Subject:
Investigation
of
Consequences
of
FRS
Conversion
for
RCRA
ID
Information
From:
Ben
Smith
To:
Docket
for
Forms
Modification
Rule
In
response
to
OSW
concerns
with
the
accuracy
of
FRS
information,
I
conducted
an
independent
investigation
of
the
spreadsheets
developed
by
Bill
Kline.
Bill
had
conducted
a
rather
thorough
analysis
which
is
detailed
in
five
spreadsheets
(
attached).
Because
Bill's
analysis
was
not
based
on
a
random
subset,
I
then
used
an
analysis
by
Tim
Antisdel
to
attempt
to
extrapolate
the
results
to
the
TRI
universe.

Speaking
first
in
terms
of
the
Kline
analysis,
Tab
A
merely
displays
a
set
of
10139
TRI
Facility
IDs
(
TRIFD)
for
which
he
requested
all
associated
RCRA
IDs
out
of
FRS.
These
RCRA
IDs
came
from
RCRA
Info.
The
study
was
undertaken
with
2002
Reporting
Year
information
because
RY
2003
data
were
not
yet
released.

In
Tab
D
are
listed
the
1530
TRIFD
for
which
there
was
no
corresponding
RCRA
ID
in
Form
R
nor
FRS.
Tab
B
identifies
the
8609
TRIFD
which
had
at
least
one
RCRA
ID.
Tab
C
presents
10190
RCRA
ID
found
to
be
associated
with
those
TRIFD.
Tab
E
was
an
examination
of
the
1432
TRIFD
for
which
no
FRS
ID
was
reported.
Additional
analysis
was
conducted
on
2021
of
the
Tab
C
IDs
to
determine
the
nature
of
discrepancies
between
FRS
and
TRI
Form
R
information.
Tab
E
contains
an
analysis
of
501
of
the
facilities
without
RCRA
ID
in
FRS.
130
of
these
had
a
RCRA
ID
reported
on
a
Form
R.

Principal
findings
are
as
follows:

From
Tab
C:
1628
(
81%)
had
no
ID
related
issues
241
(
12%)
had
RCRA
ID
not
reported
on
the
Form
R
152(
7.5%)
had
discrepancies
in
TRI
and
RCRA
Info
ID.
These
could
be
further
sub­
evaluated
to
determine
that
9
of
the
TRI
forms
had
clearly
invalid
RCRA
ID,
2
RCRA
Info
ID
were
clearly
invalid,
57
had
RCRA
ID
on
Form
R
that
differed
from
RCRA
Info
reported
ID
(
although
neither
was
necessarily
invalid),
and
84
had
ID
differences
which
could
have
been
a
consequence
of
changes
in
ownership
(
FRS
had
the
more
current
ID).

From
this
it
was
concluded
that
no
more
than
59
out
of
the
2021
ID
investigated
could
potentially
have
RCRA
Info
related
problems,
but
334
TRI
reported
RCRA
ID
had
problems.
Clearly
FRS
was
a
superior
source.

From
Tab
E:

23
of
the
TRI
reported
ID
were
invalid
107
of
the
ID
could
not
be
confirmed
in
FRS
(
could
be
invalid
or
expired
or
small
quantity
generator
related).

Consequently,
from
this
analysis,
one
would
infer
that
there
are
a
maximum
of
59(
10190/
2021)+
107(
1432/
501)=
603
maximum
potential
problematic
ID
in
the
facility
pool,
and
334(
10190/
2021)+
23(
1432/
501)=
1750
potentially
problematic
RCRA
ID
in
TRI
information.

A
subsequent
attempt
was
made
to
evaluate
this
number
using
the
Antisdel
analysis.
In
his
analysis
(
attached),
he
examined
RCRA
ID
reporting
by
all
TRI
facilities
in
RY
2003.
There
were
19106
RCRA
ID
reported.
1770
of
the
RCRA
ID
were
not
in
FRS.
If
the
same
proportion
of
these
ID
(
107/
130)
are
FRS
rather
than
TRI
issues
as
was
seen
in
the
Kline
study,
we
would
project
1457
to
be
due
to
non
TRI
related
issues.
They
could,
for
instance,
be
ID
for
small
quantity
generators
not
required
to
report
to
RCRA
Info.
When
this
number
is
scaled
down
to
reflect
the
smaller
number
of
facilities
in
the
Kline
survey,
one
projects
(
1457*
10190/
19106)
or
776
potential
issues.
To
this
estimate,
one
would
add
the
number
of
RCRA
ID
in
the
Kline
survey
with
potential
FRS
use
related
issues
(
59*
10190/
2021),
or
297
ID.
Together,
the
estimated
level
of
potential
issues
is
1073
ID.

Since,
it
is
reasonable
to
assume
conditions
could
change
from
year
to
year,
an
average
of
these
two
estimates
is
applied
to
come
up
with
a
best
estimate.
That
value
is
838
potential
issue
related
ID.

Numerous
caveats
must
be
applied
to
this
estimate.

­
A
number
of
the
ID
discrepancies
appear
to
be
issues
related
to
transcription
errors
(
ie
reversing
digits
or
entering
0
for
o.

­
Small
quantity
generator
ID
are
known
to
be
entered
by
some
respondents
but
are
not
kept
in
RCRA
Info.
This
would
account
for
some
portion
of
the
discrepancies.

­
Response
space
for
ID
entry
is
limited
on
the
TRI
paper
form.
Many
discrepancies
could
result
from
selective
entry
of
RCRA
ID
on
the
form.
