Minutes of the February 27, 2007 Exposure Modeling Public Meeting 

Office of Pesticide Programs (OPP) Conference Room S-4370

2777 Crystal Drive – Potomac Yard One, South

Arlington, VA 22202

Attendees

Name	Association

Allan F. Rose	Valent U.S.A. Corporation

Amy Ritter	Waterborne Environmental Inc.

Andy Newcombe	LFR Inc.

Chris Holmes	Waterborne Environmental Inc.

Cliff Hasig	Exponent

Clifford Rice	USDA/ARS

Dirk Young	USEPA/OPP

Elizabeth Behl	USEPA/OPP

Fiona K. Lau	Rohm and Haas Company

Greg Orrick	USEPA/OPP

James Breithaupt	USEPA/OPP

James L Kunstman	PBI/Gordon Corporation

Jennifer Long	USDA/FAS

Jim Warren	USDA/APHIS

Jim Wolf	USEPA/OPP

Joe Massey	Mississippi State University 

John Weeks	SC Johnson

Jonathan Angier	USEPA/OPP

Julie Callahan	USDA/Foreign Ag. Service

Katherine Carr	Monsanto Company

Keara Moore	USEPA

Kenneth W. Chisholm	Nichino America, Inc.

Linda Abbott	USDA

Lisa Eisenhauer	USEPA/OPP

Lizanne Avon	Health Canada, PMRA

Lucy Shanaman	USEPA/OPP

Marietta Echeverria	USEPA/OPP

Mark Cheplick	Waterborne Environmental Inc.

Mark Corbin	USEPA/OPP

Mark J. Schocken	Chemtura Corporation

Mary Nett	Water Quality Consulting

Michael Barrett	USEPA/OPP

Michael Winchell	Stone Environmental, Inc.

Mohammed Ruhman	USEPA/OPP

Nasser Assaf	Valent Biosciences Corporation

Natalia Peranginangin	Syngenta Crop Protection, Inc.

Nicole Buffa	LW

Pamela Rao	Farmworker Justice

Patrick L. Havens	Dow AgroSciences LLC

Paul Hendley	Syngenta Crop Protection Inc.

Paul Paquin	Hydroqual

Qingli Ma	Environmental & Turf Services, Inc

Ralph L. Warren	DuPont Crop Protection

Reuben Baris	Environmental & Turf Services, Inc

Robert C Everich	Makhteshim-Agan of North America Inc.

Ronald Parker	USEPA/OPP

Russell Jones	Bayer CropScience

Scott Jackson 	BASF Corporation

Stuart Cohen	Environmental & Turf Services, Inc

Subijoy Dutta	USEPA/OPP

Tammara L. Estes	Stone Environmental Inc.

Teung Chin	USDA/FAS

Tharacad Ramanarayanan	Bayer CropScience

Uwe Wanner	Chemtura Corporation

Wenlin Chen	Syngenta Crop Protection, Inc.

William P. Eckel	USEPA/OPP



Dial In Participants

Name	Association

Adrian Wadley	Stone Environmental, Inc.

Ann Pitchford	U.S. EPA, NERL/ESD/LEB

Cecil Dharmasri	Syngenta Crop Protection, Inc.

Christopher S. Warren	ExxonMobil Biomedical Sciences, Inc.

Elise D.G.-McCoy	FMC Corporation

James Cooper	FL DACS

Jim McFadden	Wellmark International

Jimmy Williams	USDA/ARS

John Troiano	CA DPR

Joseph K. Bagdon	USDA /NRCS

Julie Wagner	Safer Pest Control Project, IL

Kelly Moran	TDC Environmental

Kevin L. Armbrust	State of Mississippi, MSU

Monica A. Ball	DuPont Crop Protection

Roy W. Meyer	NJ Dept of Env Prot, SFIREG

Susan Hunter Youngren	Bergeson & Campbell, P.C.

Welcome and Introductions

The USEPA/OPP/EFED hosts triannual meetings that are open to the public
to provide a forum for presentation and discussion between EFED and
stakeholders with similar technical expertise of current issues related
to pesticide exposure modeling in support of risk assessment.  This
meeting did not include an overarching theme or general topic.  Greg
Orrick (USEPA/OPP/EFED) chaired the meeting in the capacity of co-chair
of the Water Quality Tech Team (WQTT).

The agenda, minutes, and presentations for the February 27, 2007 EMPM
are posted in the docket (ID # EPA-HQ-OPP-2006-0994), which can be
accessed at  HYPERLINK "http://www.regulations.gov/"
http://www.regulations.gov/   The docket ID # is reported on the EMPM
webpage:   HYPERLINK
"http://www.epa.gov/oppefed1/models/water/emwg_top.htm" 
http://www.epa.gov/oppefed1/models/water/emwg_top.htm  

Brief Updates

Tier I Rice Model v1.0

The Interim Rice Model is being replaced by the Tier I Rice Model v1.0,
which uses the same formula.  The only change in approach is that the
assumed organic carbon content of sediment, which is 2% for the Interim
Rice Model, is being lowered to 1% to reflect the average organic carbon
content of rice paddies in submitted paddy dissipation studies conducted
in the Southeastern U.S. and California.  Once the Tier I Rice Model
v1.0 documentation is finalized, it will be posted on the Water Models
webpage:   HYPERLINK
"http://www.epa.gov/oppefed1/models/water/index.htm" 
http://www.epa.gov/oppefed1/models/water/index.htm 

FIRST v1.1.0

An update to FIRST (v1.1.0) that allows incorporation of ground spray
applications and converts units carrying more significant figures than
v1.0 is under review.  The updated model will be posted on the Water
Models webpage when it is released.

PRZM v3.12.2, EXAMS v2.98.04.06, and pE5

Issues exist with the QA of model updates to PRZM and EXAMS and the pE5
shell.  Estimates of when the QA issues will be resolved are elusive.

Presentations

Degradation Influenced by Soil Temperature under Cropped and Bare Soil
Conditions: A Conceptual Model Approach (Dr. Natalia Peranginangin,
Syngenta Crop Protection, Inc.)

Dr. Peranginangin and Dr. Chen presented an analysis of the effects of
soil temperature on degradation of two pesticides and related this to
the performance of PRZM in simulating the degradation / dissipation of
the pesticides.  This was performed without calibration (i.e., the
temperature adjustments were based upon laboratory data, not on the
field results the PRZM model was compared to).  They further tested the
impact of foliar washoff coefficient adjustment (in this case, based
upon some literature recommendations, but not on specific data for the
two pesticides in this study).

The temperature routine modeled closely followed field measurements of
soil temperature at different depths.  No specific statistical tests
were proposed or utilized for the goodness of fit of modeled dissipation
compared to measured dissipation.  Nonetheless, the graphics presented
do appear to show that utilization of specific heat transfer
coefficients (calibrated from site-measured soil temperature) for a
pesticide along with modeled soil temperature has the potential for
substantially improving the accuracy of predicted dissipation patterns
of pesticides.  Improved accuracy in the inputs for foliar wash off can
significantly increase the accuracy of the prediction of the timing and
magnitude of peak residues in soil (and presumably also in pesticide
runoff, although this was not examined in this study).  The authors
hypothesize that their analysis supports the conclusion that “aerobic
soil degradation half-lives are considered valid under both laboratory
and field conditions and that laboratory measured Koc may underestimate
soil adsorption under actual field condition (as it ages in soil).”

Development of MUSS (Dr. Jimmy Williams, USDA/ARS) 

Dr. Williams presented a history of the development of the erosion
simulation models based on the Universal Soil Loss Equation (USLE).  His
group at USDA/ARS modified the USLE to predict soil erosion on a single
storm basis rather than the original multi-year basis.  The new Modified
Universal Soil Loss Equation (MUSLE) methodology used a factor based on
the product of the runoff volume times the runoff peak rate rather than
on the original erosivity (R) factor.  The formula was based upon data
from almost 800 runoff events in eighteen watersheds that varied in
slope from 1 to 6 percent.  The MUSLE equation and its derivatives are
currently used in the USDA models ALMANAC, APEX, EPIC and SWAT as well
as EPA’s PRZM.

The second major version of MUSLE was based on converting the equation
to metric units and to a per-unit-area basis.  In this formulation,
almost all of the erosion is sheet and rill erosion and very little is
channel erosion.

The next major step was the development of MUSS.  This model was based
on a subset of the rainfall events used in MUSLE and only the watersheds
with surface areas of less than 71 hectares and no defined channel.

The group then developed a theoretical version which they named MUST. 
This version was based on the ratio of the peak rate to the volume of
runoff for each storm and kept the original K, C, P and LS factors.

Additional work could still be done using the 3695 rainfall events and
63 watersheds in 7 states which are now available.  New work would
include development of separate equations for sheet erosion, transport
and gully erosion and might also include collection of additional data
for model validation.  

High Priority Changes for PRZM (Dr. Dirk Young, USEPA/EFED) 

Dr. Dirk Young gave an overview of discussions between the EPA Office of
Pesticides and the EPA Office of Research and Development on changes to
PRZM that would improve PRZM's usability for risk assessments.  These
changes include the ability for a user to have more control over
processes such as erosion, irrigation, and contaminant interaction with
soil.  Some of the work may involve answering research questions such as
how does runoff interact with surface soils and how does partitioning of
organic compounds to surface soils differ from partitioning to eroding
soils.  The consultations with ORD will continue in an effort to
maximize the usefulness of PRZM towards pesticide exposure assessments.

Sediment Concentrations: Implications of the Current Conceptual Model
(Dr. Paul Hendley, Syngenta Crop Protection Inc.)

Dr. Hendley presented the following collection of issues with the
current conceptual model for estimation of sediment concentrations with
PRZM-EXAMS.  The constant mass of sediment in EXAMS versus the gradual
deposition/burying of sediment in real world water bodies leads to
overestimates of residues of relatively persistent compounds in pond
sediments.  Background data were presented on erosion and sedimentation
in water bodies.  For example, sediment is the most common reason for
TMDL listings.

Another issue raised is that all edge-of-field adsorbed residues from
PRZM are instantly transferred to the water body in EXAMS with assumed
instant desorption and distribution of the adsorbed material between
water column and the sediment bed being forced by the PRBEN parameter. 
Water and sediment leaving the field do not enter the pond; and this
outcome may not be appropriate for compounds of high Koc.  Also, some
PRZM scenarios predict unreasonably high edge-of-field losses of
sediment, which can lead to overestimation of exposure to adsorbed
residues in sediment and/or pore water.  Furthermore, neither
differential deposition of sediment nor the increase in suspended
sediment concentrations is addressed in EXAMS.  These deficiencies in
the current static aquatic water body conceptual model can cause
sensitivity analyses on EXAMS to be misleading.

Topics for further consideration include sediment sampling, which is
orders of magnitude more error prone than water sampling; the organic
content of suspended sediment; the ability of NHD+ to address these
issues; sediment transport in irrigation return flow; and variation of
Koc due to the “age” of sediments.  Furthermore, consideration of a
variable TSS, variable sediment depth, variable volume pond and/or
flowing water body is encouraged as a science issue, not a policy issue.
 Model users are encouraged to re-examine PRZM inputs when sediment
losses are unreasonable and when residue concentrations rank by year.

Factors Impacting Pesticide Runoff from Warm-Season Turf (Dr. Joe
Massey, MSU) 

Dr. Joe Massey described a multi-state effort to improve pesticide
runoff estimations from managed turf grasses.  Presently the project is
in the pre-modeling, and dataset development stage where two cool season
(MN, MD) and two warm season (MS, OK) locations have been selected for
conducting field studies.  This presentation covered only the field
studies on warm-season turf conducted in Mississippi.  The overall goal
of the project is to generate datasets that will improve estimation of
pesticide runoff from warm-season turf.  The specific objectives were to
determine how pesticide runoff is impacted by the following:

1.	Mowing height (0.5 in. vs. 2.0 in.)

2.	Grass species (Bermuda vs. Zoysia) 

3.	Plot size (four turf plot sizes ranging from 40 ft2 to 5000 ft2)

The site contained Brooksville silty clay soil, Hydrologic soil group:
Group D (high runoff potential), with an average infiltration rate of
0.7 ± 0.3 cm h-1 (N = 78), and slope: 3% with no cross-slope.  Thirty
different plots with were used with varying grass (Bermuda and Zoysia)
heights and plot size.  Considering a wide range of runoff potential and
relative ease of analysis, dicamba DMA, 2,4-D DMA, MCPP DMA, flutolanil,
and chlorpyrifos were selected as the test chemicals/pesticides for the
study.  A number of Quality Control measures were undertaken such as
washing and covering of diverters, troughs, and flumes prior to chemical
applications, verifying the rainfall rate by installing rain gauges
around the plots, and application monitoring.

Results from the study showed that chemicals behaved differently, as
expected.  The species effect (flutolanil) and mowing height interaction
(chlorpyrifos) could not be attributed to differences in water flow,
suggesting that retention mechanisms were the dominant factor in runoff
behavior of these chemicals.  It was observed from the data from the
different plot sizes that smaller plots which allowed direct collection
of runoff had the highest concentrations of flutolanil and chlorpyrifos.


It was concluded from the study that – (a) the effects of grass
species, mowing height, and plot size were compound-dependent; (b)
differences in retention rather than differences in water flow appeared
to govern pesticide behavior in these grasses; (c) runoff from Bermuda
grass turf was ‘scalable’ across a range of chemicals and plot
sizes; and (d) ‘micro’ plots may be most appropriate for low
solubility compounds since they may reduce sorption to plastic
collection components that were necessary for larger plot sizes. 

Kinetic Analysis of Metabolism Data (Dr. William Eckel, USEPA/EFED) 

Dr. William Eckel presented work completed to date on the harmonization
between USEPA and Canada’s PMRA, of degradation kinetics calculations.
 The basic problem is that metabolism data sets that do not obey
first-order kinetics must be used to derive first-order rate constants
for model input.  The approach taken is to divide the problem into two
tracks: Descriptive and Model Input.  The Descriptive track seeks to
reproduce the data as accurately and simply as possible, and uses
non-first order techniques if necessary.  The Model Input track seeks to
find the best compromise first-order fit to the data, while minimizing
over- or under-prediction of concentrations.

Status Report on Field Evaluation and REMM Modeling of a Pesticide
Runoff Buffer (Dr. Rob Everich, Makhteshim-Agan of North America) 

Dr. Robert Everich gave a presentation on the status of the Riparian
Ecosystem Management Model (REMM).  The REMM model was developed at the
Tifton, GA USDA-ARS lab by Don Wauchope and Richard Lowrance.  REMM is
intended to model the effectiveness of a pesticide runoff buffer to
reduce water and chemical loss.  The chemical studied was novaluron, an
insect growth regulator (IGR) that inhibits chitin synthesis.  Novaluron
is stable to hydrolysis and photolysis, essentially immobile in soil and
sediment, and degrades by metabolism in aerobic soil and water-sediment
systems.  A study is being conducted in Greenville, MS to provide data
for modeling of a cotton field with a Bermuda grass buffer.  The
objectives of the study are to quantify the volume of runoff from a
representative typical field, quantify the efficacy of the vegetative
buffer for chemical removal, and estimate edge of field concentrations
for use in risk assessment.  For the model to be used in risk
assessment, an interface between PRZM and REMM must be created, as well
as an interface between REMM and EXAMS. 

Forest Canopy Delivery of Pesticides to a Riparian Buffer Area (Dr.
Cliff Rice, USDA/ARS) 

Riparian buffers are used for their ability to limit the escape of
nutrients and agricultural chemicals from sites of application.  They
cause runoff water to enter the shallow ground water pool where sorption
and degradation can occur more rapidly.  Vegetative canopies of riparian
areas can sorb pesticides that are drifted or volatilized from treated
areas, and then release the pesticides when rainfall events occur. 
Sampling of rainfall using rain buckets in open areas, under trees, or
from water flowing down tree trunks has demonstrated this phenomenon. 
Concentrations of atrazine, metolachlor, and degradates of each were
lowest in direct rainfall from open areas, followed by rainfall falling
through the tree canopy, and stem flow.  Metolachlor was found in higher
concentrations in stem flow than atrazine, and stem flow had higher
metolachlor concentrations than rain falling through the canopy or
rainfall in open areas.  Riparian buffers that are near sites of
application can receive considerable loading of pesticides through
atmospheric pathways.  

GeoSTAC: GEOSpatial Tools and ACess (Dr. Patrick Havens, Dow
AgroSciences LLC, on behalf of the CropLife America Environmental
Exposure Working Group (CLA EEWG))

Dr. Havens presented and demonstrated the GeoSTAC (GEOSpatial Tools and
Access) product that was developed to overcome difficulties with the use
of geospatial information by providing a standard, freely distributable
set of citable data and tools.  Most of the data in GeoSTAC are federal:
National Atlas, HUCS, NHD, NLCD (USGS); AgCensus, NRCS land resource
layers, NASS (USDA); Bailey’s Ecoregions (USFS); USGS gauge locations
(NWIS); ERFI v.2 watersheds (USEPA); and more.  Additional preprocessed,
“value-added” datasets are also included.  The full data set takes
over 75 GB of disk space and the data are contained in
logically-arranged ESRI personal geodatabases.  

The analysis tools component of GeoSTAC is designed to increase
efficiency for routine analyses.  The tools were developed in ESRI
ArcGIS 9.1 (Service Pack 1+) format on the ArcView level and the source
code is freely available to CLA members and government agencies; code is
available to others upon request with CLA approval.  Five tools are
currently available; demonstrations accompany the presentation.  All of
the tools automatically produce new metadata for any datasets they
produce.

1.	The Feature and Attribute Extraction tool is a complex query tool
with multiple input layers that can, for example, isolate counties where
a given crop is grown.

2.	The Table Join tool automates joining Excel or other tabular data to
geospatial layers, along with some simple error checking.

3.	Doane’s Aggregation/Disaggregation tool can aggregate county-level
market research data to the CRD level and reapportion based on land use.

4.	The Data Aggregation and Summary tool aggregates and summarizes
datasets over a spatial dataset with different features (e.g.,
intersecting county sales estimates with salmon ESUs).

5.	The Linear Feature Summary tool summarizes the length and density of
linear features using polygon features, such as approximating the total
NHD stream length within a HUC.

GeoSTAC is currently being ported to ArcGIS 9.2.  The data are refreshed
annually and tool updates are under development.  Submission of the
GeoSTAC package to USEPA is planned for the first half of calendar 2007.
 A Quality Assurance/Quality Control (QA/QC) analysis of the data and
tools has been completed by the Research Triangle Institute.  The QA
report for the currently available data and tools will also be submitted
to USEPA within the year.

Wrap up

EFED is looking into dates for the next meeting and will announce the
date through the “empmlist” forum on the LYRIS list server at:  
HYPERLINK "https://lists.epa.gov/read/all_forums/" 
https://lists.epa.gov/read/all_forums/  

