1
Guidance
on
Demonstrating
Compliance
With
the
Land
Disposal
Restrictions
(LDR)
Alternative
Soil
Treatment
Standards
­
Interim
Guidance
(EPA530­
D­
00­
002):

June,
2002,
Response
to
Comments
on
66
FR
52918
(October
18,
2001)

Commenter:
U.
S.
DEPARTMENT
OF
ENERGY
Comment:
The
U.
S.
Department
of
Energy
(DOE
or
the
Department)
appreciates
the
effort
of
the
U.
S.
Environmental
Protection
Agency
(EPA)
in
developing
and
providing
the
"Interim
Guidance
on
Demonstrating
Compliance
With
the
Land
Disposal
Restrictions
(LDR)
Alternative
Soil
Treatment
Standards."
Generally,
DOE
believes
the
Interim
Guidance
should
help
the
regulated
community
to
better
prepare
demonstrations
of
compliance
with
the
alternative
LDR
treatment
standards
for
contaminated
soils.

Response:
The
Agency
appreciates
the
commenter's
support.

Comment:
DOE
notes
that
much
of
the
Interim
Guidance
focuses
on
use
of
the
Data
Quality
Objectives
(DQO)
process
to
develop
a
sampling
strategy
for
characterizing
the
contaminant
concentrations
within
soil
to
which
the
alternative
LDR
treatment
standards
will
be
applied.
DOE
believes
targeted
DQO
guidance
documents
of
this
type
could
also
be
helpful
to
members
of
the
regulated
community
who
must
develop
sampling
strategies
for
demonstrating
compliance
with
LDR
treatment
standards
other
than
the
alternative
soil
standards.
Accordingly,
DOE
suggests
that
EPA
consider
preparing
a
similar
guidance
document
targeted
at
development
of
sampling
strategies
for
characterizing
contaminant
concentrations
within
wastes
subject
to
the
LDR
treatment
standards
for
hazardous
wastes
identified
in
40
CFR
268.40.

Response:
EPA
OSW
currently
is
in
the
process
of
revising
the
sampling
guidance
found
in
Chapter
Nine
of
"Test
Methods
for
Evaluating
Solid
Waste"
EPA
publication
SW­
846.
While
a
draft
is
not
yet
available,
we
anticipate
that
the
document
will
include
guidance
on
sampling
for
compliance
with
LDR
standards.

Comment:
While
DOE
is
an
advocate
of
the
DQO
process,
the
Department
has
found
that
the
process
can
become
very
complex,
time
consuming,
and
expensive,
especially
when
it
is
applied
to
develop
multiple,
statistically­
based
sampling
approaches
at
a
site
undergoing
cleanup.
This
may
occur
when
multiple
contaminants
are
present
in
different
media
at
a
site,
and
each
contaminant
has
multiple
action
levels
(e.
g.,
cleanup
levels,
hazardous
characteristic
levels,
and
LDR
treatment
standards).
The
problem
can
be
compounded
even
further
at
facilities
that
must
consider
eco­
risk
issues
or
contain
numerous
cleanup
locations.
DOE
recommends
that
EPA
consider
devoting
future
resources
to
evaluating
the
need
for
statistically­
based
sampling
approaches
with
an
eye
toward
simplifying
the
task
2
of
applying
the
DQO
process
and
reducing
its
costs.
DOE
would
be
pleased
to
cooperate
with
EPA
in
such
an
effort.

Response:
The
DQO
Process
is
designed
to
be
flexible.
EPA
encourages
a
"graded"
approach
to
the
DQO
Process
such
that
the
level
of
planning,
management,
and
QA/
QC
is
commensurate
with
the
end
use
of
the
data
and
the
level
of
confidence
required
in
decisions
to
be
made
from
the
data.
EPA
has
added
text
to
the
guidance
in
two
places
to
reinforce
this
point.

Comment:
1.
INTRODUCTION
AND
BACKGROUND,
1.1
What
Is
the
Purpose
of
This
Guidance?
p.
1,
1
st
paragraph
in
the
section
–
The
Interim
Guidance
explains
that
40
CFR
268.49
allows
the
generator
of
contaminated
soils
that
will
be
land
disposed
to
elect
to
comply
with
either
the
LDR
alternative
soil
treatment
standards
at
40
CFR
268.49
or
the
generic
treatment
standards
at
40
CFR
268.40.

DOE
believes
it
would
be
helpful
to
the
regulated
community
for
this
section
of
the
Interim
Guidance
to
acknowledge
that
a
third
alternative
might
be
available
on
a
site­
specific
basis.
Specifically,
the
Interim
Guidance
should
mention
that,
if
a
generator
can
demonstrate
that
the
LDR
alternative
soil
treatment
standards
would
impose
treatment
of
contaminated
soil
beyond
the
point
at
which
threats
are
minimized,
the
generator
can
seek
a
site­
specific
risk­
based
treatment
variance
in
accordance
with
40
CFR
268.44(
h)(
3).

Response:
Site­
specific
LDR
variances
are
described
in
Appendix
A
of
the
guidance
(page
6)
in
the
memo
entitled
"Management
of
Remediation
Waste
Under
RCRA."
Detailed
guidance
also
is
given
in
EPA's
recent
(August
2001)
publication
"Land
Disposal
Restrictions:
Summary
of
Requirements"
(http://
www.
epa.
gov/
epaoswer/
hazwaste/
ldr/
guidance.
htm).
Note
also
that
variances
are
described
in
Section
1.5
(page
3)
of
the
Interim
Guidance.
The
Agency
appreciates
the
comment
and
will
add
additional
language
to
highlight
availability
of
variances
under
40
CFR
268.44(
h).

Comment:
p.
1,
2
nd
paragraph
in
the
section
–
The
Interim
Guidance
states
that
it
should
not
be
used
to
establish
site­
specific
soil
cleanup
standards.

DOE
believes
it
would
be
helpful
to
the
regulated
community
for
this
paragraph
to
briefly
explain
why
the
Interim
Guidance
should
not
be
used
to
establish
site­
specific
soil
cleanup
standards.
For
example,
a
discussion
similar
to
that
provided
in
the
preamble
to
the
final
rule
promulgating
the
LDR
alternative
soil
treatment
standards
would
be
useful
(see
63
FR
28556,
28606
(column
2);
May
26,
1998).

Response:
The
Agency
appreciates
the
comment
and
will
include
additional
language
in
the
guidance
to
explain
why
the
soil
treatment
standards
should
not
be
used
to
establish
site­
specific
soil
cleanup
standards.

The
purpose
of
the
land
disposal
restriction
treatment
standards
is
to
ensure
that
prohibited
hazardous
3
wastes
are
properly
pre­
treated
before
disposal
(i.
e.,
treated
so
that
short­
and
long­
term
threats
to
human
health
and
the
environment
posed
by
land
disposal
are
minimized).
The
soil
treatment
standards,
like
other
land
disposal
restriction
treatment
standards,
are
based
on
the
performance
of
specific
treatment
technologies.
In
contrast,
most
soil
cleanup
levels
are
based
not
on
the
performance
of
specific
treatment
technologies
but
on
an
analysis
of
risk.
Technology­
based
treatment
standards
are
not
necessarily
appropriate
surrogates
for
site­
specific
risk­
based
cleanup
levels.
In
a
circumstance
where
the
soil
treatment
standards
result
in
constituent
concentrations
that
are
higher
than
those
determined,
on
a
site­
specific
basis,
to
be
required
for
soil
cleanup,
existing
remedial
programs
such
as
RCRA
Corrective
Action,
CERCLA
and
state
cleanup
programs
could
be
applied
to
ensure
that
remedies
are
adequately
protective
(e.
g.,
require
a
site­
specific
cleanup
standard
that
is
lower
than
the
soil
treatment
standard).

Comment:
1.3
Why
Did
EPA
Develop
Alternative
Soil
Treatment
Standards?
p.
2
–
Section
1.3
of
the
Interim
Guidance
explains
that
EPA
developed
the
alternative
soil
treatment
standards
because
the
traditional
technology­
based
standards
were
intended
for
industrial
hazardous
waste
and
were
often
not
appropriate
or
not
achievable
when
applied
to
hazardous
constituents
present
in
soil.

DOE
suggests
that
Section
1.3
be
expanded
to
briefly
describe
the
LDR
compliance
option
of
developing
site­
specific,
risk­
based
treatment
standards
through
the
variance
process
in
40
CFR
268.44(
h)(
3).
The
Department
believes
this
information
would
be
helpful
to
the
regulated
community
in
evaluating
options
for
treatment
of
hazardous
soils.
In
addition,
the
discussion
could
explain
why
EPA
based
the
alternative
soil
treatment
standards
on
technology
rather
than
risk.

Response:
Site­
specific
LDR
variances
are
described
in
Appendix
A
of
the
guidance
(page
6)
in
the
memo
entitled
"Management
of
Remediation
Waste
Under
RCRA."
Detailed
guidance
also
is
given
in
EPA's
recent
(August
2001)
publication
"Land
Disposal
Restrictions:
Summary
of
Requirements"
(http://
www.
epa.
gov/
epaoswer/
hazwaste/
ldr/
guidance.
htm).
Note
also
that
variances
are
described
in
Section
1.5
(page
3)
of
the
Interim
Guidance.
The
Agency
appreciates
the
comment
and
will
add
additional
language
to
highlight
availability
of
variances
under
40
CFR
268.44(
h)(
3).

Comment:
2.1
What
Steps
Should
I
Use
to
Plan
the
Sampling
and
Analysis
Program?
p.
6,
Step
3:
Identify
Inputs
to
the
Decision,
1
st
paragraph
in
the
section
–
The
Interim
Guidance
lists
the
following
informational
inputs
needed
to
resolve
the
question
of
whether
compliance
with
the
LDR
alternative
soil
treatment
standards
has
been
achieved:
a
list
of
the
underlying
hazardous
constituents;
the
units
of
measure
(e.
g.,
mg/
kg
or
mg/
L);
and
a
listing
of
appropriate
analytical
methods,
method
performance
criteria
(e.
g.,
for
precision
and
accuracy),
and
required
quantitation
limits.

Noticeably
absent
from
the
list
of
informational
inputs
described
in
the
Interim
Guidance
as
necessary
to
resolve
the
question
of
compliance
with
the
LDR
alternative
soil
treatment
standards
is
input
concerning
background
constituent
concentrations,
especially
metals.
Accordingly,
DOE
requests
that
EPA
address
this
issue
in
the
Interim
Guidance
by
discussing
acceptable
methods
for
determining
4
background
constituent
concentrations
and
for
excluding
them
from
the
determination
of
whether
compliance
has
been
achieved.
In
addition,
DOE
requests
that
the
Interim
Guidance
discuss
the
provisions
of
40
CFR
268.44(
h)(
4),
which
allow
EPA
or
the
authorized
state
agency
to
grant
a
variance
capping
treatment
at
natural
background
concentrations
in
circumstances
where
treatment
of
contaminated
soil
to
meet
the
LDR
alternative
soil
treatment
standards
would
result
in
concentrations
of
hazardous
constituents
that
are
below
(i.
e.,
lower
than)
natural
background
concentrations
at
the
site
where
the
treated
soil
will
be
disposed
of
on
land.

Response:
The
regulations
at
40
CFR
Part
268.44(
h)(
4)
allow
EPA
and
authorized
states
to
grant
a
site­
specific
LDR
treatment
variance
for
contaminated
soil
if
the
level
or
the
method
specified
in
the
soil
treatment
standards
would
result
in
concentrations
of
hazardous
constituents
that
are
below
(i.
e.,
lower
than)
natural
background
concentrations
at
the
site
where
the
contaminated
soil
will
land
disposed.
Natural
background
concentrations
are
constituent
concentrations
that
are
present
in
soil
which
have
not
been
influenced
by
human
activities
or
releases.
Because
natural
background
concentrations
may
vary
across
geographic
areas,
and
to
ensure
that
LDRs
will
only
be
capped
at
background
where
appropriate,
EPA
requires
that
individuals
who
wish
to
cap
LDR
treatment
at
natural
background
concentrations
apply
for
and
receive
a
treatment
variance.
Information
on
how
to
determine
background
concentrations
can
be
found
in
an
issue
paper
entitled
Determination
of
Background
Concentrations
of
Inorganics
in
Soils
and
Sediment
at
Hazardous
Waste
Sites
(EPA/
540/
S96
500)
published
by
EPA's
Office
of
Research
and
Development
and
the
Office
of
Solid
Waste
and
Emergency
Response
(http://
www.
epa.
gov/
nerlesd1/
pdf/
engin.
pdf).
In
addition,
consultation
with
a
professional
statistician
is
recommended
before
preparing
a
request
for
a
variance
from
LDR
treatment
standards
for
soil
based
on
background
concentrations.
EPA
edited
the
document
to
include
this
information.

Comment:
p.
7,
Step
4:
Define
the
Boundaries,
last
paragraph
on
p.
7
–
The
Interim
Guidance
indicates
that,
with
certain
exceptions,
mixing
hazardous
soil
with
nonhazardous
soil
may
be
impermissible
dilution.
The
first
exception
is
stated
in
part
as
follows:
"If
mixing
occurs
through
the
normal
consolidation
of
contaminated
soil
from
various
portions
of
a
site
that
typically
occurs
during
the
course
of
remedial
activities
or
in
the
course
of
normal
earthmoving
and
grading
activities,
then
the
Agency
does
not
consider
this
to
be
intentional
mixing
of
soil
with
nonhazardous
soil
for
the
purposes
of
evading
LDR
treatment
standards."

DOE
notes
that
the
delineation
and
excavation
of
soils
subject
to
remedial
action
thresholds
is
one
of
the
more
difficult
aspects
of
a
cost
effective
cleanup.
Costs
associated
with
leaving
staff
and
equipment
idle
while
waiting
for
the
receipt
of
cleanup
verification
results
can
be
so
prohibitive
that,
sometimes,
it
is
more
cost­
effective
to
"over
excavate"
into
"clean"
soil
in
order
to
ensure
that
all
soil
containing
contaminants
in
excess
of
proscribed
concentrations
has
been
removed
quickly.
Because
it
is
possible
that
over
excavation
such
as
this
could
be
interpreted
by
a
responsible
regulator
as
intentional
dilution,
since
cleaner
soil
is
being
mixed
with
the
soil
volume
targeted
for
cleanup,
DOE
requests
that
the
discussion
of
Step
4
in
Section
2.1
of
the
Interim
Guidance
be
enhanced.
Specifically,
DOE
5
recommends
that
EPA
clarify
that
over
excavation,
when
performed
as
described
above,
does
not
constitute
impermissible
dilution.
In
addition,
DOE
recommends
that
the
Interim
Guidance
provide
instructions
or
examples
of
how
removal
and
remediation
plans
should
be
structured
to
avoid
disagreements
about
whether
over
excavation
constitutes
impermissible
dilution.

Response:
The
Agency
understands
that
it
is
sometimes
advantageous
to
over­
excavate
contaminated
soils
as
part
of
implementing
a
cost­
effective
cleanup
and
minimize
the
need
for
multiple
mobilizations
of
a
field
team
for
sampling,
analysis,
and
removal.
Because
each
site­
specific
situation
is
unique,
the
Agency
cannot
provide
detailed
guidance
on
the
extent
to
which
over­
excavation
can
be
performed
before
it
would
be
considered
impermissible
dilution.
However,
gross
over­
excavation
could
be
viewed
as
impermissible
dilution
and
should
be
avoided.
EPA
added
a
discussion
of
this
topic
to
the
guidance.

Comment:
p.
8,
Step
4:
Define
the
Boundaries,
last
paragraph
on
p.
8
–
The
Interim
Guidance
states:
"Note
that
if
the
90­
percent
reduction
standard
is
used,
then
the
estimate
of
post­
treatment
concentrations
should
apply
to
the
same
unit
of
soil
characterized
initially."

When
relying
on
a
90­
percent
reduction
soil
treatment
standard,
a
strategy
is
needed
to
ensure
consistent
comparison
of
soil
before
and
after
treatment.
However,
some
soil
handling
technologies
and
treatment
technologies
may
significantly
change
both
the
volume
and
the
mass
of
soil
being
treated.
For
example,
thermal
treatment
technologies
tend
to
reduce
soil
mass
and
volume
as
moisture
and
combustible
fractions
are
driven
off
during
the
treatment
process.
Stabilization
technologies
can
result
in
an
increase
in
soil
mass
and
volume
because
these
technologies
rely
on
mixing
additives
with
the
unit
of
soil
being
treated.
Excavation
technologies
can
also
affect
soil
volume;
the
volume
of
in
situ
soil
tends
to
increase
after
the
soil
is
excavated.
Clearly,
volume
or
mass
of
soil
can
change
as
a
result
of
excavation
and
treatment.
As
a
result,
DOE
seeks
additional
clarification,
perhaps
through
the
use
of
hypothetical
examples,
to
clearly
communicate
to
the
regulated
community
whether
or
not
mass
and
volume
changes
caused
by
excavation
and
treatment
needs
to
be
addressed
in
the
post­
treatment
sampling
effort.

Response:
The
Agency
recognizes
that
some
soil
handling
technologies
and
treatment
technologies
may
significantly
change
both
the
volume
and
the
mass
of
soil
between
the
time
the
soil
is
generated
to
after
treatment
is
completed.
Note
that
the
regulation
at
40
CFR
268.49(
a)
indicates
that
the
treatment
standards
apply
to
"any
given
volume"
of
hazardous
soil
(see
chart
note
at
§268.49(
a)),
and
the
regulation
does
not
require
that
the
volume
or
mass
of
the
soil
subject
to
the
treatment
standards
remain
unchanged
after
treatment.
The
"identity"
of
the
parcel
of
the
soil,
however,
should
remain
intact
after
treatment
to
facilitate
comparison
of
the
pre­
treated
to
post­
treated
concentrations
in
the
soil.
For
this
reason,
the
guidance
document
simply
refers
to
the
"unit"
of
soil
that
should
be
used
for
comparing
pretreated
to
post­
treated
concentrations.
The
guidance
does
not
indicate
that
the
volume
or
mass
of
soil
needs
to
remain
the
same
before
and
after
treatment.
The
Agency
has
modified
the
language
describing
Step
4
of
the
DQO
Process
to
clarify
that
the
volume
of
soil
may
change
as
a
result
of
treatment.
6
Comment:
p.
9,
Text
Box:
Hypothetical
Example
–
The
Interim
Guidance
provides
an
example
of
defining
a
"given
volume"
of
contaminated
soil
subject
to
LDR
treatment
standards.

DOE
believes
it
would
be
helpful
for
the
Interim
Guidance
to
remind
the
regulated
community
that
the
number
of
samples
that
must
be
tested
to
determine
whether
soil
exhibits
a
hazardous
characteristic
might
be
reduced
using
the
following
relatively
simple
screening
technique.
Assume
that
the
total
amount
of
each
hazardous
constituent
of
concern
in
a
volume
of
contaminated
soil
(in
this
example,
benzene)
would
partition
to
the
extract
during
the
TCLP
(which
is
often
a
very
conservative
assumption).
Then,
using
the
fact
that
the
TCLP
has
an
inherent
20­
fold
dilution
factor
(liquid
to
solid
ratio),
calculate
the
highest
concentrations
that
could
be
present
before
hazardous
constituents
of
concern
would
cause
the
soil
to
exhibit
the
toxicity
characteristic.
For
example,
in
the
case
of
benzene,
the
TCLP
toxicity
characteristic
concentration
limit
is
0.5
mg/
L.
Therefore,
10
mg/
kg
of
benzene
would
have
to
be
present
in
soil
before
it
would
exhibit
the
toxicity
characteristic
for
benzene.
Compare
the
calculated,
highest
allowable
concentrations
of
hazardous
constituents
of
concern
with
total
concentrations
measured
in
each
volume
of
soil.
Only
volumes
of
soil
containing
total
concentrations
of
one
or
more
hazardous
constituents
in
excess
of
the
highest
allowable
concentrations
must
be
subjected
to
the
TCLP
in
order
to
determine
whether
such
volumes
are
hazardous.
In
the
hypothetical
example,
only
those
volumes
of
soil
containing
greater
than
10
mg/
kg
of
benzene
would
need
to
be
subjected
to
the
TCLP
in
order
to
determine
whether
they
were
hazardous.

Response:
The
commenter
is
correct
that
a
20:
1
dilution
factor
can
be
used
to
screen
samples
for
the
TC.
This
approach
is
specifically
allowed
in
SW­
846
Method
1311
(TCLP).
The
example
has
been
modified
to
show
how
total
concentrations
can
be
used
to
screen
soil
for
the
TC.

Comment:
p.
9,
Text
Box:
Hypothetical
Example
–
The
Interim
Guidance
provides
an
example
of
defining
a
"given
volume"
of
contaminated
soil
subject
to
LDR
treatment
standards.

DOE
requests
that
EPA
consider
including
at
least
one
other
hypothetical
example
that
is
more
complicated
and
more
typical
of
an
actual
cleanup
action.
Such
an
example
would
depict
a
cleanup
site
containing
several
different
locations
of
contamination
with
multiple
contaminants
having
various
types
of
action
levels.

Response:
At
this
time,
EPA
has
very
limited
experience
with
application
of
the
alternative
soil
treatment
standards
at
actual
sites.
EPA
believes
that
the
inclusion
of
additional
more
complex
hypothetical
examples
is
unnecessary
at
this
time,
and
may
in
fact
detract
from
the
flexibility
that
alternative
treatment
standards
offer.

Comment:
p.
11,
Step
7:
Optimize
the
Design
for
Obtaining
the
Data
–
The
Interim
Guidance
provides
references
to
several
other
documents
that
contain
detailed
instructions
on
development
and
optimization
of
a
sampling
plan.
7
Define
the
"given
volume
of
soil"
subject
to
LDRs
per
268.49(
a)
(use
DQO
Process
Step
4)

If
the
90%
reduction
standard
is
selected,
then
obtain
nu
random
samples
representing
the
untreated
soil.

Does
the
soil
attain
the
standard
of
10xUTS?
Start
Yes
NO
Because
Step
7
of
the
DQO
process
is
so
important,
DOE
recommends
that
(in
addition
to
referencing
other
available
guidance)
this
discussion
in
the
Interim
Guidance
be
expanded
to
include
information
focusing
on
the
development
and
optimization
of
sampling
plans
that
deal
specifically
with
the
management
of
contaminated
soil.

Response:
The
Agency
agrees
that
Step
7
of
the
DQO
Process
is
important
because
it
involves,
among
other
activities,
a
determination
of
the
appropriate
number
of
samples.
Users
of
the
guidance
should
be
aware
that
some
of
the
measurement
data
required
to
determine
compliance
with
the
treatment
standards
may
be
generated
as
part
of
normal
site
characterization
sampling
and
analysis
activities.
Inputs
into
Step
7
of
the
DQO
Process
need
to
take
this
existing
characterization
data
into
account.
Based
on
the
comment,
EPA
has
expanded
the
discussion
of
the
Step
7
of
the
DQO
Process
and
provide
reference
to
more
specific
guidance
on
how
to
calculate
the
appropriate
number
of
samples.

Comment:
2.3
How
Should
I
Evaluate
the
Data
to
Determine
Attainment
of
the
Treatment
Standards?
p.
13,
Figure
2
–
Figure
2
provides
a
generalized
flow
diagram
indicating
the
decisionmaking
process
for
determining
attainment
of
the
alternative
soil
treatment
standards.

Reproduced
below
is
a
portion
of
the
flow
diagram
contained
in
Figure
2
of
the
Interim
Guidance.

Regarding
this
portion
of
the
diagram,
DOE
offers
the
following
observations:

(1)
DOE
believes
the
flow
would
be
more
logical
if
the
diamond
box
("
Does
the
soil
attain
the
standard
of
10xUTS?")
were
placed
upstream
rather
than
downstream
of
the
box
containing
the
statement,
"If
the
90%
reduction
standard
is
selected,
then
obtain
nu
random
samples
representing
the
untreated
soil."
The
determination
of
whether
or
not
soil
attains
the
standard
of
10xUTS
should
be
the
first
step
in
the
flow
diagram,
because
if
the
soil
attains
this
standard,
it
can
be
disposed
as
hazardous
waste
without
further
treatment.
8
(2)
DOE
believes
the
box
containing
the
statement
"If
the
90%
reduction
standard
is
selected,
then
obtain
nu
random
samples
representing
the
untreated
soil"
is
misleading
because
the
statement
implies
that
random
sampling
of
the
untreated
soil
is
the
only
approach
that
can
be
used
to
determine
whether
the
90%
reduction
standard
has
been
met.
However,
as
Section
2.3.3.2
(Welch's
t­
Test)
indicates,
sometimes
systematic
sampling
designs
are
equally
acceptable
(see
p.
21,
discussion
on
"Procedure").

In
light
of
these
observations,
DOE
suggests
that
EPA
amend
the
flow
chart
to
better
reflect
the
narrative
descriptions
in
the
Interim
Guidance
of
the
various
methods
that
can
be
used
to
evaluate
attainment
of
the
alternative
soil
treatment
standards.

Response:
EPA
agrees
with
the
commenter
and
made
the
suggested
changes
to
the
figure.

Comment:
p.
13,
Figure
2
–
According
to
the
flow
diagram
in
Figure
2,
if
the
soil
does
not
attain
either
the
90­
percent
reduction
standard
or
the
standard
of
10xUTS,
then
further
treatment
is
always
required.

DOE
suggests
that
Figure
2
and
the
text
of
Section
2.3
be
revised
to
reflect
the
availability
of
certain
variances
in
the
event
that
treatment
of
soils
to
meet
the
LDR
alternative
soil
treatment
standards
is
either
not
possible
or
not
appropriate.
DOE
recognizes
that
the
alternative
LDR
treatment
standards
for
contaminated
soil
are
intended
to
be
achievable
by
well­
designed
and
well­
operated
technologies
appropriate
to
the
soil
matrix
and
constituents
of
concern.
However,
in
the
event
that
a
particular
soil
cannot
be
treated
by
such
a
system
to
meet
the
LDR
alternative
soil
treatment
standards,
it
is
DOE's
understanding
that
a
site­
specific
treatability
variance
based
on
the
"not
physically
possible"
test
under
40
CFR
268.44(
h)(
1)
would
still
be
available.
It
is
also
DOE's
understanding
that
a
"not
appropriate"
treatability
variance
under
40
CFR
268.44(
h)(
2)
could
be
approved
for
contaminated
soil
if
applying
the
LDR
alternative
soil
treatment
standards
would
present
unacceptable
risks
to
on­
site
workers
(for
example,
if
certain
explosive
or
radioactive
contaminants
are
present
in
the
soil).
Accordingly,
DOE
believes
it
would
be
helpful
to
the
regulated
community
if
the
Interim
Guidance
contained
instructions
for
evaluating
whether
one
of
these
variances
could
be
justified
for
contaminated
soil
in
a
particular
set
of
circumstances.

Response:
The
flow
chart
is
presented
in
the
guidance
for
use
in
data
quality
assessment
(DQA).
Use
of
the
flow
chart
assumes
that
the
generator
or
treater
has
determined
that
the
soil
is
subject
to
the
LDR
treatment
standards
and
has
elected
to
use
the
alternative
soil
treatment
standards
available
at
40
CFR
268.49.
For
clarity,
however,
EPA
added
information
to
the
chart
as
a
footnote
to
indicate
that
one
option
is
to
petition
for
a
variance
from
the
treatment
standards.
9
Commenter:
U.
S.
Army
Corps
of
Engineers
Comment:
Method
quality
objectives
for
sensitivity
are
not
adequately
addressed.
In
order
to
satisfy
the
LDR
treatment
standard,
the
concentration
of
contamination
must
be
reduced
by
90%
on
the
average
or
must
be
less
than
10
x
UTS.
For
example,
using
a
statistical
approach
the
null
hypothesis
is
H0
:
(Treated)
–
0.1
(Untreated)
>
0
m
m
This
suggests
that
the
laboratory
method
must
be
capable
of
reporting
quantitative
results
at
concentrations
less
than
0.1
(Untreated).
Similarly,
if
the
null
hypothesis
is:
m
H0
:
(Treated)
>
10
x
UTS
m
then,
quantitative
values
would
be
required
for
concentrations
less
than
10
x
UTS.
However,
Page
6
of
the
document
requires
only
that
the
analytical
method
be
"capable
of
detecting
constituents
of
concern
at
concentrations
less
than
10
UTS."
In
general,
the
requirement
that
the
detection
limit
(e.
g.,
as
opposed
to
the
quantitation
limit)
must
be
less
than
the
decision
limit
(e.
g.,
10
UTS)
will
not
be
adequate
to
support
the
decision­
making
process.
The
detection
limit
establishes
only
presence­
absence.
Results
near
the
detection
limit
not
quantitatively
reliable.
The
decision
limit
should
be
significantly
higher
than
the
detection
limit
(e.
g.,
by
a
factor
of
10).

Response:
The
commenter
has
correctly
outlined
the
hypothesis
framework
used
in
the
guidance
document.
In
the
Interim
Guidance,
EPA
used
the
term
"detecting"
on
Page
6
where
more
appropriately
the
term
"quantitating"
(in
reference
to
the
analytical
quantitation
limit)
should
have
been
used.
EPA
has
modified
the
language
in
the
guidance
so
that
use
of
the
term
"detecting"
is
not
confused
with
the
term
"quantitation"
or
"quantitation
limit."

Comment:
The
"non­
statistical"
approach
discussed
in
Section
2.3.1
of
the
document
(page
15)
for
determining
whether
contamination
is
less
than
either
10
x
UTS
or
whether
a
90%
reduction
has
occurred
does
not
appear
to
be
scientifically
defensible.
It
not
valid
to
infer
that
:(
Treated)
<
10
x
UTS
or
0.1
:(
Treated)

(where
:
denoted
the
"true"
concentration)
when
C(
Treated)
<
0.1C(
Untreated)

or
10
C(
Treated)
<
10
x
UTS
unless
the
uncertainties
associated
with
the
measurements
C(
Treated)
and
C(
Untreated)
are
known
or
can
be
estimated.
For
example,
assume
that
for
some
single
metal
contaminant
10
x
UTS
=
100
mg/
kg
and
C(
Treated)
=
95
mg/
kg
Since
instrumental
error
alone
for
the
metal
analysis
is
about
±10%,
the
uncertainty
associated
with
the
single
measurement
C(
Treated)
=
95
mg/
kg
is
at
least
±
10
mg/
kg.
The
measurement
does
not
demonstrate
that
contamination
is
either
above
or
below
10
UTS.
A
single
measurement,
without
some
estimate
of
the
uncertainty
is
meaningless.

In
addition,
according
to
the
guidance,
the
"non­
statistical"
approach
should
be
used
only
when
"the
soil
is
relatively
homogeneous."
What
constitutes
"homogeneous
soil"?
Natural
soils
are
inherently
heterogeneous.
An
operational
definition
for
"homogeneous
soil"
is
not
presented;
that
some
practical
approach
for
evaluating
whether
the
soil
is
"homogeneous"
is
required.
This
would
typically
require
the
analysis
of
multiple
samples,
but
if
multiple
samples
were
taken,
then
advantage
of
using
a
"non­
statistical"
approach
is
unclear.

The
approach
presented
in
2.3.1
appears
to
be
viable
only
if
some
estimate
of
the
total
measurement
uncertainty
exists
and
the
measured
value
is
much
less
than
the
decision
limit
(in
the
context
of
the
magnitude
of
the
estimated
uncertainty).
For
example,
if
it
is
estimated
that
the
total
uncertainty
is
approximately
±
30%
of
the
measured
value,
then
a
result
of
C(
Treated)
=
10
mg/
kg
could
be
considered
to
be
less
than
10
x
UTS
=
100
mg/
kg.
In
general,
unless
the
uncertainty
is
known
or
can
be
estimated,
single
point
comparisons
will
be
invalid
(e.
g.,
unless
the
unless
[sic]
the
measured
results
are
orders
of
magnitude
less
than
the
decision
limit).

Response:
[For
the
purpose
of
responding
to
the
comment,
EPA
assumes,
where
the
commenter
wrote
":(
Treated)
<
10
UTS
or
0.1
:(
Treated)"
that
the
commenter
actually
meant
":(
Treated)
<
10
UTS
or
0.1
:(
Untreated)."]

EPA
agrees
that
there
are
some
limitations
to
the
"nonstatistical"
approach
given
in
the
Interim
Guidance.
These
limitations
are
stated
in
Table
1
(page
14):
the
method
"only
provides
a
`point
estimate'
of
the
constituent
concentration,
does
not
provide
information
about
variability,
and
does
not
quantify
the
uncertainty
associated
with
the
estimate."
11
EPA
agrees
that
it
is
prudent
to
identify
and
quantify
sampling
and
measurement
error
and
to
take
that
information
into
account
when
using
the
data
to
make
a
decision
about
the
status
of
a
waste.
Variability
and
bias
(collectively
known
as
"error")
introduced
in
the
sampling
and
measurement
processes
can
cause
decision
errors
to
be
made.
Waste
generators
should
(but
are
not
required
to
do
so
by
LDR
regulations)
specify
tolerable
limits
for
such
decision
errors
as
part
of
the
DQO
Process.
Tolerable
limits
for
error
should
balance
the
consequences
of
decision
errors
against
the
cost
of
limiting
the
possibility
of
those
errors.
When
the
soil
subject
to
the
treatment
determination
is
(1)
relatively
small
in
volume
(e.
g.,
a
single
drum),
(2)
is
"relatively
homogeneous,"
and
(3)
sampling
and
measurement
error
can
be
minimized,
a
single
representative
sample
may
provide
a
reasonable
point
estimate
of
the
concentration
of
the
constituent
of
concern
in
that
given
volume
of
soil,
and
the
consequences
of
a
decision
error
would
not
be
severe
due
to
the
small
quantity
of
soil
affected.

Finally,
EPA
has
noted
the
comment
that
natural
soils
are
inherently
heterogeneous,
and
that
an
operational
definition
for
"homogeneous
soil"
is
not
presented
in
the
Interim
Guidance.
The
phrase
"relatively
homogeneous"
is
used
in
a
qualitative
manner
to
describe
an
ideal
condition.
In
fact,
soils
are
often
described
in
the
literature
in
terms
of
their
type
and
degree
of
heterogeneity.
EPA
agrees
with
the
commenter
that
a
practical
approach
for
dealing
with
heterogeneous
soils
should
be
given.
As
noted
in
the
footnote
on
page
15
of
the
Interim
Guidance,
the
sampling
error
(which
is
caused
in
part
by
heterogeneity)
can
be
controlled
by
using
optimal
sample
mass,
obtaining
the
correct
shape
and
orientation
of
individual
samples
(known
as
the
sample
"support"),
and
by
using
sampling
devices
and
sub­
sampling
procedures
that
will
minimize
bias.
Several
references
are
provided
in
the
footnote.
In
response
to
the
comment,
EPA
has
removed
the
phase
"relatively
homogeneous"
and
emphasize
that
the
effects
of
heterogeneity
can
be
controlled
by
means
of
certain
sampling
and
sample­
handling
techniques.

Comment:
All
the
assumptions
for
the
Wilcoxon
Rank
Sum
test
should
also
be
presented
in
Section
2.3.3.
It
is
my
understanding
that
this
test
requires
similar
variances
for
the
two
populations
(i.
e.,
the
underlying
distributions
are
required
to
possess
similar
shapes
and
dispersion).
Furthermore,
a
lower
limit
but
not
an
upper
limit
is
specified
for
the
number
of
nondetections.
This
test
will
probably
be
inappropriate
if
a
large
number
of
nondetections
are
being
reported.

Response:
The
assumptions
for
the
Wilcoxon
Rank
Sum
test
include
the
following:
(1)
both
samples
are
random
samples
from
their
respective
populations,
(2)
in
addition
to
independence
within
each
sample,
there
must
be
mutual
independence
between
the
two
samples
(i.
e.,
there
can
not
be
spatial
correlation
between
observations
and
the
samples
must
not
be
"paired"),
and
(3)
the
measurement
scale
is
at
least
ordinal
(i.
e.,
you
can
rank
the
sample
values
from
highest
to
lowest)
(from
Practical
Nonparametric
Statistics,
Third
Addition,
W.
J.
Conover,
1999).
In
addition,
variances
of
the
two
populations
are
assumed
to
be
the
same,
however,
the
rank
sum
test
is
relatively
robust
with
respect
to
violations
of
the
equal
variance
assumption
–
that
is,
the
test
is
approximately
correct
even
when
the
variances
of
the
two
populations
differ
(from
Statistics
in
Plain
English
with
Computer
Applications,
Robert
S.
Schulmam,
1992).
EPA
will
add
these
assumptions
for
the
Wilcoxon
Rank
12
Sum
test
to
Section
2.3.3.

The
upper
limit
of
the
percentage
of
nondetects
is
about
90%.
This
statement
was
added
to
the
guidance
document.
13
Commenter:
Environmental
Technology
Council
(ETC)

Comment:
Section
2:
Determining
if
Soil
Needs
to
be
Treated.
On
page
4
of
the
guidance,
EPA
states
that
the
generator
must
determine
if
the
soil
has
to
be
treated
before
it
can
be
land
disposed.
The
Guidance
states
that
this
determination
can
be
based
on
knowledge
or
laboratory
analysis.
If
knowledge
is
used,
acceptable
knowledge
can
be
based
on
any
of
the
following
sources:

°
Process
knowledge
or
information
on
waste
generated
from
similar
processes;
°
Waste
analysis
data
from
facilities
that
send
wastes
off­
site
for
treatment,
storage
or
disposal;
°
Facility
records
of
analyses
performed
before
the
effective
date
of
RCRA
regulations.

The
ETC
finds
this
language
both
problematic
and
puzzling.
With
regard
to
the
first
two
sources,
we
are
concerned
that
knowledge
or
data
on
waste
generated
from
similar
processes
may
not
be
sufficient
for
contaminated
soil
at
a
given
site.
Such
information
may
be
used
as
a
starting
point;
however,
other
contaminants
may
be
present
based
on
unique
aspects
of
the
process
at
the
given
site.
Therefore,
when
using
such
knowledge,
the
generator
or
operator
should
have
to
reasonably
ascertain,
if
possible,
whether
other
contaminants
may
be
present
in
the
contaminated
soil.

With
regard
to
the
third
source,
why
would
facility
records
be
limited
to
analytical
data
from
before
the
effective
date
of
RCRA
regulations?
If
more
recent
data
is
available,
why
could
that
not
be
used?
We
suspect
EPA
means
that
facility
records,
including
records
of
waste
analyses
performed
before
RCRA,
can
be
used.
To
avoid
confusion,
we
recommend
that
the
Guidance
simply
state
the
credible
facility
records
of
waste
analyses
may
be
relied
upon
to
support
acceptable
knowledge.

If
process
information
is
limited
or
of
questionable
validity,
then
the
Guidance
must
require
that
new
analytical
data
be
obtained.
The
Guidance
needs
to
define
specific
circumstances
when
new
analytical
data
on
contaminated
soil
must
be
obtained.
This
would
include
situations
when
process
information
is
limited,
or
the
site
history
has
involved
various
types
of
owners
and
processes.
Also,
if
analytical
data
on
the
site
is
limited
or
of
questionable
quality,
a
soil
sampling
survey
should
be
performed.

Response:
The
language
to
which
the
commenter
is
referring
was
drawn
from
EPA's
Waste
Analysis
Plan
Guidance
Manual
(OSWER
9938.4­
03,
pages
1­
11
and
1­
12).
The
Agency
agrees
with
the
commenter
in
that
the
language
could
be
taken
out
of
context
to
imply
that
acceptable
knowledge
for
waste
analysis
purposes
allows
for
the
use
of
information
that
is
of
questionable
validity.
EPA
has
revised
this
section
to
be
clearer
about
the
responsibilities
of
all
hazardous
waste
handlers
and
to
be
clearer
about
the
extent
of
the
usefulness
of
acceptable
knowledge.
Further,
the
revised
section
encourages
the
reader
to
reference
the
Waste
Analysis
Plan
Guidance
Manual
for
more
detailed
information.
14
Comment:
Section
2.1:
Data
Quality
Objectives.
The
ETC
is
supportive
of
using
the
DQO
Process
to
develop
the
sampling
strategy.
However,
allowance
for
"other
systematic
planning
processes"
as
stated
on
page
5
invites
potentially
less
effective
sampling
programs.
The
ETC
believes
that
the
DQO
Process
can
be
followed
in
all
cases,
without
imposing
unreasonable
costs.
Response:
The
Data
Quality
Objectives
(DQO)
Process
is
the
Agency's
recommended
systematic
planning
tool
that
is
part
of
EPA's
Quality
System.
EPA
organizations
and
organizations
with
extramural
agreements
with
EPA
must
follow
a
systematic
planning
process
as
required
by
EPA
Order
5360.1
A2;
however,
waste
generators
and
treaters
have
no
regulatory
obligation
under
RCRA
to
use
the
Data
Quality
Objectives
(DQO)
Process
or
any
other
systematic
planning
process
when
complying
with
the
LDR
regulations.
Regulated
entities
are
free
to
use
any
planning
process
they
desire.

Comment:
Under
Step
4
of
the
process
(pages
6
through
8)
the
ETC
has
concerns
that
the
wording
will
not
adequately
protect
against
impermissible
dilution
through
the
mixing
of
different
segments
of
soil
contamination.
The
Guidance
should
be
specific
about
defining
hot
spots
through
analysis,
and
prohibiting
mixing
such
hot
spots
with
soil
containing
low
levels
of
contamination.
To
do
otherwise
will
be
encouraging
impermissible
dilution.

Response:
The
Agency's
position
on
impermissible
dilution
and
mixing
of
contaminated
is
not
changed
by
the
guidance
document.
The
guidance
document
simply
reiterates
the
Agency's
position
on
these
issues
given
previously
in
the
Federal
Register.
See
63
FR
28605
and
28621
(May
26,
1998),
51
FR
40592
(November
7,
1986),
and
53,
FR
30911
(August
16,
1988).

Regarding
the
identification
of
"hot
spots,"
the
guidance
provides
several
specific
references
that
provide
detailed
procedures
for
the
identification
of
hot
spots
using
statistical
methods.

Comment:
Also
on
page
7,
EPA
states
that
"Generally,
subject
to
limited
exceptions,
you
should
not
mix
hazardous
soil
with
nonhazardous
soils".
The
ETC
disagrees
with
the
use
of
the
term
"generally",
and
mixing
of
hazardous
with
nonhazardous
soil
should
always
be
viewed
as
impermissible
dilution.
On
page
8,
EPA
states
that
mixing
such
soil
may
be
needed
to
adjust
BTU
or
water
content.
Yet
the
BTU
value
of
various
soils
does
not
change
significantly
to
affect
a
given
treatment
process,
and
water
can
always
be
addressed
easily
in
other
ways.
Given
the
broad
flexibility
already
provided
by
the
90%
reduction
or
10xUTS
standard,
there
is
no
need
to
risk
promoting
impermissible
dilution
by
allowing
the
mixing
of
nonhazardous
with
hazardous
soils,
based
on
justifications
that
are
both
weak
and
avoidable.

Response:
In
response
to
the
comment,
EPA
has
removed
the
phrase
"Generally."
EPA
does
not
agree
with
the
comment
that
mixing
of
hazardous
soil
with
nonhazardous
soil
"should
always
be
viewed
as
impermissible
dilution."
As
stated
in
the
preamble
to
the
Land
Disposal
Restrictions
Phase
IV:
Final
Rule
(May
26,
1998),
the
"Agency
notes
that
the
normal
mixing
of
contaminated
soil
from
various
portions
of
a
site
that
typically
occurs
during
the
course
of
remedial
activities
or
in
the
course
of
normal
earthmoving
and
grading
activities
is
not
considered
intentional
mixing
of
soil
with
non­
media
or
prohibited
soil
with
non­
prohibited
soil
and,
therefore,
is
not
a
type
of
impermissible
dilution."
15
Comment:
On
the
bottom
of
page
8,
EPA
states
that
when
using
the
90%
reduction
criteria,
that
an
initial
study
could
be
done
to
demonstrate
the
90%
reduction,
and
then
other
process
variables,
controls
and
operating
conditions
can
be
used
along
with
knowledge
to
support
the
90%
standard
in
lieu
of
testing.
The
ETC
disagrees
with
the
allowance
for
this
alternative.
There
is
nothing
said
in
the
Guidance
about
the
degree
of
trial
treatment
required
to
demonstrate
the
correlation
between
90%
reduction
and
process
variables.
For
example,
how
many
trial
runs
are
required
to
demonstrate
the
correlation
between
a
given
process
condition
and
90%
reduction?
How
representative
does
the
waste
need
to
be
during
the
initial
study
correlating
process
conditions
and
percent
reduction?
In
order
to
ensure
effective
treatment,
the
initial
study
should
be
based
on
a
worse
case
contaminated
soil,
which
is
spiked
with
the
hazardous
constituents
at
levels
above
the
maximum
level
known
to
exist.
But
what
if
a
hot
spot
is
encountered
that
exceeds
the
concentration
demonstrated
during
the
initial
study?
There
are
too
many
variables
and
sources
of
uncertainty
to
allow
use
of
process
variables
to
demonstrate
90%
reduction
in
lieu
of
batch
testing.

Instead,
EPA
should
still
require
batch­
by­
batch
testing,
but
provide
flexibility
by
allowing
use
of
surrogate
indicator
constituents.
For
example,
if
a
petroleum
contaminated
site
is
under
remediation,
the
batch­
by­
batch
testing
can
be
limited
to
benzene
or
TPH.
If
chlorinated
contaminated
soil
is
being
treated,
testing
for
total
halogens
or
for
the
highest
chlorinated
constituent
can
be
performed.
This
would
reduce
cost
while
ensuring
that
each
batch
achieves
the
90%
reduction
level.
Then
on
a
periodic
basis,
such
as
monthly,
a
sample
can
be
verified
for
all
of
the
constituents.

Response:
EPA
recognizes
that
there
can
be
many
site­
specific
variables
that
influence
the
frequency
of
testing,
thus
EPA
cannot
specify
in
guidance
documents
or
regulations
rigid
sampling
and
testing
requirements.
Instead,
a
hazardous
waste
treater
must
test
their
waste
according
to
the
frequency
specified
in
the
their
facility­
specific
Waste
Analysis
Plan
(WAP)
as
required
by
40
CFR
264.13
(for
permitted
facilities)
or
40
CFR
265.13
(for
interim
status
facilities).

Comment:
Section
2.3:
Methods
for
Determining
Attainment
of
90%
Reduction
or
10xUTS.
The
non­
statistical
approach
defined
in
Section
2.3.1
allows
for
collection
of
a
single
soil
sample
with
the
critical
assumptions
that
the
soil
is
homogenous
and
that
the
volume
of
soil
is
"relatively
small".
These
criteria
are
very
subjective
and
weak.
The
Guidance
should
be
more
specific
in
setting
a
maximum
volume
above
which
the
non­
statistical
approach
cannot
be
used.
In
addition,
the
non­
statistical
approach
could
require
collection
of
a
certain
number
of
duplicate
samples
(such
as
every
5th
batch)
to
verify
that
the
single
sample
non­
statistical
approach
is
valid
and
that
the
soil
is
homogenous
and
the
treatment
is
consistent.

Response:
Through
application
of
the
DQO
Process,
EPA
encourages
–
but
does
not
require
by
regulation
–
waste
generators
and
treaters
to
strike
a
balance
between
the
costs
of
limiting
the
possibility
of
making
a
waste
classification
error
(or
"decision
error")
and
the
cost
of
sampling
and
analysis
by
considering
the
consequences
of
a
decision
error.
If
a
soil
subject
to
the
treatment
determination
is
(1)
relatively
small
in
volume
(e.
g.,
a
single
drum)
and
(2)
sampling
and
measurement
16
error
can
be
minimized,
then
a
single
representative
sample
may
provide
a
reasonable
point
estimate
of
the
concentration
of
the
constituent
of
concern
in
that
given
volume
of
soil,
and
the
consequences
of
a
decision
error
would
not
be
severe
due
to
the
small
quantity
of
soil
affected.
Several
examples
of
what
constitutes
a
"small
volume"
are
given
in
the
guidance;
however,
waste
generators
and
treaters
have
an
obligation
to
ensure
that
all
portions
of
the
waste
meet
the
treatment
standard.
Otherwise,
there
is
evidence
that
the
treatment
was
not
effective
and
there
is
noncompliance
with
the
LDR
treatment
standard.

EPA
agrees
that
the
term
"homogeneous"
is
subjective,
as
used
in
the
Interim
Guidance
document.
EPA
removed
the
phrase
"relatively
homogeneous"
and
emphasizes
that
the
effects
of
heterogeneity
can
be
controlled
(i.
e.,
sampling
error
can
be
minimized)
by
means
of
certain
sampling
techniques
referenced
in
footnote
7
at
the
bottom
of
page
15.

Comment:
In
Section
2.3.2
EPA
discusses
methods
that
can
be
used
to
determine
attainment
of
the
UTS
or
10xUTS
standard.
The
ETC
supports
use
of
a
99%
Upper
Confidence
Limit
approach
to
ensuring
compliance
with
the
UTS
or
10xUTS
standard.
However,
such
a
statistical
evaluation
should
only
be
necessary
if
the
levels
measured
in
the
soil
exceed
50%
of
the
compliance
threshold.
In
general,
if
a
single
test
shows
that
the
treated
soil
is
less
than
50%
of
the
compliance
threshold,
then
no
further
testing
should
be
necessary.
However,
if
the
result
exceeds
50%
of
the
threshold,
then
at
least
3
additional
samples
should
be
taken
and
an
upper
confidence
limit
calculated.
Then
the
confidence
limit
can
be
evaluated
using
the
two
bulleted
criteria
at
the
top
of
page
18.
In
the
case
of
the
second
bullet,
where
the
four
samples
comply
but
the
upper
confidence
limit
exceeds
the
standard,
it
would
be
prudent
to
collect
another
set
of
4
samples
to
verify
that
the
batch
is
in
compliance.
The
results
for
these
4
could
be
combined
with
the
first
4
to
obtain
a
better
standard
deviation
and
thus
a
more
accurate
calculation
of
the
upper
confidence
limit.
As
a
result,
it
may
be
found
that
the
upper
confidence
limit
drops
below
the
compliance
threshold.

Response:
Waste
generators
and
treaters
have
no
obligation
under
the
LDR
regulations
to
conduct
statistical
sampling
or
to
apply
statistical
methods
to
evaluate
attainment
of
the
LDR
treatment
standards.
If
a
generator
or
treater
elects
to
use
statistical
sampling
to
estimate
the
uncertainty
associated
with
a
waste
classification
decision,
then
use
of
an
upper
confidence
limit
on
an
upper
percentile
is
one
data
assessment
tool
that
could
be
used.
(Note
that
a
nonparametric
version
of
the
test
does
not
require
knowledge
of
the
standard
deviation).
Other
methods
such
as
a
test
of
proportions
or
"acceptance"
sampling
methods
also
could
be
used,
though
none
of
these
methods
are
required
by
regulation.
As
noted
previously,
a
hazardous
waste
treater
must
test
their
waste
according
to
the
frequency
specified
in
the
their
facility­
specific
Waste
Analysis
Plan
(WAP)
as
required
by
40
CFR
264.13
(for
permitted
facilities)
or
40
CFR
265.13
(for
interim
status
facilities).

Comment:
In
Section
2.3.3
EPA
describes
various
statistical
methods
that
can
be
used
to
determine
attainment
of
the
90%
reduction
standard.
The
ETC
is
in
agreement
with
the
various
"two­
sample"
statistical
tests
allowed
to
demonstrate
compliance
with
90%
reduction.
The
ETC
also
supports
the
17
two­
tiered
approach,
starting
with
the
Non­
Parametric
test
and
progressing
to
the
second
tier
using
either
the
Welch's
t­
Test
or
the
Wilcoxon
Rank
Sum.
The
only
improvement
warranted
to
this
section
is
to
require
for
all
tiers
that
a
minimum
of
4
replicates
be
obtained
of
soil
before
treatment
and
after
treatment.
This
would
provide
greater
confidence
that
the
90%
reduction
standard
has
been
achieved,
no
matter
what
statistical
approach
is
taken.
All
three
statistical
approaches
are
dependent
on
an
accurate
standard
deviation,
and
a
minimum
of
three
samples
are
needed
to
establish
a
reasonably
accurate
standard
deviation.

Response:
EPA
assumes
the
term
"replicates"
used
by
the
commenter
refers
to
individual
field
samples
and
not
quality
control
samples
sometimes
referred
to
as
"replicates"
or
"duplicates."
The
number
of
samples
used
in
the
various
two­
sample
tests
influences
the
statistical
"power"
of
the
test.
The
power
of
a
statistical
test
is
the
probability
of
rejecting
the
null
hypothesis
when
it
is
false
(a
correct
decision).
In
the
context
of
a
90­
percent
reduction
determination,
power
is
the
probability
of
correctly
deciding
the
standard
has
been
attained.
This
measure
differs
from
"confidence"
(1
­
"),
where
"
is
the
"level
of
significance"
or
the
probability
of
rejecting
the
null
hypothesis
when
it
is
true.
The
value
of
",
and
hence
(1
­
")
is
set
in
advance
by
the
planning
team.
Setting
the
number
of
samples
to
some
minimum
value
would
control
"power"
but
would
not
"provide
greater
confidence"
that
the
90%
reduction
standard
has
been
achieved.

Aside
from
the
technical
definition
of
"confidence,"
the
ability
to
statistically
conclude
that
90%
reduction
has
been
attained
(when
in
fact
it
has)
is
indeed
improved
by
increasing
the
number
of
samples
representing
each
population.

Note
also
that
only
Welch's
t­
Test
relies
on
estimates
of
the
standard
deviation.
The
other
two­
sample
tests
(the
nonparametric
test
of
location,
and
the
Wilcoxon
Rank
Sum
test)
do
not
require
estimates
of
the
standard
deviation.
However,
a
minimum
of
three
to
five
samples
(per
population)
typically
is
needed
to
perform
any
of
the
two­
sample
tests
given
in
the
guidance
document.

Comment:
Section
3.
Notification,
Certification
and
Recordkeeping
Requirements.
The
recordkeeping
requirements
need
to
be
more
emphatically
stated,
eliminating
the
word
"should"
and
replacing
it
with
"must."
This
is
consistent
with
the
language
in
40
CFR
§
268.7.
The
facility
or
generator
must
have
documentation
to
support
that
the
alternative
treatment
standards
were
achieved
including
the
list
of
constituents
subject
to
treatment
and
how
this
was
determined.
The
documentation
at
the
bottom
of
page
27
lists
data
associated
with
the
90%
reduction
standard,
but
does
not
list
any
data
for
the
10xUTS
alternative.
All
statistical
calculations
must
also
be
maintained
to
document
the
degree
of
confidence
in
achieving
the
alternative
soil
treatment
standards.
Finally,
in
cases
where
there
is
no
central
file
applicable
to
a
given
generator
or
facility
location,
then
this
data
must
be
submitted
to
the
Agency
as
part
of
a
Closure,
Remedial
Action,
Corrective
Action
or
CERCLA
response
report.
The
appropriate
treatment
certification
under
40
CFR
§
268.7(
b)(
5)
must
also
be
maintained
on
file
and
included
with
the
Report
to
the
Agency.
18
Response:
EPA
agrees
that
treatment
data
used
to
verify
compliance
with
either
the
standard
of
10
x
UTS
or
90­
percent
reduction
should
be
listed.
EPA
has
modified
the
guidance
accordingly.
