ENVIRONMENTAL
MODELING
for
HALOHYDANTOINS
PDM4
MODEL
Page
2
of
10
INTRODUCTION
This
document
provides
estimated
environmental
concentrations
(
EECs)
in
the
receiving
body
of
water
from
the
use
of
hydantoins
in
once­
through
industrial
water
systems.
The
Probabilistic
Distribution
Model
version
4
(
PDM4),
part
of
the
new
EFAST
model
that
Versar
is
currently
developing,
to
be
publically
released
in
the
future,
was
used
to
evaluate
the
hydantoin
concentration
in
stream
water.
The
probabilistic
dilution
algorithm
remains
the
same
as
that
used
in
EFAST
v2.0,
but
the
underlying
dataset
of
facilities/
reaches
for
this
version
is
more
accurate.
Using
a
steam
electric
power
plant
as
the
representative
facility,
the
number
of
days
that
the
concentration
in
the
water
body
exceeded
threshold
levels
was
determined.

Overview
Of
PDM4
PDM4
calculates
downstream
chemical
concentrations
from
a
chemical
discharge
using
the
probabilistic
distribution
model.
The
algorithm
used
is
taken
from
DiToro,
1984.
A
simple
mass
balance
approach
forms
the
basis
of
PDM;
however,
the
input
variables
are
not
single
point
estimates,
because,
in
reality,
these
variables
are
not
constant.
Streams
follow
a
highly
variable
seasonal
flow
pattern
and
numerous
variables
in
a
manufacturing
process
can
affect
the
chemical
concentration
and
flow
rate
of
the
effluent.
PDM4
models
the
stream
flow
as
constantly
changing
and
calculates
the
probability
that
the
concentration
in
given
target
stream
will
exceed
some
target
value.
Specifically,
it
calculates
the
percent
of
days
per
year
that
the
concentration
of
concern
is
exceeded.
PDM4
uses
probability
distributions
as
inputs
and
calculates
the
resulting
probability
distribution
of
the
concentration
in
the
stream.

INPUTS
Power
Plant
Information
In
modeling
the
fate
of
Hydantoins
in
the
receiving
body
of
water,
the
following
assumptions
were
made
regarding
the
representative
power
plant
to
be
modeled:

°
The
7Q10
flow
rate
(
a
flow
rate
that,
once
every
ten
years,
a
stream
is
expected
to
be
below
for
seven
consecutive
days)
was
assumed
to
be
the
normal
cooling
water
flow
rate.
This
value
was
chosen
because
it
is
assumed
that
electric
plants
would
need
to
have
a
steady
supply
of
cooling
water,
and
the
7Q10
flow
reflects
a
rate
that
could
be
maintained
continuously
by
the
power
plant.
This
may
is
not
a
conservative
assumption,
since
electric
plants
may
use
more
cooling
water
under
normal
conditions,
though
at
a
greater
risk
of
running
out
of
usable
water.
For
lack
of
better
data,
these
values
were
used.
°
It
is
assumed
that
the
system
is
running
without
any
noticeable
fouling,
so
that
only
maintenance
applications
of
the
chemical
are
necessary.
°
The
application
rate
of
the
chemical
is
the
maximum
rate
listed
on
the
label
for
use
in
maintaining
once­
through
cooling
water
systems.
The
application
rate
is
0.3
lbs
product/
1000
gallons
of
water
contained
in
the
system.
Since
the
product
is
98%
a.
i.,
the
application
rate
can
be
expressed
as
0.294
lbs
a.
i./
1000
gallons.
(
The
label
also
notes
that,
Page
3
of
10
for
systems
that
are
noticeably
fouled,
up
to
0.6
lbs
product/
1000
gallons
should
be
used
as
an
initial
dose,
to
be
followed
later
by
the
regular
maintenance
dose.
This
initial
dose
application
rate
was
not
modeled).
°
The
label
specifies
that
the
application
of
the
chemical
should
be
repeated
as
needed
"
to
maintain
one
to
three
ppm
chlorine
residual
for
at
least
4
hours."
For
this
model,
it
is
assumed
that
one
application
per
day
is
sufficient
to
attain
this
level.

PDM4
Model
Inputs
A
summary
of
the
input
parameters
is
described
below.
To
see
the
complete
input
and
output
files,
please
refer
to
the
appendix
of
this
report
°
Release
Days
­
Dang
(
1996)
has
produced
a
document
that
includes
estimates
for
typical
operating
conditions
of
once­
through
industrial
cooling
systems.
According
to
Dang
(
1996),
it
is
expected
that
the
product
will
be
applied
between
100
and
250
days
a
year,
though
there
may
be
seasonal
differences
in
the
rate
of
use.
Although
typical
applications
are
applied
within
the
time
frames
above,
PDM4
runs
the
model
on
the
basis
of
365
days
or
a
full
year.

°
Pretreatment
Release
­
The
amount
released
per
day
can
be
calculated
by
multiplying
the
application
rate
(
0.294
lbs
a.
i./
1000
gallons)
by
the
amount
of
water
contained
in
the
system.
This
volume
(
the
water
contained
between
the
inlet
and
the
outlet)
is
difficult
to
determine,
since
power
plants
will
vary
greatly.
In
general,
it
is
expected
that
there
will
be
pipes
leading
from
the
water
body
to
the
heat
exchanger,
that
the
heat
exchanger
must
be
of
sufficient
volume
to
adequately
condense
the
boiler
steam,
and
that
there
will
be
pipes
leading
from
the
heat
exchanger
back
to
the
water
body.
The
variables
that
are
difficult
to
determine
include
the
number
of
heat
exchangers
and
inlet/
outlet
pipes
in
the
system,
the
existence
of
reservoirs
and
sumps,
the
types
of
heat
exchangers
that
are
used,
and
the
distance
between
the
heat
exchangers
and
the
water
body.
A
search
on
the
Internet
revealed
no
estimates
of
the
volume
in
question,
nor
did
it
reveal
typical
dimensions
for
either
the
inlet
pipes,
the
heat
exchangers,
or
any
of
the
variables
listed
above.
Therefore,
a
different
approach,
estimating
the
volume
by
using
retention
times,
was
used.
A
study
was
performed
in
which
a
conventional
cooling
water
system
was
used
as
a
biofilm
reactor
for
wastewater
(
Cloete
et
al,
1999).
In
this
study,
the
cooling
system
was
operated
at
different
wastewater
flow
regimes.
The
hydraulic
retention
times
used
in
this
model
varied
between
8
minutes
20
seconds
and
26
minutes
2
seconds.
Virginia's
Dominion
Power
Company
was
contacted
and
estimates
on
retention
times
of
8
to
12
minutes
were
also
obtained.
The
range
was
based
on
the
number
of
units
in
operation
and
stream
flow
conditions
at
time
of
operation.
Data
from
Dominion
Power
Company,
though
anecdotal,
were
chosen
over
the
data
from
Cloete
et
al,
1999,
since
Cloete
et
al
describes
retention
times
for
much
smaller
scale
cooling
systems.
Loads
were
calculated
assuming
a
retention
time
of
12
minutes,
to
be
conservative.
The
volumes
of
the
cooling
water
systems
were
calculated
by
multiplying
the
retention
time
by
the
flow
rate
of
cooling
water.
For
Page
4
of
10
example,
for
a
water
system
with
a
flow
rate
of
725
MGD,
the
calculations
yield
a
volume
of
6.04
million
gallons
for
the
cooling
water
system.
Multiplying
by
the
application
rate
of
0.294
lbs
a.
i.
per
1000
gallons
of
water
leads
to
a
final
input
load
of
807
kg/
day.

°
Concentration
of
Concern
­
PDM4
is
designed
to
report
the
number
of
days
that
the
concentration
in
the
water
body
exceeds
a
concentration
of
concern
that
has
been
specified
by
the
user.
For
this
study,
the
model
was
run
multiple
times
using
different
concentrations
of
concern
to
determine
a
curve.
The
concentration
of
concern
for
each
flow
regime
was
as
follows:
0.5,
0.025,
0.005,
0.025,
0.050,
0.25,
0.50,
2.5,
5.0,
25,
50,
250,
500,
2500,
and
5000
ppb.

The
input
parameters
are
summarized
in
Table
1
below.

Table
1.
PDM4
Model
Inputs
Parameter
Value
Rationale
Industry
Type
Steam
Electric
Power
Plants
(
SIC
#
4911)
Various
NPDES
EPA
Assumption
as
being
the
representative
facility
for
once­
through
industrial
water
systems
using
hydantoins.

Release
Days
250,
365
Dang,
1996,
and
PDM4
Model.

Pretreatment
Release
807
kg/
day
for
725
MGD
cooling
water
system.
Lower
values
used
for
smaller
systems.
Based
on
an
12
minute
hydraulic
residence
time
and
the
maximum
application
rate
listed
on
the
label.

Concentration
of
Concern
0.5
ppb
to
5000
ppb
A
range
was
used
to
determine
a
curve.

The
three
different
flow
regimes
consisted
of
power
plants
with
stream
flow
rates
of
100
±
10
MGD
(
million
gallons
per
day),
500
±
50
MGD,
and
1000
±
50
MGD.
The
low,
medium,
and
high
stream
flow
rates
correspond
to
378.5
±
37.85,
1982.5
±
189.25,
and
3785
±
189.25
million
liters
per
day.
These
plants
were
pulled
from
a
database
of
NPDES
plant
codes
based
on
the
above
criterion.
Tables
2,
3,
and
4
below
show
details
regarding
the
power
plants
and
their
cooling
streams.
Page
5
of
10
Table
2.
Low
Flow
(
100
MGD)
Stream
Specifications
NPDES
Mean
Stream
Flow
(
MLD)
Mean7Q101
Stream
Flow
(
MLD)
Mean
Effluent
Flow
(
MLD)
Pretreatment
Release
(
kg/
day)
IA0033235
401.08
2.84
0.11
0.833
PA0002062
391.05
4.09
51.97
1.20
LA0003042
383.57
44.57
2.56
13.1
MI0038172
379.25
13.97
1980.49
4.10
OK0002682
363.06
3.37
4.66
1.00
WV0005525
358.67
14.56
4212.65
4.28
IL0036919
355.83
3.89
2407.06
1.14
LA0036145
354.06
13.56
0.23
3.99
UT0000116
351.97
32.39
0.59
9.52
TX0054500
351.57
26.85
9.08
7.89
PA0008443
336.52
12.64
28.94
3.71
IL0048321
336.36
97.05
46.38
28.5
1.
Seven
consecutive
days
of
lowest
stream
flows
over
a
ten
year
period
Table
3.
Medium
Flow
(
500
MGD)
Stream
Specifications
NPDES
Mean
Stream
Flow
(
MLD)
Mean7Q101
Stream
Flow
(
MLD)
Mean
Effluent
Flow
(
MLD)
Pretreatment
Release
(
kg/
day)
MA0004367
2029.15
226.776
21.259
66.6
IA0000108
1973.2
61.301
1.737
18.0
NM0000108
1970.38
5.924
4.353
1.74
FL0025526
1959.16
653.078
443.652
192
IN0032948
1948.06
152.814
0.053
44.9
TX0001163
1867.21
11.373
1911.43
3.34
OH0010421
1840.04
59.802
34.878
17.6
IN0041246
1819.99
167.187
90.014
49.1
PA0002054
1792.41
114.968
711.913
33.8
MN0000906
1749.68
28.936
3.693
8.50
NH0001431
1733.51
75.685
30.28
22.2
IN0038806
1715.15
58.192
0.32
17.1
1.
Seven
consecutive
days
of
lowest
stream
flows
over
a
ten
year
period
Table
4.
High
Flow
(
1000
MGD)
Stream
Specifications
NPDES
Mean
Stream
Flow
(
MLD)
Mean7Q101
Stream
Flow
(
MLD)
Mean
Effluent
Flow
(
MLD)
Pretreatment
Release
(
kg/
day)
GA0004341
3964.96
837.511
0.545
246
WA0003280
3831.92
591.866
129.447
174
KS0079057
3757.2
20.42
2452.56
6.00
NC0005088
3693.64
812.478
165.318
239
SC0001104
3652.9
50.251
3.687
14.7
IL0002186
3635.95
1169.28
1557.46
344
1.
Seven
consecutive
days
of
lowest
stream
flows
over
a
ten
year
period
Page
6
of
10
Twelve
different
sites
throughout
the
United
States
were
tested
in
the
model
from
the
low
and
medium
stream
regimes,
and
six
were
tested
from
the
high
flow
regime.

RESULTS
PDM4
calculates,
for
each
reach,
the
percent
of
days
per
year
that
a
reach
would
have
concentrations
above
a
particular
concentration
of
concern.
These
values
were
averaged
by
Versar
among
the
three
different
flow
regimes,
giving
average
percentages
of
exceedence.
Versar
also
looked
at
the
worst
case
scenarios
for
the
low,
middle,
and
high
stream
flows
considered.

Results
are
presented
in
Tables
5
and
6.
Taking
these
percentages
and
multiplying
by
365
days
a
year,
one
can
calculate
the
number
of
days
out
of
the
year
the
hydantoin
ion
concentration
in
the
water
will
be
greater
than
any
given
COC.
For
all
stream­
flow
categories,
the
average
exceedance
rate
is
less
than
one
day
per
ten
years
(
0.027%)
if
the
concentration
of
concern
is
2500
ppb
or
above.
The
relationship
between
the
concentration
of
concern
and
the
exceedance
rates
appear
to
be
very
similar
for
all
three
steam
sizes
considered.
Standard
deviations
are
presented
in
Table
5,
but
should
be
interpreted
with
care.
Coefficients
of
variation
(
standard
deviation
divided
by
average)
are
small
for
small
COCs
because
the
exceedance
rates
are
very
near
100%
in
all
cases.
At
larger
COCs,
the
coefficients
of
variation
become
larger,
as
differences
in
stream
flows
become
more
pronounced.

Table
5.
Average
Percent
of
Days
COC
Exceeded
COC1
(
ppb)
LF2
Percent
Days
COC
Exceeded
(%)
LFStandard
Deviation
(%)
MF3­
Percent
Days
COC
Exceeded
(%)
MFStandard
Deviation
(%)
HF4­
Percent
Days
COC
Exceeded
(%)
HFStandard
Deviation
(%)
0.5
82.9
30.1
96.1
7.93
97.4
4.80
2.5
65.1
40.4
83.1
26.1
82.0
33.6
5
56.9
39.6
74.6
28.7
76.8
39.3
25
31.1
29.7
41.3
24.7
57.8
37.0
50
19.1
21.3
25.4
19.4
40.8
28.0
250
2.49
3.52
3.33
2.99
5.52
4.44
500
0.663
1.05
0.875
0.929
1.31
1.31
2500
0.0118
0.0253
0.0188
0.0375
0.0119
0.0190
5000
1.24x10­
3
2.75x10­
3
2.56x10­
3
5.56x10­
3
8.29x10­
4
1.54x10­
3
1.
COC
=
Concentration
of
Concern
2.
LF
=
Low
Flow
Regime
(
100
MGD)
3.
MF
=
Medium
Flow
Regime
(
500
MGD)
4.
HF
=
High
Flow
Regime
(
1000
MGD)

Figure
1
shows
the
same
data
as
Table
1.
As
expected,
the
general
trend
in
all
three
flow
regimes
is
that
the
percent
of
days
the
COC
is
exceeded
tends
to
decrease
faster
as
the
COC
is
increased.
Figure
1
also
shows
clearly
the
similarities
between
the
three
steam
flow
size
categories
modeled.
Page
7
of
10
Figure
1.
Average
Exceedance
Rates
0.0001
0.001
0.01
0.1
1
10
100
0.1
1
10
100
1000
10000
Concentration
of
Concern
(
ppb)
Percent
of
Days
Exceeded
per
Year
(%)

Low
(
100
MGD)

Middle
(
500
MGD)

High
(
1000
MGD)

Table
6
lists
the
calculations
for
those
power
plants
with
the
highest
exceedance
rates
for
each
of
the
three
stream
flow
sizes
considered.
If
the
COC
is
2500
ppb,
the
worst­
case
power
plant
considered
in
this
model
(
see
MF
column
in
Table
7)
exceeds
the
COC
about
once
every
two
years.

Table
6.
Worst
Case
Scenarios:
Percent
of
Days
COC
Exceeded
COC1
(
ppb)
LF2­
Percent
Days
COC
Exceeded
(%)
MF3­
Percent
Days
COC
Exceeded
(%)
HF4­
Percent
Days
COC
Exceeded
(%)
0.5
100
100
100
2.5
100
100
100
5
99.8
99.9
100
25
87.7
91.2
91.4
50
65.8
70.6
68.2
250
9.78
8.97
11.3
500
3.33
3.21
3.39
2500
0.0756
0.126
0.0494
5000
7.22x10­
3
0.0186
3.93x10­
3
1.
COC
=
Concentration
of
Concern
2.
LF
=
Low
Flow
Regime
(
100
MGD)
3.
MF
=
Medium
Flow
Regime
(
500
MGD)
4.
HF
=
High
Flow
Regime
(
1000
MGD)
Page
8
of
10
Table
7
shows
the
average
of
all
power
plants
modeled,
and
compares
these
values
to
the
average
of
the
three
worst­
case
scenarios
that
were
presented
in
Table
6.
At
high
COCs,
the
difference
between
the
average
and
worst­
case
scenarios
is
about
one
order
of
magnitude.
This
table
can
be
used
to
determine
an
appropriate
COC,
based
on
the
level
of
risk
that
is
deemed
acceptable.
The
choice
between
using
either
the
total
average
column
or
the
worst­
case
scenarios
average
column
will
depend
on
how
conservative
a
value
is
deemed
necessary.

Table
7.
Total
Average
Exceedance
Rates
vs.
Worst­
Case
Exceedance
Rates
COC
(
ppb)
Percent
of
Days
COC
Exceeded
(%)

Total
Average
Average
of
the
Three
Worst
Case
Scenarios
(
Table
6)

0.5
91.1
100
2.5
75.7
100
5
68.0
99.9
25
40.5
90.1
50
26.0
68.2
250
3.43
9.90
500
0.876
3.31
2500
0.0146
0.0837
5000
0.00169
0.00993
Point
Estimates
of
Peak
Concentrations:
In
order
to
provide
point
estimates
for
calculating
risk
quotients
(
RQs)
to
be
used
in
the
risk
assessment,
point
estimates
of
peak
concentrations
over
time
were
determined
from
the
model
output.
The
low­
flow
power
plant
facility
treating
at
the
application
rate
once
per
day
was
used
as
a
worst­
case
scenario,
and
provided
the
following
estimated
environmental
concentrations
(
EECs):

Table
8:
Summary
of
Estimated
Environmental
Concentrations
of
DMH
in
Rivers
Receiving
Outfall
from
Low­
Flow
Power
Plants
Using
Once­
Through
Cooling
Systems
Time
Period
Modeled
Peak
Concentration
of
DMH
Duration
of
Peak
Concentration
4
days
36.0
ppb
24
hours
30
days
210
ppb
24
hours
60
days
313
ppb
24
hours
Actual
concentrations
in
receiving
waters
are
likely
lower,
and
will
likely
not
show
the
increasing
trend
indicated
in
Table
8,
due
to
higher
flow
rates
and
possible
degradation/
dissipation
of
DMH
Page
9
of
10
by
mechanisms
other
than
hydrolysis.

UNCERTAINTIES/
LIMITATIONS
The
following
limitations
apply
to
the
results
of
this
model:

°
Dominion
Power
Company
estimates
for
retention
times
vary
according
to
stream
flow
conditions
and
the
number
of
units
in
operation
on
that
given
day.

°
The
PDM4
database
contains
data
on
steam
electric
powerplants
in
the
United
States.
In
the
database,
no
differentiation
was
made
between
those
powerplants
that
used
a
oncethrough
cooling
water
system
and
those
that
did
not.
In
the
absence
of
better
data,
it
was
assumed
that
all
powerplants
in
the
database
used
once­
through
cooling
water
systems,
and
that
the
amount
of
cooling
water
used
in
normal
conditions
was
equal
to
the
7Q10
flow.
These
assumptions
may
not
be
conservative.

°
It
has
been
assumed
that
a
maintenance
dose
of
the
chemical
is
required
daily.
It
is
unclear
from
the
label
whether
or
not
treatment
is
needed
this
frequently.

°
PDM4
is
a
screening­
level
model.
Screening­
level
models
are
rarely
if
ever
used
as
the
sole
justification
for
regulatory
decision­
making
at
EPA.
Additional
data
and
more
rigorous
tools
are
used
to
improve
the
estimates
of
exposures
and
risks
for
such
decisions.
Thus,
although
PDM4
may
be
very
useful
to
industry,
its
results
may
not
accurately
reflect
all
of
the
information
and
data
used
by
EPA
to
make
a
regulatory
decision
on
a
chemical.
Page
10
of
10
REFERENCES
Cloete
TE,
Smith
Z,
Saayman
G.
A
Cooling
Water
System
as
a
Biofilm
Reactor
for
the
Treatment
of
Municipal
Wastewater.
Water
SA
Vol.
25
No.
3
July
1999.
Available
on
website
http://
www.
wrc.
org.
za.

Dang
W,
1996.
Antimicrobial
Pesticides,
Uses,
Human
Exposures,
and
Risk
Assessments.
March,
1996.

DiToro,
D.
M.
1984.
Probability
Model
of
Stream
Quality
Due
to
Runoff.
ASCE.
Journal
of
Environmental
Engineering.
110(
3):
607­
628.

EFAST
Help,
beta
version,
2004.

Genest,
Dan,
Dominion
Power.
Telephone
interview.
June
14,
2004.
