<%BANNER%>

Record for a UF thesis. Title & abstract won't display until thesis is accessible after 2009-12-31.

Permanent Link: http://ufdc.ufl.edu/UFE0021566/00001

Material Information

Title: Record for a UF thesis. Title & abstract won't display until thesis is accessible after 2009-12-31.
Physical Description: Book
Language: english
Creator: Kim, Taeyun
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2007

Subjects

Subjects / Keywords: Civil and Coastal Engineering -- Dissertations, Academic -- UF
Genre: Coastal and Oceanographic Engineering thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Statement of Responsibility: by Taeyun Kim.
Thesis: Thesis (Ph.D.)--University of Florida, 2007.
Local: Adviser: Sheng, Y. P.
Electronic Access: INACCESSIBLE UNTIL 2009-12-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2007
System ID: UFE0021566:00001

Permanent Link: http://ufdc.ufl.edu/UFE0021566/00001

Material Information

Title: Record for a UF thesis. Title & abstract won't display until thesis is accessible after 2009-12-31.
Physical Description: Book
Language: english
Creator: Kim, Taeyun
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2007

Subjects

Subjects / Keywords: Civil and Coastal Engineering -- Dissertations, Academic -- UF
Genre: Coastal and Oceanographic Engineering thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Statement of Responsibility: by Taeyun Kim.
Thesis: Thesis (Ph.D.)--University of Florida, 2007.
Local: Adviser: Sheng, Y. P.
Electronic Access: INACCESSIBLE UNTIL 2009-12-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2007
System ID: UFE0021566:00001


This item has the following downloads:


Full Text

PAGE 1

MODELING OF FLORIDAS ESTUARIES: UPPER CHARLOTTE HARBOR AND INDIAN RIVER LAGOON By TAEYUN KIM A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2007

PAGE 2

Taeyun Kim

PAGE 3

To the memory of my grandmothers: Sunlee Park on my fathers side and Kumhyo Lee on my mothers side.

PAGE 4

iv ACKNOWLEDGMENTS I would like to first thank m y advisor, Dr. Y. Peter Sheng, for his unlimited guidance and financial assist ance throughout my Ph.D. study. I would also like to thank the other committee members, Dr. Robert Dean, Dr. K. Ramesh Reddy, Dr. Kirk Hatfield, and Dr. Robert J. Thieke, for their review of my dissertation. In addition, I am greatly indebted to the faculty members in the coastal and oceanographic program for their support and encouragement. I am thankful to the Southw est Florida Water Management District and St. Johns River Water Management District for funding research projects, which provided me the opportunity to study integrated 3D hydr odynamic-ecological model for estuarine ecosystems. I am grateful to my colleagues, Jun Lee, Kijin Park, Yeonsik Chang, Yangfeng, Justin, Dave, Jeff, Vadim, Vladim ir, Bilge, Ma, Andrew, Detong, Chenxia, Sangdon, Jungwoo, and all students in the coastal program for their help and friendship. A big gratitude is owed to Sidney, Becky, Ki m, Helen, Tony, Nancy, Carol, Doretha, and some Korean students in our de partment for making life easier. Most importantly, none of this would ha ve been possible without the love and patience of my family. My father and mother have been a constant source of love, concern, support and strength all these years. I am grateful to my brother for his aide and support me throughout this endeavor. Finall y, I would like to express my heart-felt gratitude to my wife, Myoungju Jeoung, and parents-in-law.

PAGE 5

v TABLE OF CONTENTS page ACKNOWLEDGMENTS................................................................................................. iv LIST OF TABLES............................................................................................................. ix LIST OF FIGURES.......................................................................................................... xii ABSTRACT.....................................................................................................................xxi CHAP TER 1 INTRODUCTION AND OBJECTIVES...................................................................... 1 1.1 Introduction............................................................................................................. 1 1.2 Objectives...............................................................................................................4 2 LITERATURE REVIEW.............................................................................................6 2.1 Freshwater Inflow and Salinity............................................................................... 6 2.2 Multi-Phytoplankton and Silica Cycle.................................................................... 7 2.3 Sediment Oxygen Demand................................................................................... 13 3 CHARACTERIZATION OF STUDY AREAS......................................................... 17 3.1 Upper Charlotte Harbor Characterization............................................................. 17 3.2 Indian River Lagoon Characterization.................................................................. 22 4 MODEL DESCRIPTION...........................................................................................27 4.1 Hydrodynamic Model...........................................................................................27 4.2 Sediment Model.................................................................................................... 28 4.3 Water Quality Model............................................................................................ 28 4.3.1 Phytoplankton and Zooplankton Dynamics............................................... 29 4.3.2 Oxygen Balance..........................................................................................34 4.3.3 Nutrient Dynamics......................................................................................39 4.3.4 Water Quality Model Parameters............................................................... 44 4.4 Light Attenuation Model......................................................................................47

PAGE 6

vi 5 MODEL APPLICATION OF CIRCULATION AND TRANSPORT TO UPPER CHARLOTTE HARBOR ...........................................................................................49 5.1 A High-Resolution Curvilinear Grid.................................................................... 49 5.2 Measured Data......................................................................................................50 5.3 Initial and Boundary Conditions........................................................................... 54 5.4 Model Results....................................................................................................... 59 5.4.1 Water Level................................................................................................ 59 5.4.2 Currents......................................................................................................64 5.4.3 Salinity and Temperature...........................................................................66 5.4.4 Correlations between Simulated and Measured Parameters...................... 73 5.4.5 Spectral Analysis........................................................................................ 77 5.5 Establishment of Freshwater Flows and Levels................................................... 81 6 WATER QUALITY MODEL IN THE UPPER CHARLOTTE HARBOR.............. 97 6.1 Forcing Mechanism and Boundary Conditions of Circulation and Transport..... 98 6.2 Measured Data....................................................................................................100 6.3 Boundary Conditions for Water Quality Model................................................. 102 6.3.1 Pollutant Loads from Freshwater Discharge............................................ 102 6.3.2 Temperature..............................................................................................103 6.3.3 Light and Color.........................................................................................103 6.3.4 Sediment Type and Nutrient Distribution................................................104 6.4 Model Parameters and Calibration..................................................................... 104 6.5 Water Quality Model Simulation........................................................................ 113 6.6 Sediment Oxygen Demand and Pollutant Load Reduction................................ 125 6.7 Sensitivity Test for Water Quality Model.......................................................... 137 7 MODEL APPLICATION OF WATER QUALI TY TO INDIAN RIVER LAGOON................................................................................................................. 141 7.1 Basic Conditions................................................................................................. 141 7.1.1 Pollutant Loads from Freshwater Discharge............................................ 142 7.1.2 Temperature..............................................................................................144 7.1.3 Light and Color.........................................................................................144 7.1.4 Sediment Type and Nutrient Distribution................................................144 7.1.5 Open Boundary Condition........................................................................148 7.2 Measured Data....................................................................................................148 7.2.1 WQMN Data............................................................................................ 148 7.2.2 FASUF Data.............................................................................................150 7.2.3 UF Episodic Data..................................................................................... 152 7.2.4 UF Synoptic Data.....................................................................................152 7.3 Model Parameters and Calibration..................................................................... 154 7.4 Water Quality Model Simulation........................................................................ 163 7.4.1 Long-Term Simulation............................................................................. 163 7.4.2 Episodic Simulation.................................................................................. 180 7.4.3 Synoptic Simulation................................................................................. 186

PAGE 7

vii 7.4.4 The ROC Scores of Water Quality Model Simulation............................. 194 7.5 Sensitivity Test for IRL Water Quality Model................................................... 194 7.5.1 Sensitivity Test for Water Quality Nutrients............................................ 194 7.5.2 Sensitivity Test for Sediment...................................................................200 8 CONCLUSION AND DISCUSSION...................................................................... 204 APPENDIX A A THREE-DIMENSIONAL CURVILINEAR-GRID HYDRODYNAMIC MODEL ....................................................................................................................212 B SEDIMENT MODEL...............................................................................................215 C DISSOLVED OXYGEN SATURA TION AND REAERATI ON........................... 219 D A NAIVE STREETER-PHELPS SOD M ODEL...............................................221 E NUTRIENT DYNAMICS........................................................................................ 223 E.1 Nitrogen Cycle...................................................................................................223 E.2 Phosphorous Cycle.............................................................................................226 F LIGHT ATTENUATION COEFFICIENT..............................................................229 G HEAT FLUX............................................................................................................232 G.1 Short-Wave Solar Radiation.............................................................................. 232 G.2 Long-Wave Solar Radiation..............................................................................233 G.3 Latent Heat Flux................................................................................................ 233 G.4 Sensible Heat Flux.............................................................................................234 H TEMPERATURE, LIGHT, WIND AND RIVE R NUTRIENT LOADS............... 235 I BOTTOM SHEAR STRESS IN THE INDIAN RIVER LAGOON ........................ 243 J TAYLOR DIAGRAM.............................................................................................. 252 K WATER QUALITY PARAMETER ESTIMATION...............................................257 K.1 Modified Gauss-Newton Method......................................................................257 K.2 Examples of Modified Gauss-Newton Method.................................................259 K.2.1 Homogeneous Gas Phase Reaction.........................................................259 K.2.2 Toluene Hydrogenation........................................................................... 261 K.2.3 Methyl Ester Hydrogenation...................................................................263 K.3 Parameter Estimation of Water Quality Kinetic Equations............................... 265 K.4 Application of Estimated Parameters................................................................ 272 L KINETIC TERM EFFECTS IN THE WATER QUALITY MODEL..................... 278

PAGE 8

viii M NEAR BOTTOM DISSOLVED OX YGEN CONCENTRATION IN T HE UPPER CHARLOTTE HARBOR........................................................................... 285 N CORRELATION OF SALINITY DIFFERENCES BETWEEN BOTTOM AND SURFACE LAYERS ................................................................................................ 287 O TEMPORAL SOD VALUES IN TERM S OF DIFFERENT SCENARIOS............ 289 LIST OF REFERENCES.................................................................................................291 BIOGRAPHICAL SKETCH...........................................................................................305

PAGE 9

ix LIST OF TABLES Table page 2-1 Comparison of multi-phytoplankton models............................................................ 12 2-2 Sediment type........................................................................................................... 15 2-3 Notable features of SOD model............................................................................... 16 4-1 Algal maximum growth rates................................................................................... 31 4-2 Half saturation constants for each limiting nutrient................................................. 32 4-3 Settling velocity, respiration, a nd non-predatory mortality for each phytoplankton ...........................................................................................................33 4-4 Water quality model parameters.............................................................................. 44 5-1 Dates when conductivity/temperature sensors were cleaned...................................53 5-2 Time periods with good field data............................................................................ 54 5-3 Description of wind stations..................................................................................... 55 5-4 Description of river stations..................................................................................... 55 5-5 Description of precipitati on and evaporation stations ..............................................57 5-6 Comparison of phase errors in terms of m easured and predicted water level at UF station.................................................................................................................62 5-7 Comparison of amplitude errors tida l com ponents in terms of measured and predicted water level at UF station........................................................................... 62 5-8 Comparison of phase errors in terms of m easured and predicted water level at Punta Gorda station..................................................................................................62 5-9 Comparison of amplitude errors in term s of m easured and predicted water level at Punta Gorda station.............................................................................................. 62 5-10 Comparison of phase errors in terms of m easured and predicted water level at El Jobean station...........................................................................................................63

PAGE 10

x 5-11 Comparison of amplitude errors in term s of m easured and predicted water level at El Jobean station...................................................................................................63 5-12 Comparison of phase errors in terms of m easured and predicted water level at Harbor Height station...............................................................................................63 5-13 Comparison of amplitude errors in term s of m easured and predicted water level at Harbor Height station........................................................................................... 63 6-1 Description of water quality m easured stations..................................................... 101 6-2 Description of sensitivity tests............................................................................... 106 6-3 Sensitivity analysis results..................................................................................... 108 6-4 Water quality model coefficients used for the Charlotte Harbor sim ulation......... 110 6-5 Relative error of each station................................................................................. 122 6-6 Sediment Oxygen Demand in the Upper Charlotte Harbor (CDW, 1998)............ 125 6-7 Description of sensitivity tests............................................................................... 140 6-8 The results of sensitivity tests. Va lu es shown are the ROC score of each nutrient. Values shown in italics indicate the difference between the baseline and test simulations with a positive value i ndicating improvement of test simulation and a negative value indicating dete rioration of test simulation............................ 140 7-1 Description of te m perature stations........................................................................ 144 7-2 Description of WQMN stations.............................................................................. 149 7-3 Description of FASUF stations.............................................................................. 152 7-4 Description of episodic stations............................................................................. 152 7-5 Description of synoptic stations............................................................................. 153 7-6 Description of sensitivity tests............................................................................... 156 7-7 Sensitivity analysis results..................................................................................... 159 7-8 Water quality model coefficients used for the IRL simulation.............................. 161 7-9 Relative error of each station................................................................................. 165 7-10 Relative errors of FASUF stations.........................................................................180 7-11 Relative errors of episodic stations........................................................................181

PAGE 11

xi 7-12 Relative error of each segment for synoptic simulation......................................... 187 7-13 ROC scores of IRL water quality model simulation.............................................. 194 7-14 Description of sensitivity tests............................................................................... 195 7-15 The results of sensitivity tests. Va lu es shown are the ROC score of each nutrient. Values shown in italics indicate the difference between the baseline and test simulations with a positive value i ndicating improvement of test simulation and a negative value indicating dete rioration of test simulation............................ 199 7-16 Description of sedime nt sensitivity tests ................................................................ 201 7-17 The results of sensitivity tests. Va lu es shown are the ROC score of each nutrient. Values shown in italics indicate the difference between the baseline and test simulations with a positive value i ndicating improvement of test simulation and a negative value indicating dete rioration of test simulation............................ 203 8-1 Main conclusions from this study..........................................................................210 D-1 SOD in g m-2d-1 calculated with the "naive" Streeter-Phelps SOD model.............222 F-1 Spectrum of incident sunlight data......................................................................... 231 K-1 Data for the homogeneous gas phase reaction....................................................... 260 K-2 Data for the hydrogenation of Toluene.................................................................. 262 K-3 Data for the hydrogenation of methylesters........................................................... 263 K-4 Definition of symbols in kinetic equations ............................................................ 268 K-5 Values of initial unknown parameters and estimated parameters.......................... 269 K-6 Description of episodic stations............................................................................. 272 K-7 Values of parameters from modified Gauss-Newton m ethod and trial-and-error method....................................................................................................................275

PAGE 12

xii LIST OF FIGURES Figure page 2-1 Model Silica Cycle................................................................................................... 11 2-2 Schematic diagram of the sulfide oxidation............................................................. 13 3-1 Drainage Basins........................................................................................................ 18 3-2 Map of Charlotte Harbor Estuarine System............................................................. 19 3-3 Northern Charlotte Harbor seagrass stations............................................................ 21 3-4 Map of Indian River Lagoon....................................................................................23 4-1 Temperature growth curves for major algal groups................................................. 31 4-2 Dissolved oxygen and..............................................................................................36 4-3 A "naive" Streeter-P helps m odel of SOD................................................................ 37 4-4 Schematic of Nitrogen Cycle................................................................................... 40 4-5 Schematic of Phosphorous Cycle............................................................................. 41 4-6 Schematic of Silica Cycle........................................................................................ 42 5-1 Boundary-fitted grid (188 by 176) used for the study.............................................. 50 5-2 Bathymetry in boundary-fitted grid..........................................................................51 5-3 Locations of UF, Punta Gorda, Harbor Height, El Jobean, HBMP and Mote Marine recorders ...................................................................................................... 52 5-4 Wind speed at UF, Venice and Naples stations....................................................... 56 5-5 River discharge at the Peace River, S hell Creek, Myakka River, and Caloosahatchee River...............................................................................................56 5-6 Precipitation data at 211, 420, 421, 502, 505 and 537 stations................................57 5-7 Evaporation at Naples stations.................................................................................58 5-8 Air temperatures at UF, Venice and Naples stations ............................................... 59 5-9 Comparison of measured and simula ted water levels at UF station ........................ 60 5-10 Comparison of measured and simulated water levels at Punta Gorda station ......... 60

PAGE 13

xiii 5-11 Comparison of measured and simulated water levels at El Jobean station .............. 61 5-12 Comparison of measured and simulated wa ter levels at Harbor Height station ...... 61 5-13 Lower current velocity directed by East/West at UF station................................... 64 5-14 Upper current velocity directed by East/West at UF station.................................... 65 5-15 Lower current velocity directed by N orth/South at UF station................................65 5-16 Upper current velocity directed by N orth/South at UF station................................66 5-17 Comparison between measured and si m ulated salinity at UF station...................... 67 5-18 Comparison between measured and si m ulated temperature at UF station............... 68 5-19 Comparison between measured and simula ted salinity at El Jobean station ........... 69 5-20 Comparison between measured and simula ted temperature at El Jobean station .... 69 5-21 Comparison between measured and simula ted salinity at Punta Gorda station .......70 5-22 Comparison between measured and si m ulated temperature at Punta Gorda station.......................................................................................................................70 5-23 Comparison between measured and simu lated salinity at (13, 34) m easured by Mote Marine Laboratory..........................................................................................71 5-24 Comparison between measured and simu lated salinity at (19, 43) m easured by Mote Marine Laboratory..........................................................................................71 5-25 Comparison between measured and simu lated salinity at (19, 53) m easured by Mote Marine Laboratory..........................................................................................72 5-26 Comparison between measured and simu lated salinity at (22, 47) m easured by Mote Marine Laboratory..........................................................................................72 5-27 Comparison between measured a nd sim ulated salinity at HBMP........................... 73 5-28 Correlation between measured and simula ted water elevation at UF, El Jobean and Punta Gorda stations ..........................................................................................74 5-29 Correlation between measured and si m ulated salinity at UF station....................... 75 5-30 Correlation between measured and sim ulated temperature at UF station................75 5-31 Correlation between measured and simula ted salinity at El Jobean and Punta Gorda stations ...........................................................................................................76

PAGE 14

xiv 5-32 Correlation between measured and simulated temperature at El Jobean and Punta Gorda stations .................................................................................................76 5-33 Power spectrum density of the simula ted (left hand side) and m easured (right hand side) water elevation at UF station.................................................................. 77 5-34 Power spectrum density of the simula ted (left hand side) and m easured (right hand side) water elevation at El Jobean station........................................................ 78 5-35 Power spectrum density of the simula ted (left hand side) and m easured (right hand side) water elevation at Punta Gorda station................................................... 78 5-36 Power spectrum density of the simula ted (left hand side) and m easured (right hand side) surface EAST/WEST current at UF station............................................ 79 5-37 Power spectrum density of the simula ted (left hand side) and m easured (right hand side) bottom EAST/WEST current at UF station............................................ 79 5-38 Power spectrum density of the simula ted (left hand side) and m easured (right hand side) surface NORTH/SOUTH current at UF station...................................... 80 5-39 Power spectrum density of the simula ted (left hand side) and m easured (right hand side) bottom NORTH/SOUTH current at UF station...................................... 80 5-40 Plainview of vertical -longitudinal profiles ............................................................... 82 5-41 Vertical-longitudinal salinity profiles along the axis of the Myakka River during high runoff season in 2003 .......................................................................................83 5-42 Vertical-longitudinal salinity profiles along the axis of the Myakka River during low runoff s eason in 2003........................................................................................ 84 5-43 Vertical-longitudinal salinity profiles along the axis of the Peace River during high runoff season in 2003 .......................................................................................85 5-44 Vertical-longitudinal salinity profile s along th e axis of the Peace River during low runoff season in 2003........................................................................................ 86 5-45 Time histories of Myakka river di scharge and the locations of 1, 10, 20 PSU surface salin ity along the Myakka River.................................................................. 87 5-46 Time histories of Peace river disc harge and the location s of 1, 10, 20 PSU surface salinity along the Peace River...................................................................... 88 5-47 Locations where average surface salinity is 5, 10, 20 PSU for 30 days along the Myakka River...........................................................................................................89 5-48 Locations where average surface salinity is 5, 10, 20 PSU for 30 days along the Peace River............................................................................................................... 90

PAGE 15

xv 5-49 Locations of 10 PSU surface salinity al ong the Myakka River with respect to Myakka river discharge ............................................................................................91 5-50 Locations of 10 PSU surface salinity along the Peace River with respect to Peace River discharge.............................................................................................. 92 5-51 Relationship between locations at spec ific salinity values vs. Myakka River discharge ...................................................................................................................94 5-52 Relationship between locations at specific salinity values vs. Peace River discharge ...................................................................................................................95 5-53 Relationship between locations at sa linity 6 PSU and 12 PSU vs. Peace Riv er discharge...................................................................................................................96 6-1 Segments for Charlotte Harbor estuarine system .....................................................98 6-2 Air temperatures at Venice, Fo rt Myers, and Naples stations ..................................99 6-3 Wind speed at Venice, Fort Myers, and Naples stations ........................................100 6-4 Locations of 2000 water quality meas ured stations operated by SFW MD and SWFWMD............................................................................................................. 102 6-5 Temporal water quality vari ations at CH002 station in 2000 ................................ 114 6-6 Temporal water quality vari ations at CH005 station in 2000 ................................ 114 6-7 Temporal water quality vari ations at CH006 station in 2000 ................................ 115 6-8 Temporal water quality vari ations at CH007 station in 2000 ................................ 115 6-9 Temporal water quality vari ations at CH008 station in 2000 ................................ 116 6-10 Temporal water quality vari ations at CH009 station in 2000 ................................ 116 6-11 Temporal water quality vari ations at CH011 station in 2000 ................................ 117 6-12 Temporal water quality vari ations at CH013 station in 2000 ................................ 117 6-13 River discharge and DO at CH005 and CH006 in 2000........................................118 6-14 Correlation between measured/simulat ed DO at the bottom layer and Peace River discharge at CH005 and CH006 stations...................................................... 119 6-15 Correlation between measured/simulated DO at the bottom layer and simulated SOD at CH005 and CH006 stations .......................................................................120

PAGE 16

xvi 6-16 Correlation between measured and si m ulated DO at the bottom layer at CH005 and CH006 stations................................................................................................121 6-17 ROC curve for Upper Charlotte Harbor nutrients.................................................. 124 6-18 Temporal SOD variations at SOD #1 and SOD #2 stations in 2000 ...................... 127 6-19 Temporal water quality vari ations at CH002 station in 2000 ................................ 128 6-20 Temporal water quality vari ations at CH005 station in 2000 ................................ 129 6-21 Temporal water quality vari ations at CH006 station in 2000 ................................ 130 6-22 Temporal water quality vari ations at CH007 station in 2000 ................................ 131 6-23 Temporal water quality vari ations at CH008 station in 2000 ................................ 132 6-24 Temporal water quality vari ations at CH009 station in 2000 ................................ 133 6-25 Temporal water quality vari ations at CH011 station in 2000 ................................ 134 6-26 Temporal water quality vari ations at CH013 station in 2000 ................................ 135 6-27 Temporal SOD values at CH005 and CH009 in 2000........................................... 136 6-28 Temporal DO values at CH005 and CH009 in 2000............................................. 137 7-1 Boundary-fitted grid (199 by 23) and bathymetry................................................. 142 7-2 IRL segment definition...........................................................................................143 7-3 Distribution of sediment size.................................................................................. 145 7-4 Location of sediment sa mpling stations: Northern st ations are 13 to 24; Mud stations are 8, 9 and 11 in Melbourne area ; Southern stations are 1 to 7, 10 and 12 ............................................................................................................................147 7-5 Location of WQMN stations.................................................................................. 150 7-6 Location of FASUF sites........................................................................................ 151 7-7 Location of UF episodic and synoptic stations ...................................................... 154 7-8 ROC curve for IRL nutrients from 1997 to 1999................................................... 167 7-9 Temporal water quality variations at IRLV17 station from 1997 to 1999............. 169 7-10 Temporal water quality variations at IRLI07 station from 1997 to 1999..............170

PAGE 17

xvii 7-11 Temporal water quality variations at IRLB09 station from 1997 to 1999............. 171 7-12 Temporal water quality variations at IRLI18 station from 1997 to 1999..............172 7-13 Temporal water quality variations at IRLI23 station from 1997 to 1999..............173 7-14 Temporal water quality variations at IRLIRJ01 station from 1997 to 1999.......... 174 7-15 Temporal water quality variations at IRLIRJ05 station from 1997 to 1999.......... 175 7-16 Temporal water quality variations at FASUF1 station from 1997 to 1999............ 176 7-17 Temporal water quality variations at FASUF2 station from 1997 to 1999............ 176 7-18 Temporal water quality variations at FASUF3 station from 1997 to 1999............ 177 7-19 Temporal water quality variations at FASUF4 station from 1997 to 1999............ 177 7-20 Temporal water quality variations at FASUF5 station from 1997 to 1999............ 178 7-21 Temporal water quality variations at FASUF6 station from 1997 to 1999............ 178 7-22 Temporal water quality variations at FASUF7 station from 1997 to 1999............ 179 7-23 Temporal water quality variations at FASUF8 station from 1997 to 1999............ 179 7-24 Wind vector at stations for episodic #1 & #2......................................................... 182 7-25 Temporal water quality va riations at nor th station ................................................. 183 7-26 Temporal water quality vari ations at central station .............................................. 184 7-27 Temporal water quality va riations at south station ................................................ 185 7-28 Temporal water quality varia tions at Titusville station .......................................... 186 7-29 Temporal water quality varia tions at Synoptic #1 station ...................................... 188 7-30 Temporal water quality varia tions at Synoptic #2 station ...................................... 189 7-31 Temporal water quality varia tions at Synoptic #7 station ...................................... 190 7-32 Temporal water quality varia tions at Synoptic #13 station .................................... 191 7-33 Temporal water quality varia tions at Synoptic #26 station .................................... 192 7-34 Temporal water quality varia tions at Synoptic #28 station .................................... 193 H-1 The surface water temperature, surfa ce ligh t, and wind at IRLV17. Total daily nitrogen and phosphorous loads from rivers in segment 1..................................... 236

PAGE 18

xviii H-2 The surface water temperature, surfa ce ligh t, and wind at IRLI07. Total daily nitrogen and phosphorous loads from rivers in segment 2..................................... 237 H-3 The surface water temperature, surfa ce ligh t, and wind at IRLB09. Total daily nitrogen and phosphorous loads from rivers in segment 3..................................... 238 H-4 The surface water temperature, surfa ce ligh t, and wind at IRLI18. Total daily nitrogen and phosphorous loads from rivers in segment 4..................................... 239 H-5 The surface water temperature, surfa ce ligh t, and wind at IRLI23. Total daily nitrogen and phosphorous loads from rivers in segment 5..................................... 240 H-6 The surface water temperature, surfa ce ligh t, and wind at IRLIRJ01. Total daily nitrogen and phosphorous loads from rivers in segment 6..................................... 241 H-7 The surface water temperature, surfa ce ligh t, and wind at IRLIRJ05. Total daily nitrogen and phosphorous loads from rivers in segment 7..................................... 242 I-1 The simulated bottom shear stress at IRLV17....................................................... 244 I-2 The simulated bottom shear stress at IRLI07.........................................................245 I-3 The simulated bottom shear stress at IRLB09....................................................... 246 I-4 The simulated bottom shear stress at IRLI18.........................................................247 I-5 The simulated bottom shear stress at IRLI23.........................................................248 I-6 The simulated bottom shear stress at IRLIRJ01..................................................... 249 I-7 The simulated bottom shear stress at IRLIRJ05..................................................... 250 I-8 The percentage of simulated bottom sh ear stress exceeded 1 dyne/cm 2 over the simulation period in the IRL.................................................................................. 251 J-1 Taylor diagram displaying a statistic al com parison in te rms of phytoplankton.... 252 J-2 Taylor diagram displaying a statistical com parison in terms of dissolved oxygen. 253 J-3 Taylor diagram displaying a statis tical com parison in terms of TOC.................... 253 J-4 Taylor diagram displaying a sta tis tical comparison in terms of NO3....................254 J-5 Taylor diagram displaying a statistical com parison in terms of dissolved TKN...254 J-6 Taylor diagram displaying a statis tical com parison in terms of SRP.................... 255 J-7 Taylor diagram displaying a statistic al com parison in te rms of particulate nitrogen...................................................................................................................255

PAGE 19

xix J-8 Taylor diagram displaying a statistic al com parison in te rms of particulate phosphorous...........................................................................................................256 J-9 Taylor diagram displaying a statistical com parison in terms of dissolved silica... 256 K-1 Homogeneous gas phase reaction.......................................................................... 260 K-2 Toluene hydrogenation........................................................................................... 262 K-3 Methyl ester hydrogenation.................................................................................... 264 K-4 Temporal CBOD and ph ytoplankton variation ...................................................... 270 K-5 Temporal NH4 and SRP variation.......................................................................... 270 K-6 Temporal CBOD and dissolved nitrogen variation................................................ 271 K-7 Temporal particulate nitr ogen and phosphorous variation ..................................... 271 K-8 Comparison of measured and simula ted (using water quality param eters obtained by modified Gauss-Newton method) water qual ity state variables at north station............................................................................................................273 K-9 Comparison of measured and simula ted (using water quality param eters obtained by modified Gauss-Newton method) water qual ity state variables at central station.........................................................................................................273 K-10 Comparison of measured and simula ted (using water quality param eters obtained by modified Gauss-Newton method) water qual ity state variables at south station............................................................................................................ 274 K-11 Comparison of measured and simula ted (using water quality param eters obtained by trial-and-error method) water quality state variables at north station 276 K-12 Comparison of measured and simula ted (using water quality param eters obtained by trial-and-error method) wate r quality state variables at central station.....................................................................................................................276 K-13 Comparison of measured and simula ted (using water quality param eters obtained by trial-and-error method) water quality state variables at south station277 L-1 Temporal water quality variations at IRLV17 station from 1997 to 1999 including water qual ity kinetic terms..................................................................... 279 L-2 Temporal water quality variations at IRLV17 station from 1997 to 1999 without water quality kinetic terms.....................................................................................280 L-3 Temporal water quality variations at IRLB09 station from 1997 to 1999 including water qual ity kinetic terms..................................................................... 281

PAGE 20

xx L-4 Temporal water quality variations at IRLB09 station from 1997 to 1999 without water quality kinetic terms.....................................................................................282 L-5 Temporal water quality variations at IRLIRJ01 station from 1997 to 1999 including water qual ity kinetic terms..................................................................... 283 L-6 Temporal water quality variations at IRLIRJ01 station from 1997 to 1999 without water quality kinetic terms........................................................................ 284 M-1 Simulated near-bottom dissolved oxyge n concentration in the Upper Charlotte Harbor on July 3, 2000...........................................................................................285 M-2 Simulated near-bottom dissolved oxyge n concentration in the Upper Charlotte Harbor on July 12, 2000.........................................................................................286 M-3 Simulated near-bottom dissolved ox ygen concentration below 2mg/l in the Upper Charlotte Harbor on July 12, 2000..............................................................286 N-1 Correlation of salinity difference betw een the near bottom and near surface layers in terms of measured data an d simulated results at UF station....................287 N-2 Correlation of salinity difference betw een the near bottom and near surface layers in terms of measured data and si mulated results at Punta Gorda station.....288 N-3 Correlation of salinity difference betw een the near bottom and near surface layers in terms of measured data and si mulated results at El Jobean station.........288 O-1 Temporal SOD values at CH005 and CH009 stations in the U pper Charlotte Harbor in terms of different scenarios....................................................................289 O-2 Temporal SOD variations at CH005 a nd CH009 in the Upper Charlotte Harbor in term s of different scenarios................................................................................ 290

PAGE 21

xxi Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy MODELING OF FLORI DAS ESTUARINE SYSTEM S: UPPER CHARLOTTE HARBOR AND INDIAN RIVER LAGOON By Taeyun Kim December 2007 Chair: Y. Peter Sheng Major: Coastal and Oc eanographic Engineering This study focuses on the enhancement and ap plication of a science-based tool an integrated 3D hydrodynamic-ecological model to enable improved understanding of estuarine ecosystems and help ecosystem mana gers to restore, protect, and manage estuarine ecosystems. The integrated m odeling system, CH3D-IMS, includes coupled models of hydrodynamics, sediment transport, water quality, and light attenuation. In this study, CH3D-IMS is enhanced by the addition of several new features: a multi-algal group phytoplankton model, a sediment oxygen demand (SOD) model, a temperature model, a silica cycle, and the exchange of pa rticulate nutrients betw een the sediment and water columns. This enhanced integrated modeling system was applied to Florida estuaries (Indian River Lagoon and Upper Char lotte Harbor) where extensive field data exist. In the Upper Charlotte Harbor, circulati on and transport simulation from June 12th, 2003 to July 10th, 2004 were conducted. Due to the enormous size of the river basins,

PAGE 22

xxii dramatically large salinity fluctuations are found in response to daily and seasonally varying freshwater inflows. The model pr edictions are validated with field data measuring water level, currents, salinity, a nd temperature. The low salinity during the high-runoff period often causes the diminishing of Halodule Wrightii (a dominant seagrass species in this area) whose growth rate quickly decreases at salinity < 12 psu. According to model results, the salinity values were above 12 psu starting 5km downstream beyond the mouth of Peace River, indicating that the 30-day constant freshwater flow rate probably is below 70.2 m3/s. One of the most salient issues in the Upper Charlotte Harbor is hypoxia phenomen a in the bottom water. The 2000 water quality simulation in the study area produ ced the observed phe nomena of hypoxia. The model results demonstrated that hypoxia phenomena are strongly related to high freshwater flows and SOD. According to model simulations, reducing 50 % or 100 % of nutrient loads at the river boundaries did not eliminate th e low DO condition since the vertical stratification rema ined strong during the high-runoff period even though the SOD was reduced. A two year (1997 1999) simulation of th e water quality in the Indian River Lagoon (IRL) was conducted. CH3D-IMS reproduced the observed seasonal variations of phytoplankton and dissolved oxygen high phytop lankton values and low DO values in the summer and low phytoplankton values and high DO values in the winter. The skill of the CH3D-IMS water quality was quantitatively assessed by comparing model results vs. field data using a new skill assessment met hod the relative operating characteristic (ROC) method. It ranges from 1.0 (for a perfect model system ) to 0.0 (for a perfectly bad model system), with 0.5 indicating no skill. Most of the ROC scores of water quality state

PAGE 23

xxiii variables are over 0.6, and the maximum sc ore is 0.862. The water quality simulation with the multi-algal group module produced better results than the one-algal group module, and reproduced more algal successi ons. In addition, model simulations of episodic events in the IRL successfully repr oduced the wind and wave induced erosion of sediments and particulate nutrients.

PAGE 24

1 CHAPTER 1 INTRODUCTION AND OBJECTIVES 1.1 Introduction Florida has diverse estuaries appearing in all shapes and sizes, including bays and coasta l lagoons. These estuaries provide majo r natural resources and economic benefits: (1) Young fish and shellfish can find food a nd shelter avoiding predators, (2) The marshes and mangroves surrounding many estuaries become habitat for birds, and (3) Estuaries support commercial and recreational fisheries as we ll as attract many tourists (Humphreys et al., 1993; Florida Sea Grant, 2006). In addition, estuaries reduce the damage caused by storms. Due to the valuable coastal and estuarine resources, Floridas coast has developed rapidly. However, ra pid developments have caused tremendous stress which has severely damaged estuarine habitat, water quality, and the overall health of estuaries (Dixon and Kir kpatrick, 1999; Alber 2002; Steidinger et al., 1998; Badylak and Phlips, 2004; and Phlips et al., 2004). Estuarine habitat loss is a serious problem in Florida. The loss of habitat between the 1940s and 1980s measures at up to 50 % of salt marsh; nearly 60 % of seagrass; and up to 85 % of mangroves from a survey of 10 selected estuaries. In the Indian River Lagoon (IRL), one-fourth of the seagrass has been destroyed since 1940 (Humphreys et al., 1993; Florida Sea Grant, 2006). In some parts of the lagoon, losses have been as much as 83 %. One of Florid as largest estuarie s since 1945, Charlotte Harbor has lost more than half of the salt marsh and 29 % of the seagrass (Humphreys et al., 1993;

PAGE 25

2 Humphreys and Grantham, 1995). These losses jeopardize endangered and commercial species as well as environmental balance. The decline in seagrass is directly related to the reduction of light. The availability of light is associated with phytoplankton, suspended particulate, color, and water itself (Kirk, 1984; Christian and Sheng, 2003). Therefor e, as water quality declines, the amount of light reaching seagrass is reduced, which adversely affect the growth of seagrass. Many of Floridas estuaries are confronted with water quality problems. Due to the nutrient and suspended particle loading, IRL transforms a macrophyte-based system to an algae-based system (Sigua et al., 2000). Water quality in Charlotte Harbor is considered to be in the good range except the Upper Charlotte Harbor which receive s river flow from the Peace and Myakka Rivers. These nutrient-rich river discharges have decreased the clarity of water and promoted phytoplankton blooms (Humphreys and Grantham, 1995). In addition, bottomwater hypoxia, a condition of low dissolved oxy gen concentration (< 2.0 mg/l), has been observed 3 15 % of the time in the Uppe r Charlotte Harbor since 1975 (Heyl, 1998). The degradation of water quality and bottom-water hypoxia may cause many harmful impacts in this estuary. In addition to water quality degradation, fr eshwater flow causes loss of habitats and the decline of fish and wildlife species. For instance, Halodule wrightii one popular seagrass species in Florida can tolerate a wi de range of salinity (17 36 PSU). However, excessive freshwater may sharply decrease the growth rate of Halodule wrightii when salinity is reduced to below 12 PSU. Mortalit y could occur at salin ity < 6 PSU (Doering et al., 2002; Zimmerman and Livingston, 1976). Therefore, Flor ida legislature recognized

PAGE 26

3 the potential problems of alte red freshwater flow and has designed rules of water management such as minimum flow a nd level (MFL) criteria (CHNEP, 2002). Acknowledging serous estuarine problems is a first step toward their solutions, which consequently necessitate s a management tool. This study focuses on enhancing a useful tool to aid ecosystem managers in not only addressing current issues, such as degradation of water quality, hypoxia, and impaction of freshwater but also making important decisions for future planning. Th ese decisions include establishing minimum flow and level (MFL) rules, setting pollutant load reduction goal (PLPG), prediction oil spill dispersion, and tracking of other drifting an d floating objects. The tool is the threedimensional integrated modeling system CH 3DIMS developed by Sheng et al (2002). CH3D-IMS consists of CH3D (Sheng 1987, 1990), a curvilinear-grid hydrodynamic model, and coupled models of wave, sediment transport, water quality, and light attenuation. In this study, CH3D-IMS is enhanced by adding a sediment oxygen demand (SOD) model and a multi-algal group model. This enhanced CH3D-IMS model was applied to the IRL and the Upper Charlotte Harbor in Florida where multi-phytoplankton communities exist. The phytoplankton groups show seasonal succession and biomass changes because each phytoplankton species has its own dynamics and unique eutrophication characteristics (Walsby a nd Reynolds, 1980; Smayda and Reynolds, 2001; Kim et al, 2004). These seasonal variations ar e closely related to the nutrient cycling through uptake during growth, excretion, a nd mortality. The CH3D-IMS with the multialgal group model better simulates water quality species in the study areas. In addition,

PAGE 27

4 the new SOD model enables the simulation of DO changes in bottom water associated with changes in nutrient loads. The following chapter (Chapter 2) review s the previous research related to freshwater inflow and salinity dynamics multi-phytoplankton modeling, silica cycle modeling, and SOD modeling. Chapter 3 deals with the characte ristics of the IRL and the Upper Charlotte Harbor estuarine system. In Chapter 4, water quality model of CH3DIMS is mainly presented, while the other models, such as hydrodynamics, sediment transport, and light attenua tion are briefly described. Ch apter 5 presents hydrodynamic field data and model calibration and verification using 2003 to 2004 data in the Upper Charlotte Harbor. In Chapter 6, water quality simulation of Upper Charlotte Harbor is described. The simulation demonstrates th at hypoxia events ar e caused by vertical stratification and sediment oxygen demand (SOD ). Measured water quality data in 2000 are employed for verifying model simulation. Ch apter 7 presents IRL water quality field data and model simulation. This chapter verifies IRL water quality model and demonstrates the succession of phytoplankton communities using field data. Conclusions and discussions are presented in Chapter 8. 1.2 Objectives This study aim s to achieve the following objectives: Calibrate and validate the CH3D-IMS with hydrodynamic data for the Upper Charlotte Harbor. Demonstrate salt-wedge dynamics near th e mouths of Peace and Myakka Rivers. Develop a management tool to assist in the determination of minimum flow and level (MFL) criterion. Investigate the effect of changing rive r flow rates on salt-wedge dynamics and suggest the river flow rate for sustaini ng seagrass growth in the Upper Charlotte Harbor.

PAGE 28

5 Validate CH3D-IMS with multi-species phytoplankton model using measured water quality state variables in th e Upper Charlotte Harbor during 2000. Demonstrate the dynamics of bottom water hypoxia in the Upper Charlotte Harbor using CH3D-IMS with a new SOD model Evaluate the hypothesis that re duction of river nutrient loads, which is directly related to SOD, can diminish the bottom-water hypoxia. Conduct model sensitivity te sts to quantify the influe nce on water quality in the Upper Charlotte Harbor by sediment nutri ents, nutrients at open boundary, river boundary nutrients, and grid resoluti on in the Upper Charlotte Harbor. Simulate long-term (1997 to 1999) IRL water quality dynamics and validate CH3D-IMS with water quality data. Simulate IRL episodic water quality dynamics and validate CH3D-IMS with episodic event data, and evaluate the shor t term effect of so rption/desorption as well as particulate nutrient er osion processes in the IRL. Quantify the seasonal dynamics of nu trients, phytoplankton succession, and suspension of sediments and part iculate nutrients in the IRL. Conduct sensitivity tests to determine how IRL water quality is affected by such factors as sediment nutrient concentra tions, nutrients at open boundary and river boundary, surface light intensity, grid resolution, and sediment properties. Apply a new skill assessment method, the re lative operating characteristic (ROC), to the performan ce of CH3D-IMS wate r quality model. Test the feasibility of using a modified Gauss-Newton method to estimate water quality parameters of CH3D-IMS.

PAGE 29

6 CHAPTER 2 LITERATURE REVIEW 2.1 Freshwater Inflow and Salinity It is becom ing common for policies to be en acted that mandate the establishment of freshwater inflow criteria th at will preserve and protect estuarine ecosystems. These policies to maintain an ecol ogically sound environment in Texas, to ensure that no harm comes to the water resources or ecology of the area in Florida, and to restore ecological health in San Francisco Bay Estuary are broad. To meet these broad policy objectives, an understanding of the connections between freshwater inflow, estuarine conditions, and resources must be made (Alber, 2002). The Texas Water Development Board (TWDB) and the Texas Parks and Wildlife Department (TPWD) jointly established and currently maintain a data collection and analytical studies programs focused on de termining the effects of and needs for freshwater inflow into the states ten bay and estuary systems. They employ the Texas Estuarine Mathematical Programming (TxEMP ) model and apply it to the Galveston Bay and the Trinity-San Jacinto Es tuary. The monthly optimized river flow ranges from 5.13 x 109 m3 to 7.65 x 109 m3, and the peak performance measures at 6.44 x 109 m3 (Powell, 2002). In the San Francisco Estuary, freshwater flow has been altered by shifts in seasonal patterns of river flow causing concern for seve ral species. It is essential to make more effective and applicable the salinity standard (Kimmerer, 2002). South Florida Water Management District (SFWMD) estimated a minimum flow in the Caloosahatchee

PAGE 30

7 Estuary to maintain the salt-t olerant freshwater species (Val lisneria Americana) at the head of the estuary and a maximum flow to prevent mortality of the marine species (Halodule Wrightii) at its mouth (Doering, 2002). Park (2004) applied CH3D to assess the impact of the removal of the Sanibel Causeway and Intra Coastal Waterway (ICW) 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, FL. The results show that these hydrologic alterations do not appear to impact noticeably the flow and salinity patterns in the San Carlos Bay and Pine Island Sound. The salinity at the Caloosahat chee River Mouth is reduced by about 1.36 PSU in the absence of the ICW. The relationshi p between salinity at Fort Myers, FL and river discharge at S79 was established by comparing 1-day and 30-day averaged salinity values at Fort Myers and fres hwater inflow at S79. Accordi ng to this relationship, a total river discharge of 18m2/s at S79 produces a 1-day aver aged salinity of 20 PSU at Fort Myers. 2.2 Multi-Phytoplankton and Silica Cycle Water quality m odels have been develope d by numerous studies in the past three decades. DiToro et al. (1983) developed the water quality analysis simulation program (WASP) using the explicit scheme to quan tify advective mass tr ansport. HydroQual (1987) developed a water quality model based on the box-model approach. Water quality models, which have improved representation of transport processes, have been developed and applied (e.g., Cerco and Colo, 1995; Park et al., 1995). Other water quality models have been developed based on improvement of the above referred models (Wang and Johnson, 2000; Zheng et al., 2004; Tillman et al., 2004).

PAGE 31

8 Park et al. (1995) developed a water quality model with twentyone state variables integrated with EFDC to form a threedimensional Hydrodynamic-Eutrophication Model (HEM-3D) of the Virginia Institute of Ma rine Science (VIMS). The HEM-3D model was applied to the Chesapeake Bay (Park et al., 1996), York Ri ver (Shen et al., 1998), James River (Shen et al., 1999), K yunggi Bay (Kim et al., 2000), and Kwang-Yang Bay (Park et al., 2005). This model, upon receiving the information of physical transport from EFDC, simulates the spatial and temporal distributions of water quality state variables, including dissolved oxygen, suspended algae (3 groups), various components of carbon, nitrogen, phosphorus, silica cycle, and fecal coliform bacteria. Algae, which occupy a central role in the model, are grouped in to three model state variables: cyanobacteria (blue-green algae), diatoms, and green algae. Sources and sinks included in the model are growth (production) basal metabolism, predation, settling, and external loads. Equations describing these pr ocesses are largely comparable to the three algal groups, with differences arising in the equation parameter values. The kinetic equation describing these processes is V WB z BWS BPRBMP t Bx xx xxx x x *) ( (2-1) where Bx is the algal biomass of algal group x (g C m-3), t is time (day), Px is the production rate of algal group x (1/day), BMx is the basal metabolism rate of algal group x (1/day), PRx is the predation rate of algal group x (1/day), WSx is the settling velocity of algal group x (m/day), WBx is the external loads of algal group x (g C/day), and V is the cell volume of the body of water (m3). HEM-3D has two state variables for silica: particulate biogenic silica and available silica. Sources and sinks of particulate bioge nic silica included in the model are diatom

PAGE 32

9 basal metabolism and predation, dissolution to available silica, se ttling, and external loads. Those of available silica included in the model are diatom basal metabolism, predation and uptake, settling of particulate available silica, dissolution from particulate biogenic silica, and sediment-water exchange of dissolved silica for the bottom layer only. Analysis of Chesapeake Bay monitoring data indicates that silica shows similar behavior as phosphate in the adsorption-deso rption process (Cerco and Cole, 1994). As in phosphate, available silica is defined to include both dissolved and particulate fractions. Treatment of available silica is the same as total phospha te, and the same method of partitioning dissolved and particulate phosphate is used to partition dissolved and particulate available silica. The kinetic e quations of particulat e biogenic silica and available silica are as follows: V WSU z nSUWS SUKBASCPRFSPP BMFSP t SUd SUA dd d d d *** (2-2) V WSA z BFSA z nSAWS SUKBASCPPRFSIP BMFSI t SAd P TSS SUA dd dd d d ** (2-3) where SU is concentration of particulate biogenic silica (g Si m-3); FSPd is a fraction of metabolized silica by diatoms produced as par ticulate biogenic silica; FSPP is a fraction of predated diatom silica produced as particulate biogenic silica; ASCd is silica-to-carbon ratio of diatoms (g Si per g C); KSUA is dissolution rate of particulate biogenic silica (1/day); WSd is settling velocity of particulate bioge nic silica; WSU is external loads of particulate biogenic silica (g Si /day); and SA is concentration of available silica ( = SAd + SAp, g Si m-3); SAd is dissolved available silica (g Si m-3); SAp is particulate available

PAGE 33

10 silica (g Si m-3); FSId is fraction of metabolized sili ca by diatoms produced as available silica; FSIP is fraction of predated diatom silica produced as available silica; WSTSS is settling velocity of total suspended solid; BFSAd is sediment-water exchange flux of available silica (g Si m-2 day-1), applied to the bottom layer only; and WSA is external loads of available silica (g Si /day) (Park et al.1995). The Corps of Engineers Integrated Compartment Water Quality Model (CEQUAL-ICM) was designed and applied to Ch esapeake Bay (Cerco and Cole, 1994). The model was used for other areas: the Delaware Inland Bays (Cerco et al., 1994), Newark Bay (Cerco and Bunch, 1996), the San Juan Es tuary (Bunch et al., 2000 ), and Florida Bay (Cerco et al., 2000). Tillman et al. (2004) re cently applied the model to the St. Johns River. Algal sources and sinks are mechan ical processes that consist of production, respiration, predation, and settling (Equation 2-4). Equations governing the three groups are largely the same. Differences are expre ssed through the magnitude of parameters in the equation. PRB z WRP t Ba *) ( (2-4) where B is algal biomass, expressed as carbon (g C m-3); P is production (1/day); R is respiration (1/day); Wa is algal settling velocity (m/day); and PR is predation (g C m-3 d1). The model incorporates two siliceous stat e variables, which consist of dissolved silica and particulate biogenic silica. Disso lved silica is available for utilization by diatoms. Particulate biogenic silica is produced through diatom mortality. Sources and sinks represent predation, diatom production an d respiration, dissolution of particulate to

PAGE 34

11 dissolved silica, and settling (Figure 2-1). Equations 2-5 and 2-6 are the kinetic equations of dissolved silica and particulat e biogenic silica respectively. z BENSA PBSKPRFSAPBPA t SApbs SC *)* *( (2-5) z PBS WPBSKPRFSAP BRA t PBSpbs pbs SC *)*)1(*(* (2-6) where SA is dissolved silica (g Si m-3); Asc is algal silicato-carbon ratio (g Si g-1 C); FSAP is fraction of predation released as dissolved silica (0 < FSAP < 1); PBS is particulate biogenic silica (g Si m-3); Kpbs is particulate biogenic silica dissolution rate (1/day); BENSA is release of SA from sediments (g Si m-3 d-1); and Wpbs is the settling velocity of particulate biogenic sili ca (m/day) (Tillman et al., 2004). Figure 2-1. Model Silica Cycle (Tillman et al., 2004)

PAGE 35

12 The HEM-3D model and CE_QUAL_ICM mode l, which have been used in many areas, are compared to my proposed mode l in Table 2-1. Those models divide phytoplankton into three categories. These tw o models use similar algae dynamics and silica cycle. Zooplankton is not considered but treated with a consta nt predation rate for phytoplankton dynamics. However, the proposed water quality model considers zooplankton dynamics and its effects on multialgae dynamics. This consideration better represents multi-phytoplankton dynamics b ecause phytoplankton and zooplankton are so closely tied through predator-prey interaction. Table 2-1. Comparison of multi-phytoplankton models Park et al. (1995) Tillman et al. (2004) This study Water Quality Model HEM-3D CE_QUAL_ICM CH3D-IMS Algae Three groups (Diatom, cyanobacteria, and green algae) Three groups (Diatom, cyanobacteria and other phytoplankton) Multi groups Sources and sinks of algae Growth, basal metabolism, predation, settling and external loads Production, respiration, predation, settling, and external loads Growth, respiration, non-predatormortality, grazing by zooplankton, settling, and external loads Zooplankton Constant predation rate Constant predation rate Dynamic equation Silica state variables Particulate biogenic silica and available silica Particulate biogenic silica and dissolved silica Particulate organic silica and soluble organic silica Sources and sinks of silica Diatom basal metabolism, uptake, predation, dissolution to available silica, settling, external loads Diatom basal metabolism, uptake, predation, dissolution to available silica, settling, and external loads Diatom basal metabolism, uptake, predation, sorption/desorption reactions, settling, respiration mortality by zooplankton, and external loads

PAGE 36

13 2.3 Sediment Oxygen Demand SOD is due to the oxidation of organic matter in bottom sediments. These organic matters are derived from nutrient loads, pl ant matter, phytoplankton mortality, and other sources. Estuaries tend to exhibit a different carbon chemistry from freshwater systems because of the dominance of sulfur. Sulfate is used before carbon dioxide as the electron acceptor. Therefore, SOD in estuaries is dependent on the reduction of sulfate (SO4) to produce sulfide (H2S) in the anaerobic zone. Sulfate is depleted as a consequence of the reaction (Barnes and Goldberg, 1976; Chapra, 1997; DiToro, 2000) OHSHCO SOHOCH2 22 42 22 2 2 A portion of sulfide in the anaerobic zone reacts with iron to form particulate iron sulfide (Morse et al., 1987). The remaining sulfide diffu ses into the aerobic z one where sulfide is oxidized in Figure 22 (DiToro, 2000). Figure 2-2. Schematic diagram of sulfide oxida tion (DiToro,2000)

PAGE 37

14 In addition, observations reveal ed that SOD in the intertidal salt marsh was higher than in the estuary bed (Pomeroy et al., 1972). Zheng et al. (2004) developed a coupled three-dimensional physical and water quality model for the Satilla River Estuary in Georgia. The physical model is a modified ECOM-si version with inclus ion of flooding/draining proces s over the inte rtidal salt marsh. The water quality model is a modified WASP5 with in clusion of nutrient fluxes from the bottom sediment layer. This paper concluded that the low dissolved oxygen (DO) concentration in the Satilla River Estuary was mainly due to high sediment oxygen demand (SOD) over the intertidal salt mars h. In the model experiment, SOD was specified as 2.5 gO2m-2day-1 in the intertidal salt marsh and 1.2 gO2m-2day-1 in the estuary bed (Zheng et al., 2004). The SOD kinetic equation is written as 20* TH SOD t DO (2-7) where H is water depth (m), DO is oxyge n concentration, SOD is sediment oxygen demand (gO2/m2-day), and is temperature coefficient (=1.08). Park (2004) used the diffusing SOD ki netic equation in his modeling study of hypoxia in the Upper Charlotte Harbor using CH3D-IMS: DOK DO STS HH SOD t DOSOD T B **** 120 20, (2-8) where H is water depth (m); DO is oxygen concentration (mg/l); SOD is sediment oxygen demand (gO2/m2-day); SB,20 is an areal SOD rate at 20C (gO2/m2-day) which is user defined value; ST is a fractional coefficien t for sediment type (When sediment type 2, ST=1.3; when sediment type>3, ST=1.0) (Table 2-2); is temperature coefficient, which Zison et al. (1978) have reporte d a range of 1.04 to 1.13 for A value 1.065 is

PAGE 38

15 commonly employed; and KSOD is half saturation rate for SOD, which Lam et al. (1983) suggested a value of KSOD (1.4 mg/l) for Lake Erie. The SOD is negligible when DO is much lower than KSOD. When dissolved oxygen increase s in the water column, the maximum oxygen demand is released to the water as sediment oxygen demand. Table 2-2. Sediment type (Park, 2004) Type D50 Range ( mm) Category Very coarse D50 >0.50 5 Coarse 0.25 > D50 > 0.50 4 Medium 0.125 > D50 > 0.25 3 Fine 0.0625 > D50 > 0.125 2 Silts or clay 0.0625 > D50 1 A water quality model of Florida Bay was developed by Cerco et al. (2000). To obtain SOD, Cerco et al. (2000) used a sediment diagenesis model which computes the diagenesis (decay) of organic matter in the se diments, the resulting production of oxygen demand, and the movement of diagenesis products between sediments and water column. The diagenesis model consists of three pro cesses: (1) Deposition of particulate organic matter, (2) Diagenesis, and (3) Flux of substa nces produced by diagenesis to the sediment layer, to the water column, and to deep se diment. The flux portion is the most complex. Table 2-3 shows the notable features of the SOD models used by Zheng et al. (2004), Park (2004), Cerco et al. (2000), a nd this study. Empirica l SOD models do not relate SOD to the external nutrient loads because constant SOD values are employed. Although the sediment diagenesis model is more robust than the simple empirical formula, it could create more uncertainty with the absence of data for CH4, H2S, Iron, CO2, and more obscure variables, such as burial rate, decay rate, sediment-water masstransfer coefficient, particle mixi ng velocity, and other coefficients.

PAGE 39

16 Table 2-3. Notable features of SOD model Zheng, et al. (2004) Park (2004) Cerco, et al. (2000) This study Method Empirical formula Empirical Formula Sediment diagenesis model A naive Streeter-Phelps SOD model (Chapra, 1997) SOD directly related to organic matter of water column No No Yes Yes SOD related to external loading No No Yes Yes Complexity Simple Simple Very complex Less complex Data required Areal SOD rate and temperature Areal SOD rate, temperature, DO, half concentration rate of SOD Methane, sulfide, iron, nutrients, temperature, settling velocity, and many variables Areal SOD rate, temperature, DO, half concentration of SOD, particulate CBOD, and settling velocity

PAGE 40

17 CHAPTER 3 CHARACTERIZATION OF STUDY AREAS 3.1 Upper Charlotte Harbor Characterization Charlotte Harbor, a coastal-plain estuarine system, is one of the largest estuarine systems on the Southwest Florida coast and an important part of the Gulf of Mexico watershed. As shown in Figure 3-1, the drainage area consists of the Peace, Myakka, and Caloosahatchee River watersheds that drain directly into the harbor. 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 in Figure 3-2 (Park, 2004) Charlotte Harbor has a surface area of 768 km2, a drainage area of 11,956 km2, and an average depth of 2.1 m (Stoker, 1992). The majority of the freshwater that en ters Charlotte Harbor comes from the Myakka River, Peace River, and Caloosah atchee River. Average flows are 17.8 m3/s, 56.9 m3/s, and 56.7 m3/s, respectively (Park, 2004). Flows in the Myakka and Peace Rivers are largely unregulated, while flow in the Caloosahatchee River is controlled by operation of the Franklin Lock about 43 km upstream from the mouth. Therefore, the Caloosahatchee River discharge does not al ways correspond to rain fall patterns in the basin since it is controlled by S-79. Upper Charlotte Harbor located in the nor thern portion of Charlotte Harbor is defined as the area from Boca Grande east and north to the Peace and Myakka Rivers. It has about 460 km2 in the surface area and an averag e depth of 2.6 m (Stoker, 1992).

PAGE 41

18 Upper Charlotte Harbor receives freshwater inflows from the Myakka and Peace Rivers. The headwaters of the Peace River are loca ted approximately 100 miles north of the Harbor in Polk County. The Myakka River st arts in eastern Mana tee County and flows through Sarasota and Charlotte Counties to the northwest corner of Upper Charlotte Harbor. Figure 3-1. Drainage Basins (Stoker, 1992)

PAGE 42

19 StudyArea MyakkaRiver PeaceRiverGasparilla PassCaloosahatcheeRiverEsteroBay Sanibel Island Redfish Pass Captiva Pass BocaGrande PassCharlotte HarborSanibel Causeway PineIsland Pine Island SoundDeSotoCounty LeeCountyN FortMyers BeachMatlachaPassG U L F O F M E X I C OSanCarlos Bay Gasparilla Sound Blind Pass Naples FortMyers Cape CoralShellCreek PuntaGorda Venice1001020 KilometersCharlotteCounty SarasotaCountyF l o r i d a Figure 3-2. Map of Charlotte Har bor Estuarine System (Park, 2004) The climate of the Upper Charlotte Harbor is subtropical and humid with a mean temperature of 72 F. The high mean temper ature is 80 F during the summer and low mean temperature is 60 F in December and January (McPherson et al. 1996). The annually average wind speed is 3.9 m/s from the east. Wind speed can exceed this average during the passage of winter storms or summer hurricanes (Wolfe and Drew, 1990). Annual rainfall averages 132 cm with more than half occurring from June through September (Goodwin, 1996). Low rainfall in Apr il and May with high evaporation results in the lowest streamflow, lake stage, a nd ground water levels of the year (Hammett, 1990).

PAGE 43

20 Due to rainfall and enormous size of rive r basins, seasonal freshwater inflow plays an important role in maintaining the salinity ba lance in the estuary. Salinity fluctuates in response to seasonal variation in riverine freshwater infl ow. Salinity is generally the lowest during the July thr ough September wet season, and the highest between January and March. Salinity is vertically stratified lower at the surfa ce and higher near the bottom during the periods of high freshwater in flow (Stoker, 1992; Turner et al., 2001) Evaluation of the historical data indi cates a significant relationship between cumulative discharge from the Peace and Myakka Rivers and hypoxia conditions (Heyl, 1998; CDM, 1998, Tomasko et al., 2006; Turner et al., 2006). During large runoff events, vertical stratification occurs in Upper Charlotte Harbor. These events supply nutrients to the estuary which promote phytoplankton production. Excess production settles to the bottom of the estuary, decays, and then consumes oxygen (Cerco et al., 1994). This suggests that sediment and be d-load oxygen demands remain likely forcing functions for hypoxia (Dixon et al., 2003). Also, Hubertz et al. (2004) reports that the presence of even a slight stratification allows the sedi ment oxygen demand and other biochemical processes to create hypoxia conditions. The hypoxia is not c onsidered to be a totally natural phenomenon, as evidence indicates th at organic loads to the bottom sediments have increased over the past century (Turner et al., 2006). There are multiple phytoplankton species in the Charlotte Harbor estuarine system. Diatoms were dominant in 55 percent of 289 phytoplankton samples collected in the Charlotte Harbor estuaries system in 1983 and 1984; cryptophyt es in 35 percent; cyanophytes in about 6 percent; dinoflagellates in about 4 percent; and other classes in 1 percent (McPherson et al., 1996). Surveys of the Upper Charlotte Harbor were conducted

PAGE 44

21 on June 17, July 8, and July 29, 2003 in an effort to monitor phytoplankton community composition. These surveys show that diatom dominated communities in the upper harbor changed to higher abundances of flag ellates and cyanophytes during the peak of the freshwater event (K irkpatrick et al., 2003). Seagrass data from the Upper Charlotte Harbor are not abundant. However, the available data indicate that Thalassia testudinum and Halodule wrightii are dominant in this area. Syringodium Filiforme was only present as a minor component near the mouth of Captiva Pass. Spatial variation appears that Halodule wrightii was barely observed at Punta Gorda (PUN) in Figure 3-3. On the other hand, Thalassia testudinum was widely distributed in the Upper Charlotte Harbor. Figure 3-3. Northern Charlo tte Harbor seagrass stations (Dixon and Kirkpatrick, 1999)

PAGE 45

22 In particular, seagrass near the mouth of the Myakka and Peace Rivers has stresses of both low light and low sa linity (Dixon and Kirkpatric k, 1999). Each species of seagrass has different ranges of salinity tolerance. The tolerance hierarchy was Halodule > Thalassia >> Syringodium (Fong and Harwell, 1994). Especially, Thalassia testudinum is positively correlate d with salinity (Tomasko and Hall-Ruark, 1998). 3.2 Indian River Lagoon Characterization The Indian River Lagoon is a coastal la goon estuary, separated from the ocean by barrier islands and limited exchanges with the ocean through inlets (Figure 3-4). It is located on the east coast of Florida. The sh ape of the lagoon is long and narrow except the inter-coastal waterway (ICW) whose width varies from 500 meters to 7 km. The lagoon extends around 255 km from Ponce de Le on inlet in the north to Jupiter Inlet through four coastal counties along Floridas Atlantic coast. Th ere are five inlets Ponce de Leon Inlet, Sebastian Inlet, Ft. Pierce In let, St. Lucie Inlet, and Jupiter Inlet connected with the Atlantic Ocean. The averag e depth is 1.5 to 2 meters, and the width varies between 0.4 and 4 km. The lagoon is composed of three interconnected major water bodies Mosquito Lagoon, Banana River, and Indian River Lagoon proper (Steward et al., 1994; Sheng et al., 2003c). Indian River Lagoon system has a long summe r, which starts in April or May and extends for about nine months. Based on the Florida Department of Environmental Protection (FDEP) temperature data in 1998, the warmest mo nth is July with a mean temperature of 23 to 31C. February is the coldest month with a mean temperature of 11 to 20C. The fluctuations of temperature are moderate due to low latitude and proximity to the ocean (Hinkle et al., 1995). During th e winter, winds are from the North to

PAGE 46

23 Northwest, while southerly winds dominate in the summer months. Average monthly wind speeds range from 3 to 4.6 m/s (Hi nkle et al., 1995; Echternacht, 1975). Figure 3-4. Map of Indian River Lagoon (Sheng et al., 2003c) The sources of freshwater in the lagoon are precipitation, gro undwater seepage, surface water runoff, and discharge from creeks and streams, while sa lt water enters into the system from ocean boundaries. Salinity distributions in the lagoon are quite different during the dry and wet seasons. From northern IR L proper to Sebastian Inlet, the salinity

PAGE 47

24 values are lower than central and southern IRL proper during the wet seasons. However, under the dry seasons, hypersaline conditions oc cur in the northern IRL proper. Due to the tidal exchange, the salinity values of the central and southern IRL proper do not change in response to the fr eshwater input (Smith, 1993). Currents are primarily driven by tides, wi nd, density, and river discharge in the IRL. In the northern part, semi-diurnal and diurnal constituent amplitudes are insignificant (around 1 to 8 % of the total) Non-tidal and low-frequency variances dominate, and therefore currents may be driv en by wind, river discharge, evaporation, baroclinic pressure and density. On the other hand, tidal constituent amplitudes are substantial in the southern part (Fort Pierce, St. Lucie, and Jupiter Inlets). Currents may be mainly affected by tides and rivers. In th e central portion of the IRL proper, tidal and non-tidal components coex ist. The low-frequency and non-tidal variances are between 40 to 60 % of the total (Smith, 1987). As a sali ent feature in IRL, due to the causeways restricting flow to a small area, some artifici al gyres are formed and trapped pollutant or floating bodies (Sheng et al., 2003a). Some flushing studies of IRL have conducted in terms of methodology and segmentation (Simth 1993; Virnstein et al., 1994; Sheng et al., 2003d) Flushing time is the fastest in the southern and central IRL proper which includes Sebastian Inlet, Ft. Pierce, St.Lucie, and Jupiter Inlet. The 50 percent renewal time (R50) is less than eleven days. Even though Mosquito Lagoon is connect ed to the Atlantic Ocean through Ponce de Leon Inlet, R50 is roughly four times longer than the southern IRL proper because of the narrow connection and nume rous islands. The northern IRL proper, which includes

PAGE 48

25 no inlets, has slower flushing time than othe r areas except the Bana na River Lagoon. The Banana River Lagoon has been the longest R50 about one year (Sheng et al., 2003d). Recent surveys indicate the sediments in the IRL range in size from clay to coarse sand. Bulk density values range from 1.29 to 1.55 g/cm3. The dominant type of bottom sediment is fine sand. Clay and silt cover at most five percent of total bottom sediments even though most of the sediment loads from river consist of muck (Sun, 2001). The sediments in IRL are accumulating and becoming a major concern. 123 million pounds of sediment a year are flushed into the IRL from many sources including unpaved roads, construction sites, parking lots, and residen tial areas. These sediment loads may diminish the trapping capacity of lagoons tributarie s and transport muck to the estuary. Transported muck causes degradation to the wate r quality and habitat. In addition, a large storm such as Hurricane Erin in 1995, which pr oduced 6-10 inches of rainfall, can cause flooding problems because sediment-clogged cr eeks and ditches blocked the flow of water (Smithson, 1999). More than two thousand identified species live in IRL making it one of the most diverse estuaries in the world (Barile et al., 1987). Seagrass beds in the shallow waters and tidal swamp forests provide food for th ese species as well as protection from predators. There are currently between 70,000 and 90,000 acres of seagrass beds in the IRL. Seven species of seagra ss are present in the lagoon: Halodule wrightii Syringodium filiforme Halophila engelmannii Thalassia testudinum Halophia decipiens Halophia johnsonii and Ruooia maritima. Halophila johnsonii is founded only in the southern IRL proper beds (Virnstein and Morris, 1996). Ho wever, the amount of seagrass beds has decreased from past years (V ernstein et al., 1994).

PAGE 49

26 The declines of seagrass are believed to be due in part to con tinued degradation of water quality. Assessment of water quality data from 1988 to 1994 was conducted (Sigua et al., 2000). This assessment confirms water quality vari ations exist spatially and temporally in the IRL. Higher nutrient enrichment and algal productivity appear in the wet season (summer through early fall). The wet season also impart s higher dissolved organic loads in the IRL than in the dry s eason (winter through spring). Total Kjeldahl Nitrogen (TKN) was high in the north and lo w in the south. The reverse pattern was observed for Total Phosphorous (TP). A two year study of the phytoplankton community illustrated that it was dominated by dinofla gellates, diatoms, and cynobacteria. Each phytoplankton group has its own sp atial and temporal pattern (Badylak and Phlips, 2004).

PAGE 50

27 CHAPTER 4 MODEL DESCRIPTION The chapte r provides a description of the CH3DIMS. This model is completely integrated with a curvilinea r-grid hydrodynamics 3D model CH3D, a sediment transport model, a water quality model, and a light a ttenuation model. An enhanced water quality model is a main focus of the discussion. The model includes multi-species phytoplankton model as well as new SOD model. 4.1 Hydrodynamic Model CH3D IMS is based on CH3D, a three-dimensional curvilinear-grid hydrodynamic model, which was developed by Sheng (1983) and Choi (1992). The model uses a transformed vertical coordinate and a boundary-f itted coordinate system. More information on this model can be found in Appendix A. The continuity, momentum and transport equations are as follows (Sheng, 1983): 0 z w y v x u z u A zy u x u Afv x g z uw y uv x uu t uV H 2 2 2 2 z v A z y v x v Afu y g z vw y vv x vu t vV H 2 2 2 2 (4-1) z S D zy S D yx S D xz wS y vS x uS t SV H H z T K zy T K yx T K xz wT y vT x uT t TV H H

PAGE 51

28 where u, v, and w are the velocity com ponents in the horizontal, xand y-, and zdirections; t is the time; is the free surface elevation; g is the gravitational acceleration; AH and AV are the horizontal and vertical turb ulent eddy coefficients; S is salinity; DH and DV are the horizontal and vertical diffusi vity coefficients for salinity; T is temperature; and KH and KV are the horizontal and vertical diffusivity coefficients for temperature. 4.2 Sediment Model As a part of CH3D IMS, a suspended sediment model was developed by Sheng (1983). This model consists of settling, flocculation, erosi on, and deposition mechanism. This model has developed over the past decade (Sheng and Chen, 1992; Yassuda and Sheng, 1997; Sun 2001). Sun (2001) separated su spended sediment size into 2 groups and the governing equation is the following: z c D zy c D yx c D xz cww y vc x uc t ci V i H i H isi i i i)( (4-2) where u, v, and w are the velocity com ponents in the horizontal, xand y-, and zdirections; t is the time; ci is suspended sediment concentration for group i, wsi is settling velocity for group i, DH and DV are the horizontal and verti cal diffusivity coefficients respectively; and i=(1,2) w ith 1 for the fine group and 2 for the coarse group. More information on this model can be found in Appendix B. 4.3 Water Quality Model The water quality model in CH3D-IMS wa s developed and applied in many areas: Lake Okeechobee (Chen and Sheng, 1995), Ta mpa Bay (Yassuda and Sheng, 1997), and the Indian River Lagoon (Sheng et al., 2003c). This model recapitula tes the interactions between oxygen balance, nutrient dynamics, phytoplankton and zooplankton dynamics,

PAGE 52

29 light, temperature, and salinity. The water quality equations are derived from an Eulerian approach, using a control volume formulation. The equation for each of the water quality parameters is as follows (Yassuda, 1996): )()()()( )]([)( iviii iii QuDu t (4-3) where (i) is the evolution term (rate of change of concentration in the control volume), (ii) is the advection term (fluxes in to/out of the control volume due to advection of the flow field), (iii) is the dispersion term (fluxes into /out of the control volume due to turbulent diffusion of the flow field), and (iv) is the sink/sour ce term, representing biogeochemical processes due to sorption/desorption, oxidation, excretion, decay, growth, biodegradation, etc. In the followi ng sections, the biogeoc hemical processes in Equation (4-3) will be discussed in detail for phytoplankton and zooplankton dynamics, oxygen balance, and nutrient dynamics. In addi tion, water quality parameters in the model will be described. 4.3.1 Phytoplankton and Zooplankton Dynamics Phytoplankton kinetics play the major part of this water quality model and remarkably influence the overall water quality in the system (Park, 2004). These kinetics are represented by growth, respiration, nonpredator-mortality, grazing by zooplankton, and a settling term. The entire phytoplankton community is represented as a carbonaceous biomass. Chlorophyll a concentrations, for comparison with observations, are obtained through division of computed carbonaceous biomass by the carbon-tochlorophyll a ratio. The sources and sinks of multi-species phytoplankton in the conservation equation can be written as:

PAGE 53

30 ,....)3,2,1(* iZOOC PHYC WS z KasKax t PHYzi i i i i i (4-4) where, i represents the multi-species phytopl ankton; PHYC is the multi-species phytoplankton biomass, expressed as carbon (gCm-3); is the multi-species phytoplankton growth rate (1/day); Kax is re spiration of the multi-species phytoplankton (1/day); Kas is non-predator mortality of the multi-species phytoplankton (1/day); WS is the multi-species phytoplankton settling velocity (m/day); z is the zooplankton growth rate (1/day); ZOOC is the zooplankt on biomass, formulated as carbon (gCm-3). Each phytoplankton has its own growth ra te depending on intensity of light, availability of nutrients, and temperature. Th e growth rate equation is as follows (Bowie et al., 1985): )(),(),()(),(min*)(*)(*)()(maxSifPfNforpfNfIfTfTref ii (4-5) where, ( i)max(Tref) is the multi-species phytoplankton maximum growth rate at a particular reference temperature Tref under optimal conditions of saturated light intensity and excess nutrients (1/day); T is temperature ( C); I is the light intensity, obtained by the light attenuation model (C hristian, 2001); N is nitrogen ; P is phosphorous; and Si is silica. The maximum algal growth rates must be specified at a reference temperature Tref which is consistent with the particular temper ature function, f (T), used in the model. The reference temperature may correspond to 20 C, the optimum temperature condition, or other temperature conditions, depending on th e form of the temperature function. Therefore, maximum growth rate coefficients obtained from one model may have to be adjusted before using the coefficients in the other model, which has a different

PAGE 54

31 temperature adjustment functi on. Maximum growth rates for algae are tabulated in Table 4-1, along with the corresponding reference temperatures (Bowie et al., 1985). Table 4-1. Algal ma ximum growth rates (B owie et al., 1985) Algal Type Maximum Growth Rate (1/day) Reference Temperature ( C) Total phytoplankton 0.2 8.0 20 C or Topt Diatoms 0.55 5.0 20 C or Topt Green algae 0.7 3.0 20 C or Topt Blue green algae 0.41 2.5 20 C or Topt Flagellates 1.2 1.6 20 C or Topt Dinoflagellates 0.2 2.1 20 C or Topt Canale and Vogel (1974) developed a se t of temperature-growth curves for diatoms, green algae, blue-green algae, and flagellates based on a literature review of growth data for many species (Figure 4-1). The diatom grow th rate sharply falls as temperature rises over 30 C. Therefore, high water temperature in the system may cause decreased diatom biomass. Figure 4-1. Temperature growth curves for major algal gr oups (Canale and Vogel, 1974)

PAGE 55

32 Algal growth rate is limited by nutrient av ailability depending on algal species. For example, silicon limitation affects only diat oms. Nitrogen limitati on can generally be omitted for nitrogen-fixing blue-green algae. The specific nutrient limitation factors are (Bowie et al., 1985): SiH Si Sif NONHH NONH Nf SRPH SRP PfSi N P )(, )(, )(3 4 3 4 (4-6) where SRP is soluble reactive phosphorous concentration, HP is half saturation concentration for phosphorous uptake, NH4 is ammonium concentration, NO3 is nitrate concentration, HN is half saturation concentration fo r nitrogen uptake, Si is available dissolved silicon concentration, and HSi is half saturation concentration for silica uptake. Values of the half-saturation constants for each limiting nutrient are different as algal groups. Literature values of the half satu ration constants are s hown in Table 4-2. Table 4-2. Half saturation constants fo r each limiting nutrient (Bowie et al., 1985) Half saturation constants Nitrogen Phosphorus Silica Total Phytoplankton 0.01 0.4 mg/L 0.0005 0.08 mg/L Diatom 0.015 0.12 mg/L 0.001 0.163 mg/L 0.03 0.1 mg/L Green algae 0.005 0.15 mg/L 0.001 0.15 mg/L Blue green algae 0.0 4.34 mg/L 0.0025 0.06 mg/L Flagellates 0.001 0.13 mg/L 0.012 mg/L Dinoflagellates 0.005 0.13 mg/L 0.06 mg/L Respiration and non-predator mortality of each phytoplankton are considered exponentially increasing functions of temperature (Bowie et al., 1985): TrT axTri i TrT asTri iKax Kax KasKas *)( *)( (4-7) where i represents multi-species phytoplankton, (Kasi)Tr and (Kaxi)Tr are respiration and non-predator mortality rate at Tr (1/day), Tr is reference temperature of respiration and

PAGE 56

33 non-predator mortality, and as and ax are temperature adjustment coefficients of respiration and non-predator mortality, respectively. Phytoplankton settling velocity accounts for limited vertical motion of its organisms. The settling rates depend on dens ity, size and shape of each species. Table 4-3 shows literature values of settling velocity as well as respiration and non-predatory mortality. Table 4-3. Settling velocity, respirati on, and non-predatory mortality for each phytoplankton (Bowie et al., 1985) Settling Velocity Respiration & Excretion Non predatory Mortality Total Phytoplankton 0.0 40.0 m/day 0.005 0.2 1/day 0.003 0.17 1/day Diatom 0.02 17.1 m/day 0.04 0.59 1/day 0.03 1/day Green algae 0.05 0.89 m/day 0.01 0.46 1/day Blue green algae 0.0 0.2 m/day 0.03 0.92 1/day Flagellates 0.05 0.39 m/day 0.05 0.06 1/day Dinoflagellates 2.8 6.0 m/day 0.047 1/day Zooplankton dynamics are closely related to phytoplankton dynamics through predator-prey inte raction. Zooplankton is only co nsidered as the predator of phytoplankton, utilizing its av ailable biomass as a food s upply. Zooplankton uptake of each algal group is proportional to the biomass of each species. There is no consideration of preference uptake. Zooplankton kinetics are influenced by grow th, respiration and mortality (Bowie, 1985). ZOOC KzsKzx t ZOOCz* (4-8) where z is zooplankton growth rate (1/day), Kzx is respiration rate of zooplankton (1/day), and Kzs is mortality rate of zooplankton (1/day).

PAGE 57

34 Zooplankton growth is represented by the temperature-dependent maximum growth rate and phytoplankton biomass. PHY PHY PHY T z zzTrsPHY H TrsPHY **)(20 max (4-9) where ( z)max is the zooplankton maximu m growth rate (1/day), is temperature adjustment coefficient, HPHY is half saturation concentration for phytoplankton uptake (gC-3), and TrsPHY is threshold phytoplankton concentration for zooplankton uptake ( g/l). 4.3.2 Oxygen Balance Living organisms in water are dependent on dissolved oxygen (DO) to produce energy for growth and reproduction. The source of DO is from the atmosphere (diffusion and mixing at air-water interface), rainfa ll, and photosynthesis by algae. Oxygen is consumed by microbial respiration, nitrificat ion, and oxidation of organic matter in water and bottom sediments. In addition, vertical and horizontal disp ersion and diffusion are crucial mechanisms influencing DO concentration. The formation of the oxygen balance in this study is based on that of the WASP5 and EFDC models (Ambrose et al., 1994; Park, et al ., 1995). The oxygen balance includes the following processes: reaeration, oxidation, nitrification, sediment oxygen demand (SOD), photosynthesis, a nd respiration (Figure 4-2). PHYCAocKasKax P H SOD NH DOH DO K CBOD DOH DO KDODO z K t DOan NIT NN CBOD D s AE**] *)*3.03.1[( 14 64 ) (4 (4-10)

PAGE 58

35 where DOs is the equilibrium oxygen concentration at standard pressure (mg/l), KAE is the reaeration coefficient, KD is the oxidation coefficient, HCBOD is a half saturation constant for oxygen limitation, CBOD is carbonaceous biochemical oxygen demand, KNN is nitrification rate, HNIT is a half saturation constant for oxygen limitation of nitrification, H is water depth (m), Pn is the preference for ammonia uptake, and AOC is oxygen-tocarbon ratio (gO2/gC). The rate of DO production is assumed to be proportional to the growth rate of the phytoplankton in a fixed stoichiometry reac tion. An additional source of oxygen from phytoplankton growth occurs when the availa ble ammonia nutrient source is exhausted and the phytoplankton begins to utilize the available nitrate (Ambrose et al., 1994). To quantify the rate of oxidation, CBOD is us ed as a measure of the quantity of oxygen demanding materials. This approach simplifies modeling efforts by consolidating their potential efforts (Ambrose et al., 1994). The CBOD kinetic equation comprises oxidation of organic matter, denitrification, nonpredatory mortality, and respiration by zooplankton and phytoplankton (Figure 4-2). )* *(* 14 32 4 5 ]*)1(*[3 3 3ZOOCKzs PHYCKasAocNO DOH H K CBOD DOH DO KCBOD fd ws zt CBODno no DN CBOD D CBOD CBOD (4-11) where wsCBOD is the settling velocity for th e particulate fraction of CBOD, fdCBOD corresponds to the fraction of the dissolved CBOD, KDN is denitrification rate, and Hno3 is a half saturation constant for oxyge n limitation of denitrification. Reaeration, the process of oxygen exchange between the atmosphere and sea surface, is a source of oxygen in the water because dissolved oxygen levels in most

PAGE 59

36 natural waters are below saturation. However, when water is super-satur ated as a result of photosynthesis, dissolved oxygen returns to the atmosphere. Details of the dissolved oxygen saturation and reaeration coefficient calculations can be found in Appendix C. Figure 4-2. Dissolved oxygen and carbon aceous biochemical oxygen demand cycles

PAGE 60

37 Oxygen loss occurs as a result of nitrification and SOD. Nitrification is a biological process; toxic ammonia nitrogen oxidizes to less harmful nitrate. Even though this process consumes oxygen, it can re duce the potential threat to fish health. SOD is due to the oxidation of organic matter in bottom sediments. These benthic deposits derive from wastewater particulate, plant matter, alloch thonous particulate, a nd particulate organic matter. The main source is particulate orga nic matter in the water. Therefore, the relationship between areal SOD rate and particulate organic matter concentration requires further development. Chapra (1997) suggested a "naive" Street er-Phelps SOD model, which yields the relationship between SOD rate and particulate CBOD concentration in the water. In this study, the "naive" Streeter-Phelps SOD model has been incorporated into the CH3DIMS; SOD values change in response to al terations of particulate CBOD concentration and settling velocity. The segmentation scheme for the model is depicted in Figure 4-3. BOD CBOD NBOD DOsed Kd2 KN2Ks DOwater SODWater Sediments Figure 4-3. A "naive" Streeter-Phe lps model of SOD (Chapra, 1997)

PAGE 61

38 As depicted in Figure 4-3, SOD is direct ly associated with settling particulate organic matter, that is, particulate CBOD. CBO D is changed by external loads, oxidation, mortality of zooplankton and phytoplankton, a nd settling velocity. If external loads increase, phytoplankton and zooplankton biomass augments due to the supply of enough nutrients. Increased phytopla nkton and zooplankton biomass creates more CBOD with their mortality. In addition, CBOD is added by external loads, such as river runoff. Increased particulate CBOD consumes more dissolved oxygen at the sediment-water interface as SOD. However, if SOD represents a constant valu e or has a spatially variable value, consumption of dissolved oxygen as SOD will be the same value at the sedimentwater interface no matter how ma ny external loads enter into the estuary. The method of using SOD as a constant is not suitable to simulate hypoxia to be changed in external loads. The "naive" Streeter-Phelps SOD model implemented into CH3D-IMS is a linear function of particulate CBO D and settling velocity. DOK DO ST CBODv HH SODSOD T pw s *****11.1* 120 (4-12) where vs is settling velocity of particulate CBOD (m/day); CBODpw is particulate CBOD concentration in the water (mg/L); ST is a fr actional coefficient for sediment type (When sediment type 2, ST=1.3, when sediment type>3, ST=1.0) (Park, 2004); is temperature coefficient, which Zison et al. (1978) have reported a range of 1.04 to 1.13 for while a value 1.065 is commonly employed; and KSOD is half saturation rate for SOD, which Lam et al. (1983) have s uggested a value of KSOD (1.4 mg/l) for Lake Erie as well as Thomann and Mueller (1987) have proposed a value of 0.7mg/l.

PAGE 62

39 Particulate CBOD is directly related to external nutrient loads and indirectly associated with phytoplankton biomass, algae growth rate, particulate nutrient dynamics, and other changed nutrients. CH3D-IMS with new SOD model shows that diminished particulate CBOD brings less SOD at the sediment-water interface. The bottom oxygen will be increased because of reduced SOD. Detailed model explanation can be found in Appendix D. 4.3.3 Nutrient Dynamics Nutrients play a critical role in life pr ocess of aquatic organisms as well as eutrophication in natural water. The major nutrients of concern are carbon, nitrogen, phosphorous, and silica. In general nitr ogen, phosphorous, and silica are the major nutrients regulating the ecological balance in an estuarine system except carbon. Even though in same cases it may be limiting, carbon is usually in excess (Bowie et al., 1985). While nitrogen is seen to be an essent ial component, excessive concentrations of certain nitrogen species lead to environmen tal problems. For example, high concentration of ammonia can be toxic to fish and othe r aquatic organisms as well as reduce the dissolved oxygen levels in rivers and estuaries (Sawyer et al., 2003). In CH3D-IMS, nitrogen species cons ist of dissolved amm onium nitrogen (NH4), ammonia nitrogen (NH3), nitrate and nitrite (NO3), soluble organic nitrogen (SON), particulate organic nitrogen (PON), and particulate inorga nic nitrogen (PIN) in Figure 4-4. Phosphorous is considered the primary nut rient limiting the biomass in many fresh waters and is responsible for observed cultura l eutrophication of lake s (Morris, 1980). In addition, coastal areas of the eas tern United States which are now in a state of nitrogen limitation may have been in a state of phosphorous limitation depending on phosphorous and nitrogen ratio (Ryther & Dunstan, 1971). The phosphorous cycle in the water quality

PAGE 63

40 model contains soluble reactive phosphorous (SRP), soluble organic phosphorous (SOP), particulate organic phosphorous (POP), and pa rticulate inorganic phosphorous (PIP) in Figure 4-5. Details on the nitrogen and phos phorous dynamics including all equations used by water quality model for algal groups can be found in Appendix E. NH3 NH4 PIN SON PON Diatom Dinoflagellate NO3 Zooplankton Cyanobacteria Sett ling sett ling Uptake Excretion Uptake Uptake Excretion Excretion Mortality Excretion Predation NITROGEN CYCLEMortality Mortality Mortality Predation Predation Volatilization Instalbility Nitrification Mortality Mortality Aerobic Layer Anaerobic Layer Water Column NH4 PIN SON PON NO3Nitrification NH3 Instalbility NO3 N2 Denitrification Figure 4-4. Schematic of Nitrogen Cycle Silica is an essential element for the skeletal growth of diatoms, radiolarians, and certain sponges. Reduction of silica in na tural water can limit phytoplankton production. Spring diatom blooms may be co rrelated with silica availabil ity because silica is critical

PAGE 64

41 to the cell-wall formation of diatoms (Kenni sh, 1986). In this study, the silica cycle is incorporated into CH3D-IMS to better unders tand nutrient dynamics in Figure 4-6. Silica can be classified into two groups: soluble organic silica (SOS) and particulate organic silica (POS). PO4 PIP SOP POP Diatom Dinoflagellate Zooplankton Cyanobacteria Sett ling sett ling Uptake Excretion Uptake Uptake Excretion Excretion Mortality Excretion MortalityPredationPHOSPHOROUS CYCLEPredation PredationMortality Mortality Mortality Mortality Aerobic Layer Anaerobic Layer Water Column PO4 PIP SOP POP PO4 PIP SOP POP Figure 4-5. Schematic of Phosphorous Cycle In the finite difference solution of the silica cycle, the horizontal diffusion and advection terms are treated explicitly, wher eas the vertical diffusion and biochemical

PAGE 65

42 processes are implicit. By solving the sorp tion/desorption terms separately from other terms, it is possible to treat terms implic itly. Equation 4-13 shows a schematic of the numerical solution algorithm method. 1 1 21 2 2 12 1... n n n n n n nn n nnn Desporptio Sorption t Q Term Diffusion Vertical t DiffHADVADH t (4-13) DOS POS Diatom Zooplankton sett ling uptake (growthDia*DiatomS) excretion (Kax*DiatmoS) mortality (Kzs*ZOOS) excretion (Kzx*ZOOS) Mortality DiatomSPredation *ZOOS DOS POS DOS POS Sorption/Desorption Sorption/Desorption Sorption/DesorptionAerobic Layer Anaerobic Layer Water Column SILICA CYCLE Figure 4-6. Schematic of Silica Cycle

PAGE 66

43 where H.AD is horizontal advection term, V. AD is vertical advection term, H.Diff is horizontal diffusion term, Q is biochemical reaction term, and is concentration of dissolved or particulate silica. The mass balance equations for SOS and POS are written by combining biochemical processes. (1) Water Column POS WS z ZOOSKzs Diatom KasSOScpPOSd t POS ZOOSKzx Diatom Kax Diatom SOScpPOSd t SOSPOS DiatomS os os DiatomS DiatomS os os* )** (* ** (4-14) (2) Sediment Column )** (* ** SOScpPOSd t POS SOScpPOSd t SOSos os os os (4-15) where dos is the desorption rate of POS (1/day), pos is the partition coefficient between POS and SOS (1/ g), DiatomS is growth rate of diatom (1/day), KaxDiatomS is respiration rate of diatom (1/day), Kzx is respiration rate of zooplankton (1/day), ZOOS is the biomass of zooplankton incepted diatom, KasDiatomS is non-predator rate of diatom (1/day), Kzs is mortality rate of zooplankton (1/day), WSpos is the phytoplankton settling velocity (cm/s), and c is sediment concentration in the water column.

PAGE 67

44 4.3.4 Water Quality Model Parameters CH3D-IMS requires many model parameters to embody biogeochemical processes in an estuarine system. It is very difficult to determine model parameters because they are dependent on physical and biochemical factor s, such as the type of the estuary, circulation, waves, tidal va riation, temperature, light, hydrostatic pressure, salinity, benthic nutrient concentrations property of sediment, point or non-point source loadings of nutrients, and other chemical materials. In general, parameters are selected from a range of feasible values: field observation, laboratory experimentation, and adjustment until an optimal agreement between simulate d and measured values. A list of model parameters and their literature values are described in Table 4-4. Table 4-4. Water quality model parameters Parameter Description Unit Literature range Source ( dia)T-20 Temperature coefficient for diatom growth (1.01 1.2)** Di Toro et al. (1980) ( dino)T-20 Temperature coefficient for dinoflagellates growth (1.01 .2)** Di Toro et al. (1980) ( cyano)T-20 Temperature coefficient for cynobacteria growth (1.01 .2)** Di Toro et al. (1980) ( AD)T-20 Temperature coefficient for NH4 desorption 1.08 Sheng et al. (2003c) Park (2004) ( AI)T-20 Temperature coefficient for Ammonium instability 1.08 Sheng et al. (2003c) Park (2004) ( BOD)T-20 Temperature coefficient for CBOD oxidation 1.02 1.15 Bowie et al. (1985) ( DN)T-20 Temperature coefficient for denitrification 1.02 1.09 Bowie et al. (1985) ( NN)T-20 Temperature coefficient for nitrification 1.02 1.08 Bowie et al. (1985) ( OD)T-20 Temperature coefficient for SON desorption 1.08 Sheng et al. (2003c) Park (2004) ( ONM)T-20 Temperature coefficient for mineralization 1.02 1.09 Bowie et al. (1985)

PAGE 68

45 Table 4-4. Continued Parameter Description Unit Literature range Source ( RESP)T-20 Temperature coefficient for algae respiration 1.045 Bowie et al. (1985) ( z)T-20 Temperature coefficient for zooplankton growth 1.01 1.2 Di Toro et al. (1980) dia Diatom maximum growth rate 1/day 0.55 5.0 Bowie et al. (1985) dino Dinoflagellates maximum growth rate 1/day 0.2 2.16 Bowie et al. (1985) cyano Cyanobacteria maximum growth rate 1/day 0.2 4.9 Bowie et al. (1985) z Zooplankton maximum growth rate 1/day 0.1 0.3 Bowie et al. (1985) achla Algal carbon and chlorophyll a ratio mg C / mg Chla 20 1000 Jin et al. (1998) anc Algal nitrogen and carbon ratio mg N / mg C 0.05 0.43 Bowie et al. (1985) apc Algal phosphorous and carbon a ratio mg P / mg C 0.005 0.03 Bowie et al. (1985) aoc Algal oxygen and carbon ratio mg O / mg C 2.67 Ambrose (1991) adsc Diatom silica and carbon ratio mg Si / mg C 0.06 0.77 Bowie et al. (1985) dan Desorption rate of adsorbed ammonium nitrogen 1/day 0.01 0.02 Sheng et al. (2003c) Park (2004) don Desorption rate of adsorbed organic nitrogen 1/day 0.005 0.08 Bowie et al. (1985) Park (2004) dos Desorption rate of adsorbed organic silica 1/day 0.005 0.1 Bowie et al. (1985) dip Desorption rate of adsorbed inorganic phosphorous 1/day 0.01 0.02 Sheng et al. (2003c) Park (2004) dop Desorption rate of adsorbed organic phosphorous 1/day 0.01 0.08 Bowie et al. (1985) Park (2004) dmol Molecular diffusion coefficient for dissolved species cm2/s 4.E-6 1.E-5 Rao et al. (1984) fdCBOD Fraction of dissolved CBOD 0.5 Ambrose (1991) Hbod Half-saturation constant for CBOD oxidation mg O2 0.02 5.6 Bowie et al. (1985)

PAGE 69

46 Table 4-4. Continued Parameter Description Unit Literature range Source Hn_dia Half-saturation constant for diatom uptake nitrogen mg/l 0.015 0.12 Bowie et al. (1985) Hn_dino Half-saturation constant for dinoflagellates uptake nitrogen mg/l 0.005 0.13 Bowie et al. (1985) Hn_cyano Half-saturation constant for cyanobacteria uptake nitrogen mg/l 0.01 4.34 Bowie et al. (1985) Hp_dia Half-saturation constant for diatom uptake phosphorous mg/l 0.001 0.163 Bowie et al. (1985) Hp_dino Half-saturation constant for dinoflagellates uptake phosphorous mg/l 0.06* (0.001 0.03)** Bowie et al. (1985) Hp_cyano Half-saturation constant for cyanobacteria uptake phosphorous mg/l 0.0025 0.02 Bowie et al. (1985) Hs Half-saturation constant for diatom uptake silica mg/l 0.08 0.1 Bowie et al. (1985) Ha Half-saturation constant for zooplankton mg/l 0.01 2.0 Bowie et al. (1985) hv Henrys constant mg/l atm 43.8 45 Sawyer et al. (2003) Idia Optimum light intensity for diatom growth E/m2/s 88 350 Bowie et al. (1985) Idino Optimum light intensity for dinoflagellates growth E/m2/s (300 350)** Bowie et al. (1985) Icyano Optimum light intensity for cyanobacteria growth E/m2/s 43 600 Bowie et al. (1985) (NH3)air Ammonia concentration in the air g/l 0.1 Freney et al. (1981) Kax_dia Respiration rate by diatom 1/ day 0.03 0.6 Bowie et al. (1985) Kax_dino Respiration rate by dinoflagellates 1/day 0.047 Bowie et al. (1985) Kax_cyano Respiration rate by cyanobacteria 1/day 0.03 0.9 Bowie et al. (1985) Kas_dia Mortality rate by diatom 1/day 0.03 (0.003 0.1)** Bowie et al. (1985) Kas_dino Mortality rate by dinoflagellates 1/day (0.003 0.1)** Bowie et al. (1985) Kas_cyano Mortality rate by cyanobacteria 1/day (0.003 0.1)** Bowie et al. (1985) KD CBOD oxidation rate 1/day 0.02 0.6 Bowie et al. (1985)

PAGE 70

47 Table 4-4. Continued Parameter Description Unit Literature range Source KDN Denitrification rate constant 1/day 0.0 1.0 Bowie et al. (1985) KNN Nitrification rate constant 1/day 0.004 0.11 Bowie et al. (1985) Kon Ammonification rate of SON 1/day 0.001 0.4 Bowie et al. (1985) Kop Mineralization rate of SOP 1/day 0.001 0.6 Bowie et al. (1985) Kvol Constant rate for nitrogen volatilization 1/day 3.5 9.0 Fillery et al. (1986) Kzx Respiration rate of zooplankton 1/day 0.001 0.16 Bowie et al. (1985) Kzs Mortality rate of zooplankton 1/day 0.0065 0.0326 Li et al. (1985) Pan Partition coefficient between NH4 and PIN 1/ g 1.E-5 1.E-6 5E-3 Simon (1989) Sheng et al. (2003c) Pon Partition coefficient between SON and PON 1/ g 5.E-6 1.E-5 3.E-5 5.E-3 Simon (1989) Sheng et al. (2003c) Pip Partition coefficient between SRP and PIP 1/ g 1.E-6 1.E-3 Sheng et al. (2003c) Pop Partition coefficient between SOP and POP 1/ g 8.E-6 5.E-3 Sheng et al. (2003c) WSdia Diatom settling velocity cm/day 1 1700 Bowie et al. (1985) WSdino Dinoflagellates settling velocity cm/day 280 600 (0 3000)** Bowie et al. (1985) WScyano Cyanobacteria settling velocity cm/day 0 20 Bowie et al. (1985) ** All phytoplankton 4.4 Light Attenuation Model A light attenuation model, based on physics, was developed as part of CH3D-IMS (Christian, 2001) and applied to the Indian River Lagoon (S heng et al., 2003c; Christian and Sheng, 2003) as well as Charlotte Har bor (Park, 2004). Light is considered an important factor controlling the health of seagrass beds as well as phytoplankton photosynthesis. The main purpose of this model calculates photosynthesis active

PAGE 71

48 radiation (PAR), that is, li ght. The model equation is base d on the Beer-Lambert equation (Dennison et al., 1993). zPARKIzId*)(exp*)(0 (4-16) where z is the depth below the water su rface, I(z) is the PAR at depth z, I0 is the PAR just below the water surface, and Kd (PAR) is the light attenuation coefficient for downward PAR. More information on this model can be found in Appendix F.

PAGE 72

49 CHAPTER 5 MODEL APPLICATION OF CIRCULATION AND TRANSPORT TO UPPER CHARLOT TE HARBOR The baroclinic simulation of Upper Charlotte Harbor, FL was performed from June 12th, 2003 to July 11th, 2004 using CH3D-IMS described in Chapter 3. The main objectives of this simulation were the calibra tion and validation of the Upper Charlotte Harbor circulation and transport model using measured data as well as the establishment of Freshwater Flows and Levels (FFL). This FFL will serve to preserve and protect the ecosystem of this study area. To allow sufficien t spin-up to remove the transient effect of initial condition, the model simulation ac tually started in May, 2003. We conducted sensitivity studies to determine the optimal model coefficients, such as horizontal diffusivity, bottom roughness, surface wind drag coefficient, vertical layers, and other coefficients, to minimize the difference between measured data and model predictions. The time step used in the model simulation was 90 seconds. 5.1 A High-Resolution Curvilinear Grid A boundary fitted grid, which contains 188 x 176 horizontal cells and eight vertical layers, was used for this study (Figure 5-1) The bathymetry of the Charlotte Harbor estuarine system except the Upper Charlotte Ha rbor and the near shor e region of the Gulf of Mexico was extracted from the Geophysical Data System of the National Geophysical Data Center (Figure 52). Bathymetry for the Upper Char lotte Harbor area was collected by University of South Florida (Wang, 2004). An inverse distance interpolation followed by a simple smoothing scheme was the primary method for determining bathymetry.

PAGE 73

50 Easting(UTM,meter) Northing(UTM,meter) 350000 400000 450000 2.9E+06 2.95E+06 3E+06 Figure 5-1. Boundary-fitted grid (188 by 176) used for the study 5.2 Measured Data The Coastal Engineering Laboratory of the Civil and Coastal Engineering Department, University of Florida, installed an instrumentation tower (UF station) in the Upper Charlotte Harbor at Latitude 26 52' 30'' N and Longitude 82 9' E (Figure 5-3) and collected continuous hydrodynamic data between June 12th, 2003 and July 15th, 2004. The instruments included thr ee conductivity/temperature sensors, one pressure transducer, and one Acoustic Doppler Current Profiler (ADCP). Data were collected every 30 minutes. The depth relative to NAVD 88 at the UF station is 3.37 meters as obtained from 5 hours of GPS measurements. The error range is within 4.6 cm. Water

PAGE 74

51 temperature and conductivity were measured at three vertical la yers: lower (77.5cm above bottom), middle (176.5cm above bo ttom), and upper (263cm above bottom). Easting(UTM,meter) Northing(UTM,meter) 350000 400000 450000 2.9E+06 2.95E+06 3E+062000 1900 1800 1700 1600 1500 1400 1300 1200 1100 1000 900 800 700 600 500 400 300 200 100 0 Bathymetry(cm) Figure 5-2. Bathymetry in boundary-fitted grid The conductivity/temperature sensors were cl eaned periodically in order to ensure accurate data (Table 5-1). While we origin ally envisioned a cleaning interval of one month, we found out that 1-2 week cleaning interval was necessary due to the warm water and excessive growth of barnacles. This created fouling problem for the conductivity sensors and caused major damage to some of the sensors. The result is

PAGE 75

52 missing salinity data during periods of excessi ve barnacle growth a nd periods of sensor repair. The periods with good da ta are shown in Table 5-2. Easting(UTM,meter) Northing(UTM,meter) 370000 380000 390000 400000 410000 2.96E+06 2.98E+06550 525 500 475 450 425 400 375 350 325 300 275 250 225 200 175 150 13,34 ELJobean HBMPBathymetry(cm)HarborHeight 19,53 22,47 UF 19,43 PuntaGorda Figure 5-3. Locations of UF, Punta Gorda, Harbor Height, El Jobean, HBMP and Mote Marine recorders The ADCP transducers collected current da ta at six vertical layers: layer one (105.5cm above bottom), layer two (150.5cm above bottom), layer three (195.5cm above bottom), layer four (240.5cm above bottom), layer five (285.5cm above bottom), and layer six (330.5cm above bottom). Measured data were stored every 30 minutes. However, data at layer five and six were not useful because water level at the UF station was lower than those layers.

PAGE 76

53 The US Geological Survey (USGS) measured water elevation and salinity at three stations, El Jobean near th e Myakka River, Punta Gorda along the Peace River, and Harbor Height in the mouth of Peace River. Th e El Jobean station is located at Latitude 26 57' 28'' N and Longitude 82 12' 46'' E, the Punta Gorda station is at Latitude 26 56' 37'' N and Longitude 82 3' 31'' E, and the Ha rbor Height station is located at Latitude 26 59' 14'' N and Longitude 81 59' 40'' E (Figure 5-3). Table 5-1. Dates when conductivity /temperature sensors were cleaned Date Responsible Agency Activity May 8th, 2003 University of Fl orida Installed sensors June 11th, 2003 University of Florida Cleaned sensors August 7th, 2003 University of Florida Cleaned sensors September 4th, 2003 Mote Marine Laboratory Cleaned sensors September 18th, 2003 University of Florida Cleaned sensors October 2nd, 2003 Mote Marine Laboratory Cleaned sensors October 16th, 2003 University of Florida Cleaned sensors November 3rd, 2003 Mote Marine Laboratory Cleaned sensors November 20th, 2003 University of Florida Cleaned sensors December 2nd, 2003 Mote Marine Laboratory Cleaned sensors December 22nd, 2003 University of Florida Cleaned sensors January 5th, 2004 Mote Marine Laboratory Cleaned sensors January 21st, 2004 University of Florida Cleaned sensors February 3rd, 2004 Mote Marine Laboratory Cleaned sensors February 5th, 2004 University of Florida Cleaned & checked sensors February 17th, 2004 University of Florida Cleaned sensors March 3rd, 2004 Mote Marine Laboratory Cleaned sensors April 1st, 2004 University of Florida Cleaned sensors April 29th, 2004 University of Florida Cleaned sensors May 19th, 2004 University of Florida Cleaned sensors June 1st, 2004 Mote Marine Laboratory Cleaned sensors June 9th, 2004 University of Florida Cleaned sensors June 18th, 2004 Mote Marine Laboratory Cleaned sensors June 23rd, 2004 University of Florida Cleaned sensors June 30th, 2004 Mote Marine Laboratory Cleaned sensors July 7th, 2004 University of Florida Cleaned sensors July 15th, 2004 Mote Marine Laboratory Cleaned sensors Not only continuous data but also periodic da ta were collected in the estuary. The Peace River Hydrobiological Monitoring Program (HBMP) monthly collected salinity

PAGE 77

54 data from surface to bottom at Latitude 26 53' 56'' N and Longitude 82 07' 15'' E (Figure 5-3). Salinity data were also obtained for four locatio ns within Charlotte Harbor that were sampled by Mote Mari ne Laboratory. They divided th e Harbor into a series of small grids and collected data at (19, 53), (22, 47), (19, 43) and (13, 34) in Figure 5-3. Table 5-2. Time peri ods with good field data Measured parameters (every 30 minutes) The period of good data Water level obtained by pressure transducer June 12th, 2003 to July 15th, 2004. Current obtained by ADCP June 12th, 2003 to September 18th, 2003 October 10th, 2003 to July 15th, 2004 Salinity/temperature (surface) obtained by conductivity/temperature sensors June 12th, 2003 to July 9th, 2003 August 8th, 2003 to November 14th, 2003 January 21st, 2004 to July 15th, 2004 Salinity (middle) obtained by conductivity/temperature sensors June 12th, 2003 to July 9th, 2003 August 8th, 2003 to October 21st, 2003 January 21st, 2004 to July 15th, 2004 Salinity (bottom) obtained by conductivity/temperature sensors June 12th, 2003 to June 25th, 2003 August 8th, 2003 to September 4th, 2003 October 15th, 2003 to July 15th, 2004 Temperature (middle/bottom) obtained by conductivity/temperature sensors June 12th, 2003 to July 9th, 2003 August 8th, 2003 to July 15th, 2004 5.3 Initial and Boundary Conditions Hydrodynamic monitoring of Charlotte Ha rbor has been conducted by the South Florida Water Management District (SFWMD), the Southwest Florida Water Management District (SWFWM D), the National Oceanic & Atmospheric Administration (NOAA), the National Data Buoy Center ( NDBC) and the United States Geological Survey (USGS). Water level data at offshore boundary locati ons come from tidal constituents of the Advanced Circulation (ADCIRC) model a nd non-tidal components of the Naples and Clearwater station (NOAA). The ADCIRC model is a two-dimens ional, depth-integrated, barotrophic time-dependent long wave, hydrodyna mic circulation model. This model can

PAGE 78

55 provide tide and storm surge elevations corresponding to each node over the Gulf of Mexico that covers our offs hore grid. An interpolation method was employed to obtain tidal information for our open boundary grid s because the nodes of the ADCIRC model didnt exactly match with our boundary grids. The surface wind data were imposed usi ng hourly wind magnitude and direction data from the Venice station, the UF stat ion in the Upper Charlotte Harbor, and the Naples station (Table 5-3). Figure 5-4 shows wind speed and directi on at those stations. These values were interpolated to the entire computational grid w ith a weighting function inversely proportional to the square of distance. Table 5-3. Descript ion of wind stations Name Agency Latitude Longitude Period Venice NDBC 27 04' 01" 82 27' Hourly UF University of Florida 26 52' 30" 82 09' Hourly Naples NOAA 26 07' 05" 81 48' 25" Hourly Discharge and runoff boundary conditions were based on daily measured discharge data at Myakka River (near Sarasota and at Big Slough Canal), Peace River (at Arcadia, at Joshua Creek and at Horse Creek), Shell Creek, and Caloosahatchee River (Table 5-4). Figure 5-5 demonstrates the ri ver discharges from these ri vers during the year 2003 and 2004. Table 5-4. Descripti on of river stations Station Name Agency Latitude Longitude Period 02298830 02299410 Myakka River near Sarasota at Big Slough Canal USGS USGS 27 14' 25" 27 11' 35" 82 18' 50" 82 08' 40" Daily Daily 02296750 02297100 02297310 Peace River At Arcadia At Joshua Creek At Horse Creek USGS USGS USGS 27 13' 19" 27 09' 59" 27 11' 57" 81 52' 34" 81 52' 47" 81 59' 19" Daily Daily Daily 02298202 Shell Creek USGS 26 59' 04" 81 56' 09" Daily S29_S Caloosahatchee River SFWMD 26 43' 26" 81 41' 54" Daily

PAGE 79

56 Y ea r Da y 2003&2004 200 300 400 500 600 5m/s UFstation Y ea r Da y 2003&2004 200 300 400 500 600 5m/s Venice Y ea r Da y 2003&2004 200 300 400 500 600 5m/s Naples Figure 5-4. Wind speed at UF Venice and Naples stations YearDay2003&2004 RiverDischarge(m3/s) 100 200 300 400 500 0 100 200 300 400 500 600 700 PeaceRiver ShellCreek MyakkaRiver CaloosahatcheeRiver Figure 5-5. River discharge at the Peace River, Shell Creek, Myakka River, and Caloosahatchee River

PAGE 80

57 Precipitation values for the model were ba sed on daily precipitation data recorded at SWFWMD rain gauges at 6 locations in the study area (Figures 5-6a and 5-6b). The same wind interpolation method was used to ge t values for the entire computation grid. The SFWMD daily pan evaporation data (Penma n, 1956) measured near Naples, FL were used for model simulation in Figure 5-7. Ta ble 5-5 shows information of precipitation and evaporation stations. Table 5-5. Description of preci pitation and evaporation stations Station Agency Latitude Longitude Data Type Period 211 SWFWMD 26 59' 03" 81 56' 05" Precipitation Daily 420 SWFWMD 26 58' 38" 81 56' 10" Precipitation Daily 421 SWFWMD 26 56' 39" 82 13' 06" Precipitation Daily 502 SWFWMD 26 50' 27" 81 58' 56" Precipitation Daily 505 SWFWMD 26 55' 32" 82 19' 47" Precipitation Daily 527 SWFWMD 27 01' 53" 82 00' 27" Precipitation Daily BCBNAPLE_E SFWMD 26 43' 26" 81 41' 54" Evaporation Daily YearDay2003&2004 Rainfall(cm) 100 200 300 400 500 600 0 5 10 15 20Station211 Y earDa y 2003&2004 Rainfall(cm) 100 200 300 400 500 600 0 5 10 15 20Station420 Y earDa y 2003&2004 Rainfall(cm) 100 200 300 400 500 600 0 5 10 15 20Station421 Figure 5-6. Precipitatio n data at 211, 420, 421, 502, 505 and 537 stations A

PAGE 81

58 Y ea r Da y 2003&2004 Rainfall(cm) 100 200 300 400 500 600 0 5 10 15 20Station527 Y ea r Da y 2003&2004 Rainfall(cm) 100 200 300 400 500 600 0 5 10 15 20Station505 Y ea r Da y 2003&2004 Rainfall(cm) 100 200 300 400 500 600 0 5 10 15 20Station502 Figure 5-6. Continued YearDay2003&2004 Evaporation(cm) 100 200 300 400 500 600 0 0.2 0.4 0.6 0.8 1 Figure 5-7. Evaporatio n at Naples stations Air temperature measured at the Venice station (NDBC), the UF station in the Upper Charlotte Harbor, and the Naples stat ion (NOAA) were used for model simulation (Figure 5-8). Data sources were the same as those of wind data. These values were interpolated the same way as wind values were used. B

PAGE 82

59 YearDay2003&2004 AirTemperature(Celcius) 200 300 400 500 0 5 10 15 20 25 30 35 40 UF Venice Naples Figure 5-8. Air temperatures at UF, Venice and Naples stations 5.4 Model Results 5.4.1 Water Level Figures 5-9 to 5-12 show the comparison between measured and simulated water levels at the UF, Punta Gorda, El Jobean, a nd Harbor Height stati ons, respectively. The Root Mean Square (RMS) error was calculated to measure statistica lly the magnitude of differences between measured and simulate d values. The RMS error is defined as: N value measured value simulated RMS2) ( where N is total number of data points. The RMS errors for water level are 7.71 cm at UF, 7.16 cm at Punta Gorda, 7.75 cm at El Jobean, and 8.12 cm at Harbor Height. The RMS errors normalized by the tidal range are less than 8.0 % at three stations. Tables 5-6 to 5-13 show the comparison of tidal compone nts in terms of measured and predicted water levels at UF, Punta Gorda, El Jobean and Harbor Height, respectively.

PAGE 83

60 Y earDa y 2003&2004 WaterLevel(cmNAVD88) 200 250 300 350 400 450 500 550 -100 -50 0 50 Simulated Measured Belowplot YearDay2004 WaterLevel(cmNAVD88) 425 430 435 440 -100 -50 0 50 Simulated Measured Figure 5-9. Comparison of measured and simulated water levels at UF station Y earDate2003&2004 WaterLevel(cm,NAVD88) 200 250 300 350 400 450 500 550 -100 -75 -50 -25 0 25 50 75 Simulated Measured Belowplot Y earDate2003 WaterLevel(cm,NAVD88) 240 242 244 246 248 250 252 254 256 258 260 -100 -75 -50 -25 0 25 50 75 Simulated Measured Figure 5-10. Comparison of me asured and simulated water levels at Punta Gorda station

PAGE 84

61 Y earDate2003&2004 WaterLevel(cm,NAVD88) 200 250 300 350 400 450 500 550 -100 -75 -50 -25 0 25 50 75 Simulated Measured Belowplot Y earDate2003 WaterLevel(cm,NAVD88) 240 242 244 246 248 250 252 254 256 258 260 -100 -75 -50 -25 0 25 50 75 Simulated Measured Figure 5-11. Comparison of me asured and simulated water levels at El Jobean station Y ea r Da y 2003 WaterLevel(cm,NAVD88) 206 208 210 212 214 -100 -75 -50 -25 0 25 50 75 Measured Simulated Y ea r Da y 2003&2004 WaterLevel(cm,NAVD88) 180 200 220 240 -100 -75 -50 -25 0 25 50 75 Measured Simulated Belowplot Figure 5-12. Comparison of measured and si mulated water levels at Harbor Height station

PAGE 85

62 Table 5-6. Comparison of phase errors in te rms of measured and predicted water level at UF station Constituents K1 O1 Q1 M2 N2 S2 K2 Measured (degree) 350.00 345.72 349.31 94.54 95.42 97.87 95.34 Simulated (degree) 347.17 342.74 331.09 91.66 88.34 95.09 86.67 Measured Simulated (degree) 2.83 2.8 18.22 2.88 7.08 2.78 8.67 abs (MeasuredSimulated)/360 (Dimensionless-%) 0.79 0.83 5.06 0.80 1.97 0.77 2.41 Table 5-7. Comparison of amplitude errors tidal components in terms of measured and predicted water level at UF station Constituents K1 O1 Q1 M2 N2 S2 K2 Measured (cm) 12.18 11.72 2.51 14.65 2.23 3.45 2.42 Simulated (cm) 12.31 11.90 1.91 15.30 2.68 3.82 2.32 Measured Simulated (cm) -0.13 -0.18 0.60 -0.65 -0.45 -0.37 0.10 Relative Error abs (MeasuredSimulated)/Measured 1.07 1.54 23.90 4.44 20.18 10.72 4.13 Table 5-8. Comparison of phase errors in te rms of measured and predicted water level at Punta Gorda station Constituents K1 O1 Q1 M2 N2 S2 Measured (degree) 343.69 335.24 343.41 77.65 83.11 85.49 Simulated (degree) 348.20 342.47 330.53 92.28 93.07 97.57 Measured Simulated (degree) -4.51 -7.23 12.88 -14.63 -9.96 -12.08 abs (MeasuredSimulated)/360 (Dimensionless-%) 1.25 2.01 3.58 4.06 2.77 3.36 Table 5-9. Comparison of amplitude errors in terms of measured and predicted water level at Punta Gorda station Constituents K1 O1 Q1 M2 N2 S2 Measured (cm) 12.72 12.91 2.63 17.58 2.47 5.30 Simulated (cm) 12.54 12.39 2.03 17.13 3.03 4.52 Measured Simulated (cm) 0.18 0.52 0.60 0.45 -0.56 0.78 Relative Error abs (MeasuredSimulated)/Measured 1.42 4.03 22.81 2.56 22.67 14.72

PAGE 86

63 Table 5-10. Comparison of phase errors in terms of measured and predicted water level at El Jobean station Constituents K1 O1 Q1 M2 N2 S2 K2 Measured (degree) 343.33 339.62 342.36 83.63 81.59 89.31 80.84 Simulated (degree) 350.96 346.74 335.29 100.00 94.86 102.43 95.98 Measured Simulated (degree) -7.63 -7.12 7.07 -16.37 -13.27 -13.12 -15.14 abs (MeasuredSimulated)/360 (Dimensionless-%) 2.12 1.98 1.96 4.55 3.69 3.64 4.21 Table 5-11. Comparison of amplitude errors in terms of measured and predicted water level at El Jobean station Constituents K1 O1 Q1 M2 N2 S2 K2 Measured (cm) 12.92 12.49 2.55 16.85 2.59 4.85 2.87 Simulated (cm) 12.53 12.09 1.98 16.37 2.98 3.89 2.65 Measured Simulated (cm) 0.39 0.40 0.57 0.48 -0.39 0.96 0.22 Relative Error abs (MeasuredSimulated)/Measured 3.02 3.20 22.35 2.85 15.06 19.79 7.67 Table 5-12. Comparison of phase errors in terms of measured and predicted water level at Harbor Height station Constituents K1 O1 Q1 M2 N2 S2 Measured (degree) 12.68 351.22 359.34 97.18 98.17 136.38 Simulated (degree) 353.78 345.73 334.21 95.28 95.13 132.61 Measured Simulated (degree) 18.90 5.49 25.13 1.90 3.04 3.77 abs (MeasuredSimulated)/360 (Dimensionless-%) 5.25 1.53 6.98 0.53 0.84 1.05 Table 5-13. Comparison of amplitude errors in terms of measured and predicted water level at Harbor Height station Constituents K1 O1 Q1 M2 N2 S2 Measured (cm) 13.39 12.11 2.55 15.68 2.51 4.31 Simulated (cm) 12.80 12.64 2.00 17.67 3.59 4.01 Measured Simulated (cm) 0.59 -0.53 0.55 -1.99 -1.08 0.30 Relative Error abs (MeasuredSimulated)/Measured 4.41 4.38 21.57 12.69 43.03 6.96

PAGE 87

64 5.4.2 Currents For comparison between measured and si mulated currents, we separated the currents into North/South and East/West current components in Figures 5-13 to 5-16. A positive sign of North/South and East/West current indicates water flows from south to north and from west to east, respectively Figures 5-13 and 5-15 show lower-level (105.5cm above bottom) currents, while Figures 5-14 and 5-16 show upper-level (240.5cm above bottom) currents. Current data were missing during Sep. 18th through Oct. 2nd 2003 while the ADCP sensor was being repaired. The RMS errors for lowerlevel East/West current, upper-le vel East/West current, lower-l evel North/South current, and upper-level North/South current are 3.60 cm/sec, 4.29 cm/sec, 5.73 cm/sec, and 7.25 cm/sec, respectively. The RMS error of Ea st/West current is higher than that of North/South current, but the RM S errors normalized by the range of current show similar values (<15%) for both East/Wes t and North/South currents. YearDay2003&2004 Currentatbottom(East/Westdirection,cm/s) 200 250 300 350 400 450 500 550 -50 -40 -30 -20 -10 0 10 20 30 40 50 BelowPlot Simulated Measured YearDay2003 Currentatbottom(East/Westdirection,cm/s) 240 242 244 246 248 250 252 254 256 258 260 -50 -40 -30 -20 -10 0 10 20 30 40 50 Simulated Measured Figure 5-13. Lower current velocity directed by East/West at UF station

PAGE 88

65 YearDay2003&2004 Currentatsurface(East/Westdirection,cm/s) 200 250 300 350 400 450 500 550 -50 -40 -30 -20 -10 0 10 20 30 40 50 BelowPlot Simulated Measured YearDay2003 Currentatsurface(East/Westdirection,cm/s) 240 245 250 255 260 -50 -40 -30 -20 -10 0 10 20 30 40 50 Simulated Measured Figure 5-14. Upper current velocity directed by East/West at UF station YearDay2003&2004 Currentatbottom(North/Southdirection,cm/s) 200 250 300 350 400 450 500 550 -60 -40 -20 0 20 40 BelowPlot Simulated Measured YearDay2003 Currentatbottom(North/Southdirection,cm/s) 240 245 250 255 260 -60 -40 -20 0 20 40 Simulated Measured Figure 5-15. Lower current velocity di rected by North/South at UF station

PAGE 89

66 YearDay2003&2004 Currentatsurface(North/Southdirection,cm/s) 200 250 300 350 400 450 500 550 -60 -40 -20 0 20 40 BelowPlot Simulated Measured YearDay2003 Currentatsurface(North/Southdirection,cm/s) 240 245 250 255 260 -60 -40 -20 0 20 40 Simulated Measured Figure 5-16. Upper curren t velocity directed by No rth/South at UF station 5.4.3 Salinity and Temperature Figures 5-17 and 5-18 show the comparison between measured and simulated salinity as well as temperature at the UF station at three different layers: upper layer (263 cm above bottom), middle layer (176.5 cm above bottom), and lower layer (77.5 cm above bottom). Although some measured salinity da ta are missing during periods of sensor replacement and excessive growth of barnacles, the measured salinity data illustrate the vivid trend during the wet and dry seasons, with strong vertical strati fication in the period of high river flow and well mixed salinity in the period of low river flow. The model results compare well with the measured salinity data. The RMS errors for upper-level, middle-level, and lower-level salinity are 2.13 PSU, 1.84 PSU, and 2.14 PSU, respectively. Based on comparison between si mulated and measured water temperature, the RMS errors are 1.99, 1.10, and 1.30 C at surface, middle, and bottom layer at the UF station, respectively.

PAGE 90

67 The salinity at the El Jobean station is influenced by river runoff from the Myakka River and bathymetry. The RMS errors for surface and bottom salin ity are 4.51 PSU and 4.14 PSU at the El Jobean station (Figure 519). In the case of temperature, RMS errors of surface and bottom are 1.47 C and 1.45 C, respectively (Figure 5-20). At the Punta Gorda station, the RMS errors for surface sali nity, bottom salinity, surface temperature, and bottom temperature are 3.81 PSU, 3.95 PSU, 1.47 C, and 1.45 C, respectively (Figures 5-21 and 5-22). YearDay2003&2004 SurfaceSalinity(ppt) 200 300 400 500 10 20 30 SurfaceSalinity(263cmfrombottom) Simulated Measured YearDa y 2003&2004 MiddleSalinity(ppt) 200 300 400 500 10 20 30 MiddleSalinity(176.5cmfrombottom) Simulated Measured YearDa y 2003&2004 BottomSalinity(ppt) 200 300 400 500 10 20 30 BottomSalinity(77.5cmfrombottom) Simulated Measured Figure 5-17. Comparison between measured and simulated salin ity at UF station

PAGE 91

68 YearDay2003&2004 SurfaceTemperature(Celsius) 200 250 300 350 400 450 500 550 10 20 30 SurfaceTemperature(263cmfrombottom) Simulated Measured YearDa y 2003&2004 BottomTemperature(Celsius) 200 250 300 350 400 450 500 550 10 20 30 BottomTemperature(77.5cmfrombottom) Simulated Measured YearDa y 2003&2004 MiddleTemperature(Celsius) 200 250 300 350 400 450 500 550 10 20 30 MiddleTemperature(176.5cmfrombottom) Simulated Measured Figure 5-18. Comparison between measured and simulated temperature at UF station Figures 5-23 to 5-27 show the comparison between measured and simulated salinity at HBMP and the four stations m easured by Mote Marine Laboratory. Model predictions and measured data have good agre ements except bottom salinity at HBMP. The values of measured bottom salinity in dicate over 25 PSU from Julian Days 230 to 270. However, the values of bottom salinity at other stations which are located in further offshore area than HBMP were below 20 PSU du ring that period. This suggests measured bottom salinity data at HBMP seem to be incorrect.

PAGE 92

69 YearDa y 2003&2004 Salinity(ppt) 200 300 400 500 0 5 10 15 20 25 30 Simulated Measured ElJobeanSurface YearDa y 2003&2004 Salinity(ppt) 200 300 400 500 0 5 10 15 20 25 30 ElJobeanBottom Simulated Measured Figure 5-19. Comparison between measured a nd simulated salinity at El Jobean station YearDa y 2003&2004 Temperature(Celcius) 200 300 400 500 10 15 20 25 30 35 40 ElJobeanBottom Measured Simulated YearDa y 2003&2004 Temperature(Celcius) 200 300 400 500 10 15 20 25 30 35 40 Measured Simulated ElJobeanSurface Figure 5-20. Comparison between measured and simulated temperature at El Jobean station

PAGE 93

70 YearDa y 2003&2004 Salinity(ppt) 200 250 300 350 400 450 500 550 0 10 20 30 PuntaGordaBottom Simulated Measured YearDa y 2003&2004 Salinity(ppt) 200 250 300 350 400 450 500 550 0 10 20 30 Simulated Measured PuntaGordaSurface Figure 5-21. Comparison between measured and simulated salinity at Punta Gorda station YearDa y 2003&2004 Temperature(Celcius) 200 250 300 350 400 450 500 550 10 15 20 25 30 35 40 PuntaGordaBottom Measured Simulated YearDa y 2003&2004 Temperature(Celcius) 200 250 300 350 400 450 500 550 10 15 20 25 30 35 40 Measured Simulated PuntaGordaSurface Figure 5-22. Comparison between measured and simulated temperature at Punta Gorda station

PAGE 94

71 YearDay2003&2004 Surfacesalinity(ppt) 200 250 300 350 400 450 500 0 5 10 15 20 25 30 35 40 Simulated Measured YearDay2003&2004 Bottomsalinity(ppt) 200 250 300 350 400 450 500 0 5 10 15 20 25 30 35 40 Simulated Measured Figure 5-23. Comparison between measured a nd simulated salinity at (13, 34) measured by Mote Marine Laboratory YearDay2003&2004 Surfacesalinity(ppt) 200 250 300 350 400 450 500 0 5 10 15 20 25 30 35 Simulated Measured YearDay2003&2004 Bottomsalinity(ppt) 200 250 300 350 400 450 500 0 5 10 15 20 25 30 35 Simulated Measured Figure 5-24. Comparison between measured a nd simulated salinity at (19, 43) measured by Mote Marine Laboratory

PAGE 95

72 YearDay2003&2004 Surfacesalinity(ppt) 200 250 300 350 400 450 500 0 5 10 15 20 25 30 35 Simulated Measured YearDay2003&2004 Bottomsalinity(ppt) 200 250 300 350 400 450 500 0 5 10 15 20 25 30 35 Simulated Measured Figure 5-25. Comparison between measured a nd simulated salinity at (19, 53) measured by Mote Marine Laboratory YearDay2003&2004 Surfacesalinity(ppt) 200 250 300 350 400 450 500 0 5 10 15 20 25 30 35 Simulated Measured YearDay2003&2004 Bottomsalinity(ppt) 200 250 300 350 400 450 500 0 5 10 15 20 25 30 35 Simulated Measured Figure 5-26. Comparison between measured a nd simulated salinity at (22, 47) measured by Mote Marine Laboratory

PAGE 96

73 YearDay2003&2004 Surfacesalinity(ppt) 200 250 300 350 400 450 500 0 5 10 15 20 25 30 35 Simulated Measured YearDay2003&2004 Bottomsalinity(ppt) 200 250 300 350 400 450 500 0 5 10 15 20 25 30 35 Simulated Measured Figure 5-27. Comparison between measur ed and simulated salinity at HBMP 5.4.4 Correlations between Simulated and Measured Parameters Figures 5-28 through 5-32 provide the correlation be tween measured and model predictions. We notice high correlation of wate r elevation at the UF, El Jobean, and Punta Gorda stations. The location of spots indicates the correlation between model predictions and measured data. A perfect match between mo del and measured data is indicated by the diagonal line on each graph. The spots above the line illustrate that the model predictions over-predict. On the other hand, the spots belo w the line indicate that the model results under-predict. The correlation coefficient (r) whic h is a measure of how well trends in the predicted values follow trends in the measur ed values is shown with each graph. The correlation coefficient is a number between 0 and 1. If there is no relationship between the predicted values and the measured values, the correlation coefficien t is 0 or very low. As the strength of the relationship between the predicted values and measured values

PAGE 97

74 increases, so does the correlation coefficient. A perfect fit gives a coefficient of 1. The values of r are over 0.939 for water elevation. In the case of salinity and temperature, the values of r are as good as water elevation. Therefore, there are si gnificant co rrelation between measured values and simulated predictions. Measured Simulated -100 -50 0 50 -120 -100 -80 -60 -40 -20 0 20 40 60ElJobeanwaterelevation r=0.939 Measured Simulated -100 -50 0 50 -120 -100 -80 -60 -40 -20 0 20 40 60UFwaterelevation r=0.941 Measured Simulated -100 -50 0 50 -120 -100 -80 -60 -40 -20 0 20 40 60PuntaGordawaterelevation r=0.955 Figure 5-28. Correlation between measured and simulated water elevation at UF, El Jobean and Punta Gorda stations (Dia gonal line is the perfect match line)

PAGE 98

75 Measured Simulated 0 5 10 15 20 25 30 35 0 5 10 15 20 25 30SurfacesalinityatUFsite r=0.973 Measured Simulated 0 5 10 15 20 25 30 35 0 5 10 15 20 25 30MiddlessalinityatUFsite r=0.981 Measured Simulated 0 5 10 15 20 25 30 35 0 5 10 15 20 25 30BottomsalinityatUFsite r=0.930 Figure 5-29. Correlation between measured and simulated salini ty at UF station (Diagonal line is the perfect match line) Measured Simulated 0 5 10 15 20 25 30 35 0 5 10 15 20 25 30 35SurfacetemperatureatUFsite r=0.918 Measured Simulated 0 5 10 15 20 25 30 35 0 5 10 15 20 25 30MiddlestemperatureatUFsite r=0.972 Measured Simulated 0 5 10 15 20 25 30 35 0 5 10 15 20 25 30 35BottomtemperatureatUFsite r=0.968 Figure 5-30. Correlation between measured and simulated temperature at UF station (Diagonal line is the perfect match line)

PAGE 99

76 Measured Simulated 0 5 10 15 20 25 30 35 0 5 10 15 20 25 30BottomsalinityatPuntaGorda r=0.881 Measured Simulated 0 5 10 15 20 25 30 35 0 5 10 15 20 25 30SurfacesalinityatPuntaGorda r=0.912 Measured Simulated 0 5 10 15 20 25 30 35 0 5 10 15 20 25 30BottomsalinityatElJobeansite r=0.936 Measured Simulated 0 5 10 15 20 25 30 35 0 5 10 15 20 25 30SurfacesalinityatElJobeansite r=0.940 Figure 5-31. Correlation between measured a nd simulated salinity at El Jobean and Punta Gorda stations (Diagonal li ne is the perfect match line) Measured Simulated 0 5 10 15 20 25 30 35 0 5 10 15 20 25 30SurfacetemperatureatElJobeansite r=0.970 Measured Simulated 0 5 10 15 20 25 30 35 0 5 10 15 20 25 30BottomtemperatureatElJobeansite r=0.971 Measured Simulated 0 5 10 15 20 25 30 35 0 5 10 15 20 25 30SurfacetemperatureatPuntaGorda r=0.972 Measured Simulated 0 5 10 15 20 25 30 35 0 5 10 15 20 25 30BottomtemperatureatPuntaGorda r=0.975 Figure 5-32. Correlation between measured a nd simulated temperature at El Jobean and Punta Gorda stations (Diagonal li ne is the perfect match line)

PAGE 100

77 5.4.5 Spectral Analysis To compare measured data and predicted re sults, spectral analyses were conducted, and the results are shown in Figures 5-33 to 5-39. From spectral analyses of water elevation, it is clear that ti dal constituents of diurnal and semi-diurnal periods dominate this study area. The spectrum density values of simulated water elevat ion are pretty close to those of measured data at the three stations In the case of currents, there are not only diurnal and semi-diurnal components but also other components such as density, wind, and river runoff. Figure 5-33. Power spectrum density of th e simulated (left hand side) and measured (right hand side) water el evation at UF station

PAGE 101

78 Figure 5-34. Power spectrum density of th e simulated (left hand side) and measured (right hand side) water eleva tion at El Jobean station Figure 5-35. Power spectrum density of th e simulated (left hand side) and measured (right hand side) water eleva tion at Punta Gorda station

PAGE 102

79 Figure 5-36. Power spectrum density of th e simulated (left hand side) and measured (right hand side) surface EAST/W EST current at UF station Figure 5-37. Power spectrum density of th e simulated (left hand side) and measured (right hand side) bottom EAST/W EST current at UF station

PAGE 103

80 Figure 5-38. Power spectrum density of th e simulated (left hand side) and measured (right hand side) surf ace NORTH/SOUTH current at UF station Figure 5-39. Power spectrum density of th e simulated (left hand side) and measured (right hand side) botto m NORTH/SOUTH current at UF station

PAGE 104

81 5.5 Establishment of Freshwater Flows and Levels The first step to conduct the freshwater fl ows and levels estimation is to understand the vertical salinity distribution in the Uppe r Charlotte Harbor. We made two verticallongitudinal salinity profiles along the Myakka and Peace Rivers (Figure 5-40). Figures 5-41 and 5-42 show a series of vertical-long itudinal salinity profil es along the Myakka River during one day for high and low runoff periods, respectively. During the high runoff period, a mouth of river flow pushes fr esh water to the middle of Upper Charlotte Harbor. This causes the values of surface and bottom salinity to be less than 1 PSU until 13 km downstream from the mouth of the Myakka River, and the salt-wedge is formed around 15 to 25 km downstream the mouth of the Myakka River. On the other hand, salinity near the mouth of the Myakka Rive r is about 9 to 10 PSU during the low runoff period. We can see strong vertical mixi ng and gradual salinity variation along the Myakka River. There are huge differences in salinity between high and low runoff periods. The differences are 1 to 20 PSU at 10 km downstream from the mouth of the Myakka River. A series of vertical-longitudi nal salinity profiles along th e Peace River is shown in Figures 5-43 and 5-44 during one day for high and low runoff periods, respectively. There is a channel from the middle of Upper Charlotte Harbor to near the mouth of the Peace River in Figure 5-3. The channel allows salty water to move near the mouth of the Peace River. For example, as shown Figure 544 in low runoff period, salinity near the mouth of the Peace River fluctuates between 14 and 19 PSU due to the channel. During the high runoff period, fresh water marches downstream to 8km from the mouth of the Peace River and develops a salt-wedge. We ca n also see large differences of salinity between the high and low runoff periods along the Peace River.

PAGE 105

82 Figure 5-40. Plainview of ve rtical-longitudinal profiles The second step to conduct the freshwat er flows and levels estimation is to establish the quantifying relati onship between freshwater in flow and the temporal and spatial distribution of salinity along the My akka and Peace Rivers. Figure 5-45 shows the time histories of the river discharge rate and the locations of 1, 10, 20 PSU surface salinity along the Myakka River. When the Myakka river discharg e rate is over 20 m3/s, the surface salinity value is below 20 PSU fr om Myakka A to B point. When Myakka flow is over 40 m3/s, the surface salinity va lue is less than 10 PSU Myakka A to B point (Figure 5-40). The locations of 1 PSU su rface salinity along the Myakka River are directly affected by Myakka River discharge. During the low runoff period, the front of

PAGE 106

83 20 PSU surface salinity is formed around 5 to 10 km downstream from near the mouth of the Myakka River. The surface salinity near the mouth of the Myakka River is less than 10 PSU. Salinity is around 2 to 15 km from n ear mouth of the Myakka River fluctuated from 1 to 20 PSU during simulation period. 2 2 4 6 DistancefromM y akka A p ointtoM y akkaB p oint ( m ) Depth(cm) 0 5000 10000 15000 20000 25000 -600 -400 -200 014 12 10 8 6 4 2 Salinity (ppt)7/2/200312:00MyakkaBpoint MyakkaApoint 2 2 4 6 6 8 DistancefromMyakka A pointtoMyakkaBpoint(m) Depth(cm) 0 5000 10000 15000 20000 25000 -600 -400 -200 014 12 10 8 6 4 2 Salinity (ppt)7/2/200316:00MyakkaBpoint MyakkaApoint 2 2 4 6 8 DistancefromMyakka A pointtoMyakkaBpoint(m) Depth(cm) 0 5000 10000 15000 20000 25000 -600 -400 -200 014 12 10 8 6 4 2 Salinity (ppt)7/2/200320:00MyakkaBpoint MyakkaApoint 2 4 6 DistancefromM y akka A pointtoM y akkaBpoint ( m ) Depth(cm) 0 5000 10000 15000 20000 25000 -600 -400 -200 014 12 10 8 6 4 2 Salinity (ppt)7/3/200300:00MyakkaBpoint MyakkaApoint 2 2 4 DistancefromMyakka A pointtoMyakkaBpoint(m) Depth(cm) 0 5000 10000 15000 20000 25000 -600 -400 -200 014 12 10 8 6 4 2 Salinity (ppt)7/3/200304:00MyakkaBpoint MyakkaApoint 2 4 6 8 Distance f r omM y akka A pointtoM y akkaBpoint ( m ) Depth(cm) 0 5000 10000 15000 20000 25000 -600 -400 -200 014 12 10 8 6 4 2 Salinity (ppt)7/3/200308:00MyakkaBpoint MyakkaApoint Figure 5-41. Vertical-longit udinal salinity profiles along the axis of the Myakka River during high runoff season in 2003

PAGE 107

84 1 2 2 0 2 2 2 4 2 4 Distance f romM y akka A pointtoM y akkaBpoint ( m ) Depth(cm) 0 5000 10000 15000 20000 25000 -600 -400 -200 024 22 20 18 16 14 12 Salinity (ppt)11/21/200300:00MyakkaBpoint MyakkaApoint 1 6 1 8 2 0 2 2 2 4 Distance f romMyakka A pointtoMyakkaBpoint(m) Depth(cm) 0 5000 10000 15000 20000 25000 -600 -400 -200 024 22 20 18 16 14 12 Salinity (ppt)11/21/200304:00MyakkaBpoint MyakkaApoint 1 2 1 4 1 6 2 0 2 2 2 4 Distance f r omMyakka A pointtoMyakkaBpoint(m) Depth(cm) 0 5000 10000 15000 20000 25000 -600 -400 -200 024 22 20 18 16 14 12 Salinity (ppt)11/21/200308:00MyakkaBpoint MyakkaApoint 1 0 1 4 1 6 2 0 2 2 2 4 Distance f romM y akka A pointtoM y akkaBpoint ( m ) Depth(cm) 0 5000 10000 15000 20000 25000 -600 -400 -200 024 22 20 18 16 14 12 Salinity (ppt)11/20/200320:00MyakkaBpoint MyakkaApoint 1 6 1 8 2 0 2 2 2 4 2 4 Distance f romMyakka A pointtoMyakkaBpoint(m) Depth(cm) 0 5000 10000 15000 20000 25000 -600 -400 -200 024 22 20 18 16 14 12 Salinity (ppt)11/20/200316:00MyakkaBpoint MyakkaApoint 1 4 1 8 2 0 2 2 2 4 1 1 Distance f romM y akka A pointtoM y akkaBpoint ( m ) Depth(cm) 0 5000 10000 15000 20000 25000 -600 -400 -200 024 22 20 18 16 14 12 Salinity (ppt)11/20/200312:00MyakkaBpoint MyakkaApoint Figure 5-42. Vertical-longit udinal salinity profiles along the axis of the Myakka River during low runoff season in 2003

PAGE 108

85 0 5 3 5 Distance f romPeace A pointtoPeaceBpoint(m) Depth(cm) 2500 5000 7500 10000 -600 -400 -200 014.5 12.5 10.5 8.5 6.5 4.5 2.5 Salinity (ppt)7/2/200320:00PeaceBpoint PeaceApoint 0 5 4 5 Distance f romPeace A pointtoPeaceBpoint ( m ) Depth(cm) 2500 5000 7500 10000 -600 -400 -200 014.5 12.5 10.5 8.5 6.5 4.5 2.5 Salinity (ppt)7/2/200316:00PeaceBpoint PeaceApoint 0 5 3.5 5.5 Distance f romPeace A pointtoPeaceBpoint ( m ) Depth(cm) 2500 5000 7500 10000 -600 -400 -200 014.5 12.5 10.5 8.5 6.5 4.5 2.5 Salinity (ppt)7/3/200304:00PeaceBpoint PeaceApoint 1 5 6 5 Distance f romPeace A pointtoPeaceBpoint ( m ) Depth(cm) 2500 5000 7500 10000 -600 -400 -200 014.5 12.5 10.5 8.5 6.5 4.5 2.5 Salinity (ppt)7/2/200312:00PeaceBpoint PeaceApoint 0 5 4 5 Distance f romPeace A pointtoPeaceBpoint ( m ) Depth(cm) 2500 5000 7500 10000 -800 -600 -400 -20014.5 12.5 10.5 8.5 6.5 4.5 2.5 Salinity (ppt)7/3/200300:00PeaceBpoint PeaceApoint 0 5 5.5 3.5 Distance f romPeace A pointtoPeaceBpoint(m) Depth(cm) 2500 5000 7500 10000 -600 -400 -200 014.5 12.5 10.5 8.5 6.5 4.5 2.5 Salinity (ppt)7/3/200308:00PeaceBpoint PeaceApoint Figure 5-43. Vertical-longitudinal salinity profiles along the axis of the Peace River during high runoff season in 2003

PAGE 109

86 22 26 26 1 9 Distance f r omPeace A pointtoPeaceBpoint ( m ) Depth(cm) 2500 5000 7500 10000 -600 -400 -200 027 25 23 21 19 17 15 Salinity (ppt)11/20/200312:00PeaceBpoint PeaceApoint 26 26 2 5 2 3 1 9 Distance f r omPeace A pointtoPeaceBpoint(m) Depth(cm) 2500 5000 7500 10000 -600 -400 -200 027 25 23 21 19 17 15 Salinity (ppt)11/21/200300:00PeaceBpoint PeaceApoint 2 6 26 26 26 1 9 2 2 1 6 Distance f r omPeace A pointtoPeaceBpoint ( m ) Depth(cm) 2500 5000 7500 10000 -600 -400 -200 027 25 23 21 19 17 15 Salinity (ppt)11/21/200304:00PeaceBpoint PeaceApoint 2 5 26 2 3 1 6 1 9 1 4 Distance f r omPeace A pointtoPeaceBpoint(m) Depth(cm) 2500 5000 7500 10000 -600 -400 -200 027 25 23 21 19 17 15 Salinity (ppt)11/21/200308:00PeaceBpoint PeaceApoint 2 2 2 2 2 4 1 8 2 6 1 5 Distance f r omPeace A pointtoPeaceBpoint(m) Depth(cm) 2500 5000 7500 10000 -600 -400 -200 027 25 23 21 19 17 15 Salinity (ppt)11/20/200316:00PeaceBpoint PeaceApoint 2 6 26 1 9 1 7 2 2 1 4 Distance f r omPeace A pointtoPeaceBpoint(m) Depth(cm) 2500 5000 7500 10000 -600 -400 -200 027 25 23 21 19 17 15 Salinity (ppt)11/20/200320:00PeaceBpoint PeaceApoint Figure 5-44. Vertical-longitudinal salinity profiles along the axis of the Peace River during low runoff season in 2003

PAGE 110

87 YearDate2003&2004 Riverdischarge(m3/sec) 200 250 300 350 400 450 500 550 100 200 300 400 MyakkaRiver YearDate2003&2004 DistancefromMyakkaAtoBpoints(m) 200 250 300 350 400 450 500 550 0 5000 10000 15000 20000 25000 20ppt 10ppt 1ppt Figure 5-45. Time histories of Myakka river discharge and the locati ons of 1, 10, 20 PSU surface salinity along the Myakka River The time histories of the river discharge rate and the locations of 1, 10, 20 PSU surface salinity along the Peace River are shown in Figure 5-46. Most of salinity values are less than 10 PSU from Peace A to B point during the high runoff period. In the period of low runoff, salinity values are around 10 PSU near the mouth of the Peace River. This implies that some salty water can reach far up near the Peace River. From Peace A to B point, salinity values changed from 1 to 20 PSU be tween the wet and dry seasons.

PAGE 111

88 YearDate2003&2004 Riverdischarge(m3/sec) 200 250 300 350 400 450 500 550 100 200 300 400 500 600 700 800 PeaceRiver YearDate2003&2004 DistancefromPeaceAtoBpoints(m) 200 250 300 350 400 450 500 550 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000 11000 20ppt 10ppt 1ppt Figure 5-46. Time histories of Peace river discharge and the locations of 1, 10, 20 PSU surface salinity along the Peace River Figures 5-47 and 5-48 show the locations at which the average surface salinity is 5, 10 and 20 PSU for 30 days along the Myakka a nd Peace Rivers, respectively. These show the locations of salinity without tidal fluctu ation effect. During the high runoff period, the locations of 30-day average 5 and 10 PSU salinity were formed around 13 and 20 km from near the mouth of the Myakka River in Figure 5-49. Near the mouth of the Myakka River, the average salinity for 30 days is a bout 5 to 10 PSU in the period of low runoff.

PAGE 112

89 Along the Peace River, the locations of aver age 5 and 10 PSU salinity for 30 days are roughly 7 to 10 km from near the mouth duri ng the high runoff period. In the period of low runoff, the front of average 20 PSU salin ity for 30 days exits around 4km from near the mouth of the Peace River. YearDate2003&2004 Riverdischarge(m3/sec) 200 250 300 350 400 450 500 550 0 100 200 300 400 MyakkaRiver YearDate2003&2004 DistancefromMyakkaAtoBpoints(m) 200 250 300 350 400 450 500 550 0 5000 10000 15000 20000 25000 Thelocationswhereaveraged surfacesalinityis20pptfor30days Thelocationswhereaveraged surfacesalinityis10pptfor30days Thelocationswhereaveraged surfacesalinityis5pptfor30days Figure 5-47. Locations where average surface salinity is 5, 10, 20 PSU for 30 days along the Myakka River

PAGE 113

90 YearDate2003&2004 Riverdischarge(m3/sec) 200 250 300 350 400 450 500 550 0 100 200 300 400 500 600 700 800 PeaceRiver YearDate2003&2004 DistancefromPeaceAtoBpoints(m) 200 250 300 350 400 450 500 550 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000 11000 Thelocationswhereaveraged surfacesalinityis20pptfor30days Thelocationswhereaveraged surfacesalinityis10pptfor30days Thelocationswhereaveraged surfacesalinityis5pptfor30days Figure 5-48. Locations where average surface salinity is 5, 10, 20 PSU for 30 days along the Peace River As an example for estimating the effects of freshwater flow on sa linity distribution, we plot the locations of 10 PSU surface salin ity along the Myakka and Peace Rivers with respect to each river discharge in Figures 5-49 and 5-50, respectiv ely. The logarithmic regression lines are added to show the relationship betw een the locations of 10 PSU surface salinity and each river di scharge. When discharge rates from the Myakka River is more than 100 m3/s, the surface salinity values ar e under 10 PSU down to 10km from the mouth of the Myakka River. When the freshwater flow is 13 m3/s, surface salinity below 10 PSU is found between the mouth of the My akka River and the El Jobean station.

PAGE 114

91 Along the Peace River, when the river discharge rate is 245 m3/s, 10 PSU surface salinity to if found 11 km downstream of the mouth. From near the mouth to the Punta Gorda station, surface salinity goes below 10 PSU when the river flow is over 20 m3/s. We can see the discrepancies of freshwater flow even though the distances from each river mouth are the same. These discrepancies are from the size and shape of each river path. MyakkaRIverdischarge(m3/s) DistanceformMyakkaAtoCpoint(m) 0 20 40 60 80 100 120 0 2000 4000 6000 8000 10000 Locationof10pptsalinity Logarithmicregressionline MyakkaCpoint ElJobeanstation Freshwaterflow=13m3/s Freshwaterflow=100m3/s Figure 5-49. Locations of 10 PSU surface sa linity along the Myakka River with respect to Myakka river discharge

PAGE 115

92 PeaceRiverdischarge(m3/s) DistancefromPeaceAtoBpoint(m) 0 50 100 150 200 250 300 350 0 2000 4000 6000 8000 10000 Locationof10pptsalinity Logarithmicregressionline PeaceBpoint PuntaGordastation Freshwaterflow=245m3/s Freshwaterflow=20m3/s Figure 5-50. Locations of 10 PSU surface sa linity along the Peace River with respect to Peace River discharge Some studies employed the 30-day average method to establish minimum or maximum flows and levels (Doering et al ., 2002; Park, 2004). This method is based on calculating the locations of 30day average surface salinity concentration using constant river discharge. This study follows the met hod to estimate the minimum freshwater flow criterion. In order to obtain the locations of 30-day averag e surface salinity concentration along the Myakka and Peace Rivers, 10 scenarios w ith constant river discharge of 0, 1, 2, 5, 10, 15, 20, 30, 40, 50 m3/s at the Myakka River are simulated for 60 days in the period

PAGE 116

93 of low runoff. In addition, 10 scenarios with c onstant river discharge of 5, 10, 15, 20, 30, 50, 75, 100, 150, 200 m3/s at the Peace River are simula ted. When each river uses the constant value, the other rivers are given the measured value. Figures 5-51 and 5-52 show the locations of 5, 10, 15, and 20 PSU salinity versus Myakka and Peace River discharge, respective ly. For 30-day average surface salinity to be below 5 and 10 PSU at the El Jobean st ation located around 6k m downstream from the mouth of Myakka River should be greater than 40 and 20 m3/s, respectively. When river runoff at the Myakka River is 20 m3/s, the locations of 30-day average surface salinity 5, 10, 15, and 20 PSU are 3, 6.1, 8.2, and 9.6 km downstream from the Myakka River mouth, respectively. The locations of 30-day average surface salinity noticeably shifted downstream as river flow at the Mya kka River is altered from 2 to 20 m3/s. For example, a difference in 10 m3/s flow rate from the Myakka River results in 4 to 5 PSU differences in the salinity at the El Jobean station. As shown in Figure 5-52, 30-day average surface salinity at Peace River mouth is less than 5 PSU until river discharge at the Peace River is over 30 m3/s. If the river runoff at the Peace River is 10, 20, 50, and 100 m3/s, 30-day average surface salinity at the Punta Gorda station is 20, 15, 10, and 5 PSU, respectively. As river runoff at the Peace Rive r is increased from 30 to 100 m3/s, the locations of 30day average surface salinity move d to much further downstream. According to recent study, seagrass loss in the Upper Charlotte Harbor is considered a serious problem The Halodule wrightii which is the dominnant seagrass species in this area, was ba rely observed along the Peace Ri ver (Dixon and Kirkpatrick, 1999). This phenomenon may be related to the large freshwater inflow from the Peace River because the growth rate of Halodule wrightii quickly decreases at salinity < 12

PAGE 117

94 PSU. Mortality could occur when salinity becomes below 6 PSU (Doering et al. 2002; Zimmerman and Livingston, 1976). From Figure 5-53, di scharge around 74.9 and 136.9 m3/s maintained salinity about 6 PSU at 2 a nd 5 km from Peace A point, respectively. For increasing the amount of Halodule wrightii from 5km downst ream of Peace A point to Peace B point, 30-day constant freshwater flows need to be less than 70.2 m3/s. MyakkaRiverDischarge(m3/s) DistancefromMyakkaAtoCpoint(m) 0 20 40 60 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000 11000 MyakkaCpoint ElJobeansiteThevalueofaveraged salinityis20pptfor30day Thevalueofaveraged salinityis5pptfor30day Thevalueofaveraged salinityis10pptfor30day Thevalueofaveraged salinityis15pptfor30day Figure 5-51. Relationship betw een locations at specific salinity values vs. Myakka River discharge

PAGE 118

95 PeaceRiverDischarge(m3/s) DistancefromPeaceAtoBpoint(m) 50 100 150 200 250 0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000 11000 PeaceBpoint PuntaGordasiteThevalueofaveraged salinityis20pptfor30day Thevalueofaveraged salinityis5pptfor30day Thevalueofaveraged salinityis10pptfor30day Thevalueofaveraged salinityis15pptfor30day Figure 5-52. Relationship between location s at specific salinity values vs. Peace River discharge

PAGE 119

96 PeaceRiverDischarge(m3/s) DistancefromPeaceAtoBpoint(m) 50 100 150 200 250 0 2000 4000 6000 8000 10000 2kmfromPeaceApointThevalueofaveraged salinityis6pptfor30day Thevalueofaveraged salinityis12pptfor30dayPeaceBpoint 5kmfromPeaceApoint74.9m3/s 37.1m3/s 70.2m3/s 136.9m3/s Figure 5-53. Relationship between location s at salinity 6 PSU and 12 PSU vs. Peace River discharge

PAGE 120

97 CHAPTER 6 WATER QUALITY MODEL IN THE UPPER CHARLOTTE HARBOR This chapter describ es the application of the water quality model in the Upper Charlotte Harbor. Due to the availability of nutrient data, the period of simulation was altered to the year 2000. However, in th e hydrodynamic and transport model, forcing mechanism and boundary conditions were near ly the same as those of the 2003 2004 simulation. Water quality simula tion was conducted from January 10th to December 13th 2000 using CH3D-IMS described in Chapter 4. The boundary-fitted grid and bathymetry are the same as those of the previous chap ter. Even though the entire Charlotte Harbor estuarine system was simulated, the Upper Ch arlotte Harbor area is the primary focus of this study. To represent hydrology, geomorphol ogy, and water quality characteristics, the entire estuary was divided into 15 segments in Figure 6-1. According to the phytoplankton sampling analys is, this study classified three algal groups, such as diatom, crytophytes, and cy anophytes (McPherson et al., 1996). In addition, a new SOD method and silica cycle were used to better represent water quality dynamics of the Upper Charlotte Harbor ar ea. This simulation provides three major objectives: (1) Calibration and validation of Upper Charlotte Harbor water quality model using multi algal groups, (2) Observation of the impact of river pollutant loading and its reduction on the water quality in this sy stem applying the new SOD model, and (3) Understanding the mechanism of hypoxia event at bottom water in the upper Charlotte Harbor.

PAGE 121

98 1 2 3 4 56 78 9 10 11 12 13 14 15 Figure 6-1. Segments for Charlotte Harbor estuarine system 6.1 Forcing Mechanism and Boundary Conditions of Circulation and Transport There is one major difference betw een the 2003 2004 simulation and the 2000 simulation. Data at the UF site (located Latitude 26 52' 30'' N and Longitude 82 9' E) are not available during the 2000 simulation. However, a NOAA station at Fort Myers (26 36' N, 81 52' E) provided hourly wind and air temperature information instead of the UF site. The other forcing mechanism and boundary conditions were the same as the 2003 2004 simulation.

PAGE 122

99 Water level data at the offshore boundary condition came from tidal constituents of the Advanced Circulation (ADCIRC) model and non-tidal components of the Naples and Clearwater station (NOAA). Ai r temperature was obtained at Venice (NDBC), Fort Myers, and Naples stations (NOAA) in Figure 6-2. Discharge and runoff boundary conditions were implied using daily measured discharge data at the Myakka River (near Sarasota and at Big Slough Canal), Peace River (at Arcade, at Joshua Creek, and at Horse Creek), Shell Creek, and Caloosahatchee River. Daily precipitation data were recorded at SWFWMD rain gauges at si x locations through the study area. The only evaporation station data in the study domain was the SF WMD daily pan evaporation data measured near Naples, FL. The surface wind data were imposed using hourly wind magnitude and direction data at Venice st ation (NDBC), Fort Myers, and Naples stations (NOAA) in Figure 6-3. More detailed informat ion was explained in Chapter 5. YearDay2000 AirTemperature(Celcius) 100 200 300 0 5 10 15 20 25 30 35 40 Venice FortMyers Naples Figure 6-2. Air temperatures at Venice, Fort Myers, and Naples stations

PAGE 123

100 Y earDa y 2000 100 200 300 5m/s Venice Y earDa y 2000 100 200 300 5m/s FortMyers Y earDa y 2000 100 200 300 5m/s Naples Figure 6-3. Wind speed at Venice, Fort Myers, and Naples stations 6.2 Measured Data SWFWMD collected water samples monthl y for one year at 12 stations. Those stations are mostly located in the Upper Charlotte Harbor. One of them is outside the estuary (Table 6-1). Eight stations along the Caloosahatchee River were sampled by SFWMD in Figure 6-4. Those da ta were also collected on a monthly basis. The water samples provided light, water temperature, salinity, TSS, TOC, TKN, Nitrate, Ammonium (NH4), TP, Phosphates (PO4), Dissolved Silica, DO, and Corrected Chlorophyll a In the Upper Charlotte Harbor Syst em, the amount of Total Nitrogen is generally greater than that of Total Phosphorous. However, the results of the data

PAGE 124

101 collected at stations near the river (CH 002, CH004 and CH005) demonstrated higher phosphates than ammonium. This suggests that the system may be limited by N availability. Many other studies showed that phytoplankton growth in the estuary is controlled by the availability of nitrogen (Froelich et al. 1985; Mc Person and Miller 1994; and Montgomery et al. 1991). Remarkable features of these data were low DO concentration in the Upper Charlotte Harbor area during the summer. The results of the data collected at CH004, CH005, and CH006 st ations indicated low DO values, and its values were sometimes below 2mg/l. From the comparison between DO concentration and phytoplankton biomass, low DO concentrations do not seem to strongly relate to the bloom of phytoplankton. Low DO values in the Upper Charlotte Harbor are commonly discovered though other studies (Heyl 1998; Leverone and Culter 2003; Dixon et al. 2003; and Montgomery and Stone 2003). Table 6-1. Description of wa ter quality measured stations Station Agency Latitude Longitude Easting (UTM, meter) Northing (UTM, meter) CH-001 SWFWMD 27 00' 07" 82 15' 11" 375674 2987267 CH-002 SWFWMD 26 57' 22" 82 12' 30" 380063 2982147 CH-029 SWFWMD 27 00' 58" 81 59' 03" 402367 2988631 CH-004 SWFWMD 26 56' 39" 82 03' 32" 394887 2980722 CH-005 SWFWMD 26 55' 56" 82 06' 15" 390352 2979406 CH-005B SWFWMD 26 57' 11" 82 06' 33" 389904 2981718 CH-006 SWFWMD 26 54' 01" 82 07' 09" 388859 2975881 CH-007 SWFWMD 26 52' 39" 82 04' 07" 393859 2973345 CH-008 SWFWMD 26 53' 13" 82 09' 28" 384983 2974439 CH-009 SWFWMD 26 49' 13" 82 05' 29" 391542 2966995 CH-011 SWFWMD 26 44' 12" 82 10' 00" 383975 2957800 CH-013 SWFWMD 26 41' 30" 82 18' 01" 370636 2952944 CES001 SFWMD 26 43' 20" 81 41' 23" 431400 2955824 CES002 SFWMD 26 43' 35" 81 42' 28" 429607 2956326 CES003 SFWMD 26 43' 00" 81 45' 38" 424351 2955279 CES004 SFWMD 26 40' 54" 81 50' 01" 417059 2951448 CES005 SFWMD 26 38' 12" 81 53' 19" 411552 2946470 CES006 SFWMD 26 34' 56" 81 54' 36" 409380 2940485 CES007 SFWMD 26 31' 48" 81 57' 56" 403804 2934742 CES008 SFWMD 26 31' 24" 82 00' 31" 399508 2934006

PAGE 125

102 CH-029 CES001 CES002 CES003 CES004 CES005 CES006 CES007 CES008 CH-001 CH-002 CH-029 CH-005B CH-006 CH-007 CH-009 CH-008 CH-011 CH-013CH-004CH-005 Easting(UTM,meter) Northing(UTM,meter) 320000 340000 360000 380000 400000 420000 440000 2880000 2900000 2920000 2940000 2960000 2980000 Figure 6-4. Locations of 2000 water quality measured stations operated by SFWMD and SWFWMD 6.3 Boundary Conditions for Water Quality Model 6.3.1 Pollutant Loads from Freshwater Discharge River nutrient loads and to tal suspended solids were provided by SWFWMD and SWWMD water quality monito ring stations. Data collete d at CH001, CH029, CH004, and CH005B stations are representative of th e river loads of the Myakka River, Peace River, Shell Creek, and a small creek between the Myakka and Peace Rivers, respectively. There were thr ee river boundary conditions at Caloosahatchee River, which are S-79 spillway, Whiskey Cr eek, and Cape Coral. For these river boundary conditions, data at CES001, CES005, and CES006 were employed. The load s of each phytoplankton

PAGE 126

103 were calculated multiplying chlorophyll a value by mean biovolume ratio (diatom in 55 percents, cryptophytes in 35 percent; cyanophytes in 10 percen t) of the whole Charlotte Harbor (McPherson et al., 1996). 6.3.2 Temperature Temperature was calculated by solving th e temperature equation in Appendix A. Temperature varied from place to place and from time to time. The temperature variations depend on the net rate of heat fl ow into or out of a water body (Pickard and Emery, 1990). The temperature equation in th e model includes the net heat flux across the surface water as a su rface boundary condition. HEL S HpQQQ Q z T KC )1(* (6-1) where is water density, CP is the specific heat, KH is an empirically determined turbulence exchange coefficient, QS is the incident short-wave heat flux, is the surface albedo, QL is the long-wave heat flux, QE is the latent heat flux, and QE is the sensible heat flux. Other sources of heat flow, such as heat from the earths interior, change of kinetic energy of waves into heat at the surf, heat from chemical or nuclear reactions, are relatively small and would be neglected. Details of the heat flux can be found in Appendix G. 6.3.3 Light and Color The light intensity data were converted from the global and diffusion horizontal solar irradiance data which were processed at the National Renewa ble Energy Laboratory (NREL). These data were used as the su rface boundary condition to calculate PAR (Photosynthetically Active Radiance) in th e light attenuation model (Park 2004). Color data were obtained by SWFWMD and SF WMD water quality monitoring stations.

PAGE 127

104 6.3.4 Sediment Type and Nu trient Distribution A total of 215 sediment grab samples and 28 shallow cores were collected from the Charlotte Harbor estuarine system during the period December 27th, 1964 to January 1st, 1965 by Huang (1966). Based on the sediment si ze distribution, the entire study area was characterized into five sediment types. In Upper Charlotte Harbor, the sediments are primarily sands (mostly greater than 97 %). The finer fractions, mostly clays, are observed at the mouth of the Peace River (Turner et al., 2001). 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 included organic carbon, total nitrogen, and total phosphor ous. Silica data in the sedi ment column were obtained from Turner et al. (2001) who suggested th e relationship between %Carbon and %N, %P, and %S from core samples in the Upper Charlotte Harbor. 6.4 Model Parameters and Calibration The determination of water quality mode l parameters described in Chapter 4 generally requires effort because they are dependent on physical and biochemal factors, such as geological characterist ics of estuaries, temperature, tidal variation, freshwater inflows, point/non point nutrient loads, species of algal group, nutrient concentrations at the sediment column, etc. Therefore, it is essential to apply adequate procedure for adjusting water quality parameters. This procedure should begin with deciding which parameters are sensitive and important to the water quality species. Substantial parameters are employed to reproduce major pa tterns of all water qu ality species, and less considerable parameters are applied to calibrate detailed characteristics of certain species. In order to determine the significance of each water quality parameter, sensitivity tests were conducted by varying each parame ter within l iterature values. This study

PAGE 128

105 conducted 48 sensitivity tests using 24 para meters. For each parameter, two sensitivity tests are performed: to double parameter values and to halve parameter values (Table 62). This study desires a baseline simulation for comparing sensitivity test simulations. The baseline simulation employed the model parameters obtained by field observation, estimation from nutrient kinetic equations, averaged literature values, and previous model applications. Pribble et al. (1998) and Park (2004) suggested wate r quality parameters in the study area. Some parameter values of baselin e simulation had a range and some were a constant number (Table 6-2). This study used a multi-phytoplankton feature. Each species has its own parameter values, thus the parameters which are related to the phytoplankton have a range. The partition co efficient values were estimated by TSS values and particulate nutrient concentrati on. The values are also given by variant numbers for different segments. Each sensit ivity test employed the same initial and boundary conditions as well as phys ical forcing as those for the baseline simulation. The 90-day simulation was conducted and comp ared with the baseline simulation. The relative error was employe d as a skill assessment to indicate the discrepancy between baseline and sensitivity test simulations. N i i N i i i relBaseline testySensitivit Baseline E1 1 (6-2) where N is the total number of model predic tions (i=1,2,3,,N). The results of the sensitivity tests showed maximum algae growth rate as the most sensitive parameter in the water quality model (Table 6-3).

PAGE 129

106 Table 6-2. Description of sensitivity tests Test run Parameter Literature range Baseline values Test values Unit Test_agrm1 Test_agrm2 Maximum algae growth rate 0.2 8.0 1.2 1.9 2.4 3.8 0.6 0.95 1/day Test_halfn1 Test_halfn2 Nitrogen half saturation rate 0.005 4.34 0.025 0.05 0.0125 mg/l Test_halfp1 Test_halfp2 Phosphorous halfsaturation rate 0.001 0.163 0.005 0.01 0.0025 mg/l Test_kaex1 Test_kaex2 Algae respiration rate 0.03 0.9 0.065 0.09 0.13 0.18 0.0325 0.045 1/day Test_kas1 Test_kas2 Algae mortality rate 0.003 0.1 0.03 0.06 0.06 0.1 0.015 0.025 1/day Test_was1 Test_was2 Algae settling velocity 0 1700 5 10 10 20 2.5 5 cm/sec Test_sonm1 Test_sonm2 Ammonification rate of SON 0.001 0.4 0.015 0.02 0.03 0.04 0.0075 0.01 1/day Test_nitr1 Test_nitr2 Nitrification rate 0.004 0.11 0.05 0.08 0.1 0.16 0.025 0.04 1/day Test_dron1 Test_dron2 Sorption/desorption rate of SON 0.005 0.08 0.01 0.02 0.005 1/day Test_dran1 Test_dran2 Sorption/desorption rate of NH4 0.01-0.02 0.01 0.02 0.005 1/day Test_sopm1 Test_sopm2 Mineralization rate of SOP 0.001 0.6 0.02 0.03 0.04 0.06 0.01 0.015 1/day Test_drop1 Test_drop2 Sorption/desorption rate of SOP 0.01 0.08 0.01 0.02 0.005 1/day Test_drip1 Test_drip2 Sorption/desorption rate of SRP 0.01 0.02 0.01 0.02 0.005 1/day Test_dros1 Test_dros2 Sorption/desorption rate of DOS 0.005 0.1 0.02 0.04 0.01 1/day Test_akd1 Test_akd2 Oxidation rate 0.02 0.6 0.05 0.1 0.025 1/day Test_zgrm1 Test_zgrm2 zooplankton growth rate 0.1 0.3 0.16 0.32 0.08 1/day Test_halfa1 Test_halfa2 Half-saturation rate for zooplankt on 0.01 2.0 1.0 2.0 0.5 mg/l Test_kzex1 Test_kzex2 zooplankton respiration rate 0.001 0.16 0.015 0.03 0.0075 1/day Test_kzs1 Test_kzs2 zooplankton mortality rate 0.001 0.125 0.02 0.04 0.01 1/day Test_pcon1 Test_pcon2 Partition coefficient (SON and PON) 5.E-6 5.E-3 1.E-4 3.E-4 2.E-4 6.E-4 5.E-5 1.5E-4 1/ g Test_pcan1 Test_pcan2 Partition coefficient (NH4 and PIN) 5.E-7 5.E-3 1.E-4 2.E-4 5E-5 1/ g Test_pcop1 Test_pcop2 Partition coefficient (SOP and POP) 8.E-6 5.E-3 5.E-5 1.E-4 1.E-4 2.E-4 2.5E-5 5.E-5 1/ g Test_pcip1 Test_pcip2 Partition coefficient (SRP and PIP) 1.E-6 1.E-3 5.E-5 1.E-4 1.E-4 2.E-4 2.5E-5 5.E-5 1/ g Test_pcos1 Test_pcos2 Partition coefficient (DOS and POS) 1.E-5 2.E-5 2.E-5 4.E-5 5E-6 1.E-5 1/ g

PAGE 130

107 The maximum algae growth rate is re lated in the phytoplankton biomass. As maximum growth rate increase or decrease, nitrat e, nitrite, SRP, and dissolved silica were promptly changed. The effect of reducing the maximum growth rate (Test_agrm2) was more pronounced in nutrients, phytoplankton bi omass, dissolved oxygen, and carbon than that of an increased rate (Test_agrm1). Not only maximum algae growth rate but also parameters related to phytoplankton biomass pr oved vital to the model. Those parameters include algae respiration/mort ality rate, settling velocity, nitrogen/phosphorous half saturation, zooplankton maximum growth ra te, and zooplankton respiration/mortality rate. The second most important parameter rev ealed by the sensitivity test was the ammonification rate of SON, follo wed by the partition coefficient and sorption/desorption rate between SON and P ON. SON is one of the main water quality species because it directly affects ammonium nitrogen (NH4), nitrite, and nitrate. As the ammonification rate increases, S ON is rapidly mineralized to NH4. This ammonium nitrogen transfers to NO3 due to nitrification. Sinc e inorganic-nitr ogen-containing compounds (NH4 and NO3) are a major food for phytoplankton growth, more available inorganic nitrogen in the water column leads to larger phytoplankton production. Conversely, decreasing the a mmonification rate promotes a reduction for inorganic nitrogen as well as phytoplankton. Organic nitrogen contained in the soil part icle is brought from the soil surface to the water column due to erosion; that is, the excess of bottom shear st ress over the critical shear stress of the bed. During the eros ion events, the suspended sediment which embraces a large amount of organic nitrogen can partially change into SON due to

PAGE 131

108 sorption/desorption. This process will not cease until the concentration of organic nitrogen on sediment particle s and SON in the water reaches equilibrium. In the water quality model, the partition coefficient and sorption/desorption rate mainly control the interaction between SON and PON. Therefore, those two parameters play an important role in changing SON values in the water column, as does ammonification rate. Results of sensitivity tests for other parameters indicate that the mineralization rate of SOP and partition coeffi cient of SRP/PIP became the most significant among the parameters regarding phosphorous. DO is dir ectly altered by phytoplankton biomass due to its photosynthesis. However, DO is more sensitive to oxidation rate than other parameters. Nitrogen half saturation constant (HALFN), algae respiration/mortality rate and maximum zooplankton growth rate prove d a major impact on the water quality species. Each parameter closely related to th e specific water qual ity species played a major role in that species. Table 6-3. Sensitivity analysis results Test run DO Chl. a TOC NO3 DTKN SRP PN PP SOS AVG Baseline 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Test_agrm1 0.36 2.69 0.07 44.58 1.90 2.02 0.10 0.02 3.39 6.13 Test_agrm2 0.71 14.35 0.39 182.5 9.06 9.14 0.35 0.08 19.41 26.22 Test_halfn1 0.25 3.30 0.14 68.73 3.09 3.24 0.10 0.02 6.40 9.47 Test_halfn2 0.20 1.73 0.06 32.35 1.39 1.49 0.05 0.01 1.87 4.35 Test_halfp1 0.05 0.38 0.01 1.86 0.10 0.11 0.01 0.00 0.20 0.30 Test_halfp2 0.00 0.00 0.00 0.01 0.00 0.00 0.00 0.00 0.00 0.00 Test_kaex1 0.52 4.41 0.34 78.42 3.37 3.66 0.46 0.51 4.87 10.73 Test_kaex2 0.26 1.62 0.34 29.84 1.28 1.35 0.41 0.50 5.36 4.55 Test_kas1 1.07 27.50 1.26 12.58 3.99 2.08 2.52 0.77 4.96 6.30 Test_kas2 1.00 18.30 0.98 5.05 2.98 1.33 2.13 0.66 2.71 3.90 Test_was1 0.09 1.05 0.33 0.66 0.18 0.20 0.41 0.50 3.24 0.74 Test_was2 0.08 0.55 0.32 0.55 0.17 0.19 0.41 0.50 3.23 0.67 Test_sonm1 1.20 25.09 1.51 32.31 12.01 26.49 2.46 0.53 50.65 16.92 Test-sonm2 0.59 18.79 0.92 15.28 7.67 16.81 1.58 0.52 39.27 11.27 Test_nitr1 0.09 0.15 0.32 3.83 0.17 0.19 0.41 0.50 3.10 0.97 Test_nitr2 0.09 0.14 0.32 2.25 0.17 0.21 0.41 0.50 3.31 0.82 Test_dron1 0.37 10.51 0.57 8.46 28.09 8.91 4.66 0.51 22.19 9.36 Test_dron2 0.43 10.47 0.76 8.77 31.07 9.17 5.91 0.51 16.32 9.27 Test_dran1 0.10 0.97 0.31 1.17 0.23 0.93 0.47 0.50 5.08 1.08 Test_dran2 0.09 0.55 0.33 0.80 0.19 0.49 0.41 0.50 2.25 0.62

PAGE 132

109 Table 6-3. Continued Test run DO Chl. a TOC NO3 DTKN SRP PN PP SOS AVG Test_sopm1 0.00 0.00 0.00 0.01 0.00 25.79 0.00 0.00 0.00 2.87 Test_sopm2 0.01 0.05 0.00 0.29 0.01 21.16 0.00 0.00 0.03 2.39 Test_drop1 0.00 0.00 0.00 0.01 0.00 3.49 0.00 0.58 0.00 0.45 Test_drop2 0.00 0.00 0.00 0.01 0.00 2.24 0.00 0.33 0.00 0.29 Test_drip1 0.00 0.00 0.00 0.01 0.00 10.17 0.00 0.85 0.00 1.23 Test_drip2 0.00 0.00 0.00 0.01 0.00 6.98 0.00 0.61 0.00 0.85 Test_dros1 0.00 0.04 0.00 0.06 0.00 0.00 0.00 0.00 69.09 7.69 Test_dros2 0.00 0.04 0.00 0.10 0.01 0.01 0.00 0.00 38.44 4.29 Test_akd1 4.16 0.00 9.97 0.03 0.00 0.00 0.00 0.00 0.00 1.57 Test_akd2 2.58 0.00 5.94 0.02 0.00 0.00 0.00 0.00 0.00 0.95 Test_zgrm1 0.90 32.84 0.46 13.19 0.53 0.82 0.83 0.17 1.82 6.84 Test_zgrm2 0.71 24.03 0.34 6.06 0.61 0.32 0.67 0.14 0.76 3.74 Test_halfa1 0.53 17.83 0.25 4.92 0.44 0.24 0.50 0.10 0.56 2.82 Test_halfa2 0.55 18.94 0.27 10.20 0.34 0.35 0.50 0.11 0.73 3.55 Test_kzex1 0.23 15.10 0.21 4.20 0.57 0.28 0.44 0.09 4.61 2.86 Test_kzex2 0.09 8.55 0.11 2.84 0.35 0.19 0.25 0.05 2.90 1.70 Test_kzs1 0.41 9.38 0.53 3.75 1.39 0.47 1.13 0.23 0.26 1.95 Test_kzs2 0.30 5.29 0.57 3.32 1.33 0.40 1.19 0.25 0.19 1.42 Test_pcon1 0.39 11.92 0.59 9.52 31.60 10.00 5.11 0.06 21.38 10.06 Test_pcon2 0.47 11.56 0.60 9.66 34.41 10.17 6.55 0.06 21.64 10.57 Test_pcan1 0.07 1.72 0.09 1.59 0.27 1.55 0.27 0.01 3.34 0.99 Test_pcan2 0.03 0.88 0.04 0.82 0.14 0.80 0.14 0.00 1.72 0.51 Test_pcop1 0.00 0.00 0.00 0.01 0.00 6.22 0.00 1.16 0.00 0.82 Test_pcop2 0.00 0.00 0.00 0.01 0.00 3.78 0.00 0.76 0.00 0.51 Test_pcip1 0.09 0.13 0.32 0.51 0.17 18.35 0.41 1.40 3.24 2.74 Test_pcip2 0.09 0.13 0.32 0.50 0.17 12.48 0.41 1.05 3.23 2.04 Test_pcos1 0.09 0.13 0.32 0.50 0.17 0.20 0.41 0.50 0.71 0.34 Test_pcos2 0.09 0.13 0.32 0.50 0.17 0.20 0.41 0.50 4.89 0.80 Based on the sensitivity te sts, this study decides th e eleven most important parameters including maximum algae growth ra te, algae respiration/mortality, maximum zooplankton growth rate, amm onification rate of SON, mi neralization rate of SOP, oxidation rate, sorption/desorption rate for so luble nutrients (SON, SOP, and DOS), and partition coefficients between soluble and particulate nutrients (SON/PON and SRP/PIP). Those parameters were mainly employed to ad just all water quality species. The other parameters are attuned as partially sensitiv e parameters for each specific water quality species. More than 200 simulations were ma de for calibrating model parameters. During the trial-and-error calibration procedure, the relative error and correlation coefficient (R2) were employed to decide optimal numerical measurements that compare observations of

PAGE 133

110 the state of the system with corresponding simulated predictions. N i i N i iiO PO error relative1 1 (6-3) )* (*)* ( )***(2 1 2 2 1 2 2 1 2PNP ONO PONPO RN i N i i i N i i (6-4) where N is the total number of observati ons or predictions (i=1,2,3,,N), Oi is the observation, Pi is the simulated prediction, Ois the average of observation, and P is the average of simulation prediction. Table 6-4 provides model parameters in the Charlotte Harbor as well as feasible ranges of those parameters obtained from field observation, laboratory experimentation, literature, and previous modeling studies. Table 6-4. Water quality model coefficients used for the Charlotte Harbor simulation Symbol Coefficient Unit Literature range Charlotte Harbor (dia)T-20 Temperature coefficient for diatom growth (1.01 1.2)** 1.08 (flag)T-20 Temperature coefficient for flagellates growth (1.01 1.2)** 1.08 (cyano)T-20 Temperature coefficient for cynobacteria growth (1.01 1.2)** 1.08 (AD)T-20 Temperature coefficient for NH4 desorption 1.08 1.08 (AI)T-20 Temperature coefficient for Ammonium instability 1.08 1.08 (BOD)T-20 Temperature coefficient for CBOD oxidation 1.02 1.15 1.08 (DN)T-20 Temperature coefficient for denitrification 1.02 1.09 1.045 (NN)T-20 Temperature coefficient for nitrification 1.02 1.08 1.04 (OD)T-20 Temperature coefficient for SON desorption 1.08 1.08 (ONM)T-20 Temperature coefficient for mineralization 1.02 1.09 1.08 (RESP)T-20 Temperature coefficient for algae respiration 1.045 1.05 (z)T-20 Temperature coefficient for zooplankton growth 1.01 1.2 1.04

PAGE 134

111 Table 6-4. Continued Symbol Coefficient Unit Literature range Charlotte Harbor dia Diatom maximum growth rate 1/day 0.55 5.0 1.9 flag Flagellates maximum growth rate 1/day 1.2 1.6 1.6 cyano Cyanobacteria maximum growth rate 1/day 0.2 4.9 1.2 z Zooplankton maximum growth rate 1/day 0.1 0.3 0.16 achla Algal carbon and chlorophyll a ratio mg C / mg Chla 20 1000 100 anc Algal nitrogen and carbon ratio mg N / mg C 0.05 0.43 0.16 apc Algal phosphorous and carbon a ratio mg P / mg C 0.005 0.03 0.025 aoc Algal oxygen and carbon ratio mg O / mg C 2.67 2.667 adsc Diatom silica and carbon ratio mg Si / mg C 0.06 0.77 0.5 dan Desorption rate of adsorbed ammonium nitrogen 1/day 0.01 0.02 0.01 don Desorption rate of adsorbed organic nitrogen 1/day 0.005 0.08 0.01 dos Desorption rate of adsorbed organic silica 1/day 0.005 0.1 0.02 dip Desorption rate of adsorbed inorganic phosphorous 1/day 0.01 0.02 0.01 dop Desorption rate of adsorbed organic phosphorous 1/day 0.01 0.08 0.01 dmol Molecular diffusion coefficient for dissolved species cm2/s 4.E-6 1.E.-5 1.E-5 fdCBOD Fraction of dissolved CBOD 0.5 0.5 Hbod Half-saturation constant for CBOD oxidation mg O2 0.02 5.6 0.5 Hn_dia Half-saturation constant for diatom uptake nitrogen mg/l 0.015 0.12 0.025 Hn_flag Half-saturation constant for flagellates uptake nitrogen mg/l 0.001 0.13 0.025 Hn_cyano Half-saturation constant for cyanobacteria uptake nitrogen mg/l 0.01 4.34 0.025 Hp_dia Half-saturation constant for diatom uptake phosphorous mg/l 0.001 0.163 0.005 Hp_flag Half-saturation constant for flagellates uptake phosphorous mg/l 0.012 (0.001 0.03)** 0.005 Hp_cyano Half-saturation constant for cyanobacteria uptake phosphorous mg/l 0.0025 0.02 0.005

PAGE 135

112 Table 6-4. Continued Symbol Coefficient Unit Literature range Charlotte Harbor Hs Half-saturation constant for diatom uptake silica mg/l 0.08 0.1 0.08 Ha Half-saturation constant for zooplankton mg/l 0.01 2.0 1.0 hv Henrys constant mg/l atm 43.8 45 45 Idia Optimum light intensity for diatom growth E/m2/s 88 350 350 Iflag Optimum light intensity for flagellates growth E/m2/s 100 288 350 Icyano Optimum light intensity for cyanobacteria growth E/m2/s 43 600 350 (NH3)air Ammonia concentration in the air g/l 0.1 0.1 Kax_dia Respiration rate by diatom 1/day 0.03 0.6 0.09 Kax_flag Respiration rate by flagellates 1/day 0.05 0.06 0.07 Kax_cyano Respiration rate by cyanobacteria 1/day 0.03 0.9 0.065 Kas_dia Mortality rate by diatom 1/day 0.03 (0.003 0.1)** 0.05 Kas_flag Mortality rate by flagellates 1/day (0.003 0.1)** 0.04 Kas_cyano Mortality rate by cyanobacteria 1/day (0.003 0.1)** 0.03 KD CBOD oxidation rate 1/day 0.02 0.6 0.05 KDN Denitrification rate constant 1/day 0.0 1.0 0.09 KNN Nitrification rate constant 1/day 0.004 0.11 0.05 0.08 Kon Ammonification rate of SON 1/ day 0.001 0.4 0.015 0.02 Kop Mineralization rate of SOP 1/day 0.001 0.6 0.02 0.03 Kvol Constant rate for nitrogen volatilization 1/day 3.5 9.0 7.0 Kzx Respiration rate of zoopla nkton 1/day 0.001 0.16 0.015 Kzs Mortality rate of zooplankton 1/day 0.0065 0.0326 0.02 Pan Partition coefficient between NH4 and PIN 1/ g 1.E-6 5.E-3 1.E-4 Pon Partition coefficient between SON and PON l/ g 5.E-6 5.E-3 1.E-4 3.E-4 Pip Partition coefficient between SRP and PIP 1/ g 1.E-6 1.E-3 5.E-5 1.E-4 Pop Partition coefficient between SOP and POP 1/ g 8.E-6 5.E-3 5.E-5 1.E-4 Pos Partition coefficient between DOS and POS 1/ g 1.E-5 2.E-5 WSdia Diatom settling velocity cm/day 1 1700 10 WSflag Flagellates settling velocity cm/day 5 39 10 WScyano Cyanobacteria settling velocity cm/day 0 20 5 ** All phytoplankton

PAGE 136

113 6.5 Water Quality Mode l Simulation The water quality model was simulated from January 10th to December 13th 2000 with the hydrodynamic and sediment model. Be fore conducting this simulation, a spin-up simulation was executed to create an appr opriate initial condition through the study domain. A 30-day spin-up simulation (December 11th, 1999 to January 10th, 2000) was iteratively performed until surface elevation, currents, salinity, temperature, suspended sediment, and nutrients reached a dynamic steady state. The initial nutrient values of spin-up simulation were obtained from the SWFWMD and SFWMD water quality monitoring stations. The other initial values were acquired from hydrodynamic monitoring stations maintained by SFWM D, SWFWMD, NOAA, NDBC, and USGS. Figures 6-5 through 6-12 show the simulate d results and measured data at CH002, CH005, CH006, CH007, CH008, CH009, CH011, and CH013 stations, respectively. Each figure embodies eight different time seri es of water quality parameters (Chlorophyll a, DO, TKN, PO4, CBOD, NH4, and DOS). The rectangular symbol represents measured water quality parameters near the surface layer. In the DO time series, delta and gradient shape symbolize measured surface and bottom oxygen. The solid and dashed lines represent the simulated water quality paramete rs near the surface and bottom except three species time series. In the summer season, low DO values at the bottom layer were observed at CH005 and CH006 stations, which are locate d along the Peace River. These phenomena were strongly related to vert ical stratification due to high river flow from the Peace River and low wind mixing (Figures 6-13 and 614). Figure 6-14 provides the correlation between measured/simulated DO at the botto m layer and Peace River discharge at CH005 and CH006 stations.

PAGE 137

114 Chlotophylla(ug/l) 50 100 150 200 250 300 0 5 10 15 20 DisolvedSilica(ug/l) 50 100 150 200 250 300 0 1000 2000 3000 4000 5000 TKN(ug/l) 50 100 150 200 250 300 0 500 1000 1500 2000 NH4(ug/l) 50 100 150 200 250 300 0 50 100 150 200 PO4(ug/l) 50 100 150 200 250 300 0 50 100 150 200 250 300 Chlotophylla(ug/l) 50 100 150 200 250 300 0 5 10 15 20 Diatom Flagellates Cyanophytes CBOD(mg/l) 50 100 150 200 250 300 0 5 10 15 20 25 30 DO(mg/l) 50 100 150 200 250 300 0 5 10 15 SimulatedSurfaceDO SimulatedBottomDO MeasuredSurfaceDO MeasuredBottomDO Figure 6-5. Temporal water qualit y variations at CH002 station in 2000 Chlotophylla(ug/l) 50 100 150 200 250 300 0 5 10 15 20 TKN(ug/l) 50 100 150 200 250 300 0 500 1000 1500 2000 NH4(ug/l) 50 100 150 200 250 300 0 50 100 150 200 PO4(ug/l) 50 100 150 200 250 300 0 50 100 150 200 250 300 Chlotophylla(ug/l) 50 100 150 200 250 300 0 5 10 15 20 Diatom Flagellates Cyanophytes CBOD(mg/l) 50 100 150 200 250 300 0 5 10 15 20 25 30 DisolvedSilica(ug/l) 50 100 150 200 250 300 0 500 1000 1500 DO(mg/l) 50 100 150 200 250 300 0 5 10 15 SimulatedSurfaceDO SimulatedBottomDO MeasuredSurfaceDO MeasuredBottomDO Figure 6-6. Temporal water qualit y variations at CH005 station in 2000

PAGE 138

115 Chlotophylla(ug/l) 50 100 150 200 250 300 0 5 10 15 20 TKN(ug/l) 50 100 150 200 250 300 0 500 1000 1500 2000 NH4(ug/l) 50 100 150 200 250 300 0 50 100 150 200 PO4(ug/l) 50 100 150 200 250 300 0 50 100 150 200 250 300 Chlotophylla(ug/l) 50 100 150 200 250 300 0 5 10 15 20 Diatom Flagellates Cyanophytes CBOD(mg/l) 50 100 150 200 250 300 0 5 10 15 20 25 30 DisolvedSilica(ug/l) 50 100 150 200 250 300 0 500 1000 1500 DO(mg/l) 50 100 150 200 250 300 0 5 10 15 SimulatedSurfaceDO SimulatedBottomDO MeasuredSurfaceDO MeasuredBottomDO Figure 6-7. Temporal water qualit y variations at CH006 station in 2000 Chlotophylla(ug/l) 50 100 150 200 250 300 0 5 10 15 20 TKN(ug/l) 50 100 150 200 250 300 0 500 1000 1500 2000 NH4(ug/l) 50 100 150 200 250 300 0 50 100 150 200 PO4(ug/l) 50 100 150 200 250 300 0 50 100 150 200 250 300 Chlotophylla(ug/l) 50 100 150 200 250 300 0 5 10 15 20 Diatom Flagellates Cyanophytes CBOD(mg/l) 50 100 150 200 250 300 0 5 10 15 20 25 30 DisolvedSilica(ug/l) 50 100 150 200 250 300 0 500 1000 1500 DO(mg/l) 50 100 150 200 250 300 0 5 10 15 SimulatedSurfaceDO SimulatedBottomDO MeasuredSurfaceDO MeasuredBottomDO Figure 6-8. Temporal water qualit y variations at CH007 station in 2000

PAGE 139

116 Chlotophylla(ug/l) 50 100 150 200 250 300 0 5 10 15 20 TKN(ug/l) 50 100 150 200 250 300 0 500 1000 1500 2000 NH4(ug/l) 50 100 150 200 250 300 0 50 100 150 200 PO4(ug/l) 50 100 150 200 250 300 0 50 100 150 200 250 300 Chlotophylla(ug/l) 50 100 150 200 250 300 0 5 10 15 20 Diatom Flagellates Cyanophytes CBOD(mg/l) 50 100 150 200 250 300 0 5 10 15 20 25 30 DisolvedSilica(ug/l) 50 100 150 200 250 300 0 500 1000 1500 DO(mg/l) 50 100 150 200 250 300 0 5 10 15 SimulatedSurfaceDO SimulatedBottomDO MeasuredSurfaceDO MeasuredBottomDO Figure 6-9. Temporal water qualit y variations at CH008 station in 2000 TKN(ug/l) 50 100 150 200 250 300 0 500 1000 1500 2000 NH4(ug/l) 50 100 150 200 250 300 0 50 100 150 200 Chlotophylla(ug/l) 50 100 150 200 250 300 0 5 10 15 20 Diatom Flagellates Cyanophytes CBOD(mg/l) 50 100 150 200 250 300 0 5 10 15 20 25 30 DisolvedSilica(ug/l) 50 100 150 200 250 300 0 500 1000 1500 PO4(ug/l) 50 100 150 200 250 300 0 20 40 60 80 100 DO(mg/l) 50 100 150 200 250 300 0 5 10 15 SimulatedSurfaceDO SimulatedBottomDO MeasuredSurfaceDO MeasuredBottomDO Chlotophylla(ug/l) 50 100 150 200 250 300 0 5 10 15 20 Figure 6-10. Temporal water quality variations at CH 009 station in 2000

PAGE 140

117 Chlotophylla(ug/l) 50 100 150 200 250 300 0 5 10 15 20 TKN(ug/l) 50 100 150 200 250 300 0 500 1000 1500 2000 NH4(ug/l) 50 100 150 200 250 300 0 50 100 150 200 Chlotophylla(ug/l) 50 100 150 200 250 300 0 5 10 15 20 Diatom Flagellates Cyanophytes CBOD(mg/l) 50 100 150 200 250 300 0 5 10 15 20 25 30 DisolvedSilica(ug/l) 50 100 150 200 250 300 0 500 1000 1500 PO4(ug/l) 50 100 150 200 250 300 0 20 40 60 80 100 DO(mg/l) 50 100 150 200 250 300 0 5 10 15 SimulatedSurfaceDO SimulatedBottomDO MeasuredSurfaceDO MeasuredBottomDO Figure 6-11. Temporal water quality variations at CH 011 station in 2000 NH4(ug/l) 50 100 150 200 250 300 0 50 100 150 200 DisolvedSilica(ug/l) 50 100 150 200 250 300 0 500 1000 1500 CBOD(mg/l) 50 100 150 200 250 300 0 2 4 6 8 10 PO4(ug/l) 50 100 150 200 250 300 0 10 20 30 40 Chlotophylla(ug/l) 50 100 150 200 250 300 0 2 4 6 8 10 Chlotophylla(ug/l) 50 100 150 200 250 300 0 2 4 6 8 10 Diatom Flagellates Cyanophytes TKN(ug/l) 50 100 150 200 250 300 0 200 400 600 800 1000 DO(mg/l) 50 100 150 200 250 300 0 5 10 15 SimulatedSurfaceDO SimulatedBottomDO MeasuredSurfaceDO MeasuredBottomDO Figure 6-12. Temporal water quality variations at CH 013 station in 2000

PAGE 141

118 YearDay2000 RiverDischarge(m3/s) 50 100 150 200 250 300 0 20 40 60 80 100 PeaceRiver MyakkaRiver DO(mg/l) 50 100 150 200 250 300 0 5 10 15 SimulatedSurfaceDO SimulatedBottomDO MeasuredSurfaceDO MeasuredBottomDOCH006 DO(mg/l) 50 100 150 200 250 300 0 5 10 15 SimulatedSurfaceDO SimulatedBottomDO MeasuredSurfaceDO MeasuredBottomDOCH005 Figure 6-13. River discharge and DO at CH005 and CH006 in 2000 A perfect correlation between bottom DO a nd river discharge is indicated by the diagonal line on each graph. The values of correlation coefficient (r) are 0.637, 0.483, 0.736, and 0.48 for measured bottom DO ve rsus Peace River discharge at CH005, simulated bottom DO versus Peace River discharge at CH005, measured bottom DO versus Peace River discharge at CH006, and simulated bottom DO versus Peace River discharge at CH006, respectiv ely. Vertical stratificati on caused weak water mixing, which prevented the surface water body from supplying DO to the bottom water body.

PAGE 142

119 SOD continuously consumed DO at the bottom layer without sufficient sources from the surface water. Riverdischarge(m3/s) Simulateddissolvedoxygen(mg/l) 0 5 10 15 20 25 0 1 2 3 4 5 6 7 8 9CH006 r=0.48 Riverdischarge(m3/s) Simulateddissolvedoxygen(mg/l) 0 5 10 15 20 25 0 1 2 3 4 5 6 7 8 9CH005 r=0.483 Riverdischarge(m3/s) Measureddissolvedoxygen(mg/l) 0 5 10 15 20 25 0 1 2 3 4 5 6 7 8 9CH005 r=0.637 Riverdischarge(m3/s) Measureddissolvedoxygen(mg/l) 0 5 10 15 20 25 0 1 2 3 4 5 6 7 8 9CH006 r=0.736 Figure 6-14. Correlation between measured/s imulated DO at the bottom layer and Peace River discharge at CH005 and CH006 stations These successions created low DO events in the Upper Charlotte Harbor. At stations that are far from Myakka and Peace Rivers, high DO values were measured at the bottom layer during the summer. Model simulated results illustrated vertical stratification and low DO events during the simulation period. Figure 6-15 de picts the correlation between measured/simulated DO at the bottom layer and simulated SOD at CH005 and

PAGE 143

120 CH006 stations. Due to the absence of measured SOD, simulated SOD was employed to obtain the correlation. The values of co rrelation coefficient are 0.518 and 0.673 for measured DO at the bottom layer versus simulated SOD at CH005 and CH006 stations. There are high correlations between simulated SOD versus DO at the bottom layer. The correlation coefficient values are over 0. 800. Figure 6-16 shows strong correlations between measured and simulated DO at th e bottom layer at CH005 and CH006 stations. The values of correlation coefficient (r) are 0.846 and 0.904 for CH005 and CH006 stations, respectively. SOD(gO2/m2/d) Simulateddissolvedoxygen(mg/l) 0.5 1 1.5 2 0 1 2 3 4 5 6 7 8 9CH005 r=0.800 SOD(gO2/m2/d) Measureddissolvedoxygen(mg/l) 0.5 1 1.5 2 0 1 2 3 4 5 6 7 8 9CH005 r=0.518 SOD(gO2/m2/d) Measureddissolvedoxygen(mg/l) 0.6 0.8 1 1.2 1.4 1.6 1.8 2 0 1 2 3 4 5 6 7 8 9CH006 r=0.673 SOD(gO2/m2/s) Simulateddissolvedoxygen(mg/l) 0.6 0.8 1 1.2 1.4 1.6 1.8 2 0 1 2 3 4 5 6 7 8 9CH006 r=0.825 Figure 6-15. Correlation between measured/simulated DO at the bottom layer and simulated SOD at CH00 5 and CH006 stations

PAGE 144

121 MeasuredDO(mg/l) SimulatedDO(mg/l) 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9CH006 r=0.904 MeasuredDO(mg/l) SimulatedDO(mg/l) 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9CH005 r=0.846 Figure 6-16. Correlation between measured and simulated DO at the bottom layer at CH005 and CH006 stations Surveys of the Charlotte Harbor, conducted on June 17th, July 8th, and July 29th, 2003 in an effort to monitor phytoplankton co mmunity composition, showed that diatomdominated communities in the Upper Charlo tte Harbor were outnumbered by higher abundances of flagellates duri ng the peak of the freshwater event (Kirkpatrick et al., 2003). Even though the model simulated period of this study was different than the monitored period by Kirkpatrick et al. (2003) Figures 6-5 to 612 depicted similar phenomena as Kirkpatrick reported during the freshwater event. The cryptophytes, socalled flagellates, noticeably increased their biomass over the Upper Charlotte Harbor during the summer season because the species has a high temperature tolerance. These temporal and spatial changes of species might play an important role in altering the Upper Charlotte Harbor ecosystem. Two skill assessments were employed to quantify the differences between simulated results and measured data: the relative error and th e relative operating characteristic (ROC). The relative error (Erel) is as follows:

PAGE 145

122 N i i N i i i reldata Measured prediction Model data Measured E1 1 (6-5) where N is the total number of model predic tions or observations (i=1,2,3,,N). The relative error between measured data and simulated results showed the dissolved oxygen element had the least average relative error (l ess than 13 %) in Table 6-5. The average relative error for Chlorophyll a approached 59 %. Results of nitrogen and phosphorous species ranged from 20 to 51 %, 72 to 89 %, and 28 to 121 % for TKN, NH4, and PO4, respectively. The relative errors for CBOD were rather low values except for CH007 station. All water quality parameters had less than 80 % averaged relative error. Table 6-5. Relative error of each station CH 002 CH 005 CH 006 CH 007 CH 008 CH 009 CH 011 CH 013 AVG Chlorophyll a 0.49 0.30 0.50 0.42 0.58 0.72 0.90 0.85 0.59 Surface DO 0.18 0.13 0.13 0.15 0.20 0.07 0.11 0.07 0.13 Bottom DO 0.14 0.16 0.12 0.16 0.19 0.09 0.11 0.1 0.13 CBOD 0.43 0.31 0.32 0.78 0.27 0.31 0.46 0.35 0.4 TKN 0.47 0.46 0.42 0.32 0.51 0.33 0.36 0.20 0.38 NH4 0.88 0.78 0.72 0.79 0.88 0.74 0.75 0.75 0.78 PO4 0.28 0.61 0.85 0.82 0.55 1.21 0.34 0.44 0.63 DOS 0.51 0.61 0.84 0.64 0.65 0.70 0.54 0.66 0.64 The ROC has been used for many studies like weather forecasting, medical imaging, material testing, and several other sc ientific disciplines (Swets, 1988; Pontius Jr. and Schneider, 2001). The ROC is a representation of the skill in which true positive fraction (TPF) and false positiv e fraction (FPF) are compared True positive fraction is simply sensitivity, and false positive fraction is the same as 1specificity (Metz, 1978). cases positive actuallyofNumber decisions TPPositive TrueofNumber TPFySensitivit )(

PAGE 146

123 cases negative actuallyofNumber decisions TN Negative TrueofNumber TNFySpecificit )( TPF and TNF terms suggest two other definitions: false positive fraction (FPF) and false negative fraction (FNF). cases negative actuallyofNumber decisions FPPositive FalseofNumber FPF )( cases positive actuallyofNumber decisions FN Negative FalseofNumber FNF )( In order to see how this skill assessment can be applied to the water quality state variable, consider the following examples. We decided the threshold as 7mg/l from 1200 measured DO data, which range from 2mg/l to 9mg/l. 200 actually positive data, with DO values above 7mg/l, are found in the 1200 data. The mo del results showed 140 true positive (TP) decisions when model predictions are above 7 mg/l while data are above 7mg/l. The false negative (FN) decisions become 60 when model predictions are below 7mg/l while observed data are above 7mg/l. Since the actually positive data are 200, the actually negative data are 1000 with DO values below 7 mg/l. If the model simulation predicted 900 true negative (TN) decisions when the model predictions are below 7 mg/l while measured data are below 7mg/l, the false positive (FP) decisions are 100 when the model predictions are above 7mg/l while measur ed data are below 7mg/l. Using these information, TPF, TNF, FPF, and FNF can be calculated (TPF would be 140/200 = 0.70; TNF would be 900/1000=0.90; FPF would be 100/1000=0.10; and FNF=60/200=0.30). If all cases are diagnosed as either positive or negative decisions for actual cases, the number of correct decisions plus incorrect decisions must equal the number of cases. Thus, it is easy to show that the various fr actions defined above must be related by TPF +

PAGE 147

124 FNF = 1 and TNF + FPF = 1. Sensitivity is plotted against the corresponding 1specificity to generate the ROC curve. Th e area under the curve is the most commonly used and has become known as the ROC score. If the total area under the curve is greater than 0.5, the model system is determined as skillful. It ranges fr om 1.0 (for a perfect model system) to 0.0 (for a perfectly bad model system), with 0.5 indicating no skill (Mason and Graham, 1999). The ROC curves an d scores of each nut rient are shown in Figure 6-17. Most of the ROC sc ores are over 0.65 and the maximum score is 0.986. This indicates that the water quality model is sk illful and accurate in producing the nutrient dynamic in the Upper Charlotte Harbor area. 1-Specificity Sensitivity 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1DO_surfaceA=0.727 1-Specificity Sensitivity 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1DO_bottomA=0.843 1-Specificity Sensitivity 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1DissolvedSilicaA=0.726 1-Specificity Sensitivity 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1SRPA=0.986 1-Specificity Sensitivity 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1CHLAA=0.779 1-Specificity Sensitivity 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1CBODA=0.774 1-Specificity Sensitivity 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1NH4A=0.506 1-Specificity Sensitivity 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1TKNA=0.675 Figure 6-17. ROC curve for Uppe r Charlotte Harbor nutrients

PAGE 148

125 6.6 Sediment Oxygen Demand and Pollutant Load Reduction Sediment oxygen demand is believed to be one of the factors causing low DO values in the Upper Charlotte Harbor. The EPA observe d SOD values at two locations in Upper Charlotte Harbor in Table 6-6 (CDW, 1998). Th e range of SOD values are 1.03 to 1.49 g O2/m2/d. In addition, CDM measured SOD values in the Pine Island Sound. Those values broadly ranged from 0.14 to 3.26 g O2/m2/d. Unfortunately, there are no SOD data during the simulated period because SOD is not a common environmental measurement. However, the predicted SOD values range 0.2 to 1.58 g O2/m2/d which is similar to the measured values in Figure 6-18. The model re sults of SOD indicate seasonal and spatial variations: (1) SOD values in the summer are higher than in other seasons and (2) SOD values near the rivers are higher than in other areas. Those variations were caused by nutrient loads, temperature, dissolved oxygen, sediment type, CBOD, and particulate settling velocity. Table 6-6. Sediment Oxygen Demand in the Upper Charlotte Harbor (CDW, 1998) Station Latitude Longitude Measured Value (March 1984) Measured Value (Sep. 1984) SOD #1 26 55' 00" 82 06' 18" 1.49 1.03 SOD #2 26 48' 18" 82 06' 30" N/A 1.39 Since the SOD values are somewhat dependent on the nutrient loads, the study of nutrient load reduction is usef ul for detecting the relati on between nutrient loads and SOD values. Furthermore, the study shows the po tential impact of nut rient loads to DO as well as other water species. To test the effect of river loads, model simulations were conducted using 50 % and 100 % nutrient load reduction at the Peace and the Myakka Rivers. Figures 6-19 through 626 show simulated water quality species of baseline simulation (no reduction simulation), 50 % nut rient load reduction, and 100 % nutrient

PAGE 149

126 load reduction measured at CH002, CH 005, CH006, CH007, CH008, CH009, CH011, and CH013 stations, respectively. Based on those simulation results, during the first six months, the water quality species barely responded to the river nutrient loads. There are two reasons for this: (1) it takes some time for water quality species to dynamically react to changing loads and (2) the amount of nutrient loads are relatively low and thus insubstantial enough to affect the kinetics of water quality species quickly. However, in the period of the second six months, the phytoplankton biomass decreased as nutrient loads d eclined because the reduction of river nutrient loads decrease d phytoplankton growth rates. The bottom DO slightly increases due to dimini shing SOD, which is directly related to particulate organic matter. In addition, the discrepancies of DTKN, SRP, and DOS were noticeable at the end of the simulation period. This suggests th at long term load reduction will affect the water quality species in the Upper Charlotte Harbor area. From the viewpoint of spatial distribution, the stations near the rivers (CH002, CH005, and CH006) were more influenced than other stations for river load reduction. Due to reduction of river nut rient loads, the chlorophyll a concentration decreased because of food lessening for phytoplankton. The amount of the phytoplankton decrease was 4 g/L at the CH002 and 2 to 3 g/L at the CH005 and CH006 for 50 % nutrient load reduction. In the case of 100% nutrient load reduction, phy toplankton biomass significantly reduced in the pe riod of the second six months Conversely, the stations (CH011 and CH013) located in Pine Island and Gu lf of Mexico showed very little effect for river nutrient load reduction. The differences of chlorophyll a concentration before and after load reducti on were less than 0.2 g/L at both stations.

PAGE 150

127 JulianDay SOD(gO2/m2/d) 50 100 150 200 250 300 0 0.5 1 1.5 2SOD#1 JulianDay SOD(gO2/m2/d) 50 100 150 200 250 300 0 0.5 1 1.5 2SOD#2 Figure 6-18. Temporal SOD variations at SOD #1 and SOD #2 stations in 2000 Figures 6-27 and 6-28 show the SOD va lues and bottom DO concentration at stations CH005 and CH009. Those two stations were selected for investigating the effect of load reduction to the SOD and bottom DO. Temporal SOD and bottom DO plots at CH005 (which is near the rive rs) indicated that SOD values decreased as nutrient loads were reduced, whereas bottom DO concentration increased. Since SOD is directly related to particulate organic matter, load reduction at rivers causes lessening of SOD. Due to the reduction of SOD, the bottom DO concentration affected by SOD fluxes improves. As revealed by Figure 6-27, the SOD values of the baseline simulation were approximately 0.1 to 0.2 and 0.2 to 0.4 g O2/m2/d higher than those of the 50 % and 100 % nutrient load reduction simulations, respectively.

PAGE 151

128 JulianDay CBOD(mg/l) 100 200 300 0 5 10 15 20 Baseline 50%Reduction 100%Reduction JulianDay TKN(ug/l) 100 200 300 0 500 1000 1500 2000 Baseline 50%Reduction 100%Reduction JulianDay DO(mg/l) 100 200 300 0 5 10 15 Baseline 50%Reduction 100%ReductionSurface JulianDay Chrolophylla(ug/l) 100 200 300 0 5 10 15 20 Baseline 50%Reduction 100%Reduction JulianDay DO(mg/l) 100 200 300 0 5 10 15 Baseline 50%Reduction 100%ReductionBottom JulianDay NH4(ug/l) 100 200 300 0 20 40 60 80 Baseline 50%Reduction 100%Reduction JulianDay PO4(ug/l) 100 200 300 0 100 200 300 Baseline 50%Reduction 100%Reduction JulianDay DOS(ug / l) 100 200 300 0 1000 2000 3000 4000 Baseline 50%Reduction 100%Reduction Figure 6-19. Temporal water quality variati ons at CH002 station in 2000 (Baseline, 50 %, and 100 % nutrients load reduction from the Peace and Myakka Rivers)

PAGE 152

129 JulianDay CBOD(mg/l) 100 200 300 0 5 10 15 20 Baseline 50%Reduction 100%Reduction JulianDay DO(mg/l) 100 200 300 0 5 10 15 Baseline 50%Reduction 100%ReductionSurface JulianDay DO(mg/l) 100 200 300 0 5 10 15 Baseline 50%Reduction 100%ReductionBottom JulianDay NH4(ug/l) 100 200 300 0 20 40 60 80 Baseline 50%Reduction 100%Reduction JulianDay Chrolophylla(ug/l) 100 200 300 0 5 10 15 20 Baseline 50%Reduction 100%Reduction JulianDay PO4(ug/l) 100 200 300 0 100 200 300 Baseline 50%Reduction 100%Reduction JulianDay DOS(ug / l) 100 200 300 0 1000 2000 3000 Baseline 50%Reduction 100%Reduction JulianDay TKN(ug/l) 100 200 300 0 200 400 600 800 1000 1200 Baseline 50%Reduction 100%Reduction Figure 6-20. Temporal water quality variati ons at CH005 station in 2000 (Baseline, 50 %, and 100 % nutrients load reduction from the Peace and Myakka Rivers)

PAGE 153

130 JulianDay CBOD(mg/l) 100 200 300 0 5 10 15 20 Baseline 50%Reduction 100%Reduction JulianDay DO(mg/l) 100 200 300 0 5 10 15 Baseline 50%Reduction 100%ReductionBottom JulianDay NH4(ug/l) 100 200 300 0 20 40 60 80 Baseline 50%Reduction 100%Reduction JulianDay PO4(ug/l) 100 200 300 0 100 200 300 Baseline 50%Reduction 100%Reduction JulianDay Chrolophylla(ug/l) 100 200 300 0 5 10 15 20 Baseline 50%Reduction 100%Reduction JulianDay DO(mg/l) 100 200 300 0 5 10 15 Baseline 50%Reduction 100%ReductionSurface JulianDay TKN(ug/l) 100 200 300 0 200 400 600 800 1000 Baseline 50%Reduction 100%Reduction JulianDay DOS(ug / l) 100 200 300 0 500 1000 1500 2000 Baseline 50%Reduction 100%Reduction Figure 6-21. Temporal water quality variati ons at CH006 station in 2000 (Baseline, 50 %, and 100 % nutrients load reduction from the Peace and Myakka Rivers)

PAGE 154

131 JulianDay CBOD(mg/l) 100 200 300 0 5 10 15 20 Baseline 50%Reduction 100%Reduction JulianDay TKN(ug/l) 100 200 300 0 500 1000 1500 2000 Baseline 50%Reduction 100%Reduction JulianDay DO(mg/l) 100 200 300 0 5 10 15 Baseline 50%Reduction 100%ReductionSurface JulianDay DO(mg/l) 100 200 300 0 5 10 15 Baseline 50%Reduction 100%ReductionBottom JulianDay PO4(ug/l) 100 200 300 0 100 200 300 Baseline 50%Reduction 100%Reduction JulianDay Chrolophylla(ug/l) 100 200 300 0 5 10 15 20 Baseline 50%Reduction 100%Reduction JulianDay DOS(ug/l) 100 200 300 0 500 1000 1500 2000 Baseline 50%Reduction 100%Reduction JulianDay NH4(ug/l) 100 200 300 0 20 40 60 80 Baseline 50%Reduction 100%Reduction Figure 6-22. Temporal water quality variati ons at CH007 station in 2000 (Baseline, 50 %, and 100 % nutrients load reduction from the Peace and Myakka Rivers)

PAGE 155

132 JulianDay CBOD(mg/l) 100 200 300 0 5 10 15 20 Baseline 50%Reduction 100%Reduction JulianDay DO(mg/l) 100 200 300 0 5 10 15 Baseline 50%Reduction 100%ReductionSurface JulianDay DO(mg/l) 100 200 300 0 5 10 15 Baseline 50%Reduction 100%ReductionBottom JulianDay NH4(ug/l) 100 200 300 0 20 40 60 80 Baseline 50%Reduction 100%Reduction JulianDay PO4(ug/l) 100 200 300 0 100 200 300 Baseline 50%Reduction 100%Reduction JulianDay Chrolophylla(ug/l) 100 200 300 0 5 10 15 20 Baseline 50%Reduction 100%Reduction JulianDay TKN(ug/l) 100 200 300 0 200 400 600 800 1000 Baseline 50%Reduction 100%Reduction JulianDay DOS(ug/l) 100 200 300 0 1000 2000 3000 Baseline 50%Reduction 100%Reduction Figure 6-23. Temporal water quality variati ons at CH008 station in 2000 (Baseline, 50 %, and 100 % nutrients load reduction from the Peace and Myakka Rivers)

PAGE 156

133 JulianDay CBOD(mg/l) 100 200 300 0 5 10 15 20 Baseline 50%Reduction 100%Reduction JulianDay DO(mg/l) 100 200 300 0 5 10 15 Baseline 50%Reduction 100%ReductionSurface JulianDay DO(mg/l) 100 200 300 0 5 10 15 Baseline 50%Reduction 100%ReductionBottom JulianDay NH4(ug/l) 100 200 300 0 20 40 60 80 Baseline 50%Reduction 100%Reduction JulianDay PO4(ug/l) 100 200 300 0 100 200 300 Baseline 50%Reduction 100%Reduction JulianDay Chrolophylla(ug/l) 100 200 300 0 5 10 15 20 Baseline 50%Reduction 100%Reduction JulianDay TKN(ug/l) 100 200 300 0 200 400 600 800 1000 Baseline 50%Reduction 100%Reduction JulianDay DOS(ug/l) 100 200 300 0 500 1000 1500 2000 Baseline 50%Reduction 100%Reduction Figure 6-24. Temporal water quality variati ons at CH009 station in 2000 (Baseline, 50 %, and 100 % nutrients load reduction from the Peace and Myakka Rivers)

PAGE 157

134 JulianDay CBOD(mg/l) 100 200 300 0 5 10 15 20 Baseline 50%Reduction 100%Reduction JulianDay DO(mg/l) 100 200 300 0 5 10 15 Baseline 50%Reduction 100%ReductionSurface JulianDay DO(mg/l) 100 200 300 0 5 10 15 Baseline 50%Reduction 100%ReductionBottom JulianDay NH4(ug/l) 100 200 300 0 20 40 60 80 Baseline 50%Reduction 100%Reduction JulianDay Chrolophylla(ug/l) 100 200 300 0 5 10 15 20 Baseline 50%Reduction 100%Reduction JulianDay DOS(ug/l) 100 200 300 0 500 1000 1500 2000 Baseline 50%Reduction 100%Reduction JulianDay TKN(ug/l) 100 200 300 0 200 400 600 800 1000 Baseline 50%Reduction 100%Reduction JulianDay PO4(ug/l) 100 200 300 0 100 200 300 Baseline 50%Reduction 100%Reduction Figure 6-25. Temporal water quality variati ons at CH011 station in 2000 (Baseline, 50 %, and 100 % nutrients load reduction from the Peace and Myakka Rivers)

PAGE 158

135 JulianDay CBOD(mg/l) 100 200 300 0 5 10 15 20 Baseline 50%Reduction 100%Reduction JulianDay DO(mg/l) 100 200 300 0 5 10 15 Baseline 50%Reduction 100%ReductionSurface JulianDay DO(mg/l) 100 200 300 0 5 10 15 Baseline 50%Reduction 100%ReductionBottom JulianDay Chrolophylla(ug/l) 100 200 300 0 5 10 15 20 Baseline 50%Reduction 100%Reduction JulianDay NH4(ug/l) 100 200 300 0 50 100 Baseline 50%Reduction 100%Reduction JulianDay TKN(ug/l) 100 200 300 0 200 400 600 800 1000 Baseline 50%Reduction 100%Reduction JulianDay DOS(ug/l) 100 200 300 0 500 1000 1500 2000 Baseline 50%Reduction 100%Reduction JulianDay PO4(ug/l) 100 200 300 0 20 40 60 80 100 Baseline 50%Reduction 100%Reduction Figure 6-26. Temporal water quality variati ons at CH013 station in 2000 (Baseline, 50 %, and 100 % nutrients load reduction from the Peace and Myakka Rivers)

PAGE 159

136 In spite of reducing nutrien t loads at the river bounda ry, the low DO condition at the bottom layer was maintained at the CH005 station because of strong river discharge. The river discharge formed the vertical stra tification which preven ted DO at the surface layer from transferring to the bottom layer. In the meantime, without sufficient DO sources from surface water, SOD continuously consumed DO at the bottom layer. The vertical stratification and S OD mainly contributed to the hypoxia in Upper Charlotte Harbor. On the other hand, at the stat ion CH009, there were no hypoxia phenomena during the strong river discharge events. The river discharge wa s insufficient to create the vertical stratification at the CH009 station (which is fa r from rivers). Even though SOD depleted DO at the bottom, DO at the surface la yer transferred to the bottom layer due to the vertical mixing. As can be seen in Fi gure 6-28, the DO values at the CH009 station were always over 4.5 mg/l in the cases of baseline, 50 % nut rient load reduction, and 100 % nutrient load reduction simulations. JulianDay SOD(gO2/m2/d) 260 265 270 275 280 0 0.5 1 1.5 2 CH005Baseline 50%Reduction 100%Reduction JulianDay SOD(gO2/m2/d) 260 265 270 275 280 0 0.5 1 1.5 2 CH009Baseline 50%Reduction 100%Reduction Figure 6-27. Temporal SOD values at CH005 and CH009 in 2000 (Baseline, 50 %, and 100 % nutrients load reduction from the Peace and Myakka Rivers)

PAGE 160

137 JulianDay DO(mg/l) 260 265 270 275 280 3 4 5 Baseline 50%Reduction 100%ReductionCH005(Bottom) JulianDay DO(mg/l) 260 265 270 275 280 5 5.5 6 6.5 7 Baseline 50%Reduction 100%ReductionCH009(Bottom) Figure 6-28. Temporal DO values at CH005 and CH009 in 2000 (Baseline, 50 %, and 100 % nutrients load reduction from the Peace and Myakka Rivers) 6.7 Sensitivity Test for Water Quality Model Water quality simulation in the Upper Char lotte Harbor depends on many factors: nutrients at the sediment column, open bounda ry nutrients, river boundary nutrients, and grid resolution. Several sensitivity simulations were performed in order to test the effects of each factor (Table 6-7). First, nutrients in the sediment column are considered to be a main source of particulate nutrients in the system. During th e erosion event caused by wind or currents, the sediment and particulate nutrients rise and suspend in the water column. These suspended particulate nutrients react to the dissolved nutri ents as sorption/desorption. Tests were designed to measure how changed sediment nutrients affect the study area: doubling the original nutrient va lues (Test 1) and halving th e original nutrient values

PAGE 161

138 (Test 2). As shown in Table 6-8, the results of Test 2 indicate that the average ROC score is lower than the baseline simulation due to the reduced ROC score of dissolved nitrogen and phosphorous. As the nutrients in the sedi ment column decrease, so do the eroded particulate nutrients. The number of particulate nutrients in the water column has a direct correlation to the number of soluble nutrien ts accounted for by the diminishing of desorption. This suggests that some of the dissolved nutrients in the water column are very likely from nutrients at the sediment layer. In Test 1, the TKN noticeably improved because of the acquisition of higher nutrients at the sediment layer. There is a definite possibility of underestimation regarding the amount of nitr ogen at the sediment layer. Second, the nutrients from river boundaries play a major role in nutrient dynamics in the water column. The stations near th e river runoff are mainly affected by river nutrients. In order to estimate the influence in terms of altered river nutrients, Test 3 and Test 4 were created; Test 3 ha s increased nutrient values to twice the original river inflow nutrient values, and Test 4 has decreased nutrient values to half the original river inflow nutrient values. As can be seen in Table 6-8, the comprehensive average is similar to the baseline simulation for the test of mounting river nutrients. Howeve r, for the test of lessening river nutrients, the nut rient ROC scores are lower than the baseline simulation, especially that of di ssolved silica, known as a primary source for diatom production. This indicates that the variations of soluble silica and diatom in the water column are probably involved in association with alternated river nutrients. Third, like river inflow nutrients, nu trients from the open boundary are also contemplated to be one of the nutrient sources in the study area. It is valuable to measure how open boundary nutrients a ffect nutrient dynamics. Tw o tests are carried out:

PAGE 162

139 doubling nutrient values from the original open boundary (Test 5) and halving nutrient values from the original open boundary (Test 6). As revealed by Table 6-8, in the case of the reduced open boundary nutrients, the average ROC score is nearly equivalent to the baseline simulation while, in the case of increased open boundary nutrients, the average ROC score is lower than the baseline simula tion. Even though the open boundary is far away from the Upper Charlotte Harbor, the nutrients at the open boundary considerably influence the nutrient dynamics in the Upper Charlotte Harbor. Fourth, water quality model performa nce seems to be affected by grid configuration, bathymetry, and resolved na vigation channels. In order to analyze the effects of how the horizontal grid resoluti on causes simulated water quality transport in this study area, Test 7 was performed using a coarse grid (129 x 92), which has approximately three times coarser resoluti on than the fine grid (188 x 176) for the baseline simulation. As can be seen from th e results in Table 6-8, the ROC scores of nutrients for the coarse grid are substantially lower than the baseline simulation. This indicates that the fine grid resolved the ch annels and complex bathymetry well; thus, the model simulation better embodi ed the hydrodynamic circulation and nutrient transport. Test 8 was conducted to measure the differences between the multi-species model and the one-species model. As shown in Table 6-8, the average ROC scores of this test are nearly equivalent to that of the baseline simulati on. The same model parameters as with the multi-species model were used but without phytoplankton parameters. The lack of multispecies information may have affected the performance of the multi-species model. However, The Chlorophyll a ROC score of the multi-species model is slightly higher than that of the one-species model.

PAGE 163

140 Table 6-7. Description of sensitivity tests Test run Description Test 1 Double water quality nutrien t values at sediment layer Test 2 Halve water quality nutrien t values at sediment layer Test 3 Double water quality nutri ent values at river boundary Test 4 Halve water quality nutri ent values at river boundary Test 5 Double water quality nutri ent values at open boundary Test 6 Halve water quality nutrient values at open boundary Test 7 Use coarse grid (129 x 92) Test 8 One algal group simulation Table 6-8. The results of sensitivity test s. Values shown are the ROC score of each nutrient. Values shown in italics indi cate the difference between the baseline and test simulations with a positive value indicating improvement of test simulation and a negative value indicati ng deterioration of test simulation. Test run DO(S) DO(B) Chl. a TOC TKN NH4 SRP SOS AVG Baseline 0.727 0.843 0.779 0.774 0.675 0.506 0.986 0.726 0.752 Test 1 0.723 -0.004 0.833 -0.01 0.744 -0.035 0.701 -0.073 0.804 0.130 0.483 -0.023 0.977 -0.009 0.806 0.08 0.759 0.007 Test 2 0.736 0.009 0.845 0.001 0.799 0.021 0.791 0.017 0.562 -0.112 0.493 -0.012 0.883 -0.103 0.697 -0.029 0.726 -0.026 Test 3 0.726 -0.001 0.820 -0.023 0.774 -0.004 0.759 -0.015 0.701 0.026 0.511 0.006 0.968 -0.019 0.807 0.08 0.758 0.006 Test 4 0.739 0.012 0.850 0.007 0.759 -0.019 0.775 0.002 0.646 -0.028 0.503 -0.003 0.943 -0.044 0.638 -0.088 0.731 -0.021 Test 5 0.731 0.004 0.834 -0.009 0.727 -0.052 0.767 -0.007 0.527 -0.148 0. 494 -0.011 0.988 0.002 0.653 -0.074 0.715 -0.037 Test 6 0.745 0.018 0.856 0.013 0.810 0.031 0.777 0.003 0.643 -0.032 0.502 -0.004 0.986 0.0 0.729 0.002 0.756 0.004 Test 7 0.720 -0.007 0.826 -0.017 0.780 0.001 0.780 0.006 0.548 -0.126 0.487 -0.018 0.886 -0.1 0.619 -0.107 0.706 -0.046 Test 8 0.755 0.028 0.861 0.018 0.765 -0.013 0.771 -0.003 0.687 0.013 0.466 -0.040 0.986 0.0 N/A 0.755 0.003

PAGE 164

141 CHAPTER 7 MODEL APPLICATION OF WATER QUALITY TO INDIAN RIVER LAGOON The CH 3D-IMS was used to simulate the Water Quality Monitoring Network (WQMN) water quality data in Indian River Lagoon (IRL) from 1997 to 1999. Sheng et al. (2002a) and Davis (2001) conducted and va lidated the hydrodynamic simulation using this model. The sediment simulation was conducted and calibrated with measured data (Sheng et al., 2003b). This chapter mainly deals with water quality simulation using multi-algal groups described in Chapter 4. Three phytoplankton groups were considered in terms of biovolume: diatom, dinoflagellatte s, and cyanobacteria (Badylak and Phlips, 2004). The main assessments of this chapter regard the calibration and validation of the IRL water quality model as well as dem onstrating the succession of phytoplankton communities using field data. 7.1 Basic Conditions The boundary fitted grid, which contains 199 x 23 horizontal cells and four vertical layers, was used for this study (Figure 7-1). The bathymetry of th e Indian River Lagoon was provided by SJRWMD. The primary method of determining bathymetry is an inverse distance interpolation. After the interpolati on and smoothing were performed, some areas were further adjusted to fit on measured data including cross-sectional areas at key flow restriction points (Davis, 2001). IRL was divided into eight segments based on hydrological and geomorphological characteristics in Figure 7-2.

PAGE 165

142 7.1.1 Pollutant Loads from Freshwater Discharge Time-varying loads of nutrients and to tal suspended solids were provided by SJRWMD using the Hydrologic Simulation Pr ogram Fortran (HSPF) model (Adkins et al., 2004). These data did not include silica and phytoplankton groups. Silica loads at the tributary mouths and watersheds were obtained by multiplying freshwater inflows by measured pollutant concentrations at or near mouths of tributaries. Badylak and Phlips (2004) collected phytoplankton samples at ei ght sites from September 1997 to August 1999. They calculated mean biovolume of major algal taxa at the ei ght sampling sites. Using this information, three phytoplankton species loads were calculated by multiplying the chlorophyll a value by the mean biovolume ratio near sampling sites. Easting(UTM,meter) Northing(UTM,meter) 520000 560000 600000 3E+06 3.05E+06 3.1E+06 3.15E+06 3.2E+06 Easting(UTM,meter) Northing(UTM,meter) 520000 560000 600000 3.05E+06 3.1E+06 3.15E+06 3.2E+061000 950 900 850 800 750 700 650 600 550 500 450 400 350 300 250 200 150 100 50 0Bathymetry(cm) Figure 7-1. Boundary-fitted grid (199 by 23) and bathymetry

PAGE 166

143 Figure 7-2. IRL segment definition (The shading distinguish each segment)

PAGE 167

144 7.1.2 Temperature The Florida Department of Environm ental Protection (FDEP) measured temperatures at eight sites inside IRL fr om 1997 to 2000 (Table 7-1). The coldest and warmest months are February and July. Due to shallowness and wind mixing, vertical stratification rarely happens except near th e river mouths. The highest, average, and lowest temperatures are 36C, 25C, and 11C, respectively. Table 7-1. Description of temperature stations Name Agency Latitude Longitude Period Banana River FDEP 28 08' 57" 80 36' 20" Hourly Fort Pierce Causeway FDEP 27 28' 18" 80 19' 28" Hourly Merritt Causeway (East) FDEP 28 21' 26" 80 38' 53" Hourly Merritt Causeway (West) FDEP 28 21' 22" 80 43' 08" Hourly Melbourne Causeway FDEP 28 05' 01" 80 35' 30" Hourly Mosquito Lagoon (NSB) FDEP 29 01' 23" 80 55' 05" Hourly Titusville Brewer Causeway FDEP 28 37' 14" 80 47' 52" Hourly Vero Bridge FDEP 27 37' 55" 80 22' 17" Hourly 7.1.3 Light and Color The University of Florida (UF) conducted twelve synop tic sampling trips between April 1997 and May 1998 inside the lagoon. The first six synoptic trips took place over a three-month span in the spring of 1997, and the second six synoptic trips over a sevenmonth span beginning in November 1997. In addition, WQMN took samples monthly at thirty-four sites throughout the lagoon between 1996 and 1999. The two data sets that contained color and light (photosynthetically active radiati on (PAR)) were used as our boundary conditions. The detail information can be found in Christian (2001) 7.1.4 Sediment Type and Nutrient Distribution UF collected a total of 219 marine sediment grab samples from IRL on November 5th to 11th, 1996 and February 7th, 1997 (Sheng, et al., 1998b). Sediment grab samples were collected from the top 10cm of the la goon bottom. The sediment size was divided

PAGE 168

145 by five different scal es: very coarse (d50 > 0.5 mm), coarse (0.25 mm< d50 < 0.5 mm), medium (0.125 < d50 < 0.25 mm), fine (0.0625 < d50 < 0.125 mm), and silts (d50 < 0.0625 mm) in Figure 7-3. Figure 7-3. Distribu tion of sediment si ze (Sheng et al., 1998b)

PAGE 169

146 Due to weak wave action, fine sediment s exist in Mosquito Lagoon, north Indian River Lagoon, and north Banana River. Some fine sediments can be found in restricted areas around Vero Beach. In the vicinity of Sebastian River and Sebastian Inlet, more fine sediments can be seen. Sediment nutrient analyses were perf ormed by Dr. Reddy in the Soil and Water Science Department, University of Florida (Reddy et al., 2001). Twenty-four sampling locations were selected to encompass a range of physical character istics of sediments while providing reasonab le spatial coverage of the lagoon. Intact sediment cores were obtained by driving a core tube into the sediment 0 -10 cm and 10 20 cm in depth in March and April of 1997. The sampling stations were subdivided into northern stations ( 13-24), mud stations (8, 9 and 11 in Melbourne area), and southern stations (1-7, 10 and 12) in Figure 7-4. The averaged sediment nutrients of 24 stations were 4.8 and 0.22 for porewater ammonium nitrogen (NH4-N) and soluble reactive phosphorous (SRP), 950 and 444 mg/kg for total nitrogen (TN) and total phosphorous (TP) resp ectively. TN and TP storage in the surface sediments (0-10 cm) was approximately 41,000 and 33,000 metric tons. There was significant seasonal variation of porewater NH4-N and SRP with highe r values in summer than winter (Reddy et al., 2001). Muddy sediment deposits in the vicinity of Melbourne were found to have elevated concentrations of porewater NH4-N and SRP, TN, and TP. Th e surface sediment of the southern stations generally in cluded higher TP than most of the other stations, with the exception of the mud zone stations. Most of the sediment phosphorous (78 %) was in the

PAGE 170

147 Ca and Mg-bound pool (apatite-P); this is due to high calcium carbona te and low organic matter content of these sediments (Reddy et al., 2001). Figure 7-4. Location of sedi ment sampling stations: Northe rn stations are 13 to 24; Mud stations are 8, 9 and 11 in Melbourne area; Southern stations are 1 to 7, 10 and 12 (Reddy et al. 2001)

PAGE 171

148 7.1.5 Open Boundary Condition There are four connections between the IR L domain and the Atlantic Ocean: Ponce de Leon Inlet, Sebastian Inlet, Ft. Pierce Inlet, and St. Lucie Inlet. These connections were chosen as open boundaries. Water leve l data used as boundary conditions were measured by the FDEP. Due to no long-term salinity measurements at the inlets, a constant oceanic salinity value of 35 ppt was applied to water flow ing into the lagoon at the oceanic boundaries. The University of Fl orida conducted four inlet sampling surveys to collect nutrient concentration and su spended sediment concentration (Sheng and Davis, 1999). These data were em ployed as open boundary conditions. 7.2 Measured Data 7.2.1 WQMN Data The IRL-WQMN (Sigua et al., 1998) was established as a coordinated multiagency project spanning the entire Indian River Lagoon system. The active participants of the project are the SJRWMD, SFWMD, Volusia County, Brevard County, Indian River County, and NASA-Dynamic. These agen cies collected a to tal of 150 stations resulting in a station nearly every 1.6 km of lagoon. After the modification in 1996, the station numbers were reduced to 22 and 10 SJ RWMD tributary statio ns (Figure 7-5 and Table 7-2). Water quality monitoring in th e Indian River Lagoon system consisted of sampling at regular intervals (monthly) fo r the following set of parameters: water temperature, salinity, color, turbidity, total suspended sediment (TSS), total organic carbon (TOC), total kjeldahl nitrogen (TKN), dissolved TKN, nitrate, total phosphorous (TP), dissolved TP, soluble reactive phosphorous (SRP), di ssolved silica, dissolved oxygen (DO), and corrected chlorophyll a

PAGE 172

149 In the north Indian River Lagoon, water qua lity is influenced by river pollutants, such as in Turnbull Creek, Big Flounder Creek Turkey Creek, Crane Creek, and the Eau Gallie River. These tributaries receive urban and agriculture land runoff and have a long history of impact from wastew ater treatment plants (Sigua et al., 1998). The water quality in the south Indian River Lagoon is influenced by river as well as tidal flushing because there are two inlets: Sebastian Inlet and Ft. Pierce Inlet. Generally, near the inlet, the phosphorous, nitrogen, and chlorophyll a were shown to have lower values (Crean, 2001). Table 7-2. Descripti on of WQMN stations Station Segment Easting (UTM, meter) Nothing (UTM, meter) IRLV05 One 508819 3208881 IRLV11 One 515130 3202732 IRLV17 One 515656 3194485 IRLML02 One 527533 3177705 IRLTBC Two 513524 3188113 IRLBFC Two 515132 3180852 IRLI02 Two 519474 3179073 IRLI07 Two 519716 3164056 IRLB02 Three 535308 3145290 IRLB04 Three 535929 3137876 IRLB06 Three 535957 3128644 IRLB09 Three 536748 3119291 IRLI13 Four 525852 3140742 IRLI15 Four 528098 3134346 IRLI18 Four 534460 3118792 IRLI21 Four 537646 3111109 IRLHUS Four 534910 3115707 IRLEGU Four 536283 3110951 IRLI23 Five 539849 3105023 IRLI27 Five 546289 3091260 IRLCCU Five 539083 3105851 IRLGUS Five 544778 3093747 IRLIRJ01 Six 554288 3074800 IRLSUS Six 550075 3081152 IRLIRJ04 Seven 560420 3063289 IRLIRJ05 Seven 561590 3059449 IRLIRJ07 Seven 562379 3055238 IRLIRJ10 Seven 559732 3064025 IRLIRJ12 Seven 562467 3054099 IRLVNC Seven 557819 3063308 IRLVMC Seven 558965 3058482 IRLVSC Seven 560908 3053569

PAGE 173

150 Easting(UTM,meter) Northing(UTM,meter) 520000 54000 0 3.15E+06 3.175E+06 3.2E+06IRLBFC IRLV05 IRLV11 IRL B04 IRLI15 IRLI13 IRLI02 IRLTBC IRLV17 IRLI07 IRLML02 IRL B02 SEG4 SEG3 SEG1 SEG2 Easting(UTM,meter) Northing(UTM,meter) 560000 580000 3E+06 3.025E+06 3.05E+06 3.075E+06IRLIRJ01 IRLIRJ12 IRLIRJ04 IRLIRJ05 IRLIRJ07 IRLIRJ10 SEG8 SEG6 SEG7 Easting(UTM,meter) Northing(UTM,meter) 540000 560000 3.05E+06 3.075E+06 3.1E+06 3.125E+06IRLSUS (125,6) IRLI15 IRLB06 IRLEGU IRLI18 IRLHUS IRLB09 IRLI21 IRLI23 IRLI27 IRLCCU IRLGUS (113,6) IRL B04 IRLVSC IRLIRJ01 IRLVNC IRLVMC SEG6 SEG4 SEG5 SEG3 Figure 7-5. Location of WQMN stations 7.2.2 FASUF Data The Department of Fisheries and Aquatic Sciences at the University of Florida (FASUF) collected phytoplankton samples at eight sampling sites on a monthly basis. Twenty collections were made from Se ptember 1997 to August 1999. Those samples were preserved with Lugol solution (Badylak and Phlips, 2004). The selected sampling sites extend throughout the lagoon (F igure 7-6). Fortunately, each sampling site is located in a different segment divided for this study (Table 7-3). The major algal groups in the IRL were generally diatoms, dinoflagellates, and cyanobacteria. The other algal groups are in significant in the lagoon. However, at the FASUF2 and FASUF8 stations, spherical green algae appeared at bloom levels.

PAGE 174

151 Phytoplankton composition and abundance show seasonal and inter-annual variations. The frequency of blooms generally occurred in the warm season, when temperatures exceeded 25 C. For example, dinoflagellat es which are toxic species reached bloom levels only during the summer season. Howeve r, diatom and cyanobacteria blooms were encountered in every month of the sampli ng period somewhere in the lagoon (Badylak and Phlips, 2004). These species were observed over a broad range of temperature. Other estuaries have been shown a similar pa ttern (Goldman and Carpenter, 1974). Figure 7-6. Location of FASUF sites (Badyl ak and Phlips, 2004)

PAGE 175

152 Table 7-3. Descripti on of FASUF stations Name Latitude Longitude Northing UTM (meter) Easting UTM (meter) Site One (FASUF1) 28 52' 41" 80 50 21" 3194515 515679 Site Two (FASUF2) 28 37' 25" 80 47' 43" 3166303 520011 Site Three (FASUF3) 28 11' 43" 80 37' 32" 3118891 536749 Site Four (FASUF4) 28 09' 01" 80 37' 11" 3113908 537338 Site Five (FASUF5) 28 03' 44" 80 34' 48" 3104166 541271 Site Six (FASUF6) 27 49' 17" 80 27' 44" 3077532 552963 Site Seven (FASUF7) 27 42' 05" 80 23' 25" 3064273 560114 Site Eight (FASUF8) 27 36' 23" 80 21' 38" 3053764 563099 7.2.3 UF Episodic Data UF designed four episodic events to study the resuspension of sediments and particulate nutrients during strong wind ev ents (Sheng et al., 1998a). Two of four episodic events were performed during the year 1999 (Table 7-4). Those two events lasted from 12 hours to several days. To collect field data for each episodic event, one or more in-situ platforms with numerous inst ruments (wave gauges, wind anemometers, auto-samplers, temperature/salinity gauges, and optical scatter sensors, etc.) were deployed at different locations in the lagoon (Figure 7-7). Da ta were sampled every hour. Table 7-4. Descripti on of episodic stations Event Name Latitude Longitude Northing UTM (meter) Easting UTM (meter) Date North Station 27 58' 54" 80 32' 10" 3095271 545608 Central Station 27 57' 27" 80 31' 36" 3092580 546551 1 South Station 27 56' 20" 80 31' 34" 3090516 546619 Feb. 19th to 22nd 1999 2 Titusville 28 41' 07" 80 48' 25" 3173130 518864 Nov. 29th to 30th 1999 7.2.4 UF Synoptic Data UF collected water quality data for calibrating and verifying the water quality model as well as providing statistical analysis of various types of data (Figure 7-7). The

PAGE 176

153 sampling stations were select ed throughout segments 2, 4 and 5 of the study area while taking into account spatial re solution and bottom sediment type (Sheng and Melanson, 1999). There were 28 stations chosen with sa mples during year 1998 in Table 7-5. These samples were collected via modified Nisk in bottles. The evaluated water quality parameters were nutrients (nitroge n, phosphorous, and si lica), chlorophyll a TSS, and dissolved oxygen. Table 7-5. Descripti on of synoptic stations Station Segment Easting (UTM, meter) Nothing (UTM, meter) Date Synoptic 1 Five 546981 3089768 Synoptic 2 Five 545566 3095209 Synoptic 3 Five 541853 3102489 Synoptic 4 Four 538385 3111063 Synoptic 5 Four 535668 3116409 Synoptic 6 Four 532621 3123786 Synoptic 7 Four 531552 3126368 Synoptic 8 Four 533774 3121038 Synoptic 9 Four 539458 3108389 Synoptic 11 Four 530319 3128949 Synoptic 12 Four 527162 3137250 Synoptic 13 Four 524710 3145923 Synoptic 14 Four 523166 3150998 Synoptic 15 Two 522338 3157459 Synoptic 16 Two 523391 3160877 Synoptic 17 Two 521354 3160411 Synoptic 18 Two 525272 3157926 Synoptic 19 Four 525402 3142915 Synoptic 21 Two 523031 3177864 Synoptic 22 Two 518796 3179610 Synoptic 23 Two 521569 3176199 Synoptic 24 Two 520758 3174720 Synoptic 25 Two 518565 3171393 Synoptic 26 Two 519551 3166317 Synoptic 27 Two 519637 3163363 Synoptic 28 Two 518814 3168624 Synoptic 29 Two 518237 3173239 Synoptic 30 Two 517004 3180992 Jan. 29th, Feb. 26th, March 26th, April 30th, and May 28th, 1998

PAGE 177

154 Easting(m) Northing(m) 560000 580000 3E+06 3.025E+06 3.05E+06 3.075E+06SEG8 SEG6 SEG7 Easting(m) Northing(m) 520000 54000 0 3.15E+06 3.175E+06 3.2E+06Epi.3 SEG4 SEG3 SEG1 SEG2 Syn.17 Syn.14 Syn.18 Syn.16 Syn.26 Syn.15 Syn.23 Syn.13 Syn.22 Syn.12 Syn.19 Syn.21 Syn.30 Syn.29 Syn.28 Syn.27 Syn.25 Syn.24 Easting(m) Northing(m) 540000 560000 3.05E+06 3.075E+06 3.1E+06 3.125E+06Syn.1 Epi.2(South) SEG6 SEG4 SEG5 SEG3 Syn.3 Syn.2 Epi.2(North) Epi.2(Central) Syn.4 Syn.5 Syn.6 Syn.9 Syn.11 Syn.8 Syn.7 Figure 7-7. Location of UF episodic and synoptic stations 7.3 Model Parameters and Calibration There are many parameters in the water quality model. It is challenging to determine model parameters because they are dependent on physical and biochemical factors, such as the geological characteristic s of estuary, temperature, tidal variation, freshwater inflows, point/non point nutrient loads, speci es of algal groups, nutrient concentrations at the sediment column, etc. Fortunately, those para meters have feasible ranges obtained from field observation, la boratory experimentation, and previous modeling studies. However, the ranges are so broad, necessitating parameter adjustment. This adjustment is complicated and takes a lot of time.

PAGE 178

155 The first step for the tuning procedure is to conduct sensitiv ity tests for water quality model parameters. These tests indicate which parameters are the most sensitive in the water quality model. In order to execute sensitivity tests, a baseline simulation is desired. The model parameters for the base line simulation are obt ained by measured values, estimation from nutrient kinetic e quations, averaged literature values, and previous model applications. In the IRL, Phlips (1999) estimated phytoplankton maximum growth rate based on collected data from eight sampling si tes and subsequent laboratory analyses. Reddy et al. (2001) conducted the sedime nt reaction rate experiment with core samples. In the Table 7-6, some parameter values of baseline simulation were a constant number and some had a range. This study used a multi-phytoplankton feature. Each species has its own parame ter values, thus the parameters which are related to the phytoplankton have a range. The partition co efficient values were estimated by TSS values and particulate nutrient concentrati on. The values are also given by variant numbers for different segments. Some studies conducted the sensitivity test s in the different study areas (Park, 2004; Yassuda, 1996). The tests were conducte d by varying each parameter within literature values to find which model parame ters more influenced the water quality model. This study conducted 48 sensitivity test s using 24 parameters. For each parameter, two sensitivity tests were performed: to incr ease twice the value of the parameter and to decrease the half that of the parameter (Tab le 7-6). Each test employed the same initial and boundary conditions as well as physical fo rces as those for the baseline simulation. The 90-day simulation was conducted and comp ared with the baseline simulation.

PAGE 179

156 Table 7-6. Description of sensitivity tests Test runs Parameter Literature range Baseline values Test values Unit Test_agrm1 Test_agrm2 Maximum algae growth rate 0.2 5.0 1.2 2.0 2.4 4.0 0.6 1.0 1/day Test_halfn1 Test_halfn2 Nitrogen halfsaturation rate 0.005 4.34 0.025 0.035 0.05 0.07 0.0125 0.0175 mg/l Test_halfp1 Test_halfp2 Phosphorous halfsaturation rate 0.001 0.163 0.004 0.01 0.008 0.02 0.002 0.005 mg/l Test_kaex1 Test_kaex2 Algae respiration rate 0.03 0.9 0.045 0.1 0.045 0.2 0.0225 0.05 1/day Test_kas1 Test_kas2 Algae mortality rate 0.003 0.1 0.03 0.06 0.06 0.12 0.015 0.03 1/day Test_was1 Test_was2 Algae settling velocity 0 1700 5 20 10 40 2.5 10 cm/sec Test_sonm1 Test_sonm2 Ammonification rate of SON 0.001 0.4 0.008 0.011 0.016 0.022 0.004 0.0055 1/day Test_nitr1 Test_nitr2 Nitrification rate 0.004 0.11 0.03 0.07 0.06 0.14 0.015 0.035 1/day Test_dron1 Test_dron2 Sorption/desorption rate of SON 0.005 0.08 0.005 0.01 0.0025 1/day Test_dran1 Test_dran2 Sorption/desorption rate of NH4 0.01 0.02 0.01 0.02 0.005 1/day Test_sopm1 Test_sopm2 Mineralization rate of SOP 0.001 0.6 0.01 0.02 0.005 1/day Test_drop1 Test_drop2 Sorption/desorption rate of SOP 0.01 0.08 0.01 0.02 0.005 1/day Test_drip1 Test_drip2 Sorption/desorption rate of SRP 0.01 0.02 0.01 0.02 0.005 1/day Test_dros1 Test_dros2 Sorption/desorption rate of DOS 0.005 0.1 0.01 0.02 0.005 1/day Test_akd1 Test_akd2 Oxidation rate 0.02 0.6 0.04 0.08 0.08 0.16 0.02 0.04 1/day Test_zgrm1 Test_zgrm2 zooplankton growth rate 0.1 0.3 0.12 0.14 0.24 0.28 0.06 0.07 1/day Test_halfa1 Test_halfa2 Half-saturation rate for zooplankton 0.01 2.0 0.8 1.6 0.4 mg/l Test _kzex1 Test_kzex2 zooplankton respiration rate 0.001 0.16 0.03 0.06 0.015 1/day Test_kzs1 Test_kzs2 zooplankton mortality rate 0.001 0.125 0.02 0.04 0.01 1/day Test_pcon1 Test_pcon2 Partition coefficient (SON and PON) 5.E-6 5.E-3 2.E-4 5.E-3 4.E-4 1.E-2 1.E-4 2.5E-3 1/ g Test_pcan1 Test_pcan2 Partition coefficient (NH4 and PIN) 5.E-6 5.E-3 5.E-4 3.E-3 1.E-3 6.E-3 2.5E-4 1.5E-3 1/ g Test_pcop1 Test_pcop2 Partition coefficient (SOP and POP) 8.E-6 5.E-3 1.E-4 1.E-3 2.E-4 2.E-3 5.E-5 5.E-4 1/ g Test_pcip1 Test_pcip2 Partition coefficient (SRP and PIP) 1.E-6 1.E-3 5.E-5 7.E-4 1.E-4 1.4E-3 2.5E-5 3.5E-4 1/ g Test_pcos1 Test_pcos2 Partition coefficient (DOS and POS) 3.E-5 4.E-4 6.E-5 8.E-4 1.5E-5 2.E-4 1/ g

PAGE 180

157 The relative error was employe d as a skill assessment to indicate the discrepancy between the baseline and sensitivity test simulations. N i i N i i i relBaseline testySensitivit Baseline E1 1 (7-1) where N is the total number of model predictions (i=1,2,3,,N). The results of the sensitivity tests showed all parameters directly or indirectly affect water quality species (Table 7-7). The maximum algae growth rate is the most sensitive parameter in the water quality model. Due to the increasing or decreasing of the parame ter, the biomass of phytoplankton rapidly changes. This change in fluences the phytoplan kton nutrients, such as nitrogen, phosphorous, and sili ca. In addition, DO is dir ectly altered by phytoplankton biomass due to its photosynthesis. However, DO is more sensitive to oxidation rate than maximum algae growth rate. Not only maximum algae growth rate but also parameters related directly with phytopl ankton biomass are also important. Those parameters include algae respiration/mortality ra te, settling velocity, nitrogen /phosphorous half saturation, zooplankton maximum growth rate, and zooplankton respiration/mortality rate. Despite the fact that ammonification ra te of SON (Test_sonm2) indicates the highest averaged relative error, ammonifica tion rate of SON is considered the second most important parameter because the highest averaged relative error is in part caused by a specific nutrient (NO3). SON is rapidly mineralized to ammonium nitrogen. The ammonium nitrogen converts to NO3 due to nitrification. In a relative manner, the value of Nitrite+nitrate (NO3) is low, so the relative error of NO3 is generally much higher than that of other nutrients. The other parameters (desorption rate of NH4, partition coefficient

PAGE 181

158 of NH4/PIN, and nitrification rate) related with NO3 also show high averaged relative error. Among the parameters related with phosphor ous, the mineralization rate of SOP and partition coefficient of SRP/PIP beco me the most important parameters for sensitivity tests. As the mineralization ra te increases, so does SRP because SOP is mineralized to SRP. More SRP in the water column leads to a more enlarged phytoplankton biomass. Conversely, phyt oplankton biomass decreases as does mineralization rate of SOP. The partit ion coefficient of SRP/PIP represents sorption/desorption reaction between SRP and PIP in Appendix E. The coefficient is dependent on sorption/desorption rate for SR P, sediment concentration, and SRP/PIP concentration in the water column. An increas e in the partition coefficient corresponds in a decreased SRP, hence limiting phytoplankton uptake. Results of sensitivity tests for other parameters indicate that each parameter closely related to the specific water quality species plays a major role in the species. For example, the partition coefficients for phos phorous and nitrogen mainly affect the particulate phosphorous and ni trogen, respectively. The ch ange of sorption/desorption rate for nutrients (silica, nitrogen, and phosphorous) spec ifically alters dissolved nutrients. Half-saturation constants (HAL FN, HALFP, and HALFA) show a major impact in the growth of phytoplankton. A ccording to the sensitivity tests, some parameters which are directly related to phytoplankton and main nutrients appeared to greatly affect the entire water quality model. The other parameters influenced the entire water quality state variables but to a relatively small degree. However, these parameters also play a significant role in the specific water quality species.

PAGE 182

159 Table 7-7. Sensitivity analysis results Test runs DO Chl. a TOC NO3 DTKN SRP PN PP SOS AVG Baseline 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Test_agrm1 1.65 9.72 0.91 93.14 2.09 34.83 0.47 0.62 7.43 16.76 Test_agrm2 2.40 29.62 1.89 67.72 5.06 36.78 0.66 1.02 11.05 17.36 Test_halfn1 0.60 2.99 0.19 12.23 0.70 7.94 0.07 0.10 1.67 2.94 Test_halfn2 0.45 2.47 0.11 6.09 0.40 5.47 0.04 0.07 1.87 1.88 Test_halfp1 0.47 6.10 0.51 30.2 1.01 11.21 0.25 0.34 4.32 6.05 Test_halfp2 0.34 4.69 0.33 28.02 0.62 0.68 0.21 0.24 2.50 5.07 Test_kaex1 1.34 19.01 1.07 33.13 1.85 21.63 0.57 0.62 10.02 9.92 Test_kaex2 0.73 7.91 0.52 19.66 0.91 15.08 0.33 0.32 6.46 5.77 Test_kas1 0.65 24.92 1.30 20.74 2.09 6.59 0.88 0.55 3.48 6.80 Test_kas2 0.53 9.81 1.32 12.56 1.67 4.05 0.80 0.49 1.89 3.68 Test_was1 0.18 12.39 0.81 6.37 0.77 1.84 0.59 0.53 1.52 2.78 Test_was2 2.37 7.62 13.89 5.70 0.49 5.78 6.63 5.63 2.29 5.60 Test_sonm1 1.94 3.20 0.42 69.03 12.38 28.61 3.78 0.2 4.28 13.76 Test-sonm2 0.92 5.20 0.62 260 8.65 16.64 3.15 0.31 4.55 33.40 Test_nitr1 0.13 0.18 0.02 50.71 1.48 0.83 0.44 0.02 0.15 6.00 Test_nitr2 0.08 0.11 0.01 96.14 0.7 0.46 0.24 0.01 0.08 10.87 Test_dron1 0.60 1.06 0.15 31.43 18.90 7.98 5.03 0.08 0.99 7.36 Test_dron2 0.86 1.24 0.14 14.96 11.31 10.42 5.11 0.11 1.33 5.05 Test_dran1 0.28 1.08 0.12 108 3.41 3.90 1.07 0.06 0.85 13.21 Test_dran2 0.22 0.61 0.06 48.45 4.39 2.36 1.27 0.03 0.55 6.44 Test_sopm1 0.37 5.86 0.7 55.73 0.9 15.00 0.48 0.97 5.51 9.50 Test_sopm2 0.27 5.70 0.51 28.25 0.79 12.46 0.27 0.64 3.53 5.82 Test_drop1 0.09 2.09 0.16 9.00 0.21 2.38 0.11 1.36 1.11 1.83 Test_drop2 0.07 1.54 0.13 7.21 0.15 1.76 0.11 1.27 0.95 1.47 Test_drip1 0.27 3.55 0.24 13.55 0.45 30.60 0.14 1.14 1.67 5.73 Test_drip2 0.19 2.34 0.17 10.91 0.32 21.50 0.11 1.16 1.02 4.19 Test_dros1 0.00 0.00 0.00 0.01 0.00 0.00 0.00 0.00 12.36 1.37 Test_dros2 0.00 0.00 0.00 0.01 0.00 0.00 0.00 0.00 7.15 0.79 Test_akd1 7.74 0.84 12.29 11.86 0.25 11.09 0.07 0.46 0.73 5.04 Test_akd2 5.31 0.97 6.40 7.65 0.20 17.10 0.05 0.53 0.88 4.34 Test_zgrm1 0.41 76.08 0.25 17.59 0.78 5.35 0.20 0.30 3.02 11.55 Test_zgrm2 0.24 16.88 0.10 5.47 0.21 1.51 0.07 0.12 1.18 2.86 Test_halfa1 0.19 14.45 0.08 4.56 0.18 1.23 0.06 0.09 0.99 2.43 Test_halfa2 0.28 40.35 0.13 10.04 0.40 2.49 0.11 0.18 1.87 6.20 Test_kzex1 0.28 10.76 0.09 3.85 0.18 2.04 0.07 0.18 0.68 2.01 Test_kzex2 0.24 9.59 0.08 3.29 0.14 1.87 0.05 0.16 0.4 1.76 Test_kzs1 0.34 12.15 0.39 3.77 0.21 1.81 0.28 0.23 0.75 2.22 Test_kzs2 0.42 15.42 0.56 4.46 0.34 2.38 0.31 0.34 0.63 2.76 Test_pcon1 0.63 1.10 0.17 37.11 21.42 8.57 5.45 0.09 1.07 8.40 Test_pcon2 0.90 1.27 0.15 16.48 12.11 11.12 5.44 0.12 1.41 5.44 Test_pcan1 0.32 1.21 0.14 119.7 3.63 4.58 1.15 0.07 0.98 14.64 Test_pcan2 0.24 0.65 0.07 49.86 4.61 2.72 1.33 0.03 0.60 6.68 Test_pcop1 0.11 2.51 0.20 11.25 0.27 3.28 0.13 1.63 1.35 2.30 Test_pcop2 0.12 2.31 0.17 17.89 0.49 4.04 0.10 2.32 1.34 3.20 Test_pcip1 0.38 3.82 0.44 26.67 0.88 47.53 0.23 1.37 3.45 9.42 Test_pcip2 0.25 2.15 0.29 24.22 0.56 27.10 0.19 1.44 2.08 6.48 Test_pcos1 0.00 0.03 0.01 0.17 0.01 0.07 0.00 0.00 23.29 2.62 Test_pcos2 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 11.31 1.26

PAGE 183

160 The second step of model calibration is a trial-and-error parameter adjustment procedure. Based on the sensi tivity tests, the most important parameters, such as maximum algae growth rate, ammonification rate of SON, minerali zation rate of SOP, algae respiration rate, sorpti on/desorption rate for nutrients, and algae mortality are adjusted first for all water quality species. The other parameters are attuned as partially sensitive parameters for each specific water qua lity species. To calibrate the water quality model for IRL appropriately, model parameters are slightly different at each segment because of distinctive inflow, loads of nutrients and hydrological/geomorphological characteristics. The calibration procedure involves optimiza tion of numerical m easures (objective functions) that compare obser vations of the state of th e system with corresponding simulated predictions. The relative e rror and correlation coefficient (R2) were employed as major objective functions. Those objective functions indicate the average discrepancy and overall model performan ce between measured data and model predictions. N i i N i iiO PO error relative1 1 (7-2) )* (*)* ( )***(2 1 2 2 1 2 2 1 2PNPONO PONPO RN i N i i i N i i (7-3) where N is the total number of observati ons or predictions (i=1,2,3,,N), Oi is the observation, Pi is the simulated prediction, Ois the average of observation, and P is the average of simulation prediction. More than 200 simulations were made during a trialand-error calibration procedure. During th ese simulations, model parameters were attuned as well as new processes added or modified in the water quality model when

PAGE 184

161 necessary. Table 7-8 provides model parameters in the IRL as well as feasible ranges of those parameters obtained from field observa tion, laboratory experimentation, literature and previous modeling studies. Table 7-8. Water quality model coe fficients used for the IRL simulation Symbol Coefficient Unit Literature range IRL Model (dia)T-20 Temperature coefficient for diatom growth (1.01 1.2)** 1.066 (dino)T-20 Temperature coefficient for dinoflagellates growth (1.01 1.2)** 1.066 (cyano)T-20 Temperature coefficient for cynobacteria growth (1.01 1.2)** 1.066 (AD)T-20 Temperature coefficient for NH4 desorption 1.08 1.08 (AI)T-20 Temperature coefficient for Ammonium instability 1.08 1.07 (BOD)T-20 Temperature coefficient for CBOD oxidation 1.02 1.15 1.08 (DN)T-20 Temperature coefficient for denitrification 1.02 1.09 1.045 (NN)T-20 Temperature coefficient for nitrification 1.02 1.08 1.08 (OD)T-20 Temperature coefficient for SON desorption 1.08 1.08 (ONM)T-20 Temperature coefficient for mineralization 1.02 1.09 1.07 (RESP)T-20 Temperature coefficient for algae respiration 1.045 1.04 (z)T-20 Temperature coefficient for zooplankton growth 1.01 1.2 1.04 dia Diatom maximum growth rate 1/day 0.55 5.0 1.8 2.0 dino Dinoflagellates maximum growth rate 1/day 0.2 2.16 1.2 1.6 cyano Cyanobacteria maximum growth rate 1/day 0.2 4.9 0.8 1.6 z Zooplankton maximum growth rate 1/day 0.1 0.3 0.12 0.14 achla Algal carbon and chlorophyll a ratio mg C / mg Chla 20 1000 75 anc Algal nitrogen and carbon ratio mg N / mg C 0.05 0.43 0.15 apc Algal phosphorous and carbon a ratio mg P / mg C 0.005 0.03 0.025 aoc Algal oxygen and carbon ratio mg O / mg C 2.67 2.667

PAGE 185

162 Table 7-8. Continued Symbol Coefficient Unit Literature range IRL Model adsc Diatom silica and carbon ratio mg Si / mg C 0.06 0.77 0.5 dan Desorption rate of adsorbed ammonium nitrogen 1/day 0.01 0.02 0.005 0.01 don Desorption rate of adsorbed organic nitrogen 1/day 0.005 0.08 0.005 0.01 dos Desorption rate of adsorbed organic silica 1/day 0.005 0.1 0.01 dip Desorption rate of adsorbed inorganic phosphorous 1/day 0.01 0.02 0.01 dop Desorption rate of adsorbed organic phosphorous 1/day 0.01 0.08 0.01 dmol Molecular diffusion coefficient for dissolved species cm2/s 4.E-6 1.E-5 1.E-5 fdCBOD Fraction of dissolved CBOD 0.5 0.5 Hbod Half-saturation constant for CBOD oxidation mg O2 0.02 5.6 0.5 Hn_dia Half-saturation constant for diatom uptake nitrogen mg/l 0.015 0.12 0.025 0.035 Hn_dino Half-saturation constant for dinoflagellates uptake nitrogen mg/l 0.005 0.13 0.025 0.03 Hn_cyano Half-saturation constant for cyanobacteria uptake nitrogen mg/l 0.01 4.34 0.025 0.035 Hp_dia Half-saturation constant for diatom uptake phosphorous mg/l 0.001 0.163 0.004 0.005 Hp_dino Half-saturation constant for dinoflagellates uptake phosphorous mg/l 0.06 (0.001 0.03)** 0.005 0.01 Hp_cyano Half-saturation constant for cyanobacteria uptake phosphorous mg/l 0.0025 -0.02 0.004 0.005 Hs Half-saturation constant for diatom uptake silica mg/l 0.08 0.1 0.08 Ha Half-saturation constant for zooplankton mg/l 0.01 2.0 0.8 Hnit Half-saturation constant for nitrification mg O2 0.1 2.0 2.0 hv Henrys constant mg/l atm 43.8 45 45 Idia Optimum light intensity for diatom growth E/m2/s 88 350 300 Idino Optimum light intensity for dinoflagellates growth E/m2/s (300 350)** 300 Icyano Optimum light intensity for cyanobacteria growth E/m2/s 43 600 300 (NH3)air Ammonia concentration in the air g/l 0.1 0.1

PAGE 186

163 Table 7-8. Continued Symbol Coefficient Unit Literature range IRL Model Kax_dia Respiration rate by diatom 1/day 0.03 0.6 0.08 0.1 Kax_dino Respiration rate by dinoflagellates 1/day 0.047 0.047 Kax_cyano Respiration rate by cyanobacteria 1/day 0.03 0.9 0.045 0.065 Kas_dia Mortality rate by diatom 1/day 0.03 (0.003 0.1)** 0.04 0.07 Kas_dino Mortality rate by dinoflagellates 1/day (0.003 0.1)** 0.03 0.06 Kas_cyano Mortality rate by cyanobacteria 1/day (0.003 0.1)** 0.03 0.06 KD CBOD oxidation rate 1/day 0.02 0.6 0.03 0.06 KDN Denitrification rate constant 1/day 0.0 1.0 0.09 KNN Nitrification rate constant 1/day 0.004 0.11 0.03 0.07 Kon Ammonification rate of SON 1/ day 0.001 0.4 0.008 0.011 Kop Mineralization rate of SOP 1/day 0.001 0.6 0.01 Kvol Constant rate for nitrogen volatilization 1/day 3.5 9.0 7.0 Kzx Respiration rate of zoopla nkton 1/day 0.001 0.16 0.02 Kzs Mortality rate of zooplankton 1/day 0.0065 0.0326 0.03 Pan Partition coefficient between NH4 and PIN 1/ g 1.E-6 5.E-3 2.E-4 3.E-3 Pon Partition coefficient between SON and PON 1/ g 5.E-6 5.E-3 2.E-4 5.E-3 Pip Partition coefficient between SRP and PIP 1/ g 1.E-6 1.E-3 5.E-5 6.E-4 Pop Partition coefficient between SOP and POP 1/ g 8.E-6 5.E-3 1.E-4 1.E-3 Pos Partition coefficient between DOS and POS 1/ g 3.E-5 4.E-4 WSdia Diatom settling velocity cm/day 1 1700 10 WSdino Dinoflagellates settling velocity cm/day 280 600 (0 3000)** 10 20 WScyano Cyanobacteria settling velocity cm/day 0 20 8 ** All phytoplankton 7.4 Water Quality Mode l Simulation 7.4.1 Long-Term Simulation The water quality model was conducted from September 2nd 1997 to November 31st 1999 with a hydrodynamic and sediment model. The first three month simulation is considered as a spin-up simulation that make s the surface elevation, currents, salinity,

PAGE 187

164 temperature, suspended sediment, and nutrien ts reach a dynamic steady state. The initial conditions of hydrodynamic, sediment, and nutrients are obtained by a variety of organizations, such as SJRWMD, SFWMD, USGS, FDEP, NOAA, NCDC, and COOP. These values are interpolated to the enti re computational domain, with a weighting function inversely proportional to the square of the distance. The simulated water quality species are: SRP, SOP, POP, PIP, NH4, NH3, NO3, SON, PON, PIN, CBOD, DO, diatom, dinoflagellate, cya nobacteria, and zooplankton. Figures 7-9 through 7-15 show the simulate d results and measured data at IRLV17 (Segment 1), IRLI07 (Segment 2), IRLB09 (Segment 3), IRLI18 (Segment 4), IRLI23 (Segment 5), IRLIRJ01 (Segment 6), and IRLIRJ05 (Segment 7). Not only was each station located far away from the rivers to better represent the ch aracteristics of each segment, it also was close to FASUF stations Each figure embodies ten different time series plots of water quality species (oxygen, chlorophyll a, NO3, SPR, dissolved silica, TSS, TOC, dissolved TKN, particulate ni trogen, and particulate phosphorous). The rectangular shape symbolizes measured water quality state variables. On the other side, the solid line represents the simulated water quality state variables. To quantify the differences between simulated results and measured data, two skill assessments were employed. The first skill assessment was the relative error (Erel). N i i N i i i reldata Measured prediction Model data Measured E1 1 (7-4) where N is the total number of model predic tions or observations (i=1,2,3,,N). The relative errors for disso lved oxygen are less than 22 % in all stations in Table 7-9. In the case of chlorophyll a, the range of relative error is 44 to 68%. The relative errors for

PAGE 188

165 Dissolved TKN and TOC show a good agreement of simulated predictions and measured values. The results of particulate species range from 45 to 69 %, from 50 to 104 %, and from 49 to 134 % for nitrogen, silica, and phosphorous, respectively. The relative errors of particulate species are worse than those of dissolved species. The reasons are in part the less accurate erosion and deposition mechanis m of particulate species. As we can see, the relative errors for TSS have also high values (62 to 90 %). Table 7-9. Relative error of each station Station DO TSS Chlorophyll a TOC NO3 IRLV17 0.16 0.65 0.58 0.40 0.68 IRLI07 0.17 0.90 0.48 0.37 0.92 IRLB09 0.18 0.74 0.50 0.26 1.06 IRLI18 0.19 0.80 0.62 0.25 0.63 IRLI23 0.22 0.85 0.55 0.42 0.63 IRLIRJ01 0.22 0.83 0.68 0.43 0.56 IRLIRJ05 0.17 0.77 0.62 0.39 0.74 Table 7-9. Continued Station DTKN SRP Particulate Nitrogen Dissolved Silica Particulate Phosphorous IRLV17 0.39 0.44 0.65 0.50 1.04 IRLI07 0.19 0.77 0.61 1.04 0.72 IRLB09 0.30 0.80 0.55 0.77 0.64 IRLI18 0.19 0.71 0.69 0.82 0.63 IRLI23 0.23 0.90 0.65 0.81 0.50 IRLIRJ01 0.39 0.57 0.45 0.52 0.51 IRLIRJ05 0.29 0.40 0.52 0.70 1.35 The second skill assessment is the relative operating characteristic (ROC). The ROC is a representation of the skill in which true positive fraction (TPF) and false positive fraction (FPF) are compared. True pos itive fraction is simply sensitivity, and false positive fraction is the same as 1sp ecificity. Sensitivity is plotted against the corresponding 1-specificity to ge nerate the ROC curve. The de tailed explanation of this skill assessment can be found in previous section 5 of Chapter 6.

PAGE 189

166 cases positive actuallyofNumber decisions TPPositive TrueofNumber TPFySensitivit)( cases negative actuallyofNumber decisions TN Negative TrueofNumber TNFySpecificit)( The area under the curve is the most commonly used and has become known as the ROC score. If the total area under the curve is grea ter than 0.5, the model system is determined as skillful. Its range is from 1.0 (for a perfect model system) to 0.0 (for a perfectly bad model system), with 0.5 indicating no sk ill (Mason and Graham, 1999). The ROC curves and scores of each nutrient are shown in Figur e 7-8. Most of the ROC scores are over 0.6 and the maximum score is 0.862. This indicates that the water quality model is skillful for the IRL area. Due to the lack of sediment nutrient data, the ROC scores of particulate nutrients are relatively lower than for other nutrients. During the simulation period, the water qua lity species data show the seasonal variations, and these seasonal variations ar e quite well predicted by the water quality model (Figure 7-9 to 7-15). In the case of dissolved oxygen, the s easonal variations are very vivid in all stations. The DO values ar e high in the winter and decline until summer. After summer, the values increase until winter. These variations are caused by many factors, such as photosynthesis and respiration by phytoplankt on, temperature, tide and wind mixing, decomposition of organic matter, etc. On the other hand, phytoplankton biomass becomes high in the summer period be cause of sufficient light and temperature. Those conditions offer an optimal environm ent for algal growth. The information of surface light, surface water temperature, wind and river nutrient lo ads can be found in Appendix H.

PAGE 190

167 1-Specificity Sensitivity 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1DOA=0.688 1-Specificity Sensitivity 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1CHLAA=0.658 1-Specificity Sensitivity 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1TOCA=0.722 1-Specificity Sensitivity 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1SRPA=0.858 1-Specificity Sensitivity 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1DissolvedSilicaA=0.681 1-Specificity Sensitivity 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1ParticulateNitrogenA=0.591 1-Specificity Sensitivity 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1TSSA=0.607 1-Specificity Sensitivity 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1ParticulatePhosphorousA=0.510 1-Specificity Sensitivity 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1DTKNA=0.792 Figure 7-8. ROC curve for IR L nutrients from 1997 to 1999. The particulate water quality species a ppeared to follow the wind and currentinduced temporal variation of sediment settl ing and resuspension. The time series of TSS and particulate nutrients showed that erosion of sediment s leads to resuspension of particulate nutrients. The er osion rate of sediments is significantly related to the sediment type and depth. The fine sediment group is more easily eroded than the coarse group. Even though the sediment type is same, the sediment in the shallow area tends to be plainly resuspended with small forces. Under this extreme situation, particulate nitrogen and phosphorous in the water column could reach over 1000 g/l and 300 g/l,

PAGE 191

168 respectively. Particulate species in the middle of Mosquito Lagoon, Banana River, and Northern IRL Proper are likely affected by the wind-wave shear stress because these areas are far from the inlets. Thus, the curren t shear stress is relatively weaker than in other areas. In the central and southern IRL Pr oper, particulate species are affected by the current and wind-wave shear stress. The spat ial distribution of cal culated bottom shear stress can be found in Appendix I. From the spatial distributions of wate r quality species, SRP is higher in the southern IRL proper than in the northern and central IRL proper, while the values of DTKN indicate opposite aspects. This s uggests that phosphorous maybe the limiting nutrient in the northern/central IRL proper, and nitrogen coul d be the limiting nutrient in the southern IRL proper. DO concentration in Banana River and northern IRL proper maintains high values. On the other side, in Mosquito Lagoon, a lower-than-4mg/l DO value was observed from May to October. Figures 7-16 through 7-23 show three al gal simulated predictions and measured data at all FASUF stations. The measured da ta of each algal group are scaled as algal biovolume ( m3ml-1). The 2 million m3ml-1 of algal biovolumes is roughly equivalent to 10 g/l of chlorophyll a (Phlips et al., 2004). For convenience and consistency with WQMN and UF episodic data, the algal biovolumes are changed to chlorophyll a. Each figure contains four different time series (total chlorophyll a, diatom chlorophyll a, dinoflagellate chlorophyll a, and cyanobacteria chlorophyll a). The rectangular shape symbolizes measured values, while the solid line represents the simulated predictions. Diatom dominated in the Mosquito La goon (FASUF 1), northern IRL Proper (FASUF 2), and southern IRL proper (F ASUF 7 and 8). The trend of greater

PAGE 192

169 dinoflagellate dominance was shown in the Banana River (FASUF 3) and central IRL proper (FASUF 4 and 5). There were some cy anobacteria blooms in the Banana River (FASUF 3). At the FASUF 2, 4, and 5, cya nobacteria were also observed. The abundance and relative dominance of the different algal species underg o continuous change. Diatom is dominant species but declines its biomass in the summer because of weak temperature tolerance. On the other hand, dinoflagellates, which have a high temperature tolerance, become dominant in the algal communities at some segments. The model predictions represented the process of algal succession. DAY Chlorophylla(ug/l) 400 600 800 0 5 10 15 20 DAY NO3(ug/l) 400 600 800 0 20 40 60 DAY DissovedSilica(ug/l) 400 600 800 0 1000 2000 3000 4000 5000 DAY TSS(mg/l) 400 600 800 20 40 60 80 100 DAY TOC(mg/l) 400 600 800 10 20 30 DAY DissolvedTKN(ug/l) 400 600 800 0 500 1000 1500 2000 DAY ParticulateNitrogen(ug/l) 400 600 800 0 200 400 600 800 1000 DAY ParticulatePhos.(ug/l) 400 600 800 0 100 200 300 DAY DO(mg/l) 400 600 800 5 10StationIRLV17 DAY SRP(ug/l) 400 600 800 20 40 Figure 7-9. Temporal wate r quality variations at IR LV17 station from 1997 to 1999

PAGE 193

170 DAY Chlorophylla(ug/l) 400 600 800 0 5 10 15 20 DAY NO3(ug/l) 400 600 800 0 20 40 60 DAY DissovedSilica(ug/l) 400 600 800 0 1000 2000 3000 4000 5000 DAY TSS(mg/l) 400 600 800 20 40 60 80 100 DAY TOC(mg/l) 400 600 800 10 20 30 DAY DissolvedTKN(ug/l) 400 600 800 0 500 1000 1500 2000 DAY ParticulateNitrogen(ug/l) 400 600 800 0 200 400 600 800 1000 DAY ParticulatePhos.(ug/l) 400 600 800 0 100 200 300 DAY DO(mg/l) 400 600 800 5 10StationIRLI07 DAY SRP(ug/l) 400 600 800 20 40 Figure 7-10. Temporal water quality varia tions at IRLI07 sta tion from 1997 to 1999

PAGE 194

171 DAY Chlorophylla(ug/l) 400 600 800 0 5 10 15 20 DAY NO3(ug/l) 400 600 800 0 20 40 60 DAY DissovedSilica(ug/l) 400 600 800 0 1000 2000 3000 4000 5000 DAY TSS(mg/l) 400 600 800 20 40 60 80 100 DAY TOC(mg/l) 400 600 800 10 20 30 DAY DissolvedTKN(ug/l) 400 600 800 0 500 1000 1500 2000 DAY ParticulateNitrogen(ug/l) 400 600 800 0 200 400 600 800 1000 DAY ParticulatePhos.(ug/l) 400 600 800 0 100 200 300 DAY DO(mg/l) 400 600 800 5 10StationIRLB09 DAY SRP(ug/l) 400 600 800 20 40 Figure 7-11. Temporal water quality vari ations at IRLB09 station from 1997 to 1999

PAGE 195

172 DAY Chlorophylla(ug/l) 400 600 800 0 5 10 15 20 DAY NO3(ug/l) 400 600 800 0 20 40 60 DAY DissovedSilica(ug/l) 400 600 800 0 1000 2000 3000 4000 5000 DAY TSS(mg/l) 400 600 800 20 40 60 80 100 DAY TOC(mg/l) 400 600 800 10 20 30 DAY DissolvedTKN(ug/l) 400 600 800 0 500 1000 1500 2000 DAY ParticulateNitrogen(ug/l) 400 600 800 0 200 400 600 800 1000 DAY ParticulatePhos.(ug/l) 400 600 800 0 100 200 300 DAY DO(mg/l) 400 600 800 5 10StationIRLI18 DAY SRP(ug/l) 400 600 800 20 40 Figure 7-12. Temporal water quality varia tions at IRLI18 sta tion from 1997 to 1999

PAGE 196

173 DAY Chlorophylla(ug/l) 400 600 800 0 5 10 15 20 DAY NO3(ug/l) 400 600 800 0 20 40 60 DAY DissovedSilica(ug/l) 400 600 800 0 1000 2000 3000 4000 5000 DAY TSS(mg/l) 400 600 800 20 40 60 80 100 DAY TOC(mg/l) 400 600 800 10 20 30 DAY DissolvedTKN(ug/l) 400 600 800 0 500 1000 1500 2000 DAY ParticulateNitrogen(ug/l) 400 600 800 0 200 400 600 800 1000 DAY ParticulatePhos.(ug/l) 400 600 800 0 100 200 300 DAY DO(mg/l) 400 600 800 5 10StationIRLI23 DAY SRP(ug/l) 400 600 800 20 40 Figure 7-13. Temporal water quality varia tions at IRLI23 sta tion from 1997 to 1999

PAGE 197

174 DAY Chlorophylla(ug/l) 400 600 800 0 5 10 15 20 DAY SRP(ug/l) 400 600 800 20 40 60 80 100 DAY NO3(ug/l) 400 600 800 0 20 40 60 DAY DissovedSilica(ug/l) 400 600 800 0 1000 2000 3000 4000 5000 DAY TSS(mg/l) 400 600 800 20 40 60 80 100 DAY TOC(mg/l) 400 600 800 10 20 30 DAY DissolvedTKN(ug/l) 400 600 800 0 500 1000 1500 2000 DAY ParticulateNitrogen(ug/l) 400 600 800 0 200 400 600 800 1000 DAY ParticulatePhos.(ug/l) 400 600 800 0 100 200 300 DAY DO(mg/l) 400 600 800 5 10StationIRLIRJ01 Figure 7-14. Temporal water quality varia tions at IRLIRJ01 st ation from 1997 to 1999

PAGE 198

175 DAY Chlorophylla(ug/l) 400 600 800 0 5 10 15 20 DAY SRP(ug/l) 400 600 800 20 40 60 80 100 DAY NO3(ug/l) 400 600 800 0 20 40 60 DAY DissovedSilica(ug/l) 400 600 800 0 1000 2000 3000 4000 5000 DAY TSS(mg/l) 400 600 800 20 40 60 80 100 DAY TOC(mg/l) 400 600 800 10 20 30 DAY DissolvedTKN(ug/l) 400 600 800 0 500 1000 1500 2000 DAY ParticulateNitrogen(ug/l) 400 600 800 0 200 400 600 800 1000 DAY ParticulatePhos.(ug/l) 400 600 800 0 100 200 300 DAY DO(mg/l) 400 600 800 5 10StationIRLIRJ05 Figure 7-15. Temporal water quality varia tions at IRLIRJ05 st ation from 1997 to 1999

PAGE 199

176 JulianDay Chlorophylla(ug/l) 300 400 500 600 700 800 900 0 10 20 30 JulianDay Diatom(ug/l) 300 400 500 600 700 800 900 0 10 20 30 JulianDay Cyanobacteria(ug/l) 300 400 500 600 700 800 900 0 1 2 3 4 5 JulianDay Dinoflagellates(ug/l) 300 400 500 600 700 800 900 0 10 20 30 Figure 7-16. Temporal water quality varia tions at FASUF1 st ation from 1997 to 1999 JulianDay Cyanobacteria(ug/l) 300 400 500 600 700 800 900 0 1 2 3 4 5 JulianDay Dinoflagellates(ug/l) 300 400 500 600 700 800 900 0 10 20 30 JulianDay Chlorophylla(ug/l) 300 400 500 600 700 800 900 0 10 20 30 JulianDay Diatom(ug/l) 300 400 500 600 700 800 900 0 10 20 30 Figure 7-17. Temporal water quality varia tions at FASUF2 st ation from 1997 to 1999

PAGE 200

177 JulianDay Cyanobacteria(ug/l) 300 400 500 600 700 800 900 0 1 2 3 4 5 JulianDay Dinoflagellates(ug/l) 300 400 500 600 700 800 900 0 10 20 30 JulianDay Chlorophylla(ug/l) 300 400 500 600 700 800 900 0 10 20 30 JulianDay Diatom(ug/l) 300 400 500 600 700 800 900 0 10 20 30 Figure 7-18. Temporal water quality varia tions at FASUF3 st ation from 1997 to 1999 JulianDay Cyanobacteria(ug/l) 300 400 500 600 700 800 900 0 1 2 3 4 5 JulianDay Dinoflagellates(ug/l) 300 400 500 600 700 800 900 0 10 20 30 JulianDay Chlorophylla(ug/l) 300 400 500 600 700 800 900 0 10 20 30 JulianDay Diatom(ug/l) 300 400 500 600 700 800 900 0 10 20 30 Figure 7-19. Temporal water quality varia tions at FASUF4 st ation from 1997 to 1999

PAGE 201

178 JulianDay Cyanobacteria(ug/l) 300 400 500 600 700 800 900 0 1 2 3 4 5 JulianDay Dinoflagellates(ug/l) 300 400 500 600 700 800 900 0 10 20 30 JulianDay Chlorophylla(ug/l) 300 400 500 600 700 800 900 0 10 20 30 JulianDay Diatom(ug/l) 300 400 500 600 700 800 900 0 10 20 30 Figure 7-20. Temporal water quality varia tions at FASUF5 st ation from 1997 to 1999 JulianDay Cyanobacteria(ug/l) 300 400 500 600 700 800 900 0 1 2 3 4 5 JulianDay Dinoflagellates(ug/l) 300 400 500 600 700 800 900 0 10 20 30 JulianDay Diatom(ug/l) 300 400 500 600 700 800 900 0 10 20 30 JulianDay Chlorophylla(ug/l) 300 400 500 600 700 800 900 0 10 20 30 Figure 7-21. Temporal water quality varia tions at FASUF6 st ation from 1997 to 1999

PAGE 202

179 JulianDay Cyanobacteria(ug/l) 300 400 500 600 700 800 900 0 1 2 3 4 5 JulianDay Dinoflagellates(ug/l) 300 400 500 600 700 800 900 0 10 20 30 JulianDay Diatom(ug/l) 300 400 500 600 700 800 900 0 10 20 30 JulianDay Chlorophylla(ug/l) 300 400 500 600 700 800 900 0 10 20 30 Figure 7-22. Temporal water quality varia tions at FASUF7 st ation from 1997 to 1999 JulianDay Cyanobacteria(ug/l) 300 400 500 600 700 800 900 0 1 2 3 4 5 JulianDay Dinoflagellates(ug/l) 300 400 500 600 700 800 900 0 10 20 30 JulianDay Diatom(ug/l) 300 400 500 600 700 800 900 0 10 20 30 JulianDay Chlorophylla(ug/l) 300 400 500 600 700 800 900 0 10 20 30 Figure 7-23. Temporal water quality varia tions at FASUF8 st ation from 1997 to 1999

PAGE 203

180 The relative errors of each st ation can be seen in Table 7-10. Some of the measured diatom and dinoflagellate chlorophyll a are comparatively high. These values are barely measured in the WQMN stations. The high relative errors of diatom and dinoflagellate chlorophyll a may be due to the unexpected high measured values. The average relative of total chlorophyll a is nearly 0.71. The ROC scores of total chlorophyll a, diatom chlorophyll a, dinoflagellate chlorophyll a, and cyanobacteria chlorophyll a are 0.550, 0.514, 0.536, and 0.628, respectively. Table 7-10. Relative erro rs of FASUF stations Name Total chlorophyll a Diatom chlorophyll a Dinoflagellate chlorophyll a Cyanobacteria chlorophyll a FASUF 1 0.72 0.96 2.03 0.79 FASUF 2 0.52 0.83 0.95 1.04 FASUF 3 0.59 0.97 0.70 0.87 FASUF 4 0.76 0.90 0.90 0.73 FASUF 5 0.68 0.77 0.82 0.79 FASUF 6 0.74 1.11 1.43 0.60 FASUF 7 0.83 1.18 1.12 0.75 FASUF 8 0.86 0.94 1.52 0.68 Average 0.71 0.96 1.18 0.78 7.4.2 Episodic Simulation The episodic events were short term fi eld experiments designed to study the resuspension during strong wind events. To allow sufficient model spin up, the simulation started on September 2nd 1997. Figures 7-25 through 7-28 show the simulated predictions and measured data for episodic event #1 and event #2, respectively. Those figures contain ten different time series (oxygen, chlorophyll a, NH4, SPR, dissolved silica, TSS, TOC, dissolved TKN, particulat e nitrogen, and partic ulate phosphorous). The rectangular shape symbolizes measured wate r quality parameters, while the solid line represents the simulated water quality parameters.

PAGE 204

181 Table 7-11. Relative errors of episodic stations Water quality species Episodic #1 (North) Episodic #1 (Center) Episodic #1 (South) Episodic #2 DO 0.05 0.06 0.11 0.06 TSS 0.48 0.42 0.31 0.57 Chlorophyll a 0.25 0.22 0.37 0.19 TOC 0.46 0.43 0.30 0.71 NH4 0.67 0.76 0.60 0.88 DTKN 0.11 0.25 0.44 0.08 SRP 1.65 3.17 1.18 0.41 Particulate Nitrogen 0.38 0.30 0.33 0.64 Dissolved Silica 0.19 0.10 0.10 0.06 Particulate Phosphorous 0.65 0.63 1.11 0.27 All results for episodic event #1 agree with measured data with respect to the trend and scale. There is no significant peak of par ticulate nutrients due to weak wind in Figure 7-24. However, for episodic event #2, the high peak values of sediment and particulate nutrients on Julian days 334.4 are shown in th e simulated predictions and measured data even though the measured period was less than one day. These peak values are strongly related to wind (Figure 7-24). Particulate nutrients and sediment follow the same trend because particulate phosphorous and nitrogen are contained in or absorbed into sediment, and participated in the same settling and resuspended processes as sediment. Sorption-desorption processes (which play an important role in nutrient dynamics) are present for the episodic events. Due to th e fact that adsorbed nutrients in sediments are generally higher than nut rient concentration in the water column, desorption of absorbed nutrients during a resuspension even t can cause a significant increase in nutrient concentration in the water column (Chen, 1995). As shown in Figures 7-25 through 7-28, dissolved nutrients, such as DTKN, NH4, and SRP increase as do particulate nutrients augment.

PAGE 205

182 JulianDay1999 51 52 53 54 Episodic#1(NorthStation) 5m/s JulianDay1999 51 52 53 54 5m/s Episodic#1(CentralStation) JulianDay1999 51 52 53 54 5m/s Episodic#1(SouthStation) JulianDay1999 333 334 335 Episodic#2 5m/s Figure 7-24. Wind vector at stations for episodic #1 & #2

PAGE 206

183 JulianD ay1999 Sediment(mg/l) 51 52 53 0 5 10 15 20 25 30 Ju lianD ay1999 DTKN(ug/l) 51 52 53 0 200 400 600 800 1000 1200 Ju lianD ay1999 POP(ug/l) 51 52 53 0 50 100 150 JulianDay1999 PON(u/l) 51 52 53 0 100 200 300 400 JulianD ay1999 NH4(ug/l) 51 52 53 0 20 40 60 80 JulianDay1999 SRP(ug/l) 51 52 53 0 5 10 15 20 25 30 JulianD ay1999 TOC(mg/l) 51 52 53 0 5 10 15 20 Ju lianD ay1999 Oxygen(mg/l) 51 52 53 0 2 4 6 8 10 Ju lianD ay1999 DOS(ug/l) 51 52 53 0 500 1000 1500 2000 JulianDay1999 Chlorophylla(ug/l) 51 52 53 4 6 8 10 12 Figure 7-25. Temporal water quality vari ations at north st ation (episodic #1)

PAGE 207

184 JulianDay1999 Sediment(mg/l) 51 52 53 0 5 10 15 20 25 30 JulianDay1999 DTKN(ug/l) 51 52 53 0 200 400 600 800 1000 1200 JulianDay1999 POP(ug/l) 51 52 53 0 50 100 150 JulianDay1999 PON(u/l) 51 52 53 0 100 200 300 400 JulianDay1999 NH4(ug/l) 51 52 53 0 20 40 60 80 JulianDay1999 TOC(mg/l) 51 52 53 0 5 10 15 JulianDay1999 SRP(ug/l) 51 52 53 0 5 10 15 20 25 30 JulianDay1999 DOS(ug/l) 51 52 53 0 500 1000 1500 2000 JulianDay1999 Oxygen(mg/l) 51 52 53 0 2 4 6 8 10 JulianDay1999 Chlorophylla(ug/l) 51 52 53 4 6 8 10 12 Figure 7-26. Temporal water quality varia tions at central st ation (episodic #1)

PAGE 208

185 JulianDay1999 Sediment(mg/l) 51 52 53 0 5 10 15 20 25 30 JulianDay1999 DTKN(ug/l) 51 52 53 0 200 400 600 800 1000 1200 JulianDay1999 POP(ug/l) 51 52 53 0 50 100 150 JulianDay1999 PON(u/l) 51 52 53 0 100 200 300 400 JulianDay1999 NH4(ug/l) 51 52 53 0 20 40 60 80 JulianDay1999 TOC(mg/l) 51 52 53 0 5 10 15 JulianDay1999 SRP(ug/l) 51 52 53 0 5 10 15 20 25 30 JulianDay1999 DOS(ug/l) 51 52 53 0 500 1000 1500 2000 JulianDay1999 Oxygen(mg/l) 51 52 53 0 2 4 6 8 10 JulianDay1999 Chlorophylla(ug/l) 51 52 53 2 4 6 8 10 Figure 7-27. Temporal water quality vari ations at south st ation (episodic #1)

PAGE 209

186 JulianDay1999 Sediment(mg/l) 333.5 334 334.5 0 20 40 60 80 100 120 JulianDay1999 DTKN(ug/l) 333.5 334 334.5 0 200 400 600 800 1000 1200 JulianDay1999 NH4(ug/l) 333.5 334 334.5 0 20 40 60 80 JulianDay1999 SRP(ug/l) 333.5 334 334.5 0 5 10 15 20 JulianDay1999 DOS(ug/l) 333.5 334 334.5 0 1000 2000 3000 4000 5000 JulianDay1999 POP(ug/l) 333.5 334 334.5 0 50 100 150 JulianDay1999 Oxygen(mg/l) 333.5 334 334.5 0 2 4 6 8 10 JulianDay1999 TOC(mg/l) 333.5 334 334.5 0 5 10 15 JulianDay1999 PON(u/l) 333.5 334 334.5 0 200 400 600 800 1000 JulianDay1999 Chlorophylla(ug/l) 333.5 334 334.5 2 4 6 8 10 Figure 7-28. Temporal water quality variati ons at Titusville station (episodic #2) 7.4.3 Synoptic Simulation Like the long-term simulation and epis odic event simulation, not only boundary and initial conditions but also model parameters are the same for the synoptic simulation because model simulation must be consistent for proving its performance. Even though synoptic data were collected during the y ear 1998, the synoptic simulation began on September 2nd 1997 in order to have adequate spin -up time. Due to a small amount of data at each station, the relati ve errors are calculated in terms of segments which include some stations (Table 7-12).

PAGE 210

187 Table 7-12. Relative error of each segment for synoptic simulation Water quality species Segment Two Segment Four Segment Five DO 0.11 0.09 0.11 TSS 0.68 0.71 0.70 Chlorophyll a 0.79 0.63 0.79 TOC 0.48 0.28 0.18 NO3 0.80 1.77 1.16 DTKN 0.18 0.16 0.15 SRP 0.63 0.55 0.63 Particulate Nitrogen 0.95 0.74 0.78 Dissolved Silica 0.96 1.02 0.44 Particulate Phosphorous 1.00 1.14 1.24 As can be seen from Table 7-12, overall, the dissolved nutrient results are generally better than the particul ar nutrient results with the exception of NO3. The dynamic processes of particular nutrie nts include not only the reaction of particular and dissolved nutrients but also the erosion/ deposition mechanism. Therefore, the less accurate results of particular nutrients would seem to stem from the complicated dynamic processes. In the case of NO3, the large relative error can be attributed to a relatively small quantity of measured value and over-prediction. However, the temporal variation trend of model prediction for NO3 is well-matched with that of measured data in Figures 7-29 to 7-34. Figures 7-29 to 7-34 show the temporal wa ter quality variations at synoptic station 1, station 7, and station 28, respectively. Due to many synoptic stations, only two stations are selected from each segment. Those figur es are comprised of ten different water quality time series (oxygen, chlorophyll a, NO3, SPR, dissolved silica, TSS, TOC, dissolved TKN, particulate nitrogen, and particulate phosphorous). The rectangular shape symbolizes measured water quality parameters, while the solid line represents the simulated water quality parameters. The resu lts given in those fi gures depict that simulated DO, TOC, phytoplankton, SRP, dissolved silica, and DTKN are in good

PAGE 211

188 agreement with the measured values. Particul ate nutrients and sedime nt indicate similar variations because they are affected by th e same erosion and deposition mechanisms. DAY Chlorophylla(ug/l) 400 450 500 0 5 10 15 20 DAY NO3(ug/l) 400 450 500 0 20 40 60 DAY DissovedSilica(ug/l) 400 450 500 0 1000 2000 3000 4000 5000 DAY TSS(mg/l) 400 450 500 20 40 60 80 100 DAY TOC(mg/l) 400 450 500 10 20 30 DAY DissolvedTKN(ug/l) 400 450 500 0 500 1000 1500 2000 DAY ParticulateNitrogen(ug/l) 400 450 500 0 200 400 600 800 1000 DAY ParticulatePhos.(ug/l) 400 450 500 0 100 200 300 DAY DO(mg/l) 400 450 500 5 10Synoptic#1 DAY SRP(ug/l) 400 450 500 20 40 Figure 7-29. Temporal water quality variati ons at Synoptic #1 stat ion (Segment Five)

PAGE 212

189 DAY Chlorophylla(ug/l) 400 450 500 0 5 10 15 20 DAY NO3(ug/l) 400 450 500 0 20 40 60 DAY DissovedSilica(ug/l) 400 450 500 0 1000 2000 3000 4000 5000 DAY TSS(mg/l) 400 450 500 20 40 60 80 100 DAY TOC(mg/l) 400 450 500 10 20 30 DAY DissolvedTKN(ug/l) 400 450 500 0 500 1000 1500 2000 DAY ParticulateNitrogen(ug/l) 400 450 500 0 200 400 600 800 1000 DAY ParticulatePhos.(ug/l) 400 450 500 0 100 200 300 DAY SRP(ug/l) 400 450 500 20 40 DAY DO(mg/l) 400 450 500 5 10Synoptic#2 Figure 7-30. Temporal water quality variati ons at Synoptic #2 stat ion (Segment Five)

PAGE 213

190 DAY NO3(ug/l) 400 450 500 0 20 40 60 DAY DissovedSilica(ug/l) 400 450 500 0 1000 2000 3000 4000 5000 DAY TSS(mg/l) 400 450 500 20 40 60 80 100 DAY TOC(mg/l) 400 450 500 10 20 30 DAY DissolvedTKN(ug/l) 400 450 500 0 500 1000 1500 2000 DAY ParticulateNitrogen(ug/l) 400 450 500 0 200 400 600 800 1000 DAY ParticulatePhos.(ug/l) 400 450 500 0 100 200 300 DAY SRP(ug/l) 400 450 500 20 40 DAY Chlorophylla(ug/l) 400 450 500 0 5 10 15 20 DAY DO(mg/l) 400 450 500 5 10Synoptic#7 Figure 7-31. Temporal water quality variati ons at Synoptic #7 station (Segment Four)

PAGE 214

191 DAY NO3(ug/l) 400 450 500 0 20 40 60 DAY DissovedSilica(ug/l) 400 450 500 0 1000 2000 3000 4000 5000 DAY TSS(mg/l) 400 450 500 20 40 60 80 100 DAY TOC(mg/l) 400 450 500 10 20 30 DAY DissolvedTKN(ug/l) 400 450 500 0 500 1000 1500 2000 DAY ParticulateNitrogen(ug/l) 400 450 500 0 200 400 600 800 1000 DAY ParticulatePhos.(ug/l) 400 450 500 0 100 200 300 DAY SRP(ug/l) 400 450 500 20 40 DAY Chlorophylla(ug/l) 400 450 500 0 5 10 15 20 DAY DO(mg/l) 400 450 500 5 10Synoptic#13 Figure 7-32. Temporal water quality variati ons at Synoptic #13 stat ion (Segment Four)

PAGE 215

192 DAY DissovedSilica(ug/l) 400 450 500 0 1000 2000 3000 4000 5000 DAY TSS(mg/l) 400 450 500 20 40 60 80 100 DAY TOC(mg/l) 400 450 500 10 20 30 DAY DissolvedTKN(ug/l) 400 450 500 0 500 1000 1500 2000 DAY ParticulateNitrogen(ug/l) 400 450 500 0 200 400 600 800 1000 DAY ParticulatePhos.(ug/l) 400 450 500 0 100 200 300 DAY SRP(ug/l) 400 450 500 20 40 DAY Chlorophylla(ug/l) 400 450 500 0 5 10 15 20 DAY NO3(ug/l) 400 450 500 0 5 10 15 20 DAY DO(mg/l) 400 450 500 5 10Synoptic#26 Figure 7-33. Temporal water quality variati ons at Synoptic #26 stat ion (Segment Two)

PAGE 216

193 DAY DissovedSilica(ug/l) 400 450 500 0 1000 2000 3000 4000 5000 DAY TSS(mg/l) 400 450 500 20 40 60 80 100 DAY TOC(mg/l) 400 450 500 10 20 30 DAY DissolvedTKN(ug/l) 400 450 500 0 500 1000 1500 2000 DAY ParticulateNitrogen(ug/l) 400 450 500 0 200 400 600 800 1000 DAY ParticulatePhos.(ug/l) 400 450 500 0 100 200 300 DAY SRP(ug/l) 400 450 500 20 40 DAY Chlorophylla(ug/l) 400 450 500 0 5 10 15 20 DAY NO3(ug/l) 400 450 500 0 5 10 15 20 DAY DO(mg/l) 400 450 500 5 10Synoptic#28 Figure 7-34. Temporal water quality variati ons at Synoptic #28 stat ion (Segment Two)

PAGE 217

194 7.4.4 The ROC Scores of Water Quality Model Simulation The long-term, synoptic, and episodic si mulated predictions were separately compared to the measured data. Table 713 shows the comparison between combined model predictions and measured data. Ther e are four data cate gories: seven WQMN stations (IRLV17, IRLI07, IRLB09, IRLI 18, IRLI23, IRLIRJ01, and IRLIRJ05); all WQMN stations; combined WQMN and s ynoptic stations; and combined WQMN, synoptic, and episodic stations. As can be s een, the average ROC scores of the seven WQMN stations and all WQMN stations ar e comparable. The comparison of each ROC score of water quality state variables showed insubstantial differences between the seven WQMN stations and all WQMN stations. Th e average ROC score of combined model predictions and measured data is nearly 0.7. This suggests that the water quality model has a skill to reproduce the ecolo gical dynamics in the IRL. Table 7-13. ROC scores of IRL water quality model simulation Name TSS DO Chl. a TOC NO3 DTKN SRP PN PP SOS AVG WQMN (7 stations) 0.607 0.688 0.658 0.722 0.548 0.792 0.858 0.591 0.510 0.681 0.666 WQMN All stations 0.539 0.703 0.662 0.772 0.592 0.792 0.746 0.596 0.491 0.726 0.662 WQMN(ALL) + Synoptic 0.560 0.720 0.640 0.780 0.597 0.800 0.779 0.539 0.797 0.718 0.693 WQMN(ALL) + Synoptic + Episodic 0.586 0.679 0.630 0.778 0.597 0.813 0.761 0.536 0.804 0.758 0.694 7.5 Sensitivity Test for IRL Water Quality Model 7.5.1 Sensitivity Test for Wat er Quality Nutrients Water quality simulation is not only depe ndant on hydrodynamic circulation, model parameters, and nutrient kinetic equations, but also other fact ors, such as nutrients at sediment column, open boundary nutrients, river boundary nutrients, surface light

PAGE 218

195 intensity, and grid resolution. Therefore, se veral sensitivity simula tions were performed in order to test the effects of each factor (Table 7-14). Table 7-14. Description of sensitivity tests Test run Description Test 1 Double water quality nutrient values at sediment layer Test 2 Halve water quality nutrient values at sediment layer Test 3 Double particulate nitrogen va lues and half particulate phosphorous values at sediment layer Test 4 Double water quality nutrient values at river boundary Test 5 Halve water quality nutrient values at river boundary Test 6 Double water quality nutrient values at open boundary Test 7 Halve water quality nutrient values at open boundary Test 8 Double surface light intensity Test 9 Halve surface light intensity Test 10 Use high resolution grid (478 x 44) Test 11 One algal group simulation Test 12 Extend some river boundaries Test 13 Time-varying salinity at inlets First, nutrients at the sediment column play a major role in nutrient dynamics in water column, especially particulate nutrien ts. During the erosion event caused by wind or current, the sediment and particulate nut rients suspend into the water column. These suspended particulate nutrients react to the dissolved nutrients as sorption/desorption and directly increase the concentra tion of particulate nutrients in the water column. Therefore, tests were designed to measure how changed sediment nutri ents affect the study area: doubling the original nutrient va lues (Test 1) and halving th e original nutrient values (Test 2). As can be seen in Table 7-15, the results of Test 1 indicate that most of nutrient ROC scores are lower than the baseline si mulation besides particulate nitrogen. This suggests that model simulation may overly predict nutrients on account of affluent nutrient source from the sediment layer, while particulate nitrogen calculated by the model compare well with measured data. It is possible that some measured particulate nitrogen at the sediment laye r is underestimated owing to lack of data. The range of

PAGE 219

196 measured values is markedly broad from 27 to 4260 mg/kg for PON and from 1.61 to 40.95 mg/kg for PIN (Reddy et al. 2001). Test 2 shows that the overall ROC score is comparable to that of the baseline simulation. Total organic carbon and part iculate phosphorous demonstrate noticeable improvement. It is likely that the baselin e simulation overestimated TOC and PP in certain areas. In the case of PP, eroded pa rticulate phosphorous from sediment layer is considered the main source. Therefore, PP c oncentration in the water column tends to be decreased as PP concentration at the sediment layer dwindles in this test. Analyzing from Test 1 and Test 2, a hypothesis is sugg ested: increasing PN and decreasing PP concentration at the sediment layer may am eliorate model performance. Based on this hypothesis, a sensitivity test is designed to double particulate nitrogen values and halve particulate phosphorous values at the sediment layer (Test 3). From this test, even model simulation compares well with measured data in terms of particulate nutrients. The other nutrient results are markedly worse than the baseline simulation. It is certain that all nutrients in the water quality model are dynamically related to each nutrient. As particulate nitrogen increases, so does the dissolved nitrogen nutri ent, while decreased particulate phosphorous lessens the concentr ation of soluble phosphorous in the water column. These changed nutrients affect not only phytoplankton but also DO and other nutrients. Second, the nutrients from the river boundary are consid ered to be one of the nutrient sources in the system. The stations n ear the river runoff are mainly affected by river nutrients. In order to estimate the influen ce in terms of altered river nutrients, Test 4 and Test 5 are created: increasing nutrient va lues as twice the original river inflow

PAGE 220

197 nutrient values (Test 4) and d ecreasing nutrient values as ha lf the original river inflow nutrient values (Test 5). As re vealed by Table 7-15, the comprehensive average is similar to the baseline simulation for both tests. Sol uble silica, in particular, has varied in association with alternated ri ver nutrients. This variation directly affects the diatom, which is one of the major algal groups, b ecause the nutrient is a primary source for diatom production. This indicate s that soluble silica and diatom in the water column are probably dependent on the river inflow silica. Third, like river inflow nutrients, nu trients from the open boundary are also contemplated to be one of the nutrient sour ces in the study area. The study domain has four inlets connected with the Atlantic Ocean and a dynamic nutrient exchange between IRL and the ocean. Therefore, tw o tests were conducted to anal yze the effects in terms of altered open boundary nutrients: doubling nutrient values from the original open boundary (Test 6) and halving nutrient values from the original ope n boundary (Test 7). As shown in Table 7-15, in the case of reduced open boundary nut rients, the average ROC score is nearly equivalent to the baseline simulation, while in the case of increased open boundary nutrients, the average ROC score is substantially lowe r than the baseline simulation. Phytoplankton, DO, and SRP have noticeably changed in association with alternated open boundary nutrients These results show this area definitely to be affected by open boundary nutrients and likely more sensitive for open boundary nutrients than river inflow nutrients. Fourth, surface light intensity is one of the most important variables controlling phytoplankton photosynthesis, which provides the predominant source of energy for autotrophic organisms (Day et al., 1989). It is valuable to measure how surface light

PAGE 221

198 intensity affects nutrients. Two tests are carried out: doubling surf ace intensity (Test 8) and halving surface light intensity (Test 9). From the Table 7-15, the surface light intensity is slightly more influential to nitrogen nutrients th an phosphorous nutrients. Even though surface light intensity is re duced by half, the amount of phytoplankton photosynthesis is likely not decreased because of the relatively high strength of surface light intensity at IRL, which has a true tropical climate. Fifth, water quality model performance seem s to be affected by grid configuration, bathymetry, and resolved navigation channel. In order to examine how the horizontal grid resolution causes simulated water quality transport in this study area, Test 10 was performed using the fine grid (478 x 44), wh ich has a resolution more than four times finer the coarse grid (199 x 23) for the baseline simulation. As can be seen from the results in Table 7-15, the ROC scores of so me nutrients show differences between the coarse grid and fine grid si mulation. This indicates that the horizontal grid resolution induces the change of hydrodynamic circulatio n and water quality transport. One of interesting results from this test is that the average ROC score for the fine grid is slightly lower than that for the coarse grid. It is possible that initial and boundary water quality data are not sufficient for covering the fine grid, and calibra ted water quality parameters are more suitable for the coarse grid. One of the main objectives for this study is to develop the multi-species water quality model. Test 11 was conducted to rev eal why the multi-species model would be better than the one-species model. In Test 11, the same model parameters as with the multi-phytoplankton model were used but without phytoplankton parameters. Phytoplankton parameters in association w ith the one-species model are obtained by

PAGE 222

199 averaged multi-species values. As revealed by Table 7-15, the aver age ROC score of the one-species model is lower than that of the baseline simulation (multi-species model). The salient difference of the tw o models relates to Chlorophyll a. The Chlorophyll a ROC score of the multi-species model is much higher than that of the one-species model. This suggests that the multi-species model be tter performs than the one-species model. The graphical summary of these tests are provided by Taylor diagrams in Appendix J. Table 7-15. The results of sensitivity test s. Values shown are the ROC score of each nutrient. Values shown in italics indi cate the difference between the baseline and test simulations with a positive value indicating improvement of test simulation and a negative value indicati ng deterioration of test simulation. Test run DO Chl. a TOC NO3 DTKN SRP PN PP SOS AVG Baseline 0.687 0.655 0.722 0.566 0.794 0.863 0.591 0.497 0.693 0.674 Test 1 0.693 0.006 0.577 -0.078 0.696 -0.026 0.545 -0.020 0.651 -0.143 0.782 -0.081 0.605 0.014 0.442 -0.055 0.707 0.014 0.633 -0.059 Test 2 0.698 0.011 0.597 -0.058 0.756 0.034 0.586 0.020 0.770 -0.024 0.839 -0.023 0.547 -0.043 0.539 0.042 0.685 -0.008 0.669 -0.005 Test 3 0.661 -0.026 0.599 -0.056 0.702 -0.019 0.325 -0.240 0.630 -0.164 0.787 -0.076 0.605 0.014 0.530 0.033 0.610 -0.083 0.606 -0.068 Test 4 0.683 -0.004 0.641 -0.013 0.672 -0.05 0.582 0.016 0.745 -0.049 0.870 0.008 0.591 0.0 0.504 0.007 0.663 -0.030 0.661 -0.013 Test 5 0.695 0.008 0.629 -0.025 0.705 -0.016 0.558 -0.008 0.808 0.014 0.870 0.007 0.603 0.013 0.503 0.006 0.683 -0.009 0.672 -0.002 Test 6 0.580 -0.107 0.598 -0.056 0.715 -0.007 0.648 0.082 0.682 -0.112 0.735 -0.128 0.595 0.004 0.498 0.001 0.702 0.010 0.639 -0.035 Test 7 0.771 0.084 0.666 0.011 0.714 -0.008 0.528 -0.038 0.754 -0.040 0.859 -0.004 0.587 -0.004 0.498 0.001 0.694 0.001 0.674 0.0 Test 8 0.691 0.004 0.640 -0.015 0.725 0.003 0.532 -0.034 0.797 0.003 0.848 -0.015 0.607 0.016 0.489 -0.008 0.694 0.001 0.669 -0.005 Test 9 0.702 0.015 0.659 0.005 0.719 -0.003 0.661 0.095 0.794 0.001 0.858 -0.005 0.607 0.016 0.499 0.002 0.695 0.002 0.688 0.014 Test 10 0.734 0.047 0.557 -0.098 0.676 -0.046 0.613 0.047 0.793 -0.001 0.841 -0.022 0.581 -0.009 0.545 0.048 0.655 -0.038 0.666 -0.008 Test 11 0.695 0.008 0.589 -0.066 0.709 -0.013 0.534 -0.032 0.807 0.014 0.849 -0.014 0.611 0.020 0.493 -0.004 N/A 0.661 -0.013 Test 12 0.690 0.003 0.658 0.003 0.721 -0.001 0.566 0.0 0.793 -0.001 0.866 0.003 0.591 0.0 0.492 -0.005 0.679 -0.014 0.673 -0.001 Test 13 0.692 0.005 0.655 0.0 0.723 0.001 0.566 0.0 0.796 0.002 0.864 0.001 0.591 0.0 0.497 0.0 0.688 -0.005 0.674 0.0 In the computation domain, some main rive rs did not have ade quate resolution, and thus the grid area was expanded to incorporat e more of the river. Test 12 was conducted to analyze how the extended river boundary a ffects the water quality model results. As

PAGE 223

200 can be seen in Table 7-15, the difference of the average ROC was insubstantial. Dissolved silica was influenced somewhat by extending some of river boundaries. The salinity values at inlets were obtained from the model simulation with a large domain, which includes IRL and further offshore regi on. Test 13 was create d to estimate how the time-varying salinity at inlets influences the nutrient dynamics in the study area. From Table 7-15, most of ROC scores of this test are nearly equivalent to that of the baseline simulation. This suggests that the circulati on and transport alternations due to timevarying salinity at inlets are not stro ng enough to change the nutrient dynamics. 7.5.2 Sensitivity Test for Sediment Sediment transport is quite significant for nutrient cycling in the water column because sediment contains a large amount of nutrients. It is common that nutrient concentration in the bottom se diments is one or two orders of magnitude higher than the concentration of the water column (Simon, 1988 and 1989). Nutrients in particulate form suspended with sediment change into a di ssolved form due to sorption/desorption. Therefore, the resuspension/deposition of sediment from the bottom not only increases/decreases the nutrient concentr ation in the water column, but may also accelerate/decelerate the nutrient cycling. In this study, sediments are divided into two groups: fine group (cohesive) and coarse group (non-cohesive), assuming no in teraction between the two groups. Based on the grab samples, sediments are classified into 5 categories: silt, fine sand, medium sand, coarse sand, and very coarse sand. Sediment transport is significantly influenced by the following factors: settling velocity, river or open boundary sediment concentration, critical shear stress, and init ial fine/coarse group ratio in as sociation with the sediment

PAGE 224

201 category. Fourteen sensitivity tests were designed to measure how the sediment-related factors influence water quality nutrients and sediment itself (Table 7-16). Table 7-16. Description of sediment sensitivity tests Test runs Description Test 1 Double fine group settling velocity Test 2 Half fine group settling velocity Test 3 Double coarse group settling velocity Test 4 Half coarse group settling velocity Test 5 Double sediment values at river boundary Test 6 Half sediment values at river boundary Test 7 Double sediment values at open boundary Test 8 Half sediment values at open boundary Test 9 Increase by 25 % critical shear stress Test 10 Decrease by 25 % critical shear stress Test 11 Increase initial fine/coars e group ratio for mud and silt/mud Test 12 Decrease initial fine/coars e group ratio for mud and silt/mud Test 13 Increase initial fine/coarse group ratio for sand and silt/sand Test 14 Decrease initial fine/coarse group ratio for sand and silt/sand Settling is a process by which sediment pa rticles fall through th e water column due to density differences. The settling velocity relates to the representative size of suspended sediments and hydrodynamic conditions. Four te sts are conducted in association with settling velocity: doubling fine group settling ve locity (Test1), halving fine group settling velocity (Test 2), doubling coarse group settli ng velocity (Test 3), and halving coarse group settling velocity (Test 4). As revealed by Table 7-17, the water quality nutrients are more affected by fine group settling velocity than by coarse group settling velocity. The interesting results are that TSS ROC score decreases as coarse group settling velocity decreases, and as fine group settling veloci ty increases. This suggests that coarse sediment of TSS is likely affected by its settling velocity, while fine sediment of TSS is probably dependent on other fact ors, such as erosion, critical shear stress, and other hydrodynamic conditions. Sediment from river and open boundaries ar e considered to be one of the main sources in the system. Four tests are carried out to analyze the effect s in terms of altered

PAGE 225

202 river and open boundary sedime nts: doubling sediment values from the original river boundary (Test 5), halving sediment values from the original river boundary (Test 6), doubling sediment values from the original open boundary (Test 7), and halving sediment values from the original open boundary (Test 8). As shown in Table 7-17, the average ROC scores of those tests are nearly equivalent to that of the baseline simulation. The changes of sediment concentration at rive r and open boundaries ar e not substantially attributed to the nutrients because sediment s indirectly affect the nutrient dynamics. Critical shear stress is one of the most important parameters which govern sediment erosion. The greater the value of critical shear stress the model uses, the less bottom sediment will be eroded. Two tests are conduc ted to analyze how critical shear stress induces dynamic alternations in terms of nutrients and sediment: increasing by 25 % of original critical shear stress (Test 8) and decreasing by 25 % of original critical shear stress (Test 9). As can be seen in Table 7-17, increased critical stress prevents sediment as well as particulate nutrients from suspe nding. From the partic ulate phosphorous pointof-view, less-eroded particulate phosphorous co ntributes better agr eement with measured data which leads to a consistency with the sensitivity test (Test 3) for water quality nutrients. The TSS is more sensitive for increasing critical shear stress than for decreasing critical shear stress. Based on the sediment grab samples, the in itial fine/coarse sedi ment ratio of each sediment group was decided. Even though the fine/coarse sediment ratio at IRL is dynamically changed by horizontal/vertical sedi ment exchange, sediment inflow/outflow, and other forces, initial fine/coarse sediment ratio is highly consider ed to influence the sediment dynamics. Four tests are carried out for measuring the effects in terms of altered

PAGE 226

203 initial fine/coarse sediment ratio: increasing initial fine/c oarse group ratio for mud and silt/mud (Test 10), decreasing initial fine/c oarse group ratio for mud and silt/mud (Test 11), increasing initial fine/c oarse group ratio for sand a nd silt/sand (Test 12), and decreasing initial fine/coarse group ratio for sa nd and silt/sand (Tes t 13). As revealed by Table 7-17, the initial fine /coarse group ratio does not substantially induce dynamic alterations of sediment and water quality species. However, sediment and nutrients have a lot to do with the change of the initial fine/coarse group rati o than that of sediment at open and river boundaries. Table 7-17. The results of sensitivity test s. Values shown are the ROC score of each nutrient. Values shown in italics indi cate the difference between the baseline and test simulations with a positive value indicating improvement of test simulation and a negative value indicati ng deterioration of test simulation. Test runs TSS DO Chl. a TOC NO3 DTKN SRP PN PP SOS AVG Baseline 0.607 0.687 0.655 0.722 0.566 0. 794 0.863 0.591 0.497 0.693 0.667 Test 1 0.594 -0.013 0.600 -0.087 0.517 -0.138 0.651 -0.070 0.433 -0.133 0.797 0.004 0.849 -0.014 0.495 -0.096 0.548 0.051 0.689 -0.004 0.617 -0.05 Test 2 0.621 0.013 0.705 0.018 0.600 -0.055 0.684 -0.038 0.569 0.004 0.719 -0.074 0.843 -0.020 0.593 0.003 0.406 -0.091 0.715 0.022 0.646 -0.021 Test 3 0.610 0.002 0.696 0.009 0.601 -0.054 0.729 0.007 0.569 0.004 0.764 -0.030 0.835 -0.028 0.601 0.010 0.456 -0.041 0.722 0.029 0.658 -0.009 Test 4 0.590 -0.017 0.682 -0.005 0.568 -0.087 0.756 0.034 0.513 -0.052 0.784 -0.009 0.846 -0.017 0.550 -0.041 0.534 0.037 0.690 -0.003 0.651 -0.016 Test 5 0.615 0.008 0.692 0.005 0.646 -0.009 0.725 0.003 0.566 0.0 0.797 0.004 0.869 0.006 0.591 0.0 0.488 -0.009 0.700 0.007 0.669 0.002 Test 6 0.603 -0.004 0.687 0.0 0.645 -0.010 0.720 -0.002 0.561 -0.005 0.799 0.006 0.869 0.006 0.607 0.016 0.502 0.005 0.698 0.005 0.669 0.002 Test 7 0.601 -0.006 0.716 0.029 0.655 0.0 0.709 -0.013 0.570 0.004 0.802 0.008 0.864 0.001 0.593 0.003 0.495 -0.002 0.698 0.005 0.670 0.003 Test 8 0.600 -0.007 0.669 -0.018 0.615 -0.039 0.747 0.026 0.562 -0.004 0.784 -0.010 0.853 -0.009 0.594 0.004 0.499 0.002 0.695 0.002 0.662 -0.005 Test 9 0.592 -0.015 0.677 -0.010 0.588 -0.067 0.715 -0.007 0.506 -0.059 0.794 0.0 0.861 -0.002 0.573 -0.018 0.518 0.021 0.700 0.008 0.653 -0.014 Test 10 0.604 -0.003 0.704 0.018 0.604 -0.051 0.710 -0.012 0.569 0.004 0.796 0.002 0.847 -0.016 0.561 -0.030 0.441 -0.056 0.710 0.017 0.655 -0.012 Test 11 0.595 -0.013 0.678 -0.009 0.611 -0.044 0.749 0.028 0.557 -0.009 0.797 0.003 0.855 -0.008 0.611 0.02 0.474 -0.023 0.711 0.018 0.663 -0.004 Test 12 0.618 0.011 0.704 0.017 0.624 -0.031 0.703 -0.019 0.518 -0.048 0.754 -0.039 0.860 -0.003 0.584 -0.007 0.514 0.017 0.694 0.001 0.657 -0.01 Test 13 0.622 0.015 0.755 0.068 0.645 -0.01 0.668 -0.054 0.546 -0.02 0.775 -0.019 0.836 -0.027 0.589 -0.002 0.469 -0.028 0.682 -0.011 0.659 -0.008 Test 14 0.601 -0.006 0.669 -0.018 0.611 -0.044 0.761 0.039 0.570 0.004 0.788 -0.006 0.827 -0.036 0.585 -0.006 0.528 0.031 0.706 0.013 0.664 -0.003

PAGE 227

204 CHAPTER 8 CONCLUSION AND DISCUSSION CH3D-IMS has been enhanced and applied to the Upper Charlotte Harbor and Indian River Lagoon. The circulation and tran sport in the Upper Charlotte Harbor have been sim ulated using CH3D-IMS and field data from June 12th, 2003 to July 10th, 2004. This simulation focused on the temporal and spatial salinity dynami cs in response to time-varying freshwater flows. Due to the lack of measur ed nutrient data in 2003 2004, water quality simulation was conducted duri ng 2000 instead of 2003 2004. However, the forcing functions and boundary conditions fo r the circulation and transport simulation are nearly equal to those of the 2003 2004 si mulation. This water quality simulation has included a more robust SOD model and a mu lti-algal group model to better simulate bottom water hypoxia and nutrient dynamics in the Upper Charlotte Harbor. For the Indian River Lagoon, a long-term (1997 1999) water quality model was conducted using CH3D-IMS and model predicti ons compared with field data. This simulation reproduced observed seasonal nut rient variation and phytoplankton succession in this area. In addition, the 3-D water quality model was employed to simulate the episodic event data collected in February and November 1999. Based on the simulations, it is apparent that the short-term effect of sorption/desorption as well as particulate erosion processes are significant during episodi c events. Major conclusions of this study are as follows: (1) A fine-resolution numerical grid that represents the complex geometrical and bathymetric features in the Upper Char lotte Harbor was employed to simulate

PAGE 228

205 hydrodynamics, salinity and temperature characteristics during 2003 2004. The normalized RMS errors between measured and model predictions were less than 7.5 % at three stations: UF, Punta Gorda, and El Jobean Simulated currents at the UF station have less than 15 % RMS errors normalized by the range of currents at three levels. For salinity, the RMS errors at the surface layer are 2.51, 3.08, and 2.42 PSU at the UF, Punta Gorda, and El Jobean stations during the low-runoff period, respectively. (2) Large amount of river flow from the Myakka River brings fresh water to the middle of Upper Charlotte Harbor during the high runoff period. This causes the values of surface and bottom salinity to be less than 1 PSU until approximately 13 km downstream from the mouth of the Myakka River, and form s the salt-wedge between 15 and 25 km downstream from the mouth of the My akka River. In cont rast, salinity values near the mouth of the Myakka River are a bout 9 to 10 PSU, during the low-runoff period, with strong vertical mixing and salinity grad ient along the Myakka River. There are large differences between salinity during th e high-runoff and low-runoff periods. The differences are 1 to 20 PSU at 10 km downs tream from the mouth of the Myakka River. (3) There is a channel from the middle of the Upper Charlotte Harbor to the mouth of the Peace River. The channe l allows transport of salty water to near the mouth of the Peace River. For example, in the low runoff period, salinity near the mouth of the Peace River fluctuates between 14 and 19 PSU due to the presence of the channel. During the high runoff period, fresh water marches downstream to nearly 8 km from the mouth of the Peace River and forms a salt-wedge. Large differences are formed between salinity during the high and low runoff periods along the Peace River.

PAGE 229

206 (4) The dominant seagrass, Halodule wrightii, in the Upper Charlotte Harbor is very sparse along the Peace River. The la rge amount of freshwater may reduce the growth of Halodule wrightii because this seagrass has a low tolerance for low salinity and its growth rate quickly decreases at salinity < 12 PSU. Mortality begi ns to occur when the salinity goes below 6 PSU. The model simulation showed that, to maintain the salinity values above 12 PSU at 5km below the m outh of the Peace River, 30-day constant freshwater flow rate at the P eace River must stay below 70.2 m3/s. (5) Model simulation reproduced the summe rtime low DO values at the bottom layer at CH005 and CH006 stat ions, which are located near the Peace River. The model results showed that SOD values in the summ er are higher than other seasons, and their values near the rivers are hi gher than other areas. These spat ial and temporal variations are affected by variations in nutrient loads, temperature, di ssolved oxygen, sediment type, CBOD, and particulate settling velocity. W ith 50 % and 100 % nutrien t load reduction at the Peace and Myakka Rivers, model results sh owed that bottom DO slightly increases due to diminishing SOD, which is directly re lated to particulate organic matter. However, hypoxia condition at the bottom layer still re mains during the high-runoff period. This suggests that bottom-water hypoxia is strongly related to vertical stratification caused by high river flow from the Peace River. Vertical stratification weakened water mixing thus preventing surface water body from supplying DO to the bottom water, while DO continued to be consumed in the bottom wa ter. These successions created low bottom DO events in the Upper Charlotte Harbor, wh ile at stations far from Myakka and Peace Rivers high bottom DO values were measured during the summer.

PAGE 230

207 (6) Model predictions of water quality species have substantially agreed well with measured data in the Indian River La goon during 1997 1999. The skill of the water quality model was quantified by comparing m odel results with field data using a new skill assessment method: relative operating characteristic (ROC) test. Most of the ROC scores of water quality state variables are over 0.6, and the maximum score is 0.862. The absolute relative errors for di ssolved species (except NO3 a nd SRP) are less than 43% at all stations. The particulate species shows larger absolute relative errors than the dissolved species. The reason is due to difficu lty in accurately simulating the erosion of sediment and particulate nutrien ts caused by wave and currents. (7) Seasonal variations of phytoplan kton and dissolved oxygen were obvious during the simulation period. The phytoplankton values were high in the summer and low in the winter. It is believed that light and temperature are among main factors for phytoplankton growth. Therefore, deficient light and temperat ure in the winter period could prevent phytoplankton from thriving. On the other hand, DO became high in the winter, and declined until summer. After summer, the values increased until winter. From the point of view of spatial distributions, SRP was higher in the southern than in the northern and central IRL proper, while the values of DTKN indicated opposite aspects. This suggests that phosphorous and nitrogen may be the limiting nutrients in the northern/central and southern IRL proper, respectively. (8) The abundance and relative domin ance of the algal species underwent continuous change in the IRL. Diatom was mostly dominant but its biomass declined in the summer because of weakened temper ature tolerance. On the other hand, dinoflagellates, which have a high temperat ure tolerance, became the dominant algal

PAGE 231

208 species in the summer period. In addition, there were some cyanobacteria blooms in the Banana River. These algal successions were reproduced by the water quality model. (9) Model successfully simulated episodic events in the IRL during which the erosion of sediments and particulate nutrie nts were caused by wave and current induced bottom shear stress. The spatial distribution of calculated bottom shear stress indicated sediments and particulate nutrients in the Mosquito Lagoon, Banana River, and Northern IRL Proper are mostly affected by the wind waves because these areas are far from the inlets. Thus, the current-induced shear stress is relatively weaker than in other areas. In the central and southern IRL Proper, particulate species are eroded by the current and wave bottom shear stress. (10) From the IRL model sensitivity tests for water quality nutrients, it is clear that all nutrients in the water quality model ar e dynamically related to each other. Soluble silica and diatom in the water column are pr obably dependent on the silica in the river inflow. Even though surface light intensity is reduced by half, the amount of phytoplankton photosynthesis is likely not de creased because of the relatively high strength of surface light intensity at IRL, wh ich has a true tropical climate. Due to the insufficiency of initial and boundary water quality data, the fine grid system was not as precise as the coarse grid system. Also, th e calibrated water quality parameters were more suitable for the coarse grid system. One of the salient consequences from sensitivity tests was that the multi-species model better performed than the one-species model. The ROC score of Chlorophyll a in the multi-species is 10 % higher than that of the onespecies model.

PAGE 232

209 (11) The study introduced a modified Gauss-Newton method to estimate water quality model parameters, as reported in a ppendix K. The method was applied to kinetic equations of water quality state variables. To estimate water quality parameters, the method employed IRL episodic data. Two IRL episodic simulations were conducted: A simulation with water quality parameters calculated by the modified Gauss-Newton method and another simulation with water qua lity parameters obtained by the trial-anderror method. The average relative error and ROC score from model results obtained with parameters calculated by the parameter estimation method are 58.3 % and 0.565, respectively. Using parameters obtained by trial-and-error process, the average relative error and ROC scores are 60.5% and 0.555, respectively. The results of two skill assessments indicate that the model predictions with parameters from the modified Gauss-Newton method performed better than those with current IRL simulation parameters obtained by trial-and-error proces s. The summarized main conclusions can be seen in Table 8-1. Even though the robust ecological model CH3D-IMS has been enhanced and applied to a couple of Florida estuaries, the model still requires further development. The model contains various assumptions and simplif ications due to the insufficient field and laboratory data and inadequate unders tanding of complex nutrient dynamics. Recommendations for future developments are given: (a) The sediment layer has rich nutrient s which influence nutrient dynamics in the water column, especially during strong wind events. The model did not sufficiently capture particulate nutrient dynamics due to th e lack of spatial va riation of sediment

PAGE 233

210 nutrient data. More detailed sediment nutrient data with fine horizontal resolution are needed to improve the model boundary conditions and performances. Table 8-1. Main conclusions from this study Study area Main conclusions Upper Charlotte Harbor 1. CH3D-IMS was calibrated and valid ated with field data measuring water level, currents, salinity, te mperature, and water quality state variables in the Upper Charlotte Harbor. 2. The river flows from the Myakka and Peace Rivers play a major role in the salinity variation and vertical mixing. 3. 30-day constant river flow with below 70.2 m3/s from the Peace River may sustain a seagrass (Halodule wrightii) growth rate. 4. Hypoxia phenomena at the bottom water are related to freshwater flows and SOD. According to model simulations, reducing 50 % or 100 % of nutrient loads at the river bound aries did not eliminate the low DO condition. Indian River Lagoon 1. CH3D-IMS reproduced the seas onal dynamics of nutrients and phytoplankton succession in IRL. 2. The erosion of particulate nutr ients and sediment is significantly related to wind wave in Mosquito Lagoon, Banana River, and Northern IRL Proper, while in the central and southern IRL Proper, due to currents and waves. 3. The multi-species model produced better results than the one-species model (ROC score of Chlorophyll a in the multi-species is 10 % higher). 4. In the IRL short-term simulation, the model predictions with water quality parameters calculated by the modified Gauss-Newton method better performed than those with wa ter quality parameters obtained by the trial-and-error process. (b) Even though CH3D-IMS includes multip le phytoplankton species, there is insufficient information about multi-algal sp ecies at the river and open boundaries. In addition, constant values are used for open boundary water quality data during the simulation period. These factors cause increas ing uncertainty in the water quality model simulation. To improve model simulation, te mporal water quality data and multi-algal species data should be collected at the river and open boundaries. (c) The interaction between the water co lumn and the sediment column is an important feature in estuaries. The excha nge of dissolved nutrients is primarily a

PAGE 234

211 chemical and biological process involving methane, sulfide, organic matter, nutrient, and other metals albeit also affected by physical diffusion in the water and sediment columns. More data is needed to further elucidate th e exchange of dissolve d species between the sediment and water columns.

PAGE 235

212 APPENDIX A A THREE-DIMENSIONAL CURVILINEAR-GRID HYDRODYNAMIC MODEL Sheng (1983) used -transfor mation to simulate circul ation and transport in water bodies with gradual bathymetric variation. ),,(),( ),,(tyxyxh tyxz where h is water depth and is the surface elevation. The go verning equations are transformed into -stretched coordinate system. Whereas w=dz/dt in the z-plane, =d /dt in the -plane. In order to simplify the model equations under cert ain circumstances, the dimensionless equations are adopted. These equations allow one to comp are the relative importance of various terms in the equations. The dimensionless vari ables and parameters are written as )/,/,/,/,/(),,,,( )/,/,/,/,/(),,,,( )/,/,/,/,/,/(),,,,,( ** ),() *** *** (),( ,* ) (),( )/,/(),( )/,/,/(),, (22 0 2 2 0 0 0 ** 0 0 ** ** ***r HrHrVrVrHrHrVrVr rHrVCHCV rHrrVr rr rr HVrdr VrVHrHVrVHrHVrVHrH VHVHVH rrr r w y r w x rr y rr w x yx r r r rr r rr rr rFRKAKADADAPPSS fXAfZAFgZUrfXUEEFFR KKKKDDDDAAAAKKDDAA SZUf g UZfUZf ftt U X UZ Xw w UvUuvu ZzXyXxzyx

PAGE 236

213 where Xr and Zr are the referen ce lengths in the vertical and hor izontal directions; Ur is the reference velocity; 0, r and = 0 are the mean density, refe rence density, and density gradient in a stratified flow; AHr and AVr are the reference eddy viscosities in the horizontal and vertical directions; DHr and DVr are the reference eddy diffusivities in the horizonta l and vertical directions for salinity; KHr and KVr are the reference eddy diffusi vities in the horizontal and vertical directions for temperature; R0 is Rossby Number; Fr is Froude Number; Frd is Densimetric Froude Number; EV and EH are Vertical and Lateral Ekman Number; SCV and SCH are Vertical and Lateral Schmidt Number; and PrV and PrH are Vertical and Lateral Prandtl Number. The governing equations transformed into the -stretched coordinate are nondimensionalized using the referenc es scales system. These equati ons once again are transformed from original coordinates (x,y) to curvilinear coordinates ( ) to better represent complex shoreline geometrics. The final continuity momentum, and transport equations are 00 0 0 H Hvg Hug g t u A H E d H g H gdggH F R Hu gHvvgxHuvgx HvvgxHuugxy HvvgyHuvgy HuvgyHuugyx Hg R uofTerms Diffusion Horizontal AE g vg g ug gg t Hu HV V r HH2 0 0 12 11 12 11 2 0 0 0 0 0 0 0 0 0 0 0 0 0 22 0 12 12 11) () ( ) ( ) ( 1

PAGE 237

214 v A H E d H g H gdggH F R Hv gHvvgxHuvgx HuvgxHuugxy HvvgyHuvgy HuvgyHuugyx Hg R vofTerms Diffusion Horizontal AE g vg g ug gg t Hv HV V r HH2 0 0 22 21 22 21 2 0 0 0 0 0 0 0 0 0 0 0 0 0 21 0 11 22 21) () ( ) ( ) ( 1 S Hgg S Hgg S Hgg S Hgg gS E HvSg HuSg g R SH R S D HS E t HSCH H V CV V22 0 21 0 12 0 11 0 0 0 0 0 0 0)()( T Hgg T Hgg T Hgg T Hgg gP E HvTg HuTg g R TH R T K HP E t HTrH H V rV V22 0 21 0 12 0 11 0 0 0 0 0 0 0)()( where 0g is the Jacobian (J). The determinant of the metric tensor, gij, is defined as 2221 1211 22 22gg gg yxyyxx yyxxyx gij whose inverse is 2221 1212 22 22 2) ( ) ( 1 gg gg yx yyxx yyxx yx J gij More detail information and terms are found in Choi (1992) and Davis (2001).

PAGE 238

215 APPENDIX B SEDIMENT MODEL B.1 Vertical and Lateral Boundary Conditions Vertical boundary conditions play a major role in solving the sediment transport equation at the sediment-water interface. No net sediment flux is considered at surface boundary. The erosion and deposition rate is expected net sediment flux at sedimentwater interface. 0 surface i visiz C DCw ii bottonm i visiED z C DCw where wsi is settling velocity for group i, Ci is suspended sediment concentration for group i, DV are the vertical diffusivity coefficients, Ei and Di are erosion and deposition rates for group i respectively and i=(1,2) with 1 for the fine group and 2 for the coarse group. There is no sediment flux at the solid boundary. At a river mouth and the ocean boundary, the observed sediment concentration values are used. B.2 Settling Velocity There are two settling velocities: fine group (cohesive) and coarse group (sandy particles). In case of fine group, three methods are suggested (Sheng and Chen 1992). A constant settling velocity from either laboratory or field data Stokess formula for a single particle

PAGE 239

216 v D gws ws s2*)(** 18 1 where ws is settling velocity, g is gravity, s is sediment density,w is water density, Ds is the floc size in suspension, and is kinematic viscosity coefficient for water. An empirical settling velocity formula which considers flocculation and hindered settling (Wolanskiet al., 1989) m n sbc ca w )( *22 where a, b, m, and n are empirical constants from laboratory experiments; and c is sediment concentration. When c is below 100 mg/l, ws becomes constant and can be described by Stokes law (free settling). As sediment concentration increases, so does settling velocity. However, ws starts to decrease when c is over 3000 mg/l due to hindered settling. For the coarse group, two methods are used to calculate settling velocity. A constant settling velocity from either laboratory or field data An empirical formula (van Rijn, 1994), if the representative size of the sediments in suspension is known. 1) **)1(*01.0 1( *105.0 2 3 s s sDgs D w where Ds is the representative particle size fo r sandy particles in suspension, and s is specific density. B.3 Erosion Rate

PAGE 240

217 Erosion rate depends on bottom sediment type. Two kinds of bed type are considered: cohesive sediment bed and non-c ohesive sediment bed. Erosion rate of cohesive sediment bed is calculated by th e power law or the exponential law (Sheng and Chen, 1992). The power law p c b c b dT E E |1|* 2 1 )1(* 2 12 0 where E0 is the erosion rate constant, Td is a dimensionless erosion time constant, b is the bottom shear stress, c is the critical shear stress, and p is an empirical constant which can be determined from laboratory or field experiments. The exponential law 5.0 2 0|1|* 2 1 )1(* 2 1 *expc b c b dT E E where is an empirical constant. In the case of erosion rate of non-cohesive sediment, the equation developed by van Rijn 1994) is used. 3 3/1 2 50 5.1 0 50/*)1(* **015.0** gsD T z D wEs where is a calibration parameter, D50 is the median size of the bottom sediments, T is a dimensionless bed-shear parameter, and z0 is bottom roughness. B.4 Deposition Rate Deposition rate also divided two groups: fine group and coarse group (Chen and Sheng, 1995)

PAGE 241

218 Fine group fine dc ds dhC p D )()()(1 1 1 where p is the probability function; vdh, vds, and vdc are the deposition velocity within the logarithmic layer, the vi scous sub-layer, and the ca nopy layer, respectively; and Cfine is the fine sediment concentration. Coarse group coarse sCwD where ws is settling velocity of coarse sediment, and Ccoarse is the coarse sediment concentration.

PAGE 242

219 APPENDIX C DISSOLVED OXYGEN SATU RATION AND REAE RATION Dissolved oxygen saturation in seawater is determined as a function of temperature and salinity (APHA, 1985). 2 3 2 4 11 3 10 2 7 5108673.3109428.1 101929.3 80655.1 10621949.8 102438.110642308.610575701.1 34411.139 ln T T S T T T T DOs where DOs is the equilibrium oxygen concentration at standard pressure (mg/l), T is temperature (K, K=C+273. 15), and S is salinity. The reaeration process is modeled as the flux of dissolved oxygen across the water surface. ) (*) (** DODO z K DODO V A K dt dDO Vs AE s s AE where V is volume of the water body, KAE is the reaeration coefficient (m/day), As is surface area of the water body, and z is volume divided by surface area ( z=V/As). Many empirical formulas have been suggested for estimating reaeration coefficient (OConnor and Dobbins, 1958; Banks and Herrera, 1977; Bowie et al., 1985). Since estuary gas transfer can be affected by both water and wind velocity, elements of current and wind should be considered fo r estimating reaeration rate. Thomann and Fitzpatrick (1982) suggested a formula incl uding current and wind effects. Park (2004) implemented this formula into CH3D-IMS.

PAGE 243

220 2 5.0 00372.0317.0728.093.3w w w AEU U U H U K Where U0 is depth averaged velocity (m/s), H is a depth (m), and Uw is wind speed (m/s).

PAGE 244

221 APPENDIX D A NAIVE STREETER-PHELPS SOD M ODEL For CBOD the equation would be CBOD VkBODVk dt dCBOD Vsed d ws sed** **2 where Vsed is volume of the sediment column, ks is settling removal rate (1/day), Vw is volume of the water column, and kd2 is decomposition rate in the sediments (1/day). At steady-state this balance can be solved for BOD Hk Hk CBODsedd ws2 where Hw and Hsed are layer thickness (m) of th e water and sediment column, respectively. A similar balance for NBOD can be written as CBOD VkBODVkra dt dNBOD Vsedn wsonno sed** ****2 where ano is stoichiometric yield of nitrogen from the decom position of settling BOD, ron is oxygen demand to nitrogen ratio due to nitrification (4.57 gBOD/gN), and kn2 is nitrification rate in the sediments (1/d ay). At steady-state, the solution is BOD Hk Hk NBODsedn ws2*3.0 According to simple model at steady-state, the settling particulate CBOD carries an additional 30% of oxygen demand due to nitr ification (Chapra, 1997). The oxygen deficit balance is as follows:

PAGE 245

222 s B sedn sed d sed sedASNBOD VkCBOD Vk dt dDO V ** **20, 2 2 where As is area of the sediment water interface (m2). At steady-state this balance can be solved for pw s ws BCBODv BODHk S **3.1***3.120, where SB,20 is SOD rate at 20 C, vs is settling velocity of th e particulate CBOD (m/day), and CBODpw is particulate CBOD concentration in the water (mg/L). SOD rate at 20 C should be adjusted because reduction occu rs when methane is consumed in the denitrification process (Chapra, 1997). pw s BCBODv S **11.120, Table D-1 shows a matrix that applies this result with a rang e of organic settling velocities and particulate CBOD that are typi cally encountered in natural waters. From this table, SOD in natural waters sh ould range between about 0.05 to 65 g m-2d-1. Table D-1. SOD in g m-2d-1 calculated with the "naive" Streeter-Phelps SOD model (Chapra, 1997) Settling velocity (m/day) Particulate CBOD (mg/L) 0.1 0.2 0.3 0.5 0.065 0.13 0.325 1 0.13 0.26 0.65 5 0.65 1.3 3.25 10 1.3 2.6 6.5 50 6.5 13 32.5 100 13 26 65

PAGE 246

223 APPENDIX E NUTRIENT DYNAMICS E.1 Nitrogen Cycle Nitrogen can be classified into two gr oups: dissolved nitrogen and particular nitrogen. The nitrogen cycle of CH3D-IM S includes the following processes: Ammonification of organic nitrogen Nitrification of ammonium Volatilization of ammonia Denitrification of nitrate Uptake of ammonia and nitrat e by multi-species phytoplankton Conversion of multi-species phytoplankton nitrogen to zooplankton nitrogen by gazing Excretion and mortality by multi-sp ecies phytoplankton and zooplankton Settling for particulate phosphorous Sorption/desorption reactions The mass balance equation for nitrogen state variables is written by combining nitrogen transformation processes. Ammonia nitrogen (NH3) includes ammonia conversi on and the volatilization process. For water column: atm v VOL al alNHNHhKNH pHH pH K t NH3 3 4 3** For sediment column:

PAGE 247

224 4 3* NH pHH pH K t NHal al where Kal is the ammonia conversion rate consta nt (1/day), which is a temperature function; Hal is the half saturation consta nt for ammonia conversion (mg O2); KVOL is the volatilization constant (1/day); and hv is Henry s constant (mg/L-atm). Dissolved ammonium nitrogen (NH4) includes multi-species phytoplankton uptake and respiration, zooplankton respirati on, ammonification, n itrification, ammonia conversion, and sorption/desorption reaction. For water column: )**(* * **4 4 4 4NHcPPINdNH pHH pH KNH DOH DO K SONKZOOCKPHYC KPA t NHan an al al nit NN ONM zxi axi aini NC For sediment column: )**(* * *4 4 4 4NHcPPINdNH pHH pH K NH DOH DO KSONK t NHan an al al nit NN ONM where ANC is the nitrogen-to-carbon ratio of phytoplankton, i represents multi-species phytoplankton; PHYC is multi-species phyt oplankton biomass, expressed as carbon (gCm-3); is the multi-species phytoplankton growth rate (1/day); Kax is the respiration rate of multi-species phytoplankton (1/day); ZOOC is zooplankton biomass incepted multi-species phytoplankton, formulated as carbon (gCm-3); Kzx is the respiration rate of zooplankton (1/day); KONM is the rate of organic nitr ogen mineralization (1/day); KNN is the nitrification rate constant (1/day); Hnit is the half saturation constant for oxygen

PAGE 248

225 limitation (mg O2 ); dan is the sorption/desorption rate of NH4 from sediment particles (1/day); pan is the partition coefficient between NH4 and PIN (1/ g); c is suspended sediment concentration; and Pn is the ammonia uptake prefer ence factor for each species (0~1). Nitrate and nitrite nitrogen (NO3) include nitrification, denitrification and multispecies phytoplankton uptake processes. For water column: 3 3 3 4 3* **1* NO DOH H K NH DOH DO KPHYC PA t NOno no DN nit NN i aini NC For sediment column: 3 3 3 4 3* NO DOH H KNH DOH DO K t NOno no DN nit NN where KDN is the denitrification rate constant (1/day) and Hno3 is the half saturation constant for denitrification. Soluble organic nitrogen (SON) includes ammonification, mortality of multispecies phytoplankton, mortal ity of zooplankton and sorp tion/desorption reaction. For water column: )** (* **)1(* SONcPPONd SONKZOOCKPHYCKKPDN A t SONon on ONM zsi asi NC For sediment column: )** (* SONcPPONdSONK t SONon on ONM

PAGE 249

226 where KPDN is the mortality ratio of multi-species phytoplankton and zooplankton for PON(0~1), don is the sorption/desorption rate of S ON from sediment particles (1/day) and pon is the partition coefficient between SON and PON (1/ g). Particulate organic nitrogen (PON) includes mortality of multi-species phytoplankton and zooplankton, settling, and a sorption/desorption reaction. For water column: )** (* *** SONcPPONd z PONws ZOOCKPHYCKKPDN A t PONon on p zsi asi NC For sediment column: )** (* SONcPPONd z PONws t PONon on p where Kas is the mortality rate of multi -species phytoplankton (1/day); Kzs is the mortality rate of zooplankton (1/day); and wsp is a settling velocity for particulate species, which is the same as that of suspended sediment particles. Particulate inorganic nitrogen (PIN) in cludes settling, and a sorption/desorption reaction. For water and sediment columns: )**(* *4NHcPPINd z PINws t PINan an p E.2 Phosphorous Cycle Phosphorous can be classified into two groups: dissolved phosphorous and particulate phosphorous. The phosphorous cycl e of CH3D-IMS includes the following processes:

PAGE 250

227 Mineralization of organic nitrogen Uptake of soluble reactive phosphorous by multi-species phytoplankton Conversion of multi-species phyt oplankton phosphorous to zooplankton phosphorous by gazing Excretion and mortality by multi-sp ecies phytoplankton and zooplankton Settling for particulate phosphorous Sorption/desorption reactions The mass balance equation for phosphorous st ate variables are wr itten by combining nitrogen transformation processes. Soluble reactive phosphorous (SRP) in cludes mineralization, multi-species phytoplankton uptake, respiration of multispecies phytoplankton, and mortality of zooplankton. For water column: )**(* ** SRPcPPIPd ZOOCKPHYCKPHYC ASOPK t SRPip ip zxi axii ai PC opm For sediment column: )**(* SRPcPPIPdSOPK t SRPip ip opm where Kopm is the rate constant for mineralizati on of SOP (1/day), which is the function of pH and temperature ; APC is the phosphorous-to-carbon ratio of phytoplankton; dip is the sorption/desorption rate of SRP fro m sediment particles (1/day); and pip is the partition coefficient between SRP and PIP (1/ g). Soluble organic phosphorous (SOP) include s mineralization, mortality of multispecies phytoplankton, mortality of zoopla nkton, and the sorption/desorption reaction.

PAGE 251

228 For water column: )** (* **)1(* SOPcPPOPd SOPKZOOCKPHYCKKPDP A t SOPop op opm zsi asi PC For sediment column: )** (* SOPcPPOPdSOPK t SOPop op opm where KPDP is the mortality ratio of multi-species phytoplankton and zooplankton for POP (0~1), dop is the sorption/desorption rate of SO P from sediment particles (1/day), and pop is the partition coefficient between SOP and POP (1/ g). Particulate organic phosphorous (POP) in cludes mortality of multi-species phytoplankton and zooplankton, settling, and a sorption/desorption reaction. For water column: )** (* *** SOPcPPOPd z POPws ZOOCKPHYCKKPDPA t POPop op p zsi asi PC For sediment column: )** (* SOPcPPOPd z POPws t POPop op p Particulate inorganic nitrogen (PIP) in cludes settling, and a sorption/desorption reaction. For water and sediment columns: )**(* SRPcPPIPd z PIPws t PIPip ip p

PAGE 252

229 APPENDIX F LIGHT ATTENUATION COEFFICIENT Kirk (1984) developed th e equation calculating vert ical light attenuation coefficient. 2/1 201 2 0)(*)(*)*()( 1 )( bagga Kt t d where is the wavelength between 400 nm and 700 nm, Kd( ) is a vertical light attenuation coefficient, 0 is a solar zenith angle, at ( ) is the total absorption, g1 is a numerical constant (=0.473), g2 is a numerical cons tant (=0.218), and b ( ) is a scattering Total absorption at ( ) is partitioned into attributes of water ( aw), phytoplankton ( aph), dissolved color (adc), and detritus ( ad). )()()()()( d dc ph w taaaaa The absorption of water, aw ( ), can be determined from literature values (Smith and Baker, 1981). Phytoplankton absorption, aph ( ), is calculated from the linear relationship of normalized spectra ( ) and chlorophyll a concentration (Dixon and Kirkpatrick, 1999): )( *), (*0209.0)( Spectra Normalized corrected alChlorophyl aph To obtain normalized spectra (), 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, adc ( ), for each wavelength in the visible spectrum can be found using a negative exponen tial function (Bricaud et al., 1981).

PAGE 253

230 )440(*exp*)(440 dc dcs ga where g440 is the absorption by dissolv ed color at 440 nm and sdc is spectral slope. Dixon and Kirkpatrick (1999) calculat ed empirical absorption at 440 nm and spectral slope as a function of color (in PCU): 9329.0,129, *0667.02 440 rn Color g 5111.0,129,0178.0*00003.002 rn Color sdc Absorption due to organic and mineral detritus is represented as a function of turbidity (Gallegos, 1993): )]440(*[exp* )(400 d bl ds Turbidity a where turbidity is in NTUs and bl is the longwave absorption cross section, 400 is the maximum detritus absorption at 400 nm, and sd is exponential slope. As the absorption by detritus, scattering can be described as a function of turbidity (Morel and Gentili, 1991): Turbidity b *)/550()( The values of vertical light attenuation coefficient, Kd ( ), which is calculated by total absorption and scattering, are used to calculate irradiance Ez ( ): ]*)([exp*)()(0r d zzK EE where E0 ( ) is the incident irradiation and ( ) and zr is the referecne depth. The spectrum of incident sunlight data in Weast (1977) is used as incident irradiation, E0 ( ), in Gallegos's work (1993). These da ta are shown in Table F-1. E0 ( ) and Ez ( ) are integrated over the visible spectrum getting PAR0 for the incident PAR and zPAR for the calculated PAR at the reference depth. These integrated

PAGE 254

231 values are used to obtain spectrally sensitive attenuation coefficient Kd ( PAR) by using the rearranged form of the Lambert-Beer equation: 0ln* 1 )( PAR PAR z PARKz r d This Kd ( PAR ) is used in the model to calculate light levels throughout the water column as a function of measured incident light inte nsity. More detailed information is found in Christian (2001) and Park (2004). Table F-1. Spectrum of incident sunlight data (G allegos, 1994) (nm) E0 (nm) E0 (nm) E0 400 4.780 405 5.568 505 8.108 605 8.332 410 6.003 510 8.026 610 8.340 415 6.052 515 7.894 615 8.323 420 6.135 520 7.970 620 8.305 425 6.017 525 8.130 625 8.289 430 5.893 530 8.163 630 8.271 435 5.940 535 8.133 635 8.267 440 6.659 540 8.051 640 8.263 445 7.152 545 7.993 645 8.239 450 7.548 550 7.993 650 8.213 455 7.826 555 7.982 655 8.207 460 7.947 560 7.937 660 8.201 465 7.963 565 8.055 665 8.180 470 7.990 570 8.160 670 8.157 475 8.119 575 8.265 675 8.136 480 8.324 580 8.318 680 8.114 485 8.014 585 8.375 685 8.102 490 7.990 590 8.387 690 8.089 495 8.113 595 8.369 695 8.052 500 8.119 600 8.359 700 8.013

PAGE 255

232 APPENDIX G HEAT FLUX G.1 Short-Wave Solar Radiation Short wave solar radiation is the main source of heat fl ux through the air-sea surface. This study uses an empirical form ula suggested by Reed (1977). The rate at which short-wave solar energy enters the sea is dependent on a number of factors, such as length of day, absorption in the atmosphere the sun altitude, the cloud cover, and the reflection at the sea surface. )1(*)*0019.0*62.01(*)*2.061.0(**1353 C s s Qs where Qs is short-wave solar radiation, s is the sine of the solar elevation angle, C is cloud cover (tenths), is noon solar elevation angle, and is albedo. The solar elevation angle is computed to th e nearest 0.1 from date and time for the latitude and longitude of the study area using the following equations (Miller and McPherson, 1995) )cos(*)cos(*)cos()sin(*)sin()sin( ))sin(*39785077.0arcsin( )2cos(*0016.0)2sin(*0199.0)cos(*0795.0)sin(*9148.19348.79 )(*15 )2cos(*0608.0)2sin(*1538.0)cos(*0043.0)sin(*1236.012 242.365 360 *)1( s d where is the angular fraction of the y ear (in degrees), d is Julian date, is true solar noon (in hours), is the solar hour angle (in degrees), is the Greenwich Mean Time (in

PAGE 256

233 hours), is the longitude (in degrees), is an estimate of the true longitude of the sun (in degrees), is the solar declin ation (in degrees), is the latitude (in degrees), and is the solar elevation angle (in degrees). G.2 Long-Wave Solar Radiation The back radiation term, QL, is the net amount of energy lost by the sea as longwave radiation. The outward radiation from the sea is always greater than the inward radiation from the atmosphere. Therefore QL presents a loss of energy from the sea. The amount of long-wave flux depends on surface water/atmospheric temperature, humidity, and cloud cover (Stevenson and Niiler, 1983). )7.01()(4)05.039.0(10 3 5.0 4N TTTe TQss s L where is the emissivity of the sea surface (=0.97), is Stefan-Boltzmans constant, e is the water vapor pressure (mb) at a height of 10m, Ts is the sea surface temperature (K), T10 is the atmospheric temperature (K), and N is the fraction of the sky that is covered by clouds. The factor (1-0.7N) is the cloud correction factor suggested by Reed (1976) for the tropics. The Antoine constants ( http://www.owlnet.rice.edu/~chbe301/31.html ) give the vapor pressure as a function of tem perature so that the approximate value for water vapor pressure is: 997.38 44.3985 5362.16expsT e G.3 Latent Heat Flux Latent heat flux is one of the dominant components of air sea energy exchange, and it occurs as a heat transfer from the sea to the atmosphere. The flux is estimated by using the bulk aerodynamic equation (Mellor, 1996)

PAGE 257

234 )(10 10qqLUCQs E E where is the air density, CE is the evaporation coefficien t, L is the latent heat of vaporization (2.4x106 J/kg), U10 is the wind speed at a height of 10m, qs is the specific humidity of saturated air at th e sea surface temperature, and q10 is the specific humidity of the air at a height of 10m. The air density is determined from the ideal gas equation in the form, =Pa/(RTa), where Pa is air pressure, R is the universal gas constant (287.04 J/kg K), and Ta is air temperature. The evaporation coefficient is a constant value over a wide range of wind speed. Smith (1989) recommended CE=(1.2 0.1)x10-3 for winds between 4 and 14m/s. For winds up to 18m/s, the value CE=(1.12 0.24)x10-3 was suggested (DeCosmo et al., 1996). G.4 Sensible Heat Flux Sensible heat flux is due to temperature gr adient in the air above the sea. The rate of loss or gain of heat is proportional to th e temperature gradient, heat conductivity, and specific heat of air at consta nt pressure (Mellor, 1996). )(10 as Hp HTTUCcQ where cp is the specific heat of air (1004.8 J/kg K) and CH is the sensible heat transfer coefficient (Stanton number). Friehe and Sc hmitt (1976) obtained slightly different CH values for unstable and stable conditions, 0.97x10-3 and 0.86x10-3

PAGE 258

235 APPENDIX H TEMPERATURE, LIGHT, WIND AND RIVE R NUTRIENT LOADS The following figures illustrate the info rmation of surface water temperature, surface light, wind, and river nutrient loads at IRLV17 (Segment 1), IRLI07 (Segment 2), IRLB09 (Segment 3), IRLI18 (Segment 4), IR LI23 (Segment 5), IRLIRJ01 (Segment 6), and IRLIRJ05 (Segment 7). The river nutrient loads represent total daily nutrient loads from rivers in that segment.

PAGE 259

236 JulianDay1997through1999 Phosphorousloads(kg/day) 400 600 800 1000 0 1000 2000 3000 4000 5000 JulianDay1997through1999 Nitrogenloads(kg/day) 400 600 800 1000 0 5000 10000 15000 20000 JulianDay1997through1999 Windspeed(m/s) 400 600 800 1000 5m/s JulianDay1997through1999 Surfacetemperature(C) 400 600 800 1000 0 10 20 30 40IRLV17 JulianDay1997through1999 Surfacelight(Langleys/day ) 400 600 800 1000 0 500 1000 1500 Figure H-1. The surface water temperatur e, surface light, and wind at IRLV17. Total daily nitrogen and phosphorous loads from rivers in segment 1.

PAGE 260

237 JulianDay1997through1999 Surfacelight(Langleys/day ) 400 600 800 1000 0 200 400 600 800 1000 1200 JulianDay1997through1999 Windspeed(m/s) 400 600 800 1000 5m/s JulianDay1997through1999 Nitrogenloads(kg/day) 400 600 800 1000 0 2000 4000 6000 8000 10000 JulianDay1997through1999 Phosphorousloads(kg/day) 400 600 800 1000 0 1000 2000 JulianDay1997through1999 Surfacetemperature(C) 400 600 800 1000 0 10 20 30 40IRLI07 Figure H-2. The surface water temperatur e, surface light, and wind at IRLI07. Total daily nitrogen and phosphorous loads from rivers in segment 2.

PAGE 261

238 JulianDay1997through1999 Surfacetemperature(C) 400 600 800 1000 0 10 20 30 40IRLB09 JulianDay1997through1999 Surfacelight(Langleys/day ) 400 600 800 1000 0 500 1000 1500 2000 JulianDay1997through1999 Windspeed(m/s) 400 600 800 1000 5m/s JulianDay1997through1999 Nitrogenloads(kg/day) 400 600 800 1000 0 2000 4000 6000 8000 10000 JulianDay1997through1999 Phosphorousloads(kg/day) 400 600 800 1000 0 1000 2000 3000 4000 5000 Figure H-3. The surface water temperatur e, surface light, and wind at IRLB09. Total daily nitrogen and phosphorous loads from rivers in segment 3.

PAGE 262

239 JulianDay1997through1999 Surfacetemperature(C) 400 600 800 1000 0 10 20 30 40IRLI18 JulianDay1997through1999 Surfacelight(Langleys/day ) 400 600 800 1000 0 200 400 600 800 1000 JulianDay1997through1999 Windspeed(m/s) 400 600 800 1000 5m/s JulianDay1997through1999 Phosphorousloads(kg/day) 400 600 800 1000 0 1000 2000 3000 4000 5000 JulianDay1997through1999 Nitrogenloads(kg/day) 400 600 800 1000 0 2000 4000 6000 8000 10000 Figure H-4. The surface water temperatur e, surface light, and wind at IRLI18. Total daily nitrogen and phosphorous loads from rivers in segment 4.

PAGE 263

240 JulianDay1997through1999 Surfacetemperature(C) 400 600 800 1000 0 10 20 30 40IRLI23 JulianDay1997through1999 Windspeed(m/s) 400 600 800 1000 5m/s JulianDay1997through1999 Phosphorousloads(kg/day) 400 600 800 1000 0 1000 2000 3000 4000 5000 JulianDay1997through1999 Nitrogenloads(kg/day) 400 600 800 1000 0 5000 10000 15000 20000 JulianDay1997through1999 Surfacelight(Langleys/day ) 400 600 800 1000 0 200 400 600 800 1000 1200 Figure H-5. The surface water temperatur e, surface light, and wind at IRLI23. Total daily nitrogen and phosphorous loads from rivers in segment 5.

PAGE 264

241 JulianDay1997through1999 Surfacetemperature(C) 400 600 800 1000 0 10 20 30 40IRLIRJ01 JulianDay1997through1999 Windspeed(m/s) 400 600 800 1000 5m/s JulianDay1997through1999 Nitrogenloads(kg/day) 400 600 800 1000 0 5000 10000 15000 20000 JulianDay1997through1999 Phosphorousloads(kg/day) 400 600 800 1000 0 1000 2000 3000 4000 5000 JulianDay1997through1999 Surfacelight(Langleys/day ) 400 600 800 1000 0 200 400 600 800 1000 1200 Figure H-6. The surface water temperature, surface light, and wind at IRLIRJ01. Total daily nitrogen and phosphorous loads from rivers in segment 6.

PAGE 265

242 JulianDay1997through 1999 Windspeed(m/s) 400 600 800 1000 5m/s JulianDay1997through1999 Nitrogenloads(kg/day) 400 600 800 1000 0 5000 10000 15000 20000 JulianDay1997through1999 Phosphorousloads(kg/day) 400 600 800 1000 0 1000 2000 3000 4000 5000 JulianDay1997through1999 Surfacetemperature(C) 400 600 800 1000 0 10 20 30 40IRLIRJ05 JulianDay1997through1999 Surfacelight(Langleys/day ) 400 600 800 1000 0 200 400 600 800 1000 1200 Figure H-7. The surface water temperature, surface light, and wind at IRLIRJ05. Total daily nitrogen and phosphorous loads from rivers in segment 7.

PAGE 266

243 APPENDIX I BOTTOM SHEAR STRESS IN THE INDIAN RI VER L AGOON The following figures illustrate the botto m shear stress at IRLV17 (Segment 1), IRLI07 (Segment 2), IRLB09 (Segment 3), IRLI18 (Segment 4), IRLI23 (Segment 5), IRLIRJ01 (Segment 6), and IRLI RJ05 (Segment 7). In addition, to better demonstrate the entire IRL bottom shear stress, a figure pres ent the percentage of simulated bottom shear stress exceeded 1 dyne/cm2 over the simulati on period in the IRL. In the bottom shere stress plot, there are two solid lines wh ich represent the critical shear stress 0.15 dyne/cm2 and 1.2 dyne/cm2 for sediment type 1 (silt or clay) and type 5 (very coarse), respectively.

PAGE 267

244 YearDay1997through1999 Bottomshearstress(dyne/cm2) 400 600 800 1000 0 2 4 6 8IRLV171.2dyne/cm20.15dyne/cm2 YearDay1997through1999 Current-inducedstress(dyne/cm2) 400 600 800 1000 0 2 4 6 8 YearDay1997through1999 Wave-inducedstress(dyne/cm2) 400 600 800 1000 0 0.5 1 1.5 2 Figure I-1. The simulated bottom shear stress at IRLV17

PAGE 268

245 YearDay1997through1999 Bottomshearstress(dyne/cm2) 400 600 800 1000 0 0.5 1 1.5 2IRLI071.2dyne/cm20.15dyne/cm2 YearDay1997through1999 Current-inducedstress(dyne/cm2) 400 600 800 1000 0 0.5 1 1.5 2 YearDay1997through1999 Wave-inducedstress(dyne/cm2) 400 600 800 1000 0 0.5 1 1.5 2 Figure I-2. The simulated bottom shear stress at IRLI07

PAGE 269

246 YearDay1997through1999 Bottomshearstress(dyne/cm2) 400 600 800 1000 0 0.5 1 1.5 2 2.5 3IRLB091.2dyne/cm20.15dyne/cm2 YearDay1997through1999 Current-inducedstress(dyne/cm2) 400 600 800 1000 0 0.5 1 1.5 2 2.5 3 YearDay1997through1999 Wave-inducedstress(dyne/cm2) 400 600 800 1000 0 0.5 1 1.5 2 2.5 3 Figure I-3. The simulated bottom shear stress at IRLB09

PAGE 270

247 YearDay1997through1999 Bottomshearstress(dyne/cm2) 400 600 800 1000 0 1 2 3 4IRLI181.2dyne/cm20.15dyne/cm2 YearDay1997through1999 Current-inducedstress(dyne/cm2) 400 600 800 1000 0 1 2 3 4 YearDay1997through1999 Wave-inducedstress(dyne/cm2) 400 600 800 1000 0 1 2 3 4 Figure I-4. The simulated bottom shear stress at IRLI18

PAGE 271

248 YearDay1997through1999 Bottomshearstress(dyne/cm2) 400 600 800 1000 0 1 2 3 4IRLI231.2dyne/cm20.15dyne/cm2 YearDay1997through1999 Current-inducedstress(dyne/cm2) 400 600 800 1000 0 1 2 3 4 YearDay1997through1999 Wave-inducedstress(dyne/cm2) 400 600 800 1000 0 1 2 3 4 Figure I-5. The simulated bottom shear stress at IRLI23

PAGE 272

249 YearDay1997through1999 Wave-inducedstress(dyne/cm2) 400 600 800 1000 0 1 2 3 4 YearDay1997through1999 Bottomshearstress(dyne/cm2) 400 600 800 1000 0 1 2 3 4IRLIRJ011.2dyne/cm20.15dyne/cm2 YearDay1997through1999 Current-inducedstress(dyne/cm2) 400 600 800 1000 0 1 2 3 4 Figure I-6. The simulated bo ttom shear stress at IRLIRJ01

PAGE 273

250 YearDay1997through1999 Bottomshearstress(dyne/cm2) 400 600 800 1000 0 2 4 6 8IRLIRJ051.2dyne/cm20.15dyne/cm2 YearDay1997through1999 Current-inducedstress(dyne/cm2) 400 600 800 1000 0 2 4 6 8 YearDay1997through1999 Wave-inducedstress(dyne/cm2) 400 600 800 1000 0 0.5 1 1.5 2 Figure I-7. The simulated bo ttom shear stress at IRLIRJ05

PAGE 274

251 EastingUTM,meter) Northing(UTM,meter) 550000 600000 3.05E+06 3.1E+06 3.15E+06 3.2E+0620 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 Percent(%) to100 Figure I-8. The percentage of simulated bottom shear stress exceeded 1 dyne/cm2 over the simulation period in the IRL

PAGE 275

252 APPENDIX J TAYLOR DIAGRAM Taylor diagram s (Taylor, 2001) provide a way of graphically summarizing how closely simulated predictions match observations. The similarity between model predictions and observations is quantified in terms of their correlation and standard deviations. The following figures illustrate the diagrams in association with water quality nutrients at the IRL. Each diagram includes statistical values of obs ervation, the baseline simulation, and sensitivity test simulations. Figure J-1. Taylor diagram displaying a st atistical comparison in terms of phytoplankton (M is the observation, B is the base line simulation, and T1 though T11 are the sensitivity test 1 to test11, respectively).

PAGE 276

253 Figure J-2. Taylor diagram displaying a st atistical comparison in terms of dissolved oxygen (M is the observation, B is the baseline simulation, and T1 though T11 are the sensitivity test 1 to test11, respectively). Figure J-3. Taylor diagram di splaying a statistical comparison in terms of TOC (M is the observation, B is the baseline simulation, and T1 though T11 are the sensitivity test 1 to test11, respectively).

PAGE 277

254 Figure J-4. Taylor diagra m displaying a statistical comparison in terms of NO3 (M is the observation, B is the baseline simulation, and T1 though T11 are the sensitivity test 1 to test11, respectively). Figure J-5. Taylor diagram di splaying a statistical comparison in terms of dissolved TKN (M is the observation, B is the base line simulation, and T1 though T11 are the sensitivity test 1 to test11, respectively).

PAGE 278

255 Figure J-6. Taylor diagram di splaying a statistical comparison in terms of SRP (M is the observation, B is the baseline simulation, and T1 though T11 are the sensitivity test 1 to test11, respectively). Figure J-7. Taylor diagram displaying a statistical comparison in terms of particulate nitrogen (M is the observation, B is the baseline simulation, and T1 though T11 are the sensitivity test 1 to test11, respectively).

PAGE 279

256 Figure J-8. Taylor diagram displaying a statistical comparison in terms of particulate phosphorous (M is the observation, B is the baseline simulation, and T1 though T11 are the sensitivity test 1 to test11, respectively). Figure J-9. Taylor diagram displaying a st atistical comparison in terms of dissolved silica (M is the observation, B is the baseline simulation, and T1 though T11 are the sensitivity test 1 to test11, respectively)

PAGE 280

257 APPENDIX K WATER QUALITY PARAMETER ESTIMATION The water quality m odel includes many ki netic equations which have uncertain parameters. Parameters are not a common e nvironmental measurement, and thus the values of parameters are selected from a range of feasible values: field observation, laboratory experimentation, previous st udy, etc. The popular method of parameter selection is the trial-and-error process; that is, adjustment is achieved until there is an optimal agreement between predicted nutrient values and measured nutrient data. This process is known as subjective and labori ous approach. More efficient and robust processes are required to obtain a reliable water quality model performance. Some algorithms are employed to estimate the unknown parameters: a genetic algorithm (Whingham and Recknagel, 2001), an exergy method (Jorgensen, 2001), an inverse method (Sun, 1994), etc. In this st udy, the modified Ga uss-Newton method is used to determine optimal parameters. This is considered the most efficient method for estimating parameters in nonlinear models (Bar d, 1970). The application of the method is diverse (Basulto et al., 1978; Yeh, 1986; Linga et al., 2006). K.1 Modified Gauss-Newton Method The modified Gauss-Newton method is de signed for minimi zing the objective function E(k). The objective f unction has the form of least squares and is used as a criterion to measure the difference between measured data and predicted values. )())(()(1 2 2 1kf ukuwkEL l l L l obs l cal ll

PAGE 281

258 where k is unknown parameters, wl is weighting factor, ul cal is predicted values, and ul obs is measured data. The first order derivatives of E(k) are i l L l l ik f f k E 12 (i=1,2,3,,M) After neglecting high order terms, the second or der derivatives of E(k) can be written as j l L l i l jik f k f kk E 1 22 (i=1,2,3,,M; j=1,2,3,,M) It is convenient to define a new matrix A M L LL M Mk f k f k f k f k f k f k f k f k f A ........ .................................. ........ ........21 2 2 2 1 2 1 2 1 1 1 Using matrix A, the first and second deri vatives of E(k) are in the following: fAET2 and AAET22 For kn+1 being a minimum of E(k), 0)(1 nkE is the necessary and sufficient condition. Let nnnkkk1 and use Taylor expansion of E(kn+1) 0*)()()(2n n n nnkkEkEkkE n T n T n n n nfAAAkEkEk1 1 2)()()( where kn is the Gauss-Newton dire ction in nth iteration.

PAGE 282

259 The search sequence does not converge, i.e., E(kn+1) > E(kn) for some n. It is necessary to modify the Gauss-Newton algorithm. n T n n T n n n nfAIAAkEIkEk1 1 2) ()( )( where I is the unit matrix, is a coefficient. E(kn+1) > E(kn) can always be expected by increasing the value of (Sun, 1994). K.2 Examples of Modified Gauss-Newton Method There are three examples to illustrate the performance of modified Gauss-Newton method: homogeneous gas phase reaction, to luene hydrogenation, and methyl esters hydrogenation. K.2.1 Homogeneous Gas Phase Reaction The first example is the hom ogeneous gas phase reaction. 2 22 2 NOONO At constant volume and temper ature, the experiment for homogeneous gas phase reaction was conducted. The observed data are shown in Table K-1. This reaction can be written by a differential equation (Bellman, 1967). 2 2 2 1*)9.91(*)2.126(* xkx x k dt dx where x is the concentration of NO2, and k1and k2 are the rate constants. The initial parameter values of k1and k2 are 0.000001 and 0.000001, respectively. Within 10 iterations, the optimal parameters are obtained, 0.0000046 and 0.00028 for k1and k2, respectively. The predicted values using the parameters obtained by modified GaussNewton method and observed data can be seen in Figure K-1. To indicate the discrepancy between predicted values and observed data, the relative error was employed as a skill assessment.

PAGE 283

260 Observed edicted Observed ErelPr The relative error between predicted values and observed data is less than 3.8 %. Table K-1. Data for the homogene ous gas phase reaction (Bellman, 1967) Time Concentration of NO2 0 0 1 1.4 2 6.3 3 10.5 4 14.2 5 17.6 6 21.4 7 23.0 9 27.0 11 30.5 14 34.4 19 38.8 24 41.6 29 43.5 39 45.3 TIME Concentration(NO2) 0 10 20 30 40 50 0 10 20 30 40 50 Data Model Figure K-1. Homogene ous gas phase reaction

PAGE 284

261 K.2.2 Toluene Hydrogenation As the second example, the toluene hydrogenation is considered. The reaction kinetic of toluene hydrogenation is 3 2 & 12 11xx xr rr where x1 is toluene, x2 is 1-methyl-cyclohexane, x3 is methyl-cyclohexane, r1 is the hydrogenation rate (for ward reaction), r-1 is the disproportionation rate (backward reaction), and r2 is the hydrogenation rate. These reactions are represented as differen tial equation forms (Bel ohlav et al., 1997). 35214 22 35214 141 11 1* * ** xkxxk xk xkxxk xkk rr dt dx 35214 23 35214 22 35214 141 211 2* * ** xkxxk xk xkxxk xk xkxxk xkk rrr dt dx 35214 23 2 3* xkxxk xk r dt dx where k1, k2, k3, k4, and k5 are unknown parameters. The experiment of toluene hydrogenation was performed at ambient temperature and atmospheric pressure in a semi-batch isothermal stirred reactor (Table K-2). Luss (2001) calculated unknown parameters of toluene hydrogenation using an LJ optimization procedure. All initial unknown parameter valu es were set by 0.25. The reported parameter values are 0.02511, 0.00326, 0.00873, 1.27258, and 1.24230 for k1, k2, k3, k4, and k5, respectively. This study set the sa me number (0.25) as all initial parameter values. The values for parameters k1, k2, k3, k4, and k5 were 0.02813, 0.00369, 0.00696, 0.92223, and 0.82247, respectively. Even though these parameter values differ from those given by Luss (2001), the diffe rences in the responses are visually

PAGE 285

262 insignificant. Figure K-2 shows a visual difference between predicted values and measured data. The relative errors of to luene, 1-methyl-cyclohexane, and methylcyclohexane are 10.5, 4.86, and 4.01 %, respectively. Table K-2. Data for the hydrogenation of Toluene (Belohlav et al., 1997) Time(min) Toluene (x1) 1-methyl-cyclohexane (x2) Methyl-cyclohexane (x3) 0 1.000 0.000 0.000 15 0.695 0.312 0.001 30 0.492 0.430 0.08 45 0.276 0.575 0.151 60 0.225 0.570 0.195 75 0.163 0.575 0.224 90 0.134 0.533 0.330 120 0.064 0.462 0.471 180 0.056 0.362 0.580 240 0.041 0.211 0.747 320 0.031 0.146 0.822 360 0.022 0.080 0.898 380 0.021 0.070 0.909 400 0.019 0.073 0.908 Time(min) Concentration 0 100 200 300 400 0 0.2 0.4 0.6 0.8 1 1.2 Measured(x1) Measured(x2) Measured(x3) Model(x1) Model(x2) Model(x3) Figure K-2. Toluene hydrogenation

PAGE 286

263 K.2.3 Methyl Ester Hydrogenation The third example considered is the me thyl esters hydrogenation. The reaction scheme for the methyl ester hydrogenation is 4 3 2 13 2 1xxxxr r r where x1, x2, x3, and x4 are the methyl esters of linolenic linoleic, oleic, and stearic acids; and r1, r2, and r3 are the hydrogenation rates. The di fferential equations of methyl ester hydrogenation are as follows, 4635241 11 1 1*** xkxkxkx xk r dt dx 4635241 242 4635241 11 21 2*** ** *** xkxkxkx xkk xkxkxkx xk rr dt dx 4635241 353 4635241 242 32 3*** ** *** ** xkxkxkx xkk xkxkxkx xkk rr dt dx 4635241 353 3 4*** ** xkxkxkx xkk r dt dx where k1, k2, k3, k4, k5, and k6 are unknown parameters. The experiment for the methyl ester hydroge nation was carried out in an autoclave at elevated pressure and temperature. The data for this experiment are given in Table K-3 (Belohlav et al., 1997). Table K-3. Data for the hydrogenation of methylesters (Bel ohlav et al., 1997) Time(min) Linolenic acid (x1) Linoleic acid (x2) Oleic acid (x3) Stearic acid (x4) 0 0.1012 0.2210 0.6570 0.0208 10 0.0150 0.1064 0.6941 0.1977 14 0.0044 0.0488 0.6386 0.3058 19 0.0028 0.0242 0.5361 0.4444 24 0.0029 0.0015 0.3956 0.6055 34 0.0017 0.0005 0.2188 0.7808 69 0.0003 0.0004 0.0299 0.9680 124 0.0001 0.0002 0.0001 0.9982

PAGE 287

264 All initial values of unknow n parameters are given as 1.0. The parameter estimation using modified Gauss-Newton me thod produced the parameter values which are 0.27363, 0.10458, 0.28572, 1.82591, 0.13683, and 0.21482 for k1, k2, k3, k4, k5, and k6, respectively. The minimum value of objective function was 0.000507. This value is nearly equivalent to that of other studies: 0.000458 (Luss, 2001), 0.000459 (Linga et al., 2006), and 0.000694 (Belohlav et al., 1997). The methyl ester relative errors are 32.9, 10.6, 0.9, and 0.7 % for linolenic, linoleic, ol eic, and stearic acids, respectively. The high relative error of linolenic aci d is a consequence of its measured values which are substantially lower than other methyl esters. Time(min) Concentration 0 25 50 75 100 125 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 Measured(x1) Measured(x2) Measured(x3) Measured(x4) Model(x1) Model(x2) Model(x3) Model(x4) Figure K-3. Methyl ester hydrogenation

PAGE 288

265 K.3 Parameter Estimation of Water Quality Kinetic Equations The parameter estimation method applied to kinetic equations of water quality state variables includes: dissolv ed ammonium nitrogen (NH4), nitrate and nitrite nitrogen (NO3), soluble organic nitrogen (SON), particulate organic nitrogen (PON), soluble reactive phosphorous (SRP), soluble organic phosphorous (SOP), particulate organic phosphorous (POP), particulate inorganic nitrogen (PIP), phytoplankton (PHYC), zooplankton (ZOOC), dissolve oxygen (DO), and carbonaceous biochemical oxygen demand (CBOD). The symbols of each e quation are defined in Table K-4. )**(* * **4 4 4 4NHcPPINd NH pHH pH KNH DOH DO KSONK ZOOCKPHYC KPA t NHan an al al nit NN ONM zx axan NC 3 3 3 4 3* **1* NO DOH H K NH DOH DO KPHYC PA t NOno no DN nit NN an NC )** (* **)1(* SONcPPONdSONK ZOOCKPHYCKKPDN A t SONon on ONM zs as NC )** (* *** SONcPPONdPONws ZOOCKPHYCKKPDN A t PONon on p zs as NC )**(* *4NHcPPINdPINws t PINan an p ) /()**(* ** reaction desorption Sorption SRPcPPIPd ZOOCKPHYCKPHYC ASOPK t SRPip ip zx ax a PC opm

PAGE 289

266 )** (* **)1(* SOPcPPOPdSOPK ZOOCKPHYCKKPDP A t SOPop op opm zs as PC )** (* *** SOPcPPOPdPOPws ZOOCKPHYCKKPDPA t POPop op p zs as PC )**(* SRPcPPIPdPIPws t PIPip ip p ZOOC PHYC WSKasKax t PHYCz* SRPH SRP NONHH NONH IfTfTp N ref, min*)(*)(*)()(3 4 3 4 max ZOOC KzsKzx t ZOOCz* PHY PHY PHY T z zzTrsPHY H TrsPHY **)(20 max PHYCAocKasKax P NH DOH DO K CBOD DOH DO KDODOK t DOan NIT NN CBOD D s AE**] *)*3.03.1[( 14 64 )(*4 )* *(* 14 32 4 5 *)1(*3 3 3ZOOCKzs PHYCKasAocNO DOH H K CBOD DO H DO KCBOD fd ws t CBODno no DN CBOD D CBOD CBOD This study estimated 25 unknown parameters fr om 13 water quality state variables: phytoplankton maximum growth rate (( )max), respiration rate of phytoplankton (Kax), mortality rate of phytoplankton (Kas), phytoplan kton settling rate (W S), half-saturation constant for phytoplankton uptake nitrogen (HN), half-saturation constant for

PAGE 290

267 phytoplankton uptake phosphorous (Hp), reaeration rate constant (KAE), oxidation rate constant (KD), nitrification rate constant (KNN), settling rate for particulate CBOD (wsCBOD), denitrification rate constant (KDN), ammonification rate of SON (Konm), rate constant for mineralization of SOP (Kopm), sorption/desorption rate of SON from sediment particles (don), partition coefficient between SON and PON (pon), settling rate for particulate species (wsp), sorption/desorption rate of NH4 from sediment particles (dan), partition coefficient between NH4 and PIN (pan), sorption/desorption rate of SOP (dop), partition coefficient between SOP and POP (pop), sorption/desorption rate of SRP from sediment particles (dip), partition coefficient between SRP and PIP (dip), zooplankton maximum growth (( z)max), respiration rate of zooplankton (Kzx), and mortality rate (Kzs). The Department of Civil and Coastal Engi neering at the University of Florida collected field data to study the resuspension of sediments and particulate nutrients (Sheng et al., 1998a). The in-situ platform with numerous instruments (wave gauges, wind anemometers, auto-samplers, temperature/salinity gauges, and optical scatter sensors, etc.) was deployed at Latitude 27 56' 33'' N and Longitude 80 31' 57'' E in the IRL from Feb. 19th to 22nd, 1999. The data included nutri ents, algae, CBOD, DO, TSS, and other measurements and were sampled every hour. The initial unknown parameter values and the estimated parameter values ar e shown in Table K-5. To better and faster converge to the optimum, the initial value of each parameter was chosen from a range of feasible values instead of random values. Due to the unavailability of measured zooplankton, NO3, and SOP, these nutrients were not c onsidered to be a part of objective function.

PAGE 291

268 Table K-4. Definition of sy mbols in kinetic equations Name Definition ANC Nitrogen-to-carbon ra tio of phytoplankton Phytoplankton growth rate (1/day) Kax Respiration rate of phytoplankton (1/day) Kzx Respiration rate of zooplankton (1/day) Konm Ammonification rate of SON (1/day) KNN Nitrification rate constant (1/day) Hnit Half saturation constant for oxygen limitation (mg O2 ) dan Sorption/desorption rate of NH4 from sediment particles (1/day). pan Partition coefficient between NH4 and PIN (1/ g) C Suspended sediment concentration Pn Ammonia uptake preference f actor for each species (0~1) KDN Denitrification rate constant (1/day) Hno3 Half saturation constant for denitrification KPDN Mortality ratio of phytoplankton and zooplankton for PON (0~1) don Sorption/desorption rate of SON from sediment particles (1/day) pon Partition coefficient between SON and PON (1/ g) Kas Mortality rate of phytoplankton (1/day) Kzs Mortality rate of zooplankton (1/day) wsp Settling rate for particulate species Kopm Rate constant for mineralization of SOP (1/day) APC Phosphorous-to-carbon ratio of phytoplankton. dip Sorption/desorption rate of SRP from sediment particles (1/day) pip Partition coefficient between SRP and PIP (1/ g). KPDP Mortality ratio of phytoplankt on and zooplankton for POP (0~1) dop Sorption/desorption rate of SOP from sediment particles (1/day) pop Partition coefficient between SOP and POP (1/ g) WS Phytoplankton settling rate (m/day) z Zooplankton growth rate (1/day) I Light intensity ( z)max Zooplankton maximum growth rate (1/day) ( )max Phytoplankton maximum growth rate (1/day) N Nitrogen P Phosphorous HN Half-saturation constant for phytoplankton uptake nitrogen (mg/l) Hp Half-saturation constant for phytoplankton uptake phosphorous (mg/l) TrsPHY Threshold phytoplankton concen tration for zooplankton uptake ( g/l) HPHY Half saturation concentrati on for phytoplankton uptake (gC-3) DOs Equilibrium oxygen concentration at standard pressure (mg/l) KAE Reaeration rate constant (1/day) KD Oxidation rate constant (1/day) HCBOD Half-saturation constant for oxygen limitation (mgO2) wsCBOD Settling rate for particulate CBOD AOC Oxygen-to-carbon ratio (gO2/gC) fdCBOD Fraction of the dissolved CBOD

PAGE 292

269 Table K-5. Values of initial unknown parameters and estimated parameters Name Initial value Estimated value ( )max 1.90 E00 (1/day) 1.90 E00 (1/day) Kax 5.00 E-01 (1/day) 1.00 E-01 (1/day) Kas 5.00 E-01 (1/day) 7.99 E-02 (1/day) WS 1.00 E-05 (1/day) 1.00 E-04 (1/day) Hp 5.00 E00 (g/l) 1.00 E00 (g/l) HN 2.00 E01 (g/l) 1.99 E01 (g/l) KAE 1.00 E-04 (1/day) 1.05 E-02 (1/day) KD 9.00 E-02 (1/day) 7.15 E-02 (1/day) KNN 9.00 E-02 (1/day) 9.07 E-02 (1/day) wsCBOD 1.00 E-05 (1/day) 1.00 E-04 (1/day) KDN 1.00 E-05 (1/day) 1.00 E-05 (1/day) Konm 1.00 E-02 (1/day) 1.99 E-02 (1/day) Kopm 1.00 E-02 (1/day) 1.01 E-02 (1/day) don 1.00 E-02 (1/day) 9.0 E-03 (1/day) pon 1.00 E-07 1.20 E-03 wsp 1.00 E-05 (1/day) 1.00 E-05 (1/day) dan 1.00 E-02 (1/day) 1.02 E-02 (1/day) pan 1.00 E-07 1.00 E-06 dop 1.00 E-02 (1/day) 9.96 E-03 (1/day) pop 1.00 E-07 1.02 E-04 dip 1.00 E-02 (1/day) 1.00 E-02 (1/day) pip 1.00 E-07 5.41 E-04 ( z)max 1.20 E-01 1.18 E-01 (1/day) Kzx 2.0 E-02 2.99 E-02 (1/day) Kzs 3.0 E-02 2.99 E-02 (1/day) The relative errors of water quality state variables are 59.9, 6.1, 11.8, 30.7, 46.4, 13.8, 39.4, 32.3 % for chlorophyll a DO, CBOD, NH4, SRP, dissolved TKN, particulate nitrogen, and particulate phosphorous, respec tively. Figures K-4 through K-7 show the model predictions and measured data for wate r quality state variables. As can be seen from particulate nutrient plots, the variation of m odel predictions was not as large as that of measured data. The weak fluctuation of model results would seem to stem from the limited consideration of the particulate nut rient dynamics due to wind and currentinduced shear stress.

PAGE 293

270 JulianDay1999 Dissolvedoxygen(mg/l) 51 51.5 52 52.5 53 0 2 4 6 8 10 JulianDay1999 Chlorophylla(ug/l) 51 51.5 52 52.5 53 0 4 8 12 16 20 Figure K-4. Temporal CBOD and phytoplankton variation JulianDay1999 NH4(ug/l) 51 51.5 52 52.5 53 0 20 40 60 80 100 JulianDay1999 SRP(ug/l) 51 51.5 52 52.5 53 0 10 20 30 40 50 Figure K-5. Temporal NH4 and SRP variation

PAGE 294

271 JulianDay1999 CBOD(mg/l) 51 51.5 52 52.5 53 0 2 4 6 8 10 12 JulianDay1999 Dissolvednotrogen(ug/l) 51 51.5 52 52.5 53 0 200 400 600 800 1000 Figure K-6. Temporal CBOD and dissolved nitrogen variation JulianDay ParticulateNitrogen(ug/l) 51 51.5 52 52.5 53 0 50 100 150 200 250 300 350 400 450 500 JulianDay ParticulatePhosphorous(ug/l) 51 51.5 52 52.5 53 0 10 20 30 40 50 Figure K-7. Temporal particulate nitrogen and phosphorous variation

PAGE 295

272 K.4 Application of Estimated Parameters In order to prove the reliability of esti mated parameters calcu lated by the modified Newton-Gauss method, the parameters were applied to a short-te rm IRL water quality simulation (episodic event #1). For sufficient model spin-up, the wa ter quality simulation started from September 2nd, 1997 using parameters obtained by trial-and-error method. The spin-up simulation finished before the period of episodic event #1, and all model variables and coefficients are saved to simula te a short-term IRL simulation. The episodic simulation was conducted from February 19th to 22nd, 1999, using estimated parameters calculated by the modified Newton-Gauss method. The Department of Civil and Coastal Engi neering at the University of Florida installed three in-situ platforms in the IRL during episodic event #1. The platforms collected eight water quality state variables every hour (Table K-6). The model predictions and measured data are shown in Figures K-8 through K-10. Each figure embodies eight different time series plots of water quality state variables (chlorophyll a oxygen, NH4, SPR, dissolved silica, TOC, dissolv ed TKN, particulate nitrogen, and particulate phosphorous). The rectangular shap e symbolizes measured water quality state variables. On the other side, the solid line represents the simulated water quality state variables. Table K-6. Description of episodic stations Name Latitude Longitude Northing UTM (meter) Easting UTM (meter) Date North Station 27 58' 54" 80 32' 10" 3095271 545608 Central Station 27 57' 27" 80 31' 36" 3092580 546551 South Station 27 56' 20" 80 31' 34" 3090516 546619 chlorophyll a DO, CBOD, NH4, SRP, dissolved TKN, particulate nitrogen, and particulate phosphorous

PAGE 296

273 JulianDay1999 Chlorophylla(ug/l) 51 52 53 0 2 4 6 8 10 JulianDay1999 NH4(ug/l) 51 52 53 0 20 40 60 80 JulianDay1999 SRP(ug/l) 51 52 53 0 5 10 15 20 25 30 JulianDay1999 DTKN(ug/l) 51 52 53 0 200 400 600 800 1000 1200 JulianDay1999 Oxygen(mg/l) 51 52 53 0 2 4 6 8 10 JulianDay1999 TOC(mg/l) 51 52 53 0 5 10 15 20 JulianDay1999 POP(ug/l) 51 52 53 0 50 100 150 JulianDay1999 PON(u/l) 51 52 53 0 100 200 300 400 Figure K-8. Comparison of measured and simulated (using water quality parameters obtained by modified Gauss-Newton method) water qual ity state variables at north station JulianDay1999 Chlorophylla(ug/l) 51 52 53 0 2 4 6 8 10 JulianDay1999 NH4(ug/l) 51 52 53 0 20 40 60 80 JulianDay1999 SRP(ug/l) 51 52 53 0 5 10 15 20 25 30 JulianDay1999 DTKN(ug/l) 51 52 53 0 200 400 600 800 1000 1200 JulianDay1999 Oxygen(mg/l) 51 52 53 0 2 4 6 8 10 JulianDay1999 TOC(mg/l) 51 52 53 0 5 10 15 20 JulianDay1999 POP(ug/l) 51 52 53 0 50 100 150 JulianDay1999 PON(u/l) 51 52 53 0 100 200 300 400 Figure K-9. Comparison of measured and simulated (using water quality parameters obtained by modified Gauss-Newton method) water qual ity state variables at central station

PAGE 297

274 JulianDay1999 Chlorophylla(ug/l) 51 52 53 0 2 4 6 8 10 JulianDay1999 NH4(ug/l) 51 52 53 0 20 40 60 80 JulianDay1999 SRP(ug/l) 51 52 53 0 5 10 15 20 25 30 JulianDay1999 DTKN(ug/l) 51 52 53 0 200 400 600 800 1000 1200 JulianDay1999 Oxygen(mg/l) 51 52 53 0 2 4 6 8 10 JulianDay1999 TOC(mg/l) 51 52 53 0 5 10 15 20 JulianDay1999 POP(ug/l) 51 52 53 0 50 100 150 JulianDay1999 PON(u/l) 51 52 53 0 100 200 300 400 Figure K-10. Comparison of measured and simulated (using water quality parameters obtained by modified Gauss-Newton method) water qual ity state variables at south station The relative errors between measured data and simulated results are 7.2, 27.9, 40.0, 88.1, 23.7, 176.1, 30.1, and 73.3 % for DO, chlorophyll a TOC, NH4, dissolved TKN, SRP, particulate nitrogen, and particulate phosphorous, respectively. The average relative error of these water quality st ate variables is 58.3 %. The la rge SRP error may be due to the relatively low SRP measured values. Another skill assessment, relative operating characteristic (ROC) score, shows that th e average ROC score of water quality state variables is 0.565. This score suggested that the water quality mode l using the estimated parameters is skillful for reproducing th e IRL ecological dynamics Each of the ROC score of water quality state variables is 0.721, 0.368, 0.587, 0.382, 0.780, 0.611, 0.514, and 0.558 for DO, chlorophyll a TOC, NH4, dissolved TKN, SRP, particulate nitrogen, and particulate phosphorous, respectively.

PAGE 298

275 Another episodic simulation was conducted from February 19th to 22nd, 1999 using parameters obtained by trial-a nd-error method. All model vari ables, initial conditions, and model coefficients are the same as th e previous simulation except for water quality model parameters (Table K-7). Figures K-11 to K-13 show the model predictions and measured data. The relative errors of DO, chlorophyll a TOC, NH4, dissolved TKN, SRP, particulate nitrogen, a nd particulate phosphorous are 8.7, 24.7, 42.9, 59.1, 24.2, 200.1, 31.5, and 84.3 %, respectively. The av erage relative error of these water quality state variables is 60.5 %. Th e ROC score of DO, chlorophyll a TOC, NH4, dissolved TKN, SRP, particulate nitr ogen, and particulate phosphorous prove 0.566, 0.367, 0.587, 0.489, 0.793, 0.618, 0.503, and 0.522, respectively. The average ROC score is 0.555. Table K-7. Values of parameters from modified Gauss-Newton method and trial-anderror method Name Trial-and-error method Gauss-Newton method ( )max 1.80 E00 (1/day) 1.90 E00 (1/day) Kax 8.00 E-02 (1/day) 1.00 E-01 (1/day) Kas 6.00 E-02 (1/day) 7.99 E-02 (1/day) Hp 5.00 E00 ( g/l) 1.00 E00 ( g/l) HN 3.00 E01 ( g/l) 1.99 E01 ( g/l) KD 5.00 E-02 (1/day) 7.15 E-02 (1/day) KNN 5.00 E-02 (1/day) 9.07 E-02 (1/day) KDN 0.00 E00(1/day) 0.00 E00 (1/day) Konm 1.00 E-02 (1/day) 1.99 E-02 (1/day) Kopm 1.00 E-02 (1/day) 1.01 E-02 (1/day) don 5.00 E-03 (1/day) 9.0 E-03 (1/day) pon 2.00 E-04 1.20 E-03 dan 5.00 E-03 (1/day) 1.02 E-02 (1/day) pan 5.00 E-04 1.00 E-06 dop 1.00 E-02 (1/day) 9.96 E-03 (1/day) pop 1.00 E-04 1.02 E-04 dip 1.00 E-02 (1/day) 1.00 E-02 (1/day) pip 1.00 E-04 5.41 E-04 (z)max 1.40 E-01 1.18 E-01 (1/day) Kzx 3.0 E-02 2.99 E-02 (1/day) Kzs 2.0 E-02 2.99 E-02 (1/day)

PAGE 299

276 JulianDay1999 Chlorophylla(ug/l) 51 52 53 0 2 4 6 8 10 JulianDay1999 NH4(ug/l) 51 52 53 0 20 40 60 80 JulianDay1999 SRP(ug/l) 51 52 53 0 5 10 15 20 25 30 JulianDay1999 DTKN(ug/l) 51 52 53 0 200 400 600 800 1000 1200 JulianDay1999 Oxygen(mg/l) 51 52 53 0 2 4 6 8 10 JulianDay1999 TOC(mg/l) 51 52 53 0 5 10 15 20 JulianDay1999 POP(ug/l) 51 52 53 0 50 100 150 JulianDay1999 PON(u/l) 51 52 53 0 100 200 300 400 Figure K-11. Comparison of measured and simulated (using water quality parameters obtained by trial-and-erro r method) water quality state variables at north station JulianDay1999 Chlorophylla(ug/l) 51 52 53 0 2 4 6 8 10 JulianDay1999 NH4(ug/l) 51 52 53 0 20 40 60 80 JulianDay1999 SRP(ug/l) 51 52 53 0 5 10 15 20 25 30 JulianDay1999 DTKN(ug/l) 51 52 53 0 200 400 600 800 1000 1200 JulianDay1999 Oxygen(mg/l) 51 52 53 0 2 4 6 8 10 JulianDay1999 TOC(mg/l) 51 52 53 0 5 10 15 20 JulianDay1999 POP(ug/l) 51 52 53 0 50 100 150 JulianDay1999 PON(u/l) 51 52 53 0 100 200 300 400 Figure K-12. Comparison of measured and simulated (using water quality parameters obtained by trial-and-error method) wate r quality state variables at central station

PAGE 300

277 JulianDay1999 Chlorophylla(ug/l) 51 52 53 0 2 4 6 8 10 JulianDay1999 NH4(ug/l) 51 52 53 0 20 40 60 80 JulianDay1999 SRP(ug/l) 51 52 53 0 5 10 15 20 25 30 JulianDay1999 DTKN(ug/l) 51 52 53 0 200 400 600 800 1000 1200 JulianDay1999 Oxygen(mg/l) 51 52 53 0 2 4 6 8 10 JulianDay1999 TOC(mg/l) 51 52 53 0 5 10 15 20 JulianDay1999 POP(ug/l) 51 52 53 0 50 100 150 JulianDay1999 PON(u/l) 51 52 53 0 100 200 300 400 Figure K-13. Comparison of measured and simulated (using water quality parameters obtained by trial-and-erro r method) water quality state variables at south station The average relative error and ROC score from the model results with parameters, which are calculated by the parameter estima tion method, are 58.3 % and 0.565. In the case obtained by trial-an d-error process, the average rela tive error and ROC scores are 60.5% and 0.555. The results of two skill assessm ents indicate that the model predictions with parameters from the modified GaussNewton method better performed than those with current IRL simulation parameters obt ained by trial-and-error process. This suggested that the parameter estimation met hod can be used to determine the optimum parameters as well as improve the efficiency in estimating model parameters in water quality models.

PAGE 301

278 APPENDIX L KINETIC TERM EFFECTS IN THE WATER QUALITY MODEL Kinetic equations in the water quality model are considered to play a major role in the dynamics of water quality state variables. In order to demons trate the effects of kinetic terms in the water qua lity model, the IRL long-term simulation, which has no kinetic terms, was conducted. Figures L-1, L-3, and L-5 showed the simulated results without kinetic terms, and Figur es L-2, L-4, and L-6 displaye d the simulated results with kinetic terms at IRLV17, IRLB 09, and IRLIRJ09, respectively. Each figure embodies ten different time series plots of wate r quality species (oxygen, chlorophyll a NO3, SPR, dissolved silica, TSS, TOC, dissolved T KN, particulate nitrogen, and particulate phosphorous). The rectangular shap e symbolizes measured wate r quality state variables. On the other side, the solid line represents the simulated water quality state variables. The model results, which have no kinetic terms, illustrated insignificant seasonalvariations of water quality state variables. The simulation considered horizontal/vertical mixing and river/open boundary nutrient loads. Thus, the values of the simulated water quality state variables are within the range of measured values. However, the simulation results didnt capture the dynamics of wate r quality state variables. Conversely, the model results with kinetic te rm process obviously demonstrated temporal variations of water quality state variables. This suggested that the kinetic terms in the water quality model play a major role in the dynamics of water quality state variables.

PAGE 302

279 DAY Chlorophylla(ug/l) 400 600 800 0 5 10 15 20 DAY NO3(ug/l) 400 600 800 0 20 40 60 DAY DissovedSilica(ug/l) 400 600 800 0 1000 2000 3000 4000 5000 DAY TSS(mg/l) 400 600 800 20 40 60 80 100 DAY TOC(mg/l) 400 600 800 10 20 30 DAY DissolvedTKN(ug/l) 400 600 800 0 500 1000 1500 2000 DAY ParticulateNitrogen(ug/l) 400 600 800 0 200 400 600 800 1000 DAY ParticulatePhos.(ug/l) 400 600 800 0 100 200 300 DAY DO(mg/l) 400 600 800 5 10StationIRLV17 DAY SRP(ug/l) 400 600 800 20 40 Figure L-1. Temporal wate r quality variations at IR LV17 station from 1997 to 1999 including water quality ki netic terms (A: di ssolved oxygen, B: total suspended sediment, C: chlorophyll a D: total organic carbon, E: nitrate-nitrite, F: dissolved TKN, G: soluble reactive phos phorous, H: particulate nitrogen, I: dissolved silica, and J: particulate phosphorous) A B C D E F G H I J

PAGE 303

280 DAY Chlorophylla(ug/l) 400 600 800 0 5 10 15 20 DAY NO3(ug/l) 400 600 800 0 20 40 60 DAY DissovedSilica(ug/l) 400 600 800 0 1000 2000 3000 4000 5000 DAY TSS(mg/l) 400 600 800 20 40 60 80 100 DAY TOC(mg/l) 400 600 800 10 20 30 DAY DissolvedTKN(ug/l) 400 600 800 0 500 1000 1500 2000 DAY ParticulateNitrogen(ug/l) 400 600 800 0 200 400 600 800 1000 DAY ParticulatePhos.(ug/l) 400 600 800 0 100 200 300 DAY DO(mg/l) 400 600 800 5 10StationIRLV17 DAY SRP(ug/l) 400 600 800 20 40 Figure L-2. Temporal wate r quality variations at IR LV17 station from 1997 to 1999 without water quality kinetic terms (A : dissolved oxygen, B: total suspended sediment, C: chlorophyll a D: total organic carbon, E: nitrate-nitrite, F: dissolved TKN, G: soluble reactive phos phorous, H: particulate nitrogen, I: dissolved silica, and J: particulate phosphorous) B A C D E F G H I J

PAGE 304

281 DAY Chlorophylla(ug/l) 400 600 800 0 5 10 15 20 DAY NO3(ug/l) 400 600 800 0 20 40 60 DAY DissovedSilica(ug/l) 400 600 800 0 1000 2000 3000 4000 5000 DAY TSS(mg/l) 400 600 800 20 40 60 80 100 DAY TOC(mg/l) 400 600 800 10 20 30 DAY DissolvedTKN(ug/l) 400 600 800 0 500 1000 1500 2000 DAY ParticulateNitrogen(ug/l) 400 600 800 0 200 400 600 800 1000 DAY ParticulatePhos.(ug/l) 400 600 800 0 100 200 300 DAY DO(mg/l) 400 600 800 5 10StationIRLB09 DAY SRP(ug/l) 400 600 800 20 40 Figure L-3. Temporal wate r quality variations at IR LB09 station from 1997 to 1999 including water quality ki netic terms (A: di ssolved oxygen, B: total suspended sediment, C: chlorophyll a D: total organic carbon, E: nitrate-nitrite, F: dissolved TKN, G: soluble reactive phos phorous, H: particulate nitrogen, I: dissolved silica, and J: particulate phosphorous) A B C D E F G H I J

PAGE 305

282 DAY Chlorophylla(ug/l) 400 600 800 0 5 10 15 20 DAY NO3(ug/l) 400 600 800 0 20 40 60 DAY DissovedSilica(ug/l) 400 600 800 0 1000 2000 3000 4000 5000 DAY TSS(mg/l) 400 600 800 20 40 60 80 100 DAY TOC(mg/l) 400 600 800 10 20 30 DAY DissolvedTKN(ug/l) 400 600 800 0 500 1000 1500 2000 DAY ParticulateNitrogen(ug/l) 400 600 800 0 200 400 600 800 1000 DAY ParticulatePhos.(ug/l) 400 600 800 0 100 200 300 DAY DO(mg/l) 400 600 800 5 10StationIRLB09 DAY SRP(ug/l) 400 600 800 20 40 Figure L-4. Temporal wate r quality variations at IR LB09 station from 1997 to 1999 without water quality kinetic terms (A : dissolved oxygen, B: total suspended sediment, C: chlorophyll a D: total organic carbon, E: nitrate-nitrite, F: dissolved TKN, G: soluble reactive phos phorous, H: particulate nitrogen, I: dissolved silica, and J: particulate phosphorous) B C A D E F G H I J

PAGE 306

283 DAY Chlorophylla(ug/l) 400 600 800 0 5 10 15 20 DAY SRP(ug/l) 400 600 800 20 40 60 80 100 DAY NO3(ug/l) 400 600 800 0 20 40 60 DAY DissovedSilica(ug/l) 400 600 800 0 1000 2000 3000 4000 5000 DAY TSS(mg/l) 400 600 800 20 40 60 80 100 DAY TOC(mg/l) 400 600 800 10 20 30 DAY DissolvedTKN(ug/l) 400 600 800 0 500 1000 1500 2000 DAY ParticulateNitrogen(ug/l) 400 600 800 0 200 400 600 800 1000 DAY ParticulatePhos.(ug/l) 400 600 800 0 100 200 300 DAY DO(mg/l) 400 600 800 5 10StationIRLIRJ01 Figure L-5. Temporal wate r quality variations at IR LIRJ01 station from 1997 to 1999 including water quality kinetic terms (A: dissolved oxygen, B: total suspended sediment, C: chlorophyll a D: total organic carbon, E: nitrate-nitrite, F: dissolved TKN, G: soluble reactive phos phorous, H: particulate nitrogen, I: dissolved silica, and J: particulate phosphorous) B C A D E F G H I J

PAGE 307

284 DAY Chlorophylla(ug/l) 400 600 800 0 5 10 15 20 DAY SRP(ug/l) 400 600 800 20 40 60 80 100 DAY NO3(ug/l) 400 600 800 0 20 40 60 DAY DissovedSilica(ug/l) 400 600 800 0 1000 2000 3000 4000 5000 DAY TSS(mg/l) 400 600 800 20 40 60 80 100 DAY TOC(mg/l) 400 600 800 10 20 30 DAY DissolvedTKN(ug/l) 400 600 800 0 500 1000 1500 2000 DAY ParticulateNitrogen(ug/l) 400 600 800 0 200 400 600 800 1000 DAY ParticulatePhos.(ug/l) 400 600 800 0 100 200 300 DAY DO(mg/l) 400 600 800 5 10StationIRLIRJ01 Figure L-6. Temporal wate r quality variations at IR LIRJ01 station from 1997 to 1999 without water quality kinetic terms (A : dissolved oxygen, B: total suspended sediment, C: chlorophyll a D: total organic carbon, E: nitrate-nitrite, F: dissolved TKN, G: soluble reactive phos phorous, H: particulate nitrogen, I: dissolved silica, and J: particulate phosphorous) B C A D E F G H I J

PAGE 308

285 APPENDIX M NEAR BOTTOM DISSOLVED OXYGEN CONCENTRATI ON IN THE UPPER CHARLOTTE HARBOR The following figures illustrate the botto m hypoxia areas in the Upper Charlotte Harbor during summer time in 2000. Figures M-1 and M-2 show the simulated nearbottom dissolved oxygen concentration in th e Upper Charlotte Harbor on July 3rd and July 12th, 2000, respectively. Figur e M-3 depicts the simulate d near-bottom dissolved oxygen concentration below 2mg/l in th e Upper Charlotte Harbor on July 12th, 2000. Easting(UTM,meter) Northing(UTM,meter) 375000 380000 385000 390000 395000 400000 2.97E+06 2.98E+06 2.99E+068 7.5 7 6.5 6 5.5 5 4.5 4 3.5 3 2.5 2 1.5 Time=185.5417day DO(mg/l) Figure M-1. Simulated near-bottom disso lved oxygen concentr ation in the Upper Charlotte Harbor on July 3, 2000

PAGE 309

286 Easting(UTM,meter) Northing(UTM,meter) 375000 380000 385000 390000 395000 400000 2.97E+06 2.98E+06 2.99E+068 7.5 7 6.5 6 5.5 5 4.5 4 3.5 3 2.5 2 1.5 Time=194.2917day DO(mg/l) Figure M-2. Simulated near-bottom disso lved oxygen concentr ation in the Upper Charlotte Harbor on July 12, 2000 Easting(UTM,meter) Northing(UTM,meter) 375000 380000 385000 390000 395000 400000 2.97E+06 2.98E+06 2.99E+06Time=194.2917day Figure M-3. Simulated near-bottom dissolved oxygen concentration below 2mg/l in the Upper Charlotte Harbor on July 12, 2000

PAGE 310

287 APPENDIX N CORRELATION OF SALINITY DIFFERENCES BETWEEN BOTTOM AND SURF ACE LAYERS The following figures show the correlation of salinity difference between the near bottom and near surface layers in terms of measured data and simulated prediction at UF, Punta Gorda, and El Jobean stations. In order to illustrate the large salinity differences at each station, this study selected the following periods: June 19 25, 2003 (UF); August 5 9, 2003 (Punta Gorda); and February 13 18, 2004 (El Jobean). As can be seen in Figures N-1 and N-2, the mode l well simulated the vertical stratification at the UF and Punta Gorda stations. In the El Jobean stati on, which is near the mouth of the Myakka River, measured data and model predictions re presented very weak ve rtical stratification during the simulation period. Ther efore, the correlation value in the El Jobean station is not as good as those of UF and Punta Gorda stations. Measured(psu) Simulated(psu) 0 5 10 15 20 0 5 10 15 20r=0.807 Figure N-1. Correlation of sa linity difference between the ne ar bottom and near surface layers in terms of measured data an d simulated results at the UF station (Diagonal line is the perfect match line)

PAGE 311

288 Measured(psu) Simulated(psu) 0 5 10 0 5 10r=0.655 Figure N-2. Correlation of sa linity difference between the ne ar bottom and near surface layers in terms of measured data an d simulated results at the Punta Gorda station (Diagonal line is the perfect match line) Measured(psu) Simulated(psu) 0 1 2 3 4 5 6 0 1 2 3 4 5 6r=0.163 Figure N-3. Correlation of sa linity difference between the ne ar bottom and near surface layers in terms of measured data and si mulated results at the El Jobean station (Diagonal line is the perfect match line)

PAGE 312

289 APPENDIX O TEMPORAL SOD VALUES IN TERM S OF DIFFERE NT SCENARIOS The following figures illustrate temporal SOD values at stations CH005 and CH009 in the Upper Charlotte Harbor in terms of different scenarios during the year 2000: baseline; 100 river nutrient load reduction; 100 river nutrient load reduction and zero algal mortality; and 100 % river nutrient load reduction and zero algal mortality and 100 % nutrient load reductio n at open boundary. JulianDay SOD(gO2/m2/d) 274 275 276 277 0 0.5 1 1.5 2 CH009Baseline 100%Reduction 100%Reduction+zeromortality JulianDay SOD(gO2/m2/d) 274 275 276 277 0.5 1 1.5 CH005Baseline 100%Reduction 100%Reduction+zeromortality Figure O-1. Temporal SOD values at CH005 and CH009 st ations in the Upper Charlotte Harbor in terms of different scenario s (baseline; 100 rive r nutrient load reduction; and 100 % river nutrient load reduction and zero algal mortality)

PAGE 313

290 JulianDay SOD(gO2/m2/d) 274 274.5 275 275.5 276 276.5 277 0.5 1 1.5 CH005Baseline 100%Reduction+zeromortality 100%Reduction+zeromortality+noopenboundarynutrient JulianDay SOD(gO2/m2/d) 274 274.5 275 275.5 276 276.5 277 0 0.5 1 1.5 2 CH009Baseline 100%Reduction+zeromortality 100%Reduction+zeromortality+noopenboundarynutrient Figure O-2. Tempor al SOD variations at CH005 a nd CH009 in the Upper Charlotte Harbor in terms of different scenario s (baseline; 100 rive r nutrient load reduction and zero algal mo rtality; and 100 % river nutrient load reduction and zero algal mortality and 100 % nutr ient load reduction at open boundary)

PAGE 314

291 LIST OF REFERENCES Adkins, W .A., Mao, M., Taylor, M., Green, W ., Basci, C., Bergman, M., and Smith, D. (2004). Watershed model developm ent for the Indian River Lagoon basin: providing simulated runoff and pollution load to the In dian River Lagoon pollution load reduction model Tech. Mem. 50, St. Johns River Ma nagement District, Palatka, FL. Alber, M. (2002). A conceptual model of estuarine freshwater inflow management. Estuaries, 25(6B), 1246-1261. Ambrose, R.A., Wool, T.A., Martin, J.L., Connolly, J.P., and Schanz, R.W. (1994). WASP5, a hydrodynamic and water quality m odel: model theory users manual, and programmers guide EPA/600/3-87/039, U.S.Environmen tal Protection Agency, Athens, GA. American Public Health Association (APHA) (1985). Standard methods for the examination of water and wastewate r, APHA, Washington, D.C. Badylak, S., and Phlips, J. (2004). Spatia l and temporal patt erns of phytoplankton composition in a subtropical coastal lagoon, the Indian River Lagoon, Florida, USA. Journal of Plankton Research 26(10), 1229-1247. Banks, R.B., and Herrera, F.F. (1977). Effect of wind and rain on surface reaeration. Journal of Environmenta l Engineering division 103(EE3), 489-504. Bard, Y. (1970). Comparison of gradient methods for the so lution of nonlinear parameter estimation problems. SIAM J. Numerical Analysis 7, 157-186. Barile, D., Panico, C.A ., and Heath, G. (1987). Management plan and implementation strategy for the Indian River Lagoon systems Technical Report, Marine Resources Council, Florida Institute of Technol ogy. Sea Grant Contract IRL IR-1. Barnes, A.F., and Goldberg, J.H. (1978). M ethan production and c onsumption in anoxic marine sediments. Geology 4, 297-300. Basulto J., Morcillo E., Rubio, E., and Espl ugues, J. (1978). Applica tion of the modified Gauss-Newton method and elimination of linear parameters to drug-re ceptor interactions of n-order kinetic. Arch Farmacol Toxicol. 4(3), 313-318. Bellman, Richard (1967). Quasilinearizati on and the estimation of chemical rate constants from raw kinetic data Mathematical Biosciences 1, 71-76.

PAGE 315

292 Belohlav, Z., Zamostny, P., Kluson, P., a nd Volf, J. (1997). Application of randomsearch algorithm for regression anal ysis of catalytic hydrogenations. The Canadian Journal of Chemical Engineering 75, 735-742. 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 formulati ons in surface water quality modeling In: (Editors), U.S. Environmental Protection Agency, Office of Research and Development, Athens, GA. 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 26(1), 43-53. Bunch, B.W., Cerco, C.F., Dortch, M. S., Johnson, B.H., and Kim, K. (2000). Hydrodynamic and water quality model study of San Juan bay estuary ERDC TR-00-1, U.S. Army Engineer District, Jacksonville, FL. Camp Dresser & McKee (CDM). (1998). The study of seasonal and spatial patterns of hypoxia in Upper Charlotte Harbor Prepared by Camp Dresser & McKee Inc, For Southwest Florida Water Management Dist rict, Venice, Florida. Under SWFWMD Contract 95CON000123. Canale, R.P., and Vogel, A.H. (1974). Eff ects of temperature on phytoplankton growth. Journal of the Environmental Engineering Division 100(EE1), 231-241. Cerco, C., and Bunch, B. (1996). Passaic river tunnel divers ion model study; report 5, water quality modeling Technical Report HL-96-2, US Army Engineer Waterways Experiment Station, Vicksburg, MS. Cerco, C., Bunch, B., Cial one, M., and Wang, H. (1994). Hydrodynamic and eutrophication model study of In dian River and Rehoboth bay, Delaware Technical Report EL-94-5, US Army Engineer Waterway s Experiment Station, Vicksburg, MS. Cerco, C.F., Bunch, B.W., Teeter, A.M., and Dortch, M.S. (2000). Water quality model of Florida Bay. ERDC/EL TR-00-10, U.S. Army Engi neer District, Jacksonville, FL. Cerco, C.F., and Cole, T.M. (1994). Three-dimensional eutr ophication model of Chesapeake Bay Technical Report El-94-4, U.S. Ar my Engineer Waterways Experiment Station, Vicksburg, M.S. Cerco, C.F., and Cole, T.M. (1995). Users guide to the CE-QUAL-ICM three dimensional eutrophication model, Technical Report El-95-15, U.S. Army Engineer Waterways Experiment Station, Vicksburg, M.S. Chapra, S.C. (1997). Surface water quality modeling McGraw-Hill, New York.

PAGE 316

293 Charlotte Harbor National Estuary Program (2002). 2002 implementation review report, Charlotte Harbor National Estuary Program Technical Report 02-2, June 7, 2002. Chen, X. and Sheng, Y.P. (1995). Recent studies on the effects of hydrodynamics and sediment dynamics on nutrient transport in shallo w lakes and estuaries Coastal and Oceanographic Engineering Department Tec hnical Report, University of Florida. Choi, Jei-Kook (1992). Three-dimensional curvilinear grid modeling of baroclinic circulation and mixing in a partia lly-mixed estuary, Florida Ph D dissertation, University of Florida, Gainesville, Florida Christian, D.J. (2001). Modeling the effects of hydrodynamics, suspended sediments, and water quality on light attenuation in Indian River Lagoon, Florida Master Thesis, University of Florida, Gainesville, Florida. Christian, D., and Sheng, Y. (2003). Relative influence of various water quality parameters on light attenuati on in Indian River Lagoon. Estuarine, Coastal and Shelf Science, 57, 961-971. Crean, D. (2001). Preliminary water quality targets for the southern Indian River Lagoon, Technical document, South Florida Wa ter Management District, West Palm Beach, Florida. Davis, J.R. (2001). High performance modeling of circul ation and transport in the Indian River Lagoon, Florida Ph D 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., A nderson, 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 th e sea (HEXOS) results. Journal of Geophysical Research 101, 12001-12016. Dennison, W.C., Orth, R.J., Moore, K.A., Stevenson, J.C., Carter, V., Kollar, S., Bergstrom, P.W., and Batiuk, R.A. (1993). Assessing water quali ty with submerged aquatic vegetation. Bioscience, 43, 86-94. DiToro, D.M. (2000). Sediment flux modeling Wiley-Interscience. DiToro, D.M., and Connoly, J.F. (1980). Mathematical models of water quality in large lakes, part II: Lake Erie U.S. EPA Ecological Res earch Series. EPA-600/3-80065.Athens, GA.

PAGE 317

294 DiToro, D.M, Fitzpatrick, J.J., Thoma nn, R.V., and Hydroscience, Inc. (1983). Documentation for water quality analysis simulation Program (WASP) and model verification program (MVP). U.S. EPA Ecological Resear ch Series. EPA-600/3-81-044. Dixon, L.K., and Kirkpatrick, G.J. (1999). Causes 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. Dixon, L.K., Nissanka, A., Perry, J., and Toutant, T. (2003). Observations on water column oxygen demand and hypoxia in Ch arlotte Harbor, Florida, 2002-2003. Charlotte Harbor Conference AFocus in 2002-2003" Program, Mote Marine Laboratory Doering, P.H., Chamberlain, R.H., and Ha unert, D.E. (2002). Using submerged aquatic vegetation to establish minimum and maximum freshwater inflows to the Caloosahatchee estuary, Florida. Estuaries 25(6B), 1343-1354. Echternacht, K.L. (1975). A study of the precipitation regi mes of the Kissimmee River Lake Okeechobee watershed Florida Department of Envi ronmental Regulation Technical Service (3). Tallahassee, Florida. Fillery, I.R.P., and DeDatta, S.K. (1986). Ammonia volatilization from nitrogen sources applied to rice fields: I. methodology, ammonia fluxes and nitrogen-15 loss. Soil Science, 50, 80-86. Florida Sea Grant (2006). Publications Library < http://www.flseagrant.org > (Decem ber, 1, 2006) Fong, P., and Harwell, M.A. (1994). Mode ling seagrass communities in tropical and subtropical bays and estuarie s: A mathematical model synt hesis of current hypotheses. Bulletin of Marine Science 54(3), 757-781. Freney, J.R., Simpson, J.R., and Denmead O.T. (1981). Ammonia Volatilization. Ecological Bulletin 33:291-302. Friehe, C.A., and Schmitt, K.B. (1976). Param eterization of air-s ea interface fluxes of sensible heat and moisture by the bulk aerodynamic formulas. Journal of Physical Oceanography 6, 801-809. Froelich, P.N., Kaul, L.W., Byrd, J.T., Andreae, M.O., and Roe, K.K. (1985). Arsenic, barium, germanium, tin, dimethysulfide and nutrient biogeochemistry in Charlotte Harbor, Florida, A phosphor us-enriched estuary. Estuarine, Coastal and Shelf Science 20, 239-264. 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 Florid a Water Management District.

PAGE 318

295 Goldman, J.C., and Carpenter, E.L. (1974). A kinetic approach to the effect of temperature on algal growth. Limnology and Oceanography, 19(5), 756-766. Goodwin, C.R. (1996). Simulation of tidal-flow, circulati on, and flushing of the Charlotte Harbor estuarine system, Florida U.S. Geological Survey Water-Resources Investigations Report 93-4153. Hammett, K.M. (1990). Land use, water use, streamflo w characteristics, and water quality characteristics of the Char lotte Harbor inflow area, Florida U.S. Geological Survey Water Supply Paper 2350-A:64. Heyl, M.G. (1998). Hypoxia in upper Charlotte Harbor. Proceedings of the Charlotte Harbor Public Conference and Technical Symposium, Technical Report No.98-02 219228. Hinkle, R., Kilburn, C., and Reddick, R. (1995). Kennedy space center (KSC) ecological resources. < http://www.ksc.nasa.gov > (July, 1, 2007) Huang, T.C. (1966). A sedimentologic study of Charlo tte Harbor, Southwestern Florida Master s Thesis, Florida State University, Tallahassee, Florida. Hubertz, J.M., Krallis, G., Prak ash, S., and Kolluru, V. (2004). Relationship between physical and biochemical processes in an estuary The Seventh International Marine Environmental Modeling Semi nar. Washington DC. USA. Humphryes, J., Franz, S., and Bill, S. (1993). Floridas estuaries: A citizens guide to coastal living and conservation Florida Sea Grant College Program, University of Florida. Sea Grant Extension Bulletin 23. Humphryes, J. and Grantham, S. B. (1995). Floridas estuaries FAMTOM Magazine, winter issue, 7, Florida Sea Grant, < http://flseagrant.org/library/publications/fathom_magazine/volume-7_issue-4 > (July, 1, 2007). HydroQual (1987). A steady-state coupled hydrody namic/water quality m odel of the eutrophicatin and anoxia process in Chesapeake Bay. Final Report to the US EPA Chesapeake Bay Program. HydroQual, Mahmah, NJ. Jin, H., Egashira, S., and Chau, K.W. (1998). Carbon to chlorophylla ratio in modeling long-term eutrophication phenomena. Wat. Sci. Tech., 18(11), 227-235. Jorgensen, S.E. (2001). Parameter estim ation and calibration by use of exergy. Ecological Modelling 191, 299-302.

PAGE 319

296 Kennish, M.J. (1986). Ecology of Estuaries (Volumn I) : Physical and Chemical Aspects CRC Press, Florida. Kim, D., Matsuyama, Y., Nagasoe, S., Yamaguc hi, M., Yoon, Y., Oshima, Y., Imada, N., and Honjo, T. (2004). Effects of temperatur e, salinity and irradi ance on the growth of the harmful red tide dinoflagellate cochl odinium polykrikoides margalef (dinophyceae). Journal of Plankton Research 26(1), 61-66. Kim, S.C., Shen, J., Kim, C.S., and Kuo, A. Y. (2000). Application of VIMS HEM-3D to a macro-tidal environment Estuarine and Coastal Modeling, 6, 238-249 Kimmerer, W.J. (2002). Physic al, biological, and management responses to variable freshwater flow into the San Francisco estuary. Estuaries, 25(6B), 1275-1290. Kirk, J.T.O. (1984). Dependence of relations hip between inherent and apparent optical properties of water on solar altitude. Limnology and Oceanography, 29(2), 350-356. Kirkpatrick, G.J., Berg, B., Higham, C ., Ordonez, D., and Pederson, B. (2003). Phytoplankton community composition associ ated with hydrographic variability during July, 2003. Charlotte Harbor Conference AFocus in 2002-2003" Program, Mote Marine Laboratory 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. Leverone, J.R., and Culter J.K. (2003). Ben thic infaunal dynamics in upper Charlotte Harbor. temporal response to hypoxia during 2003. Charlotte Harbor Conference AFocus in 2002-2003" Program, Mote Marine Laboratory Li, M., Gargett, A., and Denman, K. (2000). What determines seas onal and interannual variablility of phytpoplankton and zoopla nkton in strongly estuarine systems? Estuarine, Coastal and Shelf Science 50, 467-488. Linga, P., Al-Saifi, N., and Englezos, P. (2006). Comparison of the Luus-Jaakola optimization and Gauss-Newton methods fo r parameter estimation in ordinary differential equation models. Ind. Eng. Chem. Res. 45, 4716-4725. Luss, R. (2001). Iterative dynamic programming Chapman & Hall/CRC, New York. Mason, S.J., and Graham, N.E. (1999). Conditional probabilitie s, relative operating characteristics, and relative operating levels. American Meteorological Society 14, 713725. McPherson, B.F. and Miller, R.L. (1994). Causes of light attenuation in Tampa Bay and Charlotte Harbor, Southwestern Florida. Water Resource Bulletin 30, 43-53.

PAGE 320

297 McPherson, B.F., Miller, R.L., and Stoker, Y.E. (1996). Physical, chemical, and biological characteristics of the Charlott e Harbor basin and estuarine system in Southwestern FloridaA summary of the 1982-1989 U.S. geological survey Charlotte Harbor assessment and other studies, U.S. Geological Survey Water-Supply Paper 2486. Mellor, G.L. (1996). Introduction to phy sical oceanography, Springer, New York. Metz, Charles E. (1978). Basic principles of ROC analysis. Seminars in Nuclear Medicine, 8(4), 283-298. Miller R.L. and McPherson B.J. (1995). Mode ling photosynthetically active radiation in water of Tampa Bay, Florida, with emphasi s on the geometry of incident irradiance. Estuarine, Coastal and Shelf Science 40, 359-377. Montgomery, R.T., McPherson, B.F., and Emmons, E.E. (1991). Effects of nitrogen and phosphorus additions on phytoplankton producti vity and chlorophyll a in a subtropical estuary, Charlotte Harbor, Florida U.S. Geological Survey Water Resources Investigation Report 91-4077. Florida Depa rtment of Environmental Regulation, Tallahassee, Florida. Montgomery, R., and Stone, S. (2003). Di fferences in long-term Peace River water quality data. Charlotte Harbor Conference AFocus in 2002-2003" Program, Mote Marine Laboratory. 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, 2227-2238. Morris, Ian. (1980). The physiological ecolo gy of phytoplankton University of California Press. Morse, J.W., Miller, F.J., Cornwell, J.C., and Rickard, D. (1987). The chemistry of the hydrogen sulfide and iron sulfide systems in natural waters. Earth Science Reviews 24, 1-42. OConnor, D.J. and Dobbins, W.E. (1958). Mechanism of reaeration in natural streams. ASCE Transactions, Paper No. 2934 641-684. Park, Kijin. (2004). Modeling the circulation and wate r quality in Charlotte Harbor estuarine system, Florida Ph.D. dissertation, University of Florida, Gainesville, Florida. Park, K., Jung, H., Kim, H., and Ahn, S. (2005). Three-dime nsional hydrodynamiceutrophication model (HEM -3D): application to Kwang-Yang Bay, Korea. Marine environmental research 60(2), 171-193.

PAGE 321

298 Park, K., Kuo, A.Y., and Neilson, B.J. ( 1996) A numerical mode l study of hypoxia in the tidal Rappahannock River of Chesapeake Bay Estuarine, Coasta l and Shelf Science 42(5), 563-581. Park, K., Kuo, A.Y., Shen, Jian, and Hamrick, J.M. (1995). A three-dimensional hydrodynamic-eutrophication model (HEM-3 D): description of water quality and sediment process submodels. Special Report in Applied Marine Science and Ocean Engineering No.327. School of Marine Scien ce Virginia Institute of Marine Science, College of William and Mary, Gloucester Point, VA 23062. Penman, H.L. (1956). Evaporation; an introductory surv ey, Netherland. Journal of Agricultural Science, 4, 9-29. Phlips, E.J. (1999). Nutrient limitation of phytoplankton communities in the Indian River Lagoon, Final Report to the St. Johns River Wate r Management Distri ct, University of Florida. Phlips, E.J., Badylak, S., Youn, S., and Kelle y, K. (2004). The o ccurrence of potentially toxic dinoflagellates and diatoms in a s ubtrophical lagoon, the Indian River Lagoon, Florida, USA. Hamful Algae, 3, 39-49. Pickard, J.R., and Emery, W.J. (1990). Descriptive Physical Oceanography Pergamon Press, New York. Fifth (SI) English Edition: 320. Pomeroy, L.R., Shenton, L.R., Jones, R.D.H ., and Reimold, R.J. (1972). Nutrient flux in estuaries. In: Likens, G.E. (ed.), nutri ents and eutrophication: the limiting-nutrient controversy. American Society of Limnology and Oceanography, Special Symposium 1. Lawrence, Kansas, 274-191. Pontius Jr., R.G., and Schneider, L.C. (2001) Land-cover change model validation by an ROC method for the Ipswich wa tershed, Massachusetts, USA. Agriculture, Ecosystems and Environmen t, 85, 239-248. Powell, G.L., Matsumoto, J., and Brock, D. (2002). Methods for determining minimum freshwater inflow needs of Texas bays and estuaries. Estuaries 25(6B), 1262-1274. Pribble, J.R., Wade, D.L., Squires, A.D., a nd 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 242258. Rao, P.S.C., Jessup, R.E., and Reddy, K.R. (1984). Simulation of nitrogen dynamics in flooded soils. Soil Science 138, 54-62. Reddy, K.R., Fisher, M.M., Pant, H., Inglett, P., and White, J.R. (2001). Indian River Lagoon hydrodynamics and water quality model: nutrient storage and transformations in

PAGE 322

299 sediments Final Report to the St. Johns River Water Management District, Soil and Water Science Department, University of Florida, Gainesville, Florida. Reed, R.K. (1976). On estimation of net long-wave radiation from the oceans. Journal of Geophysical Research 81, 5793-5794. Reed, R.K. (1977). On estimation insolation over the ocean. Journal of Physical Oceanography 7, 482-485. Ross, M. A., Said, A., Trout, K., and Zhang, J. (2005). Hydrologic modeling of streamflow from ungauged areas in th e Upper Charlotte Harbor basin Report of the University of South Florida Center for M odeling and Hydrologic and Aquatic Systems, University of South Florida, submitted to the Southwest Florida Water Management District. Ryther J., and Dunstan W.M. (1971). Nitr ogen, phosphorous and eu trophication in the coastal marine environment. Science 171(3975), 1008 1013. Sawyer, C.N., McCarty, P.L., and Parkin, G.F. (2003). Chemistry for Environmental Engineering and Science. 5th edition McGraw-Hill, Inc. New York, NY. Schropp, S.J. (1998). Charlotte Harbor sedi ment quality data review and evaluation. Proceedings of the Charlotte Harbor Pub lic Conference and Technical Symposium, Technical report No. 98-02, 35-47. Shen, J., Sisson, G.M., Kuo, A.Y., Boon, J. D., and Kim, S.C. ( 1998). Three-dimensional numerical modeling of the tidal Yo rk River system, Virginia. Estuarine and Coastal Modeling 5, 495-510. Shen, J., Boon, J.D., and Kuo, A.Y. (1999). A modeling study of a tidal intrusion front and its impact on larval dispersion in the James River estuary, Virginia. Estuaries, 22(3A), 681-692. Sheng, Y.P. (1983). Mathematical Modeling of ThreeDimensional Coastal Currents and Sediment Dispersion: Model Development and Application Coastal Engineering Research Center, University of Florid a, Fla., Technical Report CERC-TR-83-2 DACW39-80-C-0087. Sheng, Y.P. (1987). On modeling three-dimensional estuarine and marine hydrodynamics. In Nihoul and Jamart, Three-Dime nsional Models of Marine and Estuarine Dynamics, 35-54, Elsevier, Amsterdam. Sheng, Y.P. (2000). A framework for in tegrated modeling of coupled hydrodynamicsedimentary-ecological processes. Estuarine and Coastal Modeling 6, 350-362.

PAGE 323

300 Sheng, Y.P., and Chen, X.J. (1992). A three-dimensional numerical model of hydrodynamics, sediment transport and phosphorous dynamics in Lake Okeechobee: theory, model development, and documentation, spring 1989 Final Report to the South Florida Water Management District, Coastal and Oceanographic Engineering Department, University of Florida, Gainesville, Florida. Sheng, Y.P., Chenxia, Qiu, Park, K.J., and Kim, T. (2003c). Indian River Lagoon Pollutant Load Reduction (IRLPLR) Model Development Project. Final Report Volume III (A 3-D Water Quality Model) to St. Johns River Water Management District. Civil and Coastal Engineering Department, Univer sity of Florida, Gainesville, Florida. Sheng, Y.P. and Davis, J.R. (1999). Measure nutrient concentra tion/ssc/salinity at ocean boundary, task 5.4 of IRLPLR model development Interim report, Coastal and Oceanographic Engineering Depart ment, University of Florid a, Gainesville, Florida. Sheng, Y.P. and Davis, J.R. (2003a). Indian River Lagoon Pollu tant Load Reduction (IRLPLR) Model Development Project. Final Report Volume I (A 3-D Hydrodynamic/Salinity Model) to St. Johns Ri ver Water Management District. Civil and Coastal Engineering Department, University of Florida, Gainesville, Florida. Sheng, Y.P., Davis, J.R., Sun, D., Qiu, C ., Park, K.J., Kim, T., and Zhang, Y. (2001). Application of an integrated modeling system of estuarine and coastal ecosystem to Indian River Lagoon, Florida. Estuarine and Coastal Modeling 4B, 329-343. Sheng, Y.P., Du, H., Kornick, A., Davis, J.R., Sun, D., Qiu, C., Melanson, J., and Christian, D. (1998a). Resuspension of sediments and nut rients in the Indian River Lagoon. Interim report on first episodic event, Civil and Coastal Engineering Department, University of Florida. Sheng, Y.P. and Melanson, J.(1999). IRL-PLR model development (task 5.3-enhanced synoptic water quality sampling) Interim Report #2, Civil and Coastal Engineering Department, University of Florida. Sheng, Y.P., Sun, D.T., Henderson, M., and Melanson, J. (1998b). Sediment map of the Indian River Lagoon Interim Report to the St. Johns River Water Management District, Civil and Coastal Engineering Department, University of Florida, Gainesville, Florida. Sheng, Y.P., Sun, D.T., and Zhang, Y.F. (2003b). Indian River Lagoon pollutant load reduction (IRLPLR) model development project Final Report Volume II (A 3-D Suspended Sediment Transport Model) to St Johns River Water Management District. Civil and Coastal Engineering Department, University of Florida, Gainesville, Florida. Sheng, Y.P.and Vladimir, P. (2003d). Indian River Lagoon pllu tant load reduction (IRLPLR) model development Final Report (Refined Segmen tation of the Indian River Lagoon Estuarine System) to St. Johns River Water Management District. Civil and Coastal Engineering Department, University of Florida, Gainesville, Florida.

PAGE 324

301 Sigua, G.C., Steward, J.S., Miller, J.D., and Tweedale, W.A. (1998). Water Quality Assessment Program in the In dian River Lagoon, Florida: II Redesigning a monitoring network, 1998 Conference Proceedings, National Water Quality Monitoring Council. Sigua, G.C., Steward, J.S., and Tweedale, W.A. (2000). Water-quality monitoring and biological integrity assessment in the IRL, Florida: stat us, trends, and loadings (19881994). Environmental Management 25(2), 199-209. Simon, N.S (1988). Nitrogen cycling between sediment a nd the shallow-water column in the transition zone of the Potomac Rive r and estuary. I. Nitrate and ammonium fluxes. Estuarine, Coastal and Shelf Science 26, 483-497. Simon, N.S (1989). Nitrogen cycling between sediment a nd the shallow-water column in the transition zone of the Potomac River and estuary. II. The role of wind-diven resuspension and adsorbed ammonium. Estuarine, Coastal and Shelf Science 28, 531547. Smayda, T.J., and Reynolds, C.S. (2001). Community assembly in marine phytoplankton: application of recent models to harmful dinoflagellate blooms. Journal of Plankton Research 23(5), 447-461. Smith, N.P. (1987). An introduction to th e tide of Floridas Indian River Lagoon. Florida Scientist 50, 49-61. Smith, N.P. (1993). Tidal and nontidal fl ushing of Floridas Indian River Lagoon. Estuaries, 16(4), 739-746. Smith, R.C., and Baker, K. (1981). Optical properties of the clearest waters. Applied Optics 20(2), 177-184. Smith, S.D. (1989). Coefficients for sea surf ace wind stress, heat flux, and wind profiles as a function of wind speed and temperature. Journal of Geophysical Research 93, 15467-15472. Smithson, M.S. (1999). The Indian River Lagoons problems are as common as dirt. Indian River Lagoon Program Update, winter issue, < http://sjr.state.fl.us/ > (July 1, 2007). Steward, J., Virnstein, R., Ha unert, D., and Lunc, F. (1994). Indian river lagoon surface water improvement and m anagement (SWIM) plan Technical report, South Florida Water Management District and St. J ohns River Water Management District. Steidinger, D.A., Landsberg, J.H., Truby, E.W., and Roberts, B.S. (1998) First report of Gymnodinium pulchellum (Dinophyceae) in North America a nd associated with kills in the Indian River, Florida. Journal of Phycology 34, 431-437.

PAGE 325

302 Stevenson, J.W. and Niiler, P.P. (1983). Upper ocean heat budget during the hawaii to tahiti shuttle experiment. Journal of Physical Oceanography, 13, 1894-1907. Stoker, Y.E. (1992). Salinity distribution and variation with fresh-water inflow and tide, and potential changes in salinity due to altered freshwater infl ow in the Charlotte Harbor Estuarine system, Florida. U.S. Geological Survey Water-R esource Investigations Report 92-4062:30. Sun, Detong (2001). Modeling Suspended Sediment Tr ansport Under Combined Wave Current Action in Indian River Lagoon, Florida Ph.D. dissertation, University of Florida, Gainesville, Florida Sun, N.Z. (1994). Inverse Problems in Groundwater Modeling Kluwer Academic Publishers, Dordrecht, The Netherlands. Swets, John A. (1988). Measuring th e accuracy of diagnostic systems. Science, 240(4857), 1285-1293. Taylor, K.E. (2001). Summarizing multiple as pects of model performance in a single diagram. Journal of Geophysical Research 106, 7183-7192 Tillman, D.H., Cerco, C.F., Noel, M.R ., Martin, J.L., and Hamrick J. (2004). Threedimensional eutrophication model of th e lower St. Johns River, Florida Technical Report ERDC/El TR-04-13, U.S. Army Engineer Wa terways Experiment station, Vicksburg, M.S. Thomann, R.V., and Fitzpatrick, J.F. (1982). Calculation and verification of a mathematical model of the eutroph ication of the Potomac Estuary Prepared for the Department of Environmental Service, Government of the District of Columbia. Thomann, R.V., and Mueller, J.A. (1987). Principles of surface water quality modeling and control Harper & Row, New York. Tomasko, D.A., and Hall-Ruark, M.O. (1998). Abundance and productivity of the seagrass thalassia testudinum along a gradien t of freshwater influence in Charlotte Harbor, Florida (USA) Proceedings of the Charlotte Harbor Public Conference and Technical Symposium, Technical Report No.98-02: pages 111-119. Tomasko, D.A., ANASTASIOU, C., and KOVACH, C. (2006). Dissolved oxygen dynamics in charlotte harbor and its contri buting watershed, in response to hurricanes Charley, Frances, and Jeanne-impacts and recovery. Estuaries and Coasts 29, 932-938. Turner, R.E., Rabalais, N.N., Atilla, N., Normandeau, C., Fry, B. Lee, J.M., Milan, C.S., Oswald, T.A., and Swenson, E.M. (2001). Paleo-reconstruction of water quality in the

PAGE 326

303 Charlotte Harbor estuary (Florida) Final Report to the Southwest Florida Water Management District. Turner, R.E., Rabalais, N.N., Fry, B., Atilla, N., Milan, C.S., Lee, J.M., Normandeau, C., Oswald, T.A., Swenson, E.M., and Tomasko, D.A. (2006). Paleo-indicators and water quality in the Charlotte Harbor estuary (Florida). Limnology and Oceanography, 51, 518-533. van Rijn, L.C. (1994). Handbook, sediment transport by currents and waves, Delft Hydraulics Report H461, Delft Hydr aulics, Delft, The Netherlands. Virnstein, R.W., Ragasekhar, D.S., and Cart er IV, E.W. (1994). Spatial analysis of seagrass changes versus changes in loadi ng to the Indian Rive r Lagoon basin over the past 50 years. Virnstein, R.W. and Morris, L.J. (1996). Seagrass preservation and restoration: a diagnostic plan for the Indian River Lagoon Technical Memorandum 14, St. Johns River Water Management District, Palatka, FL. Walsby, A.E. and Reynolds, C.S. (1980). Si nking and floating, in physiological ecology of phytoplankton (Ed by Morris, L). University of California Press, Berkeley and Los Angeles, 371-412. Wang, H.V., and Johnson, B.J. (2000). Validation and application of the second generation three dimensional hydrod ynamic model of Chesapeake Bay. Water Quality and Ecosystem Modeling 1, 51-90. Wang P. (2004). Bathymetric survey of Charlotte Harbor area Report by the Department of Geology, University of South Florida prepared for the Sout hwest Florida Water Management District. Weast (1977). Handbook of chemistry and physics, 58th edition CRC Press. Whigham, P.A. and Recknagel, F. (2001). Predi cting chlorophylla in freshwater lakes by hybridizing process-based m odels and genetic algorithms. Ecological Modelling, 146, 243-251. Wolanski, E., Aseda, T., and Imberger, J. (1989). Mixing across a lutocline. Limnology and Oceanography 34(5), 931-938. Wolfe, S.H., and Drew, R.D. (1990). An ecological characterization of the Tampa Bay watershed U.S. Fish Wildl. Serv. Biol. Rep. 90:20. Yassuda, E.A. (1996). Integrated modeling of the Tampa Bay estuarine System Ph D dissertation, University of Florida, Gainesville, Florida.

PAGE 327

304 Yassuda, E.A., and Sheng, Y.P. (1997). M odeling dissolved oxygen dynamics of Tampa Bay during summer of 1991. Estuarine and Coastal Modeling, 5, 35-58. Yeh, W.W-G. (1986). Review of parameter identification procedures in groundwater hydrology: The inverse problem. Water Resources Research, 22, 95-108. Zheng, L., Chen, C., and Zhang, F.Y. (2004) Development of water quality model in the Satilla River estuary, Georgia. Ecological Modelling 178, 457-482. Zimmerman, M.S., and Livingst on, R.J. (1976). Seasonality and physio-chemical ranges of bentic macrophytes from a north fl orida estuary (Apalachee Bay). Contributions in Marine Science, 20, 33-45. Zison, S., Mills, W.B., Deimer, D., and Chen, C.W. (1978). Rates, constants and kinetic formulations in surfac e water quality modeling. Prepared by Tetra Tech, Inc., Lafayette, CA for Environmental Research Laboratory, U.S. Environmental Protection Agency, Athen, GA. EP A-600/3-78-105:335

PAGE 328

305 BIOGRAPHICAL SKETCH Taeyun Kim was born on the 5th of July in Pusan, the San Francisco of South Korea. He loved the beach and ocean. He received the Bachelor and Master of Engineering degrees in the Department of Ocean Engineering and Interdisciplinary Program of Ocean Industrial Engineering at Pukyoung National University in February, 1995 and 1998, respectively. In order to learn th e-state-of-art numeri cal model, he moved to the Gainesville, FL, where he started hi s Ph.D. studies in co astal and oceanographic program at the University of Flor ida. He earned a doctorate in 2007.