with thifensulfuron and thifensulfuron:tribenuron (2:1). Weed
Technol., 9, 468-476.
Wax, L.M., L.A. Knuth, F.W. Slife (1969). Response of soybeans to 2,4-D, dicamba and
picloram. Weed Sci 17, 388-393.
Weaver, R.J., O.A. Leonard, S.B. McCune. (1961). Response of Tokay grapes to spray
applications of 2,4-D. Hilgardia, 31, 419-433.
Webster, T.M. (2000). Weed survey - Louisiana. Proc. South. Weed Sci. Soc., 53, 261.
Weidenhamer, J.D., G.B. Triplett Jr., F.E. Sobotka. (1989). Dicamba injury to soybean. Agron.
J., 81,637-643.
Westra, P., G. Franc, B. Cramner, T. D'Amato (1991). Research report on 1988 potato-
herbicide injury report. Pp. 98-104 in J. Fletcher and H. Ratsch, eds. Plant Tier Testing:
A Workshop to Evaluate Nontarget Plant Testing in Subdivision J Pesticide Guidelines.
U.S. Environmental Prot. Agency, Corvallis.
Wiese, A.F., E.B. Hudspeth Jr. (1968). Effects of DSMA and MSMA on cotton yield and
arsenic content of cottonseed. Texas Agric. Exp. Stn. MP-877. 4 pp.
Wilson, A.G.L., L.A. Harper, H. Baker. (1986). Evaluation of insecticide residues and droplet
drift following aerial application to cotton in New South Wales. Aust. J. Exp. Agric., 26,
237-243.
85
-------
Wilson, H.P., I.E. Mines, R.R. Bellinder, J.A. Grande. (1985). Comparison of HOE-39866,
SC-0224 paraquat, and glyphosate in no-till corn (Zea mays). Weed Sci., 33, 531-536.
Wilson, H.P., A.D. Worsham. (1988). Combinations of nonselective herbicides for difficult to
control weeds in no-till corn, Zea mays, and soybeans, Glycine max. Weed Sci., 36, 648-
652.
Woods, N., I.P. Craig, G. Dorr, B. Young. (2001). Spray drift of pesticides arising from aerial
application in cotton. J. Environ. Qual., 30, 697-701.
Yarborough, D.E., Bhowmik, P.C. (1989). Evaluation of sulfonylurea and imidazoline
compounds for bunchberry control in lowbush blueberry fields. Proc. Northeastern
Weed Sci. Soc., 43, 142-148.
Yates, W.E., N.B. Akesson, D.E. Bayer. (1978). Drift of glyphosate sprays applied with aerial
and ground equipment. Weed Sci., 26, 597-604.
Young, B.G., J.M. Young, J.L. Matthews. (2003). Soybean (Glycine max) response to foliar
applications of mesotrione. Weed Technol., 17, 651-654.
86
-------
Appendix A
Abbreviations and Acronyms
2,4-D
2,4-DB
a.i.
AAPCO
ALS
APFO
BXN®
CASS
CDPR
CDWR
CES
CLU
CPUR
CYDA
DSMA
EC10
ESDC
FSCBG
FYIA
g
GIS
GM
GPS
ha
HYSPLIT model
ID
2,4 dichlorophenoxy acetic acid
2,4 dichlorophenoxy butonoic acid
active ingredient
American Association of Pesticide Control Officials
acetolactate synthase
Aerial Photography Field Office
bromoxynil tolerant
California Agricultural Statistics Service
California Department of Pesticide Regulation
California Department of Water Resources
Cooperative Extension Service
Common Land Unit
California Pesticide Use Reporting
County of Yolo Department of Agriculture
disodium methylarsonate
effective concentration associated with 10% yield loss
expected spray drift concentration
Forest Service Cramer-Barry-Grim
Fresno Yosemite International Airport
gram
Geographic Information System
genetically modified
Global Positioning System
hectare
HYbrid Single-Particle Lagrangian Integrated Trajectory Model
Identification
87
-------
ISIS Center
kg
m
m/s
MCPA
mPa
MSMA
NASS
NCDC
NCEP
NCFAP
NNW
NOAA
NAPIAP
NWS
OPP
OPPTS
PLSS
PNSP
ppm
PUR
RR
RUP
SDTF
SETAC
SJV
su
USDA
USEPA
Interdisciplinary Spatial Information Systems Center
kilogram
meter
meters per second
(4-chloro-2-methylphenoxy)acetic acid
milliPascals
Monosodium methylarsonate
National Agricultural Statistics Service
National Climate Data Center
National Centers for Environmental Prediction
National Center for Food and Agricultural Policy
north northwest
National Oceanic and Atmospheric Administration
National Agricultural Pesticide Impact Assessment Program
National Weather Service
Office of Pesticides Programs
Office of Prevention, Pesticides, and Toxic Substances
Public Land Survey System
Pesticide National Synthesis Project
parts per million
Pesticide Use Reporting
Roundup Ready™
Restricted Use Pesticide
Spray Drift Task Force
Society of Environmental Toxicology and Chemistry
San Joaquin Valley
sulfonylurea
United States Department of Agriculture
United States Environmental Protection Agency
88
-------
USFWS United States Fish & Wildlife Service
USGS United States Geological Service
W west
WNW west northwest
89
-------
Appendix B
Colloquial and Latin Names
Alfalfa
Almond
Amaranth, palmer
Apple
Apricot
Asparagus
Barley
Barnyardgrass
Bean
Beet, sugar
Bentgrass
Blueberry
Buckwheat
Bulrush
Cabbage
Canola (rape)
Canarygrass, reed
Cantaloupe
Carrot
Cauliflower
Cherry
Cocklebur
Composite
Corn
Cotton
Crabgrass
Medic ago L.
Prunus amygdalus Batsch cv. Nonpariel
Amaranthus spp.
Mains spp.
Prunus armeniaca L.
Asparagus officinalis
Hordeum spp.
Echinochloa spp.
Phaseolus vulgaris L.
Beta vulgaris L.
Agrostis spp.
Vaccinium corymbosum
Eriogonum atrorubens
Scirpus spp.
Brassica oleracea var. capitata
Bras sic a spp.
Phalaris arundinacea
Cucurbita melo var. reticulatus
Daucus carota
Brassica oleracea var. Botrytis
Prunus spp.
Xanthium spp.
Composita spp.
Zea mays L.
Gossypium hirsutum L.
Digit aria spp.
90
-------
Cucumber
Flax
Foxtail
Garlic
Grape
Grass, bahia
Grass, bermuda
Grass, dallis
Grass, rye
Hay
Johnsongrass
Kiwifruit
Lambquarters
Legume
Lentil
Lettuce
Mint
Morningglory
Mustard
Nectarine
Nutsedge
Oats
Onion
Orange
Panicum
Pea
Peanut
Pecan
Pepper
Cucumis sativus L.
Linum spp.
Alopecurus spp.
Allium sativum
Vitis spp.
Paspalum notatum
Cynodon dactylon
Paspalum dilatatum
Lolium perenne L.
Car ex spp.
Sorghum halepense
Actinidia chinensis
Chenopodium
Leguminosae
Lens culinaris
Lactuca saliva L.
Mentha spp.
Ipomoea spp.
Brassica spp.
Prunus persica
Carex castanea
Avena sativa L.
Allium cepa L.
Citrus sinensis
Panicum spp.
Pisum sativum L.
Arachis L.
Gary a spp.
Capsicum annuum L.
91
-------
Pigweed
Pistachio
Potato
Potato, sweet
Prune
Pumpkin
Quackgrass
Radish
Ragweed
Rice
Rutabaga
Sedge
Shattercane
Sicklepod
Smartweed
Sorghum
Soybean
Sprangletop
Squash
Sugarcane
Sunflower
Thistle, Canada
Thistle, yellow star
Tobacco
Tomato
Wheat
Amaranthus spp.
Pistacia spp.
Solanum tuberosum L.
Ipomea batatas
Prunus domestica L.
Cucurbita maxima
Elytrigia repens
Raphanus sativus L.
Ambrosia spp.
Oryza saliva L.
Brassica napus var. Napobrassica
Car ex spp.
Sorghum bicolor (L.) Moench
Arabis canadensis
Polygonum spp.
Sorghum spp.
Glycine max (L.) Merr.
Leptochloa
Cucurbita spp.
Saccharum officinarum
Helianthus annuus L.
Cirsium arvense
Centaurea solstitialis
Nicotiana tabacum L.
Lycopersicon esculentum Mill.
Triticum aestivum L.
92
-------
Appendix C
List of the High Potential Risk Herbicides and Associated Sensitive Crop Species Reported in the Published Literature
Table Cl: High potential risk herbicides that have been shown to cause yield reductions of 10% or greater in susceptible plants when
exposed to 10% of typical application rate. The California chemical code for the most widely used formulation of a herbicide is
indicated by the bold font.
Herbicide
Mode of Action and Use1
Application rate (g ai ha"1)
Drift Problem
2,4-DB
Chemcode = 5020,
1385,838
(dimethylamine salt),
837
Growth regulator.
Seldom used in CA.
Growth regulator or
auxenic herbicides which
include 2,4-D, 2,4-DB
and MCPA are restricted
from use in S JV in CA
from March 16 to Oct 15.
Usage rates ( % CA use)
(Chemcode = 838)
Alfalfa =940 (99%)
Clover =1,310(1%)
Avg. app rate = 1,065
Effects expected to be similar to those for 2,4-D
93
-------
2,4-D
Chemcode = 636,
1622,801,980,981,
1255, 802, 803, 804,
805, 875, 806
(dimethylamine salt),
807, 1962, 809, 810,
3953,811,1032,1259,
1096, 1999, 1275,
812,813,814,815,
1138,5538,816
Growth regulator.
Selective herbicide. Dicot-
specific (amine
formulation).
% national use (USDA
NASS, 1997)
Pasture = 42%
Winter Wheat = 20%
Corn = 9%
Soybean = 8%
Rice=l%
Growth regulator or
auxenic herbicides which
include 2,4-D, 2,4-DB
and MCPA are restricted
Usage rates ( % CA use)
(Chemcode = 806)
Pasture = 960 (3%)
Wheat = 970 (21%)
Corn =710(3%)
Rice = 460 (2%)
Almond = 720 (15%)
Barley = 910 (6%)
Rights-of-way = 1,220
(8%)
Landscape maintenance =
1,550(6%)
Avg. app rate = 800
IPM recommended rates
Barley = 280-1120 gae
ha-1
Corn = 265-532 gai ha-1
Potential damage to alfalfa (Al-Khatib et al, 1992a),
bushbean (Hemphill and Montgomery, 1981), field bean
(Smith and Geronimo, 1984, Lyon and Wilson, 1986),
buckwheat (Wall, 1994b), cabbage (Hemphill and
Montgomery, 1981), canola (Wall, 1996a), carrot (Hemphill
and Montgomery, 1981), cauliflower (Hemphill and
Montgomery, 1981), cotton (Miller et al., 1963, Smith and
Wiese, 1972, Smith and Geronimo, 1984, Banks and
Schroeder, 2002), cucumber (Hemphill and Montgomery,
1981), wine grape (Weaver et al., 1961, Lobb and Woon,
1983, Ogg et al., 1991, Al-Khatib et al., 1993), grape
(Smith and Geronimo, 1984), kiwifruit (Lobb and Woon,
1983), lettuce (Hemphill and Montgomery, 1981), lentil
(Wall, 1996a), onion (Hemphill and Montgomery, 1981),
peanut (Smith and Geronimo, 1984), pepper (Hemphill and
Montgomery, 1981, Gilreath et al., 200Ib), potato (Lobb
and Woon, 1983, Haderlie et al., 1986, Leino and Haderlie,
1985), radish (Hemphill and Montgomery, 1981), rutabaga
(Hemphill and Montgomery, 1981), soybean (Slife, 1956,
94
-------
Bromoxynil
Chemcode = 2163,
5036,834
(octanoate), 2429
Photosynthesis inhibitor.
% national use (USDA
NASS, 1997)
Winter Wheat = 43%
Corn = 35%
Barley = 8%
Sorghum = 4%
Cotton = 2%
Usage rates ( % CA use)
(Chemcode = 834)
Wheat = 415 (28%)
Corn = 280 (2%)
Oats = 325 (16%)
Alfalfa = 260 (10%)
Barley = 340(1%)
Sorghum = 280 (<1%)
Cotton = 400 (12%)
Avg. app rate = 415
IPM recommended rates
Alfalfa = 280-420
Barley = 420-560
Corn = 280-425
Cotton = 560
Oats = 280-425
Sorghum = 280-425
Potential damage to cotton (Smith and Wiese, 1972), wine
grape (Al-Khatib et al, 1993), potato (Haderlie et al, 1986,
Leino and Haderlie, 1985).
Applied to wheat in early spring in WA. Grape vineyards
close to wheat fields may be damaged by drift (Al-Khatib,
1993).
95
-------
CGA-152005 or
prosulfuron
Chemcode = 5115
ALS inhibitor.
Not registered for use in
CA.
% national use (USDA
NASS, 1997)
Corn = 90%
Sorghum = 9%
Winter Wheat = 1%
IPM recommended rates
Oats =170-325
Winter wheat = 10-20
+2nd herbicide
Sorghum = 20-40
Potential damage to soybean (Al-Khatib and Peterson,
1999).
Soybean planted later than corn so post-emergent
application to com may occur when soybean is at an early
growth stage (Young et al., 2003).
96
-------
Chlorsulfuron
(Glean)
Chemcode = 2143
ALS inhibitor.
Controls a wide range of
broadleaf annuals and
inhibits Canada thistle.
Barley -Apply in fall or
spring any time after most
weeds have emerged and
after crop is in two-leaf
stage through second joint
Usage rates ( % national
use) (USDA NASS)
Winter Wheat = 94%
Barley = 5%
Oats= 1%
Usage rates ( % CA use)
Wheat =10 (2%)
Barley = 11 (1%)
oats = 9(1%)
Rights-of-way = 56 (72%)
Landscape = NA (19%)
Avg app rate (crops) =11
IPM recommended rates
Barley =8-18
Oats = 8-18
Winter wheat = 8-18
Potential damage to alfalfa (Al-Khatib et al, 1992a), canola
(Fletcher et al., 1996), cherry (Al-Khatib et al., 1992b,
Bhatti et al., 1995), cotton (Baker and Barrentine, 1986),
wine grape (Al-Khatib et al., 1993), mustard (Derksen,
1989), pea (Al-Khatib and Tamhane, 1999), smartweed
(Fletcher et al., 1996), soybean (Fletcher et al., 1996),
sunflower (Fletcher et al., 1996).
97
-------
Clomazone
Chemcode = 3537
Registered for use by
CDPR in 2002.
% national use (USDA
NASS, 1997)
Soybean = 65%
Cotton = 23%
Tobacco = 9%
Sweet potatoes = 2%
Usage rates ( % CA use)
(Chemcode = 3537)
Rice = 650 (100%)
Avg app rate = 650
IPM recommended rates
Potential damage to pecans (Carter, 1996).
98
-------
Clopyralid
Chemcode = 5135,
5050
(monoethanolamine
salt), 2339
Growth regulator.
Controls Canada thistle
and other broadleaf
weeds. Selective
herbicide. Dicot-specific
(esp. composites, legumes,
and smartweeds).
Alternative to picloram.
% national use (USDA
NASS, 1997)
Corn = 38%
Wheat = 33%
Sugarbeet= 16%
Asparagus < 1%
Usage rates ( % CA use)
(Chemcode = 5050)
Sugarbeet= 120(12%)
Rights-of-way = 56 (39%)
Rangeland= 135(20%)
Landscape = 34 (17%)
Avgapprate= 150
IPM recommended rates
Barley = 100-140 g ae ha1
Oats= lOS-WOgaeha1
Sugarbeets= 105-210
Orchard crop = 140-280 g
ai ha-1
Corn =135-280
Potential damage to field bean (Smith and Geronimo,
1984), cotton (Smith and Geronimo, 1984, Lanini, 1999),
wine grape (Lobb and Woon, 1983, Smith and Geronimo,
1984), kiwifruit (Lobb and Woon, 1983), lentil (Derksen,
1989), sunflower (Lanini, 1999), potato (Lawson et al.,
1992, Wall, 1994a), safflower (Lanini, 1999), soybean
(Smith and Geronimo, 1984), sunflower (Lanini, 1999),
tobacco (Smith and Geronimo, 1984), tomato (Smith and
Geronimo, 1984, Ray et al, 1996, Lanini, 1999).
Long-lasting herbicide can contaminate compost, resulting
in damage to asters, sunflowers, beans, peas and tomatoes
growing in the enriched soil.
99
-------
Dicamba
Chemcode = 90200,
780,849
(dimethylamine salt),
90849,5098,1829,
91829,90200,5110,
5057, 1113
Growth regulator.
Dicamba, an auxerric
herbicide, is regularly used
to control pigweed in
corn.
% national use (USDA
NASS, 1997)
Corn = 64%
Pasture = 14%
Wheat = 7%
Sorghum = 2%
Barley = 1%
Usage rates ( % CA use)
(Chemcode = 849)
Corn = 360 (36%)
Pasture = 290 (<1%)
Wheat =160 (27%)
Rights-of-way =112 (4%)
Barley =100 (5%)
Landscape = 145(10%)
Avg app rate = 220
IPM recommended rates
Barley = 67-105 g ae ha-1
Corn = 280-560 gai ha-1
Oats = 67-140 gae ha-1
Sorghum = 280 g ai ha-1
Winter wheat = 67-140 g
aiha-1
Winter wheat = 560-2,240
Potential damage to bean (Smith and Geronimo, 1984),
cotton (Smith and Wiese, 1972, Hamilton and Arle, 1979,
Smith and Geronimo, 1984), fieldbean (Smith and
Geronimo, 1984, Lyon and Wilson, 1986), grape (Smith
and Geronimo, 1984), pea (Al-Khatib and Tamhane, 1999),
peanut (Smith and Geronimo, 1984), pepper (Gilreath et al.,
200Ib), potato (Haderlie et al., 1986, Leino and Haderlie,
1985, Wall, 1994a), soybean (Auch and Arnold, 1978,
Smith and Geronimo, 1984, Lyon and Wilson, 1986, Wax
et al, 1969, Weidenhamer et al., 1989, Al-Khatib and
Peterson, 1999), sunflower (Derksen, 1989), tobacco
(Smith and Geronimo, 1984), tomato (Jordan and
Romanowski, 1974, Smith and Geronimo, 1984).
100
-------
Glufosinate
Chemcode = 3946
Glutamine synthetase
inhibitor.
Pre- and post-emergent
control in glufosinate-
resistant rice.
Seldom used in CA.
Usage rates (% CA use
2002)
Wine grapes = 500 (67%)
Almonds = 325 (17%)
Grapes = 380 ( 6%)
Avg app rate = 430
IPM recommended rates
Corn = 290-415
Potato = 420
Orchard crop = 840-1,680
Soybean = 290-415
Potential damage to corn (Ellis et al, 2003), rice (Ellis et al,
2003).
101
-------
Glyphosate
Chemcode = 2997,
1855 (isopropylamine
salt), 2301,2275,
2327 (trimesium),
3946
EPSP inhibitor.
Non-selective systemic
herbicide used to control
most annual and perennial
plants. Approved for
forestry applications
% national use (USDA
NASS, 1997)
Soybean = 43%
Corn= 13%
Winter Wheat = 5%
Cotton = 5%
Sorghum = 4%
Grapes = 1%
Usage rates ( % CA use)
(Chemcode =1855)
Rights-of-way = 1,460
(29%)
Corn =1,060 (2%)
Wheat = 1,030(<1%)
Wine Grapes = 1,000
Almonds = 1,050(15%)
Cotton =1,090 (7%)
Tomato = 1,110(1%)
Landscape = 720 (7%)
Oranges = 850 (5%)
Avg app rate = 970
IPM recommended rates
Winter wheat = 425-840
(hard dough stage)
Potato = 420-1240 gae
ha-1
May damage alfalfa (Al-Khatib et al, 1992a), corn (Ellis et
al, 2000, 2002, 2003, Al-Khatib et al, 2000, Hartzler,
2001), cotton (Miller et al, 2004), wine grape (Al-Khatib et
al., 1993), onion (Lobb, 1989), pepper (Gilreath et al.,
2000), potato (Haderlie et al., 1986; Leino and Haderlie,
1985, Lobb, 1989), rice (Kurtz and Street, 2003, ElHs et al,
2000, 2003), sorghum (Al-Khatib et al., 2003), soybean
(Al-Khatib and Peterson, 1999, Ellis and Griffin, 2002,
Ellis et al, 2002), tomato (Romanowski, 1980, Lobb,
1989, Gilreath et al., 2001a).
102
-------
Imazapyr
Chemcode = 2256,
2257( isopropylamine
salt)
ALS inhibitor.
Broad-spectrum herbicide
that controls terrestrial
annual and perennial
grasses and broadleaved
herbs, woody species, and
riparian and emergent
aquatic species.
Usage rates ( % CA)
(Chemcode = 2257)
Forest trees = 460 (90%)
Avg app rate = 500
Potential damage to blueberry (Yarborough and Bhowmik,
1989), corn, potato (Eberlein and Guttieri, 1994), soybean.
103
-------
Imazethapyr
(Pursuit)
Chemcode = 2340,
2341
ALS inhibitor.
Controls certain annual
grasses and broadleaf
weeds. Soybean and
peanut herbicide applied
PPI, early preplant,
preemergence, or
postemergence.
% national use (USDA
NASS, 1997)
Soybean = 92%
Corn = 4%
Alfalfa = 2%
Usage rates ( % CA)
Alfalfa = 90 (99%)
Avg app rate = 90
IPM recommended rates
Alfalfa = 53-105
Beans = 53
Soybean = 70
Potential damage to corn (Al-Khatib et al, 2000), potato
(Eberlein and Guttieri, 1994), rice (Bond et al, 2000),
sorghum (Al-Khatib et al., 2003), sunflower (Wall, 1996b).
Grapes are sensitive to most common growth regulator and
ALS inhibitor herbicides (Ball et al., 2004).
104
-------
MCPA
Chemcode = 2326,
5059, 784, 785, 786
(dimethylamine salt),
787, 788
Growth regulator.
% national use (USDA
NASS, 1997)
Wheat = 78%
Barley = 12%
Rice = 2%
Growth regulator or
auxenic herbicides which
include 2,4-D, 2,4-DB
and MCPA are restricted
from use in SJV in CA
from March 16 to Oct 15.
Usage rates ( % CA)
(Chemcode = 786)
Wheat = 790 (63%)
Oats = 800 (21%)
Rice = 300 (5%)
Avg app rate = 730
IPM recommended rates
Barley = 280-840 g ae ha1
Oats = 280-560 gae ha-1
Winter wheat = 560-1,120
Potential damage to cotton (Smith and Wiese, 1972),
sugarbeet (Byford and Prince, 1976).
Grapes are sensitive to most common growth regulator and
ALS inhibitor herbicides (Ball et al, 2004).
105
-------
Metribuzin
Chemcode = 1692,
4079
Photosynthesis inhibitor.
Controls annual broadleaf
weeds.
% national use (USDA
NASS, 1997)
Soybean =50%
Potato = 11%
Wheat = 10%
Corn = 5%
Tomatoes =1%
Not used in Kern County,
CA or in southern desert
regions that have high
alkaline soils.
Usage rates ( % CA)
Potato = 460 (14%)
Wheat = 520 (<1%)
Alfalfa = 500 (25%)
Tomato = 350 (37%)
Asparagus = 1,000(13%)
Avg app rate = 400
IPM recommended rates
Alfalfa = 280-1,120
Barley = 50-560
Corn= 105
Potato = 280-1,120(PRE),
280-560 (POST)
Tomato = 560-1,120
Winter wheat = 212-560
Potential damage to cotton (Hurst, 1982).
106
-------
Met sul fur on- methyl
Chemcode = 2222
ALS inhibitor.
Not registered for use in
CA.
% national use (USDA
NASS, 1997)
Winter Wheat = 51%
Pasture = 43%
Barley = 4%
Hay= 1%
IPM recommended rates
Barley = 4.2
Potential damage to bean (Boutin et al., 1999), onion
(Lobb, 1989), potato (Lobb, 1989), tomato (Ray et al, 1996,
Lobb, 1989).
Grapes are sensitive to most common growth regulator and
ALS inhibitor herbicides (Ball et al., 2004).
107
-------
MSMA
Chemcode = 34
Used to control emerged
annual grasses, bahiagrass
and dallisgrass, and
nutsedge.
% national use (USDA
NASS, 1997)
Cotton = 97%
Usage rates ( % CA)
(Chemcode = 34)
Cotton = 1,270(53%)
Bermudagrass = 2,410
(21%)
Landscape = NA (16%)
Avg app rate = 1,800
Potential damage to corn, cotton, soybean (Bode and
McWhorter, 1977).
Nicosulfuron
Chemcode = 3829
ALS inhibitor.
% national use (USDA
NASS, 1997)
Corn = 99%
Usage rates ( % CA)
(Chemcode = 3829)
Corn = 35 (97%)
Avg app rate = 35
IPM recommended rates
Corn = 35.
Potential damage to soybean (Al-Khatib and Peterson,
1999).
108
-------
Paraquat dichloride
Chemcode = 1601
Restricted-use herbicide.
% national use (USDA
NASS, 1997)
Corn = 29%
Soybeans = 20%
Cotton = 15%
Wheat = 6%
Alfalfa = 5%
Grapes = 5%
Usage rates ( % CA)
(Chemcode =1601)
Cotton = 460 (23%)
Grapes = 695 (22%)
Almonds = 852 (18%)
Alfalfa = 706 (14%)
Avg app rate = 661
Potential damage to corn, cotton, potatoes, soybean. Drift
symptoms often appear as spots on leaves, with negligible
long lasting effects.
109
-------
Picloram
(Tordon)
Chemcode = 593, 835,
5330,1099
Growth regulator.
Dicot-selective persistent
herbicide used to control
annual and perennial
broadleaved herbs and
woody species.
Seldom used in CA 2000.
Restricted-use herbicide.
% national use (USDA
NASS, 1997)
Pasture = 91%
Hay = 5%
Winter wheat = 2%
Barley <1%
Usage rates ( % CA)
(Chemcode = 593)
Regulatory = NA (90%)
Avg app rate = 1,360
Potential damage to field bean (Smith and Geronimo,
1984), cotton (Smith and Wiese, 1972, Smith and
Geronimo, 1984), wine grape (Lobb and Woon, 1983),
grape (Smith and Geronimo, 1984), kiwifruit (Lobb and
Woon, 1983), peanut (Smith and Geronimo, 1984), potato
(Lobb and Woon, 1983), soybean (Smith and Geronimo,
1984), tobacco (Smith and Geronimo, 1984), tomato (Smith
and Geronimo, 1984).
Picloram can be released into the soil through the roots of
treated plants and can be readily absorbed by neighboring
plants.
Soybean planted later than corn so post-emergent
application to com may occur when soybean is at an early
growth stage (Young et al, 2003).
110
-------
Primisulfuron
Chemcode = 5103
ALS inhibitor.
Not registered for use in
CA.
% national use (USDA
NASS, 1997)
Corn= 100%
IPM recommended rates
Corn = 40
Potential damage to soybean (Bailey and Kapusta, 1993).
Soybean planted later than corn so post-emergent
application to com may occur when soybean is at an early
growth stage (Young et al., 2003).
Propanil
Chemcode = 503
Photosynthesis inhibitor.
Primary herbicide for rice
weed control.
% national use (USDA
NASS, 1997)
Rice= 100%
Usage rates ( % CA)
(Chemcode = 503)
Rice = 4,550 (100%)
Avg app rate = 4,550
Potential damage to cotton (Hurst, 1982), grapes, pistachio,
prunes, soybean (Bode and McWhorter, 1977) and
sorghum.
Ill
-------
Quinclorac
Chemcode = 5104
Growth regulator.
% national use (USDA
NASS, 1997)
Rice= 100%
Usage rates ( % CA) -
Landscape = 840 (100%)
Avg app rate = 840
Potential damage to corn, cotton (Snipes et al., 1992),
soybean, tomato, vegetables.
Rimsulfuron+thifensul
furon
Chemcode = 3835+
2237
ALS inhibitor.
% national use (USDA
NASS, 1997)
Corn = 84%
Potatoes = 13%
Tomato = 2%
Usage rates ( % CA)
(Chemcode = 3835)
Potatoes = 34 (18%)
Tomatoes = 11(80%)
Avg app rate =11
IPM recommended rates
Corn = 11
Potato = 17-26
Tomato = 18-35
Potential damage to soybean (Al-Khatib and Peterson,
1999).
Small grains, canola, sugar beets, peas, and onions are very
sensitive to rimsulfuron drift (PNW Weed Management
Handbook).
Soybean planted later than corn so Post-emergent
application to com may occur when soybean is at an early
growth stage (Young et al., 2003).
112
-------
Sethoxydim
Chemcode = 2177
Lipid synthesis inhibitor.
Selective herbicide is used
to kill and suppress annual
and perennial grasses.
% national use (USDA
NASS, 1997)
Soybean = 64%
Alfalfa = 8%
Cotton = 7%
Sugarbeets = 4%
Usage rates ( % CA)
Alfalfa = 360 (46%)
Cotton = 310 (9%)
Sugarbeets = 240 (7%)
Wine grapes = 260 (5%)
Tomatoes = 220 (4%)
Avg app rate = 280
IPM recommended rates
Alfalfa =110-560
Beans = 325-540
Cotton = 100-420
Potato = 320-530
Soybean = 110-315
Sugarbeets = 210-530
Orchard crop = 310-530
Tomato = 320
Potential damage to corn (Al-Khatib et al, 2000), sorghum
(Al-Khatib et al., 2003).
113
-------
Sulfometuron methyl
(Oust)
Chemcode = 2149
ALS inhibitor.
Sulfometuron is used to
control herbaceous weeds
in newly established forest
plantations (Cameron and
Turvey, 1977,
Constantini, 1986).
Usage rates ( % CA)
Rights-of-way =125
(79%)
Landscape maintenance
125(11%)
Avg app rate =125
Potential damage to potato (Westra et al., 1991).
Stunting of newly planted loblolly pine seedlings (Gjerstad
and Nelson, 1983); greater mortality in slash pine at rate of
280 g ai ha-1 (Knowe, 1984); reduced survival of blue
spruce at rates of 100 g ai ha-1 or higher and Scots pine at
300 g ai ha-1 or higher (Bacon et al., 1983). Sulfometuron
is known to reduce cell division in herbaceous plants
(DuPont, 1984) and may stunt the growth of newly planted
seedlings. Grapes are sensitive to most common growth
regulator and ALS inhibitor herbicides (Ball et al., 2004).
114
-------
Thifensulfuron
(Harmony GT)
Chemcode = 2237
ALS inhibitor.
Not registered for use in
CA. Controls broadleaf
weeds.
% national use (USDA
NASS, 1997)
Winter Wheat = 69%
Soybean =15%
Barley = 9%
Corn = 6%
IPM recommended rates
Barley = 16-32
Soybean = 5
Winter wheat = 16-32 (+)
Potential damage to alfalfa (Al-Khatib et al, 1992a), canola
(Wall et al., 1995), wine grape (Al-Khatib et al., 1993), pea
(Al-Khatib and Tamhane, 1999), potato (Lawson et al.,
1992).
Wheat herbicide in WA applied in early spring. Grape
vineyards close to wheat fields may be damaged by drift
(Al-Khatib, 1993).
Thifensulfuron:tribenu
ron
(Harmony Extra)
Chemcode = 2237 +
2338
ALS inhibitor.
Not registered for use in
CA.
IPM recommended rates
Barley = 16-32
Oats= 16-21
Winter wheat =10-21+5-
10
Potential damage to buckwheat (Wall, 1994b), canola
(Wall, 1994b, Ball, 1994, Brammer et al., 1996, Wall,
1997), lentil (Wall, 1994b, Gealy et al., 1995), pea
(Mallory-Smith, 1993, Wall 1994b, Gealy et al., 1995),
potato (Lawson and Wiseman, 1991), sunflower (Wall,
1994b, Wall, 1997).
115
-------
Tribenuron
(Express +)
Chemcode = 2338
ALS inhibitor.
Not registered for use in
CA.
Usage rates ( % national
use)(USDANASS)
Winter Wheat = 11(85%)
Barley = 11(14%)
Oats =11(1%)
IPM recommended rates
Barley = 9-18
Winter wheat = 8-17
Potential damage to potato (Wall, 1994a).
Grapes are sensitive to most common growth regulator and
ALS inhibitor herbicides (Ball et al, 2004).
116
-------
Trick) pyr
Chemcode = 2170,
2131 (triethylamine
salt)
Growth regulator.
Selective systemic
herbicide used to control
woody and herbaceous
broadleaf plants.
Dicot-specific. Rice
herbicide. Propanil and
triclopyr are widely used
rice herbicides in CA.
% national use (USDA
NASS, 1997)
Pasture = 68%
Rice = 29%
Hay= 1%
Usage rates ( % CA)
Pasture = 470(1%)
Rice = 240 (80%)
Landscape maintenance :
45 (8%)
Avg app rate = 340
Potential damage to bean (Smith and Geronimo, 1984),
cotton (Smith and Geronimo, 1984, Jacoby et al, 1990,
Snipes et al., 1991), wine grape (Lobb and Woon, 1983),
grape (Smith and Geronimo, 1984), peanut (Smith and
Geronimo, 1984), potato (Lobb and Woon, 1983, Lobb,
1989), soybean (Smith and Geronimo, 1984), tomato
(Smith and Geronimo, 1984, Lobb, 1989, Ray et al, 1996,
Ray et al, 1996), tobacco (Smith and Geronimo, 1984).
PNW Weed Management Handbook.
117
-------
Table C2. High potential risk herbicides and the associated susceptible crops ranked by reported
EC 10 values (given in parentheses in g/m2) in the published literature. Crops ranked as high sensitivity
had EC 10 values less than about 3.3% of the application rate; medium sensitivity corresponds to
EC 10 values between 3.3% and 6.7% of the application rate and low sensitivity corresponds to EC 10
values between 6.7% and 10% of the application rate.
Chemical
2,4-DB
2,4-D
Bromoxynil
High Sensitivity
bushbean (22)
cabbage (21)
cotton (1-11)
wine grape (1-20)
grape (1)
kiwifruit(l-14)
pepper (1-11)
potato (1;1 1-1 12)
sunflower (10)
tomato (1-35)
cotton (11)
potato (11)
Medium Sensitivity
alfalfa (11-1 12)
field bean (11. 2- 11 2)
buckwheat (38)
cauliflower (21-104)
cucumber (21-1 04)
onion (2 1-1 04)
peanut (11-1 12)
radish (21-104)
rutabaga (21-104)
soybean (11 -112)
sugarbeet (35-70)
tobacco (11-1 12)
turnip (21-104)
wine grape (11-43)
Low Sensitivity
canola (75)
carrot (44-440)
lettuce (21-208)
118
-------
Chlorsulfuron
Clomazone
Clopyralid
canola (<0.05)
cherry (0.1 -0.3)
wine grape (<0.3)
mustard (0.06-0.6)
pea (0.04-0.09)
smartweed (0.05-
0.09)
soybean (<0.05)
sunflower (<0.05)
cotton (0.1 -0.3)
wine grape (0.2)
grape (1-11)
kiwifruit (0.02)
lentil (<3)
potato (1)
safflower (0.3-3)
soybean (0.1-11)
sunflower (<0.3)
tobacco (0.1-11)
tomato (1-11)
alfalfa (0.9-2.6)
pecans
119
-------
Dicamba
Glufosinate
Glyphosate
Imazapyr
Imazethapyr
Metribuzin
Met sul fur on- methyl
MCPA
MSMA
field bean (1-11)
cotton (1-11)
grape (<0.1)
peanut (1-11)
pepper (1-11)
potato (3; 11-56)
soybean (<11)
sunflower (0.4)
tobacco (1)
tomato (1)
corn(18;46)
wine grape1 (4)
pepper (<30)
potato (10-50)
tomato (1-10)
blueberry (<12)
sunflower (1.5-3)
bean (O.045)
onion (0.1)
potato (0.1)
tomato (<0.1)
corn (26)
rice (26)
alfalfa (43)
cotton (43)
onion (50)
sorghum (34- 11 2)
potato (<22)
rice (4-8)
cotton (<20)
cotton (<56)
pea (25)
rice (70)
soybean (70-
140)
corn (7-9)
potato (7)
sorghum (7)
sugarbeet (70)
soybean (< 100)
120
-------
Nicosulfuron
Paraquat dichloride
Picloram
Pri mis ulf uro n-m ethyl
Propanil
Prosulfuron
Quinclorac
Rimsulfuron
Sethoxydim
Sul fom eturon -me thyl
Thi fens ulfuro n-m ethyl
Tribenuro n-m ethyl
soybean (1)
field bean (0.1)
grape (<0.1)
tobacco (0.1 -1.0)
cotton (100)
soybean (0.2)
cotton (9)
potato (3.8)
buckwheat (<0.23)
canola (<0.23)
wine grape (<0.3)
lentil (<0.23)
pea (0.23-0.55)
sunflower (0.23-0.45)
buckwheat (0.23)
canola (<0.23)
lentil (<0.23)
pea (0.23-0.55)
potato (0.3)
sunflower (0.23-0.45)
wine grape (1-10)
kiwifruit(l-lO)
peanut (1-11)
soybean (1-11)
soybean (200)
corn (17)
soybean (1.2)
soybean (<4)
soybean (2.4)
alfalfa (2.6)
potato (3)
121
-------
Tricio pyr
cotton (1-11)
grape(1)
tomato (1-10)
field bean (11-112)
peanut(11-112)
potato (10-50)
soybean (11-112)
tobacco (11-112)
1 Glyposate effects greater on newly established and 1-year-old plantings than on 2-year-old plantings
(Al-Khatib et al, 1993).
122
-------
Appendix D
Background Information on the High Potential Risk Herbicides
D.I. 2,4-D
The herbicide 2,4-D (2,4-dichlorophenoxy acetic acid) was the sixth most widely used active
ingredient (third among herbicides) in the agricultural sector with 18 million kg a.i. applied in the United
States in 1997 (Gianessi and Marcelli, 2000). The compound is the top herbicide worldwide because
it is low in price and is effective in controlling many types of broadleaf weeds in many agricultural
crops. 2,4-D has moderate to low acute toxicity, low reproductive toxicity, does not cause birth
defects, is not carcinogenic (USEPA, 1997), has low potential to cause neurotoxicity, does not cause
genetic damage, and is not a threat to wildlife at the label rate (USEPA, 1988a, 1988b). The
compound is used in cultivated agriculture, in pasture and rangeland applications, forest management,
2,4-d - Herbicides
Estimated Annual Agricultural Use
Average use of
Active Ingredient
Grams per Hectare
of county per year
Q No Estimated Use
D <0.019
D 0.019-0.109
D 0.110-0.627
D 0.628 - 3.085
• >=3.086
Crops
pasture
wheat
corn
soybeans
fallowland
other hay
sugarcane
barley
sorghum
rice
Total
Kilograms
Applied
7,833,212
3,751.702
1,758,810
1,468.219
1,058,334
662,851
568,201
346,003
251,873
222,773
Percent
National Use
42.32
20.66
9.69
8,09
5.83
3.65
3.13
1.91
1.39
1.23
Figure Dl. Agricultural use of 2,4-D in the United States in 1997 (USGS,
PNSP available at http://ca.water.usgs.gov).
123
-------
home, garden, and to control aquatic vegetation. 2,4-D use is highest in the Corn Belt and wheat
states including North Dakota, Texas, Washington, Oregon and Montana (Figure Dl). In comparison,
the amount of 2,4-D use in California is low because 2,4-D is banned from use in the SJV from March
16 to October 15 in order to avoid off-site spray drift to susceptible plants. Drift of auxinic herbicides
such as 2,4-D can cause considerable damage to many crops including cotton (Miller et al., 1963;
Smith and Wiese, 1972; Smith and Geronimo, 1984; Banks and Schroeder, 2002), grapes (Smith
and Geronimo, 1984), wine grapes (Weaver et al., 1961; Lobb and Woon, 1983; Ogg et al., 1991;
Al-Khatib et al., 1993), beans (Smith and Geronimo, 1984), alfalfa (Al-Khatib et al., 1992a), lettuce
(Hemphill and Montgomery, 1981), ornamentals, and all broadleaf species which comprise a
considerable portion of the total agronomic production value in California. In particular, cotton,
grapes, wine grapes and tomatoes are highly sensitive to 2,4-D and susceptible to drift damage at 3%
of the label rate (Table C2). 2,4-D alters the plant's metabolism and growth characteristics, often
causing a proliferation of abnormal growth that interferes with transport of nutrients throughout the
plant. The chemical can damage plants on contact, cause abnormalities to existing plant parts, and
affect new and future growth and development. The compound is absorbed through the cuticles of
leaves and shoots and is translocated throughout the plant. There are many forms or derivatives
(esters, amines, salts) of 2,4-D of which the esters are the most volatile and prone to drift
D.2. 2,4-DB
4-(2,4-dichlorophenoxy)butyric acid (2,4-DB) is used for the control of many annual and
perennial broadleaf weeds in alfalfa, peanuts, soybeans and other crops. In the plant, 2,4-DB changes
to 2,4-D and inhibits growth at the tips of stems and roots. It has a very low toxicity to fish, birds and
mammals and has a half-life of about 7 days. The compound may be found in formulations with other
herbicides such as cyanazine, MCPA, benazolin, linuron and mecoprop. The dominant issue with
using 2,4-DB is cotton injury. Adjacent cotton fields are extremely susceptible to 2,4-DB drift.
Although 2,4-DB is restricted from use in the SJV from March 16 to October 15, injury to sensitive
crops including cotton and grapes due to off-site drift may still occur. In California, the injury is likely
to occur when 2,4-DB is used to control weeds in seedling alfalfa in adjacent fields. Agricultural use of
124
-------
2,4-db - Herbicides
Estimated Annual Agricultural Use
Average use of
Active Ingredient
Grains per Hectare
of county per year
No Estimated Use
Crops
alfalfa
peanuts
soybeans
other hay
mint
Total
Kilograms
Applied
173,180
72,501
14,468
3,096
3,053
429
Percent
National Use
54.93
27,18
5.42
1,16
1 14
0.16
D 0.047-0114
D 0.115-0.226
D 0,227-0490
>=0.491
Figure D2. Agricultural use of 2,4-DB in the United States in 1997 (USGS,
PNSP available at http://ca.water.usgs.gov).
the compound in California represents about 8% of the 272,000 kg a.i. applied domestically. It is
widely used in all states in the continental U.S. (Figure D2).
D.3. Bromoxynil
Bromoxynil was first registered in the United States in 1965 for use as a herbicide to control
grassy and broadleaf weeds in wheat and barley and was reregistered in 1998. When bromoxynil was
registered for use in transgenically modified BXN® cotton in 1995, this breakthrough allowed safe
over-the-top application of a selective postemergence herbicide in cotton. This then-new technology
dramatically impacted weed management practices in cotton. This breakthrough was followed in 1997
with the registration of glyphosate-resistant cotton which provided greater control of many grasses as
well as broadleaf weeds. Consequently, the acreage planted of Roundup Ready™ cotton far exceeds
that of BXN® cotton in California and the United States. Bromoxynil usage in the agricultural sector
125
-------
was about 1,360,000 kg a.i. in the United States in 1997 (Gianessi and Marcelli, 2000). Bromoxynil
is most heavily used in the Corn Belt and the wheat states (Figure D3). In California, bromoxynil is
used primarily on BXN® cotton, wheat and garlic. BXN® cotton represents less than 5% of total ha
planted of cotton in California.
At the time of reregistration in 1998, the EPA had concerns for potential occupational risks to
workers handling bromoxynil products and toxicological risks to aquatic and terrestrial organisms. The
EPA classifies bromoxynil phenol and bromoxynil octanoate in Toxicity Category II/III for acute
effects via oral and either inhalation or dermal routes (USEPA, 1998). Bromoxynil phenol has been
classified as a Group C, possible human carcinogen. The compound is considered to be
developmentally toxic. These factors are reflected in the declining trend in agricultural usage of
bromoxynil in the United States. For example, bromoxynil use on winter wheat in the U.S. decreased
Bromoxynil - Herbicides
Estimated Annual Agricultural Use
Average use of
Active Ingredient
Grams per Hectare
of county per year
fj No Estimated Use
Q 0.012
D 0012-0.047
D 0046-0.187
D 0188-1517
• >=1518
Crops
wheat
com
barley
sorghum
cotton
alfalfa
oats
onions
mint
Total
Kilograms
Applied
567,751
473,948
110,400
60,550
34,177
16,695
16,411
11,778
7,012
6,669
Percent
National Use
43,10
3598
8.33
460
2.59
1.27
1 25
0.89
053
0.51
Figure D3. Agricultural use of bromoxynil in the United States in 1997
(USGS, PNSP available at http://ca.water.usgs.gov).
126
-------
from 91,000 kg a.i. in 1992 to 11,300 kg a.i. in 2002 (USDA Pesticide Use Data). Similar declines in
bromoxynil use on corn were also noted. Agricultural usage of bromoxynil octanoate in California
declined from 67,000 kg a.i. in 1996 to 33,000 kg a.i. in2002 while usage of bromoxynil heptanoate
increased from 4 kg a.i. in 1997 to 20,000 kg a.i. in 2002 (CPUR).
D.4. Chlorsulfuron
Chlorsulfuron was the first SU herbicide developed by DuPont in the mid-1970s and
promoted to wheat producers as "Glean" in 1982. The compound is a pre- or post-emergent
systemic herbicide used to control many broadleaf and some annual grass weeds in cereal crops on
lands having a soil pH of 7.5 or lower. It is a low toxicity herbicide but is persistent in some types of
soils. Shortly after the introduction of Glean for wheat, crop failures in fruits and flowers in south-
central Washington were suspected to be caused by SUs drifting from wheat fields in the Horse
Heaven Hills. It was impossible to detect any residual traces of the SU herbicide in the damaged
plants. However, in a controlled experiment, EPA scientists found that cherry trees showed no visible
foliar injury but bore no fruit when exposed to .2% of the application rate for Glean (Fletcher et al.,
1996). In fact, many crops such as alfalfa (Al-Khatib et al, 1992a), canola (Fletcher et al., 1996),
cherry (Al-Khatib et al., 1992b, Bhatti et al., 1995), cotton (Baker and Barrentine, 1986), wine grape
(Al-Khatib et al., 1993), mustard (Derksen, 1989), pea (Al-Khatib and Tamhane, 1999), smartweed
(Fletcher et al., 1996), soybean (Fletcher et al., 1996) and sunflower (Fletcher et al., 1996) are
susceptible to damage from exposure to Chlorsulfuron at sublethal levels. These early drift problems
with SU herbicides raised national concerns about the safe use of these low dose, high potency
herbicides. In 1992, DuPont ceased sales of Glean in seven Great Plains states and now requires
Glean be mixed with other herbicides to ensure that resistant weeds are killed at lower application
rates. Because of restrictions in sales and usage, agricultural use of Chlorsulfuron is limited to the
wheat-producing states in the western U.S. (Figure D4). In California, Chlorsulfuron is primarily used
for weed control in landscape maintenance and rights-of-way.
127
-------
Chlorsulfuron - Herbicides
Estimated Annual Agricultural Use
Average use of
Active Ingredient
Grams per Hectare
of county per year
Q No Estimated Use
D <-0.000
D -0.001 - o.ooo
n 0.000-0.001
• 0 002 - 0.005
• >=0.006
Crops
wheat
barley
oats
fallowland
Total
Kilograms
Applied
25,408
1,249
275
99
Percent
National Use
9400
4.62
1.02
0.37
Figure D4. Agricultural use of chlorsulfuron in the United States in 1997
(USGS, PNSP available at http://ca.water.usgs.gov).
D.5. Clomazone
Clomazone is a broad-spectrum herbicide used for control of annual grasses and broadleaf
weeds in various crops including cotton, soybeans and tobacco mainly in the Corn Belt, Texas and the
eastern part of the United States (Figure D5). Only 227 kg a.i. were applied in California in 2002
when clomazone was first registered for use in the state. It can be applied early preplant, preemergent,
or preplant-incorporated, depending upon the crop, geographic area and timing. Acute health effects
from exposure to clomazone are minimal due to its low mammalian toxicity. The compound is
expected to have minimal impact on the environment since it neither accumulates in the soil nor moves
into groundwater. However, its relatively high vapor pressure (19.2 mPa @ 25°C) makes it
susceptible to off-target volatility. Off-site spray drift damage to pecans and other vegetation were
128
-------
Clomazone - Herbicides
Estimated Annual Agricultural Use
Average use of
Active Ingredient
Grams per Hectare
of county per year
n
D
n
n
No Estimated Use
0.016-0.102
0.103 -0764
0.765 - 3.853
>=3 954
Crops
soybeans
cotton
tobacco
sweet potatoes
pumpkins
green peas
cucumbers
squash
sweet peppers
green beans
Total
Kilograms
Applied
742,716
265,679
98,479
17,303
6,660
4,897
2,165
1,202
562
-------
Research Service; Benbrook, 2004).
D.6. Clopyralid methyl
Clopyralid methyl is used to control annual and perennial broadleaf weeds in pastures and
rangelands and some agricultural crops including corn, wheat and sugarbeets. In California, clopyralid
is a key herbicide applied by air to control yellow starthistle, a noxious weed (DiTomaso et al., 2004).
Agricultural use of clopyralid was limited to 625 kg a.i. applied to sugarbeets in California in 2002 and
was highest in the Corn Belt, in particular Minnesota, Iowa, Nebraska, Wisconsin and Illinois (Figure
D6). Domestic use on corn increased more than twofold from 1997 to 2001. It is used at low rates
and has very low toxicity to fish, birds and mammals. The compound is not strongly adsorbed by the
soil and may be persistent in soils under anaerobic conditions and in soils with a low microorganism
content. Special precaution is warranted in the use of clopyralid because many crop species are highly
sensitive to the herbicide (Table C2). Potatoes are extremely sensitive to clopyralid with damage
Clopyralid - Herbicides
Estimated Annual Agricultural Use
Average use of
Active Ingredient
Grams per Hectare
of county per year
Q No Estimated use
D <0.014
D 0.014-0088
D 0.089-0.401
LJ 0.402 • 1 259
B >=1.260
Crops
corn
wheat
sugarbeets
pasture
barley
oats
other hay
mint
asparagus
Total
Kilograms
Applied
152.841
134.254
68,056
24,976
11.500
4.387
2,737
1,269
616
93
Percent
National Use
3813
33.49
16.9S
623
2.87
1.09
068
0,32
015
002
Figure D6. Agricultural use of clopyralid in the United States in 1997 (USGS,
PNSP available at http://ca.water.usgs.gov).
130
-------
occurring at 7% of typical application rate (Lawson et al., 1992, Wall, 1994a). The compound is
similar in structure and mode of action to the herbicide picloram (Cremlyn, 1991).
D.7. Dicamba
Dicamba, first registered in the United States in 1967 (USEPA, 1983), is a selective herbicide
used to control annual and perennial broadleaf weeds in grain crops and grasslands. Dicamba was the
10th most heavily used herbicide in the United States in 1997 with 4.5 million kg a.i. applied to grain
crops (Gianessi and Marcelli, 2000). Similar to 2,4-D, dicamba usage is highest in the Corn Belt and
wheat states including North Dakota, Texas and Montana (Figure D7). Soybeans are particularly
sensitive to dicamba at time of post-emergent application to corn (Auch and Arnold, 1978; Smith and
Geronimo, 1984; Weidenhamer et al., 1989; Al-Khatib and Peterson, 1999), more so than to 2,4-D
(Burnside and Lavy, 1966). Soybean are planted later than corn so Post-emergent application to corn
Dicamba - Herbicides
Estimated Annual Agricultural Use
Average use of
Active Ingredient
Grams per Hectare
of county per year
n
a
n
No Estimated Use
<0,014
0.014-0.072
0.073 - 0 329
0.330-2.627
=•=2628
Crops
corn
pasture
wheat
fallowland
sorghum
barley
other hay
sugarcane
sod
Total
Kilograms
Applied
3 053 526
686,898
366,355
353,325
33.904
S7.407
26.895
24,435
11,061
7,432
Percent
National Use
6497
14.62
7.79
7.52
2.00
1.43
0.57
0.52
0.24
0.16
Figure D7. Agricultural use of dicamba in the United States in 1997 (USGS,
PNSP available at http://ca.water.usgs.gov).
131
-------
may occur when soybean is at an early growth stage (Young et al., 2003). Other crops including field
bean (Smith and Geronimo, 1984; Lyon and Wilson, 1986), cotton (Smith and Wiese, 1972; Hamilton
and Arle, 1979; Smith and Geronimo, 1984), grape (Smith and Geronimo, 1984), pepper (Gilreath et
al., 2001b) and tomato (Jordan and Romanowski, 1974, Smith and Geronimo, 1984) are highly
sensitive to off-site dicamba drift (Table C2). In California, dicamba is a state-limited-use pesticide
registered for use on corn, wheat, landscape management and rights-of-way (California Code of
Regulations Title 3 Section 6400).
Dicamba is highly mobile in most soils but has a half-life of only 1 to 6 weeks. The compound
is relatively volatile and can evaporate from leaf surfaces and from the soil. Dicamba is slightly toxic to
mammals and fish, non-toxic to aquatic invertebrates and birds, does not accumulate in terrestrial and
aquatic animals, and does not cause birth defects. These factors make dicamba a safe herbicide for
weed management. On the other hand, dicamba has moderate to high volatility as indicated by a
vapor pressure of 4.5 mPa @ 25°C. Consequently, the potential damage to nontarget crops from off-
site drift of dicamba is a concern, more so in the Midwest where more croplands are used to produce
grain crops than in California.
D.8. Glufosinate
Glufosinate ammonium is a broad-spectrum contact herbicide used extensively in Europe and
North America with GM crops that are resistant to the herbicide. These GM crops include corn, rice,
soybeans, canola, sugar beets, sugar cane and sweet potato. It is also used in landscape maintenance,
rights-of-way and commercial areas where complete vegetation kill is desired. A concern in the use of
glufosinate is its potential to damage conventional plants exposed to off-site spray drift. The
compound, like glyphosate, is readily adsorbed by organic soil particles, decomposes rapidly and is
not a danger to leaching and contamination of groundwater. The compound has lower toxicity to
humans and animals than many other herbicides in current use but may have some side effects to
humans, animals and beneficial insects. Glufosinate inhibits photorespiration by causing a buildup of
ammonia and a reduction of the amino acid glutamine. It inhibits the same enzyme called glutamine
synthetase in animals. Exposure to high levels of glufosinate may result in neurotoxicity (Matsumura et
132
-------
Figure D8. Glufosinate use in kg a.i. applied in
California in 1997-2002 (CPUR).
al., 2001) and birth defects (Garcia et al., 1998) in humans and animals. Glufosinate has been found to
be toxic to some aquatic organisms and beneficial soil micro-organisms (USEPA, 1990a, 1990b,
1990c, 1993).
Glufosinate usage in California has increased dramatically from 5 kg a.i. in 1997 to 2,651 kg
a.i. in 2002 (Figure D8, CPUR) as a result of the development of glufosinate-resistant crops.
Applications on wine grapes and alfalfa accounted for 67% and 17% of the state total in 2002. No
glufosinate applications were registered in Fresno County in 2000 and only 120 kg a.i. were applied in
Fresno County in 2002. Agricultural usage of glufosinate for the United States in 1997 were not
reported by the USDA NASS because it is a relatively new herbicide. The adoption of glufosinate-
resistant cotton, soybean and corn lags behind that of glyphosate-resistant hybrids in the United States
because of the higher cost of Liberty® herbicide (about $85 per gallon) and, therefore lower
profitability.
D.9. Glyphosate
Glyphosate was regulated for agricultural use in the United States in 1971 and was limited to
nonselective applications prior to crop emergence for control of annual and perennial broadleaf weeds
and grasses in no-till crops (Wilson et al., 1985; Wilson and Worsham, 1988). Its use increased
dramatically following the introduction of glyphosate-resistant varieties of soybean in 1996 and cotton
133
-------
in 1997 in the United States (Bradshaw et al., 1997). Over 15 million kg of glyphosate were applied
to crops in 1997 (Gianessi and Marcelli, 2000), primarily to soybeans, corn, and citrus in the Midwest
and to cotton, grapes, almonds and citrus in California. By 1999, glyphosate usage in the agricultural
sector more than doubled to 30-33 million kg a.i. (Donaldson et al., 2002) as more producers
converted to glyphosate-resist ant soybean and cotton from the conventional varieties. Currently,
approximately 85% of soybean and 60% of cotton acreage are glyphosate-resistant in the United
States (USDA Economic Research Service; Benbrook, 2004). Agronomic application of glyphosate
is highest in California in terms of total ha treated and kg a.i. used (Table 4). Substantial glyphosate
use also occurs in the Corn Belt, Texas and other states where soybean, wheat, small grains and
cotton are grown (Figure D9).
Producers of RR soybean have come to rely solely on postemergent weed management
Glyphosate - Herbicides
Estimated Annual Agricultural Use
Average use of
Active Ingredient
Grams per Hectare
of county per year
D
D
D
No Estimated Use
<0.019
0.019-0.062
0.083 - 0.236
0.237 - 0.865
>=0.866
CrooR
soybeans
corn
fallowland
wheat
cotton
citrus
sorghum
pasture
almonds
grapes
Total
Kilograms
App-'iec
6,753.753
2,114,673
1.536,111
854,553
840,668
830,803
677,716
400,058
270,453
248,281
Percent
National Use
4324
1354
10.15
5.66
5. 35
532
4.34
2.56
1.73
1.59
Figure D9. Agricultural use of glyphosate in the United States in 1997
(USGS, PNSP available at http://ca.water.usgs.gov).
134
-------
programs which can have serious consequences of off-site drift on adjacent sensitive crop fields. Since
1999, the chemical has been among the top herbicides in agricultural drift complaints reported in many
states including Arkansas, Iowa, Mississippi and South Dakota.
Of the high potential risk herbicides listed in Table 4, glyphosate is of primary concern in
California because: (1) glyphosate is the leading herbicide in California and Fresno County (Tables 1
and 4) based on agricultural usage in terms of kg of chemical used or ha treated; and (2) glyphosate is
acutely toxic to most plants and can cause severe injury or destruction to nontarget plants. The United
States Fish & Wildlife Service (USFWS) has recognized 74 endangered plants that maybe threatened
by glyphosate alone (USEPA, 1986). Exposure to sublethal levels of glyphosate can severely reduce
seed quality (Locke et al., 1995) and increase susceptibility of certain plants to disease (Brammall and
Higgins, 1988). When exposed to sublethal levels of glyphosate, the most sensitive crop species were
tomato (Lobb, 1989; Gilreath et al., 200la), newly planted and 1-year-old wine grapes (Al-Khatib et
al., 1993), corn (Ellis et al., 2000, 2003; Hartzler, 2001) and pepper (Gilreath et al., 2000) (Table
C2). The next most sensitive species with EC 10 values between 40 and 70 g a.i./ha were cotton
(Miller et al., 2004), onion (Lobb, 1989), potato (Lobb, 1989) and rice (Ellis et al., 2003), followed
by soybeans with EC 10 of 94 g a.i./ha (Al-Khatib and Peterson, 1999). Rice is sensitive to
glyphosate drift at nearly all stages of growth after emergence. However, as rice development
progresses, the sensitivity to glyphosate also increases. When glyphosate is applied to rice in the
vegetative stage, the crop can overcome much of the injury, if it doesn't die completely. On the other
hand, rice exposed to glyphosate after midseason will likely suffer extreme yield losses, even at drift
rates. Note, as plant development progresses, less glyphosate is required to result in significant yield
reductions.
Glyphosate exhibits favorable environmental characteristics (Franz et al., 1997) and offers
comparable or improved control of both grass and broadleaf weeds with total postemergence
programs as that with standard herbicide programs (Baldwin, 1995; Clay et al., 1995; Culpepper and
York, 1998). Glyphosate has no significant vapor pressure and is unlikely to volatilize. Glyphosate is
strongly sorbed to most soils. These factors make glyphosate an effective and popular tool for weed
135
-------
management.
D.10. Imazapyr
Imazapyr is a non-selective herbicide from the imadazolinone herbicide family and is used to
control annual and perennial grasses, broadleaf herbs, woody species and riparian and emergent
aquatic species in non-crop land. It is used at low rates and has a similar mode of action as the SU
herbicides. It is not highly toxic to birds and mammals but some formulations can cause severe
irreversible eye damage. It is readily absorbed through foliage and roots and, therefore, could be
injurious by drift, runoff or leaching from roots of treated plants. Like the broad-spectrum SU
herbicides, imazapyr kills most nontarget plants exposed at the application rate and can damage plants
at levels too low to detect by standard laboratory procedures. It is persistent in many types of soils.
Potatoes are particularly susceptible to imazapyr at amounts as small as 2% of the normal application
rate (Eberlein and Guttieri, 1994). California uses imazapyr primarily in forest management (CPUR).
Imazapyr usage in Fresno County, CA was minimal in 2000.
D.ll. Imazethapyr
Imazethapyr is a selective herbicide used to control grasses and broadleaf weeds including
barnyardgrass, crabgrass, cocklebur, smartweed, morningglory and others in primarily soybeans, corn
and alfalfa. The compound is used at low rates and is applied preplant-incorporated, preemergence,
at cracking and postemergence. Agricultural use of the compound is highest in the soybean-growing
states including Iowa, Minnesota, South Dakota and Nebraska (Figure D10). Its use on soybeans has
declined from 561,000 kg a.i. in 1997 to 155,000 kg a.i. in 2002 due to shifts in cropping systems
toward genetically-modified soybean and the increased use of glyphosate. Sunflower (Wall, 1996b) is
highly sensitive to drift damage from postemergent applications of imazethapyr. Rice has medium
sensitivity to imazethapyr (Bond et al, 2000) and corn, potato and sorghum are less sensitive to the
herbicide (Table C2).
136
-------
Imazethapyr - Herbicides
Estimated Annual Agricultural Use
Average use of
Active Ingredient
Grains per Hectare
or county per year
d No Estimated Use
D <0004
• 0.004-0.009
D 0010-0.026
• 0.027-0.070
• >=0.071
Crops
soybeans
corn
alfalfa
dry beans
peanuts
green peas
green beans
dry peas
lettuce
Total
Kilograms
Applied
522,622
23,511
12,546
3,364
3,014
aeo
175
47
7
Percent
National Use
92.30
4.15
2.23
0.59
053
015
003
0.01
000
Figure D10. Agricultural use of imazethapyr in the United States for 1997
(USGS, PNSP available at http://ca.water.usgs.gov).
D.12. MCPA
MCPA is a systemic postemergence phenoxy herbicide used to control annual and perennial
weeds in cereals, flax, rice, vines, peas, potatoes, grasslands, forestry applications and on rights-of-
way. This herbicide is often mixed with other products including bromoxynil, 2,4-D, dicamba,
thifensulfuron and tribenuron. MCPA is slightly toxic (i.e., classified as an EPA toxicity class HI), has
low persistence, degrades rapidly in soils and does not cause birth defects in rats. Its use on wheat
accounts for 78% of its domestic applications. The compound is heavily used in the grain states, in
particular Montana, North Dakota and Minnesota (Figure Dl 1). Cotton is sensitive to MCPA drift
(Smith and Wiese, 1972). Both California and Texas have severely restricted the use of MCPA in
order to avoid spray drift damage to nontarget cotton fields. Consequently, MCPA use in California
is negligible (Table 4) because the compound is banned from use between March 16 and October 15.
137
-------
MCPA - Herbicides
Estimated Annual Agricultural Use
Average use of
Active Ingredient
Grams per Hectare
of county per year
CH No Estimated Use
D O.012
D 0.012 - 0.033
D 0.034 - 0.084
D 0.085 - 0.298
• >=0.299
Crops
wheat
barley
pasture
oats
flax
green peas
rye
dry peas
Total
Kilograms
Applied
1,888,244
284,303
90,198
62,461
58,068
28,304
5,175
2,433
1,037
49
Percent
National Use
78.02
11.75
3.73
2.58
2.40
1,17
0.21
0.10
0.04
3 DD
Figure Dll. Agricultural use of MCPA in the United States in 1997 (USGS,
PNSP available at http://ca.water.usgs.gov).
Domestic use of MCPA on winter wheat declined from a peak of 356,000 kg a.i. in 1998 to a low of
109,000 kg a.i. in 2002 (USDA Pesticide Use Data).
D.13. Metribuzin
Metribuzin is a triazinone herbicide used primarily to control germinating and newly emerging
grasses and broadleaf weeds in soybeans, field crops, and potatoes. Metribuzin may be applied
preplant, preemergence, postemergence, or post transplant using ground or aerial equipment (except in
tomatoes and asparagus) as well as chemigation. The compound is slightly toxic to humans, essentially
nontoxic to fish and aquatic invertebrates, and moderately toxic to birds. The compound is used in
every state in the continental U.S., most notably in the Corn Belt and mid-southern states where
soybeans are prevalent (Figure D12). Domestic use of metribuzin on soybeans declined from a peak
of 767,000 kg a.i. in 1997 to 200,000 kg a.i. in 2002 due to the commercial success of RR soybeans
138
-------
Metribuzin - Herbicides
Estimated Annual Agricultural Use
Average use of
Active Ingredient
Grams per Hectare
of county per year
] No Estimated Use
H <0.025
P 0025-0,137
H 0-138-0.389
3 0390-1.035
I >=1.036
Crops
soybeans
potatoes
wheat
alfalfa
sugarcane
corn
tomatoes
asparagus
green peas
fetal
Kilograms
Applied
740,805
173,979
150.032
143,817
140,394
78,146
20.927
10,676
9508
824
Percent
National Use
50.41
11.34
10.21
979
9.55
5.32
142
0.73
065
:i x
Figure D12. Agricultural use of metribuzin in the United States in 1997
(USGS, PNSP available at http://ca.water.usgs.gov).
(USDA Pesticide Use Data). Metribuzin use in California was minimal (i.e., <1% of the domestic
usage) because cotton plants are very sensitive to off-site spray drift of metribuzin (Hurst, 1982). The
majority of metribuzin applications in California are on tomatoes and asparagus in which aerial spraying
is prohibited by the USEPA. Consequently, the potential risk of off-site metribuzin spray drift to
nontarget cotton fields is low in California.
D.14. Metsulfuron-methyl
Similar to chlorsulfuron, metsulfuron-methyl is a selective pre- and post-emergence SU
herbicide used to control some grasses and many broadleaf weeds in wheat and pastures. It is usually
applied as a sole herbicide and is occasionally mixed with other herbicides such as 2,4-D and
139
-------
Metsulfuron - Herbicides
Estimated Annual Agricultural Use
Average use of
Active Ingredient
Grams per Hectare
of county per year
n
n
•
n
No Estimated Use
•=0.002
0 002 - 0.002
0.003-0.013
0.014-0.053
>=0.054
ClCLlS
wheat
pasture
barley
other hay
sod
Total
Kilograms
Applied
10.521
8,880
854
m
i
Percent
National Use
51.44
43.41
4.18
0.97
001
Figure D13. Agricultural use of metsulfuron in the United States in 1997
(USGS, PNSP available at http://ca.water.usgs.gov).
picloram. The compound is widely toxic to terrestrial and aquatic plants and may cause skin and eye
irritation. Several crops such as bean (Boutin et al, 1999), onion (Lobb, 1989), potato (Lobb, 1989)
and tomato (Ray et al, 1996, Lobb, 1989) are susceptible to unintentional exposure to metsulfuron
(Table C2). It is not registered for agricultural use in California and, like chlorsulfuron, is used
primarily in states west of the Mississippi as well as North and South Carolina, Pennsylvania and
Maryland (Figure D13).
D.15. MSMA
Monosodium methanearsonate (MSMA) is an arsenic-based herbicide often mixed with other
herbicides to increase control of emerged broadleaf weeds, grasses, and sedges in a postemergence-
directed application in cotton. The compound is widely used in cotton cropping systems in the Cotton
Belt (Figure D14). It is also used in turfgrass maintenance on golf courses to control annual glass
weeds including crabgrass and bahiagrass. The compound is the most widely used arsenical herbicide
in Florida golf courses (Johnson, 1997). The use of MSMA has been linked to elevated
140
-------
MSMA - Herbicides
Estimated Annual Agricultural Use
Crops
cotton
sod
citrus
Total
Kilograms
Applied
2,106,093
49.780
3,437
Percent
National Use
97, 54
2,31
0.16
Average use of
Active Ingredient
Grams per Hectare
of county per year
n No Estimated Use
D <0.529
• 0.529 - 3 478
n 3.479-11758
n 11.759-40.897
• =-=40.898
Figure D14. Agricultural use of MSMA in the United States for 1997
(USGS, PNSP available at http://ca.water.usgs.gov).
concentrations of arsenic in both soil and groundwater at South Florida golf courses (Ma et al., 2000).
Because of the risk of arsenic contamination in groundwater, the Florida Department of Environmental
Protection recommends that the use of MSMA be severely restricted. MSMA usage in California
accounted for less than 3% of the domestic usage in 1997 (Table 4) and declined to 1.5% in 2002
(Figure D15). Similarly, domestic usage of MSMA on Upland cotton declined from a peak of
2,222,000 kg a.i. in 1997 to 525,000 kg a.i. in 2003 (USDA Pesticide Use Data). Off-site MSMA
spray drift is a greater problem in the mid-southern states in the Cotton Belt where postemergent
applications of MSMA in cotton is more likely to damage soybean plants growing in adjacent fields
than in California.
D.16. Nicosulfuron and Primisulfuron-methyl
Nicosulfuron and primisulfuron-methyl are members of the SU family of herbicides and are
used to control weeds in corn. Nicosulfuron is applied postemergence to control weeds such as
141
-------
Q- o
1992
1994
1996
1998
2000
2002
Year
Figure D15. MSMA usage in kg a.i. applied in California
for 1991-2002 (CPUR).
Johnsongrass, quackgrass, foxtails, shattercane, panicums, barnyardgrass, and morningglory.
Primisulfuron is applied postemergence to control grassy and broadleaf weeds. These SU compounds
may be found in mixes with other herbicides such as 2,4-D, dicamba, cyanazine, bromoxynil and
atrazine. The toxicity of these SU herbicides to humans, animals, birds and fish is minimal. The main
concern is the potential damage to soybeans which are particularly sensitive to nicosulfuron (Al-Khatib
and Peterson, 1999) and primisulfuron (Bailey and Kapusta, 1993) at drift levels. Soybeans are
planted later than corn so postemergence application of these SU herbicides to corn may occur when
soybean is at an early growth stage (Young et al., 2003). Off-site drift damage to soybean from these
SU herbicides is a greater concern in the Corn Belt than in California because agricultural use is
considerably higher (Figure D16) and more soybean fields are adjacent to corn fields. Nicosulfuron
use on corn fluctuated between 64,000 kg a.i. and 113,000 kg a.i. in the years 1992-2000 while
primisulfuron use on corn has steadily increased from 14,000 kg a.i. to 64,000 kg a.i. in 2000 (USDA
Pesticide Use Data). About 14% of the corn acreage was treated with nicosulfuron in 2001.
Agricultural use of nicosulfuron of 691 kg a.i. in California in 1997 represents less than 1% of the
domestic total use. Primisulfuron is not registered for use in California. When compared with higher
use-rate conventional herbicides, the application rates producing an effect in sensitive plants are often
142
-------
Nicosulfuron - Herbicides
Estimated Annual Agricultural Use
Crops
corn
sweet corn
Total
Kilograms
Applied
95,066
750
Percent
National Use
9922
078
Average use of
Active Ingredient
Grams per Hectare
of county per year
Q| No Estimated Use
D <0.007
• 0.007 - 0 032
D 0.033-0.107
D 0.108-0.389
• >=0390
Figure D16. Agricultural use of nicosulfuron in the United States in 1997
(USGS, PNSP available at http://ca.water.usgs.gov).
lower for SU herbicides; however, when expressed as a percentage of the labeled application rate, the
SU herbicides were often not different from other herbicides (Wall, 1994a; Derksen, 1989; Bhatti et
al., 1995; Al-Khatib et al., 1993). This is an important consideration for drift effects because the
response of plants will be dependent upon the volume of spray mixture moving off-target. As a class,
the SU herbicides should not be expected to present more of a nontarget risk than many other
herbicides.
D.17. Paraquat dichloride
Paraquat is a Restricted Use Pesticide (RUP) because the compound is highly toxic to humans
via ingestion. Paraquat has low mammalian toxicity via dermal route, is strongly adsorbed in clay soils,
is essentially environmentally benign but has high occupational risk to agricultural workers. The
compound is damaging to the lungs, eyes, skin, nose, fingernails and toenails and may result in serious
143
-------
illness when inhaled. Paraquat controls broadleaf weeds and grasses and is found in many
formulations with other herbicides including 2,4-D, simazine, sodium chlorate and diquat dibromide for
expanded coverage. The compound is the most widely used herbicide in the developing world and is
also widely used in California on cotton, alfalfa, fruits and nuts. The compound is also registered for
use as a crop dessicant and defoliant and as an aquatic herbicide. Paraquat use of 413,000 kg a.i. in
California in 1997 represents about 13% of the domestic use. Paraquat ranks second in agricultural
usage in Fresno Co, CA among the high potential risk herbicides listed in Table 4. The compound is a
commonly used herbicide in cropping systems in the Corn Belt, the northern grain states, Oregon,
Washington and the cotton states (Figure D17).
Paraquat is a quaternary nitrogen herbicide that destroys green plant tissue on contact and by
translocation within the plant. The compound disrupts the intracellular electron transfer systems,
Paraquat - Herbicides
Estimated Annual Agricultural Use
Average use of
Active Ingredient
Grams per Hectare
of county per year
^J No Estimated Use
ID <0.012
• 0.012 • 0.065
HI 0.066 - 0.224
Z3 0.225 - 0.660
• >=0,661
Crops
corn
soybeans
cotton
faliowland
wheat
alfalfa
grapes
almonds
applies
peanuts
Total
Kilograms
Applied
901,391
627,308
472,916
185,734
185,527
16' 4C9
93,303
71,040
68.643
35,351
Percent
National Use
29.17
20.30
15,30
6.01
6.00
5.22
3.02
2.30
2.22
1,14
Figure D17. Agricultural use of paraquat in the United States in 1997 (USGS,
PNSP available at http://ca.water.usgs.gov).
144
-------
inhibiting the reduction of an enzyme during photosynthesis. This will result in the accumulation of
superoxide radicals which cause destruction of lipid cell membranes of the leaves. Symptoms of
paraquat drift include brown spots on leaves of nontarget plants upon contact, with little or no long
lasting effects.
D.18. Picloram
Picloram is a systemic herbicide used for control of woody plants and broadleaf weeds on
rangelands and pastures in the western U.S. excluding California, Arizona and New Mexico (Figure
D18). Most grasses are resistant to picloram. The compound is listed as a RUP by the USEPA due
to its mobility in water and the extreme sensitivity of many important crops to the herbicide. Crops
susceptible to damage from exposure to picloram include field bean (Smith and Geronimo, 1984),
cotton (Smith and Wiese, 1972, Smith and Geronimo, 1984), wine grape (Lobb and Woon, 1983),
grape (Smith and Geronimo, 1984), kiwifruit (Lobb and Woon, 1983), peanut (Smith and Geronimo,
1984), potato (Lobb and Woon, 1983), soybean (Smith and Geronimo, 1984), tobacco (Smith and
Picloram - Herbicides
Estimated Annual Agricultural Use
Average use of
Active Ingredient
Grams per Hectare
of county per year
D
n
No Estimated Us
•=0007
0.007-0.025
D 0.026-0.100
O 0.101-0.361
• >=0 362
Crops
pasture
other hay
fallowland
wheat
barley
oats
Total
Kilograms
Applied
537,608
30.C76
10,976
9,945
1,620
178
Percent
National Use
90.97
5.18
1.86
1.68
0.27
0.03
Figure D18. Agricultural use of picloram in the United States in 1997 (USGS,
PNSP available at http://ca.water.usgs.gov).
145
-------
Geronimo, 1984) and tomato (Smith and Geronimo, 1984).
D.19. Propanil
Propanil is an anilide foliar active herbicide for the selective postemergent control of broadleaf
weeds and grass in rice, barley, oats and spring wheat. Propanil provides control of a broad spectrum
of ricefield weeds, including smallfiower umbrella sedge and ricefield bulrush. The compound has low
human, avian, mammalian and aquatic toxicity but has high occupational risk to workers handling and
applying the herbicide. It was the most commonly used herbicide in rice with 62% of the reported rice
acreage in the U.S. being treated (USDA, 2001), primarily in the Mississippi Delta region and the rice-
producing counties of Texas and California (Figure D19). The chemical is also registered for use on
barley, oats and spring wheat in Minnesota, Montana and North Dakota but no propanil sales and use
were reported in these states in 1997 by the USDA NASS. California produces approximately 25%
Propanil - Herbicides
Estimated Annual Agricultural Use
Average use of
Active Ingredient
Crams per Hectare
of county per year
] No Estimated Use
] <15.170
I 15.170-35.872
] 35.873-111511
I 111.512-311851
I >=311.852
Figure D19. Agricultural use of propanil in the United States for 1997
(USGS, PNSP available at http://ca.water.usgs.gov).
146
-------
of the rice grown in the United States and represents about 17% of the national propanil use on rice
(Table 4). In the Central Valley of California, propanil is a state limited-use herbicide and can be
applied 1.6 km or more from a cotton field and 4.8 km from prunes, pistachios, and grapevines
(California Code of Regulations Title 3 Section 6400). Propanil use in Fresno County, CA is
negligible because very few hectares of cropland are used for rice production.
Cotton (Hurst, 1982) and soybean (Bode and McWhorter, 1977) are sensitive to propanil
drift. The vast majority of chemicals are applied on rice at the early plant development stages prior to
grain exposure so that the applied chemicals have sufficient time to degrade or dissipate prior to
harvest. Chemicals are frequently applied on rice in late tillering stage, which typically corresponds to
the early reproductive stages of cotton (Snipes et al., 1991). In addition, most rice herbicides are
applied by fixed-wing aircraft (Smith and Shaw, 1966; Snipes et al., 1991). Consequently, the
likelihood of damage to susceptible crops, in particular cotton, due to off-spray drift of propanil may
be high in Arkansas, Louisiana, Missouri, Mississippi, Texas and California. The drift problem may be
greater in the mid-southern states than in California because of the greater acreage of soybean and
cotton fields that are in close proximity to rice fields.
D.20. Prosulfuron
A 1:1 mixture of two SU herbicides, prosulfuron and primisulfuron, was registered in 1995 and
marketed as "Exceed®" for postemergent control of broadleaf weeds in corn. Exceed® is often mixed
with other herbicides such as dicamba and 2,4-D to provide a broader spectrum of weed control and
to slow down the development of herbicide-resistant weed populations. Soybeans are extremely
sensitive to prosulfuron (Al-Khatib and Peterson, 1999) and are susceptible to carryover injury in the
following spring in areas with high pH soils. Prosulfuron is much more persistent as pH reaches levels
> 7.2 and remains active in the following growing season in some types of soils. Annual domestic use
of prosulfuron in 1997 was about 33,000 kg a.i., primarily on corn in Nebraska, Minnesota, Iowa,
Illinois and Indiana (Figure D20). Prosulfuron usage on corn has declined to 11,000 kg a.i. by 2000
(USDA Pesticide Use Data). About 1360 kg a.i. was applied on corn in Illinois in 2001 (Illinois
Agricultural Statistics Service). The compound is not registered for use in California.
147
-------
Prosulfuron - Herbicides
Estimated Annual Agricultural Use
D
D
Average use of
Active Ingredient
Grams per Hectare
of county per year
No Estimated Use
O.002
0.002 - 0.009
0010-0.035
0036-0.219
>=0220
Crops
corn
sorghum
wheat
Total
Kilograms
Applied
29,608
2,965
495
Percent
National Use
89.54
3.97
1.50
Figure D20. Agricultural use of prosulfuron in the United States in 1997
(USGS, PNSP available at http://ca.water.usgs.gov).
D.21. Quinclorac
Quinclorac belongs to a new class of highly selective auxin herbicides and is used to control
dicotyledonous and monocotyledonous weeds, particularly barnyardgrass, in rice (Grossmann, 1998).
The compound may not be applied by air in certain counties of Idaho, Oregon, and Washington to
avoid off-site spray drift damage to vegetation. Quinclorac-resistant barnyardgrass has been observed
in Louisiana and Mississippi. Quinclorac is not effective against sprangletop, a problem weed in
Missouri. Many other rice herbicides are more cost-effective and provide a broader range of weed
control than quinclorac. Consequently, agricultural use of quinclorac in the United States is low and is
mainly in the Mississippi Delta region (Figure D21).
D.22. Rimsulfuron
Rimsulfuron is a selective post-emergent SU herbicide used to control annual grasses and
148
-------
Quinclorac - Herbicides
Estimated Annual Agricultural Use
Crops
Total
Kilograms
Applied
Percent
National Use
rcu
130,071
1DDDC
Average use of
Active Ingredient
Grams per Hectare
of county per year
Q No Estimated Use
n <0585
D 0.585-1.984
D 1985-4312
D 4,313-13892
• >= 13.893
Figure D21. Agricultural use of quinclorac in the United States in 1997
(USGS, PNSP available at http://ca.water.usgs.gov).
broadleaf weeds in corn, sorghum and wheat. Rimsulfuron can be more injurious to corn and has less
control on larger grasses compared to other corn herbicides. In corn, broadleaf weeds are easier to
control than grasses so the timing of post-emergent herbicide application is determined by grass size.
The best corn herbicide to use depends upon its degree of residual activity in relation to weed
pressure, relative size of the corn and weeds, and planting date. Products containing rimsulfuron can
provide several weeks of control of annual grasses and broadleaf weeds versus little or no residual
activity in other post-emergent corn herbicides. Annual domestic agricultural use was about 9,000 kg
a.i. in 1997, primarily on corn in the Corn Belt (Figure D22). Rimsulfuron use on corn increased from
7,300 kg a.i. in 1997 to 37,200 kg a.i. in 2000 (USDA Pesticide Use Data). Similarly, rimsulfuron
use in California increased from 119 kg a.i. in 1997 to 807 kg a.i. in 2002 (CPUR). The compound is
very active even at low dosages and soybeans are sensitive to rimsulfuron drift.
149
-------
Rimsulfuron - Herbicides
Estimated Annual Agricultural Use
Average use of
Active Ingredient
Grams per Hectare
of county per year
n
n
No Estimated Use
<0.002
0.002-0.011
0,012-0.044
0,045 - 0.077
>=0 073
Crops
corn
potatoes
tomatoes
Total
Kilograms
Applied
7,458
1,219
188
Percent
National Use
34 13
13.75
2.12
Figure D22. Agricultural use of rimsulfuron in the United States in 1997
(USGS, PNSP available at http://ca.water.usgs.gov).
D.23. Sethoxydim
Sethoxydim is a selective postemergence herbicide used to control annual and perennial grass
weeds in primarily soybeans. The compound is of low toxicity to birds, mammals and aquatic animals
but is water-soluble (4,000 ppm) and can be highly mobile. Sethoxydim is most effective against reed
canarygrass (Phalaris arundinacea) in Illinois but less effective against quackgrass (Elytrigia repens)
than fiuazifop-p-butyl in Oregon. Sethoxydim is used in all states in the continental U.S., less in
California and more in the Corn Belt and the mid-southern states (Figure D23). Sethoxydim use in
California peaked at 29,227 kg a.i. in 1998 and has steadily declined to 11,142 kg ai. in 2002 (Figure
D24). Similarly, domestic use of the compound on soybeans declined from a peak of 525,000 kg a.i.
in 1996 to 209,000 Ib ai. in 2002 (USDA Pesticide Use Data) due to the commercial success of RR
150
-------
Sethoxydim - Herbicides
Estimated Annual Agricultural Use
Average use of
Active Ingredient
Grams per Hectare
of county per year
H No Estimated Use
I] <0.007
J 0.007 - 0.030
H 0.031 - 0.091
H 0.092-0215
• =-=0.216
Crops
soybeans
alfalfa
cotton
sugarbeets
peanuts
sunflowers
dry beans
mint
citrus
potatoes
Total
Kilograms
Aool ed
480,175
60,271
58,846
35,169
25.110
15,590
12,680
10,052
6,750
5,205
Percent
National Use
64 44
8.09
7,90
4,72
3.37
2.09
1.70
1.35
0.91
C73
Figure D23. Agricultural use of sethoxydim in the United States in 1997
(USGS, PNSP available at http://ca.water.usgs.gov).
1994
1996
Year
1998
2000
2002
Figure D24. Sethoxydim usage in kg a.i. applied in
California in 1991-2002 (CPUR).
151
-------
soybeans. Corn (Al-Khatib et al, 2000) and sorghum (Al-Khatib et al., 2003) are most likely to be
impacted by off-site spray drift of sethoxydim. Consequently, the likelihood of damage from
sethoxydim spray drift is minimal in California and high in the Corn Belt and mid-northern states where
soybean, corn and sorghum fields are in close proximity.
D.24. Sulfometuron-methyl
Sulfometuron-methyl is a broad-spectrum SU herbicide used to control annual and perennial
grasses and broadleaf weeds in non-crop land. It is also used in forestry applications to control
woody tree species. Sulfometuron is one of the most potent of the SU family and is marketed under
the brand name Oust®. The Oust® label prohibits using equipment that previously has been used to
apply Oust® to apply any other herbicide because "low rates of OUST® XP can kill or severely injure
most crops" (E.I. du Pont de Nemours and Company, 2000-2002). It is applied either pre- or post-
emergent. The compound has low toxicity to humans, animals, birds and fish but is an eye irritant. It
has the potential to contaminate groundwater because it is relatively mobile in soil, is moderately
persistent and has a high intrinsic leaching potential. Sulfometuron is active in the soil and is usually
absorbed from the soil by plants. Off-target movement of Oust® may have caused significant crop
damage in several large-scale drift instances in Idaho and Washington. In Idaho, reported damage of
several million dollars worth of crops over 40,500 ha allegedly was caused by wind transport from an
aerial Oust® application by the Bureau of Land Management to kill cheatgrass following a wildfire
(Idaho Dept of Agriculture, 2002). Another well-documented, large-scale drift event occurred in
1985 when Oust® was applied to over 1,126 km of roadsides by county and state road crews in
Franklin County, Washington. Off-site drift of Oust® caused over a million dollars of damage including
damage to over 300,000 young trees in one nursery (Turner, 1987). Both of these drift incidences
were the result of wind-borne soil particles containing Oust® rather than the movement of liquid
particles of the herbicide in the air at the time of application.
D.25. Thifensulfuron-methyl and Tribenuron-methyl
Thifensulfuron-methyl, alone or in combination with tribenuron-methyl, has been the most
commonly studied of the SU herbicides because of its widespread use to control annual and perennial
152
-------
weeds in small grain crops in areas where sensitive crops are also grown (Obrigawitch et al., 1998).
Thifensulfuron and tribenuron contain the new SU a.i. DPX-L5300. This compound has low toxicity to
humans, animals, and beneficial insects and is not an eye or skin irritant. Many crops such as alfalfa (Al-
Khatib et al, 1992a), canola (Wall et al., 1995), wine grape (Al-Khatib et al., 1993), pea (Al-Khatib
and Tamhane, 1999) and potato (Lawson et al., 1992) are susceptible to damage from exposure to
thifensulfuron/tribenuron at drift levels (Table C2). Grape vineyards close to wheat fields may be
damaged by off-site drift from a spring application (Al-Khatib et al., 1993). In general, foliar
applications of a SU herbicide at or shortly before flowering caused greater reductions in yield than at
other growth stages (Obrigawitch et al., 1998). Consequently, a number of states have severely
restricted or banned their use due to their potential to damage economically-important crops. For
example, thifensulfuron and tribenuron are not registered for use in California. The New York State
Thifensulfuron - Herbicides
Estimated Annual Agricultural Use
Average use of
Active Ingredient
Grams per Hectare
of county per year
n
No Estimated Use
<0.002
0.002 - 0.004
0.005 - 0.021
0.022 - 0.054
>=0.055
Crops
wheat
soybeans
barley
com
oats
Total
Kilograms
Applied
32,769
7,233
4,272
2,931
349
Percent
National Use
6334
1530
8.98
6 15
073
Figure D25. Agricultural use of thifensulfuron in the United States in 1997
(USGS, PNSP available at http://ca.water.usgs.gov).
153
-------
Department of Environmental Conservation approved the registration of thifensulfuron/tribenuron for
ground application only to wheat and barley crops in April 1993; an amended registration for aerial
application to wheat and barley crops was approved in May 1996. Agricultural use of thifensulfuron is
highest in North Dakota, Illinois, Indiana, Arkansas, Tennessee and Oklahoma (Figure D25). Use of
tribenuron mirrors that of thifensulfuron. In Illinois, Common Waterhemp (Amaranthus rudis) infests >
800,000 hectares, is found primarily in corn and soybean, and is resistant to several widely used
herbicides such as thifensulfuron and atrazine.
D.26. Triclopyr
Triclopyr is a selective systemic herbicide used to control unwanted woody and herbaceous
weeds in pastures, rangelands, forests and rice. The compound maybe found in formulations with other
herbicides such as 2,4-D and clopyralid. The ester form of triclopyr is highly toxic to fish but, under
normal conditions, rapidly breaks down in water to a less toxic form. Triclopyr is a RUP because many
crop species are susceptible to damage when exposed to the compound (Table Al). In California, the
butoxyethyl ester formulation of triclopyr is used in landscape management, rights-of-ways and forests
while the triethylamine salt formulation is used primarily on rice. Combined, triclopyr use in California of
60,000 kg a.i. in 1997 represents about 23% of the domestic total. Agricultural use of the compound is
also high in Oregon, Washington, Texas, Oklahoma and the rice-producing states in the Mississippi
Delta (Figure D26). Triclopyr use in Fresno County is minimal because very little cropland is used for
rice production.
The triethylamine salt (found in Garlon 3A) is a skin irritant and a severe eye irritant. Over the
summer of 1995, a number of residents in timber lands near Comptche, CA developed mysterious flu-
like symptoms that were suspected to be caused by the use of Garlon (Coast News Service Jan 6,
1996). These affected people were in areas that were bordering harvested timber areas where there
was a major clear cut, followed by a spraying of Garlon, and more recently was burned.
154
-------
Triclopyr - Herbicides
Estimated Annual Agricultural Use
Average use of
Active Ingredient
Grams per Hectare
of county per year
£] No Estimated Use
D <0.065
Q 0.065-0.196
D 0.197-0.555
D 0.556-1.289
• >=1.290
Crops
pasture
rice
other hay
Total
Kilograms
Applied
180,892
77,981
4,903
Percent
National Use
68.58
29.56
1.86
Figure D26. Agricultural use of triclopyr in the United States in 1997 (USGS,
PNSP available at http://ca.water.usgs.gov).
155
-------
Appendix E
Crop Planting and Harvest Dates in Fresno County, CA
Table El. Relevant dates of importance for crops in Fresno County, CA.
Crop species
Alfalfa
Dried beans
Corn
Planting dates
(USDA, 1997)
Spring or fall,
perennial
04/02-07/02
03/16-07/16
04/02-07/02
most active
Harvest dates
(USDA,
1997)
Multiple
cuttings from
03/02-11/06
08/02-11/16
09/02-12/02
Time of
increased
sensitivity
03/04-
07/29
04/15-
07/29
04/02-
09/21
correspon
ds to V5
-VT
stages
Comments
Yield loss at immediate and
subsequent harvests in year o
exposure. (Al-Khalib et al,
1992a).
Most sensitive to metsulfuron
at cotyledon to flower bud
initiation stages (Boutin et al.
1999).
Greater yield losses at 4-9
leaf stage or tasseling stages
(Swanton et al., 1996;
Rowland et al., 1999; Ellis et
al, 2000, 2003; Hartzler,
2001 web).
156
-------
Cotton
Grapes, table
04/01-05/16
(SJV weekly
reports)
03/24-04/22
(bud break based
on weekly crop
reports for
Fresno Co)
10/10-11/24
(60% open
boll 165
DAE)
05/26-12/15
(based on
weekly crop
reports for
Fresno Co)
05/23-
07/29
correspon
ds to 1st
square to
1st bloom
03/23-
06/06
Yield reductions at all rates al
early square to full bloom
stages, greatest damage at
time of flowering (Miller et a
1963; Wiese and Hudspeth,
1968; Jeffery et al, 1972;
Smith and Wiese, 1972;
Hurst, 1977, 1982; Hamilton
and Arle 1979; Jacoby et al.,
1990; Snipes et al., 1991;
Lanini, 1999; Miller et al,
2004).
Greatest damage when
sprayed with 2,4-D at early
cluster development (Weaver
et al., 1961). Grapevines
sensitive to phenoxy
herbicides from early April to
early June (Hellman, TX
A&M).
157
-------
Grapes, wine
Onion
Sorghum
Sugarbeets
03/24-04/22
(bud break based
on weekly crop
reports for
Fresno Co)
Transplanted or
direct seeded
04/03-06/01
(Kansas State
Univ)
Sugarbeets are
planted
throughout the
year in California.
09/08-12/15
(based on
weekly crop
reports for
Fresno Co)
05/21-07/01
(transplanted
short-
intermediate
varieties)
07/01-10/01
(direct
seeded)
07/13-09/11
(maturity)
(KSU)
07/02-12/11
(spring
harvest)
04/02-08/16
(fall harvest)
03/23-
06/06
03/18-
07/29
04/15-
07/29
03/05-
07/29
Greatest damage when
sprayed with 2,4-D at early
cluster development (Weaver
etal, 1961). Grapevines
sensitive to phenoxy
herbicides from early April to
early June (Hellman, TX
A&M).
2,4-D at 20 cm tall caused
significant damage (Hemphil
and Montgomery, 1981).
Metsulfuron at early bulbing
caused significant damage
(Lobb, 1989).
Imazethapyr and sethoxydim
caused significant damage
when applied 25 DAP (Al
Khaittib et al, 2003).
Sugarbeets are most sensitive
to 2,4-D and other herbicides
at 6 WAP.
158
-------
Tomatoes, fresh
Tomatoes,
processed
Tomatoes are
planted
throughout the
year in California.
Tomatoes are
planted
throughout the
year in California.
05/16-12/31
06/21-11/11
04/29-
07/29
04/29-
07/29
Tomato plants were damaged
by 2,4-D and dicamba at
prebloom stage (Jordan and
Romanowski, 1974).
Tomato plants were damaged
by 2,4-D and dicamba at
prebloom stage (Jordan and
Romanowski, 1974).
159
------- |