Appendix
B,
Page
1
of
5
Appendix
B.
Executive
Summary
from:
A
Probabilistic
Model
and
Process
to
Assess
Risks
to
Aquatic
Organisms
Full
document
is
available
at:
http://
www.
epa.
gov/
scipoly/
sap/
2001/
march/
aquatic.
pdf
Appendix
B,
Page
2
of
5
A
Probabilistic
Model
and
Process
to
Assess
Risks
to
Aquatic
Organisms
May
13­
16,
2001
FIFRA
Scientific
Advisory
Panel
Meeting
Kathryn
Gallagher,
Ph.
D.,
Leslie
Touart,
Ph.
D.,
James
Lin,
Ph.
D.
US
EPA
Office
of
Pesticide
Programs
Environmental
Fate
and
Effects
Division
Timothy
Barry,
Sc.
D.
US
EPA
Office
of
the
Administrator
Office
of
Policy,
Economics
and
Innovation
February
19,
2001
Appendix
B,
Page
3
of
5
Executive
Summary
Background
The
U.
S.
EPA
is
implementing
a
new
tiered
process
for
conducting
ecological
risk
assessment,
which
will
be
used
under
the
Federal
Insecticide,
Fungicide,
and
Rodenticide
Act
(
FIFRA)
regulatory
framework.
This
process
is
being
carried
out
by
the
Environmental
Fate
and
Effects
Division
(
EFED)
of
the
Office
of
Pesticide
Programs.
This
approach
will
include
the
use
of
probabilistic
tools
at
the
more
refined
levels
(
tiers)
of
the
risk
assessment
process.
The
aim
of
this
process
is
to
provide
information
on
the
probability
or
likelihood
of
ecological
impact,
and
on
the
magnitude
or
severity
of
the
effect.
The
results
of
probabilistic
analyses
will
be
used
by
risk
managers
in
their
decision­
making
process.

In
1996,
the
Agency's
Scientific
Advisory
Panel
(
SAP)
endorsed
the
risk
quotient
approach
as
a
suitable
screening
method
to
identify
pesticide
uses
of
potential
concern
that
warrant
refined
assessment.
The
SAP
suggested
that
EFED
incorporate
probabilistic
methods
into
such
refined
assessments,
noting
that
such
approaches
can
provide
estimates
of
the
probability
and
magnitude
of
effects.

Over
the
past
three
years,
EFED,
along
with
individuals
from
academia,
industry,
and
nongovernmental
organizations,
has
worked
to
develop
a
suite
of
refined
risk
assessment
methods
through
the
Ecological
Committee
on
FIFRA
Risk
Assessment
Methods
(
ECOFRAM)
initiative.
Concluded
in
1999,
reports
originating
out
of
the
ECOFRAM
initiative
served
as
the
basis
for
development
of
an
EFED
implementation
plan
for
conducting
probabilistic
ecological
risk
assessments.
This
implementation
plan,
presented
to
the
SAP
in
April
of
2000,
outlined
a
proposed
general
approach
for
assessing
pesticide
risks
to
birds
and
aquatic
organisms,
including
the
use
of
probabilistic
tools
in
a
tiered
manner.

As
part
of
the
process
of
developing
and
implementing
a
probabilistic
approach
for
ecological
risk
assessment,
EFED
has
completed
a
case
study.
The
case
study
involved
both
deterministic
and
probabilistic
risk
assessments
for
ChemX.

In
the
deterministic,
screening­
level
assessment,
risks
to
birds
and
aquatic
organisms
from
exposure
to
ChemX
were
concluded
to
be
high.
It
was
determined
that
a
refinement
of
the
risk
assessment
was
warranted.
The
refined
risk
assessment
presented
herein
follows
the
approaches
outlined
in
EFED=
s
implementation
plan
for
conducting
probabilistic
terrestrial
and
aquatic
organism
risk
assessments.
The
case
study
is
designed
to
express
risk
in
more
quantitative
terms
than
in
the
deterministic
approach,
to
better
understand
the
likelihood
and
magnitude
of
expected
effects
to
fish
and
aquatic
invertebrates.

For
this
case
study,
the
aquatic
assessment
was
limited
to
4
crops
for
a
single
chemical
example
and
product
formulation.
Specific
usage
characteristics
were
considered
important
in
deciding
which
crops
to
examine:
(
1)
application
rate,
(
2)
associated
estimated
exposure
concentrations,
Appendix
B,
Page
4
of
5
(
3)
acreage
treated,
and
(
4)
representativeness
of
the
use
site.
Corn,
cotton,
potatoes,
and
grapes
were
the
use
scenarios
selected
for
the
case
study
analysis.
The
variety
of
crops,
regions,
use
rates,
and
scenarios
selected
represent
a
broad
and
reasonably
comprehensive
selection
for
uses
of
a
chemical.

For
each
acute
effects
assessment
for
a
given
species,
a
distribution
of
estimated
exposure
and
a
distribution
of
toxicity
were
combined
in
a
joint
probability
function
through
Monte
Carlo
analysis.
The
resultant
function
allows
the
expression
of
risk
in
terms
of
the
likelihood,
or
probability,
of
an
exposure
concentration
occurring,
and
the
related
magnitude
of
effect
in
percent
mortality
within
a
given
species
expected
to
occur
at
that
level.
For
each
chronic
assessment,
a
distribution
of
exposure
concentrations
was
compared
to
a
chronic
effect
point
estimate.
The
chronic
analyses
give
information
only
on
the
probability
that
the
chronic
endpoint
assessed
would
be
exceeded,
not
on
the
magnitude
of
chronic
effect
expected.
In
both
acute
and
chronic
assessments,
estimated
environmental
concentrations
generated
from
PRZM/
EXAMS
simulation
modeling
were
directly
compared
to
the
toxic
effect
in
the
laboratory
at
the
same
exposure
concentration
(
i.
e.,
internal
dose
to
the
animal
was
not
considered).

A
deterministic
screen
conducted
on
ChemX
concluded
qualitatively
that
it
can
pose
a
high
risk
to
aquatic
organisms.
Through
the
deterministic
risk
quotient
method,
the
acute
restricted
use
Level
of
Concern
(
LOC)
was
exceeded
for
freshwater
fish
for
all
crops
examined.
For
freshwater
invertebrates,
the
acute
high
risk
LOC
was
exceeded
for
most
of
the
crops
examined.
The
chronic
LOC
was
exceeded
for
fish
for
potatoes,
sorghum
,
and
grapes,
in
a
comparison
of
the
peak
estimated
environmental
concentrations
to
chronic
no­
observed­
effects­
concentration,
while
chronic
the
LOC
was
exceeded
for
all
crops
for
freshwater
invertebrates.
Based
on
the
probabilistic
analysis
undertaken
herein,
it
was
concluded
that
the
use
of
ChemX
was
expected
to
infrequently
(
5%
of
the
time
or
less)
result
in
significant
freshwater
fish
mortalities,
but
routinely
result
in
reduced
growth
and
other
chronic
effects
in
exposed
fish.
Substantial
mortalities
and
chronic
effects
to
sensitive
aquatic
invertebrates
were
predicted
to
routinely
occur
after
peak
exposures.
It
was
concluded
that
estuarine
fish
and
invertebrates,
which
are
more
sensitive
than
their
freshwater
relatives,
are
likely
to
experience
substantial
mortality
when
exposed
to
concentrations
in
their
habitat
equivalent
to
what
has
been
modeled
for
farm
ponds
in
the
use
scenarios
studied.

The
probability
of
high
mortality
rates
for
invertebrates
would
be
likely
to
have
an
indirect
effect
on
fish,
as
would
the
high
probability
of
exceeding
invertebrate
chronic
no­
effects
concentrations.
Indirect
effects
on
fish
could
include
reduced
juvenile
fish
survival
due
to
reduced
invertebrate
food
resources,
with
concomitant
potential
alterations
in
fish
population
structure,
and
potential
decreases
in
fish
populations
or
other
higher
trophic
levels.

Sublethal
effects
on
fish
from
acute
exposures
to
ChemX
were
not
addressed.
These
endpoints
are
generally
not
assessed
by
the
standard
toxicity
test
protocols
used
by
EFED,
but
sublethal
effects
from
acute
exposure
should
be
considered,
at
least
qualitatively,
when
evaluating
the
risk
to
fish
from
a
chemical.
Were
dose­
response
data
for
acute
effects
of
ChemX
available,
these
data
Appendix
B,
Page
5
of
5
could
have
been
used
in
the
probabilistic
model
developed.

There
are
a
number
of
uncertainties
associated
with
both
the
refined
and
deterministic
risk
assessments
that
are
not
included
in
this
case
study
risk
assessment.
These
include
uncertainty
regarding
the
error
introduced
when
extrapolating
from
laboratory
to
field
effects.
For
example,
mortality
in
the
field
could
be
greater
in
populations
previously
stressed
by
other
pesticide
exposures,
habitat
loss,
predation,
temperature
stress,
or
competition
for
limited
resources,
among
other
things.
Overall
field
mortality
could
be
lower
if
the
laboratory
population
were
to
represent
an
unusually
sensitive
species.
There
is
uncertainty
associated
with
the
use
of
single
laboratory
toxicity
tests
for
a
species.
The
experimental
uncertainty
regarding
the
representativeness
of
the
test
results
could
be
reduced
through
replication
of
the
toxicity
tests
for
each
species.
There
is
uncertainty
regarding
the
extrapolation
of
expected
effects
to
more
sensitive
species,
and
the
frequency
with
which
these
more
sensitive
species
occur
in
ChemX
use
areas.
There
is
uncertainty
regarding
the
extrapolation
to
all
invertebrates
from
a
data
set
containing
only
arthropods.
There
is
also
uncertainty
regarding
the
representativeness
of
the
fate
input
parameters
used
in
the
exposure
modeling;
some
input
parameters
were
based
on
limited
physico­
chemical
data.
Another
source
of
uncertainty
is
the
use
of
exposure
model­
derived
estimated
concentrations
to
represent
actual
environmental
concentrations.
These
uncertainties
could
,
in
general,
be
reduced
through
additional
field
testing
and
measurement.

The
uncertainties
that
were
addressed
in
the
probabilistic
analyses,
and
which
are
included
in
the
confidence
bounds
given
for
the
risk
estimates,
reflect
uncertainty
in
the
risk
analysis
construct
and
primarily
regard
the
fits
of
the
applied
distributions
to
the
data.
The
risk
analysis
construct
for
this
assessment
involved
the
use
of
PRZM/
EXAMS
model
exposure
data
and
laboratory
toxicity
data
as
fit
by
a
log­
probit
model.
As
noted
above
these
uncertainty
estimates
reflect
only
a
portion
of
the
true
uncertainty.
However,
the
uncertainty
expressed
does
address
the
fact
that
there
is
uncertainty
in
the
estimates
of
risks
indicated
by
the
RQs,
and
is
an
effort
to
provide
quantitative
estimates
of
detrimental
effect
and
bounds
on
those
effect
estimates.

The
ChemX
incident
history
for
fish
is
relatively
scant
in
comparison
to
birds,
and
the
probabilistic
analysis
explains
that
large
die­
offs
would
be
an
infrequent,
though
real,
occurrence.
Invertebrates
can
be
expected
to
suffer
more
routine
mortalities
and
other
ill
effects,
although
these
would
be
unlikely
to
be
captured
in
an
incident
report
due
to
their
subtlety.
Similarly,
sublethal
and
chronic
effects
to
fishes
would
not
be
tractable
through
incident
reports.
Nevertheless,
use
of
ChemX
would
be
expected
to
pose
real
and
significant
risks,
acutely
and
chronically,
to
aquatic
organisms.
