02_EFDC_Final_Report

1/114

TECHNICAL MEMORANDUM #6:
Hydrodynamic Documentation of the Lower Tar River and Pamlico Estuary using the Environmental Fluid Dynamics Code (EFDC) Lower Tar-Pamlico River Model (LTPR)
To: Copies:
Greenville Utilities Commission
Division of Water Quality
Division of Water Resources
Mary Sadler, Hazen & Sawyer
From:
Lauren Elmore, Paul Leonard, and Alix Matos,
Cardno ENTRIX

Date / Rev: ARCADIS Project No.:
March 2013 NC706015.1000
Subject:
Tar River Flow Study
Revisions for House Bill 609
Greenville Utilities Commission

2/114
This page left intentionally blank.

3/114
Table of Contents
1. Introduction 9
2. Model Description 10
3. Model Setup 11
3.1 Model Grid Extension 13
3.2 Atmospheric Forcing Functions 16
3.3 Boundary Conditions 16
3.3.1 Tar River at Greenville Inflows 16
3.3.2 Tributary Inflows 16
3.3.3 Point Sources Discharges 17
3.3.4 Groundwater Inputs 18
3.3.5 Surface Water Withdrawal 18
3.3.6 Open Boundary Conditions 18
3.3.7 Modeling Years 19
3.4 Model Calibration and Validation 21
3.4.1 Stage 21
3.4.2 Temperature 26
3.4.3 Salinity 59
4. Modeling Scenarios 91
4.1 Development 92
4.2 Results 95
4.2.1 Salinity Habitat Impacts 95
4.2.2 Temperature and Salinity Impacts at Hourly Scale 104
5. LTPR Model Summary 112
6. References 113

4/114
Figures
Figure 3-1: Tar River Flow Study Area 12
Figure 3-2: LTPR Modeling Grid from the Tar River upstream of Greenville to the Mouth of the Pamlico Estuary 14
Figure 3-3: LTPR Modeling Grid between Greenville and Washington, NC 15
Figure 3-4: Annual and Seasonal Median Flows (Sorted by Annual Median) for the Tar River at
Greenville (1998 through 2010) 20
Figure 3-5: Flow Hydrographs for Four Modeling Years 20
Figure 3-6: Simulated and Observed Stage at Washington, NC (2001) 23
Figure 3-7: Simulated and Observed Stage at Washington, NC (2003) 24
Figure 3-8: Simulated and Observed Stage at Washington, NC (2008) 25
Figure 3-9: Time Series Comparison of Simulated and Observed Temperature at Grimesland
in 2001 27
Figure 3-10: Time Series Comparison of Simulated and Observed Temperature at Grimesland
in 2003 28
Figure 3-11: Time Series Comparison of Simulated and Observed Temperature at Grimesland
in 2008 29
Figure 3-12: Box Plot Comparison of Simulated and Observed Temperature at Grimesland in
2001 30
Figure 3-13: Box Plot Comparison of Simulated and Observed Temperature at Grimesland in
2003 30
Figure 3-14: Box Plot Comparison of Simulated and Observed Temperature at Grimesland in
2008 31
Figure 3-15: Time Series Comparison of Simulated and Observed Temperature at St 10 in
2001 32
Figure 3-16: Time Series Comparison of Simulated and Observed Temperature at St 10 in
2003 33
Figure 3-17: Time Series Comparison of Simulated and Observed Temperature at O787000C
in 2008 34
Figure 3-18: Box Plot Comparison of Simulated and Observed Temperature at St 10 in 2001
(Top and Bottom) 35
Figure 3-19: Box Plot Comparison of Simulated and Observed Temperature at St 10 in 2003
(Top and Bottom) 36
Figure 3-20: Box Plot Comparison of Simulated and Observed Temperature at O787000C in
2008 (Top and Bottom) 37
Figure 3-21: Profile Comparison of Simulated and Observed Temperature at O787000C in
2001 38
Figure 3-22: Profile Comparison of Simulated and Observed Temperature at O787000C in
2003 39
Figure 3-23: Profile Comparison of Simulated and Observed Temperature at O787000C in
2008 40

5/114
Figure 3-24: Time Series Comparison of Simulated and Observed Temperature at St 07 in 2001 41
Figure 3-25: Time Series Comparison of Simulated and Observed Temperature at St 07 in 2003 42
Figure 3-26: Time Series Comparison of Simulated and Observed Temperature at O8498000 in 2008 43
Figure 3-27: Box Plot Comparison of Simulated and Observed Temperature at St 07 in 2001 (Top and Bottom) 44
Figure 3-28: Box Plot Comparison of Simulated and Observed Temperature at St 07 in 2003 (Top and Bottom) 45
Figure 3-29: Box Plot Comparison of Simulated and Observed Temperature at O8498000 in 2008 (Top and Bottom) 46
Figure 3-30: Profile Comparison of Simulated and Observed Temperature at O8498000 in 2001 47
Figure 3-31: Profile Comparison of Simulated and Observed Temperature at O8498000 in 2003 48
Figure 3-32: Profile Comparison of Simulated and Observed Temperature at O8498000 in 2008 49
Figure 3-33: Time Series Comparison of Simulated and Observed Temperature at St 03 in 2001 50
Figure 3-34: Time Series Comparison of Simulated and Observed Temperature at St 03 in 2003 51
Figure 3-35: Time Series Comparison of Simulated and Observed Temperature at O9059000 in 2008 52
Figure 3-36: Box Plot Comparison of Simulated and Observed Temperature at St 03 in 2001 (Top and Bottom) 53
Figure 3-37: Box Plot Comparison of Simulated and Observed Temperature at St 03 in 2003 (Top and Bottom) 54
Figure 3-38: Box Plot Comparison of Simulated and Observed Temperature at O9059000in 2008 (Top and Bottom) 55
Figure 3-39: Profile Comparison of Simulated and Observed Temperature at O9059000 in 2001 56
Figure 3-40: Profile Comparison of Simulated and Observed Temperature at O9059000 in 2003 57
Figure 3-41: Profile Comparison of Simulated and Observed Temperature at O9059000 in 2008 58
Figure 3-42: Time Series Comparison of Simulated and Observed Salinity at Grimesland in 2001 59
Figure 3-43: Time Series Comparison of Simulated and Observed Salinity at Grimesland in 2003 60
Figure 3-44: Time Series Comparison of Simulated and Observed Salinity at Grimesland in 2008 61

6/114
Figure 3-45: Box Plot Comparison of Simulated and Observed Salinity at Grimesland in 2001 (Top) 62
Figure 3-46: Box Plot Comparison of Simulated a nd Observed Salinity at Grimesland in 2003 (Top) 62
Figure 3-47: Box Plot Comparison of Simulated and Observed Salinity at Grimesland in 2008 (Top) 63
Figure 3-48: Time Series Comparison of Simulated and Observed Salinity at St 10 in 2001 64
Figure 3-49: Time Series Comparison of Simulated and Observed Salinity at St 10 in 2003 65
Figure 3-50: Time Series Comparison of Simulated and Observed Salinity at O787000C in 2008 66
Figure 3-51: Box Plot Comparison of Simulated and Observed Salinity at St 10 in 2001 (Top and Bottom) 67
Figure 3-52: Box Plot Comparison of Simulated and Observed Salinity at St 10 in 2003 (Top and Bottom) 68
Figure 3-53: Box Plot Comparison of Simulated and Observed Salinity at O787000C in 2008 (Top and Bottom) 69
Figure 3-54: Profile Comparison of Simulated and Observed Salinity at O787000C in 2001 70
Figure 3-55: Profile Comparison of Simulated and Observed Salinity at O787000C in 2003 71
Figure 3-56: Profile Comparison of Simulated and Observed Salinity at O787000C in 2008 72
Figure 3-57: Time Series Comparison of Simulated and Observed Salinity at St 07 in 2001 73
Figure 3-58: Time Series Comparison of Simulated and Observed Salinity at St 07 in 2003 74
Figure 3-59: Time Series Comparison of Simulated and Observed Salinity at O8498000 in
2008 75
Figure 3-60: Box Plot Comparison of Simulated and Observed Salinity at St 07 in 2001 (Top
and Bottom) 76
Figure 3-61: Box Plot Comparison of Simulated and Observed Salinity at St 07 in 2003 (Top
and Bottom) 77
Figure 3-62: Box Plot Comparison of Simulated and Observed Salinity at O8498000 in 2008
(Top and Bottom) 78
Figure 3-63: Profile Comparison of Simulated and Observed Salinity at O8498000 in 2001 79
Figure 3-64: Profile Comparison of Simulated and Observed Salinity at O8498000 in 2003 80
Figure 3-65: Profile Comparison of Simulated and Observed Salinity at O8498000 in 2008 81
Figure 3-66: Time Series Comparison of Simulated and Observed Salinity at St 03 in 2001 82
Figure 3-67: Time Series Comparison of Simulated and Observed Salinity at St 03 in 2003 83
Figure 3-68: Time Series Comparison of Simulated and Observed Salinity at O9059000 in 2008 84
Figure 3-69: Box Plot Comparison of Simulated and Observed Salinity at St 03 in 2001 (Top and Bottom) 85
Figure 3-70: Box Plot Comparison of Simulated and Observed Salinity at St 03 in 2003 (Top and Bottom) 86

7/114
Figure 3-71: Box Plot Comparison of Simulated and Observed Salinity at O9059000 in 2008 (Top and Bottom) 87
Figure 3-72: Profile Comparison of Simulated and Observed Salinity at O9059000 in 2001 88
Figure 3-73: Profile Comparison of Simulated and Observed Salinity at O9059000 in 2003 89
Figure 3-74: Profile Comparison of Simulated and Observed Salinity at O9059000 in 2008 90
Figure 4-1: Hydrographs for Years 2007 and 2008 91
Figure 4-2: Daily Time Series for GUC Facilities - Existing Scenario 94
Figure 4-3: Daily Time Series for GUC Facilities - 2050 Scenario 94
Figure 4-4: Percent of Time 5 PSU is Exceeded by River Mile for Top, Middle, and Bottom Layers 97
Figure 4-5: Percent of Time 1 PSU is Exceeded by River Mile for Top, Middle, and Bottom Layers 98
Figure 4-6: Percent of Time 0.5 PSU is Exceeded by River Mile for Top, Middle, and Bottom Layers 99
Figure 4-7: Maximum Daily Salinity in the Bottom Layer at River Mile 44.5 (1/2 mile downstream of Grimesland) for Existing and 2050 Scenarios 101
Figure 4-8: Maximum Daily Salinity in the Bottom Layer at River Mile 49.0 (5 miles downstream of GUC WWTP discharge) for Existing and 2050 Scenarios 101
Figure 4-9: Maximum Daily Salinity in the Bottom Layer at River Mile 56.5 (USGS Gage at Greenville) for Existing and 2050 Scenarios 102
Figure 4-10: Maximum Daily Salinity in the Bottom Layer at River Mile 57.7 (0.3 miles downstream of GUC WTP withdrawal) for Existing and 2050 Scenarios 102
Figure 4-11: Maximum Daily Salinity in the Bottom Layer at River Mile 60.6 (0.6 miles upstream of GUC WTP withdrawal) for Existing and 2050 Scenarios 103

8/114
Tables
Table 3-1: Tar River Flow Study Segments 12
Table 3-2: Flow Estimation for Tributaries Between USGS Gage Tar River at Greenville and the Mouth of the Pamlico River Estuary 17
Table 3-3: Permitted Discharges Accounted for in the LTPR Model 18
Table 3-4: Flow Rankings for Four Modeling Years (Out of the 79-yr Extended Flow Record) 19
Table 3-5: Error Statistics for Simulated Stage at Washington, NC 22
Table 4-1: Average Annual Flows and Nutrient Loads Associated with the LTPR Modeling
Scenarios 92
Table 4-2: Monthly Average Ratio of GUC Demand to Raw Water Withdrawal and Return
Rates 93
Table 4-3: Hourly Temperature (°C) Percentiles in the Top Two Layers 105
Table 4-4: Hourly Temperature (°C) Percentiles in the Middle Two Layers 106
Table 4-5: Hourly Temperature (°C) Percentiles in the Bottom Two Layers 107
Table 4-6: Hourly Salinity (PSU) Percentiles in the Top Two Layers 109
Table 4-7: Hourly Salinity (PSU) Percentiles in the Middle Two Layers 110
Table 4-8: Hourly Salinity (PSU) Percentiles in the Bottom Two Layers 111

Appendices (on CD)
A Hydrodynamic and Water Quality Model Selection
B EFDC Model Description and Discussion of Capabilities
C Flow Rankings for Period of Record
D 2001 Time Series Calibration Plots
E 2001 Calibration Box Plots
F 2003 Time Series Validation Plots
G 2003 Validation Box Plots
H 2008 Time Series Supplemental Validation Plots
I 2008 Supplemental Validation Box Plots
J Profile Plots for Selection Stations (2001, 2003, 2008)

9/114
1. Introduction
The State of North Carolina typically limits water withdrawals to no more than twenty percent of the 7Q10
flow unless a flow study is conducted that determines additional withdrawals are acceptable. Greenville
Utilities Commission’s (GUC) water treatment plant (WTP) is permitted to withdraw 22.5 mgd from the Tar
River for public water supply, which equals 32 percent of the 7Q10 (109 cfs for summer months; 258 cfs
for winter months). The raw water intake is one of the few tidally influenced water supply intakes in North
Carolina, so the last WTP expansion was able to proceed without a flow study due to the tidal influence.
The Division of Water Resources (DWR) noted that any future plant expansion requests would have to be
accompanied by a flow study of the Tar River.
GUC conducted a planning study to evaluate the issues associated with future flow conditions and a
range of water withdrawals. The intent of this study was to develop an agency approved approach for
evaluating the Tar River’s instream flow needs. The study developed models and methods to assess
water quality, water quantity, and habitat conditions over a range of flow and metrological conditions.
Groundwater withdrawal reductions mandated by the Central Coastal Plain Capacity Use Area
(CCPCUA) Rules, recent droughts, encroachment of saltwater, and increasing regional water supply
needs have further emphasized the need for a better understanding of water availability in the Tar River.
The Environmental Fluid Dynamics Code (EFDC) hydrodynamic and water quality model developed for
the Pamlico River Estuary (the “PRE” Model; Xu et al., 2008) was selected for use in the Tar River Flow
Study. Of the available models, the PRE model provided the most robust and suitable model for the
analyses required for the Tar River Flow Study.
Currently, GUC is working cooperatively with NCDENR under the House Bill 609 (HB 609) process to
continue investigating the capacity of the Tar River to support additional future water withdrawal under
various flow conditions, particularly low flows. As part of this process, NCDWR wants to present the
LTPR model to the EMC to extend the Tar River Basin hydrologic model to cover the tidally influenced
sections of the basin. To support this objective, NCDWQ and NCDWR requested that GUC revise the
model documentation to focus on the hydrodynamics of the system: flows, stage, temperature, and
salinity.
GUC has investigated the capacity of the Tar River to support additional future water withdrawal under
various flow conditions, particularly low flows. The primary objectives of the Study include quantifying the
relationship between flow and aquatic habitat, identifying potential constraints on flow withdrawals, and
addressing potential changes in the location and movement of the freshwater-saltwater interface in the
lower Tar River due to withdrawals. To address the complexity of the hydrodynamic conditions in the tidal-
estuarine transition of the Tar River, the TAG chose to simulate the system with a hydrodynamic model.
The Environmental Fluid Dynamics Code (EFDC) hydrodynamic model developed for the Pamlico River
Estuary (the “PRE” Model; Xu et al. 2008) was selected for use in the Tar River Flow Study. Of the
available models, the PRE model provides the most robust and suitable model for the analyses required
for the Tar River Flow Study. This technical memorandum (TM) describes how the PRE Model was
modified and expanded to meet the needs of this study. The following sections describe the
hydrodynamic model setup, calibration, and performance.

10/114
2. Model Description
The EFDC model is a public domain, multi-functional surface water modeling system that includes
hydrodynamic, sediment-contaminant, and eutrophication components. The model can be used to
simulate aquatic systems in one, two, and three dimensions. It has evolved over the past two decades to
become one of the most widely used and technically defensible hydrodynamic models in the world (EPA,
2009).
The EFDC software was developed at the Virginia Institute of Marine Science for estuarine and coastal
applications, and is presently supported and made available by the EPA. A description of the model
selection process is included as Appendix A and a general presentation of the EFDC modeling software
and its capabilities is included as Appendix B) of this TM. The EFDC model has been applied to several
hydrodynamic and water quality numerical modeling studies in North Carolina by researchers at North
Carolina State University in the Department of Marine, Earth, and Atmospheric Sciences.
In one application of the PRE model, Xu et al. (2008) investigated circulation, salt intrusion, and vertical
stratification of salinity under different river flow and wind conditions in the Pamlico River Estuary (PRE).
The model domain for this application extended from Washington to the Pamlico Sound, and includes the
Pamlico River and its tributary, the Pungo River. The model was calibrated and verified against water
level, temperature, and salinity measured during 2001 and 2003. Eight sensitivity tests were conducted
with different river flow and wind conditions specified in the model (Xu et al., 2008). Model results show
that the salinity distribution response to various environmental forcing factors is dependent upon location
within the estuary. Additionally, in the PRE, salinity stratification tends to be most sensitive to changes in
river discharge, while the upstream distance of salinity intrusion is more sensitive to factors such as water
level set-up/set-down (increase/decrease in mean water level at the estuary mouth) and along river wind.

11/114
3. Model Setup
This section describes the various data sources and preprocessing steps required to extend the original
PRE model into the Tar River study areas and set up the hydrodynamic components of the full EFDC
based Lower Tar-Pamlico River Model (LTPR) for the Tar-Pamlico system. The hydrodynamic portion of
the LTPR model is capable of predicting river salinity and temperature at different river flows and water
use conditions.
The Tar River Flow Study includes four segments in the general vicinity of Greenville. These include the
Freshwater Non-tidal, Tidal Freshwater, Estuarine Transition, and Pamlico River Estuary segments
(Figure 3-1). These segments have dynamically changing boundaries. The boundaries are influenced by
river discharge, tidal cycles, and meteorological conditions. Near Greenville, the river is freshwater, but in
the Estuarine Transition segment salinity levels typically increase in a downstream direction to
Washington. At Washington, the Tar River becomes the Pamlico River Estuary. The LTPR model
simulates conditions within the Tidal Freshwater, Estuarine Transition, and Pamlico River segments of the
Study. Table 3-1 describes the different segments included in the LTPR model, their characteristics, and
the full set of analyses that are being conducted in each.
The remainder of this section describes the model grid extension, atmospheric inputs, boundary
conditions, and modeling years used to develop and calibrate the model. A summary of the model
calibration and validation results is also provided.

12/114
Figure 3-1: Tar River Flow Study Area

Table 3-1: Tar River Flow Study Segments
Tar River Segment Attributes, Approach, Key Issues
Tidal Freshwater
Segment • GUC Water Treatment Plant (WTP) to GUC Wastewater Treatment Plant (WWTP).
• Subject to total GUC withdrawals.
• Lowest net freshwater flows may occur in this segment due to withdrawals; non consumptively used water returned to river at GUC
WWTP.
• Hydraulic habitat conditions strongly influenced by tidal fluctuations,
especially at low flows.
Estuarine Transition
Segment
• GUC WWTP to Washington, NC.
• Tidally dominated; fluctuating salinity levels and salt wedge location.
• Subject to consumptive use only, not total GUC withdrawals.
Pamlico River Estuary
Segment • Estuarine circulation dominated; freshwater inflows can be important.
• Periodic summer bottom hypoxia/anoxia.

13/114
3.1 Model Grid Extension
The EFDC hydrodynamic model represents a water body as an assemblage of discrete volumetric cells.
The model’s horizontal grid defines the boundaries of these cells on a map projection. The EFDC model
uses a boundary fitted curvilinear-orthogonal horizontal grid to represent shoreline and interior features of
a water body. The LTPR model represents the vertical dimension with six discreet layers. The original
PRE model required a significant extension in the upstream direction to simulate the Tar River mainstem
(Tidal Freshwater and Estuarine Transition Segments).
Several datasets and processing steps were required to create the grid-based representation of the Tar-
Pamlico system from Washington to US 264 upstream of the GUC raw water intake at Greenville. As
described in the Study Plan, bathymetric and side scan surveys were conducted from a shallow-draft
vessel. The vessel used a Trimble GPS system operating with hydrographic survey software to develop
the modeling grid for the Tar River section of the model. In July 2009, the initial bathymetry data was
collected using at least three longitudinal survey lines (both depth data and side scan sonar images were
collected). Survey lines were spaced approximately 50 to 70 feet apart across the width of the channel.
The bathymetry data was collected with an echo-sounder approximately every two feet along each survey
line. The side scan sonar surveys were conducted at the same time as the bathymetric surveys to
characterize and delineate the substrate types. Objects such as snags, logs, wrecks, and other debris
exposed above the river bottom were also identified. In February 2010, cross sectional bathymetry was
collected at selected transects. In straight river sections, cross section data were collected approximately
once every 500 feet. Spacing between transects was shorter in areas with more diverse habitat,
especially in river bends. In the primary and secondary study areas, transects were spaced every 50 feet
for one-half mile within each area.
The bathymetric survey data for the Tar River reaches was interpolated to the original PRE grid and
enhanced using existing data sets for the transition area (e.g., LIDAR data obtained from the North
Carolina Floodplain Mapping Program). Data obtained from National Oceanic and Atmospheric
Administration (NOAA) was used to verify the original PRE model bathymetry, and some modifications
were made to more accurately reflect the main channel of the Estuary. Detailed digital elevation model
(DEM) data (1.5 m by 1.5 m) developed by USGS for their hydraulic modeling of areas around gage
stations (Bales et al. 2007) were used to refine the bathymetry in the vicinity of the gages.
Figure 3-2 shows the grid for the entire LTPR model. Figure 3-3 focuses on the area between the GUC
raw water intake and Washington. Climate stations, water quality monitoring stations, and the GUC
surface water withdrawal and wastewater discharge are labeled on each figure.

14/114
Figure 3-2: LTPR Modeling Grid from the Tar River upstream of Greenville to the Mouth of the Pamlico Estuary

15/114
Figure 3-3: LTPR Modeling Grid between Greenville and Washington, NC

16/138
3.2 Atmospheric Forcing Functions
Atmospheric forcing refers to forces that drive circulation in a particular waterbody, including tides, wind,
atmospheric pressure, and solar radiation. Data on atmospheric forcing are required to represent wind
driven circulation and to predict water temperature. Wind speed and solar radiation are also required by
the water quality modeling component to predict reaeration and primary production. Hourly wind data
observed at the PCS station on the southern bank of the PRE were used in the model. Hourly
meteorological data sets, including air pressure and temperature, relative humidity, rainfall, solar
radiation, and cloud cover were obtained from the Aurora station maintained by the N.C. State Climate
Office. Missing cloud cover data at Aurora were patched with observations from the KOCW station
operated by NOAA. Climate station locations with respect to the modeling grid are shown in Figure 3-2.
3.3 Boundary Conditions
Inflow boundary conditions are used by the EFDC model to simulate point and nonpoint source flows and
loadings to a system. This section describes how the boundary conditions were defined for the
headwaters, tributaries, and point-source discharges between Greenville and the mouth of the PRE.
3.3.1 Tar River at Greenville Inflows
The Tar River at US 264 upstream of Greenville is the upstream input for this system. The watershed
area draining to the most upstream grid cell of the Tar-Pamlico EFDC model is 2,660 square miles.
Simulated daily mean flows for the Tar River were based on data published by the USGS for the Tar
River at Greenville (Gage 02084000). Mean daily discharge has been recorded at this site since April
1997. Because this gage is located downstream of the GUC surface withdrawal, simulated withdrawals
were added to the flow series to represent headwater conditions at the US 264 location.
3.3.2 Tributary Inflows
Estimation of flows from tributaries in the watershed downstream of the upstream boundary were
dependent on whether or not a USGS gage was present on the tributary and what 8-digit HUC the
tributary was located in. There are three major tributaries downstream of the Tar River at Greenville
Gage that are located in Tar River Subbasin (HUC 03020103): Chicod, Grindle, and Tranters Creeks. Of
these, only Chicod is gaged by the USGS (Station 02084160 Chicod Creek near Simpson, NC). The
drainage area at the gage on Chicod Creek is 45 square miles, and flows have been recorded at this site
since October 1975. Flow data based on flow per unit watershed area from the Chicod USGS gage were
used to estimate flow contributions from Grindle and Tranters Creeks.
In the Pamlico River Basin (HUC 03020104) there are nine additional tributaries delivering flows and
loads to the estuary (including the Pungo River branch), none of which have an operating USGS gage. To
select the best index gage for these tributaries, statistical analyses of area-weighted average daily flows
by month were compared for the periods of record at six USGS gages located in coastal North Carolina.
Results for each gage were compared to the average and the gage representing the closest to average
condition for the coastal region was selected. Based on this analysis, USGS gage 02091000 Nahunta
Swamp near Shine was selected (drainage area of 80.4 square miles).

17/114
Daily mean flows for each tributary and the respective balance of watershed were then estimated based
on the selected index gage and the ratio of gaged drainage area to tributary drainage area. Table 3-2
lists each major tributary, index gage, and drainage area used to calculate tributary flows to the Tar-
Pamlico system. To account for flows from the balance of watershed (overland flow or small tributaries
not draining to a modeled tributary), additional area was assigned to the closest tributary. The balance of
watershed in the Tar Basin is 74.2 square miles and in the Pamlico Basin is 321.9 square miles.
Table 3-2: Flow Estimation for Tributaries Between USGS Gage Tar River at Greenville and the Mouth
of the Pamlico River Estuary
Tributary
Tributary Drainage Area (square miles)
Additional Drainage Area for Balance of Watershed (square miles)
Total Simulated Area (square miles)
Index Gage (Drainage Area square miles)
Tar River 2660 70 2730 Tar River at Greenville
Chicod 57.1 0 57.1 Chicod
Grindle 80.0 10 90 Chicod
Tranters 246.0 97 343 Chicod
Chocotowinity 38.5 45.4 83.9 Nahunta
Blounts 59.1 0 59.1 Nahunta
Durham 60.8 25.0 85.8 Nahunta
South 91.2 116.2 207.4 Nahunta
Goose 37.8 0 37.8 Nahunta
Bath 39.2 39.7 78.9 Nahunta
Pungo Creek 100.1 0 100.1 Nahunta
Pantego 173.4 0 173.4 Nahunta
Pungo River 480.2 0 480.2 Nahunta

3.3.3 Point Sources Discharges
Separate model inputs were created for permitted discharges that were not accounted for in the upstream
boundary or tributary inputs to the model. In general, the USEPA Point Source Compliance Database was
used to estimate monthly average flow from each permitted discharge. Discharge monitoring reports
obtained directly from PCS Phosphate Mine (PCS) were used to estimate loading from this source. Table
3-3 lists the permitted discharges accounted for explicitly in the LTPR model. Each discharge was
applied equally to the top two layers of the model grid.

18/114
Table 3-3: Permitted Discharges Accounted for in the LTPR Model
Facility Name
Permit
Number
Permitted
Flow (mgd)
Facility
Type
Receiving
Stream
Latitude,
Longitude
PCS Phosphate
Company Incorporated
NC0003255 NA Phosphatic
Fertilizers
Pamlico River 35.376981,
−76.748524
Greenville Utilities
Comm. WWTP
NC 0023931 17.5 Sewerage
Systems
Tar River 35.5989,
−77.3017
City of Washington
WWTP
NC0020648 3.65 Sewerage
Systems
Tar River 35.552222,
−77.072639
Greenville WTP NC0082139 1.2 Water Supply Tar River 35.634444,
−77.398611
Belhaven WWTP NC0026492 1 Sewerage
Systems
Battalina Creek 35.542365,
−76.609954

3.3.4 Groundwater Inputs
Groundwater inputs were specified separately for the river and estuarine segments. For the Tar River, it
was assumed that groundwater flows into the system are 3.99E-4 m/d based on data presented in
O’Driscoll et. al (2008). Spruill and Bratton (2008) estimate groundwater flows into the estuary as a
velocity of 0.09 m/d.
3.3.5 Surface Water Withdrawal
Greenville Utilities withdraws fresh water from the Tar River at 35.636034°, -77.401569°. Daily raw water
withdrawals and finished water pump rates were provided by GUC. Water withdrawals were assumed to
occur equally over the 2nd, 3rd, and 4th layers counting up from the bottom of the modeling grid.
3.3.6 Open Boundary Conditions
Open boundary conditions at the ocean boundary and truncated open water interior boundaries are
required for water surface elevation, salinity, and temperature. For modeling years 2001 and 2003, the
15-minute surface elevation data near the mouth of the PRE and biweekly data for salinity and
temperature obtained from East Carolina University (ECU) were specified as the open boundary
conditions for the EFDC hydrodynamic model. Water quality measurements obtained from ECU station
ST01 were interpolated temporally between measurements and vertically between layers.
For modeling years 2007 and 2008, the ECU data were not available, so data collected by DWQ at
station O982500C were used in conjunction with stage data observed by USGS at Washington as an
approximation of the open boundary condition. Methods described by Xu et al. (2008) were used to
extrapolate the observed stage at Washington to the open boundary condition at the mouth of the estuary
based on lag time and distance.

19/114
3.3.7 Modeling Years
The LTPR model is being used to simulate changes in water quality as well as salinity. The original PRE
model was calibrated to data observed in 2001 and 2003 to simulate the range of hydrologic conditions
observed in the watershed. The year 2001 was not the driest year on record, but it was the driest year
with the most available data. To compare the performance of the extended model to that of the original,
input files were created or extended for these years as needed. Year 2008 was also simulated with the
LTPR model for comparison to extensive salt-wedge monitoring data collected during an extreme drought
year.
Based on initial River 2-D and EFDC modeling results, in April 2011, the Habitat Subgroup decided that
habitat quality is more sensitive to changes in salinity than to changes in depth, velocity, and cover.
Depth, velocity, and cover are dominated by tidal conditions, especially during low flows. To evaluate
habitat changes under extreme low flow conditions, years 2007 and 2008 were used. As part of the
Environmental Assessment for the Interbasin Transfer Certificate (IBT) (ARCADIS, 2008), the Tar River at
Greenville flow record was extended back to 1932 based on historical data collected at the Tar River
USGS gage at Tarboro (ENTRIX, 2007).
Table 3-4 summarizes how the modeling years rank in terms of the extended flow dataset, with 1 being
the driest ranking and 79 the wettest. Rankings are based on median daily flow for each period. Out of
the 79-year extended flow record, 2007 was the fifth driest year with the driest fall and the fifth driest
summer. The year 2008 was the third driest year with the fourth driest summer and the 26th driest fall.
Low flow conditions in 2007 were exacerbated by operational errors of the upstream Tar River Reservoir
in Rocky Mount. Flow rankings for the extended period of record are provided in Appendix C.
Table 3-4: Flow Rankings for Four Modeling Years (Out of the 79-yr Extended Flow Record)
Year Annual Rank Winter Rank Spring Rank Summer Rank Fall Rank
2008 3 6 42 4 26
2007 5 52 32 5 1
2001 25 9 30 52 18
2003 79 76 78 77 69

The four modeling years cover a range of hydrologic conditions for the watershed. Figure 3-4 shows the
median daily flow recorded for each season, as well as annually, for the 13-year period of record for the
USGS gage on the Tar River at Greenville (years are sorted from lowest to highest annual median flow).
Over this 13-year period, 2008 and 2007 were the driest years (on an annual basis), and 2003 was the
wettest year; 2001 fell in the middle of the range for this dataset. Figure 3-5 shows the flow hydrographs
for these years, separated by vertical lines.

20/114
Figure 3-4: Annual and Seasonal Median Flows (Sorted by Annual Median) for the Tar River at
Greenville (1998 through 2010)
Figure 3-5: Flow Hydrographs for Four Modeling Years
0
1000
2000
3000
4000
5000
6000
7000
8000
9000
2008 2007 2009 2010 1998 2002 2001 2005 1999 2006 2000 2004 2003Median Daily Flow (cfs)Annual Winter Spring Summer Fall

21/114
3.4 Model Calibration and Validation
The LTPR model was calibrated to year 2001 data and validated to year 2003. These years were
selected due to their differences in hydrologic regime (2001 was relatively dry and 2003 was wet) and the
availability of data to develop boundary conditions and modeling inputs. For 2001 and 2003 inputs, water
column top and bottom measurements of temperature and salinity were available. Year 2008 was used
as a supplemental validation year to compare with profile data collected by GUC during salt wedge
tracking. Open boundary condition data for 2008 were not available at the same level as that in 2001 and
2003 (Section 3.3.6).
Calibration/validation for years 2001 and 2003 was assessed using data collected by ECU which were
collected at the surface and bottom layers at each site. Unfortunately, the time and exact sampling depth
associated with the ECU data collection efforts was not available, so correlating sampling data with a
specific hour and depth layer of model output was not possible. For the Tar River station at Grimesland
where no ECU data were available, calibration/validation was assessed using data collected by DWQ for
all three modeling years. Only surface grab samples were collected by DWQ at Grimesland, so bottom
comparisons were not generated. NCDWQ profile data for temperature and salinity were also used to
compare model simulations to observations. For year 2008, the supplementation validation data for
temperature and salinity were obtained from NCDWQ, and the ECU data were no longer being collected.
Time series plots for the calibration and validation compare hourly model output to grab sample data. For
the DWQ data (Grimesland station in 2001 and 2003; all stations in 2008), recorded sampling time was
used to generate the observed data plots. Time stamps were not provided for the ECU data, so these
data were plotted as being collected at noon each day (2001 and 2003).
This section of the TM describes the stage, temperature, and salinity calibration results. Five stations
(one stage and four water quality) have been selected as the focus of this main report. Simulated hourly
values are compared to point in time grab samples using both time series plots and box plots. The latter
format is more appropriate when comparing hourly simulations to observations that are collected once or
twice per month. Profile plots of temperature and salinity are also provided to show how the model
stratification compares to observations. Appendices D through G present the results for each monitoring
station in the modeled system.
3.4.1 Stage
USGS stage data at Washington, NC (Gage 02084472) were used to compare simulated water levels to
those observed. The datum for the gage is minus feet relative to the North American Vertical Datum of
1988 (NAVD88). Figure 3-6, Figure 3-7, and Figure 3-8 present the comparisons of water level for years
2001, 2003, and 2008, respectively. The top half of each figure shows the simulated values for one year.
The bottom half of each figure zooms in over a 20 day period. Error statistics are presented in Table 3-5.
During each modeling year, the error and bias statistics are relatively low (near zero) and the correlation
coefficients are relatively high (near one).

22/114
Table 3-5: Error Statistics for Simulated Stage at Washington, NC
Year
Root-Mean-Squared Error
Model Bias (Mean Error)
Standard Deviation Error
Correlation Coefficient (R2)
Relative Average Skill (percent)
Relative Average Error (percent)
Cond-itional Bias Uncond-itional Bias Skill Score
2001 0.042 0.002 0.022 0.958 98.48 1.52 0.033 0.000 0.883
2003 0.037 0.010 -0.006 0.972 99.21 0.79 0.000 0.002 0.942
2008 0.033 -0.001 0.005 0.971 99.238 0.762 0.003 0.000 0.939

23/114
Figure 3-6: Simulated and Observed Stage at Washington, NC (2001)

24/114
Figure 3-7: Simulated and Observed Stage at Washington, NC (2003)

25/114
Figure 3-8: Simulated and Observed Stage at Washington, NC (2008)

26/114
3.4.2 Temperature
Temperature data in the Lower Tar Pamlico system used for model calibration and validation were
obtained from ECU and NCDWQ. Three types of graphs are used to compare simulated and observed
values. Time series plots compare hourly model output in either the surface or bottom layer to point in
time measurements represented by pink diamonds. Box plots show the distribution of simulated values
over a month compared to the observations (pink diamonds) collected during the month. Profile plots
show the simulated value for a specific model layer at a specific hour and day compared to
measurements collected over the depth of the water column.
The majority of the data used for calibration and validation comes from ECU in 2001 and 2003 and from
NCDWQ in 2008. NCDWQ data were used in all three years for Grimesland station, since ECU only
collected data in the estuary. NCDWQ data are also used for the profile plots since ECU only collected
field parameters at the surface and bottom and not as profiles over the depth of the water column.
This section of the report provides comparisons of simulated and observed data for four regions in the
system (Grimesland, Upper Estuary, Middle Estuary, and Lower Estuary). Comparisons for additional
locations are provided in Appendices D through J:
• Appendix D 2001 Time Series Calibration Plots
• Appendix E 2001 Calibration Box Plots
• Appendix F 2003 Time Series Validation Plots
• Appendix G 2003 Validation Box Plots
• Appendix H 2008 Time Series Supplemental Validation Plots
• Appendix I 2008 Supplemental Validation Box Plots
• Appendix J Profile Plots for Selection Stations (2001, 2003, 2008)
Note that temperature is simulated over the entire 24-hour period, but observations are only collected
during daylight hours. Thus simulated values capture more diurnal variability than the point in
measurements.
3.4.2.1 Tar River at Grimesland (NC DWQ Station O6500000)
The surface of the Tar River at Grimesland is monitored by NCDWQ. Simulated hourly values in the
surface layer follow the same seasonal patterns as the point in time measurements (Figure 3-9 through
Figure 3-14). There are some periods when simulated values are a few degrees cooler than those
observed. These cooler temperatures may be due to simulated values representing the average value
over a model layer being compared to a measurement taken at a specific depth near the surface. Input
files based on monthly measurements for the headwater and a limited number of tributary inputs may also
lead to a discrepancy in simulated and observed values.

27/114
The agency does not collect monitoring data in the middle or bottom layers at this location, so
comparisons of simulated and observed values in the bottom layers or along profiles are not available at
Grimesland.

Figure 3-9: Time Series Comparison of Simulated and Observed Temperature at Grimesland in
2001

28/114
Figure 3-10: Time Series Comparison of Simulated and Observed Temperature at Grimesland in
2003

29/114

Figure 3-11: Time Series Comparison of Simulated and Observed Temperature at Grimesland in
2008

30/114
Figure 3-12: Box Plot Comparison of Simulated and Observed Temperature at Grimesland in 2001

Figure 3-13: Box Plot Comparison of Simulated and Observed Temperature at Grimesland in 2003

31/114

Figure 3-14: Box Plot Comparison of Simulated and Observed Temperature at Grimesland in 2008

3.4.2.2 Upper Estuary (ECU Station St 10 and NCDWQ O787000C)
Figure 3-15 through Figure 3-23 compare simulated and observed temperature at two stations located in
the upper estuary. For year 2001 and 2003, the Upper Estuary is primarily represented by Station 10
which is monitored by ECU. Observations are available for the surface and the bottom at this location.
Depth profile data in the Upper Estuary were collected by NCDWQ at station O787000C. In 2001 and
2003, simulated seasonal trends follow those observed, but there are some periods when the simulated
values in the surface layer are generally cooler than those observed. In 2008, the simulated surface
values match those observed more closely than in 2001 and 2003. Simulated bottom temperatures are
more consistent with observations compared to the surface layer. This may be due to the homogeneity of
bottom temperatures relative to temperatures in the surface layer which may be more variable within the
layer.

32/114
Figure 3-15: Time Series Comparison of Simulated and Observed Temperature at St 10 in 2001

33/114

Figure 3-16: Time Series Comparison of Simulated and Observed Temperature at St 10 in 2003

34/114
Figure 3-17: Time Series Comparison of Simulated and Observed Temperature at O787000C in
2008

35/114

Figure 3-18: Box Plot Comparison of Simulated and Observed Temperature at St 10 in 2001 (Top
and Bottom)

36/114

Figure 3-19: Box Plot Comparison of Simulated and Observed Temperature at St 10 in 2003 (Top
and Bottom)

37/114

Figure 3-20: Box Plot Comparison of Simulated and Observed Temperature at O787000C in 2008
(Top and Bottom)

38/114

Figure 3-21: Profile Comparison of Simulated and Observed Temperature at O787000C in 2001

0.0
1.0
2.0
3.0
4.0
0 10 20 30
Depth (m)Temperature (oC)
02/06/2001 10:42
0.0
1.0
2.0
3.0
4.0
0 10 20 30
Depth (m)Temperature (oC)
05/01/2001 11:27 Data
Model
0.0
1.0
2.0
3.0
4.0
0 10 20 30
Depth (m)Temperature (oC)
08/07/2001 11:10
0.0
1.0
2.0
3.0
4.0
0 10 20 30
Depth (m)Temperature (oC)
11/07/2001 11:55

39/114

Figure 3-22: Profile Comparison of Simulated and Observed Temperature at O787000C in 2003

0.0
1.0
2.0
3.0
4.0
0 10 20 30
Depth (m)Temperature (oC)
02/11/2003 10:20
0.0
1.0
2.0
3.0
4.0
0 10 20 30
Depth (m)Temperature (oC)
05/14/2003 09:45 Data
Model
0.0
1.0
2.0
3.0
4.0
0 10 20 30
Depth (m)Temperature (oC)
08/04/2003 12:21
0.0
1.0
2.0
3.0
4.0
0 10 20 30
Depth (m)Temperature (oC)
09/16/2003 10:15

40/114

Figure 3-23: Profile Comparison of Simulated and Observed Temperature at O787000C in 2008

0.0
1.0
2.0
3.0
4.0
0 10 20 30
Depth (m)Temperature (oC)
02/05/2008 10:30
0.0
1.0
2.0
3.0
4.0
0 10 20 30
Depth (m)Temperature (oC)
05/07/2008 10:02 Data
Model
0.0
1.0
2.0
3.0
4.0
0 10 20 30
Depth (m)Temperature (oC)
08/05/2008 10:20
0.0
1.0
2.0
3.0
4.0
0 10 20 30
Depth (m)Temperature (oC)
11/12/2008 10:20

41/114
3.4.2.3 Middle Estuary (ECU Station St 07 and NCDWQ O8498000)
Figure 3-24 through Figure 3-32 provide comparisons of simulated and observed temperatures at two
monitoring stations in the middle estuary. Station 07 is monitored by ECU and provides comparisons for
years 2001 and 2003. Station O8498000 is monitored by NCDWQ and provides comparisons for year
2008 and profile data for all three years. At these stations, simulated hourly output in the surface layer
tracks with observed seasonal trends, but the simulated values are cooler during certain periods
compared to those observed. This is likely due to the diurnal output associated with the simulation, the
variability of temperature within the surface layer, and the frequency of data collected and used to build
input files. Simulated and observed temperatures in the bottom layer match fairly well. Simulated
temperature profiles are lower in 2001 and 2003 compared to observed, but match well in 2008.

Figure 3-24: Time Series Comparison of Simulated and Observed Temperature at St 07 in 2001

42/114

Figure 3-25: Time Series Comparison of Simulated and Observed Temperature at St 07 in 2003

43/114

Figure 3-26: Time Series Comparison of Simulated and Observed Temperature at O8498000 in
2008

44/114

Figure 3-27: Box Plot Comparison of Simulated and Observed Temperature at St 07 in 2001 (Top
and Bottom)

45/114

Figure 3-28: Box Plot Comparison of Simulated and Observed Temperature at St 07 in 2003 (Top
and Bottom)

46/114

Figure 3-29: Box Plot Comparison of Simulated and Observed Temperature at O8498000 in 2008
(Top and Bottom)

47/114

Figure 3-30: Profile Comparison of Simulated and Observed Temperature at O8498000 in 2001

0.0
1.0
2.0
3.0
4.0
0 10 20 30
Depth (m)Temperature (oC)
02/06/2001 09:45
0.0
1.0
2.0
3.0
4.0
0 10 20 30
Depth (m)Temperature (oC)
05/101/2001 10:20 Data
Model
-0.5
0.5
1.5
2.5
3.5
4.5
0 10 20 30
Depth (m)Temperature (oC)
08/07/2001 10:00
0.0
1.0
2.0
3.0
4.0
0 10 20 30
Depth (m)Temperature (oC)
11/07/2001 11:39

48/114

Figure 3-31: Profile Comparison of Simulated and Observed Temperature at O8498000 in 2003

0.0
1.0
2.0
3.0
4.0
0 10 20 30
Depth (m)Temperature (oC)
02/11/2003 13:00
0.0
1.0
2.0
3.0
4.0
0 10 20 30
Depth (m)Temperature (oC)
05/14/2003 10:00 Data
Model
-0.5
0.5
1.5
2.5
3.5
4.5
0 10 20 30
Depth (m)Temperature (oC)
08/04/2003 09:48
0.0
1.0
2.0
3.0
4.0
0 10 20 30
Depth (m)Temperature (oC)
09/16/2003 13:19

49/114

Figure 3-32: Profile Comparison of Simulated and Observed Temperature at O8498000 in 2008

0.0
1.0
2.0
3.0
4.0
0 10 20 30
Depth (m)Temperature (oC)
02/05/2008 13:28
0.0
1.0
2.0
3.0
4.0
0 10 20 30
Depth (m)Temperature (oC)
05/07/2008 12:40 Data
Model
-0.5
0.5
1.5
2.5
3.5
4.5
0 10 20 30
Depth (m)Temperature (oC)
08/05/2008 12:46
0.0
1.0
2.0
3.0
4.0
0 10 20 30
Depth (m)Temperature (oC)
11/12/2008 13:10

50/114
3.4.2.4 Lower Estuary (ECU Station St 03 and NCDWQ O9059000)
Figure 3-33 through Figure 3-41 present simulated and observed temperature data at two locations in the
lower estuary. Station 03 is monitored by ECU and data are available for the surface and bottom in years
2001 and 2003. Depth profile data in the Lower Estuary were collected by NCDWQ at station O9059000,
which is also used to compare year 2008 simulated values to those observed. Patterns in the lower
estuary are similar to those in the upper reaches with bottom temperature simulations match observations
fairly well and surface simulations sometimes cooler than observed.

Figure 3-33: Time Series Comparison of Simulated and Observed Temperature at St 03 in 2001

51/114
Figure 3-34: Time Series Comparison of Simulated and Observed Temperature at St 03 in 2003

52/114

Figure 3-35: Time Series Comparison of Simulated and Observed Temperature at O9059000 in
2008

53/114

Figure 3-36: Box Plot Comparison of Simulated and Observed Temperature at St 03 in 2001 (Top
and Bottom)

54/114

Figure 3-37: Box Plot Comparison of Simulated and Observed Temperature at St 03 in 2003 (Top
and Bottom)

55/114

Figure 3-38: Box Plot Comparison of Simulated and Observed Temperature at O9059000in 2008
(Top and Bottom)

56/114
Figure 3-39: Profile Comparison of Simulated and Observed Temperature at O9059000 in 2001

0.0
1.0
2.0
3.0
4.0
5.0
0 10 20 30
Depth (m)Temperature (oC)
02/06/2001 10:55
0.0
1.0
2.0
3.0
4.0
5.0
0 10 20 30
Depth (m)Temperature (oC)
05/01/2001 11:40 Data
Model
0.0
1.0
2.0
3.0
4.0
5.0
0 10 20 30
Depth (m)Temperature (oC)
08/07/2001 11:35
0.0
1.0
2.0
3.0
4.0
5.0
0 10 20 30
Depth (m)Temperature (oC)
11/07/2001 12:49

57/114

Figure 3-40: Profile Comparison of Simulated and Observed Temperature at O9059000 in 2003

0.0
1.0
2.0
3.0
4.0
5.0
0 10 20 30
Depth (m)Temperature (oC)
02/11/2003 11:25
0.0
1.0
2.0
3.0
4.0
5.0
0 10 20 30
Depth (m)Temperature (oC)
05/14/2003 11:10 Data
Model
0.0
1.0
2.0
3.0
4.0
5.0
0 10 20 30
Depth (m)Temperature (oC)
08/04/2003 10:55
0.0
1.0
2.0
3.0
4.0
5.0
0 10 20 30
Depth (m)Temperature (oC)
09/16/2003 11:34

58/114

Figure 3-41: Profile Comparison of Simulated and Observed Temperature at O9059000 in 2008

0.0
1.0
2.0
3.0
4.0
5.0
0 10 20 30
Depth (m)Temperature (oC)
02/05/2008 11:56
0.0
1.0
2.0
3.0
4.0
5.0
0 10 20 30
Depth (m)Temperature (oC)
05/07/2008 10:28 Data
Model
0.0
1.0
2.0
3.0
4.0
5.0
0 10 20 30
Depth (m)Temperature (oC)
08/05/2008 11:28
0.0
1.0
2.0
3.0
4.0
5.0
0 10 20 30
Depth (m)Temperature (oC)
11/12/2008 11:47

59/114
3.4.3 Salinity
Salinity data in the Lower Tar Pamlico system are primarily from ECU in 2001 and 2003 and from
NCDWQ in 2008 (with the exception of the station at Grimesland which is used for comparison in all three
years). This section of the report provides comparisons of simulated and observed data for four locations
in the system (Grimesland, Upper Estuary, Middle Estuary, and Lower Estuary). Comparisons for
additional locations are provided in Appendices D through I.
3.4.3.1 Tar River at Grimesland (NC DWQ Station O6500000)
The surface of the Tar River at Grimesland is monitored by NCDWQ. Two grab samples were collected
in the winter of 2001 with more frequent sampling in 2003 and 2008 (Figure 3-42 through Figure 3-47).
The agency does not collect monitoring data in the bottom layers at this location. The model generally
overpredicts surface salinities at this location in 2008 which results in a conservative prediction of habitat
impacts during increase withdrawal scenarios (Section 4).
Figure 3-42: Time Series Comparison of Simulated and Observed Salinity at Grimesland in 2001

60/114
Figure 3-43: Time Series Comparison of Simulated and Observed Salinity at Grimesland in 2003

61/114
Figure 3-44: Time Series Comparison of Simulated and Observed Salinity at Grimesland in 2008

62/114
Figure 3-45: Box Plot Comparison of Simulated and Observed Salinity at Grimesland in 2001
(Top)
Figure 3-46: Box Plot Comparison of Simulated and Observed Salinity at Grimesland in 2003
(Top)
Top
Top

63/114
Figure 3-47: Box Plot Comparison of Simulated and Observed Salinity at Grimesland in 2008 (Top)

3.4.3.2 Upper Estuary (ECU Station St 10 and NCDWQ O787000C)
Figure 3-48 through 3-56 compare salinities in the upper estuary at two monitoring stations. For year
2001 and 2003, the upper estuary is primarily represented by Station 10 which is monitored by ECU.
Observations are available for the surface and the bottom at this location. Depth profile data in the Upper
Estuary were collected by NCDWQ at station O787000C; this station is also used for year 2008
comparisons. In 2001, the model tends to over predict salinity during some times of the year in both the
surface and bottom layers; the fit in 2003 and 2008 is better. Simulated profiles tend to match those
observed.

64/114

Figure 3-48: Time Series Comparison of Simulated and Observed Salinity at St 10 in 2001

65/114

Figure 3-49: Time Series Comparison of Simulated and Observed Salinity at St 10 in 2003

66/114

Figure 3-50: Time Series Comparison of Simulated and Observed Salinity at O787000C in 2008

67/114

Figure 3-51: Box Plot Comparison of Simulated and Observed Salinity at St 10 in 2001 (Top and
Bottom)

68/114

Figure 3-52: Box Plot Comparison of Simulated and Observed Salinity at St 10 in 2003 (Top and
Bottom)

69/114

Figure 3-53: Box Plot Comparison of Simulated and Observed Salinity at O787000C in 2008 (Top
and Bottom)

70/114

Figure 3-54: Profile Comparison of Simulated and Observed Salinity at O787000C in 2001

0.0
1.0
2.0
3.0
4.0
0 10 20
Depth (m)Salinity (psu)
02/06/2001 10:42
0.0
1.0
2.0
3.0
4.0
0 10 20
Depth (m)Salinity (psu)
05/01/2001 11:27 Data
Model
0.0
1.0
2.0
3.0
4.0
0 10 20
Depth (m)Salinity (psu)
08/07/2001 11:10
0.0
1.0
2.0
3.0
4.0
0 10 20
Depth (m)Salinity (psu)
11/07/2001 11:55

71/114

Figure 3-55: Profile Comparison of Simulated and Observed Salinity at O787000C in 2003

0.0
1.0
2.0
3.0
4.0
0 10 20
Depth (m)Salinity (psu)
02/11/2003 10:20
0.0
1.0
2.0
3.0
4.0
0 10 20
Depth (m)Salinity (psu)
05/014/2003 09:45 Data
Model
0.0
1.0
2.0
3.0
4.0
0 10 20
Depth (m)Salinity (psu)
08/04/2003 12:21
0.0
1.0
2.0
3.0
4.0
0 10 20
Depth (m)Salinity (psu)
09/16/2003 10:15

72/114

Figure 3-56: Profile Comparison of Simulated and Observed Salinity at O787000C in 2008

0.0
1.0
2.0
3.0
4.0
0 10 20
Depth (m)Salinity (psu)
02/05/2008 10:30
0.0
1.0
2.0
3.0
4.0
0 10 20
Depth (m)Salinity (psu)
05/07/2008 10:02 Data
Model
0.0
1.0
2.0
3.0
4.0
0 10 20
Depth (m)Salinity (psu)
08/05/2008 10:20
0.0
1.0
2.0
3.0
4.0
0 10 20
Depth (m)Salinity (psu)
11/12/2008 10:20

73/114
3.4.3.3 Middle Estuary (ECU Station St 07 and NCDWQ O8498000)
Figure 3-57 through 3-65 compare simulated and observed salinities in the middle estuary at two stations.
Station 07 is monitored by ECU, and data are available for the surface and bottom in 2001 and 2003.
Year 2008 observations and depth profile data were collected by NCDWQ at station O8498000. Model fit
in this location is relative good; there are some periods when the model over predicts salinity.

Figure 3-57: Time Series Comparison of Simulated and Observed Salinity at St 07 in 2001

74/114
Figure 3-58: Time Series Comparison of Simulated and Observed Salinity at St 07 in 2003

75/114
Figure 3-59: Time Series Comparison of Simulated and Observed Salinity at O8498000 in 2008

76/114

Figure 3-60: Box Plot Comparison of Simulated and Observed Salinity at St 07 in 2001 (Top and
Bottom)

77/114

Figure 3-61: Box Plot Comparison of Simulated and Observed Salinity at St 07 in 2003 (Top and
Bottom)

78/114

Figure 3-62: Box Plot Comparison of Simulated and Observed Salinity at O8498000 in 2008 (Top
and Bottom)

79/114

Figure 3-63: Profile Comparison of Simulated and Observed Salinity at O8498000 in 2001

0.0
1.0
2.0
3.0
4.0
0 10 20
Depth (m)Salinity (psu)
02/06/2001 09:45
0.0
1.0
2.0
3.0
4.0
0 10 20
Depth (m)Salinity (psu)
05/01/2001 10:20 Data
Model
0.0
1.0
2.0
3.0
4.0
0 10 20
Depth (m)Salinity (psu)
08/07/2001 10:00
0.0
1.0
2.0
3.0
4.0
0 10 20
Depth (m)Salinity (psu)
11/07/2001 11:39

80/114

Figure 3-64: Profile Comparison of Simulated and Observed Salinity at O8498000 in 2003

0.0
1.0
2.0
3.0
4.0
0 10 20
Depth (m)Salinity (psu)
02/11/2003 13:00
0.0
1.0
2.0
3.0
4.0
0 10 20
Depth (m)Salinity (psu)
05/14/2003 10:00 Data
Model
0.0
1.0
2.0
3.0
4.0
0 10 20
Depth (m)Salinity (psu)
08/04/2003 09:48
0.0
1.0
2.0
3.0
4.0
0 10 20
Depth (m)Salinity (psu)
09/16/2003 13:19

81/114

Figure 3-65: Profile Comparison of Simulated and Observed Salinity at O8498000 in 2008

0.0
1.0
2.0
3.0
4.0
0 10 20
Depth (m)Salinity (psu)
02/05/2008 13:28
0.0
1.0
2.0
3.0
4.0
0 10 20
Depth (m)Salinity (psu)
05/07/2008 12:40 Data
Model
0.0
1.0
2.0
3.0
4.0
0 10 20
Depth (m)Salinity (psu)
08/05/2008 12:46
0.0
1.0
2.0
3.0
4.0
0 10 20
Depth (m)Salinity (psu)
11/12/2008 13:10

82/114
3.4.3.4 Lower Estuary (ECU Station St 03 and NCDWQ O9059000)
Figure 3-66 through Figure 3-74 compare salinities at two monitoring stations in the lower estuary.
Station 03 is monitored by ECU and data are available for the surface and bottom in 2001 and 2003.
Data for year 2008 as well as depth profile data are represented by NCDWQ station O9059000.
Salinities in the surface layer generally matches those observed; in the bottom layer the model tends to
over predict salinity during certain periods. Profiles match fairly well at this location.

Figure 3-66: Time Series Comparison of Simulated and Observed Salinity at St 03 in 2001

83/114

Figure 3-67: Time Series Comparison of Simulated and Observed Salinity at St 03 in 2003

84/114

Figure 3-68: Time Series Comparison of Simulated and Observed Salinity at O9059000 in 2008

85/114

Figure 3-69: Box Plot Comparison of Simulated and Observed Salinity at St 03 in 2001 (Top and
Bottom)

86/114

Figure 3-70: Box Plot Comparison of Simulated and Observed Salinity at St 03 in 2003 (Top and
Bottom)

87/114

Figure 3-71: Box Plot Comparison of Simulated and Observed Salinity at O9059000 in 2008 (Top
and Bottom)

88/114

Figure 3-72: Profile Comparison of Simulated and Observed Salinity at O9059000 in 2001

0.0
1.0
2.0
3.0
4.0
5.0
0 10 20
Depth (m)Salinity (psu)
02/06/2001 10:55
0.0
1.0
2.0
3.0
4.0
5.0
0 10 20
Depth (m)Salinity (psu)
05/01/2001 11:40 Data
Model
0.0
1.0
2.0
3.0
4.0
5.0
0 10 20
Depth (m)Salinity (psu)
08/07/2001 11:35
0.0
1.0
2.0
3.0
4.0
5.0
0 10 20
Depth (m)Salinity (psu)
11/07/2001 12:49

89/114

Figure 3-73: Profile Comparison of Simulated and Observed Salinity at O9059000 in 2003

0.0
1.0
2.0
3.0
4.0
5.0
0 10 20
Depth (m)Salinity (psu)
02/11/2003 11:25
0.0
1.0
2.0
3.0
4.0
5.0
0 10 20
Depth (m)Salinity (psu)
05/14/2003 11:10 Data
Model
0.0
1.0
2.0
3.0
4.0
5.0
0 10 20
Depth (m)Salinity (psu)
08/04/2003 10:55
0.0
1.0
2.0
3.0
4.0
5.0
0 10 20
Depth (m)Salinity (psu)
09/16/2003 11:34

90/114

Figure 3-74: Profile Comparison of Simulated and Observed Salinity at O9059000 in 2008

0.0
1.0
2.0
3.0
4.0
5.0
0 10 20
Depth (m)Salinity (psu)
02/05/2008 11:56
0.0
1.0
2.0
3.0
4.0
5.0
0 10 20
Depth (m)Salinity (psu)
05/07/2008 10:28 Data
Model
0.0
1.0
2.0
3.0
4.0
5.0
0 10 20
Depth (m)Salinity (psu)
08/05/2008 11:28
0.0
1.0
2.0
3.0
4.0
5.0
0 10 20
Depth (m)Salinity (psu)
11/12/2008 11:47

91/114
4. Modeling Scenarios
Following setup and calibration, the LTPR model was used to evaluate the impacts of future water
withdrawal scenarios on salinity in the Tar-Pamlico system and compare them to existing conditions.
Model setup files for years 2007 and 2008 were selected because flows during these years were some of
the driest on record (the 5th and 3rd driest years out of the 79-year extended flow record, respectively,
Figure 4-1) and previous hydrologic and screening analyses indicated that changes in GUC operations
could have the greatest impacts when river flows are low.
Figure 4-1: Hydrographs for Years 2007 and 2008
Three scenarios were developed for this assessment. The existing conditions scenario used facility
information reported for year 2010 to develop operational time series and provide a baseline condition. A
future year was developed to simulate changes due to increases in facility operations by GUC. Year
2050 was selected for this analysis because the 2050 withdrawal rate (47.1 mgd) was greater than
preliminary projections for facility expansions (30 mgd to 35 mgd) and would therefore cover uncertainties
with respect to changes in upstream river flows in the future, etc. Lastly, a scenario where GUC neither
withdraws from nor discharges to the Tar River (No GUC) was developed to assess conditions that would
likely be present if the GUC facilities were not active. This scenario was not run for the salinity-habitat
assessments that were performed with year 2008 model runs. Simulated temperature and salinity data for
the three scenarios was generated as hourly output at each of the monitoring stations.

92/114
4.1 Development
For the three modeling scenarios, only the modeling inputs for the GUC facilities were altered: all
boundary conditions, meteorological files, and other facility discharges were kept at year 2007/2008
conditions. This allows comparison of changes associated with increasing GUC withdrawals during low
flow conditions. Table 4-1 summarizes the average annual flows simulated under the two scenarios.
Table 4-1: Average Annual Flows and Nutrient Loads Associated with the LTPR Modeling Scenarios
Scenario Existing 2050 No GUC
Withdrawal (mgd) 13.3 47.1 0
WTP Return( mgd) 0.5 1.9 0
WWTP Discharge (mgd) 10.2 30.4 0
Consumptive Use (mgd) 2.6 14.8 0

To simulate the impact of increasing GUC withdrawals on the system, daily time series for each scenario
were generated for the WTP withdrawal and return and the WWTP discharge. Daily WTP withdrawal
rates were generated by Hazen and Sawyer (2011). A managed water system demand strategy was
developed in support of the IBT Certificate for Greenville Utilities, Greene County, and the Towns of
Farmville and Winterville. A daily water use pattern was created for each community over multiple
consecutive years. The patterns were applied to the annual average water demand projections to
produce a daily water demand projection curve over the planning period. The resulting daily water
demand curves were added consecutively to produce a managed system demand that reflects the total
water use for all four communities. This same methodology was used for the water withdrawal projections
for the Tar River Flow Study. The only change made to the data supporting the IBT effort was a revision
to the water demand projections for Greenville Utilities in 2011. The managed system demand was
converted into a managed water withdrawal by considered process water use at the WTP. The Tar River
Flow Study Water Withdrawal Analysis Technical Memorandum (TM #5) provides the annual and monthly
water demand projections, consumptive use, managed system demand, and managed system withdrawal
from 2011 through 2090.
For this analysis, WTP returns were assumed to be 4 percent of withdrawals based on monthly average
data for year 2010. WWTP discharges for existing conditions were calculated by multiplying the raw
water intake by 1) the monthly average return flow factor based on historic data and 2) a correction factor
to ensure that monthly average WWTP discharges match those observed in 2010 (Table 4-2). WWTP
discharges for 2050 were calculated by multiplying the raw water intake by 1) the ratio of water used by
Greenville to total water use (equals 1 minus fraction of water transferred outside of the basin) and 2) the
monthly average return flow ratio based on historic data (Table 4-2). Resulting daily time series for the
GUC operations are shown in Figures 4-2 and 4-3 for the existing and 2050 scenario, respectively.

93/114
Table 4-2: Monthly Average Ratio of GUC Demand to Raw Water Withdrawal and Return Rates
Existing Scenario (2010) 2050
Month
Ratio of Return (WWTP Discharge/WTP
Withdrawal)
Correction Factor To Match Actual Monthly Average
Discharge Rates
Ratio of GUC Demand to Total
Demand
Ratio of Return (WWTP Discharge/WTP
Withdrawal)
January 0.899 1.097 0.78 0.899
February 0.968 1.283 0.79 0.968
March 0.972 1.038 0.78 0.972
April 0.866 0.862 0.79 0.866
May 0.773 0.808 0.79 0.773
June 0.733 0.777 0.78 0.733
July 0.731 0.778 0.78 0.731
August 0.759 0.789 0.77 0.759
September 0.850 0.765 0.76 0.850
October 0.877 1.029 0.75 0.877
November 0.823 0.885 0.76 0.823
December 0.912 0.825 0.75 0.912

94/114
Figure 4-2: Daily Time Series for GUC Facilities - Existing Scenario
Figure 4-3: Daily Time Series for GUC Facilities - 2050 Scenario

95/114
4.2 Results
Model output for the three scenarios was generated for two main purposes: 1) compare daily average
salinity at every cell and layer in the model grid to provide a basis from which to assess habitat changes
due to impacts on salinity and 2) compare hourly output at monitoring locations to assess changes in
temperature and salinity at a more refined temporal scale. The requirements of these assessments
resulted in distinct output cells, time steps, and data processing steps. This section describes the
methods for developing these assessments as well as the modeling results.
4.2.1 Salinity Habitat Impacts
The objectives of the salinity modeling were to provide an assessment of changes in salinity due to
altering GUC facility operations in terms of 1) salinity encroachment, and 2) absolute changes at the grid
cell level for identified habitat areas (fish spawning areas, submerged aquatic vegetation beds, etc.). In
order to assess these changes throughout the system, the LTPR model was used to output daily average
salinity values at every grid cell in the model for each of the six depth layers. The Tar River Flow Study
Habitat Modeling Technical Memorandum (TM #7) describes in detail how the salinity modeling was used
to assess impacts on habitats.
To provide an assessment of salinity encroachment due to increased river withdrawals, daily average
salinity values for the existing, 2050, and No GUC scenarios were output for years 2007 and 2008, which
represent very dry conditions and some of the lowest flows on record. This phenomenon is particularly
evident in 2007 when the operation of the Tar River Reservoir was managed using un-calibrated USGS
gage data. Percent exceedance graphs were generated for the top, middle, and bottom layers along the
centerline of the system showing the amount of time that each river mile exceeded daily average salinity
thresholds of 0.5 PSU, 1 PSU, and 5 PSU under the existing, 2050, and No GUC scenarios (Figures 4-4
through 4-6). Inset boxes show the river miles where the difference in percent exceedance between the
2050 and No GUC scenario was at least one percent.
For the 5 PSU threshold, the greatest simulated change in percent exceedance (3.4 percent) occurred at
river mile 40.4 in the surface layer where the No GUC exceedance occurred 7.6 percent of the time and
the 2050 exceedance occurred 11 percent of the time. The existing exceedance of the 5 PSU threshold
occurs 7.8 percent of the time. River mile 40.4 is located approximately 2 miles upstream of Washington.
The greatest predicted change in percent exceedance for the 1 PSU threshold occurs at river mile 45.1,
which is near Grimesland and downstream of the GUC WWTP. The 3.2 percent difference occurs in the
surface layer. Under the No GUC scenario, the 1 PSU threshold is exceeded 13.7 percent of time.
Under the 2050 scenario, the threshold is exceeded 16.9 percent of the time, and under the existing
scenario it is exceeded 14.3 percent of the time.
The simulated percent exceedance of the 0.5 PSU threshold is predicted to increase by approximately
4 percent in the surface, middle, and bottom layers between river miles 46.4 to 46.8. In the top layers,
this threshold is exceeded on average 10.6 percent of the time under the No GUC scenario, 14.2 percent
of time under the 2050 scenario, and 11.1 percent of time under the existing scenario. In the middle
layers, the percent exceedance for these three scenarios (as an average for the river miles 46.4 to 46.8)

96/114
is 11.1, 15.6, and 11.7, respectively. In the bottom layers, the percent exceedance for the 0.5 PSU
threshold in this section of the river is 12.0 for No GUC, 15.8 for 2050, and 12.8 for existing.
Thus, based on daily average salinities, the expected increase from existing conditions to 2050 is not
expected to exceed 3.2 percent, and the simulated impact of 2050 relative to No GUC is not expected to
exceed 4 percent at any location along the river.

97/114
Figure 4-4: Percent of Time 5 PSU is Exceeded by River Mile for Top, Middle, and Bottom Layers

98/114
Figure 4-5: Percent of Time 1 PSU is Exceeded by River Mile for Top, Middle, and Bottom Layers

99/114
Figure 4-6: Percent of Time 0.5 PSU is Exceeded by River Mile for Top, Middle, and Bottom Layers

100/114
Based on the comparison of the daily exceedances, the section of the system most impacted by
increases in GUC withdrawals is between river miles 40 to 50. To assess the impacts in this section of the
river more closely and to ensure protection of the GUC WTP withdrawal, hourly salinity data was output
for the existing and 2050 scenarios at select river miles for the bottom layers where saline waters tend to
accumulate and migrate upstream. [Note: hourly output was only generated for year 2008 because hourly
output was only developed for the water quality simulations, and sufficient water quality data were not
available to set up the water quality model for 2007.] Time series plots of the maximum daily salinities
(maximum of the hourly output generated each day) under existing and 2050 scenarios are shown in
Figures 4-7 through 4-11. Daily average Tar River at Greenville flows are shown on the top each figure to
show the correlation between river flow and salinity. As Tar River flows at Greenville decrease, salinity
increases at each location under existing and 2050 scenarios.
At river mile 44.5, there was little difference in the magnitude of bottom salinity between the two
scenarios. Natural swings in salinity due to tidal influence are an order of magnitude greater than the
maximum difference observed (1.52 PSU) between the two withdrawal scenarios.
At river mile 49.0 between Washington and the Greenville WWTP discharge, the trend is similar with tidal
influence dominating salinity impacts. The primary difference between existing and 2050 scenario is
timing, with higher salinities sometimes reaching river mile 49 approximately one day earlier under the
2050 scenario relative to the existing scenario. The difference in the magnitude of the salinities is
insignificant.
At river mile 56.5 upstream of the WWTP discharge and below the USGS Gage at Greenville, the
greatest difference in maximum daily bottom salinity occurs in early November. Under the 2050 scenario,
the value is approximately 0.2 PSU greater than under existing scenario, but does not exceed 0.4 PSU.
There is no significant difference in simulated daily maximum bottom salinities between existing and 2050
scenarios in the vicinity of the GUC WTP intake when year 2008 is used to drive the model.
In summary, there is little impact on salinity encroachment up the river under the 2050 scenario.
Additional analyses specific to habitat assessments are described in the Habitat Modeling TM (TM #7).

101/114
Figure 4-7: Maximum Daily Salinity in the Bottom Layer at River Mile 44.5 (1/2 mile downstream of
Grimesland) for Existing and 2050 Scenarios
Figure 4-8: Maximum Daily Salinity in the Bottom Layer at River Mile 49.0 (5 miles downstream of
GUC WWTP discharge) for Existing and 2050 Scenarios

102/114
Figure 4-9: Maximum Daily Salinity in the Bottom Layer at River Mile 56.5 (USGS Gage at
Greenville) for Existing and 2050 Scenarios
Figure 4-10: Maximum Daily Salinity in the Bottom Layer at River Mile 57.7 (0.3 miles downstream
of GUC WTP withdrawal) for Existing and 2050 Scenarios

103/114
Figure 4-11: Maximum Daily Salinity in the Bottom Layer at River Mile 60.6 (0.6 miles upstream of
GUC WTP withdrawal) for Existing and 2050 Scenarios

104/114
4.2.2 Temperature and Salinity Impacts at Hourly Scale
To assess changes in simulated conditions at a more refined temporal scale, the LTPR model was used
to output hourly values of temperature and salinity at select locations in the river and estuary, mostly
corresponding to DWQ ambient monitoring stations. The following sections show the simulated
differences for each parameter as percentiles for the surface, middle, and bottom two layers as averages.
The 0 percentile column shows the simulated minimum, and the 100 percentile shows the simulated
maximum. The 50th percentile is the median, which means that half the time values were less than and
half the time they were greater.
The tables are ordered by river mile with stations located at the mouth of the Pamlico at the top of the
table and stations located near Greenville at the bottom of each table.
4.2.2.1 Temperature
Tables 4-3 through 4-4 show the hourly percentiles for the existing, 2050, and No GUC scenarios in the
top, middle, and bottom layers, respectively. In terms of protecting aquatic life, increases in temperature
may be detrimental due to changes in biological and chemical reaction rates and alteration of saturation
levels for dissolved gases, such as oxygen. There are no increases in percentile values for temperature
greater than 1°C for the 2050 scenario relative to the No GUC scenario. The largest changes occurs at
river mile 53.3 downstream of the GUC WWTP in the top two layers where the 10th percentile temperature
under the No GUC scenario is 7.2°C (45.0°F) and the 10th percentile temperature under the 2050
scenario is 8.2°C (46.8°F).

105/114
Table 4-3: Hourly Temperature (°C) Percentiles in the Top Two Layers
WQ Station Scenario
Top Two Layers
Percentile
0 10 25 50 75 90 100
O982500C (RM 7.1)
Existing 1.99 7.15 10.40 17.05 26.00 27.30 29.80
2050 2.45 7.09 10.34 17.15 25.95 27.35 29.80
No GUC 1.95 7.18 10.40 17.10 25.90 27.30 29.90
O9059000 (RM 13.0)
Existing 1.26 6.79 10.56 17.15 25.30 26.85 29.35
2050 1.82 6.80 10.55 17.30 25.40 26.80 29.20
No GUC 1.77 6.84 10.50 17.25 25.40 26.70 29.20
O865000C (RM 18.1)
Existing 2.01 7.16 10.65 17.25 25.50 26.90 29.90
2050 2.03 7.14 10.60 17.25 25.50 26.90 29.70
No GUC 2.27 7.07 10.60 17.20 25.50 26.90 29.80
O8498000 (RM 22.4)
Existing 2.31 7.04 10.90 17.35 25.15 26.80 29.70
2050 2.34 6.97 10.90 17.30 25.15 26.80 29.85
No GUC 2.37 7.01 10.85 17.35 25.20 26.80 29.70
O787000C (RM 29.5)
Existing 3.55 7.59 11.10 17.45 25.30 26.90 30.00
2050 3.40 7.57 11.10 17.35 25.30 26.90 29.80
No GUC 3.55 7.66 11.10 17.35 25.25 27.00 29.80
O7680000 (RM 34.2)
Existing 4.64 8.57 11.55 17.30 25.35 26.90 30.00
2050 4.74 8.59 11.60 17.45 25.30 26.90 30.00
No GUC 4.62 8.54 11.55 17.30 25.30 27.00 30.20
O7650000 (RM 37.6)
Existing 6.08 9.35 12.20 17.40 26.00 27.65 32.10
2050 6.10 9.39 12.25 17.30 25.95 27.50 32.10
No GUC 6.07 9.31 12.15 17.40 26.00 27.70 32.20
O6500000 (RM 44.4)
Existing 5.11 8.89 11.55 18.00 26.30 28.10 31.30
2050 5.17 9.11 11.65 18.00 26.25 28.05 31.25
No GUC 5.04 8.74 11.50 18.00 26.30 28.10 31.20
Downstream of GUC WWTP (RM 53.3)
Existing 3.48 7.45 11.25 17.30 25.00 26.90 29.90
2050 3.81 8.19 11.50 17.80 25.40 27.00 29.60
No GUC 3.23 7.23 11.18 17.00 24.90 26.90 30.30
Greenville Gage (RM 57.7)
Existing 3.96 7.70 10.90 16.80 24.90 27.40 30.10
2050 3.95 7.71 10.90 16.80 24.80 27.30 30.20
No GUC 3.96 7.70 10.90 16.80 25.00 27.50 30.10
Above GUC WTP (RM 60.4)
Existing 4.67 7.81 10.90 16.90 25.60 28.40 31.90
2050 4.66 7.80 10.90 16.85 25.60 28.40 31.90
No GUC 4.66 7.81 10.90 16.90 25.65 28.40 31.85
02083893 - US264 Bypass Gage (RM 61.8)
Existing 4.96 7.93 11.00 17.00 26.40 29.50 31.90
2050 4.96 7.92 11.00 17.00 26.40 29.50 31.90
No GUC 4.97 7.93 11.00 17.00 26.40 29.50 31.90

106/114
Table 4-4: Hourly Temperature (°C) Percentiles in the Middle Two Layers
Middle Two Layers
WQ Station Scenario
Percentile
0 10 25 50 75 90 100
O982500C (RM 7.1)
Existing 4.09 8.16 10.80 17.55 26.40 27.40 29.50
2050 4.53 8.23 10.80 17.50 26.40 27.45 29.45
No GUC 4.00 8.19 10.75 17.50 26.40 27.40 29.45
O9059000 (RM 13.0)
Existing 3.92 8.24 10.65 18.10 26.10 27.45 29.65
2050 4.40 8.22 10.61 17.95 26.10 27.45 29.55
No GUC 4.24 8.26 10.68 18.00 26.15 27.45 29.60
O865000C (RM 18.1)
Existing 4.79 8.18 10.75 18.05 26.40 27.75 29.95
2050 4.68 8.40 10.80 18.15 26.45 27.70 29.85
No GUC 4.88 8.41 10.75 18.10 26.40 27.85 29.90
O8498000 (RM 22.4)
Existing 5.07 8.45 11.00 18.65 26.15 27.35 30.15
2050 5.04 8.51 10.95 18.55 26.10 27.30 29.70
No GUC 5.01 8.48 11.00 18.70 26.10 27.40 30.00
O787000C (RM 29.5)
Existing 5.40 8.69 11.30 18.35 25.95 27.55 30.15
2050 5.46 8.66 11.35 18.40 26.05 27.50 29.80
No GUC 5.44 8.70 11.35 18.40 26.05 27.50 30.00
O7680000 (RM 34.2)
Existing 6.70 9.34 11.95 18.85 26.45 27.85 30.40
2050 6.69 9.34 12.00 18.85 26.40 27.80 30.30
No GUC 6.63 9.35 12.05 18.90 26.40 27.90 30.45
O7650000 (RM 37.6)
Existing 6.93 10.05 12.65 18.20 27.85 29.55 32.30
2050 6.93 10.04 12.60 18.20 27.70 29.50 31.90
No GUC 6.93 10.04 12.65 18.20 27.90 29.55 32.15
O6500000 (RM 44.4)
Existing 5.12 10.20 12.20 18.40 27.55 30.30 32.70
2050 5.18 10.25 12.30 18.40 27.50 30.20 32.55
No GUC 5.04 10.10 12.20 18.40 27.35 30.35 32.60
Downstream of GUC WWTP
(RM 53.3)
Existing 3.48 7.36 11.30 18.10 25.00 26.80 29.80
2050 3.81 7.59 11.40 18.40 25.20 26.80 29.50
No GUC 3.24 7.23 11.20 18.00 24.90 26.80 30.00
Greenville Gage (RM 57.7)
Existing 3.97 7.71 10.90 16.80 24.90 27.40 30.10
2050 3.96 7.72 10.90 16.80 24.80 27.30 30.20
No GUC 3.97 7.70 10.90 16.80 25.00 27.50 30.10
Above GUC WTP (RM 60.4)
Existing 4.67 7.81 10.90 16.90 25.60 28.40 31.90
2050 4.66 7.80 10.90 16.90 25.60 28.40 31.90
No GUC 4.67 7.81 10.90 16.90 25.70 28.40 31.90
02083893 - US264 Bypass Gage
(RM 61.8)
Existing 4.96 7.93 11.00 17.00 26.40 29.50 31.90
2050 4.96 7.93 11.00 17.00 26.40 29.50 31.90
No GUC 4.97 7.94 11.00 17.00 26.40 29.50 31.95

107/114
Table 4-5: Hourly Temperature (°C) Percentiles in the Bottom Two Layers
Bottom Two Layers
WQ Station Scenario
Percentile
0 10 25 50 75 90 100
O982500C (RM 7.1)
Existing 6.00 7.93 10.65 18.20 26.40 27.80 29.35
2050 5.85 7.95 10.75 18.20 26.40 27.70 29.40
No GUC 6.02 7.97 10.70 18.20 26.40 27.70 29.40
O9059000 (RM 13.0)
Existing 5.20 8.19 10.90 17.40 26.30 27.80 29.70
2050 5.56 8.14 11.00 17.50 26.45 27.75 29.60
No GUC 5.57 8.14 10.90 17.50 26.50 27.80 29.50
O865000C (RM 18.1)
Existing 5.44 8.23 11.00 17.60 26.50 27.70 29.90
2050 5.42 8.26 11.00 17.55 26.50 27.65 29.90
No GUC 5.51 8.29 11.00 17.65 26.50 27.75 29.80
O8498000 (RM 22.4)
Existing 5.93 8.51 11.10 18.50 26.60 27.85 30.30
2050 5.77 8.52 11.15 18.10 26.60 27.90 30.30
No GUC 5.79 8.48 11.10 18.25 26.70 27.85 30.50
O787000C (RM 29.5)
Existing 6.42 8.88 11.60 18.85 26.75 27.65 30.25
2050 6.35 8.89 11.55 18.85 26.70 27.60 30.05
No GUC 6.44 8.88 11.55 18.85 26.75 27.65 30.30
O7680000 (RM 34.2)
Existing 6.95 9.28 12.05 18.80 26.70 27.75 30.00
2050 6.95 9.29 12.10 18.80 26.65 27.70 30.20
No GUC 6.95 9.31 12.10 18.90 26.70 27.80 30.20
O7650000 (RM 37.6)
Existing 6.88 9.77 13.00 18.50 27.80 29.25 32.10
2050 6.88 9.75 13.00 18.50 27.70 29.10 31.90
No GUC 6.88 9.74 13.00 18.50 27.80 29.20 32.00
O6500000 (RM 44.4)
Existing 5.12 10.30 12.40 18.60 28.30 30.65 32.75
2050 5.18 10.35 12.40 18.70 28.15 30.55 32.65
No GUC 5.04 10.24 12.40 18.60 28.15 30.75 32.70
Downstream of GUC WWTP (RM 53.3)
Existing 3.48 7.37 11.30 19.00 24.95 26.70 29.40
2050 3.81 7.68 11.40 19.20 25.15 26.70 29.30
No GUC 3.24 7.24 11.20 18.85 24.90 26.70 29.35
Greenville Gage (RM 57.7)
Existing 3.97 7.71 10.90 16.80 24.90 27.40 30.10
2050 3.96 7.72 10.90 16.80 24.80 27.30 30.20
No GUC 3.97 7.70 10.90 16.80 25.00 27.50 30.10
Above GUC WTP (RM 60.4)
Existing 4.67 7.81 10.90 16.90 25.60 28.40 31.90
2050 4.66 7.80 10.90 16.90 25.60 28.40 31.90
No GUC 4.67 7.81 10.90 16.90 25.70 28.40 31.90
02083893 - US264 Bypass Gage (RM 61.8)
Existing 4.97 7.93 11.00 17.00 26.40 29.50 31.90
2050 4.96 7.93 11.00 17.00 26.40 29.50 31.90
No GUC 4.97 7.94 11.00 17.00 26.40 29.50 32.00

108/114
4.2.2.2 Salinity
Tables 4-6 through 4-8 show the hourly salinity percentiles for the three scenarios for the top, middle, and
bottom two layers. The largest difference between the salinity percentiles occurs between the No GUC
and 2050 scenario in the bottom layer at river mile 53.3, which is downstream of the GUC WWTP. The
90th percentile salinity value at this location in the bottom layer under the No GUC scenario is 0.08 PSU
while under the 2050 scenario the 90th percentile value is approximately 1 PSU. The simulated
maximums at this location are similar with values of 7.37 PSU for No GUC and 7.54 PSU for 2050.

109/114
Table 4-6: Hourly Salinity (PSU) Percentiles in the Top Two Layers

Top Two Layers
WQ Station Scenario
Percentile
0 10 25 50 75 90 100
O982500C (RM 7.1)
Existing 8.88 12.20 14.25 17.85 19.50 21.00 22.30
2050 8.84 12.10 14.25 17.80 19.55 21.00 22.30
No GUC 8.85 12.20 14.20 17.80 19.50 20.95 22.30
O9059000 (RM 13.0)
Existing 6.79 10.14 12.30 15.85 17.70 19.35 21.40
2050 6.80 10.35 12.20 15.85 17.70 19.40 21.40
No GUC 6.80 10.29 12.15 15.80 17.70 19.40 21.35
O865000C (RM 18.1)
Existing 5.47 8.97 11.10 14.85 16.70 18.55 20.70
2050 5.53 8.93 11.20 14.90 16.75 18.60 20.65
No GUC 5.54 8.98 11.10 14.80 16.70 18.55 20.45
O8498000 (RM 22.4)
Existing 4.00 6.97 9.77 13.25 15.40 16.95 20.00
2050 4.02 6.96 9.94 13.30 15.40 16.95 20.00
No GUC 3.98 6.95 9.84 13.20 15.35 16.95 19.90
O787000C (RM 29.5)
Existing 2.06 4.61 7.91 11.50 13.45 15.20 17.60
2050 2.05 4.61 8.03 11.50 13.50 15.30 17.65
No GUC 2.03 4.59 7.91 11.40 13.40 15.20 17.65
O7680000 (RM 34.2)
Existing 0.24 1.91 5.47 9.84 11.90 13.80 17.15
2050 0.24 1.91 5.52 9.90 11.95 13.90 17.20
No GUC 0.23 1.88 5.48 9.77 11.80 13.75 17.20
O7650000 (RM 37.6)
Existing 0.02 0.12 2.94 7.39 10.40 12.55 18.10
2050 0.02 0.13 2.99 7.51 10.50 12.65 17.25
No GUC 0.02 0.11 2.94 7.36 10.33 12.50 17.50
O6500000 (RM 44.4)
Existing 0.02 0.04 0.05 1.67 4.04 6.10 10.70
2050 0.02 0.04 0.06 1.75 4.22 6.38 11.05
No GUC 0.02 0.04 0.05 1.64 3.95 6.05 10.70
Downstream of GUC WWTP (RM 53.3)
Existing 0.02 0.04 0.04 0.06 0.08 0.12 4.32
2050 0.02 0.04 0.05 0.07 0.13 0.24 4.74
No GUC 0.02 0.04 0.04 0.05 0.07 0.08 5.23
Greenville Gage (RM 57.7)
Existing 0.02 0.04 0.04 0.05 0.07 0.07 0.16
2050 0.02 0.04 0.04 0.05 0.07 0.07 0.16
No GUC 0.02 0.04 0.04 0.05 0.07 0.07 0.16
Above GUC WTP (RM 60.4)
Existing 0.02 0.04 0.04 0.05 0.07 0.07 0.09
2050 0.02 0.04 0.04 0.05 0.07 0.07 0.09
No GUC 0.02 0.04 0.04 0.05 0.07 0.07 0.09
02083893 - US264 Bypass Gage (RM 61.8)
Existing 0.02 0.04 0.04 0.05 0.07 0.07 0.09
2050 0.02 0.04 0.04 0.05 0.07 0.07 0.09
No GUC 0.02 0.04 0.04 0.05 0.07 0.07 0.09

110/114
Table 4-7: Hourly Salinity (PSU) Percentiles in the Middle Two Layers

Middle Two Layers
WQ Station Scenario
Percentile
0 10 25 50 75 90 100
O982500C (RM 7.1)
Existing 11.70 14.35 16.05 19.00 20.60 22.05 22.60
2050 11.65 14.35 16.10 19.00 20.60 22.10 22.65
No GUC 11.70 14.35 16.05 19.00 20.55 22.05 22.75
O9059000 (RM 13.0)
Existing 10.20 12.95 14.55 17.90 19.65 21.20 22.10
2050 9.95 12.95 14.45 18.00 19.65 21.20 22.05
No GUC 10.27 12.95 14.50 17.95 19.65 21.15 22.25
O865000C (RM 18.1)
Existing 9.53 12.20 14.00 17.15 19.10 20.45 21.55
2050 9.46 12.20 14.00 17.15 19.10 20.45 21.60
No GUC 9.48 12.20 14.00 17.10 19.10 20.35 21.60
O8498000 (RM 22.4)
Existing 7.07 10.40 12.45 15.80 18.10 19.30 21.05
2050 7.14 10.41 12.50 15.90 18.10 19.35 20.90
No GUC 7.06 10.40 12.45 15.75 18.10 19.35 21.00
O787000C (RM 29.5)
Existing 5.32 8.31 10.81 13.85 16.75 18.05 19.55
2050 5.37 8.36 10.90 13.85 16.75 18.05 19.65
No GUC 5.33 8.33 10.85 13.80 16.80 18.05 19.55
O7680000 (RM 34.2)
Existing 1.24 6.84 9.89 13.15 16.35 17.85 20.10
2050 1.17 6.88 9.94 13.15 16.30 17.85 20.10
No GUC 1.25 6.88 9.89 13.05 16.35 17.85 19.80
O7650000 (RM 37.6)
Existing 0.02 0.49 7.49 11.95 15.50 17.35 19.50
2050 0.03 0.56 7.55 12.00 15.50 17.35 19.50
No GUC 0.02 0.41 7.51 11.90 15.50 17.35 19.50
O6500000 (RM 44.4)
Existing 0.02 0.04 0.05 4.74 8.78 11.55 14.40
2050 0.02 0.04 0.06 4.78 8.88 11.55 14.40
No GUC 0.02 0.04 0.05 4.66 8.81 11.55 14.30
Downstream of GUC WWTP (RM 53.3)
Existing 0.02 0.04 0.04 0.05 0.08 0.11 5.66
2050 0.02 0.04 0.05 0.06 0.11 0.32 5.80
No GUC 0.02 0.04 0.04 0.05 0.07 0.08 5.60
Greenville Gage (RM 57.7)
Existing 0.02 0.04 0.04 0.05 0.07 0.07 0.21
2050 0.02 0.04 0.04 0.05 0.07 0.07 0.21
No GUC 0.02 0.04 0.04 0.05 0.07 0.07 0.21
Above GUC WTP (RM 60.4)
Existing 0.02 0.04 0.04 0.05 0.07 0.07 0.10
2050 0.02 0.04 0.04 0.05 0.07 0.07 0.10
No GUC 0.02 0.04 0.04 0.05 0.07 0.07 0.10
02083893 - US264 Bypass Gage (RM 61.8)
Existing 0.02 0.04 0.04 0.05 0.07 0.07 0.09
2050 0.02 0.04 0.04 0.05 0.07 0.07 0.09
No GUC 0.02 0.04 0.04 0.05 0.07 0.07 0.09

111/114
Table 4-8: Hourly Salinity (PSU) Percentiles in the Bottom Two Layers

Bottom Two Layers
WQ Station Scenario
Percentile
0 10 25 50 75 90 100
O982500C (RM 7.1)
Existing 14.10 15.55 16.90 19.70 22.10 23.05 24.05
2050 14.10 15.50 16.95 19.70 22.10 23.05 24.05
No GUC 13.85 15.55 16.95 19.70 22.10 23.00 24.05
O9059000 (RM 13.0)
Existing 13.25 15.15 16.30 19.10 21.45 22.15 23.25
2050 13.25 15.10 16.30 19.10 21.50 22.20 23.25
No GUC 13.10 15.10 16.30 19.05 21.45 22.20 23.25
O865000C (RM 18.1)
Existing 12.15 14.20 15.40 18.10 20.50 21.60 22.25
2050 12.40 14.20 15.35 18.10 20.45 21.60 22.30
No GUC 12.20 14.20 15.40 18.00 20.50 21.60 22.20
O8498000 (RM 22.4)
Existing 11.50 13.20 14.70 17.40 19.95 21.10 21.70
2050 11.45 13.20 14.70 17.40 19.90 21.10 21.75
No GUC 11.50 13.20 14.70 17.35 19.90 21.10 21.70
O787000C (RM 29.5)
Existing 9.44 11.00 12.85 15.60 18.25 19.75 20.80
2050 9.43 11.00 12.85 15.55 18.20 19.70 20.80
No GUC 9.45 10.95 12.85 15.60 18.25 19.70 20.90
O7680000 (RM 34.2)
Existing 5.45 8.86 11.40 14.20 17.30 19.00 20.15
2050 5.49 8.90 11.40 14.20 17.30 18.95 20.15
No GUC 5.38 8.87 11.40 14.15 17.30 19.00 20.15
O7650000 (RM 37.6)
Existing 0.02 1.93 9.28 13.25 16.20 18.20 19.90
2050 0.03 2.04 9.34 13.10 16.10 18.30 19.90
No GUC 0.02 1.80 9.26 13.30 16.20 18.20 19.90
O6500000 (RM 44.4)
Existing 0.02 0.04 0.05 6.01 10.30 12.70 14.80
2050 0.02 0.04 0.06 6.15 10.52 12.70 14.90
No GUC 0.02 0.04 0.05 5.94 10.20 12.70 14.95
Downstream of GUC WWTP (RM 53.3)
Existing 0.02 0.04 0.04 0.05 0.08 0.11 7.41
2050 0.02 0.04 0.05 0.06 0.10 0.99 7.54
No GUC 0.02 0.04 0.04 0.05 0.07 0.08 7.37
Greenville Gage (RM 57.7)
Existing 0.02 0.04 0.04 0.05 0.07 0.07 0.22
2050 0.02 0.04 0.04 0.05 0.07 0.07 0.22
No GUC 0.02 0.04 0.04 0.05 0.07 0.07 0.22
Above GUC WTP (RM 60.4)
Existing 0.02 0.04 0.04 0.05 0.07 0.07 0.10
2050 0.02 0.04 0.04 0.05 0.07 0.07 0.10
No GUC 0.02 0.04 0.04 0.05 0.07 0.07 0.10
02083893 - US264 Bypass Gage (RM 61.8)
Existing 0.02 0.04 0.04 0.05 0.07 0.07 0.09
2050 0.02 0.04 0.04 0.05 0.07 0.07 0.09
No GUC 0.02 0.04 0.04 0.05 0.07 0.07 0.09

112/114
5. LTPR Model Summary
This section of the memorandum provides a summary of the LTPR model limitations, general findings
with respect to temperature and salinity, and recommendations for future analysis. The goals of the Tar
River Flow Study were to develop the tools to better understand the hydrodynamic, salinity, and habitat
conditions in the river, and to quantify the effects of potential future water withdrawals and possible
constraints to future water use in the Tar River. GUC has stated that an ultimate goal was to develop of
set of analytical evaluative tools that could be used now and in the future. The approaches and models
were developed to serve as sound basis for current and future management evaluation and decisions, to
provide expandability through improvement and refinement, and provide a basis for permitting of the
future expansion of the Water Treatment Plant and other water management actions.
The LTPR Model has met all of these goals and provides an excellent tool to assess changes in
hydrodynamics of the system, particularly during low flow years. The LTPR model effectively and
accurately simulates the dynamics of river flow and hydrodynamic behavior of the Tar-Pamlico system,
including tidal movements, hydraulics, salinity stratification, and salinity regime. Prior to the development
of the LTPR model, there was not a current or completely functional model of the lower Tar River and
Upper Pamlico. The LTPR model was developed with the state-of-the-art EFDC model, which is a sound
platform for future use and enhancement to its hydrodynamic and water quality modeling capabilities.
The following summarizes the salinity impacts resulting from a 2050 scenario relative to existing
conditions. Habitat impacts are described in the Habitat Modeling TM.
• The percent of time that salinity thresholds of 5 PSU, 1 PSU, and 0.5 PSU are exceeded
at any given river mile increases by less than four percent from existing conditions to
2050 conditions.
• Daily maximum salinities, based on hourly salinity output, indicate that salinity changes
downstream of the GUC WWTP due to increased consumptive use are an order of
magnitude less than the changes from tidal fluctuations under existing conditions.
Upstream of the WWTP, daily maximum salinities are not predicted to increase by more
than 0.2 PSU from the existing to 2050 scenario.
• The 2050 scenario has little impact on simulated temperatures.
• Salinity impacts due to a consumptive use of 14.7 mgd appear insignificant in this tidally
influenced system.

113/114
6. References
ARCADIS, 2008. Final Environmental Assessment for Greenville Utilities Interbasin Transfer.
Cerco, C., and Cole, T. 1994. “Three-dimensional eutrophication model of Chesapeake Bay,”Technical
Report EL-94-4, U.S. Army Engineer Waterways Experiment Station, Vicksburg, MS.
Di Toro, D.M. Sediment Flux Modeling. J. Wiley and Sons., New York: (2001), 624p.
Di Toro, D.M. 1978. Optics of turbid estuarine waters: approximations and applications. Water Res.
12:1059-1068.
ENTRIX. 2007. Analysis of Greenville Utilities Commission’s Proposed Interbasin Transfer Withdrawals
on Tar River Flows at Greenville, North Carolina, October 2007, revised April 2008.
Environmental Protection Agency (EPA). 2009. Environmental Fluid Dynamics Computer Code.
Environmental Protection Agency, Ecosystems Division. Available:
http://www.epa.gov/athens/research/modeling/efdc.html. (March 2009).
Hazen and Sawyer. 2012. Water Distribution System Master Plan.
Lin, J., L. Xie, and L. J. Pietrafesa. 2007. Water quality gradient across Albemarle-Pamlico Estuarine
System: seasonal variations and model applications. Journal of Coastal Research 23:213-229.
Lin, J., H. Xu, C. Cudaback, and D. Wang. 2008. Inter-annual variability of hypoxic conditions in a shallow
estuary. Journal of Marine Systems 23:169-184.
Lung, W.S. and H.W. Paerl 1988 Modeling blue-green algal blooms in the Lower Neuse River Estuary.
Water Research, Vol. 22, Issue 7, July 1988, Pages 895-905.
Matson, Brinson, Cahoon, Davis (1983) Biogeochemistry of the sediments of the Pamlico and Neuse
River estuaries, North Carolina. Water Resources Research Institute, University North Carolina,
WRRI Report No. 191, NC (UNC-WRRI-83-191).
O’Driscoll, M.A., D.J. Mallinson, and P.K. Johnson. 2008. Surface Water/Ground Water Interactions Along
the Tar River, NC. Water Resources Research Institute of the University of North Carolina,
Report No. 370.
ODEQ (Oklahoma Department of Environmental Quality), (2006), TMDL Development For Cobb Creek
Watershed And Fort Cobb Lake, Final Report.
Park, K., Kuo, A.Y., Butt, A.J., 1995. Application of a tidal prism water quality model to the Lynnhaven
River, Virginia. A report to the Virginia Coastal Resources Management Program, Virginia
Department of Environmental Quality, Special Report No. 329 in Applied Marine Science and

114/114
Ocean Engineering, Virginia Institute of Marine Science/School of Marine Science, The College of
William and Mary, Virginia.
Park, K., Jung, H.-S., Kim, H.-S., Ahn, S.-M., 2005. Three-dimensional hydrodynamic-eutrophication
model (HEM-3D): application to Kwang-Yang Bay, Korea. Mar. Environ. Res. 60, 171–193.
Spruill, T.B. and J.F. Bratton. 2008. Estimation of Groundwater and Nutrient Fluxes to the Neuse River
Estuary, North Carolina. Estuaries and Coasts 31: 501-520.
Thomann, R.V. and J. A. Mueller. 1987. Principles of Surface Water Quality Modeling and Control. New
York: Harper & Row, Pub., Inc.
Xu, H., J. Lin, and D. Wang. 2008. Numerical study on salinity stratification in the Pamlico River Estuary.
Estuarine, Coastal and Shelf Science 80:74-84.