Skip to main content

Full text of "Modeling the circulation and water quality in Charlotte Harbor Estuarine System, Florida"

See other formats


MODELING THE CIRCULATION AND WATER QUALITY 
IN CHARLOTTE HARBOR ESTUARINE SYSTEM, FLORIDA 



By 

KIJFN PARK 



A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL 

OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT 

OF THE REQUIREMENTS FOR THE DEGREE OF 

DOCTOR OF PHILOSOPHY 



UNIVERSITY OF FLORIDA 
2004 



ACKNOWLEDGMENTS 

I would like to thank my advisor, Dr. Y. Peter Sheng, for his guidance, support and 
financial assistance throughout my study. In addition, much appreciation is owned to my 
other committee members, Dr. Robert G. Dean, Dr. Robert J. Thieke, Dr. Louis H. Motz and 
Dr. K. Ramesh Reddy, for their review of my dissertation. 

I would like to thank the South Florida Water Management District and Southwest 
florida Water Management District for sponsoring research project and providing data for 
Charlotte Harbor estuarine system. I would like to express my thanks to Justin, Jeffery, 
Yangfeng, Taeyoon, Jun, Vadim, and Vladimir whose help with class, research and writing 
this dissertation. Many thanks go to Becky, Lucy, Sonna, Sidney, Kim and Halen for making 
life easier. 

I would like to dedicate this dissertation to my parents whose love and support made 
this degree possible. Last, but not least, I would like to thank my wife, Kyung-Mi, who have 
been praying for me to be faithful, kind, and honest all the time. 



11 



TABLE OF CONTENTS 

Page 

ACKNOWLEDGMENTS » 

LIST OF TABLES vi 

LIST OF FIGURES ix 

ABSTRACT xvil 

CHAPTER 

1 INTRODUCTION 1 

2 CHARLOTTE HARBOR CHARACTERIZATION 9 

2.1 Climate U 

2.2 Hydrodynamics 1 1 

2.2.1 Tidal Stage, Discharge at Inlet, and Tidal Circulation 13 

2.2.2 Freshwater Flow 14 

2.2.3 Salinity 14 

2.3 Water Quality 15 

2.3.1 Nutrients 16 

2.3.2 Dissolved Oxygen 18 

2.3.3 Phytoplankton 19 

2.4 Sediment 20 

2.5 Light Environment 21 

3 HYDRODYNAMIC AND SEDIMENT TRANSPORT MODEL 23 

3.1 Governing Equation 24 

3.1.1 Hydrodynamic Model 24 

3.1.2 Sediment Transport Model 26 

3.2 Boundary and Initial Conditions 27 

3.2.1 Boundary Conditions 27 

3.2.2 Initial Conditions 30 

3.3 Heat-Flux at Air-Sea Interface 30 

3.3.1 Short-Wave Solar Radiation 31 

3.3.2 Long-Wave Solar Radiation 32 

in 



3.3.3 Sensible and Latent Heat Fluxes 33 

4 WATER QUALITY MODEL 36 

4.1 Mathematical Formulae 38 

4.2 Phytoplankton Dynamics 40 

4.2.1 Modeling Approach 40 

4.2.2 Relationship between Phytoplankton and Nutrients 42 

4.3 Nutrient Dynamics 43 

4.4 Oxygen Balance 45 

4.5 Effects of Temperature and Light Intensity on Water Quality Model 59 

4.5.1 Temperature 59 

4.5.2 Light intensity 61 

4.6 Light Attenuation Model 61 

4.7 Model Parameters and Calibration Procedures 69 

5 APPLICATION OF CIRCULATION AND TRANSPORT MODEL 81 

5.1 A High-Resolution Curvilinear Grid for Charlotte Harbor Estuarine System 82 

5.2 Forcing Mechanism and Boundary Condition 86 

5.3 Simulations for 1986 Hydrodynamics 94 

5.3.1 Sensitivity and Calibration Simulations 99 

5.3.2 Results of the 1986 simulation 101 

5.4 Simulations for 2000 Hydrodynamic 118 

5.4.1 Sensitivity and Calibration Simulations 118 

5.4.2 Results of the 2000 Simulation 130 

5.4.3 Application of the 2000 hydrodynamic Simulations 146 

6 APPLICATION OF WATER QUALITY MODEL 167 

6.1 Forcing Mechanism and Boundary Condition for Circulation 167 

6.2 Initial and Boundary Condition for the Water Quality Model 172 

6.3 Simulations of Water Quality in 1996 180 

6.3.1 Calibration 180 

6.3.2 Results for 1996 Water Quality Simulations 191 

6.4 Simulations of Water Quality in 2000 212 

6.4.1 Verification 212 

6.4.2 Results of 2000 Water Quality Simulations 215 

6.4.3 Application of 2000 Water Quality Simulations 235 

7 CONCLUSION AND DISCUSSION 258 

APPENDIX 

A FLOW CHARTS FOR CH3DJMS 263 



IV 



B DMENSIONLESS EQUATIONS IN CURVILINEAR BOUNDARY-FITTED AND 
SIGMA GRID 278 

C COMPARISON OF WATER QUALITY MODELS 284 

D NUMERICAL SOLUTION TECHNIQUES FOR WATER QUALITY MODEL . . 289 

E NUTRIENT DYNAMICS 294 

F SEDIMENT FLUX MODEL 306 

G MODEL PERFORMANCE TEST WITH PARALLEL CH3D_IMS 309 

H TIDAL BENCH MARKS FOR CHARLOTTE HARBOR 313 

REFERENCES 314 

BIOGRAPHICAL SKETCH 325 



LIST OF TABLES 
Table E^e 

1.1 Component models of the CH3D-IMS 6 

2.1 Ratios of nitrogen and phosphorous constituents 18 

3.1 Mean latitudinal values of the coefficient X 33 

4. 1 Average values of oxygen uptake rates of bottom 54 

4.2 The spectrum of incident sunlight data 64 

4.3 Coefficient ranges to use in stand along light model 67 

4.4 The best fit light model coefficients for Charlotte Harbor estuarine system 67 

4.5 Dixon and Gray's model coefficients for the Charlotte Harbor estuarine system.. . 67 

4.6 Temperature adjustment functions for water quality parameters 70 

4.7 Water quality parameters related to conversion rate 71 

4.8 Water quality parameters related to phytoplankton and zooplankton 71 

4.9 Water quality parameters in the nitrogen dynamics 72 

4.10 Water quality parameters in the phosphorous dynamics 73 

4.11 Water quality parameters in the oxygen balance 74 

4.12 The relationship between water quality parameters and model constituents 79 

5.1 Descriptions of 1986 and 2000 river boundary conditions for Charlotte Harbor 

estuarine system 87 

5.2 Locations of tidal stage and velocity-salinity measured stations by USGS 96 



VI 



5.3 The effect of removing selected boundary conditions on the accuracy of simulated 

water level, velocity and salinity in July 1986. Value shown are average RMS 
differences vs. baseline simulation at all data stations 99 

5.4 The effect of varying bottom roughness, z , on the accuracy of simulated water level, 

velocity and salinity in July 1986. Value shown are average RMS errors at all 
data stations 100 

5.5 The effect of varying horizontal diffusion, A H , on the accuracy of simulated water 

level, velocity and salinity in July 1986. Value shown average RMS errors at all 
data stations 101 

5.6 A summary of boundary conditions and model parameters used in the 1986 

simulation 102 

5.7 Calculated RMS errors between simulated and measured water level for 1986 

simulation 102 

5.8 Calculated RMS errors between simulated and measured current velocity for 1986 

simulation 104 

5.9 Calculated RMS errors between simulated and measured salinity for 1986 

simulation 105 

5.10 The effect of horizontal grid resolution, on the accuracy of simulated water level 

and salinity. Values shown are average RMS errors for 2000 calibration at all 
available stations. Values shown in parenthesis are % RMS error normalized by 
maximum values 121 

5.1 1 The effect of vertical grid resolution, on the accuracy of simulated water level and 

salinity. Values shown are average RMS errors for 2000 calibration at all 
available stations. Values shown in parenthesis are % RMS error normalized by 
maximum values 122 

5.12 The effect of varying bottom roughness, z , on the accuracy of simulated water level 

and salinity in 2000. Values shown are average RMS errors all data stations. . 123 

5.13 The effect of varying salinity advection scheme on the accuracy of simulated water 

level and salinity in 2000. Values shown are average RMS errors at all data 
stations 128 

5.14 The effect of modifying bathymetry on the accuracy of simulated water level and 

salinity in 2000. Values shown are average RMS errors for all data stations. . 129 



vn 



5. 15 A summary of boundary conditions and model parameters used in the 2000 

simulation ljyj 

5.16 Calculated RMS errors between simulated and measured water level for 2000 

simulation "1 

5.17 Calculated RMS errors between simulated and measured salinity for 2000 

simulation "3 

5.18 The effect of hydrologic alternations on 2000 water level. Value shown are 

average RMS differences with baseline simulation for all selected stations. . . 154 

5.19 The effect of hydrologic alternations on 2000 salinity. Value shown are average 

RMS differences with baseline simulation for all selected stations 155 

6.1 Locations of water quality measured stations 173 

6.2 Sediment types for Charlotte Harbor water quality simulations 174 

6.3 Water quality parameters, baseline values, and ranges used in the sensitivity 

analysis 1" 

6.4 Sensitivity analysis results in RMS difference w.r.t. baseline for 1996 water quality 

calibration simulation 184 

6.5 The water quality model coefficients used for the Charlotte Harbor simulation. . . 185 

6.6 The temporally averaged water quality species concentrations for baseline 2000 

simulation 240 

6.7 Normalized RMS differences of water quality species concentrations at 1 1 stations 

during April to June 2000, showing the effect of no causeway islands 240 

6.8 Normalized RMS differences of water quality species concentrations at 11 stations 

during April to June 2000, showing the effect of no ICW 241 

6.9 Normalized RMS differences of watei quality species concentrations at 1 1 stations 

during April to June 2000, showing the effect of no causeway islands and ICW241 

C.l Comparison of water quality models 287 

G. 1 CPU time for the parallel, shared memory, CH3D procedures on SGI origin 

platform 311 

H.l Tidal datum referred to Mean Low Low Water (MLLW), in meter 313 

viii 



LIST OF FIGURES 

Figure Page 
1.1 Charlotte Harbor estuarine system and its subarea boundaries 2 

2.1 Drainage basins of Charlotte Harbor estuarine system 10 

2.2 Seasonal wind pattern in Florida 12 

2.3 Average monthly concentration of dissolved oxygen in upper Charlotte Harbor, site 

CH-006, 1976-84 19 

4.1 The connection between nitrogen, phosphorous and carbon cycle 44 

4.2 CBOD cycle and DO cycle 46 

4.3 Comparison of wind-dependent re-aeration formula 49 

4.4 The relationship between POM flux and SOD flux related in the oxidation and 

reduction of organic matter in sediment column 53 

4.5 Effect of dissolved oxygen on sediment consumption and SOD release 57 

4.6 The scatter plots for k d (PAR) during calibration period with best fit coefficients . . 68 

4.7 Systematic calibration procedure 80 

5.1 Boundary-fitted grid (92 x 129) used for numerical simulation for Charlotte Harbor 

estuarine system 84 

5.2 Bathymetry in the boundary-fitted grid for Charlotte Harbor estuarine system (92 x 

129) 85 

5.3 Tidal forcing and river discharges for 1986 simulations of Charlotte Harbor 

circulation 89 

5.4 Wind velocity for 1986 simulations for Charlotte Harbor circulation 90 

ix 



5.5 Tidal forcing and river discharges for 2000 simulations of Charlotte Harbor 

circulation 91 

5.6 Wind velocity for 2000 simulation of Charlotte Harbor circulation 92 

5.7 Air temperature for 2000 simulation of Charlotte Harbor circulation 93 

5.8 Locations of the available 1986 water level and discharge measurement stations of 

USGS 97 

5.9 Locations of available 1986 velocity and salinity measurement stations of USGS . 98 

5.10 Comparison between simulated and measured water level in July 1986 108 

5. 1 1 Comparison between simulated and measure spectra of water level in July 1986 1 10 

5.12 Comparison between simulated and measured current velocity in July 1986 ... 1 1 1 

5.13 Comparison between simulated and measured salinity in July 1986 115 

5.14 Typical flow pattern of Charlotte Harbor estuarine system during one tidal cycle for 

August 6, 1986 H6 

5.15 The 29-day residual flow and salinity for Charlotte Harbor estuarine system during 

July 2 to July 30, 1986 117 

5.16 Locations of the available 2000 water level and salinity measured stations at 

Caloosahatchee River operated by SFWMD 119 

5.17 The comparison of the coarse grid (71x92) and the fine grid (92x129) for Charlotte 

Harbor estuarine system 120 

5.18 A comparison between simulated and measured salinity at Shell Point using 

ultimate QUICKEST, QUICKEST, and upwind advection schemes 124 

5.19 A comparison between simulated and measured salinity at Fort Myers using 

ultimate QUICKEST, QUICKEST, and upwind advection schemes 125 

5.20 A comparison between simulated and measured salinity at BR31 using ultimate 

QUICKEST, QUICKEST, and upwind advection schemes 126 

5.21 Simulated longitudinal-vertical salinity along the Caloosahatchee River at slack- 

water before flood tide on September 7, 2000 127 

5.22 Comparison between simulated and measured water level in 2000 136 

x 



5.23 Comparison between simulated and measured salinity at S79 in 2000 137 

5.24 Comparison between simulated and measured salinity at BR31 in 2000 138 

5.25 Comparison between simulated and measured salinity at Fort Myers in 2000. . . 139 

5.26 Comparison between simulated and measured salinity at Shell Point in 2000. . . 140 

5.27 Comparison between simulated and measured salinity near Sanibel Causeway in 

2000 141 

5 28 Comparison between simulated and measured temperature at Fort Myers in 2000. 

142 

5.29 Typical flow pattern of San Carlos Bay during ebb and flood tide for August, 7 on 

2000 143 

5.30 One-year residual flow in San Carlos Bay in 2000 144 

5.31 One-year residual salinity distribution in San Carlos Bay in 2000 145 

5.32 The locations of Sanibel Causeway and IntraCoastal Waterway and stations for 

comparing the effect of hydrologic alterations 148 

5.33 The comparison of bathymetry and shoreline for each hydrologic alteration case 

scenarios which are Baseline, the absence of IntraCoastal Waterway, and the 
absence of causeway l 4y 

5.34 The comparisons of water level for three cases at three selected stations: ST05 (Pine 

Island Sound), ST08 (San Carlos Bay), and ST10 (Caloosahatchee River mouth). 
150 

5.35 The comparisons of surface and bottom salinity for three cases at three selected 

stations: ST05 (Pine Island Sound), ST08 (San Carlos Bay), and ST 10 
(Caloosahatchee River mouth) 1*1 

5.36 The comparisons of surface residual flow and salinity fields for three cases. ... 152 

5.37 The comparisons of bottom residual flow and salinity fields for three cases. ... 153 

5.38 The vertical-longitudinal salinity profiles along the axis of the Caloosahatchee River 

during wet season in 2000 l->9 



XI 



5.39 The vertical-longitudinal salinity profiles along the axis of the Caloosahatchee River 

during dry season in 2000 160 

5.40 Time histories of river discharge at S-79 and the locations of 1, 10, and 20 ppt 

surface salinity along the Caloosahatchee River during 2000 simulation period.161 

5.41 Time histories of river discharge at S-79 and the 1 ppt salinity location along 

Caloosahatchee River 162 

5.42 The 1-day averaged 20 ppt surface salinity location and 30-day averaged 10 ppt 

surface salinity location during 2000 simulation period 163 

5.43 The locations of 10 ppt surface salinity due to river discharge rate at S-79 during 

2000 baseline simulation 164 

5.44 The relationship between location of specific salinity value vs. river discharge at 

S-79 165 

5 45 The relationship between salinity at Fort Myers station vs. river discharge at S-79. 

166 

6.1 Tidal forcing and river discharges for 1996 simulations of Charlotte Harbor 169 

6.2 Wind velocity for 1996 simulations of Charlotte Harbor 170 

6.3 Air temperature for 1996 simulations of Charlotte Harbor 171 

6.4 Locations of 1996 water quality measurement stations operated by EPA 176 

6.5 Locations of 2000 water quality measurement stations operated by SFWMD and 

SWFWMD 177 

6.6 Light intensity at water surface for 1996 and 2000 simulations 178 

6.7 Segments for Charlotte Harbor estuarine system 179 

6.8 The scatter plots for water quality constituents during calibration period 189 

6.9 Temporal water quality variations at CH002 station in 1996 194 

6.10 Temporal water quality variations at CH004 station in 1996 195 

6.1 1 Temporal water quality variations at CH005 station in 1996 196 

6.12 Temporal water quality variations at CH006 station in 1996 197 

xii 



6.13 Temporal water quality variations at CH007 station in 1996 198 

6.14 Temporal water quality variations at CH09B station in 1996 199 

6.15 Temporal water quality variations at CH009 station in 1996 200 

6.16 Temporal water quality variations at CH010 station in 1996 201 

6.17 Temporal water quality variations at HB002 station in 1996 202 

6.18 Temporal water quality variations at HB006 station in 1996 203 

6.19 Temporal water quality variations at HB007 station in 1996 204 

6.20 Temporal water quality variations at CH013 station in 1996 205 

6.21 Simulated dissolved oxygen concentration in Charlotte Harbor estuarine system for 

August 21, 1996 206 

6.22 Simulated chlorophyll a concentration in Charlotte Harbor estuarine system for 

August 21, 1996 207 

6.23 Simulated dissolved ammonium nitrogen concentration in Charlotte Harbor 

estuarine system for August 21, 1996 208 

6.24 Simulated soluble organic nitrogen concentration in Charlotte Harbor estuarine 

system for August 21, 1996 209 

6.25 Simulated soluble reactive phosphorous concentration in Charlotte Harbor estuarine 

system for August 21, 1996 210 

6.26 Simulated soluble organic phosphorous concentration in Charlotte Harbor estuarine 

system for August 21, 1996 211 

6.27 The scatter plots for water quality constituents in 2000 213 

6.28 Temporal water quality variations at CH002 station in 2000 218 

6.29 Temporal water quality variations at CH004 station in 2000 219 

6.30 Temporal water quality variations at CH005 station in 2000 220 

6.31 Temporal water quality variations at CH006 station in 2000 221 



xm 



6.32 Temporal water quality variations at CH007 station in 2000 222 

6.33 Temporal water quality variations at CH09B station in 2000 223 

6.34 Temporal water quality variations at CH009 station in 2000 224 

6.35 Temporal water quality variations at CH010 station in 2000 225 

6.36 Temporal water quality variations at CES02 station in 2000 226 

6.37 Temporal water quality variations at CES03 station in 2000 227 

6.38 Temporal water quality variations at CES08 station in 2000 228 

6.39 Temporal water quality variations at CHOI 3 station in 2000 229 

6.40 The comparison between simulated dissolved oxygen concentration and the 

possible causes hypoxia: river discharge, salinity, temperature, and re-aeration and 
SOD fluxes at CH006 water quality measured station 230 

6.41 Simulated longitudinal-vertical salinity and dissolved oxygen concentration along 

the Peace River at 1 pm on June 18 (Julian Day 170), 2000 231 

6.42 Simulated longitudinal-vertical salinity and dissolved oxygen concentration along 

the Peace River at 1 pm on October 6 (Julian Day 280), 2000 232 

6.43 Simulated near-surface chlorophyll a concentration in Charlotte Harbor estuarine 

system for February 9, May 9, August 7, November 5, 2000 233 

6.44 Simulated near-bottom dissolved oxygen concentration in Charlotte Harbor 

estuarine system for February 9, May 9. August 7, November 5, 2000 234 

6.45 Comparisons of simulated surface chlorophyll a concentration for three cases at 

three selected stations: ST05 (Pine Island Sound), ST08 (San Carlos Bay), and 
ST10 (Caloosahatchee River mouth) 237 

6.46 Comparisons of simulated surface chlorophyll a concentration fields in San Carlos 

Bay after 90 days of simulation for three cases 238 

6.47 Comparisons of simulated surface dissolved ammonium nitrogen (NH4) 

concentration fields in San Carlos Bay after 90 days of simulation for three cases. 
239 

6.48 The water quality species at CH004 water quality measured station before and after 

100 % nitrogen load reduction from Peace River 247 

xiv 



6.49 The water quality species at CH006 water quality measured station before and after 

100 % nitrogen load reduction from Peace River 248 

6.50 The water quality species at CH004 water quality measured station before and after 

100 % phosphorous load reduction from Peace River 249 

6.51 The water quality species at CH006 water quality measured station before and after 

100 % phosphorous load reduction from Peace River 250 

6.52 The water quality species at CES02 water quality measured station before and after 

100 % nitrogen load reduction from Caloosahatchee River 251 

6.53 The water quality species at CES08 water quality measured station before and after 

100 % nitrogen load reduction from Caloosahatchee River 252 

6.54 The water quality species at CES02 water quality measured station before and after 

100 % phosphorous load reduction from Caloosahatchee River 253 

6.55 The water quality species at CES08 water quality measured station before and after 

100 % phosphorous load reduction from Caloosahatchee River 254 

6.56 Dissolved oxygen and Chlorophyll a concentrations at CH006 water quality 

measured station before and after 100 % organic matter load reduction from Peace 
River using DiToro's sediment flux model 255 

6.57 Dissolved oxygen concentrations at CH004 and CH006 water quality measured 

stations before and after 50 % SOD reduction and 75% SOD reduction 256 

6.58 The comparison of hypoxia area at Upper Charlotte Harbor according to varying 

SOD constant rate at 20°C 257 

A.l Flow chart for main CH3D program 264 

A.2 Flow chart for driving subroutine for the time stepping of the solution 266 

A. 3 Flow chart for initializing sediment transport model 271 

A.4 Flow chart for main sediment transport model 272 

A.5 Flow chart for initializing water quality model 277 

A.6 Flow chart for main water quality model 275 



xv 



A.7 Flow chart for computing temperature and light attenuation functions for water 

quality model 277 

D.l The vertical one-dimensional z-grid 292 

E. 1 Nitrogen cycle 296 

E.2 Phosphorous cycle 302 

G.l Parallel speedup gained in performing the simulation on the SGI Origin platform.312 



xvi 



Abstract of Dissertation Presented to the Graduate School 

of the University of Florida in Partial Fulfillment of the 

Requirements for the Degree of Doctor of Philosophy 

MODELING THE CIRCULATION AND WATER QUALITY 
IN CHARLOTTE HARBOR ESTUARINE SYSTEM, FLORIDA 

By 

Kijin Park 
August 2004 

Chair: Y. Peter Sheng 

Department: Civil and Coastal Engineering 

This study aims to develop an enhanced version of a three-dimensional curvilinear- 
grid modeling system, CH3D-IMS, which include a 3-D hydrodynamics model, a 3-D 
sediment transport model, and a 3-D water quality model, to simulate circulation and water 
quality of the Charlotte Harbor Estuarine System and to provide quantitative assessment of 
various management practices. 

In the past decade, the upper Charlotte Harbor system has been suffering summer 
hypoxia in bottom water. Field study indicated that hypoxia in the upper Charlotte Harbor 
is related to a strong stratification caused by high freshwater flows and dissolved oxygen 
fluxes at the air-sea and sediment-water interfaces. To simulate the hypoxia event, models 
of oxygen balance and oxygen fluxes at the air-sea and sediment-water interfaces in previous 
versions of CH3D-IMS are enhanced. The three dimensional temperature model and 
physics-based light model in CH3D-IMS are also enhanced to enable better understanding 



xvn 



of the temporal and spatial variations of temperature and light and their effect on water 
quality processes. 

The hydrodynamics component of the integrated model CH3D-IMS for Charlotte 
Harbor has been successfully calibrated, using hydrodynamic data gathered by the National 
Oceanic and Atmospheric Administration (NOAA) and the United States Geological Survey 
(USGS) in 1986 and 2000. This calibrated model was applied to assess the impact of the 
removal of the Sanibel Causeway and the IntraCoastal Waterway on the circulation in the 
San Carlos Bay area and to provide a quantitative evaluation of the minimum flow and level 
(MFL) for Caloosahatchee River. The results show that the hydrologic alterations would not 
noticeably affect the circulation, salinity, and water quality in San Carlos Bay except near the 
mouth of Caloosahatchee River. The minimum flow required to produce a salinity of no 
more than 10 ppt at Fort Myers is about 18 m 3 /s. 

The water quality model of the CH3D-MS was calibrated with the systematic 
calibration procedure and validated using hydrodynamic, sediment, and water quality data 
provided by the USGS and the United States Environmental Protection Agency (USEPA) 
in 1996, and by the USGS, the South Florida Water Management District (SFWMD) and the 
Southwest Florida Water Management District (SWFWMD) in 2000. This validated model 
was used to examine the temporal and spatial dynamics of factors which can affect hypoxia 
in upper Charlotte Harbor, such as freshwater inflow, tidal variation, sediment oxygen 
demand (SOD), water column oxygen consumption, and dissolved oxygen (DO) re-aeration. 
The model results suggest that hypoxia in the upper Charlotte Harbor System is primarily 
caused by a combination of vertical salinity stratification and SOD, while water quality also 
affected by the oxygen re-aeration and water column oxygen consumption. This model was 



xvin 



applied to assess the effects of hydrologic alterations and to provide a preliminary evaluation 
of pollutant load reduction goal (PLRG). Due to lack of detailed data and insufficient 
understanding on the causes of SOD and how SOD is related to the loading, the present 
model cannot simulate the complete effect of nutrient load reduction on the DO 
concentration in the estuary. Further research and more complete data are needed 

A systematic calibration procedure has been developed for a more efficient and more 
objective calibration of the water quality model. A consistent framework for systematic 
calibration is formulated, which include in the following steps: model parameterization, 
selection of calibration parameters, and formulation of calibration criteria. 



xix 



CHAPTER 1 
INTRODUCTION 

Charlotte Harbor (Figure 1 . 1 ) is a shallow estuarine system in southwest Florida. The 

estuary receives freshwater from the Caloosahatchee, Peace, and Myakka Rivers; is 

connected to the Gulf of Mexico through the Boca Grande Pass, Gasparilla Pass, Captiva 

Pass, Blind Pass, and San Carlos Bay; and provides water resources for several counties in 

Southwest Florida. Charlotte Harbor is dominated by rivers that flow into the coastal area. 

While most estuaries in Southwest Florida are influenced by the Gulf of Mexico, the 

characteristics of the Charlotte Harbor system are strongly influenced by large rivers such as 

the Caloosahatchee and Peace Rivers. Large fluctuations of river flow between wet and dry 

seasons strongly affect the salinity and water characteristics in Charlotte Harbor (Estevez, 

1998). 

Population growth and development in the surrounding areas during the past few 
decades have led to concerns over human impacts on the quality of the estuarine system. 
Industrial and agriculture development also increase environment pollution. Growth and 
development will cause an increased demand for fresh water and a corresponding increase 
in urban, agricultural, and industrial waste. The inflow of freshwater is essential to the 
integrity and health of the estuarine system. Increased freshwater withdrawal or diversion, 
or increased wastewater discharges to the rivers and streams that flow into the estuary will 
create environmental stress in the estuary (McPherson et al., 1996). 



1 



Myakka River 




%> Deer Prairie Creek 

\ Dirt Cl^i tr 



Big Slough Canal 



h& J Sarasota Co. 

Nortlmortf*Z- •--- 

■y Char 



Peace River 



!\ V El Jpbean 



i j« De_SotoC_o. 

clia7iotteCo.*W'~ 

-. 11 



\^Gaspar1lla\ "«*» i 
w^ Sound ~ Charlotte jr 

Gasparilla \h£ ,.) M Harbor It- 
Pass -^ »i w /''fi £ 





V^Lower X 

- rravfa Charlotte HarborV / 
Boca Grande^ Bokeelia y. 

Pass , -. ^^M^f% 

CayoCostit.': S ^ 



S79 
Caloosahatchee River. fJmQj <T'" 



Charlotte Co. 
Lee Co. 



'u^ & ^ -\ Matlacha Pass J- * ~ X 
Captiva ' Pin _,-^i \ i*„ ■; --. Fort Myers 

Pjjqq ^- I II IC -i .- j, ■■■■■ i, < _, 

rass Island .. v &M r„„„ tJP ;' 



■Island .\ \j?f>t Cape jgP / 
v Sound i # a* C'or«/. >f ,> 
RedfishV -T -.v aia^.^vv 






'-■ : , Sound-'^, %"' £* Cora/ ^"|f ,> 
^ Cf V ':\-. StlJ^ /^Whiskey Creek 

f , PineSf 1^-^r^^ 
Captiva ,-, Islanq Ajlffr ._ *p iVt r.. _ . f 



Pass 



Blind 
Pass- 



o 



$iell PoiiU j 

-■ .- „-* 

Sanibel/ ./'' Causeway '%r^t 

Island San Carlos Fort M s^fc ^ 
Bay &at ./i <$ j 






>P- •■'■>-> M 

Estero Bay *5r Imperial River ■ * 

: %"~ A 



EXPLANATION 

— — Subarea Divide 



1 
» Naples 



Kilometers 



L b ^ ^ e 

ho 10 



20 



_^__ i . 

Figure 1.1 Charlotte Harbor estuarine system and its subarea boundaries 



Water quality in the southern Charlotte Harbor estuarine system (the Caloosahatchee 
Estuary, San Carlos Bay, and Pine Island Sound) appear to be influenced by freshwater 
discharge from the Franklin Lock and Dam, also known as United States Army Corps of 
Engineers Structure Number 79 (S-79). Waste load allocation studies conducted by the 
Florida Department of Environmental Regulation (Degrove, 198 1 ; Degrove, and Nearhoof, 
1987; Baker, 1990) concluded that the Caloosahatchee Estuary had reached its nutrient 
loading limits as indicated by elevated chlorophyll-a and depressed dissolved oxygen (DO) 
levels. Similarly, McPherson and Miller (1990) concluded that increased nitrogen loading 
would result in undesirable increases in phytoplankton and benthic algae. In order to 
effectively manage the loading of pollutants such as nutrients in a shallow estuary such as 
Charlotte Harbor, and control eutrophication, it is necessary to have a quantitative 
understanding of the transport and transformation processes of nutrients in system. 

Increased nutrient loading from tributaries, the atmosphere, and bottom sediments 
have been known to cause eutrophication in estuaries. Eutrophication varies both spatially 
and temporally. Eutrophication does not necessarily occur in regions with high nutrient 
loading because the circulation and sediment dynamics affect nutrient fate and transport. In 
the Upper Charlotte Harbor where a pycnocline develops frequently during high river flow 
conditions, hypoxia usually develops within two days of the formation of the pycnocline, 
when coupled with significant sediment oxygen demand (SOD) (CDM, 1998). In fact, 
hypoxia can occasionally extend all the way to the Boca Grande area. This strong linkage 
between hydrodynamics and water quality dynamics suggests the importance of having a 
comprehensive understanding of hydrodynamics. 

Nutrient dynamics in estuaries are not only determined by biological and chemical 



processes, but is also strongly affected by weather and climate (wind, air temperature, etc.), 
hydrodynamics (wave, tide, current, turbulence mixing, etc.), and sediment transport 
processes (resuspension, deposition, flocculation, etc.). In this study, integrated modeling 
of the system is conducted to understand the complex water quality processes, which are 
strongly linked to hydrodynamics and sediment processes. The primary objective of this 
study is to use models and field data to produce a detailed characterization of 
hydrodynamics, sediment, and water quality dynamics within the Charlotte Harbor estuarine 
system. 

The Chesapeake Bay model (Cerco and Cole, 1994) and Moreton Bay model 
(McEwan and Garbric, 1998) are coupled hydrodynamics-water quality models. These 
coupled models are more suitable for managing nutrient loads and predicting eutrophication- 
related problems than uncoupled models. Because these models were not coupled with a 
dynamic model for sediment transport, however, they could not accurately consider 
sediment-process effects such as resuspension, deposition, flocculation, and settling on 
nutrient dynamics in estuaries. Therefore, these loosely coupled models cannot account for 
nutrient release by sediments in episodic events (Chen and Sheng, 1994) 

Chen and Sheng (1994) developed a coupled hydrodynamic-sediment-water quality 
model and applied it to Lake Okeechobee. Their 3-D model includes a hydrodynamic model, 
a sediment transport model, and a water quality model with a nitrogen and a phosphorous 
cycle on a rectangular grid. Measured nitrogen and phosphorous dynamics during episodic 
events in 1993 and monthly sampling events in 1989 were accurately simulated with 
absorption-desorption reactions and the exchange of nutrients between the sediment and 
water column. 



While Chen and Sheng (1994) developed a rectangular grid model, Yassuda and 
Sheng (1996) developed an integrated model in curvilinear grids and applied it to Tampa 
Bay. Their water quality model, based on the Chen and Sheng (1994) model, includes a 
nitrogen cycle and an oxygen cycle, but not a phosphorous cycle. Zooplankton distribution 
is allowed to influence the phytoplankton dynamics. Their integrated model also includes 
a wave model, a light model, and a seagrass model. Although Yassuda and Sheng (1996) 
simulated the observed hypoxia in Tampa Bay during 1991, their model did not include 
dynamic fluxes of oxygen at the air-sea interface (re-aeration) and the sediment-water 
interface (sediment oxygen demand). Their model could not simulate the daily fluctuation 
and vertical stratification of DO. Although the phytoplankton growth rate in their model is 
controlled by temperature and light, it is difficult to reproduce the vertical distribution of 
phytoplankton because the temperature is not simulated. The light model in Yassuda and 
Sheng (1996) has not been sufficiently validated with Tampa Bay data, thus limiting the 
predictability of their model for vertical distribution of phytoplankton and dissolved oxygen. 
Their model could not produce the low dissolved oxygen phenomena in upper Charlotte 
Harbor because there are no dissolved oxygen fluxes at the air-water and sediment-water 
interfaces, such as reaeration and sediment oxygen demand (SOD). 

Sheng (2000) developed the framework for an integrated modeling system: CH3D- 
EVIS (http://ch3d.coastal.ufl.edu). The integrated modeling system (Table 1.1) is based on 
the curvilinear-grid hydrodynamic model CH3D (Sheng, 1987, 1989; Sheng et al., 2002) and 
also includes a wave model, a sediment transport model (Sheng et al., 2002a), a water quality 
model, (Sheng et al., 2001b), a light attenuation model (Sheng et al., 2002c, Christian and 
Sheng, 2003), and a seagrass model (Sheng et al., 2002d). Sheng et al. (2002) applied the 



CH3D-IMS to the Indian River Lagoon (IRL), Florida. The water quality model (Sheng et 
al., 2001b) includes a nitrogen cycle, a phosphorous cycle, a phytoplankton cycle, and a DO 
cycle. The CH3D-IMS was validated with comprehensive data collected from the IRL. 
Table 1.1 Component Models of the CH3D-IMS 



Component Model Model Name 



Hydrodynamic Model CH3D 

Flow Model 
Salinity Transport Model 

Wave Model SMB 

Sediment Transport Model CH3D-SED3D 

Water Quality Model CH3D-WQ3D 

Dissolved Oxygen Model 

Phytoplankton and zooplankton 
Model 

Nitrogen Model 

Phosphorous Model 

Light Attenuation Model CH3D-LA 

Seagrass Model CH3D-SAV 



The CH3D-IMS was able to successfully simulate the observed circulation, wave, 
sediment transport, water quality, light attenuation, and seagrass biomass in IRL during 1998. 
However, several aspects of the model were identified for further improvement. For 
example, although a temperature model was developed during the later phase of the IRL 
study, it was not sufficiently validated with data. The water quality model, while capable of 
simulating the annual variation in DO, could not simulate the diurnal DO variation. 

In this study, the CH3D-IMS is enhanced and used to study the circulation and water 
quality dynamics in the Charlotte Harbor estuarine system. The validated model will be used 
to determine minimum flow and level and pollutant load reduction goal by management 
agencies. Specially, the CH3D-IMS used for this study includes the simulation of 



temperature in addition to the flow and salinity simulations. An air-sea heat flux model is 
implemented along with the temperature equation. The water quality model is enhanced to 
include a re-aeration model and a sediment oxygen demand (SOD) model, such that hypoxia 
and daily fluctuation and vertical stratification of DO can be simulated. Moreover, the same 
basic light attenuation model developed for the IRL is used for this study. 

Although the factors that cause hypoxia have not been positively identified, it is 
generally accepted that hypoxia is related to strong stratification caused by high freshwater 
inflow. The modified CH3D-IMS is used in this study to examine the major factors of the 
observed hypoxia in 2000, including freshwater inflow, tidal variation, SOD, water column 
oxygen consumption, and DO reaeration. 
This study includes the following objectives: 

• To validate the CH3D-IMS with extensive hydrodynamic and water quality data from 
Charlotte Harbor, 

To improve the simulation of dissolved oxygen processes, including the observed 
daily dissolved oxygen fluctuation, and hypoxia in bottom water during high flow 
event with vertical stratification, 

• To validate the ability of temperature model and heat flux model to simulate seasonal 
and spatial temperature distribution in Charlotte Harbor, 

• To apply the physics-based light attenuation model to improve Charlotte Harbor 
simulations, and 

To demonstrate the feasibility of the validated modeling system for simulating the 
effects of various anthropogenic impacts on the Charlotte Harbor estuarine system. 

More specifically, the objectives of my research are to: 

• Reproduce observed circulation and transport dynamics in 1986 and 2000, 

• Reproduce observed water quality dynamics in 1996 and 2000, 

• Investigate the effect of the Sanibel causeway islands and IntraCoastal Waterway on 
the flow, salinity, and water quality distribution in San Carlos Bay and Pine Island 



Sound, 

Reproduce the main factors for bottom water hypoxia and habitat loss in Charlotte 
Harbor estuarine system, 

• Provide a preliminary determination of pollutant load reduction goal (PLRG) and 
minimum flow and levels (MFL), 

Perform sensitivity tests to quantify the influence of each model parameter and to 
determine the most sensitive parameters and most plausible parameter set, and 

• Establish a systematic calibration procedure. 

To achieve these goals, the characteristics of Charlotte Harbor estuarine system are 
presented in Chapter 2. Chapter 3 discusses the general hydrodynamic circulation and 
transport of temperature, salinity, and sediment in the estuarine system. This chapter will 
include initial and boundary conditions for the hydrodynamic and transport models. As an 
air/sea interface boundary condition, a heat flux model is also presented in this chapter. 
Chapter 4 describes the water quality processes and models of phytoplankton, zooplankton, 
nitrogen, phosphorous, dissolved oxygen, and light attenuation processes in estuarine system. 
The effects of temperature and light intensity, numerical solution technique, and model 
parameters and calibration procedure are also discussed in this chapter. Chapter 5 presents 
hydrodynamic field data and model simulations of hydrodynamics in Charlotte Harbor 
estuarine system. These model simulations include model calibration and verification of 
1986 and 2000 data, simulation of the effect of causeway islands and navigation channel, and 
determination of minimum flow and levels (MFL). Chapter 6 presents water quality and 
sediment field data and model simulations of circulation, sediment transport and water 
quality processes in the study area. These model simulations include calibration and 
verification of 1996 and 2000 data and determination of pollutant load reduction goal 
(PLRG). Discussions and conclusions are presented in Chapter 7. 



CHAPTER 2 
CHARLOTTE HARBOR CHARACTERIZATION 

Charlotte Harbor, a coastal-plain estuarine system is one of the largest estuarine 
systems on the southwest florida coast and is an important part of the Gulf of Mexico 
watershed. It is located on the southwest corner of the Florida peninsula, between 26° 20' 
and 27° 10'N and 81° 40' and 82° 30' W. As shown in Figure 1.1, the Charlotte Harbor 
estuarine system is sub-divided into Upper Charlotte Harbor, Lower Charlotte Harbor, Pine 
Island Sound, Matlacha Pass, San Carlos Bay, Gasparilla Bay, Peace River, Myakka River 
and Caloosahatchee River. The drainage area (Figure 2.1) consists of the Peace, Myakka, 
and Caloosahatchee River Watersheds and the coastal area and islands that drain directly into 
the harbor. The estuary has a surface area of 648 km 2 , a drainage area of a more than 10000 
km 2 , and a average depth of 2. 1 m. The upper harbor has an average depth of 2.6 m, and the 
lower harbor has an average depth of 1.6 m (Stoker, 1986). The estuary is separated from 
the Gulf of Mexico by barrier islands and is connected to the Gulf through two major inlets: 
Boca Grande and San Carlos; and through several smaller passes: Gasparilla, Captiva, 
Redfish (McPherson et al., 1996). 

In 1995, the Charlotte Harbor National Estuary Program (CHNEP) was established 
by U.S. Environmental Protection Agency (USEPA) and the State of Florida. In its planning 
documents, CHNEP identified three draft major problems: (1) hydrologic alterations, (2) 
nutrient enrichment, and (3) habitat loss. In order to address these problems, it is useful to 



10 
have a comprehensive understanding of the climate, hydrodynamics, sediment properties, and 
biological and chemical characteristics of the system. 



27°45' 



27°00' - 




Figure 2.1 Drainage basins of Charlotte Harbor estuarine system (Stoker, 1992) 



11 

2.1 Climate 

The climate of the Charlotte Harbor estuarine system is subtropical and humid. The 
mean annually average temperature is 72 °F, with low mean of 60°F in December and 
January and a high mean of 80 °F during the summer (McPherson et al., 1996). 

Annual rainfall averages 132 cm, of which more than half occurs from June through 
September, during local thundershowers and squalls. Rain during the fall, winter, and spring 
seasons is usually the result of large frontal systems and tends to be more broadly distributed 
than rain associated with local thunderstorms and squalls. The period from October through 
February is characteristically dry, with November usually being the driest month. The 
months of April and May also are characteristically dry. Low rainfall in April and May 
coincides with high evaporation and generally results in the lowest streamflow, lake stage, 
and ground-water levels of the year (Hammett, 1990). 

The annual average wind speed is 3.9 m/s from the east. Four typical seasonal wind- 
field patterns are shown in Figure 2.2. In the Winter months, the easterly trade winds 
dominate the region south of latitude 27° N, while the westerlies dominate the area north of 
latitude 29°N. Spring and Summer generally exhibit more southerly winds, and Fall is 
characterized by easterly or northeasterly winds. Wind speed can exceed 10 m/s during the 
passage of Winter storms or during Summer squalls, hurricanes and tornadoes (Wolfe and 
Drew, 1990). 

2.2 Hydrodynamics 

The hydrodynamic behavior of the Charlotte Harbor estuarine system is affected by 
physical characteristics (bathymetry and geometry) as well as weather, climate, and 
oceanographic and hydrologic characteristics, such as wind, atmosphere heating and cooling, 



12 
evaporation and precipitation, tidal-stage oscillations, discharge through tidal inlets, tidal 
velocity, and freshwater inflow. 




Figure 2.2 Seasonal wind pattern in Florida (Echternacht, 1975) 



13 
2.2.1 Tidal Stage, Discharge at Inlet, and Tidal Circulation 

Tides along the Gulf coast of West-Central Florida in the vicinity of Charlotte Harbor 
have a range of 30 to 140 cm and are of the mixed type with both diurnal and semidiurnal 
characteristics (Goodwin and Michaelis, 1976). Spring tides, which have the largest range, 
sometimes have only a diurnal fluctuation, whereas neap tides, which have the smallest 
range, approach semidiurnal conditions of two nearly equal high and low water levels per 
day. 

The boundary between the Charlotte Harbor estuarine system and the Gulf of Mexico 
extends about 40 mi from Gasparilla Pass on the north to San Carlos Pass on the south. 
Tidal characteristics in the Gulf of Mexico are nearly uniform from Gasparilla Island to the 
western face of Sanibel Island, but are of a larger range off the southern shore of Sanibel 
Island (Goodwin, 1996). 

Circulation in the system is primarily driven by Gulf tides, entering the system 
through San Carlos Pass, Boca Grande Pass, Captita Pass, Redfish Pass, Gasparilla Pass and 
Blind Pass. San Carlos Pass has a maximum depth of 5.3 m and a width of 3.3 km; Boca 
Grande Pass has a maximum depth of a 18.3 m and a width of 1.28 km. The discharge 
through Boca Grande Pass is about twice the discharge through San Carlos Bay and three to 
four times the discharge through Captiva and Redfish Passes (Goodwin, 1996). The 
geometric narrowing that occurs at passes focuses tidal energy, resulting in high velocities 
associated with the large volume of water moving through the pass. This tidal energy is 
dispersed inside the harbor and influences the harbor and tidal rivers as much as 40 to 44 km 
upstream from the Peace River mouth. There is about a 2-hour lag between tide phases at 
Boca Grande and tide phase in the upper harbor near the mouth of the Peace River (Stoker, 



14 
1992). Tide and wind keep the water column well mixed in the southern part of the estuary. 
In the northern part of the estuary, vertical stratification can develop during moderate to high 
fresh water inflows and can persist for weeks after a high freshwater inflow event (Sheng, 
1998). 

2.2.2 Freshwater Inflow 

The majority of the freshwater that enters Charlotte Harbor come from the Myakka 
River, Peace River and Caloosahatchee River. Average flows are 17.8 m 3 /s, 56.9m 3 /s and 
56.7 m 3 /s, respectively. Flows in the Myakka and Peace Rivers are largely unregulated, 
while flow in the Caloosahatchee is controlled by operation of the Franklin Lock about 43 
km upstream from the mouth. Discharge in the Peace and Myakka Rivers tend to peak in 
August and September when rainfall totals are generally the greatest. Discharges are usually 
lowest in April and May (Stoker, 1992). The Caloosahatchee River discharge does not 
always correspond to rainfall patterns in the basin, since it is controlled by S-79. 

Analyses of long-term streamflow trends in the Charlotte Harbor area have indicated 
statistically significant decreases in streamflow at several gages in the Peace River basin 
from 1931 to 1984 (Hammett, 1990). The long-term decrease in streamflow of the Peace 
River is probably related to the increased use of ground water and subsequent decline of the 
potentiometric surface of the upper Florida aquafer (Hammett, 1990). A sustained 
significant reduction in streamflow could result in an increase of salinity in Upper Charlotte 
Harbor, possibly approaching the Gulf of Mexico salinity (Sheng, 1998). The impact of 
freshwater reduction is of major ecological and economic significance. 

2.2.3 Salinity 

As in any other typical estuarine system, Charlotte Harbor generally exhibits 



15 
significant horizontal gradients in salinity. The higher salinity values in the adjacent Gulf 
of Mexico fluctuate around 36 ppt, whereas the lower salinity levels nearly zero in wet 
season occur near the mouth of creeks and rivers. Seasonal changes in salinity occur 
primarily in response to changes in freshwater inflow from the Peace, Myakka, and 
Caloosahatchee River basins. Other sources of freshwater, including direct rainfall, runoff 
form coastal areas, ground-water seepage, and domestic influence, have smaller and usually 
more local effects on salinity in the estuary (McPherson et al., 1996). 

Stoker (1992) described salinity characteristics in the system based on data collected 
from June 1982 to May 1987. Salinity generally is the lowest during the wet season between 
July and September, and is the highest during the dry season from January through March. 
Salinity also varies daily in response to tidal fluctuation. Peak salinity is near the flood-tide 
stage, and lowest salinity is near the ebb-tide stage. 

Due to the shallow depth and because of significant vertical mixing, salinity is 
generally not stratifed in the southern part of the estuary. Vertical salinity stratification in 
the Upper Charlotte Harbor is a common seasonal occurrence (Environmental Quality 
Laboratory, Inc., 1979). In high river inflow events, a stable vertical salinity gradient is 
created which suppresses vertical mixing unless there are sufficient mixing by wind or tide. 

2.3 Water Quality 

Water quality refers to the condition of water (e.g., dissolved oxygen concentration, 
chlorophyll concentration, light attenuation, etc.) relative to legal standard, social 
expectations or ecological health. Overall, water quality in the Charlotte Harbor estuarine 
system is fair or good, but some areas have poor water quality or declining trends (Estevez, 
1998). The water quality of an estuary is strongly influenced by hydrodynamic processes 



16 

(e.g., circulation and flushing), and chemical and biological processes (e.g., ammonification, 
mineralization, decomposition, algae uptake, excretion, and mortality, etc.) in the estuary and 

basin. 

Color levels and concentrations of nitrogen, phosphorous and chlorophyll-a in 
Charlotte Harbor exhibit pronounced salinity-related gradients which extend from the head 
of the estuary to its mouth (Morrison, 1997). The water quality in the Caloosahatchee system 
is more degraded than the water quality in the Myakka or Peace systems. Oxygen depletion 
is common upstream of Franklin Lock. Nutrient and chlorophyll levels are high, and algal 
blooms occur regularly in the tidal river (Estevez, 1998). 
2.3.1 Nutrients 

Nutrient availability is a key factor in the regulation of primary productivity in 
estuarine and coastal water (Ketchum, 1967). Increased nutrient loads related to the urban 
development of coastal basins have been implicated in estuarine nutrient enrichment, 
increased phytoplankton productivity, and incresed phytoplankton biomass (Jaworski, 1981). 
and in declines of seagrass communities (Orth and Moore, 1983). 

The distribution of nutrients in the system is mainly the result of nutrient input from 
rivers, freshwater and tidal flushing, and recycling processes in the estuary (McPherson and 
Miller, 1990). The major factor that influences estuarine nutrient distribution is freshwater 
inflow from rivers, which contributes substantial nutrient loads and flushes nutrients 
seaward. NOAA estimates that Charlotte Harbor receives about 2,500 tons of nitrogen as 
total Kjeldahl nitrogen, or TKN, and 1,000 tons of phosphorous per year. Relative to its 
dimensions and flushing characteristics, phosphorous loads are high, signifying a nitrogen 
limited system (Estevez, 1986). 



17 
Molar ratios of dissolved inorganic nitrogen to dissolved inorganic phosphorous were 
below the Redfield ratio of 16, averaging 5.7 at Caloosahatchee River during two sampling 
periods, which are from 1985 to 1989 and from 1994 to 1995 (Doering and Chamberlain, 
1997). By contrast, total nitrogen to total phosphorous ratios were above 16, averaging 35 
in Caloosahatchee River. With these results, dissolved inorganic N:P ratios (<16) suggest 
that nitrogen could limit phytoplankton productivity in agreement with McPherson and 
Miller (1990), while total N:P ration (>16) indicated that phosphorous could limit 
productivity (Doering, 1996). In fact, nutrient addition studies conducted by FDER found 
N and P to be co-limiting (Degrove, 1981) 

The distributions of phosphorous in the system were nearly conservative and a 
function of river phosphorous concentration, flow, and physical mixing (McPherson and 
Millor, 1990). A large amount of phosphorous from the watershed is carried by freshwater 
discharge into the tidal reaches of the Peace River, with subsequent dilution by the lower- 
nutrient seawater entering the estuary from the Gulf of Mexico. Concentrations of total 
phosphorous averaged about 0.08 mg/L in Pine Island Sound, 0.15 mg/L in the tidal 
Caloosahatchee River, and 0.62 mg/L in the tidal Peace River according to USGS nutrient 
data from 1982 to 1989 (McPherson and Miller, 1990). 

Most of the nitrogen in the rivers and estuary is organic nitrogen (McPherson and 
Miller, 1990). Organic nitrogen concentrations decreased over the salinity gradient, 
indicating river input as a source. The relatively low concentrations of inorganic nitrogen 
could limit plant growth in the estuary (McPherson and Miller, 1990). Ratios of nitrogen and 
phosphorous constituents at the Peace River, Myakka River, and Upper Charlotte Harbor 
determined by USGS nutrient data during 1975 through 1990 are shown in Table 2.1 (Pribble 



IS 



etal., 1998) 

Table 2.1 Ratios of nitrogen and phosphorous constituents (Pribble et al ., 1998) 
RATIO Peace River Myakka River Upper Harbor 



NH3 : TN 


0.0327 


0.0349 


0.0338 


N03 : TN 


0.4378 


0.0505 


0.2442 


ON:TN 


0.5614 


0.9096 


0.7355 


P04 : TP 


0.9384 


0.9371 


0.9361 


OP:TP 


0.0616 


0.0683 


0.0650 



Concentrations of ammonia were highly variable along the salinity gradient and were 
in about the same range as concentrations in the rivers (McPherson and Miller, 1990). 
Ammonia concentrations increased in the deeper water of Charlotte Harbor during summer 
(Fraser, 1986). Ammonia enrichment probably was related to density stratification and to 
low concentrations of dissolved oxygen in bottom waters (McPherson et al, 1996). 
Concentrations of nitrate and nitrite nitrogen were nonconservative and decreased sharply 
along the salinity gradient (McPherson and Miller, 1990). The sharp decline of the nitrate 
and nitrite nitrogen in the low salinity regions indicates a substantial removal of nitrogen 
from the water column due to biological uptake. 
2.3.2 Dissolved Oxygen 

Dissolved oxygen is critical for survival of plants and animals in fresh and salt water, 
and a major constituent of interest in water quality study. Dissolved oxygen concentrations 
in the near surface water of the system ranged from about to 6 to 8 mg/L during daylight 
samplingin 1982 -84 (Stoker, 1986). Dissolved oxygen concentrations of near bottom water 
of the estuary generally are lower than near surface concentrations. 

Bottom water hypoxia (dissolved oxygen <2 mg/L) in Charlotte Harbor has been 
reported periodically by the Environmental Quality Laboratory (EQL) since 1975 (Heyl. 



19 

1996). Dissolved oxygen concentrations were measured monthly from 1976 to 1984 in the 
river mouth of Peace River near PuntaGorda (CH-006) (Fraser, 1986). The average monthly 
near-surface concentrations declined from 8.5 to 6.7 mg/L from January to July and then 
began to rise (Figure 2.3). Near-bottom average monthly concentrations at this area were 
highest in February, declined slowly through May, and then declined more rapidly until July 
(Fraser, 1986). The hypoxia conditions during summer are attributed to strong stratification, 
which cause restricted reaeration, and to SOD. After breakup of the stratification, the 
concentration increased from October to December. 




F M 



M J 



A S O N 



Figure 2.3 Average monthly concentration of dissolved oxygen in upper Charlotte 
Harbor, site CH-006, 1976-84 (Fraser, 1986) 

2.3.3 Phytoplankton 

Phytoplankton is an important component of water quality processes and a major 
primary producer in coastal and estuarine waters. The temporal and spatial variability of 



20 
phytoplankton productivity and biomass in Charlotte Harbor have been investigated by 
Environmental Quality Laboratory, Inc (EQL) (1987). Phytoplankton productivity and 
biomass (as chlorophyll _a) in the system are relatively low most of time. Productivity 
ranged from 5 to 343 (mgC/m 3 )/h and averaged 59 (mgC/m 3 )/h from 1985 to 1986 
(McPherson et al., 1990). Chlorophylljx concentrations ranged from 1 to 46 mg/m 3 and 
averaged 8.5 mg/m 3 . Both productivity and biomass were greater during summer near the 
mouth of tidal rivers which has middle range salinity of 6 to 12 ppt (McPherson et al., 1990). 
Phytoplankton productivity and biomass in the system are affected by freshwater 
inflow that lowers salinity, increases nutrient availability, and reduces light penetration in the 
water column. The nutrient rich colored water is diluted by seawater at middle range salinity 
of 10 to 20 ppt, so that availability of light increases and sufficient nutrient concentrations 
remain available from runoff, to stimulate productivity and growth of phytoplankton in these 
areas (McPherson et al., 1990). Of the major nutrients for phytoplankton productivity, 
inorganic nitrogen is in lowest supply and most critical in limiting phytoplankton 
productivity and growth in the system (Fraser and Wilcox, 1981). 

The composition of the phytoplankton in the system varied with location and season 
(McPherson et al., 1990). Diatoms were dominant in 55 % of 289 phytoplankton samples 
collected in the system between 1983 and 1984, cryptophytes in 35 %, cyanophytes (blue- 
green algae) in about 6 %, dinoflagellates in about 4 %, and other classes in 1 % (McPherson 
et al., 1996). 

2.4 Sediment 

Sediment quality data from Charlotte Harbor are scarce. Organic carbon, nitrogen 
and phosphorous data exist in only two studies. Hwang (1966) obtained data in 1965 from 



21 
119 stations throughout the study area. The FDEP Coastal Sediment Contaminant Survey 
obtained carbon, nitrogen phosphorous, and metal data in the sediment column from 33 
stations during 1985 through 1989, in all areas except the Gulf of Mexico (Sloane, 1994). 
Mean carbon/nitrogen ratios in the six different areas of Charlotte Harbor ranged from 8.9 
at Lower Charlotte Harbor to 16.4 at Gasparilla Sound (Schropp, 1998). Despite data 
limitation, the available data indicate that Charlotte Harbor is relatively free of sediment 
contamination in comparison with Tampa Bay and Biscayne Bay, where sediment 
contaminations are widespread and are present in higher concentrations (Schropp, 1998). 

2.5 Light Environment 

The amount of photosynthetically active radiation (PAR) in natural water is 
fundamentally important in determining the growth and vigor of aquatic plants (McPherson 
et al., 1996). Light is reflected, absorbed, and refracted by dissolved and suspended 
substances in the water column, and by water itself. The controlling factors are chlorophyll- 
a, dissolved substances, and non-algal particulate matter (Christian and Sheng, 2003). 

Dissolved and suspended matter are the major causes of light attenuation in the 
system: phytoplankton and chlorophyll-a are generally minor causes of attenuation 
(McPherson and Miller, 1990). On average, non-chlorophyll suspended matter accounted 
for 72% of light attenuation, dissolved organic matter (color) accounted for 21%, 
phytoplankton chlorophyll for 4%, and water for the remaining 3% (McPherson et al., 1996). 
Water color can cause light attenuation at tidal rivers because the source of the water color 
is terrestrial dissolved organic matters. Dissolved organic matter has little effect on light 
attenuation in much of the southern part of the estuary, where suspended matter is the major 
cause of light attenuation (McPherson and Miller, 1996). The source of suspended matter 



22 

includes the bottom of the estuary, which consisted of very fine to fine sand (Hwang, 1966) 
and organic detrital -material (McPherson and Miller, 1990), and the major rivers. 



CHAPTER 3 
HYDRODYNAMICS AND SEDIMENT TRANSPORT MODEL 

Biological processes in an estuarine system are strongly affected by hydrodynamics 
and sediment transport processes. Therefore, as the first step to study nutrient cycling in an 
estuary, hydrodynamics and sediment transport processes must be investigated and 
understood. 

The three-dimensional hydrodynamics model CH3D (Sheng et al., 2001) and the 
associated sediment transport model CH3D-SED3D (Sheng et al., 2000a) are used for 
numerical simulations in this study. The model framework has been improved and modified 
from earlier versions in order to develop an integrated model that couples hydrodynamics, 
sediment and water quality dynamics. The code of the enhanced integrated model is 
optimized to achieve more efficient coupling and more systematic structure. The detail 
structure of enhanced model is explained in Appendix A as flow chart. The application of 
the circulation and transport model to produce a detailed characterization of hydrodynamics 
within system is the first step in the development of the integrated model of the system. 

The CH3D model for Tampa Bay and Indian River Lagoon did not include the 
simulation of temperature distribution. In this study, the temperature field is simulated by 
solving the temperature equation with an air-sea interface boundary condition. Temperature 
is an important factor for baroclinic circulation and water quality processes, in which almost 
all reaction parameters in the nutrient cycle are a functions of temperature. Therefore, 



23 



24 
improved simulation of temperature is expected to produce more accurate baroclinc and 
water quality simulations. 

3.1 Governing Equation 

Hydrodynamics and sediment transport processes in estuaries are complicated, three- 
dimensional and time-dependent. Mathematical descriptions of these processes generally 
require simplifying assumptions. 
3.1.1 Hydrodynamic Model 

The governing equations that describe the velocity and surface elevation fields in 
shallow water are derived from the Navier-Stokes equations. In general, four simplifying 
approximations are applied: 1) the flow is incompressible, which results in a simplified 
continuity equation, 2) the horizontal scale is much larger than the vertical scale such that 
the hydrostatic pressure distribution is valid, 3) the Boussinesq approximation can be used 
to simplify the treatment of the baroclinic terms, 4) the eddy-viscosity concept, which 
assumes that the turbulent Reynolds stresses are the product of mean velocity gradients and 
"eddy viscosity", can be employed. With the above assumptions, the continuity equation and 
x- and y- momentum equations have the following form (Sheng, 1983): 

du dv dw _ 

— + — + — = (3.1) 

dx dy oz 



+ 
dz 



du duu duv duw _ dC, 

dt dx dy dz dx H \dx 2 dy 

dv dvu dvv dvw dC, ( d 2 v 3 2 v| df A oV 



f d 2 u aV 

J 



d ( du 



^ 



(3.2) 



UVU UVV UVYV l/L, V V V r , w \ A v /TT\ 



dt dx dy dz ' dy {dx 2 dy 



J 



dz\ dz 



where u(x,y,z,t), v(x,y,z,t), w(x,y,z,t) are the velocity components in the 
horizontal x- and y-directions, and vertical z-direction; t is time; g(x, y, t) is the free surface 



25 



elevation; g is the gravitational acceleration; and A H and \ are the horizontal and vertical 

turbulent eddy coefficients, respectively. 

In Cartesian coordinates, the conservation of salt and temperature can be written as 



dS duS dvS dwS d 

— + + + = — 

dt dx dy dz dx 
dT duT dvT dwT d 



D H — 

ox 



d' 

+ — 
dy 



dy 



dz 



+• 



+ ■ 



■ + ■ 



dt dx dy dz 



K t 



dT_ 
dx 



+ ■ 



< dT 



dy[ dy 



+ ■ 



dz 



Dv Tz 

v V dz 



(3.4) 



(3.5) 



where S is salinity; T is temperature; D H and K H are the horizontal turbulent eddy 
diffusivity coefficients for salinity and temperature, respectively; and D v and K v are the 

vertical turbulent eddy diffusivity coefficients for salinity and temperature, respectively. 

Since the length scales of horizontal motion in estuarine systems are much greater 
than those of vertical motion, it is common to treat vertical turbulence and horizontal 
turbulence separately. In shallow estuaries, the effect of the horizontal eddy viscosities on 
circulation are much smaller than the effect of the vertical eddy viscosity, although the 
horizontal eddy viscosity is typically 2-3 orders of magnitude larger than the vertical eddy 
viscosity. 

Vertical turbulent mixing is an important process, which can significantly affect 
circulation and transport in an estuary. Since turbulence is a property of the flow instead of 
the fluid, it is essential to use a robust turbulence model to parameterize the vertical turbulent 
mixing. In this study, the vertical eddy coefficients ( A v , D v and K v ) are computed from 

a simplified second-order closure model developed by Sheng and Chiu (1986), and Sheng 
and Villaret (1989). 



26 

Various form forms of the equation of state can be used. The present model uses the 
equation given by (Eckart, 1958) 

P 

P = 



(a + 0.698P) 

P = 5890 + 387/ -0.357r 2 + 3S (3.6) 

a = 1779.5 + 1 1.257 -0.04757/ 2 - (3.8 + 0.017)5 

where T is in °C , S is in ppt and p is in g I cm . 

The complete details of model equations in the curvilinear boundary-fitted and sigma 
coordinates have been derived (Sheng, 1989), and are presented in Appendix B. 
3.1.2 Sediment Transport Model 

The suspended sediment model includes the advection-diffusion processes, which are 
computed by the hydrodynamics model, as well as such processes as erosion, deposition, 
flocculation, settling, consolidation, and entrainment (Sheng, 1986; Metha, 1986). 

The governing equation that represents the transport of suspended sediments is given 
by: 



dc due dvc d(w + w)c d 

— + + ^ + 



dt dx dy dz dx \ H dx ) dy\ dy 



B* 



+ ■ 



'B^ 



K dz 



c)z 



(3.7) 



where c is the suspended sediment concentration, w 5 is the settling velocity of suspended 
sediment particles (negative downward), B H is the horizontal turbulent eddy diffusivity, and B v 

is the vertical turbulent eddy diffusivity. 

Four simplifying approximations are implied in the above equation: 1) the concept 
of eddy diffusivity is valid for the turbulent mixing of suspended sediments, 2) the suspended 
sediment dynamics are represented by the concentrations of two particle size groups (Sheng 
et al., 2002a), 3) the suspended sediment concentrations are sufficiently small that all 



27 
particles follow turbulent eddy motions, 4) the SSC is sufficiently low that non-Newtonian 
behavior can be neglected. 

In this study, the determination of settling, flocculation, deposition, erosion, 
flocculation, and consolidation processes are based on the previous work of Sheng and Lick 
(1979), Sheng(1986), Metha (1989), Sheng et al. (1990), Chen and Sheng (1994) and Sun 
and Sheng (2001). 

3.2 Boundary and Initial Conditions 
Obtaining solutions for Equations (3.1) Through (3.7) requires the specification of 
appropriate boundary and initial conditions. 
3.2.1 Boundary Conditions 

The boundary conditions at the free surface (a=0) in a non-dimensional, vertically 
stretched, boundary-fitted coordinate system are: 

du H „ 



For Hydrodynamics: \ 



dcr Ey 



*" 3.7 E, " 

For Salinity: = (3.8) 

dcr 



dT H Pr, 



v 



For Temperature: — — = q T 



dcr E 



v 



D v fci ^ 

For Sediment: w.c. + —^—^ = 

" ' H dcr 
where E v is a vertical Ekman number and Pr v is a vertical Prandtl number; The Cartesian 

wind stress, t", is calculated using 

T w =pC,v Ju 2 +v 2 (3.9) 

y fa as w v w w v ' 



28 
where p a is the air density (0.0012 g/cm 3 ), u w and v w are the components of wind speed 
measured at some height above the sea level. C ds , the drag coefficient, is given as a function 
of the wind speed measured at 10 meters above the water surface by (Garrat, 1977). 

C ds = 0.001(0.75 + 0.067 W s ) (3.10) 

Boundary conditions, if specified in a Cartesian coordinate system, such as wind 
stress, must be transformed before being used in the boundary-fitted equations. For example, 
the surface stress in the transformed system is given by 



3£ 3£ 
dx dy 

drj drj 



r <-ar- + * r < 



r,„=-^,+— r, v (3 - u) 

ox oy 

The second surface boundary condition is the kinematic free surface boundary 
condition which states 

BC d£ dC 

w = -?- + u-^ + v^f- (3.12) 

dt ox oy 
The boundary conditions at the bottom in a non-dimensional, vertically stretched (o 

= -1), boundary-fitted coordinate system are 

For Hydrodynamics: 

\ -£- = —T^ =-^H r Z r C d [g u u 2 b + 2g l2 u b v b + g 22 v;~\u b 



da E v "' A, 

A ^L = JL t - = £j-HZC 

da E v \ r 



g u u; + 2g n u b v b + g 22 v 2 b ~\v b 



For Salinity: ^ = ° ( 313 ) 

dcr 

dT n 
For Temperature: — — = U 

da 

D v dc 

For Sediment: w.c i + — -— - 1 - = A _ E t 

si ' H da ' ' 



29 
where u b and v h are the contravarient velocity components at the first grid point above the 
bottom. Ej is the erosion rate and D, is the deposition rate for sediment group i. C d is drag 
coefficient which is a function of the size of bottom roughness elements, z , and the height 
at which u h is measured, so long as z, is within the constant flux layer above the bottom. The 
drag coefficient is given by (Sheng, 1983) 



c d = 



K 



(3.14) 



ln(z,/z ) 
where k=0.4 is the von Karman constant, z =k/30 and k s is the bottom roughness. 

Along the shoreline where river inflow may occur, the conditions are generally 

u = u(x,y,(J,t) 

v = v(x,y,cr,t) 

w = 

(3.15) 
S = S(x,y,cr,t) 

T = T(x,y,cr,t) 

c, =c i (x,y,a,t) 

Contravarient velocity components provide lateral boundary conditions similar to 

those in Cartesian systems (x,y). Along solid boundaries, the normal velocity component 

must be zero to satisfy the no-slip condition. In addition, normal derivatives of salinity, 

temperature and suspended sediment concentration are assumed to be zero. When flow is 

specified at a boundary, the normal velocity component is prescribed. Tidal boundary 

conditions are specified using water level, (, directly When tidal boundary conditions are 

given in terms of C the normal velocity component is assumed to be of zero slope, while 

tangential velocity component may be either zero, of zero slope, or computed from the 

momentum equations. During an ebb tide, the concentrations of salinity, temperature, and 

suspended sediment flowing out are calculated using a 1-D advection equation while during 



30 
a flood tide the offshore concentrations are generally prescribed as either fixed or time 
varying. 
3.2.2 Initial Conditions 

To initiate a simulation, the initial spatial distribution of (,, u, v, w, S, T, and c, must 
be specified. When these values are unknown, "zero" initial fields can be used. When these 
values are known at a limited number of locations, an initial field can be generated by spatial 
interpolation. In the principle, the interpolated field should satisfy the conservation equation 
for a particular variable. For practical simulations, a "spin-up" period is required to damp 
out transients caused by the initial condition, which does not satisfied conservation. The 
length of a spin up period is variable and depends on such factors as basin size, flushing 
time, and current velocity. 

3.3 Heat Flux at the Air-Sea Interface 

The ocean receives energy through the air-sea interface by exchange of momentum 
and heat. The temperature of ocean waters varies from place to place and from time to time. 
Such variations are indications of heat transfer by currents, absorption of solar energy, and 
loss by evaporation. The size and character of the temperature variations depends on the net 
rate of heat flow into or out of water body (Pickard and Emery, 1990). The transfer of heat 
across the air-sea interface determines the distribution of temperature in the ocean as a 
surface boundary condition for the temperature equation, as follow: 

dT 
PoK v — = q T (3-16) 

az 

where q T is the net heat flux across the surface water. 

To estimate the net heat flux at the air-sea interface, it is necessary to consider the 
following processes: net heat flux across the surface, q T \ heating by incoming solar radiation 



31 

(insolation), q s ; warming by conductive heat exchange across the surface (sensible heat 
flux), q h ; cooling by net outgoing long wave radiation, q h ; heat loss as water evaporated 
(latent heat flux), q, ; heat flux from precipitation. Other source of heat flow, such as that 
from the earth's interior, change of kinetic energy of waves into heat at the surf; heat from 
chemical or nuclear reactions, are small and can be neglected. The net heat exchange at the 
surface can be divided into four terms: 

qT = <2s-<ib-<ik-<ii (3 - 17) 

Direct measurement of these fluxes is the best way to provide the net heat flux at the 
air-sea interface. However, directly measured air-sea fluxes are only available at very few 
stations to allow calculation of air-sea interface over a large area. Instead, directly measured 
air-sea fluxes are used for developing, calibrating, and verifying the parametric formulae for 
estimating the fluxes from the primary variables including wind speed, air temperature, water 
temperature, and cloud cover (Taylor et al., 2000). These required parametric formulae are 
then used to compute heat fluxes over a large area. 
3.3.1 Short- Wave Solar Radiation 

The main source of heat flux through the air-sea interface is short-wave solar 
radiation, g v (incoming solar radiation), received either directly or by reflection and scattering 
from clouds and the atmosphere. The incoming solar radiation is based on the empirical 
formula of Reed (1977): 

q s = S sin y ( A + fisin y)(\ - 0.62n - 0.0019r)(l - a) (3.18) 

where S is solar constant =1353 w/m 2 
Y is solar elevation angle 
n is cloud cover = 0-1 



32 

a is the albedo = 0.06 

A and B are empirically determined coefficients for each category of reported total 
cloud amount and the solar elevation 

The solar elevation angle was computed to the nearest 0. 1 ° from date and time for the 

latitude and longitude of the study area, using the following equations (Miller and 

McPherson, 1995): 

360 
<p = (d-l) 

365.242 

J = 12 + 0.1236sin^-0.0043cos^ + 0.1538sin2^ + 0.0608cos2^ 
<t = 279.9348 + p + 1.9148 sin p-0.0795cos^ + 0.00199 sin 2^ + O.OO16cos20 

Y = \5(t-S)-A (3.19) 

k = arcsin (0.39785077 x sin a) 
sin /3 - sin /sin k + cos y cos /f cos Y 

where (p is the angular fraction of the year, in degree; d is Julian date; 6 is true solar noon, 
in hours; T is the solar hour angle, in degrees; T is the Greenwich Mean Time, in hours; X 
is the longitude, in degrees; o is an estimate of the true longitude of the sun, in degrees; K is 
the solar declination, in degrees; y is the latitude, in degrees; and P is the solar elevation 
angle, in degrees. 
3.3.2 Long-Wave Solar Radiation 

The back radiation term, q h , is the net amount of energy lost by the sea as long-wave 
radiation. The amount of long wave flux is dependent on surface water temperature, 
atmospheric temperature, humidity and cloud cover (Clark et al, 1974) 

q b = eoT* (0.39 - 0.05e 05 ) (l - An 2 ) + 4eaT* (T s - T a ) (3.20) 

where e is the emittance of the sea surface = 0.98 

o is Stefan-Boltzman constant = 5.673 x 10" 8 W/m 2 



33 

e is the water vapor pressure 
T a and T s are the air and sea temperatures in K. 
A is a cloud cover coefficient which varies with latitude 

The Antoine constants (http://www.owlnet.rice.edu/~ceng301/31.html) give the vapor 
pressure as a function of temperature so that the approximate value for water vapor pressure 



is: 



( 3985.44 A 

16.5362- 



e = exp 

T s -38.997 

where T s is the surface water temperature (°K). 



(3.21) 



A theoretical calculation of the mean values of the coefficient X for different latitudes 
has been made by M. E. Berliand (1952). In this calculation, the mean frequency of clouds 
of different layers at each latitude is taken into account. The values obtained for the 
coefficient X are given in Table 3.1 (Budyko, 1974). The reduction of values of this 
coefficient in low latitudes is explained mainly by a higher mean altitude of clouds in these 
regions. 
Table 3.1 Mean latitudinal values of the coefficient X 



* 


75° 


70° 


65° 


60° 


55° 


50° 


45° 


40° 


X 


0.82 


0.80 


0.78 


0.76 


0.74 


0.72 


0.70 


0.68 


d> 


35° 


30° 


25° 


20° 


15° 


10° 


5° 


0° 


X 


0.65 


0.63 


0.61 


0.59 


0.57 


0.55 


0.52 


0.50 



3.3.3 Sensible and Latent Heat Fluxes 

The most significant part in terms of heat transfer from the sea to the atmosphere is 
latent heat flux. The rate of heat loss is equal to the rate of vaporization times the latent heat 
of evaporation. Sensible heat flux is due to temperature gradient in the air above the sea. 
The rate of loss or gain of heat is proportional to the temperature gradient, heat conductivity 



34 
and the specific heat of air at constant pressure. Sensible and latent heat fluxes are estimated 
by using the bulk aerodynamic equations (Mellor, 1996): 

q k =pc p C H U w (T s -T a ) (3.22) 

qi =pC E LU l0 (h s -h a ) (3.23) 

where p is the density of air 

C H and C E are transfer coefficients for sensible and latent heat respectively 

U I0 is the wind speed at 10 m height above surface 

T s and T a are the sea surface and air temperatures 

c p is the specific heat of air, 1.0048 x 10 3 J/kg°K 

L is the latent heat of evaporation, 2.4 x 10 6 J/kg 

h s and h a are specific humidities at the surface and at a 10 m height above the surface 

The air density is determined from the ideal gas equation in the form, 

p = pJ(RT a ) , where P a is air pressure and R is universal gas constant, 287.04 J/kg °K. 

The traditional estimates of C H and C E over the ocean tend to support a fairly constant value 
over a wide range of wind speed. Smith (1989) recommended a constant "consensus" value 
C E = (1.2 ± 0.1)xl0" 3 for winds between 4 and 14 m/s. DeCosmo et al. (1996) also suggested 
a near constant value with C E = (1.12 ± 0.24)xl0 3 for winds up to 18 m/s. For the Stanton 
number, C H , Friehe and Schmitt (1976) obtained slightly different values for unstable and 
stable conditions, 0.97 xlO" 3 and 0.86 xlO' 3 respectively. Smith (1989) suggested C H = 1.0 
x 10 3 . C E - 1.2 xlO" 3 and C H = 1.0 x 10" 3 were used for this study. 

The formulae used for air specific heat, c p , and latent heat for evaporation, L, have 
been taken from Stull (1988), 

c p =1004.67 + 0.84^ (3.24) 



35 

L = 10 6 [2.501 -0.00237(7; -273.16)] (3.25) 

An empirical relation (based on the Clausius-Clapeyron theory) provides the 

• i ir .(0.7859+O.O3477r)/(l+O.0O4l2r) , » „-„_-„! 

saturation, water vapor partial pressure, e s = 10 v mb. A general 

relation between specific humidity and partial pressure is q - (Q.622e/P)/(l - 0.31Se/P) 

, where P is the atmospheric pressure (Mellor, 1996). Specific humidities at surface and at 
a 10 m height above the surface, h s and h a , can be calculated using these empirical formulae 
with T s and T a respectively. 



CHAPTER 4 
WATER QUALITY MODEL 

In this chapter, a water quality model for simulating water quality processes occurring 
in the Charlotte Harbor estuarine system in both the water and sediment columns is 
presented. This is an extension of an earlier model developed to simulate water quality for 
Lake Okeechobee (Sheng et al., 1993; Chen and Sheng, 1994), Tampa Bay (Yassuda and 
Sheng, 1996) and the Indian River Lagoon (Sheng et al., 2001 and 2002). These models 
include the effect of sediment transport on nutrient dynamics through the explicit use of a 
sediment transport model and the incorporation of a nutrient resuspension flux. 

The water quality model incorporates on the interactions between oxygen balance, 
nutrient dynamics, light attenuation, temperature, salinity, phytoplankton and zooplankton 
dynamics. To develop the water quality model, the mass conservation principle can be 
applied to each water quality parameter, as it relates to phytoplankton and zooplankton 
dynamics, nitrogen and phosphorous cycles, and oxygen balance. 

The nitrogen and phosphorous cycles in an estuarine system, are modeled through a 
series of first order kinetics. Nutrient concentrations in the estuarine system are constantly 
changing in time and space due to loading from rivers, exchanges with the ocean, seasonal 
climatic changes, biogeochemical transformations, hydrodynamics, and sediment dynamics. 
Nitrogen species include ammonia nitrogen (NH3), soluble ammonium nitrogen (NH4), 
nitrate and nitrite nitrogen (N03), particulate ammonium nitrogen (PIN), soluble organic 



36 



37 

nitrogen (SON), particulate organic nitrogen (PON), phytoplankton nitrogen (PhyN), and 
zooplankton nitrogen (ZooN). Similar to the nitrogen species, phosphorous species include 
soluble reactive phosphorous (SRP), particulate inorganic phosphorous (PIP), soluble organic 
phosphorous (SOP), particulate organic phosphorous (POP), phytoplankton phosphorous 
(PhyP), and zooplankton phosphorous (ZooP). 

Phytoplankton kinetics are the central part of this water quality model, since the 
primary water quality issue in the estuarine system is eutrophication (Boler et al., 1991). 
Phytoplankton population is a complex variable to measure in the field. However, the lack 
of data on each specific species prevented a more detail characterization, the entire 
phytoplankton community is represented in this study by a single state variable and 
quantified as carbonaceous biomass. Chlorophyll _a concentrations, for comparison with 
observations, are obtained through division of computed carbonaceous biomass by the 
carbon-to-chlorophyll_a ratio. 

The oxygen balance couples dissolved oxygen to other state variables. Reaeration 
through the air-sea interface, and phytoplankton production during photosynthesis are the 
main source for oxygen. Oxidation, nitrification, respiration, mortality and SOD reduce 
oxygen in the system. Oxidation of organic matter and carbonaceous material, respiration by 
zooplankton and phytoplankton, and oxygen consumption during the nitrification process are 
collectively grouped into the CBOD (Carbonaceous-Biogeochamical Oxygen Demand) 
variable, which is a sink for dissolved oxygen (Ambrose et al., 1994). 

The methods of coupling with hydrodynamic and sediment transport models, the 
simulated parameters, the assumptions, the chemical/biological processes of the CH3D water 
quality model (CH3D-WQ3D) was compared with those for existing water quality models, 



38 
specially Water Quality Analysis ans Simulation Program (WASP), the integrated 
Compartment water quality model developed by the US Army Corps (CE-QUAL-ICM) in 
Appendix C. 

Temperature, salinity, and light are important parameters that effect the rate of 
biogeochemical reactions. Most of the transformation processes in the nutrient cycle are 
affected by temperature. Salinity also influences DO saturation concentration and is used in 
the determination of kinetics constants that differ in saline and fresh water. Light intensity 
affects the photosynthesis process and thus algae growth rate. 

The water quality model is enhanced to include a reaeration model and a sediment 
oxygen demand (SOD) model and newly coupled with temperature model and physics-based 
light attenuation model. Applying enhanced water quality model and a more accurate light 
model and temperature model could improve the vertical distribution and daily fluctuation 
of both phytoplankton and dissolved oxygen. Furthermore, the bottom water hypoxia at 
upper Charlotte Harbor which is caused by SOD and vertical stratification can be reproduced 
and analyzed in this study. 

4.1 Mathematical Formulae 

The water quality equations are derived from a Eulerian approach, using a control 
volume formation. In this method, the time rate of the concentration of any substance within 
this control volume is the net result of (i) concentration fluxes through the sides of the 
control volume, and (ii) production and sink terms inside the control volume. The 
conservation equation for each of the water quality parameters is given by: 



M + V .(fl«) = V-[DV(fl«)] + Q, 
(i) («) (Hi) (iv) 



(4.1) 



39 

where (I) is the evolution term (rate of change of concentration in the control volume), (ii) 
is the advection term (fluxes into/out of the control volume due to the advection of the flow 
field), (iii) is the diffusion term (fluxes into/out of the control volume due to turbulent 
diffusion of the flow field), and (iv) is the sink and source term, representing the kinetics and 
transformations due to sorption/desorption, oxidation, excretion, decay, growth, and bio- 
degradation. In the finite difference solution of the water quality model, the advection and 
horizontal diffusion terms are treated explicitly, whereas the vertical diffusion and 
biogeochemical transformations are treated implicitly. The detail descriptions of numerical 
solution technique are described in Appendix D. The water quality equations in the 
curvilinear non-orthogonal boundary fitted system (£, X], o) are given by: 



dt 



HS CV da 



3 < a* ^ 



V 



'V -s 

da 



-R, 



d(H(D)(p t 
da 



/?„ 



**\*Z 



—( y fg~ Hu<p i ) + — (Jg~ Hv(p i ) 



+ 



^CHyJSo 



^ 



go H 8 



+ - 



^CH V #0 



drj 



8o H S 



drj 
21 d<P, 



+ Jg Hg l 



+ Jgo H g' 



drj 



(4.2) 



J] 



+ Q, 



where (p i represents any water quality parameter, -^g^ x% me Jacobian of horizontal 



transformation, (g ,g ,g ) are the metric coefficients of coordinate transformation, 



and Q. represent biogeochemical processes. Equation (4.2) is in dimensionless form and the 



dimensionless constants are defined in Appendix B. 

In the following sections, the biogeochemical processes controlling the sink/source 
term of Equation (4.2) will be discussed in detail for nutrient dynamics, zooplankton and 



40 
phytoplankton dynamics, and oxygen balance in system. 

4.2 Phytoplankton Dynamics 

The overall water quality in the system is markedly influenced by the dynamics of 
zooplankton and phytoplankton communities (Boleretal., 1991). Phytoplankton dynamics 
and nutrient dynamics are closely linked, since nutrient uptake during phytoplankton growth 
is the main process to remove dissolved nutrients from the water, and phytoplankton and 
zooplankton respiration and mortality are major components of nutrient recycling. The 
diurnal variation of dissolved oxygen is related with photosynthetic oxygen production 
during the day and oxygen consumption due to phytoplankton respiration during the night. 
4.2.1 Modeling Approach 

Phytoplankton kinetics are represented by growth, respiration, non-predator-mortality, 
grazing by zooplankton, and a settling term. The phytoplankton sources and sinks in the 
conservation equation can be written as: 

dPhvC ( d s \ 

^1= Ma -K^-K as+ -WS aigae -PhyC-^ZooC (4.3) 

at \ az J 

Where PhyC is phytoplankton biomass, expressed as carbon (gCm 3 ); /J a is phytoplankton 
growth rate (1/d); K as is respiration rate (1/d); K^ is non-predator mortality (1/d); WS algae is 
the phytoplankton settling velocity(m/d); ZooC is zooplankton biomass (gCm 3 ); and /u z is 
zooplankton growth rate (1/d). 

Phytoplankton growth is determined by the intensity of light, by the availability of 
nutrients, and by the ambient temperature. Light limitation is formulated according to the 
photo inhibition relationship (Steel, 1965). The quantity of the growth limitation factor for 
nutrients is related with a half-saturation constant. The half-saturation constant refers to the 
concentration of the nutrient at which the growth rate is one half its maximum value. This 



41 
results in a hyperbolic growth curve. An exponential increasing function is applied for 
temperature limitation.. 

The minimum formation approach has been used to combine the limiting factor of 
light and each limiting nutrient. The minimum formation is based on "Liebig's law of the 
minimum" which states that the factor in shortest supply will control the growth of algae 
(Bowie et al., 1980) 

Va=(M a ) aaK -f(T)'M)'f(N,P) 

(4.4) 



= (/0™C 2 °-min 



— exp 
L 



Is J 



NH 4 + N0 3 SRP 



H n +NH 4 + N0 3 H p +SRP 

where OuJ max is the phytoplankton maximum growth rate (1/day); d is temperature 
adjustment coefficient; Tis temperature (°C); /is the light intensity, calculated by the light 
attenuation model; I s is the optimum light intensity for algae growth; H n is half saturation 
concentration for nitrogen uptake (gN — 3 ); H p is half saturation concentration for 
phosphorous uptake (gP — 3 ); NH4 is ammonium concentration (gN — 3 ); N0 3 is nitrate 
concentration (gN ~~ 3 ); and SRP is soluble reactive phosphorous concentration (gP — 3 ). 

Respiration and mortality are considered to be an exponentially increasing functions 
of temperature: 

T T (4.5) 

where (K a J Tr and (K M ) Tr are respiration and mortality rate at Tr (1/day); and Tr is reference 
temperature of respiration and mortality. 

For phytoplankton, literature values of algae settling velocity, which account for the 
limited vertical motion of these organisms will be used. 

Zooplankton are included in water quality models primarily because of their effects 



42 
on algae and nutrients. Phytoplankton and zooplankton dynamics are closely tied through 
predator-prey interaction. Phytoplankton dynamics are of major concern in this study while 
no attempt is made to investigate zooplankton dynamics due to lack of zooplankton data. 
Zooplankton is only considered as the predators of phytoplankton, utilizing their available 
biomass as food supply. Zooplankton kinetics, influenced by growth, respiration and 
mortality, are represented in a source and sink term as (Bowie, 1985): 

l^-fa-K.-K.yZooC (4.6) 

at 

where //, is zooplankton growth rate (1/day); K as is respiration rate (1/day); and K^ is 

mortality (1/d) 

Zooplankton growth is represented by a temperature-dependent maximum growth 

rate, which is limited by phytoplankton availability: 

K '~ ' H phr+ PhyC-Trs ph:1 

where (/u z ) m . M is the zooplankton maximum growth rate (1/day); 6\s temperature adjustment 
coefficient; H phy is half saturation concentration for phytoplankton uptake (gC — 3 ); and Trs phy 
is threshold phytoplankton concentration for zooplankton uptake (|ig/l). 
4.2.2 Relationship between Phytoplankton and Nutrients 

Phytoplankton biomass is quantified in units of carbon. In order to express the effects 
of phytoplankton on nitrogen and phosphorous, the ratio of nitrogen-to-carbon and 
phosphorous-to-carbon in phytoplankton biomass must be specified. Global mean values of 
these ratios are well known (Redfield et al., 1966). The amounts of nitrogen and 
phosphorous incorporated in algae biomass is quantified through a stoichiometric ratio. 
Thus, total nitrogen and total phosphorous in the model are expressed as: 



43 

TotN = NH. + NH, + NO, + SON + PON + PIN + Anc ■ PhyC + Anc ■ ZooC 

4 J (4.8) 

TotP = SRP + SOP + POP + PIP + Ape ■ PhyC + Ape ■ ZooC 

where TotN is total nitrogen (gN ~ 3 ); NH4 is dissolved ammonium nitrogen (gN ~ 3 ); NH3 
is ammonia nitrogen (gN ~ 3 ); N03 is nitrate and nitrite nitrogen (gN ~ 3 ); SON is soluble 
organic nitrogen (gNm~ 3 ); PON is particulate organic nitrogen (gN _3 ); PEN is particulate 
inorganic nitrogen (gN ~ 3 ); Anc is Algae nitrogen-to-carbon ratio (gN/gC) ; TotP is total 
phosphorous (gP ~ 3 ); SRP is soluble reactive phosphorous (dissolved phosphate) (gPrn 3 ); 
SOP is soluble organic phosphorous (gP -3 ); POP is particulate organic phosphorous (gP -3 ); 
PIP is particulate inorganic phosphorous (gP _3 ); and Ape is algae phosphorous-to carbon 
ratio (gP/gC). 

The connection between the carbon, nitrogen and phosphorous cycle is shown in 
Figure 4.1. Phytoplankton uptakes dissolved ammonium nitrogen, nitrate and nitrite 
nitrogen, and soluble reactive phosphorous during production and releases dissolved 
ammonium nitrogen, soluble reactive phosphorous and organic nitrogen, organic 
phosphorous during respiration and mortality processes. Zooplankton has similar kinetic 
processes as phytoplankton. The measured phytoplankton as algae mass per volume, was 
converted to phytoplankton carbon with a algae to carbon ratio. The amounts of nitrogen and 
phosphorous from phytoplankton can be converted with a nitrogen to carbon ratio and a 
phosphorous carbon ratio, respectively. 

4.3 Nutrient Dynamics 

Nutrients are essential elements for life processes of aquatic organisms. Nutrients 
of concern include carbon, nitrogen, phosphorous, silica and sulfur. Among these nutrients, 
the first three elements are utilized most heavily by zooplankton and phytoplankton. Since 
carbon is usually available in excess, nitrogen and phosphorous are the major nutrients 



44 
regulating the ecological balance in an estuarine system. Nutrients are important in water 
quality modeling for several reasons. For example, nutrient dynamics are critical 
components of eutrophication models since nutrient availability is usually the main factor 
controlling algae bloom. Algae growth is typically limited by either phosphorous or nitrogen 
(Bowie et al., 1980). Details on the nutrient dynamics including all the equations used by 
the water quality model to calculate the nitrogen and phosphorous, can be found in 
Appendix-E. 



Excretion 



NH3 



Volatilization 



I Sorption/ 

Desorption w i r 1 1 



Mortality 



Ammonification 

*■ 



SON 



Sorption/ 



Desorption 



NITROGEN 
A 



Nitrification 



NH4 



N03 



PhyN=a r ,'PhyC 



CARBON 

i, 



Excretion 



Uptake 



Mortality 



Mortality 



»- PON 



Mortality 



Uptake 



Phytoplankton (PhyC) 



Uptake 



PhyP^ 'PhyC 



Uptake 



PHOSPHOROUS 



Excretion 



SRP 



n i 4 



PIP 



Sorption/ 



Desorption 



Zooplankton (ZooC) 



Mortality 



•H POP r* 



Mortality 



Mortality 



Mineralization 



SOP 



Sorption/ 



Desorption 



Mortality 



Excretion 



Figure 4.1 The connection between nitrogen, phosphorous and carbon cycle 



45 
4.4 Oxygen Balance 

Dissolved oxygen (DO) refers to the volume of oxygen contained in water. Five state 
variables participate in the dissolved oxygen balance: phytoplankton carbon (PhyC), 
ammonia (NH4), nitrate (N03), carbonaceous biochemical oxygen demand (CBOD), and 
dissolved oxygen (DO). A summary is illustrated in Figure 4.2. The methodology for the 
analysis of dissolved oxygen dynamics in natural water, particularly in streams, rivers, and 
estuaries is reasonably well-developed (O'Connor and Thomann, 1972). Dissolved oxygen 
evolution depends on the balance between production from photosynthesis, consuming from 
respiration and mortality, and exchanges with the atmosphere and sediment. 

The main physical mechanisms influencing DO concentration are horizontal and 
vertical dispersion and diffusion. Vertical diffusion occurs across the air-sea interface as a 
function of wind, waves, currents, and DO saturation rate. Laboratory experiments show that 
the bottom shear stress controls the dissolved oxygen diffusive layer thickness and the flux 
at the sediment-water interface (Steinberger andHondzo, 1999). Slow oxygen diffusion rates 
and high oxygen demand by sediment results in a thin aerobic layer. Two distinct sediment 
zones are created in the sediment column: an aerobic layer and an anaerobic layer. The 
thickness of the aerobic soil zone is influenced by oxygen concentration in the overlying 
water column, and concentration of the reduced compounds in the anaerobic soil zone. The 
model dissolved oxygen cycle includes the following processes 

1 ) Reaeration 

2) Carbonaceous oxygen demand (CBOD) 

3) Nitrification 

4) Sediment oxygen demand (SOD) 



46 



5) Photosynthesis and respiration 



Carbonaceous Oxygen Demand 









Zooplankton 


Oxidation 






Mortality 

K^'ZOOC 




Haunt + DO 


CBOD 








CBOD 

1 
















Mortality 

K^'PHYC 






c 
1 

1 

^ 


f 

r 






Phytoplankton 



dz 



i Diffusion 



CBOD 



Oxidation 



Denitrification 



DO 



»™» +DO 



-CBOD K, 



H ,»;) 



«„,,+ DO 



-NO 3 



Dissolved Oxygen 

Air 



Oxidation 



K D — CBOD 

"H iM + DO 



Reaeration K^DOg-DO) 



Phytoplankton 



DO 



Photosynthesis & Respiration 



,->• 



Nitrification 



6f^_22_AW4 

14 m H xrT +DO 



1(1.3-0. 



3P.)^-(^+^.)l*PWC»«, 



Sediment Oxygen Demand 



KO2+D0 

Water column 



* Diffusion 



DO 



Nitrification 



DO 



14 " H^+DO 



NH4 



Aerobic Layer 

Figure 4.2 DO and CBOD cycles 
Reaeration 

Reaeration is the process of oxygen change between the atmosphere and sea surface. 
Typically, dissolved oxygen diffuses into surface waters because dissolved oxygen levels in 
most natural waters are below saturation. However, when water is super-saturated as a result 
of photosynthesis, dissolved oxygen returns the atmosphere. Dissolved oxygen saturation 
in seawater is determined as function of temperature and salinity (APHA, 1985) 



47 



1.575701xl0 5 6.642308xl0 7 1.2438x10'° 8.621949x10" 
In DO. =-139.34411 + — - + ~ 3 =5 



r 

(4.9) 

Sa I , 1.9428x10 

3.1929x10- + • 



, 1.9428x10 3.8673xl0 3 ^ 



1.80655 



T 2 



where DO s is equilibrium oxygen concentration, mg/1, at standard pressure 
T is temperature, °K, °K = °C + 273.150 
Sa is salinity, ppt 
The reaeration process is modeled as the flux of dissolved oxygen across the water 

surface: 

V^- = K AE A s (DO I -DO) (4-10) 

at 

where V and As are volume and surface area of the water body 

In case where the air-sea interface is not constricted, the volume is V = A s ■ Az . The 

equation for reaeration can be expressed as 

^. = ^AL( D0 -DO) (4.11) 

dt Az V v 

where DO is dissolved oxygen concentration (mg/1); K AE is reaeration coefficient (m/day). 
Many empirical formulas have been suggested for estimating reaeration rate 
coefficient specially in the river. Bowie et al. (1985) have reviewed thirty-one reaeration 
formulas, and have tried to evaluate the performance of the each formulas. Most formulae 
have been developed based on hydraulic parameters, most often depth and velocity. This 
review of stream reaeration has shown that no one formula is best under all conditions, and 
depending on the data set used, the range of the reaeration coefficients in the data set, and 
error measurement selected, the best formula may change. Among these formulas, the most 



48 
common method of simulating reaeration in rivers is the O'Connor-Dobbins formula. This 
method has the widest applicability being appropriate for moderate to deep streams with 
moderate low velocities. With approximately 2.09xl0" 5 cm 2 /s diffusivity of oxygen in 
natural waters, the O'Connor-Dobbins formula can be expressed as 

jjO.5 

K u =3.93— (4.12) 

For standing water, such as lake, impoundments, and wide estuaries, wind becomes 

the predominant factor in causing reaeration. The oxygen-transfer coefficient itself can be 

estimated as a function of wind speed by a number of formulas. Chapra (1997) compared 

four common wind-dependent reaeration formulas: Broeckeret al. (1978), Banks andHerrera 

(1977), O'Connor (1983), and Wanninkhof et al. (1991). The comparison shown in Figure 

4.3 show all these methods except Broeckeret al.'s have similar reaeration coefficients when 

wind speed is less than 5 m/s. When wind speed is greater than 5 m/s, Bank and Herrera's 

formula produce the middle range of reaeration coefficient among these three formulas. 

This formula uses various wind dependencies to attempt to characterize the difference 

regimes that result at air-water interface as wind velocity increase (Banks 1975; Banks and 

Hen-era, 1977). 

K l =0.72SU° w 5 -03llU w + 0mi2Ul (4.13) 

Since estuary gas transfer can be affected by both water and wind velocity, effort to 
determine reaeration in estuaries combines elements of current and wind-driven approaches. 
Thomann and Fitzpatrick (1982) combined the two approaches for estuaries affected by both 
tidal velocity and wind, 



K AE = 3.93J^ + 0.728£/° 5 - 0.3 17£/„ -0.0372£/ H 2 (4.14) 



49 

where U is depth averaged velocity (m/s); H is a depth (m); U w is wind speed (m/s). 

For the Charlotte Harbor estuarine system, The aeration coefficient is assumed to be 
proportional to the water velocity, depth, and wind speed following Thomman and 
Fitzpatrick (1982). 



10r- 

9 - 
8 - 
7 - 



n 
? 5 



* 4 

3 
2 
1 




Broeckeretal. (1978) ,' 



Wanninkhofetal(1991). 




Banks and 
Herrera(1977) 



4 6 

U w (m/s) 



Figure 4.3 Comparison of wind-dependent reaeration formulas. 
Carbonaceous Oxygen Demand 

The use of carbonaceous oxygen demand (CBOD) as a measure of the oxygen- 
demanding processes simplified modeling efforts by aggregating their potential efforts 
(Ambrose et al., 1994). Oxidation organic matter, nitrification, non-predatory mortality and 
respiration by zooplankton and phytoplankton are nitrogenous-carbonaceous-oxygen- 
demand, collectively combined into the state variable CBOD. 



50 
The kinetic pathway of CBOD is represented in the source term of the equation as 
(Ambrose et al., 1994): 



For water column: 

?-CBOD = -*- 
dt dz 



IcBOD = -f [ws CB0D • (1 - fd CB0D ) ■ CBOD] - K D ■ °° ■ CBOD 
at dz tl r-o^-r UU 

(4.15) 

--■ — ■K DN ^ NO,+ A oc (K ar PhyC + K a ZooC) 

4 14 H no3 +DO V 

For sediment column: 

IcBOD = +^-[ws CBOD ■ (1 - fd CB0D ) ■ CBOD] - K D ■ °° ■ CBOD 

(4.16) 

--■ — ■K DN ^ NO, 

4 14 H nu3 + DO 

where fd CB0D corresponds to the fraction of the dissolved CBOD; ws CBOD is the settling 
velocity for the particulate fraction of CBOD ; K D is oxidation coefficient which is a 
temperature function; H CB0D is a half saturation constant for denitrification; K DN is 
denitrification constant; H m3 is a half saturation rate for denitrification; and A oc is oxygen- 
carbon ratio (g0 2 /gC). 

The consumption of oxygen, as a function of water column CBOD decay, can be 
expressed as CBOD oxidation; 

— DO = K D — CBOD (4.17) 

dt H CB0D + DO 

Note that non-oxidative processes such as settling, denitrification, and mortality do 
not contribute to dissolved oxygen depletion, and are not included in the expression. 

Nitrification 

The transformation of reduced forms of nitrogen to more oxidized forms 
(nitrification) consumes oxygen. Although nitrification is also a nutrient transformation, this 
section addresses oxygen consumption. First order kinetics is the most popular approach for 



51 
simulating nitrification in natural system: 

±DO—£k„— ^— -NH 4 (4.18) 

dt 14 m H nil +DO 

where K NN is nitrification rate which is a function of temperature; and H nil is the half- 
saturation constant for the bacteria growth. 

Sediment Oxygen Demand 

Sediment oxygen demand is a dissolved oxygen flux at water/sediment interface due 
to the oxidation of organic matter in bottom sediments. The particulate organic matters are 
from a source outside the system such as wastewater particulate or leaf litter materials, and 
generated inside system as occurs with plant growth in highly productive environments. 
According to accumulate these particulate organic matter in the bottom sediment, the 
sediment oxygen demand will increase due to oxidation of the accumulated organic matter 
at aerobic sediment layer. Figure 4.4 shows the mechanism of SOD flux with particulate 
organic matters in the sediment column. Particulate organic matter (POM) is delivered to 
the sediments by settling. Within the anaerobic sediment layer, the organic carbon, sulfate, 
nitrate undergo reduction reactions to yield dissolved methane, sulfide, dissolved ammonium 
nitrogen. These reduced species (CH4, H2S, NH4) diffuses upward to the aerobic layer 
where these are oxidized. During these oxidation reactions, SOD is generated. Any residual 
reduced species that are not oxidized in the aerobic layer is diffused back into the water 
column where additional oxygen is consumed by oxidation. Developing an SOD model is 
quite complicated task since many aerobic and anaerobic reactions in the sediments are 
involved and many chemical species are included in mass balance equation. Therefore, these 
redox chemistry with POM are usually treated as a composite characteristics of the particular 
system. Recently, techniques have been developed for investigating these factors. 



52 
DiToro et al. (1990) developed a model of the SOD process in a mechanistic fashion 
using the square-root relationship of SOD to sediment oxygen carbon content. Using similar 
analysis as applied to carbon, they also evaluate the effect of nitrification on SOD. In this 
model, carbon and nitrogen diagenesis are assumed to occur at uniform rates in a 
homogeneous layer of the sediment of constant depth (active layer). The sediment oxygen 
demand and sediment fluxes are calculated by the concentrations of particulate organic 
carbonaceous material and of particulate organic nitrogenous material in this active layer. 
The more detail description of sediment flux model developed by DiToro and Fitzpatrick 
(1993) is presented in Appendix F. Their framework has been applied to the Chesapeake 
Bay (Cerco and Cole, 1994; DiToro and Fitzpatrick, 1993) with sediment flux model which 
include ammonia and nitrate flux and the sulfide, oxygen, phosphorous and silica flux. 

A calculation of the detailed redox chemistry of the sediment interstitial water is 
required for a detail understanding of the situation for each reduced species and chemical 
parameters for redox reaction such as pH, Eh in the system. In the absence of data for CH4, 
H2S, Iron, Methane, and C02, and detail understanding of redox chemistry of the system, 
applying this model could create more uncertainty than the simple empirical formula based 
on the measurement technique. 

The process of oxygen demand in the sediment/water interface is usually referred to 
as sediment oxygen demand (SOD) because of the typical mode of measurement: enclosing 
the sediments in the chamber and measuring the change in the dissolved oxygen 
concentration at several time increments. The major factors affecting SOD are: temperature, 
oxygen concentration at the sediment water interface, organic and physical characteristics of 
the sediment, and current velocity over the sediments (Bowie, 1985). 



Water Column 



SL 



53 




02 



Aerobic Layer 



Oxidation 




CH4, NH4, H2S 



Anaerobic Layer 



Diffusion 



Reduction 



^ CH4, NH4, H2S 



Production of 
CH4, NH4, H2S 




Figure 4.4 The relationship between POM flux and SOD flux related in the oxidation and 
reduction of organic matter in sediment column. 



54 



Typical values of SOD are listed in Table 4.1 (Chapra, 1997). In general, values from 
about 1 to 10 g0 2 /m 2 -day are considered indicative of enriched sediments. According to this 
table, SOD at a mud bottom is higher than SOD at a sandy bottom. In Charlotte Harbor 
estuarine system, the sediment type near upper Charlotte Harbor is finer and includes more 
clay and mud than those for the other study area. The finer sediment which has clay, silt, and 
mud would have higher SOD as more organically rich sediments. From this assumption, 
SOD can be applied as a function of sediment bottom type (sediment size). 
Table 4.1 Average Values of Oxygen Uptake Rates of Bottom (Chapra, 1997) 



SOD rate at 20°C (g0 2 /m 2 -day ) 



Bottom type and Location 



Average Value 


Range 


7 




4 


2-10 


1.5 


1-2 


1.5 


1-2 


0.5 


0.2-1 


0.07 


0.05-1 



Sphaerolitus (10 g-dry wt — 2 ) 

Municipal sewage sludge: 
Outfall vicinity 
Downstream of outfall 

Estuarine mud 

sandy bottom 

Mineral soil 



The effect of temperature and sediment type on SOD can be represented by 



S b(T) = S b,20' ST '& 



7-20 



(4.19) 



where S B20 is an areal SOD rate at 20 °C (g0 2 /nr-day ) which is user defined value; ST is 
a fractional coefficient for sediment type. (When sediment type <2, ST=1, when sediment 
type >3, ST=0.5) ; is temperature coefficient. Zison et al. (1978) have reported a range of 
1.04 to 1.13 for 0. A value 1.065 is commonly employed and is used in this study. 

Oxygen is another factor that affects sediment oxygen demand. Sediment oxygen 
consumption is reduced as oxygen concentration in the overlying water decrease. Lam et al. 
(1984) use a Michealis-Menten relationship to represent the dependence, by a saturation 
relationship, 



55 
S B (DO)= D ° S B (T) (4.20) 

K SOD +UU 

where K SOD is half saturation rate for SOD. Lam et al. (1984) have suggested a value for K S0D 
of 1.4 mg/1. 

The decay of substrate is assumed to balance continued settling resulting in a steady- 
state sediment concentration of oxygen demand substrate. According to this assumption, the 
kinetic equation for sediment oxygen demand is (Ambrose et al., 1994): 

dDO__SOD (421) 

dt H 

where H is water depth (m); and SOD is sediment oxygen demand (as measured), g0 2 /m 2 - 
day. SOD can be calculated as a function of temperature, dissolved oxygen at the water- 
sediment interface, and sediment bottom type based on characteristics of SOD measurement. 

Exchange of material between the water column and benthic sediment is an important 
component of the eutrophication process. Sediment oxygen demand may comprise a 
substantial fraction of total system oxygen consumption (Cerco and Cole, 1995). 

Oxygen consumption in the sediments depends upon water-column temperature and 
oxygen availability, and sediment type. As temperature increases, respiration in the sediment 
increases. Sediment oxygen consumption is reduced as oxygen concentration in the 
overlying water decreases. Therefore, the kinetic equation for sediment oxygen consumption 
(SOC) in sediment column can be represented as (Cerco and Cole, 1995): 



dDO sed SOC _ 1 s QT ^ DO, 

dt H H B - 2 ° K S0D +1 



ML - -^1 = - J- . 5 . flr-20 . _ f^Mfc (422) 



where S B20 is a function of sediment bottom type (from table 4.1). 

The processes that create sediment oxygen demand are little affected by the 



56 

concentration of oxygen in the overlying water. When oxygen is unavailable to fulfill 
sediment oxygen demand, the demand is exported to the water column. The exported 
demand may be in the form of reduced iron, manganese, methane, or sulfide, which are 
represented in the model as sediment oxygen demand and provides a function which 
computes additional release as oxygen consumption in the sediment is restrained (Cerco and 
Cole, 1995). 

d£O_SOD = _l_ T . 20t _K^_ 

dt H H Bao K SOD + DO 

The relationship between sediment oxygen consumption and sediment oxygen 

demand is represented in Figure 4.5. The SOD is negligible when DO much higher than 

K SOD . When dissolved oxygen is absent from the water column, the maximum oxygen 

demand is released to the water as sediment oxygen demand. 



57 



SOC (Sediment Oxygen Consumption) at sediment column 
SOD (Sediment Oxygen Demand) at water column 

where S B (T= 20°C) = 2 g/m 2 /day 
K S0D = 2 g/m 3 




X 

2-0.5 



■1 



■1.5 

-2 



Dissolved Oxygen (g/m ) 



Figure 4.5 Effect of dissolved oxygen on sediment oxygen consumption and SOD release 



58 
Photosynthesis and Respiration 

The photosynthesis and respiration of phytoplankton can add and deplete significant 
quantities of oxygen from natural systems. The produced oxygen concentration by 
photosynthesis depends on the form of the nitrogen species accessed for phytoplankton 
growth. One mole carbon dioxide can produce one mole oxygen when ammonium is the 
nitrogen source, while one mole carbon dioxide produces 1.3 moles oxygen when nitrate is 
the nitrogen source, according to Morel's equation (1983) 

106Ca +\6NH: + H.PO; + 106ff 2 O -» protoplasm + l06O 2 + 15H + 

(4.24) 

\06CO 2 +\6NO~ +H 2 PO; +\22H 2 + UH^ -> protoplasm + 1380 2 

The simple representation of the respiration process can be used to determine how 
much oxygen would be consumed in the decomposition of a unit mass of organic carbon, 

6C0 2 + 6H 2 ^ C 6 H l2 6 + 60 2 (4.25) 

The dissolved oxygen-to-carbon ratio in respiration can be calculate from this 
equation. 

A oc = ( ^- = 2.67gO/gC (4.26) 

oc 6(12) 

The equation that describes photosynthesis and respiration on dissolved oxygen is: 
^ = [(1.3 - 0.3 • P H )jU a ~ K„ - K M ] • A oc ■ PhyC (4.27) 

where P n is nitrogen preference coefficient; // a is phytoplankton growth rate, which is a 
function of the intensity of light, the availability of nutrients, and the ambient temperature; 
and K as and K^ are respiration and mortality, which are functions of temperature. 

The mass balance equation for dissolved oxygen is written by combining all oxygen 
transformation processes. 



59 



For water column: 



^ = + ^ {D0 -DO)-K D 52 CBOD 

dl te K ' ' H CBm + DO 

14 m H NIT +DO Az 

+[(l3-03.P n )ju a -K (U -K ax }A oc -PhyC 

which include reaeration, oxidation by CBOD, nitrification, sediment oxygen demand, and 

photosynthesis and respiration terms. 
Fore sediment column: 

dDO DO „ _ 64 „ DO XTr , SOC 

= -K D CBOD K NN NH 4 (4.29) 

& H CBOD +DO 14 NN H NIT + DO Az 

which include oxidation by CBOD, nitrification, and sediment oxygen consumption. 
4.5 Effects of Temperature and Light Intensity on Water Quality Processes 

Temperature and light intensity are the most important parameters for transformation 
processes. To achieve a better understanding of water quality processes, it is necessary to 
improve the spatial and temporal variation of these parameters. 
4.5.1 Temperature 

In the nutrient cycle, almost all the reaction parameters are affected by temperature, 
such as zooplankton and phytoplankton growth, respiration and mortality, nitrification, 
denitrification, NH3 stability, mineralization, oxidation, sediment oxygen demand, and 
sorption/desorption reactions. The effect of temperature on reaction rates can be explained 
by the Van't Hoff-Arrhenius equation, as follows: 

^^1 = ^L ( 4.30) 

dt RT 2 

where K is reaction rate at temperature T, AH is the amount of heat required to bring the 

molecules of the reactant to the energy state required for the reaction, and R is universal gas 



60 



constant. 



Integrating Equation (4.44) from temperature T, to T 2 gives: 



K 2 

— - = exp 



AH 
RT{T 2 



(r a -r t ) 



(4.31) 



where K, and K 2 are reaction rate at temperature T, and T 2 , respectively. The temperature 



adjustment function ( 6 = exp 



A// 
RTJ 2 



) is almost constant at the temperature range of 



interest (0° ~ 30°C), ranging from 1.01 to 1.2. This equation can be rearranged into a more 
useful form as: 



K(T) = K(T ref )-0 {T ~ T " /> 



(4.32) 



where K(T) is the reaction coefficient at temperature T, such as /u a (phytoplankton growth 
rate), K^ (phytoplankton respiration rate), K m (phytoplankton mortality rate),//, (zooplankton 
growth rate), K a (zooplankton respiration rate), K^ (zooplankton mortality rate), K M 
(ammonia instability), K om (ammonification rate), K NN (nitrification rate), d un 
(sorption/desorption rate for organic nitrogen), d m (sorption/desorption rate for inorganic 
nitrogen), K 0PM (mineralization rate), d op (sorption/desorption rate for organic phosphorous), 
d ap (sorption/desorption rate for inorganic phosphorous), K D (oxidation rate),or SOD 
(sediment oxygen demand). Each rate term has a unique temperature adjustment function. 
Most models, which use exponential temperature functions, assume a reference 
temperature of 20 °C (Chen and Orlob, 1975; Thomann and Frizpatrick, 1982). Eppley 
(1972) showed that an exponential relationship describes the envelope curve of growth rate 
versus temperature data. The determination was made with a large number of studies, with 



61 
many different species. 

4.5.1 Light intensity 

Light intensity affects the photosynthesis process and thus the phytoplankton growth 
rate. The effects of light intensity on nutrient cycling is often modeled by a light intensity 
limiting function as follow (Steele, 1974) 



/(/) = y-exp 



1-1 



(4.33) 



where / is the light intensity, and Z, is the optimum light intensity for phytoplankton growth 
. According to the Lambert-Beer equation, the light intensity over the water depth is: 

I(z) = I -exp[-K d (PAR)-z] (4.34) 

where I(z) is the light intensity at depth z, and I is the light intensity at the water surface. 
KJPAR) is a function of suspended sediment concentration, algae concentration, and color. 
This value was calculated by the light attenuation model, which will be discussed in the next 
section. 

4.6 Light Attenuation Model 

One of the most important variables controlling phytoplankton photosynthesis is 
"photosynthetically active radiation" (PAR), or light, in the range of wavelengths from 400- 
700nm, which provides the predominant source of energy for autotrophic organisms (Day 
et al., 1989). Absorption and scattering of light by water and dissolved and suspended 
matter determine the quantity and spectral quality of light at a given depth (Jerlov, 1976; 
Prieur and Sathyendranath, 1981), which in tern affect the photosynthesis of aquatic plants. 

One way to develop a light attenuation model is to find a simple regression 
relationship between K d (PAR), calculated from light measurements, and water quality 



62 
measurements collected at the same instant as the K d (PAR) vales. This type of empirical 
model is a simple way to relate light attenuation to water quality, at a certain time. This 
method was used by Mcpherson and Miller (1994) in Tampa Bay and Charlotte Harbor. A 
physics-based light attenuation model was developed as part of the CH3D-EVIS (Sheng et al. 
2001c, Christian and Sheng, 2003) and successfully applied to the Indian River Lagoon. This 
light model is adopted for the Charlotte Harbor study. 

In the light model, the vertical light attenuation coefficient (K d ) is a function of solar 
zenith angle (yu ), scattering (b) and total absorption (a,) (Kirk, 1984) 

1 r , , N -|l/2 

Mo J (4.35) 

and G(p ) = g r fi -g 2 

with g, =0.473 and g 2 =0.218 determined for the mid point of euphotic zone (Kirk, 1984). 
The scattering coefficient, b, is determined only using particles since scattering due 
to particles is much greater than scattering due to water (Gallegos, 1994). Scattering can be 
described as a function of turbidity (Morel and Gentili, 1991) as follow: 

b(A) = (550/ A) ■ [Turbidity] (4.36) 

Total absorption (a,) is partitioned into attributes of water (a w ), phytoplankton (a ph ), 
dissolved color (a d( ), and detritus {a d ). 

a , = «»• + a P h + a dc + a d (4.37) 

The absorption of water (a J can be determined from literature values (Smith and 
Baker, 1981) with 1 nm linearly interpolated from 5 nm found in the literature. Chlorophyll- 
specific absorption (a ph ) is calculated for the model from the linear relationship between the 
maximum absorption and analytical chlorophyll_a concentration (Dixon and Gary, 1999): 



63 

a nh C^) = \ a oh ) ' formalized _ Spectra(A) 

MAX (4.38) 

fa . J =0.0209 -(Chlorophyll a, corrected) 

where (a ph ) MAX is the maximum absorption of chlorophyll-a. To calculate normalized_spectra 
(X), individual spectra are normalized to the maximum absorption (437 - 440 nm), and 
averaged for all samples for an overall normalized spectra. 

The absorption by dissolved color, a dc , for each wavelength in the visible spectrum 
can be found using a negative exponential function (Bricaud et al., 1981) 

a dl U) = 8uo -«p[-** (A -440)] (4.39) 

where g440 is the absorption by dissolved color at 440 nm and s dc is spectral slope. Dixon 
and Gary (1999) calculated empirical absorption at 440 nm and spectral slope as a function 
of color (in PCU) at Charlotte Harbor in the form: 

g 44() = 0.0667 ■ f color] , n = 129, r = 0.9329 

(4.40) 
s Jc = 0.00003 • [color] - 0.0178, n = 129, r = 0.5 1 1 1 

Absorption due to organic and mineral detritus is represented as a function of 
turbidity (Gallegos, 1994): 

a d = G d (A) • [Turbidity] (4.41) 

where turbidity is in NTUs and o d (A) is the wavelength specific absorption cross section of 
turbidity as calculated in: 

<?d U) = <*bi + ^oo • ex P [~ s d ( A ~ 40 °)] (4-42) 

in which a bl is the longwave absorption cross section, a 400 is the maximum detritus 
absorption at 400 nm, and s d is exponential slope. 

The total absorption and scattering are used in Equation (4.49) to calculate the 



64 
vertical attenuation coefficient, KJA), for each wavelength depending on color, 
chlorophylls , and turbidity. This KJA) value and incident irradiation EJA) can be used in 
the equation for calculating irradiance E Z (A) at the reference depth, z r : 

E z (A) = E (A)-exp[-K d (A)-z r ] (4.43) 

The spectrum of incident sunlight data from table F-200 in Weast (1977) is used for 
incident spectral information, EJA), as in Gallegos's work (1994) . These data are shown 
in Table 4.2. 



Table 4.2 The spectrum of incident sun 



X (nm) 



400 

405 

410 

415 

420 

425 

430 

435 

440 

445 

450 

455 

460 

465 

470 

475 

480 

485 

490 

495 

500 



X (nm) 



4.780 




5.568 


505 


6.003 


510 


6.052 


515 


6.135 


520 


6.017 


525 


5.893 


530 


5.940 


535 


6.659 


540 


7.152 


545 


7.548 


550 


7.826 


555 


7.947 


560 


7.963 


565 


7.990 


570 


8.119 


575 


8.324 


580 


8.014 


585 


7.990 


590 


8.113 


595 


8.119 


600 



ight data (Gallegos, 1994) 



X (nm) 



8.108 


505 


8.026 


510 


7.894 


515 


7.970 


520 


8.130 


525 


8.163 


530 


8.133 


535 


8.051 


540 


7.993 


545 


7.993 


550 


7.982 


555 


7.937 


560 


8.055 


565 


8.160 


570 


8.265 


575 


8.318 


580 


8.375 


585 


8.387 


590 


8.369 


595 


8.359 


600 



8.332 
8.340 
8.323 
8.305 
8.289 
8.271 
8.267 
8.263 
8.239 
8.213 
8.207 
8.201 
8.180 
8.157 
8.136 
8.114 
8.106 
8.089 
8.052 
8.013 



EJA) and E,(A) are integrated over the visible spectrum to get PAR for the incident 



PAR and PAR Z for the calculated PAR at the reference depth. The spectrally sensitive 



65 

attenuation coefficient K d {PAR) is calculated from these integrated values by using the 

rearranged form of the Lambert-Beer equation: 



~ 1 < 

K d (PAR) = —\n 

Z. 



PAR Z 



V PAR o j 



(4.44) 



This Kj(PAR) will be used in the model to calculate light levels throughout the 

water column as a function of measured incident light intensity. Dixon and Gray (1999) 
calibrated the light attenuation model with measured data and applied the model to determine 
the light requirements for seagrasses of the Charlotte Harbor estuarine system. The results 
of the optical model are very good, with mean percentage agreement between modeled and 
observed K tl of 1 10%. Using water quality data from all stations as inputs, Dixon and Gray 
(1999) determined that chlorophyll, color, and turbidity account for 4%, 66%, and 31% of 
water column light attenuation, respectively. The maximum annual average chlorophyll 
contribution is 6%, with an individual maximum of 18% during a phytoplankton bloom. 
Color dominated water column attenuation, ranging from 40% to 78%. Stations in the lower 
Harbor and southern sites showed increased attenuation due to turbidity, compared to the 
upper Harbor site (Dixon and Gary, 1999). For the southern sites, a much larger portion of 
light attenuation is produced by turbidity, up to 55% for the station near Captiva pass. 

To provide a light attenuation component which can be coupled with water quality, 
hydrodynamic, and sediment model components in the Charlotte Harbor estuarine system. 
A stand alone light model needs to be developed and calibrated with measured data. The 
stand alone light model has been calibrated by finding best fit between simulated and 
measured light attenuation with various coefficient sets. 

For calibration of the light model, a M , a 400 , s d , and s are allowed to vary within 



66 
certain ranges. The predicted kJPAR) values for the stand alone light model are compared 
to the corresponding kJPAR) values from data provided by the SFWMD. If the predicted 
values do not match data well, then attempts are made to try to find model coefficients to 
produce better fit. In this case, the model is run and the coefficients are varied between 
maximum and minimum literature values specified in Table4.3 (Christian and Sheng, 2003). 
More than one thousand runs of light model simulations were conducted with this data set. 
The RMS errors were calculated for each run to test model performance of each set of light 
model coefficients. The root mean square error is an indication of the average discrepancy 
between observations and model results. In addition to root mean square error, model 
calibration was assessed via plots of model output and observation. Scatter plots of model 
output and observed data provide an indication of overall model performance. The best 
RMS errors for each of the tests is 0.61 1 1 — ' with best fit coefficients shown in Table 4.4 
Dixon and Gray (1999) found the coefficients in Table 4.5 provided the best fit to all the data 
they examined. They applied a 400 and s y as a function of color. The RMS error with this 
coefficient is 0.65 11m" 1 . Even though, they reduced adjustable coefficients with relationship 
of a 400 ands y with color, the RMS error is greater than that for adjusting all four coefficients. 
In this study, the best fit coefficients were used for simulating light attenuation. Figure 4.6 
contains calibration period scattering plots with best fit of coefficient sets. The location of 
circles indicate the correlation between model predictions and observed data. A perfect 
match between model and observed data is indicated by the diagonal line on each graph. 
Circle above the line shows over prediction, while circles below the line indicate that under 
prediction. 



67 



Table 4.3 Coefficient ranges for use in stand along light model. 



Coefficients 



Minimum value 



o,. 



o_ 



41)11 



0.004 
0.255 
0.011 
0.011 



Maximum value 



0.090 
0.600 
0.019 
0.019 



Table 4.4 Best fit light model coefficients for Charlotte Harbor estuarine system. 



Coefficient 



<Ju 



a A 



401) 



Value 



0.01 m-'NTTJ- 1 

0.59 m'NTU'' 

0.014 nm 1 

0.013 nm' 1 



Table 4.5 Dixon and Gray's model coefficients for the Charlotte Harbor estuarine system. 


Coefficient 


Value 


°u 


0.064 m'NTU- 1 


°400 


0.0014xcolor+0.0731 m'NTU' 1 


*i 


0.01125 nm 1 


S v 


0.00003xcolor+0.0178 nm ' 



68 



T3 




■o 



cr 
< 

Q. 



5 - 



4 - 



3 - 



2 - 



1 - 







• PARPS model 




• 

* 

— • 



_ 





_ 





- 

^m 





— 










_ 

• 






• 

— ■ 




• 


• / 



_ # 






• • •+' • 


•*''• ••• • 


• %T\0 • # 


• 
_ * 




'**' I I I I J 1 1 I I I I I I 1 I I I I 1 I I I I 1 I I I I 



2 3 4 

Kp AR observed 



Figure 4.6 The scatter plots for Kd(PAR) during calibration period with best fit coefficients. 



69 

The water quality and sediment model outputs needed for the light model are 

turbidity in NTUs, chlorophyll_a concentration in Hg/L, and color in Pt. Units. To calculate 

solar angle, this model needs the latitude and simulation day and time. Because the sediment 

model simulates the total suspended solids (TSS), it is necessary to develop a regression 

between TSS (mg/L) and turbidity (NTU) as follow: 

Turbidity = 0.2677 x TSS + 0.9665 for 1996 data 

Turbidity = -0.0 1 53 x TSS + 2.602 for 2000 data 

The stand alone light model was tested, calibrated, and then coupled with the models 
of hydrodynamics, sediment and water quality. 

4.7 Model Parameters and Calibration Procedures 

Reaction terms described in the previous section contain many model parameters 
which must be determined before using the model. The determination of model parameters 
is generally very difficult because they depend on many physical and biochemical factors, 
such as the location of the estuary, temperature and tidal variation, and point or non-point 
source loadings of nutrients and other chemical materials. In practice, parameters are 
selected from a range of feasible values, tested in the model, and adjusted until an optimal 
agreement between simulated and measured values is obtained. Two ways to determine 
feasible ranges of model parameters are field observation and laboratory experimentation. 
When field observations and laboratory experimentation are not available, the feasible ranges 
are obtained from literature or previous modeling studies. A list of the important kinetic 
parameters and their literature values are given in Table 4.6 ~ 4.11. 



70 



Table 4.6 Temperature adjustment functions for water quality parameters 



Paramete 

r 


description 


Unit 


literature 
value 


Reference 


(»*) 


temperature function for 
phytoplankton growth 


- 


1.01 ~ 1.2 
1.09 
1.08 


Di Toro et al. (1980) 
Pribble et al.(1997) 
Sheng etal. (2001) 


K») 


temperature function for 
phytoplankton 
respiration and mortality 


- 


1.045 

1.05 

1.08 


Ambrose (1991) 
Pribble etal.(1997) 
Sheng et al. (2001) 


(»«) 


temperature function for 
zooplankton growth, 
respiration and mortality 


- 


1.01 - 1.2 
1.04 


Di Toro etal. (1980) 
Sheng et al. (2001) 


\"oNM ) 


temperature function for 
ammonification 


- 


1.0- 1.04 
1.07 
1.02 


Bowie et al. (1985) 
Pribble etal.(1997) 
Sheng et al. (2001) 


ifim) 


temperature function for 
nitrification 


- 


1.02-1.08 
1.08 
1.08 


Bowie et al. (1985) 
Pribble et al.(1997) 
Sheng et al. (2001) 


(<V) 


temperature function for 
denitrification 


- 


1.02- 1.09 

1.04 
1.045 


Bowie etal. (1985) 
Pribble et al.(1997) 
Sheng et al. (2001) 


M 


temperature function for 
ammonia instability 


- 


1.08 


Sheng etal. (2001) 


( OPM ) 


temperature function for 
mineralization 


- 


1.08 


Sheng etal. (2001) 


{&s/d) 


temperature function for 
sorpti on/desorpti on 


- 


1.08 


Sheng etal. (2001) 


M 


temperature function for 
oxidation 


- 


1.02-1.15 
1.08 


Bowie et al. (1985) 
Sheng etal. (2001) 


(8 SOD ) 


temperature function for 
sediment oxygen 
demand 


- 


1.045 
1.08 


Bowie et al. (1985) 
Sheng et al. (2001) 



71 



Table 4.7 Water quality parameters related to conversion rate 




Paramete 


description 


Unit 


literature 


Reference 


r 






value 




Ajc 


Phytoplankton Carbon 


go 2 /gc 


2.67 


Ambrose(1991) 




/ Oxygen rate 




2.67 
2.67 


Cerco and Thomas 
(1995) 
Shengetal. (2001) 


ChlaC 


Phytoplankton Carbon 


gC/gChl 


10- 112 


Bowie et al. (1985) 




/ Chlorophyll_a rate 


a 


100 

60 

50 


Pribble et al.(1997) 
Cerco and Thomas 
(1995) 
Shengetal. (2001) 


A:/V 


Phytoplankton Carbon 


gN/gC 


0.05 -0.43 


Jorgensen (1976) 




/ Nitrogen rate 




0.15 
0.167 
0.15 


Pribble et al.(1997) 

Cerco and Thomas 

(1995) 

Sheng et al. (2001) 


\p 


Phytoplankton Carbon 


gP/gC 


0.005-0.03 


Jorgensen (1976) 




/ Phosphorous rate 




0.027 
0.025 


Cerco and Thomas 
(1995) 
Shengetal. (2001) 



Table 4.8 Water quality parameters related to phytoplankton and zooplankton 


Paramete 
r 


description 


Unit 


literature 

value 


Reference 


ta>L 


maximum phytoplankton 
growth rate 


1/day 


0.2-8 
2.25-2.5 
1.06-2.68 


Bowie et al. (1985) 
Cerco and Thomas 
(1995) 
Shengetal. (2001) 


H n 


Nitrogen half saturation 
rate for phytoplankton 
uptake 


ng/i 


1.5-400 
0.5 

1 
10 


Bowie etal. (1985) 

Pribble et al.(1997) 

Cerco and Thomas 

(1995) 

Sheng et al. (2001) 


H P 


Phosphorous half 
saturation rate for 
phytoplankton uptake 


Mfl 


1. - 105 
1 

1 

2-4 


Bowie etal. (1985) 
Pribble et al.( 1997) 
Cerco and Thomas 
(1995) 
Shengetal. (2001) 



72 



opt 


optimum light intensity 
for phytoplankton 
growth 


ly/day 


225-600 
300 


Canale et al. (1976) 
Sheng et al. (2001) 


K M 


phytoplankton 
respiration rate 


1/day 


0.02- 0.24 
0.03- 0.09 
0.03- 0.05 


Jorgenson (1976) 

Cerco and Thomas 

(1995) 

Sheng etal. (2001) 


K us 


phytoplankton non- 
predator mortality 


1/day 


0.01-0.22 
0.03- 0.09 
0.02- 0.06 


Jorgenson (1976) 

Cerco and Thomas 

(1995) 

Sheng etal. (2001) 


ws P*y 


phytoplankton settling 
rate 


m/da 

y 


0. - 3. 

0. - 0.25 

0.05-0.1 


Bowie etal. (1985) 

Cerco and Thomas 

(1995) 

Sheng et al. (2001) 


H P ky 


phytoplankton half 
saturation rate for 
zooplankton uptake 


ng/i 


200-2000 
800-1200 


Bowie et al. (1985) 
Sheng et al. (2001) 


TrS P»y 


phytoplankton threshold 
for zooplankton uptake 


ran 


1-200 
200 


Bowie et al. (1985) 
Sheng et al. (2001) 


V * /max 


maximum zooplankton 
growth rate 


1/day 


0.15-0.5 
0.18-0.2 


Bowie et al. (1985) 
Sheng et al. (2001) 


K„ 


zooplankton respiration 
rate 


1/day 


0.003-0.07 

5 

0.01 


Bowie etal. (1985) 
Sheng et al. (2001) 


K a 


zooplankton non- 
predator mortality 


1/day 


0.001-0.36 
0.015-0.05 

5 


Jorgensen (1976) 
Sheng et al. (2001) 



Table 4.9 Water quality parameters in the nitrogen dynamics 




Paramete 
r 


description 


Unit 


literature 

value 


Reference 


K-onm 


Ammonification rate 


1/day 


0.001-0.4 

0.1 

0.015 

0.01 


Bowie et al. (1985) 

Pribble et al. (1997) 

Cerco and Thomas 

(1995) 

Sheng etal. (2001) 



73 



K NN 


nitrification rate 


1/day 


0.004-0.11 

0.08 

0.07 
0.01-0.02 


Bowie et al. (1985) 
Pribble et al.(1997) 
Cerco and Thomas 
(1995) 
Shengetal. (2001) 


H nil 


DO saturation rate for 
nitrification 


mg/1 


0.1-2.0 

2 
2 


Ambrose (1994) 
Pribble et al.(1997) 
Sheng et al. (2001) 


K DN 


denitrification rate 


1/day 


0.02- 1.0 
0.09 
0.09 


Bowie et al. (1985) 
Pribble et al.(1997) 
Sheng et al. (2001) 


H noi 


DO saturation rate for 
denitrification 


mg/1 


0.- 1.0 
0.1 


Bowie et al. (1985) 
Shengetal. (2001) 


P m 


partition coefficient of 
PON/SON 


- 


l.E-5 
1.E-6-9E-6 


Simon (1989) 
Sheng et al. (2001) 


Pan 


partition coefficient of 
PIN/NH4 




5.E-6-1.E- 

5 
3.E-5-4.E- 

3 


Simon (1989) 
Sheng et al. (2001) 


d on 


sorption/desorption rate 
for SON/PON 


1/day 


0.02 

0.08 

0.01-0.02 


Bowie etal. (1985) 
Cerco and Thomas 
(1995) 
Shengetal. (2001) 


d an 


sorption/desorption rate 
for PIN/NH4 


1/day 


0.01 


Shengetal. (2001) 


^■PDN 


preference partition 
coefficient of mortality 
for SON / PON 


- 


0.5 
0.5 


Cerco and Thomas 

(1995) 

Sheng etal. (2001) 



Table 4.10 Water quality parameters in 


the phos] 


Dhorous dynamics 


Paramete 
r 


description 


Unit 


literature 

value 


Reference 


"~OPM 


mineralization rate 


1/day 


0.001 - 0.6 

2.27 

0.1 

0.1 


Bowie et al. (1985) 

Pribble et al. (1997) 

Cerco and Thomas 

(1995) 

Sheng et al. (2001) 


Pop 


partition coefficient of 
POP/SOP 


- 


8.E-6-1.E- 

4 


Shengetal. (2001) 



74 



% 


partition coefficient of 
PIP/SRP 


- 


1.E-6-6E- 
4 


Sheng et al. (2001) 


< p 


sorption/desorption rate 
for SOP/POP 


1/day 


0.08 
0.01 


Cerco and Thomas 

(1995) 

Sheng etal. (2001) 


d ap 


sorption/desorption rate 
for PIP/SRP 


1/day 


0.01 


Sheng etal. (2001) 


PpDP 


preference partition 
coefficient of mortality 
for SOP / POP 


- 


0.5 
0.5 


Cerco and Thomas 

(1995) 

Sheng etal. (2001) 



Table 4.1 1 Water quality parameters in 


the oxygen balance 




Paramete 
r 


description 


Unit 


literature 
value 


Reference 


K D 


oxidation rate 


1/day 


0.02 - 0.6 
0.05 


Bowie et al. (1985) 
Sheng et al. (2001) 


" CBOD 


DO half saturation rate 
for oxidation 


mg/1 


1.5-400 
0.5 

1 
10 


Bowie et al. (1985) 

Pribbleetal.(1997) 

Cerco and Thomas 

(1995) 

Sheng et al. (2001) 


J®CBOD 


partition coefficient of 

particular/dissolved 

CBOD 


- 


0.5 
0.3-0.5 


Bowie et al. (1985) 
Sheng et al. (2001) 


SOD 


Sediment oxygen 
demand 


g0 2 /m 

2 -day 


0.02-10. 
0.0 - 10.7 


Thomann (1972) 
Bowie et al. (1985) 


" SOD 


DO half saturation rate 
for sediment oxygen 
demand 


mg/1 


0.01-0.22 
0.03- 0.09 
0.02- 0.06 


Jorgenson (1979) 

Cerco and Thomas 

(1995) 

Sheng et al. (2001) 


K AE 


reaeration rate 


1/day 


0. - 3. 

0. - 0.25 

0.05-0.1 


Bowie etal. (1985) 

Cerco and Thomas 

(1995) 

Sheng etal. (2001) 



75 
Model calibration is the first stage testing or tuning of the model to a field data not 
used in the original construction of the model. Such tuning is to include consistent and 
rational set of theoretically defensible parameters and inputs (Thomann, 1992). Proper 
calibration of the water quality model requires having accurate representation of the inflow 
and loads of nutrients into the water body and selecting appropriate model parameters. 

The reaction equation shown in previous section is a function of its concentration and 
the water quality parameters connected to it by the indicated processes. Within each reaction 
equation, there are numerous kinetic parameters and additional parameters. Water quality 
model in CH3D-EVIS consists of 13 state variable equations with over 40 interrelated 
parameters. The interactions of water quality model parameters and the constituent equations 
as shown in Table 4.12 clearly indicate the complexity of the calibrating these types of 
models. The column of Table 4.12 show that each modeled constituent equation contains 
between 2 and 15 different water quality parameters. According to each row, each parameter 
can be found in up to 10 different constituent equations. Therefore, changing one parameter 
to improve the calibration of one constituent will simultaneously affect many other 
constituents. Traditionally, calibration of water quality models has been performed manually 
using a trial-and-error parameter adjustment procedure. The process of manual calibration 
depending on the number of model parameters and the degree of parameter interaction. With 
complexity of water quality model, the traditional process is a very tedious and time 
consuming task. It is necessary to apply more systematic and efficient calibration procedure 
for reducing calibration time and effort. 

Based on the cascading effect of adjusting interrelated parameters, the efficient 
calibration of the water quality model should begin with the parameters that affect the more 



76 
constituents and the more sensitive parameters. First of all, each water quality parameter 

can be ranked with these sensitivity and relativity as parameterization. According to this 

order, high ranked parameters will be adjusted to reproduce major pattern of all constituents, 

and then lower ranked parameters will be calibrated for detail characteristics of local 

constituents. This procedure will reduce numerous calibration efforts and increase efficiency 

and effectiveness. 

The calibration procedure involves optimization of numerical measures (objective 

functions) that compare observations of the state of the system with corresponding simulated 

values. The most commonly used objective function adopted in calibration is the root mean 

square errors between the observed and simulated model response. The root mean square 

error is an indication of the average discrepancy between observations and model results. 

It is computed as follow: 



T(o-p) 2 

RMS = J^- (4.45) 

n 
where RMS is root mean square error 

O is observation 

P is model prediction 

n is number of observation 

In addition with root mean square error, model calibration was assessed via plots of 
model output and observation with correlation coefficient R 2 . Scatter plots of model output 
and observed data provide an indication of overall model performance. The correlation 
coefficient is defined by 



77 



ssl C£0P-nOP) : 



R* = 



(4.46) 



SS^-SS^ (£0 2 -n.0 2 ).(£P 2 -n-P 2 ) 
SS xx =Z(O l -d) 2 =Y J 2 -n.0 2 
SS^ZiP.-Pf-TP'-n-P 2 
SSv^iOt-OM-P^J^O-P-nO-P 

The process of model calibration is illustrated in Figure 4.7. In systematic calibration 
procedure, parameters are adjusted according to order of sensitivity and relativity according 
to parameterization for optimization of certain criteria (objective functions) that measure the 
goodness-of-fit of the simulation model. The process is repeated until a specified stopping 
criterion is satisfied. 

Formation of a proper framework for systematic calibration involves the following 
key elements: 

• Sensitivity analysis 

• Model parameterization and choice of calibration parameters 

• Specification of calibration criteria 

The best way to calibrate water quality parameters is the automatic calibration, in 
which parameters are adjusted automatically according to a specific search scheme for 
optimization of certain calibration criteria. The process is repeated until a specified stopping 
criterion is satisfied. In this study, automatic calibration procedures have been developed 
and tested, using a Gauss-Newton method. 

To test the automatic calibration procedure of the CH3D water quality model, a 1996 
baseline run was adopted as representative of the true and valid field condition. All 
parameters from baseline run were considered representative of conditions that could 
hypothetically exist in the field. A small number of these parameters were perturbed slightly, 
and the automatic calibration procedure was employed. If the procedure is correct, these 



78 
perturbed parameters should converge on pre-perturbed values. The results of the test show 
that the changed parameters did converge on pre-perturbed baseline values, with the Gauss- 
Newton method. The method is therefore considered valid and accurate. 

In real case, a trial run was conducted by using all the measured data to find best fit 
parameters. The calibration procedure continuously failed as calibration parameters were 
automatically adjusted outside the upper and lower bounds. The accuracy of the model 
calibration generated by this procedure, relies heavily on the quality and quantity of field 
observation, the model structure, and the nature of the system. Limited data, and 
uncertainties in water quality processes caused the procedure to fail to generated a valid 
calibration condition. To apply an automatic calibration procedure to this, or other water 
quality models, further investigation is required. These investigations should be focus on the 
optimization algorithm and refinement of the water quality processes. 



79 



Table 4.12 The relationship 


> between water quality parameters and model constituents 




CA CZ PIN NH4 N03 SON PON PIP SRP SOP POP DO CBOD 


#of 
Eqns 


AGRM 
HALFN 
HALFP 
KAEX 
KAS 
WAS 
HALFA 
TRESHA 
ZGRM 
KZEX 
KZS 
SONM 
NITR 
DENR 
PCON 
DRON 
KPDN 
PCAN 
DRAN 
SOPM 
PCOP 
DROP 
KPDP 
PCIP 
DRIP 
SODM 
FD5 
AKD 
AOC 
CHLAC 
CAN 
ACP 
TOPT 
OPTL 
AKNIT 
AKDEN 
AKBOD 
AKSOD 
AKAIR 


1 i ; 


! l 


1 I 


l i 1 ; 


5 
1 
1 

5 

7 

1 

2 

2 

2 

3 

6 

2 

3 

2 

2 

2 

2 

2 

2 

2 

2 

2 

2 

2 

2 

1 

2 

2 

10 

2 

6 

5 

1 

1 

3 

2 

2 

1 

1 


i i 

1"! i 












i i l 




l ! 


l 


l l 


l ; 


i ■ ' : 


; l ! l 


l 


: 


: : 


: i : : 




i I i i 








i i ; 






1 ; i ! 


! 


: i i : 






i l 






i 




i i i 


...........J 

l i 

1 : 


1 


L.1...L. ...L 


l 


l 




i i 
i 




l 
l 




i 




1 i 1 








! 






1 1 














1 i 


1 i 










1 


i 1 




1 


l i 














1 1 














1 


l 






: 




1 


l 












1 


l 


: 










l 


1 




i 




l 


1 






: 




"T'"? 






1 




1 


l 


i 


l 

l 


1 
j- f j- 1 

l 1 ! 
l i 1 1 


1 1 ! 


i 11 ii 


11 = 1 


i 






111 





1 j 
l j 








•1 i ; 
! l ; 




t "T" "I 

t t m 


i i 


1 ] 






! ! I ! ! i i i ! i'T" 


#of 
Parameters 


15 82 10 7872887 14 7 


103 



80 



INITIAL VALUE for WQ parameter 



I 



Execute WQ model 



I 



Adjust WQ river boundary Condition 



I 



Execute WQ model 




change i parameter for sensitivity test 



I 



Execute WQ model 




Sensitivity analysis : 
check RMS differneces between base run and each sensitivity test 



I 



parameterization : 
make order with sensitivity result and relativity table 



1 



Adjust i ranked parameter 



Execute WQ model 





Calculate statistics : scatter plot, RMS, R2 
Plot time series for each species with measured datat 



Figure 4.7 Systematic calibration procedure 



CHAPTER 5 
APPLICATION OF CIRCULATION AND TRANSPORT MODEL 

Circulation in Charlotte Harbor is driven by tide, wind, and density gradient. The 
numerical model of circulation and transport described in Chapter 3 was applied to the 
Charlotte Harbor estuarine system, a large shallow estuary in southwest Florida. Model 
applications include 1) the simulation of three-dimensional circulation during summer 1986 
2)one year simulation of flow, salinity, temperature, and sediment transport in 2000 and 3) 
simulation of the impact of removal of Sanibel Causeway on the circulation in San Carlos 
Bay and Pine Island Sound 

A major purpose of the first two simulations is the calibration and validation of the 
Charlotte Harbor circulation and transport model using field data obtained by USGS, NOAA, 
SWFWMD, and SFWMD in Summer 1986 and January to December in 2000. During the 
calibration process, a few model coefficients and/or boundary/initial conditions are adjusted 
to produce more accurate overall simulation of observed data collected in July 1986 by 
USGS (Hammet, 1992; Stoker, 1992; Goodwin, 1996) covered the entire Charlotte Harbor 
estuarine system and include water level, horizontal currents and salinity. Hence July 1986 
data were used for short-term calibration. To supplement the short-term calibration of July 
1986, the long-term model calibration was conducted using one year water level and salinity 
data in Caloosahatchee River during 2000. 



81 



82 
5.1 A High-Resolution Curvilinear Grid for Charlotte Harbor Estuarine System 

To procedure a successful numerical simulation of the Charlotte Harbor estuarine 
system, it is necessary to design a numerical grid which represents the dominant geographic/ 
bathymetric features with sufficient spatial resolution. Since CH3D use a boundary-fitted 
curvilinear grid in the horizontal directions, it is possible to align the grid lines to coincide 
with shorelines, causeways, and bridges. The Charlotte Harbor estuarine system 
encompasses about 735 km 2 on Florida's southwest coast. The model domain includes all 
the sub-basins (Upper Charlotte Harbor, Lower Charlotte Harbor, Pine Island Sound, 
Matlatcha Pass, and San Carlos Bay), the major tributaries (Peace, Myakka, and 
Caloosahatchee up to the Franklin Dam), Estero Bay, and some offshore water. The model 
should be run with a spatial grid sufficiently fine to resolve the Sanibel Causeway. The 
eastern boundary includes many of the major rivers with specified flow rates, while the 
western boundaries and the western portion of north and south boundaries are open tidal 
boundaries where tidal elevations are prescribed. The interior of the model domain is 
represented with a boundary-fitted grid which is non-orthogonal but as orthogonal as 
possible, i.e. the grid analysis are usually between 60° and 120°. 

The final boundary fitted grid used for the three-dimensional curvilinear-grid model 
(CH3D) of Charlotte Harbor estuarine system was generated using a grid generation program 
originally developed by Thompson et al. (1985) and further enhanced for this study. This 
grid (Figure 5.1) contains 92 x 129 horizontal cells and eight vertical layers, with a total of 
1 1 648 horizontal grid cells which include 5367 water cells and 628 1 land cells. Grid spacing 
varies from 40 to 2876 meters (average 598 meters). 

The bathymetry of the Charlotte Harbor estuarine system and the nearshore region 



83 
of the Gulf of Mexico is derived from data obtained from the Geophysical Data System of 
National Geophysical Data Center. Bathymetry of navigation channels in San Carlos Bay 
and the vicinity of Sanibel Causeway are based on the recent survey data provided by Lee 
County in December 1 999. Bathymetry for Peace River and the upper Charlotte Harbor area 
was collected by SWFWMD and that for Caloosahatchee River was collected by SFWMD. 
All bathymetric data were converted to NAVD88 to unify vertical datum level (Appendix 
H). Charlotte Harbor grid bathymetry (Figure 5.2) was developed using all these bathymetric 
data. While an inverse distance interpolation followed by simple smoothing scheme was the 
primary method for determining bathymetry in the study area, several areas were added and 
the bathymetry further adjusted after the interpolation and smoothing was performed to 
ensure the proper passage of flow through a navigation channels. 



84 



C 

o 

z 





\ 








J? 






? 




A V 




o 




h 




^^fc^ffiE^ 




o 




Sv»T- 








o 




AVC*^ 








in 




\yvow**' 






Y ,--- 


r- 












co 






v" 






CM 




















j i^^H 




O 












o 












o 












o 


^m \s&& 










LO 












CD 








SgS&SRi s^^W 


^■rstLrf/fiv 


C\J 


w 






§11 \1 




o 








illlllll 


WELug; \ L 


o 








?Crt>, Vv^H| 




o 










■~''s$5r>SJtit. 4$K2> 


LO 


— 










C\J 












en 










SSsc$v---" 


CM 








v ' '' '~^*^$x?xX& 








il 






o$x38BKv\S^^^^S^Stfcj j * t -- > 












v^^sSSSKS^^^^ww^ 


o 






o 












o 










s/?c^>otjo?v^0cfilS^^B^ 


o 
o 

CO 










liPH 


CM 


lilt 


i i 


1 


1 1 _J_S^ 


r 1 i i i i 1 i i i i 



350000 



375000 400000 

Easting (m) 



425000 



Figure 5.1 Boundary-fitted grid (92 x 129) used for numerical simulation for Charlotte 
Harbor estuarine system. 



85 



£ H>£J 



r 




J I I L 



J i i i i 



350000 



375000 400000 

Easting (m) 



425000 



2000 

1900 

1800 

1700 

1600 

1500 

1400 

1300 

1200 

1100 

1000 

900 

800 

700 

600 

500 

400 

300 

200 

100 





Figure 5.2 Bathymetry in boundary-fitted grid for Charlotte Harbor estuarine system (92 
x 129). 



86 
5.2 Forcing Mechanism and Boundary Conditions 

The forcing mechanisms for the hydrodynamic model include tide, wind, and density 
gradients. To incorporate these mechanisms in the model simulations, tides and oceanic 
salinity are specified along the offshore boundary, wind is specified at the air-sea interface, 
and river discharges are specified at the river boundaries. 

The boundary between the Charlotte Harbor estuarine system and the Gulf of Mexico 
extends about 64 km from Gasparilla Pass on the north to San Carlos Pass on the south. One 
of the most important features is the similarity between the tidal curves at Venice and Cayo 
Costa which are located at northern end and near Boca Grande Pass (Figure 1.1). The 
phasing of the Gulf tides is nearly synchronous along the boundary of the estuarine system 
between Venice and Naples, which are located at the northern and southern ends. 

For the 2000 simulation, tidal forcing along the Gulf of Mexico boundary was 
prescribed with water level data measured at the Naples station by the NOAA CO-OPS. For 
the 1986 simulation, water level data at Venice measured by USGS were used to provide the 
tidal boundary condition, since there was no data from the Naples station in this period. To 
unify the datum level with the bathymetry, these water level data were converted to NAVD88 
datum level. 

Discharge and runoff boundary conditions were imposed using the daily measured 
discharge data described in Table 4.1. Because no flow data were available for the Estero 
Bay and Caloosahatchee River except at S-79 in 1986, the discharges were assumed to be 
one half and a quarter of the Myakka River discharges, respectively. 

The surface wind boundary condition is supplied using hourly wind magnitude and 
direction data collected at three stations, which are the National Data Buoy Center C-MAN 



87 
stations at Venice (27.07 N, 82.45 W ), NOAA CO-OPS stations at Fort Myers (26° 36' N, 
81° 52' W), and Naples (26.13 N, 81.807 W ). For the 2000 simulation, hourly wind data 
from all three stations are available, while only the Venice and Fort Myers stations are 
available for 1986 simulation. The hourly wind magnitude and direction are converted into 
x-(EastAVest) and y- (North/South) velocity components using Garrat's formula (Garrat, 
1977). These values are then interpolated to the entire computational grid, with a weighting 
function inversely proportional to the square of distances. Figures 5.3 shows the water level 
at Venice station for tidal boundary condition and the river flows at river boundaries in 1986. 
Figure 5.4 shows wind speed and direction at Venice and Fort Myers station in 1986. Figure 
5.5 shows the water level at Naples station for tidal boundaries and the river flows at river 
boundaries in 2000. Figure 5.6 shows wind speed and direction at Venice, Fort Myers and 

Naples in 2000. 

Table 5.1 Descriptions of 1986 and 2000 river boundary conditions for Charlotte Harbor 



wluu "" v "J" 
Station 


Name 


Agency 


Latitude 


Longitude 


drainage 


perio 








degree W 


degree N 


area:mi 2 


d 




Myakka River 












02298830 


near Sarasota 


USGS 


27°14'25" 


82°18'50" 


229.0 


daily 


02299120 


at Deer Prairie Slough 


USGS 


27°08'06" 


82°15'24" 


- 


daily 


02299410 


at Big Slough Canal 


USGS 


27°11'35" 


82°08'40" 


36.5 


daily 




Peace River 












02296750 


at Arcade 


USGS 


27°13'19" 


81 °52'34" 


1 ,376.0 


daily 


02297100 


at Joshua Creek 


USGS 


27°09'59" 


81°52'47" 


132.0 


daily 


02297310 


at Horse Creek 


USGS 


27°11'57" 


81°59'19" 


218.0 


daily 


02298202 


Shell Creek 


USGS 


26°59'04" 


81 °56'09" 


373.0 


daily 




in Caloosahatchee River 












02292900 


S79 spillway 


USGS 


26°43'25" 


81°41'55" 


- 


daily 


02293214 


Cape Coral 


USGS 


26°38'10" 


81 °5548" 


- 


daily 


02293230 


Whiskey Creek 


USGS 


26°34'29" 


81 °53'29" 


" 


daily 




in Estero Bay 












02291500 


Imperial 


USGS 


26°20'07" 


81°44'59" 


65.0 


daily 


02291524 


Spring Creek 


USGS 


26°21'42" 


81 °47'27" 


- 


daily 


02291580 


Estero River 


USGS 


26°26'30" 


81°47'45" 


- 


daily 


02291673 


Mullock & Henry Creek 


USGS 


26°30'19" 


81°51'00" 


~ 


daily 



88 
Baroclinic simulations require realistic initial salinity and temperature fields in order 
to minimize the effects of the initial conditions on the solution, and to minimize the "spin- 
up" time. Salinity along the offshore boundary is assumed to be approximately constant at 
36.5 ppt. Salinity at the river boundaries was set to ppt. The near bottom salinity data 
measured at the nine USGS stations in 1986 were used to determine the initial salinity at the 
corresponding grid cells. These salinity data were interpolated over the entire computational 
grid, with a weighting function inversely proportional to the square of the distance to the 
three nearest stations. Initial temperature field for 2000 simulation was interpolated using 
water temperature data at Fort Myers and Naples stations measured by NOAA CO-OPS and 
Venice stations measured by National Data Buoy Center (NDBC). Air temperature data at 
the same locations were used to calculate heat flux for air-sea interface (Figure 5.7). These 
air temperature data were also interpolated over the study area. 



89 




160 



1 80 200 

Julian Day 



220 



300 r 



Caloosahatchee River 
Peace River 
Shell Creek 
Myakka River 




160 



180 200 

Julian Day 



220 



Figure 5.3 Tidal forcing and river discharges for 1986 simulation of Charlotte Harbor 
circulation. 



at Venice 




90 



I ; 



' I I 



I i 



150 160 170 180 190 200 

Julian Day 



210 



220 



230 



at Fort Myers 




j_ 



150 160 170 180 190 200 

Julian Day 



210 



220 



230 



Figure 5.4 Wind velocity for 1986 simulation of Charlotte Harbor circulation. 



91 




Julian Day 



200 r 



100 



Caloosahatchee River 
Peace River 
Shell Creek 
Myakka River 




200 

Julian Day 



300 



Figure 5.5 Tidal forcing and river discharges for 2000 simulation of Charlotte Harbor 
circulation 



Wind speed and direction at Venice 



92 




36850 



36875 



36900 



36925 



36950 



Wind speed and direction at Fort Myers 




36850 



36875 



36900 



36925 



36950 



Wind speed and direction at Naples 




36660 



36720 



36680 36700 

Julian Day since 1 900 

Figure 5.6 Wind velocity for 2000 simulation of Charlotte Harbor circulation. 



93 



at Venice 




100 



200 

Julian Day 



300 



at Fort Myers 




200 

Julian Day 




100 200 300 

Julian Day 

Figure 5.7 Air temperature for 2000 simulation of Charlotte Harbor circulation. 



94 
5.3 Simulations for 1986 Hydrodynamics 

During the calibration process, a few model coefficients and/or boundary and initial 
conditions were adjusted to allow accurate simulation of observed data. As shown in Table 
5.2, measured data of tidal stage, current, and salinity at several locations throughout the 
Charlotte Harbor were provided by USGS (Goodwin, 1992). Figure 5.8 shows the locations 
of tidal stages and Figure 5.9 shows the locations of current velocity and salinity 
measurement. Water elevation data at USGS station 4 (Venice) were compared with those 
at USGS stations 5 (Cayo Costa) and 8 (Gulf of Mexico at Fort Myers) which are located in 
the Gulf of Mexico (Figure 1.1). The tidal elevation and phase at stations 4 and 5 are almost 
identical. Tidal range at Fort Myers beach (station 8) is larger than those at stations. The 
reason for the larger tidal range at Fort Myers is probably due to the abrupt change in the 
direction of shoreline at the southern end of Sanibel Island, rather than different offshore 
tides. Therefore, water level data at Venice station were used as tidal boundary conditions 
along all open boundaries. All specific conductance measurements were converted to salinity 
concentrations in parts per thousand using the conversion equation by Miller (Goodwin. 
1992). These data were used for comparison with model simulations of water level, velocity 
and salinity. 

Before baroclinic simulations are performed, it is necessary to generate an initial 
salinity field because, in general, salinity is much slower to adjust to initial transients than 
water level or currents. The initial salinity field for the 1986 simulation is generated by 
"spinning up" a prescribed salinity field for a sufficiently long period. The salinity values 
at the end of the simulation are then used the initial values for the 1986 simulations. This 
process is discussed below in more detail. 



95 

First, a salinity field is created by linearly interpolating measured salinity onto the 
Charlotte Harbor grid. The bottom salinity data measured by USGS during 1986 are used 
to create this salinity field. Next, a salinity assimilation term is added to the salinity transport 
equation in the model. This term forces the salinity in a selected cell to approach a specific 
value, in this case, the initial value. The six USGS measured bottom salinity values shown 
in Figure 5.6 are chosen to be assimilated into the simulation. These values were chosen 
because they were measured closest to the July 9 starting date of the 1986 simulations. The 
assimilation term , which appears on the right side of the salinity equation, takes on the 
following form (Sheng and David, 2002): 

T 

where S" iJk is the simulated salinity at the n-th time level, (S A ) iJk is the value to be 

assimilated, i.e., the initial salinity values, and T is the assimilation period, which is set to 
30 days for Charlotte Harbor simulation. It can be seen, when S" ijk is not equal to (S A ) iJk a 
force is created to drive S" iJk , toward the value of (S A ) ijk . Because of the addition of the 
salinity assimilation term, the salinity at the six USGS sites at the end of a simulation will 
be very close to their measured values at the beginning of the simulation. 

A 30-day spin-up simulation was performed from May 28 to June 27 during a dry 
season with all barotrophic forcing mechanisms (tides, river discharges, wind) to allow water 
level, velocity and salinity field to reach a dynamic steady-state throughout the computational 
domain. Using the last time step surface elevation, velocity, and salinity of the spin-up 
simulation as the initial condition, the Charlotte Harbor circulation and salinity transport 
from June 27 to July 30 was simulated with tidal forcing, wind field and river discharges as 
boundary condition. 



96 



Site 


(I ,J) Location in 


Duration 


Data Type 


Number 


Computational 
domain 






1 


(29,106) 


6/15-8/31,1986 


Tidal Stage 


2 


(40,102) 


7/15-8/31,1986 


Tidal Stage 


3 


(23, 81) 


6/15-8/31, 1986 


Tidal Stage 


4 


Outside, at Venice 


6/30-8/31,1986 


Tidal Stage 


5 


(12, 81) 


6/30-8/31, 1986 


Tidal Stage 


6 


(14, 62) 


7/10-8/31,1986 


Tidal Stage 


7 


(28, 47) 


8/14-8/31,1986 


Tidal Stage 


8 


(33, 29) 


6/26 - 8/20, 1986 


Tidal Stage 


SI-1 


(33,100) 


7/9-8/ 6, 1986 


Velocity & Salinity 


SI-2 


(30, 90) 


11 9-8/ 8, 1986 


Velocity & Salinity 


SI-3 


(35, 86) 


7/9-8/ 6, 1986 


Velocity & Salinity 


SI-4 


(11, 96) 


7/10-7/15, 1986 


Velocity & Salinity 


SI-6 


(18, 76) 


7/10 - 7/23, 1986 


Velocity & Salinity 


SI-7 


(17, 65) 


7/10 - 7/23, 1986 


Velocity & Salinity 


SI-8 


(20, 51) 


7/16 - 7/20, 1986 


Velocity & Salinity 



97 



o 
o 
o 
o I— 

o 

o 

CO 



o 

o 
o 
m 

(j> 

CM 



£ 8 

w o 

£ CM 

o 



o 
o 
o 
o 
o 
a> 

CM 



_ 




J_ 



350000 






• Tidal Measurement 
Stations (1 to 8) 

Discharge Measurement 
Stations (9 to 15) 



ite 



r 



4 



100 






Y>U r 




7 ^# 



J I . L 



J_ 



J L 



J I I L 



J__l L 



J_ 



375000 



400000 

Easting (m) 



425000 



450000 



Figure 5.8 Locations of 1986 water level and discharge measurement stations of USGS 
(Goodwin, 1992). 



98 



o 
o 
o 
o 
o 
o 

CO 



w o 

U) o 

.b en 

.C CM 

t 

o 

z 

o 
o 
o 
to 

CM 

en 

CM 



o 
o 
o 
o 
o 
en 
c\j 




J L 



± 



Velocity and Salinity 
Measurement Stations 
(SI-1 toSI-10) 



SI-3 / 







s| - 9 \ %¥%&* 



V 



SI-10 




J I l_L 



J_ 



J I L 



_L 



J I I L 



-L 



J__l I L 



350000 375000 400000 425000 450000 

Easting (m) 



Figure 5.9 Locations of 1986 velocity and salinity measurement stations of USGS 
(Goodwin, 1992). 



99 
5.3.1 Sensitivity and Calibration Simulations 

Several simulations were performed in order to test the consistency and sensitivity 
of numerical grid, bottom roughness, horizontal diffusion coefficient, and bathymetry. To 
quantify the model sensitivity to variations of these model parameters, time series statistics 
are presented in terms of the root-mean-square error (RMS) and normalize RMS error, define 
as the ratio between the RMS error and the observed range. The normalized RMS error gives 
a more meaningful indication on model's ability to reproduce the tidal signal and salinity at 
each station. The RMS error is calculated according to: 



E = 



■ , N -11/2 

,r / j \ simulated data ) 

TV i 



(5.2) 



where N is the total number of measured data, S simulated is the model result, and S data is 

measured data at each station. 

To determine which non-tidal boundary and forcing conditions are most important 

to Charlotte Harbor estuarine system modeling, a series of 1986 simulations were performed 

with a different boundary and forcing conditions removed in each run (Table 5.3). These 

simulations show that baroclinc forcing is the most important factor in simulating water 

level, flow, and salinity at overall available measured stations. 

Table 5.3 The effect of removing selected boundary conditions on the accuracy of simulated 
water level, velocity and salinity in July 1986. Values shown are average RMS differences 
vs. baseline simulation at all data stations. 



Boundary and Forcing 


Water level 


U Velocity 


V Velocity 


Salinity (ppt) 


Condition Removed 


(cm) 


(cnr/s) 


(cnr/s) 




Baroclinic Forcing 


2.35 


3.17 


2.05 


3.08 


Wind 


1.82 


2.64 


2.08 


1.03 


River Discharge 


0.57 


1.15 


0.84 


1.05 



As shown in Equation 3.28, the bottom drag coefficient is defined as a function of 



100 
the bottom roughness, z . The default value of bottom roughness, z , used for Charlotte 
Harbor simulations is chosen to be a constant 0.3 cm over the entire domain based on 
running with the several bottom roughness values which range between 0.01 and 1 (Table 
5.4). Because of the numerous different bottom types in the estuary, a constant value of 
bottom roughness may not be appropriate. To determine whether a spatially varying bottom 
roughness or a constant bottom roughness produces better simulated circulation and 
transport, a series of 1998 simulation was performed for the Indian River Lagoon (Davis and 
Sheng, 200 1 ). The results show that simulations with the spatially varying bottom roughness 
have slightly smaller errors in water level but slight larger errors in salinity and flow. Since 
the varying bottom roughness does not improve the simulated circulation and transport 
significantly, the constant bottom roughness of 0.3 cm is used for all simulations in this 

study. 

Table 5.4 The effects of varying bottom roughness, z , on the accuracy of simulated water 
level velocity, and salinity in July 1986. Values shown are average RMS errors at all data 
stations. 



Bottom roughness (cm) 


Water level 
(cm) 


Velocity 
(cm : /s) 


Salinity (ppt) 


0.1 


8.11 


7.87 


1.71 


0.3 


8.12 


7.88 


1.71 


1 


8.11 


7.87 


1.71 



As shown in Equations 3.2 and 3.3, the sub-grid scale motion is estimated with a 
horizontal diffusion coefficient, A H . A default value of 10,000 cm 2 /s was used for Charlotte 
Harbor simulations after performing several simulations using different coefficient values 
(Table 5.5). The results show little difference in water level and salinity. 

Besides adjusting boundary conditions and model coefficients, it is also important to 
use more accurate grid and bathymetry. To improve the accuracy of bathymetry in Charlotte 



101 

Harbor, bathymetric surveys were performed in Caloosahatchee River and San Carlos Bay, 

and upper Charlotte Harbor by SFWMD and SWFWMD, respectively. 

Table 5.5 The effect of varying horizontal diffusion, A H , on the accuracy of simulated water 
level, velocity and salinity in July 1986. Values shown are average RMS errors at all data 



Horizontal diffusion 
Coefficient (cm2/s) 


Water level 

(cm) 


Velocity 
(cm 2 /s) 


Salinity (ppt) 


5000 


8.12 


7.86 


1.75 


10000 


8.12 


7.88 


1.71 


20000 


8.07 


7.88 


1.64 



Overall, the model is able to simulate the surface elevation and salinity within 10% 
normalized error with maximum value. This provides validation that the hydrodynamic 
model reproduces the basic circulation of the Charlotte Harbor estuarine system. 

To supplement this short term calibration of July 1986, long term model calibration 
was conducted using one year water level and salinity data in Caloosahatchee River during 
2000. This 2000 model calibration include several test simulations to investigate effects of 
bottom roughness, salinity advection schemes, and grid resolution and bathymetry. 
5.3.2 Results of the 1986 simulation 

For the 1986 simulation, the 92x129 grid and updated bathymetry were used along 
with boundary conditions and model parameters described in Table 5.6. Water level, current 
velocity, and salinity are compared both qualitatively and quantitatively with measured data, 
where available. 

Water Level 

Calculated RMS errors between simulated and measured water level during the July 
1986 simulation are shown in Table 5.7. The normalized RMS errors are less than 10% at 
all water level stations, demonstrating the model's ability to accurately reproduce surface 



102 

elevation in the system. The highest errors at stations 1 and 2 can be attributed to the 

relatively coarse horizontal grid near the river where the gages were located. 

Table 5.6 A summary of boundary conditions and model parameters used in the 1986 

simulation. . 

Boundary Condition or Model Parameter Value 

Tidal Forcing Measured 

Wind speed and direction Measured at 2 stations 

Fresh Water Discharge Estimated (for Estero Bay) and Measured 

Bottom Roughness Constant (0.3 cm) 

Horizontal Diffusion Constant (10000 cm 2 /s) 

Horizontal grid 92 x 129 

Vertical Layers 8 



Table 5.7 Calculated RMS errors between simulated and measured water level in July 1986. 
Station Number RMS error (cm) Range (cm) % RMS error 



1 


9.59 


104.24 


9.20 


2 


9.61 


104.55 


9.19 


3 


6.61 


90.83 


7.28 


5 


9.55 


117.05 


8.16 


6 


7.56 


107.90 


7.00 


7 


5.78 


144.48 


4.00 


Average 


8.12 


111.51 


7.47 



Figure 5.10 shows the comparison between measured and simulated water level at 
six Charlotte Harbor stations. The result shows a good agreement between data and model 
results, in both amplitude and phase. From these figures, it can be seen that tidal range 
decreases about 15% from Boca Grande Pass (site 5) to Peace and Myakka Rivers (sites 1 
and 2). Also, the tidal wave takes about two to three hours to reach these rivers. 

Figure 5.11 shows the spectra of water level (data and model results) for the same 



103 
stations. The spectral curves reveal major energy bands centered around two tidal 
frequencies: the diurnal band (between 0.8 to 1.2 cycle per day) and a semi-diurnal band (1.8 
to 2.2 cycles per day). These two bands combined to represent the tides propagating from 
Gulf of Mexico into the estuarine system. The secondary energy bands, representing 
nonlinear interactions between diurnal and semi-diurnal tides with complex geometries and 
bathymetric features inside the estuary, are represented by the relatively small third and 
fourth diurnal peaks. Comparisons show that the model is able to capture both the 
amplitudes and phases well. 

Velocity 

The simulated and measured horizontal velocities at selected sites are shown in 
Figure 5.12. The angle of the vector shows the azimuth of the flow vector at a specific time, 
and the length of the vector represents the magnitude of the flow. It is important to point out 
that the observed velocity is measured at a discrete point in space and is subjected to many 
local influences that are not well represented in the model. These influences include physical 
features such as small channels, depression shoals, and mounds that are not resolvable at the 
model scale, but often influence the distribution of velocity at that location. Because of local 
influences, discrete measurement points are not consistently reliable indicators of the general 
velocity characteristics (Goodwin, 1996). The model results are spatially averaged values 
over a grid cell, hence are not expected to agree exactly with discrete data. 

Site SI-1 is located in northern Charlotte Harbor in an area of converging river flow 
and having a complex bathymetry that creates a complex lateral distribution of velocity. The 
simulated velocity at SI-1 agrees well with measured data. However, some simulated 
velocity magnitudes were 50% smaller than measured magnitudes and there was some 



104 
deviation in the azimuth of more than 30 degrees. Since the current meter was located near 
the surface at this station, this station could be influenced by freshwater inflow, wind and any 
other effects as indicated by Goodwin (1996). 

At Si-2 and Si-3, there is good agreement between simulated and measured velocities. 
In the Pine Island Sound, site-7 is located in the middle of several islands, and the observed 
data were strongly affected by local features that cannot be resolved at the scale modeled. 
However, agreement between simulated and measured data for this site is quite good. 

Calculated RMS errors between simulated and measured current velocity are shown 
in Table 5.8. The average normalized RMS error is less than 20%. The RMS error is quite 
reasonable in comparison to the range of velocity, considering the fact that many local 
influences that are not well represented in the model. 
Table 5.8 Calculated RMS errors between simulated and measured current velocity for 1986 



Station 




RMS error (cm/s) 


Range (cm/s) 


% RMS error 


Number 












U 


4.38 




14.34 


SI-1 


V 


5.97 


30.54 


19.56 




u 


11.70 




21.20 


SI-2 


V 


13.14 


55.19 


23.82 




V 


7.16 




18.72 


SI-3 


V 


6.78 


38.25 


17.72 




u 


3.06 




8.79 


SI-6 


V 


10.82 


34.79 


31.10 


Average 




7.88 




19.40 


Salinity 











Figure 5.13 shows the near-bottom simulated and measured salinity at USGS stations 



105 

SI-1,2,3,6,7, and 8 inside of Charlotte Harbor. The difference between simulated and 

measured salinity does not exceed 2 ppt except at station SI-7. 

Calculated RMS errors between simulated and measured salinity during July 1986 

simulation are shown in Table 5.9. The results show the model's ability to simulate salinity 

within 10% accuracy. The highest error is at station SI-7. The measured salinity shows very 

strong daily fluctuation which is five times greater than simulated values. This location is 

near the Captiva pass which is possibly affected by local features such as small channels, 

depressions, shoals, and mounds as mentioned before. 

Table 5.9 Calculated RMS errors between simulated and measured salinity for 1986 

simulation . 

Station Number RMS error (ppt) Maximum (ppt) % RMS error 



SI-1 


1.23 


24.01 


5.15 


SI-2 


1.47 


25.42 


5.77 


SI-3 


1.51 


27.05 


5.60 


SI-6 


0.61 


33.83 


1.81 


SI-7 


3.52 


33.82 


10.42 


SI-8 


1.90 


25.38 


7.50 


Average 


1.71 


26.55 


6.04 


Flow Patterns 









Typical averaged flow patterns during one tidal cycle, based on the water level at 
open boundary, for August, 6 on 1986 are shown in figure 5.14. Easterly flow on the western 
Florida shelf is dominant during the flood tide, whereas northerly and westerly shelf flows 
are dominant during the ebb tide. Two major passes in the Charlotte Harbor estuarine system 
are Boca Grande Pass and San Carlos Pass. The large flood and ebb flows through the Boca 
Grande Pass affects the tidal prism in upper and lower Charlotte Harbor and the flow through 



106 
San Carlos Pass satisfies the tidal prism in the Caloosahatchee River, San Carlos Bay, the 
lower part of Matacha Pass and lower extremity of Pine Island. The flow through Gasparilla, 
Captiva, and Redfish passes have effects that appear to be limited to the local area. In Pine 
Island Sound, the flow is very low during flood and ebb tide periods. However, there is some 
water transport through Pine Island Sound during high slack and low slack tide periods. 
During near high slack tide, there is northerly flow from San Carlos Bay, while there is 
southerly flow during near low slack tide. 

Residual Flow and Salinity Patterns 

Figure 5.15 show the 29-day residual circulation during July 2 to July 30, 1986. In 
the upper Charlotte Harbor, there are opposite direction surface residual flows which are the 
landward surface residual flow in relatively deep channel along the right side channel and 
seaward surface residual flow along the left side shoreline. In the bottom residual flow, the 
landward flow is dominant across the channel. 

In the Pine Island Sound and Matlacha Pass, the northward residual flow is dominant 
in the surface, while there is a week southward flow in the bottom. The surface flow come 
from upper Charlotte Harbor, Pine Island Sound, and Matlacha Pass create a strong seaward 
surface residual flow at the lower Charlotte Harbor. 

The offshore residual flow in the Gulf of Mexico is northward. The strong outflows 
from the Boca Grande Pass and San Carlos Pass create the clockwise gyres near the surface 
in the Gulf of Mexico. 

The residual salinity has a uniform vertical and horizontal distribution in the lower 
Charlotte Harbor, while the salinity has horizontal inclination across the channel and 
vertically stratification in the upper Charlotte Harbor. In upper Charlotte Harbor, the surface 



107 
salinity along the right side shoreline has greater value than that along the left side shoreline 
because of the residual surface flow pattern. The residual salinity field shows a vertical 
stratification of 2 ppt at the Peace River mouth since the strong river discharge and mixed 
tide from Gulf of Mexico create two layered flow (seaward flow at surface and landward 
flow at bottom) in this area. A two layered flow and salinity stratification is often presented 
at the Charlotte Harbor estuarine system. The significant vertical stratification can suppress 
vertical mixing and lead to the formation of pycnocline and hypoxia in the upper Charlotte 
Harbor (Stoker, 1992; CDM, 1998). Therefore, understanding of the density-driven 
circulation is critical to the management of Charlotte Harbor estuarine system. 



108 



100 



100 r 



■100 



190 



Myakka River at El Jobean (site 1) July, 1986 




Julian Day 
Peace River at Punta Gorda (site 2) July, 1 986 




200 



210 



Julian Day 
Charlotte Harbor at Bokeelia (site 3) July, 1 986 





IUU 




? 

o 


50 




i 

<D 
| 







i- 
CO 


-50 


'. 



210 



Julian Day 

Figure 5.10 Comparison between simulated and measured water level in July 1986. 



109 



Gulf of Mexicoat Cayo Costa near Boca Grande (site 5) July, 1 986 




100 r 



-100 



100 r 



Julian Day 
Pine Island Sound near Captiva (site 6) July, 1 986 




200. 



Julian Day 
B. Fort Myers Beach (site 8) July, 1 986 



210 




190 

Figure 5.10 continued 



ulian Day 



110 



Site 1 




Site 2 



>- 
a , 

T10 5 

to' 

■D 

c 

fto 1 

o 

a 



12 3 4 

Frequency (cycle/day) 



Site 6 




Simulated 
Measured 



Frequency (cycle/day) 




||Q 
o 

Q. 



710- 



5iry 
f 

C10 3 

0) 

TJ 

|io 2 

w 

|10° 

o 

CL 



2 3 4 

Frequency (cycle/day) 



Site 5 




Frequency (cycle/day) 



Site 8 




Simulated 
Measured 



12 3 4 

Frequency (cycle/day) 



Figure 5.1 1 Comparison between simulated and measure spectra of water level in July 
1986. 



Simulated Velocity 



111 



Measured Velocity 



m161910(SI-1), July 1986 



o 

<v 



o 
o 




I I I I I I 



J l_ 



10 11 12 13 



14 



15 16 



o 



o 

o 



W^w 



y/Mi 



WYV 




LP 



-i I I L. 



9 10 11 12 13 14 15 16 

Figure 5.12 Comparison between simulated and measured velocity data in July 1986. 



112 



Measured Velocity 



m809110(SI-2), JULY 1986 



o 

I 

o 

o 




Simulated Velocity 



o 

o 




Figure 5.12 continued. 



Measured Velocity 



m810110(SI-3), July 1986 



o 

111 

i 

o 




I l_J I ^_l l_-l 1 L_J- 



J 



10 



11 



12 



13 



14 



15 



16 



Simulated Velocity 



o 

0) 

I 

o 

o 




_J I I I I 1 1 L- 

10 11 



12 



13 



14 



15 



16 



Figure 5.12 continued. 



113 



114 



Measured Velocity 



S 

E 
o 

o 



m149614(SI-6), July 1986 




i ' i i i i i i i i i i — i — i — i — i — i — i — i — i — i — i 

9 10 11 12 13 14 15 16 



Simulated Velocity 



1 



\V\ 



- 



i« M Ill 



ftfMi 



i i i i i i 



■ .... i . . i i i i i i i i 



9 10 11 12 13 14 15 16 



Figure 5.12 continued. 



115 



SI-1 station, July 1986 



SI-2 station, July 1986 



35 
30 
25 

P 

L 

L 

£ 

= 15 
s 

10 

5 



35 
30 

25 

*■» 
D. 
320 

= 15 

10 

5 

9 



simulated salinity 
measured salinity 



f90 195 . .. 200 

Julian Day 



205 



SI-3 station, July 1986 



; 






H^^ r .^y^r^^X'yW^''%- A 


- 




- 


simulated salinity 

measured salinity 

. i 



90 195 . ,. 200 

Julian Day 



205 



35 




35 
30 
25 h 



Q. 
S20 

£15 

s 

10 



5 - 
9 



90 



simulated salinity 
measured salinity 



1 95 200 205 

Julian Day 
S 1-6 station, July 1986 



simulated salinity 
measured salinity 



195 , .. 200 

Julian Day 



205 



35 
30 

25 

Q. 

S20 

£ 

£15 
10 

5 

9 



SI-7 station, July 1986 



1 i ill ;i."iiV : c 



i nihil! 



\m\i 



>j ys.-^ 



90 



simulated salinity 
measured salinity 



195 . . . 200 

Julain Day 



205 



35 



SI-8 station, July 1986 



30 - 

25 
C 
a. 

81 

10 






simulated salinity 
measured salinity 



90 



195 , .. 200 

Julian Day 



Figure 5.13 Comparison between simulated and measured salinity in July 1986. 



205 



116 



l Vw 




Low Slack Tide 
8/06/86 10:00 

1 m/sec 











Flood Tide 
8/06/86 13:00 









!»>»-•"., 



/WW-'.'lf"':-* 




j^.^ 






High Slack Tide 
8/06/86 16:00 




••• '-••'•-•-'■■••'••••••• '•■y/4 / >y^y^k 

••.••.•'.'•.'/.v:.;;^ : $^ 




Ebb Tide 
8/06/86 19:00 



I ^r < 



'•.\\VWSavV 







Figure 5.14 Typical flow pattern of Charlotte Harbor estuarine system during one tidal 
cycle for August 6, 1986. 



117 



I 










& 



X. 



V; v 



<A fe*d 



Residual Flow 
at Surface 

1 cm/sec 




Residual Flow 
at bottom 

1 cm/sec 



\x, \t h 




Residual Salinity 
at Surface 







rca* 



Residual Salinity 
at bottom 



,4r* 




Figure 5.15 The 29-day residual flow and salinity for Charlotte Harbor estuarine system 
during July 2 to July 30, 1986 



118 
5.4 Simulations for 2000 Hydrodynamics 

1986 hydrodynamic simulation of Charlotte Harbor estuarine system was conducted 
for 30 days to calibrate the major model coefficients and inputs. The field data did not cover 
the Caloosahatchee River which is a very important segment of the study area. To 
supplement 1986 calibration simulation, long-term model calibration was conducted using 
one year water level and salinity data in the Caloosahatchee River during 2000. Figure 5.16 
shows the salinity measurement stations in Caloosahatchee River. Water levels were 
measured at Fort Myers and Shell Point. The validated CH3D model was then used to assess 
the effects of Sanibel Causeway and the navigation channel in San Carlos Bay, and to 
quantify the relationship between freshwater inflow and spatial and temporal salinity 
distribution in the Caloosahatchee River. 
5.4.1 Sensitivity and Calibration Simulations 

Several simulations were performed in order to test the effects of grid configuration, 
bottom roughness, bathymetry, and salinity advection scheme on salinity distribution in the 
Caloosahatchee River. To study how the horizontal grid resolution affects simulated 
circulation and transport within the estuarine system, two 2000 simulations are performed 
using two different grids. Figure 5.17 shows the comparison of the coarse grid (71x92) and 
the fine grid (92x 1 29). The fine grid has about two times finer resolution than the coarse grid 
in San Carlos Bay and near the Caloosahatchee River mouth. Table 5.10 shows that the two 
different horizontal grids affected the salinity less significantly than the water level. The 
geometry and bathymetry at the river mouth play a very important role for water circulation 
inside the river. The coarse grid tends to produce more error in water level simulation than 
fine grid because the coarse grid does not accurately to represent the islands and navigation 



119 
channels in the San Carlos Bay and Caloosahatchee River mouth. The fine grid (92x129) 
is used for all simulations for Charlotte Harbor estuarine system. 



o 
o 
o 
o 

CD 
CD 

CM 



T 



BR31r ^#^ 






§ityf Fort Myers 







JKTSF*** 






1 



J L 






h.-B-K^ I I I 



400000 



410000 420000 

Easting (m) 



S79 



-L 



430000 



Figure 5.16 Locations of the available 2000 water level and salinity measured stations at 
Caloosahatchee River operated by SFWMD. 



120 







depth 
(cm) 

■i 1000 



900 
800 
700 
600 
500 
400 
300 
200 
100 




t 



depth 
(cm) 

1000 



900 
800 
700 
600 
500 
400 
300 
200 
100 




•A> 



Figure 5.17 The comparison of the coarse grid (71x92) and the fine grid (92x129) for 
Charlotte Harbor estuarine system. 



121 

Table 5.10 The effect of horizontal grid resolution, on the accuracy of simulated water level 
and salinity. Values shown are average RMS errors for 2000 calibration at all available 
s tations. Values shown in parenthesis are % RMS error normalized by maxim um values. 

Variables Station 71x92 92 x 129 





S79 


S 
B 


1.69(4.62) 
2.54(6.93) 


1.46(3.98) 
1.84(5.03) 


Salinity 
(Ppt) 


BR31 


S 
B 


1.17(3.20) 
1.89(5.16) 


1.55(4.22) 
2.06(5.61) 




Fort Myers 


S 
B 


1.55(4.23) 
2.07(5.63) 


1.85(5.03) 
2.63(7.18) 




Shell Point 


S 
B 


3.78(10.29) 
3.37(9.17) 


3.76(10.25) 
3.53(9.64) 




Sanibel 


S 
B 


1.85(5.05) 
2.05(5.59) 


1.94(5.28) 
2.00(5.45) 




average 




2.20(5.99) 


2.26(6.17) 


Water level 


Fort Myers 




9.08(5.44) 


5.51(3.30) 


(cm) 


Shell Point 




6.92(4.83) 


4.25(2.96) 




average 




8.00(5.14) 


4.58(3.13) 



Vertical salinity stratification in the Upper Charlotte Harbor and Caloosahatchee 
River is a common seasonal occurrence (Environmental Quality Laboratory, Inc., 1979). In 
high river inflow events, a stable vertical salinity gradient is created which suppresses 
vertical mixing unless there are sufficient mixing by wind or tide. Therefore, the vertical 
grid resolution is an important factor to reproduce vertical salinity distribution which is a one 
of major cause effects of hypoxia in upper Charlotte Harbor. To study the effect of varying 
the number of vertical layers used by CH3D model, 2000 simulations were performed with 
4 and 8 vertical layers (Table 5.11). Since a one year simulation of hydrodynamics, sediment 
transport, and water quality took several days, the simulation with more than 8 vertical layers 
which required the additional computational time, is not feasible for this study. While the 
simulated water level RMS error changes little, the averaged simulated salinity RMS error 



122 



improves about 1 ppt with 8 layer simulation. The largest improvement in simulated salinity 

was achieved at Shell Point station with 3 ppt using 8 vertical layers. To reproduce vertical 

stratification for salinity and water quality simulation, eight layer simulations are deemed 

appropriate for the Charlotte Harbor estuarine system. 

Table 5.1 1 The effect of vertical grid resolution, on the accuracy of simulated water level 
and salinity. Values shown are average RMS errors for 2000 calibration at all available 
s tations. Values shown in parenthesis are % RMS error normalized by maximum values. 

8 layers 



Variables 



Station 



4 layers 





S79 


S 
B 


1.85(5.08) 
2.09(5.73) 


1.46(3.98) 
1.84(5.03) 


Salinity 
(Ppt) 


BR31 


S 
B 


1.86(5.11) 
2.37(6.51) 


1.55(4.22) 
2.06(5.61) 




Fort Myers 


S 
B 


2.34(6.43) 
2.30(6.31) 


1.85(5.03) 
2.63(7.18) 




Shell Point 


S 
B 


6.82(18.70) 
5.31(14.55) 


3.76(10.25) 
3.53(9.64) 




Sanibel 


S 
B 


2.48(6.80) 
4.02(11.02) 


1.94(5.28) 
2.00(5.45) 




average 




3.14(8.63) 


2.26(6.17) 


Water level 


Fort Myers 




4.50(2.70) 


5.51(3.30) 


(cm) 


Shell Point 




4.42(3.08) 


4.25(2.96) 




average 




4.46(2.89) 


4.58(3.13) 



In Caloosahatchee River, water level and salinity are very sensitive with the varying 
bottom roughness while the varying bottom roughness does not affect the simulated 
circulation and transport significantly in the estuary in July 1986. However, several 2000 
simulations were performed using different coefficient value to determined the effect of 
varying the bottom roughness on the accuracy of the simulation circulation and transport. 
(Table 5.12). With increasing bottom roughness, accuracy of water level is improved while 
that for salinity is worse. A constant bottom roughness of 0.3 cm is used for all simulations. 



123 

Table 5.12 The effect of varying bottom roughness, z , on the accuracy of simulated water 
l evel and salinity in 2000. Values shown are average RMS errors at all data stations. 

Bottom roughness (cm) Water level (cm) Salinity (ppt) 

0.2 4.94 2.28 
0.3 4.87 2.29 
04 4J38 231 

Model sensitivity study was conducted to investigate the relative accuracy of the 
salinity advection schemes including upwind, QUICKEST, and Ultimate QUICKEST 
methods. Time series comparisons of measured and simulated surface salinity using three 
advection schemes at three stations, Shell Point, Fort Myers, andBR31, are shown in Figures 
5.18 to 5.20. The salinity comparison at Shell point does not show much difference for all 
three schemes, while at Fort Myers and BR31, the simulated salinity with upwind scheme 
is much higher than those for the other advection schemes and measured salinity. The 
advection scheme plays a very important role for salinity simulation at cells inside river 
where there is usually very strong salinity gradient from the river mouth at Shell Point to the 
upstream. Lower order advection scheme such as upwind scheme tends to produce more 
error in salinity simulation because of its inherently high numerical diffusion. A series of 
vertical-longitudinal salinity profiles along the axis of the river during slack water is shown 
in Figure 5.21 to compare salinity distribution along the river simulated by various advection 
schemes. The salinity distribution obtained with the QUICKEST method shows similar 
pattern with that for Ultimate QUICKEST method. The upstream salinity near BR3 1 shows 
much higher value with upwind scheme because salinity is quickly diffused from river mouth 
to the upstream. With the upwind scheme, it is difficult to reproduce the vertical salinity 
stratification which often occurs during high river flow period, while the other advection 
schemes produced the stratification. 



124 



Ultimate QUICKEST Scheme 




200 250 

QUICKEST Scheme 



D. 

a. 

>. 

4-1 

c 



200 
Upwind Scheme 

35 r 



250 



simulated salinity 
measured salinity 



300 



300 



200 



250 



300 



350 




350 




350 



Julian Day 

Figure 5.18 A comparison between simulated and measured salinity at Shell Point using 
Ultimate QUICKEST, QUICKEST, and upwind advection schemes. 



125 



Ultimate QUICKEST Scheme 



35 

30 

a- 25 

Q. 

S 20 

i 15 
5 

w 10 

5 




200 250 

QUICKEST Scheme 



35 

30 

S? 25 

Q. 

3 20 

I 15 

CO 

w 10 

5 




200 

Upwind Scheme 



250 



^A^^ 






- 




simulated salinity 

measured salinity 









300 



300 



350 




350 







200 



250 



300 



350 



Julian Day 

Figure 5.19 A comparison between simulated and measured salinity at Fort Myers using 
Ultimate QUICKEST, QUICKEST, and upwind advection schemes. 



126 



Ultimate QUICKEST Scheme 



20 



-, 15 

a 

S 

.IT 10 

c 

TO 



200 250 

QUICKEST Scheme 



20 



^ 15 

a 
a. 

£• 10 

c 

nj 







200 

Upwind Scheme 



250 



~ 15 - 



Q. 

a 



£* 10 






200 



250 



simulated salinity 
measured salinity 




300 



300 



350 




350 




300 



350 



Julian Day 

Figure 5.20 A comparison between simulated and measured salinity at BR31 using 
Ultimate QUICKEST, QUICKEST, and upwind advection schemes. 



127 



Current Time : 9/ 7/2000 1 9:00 

100 Shell Fort 

Point Myers 



-40000 



BR31 



-30000 -20000 -10000 

Distance froma S79 (cm) 



Salinity 
S79 <PP*> 




100 Shell 
Point 



Fort 
Myers 



BR31 



-40000 



-30000 -20000 -10000 

Distance from S79 (cm) 



Salinity 
S79 (ppt) 




BR31 



-40000 



10000 



Salinity 
S79 (P pt ) 




-30000 -20000 

Distance from S79 (cm) 

Figure 5.21 Simulated longitudinal-vertical salinity along the Caloosahatchee River at 
slack water before flood on September 7, 2000. 



128 

Table 5.13 shows the effect of salinity advection scheme on the accuracy of simulated 

water level and salinity. The results show slight difference between water level and salinity 

simulated with the Ultimate QUICKEST scheme and the QUICKEST scheme, while the 

simulated salinity RMS errors with upwind scheme is much greater. Maximum 

improvement in simulated salinity is achieved when the QUICKEST scheme is used because 

most of the model parameters and boundary conditions are calibrated using the QUICKEST 

scheme. However, because of the higher cost of model simulations using the Ultimate 

QUICKEST scheme, coupled with only a marginal improvement in the simulated results, the 

QUICKEST scheme is deemed appropriate for this study. 

Table 5.13 The effect of varying salinity advection scheme on the accuracy of simulated 
water level and salinity in 2000. Values shown are average RMS errors at all data stations. 

Advection scheme Water level (cm) Salinity (ppt) 

2.35 
2.34 
3/73 

Beside adjusting boundary condition and model coefficients, it is also interesting to 
see how varying the model grid and bathymetry affects simulated circulation and transport 
within the estuary. Because flow through the Caloosahatchee River mouth is crucial to the 
water level and salinity inside the river basin, it is useful to perform a numerical simulation 
to study how the flow through the river mouth is affected by its cross-sectional area. As was 
discussed in the previous section, the Charlotte Harbor grid bathymetry was developed by 
interpolating measured bathymetry onto the entire grid followed by a simple smoothing 
scheme. Although the bathymetry in upper Charlotte Harbor and inside Caloosahatchee 
River were updated by measured data by SWFWMD and SFWMD, the resolution of 
bathymetry in San Carlos Bay area, including the Caloosahatchee River mouth, is very coarse 



Ultimate QUICKEST 


5.56 


QUICKEST 


5.63 


Upwind 


5.36 



129 
(about 300 m resolution). To study how the accuracy of grid bathymetry affects simulated 
the circulation and transport, several 2000 simulations were performed using grid systems 
with modified bathymetry data. The first simulation used a modified grid bathymetry which 
had a minimum depth of 2.5 m (1.5 m for baseline simulation). This simulation was 
performed to determine how important the "shallowness" of the river is to the circulation and 
transport. The second simulation used a bathymetry that included an artificially dredged 
channel through out the entire river. The navigation channel included in the grid bathymetry 
at a depth of 3.5 m (NAVD88). Since the grid system is too coarse to adequately resolve the 
navigation channel, forcing a channel into the grid system resulted in over estimation of the 
cross-sectional areas within the river. Table 5.14 compares the results of the year-long 
simulations performed using the original and two modified grid bathymetries. While forcing 
an deeper navigation channel into the grid system had little effect, imposing a 2.5 m 
minimum cell depth slightly improved simulated salinity and greatly worsened water level 
through the Caloosahatchee River. Since neither modified grid system improved the 2000 
simulation, the standard bathymetry is used for all subsequent simulations. 

Table 5. 14 The effect of modifying bathymetry on the accuracy of simulated water level and 
salinity in 2000. Values shown are average RMS errors at all data stations. 



Simulation description 


Water level (cm) 


Salinity (ppt) 


Standard bathymetry 


4.88 


2.29 


Minimum depth of 2.5 m 


6.14 


1.90 


Deeper navigation 
channel 


4.90 


2.32 



Overall, the model is able to simulate water level and salinity within 10 % normalized 
RMS error with the maximum measured values. This indicates that the hydrodynamic model 
reproduces the basic circulation and salinity transport of the Caloosahatchee River so that 



130 
this model can be used to simulate water quality processes. 
5.4.1 Results of the 2000 Simulation 

This section summarizes the calibration of the long term Charlotte Harbor circulation 

and transport model using the field data collected in January to December 2000. In addition, 

the mathematical model of the temperature transport with heat flux model as surface 

boundary condition was applied to improve the simulation of circulation in the study area. 

Based on the results, it is apparent that the model accurately simulated the observed water 

level, salinity, and temperature distribution in the Caloosahatchee River. Table 5.15 

summarizes the boundary conditions and model parameters for 2000 simulation. 

Table 5.15 A summary of boundary conditions and model parameters used in 2000 
simulation. 

Boundary Condition or Model Parameter Value 

Tidal Forcing Measured at Naples 

Wind speed and direction Measured at 3 stations 

Fresh Water Discharge Estimated for Estero Bay and Measured at 

Peace, Myakka, and Caloosahatchee Rivers 

Bottom Roughness Constant (0.3 cm) 

Horizontal Diffusion Constant (10000 cm 2 /s) 

Horizontal grid 92 x 129 

Vertical Layers 8 

Water Level 

When measured water level at an open sea is used as a tidal boundary condition, it 
is necessary to unify datum level with bathymetry and water level measured in the estuary. 
For the 2000 simulation of Charlotte Harbor, the datum level for all measured water levels 
and the bathymetry were converted to NAVD88. Water level at the Naples station, which 
is the tidal boundary condition, was leveled to MLW by NOAA CO-OPS and converted to 



131 
NAVD88 according to tidal bench mark at Naples (26° 7.8'N, 81° 48.4W). 

Calculated RMS errors between simulated and measured water level for 2000 

simulations are shown in Table 5.16. The normalized RMS errors are less than 3% at all 

available water level stations, demonstrating the model's ability to accurately reproduce 

surface elevation in the system. 

Table 5.16 Calculated RMS errors between simulated and measured water level for 2000 
simulation 



Station Name 


RMS 


! error 


(cm) 


Max 


imum Range 
(cm) 


% RMS error 


Fort Myers 




5.51 






166.80 


3.30 


Shell Point 




4.25 






143.26 


2.96 


Average 




4.88 






155.03 


3.13 



Figure 5.22 shows a year long and 20 days' comparison between simulated and 
measured water level at Shell point and Fort Myers. Although the normalized RMS errors 
are very small, the model constantly overestimated at Fort Myers and underestimated at Shell 
Point station. A probable sources of these differences could be the bathymetry, grid 
resolution, bottom roughness, and open boundary condition. In the Caloosahatchee River, 
there is a very narrow navigation channel which the current grid system could not resolved. 
The bottom roughness is a very sensitive parameter, specially in the river environment, as 
shown in previous calibration. This study uses a constant bottom roughness of 0.3 value, but 
there is no direct measurement of the bottom roughness for this estuarine system. This might 
cause some error in model simulation. In this study, CH3D does not include calculation of 
flooding/drying cell. Therefore, this study uses a minimum depth of 1.5 m which could not 
resolve any shallow water below this depth. Once again, it has been demonstrated that model 
is sensitive to the bathymetry and grid resolution, and the model accuracy will improve if 



132 
more accurate bathymetry and grid are used. 

Salinity and Temperature 

The circulation in Charlotte Harbor estuarine system is driven primarily by the mixed 
(diurnal and semi-diurnal) tides from the Gulf of Mexico, as well as by wind and density 
gradient. During periods of high freshwater inflow from the rivers, significant vertical 
salinity stratification can be found in the Upper Charlotte Harbor and Caloosahatchee River. 
A two layered flow and salinity structure characteristic of the "classic estuarine circulation" 
is often present (Sheng, 1998). Therefore, understanding of vertical salinity distribution is 
a very important part of hydrodynamic and transport simulations. The salinity is measured 
at two locations in the water column referred to as "upper" and "lower". The vertical 
positions of the salinity measurements taken by SFWMD are given in Table 5.17. For this 
study, eight vertical layers are used by the CH3D model. The simulated salinity values at 
eight vertical layers are interpolated vertically to allow comparison of salinity at the exact 
location of the measured stations. 

Table 5.17 presents the maximum and minimum measured salinity, the RMS error, 
and the normalized RMS error (with respect to maximum salinity) for all stations during 
2000. The results show the model's ability to simulate salinity is within 7% error except at 
the Shell Point station. The Shell Point station is located at the Caloosahatchee River mouth, 
which is very sensitive to bathymetry and grid resolution. As mentioned before, to improve 
salinity at this point, finer grid resolution and bathymetry data will be needed. It should be 
pointed out that % RMS error is not a good measure of model accuracy. At S79 and BR3 1, 
although the % RMS error is low, the actual error is quite significant. 

Figure 5.23 to 5.27 show the near-bottom and near-surface simulated and measured 



133 
salinity at SFWMD stations in Caloosahatchee River. The results show reasonable 

agreement with measured data for both wet season and dry season except at S-79 and BR3 1 . 

S-79 and BR31 stations are directly affected by the river boundary condition because these 

two stations are located within a few grid cells from S-79. The water level and discharge at 

the Lock and S-79 is controlled by 8 tainter gates and 2 sector gates whereas the river 

boundary condition in CH3D is specified by the averaged flow rate. At these two stations, 

there is only one grid cell across the relatively narrow width. Therefore, the current boundary 

condition and grid resolution at S-79 and BR31 stations could not resolve the detail 

characteristics of circulation and salinity transport. Furthermore, the bathymetry in this area 

was produced from only a few cross sectional bathymetry data provided by SFWMD. 

Table 5.17 Calculated RMS errors between simulated and measured salinity for 2000 
simulation 



Station 
Name 


Mean 
depth 
(cm) 


Layers 


Location 
from 

bottom 
(cm) 


RMS 
error 
(PPt) 


Minimum 

Salinity 

(PPt) 


Maximum 

Salinity 

(PPt) 


% 
RMS 
error 


S-79 


591 


upper 


298 


1.49 


0.3 


13.85 


4.07 






lower 


103 


1.88 


0.3 


14.55 


5.11 


BR31 


678 


upper 


386 


1.60 


0.3 


13.96 


4.36 






lower 


251 


2.11 


0.3 


15.51 


5.76 


Fort Myers 


261 


upper 


227 


1.84 


0.3 


18.66 


5.01 






lower 


101 


2.56 


0.3 


20.83 


6.98 


Shell Point 


304 


upper 


209 


3.82 


2.4 


34.50 


10.41 






lower 


93 


3.57 


3.1 


36.13 


9.72 


Sanibel 


254 


upper 


135 


1.96 


11.2 


36.69 


5.34 






lower 


19 


2.05 


14.3 


36.45 


5.59 


Average 








2.29 






6.23 



Seasonal salinity patterns occurred in response to the variation in volume of 



134 
freshwater inflow. The highest measured salinity occurred during an extended period of low 
flow in December 2000. During the wet season, the fresh water from S-79 reached into San 
Carlos Bay. The salinity in San Carlos Bay was directly affected by the variation of river 
discharge from S-79 station. 

Figure 5.28 shows the simulated and measured temperature near the surface at the 
Fort Myers station. The measured and simulated temperature show significant annual 
variations in the Charlotte Harbor estuarine system. Water temperature ranges from an 
average of about 30 °C during the summer to about 15 °C in December and January. The 
daily fluctuation of water temperature is about 1 to 3 °C. The simulation result shows very 
good agreement with measured data over seasonal and daily scales. The RMS error is less 
than 2.5 °C at the Fort Myers station which is about 7% RMS error of the maximum 
temperature of 32.20 °C. 

Flow Patterns 

Typical flow pattern (vertically averaged flow) in San Carlos Bay during flood tide 
and ebb tide on August, 7 2000 are shown in figure 5.29. The large flood and ebb flows 
through San Carlos Pass satisfy the tidal prism in the Caloosahatchee River, San Carlos Bay, 
the lower part of Matacha Pass and lower extremity of Pine Island Sound. Due to Sanibel 
Causeway and three navigation channels, the flood tide is separated into three major flows 
along the navigation channel. The flow through the right side of causeway moves toward the 
Caloosahatchee River , the flow through left side of causeway moves to Pine Island Sound, 
and flood tide across the center channel of causeway follows navigation channel toward the 
Matlacha Pass. During ebb tide, the flows are relatively weak and follow the opposite 
directions. Along the East-West direction navigation channel, there are two direction flows: 



135 
the eastward flow during flood tide and the westward flow during ebb tide. 
Residual Flow and Salinity Patterns 

Figures 5.30 and 5.3 1 show the one year residual flow and salinity in 2000. In the San 
Carlos Bay, the westward residual flow is dominant in the surface and bottom, while there 
is a relatively week northward flow along the left and right side of shoreline. The strong 
outflows from the left side of San Carlos Pass create the anti-clockwise gyres near the surface 
at the outside of Sanibel Causeway. This anti-clockwise gyre is associated with the 
clockwise gyre in the Gulf of Mexico which is shown in Figure 5.15 for the 1986 simulation. 

The residual salinity has a relatively uniform vertical and horizontal distribution in 
the San Carlos Bay, while the salinity has a strong vertical stratification at the 
Caloosahatchee River mouth due to the strong river discharge and mixed tide from Gulf of 
Mexico. Therefore, the density-driven circulation is a very important factor for 
understanding the circulation near such large rivers as Peace and Caloosahatchee. 



136 



at Fort Myers 




100 



80 85 

at Shell Point 



3 



85 



200 



300 

simulated 
measured 




95 



100 





> 


• 




1 , 'S^fi»*^ $2 




t r ff 

t 






k 


30 




200 


300 



simulated 
measured 




95 



JuliarrDay 

Figure 5.22 Comparison between simulated and measured water level for 2000 
simulation. 



100 



137 



atS79 

30 r 



25 



S 20 

E 

ra 10 
CO 



u i00 



30 
25 

1 2 ° 

5 15 

c 

ro 10 
CO 

5 



loo 



S79 at Caloosahatchee River 




20 ?Julian Day 



Measured at near Surface 
Simulated at near Surface 



£±~ 



-*/v^ ■*■ 



w 



200, ,. „ 

Julian Day 



300 

Measured at near Bottom 
Simulated at near Bottom 



n> 



a - 



200, ,. 
Julian Day 



w~ 



w 



300 



Figure 5.23 Comparison between simulated and measured salinity at S-79 in 2000. 



138 




S79 at Caloosahatchee River 



a 20 

c 

■ 10 
CO 



^^\_ 



100 



30 
25 



§■ 20 



C 
CO 



15 



10 




100 




20( 3ulian Day 



300 

Measured at near Surface 
Simulated at near Surface 



ZL 



/T* 



200, „ _ 

Julian Day 



300 

Measured at near Bottom 
Simulated at near Bottom 



-&. 



,- -^T^i ta 



200, ,. „ 

Julian Day 



4* 

300 



Figure 5.24 Comparison between simulated and measured salinity at BR31 in 2000. 



139 



S79 at Caloosahatchee River 




"100 



20l 2)ulian Day 



at Fort Myers 



300 

Measured at near Surface 
Simulated at near Surface 




200 i . 
Julian Day 



300 

Measured at near Bottom 
Simulated at near Bottom 




100 



200 

Julian Day 



300 



Figure 5.25 Comparison between simulated and measured salinity at Fort Myers in 2000. 



140 




S79 at Caloosahatchee River 




Julian Day 



at Shell Point 

40 r 



300 

Measured at near Surface 
Simulated at near Surface 




TOO 



20 °, .. « 

Julian Day 



300 

Measured at near Bottom 
Simulated at near Bottom 




Q. 

a. 

i 

_£ 



100 200, ,. _ 300 

Julian Day 

Figure 5.26 Comparison between simulated and measured salinity at Shell Point in 2000. 



J41 




S79 at Caloosahatchee River 



a. 
a, 

5 

c 
ro 




20 julian Day 



near Sanibel Causeway 



300 

Measured at near Surface 
Simulated at near Surface 



Q. 
Q. 

5 

15 
(J) 




200. , 

Julian Day 



300 

Measured at near Bottom 
Simulated at near Bottom 




TOO 



Julian Day 



300 



Figure 5.27 Comparison between simulated and measured salinity near Sanibel 
Causeway in 2000. 



142 



at Fort Myers 




100 



200 

Julian Day 



300 



35 



at Fort Myers 



o 

o 



3 30 

+-• 
w 

I— 
Q) 
Q. 

£ 

0) 



25 - 



o 

TO 



? 9 




Simulated 
Measured 



40 



145 150 

Julian Day 



155 



160 



Figure 5.28 Comparison between simulated and measured temperature at Fort Myers in 
2000. 



143 



/ 


■ «c» 




q V* 




he I 




" *& 




& 




7 o>St[ 




n 'A 




1 \VW 




" 




i .^\ 




» • > 




°!?> . ' - 



" : IIIIU 1 '' 



mitt 






bo 









fe7 ,>' ;V .*• £ 

I: 7 



n\, 



ho,*** 

! > ' O^" 



jus p ,.r . - './ 



«■■■ / .'■ >' / '' /y,^^p j)t)£ - 



rf ^f" 



Ebb Tide 
8/7/2000 0:0 







Figure 5.29 Typical flow pattern of San Carlos Bay during ebb and flood tide for August, 
7 on 2000. 



144 



/ '• K 



w 



• "it 



I""' 

s V * * 

f. 



us / 



, ■ y-ugf .... \^' 




"^s 



1' y'y/W4mf§^ * r^ 















& 






... 



s 





Residual surface flow 



> 



>^d 



H 






iw 



J J) 



Jliijf 












A 



->• •' ■■ -J3£u ^ >•<*/ -" •' -' / 






SB^-fcN' 



Residual bottom flow 

— > 
1 cm/s fen 



n 












■* <L : M>& 



\v u 



1 



^ := s 



^V-,7 W 



Figure 5.30 One-year residual flow in San Carlos Bay in 2000. 



145 







Figure 5.31 One year residual salinity distribution in San Carlos Bay in 2000. 



146 
5.4.2 Applications of 2000 Hydrodynamic Simulations 
Hydrologic Alterations 

Hydrologic alterations in the San Carlos Bay and Caloosahatchee River have taken 
many forms such as navigation channel and causeway. It has been suggested such large 
transportation projects as the dredging of the IntraCoastal Waterway (ICW) and the 
construction of Sanibel Causeway (SC) are linked to the decline of scallop populations in 
Pine Island Sound (Estevez, 1998). A quantitative assessment of this suggestion is long 
overdue. Moreover, to improve the natural environment of this area, it is necessary to 
quantify the effects of these artificial hydrologic alterations. In this study, the calibrated 
Charlotte Harbor model is used to evaluate the effects of these hydrologic alterations. 

Figure 5.32 shows the location of eleven stations which are selected to quantify the 
hydrologic alteration, and the locations of Sanibel Causeway (A) and IntraCoastal Waterway 
(B). Three month simulations from April 9 to July 8, 2000 were conducted with two 
hydrologic alterations in the Charlotte Harbor estuarine system. To test these effects, three 
cases were considered: (1) present condition with Sanibel Causeway and existing bathymetry 
(BASELINE), (2) hypothetical condition with Sanibel Causeway removed (NSC), (3) 
hypothetical condition with ICW removed (NICW). Figure 5.33 shows the bathymetry and 
shoreline for each case. 

The simulated Charlotte Harbor circulation in the presence of and in the absence of 
the Sanibel Causeway and IntraCoastal Waterway are compared in terms of the instantaneous 
flow from July 3 to July 8, 2000. Figures 5.34 and 5.35 show the comparisons of water level 
and salinity for each case at three selected stations: ST05 (Pine Island Sound), ST08 (San 
Carlos Bay), and ST10 (Caloosahatchee River mouth). 



147 
The water level at all three stations show little difference among the three cases. The 
salinity at Pine Island Sound (ST05) show little difference among the three cases, because 
the impacts of these hydrologic alterations are very small and local. The results show that 
water level and salinity transport are more affected by the absence of Intracoastal Waterway 
(NICW) than by the absence of the Sanibel Causeway (NSC) except salinity at ST08 which 
is located near the Sanibel Causeway. The freshwater from Caloosahatchee River and 
saltwater from Gulf of Mexico are exchanged through the ICW. In the absence of the ICW, 
salinity at San Carlos Bay (ST08) is increased, while salinity at Caloosahatchee River Mouth 
(ST 10) is decreased at both surface and bottom layers because of the reduced flow and 
salinity transport between the Caloosahatchee River and San Carlos Bay. 

Calculated water level RMS differences between the baseline simulation and two 
alteration cases from April to July 2000 are shown in Table 5.18 for 11 selected stations. 
The RMS differences are less than 2 cm at all selected stations. The highest difference is 
found at station 10 which is located at the Caloosahatchee River mouth. Calculated salinity 
RMS differences between baseline simulation and the two cases during this period are shown 
in Table 5.19. Salinity for the two alteration runs did not show much difference with the 
baseline simulation except at station 10. 

The 29-day residual flow and salinity patterns in San Carlos Bay are shown in Figure 
5.36 (at a surface) and Figure 5.37 (at a bottom) during June 9 to July 8, 2000. Although the 
results show some slight impact of the causeway on the residual flow as manifested by the 
circulation gyres in the immediate vicinity of the causeway, there is no noticeable impact on 
the salinity distribution in the San Carlos Bay area. The causeway islands did not block the 
flow of saline ocean water from entering into the San Carlos Bay and reaching Pine Island 



148 
Sound because the causeway islands are already located in a very shallow region. While 
there is strong residual flow along the IntraCoastal Waterway for the baseline simulation, this 
strong residual flow vanished without this waterway. This can explain the effect of the 
absence of IntraCoastal Waterway, which reduced the flow and salinity transport between 
Caloosahatchee River and San Carlos Bay. Overall, the IntraCoastal Waterway and Sanibel 
Causeway did not appear to show noticeable impact on the flow and salinity patterns in the 
San Carlos Bay and Pine Island Sound. 




o 
o 
o 
un 

CM 
OJ 

C\J 



st-06 



st-03 



st-05 A 



\ 






r~* 






J L 



J I L 



± 



J I I L 



± 



J I L 



390000 



395000 

Easting (m) 



400000 



Figure 5.32 The locations of Sanibel Causeway and Intracoastal Waterway and stations 
for comparing the effects of hydrologic alterations. 



149 



Depth 
(NAVD88 :cm) 

1500 
450 
400 
350 
300 
250 
200 
150 
100 




Depth 
(NAVD88 :cm) 

■ 500 

450 
400 
350 
300 
250 
200 
150 
100 






Figure 5.33 The comparison of bathymetry and shoreline for each hydrologic alteration case 
scenarios which are Baseline, the absence of IntraCoastal Waterway, and the absence of 
causeway. 



150 



ST05 



E 
o 



CD 
> 
CD 



S 

CO 



£ 

> 

CD 



CD 

*-» 
CO 



2 

CO 




185 



ST08 



60 
40 E- 

20 

o r 

-20 
-40 
-60 



185 



186 



186 



187 Julian Day 388 



Julian Day 



189 



190 




190 



BASELINE 
No Causeway 
NolCW 




189 



190 



^JulianDay 188 

Figure 5.34 The comparisons of water level for three cases at three selected stations: 
ST05 (Pine Island Sound), ST08 (San Carlos Bay), and ST10 (Caloosahatchee River 
mouth). 



151 



ST05 



35 



Q. 

330 

c 

I 25 

o 

o 

to 

•C20 
3 
W 



15 



BASELINE 
No Causeway 
NolCW 



186 



Julian Day 



ST08 



35- 



Q. 

-S30 
& 

C 

I 25 

a> 
o 

CO 

"C20 
3 
3) 




190 



ST05 



35 



S30 



c 

I 25 

£ 
o 
§20 1- 



15 



BASELINE 
No Causeway 
NolCW 



186 



. .- rO 88 

Julian Day 



1 n rJ 88 

Julian Day 



190 





ST08 






35 








a. 
-S30 








c 

s 25 




/ v V v 


V ' 


E 


?20 

n 


1 


1 


1 



190 




186 



1 h J 88 
Julian Day 



190 



186 



. ■• ,-J 88 
Julian Day 



190 



Figure 5.35 The comparisons of surface and bottom salinity for three cases at three 
selected stations: ST05 (Pine Island Sound), ST08 (San Carlos Bay), and ST10 
(Caloosahatchee River mouth). 



152 



Salinity 
(PPt) 

36 




Figure 5.36 The comparisons of surface residual flow and salinity fields for three cases. 









153 



'p$r*r^&*. '-. ^'t^-— 



Salinity 
(PPt) 

36 




35 

34 
33 
32 
31 
30 
29 
28 

fe 27 

i*ffim 26 

Figure 5.37 The comparisons of bottom residual flow and salinity fields for three cases. 



154 

Table 5.18 The effects of hydrologic alteration on 2000 water levels. Values shown are 
average RMS differences with baseline simulation for all selected stations. 

Station Range (cm) RMS error RMS error RMS error for 

Number for NSC forNICW NSCICW 



1 


126.58 


0.41 


0.16 


0.38 


2 


115.75 


0.76 


0.31 


0.93 


3 


115.01 


1.11 


0.40 


1.31 


4 


115.19 


0.55 


0.44 


0.92 


5 


116.69 


0.36 


0.19 


0.49 


6 


114.70 


0.43 


0.23 


0.60 


7 


114.53 


0.55 


0.30 


0.80 


8 


112.49 


0.63 


0.33 


0.90 


9 


110.60 


0.62 


0.68 


1.15 


10 


84.55 


0.40 


1.36 


1.22 


11 


111.49 


0.61 


0.39 


0.94 


Average 


126.58 


0.58 


0.44 


0.88 



155 



Table 5.19 The effects of hydrologic alteration on 2000 salinity. Values shown are average 
RMS differences with baseline simulation for all selected stations. 



Station 


Layers 


Range (ppt) 


RMS error 


RMS error 


RMS error for 


Name 






for NSC 


for NICW 


NSCICW 


1 


surface 


19.48 


0.51 


0.21 


0.49 




bottom 


11.07 


0.32 


0.08 


0.34 


2 


surface 


18.16 


0.78 


0.20 


0.80 




bottom 


11.80 


0.24 


0.06 


0.24 


3 


surface 


19.90 


0.66 


0.33 


0.70 




bottom 


18.46 


0.39 


0.20 


0.44 


4 


surface 


28.45 


0.40 


0.41 


0.50 




bottom 


26.58 


0.32 


0.21 


0.31 


5 


surface 


14.42 


0.22 


0.09 


0.21 




bottom 


9.53 


0.25 


0.07 


0.29 


6 


surface 


16.46 


0.27 


0.14 


0.28 




bottom 


12.07 


0.23 


0.08 


0.24 


7 


surface 


20.69 


0.44 


0.23 


0.50 




bottom 


11.92 


0.25 


0.10 


0.28 


8 


surface 


20.95 


0.44 


0.43 


0.61 




bottom 


20.26 


0.42 


0.40 


0.57 


9 


surface 


30.44 


0.44 


0.71 


0.69 




bottom 


30.37 


0.33 


0.51 


0.66 


10 


surface 


30.01 


0.32 


1.62 


1.79 




bottom 


30.45 


0.38 


1.33 


1.24 


11 


surface 


20.83 


0.21 


0.69 


0.80 




bottom 


18.18 


0.33 


0.37 


0.42 


Average 


surface 




0.43 


0.46 


0.67 




bottom 




0.32 


0.31 


0.46 



156 
Freshwater Inflow and Salinity in the Caloosahatchee River 

The Caloosahatchee River has been drastically altered for channelized flood-control 
and navigational waterway. These changes have caused large fluctuations in freshwater 
inflow volume, frequency of inflow events, timing of discharges, and water quality in the 
downstream estuary (Chamberlin and Doering, 1998). Therefore, it is necessary to quantify 
the impacts of freshwater inflow from S-79 on downstream estuarine system. 

A minimum flow is defined by Ch.373.042(l) F.S (Florida State Law) as "the limit 
at which further withdrawals would be significantly harmful to the water resources or 
ecology of the area." Significant harm is defined in Chapter 40E-8 F.S. as "the degree of 
impact requiring more than two years for the water (or biological) resource to recover". 
Establishing quantifying relationship between freshwater inflow and the temporal and spatial 
distribution of salinity in the river is the first step to determine MFL condition. 

Salinity distribution in San Carlos Bay and adjoining water directly responds to the 
fresh water inflow from Caloosahatchee River (Sheng and Park, 2001). In this study, the 
calibrated Charlotte Harbor model is used to evaluate the effects of the fresh water inflow 
from S79 to the salinity distribution in Caloosahatchee River. Using this calibrated model, 
a series of vertical-longitudinal salinity profiles along the axis of the Caloosahatchee River 
during one tidal cycle for wet and dry periods of river discharge at S-79 in 2000 are shown 
in Figures 5.38 and 5.39, respectively. The fresh water (less than 1 ppt salinity) from S-79 
reached the river mouth near Shell Point in wet season, while the fresh water stayed upstream 
near BR3 1 during dry season. The salinity distribution at Caloosahatchee River shows much 
difference corresponding to river discharge from S-79. 

To quantify this relationship for Caloosahatchee River, the locations of specific 



157 
salinity values were calculated during the 2000 simulation periods. Figure 5.40 shows the 
time histories of the river discharge rate and the locations of 1, 10, and 20 ppt surface salinity 
along the Caloosahatchee River. As the river discharge varied, the 1 ppt location is varied 
from S79 to near Shell Point. The tidal excursion of the salinity location is about 2 km for 
lppt salinity and 5 km for 20 ppt salinity. The salinity locations result from a combination 
of river discharge and tide and wind driven water circulation. 

To test the relationship between river discharge and salinity distribution, the river 
discharge at S79 was reduced 50% and increased 50% of current river discharge condition. 
To remove tidal effect, 1-day averaged salinity location was compared. Figure 5.41 shows 
the locations of 1-day averaged 1 ppt salinity location during the simulation period in 
response to varying river discharge. The results show that 1 ppt salinity location move 3.4 
km downstream and 8.5 km upstream, corresponding to a 50% reduction and a 50% increase 
in river discharge, respectively. 

The Caloosahatchee MFL rule (SFWMD) states that: "A MFL exceedance occurs 
during a 365-day period, when (a) 30-day average salinity concentration exceeds 10 parts per 
thousand at the Fort Myers salinity station or (b) a single daily averaged salinity exceeds a 
concentration of 20 parts per thousand at the Fort Myers salinity station. Exceedance of 
either subsection (a) or subsection (b), for two consecutive years are a violation." Figure 
5.42 show current conditions of salinity distribution to compare MFL rule for 
Caloosahatchee River during 2000 simulation periods. According to this result, the salinity 
does not exceed the Caloosahatchee MFL rule during 2000 simulation periods except winter 
dry season. 

To quantify the Caloosahatchee MFL rule due to river discharge, the locations of 10 



158 
ppt surface salinity values were calculated during the 2000 simulation periods. Figure 5.43 
shows the locations of 10 ppt surface salinity along the Caloosahatchee River due to river 
discharge rate. The polynomial regression line was calculated from the relationship between 
the locations of 10 ppt salinity and river discharge rate. According to this regression line, a 
total river discharge of 15 m 3 /s at S79 produces 10 ppt salinity at the Fort Myers station. 

An alternative and more acceptable approach to determine if the MFL has been 
exceeded, is to use a numerical mass-balanced model in which flows from different sources 
can be specified (Edwards et al., 2000). Therefore, the total river discharge at S79 required 
to produce a given salinity at Caloosahatchee River can be estimated with a calibrated model. 
Using the calibrated model, a diagram describing the relationship between river discharge 
and the locations of two salinity values (1-day averaged 20 ppt and 30-day averaged 10 ppt) 
in the river were generated in Figure 5.44. Twelve scenarios with constant river discharges 
of 5, 10, 15, 20, 30, 50, 75, 100, 150, 200 m 3 /s at S79 were simulated for 60 days. The 60 
day simulations allowed the model salinity to reach equilibrium condition for all specified 
river discharges. To compare current Caloosahatchee MFL rule with simulated salinity 
results, 30-day averaged 10 ppt salinity location and 1-day averaged 20 ppt salinity location 
were calculated for each case. According to this diagram, a total river discharge of 18 m 3 /s 
at S79 produces 30-day averaged 10 ppt salinity at the Fort Myers station. 

To quantify the relationship between salinity at Fort Myers station and river 
discharge, 1-day and 30-days averaged salinity at Fort Myers station were plotted vs. the 
fresh water inflow at S79 (Figure 5.45). The result show the minimum flow to produce a 
salinity of a 10 ppt at Fort Myers is about 18 m 3 /s, which is the same as that obtained form 
Figure 5.44. The contribution of river discharge at S79 to spatial and temporal salinity 



159 

distribution was successfully quantified with the integrated model for Charlotte Harbor 
estuarine system. This modeling approach would develop a management tool to establish 
the MFL criteria, with long term salinity and river discharge data. 



rShell Point 




5/10/2000 01:00 



-700 



40000 



-30000 




5/10/2000 04:00 



-700 



-40000 



-30000 



100 



-100 

-200 



j= -300 

8" -400 
Q 

-500 

-600 



-700 



5/10/2000 07:00 



-30000 




Fort Myers 



-700 



40000 



-30000 



-20000 
Distance from S79 



-20000 
Distance from S79 



-20000 
Distance from S79 



-20000 
Distance from S79 



BR31 



-10000 



-10000 



-10000 



-10000 



S-79 Salinity 
-* (PPt) 



Figure 5.38 The vertical-longitudinal salinity profiles along the axis of the 
Caloosahatchee River during wet season in 2000 




Salinity 
(PPt) 




Salinity 
(PPt) 

28 

24 

20 

16 

12 

8 

4 





Salinity 
(PPt) 




160 



-700 



40000 



-30000 



-20000 

Distance from S79 



-10000 



-700 



40000 



-30000 



-20000 

Distance from S79 



-10000 



-40000 



-30000 



-20000 
Distance from S79 



-10000 



-40000 



-30000 



-20000 
Distance from S79 



-10000 



S-79 Salinity 
> (PPt) 




Figure 5.39 The vertical-longitudinal salinity profiles along the axis of the 
Caloosahatchee River during dry season in 2000 



Salinity 
(PPt) 




Salinity 
(PPt) 




Salinity 
(PPt) 




161 



E 1 
ro 
.n 
o 
to 

I 



150 r 



100 - 




150 



200 250 

Julian Day 



300 



20 ppt 




Shell 
"^oint 



Fort 



ppt M y ers 



BR31 



S79 



Figure 5.40 Time histories of river discharge at S79 and the locations of 1, 10, and 20 ppt 
surface salinity along the Caloosahatchee River during 2000 simulation period. 



162 





i- 



o 150 



> 





200 



300 



Location of 1 ppt surface salinity 



45000 r 



40000 - 



£35000 r 



K30000 

E25000 
o 

^20000 


C15000 

*S 

■^10000 
Q 

5000 







current condition 

50% increased river discharge 

50% reduced river discharge 




Julian Day 



300 



Figure 5.41 Time histories of river discharge at S79 and the 1 ppt salinity location along 
Caloosahatchee River. 



163 



£ 

Q 



45000 
40000 
35000 
"'30000 
25000 




0)20000 
o 

c 

*)j 15000 

Q 

10000 



5000 - 



1-day averaged surface salinity location of 20 ppt 






rvn. 






Shell 
Point 




Fort 
Myers 



30-day averaged surface salinity location of 1 ppt\ 



BR31 



J L 



X 



J I I L 



200 

Julian Day 



300 



J L_ S79 



Figure 5.42 The 1-day averaged 20 ppt surface salinity location and 30-day averaged 10 ppt 
surface salinity location during the 2000 simulation period. 



164 




20000 

O 

C i / 

ra §K«-\ • Minimum Flow = 15 m /s 

■§15000 
Q 



10000H 
5000 



_L 



Shell Point 



Location of 10 ppt salinity 
Polynomial regression line 



-Fort Myers 



•BR31 



J I L 



_L 



J L 



S79 







50 



100 



150 



200 



River discharge (m /s) 



Figure 5.43 The locations of 10 ppt surface salinity due to river discharge rate at S-79 
during 2000 baseline simulation. 



165 



30 day averaged 1 ppt salinity 
1 day averaged 20 ppt salinity 




Shell Point 



Fort Myers 



■§15000 

Q 

10000 



BR31 



i i i i 



I I I L 



S79 



50 100 150 

River discharge (m /s) 



200 



Figure 5.44 The relationship between locations of specific salinity value vs. river 
discharge at S-79. 



166 




♦ 1 -day averaged salinity 

B 30-day averaged salinity 



Minimum Flow = 18 m /s 



- »--- 



■ | <■ ■ f ■ .|„ ■ | ,„ ■!■ - t - --t~ 



100 



150 



200 



River discharge (m /s) 



Figure 5.45 The relationship between salinity at Fort Myers station vs. river discharge at 
S-79 



CHAPTER 6 
APPLICATION OF WATER QUALITY MODEL 

The water quality model described in Chapter 4 was applied to the Charlotte Harbor 
estuarine system. Model applications include the simulation of the three-dimensional 
circulation, sediment transport and water quality processes during summer 1996, the entire 
2000. These simulations provide calibration and validation of the Charlotte Harbor water 
quality model using field data obtained by USGS, SWFWMD, and SFWMD. In addition, 
model simulation provide assessment of (1) the effects of hydrologic alteration of causeway 
and navigation channels, (2) the impact of river pollutant loading and its reduction on the 
water quality in the Charlotte Harbor estuarine system, and (3) the causes for hypoxia of 
bottom water in the upper Charlotte Harbor. 

To reduce the computational time necessary to simulate annual response of the 
Charlotte Harbor estuarine system, the parallel CH3D model developed by Sheng et al. 
(2003) is used. The detail description of validation, CPU time, and speedup of the parallel 
CH3D code are given in Appendix G. 

6.1 Forcing Mechanism and Boundary Condition of Circulation 

The tidal forcing along the Gulf of Mexico boundary is prescribed by the water level 
data measured at Naples by NOS (http://www.co-ops.nos.noaa.gov/data_res.html). The 
surface wind boundary condition is produced by using the hourly wind magnitude and 
direction data collected at the National Data Buoy Center C-MAN stations at Venice and 



167 






168 
USGS station at Naples and Fort Myers. The hourly wind magnitude and direction are 
converted into EastAVest(-x) and North/South (-y) wind velocity components and then 
interpolated onto the entire computational grid. 

Daily river discharges for Peace River, Shell Creek, Myakka River, Caloosahatchee 
River (S-79 spillway, Cape Coral and Whiskey Creek) and Estero Bay (Mullock & Hendry 
Creek, Estero River, Spring Creek, and Imperial River) were measured by the USGS daily 
(Table 5.1). Figures 6.1 to 6.3 show the water level at the tidal boundary, river flows at the 
river boundary, wind speed and direction at the surface boundary, and air temperature for the 
air-sea heat flux in 1996. The forcing mechanism and boundary condition for circulation 
of 2000 simulation were explained at Chapter 5. 



169 




1 50 200 

Julian Day 



300 r 



Caloosahatchee River 
Peace River 
Shell Creek 
Myakka River 




150 



200 



Julian Day 

Figure 6.1 Tidal forcing and river discharges for 1996 simulations of Charlotte Harbor. 



at Venice 



170 




_1_ 



-J I : i_ 



150 160 170 180 190 200 

Julian Day 



210 



220 



230 



at Fort Myers 




150 160 170 180 190 200 210 220 230 

Julian Day 

Figure 6.2 Wind velocity for 1996 simulations of Charlotte Harbor. 



171 



at Venice 




150 



Julian Day 



200 



at Fort Myers 




150 



200 



Julian Day 

Figure 6.3 Air temperature for 1996 simulations of Charlotte Harbor. 



172 
6.2 Initial and Boundary Condition for the Water Quality Model 

For 1996 simulations, there were 21 stations sampled monthly for water quality data 
in 1996 by SWFWMD and SFWMD (Table 6.1). The locations of the measurement sites are 
shown in Figure 6.4. Most of the data were collected from May to July 1996. For Estero 
Bay, there were 14 sites sampled for total nitrogen, total phosphorous, pH, dissolved oxygen, 
temperature, turbidity and chlorophyll_a data in June 1996 by SFWMD. The initial water 
column concentrations of several water quality parameters are determined from the EPA 
data. The water quality data for May are used to produce the initial condition of 1996 
simulations. The water quality data at each grid cell are calculated by interpolation of the 
data at the three closest data stations, with a weighting function inversely proportional to the 
distance from each of the three stations. Data collected at station CH001, CH029, and 
CH004 are used to provide river loading data of Myakka River, Peace River, and Horse 
Greek, respectively. There are three river boundary conditions at Caloosahatchee River, 
which are S-79 spillway, Cape Coral, and Whiskey Creek. For these three river boundary 
conditions, data at HB01, HB03, and HB04 were used. The water quality data at stations 
EB002, EB004 EB01 3 and EBO 1 2 are used for four river boundary conditions of Estero Bay. 

For 2000 simulation, the initial water column concentrations of several water quality 
parameters were determined from SWFWMD and SFWMD data of January 2000. Figure 
6.5 shows the locations of SWFWMD (CH-001 to CH-014) and SFWMD (CES01 to CES08) 
water quality monitoring stations (Table 6.1). Data collected at station CH001, CH029, 
CH004, CES01, CES05, and CES08 are used to provide river loading data of each river 
boundary. 



173 



Table 6.1 Locations of water quality measured stations 



station 



Latitude Longitude 



XUTM 



YUTM 



(l,J) in Grid Agency 



CH-001 


27 00 072 


82 15 108 


CH-002 


26 57 216 


82 12 300 


CH-02B 


26 58 048 


82 1 1 048 


CH-029 


27 00 588 


81 59 030 


CH-004 


26 56 396 


82 03 324 


CH-005 


26 55 558 


82 06 156 


CH-05B 


26 57 114 


82 06 330 


CH-006 


26 54 006 


82 07 090 


CH-007 


26 52 396 


82 04 072 


CH-009 


26 49 132 


82 05 294 


CH-09B 


26 53 132 


82 09 288 


CH-011 


26 44 120 


82 10 000 


CH-013 


26 41 300 


82 18 010 


CH-014 


26 39 310 


82 19 210 


HB-001 


26 41 480 


81 49 277 


HB-002 


26 40 007 


81 52 213 


HB-003 


26 38 090 


81 54 294 


HB-004 


26 33 576 


81 54 534 


HB-005 


26 30 570 


81 59 024 


HB-006 


26 29 250 


82 01 150 


HB-007 


26 32 264 


82 07 324 


CES01 


26 43 199 


81 41 233 


CES02 


26 43 354 


81 42 284 


CES03 


26 43 001 


81 45 382 


CES04 


26 40 541 


81 50 017 


CES05 


26 38118 


81 53 193 


CES06 


26 34 563 


81 54 367 


CES07 


26 31 488 


81 57 562 


CES08 


26 31 239 


82 00 312 



375674.130 2987267.910 

380063.200 2982147.510 

382419.400 2983448.440 

402367.040 2988631 .380 

394887.010 2980722.130 

390352.680 2979406.960 

389904.080 2981718.790 

388859.690 2975881.310 

393859.500 2973345.640 

391542.120 2966995.710 

384983.960 2974439.100 

383975.740 2957800.680 

370636.040 2952944.680 

368386.980 2949305.790 

417982.560 2953104.130 

413179.170 2949843.840 

409616.420 2946422.590 

408897.610 2938704.790 

401965.780 2933186.980 

398261.910 2930385.140 

387873.180 2936041.360 

431400.582 2955824.308 

429607.466 2956326.423 

424351.809 2955279.834 

417059.745 2951448.633 

411552.356 2946470.516 

409380.629 2940485.917 

403804.420 2934742.217 

399508.606 2934006.048 



(38,123) 


SWFWMD 


(37,120) 


SWFWMD 


(42,119) 


SWFWMD 


(71,117) 


SWFWMD 


(60,116) 


SWFWMD 


(53,116) 


SWFWMD 


(54,118) 


SWFWMD 


(47,113) 


SWFWMD 


(57,111) 


SWFWMD 


(47,106) 


SWFWMD 


(38,113) 


SWFWMD 


(29,97) 


SWFWMD 


(9,98) 


SWFWMD 


(7,95) 


SWFWMD 


(79,58) 


SFWMD 


(76,58) 


SFWMD 


(73,58) 


SFWMD 


(70,55) 


SFWMD 


(63,50) 


SFWMD 


(45,43) 


SFWMD 


(24,70) 


SFWMD 


(91 ,56) 


SFWMD 


(89,56) 


SFWMD 


(84,56) 


SFWMD 


(78,57) 


SFWMD 


(74,56) 


SFWMD 


(71,55) 


SFWMD 


(65,54) 


SFWMD 


(58,51) 


SFWMD 



174 
Light and temperature are the main limiting factors for the phytoplankton growth rate. 
Temperature is calculated by solving the heat equation with the heat flux model as a surface 
boundary condition, which was explained in chapter 3 and 4. The input data for the 
temperature model include an air temperature and cloud cover rate. Continuous hourly air 
temperature data collected at the NOS data collection platforms (DCP) and stored in the CO- 
OPS databases are available for the Naples and Vince stations. Since there is no available 
data for cloud cover, a constant cloud cover of 0.2 is used for all simulations. 

To calculate PAR (Photosynthetically Active Radiance) in the light attenuation 
model, the light intensity at the surface is used as the surface boundary condition. The light 
intensity data (Langleys/day) are converted from the global and diffuse horizontal solar 
irradiance data (W/m 2 ) which were processed at the National Renewable Energy Laboratory 
(NREL). Figure 6.6 shows the light intensity data used in the 1996 and 2000 water quality 
model simulations. 

A total of 215 sediment grab samples and 28 shallow cores were collected from the 
Charlotte Harbor estuarine system during the period of December 27, 1964, to January 1, 
1965 by Huang (1965). Based on the sediment size distribution, the entire study area is 
characterized into five sediment types (Table 6.2). 
Table 6.2 Sediment types for Charlotte Harbor water quality simulations 



Type 


D 50 Range ( mm) 


Category 


Remark 


very coarse 


D 50 >0.50 


5 




coarse 


0.25 > D 50 > 0.50 


4 




medium 


0.125 >D 50 >0.25 


3 




fine 


0.0625 >D 50 > 0.125 


2 




silts or clay 


0.0625 > D 50 


1 





Sediments were predominantly sand, with low amounts of silt, clay, and organic 



175 
matter. Silt or clay deposits were only observed at the mouth of Peace River. This type of 
sediment can be considered fine or cohesive sediment. Fine sediments exist in upper 
Charlotte Harbor, the southern part of Pine Island Sound, and the northern part of San Carlos 
Bay due to weak wave actions there. Sediment nutrient analyses were performed by the 
FDEP Sediment Contaminant Survey with data from 33 sample stations from 1985 to 1989 
(Schropp, 1998). These data include organic carbon, total nitrogen, and total phosphorous. 
Based on water quality data, sediment data, bathymetry, and geometry, the Charlotte 
Harbor study area is divided into 15 segments as shown in Figure 6.7 for the water quality 
simulations. 



176 



o 
o 
o 

O) 

cm 



C 

IE 

■c 
o 

Z o 
o 

o 

CO 
CM 
OJ 

CM 



o 
o 
o 
o 
o 

CD 
CM 



I 



X 



CH-lp'l 

\- : . CH-02| 



fCH-029 



V4,. ^hdS iKU-.n 



^ . CH-09B* •^r 07 



At' 



% 



ak •cfi-009 

vj f 



• ch-01 r 

CH-013* (k« r €*^ J 
CH-01 4 • W%f &*\ % i 



HB-001 Jft^-yr- 

v.fc*" 



. •'HB-002 



^^? or * 



fl 



<K: 



v 



1. 



i i i i 



_L 



J I I L 



J I I L 



± 



i i i 



350000 



375000 400000 

Easting (m) 



425000 



Figure 6.4 Locations of 1996 water quality measurement stations operated by EPA 



177 



c 
o 



o 
o 
o 
m 
t-~ 

CM 



o 
o 
o 
m 
cm 

CD 
CM 



o 
o 
o 
o 
o 
<J1 

C\J 



_L 



X 



CH 



GH-02B, 



fcH-029 



X 



„ •CH-Q06 
CH-09B* ^Hrd07 

WMEjjfaiJK •Cf-009 




• CH-01,1 



J_ 




CES003 C i? 002 

*&£ cisob-i 

ES004 



J_ 



J L 



-L 



J L 



-L 



_ 



350000 



375000 



400000 
Easting (m) 



425000 



Figure 6.5 Locations of 2000 water quality measurement stations operated by SFWMD 
and SWFWMD 



178 



2500 - 



2000 



£ 1500 
c/) 

C 

S, 

C 1000 

+■> 
JC 

.2> 500 



— 



; 'Hi , ! [ , 



120 140 160 180 200 

Julian Day 



For 1996 



220 240 

For 2000 




50 100 150 200 250 300 

Julian Day 

Figure 6.6 Light intensity at water surface for 1996 and 2000 simulations 



350 



179 



\ 



& 



\ 



u 



"V 



fe 



i~\^--Y 



WW 



kV' 2 /-^ 




-*<w~" 










4,. 







# 



K 



# 






45*sr 



P 



*ui/ 



,/t 



X S*^*K*f k «fc 






r 









Figure 6.7 Segments for Charlotte Harbor estuarine system. 



180 
6.3 Simulations of Water Quality in 1996 

CH3D-IMS are used to simulate the circulation, sediment transport, and water quality 
dynamics of the Charlotte Harbor estuarine system during May 23 to August 21, 1996. To 
create an appropriate initial condition throughout the computational domain, a spin-up 
simulation was executed until the flow and salinity field reached a dynamic steady state. The 
river discharge data for this area showed that the simulation period can be divided into a dry 
season which lasts from January 1 until June 8 and a wet season which spans from June 9 
until July 22 (Sheng and Park, 2002). A 30-day spin-up simulation was performed from 
April 23 to May 23 (dry season) with all forcing mechanisms (tides, river discharges, wind) 
to allow water level, velocity and salinity field to reach dynamic steady-state throughout the 
computational domain. Using the surface elevation, velocity, and salinity at the end of the 
spin-up simulation as the initial condition, the Charlotte Harbor circulation and water quality 
models are then run from May 23 to August 21 with hydrodynamic, sediment and water 
quality input data. 
6.3.1 Calibrations 

The calibration of water quality model was conducted using the monthly water 
quality data obtained from the EPA STORET database. To compare the water quality results 
with measured data, fourteen stations were selected: CH002 (near Myakka), CH005 (near 
Peace), HB02, HB03 and HB06 (near Caloosahatchee), CH006, CH007, CH09B and CH009 
in upper Charlotte Harbor, CHOI 1 in lower Charlotte Harbor, HB007 in Pine Island Sound, 
HB006 in San Carlos Bay, and CHOI 3 and CH014 in the open sea. 

Several initial water quality parameters for the Charlotte Harbor simulation are 
obtained from "A Mechanistic Water Quality Model for the Tidal Peace and Myakka Rivers" 



181 
(Pribble et al, 1997) and Indian River Lagoon simulations (Sheng et al., 2001) shown in 
Chapter 4. 

In order to test the model response to variation of these specific parameters, the 
sensitivity test was conducted for each water quality parameter by performing 90-day 
simulations and comparison with the baseline simulation. These simulations used the same 
initial and boundary conditions, and external forcing as those for the baseline simulation. 
Therefore, variations in nutrient concentrations between the sensitivity test and the baseline 
simulation can be directly related to variation of each parameter in a sensitivity test. Table 
6.3 shows the parameters considered in the sensitivity analysis, baseline values in the 
simulations and the parameter variations of each sensitivity test. The tests were performed 
by varying each parameter within reasonable ranges according to literature survey. Table 6.4 
shows the sensitivity analysis results which are presented in terms of percent RMS difference 
in concentration, normalized by concentrations of each water quality species in the baseline 
simulation. The SUM in Table 6.4 is linearly averaged value of RMS errors for each 
sensitivity test. 

The results of the tests showed the nitrogen half saturation constant (HALN) for 
uptake as the most sensitive parameter in the water quality model, followed by the maximum 
algae growth rate (AGRM). These two parameters are related and their major impact should 
be detected in the phytoplankton biomass. However, the chlorophyll_a concentration is more 
sensitive to ammonification than maximum algae growth rate or nitrogen half saturation 
constant, which reveals the extension of nitrogen limitation to phytoplankton growth. 
Morever, the effect of reducing the maximum growth rate (AGRM0.5) is more pronounced 
in chlorophyll_a concentration than increasing this rate (AGRM2.0). Result of test 



182 
AGRM2.0 show that ammonia nitrogen (NH4) and nitrate+nitrite (N03) is rapidly uptaked 
by phytoplankton and phytoplankton is increased due to increasing growth rate. If more 
ammonia nitrogen and nitrate+nitrite are available in the system, phytoplankton will increase 
more than that for this sensitivity test simulation. Related with algae growth rate and 
nitrogen half saturation constant, algae respiration rate and algae mortality are also important 
parameters. Therefore, the coefficients related with algae growth rate are most important 
factors for overall water quality processes. 

The third most important parameter revealed by the sensitivity test is an 
ammonification rate. Soluble organic nitrogen (SON) is rapidly mineralized to ammonium 
nitrogen. Nitrate (N03) level also increased due to nitrification. Due to high concentrations 
of inorganic nitrogen, which is a food for phytoplankton growth, chlorophyll_a concentration 
increased. On the opposite side, decreasing an ammonification rate promoted an increase in 
SON, and decrease in NH4, N03, and chlorophyll_a. 

The information obtained from the sensitivity tests enabled a more systematic and 
efficient calibration of model coefficients described in Table 4.9 of Chapter 4. According 
to the sensitivity test, the six most important parameters, which include maximum algae 
growth rate, nitrogen half saturation constant, ammonification rate, algae mortality, algae 
respiration rate, and sorption/desorption rate for SON and PON, are adjusted first for all 
water quality species as part of the systematic calibration procedure. The other parameters 
are adjusted as partially sensitive parameters for each specific water quality species. More 
than 100 simulations were made during calibrations which include all the sensitivity 
analyses. During these simulations, kinetic coefficients were adjusted within accepted 
tolerances, estimated loads were reviewed and adjusted, and new processes were added or 



183 



modified in water quality model, if necessary. Listed in Table 6.5 are values for the 
calibration parameters described in Chapter 4 and Table 4.3. 

Table 6.3 Water quality parameters, baseline values, and range used in the sensitivity analysis 

Test Run Parameter Literature Baseline Test run parameter 

Range Parameter =Baseline parameter 
*multiplier 



AGRM2.0 
AGRM0.5 


Maximum algae 
growth rate 


0.2-8 


2.0 - 2.2 


2.0 
0.5 


HALN10.0 
HALN0.5 


Nitrogen half 
saturation rate 


1.5-400 


25 


10.0 
0.5 


HALP10.0 
HALP0.5 


Phosphorous half 
saturation rate 


1.0- 105 


2 


10.0 
0.5 


KAEX2.0 
KAEX0.5 


Algae respiration 
rate 


0.02 - 0.24 


0.06 


2.0 
0.5 


KAS2.0 
KAS0.5 


Algae mortality 


0.01-0.22 


0.07 


2.0 
0.5 


WAS 10.0 
WAS0.1 


Algae settling 
velocity 


0.0 - 300 


10 


10.0 
0.1 


HALA2.0 
HALA0.5 


Algae half 
saturation rate 


200 - 2000 


200 


2.0 
0.5 


SONM2.0 
SONM0.2 


Ammonification 
rate 


0.001 - 1.0 


0.015 


2.0 
0.2 


NTTR2.0 
NITR0.5 


Nitrification Rate 


0.004-0.11 


0.08 


2.0 
0.5 


DRON2.0 
DRON0.5 


Sorption/desorption 
rate of SON/PON 


0.02 - 0.08 


0.03 


2.0 
0.5 


DRAN2.0 
DRAN0.5 


Sorption/desorption 
rate of PIN/NH4 


0.02 - 0.08 


0.03 


2.0 
0.5 


SOPM2.0 
SOPM0.2 


Mineralization rate 


0.001-0.6 


0.02 


2.0 
0.2 


DROP2.0 
DROP0.5 


Sorption/desorption 
rate of SOP/POP 


- 


0.02 


2.0 
0.5 


DRIP2.0 
DRIP0.5 


Sorption/desorption 
rate of SRP/PIP 


- 


0.02 


2.0 
0.5 


SODM2.0 


Sediment oxygen 
demand 


0- 10.7 


0.5 - 2.0 


2.0 


AKD2.0 
AKD0.5 


Oxidation rate 


0.02 - 0.6 


0.05 


2.0 

0.5 



184 



Table 6.4 Sensitivity analysis results in 
quality calibration simulations 


RMS d 


ifference 


w.r.t. b 


aseline 1 


or 199 


6 water 




ChlA 


DO 


NH4 


NOX 


TKN 


P04 


PHOS 


TOC 


SUM 


BASE 


0.00 


0.00 


0.00 


0.00 


0.00 


0.00 


0.00 


0.00 


0.00 


AGRM0.5 


17.63 


4.29 


55.20 


171.6 


3.31 


25.53 


8.47 


9.51 


36.9 


AGRM2.0 


5.68 


2.54 


12.68 


12.72 


0.97 


5.00 


1.91 


1.81 


5.42 


HALN0.5 


2.03 


0.73 


6.23 


5.92 


0.47 


1.99 


0.78 


0.89 


2.38 


HALN10.0 


18.56 


4.26 


65.73 


204.0 


3.71 


37.54 


10.61 


10.99 


44.4 


HALPO.5 


0.31 


0.07 


0.35 


0.67 


0.03 


0.56 


0.10 


0.11 


0.28 


HALP10.0 


4.02 


0.77 


6.07 


13.27 


0.45 


12.03 


1.89 


1.76 


5.03 


KAEXO.5 


2.13 


0.92 


7.17 


5.24 


0.54 


2.58 


0.87 


0.92 


2.55 


KAEX2.0 


5.43 


2.08 


18.59 


17.27 


1.35 


6.87 


2.33 


2.55 


7.06 


KASSO.5 


14.75 


6.23 


2.64 


4.69 


3.47 


1.93 


1.49 


9.11 


5.54 


KASS2.0 


19.58 


5.59 


7.56 


13.49 


3.99 


2.47 


2.20 


9.29 


8.02 


WASSO.l 


0.57 


0.08 


0.22 


0.26 


0.04 


0.08 


0.05 


0.17 


0.18 


WASSOIO 


5.51 


0.81 


2.43 


2.94 


0.47 


0.84 


0.50 


1.67 


1.90 


HALAO.5 


13.47 


3.03 


7.88 


16.84 


0.81 


2.89 


0.84 


4.09 


6.23 


HALA2.0 


13.13 


3.11 


2.76 


4.54 


1.00 


1.20 


0.28 


3.62 


3.70 


SONM0.2 


22.22 


3.25 


22.71 


15.43 


19.75 


72.50 


16.26 


17.26 


23.6 


SONM2.0 


12.47 


3.66 


32.01 


42.19 


13.49 


19.84 


7.96 


8.96 


17.5 


NITRO.5 


0.12 


0.11 


2.51 


10.39 


0.35 


0.17 


0.06 


0.07 


1.72 


NTTR2.0 


0.20 


0.17 


4.09 


17.36 


0.55 


0.29 


0.11 


0.12 


2.86 


DRONO.5 


3.19 


0.90 


6.03 


6.27 


10.87 


6.46 


2.44 


3.13 


4.91 


DRON2.0 


4.47 


1.60 


11.52 


15.81 


17.05 


9.31 


3.58 


4.42 


8.47 


DRANO.5 


1.46 


0.40 


1.74 


1.65 


0.71 


2.94 


1.12 


1.31 


1.42 


DRAN2.0 


2.32 


0.68 


2.96 


3.01 


1.12 


4.58 


1.74 


1.99 


2.30 


SOPM0.2 


2.24 


0.29 


3.42 


7.15 


0.29 


10.14 


2.96 


1.07 


3.44 


SOPM2.0 


1.68 


0.21 


2.68 


4.21 


0.23 


9.15 


2.68 


0.91 


2.72 


DROPO.5 


0.70 


0.08 


1.07 


1.94 


0.09 


2.61 


2.73 


0.37 


1.20 


DROP2.0 


0.55 


0.07 


0.83 


1.61 


0.07 


2.13 


2.21 


0.29 


0.97 


DRIPO.5 


1.99 


0.27 


3.16 


6.42 


0.27 


8.43 


3.32 


0.94 


3.10 


DRIP2.0 


2.34 


0.31 


3.94 


5.74 


0.34 


14.75 


5.73 


1.28 


4.30 


SODM2.0 


0.01 


13.06 


0.24 


1.72 


0.03 


0.02 


0.01 


0.58 


1.96 


AKDDO.5 


0.00 


3.26 


0.05 


0.26 


0.01 


0.00 


0.00 


6.14 


1.21 


SUM 


5.96 


2.09 


9.82 


20.49 


2.86 


8.83 


2.84 


3.51 


7.05 



185 



Table 6.5 The water quality model coefficients used for the Charlotte Harbor simulation 


Coefficient 


Description 


Units 


Literature 
Range 


Charlotte 
Harbor 


(6 AD ) T - 20 


temperature coefficient for 
NH4 desorption 


- 


1.08 


1.08 


(6J T - 20 


temperature coefficient for 
algae growth 


- 


1.01-1.2 


1.08 


(0 A ,) T - 20 


temperature coefficient for 
ammonium instability 


- 


1.08 


1.08 


(e BOD ) T - 20 


temperature coefficient for 
CBOD oxidation 


- 


1.02-1.15 


1.08 


(6 D n) t - 20 


temperature coefficient for 
denitrification 


- 


1.02-1.09 


1.08 


(6 NN ) T - 20 


temperature coefficient for 
nitrification 


- 


1.02-1.08 


1.04 


Ood) 7 " 20 


temperature coefficient for 
SON desorption 


- 


1.08 


1.045 


/o xT-20 
V u ONM/ 


temperature coefficient for 
mineralization 


- 


1.02-1.09 


1.08 


re V- 20 

\VreSPS 


temperature coefficient for 
algae respiration 


- 


1.045 


1.08 


(0z) T-2O 


temperature coefficient for 
zooplankton growth 


- 


1.01-1.2 


1.04 


VH'a/max 


algae maximum growth 
rate 


1/day 


0.2-8. 


2.0-2.1 


VH'z/max 


zooplankton maximum 
growth rate 


1/day 


0.15-0.5 


0.16 


(NH 3 ) air 


ammonia concentration in 
the air 


Hg/L 


0.1 


0.1 


a chla 


algal carbon-chlorophyll-a 
ratio 


mg C/mg 
Chla 


10- 112 


100 


a nc 


algal nitrogen-carbon ratio 


mg N /mg C 


0.05-0.43 


0.16 


^c 


algal phosphorous-carbon 
ratio 


mg N /mg C 


0.005-0.03 


0.025 


a oc 


algal oxygen-carbon ratio 


mg 2 /mg C 


2.67 


2.667 



186 



Coefficient 


Description 


Units 


Literature 

Range 


Charlotte 
Harbor 


dan 


desorption rate of adsorbed 
ammonium nitrogen 


1/day 


- 


0.03 


don 


desorption rate of adsorbed 
organic nitrogen 


1/day 


- 


0.03 


dip 


desorption rate of adsorbed 
inorganic phosphorous 


1/day 


- 


0.02 


d op 


desorption rate of adsorbed 
organic phosphorous 


1/day 


- 


0.02 


d^l 


molecular diffusion 
coefficient for dissolved 
species 


cm 2 /s 


4.E-6-1.E-5 


l.E-5 


Hbod 


half-saturation constant for 
CBOD oxidation 


mgO-, 


0.02-5.6 


0.5 


H n 


half-saturation constant for 
algae uptake nitrogen 


g/L 


1.5-400 


25 


H p 


half-saturation constant for 
algae uptake phosphorous 


g/L 


1.-105 


2 


Hnit 


half-saturation constant for 
nitrification 


mg0 2 


0.1-2.0 


2.0 


h v 


Henry's constant 


mg/L-atm 


43.8 


45 


Is 


optimum light intensity for 
algal growth 


|J.E /m 2 /s 


300-350 


350 


K ax 


excretion rate by algae 


1/day 


0.02-0.24 


0.06 


K as 


mortality rate of algae 


1/day 


0.2-0.22 


0.07 


K AI 


ammonia conversion rate 
constant 


1/day 


0.01-0.1 


0.01-0.02 


K D 


CBOD oxidation rate 


1/day 


0.02-0.6 


0.05 


^DN 


denitrification rate constant 


1/day 


0.02-1.0 


0.09 


^NN 


nitrification rate constant 


1/day 


0.004-0.11 


0.08 


K on 


rate of ammonification of 
SON 


1/day 


0.001-1.0 


0.015 



187 



Coefficient 


Description 


Units 


Literature 
Range 


Charlotte 
Harbor 


K op 


rate of mineralization of 
SOP 


1/day 


0.001-0.6 


0.02 


K V oi 


rate constant for nitrogen 
volatilization 


1/day 


3.5-9.0 


7. 


K zs 


mortality rate of 
zooplankton 


1/day 


0.001-0.36 


0.02 


Kzx 


excretion rate of 
zooplankton 


1/day 


0.003-0.075 


0.01 


Pan 


partition coefficient 
between SAN and PEN 


I/Jig 


.5E-7-1.E-5 


1E-4 


Pon 


partition coefficient 
between SON and PON 


1/M-S 


1.0E-5 


LE-5 


Pip 


partition coefficient 
between SRP and PEP 


i/ng 


- 


1E-5 


Pop 


partition coefficient of 
SOP and POP 


m 


- 


E-4 


W^CBOD 


CBOD settling velocity 


cm/s 


- 


0.01 


^*^algae 


algae settling velocity 


cm/s 


0.0-300. 


10. 



Figure 6.8 contains the scattering plots for the calibration period. The location of 
circles indicates the correlation between model predictions and observed data. A perfect 
match between model and observed data is indicated by the diagonal line on each graph. The 
circle above the line is over predicting the observation. Circles below the line indicate that 
the model is under predicting the observation. Shown with each plot are the root mean 
square (RMS) error and correlation coefficient (R 2 ) mentioned before. Values shown in 
parenthesis are RMS errors normalized with maximum measured data. 

Overall, the model results show reasonable RMS errors for all constituents, although 
the R 2 values are not very high. The normalized RMS errors are less than 50%, except NH4, 
P04, and TSS. Dissolved oxygen has the best agreement with measured data as shown by 



188 
the RMS error. Total suspended sediment concentration changed quickly with time on the 
order of minutes and hours, hence comparison with monthly data did not show reasonable 
RMS error. Dissolved ammonium nitrogen and soluble reactive phosphorous are strongly 
affected by sorption and desorption processes, while particulate species are affected by 
dynamic processes of settling and erosion of suspended sediment. The water quality model 
calculates CBOD instead of total organic carbon. The simulated CBOD was compared with 
measured total organic carbon because there is no measured CBOD data. Hence, the amount 
of CBOD was under predicted as shown in the scatter plot, due to the difference between 
CBOD and total organic carbon concentration in the system. 

The correlation coefficient (R 2 ) measures the strength of linear association between 
simulated and measured data. Correlation coefficients for all species vary form 0.2 to 0.88. 



189 



A) Chrolophyll a 

25 



B) Dissolved Oxygen (DO) 



RMS=7.51 (33.56%) 
R2 =0.519 



20 

£ 15 
w 

1 10 







C) Total Kjeldahl Nitrogen (TKN) 



500 1000 

measured 





IU 


RMS= 1.40 (17.63%) # X 






R2 =0.308 ** ,X 






* •<*<%/. 






•* ,4r 






• ui?v. 


(1) 






3 


5 




E 




• -X' • 


c/) 


n 





5 10 15 20 25 

measured 



IOUU 


RMS=0.228(21.98 


%) / 




R2 =0.462 


•^r 




• 




simulated 

o o 

o o 


. . *x • 

*3i» • • • 


• 

• 
• 

• 
i , , , , 



measured 

D) Dissolved Ammonium Nitrogen (NH ,) 

250 



10 



RMS= 0.040 (44.48%) 
R2 = 0.362 



150( 




100 200 

measured 



Figure 6.8 The scatter plots for water quality constituents during calibration period 



190 



E) Total Phosphorous 

450 



400 
350 



RMS= 0.0783 (31.26%) 
7 R2 =0.242 



F) Soluble Reactive Phosphorous (P0 4 ) 

250 1 




100 200 300 400 

measured 

G) Total Organic Carbon (CBOD) 



1 ■•* » i i ■ I ■ . i i 1 i i i ■ I i ■ 





£U 








RMS= 3.320 (32.13%) / 






R2 = 0.806 »A* 




15 


/>•'• 


■a 




f • 


0) 




/ * • 


*•> 




X • 


m 




/ • . • 


=j 


10 


s • 


£ 




/ •*:.# 


w 




A* • \ . 




5 
n 


/ J* *'» * 
AW. 



50 100 150 200 250 

measured 

H) Total Suspended Sediment (TSS) 

150 



RMS= 25.67 (36.45%) 
R2 =0.124 



"0 5 10 15 

measured 

Figure 6.8 Continued 



20 




««h< 



50 100 

measured 



150 



191 
6.3.2 Results of 1996 Water Quality Simulation 

The temporal variation of water quality species at each measured station is compared 
with measured data in Figures 6.9 to 6.20. The thicker solid lines represent the simulated 
water quality parameters near surface and dashed lines are those for bottom layer. The 
measured water quality parameters are represented with rectangular symbols which represent 
several vertical layers. The numbers of vertical layer of measured data are all different for 
each species and each station. The measured data were plotted at all available vertical 
measured location. The results show simulated and measured Chlorophyll_a, dissolved 
oxygen, TKN, dissolved ammonium nitrogen, total phosphorous, and soluble reactive 
phosphorous. During the simulation period, there was relatively little temporal variation in 
the water quality parameters. One major exception is at the CH-005 and CH-006 station 
which showed a significant vertical stratification of dissolved oxygen concentration. 

The dissolved oxygen processes are important in any aquatic environment, because 
living organisms depend on oxygen in one form or another to maintain their metabolic 
processes. In Charlotte Harbor estuarine system, bottom water hypoxia has been reported 
periodically by Environmental Quality Laboratory (EQL) since the mid-1970 (Heyl, 1996). 
In the water quality model, dissolved oxygen is a function of photosynthesis and respiration 
by phytoplankton organisms, sediment oxygen demand, reaeration, nitrification and 
denitrification, decomposition of organic matter, tide and wind mixing, and river loading. As 
shown in Figures 6.9 to 6.20, concentrations of dissolved oxygen exhibit a temporal and 
vertical variation in response to variations in phytoplankton biomass and nitrogen species in 
the northern part of Charlotte Harbor, while relatively little variation in the southern part of 
Charlotte Harbor during the simulation period. Figure 6-21 shows the snapshots of the near 



192 
simulated bottom dissolved oxygen distribution in the study area on August 22, 1996 (at the 
end of 90-day simulation). The result corresponds well with the measured low (< 2mg/L) 
dissolved oxygen in the bottom water at the Peace River mouth. The rest of estuary does not 
show low dissolved oxygen less than 2 mg/1. 

Due to wind and tidal mixing and the shallow water depth, the Charlotte Harbor 
estuarine system generally exhibits a vertically well mixed distribution of DO. In the upper 
Charlotte Harbor, where the lowest DO in the system is usually found, stratification may 
occur and a pycnocline may form if the condition is right - high river inflow from the Peace 
river and low wind mixing (Sheng and Park, 2002). With this strong stratification, surface 
water DO from reaeration would be blocked and high sediment oxygen demand in summer 
season could create the hypoxia observed in this area. Model results showed that strong 
stratification and low DO developed during the simulation period. 

Phytoplankton dynamics are very important processes simulated in this study. It is 
closely related to the nutrient recycling through uptake during growth, and excretion/decay 
during respiration. The measured phytoplankton as algal mass per volume, was converted 
to phytoplankton carbon with an algae to carbon ratio of 100. The chlorophyll_a 
concentrations of Myakka and Peace River have maximum value of on July 23 (Julian Day 
204) and those for the other stations have no trends during the simulation periods. The 
highest daily fluctuation of chlorophyll_a is occurred in HB006 due to strong tidal 
fluctuation and high chlorophyll_a concentration from the Caloosahatchee River. 

In order to compare the nitrogen cycle simulations with the nitrogen species data 
provided by EPA, simulated soluble organic nitrogen, dissolved ammonium and ammonia 
nitrogen concentrations are combined and compared with the EPA total Kjeld nitrogen data. 



193 
The dissolved ammonium nitrogen data were also compared with measured data. For 
phosphorous species, total phosphorous concentration and soluble reactive phosphorous are 
used to compare with the EPA data. According to time series plot of each species in Figures 
6.9 to 6.20, model results appear to capture the overall trend of the EPA data. 

Figure 6-22 shows the snapshots of the near surface chlorophyll_a distribution in 
study area for August 22 (the end of 90-day simulation). Only the Peace River mouth area 
exhibits very high chlorophyll_a concentration of 3000p-g (phytoplankton carbon). Higher 
phytoplankton concentrations are generally found near the river mouths and low 
concentrations are found in the Gulf of Mexico. 

Figures 6.23 to 6.26 show the snapshots of the near surface dissolved ammonium 
nitrogen, soluble organic nitrogen, soluble reactive phosphorous, and soluble organic 
phosphorous distributions in the study area at 2 pm on August 21, 1996 (at the end of 90-day 
simulation). It is interesting to note that high dissolved ammonia concentrations but low 
dissolved organic nitrogen concentrations are found in the area between upper Charlotte 
Harbor and the Boca Grande Pass. High concentrations of particulate organic nitrogen and 
adsorbed ammonium are first resuspended by the strong currents in the Boca Grande Pass 
area, causing high dissolved ammonium concentration in the water column which are then 
transported towards the upper Charlotte Harbor area. High concentrations of phosphorous 
species are found in the upper Charlotte Harbor because Peace River drains the Hawthorn 
phosphatic formations. 



194 




160 



180 



200 



220 



| 30 
<o 20 



Q. 
O 



o 



10 




160 



180 



200 



220 



2000 



§ 1500 

— ' 1000 - 

H 500 • 



160 



i ^ w ^wv»/^ rf ' ^^m^^^ vvv - J ~^~-"''^v^.' Vvwm ^ 



180 



200 



220 



500 
g. 400 
J 300 
x - 200 
Z 100 




160 180 200 " 220 "~ 





1000 




800 






■a 


ROO 


d 








Q. 


400 


t- 






200 









500 


^^ 


400 


"5> 

-i 


300 






** 

o 


200 


□. 


100 



160 



180 



200 



220 



160 180 200 220 

Figure 6.9 Temporal water quality variations at CH002 station in 1996. 



195 



10 r 



O 

Q 2 





*rwww\^I^ 




»wv \\jf 'vwq 



160 




180 



200 



220 




160 



180 



200 



220 



2000 



f* 1500 



500 





1000 *^^-^' 






/WV\*|^ 



500 r 
.-. 400 
J 300 
£• 200 
z 100 




160 



180 



200 



220 



I 




uMMjU 


■ 




a(L « 


mMmaw 


iMMMy^M^yvvWito 




160 






180 




200 


■ 220 " 



1000 

800 

"9> 600 

i 400 

200 





500 



160 



180 



200 



220 




160 180 200 220 

Figure 6.10 Temporal water quality variations at CH004 station in 1996. 



196 




\m 



■• 



180 200 




y^mfki' V) 



220 



160 



"&> 


au 


















-i 




















"■ — 


















■ 


oj 


20 


















>. 




















.c 

CL 

o 

c 

o 


10 


JP«MA/ 


tfjtW/Wwwi 


■ 


™rv«*\\A/v' w ^ 




WJMK 


mMJ}M^ 


flL .iA-' 


.c 





















u 




160 




180 


200 




220 





2000 



f* 150 ° 
-' 1000 



1^^-^.^WaV 






P 500 



500 
^ 400 
J 300 f- 
3-' 200 
Z 100 



160 



180 



200 



"220" 



m»AA*AAWjmmy*^^ 



200 



220 




160 



180 



200 



220 



500 




160 180 200 220 

Figure 6.1 1 Temporal water quality variations at CH005 station in 1996. 



197 






160 



180 



200 



220 



re 

>> 
.c 
D. 

2 
o 

.c 
o 



JU 
















20 














■ 


10 




^!«VAAflMW!©5M^ 


1 " 


,vJk NW 3 v « Jlw Y 


vaII M/ - -^^ ^ArvvJ 


T\-V? **\- 


^-H"— vvK\j 


L^-'' r 


160 




180 


200 




220 





2000 



1* 1500 
* 1000 
£ 500^ 



ltfiMwJW^r v ^' t J*JWM^^ 



500 



160 



160 



180 



200 



220 






1UUU 












800 










3 

a 

i- 


600 
400 
200 






.aft, . miw/ 


ICuhJU ■ ftt 


w^^^j^^^A^ 




160 


180 


200 


220 



500 r 




160 180 200 220 

Figure 6.12 Temporal water quality variations at CH006 station in 1996. 



198 



E 

O 
Q 



10 
8 
6 
4 
2 




500 
g. 400 
1 300 



X 


200 


z 


100 









1000 


f 

2 
a. 


800 
600 
400 
200 









500 


1 


400 
300 


O 
a 


200 
100 








T"'* 



160 



180 



200 



220 




160 



160 180 200 220 



160 



180 



200 



220 




160 180 200 220 

Figure 6.13 Temporal water quality variations at CH007 station in 1996. 



199 



O 
a 



| 30 

m 20 ■ 

> 

| 10 

o 

5 o 




160 



180 



200 



220 



■g*^^^^^ 



160 



180 



200 



220 



2000 



U 1500 

■3" 1000 

z 

H 500 



500 
^ 400 
t 300 
x "» 200 

Z 100 



KJLS**-*** 



^^/V^^'-J 






160 



180 



200 



220 



laaaaaftttMawwvflt^^ 



160 



180 



200 



220 





1000 




800 






■& 


600 


3 








Q- 


400 


1- 






200 









500 


/— s 


400 


-1 


300 






o 


200 


a 


100 








160 



180 



200 



220 



160 180 200 220 

Figure 6.14 Temporal water quality variations at CH09B station in 1996. 



200 



10 

«-. 8 



g 6 fc^v^^^ 5 ^^^ 



O 
Q 



5 

>. 

a. 
o 



4 t 
2 




160 



180 



200 



220 




160 



180 



200 



220 



2000 r 



ffc 1500 
w 1000 



500 



BjVA^V,— j~*^V*WW*. 



NVM^WW^f 



r ^ J vJWV-v^v/'/*/v^W' 



,^.^^Vv*WVW«^'ww<M^wn^/^'^^*^^^ 



160 



180" 



200 



220 



500 F 
j-. 400 
J 300 f 
_.* 200 



°"^ ~T60 180 ' 200 220 





1000 




800 






■ft 


600 


Ti 








a. 


400 


i- 






200 









500 


^^ 


400 


a 


300 






O 


200 


Q. 


100 








160 



180 



200 



220 



160 180 200 220 

Figure 6.15 Temporal water quality variations at CH009 station in 1996. 



201 



■a 

E 
o 

D 



10 

8 

6 | 
4 

2 




160 



180 



200 



220 



5 30 



TO 
~>. 

SI 

a. 

2 
o 

.c 
O 



20 



10 



■a. 



5 

I 

z 





2000 

1500 

1000 

500 



500 
400 
300 
200 
100 



T60 - 180 200 220 



***~~*<w*^*i»*>iW<t1fiM^^ 



160 



180 



200 



220 



™^ 1 «n ' Tori ' oAa ™ ' oon " 



160 



180 



200 



220 



=5. 



1000 
800 
600 



o 

a. 



400 
200 



500 
400 
300 
200 



| l^-^WW^'^W 



160 



180 



200 



220 



l|~~""^^^ 



Figure 



160 180 200 220 

6.15 Temporal water quality variations at CH010 station in 1996. 



202 





10 




8 


E 


6 






O 


4 


Q 


2 







I 


30 


s 




"— ^ 




ro 


20 


>. 




.n 




Q. 
O 


10 






O 




Z 




o 







2000 





1500 







«. a hi 


At iLllf 


mi 1 1 AA a 






^VvvH^ 






*%P 


^jW^ 


^#% 


m!0&i^d>MmN 












■ 






160 


180 




200 




220 




160 



180 



200 



220 



i 

— 1000 

z 

H 500 



'VVVwA^^^ 



160 



180 



200 



220 



500 r 



1 

3 


400 
300 


z 


200 
100 









1000 


f 

Q. 

H 


800 
600 
400 
200 









500 


1 

2 


400 
300 


if 

O 

Q. 


200 
100 








/^M/I/VI/ma/mm/m^^ 



160 



180 



200 



220 



160 



180 



200 



220 



160 180 200 

Figure 6.17 Temporal water quality variations at HB002 station in 1996 



220 



203 



10 

Q 

o 4 



160 



| 30 

n 20 

>. 

f 10 

5 o 

2000 



fi 1500 
& 1000 

z 

K 500 [■ 



160 



kW.'.^*™^ 





500 
-. 400 
1 300 
j? 200 ■ 
Z 100 • 




160 



160 



1000 
^ 800 
"^ 600 
a 400 

H 200 t 



500 
^ 400 
§ 300 
Q ' 200 

a 100 





160 



180 



200 



220 




180 



200 



220 



T80~ 200 



WW 



220 



180 



200 



220 



180 



200 



220 



'^ W-Vw 'B s, ~*^^ 



160 180 200 220 

Figure 6.18 Temporal water quality variations at HB006 station in 1996. 



204 



E 

O 
Q 



10 
8 
6 
4 
2 




f **ZXf^ 



^$^*^ 



K/\M 



ns/vj 



160 



180 



"200" 



220 




160 ~~^80~~ ~200 220" 

Figure 6.19 Temporal water quality variations at HB007 station in 1996. 



205 



"a 
J, 

o 




160 



180 



200 



220 



■a 



Q. 
O 

I— 

O 

r 
o 



30 

20 

10 



2000 



=• 1500 

-" 1000 



500 



500 
400 
300 
200 
100 


1000 
800 
600 
400 
200 ■ 


500 
400 
300 
200 
100 




160 



180 



200 



220 



■a. 



a. 



160 



■a 

3 

t 

o 
a. 



160 180 200 ~ 220 

Figure 6.20 Temporal water quality variations at CH013 station in 1996. 



206 



Near-Surface Dissolved Oxygen (mg/L) 



Near-Bottom Dissolved Oxygen (mg/L) 







August 21, 1996 



5. \ 

Figure 6.21 Simulated dissolved oxygen concentration in Charlotte Harbor estuarine 
system on August 21, 1996. 



207 



Near-Surface Chlorophyll a (ug/L) 



Near-Bottom Chlorophyll a (ug/L) 




Figure 6.22 Simulated Chlorophyll_a concentration in Charlotte Harbor estuarine system 
on August 21, 1996. 



208 



Near-Surface NH4 (ug/L) 



Near-Bottom NH4 (ug/L) 




Figure 6.23 Simulated dissolved ammonium nitrogen concentration in Charlotte Harbor 
estuarine system on August 21, 1996. 



209 



Near-Surface SON (ng/L) 



Near-Bottom SON (ug/L) 




Figure 6.24 Simulated soluble organic nitrogen concentration in Charlotte Harbor 
estuarine system on August 21, 1996. 



210 



Near-Surface SRP (ng/L) 



Near-Bottom SRP (|ig/L) 



500 / 
400 \ 




\ 



X \ 







J 



500 
400 
300 
200 
100 



V 






August 21, 1996 










Figure 6.25 Simulated soluble reactive phosphorous concentration in Charlotte 
Harbor estuarine system on August 21, 1996. 



211 



Near-Surface SOP (ng/L) 



Near-Bottom SOP (ug/L) 







60 
50 
40 
30 
20 
10 



August 21, 1996 



v *^ 








Figure 6.26 Simulated soluble organic phosphorous concentration in Charlotte 
Harbor estuarine system on August 21, 1996. 



212 
6.4 Simulations of Water Quality in 2000 

Model simulation of circulation, sediment transport and water quality dynamics in 
Charlotte Harbor estuarine system during the summer 1996 was conducted as calibration 
process tunning of model parameters and inputs followed systematic calibration procedure. 
Based on the results presented, the water quality model was considered calibrated and 
validated, and additional one year simulation was performed to validate the model. This 
validated model was used to assess the effects of Sanibel Causeway and the navigation 
channel in San Carlos Bay as well as river load reductions. 
6.4.1 Validation 

For validation runs, hydrodynamic and water quality model coefficients were held 
fixed at the calibration values, and results of a one-year 2000 simulation were compared with 
field data collected in January to December 2000. The data include phytoplankton, nitrogen, 
phosphorous, dissolved oxygen at several locations in the Charlotte Harbor estuarine system. 

Figure 6.27 contains calibration period scattering plots which include the root mean 
square (RMS) error and correlation coefficient (R 2 ). Values shown in parenthesis are 
normalized RMS errors with maximum measured data. Without adjusting any water quality 
model coefficients, the normalized RMS errors are less than 45%. Dissolved oxygen has the 
best agreement with measured data as shown by the RMS error. Correlation coefficients for 
all species vary form 0. 124 to 0.80. These plots allow us to assess weaknesses of the model, 
and to suggest areas needing further improvement. 



213 



A) Chrolophyll a 

25 



to 



RMS= 7.51 (33.56 %) 
R2 =0.519 

20h * 

■a 

$> 15- 
co 

E - ' I- / 

- iot. l-jyzs** • • • 

• . v v n >s • • • 







"0 5 10 15 20 

measured 

C) Total Kjeldahl Nitrogen (TKN) 

1500 



■D1000 

3 

E 

« 500 



B) Dissolved Oxygen (DO) 

10 



CD 

£ 

CO 



RMS= 1.40 (17.63%) 
R2 = 0.308 



25 




0, 



RMS=0.228(21.98 


%) / 


R2 


= 0.462 


mf 




• 




• • 

• 
• 


• I • • 4 S m • 

• *y * 


Vm 

9 

•; 

• 
• 

• 


" 


pA- 






r * 




/ 


,,!,,,, 


, , , , 



measured 

D) Dissolved Ammonium Nitrogen (NH 4 ) 

250 



RMS= 0.040 (44.48 %) 
R2 = 0.362 




500 1000 150( 100 200 

measured measured 

Figure 6.27 The scatter plots for water quality constituents in 2000. 



214 



E) Total Phosphorous 

450 



400 
350 



RMS= 0.0783 (31.26%) 
R2 = 0.242 




■o 
E 



RMS= 3.320 (32.13%) 
R2 = 0.806 






100 200 300 400 

measured 

G) Total Organic Carbon (CBOD) 

20 




5 10 15 

measured 

Figure 6.27 continued 



20 



F) Soluble Reactive Phosphorous (P0 4 ) 

250 



RMS= 0.0294 (42.87%) 
R2 = 0.280 



200- 




50 100 150 200 250 

measured 

H) Total Suspended Sediment (TSS) 

150 



RMS= 25.67 (36.45 %) 
R2 =0.124 







50 100 

measured 



150 



215 
6.4.2 Results of 2000 Water Quality Simulation 

The simulated water quality species at each measured station in 2000 as shown with 
measured data in Figure 6.28 to Figure 6.39. During the simulation period, the water quality 
data show the seasonal variation, and these seasonal variations are produced quite well by 
the water quality model. In the water quality model, dissolved oxygen is a function of 
photosynthesis and respiration by phytoplankton organisms, sediment oxygen demand, 
reaeration, nitrification and denitrification, decomposition of organic matter, tide and wind 
mixing, and river loading. As shown in Figures 6.28 to 6.39, simulated dissolved oxygen 
concentrations show spatial and temporal variations in reasonable agreement with measured 
data. The average monthly near-surface concentrations declined from 8.5 to 6.7 mg/L from 
January to July and then began to rise in upper Charlotte Harbor. Near-bottom average 
monthly concentrations in this area were highest in February, declined slowly through May, 
and then declined more rapidly until September. The hypoxia conditions during summer are 
attributed to strong stratification, which cause restricted re-aeration, and SOD. After the 
breakup of the stratification, the DO concentration increased from October to December. 

Algae photosynthesis and re-aeration maintain surface dissolved oxygen level, while 
vertical mixing controls the transfer of dissolved oxygen to bottom water. Sediment oxygen 
demand is a function of temperature which has strong seasonal variation. The high water 
temperature in summer season will increase SOD in bottom sediment. If there is no 
stratification, surface water dissolved oxygen will become mixed with bottom water 
dissolved oxygen quickly. The Charlotte Harbor estuarine system generally exhibits a 
vertically well mixed distribution of DO due to wind and tidal mixing and the shallow water 
depth. The dissolved oxygen in most part of the estuary does not show any strong 



216 
stratification except in Peace and Caloosahatchee Rivers. In the upper Charlotte Harbor, 
which usually has the lowest level of DO, some stratification may occur due to high 
consumption by SOD near the bottom and super-saturation near the surface with strong river 
discharge from Peace River. To quantify the causes of the bottom water hypoxia in this area, 
dissolved oxygen concentration at CH006 was compared with river discharge at Peace River, 
salinity, temperature and, reaeration and SOD fluxes in Figure 6.40. Although, the SOD flux 
is higher than the reaeration flux in summer (Julian Day 120 to 280) due to high 
temperature, the DO stratification did not occur during this period. The DO stratification 
period (Julian Day 220-280) matches with the salinity stratification caused by strong river 
discharge from Peace River. 

To compare salinity stratification and vertical DO distribution, Simulated 
longitudinal-vertical salinity and dissolved oxygen concentration along the Peace River at 
1 pm on June 18 (Julian Day 170) and October 6 (Julian Day 280), 2000 were plotted in 
Figures 6.41 and 6.42, respectively. The salinity at these two time period represent the effect 
of salinity stratification on vertical DO distribution since the SOD and re-aeration fluxes are 
similar at these periods. The results show the strong relationship between salinity and DO 
stratifications. Therefore, the hypoxia in the upper Charlotte Harbor is primarily caused by 
the combination effects of SOD and stratification with strong river discharge and high 
temperature. 

In the Caloosahatchee River upstream near the S79 (CES02), there is very strong 
daily fluctuation in the surface dissolved oxygen concentration due to tidal fluctuation and 
high dissolved oxygen concentration from the Caloosahatchee River. Chlorophyll_a 
concentrations ranged from 1 to 100 mg/L and averaged 8.5 mg/L. Both productivity and 



217 
biomass were greater during summer near the mouth of tidal rivers which has middle range 

salinity of 6 to 12 ppt. The chlorophyll_a concentration of Caloosahatchee River has a 
maximum value of 98 mg/L on July 12 (Julian Day 193), 2000 and then quickly drop to 5 
mg/L because of low chlorophyll_a concentration of river loading and fast flushing due to 
strong river discharge from Caloosahatchee River. Simulated nitrogen and phosphorous 
concentrations appear to capture the overall trend of the measured data collected by 
SFWMD and SWFWMD. 

Figures 6.43 and 6.44 show the snapshots of the near surface chlorophyll_a and the 
near bottom dissolved oxygen concentration distribution in the study area on February 9, 
May 9, August 7, and November 5, to represent seasonal characteristics. In February, 
phytoplankton was low in the entire estuary and DO was a generally high. In May, 
phytoplankton significantly increased in the Caloosahatchee River due to strong river 
discharge with high nutrient concentrations. Hence, DO was reduced substantially in all 
portions but was higher in the north than in the south, except in Pine Island Sound. In 
August, there was significant increase in phytoplankton and decrease in dissolved oxygen 
concentration in the bottom water. This low dissolved oxygen in bottom water reached the 
north portion of upper Charlotte Harbor. The new water quality model, which contains 
improved dissolved oxygen processes at the air/sea interface and water-sediment interface, 
successfully reproduced the bottom water hypoxia in both temporal and spatial plots. 



218 



1 

E 

o 

Q 



I 30 

3. 




>» 

a 
o 

S 

o 






20 

10 



1500 

1000 
500 



^— i -*-«-^^^ 



^»^ 



•6) 



200 
150 
100 - 
50 ■ 

400 
300 
200 

100 



200 



=; 150 

■3 100 
o 

0. 50 



a. 



**** 



100 



"500" 



300 



-^^^>-^^_ 



100 



^00 



300 




[- 


100 






200 




300 


■ 


1 — .1^ 


"*"**«» 


""■■ DlJlljlh 


1 -y. 
in iiiiiiiiiiiiiJiiiJ^rfP^ 


1 OKU 


«*JLgI 




"^ H^^»» . a. .^._ ^B 



100 



200 



300 



^^^Jj>,y*L-, 



100 200 300 

Figure 6.28 Temporal water quality variations at CH002 station in 2000. 



219 



§ 

o 

Q 



| 3 ° 
3. 



TO 

>. 
.C 
Q. 

2 

o 

O 




100 



200 






20 

10 



1500 

1000 



ksr^*^^^^^^ 



* 500 


200 



i 

z 



150 
100 

50 



o PVwJl 



5 



■B) 

if 

o 



100 



200 



100 



200 



300 




300 




300 



WHM'Jftl 



100 



200 



iiiiwmwimwi 



A 






300 




Figure 



100 200 300 

6.29 Temporal water quality variations at CH004 station in 2000. 



220 



O 
Q 




■a 

a, 



Q. 
O 

i- 

o 

i 

o 



1 

3 



30 

20 

10 



1500 

1000 

500 



200 
150 

100 

I* 

Z 50 



400 

O 300 

3 200 
Q. 

*- 100 



i*-ff^l — nr^J^^^*^^ 



100 



200 



300 



/-T 



»**"*" 




100 



200 



"300" 



"5> 
2 



wJLfciriffMjLt^r riv 



100 




200 



300 



•-if"**^^ 



100 



200 



300 



5 

o 

a 




Figure 



100 200 300 

6.30 Temporal water quality variations at CH005 station in 2000. 



221 



■5) 
E 

O 
a 



% 30 




100 



200 



300 



Q. 

2 
o 

o 



20 



10 




Piftw** », 



100 



200 



300 



1500 



"S> 1000 

1 500 

I- 



200 



yw' 



,_^^|^ 



«U#W/»fc 



m i «njl» , w ^ i ^^ ' ^i # ■*![ in i 



100 



200 



300 




100 



200 



300 



400 r 



~ 300 [ 

% 

3 200 

Q. 

*- 100 



^t-^-»*-^^jl^a*^^ — . 



100 



200 



300 



200 



=; 150 
o) 

« 100 

o 

Q. 50 



100 




200 



300 



Figure 6.31 Temporal water quality variations at CH006 station in 2000. 



222 





10 


£ 


8 
6 


-^ 




o 


4 


Q 


2 







1 


30 


3 




-w' 




TO 


20 


>< 




J= 




Q. 
O 


10 


i- 




O 








£ 




o 







1500 




"S> 1000 
* 500 

I- 



■5) 



■a. 



"ft 

o 

Q. 



150 
100 

50 


400 
300 
200 
100 



200 r 

150 

100 

50 




100 



200 



300 



100 



200 



300 



~_*f-'-" vJ 




100 



200 



300 



k^Nrt* 



100 



200 



300 




JUttAm " i - 



ll^Mt^^k 



100 200 30( 

Figure 6.32 Temporal water quality variations at CH007 station in 2000. 



223 




100 



200 



300 



% 


30 


3 




w 




TO 


20 


>< 




.C 




Q. 
O 


10 






O 




.C 




o 







1500 


^ 




"& 


1000 


d 




, — •* 




7 




* 


600 


h- 





'*>i l *i«mmm*M^ m >i'm^f0>^ 



«^*^^^ 



100 



200 



300 



> „^~~»v-'" 






100 



200 



300 



200 



f 

2 

I 
2 


150 
100 

50 









400 


1 

zt 


300 
200 


i- 


100 



)J ^^ k ^mm»^^ 




100 



200 



300 



100 



200 



300 



200 



^ 150 



3 

O 

a 



100 

50 








fr ^""""If**^*"^ 



Figure 



100 200 300 

6.33 Temporal water quality variations at CH008 station in 2000. 



224 



E 

O 
Q 



10 
8 
6 
4 
2 







100 



200 



300 



i 30 

• 20 

§" 10 



1500 



"3) 1000 
* 500 



^jj**< lim ^ m »< , ' i -»iii«*^ 



riMHP 



H^.^^**--^^^ 



100 



200 



300 




200 



=; 150 
* 100 



_■ ■ "*** J ^' II> *J" > I ' |L M iii Irt.nf-J-l 



100 



200 



"300" 



I 

z 



50 


400 



TOO ' ' 200 ' ' ' ' 500 ~ 



ss- 300 

3 200 

Q. 

*- 100 ■ 


200 



g 150 

3 100 

o 

Q. 50 



100 



200 



300 




>^JL*^ I 



100 200 300 

Figure 6.34 Temporal water quality variations at CH009 station in 2000. 



225 



f 

o 

Q 



"ft 

(0 

>. 

Q. 

2 



10 
8 
6 
4 
2 








100 



200 



300 



30 



20 



W ' 200 500 ^~ 






1500 

1000 

500 



200 
=• 150 

•S 100 

X 

Z 50 



400 
300 
200 
100 



■^J^^^^^ 



100 



200 



^00" 



1 00 200 300 






vtrfiyWMllW,^^^, 



100 



200 



300 



1 

o 



200 
150 
100 [ 



50 



^y*%%»^^ ■ 



Figure 



100 200 300 

6.35 Temporal water quality variations at CH010 station in 2000. 



226 



o 

Q 



3 

(0 

>< 

Q. 

2 
o 

.c 
o 




100 



200 



3. 






2000 

1500 

1000 

500 



400 
300 
200 
100 





600 



100 



200 



;to; * S; * s *^^ 



100 



200 



=; 150 
~ 100 

o 

a 



50 ■ _ 

u 100 



300 





300 




300 





200 300 

Figure 6.36 Temporal water quality variations at CES02 station in 2000. 



227 




400 r 

~. 300 

■5. 

3 200 

Q. 

*- 100 



200 



^ 150 

« 100 

o 

Q. 50 



100 



200 



300 





4UU 






a 


i | 




^ 


300 


- 




/P l/V ,,| 


? *V 




eft 

5 


200 




A 


^ l^n/1 


■ " 


I 




*v^ ■ 


A IM 


Jj jtitMiiiM 


liiiiAlii ujAl 




z 


100 




i'-vHi 1 'MiWmma 


mnm 11 


iiliiiJlWW *v^_ 


^^f^y- 









pff^fmPHH 


mf if 


iilllMi" 


■ V^* 






100 




200 


300 



■ 

■ 




■ 








""^^^ J^^^Jj^^i''' | 


■ 






\~s 




100 




200 






300 



jmBS^jmL 



A 

/^ W-v I 



100 "" 200 300 

Figure 6.37 Temporal water quality variations at CES03 station in 2000. 



228 



| 

E 

O 
Q 



§ 10 ° 

3 80 F 

a 

= 60 




100 



200 



300 



•q. 40 

o 

o 20 

5 o 



" 100 200 %6o 



"ft 

z 



2000 

1500 

1000 

500 



400 
=; 300 



^*«<€Mft 



MPp 



0H 



l ^P*»*^» w > B ^^ 



100 



200 



300" 



200 



■a. 

2 





400 
300 
200 
100 



100 -^^^^^ 

TOO ' 2$0 300 



"-*****, 



W Wj W i f& HR K lmmim 



100 



illlili W i l liiMHjI W IlWM 



200 



ggg g ^g g gggg g i ^ £»f*>m 



300 



i 

3 

o 

Q. 



Figure 



200 
150 
100 i 

50 




J-« "^ 



100 200 300 

6.38 Temporal water quality variations at CES08 station in 2000. 



229 



10 



s=> 8 
% 

E 

O 

Q 



E 6 " ** to, *J">>**^ 



100 



200 



300 



% 30 



3 

>. 

sz 
a. 
o 



20 



10 







.^""^^frlt)^ 



1500 



^> 1000 



200 



250 



300 



lI v^^^"^'^^^ifc»ly l ^V^*** w *^^^l J .n ni'M i i**mj 



100 



200 



300 



200 r 






150 
100 

50 



hy^r^V-^^mt^K^^s^^ *V*^ 



fti**mH**\nrt> , "«'*r'iH/(f 



100 



200 



300 



3. 



200 r 
150 



100 

50 



i 

o 

a 





200 
150 
100 



H^4w»"*^^ 






100 



200 



300 



50 



' rr 



Figure 



100 200 300 

6.39 Temporal water quality variations at CHOI 3 station in 2000. 



230 



"e 

(D 



O 

.2 
Q 

> 




35 



g 30 

Q. 

£ 25 



W 20 



15 









Surface 

Bottom 




*¥*V 



50 



100 



150 



200 



250 



300 



O 

E 

re 
k_ 
0) 

a 

E 

0) 



•S 15 



re 

E 

S 

c 
o 

E 
a> 
re 
v 
O. 



1> 

s 

c 



T3 
1 

o 
in 




re 

"e 

i 

Q 
O 
C/3 



1 50 200 

Julian Day 



300 



Figure 6 
possible 
fluxes at 



.40 The comparison between simulated dissolved oxygen concentration and the 
causes of hypoxia: river discharge, salinity, temperature, and re-aeration and SOD 
CH006 water quality measurement station. 



231 



6/18/2000 13:00 (Julian Day 170) 
CH006 CH005 CH004 




-22500 -20000 -17500 -15000 -12500 

Distance from Peace River upstream (m) 



10000 



>5000 



CH006 



CH005 



CH004 




-22500 -20000 -17500 -15000 -12500 

Distance from Peace River upstream (m) 



-10000 



Figure 6.41 Simulated longitudinal-vertical salinity and dissolved oxygen concentration 
along the Peace River at 1 pm on June 18 (Julian Day 170), 2000. 



232 



10/6/2000 13:00 (Julian Day 280) 

CH005 CH004 




-22500 -20000 -17500 -15000 -12500 

Distance from Peace River upstream (m) 



10000 




25000 



-22500 -20000 -17500 -15000 -12500 

Distance from Peace River upstream (m) 



10000 



Figure 6.42 Simulated longitudinal-vertical salinity and dissolved oxygen concentration 
along the Peace River at 1 pm on October 6 (Julian Day 280), 2000. 



233 




Chlorophyll a (p.g/L) 



Chlorophyll a (ug/L) 






ft • ^ 







Chlorophyll a (ug/L) 



Chlorophyll a (ug/L) 




Figure 6.43 Simulated near-surface chlorophyll_a concentration in Charlotte Harbor 
estuarine system on February 9, May 9, August 7, and November 5, 2000. 



234 



Near-Bottom Dissolved Oxygen (mg/L) 



Near-Bottom Dissolved Oxygen (mg/L) 



V \ . > 



$e 




Near-Bottom Dissolved Oxygen (mg/L) 




Near-Bottom Dissolved Oxygen (mg/L) 



August 7, 2000 




I 



November 5, 2000 



&*$■%! 



J. tAa? 




Bvl 





J 



^ 






Figure 6.44 Simulated near4Dottom dissolved oxygen concentration in Charlotte Harbor 
estuarine system on February 9, May 9, August 7, and November 5, 2000. 







v 



235 
6.4.3 Application of 2000 Water Quality Simulations 

Hydrologic alteration 

The hydrodynamic and water quality models of the Charlotte Harbor estuarine system 
have been successfully developed and validated. The models can be used to address the 
effect of hydrologic alteration on the water quality as shown in Chapter 5. Using the 
validated water quality model, we performed several model simulations with the causeway 
islands removed and with the IntraCoastal Waterway removed during April 9 to June 10, 
2000. The results are compared with those under existing conditions with both of them in 
place. The scenarios and stations for comparison are the same as those for comparison of 
flow and salinity in Chapter 5. 

The chlorophyll_a concentrations for both cases were compared with those for the 
baseline simulation in Figure 6.45. The results at all three stations show that chlorophyll_a 
concentration is not noticeably affected by the absence of Intracoastal Waterway (NICW 
case) or the causeway (NSC case). Just like the chlorophyll_a concentration, the other water 
quality species do not show much effect by these hydrologic alterations. 

To quantify the effect of these hydrologic alterations on the spatial distribution of 
water quality species, the snapshots of chlorophyll_a and dissolved ammonium nitrogen 
concentrations for these two cases were compared with those for the baseline simulation in 
Figures 6.46 and 6.47. There is not noticeable impact on the water quality species 
distribution in the San Carlos Bay area, consistent with the negligible effect of hydrologic 
alteration on flow and salinity in Chapter 5. 

Table 6.6 shows the temporal average water quality species concentrations at the 
eleven comparison stations for the baseline run. Results of the hydrologic alteration cases 



236 
were compared to the baseline results and RMS differences calculated, then normalized with 
these averaged water quality species concentration in Table 6.6. Normalized RMS 
differences of water quality species concentrations for the no causeway case from April to 
June, 2000 are shown in Table 6.7. Table 6.8 shows the normalized RMS differences for the 
no ICW case, while Table 6.9 shows the normalized RMS differences when both the 
causeway islands and the ICW are removed. The RMS differences are less than 2% at all 
selected stations for hydrologic alteration. Therefore, it can be concluded that neither the 
causeway islands nor the Intracoastal Waterway had noticeable effect on the water quality 
in the San Carlos Bay and Pine Island Sound area. 



237 





ST05 


BASELINE 


30 




Causeway removed (NSC) 


i 25 




Intra Coastal Waterway removed (NICW) 


a. 






"-" 20 






CO 






^ 15 






£ 






a 




,aAWW/Vu, u 


g 10 




^ ^^w 


o 


,. AAAA/W 


1 ^^,,jy/W\MM/\IVVw^ 


■= 5 


-,^a^A*W^ 


v V\jW™ 'V'* " 


O 


1 


II 1 





12C 


) 140. _ 160 180 



"9> 

CO 



2 
o 



30 



ST08 



25 - 



"a. 

a. 

•— ' 20 

CO 

% 15 
Q. 

2 10 
o 

■= 5 
O 






ST10 



120 



120 



140. _ 160 

Julian Day 



180 




160 



180 



14Q. _ 
Julian Day 

Figure 6.45 Comparisons of simulated surface chlorophyll_a concentration for three cases 

at three selected stations: ST05 (Pine Island Sound), ST08 (San Carlos Bay), and ST 10 

(Caloosahatchee River mouth) 



238 




Chlorophyll a 

20 

15 
10 

5 

• S xk 

Figure 6.46 Comparisons of simulated surface chlorophyll_a concentration fields in San 
Carlos Bay after 90 days of simulation for three cases. 






239 



' m 




Figure 6.47 Comparisons of simulated surface dissolved ammonium nitrogen (NH4) 
concentration fields in San Carlos Bay after 90 days simulation for three cases. 






240 



Table 6.6 The temporally averaged water quality species concentrations for the baseline 
2000 simulation. 



Station number 


ChlA 


DO 


NH4 


NOX 


TKN 


P04 


PHOST 


TOC 




(mg/L) 


(mg/L) 


(u-g/L) 


(Hg/L) 


(Hg/L) 


(Hg/L) 


(lig/L) ( 


mg/L) 


1 


6.04 


6.39 


99.53 


31.69 


728.98 


9.11 


103.50 


1.61 


2 


6.51 


6.39 


103.06 


29.60 


765.36 


9.62 


105.38 


1.68 


3 


7.28 


6.46 


101.99 


27.46 


763.13 


9.63 


110.97 


1.69 


4 


9.14 


6.50 


103.79 


24.03 


784.51 


9.69 


141.30 


2.07 


5 


7.37 


6.18 


115.69 


26.32 


864.87 


10.60 


90.78 


2.35 


6 


7.59 


6.33 


110.88 


25.46 


839.52 


10.39 


109.21 


1.97 


7 


7.24 


6.40 


107.63 


26.88 


809.50 


10.10 


108.62 


1.79 


8 


8.57 


6.48 


103.77 


22.69 


785.02 


10.35 


116.44 


1.92 


9 


11.10 


6.50 


103.05 


20.03 


787.49 


9.61 


145.89 


2.72 


10 


13.71 


6.44 


101.44 


15.66 


784.27 


9.35 


148.24 


3.85 


11 


10.14 


6.38 


100.42 


16.16 


771.49 


11.44 


117.82 


2.38 


AVG 


8.61 


6.40 


104.66 


24.18 


789.47 


9.99 


118.01 


2.18 


Table 6.7 Normalized RMS differences of water quality species 


concentrations at 1 1 


stations during April to June 2000, i 


showing the effect of no causeway islands. 




Station number 


ChlA 


DO 


NH4 


NOX 


TKN 


P04 


PHOST 


TOC 




(%) 


(%) 


(%) 


(%) 


(%) 


(%) 


(%) 


(%) 


1 


0.27 


0.25 


0.83 


1.33 


1.36 


0.26 


1.58 


0.24 


2 


0.37 


0.28 


1.17 


1.16 


1.91 


0.29 


1.29 


0.27 


3 


0.65 


0.24 


0.83 


1.56 


1.36 


0.39 


1.9 


0.34 


4 


0.85 


0.24 


1.04 


0.58 


1.77 


0.21 


2.25 


0.31 


5 


0.58 


0.68 


1.35 


1.14 


2.55 


0.45 


2.2 


0.55 


6 


0.47 


0.43 


1.15 


1.19 


2.02 


0.33 


1.95 


0.3 


7 


0.44 


0.32 


1.15 


1.17 


1.95 


0.25 


1.62 


0.23 


8 


0.5 


0.27 


0.94 


1.59 


1.46 


0.4 


2.22 


0.41 


9 


0.93 


0.2 


0.9 


0.41 


1.54 


0.2 


1.93 


0.39 


10 


0.76 


0.18 


0.74 


0.29 


1.23 


0.15 


1.44 


0.36 


11 


0.33 


0.24 


1.09 


1.22 


1.44 


0.28 


1.39 


0.25 


AVG 


0.56 


0.3 


1.02 


1.06 


1.69 


0.29 


1.8 


0.33 



241 

Table 6.8 Normalized RMS differences of water quality species concentrations at 1 1 
stations during April to June 2000, showing the effect of no ICW. 

Station number ChlA DO NH4 NOX TKN P04 PHOST TOC 

(%) (%) (%) (%) (%) (%) (%) (%) 



1 


0.28 


0.05 


0.12 


0.42 


0.24 


0.11 


0.46 


0.13 


2 


0.33 


0.08 


0.23 


0.65 


0.38 


0.11 


0.33 


0.19 


3 


0.34 


0.08 


0.18 


0.41 


0.33 


0.22 


0.63 


0.19 


4 


0.97 


0.12 


0.37 


0.75 


0.41 


0.24 


1.89 


0.37 


5 


0.37 


0.30 


0.44 


1.13 


0.75 


0.16 


0.53 


0.49 


6 


0.52 


0.19 


0.39 


1.16 


0.59 


0.16 


0.53 


0.35 


7 


0.46 


0.12 


0.34 


0.95 


0.51 


0.14 


0.56 


0.26 


S 


0.50 


0.17 


0.24 


0.38 


0.47 


0.37 


0.84 


0.30 


9 


1.39 


0.20 


0.62 


0.86 


0.51 


0.29 


2.67 


0.68 


10 


1.51 


0.41 


0.74 


0.82 


0.68 


0.51 


5.31 


2.11 


11 


0.48 


0.13 


0.32 


0.41 


0.56 


0.26 


1.03 


0.41 



AVG 0.65 0.17 0.36 0.72 0.49 0.23 1.34 0.50 

Table 6.9 Normalized RMS differences of water quality species concentrations at 1 1 
stations during April to June 2000, showing the effect of no causeway island and ICW 

Station number ChlA DO NH4 NOX TKN P04 PHOST TOC 

(%) (%) (%) (%) (%) (%) (%) (%) 



1 


0.40 


0.27 


0.95 


1.69 


1.36 


0.26 


1.30 


0.30 


2 


0.39 


0.34 


1.39 


1.77 


2.11 


0.30 


1.33 


0.39 


3 


0.72 


0.28 


0.96 


1.89 


1.25 


0.43 


1.46 


0.40 


4 


0.57 


0.24 


1.34 


1.24 


1.47 


0.30 


1.69 


0.39 


5 


0.40 


0.93 


1.77 


2.18 


3.23 


0.54 


2.58 


0.97 


6 


0.42 


0.59 


1.51 


2.22 


2.43 


0.39 


2.19 


0.59 


7 


0.40 


0.42 


1.44 


2.01 


2.23 


0.26 


1.75 


0.42 


8 


0.69 


0.38 


1.03 


1.96 


1.12 


0.48 


2.06 


0.52 


9 


1.07 


0.23 


1.42 


1.15 


1.17 


0.35 


1.80 


0.74 


10 


1.40 


0.35 


1.36 


0.89 


0.75 


0.50 


4.28 


2.19 


11 


0.47 


0.30 


1.26 


1.58 


0.86 


0.37 


0.75 


0.32 



AVG 0.63 0.39 1.31 1.69 1.63 0.38 1.93 0.66 



242 
River load reduction 

To provide a tool for studying management options and corresponding responses of 
the Charlotte Harbor estuarine system, which is one of the primary objectives of this study, 
the validated integrated model was used to evaluate the effectiveness of load reductions for 
improving estuarine water quality. To achieve this goal, it is necessary to use a common set 
of initial conditions for the water column and the sediments so that any differences observed 
between the results of different load reduction simulations would be attributable to the 
differences in load reduction. To analyze the potential impact of reduced nutrient loadings 
to the system, model simulations were carried out using 100% load reductions of nitrogen 
species concentrations, 100% load reductions of phosphorous species concentrations, and 
100% load reduction of nitrogen and phosphorous species concentrations at Peace River and 
Caloosahatchee River. 

One major assumption in the load reduction runs is that the SOD values are assumed 
to be the same as the baseline run. This is because the fact that SOD values for the baseline 
run were provided from field experiments and the model's inability to directly related the 
SOD values inside estuary to nutrient loading from the rivers. Hence the results presented 
in the following should be interpreted with caution, since large errors may be associated with 
the above assumption. In particular, since the SOD values are unchanged, load reduction is 
not expected to produce any improvement in hypoxia, hence the DO results are practically 
unchanged in the model results. In reality, load reduction is expected to lead to increased 
DO. 

For the Peace River load reduction simulations, CH004 and CH006 stations are 
selected to compare the water quality species before and after load reduction. Figures 6.48 



243 
and 6.49 show the water quality species at CH004 and CH006 before and after 100% 

nitrogen load reduction. Due to reduction of nitrogen loading, the chlorophyll_a 
concentration decreased at both stations because the reduction of dissolved ammonium 
nitrogen as the limiting factor for phytoplankton. The amount of the decrease was 12 |ig/L 
at the CH004 station and 4 [igfL at the CH006 station. The dissolved oxygen concentration 
is little decreased at the CH004 station because the amount of dissolved oxygen from 
photosynthesis is reduced due to decreasing of phytoplankton biomass, while those from re- 
aeration and SOD fluxes remained the same as the baseline simulation. 

Figures 6.50 and 6.51 show the water quality species at CH004 and CH006 before 
and after 100% phosphorous load reduction from the Peace River. The chlorophyll_a 
concentrations decreased at both stations. The phytoplankton food limiting factor is changed 
to soluble reactive phosphorous after SRP concentration reached to less than phosphorous 
half saturation rate. The amounts of decrease 8 [ig/L at CH004 station and 4 |ig/L at CH006 
station were smaller than those for 100% nitrogen load reductions. Dissolved ammonium 
nitrogen concentration was increased because phytoplankton consumed a smaller amount of 
NH4 concentration, as phytoplankton food limit was decreased by SRP concentration. 

For the Caloosahatchee River load reduction, CES02 and CES08 stations are selected 
to compare the water quality species before and after load reduction. Figures 6.52 to 6.53 
show the water quality species at CES02 and CES08 before and after 100% nitrogen load 
reduction, while Figures 6.54 to 6.55 show the corresponding results for 100% phosphorous 
load reduction. With reduction of nitrogen or phosphorous loading, the chlorophyll_a 
concentration decreased at CES02 while there is no significant difference at CES08. The 
Chlorophyll_a concentration at CES02 reached 94 |ig/L due to large amount of chlorophyll_a 



244 
from the river loading. Most of the water quality species are strongly affected by the loading 
from Caloosahatchee River, as shown in the results at these two stations. The impacts 
resulting from load reductions were confined in Caloosahatchee River and became 
insignificant once outside the Caloosahatchee River. 

As mentioned earlier, dissolved oxygen concentration was relatively unaffected by 
the loading reduction because the SOD kinetic process used in water quality model is an 
empirical formula which is a function of temperature, dissolved oxygen, and sediment type. 
Sediment oxygen demand (SOD) depends on the deposition and decomposition of organic 
matter on the seabed, and the exchange of nutrients and oxygen across the sediment-water 
interface. To provide detailed trends in hypoxia in response to organic loads, it is necessary 
to apply sediment flux model (DiToro and Fitzpatrick, 1993) with observed data for CH 4 , 
H 2 S, and organic matter in sediment column and river boundary as described in Figure 4.4. 
Although applying of sediment flux model provides rational predictions of sediment response 
to environmental alterations, it requires additional information as compared to the use of 
empirical SOD model. 

Without any available data, the sediment flux model was tested for organic matter 
river load reduction in Peace River using DiToro's method described in Appendix F. Figure 
6.56 shows the comparison of chlorophyll_a and dissolved oxygen concentrations at CH006 
station between baseline and 100% organic matter loading reduction simulations. The result 
does not show much difference in dissolved oxygen concentrations between the baseline and 
100% organic matter load reduction simulations, because SOD flux from methane and 
nitrogen calculated by the sediment flux model is not enough to create hypoxia in CH006. 
The measured total organic carbon concentration at CH006 is much lower than that at the 



245 
Peace River boundary (CH029). Therefore, the organic carbon concentration from Peace 
River does not reach to the hypoxia area (CH006). In the summer season, the dissolved 
oxygen concentration of the river load reduction scenario is even lower than that for the 
baseline simulation due to reduced chlorophyll_a concentration. Chlorophyll_a 
concentration from river loading in summer season is quickly reduced because of the nutrient 
load reduction from the river. The photosynthesis is decreased due to chlorophyll_a 
decrease, but CBOD is increased in the water column due to mortality of phytoplankton and 
settling to the sediment layer. This increase in CBOD will increase the SOD flux at the 
sediment-water interface. 

The measured data show that total organic carbon concentration at CH006 was very 
low during the simulation period. Hence, the carbonaceous oxygen demand may not be a 
major factor of SOD flux in upper Charlotte Harbor. Sulfide flux could be a very important 
component of SOD in anoxic estuarine water (Chapra, 1997) such as the upper Charlotte 
Harbor. Hence, sulfide and iron fluxes should be included in the sediment flux model to 
better represent sediment oxygen demand for river load reduction simulations. In addition, 
more field data and modeling effort to determine sediment fluxes should be focused on the 
specific conditions of upper Charlotte Harbor, especially specifying the parameters for the 
sediment layer and the sediment- water interface. 

To test the effect of reducing organic matter in river loading with current water 
quality model, the analysis of the potential impact of reduced organic matter loadings to the 
system was carried out by model simulations using 50% and 75% SOD load reductions at 
Peace River. Figure 6.57 shows the comparison of dissolved oxygen concentrations at 
CH004 and CH006 before and after 50% SOD reduction and 75% SOD reduction. The 



246 
dissolved oxygen concentration in the upper Charlotte Harbor is very sensitive with SOD 

reduction. With 50% SOD load reduction, the surface and bottom dissolved oxygen 

concentration are increased from maximum 2 mg/L to 0.2 mg/L. With 75% load reduction, 

no hypoxic event was found during this simulation and the dissolved oxygen distribution 

exhibit minimal vertical stratification. Although there is no hypoxia event, some localized 

low-levels of near-bottom dissolved oxygen and vertical stratification were maintained even 

during the 75% SOD load reduction scenario. 

To quantify the relationship between SOD coefficient and hypoxia, the area which 
exposed in bottom water hypoxia condition (DO is less than 2 mg/L) was calculated with 0.5 
(75% reduction), 1.0 (50% reduction), and 2.0 (baseline) rate constant of SOD at 20 °C 
during the simulation period (Figure 6.56). With 2.0 as the rate constant of SOD, over 60 
km 2 of the upper Charlotte Harbor was hypoxic during the summer season, while hypoxia 
condition was not observed with 0.5 as the SOD rate constant. The maximum areal extent 
of hypoxia 67 km 2 was on August 21, 2000. This was also accompanied by a high degree 
of stratification as indicated in Figure 6.40. 

With further scientific understanding, the integrated modeling approach would enable 
the development of a science-based management tool which is built on process-based 
understanding rather than simple regression. Subsequent refinement of this integrated model 
can be used to address ecosystem management issues such as controlling estuarine 
eutrophication and determining allowable external nutrient loading levels to restore water 
quality in estuarine system. 



247 





10 


.— - 


8 


E 


6 


•^^ 




O 


4 


Q 


2 




°, 


1 


50 


3 


40 


(D 




^T 


30 


>. 




SZ 
Q. 


20 


O 




O 


10 


SZ 




O 


U , 




260 



280 



baseline near surface 

baseline near bottom 
-100 % N load reduction near surface 
' 1 00 % N load redcction near bottom 



VA*vvw*w*s#//w\tf*MMwWA 



60 



2000 



200 



^ 150 

* 100 



50 



i60 



180 



200 



220 



240 



260 



280 



260 



280 



180 



200 



ijiiiiliiijiiiljjiiiiiiiiliiiiiiiiiii 
220 



i 



dlWilillLi '■'■ 



liiuilVai'i'iili'uiiiiii'ii'iVuimiii 



240 



260 



280 



300 



320 




300 



320 




300 



320 



300 



320 




u i60 180 200 220 240 260 280 300 ~ 320 

Figure 6.48 The water quality species concentrations at CH004 water quality measured 
station before and after 100 % nitrogen load reduction from the Peace River. 



248 





10 


^^ 


8 


E 


6 






O 


4 


D 


2 




°, 


1 


50 


3 


40 


CO 




— 


30 


>> 




a. 


20 


o 






- ***$* ' SJrS 



60 



180 



200 



'^%M^r% 



220 



240 



'fi ^ 



260 



280 



10 



o 



baseline near surface 

baseline near bottom 

100 % N load reduction near surface 

1 00 % N load redcction near bottom 



^VWtilii^^HWWi 



1 60 



180 



200 



220 



240 



260 



280 



200 



220 



240 



260 



280 





400 


i 

s 


300 
200 


Q. 


100 



Q 



300 



320 



300 



320 




320 



300 



320 




160 



180 



Tor 



220 



240 



260 



280 



300 



320 



1 

O 

Q. 



200 
150 
100 

50 
Q 



V^. 




Figure 
station 



160 180 200 220 240 260 280 300 320 

6.49 The water quality species concentrations at CH006 water quality measured 
before and after 100 % nitrogen load reduction from the Peace River. 



249 



5 

o 

Q 



■"TW ^ 



u i60 



ML . \^x^ %^jM'ku^..jf II 

200 220 240 260 280 



1 '•».., 



*% 






50 
40 

m 

= 30 
q. 20 

2 

o 



10 



o 



180 



baseline near surface 

baseline near bottom 

100 % N load reduction near surface 

1 00 % N load redcction near bottom 



300 



320 



MwtomHttmi&NNtMNHHNlkNItlllt 






160 



180 



200 



"221T 



240 



260 



280 



300 



320 




320 



300 



320 



1000 

800 

"j* 600 

a 400 I 

200 



u i60 

500 
e. 400 r 
2 300 



mftmt ^ mt ^^ 



180 



200 



220 



240 



260 



280 



300 



320 



O 
a 



200 
100 

q 



^*^ m* kmi iiML 



Hm- 



m^^^M j: ^ lm ^ 



iMMftto 



1 60 180 200 220 240 260 280 300 320 

Figure 6.50 The water quality species concentrations at CH004 water quality measured 
station before and after 100 % phosphorous load reduction from the Peace River. 



250 



o 



10 
8 
6 
4 
2 
Q 




'^^^^m^mL^' 



ML3 



W^^ 



60 



180 



200 



220 



240 



260 



280 



300 



320 



*> 50 

3 40 
n 

= 30 

q. 20 
o 

S 10 

5 °.6ir 

2000 



fi 1500 
w 1000 l 

z 

* 500 



baseline near surface 

baseline near bottom 
- 1 00 % N load reduction near surface 
' 1 00 % N load redcction near bottom 



^*Vftnt»»»»» 



160 



m mmmm^0^Smmmmm0S»j 



ttUUtU MttHUUk,,, ,/ ( |i 



180 200 



220 



240 260 



280 



300 



320 



•" " '••vm u t HMinmii 



4 -»»»» 




1000 


• 




800 


. 










1> 600 


. 




i 












q. 400 


■ 




H 






200 







180 



200 



<G&m0B aaai£ m^kmu* 



220 



240 



260 



320 



o 

Q. 


SUU 

400 
300 
200 
100 


■ 






..^..^Wf^few.-. 


4* 




>0 


180 


200 


220 240 


260" 280 360 32 



Figure 6.51 The water quality species concentrations at CH006 water quality measured 
station before and after 100 % phosphorous load reduction from the Peace River. 



251 




100 



baseline near surface 

baseline near bottom 

1 00 % N load reduction near surface 

1 00 % N load redcction near bottom 




400 
^ 300 

* 200 

X 

Z 100 











• J 


i 


tyi|i||jjijj|tj|j ' 

ij|f|fUj|L, | 


*'i 


60 


180 


200 


220 240 



260 



280 



300 



320 




u i60 180 200 220 240 260 280 300 320 

Figure 6.52 The water quality species concentrations at CES02 water quality measured 
station before and after 100 % nitrogen load reduction from Caloosahatchee River. 



252 



■a 

E 

O 

D 






>> 
SZ 
D. 
O 



10 
8 
6 
4 
2 



M WWW I flM W M 



w^w *^«HW t lHw» w^ ^ 



H mm+tw+ ^H 



'i60 



180 



200 



220 



240 



260 



280 



300 



320 



50 
40 
30 
20 

10 f- 



baseline near surface 
baseline near bottom 
- 1 00 % N load reduction near surface 
1 00 % N load redcction near bottom 



1 **M* w< *nv™,w/A^/llfl 



WWWM l w il WHH i^ 



60 



180 



200 



220 



240 



260 



280 



300 



320 



3 
z 
i- 



2000 

1500 

1000 

500 



^ m ^mm#m t umm\w* 



i60 



180 



200 



220 



240 



260 



280 



300 



320 



5 

a. 



400 
300 
200 
100 



^^m^^^^^S^^^SSZ 



w *" M **Hvw'wu 







u i60 
400 r 
300 
200 
100 



180 



200 



220 



240 



260 



280 



300 



320 



ikMHUuHHn ttn m t tm " > tf * m '* " ■« *" * * *mfiWtm*nM ' " *% l »M»%^ | /r^^ 



1 

O 

Q. 



u i60 
200 r 
150 
100 

50 



180 



200 



220 



240 



260 



"280* 



300 



*320 







MHHttttt 0lto* i * i W MU M HH w 



.ji ^fc i tt ii 



i 60 180 200 220 240 260 280 300 320 

Figure 6.53 The water quality species concentrations at CES08 water quality measured 
station before and after 100 % nitrogen load reduction from Caloosahatchee River. 



253 



10 

8 

E 

O 

Q 



_ : 


.ll/V I 

llnJn ' 






60 


180 


200 


220 240 260 280 300 32 



baseline near surface 

baseline near bottom 

1 00 % N load reduction near surface 

1 00 % N load redcction near bottom 




1000 
800 r 
"^ 600 
a 400 

200 r 







,«"•-'" 









"* ^ **W. 



1 60 



180 



200 



220 



240 



260 



280 



300 



320 



500 
£. 400 
J 300 

J 200 

°- 100 



A, 



, jfcyTf , , fa , 



.v'.':' 



sj^%i 



#T'W| 



W W ""%** 



*~J .»— «<ft*»W»»H»Vm~- 



u i60 180 200 220 240 260 280 300 320 

Figure 6.54 The water quality species concentrations at CES02 water quality measured 
station before and after 100 % phosphorous load reduction from Caloosahatchee River. 



254 



E 
o 

Q 



=1 



a 

2 
o 

O 



10 

8 

4 
2 f 






immiiim w m w ww 



240 



260 



280 



300 



320 



50 
40 r 

30 
20 
10 



60 



•5. 

a. 



2000 

1500 

1000 

500 



%*. — 



1 60 






"a. 

3 



400 
300 
200 
100 
Cj 

400 
300 
200 
100 



60 



1 60 



o 

Ql 



200 
150 
100 

50 



baseline near surface 

baseline near bottom 

100 % N load reduction near surface 

1 00 % N load redcction near bottom 



*— iw w»^». M<>(#ffWW ^NW^ MMi w *< »* w«tMwwM MM1 — riff**^ jt^'-^ ^^ g j i j '^i ^V i Ify^***' Wi ll I 'f^ l Wf'^TyV l ' i W ' VttU i .u 



180 



200 



220 



240 



260 



280 



300 



320 



^^iimmmm^''fmtmm^ 



1 ■ 'numwmwm 



180 



200 



"22T 



240 



260 



280 



300 



320 



180 



200 



"22T 



240 



260 



~2&0 



300 



320 



HH ^ M i n^ llll^^ 



180 



200 



220 



240 



260 



280 



300 



320 



IW MXI Mi litJl fc WW lWWMMMM 



, m i, ) .-v,'i , "i-,-,--^-- l '-;ri , , ' ,; "' ■"■ ■■"■->-- 



u i60 180 200 220 240 260 280 300 320 

Figure 6.55 The water quality species concentrations at CES08 water quality measured 
station before and after 100 % phosphorous load reduction from Caloosahatchee River. 



255 




50 



100 



150 200 250 

Julian Day 



300 



10r 



c 

D) 



I 

O 



8 








50 



at surface for Baseline simulation 

at bottom for Baseline simulation 

at surface for 1 00 % organic matter reduction 

at bottom for 1 00 % organic matter reduction 



100 



250 



-I I I !_ 



300 



i i i i 



150 200 

Julian Day 

Figure 6.56 Dissolved oxygen and chlorophyll_a concentrations at CH006 water quality 
measured station before and after 100 % organic matter load reduction from Peace River 
using DiToro's sediment flux model. 



256 



50 % SOD reduction at CH004 station 



1 
E 

c 

<D 
U> 
>• 
X 

O 

■a 

I 

o 
en 
w 



6 

c 
>< 

X 

O 

J 

O 

(/> 

W 



§ 

c 
a) 

5. 
x 

O 

-o 

I 
o 
</> 




ian Day 
50 % SOD reduction at CH006 station 




10r 

8 

6 
4 
2 




luflan Day 200 250 300 

75 % SOD reduction at CH004 station 




Surface, for Baseline 
Bottom, for Baseline 
Surface, for Reduction 
Bottom, for Reduction 



m 



ian Day 
75 % SOD reduction at CH006 station 




irian Day 

Figure 6.57 Dissolved oxygen concentrations at CH004 and CH006 water quality 
measured stations before and after 50 % SOD reduction and 75% SOD reduction. 



257 



80 r 



70 



eg 

E 


60 


-* 




(0 


50 


<D 




l_ 




< 


40 


A3 




X 

o 


30 


Q. 




> 
X 


20 




10 








SOD = 2.0 g/m 3 /day 

SOD = 1.0g/m 3 /day 
- - SOD = 0.5 g/m 3 /day 




150 200 

Julian Day 

Figure 6.58 The comparison of hypoxia area at Upper Charlotte Harbor according to 
varying SOD constant rate at 20°C. 



CHAPTER 7 
CONCLUSION AND DISCUSSION 

An integrated modeling system, CH3D-HMS (Sheng et al.,2002), which includes a 3- 
D hydrodynamics model, a 3-D sediment transport model, and a 3-D water quality model, 
has been enhanced and applied to the Charlotte Harbor estuarine system. Circulation and 
water quality in the entire estuarine system have been simulated and validated with field data 
from numerous sources in 1986, 1996, and 2000. To reproduce bottom water hypoxia in 
upper Charlotte Harbor, models of oxygen balance and oxygen fluxes at the air-sea and 
sediment-water interfaces are enhanced. To achieve a better understanding of the temporal 
and spatial variation of temperature and light, and their effect on water quality processes, 
the 3-D temperature model, with a heat flux model at the air-sea interface, and the physics- 
based light attenuation model (Christian and Sheng, 2003) were applied. Major conclusions 
of this study are summarized in the following: 

1) A fine-resolution numerical grid which accurately represents the complex 
geometrical and bathymetric features in Charlotte Harbor estuarine system was generated and 
used to simulate the hydrodynamic, salinity and temperature characteristics in 1986 and 
2000. A sensitivity study was conducted to examine the sensitivity of model results to such 
factors as boundary conditions, model coefficients, bathymetry, advection scheme, and grid 
resolution. Flow, salinity and temperature patterns produced by hydrodynamic model agree 
well with existing data. The normalized RMS error analysis demonstrated model's ability 



258 



259 
to simulate water level, currents, salinity and temperature within 7.5 %, 20%, 6.5 % and 7% 
accuracy, respectively. 

2) The calibrated circulation model was applied to assess the impact of the removal 
of the Sanibel Causeway and IntraCoastal Waterway on the flow and salinity pattern in the 
San Carlos Bay and Pine Island Sound and to develop minimum flow criteria for the 
Caloosahatchee River. The results show that these hydrologic alterations do not appear to 
show noticeable impact on the flow and salinity patterns in the San Carlos Bay and Pine 
Island Sound. The salinity at the Caloosahatchee River Mouth is reduced by about 1.36 ppt 
in the absence of the IntraCoastal Water way. 

3) The relationship between salinity at Fort Myers and river discharge was established 
by comparing 1 -day and 30-day averaged salinity values at Fort Myers and fresh water inflow 
at S79 (Figure 5.35). According to this relationship, a total river discharge of 18 m 3 /s at S79 
produces a 1-day averaged salinity of 20 ppt at Fort Myers. 

4) The newly enhanced integrated model, CH3D-IMS, was used to perform 1996 and 
2000 hydrodynamic, sediment transport, and water quality simulations of the Charlotte 
Harbor estuarine system. The normalized RMS errors of model simulation vary generally 
within 10 - 45% for all stations and all species. The agreement between simulated and 
measured DO is particularly good with less than 20% error everywhere. The daily and 
seasonal vertical distribution and fluctuation of DO are successfully simulated by applying 
measured SOD at the upper Charlotte Harbor. 

5) The upper Charlotte Harbor system has been suffering summer hypoxia in bottom 
water in the past decade. In this study, CH3D-IMS was enhanced and calibrated to analyze 
this phenomenon. The calibrated model was used to examine the dynamics of various factors 



260 
that can affect hypoxia, including freshwater inflow, tides, SOD, and water column oxygen 

consumption. The model results suggest that hypoxia in the upper Charlotte Harbor is 

primarily caused by the combined effects of SOD and vertical salinity stratification, and also 

enhanced by water column oxygen consumption. 

6) The integrated model was applied to assess the effects of hydrologic alterations 
and to provide a preliminary evaluation of pollutant load reduction goal (PLRG). The results 
show that these hydrologic alterations do not appear to show noticeable impact on the water 
quality patterns in the San Carlos Bay and Pine Island Sound. The dissolved oxygen 
concentration of the estuarine system does not response with different load reduction because 
of the use of the same SOD values for all load reduction simulation. In reality, SOD is 
expected to be reduced when nutrient loading is reduced. However, due to lack of detailed 
data of the sediment diagenesis and lack of understanding of the relationship between SOD 
and nutrient loading, the model cannot simulate the response of SOD to load reduction. 
Further research is needed to develop the ability to simulate the effect of nutrient load 
reduction. 

7) A systematic calibration procedure has been developed for a more efficient and 
more objective calibration of the water quality model. A consistent framework for 
systematic calibration is formulated which include following step: model initialization, 
sensitivity analysis, parameterization, and formulation of calibration criteria. 

8) Although the integrated model developed and applied in this study have performed 
well simulating hydrodynamic and water quality components in the Charlotte Harbor 
estuarine system, the model contains various uncertainties, assumptions and simplifications 
which need further investigation and improvement. Several possible improvements in the 



261 
developed integrated model and year-long simulation of hydrodynamic and water quality 
dynamics in the system are: 

a) A finer grid system is needed to represent the navigation channel, causeway, small 
islands, and complex shoreline. With grid resolution used in this study, the model cannot 
solve the dynamics of the shallow regions less than 1.5 m in depth of the system, where wave 
interactions become important. 

b) Sediment oxygen demand (SOD) depends on the deposition and decomposition 
of organic matter on the seabed, and the exchange of nutrients and oxygen across the 
sediment-water interface. In order to predict trends in hypoxia in response to organic loads, 
it is necessary to include oxidation-reduction processes with CH 4 , H 2 S, and organic matter 
(DiToroetal. 1990). 

c) More accurate river discharge data are needed. The river discharge data at Peace 
and Myakka rivers are underestimated since not all freshwater inflows are accounted for by 
the data. To solve this problem, the integrated model need to be coupled with watershed 
model to supply missing freshwater discharge 

d) Phytoplankton dynamics are different with different phytoplankton species. To 
better understand phytoplankton process and eutrophication characteristics, gathering more 
phytoplankton species data and applying models of multiple phytoplankton species are 
needed 

e) Submarine groundwater is an important mechanism in delivering chemical 
elements to surface waters, and can be of the same order of magnitude as surface water 
sources (Oliveria et al, 2003). Rutkowski et al. (1999) suggest that the nutrient flux by 
submarine groundwater is on the same order of magnitude as a small river. To address this 



262 
issue, a nutrient advection by coupling with groundwater model with the integrated model. 

f) Spatially more detailed and temporally more frequent hydrodynamic and water 
quality data inside the Charlotte Harbor and at the river mouths are needed. Data of nutrients 
and organic material in the sediment column should be gathered to enable simulation of 
hypoxia in the upper Charlotte Harbor. 



APPENDIX A 
FLOWCHARTS FOR CH3D-IMS 

The following flow charts illustrate the order of operations performed in the 

integrated modeling for Charlotte Harbor estuarine system. Figures A. 1 illustrate the flow 

chart of the main CH3D program (ch3d.f). Main program, ch3d.f, call the subroutine 

ch3dm2.f which is the driving subroutine for time stepping of the solution as well as defining 

the time varying forcing functions and generating output shown in Figure A.2. Water quality 

model and sediment transport model are coupled in this subroutine. Figure A.3 and A.4 

show the initializing sediment parameters and main routine of sediment transport model, 

respectively. Water quality model was illustrated in Figure A.5 through A.7. 



263 



Chart 1: main program (ch3d.f) 



264 




Set screen output destination 
(ch3dfileopen.f) 



Output the version number of the fortran/script files 
(ch3d_versionsf) 



Read optional command line arguments 
(ch3dclo.f) 



Read grid & bathymetry 

(ch3dgdep f) 



Read hydrodynamic input file 
(ch3dir_hyd.f) 



Read salinity input file 
(ch3dir_salf) 




Curvilinear coordinate transformation constants 
(ch3dtr.f) 



Modify bathymetry (hmin and hadd) 
(ch3dih.f) 



Nondimensionalize and set constants 
(ch3dnd.f) 



Set cell boundary direction flag and sweeping indices 
(ch3dii.f) 



Read temperature input file 
(ch3dir_tem.f) 



Figure A.l Flow chart for the main CH3D program. 



265 



Initialize arrays 
(ch3dif.f) 



Compute constants 
(ch3div.f) 



Set wind stress 
(ch3dws.f) 



Is flushing model included? 
(_FLUSHING_MODEL_) 




F * 



^ Is tidal forcing ~~^_ 
<C_ read from file? 

-^(ITIDE < 0) ^ " 


--^ T 


Read tidal forcing from file 
(ch3dtd.f) 




F 






















1 


r 


Details in Chart 2 






Main time iteration loop 
(ch3dm2.f) 

















f END J 

Figure A.l continue. 



Read flushing input file 
(ch3dirjlu.f) 



Set initial concentration 
field (ch3dics.f) 



266 



Chart 2: main time-iteration loop (ch3dm2.f) 



Called from ch3d.f : Chart 1 



-lag for initiaT 

surface elevation 

XISURF = 3L 



Interpolate the initiatial 

surface elevation 

(ch3dinitsurf.f) 



Set depth array 
(ch3ddp.f) 




Initializing sediment model 
Chart 3-1 



Initialize sediment 

model 

(ch3dsedi.f) 



Initializing water quality model: 
Chart 4-1 



initialize 

water quality model 

(ch3dwaini.f) 




Set temperature 

at river boundary 

(ch3dtk.f) 




Initialize river 
arrays(ch3dri.f) 




Calculate 3-D velocity 

at river boundaries 

(ch3dri_3d.f) 



Is precipiation 

^/evaporation included?^ 

(IRAIN * 0) 



Initialize rainfall 

/evaporation arrays 

(ch3drain.f) 



Figure A.2 Flow chart for the driving subroutine for the time stepping of the solution. 



267 




Initialize heavily used 

transport coefficients 

(ch3dcoef.f) 



Start of time loop 



Time varying inflow 

river discharge routine 

(ch3drv.f) 




Calculate water 
density&arroclinic 
pressure in 
sigma plane 
(ch3dde_i.f) 
(ch3ddej.f) 




ISIGZ=1 



CSD 



Calculate water 

density&arroclinic 

pressure in 

z plane 
(ch3dde_z_i.f) 
(ch3dde_z_j.f) 



1s vertical eddy- 
coefficient varying? 
(IEXP*01 



Computer variable 

turbulent eddy coefficients 

(ch3ded.f) 



Figure A. 2 continue. 



268 



Update internal and external variables 
(ch3dreset.f) 




Set wind field 
(ch3ddwt.f) 





T 






oes rainfall vary witff-^ 


Update the time 
varying rain (ch3drt.f) 


time?(IRAIN > \^ 






^F 







Is real tide being 
used?(ITIDE<0) 



Read tide data and 
interpolate (ch3dti.f) 



Ts harmonic tide beinc] 
used? (ITIDE>0L 



Update old depth 



Tidal forcing calculated 

from harmonic input file 

(ch3dtd.f) 



Compute Ul"* 1 and S* for2-D model 
(ch2dxy_i.f) 



Compute Vl~' and S"- 1 for 2-D model 
(ch2dxy_j.f) 



Update new depth arrays 
(ch3ddp.f) 



Figure A. 2 continue. 



269 





Update velocities 

at river boundaries 

(Ch3dri_3d.f) 



Compute velocities in x/y/z directions 
(ch3dxyz_i.f) 
(ch3dxyzj.f) 
(ch3dxyz_k.f) 



Compute inertia&diffusion 

for 3-D model in x/y directions 

(ch3ddi_i.f) 

(ch3ddij.f) 



Compute inertia&diffusion 

for 2-D model in x/y directions 

(ch2ddi_i.f) 

(ch2ddi_j.f) 



Is salinity 

being simulated? 

(IT>ISALT & 

ISALT#0) 



Salinity transport 
(ch3dsa.f) 




-Is temperature^ 

being simulated? 

(IT>ITEMP & 

ITEMP*0) 



Temperature 
transport (ch3dte.f) 




Suspended sediment 
transport (ch3dsed.f) 



Sediment model: 
Chart 3-2 



Figure A. 2 continue. 



(2D_2) 



270 




Water quality model 
Chart 4-2 



Water quality transport 

and chemical reactions 

(ch3dwq.f) 



Conservative 

species transport 

(ch3dcs.f) 




Conservative 

species transport 

for 2-D model 

(ch3dcs.f) 




Check conservation 
(ch3dcheckcons.f) 



Output Hydrodynamics&salinity 
model results(ch3dot.f) 




Output sediment&wave 

model results 

(ch3dot_sed.f) 




Run "poor mans RPC 
via tecplot (ch3dpmrpc.O 



End of time loop 




Return to chart 



Figure A. 2 continue. 



271 



Chart 3-1: Sediment initialization and fetch (ch3dsedi.f) 



Called from ch3dm2.f: Chart 2 



Read sediment input file and wave 

induced bottom shear stress 

(ch3dirs.f) 



Initialization 
(ch3disd.f) 



Compute fetch 
(ch3dft.f) 




.Return to ch3dm2.f: Chart 2) 



Compute wave induced bottom 

shear stress 

(ch3dbs.f) 



Figure A.3 Flow chart for the initializing sediment transport model 



272 



Chart 3-2: Sediment model (ch3dsed.f) 



Called from ch3dm2.f: Chart 2 




T 



Compute wave induced bottom 
shear stress 

(ch3dbs.f) 



Calculate settling, deposition velocities and erosion rate for 
fine and coarse sediment 



Fine group advection, diffusion 
(ch3dsd1 .f) 



Coarse group advection, Diffusion 
(ch3dsd2.f) 



Calculate sediment bottom change 



Return to ch3dm2.f: Chart 2) 



Figure A.4 Flow chart for main sediment transport model 



273 



Chart 4-1 : Initializing Water Quality Model (ch3dwqini.f) 



Called from ch3dm2.f :chart 2 



Initialize WQ variables 



Read water quality input file 
(ch3dir_nut.f) 




Read dissolved oxygen data for several 

stations (fort. 310) and time and space 

interpolate (ch3drdoxy.f) 



F •* 



Read incident light data (fort.307) 
and time interpolate (ch3drlight.f) 



Read temperature data 

for each segment(fort.306) 

and time interpolate 

(ch3drtemp.f) 




Read algae growth rate for each segment 
(fort.309) and time interpolate (ch3dragrm.f) 



Read color data for each segment 
(fort.308) and time interpolate (ch3drcolor.f) 



Figure A.5 Flow chart for initializing water quality model 



274 




Sgad initial WQ at aerobic sediment layer. 
(INIT_WATER t 0) 




Read light information:Table for absorption rate 
according to wave length from fort.321 



Read all initial WQ parameters 
at computational grid points (fort.99) 



Read initial WQ for water column(fort.303) 
and spatial interpolate (ch3dinterp.f) 



Read initial WQ for aerobic sediment column 
(fort.301) and spatial interpolate (ch3dinterp.f) 



Read initial WQ for anaerobic sediment layer 
(fort.302) and spatial interpolate (ch3dinterp.f) 



Set initial WQ area for conservation check 
and initialize assimilation terms 




Read river boundary data from fort.304 or fort.305 
time interpolate (ch3drrivwq.f) 



F K 




Return to ch3dm2.f : chart 2 



Figure A.5 Continue. 



275 



Chart 4-2 : Water Quality Model (ch3dwq.f) 



Called from ch3dm2.f :chart 2 



i 



Set up assimilation terms 
and Reset WQ variables 




Read river boundary WQ data (fort.305) 
and time interpolate (ch3drrivwq.f) 



Calculate horizontal & vertical advection, horizontal diffusion 
in water column and specify tidal and river boundary (ch3dtrp.f) 



Calculate solar radiation angle 
(solar_angle.f) 



Compute incident light 
(ch3dlig.f) 




Read incident light data (fort.307) 
and time interpolate (ch3drlight.f) 




Read color data for each segment 
(fort.308) and time interpolate (ch3drcolor.f) 



Read dissolved oxygen data for several 

stations (fort.310) and time and space 

interpolate (ch3drdoxy.f) 



Read temperature datafor each segment 
(fort.306)and time interpolate(ch3drtemp.f) 



Read algae growth rate for each segment 
(fort.309) and time interpolate (ch3dragrm.f) 



Figure A.6 Flow chart for main water quality model. 



276 



Setup Z- vertical grid 

include water and sediment column 

(gridld.f) 



Compute temperature 

and light functions 

(temlit.f) 



Temperature and light coefficient 
: Chart 5 



Compute vertical diffusion and chemical reaction 

for plankton species in water column 

(ch3dalg.f) 



Prepare variables and convert to z-direction 

1 dimensional variables in water and sediment column 

for WQ simulations (preset.f) 



Compute vertical diffusion and chemical reaction 

for nitrogen species in water and sediment column 

(ch3dnitro.f) 



Compute vertical diffusion and chemical reaction 

for phosphorous species in water and sediment column 

(ch3dphosp.f) 




Compute DO and CBOD 
(ch3doxy.f) 



Post processing of WQ variables 
(postset.f) 



Check conservation fro WQ species 
(ch3cons_wq.f) 



Output water quality model results 
(ch3ot_nut.f) 




Return to ch3dm2.f : chart 2 




Figure A. 6 continue. 



Chart 5: Sets up temperature and light functions for water quality model 
and finds light intensity in a particular water column (temlit.f) 



277 



Set up temperature functions 
For water quality 




Set coefficient for algal growth 
based on light intensity 




Return to ch3dwq.f: Chart 4-1 




Figure A.7 Flow chart for computing the temperature and light attenuation functions for 
water quality model. 



APPENDIX B 
DIMENSIONLESS EQUATIONS IN CURVILINEAR BOUNDARY-FITTED AND 

SIGMA GRID 

Non-dimensionalization of governing equations make it easy to compare the relative 

importance of various terms in the equations. The governing equations are non- 

dimensionalized using the following reference scales: X r and Z r are the reference lengths 

in the vertical and horizontal directions; [/, is the reference velocity; p , p r and 
Ap = p- p are the reference density, mean density and density gradient in a stratified 
flow; A Hr and A Vr are the reference eddy viscosities in the horizontal and vertical 

directions; D w and D Vr are the reference eddy diffusivities in the horizontal and vertical 
directions. The dimensionless variables can be written as (Sheng, 1983) 

(x\y\z*) = (x,y,zX r /Z r )/X, 

(u\v\w*) = (u,v,wX r /Z r )/U r 
m =coX r IU r 
t*=tf (B.l) 

(f>;) = (r;,<)/( Po iz r [/ r ) = (r;>;j/r r 
C = gCi(fu r x r )=cis r 

A\r = A/ ' A Vr 

D' H =D H / D Hr 
D* v =D v /D Vr 

278 



279 

These dimensionless variables can be combined to yield the following dimensionless 
parameters 

Rossby number : R = 



Froude Number : F = 



° 4^ 
u. 



r M 

F 

Densimetric Froude Number : J7 = !_ /g ? ^ 



Vertical and Horizontal Ekman Number : E v = —^- , E H = —&- 



ft ft 



A A 

Vertical and Horizontal Schmidt Number : S rv =—^-,S ru = - Jt - 

Dvr D Hr 



(3 = ^ 



A, 

2 v 2~ — 



K F rJ 



In three-dimensional modeling, complex bottom topographies can be better 
represented with the application of o-stretching (Sheng, 1983). This transformation allows 
the same vertical resolution in the shallow coastal areas and the deeper navigation channels. 
The vertical coordinate, z, is transformed into a new coordinate, o, by (Phillips, 1957). 

^_ z-£(x,y,t) 

h(x,y) + C(x,y,t) (B3) 

where h is the water depth and Q is the surface elevation. 

In this new vertical coordinate, the vertical velocity is calculated by the following 
equation. 



280 



DC 
w = H co + (\ + a)—?- + a 

Dt 



dh dh 
u — + v — 
dx dy 



) 



(B.4) 



u dz da . , 

where w = — in the z-plane, CO = in the o-plane 

dt dt 

Using non-orthogonal boundary-fitted horizontal grid, it is possible to better represent 
the circulation and transport processes in estuarine systems with complex shoreline 
geometries. Using the elliptic grid generation technique developed by Thompson (1982) and 
Thompson et al. (1985), a non-orthogonal boundary-fitted grid can be generated in the 
horizontal dimension. To solve for flow in a boundary-fitted grid, it is necessary to 
transform the governing equations from original coordinates (x, y) to the transformed 
coordinates ($, r\). The spatial coordinate system in the computational plane (£, r|) is 
dimensionless while the coordinates in the physical plane (x, y) have dimensions of length. 
During the transformations, the velocities are transformed into contra-variant velocities. 

In the boundary-fitted, curvilinear, o-stretched, non-dimensional coordinate system, 
the continuity and momentum equations are 



dt 



^(^»+4(^» 



do 



(B.5) 



281 



1 dHu 
H~dT 



( 



s d£ * drj 



S~ , 8 

rll + , V 



^V 60 



^ 



■Vn 



_a_ 
a 



( y{ 4^ HUU + y„ \[8~0 HUV ) + T- ( Vf V^# "V + ^ yfg^Hw} 



Ro 



-So" 



«£( 



a//i 



(B.6) 



wv 



ao- 



ap + 12 a^ 
a<f a;/. 



Jcr + 



^fr^flfr^H 



+ 



— 7- — A^ — +E H A H (Horizontal Diffusion of u) 
H da\ da j 



i a//v 

ff ar 



d# * a^ 



+ 



.21 



\ 



U + , V 

"n/So 



+ AL 



So# 



a_ 
a_ 

dHvw 



So 



f: 



«t 



da + 



+ 



e u a 



H 2 do- 



da j 

21 ^P 22 ^/° 

dv\ 
\- — +E H A H (Horizontal Diffusion of v) 
da ) 



2l dH , 2 dH 



[l#° 



+ crp 



(B.7) 



is the determinant of the matric tensor, g u , which is defined as 



(B.8) 



§"- 



x: + 



i + y e 



x s x , + y*y n 



_ x n x { + y,yz x v + y, 



Sn 812 

621 <?22 



(B.9) 



whose inverse is 



282 



g' J = 



4 + yi -( x ^ n + y { y n ) 

-<XnXe+y v ye) x ]+y] 



11 12 

8 8 
8 8 



(B.10) 



As shown in Sheng (1986), the contravarient components («') and physical 
components (u(I)) of the velocity vector in the non-Cartesian system are locally parallel or 
orthogonal to the grid lines, while the covarient components («,) are generally not parallel 
or orthogonal to the local grid lines. The relationship between the physical velocity and 
contravarient velocity is given by (Sheng, 1986) 



u(i) 



with summation on i. 



The salinity and temperature transport equations can be written as 



dHS 

dt 



HS CV da 



3 f D v ^ 
v da 



K 



dHwS 
da 



& 



VioW 



±(4J- HuS) + ±(^ HvS) 



+ 



+ 



Sch\I8o 



d ( dS 



2 3S" 
) 



d ( l—rr 2. dS f— „ 22 dS ^ 



(B.ll) 



(B.12) 



and 



283 



BHT 

dt 



HS CV da 



3 ( dX 

v da 



K 



H 



+ 



go 



dHwT 
da 



ToHuT) + — (JT HvT) 



*Z 



S cny[g~oldZ 



drj 

8oHg n —+Jg Hg 12 — 

d£ drj 



s cH^l dr l 






i — _. 21 dT / — „dT^ 
d£ drj 



(B.13) 



The sediment transport equation can be written as 



dHc : 



dt H ■ S cv da 



H 



I ( D V 



V 



-/?, 



y 



d(Hco-wJc i 
da 






+ 





Ev 


ScHyJgo 


ScH\jgo 



f 



K 



dc 



go H g "-j: + J go Hg i --± 



V 

d f 



^ 



, 3c, 



3?7 



drj 

„dc" 



\ 



g Hg 2 ^ + Jg~ Hg 



(B.14) 



where c, represents cohesive (i=l) and non-cohesive (i=2) sediment concentrations and w Vi 
is settling velocity for sediment group i. 



APPENDIX C 
COMPARISON OF WATER QUALITY MODELS-WASP, CE-QUAL-ICM, and 

CH3D-WQ3D 



The CH3D water quality model (CH3D-WQ3D) was compared with existing water 
quality models, specially Water Quality Analysis ans Simulation Program (WASP), the 
integrated Compartment water quality model developed by the US Army Corps (CE-QUAL- 
ICM). The methods of coupling with hydrodynamic and sediment transport models, the 
simulated parameters, the assumptions, the chemical/biological processes, and the limitations 
of each model are discussed and compared in Table C.l. The first criterion in model 
comparison is its ability to simulated hydrodynamic, sediment transport, and water quality 
with the efficient coupling. 

The WASP box modeling framework has proven to be an excellent water quality 
model for riverine systems, where the steady state assumption is applicable. In addition, this 
simple box model can be successfully used to perform numerical experiments like sensitivity 
test. However, in marine environments, it should not be used without proper linkage with 
3-D hydrodynamics and sediment models, because tide wind and baroclinic forcing underact 
in an unsteady balance. 

As pointed out by Chen and Sheng (1994), the loosely coupled models such as CE- 
QUAL-ICM, cannot account for nutrient release by sediments in episodic events. Because 
these models were not coupled with a dynamic model for sediment transport, they could not 



284 



285 
accurately consider sediment-process effects such as resuspension, deposition, flocculation, 

and settling on nutrient dynamics in estuaries. Furthermore, this model use equilibrium 

partition with function of biomass for hydrolysis process. This may be a reasonable 

assumption when the time step of the water quality simulation is large compared to the time 

it takes to reach equilibrium. However, sometimes a water quality model may use a small 

time step during which the absorbed and the dissolved nutrient may not achieve equilibrium. 

In this case, a kinetics model, such as sorption/desorption kinetic process in CH3D-WQ3D, 

is needed. Analytical chemists repeatedly found that complete recovery of contaminants 

from soil/sediments frequently requires lengthy extraction periods, abrasive mixing, and 

strong solvents (Witkowski and Jaffe, 1978). These observations are in contradiction with 

the equilibrium models which assume that sorption-desorption reactions are accomplished 

instantaneously. 

The light attenuation model need chlorophyll a and suspended sediment 
concentrations to calculate light attenuation in water column. The water quality model use 
light intensity for limitation of algae growth rate and photosynthesis processes. Therefore, 
without fully coupling these models, it could not be accomplished to communicate each 
others with both direction. 

The chemical/biological processes in all three models are basically similar and 
compatible with each others. Most of them are the adjustable coefficients. Thus, the ability 
to simulate water quality depends on modeler experience, and necessity of model complexity. 
Also the data availability is the key to ensure reliability and accuracy of the results. The CE- 
QUAL-ICM include much more simulating species than those for the other model. Without 
measured data for those species, these complex nutrient cycles could create more uncertainty 



286 
than simple nutrient species model. CE-QUAL-ICM use sediment flux model developed by 

Ditoro and Fitzpatric (1990) for sediment oxygen demand. Although, applying of sediment 
flux model provides rational predictions of sediment response to environmental alterations, 
it increases information requirements and computation time compare to employment of user- 
specified fluxes (Cerco and Cole, 1995). They point out that employment of user-specified 
fluxes show the better calibration of the water quality model than employment of the 
sediment flux model. 

Overall, with the similarity and compatibility of chemical/biological processes which 
is depend on data availability, coupling with hydrodynamic model, sediment transport model, 
temperature model, and light attenuation model is the key to ensure reliability and accuracy 
of model. Specially, the Charlotte Harbor estuarine system which has a strong linkage 
between hydrodynamics and water quality dynamics suggests the importance of coupling 
with hydrodynamics and sediment transport. Therefore, CH3D-WQ3D is better to simulate 
water quality in Charlotte Harbor estuarine system than the other water quality models. 



287 



Table C.l Comparison of water quality models 



model 


WASP5.X 


CE_QUAL_ICMvl.O 


CH3D-WQ3D 


Hydrodynamic 
model 


Read into the 
model as input 
parameters 


read into the model 
as input parameters 


fully coupled with 
circulation, salinity 
and temperature 
transport model 


Sediment 
transport model 


Read into the 
model as input 
parameters 


loosely coupled with 
simple model 


fully coupled with 
sediment dynamic 
model 


Light attenuation 
model 


loosely coupled 
with empirical 
formula 


loosely coupled with 
empirical formula 


fully coupled with 
physics-based light 
attenuation model 


Nitrogen cycle 


NH3, N03, ON 


NH4, N03, SON, 
LPON, RPON 


NH4, N03, SON, 
PON, PIN, NH3 


phosphorous 
cycle 


SRP, OP 


SRP, SOP, LPOP, 
RPOP 


SRP, SOP, POP, 
PIP 


carbon cycle 


Using CBOD 


SOC, LPOC, RPOC 


Using CBOD 


silica cycle 


No 


Available Silica, 
Particulate Biogenic 
silica 


No 


metal cycle 


No 


Iron & Manganese 


No 


phytoplankton 


One species 


Three species 


One species 


zooplankton 


Using zooplankton 
grazing rate 


Using phytoplankton 
predation rate 


Yes 


Metabolism 


Respiration and 

non-predator 

mortality 


One parameter 


Respiration and 

non-predator 

mortality 


re-aeration 


Function of current 
velocity and depth 


Constant 


Function of current 
velocity, depth, and 
wind speed 


sediment oxygen 
demand 


Empirical formula 


Sediment flux model 


Empirical formula 


Variation of 
Ammonification 


Function of 
temperature 


Function of algae 
biomass 


Function of 
temperature 


Variation of 
mineralization 


Function of 
temperature 


Function of algae 
biomass 


Function of 
temperature 



288 



Hydrolysis 


No particulate 


Equlibrium partition 


Sorption/desorption 




species 


with function of 
algae biomass 


kinetic 


phosphorous 


Constant 


Variable ratio 


Constant 


/carbon ratio 




described by 

empirical 

approximation 





APPENDIX D 
NUMERICAL SOLUTION TECHNIQUE FOR WATER QUALITY PROCESSES 



In the finite difference solution of the water quality model, the advection and 
horizontal diffusion terms are treated explicitly, whereas the vertical diffusion and 
biogeochemical transformations are treated implicitly. Fractional step methods, which 
guarantee numerical stability and prevent negative concentrations, are applied in the 
numerical solution (Chen and Sheng, 1994). The horizontal diffusion and horizontal and 
vertical advection terms are solved first. The numerical solutions proceeds with the 
calculation of vertical diffusion, and then biogeochemical transformation reactions. Finally, 
the sorption/desorption reaction terms are solved. Equation (D.l) shows a schematic of the 
numerical solution algorithm method used in this study. 



At 

N n2 -N n{ 



- [Horiz. Advection + Vertical Advection + Horiz. Diffusion]" 

■ = [Vertical Diffusion]" 2 + [Q]" 2 (D.l) 



At 

= [Sorption]" + [desorption] 



N n+l -N" 2 r „ . in+1 r , . in+1 



At 

By solving the sorption/desorption terms separately from other terms, it is possible 
to treat these terms implicitly. To illustrate this, the difference equations for the 
sorption/desorption reactions are examined: 



289 



290 
d ^ d =Hd x N?-d x - Px .c-N? 1 ) 

N> Hl _ N n2 (D.2) 

' ^ ' =<d x N?-d x .p x -c-N? x ) 

where N d and N p are concentrations of dissolved and particulate nutrients such 
as,NH4 and PIN, SON and PON, SRP and PIP, and SOP and POP; dx is desorption rate for 
a specific nutrient species; andp^ is partition coefficient for a specific nutrient species. 

In solving the second step of the fractional step method for organic/inorganic 
phosphorous and nitrogen species, the difference equations for the entire water column and 
sediment column are solved simultaneously. The kinetic processes for particulate nutrient 
and plankton species need to include a settling process, which accounts for the limited 
vertical motion of these species. 



For dissolved species 



For particulate species 



dN 



d _ 



dt dz 



D. 



dN 



\ 



<i 



dz 



+ Q (D.3) 



dN D a 



f 



dz J 



+ Q (DA) 



dt dz 

The exchanges of particulate nutrients at the water-sediment interface are determined 
by the sediment resuspension and deposition fluxes while the exchange of dissolved nutrient 
species between the water and sediment column are automatically included in the numerical 
solutions, with diffusion term. In order to ensure that the water quality model is consistent 
with the sediment model, in terms of bottom exchanges, the erosion and deposition rates 
calculated in the sediment model can be imported into the governing equations for 
particulate nutrients. Therefore, the particulate nutrient concentrations are solved with the 
unit of percentage (jig/jig). Let p be any particulate species (N p ) mass per unit mass of 
sediments, then, 



N=p-c 



291 
(D.5) 



where c is the suspended sediment concentration. Substitute N p with pc into second step of 
fractional step method: 



dpc _ d 
dt dz 



w s pc + D v 



dpc 

~dz~ 



+ Q 



dc dp d 
p — + c— = — 
dt dt dz 



wc + D V — 

v dz 



dz 



v dz 



+ Q 



(D.6) 



(D.7) 



Since the vertical one-dimensional sediment equations are: 



dc d ( dc^ 

— = — w t c + D v — 
dt dzy ' dz 



(D.8) 



Equation (D.7) becomes: 

dt 



wc + D — 

'dz 



dz dz 



dA 

v dz 



+ Q 



(D.9) 



This equation allows us to import the erosion and deposition rates calculated in the 
sediment model into the particulate nutrient species. Since the boundary condition for the 
sediment model at the water-sediment interface is: 



w s c + D v -^- = D-E 
dz 



Equation (D.9) at the water-sediment interface becomes: 



(D.10) 



dt dz dz 



dz 



+ Q 



(DM) 



where D and E are deposition and erosion rate, which are calculated at sediment model. 

Due to the difference in the partial pressures of oxygen and carbon and because of 
other physical and biochemical factors, the transformation processes in the water column and 
in the sediment column are different. There are no phytoplankton and zooplankton species 



292 



in the sediment column. When these plankton species die, they are treated as particulate 
organic species. In the sediment column, there exist two distinctive layers: an aerobic layer 
and an anaerobic layer, within which the transformation processes are not the same. For 
example, in the anaerobic layer, there is no nitrification process due to a lack of oxygen, 
while in the aerobic layer, there is no denitrification process due to the availability of oxygen. 
The vertically, one dimensional z-grid is shown in Figure D.l. 



a> 
o 

Q. 

-o 



KMG 



KMG-1 



CO 

o 

3 



o 
o 
3 



KMS+2 



KMS+1 



KMS 



KMR+1 



KMR 



V 



KMG=KM+KMS+KMR 



Water Column 
KMS=KMR+KMO 



7^ 

o 



-V 



Aerobic Sediment Column 
KMR 



J3 



Anaerobic Sediment Column 



Figure D.l The vertical one-dimensional z-grid 

Since the horizontal transport is generally very weak in the sediment column, the 
mass flows of nutrient species are mainly vertical, and the governing equations are: 



For dissolved species in a sediment layer 



dt dz 



0M 



dz > 



+ Q (D.12) 



293 



For particulate species in a sediment layer 



dN n 

w r N+M p - 

c P dz 



+ Q (D.13) 



dt dz 

where N d is dissolved nutrient species such as NH4, SON, SRP, SOP concentrations (per unit 
volume of porewater), w c is the consolidation velocity of sediments, M is the molecular 
diffusivity, and 0is the porosity. 



APPENDIX E 
NUTRIENT DYNAMICS 

Nutrients are essential elements for life processes of aquatic organisms. Nutrients 
of concern include carbon, nitrogen, phosphorous, silica and sulfur. Among these nutrients, 
the first three elements are utilized most heavily by zooplankton and phytoplankton. Since 
carbon is usually available in excess, nitrogen and phosphorous are the major nutrients 
regulating the ecological balance in an estuarine system. Nutrients are important in water 
quality modeling for several reasons. For example, nutrient dynamics are critical 
components of eutrophication models since nutrient availability is usually the main factor 
controlling algae bloom. Algae growth is typically limited by either phosphorous or 
nitrogen. (Bowie et al., 1980) 

Nutrient inputs to estuarine systems are related to point and non-point sources from 
land, atmospheric deposition, and fixation. Additionally, internal loadings such as from 
resuspended sediments containing inorganic and organic forms are also important. The 
specification and quantification of each of these contributions are the first steps towards the 
determination of nitrogen and phosphorous budgets in an estuarine system. 

Nutrient cycles are highly dependent on the hydrodynamics and sediment dynamics 
of the estuarine system. Resuspension events, combined with desorption processes, can 
significantly change the input and budget of nitrogen and phosphorous in the system. On the 
other hand, deposition and sorption may contribute to major losses of nitrogen and 



294 



295 
phosphorous from the water column. The hydrodynamics not only derive the sediment 

processes, but also affect the sorption/desorption reactions, through turbulent mixing. 
E.l Nitrogen Cycle 

Nitrogen can be classified into two groups: dissolved nitrogen and particular nitrogen. 
The criterion of this division is established in the laboratory using filtering technique. The 
dissolved nitrogen include ammonia nitrogen (NH3), dissolved ammonia nitrogen (NH4), 
nitrite and nitrate nitrogen (N03), and dissolved organic nitrogen (SON). Particulate 
nitrogen includes particulate inorganic nitrogen (PIN), and particulate organic nitrogen 
(PON). Phytoplankton nitrogen (PhyN) and zooplankton nitrogen (ZooN) related biomass 
to nitrogen concentration through a fixed stoichiometric ratio: nitrogen-to-carbon ratio (A NC ). 

The model nitrogen cycle (Figure E.l) includes the following processes. 

1) Ammonification of organic nitrogen 

2) Nitrification of ammonium 

3) Volatilization of ammonia 

4) Denitrification of nitrate 

5) Uptake of ammonia and nitrate by phytoplankton 

6) Conversion of phytoplankton nitrogen to zooplankton nitrogen by grazing 

7) Excretion and mortality by phytoplankton and zooplankton 

8) Settling for particulate nitrogen 

9) Sorption/desorption reactions 



296 



tVolatili 



NH3 



Volatilization 

NH3-(NH3),J 

Instalbility 



Excretion (Kzx'ZOON) 



PIN 



Kal — SS. — N H4 

Hal + pH 



Sorption/Desorption 

d„(PIN-p„*C-NH4 



" " 



Zooplankton 



Mortality 

(1-KPDN)*K -ZOON 



NH4 



Nitrogen 



Diffusion* Erosion/ 



Amonification 



Mortality 

KPDhTK 'ZOON 



SON 



Mortality 

(1-KPDN)*K as 'PHYN 



Uptake P m "u a *PHYN 



Excretion K./PHYN 
Nitrification 



Sorption/Desorptio 



,(PON-p 'C'SO 



on 



PON 



Mortality 

KPDN'K -PHYN 



Phytoplankton 



Kim ^?— AW4 

Hnit+DO 

♦ Diffusion 



N03 



Uptake 



(1-PJ'u-PHYN 



f.|iIhs S :-;:< 4 



•■ . .<■ "' .:, 



PIN 



Sorption/Desorption 



°,„(PIN-P„-C-NH * 
Instalbility 



NH4 



Amonification 



SON 



1 



Water Column 

fcrosioii/ ^Diffusion 



iorption/Desorptio 



„(PON-p "C 



>rptio n 

•SON! 



NH3 



Kal ^ AW 4 r 

Hal + pH 



PON 



Nitrification 

Knn - AW 4 

< < Hnil + DO 



N03 



♦ Diffusa 



"t"""*"""""" 



N03 



Denitrification 



Kiln " n ° 3 A-03 
Hno3+ DO 



N2 



Aerobic Layer 



Anaerobic Layer 



Figure E.l Nitrogen Cycle 

Ammonification is the biological process of formation of ammonium from soluble 
organic nitrogen. It is the first step of nitrogen mineralization, in which organic nitrogen is 
converted to the more mobile, inorganic state. The rate of ammonification is expressed as 
a first order reaction (Rao et al, 1984) : 



0, 



ammonification ** ONM ' " " ^ 



(E.l) 



where K om is the rate constant of ammonification which is a function of water temperature, 
pH, and the C/N ratio of the residue (Reddy and Patrick, 1984). 

The second step of mineralization of organic nitrogen is nitrification, which is an 



oxidation of ammonia to nitrate( NO^ ) directly or to nitrite ( NO: ) and then to nitrate: 



297 

nh: + 1 .5a -> no; +ih + + h.o 

(E.2) 

Nitrification requires oxygen as the electron acceptor. Therefore, nitrification is a 
strictly aerobic process, occurring only in the water column and in the aerobic layer of the 
sediment column. This process is related to a sink of dissolved oxygen in the system as 
shown in eq 4.11. The kinetics of nitrification are modeled as function of available 
ammonia, dissolved oxygen, and temperature. 

^nitrification ~ ~ ^ NN ~7l nTTi" 4 (E-3) 

M nit +DO 

where K NN is nitrification rate which is a function of temperature; and H ni , is the half- 
saturation constant for the bacteria growth. 

The dissolved form of ammonium in water is generally not stable and can exist in its 
gaseous form, or ammonia nitrogen. Because the ammonia concentration in the atmosphere 
is very low, ammonia in the water column can escape to the air. This is a volatilization 
process of ammonia. The volatilization of ammonia is a sink for nitrogen in an aquatic 
system. A first order rate equation can be used to describe the kinetics of the ammonia 
volatilization process (Chen and Sheng, 1994): 

&***. = Kvol [KNH 3 - NH°°* ] (E.4) 

where K V0L is rate constant of volatilization which is a function of temperature. It can 

be derived from the so-called two-film model (Jorgensen, 1983); h v is henry's constant; and NH"' m 

is the ammonia concentration in the air. 

Denitrification is defined as the biogeochemical transformation of nitrate nitrogen to 
gaseous end products such as molecular nitrogen or nitrous oxide (Reddy and Patrick, 1 984). 



298 
While nitrification occurs in the water column and aerobic layer of the sediment column, 

denitrification occurs only in the anaerobic layer of the sediment column. The denitrification 
process can be described by the standard Michaelis-Menten equation (Bowman and Focht, 
1974). 

jj 

U denitrification = ~^dn~ " ~ ~ "0 3 (E.5) 

H no3 + DO 

where K dn is the denitrification rate, which is a temperature function; and H n ^ is the half- 
saturation constant for denitrification. 

In the nitrogen species, sorption processes refer to conversion from a soluble to a 
solid phase of inorganic (NH4 to PIP) and organic (SON to PON) species, while desorption 
reactions describe the inverse process. Sorption/desorption processes, combined with 
resuspension events can significantly alter the nitrogen cycle in the system. The kinetics of 
sorption/desorption reactions are dependent on nitrogen species characteristics, sediment 
properties, pH, temperature, and dissolved oxygen concentration (Simon, 1989) The most 
commonly used mathematical representation of sorption/desorption processes is the linear, 
reversible, isotherm (Berkheiser et al., 1980; Reddy et al., 1988): 

J t N ad --D r -N ud+ S r -N s (E.6) 

where D r is the desorption rate constant which is a temperature function; S r is the sorption 
rate constant; N ad is the adsorbed nitrogen concentration such as particulate inorganic and 
organic nitrogen; N, is the dissolved nitrogen concentration such as dissolved ammonium 
nitrogen and dissolved organic nitrogen. 

The ratio between the desorption and sorption rates gives the partition coefficient 
dissolved and particulate forms, because dN ad /dt = at equilibrium. 



299 



S r _ < d _ 



o = Pc (E.7) 



D r N 
where A/^and AT" are the adsorbed and dissolved nitrogen, respectively, at the equilibrium 

condition; and /?,. is the partition coefficient. Therefore, the kinetic equation for 
sorption/desorption reaction is: 

J t N ad =-D r -(N ad -p c -N s ) (E.8) 

Inorganic nitrogen is incorporated by phytoplankton during growth and release as 
ammonium and organic nitrogen through respiration and non-predatory mortality. The 
phytoplankton nitrogen can be converted to zooplankton nitrogen by grazing process. The 
kinetic processes for particulate nitrogen species need to include settling process, which 
accounts for the limited vertical motion of particulate nitrogen. For this species, it is 
reasonable to assume the same settling velocity of the suspended sediment particles. 

The mass balance equations for nitrogen state variables are written by combining 
nitrogen transformation processes. 

Ammonia nitrogen (NH3) 

include ammonia conversion and volatilization processes 



For water column 

d ._„ v pH 



dt H al + pH 



h v -NH 3 -NH u 3 ""] (E.9) 



For sediment column: 



dt H al + pH 

where K al is the ammonia conversion rate constant which is a temperature function; H al is 

half-saturation constant for ammonia conversion. 
Dissolved ammonium nitrogen (NH4) 



300 
include phytoplankton uptake and respiration, zooplankton respiration, 
ammonification, nitrification, ammonia conversion, and sorption/desorption reaction. 
For water column: 

-NH 4 =-[(P n . Ma - K ax ) • PhyC - K u ■ ZooC] ■ A NC + K 0NM ■ SON 

-K m — NH 4 -K al P — NH 4 (E- 11 ) 

H nil+ DO 4 - H al+P H 4 

+d an (PIN-p an -c-NH 4 ) 
For sediment column: 

-NH 4 = + K ONM .SON-K NN .-^.NH 4 -K ar --^--.NH 4 

m H ni,+ D0 H al + pH ( E . 12 ) 

+d an (PIN- Pan -c-NH 4 ) 

where d an is sorption/desorption rate of NH4 from sediment particles; p an is the partition 
coefficient between NH4 and PIN; and c is the suspended sediment concentration. 

Nitrate and nitrite nitrogen (NQ3) 

include nitrification, denitrification and phytoplankton uptake processes. 

For water column: 

|-/V0 3 = +K NN ■ D ° ■ NH 4 - K DN ^ NO, 

dt H nil+ DO H lw3+ DO 3 (e. 13) 

-A NC ia-P n >M a 'PhyC] 

For sediment column: 

Soluble organic nitrogen (SON) 

include ammonification and sorption/desorption reaction. 

For water column: 

-SON = -K ONM -SON + d on (PON-p m -C'SON) (E.15) 

For sediment column: 



301 

^SON = -K 0NM ■ SON + d (m -(PON- Pon -c-SON) (E.16) 

where d (m is sorption/desorption rate of SON from sediment particles; andp on is the partition 
coefficient between SON and PON 

Particulate organic nitrogen (PON) 

include mortality of phytoplankton and zooplankton, settling, and a sorption- 

desorption reaction. 

For water column: 

-PON = A NC \K as -PhyC + K zs .ZooC]--ws p -PON 

-d on (PON- Pon -c-SON) 
For sediment column: 

— PON = ws • PON - d on ( PON - p on ■ c ■ SON ) (E. 18) 

dt dz 

where ws p is settling velocity for particulate species, which is same with that of suspended 

sediment particles. 

Particulate inorganic nitrogen (PIN) 

include settling and a sorption/desorption reaction. 

For both water and sediment columns: 

^PIN = -j-ws p -PIN-d an (PIN-p an -c-NH 4 ) (E.19) 

E.2 Phosphorous Cycle 

Phosphorous can be classified into two groups: dissolved phosphorous and particulate 
phosphorous. The criterion for this division is established in the laboratory, using a filtering 
technique. The dissolved phosphorous include soluble reactive phosphorous (SRP) and 
dissolved organic phosphorous (SOP). Particulate phosphorous includes particulate 
inorganic phosphorous (PIP), and particulate organic phosphorous (POP). Phytoplankton 



302 
phosphorous (PhyP) and zooplankton phosphorous (ZooP) related biomass to phosphorous 

concentration, through a fixed stoichiometric ratio: phosphorous-to-carbon ration (A PC ). 

The model phosphorous cycle (Figure E.2) includes the following processes: 

1) Mineralization of organic phosphorous 

2) Uptake of soluble reactive phosphorous by phytoplankton 

3) Conversion of phytoplankton phosphorous to zooplankton phosphorous by grazing 

4) Excretion and mortality by phytoplankton and zooplankton 

5) Settling for particulate phosphorous 

6) Sorption/desorption reactions 



Excretion (Kzx'ZOOP) 



PIP 



Sorption/Desorption „_„ 

<upip-p -c-srpT " HK 



Zooplankton 



Mortality 

(1-KPDP)*K 'ZOOP 



Mineralization 



Mortality 
KPDP'K -ZOOP 



SOP 



Mortality 

(1-KPDPCK 'PHYP 



Uptake P -u "PHYP 



Excretion K. /PHYP 



Sorption/Desorption 

V(pop-p -c-sopT 



POP 



Mortality 

KPDP'K "PHYP 



Phytoplankton 



Phosphorous 



Diffusion ♦ Erosion/ 



* Diffusion 



Diffusion * 



Water Column 

Erosion/ 4 Diffusion 



~ 



t deposition 



PIP 



Sorption/Desorption 



"dJPIP-p -C'SRPy 



_ 



deposition 



SRP 



Mineralization 



SOP 



Sorption/Desorption „_„ 
VJPOP-p-C'SOrf KUK 



♦ Diffusion 

T" *" 



p.p Sorption/Desorption 



♦ Diffusion Diffusion^ 

Mineralization 



SRP 



Aerobic Layer 
Diffusion^ 

"T ' 



SOP 



Sorption/Desorptio n 

a„(POP- P -c-sopT 



POP 



op v ^op 



Anaerobic Layer 

Figure E.2 Phosphorous Cycle 

The mineralization process is mediated by bacteria, which transfers dissolved organic 
phosphorous to soluble reactive phosphorous, through the uptake of SOP and excretion of 



303 
SRP. Since bacteria abundance is related to algae biomass, the rate of organic phosphorous 

mineralization is related to algae biomass. The mineralization of SOP is a relatively fast 

process, it can take a few hours, compared to the mineralization of carbon and nitrogen, 

which takes place in a few days (Golterman, 1973). Mineralization is highest when algae are 

strongly phosphorous limited and is lowest when no limitation occurs. Thompson et al. 

(1954) found that the mineralization of dissolved organic phosphorous is influenced by pH 

value. An increase in pH causes a temporary increase in the rate of mineralization of 

dissolved organic phosphorous. Temperature can also affect the speed of the mineralization 

by stimulating the mineralization process with high temperatures. 

The mineralization rate of SOP is usually modeled by a first order equation, as 

follows (Jorgensen, 1983) 

^mineralization ~ "■ opm ' ^U* (E.20) 

where Koptn is a rate constant for mineralization of SOP, which is a function of pH and 
temperature. 

Soluble reactive phosphorous is incorporated by phytoplankton, during growth and 
release, as soluble reactive phosphorous; and organic phosphorous through respiration and 
mortality. The phytoplankton phosphorous is converted to zooplankton phosphorous by 
grazing processes. The settling and sorption/desorption processes are similar to the nitrogen 
species. 

The mass balance equations for phosphorous state variables are written by combining 
these phosphorous transformation processes. 

Soluble reactive phosphorous (SRP) 

include mineralization, uptake by phytoplankton, mortality of zoo and phytoplankton, 



304 
and sorption/desorption reaction. 

For water column: 

T- SRP = K o P> „ ■ SOP + A pc [-ju a ■ PhyC + K^ ■ PhyC + K. x ■ ZooC] 
m (E.21) 

+d ip (PIP-p ip -c-SRP) 

For sediment column: 

J t SRP = K opm ■ SOP + d ip (PIP - Pip -c- SRP) (E.22) 

where d ip is sorption/desorption rate of SRP from sediment particles; and/?, p is the partition 
coefficient between SRP and PIP. 

Soluble organic phosphorous (SOP) 

include mineralization and sorption/desorption reaction. 
For both water and sediment columns: 

-SOP = -K opm -SOP-d^POP-p^c-SOP) (E.23) 

where d l)p is sorption/desorption rate of SOP from sediment particles; andp op is the partition 
coefficient between SOP and POP. 

Particulate organic phosphorous (POP) 

include respiration of zooplankton and phytoplankton, settling, and a 
sorption/desorption reaction. 



For water column: 

d d 

— POP = A PC ( K ux ■ PhyC + K^ -ZooC)- — ws p - POP 

-d op -(POP- Pop -c-SOP) 
For sediment column: 



(E.24) 



j t POP = - — ws p POP-d op \POP-p op -c-SOP) (E.25) 



305 
Particulate inorganic phosphorous (PIP) 

include settling and a sorption/desorption reaction 
For both water and sediment columns: 

-PIP = -—w Sp -PIP-d ip -(PIP- Pip .cSRP) (E.26) 



APPENDIX F 
SEDIMENT FLUX MODEL 

DiToro et al. (1990) developed a model of the SOD process in a mechanistic fashion 

using the square-root relationship of SOD to sediment oxygen carbon content. Using similar 

analysis as applied to carbon, they also evaluate the effect of nitrification on SOD. In this 

model, carbon and nitrogen diagenesis are assumed to occur at uniform rates in a 

homogeneous layer of the sediment of constant depth (active layer). The sediment oxygen 

demand and sediment fluxes are calculated by simple first-order decay processes of the 

concentrations of particulate organic carbonaceous material and of particulate organic 

nitrogenous material in this active layer as follow: 

dC 

poc 



= -k poc C poc +M c (F.l) 

at 



dC 



, K pon^ pon ~ ln N \ L --J 

where C P0C is a concentration of POC in sediment (g 2 /m 3 ) 
C P0N is a concentration of PON in sediment (g 2 /m 3 ) 
K P0C is a decay rate of POC in sediment (g 2 /m 3 ) 
K PON is a decay rate of POC in sediment (g 2 /m 3 ) 
M N is a source term for C P0C (g/m 3 -day) 
M c is a source term for C PON (g/m 3 -day) 
Sediment carbon and nitrogen diagenesis fluxes, J c and J N , are the most important 



306 



307 
parameters in the equations for sediment oxygen demand and sediment fluxes as defined by 

h = ~k poc C poc H 

i -b r h F 3) 

J N IC pon ( ^pon" 

where H is the depth of the active layer (m) 

Using these diagensis fluxes, DiToro and Pitzpatrick developed sediment flux model 
which include NSOD, CSOD, ammonia and nitrate flux, methane flux, and sulfide flux. 
F.l Sediment Flux Equations inCH3D-WQ3D 

The reactive portion of particulate organic carbon and particulate organic nitrogen 
in the sediment are presented by CBOD and NH4 in CH3D-WQ3D model system. The 
equation (F.3) can be rewritten as 

J c =-k gxy -Q™CBOD std .H 

( R4 ) 
J N =-K M <i,-°-NHA sed -H 

where k oxy is a oxidation rate in sediment (g 2 /m 3 ) 
k ni[ is a nitrification rate in sediment (g 2 /m 3 ) 

®l~y° is a temperature effect of oxidation 

@J t ' is a temperature effect of nitrification 

In this study, sediment fluxes include sediment oxygen demand, benthic dissolved 
methane flux, and benthic methane gas flux. Sediment oxygen demand (SOD) can be 
determined for the set of equations developed by DiToro and Fitzpatrick (1993). 



Carbonaceous sediment oxygen demand (g/m 2 /d): 



CSOD = J2K D C S J C 



f 



1-sec/z 



K r O, 



SOD 

Note: The square root term is replaced by J c ifJ c <2K D C, 

Nitrogenous sediment oxygen demand (g/m 2 /d): 



(F.5) 



308 



/ 



NSOD = 1.714 J> 



Total sediment oxygen demand (g/m 2 /d): 



1 - sec h 



V 



K N Q 2 
SOD 



(F.6) 



SOD = CSOD + NSOD (F.7) 

where 02 is a concentration of dissolved oxygen in overlying water column (g 02/m3). 

k D is a dissolved methane diffusion mass transfer coefficient (g 2 /m 3 ) 

k c is a reaction velocity for methane oxidation (g 2 /m 3 ) 

k N is a reaction velocity for ammonia oxidation (g 2 /m 3 ) 

C, is a methane solubility (g 2 /m 3 ) 

The magnitude of the fluxes at the sediment water interface of aqueous methane and 
gaseous methane are predicted to be 



•* CHA(aq) V *> S C 



sec/z 



K c 2 
SOD 



"'CH4(g)~''C V O 5 C 



(F.8) 



(F.9) 



The original DO kinetic equation in CH3D-WQ3D is modified to incorporate the 
sediment exchange predictions of the DiToro model, given by equation (F.10). 



dDO 
dt 



-Oxydation - Nitrification + reaeration + photosynthesis - respiration 



SOD JcH4(aq) ^ "* CH4(g) 



(F.10) 



H H J,u ' H 

where G frac is a fraction of gaseous methane produced in the sediment. 

His a water depth. 

The first five source and sink terms in the equation (F.10) are the same with the 
original DO kinetic equation, while the last three terms represent oxygen demand due to 
sediment exchange processes, where SOD and aqueous and gaseous methane flux rates. 



APPENDIX G 
MODEL PERFORMANCE TEST WITH PARALLEL CH3D-IMS 

To reduce the computational time necessary to simulate multiple-year seasonal 
response of the Charlotte Harbor estuarine system with a serial CH3D model, a parallel 
CH3D model was developed and validated. The parallel approach is applied to the parallel 
CH3D model via parallel constructs added to the serial CH3D model (Davis and Sheng, 
2000; Sheng et al., 2003). These parallel constructs are implemented by adding additional 
macros to the original serial source code. By defaults, the parallel CH3D model uses 
OpenMP constructs although due to the flexible nature of the implementation process, either 
Sun Microsystems-style or Clay-style constructs can be used. The parallel source code 
closely resembles the serial source code and maintains perfect compatibility with the serial 
code. In other words, the parallel CH3D model can be executed using the same input files 
as the serial CH3D model and parallel CH3D output format is identical to the output format 
of the serial CH3D model. To determine how well the parallel CH3D model performs in 
Charlotte Harbor estuarine system, the serial and parallel CH3D simulations performed for 
the parallel model validation. 
G.l Validation of Parallel Model 

To validate the parallel CH3D model, simulations are performed using both the serial 
and parallel CH3D model. CH3D model output from all of the parallel simulations are then 
compared to their respective serial simulations to fully validate the parallel CH3D model. 



309 



310 
The computational grid for Charlotte Harbor estuarine system contains 92 x 129 horizontal 

cells (Figure 5 . 1 ) and 8 vertical layers, with a total 1 1 648 grid cells which include 5367 water 

cells and 6281 land cells. Grid spacing varies from 40 to 2876 meters (average 598 meters). 

Using this grid, the 3-D hydrodynamic and water quality simulation simulate the circulation, 

sediment transport, and water quality dynamics of the Charlotte Harbor estuarine system 

from May 23 to June 22, 1996. The initial condition for simulation is provided by a 30-day 

spin up simulation (April 23 to May 23) previously performed during the dry season with all 

forcing mechanisms (tides, river discharges, wind) to allow water level, velocity and salinity 

field to reach dynamic steady-state throughout the computational domain. A more detailed 

description of the simulation can be found in Chapter 6. 

Simulated time series and field parameters at the end of 30-day simulation of each 
parallel simulation are compared to the all corresponding simulated parameters from the 
serial CH3D simulation which include water level, current velocity, temperature, salinity, 
sediment concentration, and all water quality species. For all comparisons, the simulated 
time series and field parameters obtained from the serial CH3d model are identical to the 
simulated parameters obtained from the parallel CH3D model. Thus the parallel CH3D 
model is completely validated. 
G.2 CPU Times of Parallel Routines 

The 1 -month simulation is simulated using both serial and parallel CH3D models on 
the SGI orgin system witch is Silicon Graphyics Origin 3400, 400 MHz MIPS R 12000 (IP 
35). This platform has 16 processors and 8 GB main memory size. The parallel model is 
executed using from 1 to 4 processors on this system. Table F.l report the CPU times per 
iteration for the parallel CH3D model executed on this platform. This timing results show 



311 
that the time necessary for CH3D simulation can be significantly reduced using multi- 
processor computers coupled with parallel techniques. 

Table G.l CPU time for the parallel, shared memory, CH3D procedures on SGI origin 
platform. Time shown are per time step iteration of the model using computational grid 
(92x 1 29) and are given in seconds, n is the number of processors used and speed up is shown 
in parenthesis. 



Serial 



n=1 



n=2 



n=3 



n=4 



Main WQ 


1.221(1.00) 


1.287(0.95) 


0.680(1.80) 


0.469(2.60) 


0.370(3.30) 


Main Sediment 


0.126(1.00) 


0.125(1.01) 


0.078(1.62) 


0.060(2.10) 


0.052(2.41) 


Turbulence 


0.100(1.00) 


0.107(0.93) 


0.059(1.70) 


0.045(2.21) 


0.038(2.61) 


Sediment transport 


0.090(1.00) 


0.089(1.01) 


0.047(1 .94) 


0.032(2.85) 


0.025(3.62) 


Dimensionalize 


0.075(1 .00) 


0.083(0.90) 


0.047(1.62) 


0.033(2.28) 


0.026(2.89) 


Baroclinic (J) 


0.071(1.00) 


0.075(0.95) 


0.042(1 .70) 


0.038(1.87) 


0.036(1.97) 


Baroclinic (I) 


0.070(1.00) 


0.067(1.05) 


0.038(1.86) 


0.025(2.75) 


0.020(3.49) 


Salinity 


0.055(1.00) 


0.059(0.95) 


0.030(1.85) 


0.020(2.51) 


0.016(3.43) 


N.L/Diffusion (J) 


0.055(1.00) 


0.055(1.00) 


0.033(1.65) 


0.030(1 .84) 


0.029(1.87) 


N.L/Diffusion (I) 


0.042(1 .00) 


0.043(0.98) 


0.025(1.65) 


0.018(2.34) 


0.015(2.82) 


Layer Vel. (v) 


0.040(1.00) 


0.041(0.97) 


0.027(1.48) 


0.021(1.86) 


0.019(2.12) 


Layer Vel. (u) 


0.028(1 .00) 


0.029(0.97) 


0.016(1.73) 


0.011(2.60) 


0.009(3.28) 


Layer Vel. (w) 


0.011(1.00) 


0.011(0.99) 


0.006(1.70) 


0.004(2.37) 


0.003(2.85) 


Interpolation 


0.011(1.00) 


0.011(1.00) 


0.006(1.79) 


0.005(2.62) 


0.004(3.52) 


integrate Vel. (U) 


0.008(1 .00) 


0.008(1.00) 


0.006(1.34) 


0.004(1.69) 


0.004(1.87) 


integrate Vel. (V) 


0.006(1.00) 


0.006(0.97) 


0.004(1.35) 


0.005(1 .32) 


0.005(1 .26) 


Bottom sheer stress 


0.001(1.00) 


0.001(1.01) 


0.001(1.89) 


0.000(2.76) 


0.000(3.60) 


Wave H/T 


0.001(1.00) 


0.001(1.03) 


0.000(1.95) 


0.000(2.93) 


0.000(3.87) 


All Parallel Routine 


1.973(1.00) 


2.060(1.00) 


1.152(1.71) 


0.843(2.34) 


0.701(2.82) 


Total Routine 


2.022(1 .00) 


2.107(1.00) 


1.204(1.68) 


0.895(2.26) 


0.753(2.69) 



G.3 Parallel Speedup 

The parallel speed up for the Charlotte Harbor 1 -month simulation described earlier 
on the SGI Origin platform is shown in Figure G.l. As the lines get close to the theoretical 
maximum, the parallel model is performing better. The speedups of the individual parallel 
procedures are shown as values inside parenthesis in the previous table G. 1 and also illustrate 



312 
how computationally intense routines have higher speedups. Applying parallel method for 
CH3D model achieved a 1.71x and 2.82x speedups using 2 processors and 4 processors, 
respectively. 




2 3 

Numbers of Processors 

Figure G.l Parallel speedup gained in performing the simulation on the SGI Origin 
platform 



APPENDIX H 
TIDAL BENCH MARKS FOR CHARLOTTE HARBOR 

Table H.l Tidal datum referred to Mean Low Low Water (MLLW), in meter. 

Station Station Latitude Longitude MHHW MHW MLW NAVD 

Number Name 



8725853 


Venice 


27 04.3 


82 27.3 


0.671 


0.589 


0.113 


0.496 


8725791 


Peace 
River 


26 59.3 


81 59.6 


0.615 


0.537 


0.119 


0.501 


8725781 


Shell 
Creek 


26 58.8 


81 57.6 


0.668 


0.570 


0.135 


0.472 


8725541 


Bokeelia 


26 42.4 


82 09.8 


0.526 


0.481 


0.072 


0.495 


8725520 


Fort Myers 


26 38.8 


81 52.3 


0.401 


0.335 


0.191 


0.318 


8725391 


Sanibel 


26 29.3 


82 00.8 


0.689 


0.614 


0.145 


0.570 


8725110 


Naples 


26 07.8 


81 48.7 


0.874 


0.797 


0.184 


0.696 



313 



REFERENCES 

Ambrose, R. B. Jr., Wool, T.A., Martin, J.L., Connolly, J.P., Schanz, R.W. (1994) WASP5, 
A hydrodynamic and water quality model-model theory, user's manual, and programmer's 
guide. U.S. EPA Environmental Research Laboratory. Athens, Georgia. 

American Public Health Association(APHA) (1985). Standard methods for the examination 
of water and wastewater. APHA. Washington, D.C. 

Banks, R.B. (1975). Some features of wind action on shallow lakes, ASCE, Journal of 
Environmental Engineering Division, Vol. 101, No. EE5, p8 13-827 

Banks, R.B. and Herrera, F.F. (1977). Effect of wind and rain on surface reaeration, ASCE, 
Journal of Environmental Engineering Division, Vol. 103, No. EE3, p489-504 

Baker, B. (1990). Draft Caloosahatchee water quality based effluent limitations 
documentation (Lee County). Florida Department of Environment Regulation. Water Quality 
Tech. Ser. 2. No. 121: 18. 

Berkheiser, V.E., Street, J.J., Rao, P.S.C., and Yuan, T.L. (1980) Partitioning of inorganic 
orthophosphate in soil-water systems. CRC Critical Review in Environmental Control. 

Berliand, M.E., and Berliand, T.G. (1952). Determining the net longwave radiation of the 
Earth with the consideration of the effect of cloudiness, Isv. Acad. Nauk., SSSR Ser. Geofis, 
No. 1. 

Boler, R.N., Malloy, R.C., and Lesnett, E.M. (1991) Surface water quality monitoring by the 
Environmental Protection Commission of Hillsborough County. In (S.F. Treat and Clark 
P.A., eds.) Proceedings of Tampa Bay Area Scientific Information Symposium 2, 1991. 
Tampa, FL. 

Bowie, G.L., Mills, W.B., Porcella, D.B., Campbell, C.L., Pagenkopf, J.R., Rupp, G.L., 
Johnson, K.M., Chan, P.W.H., Gherini, S.A. and Chamberlin, C.E. (1985). Rates, constants, 
and kinetics formulations in surface water quality modeling. In: (Editors), U.S. 
Environmental Protection Agency, Office of Research and Development, Athens, GA 

Bowie, G.L., Chen, C.W. and Dykstra, D.H. (1980) Lake Ontario ecological modeling, phase 
HI. Tectra-Tech, Inc., Lafayete, CA. 



314 



315 

Bowman, R.A. and Focht, D.D. (1974). The influence of glucose and nitrate concentrations 
upon denitrification rates in sandy soils. Soil Biol. Biochem. 6: 297-301. 

Bricaud, A., Morel, A., and Prieur, L. (1981). Absorption by dissolved organic matter of the 
sea (yellow substance) in the uv and visible domains. Limnology and Oceanography, 
28(5):43-53. 

Budyko, M. I. (1974). Climate and Life. Academic Press Inc., New York, 508pp. 

Camp, Dresser, & Mckee (1998). The Study of Seasonal and Spatial Patterns of Hypoxia in 
the Upper Charlotte Harbor. Prepared for S WFWMD, Prepared by Camp, Dresser, & Mckee 

Canale, R.P., DePalma, L.M., and Vogal, A.H. (1976). A plankton-based food web model 
for Lake Michigan, in (R.P. Canale, ed.) Modeling Biochemical Processes in Aquatic 
Ecosystem, Ann Arbor Science Publishers Inc., P.O. Box 1452, Ann Arbor, Michigan 48 160. 

Cerco, D.F., and Cole, T.M. (1994). Three-dimensional eutrophication model of Chesapeake 
Bay. Technical Report El-94-4, U.S. Army Engineer Waterways Experiment Station, 
Vicksburg, MS. 

Cerco, D.F., and Cole, T.M. (1995). User Guide to the CE-QUAL-ICM three-dimensional 
eutrophication model, Technical Report EL-95-15, U.S. Army Corps of Engineers, 
Washington, DC. 

Chamberlain, R.H. and Doering, P.H. (1998). Freshwater inflow to the Caloosahatchee 
Estuary and the resource-based method for evaluation. Proceedings of the Charlotte Harbor 
public conference and technical symposium, Technical Report No. 98-02: 81-90. 

Chapra, S.C. (1997). Surface water-quality modeling. The McGRAW-HILL companies, 
INC., New York. 

Chen, C.W., andOrlob, G.T. (1975). Ecologic simulation of aquatic environments. Systems 
Analysis and Simulation in Ecology, Vol. 3, B.C. Pattern (ed.). Academic Press, New York: 
475-588. 

Chen, X. and Sheng, Y.P.(1994). Effects of hydrodynamics and sediment transport processes 
on nutrient dynamics in shallow lakes and estuaries. Ph.D. dissertation. University of Florida, 
Gainesville, Florida 

Christian, D.J. and Sheng, Y.P.(2003). Modeling the Effects of Hydrodynamics, Suspended 
Sediments, and Water Quality on Light Attenuation in Indian River Lagoon, Florida. PhD 
dissertation, University of Florida., Gainesville, Florida. 

Clark, N. E., Eber R. M. Renner J. A., and Saur J. F. T. (1974) Heat exchange between ocean 
and atmosphere in the eastern North Pacific for 196-71. NOAA Tech. Rep. NMFS SSRF- 
682, U.S.Dept. Commerce, Washington. ,D.C. 



316 

Davis, J.R. and Sheng, Y.P. (2000). High performance estuarine and coastal environmental 
modeling: The CH3D example. In Estuarine and Coastal Modeling, Vol. 6, p470-486. 
American Society of Civil Engineers. 

Davis, J.R. and Sheng, Y.P. (2001). High Performance Modeling of Circulation and 
Transport in the Indian River Lagoon, Florida. PhD dissertation, University of Florida., 
Gainesville, Florida. 

Day, J.W. Jr., Hall, C.A.S, Kemp, K.M., and Arancibia, A.J. (1989). Estuarine Ecology. 
John Wiley & Sons. New York. 

DeCosmo, J., Katsaros, K.B., Smith, S.D., Anderson, R.J., Oost, W.A., Bumke, K., and 
Chadwick, H.M. (1996). Air-sea exchange of water vapor and sensible heat: The Humidity 
Exchange Over the Sea (HEXOS) results. J. Geophys. Res.-Oceans, 101: 12,001-12,016. 

Degrove, B. (1981). Caloosahatchee River wasterload allocation documentation. Florida 
Dept. Of Env. Reg., Water Quality Tech ser. 2. No. 52: 17. 

Degrove, B. and Nearhoof, F. (1987). Water quality assessment for Caloosahatchee River. 
Florida Dept. Of Env. Reg. Water Quality Tech. Ser. 3. No. 19. 

DiToro, D.M., Paquin, P.R., Subburamu,K., and Gruber, D.A. (1990). Sediment oxygen 
demand modehMethan and ammonia oxidation. Journal of Evironmental Engineering, 
116(5), p945-986. 

DiToro, D.M. and Fitzpatrick, J.J. (1993). Chesapeaka bay sediment flux model. U.S. Army 
Corps of Engineers, Waterways Experiment Station, Tech. Report EL-93-2. 

Dixon, L.K., and Gary, J.K. (1999). Cause of light attenuation with respect to seagrasses in 
upper and Lower Charlotte Harbor. Marine Laboratory Technical Report No. 650. Prepared 
for Southwest Florida Water Management District. 

Doering, P.H. (1996). Temporal variability of water quality in the St. Lucie Estuary, South 
Florida. Water Resource Bulletin 32:1293-1306. 

Echternacht, K.L. (1975). A study of the precipitation regimes of the Kissimmee River-Lake 
Okeechobee watershed. Florida Department of Environmental Regulation Technical Service 
(3)1. Tallahassee, Florida. 

Eckart, C. (1958). Properties of water. Part H The equation of state of water and sea water 
at low temperatures and pressures. American Journal of Science, 256(4): 224-240. 

Edwards, R.E., Lung, W., Montagna, P. A., and Windom, H.L. (2000). Final review report. 
Caloosahatchee Minimim Flow Peer Review Panel, September 27-29, 2000. Report to the 
south Florida Water Management District, West Palm Beach, Florida. 



317 

Environmental Quality Laboratory, Inc. (1979). Hydrobiological monitoring report to the 
Southwest Florida Water Management District, Port Charlotte, Florida. 

Environmental Quality Laboratory, Inc. (1987). Hydrobiological monitoring program, data 
report for the period from March 1986 through Feburaury 1987 covering the lower Peace 
River abd Charlotte Harbor. Port Charlotte, Florida, Environmental Quality Laboratory, Inc.: 
92. 

Eppley, R.W. (1972). Temperature and phytoplankton growth in the sea. Fish. Bull., 
70:1063-1085. 

Estevez, E. D. (1986). Infaunal macro invertebrates of the Charlotte Harbor estuarine system 
and surrounding inshore waters, Florida.. U.S. Geological Survey Water-Resources 
Investigations Report 85-4260: 116. 

Estevez, E.D. (1998). The story of the greater Charlotte Harbor Watershed, Charlotte Harbor 
Estuarine Program, North Fort Myers, Florida. 

Fraser, T.H. (1986). Long-term water-quality characteristics of Charlotte Harbor, Florida. 
U.S. Geological Survey Water-Resources Investigations Report 86-4180: 43. 

Fraser, T.H. and Wilcox, W.H. (1981). Enrichment of a subtropical estuary with nitrogen, 
phosphorous, and silica, in Neilson, B.J., and Cronin, L.E., eds., Estuarine and Nutrients, 
Humana Press, New Jersey: 481-498. 

Friehe, C. A. and Schmitt, K. B. (1976). Parameterization of air-sea interface fluxes of 
sensible heat and moisture by the bulk aerodynamic formulas. Journal of Physical 
Oceanography, 6, 801-809. 

Gallegos, C.L. (1993). Determination of optical water quality requirements in the Indian 
River near Ft. Pierce, FL, with emphasis on the impact of colored water discharges. 
Technical report, South Florida Water Management District. 

Gallegos, C.L. and Correl, D.L. (1990). Modeling spectral diffuse attenuation, absorption, 
and scattering coefficients in a turbid estuary. Limnology and Oceanography, 35(7): 1484- 
1502. 

Gallegos, C.L. and Kenworthy, W.J. (1996). Seagrass depth limits in the Indian River 
Lagoon (Florida, USA). Estuarine, Coastal, and Shelf Science, 42:267-288. 

Garret, J.R. (1977). Review of drag coefficients over oceans and continents. Monthly 
Weather Review 105:915-929. 

Golterman, H.L. (1973). Vertical Movement of Phosphorous in Freshwater, in (E.J. Griffith, 
A. Beeton, J.M Spencer, and D.T. Mitchell, eds.) Environmental Phosphorous Handbook, 
John Wiley & Sons., New York. 



318 

Goodwin, C.R. (1996). Simulation of tidal-flow, circulation, and flushing of the Charlotte 
Harbor estuarine system, Florida. U.S. Geological Survey Water-Resources Investigations 
Report 93-4153. 

Goodwin, C. R. and Michaelis, D.M. (1976). Tides in Tampa Bay, Florida: June 1971 to 
December 1973. U.S. Geological Survey Open-File Report FL-75004: 338. 

Hammet, K. M. (1990). Land use, water use, streamflow characteristics, and water-quality 
characteristics of the Charlotte Harbor inflow area, Florida. U.S. Geological Survey Water- 
Supply Paper 2350-A: 64. 

Hammet, K. M. (1992). Physical processes, salinity characteristics, and potential salinity 
changes due to freshwater withdrawals in the tidal Myakka River, Florida. U.S. Geological 
Survey Water-Resources Investigations Report 90-4054:20 

Heyl, M.G. (1997). Hypoxia in upper Charlotte Harbor. Proceedings of the Charlotte Harbor 
public conference and technical symposium, Technical Report No. 98-02: 219-228. 

Hwang, T.C. (1966). A Sedimentologic Study of Charlotte Harbor, Southwestern Florida. 
Master's thesis, Florida State University, Tallahassee, Florida. 

Jaworski, N.A. (1981). Source of Nutrients and the scale of eutrophication problems in 
estuaries, in Neilson, N.J., Jersey, Humana Press: p83-110. 

Jerlov, N.G. (1976). Marine Optics. Elsevier Scientific Publishing Company, New York. 

Jorgensen, S.E. (1976). A eutrophication model for a lake. Ecological Modeling 2: 147-165. 

Jorgensen, S.E. (1983). Modeling the ecological processes. In: (G.T. Orlob, ed.) 
Mathematical Modeling of Water Quality: Streams, Lakes, and Reservoirs. John Wiley and 
Sons. New York. Pp 116-149. 

Jorgensen, S.E., Sorensen, B.H., and Nielsen, S.N. (editors) (1996). Handbook of 
Environmental and Ecological Modeling. CRC Press, Inc. 

Jorgensen, S.E., and Gromiec, M.E. (editors) (1989). Mathematical sub-models in water 
quality systems. Development in Environmental Modeling, 14. Elsevier Science Publishers. 

Ketchem, B.H. (1967). Phytoplankton nutrients in estuaries, in Lauff, G. H., ed., Estuaries. 
National Oceanic and Atmospheric Administration 

Kirk, J.T.O. (1984). Dependence of relationship between inherent and apparent optical 
properties of water on solar altitude. Limnology and Oceanography, 29(2):350-356. 



319 

Lam, D.D.L., Schertzer, W.M., and Fraser, A.S. (1983). Simulation of Lake Erie water 
quality responses to loading and weather variations, I.W.D. Scientific Series, 134, 
Environment Canada, Ottawa:299. 

McEwan, J., Gabric, A.J. & Bell, P.R.F. (1998). Water quality and phytoplankton dynamics 
in Moreton Bay, Southeastern Queensland, n. Mathematical modeling. Marine and 
Freshwater Research, Vol. 40, No. 3, p215-225. 

McPherson, B. F. and Miller, R.L. (1990). Nutrient distribution and variability in the 
Charlotte Harbor estuarine system, Florida. Water Resource Bulletin 26: 67-80. 

McPherson, B. F., Miller, R. L. and Stoker, Y. E. (1996). Physical, Chemical, and Biological 
Characteristics of the Charlotte Harbor Basin and Estuarine System in Southwestern Florida- 
A Summary of the 1982-89 U.S. Geological Survey Charlotte Harbor Assessment and Other 
Studies. U.S. Geological Survey Water-Supply Paper 2486. 

Metha, A.J. (1986). Characterization of cohesive sediment properties and transport processes 
in estuaries. In:A.J. Metha (ed.), Estuarine Cohesive Sediment Dynamics. Springer, New 
York:290-325. 

Mehta, A.J. (1989). On estuarine cohesive sediment suspension behavior. Journal of 
Geophysical Research 94(C 10): 14303-14314. 

Mellor, G. L., (1996), Introduction to Physical Oceanography, Springer, Verlag New York. 

Miller R. L. and McPherson B. J., (1995). Modeling Photosynthetically Active Radiation in 
Water of Tampa Bay, Florida, with Emphasis on the Geometry of Incident Irradiance. 
Estuarine, Coastal and Shelf Science, 40, 359-377 . 

Morel, F. (1983). Principles of aquatic chemistry. John Wiley and Sons, New York, NY: 150. 

Morel, A., and Gentili, B. (1991). Diffuse reflectance of oceanic waters: Its dependence on 
sun angle as influenced by the molecular scattering contribution. Applied Optics 30:4427- 
2238. 

Morrison, G. (1997). Proposed trophic state goals and nitrogen management objectives for 
the tidal reaches of the Peace and Myakka rivers. Draft. Prepared by Southwest Florida Water 
Management District, Surface Water Improvement and Management (SWIM) Section. 

O'Connor, D.J. and Thomann, R.V. (1972). Water quality models: chemical, physical and 
biological constituents. In: Estuarine Modeling: an Assessment. EPA Water Pollution 
Control Research Series 16070 DZV, Section 702/71. 

Oliveria, J., Burnett, W.C., Mazzilli, B.P., Brags, E.S., Farias, L.A., Christoff, J., and 
Furtado, V. V. (2003). Reconnaissance of submarine groundwater discharge at ubatuba coast, 
Brazil, using rn-222 as a natural tracer, Journal of Radioact., 69:37-52. 



320 

Orth, R.J. and Moore, K.A. (1983). Chesapeake Bay: An unprecedented decline in 
submerged aquatic vegetation. Science, v. 222, No. 4619: 51-53. 

Phillips, N.A. (1957). A coordinate system having some special advantage for numerical 
forecasting. Journal of Meteor. 14:184-185. 

Pickard , G. L . and Emery, W. J. (1990). Descriptive physical Oceanography. Pergamon 
Press, New York. Fifth (SI) Enlarged Edition: 320. 

Pribble, J.R., Wade, D.L., Squires, A.D., and Janicki, A.J. (1998). A mechanic water quality 
model for the tidal Peace and Myakka rivers. Proceedings of the Charlotte Harbor public 
conference and technical symposium, Technical Report No. 98-02: 242-258. 

Prieur, L. and Sathyendranath, S. (1981). An optical classification of coastal and 
oceanographic waters based on the specific spectral absorption curves of phytoplankton 
pigments, dissolved organic matter, and other particulate materials. Limnology and 
Oceanography, 26(4):67 1-689. 

Rao P.S.C., Jessup, R.E., Reddy, K.R. (1984). Simulation of nitrogen dynamics in flooded 
soils. Soil Science 138:54-62. 

Reddy, K.R., Jessup, R.E., Rao, P.S.C. (1988). Nitrogen dynamics in a eutrophic lake 
sediment. Hydrobiologia 159:177-188. 

Reddy, K.R., and Patrick, W.H. (1984). Nitrogen transformations and loss in flooded soil and 
sediments. CRC Critical Reviews in Environmental Control 13(4):273-309. 

Redfield, A., Ketchman, B., and Richar, F. (1966). The influence of organisms on the 
composition of sea-water. The Sea VolII. Interscience Publishers, New York, 26-48. 

Reed, R. K. (1977). On estimating insolation over the ocean. Journal of Physical 
Oceanography, 7, 482-485. Research 49(3): 227-239. 

Rutkowski, C.W., Burnett, W.C., Iverson, R.L., and Chanton, J.P. (1999). The effect of 
groundwater seepage on nutrient delivery and seagrasss distribution in the northern Gulf of 
Mexico. Estuaries, 22:2033-1040. 

Schropp, S.J. (1998). Charlotte Harbor Sediment Quality Data Review and Evaluation. 
Proceedings of the Charlotte Harbor public conference and technical symposium, Technical 
Report No. 98-02: 35-47. 

Sheng, Y.P. (1983). Mathematical modeling of three-dimensional coastal currents and 
sediment dispersion: model development and application. Coastal Engineering Research 
Center Tech. Rep. CERC-TR-83-2. U.S. Army Corps of Engineers. 



321 

Sheng, Y.P. (1986). Modeling bottom boundary layers and cohesive sediment dynamics. 
In:(A.J. Mehta, ed.) Estuarine Cohesive Sediment Dynamics. Pp.360-400. Springer- Verlag. 
Berlin. 

Sheng, Y.P. (1987). On modeling three-dimensional estuarine and marine hydrodynamics. 
In Nihoul and Jamart, editors, Three-dimensional models of marine and estuarine dynamics, 
pages 35-54. Elswvier, Amsterdam. 

Sheng, Y.P. (1989). Evolution of a 3-D curvilinear grid hydrodynamic model: CH3D. 
In:(M.L. Spaulding, ed.) Proceedings of the Estuarine and Coastal Modeling Conference. 
ASCE, Newport, R.I., pp.40-49. 

Sheng, Y.P. (1994). Modeling hydrodynamics and water quality dynamics in shallow waters. 
In Proceedings of the first International Symposium on Ecology and Engineering, Taman 
Negara, Malaysia. University of Western Australia and Malaysian Technical University. 

Sheng, Y.P. (1995). On Modeling circulation in Florida Bay. In American Geophysical 
Union Conference on Circulation in Intra- American Seas, Pueto Rico. American Geophysical 
Union. 

Sheng, Y.P. (1998). Circulation in the Charlotte Harbor estuarine system. Proceedings of the 
Charlotte Harbor public conference and technical symposium, Technical Report No. 98-02: 
91-110. 

Sheng, Y.P. (2000). A framework for integrated modeling of coupled hydrodynamic- 
sedimentary-ecological processes. In Estuarine and Coastal Modeling, Volume 6, pages 350- 
362. American Society of Civil Engineers. 

Sheng, Y.P., Ahn, K.M., and Choi, J.K. (1990). Wind-wave hindcasting and estimation of 
bottom sheer stress in Lake Okeechobee. Technical Report UFL/COEL-93/02 1 , Coastal and 
Oceanographic Engineering Department, University of Florida, Gainesville, Florida. 

Sheng, Y.P., Chen, X.J., Schofield, S., and Yassuda, E. (1993). Hydrodynamics, sediment 
and phosphorous dynamics in Lake Okeechobee during an episodic event. Technical Report, 
Coastal and Oceanographic Engineering Department, University of Florida, Gainesville, 
Florida. 

Sheng, Y.P., and Chiu, S.S. (1986). Tropical cyclone generated currents. In Proceedings of 
the 20 th International Conference on Coastal Engineering, pages 737-751. American Society 
of Civil Engineers. 

Sheng, Y.P., Christian, D., and Kim, T.Y. (2002c). A IRL light attenuation model. Indian 
River Lagoon Pollutant Load Reduction (IRLPLR) Model Development Project. Final Report 
Volume II (A 3-D Water Quality Model) to St. Johns River Water Management District. 
Coastal and Oceanographic Engineering Department, University of Florida, Gainesville, 
Florida. 



322 

Sheng, Y.P., Davis, J.R., and Park, K. (2003) Development of parallel computing version of 
CH3D: Parallel CH3D User's Manual. Amendment to the three dimensional water quality 
model of the Charlotte Harbor estuarine system for the South Florida Water Management 
District. 

Sheng, Y.P., Davis, J.R., Sun, D.,Qiu, C, Park, K., Kim, T.Y., and Zhang, Y. (2001). 
Application of an integrated modeling system of estuarine and coastal ecosystem to Indian 
River Lagoon, Florida. In Estuarine and Coastal Modeling, Session 4B. American Society 
of Civil Engineers, pp. 329-343 

Sheng, Y.P., and Lick, W. (1979). Resuspension and transport of sediments in a shallow 
lake. Journal of Geophysical Research, V. 84, pp. 1809-1826. 

Sheng, Y.P. and Park, K. (2000). A 3-D Circulation and Salinity Model of Charlotte Harbor 
Estuarine System. Phase I Final Report to Southwest Florida Water Management Department 
(SFWMD), University of Florida, Gainesville, Florida. 

Sheng, Y.P. and Park, K. (2001). Modeling the Effects of Caloosahatchee Discharges on 
Circulation and Salinity in Charlotte Harbor. Interim Report Southwest Florida Water 
Management Department (SFWMD), University of Florida, Gainesville, Florida. 

Sheng, Y.P. and Park, K. (2002). 3-dimensional Water Quality model for Charlotte Harbor 
estuarine system. Phase HI Final Report Southwest Florida Water Management Department 
(SFWMD), University of Florida, Gainesville, Florida. 

Sheng, Y.P., Qiu, Q., Park, K., and Kim, T.Y. (2002b). A 3-D water quality model. Indian 
River Lagoon Pollutant Load Reduction (IRLPLR) Model Development Project. Final Report 
Volume II (A 3-D Water Quality Model) to St. Johns River Water Management District. 
Coastal and Oceanographic Engineering Department, University of Florida, Gainesville, 
Florida. 

Sheng, Y.P., Sun, D., and Zhang, Y. (2002a). A 3-D IRL suspended sediment transport 
model. Indian River Lagoon Pollutant Load Reduction (IRLPLR) Model Development 
Project. Final Report to St. Johns River Water Management District. Coastal and 
Oceanographic Engineering Department, University of Florida, Gainesville, Florida. 

Sheng, Y.P., and Villaret,C. (1989). Modeling the effect of suspended sediment stratification 
on bottom exchange processes. Journal of Geophysical Research - Oceans, 94(C10): 14229- 
14444. 

Simon, N.S. (1989). Nitrogen cycling between sediment and the shallow-water column in 
the transition zone of the Potomac River and Estuary. U. The role of wind-driven 
resuspension an adsorbed ammonium. Estuarine, Coastal and Shelf Science 28:531-547. 

Smith, R.C. and Baker, K. (1981). Optical properties of the clearest waters. Applied Optics, 
20(2): 177-184. 



323 

Smith, S. D. (1980). Wind stress and heat flux over the ocean in gale force winds. Journal 
of Physical Oceanography, 10, 709-726 

Smith, S. D. (1989). Coefficients for sea surface wind stress, heat flux, and wind profiles as 
a function of wind speed and temperature. Journal of Geophysics Research, 93, 15467-15472 

Steele, J. A. (1974). The structure of marine ecosystem. Oxford, Blackwell: p74-91. 

Steele, J.M. (1965) Notes on some theoretical problems in production ecology. In:(C.R. 
Goldman ed.) Primary Production in Aquatic Environments. University of California Press. 
Berkeley 

Steinberger, N., and Hondzo, M. (1999). Experimental study of the hydrodynamic control 
of diffusional mass transfer at the sediment-water interface. Journal of Environmental 
Engineering, 125(2): 192-200 

Stoker, Y.E. (1986). Water quality of the Charlotte Harbor estuarine system, Florida, 
November 1982 through October 1984. U.S. Geological Survey Open-File Report 85-563: 

213. 

Stoker, Y.E. (1992). Salinity distribution and variation with fresh-water inflow and tide, and 
potential changes in salinity due to altered freshwater inflow in the Charlotte Harbor 
estuarine system, Florida. U.S. Geological Survey Water-Resources Investigations Report 
92-4062: 30. 

Stull, R. B. (1988). An Introduction to Boundary Layer Meteorology. Kluwer Academic 
Publishers, Netherlands, 660pp. 

Sun, D. and Sheng, Y.P. (2001). Modeling suspended sediment transport under combined 
wave current interactions in Indian River Lagoon. PhD dissertation, University of Florida, 
Gainesville, Florida. 

Taylor, P., Gulev, S., Barnier, B.,Bradiey, F.,Charlock, T.,Gleckler, P.,Kubota, M., 
Kutsuwada, M.,Legler, D., Lindau, R., Rothrock, D., Schulz, J.da-Silva, A., Sterl, A., and 
White, G. (2000). Intercomparison and validation of ocean-atmosphere energy flux fields. 
Final Report of the joint WCRP/SCOR working group on air-sea fluxes, WCRP- 112, WMO/ 
TD-No. 1036, 303 S 

Thomann, R.V. (1992). Section 14.1. in Technical guidance manual for performing waste 
load allocations. Book III: Estuaries-Part 4, Critical review of coastal embayment and 
estuarine waste load allocation modeling. U.S. EPA Technical Report EPA-823-R-92-005. 
Washington, D.C. 

Thomann, R. V., and Fitzpatrick, J.F. (1982). Calculation and verification of a amthematical 
model of the eutrophication of the Potomac Estuary. Prepared for the Department of 
Environmental Services, Government of the Distric of Columbia. 



324 

Thompson, J.F. (1982). Numerical Grid Generation, chapter General Curvilinear Coordinate 
Systems, pages 1-30. Elsevier, Amsterdam. 

Thompson, J.F., Warsi, Z.U.A. and Mastin, C.W. (1985). Numerical Grid Generation: 
Foundations and Applications. North-Holland. New York 

Weast (1977). Handbook of chemistry and physics, 58 th ed., Table F-200 

Witkowski, J. and Jaffe, P.R. (1987). Kinetic and equlibrium formulations in modeling of 
hydrophobic organic chemicals in surface and ground waters. System Analysis in Water 
Quality Management (M.B. Beck, ed.), Advences in Water Pollution Control- Aseries of 
Conferences Sponsored by IAWPRC. 

Wolfe, S.H., and Drew R. D., eds. (1990). An ecological characterization of the Tampa Bay 
watershed. U.S. Fish Wildl. Serv. Biol. Rep. 90: 20. 

Yassuda,E.A. andSheng, Y.P. (1996). Integrated modeling of Tampa Bay Estuarine System. 
PhD dissertation, University of Florida, Gainesville, Florida. 

Zison, S., Mills, W.B., Deimer, D., and Chen, C.W. (1978). Rates, constants and kinetic 
formulations in surface water quality modeling. Prepared by Tetra Tech, Inc., Lafayette, CA 
for Environmental Research Laboratory, U.S. Environmental Protection Agency, Athen, GA. 
EPA-600/3-78-105:335. 



BIOGRAPHICAL SKETCH 
Kijin Park was born in Pusan, the largest harbor of Korea, on the 25 th of October. He 
received the Bachelor of Science degree and M.S. degree in department of Marine Science 
at the Pusan National University in February, 1991 and 1994, respectively. After finished 
M.S. degree, he had worked as research scientist at Korea Oceanography Research & 
Development Institute (KORDI) and Korea Meteorological Administration (KMA) until 
June, 1997. The following summer he left for Gainesville, Florida , and began his graduate 
studies in coastal and oceanographic engineering at the University of Florida. 



325 



I certify that I have read this study and that in my opinion it conforms to 
acceptable standards of scholarly presentation and is fully adequate, in scope and quality, 
as a dissertation for the degree of Doctor of Philosophy. 




Y. Pe^fSneng, Chair^ 

Professor of Civil and Coastal Engineering 

I certify that I have read this study and that in my opinion it conforms to 
acceptable standards of scholarly presentation and is fully adequate, in scope and quality, 
as a dissertation for the degree of Doctor of Philosophy. 

Robert G. Dean 

Graduate Research Professor of Civil and 
Coastal Engineering 

I certify that I have read this study and that in my opinion it conforms to 
acceptable standards of scholarly presentation and is fully adequate, in scope and quality, 
as a dissertation for the degree of Doctor of Philosophy. 




Robert J. Thiek£ 

Assistance Professor of Civil and Coastal 
Engineering 

I certify that I have read this study and that in my opinion it conforms to 
acceptable standards of scholarly presentation and is fully adequate, in scope and quality, 
as a dissertation for the degree of Doctor of PhilcrStoph\ 





Louis H. Motz 

Associate Professor of Civil and Coastal 
Engineering 



I certify that I have read this study and that in my opinion it conforms to 
acceptable standards of scholarly presentation and is fully adequate, in scope and quality, 
as a dissertation for the degree of Doctor of Philosopl; 




K. Ramesh Reddy 

Graduate Research Professor of Soil and 
Water Science 



This dissertation was submitted to the Graduate Faculty of the College of 
Engineering and to the Graduate School and was accepted as partial fulfillment of the 
requirements for the degree of Doctor of Philosophy. 



) Ajgw^l t/j^-t^-yv-cW^ 



August, 2004 

Pramod P. Khargonekar 
Dean, College of Engineering 



Kenneth J. Gerhardt 

Interim Dean, Graduate School 




,PA3&> 



UNIVERSITY OF FLORIDA 



3 1262 08554 5670