I MONSANTO MONSANTO COMPANY 1300 1 (Eye) Street, NW Suite 450 East Washington. DC. 20005 http://wwmonsontocom May 31, 2016 Philip W. Miller Vice President, Global Regulatory and GovernmentAffairs Docket No. Dicamba: New Use on Herbicide-Tolerant Cotton and Soybeans Environmental Protection Agency Mailcode 28221 1200 Ave, NW Washington, DC 20460 Attention: Jack Housenger Dear Mr. Housenger: Please find enclosed Monsanto? comment to Docket ID No. in support of Dicamba: New Use/on Herbicide-Tolerant Cotton and Soybeans. Sincerely, CLO Philip W. Millr ,.Ph D. Vice President, Global Regulatory and Government Affairs COMMENTS OF MONSANTO COMPANY ON EPA’S PROPOSED REGISTRATION OF DICAMBA FOR USE ON DICAMBA-TOLERANT COTTON AND SOYBEAN EPA-HQ-OPP-2016-0127-0016 Submitted by: Monsanto Company 800 North Lindbergh Blvd. St. Louis, MO 63167 TABLE OF CONTENTS Page I. INTRODUCTION AND EXECUTIVE SUMMARY.........................................................1 II. THE NEW DICAMBA USE WILL PROVIDE SIGNIFICANT BENEFITS TO U.S. AGRICULTURE AND THE ENVIRONMENT ........................................................2 III. A. Studies Confirm Dicamba Has Effective Pre-Emergent Control ............................3 B. Approval of Dicamba for Use on Dicamba-Tolerant Soybean and Cotton Will Further Facilitate No-Till Farming, Which Provides Significant Environmental Benefits ...........................................................................................4 C. Approval of Dicamba for Use On Dicamba-Tolerant Soybean and Cotton Will Provide Additional Benefits to Farmers ..........................................................5 A DOWNWIND BUFFER OF 110 FEET IS MORE THAN SUFFICIENT TO PROTECT ENDANGERED SPECIES AND NON-TARGET ORGANISMS ..................8 A. A Four-Sided Buffer Would Harm Farmers and Agriculture ..................................9 B. A Buffer Is Not Needed to Protect Against Volatilization of the Monsanto Dicamba Products. .................................................................................................10 1. Laboratory (Humidome) Relative Volatility Study Showed Low Volatility Levels, Particularly Compared to Banvel ..................................11 2. EPA Guideline Field Volatility Studies Conducted Under Varied Environmental Conditions Provide Additional Data for the Assessment of Off-Site Movement ............................................................12 3. Modeling of Off-Target Movement Due to Volatility Demonstrates That a Buffer Is Not Necessary to Protect Against Volatilization of the Monsanto Dicamba Products ...............................................................13 4. Any Effects of the Monsanto Dicamba Formulations From Off-Site Movement Are Limited to Downwind Particle Drift at the Time of Application .................................................................................................16 C. The Available Data on Particle Drift Support That a Downwind Buffer of No More Than 110-Feet at the Time of Application Is More Than Sufficiently Protective ...........................................................................................17 D. Information from Product Development Research and Extraneous Incidents Do Not Undermine the Sufficiency of a 110-Foot Downwind Buffer for Particle Drift, and No Buffer for Volatility ..........................................18 i E. IV. V. VI. 1. M1691 Herbicide Was Not Used Alone in the Product Development Research ..............................................................................18 2. EPA’s Analysis of Field Trial Data Did Not Consider Dose Response ....................................................................................................18 3. Reported Impacts From Specific Sites Are Due to Uses That Are Inconsistent With Current Proposed Label Conditions .............................23 Other Lines of Evidence Also Support the Sufficiency of 110-Foot Downwind Buffer For Particle Drift Only .............................................................25 1. The Published Literature Does Not Show High Vapor Drift for Dicamba Resulting in Non-Target Plant Injury .........................................25 2. The Buffer Distance for Vapor Movement Estimated From Egan and Mortensen Is Incorrect ........................................................................26 EPA’S RISK ASSESSMENTS ARE SCIENTIFICALLY-SOUND AND SUPPORTED BY THE LAW ...........................................................................................27 A. EPA’s Approach to the Risk Assessments Is Scientifically-Sound and Well-Supported ......................................................................................................27 B. EPA’s Authority for a Mitigated No-Effect Finding Is Well-Established .............29 C. Dicamba Does Not Pose a Risk to Terrestrial Invertebrates ..................................30 D. EPA Properly Considered Runoff in Its No Effect Rationale for Listed Species ...................................................................................................................31 E. Dicamba Acid Seedling Emergence Endpoints Should Not Have Been Considered in the Dicamba DGA Salt Risk Assessment .......................................31 MONSANTO SUPPORTS EPA’S WEED RESISTANCE PLAN, WITH CERTAIN REFINEMENTS ..............................................................................................33 A. Monsanto Suggests Adjustments That Will Help to Reinforce Practices That Assist Growers With Herbicide Resistance Management .............................34 B. Monsanto Product Recommendations Will Maintain the Durability of M1691 Herbicide ...................................................................................................36 ENABLING TANK MIXES IS CRITICAL FOR EFFECTIVE WEED RESISTANCE MANAGEMENT AND ENVIRONMENTAL STEWARDSHIP ...........37 A. Tank Mixing Provides Multiple Benefits to Growers............................................37 B. EPA Should Consider Specific Tank Mixes Where Requested.............................38 ii VII. EPA’S ANALYSIS IS TECHNICALLY AND SCIENTIFICALLY SOUND ................39 A. EPA’s Human Health Risk Assessment Is Highly Conservative and Amply Demonstrates a Reasonable Certainty of No Harm ...................................39 B. EPA’s Occupational Risk Assessment Demonstrates Ample Protection From Occupational Exposure ................................................................................40 C. The Proposed Label Is Protective of Non-Target Susceptible Plants ....................41 VIII. REFERENCES ..................................................................................................................42 IX. APPENDIX: TECHNICAL CORRECTIONS ..................................................................48 iii I. INTRODUCTION AND EXECUTIVE SUMMARY Monsanto is pleased to have the opportunity to comment on EPA’s proposed approval of M1691 Herbicide, the new use of dicamba in the Roundup Ready® Xtend Crop System for soybean and cotton. This comment discusses the following products (collectively referred to as “the Monsanto Dicamba Products”): (1) M1691 Herbicide (Clarity®)1; (2) M1768 Herbicide (XtendimaxTM with VaporGripTM Technology)2; and (3) M1769 Herbicide (Roundup® XtendTM with VaporGripTM Technology).3 The assessment that EPA has conducted for M1691 Herbicide further supports EPA’s assessment for the in-crop use of Monsanto’s M1768 and M1769 Herbicides containing the DGA salt of Dicamba. Monsanto’s comments make the following key points: First, the new use of dicamba will provide important and wide-ranging benefits to cotton and soybean production in the United States. Dicamba use on dicamba-tolerant soybean and cotton will (1) help further promote the sustainability of crop production in the U. S.; (2) provide more weed control options for farmers; (3) slow the development of resistant herbicide genotypes for other herbicide mechanisms-of-action that are currently registered for use in soybean and cotton; and (4) further facilitate the adoption of no-till and conservational tillage practices. Second, a downwind buffer of 110 feet (for 0.5 lb a.e./A) would be more than sufficient to protect non-target organisms and threatened or endangered species from any effects of dicamba. Multiple studies show that no buffer whatsoever is necessary to be protective of volatilization, while available data support that a 110-foot downwind buffer at the time of application is more than amply protective of the effects of particle drift. Third, EPA’s risk assessments are scientifically sound and supported by a large body of law. EPA’s registration of M1691 Herbicide for dicamba-tolerant soybean and cotton is based on extensive toxicity and ecological effects evaluations. As part of this process, EPA requires data on environmental fate, and ecological effects on non-target terrestrial and aquatic animals, and non-target plants. EPA also specifically focuses on potential effects a pesticide may have to listed species by conducting thorough ecological risk assessments. EPA’s Environmental Fate and Effects Division (EFED) conducted EPA’s assessments of both ecological effects and exposure, to evaluate whether M1691 Herbicide may affect listed species. Where, based on its ecological risk assessment, EPA concluded that M1691 Herbicide would have “no effect” on listed species, it ended its endangered species assessment. Where EPA concluded, instead, that M1691 Herbicide “may affect” but was “not likely to adversely affect,” a species, it conducted further analysis and informal consultation with the U.S. Fish & Wildlife Service. Where EPA concluded that M1691 Herbicide “may affect” and was “likely to adversely affect” a species, EPA imposed a label restriction preventing dicamba from being applied in that county. Additionally, EPA’s authority for a mitigated “no effect” finding is well-established. In relying upon maximum use scenarios, EPA’s no effect determinations already take into account multiple 1 Clarity® is a dicamba diglycolamine (DGA) salt containing formulation. 2 Xtendimax™ with VaporGrip™ Technology Herbicide is a dicamba DGA salt containing formulation. 3 Roundup® Xtend with VaporGrip™ is a dicamba DGA salt and glyphosate ethanolamine salt containing formulation. 1 M1691 Herbicide label mandates. The conditions on the FIFRA label mandated by EPA are legally enforceable under FIFRA and become part of the proposed registration under the Endangered Species Act. Critically, EFED’s screening-level assessments already rely upon the label’s mandatory conditions as part of the effects determination. EPA also may take into account other mandatory conditions on pesticide labels, such as mandatory appropriate downwind buffer zones. This is entirely appropriate given that failure to follow the label requirements is unlawful. Absent reliance on label application rates and frequency, it would be impossible for EPA to assess potential exposure – and thus impossible for EPA to make either a “no effect” or a “may affect” determination. Fourth, Monsanto supports a Weed Resistance Plan, with certain refinements. Monsanto greatly values maintaining the durability of our products and agrees that the benefits of tools such as M1691 Herbicide should be preserved for agriculture. Sustainability of this tool is important for American farmers and a business objective for Monsanto. As such, Monsanto supports voluntary stewardship, which enables the marketplace to adjust to emerging challenges around herbicide resistance. Fifth, EPA should consider tank mixes where specifically requested, as tank mixing is critical for weed resistance management and environmental stewardship. Tank mixing also provides economic, sustainability, efficacy and productivity benefits to U.S. agriculture and the environment. Lastly, EPA’s analysis is technically and scientifically sound. EPA’s Human Health Risk Assessment is highly conservative and demonstrates a reasonable certainty of no harm, and EPA’s Occupational Risk Assessment demonstrates that the proposed label will provide ample protection from occupational exposure. The proposed label also is protective of non-target susceptible plants. Technical corrections are included in an appendix to the comments. II. THE NEW DICAMBA USE WILL PROVIDE SIGNIFICANT BENEFITS TO U.S. AGRICULTURE AND THE ENVIRONMENT The new use of dicamba in the Roundup Ready® Xtend Crop System for soybean and cotton provides a new management tool for overall weed management, including the management of herbicide-resistant weeds. Specifically, dicamba use on dicamba-tolerant soybean and cotton will help further promote the sustainability of crop production in the United States. As discussed below, this new use of dicamba will offer a valuable weed management tool, provide more weed control options for farmers, and slow the development of resistant herbicide genotypes for other herbicide mechanisms-of-action (i.e., ALS, PPO, HPPD, glufosinate, glyphosate) that are currently registered for use in soybean and cotton, while also supporting the adoption of no-till and conservational tillage practices. Existing herbicides present challenges to growers’ ability to consistently control problematic weeds. As of April 2013, 400 herbicide-resistant weed biotypes have been reported to be resistant to 21 different herbicide modes-of-action worldwide (Heap, 2013). Dicamba-resistant and glufosinate-resistant weeds, however, account for only <1% and 0.5% of resistant biotypes, 2 respectively (Heap, 2013). Glyphosate-resistant weeds account for approximately 6% of the herbicide-resistant biotypes. Weeds resistant to herbicides that inhibit acetolactate synthase (ALS) account for 32% of the herbicide-resistant biotypes. Specifically, ALS-herbicide resistance is present in most of the major broadleaf weed species commonly found in cotton and soybeans. In several geographies—especially the mid-South—the invasive and aggressive weed Palmer amaranth, which is one of the worst pests for cotton and soybean producers, has become resistant to PPO and ALS inhibitors. As a result, PPO and ALS inhibitors standing alone often are not effective weed control options for many soybean and cotton farmers today. The widespread occurrence of ALS- and glyphosate-resistant Palmer amaranth has led to higher use of PPO-inhibiting herbicides on soybean and cotton. As a result, Palmer amaranth in the U.S. mid-South has become resistant to PPO inhibitors. Studies conducted in 2016 have confirmed resistance to foliar applied PPO inhibitors (Salas et al., 2016). Additionally, Larry Steckel, Tennessee extension weed specialist, stated in a 2016 Delta Farm Press article that “We can safely say PPO resistance will be more widespread this spring and in a higher proportion of the population in fields where it was confirmed in 2015.” He further stated that “Liberty® is the last herbicide standing that will control Palmer amaranth post in our soybean fields in 2016.” (Steckel, 2016). However, the use of Liberty for post-emergent control is only available to farmers planting glufosinate-tolerant soybeans. Also of enormous concern is the overdependence currently being placed on glufosinate for the control of Palmer amaranth, resulting in tremendous selective pressure that facilitates development of glufosinate resistance (Sosnoskie et al., 2011). In contrast, dicamba has been shown to be effective for control of multiple herbicide-resistant Palmer amaranth (Reiofeli et al., 2016). Dicamba also will be an important tool in preventing the development and further spread of other herbicide-resistant weeds, including glufosinate-tolerant weeds. The use of dicamba in the Roundup Ready® Xtend Crop System will be a benefit to farmers for controlling resistant weeds and maintaining the durability of currently available weed control options. BEAD’s review confirmed the importance of dicamba in post-emergent control of herbicide-resistant weeds: “Dicamba used during the growing season would target new flushes of weeds and could have the effect of reducing populations of these weeds and particularly would help reduce weed seed banks (i.e., viable seeds in the soil) to reduce populations of a new generation of weeds. Postemergence use of dicamba on genetically modified cotton and soybean during the growing season will expand options for broadleaf weed control.” (U.S. EPA, 2016). A. Studies Confirm Dicamba Has Effective Pre-Emergent Control BEAD’s benefits review states that “dicamba does not control weeds before they emerge.” (U.S. EPA, 2016). While dicamba is not strictly classified as a pre-emergent herbicide, it is an effective tool for pre-emergent weed management. When used on dicamba tolerant soybean or cotton, multiple mechanisms-of-action including dicamba applied pre-plant or pre-emergent can provide more consistent residual control of Palmer amaranth on dryland acres (Steckel 2013). A study by Purdue University in 2015 comparing use of dicamba alone and with multiple mechanisms-of-action found that mixtures containing dicamba applied to soil pre-emergent provided up to 91% control of Palmer amaranth species (Spaunhorst et al., 2010). Experiments using dicamba susceptible and resistant kochia seed suggested that both susceptible and resistant kochia can be over 95% controlled by pre-applied dicamba. The study concluded that “PRE application of dicamba can be a feasible option to manage kochia in rangeland or crop fields and, 3 more importantly, to minimize the spread of resistant plants.” (Ou et al., 2010). The residual weed control provided by dicamba is another reason why farmers will benefit from use dicamba in a pre-emergent weed control regimen. A 2013 study found that under moderate rainfall conditions dicamba provided 50 to 80% control of glyphosate-resistant waterhemp 21 days after treatment (Logan et al., 2013). Bare ground field experiments at 20 locations across the Midwest evaluated the length of residual control of dicamba used alone and in combination with other herbicides. Using dicamba with additional mechanisms of action improved weed control across all species. The addition of dicamba to burndown and in-season weed management in dicamba-tolerant soybean can provide more consistent weed control and a more sustainable solution to management of glyphosate-resistant and hard to control weeds (Willis et al., 2012). Waterhemp emergence was evaluated in a study of soil persistence of dicamba. The study concluded “Initial results suggest that soil applied dicamba can significantly reduce waterhemp emergence up to 35 days.” (Schlichenmayer et al., 2011). Dicamba is currently labeled for pre-emergent use on soy and cotton, but concerns over potential damage to crops and label restrictions such as waiting 21 days before planting cotton and 28 days before planting soybean (BASF Corporation, 2008) limit its use as a pre-emergent herbicide. The proposed label for use in the Roundup Ready® Xtend Crop System, however, allows for the use of dicamba before, during, or immediately after planting (Proposed Label, EPA-HQ-2016-01870015). The use of dicamba in the Roundup Ready® Xtend Crop System thus removes the preplant restrictions, providing greater flexibility to farmers. B. Approval of Dicamba for Use on Dicamba-Tolerant Soybean and Cotton Will Further Facilitate No-Till Farming, Which Provides Significant Environmental Benefits Dicamba-tolerant soybeans and cotton will also provide significant benefits by facilitating noand low-till farming, which in turn confers significant environmental and economic benefits. Tillage is used for seedbed preparation and may also be used for weed control. The primary purposes of pre-plant tillage are to incorporate residue from the previous crop, reduce wheel traffic compaction from the previous season, improve water filtration and soil aeration, control weeds, loosen the soil for root penetration, and provide a suitable environment for planting and germination (Hake et al., 1996). Conventional tillage is associated with intensive plowing and less than 15% crop residue at the soil surface; reduced tillage is associated with 15 to 30% crop residue; and conservation tillage, including no-till practices, is associated with at least 30% crop residue and substantially less soil erosion than other tillage practices (CTIC, 2011). The benefits of conservation tillage or no-till systems relative to conventional tillage are welldocumented and include: reduced soil erosion; reduced fuel and labor costs; conservation of soil moisture; improvement of soil structure; reduction of soil compaction; and improvement of soil organic matter content. A study by CropLife America demonstrated that in 2005 the use of herbicides saved U.S. farmers 337 million gallons of fuel, produced $16 billion in crop yield increases, and cut weed control costs by $10 billion as compared to production without the use of herbicides (Gianessi and Reigner, 2006). Additionally, without herbicides, growers would have to abandon no-till or other conservation tillage production practices, which reduce soil erosion. If U.S. growers stopped using herbicides and resumed tillage on the number of acres not tilled in 4 2005, an additional 356 billion pounds of sediments would be deposited in streams and rivers, resulting in an estimated $1.4 billion in downstream damage (Gianessi and Reigner, 2006). BEAD’s benefits review recognizes the benefits of no-till farming practices but states that no data was provided to support how the use of dicamba in the Roundup Ready® Xtend Crop System would support the continued use of no-till practices. However, the new use of dicamba in the Roundup Ready® Xtend Crop System will provide environmental and economic benefits by enabling the continued use of reduced tillage practices and reducing the inputs required to produce a successful crop. An analysis of farmer survey data in 2011 (Monsanto, 2012) demonstrated a preference for use of post-emergent products on reduced tillage acres compared to reliance on pre-emergent applications of soil residual herbicides. Soybean and cotton production is more productive when herbicides are used to protect the crop from the stress of weed competition. However, controlling weeds can be more difficult in a notill system due to the accumulation of residue from prior plantings, the lack of tillage in the cropping system, and the economics and effectiveness of various herbicide options. With the ability to make pre-plant and in-crop applications with effective and economical herbicides, notill farmers are able to effectively manage weeds throughout the growing season. Specifically, the approval of dicamba for in-crop use would add new control options for weeds, including herbicide resistant weeds that threaten to eliminate the benefits of no-till. Roundup Ready® crop systems provided better weed management options in no-till and reduced-tillage systems, facilitating greater adoption of conservation tillage systems in cotton and soybeans (Baldwin and Baldwin, 2002). Just as the introduction of effective post-weed control options in Roundup Ready crop systems led to an increase in no-till and reduced tillage farming, the introduction of dicamba-tolerant crop systems will provide additional effective postweed control options in the Roundup Ready® Xtend Crop System, and therefore likely will increase the use of no-till and reduced tillage farming. Studies support this conclusion. For example, a 2015 paper demonstrated the effectiveness of dicamba for controlling horseweed in a conservation tillage system (Flessner et al., 2015). Horseweed is among the most troublesome weeds in glyphosate-resistant (GR) cropping systems (Kruger et al., 2009). Control of GR horseweed prior to conservation tillage planting is particularly challenging, often necessitating the use of other herbicides in addition to, or in lieu of, glyphosate as a pre-plant treatment (Steckel et al., 2006). One such herbicide is dicamba. Only 20% GR horseweed control was reported with glyphosate at 0.84 kg a.e./ha, but 77% control was obtained when dicamba at 0.28 kg a.e./ha was tank-mixed with glyphosate applied at spring planting (Owen et al., 2009). GR horseweed was controlled 90 to 100% 8 weeks after dicamba was applied pre-plant at 0.60 kg a.e./ha (Byker et al., 2013); 93 and 98% control was reported 4 weeks after treatment with dicamba at 0.14 and 0.28 kg a.e./ha, respectively (Everitt and Keeling, 2007). C. Approval of Dicamba for Use On Dicamba-Tolerant Soybean and Cotton Will Provide Additional Benefits to Farmers A farmer’s choice of herbicide is based on several factors such as efficacy, crop safety, rotational restrictions, convenience, and the ability to use integrated pest management practices. Many farmers will choose dicamba instead of alternative herbicides for the following reasons: 5 a) Herbicide effectiveness: Dicamba provides control of over 95 annual and biennial weed species and control or suppression of over 100 perennial broadleaf and woody species. While 2,4-D is more effective than dicamba on some weeds, dicamba provides more effective pre-emergent weed control than 2,4-D on cutleaf evening primrose, clover, and chickweed (Loux et al., 2010). Furthermore, dicamba provides better control compared to 2,4-D on summer annuals including those with a prostrate growth habit such as knotweed and purslane. With regard to winter annual weeds, University of Arkansas data indicated that dicamba is more effective in controlling marestail compared to 2,4-D (University of Arkansas, 2011). Dicamba also provides excellent control of wild buckwheat, while 2,4-D has only limited activity and provides inadequate control (Zollinger, 2006). Other pre-emergent or postemergent herbicides often provide incomplete control of wild buckwheat including dinitroanilines or PPO inhibitors. The most effective herbicides for buckwheat are dicamba and some sulfonylurea products; however, some of the sulfonylurea herbicides may persist and carry over for more than one growing season, especially in high pH soils. Dicamba has been valued as more efficacious on lambsquarters than fomesafen, based on university weed control guidelines (Moechnig et al., 2010; University of Illinois, 2008). In addition, dicamba exhibits improved control of sicklepod, kochia and common ragweed, and waterhemp compared to fomesafen. Dicamba has improved post-emergent weed control and added residual efficacy compared to other commercially available herbicides to control tough broadleaf weeds and weeds that are resistant to glyphosate and other herbicides (Johnson et al., 2010). Residual herbicides, in addition to dicamba, will be recommended in the Roundup Ready® Xtend Crop System. Depending upon crop safety and best management practices for overall weed management, including resistance management, dicamba use may reduce the number of applications of soil residual herbicides needed for season long control of Palmer amaranth (e.g. fomesafen, pyrithiobac, fluometuron) (L. Steckel 2012, Personal Communication; S. Culpepper 2012, Personal Communication). Currently, multiple applications of residual herbicides are needed because of the lack of effective post-emergent options (Southeast Farm Press, 2009). b) Crop safety: The proposed use of dicamba in the Roundup Ready® Xtend Crop System in soybean and cotton will allow more flexibility for control of weeds just prior to or at planting of the crop, due to elimination of pre-plant intervals or plant back restrictions on present dicamba labels. These restrictions were in place due to concern over potential crop injury, which would be eliminated by the introduction of the dicamba tolerance trait. Current practice is to use dicamba or 2,4-D for pre-plant burndown of weeds that are present prior to planting cotton. When applied too close to cotton planting, dicamba or 2,4-D can potentially reduce crop stands and cause injury to new seedlings (Thompson et al., 2007). For most dicamba products, a pre-plant interval of 21 days per 0.25 lb a.i. and a minimum of 1” rain prior to planting are required. For 2,4-D, a minimum of 1” of rainfall and a waiting interval of 30 days is required prior to planting with rates up to 1.0 lb a.i. per acre. 6 c) Rotational restrictions: Some residual and post-emergent herbicides have extensive rotational restrictions of up to 20 months after application to avoid subsequent crop injury to the rotational crop due to herbicide remaining in the soil. These limitations reduce the choice of crops that can be re-planted the same season in case the initial crop is destroyed, or even the crops that may be planted the following growing season. Examples of replanting limitations among four of the alternative herbicides are shown in the table below. Planting Restrictions (months) for Some Alternative Herbicide Products ROTATIONAL Reflex Treflan Envoke Staple LX (fomesafen- HFP (Trifloxysulfuron- (pyrithiobacCROP sodium) (trifluralin) sodium) sodium) Field corn 10 12 - 20 7 10 - 20 Wheat 4 5 3 4 Soybean 0 same 7 4 - 12 season Peanut 10 5 7 10 Sorghum 10 - 18 12 - 20 12 More than 10 Onions 18 5 18 4 - 12 The combination of dicamba and the Roundup Ready® Xtend Crop System will remove pre-plant restrictions providing greater flexibility to farmers and will increase the adoption of dicamba use in pre-plant weed control systems. Dicamba does not have rotational restrictions for such extended time periods, and therefore provides a substantial advantage in flexibility. For the majority of crops, no rotational restrictions apply after 120 days following a dicamba application. In addition, there are no rotational restrictions for planting corn following a dicamba application. The lack of long rotational restrictions allows farmers to keep their land in productive cultivation of crops and avoids unproductive fallow periods, which is of critical importance in selecting which herbicide(s) to use. d) Greater convenience: Dicamba use in the Roundup Ready® Xtend Crop System in soybean and cotton will reduce the need for some herbicides that require use of specialized application equipment such as hooded and directed spray equipment (diuron, paraquat, flumioxazin, MSMA) to prevent crop injury. Hooded sprayers in particular require more time to make herbicide applications and application errors can lead to significant crop damage. In addition, hooded spray rigs employ shorter boom lengths compared to 100-foot spray booms requiring more passes across the field, which in turn use more resources such as labor, fuel, and wear-and-tear on equipment. e) Support integrated pest management practices: Dicamba use in the Roundup Ready® Xtend Crop System in soybean and cotton will continue to support the adoption of integrated pest management (IPM) practices by allowing farmers to continue to focus on post-emergent, in-crop weed control, as they have practiced with glyphosate- and glufosinate-tolerant crops, and making current practices more effective 7 by providing an additional mechanism-of-action. This will allow farmers to avoid unnecessary herbicide treatments and to delay some herbicide treatments until field scouting indicates a need for additional weed control, which is consistent with the principles of IPM. In sum, dicamba is a particularly effective herbicide, especially on multiple-herbicide resistant weeds and hard-to-control weeds. The use of dicamba in the Roundup Ready® Xtend Crop System will benefit farmers by controlling resistant weeds and maintaining the durability of currently available weed control options. It will provide environmental and economic benefits by expanding the use of reduced tillage agronomic practices and reducing crop inputs. Thus, the registration of dicamba for use in the Roundup Ready® Xtend Crop System represents a significant opportunity to extend the well-established benefits of herbicide-tolerant cropping systems. The new use of dicamba will improve the management of herbicide-resistant weeds, and thus overall weed management and sustainability of crop production in the United States. It will provide a valuable weed management tool, provide more weed control options for farmers, and help to slow the selection for resistant herbicide genotypes for all herbicide mechanisms-ofaction (i.e., ALS, PPO, HPPD, glufosinate, glyphosate) that are currently registered for use in soybean and cotton while supporting the adoption of no-till and conservational tillage practices. III. A DOWNWIND BUFFER OF 110 FEET IS MORE THAN SUFFICIENT TO PROTECT ENDANGERED SPECIES AND NON-TARGET ORGANISMS EPA has proposed to mandate a 110-foot buffer in all directions to protect non-target plants and threatened and endangered species from dicamba off-site movement from volatilization and particle drift. Particle drift is also known as spray drift. EPA concluded that a 100-110 foot downwind buffer was sufficient to address particle drift, but proposed to impose that buffer in all four directions to protect against off-site movement from volatilization (Ecological Risk Assessment for Cotton, EPA-HQ-OPP-2016-0187-0005). For particle drift, Monsanto data confirms that a 110-foot downwind buffer is more than sufficient to protect endangered species and non-target organisms for the TTI 11004 nozzle (MRID 49424601)4; for volatility, Monsanto data establishes that no buffer is required to protect endangered species and non-target organisms regardless of wind direction. EPA’s proposed registration stated that “… if EPA receives volatility data under varied conditions of temperature and relative humidity, as these factors play a strong role in volatility under field conditions, it may reconsider whether this mitigation requirement [in-field buffer in all directions] is necessary.” In response to this request, on April 12, 2016, Monsanto provided EPA with additional volatility data conducted under varied conditions of temperature and relative humidity in key use areas for the Monsanto Dicamba Products. The data demonstrate that a buffer is not necessary to protect against volatilization. Although earlier dicamba products (including Banvel®) had the potential to move off-site through volatilization, multiple lines of evidence demonstrate that there should be no concern of off-site movement due to volatility for the Monsanto Dicamba Products, and that a downwind buffer of 110 feet is more than adequate 4 Any nozzle that does not increase the drift potential of M1691 Herbicide relative to the nozzle for which the buffer has been determined should be acceptable 8 to protect non-target organisms from potential adverse effects when the mandatory application requirements described in the draft M1691 Herbicide label are instituted. These lines of evidence include a laboratory humidome study, flux measurements from field volatility studies, and air modeling to predict off-target air concentrations and deposition due to off-target movement via volatility, as described further below. Thus, EPA should remove the proposed requirement for an omnidirectional buffer and instead retain only a downwind buffer to protect non-target plants and threatened and endangered species in order to prevent unreasonable impacts to U.S. agriculture. As discussed further in Section III.A, a four-sided buffer would impose large costs on growers, and disproportionately impact smaller-scale farms. A. A Four-Sided Buffer Would Harm Farmers and Agriculture Requiring a buffer around the entire perimeter of the field would impose substantial complexity, management challenges and potential economic impacts on growers, and large costs on growers and commercial applicators. The percentage of a field that would be occupied by a four-sided buffer is large, regardless of field size, as shown in Table 1. This would leave a substantial area at risk of poor weed control and decreased yield, forcing growers to apply multiple herbicide mixes on every field, at significant financial cost, by using multiple applicator rigs, or frequently rinsing existing rigs. Table 1. Field Buffer As Percentage of Field5 Critically, field sizes between 10 and 120 acres are common; growers of those fields, in particular, may not be able to economically withstand leaving this portion of their field untreated. Moreover, the percentages in Table 1 assume a square or rectangular field. However, many fields are irregularly shaped, which exacerbates greatly the applicator’s ability to observe a buffer on the entire perimeter of the field. These impacts would be disproportionate on smaller scale farmers and farmers with large numbers of small or irregular-shaped fields. It will be impractical to use the technology in many small fields where such a significant portion will require alternative weed control measures. Compliance with the buffer requirements will also be 5 These data assume fields are immediately adjacent to sensitive areas on all sides. 9 challenging in irregular-shaped fields and may result in unintended herbicide application overlaps and gaps. Moreover, cotton is commonly produced on raised beds which require equipment passes over the field to be in the same direction as the rows. It will be impossible to manage weed control in buffers on the ends of the field with applicator passes perpendicular to the rows. This means the applicator will have to run across the entire field twice- once to apply a dicamba-based herbicide program and a second time to apply a non-dicamba-based program in the buffer areas, resulting in additional labor, fuel, equipment wear and mechanical damage to the crop and soil. Applicators may be able to mitigate some of the challenges posed by buffers around the entire perimeter of the field by installing direct injection systems on existing equipment or purchasing new equipment with this capability. However, this would require significant equipment investments and may not be practical for smaller scale growers. This buffer requirement will also prevent the use of dicamba-based herbicide programs for management of resistant weed populations on significant numbers of acres. This will limit M1691 Herbicide’s utility as a herbicide resistance management tool and allows potential proliferation of resistant weeds that would otherwise be controlled by this technology. Lastly, Monsanto agrees with EPA that the following areas should be included in the buffer distance calculation when adjacent to field edges:  Roads, paved or gravel surfaces.  Planted agricultural fields containing dicamba-tolerant crops and monocot crops such as corn, sorghum, proso millet, small grains and sugarcane.  Agricultural fields that have been prepared for planting.  Areas covered by the footprint of a building, shade house, silo, feed crib, or other manmade structure with walls and/or a roof. B. A Buffer Is Not Needed to Protect Against Volatilization of the Monsanto Dicamba Products. Monsanto has provided EPA with additional volatility data and exposure modeling for the Monsanto Dicamba Products. EPA previously concluded that volatility is not a major component to offsite movement (EPA-HQ-OPP-2016-0187-0005, pg. 7); this additional evidence demonstrates that there should be no concern of off-site movement due to volatility. These data provide multiple lines of evidence that the Monsanto Dicamba Products do not require any omnidirectional buffers to address volatility. The flux values, calculated from EPA guideline field volatility studies, are consistent with the assessments of low-exposure concerns previously made by the EPA from the previously submitted and reviewed Theoretical Profile Shape (TPS) method volatility study (MRID 49022501). These GLP guideline studies were conducted under varied environmental conditions (higher temperatures, varied relative humidity, and different soil types in Texas and Georgia), rates (0.5 lb/A and 1.0 lb/A) and agronomic 10 conditions (in-crop and pre-emergence) that are consistent with typical applications in cotton and soybean growing regions. These additional flux values as well as the TPS method flux value are at least two to three orders of magnitude lower than the conservative modeling estimate that was determined by the EPA using the Woodrow equation. These low flux values, obtained under varied environmental conditions, provide additional weight of evidence that M1691 Herbicide is a low-volatility formulation, and that M1768 and M1769 herbicides are also low volatility formulations. These values have been used as input parameters for standard EPA models (PERFUM and AERMOD) to estimate off-site dicamba air concentrations and deposition, respectively. These modeling results, consistent with EPA’s assessment using flux values from the TPS study, predict that there would be no effects to non-target plants outside of the treated area on any side of the field, regardless of wind direction. EPA previously concluded that volatility is not a major component to offsite movement; this additional evidence demonstrates that there should be no concern of off-site movement due to volatility. Therefore, an in-field buffer in all directions is not necessary to address volatility; only a downwind buffer for particle drift is necessary to protect non-target organisms and threatened or endangered species. 1. Laboratory (Humidome) Relative Volatility Study Showed Low Volatility Levels, Particularly Compared to Banvel A laboratory method utilizing a growth chamber under controlled environmental conditions was used to assess the relative volatility of formulated products (Gavlick et al., 2016). This published method has been used to assess the relative volatility of the Monsanto Dicamba Products and Banvel® (dicamba dimethylamine (DMA) salt containing formulation) (MRID 49888605). The relative volatility6 of four dicamba containing formulations was determined. M1768 Herbicide was approximately three orders of magnitude less volatile than Banvel®, and over one order of magnitude less volatile than M1691 Herbicide. M1769 Herbicide exhibited half the volatility of M1691 Herbicide (Figure 1). 6 The humidome method is designed to assess relative volatility of different formulations. Measured concentrations for a given formulation may vary from study to study. 11 Figure 1. Relative volatility data 2. EPA Guideline Field Volatility Studies Conducted Under Varied Environmental Conditions Provide Additional Data for the Assessment of Off-Site Movement Monsanto conducted a total of six new EPA Guideline 835.8100 field volatility studies to support the conclusion that the Monsanto Dicamba Products are low-volatility formulations (MRIDs 49888401, 49888403, 49888501, 49888503, 49888601, 49888603), which is in agreement with the already submitted TPS field volatility study described above (MRID 49022501). The field volatility studies were conducted under varied environmental conditions (higher temperatures, varied relative humidity, and different soil types in Texas and Georgia), application rates (0.5 lb/A and 1.0 lb/A), agronomic conditions (in-crop and pre-emergence) and data was collected over longer duration (3 days). Flux values from these studies were calculated using the EPA Study Profile template. A comparison of the peak flux values from Texas and Georgia for M1768 Herbicide and M1769 Herbicide shows that these formulations are lower volatility relative to M1691 Herbicide from the same locations (Figure 2) and are consistent with the same formulation volatility trends (M1691 (Clarity) > M1769 (Roundup Xtend) > M1768 (Xtendimax)) observed in the humidome study (MRID 49888605) described above and shown in Figure 1. 12 Figure 2. Peak Flux Values from field studies conducted in Texas and Georgia The results from these field volatility studies on the three formulations were used in conjunction with standard EPA models (PERFUM and AERMOD) to demonstrate that there will be no effects outside of the field due to volatilization. The analysis of potential offsite movement of dicamba due to volatility is discussed in more detail in the following section. 3. Modeling of Off-Target Movement Due to Volatility Demonstrates That a Buffer Is Not Necessary to Protect Against Volatilization of the Monsanto Dicamba Products Flux values calculated from the field volatility studies were used along with a range of weather data to predict off-target movement via volatility. The selected weather conditions used in this assessment represent widely varied weather conditions, covering a broad geographical range representative of soy and cotton-growing regions that represents conservative worst case conditions. The predicted air concentrations and deposition values were then compared to noeffect concentrations/rates to assess distances from the application area that would result in no effects to listed non-target plant species. These analyses are described further below, and demonstrate that no effects of dicamba will occur outside the boundaries of the field. a. AERMOD Volatility Deposition Modeling Demonstrates That The NOER Is Not Exceeded In Any Direction Estimates of dry and wet deposition that could potentially occur downwind of an application of the Monsanto Dicamba Products were derived using the U.S. EPA dispersion model AERMOD7 7 AERMOD reports were submitted on May 27, 2016 (MRIDs 49925701, 49925801, and 49925901 for M1691, M1768, and M1769 herbicides, respectively). 13 considering the application window for dicamba during the year.8 The major inputs to AERMOD include the flux rate, which represents the mass off-gassing on an hourly basis after an application, and the meteorological conditions. Flux rates were derived from the field volatility (flux) studies conducted in Georgia and Texas by Monsanto for all three formulations. AERMOD also requires meteorological data files with hourly values for wind speed, wind direction, air temperature, atmospheric stability, and characteristics of the planetary boundary layer. Meteorological files were developed using National Weather Service data sources for Raleigh, North Carolina; Peoria, Illinois; and Lubbock, Texas9 which represent a range of key growing areas and widely varied environments. AERMOD deposition estimates were made for an 80-acre application10 using hourly weather data from each day of the application window at three locations. Estimates were made at several downwind distances from the edge of the field, including 5, 10, 20, 30, 40, 50, 75, 100, 125, and 150 meters. The estimates represent high-end values that may occur in the direction the wind is blowing during meteorological conditions that are least conducive to gas dispersion (i.e., conditions that maximize offsite deposition – a conservative assumption). At the Georgia site, the 90th percentile dry deposition at 5 m ranged from 9.55×10-7 to 4.55×10-6 g/m2, and the 90th percentile wet deposition ranged from 2.28 ×10-9 to 4.41 ×10-8 g/m2. At the Texas site, the 90th percentile dry deposition at 5 m ranged from 5.77×10-7 to 2.12 ×10-6 g/m2, and the 90th percentile wet deposition ranged from 2.72×10-9 to 1.99 ×10-8 g/m2. In all cases, the 90th percentile deposition results were below the NOER for listed non-target plant (2.91 x 10-5 g/m2) in the downwind direction and represent worst case exposure estimates. Deposition from volatility in other directions would be even lower than levels predicted in the downwind direction, and thus the deposition would be less than the NOER in all directions from the spray area. Based on this analysis, the NOER is not exceeded in any direction; therefore, no buffer whatsoever is required to protect non-target organisms from any potential effects to due to volatility. Moreover, such a four-sided buffer would impose large and unnecessary costs on growers, and would disproportionately impact small-scale farms. 8 Since dicamba is only applied for this use during the growing season, the deposition values were only determined during the following time periods: Lubbock, TX (May 1 – Jan. 4); Peoria, IL (Apr. 1 – Jul. 31); Raleigh, NC (Apr. 1– Dec 8.). Time periods that extend into the winter are the result of the extended application window of cotton versus soybean. 9 Phoenix, AZ is included in the PERFUM modeling described below but was not included in the AERMOD assessment for two reasons: (1) cotton growing regions were already represented by the Texas and Raleigh scenarios; and (2) the air concentrations from Phoenix were comparable to the air concentrations from either Raleigh or Peoria, depending on the averaging period, and thus the Phoenix site was not modeled. 10 AERMOD deposition estimates were based on application rates of 1.0 lb a.e./A for the Georgia site and 0.5 lb a.e./A for the Texas site. 14 b. Additional Analysis of Dicamba Air Concentrations from Volatility Demonstrate That There Would Be No Effects Outside the Field From Volatilization To assess potential effects to non-target plant species as a result of exposure to dicamba in air, predicted off-target dicamba air concentrations were compared to a vapor phase no effect concentration (NOEC) for soybean, the plant species most sensitive to dicamba. The vapor phase NOEC was determined from a study designed to evaluate the relationship between dicamba vapor concentrations and plant effects. Dicamba air concentrations were estimated outside the spray area using the U.S. EPA model PERFUM.11 The NOEC was not exceeded in the downwind direction and thus would not be exceeded in any other direction; as such, a buffer in all directions is not necessary to protect threatened or endangered species and non-target organisms from the effects of volatilization from the Monsanto Dicamba Products. Additional details regarding the air concentration plant effects study used to determine the vapor phase NOEC and the off-target vapor modeling are provided below. Monsanto recently conducted and submitted a humidome study (MRID 49925703, submitted to EPA on May 27, 2016) to generate laboratory-based data to determine the relationship between dicamba vapor concentration and crop response. The results of this study determined a no-effect concentration of at least 17.7 ng/m3 for a 24-hour period, which served as a basis of comparison to the air concentration estimates from the PERFUM modeling analysis described below for the Monsanto Dicamba Products. Air concentration estimates that could potentially occur downwind of an application of the Monsanto Dicamba Products were derived using the Probabilistic Exposure and Risk model for FUMigants (PERFUM) considering the application window for dicamba for the year.12 The major inputs to PERFUM include the flux rate, which represents the mass off-gassing on an hourly basis after an application, and the meteorological conditions. Flux rates were derived from the field volatility (flux) studies conducted in Georgia and Texas by Monsanto for all three formulations, as described above. PERFUM requires meteorological data files for a multiyear period with hourly values for wind speed, wind direction, air temperature, atmospheric stability, and mixing height. Meteorological files were developed using U.S. Environmental Protection Agency data sources for Raleigh, North Carolina; Peoria, Illinois; Lubbock, Texas; and Phoenix, Arizona which represent a range of key growing regions and widely varied environments. 11 PERFUM reports were submitted on May 27, 2016 (MRIDs 49925702, 49925802, and 49925902 for M1691, M1768, and M1769 herbicides, respectively). 12 Since dicamba is only applied for this use during the growing season, the deposition values were only determined during the following time periods: Lubbock, TX (May 1 – Jan. 4); Peoria, IL (Apr. 1 – Jul. 31); Phoenix, AZ (Mar. 1 to Dec. 18); Raleigh, NC (Apr. 1– Dec 8.). Time periods that extend into the winter are the result of the extended application window of cotton versus soybean. 15 PERFUM air concentration estimates were made for an 80-acre application13 at several downwind distances from the edge of the field, including 5, 10, 25, and 50 meters. The air concentration estimates presented are the 95th percentile14 (95% of air concentration values would be lower) of all values at a given distance from the field, considering all wind directions and all the meteorological data for a given site. Thus, the estimates represent high-end values that may occur in the direction the wind is blowing during meteorological conditions that are least conducive to gas dispersion (i.e., conditions that maximize offsite air concentrations – a conservative assumption). Additionally, air concentration estimates were made for four different averaging times, including 1, 4, 8, and 24 hours. The 24-hour air concentration estimates at the nearest distance predicted by the model range from 1.5 to 16.1 ng/m3, which are less than the no-effect concentration of at least 17.7 ng/m3 determined in a laboratory (humidome) plant effects study (described above). Air concentrations from volatility in other directions would be even lower than the concentrations predicted in the downwind direction since dicamba vapor is carried with the wind in the downwind direction, and thus air concentrations would also be lower than the NOEC at all other directions from the spray area. Based on this conservative exposure estimate, the NOEC is not exceeded in the downwind direction (i.e., conditions that maximize offsite air concentrations – a conservative assumption) and would not be exceeded in any other direction, regardless of whether wind direction changes after application; therefore, a buffer in all directions is not necessary in order to sufficiently protect threatened or endangered species and non-target organisms from the effects of volatilization from the Monsanto Dicamba Products. 4. Any Effects of the Monsanto Dicamba Formulations From Off-Site Movement Are Limited to Downwind Particle Drift at the Time of Application The results from these additional lines of evidence for the Monsanto Dicamba Products – including laboratory (humidome) volatility studies, field volatility studies, and off-target movement modeling – demonstrate that there should be no concern of off-site movement due to volatility. Any potential effects from movement of the Monsanto Dicamba Products therefore only occur in a downwind direction at the time of application, and the Monsanto Dicamba Products do not leave the field in any other directions at levels that could impact non-target organisms or threatened or endangered species. Furthermore, the field volatility studies and offtarget modeling consider varied temperature and relative humidity conditions, as specified by U.S. EPA. As such, a downwind buffer for particle drift is protective of any potential effects to non-target organisms and threatened and endangered species, and no buffer at all is necessary to 13 PERFUM air concentration estimates were based on flux values from studies with application rates of 1.0 lb a.e./A for the Georgia site and 0.5 lb a.e./A for the Texas site. 14 EPA exposure assessments typically consider the 90th percentile value (US EPA, 2013) as representative of an appropriate level of conservatism for assessment of exposure. Thus, use of the 95th percentile value represents an even more conservative assessment. Use of either 90th or 95th percentile values is consistent with the recommendation of the National Academy of Science panel (NRC, 2013) that recommended the use of probabilistic risk assessment. 16 protect non-target organisms or threatened or endangered species from the effects of volatilization of the Monsanto Dicamba Products. Any buffer, other than a downwind 110-foot buffer at the time the Monsanto Dicamba Products are applied, would be unnecessary and unreasonably restrictive to growers. C. The Available Data on Particle Drift Support That a Downwind Buffer of No More Than 110-Feet at the Time of Application Is More Than Sufficiently Protective Monsanto previously submitted multiple lines of evidence to support that a downwind buffer of 110 feet will be protective of potential adverse effects to non-target organisms and threatened or endangered species (MRID 49292801, MRID 49424601). Specifically, a 70-foot buffer will be protective of potential adverse effects from ultra coarse droplets, and a 110-foot buffer will be protective of potential adverse effects from extremely coarse droplets. Therefore, a 110-foot buffer will be protective of adverse effects from both ultra coarse and extremely course droplets. The lines of evidence previously submitted to EPA included deposition curves derived from both generic15 and dicamba-specific particle drift deposition data used in conjunction with the most sensitive endpoint from the DGA salt vegetative vigor study to estimate a no effect distance. Chemical deposition data included data from a 1993 Spray Drift Task Force trial using an extremely coarse nozzle, data from Agriculture, Agri-Food Canada trials conducted in 2000 at relevant weather conditions using an extremely coarse nozzle, and data from a dicamba diglycolamine salt-specific BASF field deposition study (MRID 49067704) using the TTI nozzle. Collectively, these lines of evidence indicate that a downwind buffer distance of 110 ft for dicamba applications at a rate of 0.5 lb a.e./A is more than adequate to protect non-target organisms from potential adverse effects when the mandatory application requirements described in the draft M1691 Herbicide label are instituted16. This conclusion is further supported by plant effect studies conducted under strict data quality standards (MRID 48876001). Furthermore, the data collected to support these lines of evidence were collected under protocol following GLP methods or implemented strict data quality measures and should be given greater weight than ancillary information obtained from product development research or incident reports that lack sufficient detail to fully reproduce the conditions of the dicamba applications. 15 Particle drift is primarily influenced by the drop size distribution that results from a solution being sprayed through a nozzle. Generic deposition data are appropriate for a broad range of spray solutions provided that the data used to generate the generic deposition curve have a similar drop size classification to that produced by a spray nozzle. This approach is consistent with the approach used in standard EPA exposure assessments that utilize generic spray deposition curves in AgDRIFT. 16 Monsanto has also submitted product-specific field deposition studies for M1768 (Submitted November 19, 2015; MRID 49770301) and M1769 (Submitted April 12, 2016; MRID 49888606) herbicides that provide evidence that the no-effect distances are less than the 110-foot downwind buffer distance for both M1768 and M1769. 17 D. Information from Product Development Research and Extraneous Incidents Do Not Undermine the Sufficiency of a 110-Foot Downwind Buffer for Particle Drift, and No Buffer for Volatility 1. M1691 Herbicide Was Not Used Alone in the Product Development Research Monsanto agrees with EFED that data generated as part of product development research – including academic field trials – is not suitable for regulatory decision-making (see EPA-HQOPP-2016-0187-0005, pg. 25); no attempts were made to comply with rigorous data quality standards or EPA test guidelines, and no formal study reports were generated. The data do not meet the standards of CFR Parts 158 and 160 and are wholly unsuitable for use in EPA’s decision-making process for the registration of M1691 Herbicide. Indeed, it is important to recognize that the ‘field trials’ that EPA references were conducted to understand and refine components (nozzles, formulations and spray mixtures, label restrictions) of the Roundup Ready® Xtend Crop System – components that ultimately will become part of the label requirements for the product. As such:  M1691 Herbicide alone was not used in any of the field trials;  Experimental formulations that were evaluated are not being commercialized;  There has been no confirmation that these treatments have droplet size distributions that would fall within the acceptable range compared to the droplet size distribution of the spray mixture on which the 110 ft buffer is based; and  In all cases not all currently proposed label requirements were met. These field trials are a critical part of product development, and are essential to help determine the appropriate conditions and requirements for the product to ensure appropriate stewardship. Moreover, these field trials necessarily must include some level of “trial and error” – such that incidents are expected to occur. It therefore would be profoundly inappropriate for EPA to use these field trials – which are not representative of the product that will be sold or consistent with the requirements of the label – to make determinations about potential off-target movement, and therefore about appropriate buffers or other label conditions. Pesticide manufacturers should not be penalized for conducting critical research that informs the ultimate label requirements. 2. EPA’s Analysis of Field Trial Data Did Not Consider Dose Response Even if the field trial data were considered, a rigorous modeling approach would show no adverse effects beyond 110 feet (MRIDs 49570501 and 49570502), with one exception likely caused by important differences between the experimental formulation and M1691 Herbicide, including the surfactant loading and potential differences in the drop size distribution. These data support that a 110 ft. downwind buffer is protective of sensitive crops and threatened and endangered species. Off-target movement was evaluated by spraying a block of sensitive soybean plants in the middle 18 of a much larger field, and then taking measurements on quantitative endpoints (e.g., plant height, yield) downwind. If off-target movement was sufficient to impact plants downwind, distance-dependent variation in plant height (or yield) was observed (i.e. a dose-response, Figure 3). If off-target movement was not sufficient to impact plants downwind, variation in plant height (or yield) was independent of distance (Figure 4). The absence of a distance-dependent relationship (i.e. no dose response) indicates that the level of dicamba exposure was not sufficient to impact plants, even immediately adjacent to the spray application. Figure 3. An example of distance-dependent variation in yield (i.e. a dose-response) and illustration of the segmented regression analysis. 19 Figure 4. A lack of distance-dependent variation in yield (i.e. no dicamba effect). EFED’s use of hypothesis testing led to purported effects on yield up to approximately 250 ft., indicated by the arrow. Applying a segmented regression modeling approach (illustrated in Figure 3), there are no adverse effects on non-target plant growth or reproduction beyond 110 feet, with the exception of one spray application. This exception likely was caused by dissimilarities between the spray mixture used in the field and M1691 Herbicide, as discussed in greater detail at the end of this section. In their analysis, however, EFED employed a hypothesis testing method which used two-sample t-tests (one-tailed, α=0.1) to compare plant height and yield at each measured distance to “control groups”, represented by the two farthest distances downwind of the spray application. The “no-effect” distance for a given site was considered to be the first distance greater than the furthest distance downwind which had a significant decrease compared to the control group. Based on this analysis and without evaluating or considering dose-response, EFED concluded that several of the field trials provided evidence that a 110-foot buffer might not be sufficient to protect listed species. But consideration of a dose-response is especially important in a field situation where non-uniformity in soil type, soil moisture content, soil compaction, soil fertility level, crop emergence, etc. may cause an increase or decrease in plant height or yield in a small area of a field, completely independent of any exposure to dicamba. With the potential for this type of variability, it is important that samples not be evaluated individually against a control, but rather the evaluation should consider the response across the field in its entirety. As a result, the hypothesis testing method employed by EFED had an uncontrolled and unknown Type 1 error rate (i.e., the rate of false positives) and generated no-effect estimates that were, at least in part, an artifact of the experimental design (i.e., results were dependent on the distances 20 at which measurements were taken). Consider the analysis of the second spray swath for yield at Monmouth, Illinois (Figure 4). In this trial, measurements were taken at 35 downwind distances. Using EFED’s one-tailed t-test approach, yield was un-impacted at the highest levels of dicamba exposure – the p-values for the first three distances immediately downwind of the spray were 0.22, 0.40, and 0.18, respectively. Considering the entire dataset, variation in yield is clearly independent of distance and a conclusion of no adverse effect is warranted. Nonetheless, EFED proceeded with an analysis consisting of 33 unprotected hypothesis tests using a “control group” with 2 plants. One concern when making several comparisons in a single analysis is whether significant differences are due to real differences or simply the result of making a large number of comparisons. Making a large number of comparisons increases the chance of finding differences that appear to be significant when they are not. With an α-level of 0.10 and assuming the tests were independent, we would expect to find (0.10)(33) = 3.3 significant differences even in the absence of any real dicamba effect. The family-wise error rate is 1-0.9033 = 0.9691. This means that the probability of making at least one false claim of a dicamba effect is approximately 97%. Further complicating the interpretation is the fact that these hypothesis tests are not independent, because yield is correlated in space. This dependence between the 33 hypothesis tests and the inadequate level of replication in the control group have the potential to inflate the Type 1 error rate a very large, but unknown amount. P-values cannot be interpreted within the framework of hypothesis testing when the Type 1 error rate is unknown and uncontrolled. The conclusions at this and all other sites are therefore suspect. In addition to the limitations and weaknesses of the statistical analysis, site-specific considerations are discussed below. Rower, AR M1691 Herbicide was not evaluated in this field trial. Additionally, no dose-response was observed; yield was un-impacted at the highest levels of dicamba exposure (immediately adjacent to the spray application). The result from EPA’s analysis in Rower is not supported by other measured endpoints. Based on EPA’s analysis, the no-effect distances for plant height, the most sensitive endpoint from the vegetative vigor study with Clarity, were 7.9 and 20.6 ft. for 14 and 28 DAT, respectively. Combined with the limitations and weaknesses of the statistical analysis conducted by EFED (see above), data from this site do not undermine the sufficiency of a 110-foot downwind buffer for protection of non-target plants and threatened and endangered species. Kirkwood, IL M1691 Herbicide was not evaluated in this field trial. Additionally, variation in plant height was independent of distance after approximately 50 ft. The result from EPA’s analysis in Kirkwood is not supported by other measured endpoints. Based on EPA’s analysis, the no-effect distance for yield was 16.25 ft. Combined with the limitations and weaknesses of the statistical analysis conducted by EFED (see above), data from this site do not undermine the sufficiency of a 110-foot downwind buffer for protection of non-target plants and threatened and endangered species. 21 Monmouth, IL Swath 1 M1691 Herbicide was not evaluated in this field trial. Instead, the trial used an experimental formulation that has difference surfactant properties and potentially a different droplet size distribution than M1691 Herbicide. The result from EPA’s analysis in Monmouth Swath 1 is not supported by other measured endpoints. Based on EPA’s analysis, the no-effect distances were 74.2 and 0 ft. for plant height 14 DAT and yield, respectively. Combined with the limitations and weaknesses of the statistical analysis conducted by EFED (see above), data from this site do not undermine the sufficiency of a 110-foot downwind buffer for protection of non-target plants and threatened and endangered species. Monmouth, IL Swath 2 M1691 Herbicide was not evaluated in this trial. Instead, the trial used an experimental formulation that has difference surfactant properties and potentially a different droplet size distribution than M1691 Herbicide. Additionally, no dose-response was observed; instead, yield was un-impacted at the highest levels of dicamba exposure (immediately adjacent to the spray application). The result from EPA’s analysis in Monmouth Swath 2 is not supported by other measured endpoints. Based on EPA’s analysis, the no-effect distances for plant height, the most sensitive endpoint from the vegetative vigor study with Clarity, were 53 and 95.4 ft. for 14 and 28 DAT, respectively. Combined with the limitations and weaknesses of the statistical analysis conducted by EFED (see above), data from this site do not undermine the sufficiency of a 110-foot downwind buffer for protection of non-target plants and threatened and endangered species. Haubstadt, IN Swath 2 M1691 Herbicide alone was not evaluated in this trial. Spray application was made under environmental conditions (e.g., wind speed) that are not compliant with draft label requirements for M1691 Herbicide. The result from EPA’s analysis in Haubstadt Swath 2 is not supported by other measured endpoints. Based on EPA’s analysis, the no-effect distances for plant height, the most sensitive endpoint from the vegetative vigor study with Clarity, were 40 and 80 ft. for 14 and 28 DAT, respectively. Combined with the limitations and weaknesses of the statistical analysis conducted by EFED (see above), data from this site do not undermine the sufficiency of a 110-foot downwind buffer for protection of non-target plants and threatened and endangered species. Lastly, it is also important to note that the spray solutions included in the academic trials were not consistent with the draft M1691 Herbicide label (i.e., M1691 Herbicide was not sprayed 22 alone in any of the academic trials) and are not directly relevant to the proposed M1691 Herbicide registration. Furthermore, there has been no confirmation that the droplet size distributions in these treatments would fall within the acceptable range compared to the droplet size distribution of the spray mixture on which the 110-foot buffer is based. 3. Reported Impacts From Specific Sites Are Due to Uses That Are Inconsistent With Current Proposed Label Conditions The incident data discussed by EPA for specific sites were not compiled for the purpose of being included in EPA’s risk assessment for listed species. The site-specific incident information that was reported does not have the same quality measures (e.g., full data documentation, study protocols, QA/QC, study reports) in place as the regulatory studies, and therefore is not suitable for inclusion in EPA’s risk assessment. All incident information reported to EPA appeared to result from applications that were inconsistent with current label application requirements. In other words, there were no incidents reported when current proposed label application requirements were met. (See MRID 49570501, submitted to EPA April 13, 2015.) The Missouri Department of Agriculture (MDA) reported incidents occurring from 2013 to 2015, and the Arkansas Plant Board (APB) reported incidents occurring in 2015. As discussed by EPA, eight of the 15 total incidents were reported to be a result of a single instance of postemergent dicamba application to DT-cotton of Strut herbicide mixed with glyphosate and applied at two times the labeled rate for proposed M1691 post-emergent use (EPA-HQ-OPP-2016-0187005, p.32). The remaining incidents were reported to be from pre-emergent applications of dicamba, or it is unknown as to whether any dicamba applications were made. In its investigation of a 2014 incident, MDA determined that it is uncertain whether the damage observed was due to dicamba or another cause. A 2013 incident reported by MDA is discussed in further detail below. MO Case File #81513M00701/EIIS Incident #I026579-001 The MDA notified EPA of an incident in 2013 in which it was alleged that damage was observed in a non-DT soybean field from a nearby application of M1691 Herbicide (Clarity®). The final report executed by MDA to EPA did not determine any cause of the damage. However, EPA has indicated that, in a subsequent communication reported by EPA (2015 letter from D. Slade, MDA to Grant Rowland, EPA), MDA concluded that the reported damage was caused by off-site movement due to volatilization. It is not clear what additional information was provided to make this determination, or why it was not included in the final MDA written report of the investigation, but a number of factors indicate that the damage observed on the non-DT soybean field was not caused by the M1691 Herbicide application, consistent with the conclusion in the final MDA report. First, typically plant response in particle drift or volatilization incidents gradually diminishes as the distance from the herbicide application increases (Sosnoskie et al., 2015). In this incident, however, the plant response did not exhibit a gradual decline in the extent of plant response. Instead, there was evidence of slight crop response in soybeans directly east of the sprayed field; however, after a very short distance, all crop response stopped and was not seen in the remainder of the distance (890 yards) between the sprayed field and the non-DT soybean field. This is not 23 consistent with vapor drift, as vapor drift generally results in a gradual diminishment of symptoms as one moves further from the source. The non-DT soybean field is east of the seed production field in question, and during the time of the dicamba application, the wind was blowing from the southeast. The wind direction was verified from independent weather records from three University of Missouri Extension Centers. Wind from the southeast direction would carry any drift from the sprayed field to the northwest, not to the east where the claimant’s field is located. Last, and most importantly, crop visual response was observed on the non-DT soybean field before the dicamba application was made. Plant dicamba residue data also were inconclusive. Plant samples were taken from the non-DT soybean field for analysis of pesticide residue, and of the 3 samples taken and analyzed only one had a dicamba level above the limit of detection. Therefore, it could not be confirmed there was any dicamba present in the other two samples. A county road runs to the west of the claimant’s field, bisecting the area of crop response, which included the northeastern corner of the seed production field. No samples were taken from west of the road, in an area that exhibited similar symptoms. Therefore, the limited available plant dicamba residue data does not link dicamba exposure to the visual response observed. EPA also noted that volatilization at this site may have occurred due to climatic conditions following the application which fall outside the range of conditions for which EPA has data. Monsanto has submitted field volatility studies conducted in Texas under conditions favoring volatility, including high heat and humidity. Historical weather data for Bernie, MO, as reported at the Poplar Bluff Municipal Airport for the week following the spray application date reported in MO Case File #81513M00701, EIIS Incident report number I026579-001 was compared to temperature and relative humidity conditions from an application of M1691 Herbicide from a 2015 field volatility study conducted in Texas (June 8 – 11, 2015) (MRID 49888403). The ranges (minimum and maximum) of mean daily temperature, maximum daily temperature, mean relative humidity, and maximum relative humidity reported in the week following the application of dicamba for MO Case File #81513M00701, EIIS Incident report number I026579-001 were comparable to conditions reported for the M1691 Herbicide field volatility study. As such, based on the results of the M1691 Herbicide flux study conducted in Texas, along with the subsequent AERMOD and PERFUM modeling, a downwind buffer is more than adequate to be protective of non-target organisms. It is highly unlikely that crop damage reported in MO Case File #81513M00701, EIIS Incident report number I026579-001 was the result of off-site movement of dicamba through volatility. This comparison of temperature and relative humidity data is summarized in Table 2. 24 Table 2: Temperature and Relative Humidity Data Comparison; Bernie, MO and M1691 Herbicide Volatility Study (TX) Maximum Daily Temperature (F) Average Daily Temperature (F) Maximum Relative Humidity (%) Mean Relative Humidity (%) E. Bernie, MO July 5-11, 2013 Min Max 87 95 M1691 Volatility Study (TX) June 8-11, 2015 Min Max 78 98 76 86 73 83 85 94 89 99 66 75 51 86 Other Lines of Evidence Also Support the Sufficiency of 110-Foot Downwind Buffer For Particle Drift Only 1. The Published Literature Does Not Show High Vapor Drift for Dicamba Resulting in Non-Target Plant Injury In the Ecological Risk Assessments for Dicamba-Tolerant Cotton and Dicamba-Tolerant Soybean, EPA has stated that papers in the published literature demonstrate that there is high vapor drift from soybean fields resulting in non-target plant injury: “It is important to note that multiple literature studies show that there is a high vapor drift from soybean fields resulting in non-target plant injury.” EPA cited the following studies in support of this conclusion: Al-Khatib and Tamhane, 1999; Auch and Arnold, 1978; Everitt and Keeling, 2009; Kelley et al., 2005; Hamilton and Arle, 1979; Lanini, 2000; Marple et al., 2008; Wall, 1994; Weidenhamer et al., 1989; and Wax et al., 1969. Importantly, however, none of these literature references support the conclusion that there is “high vapor drift from soybean fields.”17 In fact, none of these studies quantified or assessed vapor drift in any way. All of the reviewed papers intentionally made direct applications of dicamba at low rates to simulate particle drift – not volatilization – in order to assess plant effects at known rates. One of the papers (Auch and Arnold, 1978) described a survey of farmers regarding drift, but did not distinguish between particle drift and volatility. In addition, only four of the nine reviewed papers were even conducted with soybeans. These peerreviewed journal articles do not support the conclusion that “there is high vapor drift from soybean fields resulting in non-target plant injury.” 17 The Lanini, 2000, reference was from the Proceedings of the Annual Meeting of the California Weed Science Society rather than a peer-reviewed journal and was not readily available for review. 25 2. The Buffer Distance for Vapor Movement Estimated From Egan and Mortensen Is Incorrect The Ecological Risk Assessments for Dicamba-Tolerant Soybean (MON 87702) and DicambaTolerant Cotton (MON 87791) rely upon a peer-reviewed literature paper by Egan and Mortensen (2012) to conclude the following regarding the buffer distance sufficient to reduce exposure below the soybean NOAEC from dicamba vapor movement: “Based on damage assessments of non dicamba-tolerant soybean plants near treated fields after spray drift from a 0.5 lb/A DGA salt application had dissipated, the authors estimated the exposure at distance by correlation to known dose-damage correlations. They estimated that the 95% upper bound vapor exposure would drop below the soybean NOAEC at approximately a distance of 25 meters (82 feet).”18 Although not directly stated in the dicamba risk assessment documents, it appears that EPA may have used this distance estimate to support the need for an in-field buffer of 110 feet in all directions to be protective of potential effects to non-target plants outside the field due to volatilization. The use of a 25 meter distance to define a buffer based on the 95% confidence interval from this study is inappropriate for three reasons. First, the bootstrap procedure used to generate this confidence interval is well-known to be inconsistent (e.g., Andrews 2000; Hall & Miller 2010) when the parameter (i.e. amount of dicamba) is at or near the boundary of the parameter space (i.e. 0 g a.e./ha). Figure 4c of Egan and Mortensen (2012) shows that all empirical measures of dicamba were at or near zero after approximately 15 meters. Thus, the bootstrap is attempting to put a confidence interval around 0 – the boundary of the parameter space – and there is no way of knowing if the procedure converged to the value that it was designed to estimate. Because of this uncertainty around estimation of the 95% confidence interval, it is inappropriate to use this endpoint to set a buffer distance. Second, if the predicted dose, rather than the 95% confidence interval, is considered, the predicted dose 3 meters from the edge of the plot (0.11 g/ha at 15 m from the center of the plot19) is well below the NOAEC for the most sensitive non-target plant endpoint (0.29 g a.e./ha for foliar exposure of soybeans), indicating no effects outside the treated area would be predicted. Third, since 25 meters is measured from the center of the plot, even if the 95% confidence interval was used, the actual distance from the edge of the plot would be 13 meters (42 feet) not 25 meters. Based on the points above, the results from the Egan and Mortensen (2012) paper support the conclusion that a 18 Ecological Risk Assessment for Dicamba DGA Salt and its Degradate, 3,6-dichlorosalicylic acid (DCSA) for the Proposed Post-Emergence New Use on Dicamba-Tolerant Soybean (MON 87702), at p. 10; Ecological Risk Assessment for Dicamba DGA Salt and its Degradate, 3,6dichlorosalicylic acid (DCSA) for the Proposed Post-Emergence New Use on Dicamba-Tolerant Cotton (MON 87701), at p. 29. 19 The distances cited in Table 3 of the Egan and Mortensen paper (2012) were measured from the center of the dicamba-treated area not the edge. The legend to Table 3 of the paper clearly states that “Distance is from the center of the treated plot, and distance >12 m are located outside of the treated plot.” Adjusting distances for distance from the edge of the plot, rather than the center, 15 meters becomes 3 meters, and 25 meters becomes 13 meters. 26 buffer is not required to protect non-target plants outside the field from dicamba vapor movement. IV. EPA’S RISK ASSESSMENTS ARE SCIENTIFICALLY-SOUND AND SUPPORTED BY THE LAW A. EPA’s Approach to the Risk Assessments Is Scientifically-Sound and WellSupported EPA’s registration of M1691 Herbicide for dicamba-tolerant soybean and cotton is based on extensive toxicity and ecological effects evaluations. As part of this process, EPA requires data on environmental fate, 40 C.F.R. §158.1300, and ecological effects on non-target terrestrial and aquatic animals, id. at §158.630, and non-target plants, id. at §158.660. EPA also specifically focuses on potential effects a pesticide may have to listed species by conducting thorough ecological risk assessments. EPA’s Environmental Fate and Effects Division (EFED) conducted EPA’s assessments of both ecological effects and exposure, to evaluate whether M1691 Herbicide may affect listed species. Where, based on its ecological risk assessment, EPA concluded that M1691 Herbicide would have “no effect” on listed species, it ended its endangered species assessment. Where EPA concluded, instead, that M1691 Herbicide “may affect” but was “not likely to adversely affect,” a species, it conducted further analysis and informal consultation with the U.S. Fish & Wildlife Service (the “Service”). Where EPA concluded that M1691 Herbicide “may affect” and was “likely to adversely affect” a species, EPA imposed a label restriction preventing dicamba from being applied in that county. For screening-level and some higher tier assessments, EPA has developed specific methods for calculation of estimated exposure concentrations. These methods are highly conservative and specifically designed to overstate—and not understate—concentrations that may be present in the environment. EPA also has defined methods for conducting toxicity studies as required by 40 C.F.R. part 158. Designated species within different taxonomic groups of non-target organisms are routinely used in the mandated toxicity studies. The designated species are surrogate species, and have been selected for a number of reasons, including relative sensitivity in a class of organisms, as well as availability and amenability to routine laboratory testing. Especially in the case of threatened or endangered species, it would not be possible to conduct toxicity tests with the species itself. Because values from the same exposure models and toxicity studies are used, EPA typically conducts the screening-level FIFRA general ecological assessment and the threatened and endangered species assessment at the same time. While the safety standards under FIFRA and the ESA are different, EPA uses the screening-level assessment as the first step of both assessments. This is because the first step is a “worst-case” assessment: If an herbicide passes this worst-case screening-level assessment, EPA can conclude that the herbicide will have “no effect” on threatened and endangered species or their habitats under the ESA, and will not cause unreasonable adverse effects to the environment under FIFRA. As a general matter, EFED’s initial screening-level assessments are based, in large part, on levels of concern (LOC) for listed species, which indicate the level of pesticide use at which listed species “may be potentially affected by use.” EPA, Office of Pesticide Programs, Overview of 27 the Ecological Risk Assessment Process in the Office of Pesticide Programs: Endangered and Threatened Species Effects Determinations, 46 (Jan. 2004). The acute LOC for listed species is just a fraction of the LOC for non-listed species. Id. at 47. If the LOC is triggered for a species within a taxonomic group, EPA conducts further analysis for that entire taxa. In performing a screening-level assessment, EPA compares the toxicity value from the most sensitive species tested for a class of organisms to exposure values derived from worst-case exposure scenarios. Further, to add further conservatism to the protection of threatened and endangered species, EPA adds an additional safety factor when evaluating acute risk (10x for aquatic animals, 5x for terrestrial animals) so that the threatened and endangered species under investigation is assumed to be far more sensitive than the most sensitive organism tested. As a general rule, this assumption will substantially overstate the potential adverse effects of the herbicide to the threatened or endangered species. In addition, in estimating the amount of exposure of the species to the herbicide, EPA assumes that the herbicide will be used in proximity to the threatened and endangered species, and that it will be applied at the highest rate of application and number of applications permissible. Again, these assumptions generally overstate the amount of exposure because the herbicide may not be used near the species’ habitat, or at the highest application rate or highest number of applications. Thus, the amount of actual exposure may be considerably less than the amount estimated using the worst-case assumptions. If, using these conservative worst-case assumptions, the screening-level assessment shows that no direct or indirect effects on threatened and endangered species are indicated, EPA can conclusively rule out the potential for adverse effects and properly conclude that there will be “no effect” on threatened and endangered species or critical habitat from use of the herbicide. Thus the screening-level assessment is used to identify herbicide use scenarios where adverse ecological effects are not expected to occur. An herbicide that passes the screening-level assessment is deemed acceptable for use under both FIFRA and the ESA, and both the FIFRA assessment and the threatened and endangered species effects determination are concluded. If the screening-level assessment does not result in a “no effect” determination, EPA does not conclude that an herbicide “may affect” threatened and endangered species. Instead, EPA conducts a more refined risk assessment, replacing some of the generic worst-case assumptions which are designed to overstate potential exposures dramatically, instead using more sitespecific, real-world information pertaining to potential exposure (such as the application rates used under normal agricultural practice; herbicide usage data compiled by the relevant State; the proximity of herbicide use in relation to the habitat of the threatened and endangered species; the results of monitoring for the herbicide in the environment, particularly in or near the species habitat; and the impact of label use restrictions in reducing exposure). Where available, the refined assessment also uses toxicity information relating to the species under examination, in order to assess whether use of an herbicide may have an effect on the listed species. In an abundance of caution, EPA adopted an even more conservative approach here. Specifically, for each state in which M1691 Herbicide is proposed to be registered, EPA evaluated which threatened or endangered species might be present in the state. EPA then evaluated whether those species might be present on or near fields where M1691 Herbicide could potentially be 28 used. Where there was overlap between the species and potential usage, EPA conducted a detailed evaluation of the species’ habitat requirements, life cycle, dietary needs, etc. Based on that analysis, EPA determined whether there would be “no effect” on the TES (in which case EPA’s inquiry concluded), M1691 Herbicide “may affect” but is “not likely to adversely affect” the TES (in which case EPA consulted informally with FWS and obtained the Service’s concurrence with its conclusion) or “may affect” and is “likely to adversely affect” the TES, in which case M1691 Herbicide usage is prohibited in the relevant county. B. EPA’s Authority for a Mitigated No-Effect Finding Is Well-Established In relying upon maximum use scenarios, EPA’s no effect determinations already take into account multiple M1691 Herbicide label mandates. The conditions on the FIFRA label mandated by EPA are legally enforceable under FIFRA, 7 U.S.C. § 136j(a)(2)(G), and become part of the proposed registration under the ESA. Cf. Center for Biological Diversity v. BLM, No. 10-72356 at *12722 (9th Cir. Oct. 22, 2012) (conservation measures should have been considered as part of the proposed action because that would have made them enforceable under the ESA); 50 C.F.R. § 402.02. Critically, EFED’s screening-level assessments already rely upon the label’s mandatory conditions as part of the effects determination. Specifically, “[s]creening-level risk assessments rely on labeled statements of the maximum rate of pesticide application, the maximum number of applications, and shortest interval between applications. Together, these assumptions constitute a maximum use scenario.” Id. at 51 (emphasis supplied). See, e.g., Nikiba Daughtry, Environmental Field Branch, Office of Pesticide Programs, Oxyfluorfen: Analysis of Risks to Endangered and Threatened Salmon and Steelhead, Supporting No Effect Determination for Oxyfluorfen, 1 (April 28, 2004) (“The use of oxyfluorfen will have no direct or indirect effect from loss of food supply or loss of cover in the 26 ESUs of Pacific salmon and steelhead when used according to labeled application directions.”).20 EPA also may take into account other mandatory conditions on pesticide labels, such as mandatory buffer zones. This is entirely appropriate given that failure to follow the label requirements is unlawful. Absent reliance on label application rates and frequency, it would be impossible for EPA to assess potential exposure – and thus impossible for EPA to make either a “no effect” or a “may affect” determination.21 Courts have held that mitigation measures or other enforceable protections should be taken into account when determining the impact of the proposed action on listed species. See, e.g., Sierra Club v. Van Antwerp, 661 F.3d 1147 (D.C. Cir. 2011) (finding “no reason why the general principle of taking mitigation into account should not apply to the decision whether the ESA requires formal consultation.”); Selkirk Consfervation Alliance v. Forsgren, 336 F.3d 944, 956 (9th Cir. 2003) (“If a Conservation Agreement is in place, then the reviewing agencies ought to consider it when evaluating the impact of the proposed action [in the ESA context]”); Sierra 20 http://www.epa.gov/espp/litstatus/effects/oxyfluorfen/oxyfluorfen_analysis.pdf. 21 EPA would either have to make unsupported assumptions about potential exposures – which almost by definition would be arbitrary – or, in the alternative, without any boundaries for analysis, EPA arguably would have to consult with the U.S. Fish & Wildlife Service or the National Marine Fisheries Service regarding every single pesticide and every single endangered species. No court has ever held that this is required. 29 Club v. Marsh, 816 F.2d 1376, 1379-80 (9th Cir. Cal. 1987) (explaining that agency relied upon mitigation to support conclusion that agency action was not likely to jeopardize the species’ continued existence); Center for Biological Diversity v. BLM, No. 10-72356 at * 12739 (9th Cir. Oct. 22, 2012). In fact, outside of the ESA context, case law has long supported the notion that agencies may rely upon mitigation measures to contribute to compliance with environmental laws even where mitigation measures are actually carried out by private parties. See Selkirk Conservation Alliance, 336 F.3d at 955 (listing a number of such cases). Action agencies commonly take into account mitigation or other enforceable requirements that are part of the proposed action to support a no effect determination. So does the U.S. Fish and Wildlife Service, which itself recognizes the propriety of relying on mitigation in the no effect context. For example, as the action agency (rather than the consulting agency) the U.S. Fish and Wildlife Service has, in a number of instances, relied upon herbicide label application rates as well as binding agreements between the agency and refuge farmers to support a no effect determination for the use of herbicides in cooperative farming on the National Wildlife Refuge System. See, e.g., U.S. Fish and Wildlife Service, Midwest Region, Environmental Assessment: Use of Row Crop Farming and Genetically-modified, Glyphosate-tolerant Corn and Soybeans on National Wildlife Refuges and Wetland Management Districts, 26-27 (March 2011); U.S. Fish and Wildlife Service, Mountain-Prairie Region, Environmental Assessment: Use of Genetically Modified, Glyphosate-Tolerant Soybeans and Corn on National Wildlife Refuge Lands in the Mountain–Prairie Region (Region 6), 15 (April 2011). As another example, the U.S. Food and Drug Administration’s no effect determination with respect to its approval of genetic modifications to Atlantic salmon relied on numerous containment measures that were part of the proposed action. See Appendix D, U.S. Food and Drug Administration, Draft Environmental Assessment: AquAdvantage Salmon (May 2012). The U.S. Food and Drug Administration submitted that no effect determination to the Services to amend its previous “may affect” determination. The U.S. Fish and Wildlife Service deemed the no effect determination “well supported” in light of the containment facilities and NOAA had no objections. See id. C. Dicamba Does Not Pose a Risk to Terrestrial Invertebrates In addition to the lines of evidence that EPA has used to conclude that M1691 Herbicide does not pose a risk to threatened or endangered terrestrial invertebrates, Monsanto has identified two publications that also support EPA’s conclusion of low acute and chronic risk for honey bees (Apis mellifera L.) that serve as the surrogate species for terrestrial invertebrates. Scientists at the United States Department of Agriculture, Agricultural Research Service Laboratory in Tucson, Arizona concluded that the survival of adult worker bees was not affected after chronic exposure to dicamba technical active ingredient or to a dicamba salt formulation at dietary concentrations up to 1000 ppm in a sucrose solution (Morton et al., 1972). Scientists from the same lab also concluded that honey bee brood production was not affected when the colony was fed concentrations of a dicamba salt formulation at concentrations up to 1000 ppm a.e. in a sucrose solution (Morton et al., 1972). Based on predicted dicamba residues in pollen and nectar from the new use of dicamba on dicamba tolerant soy and cotton, exposure to honey bee brood would not be expected to exceed 30 the levels tested in this study. In addition, dicamba residues decline rapidly with a half-life of less than 9 days on foliage, and plant species sensitive to dicamba to which dicamba is applied will die and no longer be a source of nectar or pollen. For these reasons, the new use of dicamba would have no effect on brood production, survival and development. Data available in the dicamba registration package demonstrate that dicamba does not pose an acute contact or oral risk to honey bees. These two papers from the published literature demonstrate that dicamba does not pose a chronic risk to either adult honey bees or to brood production at doses up to 1000 ppm dicamba acid equivalent. Dicamba, therefore, does not pose an acute or chronic risk to honey bees. Since honey bees are the surrogate species for other terrestrial invertebrates, dicamba would also not pose an acute or chronic risk to other terrestrial invertebrates. D. EPA Properly Considered Runoff in Its No Effect Rationale for Listed Species EPA’s screening-level risk assessment for cotton and soybean characterized risk following exposure to dicamba residues in runoff and found that the predicted concentrations from modeling were lower than the most sensitive taxa’s endpoint (soybean plant height). EFED concluded that all available lines of evidence supported a “no effects” determination for runoff exposure for off-field listed plants for the proposed labeled use of dicamba DGA. In addition, a “no effect” conclusion from run-off can be extended to animals. The levels of dicamba off-field resulting from runoff and erosion are estimated to be only a small fraction of the application rate; therefore, none of the acute or chronic levels of concern for dicamba would be exceeded. Further, the rainfast mitigation on the label would further protect all listed species in off-field habitats from exposure. E. Dicamba Acid Seedling Emergence Endpoints Should Not Have Been Considered in the Dicamba DGA Salt Risk Assessment In the soybean and cotton risk assessments for use of dicamba on dicamba-tolerant soybean (EPA-HQ-OPP-2016-187-0008) and cotton (EPA-HQ-OPP-2016-187-0005), EPA has not only evaluated effects to non-target plants based on endpoints from the non-target plant studies conducted with the proposed M1691 (dicamba DGA salt) end use product, but has also evaluated effects to non-target plants considering endpoints from the dicamba acid seedling emergence study. In these risk assessments EPA has stated: “For dicamba acid, which DGA salt may dissociate to and which has more sensitive seedling emergence values, RQ values would exceed the LOC of 1.0 for all listed and non-listed monocots and dicots in semi-aquatic areas and for listed monocots and listed and non-listed dicots in dry areas.” However, the dicamba acid study – conducted in 1993 under the previous EPA test guidelines for non-target plant studies – is not relevant to the current risk assessment. A number of differences in methodology between previous and current test guidelines support the use of the most current study which used the dicamba DGA salt, rather than dicamba acid. The table below illustrates some significant differences in test methodology between the two studies: 31 Table 3. Comparison of non-target plant testing methodology Subdivision J guideline Dicamba Acid Seedling Emergence Study (Hoberg, 1993) Current Series 850 Guideline Clarity® (Dicamba DGA salt) Seedling Emergence Study (Porch, 2009) Silica sand Natural or artificial soil OM = 0.11 – 0.17% Up to 1.5% OC (approx 3% OM) Growth chamber Growth chamber, glasshouse, semi-field ¼ strength AAP nutrient solution Fertility as needed based on controls 200 ml solution “added” to sand Surface application Seeds planted 1 cm deep Seeding depth appropriate for seeds Onion seedling survival in control = 83% Control survival of at least 90% needed 30 ml test solution sprayed ( 1026 gal/A) Typical application volumes (10-100 gal/A) “not to exceed runoff” NOEC not determined for all species 1 dose below EC25, 1 dose above EC50, determine NOEC (or calculate EC05 if necessary) The dicamba acid seedling emergence study was conducted in a growth chamber, and the test substance for this study was dicamba acid rather than a typical end use product as is currently required. In addition, the study report from the initial study (Hoberg, 1993) documents numerous observations that cast doubt on the validity and interpretation of the test results. For example, the study report notes visual injury symptoms of brown leaf tips, necrosis, and chlorosis. However, these plant responses are different than typical sublethal plant symptoms associated with exposure to dicamba such as leaf curling, epinasty, thickened internodes, and shortened plants (Auch and Arnold, 1978; Weidenhamer et al.,1989). It is likely that the observed plant responses in the 1993 study were the result of unfavorable test conditions rather than the test material itself. The tip browning and necrosis observed in the studies were likely caused by excess salt accumulation in the leaf tips and leaf margins. The excess salt accumulation, in turn, was likely the result of growing the plants in silica sand, a medium with little or no buffering capacity, coupled with the use of a nutrient solution that was one-quarter strength AAP nutrient solution and applied daily without regard to nutritional requirements for the plants.22 This continued application of nutrient solution likely resulted in the accumulation of excess salt in the leaf tips and leaf margins and in the desiccation and 22 The 1993 study did not follow the most current guidance which requires that nutrient and fertilizer needs should be based on observations of the control plants. 32 browning of the tissue.23 Another key test condition was the use of KNO3 in the media in place of NH4NO3. The use of KNO3 commonly leads to alkalinization over time and plant chlorosis due to iron and magnesium deficiencies. The AAP medium was designed for use in hydroponic systems and in algal studies where the solution is replaced at regular intervals rather than allowed to accumulate in a closed container. There are also differences between the guidelines used in the 1993 study as compared to the current guidelines, thus making the current study more relevant for this risk assessment: the previous use of growth chambers to grow plants versus the current use of greenhouses or field-grown plants; the previous planting of seeds from all species at the same depth (1 cm) versus the current guidance to plant seeds at a depth appropriate to the seed size; and the previous germination of seeds first in paper “rag dolls”, followed by transfer to pots, versus the current guidance to allow germination to occur in soil. In summary, the dicamba DGA salt seedling emergence study endpoints are the appropriate endpoints to use in the current risk assessments because the dicamba DGA salt seedling emergence study was conducted according to the current EPA seedling emergence test guideline, and the methodology used in the dicamba acid seedling emergence study likely resulted in plant injury due to causes other than dicamba treatment resulting in endpoints that should not be attributed to dicamba. Furthermore, based on EPA’s refined analysis using the DGA salt seedling emergence study no effects were observed. V. MONSANTO SUPPORTS EPA’S WEED RESISTANCE PLAN, WITH CERTAIN REFINEMENTS Monsanto greatly values maintaining the durability of our products and agrees that the benefits of tools such as low-volatility formulations of M1691 Herbicide should be preserved for agriculture. Sustainability of this tool is important for American farmers and a business objective for Monsanto. As such, Monsanto supports voluntary stewardship (including as described in PR 2001-5), which enables the marketplace to adjust to emerging challenges around herbicide resistance. For example, over the last 10 years market flexibility has allowed/encouraged farmers to incorporate the use of soil residual herbicides as needed and other complementary postemergent herbicides to address the emerging weed management challenges posed by glyphosateresistant weeds. If proposed weed management plans were to become overly prescriptive, the ability of the market to adjust to new or future emerging weed control challenges could be negatively impacted. Voluntary stewardship should therefore be the basis of EPA’s proposed management of herbicide resistance. 23 The fact that tip browning noted in the 1993 study was reported as occurring on corn, a plant that is not normally sensitive to dicamba effects, is a further demonstration that the effects observed were not dicamba specific effects. 33 With this in mind we provide comments on two areas of the proposed resistance management plans including A) The Herbicide Resistance Management Plan for M1691; B) Maintaining the durability of M1691. A. Monsanto Suggests Adjustments That Will Help to Reinforce Practices That Assist Growers With Herbicide Resistance Management Monsanto is generally supportive of an appropriate Herbicide Resistance Management Plan. For example, Monsanto supports offering education, consultation, and recommendations for implementing additional weed control measures, when requested. Monsanto also supports reporting weed species resistant to M1691 Herbicide under appropriate circumstances. However, Monsanto is opposed to restrictions that go beyond these principles, as discussed below.  Herbicide categorization. Monsanto opposes the categorization scheme proposed by EPA entitled “Herbicide Resistance Categories of Concern and Resistance Management Elements for Use by Risk Managers.” In particular, the implication that because an herbicide is used on a herbicide tolerant plant produced through biotechnology (genetically engineered {GE}) or other breeding methods that the herbicide is automatically categorized as “site of high concern for resistance.” This implies that adoption of biotechnology has caused a rapid onset of resistance in weeds. There is nothing inherently different about GE crops such that the EPA should regulate for herbicide resistance any differently than it should do for herbicides in other crops. Additionally, separating herbicides into different groups will cause confusion about herbicide resistance management and confusion about why and how an herbicide is included in a particular category. Herbicide resistance in weeds is an herbicide and herbicide-use issue. It is not associated with how the crop was bred or developed or if the crop is tolerant, resistant or susceptible to an herbicide. Monsanto urges EPA to eliminate 34 the categorization and rather review as appropriate each herbicide against the 11 elements outlined in Table A. This will reinforce practices to help with herbicide resistance management, will be equitable across manufacturers, and will avoid confusion about the categorization.  Specific elements of the Herbicide Resistance Management Plan Checklist. EPA proposes to mandate all of the elements in Table A in the final resistance management plan. While Monsanto will endeavor to address each element, it is important to consider that some do not apply to this product. For example, Element 10 is an element specific to combination products with multiple Mechanisms of Action. Element 11 is not applicable, as no additional specific requirements for resistance management were identified that were not already stated in the registration. o Element 7. We oppose the implementation of Element 7: “ List confirmed resistant species in separate table and list effective or recommended rates for these weeds with the table.” While Monsanto supports making information about weeds resistant to the Monsanto Dicamba Products available to stakeholders and farmers, there are a number of communication challenges. For example, if this information were to be listed on product labels, the label could be out of date as soon as a new resistant species were identified. In many cases labels are not updated for a period of time permitted by EPA. As such, product users may not gain complete awareness to the full list of weeds to have developed resistance in the interim from the product label itself. If EPA suggests a separate website listing, it is unclear if this website then becomes a legal extension of the label. Furthermore, since the occurrence of confirmed resistant weeds are reported to EPA under 6(a)(2) reporting by various registrants, it is unclear how different dicamba registrants would provide consistent information about resistant weeds to growers and stakeholders. o Element 9. Monsanto is also concerned about the implementation of Element 9: “Provide growers with: Resistance Management Plan, Remedial Action Plan, Educational materials on resistance management.” Monsanto is willing to make such information available to farmers to develop and implement appropriate resistance management plans. Furthermore, farmer decisions on purchasing herbicides for their intended use is based on a variety of weed management decisions and practices. Generalized considerations for weed management and management of herbicide-resistant weeds are encompassed by best management practices and are appropriately provided to farmers through university extension, company informational, and other educational materials. However, each farmer must make these decisions based on their crop, the weeds present, equipment, economics, other herbicide choices, and many other farm or field specific needs/requirements. Therefore providing information pertinent to a resistance management plan is an appropriate approach to this element. In summary, Monsanto is supportive of an appropriate herbicide resistance management plan. Herbicide-resistant management plans that enable and support farmer best management practices, maintain a farmers’ ability to adjust practices to address the challenges of weed 35 management including the management of herbicide-resistant weeds, and allow the use of new tools for weed management, are critical to maintaining our agricultural productivity. B. Monsanto Product Recommendations Will Maintain the Durability of M1691 Herbicide Maintaining the durability of M1691 Herbicide was carefully considered as the herbicide resistance management plan was developed. Although Monsanto believes that the risk of resistance to M1691 Herbicide in many weed species is inherently low due to the high probability that the herbicide would be used as part of a diversified weed management program including multiple the use of other herbicides and a low frequency of resistant alleles present in weed species/populations, product and weed control program recommendations will strongly reinforce and support best management practices for diversified weed management to reduce selection for – and manage existing populations of – weeds resistant to this herbicide. Dicamba, the active ingredient in M1691 Herbicide, has been used for more than 50 years in agriculture with only two known resistant species in the United States, and a total of five resistant species worldwide. Furthermore, auxin-based classes of herbicides are known for having low rates of resistance. Possible reasons for low level of resistance are: resistant alleles occur infrequently; the manifestation of resistance is likely multi-genic; and resistance genes are associated with a fitness penalty (Mithlia et. al. 2011). Notably, M1691 Herbicide will be used in systems that include the use of other herbicides. Since M1691 Herbicide is primarily used to control broadleaf weeds, other herbicides will be needed in the overall system to control other species, including grasses. Recommendations will include the use of soil-applied residual and foliar applied post-emergent herbicides, as appropriate to provide multiple effective mechanisms of action on target weed species. Dicamba tolerance is stacked with glyphosate tolerance in the soybean and with glyphosate and glufosinate tolerance in the cotton product thus enabling, where appropriate, the use of these herbicides to manage weeds. Furthermore, M1691 Herbicide will be part of the Roundup Ready plus program. This program currently includes education, recommendation, and incentive components that have contributed to the diversification of weed management programs including increased in the use of soil applied residual herbicides and herbicides with different mechanisms of action to control weeds – both considered Best Management Practices for weed management and to manage and/or delay the selection of herbicide-resistant weeds. The product label will also require a minimum single application use rate of 0.5 lb a.e. to further reduce potential selection for resistant weeds. 36 VI. ENABLING TANK MIXES IS CRITICAL FOR EFFECTIVE WEED RESISTANCE MANAGEMENT AND ENVIRONMENTAL STEWARDSHIP A. Tank Mixing Provides Multiple Benefits to Growers The benefits of herbicide tank mixtures can be grouped into three general categories: weed resistance management, economics and sustainability, and efficacy and productivity. Weed Resistance Management The use of herbicide tank mixes facilitates one of the most practical, convenient and reliable methods to ensure use of multiple herbicide sites of action in weed control programs, which is a fundamental component of weed resistance management best practices (Beckie 2006, Dill et al. 2008, Gustafson 2008, Green and Owen 2011, Norsworthy et al. 2012). Recent research indicates that herbicide tank mixtures are more effective at delaying the development of herbicide resistance or managing herbicide resistant populations than herbicide rotation (Diggle et al., 2003, Beckie 2006, Beckie and Reboud 2009, Evans et al 2015). Tank mixtures also facilitate use of herbicide combinations to provide both foliar and residual activity to control emerged weeds and subsequent flushes of newly germinating weeds in a single application (Spaunhorst 2012). The use of tank mixtures with herbicides that provide residual control can minimize weed species shifts and resistance selection pressure (Dill et al. 2008, Gustafson 2008) and can extend the effective window of application for subsequent herbicide applications to be made at proper weed growth stage timings (VanGessel et al. 2000). Herbicide tank mixtures are a key enabler for the use of multiple sites of action and effective and sustainable weed management systems. Economics and Sustainability Tank mixing herbicides provides convenience, efficiency and economic benefits for growers and applicators. “The use of mixtures decreases the number of trips across the field, saves fuel, decreases labor, reduces equipment wear and lessens the mechanical damage to the crop and soil. These economical aspects are increasingly important in the U.S. where the trend is toward larger farms, fewer workers, and higher application costs” (Green and Bailey 1987). Tank mixing herbicides that would otherwise be applied in sequential applications also facilitates time management across farm operations with multiple fields and crops that require management practices within the same windows of time. This enables growers to manage more acres more efficiently with existing equipment and resources. Environmental benefits can also be realized with fewer applications across the field. As an example, each application trip eliminated across a single 100-acre field would conserve between 1,000 – 2,000 gallons of water carrier and 2.5-20 gallons of a single adjuvant at 10 - 20 gallon per acre spray volumes and typical adjuvant use rates of 0.25 to 1.0% v/v concentration. Less fuel use through fewer applications trips also corresponds to reduced CO2 emissions. 37 Efficacy and Productivity Tank mixing herbicides enables the farmer to choose the best combination and ratios of herbicide components for his individual, unique weed control needs and conditions. Herbicide mixtures improve control and provide a broader weed control spectrum including control of herbicideresistant weeds (Beckie 2006, Green and Owen 2011, Barnett 2013). Proper timing of post-emergence weed control applications based on weed height is critical to optimize weed control and minimize yield losses (Gower et al., 2002, Gower et al. 2003, Dalley et al. 2004). Herbicide programs that contain tank mix combinations and are applied to smaller plants earlier in the season provide generally greater soybean yield than sequential application programs (Vink et al. 2012, Riley and Bradley 2014). Sequential application of multiple herbicide program components requires an interim period of time between applications. The planned time period can often be extended or the sequential application missed altogether due to weather, time constraints or other unplanned issues. Uncontrolled weeds can rapidly grow beyond the planned target growth stage reducing yield potential (Carey and Kells 1995). “Palmer amaranth plants can grow 2-3 inches per day under good growing conditions. The effectiveness of most foliar-applied herbicides dramatically decreases when Palmer amaranth plants are taller than 4 inches.” (Hager 2014). Compared to sequential applications, tank mixing allows for more consistently optimal application timings for multiple program components and reduces potential for mistimed or missed applications, inadequate weed control and yield loss. B. EPA Should Consider Specific Tank Mixes Where Requested Monsanto intends to submit additional data to demonstrate that tank mixing the Monsanto Dicamba Products with other Active Ingredients would be consistent with EPA’s risk assessment. The practice of tank-mixing herbicides is an important tool used in agriculture that gives growers flexibility in a comprehensive weed resistance management program to include multiple modes of action while expanding the spectrum of weed control. Specifically, in the proposed registration of M1691 Herbicide for uses on Dicamba-Tolerant Cotton and Soybean, EPA has expressed concern about potential “synergistic” effects from tank mixing dicamba with other active ingredients With regard to EPA’s concern about “synergistic effects” Monsanto has evaluated various sources of information for reliability and relevance to an endangered species and non-target organism assessment. This evaluation resulted in the following conclusions:   True synergistic interactions between chemicals are rare and are generally not predicted to occur under low exposure scenarios (Verbruggen and van den Brink 2010, Cedergreen, 2014). For synergistic interactions to occur among herbicides, or other pesticides, in the environment, interacting substances have to co-occur at high enough levels to exceed the toxic threshold for each product (Cedergreen, 2014, COT, 2002; Kortenkamp and Altenburger; 2011). The NAS panel report (NRC 2013, p. 134) is consistent on this point by also stating, “…components do not need to be considered when present at concentrations below their toxic thresholds.” Effects should only be assessed with biologically relevant endpoints (i.e., growth and survival for non-target plants (NTPs)). 38   The buffer distance of 110 feet for M1691 Herbicide particle drift is being implemented to keep exposure levels below the threshold effect level (NOAER) for non-target plants – and therefore would also protect against potential synergistic effects. When evaluating tank mixtures it’s important that the conclusions of the data are limited to application rates relevant to the threatened and endangered species analysis. Conclusions of synergy at field rates, often the focus of literature and patent data, are of little use and not appropriate in identifying and quantifying synergy at low, environmentally relevant concentrations outside of the spray application area. An acceptable statistical test must be used to distinguish whether the response produced by a dose combination is larger than that predicted by the “no-interaction” hypothesis. In sum, we believe that EPA should be willing to consider specific tank mixtures where requested. VII. EPA’S ANALYSIS IS TECHNICALLY AND SCIENTIFICALLY SOUND A. EPA’s Human Health Risk Assessment Is Highly Conservative and Amply Demonstrates a Reasonable Certainty of No Harm EPA’s Human Health Risk Assessment (HHRA) demonstrates that there is a reasonable certainty of no harm to the general public, including infants and children, from the proposed new uses of dicamba on dicamba-tolerant soybean and cotton.24 However, EPA’s decision to establish a chronic reference dose (cRfD) for dicamba of 0.04 mg/kg/day based on a Point of Departure (POD) of 4 mg/kg/day determined from the 2-generation rat reproduction study with the DCSA metabolite is extremely conservative and appears to lead to inconsistencies both within the HHRA and between the HHRA and Environmental Fate and Ecological Risk Assessment documents. This decision is also inconsistent with (i.e., more conservative than) the conclusions of the Joint FAO/WHO Meeting on Pesticide Residues (JMPR). Section 5.1.4 of the HHRA states, “Based on available toxicity studies and structural similarities, HED considers the parent and all three metabolites [DCSA, DCGA and 5-OH dicamba] to be of comparable toxicity.” This is similar to the JMPR (2010) conclusion that “DCSA and DCGA have toxicity similar to or lower than that of dicamba” while 5-OH dicamba “appears to be of lower toxicity than the parent.” In contrast, Section 4.3 of the HHRA indicates that DCSA is approximately 12-fold more toxic to offspring than dicamba acid. The latter statement is apparently based on the most recent EPA conclusions regarding the multi-generation rat reproduction studies with both compounds that were conducted at different times in different laboratories. Although full details are not available, this conclusion appears to be based, at least in part, on different Agency approaches used in evaluating or re-evaluating pup weights. Slight decreases in pup weight or pup weight gain relative to concurrent control were apparently reported in only one generation of both studies at 500 ppm. For DCSA, F1 pup weights at 500 ppm were lower than the concurrent control but were similar to those observed in the F2 controls as well as the laboratory’s historical control values and were thus not considered to be a treatment related adverse effect. Although JMPR agreed with this conclusion, EPA apparently did not and instead concluded that 500 ppm was a Lowest Observed Adverse Effect Level 24 M1768 and M1769 herbicides would also be covered under the current HHRA for dicamba. 39 (LOAEL). This resulted in a No Observable Adverse Effect Level (NOAEL) of 50 ppm, the next lowest concentration tested. In contrast, for dicamba, EPA apparently recently concluded that the decreased pup weights reported in one generation at both 500 and 1500 ppm were not treatment related due to the values being comparable to historical control data. EPA thus raised the dicamba offspring NOAEL from 500 ppm to 1500 ppm. It is not clear why comparisons to historical control data were acceptable for dicamba but not for DCSA. However, the end result is that although both the DCSA and dicamba studies appear to have similar marginal responses at 500 ppm, EPA conclusions regarding offspring NOAELs for these molecules are now quite different, 50 ppm and 1500 ppm, respectively. This does not appear to be consistent with the EPA conclusion that the two molecules exhibit comparable toxicity. The EPA decision to reduce the chronic reference dose (cRfD) for dicamba from 0.45 to 0.04 mg/kg/day and to use this value for dietary risk assessment also does not appear to be justified. DCSA is only a very minor metabolite in conventional crops. Accordingly, the primary source for DCSA in the diet from the proposed uses will be from use of dicamba on dicamba-tolerant soybean (cotton consumption is negligible). However, based on the EPA analysis, potential residues in soybean represent only a very small percentage of the total dietary intake of dicamba. As a result, the vast majority of dietary exposure will be to parent dicamba, not DCSA. Therefore, it would seem more appropriate to assess the risks from these residues utilizing a cRfD based on dicamba data rather than an 11-fold lower cRfD based on a marginal response seen in a DCSA study. The cRfD of 0.04 mg/kg/day proposed by EPA for use in dietary risk assessment is much lower than the value recommended by JMPR. JMPR concluded that the NOAEL for both the dicamba and DCSA rat reproduction studies was 500 ppm (~35 mg/kg/day for dicamba and ~37 mg/kg/day for DCSA). JMPR then concluded that an Acceptable Daily Intake (ADI) of 0.3 mg/kg/day was appropriate to characterize potential risks to both dicamba and its metabolites. This value was based on the NOAELs from the rabbit teratology and rat reproduction studies with dicamba and a 100-fold uncertainty factor. Finally, the use of a Point of Departure (POD) of 4 mg/kg/day from the DCSA rat reproduction study for determining the cRfD for dicamba seems inconsistent with information included in the Second Addendum to the Environmental Fate and Ecological Risk Assessment (March 24, 2016). According to that document, EPA has conducted a benchmark dose analysis of the DCSA reproduction study and concluded that the threshold value for the NOAEL would be 8 mg/kg/day and that the lower 95% confidence limit on the benchmark dose resulting in a 5% change from background (BMDL5) would be 34.9 mg/kg/day. It is not clear why these analyses were not summarized and utilized in the selection of the POD and cRfD in the HHRA. B. EPA’s Occupational Risk Assessment Demonstrates Ample Protection From Occupational Exposure EPA’s occupational risk assessment of dicamba DGA salt demonstrates that the proposed label restrictions are adequately protective of human health. For occupational mixer, loader and applicator assessments, EPA’s selection of relevant exposure scenarios, baseline personal protective equipment (i.e., long-sleeved shirt, long pants, shoes and socks, but no respiratory protection), an application rate consistent with the proposed label, best available unit exposure 40 data, ExpoSAC Policy 9.1-consistent assumptions regarding number of acres treated per day, relevant exposure durations, toxicological endpoint and level of concern result in calculations of sufficient margins of exposure. For the occupational post-application exposure assessment, potential inhalation exposure to individuals performing post-application activities in previously treated fields would be expected to be less than inhalation exposure related to occupational activities involving mixing, loading and applying. Therefore, EPA’s occupational risk assessment of dicamba DGA salt is adequately protective of health.25 C. The Proposed Label Is Protective of Non-Target Susceptible Plants The Proposed Dicamba Label sufficiently protects sensitive crops from the offsite movement of the Monsanto Dicamba Products. As demonstrated in sections III.B-C, a downwind buffer of 110 feet is more than sufficiently protective of non-target organisms, which include sensitive crops. First, volatility data provided by Monsanto on April 12, 2016, demonstrate that volatility is not a major component of offsite movement; additional evidence demonstrates that there should be no concern of off-site movement due to volatility, and that a downwind buffer of 110 feet is more than adequate to protect non-target organisms from potential adverse effects when the mandatory application requirements described in the draft M1691 Herbicide label are instituted. Second, Monsanto previously submitted multiple lines of evidence to support that a downwind buffer of 110 feet will be protective of potential adverse effects to non-target organisms (MRID 49292801, MRID 49424601). However, as an additional layer of protection for sensitive crops, the Proposed Label for M1691 Herbicide specifies the following application restrictions: “Do not apply under circumstances where spray drift may occur to food, forage, or other plantings that might be damaged or the crops thereof rendered unfit for sale, use or consumption. Avoid contact of herbicide with foliage, green stems, exposed non-woody roots of crops, and desirable plants, including beans, cotton, flowers, fruit trees, grapes, ornamentals, peas, potato, soybean, sunflower, tobacco, tomato, and other broadleaf plants, because severe injury or destruction may result, including plants in a greenhouse. Small amounts of spray drift that may not be visible may injure susceptible broadleaf plants. Applicators are required to ensure that they are aware of the proximity to sensitive areas, and to avoid potential adverse effects from off-target movement of M1691 Herbicide. The applicator must survey the application site for neighboring sensitive areas prior to application. The applicator also should consult sensitive crop registries for locating sensitive areas where available. Failure to follow the requirements in this label could result in severe injury or destruction to desirable sensitive broadleaf crops and trees when contacting their roots, stems or foliage. Specifically, commercially grown tomatoes and other fruiting vegetables (EPA crop group 8), cucurbits (EPA crop group 9), and grapes are sensitive to dicamba. In order to prevent unintended damage from any drift of this product, do not apply this product when the wind is 25 EPA’s current occupational exposure assessment for use of M1691 Herbicide on dicambatolerant cotton and soy would also cover occupational exposure assessments for similar use patterns of other dicamba DGA salt formulations containing VaporGrip™ technology, M1768 Herbicide (EPA Reg. No. 524-617) and M1769 Herbicide (EPA Reg. No. 524-616). 41 blowing toward adjacent commercially grown sensitive crops.” Extensive data further illustrates that use of the Monsanto Dicamba Products will not have a negative impact on yield of other crops since other crops may be exposed to higher rates of dicamba than soybean without having a negative impact on growth or yield (MRID 48892302). VIII. REFERENCES Andrews, D. 2000. Inconsistency of the bootstrap when a parameter is on the boundary of the parameter space. Econometrica 68:399-405. Auch, D.E. and W.E. Arnold. 1978. Dicamba use and injury on soybeans Glycine max in South Dakota. Weed Science 26:471-475. Baldwin, F.L. and T.L. Baldwin. 2002. Impact of herbicide resistant crops in the Americas - A southern prospective. Pages 650-654 in Thirteenth Australian Weeds Conference, Council of Australian Weed Societies Inc., Perth, Western Australia. Barnett, K. A., T. C. Mueller and L. E. Steckel. 2013. Glyphosate-resistant giant ragweed (Ambrosia trifida) control with glufosinate or fomsafen combined with growth regulator herbicides. Weed Tech. 27:454-458. BASF Corporation. 2008. Clarity® herbicide label. BASF Corporation, Research Triangle Park, North Carolina. Beckie, H. J. 2006. Herbicide-resistant weeds: management tactics and practices. Weed Tech. 20:793-814. Beckie, H.J. and X. Reboud. 2009. Selecting for weed resistance: herbicide rotation and mixture. Weed Tech. 23:363-370. Byker, H.P., N. Soltani, D.E. Robinson, F.J. Tardif, M.B. Lawton and P.H. Sikkema. 2013. Control of glyphosate-resistant horseweed (Conyza canadensis) with dicamba applied preplant and postemergence in dicamba-resistant soybean. Weed Technology 27:492-496. Carey, J.B. and J.J. Kells. 1995. Timing of total postemergence herbicide applications to maximize weed control and corn (Zea mays) yield. Weed Tech. 9:356–361. Cedergreen N. 2014. Quantifying synergy: a systematic review of mixture toxicity studies within environmental toxicology. PLoS One. 2;9(5):e96580. doi: 10.1371/journal.pone.0096580. COT. 2002. Risk assessment of mixtures of pesticides and similar substances. London: Committee of toxicity of chemical in food, consumer products and the environment. Available from: https://cot.food.gov.uk/sites/default/files/cot/reportindexed.pdf . Last accessed April 21, 2016. 42 CTIC. 2011. Top 10 conservation tillage benefits. Conservation Technology Information Center, West Lafayette, Indiana. Dalley, C.D., J.J. Kells and K.A. Renner. 2004. Effect of glyphosate application timing and row spacing on corn (Zea mays) and soybean (Glycine max) yields. Weed Tech. 18:165-176. Diggle, A. J., P.B. Neve and F.P. Smith. 2003. Herbicides used in combination can reduce the probability of herbicide resistance in finite weed populations. Weed Res. 43:371-382. Dill, G. M., C. A. CaJacob and S. R. Padgette. 2008. Glyphosate-resistant crops: adoption, use and future considerations. Pest Management Sci. 64:326-331. Egan, J.F. and D.A. Mortensen. 2012. Quantifying vapor drift of dicamba herbicides applied to soybean. Environmental Toxicology and Chemistry 31:1023-1031. Evans, J.A., P. J. Tranel, A. G. Hager, B. Schutte, C. Wu, L.A. Chatham and A.S. Davis. 2015. Managing the evolution of herbicide resistance. Pest Management Sci. 72:74-80. Everitt, J.D. and J.W. Keeling. 2007. Weed control and cotton (Gossypium hirsutum) response to preplant applications of dicamba, 2,4-d, and diflufenzopyr plus dicamba. Weed Technology 21:506-510. Everitt, J.D. and J.W. Keeling. 2009. Cotton growth and yield response to simulated 2,4-D and dicamba drift. Weed Technology 23:503-506. Flessner, M.L., J.S. McElroy, J.D. McCurdy, J.M. Toombs, G.R. Wehtje, C.H. Burmester, A.J. Price and J.T. Ducar. 2015. Glyphosate-resistant horseweed (Conyza canadensis) control with dicamba in Alabama. Weed Technology 29:633-640. Gavlick, W.K., D.R. Wright, A. MacInnes, J.W. Hemminhaus, J.K. Webb, V.I. Yermolenka and W. Su. 2016. A method to determine the relative volatility of auxin herbicide formulations. Pages 24-32 in Pesticide Formulation and Delivery Systems. Volume 35. G.R. Goss (ed.). ASTM International, West Conshohocken, Pennsylvania. Gianessi, L. and N. Reigner. 2006. The value of herbicides in U.S. crop production: 2005 update. CropLife Foundation, Washington, D.C. Gower, S.A., M.M. Loux, J. Cardina, and S.K. Harrison. 2002. Effect of planting date, residual herbicide, and postemergence application timing on weed control and grain yield in glyphosatetolerant corn (Zea mays). Weed Tech. 16:488-494. Gower, S.A., M.M. Loux, J. Cardina, S.K. Harrison, P.L. Sprankle, N.J. Probst, T.T. Bauman, W. Bugg, W.S. Curran, R.S. Currie, R.G. Harvey, W.G. Johnson, J.J. Kells, M.D.K. Owen, D.L. Regehr, C.H. Slack, M. Spaur, C.L. Sprague, M. VanGessel, and B. G. Young. 2003. Effect of postemergence glyphosate application timing on weed control and grain yield in glyphosateresistant corn: results of a 2-yr multistate study. Weed Tech. 17:821-828. 43 Green, J. M. and S.P. Bailey. 1987. Chapter 4 in Methods of applying herbicides. WSSA monograph 4. C.G. McWhorter and M. R. Gebhardt eds. Green, J.M. and M.D.K. Owen. 2011. Herbicide-resistant crops: utilities and limitations for herbicide-resistant weed management. J. Agric. Food Chem. 59:5819-5829. Gustafson, D. I. 2008. Sustainable use of glyphosate in North American cropping systems. Pest Management Sci. 64:409-416. Hager. A.G. 2014. Management of Palmer amaranth in Illinois. University of Illinois, Weed Science Extension update April 2014. Hake, S.J., T.A. Kerby and K.D. Hake. 1996. Preparation for the new crop season - Fall/winter. Pages 6-14 in Cotton Production Manual. S.J. Hake, T.A. Kerby, and K.D. Hake (eds.). University of California Division of Agriculture and Natural Resources, Davis, California. Hall, P. and H. Miller. 2010. Bootstrap confidence intervals and hypothesis tests for extrema of parameters. Biometrika 97:881-892. Heap, I. The International Survey of Herbicide Resistant Weeds. Online. Internet. Accessed 2013. Available www.weedscience.com Hoberg JR. 1993. Dicamba Technical – Determination of the Effects on Seed Germination, Seedling Emergence and Vegetative Vigor of Ten Plant Species. BASF Reg DOC. No. 1993/5257. MRID 42846301. Johnson, B., B. Young, J. Matthews, P. Marquardt, C. Slack, K. Bradley, A. York, S. Culpepper, A. Hager, K. Al-Khatib, L. Steckel, M. Moechnig, M. Loux, M. Bernards and R. Smeda. 2010. Weed control in dicamba-resistant soybeans. Plant Management Network, St. Paul, Minnesota. Kelley, K.B., L.M. Wax, A.G. Hager and D.E. Riechers. 2005. Soybean response to plant growth regulator herbicides is affected by other postemergence herbicides. Weed Science 53:101-112. Kortenkamp, A. and R. Altenburger. 2011. Toxicity of combined exposure to chemicals. Mixture Toxicity, LInking Approaches from Ecological and Human Toxicology. C.A.M. Van Gestel, M.J. Jonker, J.E. Kammenga, R. Laskowski, and C. Svendsen (eds.). CRC Press, Boca Raton, Florida. Kruger, G.R., W.G. Johnson, S.C. Weller, M.D.K. Owen, D.R. Shaw, J.W. Wilcut, D.L. Jordan, R.G. Wilson, M.L. Bernards and B.G. Young. 2009. U.S. grower views on problematic weeds and changes in weed pressure in glyphosate-resistant corn, cotton, and soybean cropping systems. Weed Technology 23:162-166. 44 Lanini, W.T. and V. Carrithers. 2000. Simulated drift of herbicides on grapes, tomatoes, cotton and sunflower. 2000 Proceedings of the California Weed Science Society, Monterey, California. 52:107-110. Logan, S.T., B.G. Young, J.M. Young, S. Seifert-Higgins and S.M. Allen. 2013. Dicamba in a residual system for glyphosate-resistant waterhemp control in soybean. 2013 North Central Weed Science Society Proceedings 68:17. Loux, M.M., D. Doohan, A.F. Dobbels, W.G. Johnson, G.R.W. Nice, T.N. Jordan and T.T. Bauman. 2010. Weed control guide for Ohio and Indiana. Bulletin 789. Purdue University and Ohio State University Extension, West Lafayette, Indiana. Marple, M.E., K. Al-Khatib and D.E. Peterson. 2008. Cotton injury and yield as affected by simulated drift of 2,4-D and dicamba. Weed Technology 22:609-614. Moechnig, M., D.L. Deneke and L.J. Wrage. 2010. Weed control in cotton: 2010. South Dakota State University, Brookings, South Dakota. Monsanto. 2012. Unpublished grower survey data performed by a contracting agency. Monsanto Company, St. Louis, Missouri. Morton, H.L., J.O. Moffett and R.H. McDonald. 1972. Toxicity of herbicides to newly emerged honey bees. Environmental Entomology 1:102-104. Norsworthy, J. K. , Ward, S. M., Shaw, D.R., Llewellyn, R. S. , Nichols, R. L., Webster, T. M., Bradley, K. W., Frisvold, G., Powles, S. B., Burgos, N. R., Witt, W. W., Barrett, M. 2012. Reducing the risks of herbicide resistance: Best Managment Practices and Recommendations. Weed Science Special Issue: 31-62. http: //wssajoumals.org/doi/pdf/10.1614/WS-D-11-00155.1. NRC. 2013. Assessing risks to endangered and threatened species from pesticides. Natural Resources Council, National Academy of Sciences, Washington, D.C. http://www.nap.edu/catalog.php?record_id=18344 [Accessed April 28, 2016]. Ou, J., C.R. Thompson, M. Jugulam and P.W. Stahlman. 2010. Kochia response to pre and post applied dicamba. 2010 North Central Weed Science Society Conference Proceedings 65:92. Owen, L.N., L.E. Steckel, C.H. Koger, C. Main and T.C. Mueller. 2009. Evaluation of Spring and Fall burndown application timings on control of glyphosate-resistant horseweed (Conyza canadensis) in no-till cotton. Weed Technology 23:335-339. Riley, E. B. and K. W. Bradley. 2014. Influence of Application Timing and Glyphosate TankMix Combinations on the Survival of Glyphosate-Resistant Giant Ragweed (Ambrosia trifida) in Soybean. Weed Tech. 28:1-9. 45 Salas, R.A., N.R. Burgos, P.J. Tranel, S. Singh, L. Glasgow, R.C. Scott and R.L. Nichols. 2016. Resistance to PPO-inhibiting herbicide in Palmer amaranth from Arkansas. Pest Management Science 72:864-869. Schlichenmayer, A.A., T.C. Shauck, S.A. Riley, C.F. Page and R.J. Smeda. 2011. Soil persistence of dicamba. 2011 North Central Weed Science Society Proceedings 66:28. Sosnoskie, L.M., J.M. Kichler, R.D. Wallace and A.S. Culpepper. 2011. Multiple resistance in Palmer amaranth to glyphosate and pyrithiobac confirmed in Georgia. Weed Science 59:321325. Sosnoskie, L.M., A.S. Culpepper, L.B. Braxton and J.S. Richburg. 2015. Evaluating the volatility of three formulations of 2,4-D when applied in the field. Weed Technology 29:177184. Southeast Farm Press. 2009. Resistant pigweed control programs updated. Southeast Farm Press, Prism Business Media, Lewisville, Texas. http://southeastfarmpress.com/management/resistantpigweed-control-programs-updated [Accessed April 10, 2013]. Spaunhorst, D.J., B.G. Young and W.G. Johnson. 2010. Palmer amaranth (Amaranthus palmeri) control with soil-applied herbicide programs which contain dicamba, isoxaflutole, and 2,4-D. 2010 North Central Weed Science Society Conference Proceedings 65:140. Spaunhorst, D.J., S. Seifert-Higgins, C.M. Mayo, E.B. Riley and K.W. Bradley. 2012. Programs for the management of glyphosate-resistant waterhemp and giant ragweed in dicambaresistant soybean. Proc. North Central Weed Science Vol. 67 Steckel, L. 2013. In fight against resistance – New weed control technology promising, deltafarmpress.com. Steckel, L. 2016. PPO/glyphosate resistance in Palmer not an overnight event. Delta Farm Press, Kansas City, Kansas. http://deltafarmpress.com/soybeans/ppoglyphosate-resistance-palmer-notovernight-event [Accessed May 10, 2016]. Steckel, L.E., C.C. Craig and R.M. Hayes. 2006. Glyphosate-resistant horseweed (Conyza canadensis) control with glufosinate prior to planting no-till cotton (Gossypium hirsutum). Weed Technology 20:1047-1051. Thompson, M. A., L. E. Steckel, A. T. Ellis, and T. C. Mueller. 2007. Cotton Tolerance to early Preplant Application of 2,4-D Ester, 2,4-D Amine, and Dicamba. Weed Technology 21:882-885. University of Arkansas. 2011. Recommended chemicals for weed and brush control. University of Arkansas, Cooperative Extension Service, Division of Agriculture, Fayetteville, Arkansas. University of Illinois. 2008. 2008 Illinois agricultural pest management handbook. University of Illinois Extension, Champaign, Illinois. 46 U.S. EPA. 2013. EFED Environmental Risk Assessment of Proposed Label for Enlist (2,4-D Choline Salt), New Uses on Soybean with DAS 68416-4 (2,4-D Tolerant) and Enlist (2,4-D + Glyphosate Tolerant) Corn and Field Corn. EPA-HQ-OPP-2014-0195-0002. U.S. Environmental Protection Agency, Washington, D.C. U.S. EPA. 2016. Biological and Economic Analysis Division (BEAD) review of the benefits document submitted by Monsanto to support the registration of the use of dicamba on dicamba tolerant soybean and cotton. EPA-HQ-OPP-2016-0187-0012. U.S. Environmental Protection Agency, Washington, D.C. VanGessel, M.J., A.O. Ayeni, and B.A. Majek. 2000. Optimum glyphosate timing with or without residual herbicides in glyphosate-tolerant soybean (Glycine max) under full-season conventional tillage. Weed Technology 14:140-149. Verbruggen, E.M.J. and P.J. van den Brink. 2010. Review of recent literature concerning mixture toxicity of pesticides to aquatic organisms. RIVM Report 601400001/2010. RIVM, National Institute for Public Health and the Environment, Bilthoven, the Netherlands. Vink, J. P., N. Soltani, D. E. Robinson, F. J. Tardif, M. B. Lawton and P. H. Sikkema. 2012. Glyphosate-Resistant Giant Ragweed (Ambrosia trifida) Control in Dicamba-Tolerant Soybean. Weed Tech. 26:422-428. Wall, D.A. 1994. Potato (Solanum tuberosum) response to simulated drift of dicamba, clopyralid and tribenuron. Weed Science 42:110-114. Wax, L.M., L.A. Knuth and F.W. Slife. 1969. Response of soybeans to 2,4-D, dicamba, and picloram. Weed Science 17:388-393. Weidenhamer, J.D., G.B. Triplett and F.E. Sobotka. 1989. Dicamba injury to soybean. Agronomy Journal 81:637-643. Willis, J.B., C.D. Kamienski, M.S. Malik and S. Seifert-Higgins. 2012. Dicamba contributes residual weed control to Roundup Ready® 2 Xtend soybean systems. 2012 North Central Weed Science Society Proceedings 67:16. Zollinger, R. 2006. The glyphosate weeds and crop series: Biology and management of wild buckwheat. GWC-10. Purdue Extension, West Lafayette, Indiana. 47 IX. APPENDIX: TECHNICAL CORRECTIONS Docket ID No. Document Name & Text Section Page Comment/Data Needed EPAHQOPP201601870002 Addendum to Dicamba Diglycolamine Salt (DGA) and its Degradate, 3,6dichlorosalicylic ac id (DCSA) Section 3 Risk Assessment: Refined Endangered Species Assessment for Proposed New Uses on HerbicideTolerant Soybean and Cotton in 16 states (Arkansas, Illinois, Iowa, Indiana, Kansas, Louisiana, Minnesota, Mississippi, Missouri, 2 Replace: “seeds” with “crops.” “EPA has a specific process based on sound science that it follows when assessing risks to listed species for pesticides like dicamba that will be used on seeds that have been genetically modified to be tolerant to the pesticide.” Dicamba is not a seed treatment. It will not be used on seeds that have been genetically modified to be tolerant to the pesticide. It will be used either by (1) application to soil either, pre-plant or pre-emergence of the dicamba tolerant crop, or (2) by foliar application over the dicamba tolerant crop. 48 Docket ID No. EPAHQOPP201601870002 Document Name & Text Section Nebraska, North Dakota, Ohio, Oklahoma, South Dakota, Tennessee, and Wisconsin). Addendum to Dicamba Diglycolamine Salt (DGA) and its Degradate, 3,6dichlorosalicylic ac id (DCSA) Section 3 Risk Assessment: Refined Endangered Species Assessment for Proposed New Uses on HerbicideTolerant Soybean and Cotton in 16 states (Arkansas, Illinois, Iowa, Indiana, Kansas, Louisiana, Minnesota, Mississippi, Missouri, Nebraska, North Dakota, Ohio, Page Comment/Data Needed “Similar modeling of 7 DCSA residues, which are formed inside the tolerant-soybean and tolerant-cotton plants through plant metabolism, is not feasible at this time due to a lack of sufficient data tracking DCSA residues in plant tissues over time to ascertain degradation rates.” Analysis needed: Data are available and decline of DCSA residues should be considered in the risk assessment. Decline of DCSA residues has been reported for three sites in MRID 48644205: Determination of Dicamba Residue Decline in Forage after Application to Dicamba-Tolerant Soybean MON 87708 × MON 89788. Although a half-life has not been calculated, residues appear to have declined by approximately 50% or more within 7 days after application. 49 Docket ID No. EPAHQOPP201601870002 Document Name & Text Section Oklahoma, South Dakota, Tennessee, and Wisconsin) Addendum to Dicamba Diglycolamine Salt (DGA) and its Degradate, 3,6dichlorosalicylic ac id (DCSA) Section 3 Risk Assessment: Refined Endangered Species Assessment for Proposed New Uses on HerbicideTolerant Soybean and Cotton in 16 states (Arkansas, Illinois, Iowa, Indiana, Kansas, Louisiana, Minnesota, Mississippi, Missouri, Nebraska, North Dakota, Ohio, Oklahoma, South Dakota, Tennessee, “Mass of soybean plants consumed per day = 22.44 kcal/day/(0.63 kcal/gX0.47 AE) = 75.79 g/day.” Page Comment/Data Needed Pg 15 Replace: Energy requirements for soybean plants should be 2.2 kcal/g (broadleaf forage). 50 Docket ID No. EPAHQOPP201601870002 Document Name & Text Section and Wisconsin) Addendum to Dicamba Diglycolamine (DGA) Salt and its Degradate, 3,6dichlorosalicylic acid (DCSA) Section 3 Risk Assessment: Refined Endangered Species Assessment for Proposed New Uses on HerbicideTolerant Cotton and Soybean in 16 states (Arkansas, Illinois, Iowa, Indiana, Kansas, Louisiana, Minnesota, Mississippi, Missouri, Nebraska, North Dakota, Ohio, Oklahoma, South Dakota, Tennessee, and Wisconsin) Page Comment/Data Needed “The screening level 24 risk assessment found that DCSA residues in arthropods in cotton fields (based on the empirical residues in broadleaf plant tissues and extrapolated via the Kenaga nomogram to residues in arthropods) would not exceed any chronic levels of concern for mammals. The analysis of the Louisiana Black Bear’s recovery plan described above indicates that in soybean fields, the attractive food source in these fields would be soybean grain (seeds). On the basis of this information, the refinement of the soybean screening level assessment was initiated to reflect the Add: “Using arthropods as a food source would not alter the conclusion of the screening level assessment for the Louisiana Black Bear.” 51 Docket ID No. Document Name & Text Section Page Comment/Data Needed 8 Analysis needed: Data are available and decline of DCSA residues should be considered in the risk assessment. Decline of DCSA residues has been reported for three sites in MRID 48644205: Determination of Dicamba Residue Decline in Forage after Application to Dicamba-Tolerant Soybean MON 87708 × MON 89788. Although a half-life has not been calculated, residues appear to have declined by approximately 50% or more within 7 days after application. conservative assumption of exclusive consumption of exposed soybean grain containing the maximum measured DCSA residues.” EPAHQOPP201601870003 Addendum to Dicamba Diglycolamine (DGA) Salt and its Degradate, 3,6dichlorosalicylic acid (DCSA) Section 3 Risk Assessment: Refined Endangered Species Assessment for Proposed New Uses on HerbicideTolerant Cotton and Soybean in 7 U.S. States (Alabama, Georgia, Kentucky, Michigan, North Carolina, South Similar modeling of DCSA residues, which are formed inside the tolerant-soybean and tolerant-cotton plants through plant metabolism, is not feasible at this time due to a lack of sufficient data tracking DCSA residues in plant tissues over time to ascertain degradation rates. 52 Docket ID No. Document Name & Text Section Page Comment/Data Needed 77 Add: Rationale should include “streams and creeks.” Carolina, and Texas) EPAHQOPP201601870003 Addendum to Dicamba Diglycolamine (DGA) Salt and its Degradate, 3,6dichlorosalicylic acid (DCSA) Section 3 Risk Assessment: Refined Endangered Species Assessment for Proposed New Uses on HerbicideTolerant Cotton and Soybean in 7 U.S. States; Appendix 2: Listed Species Rationale for NO Effects When Action Area is Limited to Treated Agricultural Field – Accounting for Spray Drift Mitigation Labeling Rationale: “The proposed dicamba DGA uses are not expected to overlap with rivers or other water bodies.” 53 Docket ID No. Document Name & Text Section Page Comment/Data Needed 109 Add: “The U.S. Fish & Wildlife Service concurs that fleshy-fruit gladecress will not occur in row crop agricultural fields.” Restrictions; Lacy elimia EPAHQOPP201601870003 Addendum to Dicamba Diglycolamine (DGA) Salt and its Degradate, 3,6dichlorosalicylic acid (DCSA) Section 3 Risk Assessment: Refined Endangered Species Assessment for Proposed New Uses on HerbicideTolerant Cotton and Soybean in 7 U.S. States; Appendix 2: Listed Species Rationale for NO Effects When Action Area is Limited to Treated Agricultural Field – Accounting for Spray Drift Mitigation Rationale for fleshyfruit gladecress: “Technical consultation with USFWS biologist indicated that this species will not persist in soy or cotton fields due to the competing vegetation.” 54 Docket ID No. Document Name & Text Section Page Comment/Data Needed Labeling Restrictions EPAHQOPP201601870003 Addendum to Dicamba Diglycolamine (DGA) Salt and its Degradate, 3,6dichlorosalicylic acid (DCSA) Section 3 Risk Assessment: Refined Endangered Species Assessment for Proposed New Uses on HerbicideTolerant Cotton and Soybean in 7 U.S. States; Appendix 2: Listed Species Rationale for NO Effects When Action Area is Limited to Treated Agricultural Field – Accounting for Spray Drift Mitigation “The proposed 119 dicamba DGA uses are not expected to overlap with plains.” Replace “plains” with “South Texas Plains vegetation area.” 55 Docket ID No. Document Name & Text Section Page Comment/Data Needed 125 Add: “An in-field buffer will be protective of species at the field margins.” Labeling Restrictions; Ashy dogweed EPAHQOPP201601870003 Addendum to Dicamba Diglycolamine (DGA) Salt and its Degradate, 3,6dichlorosalicylic acid (DCSA) Section 3 Risk Assessment: Refined Endangered Species Assessment for Proposed New Uses on HerbicideTolerant Cotton and Soybean in 7 U.S. States; Appendix 2: Listed Species Rationale for NO Effects When Action Area is Limited to Treated Agricultural Field – Accounting for Spray Drift Hairy rattleweed rationale: “The proposed dicamba DGA uses are not expected to overlap with the margins of cultivated land.” 56 Docket ID No. Document Name & Text Section Page Comment/Data Needed 136-37 Add: “An in-field buffer will be protective of species at the field margins.” Mitigation Labeling Restrictions EPAHQOPP201601870003 Addendum to Dicamba Diglycolamine (DGA) Salt and its Degradate, 3,6dichlorosalicylic acid (DCSA) Section 3 Risk Assessment: Refined Endangered Species Assessment for Proposed New Uses on HerbicideTolerant Cotton and Soybean in 7 U.S. States; Appendix 2: Listed Species Rationale for NO Effects When Action Area is Limited to Treated Agricultural Field – Accounting for Spray Drift Sensitive joint-vetch habitat: “Majority are found in natural tidal marsh habitats, but also a few documented cases of a pocket marsh wetland, edge of a moist soybean field, and a mowed grassy strip between a manmade drainage channel and dirt road.” 57 Docket ID No. Document Name & Text Section Page Comment/Data Needed Mitigation Labeling Restrictions EPAHQOPP201601870004 Addendum to Dicamba Diglycolamine Salt (DOA) and its Degradate, 3,6dichlorosalicylic acid (DCSA) Section 3 Risk Assessment: Refined Endangered Species Assessment for Proposed New Uses on HerbicideTolerant Soybean and Cotton in in 11 U.S. States: (Arizona, Colorado, Delaware, Florida, Maryland, New Mexico, New Jersey, New York, Pennsylvania, Virginia and West “Similar modeling of 8 DCSA residues, which are formed inside the tolerant-soybean and tolerant-cotton plants through plant metabolism, is not feasible at this time due to a lack of sufficient data tracking DCSA residues in plant tissues over time to ascertain degradation rates.” Analysis needed: Data are available and decline of DCSA residues should be considered in the risk assessment. Decline of DCSA residues has been reported for three sites in MRID 48644205: Determination of Dicamba Residue Decline in Forage after Application to Dicamba-Tolerant Soybean MON 87708 × MON 89788. Although a half-life has not been calculated, residues appear to have declined by approximately 50% or more within 7 days after application. 58 Docket ID No. EPAHQOPP201601870004 Document Name & Text Section Virginia). Addendum to Dicamba Diglycolamine Salt (DOA) and its Degradate, 3,6dichlorosalicylic acid (DCSA) Section 3 Risk Assessment: Refined Endangered Species Assessment for Proposed New Uses on HerbicideTolerant Soybean and Cotton in in 11 U.S. States: (Arizona, Colorado, Delaware, Florida, Maryland, New Mexico, New Jersey, New York, Pennsylvania, Virginia and West Virginia). Audubon’s Crested Caracara Page Comment/Data Needed 12 Add: “There is currently no commercial cotton production in the southern part of Florida. Cultivation of B.t.-cotton is not permitted south of Florida State Highway 60 due to potential cross-pollination with a wild relative, G. hirsutum. DT-cotton will therefore not be grown in Palm Beach County since it will only be available commercially as a stacked product that also possesses a B.t. trait and Palm Beach County is south of Highway 60. Since no DT cotton will be grown in proximity to the Audubon Crested Caracara (and thus no dicamba applied to DT cotton there), a no-effect determination is amply supported. Bt Plant-Incorporated Protectants October 15, 2001 Biopesticides Registration Action Document. http://www.epa.gov/oppbppd1/biopesticides/pips/bt_brad2/3ecological.pdf.” 59 Docket ID No. Document Name & Text Section Page Comment/Data Needed EPAHQOPP201601870004 Addendum to Dicamba Diglycolamine (DGA) Salt and its Degradate, 3,6dichlorosalicylic acid (DCSA) Section 3 Risk Assessment: Refined Endangered Species Assessment for Proposed New Uses on HerbicideTolerant Cotton and Soybean in 11 U.S. States; Birds; California Condor; DCSA Assessment for California condor consuming prey that had previously consumed soybean forage 12 Add: “A Chronic Dose-Based RQ of 0.02 does not exceed the LOC of 0.1. Consequently, a ‘no effect’ determination is appropriate for the California condor.” The first step in the refinement process is to calculate DCSA residues in the prey species. Using the assumption that the prey species is represented by a 1000 g mammal and the conservative assumptions that the prey animal feeds exclusively on exposed soybean forage containing the maximum measured residues of 61.1 ppm, EFED calculated the residues based on the following allometric equations (USEPA, 1993): 1000 g mammal prey ingestion rate (dry) = 0.621(1000)0.564 =30.56 g /day 1000 g mammal prey ingestion rate (wet) = 30.56/0.2 = 152.8 60 Docket ID No. Document Name & Text Section Page Comment/Data Needed g/day DCSA residue in prey eating soybean forage/hay 61.1 mg DCSA/kg-food (ww) x 0.1528 kg food/kg-bw = 9.34 mg/kg-bw/day The next step is to calculate the expected daily dose for a typical 10 kg (10000 g, Dunning 1984) condor, the adjusted LD50 value, and the acute dose-based RQ for the condor based on the following allometric equations: Food Intake (wet) = (0.301(10000)0.75)/(10.69)/1000 = 0.97 kg wet/day Dose-based EEC in condor eating large mammal= 9.34 mg/kg wet x 0.97/(10000/1000) = 61 Docket ID No. Document Name & Text Section Page Comment/Data Needed 21-24 Add: “A no effect determination is supported by limited exposure in cotton fields, no exposure in soy fields, and screening-level risk assessment results that identify no risk to mammalian species for exposure in cotton fields.” 0.91 mg/kg-bw/day Avian Chronic Endpoint of 695 mg/kg-diet (from mallard duck study for parent dicamba) modified by ratio of parent dicamba to metabolite DCSA from chronic rat studies (17x) results in Avian chronic NOAEC of 40.88 mg/kg-diet. Chronic Dose-Based RQ = 0.91/40.88 = 0.02 EPAHQOPP201601870004 Addendum to Dicamba Diglycolamine (DGA) Salt and its Degradate, 3,6dichlorosalicylic acid (DCSA) Section 3 Risk Assessment: Refined Endangered "This analysis suggests that if a pronghorn NE for Sonoran pronghorn is feeding in a soybean field there is a potential for a lethal event. Establishing a potential for overlap between species range and the cropped areas 62 Docket ID No. EPAHQOPP201601870005 Document Name & Text Section Species Assessment for Proposed New Uses on HerbicideTolerant Cotton and Soybean in 11 U.S. States; Mammals; Sonoran pronghorn; DCSA assessment proposed for treatment is an important consideration in how likely an exposure event might be for individual pronghorn....” Ecological Risk Assessment for Dicamba DGA Salt and its Degradate, 3,6dichlorosalicylic acid (DCSA), for the Proposed PostEmergence New Use on DicambaTolerant Cotton (MON 87701) “Based on the proposed maximum application rates and exceedances of the Agency’s Levels of Concern (LOCs), at the screening level there is a potential for direct adverse effects to Federally endangered and threatened (referred to hereafter as “listed”) and non-listed birds (acute exposure only), listed vascular aquatic plants, and listed and non-listed terrestrial dicots from the Page Comment/Data Needed 2 Replace “vascular plants” with “non-vascular plants.” 63 Docket ID No. Document Name & Text Section Page Comment/Data Needed 7 Replace “drift from particles volatilized” with “volatilization”. The use of the wording “particles” is not accurate for volatilization” occurring after deposition on soil or plants. The only time that volatilization of particles could occur from drift would be during application. proposed new use.” EPAHQOPP201601870005 EPAHQOPP201601870005 EPAHQ- Ecological Risk Assessment for Dicamba DGA Salt and its Degradate, 3,6dichlorosalicylic acid (DCSA), for the Proposed PostEmergence New Use on DicambaTolerant Cotton (MON 8770 1) Ecological Risk Assessment for Dicamba DGA Salt and its Degradate, 3,6dichlorosalicylic acid (DCSA), for the Proposed PostEmergence New Use on DicambaTolerant Cotton (MON 8770 1) “Therefore an analysis of drift from particles volatilized from the treated field was completed…” BAPMA salt 13 endpoints are indicated Table 7 to have been used for risk assessment for Fathead minnow acute (LC50) endpoint, Daphnia chronic (NOAEC) endpoint, green algae acute and chronic endpoints. Mysid shrimp chronic value used dicamba acid. Comment for consideration: DGA salt endpoints should be used in the assessment. Ecological Risk Assessment for Avian chronic values used is indicated as Replace: Avian chronic values for the mallard duck of 800 mg a.e./kg diet with 695 mg a.e./kg diet for mallard duck that was 13 Table 7 64 Docket ID No. Document Name & Text Section OPP201601870005 Dicamba DGA Salt and its Degradate, 3,6dichlorosalicylic acid (DCSA), for the Proposed PostEmergence New Use on DicambaTolerant Cotton (MON 8770 1) Ecological Risk Assessment for Dicamba DGA Salt and its Degradate, 3,6dichlorosalicylic acid (DCSA), for the Proposed PostEmergence New Use on DicambaTolerant Cotton (MON 8770 1) EPAHQOPP201601870005 Page Comment/Data Needed mallard duck value of 800 mg a.e./kg diet used. “LOC for any size class of mammal or bird (RQs would range from <0.01—0.34; Table 10). Residues in arthropods (as a dietary item for birds and mammals consuming insects that have consumed cotton tissues with DCSA residues) were assumed to follow the Kenaga nomogram relationship between broadleaf plants and arthropods and therefore were considered to contain 4.4 ppm which also actually used. Pg 16 Comment for consideration: DCSA arthropod data is based on Kenega values, which are extremely conservative considering that those values are based on direct sprays, not through ingestion. 65 Docket ID No. Document Name & Text Section Page Comment/Data Needed 31 Replace “XR-T-Jet 11004” with “TTI-11004.” would not result in any exceedances (RQ’s range from 0.11— 0.24)”. EPAHQOPP201601870005 EPAHQOPP201601870005 Ecological Risk Assessment for Dicamba DGA Salt and its Degradate, 3,6dichlorosalicylic acid (DCSA), for the Proposed PostEmergence New Use on DicambaTolerant Cotton (MON 8770 1) Ecological Risk Assessment for Dicamba DGA Salt and its Degradate, 3,6dichlorosalicylic acid (DCSA), for the Proposed PostEmergence New Use on DicambaTolerant Cotton (MON 8770 1) “Pesticides were applied using XR-TJet 11004 nozzle which is the same nozzle proposed for the new dicamba uses on DT cotton.” “Using these 34 assumptions in TerrPlant (total 2 lb ae/A application and a 0.055% runoff fraction), and the most sensitive endpoint of 0.000261 for the NOAEC for soybeans, the maximum RQ is less than the LOC of 1.0 by Replace: “the most sensitive endpoint of 0.000261” with “the most sensitive endpoint of 0.0673 lb a.e./A.” Runoff exposure estimates should be compared to the seedling-emergence endpoint not vegetative-vigor endpoint. 66 Docket ID No. EPAHQOPP201601870005 EPAHQOPP201601870005 EPAHQ- Document Name & Text Section Ecological Risk Assessment for Dicamba DGA Salt and its Degradate, 3,6dichlorosalicylic acid (DCSA), for the Proposed PostEmergence New Use on DicambaTolerant Cotton (MON 8770 1) Ecological Risk Assessment for Dicamba DGA Salt and its Degradate, 3,6dichlorosalicylic acid (DCSA), for the Proposed PostEmergence New Use on DicambaTolerant Cotton (MON 8770 1) Ecological Risk Assessment for a factor of at least 2 (RQs range from <0.1 to 0.48, see Appendix 4).” Table 14 Row for Mammals: “Yes (Chronic)” and “Yes2” “Prior to development of glyphosate-resistant crops, there were no known cases of evolved glypohsate resistant weeds (Dyer, 1994).” Page Comment/Data Needed 35 Replace “Yes (Chronic)” and “Yes2” with “No” and “No.” 36 This sentence should be deleted. There were known cases of Glyphosate-resistant weeds in other locations prior to the introduction of glyphosate-tolerant crops (http://www.weedscience.org). 44 Add table notes for Appendix 2. LAFT and HAFT should be defined as Lowest Average Field Trial and Highest Average 67 Docket ID No. Document Name & Text Section OPP201601870005 Dicamba DGA Salt and its Degradate, 3,6dichlorosalicylic acid (DCSA), for the Proposed PostEmergence New Use on DicambaTolerant Cotton (MON 8770 1) Ecological Risk Assessment for Dicamba DGA Salt and its Degradate, 3,6dichlorosalicylic acid (DCSA), for the Proposed PostEmergence New Use on DicambaTolerant Cotton (MON 8770 1) Memorandum: Dicamba DGA: Second Addendum to the Environmental Fate and Ecological Risk Assessment for Dicamba DGA EPAHQOPP201601870005 EPAHQOPP20160187007 Page Comment/Data Needed Field Trial, respectively. Appendix 4 47-50 Analysis needed: Considering the in-field buffer, spray drift should not be considered in TerrPlant calculations. “EFED concluded that the label should be modified to include language to maintain a 100 to 110 foot downwind buffer when applying at the 0.5 lbs a.e./A 6 Replace “restricting the droplet spectra extra-coarse and ultracoarse” with “…restricting the nozzles to those that do not have greater drift potential than TTI11004 at 63 psi for M1691.” 68 Docket ID No. EPAHQOPP201601870007 EPAHQOPP201601870007 Document Name & Text Section Page Comment/Data Needed salt and its Degradate, 3,6dichlorosalicyclic acid (DCSA) for the Section 3 New Use on DicambaTolerant Soybean application rate and with the described nozzles restricting the droplet spectra extracoarse and ultracoarse.” Dicamba DGA: Second Addendum to the Environmental Fate and Ecological Risk Assessment for Dicamba DGA salt and its Degradate, 3,6dichlorosalicylic acid (DCSA) for the Section 3 New Use on DicambaTolerant Soybean Dicamba DGA: Second Addendum to the Environmental Fate and Ecological Risk Assessment for Dicamba DGA salt and its “Appendix 2 and MRID 47899524” 14 Replace “MRID 47899524” with “MRID 48219901.” MRID 48219901 is an amended version of the MRID 47899524, and contains corrected residue values that are slightly different from the values in the original report. “The empirical data from MRID 47899524 found means and maximums, respectively, of DCSA concentrations of 17.0 and 51.3 ppm, in forage 7-10 days 14 Analysis needed: Mean and Maximum DCSA residue values do not match those of MRID 48219901. Forage Mean / Max 15.8 ppm /47.9 ppm; Hay Mean / Max 30.1 ppm/ 57.1 ppm; Seed Mean / Max 0.055 ppm/ 0.411 ppm. 69 Docket ID No. Document Name & Text Section Page Comment/Data Needed Degradate, 3,6dichlorosalicylic acid (DCSA) for the Section 3 New Use on DicambaTolerant Soybean EPAHQOPP201601870007 EPAHQOPP20160187- following the last application, 32.2 and 61.1 ppm in hay 13-15 days following the last application and 0.059 and 0.440 ppm in seeds 73-98 days after the last application.” Dicamba DGA: “Using these 21 Second Addendum assumptions in to the TerrPlant (total 2 lb Environmental Fate ae/A application and a and Ecological 0.055% runoff Risk Assessment fraction), and the most for Dicamba DGA sensitive salt and its endpoint of 0.000261 Degradate, 3,6for the NOAEC for dichlorosalicylic soybeans, the acid (DCSA) for maximum RQ is less the Section 3 New than the LOC of 1.0 by Use on Dicambaa factor of at least 2 Tolerant Soybean (RQs range from <0.1 to 0.48, see Appendix 3).” Dicamba DGA: Herbicide interactions 22 Second Addendum Section to the Environmental Fate and Ecological Risk Assessment Replace “the most sensitive endpoint of 0.000261” with “the most sensitive endpoint of 0.0673 lb a.e./A.” Runoff exposure estimates should be compared to the seedling-emergence endpoint not vegetative-vigor endpoint. Analysis needed: Synergy should include discussion of rates considered. Only rates that would be expected off field are appropriate (i.e., not field rates). 70 Docket ID No. Document Name & Text Section 0007 for Dicamba DGA salt and its Degradate, 3,6dichlorosalicylic acid (DCSA) for the Section 3 New Use on DicambaTolerant Soybean EPAHQOPP201601870007 Dicamba DGA: Second Addendum to the Environmental Fate and Ecological Risk Assessment for Dicamba DGA salt and its Degradate, 3,6dichlorosalicylic acid (DCSA) for the Section 3 New Use on DicambaTolerant Soybean Memorandum. Dicamba and Dicamba BAPMA Salt: HumanHealth Risk Assessment for Proposed Section 3 EPAHQOPP201601870009 Page Comment/Data Needed Appendix 3 31-33 Analysis needed: Considering the in-field buffer, spray drift should not be considered in TerrPlant calculations. “..most highly exposed population subgroup is children ages 1-2…42% of the cPAD.” 5, 37 Analysis needed: The text, while consistent with the select data presented in the summary table in 5.4.6, appears to be inconsistent with the DEEM-FCID results presented in EPAHQ-OPP-2016-0187-0011Dietary Exposure Assessment, Attachment 6, which show population subgroup non-nursing infants at 45.4% of the cPAD. 71 Docket ID No. Document Name & Text Section Page Comment/Data Needed 2 Analysis needed: The text, while consistent with the select data presented in summary Table 5 (p 10) appears to be inconsistent with the DEEM-FCID results presented in EPA-HQ-OPP-20160187-0011Dietary Exposure Assessment, Attachment 6, which show population subgroup non-nursing infants at 45.4% of the cPAD. New Uses on Dicamba-tolerant Cotton and Soybean EPAHQOPP201601870011 Executive Summary and 5.4.4 Chronic Dietary Risk Assessment Memorandum. Dicamba. Acute and Chronic Dietary Exposure Assessments of Food and Drinking Water to Support the Use of Dicamba on DicambaTolerant Cotton and Soybean for Amended Section 3 Registration, and Registration of the New N,N-Bis-(3aminopropyl) methylamine (BAPMA) Salt Formulation “…most highly exposed…children ages 1-2…42% of the cPAD.” “…chronic…children 1-2 years old had the highest chronic dietary risk at 42% of the cPAD.” 9 10 “…chronic…Children 1-2…42% cPAD.” Executive 72 Docket ID No. Document Name & Text Section Page Comment/Data Needed 23 Add: Iowa to the list of states. Summary VII. Results/Discussion EPAHQOPP20160187016 IX. Conclusions Proposed Registration of Dicamba on Dicamba-Tolerant Cotton and Soybean “This registration for dicamba is being proposed for registration for use in the states of Alabama, Arkansas, Arizona, Colorado, Delaware, Florida, Georgia, Illinois, Indiana, Kansas, Kentucky, Louisiana, Maryland, Michigan, Minnesota, Mississippi, Missouri, Nebraska, New Mexico, New Jersey, New York, North Carolina, North Dakota, Ohio, Oklahoma, Pennsylvania, South Carolina, South Dakota, Tennessee, Texas, Virginia, West Virginia, and 73 Docket ID No. Document Name & Text Section Page Comment/Data Needed 17 and other places in document Buffer is listed as 100 feet rather than 110 feet. Wisconsin.” EPAHQOPP20160187016 Proposed Registration of Dicamba on Dicamba-Tolerant Cotton and Soybean EPAHQOPP20160187016 Proposed Registration of Dicamba on Dicamba-Tolerant Cotton and Soybean “There is no evidence 9 of susceptibility to the young following in utero exposure to dicamba acid, dicamba BAPMA or DCSA.” Replace with “Susceptibility to young following in utero exposure to dicamba acid, dicamba BAPMA or DCSA is unlikely.” 74