Page
1
of
2
UNITED
STATES
ENVIRONMENTAL
PROTECTION
AGENCY
WASHINGTON,
D.
C.
20460
Date:
September
29,
2005
PC
Code:
038901,
038903,
038904,
038905
DP
Barcode:
D322027
MEMORANDUM
Subject:
Endothall:
Addendum
to
the
drinking
water
assessment
dated
May
5,
2004.

To:
Robert
McNally,
Chief
Special
Review
Branch
Special
Review
and
Reregistration
Division
(
7508C)

From:
James
Breithaupt,
Agronomist
Dana
Spatz,
Senior
Chemist
Environmental
Risk
Branch
II
Environmental
Fate
and
Effects
Division
(
7507C)

Thru:
Tom
Bailey,
Chief
Environmental
Risk
Branch
II
Environmental
Fate
and
Effects
Division
(
7507C)

For
most
pesticides,
modeled
concentrations
in
surface
water
and
ground
water
are
generally
higher
than
monitored
concentrations.
However
for
endothall,
the
apparent
measured
concentrations
in
surface
water
and
ground
water
were
higher
than
modeled
values
for
at
least
some
samples
reported
above
the
minimum
reporting
limit
(
MRL).
When
this
occurs,
further
analysis
and
discussion
is
warranted.

Modeling
for
surface
water
concentrations
is
normally
based
on
PRZM­
EXAMS
where
a
treated
terrestrial
field
drains
to
a
farm
pond
or
Index
Reservoir.
It
does
not
address
direct
application
to
water.
For
these
aquatic
uses,
the
peak
concentration
is
the
target
application
rate,
while
a
simple
first­
order
half
life
model
is
applied
to
estimate
an
annual
mean
concentration.
The
direct
application
of
endothall
to
water
bodies,
including
drinking
water
reservoirs,
was
modeled
assuming
uniform
application
over
the
entire
reservoir
at
the
maximum
label
rate.
Based
on
the
result
of
the
direct
application
scenario,
the
estimated
maximum
surface
water­
derived
drinking
water
concentrations
are
5
mg/
L
(
peak)
and
0.545
mg/
L
(
annual
mean).

Based
on
NCOD
data
from
the
Office
of
Water,
the
peak
surface
water
source
drinking
water
concentration
was
2.9
mg/
L,
while
the
mean
of
the
8
detections
out
of
5,112
samples
(
a
0.16%
Page
2
of
2
detection
frequency)
was
0.87
mg/
L.
This
mean
value
is
higher
than
the
model­
predicted
annual
mean.
However,
if
the
entire
dataset
was
considered
in
these
statistics,
rather
than
just
the
8
detections,
mean
concentrations
would
be
considerably
lower.
Albeit,
concentrations
in
this
range
may
be
the
result
of
direct
application
to
water
(
with
application
rates
up
to
5
mg/
L)
and
withdrawal
from
an
intake
in
the
flowpath
from
the
treatment
zone.

For
ground
water,
monitored
values
were
also
greater
than
modeled
concentrations.
In
NCOD
monitoring
data,
the
observed
peak
water
concentration
was
4.6
mg/
L,
while
the
mean
of
the
7
detections
out
of
27,494
samples
(
0.03%
detection
frequency)
was
0.67
mg/
L.
Again,
if
the
entire
dataset
was
considered
in
these
statistics,
rather
than
just
the
7
detections,
mean
concentrations
would
be
considerably
lower.

SCI­
GROW
modeling
assumes
a
terrestrial
application
with
subsequent
leaching
through
the
soil
profile
into
ground
water.
For
endothall,
the
result
was
0.086
ug/
L
(
ppb).
This
SCI­
GROW
estimate
assumes
application
of
two
pounds
of
acid
equivalents/
acre,
which
is
the
maximum
labeled
seasonal
rate
for
endothall
for
terrestrial
use.
Terrestrial
application
of
at
least
7,800
lbs
of
acid
equivalent
per
acre
would
be
required
to
reach
the
mean
detection
reported
in
NCOD.
Assuming
these
monitored
values
are
valid
exposure
estimates,
it
is
virtually
impossible
that
terrestrial
use
led
to
these
concentrations.
The
persistence
of
endothall
in
the
field
is
only
7­
19
days,
which
would
allow
degradation
prior
to
reaching
ground
water
in
almost
all
environmental
conditions.
In
terrestrial
field
dissipation
studies,
endothall
only
leached
into
the
4­
8
inch
soil
depth
in
a
California
study
and
into
the
8­
12
inch
soil
depth
at
a
Pennsylvania
site.
In
comparison,
wells
would
be
drawing
water
from
much
greater
depths
than
8­
12
inches.

One
possible
explanation
for
these
monitored
ground
water
concentrations
would
be
that
the
well
drew
water
directly
from
the
treated
water
body
carrying
residues
of
endothall
acid
through
the
soil.
Unless
there
is
a
direct
hydrologic
link
from
surface
water
to
a
ground
water
well,
direct
applications
to
water
could
not
lead
to
this
occurrence.
A
direct
hydrologic
link
between
surface
water
and
high
production
municipal
wells
would
be
rare.

Uncertainties
There
are
a
number
of
variables
which
may
contribute
to
the
uncertainty
in
reported
drinking
and
ground
water
monitoring
data.
These
may
include,
but
are
not
limited
to,
unconfirmed
errors
in
analysis,
unverifiable
transcription
errors,
reporting
discrepancies,
and
reliance
upon
data
from
samples
collected
before
current
procedures
were
in
place.
Historically,
it
has
been
demonstrated
from
compliance
and
occurrence
data
from
treated
water
systems
that
detections
of
very
low
frequency
may
be
generally
categorized
as
'
outliers'.
Although
this
trend
may
not
apply
to
endothall,
the
potential
for
the
above
errors
is
plausible.
