Page
1
of
23
UNITED
STATES
ENVIRONMENTAL
PROTECTION
AGENCY
WASHINGTON
D.
C.,
20460
OFFICE
OF
PREVENTION,
PESTICIDES
AND
TOXIC
SUBSTANCES
PC
Codes:
012501,
012502,
013802,
013803,
013806
DP
Barcode:
D309099,
D329699
Date:
June
27,
2006
MEMORANDUM
SUBJECT:
RESPONSE
TO
PUBLIC
COMMENTS.
The
Environmental
Fate
and
Effects
Division's
Response
to
Comments
on
EPA's
Phase
3
Reregistration
Eligibility
Decision
Document
for
the
Organic
Arsenic
Herbicides
(
OPP
Docket#
OPP­
2006­
0201).

TO:
Lance
Wormell,
Chemical
Review
Manager
Reregistration
Branch
II
Special
Review
and
Reregistration
Division
(
7508C)

FROM:
Keara
Moore,
Chemist
Environmental
Risk
Branch
III
Environmental
Fate
and
Effects
Division
(
7507C)

THROUGH:
Mark
Corbin,
Senior
Environmental
Scientist
Daniel
Rider,
Branch
Chief
Environmental
Risk
Branch
III
Environmental
Fate
and
Effects
Division
(
7505C)

The
attached
document
titled,
"
EFED's
Response
to
Comments
on
EPA's
Phase
3
Reregistration
Eligibility
Decision
Document
for
the
Organic
Arsenic
Herbicides"
was
generated
in
the
Phase
4
period
of
the
Proposed
Public
Participation
Process
to
address
comments
submitted
to
the
Agency
following
the
publication
of
the
Agency's
Phase
3
Reregistration
Eligibility
Decision
(
RED)
Document
for
the
Organic
Arsenic
Herbicides
in
the
Federal
Register
April
5,
2006.
EFED's
portion
of
this
response
addresses
comments
from
The
MAA
Research
Task
Force
which
consists
of
APC
Holdings
Page
2
of
23
Company/
Drexel
Chemical
Company,
KMG­
Bernuth,
and
Luxembourg­
Pamol,
Inc.;
The
Florida
Department
of
Agriculture
and
Consumer
Services
(
FDACS);
and
Florida
Department
of
Environmental
Protection.
The
attached
document
is
EFED's
response
to
those
comments.

EFED's
chapters
in
support
of
the
Agency's
Phase
RED
are
the
Drinking
Water
Assessment
(
DP
Barcode
D309097)
and
an
Ecological
Risk
Chapter
(
DP
Barcode
D327974).
Comments
relevant
to
those
documents
are
addressed
in
this
response
and
no
revisions
have
been
made
to
the
original
document.
Page
3
of
23
TABLE
OF
CONTENTS
I.
Introduction
...................................................................................................................
4
of
23
II.
MAATF
comments
......................................................................................................
4
of
23
Comment
IV­
A:
Modeling
Approach
Comment
IV­
B:
Monitoring
Data
Comment
V­
A:
Percent
Cropped
Area
for
Turf
Comment
V­
B:
Spot
Treatment
on
Turf
Comment
VI:
Use
of
Open
Literature
Studies
Comment
VII:
Comparison
to
MCL
and
SSLs
Comment
XI:
Ecological
Risk
Assessment
III.
FDACS
comments.....................................................................................................
21
of
23
IV.
FDEP
comments
.......................................................................................................
23
of
23
Page
4
of
23
EFED'S
RESPONSE
TO
COMMENTS
FOR
THE
ORGANIC
ARSENIC
HERBICIDES
I.
Introduction
The
following
is
EFED's
response
to
comments
on
the
Phase
3
Reregistration
Eligibility
Document
(
RED)
for
the
organic
arsenic
herbicides,
generated
in
response
to
the
comments
submitted
to
the
public
docket
by
the
registrants
and
by
departments
of
the
State
of
Florida
to
the
Agency
following
the
publication
of
the
preliminary
risk
assessment
(
Phase
3)
on
the
organic
arsenic
herbicides
in
the
Federal
Register
April
5,
2006
to
begin
Phase
4
of
the
Public
Participation
Process.
Some
of
the
responses
serve
as
clarification
or
a
restatement
of
Agency
policies
and
guidance
and
it
is
hoped
that
this
will
provide
a
greater
understanding
of
the
Agency's
position
and
procedures
on
these
matters.
Comments
concerning
Agency
assessment
of
the
environmental
fate
and
persistence
of
the
organic
arsenic
herbicides,
interpretation
of
available
data,
generation
of
drinking
water
estimates,
and
assessment
of
ecological
risk
were
submitted.
These
comments
are
addressed
individually
below.

Comments
addressing
the
availability/
efficacy
of
substitute
pesticides,
the
toxicology
of
arsenicals,
and
the
dietary
and
occupational/
residential
exposure
will
be
addressed
in
separate
documents
by
the
appropriate
divisions:
the
Biological
and
Economics
Assessment
Division
(
BEAD)
and
the
Health
Effects
Division
(
HED).

II.
MAATF
Comments
IV
 
A.
EPA
should
adopt
a
modeling
approach
that
is
consistent
with
actual
data
and
real­
world
processes.

OPP
models
the
organic
arsenicals
reaching
drinking
water
and
arbitrarily
assumes
that
they
are
totally
converted
to
inorganic
arsenic
in
the
water.
This
conceptual
approach
does
not
account
for
the
data
demonstrating
that
transformation
of
organic
arsenic
to
carcinogenic
inorganic
arsenic
can
occur
only
in
the
soil,
and
only
to
a
very
limited
extent,
and
that
only
inorganic
arsenic
that
is
not
bound
to
the
soil
can
be
transported
to
surface
water.
EPA
should
adopt
a
different
modeling
approach,
one
that
is
consistent
with
the
actual
data
and
real­
world
processes
and
that
focuses
only
on
inorganic
arsenic
for
cancer
risk
assessment.
The
cancer
risk
assessment
should
be
based
on
the
estimated
concentrations
of
inorganic
arsenic
that
could
be
produced
in
soil
and
then
transported
to
water
(
i.
e.,
excluding
the
fraction
that
is
bound
to
the
soil).
This
is
the
only
approach
that
is
consistent
with
the
data
and
with
science.
See
Attachments
5
and
6.

(
Please
note:
Some
statements
included
in
the
referenced
attachments
are
more
relevant
to
other
sections
of
these
registrant
comments
and
will
be
addressed
in
those
sections:
Comparison
of
modeling
results
with
monitoring
data,
section
IV­
B;
use
of
a
PCA
for
turf,
section
V­
A;
use
as
a
spot
treatment
on
turf,
section
V­
B)
Page
5
of
23
EFED
answer:
The
registrant
presents
two
primary
objections
to
EFED's
general
aquatic
modeling
approach:
that
the
modeling
is
not
consistent
with
actual
data
and
realworld
processes
and
that
it
arbitrarily
assumes
total
conversion
of
organic
arsenicals
to
inorganic
arsenic.
These
comments
are
directed
at
the
approach
used
in
modeling
inorganic
arsenic
as
a
metabolite
of
applied
MMA
and
DMA.
In
addition
to
commenting
on
the
general
conceptual
modeling
approach,
the
registrant
comments
on
specific
modeling
inputs
used
in
estimating
exposure
to
MMA
and
DMA.
These
comments
will
be
addressed
individually
below.

Conceptual
model
for
estimating
exposure
to
inorganic
arsenic
(
Attachment
5,
p.
1)
For
assessing
risk
from
pesticides
with
toxic
metabolites
when
data
are
not
robust
enough
to
model
parents
and
degradates
separately,
a
typical
approach
is
to
estimate
exposure
to
the
total
toxic
residues,
including
in
the
estimate
both
the
parent
compounds
and
the
toxic
metabolites.
When
organic
arsenicals
are
applied,
the
arsenic
may
change
forms
but
is
otherwise
persistent,
so
this
typical
approach
would
involve
modeling
with
no
degradation
in
order
to
estimate
exposure
to
total
arsenic,
the
sum
of
all
arsenical
species
present.
In
this
case,
however,
refinement
beyond
the
typical
approach
is
required;
because
toxicity
for
each
arsenic
species
needs
to
be
assessed
independently,
exposure
to
each
species
must
also
be
assessed
independently.
These
individual
exposures
must
be
estimated
in
ways
that
are
sufficiently
conservative
for
each
species.
Because
a
conservative
exposure
estimate
for
a
parent
compound
is
likely
to
underestimate
exposure
to
the
metabolite
and
vice
versa,
separate
approaches
had
to
be
used
to
estimate
exposure
to
parents
and
to
metabolites.

Limited
data
are
available
to
describe
the
transformation
of
applied
organic
arsenicals
to
inorganic
arsenic,
a
process
which
appears
to
be
quite
variable
with
variable
environmental
conditions.
Without
data
to
adequately
define
input
parameters,
direct
modeling
of
exposure
to
inorganic
arsenic
as
a
metabolite
was
not
an
option.
Instead,
EFED
assumed
that
the
modeled
estimates
of
exposure
to
applied
parent
compounds
represented
a
reasonable
upper
bound
on
the
total
amount
of
arsenic
that
would
reach
surface
water
in
any
form.
The
modeled
EDWCs
for
parent
compounds
were
therefore
also
reported
as
the
EDWCs
for
total
arsenic,
with
the
units
converted
from
µ
g
a.
i./
L
to
µ
g
arsenic/
L
in
order
to
be
comparable
to
the
relevant
endpoints
and
regulatory
limits.
The
variability
of
arsenicals'
environmental
fate
and
the
lack
of
adequate
data
to
describe
all
of
the
relevant
processes
make
it
impossible
to
define
with
certainty
what
portion
of
this
total
arsenic
in
surface
water
will
be
present
as
any
individual
arsenical
species.
Therefore,
the
risk
assessment
states
that
the
total
arsenic
EDWC
also
represents
the
maximum
potential
concentration
of
inorganic
arsenic.
As
acknowledged
in
the
assessment,
this
conceptual
approach
"
does
not
necessarily
represent
the
actual
physical
process
[
since]
inorganic
arsenic
is
likely
formed
through
metabolism
in
the
soil,
rather
than
after
reaching
surface
water."
Despite
this,
this
approach
is
expected
to
provide
a
reasonable
estimate
of
the
amount
of
inorganic
arsenic
that
may
reach
surface
water.
The
modeling
accounts
for
binding
to
soil
and
other
routes
of
dissipation.

This
discussion
of
potential
exposure
to
inorganic
arsenic
is
not
equivalent
to
assuming
total
conversion
of
organic
arsenicals
to
inorganic
arsenic,
as
stated
by
the
registrant.
Page
6
of
23
This
estimated
total
arsenic
value
is
presented
as
the
maximum
possible
concentration
of
inorganic
arsenic,
with
the
likely
extent
of
possible
conversion
characterized
extensively
based
on
available
open
literature
reports
and
registrant
submitted
studies.
Contrary
to
the
registrant's
assertion
that
this
transformation
occurs
"
only
to
a
very
limited
extent,"
observed
metabolism
of
MMA
and
DMA
to
inorganic
arsenic
in
the
laboratory
has
ranged
from
undetected
after
several
years
to
more
than
80%
transformation
in
several
months.
To
account
for
100%
conversion,
maximum
exposure
to
inorganic
arsenic
would
have
been
estimated
by
modeling
the
applied
herbicide
as
arsenic
with
no
metabolism.
Comparing
the
reported
EDWCs
to
values
calculated
this
way
shows
that
the
peak
value
represents
88%
conversion
while
the
cancer
value
represents
72%
conversion,
reasonable
upper
bounds
given
the
range
of
conversion
observed
in
the
laboratory.

The
registrant
recommends
an
alternative
modeling
approach
for
assessing
exposure
to
inorganic
arsenic.
The
first
step
of
their
approach
is
to
model
exposure
as
if
no
metabolism
occurs,
in
order
to
define
the
maximum
amount
of
applied
arsenic
that
might
reach
surface
water
in
any
form.
The
second
step
is
to
estimate
the
amount
of
arsenic
that
will
be
present
as
parent
compound
by
running
the
model
again,
this
time
with
potential
metabolism
accounted
for.
The
difference
between
the
modeled
concentration
of
total
arsenic
with
no
metabolism
and
the
concentration
with
parent
compound
only,
the
registrant
suggests,
would
represent
the
maximum
potential
concentration
of
inorganic
arsenic.
To
demonstrate
this
approach,
the
registrant
models
exposure
to
total
arsenic
from
the
turf
use
and
subtracts
the
parent
compound
EDWC
reported
in
OPP's
assessment.

If
it
were
possible
to
determine
the
precise
amount
of
parent
compound
that
would
reach
surface
water,
this
suggested
approach
would
provide
useful
results.
The
variability
of
arsenicals'
environmental
fate,
though,
as
well
as
the
limitations
inherent
in
modeling,
make
precise
determination
impossible.
Instead,
in
order
to
be
protective
in
considering
potential
risks,
OPP
must
make
conservative
estimates
of
potential
exposure.
A
modeling
approach
that
is
conservative
for
the
parent
compound
is
likely
to
underestimate
exposure
to
the
metabolite,
and
so
use
of
the
conservatively
modeled
parent
EDWC
in
this
approach
will
lead
to
underestimation
of
the
inorganic
arsenic
EDWC.
The
registrant's
proposed
modeling
approach
does
not
adequately
describe
how
to
assure
that
EDWCs
for
inorganic
arsenic
as
a
metabolite
are
sufficiently
conservative.
An
appropriately
conservative
approach
similar
to
that
suggested
by
the
registrant
would
start
with
estimating
total
arsenic
exposure
as
described
in
their
first
step.
To
determine
an
upper
bound
on
the
amount
of
that
arsenic
that
may
be
present
in
the
inorganic
form,
instead
of
subtracting
a
conservative
estimate
of
exposure
to
the
parent
compound,
the
percent
conversion
observed
in
laboratory
studies
could
be
applied
to
this
total
arsenic
estimate.
As
discussed
above,
that
approach
shows
that
the
Agency's
EDWCs
reasonably
represents
the
potential
for
conversion
to
inorganic
arsenic.

Biphase
dissipation
of
MMA
and
DMA
(
Attachment
5,
p.
7)
The
registrant
notes
that
OPP's
modeling
does
not
account
for
the
biphasic
nature
of
arsenicals'
aerobic
soil
metabolism
as
observed
in
several
studies.
The
biphasic
nature
of
Page
7
of
23
arsenical
metabolism
is
characterized
in
the
risk
assessment,
but
the
Agency
states
that
"
because
of
the
variability
of
sorption
processes,
there
are
insufficient
data
to
allow
for
determination
of
two
separate
rates."
Instead,
metabolism
is
modeled
as
a
first
order
process,
a
choice
expected
to
be
protective
given
the
importance
of
chronic
exposure
to
arsenicals
because
"
a
long
term
half­
life
that
is
less
accurate
in
representing
the
faster
initial
rate
of
metabolism
 
better
portrays
the
overall
transformation."

Based
on
the
results
of
a
single
study,
the
registrant
incorporates
biphasic
half­
lives
into
aquatic
exposure
modeling,
finding
30­
year
mean
annual
concentrations
of
MMA
on
turf
ranging
from
69.1
ppb
to
76.4
ppb,
depending
on
various
assumptions.
Comparing
these
results
to
the
value
reported
in
the
assessment,
74.6
ppb,
supports
OPP's
assumption
that
modeling
arsenical
metabolism
as
a
first­
order
process
is
appropriately
protective.

Half­
Life
for
DMA
(
Attachment
5,
p.
15)
The
registrant
asked
for
more
detail
on
how
the
DMA
aerobic
soil
metabolism
half­
life
was
determined.
As
stated
in
the
document,
the
half­
life
modeled
is
the
upper
90%
confidence
limit
on
the
mean
of
six
values
observed
in
four
studies.
These
studies
include
three
open
literature
studies
with
DMA
applied
directly
and
one
registrant
study
for
which
the
Agency
estimated
a
half­
life
for
DMA
based
on
its
formation
and
decline
following
application
of
MMA.
The
StatMost
statistical
program
was
used
to
perform
non­
linear
regression
on
this
data
using
the
Levenberg­
Marquardt
least
squares
method
for
curve
fitting.
The
data
was
fit
to
equations
for
two
first
order
processes
in
series,
based
on
the
simplifying
assumption
that
metabolism
occurred
as
two
individual
first
order
processes:
MMA
transforming
to
DMA
followed
by
transformation
to
inorganic
arsenic.
The
equations
are
shown
below
where
A
represents
parent
MMA,
B
represents
DMA
as
a
metabolite,
and
C
represents
inorganic
arsenic
as
a
metabolite.

Although
there
is
uncertainty
in
this
calculation,
it
still
provided
important
information
because
it
is
based
on
the
only
DMA
metabolism
data
available
from
a
study
conducted
according
to
GLP
standards
and
run
for
a
full
year.
The
half­
life
calculated
from
the
formation
and
decline
data,
approximately
240
days,
was
nearly
the
same
as
the
upper
bound
half­
life
determined
from
the
three
open
literature
studies,
so
inclusion
of
this
value
in
calculating
the
upper
bound
half­
life
did
not
change
the
overall
result.
Instead,
it
provided
additional
support
that
the
half­
life
estimated
from
the
open
literature
studies
was
reasonable
in
spite
of
the
uncertainty
in
those
studies.
Page
8
of
23
Modeling
of
DMA
as
a
degradate
(
Attachment
5,
p.
16)
The
registrant
argues
that
it
is
inappropriate
to
model
exposure
to
DMA
as
a
metabolite
of
MMA
by
simulating
application
of
DMA
on
the
same
day
as
MMA
despite
the
documented
lag
time
to
yield
DMA
from
MMA.
This
uncertainty
is
acknowledged
in
the
drinking
water
assessment,
which
states
that
"
additional
uncertainty
results
from
modeling
the
DMA
metabolite
as
if
it
were
applied
at
the
same
time
as
the
MMA
parent.
Metabolism
to
DMA
is
not
a
rapid
transformation
and
so
in
fact,
when
present,
the
metabolite
will
appear
at
a
later
date
dependent
on
the
environmental
conditions."
Given
the
variability
in
rates
of
transformation
of
MMA
to
DMA
and
the
slow
time
period
over
which
it
actually
occurs,
there
would
be
uncertainty
in
any
estimate
of
"
application
date"
of
DMA
as
a
metabolite
and
EFED
believes
that
the
modeling
approach
used
is
reasonable
given
the
limitations
of
the
data.

The
registrant
also
states
that
"
it
appears
that,
by
applying
the
35%­
transformed
DMA
on
the
same
date
as
the
maximum
label
rate
of
MMA,
EFED
applied
total
MMA
at
a
rate
that
exceeds
the
label
limit."
Because
it
is
impossible
to
define
with
certainty
what
portion
of
this
total
arsenic
in
surface
water
will
be
present
as
any
individual
arsenical
species,
EFED
presents
the
estimated
exposures
of
MMA
as
parent
compound
and
DMA
as
a
metabolite
as
"
maximum
potential
concentrations
for
each
species
that
are
not
additive
to
a
total
possible
concentration"
so
this
is
not
equivalent
of
modeling
in
excess
of
the
maximum
labeled
rate.

IV
 
B.
EPA
should
compare
modeled
arsenic
drinking
water
concentrations
to
actual
community
surface
water
drinking
supplies.

EPA
should
compare
the
modeled
arsenic
drinking
water
concentrations
to
arsenic
concentrations
measured
in
actual
community
surface
water
drinking
supplies,
all
of
which
contain
arsenic
levels
that
are
significantly
lower
than
the
arsenic
MCL.
Such
a
comparison
will
demonstrate
that
even
when
MSMA
and
possibly
DMA
are
used
in
cotton­
growing
states,
the
arsenic
levels
in
drinking
water
in
those
states
are
considerably
below
the
MCL.
It
is
noteworthy,
that
the
drinking
water
concentrations
modeled
by
OPP
reflect
only
the
arsenic
present
due
to
organic
arsenic
herbicide
use,
whereas
the
actual
drinking
water
contain
arsenic
compounds
from
the
herbicides,
if
any,
plus
additional
sources,
including
natural
and
anthropogenic.
Thus,
monitored
arsenic
concentrations
should
be
higher,
not
lower
than
the
modeled
concentrations.
Absent
such
a
comparison
and
corresponding
reduction
in
the
modeled
concentrations,
the
risk
assessment
is
not
scientifically,
and
hence
legally,
defensible.

In
summary,
the
estimated
drinking
water
concentrations
were
generated
using
an
erroneous
approach
with
overly
conservative
assumptions
resulting
in
modeled
results
that
do
not
reflect
actual
community
surface
water
concentrations
consumed
by
the
U.
S.
population,
as
discussed
in
Attachments
2
and
5
to
this
letter.
Even
in
the
states
that
use
the
most
MSMA,
surface
water
drinking
concentrations
are
well
below
the
MCL
for
total
arsenic.
Moreover,
comparing
the
arsenic
drinking
water
concentrations
from
various
states
show
that
there
is
no
correlation
between
the
areas
where
methylated
arsenic
Page
9
of
23
herbicides
are
used
and
the
levels
of
arsenic
in
water
(
see
Table
3
in
Attachment
2).
EPA
must
correct
the
modeling,
which
the
data
show
are
clearly
erroneous
to
ensure
a
scientifically
defensible
risk
assessment.

EFED
answer:
The
registrant
presents
Community
Water
System
(
CWS)
arsenic
data
from
26
states,
compiled
by
the
Office
of
Water,
and
compares
these
values
to
the
modeled
surface
water
EDWCs
from
the
arsenical
risk
assessment.
For
each
individual
water
supply
system
within
those
states,
the
dataset
includes
a
single
multiple
year
mean
arsenic
concentration.
Based
on
the
75th
percentile
of
these
means
for
each
state,
the
registrant
asserts
that
all
of
these
water
systems
contain
arsenic
levels
significantly
lower
than
the
MCL,
that
there
is
no
correlation
between
arsenic
levels
and
agricultural
use
of
arsenicals,
and
that
the
modeling
is
therefore
erroneous.

Comparison
of
monitoring
values
with
modeling
results
can
be
useful
but
needs
to
be
done
with
care.
Modeling
is
intended
to
estimate
1
in
10
year
peak
and
average
exposures
while
monitoring
is
likely
to
underestimate
exposures
and
the
frequency
of
occurrence
due
to
limited
number
of
samples
and
extended
time
between
sampling
events.
The
table
of
monitoring
data
cited
by
the
registrant
provides
important
information
about
the
amount
of
arsenic
that
may
be
found
in
treated
drinking
water
reaching
populations
served
by
public
water
supplies
with
surface
water
sources.
When
comparing
these
results
to
modeled
values,
however,
it
is
important
to
keep
in
mind
the
goals
of
the
sampling,
designed
by
individual
CWS,
and
of
the
subsequent
statistical
data
analysis,
performed
by
the
Office
of
Water,
which
are
different
than
the
goals
of
the
modeling
in
this
risk
assessment.
The
sampling
was
conducted
as
compliance
monitoring
while
the
statistical
analysis
was
carried
out
to
identify
long
term
trends
in
arsenic
exposure.
In
support
of
these
objectives,
the
methods
used
lead
to
results
that
are
not
directly
relevant
to
EFED's
objective
of
protecting
water
resources
from
the
localized,
temporal
impacts
of
pesticide
use.

The
primary
issue
resulting
from
the
different
data
objectives
is
that
these
CWS
data
are
mostly
for
samples
taken
after
water
treatment,
so
they
do
not
represent
the
amount
of
arsenic
actually
reaching
water
sources.
The
Agency
has
an
interest
in
protecting
source
water
as
well
as
treated
water
and
does
not
include
treatment
in
its
modeling.
The
modeled
concentrations
estimate
the
amount
of
arsenic
in
water
prior
to
treatment
and
are
therefore
not
directly
comparable
to
the
monitored
concentrations.

Another
important
difference
is
that
CWS
sampling
is
not
targeted
to
arsenical
use
areas
or
times.
Even
in
states
that
are
users
of
arsenicals,
large
numbers
of
CWS
sampling
sites
are
outside
of
arsenical
use
areas.
This
is
shown
for
cotton
in
Figure
1
(
use
of
arsenicals
on
turf
is
not
included
in
this
map).
Additionally,
only
a
single
value
for
each
system
 
the
mean
over
a
period
of
3
to
20
years
 
is
included
in
the
statistical
analysis.
Based
on
the
sampling
summary
in
the
OW
report,
the
number
of
samples
included
in
the
analysis
averaged
less
than
one
sample
per
year
for
systems
in
most
states.
The
data
from
Texas,
cited
by
the
registrant
as
evidence
of
a
lack
of
correlation
between
use
and
exposure,
can
be
considered
as
an
example.
Less
than
50%
of
the
CWS
intakes
in
Texas
are
in
counties
Page
10
of
23
!(
!(
!(

!(

!(
!(
!(

!(
!(
!(
!(

!(
!(

!(

!(
!(
!(
!(
!(
!(!(

!(
!(
!(!(

!(
!(
!(

!(
!(
!(
!(
!(
!(
!(
!(!(

!(
!(

!(!(
!(
!(
!(!(

!(
!(
!(
!(
!(!(

!(
!(
!(
!(
!(!(

!(
!(
!(
!(
!(!(
!(!(

!(!(
!(
!(
!(!
(
!(

!(!(
!(

!(
!(
!(
!(
!(
!(
!(
!
(
!(!(
!(!(
!(!(

!(
!(
!(
!(

!(
!(
!(
!(
!(
!(
!
(
!(
!(

!(
!(
!(!(

!(
!(
!(
!(
!(!(
!(!(
!(
!
(

!(!(
!(
!(

!(
!(
!(
!(!(
!(!(
!(
!(
!(
!(!(
!(
!(

!(
!(

!(
!(
!(

!(
!(
!(
!(
!(
!(!(

!(
!(
!(
!(

!(!(
!(
!(

!(
!(
!(

!(

!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(

!(
!(
!(

!(!(
!(

!(
!(

!(
!(
!(
!(
!(

!(
!(!(!
(
!(
!(

!(

!(

!(
!(
!(
!(
!(
!(
!(
!(!(
!(
!(
!(!(
!(
!(!(!
(
!(
!(!(

!(
!(

!(!(
!(
!(
!(
!(
!(!(
!(
!(
!
(
!(
!(
!(!(!(!(
!(!(
!(
!(
!(
!(
!(!(
!(!(
!
(!(
!(
!(
!(!
(
!(

!(
!(

!(
!(!(
!
(
!(
!(
!(
!(!(
!(
!(
!(
!(!(
!(!(
!(
!(
!(
!(
!(
!(!(

!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(

!(
!(

!(!(!(
!(!(
!(

!(
!(
!(!(!
(!(
!(
!(
!(
!(
!(!(
!(!(
!(
!(
!(!(
!(
!(
!(
!(
!(
!(
!(!
(!(!(
!(
!(
!(
!(
!(

!(!(!(
!(
!(

!(

!(!(!(
!(
!(
!(!
(
!(
!(
!(!(
!
(
!(

!(
!(
!(
!(!(
!(
!(
!(

!(
!(
!(
!(
!(!(

!(
!(
!(
!
(
!(!(!(
!(
!

(!(!(
!(
!(
!(
!(!
(!(
!(!(
!(!(!(!(!(

!(
!(!(
!(

!(
!(!(
!(!(!(
!(

!(
!(
!(

!(
!(
!(

!(
!(
!(

!(!
(
!(
!(
!(

!(
!(
!(
!(
!(
!(!(
!(
!(!(

!(!(!(
!(!(
!(
!(
!(
!(
!(!(!(
!(
!(!(
!(!(!(
!(!(!(
!(!(!(

!(
!(
!(
!(
!(
!(
!(!
(

!

(!
(
!(!(
!(
!(!(
!(

!(
!(
!(
!(
!(
!
(
!(!(
!(!
(
!

(!(
!(
!(
!(
!(!(
!(
!(
!(!(!(!(!(
!(!(!(
!(!(
!(
!(
!(
!(

!(!
(
!(
!(!(
!(
!(
!(
!(!(
!(
!(
!(
!(!
(
!(!(
!(!(!(
!(
!(!(
!(
!(!(!(
!(!
(
!(!(!(
!(!(!(
!(
!(
!(
!(
!(
!(

!(
!(
!(!(
!(
!(
!(

!(
!(!(
!(!(
!(
!(
!
(
!(
!(
!(!(

!(
!(
!(
!(!(!(!(
!(!(
!(!(
!(
!(
!(!(
!(
!(

!(
!(

!(!(
!(
!(
!(
!(!(!(
!(
!(!(
!(
!(
!(
!(
!(
!(
!(!(
!(
!(

!(
!(
!(
!(
!(
!(
!(
!(
!(

!(
!(!(
!(
!(
!(
!(!(
!(!(!(!(

!(
!(!(
!(
!(

!(
!(!(!(!(
!(
!(
!(
!(
!(!(!(!(
!(
!(
!(
!(
!(
!(
!(
!(!(
!(
!(!(
!(!
(!(!(!(!(
!(
!(
!(

!(!(
!(!(
!
(
!(
!(!(!(
!(

!(
!(
!(

!(
!(
!(
!(
!(
!(

!(!(
!(
!(!(
!(!(
!(
!(
!(!(!
(
!(
!(!
(!(
!(
!(!(
!(
!(
!(
!(!(
!(!(!(!(!(!(!(!(
!(
!(!(

!(
!(
!(
!(
!(!(
!(
!(!(

!(
!(!(!
(
!(!(
!(!(
!(
!(
!(

!(!(
!(!
(
!(
!(
!(
!(
!(!(
!(
!(
!(!(
!(!(
!(

!(
!(
!

(!(
!(
!(
!(!(
!(!
(!(
!(
!(!(

!(
!(
!(!(
!(
!(
!(
!(
!(

!(!(
!(!(
!(
!(
!(
!(!(!(

!(
!(
!(!
(
!(
!(
!(
!(!(
!(!(!(
!(
!(
!(
!(
!(
!(
!(!(!(

!(
!(
!(
!(
!(
!(

!(
!(!(!(!(!(!(!(!(!(!(!(!(
!(
!(!(
!(
!(

!(
!(
!(
!(
!(
!(!(
!(
!(!(
!(!(

!(!(
!(
!(
!(
!(
!(
!(
!
(
!(
!(!(!(
!(!(
!(
!(!(
!(
!(!(
!(
!(
!(
!(
!(!(
!(!(!(
!(
!(
!(
!(!(
!(
!(
!(

!(
!(!(!(!(!(!(

!(
!(
!(!(
!(
!(

!(!(
!(
!(
!(!
(!(!
(!(!(
!
(

!(

!(
!(!(!(!(
!(
!(
!(
!(!
(
!(
!(!(
!(
!(
!(
!(!(

!(
!(
!(
!(!(
!(
!(
!(
!(
!(

!(!(
!(!(!
(!(!(
!(
!(
!(!
(
!(!(!(
!(
!(
!(
!(!(!(
!(
!(!(

!(!(
!(
!(
!(
!(

!(
!(
!(
!(
!(!(

!(
!(!(
!(!
(
!(
!(
!(!(
!(!(!(!(
!(

!(
!(!(!(
!(
!(
!(
!(
!(
!(

!(
!(
!(!(
!(!(
!(
!(!(!(!(
!(
!(!(!(!(!(!(
!(!(!(!(
!(
!(!(
!(!(!(!(!
(!(
!(

!(!(
!(!(!(!(
!(
!(!(!(!(!(
!(!(
!(!(
!(!(
!(!(
!(!(!(!(!(!
(!(
!(!(
!(!(!(!(!(!(!(!(!(!(
!(!(
!(!(!(!(
!(!(
!(!(!(
!(
!(
!(
!(!(
!(!(!(
!(
!(!(!(!(
!(

!(
!(!(!(!(!(!(!(
!(!(
!(
!(!(!(!(
!(!(!(
!(!(!(

!(
!(!
(!(!(
!(!(
!(!(
!(
!(!(!(
!(!(
!(!(!(
!(!(
!(

!(
!(!(
!(

!(
!(

!(!(

!(!
(
!(!(
!(

!(!(!(
!(
!(
!(
!(!(

!(!(
!(
!(!(
!(

!(
!(!(!
(
!(

!(
!(
!(
!(!(
!(!(
!(
!(

!(
!(
!(!(
!(!(

!(
!(
!(
!(
!(
!(

!(
!(
!(
!(
!(
!(
!(!(
!(
!(
!(
!(

!(!(!(!(!(!
(
!(
!(!(
!(
!(
!(!(
!(

!(
!(
!(
!(!(
!(!(
!(

!(
!(
!(
!(!
(!(
!(
!(

!(!(
!(!(
!(

!(
!(

!(
!(!(
!(
!(
!(

!(!(
!(!(!(
!(
!(
!(

!(

!(
!(
!(!(
!(
!(
!(

!(!(
!(
!(!(
!(
!(!(

!(
!(

!(
!(
!(

!(
!(!(
!(
!(

!(!(
!(
!(
!(

!(
!(

!(!(
!(
!(
!(
!(
!(
!(

!(
!(
!(
!(
!(
!(
!(
!(!(!(
!(

!(
!(!(

!(
!(

!(!(
!(
!(

!(
!(!
(
!(
!(
!(

!(
!(

!(
!(
!(
!(!(
!(
!(
!(
!(
!(
!(
!(
!(!(
!(
!(
!(!(
!(
!(

!(
!(!(
!(
!(
!(!(
!(!(

!(
!(
!(

!(
!(
!(
!(

!(
!(
!(

!(

!(
!(
!(
!(
!(

!(
!(!(
!(
!(

!(

!(!(
!(
!(
!(

!(
!(

!(
!(
!(!(!(
!(
!(
!(
!(
!(

!(!(
!(!(
!(!(!(
!(!(

!(
!(
!(

!(
!(
!(
!(
!(!(
!(

!(!(
!(
!(!

(!
(

!(
!(!(
!(
!
(

!(
!(
!(
!(!(!(
!(!(
!(
!(!(

!(

!(!
(!(
!(
!(
!(
!(!(

!(
!(
!(
!(!(!(
!(!(!(
!(
!(
!(
!(
!(!(
!(!(
!(!(
!(
!(

!(
!(

!(
!(
!(!(
!(
!(
!(
!(!(

!(
!(!(
!(
!(!(
!(!(

!(
!(
!(!(!(

!(!
(
!(!(
!(!(
!(

!(!(!
(
!(
!(

!(

!(
!(!(
!(
!(!(!(
!(

!(!(
!(!(!(
!(
!(!(!(
!(

!(
!(
!(!(
!(
!(!(
!(
!(
!(

!(
!(!(

!(
!(
!(
!(

!(
!(!(!(
!(!(!(!(
!(!(
!(

!(!(
!(
!(!(
!(!(

!(
!(!(
!(
!(!(
!(
!(

!(
!(!(
!(
!(

!(
!(
!(!(
!(

!(
!(!(
!(
!(
!(

!(
!(
!(
!(
!(

!(
!(!(
!(
!(
!(
!(
!
(
!(
!(

!(
!(
!(
!(
!(
!(

!(!(
!(
!(

!(
!
(
!
(
!(
!(
!(
!(

!(!(!(!(
!(!(
!(!(
!(
!(!
(
!(!(
!(!(
!(
!(
!(
!(
!(

!(!(
!(
!(!(
!
(
!(
!(

!(
!(
!(
!(

!(!(!
(
!(!
(
!(
!(
!(
!(
!(!(
!(

!(
!(

!(!(
!(!(!(
!(

!(!(
!(
!(
!(
!(
!(

!(!(
!(

!(
!(
!(

!(
!(
!(
!(
!(
!(
!(

!(!(
!(
!(

!(!(
!(
!(
!(!(

!(!(

!(
!(
!(!
(

!(

!(!(
!(
!(
!(
!(
!(
!(!(
!(
!(

!(!
(
!(

!(!
(
!(!(

!(
!(
!(
!(
!(

!(
!(!(
!(
!(!(

!(
!(
!(!(
!(!(

!(!(
!(
!(
!(!
(
!(
!(!(!(
!(
!(
!(
!(
!(
!(
!(
!(

!(
!(

!(
!(
!
(
!(

!(
!(
!(!(
!(
!(
!(!(

!(
!
(

!(
!(
!(!(
!(
!(
!(!(
!(
!(
!(!(
!(!(!(
!(
!(

!(!(
!(!(
!(

!(

!(
!(
!(
!(

!(
!(
!(
!(
!(!(

!(
!(!(
!(
!(
!(
!(
!(

!(
!(

!(
!(
!(!(
!(!(

!(
!(!(
!(
!(!(!(

!(
!(!(
!(

!(
!(!
(

!(
!
(
!(!(
!(
!(
!(
!(
!(

!(
!(!(
!(

!(!
(
!(
!(!(

!(!
(
!(!(!(
!(!(!(
!(
!(!(!
(
!(
!(!(!(

!(
!(
!(
!(!(

!(

!(
!(
!(!(

!(
!(!(
!(!(

!(!(
!(!(
!(
!(
!(
!(
!(!(!(

!(
!(!(
!(!(
!(!(
!(!(
!(
!(
!(
!(!(
!(
!(
!(
!(!(
!(
!(!(

!(!(
!(
!(!(
!(
!(
!(!(
!(
!(!(
!(
!(!(
!(!(
!(!(
!(
!(!(
!(
!(
!(!
(
!(!(
!(!(!(!(
!(!(!(
!(

!(!(
!(!(

!(
!(
!(!
!(!(!(!(
!(!(!(
!(!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!
(
!

(
!(!(!(
!(!(!(!(!(
!(
!(!(!(
!(
!
(!(
!(!(!
(
!(!(
!(
!(
!(!(
!(
!(
!(
!(!(
!(!(!(!(
!(!(!(
!(!(!(
!(!(
!(
!(!(!(!(
!(
!(
!(
!(!(!(
!(!(!(!(!(!(
!(
!(!(
!(!(
!(!
(
!(
!(!(!(!(!(
!(!(
!(!(
!(
!(
!(!(!(
!(
!(!(!(!

(!
(!(!(
!(!(
!(!(!(!
(
!(

!(!(
!(!(
!(
!(
!(
!(
!(!(
!(!
(
!(!(!
(
!(!(
!(!(!(
!(
!(
!(!(!(
!(
!(
!(!(
!(!(!(
!(
!(
!(
!(!(!(
!(!(
!(
!(
!(!(!(
!(
!(!(
!(!(!(
!(!(
!(

!(
!(
!(
!(
!(
!(
!(

!(
!(
!(!(!(
!(!(
!(
!(
!(
!(
!(!(
!(!(!(
!(
!(
!(!(

!(
!(
!(!(!(
!(
!(
!(
!(
!(
!(
!(!(
!(
!(

!(
!(
!(
!(
!(
!(
!(
!(

!(
!(

!(
!(
!(

!(
!(

!(
!(
!(

!(
!(
!(
!(

!(
!(

!(

!(
!(

!(

!(
!(!(!(!(
!(
!
(
!(
!(
!(

!(
!(

!(

!(
!(

!(

!(
!(
!(

!(
!(!(
!(
!(
!(

!(
!(

!(
!(
!(

!(

!(!(
!(
!(
!(
!(

!(
!(!(
!(
!(

!(
!
(
!(!(
!(
!(

!(
!(

!(
!(

!(
!(
!(
!(

!(
!(
!(

!(
!(

!(

!(!
(
!(!(
!(!(
!(

!(

!(!(!(
!(

!(
!(
!(!(

!(!(
!(
!(
!(

!(!(
!(!(
!(
!(

!(
!(

!(
!(!(
!(

!(!(
!(
!(

!(

!(!(

!(

!(
!(
!(

!(
!(
!(

!(
!(!(
!(!(
!(!(

!(
!(
!(

!(

!(
!(!
(
!(
!(

!(
!(
!(
!(
!(
!(
!(!(
!(!(

!(
!(!(
!(
!(!(
!(
!(!(

!(
!(!(

!(
!(

!(
!(
!(

!(
!(
!(
!(
!(

!(
!(
!(
!(
!(

!(
!(
!(!(
!(
!(
!(
!(
!(

!(

!(!(
!(
!(!(
!(!(
!(
!(
!(
!(

!(
!(
!(
!(
!(

!(
!(

!(

!(
!(

!(

!(
!(

!(
!(

!(
!(

!(
!(
!(
!(
!(

!(
!(
!(

!(!(
!(
!(

!(
!(

!(!(

!(
!(
!(

!(
!(
!(

!(
!(
!(

!(
!(

!(

!(!(
!(
!(

!(
!(
!(
!(
!(
!(
!(
!(!(!(!(!(!(
!(
!(!(

!(
!(!(
!(!(
!(

!
(
!(!(
!(
!(!(
!(
!(!
(!(
!
(
!(!(
!(!(
!(!(
!(
!(
!(

!(!(
!(
!(
!(!(!(

!(
!(!(
!(
!(!(
!(!(

!(!(!(
!(
!(
!(!(
!
(

!(
!(
!(
!(
!(
!(
!(
!(!(!
(
!
(
!(
!(
!(
!(
!(
!(
!(

!(!(!(
!(
!(
!(!(
!
(
!(
!(
!(
!(
!(
!(
!(!(
!(
!(

!(!(
!(
!(!(
!(
!(
!(!
(
!(
!(!(
!(
!(
!(
!(

!(

!(
!(
!(
!(
!(
!(

!(!(
!(
!(!(

!(!(
!(
!(!
(
!(

!(
!(
!(
!(
!(
!(!(
!(!(

!(
!(!(
!(
!(
!
(
!(
!(
!(
!(

!(
!(
!(!(!(

!(
!(
!(!(
!(
!(!(!(!(
!(
!(
!(!(
!(
!(!(

!(!(!(
!(
!(
!(

!(
!(
!(
!(
!(

!(
!(
!(
!(
!(
!(

!(
!(!(
!(

!(

!(
!(
!(
!(
!(
!(
!(
!(
!(

!(
!(!(!(!(
!(
!(!(!(!(
!(
!(!(

!(
!(
!(
!(

!(!(
!(
!(
!(

!(
!(!(
!(!(

!(
!(

!(!(!(
!(
!(
!(
!(
!(

!(!(
!(
!(

!(
!(
!(

!(!(
!(
!(

!(!(
!(
!(
!(!(!(!(
!(
!(!(
!(
!(!(
!(
!(
!(
!
(
!(!(
!(
!(
!(!(
!(!(
!(
!(

!(
!(!
(
(!
(
!(!
(!(
!(

!(!(
!(!(!
(
!(

!(!(!(!(
!(
!(
!(!(

!(
!(

!(
!(!(!(
!(
!(!(
!(!(!(!(

!(!(!(!(
!(!(
!(
!(
!(
!(

!(!(
!(
!(
!(!(

!(
!(

!(

!(
!(
!(
!(
!(
!(!(
!(!(!(
!(
!(
!(
!(
!(!(

!(
!(
!(!(!(
!(
!(
!(!(
!(
!(
!(

!(!(!(
!(
!(
!(!(
!(!(
!(!(
!(
!(
!(!(
!(

!(
!(
!(
!(!(!(

!(!(!(
!(
!
(
!(

!(
!(
!(
!(
!(
!(
!(
!(!(
!(
!(
!(
!(!(

!(
!(!(!(
!(

!(
!(

!(!(!(
!(
!(!(
!(
!(!(
!(!(!(

!(
!(
!(!(
!(!(
!(!(!(!(
!(!(!(
!(!(
!(
!(

!(!(
!(!(!(!(!(!(
!(
!(!(!(!(!(
!(
!(
!(
!(
!(
!(
!(
!(
!(

!(
!(
!(!(
!(

!(
!(!(
!(
!(
!(

!(
!(
!
(!(
!(!(
!(!
(
!
(

!(
!(!(
!(
!(
!(

!(
!(
!(!(

!(!(
!(
!(!(
!(
!(
!(!(
!(
!(!(!(
!(!(!(!(
!(
!(
!(!(
!(!(!(!(
!(!(
!(
!(
!(

!(
!(
!(!(!(
!(

!(!(
!(!(
!(
!(!(

!(!(
!(!(!(!(
!(
!(!(
!(

!(!(
!(!(

!(
!
(!(!(

!
(
!(

!(!(
!(!(
!(!(!(
!(!(
!(
!(!(

!(
!(
!(
!(!(
!(
!(!(
!(
!(
!(!(!(!(!(
!(
!(!(
!(
!(
!(!(!(
!(
!(
!(
!(!(
!(!(
!(!(
!(
!(!(!(!(
!(
!(!(
!
(

!(
!
(
!(!
(!(
!(!(!(!
(!(
!(
!(!(

!(
!(!(

!(!(
!
(
!(
!(
!(
!(!
(
!(
!(
!(
!(
!(
!(
!(
!(
!(

!(

!(
!(!
(!(!(!(
!(
!(

!(!(
!(!(!(
!(!(
!(
!(
!(

!(
!(!(!(
!(
!(!(

!(!(
!(!(!(!(
!(!(!(!(!
(
!(
!(!(!(!(
!(!(
!(
!(
!(!(!(!
(!(
!(

!(!(
!(!(
!(!(
!(!(!(
!(!(!(

!(!(
!(
!(
!(
!
(!(
!(!(
!(

!(!(!(
!(
!(
!(!(
!(
!(!(
!(
!(
!(!(!(!(!(!(!(!(!(!(

!(
!(!(
!(
!(
!(!(

!(!(!(
!(
!(

!(!(
!(!(!(
!(
!(!(
!(!(

!(!(
!(
!
(
!(

!(
!(
!(!(
!(
!(!
(!(
!(

!(!(
!(

!(
!(!(
!(
!(!(
!(
!(!(
!(!(
!(
!(
!(!(

!(!(!(!(
!(
!(!(!(
!(!(
!(!(
!(
!(!(
!(!(
!(

!(
!(!(
!(
!(

!(!(
!(
!(
!(
!(
!(
!(
!(
!(

!(
!(
!(!(
!(!(
!(!(
!(!(

!(
!(
!(
!(
!(
!(

!(
!(
!(
!(!
(
!(
!(

!(!(
!(

!(
!(

!(
!(
!(
!(
!(
!(!(
!(!(
!(!(
!(
!(
!(!(

!(
!(
!(
!(
!(
!(
!(!(
!(!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!
(
!(
!(
!(
!(
!(
!(!(
!(
!(

!(
!(!(!(
!(
!(
!(
!(
!(!(!(!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(!(
!(
!(
!(
!(
!(
!(
!(
!(

!(
!(
!(
!(!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(!(
!(
!(
!(

!(
!(
!(!(
!(
!(
!(
!(
!(
!(
!(
!(!
(!(!(
!(
!(!(
!(
!(!
(
!(
!(
!(
!(
!(!(
!(!(
!(
!(
!(

!(
!(
!(
!(

!(!(
!(
!(
!(
!(

!(!
(!(
!(!(
!(
!(!(!(
!(
!(!(
!(
!(!(
!(
!(!(
!(!
(
!(
!(
!(

!(
!(!(
!(!(
!(!(!(

!(!
(
!

(!(
!(!(
!(!(

!(
!(

!(
!(!(
!(

!(
!(
!(

!(

!(

!(
!(!(
!(
!(
!(

!(
!(
!(
!(

!(!(
!(!(
!(
!(!(

!(
!(
!(

!(
!(!(!(!(

!(
!(!(
!(
!(!(

!(
!(!(
!(

!(
!(

!(
!(
!(
!(
!(

!(
!(
!(
!(

!(
!(
!(
!(
!(!(
!(!(!(

!(

!(
!(
!(

!(!
(!(!(
!(
!(

!(!
(!
(

!(
!(

!(
!(
!(
!(

!(
!(

!(
!(
!(!(

!(
!(

!(!(
!(

!(
!(!(!(
!(
!(
!(!(
!(
!(

!(
!(
!(
!(!(

!(
!(

!(!(!(
!(!(
!(

!(
!(
!(!(!(
!(

!(!(
!(

!(!(!(!(!(!(
!(!(!(!(

!(
!(
!(!(

!(

!(
!(!(
!(!(
!(
!(

!(
!(
!(

!(
!(!(!(

!(

!(
!(

!(!
(
!
(
!(
!(
!(
!(
!(!(!(
!(
!(!(
!(!(
!(
!(
!(
!(
!(
!(
!(!(
!(
!(!(
!(
!(!(
!(
!(!(
!(!(
!(!(!(!(
!(
!(
!(!(!(

!(!(!
(!(
!(
!(!(!(
!(!(
!(!(
!(
!(
!(!(!(
!(
!(
!(
!(!(!
(
!(
!(

!(!(
!(!(
!(!(!
(!(!(
!(
!(
!(!(
!(
!(

!(!(!(
!(!(!(
!(
!(!(
!(
!(!(
!(
!(
!(!(
!(!
(!(!(
!(!
(
!(
!(
!(!(
!(
!(
!(
!(!(
!(!(
!(!(!(!(!(!(
!(
!(!(!(
!(!(
!(!(
!(
!(!(!(!(!(
!(!
(!(!(
!(
!(!(
!(!(!(
(!(!(
!(!(
!(!(!(!(!(!(

!(!(!(!(
!(
!(!(

!(!
(
!(!(!(
!(
!(
!(
!(
!
(!(
!(
!(
!(!(!(!(!(!(!
(
!(!(
!(!(!(
!(!(!(!(!(
!(!(
!(!(
!(!
(!(
!(!(
!(!(
!(!(!(!(!(!(
!(!(
!(
!(!(
!(!(!(
!(
!(!(
!(
!(!(
!(
!(
!(!
(
!(
!(
!(
!(
!(
!(!(!(
!(!(
!(!(
!(

!(!(!(!(
!(
!(
!(
!(

!(!(
!(
!(!(
!(
!(!(
!(

!(!(
!(
!(
!(
!(
!(
!(!(
!(
!(
!(
!(
!(
!(!(!(

!(!(
!(!(
!(
!(
!(
!(
!(
!(!(
!(!(

!(
!(!(!
(
!(!(
!(
!(
!(!(
!(
!(
!(
!(
!

(!(
!(
!(
!(!(!(
!(!(
!(!(!(

!(
!(!(
!(
!(
!(
!(
!(
!(
!(!
(!(
!(
!(
!(!(
!(!(
!(
!(
!(
!(!(

!(
!(!(
!(!(!(
!(
!(
!(
!(!(!(
!(
!(!(!(!(!(!(
!(
!(
!(

!(
!(!(
!(
!(
!(
!(
!(!(
!(
!(!
(
!(
!(
!(!(
!(
!(
!(!(
!(!(!(
!(
!(
!(
!(
!(!(!(!(
!(
!(

!(!(
!(
!(
!(
!(!(!(!(!(!(!(
!(
!(
!(
!(!(
!(!(!(
!(!(

!(
!(!(!(!(
!(!(!(

!(
!(
!(

!(
!(!(
!(
!(
!(!(
!(
!(
!(!(
!(
!(!(
!(!(!(
!(
!(
!(
!(
!(
!(

!(
!(
!(
!(!(
!(!(
!(
!(
!(!(
!(
!(
!(
!(
!(!(!(
!(
!(
!(
!(
!(
!(

!(
!(
!(!(
!(
!(!(!(
!(!(!(
!(
!(!(!(
!(!(!(!(!
(!(
!(!(
!(
!(
!(
!(
!
(
!(!(
!(
!(
!(!(
!(

!(!
(
!(
!(
!(!(
!(!(
!(

!(
!(
!(
!(
!(
!(
!(
!(
!(!(!
(

!(
!(
!(!(
!(

!(
!(
!(
!(
!(

!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(!(!(!(
!(
!(
!(
!(

!(!(
!(
!(
!(
!(!(!(

!(
!(!(
!(
!(
!(

!(!(
!(
!(
!(
!(!(
!(!(
!(
!(!(
!(
!(

!(
!(

!(
!(

!(
!(!(!(!(!(!(!(!(
!(!(!(!(!(
!(!(!(!(!(!(!(
!(!(!(
!(!(

!(
!(

!(
!(
!(
!(

!(!(
!(
!(

!(
!(

!(

!(
!(
!(
!(
!(

!(
!(!
(
!(!(!(!(!(!(

!(

!(
!(!(

!(
!(

!(
!(!(
!(!(
!(
!(
!(

!(
!(
!(

!(

!(
!(
!(!
(
!(

!(
!(

!(
!(
!(
!(
!(
!(

!(
!(
!(
!(
!(!(

!(
!(
!(
!(

!(
!(

!(
!(
!(
!(
!(
!(
!(

!(!(
!(
!(
!(!
(
!(!(
!(
!(

!(
!(
!(
!(!(
!(

!(
!(!(
!(
!(
!(
!(

!(

!(
!(
!(

!(
!(
!(
!(
!(
!(

!(
!(

!(
!(
!(

!(
!(!(
!(
!(!
(

!(
!(!(
!(
!(
!(

!(
!(
!(
!(
!(
!(

!(
!(

!(
!(
!
(
!(
!(!(
!(!(!(

!(
!(
!(
!(
!(

!(

!(!(
!(
!(
!(!(
!(!(

!(

!(
!(
!(
!(!
(
!(
!(

!(

!(
!(
!(
!(

!(
!(
!(
!(
!(

!(
!(
!(
!(
!(
!(
!(
!(

!(
!(
!(
!(

!(
!(
!(!(
!(
!
(

!(

!(
!(
!(!(!(

!(
!(
!(
!(!(

!(
!(

!(!(
!(

!(
!(!(!(!(!(
!(
!(
!(

!(
!(
!(
!(

!(!(
!(
!(!(!(
!(!(
!(!
(
!(
!(
!(
!(
!(!(
!(
!(
!(
!(
!(
!(

!(!(
!(
!(
!
(
!(!(!(
!(!(
!(
!(
!(

!(
!(
!(
!(
!(!(

!(
!(
!(
!(
!(
!(
!(!(
!(
!(
!(!(

!(
!(!(
!(
!(
!(
!(
!(

!(
!(
!(

!(!(!(

!(
!(

!(
!(!(

!(
!(!(
!(!(
!(

!(!(
!(
!(!(

!(
!(
!(
!
(
!(
!(

!(
!(
!(!(!(!(
!(
!(!(

!(
!(
!
(!(
!(
!(
!(!(
!(
!(!(!(!(
!(!(
!(
!(!(!(
!(
!(
!(
!(
!(
!

(!
(
!(
!(!(
!(
!(
!(
!(!(
!(!(
!(
!(!(

!(!

(!
(!(
!(
!(

!(!(!
(!(!(!(!(
!(!(!
(!(!(
!(!(
!(

!(
!(
!(
!(

!(!
(
!(
!(

!(
!(
!(
!(!
(

!(
!(
!(

!(
!(
!(
!
(
!(!(
!(!(
!(

!(
!(!(
!(
!(

!(
!(
!(
!(
!(
!(

!(
!(
!(
!(!(
!(
!(

!(
!(
!(
!(
!(!(

!(!(
!(
!(!(
!(
!(
!(
!(!(
!(!(
!(
!(

!(
!(
!(
!(
!(
!(

!(
!(
!(
!(

!(!(!(
!(
!(

!(

!(
!(
!(
!(
!(
!(!(

!(
!(
!(
!(
!(

!(
!(!(
!(
!(
!(
!(
!(!(
!(!(
!(

!(
!(!(
!(!(
!(
!(
!(
!(!(
!(!
(

!(
!(
!(
!(

!(!(!
(
!(!(
!(!(!(
!(
!(
!(!(
!(!(
!(!(!(
!(

!(
!(
!(
!
(!(
!(!(!(

!(

!(!(!(!
(!(
!(
!(
!(!(

!(
!(!
(
!(!(
!(
!(
!(!(!(
!(
!(
!(
!(

!(
!(
!(
!(
!(!(!(
!(

!(!(

!(

!(
!(!(
!(
!(
!(
!(!(!(!(
!(
!(!(

!(
!(
!(
!(
!(
!(
!(!(
!(
!(
!(!(
!(
!(
!(
!(
!(
!(!(
!(!(!(

!(
!(
!(
!(
!(
!(!(
!(
!(
!(
!(
!(!(
!(
!(
!(
!(
!(
!(
!(
!(
!(

!(!(
!(
!(
!(
!(
!(
!(!(

!(
!(
!(!(
!(
!(
!(
!(
!(!(

!(
!(
!(
!(

!(
!(
!(
!(!(
!(

!(
!(

!(
!(
!(
!(
!(!(
!(!(
!(
!(

!(
!(
!(
!(!(

!(!(
!(!(
!
(!(
!(!(
!(
!(
!(
!(!(
!(!(!
(
!(
!(!(
!(
!(!(
!(

!(!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(
!(

!(!(!(
!(
!
(
!(
!(!
(

!(
!(!(
!(

!(!(
!(
!(
!(
!(
!(
!(
!(
!(
!(

!(
!(
!(
!(
!(!(
!(
!(
!(
!(!(

!(
!(
!(
!(
!(!(
!(!(!(
!(!
(
!(!(
!(
!(
!(
!(!(!
(
!(

!(
!(
!(
!(
!(

!(!(
!(!(
!(
!(!(!
(
!(
!(
!(!(

!(
!(!(!(
!(
!(
!(
!(
!(!
(!(
!(
!(
!(
!(
!(
!(
!(!(
!(
!(
!(

!(!(
!(
!(!(!(!(!(

!(
!(

!(!(
!(

!(

!(
!(

!(
!(
!(
!(!(
!(

!(
!(!(
!(

!(
!(!(

!(
!(
!(

!(
!(

!(!(

!(
!(
!(

!(

!(
!(
!(

!(

!(
!(

!(
!(!(
!(!(
!(
!(

!(!(
!(
!(
!(

!(
!(!(
!(
!(
!(

!(!(
!(!(

!(
!(
!(

!(

!(
!(
!(
!(

!(

!(
!(!(
!(!(

!(!(!(
!(

!(
!(

!(
!(

!(
!(

!(
!(
!(!(
!(

!(!(
!(
!(

!(
!(
!(!(
!(
!(
!
(
!(

!(!(
!(
!(!(

!(
!(
!(

!(!(!(!(
!(
!(!(
!(!(!(
!(
!(
!(

!(
!(
!(
!(
!(
!(
!(
!(
!(
!(

!(
!(

!(
!(
!(
!(
!(
!(

!(
!(

!(!(!(
!(!(
!(
!(
!(!(!(
!(!(

!(
!(
!(!(!(
!(!(!(
!(!(
!(
!(

!(
!(
!(
!(
!(!(

!(!(
!(!(!(!(
!(

!(!(
!(!
(
!(!(
!(!(
!(

!(!(
!
(
!(
!(
!(
!(
!(
!(!(!(!(
!(
!(!(
!(

!(!(!(

!(
!(!(
!(
!(!(
!(
!(
!(

!(!(
!(

!(
!(
!(
!(!(
!(
!(
!(
!(

!(!
(

!(!(
!(
!(!(
!(

!(
!(

!(
!(!(!(
!(
!(

!(

!(
!(
!(
!(!(!
(
!(

!(
!(
!(

!(
!(
!(

!(
!(
!(!(
!(
!(

!(
!(!(

!(
!(

!(!(
!(
!(
!(
!(

!(
!(
!(

!(
!(
!(
!(
!(
!(!(

0
500
1,000
250
Miles
4
Legend
!(
CWS
SW
Intakes
Cotton
Acres
0
1
­
40000
40001
­
80000
80001
­
200000
200001
­
300000
Cotton
Acres
Relative
to
Community
Water
System
(
CWS)
Surface
Water
Intakes
Figure
1.
Cotton
acres
relative
to
Community
Water
Systems
intakes.

where
cotton
is
grown,
and
none
of
the
intakes
are
in
the
counties
with
the
most
cotton
acreage.
The
mean
number
of
samples
taken
in
Texas
was
6.4
per
system,
taken
over
a
period
of
5
years.

The
statistical
analysis
done
on
these
data
was
designed
to
examine
long
term
systemwide
trends
rather
than
the
local,
temporal
kinds
of
effects
that
would
be
expected
to
occur
from
agricultural
use
of
pesticides.
These
results
are
not
intended
to
be
representative
of
the
impact
of
agricultural
arsenical
use.
Even
if
this
non­
targeted
dataset
of
treated
water
were
to
include
any
samples
with
arsenic
levels
elevated
by
agricultural
arsenical
use,
the
inclusion
in
the
analysis
of
so
much
data
from
non­
use
areas,
along
with
the
averaging
of
samples
widely
spaced
through
time,
would
be
expected
to
mask
these
impacts.
These
data
cannot,
therefore,
be
expected
to
demonstrate
a
correlation
between
aquatic
arsenic
levels
and
use
of
arsenicals.

The
Florida
data
presented
by
the
registrant,
although
presented
in
the
same
table
as
the
other
state
data,
is
from
a
different
source
and
any
analysis
of
this
data
was
conducted
by
the
registrant,
rather
than
by
the
Office
of
Water.
These
data
are
not
appropriate
for
comparison
with
modeled
estimates
for
the
same
reasons
stated
above
for
the
Office
of
Water
data.
The
issue
of
the
relationship
between
sampling
sites
and
usage
locations
is
very
important
for
this
Florida
data.
The
data
presented
are
only
from
surface
water
Page
11
of
23
systems,
which,
according
to
FDACS,
make
up
only
10%
of
Florida's
public
water
supply.
The
issue
of
groundwater
arsenic
exposure
will
be
discussed
in
more
detail
in
the
response
to
comments
from
FDACS.
However,
it
is
important
to
note
that
the
ground
water
arsenic
levels
in
the
Florida
dataset
are
considerably
higher
than
the
surface
water
levels.
The
2003
dataset,
for
which
the
registrant
reports
surface
water
results,
had
ground
water
detections
of
arsenic
up
to
48
ppb,
while
in
2005,
ground
water
levels
as
high
as
240
ppb
were
measured.
These
maximum
arsenic
levels
were
measured
at
intakes
near
an
apartment
complex
(
2003)
and
near
a
group
that
sells
sod
commercially
(
2005),
both
potential
users
of
arsenical
herbicides.

In
reference
to
the
cited
monitoring
data
from
Bednar
(
2002),
Swancar
(
1996),
and
DERM
(
2002),
the
registrant
states
that
"
the
[
modeling]
predictions
are
not
corroborated
by
monitoring
results."
Specifically,
the
registrant
argues
that:
1)
it
is
inappropriate
to
compare
speciated
measurements
of
inorganic
arsenic
or
unspeciated
measurements
of
total
arsenic
to
modeled
estimated
concentrations
of
organic
arsenicals,
2)
Chronic
modeled
values
should
be
compared
to
the
average
of
monitoring
data
rather
than
the
peak,
and
3)
the
monitoring
sites
do
not
mimic
the
modeled
index
scenario.
(
Attachment
5,
p.
10)

EFED
answer:
The
uncertainties
associated
with
monitoring
data,
both
in
general
and
specific
to
the
individual
studies,
are
discussed
in
the
drinking
water
assessment.
Given
the
potential
for
transformation
of
applied
organic
arsenicals
to
inorganic
arsenic,
it
is
appropriate
to
include
measurements
of
total
arsenic
and
inorganic
arsenic
in
a
consideration
of
monitoring
data.
It
is
true
that
other
sources
of
arsenic
may
be
present
in
the
monitored
concentrations;
this
uncertainty
is
characterized.

Limited
targeted
surface
water
monitoring
data
are
available.
Even
targeted
sampling
is
likely
to
underestimate
both
acute
exposures
and
the
frequency
of
occurrence
due
to
the
limited
number
of
samples
and
the
extended
time
between
sampling
events.
Modeling
of
aquatic
exposure
estimates
daily
concentrations
over
30
years.
Monitoring
occurs
much
less
frequently
on
a
shorter
time
scale
and
can't
be
expected
to
capture
the
same
exposure.
Even
with
weekly
or
biweekly
monitoring,
as
in
the
Bednar
cotton
monitoring
study,
this
represents
24
samples
over
a
single
season.
It
is
not
appropriate
to
take
a
mean
of
these
values
to
compare
to
the
modeled
chronic
result,
which
is
an
annual
average
of
daily
estimated
concentrations
expected
to
occur
only
once
every
ten
years.
Page
12
of
23
V
 
A.
The
risk
assessment
should
consider
a
realistic
percent
cropped
area
for
turf.

EPA
assumes
in
the
risk
assessment
that
100%
of
a
watershed
consists
of
turf.
This
assumption
does
not
reflect
real­
world
circumstances
and
is
contrary
to
available
data.
The
U.
S.
watershed
with
the
highest
percent
of
crop
is
covered
with
87%
cropped
area.
OPP
cannot
reasonably
ignore
these
data
and
assume
instead
that
100%
of
the
watershed
is
turf
­­
a
minor
"
crop."
Assuming
100%
percent
crop
area
results
in
an
assessment
that
is
not
scientifically
defensible.
Instead,
OPP
should
use
the
data
submitted
by
the
Task
Force
(
see
Attachments
5
and
7).

EFED
answer:
As
the
registrant
states,
aquatic
exposure
from
the
turf
use
is
modeled
in
this
assessment
with
the
assumption
that
100%
of
the
watershed
is
treated
turf,
in
accordance
with
EFED
policy.
As
acknowledged
in
the
document,
"
this
is
likely
an
overestimation
of
exposure
but
the
extent
of
overestimation
is
undefined."
The
limited
monitoring
data
available
to
characterize
the
overestimation
are
discussed
in
the
assessment.

For
agricultural
crops,
a
percent
cropped
area
factor
(
PCA)
is
used
to
adjust
modeled
EDWCs
to
account
for
the
portion
of
a
drinking
water
watershed
that
contains
fields
planted
with
a
specific
crop.
PCAs
have
been
developed
for
only
a
few
major
agricultural
crops
to
date,
largely
limited
by
availability
of
data
at
the
required
scale.
A
default
value
of
87%
is
used
for
all
other
agricultural
crops.
Based
on
land
cover
data
for
agricultural
crops,
this
default
value
is
not
applicable
to
the
turf
use.
At
the
present
time,
data
are
not
available
at
the
spatial
scale
needed
to
develop
the
non­
agricultural
equivalent
of
a
national
PCA
for
turf
uses.
The
methods
used
to
develop
the
agricultural
PCAs
were
reviewed
by
a
SAP
which
found
that
these
methods
were
unsatisfactory
for
use
with
minor
crops.
For
major
crops,
PCAs
are
based
on
8­
digit
HUC
(
average
size
over
1000
square
miles)
and
county­
based
crop
acreages.
The
SAP
expressed
concern
that
this
scale
of
data
would
"
result
in
biases
with
minor
crops,"
stating
that
"
most
drinking
water
supplies
are
fed
by
smaller
watersheds
that
may
have
different
PCAs 
The
PCA
factors
one
gets
(
especially
for
minor
crops)
for
8
digit
watersheds
versus
11
or
14
digit
watersheds
could
be
very
different."
Difficulties
associated
with
defining
a
PCA
for
minor
agricultural
crops
are
magnified
for
turf,
which
incorporates
multiple
poorly
delineated
uses,
including
golf
courses,
sod
farms,
residential
lawns,
parks,
athletic
fields,
and
highway
rights­
of­
way,
among
others.
Any
submissions
of
data
analyses
for
development
of
a
nationally
representative
turf
PCA
that
account
for
all
turf
uses
and
are
based
on
data
at
the
appropriate
spatial
scale
will
be
considered
by
EFED.

Determination
of
PCAs
requires
data
on
a
national
scale
in
order
to
establish
the
maximum
area
within
any
watershed
that
might
be
expected
to
be
treated
with
pesticide.
These
estimated
exposures
are
not
intended
to
represent
typical
values
likley
to
have
widespread
occurrence
nationally.
Rather,
the
intent
of
assessment
at
this
level
is
to
provide
high
end
estimates
of
the
exposure
that
might
result
from
maximum
use
in
the
most
vulnerable
settings.
Further
refinement
would
be
required
to
estimate
the
exposure
in
areas
where
use
might
be
less
extensive
or
areas
that
have
less
vulnerable
environmental
conditions.
Page
13
of
23
The
registrant
submitted
two
recommendations
of
PCAs
to
be
used
in
adjusting
EDWCs
resulting
from
the
turf
use.
While
both
recommendations
are
based
on
constructive
approaches,
neither
includes
sufficient
consideration
of
all
relevant
turf
areas
at
the
necessary
spatial
scale.
The
approach
presented
in
Attachment
5
primarily
relies
on
8
digit
HUC
data
to
estimate
maximum
golf
course
and
sod
farm
area
and
the
only
support
provided
for
the
residential
value
is
an
analysis
of
data
from
a
single
county.
No
discussion
is
included
of
other
turf
uses
and
no
recommendation
is
provided
for
how
these
values
should
be
combined.
The
recommended
PCA
presented
in
Attachments
6
and
7
defines
a
maximum
cropped
area
for
all
turf
uses
based
on
an
assessment
of
land
cover
within
watersheds
in
south
central
Florida.
Although
the
suggested
approach
includes
an
assessment
of
all
turf
categories
at
the
scale
of
a
16
digit
HUC,
there
is
insufficient
justification
that
the
watersheds
selected
for
evaluation
at
this
scale
are
representative
of
a
national
maximum.
The
watersheds
are
identified
based
on
data
for
golf
course
and
sod
farm
acreage
only
without
consideration
of
residential
and
other
turf
uses.

Attachment
5
also
advises
use
of
a
golf
course
adjustment
factor
(
GCAF)
to
modify
modeled
EDWCs.
The
GCAF
is
only
applicable
to
golf
courses
and
would
not
affect
EDWCs
for
the
general
turf
use.
Even
for
the
golf
course
use,
application
of
the
GCAF
would
not
significantly
change
the
EDWCs,
as
the
only
labeled
restriction
is
that
arsenicals
cannot
be
applied
to
golf
course
greens,
which
are
accounted
for
by
the
GCAF
as
only
3%
of
golf
course
area.

V
 
B.
The
risk
assessment
should
consider
that
MSMA
is
used
on
turf
mainly
in
a
spot
treatment
application.

Most
of
MSMA's
turf
uses
are
for
spot
treatment.
OPP
assumes
that
100%
of
the
turf
is
treated
with
MSMA,
arguing
that
spot
treatment
is
not
enforceable.
In
fact,
the
label
was
recently
revised
following
information
from
users
that
this
is
the
way
the
product
is
used.
Assuming
four
annual
broadcast
applications
is
wrong
and
produces
a
misleading
and
inaccurate
risk
assessment
that
is
not
scientifically
or
legally
defensible.

EFED
answer:
EFED
recognizes
that
spot
treatment
is
a
common
practice
in
applying
MSMA
to
turf.
Spot
treatment
is
enforceable
and
can
be
accounted
for
in
estimating
exposure
if
it
is
included
on
the
label
in
a
quantifiable
way.
The
discussion
of
surface
water
modeling
in
the
Drinking
Water
Assessment
includes
the
following
discussion
of
the
practice
of
spot
treatment:

Turf
EDWCs
are
also
expected
to
be
lower
if
arsenicals
are
applied
as
spot
treatments
rather
than
as
broadcast
sprays,
which
appears
to
be
a
common
practice.
There
is
a
possibility
that
this
will
be
included
as
a
restriction
on
the
labels
for
DSMA
and
MSMA.
Spot
treatments
are
applied
to
less
area
than
broadcast
sprays
and
so
lead
to
lower
exposure.
If
application
rate
or
area
is
limited
on
the
label
by
a
quantifiable
amount,
the
estimated
concentrations
will
be
reduced
by
a
proportional
Page
14
of
23
amount.
For
example,
if
treatment
of
arsenicals
on
all
turf
uses
is
limited
to
50%
of
a
field,
the
EDWCs
will
go
down
to
one
half
of
the
original
value.

It
is
important
to
keep
in
mind,
however,
that
while
this
is
true
for
surface
water,
spot
treatment
may
have
less
of
an
effect
on
ground
water
concentrations
as
described
in
more
detail
in
the
response
to
Florida's
groundwater
concerns.

VI
­
A.
The
soil
analysis
ignores
quality
data
in
favor
of
flawed
literature
studies:
EPA
Must
Correct
Its
Erroneous
Assumption
That
There
Is
100%
Transformation
of
Organic
Arsenic
Herbicides
to
Inorganic
Arsenic,
to
Reflect
the
Data
That
Show
Only
a
Small
Portion
Has
the
Potential
to
Transform
OPP
cites
two
flawed
and
unreliable
published
papers
that
are
in
contrast
to
all
acceptable
guideline
studies
and
other
peer­
reviewed
literature,
and
on
the
basis
of
those
two
papers,
assumes
that
there
is
100%
transformation
of
organic
arsenicals
in
soil 
All
of
the
substantial
body
of
scientifically
valid
data
demonstrates
that
no
more
than
20%
of
applied
MMA
would
convert
to
inorganic
arsenic
in
natural
soils.
That
figure
represents
the
worst­
case
result
among
the
available,
soundly­
conducted
studies.
Application
of
the
20%
figure
reduces
by
5­
fold
the
estimated
inorganic
arsenic
level
in
soil.
An
assessment
that
assumes
100%
transformation
in
the
face
of
these
data
is
scientifically
indefensible
and
legally
invalid.
(
Described
further
in
Attachment
8.)

EFED
answer:
Before
directly
addressing
specific
concerns
of
the
registrants
regarding
the
quality
of
data
used
in
this
assessment,
it
is
constructive
to
describe
in
general
how
data
were
considered,
from
both
registrant
submitted
studies
and
the
open
literature.
It
is
true
that
fate
studies
from
the
open
literature
are
used
in
this
environmental
fate
discussion
to
a
greater
extent
than
is
typical
for
this
type
of
assessment.
There
is
a
large
body
of
research
spanning
several
decades
which
highlights
the
complexity
of
the
fate
of
arsenical
species
and
the
processes
of
dissipation,
metabolism,
and
redistribution
that
may
occur.
It
is
essential
to
account
for
the
variability
in
these
processes
and
their
observed
sensitivity
to
environmental
conditions.
As
will
be
discussed
in
more
detail
below,
the
set
of
registrant
submitted
GLP
studies
that
meet
EFED
data
quality
requirements
are
insufficient
for
this
purpose.
The
options,
then,
for
conducting
a
fate
assessment
that
is
satisfactorily
protective
of
human
health
are
to
either
make
conservative
assumptions
about
the
fate
of
applied
pesticides
or
to
take
into
account
the
available
open
literature
reports,
acknowledging
the
uncertainty
in
those
data.

Given
the
breadth
of
the
literature
on
this
topic,
it
was
not
practical
to
conduct
a
complete
literature
review.
Instead,
the
previous
RED
chapters
(
DMA:
D210451,
D212449,
D255226,
2000;
MMA:
D277223,
2001)
were
drawn
on,
with
the
group
of
studies
considered
in
those
documents
used
as
the
foundation
for
the
current
assessment.
In
addition
to
those
studies,
literature
introduced
by
the
registrants
or
other
stakeholders
has
been
considered
along
with
any
more
recent
research
that
has
come
to
EFED's
attention.
Older
studies
were
not
revisited
unless
issues
related
to
them
were
brought
up
specifically.
EFED's
general
approach
on
the
use
of
open
literature
in
assessing
arsenical
Page
15
of
23
fate
was
laid
out
quite
clearly
in
the
previous
REDs,
which
stated
that
"
the
registrant,
through
FIFRA
GLP
studies,
presents
an
unresolved
environmental
fate
profile
for
cacodylic
acid"
and
which
found
that
laboratory
studies
from
the
open
literature
were
"
contrasting
with
the
new
GLP
lab
studies,
and
[
were]
more
consistent
with
the
global
picture"
of
arsenical
fate.
Similar
conclusions
were
drawn
about
available
data
for
MMA
fate.
In
the
five
years
that
passed
between
the
time
that
statement
was
made
and
the
release
of
the
current
RED,
the
registrants
submitted
a
review
of
the
environmental
fate
of
MMA
(
MRID
46565301)
but
have
not
otherwise
taken
the
opportunity
to
conduct
additional
studies
or
to
present
alternative
data
demonstrating
why
the
earlier
RED's
conclusions
were
inappropriate.

In
this
assessment,
available
open
literature
studies
are
primarily
employed
in
qualitative
characterization
of
arsenical
fate,
with
the
exception
of
several
studies
used
to
estimate
modeling
input
parameters
for
aerobic
soil
metabolism
in
the
absence
of
acceptable
registrant
data.
It
is
not
true
that
EFED
assumes
100%
transformation
of
organic
arsenical
herbicides
to
inorganic
arsenic.
As
described
in
more
detail
in
the
response
to
section
IV­
A,
EFED
assumed
that
modeled
estimates
of
exposure
to
applied
parent
compounds
represented
a
reasonable
upper
bound
on
the
total
amount
of
arsenic
that
would
reach
surface
water
in
any
form.
It
was
impossible
to
define
with
certainty
what
portion
of
this
total
arsenic
in
surface
water
will
be
present
as
any
individual
arsenical
species,
and
so
the
risk
assessment
states
that
the
total
arsenic
EDWC
also
represents
the
maximum
potential
concentration
of
inorganic
arsenic.
Comparing
these
reported
values
to
the
results
that
would
have
been
obtained
had
100%
transformation
been
assumed
shows
that
the
peak
value
and
cancer
value
represent
88%
and
72%
conversion,
respectively.
The
range
of
transformation
seen
in
the
open
literature
supports
these
values
as
reasonable
high
end
estimates.

It
is
well
established
that
common
bacteria
from
the
soil
environment
are
capable
of
metabolizing
organic
arsenicals
to
inorganic
forms
of
arsenic.
There
is,
however,
uncertainty
associated
with
the
extent
to
which
this
transformation
may
occur
in
the
environment.
The
registrant's
statement
that
a
"
substantial
body
of
scientifically
valid
data
demonstrates
that
no
more
than
20%
of
applied
MMA
would
convert
to
inorganic
arsenic
in
natural
soils"
is
not
supportable.
Their
comment
identifies
5
registrant
studies,
including
aerobic
soil
metabolism
studies
for
DMA
and
MMA,
aerobic
aquatic
studies
for
DMA
and
MMA,
and
an
anaerobic
aquatic
study
for
MMA
(
MRIDs
42616001,
44767601,
43036101,
43314801,
and
44767602)
as
well
as
3
open
literature
studies
in
support
of
their
statement.
Although
aquatic
metabolism
studies
can
provide
important
information
about
the
fate
of
applied
arsenicals,
transformation
to
inorganic
arsenic
is
expected
to
occur
primarily
through
soil
metabolism.
The
MMA
aerobic
soil
metabolism
study
(
MRID
44767601),
which
measures
19%
transformation
to
inorganic
arsenic
after
one
year,
is
cited
as
a
principal
source
in
EFED's
fate
assessment.
Identified
by
EFED
as
the
only
available
GLP
metabolism
study
conducted
for
a
full
year,
data
from
this
study
are
used
in
determining
model
input
half­
lives
for
both
MMA
and
metabolite
DMA
and
also
in
characterizing
the
potential
extent
of
transformation
to
inorganic
arsenic.
Despite
this
study's
value,
given
the
variability
expected
in
the
Page
16
of
23
environmental
fate
of
arsenicals,
a
single
study
in
a
single
soil
cannot
be
identified
as
"
definitive
of
the
worst
case
behavior,"
as
the
study
is
described
by
the
registrant.

The
other
4
registrant
studies
cited
in
this
comment
show
no
transformation
to
inorganic
arsenic,
as
demonstrated
by
the
fact
that
no
14CO2
is
evolved.
While
these
studies
demonstrate
that
there
are
conditions
in
which
metabolism
does
not
occur,
no
explanation
for
this
result
is
provided
in
light
of
the
well
established
evidence
of
the
potential
for
metabolism,
and
there
is
no
analysis
to
define
what
environmental
factors
contribute
to
these
conditions.
Since
metabolism
of
organic
arsenicals
is
a
bacterial
process,
one
possible
factor
could
be
non­
viable
soils
and
sediments.
The
registrant
points
to
measured
MMA
(
from
applied
DMA)
or
measured
DMA
(
from
applied
MMA)
as
evidence
of
these
soils'
viability.
In
most
cases,
though,
the
pattern
of
these
detections
led
both
the
study
author
and
the
EFED
reviewer
to
conclude
that
they
are
artifacts
of
the
analysis
rather
than
products
of
transformation.
For
applied
DMA,
this
conclusion
is
supported
by
the
lack
of
CO2
evolution,
which,
as
the
registrant
points
out
elsewhere
in
this
comment,
would
be
expected
if
demethylation
of
DMA
to
MMA
had
occurred.
Only
one
of
these
studies
is
conducted
in
a
soil
environment
and
in
all
of
them,
the
soils
and
sediments
tested
may
be
non­
viable.
These
studies,
then
cannot
be
assumed
to
include
an
adequately
representative
set
of
environmental
conditions
and
can
not
be
interpreted
as
demonstrating
that
transformation
to
inorganic
arsenic
is
not
expected
in
most
conditions.

The
registrant's
comment
also
includes
a
review
of
open
literature
laboratory
studies,
recommending
three
studies
to
be
used
quantitatively
by
EFED
to
"
help
establish
the
expected
extent
of
organic
arsenical
transformation
to
inorganic
arsenic."
All
three
of
these
studies
were
included
in
the
current
assessment's
characterization
of
the
potential
for
transformtaion
to
inorganic
arsenic
(
Von
Endt,
1968;
Dickens
and
Hiltbold,
1967;
Abdelghani,
1977)
as
were
two
additional
studies
which
the
registrant
finds
deficient
and
recommends
against
using
(
Woolson
et
al,
1982;
Gao
and
Burau,
1997).
The
uncertainty
associated
with
all
of
these
studies
is
acknowledged
in
the
assessment,
which
notes
that
they
are
all
non­
GLP,
they
were
all
conducted
for
less
than
one
year,
and
many
of
them
do
not
include
time
series
data.
As
the
registrant
states,
the
data
presented
in
open
literature
studies
are
not
as
complete
as
that
required
in
registrant
studies,
so
these
studies
can
not
be
evaluated
to
the
same
standard
as
are
registrant
submissions.
Despite
the
acknowledged
uncertainties
associated
with
these
studies,
they
are
sufficiently
robust
to
raise
doubts
that
the
single
acceptable
registrant
aerobic
soil
study
is
adequate
to
represent
the
behavior
of
arsenicals
in
all
environmental
conditions.
Given
the
variability
of
arsenical
environmental
fate
processes
and
the
limited
available
set
of
registrant
data,
the
results
of
open
literature
studies
can
provide
a
more
thorough
understanding
of
the
fate
of
applied
organic
arsenic
herbicides.

To
summarize
the
open
literature
study
results,
Von
Endt
observed
a
maximum
of
10%
transformation
from
MMA
to
inorganic
arsenic,
although
values
were
still
rising
at
the
end
of
21
days.
Dickens
observed
up
to
5.5%
transformation
in
unamended
soils
and
16%
transformation
in
amended
soils
after
30
days
incubation.
Abdelghani
measured
up
to
29%
transformation
after
49
days.
(
The
registrant
includes
the
Abdelghani
study
in
the
group
of
<
20%
transformation
because
"
six
of
seven
trials"
had
less
than
20%
Page
17
of
23
transformation.
This
is
a
misleading
statement;
in
fact,
four
of
the
trials
were
conducted
with
pure
cultures
of
individual
microorganisms
and
so
are
not
comparable
to
the
other
results.)
All
of
these
studies
estimated
transformation
to
inorganic
arsenic
based
on
14CO2
evolution,
the
method
defined
by
the
registrant
as
"
superior"
to
extraction
methods.
The
registrant's
"
substantial
body
of
scientifically
valid
data
demonstrat[
ing]
that
no
more
than
20%"
conversion
may
occur,
then,
consists
of
a
single
GLP
aerobic
soil
study
showing
19%
conversion,
an
additional
aerobic
soil
study
and
a
group
of
aquatic
metabolism
studies
that
show
no
transformation
in
possibly
non­
viable
conditions,
and
2
open
literature
studies
conducted
for
30
days
or
less
showing
up
to
16%
conversion,
with
a
third
open
literature
study
showing
up
to
29%
conversion.

Given
this
limited
data
set,
it
is
statistically
invalid
to
describe
the
two
additional
studies
which
show
greater
than
50%
transformation
as
"
outliers."
In
these
studies,
Woolson
found
up
to
52%
transformation
to
inorganic
arsenic
after
60
days
based
on
14CO2
evolution
while
Gao
directly
measured
extracted
inorganic
arsenic
of
up
to
80%
of
applied
organic
arsenic
after
70
days.
The
registrant
subjects
these
two
"
outliers"
to
a
much
higher
level
critical
analysis
than
performed
for
their
"
recommended"
studies.
The
comments
regarding
general
quality
control
and
data
presentation
issues
in
these
two
studies
will
not
be
addressed
specifically.
These
issues
do
lead
to
uncertainty,
as
acknowledged
in
the
assessment,
but
are
generally
applicable
to
all
open
literature
studies
"
recommended"
by
the
registrant
as
well
as
to
these
"
outliers."

The
more
substantial
comments
regarding
these
studies
will
be
addressed
directly.
Regarding
the
Woolson
(
1982)
study,
the
registrant
points
out
that
a
similar
study
by
the
same
author
was
submitted
to
the
EPA
in
1991
as
a
GLP
study
but
was
rejected,
and
so
the
registrant
questions
the
data
quality
standards
applied
in
use
of
open
literature
in
this
assessment.
As
discussed
above,
older
studies
not
included
in
the
2000
RED
or
introduced
by
the
registrant
were
not
re­
evaluated.
Woolson
(
1982)
is
included
in
the
2000
RED
and
discussed
by
the
registrant
in
their
review
of
the
environmental
fate
of
MMA
but
the
more
recent
submitted
study
by
Woolson
was
not
included
in
either
of
those
documents
so
it
was
not
revisited
for
the
current
assessment.
The
1991
Woolson
study
mentioned
here
by
the
registrant
was
rejected
because
sample
calculations
were
not
provided
to
describe
how
values
were
corrected
for
recovery
and
because
background
arsenic
was
not
accounted
for.
It's
true
that
these
same
issues
are
also
relevant
to
portions
of
Woolson
(
1982),
but
the
current
assessment's
characterization
of
potential
transformation
to
inorganic
arsenic
is
based
on
Woolson's
14CO2
evolution
data,
to
which
these
problems
do
not
apply.
The
registrant
argues
that
in
this
case,
14CO2
evolution
may
represent
demethylation
of
DMA
to
MMA
instead
of
complete
demethylation
to
inorganic
arsenic.
As
discussed
in
the
environmental
fate
assessment,
there
is
limited
evidence
that
transformation
of
DMA
to
MMA
can
be
a
significant
pathway,
and
additionally,
no
MMA
was
detected
in
these
soils.

The
registrant
also
submits
specific
comments
regarding
the
quality
of
data
cited
from
Gao
(
1997)
which
demonstrate
significant
transformation
to
inorganic
arsenic.
This
study
originally
came
to
EFED's
attention
because
it
was
submitted
by
the
registrant,
who
described
it
as
"
a
useful
construct
to
assess
the
behavior
of
DMA
[
which]
Page
18
of
23
systematically
examined
the
effects
of
arsenic
form,
arsenic
concentration,
soil
moisture,
and
nutrient
addition
on
the
transformations
of
arsenic
compounds
in
soil
[
and]
conclude
that
DMA
is
demethylated
and
arsenate
is
the
main
metabolite"
(
MRID
48803001).
Despite
their
previous
support
of
this
work,
the
registrant
now
finds
it
to
be
deficient
and
"
incapable
of
determining
the
extent
to
which
organic
arsenicals
transformed
to
inorganic
arsenic
in
soil."
The
primary
issue
raised
with
this
study
is
that
results
are
presented
as
"
percent
mineralization,"
calculated
based
on
the
amount
of
recovered
parent
compound
rather
than
on
directly
measured
inorganic
arsenic.
While
this
is
true
for
one
table
of
data,
all
of
the
published
figures
showing
mineralization
are
based
on
direct
measurements
of
arsenic,
as
confirmed
by
personal
communication
with
the
author.
Determining
the
percent
mineralization
based
on
these
direct
measurements
rather
than
calculated
based
on
the
remaining
parent
compound
still
shows
up
to
80%
conversion,
based
on
the
line
plot
in
Gao's
Figure
7.
With
a
reported
mean
extraction
recovery
of
91%
with
a
standard
deviation
of
9%,
it
is
possible
that
this
result
does
not
measure
all
of
the
arsenic
present
in
the
soil.

In
conclusion,
while
there
is
some
uncertainty
associated
with
the
use
of
open
literature
data
which
are
not
evaluated
to
the
same
standard
as
registrant
submissions,
the
open
literature
results
are
sufficiently
robust
to
raise
doubts
that
the
single
acceptable
registrant
aerobic
soil
study
is
adequate
to
represent
the
behavior
of
arsenicals
in
all
environmental
conditions.
The
Agency's
estimation
of
exposure
to
parent
arsenicals
is
modeled
based
on
the
measured
properties
of
those
species,
with
this
estimated
exposure
also
assumed
to
represent
a
reasonable
upper
bound
on
the
total
amount
of
arsenic
that
would
reach
surface
water
in
any
form.
It
is
impossible
to
define
with
certainty
what
portion
of
this
total
arsenic
in
surface
water
will
be
present
as
any
individual
arsenical
species
and
so
the
risk
assessment
states
that
the
total
arsenic
EDWC
also
represents
the
maximum
potential
concentration
of
inorganic
arsenic.
The
open
literature
studies
are
used
to
characterize
the
possible
extent
of
transformation.
The
range
of
transformation
observed
in
these
studies
supports
the
estimated
values
of
exposure
to
inorganic
arsenic
as
reasonable
high
end
estimates.

The
Task
Force
urges
OPP
to
account
for
the
clearly
demonstrated
biphasic
transformation
of
organic
arsenic
herbicides
to
account
for
the
clearly
demonstrated
biphasic
transformation
of
organic
arsenic
herbicides
to
inorganic
arsenic
in
its
modeling
[
and]
to
account
for
the
fractions
of
MMA
and
DMA
that
become
irreversibly
bound
to
soil.

EFED
answer:
See
section
IV­
A
for
discussion
of
the
biphasic
transformation
of
inorganic
arsenic.
The
modeled
estimates
of
exposure
to
MMA
and
DMA
use
half­
lives
estimated
from
open
literature
and
registrant
submitted
studies.
These
half­
lives
are
based
on
measurement
of
recovered
parent
compound
and
so
already
account
for
any
MMA
or
DMA
that
is
irreversibly
bound
to
soil
and
not
recovered.
Page
19
of
23
VI
­
B.
The
soil
analysis
ignores
quality
data
in
favor
of
flawed
literature
studies:
EPA
Must
Correct
Its
Erroneous
Assumption
That
Inorganic
Arsenic
in
Soil
Is
Completely
Bioavailable
to
Reflect
the
Data
That
Show
Only
a
Fraction
Is
Bioavailable
EFED
answer:
EFED
estimates
the
potential
buildup
of
total
arsenic
in
soil
following
long
term
repeated
use
of
organic
arsenicals.
No
direct
calculations
of
ecological
risk
are
calculated
based
on
these
estimates
of
soil
exposure
so
the
bioavailability
of
arsenic
in
soil
is
not
relevant.
Instead,
ecological
risk
from
soil
buildup
is
characterized
generally
by
comparison
to
OSWER
Eco­
SSLs,
values
which
are
developed
to
be
compared
directly
to
measured
soil
concentrations
as
a
screening
level
tool
for
determining
risk.

VII
­
A,
OPP
should
not
compare
total
arsenic
concentrations
in
water
to
the
VII
 
B.
arsenic
MCL,
which
is
based
on
inorganic
arsenic
and
should
not
apply
soils
screening
levels
based
on
inorganic
arsenic.

EFED
answer:
The
drinking
water
assessment
estimates
exposure
to
total
arsenic
and
states
that
"
the
total
arsenic
EDWC
can
be
compared
to
regulatory
levels,
all
of
which
are
defined
by
total
arsenic.
For
drinking
water,
the
Maximum
Contaminant
Level
(
MCL)
for
total
arsenic
is
10
ppb
and
is
based
on
concerns
for
long­
term
exposure.
As
a
conservative
assumption,
the
entire
estimated
total
arsenic
EDWC
may
be
present
as
inorganic
arsenic."

The
same
is
true
for
soil
screening
levels,
which
are
regulated
in
terms
of
total
arsenic.
The
potential
for
buildup
of
arsenicals
in
soil
is
estimated
as
total
arsenic
and
is
compared
to
the
Eco­
SSL
standards,
also
in
terms
of
total
arsenic.

The
MCL
was
established
by
EPA's
Office
of
Water
and
the
Eco­
SSLs
were
established
by
EPA's
Office
of
Solid
Waste
and
Emergency
Remediation.
OPP
believes
it
must
coordinate
with
other
offices
within
the
Agency
to
better
understand
and
anticipate
the
results
that
OPP's
actions
under
FIFRA
may
have
on
the
progress
of
other
offices.

XI
 
A.
The
Risk
Assessment
uses
an
overly
conservative
default
35­
day
half­
life
for
organic
arsenic
herbicide
residues
on
foliage
used
by
animals
as
a
food
source.

OPP
calculated
environmental
exposure
concentrations
for
MSMA
and
DSMA
using
the
unreasonable
default
assumption
that
the
half­
life
for
both
MSMA
and
DSMA
on
foliar
surfaces
is
35
days.
This
default
value
was
derived
from
Willis
and
McDowell
(
1987),
who
tabulated
half­
lives
for
many
pesticides
applied
to
many
crops,
but
no
data
for
MSMA
or
DSMA
were
included
in
the
study.
Out
of
447
half­
lives
reported,
only
two
were
greater
than
35
days,
demonstrating
this
to
be
an
extreme
value.
The
TTR
study
conducted
to
evaluate
the
amount
of
MSMA
that
can
be
transferred
from
turf
to
standard
cloth
material
(
MRID#
44958901),
however,
showed
that
the
half­
life
of
MSMA
residue
Page
20
of
23
on
turf
ranges
from
0.7
days
to
9.5
days.
Sound
risk
assessment
principles
and
the
law
both
require
that
OPP
apply
MSMA­
specific
data
from
the
TTR
study
rather
than
its
generic
default
assumptions.

EFED
answer:
The
EFED
uses
the
upper
bound
estimate
from
Willis
and
McDowell
(
1987)
when
other
appropriate
adequate
chemical
specific
data
are
not
available
to
account
for
potential
persistence
on
avian
and
mammal
food
items.
The
transferable
residue
study
cited
by
the
registrants
is
not
a
study
type
approved
for
replacing
the
default
foliar
dissipation
half­
life.
Measurement
of
total
residues,
not
transferable
residues,
is
required
to
account
for
all
pesticide
that
may
remain
embedded
on
foliage
used
by
animals
as
a
food
source.
Furthermore,
foliar
dissipation
of
total
residues
would
need
to
be
measured
on
a
variety
of
food
items,
under
a
variety
of
environmental
conditions.

XI
 
B.
EPA
inappropriately
uses
the
highest
single­
day
EEC
to
calculate
chronic
hazard
quotients.

OPP
inappropriately
uses
the
maximum
single­
day
EEC
to
evaluate
chronic
risk,
even
though
the
TREX
model
calculates
the
foliar
residue
concentration
for
each
day
of
one
year,
and
chronic
effects
arise
from
exposure
that
is
substantially
longer
than
a
single
day.
EPA
ecological
risk
assessment
guidance
notes
that,
"
the
arithmetic
average
exposure
over
the
time
period
of
toxicological
significance
is
the
appropriate
statistic"
18
for
evaluating
exposure
and
hazards.
Therefore,
chronic
risk
quotients
should
be
calculated
using
EECs
that
occur
over
a
period
of
time
that
is
appropriate
for
the
receptor
(
e.
g.,
30
days
for
organisms
with
a
short
lifecycle;
1
year
for
organisms
with
a
long
life­
cycle).
Using
the
maximum
single
day
EEC
ignores
the
reality
of
how
exposure
occurs.

EFED
answer:
The
endpoint
for
the
chronic
risk
quotient
is
the
lowest
NOAEC
from
rat
reproductive
studies.
The
reproductive
effects
observed
in
these
studies
may
result
from
a
single
exposure;
long­
term
exposure
may
not
be
necessary
to
elicit
the
observed
effects.
The
EFED
screening
risk
assessment
uses
the
single­
day
maximum
estimated
EEC
as
a
conservative
approach.
The
degree
to
which
this
exposure
is
conservative
cannot
be
determined
by
the
existing
reproduction
study.
Page
21
of
23
III.
FDACS
Comments
1a.
Despite
the
recognition
of
the
persistence
of
arsenic
in
the
organic
arsenical
herbicide
assessment,
the
EPA
does
not
provide
a
sufficiently
long­
term
quantitative
environmental
fate
analysis
for
organic
arsenical
herbicides 
The
fate
information
used
in
the
EPA
risk
assessments
considers
the
fate
of
organic
arsenical
herbicides
over
an
inadequate,
relatively
short
duration
on
the
order
of
only
months,
which
in
fact,
a
much
longer
period
on
the
order
of
several
years
to
decades
should
be
considered
to
accurately
assess
the
fate
and
risks
from
the
use
of
these
herbicides.

EFED
concurs
that
arsenicals
are
persistent
in
the
environment.
In
considering
the
environmental
fate
of
arsenicals,
EFED
is
limited
to
the
available
data
which,
as
noted
by
Florida,
are
on
the
scale
of
several
months
to
several
years.
EFED
is
not
aware
of
any
studies
providing
additional
data
on
longer
time
scales
of
several
years
to
decades,
but
will
consider
any
studies
that
are
submitted.

EFED's
modeling
inputs
are
derived
from
laboratory
studies
conducted
over
several
months
to
a
year.
First
order
half­
lives
derived
from
these
data
are
calculated
as
the
upper
90%
confidence
limit
on
the
mean
and
are
expected
to
be
sufficiently
conservative
in
representing
the
long
term
fate
of
arsenicals.
Exposures
are
estimated
based
on
30
years
of
daily
values
generated
by
models
using
these
conservative
half­
lives.
Further
characterization
of
the
environmental
fate
of
arsenicals
is
provided
based
on
field
studies
on
the
scale
of
one
to
several
years.

1b.
The
EPA
assessment
recognizes
that
organic
arsenical
herbicides
convert
to
inorganic
arsenic
over
time
under
a
wide
range
of
environmental
conditions 
FDACS
agrees
with
the
Agency's
conclusion
that [
volatilization
of
organoarsines]
seems
to
be
an
insignificant
route
of
dissipation
for
both
organic
and
inorganic
forms
of
arsenic."

EFED
concurs
that
transformation
of
organic
arsenical
herbicides
to
inorganic
arsenic
is
likely
to
occur
to
some
extent
under
a
variety
of
environmental
conditions
and
accounts
for
this
transformation
in
this
assessment.
Volatilization
is
not
considered
as
a
major
dissipation
route.

2a.
Although
the
EPA's
Drinking
Water
Assessment
recognizes
that
organic
arsenical
herbicides
could
pose
a
risk
to
groundwater
quality
"
in
highly
vulnerable
situations,"
the
risk
assessment
provides
no
quantitative
assessment
on
the
potential
acute
or
chronic
health
risks
from
ingesting
organic
arsenical
herbicides
or
their
metabolites
from
potable
groundwater
sources 
Recognizing
Florida's
vulnerable
conditions
and
the
persistence
of
arsenic
species,
FDACS
asks
EPA
to
further
evaluate
the
potential
for
exposure
to
arsenical
herbicides
via
the
groundwater
pathway.

b.
Many
forces,
both
natural
and
anthropogenic,
cause
arsenicals
to
dislodge
from
soil
and
become
remobilized
in
the
environment.
These
forces
may
be
sufficient
to
increase
arsenic
concentrations
in
groundwater
by
up
to
several
orders
of
magnitude.
Page
22
of
23
EFED
agrees
that
in
certain
areas,
use
of
organic
arsenical
herbicides
could
pose
a
risk
to
groundwater
quality.
The
risk
assessment
concludes
that
"
in
most
situations,
organic
arsenical
pesticides
should
not
contribute
significantly
to
the
already
existing
burden
of
arsenic
in
groundwater
from
all
sources,
natural
and
anthropogenic.
In
certain
vulnerable
circumstances
in
areas
with
low
background
arsenic,
application
of
organic
arsenicals
may
lead
to
an
increase
in
groundwater
total
arsenic."

The
conclusion
that
leaching
is
possible
in
certain
vulnerable
environments
is
based
primarily
on
monitoring
studies
which
detected
elevated
arsenic
in
groundwater
beneath
Florida
golf
courses.
Florida
is
described
as
"
an
area
of
concern
because
of
its
very
sandy
soils
and
shallow
water
tables"
with
golf
courses
"
particularly
susceptible
to
leaching
because
they
typically
have
well­
drained
soils
and
are
heavily
irrigated."
Similar
environments,
both
within
Florida
and
nationally,
would
be
considered
to
be
highly
vulnerable;
these
include
areas
with
shallow
water
tables,
low
background
arsenic
levels,
and
well­
drained
soils
to
which
arsenicals
do
not
sorb
strongly.
Soil
types
in
hydrologic
groups
A
and
B
are
expected
to
be
more
prone
to
leaching;
in
particular,
soil
properties
promoting
arsenical
leaching
include
low
clay
content
and
low
iron
and
aluminum
contents.
The
assessment
acknowledges
that
there
are
multiple
factors
that
may
affect
arsenical
mobility,
noting
that
"
phosphate
is
expected
to
strongly
compete
for
binding
sites"
and
that
"
remobilization
of
sorbed
arsenic
with
changing
environmental
conditions
is
also
possible."

Because
of
limitations
in
the
available
screening
models,
modeling
was
not
conducted
to
estimate
the
potential
for
exposure
to
arsenicals
in
groundwater.
Instead,
EFED
presented
the
available
targeted
monitoring
data,
with
detections
of
up
to
120
ppb
arsenic
in
groundwater
below
golf
courses.
Based
on
these
studies,
which
include
monitoring
studies
cited
in
Florida's
comment,
EFED
found
a
"
trend
of
groundwater
arsenic
at
Florida
golf
courses
well
above
local
background
levels
and
in
many
cases
exceeding
the
federal
MCL.
Although
evidence
of
the
source
of
arsenic
[
was]
generally
not
conclusive,
pesticide
application
[
was
found
to
be]
a
probable
contributor
to
these
levels,
and
in
some
cases,
appears
to
be
the
likely
source."
An
additional
data
set
cited
in
registrant
comments
includes
measurements
of
arsenic
levels
in
Florida
Community
Water
Systems.
In
groundwater
intakes
in
the
Florida
public
water
supply,
multiple
detections
of
arsenic
were
found
above
typical
background
levels,
with
a
number
of
the
detections
occurring
in
intakes
near
parks,
golf
courses,
or
residential
areas,
all
potential
users
of
arsenicals.
The
highest
detected
arsenic
level
in
the
data
set,
240
ppb,
was
measured
in
an
intake
near
a
commercial
sod
operation.
EFED
agrees
that
a
prospective
ground
water
study
in
Florida
would
be
very
useful
in
defining
the
potential
risks
from
leaching
of
applied
organic
arsenicals.
If
submitted,
EFED
will
consider
these
results
in
assessing
exposure
to
arsenicals
in
groundwater.

When
considering
how
changes
to
the
labeled
use
pattern
may
lead
to
adjustments
of
the
estimated
exposure,
it
is
important
to
keep
in
mind
the
different
ways
in
which
surface
water
and
ground
water
may
be
affected.
For
example,
the
Drinking
Water
Assesment
acknowledges
that
spot
treatment
is
a
common
practice
for
use
of
organic
arsenicals
on
turf.
If
that
practice
is
incorporated
in
the
label
in
a
quantifiable
way,
EFED
states
that
Page
23
of
23
"
the
estimated
concentrations
will
be
reduced
by
a
proportional
amount."
This
is
true
for
surface
water
but
may
have
less
of
an
impact
on
groundwater
exposure.
Unlike
surface
water,
which
can
be
impacted
by
runoff
of
pesticides
from
an
entire
watershed,
leaching
of
pesticides
to
groundwater
directly
impacts
the
local
area
of
application.
Used
as
a
spot
treatment,
the
same
amount
of
pesticide
may
reach
groundwater
in
the
treated
area,
but
the
occurrence
of
leaching
would
be
less
widespread.
Spot
treatment
then,
would
not
necessarily
decrease
estimated
concentrations
in
groundwater
but
could
cause
exposures
to
be
more
localized.

IV.
FDEP
Comments
Descriptions
of
studies
in
Florida
on
arsenic
impacts
to
groundwater
beneath
golf
courses
in
the
[
RED]
are
adequate
to
show
the
basis
for
concern
regarding
organoarsenical
use 
we
concur
with
the
conclusion
that,
"
Specific
to
arsenic,
recent
pesticide
application
data
coupled
to
concentration
trends,
along
with
known
historical
uses,
point
to
MSMA
application
as
a
likely
source
of
groundwater
arsenic
at
these
golf
courses." 
Residential
use
of
organoarsenical
herbicides,
and
its
impact
on
drinking
water,
is
also
a
potential
problem 
Most
of
the
discussion
of
problems
with
arsenic
in
soil
and/
or
groundwater
from
organoarsenical
use
in
the
risk
assessment
focuses
on
Florida
golf
courses.
The
clear
impression
from
the
risk
assessment
is
that
arsenic
accumulation
and
groundwater
contamination
problems.

EFED
answer:
As
stated
in
this
comment,
much
of
the
discussion
of
groundwater
contamination
from
the
turf
use
of
organic
arsenicals
is
based
on
monitoring
results
at
Florida
golf
courses.
This
is
because
the
available
monitoring
data
targeted
to
this
use
are
from
Florida
sites.
Similar
environments,
both
within
Florida
and
nationally,
would
be
considered
to
be
highly
vulnerable
and
have
the
same
risks
of
groundwater
contamination,
as
discussed
in
more
detail
in
the
response
to
comments
from
FDACS.
