DEVELOPMENT OF A WASTELOAD
ALLOCATION MODEL FOR THE PIGEON RIVER
BETWEEN CANTON AND HEPCO,
NORTH CAROLINA
VOLUME I - TEXT
Versnit
-------
DEVELOPMENT OF A WASTELOAD
ALLOCATION MODEL FOR THE PIGEON RIVER
BETWEEN CANTON AND HEPCO,
NORTH CAROLINA
VOLUME I - TEXT
Prepared for
U.S. Environmental Protection Agency
Region IV
345 Courtland Street
Atlanta, GA 30365
Prepared by
J. Kevin Summers
Paul F. Kazyak
Stephen B. Weisberg
Harold T. Wilson
Versar, Inc.
9200 Rumsey Road
Columbia, MD 21045
July 1989
-------
ACKNOWLEDGEMENTS
We gratefully acknowledge the cooperation of Champion
International Corporation during our study, including: Paul
Wiegand, Bill Chapman, and Mary Lee Ransmeier. We also
acknowledge Randy Dodd and Max Haner of NCNRCD for their helpful
support and contributions to the study. Thanks are also due to
Jim Greenfield, Tom Plouff, Forrest Leedy, and Kay Harris of EPA
Region IV for their participation in the field collections. In
addition, we appreciate the efforts of Gene Barker of the USGS
for timely help in providing flow data from the Canton and Hepco
gages.
i i i
-------
EXECUTIVE SUMMARY
A QUAL2E-UNCAS wasteload allocation model was constructed to
examine the relative impact of three point source inputs
(Champion Paper Mill, Clyde WWTP, and Waynesville WWTP) on
dissolved oxygen (DO) dynamics of the Pigeon River between Canton
and Hepco, North Carolina. To calibrate the model, field
measurements of dissolved oxygen, biochemical oxygen demand
(BOD), sediment oxygen demand, (nitrogen, phosphorus,
chlorophyll, flow, and light penetration were collected during a
low flow period in September 1988. Sampling was conducted at 19
stations, including the three effluents and four tributaries
entering the modeled region. The calibrated model was validated
using two data sets — the paper mill's riverwide self-monitoring
data from September 1988, and data from a synoptic water quality
survey of the Pigeon River conducted under higher flow conditions
in July 1987. The validated model was then used to simulate five
alternative loading scenarios to ascertain the impact of these
alternatives on Pigeon River DO dynamics.
The field studies identified a pattern of DO in the river
that was similar to patterns observed during previous summer
monitoring by the paper mill. Water entered the upstream border
of the study area near saturation, passed through the Champion
Mill, and exited 15 °C warmer and supersaturated by means of
artificial oxygenation. Proceeding downstream, DO levels
alternately increased and declined in response to two sidestream
oxygenation units and the associated rapid deaeration below them.
Downstream of the oxygenators, DO levels steadily declined,
reaching the lowest value near the river's confluence with
Richland Creek, about halfway through the modeled region.
Results of both the field study and the modeling efforts
demonstrated that the effluent from the Champion Paper Mill,
which has more than ten times the volume of the two WWTP's
combined, was the most important of the three point sources in
regulating Pigeon River DO dynamics. Simulated reduction of the
mill effluent by 55% increased DO at the low point in the river
by more than 1 ppm. in contrast, simulated removal of the WWTP's
from the. river resulted in no measurable change in river DO.
The sidestream oxygenation units maintained by the Champion
Mill were found to maintain state water quality standards in the
river during the time of our study, as no DO values less than 6
ppm were measured. However, violations in state water quality
standards for DO have been historically detected, suggesting that
conditions during our study did not represent worst case.
Simulated removal of these units led to a 2 ppm decline in DO
upstream of Richland Creek. However, the diluting effect of
v
-------
Richland Creek was found to be more important in regulating DO in
the river downstream of Richland Creek, as simulated removal of
the oxygenation units was found to have no effect on DO
downstream of Richland Creek.
Although ultimate carbonaceous BOD exceeded 50 mg/1 from the
Champion Mill effluent and 25 mg/1 from the two WWTPs, these
outputs did not have a major impact on DO dynamics in the river,
primarily because of the short travel time (~ 2 days) through
the study area. Presumably these high BOD loads are degraded
within Walter's Lake, which forms the downstream border of our
study area. Instead, nitrogenous demand, in the form of
oxidation from ammonia, appeared to be the dominant source of
impact on the river from the three effluent sources. Despite
abundant nitrate concentrations, photosynthetic effects on DO
were found to be negligible, presumably due to high light
extinction coefficients within the river.
vi
-------
TABLE OF CONTENTS
Page
VOLUME I - TEXT
ACKNOWLEDGEMENTS iii
EXECUTIVE SUMMARY V
I. INTRODUCTION 1-1
II. DESCRIPTION OF THE PIGEON RIVER WATERSHED II-l
III. EMPIRICAL STUDY METHODS III-l
IV. EMPIRICAL STUDY RESULTS IV-1
V. MODELING METHODS V-l
A. MODEL STRUCTURE V-l
B. MODEL VALIDATION V-13
C. MODEL UNCERTAINTY V-14
VI. MODEL RESULTS VI-1
A. CALIBRATION VI-1
B. VALIDATION VI-15
C. MODEL UNCERTAINTY VI-29
VII. PIGEON RIVER MODEL SCENARIOS VII-1
VIII. DISCUSSION.... VIII-1
IX. LITERATURE CITED IX-1
VOLUME II - APPENDICES
APPENDIX A
PIGEON RIVER QUAL2E-UNCAS CALIBRATION MODEL RUN
APPENDIX B
PIGEON RIVER QUAL2E-UNCAS VALIDATION MODEL RUN
28 SEPTEMBER 1988 DATA
APPENDIX C
PIGEON RIVER QUAL2E-UNCAS VALIDATION MODEL RUN
7 JULY 1987 DATA
vi i
-------
TABLE OF CONTENTS (CONTINUED)
APPENDIX D
RESULTS OF UNCERTAINTY ANALYSES WHEN VARYING ALL
POINT SOURCE LOADS BY 10-15% WITHOUT MODIFYING
POINT SOURCE DISCHARGE RATES
APPENDIX E
RESULTS OF UNCERTAINTY ANALYSES WHEN VARYING ALL
HEADWATER WATER QUALITY CONCENTRATIONS BY 10-15%
AND HEADWATER FLOW RATES BY 6%
APPENDIX F
RESULTS OF UNCERTAINTY ANALYSES WHEN VARYING ALL
REACTION COEFFICIENTS AFFECTING DISSOLVED OXYGEN
CONCENTRATIONS BY 3-20%
APPENDIX G
RESULTS OF UNCERTAINTY ANALYSES WHEN VARYING ALL
THE REACTION COEFFICIENT FOR THE DECAY OF CARBONACEOUS
MATERIALS BY 15%
APPENDIX H
RESULTS OF UNCERTAINTY ANALYSES WHEN VARYING ALL
REACTION COEFFICIENTS AFFECTING NITROGEN OR PHOSPHORUS
CONCENTRATIONS BY 3-20% AND REACTION COEFFICIENTS
AFFECTING NITROGENOUS OXIDATION BY 10%
APPENDIX I
RESULTS OF UNCERTAINTY ANALYSES WHEN VARYING ALL
HEADWATER WATER QUALITY CONCENTRATIONS BY 10-15%
WITHOUT MODIFYING HEADWATER FLOW RATES
APPENDIX J
RESULTS OF UNCERTAINTY ANALYSES WHEN VARYING ALL
REACTION COEFFICIENTS, MODEL PARAMETERS, AND POINT
SOURCE DISCHARGES BY 3 — 20WITHOUT MODIFYING
STREAMFLOW
APPENDIX K
PIGEON RIVER QUAL2E-UNCAS MODEL SIMULATION OF THE
EFFECTS OF REMOVING THE CLYDE AND WAYNESVILLE WWPTs
DISCHARGES
-------
TABLE OF CONTENTS (CONTINUED)
APPENDIX L
PIGEON RIVER QUAL2E-UNCAS MODEL SIMULATION OF
THE EFFECTS OF ASSIGNING MAXIMUM SECONDARY
TREATMENT STANDARDS AT THE WAYNESVILLE WWTP
APPENDIX M
PIGEON RIVER QUAL2E-UNCAS MODEL SIMULATION OF
THE EFFECTS OF REDUCING THE CHAMPION DISCHARGE
FLOW TO 30 MGD (i.e., 46.4 CFS) WHILE MAINTAINING
THE PRESENT CONCENTRATIONS IN THE CHAMPION
EFFLUENT
APPENDIX N
PIGEON RIVER QUAL2E-UNCAS MODEL SIMULATION OF THE
COMBINED EFFECT OF MAXIMUM SECONDARY TREATMENT
STANDARDS AT THE WAYNESVILLE WWTP AND REDUCTION
OF THE CHAMPION DISCHARGE TO 30 MGD
APPENDIX 0
PIGEON RIVER QUAL2E-UNCAS MODEL SIMULATION OF
REMOVING THE TWO SIDESTREAM OXYGENATORS
WP74:4205
-------
I. INTRODUCTION
During low flow periods in summer, portions of the Pigeon
River between Canton and Hepco, North Carolina have historically
experienced depressed dissolved oxygen (DO) levels. The severity
of the DO depression is such that on 21 occasions between 1985
and 1987, DO levels were below the minimum DO standard (4 ppm
minimum) established by the State of North Carolina (Randall
Dodd, North Carolina Department of Natural Resources and
Community Development, Raliegh, NC, pers. comm.). These
violations have occurred despite the presence of oxygenation
systems at several locations in the river. At present, three
point source dischargers are thought to be primarily responsible
for the DO depletion: Champion International's Canton Mill,
Waynesville Wastewater Treatment Plant (WWTP), and Clyde WWTP.
However, the relative contribution of each of the facilities to
the low DO events is not well documented.
Several studies have been conducted to describe factors
affecting DO in the Pigeon River, but attempts to verify exist-
ing models have met with limited success (North Carolina Division
of Environmental Management (NCDEM) 1984). The goal of this
study is to develop a Qual2E-UNCAS wasteload allocation model for
the Pigeon River that can be used to document the relative
importance of various inputs to the river on DO levels, and to
evaluate the consequences of several water quality management
options. A four-step approach is used to achieve the study goal.
o Conduct field studies to quantify factors that affect
oxygen content of the Pigeon River
o Calibrate and validate a QUAL2E model for the desig-
nated portion of the Pigeon River using new and/or
existing data
o Gauge model uncertainty
o Run the model to examine effects of alternative
management strategies on water quality in the Pigeon
River
1-1
-------
II.
DESCRIPTION OF THE PIGEON RIVER WATERSHED
The Pigeon River originates in a mountainous area of western
North Carolina and flows in a northwesterly direction to its
confluence with the French Broad River near Newport, Tennessee
(Fig. Il-l). Elevation gradients in the watershed vary
substantially from reach to reach, with elevations ranging from
nearly 2,000 m at the headwaters to 306 m at the French Broad
River confluence. From its headwaters to the confluence of its
East and West Forks, the Pigeon River descends nearly 1200 m.
For the remaining 111 km of its length, river descent is less
rapid, with,an average drop of 4.6 m/km. In the river reach
between Canton and Crabtree Creek, the gradient is 1.9 m/km. In
the reach below Crabtree Creek, the gradient increases to 3.4
m/km until Jonathans Creek. Below Jonathans Creek, the gradient
is 4.9 m/km until the Pigeon River enters Walters Lake, just
below Hepco.
Physically, the Pigeon River between Canton and Hepco is a
rocky, shallow, warmwater stream, predominated by short riffles
and long runs. River substrates are generally cobble/boulder,
with exposed bedrock in high gradient areas, and combinations of
sand and organic matter overlying rocks in slower depositional
areas. During low flows, most of the river is less than 1 m in
depth, with a maximum river depth of approximately 3 m recorded
at River Mile 61.5 (Nisely and Tysland 1983). At Canton (River
Mile 64.6), North Carolina's classification of the Pigeon River
shifts from coldwater trout stream to warmwater fishery/non-
contact recreation stream, with water temperatures in excess of
30' C recorded below the Champion Mill outfall during summer.
2
The Pigeon River watershed encompasses a total of 1725 km ,
more than one-half of which is above Hepco, North Carolina.
Four tributaries: Richland Creek (177 km2), Crabtree Creek (69
km2), Jonathans Creek (175 km2), and Fines Creek (66 km2)
contribute more than 85% of the areal increase in watershed size
between Canton and Hepco. Six smaller tributaries account for
the remaining increase in watershed size: Beaverdam Creek,
Thickety. Creek, Murray Branch, Bowen Branch, Chambers Branch, and
Mill Branch.
Mean annual flows in the Pigeon River increase from 319 cfs
at the Canton USGS gauge to 669 cfs at the Hepco USGS gauge, with
recorded minimum flows of 27 cfs at Canton and 81 cfs at Hepco.
In contrast, maximum flood flows at Hepco have been estimated at
42,000 cfs. Two reservoirs regulate flows on the Pigeon River
between Canton and Hepco — Lake Logan on the West Fork of the
Pigeon River, and Lake Junaluska on Richland Creek. Lake Logan
Il-l
-------
Walters Power Plan)
Nortn Carolina
Walervilie
Big Creek
Pigeon flivef
Walters Dam'
Cataioochee Creek
Walters Lake
Jonamans Creek
Pigeon River
Canton Mill
Clyde
Canton
Lake Junaluska
Lake Junaluska
Richland Creek
East Fork Pigeon River
W#$t Fork Pigeon Riv#r
Little East Fork
Pigeon River
Approximate Scale
t/2 m a i Mile
Figure 11 — 1 . Map of the Pigeon River, North Carolina (from
Weston 1983)
11-2
-------
is owned by Champion International. Discharge from the lake is
regulated to provide sufficient;, water to operate the Canton Mill
throughout the year. Lake Junaluska is owned by a non-profit
religous organization and discharge is apparently run-of-river
(Max Haner, North Carolina Department of Natural Resources and
Community Development, Asheville, NC, pers. comm.).
Land use in the Pigeon River watershed above Hepco is
predominately agricultural or undeveloped; approximately 72% of
the area is forested and 18% is crop/pasture land. Other areas
comprise only 10% of the watershed, the largest fraction being
urban (7.7%). Drainage from the upper watershed (above Canton)
is used as a public water supply by the town of Canton, and is
also used by the Champion paper mill. Withdrawals from the river
for public water are on the order of 2 cfs, while withdrawals
during full mill production are approximately 70 cfs at the paper
mill. There are a total of 60 discharge permits issued by the
State of North Carolina for the Pigeon River watershed above
Hepco, but only four facilities (Champion Mill, Waynesville WWTP,
Clyde WWTP, and Maggie Valley WWTP are required to comply with
federal NPDES regulations. Smaller dischargers are primarily
located in developed areas within the Jonathans Creek and
Richland Creek subbasins.
Dissolved oxygen (DO) is artificially introduced to the
Pigeon River at three locations: Champion Mill effluent, and
also at sites 0.9 and 2.1 miles downstream of the Mill discharge.
Champion's Mill effluent is oxygenated by means of a domed
aeration cascade which introduces pure oxygen to Mill effluent
just prior to its discharge into the river. At the downstream
locations, DO is added to the river via sidestream oxygenation
units. At each unit, gaseous oxygen is mixed with river water
under 60-70 lbs of pressure. Eductors mix the highly oxygenated
water with upstream water at a 1:3 ratio, and the mixture is
returned to the river. The total volume of water withdrawn and
returned to the stream at each sidestream location is
approximately 18.3 cfs.
II-3
-------
III. EMPIRICAL STUDY METHODS
Sampling in the Pigeon River was conducted September 27-28,
1988 between Canton and Hepco, North Carolina (Fig. III-1, Table
III-1). Samples were collected at 18 stations to provide the
following categories of data needs for the Qual2E-UNCAS model:
modeled variables, process rates, physical driving functions,
input loads, and conservative tracers.
The 18 sampling locations were selected to measure all
major inputs to the river within the study area. Stations were
located upstream of the confluence, and downstream of all major
tributaries and discharges (Fig. Ill-l). Tributaries were
sampled as close to their confluence with the Pigeon River as
access allowed, and effluent samples were collected at or near
the end of each discharge pipe. Mainstem stations located below
inputs were sited approximately 0.5 miles downstream to ensure
mixing of inputs with river water. An exception was the station
downstream of Fines Creek (S11), where the Pigeon River enters
Walters Lake before complete mixing of water from Fines Creek
occurs. Water samples for chemical analysis were collected from
the middle of the channel just below the surface. Dissolved
oxygen concentration (DO) and temperature were measured near the
middle of the channel at the surface and bottom, and near the
river margin (shoal) at a midwater depth.
The 18 sampling stations were separated into 14 nominal
stations, at which most parameters were measured, and four
intensive stations which were sampled for the same array of
parameters as nominal stations, but also included measurement of
process rates and diurnal variability (Fig. Ill-l). Locations of
intensive stations were sited to provide:
o A site above all effluents to define "background"
diurnal variability (Si)
A site downstream of the
unit to document diurnal
paper mill effluent (S4)
last Champion Mill aerator
variation associated with the
A site downstream of the Clyde and Waynesville WWTPs to
establish diurnal variation due to combined effluent
loadings (S6)
A site several miles downstream of
identify diurnal patterns in water
Lake (S10).
the DO sag zone to
entering Walters
III-l
-------
Fines Creek
to Walters Lake
T 4
Scale of Miles
S 9
T 3
T 2
Jonathans Creek
S 7
Crabtree Creek
Sidestream
Oxygenatoi
S 6
S 5
E3 (Waynesvllle WWTP)
'S3
Richland Creek
(Champion Mill)
E 2
(Clyde WWTP)
Lake Junaluska
(§) Intensive Station
Figure III-l. Nominal and intensive sampling stations in the
Pigeon River, North Carolina study area
111-2
-------
Table III-l. Description of station locations within the
Pigeon River, North Carolina study area
Station River Mile
Description
SI*
52
53
54 *
55
S6*
57
58
59
S10'
Sll
Tl
T2
63.5
62.8
61.5
59 .2
55.2
53.8
49.8
48.7
47.6
43.0
42.6
54.9
49.8
Headwater, at Champion Mill intake
Downstream of Champion effluent,
at Fiberville bridge
Between first and second sidestream
oxygenators, just upstream of
second oxygenator
Between second oxygenator and Clyde
WWTP effluent, at Thickety Road
divergence from river
Upstream of Richland Creek, at
pollution control project site
Downstream of Waynesville effluent
and Richland Creek, at bridge off
Rt. 209
Upstream of Crabtree Creek, on
Riverside Road
Downstream of Crabtree Creek, on
Riverside Road
Upstream of Jonathans Creek, on
Riverside Road
Upstream of Fines Creek at 1-40
bridge, access from 1-40
Downstream of Fines Creek, access
from dead end road on west bank
Richland Creek, on Hyder Mountain
Road bridge, creek mile 0.2
Crabtree Creek, off Riverside Road,
creek mile 0.05
*Indicates intensive sampling station
111-3
-------
Table III-l. Continued
Station River Mile
Description
T3
T4
El
E2
E3
46.0
42.7
63.3
57 .1
54.8
Jonathans Creek, on White Oak Road
below Dark Hollow Road, creek
mile 0.8
Fines Creek, at bridge at Panther
Creek Road, creek mile 0.3
Champion paper mill effluent
Clyde WWTP effluent
Waynesville WWTP effluent
111-4
-------
Sampling Parameters
Nominal stations were sampled twice on 27 September; once
between 0800 and 1130 and once between 1330 and 1530. Nominal
stations were sampled for temperature, flow rate, dissolved
oxygen, nutrients, chloride, five-day biochemical oxygen demand
(BOD), ultimate carbonaceous biochemical oxygen demand (UCBOD)
and chlorophyll during the morning sampling period (Table II1—2).
In the afternoon, nominal stations were sampled for 5-day BOD,
dissolved oxygen, temperature, and flow. Flow measurements at
mainstem and tributary stations were made prior to and after the
main portion of the study on 27 September. During the main
portion of the study, staff gauge readings were recorded each
time a station was sampled to document constancy in flows during
the study.
Intensive stations were sampled five times throughout the
day to determine diurnal patterns in the various parameters.
Intensive stations were sampled for temperature, flow rate,
dissolved oxygen, nutrients (NO^-N, NO^-N, Total Kjeldal
nitrogen, total phosphorus, ortno-phosphate), 5-day BOD, and
chlorophyll on a diurnal basis (0700, 1300, 1700, 2030, and 2400)
(Table III-2). Chloride and UCBOD were sampled only at 0700.
Water column photosynthesis and sediment oxygen demand
(SOD) were also measured on 27 September at three of the four
intensive stations (S4, S6, S10). SOD and photosynthesis were
not measured at SI because process rates at the upstream boun-
dary of the model (Si) are not incorporated into the model.
Instead, process rates were measured at S2, immediately below the
Champion discharge. Water column photosynthesis was measured
between 1300 and 1800 and SOD was measured in three consecutive
tests between 1300 and 2400. Two additional SOD measurements
(one light and one dark chamber) were taken simultaneously from
1200 to 1600 on 28 September at station SI to document
differences in SOD due to photosynthesis at the upstream study
area boundary.
Light attenuation measurements were made at all mainstem
stations (Sl-Sll) on 27 September between 1330 and 1930. In
addition, percent shading was visually estimated at mainstem
stations on 28 September.
In addition to dissolved oxygen (DO) and temperature data
collected as part of the main study on 27 September, a longi-
tudinal DO/temperature survey was conducted on 28 September at
all stations along the mainstem. In conjunction with the
longitudinal survey on 28 September, deaeration measurements were
made from 1300 to 1500 below the sidestream oxygenator at RM
61.5.
111-5
-------
Table III-2 Parameters/sampling times for intensive nominal
and special study locations sampled in the Pigeon
River study area during September 1988
Intensive Stations Nominal Stations
27 Sept 28 Sep 27 Sept 28 Sep
Parameter 0700 1300 1700 2030 2400 0800-1130 1330-1530
Flow
/
/
/
~
/
/
Dissolved Oxygen
/
~
~
~
~
/
/ ~
~
Temperature
~
~
~
~
~
/
/ ~
BOD5
~
/
/
/
~
/ /
Ultimate BOD
~
~
SOD
~
~
~
/
Atmospheric Exchange
/
~
Light Attenuation
/*
Percent Shading
/
~
Chlorophyll a
/
/
Photosynthesis
/
/
Total Kjeldahl
Nitrogen
/
/
~
~
~
/
Ammonia-nitrogen
~
~
~
/
~
/
Nitrate-nitrogen
~
/
~
/
~
~
Nitrite-nitrogen
~
~
~
/
/
Total Phosphorus
/
~
/
~
/
/
Ortho-phosphate
~
~
~
/
~
/
*Mainstem stations cnly.
111-6
-------
Measurement Methods
Flow
Flow at mainstem and tributary stations was measured using
standard USGS techniques (Buchanan and Somers 1969). At each
station, a transect was established across the stream. Depth and
velocity measurements were made at intervals of 0.5-3.0 m,
depending on stream width. Velocity measurments were taken at
each point on the transect with a Pygmy-Price current meter at a
depth below the surface equal to 0.6 times the total stream
depth. Staff gauges were installed at mainstem and tributary
stations and corrected with stream flow measurements. Stream
heights were recorded each time a station was sampled as a means
of monitoring changes in flow throughout the study. Gauging
station data from seven USGS sites within the study area were
also obtained. Paper mill plant intake and effluent discharge
rates were obtained from data reported to the NCDEM by the
Champion Mill and the WWTPs at Clyde and Waynesville.
Additionally, withdrawal rates for the town of Canton's water
supply were obtained from the town of Canton.
Dissolved Oxygen/Temperature
Dissolved oxygen concentration and temperature were measured
using a series of electronic meters, including Yellow Springs
Instruments (YSI) models 55, 56, 57, and 58, and Hydrolab
Surveyor II. To insure validity of DO and temperature
measurements, each meter was calibrated with Winkler titrations
prior to the sampling day, checked or recalibrated before the
afternoon sampling, and checked again after the 2400 sampling.
Additional calibration checks were performed at intermediate
times by a mobile quality assurance team that compared
measurements taken with a calibrated meter with those made with
the meters assigned to specific stations. Whenever measurements
from two meters differed by more than 0.4 ppm, an air calibration
check was performed on each meter. Meters which differed from
air saturation by more than 0.4 ppm were removed from use until
they could be recalibrated.
Biochemical Oxygen Demand
Two types of biochemical oxygen demand (BOD) samples were
collected during the study, five-day and ultimate carbonaceous
Five-day BOD was measured using EPA standard method 405.1 (EPA
111 -7
-------
1983) and UCBOD was calculated according to procedures estab-
lished by North Carolina Department of Environmental Monitoring
(NCDEM).
Sediment Oxygen Demand
Sediment Oxygen Demand (SOD) was measured in recirculating
chambers similar to the design used by Hickey (1988) (Figure
III-2). At each station sampled, chambers were placed at a
location with substrate composition and bottom topography which
were representative of the reach. An exception was S10, where
swift, deep water and a highly irregular bottom limited chamber
locations to a shallow pool area. After installation, sandbags
were positioned around the periphery of the chamber to provide a
watertight seal. Chambers were cleared of trapped air and
disturbed sediments by running the circulation pump with the
chamber ports opened. Chambers were then sealed and water within
the chamber was recirculated at a constant velocity with a
13,240 1/h pump for 1.5 to 4.5 h. Temperature and DO were
measured at approximately 15 min intervals with a YSI DO probe
installed in the chamber. Total oxygen consumption rates were
calculated by linear regression of DO vs. time for each
experiment. When regressions were non-significant at the a =
0.05 level, oxygen consumption was assumed equal to zero. Water
column respiration, as measured by 5-day BOD at each site, was
subtracted from total oxygen consumption to calculate SOD.
Photosynthesis
Photosynthesis rates were estimated empirically by direct
measure of oxygen production and consumption rates, and were also
estimated indirectly from other parameters measured. Empirical
measurements were made by measuring DO levels in replicate light
and dark bottles which were incubated iji situ for 3-4 h. DO
measurements were made with a YSI DO meter equipped with a BOD
probe.
Chlorophyll a was collected by filtering 100-200 ml of
sample and freezing the filters in the field. In the laboratory,
frozen filters were ground in acetone and examined fluorometri-
cally (Loftus and Carpenter 1971).
Light attenuation measurements were made at approximately
10 cm intervals with a Li-Cor model LI-185A photometer mounted on
a meter stick. Light extinction curves were then constructed
based on percent of available light at depth.
111-8
-------
Sediment Oxygen Demand (SOD) Chamber
DO/Temperature
Probe Location
Return Baffle
12v DC 3500gph Pump.
(+)
(-)
Louvered Horizontal &
Vertical Baffles
• Chamber Vent Plugs
T
0.2 m
Substrate Sampling Area
Supply Line
Throttle Valve
Top View
Flow Diffuser
Flow Bypass
DO/Temperature
Probe
Chamber Vent Plugs
Louvered Horizontal &
Vertical Baffles
Return Baffle
Flow Diffuser
Substrate Sampling Area
Border Insert
Rubber Isolation Skirt
Side View
F igure 111-2 .
In situ recirculating SOD chamber used on the
Pigeon River, North Carolina (modified from
Hickey 1988)
111-9
-------
Nutrients
Total Kjeldahl nitrogen, nitrate as nitrogen, nitrite as
nitrogen, and ammonia nitrogen, and total phosphorus and
ortho-phosphate as phosphorus were collected and analyzed using
standard EPA protocols (Table III—3) (EPA 1983). Quality control
procedures in the laboratory followed standard EPA protocol,
using complete chain-of-custody procedures, and including
control, blank, spike, and duplicate samples in each batch
processed.
Chloride
Chloride samples were collected at each station to serve as
a conservative tracer for other constituents. Laboratory
analysis followed EPA protocol 325.3.
Deaeration
The rate of deaeration of supersaturated river water below
Champion's second sidestream oxygenation unit (RM 61.5) was
measured by holding shallow pans of supersaturated river water in
a floating apparatus. Water for these pans was collected from
immediately below the oxygenation site. The rate of dissolved
oxygen decline in each pan as measured with a YSI DO meter at
approximately 15 minute intervals for 2 h. Three bottles were
concurrently filled and placed within the pans to account for the
changes in dissolved oxygen concentration associated with
metabolic processes. The DO decline in these bottles was
subtracted from the total decline in DO to calculate loss rates
due to deaeration.
111-10
-------
Table III-3. Analytical methods for Pigeon River water
chemistry sampling
Parameter
Preservat ion
Analytical
Method
Total Kjeldahl Nitrogen
Ammonia-N
NO2-N
NO3-N
Total P/Ortho-P
CBOD
NBOD
Chlor ide
H2SO/4°C Colormetric
(EPA 351.3)
H2SO/4°C Colormetric
(EPA 350.2)
H2SO/4°C Spectrophotometric
(EPA 353.2)
H2SO/4°C Spectrophotometric
(EPA 353.2)
H2SO/4°C Colorometric
(EPA 365.2)
Chill 4°C 5 day (EPA 405.1)
150 day (NCDEM long
term method)
Chill 4°C 150 day (NCDEM long
term method)
Chill 4°C Titrimetric
(EPA 325.3)
III-ll
-------
IV.
EMPIRICAL STUDY RESULTS
Flow
Flow in the Pigeon River study area was relatively constant
for the several days up to and including the study period (Fig.
IV-l). Flows ranged from approximately 72 cfs at Canton to 131
cfs at Hepco (Fig. IV-2). At Canton, nearly all of the water in
the river was withdrawn by Champion Mill; only about 3 cfs of the
total 70 cfs available spilled over the dam and was not used by
the mill. Within the mill, evaporation and other losses accounted
for an estimated reduction of about 3 cfs between intake flow and
effluent flow. Downstream, two major tributaries, Richland Creek
(Tl), and Jonathans Creek (T3) constituted about 80% of the
increase in flow between the Champion Mill outfall and Hepco.
Effluent flows contributed by Clyde and Waynesvill'e WWTPs (E2 and
E3) were small relative to the Champion effluent; the combined
total flow of Waynesville and Clyde WWTPs was 15 times less than
flow from Champion (El).
Temperature
During the study, mainstein temperatures were consistently
lowest upstream of the mill and highest just downstream of the
mill effluent (Figs. IV-3, IV-4, and IV-5). A similar longi-
tudinal pattern was observed for all three sampling periods;
temperatures rose from about in °C at the mill intake to about
33 °C below the mill outfall, and slowly declined to about 20 °C
by Hepco. However, some warming was apparent from Si to S9
(Crabtree Creek to Jonathans Creek), presumably due to an
increase in solar insolation attributed to the shallow, wide
stretches which are common in that reach.
Diurnal temperature variation measured at the four
intensive stations (SI, S4, S6, and S10) was generally small and
related to the proximity of the station to Champion Mill effluent
(Fig. IV-6). Upstream of the papermill (Si), essentially no
diurnal variation was observed. At the two intensive stations
closest to the Champion effluent (S4 and S6), diurnal
fluctuations of 3-4 °C were recorded, probably due to the greater
temperature differential between air and water at these stations
at night.
IV-1
-------
160
150
140
130
120
110
100
90
80
23
F igui
~
f
Hepco USGS Gauge
1
t
4
Canton USGS Gauge
i n i i i 1 r
EP88 24SEP88 25SEP88 26SEP88 27SEP88 28SEP88 29SEP88
DATE
iv-i.
Flow data
during 23-
from continuous USGS
29 September 1988
gauges at canton and Hepco, North Carolina,
-------
130-
120-
Fines Creek
Jonathans Creek
I
110-
100-
l/i
it—
O,
^ 90-
O
80
70-
60-
Crabtree Creek
Waynesville WWTP
Richland Creek —
Champion Mill
Clyde WWTP
' 3cfs
—r
62
~r
60
1-
54
T"
52
T~
46
64
5a
56
I
50
n
48
I
44
42
River Mile
Figure IV-2.
Flows in the Pigeon River, North Carolina between
Canton and Hepco on 27 September 1988
IV-3
-------
40
38
36
34
32
30
28
26
24
22
20
18
16
14
12
10
igu
IV-3. Water temperatures (°C) at mainstem and input sampling stations within
the Pigeon River, North Carolina study area between 0700 and 1100 on
27 September 1988
-------
40
38
36
34
32
30
28
26
24
22
20
18
16
14
12
10
igu
E01
*
S03
S04
S05
S09
S06
S08
E02
E03
SO 7
TO I
T03
T02
SOI * *
-| 1 1 1 1 I 1 1 1 1 1 l 1 i | " | i | I r
62 60 58 56 54 52 50 48 46 44 42
RIVER MILE
IV-4 .
Water temperatures (°C) at mainstem and input sampling stations within
Pigeon River, North Carolina study area between 1300 and 1700 on
27 September 1988
-------
40
38
36
34
32
30
28
26
24
22
20
18
16
14
12
10
'iqi
S02
S03
S04
S06
S05
S09
S08
S07
SOI
i 1 i 1 i 1 i 1 i 1 i ¦ i 1 i 1 i ¦ 1 1 1 ' r
4 62 60 58 56 54 52 50 48 46 44 4.2
RIVER MILE
e IV-5. Water temperatures (°C) at mainstem and input sampling stations within
Pigeon River, North Carolina study area between 1000 and 1530 on
28 September 1988
-------
40
38
36
34
32
30
28
26
24
22
20
18
16
14
12
10
STATION * * * SOI
Q--0---G S04
o~o--o S06
dr-A-a- sio
Q
Q-" ' ~
0-
.—<^— - -
—*
-¦ar
/V
-i—i—j—i—i—i—|—i—i—i—|—i—i—i—[—i i ' | 1 1 1 i 1 ' ' i 1 1 1 i 1 1 r
8 10 12 14 16 18 20 22 24
TIME
IV-6 .
Diurnal temperature variation measured at four intensive stations in
the Pigeon River, North Carolina on 27 September 1988
-------
Dissolved Oxygen (DO)
Longitudinal dissolved oxygen (DO) patterns in the mainstem
were similar among sample periods; a decline in DO to almost 2
ppm below saturation was evident near Richland Creek (S5), and
recovery to values near saturation was apparent by Hepco (Sll)
(Figs. IV-7, IV-8, and IV-9). However, even with the decline
below saturation, no DO values below 6 ppm were observed during
the study.
On 28 September, additional DO monitoring conducted near
Champion's two sidestream oxygenation units revealed DO peaks
greater than 20 ppm immediately downstream of each unit, with
rapid declines observed slightly downstream (Fig. IV-9). These
declines were attributed to dissolution of supersaturated oxygen
from the water column, as measured in the deaeration of super-
saturated river water held in shallow pans (Fig. IV-10).
Diurnal DO variation measured at the intensive sampling
stations was relatively small (less than 2 ppm), and inconsis-
tent among stations (Fig. IV-11). Diurnal fluctuation was
greatest upstream of the mill at Si, with the highest DO re-
corded in late afternoon and lowest values observed in the early
morning. Dirunal variation in DO at the three stations below the
Champion outfall did not exceed 1 ppm.
BOD
Both five-day (BOD^) and UCBOD results followed similar
longitudinal patterns in the study area; BOD^ was highest near
Champion's outfall and declined with distance downstream (Figs.
IV-12, IV-13, and IV-14). In addition, the diluting effect of
the combined addition of Richland Creek at RM 54.9 and
Waynesville WWTP at RM 54.8 was evident in all three sets of
samples.
As with DO and water temperature, diurnal variation in BOD^
was smallest at SI, above the Champion Mill outfall (Fig. IV-157
At S4, S6, and S10, BOD^ was lowest at 0700 and highest between
1300 and 1700. Conversely, BOD5 at Si was near zero at all
sample times, but the highest value of 0.5 mg/1 occurred in the
0700 samples.
IV-8
-------
<
I
o
E
Q.
CL
O
>-
X
o
o
LlJ
>
i
o
oo
00
Q
22
21
20
19
18
17
16
15
14
13
12
11
10
9
8
7
6
5
EOI
SOI
S04
64
62
60
58
56
54
52
50
48
46
44
42
RIVER MILE
Figure IV-7. Dissolved oxygen (ppm) levels at mainstem and input sampling stations
within the Pigeon River, North Carolina study area between 0700 and
1100 on 27 September 1988
-------
22
21
20
19
18
17
16
15
14
13
12
11
10
9
8
7
6
5
S04
EOI
T03
r.«T04
S"V*
T c
TOt
503
SOB S09
SOI
S1I
T02
E02
EOJ
S06
SO 7
S02
SOS
—. 1 < 1 . 1 . 1 i | ' i i i ' i i i ¦ i i r
54 62 60 58 56 54 52 50 48 46 44 42
RIVER MILE
gure IV-8. Dissolved oxygen (ppm) levels at mainstem and input sampling stations
within the Pigeon River, North Carolina study area between 1300 and
1700 on 27 September 1988
-------
22
21
20
19
18
17
16
15
14
13
12
11
10
9
8
7
6
5
SOI
S02
503
S11
T
T
T
54 62 60 58 56 54 52 50 48 46 44 42
RIVER MILE
IV-9 .
Dissolved oxygen (ppm) levels at mainstem and input sampling stations
within the Pigeon River, North Carolina study area between 1000 and
1530 on 28 September 1988
-------
Deaeration Measurement
<
i
20
19
e 18
Cu
a
z
w
o
tH
X
o
a
CJ
>
o
en
u>
M
Q
17
16
15
14
13
12
Initial
t
f
Final Bottle
I
i
i
J
Final Pan
1300
Figure IV-10.
1400
1500
1600
TIME
Decline
Carolina
shallow pans
of DO (ppm) over time in supersaturated Pigeon River, North
water collected below the oxygenation unit at RM 61.3 and held
----- for two hours
-------
11 -
10
8
5 "I
.. -Q-
Q "
-EJ
A--
O- ¦
¦•Sk
.-
-------
.0
. 5
. 0
. 5
. 0
. 5
.0
. 5
. 0
. 5
. 0
EOI
* S02
503
E03
*
504
E02
S05
SOI
54
—r~
62
SO 6
*
TO I
—I—
S07
S09
508,
T02
*
T03
*
S11
Ao4
60 58 56 54 52 50 48 46 44
RIVER MILE
sio
r
42
-12.
Five-day biochemical oxygen demand (BOD^) at sampling stations in the
Pigeon River, North Carolina between 0700 and 1100 on 27 September 1988
-------
. 0
. 5
. 0
. 5
. 0
. 5
. 0
. 5
. 0
. 5
. 0
S02
E03
S04
EOI
S05
S03
E02
S06
S08
S09
S07
SIO
T04
SI I
T02
T03
TO I
42
44
48
46
50
52
54
56
58
60
62
RIVER MILE
-13.
Five-day biochemical oxygen demand (BOD,.) at
North Carolina between 1300 and 1700 on Z7
September 1988
sampling stations in the
-------
56
52
48
44
40
36
32
28
24
20
16
12
8
4
0
ure
E01
*
S04
S02
S03
S05
E03
S06
S07
S09
E02
T02
TO I
T03
*
sot
62 60 58 56 54 52 50 48 46 44 42
RIVER MILE
-14 .
Ultimate carbonaceous biochemical oxygen demand (UCBOD) at sampling
stations in the Pigeon River, North Carolina between 0700 and 1100
on 27 September 1988
-------
<
I
I—'
-J
01
E
in
Q
O
CQ
5.0
4 . 5
4.0
3.5
3.0
2.5
2.0
1.5
1.0
0.5
0.0
STATION —*—+- soi
Q Q o S04
o S06
S10
Q-
_
D
O--
-4r'
-| 1 1 1 1 1 1 1 1 1 r
6 8 10
-I 1 1 1 1 1 1 1 1 1 1 1 1 r
Q.
~
-K>
"i 1 1 1 1 1 r
12
14 16
TIME
18
20
22
24
Figure IV-15. Diurnal variation in BOD5 (mg/1) at four stations in the Pigeon
River, North Carolina study area on 27 September 1988
-------
Sediment Oxygen Demand (SOD)
Sediment oxygen-demand (SOD) in the study area ranged from
zero to 1726 mg 02/m -day (Table IV-1). No longitudinal patterns
were evident, and there was considerable variation within
samples at a given station. In addition, no consistent differ-
nces between light and dark chamber SOD experiments were
apparent. The overall njean for the study area below the Champion
outfall was 339 mg 0,/m /day.
Photosynthesis
As measured by light/dark bottle experiments at each of the
four intensive stations, water column photosynthesis rates in the
Pigeon River were low during the study,* only at station S10 was a
net increase in DO observed (66 mg 02/m -h). At S2, S4, S6, and
the USGS gauge location at Canton, no change or a slight decrease
in DO was observed.
Water column chlorophyll a concentrations in the Pigeon
River study were consistent with in situ light/dark bottle
results; no chlorophyll a valut; above 1 mg/1 was observed
upstream of S6, while mainstem values below Crabtree Creek (S8)
were between 3-6 mg/1 (Fig. IV-16).
Light attenuation characteristics of the Pigeon River
within the study were also consistent with chlorophyll a and
primary productivity measurements (Fig. IV-17). From the
Champion Mill outfall (El) until Richland Creek (S6), essen-
tially no light penetrated water below 0.2 m. Between Richland
Creek (S6) and Jonathans Creek (S9), light penetration increased
to more than 0.34 m. Below Jonathans Creek, light penetration
increased again to more than 0.6 m.
Nutrients
Nitrogen species concentrations in the study area were
generally less than 2 mg/1, except in the discharges from the
Clyde and Waynesville WWTPs {E2 and E3), where elevated TKN and
nitrate concentrations were observed (Fig. IV-18). Ammonia and
TKN values increased sharply below the Champion Mill outfall (S2)
and slowly declined with distance downstream. In contrast,
nitrate and nitrite values were low at S2 and peaked near
Richland Creek (S6). Diurnal fluctuations in nitrogen were small
at all four intensive stations (Figs. IV-19 to IV-22).
IV-18
-------
Table IV-1. Sediment oxygen demand at selected sites in the
Pigeon River, North Carolina during 27-28
September 1988
Station River Mile Type SOD mg 02/m^/day
SO 64.8 Light (net increase) 372.2
Dark 0
S 2 62 .8 Light 358.2
Dark 558.9
Dark 0
S4 59.2 Light 0
Dark 34.5
Dark 0
S6 53.8 Light 0
Dark 1393.5
Dark 0
S10 43.0 Light 1726.3
Dark 0
Dark 0
Mean all SOD samples (excluding SO) 339.3
IV-19
-------
0
5
0
5
0
5
0
5
0
5
0
5
0
S09
SI I
S10
S08
SOT.
T02
T03
S05
S04
SOI
E02
E03
TO I
S06
S02
EOI
S03
48
42
46
44
50
52
56
54
60
58
62
RIVER MILE
Figure IV-16. Chlorophyll a concentration (mg/1) in the Pigeon River, North
Carolina study area on 27 September 1988
-------
<
I
NJ
E
u
a:
H
a.
u
a
0
20
40
60
80
100
120
.
S04
„ . S05
S02 —
S03
n. S07 S09
S06 ^
SOS ^
S10
«
u
w
w
S11
J
oi
2
u
u
U)
1 Q
Z
:
z
<
:
<
E
••
J
:
tc
<
:
u
S3
1;
H
O
S01
OS
~l I r I I I I I P
n—i—i—i—i—i—i—i—r
"i—i—i—i—r
65
60
55
50
1
45
t—i—i—i—i—i—i—i—r
40
RIVER MILE
Figure IV-17. Water depth (cm at which 99% of available surface light was attenuated
at 11 stations in the Pigeon River, North Carolina on 28 September
1-988 (NOTE: 99% depth at SI may be erroneously shallow due to an
inability to sample depths greater than lm)
-------
5.5
5.0
4.5
4.0
h 3. 5
\
Ol
3 3.0
1 'H 2.5
(0
!J £2.0
(D
= 15
o '
o
1.0
0.5
0.0
(14.2 mg/1)
I
PARAMETER
E02
E03
T03 O
*—h*- NH3-N
o-o-q N02-N
N03-N
TKN
SIO S11
..-©-e
A
-b- -a -6-
T04
64
62
60
58
56
54
52
50
48
46
44
42
RIVER MILE
Figure IV-18. Longitudinal concentrations (mg/1) of NH^-N, total Kjeldhal nitrogen,
N02-Nf and NO^-N in the Pigeon River, North Carolina study are on
27 September 1988
-------
0.6"
0.5
0.4
CO
0.3
- 0.2
0.1
0.0
-Q"
~
•Br"
STATION
-------
0
8
6
4
2
0
8
6
4
2
0
O--
ST AT I ON * * )|c SOI
G-e-O 5Q4
so6
i£sr-A-^ SIO
-¦a--"
...-Q
^ ¦ —O
\
"8""
~~I 1 1 1 1 1 1 1 I I I I I I I I I I I I I I I I I I I I I I I I p
8 10 12 14 16 18 20 22 24
TIME
IV-20. Diurnal variation in total Kjeldahl nitrogen (mg/1) at four stations
in the Pigeon River, North Carolina on 27 September 198R
-------
0
8
6
4
2
0
8
6
4
2
0
STATION "*—*—*¦ SOI
Q-B--Q SQ4
SO6
° A--A-A SIO
A--—
"~--e
— • —o
A
"O Q
-i—i—i—|—i—i—i—|—i—i—i—|—i—i—i—|—i—i—i—|—i—i—i—|—i i i | i i i | i i i r
8 10 12 14 16 18 20 22 24
TIME
IV-21. Diurnal variation in nitrate as nitrogen (mg/1) at four stations in
the Pigeon River, North Carolina on 27 September 1988
-------
10
09
08
07
06
05
04
03
02
01
00
STATION » * * sol
SQ4
SO6
siO
-Q-
~
--O
-A—*-
-A
|iii|iii|iii|iii|iii | i i i | i i i | i i i j~
8 10 12 14 16 18 20 22 24
TIME
IV-22. Diurnal variation in nitrite as nitrogen (mg/1) at four stations in
the Pigeon River, North Carolina on 27 September 1988
-------
Phosphorus concentrations in the Pigeon River mainstem were
always less than 1 mg/1, but total phosphorus values of 3.8 mg/1
and 1.5 mg/I were recorded from Clyde WWTP (E2) and Waynesville
WWTP (E3), respectively (Figure IV-23). Mainstem values
increased sharply below the Champion Mill outfall (El), declined
slowly until station S6, and remained nearly constant thereafter
No tributary values for phosphorus exceeded 0.2 mg/1. As with
nitrogen species, there were no apparent diurnal fluctuations in
total phosphorus and ortho-phosphate (Figs. IV-24 and IV-25).
Chloride
Only a single significant source of chlorides was identi-
fied during the study — the Champion Mill effluent (Fig. IV-26).
Correspondingly, mainstem values increased sharply below the
outfall and declined in response to the aperiodic addition of
flows to the river.
[V-27
-------
0
5
0
5
0
5
0
5
0
F
PARAMETER —N*—P04
Q--G-Q TP
E03
~
*
S03
..a--
S04
S02
505
S06
"13-
S07
S08
S09
--Q-
T02 9
SO
T04
TOI
56
54
50
60
58
52
48
62
46
44
42
RIVER MILE
ure IV-23. Phosphorus concentrations (mg/1) at mainstem and input stations in the
Pigeon River, North Carolina study area on 27 September 198R
-------
1.0
0.8
STATION
-0-
-o.
* * * SOI
Q-G-Q Sq4
S06
A—A-^r S10
en
e
OJ
p
u
0
x:
a
01
0
.c
01
0.6
<3 rH
I td
to -P
VD O
Eh
0.4
-a
"0-""
G—'
- —"O
. e---
o
-e--
¦A-
A- _ _ .
0.2
0.0
~i 1 1 1 1 1 r
8
—I—i—i—i—|—i—i—i—|—i—i—i—]—i—i—i—|—i—i—i—|—i—i—i—|—i i i p
10 12 14 16 18 20 22 24
TIME
Figure IV-24. Diurnal variation in total phosphorus concentrations (mg/1) at four
stations in the Pigeon River, North Carolina study area on 27 September
1988
-------
0
9
8
7
6
5
4
3
2
1
0
STATION * * * SOI
Q-G-Q SQ4
SO6
s10
~l 1 1 1 1 1 I | I I I | I I I | I I I | I I I | I I I | I I I | I I 1 1~
8 10 12 14 16 18 20 22 24
TIME
IV-25. Diurnal variation in ortho-phosphate as phosphorus concentrations (mg/1
at four stations in the Pigeon River, North Carolina study area on
27 September
-------
Cn
E
T3
•H
U
O
700
600
500
400
5 300
<
I
OJ
200
100
0
E01
*
S02
50 I
S04
E02
*
S05
S06
E03
*
TO I
*
S07
T02
*
T03
*
SI 1
T04
*
T
64 62
60
58 56
54
52 50 48 46 44 42
RIVER MILE
Figure IV-26. Chlor.ide concentration (mg/1) at mainstem and input stations in the
Pigeon River, Worth Carolina study area on 27 September 1988
-------
V. MODELING METHODS
The Pigeon River model was constructed using the QUAL2E-
UNCAS model (Brown and Barnwell 1987) and data from the September
27, 1988 field survey described in Chapter IV. The Pigeon River
system was modeled from River Mile 63.6 (i.e., just upstream from
the Champion discharge) to River Mile 42.8 (i.e., just upstream
from Walters Lake). Flow conditions during this period were low
(i.e., < 100 cfs at Champion) and were representative of low flow
periods during late summer.
A. MODEL STRUCTURE
A conceptual overview of the Pigeon River Model (PRM) is
represented schematically in Figure V-l. The wide range of
processes affecting dissolved oxygen dynamics within the 20 mile
modeled segment is evident from the figure. The processes, in
conjunction with point discharges, control the dissolved oxygen
concentrations in the river. Also shown in this figure are the
four major subsystems within the modeled Pigeon River segment:
o The upstream input subsystem is forced based on
measured concentrations of state variables and
represents a subsystem high in dissolved oxygen
concentration, and low in nutrients and BOD. This
subsystem is the initial condition of the Pigeon River
prior to the modeled point source additions.
o The riverine point sources are forced based on measured
concentrations of the state variables. These point
sources include the Champion Paper Mill effluent, three
oxygenators, the Clyde and Waynesville WWTPS, and four
tributaries to the Pigeon carrying their own point and
non-point source loads. These riverine point sources
include BOD, various nitrogenous compounds, phosphates,
chlorides, oxygen, heat, and water.
o The Pigeon River subsystem is characterized as 105 0.2-
mile segments which are dominated by advection. All
dynamic processes (e.g., degradation, dilution, and
transformation) occur within these modeled segments.
o The downstream subsystem whose inputs are controlled by
the combination of advection, transformtion, and
degradation within the modeled Pigeon River segment.
The downstream subsystem is defined as Walters Lake and
V-l
-------
Point
Sources
Nonpoint
Sources
Reaeration
Gross
Productivity
r\
Upstream
Loads
<
i
to
PIGEON RIVER DYNAMICS
Loads to
Walters Lake
CBOD
NBOD Respiration Deaeration
Nutrient
Cyding
Figure V-1. Schematic representation of Pigeon River Model
-------
is not modeled in this effort. The inputs to Walters Lake are
defined by the advective materials flows from the final modeled
Pigeon River segment at River Mile 42.6.
The physical exchange among the 105 segments used to
characterize the Pigeon River is primarily the result of
advective flow and instantaneous mixing, and is controlled by
upstream and tributary discharges. Exchange of oxygen at the
air-water interface in specifically modeled based on in-stream
dissolved oxygen measurements and experimentation (see Chapter
III). Within-stream settling of organic particles was considered
to be minimal based on the rapid travel times within the modeled
segments (i.e.,8-20 minutes) and observations during the field
sampling.
Based on the conceptual structure, the Pigeon River QUAL2E-
UNCAS model was defined with 13 biological/chemical state
variables and 6 physical state variables representing the carbon,
oxygen, nutrient, and water storages in the subsystem components
(Table V—1). A total of nine point sources were used in the
model, each with state variable concentrations of their
effluents (Table V-2). Figure V-2 is a detailed materials flow
diagram illustrating the interactions among these model
components and their relationships to the major driving forces
within the Pigeon River (i.e., upstream flow, tributary flow,
effluents, and temperature). The diagram is essentially a very
detailed view of the conceptual model described in Figure V-l.
To depict the 20.8-mile riverine system, 10 reaches of
varying lengths were incorporated into the PRM (Table V-3, Fig.
V-3). Ten reaches were considered sufficient to characterize
individual sections of the river where point sources and
tributaries could affect local concentrations of state
variables, particularly dissolved oxygen and ultimate biochemical
oxygen demand.
The QUAL2E-UNCAS model used to describe the state variable
concentrations in the Pigeon River was applied as a steady-state
formulation representing late-summer low flow conditions (i.e.,
<100 cfs). The dominant biological and chemical mechanisms
represented in the model structure include microbial/chemical
degradation and transformation, respiration, primary
productivity, and chemical oxygen demand. The primary physical
mechanisms included are advection, reaeration, deaeration, and
mixing.
V-3
-------
Table V-l. State variables used in the Pigeon River QUAL2E model
Biological/Chemical State Variables
Dissolved oxygen
Carbonaceous BOD
Chlorophyll a
Total nitrogen
Organic nitrogen
Ammonia
Nitrite
Nitrate
Organic phosphorus
Dissolved phosphorus
Total phosphorus
Chlorides
Temperature
Physical State Variables
Flow
Travel time
Veloci ty
Depth
Width
Volume
V-4
-------
Table V-2. Point sources used in the Pigeon River QUAL2E model for
September 27, 1988
Flow Temp DO BOD Cl Chla NH3
Point Sources
Champion Paper Mill 6V-4 95.9 9.7 53.5 619.0 0.5 1.1
Oxygenator #2 25.0 * 30.0 * * * *
Oxygenator #3 25.0 * 30.0 * * * *
Clyde WWTP 0.2 70.9 7.5 24.0 28.9 0.5 1.8
Waynesville WWTP 4.1 65.4 7.9 32.4 36.4 0.4 0.3
Tributaries
Richland Creek 24.7 65.0 8.8 1.6 3.6 0.4 0.1
Crabtree Creek 2.7 60.7 8.1 2.4 4.5 . 15.0 0.1
Jonathan's Creek 30.2 60.2 9.2 0.5 2.8 9.0 0.1
Fines Creek 2.9 59.9 8.7 5.4 4.4 2.4 0.4
*Defined by influent source
V-5
-------
7777
777 77
SOD
CBOD
ORG-P
ORG-N
DIS-P
Atmospheric
Reaeration
ALGAE
Cilia
Figure V-2. Detailed flow diagram of Pigeon River QUAL2E model
(from Brown and Barnwell 1987)
V-6
-------
Table V-3. Lengths and location of modeled river reaches in
Pigeon River QUAL2E model
Begin End Length Number of
Reach # River Mile River Mile (Mi.) Elements
1
63.6
62.8
0.8
4
2
62.8
61.4
1.4
7
3
61.4
59.2
2.2
11
4
59.2
55.2
4.0
20
5
55.2
53.8
1.4
7
6
53.8
49.8
4.0
20
7
49.8
48.8
1.0
5
8
48.8
47.6
1.2
6
9
47.6
43.6
4.0
20
10
43.6
42.6
1.0
5
V-7
-------
to Walters Lake
Jonathans Creek
Fines Creek
©
Scale of Miles
©
Crabtree Creek
©
Sidestream
Oxygenators
©
Richland Creek
E1 (Champion Mill)
S 1
A
E 2
(Clyde WWTP)
E3 (Waynesville WWTP)
Lake Junaluska
(§) Intensive Station
Figure V-3. Location of modeled reaches in Pigeon River QUAL2E
Model (reach numbers are circled)
V-8
-------
PRM Model Equations and Component Evaluation
The specific mathematical relationships used to represent
algal dynamics; BOD decay; nutrient transformations, uptake, and
excretion; hydrodynamics; and, transport in QUAL2E-UNCAS are
described in detail elsewhere (Brown and Barnwell 1987). The
major processes used in PRM are described below by type of
process.
Hydraulic Discharge
The QUAL2E model assumes that the stream hydraulic regime is
steady state such that the discharge at any location is the sum
of the external inflows and/or withdrawals to that element.
Given this relationship, discharge coefficients are determined
from available data concerning velocity, cross-sectional area,
and depth. The discharge coefficients used in PRM are shown in
Table V-4 and were determined from empirical data collected at
the site.
Carbonaceous BOD
The QUAL2E model assumes a first order
characterize the degradation of ultimate ca
change in UCBOD concentration is determined
degradation and settling, so that:
dL/dt = -K1L - K3L
whe re
L = Concentration of ultimate carbonaceous BOD (mg/1)
= Deoxygenation rate coefficient, temperature
dependent, (day )
= Rate of loss of carbonaceous^BOD due to settling,
temperature dependent, (day )
Based on laboratory analyses of water column oxygen demand, the
deoxygenation coefficient was calculated to be 0.11/day. The
settling rate was determined by mass balance for each stream
reach and averaged 0.001/day. The factor to convert 5-day CBOD
to ultimate CBOD was derived analytically to be 13.2 based on the
mean of UCB0D/CB0D5 ratios for stations S2 through Sll and the
Champion effluent.
linear reaction to
rbonaceous BOD. The
as the sum of the
V-9
-------
Table V-4. Discharge coefficients used in Pigeon River QUAL2E
model
Linear Power Linear Power
Manning Coefficient of Coefficient Coefficient Coefficient Coefficient
Reach # Coefficient Dispersion of Velocity of Velocity of Depth of Depth
1
0.4
60.0
0.015
0.802
1.98
0.0
2
0.4
60.0
0.016
0.802
1.92
0.0
3
0.4
60.0
0.017
0.802
1.35
0.0
4
0.4
60.0
0.024
0.802
1.58
0.0
5
0.4
60.0
0.016
0.802
1.64
0.0
6
0.4
60.0
0.013
0.802
1.50
0.0
7
0.4
60.0
0.010
0.802
1.63
0.0
8
0.4
60.0
0.011
0.802
1.61
0.0
9
0.4
60.0
u.013
0.802
1.27
0.0
10
0.2
60.0
0.020
0.802
1.24
0.0
-------
Dissolved Oxygen
The balance of dissolved oxygen in a stream reach is
determined by the sum of its sources, sinks, and reaeration/
deaeration. The major sources of oxygen within the modeled
portion of the Pigeon River are reaeration in the segments after
River Mile 50, artificial infusion of oxygen into the streamflow
by three oxygenators operated by the Champion Paper Mill, and
tributary inflows. The Champion oxygenators infuse highly
oxygenated water (i.e., 30-50 ppm) at the Champion discharge,
River Mile 62.4, and River Mile 61.0. The major sinks for
oxygen include the degradation of CBOD and NBOD and biotic
respi ration.
Reaeration
The reaeration method of Tsivoglou and Wallace (1972) was .
used to estimate reaeration rates for the Pigeon River. This
method assumes that the reaeration coefficient for a reach is
proportional to the change in elevation of the reach and is
inversely proportional to the travel time within the reach. The
reaeration coefficients used in PRM fall within the bounds
suggested by Brown and Barnwell (1987) for streams with less than
3000 cfs. In addition, the reaeration rates become "deaeration"
rates following the infusion of oxygen at the Champion aerators.
These deaeration rates were estimated using the results of the
simple experiment described in Chapter IV. These results would
underestimate the diffusion of oxygen to the atmosphere because
the oxygenated water was held motionless in the experiment and
highly oxygenated flowing water would tend to deoxygenate faster
than water in the experimental situation.
Nitrogen Cycle
The nitrogen cycle is modeled in QUAL2E as a stepwise
transformation from organic nitrogen to ammonia, to nitrite, and
finally to nitrate. The PRM models all four of these nitrogen
constituents using estimates of hydrolysis from organic nitrogen
to ammonia, settling rates, algal excretion and uptake rates,
benthic regeneration rates, and oxidation rates for nitrogenous
compounds.
V-ll
-------
Phosphorus Cycle
The phosphorus cycle is modeled in QUAL2E as a trans-
formation cycle from organic phosphorus to dissolved phosphorus.
The concentration of phosphorus is modeled based on algal uptake
and excretion rates, decay rates, settling rates, and benthic
regeneration rates.
Mass Transport
The basic equation used by QUAL2E for the transport of
materials between model elements is the one-dimensional
advection-dispersion mass transport equation. This equation is
numerically integrated over space and time for each water quality
constituent by including the effects of advection, dispersion,
dilution, constituent reactions and interactions, and sources and
sinks. For any constituent, C, this equation can be written as:
dM/dt = [(d(A D_(dC/dx))/dx]dx - [d(A uC)/dx] + (A dx)(dC/dt) + s
X L> X X
where
M = mass (M)
x = distance (L)
t = time (T)
C = concentration (ML-^)
2
A^ = cross-sectional area (L )
= dispersion coefficient ( L^T-"^ )
u = mean velocity (LT
s = external sources or sinks (MT ^).
Using this approach, a set of differential equations was
developed to describe the oxygen/BOD/nutrient dynamics of the
Pigeon River by combining all mathematical terms representing
inflows and outflows for the biological and chemical processes
described above.
The initial conditions for the PRM were determined from the
results of the field survey conducted in September 1988 (Chapter
IV). Additional coefficients and rate constants were derived
from the published literature and an unpublished time of travel
study conducted by the North Carolina Department of Environmental
V-12
-------
Monitoring (NCDEM). The values of initial conditions used in the
Pigeon River Model are listed in Appendix A.
Model Calibration
The field survey data and the results of laboratory analyses
were used to calibrate the initial QUAL2E model structure. The
primary state variables of interest were dissolved oxygen, UCBOD
and chlorides (i.e., as a conservative check for mass conserva-
tion). In addition, the model included a series of nitrogenous
compounds (i.e., total Kjeldahl nitrogen, nitrites, nitrtes, and
ammonia), total and dissolved phosphorus, and chlorphyll a. The
model was calibrated using visual correspondence between model
and field observations to approximate the best "fit" (Summers et
al. 1980). Further calibration was completed objectively by
systematically modifying each parameter over a small, defined
range to create the matrix of parameter values which represented
the smallest total sum squared error between the field and
modeled data. The result of this calibration process is listed
in Appendix A.
B. MODEL VALIDATION
Validation 1-28 September 1988
Initial validation of the Pigeon River QUAL2E model was
completed using the Champion self-monitoring data taken on
September 28, 1988, the day following the intensive field survey.
This date was selected such that the primary driving forces of
the model (e.g., flow, temperature, effluents) would be
approximately the same as those used in the nominal model
condition. The major differences between the nominal data set
and this validation data set are the locations of collection and
the personnel conducting the monitoring. Only UCBOD and DO were
validated using this data set. The validity of the model runs
for both state variables were compared by visual correspondence.
In addition, validity was examined statistically by performing a
linear regression of modeled versus observed data. Ideal
correspondence would result in a regression with a slope of 1.0
and an intercept of 0.0. Our criteria for an acceptable
validation was that the 95% confidence limits of the regression
slope and intercept not be significantly different from 1.0 and
0.0, respectively.
V-13
-------
Validation II - 7 July 1987
In addition to validating the model using late summer 1988
monitoring data, the validity of the model to describe higher
flow (200-400 cfs) conditions was examined. Using data from an
intensive field monitoring program completed for Champion in July
1987 (EA 1988), the model structure was tested to determine if
the simulation could replay the observations made at high flow
conditions. Criteria for acceptance were the same as the initial
validation. The major difference between the two validations was
that UCBOD was not used in this simulation. Rather than UCBOD,
BOD^ was simulated because the ratio between UCBOD and BODr would
not necessarily be the same as measured, in September of 1987.
Without knowing that the Champion effluent contained the same
types of carbonaceous materials in July 1987 as September 1988,
the ratio developed in 1988 would be invalid. This validation
procedure included chlorides, dissolved oxygen, BOD^, nitrogenous
compounds, and phosphorus compounds.
C. MODEL UNCERTAINTY
QUAL2E is equipped with an uncertainty assessment processor
(UNCAS) to permit evaluation of parameter sensitivity, first-
order error propagation, and overall uncertainty using Monte
Carlo analyses. A set of Monte Carlo simulations was completed
to assess the potential level of error in the Pigeon River QUAL2E
predictions. The uncertainty package (UNCAS) is based on the
assumption that the inputs to the model or its parameters are
normally distributed with known coefficients of variation. All
inputs and parameters are assumed to be independent. Values for
parameters are randomly selected for use in the Monte Carlo
simulations from these normal distributions based on the
coefficient of variation. Five hundred Monte Carlo simulations
were performed to address each uncertainty estimate. The
standard deviation of the 500 runs represents a statistical
estimate of the uncertainty associated with the model prediction
for that state variable.
Monte Carlo analyses were conducted for seven different
uncertainty scenarios to pinpoint the primary causes of
prediction uncertainty. The degree of variability in the Monte
Carlo runs was determined by the coefficient of variation of the
27 September field survey data (i.e., diurnal variability), when
possible. When no measure of diurnal variation was available,
best scientific judgement was used to estimate the coefficient of
variation. The uncertainity simulations included:
o Varying all point source loads by 10-15% without
modifying point source discharge rates
V-14
-------
o Varying all headwater water quality concentrations by
10-15% without modifying headwater flow rates
o Varying all headwater water quality concentrations by
10-15% and headwater flow rates by 6%
o Varying all reaction coefficients affecting dissolved
oxygen concentrations by 3-20%
o Varying the reaction coefficient for the decay of
carbonaceous materials by 15%
o Varying all reaction coefficients affecting nitrogen or
phosphorus concentrations by 3-20% and reaction
coefficients affecting nitrogenous oxidation by 10%
o Varying all reaction coefficients, model parameters, and
point source discharges by 3-20% without modifying
streamflow.
The QUAL2E-UNCAS model package limits the comparison of
prediction uncertainties with the nominal model predictions to
five locations within the modeled portion of the Pigeon River.
The following five comparison sites were selected:
o River Mile 63.2 - Immediately after the Champion
di scha rge
o River Mile 61.0 - Immediately after the third and final
Champion oxygenator
o River Mile 57.0 - Clyde STP discharge
o River Mile 54.8 - Waynesville STP discharge directly into
the Pigeon River
o River Mile 42.6 - Confluence of the Pigeon River with the
upstream end of Walters Lake.
These locations allow the maximum utility to be made of the
comparative data to evaluate the uncertainty associated with the
model predictions, to examine potential modeling scenarios
concerning the effluents of Champion and the two WWTPS, and to
ascertain additional information that might be needed to further
characterize the modeled segment of the Pigeon River.
V-15
-------
CHAPTER VI. MODEL RESULTS
A. CALIBRATION
The calibration of the QUAL2E Pigeon River Model "fit" the
observed field survey data remarkably well (Figures VI-1 through
VI-9). The correspondence of the simulated curves for dissolved
oxygen, ultimate biochemical oxygen demand, and chlorides are
very good and generally fall within the 95% confidence intervals
of the observed data. The dissolved oxygen simulation accurately
characterizes the rapid increases in oxygen level near the
oxygenators and the systematic deaeration following the
oxygenation. The simulated nutrient data (N and P) are also
good, with the exception of nitrite. The nitrite simulation
suggests that nitrite should be measurable at the 0.03 to 0.05
ppm level while field observations show the value to be less than
0.01 ppm.
Chlorides
The modeling of chlorides for the Pigeon River was done to
verify that the flow regimes used in the model for the Pigeon
River and its tributaries, the inflow rates of the point sources,
and the dilution processes within the river were reasonable. In
theory, chlorides added to the river should not be degraded in
any manner and any observed changes in concentration would be
simply due to dilution. Modeled chloride concentrations increase
from approximately 2 ppm upstream of the mill to 618 ppm at the
Champion discharge and remain constant until subsequently diluted
by the inflows from Richland Creek, Crabtree Creek, Jonathans
Creek, and Fines Creek (Fig. VI-1). Upstream and tributary
concentrations of modeled chlorides were minimal (i.e., 2.3-4.5
ppm) and the only appreciable sources of chloride were the
Champion discharge (i.e., 619 ppm) and the Clyde and Waynesville
WWTPs (i.e., 28.9 and 36.4 ppm, respectively). The addition of
chlorides from the Clyde and Waynesville WWTPS actually diluted
the instream concentrations of chlorides to a small degree.
The simulated chlorides confirms that the hydraulic system
used by the model conserves chlorides in the proper concen-
trations to reasonably match the observed field data. The early
over-estimates of the model from River Mile 62 to 56 suggest an
error in the field estimates of the chloride concentrations.
The field data show that the chloride input from the Champion
discharge was 619 ppm and represented 100% of the river flow, yet
regions immediatedly downstream of the discharge were measured to
VI-1
-------
CL
700
600
500
400
300
200
100
j—i—i—i—i—r
65
~~
>
**
t—i—|—i—i—i—r" i i—i i i i i i i I i i—i i i | i i i—i i i i i i | i i i i i—i—i i—r
60
55 50
RIV MILE
45
40
gu
re VI-1. Observed (*) and simulated ( ) values of chloride (ppm).
-------
BOO
100 "
Figure Vl-2. Observed (*) and simulated
oxygen demand (ppm).
* *
V
' ' I I I 1 I I | 1 I I I I I 1 I l 1 1 1 1 1 1 1 1 1 1 p
50 45 40
RIV_MILE
( ) values of ultimate carbonaceous biochemical
-------
-l—i—i—i—i—|—i—i—r~~i i i i i i J—r
65
60
55
gure VI-3. Observed (*) and simulated
*
1 1 —i I I 1 I I "I i i i i I I | ' 1 1 1 1 1 *1 1 1 [*"
50 45 40
MILE
values of dissolved oxygen (ppm).
-------
Rea
#
1
2
3
4
5
6
7
8
9
10
VI-1.
Coefficients used in Tsivoglou and
estimation of reaeration/deaeration
Wallace
(1972)
Slope of
Energy Gradient
(Se) (ft/ft)
0.0019
0.0019
0.0019
0.0019
0.0019
0.0019
0.0034
0.0034
0.0049
0.0049
Escape
Coefficient
(c) (ft-1)
0.050
0.050
0.020
0.025
0.045
0.045
0.045
0.045
0.045
0.045
Mean Tsivoglow
and Wallace
K2 (day )
3 .640
3.825
1.605
2.910
4 .632
3.657
5.446
5.856
12.440
18.232
VI -13
-------
TKN
2.5 :
2.0
1.5
1.0
0.5
0.0
1—F—I—I—I—I——I 1 1—I—I 1 1—I 1 1 1—I I—I 1 1 1 1 1 1 1 1 1 1 1—r
65
60
r
55 50
RIV MILE
"1 I I I I I | 1 1 I I 1 1 1 1 1 1~
45 40
Figure VI-4. Observed (*) and simulated ( ) values of total Kjeldahl nitrogen (ppm).
-------
I
NH3
1.0
0.8
0.6
0.4
0.2
0.0
65
t—|—i—i—I—I—i—i—I—l—I—|—i—i—i—i—i I I l—I—|—I—I—i—i—i—i—i—i—i—|—i—i—i—i—i—r~
60
1
55 50
RIV MILE
45
40
Figure Vl-5. Observed (*) and simulated ( ) values of ammonia as nitrogen (ppm).
-------
N02
0.20"
i
0.15
0. 10
0.05
0.00
* * *
* *
*1—I—I 1—I 1 1 1 1 1 1—I 1—I 1-I I ' I |~
t—|—i—i—i—i—i—I—I—I—I—|—I—I—I—I—I—i—i—i—i—1—I—I—I—I—I—I—I—i—r
65
60
55 50
RIV MILE
45
40
Figure VI 6. Observed (*) and simulated (—) values of nitrite as nitrogen (ppm).
-------
N03
2.0
0.0
65
"T—I—I—I—I—l—I—l—l—|—I—I—I—I—I—I—I—I—r
60
55
Figure VI-7. Observed (*) and simulated
50 45
RIV MILE
) values of nitrate
as nitrogen (ppm)
-------
0.0
J
~i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—1—r
65
60
55
Figure VI-8. Observed (*) and simulated (
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 j 1 1 1 1 1 1 1 1 1"
50
45
40
RIV MILE
) values of dissolved phosphorus (ppm).
-------
CHLA
10 :
8
f
^—i—i—i—i—i—i—i—i—i—|—i—i—i—i—i—i—i—i—i—|—i—I—i—i—t—i—i—i—i—|—i—i—r~
65
60
55 50
RIV MILE
~1 1 1 1
45
Figure VI-9. Observed (*) and simulated ( ) values of chlorophyll a (ppm)
-------
have between 550 and 600 ppm chlorides without any dilution flow.
Clearly, the two sets of measurements do not correspond but the
effect on the simulated results is minimal.
Ultimate Carbonaceous Biochemical Oxygen Demand
The major sources of simulated UCBOD in the river are the
Champion effluent (53.5 mg/1) and the Clyde and Waynesville WWTPs
(24.0 and 32.4 mg/1, respectively) (Fig. VI-2). Although the
concentrations of these discharges appear somewhat similar, the
slow rate of degradation of the Champion BOD materials (i.e., a
linear degradation rate of 11% per day) results in the WWTPs'
flows diluting the instream concentrations of UCBOD. The primary
mechanism affecting UCBOD concentration in the Pigeon River is
dilution. Streamflow increases by approximately a factor of 2
from the Champion discharge to Walters Lake, resulting in the
reduction of any conservative material by 50%. UCBOD decreases
from 53 ppm at the Champion discharge to 21 ppm at Walters Lake
(decrease of 60%). Given the further additions of' UCBOD
materials from the WWTPs and the tributaries, oxidation of
carbonaceous UCBOD materials can account for only about 10-20% of
the modeled reduction in UBOD. With a travel time of
approximately two days, this reduction corresponds to a
degradation rate on the order of 10% per day.
Dissolved Oxygen
Simulated upstream DO concentrations were relatively high
(7.6 ppm) and augmented at the Champion discharge by the domed
cascade which returns the riverflow at 9.7 ppm (Fig. VI-3). This
supersaturated concentration is reduced largely by deaeration to
approximately 7 ppm by River Mile 62.4, the location of the first
sidestream oxygenator. There, approximately 37% of the river is
withdrawn and supersaturated to 30 ppm resulting in an immediate
dissolved oxygen concentration of about 14 ppm at the oxygenator
discharge. Below the first sidestream oxygenator, simulated DO
concentrations are again reduced by deaeration to about 9 ppm
when a second sidestream oxygenation unit withdraws about 37% of
the flow and supersaturates it to about 30 ppm, resulting in a
local concentration of 16 ppm. Simulated dissolved oxygen
steadily decreases from deaeration to about 6 ppm when Richland
Creek enters the Pigeon River, significantly increasing
streamflow and dissolved DO concentrations. The remainder of
river is characterized by increased slope, and natural reaeration
and high-DO tributary flows combine to increase to simulated DO
about 9 ppm at the river's confluence with Walters Lake.
Simulated dissolved oxygen (DO) concentrations were
determined in the model using all sources and sinks for oxygen.
Vi-ll
-------
All of these sources and sinks were measured (i.e., photo-
synthesis, sediment oxygen demand, UCBOD, NBOD) except
reaeration/deaeration. The Tsivoglou and Wallace (1972) estimate
of reaeration is proportional to the change in elevation for a
stream reach; and represented by:
K_ = ac S„ //"
L e
whe re
a = constant to convert to units of feet/day
c = escape coefficient (ft-*)
it = mean velocity in reach (ft/sec)
Se = slope of the energy gradient (ft/ft).
The slope of the energy gradient (Sg) was estimated from
available data concerning the slopes of individual stream reaches
and are shown in Table VI-1. The value of the escape coefficient
for each stream reach was determined by calibration (i.e., by
difference) to the observed September 27, 1989 data set as all
other parameter values were available. The escape coefficients
for each reach are shown in Table VI-1. The only constraint
placed upon the calibration was that the escape coefficients be
similar to the values determined by Tsivoglou and Neal (1976).
The escape coefficient values in the region of the Pigeon River
not affected by the sidestream oxygenators are either 0.050 or
0.045 ft which are relatively close to Tsivoglou and Neal's
(1976) recommendation of 0.054 ft for streams at 20° C and
flows of 15-3000 cfs. Calibration of the escape coefficients in
the region near the sidestream oxygenators decreased to 0.020-
0.025 ft as the oxygen deficit increased substantially in these
areas (e.g., deficit is > 8 ppm above saturation at sidestream
oxygenator #3) .
The lesser escape coefficients in the area below the
oxygenators may result from an underestimate of oxygen infusion
rates. Based on information supplied by Champion, the
oxygenators were infusing 18 ppm into the withdrawn water. This
should have led to oxygen levels downstream of the final
oxygenator of less than 15 ppm. Actual measurements, though,
revealed DO levels in excess of 20 ppm. The infusion rate
necessary to account for 20+ ppm concentrations in the water
would be about 45 ppm.
Comparison of our measured deaeration rate from the pan
experiment (1.8 ppm/hr) and our modeled rate (1.2 ppm/hr) further
suggest that our modeled escape coefficients in the reaches
containing the oxygenators were low. Because the river is
subject to turbulence not experienced in the pans, the deaeration
VI-12
-------
Table VI-1. Coefficients used in Tsivoglow and Wallace (1972)
estimation of reaeration/deaeration
Reach
#
Slope of
Energy Gradient
(Se) (ft/ft)
Escape
Coefficient
(c) (ft l)
Mean Tsivoglow
and Wallace
K2 (day )
1
0.0019
0.050
3.640
2
0.0019
0.050
3.825
3
0.0019
0.020
1.605
4
0.0019
0.025
2.910
5
0.0019
0.045
4.632
6
0.0019
0.045
3.657
7
0.0034
0.045
5.446
8
0.0034
0.045
5.856
9
0.0049
0.045
12.440
10
0.0049
0.045
18.232
VI-13
-------
rate in the experiment should be less than that in the model.
Doubling of the infusion efficiency, as suggested above, would
resolve this inconsistency.
Changes to the supplied infusion rate were not made in the
model because measured rates were not available. However, this
change in oxygen infusion rate would have relatively little
effect on the simulated downstream oxygen levels since the
oxygen values at stations S3 and S4 are simulated correctly
(i.e., the low DO values after the effects of the oxyge
nators dissipates). The only effect would be to roughly double
the escape coefficients for the two stream reaches (i.e., #3 and
#4) making these values similar with the remaining escape
coefficients. The effect of this parameter change would be to
simply force the additional oxygen into the atmosphere.
Nitrogen
Below the Champion discharge, simulated TKN valves increased
from near zero to about 2 ppm (Fig. VI-4). This TKN con-
centration is reduced by about 50% (to 1 ppm) in the region
between the Champion discharge and Richland Creek and declines
steadily thereafter. Figure VI-6 shows an influx of NH,-N from
the Champion discharge (1.1 ppm) which is reduced by 70%
(converted to nitrite and then nitrate) within the region
between the Champion discharge and Richland Creek. Ample oxygen
supply is available in this region due to the oxygenators to
reduce the available ammonia. Simulated nitrite-nitrogen, a
relatively short-lived component, builds up to about 0.15 ppm and
is reduced to less than 0.05 ppm before any sewage treatment
discharges are added to the river (Fig. VI-6). Finally, Figure
VI-7 shows the steady increase in nitrate between the Champion
discharge and Richland Creek due to the oxidation of ammonia, the
dilution by Richland Creek, the continued oxidation (at a slower
rate) between Richland Creek and Jonathans Creek, and its
subsequent dilution by Jonathans Creek. The increase of
nitrates-nitrogen from 0.3 to 1.1 ppm prior to the addition of
any sewage treatment effluents shows a significant oxidation of
nitrogenous materials in the region of the river which includes
the Champion oxygenators. Oxidation of ammonia requires 4.57 mg
O2 for each milligram of ammonia-N; thus, significant dissolved
oxygen reduction can be attributed to the oxidation of ammonia in
this region of the river. Due to the degraded condition of the
Pigeon River waters (i.e., dark color), this increased nitrate
level was found not to be used to produce algal biomass. As a
result, the increased simulated nitrate load flows into Walters
Lake.
VI-14
-------
Phosphorus
The primary sources of simulated dissolved phosphorus are
the Champion effluent (0.53 ppm) and the Clyde and Waynesville
STPs (3.6 and 1.3 ppm, respectively) (Fig. VI-8). Dissolved
phosphorus is not consumed in the river due to the lack of algal
production, but is diluted from about 0.6 ppm to 0.3 ppm by the
influx of low phosphorus tributary waters).
Chlorophyll a
The simulated values of chlorophyll a are extremely low at
the upstream boundary of the model, and increase only slightly in
downstream areas (Fig. IV-9). The low simulated values are
uncharacteristic of many freshwater streams but accurately
reflect field data from the river.
Simulation of chorophyll a underestimated the observed
chlorophyll values. The field data show a 400-600% increase in
chlorophyll in the region of Crabtree Creek and Jonathans Creek.
Given the streamflows of these tributaries, chlorophyll a
concentrations in excess of 100 mg/1 would be required from
Crabtree Creek or 20-30 mg/1 from Johnathons Creek to produce to
observed 5-6 mg/1 chlorophyll at River Mile 43. Both of these
estimated chlorophyll requirements from the Pigeon River
tributaries substantially exceed the measured chlorophyll a
concentrations in these tributaries. Documentation of the
calibration model run is provided in Appendix A.
B. VALIDATION
Validation 1-28 September 1988
The values of observed UCBOD (estimated using the 5-day BOD
measurements adjusted by the UCBOD/BOD^ ratio measured on 27
September 1988 were higher than those observed on 28 September.
The simulation using these higher effluent loads matched the
observed data (Figure VI-10). Regression of observed ^nd modeled
UCBOD data produced a significant statistical model (R =0.98)
with a slope and intercept meeting the required criteria. The
simulated dissolved oxygen simulation appeared to overestimate
observed DO by less than 1.0 ppm (Fig. VI-11). Regression of
observed and simulated data met the validation acceptance
criteria with a intercept of 0.93+5.0 and a slope of 1.05+0.6.
The Pigeon River QUAL2E model was determined to validly represent
the conditions occurring in late summer 1988. Appendix B shows
the documentation of the QUAL2E Validation I run.
VI-15
-------
90
80
70
60
50
40
30
20
10
0
.0.
t—i—i—r
T 1 1—T
t—i—i—r
t—i—i—i—i—r
t—r
t—r
T
T
T
T
T
T
T
T
T
T
T
T
T
60 55 50 45 40
River Mile
Validation of observed (*) and simulated ( ) values of ultimate carbon-
aceous biochemical oxygen demand (ppm) for 28 September 1988.
-------
Figure VI-11. Validation of observed (*) and simulated ( ) values of dissolved oxygen
(ppm) for 28 September 1988.
-------
Validation II - 7 July 1987
When natural streamflow exceeds Champion withdrawal rate
(Fig. VI-12), the effects of chloride additions from the Champion
facility (507 ppm) on the Pigeon River are reduced (Fig. VI-13).
The immediate effect of the discharge on chlorides is to increase
concentrations from 2 ppm to 215 ppm. Chlorides are subsequently
diluted by tributary flows.
Simulations of carbonaceous BOD^ during the higher flow
regime underestimates observed BOD^ in the region between the
Champion discharge and Richland Creek (Fig. VI-14). However, one
of the observed BOD,- values taken just downstream of the
discharge at the FiServille Bridge shows a BOD^ value of 15.0.
Examination of Champion self-monitoring data suggest that this
field measurement was in error. In addition, the BOD^ measure
depicted in the EA report represents all biochemical demand
including nitrogenous demand. The simulation only represents
carbonaceous demand and early simulation runs (nominal case)
suggest that nitrogenous demand is high in the portion of the
river between the Champion discharge and Richland Creek.
Under the higher flow regime, Champion's oxygenators have a
much reduced effect on local DO concentrations (Fig. VI-15).
According to the observed data, water between the third
oxygenator and the Clyde WWTP deaerated too rapidly, yet the
appropriate DO concentration was reached by River Mile 54. In
general, the simulated DO concentrations match the observed DO
values.
Total Kjeldahl nitrogen and ammonia were consistently
overestimated by the model by about 1 ppm (Figs. VI-16, Vl-17),
but the simulated concentration of nitrate closely resembled the
observed data (Figure VI-18). Since nitrate is generated in the
Pigeon River system primarily by the oxidation of ammonia and it
is not being consumed by algal photosynthesis, the disparity
would suggest that at least one of the field measurements
collected in 1987 was in error.
Figures VI-19 and VI-20 describe the simulated behavior of
total organic phosphorus and dissolved phosphorus under high flow
conditions. Once the phosphorus compounds are added to the
Pigeon River at the Champion discharge, changes in phosphorus
levels are primarily the result of dilution; the river is moving
too rapidly to permit settling, and the residence time is too
short and the light penetration too minimal to permit sufficient
photosynthesis to reduce the dissolved phosphorus concentration.
With the exception of total organic nitrogen and ammonia,
all simulations of the 7 July 1987 data set meet the criteria for
validation. Documentation of this validation simulation is
provided in Appendix C.
VI-18
-------
400 -
300
c
f
s
200
100
l 1 r
65
"I 1 1 1 1 1 1 T
60
t—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—r-
1
55 50
River Mile
45
40
Figure Vl-12. Validation of observed (*) and simulated ( ) streamflow values (CFS) for
7 July 1987.
-------
300 1
*
200
P
P
m
C
L
100
0
-i—i—i—r
-i—i—i—i—i—i—r
65
60
Figure VI-13. Validation of observed
7 July 1987.
*
I 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 I 1 1 1 1 1 1 1 1 1 1 1 1 1 j~
55 50 45 40
River Mile
*) and simulated ( ) chlorides values (ppm) for
-------
15 "
10 -
i—i—i—r
1 r
t—r
t—r
i i i—r
t—r
T
t—r
t—r
T
T
T
T
1 r
T
T
T
T
T
T
T
T
65 60 55 50 45 40
River Mile
Figure VI-14. Validation of observed (*) total 5-day biochemical oxygen demand and
simulated ( ) carbonaceous BOD^ values (ppm) for 7 July 1987.
-------
River Mile
Figure VI-15. Validation of observed (*) and simulated ( ) dissolved oxygen values (ppm)
for 7 July 1987.
-------
5 :
4:
65 60 55 50 45 40
River Mile
Figure Vl-16. Validation of observed (*) and simulated total Kjeldahl nitrogen values
(ppm) on 7 July 1987.
-------
5 :
4
i
S3
m
N
H «
0
65
~I 1 1 1 1 1 1 1 ! 1 1 1 1 1 I j 1 1 1 \ 1 ( 1 1 1 J 1 I 1 1 I 1 1 1 \ 1 1 1 ( 1 1 1 1 1 1 p
60 55 50 45 40
River Mile
Figure VI-17. Validation of observed (*) and simulated ( ) NH,-N values (ppm) on
7 July 1987.
-------
River Mile
Figure Vl-18. Validation of observed (*) and simulated nitrate-nitrogen values (ddh) on
7 July 1987.
-------
1.00-
0.75 "
0.50
0. 25
¥
0.00
|—i—i—i—i—i—i—i—i—i—|—r
65 60
t—I—i—r
Figure VI-19. Validation of observed
(ppm) on 7 July 1987.
*
* *
I 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 l—T 1 1 1 1 1 1 1 1 1 1 1 1~
55 50 45 40
River Mile
*) and simulated total organic phosphorus values
-------
1.00-
0.75
50
0.25
0.00
r~
65
i—i—i—i—i—i—i—r
n—i—i—r
60
Figure VI-20. Validation of observed
(ppm) on 7 July 1987.
V
1 I I I I I I I I I I I* I " I I I I I I I I I I I I I I—I I I—I 1—p
55 50 45 40
River Mile
*) and simulated( ) dissolved phosphorus values
-------
C. MODEL UNCERTAINITY
The results of the 3500 Monte Carlo simulations comprising
the uncertainty analyses indicate that point loads have the most
pronounced effect on the prediction uncertainties of the nine
modeled constituents (Table VI-2, Appendix D). These modeled
variables responded linearly to the modification of point
discharge concentrations. The coefficients of variation for the
nine modeled state variables ranged from 9% for NO, to 15% for
BOD, TKN, and PO^. Chlorides, as would be expected, are
sensitive only to modifications in the discharge of chlorides and
streamflow (Appendices D and E). Dissolved oxygen concentrations
are most affected by point source loads (Appendix D) but are
minimally affected by changes in the reaction coefficients
(Appendix F). Regardless of the uncertainty modifications, the
dissolved oxygen concentrations at the confluence of the Pigeon
River with Walters Lake are nearly constant. UBOD concentrations
were directly affected by changes in point source loads
(Appendix D), but were unaffected by a + 15% change in the BOD
decay rate (Appendix G). Concentrations of nitrogen and
phosphorus were more sensitive to changes in reaction co-
efficients (Appendix H) than DO or UBOD. Modifications to the
upstream concentrations of DO, UBOD, nitrogen, phosphorus, or
chlorophyll a had no affect on local concentrations of these
state variables (Appendix I). Only the modification of headwater
streamflow (Appendix E) affected downstream concentrations of
these state variables and this affect was generally minor (2-6%).
VI-28
-------
Table VI-2.
Results of Monte Carlo uncertainty analysis.
Coefficient of variation represents the error in the
model prediction associated with the aggregate error
in the inputs and parameters.
Uncertainty
Coefficent
of Variation
(%)
Scenario
CHL
CL1
DO
UCBOD
TKN
nh3-n
no2-
-N
no3-n
P04
UNC-12
10
11
15
15
13
11
9
14
15
UNC-2
<1
<1
<1
<1
<1
<1
<1
<1
<1
UNC-3
6
1
3
2
2
2
6
3
4
UNC-4
0
5
2
1
20
20
10
<1
3
UNC-5
0
1
1
0
0
0
0
0
0
UNC-6
0
2
1
5
20
20
10
1
20
UNC-7
10
7
16
16
22
25
15
14
26
CL=Chlorides; DO=Dissolved Oxygen; UBOD=Ultimate Biochemical
Oxygen Demand; TKN=Total Kjeldahl Nitrogen; NHg=Ammonia;
N02=Nitrite; NO,=Nitrate; PO.=Dissolved Phosphorus;
CHL=Chlorophyll a
UNC-1: Point source concentrations varied (Appendix D)
UNC-2: Headwater water quality concentrations varied (Appendix I)
UNC-3: Headwater water quality concentrations and flow varied
(Appendix E)
UNC-4: Dissolved oxygen reaction coefficients varied (Appendix F)
UNC-5: UBOD reaction coefficients varied (Appendix G)
UNC-6: Nutrients reaction coefficients varied (Appendix H)
UNC-7: All model parameters varied except headwater streamflow
(Appendix J)
VI-29
-------
VII. PIGEON RIVER MODEL SCENARIOS
To evaluate the
the Champion, Clyde,
DO, BOD, and ammonia
the following set of
River QUAL2E model:
effects of alternative
and Waynesville dischar
concentrations, EPA Reg
5 scenarios be examined
loading conditions at
ges on Pigeon River
ion 4 requested that
using the Pigeon
o Scenario #1 - Examine the effects of removing the Clyde
and Waynesville WWTPs discharges
o Scenario #2 - Examine the effects of assigning maximum
secondary treatment standards for BOD (30 ppm) and
ammonia (15 ppm) at the Waynesville WWTP
o Scenario #3 - Examine the effects of reducing the
Champion discharge flow to 30 MGD (i.e., 46.4 cfs)
while maintaining the present concentrations in the
Champion effluent (i.e., reduced loadings)
o Scenario #4 - Examine the combined effects of maximum
secondary treatment standards at the waynesville WWTP
and reduction of the Champion discharge to 30 MGD
o Scenario #5 - Examine the effects of removing the two
sidestream oxygenators.
The validated Pigeon River QUAL2E model was used to evaluate each
of these alternative loading scenarios. The results of the model
evaluations are described below by specific scenario.
Scenario #1 - Removal of WWTP Discharges
The input values for the point source flows corresponding to
the Clyde and Waynesville WWTPs were altered to represent zero
discharge from these facilities and the model was rerun. Figures
VII-1 through VII-3 show the effects of this management scenario
on dissolved oxygen, BOD, and ammonia, respectively, and the
model run is documented in Appendix K. The removal of the WWTP
flows has no effect on the dissolved oxygen concentrations in the
river; the nominal case (i.e., 27 September validated model;
solid line in Figure VII-1) is not different from the scenario
case (i.e., dashed line on Figure VII-1). The removal of the
WWTP discharges has no effect on the the BOD concentrations in
the river (Figure VII-2) and results in very minor reduction in
VII-1
-------
I
K)
DO
16
14
12
10
8
65
—j—i—i—i—i—i—r
60
55
1
50
I I 1 I I [ I I I l
45
40
RIV MILE
Figure VII-1. The simulated effect of removing the discharges of the Clyde and Waynesville
WWTPs on Pigeon River dissolved oxygen concentrations (Nominal run = ;
Scenario = —).
-------
BOD
100 -
30 "
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
65 60 55 50 45 40
RIV MILE
Figure VII-2.
The simulated effect of removing the discharges of the Clyde and Waynesville
WWTPs on Pigeon River UCBOD concentrations (Nominal run = ;
Scenario = — ) .
-------
<
w
M
I
NH3
1.0
0.8
0.6
0.4
0.2
0.0
1—i—i—i—i—i—i—i—i i | r
65 60
-i—i—i—i—i—|—i—i—i—i—i—i—i—i—i | i i i i i i i i 1 p
55
50
RIV MILE
~i—i—i—i—i—i—I—i—r~
45
40
Figure VII 3. The simulated effect of removing the discharges of the Clvde and Wavne«;vm»
WWTPs on Pigeon River NH,-N concentrations (Nominal run =• Waynesville
Scenario = —). 3 —'
-------
ammonia downstream of Richland Creek (i.e., < 0.1 ppm; Figure
VII-3). The removal of the WWTPs from the Pigeon River system
has no significant effects on the dynamics of oxygen, BOD, or
ammonia in the river.
Scenario #2 - Secondary Treatment at the Waynesville WWTP
The input concentrations for the Waynesville point source
discharge were altered to represent secondary treatment maximums
for BOD and ammonia (i.e., 30 ppm and 15 ppm respectively) and
the model was rerun. These modifications represented a minor
decrease in the BOD load associated with the Waynesville facility
(i.e., < 3 ppm) and a significant increase in ammonia discharge
(i.e, from 0.3 ppm to 15.0 ppm). Figures VII-4 to VII-6 show the
effects of secondary treatment at Waynesville WWTP on the
dissolved oxygen, BOD, and ammonia concentrations in the Pigeon
River, respectively, and Appendix L documents the model run. The
increased ammonia load being discharged by Waynesville in this
scenario (Figure VII-6) produces an ammonia spike in the river
equivalent to that presently produced by the Champion effluent
(i.e., about 1.0 ppm). The slightly reduced BOD loading in the
scenario does not affect the simulated BOD concentrations in the
river. The increase of ammonia, however, does create a simulated
dissolved oxygen sag of about 1.0 ppm below the Waynesville
discharge for a distance of about 5 miles downstream (Figure
VII-4).
Scenario #3 - Reduction of Champion Discharge Rate to 30 MGD
The input discharge rate for the Champion facility was
reduced to 30 MGD (i.e., 46.4 cfs) and the model was rerun
(Appendix M). This management scenario represents a 55%
reduction in the discharge rate of the Champion facility.
Reductions in BOD loading from the Champion facility reduced
simulated BOD concentrations in the river by about 20 ppm near
the Champion discharge (Figure VII-7). Further downstream,
dilution acted upon this concentration in the scenario simulation
in the same manner as in the nominal simulation, reducing the
concentration of simulated BOD from about 35 ppm at the Champion
discharge to about 18 ppm at the confluence with Walters Lake.
Similarly, simulated ammonia concentrations in the river were
reduced by about 20-25% near the Champion discharge, but returned
to nominal levels by River Mile 42 (Fig. VII-8). The reduction
of the Champion discharge rate had little effect on dissolved
oxygen concentrations (Figure VII-9). In fact, reduction of the
Champion discharge rate actually lowered oxygen concentrations
slightly (< 0.5 ppm) in the vicinity of the discharge. This
reduction in DO was due to the proportional reduction in
VII-5
-------
Figure VII-4. The simulated effect of required secondary treatment standards (BOD=30 ppm
and NHt=>15 ppm) at the Waynesville STP on Pigeon River dissolved oxygen
concentrations (Nominal run = ; Scenario = —).
-------
BOD
100
90
80
70
60
50
40
30
20
10
0
re V
T—I—j—I—I—I—I—I—I—I—T
t—r
t—r
t—r
t—i—i—i—i—j—r
T
5 60 55 50 45 40
RIV MILE
-5.
The simulated effect of required secondary treatment standards (BOD=30 ppm
and NHt = 15 ppm) at the Waynesville STP on Pigeon River UCBOD concentrations
(Nominal run = ; Scenario = —)•
-------
I
oo
NH3
1 .0
0 . 8
0.6
0 . 4
0 . 2
0.0
1 I I I I T
~T 1 1 1 1 1 1 J 1 1 1 1 1 J I I I I 1 I I I I | I I I I I I 1 I I | I I I I I I I 1 1 1~
65
60
55 50
RIV MILE
45
40
Figure VII-6. The simulated effect of required secondary treatment standards (BOD=30 ppm
and NH3=15 ppm) at the Waynesville STP on Pigeon River NH,-N concentra-
tions (Nominal run = ; Scenario - —).
-------
BOD
100
90
80
70
60
50
40
30
20
10
0
re V
5 60 55 50 45 40
R I V__M I L E
1-7. The simulated effect of reducing the Champion discharge to 30 MGD while
maintaining present effluent concentrations on Pigeon River UCBOD
concentrations (Nominal run = ; Scenario = —).
-------
NH3
1.0:
0.8
0.6
0 . 4
0.2
0 . 0
1—i—i—i—i—i i 1—i I—J J i I J I i I i i J I I I I I l l I I | I I I I i I I I I—| I l—I—I—i—j—;—i—i—1~~
65 60 55 50 45 40
RIV MILE
Figure VII-8. The simulated effect of reducing the Champion discharge to 30 MGD while
maintaining present effluent concentrations on Pigeon River NH^-N
concentrations (Nominal run => ; Scenario = —).
-------
Figure VII-9. The simulated effect of reducing the Champion discharge to 30 MGD while
maintaining present effluent concentrations on Pigeon River dissolved oxygen
concentrations (Nominal run = ; Scenario = —).
-------
artificial oxygenation at the point of discharge. However,
dissolved oxygen concentrations were increased above the nominal
condition by the first downstream oxygenator, and due to de-
creased ammonia loads (Figure VII-8) and to some extent reduced
BOD loads(Figure VII-7), the rate of decrease in dissolved oxygen
between the final oxygenator and Richland Creek was slowed. This
resulted in a simulated increase of about 1 ppm at the confluence
with Richland Creek. Subsequent reaeration resulted in returning
the dissolved oxygen levels in Pigeon River under this scenario
to the same concentration observed in the nominal simulation at
the river's entrance to Walters Lake (Figure VII-9).
Scenario #4 - Reduced Champion Discharge and Secondary Treatment
at Waynesville WWTP
The input discharge rate for Champion was reduced to 30 MGD
(i.e., 46.4 cfs) and the input concentrations for the Waynesville
WWTP were altered to maximum secondary treatment levels (i.e.,
30 ppm for BOD, and 15 ppm for ammonia) and the model was rerun.
As described above, the modification in the ammonia effluent for
the Waynesville WWTP represents a substantial increase over the
nominal simulation. Figures VII-10 to VII-12 show the effects of
this management scenario on the dissolved oxygen, BOD, and
ammonia concentrations in the Pigeon River, respectively, and the"
model run is documented in Appendix M. Under this management
scenario, both simulated BOD and ammonia concentrations are
changed in the river; BOD is reduced primarily due to the reduced
Champion discharge (Figure VII-11), and ammonia is increased due
to the increased loading from the Waynesville WWTP (i.e., 15 ppm
NH, compared to 0.3 ppm in the nominal simulation; Figure VII-
12). These changes in BOD and ammonia result in an increase in
simulated dissolved oxygen in the region between the Champion
discharge and River Mile 54 and a decrease in simulated dissolved
oxygen between River Mile 54 and River Mile 47 (Figure VII-10).
Simulated dissolved oxygen in this scenario returns to the
nominal level by the Piegon River's confluence with Walters Lake
(Figure VII-10).
Scenario #5 - Removal of Sidestream Oxygenators
The inputs of the nominal Pigeon River model were altered to
eliminate the two downstream oxygenators by reducing their
withdrawal and discharge rates to zero, and by increasing the
Tsivoglou-Wallace reaeration coefficients for the reaches
associated with the sidestream oxygenators. In this scenario,
these reaches were given values equivalent to those assigned to
the remainder of the river. Figures VII-13 to VII-15 show the
VII-12
-------
DO
16
14
12
10
8
~i 3 ' ' ' ' ' ' i j—i—i—?—»—i—i—i—i—i—j—i—i—i—i—j—i—i—i—i—|—i—i—j—i—i—i—i—i—r—p
~t—i 1—i—i—i—r
65
60
55
50
45
40
RIV MILE
Figure VII-10. The simulated effects of reducing the Champion discharge to 30 MGD and re-
quiring secondary treatment standards at the Waynesville STP on Pigeon
River dissolved oxygen concentrations (Nominal run = ; Scenario = —).
-------
BOD
100 "
90
T
T
T
T
T
T
T
65 60 55 50 45 40
RIV MILE
Figure VII-11. The simulated effects of reducing the Champion discharge to 30 MGD and re-
quiring secondary treatment standards at the Waynesville STP on Pigeon
River UCBOD concentrations (Nominal run » ; Scenario - —).
-------
NH3
1 . 0
0.8
0 . 6
0 . 4
0 . 2
0 . 0
—,—i—i—i—i—i—i—i—|—i—i—i—i—i—i—i—i—i—|—i—i—i—i—i i i i i | i i > i i > > ' 1 i 1 1 1 1 > ¦ 1 1 1 r
65 60 55 50 45 40
RIY MILE
Figure VII-12.
The simulated effects of reducing the Champion discharge to 30 MGD and re-
quiring secondary treatment standards at the Waynesville STP on Pigeon
River NH^-N concentrations (Nominal run = ; Scenario = —).
-------
Figure VII-13. The simulated effects of removing the two downstream oxygenators on
Pigeon River dissolved oxygen concentrations (Nominal run = ;
Scenario = —).
-------
BOD
100
90
80
70
30
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
65 60 55 50 45 40
RIV MILE
Figure VII-14. The simulated effects of removing the two downstream oxygenators on
Pigeon River UCBOD concentrations (Nominal run = ; Scenario = —).
-------
I
I—'
00
NH3
1 . 2
1 . 0
0 . 8
0 . 6
0 . 4
0 . 2
0 . 0
"~j—i—i—i—i—i—i—i i j i i I i i I I i i j i i i i i i i i i j i i i i I I I i I | i i r~~i—i—i—i—i—i—p
65 60 55 50 45 40
RIV MILE
Figure VII-15. The simulated effects of removing the two downstream oxygenators on
Pigeon River NH^-N concentrations (Nominal run -» ; Scenario =* —).
-------
effects of sidestream oxygenator removal on dissolved oxygen,
BOD, and ammonia concentrations, respectively, and the model run
is documented in Appendix N. The removal of the oxygenators
produces a small increase in the simulated values of BOD (Figure
VII-14) and ammonia (Figure VII-15). These increases are due to
reductions in the simulated oxidation rate of these compounds
resulting from low-level inhibition processes for BOD degradation
and nitrogen oxidation. In the QUAL2E model, low-level
inhibition is initiated when dissolved oxygen levels fall below
5.0 ppm. The major parameter affected by the removal of the
oxygenators is dissolved oxygen. After initial oxygenation to
about 9.0 ppm at the Champion discharge, the reduction of oxygen
is controlled by deaeration, oxidation of nitrogen, and microbial
activity until simulated dissolved oxygen reaches a minimum
concentration of slightly less than 4.0 ppm at River Mile 60
(Figure VII-13). Due to increased elevation gradient, some
reaeration occurs between River Mile 60 and 55, and after the
inflow from Richland Creek enhances the simulated DO
concentrations in the river, there is little difference between
the scenario and the nominal simulation.
VI1-19
-------
VIII .
DISCUSSION
Under loading conditions prevailing during the time of our
study, the Champion Mill effluent had a greater effect in
determining the lowest oxygen concentrations in the river than
discharges from either the Waynesville or Clyde WWTPs. The
lowest DO value we found occurred upstream of Richland Creek.
Given its location, it is unlikely that the dissolved oxygen
minimum results from the WWTPs effluents. Most of the decline
below saturation had already occurred before the effluent from
the Clyde WWTP entered the river, and recovery was apparent prior
to introduction of the effluent from the Waynesville WWTP.
Our modeling simulations confirmed the relative importance
of Champion Mill effluent on oxygen dynamics. When the modeled
effluents from the Champion Mill were reduced by 55%, simulations
indicated there would be more than a 1 ppm improvement in
dissolved oxygen concentration in the river near Richland Creek.
When the inputs from the Waynesville and Clyde WWTPs were totally
removed from the model, there was no detectable change in river
concentration of dissolved oxygen. The dominance of the Champion
effluent was also evident from the relative insensitivity of
dissolved oxygen to other inputs in the uncertainity analyses.
These findings are not surprising given the relative volumes of
the effluents observed in our studies. Loadings of nitrogen and
ultimate BOD were 40 and 10 times higher, respectively, in the
Champion Mill discharge than in the discharge from the two WWTPs
combined.
The model run in which the Champion effluent was reduced by
55% identified that most of the improvement in downstream oxygen
conditions was as a result of reductions in nitrogen inputs.
This finding was consistent with the uncertainty analysis which
found the model to be significantly affected by changes in
nitrogen loading. Presumably the importance of nitrogen is a
result of oxidation of ammonia (i.e., consuming 4.57 ppm of
oxygen for every ppm of ammonia).
Ammonia concentrations discharged from all three effluents
during the study period were substantially less than their
yearly average concentrations (Table VIII-1). This may be the
reason oxygen values were never found to be less than 6 ppm
during our study. Oxygen values less than 5 ppm have been
frequently seen during previous low flow summer conditions.
Additional simulations in which the ammonia concentrations of the
Champion and WWTP effluents were increased to twice their annual
average value suggested that most of the depletion associated
VIII-1
-------
Table VIII-1. BOD and ammonia concentrations in the three effluents on 27 September
1988 compared to most recently available yearly average effluent
concentrations.
BOD5 Ammonia
This Yearly Yearly This Yearly Yearly
Study Average Maximum Study Average Maximum
Champion^3) 4.3 12.8 47.6 1.1 2.2 9.8
Mill
Clyde WWTP
-------
with low oxygen conditions result from quality of the Champion
Mill effluent. When ammonia concentration in the WWTP effluents
were increased, there was no significant change in oxygen
concentration in the river. However, simulated increases in
ammonia concentration from the Champion Mill effluent resulted in
a further decrease of almost 2 ppm in DO levels near the mouth of
Richland Creek to a simulated dissolved oxygen level of 4.7 ppm.
Although the discharge of the Champion Mill contained an
ultimate carbonaceous BOD of greater than 50 ppm, very little of
this material was found to degrade within our study area. The
low rate of degradation recorded for the discharge (BOD^ < 5 ppm)
and the rapid travel time to reach the mouth of Walters Lake (~ 2
days) were the primary reasons for the lack of degradation
within the study area.
in most low gradient lotic systems, algal photosynthesis and
respiration, as well as respiration of benthic biota, play an
important role in regulating instream DO patterns. In the Pigeon
River, however, these factors were found to be relatively un-
important. Measured sediment oxygen demand values were low and
had little effect on the model predictions. However, the values
we measured were consistent with that from a previous study of
sediment oxygen demand in the Pigeon River (Weston 1983).
Primary production was negligible in our study, reflecting the
low chlorophyll numbers (< 10 mg/1) seen throughout the study
area. This was confirmed by the almost total absence of diurnal
variation in the parameters that are normally affected by
photosynthesis. Presumably the low primary productivity in the
presence of abundant nitrate results from the reduced light
penetration associated with the Champion effluent.
The two sidestream oxygenation units had a substantial
influence on oxygen content of the river between the Champion
Mill and Richland Creek. Simulations removing the oxygenation
units identified that dissolved oxygen content would fall to less
than 4 ppm near River Mile 60, consistent with values of less
than 4 ppm that were observed in monitoring efforts prior to
installation of these units. However, the model run also showed
that the oxygenation units had no measurable effect on the river
downstream of Richland Creek, where reaeration and dilution by
tributaries were the dominant factors controlling recovery from
the depletion zone. The model also identified that most (60-70%)
of the oxygen added by the units is lost to the atmosphere, and
has little effect on decay rate of carbonaceous BOD from the
mill. Given that the oxygen infusion rate that was supplied to
Versar is an apparent underestimate (measured DO values
immediately downstream of the units were higher than would be
produced given the infusion rate), the percentage loss to the
atmosphere is likely to be an underestimate. Since, the observed
DO decline is mimiced by the model, and the decay rates (both
CBOD and NBOD) have sufficient oxygen to continue uninhibited at
VII1-3
-------
the simulted oxygen levels, the increase in infused oxygen would
simply be lost to the atmosphere. This would increase the
atmosphericdeaeration to about 80-85% of total infused oxygen.
Thus, it appears the major effect of the oxygenation units is to
maintain DO levels between Champion and Richland Creek above
State water quality standards.
Since the sidestream oxygenation units do not substantially
add to consumption within the river of the carbonaceous BOD load
being added by the Champion effluent, it follows that this BOD
load is being transported to the lake. A mass balance approach
comparing empirical data from the Champion effluent (El) and the
station at the headwaters of Walters Lake (Sll) shows that most
of the difference in concentration of ultimate BOD at these
stations can be attributed to dilution from tributaries; when
weighted by flow there is less than a 20% difference in BOD mass
at these sites. Our model, which leads to the same conclusion
that less than 20% of the BOD loads are being consumed in the
river, provides a semi-independent confirmation of this finding
since it is based on point source loadings and process rates,
rather than on downstream concentrations. The model conclusion
relies heavily on the high ultimate carbonaceous. BOD inputs the
UCBOD to 5-day BOD ratio was 12:1, higher than average for paper
mill effluents (NCASI 1982). However, the fact that this ratio
exceeds 10:1 for the mill effluent was confirmed by a monitoring
sample from the Champion effluent that was collected by the NCDEM"
on the same day as our study (Table VIII-2). This high ratio may
be a reflection of the lignocellulosic content and relatively
large size of fiber particles observed suspended in the water
column downstream of the Champion Mill effluent.
Very little information is available to examine the effect
of this source contribution to the lake, but what data are
available suggest that the lake is severely impacted. Data
collected in 1973 (STORET) and in 1982 (Herlong et al. 1982) both
show that DO levels near the surface were less than 2 ppm in late
summer/early fall. Values near the bottom of the reservoir are
even lower. Since Walter's Lake is used as a source for a
hydroelectric facility which has its intake located near the
bottom of the reservoir, there is reason to be concerned that
water being discharged from that facility fails to meet State
standards. Future modeling efforts that extend the downstream
boundary-of the present study are recommended to better describe
the ultimate fate of carbon inputs from the three Pigeon River
di scharges.
VII1-4
-------
Table VIII-2. Results of UBOD analyses from an undiluted Champion
Mill effluent sample collected by NCDEM on
27 September 1988 at 1015
Total
Day BOD (mg/1) NH^-N TKN-N NOX-N Total N
0 1.1 3.1 0.01 3.1
5 6.8
10 11.6
15 16.1
20 20.3
25 24.5
30 30.2
35 35.4
40 37.9
50 41.6
60 44.8
70 47.7 0.03 0.8 1.7 2.5
83 51.1
95 53.5
106 55.5
118 57.8
140 61.5
155 64.2
175 67.1
201 70.2
223 72.2
VIII-5
-------
IX. LITERATURE CITED
Brown, L.C. and T.O. Barnwell. 1987. The enchanced stream
water quality models QUAL2E and QUAL2E-UNCAS: Documentation
and user model. EPA-600-3-87-007. Prepared for United
States Environmental Protection Agency, Environmental
Research Laboratory. Athens, GA.
Buchanan, T.J., and W.F. Somers. 1969. Discharge measurements
at gauging stations. Book 3, Chapter A8 in: Techniques of
water-resources investigations of the United States
Geological Survey. Washington, D.C.: U.S. Government
Printing Office.
Engineering Science, and Technology, Inc. (EA). 1988. Synoptic
survey of physical and biological condition of the Pigeon
River in the vicinity of Champion International's Canton
Mill. Prepared for Champion International Corporation,
Stamford, Connecticut.
Herlong, D.D., K.A. MacPherson, M.A. Mallin, K.L. Stone, and
P.B. Summers, Jr. 1982. Report on Walters Reservoir
(Waterville Lake). Special sampling of October 26, 1982.
Carolina Power and Light Company, Environmental Technology
Section.
Hickey, C.W. 1988. Benthic chamber for use in rivers: testing
aqainst oxygen mass balances. J. Envir. Eng. 114(4):828-
845.
Loftus, M.E., and J. Carpenter. 1971. A fluorometric method for
determining chlorophylls a, b, and c. J. Mar. Res. 29:319-
338.
National Council of the Paper Industry for Air and Stream
Improvement, Inc. (NCASI). 1982. A review of ultimate BOD
estimation and its kinetic formulation for pulp and paper
mill effluents. Technical Bulletin No. 382.
Nisely, M., and 0. Tysland. 1983. Pigeon River dissolved oxygen
enrichment trial. 1983. Internal Memo to M.L. Ransmeier
dated 25 October 1983. Champion International Corporation,
Canton, NC.
North Carolina Department of Natural Resources and Community
Development, Division of Environmental Management (NCDEM).
1984. Dissolved oxygen modeling in the Pigeon River.
Unpub. Report.
IX-1
-------
Summers, J.K., and H.N. McKellar, Jr. 1981. A sensitivity
analysis of an ecosystem model of estuarine carbon flow.
Ecol. Modelling 13:283-301.
Tsivoglou, E.C., and J.R. Wallace. 1972. Characterization of
stream reaeration capacity. Prepared for U.S. Environmental
Protection Agency, Washington, DC.
Tsivoglou, E.C., and L.A. Neal. 1976. Tracer measurement of
reaeration: III. Predicting the reaeration capacity of
inland streams. J. of Water Poll. Control Fed. 48:2669-
2689.
United States Environmental Protection Agency (EPA). 1983.
Methods for chemical analysis of water and wastes. EPA-
600-4-79-020. Office of Research and Development,
Cincinatti, OH.
Weston, Inc. 1983. Sediment oxygen demand and long-term
biochemical oxygen demand determinations for the Pigeon
River. Prepared for Chamption international Corporation.
Canton, North Carolina.
IX-2
------- |