              OFFICE OF CHEMICAL SAFETY
AND POLLUTION PREVENTION
                 UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
                            WASHINGTON, D.C.  20460
              OFFICE OF CHEMICAL SAFETY
AND POLLUTION PREVENTION
                 UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
                            WASHINGTON, D.C.  20460
                                       
                                       
                                       
                                       
                                       
                                 June 25, 2014
                                                                PC Code: 121001
MEMORANDUM	DP Barcode: 418281

SUBJECT:	Sethoxydim: Drinking Water Exposure Assessment for IR-4 Petition on Fescue, Oilseeds, Fruits, and Vegetables

FROM:		James Lin, Environmental Engineer 
            ERB2/EFED (7507P)

REVIEWED BY:  	Greg Orrick, Environmental Scientist
            	ERB2/EFED (7507P)

THROUGH:		Brian Anderson, Chief
            ERB2/EFED (7507P)

TO:		Sydney Jackson, Risk Manager
			Barbara Madden, Team Leader
			RIMUERB/RD (7505P)

			James Parker, Risk Manager
			Neil Anderson, Branch Chief
			RMIB1/PRD (7508P)

      
This assessment provides estimated drinking water concentrations (EDWC) of sethoxydim in surface water and in ground water in support of human health risk assessment for IR-4 petitions.  Screening EDWCs (Table 1) of sethoxydim were generated with the Surface Water Concentration Calculator (SWCC) and the Provisional Cranberry Model for surface water and with the Pesticide Root Zone Model for Ground Water (PRZM-GW) for ground water.  Modeled application rates represent the maximum use patterns of proposed end-use labels and current registered labels.  Remaining model input parameters were chosen according to current guidance (USEPA, 2009).  EDWCs reflect exposure to sethoxydim residues of concern in drinking water, which include the parent compound and eight degradates of concern (DP 246356).  If the screening EDWCs listed in this memo result in dietary risk exceedances, contact James Lin (703-308-9591) of Environmental Risk Branch II (7507P) to request a refined drinking water exposure assessment.
      

Table 1.  Maximum Screening Drinking Water Exposure Estimates for Proposed Sethoxydim Uses (Based on Sethoxydim Residues of Concern)*
Source (Model)
Use (Maximum Rate)
                             Peak Exposure (μg/L)
                         Annual Mean Exposure (μg/L)
Surface water (SWCC)
                            4 apps @ 0.47 lb ai/A 
                             w/ a 14-day interval
                                     65.3
                                      6.6
Ground water (PRZM-GW)
                                       
                                     0.565
                                     0.511
Surface water (Provisional Cranberry Model)
                  2 apps @ 0.47 lb ai/A w/ a 14-day interval
                                     79.6
                                     13.9
[*]Maximum values in bold.

	The Environmental Fate and Effects Division (EFED) has produced several drinking water assessments for sethoxydim, such as the following reports:

   1. D239881, D239882, D239884, D238951 - Drinking Water Assessment for Sethoxydim for the Current Uses and for Proposed IR-4 Uses on Leafy Vegetables (excluding Brassica), Root and Tuber Crops (excluding Radish), Artichoke, and Caneberries. 

   2. D257560, D257310 - Drinking Water Assessment and Ecological Risk Summary for Sethoxydim for the Maximum Use Crop (Citrus) and the Proposed IR-4 Uses on Pistachio and Safflower.

   3. D312559 - Sethoxydim Drinking Water Assessment (Tier 1) for Reregistration Eligibility Decision.

   4. D313456, D314420 - Drinking Water Assessment for Sethoxydim: Proposed Uses on Buckwheat, Okra, Borage, Dill, Root Vegetables, and Turnip Greens.

      Due to recent updated modeling approaches, EFED is required to use new tools: the Surface Water Concentration Calculator (SWCC) for surface water and the Pesticide Root Zone Model for Ground Water (PRZM-GW) for ground water.  In addition to these two new modeling tools, this drinking water exposure assessment also uses the provisional cranberry modeling tool.
      

USE INFORMATION

	Sethoxydim is a broad-spectrum, postemergence herbicide for selective control of annual and perennial grass weeds.  It is belongs to the cyclohexenone-class herbicide.  The end use product is labelled as Poast(R) herbicide, which is formulated as an emulsifiable concentrate (EC) containing 1.5 pounds of active ingredient per gallon.  This IR-4 petition expands the proposed uses to include:

   1. New uses, R190, on fescue, Bushberry Subgroup 13-07B, Low Growing Berry Subgroup 13-07H, Small Fruit Vine Climbing Subgroup 13-07F, Rapeseed Subgroup 20A, Sunflower Subgroup 20B (except safflower), and Cottonseed Subgroup 20C.

   2. Crop group/subgroup additions, R175, for Citrus Fruit Group 10-10, Pome Fruit Group 11-10, Caneberry Subgroup 13-07A, Bulb Vegetable Group 3-07, and Fruiting Vegetable Group 8-10.

	For screening purposes, two use sites are considered for the drinking water assessment.  The first one is based on the highest label use rate among all uses.  The highest label use rate is for citrus fruit group. The maximum single application rate is 2.5 pt/A (0.47 lb ai/A).  The maximum seasonal rate is 10.0 pt/A (1.88 lb ai/A) with a 14-day retreatment interval.  The second one is based on cranberry use, which has a PHI (pre harvest interval) of 30 days.  The single application rate is same as the citrus rate.  The maximum seasonal rate is only 50% of the citrus use rate.  

FATE AND TRANSPORT CHARACTERIZATION
   
      Transport and Mobility

      Sethoxydim is a highly soluble compound (4700 mg/L in pH 7 water) with a low octanol/water partition coefficient (Kow is 45.1).  The calculated bioconcentration factors (BCF) for sethoxydim and total residues were 7X, 25X, and 21X for edible, nonedible, and whole fish, respectively.  Depuration was fast, with a half-life of 3.6 days.  For these reasons, bioaccumulation is not likely to occur.  Also, due to its low vapor pressure (1.6x10[-7] mm Hg) and Henry's Law Constant (1.47x10[-11] atm-m3/mol), sethoxydim is not expected to be volatile under field conditions and from water.  
      
      Based on batch equilibrium experiments, sethoxydim  and its transformation products, M2-SO2, M-SO, M-SO2, and M2-SO, were determined to be mobile to very mobile in sterile (autoclaved) sand, sandy loam, sandy clay loam, silt loam, and clay loam soils.  Freundlich Kd values were <1.00 for sethoxydim and its transformation products M-SO and M-SO2.  The Freundlich Kd values for M2-SO and M2-SO2 ranged from 0.06 to 9.12. The only unsterilized soils were used on M2-SO2, for the reason that M2-SO2 is thought to be the most stable among test compounds and it is the terminal degradate. The Freundlich Kd values of M2-SO2 with non-sterilized soil are 0.12, 2.89, 0.78, 5.64, and 9.42, respectively for sand, sandy loam, sandy clay loam, silt loam, and clay loam.
      
            Degradation

      Sethoxydim hydrolyzes at moderately rapid rates at low pH's, but is more stable at high pH's.  The calculated half-lives are 8.7, 155, and 284 days in pH 5, 7, and 9 solutions, respectively.  The major observed hydrolysis transformation product is M2-S or 6-(2-(ethylthio)propyl)-4-oxo-2-propyl-4,5,6,7-tetrahydrobenzoxazole.  In contrast, evaluation of the data of sethoxydim total residues shows that they remain stable at all three pHs tested.

      Sethoxydim and sethoxydim total residues degrade photolytically in both water and soil.  In pH 8.7 buffered water, the calculated photolysis half-life of sethoxydim is 5.23 days, and the major transformation product is M1-S or 2-(1-aminobutylidene)-5-(2-(ethylthio)-propyl)-cyclohex-1,3-dione.  In sandy loam soil irradiated with a xenon light source, the half-life of sethoxydim is approximately 1 hour, and the major transformation product is M-SO or 2-(1-ethoxyiminobutyl)-5-(2-(ethylsulfinyl)propyl)-3-hydroxycyclohex-2-enone.  Sethoxydim total residues photodegrade slower than parent sethoxydim in soil and water.  Using a linear regression analysis, EFED calculated a half-life of 19.8 days for the photolysis in water of sethoxydim total residues and a half-life of 20 hours in soil.

      Under aerobic conditions, parent sethoxydim transformed with short half-lives (<1 day) both in soil and aquatic environments.  It degraded with a half-life of less than one day in sandy loam and sandy clay loam soils.  The major transformation product at 2 months was M-SO, and after 12 months the major product was CO2.  Using aerobic clay loam soil:water and aerobic clay soil:water systems, it was determined that, under aerobic aquatic conditions, sethoxydim transformed with a half-life of <1 day.  After 28 days, the major transformation products were CO2, M-SO, M2-S, and M-SO2.  In contrast to parent sethoxydim, sethoxydim total residues were more persistent.  The observed half-life was 1 month for sethoxydim total residues in the aerobic sandy loam study, and 7 days in the aerobic sandy clay loam study.  In an aerobic clay loam soil:water system, the calculated half-life for sethoxydim total residues was 38.1 days, while in an aerobic clay soil:water system the half-life was 32.9 days.

      Under anaerobic conditions, parent sethoxydim is even more persistent than under aerobic conditions.  It transformed with half-lives of 11 to >60 days under anaerobic soil conditions and 25-39 days in anaerobic aquatic metabolism studies.  M-SO was the major transformation product in both studies.  It was observed that M2-S, which was a major transformation product in the hydrolysis study, was only a minor transformation product in the anaerobic aquatic metabolism study.  Sethoxydim total residues were more persistent than parent sethoxydim in the anaerobic studies.  A half-life of 91.6 days was observed in the anaerobic soil metabolism study, while half-lives of 132-187 days were observed in the anaerobic aquatic metabolism study.

      Residues of Concern
      
      The proposed degradation scheme of the parent and eight degradates has been discussed in an earlier memo (DP 312559).  The Metabolism Assessment and Review Committee (MARC) of the Health Effects Division (DP 246356) states that all degradates are to be included.  Because of the short metabolism half-lives of parent sethoxydim (< 1 day), the EDWCs represent total residues, including sethoxydim, M-SO, M-SO2, M1-S, M1-SO, M1-SO2, M2-S, M2-SO, and M2-SO2.  The proposed scheme of the transformation of sethoxydim and the eight metabolites is presented in Attachment 1. 

      Drinking Water Exposure Assessment
      
      Drinking water exposure to the residues of concern is estimated using available chemical properties and environmental fate data.  A total residues of concern (TRC) approach was used because degradation studies were not performed on the specific degradates of concern.  Table 2 lists the environmental fate exposure modeling inputs of the sethoxydim TRC.   These parameters reflect regression of the total residues at each interval of the degradation studies.  Chemical properties of the parent compound were used to represent those of the residues of concern due to the lack of data on the degradates.  The mean soil adsorption coefficient for M2-SO2 was used to represent the residues of concern because this compound was thought to be the most prominent terminal degradate.  For the batch equilibrium study (MRID 41475212), the parent and 4 degradates (M-SO, M-SO2, M2-SO, and M2-SO2) were studied separately.  Due to the possible short half-life values of the parent and degradates, the batch equilibrium experiments were tested using sterile (autoclaved) soils.  The terminal degradate (M2-SO2) was thought to be most stable, so additionally it was also tested using non-sterile soil.  The results indicate that M2-SO2 is very mobile in sand, sandy loam, sandy clay loam, and silt loam and to be mobile in clay loam.  Freundlich Kd values were 0.12 for the sand soil, 0.78 for the sandy clay loam soil, 2.89 for the sandy loam soil, 5.64 for the silt loam soil, and 9.42 for the clay loam soil.  Adsorption increased as the CEC values of the soils increased.

Table 2.  Chemical Model Inputs for Sethoxydim Total Residues of Concern
Parameter
Value
Source
Molecular Weight (g/mol)
327.48
C17H29NO3S
Water Solubility (mg/L)
4700
EFED one-liner
Vapor Pressure (torr)
1.6 x 10[-7]
EFED one-liner
Hydrolysis half-life (day) (25 [o]C)
Stable (pH 5, 7, 9)
MRID 41475207
Aqueous photolysis half-life (day)
(25 [o]C)
19.8 (total residue)
MRID 41475208
Aerobic soil metabolism half-life (day) (25 [o]C)
54 (upper 90[th] percentile on two values of 30 days and 7 days for total residues)
MRID 41475210 
Aerobic aquatic metabolism half-life (day) (25 [o]C)
44 (upper 90th percentile on two values of 38.1 days and 32.9 days for total residues)
MRID 42165604
Anaerobic aquatic metabolism half-life (day) (25 °C)
244 (upper 90th percentile on two values of 132 days and 187 days for total residues)
MRID 41475211
Adsorption coefficients (Kd) (25 [o]C)*
3.77 (Mean of five values: 0.12, 0.78, 2.89, 5.64, and 9.42)  
MRID 41475212
*The only batch equilibrium test with non-sterile soils, conducted on M2-SO2 (the terminal degradate).

	Drinking Water from Surface Water Sources

	The Surface Water Concentration Calculator (SWCC v1.106) was used to calculate EDWCs from surface water sources.  There are two citrus modeling scenarios in the SWCC database.  They are California citrus and Florida citrus.  Both scenarios were assessed to obtain EDWCs.  Table 3 lists the application related information and assumptions.  Since sethoxydim is used to control post-emergence annual and perennial grass weeds, the first application date was assumed during the spring time.   
 
Table 3.  Application Information Related to Sethoxydim Use on Citrus Scenarios
Model Input Variable
Input Value
Comments
1[st] Application Date (2 scenarios)
April 1 or April 15
In the spring time
Application Rate (maximum)
0.47 lb a.i./acre (2.5 pints)
Label Information
Maximum Number of Applications
4 (10 pints max. per season)
Label Information
Application Interval
14 days (minimum)
Label Information
Percent Cropped Area
91%
National default for all agricultural land
Application method
Ground 
Label Information
App. Efficiency / Spray Drift
0.99 / 0.066 (for ground)
USEPA, 2013 

	Standard percent cropped areas (PCA) are used as conservative default estimates of the extent of watershed on which agricultural crops of unknown specific PCA are grown.  Considering the many use sites of sethoxydim, the PCA value of 0.91 was used to represent the national default PCA for all agricultural land.  The PCA-adjusted SWCC results are tabulated in Table 4.  The modeling results of the Florida citrus scenario provided the higher EDWCs of 65.3 μg/L and 6.6 μg/L to represent the 1-in-10-year peak and mean exposure values, respectively.

Table 4.  Sethoxydim Total Residue of Concerns EDWCs from Surface Water Sources 
                                      Use
                                 PRZM Scenario
                               Initial App. Date
                      1-in-10-year Peak Exposure (μg/L)
                   1-in-10-year Annual Mean Exposure (μg/L)
                         30-year Mean Exposure (μg/L)
                                Citrus Group 10
                                  CA Citrus 
                                    Apr. 1
                                     3.89
                                     0.77
                                     3.38
                                       
                                  CA Citrus 
                                    Apr. 15
                                     3.90
                                     0.75
                                     3.55
                                       
                                  FL Citrus 
                                    Apr. 1
                                     65.3
                                      6.6
                                     0.71
                                       
                                  FL Citrus 
                                    Apr. 15
                                     56.6
                                      6.3
                                     0.68

      The Provisional Cranberry Model is a provisional refinement to the Tier I Rice Model (v1.0, May 8, 2007).  Refinements include the addition of simple degradation processes in dry and flooded conditions and a water depth of twelve inches, rather than the water depth of four inches used in the rice model.  These modifications allow estimation of screening-level peak and annual mean EDWCs of sethoxydim residues of concern that may occur in untreated surface water used as drinking water following use on cranberries.

      The Provisional Cranberry Model is based on the same assumptions as the Tier I Rice Model, with the addition of degradation processes in dry and flood conditions and a water depth of twelve inches in the flooded cranberry bog.  Degradation of sethoxydim residues of concern is assumed to occur predominantly via aerobic microbial metabolism, whether on dry cranberry bog soil or in bog flood water.
      
      The cranberry bog water depth of twelve inches is a maximum depth recommended by the Cape Cod Cranberry Growers Association.  Previous assessments of mesotrione uses that employed the Provisional Cranberry Model assumed a water depth of eighteen inches (DP barcode 339984, USEPA, 2004; DP barcode 325840, USEPA, 2006a), which is within the range of water depths traditionally used.  However, modern water conservation pressures are causing cranberry growers to reduce the flood depth to fewer inches above the vertical cranberry branches, which grow up to eight inches high.  Therefore, a twelve-inch deep flood is a reasonable refinement to model assumptions.  

	The label use of sethoxydim on cranberries at the maximum proposed application rate (0.47 lbs a.i./A) may occur no more than twice times per year, based on the annual application rate limit (0.94 lbs a.i./A).  The minimum reapplication interval is 14 days.  There is a pre-harvest interval (PHI) of 30 days requirement.
      
      Cranberry bogs are not typically flooded until the night before harvesting.  Therefore, for the parent compound, the proposed single application rate (0.47 lbs a.i./A) was aged on dry soil with the Provisional Cranberry Model for two durations: 44 and 30 days.  Peak and annual mean EDWCs were calculated beginning three days after flooding to account for transport time to a down-gradient drinking water intake following release of bog tailwater not long after flooding.   Model input parameters are listed in Table 5.  Because PCA values were not available for cranberry uses, they were not considered for exposure estimates in this cranberry use.
      
Table 5.  Provisional Cranberry Model input parameters for total residues of sethoxydim. 
Input Parameter
Value
Comments
Source
Application Rate
(lbs a.i./A)
0.47
Maximum single application rate.
Label Information
Applications per Year
2
Maximum number of applications per year at the maximum proposed single application rate.
Label Information
Application intervals (d)
14
Minimum interval
Label Information
Adsorption coefficients (Kd) (25 [o]C)*
3.77
Mean of five values: 0.12, 0.78, 2.89, 5.64, and 9.42  
MRID 41475212
Aerobic Soil Metabolism Half-life (days)
54
Upper 90[th] percentile on two values of 30 days and 7 days for total residues)
MRID 41475210 
Aerobic aquatic metabolism
half-life (days)
44
Upper 90th percentile on two values of 38.1 days and 32.9 days for total residues
MRID 42165604
      
      Based on the inputs from Table 5 and the sethoxydim application dates calculated to be 30 days (30 days PHI) and 44 days before flooding, the provisional cranberry model yields the EDWCs of 79.6 μg/L and 13.9 μg/L, respectively for peak value and annual average value.

      Drinking Water from Ground Water Sources 
      
      PRZM-GW (v1.07) is used to calculate EDWCs from ground water sources.  The chemical specific input parameters including hydrolysis half-life, aerobic soil metabolism half-life, and adsorption coefficient (Kd) are taken from Table 2 and application related inputs are taken from Table 3.  All six available scenarios were modeled as surrogates for turf scenarios, with the Florida citrus scenario producing the highest EDWCs as shown in Table 6.  The modeling results of Florida citrus scenario provide the higher EDWCs of 0.565 μg/L and 0.511 μg/L to represent the acute and chronic exposure values, respectively.

Table 6.  PRZM-GW Output of EDWCs for Sethoxydim Total Residue of Concerns
                               Modeled Scenario
                           Max. Daily Conc. (ug/L)
                        Mean Breakthrough Time (years)
                        Post-breakthrough Mean (ug/L)
                                   DMV corn
                                    0.0444
                                     86.12
                                    0.0394
                                   FL citrus
                                     0.565
                                     45.96
                                     0.511
                                   FL potato
                                    0.0656
                                     35.28
                                    0.0585
                                   GA peanut
                                     0.104
                                     67.69
                                    0.0923
                                   NC cotton
                                    0.00225
                                  Incomplete*
                                  Incomplete
                                    WI corn
                                    0.00384
                                  Incomplete
                                  Incomplete
*Incomplete indicates no breakthrough during the modeling simulation
                                       
Figure 1. PRZM-GW EDWCs over 100 years for the Florida citrus scenario

      The overall estimated drinking water concentrations from surface water sources and ground water sources are summarized in Table 1.  The surface water concentrations are about two order of magnitudes higher than the ground water concentrations.
      
Attachment 1.     Molecular Structures of Sethoxydim and Eight Metabolites
      
      The following scheme represents the relationship between parent sethoxydim (M) and its eight identified degradates.  There are three groups of degradates: a group formed from the parent by oxidation of the sulfur atom (M group), the "amines" or M1 group, and the tetrahydrobenzoxazoles, or M2 group.  The first M1 group daughter (M1-S) is formed by cleavage of the ethoxyimino group to an amino group, either by hydrolysis or aqueous photolysis.  The first member of the M2 group (M2-S) is then formed from M1-S by ring closure to produce a tetrahydrobenzoxazole ring.  Each of the three groups of degradates also has a sulfoxide (-SO) and a sulfone (-SO2), to round out the total of eight degradates.  This reaction dominates in aerobic soil metabolism and soil photolysis experiments, especially to form the M-SO and M-SO2 degradates.
      
		M Group		M1 Group	M2 Group	
		M      		      M1-S             	M2-S
		                    		                   	 
		M-SO 		      M1-SO  		         M2-SO
		                   		                    	
		M-SO2 	      M1-SO2                   M2-SO2

      The eight degradates include 2 daughter products, which are formed from the parent by a single reaction (M1-S and M-SO), 3 granddaughter products that are two reactions removed from the parent (M-SO2, M2-S, and M1-SO), 2 great-granddaughters that are three reactions removed (M1-SO2 and M2-SO), and one great-great-granddaughter that is four reactions removed (M2-SO2).  It is important to note that M1-SO, M2-SO, M1-SO2, and M2-SO2 may be formed by more than one pathway, and that the sulfur oxidation may be reversible under reducing conditions.
      
       
