Minutes of the January 11, 2011 Exposure Modeling Public Meeting

Office of Pesticide Programs (OPP) 4[th] Floor Conference Center
2777 South Crystal Drive  -  Potomac Yard Two, North
Arlington, VA  22202

	Attendees:
Name
Amanda Solliday
EPA
Andrew Jacques
RegNet
Andrew Newcombe 
ARCADIS
Avery Fellow 
Daily Environment Report
Bayazid Sarkar
EPA
Catherine Aubee
EPA
Clifford Habig
Exponent
Cody Howard
USDA/ARS
Dana Spatz
EPA
David Lieu
EPA
Greg Orrick
EPA
Jay Overmyer
Syngenta
Jeff Evans
EPA
Jenny Tao
EPA
Jim Breithaupt
EPA
Katherine Carr
Monsanto
Katrina White
EPA
Khin Oo
EPA
Kristian Paul
DuPont
Marrietta Echeverria
EPA
Mary Clock Rust
EPA
Melissa Panger
EPA
Michael Barrett
EPA
Michael Xiao Huang
DuPont
Mike Leggett
CLA
Mike Rexrode
EPA
Mohammed Ruhman
EPA
Natalia Peranginangin
Syngenta
Nathan Mottl
EPA
Nick Mastrota
EPA
Patrick Havens
Dow
Paul Hendley
Syngenta
Pete Coody
Bayer Cropscience
Phillip Villanueva
EPA
Qingli Ma
Exponent
R. David Jones
EPA
Rick Reiss
Exponent
Robert Miller
EPA
Rochelle Bohaty
EPA
Rueben Baris 
EPA
Russell Jones
ECPA Kinetics Group
Sarah Winfield
EPA
Scott Jackson
BASF
Steve Wente 
EPA
Stuart Cohen
Environmental & Turf Services
Subijoy Dutta
EPA
Tammara Estes
Stone Environmental Inc
Tanja Crk
EPA
Tiffany Mason
EPA
William Stiteler
ARCADIS US Inc.






Welcome and Introductions
	The USEPA/OPP/EFED hosts biannual Exposure Modeling Public Meetings (EMPMs), which provide a forum for exchanging information between OPP/EFED and stakeholders with similar technical expertise on current issues related to pesticide exposure modeling in support of risk assessment.  The overarching theme of this EMPM was "Spatial Context to Terrestrial Exposure Modeling."  Rochelle Bohaty, Tanja Crk, Kris Garber and Andrew Shelby chaired the meeting in the capacity of co-chairs of the Terrestrial Exposure and the Water Quality Technical Teams.  Former Water Quality Technical Team co-chair, Chuck Peck, also assisted in the coordination of the meeting.
	The minutes and presentation slides for the January 11, 2011 EMPM are posted in the docket (EPA-HQ-OPP-2009-0879), which can be accessed at http://www.regulations.gov/.   The docket ID # is reported on the EMPM webpage: http://www.epa.gov/oppefed1/models/water/empm_top.htm
Brief Updates:
None
Presentations:
Presentation 1:
"Estimating Confidence Intervals for Metabolite Degradation Rates"
Russell Jones, ECPA Kinetics Group and Bayer CropScience
	Regulatory agencies around the world are moving towards using formation decline kinetics for estimating degradation rates of metabolites.  Estimating confidence intervals of metabolite degradation rates  using such kinetic analysis has become an important element of the registration process, especially in Europe.  When confidence intervals are estimated using ordinary least squares regression techniques, the resulting confidence intervals are overly broad.  Other standard techniques such as IRLS (Iteratively Reweighted Least Squares) and MCMC (Markov ChainMonte Carlo) provide more appropriate estimates of confidence intervals since these two techniques do not assume that the error variance is the same for parent and metabolites.

      Presentation notes:

      Because confidence intervals, which are estimated using ordinary least squares (OLS) regression techniques, may be overly broad, there is a need to look at other methods for determining these intervals.  In his presentation, Russell Jones described his project for  exploring other standard techniques for determining confidence intervals, namely Iteratively Reweighted Least Squares (IRLS) and Markov Chain Monte Carlo (MCMC).  Using empirical data, Jones compared upper and lower confidence limits using OLS, IRLS, and MCMC.  Based on the data from this comparison, both IRLS and MCMC techniques are better than OLS because they account for the error variance between the parent and the metabolite or degradate.  Both IRLS and MCMC, though, are comparable. 

      The examples presented were based on single, first-order modeling; however, Jones stated that other models can be used.  He has performed similar analysis for 20-30 other individual studies (3-4 individual studies per chemical), and the results were similar (MCMC and IRLS provide similar results which are better than OLS).  In response to questions  related to the statistical analysis,  Jones referred attendees to  the following articles, which will be published in the near future:
      
      Görlitz, Linus; Gao, Zhenglei; Schmitt, Walter.   Statistical Analysis of Chemical Transformation Kinetics using Markov-Chain Monte-Carlo Methods.  Submitted to Environmental Science & Technology and accepted for publication. 

Gao, Zhenglei; Green, John; Vanderborght, Jan; Schmitt, Walter.  Improving Uncertainty Analysis in Kinetic Evaluations Using Iteratively Reweighted Least Squares.  Submitted to Environmental Toxicology and Chemistry. 

Jones RL, Beigel C, Erzgraeber B, Gao Z, Green JW, Harvey B, Hayes S, Mackay N, Paul KW, Schmitt W, Yon D  Estimating Confidence Intervals for Metabolite Degradation Rates.  Accepted for presentation and inclusion in the proceedings at the XIV Symposium in Pesticide Chemistry, 30th August - 1st September 2011,  Piacenza, Italy.  

Presentation 2:
"Determining the Fate and Transport of Pesticide Emissions in the Chesapeake Bay Region"
Cody Howard, USDA/ARS 
	Environmental scientists have primarily focused on the fate, transport, and toxicity of pesticide active ingredients, while relatively little research has been conducted on  the environmental fate of pesticide inactive ingredients, sometimes termed formulants or inerts.  Many formulants are volatile organic compounds that may contribute to ground-level ozone pollution.  In California, a system of formulation testing using thermogravimetric analysis (TGA) is used to estimate VOC emission potential for pesticide products.  In this study,  several available methods for estimating VOC emissions and ozone formation potential (OFP) of pesticide products were evaluated.  The methods examined included TGA alone; TGA with calculated OFP from confidential formulation information; TGA-FTIR of evolved gases and calculation of OFP; and gas-chromatography-mass spectrometry and calculation of OFP.  
	Further research studies are focused on measuring the concentration of selected pesticides, agricultural-related emerging contaminants, and VOCs in surface agricultural soils collected from representative sites across the agriculture-urban interface of the Chesapeake Bay.  The collected data will be used to develop emissions inventories for selected gaseous compounds using the Pesticide Emissions Model.  Input files for initial model calibration will be populated with soil concentration, soil property, soil moisture, and meteorological data from a multi-year study of pesticide fate carried out at the USDA.
--------------------------------------------------------------------------------
      Presentation notes:
--------------------------------------------------------------------------------
      Dr. Howard explained the motivation behind this work, citing the Presidential Executive Order 13508, which mandates that the governments at the local, state, and federal levels restore Chesapeake Bay to an acceptable level.  This project is currently being directed by a multi-agency government task force.  In addition, there is interest in quantifying agricultural sources of SVOC ozone-precursors since acceptable ozone standards are proposed to be more restrictive in the future.  
--------------------------------------------------------------------------------
      Throughout the U.S., urban and agricultural communities have become more spatially intertwined, resulting in blurred land use boundaries.  Emissions from urban areas such as NOx, particulate matter, and VOCs can contribute to ground-level ozone pollution, causing damage to crops.  Agricultural operations also emit pollutants to the atmosphere, including ammonia, odorous VOCs, and particulate matter which can create nuisance problems in residential communities.  Agricultural emissions can also interact with urban pollutants that contribute to poor air quality.  In addition to air pollutants, pesticides and other organic pollutants  used in both urban and agricultural communities  can have negative effects on ecosystems.  In order to mitigate the risk from these chemicals, the project team will seek  to discern the chemical nature, fate, and transport of critical agricultural pollutants emitted to the atmosphere and consider the potential risks posed by reactivity and/or deposition of these chemicals to sensitive ecosystems. 
--------------------------------------------------------------------------------
      Dr. Howard is currently working to develop a pesticide emission inventory using land use data and sales data in tandem with the Pesticide Emission Model (PEM).  The emissions inventory will be processed in the SMOKE (Sparse Matrix Operator Kernel Emission Model) pre-processor program in order to utilize the CMAQ (Community Multi-scale Air Quality) model to ultimately determine the loading in the Bay from deposition processes associated with pesticide concentrations in the air.  The PEM model for soil emissions will be calibrated with a robust USDA 10-year field volatility study on atrazine and metolachlor.  Additional field volatility studies will be needed for crop canopy and orchard type of applications in order to calibrate PEM.
--------------------------------------------------------------------------------
      
--------------------------------------------------------------------------------
Presentation 3:
--------------------------------------------------------------------------------
"Developments in Terrestrial Exposure Modeling"
--------------------------------------------------------------------------------
Ed Odenkirchen, USEPA/OPP/EFED
--------------------------------------------------------------------------------
The Environmental Fate and Effects Division, Office of Pesticide Programs, United States Environmental Protection Agency (EFED) has been working to enhance its methods for problem formulation and exposure assessment phases of the ecological risk assessment process for terrestrial wildlife to be more mindful of the potential for toxicologically significant exposures from a variety of pathways beyond the traditionally assessed dietary route.  These enhanced methods are intended to provide transparent, reasonable, concise, and consistent evaluations of the potential pesticide risks posed by the drinking water, inhalation, and dermal exposure routes.  This presentation discusses past assumptions, present problem formulation screening tools for prioritizing data collection and risk assessment efforts, and development efforts for first tier screening risk assessment tools.
--------------------------------------------------------------------------------
      Presentation notes:
--------------------------------------------------------------------------------
      Historically, EPA's ecological risk assessments for pesticides have focused solely on dietary routes of exposure to birds and mammals. Although the LD50/ft2 tool in the T-REX model addresses potential exposure to non-target organisms from all possible routes, this analysis does not distinguish which pathways (diet, drinking water, inhalation and/or dermal) are significant sources of chemical exposure. Current modeling efforts within the Office of Pesticide Programs are addressing this need to examine other potential exposure pathways. New risk assessment tools include the drinking water Screening Imbibition Program (SIP), the Screening Tool for Inhalation Risk (STIR) and a soon-to-be-developed Dermal Uptake Screening Tool (DUST). The purpose of these tools is to utilize both chemical fate and toxicity information to identify when a route of exposure can be considered negligible during the problem formulation stage of Registration Review. If an exposure pathway is determined to be a potential source of risk based on the conservative estimates provided by these screening tools, then a clear rationale exists to request additional data to address these concerns. T-REX, SIP, and STIR can be found online at http://www.epa.gov/oppefed1/models/terrestrial/
--------------------------------------------------------------------------------
Question and Answer Session:
--------------------------------------------------------------------------------
Q:  Is your ultimate goal to include these pathways in T-REX?
--------------------------------------------------------------------------------
A:  Yes
--------------------------------------------------------------------------------
Q:  For the dermal model, why do you assume that 100% of the pesticide is absorbed by the skin of birds and mammals?
--------------------------------------------------------------------------------
A:  If the potential for risk is identified with this conservative assumption, then this highlights the need for additional information. If sufficient additional data is submitted, then the percentage of pesticide absorbed could be adjusted. 
--------------------------------------------------------------------------------
Q:  Do you anticipate a peer review, such as a Scientific Advisory Panel (SAP) before employing these tools?
--------------------------------------------------------------------------------
A:  No, a Scientific Advisory Panel is not necessary as the underlying basis for these tools has previously been reviewed by former SAPs.
--------------------------------------------------------------------------------
Q:  Can STIR handle inhalation of granulars?
--------------------------------------------------------------------------------
A:  The tool can evaluate a chemical if the vapor pressure is known.
--------------------------------------------------------------------------------

--------------------------------------------------------------------------------
Presentation 4:
--------------------------------------------------------------------------------
"Dermal Contact, Movement, and Amphibian Pesticide Exposure"
--------------------------------------------------------------------------------
Tom Purucker, EPA/ORD/NERL
--------------------------------------------------------------------------------
The more terrestrial an amphibian's life cycle is, the more likely its skin will be used to regulate  water content in order to maintain hydration.  Therefore, amphibian dermal contact from enhanced skin permeability can be a key exposure pathway compared to non-amphibian receptors.  Hydrophilic and lipophilic molecules have separate pathways for dermal exposure, with neutral, lipophilic molecules traditionally receiving the most attention in humans and other mammals.  However, preferential water uptake by the seat patch in amphibians increases the effective surface area for osmosis and may enhance the potential for low Kow dermal contaminant exposures.  Although amphibian epidermis has some capability to limit exposures to contaminants, positive linear relationships between Kow and contaminant flux across amphibian skin are significantly higher in amphibians versus mammals.  In addition, sustained direct contact with soil from daily movement patterns, overnight burrowing, and aestivation/hibernation periods further complicates chronic exposure estimation for amphibians.   We reviewed recent data and methods concerning amphibian exposures to identify a set of algorithms and parameters for screening and higher-tier model selection purposes that accounts for the increased dermal contact exposure potential and movement behavior.
--------------------------------------------------------------------------------
      Presentation notes:
      The purpose of this project is to review recent literature on amphibian dermal exposure, develop a model to estimate dermal contact exposure in amphibians, and compare the relative importance of dermal exposure across animal classes.  This presentation explores factors related to dermal exposure in amphibians although knowledge is lacking on several key factors.  
      Mammals have a strong hydrophobic barrier in the epidermis that limits absorption of all but non-ionic, lipophilic molecules.  On the other hand, amphibians lack this hydrophobic layer.  Estimates of dermal exposure doses in children vary greatly, by three orders of magnitude, depending on the methodology used.  In mammals, decreasing relative dermal absorption is correlated with increasing dermal loading.  This correlation has important implications for extrapolating from lab studies (high loading) to field conditions (low loading).  Amphibians differ from mammals in the following ways: 1) they have lower metabolic demand than mammals; 2) they have greater surface area relative to body weight (due in part to the seat patch) than mammals; 3) they have thin skin without a hydrophobic barrier to chemicals, resulting in greater absorption; 4) they absorb water through the skin;  and 5) gas exchange is mainly through the skin.  
      Studies have shown that dermal exposure is an important pathway for amphibians. The seat patch on the ventral side of amphibians' skin is an adaptation that is responsible for fluid intake and osmotic release of chemicals.   With their high moisture levels, irrigation fields are especially attractive to amphibians and are sources of chemical exposure.  Amphibians may also receive high exposure to pesticides in agricultural soils when they burrow into soil at night and  over the winter.  When amphibians burrow over winter, they eat their shed skin, resulting in  a dietary route of exposure in addition to dermal exposure.  The more terrestrial an amphibian is, the more likely it will be exposed to pesticides.  Chemicals with high Kow's are likely to be more of a concern than those with low Kow's.  The dermal exposure model includes coefficients for dermal penetrability (Kp), surface area (SA), skin thickness (H), time of exposure (T), bioavailability (BF), and body weight (BW).  Walker et al. (2003) showed that there is a close linear relationship between Kp and Kow in humans, while Quaranta et al. (2009) looked at this relationship in frogs and pigs.  Both researchers found a similar relationship (i.e., similar slopes), but the log(Kp)was much greater for frogs than for pigs (i.e., the intercept was greater).  Unfortunately the results from the human study (Walker et al., 2003) are not comparable to those from the pig and frog study (Quaranta et al., 2009) because of differences in experimental design.  The ratio of frog permeability to pig permeability varied greatly, ranging from about 25x to about 300x.  The dermal exposure model was used to estimate dermal and dietary doses for mammals, birds, reptiles, and amphibians and to compare the relative contribution of the exposure routes in different classes of animals.  Metabolic rates are lower for cold-blooded reptiles and amphibians than for warm-blooded birds and mammals.  The relationship between metabolic rate and body weight is more uncertain for reptiles and amphibians than for birds and mammals.  
      For all classes, doses were higher for chemicals with higher Kow values.  The order of dermal dose by class is as follows:  amphibians > reptiles > birds and mammals.  Kow does not affect dietary dose for any class.  The order of dietary dose by class is as follows:  birds and mammals > reptiles and amphibians.  The relative contribution of dermal exposure is much greater for amphibians than for other classes.  For reptiles, the relative contribution is strongly dependent on the Kow, and is high for high-Kow chemicals.  The relative contribution of dermal exposure is small for birds and mammals.   Dermal exposure to pesticides can be estimated by using:  (1) exposure associated with moving animals, which is based on Kp, or (2) exposure based on a "finite" source in the vicinity of an aestivating animal.  
      In conclusion, amphibians and reptiles receive a significant dose of pesticides from dermal exposure, whereas birds and mammals receive very little.  These findings have important implications for the surrogate use of birds to protect amphibians and reptiles from pesticides.  Laboratory and field data sets for amphibians need to be examined to better define the parameters used in dermal exposure models.
      
--------------------------------------------------------------------------------
Question and Answer Session:
      Q:  In the plots of Kp versus Kow, most of the data were for chemicals with relatively low log(Kow) values of 2 to 3.  What would the plot show for chemicals which have much greater Kow values?  Would the line level off?
      A (Tom Purucker):  My guess is that the linear relationship probably would not hold for chemicals with high Kow values.  The slope would probably change.
      Q:  Will you consider the behavior and ecology of amphibian species when you predict their dermal exposure?  For example, some amphibians come out only at night, and seek out areas where insect densities are higher.
      A (Tom Purucker):  You could account for such ecological factors to some degree.  For example, you could adjust the time spent in field based on ecological characteristics.
      A (Edward Odenkirchen):  With all of these models, there needs to be a balance between being general and being realistic for specific species.  To be efficient, the EPA needs to keep the model fairly generic for species so that we can apply the results to protect a wide range of nontarget species.

Presentation 5:
"Habitat Classification for Ecological Risk Assessment Using Aerial Photography and GIS Data in a Two-stage Expert System"
--------------------------------------------------------------------------------
William Stiteler, PhD, ARCADIS US Inc.
--------------------------------------------------------------------------------
There are many challenges in performing a land cover classification that can be used for habitat suitability analysis of multiple species in ecological risk assessments.  The classification scheme must be detailed enough to  accomodate  several different habitat requirements, as well as flexible enough to be applicable to different species.  Because habitat suitability often has minimum area requirements, but does not require regular shapes of contiguous cover, high spatial resolution is useful, but imagery of high spatial resolution often lacks the spectral resolution useful for habitat classification.  We present the results of a land cover classification designed to allow for the modeling of habitat suitability for a variety of receptors, such as shrews, robins, and deer.  A hybrid remote sensing / GIS approach was developed that uses both aerial photography and ancillary GIS data.  A two-stage expert system is used to bring together different data sources.  This expert system was designed to work with a particular set of data sources, but it can be easily adapted to other sources.  The result is a land cover classification with fourteen classes, a minimum mapping unit of 1/20 acre, and distinct polygons that can be as narrow as four feet wide.  Ground reference data from field visits showed an overall accuracy of approximately 80% for qualitatively different classes and 60% for the full classification scheme containing quantitative class divisions, such as those between closed, intermediate, and open canopy cover classes.
--------------------------------------------------------------------------------
      Presentation notes:
--------------------------------------------------------------------------------
      A method to classify habitat based on aerial photography conditioned on GIS coverages of soils data and tax parcel data was presented. Complications in developing accurate coverages using this method are effects of shadows, compression of digital photographs, and effects due to time of the year (i.e., you can't see through a forest canopy except in winter). In some cases, boundary data are defined differently in different jurisdictions, and this situation requires special handling. In the prepared example, a dam along the boundary of two counties was described differently in the boundary data, and consequently the software had trouble classifying the river and land around the dam. The analysis was conducted in two stages: first  a draft set of polygons from the photos and GIS coverages was developed and secondly  the resulting polygons into habitat classes were aggregated. The resulting coverage had accuracy (~73%) similar to the NLCD (National Land Cover Database).
--------------------------------------------------------------------------------
      Question and Answer Session:
--------------------------------------------------------------------------------
      Q: Were standard USDA boundaries used for the assessment as it can be difficult to combine local coverages such as that presented in the paper with broader regional or national assessments? 
--------------------------------------------------------------------------------
      A: USDA boundaries were not used for the assessment.
--------------------------------------------------------------------------------

Wrap up:
The next EMPM will be held on September 13, 2011.  The meeting theme will be announced through the "empm listserve" forum on the LYRIS list serve at:
https://lists.epa.gov/read/?forum=empmlist
