UNITED
STATES
ENVIRONMENTAL
PROTECTION
AGENCY
WASHINGTON
D.
C.,
20460
OFFICE
OF
PREVENTION,
PESTICIDES
AND
TOXIC
SUBSTANCES
March
22,
2006
DP
Barcode:
D292330
PC
Code:
121601
MEMORANDUM
Subject:
EFED
Response
to
ARP
Comments
of
January
19,
2006
To:
Felecia
Fort,
Chemical
Review
Manager
Reregistration
Branch
3
(
RB3)
Special
Review
and
Reregistration
Division
(
7508C)

Donald
Stubbs,
Chief
Herbicide
Branch
Registration
Division
(
7505C)

From:
Michael
R.
Barrett,
Ph.
D.,
Senior
Chemist
Ronald
Parker,
Ph.
D.,
Senior
Environmental
Engineer
Environmental
Risk
Branch
V
Environmental
Fate
and
Effects
Division
(
7507C)

Through:
Mah
Shamim,
Ph.
D.,
Chief
Environmental
Risk
Branch
V
Environmental
Fate
and
Effects
Division
(
7507C)

Steven
P.
Bradbury,
Ph.
D.
Director
Environmental
Fate
and
Effects
Division
(
7507C)

Attached
is
the
EFED
response
to
Acetochlor
Registration
Partnership
comments
on
the
OPP
document
entitled
"
Drinking
Water
Exposure
Assessment
for
Acetochlor".
An
updated
version
of
the
document
(
February
16,
2006)
which
includes
changes
resulting
from
these
comments
has
been
reissued
under
the
title,
"
Revised
Drinking
Water
Exposure
Assessment
for
Acetochlor".
Acetochlor
Registration
Partnership
Comments
and
OPP/
EFED
Responses
ARP
Comment
(
Page
10)

The
abbreviation
"
SGW"
is
given
for
the
state
ground
water
monitoring
program;
however,
the
document
subsequently
uses
other
abbreviations
for
this
same
study,
including
"
GWM"
(
the
one
used
by
the
ARP)
and
"
SMP."
The
ARP
suggests
"
GWM"
be
used
throughout
in
order
to
maintain
consistency
with
the
borrowed
figures
and
tables.
Similarly,
the
abbreviation
"
SWM"
is
given
for
the
surface
water
monitoring
program,
but
other
acronyms
are
used,
such
as
"
SDWS."
The
ARP
suggests
"
SWM"
be
used
throughout.

EFED
Response:

EFED
has
changed
any
references
to
the
State
Groundwater
Monitoring
data
in
the
assessments
to
SGW
and
any
references
to
the
State
Drinking
Water
Supply
data
to
SDWS
to
be
consistent
from
section
to
section.
An
exception
to
this
would
be
any
reference
citations
from
the
ARP.

ARP
Comment
(
Page
11)

The
final
paragraph
suggests
that
the
SWM
monitoring
sites
"
do
not
fully
represent
the
sites
where
highest
concentrations
in
then
current
or
potential
surface
source
drinking
water
occurred."
The
ARP
reminds
the
Agency
that
the
SWM,
which
had
a
total
of
175
sites,
included
ALL
Community
Water
Systems
in
the
sampling
stratum
with
highest
corn
intensity
(>
20%),
and
that
the
Agency
participated
fully
in
site
selection
and
approved
site
selection
prior
to
the
collection
of
the
first
samples
in
March
1995.
Multiple
memos
of
conference
between
the
ARP
and
EPA
between
March
1994
and
June
1995
addressed
sites
and
conditions
relative
to
all
water
monitoring
programs.
In
a
March
30,
1995
memo
to
Donald
Stubbs
the
agreements
reached
during
the
January
13,
1995
meeting
are
detailed
as
outlined
in
the
original
March
1994
letter
setting
the
conditions
of
registration.
In
the
January
13,
1995
memo
it
states
under
Section
6.1
Surface
Water
Monitoring:
a.
"
EPA
agreed
that
the
surface
water
monitoring
(
SWM)
program
sampling
schedule, 
basic
design
elements
and
states
in
which
the
program
will
be
conducted .
are
acceptable
to
EPA."
Each
memo
of
conference
states
that
"
EPA
has
been
frequently
apprised
of
progress
through
individual
meetings
and
Quarterly
Reports"
(
January
1995
Memo
of
conference,
cited
in
the
Acetochlor
registration
agreement
and
addenda,
http://
www.
epa.
gov/
oppefed1/
aceto/
regagree.
htm).
The
ARP
stands
by
the
entire
site
selection
process
as
approved
and
regularly
reviewed
by
the
Agency.
Recent
drinking
water
monitoring
for
acetochlor
by
the
EPA
Office
of
Water
UCMR
(
http://
www.
epa.
gov/
ogwdw000/
ucmr/
data.
html#
1)
has
resulted
in
the
collection
of
well
over
30,000
drinking
water
samples
at
over
3000
sites
without
a
single
detection
of
acetochlor
in
drinking
water
>
2
ppb.
In
contrast,
the
ARP
surface
drinking
water
program
identified
several
sites
with
samples
over
2
ppb.
Clearly,
therefore,
the
SWM
effectively
captured
sites
at
the
high
end
of
the
distribution,
as
intended
by
the
Agency
when
it
approved
the
design
of
the
study.

EFED
Response:

The
ARP's
statements
reflecting
the
history
of
the
site
selection
process
for
sample
collection
appear
to
be
factually
correct.
However,
for
several
reasons,
EFED
believes,
that
these
sites
do
not
necessarily
present
a
complete
exposure
picture
for
acetochlor
usage
on
corn
during
this
period
and
that
the
SDWS
therefore
may
not
fully
represent
the
sites
where
highest
concentrations
of
acetochlor
may
have
occurred.
This
may
have
occurred
for
several
reasons:

(
1)
The
assumptions
made
by
the
ARP/
OPP
site
selection
team
were
not
entirely
accurate
in
predicting
the
eventual
area
of
highest
usage.
An
overlay
of
acetochlor
sales
data
with
the
SDWS
sampling
sites
shows
that
the
two
do
not
match
well
in
some
areas.
Future
acetochlor
usage
during
the
period
was
not
known
at
the
time
the
ARP
study
was
being
planned.
The
SDWS
design
was
based
upon
1992
corn
acreage
density,
not
on
acetochlor
usage
intensity
during
the
1995
 
2001
monitoring
period.

(
2)
Many
of
the
high
acetochlor
usage
areas
were
not
monitored
at
all.
Only
19%
of
the
total
ARP
sales
over
the
1995
to
2001
monitoring
period
were
in
counties
that
were
included
in
the
watersheds
which
encompass
the
ARP
CWS
monitoring
sites.
Only
29
of
the
top
200
acetochlor
usage
counties
over
the
first
10
years
of
registration
constituted
a
portion
of
the
watershed
for
one
or
more
ARP
surface
water
monitoring
sites.
It
is
likely
that
at
least
some
of
these
unmonitored
high
acetochlor
usage
regions
either
have
had
or
may
have
in
the
future
Community
Water
systems
utilizing
surface
water
sources.

(
3)
Many
of
the
extensively
monitored
areas
had
little
acetochlor
usage
during
the
monitoring
period,
particularly
the
sites
in
Pennsylvania
and
Delaware.

EFED
does
believe
that
the
ARP
SDWS
study
is
a
very
valuable
source
of
information
on
exposure
to
acetochlor
and
the
other
herbicides
monitored
in
many
of
the
major
corn
production
areas
of
the
United
States.
This
study
has
been
used
more
extensively
in
EFED
drinking
water
exposure
assessments
than
almost
any
other
single
monitoring
study.

ARP
Comment
(
Page
12):

The
first
paragraph
claims
that
the
focus
on
finished
drinking
water
is
"
especially
problematic
for
broader
interpretation."
The
EPA
agreed
to
the
selection
of
finished
drinking
water
as
the
only
matrix
of
interest
(
Section
6.1.
c,
Attachment
to
March
8,
1994
letter
setting
forth
conditions
of
registration
of
Acetochlor,
http://
www.
epa.
gov/
oppefed1/
aceto/
regagree.
htm)
in
a
meeting
on
January
13,
1995,
documented
in
the
addendum
to
the
acetochlor
registration
agreement.
Up
until
now,
nearly
12
years
later,
the
Agency
has
never
questioned
this
point.
The
ARP
unilaterally
added
raw
water
as
a
second
matrix
at
all
locations
that
used
granular
activated
carbon
(
GAC)
and
a
number
of
sites
using
powdered
activated
carbon
(
PAC).
As
stated
by
the
ARP
numerous
times,
we
have
no
evidence
that
conventional
drinking
water
treatment
(
without
activated
carbon)
substantially
reduces
acetochlor
concentrations
in
source
water.
Again,
the
Office
of
Water
data
confirms
the
adequacy
of
the
ARP
drinking
water
monitoring
program.

EFED
Response:

As
in
the
EFED
response
to
the
ARP
comment
on
page
11,
we
agree
with
the
ARP's
statements
reflecting
the
history
of
the
site
selection
process
for
sample
collection.
The
ARP's
citation
of
historical
meeting
results
appears
to
be
accurate.
However,
the
final
design
did
not
consider
all
potential
benefits
of
paired
finished
and
raw
water
sampling,
such
as
a
marked
increase
in
the
reliability
of
extrapolation
of
the
monitoring
program
results
to
other
sites.
Sampling
of
finished
water
alone
makes
it
impossible
to
separate
out
the
impact
of
water
treatment
on
the
observed
concentrations
and
therefore
to
separately
evaluate
the
impacts
of
acetochlor
usage
intensity,
weather,
stream
flow,
soil
characteristics,
topography,
and
other
factors
on
acetochlor
runoff.
One
cannot
assume
that
the
impact
of
water
treatment
(
whatever
type
of
treatment
system
is
used)
will
be
equivalent
for
all
CWS
(
and
indeed
it
appeared
to
be
highly
variable
for
the
few
ARP
monitoring
sites
that
did
have
paired
raw
and
finished
water
data).
EFED
is
using
this
data
set
to
the
greatest
extent
possible
in
estimating
exposure
to
acetochlor
from
drinking
water
supplies
both
from
sites
that
experienced
a
high
level
of
treatment
and
those
that
did
not.

ARP
Comment
(
Page
13)

In
the
final
paragraph,
the
Agency
questions
whether
the
observation
that
activated
carbon
lowers
parent
acetochlor
concentrations
by
50%
and
that
the
data
"
preclude
generalizing
this
as
a
predictable
effect
of
water
treatment".
The
statement
"
no
data
are
available
that
match
the
same
water
in
raw
and
finished
water"
is
both
incorrect
and
irrelevant.
It
is
incorrect
because
many
of
the
systems
utilized
reservoirs
with
residence
times
far
exceeding
the
holding
time
within
the
plant,
meaning
the
raw
and
finished
water
samples
represent
the
"
same
water,"
as
a
first
approximation.
It
is
also
incorrect
because
the
ARP
conducted
a
collaborative
study
with
several
third­
party
researchers
and
published
it
in
the
peer­
reviewed
literature
(
Gustafson
et
al.,
2003,
Activated
carbon
adsorption
of
chloroacetanilide
herbicides
and
their
degradation
products
from
surface
water
samples,
J.
Water
Supply
Res.
Technol.­
AQUA,
52:
443­
454.),
in
which
the
effect
of
PAC
on
acetochlor,
acetochlor
degradates,
other
herbicides
and
their
degradates
was
directly
measured
in
lab
jar
tests.
Finally,
it
is
irrelevant
because
of
basic
statistical
theory,
which
may
be
used
to
show
that
any
variance
introduced
by
the
effect
of
unknown
treatment
holding
time
within
the
water
plant
would
add
only
random
error
of
mean
zero,
without
introducing
any
systematic
bias
in
the
apparent
reduction
percentage.

EFED
Response:

EFED
agrees
that
the
matching
of
water
between
the
raw
and
treated
water
is
likely
to
be
better
for
most
reservoir
systems
than
river
or
stream
source
water
systems.
EFED
still
believes
that
the
inability
to
match
water
likely
introduces
significant
variability
in
some
paired
samples
that
arises
from
differences
in
the
effects
of
treatment.
EFED
has
modified
the
text
in
the
Drinking
Water
Assessment
to
reflect
that
while
specific
matching
of
raw
and
finished
water
is
not
available
for
the
ARP
study,
other
studies
of
treatment
effects
are
available
such
as
that
by
Gustafson
et
al.
(
2003).

There
does
seem
to
be
an
unusual
characteristic
of
the
dataset
with
regards
to
the
distributions
of
raw
and
finished
water
concentrations.
Most
of
the
highest
monitored
concentrations
were
found
in
treated
water.
This
means
that
one
or
more
of
the
following
situations
exist:
 
Effective
water
treatment
systems
removing
acetochlor
residues
were
not
employed
at
many
CWS
with
source
water
vulnerable
to
acetochlor
contamination
 
Effective
water
treatment
systems
were
employed
at
highly
vulnerable
sites
which
were
not
sampled
pretreatment
and
the
pretreatment
levels
would
have
been
much
higher
if
such
samples
had
been
taken
at
the
remaining
sites
 
Ostensively
effective
water
treatment
systems
were
employed,
but
the
techniques
shown
to
maximize
the
efficiency
of
activated
carbon
treatment
systems
by
Gustafson
et
al.
(
2003)
were
not
consistently
employed
and
therefore
in
some
systems
the
residues
of
raw
water,
even
if
perfectly
matched
with
finished
water,
would
not
have
been
significantly
lower
than
in
the
corresponding
finished
water
sample.

At
most
of
the
SDWS
study
sites
pretreatment
water
was
not
collected
and
there
is
no
possibility
of
comparing
the
acetochlor
levels
observed
with
the
pretreatment
levels.

ARP
Comment
(
Page
14)

In
the
first
full
paragraph,
the
ARP
respectfully
requests
that
the
Agency
produce
quantitative
data
supporting
its
implication
that
drinking
water
concentrations
might
be
up
to
3X
higher
than
those
shown
in
Table
1.
The
conservative
WARP
model
listed
in
this
table
shows
excellent
agreement
with
ARP
monitoring
results
for
both
"
peak"
and
median
concentrations.
The
lack
of
detections
in
the
Office
of
Water
monitoring
provides
further
evidence
that
no
such
"
uncertainty
factor"
should
be
applied
to
the
drinking
water
concentrations
measured
by
the
ARP.

EFED
Response
The
quantitative
assessment
of
how
residue
levels
might
vary
at
sites
the
ARP
did
not
monitor
will
be
included
in
the
new
uses
drinking
water
assessment.
The
ARP
may
wish
to
review
EFED's
detailed
assessment
at
that
time.
However,
generally,
the
following
factors
are
important
when
extrapolating
to
estimate
exposure
at
non­
ARP
monitored
sties:
 
Differences
in
usage
intensity
 
Differences
in
levels
of
water
treatment.
 
Differences
in
agricultural
practice
 
Differences
in
weather
The
concentrations
at
the
ARP
sites
are
also
expected
to
under­
estimate
the
actual
concentrations
at
the
monitored
sites.
Single
samples
taken
at
14
day
intervals,
will
almost
always
underestimate
actual
peaks
because
the
highest
concentrations
likely
to
occur
between
samples.

ARP
Comment
(
Page
28)

In
the
final
paragraph
before
section
5.7.3,
the
"
3­
2­
1"
restriction
should
be
mentioned
immediately
after
reference
is
made
to
the
3%
sands/
30
ft
restriction.

EFED
Response:

EFED
agrees
and
has
made
this
change.

ARP
Comment
(
Page
50)

The
column
labeled
"
Mean
Conc"
in
Table
14
should
be
footnoted
to
make
it
clear
these
are
not
time
weighted
means.
The
table
should
also
be
modified
to
indicate
what
time
periods
are
represented
and
the
actual
date
of
each
maximum.

EFED
Response:

EFED
notes
on
"
Mean
Conc"
and
"
Max
Conc"
columns
have
been
added
to
the
report.
