NC0000272_Analysis of Color Concentration 1-31-2018_20220401

ANALYSIS OF COLOR CONCENTRATIONS IN THE PIGEON RIVER
JANUARY 31, 2018
AQUAETER, INC.

SECTION 1.0
INTRODUCTION
Blue Ridge Paper Products Inc. doing business as "Evergreen Packaging" (Evergreen),
operates the Canton pulp and paper mill ("Canton Mill") in Canton, North Carolina. Since 1988,
the Mill has reduced color loadings to the Pigeon River by more than 90%.
Since 2001 the Canton Mill has sampled and analyzed True and Apparent Color in the
Pigeon River upstream from the Mill's discharge and approximately 0.4-mile downstream at the
Fibreville Bridge.' In 2009, Evergreen retained AquAeTer to sample and analyze True and
Apparent Color in the Pigeon River and its tributaries. Based on the Mill's data, and that collected
by AquAeTer, "A Comprehensive Assessment and Analysis of Color in the Pigeon River2" was
developed.
PURPOSE OF REPORT
This report supplements the Comprehensive Assessment and Analysis of Color in the
Pigeon River with an analysis of color concentrations in the Pigeon River upstream from the
Canton Mill at Park Street and downstream from the Mill's discharge at the Fibreville Bridge3.
The report compares the color concentrations at Fibreville and the difference in concentration
between the upstream and downstream stations from July 2010 through December 2017. This is
the period that the current NPDES permit and associated effluent limits have been in effect.
CONCLUSIONS
The long-term daily average contribution of the Canton Mill to true color in the Pigeon
River since July 2010, is 23 true color units at flows greater than or equal to the annual 30Q2.4
The long term daily average color concentration at the Fibreville Bridge at flows greater than or
equal to the annual 30Q2 is 34 true color units.5 The difference in the long-term daily average
color concentration at Park Street in Canton (14 true color units), upstream from the Canton Mill
1 NPDES Permit NC 0000272 issued to the Canton Mill by the North Carolina Department of Environmental
Quality requires the sampling. The Canton Mill also samples and analyzes color concentrations at Hepco (26 river
miles downstream) and at Brown's Bridge in Tennessee, approximately'/z mile downstream from the Duke Energy
Progress Hydro -electric plant.
2 AquAeTer, Inc. (2014). Comprehensive Assessment and Analysis of Color in the Pigeon River. Brentwood:
AquAeTer, Inc.
3 The report will also reference color concentrations at the NC / Tennessee state line and at a similar stream from the
same Ecoregion.
4 Table 5, page 10
5 Table 3, page 8
2

discharge, and the long term daily average color concentration at the North Carolina/Tennessee
state line (17 true color units), is 3 true color units.6
METHODOLOGY
AquAeTer has evaluated the color in the Pigeon River upstream from the Canton Mill
discharge and at the Fibreville Bridge. The dataset used for this analysis extends from July 1, 2010
through December 31, 20177, unless otherwise noted. The color data were also evaluated in
conjunction with the flows in the River. In particular, monthly flows were evaluated so that distinct
hydrologic/meteorologic periods that are statistically equivalent, but occur with the varying
background natural color that occurs in the River over the course of the different seasons, would
be analyzed separately but equivalently.
6 Table 4, page 9 and Table 6, page 11.
7 This is the period of the current NPDES Permit.
3

SECTION 2.0
ANALYSIS OF STREAM FLOWS
15A N.C.A.C. 2B.0206 (a) (5) specifies that aesthetic standards will be met at flows greater
than or equal to the minimum average flow for a period of 30 consecutive days that has an average
recurrence of once in two years (30Q2). The actual period for the calculation is not specified.
Historically, the period used has been an annual number, so an annual 30Q2 has been calculated
for the analyses in this report. However, the annual 30Q2 event is not reflective of the actual
stream conditions through the year for this stream. The chance for the annual 30Q2 to occur in
any given month for an equivalent 30-day period is much greater than a 2-year return interval,
which is not in keeping with the NC standard. The period that is presented in this report is for a
monthly basis, which more accurately reflects the hydrologic cycle over the course of the year. A
comparison of the monthly 30Q2's to the annual 30Q2 is presented in Figure 1.
Each of the monthly 30Q2
flows plotted in Figure 1 is
statistically equivalent to the other
months meaning they have the
same statistical probability of
occurring. However, the annual
30Q2 is not statistically likely to
occur in four months, and
represents a greater than a two-
year return interval in the other
eight months. The actual
calculated flows are provided in
Table 1. This demonstrates that
while the annual 30Q2 may
statistically have a 50% chance of
occurring in any given year, the
actual chance of the annual 30Q2
Table 1. 30Q2 Monthly and Yearly Flows and
Recurrence Intervals
MONTH
30Q2
(cfs)
Chance of Annual Flow
Occurring
(1 in _ Yrs)
(%)
January
390
57.1
1.8
February
420
>88
<1.1
March
457
>88
<1.1
April
419
>88
<1.1
May
288
>88
<1.1
June
207
57.9
1.7
July
149
15.1
6.6
August
127
4.4
22.8
September
133
4.5
22.3
October
172
4.2
23.9
November
207
9.7
10.3
December
285
44.0
2.3
Annual
90.5
2
50
flow occurring in any given month is much less.
The United States Environmental Protection Agency (USEPA) also recognizes that, in
some cases, streams need to and can be evaluated on a monthly basis8 due to the variable
hydrologic conditions found. This analysis is and has been applied in many southern states as well
8 US EPA. (1997). Technical Guidance Manual for Performing Wasteload Allocations, Book II: Streams and Rivers
- Part l: Biochemical Oxygen Demand/Dissolved Oxygen and Nutrients/Eutrophication. Washington, DC: US EPA.
4

as in some northern states and was originally proposed by the old U.S. Public Health Service, or
pre-USEPA.9
Figure 1. Comparison of Annual and Monthly 30Q2 Low Flows
500
450
400
co
z 350
300
250
200
g150
100
50
0
s
•
•
•
•
•
■ MENNEMEME
i 1 1 1 1 1 1 1 1 1 1 1
0 MONTH O o 0
♦Monthly 30Q2 ■Annual 30Q2 vt Z A
9 There are multiple NPDES Permits in the southeast which use monthly flow limits (DS Smith Riceboro on the
North Newport River and Georgia Pacific's Brunswick Mill on the Turtle River (Ga.)), (Georgia Pacific mill in
Naheola, Alabama on the Tombigbee River and West Rock Mill on the Tombigbee (Alabama)) and (Domtar Mill at
Ashdown, Arkansas).
5

SECTION 3.0
ANALYSIS OF COLOR CONCENTRATIONS
The Canton Mill has been collecting in -stream color data since 2001. AquAeTer collected
samples from 2009 through 2011 in the stream and in the tributaries. There are 1,601 days with
measurements collected at Fibreville since January 1, 2001. This report focuses on the period
beginning July 1, 2010, the effective date of the most recent NPDES Permit.'°
The analyses presented evaluate True Color in the River in various ways. The first is an
analysis of the samples collected at Fibreville. This location is approximately 0.4-mile
downstream from the Mill's outfall, and has been considered the compliance point or edge of
mixing zone.
The dataset for the period July 1, 2010 through December 31, 2017, as well as the entire
period of record was evaluated. The Comprehensive Assessment and Analysis of Color detailed
that concentration in the River is not related to flow, but the color loading or pounds per day can
be reasonably predicted based on the flow. The predicted mass loading can then be converted back
to a concentration in order to predict a concentration. However, the analysis in this report focuses
on measured concentration data.
EVALUATION OF DATA
The first evaluation is the proper identification of an average. The arithmetic mean is a
common way to calculate an average. An arithmetic mean is simply the sum of the individual
values divided by the number of values, or mathematically:
n
1=1
1
The arithmetic mean is strongly influenced by both large and small values. When a dataset
is normally distributed, the arithmetic mean is appropriate. However, when a dataset is skewed,
the arithmetic mean will no longer represent the central tendency of the dataset. Another average
is the median value.
10 This report presents an analysis of color concentrations for the period July 1, 2010 — December 31, 2017. The
specified term of the current NPDES Permit is July 1, 201 - June 30, 2015. However, Evergreen timely submitted its
renewal application on December 31 2014. As such, the current permit is administratively continued pursuant to N.C.
Gen. Stats. 150B-3 pending final action by NCDEQ on the permit.
6

A skewed dataset may follow a log -normal distribution. For these datasets, the mean and
median are calculated using the formulas:
Mean, Yin = exp
\i=1
exp (2
z
In 1 xi Ln 1 (x•• n x)2) * exp (En 1 (xi n x)
1J\
2
Median, zln = exp
i=1
Determination of Dataset Distribution
2
3
The evaluation of the full dataset that extends from January 1, 2001 to the present indicates
that most months follow a log -normal distribution. The goodness of fit results using the Anderson -
Darling (A-D) test for both normal and log -normal distributions for the full dataset and the most
recent dataset are presented in Table 2. When the result presented in the table is greater than 0.05,
relating to a 95% probability, the dataset fit the specified distribution. The best fit distribution for
each case is highlighted in Table 2. The appropriate calculations will be used for calculating the
monthly averages.
Table 2. Goodness of Fit for a 95% Probability for Dataset at Fibreville
MONTH
Normal Distribution
Log -Normal Distribution
Full Dataset
July 1, 2010 -
Present
Full Dataset
July 1, 2010 -
Present
January
0.00
0.03
0.23
0.16
February
0.00
0.00
0.20
0.34
March
0.02
0.20
0.12
0.22
April
0.00
0.00
0.40
0.10
May
0.00
0.00
0.08
0.29
June
0.00
0.00
0.22
0.37
July
0.00
0.90
0.19
0.01
August
0.00
0.00
0.21
0.24
September
0.00
0.00
0.00*
0.00*
October
0.00
0.00
0.25
0.60
November
0.00
0.01*
0.00*
0.00
December
0.00
0.04*
0.04*
0.01
Note:
* = Value does not pass A-D test, but this distribution was used.
7

EVALUATION OF IN -STREAM COLOR AVERAGES
Long-term Daily Average for each Month for In -Stream Color Concentrations Measured at
Fibreville
The long-term average of the daily concentration values for each month for the period from
July 1, 2010 through December 31, 2017 is presented in Table 3. This table demonstrates that the
color concentration measurements at Fibreville are less than or equal to 40 color units on a long-
term average for each month when flows are greater than or equal to the monthly 30Q2. The 90th
percentile column is presented since that method is used and accepted by EPA to determine if a
set of data for a pollutant parameter is meeting a numerical water quality standard. The table
shows that one month is as high as 61 color units at the 90th percentile, while the remaining months
are 53 color units or less. Though there is no North Carolina numerical water quality standard for
color, discussions of acceptable target values for the Pigeon River in the range of 50 to 100 have
been considered in the past. All of the values in Table 3 fall within or below this range.
Table 3. Long-term Daily Average Color Concentration for each Month for In -Stream
Measurements at Fibreville for July 1, 2010 through December 31, 2017
MONTH
DISTRI-
BUTION
TOTAL
NUMBER OF
VALUES
LONG-TERM
DAILY AVG.
WHEN FLOWS
>=A30Q2
(mg/L)
LONG-TERM
DAILY AVG.
WHEN FLOWS
>=M30Q2
(mg/L)
LONG-TERM
90th PER -
CENTILE
WHEN FLOWS
>=M30Q2
(mg/L)
Jan.
LN
61
27
18
29
Feb.
LN
57
27
19
30
Mar.
LN
64
26
19
27
Apr.
LN
60
22
20
29
May
LN
62
26
23
34
Jun.
LN
54
43
31
45
Jul.
N
70
48
30
40
Aug.
LN
72
46
40
61
Sep.
LN
70
51
39
53
Oct.
LN
70
45
35
51
Nov.
N
71
49
31
46
Dec.
N
71
36
25
38
Annual
LN
782
34
27
41
Note:
LN = Log Normal
A30Q2 = Annual 30Q2 Condition
M30Q2 = Monthly 30Q2 Condition
8

The values in Table 3 do not represent a separation of background color and color
contributed by the Mill. The long-term daily average Park Street measurements for each month
are presented in Table 4.
Table 4. Long -Term Daily Average Color Concentration at Park Street Upstream for
July 1, 2010 through December 31, 2017
MONTH
DISTRI-
BUTION
TOTAL
NUMBER OF
VALUES
LONG-TERM
DAILY AVG.
WHEN FLOWS
>=A30Q2
(mg/L)
LONG-TERM
DAILY AVG.
WHEN FLOWS
>=M30Q2
(mg/L)
LONG-TERM
90th PER -
CENTILE
WHEN FLOWS
>=M30Q2
(mg/L)
Jan.
LN
61
11
12
22
Feb.
LN
57
10
12
21
Mar.
LN
64
10
11
19
Apr.
LN
60
12
15
28
May
LN
62
12
12
18
Jun.
LN
61
16
15
27
Jul.
LN
70
15
14
21
Aug.
LN
72
15
15
25
Sep.
LN
69
17
19
29
Oct.
LN
70
18
21
31
Nov.
LN
70
16
17
25
Dec.
LN
71
15
18
33
Annual
LN
787
14
15
26
Note:
LN = Log Normal
A30Q2 = Annual 30Q2 Condition
M30Q2 = Monthly 30Q2 Condition
Table 4 shows the variable nature of color in the Pigeon River upstream from the Mill. In
some months the expected long term daily average when flows are greater than or equal to the
Monthly 30Q2 are as high as 21 color units. The 90th percentile values in some months are as high
as 33 color units. Background concentrations have been measured as high as 94 mg/L in the Pigeon
River upstream from the Mill." This means that no numerical water quality value below this can
be a realistic target even if the Mill were not located on the river.
Determination of In -Stream Color Delta at Fibreville
The Pigeon River coming to the Mill in Canton has been measured for color on a routine
basis. Because of the variability in background color as shown in Table 4, an in -stream delta
between the background and the downstream location is more appropriate to evaluate the Mill's
contribution to color in the River. Similarly to the in -stream values, the majority of the long-term
11 December 24, 2015 at Park Street Bridge in Canton
9

daily average delta concentrations for each month do not fit a normal distribution. The long-term
daily average in -stream delta concentrations for each month are provided in Table 5.
This table demonstrates that for any month, the long term daily average contribution of the
Canton Mill to color at Fibreville when flows are greater than or equal to monthly 30Q2 is 25 true
color units or less. The 90th percentile of color contribution of the Canton Mill does not exceed 43
true color units for any month when flows are greater than or equal to the monthly 30Q2.
Table 5. Long -Term Daily Average In -Stream Delta Color Concentration at Fibreville
Bridge downstream for Each Month for July 1, 2010 through December 31, 2017
MONTH
DISTRI-
BUTION
TOTAL
NUMBER OF
VALUES
LONG-TERM
DAILY AVG.
WHEN FLOWS
>=A30Q2
(mg/L)
LONG-TERM
DAILY AVG.
WHEN FLOWS
>=M30Q2
(mg/L)
LONG-TERM
90th PER -
CENTILE
WHEN FLOWS
>=M30Q2
(mg/L)
Jan.
LN
61
16
6
11
Feb.
LN
57
17
7
14
Mar.
LN
64
16
8
13
Apr.
LN
60
10
5
11
May
LN
62
14
11
21
Jun.
LN
53
28
17
25
Jul.
N
70
33
16
29
Aug.
LN
72
32
25
43
Sep.
N
69
34
21
41
Oct.
LN
70
27
14
30
Nov.
N
70
32
14
24
Dec.
N
71
21
7
17
Annual
LN
779
23
13
28
Note:
LN = Log Normal
A30Q2 = Annual 30Q2 Condition
M30Q2 = Monthly 30Q2 Condition
Determination of Average Color on the French Broad River and Comparison to the Pigeon
River at or near the North Carolina — Tennessee state line
For comparison purposes, the color measurements collected on the French Broad River
upstream from Newport, Tennessee are also included in Table 6. It is noted that the difference
between the upstream and downstream locations on the French Broad River and on the Pigeon
River are both 3 color units. This is consistent with the general expectation that color increases in
the downstream direction as stream order increases.
The Tennessee Department of Environment and Conservation collects samples on the
Pigeon River near the state line, as well as, on the French Broad River at the Highway 70 Bridge
10

(near Del Rio, TN) and at the Highway 321 Bridge (near Newport, TN). There are not enough
data points to calculate long-term daily averages for each month, nor to determine the proper
distribution of the datasets. Overall averages of the three datasets are presented, along with the
number of samples in each calculation.
Table 6. Comparison of Color Concentration for In -Stream Measurements on Pigeon
and French Broad Rivers
PARAMETER
PIGEON RIVER
AT PARK
STREET IN
CANTON
PIGEON RIVER
NEAR NC/TN
STATE LINE
FRENCH
BROAD RIVER
NEAR DEL RIO,
TN
FRENCH
BROAD RIVER
NEAR
NEWPORT, TN
Drainage Area (sq.
mi.)
130
538
1,713
1,858
Number of Samples
787
232
30
24
Samples Collected
by
EPI
TDEC
TDEC
TDEC
Average Color,
mg/L
14
17
22
25
Minimum Color,
1
5
8
7
mg/L
Maximum Color,
mg/L
94
43
63
110
Note: Both French Broad TDEC stations are upstream from the confluence of the Pigeon
River and the French Broad River near Newport, Tennessee.
The Pigeon River near the state line and the French Broad River near Del Rio are especially suited
for comparison purposes since 1) they both drain the identical ecosystem classification; 2) they are
both level 5 on the Schuler stream classification system (stream order); and 3) the French Broad
near Del Rio has no known large sources of color discharging to it. The comparison shows that
the Pigeon River near the state line actually has less color than the French Broad River near Del
Rio when compared to an equivalent stream system (i.e., same ecosystem and same order stream).
This indicates the Pigeon River at the state line does not have color levels outside of the expected
range for a stream with no large point source of color discharging to it. These data also show the
difference in color at the Park Street Bridge on the Pigeon River and the color at the state line on
the Pigeon River is only 3 color units which is outside the range perceptible by the human eye.
11

SECTION 4.0
FINDINGS AND CONCLUSIONS
FINDINGS FOR THE PERIOD JULY 1, 2010 THRU DECEMBER 31, 2017:
The long-term daily average color concentration at the Fibreville Bridge when flows are greater
than or equal to the annual 30Q2 is 34.
The long-term daily average contribution of the Canton Mill to true color in the Pigeon River is
23 true color units when flows are greater than or equal to the annual 30Q2.
Based on a methodology used by and recognized by EPA, 30Q2 values for each month of the year
for the Pigeon River more accurately reflects the hydrologic cycle over the course of the year.
Color concentration measurements at Fibreville are less than or equal to 40 color units on a long-
term average for each month when flows are greater than or equal to the monthly 30Q2.
The 90th percentile of the long-term daily color concentrations at Fibreville when flows are equal
to or greater than the monthly 30Q2, ranged from 27 in March to 61 in August. (90th percentile
was examined from EPA's guidance on comparing a set of data to a specific water quality standard
i.e. if the 90th percentile is less than a target value, then water quality standards are being met).
The long term daily average color concentration added, i.e., the delta, to the Pigeon River by the
Mill, at flows greater than or equal to the monthly 30Q2, is less than 25 color units for any month
except August. Also, the 90th percentile of the color concentration added by the Mill, at flows
greater than or equal to the monthly 30Q2, is less than 43 color units for any month except August.
The color concentrations in the Pigeon River upstream from the Mill are variable. In some months
the expected long term daily average color concentration, when flows are greater than or equal to
the monthly 30Q2, are as high as 21 color units; and 90th percentile values in some months are as
high as 33 color units. Also, color concentrations upstream from the Mill have been measured as
high as 94 color units. This means that no numerical water quality value below this can be a
realistic target even if the Mill were not located on the river.
OTHER FINDINGS:
Comparison of the French Broad River near Del Rio shows the Pigeon River at the Stateline does
not have color levels outside of the expected range for a stream with no large point source of color
discharging to it.
12

The difference in the long-term daily average color concentration at Park Street in Canton,
upstream from the Canton Mill discharge, and the long-term daily average color concentration at
the North Carolina -Tennessee Stateline is 3 true color units.
CONCLUSIONS:
The long term daily average contribution (delta) of the Canton Mill to true color in the Pigeon
River is 23 color units at flows greater than or equal to the annual 30Q2. The long term daily
average color concentration at the Fibreville Bridge at flows greater than or equal to the annual 30
Q2 is 34 true color units. The difference in the long term daily average color concentration at Park
Street in Canton (14 true color units), upstream from the Canton Mill discharge and the long term
daily average color concentration at the North Carolina -Tennessee state line (17 true color units)
is 3 true color units. This difference is minimal and imperceptible to the human eye.
Though there is no North Carolina numerical water quality standard for color, discussions of
acceptable target values for the Pigeon River in the range of 50 to 100 have been considered in the
past. The long-term average and 90th percentile values at Fibreville and the long-term average and
90th percentile contributed by the Mill meet this target range.
13