United States
Environmental Protection
Agency
Office Of Water
(4303)
EPA821-R-95-016
February 1995
Statistical Support Document
For The Proposed Effluent
Limitations Guidelines For The
Pharmaceutical Manufeeturin<
Industry
s^
-------
-------
STATISTICAL SUPPORT DOCUMENT FOR THE PROPOSED
EFFLUENT LIMITATIONS GUIDELINES FOR THE PHARMACEUTICAL
MANUFACTURING INDUSTRY
Submitted to:
U.S. Environmental Protection Agency
Office of Science and Technology
Engineering and Analysis Division
401 M Street, S.W.
Washington, D.C. 20460
Submitted by:
Science Applications International Corporation
Environmental and Health Sciences Group
7600-A Leesburg Pike
Falls Church, Virginia 22043
EPA Contract No. 68-C4-0046; Work Assignment No. 0-12
SAIC Project No. 01-0813-07-1676-120
February 10, 1995
-------
-------
TABLE OF CONTENTS
1. INTRODUCTION
Page
2. FACILITY-LEVEL LONG-TERM MEANS AND VARIABILITY FACTORS
2.1
2.2
2.3
Ammonia and Priority and Nonconventional Organic
Pollutants
BOD, COD, and TSS
Cyanide
3. LIMITATIONS AND LONG-TERM MEANS FOR SPECIFIC
TREATMENT TECHNOLOGIES
3.1
3.2
Treatment technologies for Ammonia and Priority
and Nonconventional Organic Pollutants
Advanced Biological Treatment of BOD, COD, TSS
1
1
12
18
19
19
26
-------
LIST OF TABLES
VARIABILITY FACTORS AND LONG-TERM MEANS FOR
ADVANCED BIOLOGICAL TREATMENT
VARIABILITY FACTORS AND LONG-TERM MEANS FOR
IN-PLANT STEAM STRIPPING
VARIABILITY FACTORS AND LONG-TERM MEANS FOR
IN-PLANT STEAM STRIPPING WITH DISTILLATION
BOD, COD, AND TSS 1- AND 30-DAY VARIABILITY FACTORS
ADJUSTED FOR AUTOCORRELATION
LONG-TERM MEANS AND LIMITATIONS FOR
ADVANCED BIOLOGICAL TREATMENT
10
11
14
21
6 LONG-TERM MEANS AND LIMITATIONS FOR
IN-PLANT STEAM STRIPPING
7 LONG-TERM MEANS AND LIMITATIONS FOR
IN-PLANT STEAM STRIPPING WITH DISTILLATION
8 BAT LONG-TERM MEANS AND LIMITATIONS FOR
AMMONIA AIR STRIPPING
g BPT LIMITATIONS AND VARIABILITY FACTORS,
SUBCATEGORY A AND C
10 BPT LIMITATIONS AND VARIABILITY FACTORS,
SUBCATEGORY B AND D
23
24
25
27
28
-------
Statistical Support Document for the Proposed Effluent
Limitations Guidelines for the Pharmaceutical Manufacturing Industry
1. Introduction
This document describes the statistical models and procedures used to estimate limitations
supporting the proposed effluent guideline regulations for the pharmaceutical industry. These effluent
limitations were estimated for technologies that formed the basis of the Best Practicable Control
Technology Currently Available (BPT), Best Conventional Pollutant Control Technology (BCT), Best
Available Technology Economically Achievable (BAT), New Source Performance Standards (NSPS),
Pretreatment Standards for Existing Sources (PSES) and Pretreatment Standards for New Sources
(PSNS).
2. Facility-level long-term means and variability factors
2.1 Ammonia and Priority and Nonconventional Organic Pollutants
Long-term means and variability factors were calculated from actual concentrations of
constituents measured in pharmaceutical wastewaters treated by BAT treatment systems when such
data were available. The data sets of daily effluent concentrations were obtained from pharmaceutical
manufacturers' monitoring data as well as from EPA sampling episodes.
The long-term mean of each dataset was estimated by the arithmetic mean of the daily
concentration values. Observations recorded as below the method detection limit were assigned a
numerical value equal to the detection limit. The daily variability factor was calculated for all
constituents (other than cyanide) by fitting a modified delta-lognormal distribution to daily concentration
data. This distributional model has previously been used by EPA for other categories including the
Organic Chemicals, Plastics and Synthetic Fibers (OCPSF), Pesticides Manufacturing, and Pulp and
Paper industries.
The daily variability factor is defined as the estimated 99th percentile of the concentration
distribution divided by the expected value of the concentration. The 4-day variability factor is defined
similarly except that the 95th percentile of the distribution of 4-day averages is used instead of the 99th
percentile of daily measurements. The description of the modified delta-lognormal model is presented
below.
-------
Modified delta-Iognorma! model
The modified delta-lognormal distribution models the effluent concentration data as a mixture
of non-detects and measured values. This distribution is appropriate because the data for most
constituents consisted of a mixture of measured values and non-detects. The modified delta-lognormal
distribution assumes that each non-detected observation represents an actual concentration equal to
the method detection limit, and the detected values are distributed according to a lognormal
distribution.
The expected value, 99th percentile, and variability factor for the concentration of each
constituent were estimated by fitting the data to a delta-lognormal distribution1, modified to
accommodate a non-zero value of the analytical method detection limit2. The following paragraphs
describe this distribution and show how the variability factors and associated parameters were
estimated from the data.
•
The modified delta-lognormal model is a mixture distribution in which the detected
concentrations follow a standard lognormal distribution (ie, the logarithm of the concentration is
assumed to be normally distributed with parameters mean// and standard deviation a). All nondetects
are assumed to have a concentration value equal to the detection limit. !
i
The cumulative distribution function, which gives the probability that an observed concentration
(C) is less than or equal to some specified level (c), can be expressed as a function of the following
quantities:
D = the detection limit,
6 = the probability of a nondetect,
l{c-D) = an indicator function which equals 1 for c>D and 0 otherwise,
/j = the mean of the distribution of log transformed concentrations,
a = the standard deviation of the distribution of log transformed concentrations.
'•Aitchison, J, and JAC Brown. 1957. The Lognormal Distribution. London:
Cambridge University Press, pp. 95-96.
2This modification of the delta-lognormal distribution was used by EPA in
establishing limitations guidelines for the Organic Chemicals, Plastics, and
Synthetic Fibers point source category. The approach is therefore sometimes
called the "Organics method." ;
-------
y = variable of integration.
The equation of the cumulative distribution function is as follows:
F(c) = P(C<:C) = dl(c-D) + (1-8)
iMM-
(In(y)-u)2
2 a2
(l)
The expected value E(C) of the actual concentration under this distribution function is given by
E(C) = 8£M-(l-8)exp [jn- —), (2)
\ 2 )
and the variance V(C) is given by the following expression:
V(C) = (1-6) exp (2fi+02) [exp (o2)-(i-5)] + 5 (l-8)I?[Z5-2exp (»i + -2i)].
The 99th percentile of the distribution can be expressed in terms of //, a, and the inverse normal
cumulative distribution function (0'1), as follows:
C99 = max ID,
(4)
Finally, the daily variability factor VF(1) is the 99th percentile divided by the expected value:
(5)
The daily variability factors for each facility-constituent dataset were estimated by the following
steps (for notational purposes let a typical dataset consist of n, detects, r,2 nondetects, and have
concentration values X,, i = 1,..,n,). The estimate, ft, of the log mean was calculated by taking the
arithmetic average of the log transformed detects:
(6)
-------
The estimate, &, of the log standard deviation was calculated by taking the square root of the sum of
the squared differences between the log concentrations and />, divided by the number of detects minus
one:
a =
(7)
The estimated probability of a nondetect, 8, was calculated by dividing the number of riondetects by
the number of observations:
s _ nz ; (8)
These quantities were then substituted into equations (2) and (4) to give estimates E(C) and C99 of the
mean concentration and the 99th percentile, respectively. Finally, the resulting estimated mean and
99th percentile were substituted into equation (5) to yield the daily variability factor estimate, vT(1).
The average daily variability factor multiplied by the median long-term mean yields the value
used by EPA as the maximum value that an individual concentration measurement can be expected to
attain. An analogous measure of the maximum value attained by the means of four daily concentration
measurements can also be defined and estimated from the data. The definition of the 4-day variability
factor, VF(4), is the 95th percentile of the distribution of 4-day means, divided by the expected value
of 4-day means.
The value of VF(4) can be estimated from the daily concentration data by iexploiting the
statistical properties of the 4-day mean, C4, and approximating the distribution of C4 by the modified
delta-lognormal model. This approximation has been shown to provide a good estimate to the actual
distribution3. To develop the estimate of VF(4), first note that the logarithm of C4 is normally
distributed with unknown mean and standard deviation denoted by//4 and a4, respectively. Also, E(C4)
= E(C) because the expected value of a sum of random variables divided by a constant, is equal to the
sum of their expectations divided by that constant. And V(C4) = V(C)/4 because the variance of a sum
of independent random variables divided by a constant is equal to the sum of their variances divided
3Barakat
1976. Sums of Independent Lognormally Distributed Random
OdiCUS-CtL. , iv.. -i--' • " • M—..M- jr — - 1 ..
Variables. Journal of the Optical Society of America. 66:211-16
-------
by the constant*. Finally, the probability that C4 is a nondetect is 8\ since the mean of four
independent concentrations is a nondetect only if all four are nondetects, and the probability of this
occurring is equal to the product of the component probabilities, or 8* if the daily nondetect probability
is 8.
The following equations therefore hold:
= E(C) =
(9)
and
1-54
(11)
Equations (9) and (10) can be algebraically solved for a4 in terms of the mean and variance of
the daily concentrations, the probability of a nondetect, and the detection limit. This expression is as
follows:
o4 = ln|l + -
(E(C) -
E(C) -
(12)
To derive an estimate, *4, of the left-hand side of equation (12), each quantity on the right-hand side
was replaced by its estimate computed from the dai.y concentration data; i.e.., E(C) was replaced by
6(0, V(C) by V(C), and 6 by S. Next, the estimated fr4 together with S and B(C) were substituted into
(9), which was solved to yield an estimate />4 of „«. Finally, „, and a4 in (11) were replaced by their
esfmates to yield an estimated value of the 95th percenti.e of the 4-day mean distribution, and this
estimate was divided by E(C) to give the estimated variability factor VF(4).
<*
PrincetOn University
-------
The results of applying the modified delta-lognormal model are shown in Tables 1, 2, and 3,
which give the estimated long-term means and variability factors of constituent concentrations for
facilities using advanced biological treatment, in-plant steam stripping, and in-p.ant steam stripping w.th
distillation, respectively. Note that variably factors could not be estimated for facility datasets that
had fewer than three detected concentrations as indicated by the dot in the VF column.
-------
Table 1. Variability factors and long-term means for
Advanced Biological Treatment
NO. PERCENT
ANALY7E
ACETONE
ACETONE
ACETONE
ACETONE
ACETONITRILE
AMMONIA
BENZENE
BENZENE
BENZENE
BENZENE
BENZENE
BENZENE
CARBON TETRACHLORIDE
CHLOROBENZENE
CHLOROFORM
CHLOROFORM
CHLOROFORM
CHLOROFORM
CHLOROFORM
CHLOROFORM
CHLOROFORM
CHLOROFORM
CHLOROFORM
CHLOROMETHANE
CHLOROMETHANE
ETHANOL
FACILITY
CODE
30540
30623
30701
30949
30623
30540
30540
30540
30600
30623
30623
30701
30050
30623
30050
30050
30540
30540
30600
30623
30623
30701
31108
30540
30623
30540
SAMPLE
SOURCE SITE
I FC
I SE
OF
DBS.
10
19
P FIL/EOP 242
3 SE
I SE
I FC
A FC
S FC
S SPE
P SE
S SE
S PPE
S SE
S SE
S SE
V SE
I FC
S FC
S SPE
I SE
P SE
S PPE
3 SE
I FC
I SE
I FC
2
19
8
1
1
1
3
1
2
' 1
1
1
3
10
1
1
19
3
2
2
10
19
10
OF
NONDET. UNITS
50.00 UG/L
42.11 UG/L
96.69 UG/L
0.00 UG/L
100.00 UG/L
0.00 MG/L
100.00 UG/L
100.00 UG/L
0.00 UG/L
100.00 UG/L
100.00 UG/L
100.00 UG/L
100.00 UG/L
100.00 UG/L
0.00 UG/L
100.00 UG/L
100.00 UG/L
100.00 UG/L
0.00 UG/L
57.89 UG/L
0.00 UG/L
0.00 UG/L
50.00 UG/L
90.00 UG/L
100.00 UG/L
90.00 UG/L
DET. MIN.
LIMIT DETECT
50.00 55.0
50.00 52.0
1000.00 1000.0
10.00 29.0
5.00
1.40 1.4
40.00
10.00
120.00 120.0
2.00
10.00
0.01
0.01
10.00
8.10 8.1
1.00
10.00
10.00
110.00 110.0
10.00 10.0
9.00 9.0
0.90 0.9
1.70 220.0
50.00 124.0
50.00
500.00 5515.0
ortnri_c
LONG-TERM SAMPLE DAILY 4-DAY
MEAN MAX. VF VF
136.. 90 800.00 8.17184 2.62475
66.,42 130.00 2.34136 1.34904
1i078.51
89.50
5.00
2.56
40.00
10.00
120.00
2.00
10.00
0.01
0.01
10.00
8.10
1.00
10.00
10.00
110.00
12.63
13.33
1.85
110.85
53,40
50.00
10P1.50
8000.00 3. 64777 1.56822
150.00 .
5.00 .
3.70 1.89382 1.26333
40.00
10.00
120.00 .
2.00
10.00 .
0.01
0.01
10.00 .
8.10 .
1.00 .
10.00
10.00 .
110.00 .
24.00 2.30398 1.33029
18.00 2.10986 1.31781
2.80
220.00 .
124.00 .
50.00 .
5515.00 .
-------
Table 1. Variability factors and long-term means for
Advanced Biological Treatment
SAMPLE
ANALYTE
ETHANOL
ETHANOL
ETHYL ACETATE
ETHYL ACETATE
FORMALDEHYDE
HEPTANE
HEXANE
ISOPROPANOL
1SOPROPANOL
ISOPROPYL ACETATE
METHANOL
METHANOL
HETHANOL
HETHYLENE CHLORIDE
METHYLENE CHLORIDE
HETHYLENE CHLORIDE
HETHYLENE CHLORIDE
HETHYLENE CHLORIDE
HETHYLENE CHLORIDE
HETHYLENE CHLORIDE
HETHYLENE CHLORIDE
N-BUTANOL
N.N-DIHETHYLFORHAHID
PHENOL
PHENOL
PHENOL
FACILITY
CODE
30623
30701
30540
30701
30623
30623
30623
30540
30949
30540
30540
30623
30701
30349
30540
30600
30623
30623
30623
30704
30949
30949
i 30623
30050
30623
30623
NU. fCKUCNI
SAHPLE OF OF
SOURCE SITE OBS. NONDET. UNITS
I SE
P FIL/EOP
I FC
P FIL/EOP
I SE
I SE
I SE
I FC
3 SE
I FC
I FC
I SE
P FIL/EOP
S SE
S FC
S SPE
I SE
P SE
S SE
3 SE
3 SE
3 SE
I SE
V SE
I SE
S SE
20
242
10
242
9
19
19
10
2
10
10
20
242
1
1
1
19
3
1
1
2
2
20
2
20
1
85.00 UG/L
93.39 UG/L
90.00 UG/L
100.00 UG/L
0.00 UG/L
100.00 UG/L
100.00 UG/L
100.00 UG/L
100.00 UG/L
100.00 UG/L
0.00 UG/L
90.00 UG/L
100.00 UG/L
0.00 UG/L
0.00 UG/L
0.00 UG/L
78.95 UG/L
0.00 UG/L
100.00 UG/L
100.00 UG/L
0.00 UG/L
100.00 UG/L
95.00 UG/L
0.00 UG/L
85.00 UG/L
100.00 UG/L
DET. HIN. LONG-TERM SAMPLE
LIMIT DETECT MEAN MAX.
500 500
1000 1000
500 600
1000
120 120
5
5
500
100
500
2260 2260
500 300
1000
250 250
140 140
2600 2600
10 26
4 4
10
40
5 17
250
10 35
3 3
10 16
10
530.00
1074.38
510.00
1000.00
343.33
5.00
5.00
500.00
62.50
500.00
13560.00
650.00
1000.00
250.00
140.00
2600.00
109.58
68.67
10.00
40.00
19.50
250.00
11.25
3.00
11.65
10.00
800
10000
600
1000
800
5
• 5
500
100
500
87560
3700
iooo
'250
140
2600
1095
:110
10
40
22
:250
35
3
25
10
DAILY 4-DAY
VF VF
1.9342 1.21086
3.0073 1.46986
•
-
4.3100 1.81518
•
•
•
•
•
6.6575 2.33730
•
-
•
-
•
18.3879 4.20528
13.4512 3.81305
•
•
•
•
•
•
2.4959 1.37214
•
-------
Table 1. Variability factors and long-term means for
Advanced Biological Treatment
ANALYTE
PYRIDINE
TETRAHYDROFURAN
TOLUENE
TOLUENE
TOLUENE
TOLUENE
TOLUENE
TOLUENE
TOLUENE
TOLUENE
TOLUENE
TR I CHLOROFLUOROHETHA
TRICHLOROFLUOROHETHA
TRIETHYLAMINE
XYLENES
XYLENES
1,2-DICHLOROETHANE
1,2-DICHLOROETHANE
1,2-DICHLOROETHANE
1,4-DICHLOROBENZENE
1,4-DIOXANE
2-BUTANONE
2-HETHYLPYRIDINE
4-HETHYL-2-PENTANONE
4-HETHYL-2-PENTANONE
FACILITY
CODE
30949
30623
30050
30050
30540
30600
30623
30623
30623
30701
30701
30540
30600
30949
30540
30540
30623
30623
30623
31108
30540
30623
30540
30540
30949
SAMPLE
SOURCE SITE
3
I
S
V
I
S
I
p
S
p
S
I
S
3
I
I
I
P
S
3
I
I
I
I
3
SE
SE
SE
SE
FC
SPE
SE
SE
SE
FIL/EOP
PPE
FC
SPE
SE
FC
FC
SE
SE
SE
SE
FC
SE
FC
FC
SE
OF
OBS.
2
20
1
3
10
1
19
3
1
242
2
10
1
2
10
10
19
3
1
2
10
19
10
10
2
OF
NONDET.
100.00
15.00
100.00
100.00
100.00
0.00
100.00
66.67
100.00
100.00
100.00
10.00
0.00
100.00
100.00
100.00
15.79
33.33
0.00
100.00
90.00
94.74
100.00
100.00
100.00
UNITS
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
DET. HIM.
LIMIT DETECT
10
5
0
1
10
180
10
2
10
1000
0
10
420
50
10
.00
.00 20
.01
.00
.00
.00 180
.00
.00 53
.00
.00
.01
.00 10
.00 420
.00
.00
10.00
10
2,
500,
1.
.00 11
.00 3
.00 500
.00
10.00 463
50.00 65
50.00
50.00
10.00
LONG-TERM
MEAN
10.00
1222.95
0.01
1.00
1CI.OO
1 SCI. 00
10.00
19.00
10.00
1000.00
0.01
18.25
420.00
50.00
10.00
10.00
64.16
38.33
500.00
1.00
55.30
50.79
50.00
50.00
10.00
SAMPLE DAILY 4-DAY
MAX. VF VF
10
3484
.00
.00 12.2719 3.55499
0.01
1.00 .
10.00
180.00
10.00
53.00
10.00
1000.00
0.01
43.00 2.7344 1.46762
420.00
50.00
10.00
10.
263.
110.
500.
1.
463.
65.
50.
50.
10.
00
00 6.8272 2.36887
00
00
00
00
00
00
00
00
-------
Table 2. Variability factors and long-term means for
In-plant Steam Stripping
ANALYTE
CHLOROFORM
ETHANOL
ISOPROPANOL
METHANOL
METHYLENE CHLORIDE
HETHYLENE CHLORIDE
HETHYLENE CHLORIDE
PYR1DINE
TETRAHYDROFURAN
TOLUENE
TOLUENE
2-BUTANONE (HEK)
2-PROPANONE (ACETONE;
FACILITY
CODE
30329
30329
30329
30329
30329
30487
30618
30487
30329
30329
30487
30329
» 30329
SAMPLE
SITE
PC
PC/PI LOT -T
PC/B45
PC/PI LOT -C
PC/B45
PC/PI LOT-C
PC/BENCH
PC/PILOT-C
PC/PI LOT -T
PC
PC
PC
PC
PC
PC/BENCH
PC/B45
PC/PILOT-C
PC/PILOT-T
PC
PC
PC
NO.
OF
OBS.
16
8
11
16
11
16
5
16
8
16
23
19
1
5
5
6
16
8
23
11
16
PERCENT
OF
NONDET.
100.00
100.00
0.00
50.00
45.45
50.00
0.00
0.00
0.00
81.25
95.65
100.00
100.00
0.00
0.00
50.00
50.00
87.50
100.00
0.00
0.00
UNITS
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
DET.
LIMIT
10
10
500
99000
500
•
•
10
100
1000
1000
•
1000
10
10
100
•
.
MIN.
DETECT
.
•
102223
3000
22357
2100
120000
46000
60000000
192
101
•
•
220
333
124
27
12
•
4697
197
SAMPLE
LONG-TERM
MEAN
10.00
10.00
693371.27
8231.25
146486.09
6125.00
1374000.00
473925.00
85000000.00
58.06
100.04
3842.11
1000.00
1542.00
473.20
1125.83
21.94
10.25
100.00
121237.91
3000.38
SAMPLE
' MAX.
'
!
1331570
1 36000
369714
: 21900
3900000
1300000
120000000
398
: 101
'
:
2570
630
2414
42
: 12
:
392212
9677
DAILY
VF
.
•
4.3821
8.1578
8.2380
7.4029
8.4810
9.8162
1.7183
8.0863
•
•
•
5.9701
1.6922
10.2576
1 .9807
•
•
11.8742
10.4546
4-DAY
VF
.
•
1 .83092
2.63772
2.69990
2.50152
2.76344
3.08106
1.21729
2.78545
•
•
•
2.18155
1.21028
3.36084
1.48040
•
•
3.54418
3.23076
10
-------
Table 3. Variability factors and long-term means for
In-plant Steam Stripping with Distillation
ANALYTE
CHLOROFORM
MET HANOI
METHYLENE CHLORIDE
METHYLENE CHLORIDE
METHYLENE CHLORIDE
PYRIDINE
TETRAHYDROFURAN
TOLUENE
TOLUENE
2-BUTANONE (MEK)
2-PROPANONE (ACETONE)
FACILITY
CODE
30329
30767
30329
30487
30618
30487
30329
30329
30487
30329
30329
SAMPLE
SITE
PC
PC/PI LOT -T
UNKNOWN
PC
PC
PC
PC
PC
PC/BENCH
PC/845
PC/PI LOT -C
PC/PI LOT -T
PC
PC
PC
NO.
OF
OBS.
16
8
65
16
23
19
1
5
5
2
16
8
23
7
8
PERCENT
OF
NONDET.
100.00
100.00
30.77
81.25
95.65
100.00
100.00
0.00
0.00
50.00
50.00
87.50
100.00
0.00
0.00
UNITS
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
DET.
LIMIT
10
10
500
10
100
1000
1000
•
1000
10
10
100
.
m
MIN.
DETECT
•
500
192
101
.
.
220
333
124
27
12
.
4697
197
SAMPLE
LONG-TERM
MEAN
10.00
10.00
1518.46
58.06
100.04
3842.11
1000.00
1542.00
473.20
562.00
21.94
10.25
100.00
25812.71
388.88
SAMPLE DAILY
MAX. VF
•
11100 5.47835
398 8.08626
101
.
.
2570 5.97012
630 1.69220
124
42 1.98075
12
.
99406 6.24302
769 3.06842
4-DAY
VF
•
2.06475
2.78545
.
.
.
2.18155
1.21028
1 .48040
.
2.24305
1.54190
11
-------
2.2 BOD, COD, and TSS
The calculation of variability factors and long-term means was based on actual concentrations
of BOD, COD, and TSS measured in pharmaceutical wastewaters treated by advanced biological
treatment systems. A 1 -day and 30-day average variability factor and limitation was estimated for each
dataset from a modified delta-lognormal distribution that was fit to the data. The facility limitations
were averaged to derive the overall 1-day and 30-day maximum limitations.
The 1-day facility-level variability factors were computed as described above in Section 2.1.
An alternative approach was taken to estimate the 30-day average variability factor in order to account
for additional variability due to the day-to-day correlation in concentrations of BOD, COD, and TSS.
The adjustment factor was based on fitting a lag-1 autocorrelation time series model from adjacent day
observations. This time-series model has been used by EPA in the OCPSF, pesticides, and pulp and
paper rulemakings.
A lag-1 autocorrelation was computed directly from each facility dataset when possible; an
average over all such computed Iag-1 autocorrelations was transferred to datasets with Insufficient
numbers of adjacent day observations for direct estimation. The autocorrelation adjustment factor is
described below.
12
-------
Adjustment of 30-day variability factor for autocorrelation
The methodology to calculate 30-day variability factors for BOD, COD and TSS assumes that
(1) the time series of daily concentration measurements for each pollutant is stationary (ie, there is no
time trend present), and (2) the lag-k autocorrelations follow a first order autoregressive model (ie, the
correlation of measurements taken k days apart is equal to the lag-1 correlation raised to the kth power).
These assumptions and the central limit theorem approximation that 30-day means are normally
distributed lead to the following formula for the 30-day variability factor:
VF{30) = 1 + 1.645
30
(13)
where the lag-1 autocorrelation, p, and the standard deviation, a, are estimated from the time series
data, and the factor f30 is given by the following expression
f (o) = 1 H-
3°(P) 1 "
2
30
(30-*) [exp(p*a2) - l]
exp«r>) -1 --
(14)
When assumption (2) does not hold, the formula for computing the f30 factor is obtained by replacing
/>k in equation (2) with the general lag-k autocorrelation, pk.
The results of applying the modified delta-lognormal model and autocorrelation adjustment are
shown in Table 3, which gives the estimated means, 1 - and 30-day variability factors, and limitations
for BOD, COD, and TSS.
13
-------
Table 4. BOD, COD, and TSS 1- and 30-day variability factors
Adjusted for autocorrelation
iNALYTE
BODS
BOOS
BODS
BODS
BODS
BODS
BODS
BODS
BODS
BODS
BODS
BOOS
BODS
BODS
BODS
BODS
BOOS
BODS
BODS
BODS
BODS
BODS
BODS
BODS
BODS
BODS
FACILITY
CODE
12053
12053
12317
50001
50001
50002
50002
50003
50003
50004
50004
50005
50005
50006
50006
50007
50007
50007
50007
50008
50008
50009
50009
50009
50009
50011
SAMPLE
SITE
E
E
E
600
200
410
110
550
130
100
410
460
100
510
100
530
510
410
100
410
100
600
078
075
015
410
NO.
OF
OBS.
39
39
51
362
363
354
348
365
364
356
356
366
366
52
52
336
159
228
339
51
48
97
82
88
79
42
UNIT
HG/L
PPD
HG/L
HG/L
HG/L
HG/L
HG/L
HG/L
HG/L
LB/D
HG/L
HG/L
HG/L
HG/L
HG/L
HG/L
HG/L
HG/L
HG/L
HG/L
HG/L
HG/L
HG/L
HG/L
HG/L
HG/L
LAG-1
AUTO
CORRELATION
0.59595
0.01154
0.59595
0.50232
0.20280
0.78486
0.78366
0.26837
0.65735
0.26331
0.34867
0.65670
0.29684
0.59595
0.59595
0.71272
0.64379
0.75653
0.39952
0.59595
0.59595
0.57677
0.64445
0.59118
0.27352
0.59595
HIN.
DETECT
2.000
0.232
1.000
1.100
60.000
3.000
10.000
15.984
230.000
699.978
0.839
16.000
566.000
5.000
180.000
2.600
5.000
3.000
140.000
0.930
49.200
5.000
8.000
117.000
826.000
1.000
SAMPLE
HEAN
5.87
0.91
4.53
9.04
1031.29
40.30
970.59
74.28
2439.72
5746.33
4.65
77.83
2621.21
15.87
1970.67
10.01
25.80
20.09
1215.37
6.68
323.21
45.54
311.20
3347.69
41656.99
7.70
SAHPLE
HAX.
13.00
2.39
18.00
62.00
12000.00
174.00
2080.00
809.36
9733.00
14119.99
19.65
660.00
5876.00
41.00
4400.00
53.00
86.40
81.60
2940.00
29.75
857.14
219.00
2400.00
40800.00
452200.00
26.00
DAILY
LIHITATION
16.33
2.96
19.91
32.86
4415.70
143.41
3086.43
211.41
5870.87
12739.99
16.61
253.23
6690.02
45.87
8419.25
34.30
75.83
62.42
2804.27
34.31
1213.21
225.37
3400.12
18669.80
242945.74
38.78
30 DAY
LIHITATION
7.63
1.09
6.70
12.21
1349.55
62.32
1476.80
87.22
3123.66
6642.17
5.96
104.56
3130.79
20.70
2998.61
14.13
34.'57
29.07
1441.86
10.18
465 .: 14
70J06
771,81
5229.95
61289.07
12^28
DAILY
VF
2.76670
3.22390
4.39433
3.63105
4.26077
3.52482
3.03134
2.94485
2.37745
2.19944
3.59315
3.39913
2.52927
2.89946
4.04568
3.55739
2.93043
3.07672
2.29497
5.31513
3.64091
4.88115
9.75144
5.80489
5.68398
4.80999
30 DAY
VF
1.29233
1.18938
1.47886
1 .34899
1 .30220
1.53186
1.45044
1.21489
1.26495
1.14671
1.29048
1.40346
1.18365
1 .30885
1.44091
1 .46534
1.33569
1.43263
1.17999
1.57758
1 .39590
1.51730
2.21353
1.62612
1.43392
1.52354
14
-------
Table 4. BOD, COD, and TSS 1- and 30-day variability factors
Adjusted for autocorrelation
ANALYTE
BODS
BOD5
BODS
BODS
BODS
COD
COD
COD
COD
COD
COD
COD
COD
COD
COD
COD
COD
COD
COD
COD
COD
COD
COD
COD
COD
COD
I-AU1L1I1
CODE
50011
50012
50012
50013
50013
12053
12053
12317
50001
50001
50002
50002
50003
50003
50004
50004
50005
50005
50006
50006
50007
50007
50007
50007
50008
50008
1 SAMPLE
SITE
140
540
100
520
110
E
E
E
600
200
410
110
550
130
100
410
460
100
510
100
530
510
410
100
410
100
OF
OBS.
41
21
24
67
99
52
51
248
362
363
359
359
365
365
50
51
366
365
52
52
614
5
605
682
63
70
UNIT
MG/L
MG/L
MG/L
MG/L
MG/L
MG/L
PPD
MG/L
MG/L
MG/L
MG/L
MG/L
MG/L
MG/L
LB/D
MG/L
MG/L
MG/L
MG/L
MG/L
MG/L
MG/L
MG/L
MG/L
MG/L
MG/L
AUTO
CORRELATION
0.59595
0.59595
0.59595
0.47187
0.62230
0.66556
0.45046
0.77814
0.58011
0.29984
0.91044
0.90209
0.33776
0.76498
0.50951
0.50951
0.54317
0.26379
0.50951
0.50951
0.90370
0.50951
0.87151
0.31155
0.61282
0.22158
MIN.
DETECT
92.00
0.40
0.40
21.00
105.00
8.00
1.36
2.00
65.00
95.00
58.00
157.00
196.80
526.00
1449.98
27.86
108.00
986.00
7.00
151.00
72.00
260.00
108.00
436.00
3.00
126.00
SAMPLE
MEAN
258.80
2.51
143.84
187.18
3413.25
79.52
11.69
16.65
150.86
1671.42
374.03
2031.80
744.23
4961.02
9192.68
98.27
883.40
5278.92
93.62
2283.48
369.65
353.80
415.95
3424.63
17.06
761.65
SAMPLE
MAX.
1100.00
9.40
373.00
753.00
5910.00
226.00
33.75
62.00
665.00
17400.00
1006.00
3750.00
10080.18
11980.00
13999.99
189.01
1702.00
11583.00
214.00
5875.00
1119.00
470.00
1156.00
28200.00
70.00
1583.00
DAILY
LIMITATION
724.47
11.33
4123.02
908.54
8568.49
263.88
', 38.68
57.22
,303 ..54
6943 ,.86
'1 150 .,00
6111,,73
1693 .,23
11i808 .71
23447.06
201.60
1676.92
12496.82
372.48
13481.39
730.70
568.35
780.57
7835.40
60.19
2554.51
30 DAY
LIMITATION
333.44
3.74
894.36
275.62
4447.56
112.63
15.30
25.29
178.39
2214.45
620.95
3347.96
836.95
6638.76
11548.25
115.72
1029.45
6174.14
136.04
3857.85
500.92
394.28
537.52
3970.92
23.91
965.24
DAILY
VF
2.8233
4.5222
.11.7601
4.7533
2.4222
3.2595
3.2682
3.4056
2.0184
4.1419
3.0401
2.9056
2.3812
2.3387
2.4959
2.0433
1.8870
2.3502
3.7293
5^2906
1 .9714
1.5994
1 .8764
2.2966
3.4964
3.2560
30 DAY
VF
1.29941
1.49266
2.55099
1.44198
1.25728
1.39123
1.29322
1.50544
1.18619
1 .32088
1.64150
1.59167
1.17702
1.31481
1.22930
1.17286
1.15842
1.16115
1 .36200
1.51395
1.35146
1.10952
1.29213
1.16390
1 .38881
1.23033
15
-------
Table 4. BOD, COD, and TSS 1- and 30-day variability factors
Adjusted for autocorrelation
IALYTE
COO
COD
COD
COD
COO
COO
COO
COD
COD
COD
TSS
TSS
TSS
TSS
TSS
TSS
TSS
TSS
TSS
TSS
TSS
TSS
TSS
TSS
TSS
TSS
FACILITY
CODE
50009
50009
50009
50009
50011
50011
50012
50012
50013
50813
12053
12053
12317
50001
50002
50003
50004
50005
50006
50007
50007
50007
50007
50008
50008
50009
SAMPLE
SITE
600
078
075
015
410
140
540
100
520
110
E
E
E
600
410
550
410
460
510
530
510
410
100
410
100
600
NO.
OF
OBS.
345
93
218
191
38
39
21
24
68
163
84
84
248
362
107
365
356
366
52
673
462
951
783
51
48
346
UNIT
HG/L
HG/L
HG/L
HG/L
HG/L
HG/L
HG/L
HG/L
HG/L
HG/L
HG/L
PPD
HG/L
HG/L
HG/L
HG/L
HG/L
HG/L
HG/L
HG/L
HG/L
HG/L
HG/L
HG/L
HG/L
HG/L
LAG-1
AUTO
CORRELATION
0.83209
0.88799
0.63628
0.19172
0.50951
0.50951
0.50951
0.50951
0.50951
0.56933
0.37161
0.43463
0.77685
0.62560
0.18041
0.38215
0.54391
0.62213
0.55093
0.59907
0.75337
0.77547
0.55539
0.55093
0.55093
0.39694
HIN.
DETECT
74.00
51.00
351.00
258.00
20.00
240.00
7.00
56.40
250.00
1500.00
0.40
0.06
1.00
12.00
8.00
16.34
0.49
8.00
1.00
2.00
4.00
4.00
24.00
1.00
48.00
1.00
SAHPLE
HEAN
267.82
754.65
5388.32
64425.75
74.61
548.21
24.60
256.23
1674.26
10838.18
6.84
1.03
5.87
63.21
101.55
154.62
18.69
105.84
37.92
11.85
24.19
30.37
655.19
7.67
213.81
11.93
SAMPLE
MAX.
822.00
7800.00
27600.00
304000.00
260.00
1100.00
69.30
462.30
3500.00
19750.00
35.00
5.26
41.00
240.00
604.00
2713.58
254.40
577.00
198.00
60.00
110.00
158.00
40434.00
41.50
672.00
95.00
DAILY
LIHITATION
769.70
4590.88
26316.54
429227.31
229.77
1219.69
88.45
775.01
4033.85
19493.77
36.38
5.54
32.20
204.34
469.17
555.22
91.47
385.07
259.75
36.76
58.98
87.25
2803.97
39.83
678.87
53.67
30 DAY
LIHITATION
395.14
1637.15
8485.32
105018.78
95.79
657.07
I
33.40
336.73
2062.04
12522.71
10.06
1.55
10.38
86.29
137.96
194.87
27.27
148.84
67.41
15.85
32.03
43.08
879.34
11.94
284.64
16.84
DAILY
VF
2.87356
6.35599
4.83076
5.77204
3.11487
2.21929
3.56248
2.94708
2.37694
1.78875
5.17731
5.25428
5.73826
3.23015
4.45756
3.74424
5.08399
3.65529
6.36306
3.09294
2.44095
2.86584
4.74914
5.06212
3.14374
4.38053
30 DAY
VF
1.47519
2.26661
1.55760
1.41224
1.29856
1.19557
1.34514
1.28045
1.21506
1.14909
1.43214
1.46848
1.84927
1.36402
1.31078
1.31418
1.51567
1.41291
1.65139
1 .33381
1 .32549
1.41502
1.48936
1.51815
1.31813
1.37483
16
-------
Table 4. BOD, COD, s
Adju
INALYTE
TSS
TSS
TSS
TSS
TSS
TSS
TSS
TSS
FACILITY
CODE
50009
50009
50009
50011
50012
50012
50013
50013
SAMPLE
SITE
078
075
015
410
540
100
520
110
NO.
OBS.
216
231
172
42
20
24
72
230
UNIT
MG/L
MG/L
MG/L
MG/L
MG/L
MG/L
MG/L
MG/L
LAG-1
CORRELATI
0.44561
0.18427
0.21235
0.55093
0.55093
0.55093
0.98000
0.81082
BOD, COD and TSS 1- and 30-day variability factors
Adjusted for autocorrelation
3.0 152.60 6630.0 815.7'9
8.6 166.49 11400.0 671.32
2.0 769.59 11000.0 14686.27
5.0 28.38 120.0 119.41
0.1 1.52 6.0 15.76
2.4 31.58 129.5 266.83
23.0 92.71 392.0 325.04
208.0 2431.22 4342.0 4805.87
198.09
168.95
2804.19
41.16
3.47
62.18
185.92
3109.81
DAILY 30 DAY
VF VF
6-5741
5.5282
12.6528
4.1297
8.7639
7.7405
3.5311
1.9590
1.59636
1.39124
2.41591
1.42339
1.93157
1.80366
2.01972
1.26765
17
-------
2.3 Cyanide
A bata rather than a daita-iognorma, distribution was used to mode, cyanide affluent j* The
was ca.cu.ated. The four-day cyanide limitation was estimated in a similar fash.on.
Beta model for cyanide limitations
The beta density function with parameters a and b is expressed as foilows:
r(a+Jb)
r(a)r(jb)
-v.
'x
(15)
These parameters are
related to the mean fj and variance a2 by the equations
a =
(16)
and
b =
L-l
(17)
M.T.
c Ksti.ation. McGraw-Hill, New York|. p. 5
18
-------
The method of moments was used to estimate the parameters of this mode,. Sample estimates
o, the mean and variance were oomputed from the data w,,h the nondetects se, ,„ the detection ,imi,
These estimates were substituted for, and S in the right-hand side of the above equa,,ons, yieldlna
estimates of a and b. The estimated a and b ware substituted into the expression ,or the beta density
which was then numerically integrated to estimate the 99" percentile, o, the daily limitation.
To estimate the 4-day limitation, the 95" percentile of 4-day means, „ was assumed tha, 4-day
means of beta random variables could be approximated by a beta random variable. The method of
moments was then used as before to estimate parameters and percentiles. The moments are readiiy
available since the 4-day mean is eoua, to the ,-day mean, and the 4-day variance is one-fourth the ,.
day variance. In fact, sums of beta random variables are no, exactly beta distributed, and as a check
on the validity „, the approximation assumption, a computer simulation approach was adopted ,„
mode, the distribution of the average o, four beta random variables. Limitations derived by these two
alternative methods agreed wlthln 2%, indicating tha, the assumption y,e,ds a reasonably dose
approximation to the percentiles of the true distribution. The estimated long-term mean, variability
factors, and limitations for cyanide destruction are as follows:
Long-term mean
(mg/L)
3. Limitations and long-term means for
3.1 Treatment .echnoiogies fo, Ammonia and Priori^ and Nonconven,ron,, Organic Po,,utants
The previous sections described how facility-level variability factors (VFs) were estimated from
daily effluen, concentration data using the modified delta-lognorma, mod,,. ,„ this section we describe
how these facility-leve, VFs were used to estimate daily and monthly average limitations for selected
treatment technologies. o«siet-«su
Each treatment technology was represented by a se, of facillty-level datasets consisting of daily
effluen, concentration measurements on a range o, different constituents. The daily limitation and
19
-------
mon,h,y averace limitafon for a giyen constituent was estimated accord.ng ,o the following
methodology:
. The long-term mean (LTM) was estimated for each facility dataset by computing the
arithmetic average of the constituent daily concentrations. Observations below the
method detection limit (DL) were set equal to the DL for the purposes of thus
calculation.
. The constituent LTM was defined to be the median of the facl.fty-.eve. constituent
LTMs.
. For those facility datasets that had at least three detected va.ues, the modified delta-
.ognorma, mode, was used to estimate daily and monthly average variability factors (VF)
as described above. ;
. The constituent daily VF was defined to be the median o, the facility-levei daily VFs;
the constituent monthly aye,ag. VF was defined to be the median o, the facility-leve,
monthly average VFs.
. The dai.y and month.y average limitations were ca.culated by multiplying the constituent
LTM by the daily and monthly constituent VFs, respectively.
A VF was transferred if it could not be direct.y estimated because of insufficient data;
the transferred VF was defined to be the median of the other organic constrtuent VFs.
Tables 5 through 8 show the results of applying this methodo.ogy to estimate option ,ong-term
means and limitations for advanced biological treatment, in-p.ant steam stripping, in-p,ant steam
stripping with distillation, and air stripping of ammonia.
20
-------
Table 5. Long-term means and limitations
Advanced biological treatment
Analyte
ACETONE
ACETONITRILE
AMMONIA
BENZENE
CARBON TETRACHLORJDE
CHLOROBENZENE
CHLOROFORM
CHLOROMETHANE
ETHANOL
ETHYL ACETATE
FORMALDEHYDE
HEPTANE
HEXANE
ISOPROPANOL
ISOPROPYL ACETATE
METHANOL
METHYLENE CHLORIDE
N-BUTANOL
N.N-DIMETHYLFORMAMID
PHENOL
PYRIDINE
TETRAHYDROFURAN
TOLUENE
TRICHLOROFLUOROMETHA
Unit
U6/L
UG/L
MG/L
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
Sample
Number of Long-Term
Facilities Mean
* 113.20
1 5.00
1 2.56
6 10.00
1 0.01
1 10.00
9 10.00
2 51.70
3 1001.50
2 755.00
1 343.33
1 5.00
1 5.00
2 281.25
1 500.00
3 1000.00
8 89.12
1 250.00
1 11.25
3 10.00
1 10.00
1 1222.95
9 10.00
2 219.13
Daily
VF
3.6478
3.9789
1.8938
3.9789
3.9789
3.9789
2.2069
3.9789
2.4707
3.9789
4.3100
3.9789
3.9789
3.9789
3.9789
6.6575
15.9195
3.9789
3.9789
2.4959
3.9789
12.2719
3.9789
2.7344
Daily
Limit
412.93
19.89
4.84
39.79
0.04
39.79
22.07
205.71
2474.43
3004.04
1479.76
19.89
19.89
1119.06
1989.43
6657.48
1418.79
994.72
44.76
24.96
39.79
15007.93
39.79
599.18
4 Day
VF
1.56822
1.69170
1.26333
1.69170
1.69170
'1.69170
1.32405
1.69170
1.34036
1.69170
1.81518
1.69170
1.69170
1.69170
1,. 69170
2.33730
4.00916
1.6917'0
1.69170
1.37214
1.69170
3.55499
1.69170
1.46762
4 Day
Limit
177.52
8.46
3.23
16.92
0.02
16.92
13.24
87.46
1342.37
1277.23
623.21
8.46
8.46
475.79
845.85
2337.30
357.31
422.92
19.03
13.72
16.92
4347.58
16.92
321.59
Transfer
VFs
From Option
Y
Y
Y
y
y
Y
Y
Y
Y
Y
Y
y
Y
Y
21
-------
Table 5. Long-term means and limitations
Advanced biological treatment
Analyte
TRIETHYLAMINE
XYLENES
1,2-DICHLOROETHANE
1,4-DICHLOROBENZENE
1,4-DIOXANE
2-BUTAKONE
2-METHYLPYRIDINE
4-METHYL-2-PENTANONE
Number of
Unit Facilities
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
1
2
3
1
1
1
1
2
Sample
Long-Term
Mean
50.00
10.00
64.16
1.00
55.30
50.79
50.00
30.00
Daily
VF
3.9789
3.9789
6.8272
3.9789
3.9789
3.9789
3.9789
3.9789
Dai ly
Limit
198.94
39.79
438.02
3.98
220.03
202.08
198.94
119.37
4 Day
VF
1.69170
1.69170
2.36887
1.69170
1.69170
1.69170
1.69170
1.69170
Transfer
4 Day VFs
Limit From Option
84.58
16.92
151.98
1.69
93.55
85.92
84.58
50.75
Y
Y
Y
Y
Y
Y
Y
22
-------
Table 6. Long-term means and limitations
In-Plant Steam Stripping
Analyte
CHLOROFORM
ETHANOL
ISOPROPANOL
HETHANOL
METHYLENE CHLORIDE
PYRIDINE
TETRAHYDROFURAN
TOLUENE
2-BUTANONE (MEK)
2-PROPANONE (ACETONE)
Unit
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
Number of
Facilities
2
2
2
3
3
1
1
5
1
1
Sample
Long-Term
Mean
10.00
350801.26
76305.55
1374000.00
100.04
1000.00
1542.00
100.00
121237.91
3000.38
Daily
VF
7.9533
6.2700
7.8204
8.4810
8.0863
7.9533
5.9701
1 .9807
11.8742
10.4546
Dai ly
Limit
79.53
2199509.27
596741.52
11652951.31
808.98
7953.34
9205.93
198.07
1439606.92
31367.78
4 Day
Vl:
2.68JJ07
?..23432
21.60071
2. 76344
2.78545
2.68207
2.18155
1.48040
3.54418
3.23076
4 Day
Limit
26.82
783803.01
198448.39
3796967.47
278.67
2682.07
3363.95
148.04
429689.10
9693.50
Transfer
VFs
From Option
Y
Y
23
-------
Table 7. Long-term means and limitations
In-Plant Steam Stripping with Distillation
Analyte
CHLOROFORM
HETHANOL
HETHYLENE CHLORIDE
PYRIDINE
TETRAHYDROFURAN
TOLUENE
2-BUTANONE (MEK)
2-PROPANONE (ACETONE)
Number of
Unit Facilities
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
UG/L
2
1
3
1
1
5
1
1
Sample
Long-Term
Mean
10.00
1518. 46
100.04
1000.00
1542.00
100.00
25812.71
388.88
Dai ly
VF
5.72424
5.47835
8.08626
5.72424
5.97012
1.83647
6.24302
3.06842
Dai ly
Limit
57.24
8318.67
808.98
5724.24
9205.93
183.65
161149.21
1.193.23
4 Day
VF
2.12315
2.06475
2.78545
2.12315
2.18155
1.34534
2.24305
1.54190
Transfer
4! Day VFs
Limit From Option
21.23 Y
3135.24
278.67
2123.15 Y
3363.95
134.53
57899.11
599.61
24
-------
Table 8. BAT long-term means and limitations
Ammonia Air Stripping
Analyte
AMMONIA
Unit
MG/L
Number of
Facilities
Sample
Long-Term
Mean
9.91429
Dai ly
VF
Daily
Limit
1.30544 12.9425
4 Day
VF
1.09986
4 Day
Limit
10.9044
25
-------
3.2 Advanced Biological Treatment of BOD, COD, and TSS
The limitations and variability factors for BOD, COD, and TSS based on advanced biological
treatment are shown in Table 7 for subcategory A and C, and in Table 8 for subcategory B and D.
They were derived by taking the arithmetic averages of limitations and variability factors of BPT facility
datasets that represent treated effluent from advanced biological treatment, selected from Table 4 (the
facility codes are listed in the first column). For example, Table 8 gives the 1-day BOD limitation for
subcategory B and D as 36.55 mg/L, which is the mean of the values 34.31 mg/L for facility 500018-
410 and 38.78 mg/L for facility 50011-410 as shown in Table 3.
26
-------
o
•a
a
o
O)
(A
2
o
I
>
.Q
(0
1
ID
O
t
oa
O)
CO
(0
Q
O
O
o
CO
il
>
c
g _
CD 13)
I-
—1
= 0
a *3
J*
c
.0 .__
!~
£•
1 S
•— m
(0 LL.
^
c
_o _
ffl D>
_l
1
CO
x
TJ
^
to
^p
O
CO
i?
"D
1
^
•f
n
>
•a
1
6
^
•o
1
I
6
CO
>,
1
(R
•§
6
CO
^,
CO
*p
si
gj
CO
CO
•*
CD
cn
1*^
CO
*—
^
»—
cri
CO
"*
^.
CD
q
CO
in
en
d
CM
CD
o
0
d
in
-
CO
in
CM
in
CO
CM
co
CM
CO
_
^~
CO
^f
CM
O
II
CO
IV
CO
^
00
^
cn
T—
CM
CM
in
in
in
00
*~^
oo
CO
CM
in
cn
CO
CO
co
CO
CM
co'
cn
CD
CM
^j-
cn
CM
CM
CM
00
_
T^-
,J
^
CM
CO
o
§s
in in
CM
in
00
o
in
^
CM
i*s!
CM
IV
,_!
cn
IV
*
o
CM
CM
IV
in
^
0
CO
^J
o
CM
cn
CM
cn
in
CO
co
en
in
eo
CO
S
o o
o «-
in «t
5
CD
CD
CO
^*
00
CO
^>
T—
IV
o
in
00
CO
CD
'
cn
00
*-
in
cri
CM
O
CM
en
CD
IV
CO
O
Q
^J"
CO
CD
in
S
r—
CO
CM
CO
in
CM
in
0
o o
O CO
in
-------
Q
•a
i
CQ
I-
f
CO
p
^3
CO
i
co
CO
1
CO
I
(0
{2
Q
O
O
Q
CQ
^
11
'§ £
^
It
1-
lj
&
Is
03 O
J>
C
0 —
|I
2
£..
ll
>
§_
•a S
II
C "•"
-J
1
o
CO
ffl*
•o
T-
1
O
CO
^
•f
r-
1
CD
CO
(3
if—
^
(Q
Tp
O
CO
Si"
*D
*
1
6
CO
^
1
>•
•a
6
CO
^,
a
•o
| |
CM
in
CO
q
in
?
CO
00
en
CO
en
CO
0
in
CO
t~~
05
CO
CM
CD
C3
CO
CO
in
CM
CO
in
00
c>
*~
r-
CO
CO
*
00
o
0 0
o «-
in
^
r-.
tn
CM
CM
CM
in
CO
^
CO
CM
CM
•~
CO
Is*
CO
CO
*
in «t
J
o
CO
"
in
in
cd
CM
CM
CO
en
in
CO
_
CO
CO
in
00
en
m
CO
ro
^ji
^
*~
in
in
^
q
in
CO
CM
T^
*~
in
in
CO
CO
a J
S>
C
o
JO
03
o
o
a>
"5
CO
O)
ca
CD
CO
c
o
'4-1
JO
03
O
O
I
CO
D5
O
'o
a>
a.
- E
CO
CM
a.
£
o
o
CO
o
I
cu
CO
CO
§
JP^
CO
4—
o
^
a>
|
c-
4-J
CO
"o
CO
-------
-------
------- |